\ X C/) ^ ^ 5 > Nj^uiis^ 2 "^ > *Z CO Z . . . t/> /CH^SS^ ^ W ^ W c _ z! _ O xiO£iiV>/ Q X^vqc^ UTION^ NOliniliSNI “*NVINOSHimS S3iavyaiT LIBRARI ES^ SMITHSONIAN”* INSTITUTION ^NOlinil. ^oX p m rr, ZTSoTXx 2 m xT5£5qX o - m I m ' m ^ (/) — CO _ CO ;r /aan libraries Smithsonian institution NoiiniiiSNi nvinoshiiws S3iavaan l Z , CO Z T CO Z CO z < S. 2 X«x < \x s ^ 1 < X g li li) I i i CO z c/5 ••■' z CO 2 CO UTION NOlinillSNI NVINOSHimS S3iavaaiT libraries SMITHSONIAN INSTITUTION NOlinil. CO 2 \ ^ = CO X CO vXx cu xSvncoX CO ^ ^ IcJ CO o z; '•^iVA22>' o “■ o 2 ■ -I 2 _l Z -J (/aan libraries Smithsonian institution NoiiniiiSNi nvinoshii/js S3iavaan librar ■ 03 X) > X3 m xg\pc^ r; ' m n^qIZs^X m 'Z ^ocs^ m CO _ CO X — CO X 2 CO UTION NOlinillSNI NviNosHims S3iavyaii libraries Smithsonian institution Noiinii CO z CO Z ..■■ CO Z V ^ X 2 '^' > MlOiixsgX 2 '*'' > to 2 CO ‘ Z CO *■ 2 vyan libraries Smithsonian institution NoiiniiiSNi nvinoshiiims saiavaan librar CO X CO — CO X Q O XOijit^tiX o N^O/ixs^;^ q X0\ pc^ 'UTION ^ NOlinillSNI “"NVINOSHimS S3iyvyan'~'LIBRARI ES^ SMITHSONIAN”* INSTITUTION ^NOlinil. Z r- z r- Z m ' X rn CO ■ — CO X CO /Han LIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHIIIMS S3iyvyai1 LIBRAR 2 , W ^ 2 , CO Z CO z , < ^ XoSTirrT^ C ^ S .^5^^ < ^ 2 ^ Z (/) 2 UTION NOlinillSNI NVINOSHIIIMS S3iyvyaiT LIBRARIES SMITHSONIAN INSTITUTION NOlinilj 2 C/> ^ ^ HSONIAN INSTITUTION NOliniliSNI NVINOSHilWS S3ldVHa io — 1 y^AW :;p. XP\ dc !/ 2 C/) 2 3 LIBRARIES SMITHSONIAN INSTITI CO O ''i2L£S^ — ' w ^ 2 josHilws S3iavaan libraries Smithsonian institution NoiinmsNi NviNosHims S3iav - ^ O H O -I 2 _j 2 mSONIAN INSTITUTION NOliniliSNI NVINOSHillMS SBIHVaail LIBRARIES SMITHSONIAN INSTIT •“ 2 r~ 2 *“ z f“ 2 CO o 2 > X^oxiis^ 2 ■^ > 2 CO‘2 CO*^2 to FHSONIAN INSTITUTION NOliniliSNI NVINOSHilWS S3iavaail LIBRARIES SMITHSONIAN _ INSTIT 5 ^ 2 Z It! ^ to XTSsoKTX CU to q: q) < cr V'"’'' Q ^ Q ^ '■v'OlIiSjV' Q O 'J0SHimS^S3 I ava a I1~'lI B RAR I ES^ SMITHSONIAN"* INSTITUTION ^ NOliniliSNI "‘NVINOSHimS S3 I a\ 2 (- 2 1“ 2 [I ' ^ - ^cy — lvA$v^>z tn X^!yos'£^ £ m CO £ CO THSONIAN INSTITUTION NOliniliSNI NVINOSHIUMS S3iavaai3 L 2 T CO 2 CO 2 O z o -t- /yJA>my t § A / ■'^^ t 5 2 XlAwsvi^ > CO ••■■ 2 CO 2 to NOSHiiws S3 lava an libraries smithsonian institution NoiiniiiSNi nvinoshiiiais S3ia I*"' t '0 ^1- i r Ki i f ' &• . t\ VOLUME 17 Palaeontology 1974 PUBLISHED BY THE PALAEONTOLOGICAL ASSOCIATION LONDON Dates of publication of parts of Volume 17 Part 1, pp. 1-202, pis. 1-27 Part 2, pp. 203-440, pis. 28-59 Part 3, pp. 441-728, pis. 60-105 Part 4, pp. 729-971, pis. 106-126 April 1974 September 1974 October 1974 November 1974 This volume edited by isles strachan, roland goldring, j. d. Hudson, D. J. GOBBETT, L. R. M. COCKS, C. P. HUGHES AND J. W. MURRAY Dates of publication o/ Special Papers in Palaeontology Special Paper No. 13 October 1974 Special Paper No. 14 December 1974 (£) The Palaeontological Association, 1974 Printed in Great Britain at the University Press, Oxford by Vivian Ridler Printer to the University CONTENTS Part Adams, C. G. and Bedford, D. J. Foraminiferal biostratigraphy of the Oligocene- Miocene limestones of Christmas Island (Indian Ocean) 3 Alvin, K. L. Leaf anatomy of Weichselia based on fusainized material 3 Andrews, H. N., Gensel, P. G. and Forbes, W. FI. An apparently heterosporous plant from the Middle Devonian of New Brunswick 2 Bassett, M. G. Review of the stratigraphy of the Wenlock Series in the Welsh Borderland and South Wales 4 Bedford, D. J. See Adams, C. G. Brunton, C. H. C. and Champion, C. A Lower Carboniferous brachiopod fauna from the Manifold Valley, Staffordshire 4 Burton, C. J. and Eldredge, N. Two new subspecies of Phacops rana [Trilobita] from the Middle Devonian of North-West Africa 2 Calef, C. E. and Hancock, N. J. Wenlock and Ludlow marine communities in Wales and the Welsh Borderland 4 Campbell, K. S. W. See Holloway, D. J. Chaloner, W. G., Mensah, M. K. and Crane, M. D. Non-vascular land plants from the Devonian of Ghana 4 Champion, C. See Brunton, C. H. C. Cholich, T. del C. See Whatley, R. C. Crane, M. D. See Chaloner, W. G. Daily, B. See Jago, J. B. Davey, R. j. and Verdier, J.-P. Dinoflagellate cysts from the Aptian type sections at Gargas and La Bedoule, France 3 Edwards, D. and Richardson, J. B, Lower Devonian (Dittonian) plants from the Welsh Borderland 2 Eldredge, N. See Burton, C. J. Fedorowski, j. The Upper Palaeozoic tetracoral genera Lophophyllidium and Timorphyllum 3 Finch, E. M. An improved method of mounting palaeontological specimens for SEM examination 2 Forbes, W. H. See Andrews, H. N. Fortey, R. a. a new pelagic trilobite from the Ordovician of Spitsbergen, Ireland, and Utah 1 Fursich, F. T. See Palmer, T. J. Gensel, P. G. See Andrews, H. N. Goldring, R. and Kazmierczak, J. Ecological succession in intraformational hardground formation 4 Hancock, N. J. See Calef, C. E. Harris, T. M. Williamsoniella lignieri: its pollen and the compression of spherical pollen grains 1 Holloway, D. J. and Campbell, K. S. W. The Silurian trilobite Onycopyge Woodward 2 Howarth, M. K. The Lower Cretaceous ammonite genera proposed by C. Jacob in 1907 3 Hurst, J. M. Selective epizoan encrustation of some Silurian brachiopods from Gotland 2 Hurst, J. M. See Fursich, F. T. Jago, J. B. and Daily, B. The trilobite Clavagnostus Howell from the Cambrian of Tasmania 1 Jell, J. S. See Pickett, J. W. Jenkins, R. J. F. A new giant penguin from the Eocene of Australia 2 A new spider-crab from the Miocene of New Zealand 4 Jenkins, T. B. H. Lower Carboniferous conodont biostratigraphy of New South Wales 4 Page 475 587 387 745 811 349 779 925 623 311 441 431 111 949 125 409 727 423 95 291 869 909 IV CONTENTS Part Page Johnson, J. G. Affinity of Dayiacean brachiopods 2 437 Kazmierczak, J. Lower Cretaceous sclerosponge from the Slovakian Tatra Mountains 2 341 Kazmierczak, j. See Goldring, R. Krassilov, V. A. Podocarpus from the Upper Cretaceous of Eastern Asia and its bearing on the theory of conifer evolution 2 365 Laing, j. F. a specimen location technique for SEM strew mounts 2 435 Lauritzen, (}). New microfossils from the Silurian (Llandovery, stage 6) of the Oslo Region, Norway 3 707 Levinton, j. S. Trophic group and evolution in bivalve molluscs 3 579 Lindstrom, M. The conodont apparatus as a food-gathering mechanism 4 729 Lord, A. Ostracods from the Domerian and Toarcian of England 3 599 Mackinnon, D. I. and Williams, A. Shell structure of Terebratulid brachiopods 1 179 McLean, D. M. Two new Paleocene dinoflagellates from Virginia and Maryland 1 65 McLean, R. A. Chonophyllinid corals from the Silurian of New South Wales 3 655 Maglio, V. j. a new proboscidean from the late Miocene of Kenya 3 699 Matthews, S. C. and Thomas, J. M. Lower Carboniferous conodont faunas from north- east Devonshire 2 371 Mensah, M. K. See Chaloner, W. G. Moody, R. T. J. See Walker, C. A. Owens, R. M. The affinities of the trilobite genus Scharyia, with a description of two new species 3 685 Palmer, C. P. A new genus of Jurassic bivalve mollusc ancestral to Globocardium 1 165 Palmer, T. J. and Fursich, F. T. The ecology of a Middle Jurassic hardground and crevice fauna 3 507 Paton, R. L. Capitosauroid labyrinthodonts from the Trias of England 2 253 Lower Permian pelycosaurs from the English Midlands 3 541 Penn, I. E. The production of stratigraphical range-diagrams by automatic methods 3 553 Pickett, J. W. and Jell, J. S. The Australasian tabulate coral genus Hattonia 3 715 Price, D. Trilobites from the Sholeshook Limestone (Ashgill) of South Wales 4 841 Rasul, S. M. The Lower Palaeozoic acritarchs Priscogalea and Cymatiogalea 1 41 Richardson, J. B. See Edwards, D. Riding, R. The Devonian genus Keega (Algae) reinterpreted as a stromatoporoid basal layer 3 565 Riva, j. a revision of some Ordovician graptolites of eastern North America 1 1 Sadler, P, M. Trilobites from the Gorran Quartzites, Ordovician of south Cornwall 1 71 Tavener-Smith, R. Early growth stages in rhabdomesoid bryozoans from the Lower Carboniferous of Hook Head, Ireland 1 149 Thomas, B. A. The lepidodendroid stoma 3 525 Thomas, J. M. See Matthews, S. C. Verdier, J.-P. See Davey, R. J. Waldman, M. Megalosaurids from the Bajocian (Middle Jurassic) of Dorset 2 325 Walker, C. A. and Moody, R. T. J. A new trionichid turtle from the Lower Eocene of Kent 4 901 Webby, B. D. Upper Ordovician trilobites from central New South Wales 2 203 Whatley, R. C. and Cholich, T. del C. A new Quaternary Ostracod genus from Argentina 3 669 Williams, A. See Mackinnin, D. I. Palaeontology VOLUME 17 • PARTI APRIL 1974 Published by The Palaeontological Association London Price £5 THE PALAEONTOLOGICAL ASSOCIATION The Association publishes Palaeontology and Special Papers in Palaeontology. Details of member- ship and subscription rates may be found inside the back cover. PALAEONTOLOGY The journal Palaeontology is devoted to the publication of papers on all aspects of palaeontology. Review articles are particularly welcome, and short papers can often be published rapidly. A high standard of illustration is a feature of the journal. Four parts are published each year and are sent free to all members of the Association. SPECIAL PAPERS IN PALAEONTOLOGY This is a series of substantial separate works. Members may subscribe to the Series; alternatively. Ordinary and Student members only may obtain individual copies at reduced rates. The following Special Papers are available : 1. (for 1967); Miospores in the Coal Seams of the Carboniferous of Great Britain, by a. h. v. smith and M. A. BUTTERWORTH. 324 pp., 72 text-figs., 27 plates. Price £8 (U.S. $22.00), post free. 2. (for 1968): Evolution of the Shell Structure of Articulate Brachiopods, by k. williams. 55 pp., 27 text- figs., 24 plates. Price £5 (U.S. $13.00). 3. (for 1968): Upper Maestrichtian Radiolaria of California, by Helen p. foreman. 82 pp., 8 plates. Price £3 (U.S. £8.00). 4. (for 1969): Lower Turonian Ammonites from Israel, by R. freund and m. raab. 83 pp., 15 text-figs., 10 plates. Price £3 (U.S. $8.00). 5. (for 1969): Chitinozoa from the Ordovician Viola and Fernvale Limestones of the Arbuckle Moun- tains, Oklahoma, by w. a. m. jenkins. 44 pp., 10 text-figs., 9 plates. Price £2 (U.S. $5.00). 6. (for 1969): Ammonoidea from the Mata Series (Santonian-Maastrichtian) of New Zealand, by R. A. HENDERSON. 82 pp., 13 text-figs., 15 plates. Price £3 (U.S. $8.00). 7. (for 1970): Shell Structure of the Craniacea and other Calcareous Inarticulate Brachiopoda, by A. WILLIAMS and a. d. WRIGHT. 51 pp., 17 text-figs., 15 plates. Price £1-50 (U.S. $4.00). 8. (for 1970); Cenomanian Ammonites from Southern England, by w. j. Kennedy. 272 pp., 5 tables, 64 plates. Price £8 (U.S. $22.00). 9. (for 1971): Fish from the Freshwater Lower Cretaceous of Victoria, Australia, with Comments on the Palaeo-environment, by m. waldman. 130 pp., 37 text-figs., 18 plates. Price £5 (U.S. $13.00). 10. (for 1971): Upper Cretaceous Ostracoda from the Carnarvon Basin, Western Australia, (iv R. h. bate. 148 pp., 43 text-figs., 27 plates. Price £5 (U.S. $13.00). 11. (for 1972): Stromatolites and the Biostratigraphy of the Australian Precambrian and Cambrian, by m. r. WALTER. 268 pp., 55 text-figs., 34 plates. Price £10 (U.S. $26.00). 12. (for 1973): Organisms and Continents through Time. A Symposium Volume of 23 papers edited by N. F. HUGHES. 340 pp., 132 text-figs. Price £10 (U.S. $26.00) (published with the Systematics Asso- ciation), 13. (for 1974); Graptolite studies in honour of O. M. B. Bulman. Edited by R. b. rickards, d. e. jackson, and C. P. HUGHES. In preparation. SUBMISSION OF PAPERS Typescripts on all aspects of palaeontology and stratigraphical palaeontology are invited. They should conform in style to those already published in this journal, and should be sent to The Secretary, P.A. Publications Committee, Department of Geology, Sedgwick Museum, Downing Street, Cambridge, CB2 3EQ, England, who will supply detailed instructions for authors on request (these are published in Palaeontology, 15, pp. 676-681). (0) The Palaeontological Association, 1973 Cover: Dactylioceras commune (J. Sowerby), Upper Lias, Jurassic, Whitby, Yorks. In collection of Professor J. E. Hemingway. The ‘head' was carved by a Whitby jet worker to illustrate the traditional story of St. Hilda of Whitby (614 80), who turned snakes into stones by the power of prayer. i t •i{vviis:k::.'Viji »i*ii (iii i ■: ^ I ■r* I' 4 '■U I, 5 ■ l» •■■•I.*'' A REVISION OF SOME ORDOVICIAN GRAPTOLITES OF EASTERN NORTH AMERICA by JOHN RIVA Abstract. In a study of part of Hall’s type material, lectotypes are proposed for CHmacograptus hicornis (Hall) and Climacograptus spinifents Ruedemann; the variation in the proximal end of the latter is amply illustrated and dis- cussed; it is also suggested that the latter may be the slightly modified descendant of the former. Climacograptus parvus Hall ( = C. phyUophorus Gurley) is considered to be conspecific with Pseudocliniacograptus scharenhergi (Lapworth) and Orthograptus amplexicaulis (Hall) is revised and the question of its priority over Orthograptus truncatus (Lapworth) is considered. In a review of part of Ruedemann’s type material, the types of Climacograptus eximius Ruedemann are interpreted as consisting of deformed specimens of Pseudocliniacograptus modestus (Ruedemann) and of Climacograptus brevis strictus (Ruedemann); Climacograptus tenuis Ruedemann is considered to be conspecific with Climacograptus pvgmaeus Ruedemann; Climacograptus lorrainensis Ruedemann is assigned to Glyptograptus, and the name Climaco- graptus putillus (W-a\\) is temporarily restricted to the types in the American Museum of Natural History. Corynoides calicularis var. americana Ruedemann is revived and raised to specific rank as Corynoides americanus- the types of Corynoides comma Ruedemann are interpreted as deformed individuals of Corynoides americanus or C. calicularis sensu lato. The systematic treatment is preceded by a review of the Middle-Upper Ordovician zonation hitherto recognized in the graptolite succession of eastern North America which is strongly provincial, containing faunas unknown outside this part of the continent. The zonation is extended downward with the addition of the Diplograptus multidens and Nemagraptus gracilis Zones and expanded with data gathered since it was first proposed (Riva 1969). This is a partial revision of a few of the graptolites first described by James Hall in volume 1 of the Paleontology of New York (1847) and of others described later by Ruedemann (1908, 1912, 1925, and 1947) from the Ordovician of New York; it includes some of the first graptolites recognized in North America. The Hall species discussed herein are all of widespread occurrence, but because of inadequate descrip- tions, figures, and, perhaps, revisions, they have been misunderstood by many specialists in this field. The Hall collections have remained untouched for many years; in fact they were last studied by Ruedemann (1908) at the beginning of this century. No holotypes or lectotypes have ever been selected; many type collections are mixed, consisting of specimens belonging to other species or contain specimens used by subsequent authors to erect new species without separating them from the original collections. The Ruedemann collections must also be approached with care, for his type material is often of a mixed nature and some of his species are really con- specific with, or variants of, other species. Others, again, have been considered synonymous with other species, but in reality they have been found to be fully valid and are redescribed. It should be kept in mind here that Ruedemann’s descriptions, especially his dimensions, do not always correspond to reality. This study has been expanded with the aid of collections from the type locality or from strata correlative with it. An effort has always been made to trace and recollect from the type locality except where it has been rendered inaccessible or it has dis- appeared under a highway. [Palaeontology, Vol. 17, Part 1, 1974, pp. 1-40, pis. 1-2.] 2 PALAEONTOLOGY, VOLUME 17 The systematic discussion is preceded by a brief discussion of the biostratigraphic framework hitherto recognized in the graptolite succession of the Middle and Upper Ordovician of eastern North America which is characterized in part by provincial faunas unlike those of other parts of North America and Europe. GRAPTOLITE BIOSTRATIGRAPHY OF THE MIDDLE AND UPPER ORDOVICIAN OF EASTERN NORTH AMERICA Regional studies have been carried out in the past ten years in order to construct a graptolite zonation which reflected the strongly provincial faunas of the Middle and Upper Ordovician of eastern North America (Riva 1968, 1969). These studies continue those begun long ago by Rudolf Ruedemann (1947, p. 52), but largely on the basis of the incomplete Ordovician of New York. The Taconic Orogeny began to affect the eastern part of North America in mid-Ordovician time causing the forma- tion of Appalachian land masses west of which endemic graptolite faunas developed. These faunas are strikingly dissimilar from those of the Pacific faunal province of western and south-western North America and those of northern Europe. The faunal succession and zonation previously proposed for the autochthonous Canajoharie- Utica Shales of eastern North America and the Upper Ordovician of Anticosti Island (Riva 1969) is here extended downward to include the D. multidens Zone, the exist- ence of which has recently been documented in Appalachian allochthonous units, and the widespread N. gracilis Zone, which is believed to correlate with the Porter- field Stage, or the Black River Group, in terms of stratigraphy. Bergstrom and Drahovzal (1972) have recently reported the discovery in the southern Appalachians of Llandeilo and Llanvirn faunas referable to the G. tere- tiuscidus and the D. murchisoni Zones. In 1972 the writer discovered a typical late Llanvirn fauna in the Stanbridge slates of southern Quebec and Vermont, and just now, 1973, he has identified Llanvirn graptolites in a collection from the Hamburg Klippe of south-eastern Pennsylvania. These discoveries represent a significant breakthrough in understanding the graptolite succession of eastern North America, for up to now Llanvirn and, possibly early Llandeilo, faunas were known only from western Newfoundland (Morris and Kay 1966; Erdtmann 1971) and scattered localities in the northern Appalachians. The graptolite zonation constructed by Berry (1960) on the basis of the graptolite succession of the Ordovician of Texas can be applied only with difficulty to the Ordovician outside the Ouachita Geosyncline. Some reasons are: (1) the Texan graptolite succession belongs to the Pacific faunal province and contains none of the Middle-Late Ordovician faunas endemic to north-eastern North America; (2) some extensive gaps seem to exist in the Texas succession, corresponding to most of the Llanvirn and Llandeilo and parts of the Caradoc and Ashgill in terms of the British succession; (3) graptolites are not common in the late Middle and Upper Ordovician of Texas as a careful examination of the data presented by Berry (op. cit.) shows; and (4) western graptolite successions are, in the writer’s experience (Riva 1970), much more varied and complex than those of Texas, suggesting the need of a separate biostratigraphic scale for that region. The Texas graptolite zones have been called the ‘North American Graptolite Zones’ by Berry (1968), who, however has since RIVA: ORDOVICIAN GRAPTOLITES 3 recognized the existence of faunal provincialism in north-eastern North America (1970^) and of a ‘scarcity of certain graptolites at some horizons’ in western suc- cessions (1971). North American graptolite successions cannot be simply expressed in terms of one ‘standard’ zonation, but a distinct biostratigraphic scale will have to be constructed in each major region in order to take fully into consideration its typical faunas and faunal provinces. Thus the task of erecting trustworthy zonal successions in the Ordovician of North America has but just begun. TABLE 1. Middle and Upper Ordovician graptolite zones of eastern North America with a suggested correlation with those of Britain AMERICAN SERIES and STAGES Eastern North America Texas (Berry, 1960) British Isles (Skevington, 1969) BRITISH SERIES Z < u C. prominens-elongatus No fauna hJ i> o Q O D. coniplauatus D. compkmatus D. anceps D. complanatus O X < oi OJ a, a. D C. manitoulinensis 1 1 O C. pygnuieus O. quadrimucronatus P. linearis — — - D C. spiniferits O. dntermedius' U o Z < U 2 > OJ ' C a o O. ruedemanni D. clingani Q < Di > o o cd 03 c U C. ainericanus No fauna u oi o tu _i Q Q Wilder- ness D. miiltidens D. nndtidens S Porter- field N. gracilis C. hiconds and N. gracilis N. gracilis Listed below are the graptolite faunas characterizing each of the zones shown in Table 1 together with the lithologic units containing them. These faunal lists are approximate as many forms are in need of revision. This zonation should not in any way be considered final or unchanging. It is only the biostratigraphic framework recognized up to the present time and will be modified, changed, further divided, or simplified as data accumulate. This zonation is correlated with the Texan zonation and a revised British zonation. 4 PALAEONTOLOGY, VOLUME 17 The N. gracilis Zone. This zone is universally recognized and constitutes one of the bases for international correlations. It contains a complex and varied fauna with horizontal didymograptids such as D. sagitticaulis, D. serratidus, D. subtemiis\ Dicellograptiis divaricatus, D. sextans and varieties; N. gracilis, N. exilis and varieties; Dicranograptus ramosus, D. spinifer, D. rectus, D. contortus, D. diapason, and D. furcatus', Orthograptus calcaratus and varieties; Pseudoclimacograptus scharenbergi, P. s. stenostoma, P. angulatus (large form), and P. modestus; Climacograptus bicornis, C. bicornis tridentatus, and C. brevis strictus; Reteograptus geinitzianus, Cryptograptus tricornis, Glossograptus ciliatus, Lasio^raptus pusillits, Lepto- graptus flaccidus trentonensis;Glyptograptus teretiusculus and G. euglyphus; Amplexograptus sp. ; Corynoides pristinus and C. calicularis, as well as Dictyonema and a few dendroids. This fauna occurs in the Mount Merino Chert and Shale, the Austin Glen Greywacke of New York, and in lithologically similar, if not identical, rocks of the northern Appalachians to the eastern end of Gaspe, in a thrust slice of the Cloridorme ‘Formation’ of Gaspe, in the lower Quebec City Formation, the lower Magog slates, the Stanbridge Forma- tion of southern Quebec, part of the Flamburg Klippe of south-eastern Pennsylvania, and part of the Athens Shale of the southern Appalachians. In Table 1, the bicornis and gracilis zones of Texas are shown as one zone because data presented by Berry ( 1 960) from the Woods Hollow Shale shows that these zones are really based on the same fauna. The D. multidens Zone. This zone contains a fauna similar to that of the D. nndtidens Zone of Britain and for this reason the zonal name is maintained. Collections from this zone contain some elements ranging up from the N. gracilis Zone such as Climacograptus bicornis, C. bicornis tridentatus, C. brevis strictus, Pseudoclimacograptus scharenbergi, P. modestus, Hallograptus nmcronatus, Reteograptus geinitzianus, varieties of Orthograptus calcaratus, Dicranograptus rectus, D. contortus, Glyptograptus teretiusculus and G. euglyphus, and a few Dicellograptus divaricatus but it lacks the didymograpti, nemagrapti, and most of the dicellograpti of that zone, as well as other less striking or common forms. New elements of this zone are: Dicranograptus nicholsoni, Diplograptus multidens, D. compactus, Cryptograptus insectiformis, and an abundance of Corynoides referable to C. calicularis s.l. This fauna was previously referred to as the ‘Magog’ fauna (Riva 1968). Graptolites of this zone occur through most of the Quebec City Formation, the upper part of the Magog, part of the Cloridorme ‘Formation’ of Gaspe; in New York they occur in parts of the Snake Hill Shale of Ruedemann, the ‘Mount Merino’ Formation of the southern part of that state (Offield 1967), and possibly in the upper Austin Glen Greywacke. The writer has recognized this fauna in collections from Middle Ordovician slates of eastern Pennsylvania, and Ruedemann (1947, pp. 79, 86) has reported it from the base of the Martinsburg Shale of Maryland. No section has yet been found showing the faunal passage between the multidens and the succeeding C. americanus Zone ; multidens Zone graptolites normally occur in allochthonous or Appalachian sequences and americanus Zone graptolites characterize autochthonous or parautochthonous sequences. An indica- tion as to the nature of this passage is given by the graptolites found at the very base of the Macasty Shale in the L.P.G.L. core from Anticosti Island (Riva 1969, pp. 534-537), the basal part of Canajoharie Shale near Amsterdam in the lower Mohawk Valley, New York, and from uncontrolled slices of Snake Hill Shale just above the mouth of the Mohawk River at Waterford, N.Y. These units uniformly yielded Lasiograptns harknessi, a new small Diplograptus, Orthograptus calcaratus hasilicus, O. amplexicaulis, Glyptograptus euglyphus, Corynoides cf. C. calicularis, and Climacograptus brevis-mohawkensis transients. These forms, all together, belong to neither the multidens nor the americanus Zones, although a few elements span both zones. A recent (late 1972) collection from a Snake Hill slice a few miles above the mouth of the Mohawk River in New York contained a new spinose Climacograptus, long, strongly curved Corynoides, and Orthograptus calcaratus hasilicus, all of which could represent a still lower level of this passage fauna. A solution of this problem awaits even more diagnostic collections. The C. americanus Zone. In this zone most elements of the N. gracilis and D. nndtidens Zones have dis- appeared with the exception of Corynoides calicularis s.l., which occurs in profusion, Climacograptus brevis strictus, Orthograptus calcaratus hasilicus, and an occasional Glyptograptus euglyphus. The thin and short Corynoides americanus occurs in great numbers near the base of the zone together with Lasiograptns harknessi, Orthograptus amplexicaulis, and Climacograptus brevis. New elements making their first appear- ence halfway up the zone are Orthograptus (piadrimucronatus micracanthus cornutus, Climacograptus caudatus, Cryptograptus in.sectiformis, and Neurograptus cf. margaritatus. Climacograptus mohawkensis (= C. minimus of Elies and Wood) and Orthograptus ruedemanni appear in profusion near the top of the zone. This fauna characterizes the lower Canajoharie of New York and Quebec, the lower Macasty Shale RIVA: ORDOVICIAN GRAPTOLITES 5 of Anticosti Island, part of the Cloridorme ‘Formation' of Gaspe, part of the flysch and wildflysch sequence (St. Germain Complex) in front of the Appalachian allochthon in Quebec, and part of the Snake Hill Shale of Ruedemann in New York, and marks the beginning of provincialism in eastern North America connected with the events accompanying the Taconic Orogeny and the rise of Appalachian lands. This provincialism is more fully developed in the C. spiniferus Zone where European and Pacific forms appear to be in a minority. The O. ruedemanni Zone. [The name O. niedemanni replaces the name C. minimus previously given to this zone (Riva 1969) since Strachan (1969, pp. 191-193) has shown that the latter name was originally given to a Silurian graptolite.] The fauna of this zone is distinguished from that of the C. americanus Zone by the absence of all Corynoides and the almost exclusive development of two small graptolites, O. ruedemanni and C. mohawkensis, often in synrhabdosomes. Orthograptus quadrinnicronatus micracanthus, Crypto- graptus insectiformis, and O. amplexicaidis pass through this zone to the C. spiniferus Zone and succeeding zones; Neurograptus margaritatus occurs sparingly. This fauna characterizes upper Canajoharie Shales, parts of the Snake Hill Shale of Ruedemann, the Cloridorme of Gaspe, and the Macasty Shale of Anticosti Island. The C. spiniferus Zone. The fauna of the ‘true Utica’ of Ruedemann as modified by Riva (1969), marks the beginning of the C. spiniferus Zone. This fauna contains great numbers of O. quadrimucronatus micra- canthiis, O. amplexicaidis in small numbers, Cryptograptus insect ifonnis, C. caudatus, and Dicranograptus ramosus (the last appearance for these two), as well as some new and diagnostic forms: Climacograptus spiniferus, C. typicalis (and its predecessor), Orthoretiolites, a new species of Glyptograpliis, Diplograptiis sp. (at the base), and Dicranograptus cf. D. nicholsoni minor. Diplograptus ingens, a form hitherto believed to be restricted to the Pacific faunal province, has been identified by the writer from the lower Utica of the Mohawk Valley. This fauna occurs in the lower half of the ‘true Utica’, part of the Macasty Shale of Anti- costi Island, the flysch sequence of the St. Germain Complex, the Iberville Formation, part of the so- called Cloridorme ‘Formation’ of Gaspe, part of the Snake Hill Shale of Ruedemann, and the whole Snake Hill of Offield (1967), and its continuation in New Jersey, Pennsylvania, and Virginia (Ruedemann 1947, pp. 78-87; and unpublished collections in the New York State Museum). The C. pygmaeus Zone. Changes in Utica faunas are gradual. Midway through the Utica C. spiniferus and D. nicholsoni cf. minor disappear and a small form of C. typicalis, C. pygmaeus, appears in profusion and ranges up to the top of the Utica. This change is also accompanied by the arrival of Leptograptus flaccidus, a new Glyptograptus, and a new Orthoretiolites with an extremely long virgella. Corynoides makes its last appearance here. Orthograptus amplexicaidis (possibly a subspecies), the typical O. quadrimucronatus, and a small member of the quadrimucronatus group known as O. eucharis (often in synrhabdosomes) appear in great numbers near the top of the Utica together with two new members of the typicalis group : C. typicalis magnificus and C. typicalis posterns. Glyptograptus lorrainensis is first seen here and passes into the succeed- ing zone. Pleurograptus linearis has been recorded once from the upper Utica of the Mohawk Valley (Ruede- mann 1908). In North America P. linearis is also known from the Phi Kappa Formation of Idaho (Churkin 1963, p. 1620) and the Point Leamington Greywacke, Fortune Peninsula, Newfoundland (J. Helwig 1967). Both collections are at Columbia University. The C. manitoulinensis Zone. The typical elements of the Utica disappear with the cessation of the black shale sedimentation of the Utica; some members, however, persist into the grey siltstones and sandstones of the succeeding Lorraine Group. An impoverished fauna appears here with elements of the previous zone such as C. pygmaeus, C. typicalis posterns, O. amplexicaidis (a subspecies), O. eucharis, O quadri- mucronatus, L. flaccidus, and Cryptograptus insectiformis, all to disappear within the zone, G. lorrainensis which increases in size and continues on into the succeeding zones and some new elements such as Pseudo- cliniacograptus cf. P. clevensis, Diplograptus sp., Climacograptus manitoulinensis, C. scalar is miserahilis, a Dicellograptus, and rare dendroids. The arrival of the succeeding zone is heralded by the appearance of Orthograptus ahhreviatus, O. socialis, and Dicellograptus complanatus. This fauna is of restricted regional extent. Generally it is confined to the lower Lorraine Group of the St. Lawrence Lowlands of Quebec and Ontario and of the upper Mohawk Valley of New York; it also occurs in the upper Macasty Shale of Anticosti Island and Lake St. John and the lower part of the succeed- ing English Head Lormation, or the lower Vaureal Formation of Bolton (1972). P. cf. clevensis is identical to Climacograptus cf. extremus reported by Ruedemann (1947, pi. 72, figs. 20-21) from the lower Whitehead Formation (Lesperance 1968, pp. 813-814) at Mont Joli, Perce, Quebec. 6 PALAEONTOLOGY, VOLUME 17 Glyptograptus hudsoni described by Jackson (1971) from Southampton Island in the North-West Terri- tories and some collections from the same island and from Akpatok Island in Ungava Bay made by members of the Geological Survey of Canada and studied by this writer probably belong to the upper part of this zone. The Apkatok collections contained Amplexograptus inuiti. The D. complanatus Zone. The characteristic graptolites of this zone are: Dicellograptus complanatus, as described by Skoglund (1963, pp. 33-36), and transients to D. anceps, Orthograptus socialis, O. abbreviatus, C. scalaris miserabilis, a large Glyptograptus descended from G. lorrainensis, a Diplograptus, and rare dendroids. The zone is restricted to Anticosti Island and, possibly, the north-east part of the Gaspe peninsula. On Anticosti, it ranges through the upper English Head Formation (= lower Vaureal of Bolton 1972) to the base of the Vaureal Formation of Twenhofel. In Gaspe, Lesperance (1968) has reported graptolites in association with a Remipyga fauna of the Whitehead Formation. This writer has restudied these collections which contain Climacograptus scalaris miserabilis, C. normalis, C. iimotatus, and a large Glyptograptus identical to that from Anticosti (identified as D. (O.) rugosusvar. apiculatus on p. 815 of Lesperance’s paper). This fauna is possibly correlative with the D. complanatus fauna of Anticosti, but little else can be said in the absence of more diagnostic elements. The Climacograptus prominens-elongatus Zone. Most elements of the D. complanatus Zone disappear abruptly at the base of the Vaureal Formation of Twenhofel (or the upper Vaureal of Bolton) and in the succeeding 2000-foot interval to the top of the Ordovician, which includes the Vaureal and the Ellis Bay Formations, graptolites are rare. Orthograptus abbreviatus, the large Glyptograptus from the older zone, occur sparingly together with Climacograptus prominens-elongatus and rare dendroids. These formations contain an extensive shelly fauna, but may be regarded as practically devoid of graptolites. The Late Ordovician graptolite succession of Anticosti resembles that of the Harjuan Series of Sweden (Jaanusson 1963, pp. 131-135; Skoglund 1963). In Scania the D. complanatus fauna occurs in a narrow zone of the Tretaspis beds (Jerrestad Stage) and it is succeeded by the Dalmanitina beds (Tommarp Stage) which are poorly graptolitiferous. The middle and upper Tretaspis beds are correlated with the lower and middle Ashgill and the Dalmanitina beds with the upper Ashgill of Britain. The interval represented by the C. prominens-elongatus Zone may be interpreted to correspond to the Dalmanitina beds and the barren interval between the D. anceps and the G. persculptus Zone of the Moffat region in Scotland and the central zone of Wales (Lawson 1971). The stratigraphic succession on Anticosti certainly shows that a long interval there separates the top of the D. complanatus Zone from the base of the Silurian. Additional light on this problem is shed by a graptolite collection reported by Lesperance (1968, p. 816) from the Dalmanitina beds of the Whitehead Formation in Gaspe. The writer has restudied this collection and identified Climacograptus rectangularis-medius transients, and fragments of Orthograptus. SYSTEMATIC DESCRIPTION Suborder diplograptina Lapworth 1880, emend. Bulman, 1963 Family diplograptidae Lapworth, 1873 Genus climacograptus Hall, 1865 Climacograptus hicoruis (Hall) Plate 1, figs. 13, 5-7; text-figs. \a-h 1847 Graptolithus hicornis Hall, pp. 268-269, pi. 73, figs. 2c-d‘l, 2f-h, 4 (non figs. 2a-b). 1865 Climacograptus hicornis (Hall); Hall, pp. Ill 112, pi. A, figs. 13-17. 1908 Climacograptus hicornis (Hall); Ruedemann, pp. 80-85, pi. A, text-figs. 12-17; pp. 433-437, text-fig. 404 (non text-fig. 405), pi. 28, figs. 24, 25 (non fig. 26). 1947 Climacograptus hicornis (Hall); Ruedemann, p. 425, pi. 72, figs. 45, 46, 49-52 (non figs. 44, 47, 48). 1947 Climacograptus hicornis (Hall); Bulman, pp. 59-62, pi. 9, figs. 10-13. 1960 Climacograptus hicornis (Hall); Berry, p. 79, pi. 16, figs. 10, 11, pi. 19, fig. 4. 1963 Climacograptus hicornis (Hall); Ross and Berry, pp. 117 119, pi. 8, figs. 4-6, 9. No attempt is made here to list synonymies from outside North America. RIVA: ORDOVICIAN GRAPTOLITES 7 The original material of Graptolithus hicornis in the American Museum of Natural History consists of specimens of two distinct but superficially similar forms: Climacogniptus hicornis (Hall) and C. spinifenis Ruedemann. Hall’s ( 1847, pp. 268, 269) original description of G. hicornis broadly covered rhabdosomes with ‘obtuse teeth’, bearing ‘two diverging forks at their base’, which are ‘sometimes thickened or expanded’, as well as rhabdosomes characterized by ‘mucronate’ thecal apertures, or flanges covering the thecal excavations. The specimens with ‘mucronate’ apertures came from Cincinnati, Ohio, and were subsequently described by Hall (1865, pi. A, figs. 1-9) as Climacograptus typicalis. The originals of the forms now known as C. hicornis came from the classic graptolite locality in ‘Hudson River beds’ (now Austin Glen Greywacke) on the Normans Kill at Kenwood, south of Albany, N.Y., and those of C. spiniferus from the lower Utica Shale at Ballston Spa, N. Y. All these specimens were listed by Whitfield and Hovey (1898, p. 18) as part of the genotype of Climacograptus and ‘species types’ of C. hicornis. Specimens of C. hicornis constitute most of the originals of G. hicornis, but Hall’s first figure of this species (1847, pi. 73, figs. 2a-b) was based on a C. spiniferus. This specimen bears a label with 'Graptolites hicornis, Ballston Spa, Sar. Co.’ Apparently Hall did not recognize the diflference in the type and origin of the ‘diverging forks’ or spines of the specimens before him when he erected G. hicornis. In C. spiniferus one spine grows out from the first thecae, th H, and the other is simply the down and outward extension of a well developed virgella or, perhaps, even the sicula; in C. hicornis a spine grows out from each of the first thecae, th H and th H. The presence or absence of a membrane on the basal spines is not in itself significant in differentiating between the two forms, for it will be seen further on that basal membranes or discs, though common to mature rhabdosomes of C. hicornis, occur also in C. spiniferus. In erecting the genus Climacograptus, Hall (1865, pp. Ill, 1 12, pi. A, figs. 13-17) refigured only original specimens from the graptolite locality on the Normans Kill as characteristic representatives of this genus, and these figures, well executed although somewhat stylized, were taken by subsequent workers as diagnostic of C. hicornis. Accordingly, the writer proposes as lectotype of C. hicornis and genolectotype of Climacograptus the first specimen (A.M.N.H. 1030a) figured by Hall (1865, pi. A, fig. 13) in erecting the genus Climacograptus. This specimen was part of the original collection of G. hicornis and was listed by Whitfield and Hovey ( 1 898, p. 1 8) as part of the genotype specimens of Climacograptus and ‘species types’ of C. hicornis. [Whitfield was then a curator of the A.M.N.H. Prior to holding that position he had been for many years Hall’s draughtsman and as such drew the 1865 figures of C. hicornis.] The specimen of Hall’s first figure (1847, pi. 73, figs. 2a-h) of G. hicornis {A.M.N.H. 1041/5) is proposed here, instead, as the lectotype of C. spiniferus. It was on this specimen that Ruedemann (1908, pp. 411, 412), upon discovering the confusion in the original material of G. hicornis, founded mut. spinifer which later he (1912, p. 84) recognized as a distinct species and named C. spiniferus. Proposed Lectotype. A.M.N.H. 1030a, from black shale in the lower part of the Austin Glen Greywacke, left bank of Normans Kill, just above an abandoned bridge at Kenwood, N.Y. For exact location, see the geologic map accompanying N.Y. State Museum Bull. 285, 1930. Other material. A.M.N.H. 1034a, on slab bearing type of Graptolithus mucronatus (Hall 1847, pi. 73, fig. Ic); all figured materia! of Graptolithus hicornis catalogued as A.M.N.H. 1041/1, 1041/5; specimens from Hall’s paratype collections in the American Museum of Natural History catalogued as 1036/5, and new collections from the type locality at Kenwood, N.Y. Most specimens are pyritized, and several are deformed or distorted to various degrees. Description. The growth and morphology of the proximal end of this species have been described to some extent by Bulman (1947) on the basis of isolated Scottish material from the Laggan Burn section; the following discussion is concerned only with the type material and new specimens from the type locality and it is intended to accompany the selection of a lectotype and help differentiate this species from C. spiniferus Ruedemann, its descendant of the lower Utica Shale. Rhabdosome long, attaining 10 cm and more in length, gradually widening from an initial width of 0‘7-0-9 mm at the level of th U-U to 1-8-2-3 mm, two cm from the proximal end, and to as much as 2-5-3 0 mm in distal parts of flattened individuals five cm or more in length. Thecae number 5-7 in the first 5 mm of the rhabdosome gradually increasing in size so as to number only 8-9 in 10 mm distally in mature individuals. Thecal excavations are wide and semicircular in immature rhabdosomes (PI. 1, fig. 6; text-fig. la), tending to become narrower and shorter in mature indivi- duals ; apertures occupy from one-quarter to three-eighths the width of the rhabdosome and are surrounded by a well-developed selvage (PI. 1 , fig. 7 ; text-fig. 1 ). The geniculum PALAEONTOLOGY, VOLUME 17 is sharply angular, the supragenicular wall generally parallel to the axis of the rhabdosome, the more so in flattened, mature individuals (PL 1, figs. 1,7; text-fig. \b) than in immature rhabdosomes (PI. 1, fig. 6; text-fig. la). A conspicuous thecal spine grows out and downward from the first two thecae and may attain a considerable length in mature individuals (a little more than 5 mm in the lectotype) (text-fig. 1^). In mature rhabdosomes the spines may be enclosed in a membrane which may extend up to the proximal thecae. The rhabdosome is strongly septate. A thin virgula may extend distally for a short distance in some specimens (PI. 1, fig. 7; text-fig. lu). The sicula is small, ranging from 0-72 to 1-3 mm in length (average for 1 1 siculae is 0-95 mm), and mostly exposed on the obverse side of the rhabdosome (text-fig. la). The virgella is short, being from 0-27 to 045 mm in length (the most common length is 0-36 mm) and pointing toward the theca^ series. Theca P originates high on the sicula, grows down toward the sicular aperture, then curves outward and gently upwards until the aperture faces outwards, almost parallel with the side of the rhabdosome (text-fig. la). The apertural excavation is shallow. Theca F grows farther upwards than th P and its aperture is set deeper into the rhabdosome, facing more distally than ventrally (PI. 1, fig. 6; text-fig. la). The succeeding thecae all grow upwards, overlapping each other by one-third. The septum begins at the th 2^-2^ level near the base of th 3^ and continues on to the end of the rhabdosome, leaving a deep septal furrow. The thecal spines are given off at the lower corners of the rhabdosome. The spine of th V is flush with the aperture, but that of th P grows slightly below the aperture (PI. 1, fig. 6; text-fig. la). A study presently being carried out on several hundred topotypes indicates that these spines grow constantly through the formation of the rhabdosome, being short, thin, and wiry in immature rhabdosomes with less than five pairs of thecae (length less 1 mm, width 0-07 mm and less) and long and thick in EXPLANATION OF PLATE 1 Figs. 1-3, 5-7. Climacograptus bicornis (Hall). 1. Proposed lectotype, A.M.N.H. 1030a, figured by Hall { 1 865, pi. A, fig. 13). Only the external mould now remains, although the rhabdosome was still preserved in 1962. Obverse view of mature specimen, distally broken. x5. 2, 3. Paratype, A.M.N.H. 1036/5a not previously figured. Mature rhabdosome with well-developed proximal spines, thickened virgella, and showing partly deformed supragenicular walls of proximal thecae. This specimen was recently recovered from the A.M.N.H. general collections. Its counterpart, A.M.N.H. 1041/lb, labelled ‘type’ of C. bicornis, is not figured because of poor and fragmentary preservation. x5 and x2. 5. Paratype, on slab bearing a type of Graptolithus miicronatus (Hall), A.M.N.H. 1034a. Reverse view showing basal spines free of membranes, slightly thickened virgella, and square thecal outlines, x 6. 6. Paratype, A.M.N.H. 1041c, not previously figured. Reverse view of immature rhabdosome showing charac- teristically thin basal spines, thin virgella, round thecal apertures surrounded by strong selvage, and virgula prolonged distally. Thecal apertures of th U and F are clearly visible. xlO. 7. Paratype, A.M.N.H. 1031 /la, not previously figured. Mature rhabdosome in obverse view, with strongly developed basal spines and virgella free of membrane, narrow thecal aperture enclosed by a strong selvage. Aperture of th 1 ' visible, x 4. Figs. 4, 8. Climacograptus spinijerus Ruedemann. Proposed lectotype, A.M.N.H. 1041/5a, and Hall’s (1847, pi. 73, figs. 2a-b) first figure of Graptolithus bicornis. 4, complete specimen, x3T2. 8, proximal end with thin and symmetrically arranged proximal spines, and with a tiny membrane covering the virgellar spine. X 8. The specimens of C. bicornis are from the Austin Glen Greywacke on the Normans Kill at Kenwood, N.Y. The lectotype of C. spinijerus is from the lower Utica Shale at Ballston Spa, N.Y. PLATE 1 RIVA, Climacograptus spp. 10 PALAEONTOLOGY, VOLUME 17 TEXT-FIG. 1. Proximal end of Clinuicugraptus bicornis (Hall) from Austin Glen Greywacke at Kenwood, N.Y. a, Topotype. Immature rhabdosome in moderate relief, obverse side showing exposed sicula, virgella, mode of growth of th U and th H, position of their apertures and point of growth of thecal spines, and point of origin of septum; x 19. b. Proximal end of lectotype, A.M.N.H. 1030a, showing size of proximal spine, virgella enclosed by membrane, and deformed apertures of proximal thecae. Obverse view. Drawn from actual specimen in 1962; x 8. mature rhabdosomes (length 5 mm and more, width 0-4 mm at the point of origin) (compare text-figs. \a and l/t). No ‘basal membranes’ were noted on immature rhabdosomes with less than 15 pairs of thecae. A slight bulge, however, begins to form on top or around each spine in individuals with 16 pairs of thecae and more. This bulge may remain small and insignificant or grow out to both ends of the spines, surrounding the virgella, and in some cases, extend up on the proximal end of the rhabdosome as far as the sixth pair of thecae, blocking oflT thecal apertures (as seen in profile at least). Specimens with ‘basal membranes’ or discs have been separated as var. peltifer by Lapworth (1876) and others with apparently even more peculiar membranes as var. signum by Ruedemann (1908), but more realistically they should be regarded as mature rhabdosomes of C. bicornis which have developed basal membranes, the function of which is purely speculative. Var. signnm does not seem to be characterized by distinct membranes; a review of its type material suggests that the peculiarities illustrated by Ruedemann (1908, pi. A, facing p. 82) are due to torn membranes, failure to remove matrix around the basal discs, deformation, and pre- servation. Not all mature rhabdosomes of C. bicornis bear basal membranes. The lectotype (PI. 1, fig. 1 ; text-fig. 16) has only a tiny membrane around the virgella, and the other mature specimens from the Hall collections illustrated on Plate 1 (figs. 2, 3, 5, 7) bear none. Specimens devoid of basal membranes are figured in this RIVA: ORDOVICIAN GRAPTOLITES 11 paper in preference to others only in order to be able to show clearly the differences between C. bicornis and C. spiniferus. A few immature individuals of the topotype collection now under study bear a long virgella (more than 100 mm long) longer than the basal spines. Rhabdosomes with long virgellas have been separated as var. tridentatm by Lapworth (1876), and this may well be a valid designation at the subspecific level because a long virgella is already present in immature individuals. Basal membranes may enclose both the basal spines and the long virgella of subspecies tridentatus or may be reduced or entirely absent. There are all variations between the two extremes. Remarks. The lectotype and one other mature specimen in the Hall collections (PI. 1, figs. 1, 2, 3) present a deformed proximal end: the apertures of the first two or three pairs of thecae are closed off by what appears to be the collapse of the infrageni- cular wall of the next succeeding theca, the supragenicular wall having become mis- shapen in the process. In the lectotype the aperture of th V is also entirely obscured by what could be secondary deposits. The rhabdosome of Plate 1, figs. 2, 3, bears a short virgella also conspicuously thickened by secondary deposits. A study of Ruedemann’s collections has revealed an interesting detail : the rhabdo- some repeatedly figured by Ruedemann (1908, text-fig. 405, pi. 28, fig. 26; 1947, pi. 72, figs. 44, 47, 48) as a typical C. bicornis but with weak proximal spines and ‘peculiar, deeply notched’ thecal apertures is in reality an Amplexograptus. Individuals belonging to this genus have been collected by this writer at the Normans Kill locality. The only species close to C. bicornis is C. spiniferus which is essentially a slightly modified descendant of C. bicornis in which th H has grown down and upwards making a full U-turn, and th H has grown outwards and upwards, losing its thecal spine. A spine in this position, however, is formed by the virgella growing down and outwards as a spine, symmetrical to that of th H (text-fig. 2a-h). In all other details the proximal ends of C. bicornis and C. spiniferus are similar, including the sicula fully exposed on the obverse side of the rhabdosome and the origin of the septal groove in the th 2^-2^ region. Stratigraphic and geographic occurrence. C. bicornis is a cosmopolitan species ranging through the N. gracilis and D. multidens Zones or their equivalents. In north-eastern North America it is restricted to units of the Appalachian sequences, such as the Mount Merino Chert and Shale, the Austin Glen Greywacke, the Quebec City Formation, the Beauceville slates, the Stanbridge slates, pan of the ‘Cloridorme Formation’ of Gaspe, the Walloomsac Formation, the Indian River Shale and Chert, and to blocks of wildflysch accumulation derived from the Appalachian sequences. Climacograptus spiniferus Ruedemann Plate 1, figs. 4, 8; text-figs. 2, 3, 4 1847 Graptolithus bicornis Hall, pp. 268-269, pi. 73, figs. 2a-b {non 2c-s). 1908 Climacograptus typicalis mut. spinifer Ruedemann, pp. 411-412, text-figs. 236, pi. 28, figs. 8-9. 1912 Climacograptus spiniferus Ruedemann, p. 84. 1947 Climacograptus spiniferus Ruedemann, p. 439, pi. 73, figs. 1-7. 1955 Climacograptus spiniferus Ruedemann; Clark and Strachan, pp. 692-693, text-figs. 2>cf f. 1963 Climacograptus spiniferus Ruedemann; Ross and Berry, p. 130, pi. 9, fig. 12. 1969 Climacograptus spiniferus Ruedemann; Riva, p. 521, text-figs. 'ik-p. 1971 Climacograptus spiniferus Ruedemann; Berry, p. 637, pi. 73, fig. 5. 12 PALAEONTOLOGY, VOLUME 17 Proposed Lectotype. A.M.N.H. 1041/5, bearing label with ‘'Graptolites biconiis, Ballston Spa, Sar. Co' (PI. 1, figs. 4, 8; text-figs. 4a, b). The specimen most likely came from an outcrop of basal Utica Shale on the left bank of an unnamed tributary of Kayaderosseras Creek in the southern part of Ballston Spa, N.Y. The outcrop is located just west of the ‘Old Iron Spring’. Material and localities. Topotype specimens from Ballston Spa, N.Y., in the collections of the P. Redpath, Museum, McGill University, Montreal (these specimens were presented to the University by James Hall at the opening of the Museum in 1882). Extensive collections from the lower Utica Shale, zone of C. spini- ferus, in New York and Quebec, the Macasty Shale of Anticosti, the Cloridorme Formation of Gaspe, the parautochthonous flysch of Quebec, and the Schenectady beds of New York. Canadian specimens figured in this paper are deposited in the type collections of the Geological Survey of Canada, numbered G.S.C. 31710 to 31729. Description. Rhabdosome long, attaining as much as 7 cm and more in length (excluding proximal spines), widening gradually from 0-70-0-85 mm at the aperture of th F to 2 mm, I to 2 cm from the proximal end, and to 2-2-2-8 mm maximum observed in flat, mature individuals (text-flgs. 4g, /). Thecae number 10-30 (average 1 1-12) in the first 10 mm of the rhabdosome, decreasing to 9-1 1 distally in individuals 4 cm and more in length. The apertural excavations are moderately deep and wide, about one-quarter to one-third the width of the rhabdosome and one-quarter to one- third the length of the free ventral of the thecae; they are generally semicircular in immature rhabdosomes, more adpressed in mature rhabdosomes, always strengthened by a well-developed selvage merging the apertural margin with the infragenicular wall. The supragenicular wall may be parallel or slightly inclined away from the axis of the rhabdosome. The proximal end characteristically bears a pair of spines, one of which is the thecal spine grown outwards from th l’^ and the other the virgella grown down and outwards, often symmetrically with the thecal spine. A membrane may enclose the proximal spines and exceptionally the virgella may bifurcate. The rhabdosome is septate. The development of the early stages of C. spiniferus does not differ essentially from that of C. dipkicanthus Bulman (1932, pp. 13-16). The sicula is almost fully exposed on the obverse side of the rhabdosome and tilted toward the th^ series (text- figs. 2b-h). The septum originates in the th 2^-2* region and continues on to the distal end, leaving a deep septal groove (text-figs. 2a, h, d, g, j). TEXT-FIG. 2. Proximal spines in C. spiniferus. a. Reverse view of immature rhabdosome, G.S.C. 31710, with thin, short, and asymmetrical basal spines; septal groove originates in th 2'-2^ region; from L.J. 2 core, depth 1800'; x 10. b. Obverse view of proximal end of mature rhabdosome, G.S.C. 31711, showing tilted, exposed sicula, symmetrical basal spines but of unequal length, and mode of growth of th H; from L.J. 2 core, depth 1680'; x 10. c, G.S.C. 31712, obverse view, with nearly symmetrical spines, and showing exposed sicula and mode of growth of th H; from L.J. 2 core, depth 1040' 1045'; X 10. d. Proximal end of mature rhabdosome, G.S.C. 31713, obverse side, with exceptionally developed but asymmetrical basal spines; from basal Utica Shale at Neuville, P.Q.; x 10. e, /, h. Broad rhabdosome, obverse view, G.S.C. 31714, with exceptionally thickened virgellar spine symmetrical to that of th 1‘; from lower Utica Shale at Neuville, P.Q.; respectively x2, x 5, x 10. g, Rhabdosome in low relief, obverse view, G.S.C. 31730, showing thickened virgellar spine symmetrical to that of th 1' ; from lower Utica Shale, Neuville, P.Q. ; x 10. /, /, Flattened immature rhabdosomes, G.S.C. 31715, 31716, with minute proximal spines of nearly equal length; respectively from L.J. core I, depth 1295' and L.J. core 2, depth 985'-990' ; both X 5. ^, /, Proximal ends of mature rhabdosomes, G.S.C. 31717, 31718, in obverse and subscalariform views; from L.J. core 2, depth 1770' and 1780'; both X 5. The Lozo-Joseph cores 1 and 2 are treated separately (Riva 1969). RIVA; ORDOVICIAN GRAPTOLITES 13 14 PALAEONTOLOGY, VOLUME 17 TEXT-FIG. 3. Proximal spines in C. spiniferus. a, b, G.S.C. 31719 and 31720, proximal ends of mature rhabdosomes with thin membranes overlying basal spines; from L.J. core 2, depth 1780' and 1785'; both x 10. c, G.S.C. 31721. proximal end of mature rhabdosome with well-developed basal membranes extending to the th P aperture; from L.J. core 2, depth 1080'; x 10. d, G.S.C. 31722, proximal end of mature rhabdosome with proximal spines enveloped in leaf-like membranes; lower Utica Shale, Portneuf, P.Q. ; coll, by T. H. Clark; x 10. ej\ G.S.C. 31723 and 31724, proximal ends of mature rhabdosomes showing bifurca- tion of virgella; respectively from L.J. core 2, depth 1780' and L.J. core 1, depth 1460'; both x 10. The Lozo-Joseph cores 1 and 2 are treated separately (Riva 1969). TEXT-FIG. 4. Climacograptiis spiniferus Ruedemann. a-c. Proposed lectotype, A.M.N.H. 1041/5 from basal Utica at Ballston Spa, N.Y., and Hall’s (1847, pi. 73, figs. 2a-h) first figure of Gmptolithus biconiis. u. Full rhabdosome showing the thin, symmetrically arranged proximal spines; x4-6; 6, detailed view of proximal end showing origin of proximal spines and the virgella covered by a residual membrane; X 10; c, the same; x 2. d-j. From the Lower Utica of Neuville, Quebec. d. Reverse view of rhabdosome, G.S.C. 31725, twisted distally to an oblique view; x 3. e. Typical oblique view of rhabdosome, G.S.C. 31726, with thickened virgella; x5. f. Reverse view of rhabdosome, G.S.C. 31727, showing septal groove originating in the th 2‘-2^ region, and thin, symmetrically arranged proximal spines; x 5. g-h. Obverse view of large rhabdosome, G.S.C. 31728. g, full rhabdosome with distal part drawn from slab counterpart; x3. /;, detailed view of proximal end showing short, thickened proximal spines; x 8. i-j, Mature rhabdosome, G.S.C. 31729, in biprolile view proximally, and twisted to an oblique view distally. /, complete rhabdosome; x2. /, detailed view of proximal end showing short, stubby proximal spines thickened up to the aperture of th F; x5. 16 PALAEONTOLOGY, VOLUME 17 Considerable variation characterizes the proximal end of C. spiniferus particularly in the development of the proximal spines. (The following comments do not con- stitute a study of the astogenetic growth of the species, but are the result of observing extensive collections from various localities and cores from north-eastern North America). Spines of immature individuals with five pairs of thecae or less (text-figs. 2a, ij) are tiny and thin, the thecal spine measuring as little as 0-5 mm in length and the virgella, or virgellar spine, 1 00 to 1-3 mm. Spines may be asymmetrically or, less com- monly, nearly symmetrically arranged around the proximal end, with the virgella directed more downwards than the thecal spine (text-figs. 2-4). The virgella is usually slightly to markedly longer than the thecal spine. Both spines develop to various degrees during the growth of the rhabdosome : some spines, like those of the lectotype (PI. 1, figs. 4, 8; text-figs. 4a-c) grow long but remain very thin, a few others (text-fig. 3d) grow to exceptional lengths (maximum observed is 3-7 mm) ; most spines, however, are shorter, from 1 -5 to 2-5 mm in length, and tend to thicken. In some mature speci- mens, thickening of the spines has proceeded to such an extent that the sicula itself appears to have been prolonged as a spine (text-figs. 2c,/, g, h \ 4c, g-j). Thickened spines tend to grow more symmetrically than thin spines. Finally, basal membranes may lie on, or enclose, one or both spines (text-figs. 3a-d). These membranes may be just barely developed or have grown upwards so as to cover the aperture of th F (text-figs. 3a, c), as in C. bicornis. Few individuals of C. spuuferus, however, bear basal spines and these occur more commonly in collections from the lower Utica Shale, which corresponds to the lower range of the species. Exceptionally the virgella may bifurcate, furnishing the proximal end with three basal spines (text-figs. 3c,/). Remarks. C. spiniferus is practically identical to C. diplaeanthus described by Bulman (1932) from collections from the Wesenberg Beds of Kurland, Estonia. The only difference between these two species is that thecal excavations are shallower in C. diplaeanthus than in C. spiniferus, occupying one-seventh the breadth of the rhabdosome in the former versus one-quarter or less in the latter. Mature rhabdo- somes of C. diplaeanthus with thickened proximal spines (Bulman 1932, pi. 3, figs. 1, 4-5) cannot be distinguished from similar rhabdosomes of C. spiniferus (text-figs. 4g-j). It might be better to regard C. diplaeanthus as a subspecies of C. spiniferus or a geographic variant of the same. C. spiniferus could be considered a descendant of C. bicornis, which has undergone some definite modifications in the proximal end. The sicula is largely exposed on the obverse side of the rhabdosome as in C. bicornis, but th 1 * has grown both down and upwards making a full U-turn, rather than only down and outwards in an inverted C, as in C. bicornis. Th U grows farther upwards, but has lost its thecal spine which is replaced by the virgella grown outwards as a second spine (compare text-fig. la with text-figs. 2a-d). Spines are generally shorter in C. spiniferus and basal membranes have either disappeared or are reduced considerably in size. The septum originates at the th 2' -2^ level as in C. bicornis', thecal spacing is similar in the two species. Stratigraphic and geographic occurrence. C. .spiniferu.s occurs widely in north-eastern North America ranging in great numbers through the zone bearing its name (Riva 1969). It is common in the lower Utica Shale, the Schenectady beds, part of the Snake Hill Shale of New York, parts of the parautochthonous sequence of Quebec (St. Germain Complex), the ‘Cloridorme Formation’ of Gaspe, the Macasty Shale of Anticosti Island, and their stratigraphic equivalents. The writer has also identified it from Unit C of the RIVA: ORDOVICIAN GRAPTOLITES 17 Summerford Group of north-central Newfoundland (Horne 1970, pp. 1772-1773) and other units from Notre Dame Bay. This species also occurs in the Viola Limestone of Oklahoma (Ruedemann and Decker 1934), the lower Maravillas Chert of Texas (Berry 1960), and from the western United States and Canada (Ross and Berry 1963; Riva 1970, p. 2712; Jackson et al. 1964). I. Strachan (written comm., 1963, 1973) has recognized it at Girvan and Moffat in Scotland. C. diplacanthus has been reported from Sinkiang, China (Mu ei al. 1960), besides Estonia. Near the top of the Utica there occurs a short, broad climacograptid with short basal spines, less than 1 mm long, one of which is the virgella and the other a spine on th 1’ (Riva 1969, figs. 4m, n). The poor material hitherto on hand has prevented a description of this form which, however, can only be interpreted as a descendant of C. spinifems characterized by extremely reduced or atrophied basal spines. Climacograptus putillus (Hall) 1865 Graptolithus putillus Hall, pp. 27, 44, pi. A, figs. 10-12u. 1908 Climacograptus putillus (Hall); Ruedemann, pp. 415-419 (partim), text-figs. 368-370 {non text-figs. 371-377). 1925 Climacograptus putillus (Hall); Ruedemann, pp. 60-64 (partim). 1947 Climacograptus putillus (Hall); Ruedemann, pp. 434-435, pi. 72, figs. 29, 32, 33, 30?, 31?, 34-42? Ruedemann (1925, pp. 60-64) attempted to separate the minute Middle Ordovician graptolites that he (1908) and other workers had previously referred to C. putillus and distinguished six forms among them; Climacograptus eximius, C. strictus, C. tenuis, C. pygmaeus, C. lorrainensis, and C. putillus s.s. A restudy of the type material of these species and the study of other collections from type localities and equivalent strata in New York and Quebec have led the writer to suggest some addi- tional modifications and revisions: 1 . Climacograptus lorrainensis. This form belongs to Glyptograptus and it is identical with C. rougensis Parks from the upper Utica (Gloucester) of Ontario (Riva 1969, p. 526, figs. 6d-f; Parks 1928, pp. 63-64, text-figs. 6-8). G. lorrainensis ranges through the upper Utica and the lower Lorraine. 2. Climacograptus pygmaeus and C. tenuis. These two names refer to the same species. Rudemann described rhabdosomes of a Climacograptus, 6-8 mm long and 0-7- 0-8 mm wide, from the upper Utica Shale at Mohawk, N.Y., as C. pygmaeus, and identical forms, but 12 mm long and 1 mm wide, from the base of the upper Utica at the Trenton-Utica contact on Big Brook near Westernville, N.Y., as C. tenuis. The study of numerous individuals from the upper Utica has amply demonstrated that this form commonly attains 1 mm in width and may, exceptionally, be as much as 15-17 mm long, although shorter individuals (6-8 mm) predominate. Both names were proposed by Ruedemann ( 1925, p. 63) at the same time but accom- panied only by descriptions. However, the 1925 description of C. pygmaeus contains a reference to previously figured specimens (Ruedemann 1908, p. 416, text-fig. 376) which he (1947, pi. 72, figs. 22-24) subsequently designated as cotypes (= syn types) of C. pygmaeus. On the other hand, no figures of C. tenuis were ever published by Ruedemann, although a search through the N.Y. State Museum revealed a specimen (N.Y.S.M. 1 1 542) from near Westernville, N.Y., that he had selected as type but never figured. Accordingly, the name C. pygmaeus stands in preference, according to article 28a of the International Rules over C. tenuis and agrees with the name usually given to these tiny climacograptids. 3. Climacograptus eximius. Here Ruedemann (1925, pp. 62, 64) raised to specific 18 PALAEONTOLOGY, VOLUME 17 rank deformed, minute climacograptids which he had previously (1908, p. 420, text-figs. 378-384) separated as C. putillus mut. eximius. The type material of eximius contains two distinct forms: the specimens of the first three 1908 figures (text-figs. 378, 379, and 380?) belong to Pseudoclimacograptus modestiis (Ruedemann) and the other four (text-figs. 38 1 -384) to C. brevis strictus as redefined below. Ruedemann apparently referred to eximius all minute climacograptids from Normanskill beds near Albany, N.Y. In raising mut. eximius to specific rank, Ruedemann (1925, p. 62) noted that C. eximius was characterized by ‘distinctly climacograptid . . . thecae’. This qualifica- tion best applies to the specimens of the first three 1908 figures of mut. eximius, rather than to the other four which have more smoothly rounded, almost sigmoidal thecae. The name C. eximius thus becomes a junior synonym of P. modestus. It is discussed further below. 4. Climacograptus strictus. The name C. strictus was intended for climacograptids with less closely spaced thecae than those referred to C. eximius and separated by ‘wide notches’. It was believed to be restricted to the lower Canajoharie Shale. Identical specimens from Normanskill beds were included in C. eximius. Morpho- logically the types of C. strictus are similar to those of C. brevis Elies and Wood, but differ from the latter in being invariably smaller, both shorter and narrower, suggest- ing a clearly distinct subspecies of C. brevis. It will be treated as such below. 5. Climacograptus putillus (Hall). Hall’s (1865, pi. A, figs. 10-12a) type is an internal mould of a climacograptid in full relief, from the Maquoketa Shale near Dubuque, Iowa. Ruedemann (1908, p. 416, text-figs. 369, 370) figured the type in full, together with a proximal fragment on the same slab, but added nothing else to the knowledge of the morphology of the species. Paratype material in the A.M.N.H. collected by Hall from a locality 12 miles west of Dubuque, Iowa, contain some flattened material with the periderm partly preserved which suggests Climacograptus crassitestus (Ruedemann). Other paratype material from Iowa and Illinois labelled D. putillus consists, instead, of internal moulds of Diplograptus peosta (Hall) (= Ortliograptus amplexicaulis), also in full relief. Climacograptus brevis strictus (Ruedemann) Plate 2, figs. U3; text-figs. 5a- j 1908 ClimacograptiLs putillus (Hall) (partiiu); Ruedemann, pp. 415-419, text-figs. 371-373 (uoji figs. 368-370, 374-377), pi. 28, figs. 14, 15, \6a. 1908 Climacograptus putillus (Hall) mut. eximius Ruedemann (partim), pp. 420, text-figs. 381- 384 (non figs. 378-380), pi. 28, fig. 16. 1925 Climacograptus strictus Ruedemann, pp. 62, 63. 1947 Climacograptus strictus Ruedemann, p. 436, pi. 72, figs. 16-19. 1947 Climacograptus eximius Ruedemann (partim), pi. 72, figs. 4-7, 13? Holotype. N.Y.S.M. 6953 (Ruedemann 1947, pi. 72, fig. 17) from Snake Hill Shale in a quarry (now partly filled) located on the right side of stream flowing through the Albany Rural Cemetery (Section 23). The quarry is about 400 feet north of the tomb of U.S. President Arthur. Other material. N.Y.S.M. 6950, 6951, 6952, all designated as cotypes (-= paratypes) of C. strictus, from the same locality as the holotype. N.Y.S.M. 6914, 6915, 6917, from Normanskill shale at station 30 (Ruede- mann 1901, pp. 541-542), north end of Lansingburg, N.Y.; and N.Y.S.M. 6916, from Normanskill shale at station 34 (op, cit., p. 543), Glenmont, N.Y. (The last four specimens are listed as cotypes of C. eximius by Kilfoyle, 1954, pp. 77-78.) A.M.N.H. 1042 on slab bearing holotype (part) of C. parvus, lower Austin RIVA: ORDOVICIAN GRAPTOLITES 19 Glen Greywacke at Kenwood, N.Y.; A.M.N.H. 1042A, 1042B, Hall’s collections. An individual on the same slab as the holotype, N.Y.S.M. 6953. Description. Rhabdosome short, attaining a maximum known length of 13 mm but commonly less than 10 mm (virgella and nema excluded), and narrow, widening from 0-7-0-8 mm at the level of th P aperture to 0-9-1 -1 mm (average 1 mm) distally. Thecae are six to seven in the first 5 mm of the rhabdosome. Apertural excavations are deep, about one-third the width of the rhabdosome, and long, half the length of the free ventral wall of the thecae. Apertural excavations are commonly semicircular, rounded on the inside and strengthened by a thin selvage. The supragenicular wall tends to be inclined away from the axis of the rhabdosome; the geniculum is usually rounded off or smooth, thus giving a sigmoidal appearance to thecae of individuals preserved in subscalariform view. The proximal end is characteristically provided with a thin virgella about 0-4-0-5 mm long; the sicula is about 0-8 mm long. A thin virgula extends through and beyond the rhabdosome. Discussion. Specimens of C. brevis strictus are commonly preserved in subscalariform view which, when flattened, give the rhabdosome the nondescript appearances of text-figs. 5a, h, d,j and of Plate 2, fig. 3. The holotype itself (text-figs. 5a, b) is pre- served similarly. No trace of a septum or septal groove has been seen on the type specimens, but it will likely show up on better material. In erecting C. eximiiis and C. strictus, Ruedemann (1925, p. 64) surmised that the former was closely related to Glyptograptus teretiusculus var. siccatus and the latter indistinguishable from Climacograptus brevis, both as figured and described by Elies and Wood (1906-1907). He indicated, however, that only ‘actual comparison could settle the question of their relationships’. The writer has studied the type and topo- type material of Elies and Wood’s species and subspecies and has noted that: (1) G. teretiusculus var. siccatus is based on rhabdosomes of C. brevis compressed diagonally so that thecal outlines appear sigmoidal; and (2) much of the type material of var. siccatus and C. brevis is from the same locality and horizon near Builth, Wales. Strachan (1971, p. 39) considers var. siccatus synonymous with C. brevis. C. brevis is morphologically identical with C. strictus, but rhabdosomes of C. brevis are invariably larger— being commonly 9 mm long, up to T2 mm wide, and with sicula 1 mm and virgella TO- 1-2 mm long — than those of C. strictus (compare text- figs. 5a-j with text-figs. 5k-I) and, consequently, C. strictus is here considered a dis- tinct subspecies of C. brevis. C. brevis strictus could be compared to C. pauperatus Bulman from the Middle Ordovician of Scandinavia, but the latter differs in being both shorter and narrower although it has a comparable thecal spacing. Strachan’s (1959) C. brevis mutabilis from Sweden is close to C. brevis strictus in length and thecae spacing, but the former is much broader and differs in details of the proximal end. Geographic and stratigraphic occurrence. In eastern North America C. brevis strictus ranges through the N. gracilis, D. multidens, and the C. aniericanus Zones. It is of infrequent occurrence in Normanskill strata but it becomes more common in Magog and lower Canajoharie beds. In the Canajoharie, C. brevis strictus gradually increases in size to the much larger C. mohawkensis (= C. minimus of Elies and Wood) which is dominant in the upper Canajoharie. No form of the brevis-mohawkensis group occurs in the Utica and younger rocks of north-eastern North America, but a form of this group, C. crassitestus (Ruedemann), persisted into the Upper Ordovician of the Ouachita Geosyncline. TEXT-FIG. 5. Climacograplus brevis slrictus(a-j) and C. brevis (k, I). a, b, Holotype, N.Y.S.M. 6953, obliquely compressed. From Snake Hill Shale, Albany Rural Cemetery, Albany, N.Y.; respectively x 10 and x 5. c, e,j. Cotypes, N.Y.S.M. 6952, from group figured by Ruedemann (1908, pi. 28, fig. 14) as C. putillus. From Snake Hill Shale, Albany Rural Cemefery, Albany, N.Y. ; all x 5. d. Specimen on same slab as fhe holofype, N.Y.S.M. 6953, opposite side; obliquely compressed; x 5. g, N.Y.S.M. 6916, figured by Ruede- mann (1908, p. 240, fig. 383) as muf. eximius. From Mt. Merino Chert and Shale at Glenmont, N.Y. ; x 5. h, N.Y.S.M. 6914, figured by Ruedemann (1908, p. 420, fig. 381) as muf. eximius. The counferpart was also figured by Ruedemann (op. cif., fig. 384). The specimen is badly deformed by lateral compression. From Mt. Merino Chert and Shale at Lansingburg, N.Y. ; x 5. i,j, Rhabdosomes on A.M.N.H. 1042. /, biprofile view, periderm partly lost, flat; j, oblique view, laterally compressed. From the lower Austin Glen Grey- wacke on the Normans Kill at Kenwood, N.Y.; both x 5. k, /, Topotypes of C. brevis for comparison with C. brevis strictus. k, rhabdosome in biprofile view (least common), BU 1849; /, rhabdosome in oblique view (most common), BU 1850. Lapworth collection at Birmingham University, England. From Llan- drindod Wells, Wales; both X 5. RIVA: ORDOVICIAN GRAPTOLITES 21 Climacograptiis pygmaeus Ruedemann Plate 2, figs. 6, 1 1 ; text-figs. 6a-h, la-h 1908 Cliniacograptus piitillus Hall; Ruedemann, pp. 415-419 {passim), fig. 376. 1925 Cliniacograptus pygmaeus Ruedemann, p. 63. 1925 Cliniacograptus tenuis Ruedemann, p. 63. 1928 Cliniacograptus prolificus Parks, pp. 61-62, fig. 5. 1947 Cliniacograptus pygmaeus Ruedemann, pp. 435-436, pi. 72, figs. 22-24. 1947 Cliniacograptus tenuis Ruedemann, p. 436, pi. 75, figs. 127-15? 1947 Cliniacograptus prolificus Parks; Ruedemann, pp. 436-437, pi. 74, fig. 58 {non fig. 59). 1969 Cliniacograptus pygmaeus Ruedemann; Riva, p. 522, figs. Af-li. 1970u Cliniacograptus pygmaeus Ruedemann; Berry, p. 212, text-fig. 16. Syntypes. N.Y.S.M. 6942. The syntypes consisted of three tiny rhabdosomes (Ruedemann 1908, p. 416, fig. 376) which, because of the fragile nature of the periderm, have flaked off. It will be, perhaps, necessary to select a neotype from the upper Utica Shale at the type locality which is on Fullmer Creek, near Mohawk, N.Y. Other material. N.Y.S.M. 11542, proposed type of C. tenuis, from the uppermost Utica on Big Brook, east of Westernville, N.Y.; R.O.M. 264 bearing types of C. prolificus, from the upper Collingwood Shale (-- upper Utica) at Camperdown, Ontario; several pyritized rhabdosomes, in full or low relief, from the upper Utica of the Lozo-Joseph 1 and 2 cores, the Bald Mountain core of the St. Lawrence Lowlands of Quebec (Riva 1969). Extensive collections of flattened specimens from the aforementioned cores and other localities in the upper Utica Shale of the St. Lawrence Lowlands. Description. Rhabdosome small, commonly 6-8 mm long but frequently attaining 10-12 mm and, exceptionally up to 17 mm in length, and thin, widening impercep- tibly from 0-5-0-6 mm at the level of th F to 0-8- 1-2 mm (average 0-9- 10 mm) between the fifth and the seventh pair of thecae. Thecae are small and closely spaced, usually 7-8-5 (average 8) in the first 5 mm of the rhabdosome (excluding the first pair of thecae), decreasing distally to about 14-15 in 10 mm in exceptionally long rhabdo- somes. Thecae are square, tending to be inclined away from the axis of the rhabdo- some in specimens preserved in relief (text-fig. 6c) or as flattened films (text-figs. Id-h) ; they overlap one-third the succeeding thecae. Apertural excavations are usually wide, deep, and semicircular, occupying at least one-third the width of the rhabdosome. The proximal end bears a short, spiny virgella about 0-2-0-5 mm long and thickened by secondary material from the rhabdosome and two sicular spines. The sicula is 1-2-1 -3 mm long (text-fig. 6d). The virgula extends through and for a short distance beyond the rhabdosome. The rhabdosome is aseptate. The sicular is mostly exposed on the obverse side of the rhabdosome as far as the aperture of th 2^ (text-figs. 6a, e,f). Th P grows first down from the middle part of the sicula to well below the sicular aperture and then it grows sharply upwards until its aperture reaches or passes the point of origin on the sicula (text-figs. 6a, c,/); th P first grows across the rhabdosome behind the sicula, and then turns upwards terminating well above the aperture of th P (text-figs. 6b, c). The periderm of C. pygmaeus is thin and brittle, flaking off readily even in fresh specimens. Rhabdosomes in relief are rare, but when so preserved they are oval, rectangular, or semicircular in cross-section (PI. 2, figs. 6, 1 1 ; text- figs. 6a-c). Remarks. C. pygmaeus is a dwarfed member of C. typicalis group which found its optimum development in the black Utica Shale of north-eastern North America. Tiixi-MG. 6. Climacograptus pygmaeus Riiedemann. a, Rhabdosome in full relief, G.S.C. 31731, obverse side, showing sicula exposed up to the level of th 2'-2^, wide apertural excavations, absence of thecal hoods and type of spines at the proximal end. From Bald Mountain core, depth 1715'; x 10. b, c, Rhabdosomes in full relief, G.S.C. 31732-31733, reverse side, showing wide apertural excavations and gently, inclined thecae. From Lozo-Joseph 2 core, depth 325'-330'; both x 10. d. Flattened sicula of C. pygmaeus with th F fully developed. From Lozo-Joseph 2 core, depth 190' 195'; x 10. e,j, Rhabdosomes in low relief, G.S.C. 31734-31735, obverse views, showing sicula exposed up to the aperture of th F and 2‘ respectively, and wide apertural knotches. From Lozo-Joseph I core; depth 655'; both x 10. g, /;, Flattened rhabdo- sonies, G.S.C. 31736-31737, obverse views, showing typical preservation of average size specimens with periderm partly Baked olT, apertural excavations enclosed by a thin selvage, and virgula showing beneath periderm. From Lozo-Joseph 2 core, depth 155' 160' and 285'-289' respectively; both x 10. RIVA: ORDOVICIAN GRAPTOLITES 23 TEXT-FIG. 7. Variation in the length of Climacograptus pygmaeus Ruedemann. a, b, Average size rhabdo- somes, G.S.C. 31736-31737, flattened, with periderm partly lost. Same as those of text-figs. 6g-/;; x5. t’. Average size rhabdosome, G.S.C. 26584, with periderm gone except for tiny fragments, parallel-sided, and showing typically square thecal outlines. From Bald Mountain core, depth 1565'; x5. d, c. Longer and slightly wider rhabdosomes, G.S.C. 26585-26586, than preceding ones with markedly inclined thecae, perideiTn largely preserved, and virgula partly visible under periderm. From Lozo-Joseph 1 core, depth 660'; both x 5. /, Long and wide rhabdosome, G.S.C. 26587. Periderm is largely preserved but thecal details are largely obscured. From Bald Mountain core, depth 1885'; x5. g, h. Exceptionally long and bent rhabdosomes, G.S.C. 26588-26589; from a level in the Bald Mountain core (1925' and 1945', respec- tively) characterized by long individuals of C. pygmaeus \ both x 5. All these specimens are deposited in the collections of the Geological Survey of Canada. The cores referred to are discussed elsewhere (Riva 1969). A Study of the development of the typicalis group through more than 2000 feet of Utica shows that C. typicalis first appears about 400 feet above the base of the true Utica, as redefined (Riva 1969, p. 514), and ranges up to the top of the formation. C. pygmaeus appears suddenly and in great numbers in the middle of the Utica. This is the first form to branch off C. typicalis, to be followed near the top of the Utica by C. typicalis magnificus, a giant biserial form with a 4-0-4-5 mm width and a narrow, whip-like proximal end, and by forms intermediate between C. typicalis and C. pygmaeus which are presently referred to as C. typiealis posterns. (The principal members of the typicalis group are figured in Riva 1969, p. 527, figs. Aa-j, o.) C. pygmaeus differs from C. typicalis by its minute size, tightly packed thecae, the imperceptible widening of the rhabdosome and the absence of a hood over the thecal apertures. It is closer to the rhabdosomes referred to as C. typicalis posterns 24 PALAEONTOLOGY, VOLUME 17 particularly in the gradual widening of the rhabdosome and the absence of thecal hoods, but it differs in being smaller and narrower and having much more closely packed thecae. Geographic and stratigraphic occurrence. C. pygmaeus is restricted to the upper Utica Shale and equivalent strata of north-eastern North America where it ranges through a zone bearing its name. Usually it is present in great profusion, literally covering and blackening whole bedding surfaces. In New York it ranges through the upper Utica Shale and the Atwater Creek, Deer River and the basal Frankfort shales and sandstones of the upper Mohawk and Black River Valley; in the Champlain Valley and the St. Lawrence it also ranges through the upper Utica into lower Lorraine strata; in Ontario it is recorded, both as C. pygmaeus and as C. prolificus, in the Collingwood Formation (uppermost Utica) (Parks 1928; Caley 1936). Farther west- wards it has been recognized as far as the northern Michigan Peninsula in shales correlative with the upper- most Utica (Ruedemann and Ehlers 1924; Berry 1970a). This writer has identified it in a collection from the Cincinnati area in Ohio (Albers coll, in the A.M.N.IT.). C. pygmaeus also occurs in great numbers in the Macasty Shale of the Lake St. John area, deep in the Precambrian Shield, and of Anticosti Island (Riva 1969). This form has not been reported from the Appalachian sequences, except for a few individuals tentatively referred to it by this writer from the Hillgrade Shale of the Notre Dame Bay, Newfoundland (coll, by Marshall Kay 1970). In the Hillgrade collections, C. cf. pygmaeus is associated with a typical British P. linearis fauna which includes P. linearis itself, the third known occurrence of this species in North America. Genus pseudoclimacograptus Pfibyl, 1947 Pseudoclimacograptus modes tus (Ruedemann) Text-figs. 8u, b 1908 Climacograptus putillus mut. eximius Ruedemann (partim), p. 420, text-figs. 378, 379, 380? (non text-figs. 381-384, pi. 28, fig. 16). 1908 Climacograptus modestus Ruedemann, pp. 432-433, text-figs. 400-403, pi. 28, fig. 30. 1925 Climacograptus eximius Ruedemann; Ruedemann (partim), pp. 62, 64. 1947 Climacograptus modestus Ruedemann; Ruedemann, p. 432, pi. 73, figs. 32-46. 1947 Climacograptus eximius Ruedemann; Ruedemann (partim), p. 435, pi. 72, figs. 2, 3 (non figs. 1, 4-15). 1948 Climacograptus modestus Ruedemann; Bulman, pp. 222-223, text-fig. 1. Material studied. N.Y.S.M. 6911, 6912, 6913, from Mt. Merino Chert and Shale blocks at station 30 (Ruedemann 1901, pp. 541-542) at the north end of Lansingburg, N.Y. TEXT-FIG. 8. Pseudoclimacograptus spp. a, b, Syntypes of C. eximius, herein referred to P. modestus (Ruede- mann). a, N.Y.S.M. 6912, deformed and fragmental rhabdosome showing deep apertural excavations; h, N.Y.S.M. 6913, laterally compressed, but showing spinose nature of proximal end. From Mt. Merino Chert and Shale at Lansingburg, N.Y.; both x 5. c, d, e, f, P. scharenbergi. c, rhabdosome on A.M.N.H. I042D, one of a group of eight short individuals labelled C. parvus’, d, rhabdosome on A.M.N.H. 1034, with type of G. mucronatus, showing zigzag septal groove, long virgella accompanied by sicular down- growth; e, rhabdosome on A.M.N.H. 1030a with lectotype of C. bicornis, clearly showing zigzag septum, long virgella, and sicular downgrowth; /, long rhabdosome on A.M.N.H. 1040C. This specimen is labelled C. parvus and is next to a specimen figured by Hall ( 1 847, pi. 73, fig. 4g) as Graptolithus scalaris, but which is really a C. parvus compressed in scalariform view; other specimens on the same slab are also of C. parvus. From the Austin Glen Greywacke on the Normans Kill at Kenwood, near Albany, N.Y. ; all X 5. g,/;,Type of C. parvus, A.M.N.H. 1042A, not previously figured. This specimen bears a label with ‘n.sp.’ From Austin Glen Greywacke on the Normans Kill at Kenwood, N.Y.; respectively X 5 and X 3. /, P. scharen- bergi stenostoma, A.M.N.H. 1 034/ la, on reverse side of slab bearing a syntype of Graptolithus mucronatus (Hall 1847, pi. 73, fig. Ic), showing slit-like thecal excavations and trace of zigzag septum at level of th 2‘. From Austin Glen Greywacke on the Normans Kill at Kenwood, N.Y.; x5. 26 PALAEONTOLOGY, VOLUME 17 Remarks. The morphology of Pseudoclimacograptus modestus has been adequately described by Ruedemann (1908) and Bulman (1948). The types of P. modestus came from Normanskill Chert and Shale of a quarry at the north end of Mt. Merino, N.Y. (Ruedemann 1908, pp. 13-15, 423). In erecting this species Ruedemann (1908, p. 423) noted that the form also occurred in collections from ‘the same horizon at the northern end of Lansingburg’, to the north of Troy, N.Y. The Lansingburg collections contain also a number of small climacograptids which Ruedemann (1908, p. 420) first dis- tinguished as mut. eximius, and later (1925, p. 62) clearly separated as C. eximius. As already pointed out in the discussion under C. putUlus, Ruedemann assigned two forms to mut. eximius: the individuals of his 1908 text-figs. 378, 379, and possibly 380 are deformed or fragmental P. modestus, while those of text-figs. 381-384 are identical to C. brevis strictus as redefined in this paper. The rhabdosome of text-fig. 378 (herein refigured as text-fig. 8^) is distorted or compressed normally to its axis, as shown by the narrow width (0-7 mm versus 1 mm in undeformed specimens), unequal though deep thecal apertures, and, especially, by the straight rather than convex supragenicular walls. The rhabdosome typically bears three proximal spines (the virgella, and mesial spines on th 1^ and th F), has eight thecae in its first 5 mm, and shows traces of a zigzag septal groove. The rhabdosome of text-fig. 379 (herein refigured as text-fig. 8a) is fragmental as well as deformed, but enough is left to show that its width barely exceeds 1 mm, thecae are nine in 5 mm, thecal excavations are deep and narrow, and supragenicular walls are distinctly convex. A faint trace of a zigzag septal groove can be seen in the middle of the rhabdosome. The proximal part is missing, but in all other respects this specimen is identical to P. modestus. The rhabdosome of text-fig. 380 is too deformed to be adequately understood. Pseudoclimacograptus scharenbergi stenostoma (Bulman) Text-fig, 8/ 1947 Climacograptus scharenbergi var. stenostoma Bulman, p. 70, pi. 7, figs. 11, 1 2, pi. 8, figs. 2-4, 8. 1960 Climacograptus scharenbergi var. stenostoma Bulman; Berry, p. 83, pi. 15, fig. 6. Material. Specimen on A.M.N.H. 1034/la bearing a type of Hallograptiis mucronatus (Hall 1847, pi. 73, fig. Ic); Austin Glen Greywacke, Kenwood, N.Y. Remarks. Bulman (1947, p. 70) distinguished as var. stenostoma a form similar in many respects to P. scharenbergi but differing in being narrower and having ‘exceed- ingly narrow, slit-like’ thecal excavations. In North America this variety has since been reported from the Normanskill of New York (Berry, 1962a), from the Woods Hollow Shale of Texas (Berry 1960), and the Valmy Formation of Nevada (Ross and Berry 1963). This writer has found it, but sparingly, in several Normanskill col- lections from the Appalachians. It is also likely that some tiny climacograptids referred to, or identified, as Climacograptus eximius, C. putillus, or Pseudoclimaco- graptus modestus may more properly belong to subsp. stenostoma. No American specimen of subspecies stenostoma has yet been figured, except for an individual from the Woods Hollow Shale tentatively identified as such (Berry 1960, pi. 15, fig. 6). The specimen here figured (PI. 2, fig. 10; text-fig. 8/) is on a slab of the Hall collection in the A.M.N.H. The rhabdosome is mostly preserved, in slight relief, 5-3 mm long (proximal spines excluded), 0-9- TO mm wide, and with eight to RIVA: ORDOVICIAN GRAPTOLITES 27 nine thecae in the first 5 mm of the rhabdosome. Thecal excavations are charac- teristically narrow, slit-like, and deep and strengthened by a thin selvage merging the infragenicular wall with the thecal aperture. The excavations of the first three thecae are slightly wider than those of the more distal thecae. The proximal end is furnished with three spines: the virgella and a mesial spine on each of the first two thecae. The supragenicular walls are markedly convex and thecal apertures slightly introverted. A trace of the zigzag septal groove can be seen in the th P-2^ region, but this feature is scarcely visible even in the holotype, which is an ‘isolated’ rhabdosome. The subspecies stenostoma approaches P. nwdestus particularly in the width of rhabdosome, number of thecae per 5 mm and over-all appearance. Distinction between these two forms may, in fact, be difficult with flattened specimens. Sub- species stenostoma differs from P. modestus, as from P. scharenhergi, principally in its characteristically slit-like apertural excavations, and in the more convex supra- genicular walls. These features alone should be sufficient criteria to separate at least relatively undeformed specimens. Pseiidoclimacograptiis scharenbergi (Lapworth) Plate 2, figs. 4, 5; text-figs. 8c-/; 1865 Climacograptiis parvus Hail, p. 57 (nom. midion). 1876 Climacograptiis Scharenbergi Lapworth, pi. 2, fig. 55. 1896 Climacograptiis phyllophorus Gurley, pp. 77-78, pi. 4, figs. 4-6. 1906 Climacograptiis Scharenbergi Lapworth; Elies and Wood, p. 206, text-fig. 116, pi. 27, figs, \4a-e. 1908 Climacograptiis scharenbergi Lapworth; Ruedemann, pp. 428-431, text-figs. 394-399, pi. 28, fig. 31. 1908 Climacograptiis parvus Hall; Ruedemann, pp. 90-96, 426-428, text-figs. 20, 24-26, 27, 28, 33-36, 388-393, pi. 28, figs. 19-23. 1932 Climacograptiis scharenbergi CaipvNOxih', Bulman, pp. 6-10, text-figs. 1-3, pi. I, figs. 1-35. 1947 Climacograptiis scharenbergi Lapworth; Bulman, pp. 65-70, text-figs. 34-38, pi. 7, figs. 1 10, pi. 8, figs. 1, 5-7. 1947 Climacograptiis scharenbergi Lapworth; Ruedemann, pp. 438-439, pi. 74, figs. 41-54. 1947 Climacograptiis parvus Hall; Ruedemann, p. 433, pi. 74, figs. 10-26. 1960 Climacograptus parvus Hall; Berry, p. 81, pi. 16, fig. 12. 1963 Climacograptus scharenbergi Lapworth; Ross and Berry, p. 129, pi. 9, figs. 14, 17. 1963 Climacograptus phyllophorusG\xx\ey\Koss,eind^&ny,p. 127, pi. 8, fig. 17?, pi. 9, figs. 137, 18? For a more complete record on the distribution of P. scharenbergi, see Strachan 1971, p. 44. Material examined. Type of Climacograptus parvus Hall, part and counterpart, A.M.N.H. 1042A and B; eight small rhabdosomes and one fragment on the type slabs; a large rhabdosome on the type slabs next to specimen illustrated by Hall (1847, pi. 73, fig. 4g) as Graptolithus scalaris', a rhabdosome on A.M.N.H. 1034/la bearing type of Graptolithus miicronatiis (Hall 1847, pi. 73, fig. lu); a rhabdosome on A.M.N.H. 1030a bearing the lectotype of C. hicornis. Remarks. The morphology and general features and abnormalities of young and mature rhabdosomes of this species have been fully described and illustrated by Bulman (1932, 1947). The following discussion is concerned rather with the type of C. parvus (or C. phyllophorus) and other material in the A.M.N.H. The name C. parvus has had a somewhat tortuous history. It was first proposed by Hall in 1865 but remained a nomen nudum until Gurley in 1896 renamed and described the specimen in question as C. phyllophorus. Ruedemann, however, on the 28 PALAEONTOLOGY, VOLUME 17 grounds that Hall’s name was long established and recognized before Gurley’s, revived C. parvus in 1908 and the name persisted unchallenged in the American literature until 1963 when Ross and Berry called Ruedemann’s action ‘ill-advised’ and reinstated C. phyllophorus. Much of this confusion stems from the failure to study critically the morphology of the type specimen of C. parvus and the associated topotype material, and the tendency to attribute to C. parvus characteristics which also belong to P. scharenbergi. The type specimen itself (PI. 2, fig. 5 ; text-figs. 8g, h) has never been figured. Ruedemann’s (1908, text-fig. 388; 1947, pi. 74, text-fig. 10) figure of C. parvus is from the impression on the counterpart of the type slab (PI. 2, fig. 4) and it is rather stylized with most of the critical features not shown. The specimen itself is 13 mm long (excluding virgella and virgula) and widens gradually from 0-8 to 1-5-1 -6 mm. Thecae are thirteen to fourteen in 10 mm proximally (ex- cluding the first pair), with supragenicular walls appearing straight rather than convex because the rhabdosome is compressed obliquely. Apertural excavations are narrow and deep occupying at least one-third the width of the rhabdosome and strengthened by a thin selvage. The proximal end bears a spine on the first two thecae and an extremely long virgella (9-5 mm). The initial portion of the virgella is partly enclosed in a 2-mm long downgrowth which is interpreted as a sicular downgrowth similar to those recognized by Bulman (1947, pp. 68-69, text-fig. 36) in Laggan Burn specimens of P. scharenbergi. The virgula is long (17 mm), thick (0-2 mm), and formed by two parallel bands; it carries no ‘disc’ or vane at the distal end. Because the rhabdosome is flattened, no trace of a septum can be recognized, but otherwise the type of C. parvus shows the characteristic features of P. scharenbergi. Several rhabdosomes on the type slab of C. parvus and in the rest of the Hall col- lection from the Normans Kill bear an original Hall label with ‘C. parvus'. Of these, eight small ones and a large one are on the type slabs : one of the small ones, shown here as text-fig. 8c, clearly shows the deep and narrow thecal excavations, the convex supragenicular walls, the width, a trace of zigzag septum, and the beginning of a sicular downgrowth along the virgella as in P. scharenbergi \ identical features are also seen on the larger rhabdosome (text-fig. 8/), except for the sicular downgrowth. The trace of the zigzag septum is best shown by two other specimens of the Hall collection (text-figs. 8<7, c); these also exhibit a long virgella accompanied by a sicular downgrowth. None of these specimens carry a trace of a ‘disc’, or vane, at the end of the virgula. It is obvious that in 1865 Hall named but did not describe as C. parvus specimens which were later named and described as C. scharenbergi by Lapworth (1876). Gurley (1896, p. 77) based C. phyllophorus on the type of C. parvus. He considered, as Ruedemann did later (1947, p. 434) for C. parvus, the ‘disc’ at the end of the virgula as the characteristic feature of this form. A ‘disc’, however, is absent on the types of C. parvus, and Bulman (1964, pp. 462-463) has shown that such feature is typical of P. seharenbergi. Ruedemann considerably expanded the concept of C. parvus as shown by his 1908 figures (p. 426) repeated in 1947. Besides the type and one other form with sicular downgrowths (p. 426, text-figs. 388-389), he figured specimens with long and narrow apertural excavations and a ‘vesicle’ at the end of a long virgula (PI. 28, figs. 19-23; text-fig. 391), as well as a specimen with more shallow thecal excavations, but pro- RIVA; ORDOVICIAN GRAPTOLITES 29 vided with a virgella and mesial spines on th P and F. He noted in the discussion of P. scharenbergi (p. 431) that C. parvus differs from P. scharenbergi in the absence of a zigzag septal grove and in having shallow rather than deep apertural excavations. But the types of C. parvus have deep apertural excavations and the zigzag septal groove cannot be recognized because the specimen is compressed. Ruedemann, in fact, shows on p. 430, text-figs. 394-399, that the zigzag septal groove of P. scharen- bergi disappears with increasing deformation. Geographic and stratigraphic occurrence. P. scharenbergi is a common species in Middle Ordovician rocks of north-eastern North America, where it ranges through both the N. gracilis and the D. multidens Zones. Individuals bearing a long virgella and an accompanying sicular downgrowth were found to be common in recent collections from the Quebec City Formation, which is almost entirely in the D. multidens Zone. In this part of North America no Pseudoclimacograptus has ever been recognized above the D. multidens Zone, i.e., from the C. americanus to the beginning of the C. manitoulinensis Zone, an interval comprising more than 3000 feet of undisturbed black Canajoharie and Utica shales. A new Pseudoclimacograptus, P. cf. P. clevensis Skoglund (Riva 1969, p. 526, figs. 6y, k), appears in the C. manitoulinensis Zone associated with a graptolite fauna similar to that reported by Skoglund (1963) from the lower Harjuan Series of Sweden, Genus orthograptus Lapworth, 1873 Ortliograptus amplexicaulis (Hall) Plate 2, figs. 7-10; text-fig. 9 1847 Graptolites amplexicaule Hall, pp. 79-80, 316, pi. 26, figs. 1 \ a-b. 1867 Diplograptus amplexicaule (Hall); Hall, pp. 24, 223, pi. 3, figs. 6, 7. 1908 Diplograptus (Glvptograptus) amplexicaulis (Hall); Ruedemann, pp. 361-365, text-figs. 302- 304, 305-306?, 307, pi. 25, figs. 11, 13, 107, 12? 1947 Diplograptus (Amplexograptus) amplexicaulis Hall; Ruedemann (partim), pp. 411-412, pi. 70, figs. 1-4, 7-9, 5-6?, 10-13? (non fig. 14). 1960 Orthograptus tnmcatus var. intermedins Elies and Wood; Berry, p. 92, pi. 17, figs. 4-5. 1960 Orthograptus aff. O. tnmcatus (Lapworth); Berry, pp. 91-92. 1969 Orthograptus amplexicaulis (Hall); Riva, fig. 3a. This synonymy is restricted to specimens from the type locality or equivalent strata. Proposed lectotype. A.M.N.H. 634/1, Trenton Limestone at Middleville, N.Y. (PI. 2, fig. 6). The type locality could be any of the following exposures, and part of the same horizon: (1) Moltentanner Creek, one mile east and above Middleville; (2) City Brook, two miles north-east of the city; (3) Stony Brook, two miles south of the city. The lectotype is the specimen figured by Hall in 1867, pi. 25, figs. 6-7. It is one of the two specimens listed by Whitfield and Hovey (1898, p. 20) as the ‘types’ of the species. Material and localities. All type material in the A.M.N.H. labelled 643/1 from Middleville, N.Y., including a recently discovered collection (still in the original bag) made by Hall in 1847 from the Trenton Limestone at Trenton Tails, N.Y. (and bearing a label with: ‘G. amplexicaule Hall, Trenton Tails, very rare, I have just got two of them'). The J. W. Hall collection in the N.Y. State Museum from the Trenton Limestone at Middleville, N.Y. catalogued as 7128 (Harris quarry), 7129 (Stony Brook), 7130 (Moltentanner Creek). All specimens from the Trenton Limestone are preserved in low to full relief. The writer has had at his disposition extensive collections from the Canajoharie Shale at Chuctenunda Creek, the Snake Hill Shale of the Hudson River Valley, and the Utica Shale of New York, and from Trenton (Cobourg) Limestone, Canajoharie-Utica and Lorraine shales and siltstones of the St. Lawrence Lowlands, and the Macasty Shale of Anticosti Island. Ruedemann’s collections in the N.Y. State Museum, those of Parks in the Royal Ontario Museum, and those of Twenhofel from Anticosti Island in the Yale Peabody Museum have also been examined. The writer has also studied most of the specimens of Orthograptus tnmcatus (Lapworth) figured by Elies and Wood (1907, pp. 233-235, pi. 29, figs. 3a, b, e) in the Lapworth collection of Birmingham University and additional material from the Hartfell Shales in the Sedgwick Museum at Cambridge Uni- versity. 30 PALAEONTOLOGY, VOLUME 17 Description. The following description is based on the type material and also flattened rhabdosomes from Canajoharie-Utica Shales. No attempt is made to define the upper stratigraphic limit of the species as such a conclusion will have to be based on much larger and varied collections. Rhabdosomes attaining several cm in length, widening from 0-8 to TOO mm at the level of th F aperture to 2-0-3-5 mm (exceptionally 4-00 mm); the maximum width may be attained in 1 cm from the proximal end or very gradually, almost impercep- tibly, with all variations in between; it is generally maintained, although some rhabdosomes tend to narrow distally. Thecae alternate, twelve to sixteen (average 13) proximally, ten to twelve (exceptionally 14-15) distally; they overlap about three- fifths proximally and seven-tenths distally and are inclined about 30°-35° to the axis of the rhabdosome (in specimens in relief), the smaller inclination being typical of proximal thecae. Thecae consist of cylindrical tubes with simple apertures inclined about 30° from the vertical. Apertural excavations are shallow distally, but deeper and more pronounced proximally, and delimited by a selvage merging the thecal aperture with the supragenicular wall. The proximal end bears a mesial spine on th P, the virgella, and one visible sicular spine. The sicula is 2 0-2-5 mm long (as shown by flattened upper Utica specimens of text-figs. 9a-d) and exposed for 1 mm on the obverse side of the rhabdosome (text-fig. 9h). The rhabdosome is circular to ovoid in cross-section (PI. 2, fig. 8), but if the periderm has collapsed it may have a roughly rectangular cross-section or be so depressed along the axis as to appear concave. A thread-like nema extends from the apex of the prosicula through the rhabdosome and for a short distance beyond. The rhabdosome is aseptate. Remarks. Flattened rhabdosomes of O. amplexicaulis present different aspects depending on the direction of compression. Thecae of flattened specimens tend to be EXPLANATION OF PLATE 2 Figs. 1, 2, 3. Climacograptiis brevis strictus (Ruedemann). 1, A.M.N.FI. 1042. 2, A.M.N.H. 1042A. 3, A.M.N.H. 1042B; all with types of Clinuicograptus parvus Hall. Figs. 1 and 2 are biprofile views showing sharply square thecae, thin virgella, and virgula. Fig. 3 is a three-quarter face view. From the lower Austin Glen Greywacke on the Normans Kill, Kenwood, N.Y. All x 10. Figs. 4, 5. Climacograptus parvus Hall. 5, A.M.N.H. 1042B, shows the ‘type’, never before figured. 4, A.M.N.H. 1042A, the counterpart and one of the ‘types’ of Graplolithus mucronatus Hall. The counter- part has been figured by Ruedemann (1908, p. 426, text-fig. 388; 1947, pi. 74, fig. 10) who designated it as the holotype, but only the part. Fig. 5 bears a label with ‘n.sp.’. From the Austin Glen Greywacke on the Normans Kill, Kenwood, N.Y. Both x3. Figs. 7-10. Orthograptus amplexicaulis (Hall) from the Trenton Limestone of Middleville, N.Y., except fig. 8 from Trenton Falls. 7, A.M.N.H. 634/1. Proposed lectotype. Rhabdosome in low relief, with proximal end missing, and thecae characteristically alternate, of the orthograptid type. Figured by Hall (1847, pi. 26, figs. I \a-h\ 1867, pi. 3, figs. 6, 7). x 6. 8, A.M.N.H. 634/ld. Specimen in full relief, circular or ovoid in cross-section; whitened, x 5. 9, A.M.N.H. 634/ld. Paratype. Proximal end of thin rhabdo- some in low relief, x 6. 10, A.M.N.H. 634/ 1 a. Paratype. Broad specimen with periderm collapsed along axis so as to appear concave. Growth lines are pronounced. x6. Figs. 6, 11. Climacograptus pygmaeus Ruedemann. A.M.N.H. 29259 and 29260. 6, Individual in three- quarter face view in full relief, and of above average length. 1 1, Obverse side of an individual of average length, in fair relief, and with nearly fully exposed sicula. From the upper Utica Shale on the St. Maurice River, north of Three Rivers, Quebec. Collected by Y. Globensky. Both x 10. PLATE 2 RIVA, Climacograptus and Orthograptus 32 PALAEONTOLOGY, VOLUME 17 inclined at a higher angle than those of specimens preserved in relief, and this inclina- tion increases noticeably toward the thecal aperture; free ventral walls of distal thecae may form sigmoidal rather than straight outlines (text-figs. 9c-d) and those of proximal thecae tend to appear sub-climacograptid because of deeper apertural excavations. This preservation probably accounts for the fact that the species was at one time referred to Glytograptus by Ruedemann (1908, p. 361, et seq.). Hall’s (1847) first figures of G. amplexicaule are rather primitive, showing cone-like struc- tures with thecae appearing as alternating scales. His later figures (1867) are more accurate but still interpretative : one side of the rhabdosome is shown as convex with straight orthograptid thecae, and the other concave with straight but quadrangular thecae. The types (PI. 2, figs. 7-10) are all distal fragments, except possibly for one (fig. 9), all characterized by simple orthograptid thecae. Their cross-section is roughly quadrangular, somewhat compressed along the centre so as to appear concave on one or both sides. These features are purely preservational and not original or primary. In erecting the genus Amplexograptus, Elies and Wood (1907, pp. 218, 221, 258) were apparently influenced by Hall’s figures of D. amplexicaule which they considered (p. 268) closely related to A. perexcavatus, the genotype. This genus has since been redefined on particular thecal characteristics (Bulman 1962), but its name is derived, oddly enough, from a common species of Orthograptus. Priority of O. amplexicaulis (Hall) over O. truncatus (Lapworth). Examination of flattened specimens of O. truncatus figured by Elies and Wood (1907, pi. 29, figs. 3a, h, e) and of additional material from the Hartfell shales— the type specimen itself is apparently lost (O. M. B. Bulman 1970, pers. comm.)— shows that they are morphologically identical to flattened specimens of O. amplexicaulis from Canajoharie shales of eastern North America. The specimens figured by Elies and Wood, including the one called ‘typical specimen’ (pi. 29, fig. 3a), lack the proximal end. The proximal end, however, is preserved on several of the Hartfell specimens in the Sedgwick Museum (text-figs. 9c-g, ij) thus allowing a comparison with topotypes of the J. W. Hall collection (text-fig. 9h) and flattened specimens from the Utica (text-figs. 9a-d). The specimens of O. truncatus differ from those of O. amplexicaulis, only in degree : they have less closely spaced thecae (10-11 proximally and 9- 1 1 distally ) and a longer sicula (from 2-6 to 3-4 mm) than those of O. amplexi- caulis, but otherwise they are indistinguishable. Such differences may be of subspecific value but hardly suggest a distinct species. O. amplexicaulis was described, figured, and refigured (Hall 1847, 1867) well before O. truncatus (Lap- worth 1877) and therefore this name has full priority over the latter and should replace it as chef de file of this long-ranging group of orthograptids. In North America the name O. truncatus was not used until 1960. Ruedemann (1947, p. 412) considered O. truncatus identical to, or the ‘vicarious’ form of, O. amplexicaulis, which he, following Elies and Wood’s (1907, p. 268) interpretation, referred to Amplexograptus. Berry, instead, departed from the American usage by stating) I960, pp. 38, 91-92) that ‘specimens at the New York State Museum . . . from Trenton Tails, near Middleville, New York, identified by Ruedemann as Amplexo- graptus amplexicaulis should be referred to the O. truncatus group or to O. truncatus var. intermedins' . The specimens in question are those of the J. W. Hall collection, one of which (N.Y.S.M. 7130) he (op. cit., pi. 17, figs. 4, 5) figured as O. truncatus var, intermedins, without mentioning its source. Stratigraphic and geographic occurrence. O. amplexicaulis occurs sparingly in limestones of the lower part of the Trenton Group (Poland Member of the Denmark Limestone) of the Mohawk Valley of New York and its equivalent of the St. Lawrence Lowlands where it is not uncommon even in upper Trenton (Cobourg) Limestone. It is common in Canajoharie shales of eastern North America, the lower Macasty Shale of Anticosti Island, and part of the Snake Hill Shale of the Hudson River Valley. It is also likely that many specimens identified as Diplograptus vespertinus really belong to O. amplexicaulis. The dif- ferences between D. vespertinus and O. amplexicaulis are not yet fully understood partly because of the impossibility to recollect from the type locality of the former which is located in the highly sheared and deformed Snake Hill Shale of the Albany, N.Y., area. TEXT-FIG. 9. Orthograptus amplexicaulis (Hall), a-g. Growth stages and proximal ends of Ortho- graptus amplexicaulis. a, h. Growth stages, G.S.C. 26590 and 31738, from the uppermost Utica Shale of the B.M. core (Riva 1969), depth 1405', showing long sicula and squarish thecae; x5. c, d. Immature individuals, G.S.C. 31739 and 32375, with long sicula and broad proximal end. Same depth and core as above; x 5. e,f, g. Growth stages of O. trimcatus (Lapworth) from the Hartfell Shales, Dumfriesshire, Scotland, with a longer sicula than the North American specimens. Sedgwick Museum A19799; X 5. /?, Topotype, N.Y.S.M. 7128, Middleville, N.Y., figured by Ruedemann (1908, p. 363, fig. 302, pi. 25, fig. 11; 1947, pi. 70, figs. 4, 7), in partial relief with periderm collapsed along axis of rhabdosome; x 3. /, /, k, Topotypes of O. rrMnca/wx (Lapworth), Elies collection, Sedgwick Museum A19799, Hartfell Shales, D. clingani Zone, Dumfriesshire, Scotland, for comparison with North America specimens of O. amplexicaulis; x 3. 34 PALAEONTOLOGY, VOLUME 17 O. amplexicaulis ranges from the C. americanus to within the C. spiniferus Zone of the lower Utica. It is found sparingly through the Utica to reappear in great numbers in the uppermost Utica and the lower Lorraine (C. pygmaeus to the C. manitoulinensis Zones) as a form which is broader proximally than the typical form, although individuals identical to the typical form also occur with it. The names Diplograptus amplexicaulis mut. uticanus, Glossograptus quadrimucronatus mut. lorrainensis, and Diplograptus recurrens have all been used at the same time by Ruedemann (1925) to describe this late appearance of O. amplexi- caulis (two of the types of D. recurrens, however, are proximal ends of O. quadrimucronatus (Hall)), while Parks (1928) named it Diplograptus montis. This group obviously needs more study. O. amplexicaulis does not occur in the Austin Glen Greywacke as reported by Berry (\962b, p. 714), which is almost entirely in the N. gracilis Zone, or in collections from the Magog slates (Berry 1962u) which are in the N. gracilis- D. multidens Zones. The metamorphism of the Magog slates is so advanced as to render specific identifica- tions of all but the most diagnostic biserial graptolites most precarious. Genus corynoides Nicholson, 1867 The morphology and development of two of the most common forms of Corynoides have been described in some detail by Bulman (1945, 1947), while Strachan (1949) has described the most common European forms of this genus and at the same time attempted to deal with the synonymy of North American forms, but purely on the basis of Ruedemann’s (1947) figures and descriptions. Berry (1960) has fully accepted Strachan’s interpretation of the North American species as have Ross and Berry (1963). The writer has made extensive collections of Corynoides in the Middle Ordovician of north-eastern North America. In general, two (or three) forms of Corynoides occur sparingly in the N. gracilis Zone : the short and stubby C. ‘'ciirtus' mut. pristinus Ruedemann and the longer and slender C. calicidaris Nicholson and C. curtus Lapworth. Mut. pristinus has been recognized only in the N. gracilis Zone and not higher; Strachan (1949, p. 155) considered this form synonymous with the Swedish C. incurvus Hadding, but this assignment is questionable since Ruedemann’s 1947 figures of mut. pristinus are not quite like those of C. incurvus as shown by TEXT-FIG. 10. Corynoides americanus (Ruedermnn). a, Syntype, N.Y.S.M. 6971, with sicula and the begin- ning of th 1. Growth bands shown as imprints in the distal part of the sicula (figured by Ruedemann 1908, pi. 13, fig. 21; 1947, pi. 58, fig. 27); x 10. b, Syntype, N.Y.S.M. 6964, with sicula largely preserved as impression and th 1 extending part way down the sicula (figured by Ruedemann 1908, p. 240, text-fig. 142); X 10. c, Syntype, N.Y.S.M. 6966, with portions of periderm preserved with growth bands. Th 1 extends halfwaydownthesicula(oneofthreespecimensfiguredby Ruedemann 1908, pi. 13, fig. 4); x 10. —«* • ®!.' ■■ , . . , ' .,( ■ ... »'..: ^--w, ■ '% '%'■:■}■ vv .ifiiWin V, •' ..'■ ''I:' H’ ■ V ->* ■ • '.V . >.;■> . -^<'1 " rw;., ( ..^, -.f' .% <• 1 1 \ ,U?- f,' /i 'i ""'" M. m yi- -■ ■> i• -r ■-, . . J'T'vi.K yma ■«JI i ii*^ ". f ■ ■ vco. TWO NEW PALEOGENE DINOFLAGELLATES FROM VIRGINIA AND MARYLAND by DEWEY M. MCLEAN Abstract. Two new species of Hystrichokolpoma (Pyrrhophyta) are described from the Aquia Formation (Upper Paleocene) of Virginia and Maryland. H. tumescens is characterized by a large, bulbous antapical process, and H. mentitwn by a main body which in size, general appearance, and outlines of process bases resembles specimens of Eisenackia. With its delicate processes removed, H. mentiuim is believed to masquerade as Eisenackia. Among the extensive dinoflagellate assemblages of the Atlantic Coastal Plain are numerous new species. The purpose of this paper is to report the discovery of two new species of Hystrichokolpoma (Pyrrhophyta) from the Aquia Formation (Upper Paleocene) of the Virginia-Maryland Coastal Plain. Each is easily recognizable, and, as presently known, is restricted to Paleocene age sediments. One species, H. tumescens sp. nov., displays typical Hystrichokolpoma morphological development, but is characterized by a unique bulbous antapical process. The other, H. mentitum sp. nov., has hollow, box-like processes typical of the genus, but has a main body resemb- ling that of specimens commonly referred to Eisenackia. With the processes broken away, the main body resembles in appearance, size, and tabulation E. circumtabulata Drugg 1967, and E. crassitabulata, as illustrated by Alberti (1961). Samples investigated are from three outcrop localities along the Potomac River south of Washington, D.C. They are: Locality 1. Prince Georges County, Maryland; U.S. Geological Survey Anacostia, Md.-D.C., quad., 7-5 minute series; 38° 45' 10" N., 76° 59' 15" W. Approximately 45 feet (14 m) of lowermost Aquia glauco- nitic quartz sands are exposed about 0-5 mile west of Friendly, Maryland, along the stream occurring immediately south of, and paralleling, the Old Fort Road. Locality 2. Stafford County, Virginia; U.S. Geological Survey Passapatanzy, Va.-Md., quad., 7-5 minute series; 38° 22' 1 5" N., 77° 17' 50" W. This is the type locality of the Aquia Formation. Approximately 70 feet (21 m) of Aquia glauconitic quartz sands are exposed in bluffs along the south bank of Aquia Creek, about 0-5 mile south-east of the Maryland-Virginia Monument No. 37. Locality 3. Stafford County, Virginia; U.S. Geological Survey Passapatanzy, Va.-Md., quad., 7-5 minute series; 38° 20' 35" N., 77° 17' 17" W. Approximately 35 feet (10 m) of Aquia glauconitic quartz sands are exposed in bluffs along the south bank of Potomac Creek, from 0 05 to 015 mile west of the Maryland-Virginia Monument No. 35. Standard acid maceration techniques were employed for all samples. Palynomorphs were concentrated by use of ZnBr (sp. gr. = 2 0), and were darkened for study and photomicrography by acetolysis. Slides are stored at Stanford University and are assigned Stanford University Paleontological Type Col- lection (SUPTC) numbers. Coordinates are measurements in millimetres to the right (R) or left (L) and toward the top ( + ) or bottom ( — ) of the slide from an index cross engraved on the coverslip near its lower left corner. [Palaeontology, Vol. 17, Part 1, 1974, pp. 65-70, pi. 8.] 66 PALAEONTOLOGY, VOLUME 17 SYSTEMATIC PALAEONTOLOGY Division pyrrhophyta Pascher Class DiNOPHYCEAE Pascher Genus hystrichokolpoma Klumpp 1953 Hystrichokolpoma tumescens sp. nov. Plate 8, figs, 6-9 Holotype. Plate 8, figs. 7-8. Loc. 2, sample 3394, SUPTC 10075 (R12-5, 14-7). Diagnosis. A species of Hystrichokolpoma with a large, bulbous antapical process. Description. Cyst main body ellipsoidal, bearing simulate, hollow, box-like processes. Cyst wall bi-layered; processes, formed of the periphragm, reflect tabulation of 4?', Oa, 6", 6c, 6'", Ip, 1"". Archeopyle apical (Type A) with simple, free operculum; corresponds to 4(?) apical plates (see Discussion). Cingulum and sulcus reflected by processes. Cingulum levorotatory, ends separated vertically up to one cingulum width and transversely up to one and one-half cingulum widths. Sulcal area (see Discussion) reflected at top and bottom by anterior and posterior sulcal processes, respectively, and in between by two pairs of small pro- cesses. Of the anterior-most pair the left member is the larger and is designated as the T"; the other pair is between processes 6'" and Ip. Individual processes of the two small process pairs occur either as single spine-like elements, or as several spine-like elements branching from a common base. TEXT-FIG. 1. Comparison of apical archeopyle outline of Hystrichokolpoma tumescens. A, with that of H. rnentitum, B. EXPLANATION OF PLATE 8 All specimens are from the Aquia Formation (Upper Paleocene) of the Virginia-Maryland Coastal Plain, from outcrops along the Potomac River south of Washington, D.C. All photographs are focused on the upper surface of the specimen unless otherwise indicated. Figs. 1-5. Hystrichokolpoma rnentitum sp. nov. Several views of holotype, x625. 1, apical view showing four apical plate-equivalents with an apical pore closing platelet in their midst. 2, sulcal view. 3, antapical view, showing pentagonal base of antapical process. 4, right-lateral view. 5, left-lateral view. Dimensions : LxW ^ 48x45 jum; process lengths up to 15 /nm. Loc. 1, sample 3372, SUPTC 10073 (R4 0, ) 11-4). Fig. 6. Hystrichokolpoma tumescens sp. nov. Ventral view, x640. Note the characteristic large, bulbous antapical process ornamented with longitudinal striations; tip drawn out and broken open. Dimensions: main body L x W = 50 X 38 jum ; antapical process LxW 42 x 27 /xm ; remaining process lengths up to 24 fxm. Loc. 2, sample 3413, SUPTC 10074 (R2L7, -f 16-4). Figs. 7-8. Hystrichokolpoma tumescens sp. nov. Holotype, x 645. Ventral views at slightly dilTerent focus levels. Antapical process has closed tip. Dimensions : main body LxW = 45 x 37 juin ; antapical process LxW =- 37x26 /xm; remaining process lengths up to 25 /xm. Loc. 2, sample 3394, SUPTC 10075 (RI2-5, I 4-7). Fig. 9. Hystrichokolpoma tumescens sp. nov. Dorsal view, x 580. Dimensions: main body LxW=- 55x42 /xin; antapical process LxW - 50x27 /xin; remaining processes up to 27 /xin. Loc. 3, sample 3429, SUPTC 10076 (R3-7, -|-8 7). PLATE 8 7 8 McLEAN, Hystrichokolpoma 68 PALAEONTOLOGY, VOLUME 17 Wall layers appressed except under processes. Endophragm ca. 0-75- 1-25 ;um thick, and smooth to faintly granulose externally. Periphragm ca. 0-5 ftm thick, and smooth to faintly granulose externally. Processes simulate, one per plate-equivalent, hollow and box-like, closed and denticulate at tips, non-branching except for Ip and sulcal processes, externally smooth, faintly striate, or faintly granulose, and may show numerous, tiny longitudinal folds; do not communicate with endocoel. Precingular processes nearly equidimensional except for relatively small 6" process. Processes 2", 4", and 6" have four-sided bases, and processes 1", 3”, and 5" have five-sided bases. Cingular process bases rectangular. Postcingular process bases charac- teristically four-sided; processes 3"', 4"', and 5'" are largest, 6'” is intermediate in size, and 1'" is smallest. Process Ip has an irregular basal outline, and may branch into several spine-like elements. Antapical process bulbous, with pentagonal base and narrow, drawn-out tip which may be broken open. Dimensions. Holotype main body LxW (excluding operculum which is missing): 45x37 (um; processes up to 25 /xm long; antapical process 37 jum long by 26 nm at base. Observed range (twenty-six specimens measured) : main body length (excluding opercula which were universally missing) : 45-62 ;ixm (mean 52 /xm) ; main body transduaneter: 37-53 /xm (mean 42 /xm); length of main body together with antapical process: 76-99 /xm. Discussion. The postcingular tabulation is interpreted as consisting of six plate- equivalents; however, uncertainty exists concerning interpretation of the anterior- most pair of small processes in the sulcal area. If both members of the pair are interpreted as sulcal processes, five postcingulars are indicated. If, however, the left, larger member of the pair is interpreted as a postcingular plate-equivalent, six post- cingulars are indicated. Because thecae of many modern dinoflagellate species display reduced V” plates (personal communication, David Wall, Woods Hole Oceanographic Institute), the latter course was followed in interpreting the fossil. Although no opercula were observed, the archeopyle outline resembles that of Hystrichokolpoma mentitum (below), the operculum of which is simple and free and displays four apical plate-equivalents and an apical pore closing platelet. Each species shows the anterior edges of the 3" and 5" plate-equivalents to be gable shaped, whereas those of the 1", 2", and 4" plate-equivalents are straight; the similarity of the archeo- pyle outlines suggests similarity of the archeopyle tabulation. Studies of several species representing various genera have shown that when the anterior edges of the precingular plate-equivalents are straight, they usually contact only one apical plate- equivalent, whereas when gable shaped, they usually contact two apical plate- equivalents. This relationship is shown in text-fig. 1 which compares the archeopyle outlines of H. turnescens sp. nov. and H. mentitum sp. nov. Comparison with similar species. Hystrichokolpoma turnescens resembles H. rigaudae Deflandre and Cookson 1955, but differs in two respects: (1) the antapical process of the new species has an inflated, bulbous structure and is convex outward along its length, whereas the antapical process of H. rigaudae is outwardly concave and, (2) the cingular processes of the new species do not bifurcate whereas those of H. rigaudae bifurcate near the base into two thin processes per plate-equivalent. Occurrence. Loc. 2, 3% of phytoplankton content throughout basal 10 feet (3 m) of section; Loc. 3, 4% through a 40 foot (12 m) interval beginning 20 feet (6 m) above the base of the section. Hystriehokolpoma mentitum sp. nov. Plate 8, figs. 1-5 Holotype. Plate 8, figs. 1-5. Loc. 1, sample 3372, SUPTC 10073 (R4 0, ! 1 1-7). MCLEAN: PALEOGENE Dl NOF L AGELLATES 69 Diagnosis. A species of Hystrichokolpoma, the main body of which, in size, shape, and tabulation, resembles specimens of Eisenackia. Description. Cyst main body spheroidal to slightly ellipsoidal, bearing simulate, hollow, box-like processes. Cyst wall bi-layered; processes, formed of the periphragm, reflect tabulation of 4', Oa, 6", 6c, 6'", Ip, I"". Archeopyle apical (Type A) with simple, free operculum; corresponds to four apical plates and an apical pore closing platelet. Cingulum and sulcus reflected by processes. Cingulum levorotatory, ends separated vertically about one-half cingulum width and transversely about two cingulum widths. Sulcal area (see Discussion) reflected at top and bottom by anterior and posterior sulcal processes, respectively, and in between by two pairs of small processes. Of the anterior-most pair the left member is the larger and is designated as the T"; the other pair is between the 6"' and Ip processes. Wall layers appressed except under processes. Endophragm ca. 2 0 ij.m thick, and externally smooth. Periphragm ca. 0-5 ^m thick, and externally microreticulate on main body; muri 0-5- TO j^m across. Processes simulate, one per plate-equivalent, hollow and box-like, closed at tips, nearly parallel sided, faintly punctate externally, and do not communicate with endocoel. Precingular processes nearly equidi- mensional except for relatively small 6"; processes 2" and 4" have four-sided bases, and 1", 3", and 5” are five-sided. Cingular process bases rectangular. Postcingular processes variable in size and shape; processes 3'", 4'", and 5'" are largest of series, are nearly equidimensional, and contain distinctive U-shaped partitions, which open upward, in their anterior portions; process 6"' is intermediate in size and T" is the smallest of the series; postcingular process bases are commonly four-sided except for 4"' and 6"' which may be five-sided. Process Ip and the posterior sulcal process bases are four-sided. Antapical process base pentagonal. Dimensions. Holotype main body L x W : 48 x 45 fim ; process lengths up to 1 5 /xm. Observed range (three specimens recovered): main body length 48-50 /xm, width 42-45 ;ixm. Discussion. Although the postcingular tabulation is interpreted as consisting of six plate-equivalents, uncertainty exists concerning interpretation of the anterior-most pair of small processes in the sulcal area. If both members of the pair are considered as sulcal processes, five postcingulars are indicated. If, however, the left, larger mem- ber of the pair is considered a postcingular plate-equivalent, six postcingulars are indicated. Because thecae of many modern dinoflagellate species display reduced V" plates (personal communication, David Wall, Woods Hole Oceanographic Institute), the latter course was followed in interpreting the fossil. Comparison with similar species. The box-like processes and distinctive U-shaped partitions of the postcingular plate-equivalents of the new species differentiate it from H. poculurn Maier 1959 (pp. 312-313, pi. 31, fig. 3), which has cylindrically shaped processes. The new species, with its delicate processes broken away, resembles Eisenackia circumtabiilata Drugg 1967 (p. 15, pi. 1, figs. 12-13), and E. crassitabulata as illustrated by Alberti (1961, p. 32, pi. 3, fig. 19). Occurrence. Loc. 1, less than 1% of the phytoplankton content of one sample 20 feet (6 m) above the base of the Tertiary section; Loc. 2, less than 1% in one sample 17 feet (5 m) above the base of the exposed section. Acknowledgements. Thanks go to Dr. Wm. R. Evitt, Stanford University, and to the National Science Eoundation which sponsored the investigation under Grants GB-4702 and GB-4711 (to Evitt). REFERENCES ALBERTI, G. 1961. Zur Kenntnis mesozoischer and alttertiarer Dinoflagellaten und Hystrichosphaerideen von Nord- und Mitteldeutschland sowie einigen anderen europaischen Gebieten. Palaeontographica, 116, A, 1-58. 70 PALAEONTOLOGY, VOLUME 17 DEFLANDRE, G. and COOKSON, 1. c. 1955. Fossil microplankton from Australian late Mesozoic and Tertiary sediments. Aust. J. Marine Freshwater Res. 6, (2), 243-313. DRUGG, w. s. 1967. Palynology of the upper Moreno Formation (Late Cretaceous-Paleocene) Escarpado Canyon, California. Palaeontographica, 120, B, 1-71. KLUMPP, B. 1953. Beitrag zur kenntnis der Microfossilien des Mittleren und Oberen Eozan. Ibid. 103, A, 377-406. MAiER, D. 1959. Planktonuntersuchungen in tertiaren and quartaren marinen Sedimenten. Ein Beitrag zur Systematik, Stratigraphie und Okologie der Coccolithophorideen, Dinoflagellaten und Hystricho- sphaerideen vom Oligozan bis zun Pleistozan. Neues Jb. Geol. Paldont. Abh. 107, (3), 278-340. D. M. MCLEAN Department of Geological Sciences Virginia Polytechnic Institute and State University Blacksburg Revised typescript received 8 January 1973 Virginia 24061, U.S.A. TRILOBITES FROM THE GORRAN QUARTZITES, ORDOVICIAN OF SOUTH CORNWALL by P. M. SADLER Abstract. The Llandeilo age of the Gorran Quartzites of south Cornwall and the Armorican affinities of the faunas suggested by some earlier authors are confirmed by a study of existing collections and material from two new localities. Revised descriptions and illustrations are given for species of Neseuretus and Kloucekia found in the quartzites of Veryan Bay and Gorran Haven. The presence of a species of Bathycheilus is recorded from there, with additional trinucleid and ogygiocaridinid material. A new species of Crozonaspis, a genus known from the Llandeilo of the Armorican Massif, Spain and Morocco, is described from the quartzites in Gerrans Bay. Since the deformation of the fossils is relatively slight, some primary variation within Neseuretus (N.) tristani seems to be demonstrable. Stratigraphical and structural investigations in Gerrans Bay and Veryan Bay in south Cornwall have indicated the presence there of several features similar to the Palaeozoic geology of the Armorican Massif. They include a confirmation and amplification of the Armorican affinities of the faunas of the Gorran Quartzites, which were first recognized by Salter ( 1 864). A Lower and Middle Devonian sequence (dated by conodont faunas) comprising slate and volcanic rocks succeeded by grey- wacke appears to rest with non-angular unconformity on the Ordovician Gorran Quartzites. Since thrust faults repeat the succession, at most only about 200 m of the Ordovician part is seen. The Ordovician trilobite-brachiopod faunas were the first reliable indication of stratigraphical age available in south Cornwall (Peach 1841; Murchison 1846). This age was at one time supposed to be applicable to the volcanic succession and much of the greywackes (Reid 1907). Peach (1841, 1842, 1844) dis- covered the fossils at Came, Diamond Rock, and Perhaver Beach and recorded the presence of Calyrnene and several orthid brachiopods. Salter (1864, 1865) described species of Calyrnene, Phacops, and Hornalonotus from the Gorran Quartzites for which he proposed Llandeilo (p. 9 of his monograph) and Arenig (pp. 100, 214) ages. He recognized the resemblance of the trilobites to fossils of the ‘May’ sandstone of Normandy, and that one of the Cornish species also occurred in the derived quartzite clasts of the Triassic Pebble Beds at Budleigh Salterton. Peach’s material was redescribed by J. Sowerby (in Murchison 1846, p. 321), M’Coy (in Sedgwick 1852, p. 13) and by Davidson (in Collins 1893). Further material was collected from Perhaver Beach by Mrs. C. Reid during preparation of the Geological Survey, New Series sheet 353 (Reid 1907). These trilobites were determined by Lake (in Reid 1907) who suggested that Cheirurus occurred in the faunas, which were probably of Llandeilo age and to be correlated with the Angus Slates or the Gres de May. Stubblefield (1939fl, 1960) confirmed this probable mid Ordovician date with a revised faunal list for the Gorran Quartzites and the discovery of a trinucleid pygidium in the material from Perhaver Beach. Whittard (1958, p. 115) noted that the specimen referred by Lake to Cheirurus sedgwieki more closely resembled Eeeoptochile clavigera [Palaeontology, Vol. 17, Part 1, 1974, pp. 71-93, pis. 9-10.] 72 PALAEONTOLOGY, VOLUME 17 TEXT-FIG. 1. Generalized geological map of south Cornwall, with out- crop of Gorran Quartzites and fossil localities indicated: 1, Came Beach; 2, Came quarries; 3, Diamond Rock; 4, East Catasuent; (5, Black Rock); 6, Perhaver quarry; 7, Perhaver Beach. (Beyrich), but advised that other fossils from the Perhaver fauna did not support a Llanvirn-Llandeilo age. Of the major quartzite outcrops, only Black Rock has failed to yield fossils. Diamond Rock (SW 9754 4083) and Perhaver Quarry (SX 0142 4230, rock face no longer available) have each yielded brachiopods and a single trilobite pygidium, while Came quarries (SW 9127 3820) have yielded brachiopods and the only record of an articulated trilobite thorax and pygidium (Fox 1908). The main Peach and Reid collections are from huge fallen blocks below the quartzite outcrops of Great Perhaver Beach (SX 0174 4234). Recent searches there have produced only poorly preserved material, but it has been possible to collect a similar fauna from a new site, in situ immediately east of Catasuent Cove (SW 9765 4076). A different, apparently monospecific, fauna has been collected from a large detached block below the quartzite outcrops of Came Beach (SW 9109 3793). The Truro city museum contains a typical pygidium of Basilicus fyraT/nwj' (Murchison), found by C. J. Lane in a sandstone block in the wall of Elerky House, Veryan (SW 9139 3949), but whereas the matrix is unlike the Gorran Quartzites, it closely resembles that of the Llandeilo Flags near Llandeilo itself (pers. comm. A. W. A. Rushton). The trilobites occur with abundant brachiopods in rare layers as ferruginous moulds, usually concentrated toward the base of laminated, and occasionally cross- laminated, quartz-sand beds. This paper describes some of the trilobites in the earlier collections as well as the author’s own material. The brachiopods have been deposited, without detailed determinations, in the collections of the Institute of Geological Sciences, London. The Devonian conodont faunas and the stratigraphy will be treated in later papers. The trilobites identified are; Came Beach: Cwzonaspis peachi n. sp. East Catasuent: Neseuretus (N.) tristani (Brongniart), Neseuretus (N.) cf. N. compUnuitus Whittard, Bathychcilus n. sp. aft'. B. perplexus (Barrande), Kloucekia (K.) niimus (Salter), Trinucleidae gen. indet., Ogygiocaridinae gen. indet. SADLER: CORNISH ORDOVICIAN TRILOBITES 73 Perhaver Beach: Neseuretus (TV.) tristani (Brongniart), Bathycheilus n. sp. aff. B. perplexus (Barrande), Kloucekia (K.) mimus (Salter), Homalonotus sp., Eccoplochile clavigera (Beyrich)? (TWhittard 1958), Trinucleidae gen. indet. (Stubblefield 1939a, p. 68), Ogygiocaridinae gen. indet. Age and Correlation. Since previously described species of Crozonaspis (Henry 1968; Clarkson and Henry 1970; Carre et al. 1972; Destombes 1972) are confined to the Llandeilo, it is probable that the Came Beach fauna is also of Llandeilo age. Neseuretus (N.) tristani occurs in the Llanvirn and Llandeilo of Brittany (Nion and Henry 1966, p. 885; Clarkson and Henry 1970, pp. 119-121). It is especially abundant near the Llanvirn-Llandeilo boundary, and again in the upper Llandeilo, but it disappears abruptly before the Caradoc. Kloucekia is recorded as beginning in the Llandeilo and extending through the upper Ordovician (Destombes 1972, p. 18), so its associa- tion with N. [N.) tristani in the Catasuent and Perhaver faunas again suggests a Llandeilo age. This means that the Gorran Quartzites, originally correlated by Salter (1864) with the ‘May’ Sandstone of Normandy can now more precisely be correlated with the gres de May inferieur (‘Petit May’) of Normandy and the gres de Kerarvail and schistes de Morgat of Finistere (Henry 1970, p. 23). In those areas there is a stratigraphic break between the Caradoc or Llandovery below and the Upper Silurian or Lower Devonian above (Renaud et al. 1968). The faunas of the Gorran Quartzites are interesting in the context of the distribu- tion of Ordovician trilobites. They contrast strongly with the contemporaneous Anglo- Welsh faunas which are dominated by trinucleid and asaphid trilobites (Stubblefield 19396, p. 55; Dean, in Whittard 1967, p. 314). An abundance of Neseuretus during TEXT-FIG. 2. Distribution of Neseuretus (Neseuretus) tristani. Inset: distribution of genus Neseuretus. 74 PALAEONTOLOGY, VOLUME 17 the Arenig characterizes the Welsh area (Whittard 1960; Whittington 1966a; Bates 1968, 1969) and the Montagne Noire (Dean 1966), where the genus is unknown above the lower Llanvirn. A Llanvirn-Llandeilo abundance of Neseuretus — usually referred to N. (N.) /ra/anz— characterizes Cornwall and the Armorican Massif, Spain, Portugal, and North Africa (Born 1916; Spjeldnaes 1961 ; Whittington \966b). In the Cornish faunas trinucleid and asaphid trilobites are a minor element only. The genus Crozonaspis seems to have had a distribution similar to that of N. (N.) tristani. During Llanvirn times, Anglo- Welsh, German (Rheinisches Schiefergebirge ; Siegfried 1969), and Bohemian (Dean, in Whittard 1967, p. 310) trilobite faunas were closely related, but this link was not again clearly in evidence until mid Caradoc time, when Kloucekia appeared in Welsh faunas (Dean 1961 ; Whittington 1962). Kloucekia is known from the Llandeilo in Bohemia (Henry 1965, p. 205), during which time it was present in Brittany and Cornwall also. The record of Bathycheilus commences in the Arenigian of the Montagne Noire (Dean 1965, 1966) and extends to Bohemia, Portugal (Thadeu 1956), and apparently Cornwall, in the later lower Ordovician. Bathycheilus is not known in Brittany, but the related genus Priofiocheilus is recorded there (Rouault 1847). The presence of Eccoptochile in the Perhaver fauna (Whittard 1958, p. 115) provides further indication of Bohemian affinity. The possibility of a link between the Welsh and Armorican Massif faunas exists in the highest Llandeilo when Marrolithus occurs in both areas (Stubblefield 1939b, p. 54). This is apparently the earliest appearance of Trinucleidae in the Armorican region (Oehlert 1895; Nion and Henry 1966). Pebbles of Ordovician Quartzites of similar appearance to the Gorran Quartzites and the gres de May and gres de Kerarvail occur in the Triassic (Warrington 1971, p. 312) Budleigh Salterton Pebble Beds, and Crozonaspis (Henry, in Carre et al. 1972, p. 781) and Neseuretus {N.) tristani (Salter 1864) have been found. SYSTEMATIC DESCRIPTIONS All the specimens treated here are deposited in the collection of the Institute of Geological Sciences, London (GSM). Suffixed letters distinguish particular moulds on a single numbered block. All descriptive terminology is from the Treatise (Moore 1959) unless otherwise stated. The use of open nomenclature and annotated synonymy lists follows the scheme proposed by R. Richter (1948). Family asaphidae Burmeister, 1843 Subfamily ogygiocaridinae Raymond, 1937 Genus and species indet. Plate 10, figs. U2 Material. East Catasuent— 1 librigena and 2 incomplete pygidia. Perhaver Beach— 2 pygidia, coll. Reid 1907 (GSM CR 1545-1548 and 1498). Remarks. Rare, large fragments which occur are not considered adequate for generic recognition, since generic differences in this subfamily are rather subtle and in some respects controversial. The pygidia are of two types. One (PI. 10, fig. 2) from East Catasuent resembles Ogygiocarella Harrington and Leanza 1957. The pleural lobes bear deep asymmetrical interpleural furrows, near the distal ends of which there are faint, divergent pleural furrows developed on the anterior part of the pleura. The SADLER: CORNISH ORDOVICIAN TRILOBITES 75 wide border, which has no furrow, bears very slight trace of concentric striation. The second type (not illustrated) resembles Ogygiocaris Briinnich 1781 and has been found at East Catasuent (GSM Zs 589) and Perhaver Beach. These pygidia are semi- circular, with a narrow, tapering, feebly segmented axis which terminates against a wide, smooth border. The pleural lobes bear faint furrows adaxially. Family trinucleidae Hawle and Corda, 1847 Genus and species indet. Plate 10, figs. 3-5 Material. East Catasuent— 7 cephala. Perhaver Beach— 1 pygidium (Stubblefield 1939fl, p. 68). Description. Cephalon with semicircular outline, about two-thirds as long as broad, except for PI. 10, fig. 5 which is only one-third as long as broad but probably compressed axially. (The available material is insufficient and too dispersed for a proper assessment of the contribution of tectonic deformation, but this individual could be a W-form (Henningsmoen 1960, p. 207) in a strained domain with a shortening of at least 1 :0-75.) Pear-shaped glabella with three, paired, lateral furrows arranged as in Trinucleus Murchison 1839. Ip and 2p lateral furrows short, deep, triangular notches separated by a narrow (exsag.) rectangular 2p lobe, which is not isolated. 3p furrows very short, inconspicuous indentations. Swollen frontal glabellar lobe continuous with 3p lobes, high and rounded, but does not overhang or transgress on to anterior fringe. Median glabellar tubercle (Whittington 1968, p. 703) sited between 3p furrows. Axial furrows very deep, narrowing and diverging forward. Occipital furrow shallow, curved, convex backward. Occipital ring narrow (sag.), without spine. Cheek lobes convex, quadrant shaped, less elevated than glabella. Lateral margins fall steeply to fringe. Abaxial posterior corner of cheek lobes pointed and directed backward to constrict posterior border furrow. Eye-tubercles a little away from brink of axial furrows, opposite (tr.) median glabellar tubercle (PI. 10, fig. 4). Posterior border furrow wide, abruptly constricted abaxially at a point beyond widest and deepest portion. Posterior border a low ridge. Pitted fringe (insufficiently complete for generic determination) appears to be narrow ( 1 mm in GSM Zs 577), comprising at least three arcs anteriorly, and is apparently unexpanded at the anterolateral angle. Only the innermost arc is satisfactorily seen showing at least fifteen pits in the half-arc; it seems to comprise the largest and deepest pits and is not transgressed by the frontal glabellar lobe. Family synhomalonotidae Kobayashi, 1960 Genus neseuretus Hicks, 1872 Type species. Neseuretus ramseyerisis Hicks, 1872. Subsequent designation by Vogdes (1925, p. 106). Whittard (1960, pp. 138-139) has argued that Synhomalonotus Pompeckj 1898 is a junior subjective synonym of Neseuretus Hicks and has given a diagnosis for the genus. This account accepts both Whittard’s diagnosis and his view of the status of Synhomalonotus. Whittard (1960, p. 140) designated N. murehisoni (Salter) as the type species— i.e. Calymene parvifronswar. murehisoni Salter, which is not synonymous with N. parvifrons as Bates (1969, p. 22) implies— believing it to be a senior synonym of N. ramseyensis. However, Bates (1969) has shown that they are not conspecific. Bates did not, unfortunately, specify the difference between N. ramseyensis and N. tristani (Brongniart), the type species proposed for Synhomalonotus by Pompeckj (1898), which is regrettable since the two are apparently to be distinguished from N. murehisoni by the same criteria — compare Whittard 1960, p. 149, and Bates 1969, p. 25. Sdzuy (1957, p. 277) has argued that Neseuretus (= Synhomalonotus) should be 76 PALAEONTOLOGY, VOLUME 17 transferred from the Calymenidae Milne-Edwards 1840 (where it was placed by, among others, Shirley 1936, p. 394) to the Homalonotidae Chapman 1890, because of the form of the glabella and anterior area (Whittard 1960, p. 143, for definition of anterior area). Sdzuy’s practice, which acknowledges the similarity between Neseuretus and Calymenella, was followed by Bates (1968, 1969). On ventral morphologies, Whittington ( 1 966a) considered Neseuretus to be a calymenid which had an anomalous hypostome associated with the relatively long preglabellar field, a proposal accepted by Henry (1970). Whittington ( 1 966a) and Whittard ( 1 960), who also placed Neseuretus in the Calymenidae, nevertheless considered that Neseuretus must belong to a stock from which the later calymenids and homalonotids were derived but not precisely referable to either family. A separate family is preferred here, following Dean (1966, p. 297; 1971, p. 9). Stratigraphical range. Arenig and Lower Llanvirn of Wales (Whittington 1966a; Bates 1968, 1969) and the Welsh Borderland (Whittard 1960), and the Arenig of the Montagne Noire (Dean 1966) and Turkey (Dean 1971). Widespread in Llanvirn and Llandeilo of Brittany (Henry 1970), the Iberian peninsula (Born 1916), and North Africa. Also recorded in Burma, central and west China (Reed 1915; Kobayashi 1951; Lu 1957), and Argentina (Harrington and Leanza 1957). Not recorded from the Ordovician of Bohemia, Scandinavia, or North America (text-fig. 2). No Caradoc record. Remarks. The longer anterior area of Neseuretus distinguishes it from Platyealymene (Llanvirn-Caradoc) and Flexicalymene (Llandeilo-Ashgill). Calymenella may closely resemble Neseuretus in the form of its anterior area and may also develop faint oblique eye ridges, but it has less distinct glabellar furrows. The two genera have very similar pygidial segmentation, but the last axial ring of Calymenella does not reach the posterior margin of the pygidium. Although Arenig occurrences of Neseuretus have been divided into several well-defined species, with emphasis on the form of the anterior area, almost all Llandeilo occurrences of the genus in Europe (and they are rather variable) appear to have been indiscriminately assigned to the inadequately defined species N. tristani (Brongniart). Subgenus neseuretus (neseuretus) Hicks, 1872 Dean (1967) has proposed two subgenera. N. {Neseuretinus) Dean differs from the nominate subgenus by a distinctly convex preglabellar field between an enlarged and arched anterior border and the frontal glabellar lobe. Apart from N. hirmanicus Reed all species described prior to 1967 belong to N. (Neseuretus). Neseuretus (Neseuretus) tristani (Brongniart, 1822) Plate 9, figs. 4-12; text-fig. 3 1808 Tristan and Bigot, p. 21. *1822 Calyniene T’rw/at;/ Brongniart, pp. 12-14, pi. I, fig. 2f-i, non 2a-e (= Colpocoryphe aragoi (Roualult)), Inon 2k. vV846 Calyniene pulchella; Sowerby in Murchison, p. 231. vV852 Calyniene parvifrons Salter? (?M’Coy); M’Coy in Sedgwick, p. 13. vl864^/ Calyniene Tristani Brongniart; Salter, p. 291, pi. 15, fig. 5. vl864 Calyniene Tristani Brongniart; Salter, pp. 99-100, pi. 9, figs. 15 18. SADLER: CORNISH ORDOVICIAN TRILOBITES 77 vV893 Calyinene parvifrons Salter; Collins, p. 470. vV893 Calymene Tristcmi Brongniart; Collins, p. 471. \I907 Calymene Tristan! Brongniart; Lake in Reid, p. 39. V 1939a Synhomalonotns; Stubblefield, p. 68. vl960 Neseuretus; Stubblefield, p. 102. 1970 Neseuretus (Neseuretus) tristani (Brongniart); Henry, pp. 5-11, pi. a, figs. 1-10. Type specimens. Brongniart’s specimens appear to be lost, and Henry (1970) has selected and figured a neotype from Hunaudiere, the type locality. Material. East Catasuent— 48 cranidia (including 17 representatives of variant two, 23 of variant three, 1 of ?variant one, 7 very incomplete), 10 librigenae, and 19 pygidia. Perhaver Beach— abundant disarticu- lated material in Reid collection. Diagnosis. A species of Neseuretus (A.) characterized by a pronounced, arched (sag.) anterior area which, in longitudinal profile, slopes distinctly and continuously downward and backward to the preglabellar furrow (text-fig. 6). The anterior area may or may not possess a weak anterior border furrow. The posterior and median lateral glabellar furrows are distinct and wide (exsag.), but the anterior lateral glabellar furrow may be very inconspicuous. In longitudinal profile the glabella is moderately convex, sloping gently to the preglabellar furrow. TEXT-FIG, 3. Reconstruction of Neseuretus (Neseuretus) tristani. a-h. variant 2 based on Zs 590a; c d. variant 3 based on Zs 609a. X 1-7. Description. Both Brongniart (1822, p. 13) and Henry (1970, p. 1 1 ) mention the varied nature of the anterior area in N. (N.) tristani, attributing this partly to compaction and deformation. However, some of the differences among previously figured specimens — such as the presence or absence of an anterior border furrow— are probably primary. Henry notes that in Brittany individuals from siliceous nodules in the slates are almost undeformed and do not show much variation. The material from the Cornish Quartzites is little deformed. Flattening is rarely evident, even in the especially vulnerable, long and very convex (tr.) posterior areas of the fixigenae, and angular shear strain is usually low and its effects obvious (text-fig. 4). Occasion- ally the anterior area is crushed backward on to the glabella but the accommodating fractures in the exoskeleton are clearly preserved on the moulds (PI. 9, fig. 8). Examination of the Cornish material shows three variants among cranidia of N. (N.) tristani'. Variant one (?P1. 9, fig. 8). Brongniart’s (1822, PI. I, fig. 2g) illustration of one of the syntypes shows an anterior border furrow and a relatively short, strongly convex glabella. Only one individual from Cornwall has an anterior border furrow. This furrow is present in individuals from Spain (Born 1916, pi. 26, figs. 46-c; 1953, pi. 7, figs. 3-4) and is weakly developed in rather squat, strongly convex cephalic shields from Brittany (Henry 1970, text-fig. 1, pi. A, figs. 1-3). Whittard (1960, p. 145) regarded the presence of an anterior border furrow as essential to N. (N.) tristani, and variant one, which includes both one of the syntypes and the neotype, could be regarded as N. (N.) tristani sensu stricto. A marked anterior border furrow and an approximately square, very convex glabella are found in N. (N.) murchisoni, but in the latter superficially similar species part of the anterior area slopes forward (text-fig. 6). With one exception, the Cornish specimens examined lack an anterior border furrow. They would therefore meet the specification of Whittard’s (1960, p. 145) working definition of N. (N.) parvifrons, but are placed under N. (N.) tristani here because of the slope of the anterior area. Subjectively, the material can readily be cast into two groups 1.0 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 I 1 1 1 1 I I I 1 Strain-Ellipse Axis Ratio at 0=45° (i.e. least extreme strain-ellipse) TEXT-FIG. 4. Shear strain and flattening in the Gorran Quartzite faunas, a, shear strain in total faunas from Came and Catasuent; h, flattening in N. {N.) tristani from East Catasuent, dots variant 2, crosses variant 3. EXPLANATION OE PLATE 9 All specimens dusted with ammonium chloride. Indian ink backgrounds have been added. Bathycheilus n. sp. aff. perplexus (Barrande). Fig. 1. GSM CR 19599, cranidium, latex cast from external mould, Perhaver, x 3-5. Fig. 2a-b. GSM Zs 609d, cranidium, internal mould, Catasuent, X 3. a, frontal view, showing right palpebral lobe; />, dorsal view. Fig. 3. GSM CR 1 556, cranidium, showing left paraglabellar area, latex cast from external mould, Perhaver, x3-5. Neseuretus (Neseuretus) tristani (Brongniart), all from Catasuent. Fig. 4. GSM Zs 609b, var. 3 cranidium, latex cast from external mould (pits in occipital ring are flaws in cast), X 3-5. Fig. 5. GSM Zs 61 la, var. 2 cranidium, internal mould, showing anterior pits and occipital apodemes, x3-5. Fig. 6. GSM Zs 631, pygidium, x3. Fig. 7. GSM Zs 609a, var. 3 cranidium, internal mould, x3. Fig. 8. GSM Zs614, ?var. I, large crushed cranidium (note fractures and telescoped preglabellar furrow), internal mould, X 3-5. Fig. 9. GSM Zs 613, glabella and anterior area, latex cast from external mould, x 3. Fig. \0a-h. GSM Zs 590a, var. 2 cranidium, internal mould, X 3-5. Fig. 1 1 . GSM Zs 632, pygidium, x 3. Fig. 12. GSM Zs 609c, var. 2 cranidium, latex cast from external mould, x3-5. Neseuretus (Neseuretus) cf. compUmatus Whittard. Fig. 13i:/ h. GSM Zs 590b, cranidium, internal mould, Catasuent, X 3-5. u, dorsal view; h, slightly oblique lateral view. PLATE 9 SADLER, Cornish Ordovician trilobites 80 PALAEONTOLOGY, VOLUME 17 according to the shape of the glabella and anterior area, and measurement seems to confirm the existence of two discrete types (text-fig. 5); these are here termed variants two and three. Variant two (PI. 9, figs. 5, 10, 12) is the slimmer of the two types that dominate the Cornish material and may also include similar material from Spain (Chauvel et al. 1969, pi. 4, figs. 1, la). Cranidium sub-triangular, about two-thirds as long as broad and strongly convex transversely. Posterior edge nearly straight but rounded genal angles may project slightly backward. Gonatoparian. Glabella sub- rectangular, longer than wide, limited by broad, conspicuous axial furrows, approximately parallel from occipital ring to 2p lobe, then converging forward at about 30°. Three pairs of glabellar furrows. Ip furrows short, deep, wide (exsag.), increasing in width adaxially and curved or angled backward at about 50° from junction with axial furrows. 2p furrows broad (exsag.), shallower, very slightly oblique, also intersect axial furrows. 3p furrows barely perceptible. Ip lobes convex and axe-shaped. 2p lobes flatter, widen slightly toward axis. 3p and frontal lobes rarely separable, truncate outline. Fossulae shallow. Central area of glabella fairly flat. Occipital furrow broadest (sag.) and shallowest axially; deep pits in the axial furrows adjacent to the occipital ring of internal moulds (PI. 1, figs. 5, 106) are not seen on external casts and represent ventral articulating bosses or apodemes without external expression. A large anterior area lies between two shallow, divergent, arcuate anterior furrows (Whittard 1960, p. 139) which curve outward through more than 90°. Anterior area flat to very gently convex, sloping distinctly and gently backward to preglabellar furrow. Anterior margin unfurrowed, moderately tightly curved to slightly pointed and inclined at 50-60° to horizontal in side view. Fixigenae narrow (tr.) adjacent to glabella, extend posteriorly as markedly down- SAGiriAL LINE I — I — . — — . — ' ' I 5 10 Glabellar Length mm £5-1 TRANSVERSE LINE IG-&1 ' ' ' 5 Glabellar Width 10 mm TEXT-FIG. 5. Differentiation of TV. (A.) tristani variant 2 and variant 3 according to shape: dots = variant 2, crosses = variant 3, open symbols = previously figured examples, a, shape of cephalic axis, uncorrected for deformation, inset replots Cornish data as a cumulative frequency; b and c, ratios from a single line, independent of deformation, b— along sagittal line, c— along transverse line through midpoints of palpebral lobes. SADLER: CORNISH ORDOVICIAN TRILOBITES 81 curved wings with distinct, broad, posterior furrows. Eyes opposite 2p lobes, with narrow (tr.) palpebral lobes. Eye ridges barely perceptible. Visual surface of eyes unknown. Variant three (PI. 9, figs. 4, 7). Cranidium relatively short, only half as long as broad, strongly convex transversely (text-fig. 4h). Posterior margin approximately straight but posterior portion of fixigenae may be directed slightly forward. Gonatoparian. Cephalic axis (glabella with occipital ring) almost square; broad axial furrows converge at 30-40° in front of Ip furrow. Three paired glabellar furrows positioned as in variant two, but narrower, more sharply incised. 3p furrows rarely seen. Glabellar lobes tend to be flatter. The furrows are similar to Calymenella (Sdzuy 1957, pi. 1, figs. 2, 5) but much more distinct. Glabellar axis somewhat more convex (sag.) than in variant two. Pits adjacent to occipital ring in internal moulds only. Wider anterior area lies between weak, rapidly divergent anterior furrows. Anterior area profile (sag.) more convex, shorter than in variant two, but without forward sloping regions such as are seen in N. (N.) parvifrons and N. (N.) monensis. Anterior margin unfurrowed, describing a very flat curve and inclined at 70-80° to horizontal in side view. Shallow semicircular depressions adjacent to Ip lobes but no distinct paraglabellar lobes. Since the curvature (tr.) of the convex posterior margin shows the same range of variation in variants two and three (text-fig. 4) the latter variant can scarcely be the result of flattening of the former. Although a few individuals show some intermediate character (e.g. PI. 9, fig. 5— variant two proportions with rather narrow (exsag.) glabellar furrows) measurement of glabellar proportions indicates that in the available Cornish material no continuous variation exists between variants two and three (text-fig. 5). Since the moulds have no preferred orientation on the bedding planes, the frequency distribution of angular shear strain (text-fig. 4) indicates that strain is variable but usually only slight. These low shear strain values and the distinct lack of intermediate shapes, whether in symmetrical or asymmetrical individuals, suggests that variants two and three are not simply L- and W-forms (Elenningsmoen 1960, p. 207) produced from a single original shape by transverse and sagittal compression respectively. If variant three had been produced by sagittal compression it ought to show consistently less transverse convexity than forms generated by transverse compression and possess relatively accentuated glabellar lobes. Neither criterion is met by the Cornish material. A single block (GSM Zs 609) bearing eleven cranidia (including PI. 9, figs. 2, 4, 7, and 12) on a 25 sq cm bedding plane offers the best means of assessing the significance of L- and W-forms. Three variant three individuals (including PI. 9, figs. 4 and 7) occur in immediate juxtaposition, with their axial directions at angles of 88° and 59° to one another, while a distinct variant two cranidium (PI. 9, fig. 12) lies with its transverse direction divergent from that of one of the former by only 21°. L- and W-forms should be oriented at approximately right angles to one another in regions of homogeneous strain . U sing Wellman’s method (1962) the best-fit strain ellipse for this bedding surface has an axial ratio of 1 :0-92. A strain ratio of 1 :0-58 produces typical symmetrical examples of variants two and three (PI. 9, figs. 10 and 7) from a single O-form, and a strain ratio exceeding 1:0-81 is necessary to generate the mean values of glabellar shape for oblique or asymmetrical variant two and three forms, even if at right angles to one another. The strain is not fully homogeneous even over this small area. The Wellman diagram shows that one variant two individual (PI. 9, fig. 12) accommodates an anomalously high strain which can be calculated as an ellipse axis ratio of 1 :0-78. However, to produce this individual and the nearest variant two from a common O-form requires a strain ratio more extreme than 1 :0-56 (the theoretical value for equivalent perpendicular shapes). Primary variation is also indicated by ratios of parallel lengths which should be unaltered by deformation but show consistent differences between variants two and three. Variant two has consistently larger ratios of fixigena width to glabellar width (opposite 2p furrows) and anterior area length (sag.) to glabellar length (text-fig. 5b and c). Apart from the possibility of locally very inhomogeneous strain, these lines of evidence indicate that variants two and three, although superficially compatible with the products of tectonic deformation, are based on a primary, discontinuous variation. Specimens figured by Chauvel el al. (1969) are larger, but plot unambiguously with variant two (text-fig. 5). However, pre- viously illustrated material with an anterior border furrow (listed above as variant one) plots close to variant three, and Henry’s (1970) material seems to indicate that continuous variation may exist between variants one and three. Eurther study of relatively undeformed material may show that variant two deserves formal separation from N. (N.) tristani sensu stricto. Eibrigenae. Only disarticulated examples found. Convex, triangular. Upturned flange at apex bordering eye. Broad lateral border furrow and pronounced raised lateral border which extends into rounded genal angle. If placed in apposition to facial suture librigena would be steeply inclined to horizontal. PALAEONTOLOGY, VOLUME 17 Pygidia. Small, triangular, about twice as wide as long. Deep axial furrows. Marked anterolateral bevels on pleural field, tapering backward. Posterior margin chevron-shaped. Segmentation less distinct toward posterior, up to ten axial rings distinguishable. Five to six pleural furrows. Shallow pleural furrows at distal ends of pleura. Remarks. Although there may be close resemblance to N. {N.) murchisoni or N. {N.) parvifrons, N. (N.) tristani can be separated from these and most other species, as Henry (1970) has shown, by the rearward sloping profile (sag.) of the anterior area (text-fig. 6). In N. {N.) antetristani Dean 1966 the glabellar furrows are much less distinct and the relatively narrow frontal lobe falls abruptly to the preglabellar furrow. N.(N.) tristani N.(N.) murchisoni 1 Whittington 1966a ) N.(N.) ramseyensis ( Bates 1969) A/. (N.) parvifrons IWhittard 1960) I Salter 1861. ) N.(N.) arenosus N (N.) attenuatus N.(N.) antetristani H(N.)monensis NM) brevisulcus N.(N.) bullatus N.(N.) complanatus (Shirley 1936) (Whittard I960) ( Whittard 1960 ) (Whittard 1960) TEXT-FIG. 6. Comparison of longitudinal cephalic profiles within the subgenus Neseuretiis {Neseuretus). Distribution. (Text-fig. 2.) N. (N.) tristani is common in the Llanvirn and Llandeilo of Normandy and Brittany, the Pyrenees, Portugal, Spain, and Morocco. It is absent from the Llanvirn and Llandeilo of Wales and the Welsh Borderland, although the genus is well represented in the Arenig there. Its appearance in the Llanvirn of Brittany is rather abrupt. This may relate to the coincident rapid transition from the quartzites of the Gres Armorican to the Courijou or Kerloc’h slates. N. (N.) tristani has been recorded from derived quartzites in the Budleigh Salterton Pebble Beds (Salter 1864«). SADLER: CORNISH ORDOVICIAN TRILOBITES 83 Neseuretus (Neseuretus) cf. N. complanatus Whittard, 1960 Plate 9, figs. \l>a-h Material. East Catasuent — 1 incomplete cranidium. Description. Narrow, conspicuously tapering glabella, limited by axial furrows converging at 25°. Occipital furrow deep adjacent to axial furrows, crossing axis as a shallow trough. Ip furrows short and deep, angled backward, defining axe-shaped Ip lobes, 2p furrows shallower and very short. 3p furrows not present. Preglabellar furrow faint, with very shallow fossulae. Glabellar profile (sag.) very low, only slightly higher than anterior portions of fixigenae, which are long, inflated, and relatively wide. Posterior portions of fixigenae not seen. Anterior portion of sutures fairly straight, only slightly convergent forwards. Anterior area about one-third glabellar length, slightly lower than glabella and slightly convex. Anterior margin a flat curve. Discussion. Whittard ( 1 960, p. 1 47) erected the species N. complanatus to accommodate two cranidia from the Mytton Flags fauna (Arenigian, Shelve inlier) which had an unusually flat profile (sag.), a narrow glabella without 3p furrows, and inflated, wide (tr.), parallel-sided anterior portions of the fixigenae. The incomplete cranidium (described above) incorporates comparable features, but lacks the somewhat elongate anterior area of the holotype, and has a more pointed frontal lobe. The species has nowhere been recorded in large numbers. Family bathycheilidae Pfibyl, 1953 Diagnosis. An emended diagnosis for this family, to include two genera— Bathycheihis Holub and Prionocheilus Rouault (= Pharostonia, see Dean 1964) — is given by Dean (1965, p. 1). Genus bathycheilus Holub, 1908 Type-species. Dalmanites perplexus Barrande 1872. By original designation. Diagnosis. Semicircular cephalic outline. Pronounced anterior border. Anteriorly tapering glabella with three pairs of furrows and lobes which diminish in size forward. Distinct paraglabellar areas. Strongly convex anterior portions of fixigenae exceed frontal lobe in height and extend farther forward. Posterior margin of cephalon concave backward. Pronounced posterior border. Opisthoparian. Librigenal spines. Distribution. Arenig of Montagne Noire (Dean 1965, 1966), Llanvirn of Bohemia (Barrande 1872; Elolub 1908), Llanvirn or Llandeilo of Portugal (Thadeu 1956) and possibly Morocco (Gigout 1951). Bathycheilus n. sp. ah'. B. perplexus (Barrande, 1872) Plate 9, figs. 1-3; text-fig. 7 Material. East Catasuent— 14 cranidia, 4 librigenae, 1 pygidium (very tentatively associated). Perhaver Beach— 7 cranidia in Reid collection. Description. Cephalon sub-semicircular, frontal margin describing a shallow curve. Maximum width twice median length. Narrow glabella comprises about two-thirds cephalic length (sag.) and one-third maximum width (across occipital ring). Glabellar margins converge forward at 30 40°. Three pairs of short glabellar lobes and furrows leaving a smooth central area. Ip furrows deep and broad (exsag.), curving inward and backward to delimit an anterolaterally rounded Ip lobe which increases in size abaxially. 2p furrows short, deep, and very slightly oblique. 3p furrows are very short, straight notches. Frontal lobe short and steeply 84 PALAEONTOLOGY, VOLUME 17 rounded. Occipital furrow deeply notched laterally, continuing across axis as a shallow, slightly curved trough. Occipital ring slightly sinuous, parallel-sided except at posterolateral extremities where, in internal moulds, ring narrows to accommodate tear-shaped apodemal pits (PI. 9, fig. 2b). Axial furrows are broad anteriorly but narrow adjacent to 2p lobes. Adjacent to Ip lobes, axial furrows widen markedly toward the posterior border furrow, producing depressed triangular areas in which a small, sub-circular paraglabellar lobe is situated (PI. 9, fig. 3). Small, low, gently concave preglabellar field with upturned anterior border continuing anterolaterally as a rim. Fixigenae include a ‘buttress’-like structure adjacent to the 2p lobe. TEXT-FIG. 7. Reconstruction of cephalon of Bathycheilus n. sp. aff. B. perplexus. a, dorsal; b, frontal; c, lateral. Approx. x2-3. Their anterior portions are conspicuously inflated, exceeding the glabella in height anteriorly and extend- ing forwards beyond the frontal lobe. Palpebral lobes sited well back, immediately behind Ip furrows, and stand up sharply from the cheek surface (PI. 9, fig. 2a). Visual surface not known. Posterior portions of fixigenae narrow (exsag.), having a pronounced border and curving markedly rearward. Suture opisthoparian, intersects posterior margin just inside genal angle, runs abaxially from close behind palpebral lobes. Librigenae relatively narrow (tr.) bearing long genal spines. Slight flange below eye. Hypostome and thorax not known. Pygidium (very tentatively associated) approximately as long as wide. Ten axial rings and tail-piece. Pleural lobes declined abaxially with six furrows. Discussion. The general style of the cranidium is very similar to B. perplexus (refigured by Dean 1965, pi. I). The segmentation of the glabella, development of paraglabellar areas, and the inflation of the fixigenae are closely comparable and the two species present almost identical frontal aspects (compare PI. 9, fig. 2a with Dean 1965, pi. I, fig. 5). However, the Cornish species has a distinctly narrower glabella with a less blunt frontal lobe. Their anterior borders are similar but in B. perplexus the preglabellar field is shorter (sag.). The full length of the genal spines of the Cornish species is not seen, but they are probably shorter and certainly less stout. In B. gallicus Dean 1965, the anterior portions of the axial furrows are more deeply incised and the anterior margin distinctly arched. The related genus Prionocheilus has a charac- teristic posterior constriction of the axial furrows behind the paraglabellar areas, and its less stout librigenal spines carry a row of short spines on their undersides. Family dalmanitidae Vogdes, 1890 Subfamily kloucekiinae Destombes, 1972 Genus kloucekia Delo, 1935 Type-.'ipecie.s. Phacops phillipsi Barrande. Designation by Delo (1935, p. 408). Discussion. Destombes (1972, p. 56) supplies a diagnosis of the genus and a dis- cussion of its taxonomic position. Following Dean (1961, p. 321) two subgenera are recognized. Kloucekia (Phaeopidina) is distinguished from the nominate subgenus by its mucronate pygidium. The Cornish material confirms the opinion of Whittington (1962, p. 7) and Dean (1961, p. 321) that small fixigenal spines occur at a subspecific level of variation. SADLER: CORNISH ORDOVICIAN TRILOBITES 85 Subgenus kloucekia (kloucekia) Delo, 1935 Kloucekia {Kloucekia) mimus (Salter, 1864) Plate 10, figs. 6-10; text-fig. 8 v*1864 Phacops (Acaste) mimus Salter, pp. 29-30, pi. I, fig. 35. vl893 Phacops mimus Salter; Collins, p. 472. vl907 Phacops mimus Salter; Lake (in Reid 1907, p. 39). vl960 Salter; Whittard, p. 131. Lectotype. (Here chosen.) GSM 19180. Cephalon, Perhaver Beach (PI. 10, fig. 9). Salter’s (1864, pi. 1, fig. 35) illustration shows a broken anterior margin and frontal lobe but reconstructs the right eye, which is missing from the specimen itself. Material. East Catasuent^ll cephala, 18 cranidia, 1 librigena, 5 pygidia. Perhaver Beach— 6 cephala (earlier collections). Diagnosis. A species of Kloucekia characterized by a cephalon with large, weakly curved eyes set close to the elongate glabella. Axial furrows slightly divergent anteriorly. TEXT-FIG. 8. Reconstruction of Kloucekia (A.) mimus. (‘L-form’ produced by transverse compression, based on GSM Zs 608b (unfigured form with short fixigenal spines); ‘W-form’ produced by sagittal compression, based on 608c (unfigured form preserved at right angles to 608b); ‘O-form’ constructed from 608b and c using a calculated strain ellipse axis ratio 1 : 894. All at x 2-7. Description. Cephalic outline sub-semicircular, length about two-thirds maximum width (across occipital furrow), anterior margin rounded or bluntly pointed. Rather elongate, pentagonal glabella, maximum width (across frontal lobe) about three-quarters length. Bluntly pointed in front. Axial furrows almost straight, diverging at 20-30°, or curving apart slightly in front of 2p lobe. Three pairs of glabellar furrows and lobes. Ip furrows straight, deep, one-third glabellar width, directed very slightly rearward. 2p furrows very short, straight, very shallow, not reaching axial f urrows, 3p furrows longest, shallow, and running almost straight and distinctly rearward from axial furrows. Glabella gently convex and in profile slopes down con- tinuously from occipital ring to anterior border, becoming progressively less convex (tr.) anteriorly. No preglabellar field. Occipital furrow deeply notched laterally, curved forward and shallow medially. Fixigenae carry deep posterior and posterolateral border furrows. Posterior border widens conspicuously toward fixigenal angle, which is usually rounded. One individual (text-fig. 8) bears short fixigenal spines. Eyes narrow and weakly curved, seen in plan view as shallow crescents, extending from immediately posterior of Ip furrow to 3p furrow. Eyes set close to axial furrows anteriorly, with distinct but narrow palpebral lobes. Palpebral area of fixigenae narrow (tr.). Schizochroal visual surfaces with at least twenty-five dorso- ventral files (Clarkson 1966, p. 3) of up to eight hexagonal facets. Posterior sections of facial sutures convex forward, running from rear ends of eyes to lateral margin, with point cv ( Richter and Richter 1 940, pp. 16-17) slightly behind point e, opposite Ip lobes. Anterior sections skirt closely around frontal glabellar lobe to anterior margin. Small, steeply angled fixigenae with conspicuous antero-lateral furrow. Hypostoma and thorax not known. Pygidia (tentatively associated) have no caudal spine. 86 PALAEONTOLOGY, VOLUME 17 Dimensions. Lectotype : length of glabella (sag.) with occipital ring (Gn) = approx. 5-7 mm (anterior margin reconstructed); width (tr.) of frontal lobe = 41 mm; length (exsag.) of eye (A) = 21 mm; width (tr.) of occipital ring = 2-7 mm; length of posterior border = 41 mm. Cephalon GSM Zs 600 (an L-form): Gn = 81 mm; length of glabella (sag.) without occipital ring (G) =- 6-6 mm ; length of frontal lobe = 3-2 mm ; width of frontal lobe = 5 0 mm ; distance between eye and posterior border furrow =11 mm; A = 30 mm; width of eye crescent =1-5 mm; A/G = 45-5%; A/Gn = 37%; width of occipital ring = 2-8 mm; width of fixigena (max.) = 4-3 mm. Cephalon Zs 597 (a W-form): Gn = 4-5 mm; G = 4 0 mm; length of frontal lobe = 1-6 mm; width of frontal lobe = 3-8 mm; A = 1-7 mm; width of eye crescent = L5 mm; A/G = 42-5%; A/Gn = 37-7%; width of occipital ring = 2-7 mm. Remarks. The species was proposed by Salter for the small, distinctive, phacopid cephala that form an important element of the south Cornwall faunas. It has never yet been recognized elsewhere. Conspicuous changes of shape occur in these cephala during only slight deformation, but these effects can be assessed and removed. The description is based upon two cephala (Zs 608b, Zs 608c) preserved 18 mm apart, at right angles, and without visible angular shear strain. Measurement of ratios of parallel lengths does not indicate any primary differences and for these two individuals to represent L- and W-forms a strain ellipse axis ratio of only 1 : 0-894 is required. Such a strain is compatible with other measurements of this fauna and so the descrip- tion is essentially based on an O-form reconstructed from these two (text-fig. 8). Salter separated his new species from Phacops apiculatus M’Coy, 1851, on cephalic proportions and on the conspicuous lateral border furrow of the former. K. {K.) rnimus is distinguished from K. (K.) apiculatus by its larger, weakly curved eyes set relatively close to the axial furrows. K. {K.) micheli (Tromelin, 1876), a species of EXPLANATION OF PLATE 10 Ogygiocaridinae gen. indet. Fig. 1 . GSM Zs 587, librigena, internal mould, Catasuent, x 1 -5. Fig. 2. GSM Zs 588, incomplete pygidium, internal mould, Catasuent, x 1. Trinucleidae gen. indet, all from Catasuent, x 3-5. Fig. 3. GSM Zs 575, cephalon, internal mould. Fig. 4. GSM Zs 577, cephalon, internal mould, showing eye tubercles and median glabellar tubercle. Fig. 5. GSM Zs 573, cephalon, internal mould. Kloucekia (Kloucekia) mimus (Salter). Fig. 6a-h. GSM Zs 597, W-form cephalon, internal mould, Catasuent, x 3-5. Fig. la-h. GSM Zs 600, incomplete L-form cephalon, internal mould, Catasuent, x 4. Fig. 8. GSM Zs 595, incomplete W-form cephalon, internal mould, Catasuent, x4. Fig. 9. GSM 19180, lectotype, cephalon, internal mould, Perhaver, x3-3. Fig. lOa-6. GSM Zs 605a, cephalon, internal mould showing angular shear strain, Catasuent, x4. a, slightly oblique. Crozonaspis peachi n. sp., all from Came, all X 3 except hg. 18x4 and hg. 20 X 3-5. Fig. 1 1 . GSM Zs 676, crushed cephalon, internal mould. Fig. 12. GSM Zs 658b, caudal spine, latex cast of external mould. Fig. 13a-6. GSM Zs 658a, paratype pygidium, latex cast of external mould. Fig. \Aa-h. GSM Zs 677, holotype, cephalon, internal mould. Fig. 15. GSM Zs 679, immature cephalon, internal mould. Fig. 16. GSM Zs 673, cranidium, internal mould. Fig. 17. GSM Zs 674, cranidium, internal mould. Fig. 18. GSM Zs 667, paratype right librigena and eye, latex cast of external mould— counter- part of Zs 666. Fig. 19. GSM Zs 675, paratype cranidium, internal mould. Fig. 20. GSM Zs 666, paratype right librigena and eye, internal mould— counterpart of Zs 667. PLATE 10 14a SADLER, Cornish Ordovician trilobites PALAEONTOLOGY, VOLUME 17 Llandeilo age and close to the type-species, was established on the basis of its much larger eyes (Henry 1965, p. 205), which are slightly larger and taller than the Cornish species. The glabella of K. {K.) micheli is more convex and relatively wide anteriorly, resembling more the supposed W-forms of K. (K.) mimus. The two species are very similar and both occasionally carry fixigenal spines (Henry 1965, p. 202). The glabellar profile resembles Phacopidella, but that genus is distinguishable by its large eyes set close to the posterior margin and by its axial furrows, which diverge forward at 45-55° (Nion and Henry 1966, p. 889). The genus Baniaspis Destombes has much smaller eyes. Subfamily dalmanitininae Destombes, 1972 Struve (1958, p. 190) distinguished a Dalmanitina group within the subfamily Zeliszkellinae (Delo 1935) on the basis of a well-developed lateral border and border furrow which are constricted or completely suppressed anteriorly by the frontal lobe of a club-shaped glabella, on the relatively small eyes situated away from the antero- lateral border, on the presence of genal spines and on a mucronate pygidium with rela- tively numerous segments. Destombes (1972, p. 36) elevated the group to the status of a separate subfamily. His diagnosis for the subfamily Zeliszkellinae is a straight translation of Struve’s diagnosis of the Zeliszkella group. His diagnosis for the sub- family Dalmanitininae is clearly based on Struve’s proposals for the Dalmanitina group, but with the presence of three, generally equally incised glabellar furrows as an added specification. Genus crozonaspis Henry, 1968 Type-species. Crozonaspis struvei Henry, 1968, by original designation. Diagnosis. (After Henry 1968, p. 368.) A genus characterized by an ogival cephalon with a very narrow anterior border, which is enlarged as a short rounded ‘beak’ (as a term to describe this projection of the anterior border, Henry’s word ‘rostre’, 1968, p. 368, cannot be directly rendered into English in accordance with the Treatise definitions) at the median line. Ip glabellar furrows wide (exsag.) and deep; 2p furrows and 3p furrows weakly defined. Eyes oblique to axial furrows so that palpebral area increases rapidly in width posteriorly. Pygidium with large caudal spine and with few pleura in lateral lobes. Remarks. Henry (1968, p. 369) considered his new genus to be very close to Dal- manitina, and set out the means of distinguishing it from the several Dalmanitina subgenera, as well as the case for establishing it at generic level. The resemblances to Dalmanitina include reduced anterior border, relatively small eyes, genal spines, bifurcation of internal extremities of Ip lateral glabellar furrows, convergence of Ip and 2p furrows and the mucronate pygidium. These features support the placing of Crozonaspis in the Dalmanitininae. However, the points of difference— relatively weak 2p and 3p glabellar furrows, and the relatively small number of pygidial pleurae, with weak interpleural furrows— introduce features of the Zeliszkellinae, especially if Destombes’s addition of glabellar segmentation to the diagnosis of the Dalmanitininae is accepted. The establishment of Cr. cliouherti Destombes adds relatively large eyes, another zeliszkellinid feature, to the scope of the genus. SADLER: CORNISH ORDOVICIAN TRILOBITES 89 Distribution. The four species already known are all of Llandeilo age. Cr. struvei Henry and Cr. kerfornei Clarkson and Henry are known from Brittany, Cr. chouberti Destombes from Morocco, and Cr. incerta (Deslongchamps) from Normandy and Spain. Crozonaspis peachi n. sp. Plate 10, figs. 1 1-20; text-fig. 9 Derivatio nominis. In tribute to C. W. Peach, who first collected trilobites in south Cornwall and mapped the distribution of the fossiliferous quartzites. Holotype. Cephalon, GSM Zs 677, Came Beach, author’s collection (PI. 10, figs. \Aa-b). Paratypes. GSM Zs 675, cranidium, PI. 10, fig. 19; GSM Zs 677b, 671a, 671b, cranidia; GSM Zs 667, 666, librigenae, PI. 10, figs. 18 and 20; GSM Zs 661, 662, librigenae with anterior border; GSM Zs 658a, pygidium, PI. 10, figs, \ia-b. Locus typicus. Came Beach, Veryan, South Cornwall (SW 9109 3793). Stratum typicum. Gorran Quartzites, Ordovician (?Llandeilo). Material. Came Beach — 4 cephala, 22 cranidia, 18 librigenae, 27 pygidia, 1 caudal spine. Diagnosis. A species of Crozonaspis characterized by a relatively wide cephalon (tr.) and moderately large eyes set away from the glabella posteriorly and not interrupting the lateral border furrow in plan view. 3p glabellar furrows deepen distally and con- nect with axial furrows. TEXT-FIG. 9. Reconstruction of Crozonaspis peachi n. sp. a-b. cephalon (internal mould); c-e, librigenae: c, external cast (ventral), d, external cast (dorsal), e, internal mould (ventral);/, cranidium, external cast (dorsal); g-/, pygidium (external cast): g, dorsal, h, posterior, /, right lateral view, a-f x3;g-i x4. 90 PALAEONTOLOGY, VOLUME 17 Description. Cephalon ogival in outline, almost half as long as maximum width (measured across occipital furrow). Glabella pentagonal, pointed anteriorly. Axial furrows moderately deep, becoming shallower anteriorly, diverging forward at about 40-45° in front of Ip glabellar lobes. Three paired glabellar lobes and furrows. Ip glabellar furrows short, straight, directed slightly backward adaxially and very deep, each about one-third glabellar width. Internal extremities bifurcate, anterior branch deeper, directed toward axis, posterior branch (more deeply incised in external casts than internal moulds) directed toward occipital furrow, tending to isolate Ip lobes. 2p furrows least well marked (especially on internal moulds), short, slightly convex forward, directed slightly forward adaxially, shallowing abaxially, not connected with axial furrows. 3p furrows distinct, but shallow adaxially, long, sinuous, directed backward adaxially. They comprise two shallow flat arcs, each convex forward, abaxial arc deepest, shallow where connected to axial furrows. 2p and 3p furrows usually visible on internal moulds and always distinct on the very fragmentary external moulds (text-figs. 9a and /). Ip glabellar lobes narrow (exsag.) adaxially tending toward isolation from median area. 2p lobes narrow (exsag.) adaxially. 3p lobes narrow adaxially, anterolateral portions may become markedly convex. Frontal glabellar lobe lozenge-shaped, twice as wide (tr.) as long. Glabellar profile (sag.) high, convex, descending steeply to anterior border. Occipital ring slightly higher than glabella. Occipital furrow convex forward, well-marked medially, very deep laterally. Anterior areas of lixigenae very small. Palpebral areas large, convex, widest posteriorly. Posterior areas wide (tr.), deep posterior border furrow, genal angles not seen. Eyes moderately large, crescentic, extending from 3p lobe to occipital furrow, diverging from axial furrows backward. Major ocular index (Struve 1958, pp. 167-168), A/G, 42- 45%; minor ocular index, A/Gn, 35-38%. Visual surface schizochroal, at least twenty-four dorso-ventral files, bearing up to seven lenses. In plan view, eyes do not interrupt anterolateral border furrows. Facial suture proparian. Anterior portion of suture closely follows anterior margin of glabella. Anterior border very narrow, swelling slightly at median line (these features are usually damaged on whole cephala but can be seen in disarticulated librigenae). This projection of the anterior border, or ‘beak’, corresponds to a distinct, rounded ventral bulge in the wide anterior cephalic doublure. Main portion of librigenae (below eyes) steeply inclined to horizontal (60-70°). Pronounced anterolateral border furrow and wide border at 45° to horizontal. Librigenae have a wide anterolateral doublure bearing a deep, oblique furrow. Small, immature cephala (PI. 10, fig. 15) have relatively large eyes. Major ocular index, 55%; minor ocular index, 50%. Hypostoma and original shell surface not known. Large, triangular pygidium. Axis stands high anteriorly between distinct axial furrows, but becomes poorly defined backward of mid length. Seven to eight distinct axial rings behind articulating half-ring, becoming smaller backward. Posterior portion of axis comprises an unsegmented, poorly defined ridge, continuing as a long, caudal spine, upturned as much as 60° from horizontal. Stout anterior end of spine bears a deep ventral furrow. Pleural lobes are set at 1 10 140° to one another and bear up to five pairs of furrows. Dimensions. Holotype cephalon: length of glabella (sag.), including occipital ring (Gn) = 7-8 mm; length of glabella without occipital ring (G) ^ 6-6 mm; length of frontal lobe — 3-3 mm; width of frontal lobe (tr.) = 7-5 mm; distance between eye and posterior border furrow = 0-9 mm; length of eye (A) = 30 mm; A/G = 45-4%; A/Gn = 38-7%; width of occipital ring = 4-7 mm; width (tr.) of fixigena (max.) = 5-4 mm. Paratype pygidium: width of anterior axial lobe = 30 mm; maximum width of pleural lobe = 3-7 mm; length of axis including caudal spine -= 9 0+ mm. Remarks. Compared with the four other species, Crozomispis peachi has a relatively short, wide cephalon. It more closely resembles Cr. clwuherti Destombes and Cr. incerta (Deslongchamps) than it does the Brittany species, which have distinctly smaller eyes. Cr. .s'trwve/Henry ; A/G 27-28%, A/Gn ^ 23%; Cr. kerforuei Clarkson and Henry: A/G = 32%, A/Gn = 28%. The eyes of these species are correspondingly farther from the posterior border furrow. The eyes of Cr. chouberti (A/G == 50-61%, A/Gn 42-52%) are larger than those of Cr. peachi. In both species the posterior extremities of the eyes lie immediately forward of the occipital furrow, but in Cr. chouberti the eyes continue as far forward as the 3p glabellar furrows. In the latter species the eyes are taller and set closer to the anterolateral border furrows, which they interrupt in plan view. The 3p glabellar furrows in Cr. peachi may become very shallow adjacent to the axial furrows, but do not show the interruption that is charac- SADLER: CORNISH ORDOVICIAN TRILOBITES 91 teristic of the Brittany species. In this respect Cr. peachi again resembles more the Normandy species and the Moroccan species. Cr. incerta most closely resembles the new species, but has a relatively narrow cephalon. The eyes of Cr. incerta are probably relatively small (A/G approx. 37-38%; A/Gn approx. 31-32%) compared with Cr. peachi, but they have not been satisfactorily illustrated. Cr. peachi is most readily distinguished by its 2p and 3p furrows which are typically distinctly marked, especi- ally 3p and on external moulds. On certain internal moulds (PI. 10, fig. 16) traces of 2p and 3p furrows are imperceptible, resembling the condition in Cr. incerta. Acknowledgements. This work was undertaken during a research assistantship at the University of Bristol, financed by N.E.R.C. and supervised by Dr. S. C. Matthews, who was a source of vital encouragement and advice during preparation of the paper. Valuable criticisms of an early draft were kindly given by Professor H. B. Whittington, F.R.S., and by Dr. A. W. A. Rushton, who also, on behalf of the Institute of Geological Sciences, generously made available material from earlier collections. For the photographs I am indebted to the expertise of Mr. F. W. Seavill. RFFFRFNCFS BARRANDE, J. 1872. Systeme silurien du centre de la Boheme. I (Supplement) Trilobites, Cnistaces divers + Poissons. Prague-Paris, 647 pp., 37 pis. BATES, D. E. B. 1968. The Lower Palaeozoic brachiopod and trilobite faunas of Anglesey. Bull. Br. Mus. nat. Hist. (Geol.) 16, 125-199, pis. 1-9. 1969. Some early Arenig brachiopods and trilobites from Wales. Ibid. 18, 1-28, pis. 1-9. BORN, A. 1916. Die Ccdymene Tristani Strife (mittleres Unter Silur) bei Almaden und ihre Fauna, Gliederung und Verbreitung. Abh. senckenb. naturforsch. Ges. 36, 309-358, pis. 24-27. 1953. FI tramo de Calymene tristani en Almaden (Ordoviciense medio) su fauna, division y extension (transl. B. Melendez from Born 1916). Publ. extranjeros geol. Esp. 7, 175-263, pis. XIII-XVIII. BRONGNIART, A. 1822. Histoire naturelle des Crustaces fossiles: Les trilobites. Paris, 154 pp., 11 pis. CARRE, D., HENRY, J.-L., POUPON, G. and TAMAiN, G. 1972. Les Quartzites Botella et leur faune trilobitique. Le probleme de la limite Llandeilien-Caradocien en Sierra Morena. Bull. Soc. geol. Fr. (7), 12, 774-785, pi. XXV. CHAUVEL, J., DROT, J., PiLLET, J. and TAMAIN, G. 1969. Precisions sur I’Ordovicien moyen et superieur de la ‘serie-type’ du Centillo (Sierra Morena orientale, Fspagne.). Ibid. (7), 11, 613-626, pis. XIII-XV. CLARKSON, E. N. K. 1966. Schizocroal eyes and vision of some Silurian acastid trilobites. Palaeontologv, 9, 1-29, pis. 1-3. and HENRY, j.-L. 1970. Sur une nouvelle espece du genre Crozonaspis (Trilobite) decouverte dans rOrdovicien de la Mayenne. Bull. Soc. geol. Fr. (7), 11, 116-123. COLLINS, J. H. 1893. A working list of the Palaeozoic fossils of Cornwall. Trans. R. geol. Soc. Corn. 11, 421-479. DEAN, w. T. 1960-1. The Ordovician trilobite faunas of South Shropshire (2 parts). Bull. Br. Mus. nat. Hist. (Geol.) 4, 73-143, 313-358, pis. 11-19, 49-55. 1964. The status of the Ordovician trilobite genera Prionocheilus and Polyeres. Geol. Mag. 101, 95-6. 1965. A revision of the Ordovician trilobite genus Bathycheilus Holub. Sb. ndr. Mus. Praze. 21 B, 1 - 10, pis. 1-2. 1966. The Lower Ordovician stratigraphy and trilobites of the Landeyran Valley and the neighbouring district of the Montagne Noire, south-western France. Bull. Br. Mus. nat. Hist. (Geol.) 12, 245-353, pis. 1-21. 1967. The correlation and trilobite fauna of the Bedinan Formation (Ordovician) in south-eastern Turkey. Ibid. 15, 81-123, pis. 1-10. 1971 . The Lower Palaeozoic stratigraphy and faunas of the Taurus Mountains near Beysehir, Turkey. II. The Trilobites of the Seydisehir formation (Ordovician). Ibid. 20, 1-24, pis. 1-5. DELO, D. M. 1935. A revision of the phacopid trilobites. J. Paleont. 9, 402-420. DESTOMBES, J. 1972. Les Trilobites du sous-ordre des Phacopina de I’Ordovicien de I’Anti-Atlas (Maroc). Notes Mmi. Serv. Mines Carte geol. Maroc, 240, 5-1 12, pis. 116. 92 PALAEONTOLOGY, VOLUME 17 FOX, H. 1908. Trilobite in the Veryan Quartzite. (Note on a specimen of Calymene from Veryan by P. Lake.) Trans. R. geol. Soc. Corn. 13, 233-236. GiGOUT, M. 1951. Etudes geologiques sur la Meseta marocaine occidentale (arriere pays de Casablanca, Mazagan et Safi). Notes Mem. Serv. Mines Carte geol. Maroc, 86, 1-507, pis. 1-18. HARRINGTON, H. J. and LEANZA, A. F. 1957. Ordovician trilobites of Argentina. Spec. Pubis. Univ. Kansas Dept. Geol. 1, 267 pp. HENNiNGSMOEN, G. 1960. The Middle Ordovician of the Oslo region 13. Trilobites of the family Asaphidae. Norsk, geol. Tidsskr. 40, 203-257. HENRY, j.-L. 1965. Revision de Kloucekia micheli (Tromelin 1876) (Trilobite, Ordovicien Moyen du massif armoricain). Bull. Soc. geol. miner. Bretagne (March 1965), 199-210, pi. I. 1968. Crozonaspis struvei n.g., n.sp., Zeliszkellinae (Trilobita), de FOrdovicien moyen de Bretagne. Senckenberg. leth. 49, 367-380, pis. 1-2. 1970. Quelques Calymenacea (Trilobites) de FOrdovicien de Bretagne. Annls. Paleont. (Invert.) 56, 3-27, pis. A-c. HICKS, H. 1872. On the Tremadoc rocks in the neighbourhood of St. David’s, South Wales, and their fossil contents. Q. Jl geol. Soc. Lond. 29, 39-52, pis. 3-5. HOLUB, K. 1908. Pfisoevek ku poznani fauny pasma Ddiy. Rozpr. ceske Akad. 17, 1-19, pi. 1. KOBAYASHi, T. 1951. On the Ordovician trilobites in Central China. J. Fac. Sci. Univ. Tokyo, 8, 1 -87, pis. I-V. LU, H. Y. 1957. Trilobita. In Index Fossils of China, Invertebrate, 3, 249-294, pis. 137-155. [In Chinese.] MOORE, R. c. 1959. Treatise on invertebrate paleontology. Part O. Arthropoda I. 579 pp. Kansas. MURCHISON, R. I. 1846. A brief review of the elassification of the sedimentary rocks of Cornwall. Trans. R. geol. Soc. Corn. 6, 317-326. NiON, J. and HENRY, j.-L. 1966. Phacopidella (Prephacopidella) Impel sp. nov.— nouveau trilobite del’Ordovo- cien du Finistere. Bull. Soc. geol. Fr. (7), 8, 884-890, pi. XXIV. OEHLERT, D. p. 1895. Sur les Trinucleus de la France. Ibid. (3), 23, 299-336. PEACH, c. w. 1841 . An account of fossil organic remains found on the south-east coast of Cornwall. Trans. R. geol. Soc. Corn. 6, 12-23. 1842. On the geology of part of the parish of Gorran in Cornwall. Ibid. 6, 51-58. 1844. On the fossil geology of Cornwall. Ibid. 6, 181-185. POMPECKJ, J. F. 1898. liber Calymene Brongniart. Neues Jb. Miner. Geol. Paldont. 1, 186-250. REED, F. R. c. 1915. Supplementary memoir on new Ordovician and Silurian fossils from the northern Shan States. Mem. geol. Surv. India, 4, 102 pp., 12 pis. REID, c. 1907. Geology of the country around Mevagissey. Mem. geol. Surv. U.K. 353 and 354. RENAUD, A. 1952. Les Collections paleontologiques du Musee vieux Chateau de Laval. Bull. Mavenne-Sci. 29-60, pis. I-IV. PONCET, J., CAVET, p., LARDEUX, H. and BABIN, c. 1968. Lc Dcvonicn du Massif Armoricain. In Inter- national symposium on the Devonian system. D. H. Oswald (ed.), Calgary, 1, 135-152. RICHTER, R. 1948. Einfuhrung in die Zoologische Nornenclatur. Kramer, Frankfurt-a.-M., 252 pp. (2nd edn.). and RICHTER, E. 1949. Die Trilobiten der Erdbach-Zone (Kulm) im Rheinischen Schiefergebirge und im Harz. Senckenbergiana 30, 63-94, pis. 1-3. ROUAULT, M. 1847. Extrait de memoire sur les trilobites du departement d’llle-et-Vilaine. Bull. Soc. geol. Fr. (2), 4, 309-328, pi. III. SALTER, J. w. 1864u. Note on the fossils from the Budleigh Salterton Pebble-Bed. Q. Jl geol. Soc. Lond. 20, 286-302. 1864-83. A monograph on the British trilobites from the Cambrian, Silurian and Devonian. Palaeontogr. Soc. [Monogr.], 5 Parts. SDZUY, K. 1957. Bemerkungen zur Familie Homalonotidae (mit der Beschreibung einer neuen Art von Calymenella). Senckenberg. leth. 38, 275-290, pi. 1. SEDGWICK, A. 1852. On the slate rocks of Devon and Cornwall. Q. Jl geol. Soc. Lond. 8, 1 19. SHIRLEY, J. 1936. Some British trilobites of the family Calymenidae. Ibid. 92, 384-422, pis. XXIX-XXXl. SIEGFRIED, p. 1969. Trilobiten aus dem Ordovizium des Ebbe-Sattels im Rheinischen Schiefergebirge. Paldont. Z. 43, 148 168. SPJELDNAES, N. 1961. Ordoviciaii climatic zones. Norsk, geol. Tidsskr. 41, 45-77. STRUVE, w. 1958. Beitriigezur Kenntnis der Phacopacea (Trilobita), 1. Die Zeliszkellinae. Senckenberg. leth. 39, 165-219, pis. 1-4. SADLER: CORNISH ORDOVICIAN TRILOBITES 93 STUBBLEFIELD, c. J. 1939(3. Some Devonian and supposed Ordovician fossils from south-west Cornwall. Bull. geol. Surv. Gt. Brit. 2, 63-71. 19396. Some aspects of the distribution and migration of trilobites in the British Lower Palaeozoic faunas. Geol. Mag. 74, 49-72. 1960. Trilobites of south-west England. Trans. R. geol. Soc. Corn. 19, 101-112. THADEU, D. 1956. Note sur le Silurien biero-durien. Bolm. Soc. geol. Port. 12, 1-42, pis. I-X. TRISTAN, I. and BIGOT, p. s. M. 1808. Notice sur les crustaces renfermes dans quelques schistes notamment dans ceux des environs de Nantes et d’Angers. J. Mines, 23, 21-32. voGDES, A. w. 1925. Palaeozoic Crustacea. Trans. S. Diego Soc. nat. Hist. 4, 1-154. WARRINGTON, G. 1971. Palynology of the New Red Sandstone sequence of the south Devon coast. Proc. Ussher Soc. 2, 307-314. WELLMAN, H. w. 1962. A graphical method for analysing fossil distortion caused by tectonic deformation. Geol. Mag. 99, 348-352. WHiTTARD, w. F. 1955-67. The Ordovician trilobites of the Shelve inlier, west Shropshire. 9 Parts. Palaeontogr. Soc. [Monogr.], 352 pp., pis. I-L. WHITTINGTON, H. B. 1962. The Ordovician trilobites of the Bala area, Merioneth. Part I. Ibid. 1-32, pis. I-XXXII. 1966a. Trilobites of the Henllan Ash, Arenig Series, Merioneth. Bull. Br. Mus. nat. Hist. (Geol.) 11, 491-505, pis. 1-5. 19666. Phylogeny and distribution of Ordovician trilobites (Presidential Address). J. Paleont. 40, 696-737. 1968. Crvptolithus (Trilobita): specific characters and occurrences in Ordovician of North America. Ibid. 42, 702-714, pis. 87-89. P. M. SADLER Geologisch-Palaontologisches Institut Georg August Universitat Berlinerstrasse 28 Revised typescript received 15 February 1973 34 Gottingen, West Germany THE TRILOBITE CLAVAGNOSTUS HOWELL FROM THE CAMBRIAN OF TASMANIA by i. B. ikGO and daily Abstract, Two new species of Clavagnostus, C. milli and C. burnsi, are described from the late Middle and early Upper Cambrian sequences of Tasmania. A third new species, ClavagnoslusO) rawlingi is questionably assigned to Clavagnostus. The genus Clavagnostus is reviewed. It is shown that the known species of Clavagnostu.s can be placed into three groups: (1) the C. repandus group which is characterized by a rounded glabellar front, no preglabellar median furrow, and a blunt pygidial axis which reaches the posterior border; (2) the C. burnsi group which is charac- terized by an angular glabellar front, a preglabellar median furrow, and a blunt pygidial axis which reaches the posterior border; and (3) the C. sidcatus group which is characterized by a preglabellar median furrow and a pointed pygidial axis. It is shown that four and probably five different species of Clavagnostus have been included in C. repandus. The pygidium figured by Poulsen (I960) as the holotype of Peronopsis ultima is shown to be a cephalon of a species of Clavagnostus, probably that of C. chipiquensis (Rusconi). Three new agnostid species are described here: two of these are referred to Clava- gnostus, a third is questionably assigned to Clavagnostus. Two isolated pygidia of Clavagnostus from separate localities are also described. All species come from the late Middle or early Upper Cambrian sequences of north-western Tasmania. Text- fig. 1 shows the correlation and stratigraphic distribution of the species described TEXT-FIG. 1 . Stratigraphic distribution of species of Clavagnostus from north-western Tasmania. Correlations between the Queensland and the North American zones are after Opik (1963, 1967). below as presently known. Text-fig. 2 indicates the geographic position of the localities noted below. The remaining trilobites from the Tasmanian localities noted herein will be described in later papers. The need for a review of Clavagnostus makes it desirable to describe the species of this genus separately. [Palaeontology, Vol. 17, Part 1, 1974, pp. 95-109, pis. 11-12.] G 96 PALAEONTOLOGY, VOLUME 17 In common with other Tasmanian Cambrian fossils the trilobites described below have undergone tectonic distortion. However, they are from localities where dis- tortion is minimal. All specimens are preserved in siltstone as external and internal moulds. In order to prepare them for description, silicone rubber casts of the external moulds were prepared. These casts were then photographed after being whitened with magnesium oxide. All catalogue numbers refer to the collections of the Geology Department, University of Tasmania. The authors had at their disposal rubber casts of all the specimens of Clavagnostiis repandus (Westergard) and C. Westergard hgured by Westergard (1946, pi. 4). Rubber casts of the specimen figured by Poulsen (1960, pi. 1, fig. 13) as the pygidium Peronopsis ultima Poulsen and of the pygidium of Clavaguostus chipiquensis (Rusconi) figured by Poulsen (1960, pi. 1, fig. 14) were also available. All these specimens are refigured herein. The taxonomic terminology used in this paper is after Opik (1961«, 1963, 1967). The agnostid trilobite classifica- tion of Opik (1967) is followed here. JAGO AND DAILY: CLAVAGNOSTVS FROM TASMANIA 97 SYSTEMATIC DESCRIPTIONS Order miomera Jaekel, 1909 Suborder agnostina Salter, 1864 Superfamily agnostacea M’Coy, 1849 Family clavagnostidae Howell, 1937 Subfamily clavagnostinae Howell, 1937 Genus clavagnostus Howell, 1937 Clavagnostus: Elowell 1937, p. 1164; 1959, p. O 174; Kobayashi 1939, p. 120; 1943, p. 307; Lermontova 1940, p. 129; Westergard 1946, p. 55; Hupe 1953, p. 64; Pokrovskaya 1960a, p. 59; 19606, p. 161 ; Opik 1967, p. 114, Tomorhachis: Resser 1938, p. 51. Culipagnostus: Rusconi 1952, p. 11. Type species. Agnostus repandus Westergard in Holm and Westergard, 1930, p. 13, pi. 4, figs. 11,12 (only). New diagnosis. Cephalon has glabella with no transverse glabellar furrow. Cephalic cheeks smooth. Cephalic spines, long to short. Glabella featureless except for a small to elongated node. Simplimarginate pygidium with long tapered axis. To the posterior of the axial centre is a pair of pits (the ‘clavagnostid pits’). Transverse axial furrows effaced on anterior part of axis. Posterior part of axis depressed. Border spines, short to long. Discussion. Species of the genus Clavagnostus have a very wide distribution in late Middle and early Upper Cambrian sediments. The type species C. repandus (Westergard) has been reported from north-west Siberia (Holm and Westergard 1930; Shabanov el al. 1967) and Sweden (Holm and Westergard 1930; Westergard 1946); C. sulcatus Westergard occurs in Sweden (Westergard 1946) and was recently reported from north-west Siberia by Lazarenko and Nikiforov (1968). Kobayashi (1943) described C. repandiformis Kobayashi and C. cf. repandiforniis from Siberia. C. ovalis Pokrovskaya is found in the Sayan-Altai region (Pokrovskaya 19606). C. aequalis Howell and C. spinosa (Resser) were described from Vermont and Alabama by Howell (1937) and Resser (1938). Drewes and Palmer (1957), Robison and Palmer (1968), and North (1971) have reported the occurrence of Clavagnostus from western North America. C. chipiquensis (Rusconi) was reported from Argentina by Rusconi (1952) and Poulsen (1960). Opik (1967) described C. bisectus Opik from Queensland. Prior to this paper Clavagnostus has been reported from Tasmania by Banks (1956, 1962), Blissett (1962), Burns (1964), Jago and Buckley (1971), and Jago (19726). The cephalon (holotype) and pygidium (both from the Andrarum Limestone, Andrarum, Scania) of Clavagnostus repandus figured by Holm and Westergard (1930, pi. 4, figs. 11, 12) and by Westergard (1946, pi. 4, figs. 19, 20) appear to be different from the cephalon and pygidium (from Skollersta, Narke) figured as C. repandus by Westergard (1946, pi. 4, figs. 21, 22). These specimens are refigured herein on Plate 1 1 . An inspection of rubber casts of these specimens reveals the follow- ing differences : ( 1 ) in the holotype cephalon (PI. 11, fig. 1 herein) the glabella expands slightly forwards whereas in the Narke specimen (PI. 11, fig. 3 herein), the glabella contracts forwards; (2) the pygidial axis of the Scania specimen (PI. 1 1, fig. 2 herein) 98 PALAEONTOLOGY, VOLUME 17 is more sharply pointed and does not extend as far to the posterior when compared with the Narke specimen (PI. 11, fig. 4 herein); and (3) the clavagnostid pits of the Scania pygidium are shallow and longitudinally elongated whereas those of the Narke pygidium are deep and only slightly elongated. The figure of the holotype cephalon given by Westergard (1946, pi. 4, fig. 19) shows a prominent elongated glabellar node. However, although our photograph (PI. 1 1, fig. 1) of a rubber cast of the holotype shows a faint trace of what may be an elongated node, the node is not as prominent as suggested by Westergard’s figure. It should be pointed out here that most of Westergard’s figures have been retouched. The differences noted above indicate that the specimens from the two different Swedish localities should be placed in different species, with the specimens from Scania being retained as the type material of Clavagnostiis repandus and the material from Narke being the basis of a new species. Also, in view of the lack of detailed know- ledge of the stratigraphy of Paradoxides forchhammeri beds in Narke (see Wester- gard, 1946, p. 17) it is questionable whether the two species are of the same age. The Siberian specimens illustrated in Chernysheva (1960, pi. 1, figs. 12, 13) as C. repandus are very similar to the type material of C. repandus from Scania. The most significant difference appears to be that the pygidial border spines of the Siberian form are placed much further to the posterior than in the form from Scania. In addi- tion the cephalic node appears to be stronger and the clavagnostid pits deeper in the Siberian material than in the type repandus. Holm and Westergard (1930, pi. 1, figs. 35-39) figured three cephala and two pygidia as C. repandus from a late Middle Cambrian sequence on Bennett Island, north of Siberia. These specimens appear to belong to a species different from the holotype cephalon and the pygidium figured by Westergard (1946) from Scania. Although the pygidia of the two forms are close, the cephala are clearly different. The glabella of the holotype expands forward and is subsquare whereas that of the Bennett Island specimens also differs from the Narke form of ‘C. repandus'. The cephalon from Narke figured by Westergard, 1946, pi. 4, fig. 21 (refigured herein, PI. 11, fig. 3), has a more angular glabellar rear than do the cephala from Bennett Island. The pygidia also differ; the pygidial axis of the Bennett Island form does not extend as far to the posterior as does the pygidial axis of the pygidium from Narke (PI. 1 1, fig. 4 herein). Holm and Westergard (1930, pi. 1, figs. 40-43) also figured four cephala question- ably assigned to C. repandus. These cephala are poorly preserved, and it is difficult to make a meaningful comparison with other material. However, as noted in Holm and Westergard (1930, p. 14), two of these cephala (figs. 42, 43) have quite narrow glabellas. The cephalon figured in Holm and Westergard ( 1930, pi. 1, fig. 40) has what appears to be a preglabellar median furrow although this is probably caused by crushing. The above discussion indicates that four (and probably five) different species of Clavagnostus have been described as C. repandus. They are the specimens from (1) Andrarum Limestone, Andrarum, Scania, Sweden; holotype cephalon and pygidium figured in Holm and Westergard (1930, pi. 4, figs. 11, 12), in Wester- gard (1946, pi. 4, figs. 19, 20), and herein as PI. 11, figs. 1, 2. (2) Skollersta, Narke, Sweden; cephalon and pygidium figured in Westergard (1946, pi. 4, figs. 21, 22), and herein as PI. II, figs. 3, 4. JAGO AND DAILY: CLAVAGNOSTUS FROM TASMANIA 99 (3) River Judoma, east Siberia, as figured in Chernysheva (1960, pi. 1, figs. 12, 13). (4) Bennett Island, north of Siberia, as figured in Holm and Westergard (1930, pi. 1, figs. 35-39). (5) Bennett Island, north of Siberia, as figured in Holm and Westergard (1930, pi. 1, figs. 40-43). As noted above, the concept of C. repandus (Westergard) should be confined to the cephalon and the pygidium from Scania. The other material needs to be re- described. The two pygidia from north-west Siberia figured as C. sidcatus Westergard in Lazarenko and Nikiforov (1968, pi. 3, figs. 13, 14) are similar to the pygidium of C. sidcatus as figured in Westergard (1946, pi. 4, fig. 26) and herein as PI. 11, figs. 6, 8. Until cephala are found with the Siberian specimens, it cannot be certain that the two pygidia figured by Lazarenko and Nikiforov do belong in C. sidcatus. Photo- graphs taken with low angled light (see PI. 11, fig. 8) reveal that the pygidium of C. sidcatus figured by Westergard (1946, pi. 4, fig. 26) has very shallow depressions along the outer lateral parts of the acrolobe. There appear to be four pairs of these depressions with the two central ones being the most prominent. The pygidium of Clavagnostus chipiquensis (Rusconi) is illustrated by Poulsen (1960, pi. 1, fig. 14) and herein PL 11, fig. 11. The cephalon of chipiquensis is not described by Poulsen. However, the pygidium figured by Poulsen (1960, pi. 1, fig. 13) as the holotype of Peronopsis ultima Poulsen appears to be a Clavagnostus cephalon. An inspection of a rubber cast of this specimen reveals the presence of a basal lobe (the other one is not preserved) and traces of long cephalic spines (see PI. 11, fig. 10). The glabellar front is pointed, and there is a preglabellar median furrow. This cephalon is similar to that of C. burnsi sp. nov. and is probably the cephalon of C. chipicjuensis (Rusconi). Opik (1967, p. 113) recognized two separate genera in the different species described as Clavagnostus, with C. repandus (Westergard 1930) and C. sidcatus Westergard, 1946 representing the different genera. The writers agree with Opik in differentiating these species into separate groups. However, there seems to be a third group of species which can be placed in Clavagnostus as presently conceived and which is intermediate between the forms represented by C. sidcatus and C. repandus. This third group is represented by C. burnsi sp. nov. from Riana, Tasmania. It is charac- terized by a pygidium similar to that of C. repandus and a cephalon similar to that of the C. sidcatus group in that it has a preglabellar median furrow and an angular glabellar front although it is distinct from sidcatus in that the cephalic anterior is not pointed. The described species of Clavagnostus can be placed in three groups. 1 . The C. repandus group which is characterized by a rounded glabellar front, no preglabellar median furrow, and a blunt pygidial axis which reaches the posterior border (see PI. 11, figs. 1-4, 12-17). The known members of this group include all the species which have previously been placed in C. repandus (Westergard), plus C. niilli sp. nov. 2. The C. burnsi group which is characterized by an angular glabellar front, a pre- glabellar median furrow, and a blunt pygidial axis which reaches the posterior border. 100 PALAEONTOLOGY, VOLUME 17 C. burnsi sp. nov. and probably C. chipiquensis (Rusconi) are members of this group. C.(?) rawlingi sp. nov. is related to this group. 3. The C. sulcatus group of species is characterized by a preglabellar median furrow and a pointed pygidial axis which does not usually reach the pygidial border. The group includes C. ^u/cfl/u.s'Westergard, 1946, C. bisectiis Opi\k, 1967, C. repandiformis Kobayashi, 1943, C. ovalis Pokrovskaya, 19606, C. aequalis Howell, 1937, and prob- ably Tomorhachis spiuosa Resser, 1938. The last two species are known only from their pygidia. The species with known cephala have angular glabellar fronts with the exception of ovalis which has a rounded or subangular glabellar front. C. sulcatus differs from other members of the group by having a pointed cephalic front (see PI. 11, figs. 5, 7, 9). C. repandiformis has very long cephalic and pygidial spines. No photograph of C. ovalis is available, but the diagram given by Pokrovskaya (19606, p. 161, fig. 44) shows that the pygidial axis extends to the border. In other species in the C. sulcatus group the pygidial axis stops short of the border. The different combination of characters noted above in groups 1, 2, and 3 and the variation within the groups emphasizes the close and gradational relationship of all the species discussed above. The discovery of C. burnsi, intermediate between the C. repandus and C. sulcatus groups, raises the question of whether or not these groups belong in separate genera as suggested by Opik (1967, p. 113). As noted by Opik (1967, p. 114), if the as-yet-unknown cephalon of Tomorhachis spinosa is found to be similar to those of the C. sulcatus group, then the species of this group could be placed in the genus Tomorhachis. It would also be possible to erect a new genus to accommodate members of the C. burnsi group and have a total of three genera for the species discussed above. However, in view of the close and gradational relation- ship between the three groups, additional generic names seem unwarranted. The authors consider that all the species noted above should be included in the single genus Clavagnostus. Opik (1967, p. 114) raised the possibility of Clavagnostus being a subjective synonym of Aspidagnostus, but the differences between Clavagnostus and Aspidagnostus seem great enough to warrant generic separation. EXPLANATION OF PLATE 1 1 Figs. 1, 2. Rubber casts of cephalon and pygidium of Clavagnostus repandus (Westergard) from Andrarum Limestone, Andrarum, Scania. 1, holotype cephalon, x 1 1-2; 2, pygidium, x 1 1-3. Figs. 3, 4. Rubber casts of cephalon and pygidium figured by Westergard 1946, pi. 4, figs. 21, 22 as Clava- gnostus repandus (Westergard) from Skollersta, Niirke. 3, cephalon, X 13-5; 4, pygidium, X 14-5. Figs. 5-9. Rubber casts of Clavagnostus sulcatus Westergard from Gudhem, Vastergotland. 5, holotype cephalon figured by Westergard 1946, pi. 4, fig. 25; x 12-3. 6, pygidium figured by Westergard 1946, pi. 4, fig. 26; x 1 1-2. Fig. 8 is the same specimen as fig. 6 photographed in low angle light in an attempt to accentuate the very shallow depressions near the acrolobe margins. 7, cephalon, figured by Wester- gard 1946, pi. 4, fig. 24; X 10. 9, cephalon figured by Westergard 1946, pi. 4, fig. 23; x 1 1-8. Figs. 10, 11. Rubber casts of Clavagnostus chipiquensis (Rusconi) from Argentina. 10, cephalon figured by Poulsen 1960, pi. 1 , fig. 1 3 is the holotype pygidium of Peronopsis ultima Poulsen ; X 10. Note the presence of the left basal lobe and the traces of the right cephalic spine. 1 1, pygidium figured by Poulsen 1960, pi. I, fig. 14; X 10. f igs. 12-17. Clavagnostus niilli sp. nov. Lower fauna at Christmas Hills. 12, UT 86860b, holotype cephalon, X15-3. 13, UT 86853b, cephalon, X 131. 14, UT 86869e, cephalon, x 13-8. 1 5, UT 86607, pygidium, X 10-9. 16, UT 92479, pygidium, x 10-8. 17, UT 92469, pygidium, x 10. PLATE 11 JACO and DAILY, Clavagnostus 102 PALAEONTOLOGY, VOLUME 17 The cephalon described by Schmidt (1942, p. 351, pi. 21, figs. lOo, b, c) from Doberlug, Germany, as Hypcigiwstus cf. parvifrons (Linnarsson) and assigned to Clavagnostusl sp. by Sdzuy (1957, p. 10) has a wider glabella than any known species of Clavagnostiis. This cephalon should be removed from Clavagnostus. Clavagnostus milli sp. nov. Plate 1 1, figs. 12-17 Material. Three cephala and three pygidia are used for descriptive purposes. A few other specimens of this rare species are known. Holotype. The best preserved cephalon, UT 86860b (PI. 1 1, fig. 12), is selected as the holotype. Diagnosis. Cephalon with well-rounded glabellar front, anteriorly placed glabellar node, angular glabellar rear, and very long, markedly divergent cephalic spines; no preglabellar median furrow. Pygidium with wide axis extending posteriorly on to border; simplimarginate, with long border spines. Description. Markedly convex cephalon about as wide as long. Moderately wide, convex, elevated rim; narrow, shallow marginal furrow. No preglabellar median furrow. Very long, markedly divergent spines arise from broad tumid bases. Glabella length 0-65 to 0-7 that of cephalon; at its widest glabella width between 0-25 and 0-3 that of cephalon. Strongly convex glabella has an elliptical outline; it is bounded by narrow, moderately deep axial furrows which shallow to the anterior. Cheeks smooth. Large basal lobes with narrow connecting band behind angular glabellar rear. Small node on anterior part of glabella. In front of this node the glabella is somewhat depressed; this depressed area probably represents an anterior glabellar segment. Glabella is widest in region of node. Moderately convex pygidium about as wide as is long. Moderately wide lateral borders; moderately wide, convex, elevated rim; shallow marginal furrow. Long divergent border spines. Wide posterior border with a narrow posterior marginal furrow and a wide, flatly convex, elevated rim. Narrow, shallow shoulder furrows; narrow shoulders with fulcra placed about midway between acrolobe margins and axial furrows. Narrow (sag.) convex articulating half-ring ; narrow articulating furrow is shallow at centre and deepens abaxially. Smooth pleural areas. Pygidial axis extends full length of acrolobe and protrudes slightly on to the border ; axis outlined by moderately wide and deep axial furrows. Large clavagnostid pits occur about two-thirds of distance from anterior to posterior of axis. Strongly convex anterior part of axis is distinctly elevated above the pleural fields. Posterior part of axis is very slightly depressed beneath level of adjacent pleural regions. Axis wide at anterior (about 0-4 width of pygidium); it is very slightly constricted in the region of a second axial segment. From a little distance anterior of the pits, the axial furrows are straight and converge evenly to the posterior marginal furrow. Width of axial posterior about 0-2 to 0-25 that of distance between spines. Prominent, centrally placed, elongated node or keel on anterior part of axis. No sign of transverse axial furrows on anterior part of axis. Discussion. Clavagnostus milli is closest to C. repamlus of the described species of Clavagnostus. The long cephalic and pygidial spines distinguish milli from repamlus. JAGO AND DAILY; CLAVAGNOSTUS FROM TASMANIA 103 The pygidium of C. chipiquensis (Rusconi), illustrated by Poulsen 1960 (pi. 1, fig. 14) and herein (PI. 11, fig. 11), is similar to that of milli. The axis of milU seems to be wider in relation to the width of the acrolobe than that of chipiquensis. As noted above, the cephalon of C. chipiquensis may be the specimen described by Poulsen as Peronopsis ultitna. If this is so, then C. ehipiquensis is in the C. hurnsi group of species whereas C. milli is in the C. repandus group of species. Occurrence and age. Lower fauna at Christmas Hills (lat. 40° 54 T S., long. 144° 29-8' E.); its age is late Middle Cambrian, probably of the Lejopyge laevigata I Zone (Jago and Buckley 1971). Clavagnostus burnsi sp. nov. Plate 12. figs. 1-10 Material. Most of the Riana material comes from an approximately 5 mm thickness of sediment. The five cephala and five pygidia illustrated show the essential features of this species. Holotype. The cephalon, UT 92584, figured as PI. 12, fig. 8, is selected as holotype. Diagnosis. Cephalon strongly convex with narrow, shallow marginal furrow and moderately wide, elevated rim and long thin cephalic spines; glabellar front pointed, rear angular; preglabellar median furrow shallows anteriorly and does not always reach marginal furrow. Pygidium moderately convex with narrow shallow marginal furrow and moderately wide convex elevated rim; posterior rim wider than lateral, with long thin border spines; pygidial axis lanceolate extending to posterior border; median elongated ridge along anterior part of axis; shallow clavagnostid pits in posterior part of axis which is depressed below level of smooth pleural areas ; posterior end of axis bluntly rounded, narrow axial rear. Description. Small strongly convex cephalon has steep acrolobe margins. Border consists of narrow, shallow marginal furrow and a moderately wide elevated rim. Border narrows slightly to posterior. Long, thin spines emerge from low on postero- lateral corners. Glabella outlined by narrow, moderately deep furrows. Smooth cheeks divided in front by a preglabellar median furrow which is moderately deep at the posterior but shallows anteriorly and does not always meet the marginal furrow. In some figured specimens the convexity of the cephalic acrolobe obscures the fading of the pre- glabellar furrow. In some specimens it tends to widen slightly at the anterior. Glabella has a length about 0-65 that of cephalon. Glabellar rear, angular. Small, simple, unconnected basal lobes. Narrow glabella is widest just in front of basal lobes where it is about 0-4 to 0-45 the width of cephalon. From its widest the glabella narrows gradually to a sharply pointed front; glabella has an over-all elongated tear-drop outline. Glabella smooth except for anterocentrally placed, low, elongated node (not visible in figured specimens). Moderately convex pygidium about as wide as is long. Moderately wide border consists of a narrow, shallow marginal furrow and a moderately wide, convex, elevated rim. Posterior rim is a little wider than lateral rims which narrow anteriorly; border spines, long and thin. Shoulder furrows continuous with marginal furrows; shoulders, convex, elevated; fulcra placed just abaxial to midpoint between axial furrows and anterolateral corners. Articulating device is nowhere well preserved. 104 PALAEONTOLOGY, VOLUME 17 Pleural areas smooth. Axis extends entire length of acrolobe; it is outlined by narrow, moderately deep axial furrows. No sign of transverse axial furrows. Clavagnostid pits occur posteriorly about 0-65 to 0-70 the distance from the anterior to the posterior of axis. Posterior part of axis is depressed below level of pleural areas; anterior part of axis stands out somewhat above pleural areas. Elongated ridge extends along centre of anterior part of axis. Discussion. The position of Clavagnostus burnsi in relation to other species of Clava- gnostus is discussed above. If the specimen figured by Poulsen (1960, pi. 1, fig. 13) as the pygidium of Peronopsis ultima is the cephalon of C. citipiquensis (Rusconi), then this is the nearest species to C. burnsi. The cephala are very similar. However, the pygidial spines of C. burnsi are larger than those of chipiquensis. The bluntly rounded pygidial axial rear of chipiquensis is wider than the rather narrow axial rear of burnsi. The clavagnostid pits of C. burnsi are placed slightly more to the posterior than those of C. chipiquensis. The differences between C. burnsi and C.(?) rawlingi sp. nov. are noted in the discussion of rawlingi. Occurrence and age. C. burnsi comes from within the upper sedimentary sequence of Radfords Creek Group of the Dial Range Trough (Burns 1964) as exposed near Riana in a quarry at lat. 41° 13 0' S., long. 146° 00-2' E. and also at lat. 41° 12-7' S., long. 146° 00 00' E. ; its age is early Upper Cambrian, Mindyallan Stage (Jago 1972a). Clavagnostus{l) rawlingi sp. nov. Plate 12, figs. 11,12 Material. One cephalon, UT 92719, and one pygidium, UT 92727, from the same locality near St. Valentines Peak are provisionally placed in the same species. The pygidium has slightly pitted pleural areas, a feature not previously seen in any species of Clavagnostus ', however, the other characters of the pygidium indicate that it belongs to Clavagnostus or a closely related genus. Holotype. The pygidium, UT 92727 (PI. 12, fig. 12), is selected as holotype. Diagnosis. Cephalon with long narrow glabella with pointed front and rounded or subrounded rear; preglabellar median furrow well developed. Pygidium with wide posterior margin; axis slightly shorter than pleural areas; wide elongated median ridge on anterior part of axis; clavagnostid pits elongated with posterior ends con- nected with axial furrows; pleural areas gently pitted near marginal furrow in posterior two-thirds of pygidium. EXPLANATION OF PLATE 12 Pigs. 1 10. Clavagnostus burnsi sp. nov. Quarry near Riana. 1, UT 92600, cephalon, X 14. 2, UT 92600, cephalon, x 13-2. 3, UT 92585, pygidium, x20-8. 4, UT 92585, pygidium, xl7-5. 5, UT 92584, cephalon, x 30. 6, UT 92594, pygidium, x 17. 7, UT 92593, cephalon showing fading of preglabellar median furrow, X 18-5. 8, UT 92584, holotype cephalon, X 23. 9, UT 92584, pygidium, x 20. 10, UT 92597, pygidium, x 15-8. Figs. 11 12. C lavagnostusO rawlingi sp. nov. Near St. Valentines Peak. 1 1, UT 92719, cephalon, X 18-7. 12, UT 92727, holotype pygidium, X 17. Fig. 13. C/«vag«fM/n.v sp. 2. Timber track on west side of Sugarloaf Gorge. 13, UT 92601, pygidium, x2L4. Fig. 14. Clavagnostus sp. I. Upper fauna at Christmas Hills. 14, UT 868721, pygidium, x22. PLATE 12 JAGO and DAILY, Clavagnostus 106 PALAEONTOLOGY, VOLUME 17 Description. Cephalon about as wide as is long. Narrow, moderately deep marginal furrow and a convex, elevated, moderately wide rim. Posterior furrows are wide and deep at the margins becoming shallower adaxially, and if they reach the glabellar furrows, they do so faintly. Posterior rims are wide, elevated, and convex ; they narrow adaxially. Posterolateral spines are not visible; the one spine base present is large. The basal lobes, which are connected behind the rounded (or subrounded) glabellar rear, are not clearly outlined and appear to merge both with the cheeks and the posterior rim. There are pits and depressions on the cheeks of the single available cephalon, but it cannot be determined whether or not these are natural or due to tectonism. Glabella, long (about 0-7 length of cephalon) and narrow (about 0-2 width of cephalon at widest part of glabella). Glabella is outlined by moderately deep axial furrows. Moderately deep preglabellar median furrow extends to marginal furrow. Glabella has an elongated spearhead outline with a pointed front. Large, centrally placed, elongated node reaches its maximum elevation at its anterior. Pygidium has a wide posterior border between short, thick, elevated border spines. Narrow, shallow, marginal furrow; rim is wide and gently convex between the spines, with a short median region which extends forward to meet the axis. Lateral rims narrow anteriorly. Shoulder and marginal furrows are continuous ; narrow, convex shoulders ; neither the facets nor the fulcra are clearly visible. Narrow, shallow articulating furrow arches gently to the posterior. Narrow, articulating half-ring has an elongated lens-shaped outline. Pleural areas gently pitted close to the marginal furrow in central and posterior parts of the pygidium. Axis does not extend as far to the posterior as do the pleural areas; it has a wide, blunt posterior margin. No transverse furrows are visible on the anterior part of the axis. Deep, narrow axial furrows; from the anterior they converge gradually to a position (0-25 length along axis) where the axis is about 0-25 the width of the pygidium. From this constriction the furrows diverge until axis is widest at about its midpoint, where it is about 0-3 the width of the pygidium. From the midpoint the axial furrows are straight and converge to the axial posterior where the axis has a width about 0- 1 that of pygidium. Strongly elevated median ridge (about one-third axial width) is most prominent between anterior waist and midpoint of axis and also extends forward less prominently to articulating furrow. Ridge is widest at axial constriction. Two deep, longitudinally elongated, clavagnostid pits occur in the posterior part of the axis. From their anteriors the pits are arched adaxially to join the axial furrows about 0-8 of the distance from the anterior to the posterior of the axis. Across the anterior end of the pits is a moderately deep, transverse furrow. Discussion. The combination of a very narrow glabella, elongated clavagnostid pits which meet the axial furrows, and pitted pleural areas probably indicates that rawiingi does not belong in Clavagnostus but in a closely related new agnostid genus. Although there are no other specimens with which either the one available cephalon or the one known pygidium can be linked, it is not absolutely certain that the cephalon and the pygidium described above belong in the one species. Hence rawiingi is placed question- ably in Clavagnostus. As noted in the generic discussion, Clavagnostus{l) rawiingi appears to be related to the C. hurnsi group of species. C.(?) rawiingi differs from all known species of Clavagnostus in that the pygidial pleural areas are slightly pitted. The unique nature JAGO AND DAILY: CLAVAGNOSTUS FROM TASMANIA 107 of the clavagnostid pits also distinguishes C. (?) rawlingi from all species of Clava- gnostus. The width of the glabella in relation to that of the cephalon is less in rawlingi than in any species of Clavagnostus. The pygidial spines and pygidial axis of rawlingi are shorter than those of either C. clnpiquensis (Rusconi) or C. burnsi sp. nov. Occurrence and age. Clavagnostiis(l) rawlingi sp, nov. comes from near St. Valentines Peak, lat. 41° 21-6' S., long. 145° 44-3' E.; its age is either late Middle Cambrian Lejopyge laevigata III Zone or the Middle Cambrian/Upper Cambrian Passage Zone (Jago 1972^). Clavagnostus sp. 1 Plate 12, fig. 14 Material. One reasonably well-preserved, partial pygidium (UT 868721) is available. Description. Moderately convex pygidium; moderately wide gently convex rim and a narrow, shallow marginal furrow. Long, well-developed border spines. Shoulder areas, not visible. Narrow (sag.), convex articulating half-ring extends almost full width of axis; narrow, shallow articulating furrow. Anterior part of axis (in front of clavagnostid pits) stands out well above the smooth pleural areas. At the pygidial anterior the axis has a width about 0-3 that of the pygidium. Axis is outlined by narrow, shallow furrows; it does not extend as far to the posterior as do the pleural fields. Clavagnostid pits, poorly outlined; they occur toward the posterior of the axis about 0-75 of the distance along the axis. On the anterior part of the axis is a long, centrally placed ridge. Discussion. As only one specimen is known, this pygidium is referred to Clavagnostus sp. 1. It differs from C. milli in that (u) the axis of Clavagnostus sp. 1 does not extend to the posterior border, whereas that of C. milli extends slightly on to the border and {b) the pits of Clavagnostus sp. 1 are placed more to the posterior than those of C. milli. Occurrence and age. Clavagnostus sp. 1 comes from the upper fauna at Christmas Hills (lat. 40° 54T' S., long. 144° 29-8' E.); its age is late Middle Cambrian, either of the Lejopyge laevigata I Zone or the L. laevigata II Zone. Clavagnostus sp. 2 Plate 12, fig. 13 Material and measurement. One poorly preserved pygidium, UT 92601, of length (including articulating half-ring), 1-4 mm is available. Description. Moderately convex pygidium about as wide as is long. Narrow, shallow marginal furrow; wide, flatly convex rim. Posterior margin extends slightly forward at its centre to meet axis. Border spines of indeterminate length are present. Arti- culating furrow seems narrow and shallow with a shallow pit at either end ; articulating half-ring appears narrow (sag.) and gently convex. Pleural areas probably smooth. Axis outlined by wide, deep furrows. Clavagnostid pits found about 0-65 of distance from anterior to posterior of axis. Anterior part of axis stands out well above pleural areas; pleural areas sit well above axial posterior. Discussion. This specimen is too poorly preserved to assign it to any particular species, and it is referred to Clavagnostus sp. 2. 108 PALAEONTOLOGY, VOLUME 17 Occurrence and age. Clavagnostus sp. 2 comes from a siltstone exposed along an old timber track on the west side of Sugarloaf Gorge, lat. 41° 15-4' S., long. 146° 04-2' E. It comes from within the lower sedimentary sequence of the Radfords Creek Group of the Dial Range Trough (Burns 1964); its age is probably of the late Middle Cambrian, Lejopyge laevigata III Zone or the Middle Cambrian/Upper Cambrian Passage Zone (Jago 1972fl). Acknowledgements. This work was done in the Department of Geology and Mineralogy, University of Adelaide. One of us (J. B. J.) was in receipt of a Commonwealth Post Graduate Award. We are indebted to Professor A. R. Palmer (State University of New York at Stonybrook), Drs. V. Jaanusson and H. Mutvei (Naturhistoriska Riksmuseet, Stockholm), and the late Professor F. Brotzen (Swedish Geological Survey) for permission for one of us (B. D.) to make rubber casts of the South American and Scandinavian agnostids discussed in this paper. Mr. M. R. Banks (University of Tasmania) arranged the transfer of some specimens to Adelaide. Mr. G. Pike and Dr. K. L. Burns assisted one of us (J. B. J.) in locating the St. Valentines Peak and Riana fossil localities. REFERENCES BANKS, M. R. 1956. The Middle and Upper Cambrian Series (Dundas Group and its Correlates) in Tasmania. El Sistema Cdmbrico, 2, Proc. 20th Int. geol. Congr. 165-212. 1962. Cambrian System. In The Geology of Tasmania. J. geol. Soc. Aust. 9, 127-145. BLISSETT, A. H. 1962. Geology of the Zeehan Sheet, 1-mile Geol. Map. Series K 55-5-50. Explan. Rep. Geol. Surv. Tasm. BURNS, K. L. 1964. Geology of the Devonport Sheet, 1-mile Geol. Map. Series K 55-6-29. Ibid. CHERNYSHEVA, N. E. (ed.) 1960. Arthropoda, Trilobitomorpha and Crustacea. Osnovy Paleontologii. Moscow. Akad. Nauk SSSR, pis. 1-12. [In Russian.] DREWES, M. and palmer, a. r. 1957. Cambrian rocks of Southern Snake Range, Nevada. Bull. Anier. Petrol. Geol. 41, 104-120. HOLLAND, c. H. (ed.) 1971. Cambrian of the New World. Wiley, London. HOLM, G. and WESTERGARD, A. H. 1930. A Middle Cambrian fauna from Bennett Island. Mem. Acad. Sci. URSS, 21, no. 8. HOWELL, B. F. 1937. Cambrian Centropleura vermontensis fauna of northwestern Vermont. Bull. geol. Soc. Amer. 48, 1147-1210, pis. 1-6. 1959. Agnostidae. In moore, r. c., 1959, pp. 172 186 (q.v.). HUPE, p. 1953. Classification des trilobites. Ann. Paleont. 39, 61 168 (1-110). JAGO, J. B. 1972fl. Biostratigraphic and taxonomic studies of some Tasmanian Cambrian Trilobites. Unpubl. thesis, Univ. Adelaide. 19726. Two new Cambrian trilobites from Tasmania. Palaeontology, 15, 226-237, pi. 44. and BUCKLEY, J. h. 1971 . An abrupt Upper Middle Cambrian faunal change, Christmas Hills, Tasmania, Australia. Pap. Proc. R. Soc. Tasm. 105, 83-85. KHALFIN, L. L. (ed.) 1960. Biostratigrapliy of the Palaeozoic of the Sayan-Altai mountain region. 1 . Lower Palaeozoic. Trudy sib. nauchno-issled. Inst. Geol. Geofiz. miner. Syr'. 19. [In Russian.] KOBAYASHi, T. 1939. On the Agnostids (Part 1). J. Fac. Sci. Tokyo Univ. Sec. 2, 5, 69-198. 1943. Cambrian Faunas of Siberia. Ibid. 6, 271-334, pis. 1 3. LAZARENKO, N. p. and NIKIFOROV, H. I. 1968. Trilobite assemblage from Upper Cambrian deposits of the Kulyurne River (north-western Siberian Platform). Uchen. Zap. naucimo-issled Inst. geol. Arkt. Paleont. Biostrat. 23, 20-80, pis. 1-15. [In Russian.] LERMONTOVA, E. 1940. Arthropoda. In vologdin, a. and others. Atlas of the leading forms of the fossil faunas of the U.S.S.R., vol. I, Cambrian. State Editorial Office for Geological Literature, Moscow. [In Russian.] MOORE, R. c. (ed.) 1959. Treatise on Invertebrate Paleontology, Part O, Arthropoda 1. Univ. Kansas Press and Geol. Soc. Am. NORTH, E. K. 1971. The Cambrian of Canada and Alaska. In Holland, c. h., 1971, 219-324 (q.v.). OPIK, A, A. 1961a. Alimentary caeca of agnostids and other trilobites. Palaeontology, 3, 410-438, pis. 68-70. 19616. Cambrian geology and palaeontology of the headwaters of the Burke River, Queensland. Bulk Bur. Min. Re.wur. Aust. 53, pis. 1-24. JAGO AND DAILY: CLAVAGNOSTUS FROM TASMANIA 109 OPiK, A. A. 1963. Early Upper Cambrian fossils from Queensland. Ibid. 64, pis. 1-9. 1967. The Mindyallan Fauna of North-Western Queensland. Ibid. 74, pis. 1-67. 1970. Nepeid trilobites of the Middle Cambrian of northern Australia. Ibid. 113, pis. 1-17. POKROVSKAYA, N. V. 1960u. Miomera. In Chernysheva, n. e., 1960, 54-61 (q.v.). 19606. Clavagnostus. In khalfin, l. l., 1960, 161 (q.v.). POULSEN, c. 1960. Fossils from the late Middle Cambrian Bolaspidella Zone of Mendoza, Argentina. Mat.-fys Meddr danske Vid. Selsk. 32, (11), 1-42, pis. 1-3. RESSER, c. E. 1938. Cambrian system (restricted) of the southern Appalachians. Spec. Pap. geol. Soc. Am. 15. ROBISON, R. A. and PALMER, A. R. 1968. Revision of Cambrian Stratigraphy, Silver Island Mountains, Utah. Bull. Amer. Ass. Petrol. Geol. 52, 167-171. RUSCONi, c. 1952. Varias especies de trilobitas del Cambrico de Canota. Revta Mus. Hist. nat. Mendoza, 6, 5-17. SCHMIDT, w. E. 1942. Die mittelkambrische Fauna von Doberlug. Jb. Reichstelle Bodenforsch. 62, 344-399, pis. 21-24. SDZUY, K. 1957. Revision der mittelkambrischen Trilobiten von Doberlug. Senckenherg. letli. 38, 7-28. SHABANOV, Y. Y., SAVITSKY, V. E. and CHERNYSHEVA, N. E. 1967. Biostratigraphy of the Maisk Stage in the Igarsk region. Trudy sib. nauclmo-issled. Inst. Geol. Geofiz. miner. Syr'. 55, 59-65. [In Russian.] WESTERGARD, A. H. 1946. Agnostidea of the Middle Cambrian of Sweden. Sver. geol. Unders. Avli. C. All, 1-140, pis. 1-16. J. B. JAGO Department of Applied Geology South Australian Institute of Technology North Terrace Adelaide, South Australia 5000 B. DAILY Department of Geology and Mineralogy University of Adelaide G.P.O. Box 498B Manuscript received 24 October 1972 Adelaide, South Australia 5001 i t .iv f"* rt ' ■ T ■ \ . .••i-.v ■ 4i»rf.>»»,v.*-a';w«rj; »' * V. M ■ 4' ■ -i,. - '■)->.'■ -■■■! .•' •'■■ >:■ ifl ..;?vi.-'.. . '• «fi- = - • " _ , « '.•■•’■{.;■ ■ . ■.«,,:=? - ,i>.l(f' -, : . • . ' • ;" ■ ' ' '•■■'.• t pi-'-o ■:■ . , ^ -' V, V 4 . .& . ■■ ... ;*■' , h >■ _•"- -A . ■ , ■ .■' ,' '..; '■ A” ..', , - ^ ' i:- t. ■ ■- -is* i; ^ "'«■ J’ i' ! . ' ^■ ' ^ ‘a •; , ■ '^7 ■■ •. '■':;j Tlrl A NEW PELAGIC TRILOBITE EROM THE ORDOVICIAN OF SPITSBERGEN, IRELAND, AND UTAH by R. A. FORTEY Abstract. Opipeuter inconnivus gen. et sp. nov. is described. It is included with Crcmastoglottos Whittard in a new family, the Opipeuteridae, which is probably closest to the Remopleurididae. It is also possible that Opipeuter is related to the enigmatic trilobite Bohemilla Barrande. The functional morphology of Opipeuter indicates a pelagic mode of life. S T u D Y of early Ordovician trilobites from the Valhallfonna Formation, Ny Friesland, Spitsbergen, and the Tourmakeady Limestone, Co. Galway, western Ireland, revealed some fragments of a trilobite which suggested that a new and peculiar species was common to both faunas. In 1971 the discovery of an almost complete dorsal exo- skeleton in Spitsbergen enabled a definite association of the known fragmentary remains. Because of its particular interest this trilobite is described and discussed separately from the rest of the systematic work on the Spitsbergen and Irish faunas. Material of the new trilobite is deposited in three institutions: Paleontologisk Museum, Oslo (PMO NF), Sedgwick Museum, Cambridge (SMA), and the British Museum (Natural Flistory) (BM It). The stratigraphy of the Spitsbergen Ordovician has been described by Fortey and Bruton (1973) and the stratigraphical terminology given therein is used in the present account. The material from the Tourmakeady Limestone is from pink, brecciated limestones of Gardiner and Reynolds’s (1909, pi. 4) locality 58. A discussion of the stratigraphy and correlation of the rocks of this area has been given by Dewey, Rickards, and Skevington (1970). SYSTEMATIC DESCRIPTION Superfamily remopleuridacea Flawle and Corda, 1847 Family opipeuteridae fam. nov. Type genus. Opipeuter gen. nov. Diagnosis. Remopleuridacea with greatly developed, convex eyes. Glabella with sub- triangular outline, tapering strongly forward, with one or two pairs of lateral glabellar furrows, and downturned glabellar tongue anteriorly. Fixed cheeks reduced or absent, posterior border modified to a spine extending backwards behind occipital ring in Opipeuter. Long narrow thorax (eleven segments) with convex axis and very short pleurae. Pygidium with entire margin, axis continued into a stout spine. Genera included in family. Opipeuter gen. nov., Cremastoglottos Whittard 1961 (known from cranidia only). [Palaeontology, Vol. 17, Part 1, 1974, pp. 111-124, pis. 13-14.] H 112 PALAEONTOLOGY, VOLUME 17 OPiPEUTER gen. nov. Type species. Opipeuter inconnivus sp. nov. Derivation of name. Greek, ‘one who stares’. Diagnosis. Opipeuteridae with one pair of lateral glabellar furrows. Occipital furrow with a fairly straight, transverse course. Surface sculpture of fine raised lines arranged in polygons. Opipeuter inconnivus sp. nov. Plates 13, 14; text-figs. 1, 3a Derivation of name. Latin, inconnivus, ‘unsleeping’. Diagnosis. As for genus. Material. Holotype. Slightly damaged dorsal exoskeleton lacking free cheeks, from the Olenidsletta Member of the Valhallfonna Formation, occurring an estimated 60 m from the base of the Member on Olenidsletta north of Papegoyeneset, PMO NF 2062. Other Spitsbergen material includes one exfoliated cranidium with a few thoracic segments attached, SMA 84010, apart from which all other material consists of isolated cranidia, free cheeks, and pygidia. Material identified so far as belonging to this species includes: Cranidia: PMO NF 700, 875, 2982-2984, 2986, 2989, 2995-2996, 2998; BM It 9799; SMA 84011. Pygidia: PMO NF 2987, 2993-2994, 2990, 2997; SMA 84012. Free cheeks: PMO NF 2980-2981, 2985, 2988. The material from western Ireland consists of three cranidia: BM It 9795-9797, and one free cheek, BM It 9798. Material from Utah consists of a single, well preserved free cheek, BM It 10279. Stratigraphic occurrence. In Spitsbergen the species is confined to the Olenidsletta Member of the Valhall- fonna Formation, the lowest occurrence being 9 m above the base of the Member, the highest 100 m from its base. Through most of this range it can be found with careful searching, but always as a rare element of the total fauna. The material from the Tourmakeady Limestone is from Locality 58 of Gardiner and Reynolds ( 1909). The free cheek from Utah is from the upper part of the Fillmore Limestone, locality H of Hintze (1953). EXPLANATION OF PLATE 13 Opipeuter inconnivus gen. et sp. nov., Olenidsletta Member, Valhallfonna Formation, Ny Friesland, Spitsbergen. Figs. 1, 10. Free cheek ( x 1 1 ) in dorsal, lateral views. 8 m from base of Olenidsletta Member. Note slender genal spine. PMO NF 2981. Figs. 2,9. Holotype, dorsal exoskeleton lacking free cheeks ( x 8) in dorsal and left-lateral views. Olenidsletta, 60 m from base of Member. Axis of thorax and median part of cranidium are exfoliated, and the ninth thoracic segment is slightly displaced on the right-hand side. PMO NF 2062. Figs. 3, 4, 5. Cranidium (x 13), retaining exoskeleton except on mid part of occipital ring, dorsal, lateral, and anterior views. 60 m from base of Member. Note characteristic polygonal surface sculpture, palpebral lobe, and backward pointing border spine on left of specimen. PMO NF 875. Figs. 6, 7, 8. Small cranidium (x 15) in dorsal, anterior, and lateral views. Horizon as fig. 1. Note narrow, horizontal anterior border in fig. 7. PMO NF 2982. Fig. 1 1. Cranidium (x 10), retaining exoskeleton. 30 m from base of Member. SMA 8401 1. Fig. 12. Imperfectly preserved, llattened internal mould of cranidium and eight thoracic segments (x2). SMA 84010. PLATE 13 FORTEY, Opipeuter 114 PALAEONTOLOGY, VOLUME 17 Description. Exoskeleton fusiform, sagittal length (excluding pygidial spine) three times the transverse width across the mid part of the thorax. Cranidium one-quarter the length of the exoskeleton, with triangular outline, transverse width across the base of the glabella equal to, or slightly exceeding sagittal length. The glabella occupies the major part of the cranidium, transversely convex, anteriorly sloping steeply downwards and slightly recurved into narrow glabellar tongue. Glabella expands in width very slightly in front of the occipital ring to a point at about half (sag.) its length, anteriorly tapering rapidly and uniformly to the anterior border, such that the trans- verse width at the anterior border is one-fifth the maximum width of the glabella. The one pair of lateral glabellar furrows meet the axial furrows just in front of the point of maximum width of the glabella, running inwards and backwards at an angle of between 40° and 60° to the sagittal line (angle measured from photo- graphs of eight cranidia in dorsal view) and extending one-fifth across the glabella. The furrows are shallow on the dorsal surface of the cranidium, and on some specimens of internal moulds may not be clearly visible, particularly if some crushing has taken place. Occipital furrow narrow and deep, medially transverse, lateral extremities sloping backwards at a low angle. Occipital ring of uniform width (sag., exsag.), about one-fifth sagittal length of glabella, not clearly differentiated from the posterior border of the fixed cheek. Axial furrows shallow adjacent to the posterior part of the glabella in front of the occipital ring, deepening forwards. text-hg. 1. Reconstruction of Opipeuter inconnivus gen. et sp. nov. in dorsal (a) and lateral (b) views (x 5). Fi.xed cheeks reduced to narrow (trans.) triangular areas adjacent to the posterior, non-tapering part of the glabella, and continuing its downward slope. On larger cranidia this triangular area almost appears to be part of the glabella, as the occipital ring extends abaxially behind it, and particularly on internal moulds the axial furrows are not deeply defined. On the smallest cranidium the extra-glabellar origin of the area is more obvious (PI. 14, fig. 3), and the triangular area is both distinct from the glabella FORTEY: PELAGIC ORDOVICIAN TRILOBITE 115 and lateral to the occipital ring. Along the lateral edge of the area there is a convex ridge. The posterior border is modified to form a flat, downward sloping spine which is directed backwards behind the posterior margin of the occipital ring at an angle of 20-30° to the sagittal line, reaching as far as the second thoracic segment. Palpebral lobe very narrow, almost horizontal, and lying at a level below that of the rest of the cranidium. The greatest width (trans.) of the palpebral lobe is adjacent to the anterior part of the triangular area. Posteriorly it narrows gradually towards the posterior spine; anteriorly it becomes rapidly narrow in front of the outer end of the lateral glabellar furrow, anteriorly not continuing as a distinct lobe but rather as an extremely narrow V-shaped ‘gutter’ as far as the anterior border. The palpebral furrow is gently bowed outwards medially, quite deep, meeting the short posterior border furrow behind the ridge border- ing the triangular fixigenal area. The cranidium has a narrow (sag., trans.) horizontal anterior border, rim- like, and with a convex anterior margin. This border is extremely narrow (sag.) on large cranidia, and difficult to prepare (the palpebral lobes may present similar problems), but one small cranidium (PI. 13, fig. 7) shows it very well. The facial suture is considered to run from the tip of the posterolateral spine inwards and forwards in an almost straight line bounding the palpebral lobe and the anterior ‘gutter’ as far as the anterior border. The dorsal surface of the glabella is covered with a very distinctive surface sculpture of fine raised lines which form a polygonal pattern. The polygons tend to become stretched out transversely on the sagittal region of the preoccipital glabella and on the occipital ring. This distinctive sculpture is present on even the smallest cranidium. Internal moulds are smooth. The free cheek consists largely of very large, subglobose eye with an elliptical outline. The eye has the greatest dorsoventral length of visual surface anteriorly. The lenses are hexagonal, arranged in files (vertic- ally) of about forty lenses. The total number of lenses on the largest eye is estimated at 2000, on an eye half the length there are an estimated 1600 lenses, so that the number of lenses does not change greatly in late growth stages. The largest individual lenses occur on the front, and towards the top of the eye. At the base of the eye, and to the rear they are closely packed and small. The base of the eye is marked by a furrow which is deeper posteriorly, faint and diffuse anteriorly. The wide and flat border has an almost semi- circular posterior outline, anteriorly curving upwards, narrowing, and becoming horizontal, rim-like. There is a slender, backward-pointing genal spine, originating from the back end of the border at a point opposite the lower edge of the posterior part of the eye. Between the lower edge of the eye and the postero- lateral part of the border there is a crescentic genal area, which is depressed below the level of the rest of the cheek. The border of the free cheek carries a surface sculpture of a few fine terrace lines, running sub- parallel to its outer margin. Thorax of eleven segments, of subequal width (sag.) along its length. Axis extremely wide (trans.) relative to the pleurae, which even at their widest are less than one-quarter the axial width; axis highly convex transversely, parallel sided anteriorly, tapering gradually for the posterior four segments. Each axial ring is similar in form to the occipital ring, band-like, of constant width (sag., exsag.). Axial furrows shallow, bowed outwards adjacent to each axial ring. Pleurae of greatest transverse width on the fourth, fifth, and sixth thoracic segments. The pleurae of the first thoracic segment are extremely narrow (trans.), pointed, triangular, with convex bands at the anterior and posterior margins. Pleural furrow of the first thoracic pleura does not reach its tip. The anterior margin directed markedly backwards, thereby allowing for the backward passage of the posterior spines of the cranidium. The pleurae of the second segment are of similar form to those of the first, but transversely wider. Posteriorly the pleurae become gradually subrectangular in outline and the pleural tips are truncate. Each pleura is crossed by a diagonal pleural furrow, which is fairly deep and reaches the tip of the pleura but is shallower abaxially. Adjacent to the axis the furrow is broad, almost equal in width (exsag.) to the adjacent axial ring. Within this broad part of the furrow there is a gently inflated triangular area adjacent to the axial furrow. The articulation between segments is sub- jacent to the axis, and consists of a boss on the most anterior part of each thoracic segment which engages with an excavation on the posterior part of the preceding pleura. Thoracic pleurae are not in contact abaxially from the articulation— there is an open triangular area between them which gives a zigzag out- line to the thorax. Articulating half-rings on the axis are wide (sag.), almost reaching as far as the front margin of the preceding axial rings. Pygidium (excluding articulating half-ring) three-quarters as long as wide, and about one-fifth total length of dorsal exoskeleton. The convex axis tapers gradually posteriorly (the axial furrows enclosing an angle between 35° and 45°) to 0-7 axial length of pygidium. Only the anterior axial ring is completely developed; 116 PALAEONTOLOGY, VOLUME 17 the ring furrow is much shallower medially. The second ring furrow is modified in a peculiar way: the outer parts, adjacent to the axial furrows, are deep and run more or less transversely up the side of the axis parallel to the deeper part of the anterior ring furrow; the inner parts curve sharply backwards parallel to the sagittal line, almost to the tip of the axis, so that the ring furrow is modified as a pair of hook-shaped furrows. The mid part of the axis posterior to the anterior ring furrow is transversely flat. The tip of the axis is prolonged into a subcircular postaxial spine of length slightly less than that of the pygidium; the spine continues in line with the tip of the axis, or is directed slightly upwards. Axial furrows shallow posteriorly, not defined around the tip of the axis. Pleural fields slope steeply away from the axis, both laterally and postaxially. Two pairs of pleural and two pairs of interpleural furrows, the posterior pleural and interpleural furrows fainter than the anterior, especially on internal moulds. The anterior interpleural furrows run almost to the margin of the pygidium, and with the well-defined anterior ring furrow on the axis, serve to separate the anterior segment from the rest of the pygidium, and give it a close resemblance to a free thoracic segment. The downward slope of the pleural fields stops abruptly at the horizontal posterior border, which has its greatest width posterolaterally, anteriorly tapering in width to the anterolateral corners of the cranidium where it is scarcely to be distinguished from the anterior parts of the first pygidial pleurae, posteriorly decreasing in width very slightly. The posterior margin of the border is convex in out- line, broadly rounded about the sagittal line, coming to a rounded point medially. Pygidial doublure doubled back horizontally beneath posterior border, of similar width and outline, with surface sculpture of terrace lines subparallel to the posterior margin of the pygidium. Ontogeny. The small cranidium (PI. 14, fig. 3) (sagittal length 0-85 mm) is readily recognizable as that of Opipeuter. It differs from the adult in the broadly rounded anterior outline of the glabella, and in the narrow, deeply incised axial furrows which are parallel to the sagittal line. A free cheek which probably belonged to a comparably sized cranidium (PI. 14, fig. 5) has the same general form as that of the adult, but differs in the following features: the border of the cheek is not clearly differentiated; the genal spine is stouter, more posteriorly positioned (only the base of the spine is present on the small cheek); eye lenses are only present dorsally (about 250 in number), the lower part of the visual surface being apparently smooth. The lenses are of equal size over the surface of the eye. EXPLANATION OF PLATE 14 Opipeuter inconnivus gen. et. sp. nov., Olenidsletta Member, Valhallfonna Formation, Ny Friesland, Spitsbergen. Fig. 1. Pygidium (x 10) in oblique lateral view, showing spine continuing backwards from tip of pygidial axis. About 80 m from base of Member. PMO NF 2987. Figs. 2, 3. Very small cranidium in lateral ( x 15) and dorsal ( X 30) views, showing essentially similar form of cranidium to that of adult. 80 m from base of Member. PMO NF 2984. Figs. 4, 6, 7. Free cheek in lateral, dorsal, and anterior views ( x 20). From the same bed as the free cheek of PI. 13, fig. 1. PMO NF2980. Fig. 5. Lateral view of extremely small free cheek ( x 37) showing the small number of dorsally concentrated lenses on the eye. Base of genal spine on right. Same bed as the small cranidium in fig. 3. PMO NF 2985. Figs. 8, 9, 10. Pygidium, preserved in full relief and retaining exoskeleton, in lateral, posterior, and dorsal views (x 14). Posterior view shows the considerable transverse convexity. From a stream section on Olenidsletta about 80-85 m from the base of the Member. PMO NF 2990. Opipeuter inconnivus gen. et sp. nov., Tourmakeady Limestone, Co. Galway, western Ireland. Fig. 1 1 . Cranidium, dorsal view ( x 6), poorly preserved on right side. BM It 9795. Figs. 12, 13. Free cheek in dorsal, lateral views from same block as cranidium in fig. 1 1 ( x 14). BM It 9798. Opipeuter sp., upper part of Olenidsletta Member, Valhallfonna Formation, Ny Friesland, Spitsbergen. Fig. 14. Free cheek (x 12) in lateral view, showing extra-marginal flange on posterior part of border. 96 m from base of Member. PMO NF 52. Figs. 15, 16. Internal mould of cranidium (x6) in dorsal, anterior views. Note long and narrow glabellar tongue. 1 02 m from base of Member. PMO NF 299 1 . PLATE 14 FORTEY, Opipeuter 118 PALAEONTOLOGY, VOLUME 17 Changes which occur during the ontogeny of the cephalon thus include a gradual relative increase in the length of the glabella associated with a more hyperbolic anterior outline, a shallowing and inward migra- tion of the posterior part of the axial furrows, and an increase in number of lenses on the eye and their ventral migration, together with a differentiation in lens size such that the larger lenses are anteriorly and dorsally positioned. The glabellar tongue of the small cranidium does not curve downwards anteriorly to the same degree as on larger cranidia. Variation. The material from Spitsbergen compares closely with that from the Tourmakeady Limestone, although a pygidium has not yet been found from western Ireland. The free cheek from the Fillmore Lime- stone is similar to that illustrated on Plate 13, fig. 1 from the Valhallfonna Formation. Variation among specimens of O. inconnivus from Spitsbergen seems to be confined to relatively small differences in the anterior outline and relative length of the glabella, some specimens being relatively shorter and slightly more truncate anteriorly than others (PI. 13, figs. 2, 11). The border of the free cheek is narrower on some specimens than others (compare PI. 13, fig. 10; PI. 14, fig. 4), One cranidium and one free cheek occurring stratigraphically above all material of O. inconnivus from the Valhallfonna Formation is believed to represent the only material of a second species of Opipeuter. The cranidium (PI. 14, figs. 15, 16) has a glabella that tapers more rapidly than that of O. inconnivus, a narrower, longer glabellar tongue, and apparently also posterolateral cranidial spines that are directed backwards almost parallel to the sagittal line. The free cheek (PI. 14, fig. 14) possesses a flange running alongside the border of the cheek at its widest part, which is not seen on cheeks of O. inconnivus. Although this material indicates the presence of a second species of Opipeuter the single cranidium and free cheek are not considered sufficient to formally name it as such. Age. The Spitsbergen and western Ireland material is of Arenig age. The earliest occurrence of the species in the Valhallfonna Formation is with an assemblage of graptolites of early Arenig (late fruticosus Zone) age (Fortey and Bruton 1973), while the latest occurrence of the species is just below a graptolite assemblage of the Isograptus Zone in North American terms, an assemblage which may be correlated with the Zone of Didymograptus hirundo of the British sequence. The species in Spitsbergen thus ranges through the pre- hirundo Arenig (probably deflexus to gibberutus Zones of the British Arenig). The Tourmakeady Limestone locality is well dated in terms of graptolite assemblages occurring above and below (Dewey, Rickards, and Skevington 1970, p. 29) as Tate Isograptus gibberulus Zone or early Didymograptus hirundo Zone in age’ (Skevington 1971, p. 80). If the former of these alternatives proves to be the case, this correlates exactly with the upper part of the range of Opipeuter in Spitsbergen. The free cheek from the Fillmore Limestone of Utah occurs with the Zone I fauna of Hintze (1953). A ffinities. A trilobite as unusual as Opipeuter poses particular problems in assessing its relationships. The various families which have been considered are briefly dis- cussed below. Remopleurididae. It is believed that Opipeuter is most closely related to the remopleuridids, particularly to Reinopleurides and allied genera (subfamily Remopleuridinae). The glabella of these remopleuridids occupies the entire distance between the palpebral lobes, and in Remopleurides, Reniopleuridiella, Rober- giella, and Reniopleuretla is produced anteriorly downwards into a narrow glabellar tongue similar to that of Opipeuter. The postocular fixed cheeks of such Remopleurididae are reduced to spines, although trans- versely directed rather than posteriorly directed as in Opipeuter. The narrow, transversely horizontal anterior border of the cranidium of Opipeuter is closely similar to the cranidial border on such remopleuridids as Robergiella sagittalis (Whittington 1959, pi. 6, figs. 20, 21) and Reniopleuridiella caudalimbata Ross 1951 (see Hintze 1953, pi. 5, figs. 10«-c). Such a border is not to be found among species of the other families (below). Palpebral rims of remopleuridids are characteristically broader (trans.) than those ot Opipeuter, but narrow anteriorly into grooves with a V-shaped cross section adjacent to the glabellar tongue in just the same way as those of Opipeuter. Like Remopleurides, Opipeuter has eleven thoracic segments, the axis being transversely wide compared with the pleurae and the articulation close to the axial furrows. In its major morphological features Opipeuter thus seems to be more closely comparable with the Remo- pleurididae than any other trilobite family. But there are a number of important features which distinguish Opipeuter from the Remopleurididae. In Opipeuter: (i) The occipital ring is transversely very wide, and the cranidium tapers anteriorly from it. Remopleurididae FORTEY: PELAGIC ORDOVICIAN TRILOBITE 1 19 have the glabella expanding forward to a point at about the mid length of the palpebral lobes. The facial suture is similarly bowed outwards medially in remopleuridids, rather than running almost uniformly inwards-forwards as in Opipeuter. (ii) The greatly developed eyes are inflated, subglobose. Among Remopleurididae the form of the eye is characteristic and quite different : it forms a flat, near vertical band-like surface of nearly uniform height along its length (see for example Whittington 1959, pi. 19, fig. 11; Shaw 1968, pi. 2, fig. 29). (iii) The modified posterior border spines of the cranidium extend backwards behind the occipital ring, this being possible because of the extreme narrowness (trans.) of the first thoracic pleurae. No such structure is to be found among the remopleuridids. (iv) The pygidial margin is entire. In Renmpleurides and allied genera a spinose margin to the pygidium is the rule. The postaxial pygidial spine of Opipeuter has no parallel among mature remopleuridid pygidia, although long axial spines are present on some small transitory pygidia of Remopleurides. (v) The defining characters of Opipeuter are present even on small cranidia less than 1 mm long which do not resemble remopleuridid cranidia of the same size. (vi) The lateral parts of the thoracic pleurae are not in contact with one another. These differences are considered of such importance that Opipeuter is here assigned to a new family. The genus Cremastoglottos Whittard 1961, which is known only from four compressed cranidia from the early Llanvirn Hope Shales of Shropshire, England, and one probable fragmentary cranidium from the Caradocian of Bohemia (Marek 19666), possesses a number of features which suggest that it should be included in the same family as Opipeuter. In the type and only species, Cremastoglottos occipitalis {V^hittard 1940) (see Whittard 1961, p. 187, pi. 25, figs. 1-5) the general cranidial outline is similar to that of Opipeuter inconuivus, being triangular, anteriorly tapering, and produced into a narrow downward-deflected glabellar tongue. A narrow palpebral rim runs along the edge of the cranidium as in Opipeuter, and the occipital ring is transversely very wide. The main differences between Cremastoglottos and Opipeuter are that the former genus has two pairs of deeply incised glabellar furrows which do not reach the axial furrows, and possesses a curiously undulating occipital furrow quite distinct from the nearly straight transverse occipital furrow of Opipeuter. It is of interest to note that Whittard originally ( 1 940, p. 1 36) assigned Cremastoglottos occipi- talis to the Cyclopygidae, but later (Whittard 1961, p. 187) to the Remopleurididae. Telephiuidae. A number of similarities may be noted between Telephina species and Opipeuter. The glabellar outline in dorsal aspect of small specimens of Opipeuter may be compared with such species as T. hipunctata (Ulrich 1930, pi. 5, fig. 1); the occipital ring is well developed in all telephinids. The posterior part of the fixed cheek may be modified as a short spine, as in T. americana (Billings 1865) (see Whittington 1965, pi. 37, figs. 9, 18), a structure comparable with the modified posterior border of Opipeuter. The thorax of Telephina, which is completely known only in T. spinifera (Ulrich) (see Fischer 1946) with nine segments, has a broad convex axis with narrow, rather blunt-tipped pleurae, each segment being similar in these respects to those of Opipeuter. Pygidia of some species of Telephina are not greatly different from that of Opipeuter. For example the pygidium of Telephina (Telephops) bicornis (Ulrich 1930, pi. 4, figs. 12, 13) has a similar outline to that of Opipeuter, is similarly transversely convex, with a prominent axis with two axial rings and a rim-like posterior border. In spite of these similarities Opipeuter is not thought to be related to the Telephinidae. The critical dif- ferences are considered to be: (i) There is no glabellar tongue developed on Telephina species, (ii) The structure of the anterior cranidial border is quite different in Opipeuter and telephinids. In the latter the border is a steeply arched convex rim which may be reduced to a pair of spines in some species; this is quite different from the narrow, horizontal rim of Opipeuter. (iii) All the telephinids have well-developed triangular fixed cheeks with broad palpebral rims, which give the inner (sutural) margin of the eye a curved outline. In addition the postaxial spine on the pygidium of Opipeuter does not seem to be found among telephinids, nor do they have such a broad border to the free cheeks. Following Nikolaisen (1963, p. 563) it is believed that certain genera of the Komaspididae, such as Carrickia, Carolinites, and Goniophrys, are closely related to Telephina', the differences considered of importance between telephinids and Opipeuter apply equally to this group of komaspidids, except that their facial sutures do not have the strongly curved outline of Telephina. Cyclopygidae. The eleven thoracic segments of Opipeuter, the wide border of the free cheek, and the convex pygidium with its relatively long axis produced posteriorly into a spine serve to exclude the genus from the 120 PALAEONTOLOGY, VOLUME 17 Cyclopygidae. The cranidium of Opipeuter, however, shows a number of similarities with cyclopygids, which are believed to be due to homeomorphy. The generally tapering glabella, which occupies almost the entire cranidium, the downward deflected anterior glabellar tongue and the narrow, gutter-like palpebral rims can be matched with such cyclopygids as Novakella bergeroni (Kloucek 1916) (see Marek 1961, pi. 4, hgs. 10-15; text-fig. 19). This species also possesses triangular fixigenal areas adjacent to the basal part of the glabella, which increases its resemblance to Opipeuter, although, like most cyclopygids, it lacks an occipital ring. An occipital ring and well-defined glabellar furrows are, however, present on the distinctive cyclopygid Ellipsotaphnis (Whittard 1961, pi. 23, figs. 3, 4). TEXT-FIG. 2. a, outline reconstruction of the cranidium of Bohemilla sp. from the Llanvirn of the Great Paxton Borehole. Based on specimens in the Institute of Geological Sciences, London, especially By 8561, By 8529; b, hypothetical cranidium derived from Opipeuter by incorporation of two thoracic segments posteriorly (see text). Both approx, x 5. Hypothetical relationship to Bohemilla. The bizarre arthropod Bohemilla Barrande 1872 (type species B. stupenda Barrande 1872) was excluded from the Trilobita by Whittard (1952, p. 320) and this inter- pretation was followed in the Treatise. The genus is known from occurrences ranging from Llanvirn to Ashgill in age. I have been able to examine some slightly crushed but otherwise well-preserved material (text-fig. 2a) of a Bohemilla species of Llanvirn age from the Great Paxton borehole, Huntingdon, England. It is suggested that Bohemilla may be related to the Opipeuteridae, and hence, more distantly, to the Remopleurididae, and that its peculiar features may be explained without having to remove it to some non- trilobite group. From a study of the type and other species of Bohemilla Marek (1966n,p. 146) also concluded that the genus was related to the Remopleuridacea. The Bohemilla cranidium has the fixed cheeks reduced to backward tapering bands adjacent to the posterior part of the glabella. The glabella is parallel sided posteriorly, tapering forwards anteriorly. The two posterior glabellar furrows pass transversely completely across the glabella, the posterior of which may be regarded as the occipital furrow; the anterior two pairs are shorter, transverse. There is a narrow, horizontal anterior rim in front of the relatively narrow anterior Tongue’ of the glabella. That part of the cranidium lying in front of the anterior transverse glabellar furrow is similar to that of Opipeuter, notably in the tapering glabella, the position and length of the anterior glabellar furrow, and the narrow anterior cranidial rim. It is possible that Bohemilla could have been derived from an Opipeuter-Wke. ancestor by incorporation of two thoracic segments into the cephalon. The backward directed posterior border spines of Opipeuter already effectively include the first thoracic segment within the area of the cephalon. Ankylosis of the spine with the pleural remnant and loss of articulation would then completely incorporate the segment within the cephalon. Like the axis of the thorax, the glabella thus formed would be parallel sided. Such a method of cephalic segmental accretion would be unique among trilobites, as so far known, but may not be so improbable given an ancestor like Opipeuter. Thoracic segments associated with Bohemilla from the Great Paxton borehole have very narrow triangular pleurae similar to those of the anterior thoracic segment of Opipeuter (see also Marek 1966a, figs. 2, 4). I would prefer to regard the Bohemillidae as a family within the Remopleuridacea, rather than as a mono- generic superfamily as considered by Marek (1966a). Note that the band-like palpebral rims of Bohemilla are more like those of remopleuridids than those of Opipeuter. Text-fig. 2b is an attempt to model a Bohemilla- like cranidium from Opipeuter by incorporation of two thoracic segments. FORTEY: PELAGIC ORDOVICIAN TRILOBITE 121 Functional morphology. Discussion of the functional morphology of Opipeuter is based on the reconstruction of the whole dorsal exoskeleton (text-fig. 1). The free cheek is believed to hang below the level of the rest of the dorsal exoskeleton. Only with the cheek in this position do the sutures on the cranidium and the inner margin of the eye correspond exactly, the result being that the border of the cheek slopes steeply downwards below the surface of the eye. The convex visual surface then com- mands a very wide field of view. There are lenses directed forwards, backwards, laterally, upwards, and even downwards. Backward vision is made possible by the narrowness (trans.) of the thoracic pleurae, downward vision is possible because of the steep slope of the cheek, and especially anteriorly where the border narrows greatly. Probably the only area not visible to the trilobite was directly below the exoskeleton. If the size of the lenses is a measure of their visual efficiency the con- centration of large lenses anteriorly and dorsally may indicate that upward (assuming the trilobite swam with the exoskeleton uppermost) and anterior vision were of principal importance to the animal. This all round vision may be contrasted with that of Remopleurides with its narrow strip-like eyes, with the lenses in near-vertical files commanding predominantly lateral field of view. In anterior view (text-fig. 3u) the front margin of the cephalon is deeply arched. Similarly, anteriorly arched profiles are to be found among the Cyclopygidae and Carolinites species (text-figs. 3b, 3c), which, as has been discussed above, are similar ^ ^ a TEXT-FIG. 3. Outline reconstructions to show similarity of anterior cephalic prohles of Opipeuter, Pricyclopyge, and Carolinites. Visual surfaces black, a, Opipeuter inconnivus gen. et sp. nov. (about x4); b, Pricyclopyge hinodosa (Salter), after Marek (1961, text-hg. 8) ( X 2); c, Carolinites sihiricus Chugae\2i, based on new material of this species from the upper part of the Valhallfonna Formation, Ny Friesland, Spitsbergen ( x 4). in many ways to Opipeuter and are regarded as homeomorphic. The vaulted anterior profile, together with the cheeks which project well below the thorax with the ‘cutting’ edge of the border facing downwards indicate that the trilobite was not adapted to rest on the sediment surface. This is in contrast to the profile of, for example, asaphids such as Ogygiocaris which are dorsoventrally flattened, and in which a ventral plat- form formed by the broad cephalic doublure lies on a level with the tips of the thoracic pleurae and the outer-part of the doublure of the pygidium; the anterior arch of the cephalon is reduced or absent. Such a morphology distributes the weight of the animal over a wide surface area, and the broad ventral doublure is suitable for rest- ing on soft mud. The all-round vision of the eyes, taken together with the distinctive profile of the cephalic region, suggests that Opipeuter was a free-swimming species. In particular. 122 PALAEONTOLOGY, VOLUME 17 the downward vision of the eyes would scarcely be pertinent to a benthonic animal. The other features of the exoskeleton are consistent with the same interpretation : (i) The exoskeleton is, for a trilobite, extraordinarily long and narrow, and laterally compressed. Text-fig. 4 shows the dorsal exoskeletal length (excluding caudal spines) compared with transverse thoracic width at the mid point of the thorax of several Ordovician species in the British Museum (Natural History). Uncrushed specimens from dilferent families within the major groups shown were chosen to give a wide taxonomic spectrum. Most show a length of about one and a half to two times the thoracic width; Trinucleina tend to be more ovate. No specimen was found which approached the relative length of Opipeiiter, three times as long as wide. Such an elongate form is almost invariable among recent actively swimming Crustacea. (ii) A very convex thoracic axis, as Richter (1919, p. 229) observed, probably allows for greater room for muscles necessary for vigorous swimming. The exoskeleton is relatively thick, which may be necessary for the support of powerful musculature. The very broad (sag.) articulating half-rings no doubt lent the thorax a great deal of flexibility in the vertical plane. TEXT-FIG. 4. Graph of sagittal length (L) (excluding caudal spines) plotted against thoracic width (W) at the mid point of the thorax for a number of perfectly preserved complete trilobites (Ordovician) in the collections of the British Museum (Natural History), o— Odonto- pleurids, p— Proetacea, a — Asaphina, /—Trinucleina, c — Cheirurina, d — Dalmanitacea, r— Remopleuracea, cu— Calymenacea, i—Illaenus, Op^Opipeuter. (iii) The transverse abbreviation of the thoracic pleurae enables a reduction in the total weight of the exoskeleton, and the pleural remnants may also have acted as stabilizing ‘fins’ along the edge of the thorax. Opipeuter exhibits the opposite tendency to the Odontopleuridae, in which the wide thoracic pleurae are greatly extended laterally by their continuation into one or more long and often slender spines such that the transverse thoracic width of the odontopleurid may exceed the sagittal length FORTEY: PELAGIC ORDOVICIAN TRILOBITE 123 (text-fig. 4). Both Richter (1919) and Raymond (1920) considered that these spines may have functioned as frictional brakes to prevent sinking and that the odonto- pleurids were planktonic. Functional studies (Seilacher 1959; Clarkson 1969) have indicated that the odontopleurids were more probably benthonic. Even if the plank- tonic view is correct it is clear that the odontopleurid mode of life must have been very different from that of Opipeuter, for spines which inhibited sinking would also have opposed active swimming. (iv) The functional significance of the gaps between thoracic pleurae is not obvious, although three suggestions may be made. The gaps may have been necessary to permit enrolment of the long, narrow thorax, or they may have served simply to further reduce the bulk of the exoskeleton, or they may have permitted a small amount of lateral ‘wriggling’ movement in the thorax. Possibly all three factors may have acted in combination. The depth at which the actively swimming Opipeuter may have lived is subject only to indirect evidence. Large eyes alone do not necessarily indicate life in the surface waters, for, as Clarkson (1967, p. 371) points out, recent arthropod species with enlarged eyes may be found at considerable depths in the sea. It is of interest to note that among recent marine arthropods the development of exceptionally large eyes is a frequent concomitant to the adoption of a pelagic (whether epi- or bathypelagic) or planktonic mode of life. Among the Amphipoda, for example, the single group Hyperiidea is almost entirely planktonic, and compared with other amphipods the eyes are very large and convex, occupying almost all the available genal region. The same relative expansion of visual area applies to the Isopoda, in such genera as Eurydice and Aega which spend at least part of their lives actively swimming and may constitute part of the plankton. These latter genera also have an elongate body form and transversely narrow, small, downward-sloping epimera, a gross form not unlike that of Opipeuter. Both pelagic isopods and the amphipods are of similar size to Opipeuter. Indirect evidence on the life habits of Opipeuter comes also from the lithology of the limestones in which O. inconnivus is found, and from the fauna associated with it. In Spitsbergen the species occurs with olenid trilobites and graptolites in dark, laminated bituminous limestones in the lower part of the Valhallfonna Formation, but also with a completely different assemblage of trilobites dominated by asaphids, nileids, and raphiophorids, often in greyish, crystalline limestones, for the upper part of its range. In western Ireland the species occurs in pure, white or pink sparites with a third assemblage of trilobites dominated by lUaenus and Cheiruridae, and associated with abundant articulate brachiopods, algal structures, and bryozoans suggestive of shallow water (possibly even ‘reef’) conditions. This independence of associated fauna and lithology is what might be expected of a pelagic animal, and the occurrence of Opipeuter with a shallow water assemblage further indicates that the species was capable of living in near-surface waters. This does not preclude the possi- bility that it was equally adapted to life in deeper waters. Acknowledgements. I am indebted to Dr. A. W. A. Rushton of the Institute of Geological Sciences for allowing me to examine undescribed material of Bohemilla. I also thank Dr. J. K. Ingham (Glasgow University) and Dr. R. J. Lincoln of the British Museum (Natural History) for their stimulating discussion. 124 PALAEONTOLOGY, VOLUME 17 REFERENCES BARRANDE, J. 1872. Systeme Silurien du Centre de la Boheme : Partie, Supplement au Vol. 1. Trilobites, Crustaces divers et Poissons, 647 pp., 35 pis. Prague and Paris. CLARKSON, E. N. K. 1967. Environmental significance of eye-reduction in trilobites and recent arthropods. Marine Geol. 5, 367-375. 1969. A functional study of the Silurian odontopleurid trilobite Leonaspis deflexa (Lake). Lethaia, 2, 329-344, 7 figs. DEWEY, J. E., RICKARDS, R. B. and SKEViNGTON, D. 1970. New light on the age of Dalradian deformation and metamorphism in western Ireland. Norsk geol. Tiddskr. 50, 19-44. FISCHER, A. G. 1946. A Carapace of the Ordovician trilobite Telephus. J. Paleont. 20, 566-569, figs. 1-3. FORTEY, R. A. and BRUTON, D. L. 1973. Cambrian-Ordovician rocks adjacent to Hinlopenstretet, North Ny Friesland, Spitsbergen. Bull. geol. Soc. Amer. 84, 2227-2242. GARDINER, c. I. and REYNOLDS, s. H. 1909. On the igneous and associated sedimentary rocks of the Tour- makeady District (County Mayo). Quart. Jour. geol. Soc. London, 65, 104-153. HAWLE, T. and CORDA, A. J. c. 1847. Prodrom einer Monographie der bohnnschen Trilobiten. 176 pp., 7 pis. Prague. HiNTZE, L. F. 1953. Lower Ordovician trilobites from western Utah and eastern Nevada. Bull. Utah geol. miner. Surv. 48, 249 pp., 28 pis. MAREK, L. 1961. The trilobite family Cyclopygidae Raymond in the Ordovician of Bohemia. Rozpr. Ustr. Ust. Geol. 28, 84 pp., 6 pis. 1966fl. Nadceld Bohemillacea Barrande, 1872 (Trilobita) v ceskem Ordoviku. Cas.narod. Mus. 135, 145-153, 2 pis. \966b. Rod Cremastoglottos Whittard, 1961 (Trilobita) v ceskem Caradoku. Ibid. 193-194, 1 pi. NiKOLAiSEN, F. 1963. The middle Ordovician of the Oslo region, Norway. 14, The trilobite family Tele- phinidae. Norsk geol. Tiddskr. 43, 345-400, 4 pis. RAYMOND, p. E. 1920. The appendages, anatomy and relationships of trilobites. Mem. Conn. Acad. Arts Sci. 7, 169 pp. RICHTER, R. 1919. Von Bau und Leben der Trilobiten. 1. Das Schwimmen. Senckenbergiana, 2, 213-238. SEiLACHER, A. 1959. Von leben der Trilobiten. Naturwissenschaften, 46, 389-393. SHAW, F. c. 1968. Early Middle Ordovician Chazy Trilobites of New York. Mem. N. Y. St. Mus. nat. Hist. 17, 1 14 pp., 24 pis. SKEVINGTON, D. 1971. The age and correlation of the Rosroe Grits, north-west Co. Galway. Proc. R. Ir. Acad. (B), 71 (5), 75-83. ULRICH, E. o. 1930. Ordovician trilobites of the family Telephidae and concerned stratigraphic correlations. Proc. U.S. Nat. Mus. 76, 101 pp., 8 pis. WHITTARD, w. F. 1940. The Ordovician Trilobite fauna of the Shelve-Corndon District, West Shropshire. II. Cyclopygidae, Dionididae, Illaenidae, Nileidae. Ann. Mag. nat. Hist. (11), 6, 129-153, 4 pis. 1952. Cyclopygid trilobites from Girvan and a note on Bohemilla. Bull. Br. Mus. nat. Hist. Geol. 1, 305-324, 2 pis. 1961. The Ordovician Trilobites of the Shelve Inlier, West Shropshire. Part V. Palaeontogr. Soc. (Mongr.), 163-196, pis. 22-25. WHITTINGTON, H. B. 1959. Silicified Middle Ordovician Trilobites: Remopleurididae, Trinucleidae, Raphio- phoridae, Endymioniidae. Bull. Mus. comp. Zool. Harv. 121, 371-496, 36 pis. 1965. Trilobites of the Ordovician Table Head Formation, Western Newfoundland. Ibid. 132, 275-442, 68 pis. R. A. FORTEY Department of Palaeontology British Museum (Natural History) Cromwell Road Revised typescript received 16 February 1973 London, SW7 5BD Note added in press. Dr. T. Tjernvik has recently shown me a pygidium of an Opipeuter species from the early Arenig (dalecarlicus zone) of Sweden. The pygidium was found in a boulder in glacial clay near Orebro, Narke, derived from a source outcrop in the region of the South Bothnian Bay. WILLIAMSONIELLA LIGNIERI: ITS POLLEN AND THE COMPRESSION OF SPHERICAL POLLEN GRAINS by TOM M. HARRIS Abstract. The type specimen of the Bennettitalean flower Williamsoniella lignieri (Nathorst) on reinvestigation proved to have an intact pollen sac yielding well-preserved pollen. The pollen grains resemble Exesipollenites scabmtus . Most of the grains are distorted by crushing and the distortion of outline is related to folds on the surface. Hollow balls made of various materials were compressed between flat surfaces and their various secondary distortions are described; one kind of ball mimicked the forms seen in W. lignieri pollen and other balls with different mechanical properties mimicked the distortions of miospores of other plants. When compressed in a matrix (whether itself com- pressible or not) the hollow balls distorted differently. There is scope for further experimental study of compression in a matrix. The specimen described here was collected by A. G. Nathorst in 1909 from the classic Lower Estuarine Plant Bed (Lower Bajocian) of Whitby, Yorkshire. Nathorst described it in 1909 as a problematic male flower but the specimen, which is preserved in the Section for Palaeobotany of the Riksmuseum, Stockholm, has remained obscure. I re-examined it at Stockholm in 1972 and realized that it is probably the same species as Williamsoniella papillosa, also from Whitby and described by Cridland in 1957. WILLIAMSONIELLA LIGNIERI (NATHORST, 1909) 1909 WilIiamsonia(l) Lignieri Nathorst, p. 20, pi. 4. 1915 Williamsoniella Lignieri (Nathorst), Thomas, p. 154. (Name, discussion.) 1933 Williamsoniella Lignieri (Nathorst), Florin, p. 11, text-figs. 3b, 5b (stomata). 1957 Williamsoniella papillosa Cridland, p. 383, text-figs. 1-3 (various organs). Mention of W. lignieri on p. 388. 1967 Williamsonia lignieri Nath., Potonie, p. 122, pi. 13, fig. 258 (pollen). 1969 Williamsoniella lignieri (Nathorst), Harris, pp. 142, 144. (Cited as possibly the same as W. coronata.) 1971 Williamsoniella papillosa Cridland, Konijnenburg van Cittert, p. 35, pi. 7, fig. 5 (pollen). General discussion. The specimen is the base of a flower compressed to form a disc of coal about 0-5 mm thick. The coaly substance consists of several layers of plant material but the matrix between them is so thin that they cannot be separated with needles though the cuticles separate on maceration. Some of the coal is missing, no doubt having been used by Nathorst and probably by Thomas for cuticle preparations, so part and counterpart do not correspond fully. The original drawings of the part and counterpart are reproduced here (PI. 1 5, figs. 2-3) together with new photographs (PI. 15, figs. 5-6). The part is a view from below morphologically (text-fig. 1b) and most of it shows lanceolate scales but on one side (PI. 15, figs. 2, 3, 6) it shows fine radiating ridges. These radiating ridges are the stalks of interseminal scales in an approximate transverse section of the gynaeceum near its base. Nathorst called them [Palaeontology, Vol. 17, Part 1, 1974, pp. 125-148, pi. 15.] 126 PALAEONTOLOGY, VOLUME 17 the disc or collar, terms he used for the corresponding part of Williamsonia flowers. All the upper interseminal scales and seeds were lost before preservation. The specimen had not been prepared by picking away rock matrix and its margins were still buried. When cleared, a few crumbs of coal came away and were macerated with useful results. The separate bract shown on the left of PI. 15, figs. 3, 6 (which Nathorst recognized as belonging to W. lignieri) also came away and also a small piece of coal just at the edge of the radiating lanceolate scales (seen at 2 o’clock in PI. 15, figs. 3, 6, and marked with an arrow). This proved to be a pollen sac. It had been drawn by Nathorst’s artist but had not been described. The original drawings are excellent but do not look quite like the photographs. They show individual bracts more clearly; no doubt varied lighting was used to show their margins. They also leave things out, for instance the imprint of the free bract is omitted from PI. 15, fig. 2 but seen in fig. 5. The rock is trimmed down artistically in both drawings but in fact both sides are fair-sized blocks which show various associated leaves. Nathorst (1909) and Thomas (1915) differed in their interpretations of the parts of the flower and sometimes I follow the one, sometimes the other. For Nathorst (1909) the radiating scales were ordinary floral bracts, the organs which when separate are called Cycadolepis, and this is here accepted fully. But he does not relate his cuticle preparations (his pi. 4, figs. 9, 10) to the two surfaces of one of these bracts. He found some unusual looking interseminal scale heads with adherent pollen, and mis- interpreted them as microsporangia. Thomas on the other hand recognized the inter- seminal scales but interpreted the floral bracts as microsporophylls. Both recognized the floral axis in the middle. This forms a rod of coal passing obliquely into the rock. No doubt compression of the matrix has greatly exaggerated the obliquity of this axis (text-fig. 1). It was the perianth bracts that first indicated that W. lignieri and W. papillosa were the same. They have marked specific peculiarities. They are unusually small, only about 1 1 mm long, a dull black, not shining like many kinds of Cycadolepis. As usual they are covered with wrinkles, these being longitudinal in the middle region but divergent near the sides. The wrinkles are exceptionally fine, only 0-5 mm from crest to crest in the middle region, but further apart near the edges. I believe that these wrinkles were produced in preservation following the decay of the inner tissues (and I have simulated them in the scales of the artichoke Cynara) but none the less they are EXPLANATION OF PLATE 15 Figs. 1-3, 5-9. Williamsoniella lignieri Nathorst. 1, Pollen grain showing the pore with a clear margin, X 1500. 2-3, Copies of original figures of W. lignieri from Nathorst 1909, pi. 4, fig. 2, 1. Fig. 3 shows the separate bract and at 2 o’clock the pollen sac now marked with an arrow. 5-6, Photos of Nathorst’s Holotype. Fig. 6 (as fig. 3) shows the pollen sac, marked with an arrow. These are bits of two associated Nilssoniopteris major leaves below the flower. Photos by Mr. H. Samuelsson, x2. 7, 8, A pollen grain at two levels of focus. Fig. 7 shows the pore as a vague pale area; there is corrosion or extraneous matter on the left. In fig. 8 the margin of the pore is partly in focus, x 1500. 9, A pollen grain with a pale area but no definite pore margin is visible, x 1500. Figs. 4, 10 14. Experimental compression of table tennis ball. 4-10, Upper and under sides of a table tennis ball compressed in dry cotton wool matrix after weakening the equator with six small cuts (p. 147). 11-12. Upper and under sides of table tennis ball compressed in dry cotton wool (p. 146), illuminated from top right. 13, 14. Upper and under sides of table tennis ball compressed in sawdust (p. 145). PLATE 15 HARRIS, Williamsoniella 128 PALAEONTOLOGY, VOLUME 17 TEXT-HG. 1. Imaginary longitudinal sections of holotype specimen. A, flower base as buried in mud but uncompressed. On the left a micro- sporophyll has one pollen sac but the rest is missing ; the microsporophyll on the right has lost its pollen sacs. The dotted line is the future plane of rock splitting. The specimen is slightly tilted, the axis being not quite vertical, b, c, part and counterpart after compression which has made the axis seem more oblique but lateral organs almost horizontal. Magnification about 2-5. characteristic of the species. Minute hairs, 0-5 mm long occur thickly on the edges of the free bract but not on its surface. They are not to be seen satisfactorily at the edges of bracts of the flower but cuticle preparations show that they were present there also. Although the bracts have thick cuticles, originally 4-8 ixm (estimated from folds), they are broken by cracks and the preparations are all small bits, but they do include margins where the two surfaces, stomatal and non-stomatal, join. The stomatal side is assumed for description to have faced downwards (or outwards) as it certainly does in some other Bennettitalean flowers. The upper (adaxial) side is of constant structure but the lower has very varied numbers of stomata ; perhaps (on analogy with artichoke scales) those with few are inner scales with little photosynthetic tissue. The two sur- faces are figured here (text-fig. 2) at a higher magnification than in Nathorst’s photo- graphs. Two peculiar features are emphasized; most of the stomata have apertures parallel with the long axis of the scale and the subsidiary cells are large and look pale. Bracts with numerous stomata show about 100 per sq. mm. which is exceptional for a Cycadolepis, but others have very few. The free bract has but few and these are near the apex. Hair bases are rare on the surface but an occasional ring-shaped print on a cell may represent a hair base. On the margin the hairs may be so crowded as to be in contact. They have very delicate cuticles and as far as is known are unicellular. Nathorst’s figures of adherent pollen show round or elongated grains about 25 /xm wide, mostly with one or two parallel folds but one with crossed folds and one round one unfolded. Potonie (1967) examined Nathorst’s slides (which had deterio- rated through desiccation) and figured a rounded grain about 30 ;u.m wide with a single mark regarded as a colpus. He remarks that most grains are spindle shaped. We only know the microsporophyll of Nathorst’s specimen from the pollen sac just beyond the bracts. This is at about 12 mm from the flower centre but from Cridland we know that the microsporophyll was over 20 mm long, so this at 12 mm should be one of the lower pair. Everything beyond it is missing. There may well be a complete HARRIS: WILLIAMSONIELLA 129 ring of microsporophyll bases concealed between the robust bracts and robust inter- seminal scales, but the microsporphylls are delicate and with delicate cuticles which would escape recognition in cuticle preparation unless deliberately searched for and even then might be missed. This pollen sac or synangium was recognized as one only after maceration. It gave a compound block of tissue nearly 2 mm x 0-4 mm and this on further maceration gave some unit blocks of pollen about 0-4 mm x 0-1 mm which are taken to be the contents of single sporangia and the best of these on still further maceration broke up to give individual pollen grains. The macerations were carried out very gently with diluted acid in the hope of saving possible delicate tissues and even after the third treatment the pollen mass was only partly disintegrated. Beside the pollen there were delicate and uninformative cuticles of about three kinds. The pollen grains agree in size and some other features with the few seen in Nathorst’s slides. TEXT-HG. 2. A, distribution orientation of stomata on bract cuticles. The base is 1 mm. Circles are possible hair bases, b, c, two interseminal scale heads in oblique compression, features at a lower level shown by broken lines. At the right of each is an imaginary section from top to bottom, both x 50. d, apex of a bract showing a few marginal hairs, x 12-5. e, adaxial cuticle of bract, x200. f, stoma of bract, it is unusual in being transverse but otherwise normal. Surface thickenings stippled, x 400. G, abaxial cuticle of bract, stomatal apertures longitudinal. There is a ring on a divided cell which may be a hair base, x 200. All preparations from type specimen in Stockholm. 130 PALAEONTOLOGY, VOLUME 17 The coal from the edge of the specimen gave some interseminal scale head cuticles. They are compressed obliquely, as though they were situated half-way from the lower pole to the equator of the gynaeceum, while Nathorst’s preparations, which were evidently from the base of the gynaeceum, are so placed that the head is pressed out flat. The new ones are useful because they are more like the scale heads prepared by Cridland from near the top of his ‘ W. papillosa gynaeceum than are those of Nathorst. In particular they are small, have a strongly raised central boss where the cuticle is thick, but it quickly becomes delicate towards the margins. Most of Nathorst’s scale heads were almost flat and their cuticles were evenly thick to their edges, but there the cuticle bends inwards and becomes delicate abruptly and adheres to the edge of the next scale. No ovules are seen in any preparations of the holotype and they are not to be expected at the base of a gynaeceum. Identification q/ Williamsoniella papillosa with W. lignieri. The specimens, although from the same locality, look different. W. lignieri is a nearly intact flower base com- pressed vertically but Cridland’s specimens are fully disintegrated. The floral axis, the bracts, and the microsporophylls were separate but found on the same blocks. Just a few interseminal scales and ovules (perhaps abortive ones) remained at the top of the floral receptacle and most of the pollen sacs were empty but one had a few pollen grains sticking to a lining membrane. Cridland (1957) mentioned the bracts but did not describe them. The bracts on Cridland’s blocks are small, 10 mm- 15 mm long, dull black, have the same fine wrinkles and minute marginal hairs and above all have similar well-characterized cuticles. As in W. lignieri some have many stomata but some very few. As mentioned the interseminal scale heads now prepared are similar to some of those in Cridland’s preparations (though less like the one he selected to figure). At first the pollen seemed to differ but it is now found to agree. The grains Cridland selected and figured at x 900 (under my supervision) were broadly oval and each shows a strongly marked groove. The wall was thin and described as smooth. They were 27 |um long which is within the range of W. lignieri. Then Konijnenburg van Cittert (1971) examined Cridland’s preparation and gave photographs of some pollen grains at x 1500 and a detailed description of the wall. She recognized a fine sculpture of a network of ridges enclosing pits about 1 p.m wide. She was doubtful whether the groove in Cridland’s grains was a true colpus. Neither author mentions a pore but it may be this was seen but dismissed as an effect of corrosion. The preparation shows only about thirty grains sticking to the pollen sac cuticles and half of these are obscure. With so few pollen grains and these only moderately clear a good deal of selection is necessary and it is hard to decide which features are original and which are effects of corrosion or distortion. On re-examination of Cridland’s slide in the British Museum, a pore-like pale area was noted in sixteen of the grains (such an area is seen also in Konijnenburg van Cittert’s pi. 7, fig. 5 top left and middle left grains but it is not indisputably a pore). The different form of Nathorst’s specimen and those of Cridland call for comment and a possible explanation. Cridland’s seem in the natural state for a fructification that has fulfilled its function; it is Nathorst’s partly intact one, still with unliberated pollen that is surprising. I suggest that it may be a slightly immature flower which HARRIS: W I LLI A M SO N I ELLA 131 had been attacked by an animal that had devoured the main part of the gynaecium and all but one of the pollen sacs but had left the tough bracts and gynaecium base. Attribution to a leaf species. A main aim of my work has been to relate fossil repro- ductive organs to the far more abundant leaves and for this association in the field gives valuable indications, at least as probabilities. Cridland’s material of Willkmi- soniella papillosa from Whitby is strikingly associated with the leaf Nilssoniopteris mfl/or (which is nowhere as abundant as at Whitby). We have long suspected or believed that Williamsoniella coronata belonged to Nilssoniopteris vittata so it was natural to look for a similar leaf for W. papillosa. But the association of W. lignieri with N. major was just as close ; the block with the holotype has five fragments of that leaf but nothing else to which the flower is likely to belong. So the two flowers seemed for a short time to be rival claimants for a single kind of leaf but with their synonymy rivalry disappears. THE POLLEN OP WILLIAMSONIELLA LIGNIERI Description. Pollen grains almost circular and mostly about 25 pm in diameter (originally nearly spherical). Wall without any ridges or furrow (colpus) but with a single round or oval pore, edge of pore neither raised nor thickened, size typically 6 pm X 8-10 pm but sometimes larger, up to 15 ;um x 12 pm in a large grain, edge of pore sometimes distinct but often ill marked. Pollen grain wall thin, up to 1 ;u,m thick but often less, when well developed showing a fine network of ridges enclosing pits. Konijnenburg van Cittert (1971) gives in addition (for Williamsoniella papillosa) wall (exine) 0-5- 10 p thick; nexine smooth, very thin; sexine with columellae — and capita— layer; columellae very short, indistinct, capita spherical, laterally fused, forming a reticulum; lumina of reticulum wider than muri; muri about 0-5 p wide, lumina 1 p". I do not doubt these additional details but while I sometimes thought I perceived them, 1 could not do so consistently. But the reticulum on one of the thicker walled grains is illustrated in text-fig. 3c. Discussion. Fortunately the grains all appear to belong to the one species without a single contaminant spore. They are as usual fully flattened, the upper and lower walls pressed together and many are obviously distorted and show compression folds of various kinds. They are neither pitted with pyrite crystals nor corroded in the manner that is attributed to bacteria. Nor have they indentations attributable to mineral crystals in the matrix. One grain showing possible corrosion is shown in PI. 15, fig. 7, but the effect may be due to extraneous matter. But although there are many clearly visible grains, the population as a whole remained rather under- macerated after its three mild treatments and a great many grains are still in compact masses and only dimly visible. There are a very few grains, less than 1%, with irregu- larities in their outline that I cannot attribute to vertical compression but dismiss as having first been compressed in other planes. The pore may be seen anywhere in the middle, half-way out, or at the edge and equally in round or elongated grains; also the pore has no obvious relation to folds. For instance it might be in the groove between two folds or on the back of the grain opposite the groove. These observa- tions seemed consistent with the idea that the original shape was nearly spherical, without any predetermined groove or folds. This conclusion was arrived at only after the pollen grains had been measured 132 PALAEONTOLOGY, VOLUME 17 a considerable number of times and recorded in increasing detail. At first merely the mean length and breadth, the total range and standard deviation were recorded, but then it seemed worth dividing them into three groups, the roundest (two axes equal within 3 i^m), the longest (two axes differing by 10 ;u.m or more), and the middle group, and giving means, ranges, and standard deviations. At the same time the kinds of folds were recorded and a relation became obvious. This led to still another survey when the grains were divided as much as possible on their shape into classes with axes differing by 0, 1, 2 up to 17 /xm. At the same time the folds were recorded in six classes — (a) no visible folds; (b) a minute concentric one at the very edge (making a double edge); (c) strong folds of nearly concentric course forming a wreath half-way between the centre and the edge; (d) crossed folds, one on the top wall, one beneath; (e) a longitudinal groove flanked by two folds, and finally (/) a single strong longitudinal fold. Certain kinds of folds seen in spores of various other plants (for instance many small wrinkles) were never met. Two grains out of two hundred were rejected as shrivelled but many still in compact masses were too obscure to measure. This analysis was successful in indicating that all folds are secondary but it is not certain that the original shape was a perfect sphere; it might have had one axis perhaps 2 ixm longer than another. It is also possible that compression might have produced a slight general change in diameter. A main fact that emerged is that the rounder a grain the smaller is its longest diameter. The perfectly round grains were not only the smallest (mean 24-5 /xm) but also the least varied in size and range recorded 22-29 fxm, standard deviation 1-5 |um. As one takes progressively more elongated grains the mean maximum diameter increases, the mean minimum decreases and the variability of both, as well as the recorded range, increases. For the group of grains where the axes differ by 3 jxm the long axis is mean 25 ixm, range 21-31 /xm. For the group where the axes differ by 10 /xm the mean of the long axis is 33 /xm, its range 26-38 /xm and the standard deviation 5 /xm. These figures may not be very significant but they show a trend. The short axis of this last group had a mean of 19 /xm, range 13-24 ^um, and standard deviation of 3 |U.m. The numbers of grains of each diameter in a count are given in a histogram in fig. 4 but the six classes of folds are reduced to two. Folds of the kinds a, h, c, d which are thought not to alter the out- line of the grain much are taken together and the single longitudinal groove with one or two folds, e, f are taken together. The pollen grains in Cridland’s preparation of WUliamsoniella papUlosa show a similar range of distortions of outline and sur- face folds. The pore of W. lignieri is the most unexpected feature of its pollen grain for nothing like it had been recorded for the Bennettitales. It was, however, likely it would be seen in some fructification since the dispersed pollen resembling W. lignieri pollen, Exesipollenites Balme, or Splieripollenites Couper is widespread and often abundant in Yorkshire plant beds. Even in grains with the clearest pores it is difficult to decide whether the pore is open or had a very thin membrane. The pollen of WUliamsoniella coronata, described in detail by Konijnenburg van Cittert ( 1971 ) agrees in its size and unusually broad shape and its thin wall, described as thinner than in most grains of W. lignieri. She could detect no sculpture on the wall, but neither can I on the thinner walled W. lignieri grains. She described it as colpate. HARRIS: W I LLI A M SO N I E LLA 133 as previous authors had all done, but again over half the grains of W. papillosa have folds which could be taken for a colpus. No one has reported a pore in its wall, but I note on one of her figures (pi. 7, fig. 4) a pale spot which could be a pore. Plainly this pollen needs fresh study. There is a real difference in shape between the pollen grains of both these Williamsoniella species and those of all other Bennettitales I have seen under the microscope or as photographs. All the others are elongated and mostly a good deal larger. Some grains of W. papillosa may indeed be just as narrow but the whole population shows that these are distorted. No doubt too the colpus of the normal Bennettitalean grain is real but the colpus-like folds of distorted William- soniella grains raise slight doubts. There is a case for the experimental compression of plastic ellipsoidal balls. No other gymnosperm with similar pollen has been described. The dispersed grains Exesipollenites scabratns which look similar have been tentatively referred to the conifers mainly because some Taxodiaceae have a single ‘pore’, but it is different, being the end of a conical papilla. It therefore is seen in surface view as a pore with a thickened rim. E. scabratus which has been figured by Couper 1958 from Yorkshire and by Tralau 1968 from the Middle Jurassic of Scania is similar in size and shape but the wall is thicker and more distinctly marked. It may be that the better characterized grains have been selected for figures but it is also possible that the name covers a number of species differing slightly in these respects. THE COMPRESSION OE VARIOUS SPHERICAE MODELS IN RELATION TO WILLIAMSONIELLA LIGNIERI AND OTHER EOSSIL POLLEN This section deals with distortions arising when spores are compressed between surfaces— the pollen sac walls for W. lignieri and boards or glass plates for the models. Although there are references to similar features to be seen among spores dispersed in rock matrix, the conditions of compression differ considerably and, while I believe many of the distortions apply to dispersed spores, others do not or only do exception- ally. My conclusions must not be applied indiscriminately to dispersed spores. I had previously thought very little about the compression of spores except to regard the secondary distortions as a nuisance, to be recognized as distortions and so to be avoided as pitfalls and to be eliminated from descriptions. I had never welcomed them as useful evidence as I try to do here. Nor can I find any papers that do anything near this except for Potonie 1962 who analyses and classifies the folds on fossil spores of many original forms. He convinced himself that the folds are by no means casual but were related to the original form in an orderly way. They were therefore to be con- sidered as features worthy of record. But I can find no record of any experiment to produce folds in models. Here I deal only with spheres, the simplest form, and as it happens the form of W. lignieri. In their survey of the form of Cicatricosisporites Hughes and Moody-Stuart (1969, pp. 103-106) regarded many of the unusual forms as distortions caused in preserva- tion. The physical distortions of compression and inflation and the chemical ones of various kinds of corrosion cause striking changes in the wall sculpture. The major distortions of these miospores are, however, different from those of W. lignieri for while bursting is common in their spores, gross folding of the wall is not. No burst grains were seen in the sample of W. lignieri but bursting is often seen in Bennettitalean 134 PALAEONTOLOGY, VOLUME 17 pollen (Nathorst 1911). In Recent plants, bursting often happens when healthy pollen falls into fresh water. Clayton (1972) describes a remarkable deformation of the miospores of Dictyotriletes admirabilis, which when compressed in a sporangial mass take strong imprints from other spores lying immediately above them but in other respects their distortions are slight. While this distortion has no similarity to those dealt with here, it is relevant in that it reveals a mechanical property of the wall. The simple, qualitative experiments described below are only justifiable in so far as they give ideas which seem useful when applied to real fossils. They can suggest good ideas and they can discredit bad ones. For me they did both. In W. liguieri the pollen is not in a matrix of mud but in the sporangium wall, itself part of a pollen sac. I imagine that before it was compressed the whole fossil had been buried in mud and the more labile materials had decayed. The pollen grains would be mere bags of water floating in the water-filled sporangium and the water both in the grain and in the sporangium, as well as in the surrounding mud, would drain away as compression occurred. And when compression did begin it would be between the firm walls of the sporangium and therefore somewhat similar to my experiments with hollow balls compressed in air between hard surfaces. For my experiments I used various hollow spheres made of different materials. The most useful was a particular plastic ball sold under the name ‘playball’ and I shall refer to it by that name. This ‘playbalf when punctured to let out air had a skin which when compressed mimicked astonishingly closely the various forms of compressed W. lignieri pollen. Other balls were constructed out of materials selected for their peculiar properties in the hope that they might throw light on the mechanics of the deformation of a sphere. The more successful experiments of this series are mentioned below. A good many of the effects produced, though not like ones seen in W. lignieri, did mimic distortions of other dispersed miospores. The specially constructed balls were made as follows. First a core of a water-soluble substance was shaped into a sphere. Then a waterproof skin of the selected materials was built on to the core and allowed to harden. Then the coated sphere had small holes cut in the skin at the top and bottom poles (these are the regions where minimal distortion is to be expected) and the whole thing was put into water till the core dis- solved. The cores were made about 10 cm wide and the skins about 2 mm thick, that is 2% of the diameter. This is similar to the purchased ‘playbalf and also to many mio- spores, including W. lignieri pollen. Magnesium sulphate, commercial Epsom salts, was found to make a good core. This salt when heated with a few drops of water seems to melt in its water of crystallization and soon boils as a liquid slush. This can be poured into suitable moulds and forms blocks on cooling. The blocks have the consistency of compact frozen snow and they can be sawn into cubes and then shaped with a coarse rasp into spheres. For the skins I used materials whose properties were familiar to me. Before dealing with the results of the experiments 1 will discuss briefly the basic principles of the compression of a sphere. When a hollow sphere is flattened by loading from above, the skin at the top is ultimately pressed against that at the bottom to give a disc of two layers, continuous at the edges. This is the primary dis- tortion of compression. But because of the geometric nature of a sphere certain secondary distortions must occur; they are unavoidable but their effects may not be obvious. If we consider such a sphere of radius r its circumference is Itti- and its HARRIS: W I LLI AM SO N I ELLA 135 surface area is It may be simpler just to consider the compression of the top hemisphere into a disc of single thickness and for this we have l-nr and If we now compress this hemisphere into a circle of radius r then awkward consequences follow; the distance over the top of the hemisphere -nr is reduced to 2r. This factor, 77/2 occurs again and again in this discussion; it is nearly 1-57 to 1. At the same time the area 2-nr^ is reduced to plainly a lot of skin must have gone somewhere and it can be taken in various different ways, each being a secondary distortion. Another possibility is for the hemisphere to spread outwards, keeping its distance -nr from side to side over the top. Its diameter is increased from 2r to ur, that is by 57% of the original diameter. When such a sphere holds its form, it is because its hollow skin has various mechani- cal strengths in adequate degree, and when it collapses under a load this happens because one or another of these mechanical strengths has failed, often several at once. It is easier to understand the effects when one strength in particular fails, and for the constructed models materials were chosen such that they should be very adequate in some respects but weak and ready to fail in one particular strength neces- sary for the maintenance of the spherical form. Provided this failure predominated in an obvious way it seemed good enough to show that the principle applied; it did not matter if other secondary distortions occurred to a slight extent, as they do indeed in some of the experiments illustrated in text-fig. 5. The secondary distortions are of five main groups, each governed by a particular mechanical failure; if others can occur I have not thought of them. I arrange them in order with the most obvious visible effects first. Failure 1. The skin is rigid and brittle and so cracks and shatters rather than flattens into a continuous disc. This failure is included only for completeness; it may never occur in fossil spores. Such shattering is familiar in the much less flexible shells of lamellibranchs preserved in soft mud. Failure 2. The sphere spreads to a disc nr wide, an increase by 57%. At the same time the circumference increases by 57% and if the skin will not extend smoothly it snaps and cracks run in toward the centre. The cracks will narrow inwards because there is less stretching in the inner part. But most materials will extend somewhat and so the cracks are neither as numerous nor as deep as they might be. It is to be noted that if the original sphere were a spore with a regularly and uniformly pitted wall, the pits would retain their size and shape over the spread-out wall. A half grapefruit skin when compressed illustrates this deformation in a hemi- sphere. For a model, a skin of commercial putty was used. This consists of powdered chalk and a viscous liquid oil. At first the putty is so soft as to be nearly liquid and it can be pulled out without snapping. But in air, the oil oxidizes, ‘dries’ to a resin and the putty becomes increasingly solid. At first it will still extend somewhat but a few days later it becomes fully solid and will not extend in a plastic way at all. Thus as the properties change putty can be used for different models illustrating different failures. The putty for this hollow model was allowed to harden for four days. It will be noticed (text-fig. 5a) that while it has split convincingly (it was allowed to flatten under its weight at first and finally pressed gently) one wedge has broken across, shattering. TEXT-FIG. 3, Pollen grains of (V. lignieri. a, b, grains with surface ornamentation sketched (pits not drawn individually), x 1000. c, free-hand drawing of edge of grain and of pore (to the left), x2000. D-P, outline drawings of grains showing the pore (dotted line) and folds, darker parts caused by folds obliquely shaded. In J and N, a pale strip, probably the groove beside the fold is outlined. Several grains have marginal folds. E, f, g, i have folds in the wreath position forming the edge or part of the edge of a broad basin. J, K, crossed folds on front and back, l p, grains with a deep groove, l, folds symmetri- cal. M, irregular, l and M, barley-grain form. N, groove to the left of the middle, no fold to the left of the groove, o, groove to the left and running on to the back, p, groove at extreme left and folds extending across to the right. All the grains are from slide ‘A’ except f, h, j, n, which are from slide ‘B’. HARRIS: W I LLI A M SO N I ELLA 137 Pollen of W. lignieri rarely shows such cracks (see however text-figs. 3b and 3e). Very likely no considerable general spreading occurred, and if any did the wall at the perimeter was able to extend. But radial cracks are known in dispersed spores, particularly in very thick walled ones, e.g. Tasmanites. No doubt authors studying cracked spores recognized the cracks as distortions, but if they recognized them as caused by peripheral extension following enlargement, they did not say so. I can see no other possible explanation. TEXT-FIG. 4. Histogram relating pollen grain shape with surface folds. Hori- zontal axis difference in n between grain length and width, vertical columns, number of grains of this shape in a count of about 180. White parts of columns, folds of the kinds that do not cause much change of shape, black parts, folds of the kinds that cause elongation and narrowing. Measurements to nearest 1 /n. Failure 3 : failure by folding. Even if it is adequate in other respects to support a load, the skin of a hollow sphere needs to have some rigidity. A very flexible skin, for instance soft leather, could never by itself form a rigid sphere. Different plastic balls sold in the shops have skins of different rigidity and probably also different extensibility and different compressibility and they varied in the way in which they folded, but by good fortune the first one purchased, the ‘playball’, happened to show just the effects seen in W. lignieri. It also had the merit of resilience and after compression regained its shape for the next trial. The ball was composed of a hydrocarbon polymer. The most pliant ball tested had a skin constructed of cotton and soft rubber. For a start cotton threads were wound in all directions on to the core (these enable rubber solution to stick to it), and when the rubber had become tacky it was rolled in cotton wool fibres till it would take up no more, let dry, again coated in rubber solution and cotton wool till the dry skin was about 2 mm thick. When the core was removed it proved just able to keep its shape in water but was so soft that it collapsed at once in air. The inextensible cotton fibres round the equator prevented any increase in diameter and the surface was thrown into many small folds. The largest of these were concentric folds at the margin itself and somewhat radial folds curving spirally out- wards from the pole, now the centre, to the edge. But there were many additional 138 PALAEONTOLOGY, VOLUME 17 smaller folds and not all are illustrated in text-fig. 5b. The folds on the two sides were unrelated. No pollen grain of fV. lignieri shows a surface crossed by many small folds and I conclude that at the time of compression its wall had rigidity comparable with that of the ‘playbair. In that ball the skin is at first very easy to deform, but as a depres- sion deepens the skin becomes more resistant and then as sharper folds are produced by compression it resists intensely. Thus even if the first depressions are so placed TLXT-HG. 5. Outline drawings of compressed hollow spheres; all drawn on the basis of the same original size (f). Below each compression is a real or imaginary section. A, failure 2 in rather firm putty. B, failure 3 in a very pliant, inextensible skin (soft rubber and cotton wool), c, d, g, h, k, l various folds, all failure 3 in a single plastic ball; skin resilient and readily forms large and rounded folds but resists small or sharp folds, k., l have two grooves, one on the front (lirm lines on one or both sides) the other on the back (dotted), e, failure 4 in an inextensible but compactible skin, chewing gum and cotton. F, the original size of all models. G, H, see above. I. failure 5 in rather soft putty, skin has spread out as in a but has stretched. (Small cracks have formed as in failure 2.) R, l, see above. HARRIS: WILLIAMSONIELLA 139 that they should produce several small folds these never develop, but instead one or two large ones are formed. However, various other miospores both from sporangia and ones dispersed in the rocks show this effect— an almost perfect outline but a wrinkled surface. Most of these are thin walled. An example from the Yorkshire flora is the outer membrane of the pollen grain of Elatides williamsoni and of the corresponding dispersed mio- spore Peritiopollenites elatoides, figured by Couper 1958 (but he has selected ones which have less folds than usual). Many similar spores have been figured, for example that of Macrostachya in Potonie 1962. The feature called ‘parallel folds’ which these delicate spores are apt to show is discussed on p. 143. Compression of the ‘playball’ readily produces a small dimple which on further pressure enlarges to transform the upper surface into a hollow hemisphere inside the lower hemisphere. When this is compressed still more the upper surface forms a cavity with overhanging margins and then the cavity becomes angular with three, four, or five corners. Finally, under heavy pressure the overhanging edges are pressed flat and we have three, four, or five small folds, which I term wreath folds; they are slightly nearer the outside than the centre (text-fig. 5c). Wreath folds (not always a complete set) are common in W. lignieri and sometimes it is easy to see that they surround what was a depressed basin (text-fig. 3g, h, i). The flattened ball has now a diameter which is slightly enlarged by up to 10% and its outline is no longer per- fectly round, but slightly elongated opposite the longer folds. Instead of a round dimple an elongated one may form and then further pressure makes it into a deep groove. The edges of this groove always overhang but they vary in different com- pressions, they may be well apart or may meet or even slightly overlap, because of diflferences in the width of the first elongated dimple. A common form of the ball when strongly compressed is the ‘barley-grain’ form (text-fig. 5d). When this forms the length of the ball is always increased and its width diminished, and so it is in IV. lignieri pollen grains (text-fig. 3l-p). The ‘playbalF was slightly translucent and by transmitted light the groove was pale and flanked by two broad dark folds where two extra layers of wall are seen. The wall is pale again at the edge but not quite as pale as in the groove. Although this barley-grain form is stable while under pressure, its skin is under considerable stress. These stresses are revealed, if crudely, by making a small slit in the skin. In tension the slit gapes, in compression the edges overlap. By manipulating the slit into different positions the distribution of tensions and compressions is revealed (text-fig. 6). It will be seen that the same piece of skin can be under tension transversely and compression longitudinally and no doubt under strong bending stresses as well. Clearly the net stress is longitudinal extension, pushing the ends of the grain outwards parallel with the groove and at the same time tension drawing in the margins opposite the groove. The length of the deformed ball reaches about 130% of the original diameter and when a slit is made in a part under tension the length at once increases to well over 130%. The forces concerned pushing out the ends and pulling in the sides are very consider- able. When the ‘playball’ was loaded with 8 kilograms of bricks the ends were thrust outward so strongly in the barley grain form that it took all the strength of my hand to push the ends back. At the same time as the ends were pushed back by hand the bricks were visibly heaved upwards. This outward thrust arises by a sort of lever 140 PALAEONTOLOGY, VOLUME 17 TEXT-FIG. 6. Tensions and compressions in compressed ball, the same as in text-fig. 5d. These stresses are displayed by a small slit, diverging arrows indicate tension, the slit gapes, converging arrows indicate compression, the sides overlap. Longitudinal compression occurs also in parts concealed by the folds. action and the lateral contraction is also powerful. Although the compression and extension forces in the ‘playball’ are powerful, its skin is not very compressible or extensible and it was difficult to measure any change in the distance between ink spots on its surface. Thus in a fossil spore that behaved like the ‘playball’, extension would not cause the wall to be appreciably thinner and paler. The ‘barley-grain’ distortion of the ‘playball’ is strikingly like some figures of fossil monocolpate pollen grains but I do not suggest that authors have made mistakes in their interpretation. In W. lignieri the distorted pollen grains with a groove flanked by two folds sometimes looked to me very like a true colpus, but often I would have doubted if they were real. I would have been deceived by the most colpus-like forms if I had seen these grains and nothing else. It is easy to form the first elongated dimple to the side of the middle line and then at first on compression it forms a groove flanked with one broad and one narrow fold (text-fig. 5g), but narrow folds are unstable in the ‘playball’ and with further com- pression the narrow one may vanish, some of it going to increase the large fold and some concealed smoothly in intense compression at the side of the groove. It causes the same elongation as the symmetrical barley-grain form. A corresponding form with a single longitudinal fold is the commonest of all in W. lignieri (text-fig. 3n, o). The first dimple can be put in a purely lateral position and as before enlarged to form a large hemispherical basin. This on compression takes a peculiar lemon-shaped out- line and its length, 140% of the original, is the longest produced by any fold (text- fig. 5h). It is rarely seen in W. lignieri (text-fig. 3p), and indeed a purely lateral first depression would seem unlikely to form. It is easy to impose distortions on to the ‘playball’ of kinds that it never shows ordinarily, for instance a ring-shaped groove or three parallel grooves but these are unstable and last only as long as they are held. No such forms were seen in W. lignieri. As mentioned, the first slight load causes a dimple and the position of this dimple and its round or elongated form determines the final form in heavily compressed distortion. In the ‘playball’ when compressed between boards the commonest forms were a wide dimple above, leading to wreath folds, or wide dimples above and below which then both became elongated and produced cross folds (text-fig. 5k, l). An elongated groove on the upper side seldom formed, yet in the pollen of W. lignieri a groove flanked by one or two folds is the commonest distortion. This difference is not explained. HARRIS: W I LLl A M SO N lELLA 141 Other purchased plastic balls gave forms which differed more or less from the ‘playbair. One rather harder ball gave very similar forms but the skin proved less able to absorb a small surplus smoothly by elastic compression. When it took the barley-grain form the floor of the groove was at first arched longitudinally and on further pressure this arch was unable to flatten smoothly but the surplus skin formed a small fold transverse to the main groove. In the still harder table tennis ball this effect was even more pronounced and two transverse folds formed (text-fig. 5j). I have not seen this distortion in any Jurassic miospore. The folded forms with a colpus-like groove are discussed later. Failure 4. The skin fails by permitting itself to be compacted and thickened under pressure and this thickening takes up the whole surplus. The final form is thus a vertical projection of the thickness of the skin at each point; the thickness in the middle is unchanged but considerably increased towards the edge. The precise way the thickness increases depends on the thickness of the skin as a fraction of the diameter and on the position. The following table gives computed figures for suc- cessive tenths of a radius measured horizontally from the edge to the centre for a sphere of external diameter 100 mm and internal diameter 98 mm. 123456789 10 4-8 3-4 2-8 2-5 2-3 2-2 2-1 2-0 2 0 2'0 It should be possible to recognize the increased thickness of the first tenth in a comparably compressed fossil spore. I also give a graph (text-fig. 7) to show the 20 15 10 5 " 1 2 3 4 5 10 15 20 25 30 TEXT-FIG. 7. Compression of a hollow hemisphere of external diameter 100 mm, internal diameter 98 mm. The wail sub- stance is compacted on to the horizontal; the left edge is at 0 and the steps are at 1 mm intervals. At 30 the thickness is 2-8 mm; at 50, the top of the hemisphere, it would be 20 mm. Vertical and horizontal scale units in mm. The thickness at 10, at 20, and at 30 correspond to those at the figures 1, 2, 3 in the table above. computed thickness in hundredths of the distance from the edge to the centre, the maximum which is at two hundredths is 19-9: I am grateful to Dr. D. T. Hopkins for these figures. If the skin had been thicker, say 5% of the whole diameter, the maximum would have been slightly further from the edge and there would have been more increase in the second and third tenths. 142 PALAEONTOLOGY, VOLUME 17 The first attempt to construct a model showing this distortion failed. This was the rubber and cotton wool skin but the rubber though very soft was too resilient. So for the next hollow model I tried the least resilient waterproof material I knew as a cement and this was the latex of chewing gum and it proved successful though it was unpleasant to use. It was prepared by boiling chewing gum in water to extract sugar and then the latex was dried and kept in benzene for some days in which it forms a sticky paste. This paste was applied in the same way as the rubber solution and the skin was built up in layers with cotton wool. The hollow ball was softened with warm water and let collapse in air. It gave a disc of the original diameter and without wrinkles except for a minute one at the very edge. The thickness was obviously greater at the edge but the skin had not been even enough for measurements to be worth while. Such a disc would certainly look darker towards the edge by transmitted light and the marginal fold would be visible as a double margin. It was evidently formed because even this wall was not sufficiently unresilient to take the great amount of compaction it needed to take. This sort of distortion may have occurred in the apparently unfolded round grains of W. lignieri but in only a few of these was thickening in the marginal tenth at all obvious. What was, however, to be seen was the double margin, a minute concentric fold right at the edge and evidently caused by the considerable compaction. Yorkshire dispersed spores were examined for this distortion. Some do show perceptible darkening at the edge and some show a small fold by the edge but many showed neither. Failure 5. The whole skin is allowed to spread out evenly because while it refuses to compress and shorten in the radial direction, it permits stretching in the tangential direction. Instead of a sphere of diameter 2r and with a distance from side to side over the top hemisphere of -nr we have a disc of diameter -nr. Its perimeter is increased in the same ratio, 77/2. The distortion is the same as in failure 2 except that there is stretching instead of splitting. The increase in the perimeter, 100 units to 157 units, is considerable and the area of the skin has increased. The disc looks altogether larger than the original sphere. At the same time there should be stresses in the skin and a certain thinning towards the edge and surface pits, while unaltered radially, should be extended tangentially. This distortion even if occurring fully would be hard to recognize and certainly unrecognizable if occurring to a small extent together with other distortions. For instance if combined with failure 4, compaction, there might be no change of wall thickness. Because it is so difficult to observe, it is difficult to say whether this distortion occurs to any considerable extent in compressed spores. It certainly occurs to a small extent in ones showing another distortion, that of failure 2 with cracked margins. In these there should in theory be a vast number of radial cracks running from the edge to the centre but instead one sees a few widely separated cracks which run only a short distance. The cracks are broad at the edge but they narrow and end much too soon and this must be because the stretched wall has pulled out smoothly. A compressed half grapefruit skin shows such stretching and splitting convincingly. For a model to illustrate failure 5 a skin is needed that combines incompressibility with extensibility and this is given by a rather soft putty. The solid part, chalk, is incompressible, but HARRIS: WILLIAMSONIELLA 143 the oil is still viscous and extensible. Text-fig. 5i shows a skin of soft putty which flattened under gravity in air; its diameter is about the theoretical one but the putty was slightly too much hardened and did crack slightly at the edge. Here we have mainly failure 5 but some of failure 2. It is unlikely that this sort of expansion occurred in the present sample of W. lignieri pollen. The effect should be maximal in a round grain without visible folds; folds by the margin or wreath folds would have taken the surplus skin. In fact the round, smooth grains are no larger than the others. Summary. In the flattening of a sphere, or indeed any solid, secondary distortions caused by the surplus skin are inevitable. These may be of several alternative forms arising from the mechanical properties of the skin. Some of these cause effects which are obviously distortions but others are much less obvious and may be invisible. There is no reason why two or more kinds of distortion should not occur together and prob- ably they do. At most these distortions would increase a spore diameter to 157% of the original but others cause no such enlargement. The spores least likely to have enlarged in this way are ones with a thickly wrinkled surface and ones with a compacted and double margin, but those with a wreath of folds or crossed folds— one front, one back— would not have expanded much. A single strong fold through the middle or a groove and two folds (like a colpus) are likely to cause much extension along the fold and contraction across it. Parallel folds. Two kinds of distortion may have been called parallel folds. At the side of a groove, whether an original colpus or a groove caused at an early stage of compression, later compression produces two separate folds which are almost parallel, but such folds are normally recognized for what they are. More commonly the term is used for the two edges of a single, elongated fold. Text-fig. 8a represents in section a simple fold in a spore wall or cuticle. Inevitably this has two edges and text-fig. 8b represents the same fold after moderate com- pression. On severe compression (text-fig. 8c) adjacent surfaces stick or fuse together. There is commonly also loss of substance, often four-fifths of a cuticle or spore wall vanishes. However, the ridge formed at the sharp bend remains and its width shows the original thickness of the membrane. Text-fig. 8d shows the same fold in surface view; the border is the original thickness and the part enclosed is slightly darker having two additional layers of membrane. In a thick or dark-coloured membrane this is obvious and the fold is recognized as one, but in a thin and colourless membrane it is scarcely darker and the two edges are the ‘parallel folds’. A clear polythene bag on which a fold is imposed and then compressed shows this convincingly by trans- mitted light, only the edges of the fold are recognizable. Even such a simple fold causes secondary distortions. A strip from the plastic ‘playbair was cut and compressed and the fold produced, as seen in side view, is drawn with some accuracy in text-fig. 8e, f. Note the small ogee shaped cavity, this becomes smaller and comma shaped with further loading. If this form on vertical compression gave a vertical projection of the thickness at each point towards the edge of the fold we would get a double membrane of the form in f, the vertical thickness being considerably increased between the almost obliterated fold cavity and the very edge. K 144 PALAFONTOLOGY, VOLUME 17 TEXT-FIG. 8. Folds in a membrane, a-c, imaginary membrane with a fold, in section. A has rounded folds, in b the fold is pressed flat, c the same as b after full compression. Much of the substance has vanished, d, the same fold in surface view. The simple fold appears as ‘parallel folds’ and the edges show the original thickness of the membrane. Light stippling marks two thicknesses of membrane, dark stippling the compressed edge of a fold where the thickness is more than double, e, compressed fold in a strip, cut from the ‘playball’ drawn to scale. The surface beneath was hard and flat but that above somewhat padded, f, projection of e on to the horizontal. COMPRESSION OF SPORES IN A MATRIX The matrix to a large extent controls the distortion of a compressed spore and is thus of the greatest importance but since this is outside the scope of the present study it is treated only lightly. A few experiments on the compression of balls in a matrix were made, however, and these indicated that much more might be done usefully. The whole of my ideas of compression of plant organs in a matrix came directly from Walton. He told me that he had compressed various solid plant organs— plant stems, apples, and the like in wet sand in a power press so constructed as to allow surplus water to drain away. But he never published his results in detail. Instead he used his findings to reassure himself when he dealt with compression in a theoretical manner (Walton 1936). His general conclusions were simple. A spongy solid such as an apple makes a hemispherical bed in the sand and under pressure it loses water to the sand and this water drains away. The apple presses down into its hemispherical bed and the sand above follows the apple. Thus we have a cup of compressed and dehydrated tissue occupied by a plug of sand. The diameter of the cup is exactly that of the apple. The final form of the ‘compression’, a term Walton put forward in his paper, is determined by the downward facing side; the top has but little eflfect. With a compressible matrix, mud, the effect is only slightly different; first the apple is sup- posed to form a cup and then the apple and matrix together compress, say to a third by dehydration of the mud. The cup becomes a saucer. Walton’s theory demands that the compressed tissue should have negligible rigidity (this is reasonable for rotted HARRIS: WILLIAMSONIELLA 145 plant organs) and he supposed that compression in mud took place first in the plant and then in the matrix. But I see no reason why the order should matter, and in my experiments the model and matrix compressed together but with the same effect. Walton went on to apply this idea to the form of certain Carboniferous plants but 1 suspect that it was the close study of these plants that aroused his thoughts about compression. 1 have used Walton’s theory many times with Jurassic compressions and always it has proved illuminating. But Walton dealt with a few cases only, all essentially spongy unresilient solids. He did not deal with the spore, a stiff empty skin. We need to study the effect on model spores, that is balls. At first the effect proved the same as that Walton found with spongy solids. Hollow wax balls (con- structed round a core — as were other balls) were buried in a matrix, extracted with warm water to remove the core and at the same time to soften the skin, compressed, and finally let cool. In sand the wax formed a hemispherical cup ; in compressible lawn- grass mowings it formed a saucer. Plastic balls, some with soft skins, some harder, were compressed in sand, and this time plaster of paris was added to the sand before it was wetted and compressed. When the compressed matrix hardened it was broken open and the ball extracted. There was a perfect hemisphere below while above there was a nearly hemispherical plug pressing down on to the ball. Another compression was made in dry sphagnum moss dusted with plaster of paris, a very compressible mixture. This was wetted before being compressed. The ball gave a mould which was essentially that of a saucer, though the surfaces showed wave-like distortions. There was no expansion that could be measured in this rough experiment. A punctured table tennis ball compressed in sawdust, a somewhat compressible matrix, gave a nearly round cup (PI. 15, figs. 13, 14). 1 was convinced that further experiments of similar kinds were unlikely to produce horizontal extension. It is hard to find a model which will expand horizontally, overcoming the Walton effect of the matrix, but since I was convinced that this does happen to spores, even if rarely, it was necessary to find and understand a model and to consider the forces involved. When Walton compressed a spongy solid, an apple, in sand there may have been no force at all favouring expansion if pressure was applied slowly enough for the juice to drain away. But a spore or a hollow ball with a resilient skin is different and we know that these balls sometimes expand when compressed between flat sur- faces in air. A spore can be regarded as a double dome, or for simplicity its top half as a simple architectural dome. A domed roof or indeed any raised roof tends to flatten under its weight and to push the walls outwards. In a roof it may be solved in various alternative ways, by cross ties (but these are not seen in spores), for a dome by a strong chain round its base, and by external buttresses. Our problem is the opposite, to construct a dome which will flatten and push the supporting walls outwards. Clearly we should choose a situation in which the outward forces develop strongly. The worst situation would be for the ball to be embedded in sand for here the top half simply collapses on to the bottom and no powerful outward forces ever develop. In a matrix we cannot escape the support given to the bottom half of the ball but we can minimize this by using a compressible matrix, for then the bed below our ball becomes a shallow saucer instead of a hemisphere. Then we must use a ball with a rigid skin, resistant to folds of any kind. Once a fold forms it accepts surplus skin and the stresses 146 PALAEONTOLOGY, VOLUME 17 caused by the surplus skin cease to be available to thrust out into the matrix. Clearly a roof dome of flexible rubber would collapse in folds without pushing the walls out- wards. The playball and other plastic balls had a rather rubber-like flexibility and this explains their failure to expand into the matrix. Then to return to the dome supported by a chain, we might cut or weaken the chain. Another kind of collapse could occur (but not in a building made of ordinary materials) by the dome changing its shape, if it changed from a circle to an ellipse it could push out the walls at its ends and drag them inwards at its sides. Finally we could make our external buttresses inadequate. Here are three different kinds of planned failure and two at least happen, as I believe, in fossil spores. Among the models so far considered the punctured table tennis ball stands out for its rigid skin. But this skin has also great tensile strength and this means that under vertical compression its equator will resist expansion. But we can weaken it— some small vertical razor cuts will do that — and the compressed dome should be able to spread out. Then for the deformation of the dome into an elliptical shape, this might follow if the first dimple on the surface became elongated rather than round. This could be imposed, or left to chance. As for providing a less resistant matrix, I do not doubt that some would resist less than others but the best could not be found without much experiment. All compressible matrices tested became more resistant when compressed, and those which like wet sphagnum or wet cotton wool stay compressed show this unmistakably. The first one used which gave success happened to be dry cotton wool. Very likely another material would have been better. To some extent the table tennis ball avoids this difficulty of consolidation of the soft matrix for under moderate load it collapses suddenly and audibly and further load (my weight) merely led to general compression. The compression of the 40-mm table tennis ball was done in a tin considerably wider than this with a wooden, piston that just fitted the tin and there was ample cotton wool above and below the ball. The effects were simple. The compressed ball was always of saucer-shaped contour and this distortion was added to the others. The height of the saucer was about 40% of the original but I believe that this is partly due to resilient recovery and that while under pressure it was flatter. For the same reason the upper surface might not be particularly hollow, but stand away from the lower though under pressure was pressed against it. The lower surface was strongly convex. A ball without any previous treatment became elliptical, 47 mm x 34 mm, that is 117% and 85% of the original diameter. The upper surface has a longitudinal groove of the ‘barley-grain’ type while the lower has a long shallow groove below one of the bulges on the upper side. This deformation of the under surface is outside what Walton’s theory contemplates. There are in addition two small inward folds at the compressed sides which have taken a good deal of skin and account for some of the narrowing. Sharply folded edges have cracked but are not displaced (PI. 15, figs. 11, 12). Another ball was weakened by six short vertical razor cuts about 3 mm long round the equator and in compression all these cuts had extended inwards by tearing, the longest gap being 15 mm on each face. The shape is a star with six blunt ended rays. The concave upper surface is mainly smooth but some rays have a deep groove in the ‘wreath fold’ position and this has evidently taken the surplus skin, for here the rays extend only about 20 mm from the centre, while elsewhere they are 23-25 mm HARRIS: W I LLI A M SO N I E LLA 147 long. The maximum diameter is 48 mm, 20% expansion. The under side is convex and smooth (PI. 15, figs. 4, 10). It will be seen that while both balls have extended considerably, the enlargement is a good deal less than the theoretical maximum of 57%. Some must be accounted for by the saucer-shaped form and some must be taken by internal strains. In both com- pressed balls the perimeter remains what it was, within the accuracy of rough measure- ment. In spores such deformation would be expected where the wall is strong and rigid and where moreover it keeps its rigidity for a long time after deposition and well into the period of compression of the sediment. And in such spores deformation of the kind giving the blunt rayed star is found, even if only occasionally. It is parti- cularly prevalent in the very thick-walled Tasmanites. I will here admit that it was the robust appearance of Tasmanites that led me to try the table tennis ball and the theo- retical treatment of stresses in a domed roof was supplied after the successful result. But for most dispersed miospores I see no evidence that any spread at all into the matrix has happened and I suppose rather that they have gone through the dis- tortions of a hemisphere and a flattening saucer and so in the end are spared serious alteration of size and shape and even the surface may be scarcely folded. But this is little more than an impression; serious study is needed. Acknowledgements. I am deeply grateful to friends who read the manuscript with great care and made corrections and also critical suggestions: Mr. C. Hill; Professor W. G. Chaloner; Dr. H. Van Konijnenburg Van Cittert. I am particularly indebted to Professor Lundblad, both for help with the manuscript and for much assistance at Stockholm. REFERENCES CLAYTON, G. 1972. Compression structures in the Lower Carboniferous miospore Dictyotriletes admirabilis Playford. Palaeontology, 15, 121-124, pi. 27. COUPER, R. A. 1958. British Mesozoic microspores and pollen grains. Palaeontogr. 103B, 75-179, pis. 15-31. CRIDLAND, A. A. 1957. WHUamsoniella papillosa, a new species of Bennettitalean flower. Ann. Mag. Nat. Hist. ser. 10, 12, 383-388. FLORIN, R. 1933. Die SpaltoflTnungsapparate der Williamsonia — , Williamsoniella — und Wielandiella — Bliiten (Bennettitales). Ark iv for hot., K Svenska Vetenskapsakad. 25A, 1-20, 1 pi. HARRIS,!. M. 1 944. A revision of Williamsoniella. Phil. Trans. Roy. Soc., London, (^) 231, 3 13-328, pis. 25, 26. 1969. The Yorkshire Jurassic Flora. III. Bennettitales. vi-|-186 pp., pis. 1-7. British Museum (Nat. Hist.) Geol. London. HUGHES, N. F. and MOODY-STUART, J. c. 1969. A method of stratigraphic correlation using early Cretaceous miospores. Palaeontology, 12, 84-111, pis. 13-22. KONIJNENBURG VAN CITTERT, J. H. A. VAN. 1971. In situ Gymnosperm pollen from the Middle Jurassic of Yorkshire. Acta Bot. Neerl. 20, 1-96, 16 pis. NATHORST, A. G. 1909. Palaobotanische Mitteilungen, 8. fiber Williamsonia, Wielandia, Cycadocephalus und Weltrichia, K. Svenska Veten.sk Akad. Handl., Stockholm, 45, 1-38, pis. 1-8. 1911. Palaobotanische Mitteilungen, 9. Neue Beitrage zur Kenntniss der Williamsonia— WuiQx\. Ibid. 46, 1-33, pis. 1-6. POTONiE, R. 1962. Regeln, nach denen sich die Sekundiirfalten der Sporen bilden. Paldont. Z, 36, 46-52, pi. 6. 1967. Versuch der Einordnung der fossilensporae dispersae in das phylogenetische System det Pflanzenfamilien 1 . Thallophyta bis Gnetales. II Teil Angiospermae. Forsch. Ber. Land. Nordrh. Westfal. 1761,11-310. 148 PALAEONTOLOGY, VOLUME 17 THOMAS, H. H. 1915. On Williamsoniella a new type of Bennettitalean flower. Philos. Trans., London, 207B, 113-148, pis. 12-14. TRALAU, H. 1968. Botanical investigations into the fossil flora of Eriksdal in Fyledalen, Scania. II. The Middle Jurassic microflora. Sveriges Geolog. Undersokning. C, 633, Avh. och. Uppsatsev, Arsbok, 62, 1-185, 26 pis. WALTON, J. 1926. On the factors which influence the external form of fossil plants. Phil. Trans. London, 2268,219-237, pis. 31, 32. TOM M. HARRIS Department of Geology The University Revised manuscript received 10 April 1973 Reading RG2 6AB EARLY GROWTH STAGES IN RHABDOMESOID BRYOZOANS FROM THE LOWER CARBONIFEROUS OF HOOK HEAD, IRELAND by RONALD TAVENER-SMITH Abstract. A number of specimens of young rhabdomesoid colonies were examined, each carrying a curious conical proximal termination. Sectioning showed that these represent the earliest growth stages, following fixation of the ancestrula. External walls of cones have a two-fold structure comprising outer, primary (recrystallized) and inner, secondary (laminar) layers. The arrangement corresponds with that in the basal walls of encrusting trepostomates. There are reasons for believing that during later development the conical structures were progressively obscured by calcification from an external mantle. Some of the young colonies grew around productid spines or other slender objects, and in such cases an internal ‘basal wall’ defines an axial tube. In others there is no such complication and zooecial tubes diverge from a simple skeletal axial rod. Differences in the origin and structure of the axial parts of rhabdomesoids and the mesotheca of ptilodictyoid cryptostomates suggest that these groups may not have been more closely related to one another than either was to certain trepostomatous stocks. About a dozen specimens representing what appeared to be minute rhabdomesoid colonies growing from conical structures were recently sent to the writer by Dr. G. Sevastopoulo (Trinity College, Dublin) and Dr. G. Lane (University of California, Los Angeles). The specimens were collected from shales of the upper Michelmia Beds of Tournaisian age (Smyth 1930, p. 533) at Hook Head, County Wexford, Ireland. They were particularly welcome, for they promised to shed light on the early develop- mental stages and mode of growth of this common but little-studied group of bryozoans. The Sub-order Rhabdomesoidea is one of three major divisions of the Cryptostomata proposed by Astrova (1964). It comprises ramose forms with slenderly cylindroid growth habit and (almost always) radial symmetry. Although the group ranges from the Ordovician to Permian, these fossils are particularly well known from certain Lower Carboniferous strata where they occur as fragments of colonies up to 30 or 40 mm long by 1 or 2 mm wide. Specimens of complete colonies or of the attachment regions between colony and substrate are, to the writer’s knowledge, as yet unrecorded. Rhabdomesoids constitute the greater part of the group sometimes loosely referred to as ‘stick bryozoans’. External morphology. The specimens varied from 1-3 to 3 0 mm in length and from 0-7 to 0-9 mm in diameter. On examination they proved to be early growth stages of rhabdomesoid bryozoans and each showed at its proximal end a conical structure. These were on average 0-8 mm long and, originating from a slender tip, each expanded distally, curving slightly along its length (PI. 16, figs. 1, 2). Microscopic examination showed that the external surfaces carried closely spaced transverse growth wrinkles. The distal (wide) end of each of these cone- or horn-shaped structures showed a gently lobate or scalloped margin, which was thickened into a notable rounded rim marking the limit of growth of the cone. Each lobe around the rim formed the lower margin of one of the first series of zooecial apertures of the new colony. The two other sides [Palaeontology, Vol. 17, Part 1, 1974, pp. 149-164, pis. 16-17.] 150 PALAEONTOLOGY, VOLUME 17 of each initial aperture combined to form a vaulted shape above the lower rim, so that the complete aperture has a rounded-triangular appearance and a basal width of about 0-2 mm. The position of each zooecial tube within the cone is evident, for it forms a rounded surface swelling with a linear shape narrowing proximally (PI. 16, figs. 1, 2). There are five to eight apertures in the first cycle around the rim of the cone. They are succeeded by others having a polygonal outline, commonly hexagonal or rhombic (PI. 16, fig. 3), though the exact shapes have in many cases been obscured by minor recrystallization or silicification. Numbers of large acanthopore nodes, 0 03 to 0 05 mm in diameter and of about the same height, project from zooecial walls at the zoarial surface. Although most of the cones studied show perfect horn shapes, a few depart from this pattern to varying degrees. In an extreme case the colonial origin is attached to a fragment of Cladochonus stem, the ‘cone’ being quite asymmetrical and tending to wrap around its tubular substrate (PI. 16, fig. 3). The young rhabdomesoid colony growing from this distorted origin nevertheless assumed an orthodox cylindroid form which grew forward parallel with and close to the Cladochonus stem. One well- developed cone with zooecial tubes just emerging showed a terminal surface with an evenly convex curvature, distal extension not yet having given rise to a cylindroid shape. The convex zooecial growth surface comprises a peripheral cycle of initial apertures, eight in number, occupying lobate projections of the cone rim. Within these are other apertures, twenty-nine in all, which become progressively smaller and more closely spaced towards the apex of the convex terminal face. It is clear that even at this early stage a vigorous growth zone, involving active subdivision to initiate new zooecia, was already established at the distal tip of the young colony. Internal structure. This was investigated using specimens embedded in an epoxy resin which were cut to provide one or more oriented surfaces. These were polished, etched with EDTA, shadowed with gold-palladium, and subsequently studied using a Cambridge scanning electron microscope. A few specimens were disappointing, for incipient silicification had destroyed much of the skeletal detail, but in others this was well preserved and permitted a full interpretation of the microstructure and growth sequence. Transverse sections close to the proximal apices of cones (text -fig. 1a) showed a small number (three to five) of the initial cycle of zooecial tubes, but the ancestrula could not be positively identified. In one case (text-fig. 1b) the tubes were arranged EXPLANATION OE PLATE 16 Figs. 1 -6. Scanning electron micrographs of whole mounts or polished sections of young rhabdomesoid (bryozoan) colonies. 1, Initial cone with zooecial apertures at distal end. x77. 2, Cone showing trans- verse growth ridges, viewed from proximal end. x70. 3, Distorted 'cone' resulting from growth of the earliest-formed zooecial lubes over a Cladochonus stem (c). Commencement of the ramose part of the colony, with zooecial apertures, is also seen. x75. 4, Longitudinal section through the peripheral wall of a cone, showing the recrystallized primary (p) and laminar secondary (s) components. X 1450. 5, Detail of peripheral wall of cone showing secondary fibres arranged in laminae oblique to the junction between primary and secondary layers. X 2900. 6, Longitudinal section at proximal end of the cone showing primary (p) and secondary (s) layers of the peripheral wall passing inward as an investment around a productid spine, only one side of which is visible (x). X 1450. PLATE 16 1 TAVENER-SMITH, young rhabdomesoid colonies 152 PALAEONTOLOGY, VOLUME 17 TEXT-FIG. 1. Transverse sections of initial growth cones, a, near proximal end. b, c, about midway along cone length. D, near distal end. Abbreviations: ha, hollow axial tube; ps, primary skeleton; ss, secondary skeleton. around a central cavity which was not a zooecium or ancestrula, but may have represented some foreign structure, perhaps entirely soft-bodied, around which the cone had developed. More distal sections showed an increase in the number and diameter of zooecia, with a few small, new tubes appearing in the axial region (text- fig. Ic). Close to the wide distal end of the cone transverse sections clearly showed the peripheral, large diameter tubes of the first cycle, and within them a considerable number of smaller tubes, decreasing in diameter towards the centre (text-fig. Id). It was evident from these transverse sections that zooecial walls within the cone were structurally continuous with the peripheral walls. This left no doubt that the cone was, in fact, an integral part of the bryozoan, and constituted an initial cup from which the mature colony took its origin. Longitudinal sections confirmed this, and showed the outer walls of the cone to be of two-fold construction: an outer, finely granular layer from 12 |um to 36 p.m wide passes internally into a laminated layer with an average width of 24 ixvn (PI. 16, fig. 4). The granular layer shows no obvious structure and may TAVENER-SMITH: RHABDOMESOID BRYOZOANS 153 be recrystallized. The laminar one consists of large numbers of roughly parallel fibres averaging about 5 ixm wide and 600 nm thick, arranged in layers, one upon the other (PI. 16, fig. 5). The resulting laminae are inclined at angles of 25 to 30 degrees to both the junction with the granular layer and the internal surface of the laminar wall. At the distal extremities of a cone sections showed that the granular layer becomes attenuated, but the laminar one thickens considerably to form the greater part of the peripheral rim. The two layers of the cone wall, granular and laminar, will hereafter be referred to as the primary and secondary layers, respectively, for they undoubtedly originated in that order. It is clear that the walls of the cone are identical in structure with the basal walls of encrusting trepostomatous colonies, and also with outer walls in many other members of the Stenolaemata (Tavener-Smith and Williams 1972). Internal zooecial walls within the cone, as well as those in later-formed parts of colonies, show the kind of ultrastructure commonly associated with such walls in the Trepostomata and Cryptostomata. Apart from certain exceptions mentioned later, they are constructed entirely of secondary fibres with widths of 3 to 5 /xm and thicknesses of between 350 nm and 1 ^m. Laminae formed from successive layers of such fibres show a distally arched arrangement within walls (PI. 17, fig. 2), and in their axial parts successive laminae commonly exhibit a selective thickening. Zooecial walls are relatively thin (average width about 9 jxm) within the cone and in the axial part (endozone) of the fully formed cylindrical stem beyond it. Where zooecial walls form the outer surface of the cylinder beyond the limits of the cone, however, it is clear that once the diameter of the initial cone had been attained these walls com- menced to thicken, signalizing the commencement of the exozone. Primary tissue is absent from zooecial walls in the endozone, though finely granular tissue not of primary origin constitutes the axial parts of large acanthopores in the exozone. Three of the specimens sectioned longitudinally showed evidence of the original presence of foreign structures within the young colony. This was particularly clear in one case where a short section of a productid spine in an early stage of growth had been lapped around and overgrown by the extending bryozoan tissues (text-fig. 2). The stump of the spine projected slightly from the pointed proximal end of the cone and sections showed that, in contact with it, the primary and secondary layers of the cone outer wall curve round and pass upwards on both sides, forming a complete sheath around the spine (PI. 1 6, fig. 6). The latter lies wholly within the cone occupying a roughly axial position : it is probable that it provided support for the young colony and helped to maintain it in a suitable growing position. The calcareous wall secreted by the bryozoan colony as an investment around the productid spine has a structure (PI. 17, fig. 1) identical to that of the cone outer wall, with which it is continuous. The thickness of this internal investing wall varies from 7 to 20 i^m, averaging 12 |Lxm. The primary wall component ranges in thickness from 4 to 16 ju.m, being most com- monly about 8 fj,m, and secondary laminae constitute the rest. Fibres composing the secondary laminae compare in dimensions with those of the outer wall of the cone. In the specimen referred to, one end of the productid spine is embedded within the young colony, and at this end it is apparent that the spine had been broken before being overgrown by the investing tissues, which unite beyond it (PI. 17, fig. 4). Many zooecial walls in this colony originated from the laminar secondary layer on the inner side of the wall around the productid spine (PI. 17, fig. 1). This is a natural 154 PALAEONTOLOGY, VOLUME 17 consequence of the formation of successive new zooecial tubes (perhaps on the pattern of a helicoid spiral around the spjne) from the axial part of the developing colony. Beyond the inner end of the spine the investing walls merge to form a poorly defined axial rod (text-fig. 2) which may, nevertheless, be traced to near the distal apex of the young colony. This structure commonly follows a somewhat irregular path, being slightly offset alternately to right and left. It is slightly wider (9 to 11 ^nm) than the zooecial walls which arise from it on either side. The structure of the axial rod is similar to that of zooecial walls, the main difference being that axial thickening of the distally arched laminae may be locally pronounced (PI. 1 7, fig. 6). In transverse sections this leads to the appearance of granular tissue (not of primary origin) in the medial part of the axial rod, and this may extend for short distances into zooecial walls which diverge from it. The morphology of these young rhabdomesoid colonies includes many features strongly reminiscent of ramose Trepostomata, to certain of which the Rhabdo- mesoidea as a group may have been closely related. These include tubular zooecia (though not as long as is general in the Trepostomata, and lacking diaphragms), the presence of recognizable endozonal and exozonal regions, large acanthopores and zooecial walls with laminar structure, the laminae being convex towards the exterior. As with so many rhabdomesoids and ptilodictyoids (bifoliate cryptostomates), differences from the Trepostomata are mainly matters of degree, rather than the presence of distinctive new structures or the entire disappearance of old ones. Mode of growth. It is clear from the situation and structure of the outer wall of the cone at the proximal end of each specimen, that this corresponds morphologically to the attached base of an encrusting stenolaemate colony. It is as if, centred on the ancestrula, a circular basal plate had been reflexed away from the substrate. There may be comparable cases among certain arthrostylid genera (for example, Arthro- styloecia, Ulrichostylus, and Sceptropora) which seem to have similar cone-like structures at their proximal ends (Bassler 1953, pp. G128, G130). Parallels between the outer cone wall and an attached colonial base emerge still more clearly if the circumstances and mode of their formation are considered. The external walls of the initial cone are also the external walls of the first cycle of zooecia that lay within it. These walls must, therefore, have been secreted from the cellular epithelium which lined the zooecial tubes and represent the calcification of EXPLANATION OE PLATE 17 Figs. I 6. Scanning electron micrographs of polished longitudinal sections through young rhabdomesoid (bryozoan) colonies. 1, Primary (p) and secondary (s) bryozoan wall tissue overlying one side of a pro- ductid spine in the axial region of a colony. The secondary tissue forms the proximal part of an internal zooecial wall, x 1400. 2, Typical zooecial wall structure, showing the distally convex secondary skeletal laminae, x 2800. 3, Three-dimensional view of the edge of a zooecial wall, showing the fibrous nature of the units composing each lamina. X 3000. 4, A view showing the continuity of bryozoan skeletal tissue around the distal end of a productid spine in the axial part of a colony, x 700. 5, Detail of Fig. 4 showing primary (p) and secondary (s) wall components overlying the fractured extremity of the productid spine (x). x 1400. 6, Part of the axial rod showing medial thickening of distally arched laminae. X 1400. PLATE 17 TAVENER-SMITH, young rhabdomesoid colonies distal end cb zo ee hs ss ps ze ar pe pr cone TEXT-HG. 2. Medial longitudinal section of young rhabdomesoid colony with proximal cone, showing main skeletal layers and reconstructed position of epithelium. Abbrevia- tions: ar, axial rod; cb, common bud; ee, eustegal epithelium; hs, hypostegal space; pe, periostracum; pr, productid spine; ps, primary skeleton; ss, secondary skeleton; ze, zooidal epithelium; zo, position of zooid. TAVENER-SMITH: RHABDOMESOID BRYOZOANS 157 TEXT-FIG. 3. Reconstruction of earliest stages of growth from the ancestrula in a rhabdomesoid colony. Abbreviations: ac, ancestrula; cb, common bud (initiated in axial position); fz, first zooid, growing directly from ancestrula; pe, perio- stracum, or external ‘cuticle’; ps, primary skeleton; se, septum, calcified proximally to form the first internal zooecial wall; ss, secondary skeleton; su, substrate; sz, second zooid; tm, terminal membrane; ze, zooidal epithelium. the originally soft limiting layer of the young colony (text-fig. 3). By analogy with modern cyclostomatous bryozoans (Borg 1926, pp. 191-194) this would have com- prised an ectodermal epithelium, sealed on its outer side by a polymerized acellular exudation here called the periostracum. Walls within the cone which served as an immediate investment around a spine or other foreign structure must also have origin- ally been bounded by periostracum, for they are continuations of the outer wall and structurally identical with it (text-fig. 2). From an anatomical point of view such walls must be considered ‘external’ and, by their direct adhesion to foreign objects, they demonstrate the same relationship as that between the basal wall of an encrusting colony and its substrate. Walls between zooecia within the cone and beyond it differ from those already mentioned both in structure and derivation. They consist entirely of laminar secondary skeleton and there is no reason to suspect the former presence 158 PALAEONTOLOGY, VOLUME 17 of primary tissue or a periostracal layer at any stage. These walls are continuous with and derive from either the inner secondary lining of the peripheral wall or from an axial rod composed of the same material. The uniform construction of common zooecial walls from superimposed, distally arched laminae which persist to the exozonal extremities, makes two clear demands in interpreting the origin of these structures. They are, firstly, that the wall was completely enclosed within a cellular envelope which must have been the zooidal epithelium; secondly, that the wall grew in length by the successive apposition of calcite laminae at the distal end. The epithelial envelopes within which mineralized zooecial walls necessarily had their inception must have arisen as plate-like invaginations of the epithelial layer lining the peripheral wall of the cone (text-fig. 3). Furthermore, the configuration of laminae at the distal terminations of shared zooecial walls indicates that the epithelium was continuous across the distal end, and therefore also between adjacent zooids. From this several further inferences may be made, namely: that the distal margins of zooecial walls were not in direct contact with the frontal wall of the colony ; that they were separated from this by what can only have been a narrow coelomic space which allowed physical continuity between the body cavities of neighbouring zooids; that, lacking skeletal evidence to the contrary, the frontal surface of the colony was entirely composed of soft tissues; and that, by the nature of their origin and construction, zooecial walls were essentially internal partitions. A logical corollary from the above is that the frontal membrane of the colony, with the associated lophophore and ancillary apparatus of individual zooids, can only have been a continuation of the epithelial- periostracal layer which, in the proximal region, secreted the outer wall of the cone (text-figs. 2, 3). Arrangements of the above kind, in which mineralized zooecial walls did not meet the frontal membrane and individual zooids were not entirely separated from one another, have been and are common to many groups of bryozoans. They are known to occur in certain cyclostome stocks (Ceramoporoidea, Cerioporina, Cancellata, and Rectangulata) and probably did so throughout the Cystoporata, Trepostomata, and Cryptostomata. They are likely to have been the rule rather than the exception and the condition, though at first sight complex, is probably primitive rather than advanced (Tavener-Smith and Williams 1972, p. 155). The arrangement and structure of walls comprising the initial cone leave no doubt that the manner of ‘budding’ off new zooids in these colonies was essentially the same as that described by Borg (1926, p. 256) from modern Cyclostomata. The same pattern may well have been general throughout the Stenolaemata. It involves the repeated division of a continuously growing cavity or cavities by calcareous septa fabricated within sheet-like invagina- tions of the zooidal epithelium in the manner described earlier for zooecial walls (text-fig. 3). In this way successive daughter cavities (new zooecial tubes) were formed. As this process continued and newly formed tubes lengthened, the cavity which under- went repeated division (the ‘common bud’ of Smitt 1865, p. 6) remained at the focal point of an increasing group of slightly divergent tubes. In this way an axis of active growth was established, with the continuous presence at the distal tip of the develop- ing colony of a group of lately formed zooecial tubes of small diameter. On the basis of the observed skeletal structure of the specimens, and the recon- struction of secretory tissues which that permits, and in the light of what is known of TAVENER-SMITH: RHABDOMESOID BRYOZOANS 159 growth processes in modern bryozoans, a developmental sequence may be suggested to account for the morphological features of the colonies examined. Settlement and fixation of the larva to form an ancestrula must have been followed by the distal growth and distension of the latter, and the formation within it of the first septum (internal zooecial wall). This duly divided the available space into two parts, con- nected only across the distal margin of the septum, which just failed to reach the terminal membrane (text-fig. 3). One of the chambers so formed would have con- stituted the first autozooid of the new colony, the other being the common bud which repeatedly divided in the same way to form a succession of slightly divergent zooecia. As these tubes lengthened and increased in diameter their outer walls, at first delimited only by combined layers of epithelium and periostracum, became mineralized. First granular, and later (more slowly, and by the addition of successive increments) laminar material was deposited to form the solid outer wall of the cone. This must, as an inherent result of its mode of origin, represent the greatly lengthened and dis- tended apertural rim of the ancestrula. The walls of the cone continued to extend distally, as indicated by successive trans- verse growth wrinkles, until the zooecial tubes of the first cycle had attained their mature, genetically ordained length. At this stage the diameter of the future zoarial cylinder was provisionally determined and zooecial extension largely ceased, though some thickening of the cone rim took place (the inception of the exozone) due to the continued deposition of calcite in more or less static circumstances. This continued as long as the zooecia were occupied by functional zooids. By the time the first cycle of zooecia, whose outer sides constituted the walls of the cone, had attained maturity it seems that repeated subdivision of the ‘common bud’ in the axial part of the young colony had already given rise to many more immature tubes. As each generation of these lengthened and attained maturity, so the colony length also extended and the ramose cylindroid shape gradually emerged. The tapering distal apex continued to mark the region of active subdivision and growth. DISCUSSION These specimens show early stages in the formation of rhabdomesoid colonies. In each of them an ancestrula gave rise to a small number of slightly divergent zooecial tubes. In growing distally these formed an initial cone from which the characteristically cylindroid main part of the colony originated. In some specimens it seems that the cone developed from the basally attached ancestrula without external support to help maintain it in a favourable growth position. This is surprising, considering the minimal nature of the attachment area. Other specimens showed a slender axial hollow within the cone (text-figs. 1b, 4), suggesting that temporary support may have been derived from the overgrowth of some soft-bodied organism which later decayed. In others the initial zooecia grew against, and then around spinose or tubular sub- strates which were skeletal parts of other organisms, so that these became incorporated within the cone. In all these instances only the earliest formed parts of colonies are concerned, and it is relevant to inquire whether such arrangements for attachment and support were adequate in fully developed colonies, or whether considerable modification would be called for in the course of subsequent growth. L 160 PALAEONTOLOGY, VOLUME 17 growth cb ar ef at pe ps ss proximal end ee cr zo ze ■?a TEXT-FIG. 4. Median longitudinal section through an initial cone contain- ing an axial hollow outlined by a thin layer of primary skeleton. The epithelium is reconstructed to show (left side only) how an external flap, or mantle, might have permitted calcification of the basal region from the exterior. Abbreviations: ar, axial rod; at, axial tube; ?a, possible ancestrular chamber; cb, common bud; cr, rim of initial cone; ee, eustegal epithelium; ef, external flap; pe, periostracum; ps, primary skeleton; ss, secondary skeleton; ze, zooidal epithelium; zo, position of zooid. Modern cheilostomatous colonies with slenderly ramose forms comparable to that of rhabdomesoid cryptostomates assume an erect growth habit in quiet water of moderate depth (Stach 1936, p. 62; Lagaaij and Gautier 1965, p. 52). Schopf (1969, p. 236), although in general agreement regarding the habitat of the erect Cheilo- stomata, suggested that at the depths at which they are commonly found (100 to 150 m) water movement may be appreciable. Cheetham (1971, pp. 7-12) examined TAVENER-SMITH: RHABDOMESOID BRYOZOANS 161 the mechanical implications of lateral stresses generated by water movements upon developing colonies of erect growth habit. He recognized that as a colony extended distally the load resulting from the weight of the structure itself and the torsional stress due to laterally directed water pressure would increase. Also, that these forces would be most pronounced at the base of the colony, and that therefore the need for greatest skeletal support would be in that region. He concluded that the skeletal pattern manifested by many of the erect Cheilostomata represents a mechanically elhcient means by which such colonies may support themselves and counter water pressures acting against them by permitting a degree of controlled flexibility. Such arrangements most commonly incorporate a strongly calcified attachment zone at the base of a pillar-like colony, the latter having a rigid outer frame (frontal walls of successive zooids) thickening progressively towards the base, and a relatively light internal structure. Fragments of rhabdomesoid colonies and of the ramose Trepostomata show what was probably an equally effective skeletal arrangement, embodying the same mechani- cal principles but differing in design. In these groups a thin-walled endozone is bordered peripherally by the more massive walls of the exozone, which were progressively added to while the colony lived and extended distally. The structural similarity to the ramose Cheilostomata strongly suggests that these colonies also were adapted to withstand stresses of the kind associated with an erect growth habit, and that they too had rigidly attached bases. In view of these considerations it is unlikely that unmodified cones of the kind seen in the proximal regions of the specimens examined would have provided adequate means of attachment or support for fully developed ramose colonies. They would have been inherently weak, and it is therefore reasonable to suppose that the basal attachment was progressively reinforced and strengthened during the later life of the colony. Such additions would have been made most effectively from the exterior, and this may well have been accomplished by secretion from the inner surface of an external mantle, formed as a flap-like proximal extension of epithelial layers from the rim of the cone (text-fig. 4). Mineralization on this pattern, in conjunction with exozonal thickening and possibly the eventual sealing of zooecial tubes in the basal region, could have provided a strongly calcified zone of attachment, as in hornerid and fenestellid bryozoans. Whether this in fact took place will not be known until specimens showing the proximal parts of fully developed rhabdomesoid colonies are available, but if such was the case it is certain that the early developmental stages shown by the present specimens would have been concealed and lost for study pur- poses in the early stages of overgrowth. Longitudinal sections of cones and subsequent cylindrical parts of the colonies examined showed certain variations of internal structure. In some sections zooecial walls took their origin from an axial rod which represented successive positions of the inner wall of the ‘common bud’ chamber. In such cases the axis of the young colony was solid (text-fig. 1a, c, d). In others zooecial walls could be traced back to the inner side of an axial tube which, as indicated earlier, was in effect an invagination of the peripheral wall of the cone. In some instances this tube was the basal bryozoan invest- ment around a productid spine, but in others no such foreign structure was present and an axial hollow or tube existed within the developing colony (text-figs. 1b, 4). 162 PALAEONTOLOGY, VOLUME 17 Such a space may originally have been occupied by some soft structure, used as a temporary support in the earliest growth stages, which later decayed. One specimen showed a tubular internal ‘basal wall’ enclosing a productid spine and closing beyond its broken end, so that the tube was succeeded by an axial rod (text-fig. 2). Although no instance was seen, it would presumably be possible for an axial hollow to pass into a solid-cored zoarial cylinder in the same fashion. This is of interest when it is recalled that the currently accepted diflference between two largely contemporaneous and commonly coexisting ramose cryptostomatous genera, Rhabdomeson Y oung and Y oung and Rhombopora Meek, is the presence in the former of an axial tube which is not found in the latter. Good examples of the early developmental stages of bryozoan colonies are not common, particularly among palaeozoic genera. Such specimens are of great interest, for they may yield structural information of value in assessing phyletic relationships between groups. This is especially important in the Cryptostomata, for stratigraphic data give little help in interpreting the affinities of the three major divisions of the Order, either to each other or to the Trepostomata, to which they are related. Certain aspects of the colonies examined seem to provide useful pointers in this connection. It is clear that the morphology of the juvenile colony was determined by budding from an attached ancestrula on a narrowly divergent pattern. The initial form evolved was therefore a cone. Subsequent cyclic budding from the distal end of the cone initiated new, tubular zooecia, each originating in the axial region and diverging towards the periphery. These zooecia were essentially similar to those of the Trepo- stomata and their continuous, systematic addition resulted in the emergence of a ramose, cylindrical form with radial symmetry. Although the cone may superficially suggest an encrusting colonial base refiected away from the substrate, it cannot have originated in that way. The development of a particular three-dimensional shape in stenolaematous bryozoans is largely depen- dent upon the size and situation of the ‘common bud’ area, that is, the part of the periphery possessing the capacity to subdivide unspecialized colonial spaces by the development of internal septa (Borg 1926, p. 256; lilies 1968, p. 225; Brood 1972, p. 41). To permit the formation of a basal disc of encrusting zooecia it is prerequisite that each of a circlet of tubes originating from and spreading radially over the sub- strate around the ancestrula should have possessed that ability. In the specimens examined it is clear, however, that the capacity to proliferate by fission was entirely restricted to a zone lying within the first-formed cycle of zooecia which was itself directed away from the substrate. In these circumstances an encrusting colonial base cannot possibly have formed. In contrast it is instructive to consider the organization of a typical ptilodictyoid such as Stictopora Hall. Here the symmetry is bilateral, not radial, and the zoarial form a flattened, bifoliate frond. Zooecia arise on either side of a medial partition, or mesotheca, and, due to the restricted width of the frond, are shorter and more compact than in the Rhabdomesoidea. Zooecia of this kind are closer to the more specialized and advanced zooecia of the Fenestelloidea than are those of the rhabdo- mesoids. The mesotheca differs in structure from zooecial walls and bears strong resemblances to that of the basal walls in many stenolaematous bryozoans, in that both primary and secondary tissue is present (Tavener-Smith and Williams 1972, TAVENER-SMITH; RHABDOMESOID BRYOZOANS 163 p. 149, pi. 28, figs. 182-184). The former is mostly seen as a series of disconnected, medially disposed lenses or pods, called ‘tubules’ by some authors (Phillips Ross 1960, p. 1063; Karklins 1969, p. 7). Superficially the appearance suggests two basal walls that were united face to face with one another, and then stretched so that the originally continuous primary layer was separated into a series of discrete lenses. However, these primary lenses show no trace of a medial partition which would have marked the position of the fused periostraca in such a case. Careful consideration of the ultrastructure makes it seem altogether more likely that the mesotheca formed within an epithelial fold drawn up from the inner side of an already partly calcified encrusting basal lamina. This being so, it follows that the periostracum was at no stage involved in the structure, which does not correspond to a doubled basal wall. A comparable case has been observed in the transverse zooecial walls of the modern cheilostome Umbonula (Tavener-Smith and Williams 1970, p. 249, fig. 34), which also contain separated lenticular pods of primary tissue. It would therefore appear that the establishment of a flat basal encrustation was an essential prerequisite for the development of a ptilodictyoid frond. If it may be assumed that the young rhabdomesoid colonies examined in this study were developmentally similar to others of the same group, it would follow that certain fundamental differences separate them from the Ptilodictyoidea. These relate to the colonial symmetry ; the presence or otherwise of an initial basal encrustation as a neces- sary astogenetic stage; the lack of structural and developmental correspondence between mesotheca and ‘axial rod’ (or internal tube wall); and, finally, the general aspect of the zooecia themselves. The ptilodictyoid attributes are such as to suggest that the group may have had close affinities with the more advanced (and strati- graphically later) fenestelloids. But it seems unlikely that the ptilodictyoids were more closely related to the rhabdomesoids than was either group to the Trepostomata, from which both may have been independently derived. Unsectioned specimens from the original collection are deposited in the Sedgwick Museum, Cambridge (SM E 20056-20069). Acknowledgement. I acknowledge with thanks a grant towards the cost of publication from the University of Natal, Durban, Research Fund. REFERENCES ASTROVA, G. G. 1964. O novem stryade Paleozoyskikh Mshanok (a new Order of palaeozoic bryozoa). Paleont. Zh. 2, 22-31. BASSLER, R. s. 1953. Treatise on Invertebrate Paleontology (ed. R. C. Moore), Part G: Bryozoa. Geol. Soc. Am. and University of Kansas Press. BORG, F. 1926. Studies on Recent cyclostomatous Bryozoa. Zool. Bidr. Upps. 10, 181-507. BROOD, K. 1972. Cyclostomatous bryozoa from the Upper Cretaceous and Danian in Scandinavia. Stock. Contr. Geol. 26, 1-464. CHEETHAM, A. H. 1971. Functional morphology and biofacies distribution of cheilostome bryozoa in the Danian stage (Paleocene) of Southern Scandinavia. Smithson. Contr. Paleobiol. 6, 1-87. KARKLINS, o. L. 1969. The cryptostome bryozoa from the Middle Ordovician Decorah Shale, Minnesota. Minn. Geol. Siirv. Spec. Pub. 6, 1-121. ILLIES, G. 1968. Multiseriale Bryozoa Cyclostomata mit gewolbtem Zweigquerschnitt aus dem Dogger des Oberrheingebietes. Oberrhein. geol. Abh. 17, 217-249. 164 PALAEONTOLOGY, VOLUME 17 LAGAAiJ, R. and GAUTIER, Y. V. 1965. Bryozoan assemblages from marine sediments of the Rhone delta, France. Micropaleont. 11, 39-58. PHILLIPS ROSS, J. 1960. Larger cryptostome bryozoa of the Ordovician and Silurian, Anticosti Island, Canada— Part 1. J. Paleont. 34, 1057-1076. SCHOPE, T. J. M. 1969. Paleoecology of Ectoprocts (Bryozoans). Ibid. 43, 234-244. SMiTT, F. A. 1865. Om Flafs-Bryozoernas utveckling och fettkropar. Ofvers K. Vetensk Acad. Fork. Stockh. 22, 5-50. SMYTH, L. B. 1930. The Carboniferous rocks of Hook Head, County Wexford. Proc. Rov. Irish Acad. 39B, 523-563. STACH, L. w. 1936. Correlation of zoarial form with habitat. J. Geol. 44, 60-66. TAVENER-SMiTH, R. and WILLIAMS, A. 1970. Structure of the compensation sac in two ascophoran bryozoans. Proc. R. Soc. Load. B175, 235-254. 1972. The secretion and structure of the skeleton of living and fossil bryozoa. Phil. Trans. R. Soc. Land. B264, 97-159. R. TAVENER-SMITH Department of Geology The Queen’s University of Belfast Present address Department of Geology University of Natal King George V Avenue Durban Typescript received 12 December 1972 S. Africa A NEW GENUS OE JURASSIC BIVALVE MOLLUSC ANCESTRAL TO GLOBOCARDIUM by C. P. PALMER Abstract. A new genus of bivalve mollusc Cryptocardia is described and assigned to the subfamily Protocardiinae. Four new species are described from the Domerian, Bajocian, and Bathonian stages of Europe, and from the Callovian stage of Africa. This Jurassic genus, characterized by the presence of posterior internal radial ridges and no external radial ornament, is ancestral to the Cretaceous genus Glohocardium. A tendency, throughout the Mesozoic, to sup- press the internal ridges and to increase the size of shell and coarseness, both of external ornament and internal hinge structure, is established as an evolutionary trend for the two genera. There is no material from the Upper Jurassic or the Caenozoic. A homeomorphous Cretaceous species of Protocardia is described, which departed from the usual subtrigonal outline of Mesozoic Protocardia and converged towards that of Glohocardium. Study of bivalves from the French Jurassic, in the British Museum (Natural History), revealed a number of heterodont bivalves that were difficult to place in any existing genus. The difficulty was caused by a unique feature seen on the internal surface of the posterior part of the shell. Beneath the thin semi-transparent shell were two lines, one in each valve, running from the posterior side of the beak, passing about 5 mm anterior to the posterior adductor muscle scar, and terminating at the postero-ventral margin (PL 18, fig. la, c). The lines were caused by the presence of three ridges, one in the left and two in the right valve, which left corresponding grooves on the surface of internal moulds. At the postero-ventral edge of the shell the end of the single left-hand ridge fitted between the ends of the paired right-hand ridges (PI. 18, fig. 4c). Sixteen specimens from the Bajocian and Bathonian of England and France clearly demonstrated that: {a) the posterior internal ridges were constantly present— two in the right valve and one in the left; {b) the fossils were referable to no previously described species or genus; (c) the Bathonian forms were morphologically distinct from those of the Bajocian. The sample contained only three specimens with the shell intact and the rest, being internal moulds, showed no hinge. Of the three intact shells only one could be separated and the hinge teeth developed. The distant anterior and posterior laterals and single tubercular cardinal proved the cardiid affinity of the shells and indicated that they were related to the subfamily Protocardiinae (PI. 18, fig. 2c). A search for possible ancestors in the Lower Jurassic produced only one internal mould from the Middle Lias, Marlstone Rock Bed, of Ilminster, Somerset. This, though lacking the posterior internal grooves of the Middle Jurassic forms, was clearly related to them. No other Cryptocardia have been seen from the European Jurassic, but grooves on some internal moulds collected by N. J. Morris from the Callovian of Tanzania show that these are related to the European forms. They are generally similar to Bathonian specimens from Ranville, France, but the grooves on the inner moulds are not so sharply incised and square-sectioned as in the French forms, being more rounded and shallower, and tending to fade toward the umbonal [Palaeontology, Vol. 17, Part 1, 1974, pp. 165-178, pis. 18-20.] 166 PALAEONTOLOGY, VOLUME 17 region. This obsolescence of the internal ridges heralds the more advanced condition found in the related Cretaceous genus Globocardium, in which the internal ridges are almost obliterated by secondary thickening of the shell so that only a shallow groove is produced on the inner mould for a short distance from the ventral edge of the shell. This is illustrated by Hayami (1956, pi. 16, figs. 3, 5, 6), and also on several British Globocardium sphaeroidiwn (Forbes) in which a shallow groove is seen for about one- quarter of the height of the shell, while a faint and fading track continues to the umbo. In contrast to the Jurassic forms, Globocardium bears a continuous external trace of the position of the internal ridges as a distinct flexure in the growth lines as they pass over them. This is sharply marked at the beak but fades to a broad flexure in the coarse concentric ornament toward the ventral edges (PI. 20, fig. 5). No examples of Cryptocardia higher than the Callovian in the Jurassic have been discovered, but a search for Cretaceous forms revealed that the Aptian Protocardia rotlipletzi, described by Krenkel from Tendaguru, Tanzania (Dietrich 1933, p. 51, pi. vi, figs. 89-91) was probably related either to the Jurassic forms or to the Cretaceous Globocardium. This was confirmed when a topotype was freed of matrix to reveal a typical protocardiid dentition and also the rounded knobs on the inner postero- ventral edge of the shell formed by the ends of the internal ridges (PI. 20, figs. 1, 2). Thus the small Jurassic internally ridged Cryptocardia was ancestral to Hayami’s Globocardium, type species Cardium sphaeroideum Forbes 1845, and the Tanzanian Cretaceous Protocardia rothpletzi was related to the British Globocardium sphaeroi- deum. This is confirmed by comparing Dietrich’s figures of Protocardia rothpletzi with Woods’s (1908, pi. 31, fig. 2), figures which show the presence of rugged con- centric ornament only and, on the posterior surface, a broad track marking the posi- tion of the internal ridges. The large size and coarse concentric ornament of the Cretaceous Globocardium contrasts strongly with the smaller and more delicately ornamented Jurassic forms; so they are not congeneric. Hence it is proposed that Globocardium Hayami 1956 be restricted to the larger, more coarsely ornamented Cretaceous forms; and that Cryptocardia be used for the smaller and more finely ornamented Jurassic forms. EXPLANATION OF PLATE 18 Fig. \a-c. Cryptocardia hajocensis sp. nov. Flolotype, B.M. 66193; Bajocian of St. Vigor near Bayeux (Calvados), France; purchased M. Tesson. lu, lateral view of right valve; \b, anterior view; Ic, posterior view with umbones tilted away to show external marks of internal ridges. X 1. Fig. 2a-c. Cryptocardia ranvilleiisis sp. nov. Holotype, B.M. 66203; Bathonian of Ranville (Calvados), France; purchased M. Tesson. \a, left side of inner mould, with some shell adhering, showing single posterior groove; \h, right side of inner mould showing paired posterior grooves; Ic, internal view of right valve showing cardinal tooth and anterior lateral posterior lateral; broken away. Figs, a-b x 1|, fig. c X 4. Fig. 2a-c. Cryptocardia bajocensis. Paratype, B.M. 66197; same locality as Fig. 1. 3u, left side of inner mould with single posterior groove; 36, right side of mould with some shell adhering and paired posterior grooves; 3c, anterior view. X 2. Fig. 4a-c. Cryptocardia ranvillemis. Paratype, B.M. 66243; Bathonian of Ranville (Calvados), France; O. Ward collection. 4u, right side of inner mould with paired posterior grooves; 46, left side with single groove; 4c, posterior view with umbones tilted away to show posterior grooves at postero-ventral edge, xli. PLATE 18 PALMER, Cryptocardia 168 PALAEONTOLOGY, VOLUME 17 SYSTEMATIC PALAEONTOLOGY Family cardiidae Lamarck 1809 Subfamily protocardiidae Keen 1951 Genus cryptocardia nov. Type species: Cryptocardia bajocensis sp. nov. Diagnosis. Small protocardiids with globose outline and fine concentric ornament only. Posterior internal radial ridges usually present ; two in right valve, one in left valve. Margins of valves smooth. Description. Small to medium sized for the family, height 26-37 mm; subquadrate to regularly ovate, umbones central or just anterior to the midline and beaks slightly prosogyrous. Valves without marginal denticulations, shell smooth apart from fine concentric growth lines, about 5-8 lines to 1 mm near the ventral edge. Internal posterior radial ridges, two in the right valve and one in the left. Upper Domerian to Callovian. The position of Cryptocardia within the Protocardiinae should be next to Globo- cardium Hayami 1956 and near both to Tendagurium Dietrich 1933 and Integri- cardium Rollier 1912. Tendagurium is Jurassic in age, has a more trigonal outline and is altogether longer; the Cretaceous Integricardium has less projected umbones. Both lack the posterior internal ridges of Cryptocardia and Globocardium. The chief differences between Cryptocardia and Globocardium are : Cryptocardia Shell small, up to 40 mm in height Shell thin, fine concentric ornament Hinge with delicate teeth Clear track of square-sectioned internal ridges on internal mould, only faint trace on external surface of shell. Globocardium Large, up to 100 mm in height Shell thick, coarse concentric ornament Hinge with coarse tubercular teeth Internal ridges almost obliterated by secondary thickening of the shell, distinct track on external surface of shell caused by flexure in growth lines over the internal ridges. The Bernard system of hinge notation, with modifications by subsequent authors, is used throughout this text. Cryptocardia(l) tutcheri sp. nov. Plate 19, fig. Aa-b Materiai. Holotype, an internal mould from the ‘Middle Lias of Ilminster, Somerset’, collected by J. W. Tutcher, B.M. 77330. The matrix clearly indicates that it is from the Marlstone, Spinatum Zone, probably from Moolham Larm. Diagnosis. Cryptocardia-\ike shell lacking posterior internal ridges, umbones broader, valves more equi- lateral and quadrate in outline than other described species. Description. Dimensions of holotype: length 2-6 cm, height 3-2 cm, inflation 2-5 cm. This sole specimen is placed with some hesitation in Cryptocardia since it lacks the internal ridges in the shell but in all other respects is clearly related to and possibly ancestral to the other species. It resembles C. ranvillensis in the less convex posterior outline, and C. bajocensis in its more broadly based umbones. The pallial line is entire and terminates at one end in a pyriform anterior adductor and, at the other, in a depressed region bounded by a bluntly rounded step (raised in the shell) which continues, past the elongate posterior adductor, up MESOZOIC PROTOCARDIIDAE 169 TEXT-FIG. 1. Comparative outlines of the right valves of four Jurassic species of Cryptocardia: a, C. tutcheri, Domerian of Ilminster, Somerset; b, C. bajo- censis, Bajocian of St. Vigor (Calvados), France; c, C. ranvillensis, Bathonian of Ranville (Calvados), France; d, C. morrisi, Callovian of Tanzania, East Africa. In figs, b-d the position of the paired, posterior, internal ridges are indicated by two converging lines. Scale approximately x to the beak (PI. 19, fig. 4b). On the mould this step presents the appearance of an incipient internal ridge but the step is also present in the holotype of C. ranvillensis and posterior to the internal ridges. Cryptocardia bajocensis sp. nov. Plate 18, figs, la-c, 3a-c Material. Flolotype, B.M. 66193, from the Bajocian of St. Vigor near Bayeux (Calvados), France, purchased M. Tesson 1857. Paratypes: B.M. 66243, same locality and horizon; six specimens collected by the author from a temporary road section in the Upper Inferior Oolite, Parkinsoni Zone, at Harford Bridge, 1 mile ESE of Naunton, Oxfordshire, B.M. LL. 31290-31295, also B.M. LL. 31251-31253. Diagnosis. Ovate Cryptocardia with sharp internal ridges, valves less equilateral, and umbones smaller than C. tutcheri, umbones longer and ornament finer than in C. ranuillensis. Description. Dimensions of holotype: length 3 cm, height 4-7 cm, inflation 2-8 cm. Smooth, ovate, slightly prosogyrous; beaks on the midline and rising 0-5 cm above the hinge line in the holotype; dorsal margin projected posteriorly, rather like Inoceramus; anterior, ventral, and posterior margins in a continuous ovoid curve with the greatest curvature at the antero-ventral edge but flattening along the dorsal margin— the hinge margin. Shell wholly lacking in radial ornament but covered with fine concentric growth lines which average five per 1 mm at the ventral edge; occasional growth halts produce a slight undulation in the other- wise perfectly curved and heart-shaped outline when viewed from the anterior. Internal surface of shell with two square sectioned ridges in the right valve and one in the left; these extend from the beak to the ventral margin, the single left-hand ridge fitting between the two right-hand ridges at the postero-ventral margin where they meet. As they pass over the radial ridge the growth lines are deflected slightly in a dorsal direction and a corresponding flexure is seen in the otherwise entire pallial line where it crosses the internal ridge. The anterior muscle scars, faintly seen on the internal moulds, are subovate, while the posterior adductor scars are elongate and set in a heart-shaped depression, bounded by a slight step, in the siphonal region. Margins without crenulations. Hinge not seen. Remarks. The belief that this species might be identical with d’Orbigny’s Cardium cryptum (prod. no. 334, p. 279, ‘species subovate, very inflated, remarkable for the striations on the interior, always very prominent on the mould’) was dispelled by M. Boule’s figure (1909; 88) of a mould with less prominent umbones and radial striations over the whole surface of the mould. Boule also observed that the radial striations are spiny and C. cryptum is probably an undescribed genus. 170 PALAEONTOLOGY, VOLUME 17 Cryptocardia ranvillensis sp. nov. Plate 18, figs. 2a-c, 4a-c; Plate 19, figs, \a-c, 2a-c, 3 Material. Holotype, B.M. 66203, from the Bathonian of Ranville (Calvados), France. Paratypes: B.M. 66243, 66183, from the same locality; and two internal moulds B.M. L. 77973, 77974, collected by D. T. Donovan from the Fullers Earth of Kelston, near Keynsham, Somerset. Diagnosis. Ovate Cryptocardia with sharp internal ridges, smaller umbones, and coarser growth lines than C. bajocensis. Description. Dimensions of holotype: length 2-1 cm, height 2-7 cm, inflation 2 cm. Similar in all respects to C. bajocensis but differing in being smaller, slightly shorter, and less inflated; umbones slightly narrower and with the posterior margin less convex. The posterior adductor scars are slightly more impressed and are bounded by a more conspicuous step. The growth lines, about five per mm at the ventral edge, are regular and more sharply incised than in C. bajocensis since each growth increment terminates in a raised lip. The valves of the holotype were disarticulated and sheared, and a development of the right hinge revealed a typical cardiid hinge with a single peg-like cardinal 3b and a socket anterior to it for the recep- tion of the cardinal 2 in the left. Anterior laterals were present, 1 -4 cm apart, as small, short lamellae, curved in the anterior and parallel in the posterior. Left hinge unknown. Remarks. One of the two Keynsham specimens has a fragment of badly eroded shell adhering to it. This shows the same incised growth lines as in the holotype. Cryptocardia morrisi sp. nov. Plate 19, figs. 5-9 Material. Holotype, B.M. LL. 31285, right valve. Paratypes: B.M. LL. 31286-31289, from the Callovian of Tanzania. Diagnosis. Ovate Cryptocardia with rounded posterior internal ridges, and with smaller umbones and a more convex posterior outline than other species of Cryptocardia. EXPLANATION OF PLATE 19 Figs. 1-3. Cryptocardia ranvillensis. Fig. 1. B.M. 66183; Bathonian of Ranville (Calvados), France; purchased M. Tesson. la, left side with single posterior groove; \b, right side of mould with paired posterior grooves ; 1 c, posterior view, tilted away to show grooves at postero-ventral edge, x 1 f . Fig. 2. B.M. L. 77973; Bathonian, Fullers Earth, at Kelston Pound Hill, near Keynsham, Somerset; collected D. T. Donovan. 2a, left side of eroded inner mould with single groove; 2b, right side with terminal ends of posterior grooves at postero-ventral edge; 2c, posterior view with umbones tilted away to show grooves at postero-ventral edge, x I f. Pig. 3. B.M. L. 77974; same locality and horizon as Pig. 2. Left side of inner mould showing pallial line and slight flexure where the posterior groove crosses it. x 1^. Pig. 4. CryptocardiaC) tutcheri sp. nov. Holotype, B.M. L. 77330; Domerian, Spinatum Zone, of Ilminster (probably Moolham’s farm), Somerset; collected J. W. Tutcher. 4a, right side of eroded inner mould; 46, posterior view showing impressed (raised on shell) posterior adductor scars, x 1. Pigs. 5-9. Cryptocardia morrisi sp. nov. Pig. 5. Holotype, B.M. LL. 31285; Middle Callovian, Manyuli stream section, Mendawa-Mahokondo anticline, Kiswere area, Tanzania. External view of right valve with trace of posterior internal ridges visible at surface of shell. X 1-L Pig. 6. Paratype, B.M. LL. 31286. Left side of inner mould showing faint and shallow posterior groove and distinct anterior adductor scar. X 1|. Pig. 7. B.M. LL. 31287. Left side of inner mould with some shell still adhering and posterior ridge visible at shell surface, x 1^. Fig. 8. B.M. LL. 31 288. Right side of inner mould with shell adhering and paired, shallow posterior grooves near postero-ventral edge, x I '. Fig. 9. B.M. LL. 31289; Longi stream section, Mendawa-Mahokondo anticline, Kiswere area, Tanzania. Left side of inner mould with shell adhering and showing distinct anterior adductor scar. X I f. PLATE 19 PALMER, Cryptocardia 172 PALAEONTOLOGY, VOLUME 17 Description. Dimensions of holotype: length 21 cm, height 2-5 cm, inflation 0-9 cm, single valve only. Smooth, ornate, slightly prosogyrous; beaks on the midline. Umbone elevated, bluntly rounded forming a right angle with slightly concave anterior and posterio-dorsal margins which merge into the nearly semicircular outline of the ventral margin, giving the shell a distinctly ‘tear-drop’ appearance. The posterior margin is straighter than the anterior which projects in a round curve. The surface of the shell is covered with fine concentric growth lines, about eight per mm at 2-2 mm height. Hinge could not be developed; however a cast taken from one of the moulds showed the hinge to be virtually identical with that of C. ranvillensis. Remarks. This species closely resembles C. ranvillensis, differing in its reduced height/length ratio, its more closely spaced ribs (which are around 8 to 1 mm compared with 5 to 1 mm at a height of 2-2 cm in C. ranvillensis), and its more projected and rounded anterior margin. At one horizon in the Callovian of Tanzania it is common, since several valves are present in pieces of rock no more than 4 cm. long. TEXT-FIG. 2. Comparative outlines of the right valves of two Cretaceous species of Globocardium: a, G. rothpletzi from the Aptian of Tanzania; b, G. sphaeroi- dium from the Aptian of the Isle of Wight, Hampshire. The broad track on the exterior of the shell marking the position of the two internal ridges is indicated by two diverging lines. Scale approximately x Genus globocardium Hayami 1956 Type species. Cardium sphaeroideiim Forbes 1845, Aptian and Albian of Western Europe and Aptian of Japan. Diagnosis. Large protocardiid with globose outline and coarse concentric ornament and track of posterior internal ridges which, except for the terminal ends, are intern- ally obsolescent. Description. ‘Strongly inflated Protocardia having globose shell, widely spaced con- centric costae on the disc and a nearly smooth posterior area without any con- spicuous radial ribs ; posterior carina weak or absent ; ventral margin smooth internal ; hinge similar to that of Protocardia' (Hayami 1956, p. 116). His diagnosis requires the addition of three internal degenerate ridges in the posterior of the shell, two in the right valve and one in the left. The external concentric ornament is modified over the internal ridges forming a continuous swollen track from the umbone to the ventral edge. The internal margin ventral margin though lacking denticulations contains the interdigitating ends of the internal ridges and forms a pronounced sinus in the ventral commissure. MESOZOIC PROTOCARDIIDAE 173 G. imbricataria was recognized as a Glohocardium by Hayami (1956, p. 116). G. diipiniaimm (d’Orbigny 1844, pi. 242, figs. 1-3) lacks radial ornament but also (on the drawing) any trace of an external track to mark the position of the internal ridges. The hinge of d’Orbigny’s fig. 3 is comparable with that of G. sphaeroideum although the tooth 3a is more strongly developed in the French examples. Glohocardium sphaeroideum (Forbes 1845) Plate 20, figs. 3-5 Cardiwn sphaeroideum E. Forbes 1845: p. 243, pi. 2, fig. 8. Protocardia sphaeroidea (Forbes) Woods 1908: p. 195, pi. 31, figs. 2, 3. Cardiimt neckerianum Pictet & Roux 1852: p. 424, pi. 30, fig. 3. Protocardia (Glohocardium) sphaeroidea (Forbes) Hayami 1956: p. 117, pi. 16, figs. 16, with synonymy. TEXT-FIG. 3. Dorsal view of G. sphaeroidium, viewed down the plane of the commissure, to show the articulation of the hinge teeth. The beaks are ‘cut-away’ to expose the position of the cardinal teeth. Solid black teeth are those projecting beyond the plane of the commissure. Dotted lines represent position of lateral teeth within the plane of the commissure. The solid black laterals All and PII articulate above, that is dorsal to, the dotted laterals AI and AIF The obsolescent cardinals 3a and 4b are omitted for clarity. Material. Eighteen specimens from the Aptian, Fower Greensand Perna Bed of Atherfield, Isle of Wight; Casey (1961, p. 497, Table 1) placed the Perna Bed in the Flssicostatus Zone of the Fower Aptian. Also one internal and one external mould from the Upper Greensand of Haldon, Devon, and three external moulds from the Upper Greensand of Wiltshire. Description. ‘Shell inflated, rather oblong and angulated posteriorly. The surface is marked by deep and regular transverse sulcations, which are cut off from the somewhat truncated side by a deep longitudinal furrow. The sulcations on a specimen of the above dimensions are about eighty in number. The shell is thick especially at the margins. The beak is very prominent. The cast is smooth’ (Forbes 1845, p. 243). The ‘deep longitudinal furrow’ is the track of the internal ridges which is partly eliminated internally by shell thickening except for a short distance from the shell margin where the terminal ends of the ridges produce three prominent rounded projections. Woods’s figured specimen (1908, pi. 31, figs. 2, 3) is a well- preserved articulated two-valved shell which shows an S -shaped sinus in the ventral commissure where the internal ridges terminate (PI. 20, fig. 5). Remarks. This species, though related to Cryptocardia, is larger than the Jurassic forms. The average height is 8 cm and ranges up to 10 cm. The concentric ribs, here much coarser than Cryptocardia, continue on to the siphonal area without radial ornament, other than the track produced by the internal ridges. Left hinge of: a strong tuberculate anterior lateral All; a strong cardinal 2 and a weak 4b; and an elongated posterior lateral PII apparently continuous with the ligamental nymph. Right hinge of: a deeply sunk, broad, anterior lateral AI (socket for the tuberculate All of the left valve); a weak cardinal 3a and a strong tuberculate 3b; and a posterior lateral PI, less deeply sunk than the anterior (forming a socket for PII in the left valve). This hinge is the same as G. rothpletzi apart from its greater size, and differs from the Crypto- cardia only in the more robust form of the anterior lateral All, to judge by the relative size and depth of the anterior lateral AI in C. ranvillensis (PI. 18, fig. 2c). Following Woods (1908, p. 195), Cardiwn neckerianum of Pictet and Roux 1852 is placed in synonymy with this species (see also Hayami 1956, p. 116). Their pi. 30, fig. 3 shows a right valve, whose outline falls 174 PALAEONTOLOGY, VOLUME 17 within the variation of specimens from the Lower Greensand of the Isle of Wight. However, there is no trace on their fig. 3 of any external track to mark the position of the internal ridge. Forms indistinguishable from G. sphaeroidewn are found in the Upper Albian Upper Greensand at Devizes, Wiltshire, and at Shaftesbury in Dorset. In the Upper Greensand of the Haldon Hills, Devon, slightly larger specimens occur, all showing the characteristic outline of G. sphaeroideum, the same density of ribbing, and the broad external track marking the position of the internal ridges. There are no records of Globocardium from the Gault Clay between the Upper and Lower Greensand, indicating the preference of G. sphaeroideum for coarser sedi- ment and higher-energy environments. Globocardium rothpletzi (Krenkel 1910) Plate 20. figs. 1-2 Protocardia rothpletzi Krenkel 1910: p. 216, pi. 21, fig- L Cardium (Tendagurium) rothpletzi (Krenkel) Dietrich 1933: p. 51, pi. 6, figs. 89-91. Material. One right value, B.M. LL. 31255, collected by N. J. Morris from the Lower Aptian, Trigonia Schwarzi Beds of Tanzania, and one left value collected by J. Parkinson, B.M. L. 51837. Description. Similar in all respects to G. sphaeroideum but differing in the finer ornament which at 5-5 cm height shows seven concentric undulations per centimetre compared with only five in G. sphaeroideum at the equivalent height. The siphonal area is smoother and the outline slightly more quadrate than the English form, and the maximum height does not appear to exceed 6 cm. Remarks. Dietrich placed P. rothpletziin Tendagurium, type species Cardium ( Tendagurium) propebanneianum Dietrich 1933 (p. 50, pi. 6, figs. 92, 93). His fig. 92 is a right valve with a straight hinge line, long laterals, and small cardinals. His fig. 93 is of an internal mould with prominently raised muscle platforms, a sharply angled carina bounding the siphonal area, and a pallial line with a moderately deep sinus in it. The internal mould of a left valve shows no trace of a groove representing internal ridges. By these characters Tenda- gurium differs from Globocardium and justifies the removal of P. rothpletzi from Tendagurium and placing it in Globocardium. This view was also expressed by Hayami (1956, p. 117) but for slightly different reasons. The left hinge (PI. 20, fig. 1) has two unequal cardinals, a strong 2 and a weak 4b, a stout, conical anterior lateral All, and an elongated, smaller posterior lateral PII. The right hinge (PI. 20, fig. 2) has two cardinals, a strong 3b, and a weak 3a fused to the lunular margin; two laterals the anterior of which AI forms a broad, deep socket for the reception of the short conical anterior lateral tooth of the left valve; the posterior lateral PI forms a much narrower socket for the posterior lateral in the left valve. The hinge of this species is the same as Cryptocardia ranvillensis and Globocardium sphaeroideum. EXPLANATION OF PLATE 20 Figs. 1, 2. Globocardium rothpletzi (Krenkel). Fig. 1. B.M. L. 51834; Lower Aptian, north of Mbemkuru River, Niongala, Tendaguru, Tanzania. Internal view of hinge of left valve cleared of matrix and terminal knob (arrow) at postero-ventral edge, x 1. Fig. 2. B.M. LL. 31255; Lower Aptian, Trigonia Schwarzi Beds at stream SE. of Nossa Stream, N. of Mbemkuru River (W. 39°, 14', 40"; S. 09°, 35'), Tanzania. Internal view of right valve showing hinge cleared of matrix and paired terminal knobs (arrow) at postero- ventral edge. xl. Figs. 3-5. Globocardium sphaeroideum (Forbes). Lower Aptian, Perna Bed, Culver Cliff, Sandown, Isle of Wight. Fig. 3. B.M. LL. 8466. Internal view of left valve with matrix cleared showing hinge and single terminal knob (arrow) at postero-ventral edge. xf. Fig. 4. B.M. 48626. Internal view of right valve showing hinge cleared of matrix and paired knobs (arrow) at postero-ventral edge. X f. Fig. 5. B.M. L. 8247; Beckles collection. Ventral view of the original of Woods 1908, pi. 31, fig- 2, showing external track and terminal knobs at postero-ventral commissure, xf. Fig. 6. Protocardia vicaryi sp. nov. Holotype, B.M. L. 17041 ; Uppermost Upper Albian (or lowest Ceno- manian), Haldon Hills, Devon ; W. Vicary collection. External view of right valve, the original of Woods 1908, pi. 31, fig. 4, slightly reduced. PLATE 20 PALMER, Cryptocardia 176 PALAEONTOLOGY, VOLUME 17 Genus protocardia von Beyrich 1845 Type species. Cardium hillanum J. Sowerby 1813. In dealing with the Cretaceous species of Globocardium it became apparent that an undescribed protocardiid, previously figured by Woods but not named, had been placed by Hayami in synonymy with G. sphaeroideum. Since this deceptive homeo- morph needed to be removed from Globocardium and placed in Protocardia, it seemed appropriate to describe and name it here. Protocardia vicaryi sp. nov. Plate 20, fig. 6 Diagnosis. Protocardia with height : length ratio about 1 30%, greater than all other species of Protocardia. Material. Holotype, B.M. L. 17041, from the Upper Cretaceous, probably Upper Albian, Haldon Hills, Devon; the original of Woods’s (1908), pi. 31, fig. 4. Paratype, B.M. LL. 31296, same locality. Both speci- mens collected by W. Vicary. Description. Dimensions of holotype: length 4-7 cm, height 6-2 cm, inflation 3-4 cm. Shell smooth, ovate, slightly prosogyrous; beaks in front of the midline and rising about 5 mm above the somewhat narrow hinge line. The anterior and ventral margins form a continuous and regular curve as far as the postero- ventral edge where it changes direction in a bluntly rounded curve and continues to the hinge line in a vertical, slightly concave, line. The valves are more noticeably inequilateral than species of Cryptocardia and Globocardium, and are covered with fine concentric growth lines and occasional growth halts. Faint radial striations are visible on the siphonal area where a slight fold makes a bluntly rounded ridge on the internal surface of the shell. The last deceptively suggests that it is a species of Globocardium but the external radial ornament, crenulate margin, and markedly inequilateral aspect shows that it is not. The hinge of the holotype, though partly obscured by matrix, is apparently similar to that of Globocardium and Protocardia, but other internal features are unknown. Remarks. This Upper Greensand protocardiid bears a superficial resemblance to G. sphaeroideum, but dilfers in that it has narrower umbones, the outline is more oblique, it displays faint radial ornament on the siphonal area, it lacks internal ridges, it has a nearly smooth shell with only fine growth lines, and the siphonal area is sharply concave and undulatory. The nearest described species to P. vicaryi is P. guerangeri (d’Orbigny 1844, p. 35, pi. 249, figs. 3, 4) from the Cretaceous of France, but it differs in its more prosogyrous beaks, smaller umbones, more rounded umbonal ridge, and its generally larger size. Most species of Protocardia have height : length ratios which are usually around 100%. However P. vicaryi and P. guerangeri have ratios closer to 130% and are noticeably taller than the majority oi Protocardia. Hayami (1956, p. 1 16) included P. vicaryi in his genus Globocardium, a position which cannot be upheld here since: (u) there is neither a trace of the internal ridges nor track on the external ornament; (b) there is faint external radial ornament on the siphonal area of the paratype and also a trace of it on the holotype; (c) the margin is finely crenulated. The siphonal area is demarcated by a bluntly rounded umbonal ridge which passes into a sharp concave fold producing a thickened ridge visible on the internal surface of the shell of the holotype. The paratype reveals that this ridge lies posterior to the adductor scar and is therefore not comparable with the internal ridges of Cryptocardia and Globocardium, in which the ridges lie anterior to the posterior adductor scar. The age of the type locality is either highest Upper Albian or basal Cenomanian— the matter is not yet settled by Cretaceous stratigraphers. DISCUSSION AND CONCLUSIONS Function of the internal ridges. No satisfactory explanation can be offered for these puzzling features. The simple hypothesis, that the terminal ends of the internal ridges provided a shear-resistant mechanism, may be refuted by the following: (a) shear stress is greatest when a bivalve is actively burrowing, for then the valves are neces- MESOZOIC PROTOCARDIIDAE 177 sarily open and the ends of the internal ridges cannot then interlock; (b) conversely when the bivalve is not actively burrowing the ends of the internal ridges interlock but shear stress is no longer present. To postulate that this feature is a response to predation is unhelpful unless a predator employing a shear action on bivalves can be identified ^starfishes do not qualify for this role. Palaeoecology. The Domerian and Bajocian species of Cryptocardia occur in oolitic limestone, the Callovian in fine sandstone, while the Cretaceous species of Globo- cardium occur in coarse gritty sandstone. All are associated with an abundant benthic fauna dominated by mollusca. The increase in coarseness of sediment, together with the presence of an abundant associated benthos, is taken as evidence that these two genera favoured a high-energy environment which increased during their evolution. The massive lateral teeth in Globocardium, and their prominence in Cryptocardia, indicate that these were active borrowers in coarse sediments. It further points to a probable connection between the increasing oxygen consumption and demands of the progressively more active bivalves and the greater availability of oxygen in the higher energy facies. Evolution. The development during Mesozoic times of the two genera Cryptocardia and Globocardium may be summarized as follows: {a) increase in size and massive- ness of shell, particularly above the Jurassic/Cretaceous boundary; (b) a coarsening of hinge characters and external ornament with increase of size — particularly in Globocardium-, (c) suppression of the square-sectioned internal ridges by an increase of secondary thickening of the internal part of the shell; {d) a gradual transference of the internal ridges from that of an internal feature of the shell in Cryptocardia to that of an external track, formed by a flexure in the growth lines, in Globocar- dium. The known range of Cryptocardia is from the uppermost Domerian to the Callovian stage of the Jurassic. No examples from the Oxfordian, Kimmeridgian, and Port- landian stages of the Jurassic, and the Neocomian stage of the Cretaceous, have come to the writer’s attention. Since Cryptocardia was probably ancestral to Globocardium it follows that intermediate forms existed during Oxfordian to Neocomian times. These, being stratigraphically intermediate between the Callovian C. morrisi and the Aptian G. rothpletzi, would have shell characters intermediate between these two forms. No examples of Globocardium younger than Upper Albian (or Lower Cenomanian) are known to the writer, and a search for possible descendants in the Caenozoic proved fruitless. Any descendants probably continued the trend towards suppression of the internal ridges, and a steady increase in shell size, coupled with a coarsening of the concentric ornament and of the hinge characters. It is probable that the Jurassic Cryptocardia was ancestral to the Cretaceous Globocardium. Cryptocardia first appeared as an offshoot of the Protocardia group in uppermost Domerian times. By the Bajocian it had lost all trace of external radial ornament and had developed unique posterior internal ridges — two in the right and one in the left valve. During Callovian times the internal ridges became less pro- nounced and this obsolescence, together with an increase in over-all shell size, led to the Cretaceous Globocardium. This evolutionary series of protocardiids tended to 178 PALAEONTOLOGY, VOLUME 17 occupy increasingly higher energy condition through their history. During mid- Cretaceous times a Protocardia vicaryi, with faint posterior radial ornament, departed from the usual subtrigonal outline of Mesozoic protocardiids and converged towards the elongated and globose shape of the CryptocardialGlobocardium group. Acknowledgements. I would like to thank Mr. C. B. Keates, of the British Museum (Nat. Hist.), for suc- cessfully handling the difficult photography. REFERENCES AiTKEN, w. G. 1961. Geology and Palaeontology of the Jurassic and Cretaceous of southern Tanganyika. Bulletin of the Geological Survey of Tanganyika, 31, vi+ 144 pp., 14 pis. BOULE, M. 1909. Types du Prodrome paleontologie stratigraphique universelle de d’Orbigny, I. Ann. Palaeont. 4 (4), 153-164, pi. 19. CASEY, R. c. 1961. The Stratigraphical Palaeontology of the Lower Greensand. Palaeontology, 3, 487-628, pis. 77-84. cox, L. R. et al. 1969. Treatise on Invertebrate Paleontology. Part N. Volume 2 (of 3), Mollusca 6, Bivalvia, Kansas. DIETRICH, w. o. 1933. Zur Stratigraphie und Palaeontologie der Tendaguruschichten. Palaeontographica, Suppl. 7, 2, 1-86, pis. 1-12. HAYAMi, I. 1956. Lower Cretaceous Marine Pelecypods of Japan, Part II. Mem. Fac. Sci. Kvushu Univ. Ser. D Geology, 17, 73-150, pis. 1-21. LEYMERiE, A. 1842. Memoire sur le terrain Cretace du Department de I’Aube, contenant des considerations generates sur le terrain Neocomien. Mem. Soc. geol. France, 5, 1-34, pis. 1-18. d’orbigny, a. 1844-1847. Palaeontologie Francaise. Terraines Cretach. 3, Lamellibranches, 807 pp., pis. 237-489. PICTET, F. J. and ROUX, w. 1849-1854. Description des Mollusques Fossiles qui se trouvant dans les gres verts des environs de Geneve. Mem. Soc. phys. et hist. nat. Geneve, 11, 12, and 13. ROLLiER, L. 1911-1918. Fossiles nouveaux ou peu connu des terrains secondaires (Mesozoiques) du Jura et des contrees environmentes. Abh. schweiz. palaeont. 37-44, 696-( 101 pp., 49 pis. WOODS, H. 1908. Cretaceous Lamellibranchia of England. Palaeont. Soc. Mongr. London, 2 (5), 181-216, pis. 28-34. C. P. PALMER 21 Upton Dene Grange Road Sutton, Surrey Typescript submitted 30 December 1972 SHELL STRUCTURE OF TEREBRATULID BRACHIOPODS by T). I. MACKINNON A. WILLIAMS Abstract. Ultrastructural studies of living and fossil brachiopods belonging to the Terebratulidae show that the shell is invariably penetrated by unbranched puncta and generally consists of a three-fold mineral succession : a primary layer composed of acicular and granular crystallites, a secondary layer of orthodoxly stacked fibres, and a prismatic tertiary layer. The constituents of the last layer are not ‘prisms' in the crystallographic sense but discrete units with interlocking boundaries arranged normal to the surface of accretion. Each prism arises from a fibre first by lateral spreading of the terminal face and then by vertical accretion so that growth banding no longer indicates an oblique increase in length as in fibres. Sections of decalcified mantle show that the epithelium secreting the tertiary layer is like that underlying the secondary shell except that it does not exude proteinous sheets between prisms. Amalgama- tion of prisms may, therefore, be inhibited by either sheets of water-soluble organic compounds or crystallographic incompatibility between adjacent prisms as indicated by non-alignment of cleavage. The tertiary layer was not developed in Lobothyris, the oldest terebratulid studied, so that the layer was possibly first acquired geronto- morphically. But it is also lacking in the shell of other genera like Rhombothyris and Terebratula and this may reflect repeated neotenous suppression. Despite the uniform texture of the tertiary layer, expansion of muscle bases across overlying epithelium causes different types of myotests to develop in Liothyrella and Gryphus as in other living articulates. The characteristic well-ordered pits in Liothyrella, in particular, suggest that the disposition and strength of muscle fibres contribute to the shaping of myotest topography. The three-fold skeletal succession characteristic of most living articulate brachiopods was first recognized by King in 1871. Carpenter had previously identified an external organic cover, named the periostracum by him, and an underlying calcareous layer which he described (1853, pp. 25-26) as consisting exclusively of ‘flattened prisms, of considerable length, arranged parallel to each other with great regularity and at a very acute angle— usually only about 10° or 12°— with the surface of the shell’. In fact an equally distinctive thin compact layer of cryptocrystalline calcite, referred to as the laminar layer by its discoverer King (1871, p. 441), intervenes between the periostracum and the very much thicker succession of so-called calcite prisms. The ‘laminar’ and ‘prismatic’ layers are now distinguished as ‘primary’ and ‘secondary’ respectively (Williams 1956, p. 246; 1968, p. 2) to indicate their relative order of deposition within the standard secretory regime of articulate brachiopods. Indeed the terms laminar and prismatic are even inappropriate as descriptions of the domi- nant fabric of each layer. The primary layer consists of crystallites usually aligned normal to the external surface, and any growth banding developed does not form discrete laminae. Similarly, the calcite components of the secondary shell are not prisms in the crystallographic sense but fibres sheathed in interconnected proteinous membrances and tightly stacked in regular alternating rows. Hobbs and Cloud (in Cloud 1942, p. 24) have shown that the optical c-axis of calcite in crushed fragments of the secondary shell makes an angle of about 25° with the long axis of constituent fibres. This arrangement is consistent with a c-axis orientation more or less normal to the terminal growth face of a fibre and, therefore, to any growth banding seen in sagittal sections of fibres. In addition to the finely crystalline and fibrous textures of the carbonate shell successions of non-strophomenide articulates, a third fabric is [Palaeontology, Vol. 17, Part 1, 1974, pp. 179-202, pis. 21-27.] 180 PALAEONTOLOGY, VOLUME 17 fairly commonly developed. So far as we are aware, it was first explicitly identified as being characteristic of a third layer succeeding the ‘inner and outer shell layers’ by Alexander (1948, p. 48) in her study of Silurian pentameraceans. Like many others authors including St. Joseph (1938, p. 241) and Amsden (1964, p. 222), Alexander referred to this fabric as prismatic, presumably because the mineral constituents are ‘prisms having their long axes perpendicular to the shell surface’ (Gauri and Boucot 1968, p. 87). In this context ‘prismatic’ is the most appropriate term available to describe the texture of the layer although few of the lateral walls of crystallites are demonstrably {1010} faces. Prismatic shell is widely distributed among extinct brachiopods being especially characteristic of pentamerides and spiriferides. In skeletal successions of species belonging to these two orders, prisms may have developed as impersistent lenses as well as a distinct tertiary layer. Under the optical microscope the fibrous and prismatic layers are seen to pass into one another without the intervention of any sharply defined interface (Williams and Rowell in Williams et al. 1965, p. H64). But the nature of the passage remained tantalizingly obscure until one of us (MacKinnon 1971) observed under the scanning electron microscope the exact relationship between fibres and prisms in the shells of the Triassic spiriferide Koninckina leonhardi (Wissman) and the living terebratulide Gryphus vitreus (Born). The occurrence of a prismatic tertiary layer in the latter genus which was first reported by Sass and Munroe (1967, p. 302), was especially significant because it indicated that it would be possible to determine the type of outer epithelium responsible for the secretion of prismatic calcite. The specimens of Gryphus used in that investigation had been pre- served in alcohol for more than 50 years and sections of decalcified shell and mantle were not well enough preserved for study under a transmission electron microscope. Fortunately, however, the more easily obtainable related species, Liothyrella neo- zelanica Thomson, also proved to have a prismatic layer in its shell, and living specimens have been successfully prepared for the study of tissue ultrastructure. Finally, although this paper is not intended to be a comprehensive survey of prismatic calcite occurring in the shell of articulate brachiopods, we thought it appropriate to ascertain whether it is found in other Terebratulidae. It is not antici- pated, however, that the development of prismatic shell among terebratulides is restricted to members of that family especially when one considers the size and thick- ness of many centronellidine shells. Materials and Methods. Specimens used in the researches described below were obtained in the living, dried, or fossilized state and have been variously treated in the following manner. For study of the mantle under the transmission electron microscope, living specimens were fixed in 4% gluteraldehyde made up in 3% sodium chloride, then decalcified in 10% EDTA, washed in sucrose, and treated for 1 hour with 2% osmic acid; all solutions were butfercd to pH 7-2 with phosphate buffer. Following dehydration, specimens were embedded in an Epoxy resin and the microtomed sections were stained with aqueous uranyl acetate and aqueous lead citrate. All shells of living and fossil species were studied under a Stereoscan electron microscope (NERC grant GR/3/443). For this purpose the valves of living and dried specimens were freed of soft parts by immersion in sodium hypochlorite for some hours. Next, any loosely adherent particles were removed from natural and fracture surfaces by brief sonication in a weak detergent and then in acetone. Many fossils examined were embedded in rock matrix and details of their skeletal successions are best seen in cut sections. Consequently, Recent shells have also been sectioned in the same way to ensure that details of fossil successions have been correctly identified. For this purpose, both fossils in matrix and free shells were embedded in a resin, like Araldite, and cut with a diamond-edged blade along any preferred direction MACKINNON AND WILLIAMS: TEREBRATULID SHELL 181 to provide surfaces containing the required sections. These were then polished with tin oxide or alumina and etched in 2% EDTA for up to 30 minutes, dependent on the texture and geological age of the shell, to accentuate topographic differences between various microscopic features. All natural surfaces, fracture surfaces, and differentially etched sections were coated with gold/palladium for examination under the Stereoscan. STRUCTURE OE LIOTHYRELLA SHELL Mantle and caeca. The mantle of Liothyrella neozelanica Thomson, with its densely distributed microscopic cylindroid outgrowths (caeca) accommodated by canals (puncta) permeating the calcareous shell and its meshwork of spicules, is essentially like that of the terebratulacean Terehratulina retusa (Linnaeus) (Williams 1968, p. 279). The mantle edge was not preserved in its entirety in the sections examined because decalcification caused the periostracum, which has not been seen, to pull away from, and rupture, the outer mantle lobe. However, enough of the outer mantle lobe survives to show that it is composed of elongate secretory cells, with large nuclei and many secretion droplets in various stages of depletion, representing immature outer epithelium. These cells are separated from the ciliated and microvillous inner epithelium by a narrow generative zone, more or less coinciding with the closure of the mantle groove, from which both kinds of epithelia are proliferated. Thus as each newly generated cell rotates around the outer mantle lobe to become part of the outer epithelium lining a valve, it successively secretes periostracum, primary crystallites, secondary fibres, and tertiary prisms. This sequence represents the secretory regime of Liothyrella. Consequently, any differences between epithelial cells associated with the various carbonate layers at the moment of death may be regarded as indicating the kind of changes each cell undergoes as it becomes increasingly distant from the radially expanding mantle edge. The unbranched caeca of Liothyrella are similar to those found in Waltonia incon- spicua (Sowerby) and evidently originate in the same way by differentiation of groups of migrating cells at the mantle edge (Owen and Williams 1969). The core cells are well developed as pendent clusters surrounded by flattened peripheral cells which form the caecal wall and are continuous basally with the outer epithelium of the mantle. The core cells are storage centres being charged with lipids and glycogen rosettes and especially spheroidal membrane-bound secretion droplets of protein up to 1 jum in diameter and much larger ellipsoidal electron-light bodies of mucoprotein without confining membranes (PI. 21, fig. 2). The distal plasmalemmas of the core cells are prolonged into microvilli, up to 180 nm in diameter, which are small enough to have filled a series of canals penetrating a canopy of primary calcite, about 4 ixm thick, between the caecum and the periostracum. The canopy canals are lined with the same electron-dense proteinous membrane, about 7 nm thick, separating the peripheral flattened cells of a caecum from the walls of its punctum. Within the canals the membrane forms polysaccharide-filled prolongations (the brush) (PI. 21, fig. 1). Elsewhere the membrane, which is attached to peripheral cells by numerous desmo- somes, probably inhibits further carbonate deposition within the punctum (PI. 21, fig. 3). Caeca are accommodated by puncta about 10 ixm in diameter proximally but widening to about 15 [xm towards the distal termination. Since they normally remain functional throughout the life of the animal, they provide reliable checks on the extent of various shell layers in decalcified sections. 182 PALAEONTOLOGY, VOLUME 17 Primary and secondary layers. Cells secreting the primary layer are horizontally elongate and highly vesicular with much glycogen in the groundmass and glyco- protein residues in the secretion droplets (PI. 21, fig. 4). Secretory plasmalemmas are less coarsely microvillous than those responsible for the exudation of the perio- stracum. Another distinctive feature of these cells is the presence of a discontinuous layer of electron-dense particles, no more than a few nanometres in size, disposed at about 10 nm external to the plasmalemmas. The primary layer is only about 30 ju.m thick and forms a rim at the edge of each valve up to 30 yim wide (PI. 22, fig. 4). Judging from the occurrence of fine banding with a periodicity of about 250 nm, the growth surfaces, on which the outer epithelial cells rest (the superficial synchronous boundary), are inclined at about 25° to the isotopic secondary junction (PI. 23, fig. 1). The layer is composed of solidly compacted acicular crystallites about 1 /xm wide and up to 15 ixm long arranged more or less normal to the growth surface. These crystallites impart a granular texture to the growth surface except where groups of them amalgamate to form cleaved nodules 3 /xm or more in size (PI. 22, fig. 5). Inwardly towards the primary-secondary junction the epithelial cells change in structure and topography as they begin simultaneously to secrete protein as well as calcite and to accommodate arrays of secondary fibres. The cells become flattened so that they do not stand much more than a micron above the basal lamina except between fibres where narrow zones may extend into the secondary shell for 2 or more microns (PI. 21, fig. 5). The chief internal differences include : more numerous mito- chondria, a reappearance of abundant glycoprotein droplets, a tendency for vesicles to become fewer but large, and especially the development of tonofibrils. The tono- fibrils normally extend from the basal lamina to the secretory plasmalemma. They are densely distributed in the inter-fibre zones where they pass into desmosomes connect- ing the plasmalemma to an external proteinous sheet about 20 nm distant. Each proteinous sheet, which is a triple-unit membrane about 7 nm thick, is continuous with those secreted by adjacent cells. They are, therefore, secreted as complete wrappings to fibres except on the terminal faces of fibres where the plasmalemma is filamentous but otherwise free of an external organic coat. The fibres are essentially like those found in the secondary shell of most articulate EXPLANATION OF PLATE 21 Figs. 1-5. Transmission electron micrographs of decalcified shell and mantle of Liotliyrella neozekmica Thomson; Recent, Farewell Spit, S. Island, New Zealand. 1. Lateral longitudinal section through the distal part of a caecum showing the brush towards the top and flattened peripheral cells with conspicuous proteinous secretion droplets below (x5150). 2. Sagittal section through a caecum at the junction of the primary and secondary carbonate layers indicated respectively by the absence (top left) or presence (bottom left) of interconnected proteinous sheets; flattened peripheral cells containing pro- teinous droplets occur in the middle and core cells filled with elliptical mucoproteinous inclusions to the right ( X 8200). 3. Sagittal section of part of a flattened peripheral cell showing the desmosomal connections between continuous proteinous sheets and the plasmalemma (x 68750). 4. Section of part of an outer epithelial cell responsible for the secretion of the primary shell showing the discontinuous proteinous sheet associated with the plasmalemma; connective tissue with collagens (below) separated from the outer epithelium by a basal lamina ( x 55000). 5. Oblique section through the outer epithelium responsible for the secretion of the secondary shell showing interconnected proteinous sheets and their desmosomal attachments to external plasmalemmas (x 17200). PLATE 21 •m" 5 '■ v.'^: : Al\ ■' V’' ■ ‘ ’4 MACKINNON and WILLIAMS, terebratulid shell 184 PALAEONTOLOGY, VOLUME 17 brachiopods, being regularly arranged in alternating rows and having keeled and convex outer and inner surfaces respectively (PI. 23, fig. 4). There are, however, note- worthy differences. The first-formed fibres, which are usually no more than 4 jxm wide, may form loosely packed aggregates disposed at a high angle to the primary- secondary junction (PI. 22, figs. 4, 5). Their terminal faces, as well as those of some mature fibres which are normally over twice as big, may have truncated or squared rather than rounded outer boundaries (PI. 23, fig. 2). All types of fibres usually bear traces of variation in calcite secretion which may represent a diurnal periodicity. On fibre surfaces, these traces consist of ridges arranged at intervals of about 300 nm parallel with the rhombic corner angles of orthodox terminal faces or with the straight outer boundaries of the square-ended ones (PI. 23, fig. 3). In sections of the secondary layer such variation in carbonate secretion is indicated by curved growth banding disposed more or less normal and tangential to the outer and inner surfaces respec- tively of fibres. In contrast to the relatively coarse texture of the primary shell, the fibres of the secondary shell are generally so finely crystalline as to appear homogeneous within the resolution limits of the scanning electron microscope. However, some exception- ally well-preserved terminal faces bear ‘spheroidal’ seeds of calcite about 100 nm in size and these appear to be the basic units in carbonate precipitation (PI. 23, fig. 3). Tertiary layer. The passage from secondary fibres to fully developed prismatic calcite is less abrupt than the change from primary to secondary shell. On the internal surface of adult valves, secondary fibres occupy a band about 1-5 mm wide. But many fibres within the innermost third of that band exhibit some of the characteristic features of the prismatic layer, the extent of which has been conveniently determined as the area in which no recognizable fibres survive. This transitional zone coincides in section with the occurrence of impersistent lenses of prismatic calcite interfingering with regularly arranged fibres (PI. 24, fig. 3) representing slight transgressions and regressions of the secondary-tertiary boundary during shell deposition. The change in the secretory regime of the outer epithelium, on the other hand, is immediately indicated by cessation in protein exudation. In decalcified sections this change is represented by the abrupt termination along a fairly constant horizon of the protein sheets ensheathing secondary fibres (PI. 22, fig. 1). Consequently cells separated by EXPLANATION OF PLATE 22 Figs. 1-3. Transmission electron micrographs of decalcified shell and mantle of Liolhyrella neozelanica. 1. Section showing terminations of interconnected proteinous sheets pervading the secondary layer at its junction with the tertiary layer (below) ( x 14000). 2. Section through part of an outer epithelial cell responsible for the secretion of the tertiary shell showing continuous proteinous sheets connected by desmosomes to the plasmalemma ( X 55000). 3. Oblique section through the outer epithelium under- lying the tertiary layer showing the absence of proteinous sheets within the layer (above); connective tissue with collagens (below) separated from the outer epithelium by a basal lamina ( X 8200). Figs. 4-5. Scanning electron micrographs of the shell o\' Liolhyrella neozelanica. 4. Internal surface of the antero-medial edge of the brachial valve showing the primary and secondary layers with puncta ( X 1 300). 5. Detail of internal surface of the antero-medial edge of the brachial valve showing the junction between the primary layer (below) and the first-formed secondary tibes (above) ( X 6700). PLATE 22 MACKINNON and WILLIAMS, terebratulid shell 186 PALAEONTOLOGY, VOLUME 17 a space of variable thickness from the free ends of the protein sheets pervading the secondary shell, are identifiable as those which secreted the prismatic layer. Cells secreting prismatic calcite do not greatly differ from those responsible for the deposition of secondary fibres, being also flattened and charged with secretion droplets of glycoproteins in various stages of depletion (PI. 22, fig. 3). Indeed despite the absence of organic membranes in the prismatic layer, even a continuous pro- teinous sheet, up to 10 nm thick, persists about 20 nm external to the secreting plasmalemma to which it is attached by septate and fibrillar desmosomes (PI. 22, fig. 2). This monolayer, like its counterparts in younger epithelium, is interpreted as the outer boundary to a film of extracellular fluid which sustains carbonate secretion. The first sign of the change in carbonate secretion, which ultimately results in the deposition of a continuous layer of prismatic calcite, is the occurrence of malformed fibres on the internal surface of a valve (PI. 23, fig. 5). The irregular outlines of these fibres are due to the differential lateral growth and resorption of the distal edges of their terminal faces. Continued lateral expansion causes these crenulated edges to interlock into a coarsely granular mosaic covering the regularly arranged fibres. Organic secretion must cease at this stage in the deposition of the prismatic layer. The stage is represented in section by a sudden swelling of the fibres into prisms three or four times as thick, and a concomitant reorientation of the carbonate constituents to lie normal rather than oblique to the plane of the outer epithelium (PI. 24, fig. 2). On the surface of this mosaic, too, there is evidence that cells no longer correspond with prisms on a one-to-one ratio as they do with fibres. Sets of arcuate grooves and ridges about 400 nm apart and about 25 |wm from another series subtending a rhombic angle commonly occur straddling the boundaries of adjacent prisms (PI. 23, fig. 6). Such complementary sets of impressions are like the outlines of the terminal faces of fibres and are believed to represent the distal and proximal boundaries respectively of an outer epithelial cell. A noteworthy feature of the prismatic layer, even in these early stages of accretion is the way most extensions of the fibres, irrespective of size or persistence, remain discrete units. Since there are no enveloping proteinous sheets present to prevent lateral intergrowth as in secondary fibres, it is likely that either a water-soluble organic material occupies the interprismatic spaces in vivo and is lost in the preparation of sections, or the general lack of epitaxial alignment in adjacent ‘prisms’ precludes their amalgamation. EXPLANATION OF PLATE 23 Figs. 1 6. Scanning electron micrographs of the internal surface and sections of a brachial valve of Liolhyrella neozelcmica. I . Submedial longitudinal section of the primary layer showing the acicular crystallites and growth bands disposed at acute angles to the primary-secondary junction below ( X 2700). 2. Internal surface of secondary layer showing fibres with truncated edges to terminal faces (x2600). 3. Internal surface of secondary layer showing fibres with rounded edges to terminal faces and growth banding (x2000). 4. Transverse section of the secondary layer showing the characteristic outline and stacking of fibres; infilled punctum on right (x2400). 5. Internal surface showing the transitional area between the secondary and tertiary layers characterized by malformed fibres (x2600). 6. Internal surface showing the transitional area between the secondary and tertiary layers with outlines of epithelial cells impressed as grooves and ridges on a mosaic of expanded terminal faces ( X 2600). PLATE 23 MACKINNON and WILLIAMS, terebratulid shell 188 PALAEONTOLOGY, VOLUME 17 Once secretion of prismatic shell has begun, it may continue and give rise to an ever-thickening layer, or it may be temporarily interrupted by reversion to secondary shell secretion. The latter phase is represented in sections of a valve by isolated lenses of prismatic calcite within a succession of fibres or as wedge-like sheets of the tertiary layer tapering peripherally to interfinger with the secondary layer (PI. 24, fig. 3). Such sheets and lenses may be more than 200 long and 10 fj,m thick. They usually appear to grade laterally and vertically into secondary shell as the relatively big, irregularly shaped constituents, by which they are distinguished, give way to large then normal- sized, orthodoxly stacked fibres. Continuous secretion of prismatic calcite gives rise to a layer of closely packed dis- crete units, up to 20 /xm thick and more than 80 long, disposed normal to the internal surface. In section they appear as subrectangular bodies each with a convex internal surface representing the surface of accretion at the moment of death (PI. 24, fig. 4). Sporadically distributed growth bands, up to 400 nm thick, are arranged parallel with these internal surfaces (PI. 24, fig. 3) and together with traces of acicular crystallites lying normal to the banding, indicate the progress of deposition of the tertiary layer. The rhombic cleavage inherent to each prism is commonly conspicuously developed on the internal surface to define small rhombohedra or crystallites about 20 nm thick (PI. 24, fig. 1 ). The cleavage reveals the general non-alignment of adjacent units and emphasizes the latitude used in calling these units ‘prisms’, because they are more frequently bounded by curved rather than prismatic, dihexagonal or rhombohedral edges. Apart from recording variation in the accretion of the tertiary layer induced by temporal or environmental changes, growth banding also illustrates the relationship between a prism and the fibre from which it arose. Reference has already been made to the highly characteristic attitude of the growth bands seen in the sagittal section of a fibre. Their disposition is, of course, parallel with the profile of the terminal face which is itself determined by the differential rates of accretion of the calcite of the fibre and the protein of its enclosing sheet. The protein sheet exuded by the distal arc of an outer epithelial cell accumulates at about one-twenty-fifth the rate of the calcite deposited by the plasmalemma immediately behind the sheet (Williams 1971, p. 60). Towards the proximal region of the cell, however, carbonate secretion diminishes to a negligible amount. The total effect of such variation in secretory rates is, therefore, the perpetuation of an inwardly bulging terminal face to the fibre accommodated by a complementary concavity in the plasmalemma of the outer epithelial cell, the borders EXPLANATION OF PLATE 24 Figs. 1 -6. Scanning electron micrographs of the internal surface and sections of a brachial valve of Liothyrella neozelanica. 1 . Internal surface of the tertiary layer showing the non-alignment of cleavage in adjacent prisms ( X 2500). 2. Oblique view of a fractured internal surface of the tertiary layer showing the relation- ship between prisms and secondary fibres (xl400). 3-4. Oblique sections showing the relationship between the tertiary layer and an earlier formed lens of prismatic calcite (fig. 3), and secondary fibres (fig. 4) (x 1350, x2600). 5. Internal surface of the outer zone of the margin of the dorsal adductor scar showing the sporadically developed fibres arising from the prismatic layer (x 1300). 6. Internal surface of the inner zone of the margin of the dorsal adductor scar showing orthodoxly stacked fibres passing into irregular adductor myotest to the left (x 1250). PLATE 24 MACKINNON and WILLIAMS, terebratulid shell 190 PALAEONTOLOGY, VOLUME 17 periostracum TEXT-FIG. 1 . Stylized longitudinal section of part of a valve and mantle of Liothyrella showing migration of an outer epithelial cell during its secretion of the mineral shell. of which are pulled up around the terminal face by desmosomes attached to the invest- ing protein sheets. Two further aspects of the disposition of growth bands in fibres are noteworthy (text-fig. 1). First, the general plane of the terminal faces they repre- sent is inclined at about 30° to the fibres. Secondly, the bands are more or less parallel to those found in the tertiary layer although the general plane of these latter are normal to the long axes of prisms. These different attitudes of growth bands relative to the mineral units of the secondary and tertiary layers are reconcilable if it is assumed that secretion of the carbonate skeleton is essentially normal to the mantle surface, but that secretion of inter-connected proteinous sheets by anterior parts of outer epithelial cells introduces a distal component into the direction of growth of secondary fibres. Cessation of organic secretion, however, removes this bias and causes reversion to the pattern of accretion found governing the growth of the primary layer as is shown by the vertical extension of tertiary prisms. Muscle fields and eardinalia. Tertiary prisms constitute the most widely distributed superficial fabric in the interior of the shell occurring everywhere except in the muscle MACKINNON AND WILLIAMS: TEREBRATULID SHELL 191 fields and cardinalia. In both these areas the grossest modification is found in those patches of shell (muscle scars) underlying the outer epithelium associated with the adductor and diductor muscle bases. The fabric of muscle scar shell (myotest of Krans 1965, p. 65) is highly distinctive especially when overlain by adductor bases. Other fabric changes found in muscle fields and cardinalia are the result of muscle tissue migration and expansion of the skeletal supports for the lophophore and, although less spectacular than the myotest, are no less significant in terms of shell growth. The adductor myotest is developed in both valves but is better displayed in the brachial valve (text-fig. 2). Here, traverses across the anterior and flanking posterior adductor scars show that the changes from a typical prismatic shell to a fully dif- ferentiated myotest delineate a number of concentric zones for each half of the muscle field. The first change to occur is the reappearance of regular fibres with rounded distal edges to their terminal faces. These fibres, which are up to 15 /xm across and sheathed in protein sheets like secondary ones, are initially sporadically distributed, flat-lying outgrowths of the prismatic layer (PI. 24, fig. 5). Inwardly, however, towards the myotest proper their frequency of occurrence increases so that the slightly raised margin of the scar itself is usually coincident with an 80 jum band of closely stacked fibres overlapping one another in the direction of the anterior margin of the valve (PI. 24, fig. 6). Such fibres could only have been deposited by outer epithelial cells identical with those responsible for the secretion of the secondary layer. Within the muscle scar margin the fibres are abruptly succeeded by a somewhat broader zone of randomly orientated bars and granules of calcite usually between 5 and 10 fxm wide. The bars may be straight and up to 40 [xm long; more frequently they are curved or bent to define shallow pits about 10 ^xm across (PI. 24, fig. 6). The general impression conveyed by such features is that this zone represents the initial stages in the differential resorption of calcite. Inwardly towards the core of the muscle scar, this zone of irregular adductor myotest passes gradually into a series of long ridges, up to 2-5 [xm wide, arranged more or less parallel to the long axis of the scar at intervals of about 5 fxm. They are linked by cross ridges of the same thickness to define regularly arranged pits about 7 [xm long (PI. 25, fig. 1). The pits are comparable in origin and dimensions with those found in the adductor scars of living Tliecidellma whose relationship to muscle tissue is known in detail (Williams 1973, p. 456). It is therefore, reasonable to assume that during life each pit accommodates the tonofibril zone of an outer epithelial cell and its bounding ridges correspond to the peripheral and intercellular parts of adjacent cells as in ThecidelUna. In view of this correlation it seems equally likely that the zone of irregular adductor myotest reflects the presence of tonofibrils in the overlying epithelium although they are not as regularly dis- tributed as they are when cells are associated with muscle bases. The posterior boundaries of the muscle scars correspond with the margin of a zone of orthodoxly stacked, protein-sheathed fibres overlapping one another anteriorly. The fibres are identical in every respect with those of the secondary shell ; but their relative position within the shell succession is significantly different because not only do they transgress forward across the myotest of the muscle scars but also antero-medially across the prismatic shell of the medial septum. The origin of these fibres is best understood in relation to the expansion of the muscle bases. N 192 PALAEONTOLOGY, VOLUME 17 prismatic shell anterior TEXi -FiG. 2. Diagrammatic representation of the left dorsal adductor scar of Liofinrella showing the variation in the fabric of the myotest and the adjoining valve Hoor. MACKINNON AND WILLIAMS: TEREBRATULID SHELL 193 adjustor myotest TEXT-FIG. 3. Diagrammatic representation of the left half of the cardinalia of Liothyrella showing the variation in shell fabric; arrows indicate the direction of growth of secondary hbres. diductor myotest cardinal process primary- Enlargement of the muscle scars during shell growth results from the antero-lateral migration of the muscle bases away from the umbo as well as an increase in the area of attachment. As the bases expand and shift, an increasing area of outer epithelium first reverts to fibrous secretion and then becomes permeated by tonofibrils. Con- comitantly each concentric zone of the muscle scar expands outwards over its neighbour so that those composed of fibres and irregular adductor myotest form two gently inclined layers separating the prismatic shell from a wedge of regularly pitted myotest. Behind the migrating muscle bases, the epithelium reverts to fibrous secre- tion and the postero-medial areas of all three zones as well as the medial area between the scars are gradually buried by an expanding fibrous layer (text-fig. 2). When first 194 PALAEONTOLOGY, VOLUME 17 observed, the occurrence of orthodoxly stacked fibres which are commonly smaller than average, was ascribed to secretion by outer epithelium proliferated in secondary generative zones behind the muscle bases (Williams 1968, p. 16). More recent work, however, has shown that the fibres are secreted by the same cells as deposited myotest at an earlier stage in growth (MacKinnon 1971, p. 38). In the pedicle valve, traverses across the posterior part of the muscle field show that development of the ventral adductor myotest is accompanied by the same changes as are found in the brachial valve, although on a subdued scale so that adductor pits are not so regularly arranged and are further blurred by closely spaced transverse growth ridges (PI. 25, fig. 3). The ventral diductor myotest is also differentiated into linearly arranged subrectangular pits, about 7-5 ixm across, with prominent transverse growth ridges (PI. 25, fig. 2). In the brachial valve, however, the dorsal ends of the diductor muscles are inserted in the cardinal process (PI. 25, fig. 5 ; text-fig. 3) and are differently accommodated. This ovoid feature situated immediately anterior of the beak consists of forty or more plates diverging on either side of the medial plane. Each plate consists of long intertwined bars, about 5 ixm thick, aligned normal to the surface of the cardinal process and culminating in a series of nodules, about 20 [xm wide, arranged along the edge of each plate. Resorption pits and slots in the sides of the plates suggest that the muscle bases were inserted not only between plates but also into their sides (PI. 25, fig. 6). The fabric of the ventral faces of the inner socket ridges which act as seats of dorsal adjustor muscles is less altered than that of the shell underlying the ventral bases which consists of irregular plates, up to 60 ;u,m across, pitted by resorption. Within the dorsal scars, the secondary fibrous mosaic is modified by the overgrowth of bars and irregular extensions of the terminal faces (PI. 25, fig. 4). Some resorption pits occur but the fibrous foundation of this myotest layer is still apparent. Indeed irregular granules and bars of calcite, developed on the surfaces of the sockets and the inner side of the loop solely in response to differential growth, form more complex patterns. All these modifications, however, serve to emphasize the prevalence of normally dis- posed secondary fibres in the cardinalia and the loop and the absence of a tertiary prismatic layer. STRUCTURE OF GRYPHUS SHELL Only one other living terebratulid, a mature specimen of Gryphus vitreus (Born) from the mid Atlantic south of Portugal, was available for study. The soft parts were too poorly preserved to show the structure of the mantle, but the shell, freed of EXPLANATION OF PLATE 25 Figs. 1-6. Scanning electron micrographs of the internal surface of the shell of Liothyrella neozekmica. 1. Internal surface of the dorsal adductor myotest showing linearly arranged pits (x 1300). 2. Internal surface of the ventral diductor myotest showing the ridges developed in linearly arranged pits (x900). 3. Internal surface of the ventral adductor myotest showing the less regularly arranged pits (x 1900). 4. Ventral surface of the inner socket ridge showing the dorsal adjustor myotest with irregular growths on the terminal faces of fibres ( x 1 500). 5-6. Ventral views of the cardinal process with details of a plate showing intertwined bars and resorption slots (x75, x 1400). PLATE 25 MACKINNON and WILLIAMS, terebratulid shell 196 PALAEONTOLOGY, VOLUME 17 adherent tissue, afforded an opportunity to compare the skeletal fabric with that of Liofhyrella in much greater detail than is possible for fossil species. The general shell succession is identical with that of Liothyrella, differing only in detail and in the relative development of various layers. The primary layer, about 20 jum thick, is mainly composed of acicular crystallites but on a smaller scale than in Liothyrella so that the internal surface of the layer at the edge of the valve is much more finely granular (PI. 26, fig. 1). The first-formed fibres of the secondary layer which is only about 50 thick also differ in being almost invariably closely and regularly stacked at low angles to the primary-secondary interface just like mature fibres (PI. 26, fig. 1). Their transformation into prisms of the tertiary layer, each up to 25 fxm thick, is usually conspicuous as is growth banding parallel to the internal surfaces of accretion (PI. 26, fig. 2). On these surfaces, the prisms again fit together like discrete jigsaw pieces and display angular differences in the dominant cleavage of adjacent units (PI. 26, fig. 3). Surface features not seen in Liothyrella are a series of pits, about 750 nm in diameter, which are fairly regularly distributed at intervals of 10 ;u,m or so in the medial region of the shell. More often than not, the pits appear to be centres of radiating furrows about 700 nm wide (PI. 26, fig. 4). These furrows are so like macroscopic solution channels that they are tentatively ascribed to differential resorption, while the pits may have been formed by inhibition of carbonate secretion around groups of tonofibrils anchoring the outer epithelium to the inner surface of the tertiary layer. Unlike those of Liothyrella or indeed most living brachiopods, the mantle canals are deeply impressed on the internal surfaces of the valves. The vascula myaria, for example, are accommodated by a pair of slightly divergent troughs, about 200 |u,m wide and 150 /xm deep (PI. 26, fig. 5). Such troughs may arise by differential resorption of the shell but in Gryphus at least this is not so. The floor and sides of the canal trough are lined with unmodified prismatic calcite (PI. 26, fig. 6) and must have developed through slower carbonate deposition along that strip of outer epi- thelium overlying the mantle canal. In contrast to the close similarity in shell suc- cession, the muscle scars of Gryphus are quite distinct from those of Liothyrella. The adductor myotest was poorly preserved in the specimens examined but it was evident that, although differential secretion and resorption led to the delineation of suboval pits, up to 10 ixm across, bounded by calcitic bars, they are irregularly not linearly arranged. The ventral diductor scars are correspondingly less distinctive. Their margins mainly consist of mosaics of calcite plates, up to 30 [xm across, usually with EXPLANATION OF PLATE 26 Figs. 1 -6. Scanning electron micrographs of the internal surface and a section of the shell of Gryphus vUreiis ( Born) ; Recent, mid Atlantic, south of Portugal. 1 . Internal surface at the edge of a brachial valve show- ing both primary and fibrous secondary layers with puncta (x 1250). 2. Section of a mature part of a brachial valve showing the primary layer (top right) and the fibrous secondary layer passing into the prismatic tertiary layer (bottom left) ; infilled puncta occur in the top left and top right ( X 650). 3. Internal surface of the tertiary layer of a brachial valve showing the discrete nature of the prisms (x 1300). 4. Internal surface of the tertiary layer of a pedicle valve showing pits with radiating furrows (x650). 5. Internal surface of a pedicle valve showing the furrow of a msciilum myarium (x 120). 6. Internal surface of a pedicle valve showing the occurrence of typical tertiary prisms in the floor of a vasculuni myarium ( x 650). PLATE 26 MACKINNON and WILLIAMS, terebratulid shell 198 PALAEONTOLOGY, VOLUME 17 Straight boundaries subtending rhombic or prismatic angles. These boundaries become blurred and irregular within the myotest proper while resorption pits are common (PI. 27, fig. 1). The ventral adjustor myotest is distinguishable from typical prismatic calcite only in the relative absence of surface relief. STRUCTURE OF FOSSIL TEREBRATULID SHELL According to Muir-Wood (in Williams et al. 1965, pp. H773-800) forty-nine genera, based on wholly extinct species, may be unequivocally assigned to the Terebratulidae. Specimens representing thirty-three of these taxa have been sectioned to determine the shell succession. They are the Jurassic species Avonothyris plicatina (Sowerby) (B27389), Bihenithyris barringtoni Muir-Wood (B46287), Cereithyris intermedia (Sowerby) (B26997), Charltonithyris uptoni Buckman (B7 1 348), Epithyris cf. maxillata (Sowerby) (B27598), Euidothyris sp. (B66893), Heimia rnayeri (Choffat) (B71369), Kutchithyris acutiplieata (Kitchin) (B95464), Loboidothyris alf. latovalis Buckman (B3706), Lobotliyris punctata (Sowerby) (B32482), Lophrothyris etlieridgii (Davidson) (B21808), Pseudoglossothyris curvifrons (Davidson) (B52008), Ptyctothyris stephani (Davidson) (B 13921), Rugithyris subomalogaster Buckman (B61079), Somalithyris macfadyeni Muir-Wood (B94335), Sphaeroidothyris globisphaeroidalis Buckman (B71645), Stiphrothyris tumida (Davidson) (B8986), Stroudithyris pisolithica Buckman (B28807), Tubithyris wrighti (Davidson) (B71258), Wattonithyris wattonensis Muir- Wood (B89861), and Weldonithyris weldonensis Muir-Wood (B6653); the Cretaceous species Carneithyris subpentagonalia Sahni (B51 154), Concinnithyris obesa (Sowerby) (B7058), Cyrtothyris cyrta (Walker) (B21407), Gibbithyris semiglobosa (Sowerby) (B5091), Musculina acuta (Quenstedt) (B97875), Neoliothyrina obesa (Davidson) (B51306), Ornatothyris sulcifera (Morris) (B24931), IPlatythyris comptonensis Middlemiss (B21155), Praelongithyris praelonga Middlemiss (B2246), Rectithyris depressa{VdL\), Rhombothyris extensa (Meyer) (B25453), and Sellithyris sella (Sowerby) (B21931); and the Tertiary species Terebratula sp. (B47986). Only in the case of Bihenithyris was replacement of the shell so complete as to destroy all traces of the original skeletal fabric. In no species, however, was it possible to determine the nature EXPLANATION OF PLATE 27 Figs. 1-6. Scanning electron micrographs of various living and fossil terebratulid species. 1. Internal surface of a lateral area of the ventral diductor myotest of Gryplnis vitreus showing the mosaic of calcitic plates and resorption pits ( x 1000). 2. Section showing puncta penetrating a fibrous secondary layer and a relatively thin tertiary layer (to the left) in the shell of Lophrothyris cZ/zenV/g// (Davidson) (B21808), Inferior Oolite, Ravensgate Hill, England ( x 650). 3. Section showing acicular and granular crystallites and first-formed fibres of the primary and secondary layers respectively in the shell of Pseudo- glossothyris curvifrons (Davidson) (B52008), Inferior Oolite, Leckhampton Hill, England (x3000). 4. Section showing fine-grained sediment (below) filling a punctum in the fibrous secondary succes- sion forming the inner layer in the shell of Carneithyris subpentagonalis Sahni (B51I54), Senonian, Catton, England (x 1400). 5. Section showing the passage of secondary fibres into tertiary prisms in the shell of Euidothyris sp.. Inferior Oolite, Misterton, England (x 1500). 6. Section showing the granular tertiary layer, with the junction with the secondary layer above and an infilled punctum to the right in the shell of Praelongithyris praelonga Middlemiss (B2246), Lower Greensand, Potton, England (x 1500). PLATE 27 1 3 2 4 MACKINNON and WILLIAMS, terebratulid shell 200 PALAEONTOLOGY, VOLUME 17 of the myotest which consists of thin layers mostly exposed on the floor of the valves and, therefore, subject to gross alteration even by non-penetrative diagenetic pro- cesses. Indeed unless techniques are developed to remove the micrite normally covering the surfaces of fossil shells even when free of entombing sediment, the prospects of carrying out a comprehensive survey of the fabric of muscle scars are poor. Comparison of the shell succession of the species listed above reveals a significant difference only in the relative development of the tertiary layer. All shells are penetrated by regularly distributed unbranched puncta with an average diameter of 10 /xm ranging from 6 to 38 /xm in Kutchithyris and Tubithyris respectively. The primary layer is, on average, 53 [xm thick varying from 16 fxm in Carneithyris to 220 ixm in Charltonithyris. The fabric of most species consists of the same mixture of acicular crystallites and granules, up to 15 and 5 ixm in size respectively, as is found in living terebratulids (PI. 27, fig. 3). In both Carneithyris and Rhombothyris, however, the texture is predominantly granular. In those species with a tertiary layer, the secondary shell has a mean thickness of 182 ^m varying from 46 /ixm in Wattonithyris to 477 /um in Platythyris. It consists exclusively of gently inclined, orthodoxly stacked fibres with an average width, when first-formed, of 5 ranging from 3 to 8 in Cyrto- thyris and Concinnithyris respectively. Identification of the tertiary layer is difficult because its components, even in the unaltered state, may be very like micritic accre- tions on the inner surfaces of valves which may actually replace parts of the layer and thereby blur its junction with the sediment. The layer, however, like the rest of the shell, is penetrated by puncta and traces of the boundaries within any carbonate underlying a fibrous succession are regarded as indicating the presence of tertiary calcite (PI. 27, fig. 3). By this means the tertiary layer has been identified in all specimens examined except those assigned to Carneithyris (PI. 27, fig. 4), Lobothyris, Neolio- thyrina, Rhombothyris, Terebratula, and Tubithyris. The layer itself may be preceded by thin lenses or even continuous layers (up to 40 jum thick in Euidothyris) inter- calated with fibres of the secondary shell. It may also contain impersistent lenses of fibres but it consists mainly of large components which, although recrystallized in some species into interlocking rhombohedra, clearly represent the prisms charac- teristic of the tertiary shell of Liothyrella (PI. 27, fig. 5). The only noteworthy textural variation occurs in Praelongithyris, the tertiary layer of which is composed of granules up to 20 jxm in size interpersed with irregular fibres (PI. 27, fig. 6). CONCLUSIONS The most interesting aspect of the structure of the terebratulid shell is the strati- graphic and taxonomic distribution of those genera lacking a tertiary layer. The Terebratulidae are usually divided into six subfamilies in the manner proposed by Muir-Wood (in Williams et al. 1965, pp. H773-800) although at that time the arrange- ment of the largest subfamily, the Terebratulinae, was regarded as only provisional (op. cit., p. H773). The family includes the Upper Triassic Plectoconcha Cooper, the earliest known terebratulacean, and is generally conceded to be ancestral to the other familial groups assigned to the Terebratulacea with the possible exception of the Orthotomidae. Although the fabric of the Plectoconcha shell has not yet been deter- MACKINNON AND WILLIAMS: TEREBRATULID SHELL 201 mined, the fact that another early terebratulid, the Liassic Lobothyris, lacks a tertiary layer is consistent with the inferred descent of the terebratulaceans from the dielas- mataceans (Stehli 1956, p. 194). The shell structure of species belonging to the latter superfamily has not been systematically investigated but information currently available (Williams 1968, p. 31 and in manuscript) suggests that a succession consisting of only the primary and secondary layers is prevalent. It is, therefore, feasible to infer that the tertiary layer found in the majority of Terebratulidae was acquired geronto- morphically after the emergence of the stock from its dielasmatacean progenitors. Moreover, even when allowance is made for the lack of information about the shell structure of those genera not available for study, the stratigraphic distribution of forms without a tertiary layer is too sporadic and their morphology too disparate to entertain their descent from Lobothyris (except for Tubithyris) or from one another. Consequently, it is likely that the skeletal succession of the Cretaceous rectithyridinids NeoHothyrina and Rhombothyris and the carneithyridinid Carueithyris, and even the Tertiary terebratulinid Terebratula, resulted from a repeated neotenous suppression of prismatic secretion. With regard to changes in shell fabric resulting from the emplacement of muscle bases, the degree of difference in the myotests of Liothyrella and Gryplius is unexpected. Studies of the myotests in other living rhynchonellides and terebratulides (Williams 1968; MacKinnon 1971) indicate that the principal modifications include a welding together of secondary fibres and loss of their characteristic outline. Since these changes occur during secretion of the prismatic layer, further modification through muscle attachment should have produced structurally similar myotests. Yet the pits of the adductor myotest of Liothyrella are characteristically strongly defined and regularly arranged, while those of the diductor myotest have no counterpart in Gryplius. Such differences suggest a deeper, more oblique insertion of the muscle bases into the Liothyrella shell. It is, therefore, possible that important factors in determining the fabric of the myotest are the disposition of the muscles relative to the valve floors as well as their strength. Acknowledgements. We are indebted to Miss Anne Brunt of the Fisheries Research Division (Wellington) of the New Zealand Marine Department and Mr. Ellis Owen of the British Museum (Nat. Hist.) for pro- viding us with living and fossil species. We also wish to thank Dr. Jean Graham, Research Assistant in the Department of Geology, The Queen’s University, Belfast, for help in the preparation of illustrations, and the Natural Environment Research Council for grants for the ‘Stereoscan’ scanning electron microscope and for the post-graduate grant held by MacKinnon from 1968 to 1971. REEERENCES ALEXANDER, F. E. s. 1948. A revision of the genus Pentameriis. James Sowerby 1813 and a description of the new species Gvpidida bnivonium from the Aymestry limestone of the main outcrop. Quart. J. geol. Soc. Loud. 103, 143-161. AMSDEN, T. w. 1964. Bracliial plate structure in the brachiopod family Pentameridae. Palaeontology, 7, 220-239. CARPENTER, w. B. 1853. On the intimate structure of the shells of Brachiopoda. In davidson, t., British fossil Brachiopoda, 1, 23-40. Palaeontograph. Soc. London. CLOUD, p. E., Jr. 1942. Terebratuloid Brachiopoda of the Silurian and Devonian. Spec. Pap. Geol. Soc. Anier. 38, 1-182. 202 PALAEONTOLOGY, VOLUME 17 GAURi, K. L. and BOUCOT, A. J. 1968. Shell structure and classification of Pentameracea M’Coy, 1844. Palaeomographica, 131, Abt. A, 79-135. KING, w. 1871 . On the histology of the test of the class Palliobranchiata. Trans. Roy. Ir. Acad. 24, 439-455. KRANS, TH. F. 1965. Etudes morphologiques de quelques spiriferes devoniens de la Chaine Cantabrique (Espagne). Leidse geol. Med. 33, 71-148. / MACKINNON, D. I. 1971 . Studies in shell growth in living articulate and spiriferide Brachiopoda. Ph.D. thesis. Queen’s Univ., Belfast. OWEN, G. and williams, a. 1969. The caecum of articulate Brachiopoda. Proc. Roy. Soc. B. 172, 187-201. ST. JOSEPH, j. K. s. 1938. The Pentameracea of the Oslo region. Norsk geol. tidsskr. 17, 225-336. SASS, D. B. and munroe, e. a. 1967. Shell-growth in recent terebratuloid Brachiopoda. Palaeontology, 10, 298-306. STEHLi, F. G. 1956. Evolution of the loop and lophophore in terebratuloid brachropods. Evolution, 10, 187-200. WILLIAMS, A. 1956. The calcareous shell of the Brachiopoda and its importance to their classification. Biol. Rev. 31, 243-287. 1968. Evolution of the shell structure of articulate brachiopods. Spec. Pap. Palaeont., No. 2, 1-55. 1971. Comments on the growth of the shell of articulate brachiopods. Smithson. Contributions to Paleobiology, 3, 47-67. 1973. The secretion and structural evolution of the’ shell of thecideidine brachiopods. Trans. Rov. Soc. B. 264, 439-478., et al. 1965. Treatise on invertebrate paleontology (ed. Moore, R. C.): Part H. Brachiopoda. 927 pp. Lawrence (Univ. of Kansas). D. I. MACKINNON Department of Geology The University of Canterbury Christchurch New Zealand ALWYN WILLIAMS Department of Geology The Queen’s University of Belfast Belfast BT7 INN- Manuscript received 29 January 1973 Northern Ireland w. \ > i » .«■ ■30 ..-•I Mu. t? THE PALAEONTOLOGICAL ASSOCIATION The Association was founded in 1957 to further the study of palaeontology. It holds meetings and demonstrations as well as publishing Palaeontology and Special Papers in Palaeontology. Membership is open to individuals and to institutions on payment of the appropriate annual subscription ; Institutional membership .... £10 00 (U.S. $26.00) Ordinary membership .... £5-00 (U.S. $13.00) Student membership .... £3-00 (U.S. $8.00) There is no admission fee. Institutional membership is only available by direct application, not through agents. Student members are persons receiving full-time instruction at educational institutions recognized by the Council. On first applying for membership, an application form should be obtained from the Membership Treasurer. Subscriptions cover one calendar year and are due each January; they should be sent to the Membership Treasurer, Dr. E. P. F. Rose, Department of Geology, Bedford College, Regent’s Park, London, NWl 4NS, England. PALAEONTOLOGY All members who join for 1974 will receive Volume 17, Parts 1-4. All back numbers are still in print and may be ordered from B. H. Blackwell, Broad Street, Oxford, 0X1 3BQ, England, at £5 per part (post free). A complete set. Volumes 1-16, consists of 63 parts and costs £315. SPECIAL PAPERS IN PALAEONTOLOGY The subscription rate is £8 (U.S. $22.00) for Institutional Members and £4 (U.S. $11-00) for Ordinary and Student Members. Subscriptions should be placed through the Membership Treasurer, Dr. E. P. F. Rose, Department of Geology, Bedford College, Regent’s Park, London, NWl 4NS, England. Ordinary and Student members only may obtain individual Special Papers from Dr. Rose at reduced rates. Non-members may obtain them at the stated prices from B. H. Blackwell, Broad Street, Oxford, 0X1 3BQ, England. COUNCIL 1973-1974 President-. Professor M. R. House, The University, Kingston upon Hull, Yorkshire, HU6 7RX Vice-Presidents-. Mr. N. F. Hughes, Department of Geology, Sedgwick Museum, Cambridge, CB2 3EQ Dr. Isles Strachan, Department of Geology, The University, Birmingham, B15 2TT Treasurer-. Dr. J. M. Hancock, Department of Geology, King’s College, London, WC2R 2LS Membership Treasurer-. Dr. E. P. F. Rose, Department of Geology, Bedford College, Regent’s Park, London, NWl 4NS Secretary: Dr. W. D. I. Rolfe, Hunterian Museum, The University, Glasgow, G12 8QQ Assistant Secretary: Dr. C. T. Scrutton, Department of Geology, The University, Newcastle upon Tyne, NEl 7RU Editors Dr. R. Goldring, Department of Geology, The University, Reading, RG6 2AB Dr. J. D. Hudson, Department of Geology, The University, Leicester, LEI 7RH Dr. D. J. Gobbett, Department of Geology, Sedgwick Museum, Cambridge, CB2 3EQ Dr. L. R. M. Cocks, Department of Palaeontology, British Museum (Natural History), London, SW7 5BD Other members of Council Dr. M. G. Bassett, Cardiff Dr. D. D. Bayliss, Llandudno Dr. E. N. K. Clarkson, Edinburgh Prof. D. L. Dineley, Bristol Dr. Julia A. E. B. Hubbard, London Dr. C. P. Hughes, Cambridge* Dr. J. K. Ingham, Glasgow Mr. M. Mitchell. Leeds Dr. Marjorie D. Muir, London Dr. J. W. Murray, Bristol* Dr. B. Owens, Leeds Dr. P. F. Rawson, London Prof. D. Skevington, Galway * Co-opted as Editors Overseas Representatives Australia: Professor Dorothy Hill, Department of Geology, University of Queensland, Brisbane Canada: Dr. B. S. Norford, Institute of Sedimentary and Petroleum Geology, 3303-33rd Street NW, Calgary, Alberta India: Professor M, R. Sahni, 98 The Mall, Lucknow (U.P.), India Ne-w Zealand: Dr. C. A. Fleming, New Zealand Geological Survey, P.O. Box 30368, Lower Hutt West Indies and Central America: Mr. John B. Saunders, Geological Laboratory, Texaco Trinidad, Inc., Pointe-a- Pierre, Trinidad, West Indies Western U.S. A.: Professor J. Wyatt Durham, Department of Paleontology. University of California, Berkeley 4, California Eastern U.S. A. : Professor J. W. Wells, Department of Geology, Cornell University, Ithaca, New York Palaeontology VOLUME 17 ■ PARTI CONTENTS A revision of some Ordovician graptolites of eastern North America JOHN RIVA 1 The Lower Palaeozoic acritarchs Priscogalea and Cymatiogalea S. M. RASUL 41 Two new Paleocene dinoflagellates from Virginia and Maryland DEWEY M. MCLEAN 65 Trilobites from the Gorran Quartzites, Ordovician of south Cornwall P. M. SADLER 71 The trilobite Clavagnostus Howell from the Cambrian of Tasmania J. B. JAGO and B. DAILY 95 A new pelagic trilobite from the Ordovician of Spitsbergen, Ireland, and Utah R. A. FORTEY 1 1 1 Williamsoniella lignieri : its pollen and the compression of spherical pollen grains TOM M. HARRIS 125 Early growth stages in rhabdomesoid bryozoans from the Lower Carboniferous of Hook Head, Ireland RONALD TAVENER-SMITH 149 A new genus of Jurassic bivalve mollusc ancestral to Globocardium C. P. PALMER 165 Shell structure of Terebratulid brachiopods D. I. MACKINNON and A. WILLIAMS 179 Printed in Great Britain at the University Press, Oxford by Vivian Ridler, Printer to the University THE PALAEONTOLOGICAL ASSOCIATION The Association publishes Palaeontology and Special Papers in Palaeontology. Details of member- ship and subscription rates may be found inside the back cover. PALAEONTOLOGY The journal Palaeontology is devoted to the publication of papers on all aspects of palaeontology. Review articles are particularly welcome, and short papers can often be published rapidly. A high standard of illustration is a feature of the journal. Four parts are published each year and are sent free to all members of the Association. SPECIAL PAPERS IN PALAEONTOLOGY This is a series of substantial separate works. Members may subscribe to the Series; alternatively. Ordinary and Student members only may obtain individual copies at reduced rates. The following Special Papers are available : 1. (for 1967): Miospores in the Coal Seams of the Carboniferous of Great Britain, by a. h. v. smith and M. A. BUTTERWORTH. 324 pp., 72 le.xt-figs., 27 plates. Price £8 (U.S. $22.00), post free. 2. (for 1968): Evolution of the Shell Structure of Articulate Brachiopods, by a. williams. 55 pp., 27 text- figs., 24 plates. Price £5 (U.S. $13.00). 3. (for 1968): Upper Maestrichtian Radiolaria of California, by Helen p. foreman. 82 pp., 8 plates. Price £3 (U.S. £8.00). 4. (for 1969): Lower Turonian Ammonites from Israel, by r. freund and m. raab. 83 pp., 15 text-figs., 10 plates. Price £3 (U.S. $8.00). 5. (for 1969): Chitinozoa from the Ordovician Viola and Fernvale Limestones of the Arbuckle Moun- tains, Oklahoma, by w. a. m. jenkins. 44 pp., 10 text-figs., 9 plates. Price £2 (U.S. $5.00). 6. (for 1969): Ammonoidea from the Mata Series (Santonian-Maastrichtian) of New Zealand, by R. a. HENDERSON. 82 pp., 13 text-figs., 15 plates. Price £3 (U.S. $8.00). 7. (for 1970): Shell Structure of the Craniacea and other Calcareous Inarticulate Brachiopoda, by A. williams and a. d. wright. 51 pp., 17 text-figs., 15 plates. Price £1-50 (U.S. $4.00). 8. (for 1970): Cenomanian Ammonites from Southern England, by w. j, Kennedy. 272 pp., 5 tables, 64 plates. Price £8 (U.S. $22.00). 9. (for 1971): Fish from the Freshwater Lower Cretaceous of Victoria. Australia, with Comments on the Palaeo-environment, by m. waldman. 130 pp., 3>1 text-figs., 18 plates. Price £5 (U.S. $13.00). 10. (for 1971 ): Upper Cretaceous Ostracoda from the Carnarvon Basin, Western Australia, by R, h. bate. 148 pp., 43 text-figs., 27 plates. Price £5 (U.S. $13.00). 11. (for 1972): Stromatolites and the Biostratigraphy of the Australian Precambrian and Cambrian, by M. R. WALTER. 268 p/7. , 55 text-figs., plates. Price £10 ( U.S. $26.00). 12. (for 1973): Organisms and Continents through Time. A Symposium Volume of 23 papers edited by N. F. HUGHES. 340 pp.. 132 text-figs. Price £10 (U.S. $26.00) (published with the Systematics Asso- ciation). 13. (for 1974) : Graptolite studies in honour of O. M. B. Bulman. Edited by r. b. rickards, d. e. jackson, and c. p. HUGHES. 261 pp., 26 plates. Price £10 (U.S. $26.00). 14. (for 1974): Palaeogene foraminiferida and palaeoecology, Hampshire and Paris Basins and the English Channel, by J. w. Murray and c. a. wright. 45 text-figs., 20 plates. In press. SUBMISSION OE PAPERS Typescripts on all aspects of palaeontology and stratigraphical palaeontology are invited. They should conform in style to those already published in this journal, and should be sent to The Secretary, P.A. Publications Committee, Department of Geology, Sedgwick Museum, Downing Street, Cambridge, CB2 3EQ, England, who will supply detailed instructions for authors on request (these are published in Palaeontology, 15, pp. 676-681). © The Palaeontological Association, 1974 Cover: Dactylioceras commune (.1. Sowerby), Upper Lias, .lurassic, Whitby, Yorks. In collection of Professor J. E. Hemingway. The 'head' was carved by a Whitby jet worker to illustrate the traditional story of St. Hilda of Whitby (614 80), who turned snakes into stones by the power of prayer. UPPER ORDOVICIAN TRILOBITES FROM CENTRAL NEW SOUTH WALES by B. D. WEBBY Abstract. Twelve trilobite species are described and illustrated from Upper Ordovician (Caradoc) successions of central New South Wales, Included among the forms is a new scutelluid genus, Heptabronteus, with two new species, H. atavus (type species) and H. major, and five other new species, Toemquistia arguta, Parkesolithus dictyotos, Sphaerocoryphe exserta, AmphiHchas nasutus, and A. encyrtos. A discussion of relationships between Heptabronteus and other scutelluid genera is presented. Triarthrus is recorded for the first time from New South Wales, and evi- dence for dimorphism in the raphiophorid, Malongullia oepiki Webby, Moors, and McLean, 1970, considered. The stratigraphical distribution of the faunas is described, and zoogeographical relationships discussed. The New South Wales fauna is included in the Heptabronteus- Pliomerina province of Australia, South-East Asia, and Kazakh- stan. one of two or more separate provinces within the much larger, band-like, probably equatorial, Remopleuridid realm of the Caradoc. The faunas are interpreted broadly as occurring in an island-arc tectono-environmental setting. They have been differentiated into a possible benthonic assemblage in the shallow-water carbonate rise facies, and into inferred benthonic and pelagic associations in the deeper, graptolitic shale slope and trough facies. Earlier descriptions of the Upper Ordovician trilobites from central New South Wales were given by Webby, Moors, and McLean (1970), Campbell and Durham (1970), Webby (1971), and Webby (1973). This paper concludes the description of the faunas based on collections in the Department of Geology and Geophysics, University of Sydney. DISTRIBUTION The distribution of trilobites in the Upper Ordovician successions of central New South Wales is represented in text-fig. 1. Locality details are given in previous papers and in the systematic descriptions of these pages. Most of the localities occur within the areas of the geotectonic ‘highs’, the Molong Rise and the Parkes Platform (text- figs. 2 and 4), in off-shore parts of the Lachlan Geosyncline (Packham 1969). In both regions extensive outpourings of Lower or Middle Ordovician andesitic volcanics accumulated on the sea floor, and seem to have built up ‘highs’ within the geo- syncline. These volcanic piles apparently became emergent and planated, providing platforms for the deposition of shallow-water carbonates. It remains to be estab- lished whether in the Upper Ordovician a deeper-water shale succession was being deposited in the Cowra Trough between contemporaneous limestone successions of the Parkes Platform and Molong Rise, or whether the carbonates originally formed on a single, extensive Parkes-Molong volcanic rise (text-fig. 4). Packham (1967; 1969, p. 217) has indicated that Upper Ordovician graptolite deposits occur in the Cowra Trough at Tomingley but, considering their location 40 miles (64 km) north of Parkes, they may not be strictly assignable to the trough. Scheibner (1972) has claimed that the Cowra Trough originated by splitting of the Parkes-Molong vol- canic rise in the early Silurian. To the west of the Parkes Platform, late Ordovician quartz-rich greywackes and slates were deposited in the Wagga Trough, a possible [Palaeontology, Vol. 17, Part 2, 1974, pp. 203-252, pis. 28-34.] A FACIES CARBONATES GRAPTOLITIC SHALES TECTONIC ELEMENT P. P. Molong Rise PP Molong Rise H.ET. FORMATION DISTRIBUTION Unnamed limestone, Billabong Creek Bowan Park Group Cliefden Caves Limestone Unnamed shales. New Durran = 8 o t 05 O Cheesemans Creek Formatiwi near Keenans Bridge Oakdale Fm., near Newrea 'Malongulli Fm', Junction Reefs Ballingoole Formation Quondong Formation Daylesford Formation Cliefden Caves (type area) Regan's Creek area North of Cheesemans 1 Creek (derived ) Upper part Lower part Geragnostus? sp vr Shumardia sp vr r Tnarthrus sp r Remopleurides saenuros ^ me R exallos ^ c R acer mo R. sp r R. sp vr R. sp C vr? R sp indet r Pseudobasilicus ^ fortis VO P ? sp A me me asaphid gen. et sp indet r lllaenus ( Parillaenus) ? incertus ve 1 iPJ? sp A vr Heptabronteus atavus e H major c Toernquistia arguta harpid gen. et sp nov. r r proteid gen, et sp indet r Parkesolithus gradyi c P dictyotos me c tnnucleid gen et sp indet r o MalonguHia oepiki (dimorph A)'^ VC r M oepiki (dimorph r me Sphaerocoryphe exerta r Phomerina austrma ve c c P pnma me VC Snennurasp/s opt/mus me vr Encrinuraspis ? sp A r r E ? sp B vr encnnurid gen et sp indet r AmphUichas nasutus r A encyrtos r TEXT-FIG. I. Chart showing trilobite distribution in Upper Ordovician of central New South Wales. The three dimorphs in the faunas are numbered 1, 2, and 3. Relative abundances at the best collecting locality of the particular unit are given by symbols: VC, very common; c, common; me, moderately common; r, rare; vr, very rare. P.P., Parkes Platform. H.E.T., Hill End Trough. ^ WELLINGTON + Newrea loc MOLONG Mirrabooka loc, north of — Cheeseman's Creek Bowan Park Group (eenans Bridge ORANGE Regan's Creek-; Trilobite Hill and Copper Mine Creek locs, diefden Caves Limestone COWRA + Junction ReefS; loc. -HVIANDURAMA'; LEGEND U Ordovician limestones Volcanics, mostly andesitic and underlying limestones Oakdale Formation Smith's 'Malongulli Fm' 0 10 □ Post-Ordovician Rocks Undifferentiated Ordovician sediments and volcanics Cheesemans Creek Fm 20 1 0 10 20 Malongulli Formation 30 miles I 30 km TEXT-FIG. 2. Generalized map showing Ordovician geology of central New South Wales, and principal trilobite localities. The major tectonic elements represented are situated in the eastern part of the Lachlan Geosyncline. Note occurrences of Upper Ordovician limestones on the Parkes Platform and Molong Rise, formed in off-shore parts of the geosyncline. Based on data from the Macquarie 1 : 500,000 Geol. Series Sheet, First Edition, 1970. 206 PALAEONTOLOGY, VOLUME 17 marginal sea (Packham and Falvey 1971). The Hill End Trough (text-fig. 4) to the east of the Molong Rise has a history of andesitic volcanism predating the earliest- determined graptolites of Darriwilian (Middle Ordovician) age (Smith 1966). The Middle-Upper Ordovician deposits comprise mainly shales, cherts, greywackes, tuffs, and andesitic volcanics. Thick limestone successions are developed on the Parkes Platform (unnamed limestone at Billabong Creek), and on the flanks of the Molong Rise (western side, typified by 560 m-thick Bowan Park Group, and eastern side by Cliefden Caves Limestone, some 250 m thick). In the thickest, continuous limestone succession at Bowan Park, three biostratigraphically distinct trilobite assemblages are recognized. Only the first and second occur in the Cliefden Caves Limestone, and only the second, in the unnamed limestone at Billabong Creek. These assemblages are named faunules after the most common constituent species. They comprise the Pliomerina prima, P. austrina, and "lUaenus’ incertus faunules (text-figs. 3, 4), and correlate reasonably well with the stromatoporoid/coral faunas I, II, and III introduced previously (Webby 1969). Directly overlying the Cliefden Caves Limestone is the Malongulli Formation, a shale and siltstone succession containing in the lower part an abundant sponge spicule, graptolite, trilobite, and brachiopod fauna. The graptolites determined by Moors (1970) suggest the Zone of Dicranograptus hians, that is. Upper Eastonian in terms of the Victorian stages. The Zone of D. hians seemingly corresponds to the Zone of D. clingani in Europe (Strachan 1972). The rich trilobite assemblage, the Malongullia oepiki faunule (text-fig. 3), is recognized in the type area near Cliefden Caves, in the Regan’s Creek area, where it lies above the Regan’s Creek Limestone, and further north, in the Cheeseman’s Creek area. In this latter area Sherwin (1971) has interpreted the occurrences of shales of the Malongulli Formation and of lime- stones near Mirrabooka homestead as exotic blocks in the Upper Silurian Wallace Shale. It seems unlikely, judging from the size of some of the blocks and the apparent ordered arrangement of shale and limestone occurrences, that they have been trans- ported far from their original sites of deposition. With the sudden change of facies from the carbonates of the Cliefden Caves Limestone type to the graptolitic shales of Malongulli Formation type on the eastern flank of the Molong Rise, there is the appearance of a completely different trilobite fauna. At least three genera, Toern- quistia, Parkesolitims, and Malongullia, make their first appearance in this shaly, possibly ‘deeper-water’, facies, and all the species are quite distinct from those of the preceding carbonates. The transgression of ‘deeper-water’ facies over the eastern flank of the rise did not continue on to the western flank where carbonate deposition continued uninterrupted (text-fig. 4), with the accumulation of the Ballingoole Formation (Bowan Park Group), and the upper parts of the Cargo Creek and Cano- modine Limestones. Limestone breccias derived from these adjacent carbonate platforms occur in the ‘deeper-water’ shales of the Malongulli Formation. Sherwin (1971) introduced the Cheesemans Creek Formation for about 900 m of interbedded siltstone, andesites, tuffs, and greywackes overlying the Reedy Creek Limestone in the Cheeseman’s Creek area, and cropping out in a broad belt to the east of the occurrences of exotic blocks in the Wallace Shale. There is a broad corre- lation of the Reedy Creek Limestone with the Cliefden Caves Limestone (Webby TEXT-FIG. 3. Correlation of the Upper Ordovician (Caradoc) trilobite faunal successions in graptolitic shale and carbonate facies of central New South Wales. C/S faunas, coral/stromatoporoid faunas of Webby 1969. 208 PALAEONTOLOGY, VOLUME 17 1969). The graptolite occurrences in the Cheesemans Creek Formation suggest an Eastonian and Bolindian age (Sherwin 1971, p. 207). It therefore correlates with the Malongulli Formation and the overlying Angullong Tuff of the Cliefden Caves area. The occurrence of Triarthrus sp. in the beds of the Cheesemans Creek Forma- tion near Keenan’s Bridge, lying stratigraphically near the middle of the formation, is possibly about the Eastonian-Bolindian boundary or Lower Bolindian age. It probably lies stratigraphically somewhat higher than the trilobite occurrences in the Malongulli Formation (text-fig. 3). Triarthrus is not known elsewhere in beds younger than the Caradoc. This particular occurrence would therefore seem, from generalized correlation, to be of late Caradoc age. As already mentioned (Webby 1973), it is likely that the records of Shumardia and Geragnostusl in the Oakdale Formation near Newrea, and of Shumardia in Smith’s ‘Malongulli Formation’ at Junction Reefs, near Mandurama, are significantly older, possibly Gisbornian occurrences (text-fig. 3). Part of these shale successions may have been laid down essentially contemporaneously with the lower part of the Cliefden Caves Limestone and other equivalent limestones on the eastern flank of the Molong Rise, in deeper waters of the Hill End Trough to the east (text-fig. 4). On the Parkes Platform, some 50 miles (80 km) to the west (text-fig. 2), another similar succession of limestones is succeeded by graptolitic shales. Although the detailed stratigraphy remains to be fully elucidated, there are two localities con- taining trilobites, one in the upper half of the limestone at Billabong Creek (Packham 1967) and the other in the stratigraphically higher shales near New Durran homestead (Campbell and Durham 1970). The rich, silicified trilobite assemblage in the lime- stone belongs to the Pliomerina austrina faunule, and is probably about Lower Eastonian in age (text-fig. 3). The overlying shales containing the trinucleid Parke- solithus gradyi Campbell and Durham, 1970 have been considered, from their associated graptolites, to have an Eastonian age by Campbell and Durham, and a Bolindian or less probably an Eastonian age by Sherwin (1970). ZOOGEOGRAPHY Whittington (1966) originally included the New South Wales ‘Caradoc’ trilobites in the "Encrinurella' fauna of South-East Asia and Australia. The fauna was sub- sequently renamed the Pliomerina fauna (Webby 1971) because Encrinurella could not be confirmed in ‘Caradoc’ successions of the Australian region. According to this conception, it was taken to represent a faunal province, and to include Aus- tralia, South-East Asia (including South Korea), and Kazakhstan. From a more recent analysis of the world-wide distribution of Caradoc trilobites, Whittington TEXT-FIG. 4. Diagram showing generalized lateral interrelationships of the carbonate and graptolitic shales facies overlying andesitic volcanics in the region between the Parkes Platform and the Hill End Trough, central New South Wales. Sections: A, Billabong Creek-New Durran; B, Bowan Park; Ci, Cheeseman’s Creek; C2, Cliefden Caves; Di, Newrea; D2, Junction Reefs. Faunules: p, Pliomerina primer, a, P. ausirina; i. Illaemis {Parillaeniis)'i incerlii.s; o, Malongiillia uepiki. Occurrences: s, Simmarditr, I, Triarthrus', g, ParkesuHthiis gradyi. Note that the thicknesses are very approximate, and that the graptolitic snale facies includes horizons of tulf and, in the sections Ci, Di, and D2, volcanics. WELLINGTON PARKES West East HILL END PLATFORM COWRA TROUGH MOLONG RISE TROUGH 210 PALAEONTOLOGY, VOLUME 17 and Hughes (1972) have considered only two provinces to be clearly recognized— a small, rather restricted Selenopeltis province (including southern Europe and North Africa) and a vast Remopleuridid province (covering North America, northern Europe, most of Asia, and Australia), the latter having essentially an equatorial distribution. The few Caradoc trilobites known from South America and New Zea- land have uncertain affinity. South-East Asia and Australia are suggested to be a possible subprovince of the Remopleuridid province. In their palaeogeographical reconstructions, Kazakhstan which has species of Pliomerina, and North-East Asia which has not, are placed together, hugely displaced from South-East Asia and Australia. Considering the New South Wales ‘Caradoc’ fauna in general terms, three out of a total of fifteen genera— Parkesolitlnis, Malongullia, and a new harpid— are believed to be endemic (Table 1). Zoogeographical relationships are closest to South-East TABLE 1. Caradoc zoogeographical relationships and endemism of the New South Wales genera. X = confirmed; ? = possible. Endemic Zoogeographical relationships elements SE. Asia Kazakh- North North New South stan Europe America Zealand America Remopleurides X X X X Amphilichas X X X X Sphaerocoryphe ? ? X X Pseudobasilicusl X ? 7 I. (Parillaenus)! ? 7 7 Toernquistia ? X X 7 Heptabronteus ? X Pliomerina X X Encrinuraspis X Parkesolithus X Malongullia X harpid gen. nov. X Geragnostusl ? 7 7 Shumardia ? X X Triarthrus ? 7 X X X X Asian and Kazakhstan regions. The genera Pliomerina and Encrinuraspis occur in South-East Asia, and Pliomerina and Heptabronteus in Kazakhstan. The presence of Toernquistia and possibly Pseudobasilicus and lUaenus (Parillaenus) suggests a less strong connection with northern Europe. Toernquistia may have had a wider distribution with its possible occurrences in Idaho and Nevada (Churkin 1963; Ross and Shaw 1972). The records of Remopleurides, Amphilichas, and Sphaero- coryphe imply a relationship within the world-wide, perhaps equatorial, Remo- pleuridid province (or, as I would prefer, realm). It seems desirable, therefore, to regard the New South Wales ‘Caradoc’ fauna, with its close relationships to South- East Asia and Kazakhstan (Table 1), as belonging to the Heptabronteus- Pliomerina province, situated within the much larger, band-like, ‘equatorial’ Remopleuridid realm. The province appears to include only shelf margin-open sea faunas (Palmer 1972). WEBBY: ORDOVICIAN TRILOBITES 211 For more complete evaluation of the zoogeographical relationships it is neces- sary to take account of the facies control, mode of life, and tectono-environmental settings of the faunas. Of the fifteen families occurring in the entire New South Wales fauna, only five— the remopleuridids, scutelluids, illaenids, trinucleids, and encri- nurids— are found in both carbonate and graptolitic shale facies. Of the genera, Remopleurides, Heptabronteus, Illaenus {Parillaenus)! , and possibly Encrinuraspis are found in both facies, though only Remopleurides and Heptabronteus are well represented in both (Table 2). At the species level there are no facies breakers. TABLE 2. Number of species represented in carbonate and graptolitic shale facies of the Gisbornian-Eastonian succession of central New South Wales. Species of Triarthnis, Shumardia, and Geragnostiisl are inferred to have had a mainly pelagic mode of life, whereas all others are thought to have been benthonic in the adult stage. Endemic elements of the fauna are indicated by asterisks. Remopleurides Platform or rise (carbonates) 3 Slope and trough (graptolitic shales) 2 Amphilichas 2 — Sphaerocoryphe 1 — Pseudobasilicusl 2 — Illaenus {Parillaenus)! 1 1 Toernquistia — 1 Heptabronteus 1 1 P Homer ina 2 — Encrinuraspis 2? 1 Parkesolithus* — 2 Malongullia* — 1 harpid gen. nov.* 1 — Geragnostusl — 1 Shumardia — 1 Triarthrus — 1 The carbonate facies of the shallow-water platform or rise exhibits a fauna inter- preted as including only benthonic elements. The abundant cosmopolitan genus Remopleurides is represented in both carbonate and graptolitic shale facies, but is not regarded as a pelagic element because no individual species of the genus is found in both facies. R. saenuros Webby from the unnamed limestone at Billabong Creek occurs in an only slightly older horizon than R. exallos Webby from shales of the Malongulli Formation in the Cliefden Caves and Cheeseman’s Creek areas (text- figs. 1, 3), yet is markedly different from it. The New South Wales species of Remo- pleurides preserved in shales typically possess a much finer pattern of ornamentation than those occurring in the limestones (Webby 1973). If they were pelagic forms one would expect to find a much less well-defined facies control of the individual species. Two distinct types of association are represented in the deeper, graptolitic shale facies. The first is best developed in shales at the base of the Malongulli Formation, in beds directly overlying the Cliefden Caves Limestone (text-fig. 4, section C2), considered to have formed in deeper waters of the slope immediately adjacent to the rise. Among elements of the fauna are sighted, zoogeographically relatively 212 PALAEONTOLOGY, VOLUME 17 restricted to endemic forms, including species of Malongullia, Parkesolithus, Encri- nuraspis, and Heptabronteus. This zoogeographically restricted fauna is inferred to have had a benthonic mode of adult life. The second type of association occurs in thick, interbedded graptolitic shale, chert, tuff, and volcanic sequences (text-fig. 4, sections Ci, Di, and D2). It comprises relatively small, mainly blind, relic elements of long-lived, cosmopolitan stocks, including species of Shumardia, Geragnostusl, and Triarthrus. This is typically an open-sea or oceanic fauna, with representatives of agnostids (Robison 1972) and olenids (Opik 1963) regarded as having had a pelagic mode of life. The elements of this fauna may have lived in even deeper waters of the trough. Triarthrus, the last of the olenids, is recognized as an extremely wide- spread stock in dark shales of Caradoc age, occurring throughout the Remopleuridid realm and also in New Zealand (Skwarko 1962) and South America (recorded by Harrington and Leanza 1957, as Porterfieldia Cooper, 1953, which was shown to be a junior synonym of Triarthrus by Whittington 1957, and an invalid genus by Whit- tard 1961). Shumardia and Geragnostus are also widespread and long-ranging genera, though apparently much less common in beds above the base of the Nemagraptus gracilis Zone. Whittington (1966) has previously noted these forms to be geo- graphically widely distributed elements. In the inferred benthonic assemblage in the shales of the slope, all the generic components, apart from Remopleurides, are restricted to having relationships with South-East Asia, Kazakhstan, or northern Europe, or are endemic. In the fauna of the carbonate facies of the adjacent rise, on the other hand, there is a markedly higher proportion of widespread elements of the Remopleuridid realm, with occur- rences of Remopleurides, Amphilichas, and Sphaerocoryphe, and possible northern European relationships suggested by records of forms tentatively assigned to Pseudo- basilicus and Illaenus (Parillaenus). The shale fauna of the slope has two endemic genera in a total of seven inferred benthonic genera, and the carbonate fauna of the rise, one endemic genus in a total of nine genera (Table 2). In terms of abun- dances of generic components at particular localities, not greatly different in age, the richly fossiliferous representatives of the shale type at Trilobite Hill and near-by Copper Mine Creek may be compared with a similarly rich occurrence of the car- bonate type at Billabong Creek, and demonstrates an even greater degree of endem- ism in the shales of the slope (Table 3). The overall percentage of endemism based on abundances of generic components is 47% in the shale, and 10% in the carbonate type. The benthonic assemblage in the shales of the slope is not significantly less diverse than the fauna in the carbonates of the rise, but it has a markedly higher proportion of endemic elements. These faunas of off-shore parts of the Lachlan Geosyncline are considered to have occupied the tectono-environmental setting of an island-arc type of geosyncline (Mitchell and Reading 1969), separated from the continental mass by the inter- vening Wagga Trough, a possible marginal sea like the present Sea of Japan (Pack- ham and Falvey 1971). The fauna in the carbonates of the rise includes a number of forms which suggest ease of widespread dispersal within the Remopleuridid realm, probably along migration routes in shallow seas bordering island chains and fring- ing continental masses, aligned east-west, in low latitudes, and with no significant climatic barriers (Valentine 1971). The presence of geographically more restricted WEBBY: ORDOVICIAN TRILOBITES 213 TABLE 3. Percentage of endemism assessed on the basis of abundances and diversity of generic components in faunas of representative occurrences of shale and limestone facies. Most abundant element of fauna scores 5 points; moderately common, 4 points; common, 3 points; rare, 2 points; and very rare occurrence, 1 point. Abundance scores for the one, two, or three genera exhibiting endemism or a particular zoo- geographical relationship are shown in columns A-D. Number of endemic genera or genera showing particular zoogeographical relationships given in brackets. Lithology Locality A Endemic B SE. Asia and Kazakh- stan C Northern Europe D Widespread in Remo- pleuridid realm Total abundance score and total number of genera in brackets Percentage of endemism in terms of abun- dance score. Percentage of endemic genera in brackets Shale Trilobite Hill and Copper Mine Creek 9(2) 7(2) -(-) 3(1) 19(5) 47(40) Limestone Billabong 2(1) 7(2) 4(1) 8(3) 21(7) 10(14) Creek relationships in the inferred benthonic slope fauna seems to imply less easy routes of migration within the deeper slope environments. No endemic elements of the deeper shale environments are known to have invaded the shallow-water environ- ments of the rise. The close faunal relationships of the New South Wales island-arc ‘benthonic’ assemblages of both shale and carbonate facies with those of South- East Asia and Kazakhstan in Upper Ordovician times implies either former close links along a continuous chain of islands, or a much nearer association of the conti- nental blocks of Australia, and South-East and Central Asia than at present. DIMORPHISM Although it remains virtually impossible to prove sexual dimorphism in trilobites, even where different forms are found together on the same bedding surfaces, there is, nevertheless, growing circumstantial evidence for dimorphism in some groups, with a steadily growing number of references to particular examples (Opik 1958; Whittington 1959, 1963, 1965; Hu 1964, 1971; Selwood 1965; Selwood and Burton 1969; Clarkson 1969; Webby 1973). One of the most convincing examples is that given by Opik (1958) for a species of Redlichia, R. forresti (Etheridge) from the lowermost Middle Cambrian of Western Australia and the Northern Territory. Opik recognized two distinct forms in R. forresti, a common, large (average length, 100 mm), slender ‘female’ form with slender dorsal spines on the axis of the thorax, short pleural spines, an unusual pygidium with a poorly differentiated axis and pygidial doublure restricted to the lateral flanks, and a rare, small (15 mm in length) ‘male’ with stout dorsal spines, long pleural spines, and a pygidium having a normal annulated axis and a doublure continuous beneath the border. The possibility of dimorphism in the New South Wales representatives of Remopleurides has been discussed recently (Webby 1973). In the shales of the Malongulli Formation at two different localities, the raphio- phorid, Malongullia oepiki Webby, Moors, and McLean, 1970, has been found to be associated in the same beds with a much larger form exhibiting a Malongullia-iypc cephalon and thorax, but much expanded pygidium. At the Copper Mine Creek 214 PALAEONTOLOGY, VOLUME 17 (and near-by Trilobite Hill) locality, near Cliefden Caves, the smaller M. oepiki is much more common, and in the Mirrabooka locality of the Cheeseman’s Creek area, though the sample is small, there is a slight dominance of the larger form with the expanded pygidium. The two types clearly seem to represent dimorphs but have been described as separate morphological forms in the following pages (PL 30, fig. 12 ; PI. 31, figs. 4-12; PI. 32, figs. 1-6). The larger form is possibly the female of the species and is referred to as M. oepiki (dimorph B). It is on average four times larger than the possible male, M. oepiki (dimorph A), the original type material of the genus and species (Webby, Moors, and McLean 1970). On the cephalon of dimorph B, the preglabellar field is longer (sag. and exsag.) and exhibits a pair of small tubercles, and the posterior border is longer (exsag.) and more flattened. Anastomosing ridges (perhaps caecae) on the antero-lateral parts of the fixed cheek of dimorph B are not seen in dimorph A. The pygidium of dimorph B is fundamentally distinct in being much longer, rounded posteriorly and in having 16-20 axial rings, eight-ten pleural ribs, more conspicuous muscle attachment areas with up to seven pairs of oval muscle scars connected across the ring furrows, nine pairs of isolated scars on axial segments behind, and a suggestion of pygidial caecae in the posterior part of the pleural fields and behind the rounded tips of the axis (PI. 31, figs. 4, 10-11; PI. 32, figs. 5-6). In the smaller dimorph A, in contrast, the pygidium is shorter, sub- triangular, it has 8-9 axial rings, 5 pleural ribs, and 6-7 pairs of transversely elon- gated apodemal pits with pairs of muscle scars confined to just in front of apodemal pits of the first two axial rings. The first pair of muscle scars are connected across the ring furrow, and the second pair, isolated (PI. 31, fig. 12; see also Webby, Moors, and McLean 1970, pi. 125, figs. 1, 5-7, 11-12). SYSTEMATIC DESCRIPTIONS Family OLENiDAE Burmeister, 1843 Genus triarthrus Green, 1832 Type species. T. beckii Green, 1832. Triarthrus sp. Plate 32, fig. 13 Material. One specimen (SUP 37999) from the Cheesemans Creek Formation in the disused quarry just east of Keenan’s Bridge, Cheeseman’s Creek area (Sherwin 1971). Another specimen (MMF 1 8638) of part of thorax and pygidium from the same locality and horizon is in collections of the Geological and Mining Museum, Geological Survey of New South Wales. Description. Only part of lateral border of left free cheek known. Thorax of fifteen segments; axial furrows deep, prominent; axis gently convex, slightly more than one-third total width, tapering very gently backward; each axial ring with large median tubercle and scattered very tiny tubercles to either side of it ; deeply indented articulating furrows, especially laterally, and well-developed articulating half rings, at least two-thirds length (sag.) of axial rings. Deep, prominent pleural furrows directed outward and backward from inner, anterior corner of pleurae towards outermost tip. Inner part of pleurae short (tr.), transverse, about one-third total WEBBY: ORDOVICIAN TRILOBITES 215 width of pleurae; outer part of pleurae, beyond fulcrum, relatively wide, directed outward and backward. Pygidium small, transversely elongated, with flattened out- line of posterior margin medially ; four (or possibly five) axial rings and small terminal piece; four pairs of pleural ribs; there appears to be a median tubercle on anterior axial ring of pygidium. Remarks. There is considerable resemblance between the thorax and pygidium of the holaspis of Triarthrus eatoni (Hall) from the Holland Patent Shale of Trenton, New York (Whittington 1957) and that of the incomplete Cheeseman’s Creek material. However, T. eatoni exhibits sixteen thoracic segments, and a relatively narrower axis with each axial ring apparently smooth apart from the slightly smaller median tubercle. In Australasia, Triarthrus has previously been recorded by Gilbert-Tomlinson (1961), and Whittington and Hughes (1972, p. 272) from the Llanvirn of the Can- ning Basin, and from the basal Gisbornian (Lower Caradoc) of New Zealand (Skwarko 1962). Family scutelluidae Richter and Richter, 1955 Genus heptabronteus gen. nov. Type species. H. atavus sp. nov. Diagnosis. Scutelluid genus with forwardly widening glabella and axial furrows deflected outwards, apparently dying out anteriorly; no clearly differentiated pre- glabellar area on anterior margin; large, raised, curved eyes situated posteriorly; relatively broad area of fixed cheek between outwardly diverging anterior branch of facial suture and axial furrow, sometimes exhibiting eye ridge directed backward and outward from axial furrow on to palpebral lobe. Sharp genal angle, not pro- longed into spine; posterior part of fixed cheek with short posterior border furrow just inside margin, dying out towards intersection with posterior branch of facial suture and also inwards. Rostral plate, hypostome, and thorax as in Kosovopeltis Snajdr. Pygidium semi-elliptical, having very short, gently elevated, subtriangular axis with up to four pairs of poorly defined oval muscle impressions set well inside shallow axial furrows. Seven pairs of depressed, lateral pleurae, and usually un- paired, median pleura; pleural furrows weak, tend to die out near axial furrow and toward border. Pygidial doublure widest posteriorly, extending inwards to about one-half total length (sag.) of pygidium, but occasionally slightly more, almost to posterior tip of axis; doublure narrows anteriorly, with inner margin evenly curved and directed towards fulcra on anterior margin. Discussion. Heptabronteus bears the closest resemblance to the Silurian genera Kosovopeltis and Planiscutellum R. and E. Richter, regarded by Snajdr (1960) as belonging to the group of ‘primitive’ scutelluids. It most closely compares with Kosovopeltis (type species, K. svobodai Snajdr, from the Upper Silurian of Bohemia), having a very similar glabella, including the arrangement of lateral glabellar muscle impressions, a similar, large eye placed posteriorly, rostral plate, hypostome, and thoracic segments. It differs principally in not having the genal angle prolonged into a spine, exhibiting a relatively longer pygidial doublure, covering virtually the 216 PALAEONTOLOGY, VOLUME 17 entire postrhachial area (only 0-42 to 0-6 of this area is occupied by the doublure of the type species of Kosovopeltis; see Campbell 1967, p. 12), and in having a less inflated and trilobed pygidial axis. Several pairs of muscle scars may show on the axis, but not the two well-defined longitudinal furrows seen on the axis of Kosovo- peltis. Also, larger specimens of Heptabronteus (notably in H. major sp. nov.) have a wider antero-lateral area of the fixed cheek, between axial furrow and anterior branch of the facial suture, and may exhibit a weakly developed eye ridge extending diagonally back into the palpebral lobe. Compared to Planiscutellum (type species, Bronteus planus Hawle and Corda, from the Middle Silurian of Bohemia; see Snajdr 1960), Heptabronteus has larger eyes and a rostral plate tapering more gently distally; it lacks a flattened, differen- tiated preglabellar area and axial rings on the pygidium, though pairs of muscle scars, possibly each pair confined to one segment, are seen. The pygidial doublure occupies a much greater part of the pleural regions posteriorly. Snajdr’s (1960, p. 233) claim that the pygidial doublure became enlarged in the course of phylogenesis in scutelluids, from the older, ‘primitive’, to the Devonian genera cannot be upheld. The earliest, ‘Caradoc’ representative of the older genera, Heptabronteus atavus sp. nov., has a doublure almost covering the entire postrhachial area, a condition not notably different from that found in the ‘advanced’ Devonian forms. None of the other Ordovician scutelluid genera exhibits seven pairs of pleurae on the pygidium. Eobronteus Reed (type species, Entomostracites laticauda Wahlen- burg, from the Boda Limestone of Dalarna, Sweden), differs in having a broad preglabellar area, a small, rounded lateral glabellar impression Ip, a diagonal furrow running backward and outward from the axial furrow across the anterior part of the fixed cheek, opposite the lateral glabellar impression Ip, another furrow curving backward and outward from the lateral muscle impression on the posterior part of the fixed cheek, broad-based genal spines, and six pairs of pleurae on the pygidium (Warburg 1925; Sinclair 1949; Snajdr 1960). Also, the hypostome seems to have a more rounded posterior outline, and less conspicuous flaring outward and upward anterior wings. Protobronteus Snajdr, 1960 closely resembles Eobronteus, but lacks a preglabellar furrow delimiting a broad preglabellar area. It is restricted to one, incompletely known, species, P. reedi (Sinclair), from the Middle Trenton of Quebec, and should perhaps be regarded as a junior synonym of Eobronteus. A broadening of the generic conception of this latter genus to include forms without a differentiated preglabellar area would seem preferable to retention of a genus based on such a small difference. Oetobronteus Weber (type species, O. klwdalevitclii Weber, from the Upper Silurian of the Urals) has even less resemblance to Hepta- bronteus. It exhibits a pygidium with eight pairs of pleurae, pleural furrows which slightly widen and deepen distally to immediately inside smooth border of variable width, and an axis which may be vaguely trilobed or shows segmentation. On the glabella there are conspicuous exsagittally elongated, anteriorly-narrowing, fused impressions of lateral glabellar furrows 2p and 3p, a tiny median node at the rear edge of the occipital ring, and prominent curved eye ridges running on to the pal- pebral lobes of the cheeks (Snajdr 1960). Heptabronteus could have been derived from Eobronteus since the latter appears somewhat earlier, in the Chazyan of eastern North America (Shaw 1968), but it WEBBY: ORDOVICIAN TRILOBITES 217 seems more likely that both developed from a common ancestor among early Middle Ordovician styginids like Bronteopsis (Whittington 1950; Skjeseth 1955; Whittington in Moore 1959, pp. 365-367). It seems probable that Heptabronteus, which is represented by two stratigraphically distinct ‘Caradoc’ species, H. atavus and H. major, in New South Wales, and by Bronteus romanovskii Weber in the Caradoc and Lower Ashgill of Kazakhstan (Tschugaeva 1958), is the forerunner of the conservative, rather unspecialized Silurian stock with seven pairs of pygidial pleurae, chiefly the genera Planiscutellum and Kosovopeltis. The Silurian forms evolved from the Heptabronteus stock occupy- ing the restricted Kazakhstan-Australian region {Heptabronteus- Pliomerina pro- vince) during the Caradoc and Ashgill, and seem to have spread out to achieve a moderately cosmopolitan distribution by Middle and Upper Silurian times (Snajdr 1960; Campbell 1967). From these arose a considerable number of specialized, relatively short-lived, endemic genera in the Lower Devonian, particularly in Bohemia (Snajdr 1960). Heptabronteus atavus sp. nov. Plate 28, figs. 1-18 Material. Holotype (SUP 29908) and twenty-four paratypes (SUP 18903, 18906, 28900, 28931-28936, 28938, 28942, 28949, 29902-29903, 29905-29907, 29909a, 29911-29915, 29943) from the ‘lower coral’ unit on Fossil Flill, lower part of the Cliefden Caves Limestone. Also, one paratype (SUP 29901 ) from the ‘mixed fauna’ unit east of Fossil Hill, lower part of the Cliefden Caves Limestone. Description. Cranidium gently convex transversely and longitudinally, becoming more strongly convex at anterior margin, with slight overhang of margin. No trace of preglabellar furrow separating frontal region of cranidium. Glabella narrowest opposite lateral muscle impressions on fixed cheeks, expanding forwards to become widest anteriorly, in front of lateral glabellar impressions Ip', widens posteriorly across occipital ring to posterior margin. Axial furrow deeply impressed through- out most of its course but appears to die out before reaching antero-lateral margin. Very weak sag (rather than furrow) extends diagonally backward and outward from anterior end of axial furrow towards anterior branch of facial suture. Occipital furrow deep, prominent; deflected forwards and less deeply impressed medially; deep, slot-like lateral occipital ?muscle impressions in furrow laterally. Occipital ring, gently convex transversely; slopes gently forward into occipital furrow, and is more steeply inclined backward. Very gentle backward arch of posterior margin. Glabella exhibits three pairs of faint, lateral glabellar impressions (PI. 28, fig. 1). Lateral glabellar impression Ip is an elongated, slot-like impression extending from just inside axial furrow inwards and forwards, parallel to antero-lateral margin. Oval-elongate lateral glabellar impression 2p is situated further inside axial furrow, near mid-length of glabella (excluding occipital ring); Ip is almost half-way between 2p and antero-lateral margin; 3p is placed on axial furrow, and is larger than lateral muscle impression on fixed cheek, but much less well defined ; placed so that posterior edge of 3p is directly opposite anterior edge of lateral muscle impression. Large, oval lateral muscle impression deeply imprinted on fixed cheek; situated a little way in front of occipital furrow. Terrace lines subparallel to margin in anterior part 218 PALAEONTOLOGY, VOLUME 17 of cranidium. Fixed cheek flattened transversely towards palpebral lobe; sloping gently backwards, and more steeply forwards towards antero-lateral corner. Palpe- bral lobe crescent-shaped, slightly raised. Anterior branch of facial suture directed forward and outward almost parallel with axial furrow, but some distance away from it, and continues to antero-lateral margin, where it turns sharply inwards and downwards to continue course between doublure and rostral plate. Behind palpe- bral lobe, posterior branch of the suture curves sharply downward and outward on to the posterior margin well inside genal angle. No trace of eye ridges seen. Width across one fixed cheek at level of palpebral lobe just less than least width across glabella measured between lateral muscle impressions. Free cheek with large, posteriorly placed, raised, crescent-shaped eye lobe, and broad, flattened, outer surface, or platform. Convex visual surface extends around lobe from forward (and slightly inward) to outward and backward. Overall align- ment of eye seems to be slightly oblique, more or less parallel to antero-lateral margin of free cheek. There may be slight forward tilt of visual surface, giving highest elevation of lobe to the rear. Antero-lateral margin of free cheek has coarse raised lines running along it, gradually becoming finer towards genal angle (PI. 28, fig. 6). Dorsal surface of flattened outer part of free cheek covered by wavy, anastomosing, rather wide-spaced lines, in the anterior half deflected backwards and outwards, and in the posterior half more typically arranged transversely with backward deflec- tion marginally. Surface between lines covered with tiny pits. Sharp discordance between coarse, continuous, raised lines on antero-lateral margin and fine, wavy transverse-diagonal lines of outer part of free cheek. Genal angle sharply pointed but not prolonged. Posterior margin wide, almost straight; no trace of posterior border furrow. Doublure occupies much of the undersurface of free cheek, with broadly spaced terrace lines similar to those on rostral plate, approximately parallel with lateral margin, except for inward deflection of lines towards posterior margin EXPLANATION OF PLATE 28 Figs. 1-18. Heptabronteus atavus sp. nov., from the lower part of the Cliefden Caves Limestone. 115, 17-18 from ‘lower coral’ unit, Fossil Hill; 16, from ‘mixed fauna’ unit east of Fossil Hill. 1, dorsal view of latex impression of cranidium; holotype SUP 29908, x 8. 2, dorsal view of incomplete crani- dium of paratype SUP 28931, x4. 3, dorsal view of fragmentary cranidium of paratype SUP 18903, X 3. 4-5, anterior and dorsal views of incomplete cranidium of paratype SUP 29907, x 4. 6, dorsal view of latex impression of left free cheek, SUP 29909a, x 5. 7, dorsal view of small left free cheek, paratype SUP 29914, x 6. 8, oblique lateral view of latex cast of exfoliated right free cheek, paratype SUP 29912, x4. 9, oblique dorso-lateral view of right free cheek of paratype SUP 28942, x4. 10, ventral view of incomplete rostral plate, paratype SUP 29902, x 5. 11, ventral view of latex cast of part of rostral plate, paratype SUP 18906, x 4. 12, ventral view of latex impression of hypostome, para- type SUP 28935, X 6. 13, oblique ventral view of latex cast of fragmentary hypostome, paratype SUP 28936, X 5. 14, dorsal view of incomplete thoracic segment, paratype SUP 28938, x4. 15, dorsal view of exfoliated pygidium showing extent of doublure, paratype SUP 28949, x 3. 16, dorsal view of pygidium, paratype SUP 29901, x3. 17, dorsal view of latex cast of partly exfoliated pygidium, SUP 28900, x4-5. 18, ventral view of incomplete pygidium showing part of broad, gently concave doublure, paratype SUP 29905, x 3. Figs. 19-20. Heptabronteus major sp. nov., from the Malongulli Formation. 19, dorsal view of latex cast of thoracic segments from near Mirrabooka homestead, Cheeseman’s Creek area, paratype SUP 20914, x4. 20, dorsal view of compressed pygidium from Trilobite Hill, paratype SUP 19934, x 1-5. PLATE 28 WEBBY, Heptahronteus 220 PALAEONTOLOGY, VOLUME 17 where doublure widens behind eye. Anteriorly, doublure has convex rolled appear- ance, but becomes progressively more flattened posteriorly. Rostral plate large, very broad, gently convex (sag. and exsag.) defined by in- wardly and backwardly directed connective sutures; gently curved anterior and posterior margins giving most expanded portion medially; in addition to backward curvature of posterior margin, it is also gently arched downwards medially. Anterior branches of facial suture turn sharply inward at antero-lateral margin into long, gently curved, horizontal rostral suture bounding anterior margin of rostral plate. Terrace lines prominent, and on outer (ventral) surface arranged parallel to posterior and anterior margins. Hypostome with shield-shaped outline, posterior border wider (sag. and exsag.) than lateral border ; posterior and lateral margins almost straight, with sharp postero- lateral angle between them; slightly projecting shoulder. Gently convex middle body divided broadly by diagonal, middle furrows into larger, anterior, and smaller, posterior crescent-shaped parts. Smooth-surfaced, crescentic macula occupies area adjacent to middle furrow; area between macula and lateral border furrow has terrace lines in continuity with those of posterior part of middle body. Faint sugges- tion of tubercle occasionally seen on raised postero-median edge of macula. Posterior border furrow runs in gentle curve in continuity with lateral border furrow. Faint pit in furrow at postero-lateral corner on each side. Lateral border furrows continue to diverge anteriorly, finally dying out in triangular base of anterior wing. Anterior margin curves gently downwards and backwards medially. Anterior wings large, triangular, directed upward and outward. Coarse terrace lines, similar to those on rostral plate, extend across hypostome; curve down off anterior wings, across middle body with convexity to the rear, and parallel to lateral and posterior borders. Only one fragmentary thoracic segment seen (PI. 28, fig. 14). Axis gently convex, both transversely and longitudinally, becoming moderately steep (about 55° to horizontal) near axial furrow. Terrace lines concentric about postero-median part of axis. Pleura consists of relatively short, horizontal, transverse, gently convex (exsag.) inner, and gently inclined (at about 40° to horizontal), slightly longer (tr.) outer parts. Terrace lines inclined diagonally forward and slightly outward on inner part, more obliquely outward on outer part except for distal extremity where lines become subparallel to outer margin of thorax. Pygidium semi-elliptical, gently convex, with evenly rounded lateral and posterior margins, and moderately sharply rounded antero-lateral corners; almost straight, transverse anterior margin except for tongue-like forward extension of gently convex articulating half ring. Axis subtriangular, occupying about one-third of total length (sag.) of pygidium, and about one-third of width at anterior margin. No segmenta- tion seen on axis behind anterior articulating furrow. Axial furrows very slightly impressed. Pleural regions broad, gently convex, with seven pairs of weakly developed pleurae and median undivided pleura. Weakly imprinted pleural furrows which tend to fade out near axial furrow and also toward margins. Width of median un- divided pleura near margin usually similar to width of adjacent paired pleurae. Terrace lines more or less transverse across entire dorsal surface of pygidium, with slight forward curvature near lateral margins. Doublure occupies about two-thirds of the total sagittal length of pygidium, virtually entire area of pleural regions behind WEBBY: ORDOVICIAN TRILOBITES 221 axis, but narrows approaching anterior margin, with inner edge of doublure meeting margin at fulcrum (PI. 28, fig. 15); medially, inner edge of doublure curved evenly behind axis. Remarks. Bronteus romanovskii Weber from the Caradoc to Lower Ashgill of Kazakhstan (Tschugaeva 1958) bears very close similarities to Heptabronteus atavus, and is undoubtedly congeneric. The only significant differences between the two species seem to be suggested by the course of the facial suture. From the outline of the suture around the palpebral lobe, it seems likely that H. romanovskii had a slightly smaller eye (Weber 1948, text-fig. 16). Secondly, though better depicted in Tschugaeva’s (1958) material, the anterior branch of the suture in H. romanovskii seems to be much less parallel to the axial furrow than in H. atavus, or more exsagit- tally than diagonally (forwardly and outwardly) directed. Heptabronteus major sp. nov. Plate 28, figs. 19-20; Plate 29, figs. 1-6 Material. Holotype (SUP 20903) and three paratypes (SUP 18904, 20912a, 20913) from the Malongulli Formation at Copper Mine Creek, and two paratypes (SUP 19934, 19936) from the same horizon at Trilobite Hill. Also two paratypes (SUP 20914, 26935) from the Malongulli Formation near Mirrabooka homestead, north of Cheeseman’s Creek. Comparative description. This larger species of Heptabronteus has a flattened crani- dium, with evenly rounded anterior margin and relatively wide fixed cheeks. Axial furrow is less deeply impressed throughout its course except at level of lateral muscle impression on fixed cheek. Lateral glabellar furrows only faintly impressed. Ip and 2p transversely elongate-oval, set inside axial furrow; 2p slightly larger than Ip, and placed just behind mid-length of glabella (excluding occipital ring) ; Ip almost midway between 2p and anterior margin of cranidium ; large 3p impression on side of axial furrow very ill-defined, about midway between 2p and occipital furrow. Occipital furrow sharp and deeply impressed medially, but widens (exsag.) and shallows laterally into attachment area with subtriangular ?muscle scars. Lateral muscle impressions on fixed cheeks only faintly shown. Fixed cheek relatively broad, about equal or slightly wider at level of palpebral lobes to width across narrower, posterior part of glabella ; much wider across fixed cheek in front of palpebral lobe than in H. atavus. Weakly impressed eye ridge runs diagonally backward and out- ward from axial furrow just in front of lateral glabellar impression 2p on to palpebral lobe (PI. 29, fig. 1). Posterior branch of facial suture curves outward and then back- ward (and slightly inward) around small posterior projection at distal end of fixed cheek adjacent to posterior margin; short discontinuous posterior border furrow developed just inside margin only near distal end of fixed cheek (PI. 29, fig. 4). Surface of cranidium covered by fine slightly anastomosing terrace lines, trans- versely arranged near anterior margin and on glabella, but deflected backward at axial furrow; on posterior part of glabella just in front of occipital furrow and on occipital ring form concentrically arranged, convex forward arcs; on fixed cheeks, somewhat anastomosing terrace lines form concentric arcs centred on palpebral lobe, running in increasingly large arcs forward and inward. Hypostome, though somewhat crushed, exhibits relatively larger anterior lobe 222 PALAEONTOLOGY, VOLUME 17 of middle body, and rather inconspicuous macula situated adjacent to inner part of middle furrow ; lateral border furrow deep, becoming shallower postero-laterally, and defining narrow, raised lateral border and its continuation, long (sag.), gently convex posterior border; postero-lateral margin more rounded than in H. atavus. External surface covered by fine, concentrically arranged lines, much finer than in H. atavus. Only four thoracic segments seen in articulation. Axis relatively narrow, gently convex, estimated to occupy between one-quarter and two-sevenths of total thoracic width; axial furrows distinctly impressed and have slight zigzag outline between articulated segments. Articulating furrows broad, well-defined, slightly obliquely inclined to posterior margin, giving axial ring slight hour-glass outline. Pleurae gently convex (exsag.), almost flat (tr.), parallel-sided from axial furrow to fulcrum, with horizontal hinge line; beyond fulcrum a barely discernible expansion, and then tapers distally to backwardly turned, pointed pleural tip. Anterior and posterior margins of pleurae with moderately sharp edges; transverse lines may show on posterior edge of outer part of pleura. Terrace lines arranged concentrically out from centre near middle-rear of each axial ring, though there may be additional more or less transverse lines along rear edge of axial ring; also transverse lines on articulating half ring. Terrace lines directed forward and outward on inner part of pleura, but beyond fulcrum, become longitudinally directed and on tapering pleural tips may be slightly inwardly and forwardly inclined. Pygidium relatively longer, narrower and flattened (probably in part by compres- sion), with well-rounded antero-lateral corners. Relatively small, subtriangular axis bounded by weak axial furrows ; occupies between one-quarter and two-sevenths of total sagittal length of pygidium, and between one-fifth and one-quarter of maximum pygidial width; tongue-like forward extension of gently convex articulating half ring, crossed by ornamentation of fine lines; articulating furrow broad, moderately deep, and smooth. Four (or ?five) pairs of elongate to oval muscle scars developed on axis behind articulating furrow. Seven lateral pairs of pleurae and broader, un- paired, median pleura; pleurae flattened to gently convex, intersected by moderately deep, narrow pleural furrows which weaken toward axial furrow and approaching margin. Doublure extends from fulcrum outward to lateral margins, but posteriorly increases in width to cover large part of pleural regions extending to more than one- third of total length of pygidium ; in small specimen (PI. 29, fig. 5) extends one-half total pygidial length; prominent, wide-spaced terrace lines more or less parallel to posterior and lateral margins. To summarize, this larger species of Heptabronteus, H. major, may be distinguished from the type species, H. atavus, by having a cranidium with an eye ridge crossing the fixed cheek diagonally on to the palpebral lobe, a relatively wider area of the fixed cheek in front of the eye ridge, a less conspicuous lateral muscle impression, less prominent axial furrows anteriorly, less deeply indented lateral ends to the occipital furrow, a hypostome with a larger, rounded anterior lobe of the middle body, less conspicuous maculae, a more rounded postero-lateral margin, and finer terrace lines on the external surface, and a relatively longer and narrower pygidium with the axis occupying a smaller proportion of the entire area of the pygidium, pairs of muscle scars on the axis, more rounded antero-lateral corners, and a relatively wider, unpaired, median pleura. WEBBY: ORDOVICIAN TRILOBITES 223 Family dimeropygidae Hupe, 1953 Genus toernquistia Reed, 1896a Type species. Cyphaspis {Toernquistia) nicholsoni Reed Toernquistia arguta sp. nov. Plate 29, figs. 7-9 Material. Holotype (SUP 26931) and two paratypes (SUP 26930, 26932) from the Malongulli Formation near Mirrabooka homestead, 2 miles north of Cheeseman’s Creek Post Office. Description. Glabella narrowing forward, rounded anteriorly, convex; outlined by deep, continuous axial and preglabellar furrows; maximum height near mid-point. Pair of weakly impressed lateral glabellar furrows Ip extend obliquely inwards and backwards at about 45° to exsagittal line, to isolate small, subtriangular lateral glabellar lobes Ip in postero-lateral corners of glabella (PI. 29, fig. 9). Glabella esti- mated to be slightly more than one-half total length of cephalon, and to have width across posterior part somewhat less than one-third total width of cephalon. Glabella measures from 0-5 to 11 mm wide and 0-6 to 1-2 mm long (sag.). Deep, broad occi- pital furrow with very slight forward convexity. Occipital ring convex, very short, with lenticular dorsal outline; posterior margin convex backwards. Surface of occi- pital ring not well enough preserved to confirm presence or absence of median tubercle; deeply impressed pits on axial furrow opposite lateral ends of occipital ring, represent large axial sockets for articulation of first thoracic segment. Preglabellar field moderately convex, sloping forward and downward to anterior furrow, and more steeply backward into preglabellar furrow; bounded by diagonal furrows which extend outward and slightly forward, weakening distally and curving forward to merge with anterior branches of facial suture (PI. 29, fig. 8). Deep median pit in preglabellar furrow has sagittally aligned, slot-like depression extending for- ward from it almost halfway across preglabellar field. Anterior border gently convex, rim-like, separated from preglabellar field by rather shallow anterior furrow. Fixed cheeks convex, moderately broad, more or less L-shaped, slightly less elevated than glabella, and separated by diagonal furrow from preglabellar field, and by deep posterior furrow from posterior border; palpebral lobe narrow, raised, but in- completely preserved. Small granules cover dorsal surface of glabella, preglabellar field, and fixed cheeks. Remarks. T. arguta may be distinguished from all the known species of Toernquistia recorded from the Caradoc and Ashgill of Britain and Sweden (Warburg 1925; Thorslund 1940; Whittington 1950; Dean 1962). T. nicholsoni Reed, the type species, from the Keisley Limestone (Ashgill) of the Cross Fell Inlier, northern England, exhibits close resemblances, but the occipital ring is longer (sag.), proportions across the glabella and fixed cheeks are different, with the fixed cheeks relatively narrower, lateral glabellar lobes Ip are not shown to be differentiated, and a coarser granula- tion is developed on the dorsal surface (Whittington 1950). On the other hand, a pair of small, subtriangular lateral glabellar lobes Ip are reported by Warburg (1925) on Swedish specimens of T. nicholsoni from the Boda Limestone of Ashgill age. T. reedi Thorslund, 1940 from the Lower Chasmops Limestone (4b/3) of Jemtland, 224 PALAEONTOLOGY, VOLUME 17 Sweden, also has much narrower fixed cheeks, but similarly shows lateral glabellar lobes Ip. T. translata (Reed) from the Balclatchie Group (Lower-Middle Caradoc) of Girvan, Scotland (Reed 1904, 1931) is in need of revision. It apparently has a less well-defined median, slot-like depression on the preglabellar field, a relatively shorter (sag.), less anteriorly tapering glabella, and slightly larger lateral glabellar lobes Ip. The diagonal, antero-lateral furrows in T. depressa Warburg from the Boda Limestone of Dalarna, Sweden, are only represented as very faint traces (Warburg 1925). Another species of Toernquistia, T. shlygini Weber, is recorded from the Upper Ordovician of Kazakhstan. However, it has a much longer (sag.) occipital ring, relatively narrower fixed cheeks, and a longer (exsag.) posterior border (Weber 1948). Toernquistial idahoensis Churkin from the late Middle-Upper Ordovician of Nevada and Idaho may be distinguished by the longer (sag.) preglabellar field and less conspicuous slot-like preglabellar depression. Family trinucleidae Hawle and Corda, 1847 Genus parkesolithus Campbell and Durham, 1970 Type species. P. gradyi Campbell and Durham, 1970. Diagnosis. Trinucleid genus with eye tubercles in adult; glabella with furrow Ip represented by large, elongate-triangular depressed area containing deeper postero- lateral and antero-median pits, and virtually imperceptible furrows 2p and 3p\ occipital spine absent ; fringe of two E arcs and from two to five continuous I arcs, with possible addition of further two I arcs laterally; auxiliary pits may occur in E2, occasionally suggesting discontinuous E3 arc laterally. Greatest regularity of pitting in antero-median parts of inner, I arcs, and usually extends about two-thirds of total distance around inner edge of fringe from mid-line to posterior border; outer I arcs, Ii and U may have irregular pitting, even antero-medially; addition of new EXPLANATION OF PLATE 29 Figs. 1-6. Heptahronieus major sp. nov., from the Malongulli Formation. 1, 4-6, from Copper Mine Creek; 3, from Trilobite Hill; 2, 7, from near Mirrabooka homestead, north of Cheeseman’s Creek. 1, dorsal view of latex impression of cranidium, holotype SUP 20903, x 3. 2, dorsal view of latex cast of part of exfoliated pygidium, paratype SUP 26935, x 3. Note two sets of terrace lines, one on dorsal surface, and the other, laterally, on doublure. 3, dorsal view of latex impression of pygidium, paratype SUP 19936, x 2. 4, dorsal view of latex impression of part of cranidium and four isolated thoracic segments (reversed orientation), paratype SUP 20912a, x2-5. Note also small, left free cheek of Remop leur ides exallos Webby. 5, dorsal view of part of small pygidium, paratype SUP 20913, x 4. 6, ventral view of compressed, rather poorly preserved hypostome, paratype SUP 18904, x 3. Figs. 7-9. Toernquistia arguta sp. nov., from the Malongulli Formation near Mirrabooka homestead, north of Cheeseman’s Creek. 7, dorsal view of internal cast of cranidium, paratype SUP 26930, x 10. 8, dorsal view of internal mould of small cranidium, paratype SUP 26932, x 15. 9, dorsal view of in- ternal cast of cranidium, holotype SUP 26931, x 10. Figs. 10-1 1. Parkesolithus dictyotos sp. nov., from the Malongulli Formation. 10, dorsal view of latex impression of external mould of small specimen from near Mirrabooka homestead, north of Cheese- man’s Creek, paratype SUP 26905, x 6. 11, dorsal view of latex cast of external mould of part of holo- type, SUP 26920, X 5, from Copper Mine Creek. PLATE 29 WEBBY, Heptahronteiis, Toernquistia, Parkesolitims 226 PALAEONTOLOGY, VOLUME 17 I arcs on external side of innermost arc. Ei and E2 separated by sharp ridge on upper lamella ; weak girder on lower lamella of equal development to Ei_2 pseudo- girder. Thorax of six segments. Pygidium subtriangular, broader than long, with eight or more weak axial rings, weakly furrowed to smooth pleural field, and pro- minent, raised posterior border. Discussion. The generic diagnosis proposed by Campbell and Durham (1970) has been widened to accommodate P. dictyotos sp. nov., a species exhibiting ornamenta- tion on the glabella, cheeks, and thorax, additional I arcs, up to five of which are continuous and usually a further two developing laterally, greater irregularity of pitting in C, and I2, and sometimes I3 arcs, numerous additional pits in E2 occasion- ally creating the impression of a third row, and fewer, usually five, faint pleural ribs on the pygidium. Campbell and Durham (1970), referring to the nature of the pit arrangement on the fringe of P. gradyi, noted the continuity of I4 and interruption of I3 in front of the glabella. A similar pattern is seen in P. dictyotos, though interruption of I arcs is not confined to the area in front of the glabella. The innermost I arc appears to be continuous around the inner margin of the fringe, and new, incomplete I arcs (usually I4 and I5) are added on its external side between the mid-line and posterior border. The slightly modified notation proposed by Ingham (1970) for use in de- scribing the Tretaspis fringe may also be applied to Parkesolithus. The new I arcs are apparently inserted on the external side of the innermost arc, supposedly the parent arc. The innermost arc is left unnumbered and given the symbol thus solving the problem of its changing I number around the inner margin. But the over- all variability in the arrangement of pitting on the fringe of Parkesolithus individuals leaves considerable doubt as to whether the conventional fringe formula is the most satisfactory method of representation. Perhaps construction of fringe maps for each individual would provide a more satisfactory means of comparison of these forms. In addition to possible relationships with Cryptolithus Green and Broeggerolithus Bancroft, discussed by Campbell and Durham (1970), Parkesolithus may be allied to Lloydolithus Bancroft. The type species, L. Iloydi (Murchison), from the Llandeilo Series of Wales and the Welsh Borderland (Whittard 1958) has some similarities, showing two E arcs, with variable numbers of auxiliary pits creating the impression of three E arcs in a few radii, complete l^.j arcs with sometimes an incomplete Ig arc, and auxiliary pits appearing in Ij, I2, and on the Ii_2 and l2_3 pseudogirders. But the prominent eye ridges in younger forms of L. Iloydi become reduced to ‘nodules’ adjacent to the curved axial furrows in the adult, whereas the eye tubercles of adult forms of Parkesolithus are well developed. Furthermore, L. Iloydi differs from Parkesolithus in lacking a clearly differentiated occiput and a sharp Ei_2 ridge on the upper lamella, in having a narrow preglabellar field, an occipital spine, an over- all greater regularity of pitting on the fringe, especially in E and E, a more pro- minent girder on the lower lamella, a less-pointed genal prolongation, and a greater number of stronger pleural ribs and axial rings on the pygidium. WEBBY; ORDOVICIAN TRILOBITES 227 Parkesolithus dictyotos sp. nov. Plate 29, figs. 10-11; Plate 30, figs. 1-11; Plate 3 1 , figs. 1-3 Material. Holotype (SUP 26920) and sixteen paratypes (SUP 26914a-b, 26915a-b, 26917-26919, 2692 la-b, 26922-26927, 26929) from the Malongulli Formation at Copper Mine Creek, and six paratypes (SUP 6910, 7914-7917, 13907) from the same horizon at Trilobite Hill, near Mandurama. Fifteen additional paratypes (SUP 26903, 26904a-c, 26905, 26906a-d, 26907-26909, 26911-26913) from the Malongulli Formation near Mirrabooka homestead, north of Cheeseman’s Creek Post Office. Description. Cephalon semicircular, twice to slightly more than twice as wide as long (sag.); somewhat crushed and flattened in specimens examined; about equal to length (sag.) of thorax and pygidium combined. Glabella pyriform, elevated above adjacent cheeks; apparently reaches maximum height near mid-length (sag.); outlined by broad, deep, gently curving axial furrows, which become very shallow to rear. Two small anterior pits mark sites of anterior fossulae, each at anterior end of axial furrow and separated by low ridge from fringe (PI. 30, fig. 1). No median tubercle. Glabella protrudes slightly in front of adjacent cheeks displacing and con- stricting inner arcs of fringe. Lateral glabellar furrow Ip consists of large, elongate to triangular depression containing a deeper, oval, postero-lateral pit inside axial furrow, just in front of apodemal pit in occipital furrow, and a sharp notch in the lateral slope of glabella antero-medially ; marks rear of anterior glabellar lobe. Furrows 2p and 3p extremely faintly impressed on slopes of glabella inside axial furrow; 2p developed as a large, very weakly imprinted, oval-shaped area in front of Ip (PI. 30, fig. 5), and 3p, as a small, faint indent situated about glabellar mid- length (sag.). Anterior glabellar lobe moderately elongate and swollen, occupies about five-sixths of length of glabella; occiput not very swollen, about one-half width of anterior glabellar lobe. Occipital ring short, convex, with sharp-ridged posterior margin, arching backwards sagittally; low, narrow ridges extend along inside margins of axial furrows from lateral ends of occipital ring, and die out opposite anterior edges of furrows Ip. Occipital furrow deepens laterally into oval apodemal pits, set inside axial furrows, and only separated from deep, postero- lateral pits of furrows Ip by low ridge. Cheeks broad, quadrant-shaped, gently convex, with steeper outer slopes; some- times exhibiting gently raised rim along antero-lateral margin adjacent to fringe. Prominent, rounded eye tubercles situated just behind glabellar mid-length, and out from axial furrow; in small specimens, curved, tapering elongation of eye tubercle to form incipient eye ridge running forward and inward toward axial furrow oppo- site furrow 3p (PI. 29, fig. 10). Faint, fine genal caecae may be seen branching out- ward across cheek from vicinity of eye tubercle; more prominent, undivided caeca extends postero-laterally from eye tubercle towards lateral pit (PI. 29, fig. 11). Raised posterior border slopes gently forwards and, from its sharply rounded crest, steeply backwards; transverse and horizontal inner course, but deflected downward and slightly backward laterally. Posterior border furrow also transverse and horizontal, extending from axial furrow to inner corner of widest part of fringe. Large lateral pit placed towards outer end of furrow, but separated from smaller pits of fringe. Glabella and cheeks have surface ornamentation of fine pits, becoming coarser and forming a more reticulate pattern posteriorly, towards posterior border furrow 228 PALAEONTOLOGY, VOLUME 17 on cheek, and along median part of glabella behind furrow 3p\ occiput has coarse reticulation which becomes more or less transversely aligned near occipital furrow. Inner part of upper lamella of fringe slopes gently downward ; outer part sharply reflected in prominent Ei_2 ridge. Fringe fairly constant in width, though narrowing slightly in front of glabella, widening across genal flange and tapering rapidly to- wards genal angle. Strong Ei_2 ridge persists around anterior and lateral margin but weakens and disappears approaching genal angle (PI. 30, fig. 11). Usually 7-8 arcs of pits developed, Ei_2 and Ii.^+n or Ii_5+n- Occasionally an additional arc, le, appears on external side of I„ arc in lateral areas (PI. 30, fig. 9). Outer I arcs, especially Ii_2, rather haphazardly arranged, but inner arcs (U-^) in anterior areas, moderately well ordered ; postero-laterally, progressively more I arcs, from outside inwards, assume irregular arrangement of pitting, completely losing their ordered radial arrangement from opposite widest part of cephalon excluding fringe to genal angle. External arcs usually have slightly larger and more widely spaced pits. Con- siderable numbers of auxiliary pits may occur in E2, especially laterally, giving impres- sion in some radii of discontinuous E3 arc. Occasional auxiliary pits may also appear on Ei_2 ridge; rarely developed in Ei except postero-laterally. Counts of from 21 to 34 ordered radial rows (with two or more radially aligned pits) in I arcs away from mid-line. Apparent continuity of innermost I„ arc. Usually two new, incomplete I arcs develop on external side of innermost arc, the first to the side of forward pro- trusion of glabella, the second further out along inner margin of fringe. Twin pits frequently formed at bifurcation of these inner arcs. On lower lamella, girder weak, ridge-like structure, no more prominent than Ei_2 pseudogirder ; appears to die out before reaching posterior extremity (PI. 3 1 , figs. 2-3) ; most easily recognized by change in nature of pitting from larger, wider-spaced pits of E arcs to smaller, more closely spaced pits in I arcs. Terrace lines developed along rolled inner edge of lower lamella (PI. 30, fig. 6). Prominent keel on ventral surface of genal spine tends to fade out at posterior corner of fringe and does not appear to continue into weakly developed girder or pseudogirders. Genal spine long, very EXPLANATION OF PLATE 30 Figs. 1-11. Parkesolithus dictyotos sp. nov., from the Malongulli Formation. 1-4, 6, 10, from Copper Mine Creek; 9, from Trilobite Hill; 5, 7-8, 11, from near Mirrabooka homestead, north of Cheeseman’s Creek. 1, dorsal view of internal mould of holotype, SUP 26920, x3-5. 2, dorsal view of latex impres- sion of external mould of thorax and pygidium, paratype SUP 26922, x 5. 3, dorsal view of internal mould of thorax of paratype SUP 26921a, x4. 4, dorsal view of latex impression of two paratypes, an incomplete exoskeleton, SUP 26915a, and a cephalon, with long genal spine (left side), SUP 26915b, X 2. 5, enlarged dorsal view of latex cast of paratype, SUP 26904a, showing pattern of ornamentation on posterior part of cephalon, x 6. 6, ventral view of latex impression of lower lamella, paratype SUP 26925, X 5. 7-8, dorsal views of internal and external (latex) moulds of paratype SUP 26908, x 6. 9, dorsal view of latex impression of part of large cephalon, paratype SUP 7915, x 3. 10, ventral view of latex impression of portions of lower lamella and genal spine, paratype SUP 26919, x 3. 11, dorsal view of latex cast of upper lamella and part of genal spine, paratype SUP 26904b, x 4. Fig. 12. Malongullia nep/A'/ Webby, Moors, and McLean 1970 (dimorph A) from the Malongulli Forma- tion near Mirrabooka homestead, north of Cheeseman’s Creek. Ventral view of latex impression of part of left free cheek showing doublure and genal spine, SUP 27907, x 6. PLATE 30 WEBBY, Parkesolithus, Malongullia 230 PALAEONTOLOGY, VOLUME 17 gently curved, presumed to project rearwards about twice length of pygidium beyond posterior tip; sharp-ridged, quadrate cross-section, in apparent continuity with lower lamella. Facial suture runs along dorso-marginal edge of fringe, crosses genal angle dorsally, and isolates genal spine on lower lamella. Keel-like ridge on dorsal side of genal spine stops abruptly against suture on genal angle (PI. 30, fig. 1 1). Hypostome not known. Thorax of six segments, each of similar length (sag. and exsag.); usually about three times wider than long. Axis occupies about one-fifth total width, tapering very slightly towards pygidium, transversely convex and elevated above adjacent, flatter pleurae; defined by well-developed axial furrows, especially internally (PI. 30, fig. 3). Axial rings relatively short (sag. and exsag.), the first four markedly sharp crested (sag.), and the last two more rounded (sag.); separated from articulating half ring by moderately deep articulating furrow which descends laterally into very deep, slot-like, transversely elongated apodemal pits set well inside axial furrows. Articu- lating half rings short, usually gently convex forwards, but on first segment almost transverse. Pleurae of third and fourth segments widest (tr.) ; pleurae of first segment taper laterally, with large facets developed on antero-lateral angles, set at angle of about 30° to transverse direction, enabling them to fit neatly in behind posterior cephalic border. Through progressive outward movement of fulcrum towards pygidium, succeeding pleurae have narrower facets and blunter, backwardly turned pleural tips. Ends of pleurae sharply downturned. Pleural furrows very broad (exsag.) and moderately deep; on pleurae of first and second segments commencing just in front of mid-length (exsag.) on axial furrow, but on remainder originate nearer anterior margin; directed towards distal extremity of each pleura, but not con- tinuous across narrow marginal rim. Coarse reticulate pattern of ornamentation developed on median part of anterior-facing slopes of sharply crested axial rings on first, second, and third segments, and less prominently on fourth segment. Pleural furrows of first and second segments exhibit finer reticulation, and those of third and fourth segments have an even finer meshwork. Fifth and sixth segments appear to be smooth. Pygidium subtriangular, approximately two-thirds length (sag.) of thorax; three to four times wider than long, with smaller specimens even relatively wider. Up to eight fused axial rings, each progressively shorter and narrower towards rear, and small terminal piece. Axial furrows broad, moderately well marked, curving inwards towards posterior margin, and separating raised axis from more or less flat pleural field. Anteriorly, pair of prominent, slot-like apodemal pits on either side of articu- lating furrow; articulating half ring, gently convex forwards (PI. 31, fig. 1). Anterior margin excluding articulating half ring, straight; posterior margin usually rounded. Up to six pairs of weakly developed pleural ribs, but more usually five; directed progressively more strongly backward to rear; often pleural field almost flat and smooth, with only gentle flexures representing pleural ribs adjacent to postero- lateral border. Narrow, crested postero-lateral border widens (sag. and exsag.) towards mid-line, and exhibits faint, closely spaced terrace lines on steeply declined posterior flank of border. Dorsal surface of pygidium smooth. Measured overall length of exoskeleton excluding prolongation of genal spines behind pygidium from 5 to 20-2 mm; width including fringe from 6-5 to 22-5 mm. WEBBY: ORDOVICIAN TRILOBITES 231 From large fragmentary specimens, estimated to reach maximum length of 26 mm and width of 31 mm. Remarks. The type species of Parkesolithus, P. gradyi, from the unnamed Upper Ordovician shales west of Parkes differs from P. dictyotos in having few if any auxili- ary pits in Ex, only two-three continuous I arcs, a relatively wider pygidium in larger specimens than in smaller, and an unornamented cephalon and thorax (Campbell and Durham 1970). Family raphiophoridae Angelin, 1854 Genus malongullia Webby, Moors, and McLean, 1970 Type species. M. oepiki Webby, Moors, and McLean, 1970. Diagnosis. Raphiophorid genus exhibiting subquadrate to clavate glabella, with large, forwardly expanding frontal lobe, often a median tubercle, and postero-lateral lobes; shallow transglabellar furrow divides frontal lobe from small basal median lobe, extending laterally into deep, slot-like linked lateral glabellar furrows Ip and 2p-, eye tubercle on fixed cheek lies close to axial furrow and opposite short, trans- verse lateral glabellar furrow 3p. Fixed cheeks united in front by typically rather narrow, flattened preglabellar field. Narrow anterior border and prominent anterior border furrow confined to free cheeks, the latter also connected across mid-line; narrow doublure not cut by suture. Well-defined posterior border and posterior border furrow. Long, slender, gently curved genal spine with dorsal and ventral longitudinal furrows giving hour-glass shaped cross-section. Thorax of six seg- ments, similar to that of Dionide, except for more zigzag-shaped axial furrows. Pygidium prominently segmented, of two distinct dimorphic types— one sub- triangular, short (sag.), less than length of cephalon, usually with eight-nine axial rings and five pleural ribs, and the other sub-semicircular, equal to cephalon or longer (sag.), typically with 16-20 axial rings and eight-ten pleural ribs. Discussion. The discovery of large, almost complete specimens exhibiting a Malongullia-type cephalon and thorax, but with much expanded pygidium, occurring in the same beds at two different localities as the original M. oepiki (now dimorph A), has necessitated the emendation of the original conception of the genus and species given by Webby, Moors, and McLean (1970). Previously, isolated, poorly preserved, large pygidia were known from Trilobite Hill but could not be assigned to any known form. Additional collecting at Copper Mine Creek, and north of Cheeseman’s Creek has revealed that there is in fact an association of the two forms in the same beds of the Malongulli Formation, and they clearly seem to represent dimorphs. Details of the differences between the two forms is given in a preceeding discussion of the dimorphism (p. 214). M. oepiki (dimorph B) exhibits close similarities to the type species of Edmund- sonia, E. typa Cooper, 1953, from the Lower Edinburg Formation and equivalents (Porterfield) of Virginia and Tennessee, but this latter form seems to lack eye tubercles and a median glabellar tubercle, and to have a much narrower (exsag.) posterior cephalic border. Although the pygidium of the type species of Cnemidopyge Whit- tard, 1955, C. nuda (Murchison) from the uppermost Llandeilo beds of the Builth area, Wales (Hughes 1969), closely resembles that of M. oepiki (dimorph B), other 232 PALAEONTOLOGY, VOLUME 17 features such as the frontal glabellar spine, the lack of eye tubercles, and the much narrower postero-lateral lobes readily distinguish it. The type species of Ampyxinella Koroleva, 1959, A. rugosa (Kolova) from the ‘Middle’ Ordovician of northern Kazakhstan (Weber 1948) is closely related to MalonguUia oepiki, but not congeneric as Kobayashi and Hamada (1971, p. 133) have claimed. It has eye tubercles and six thoracic segments as in M. oepiki, but is distinguished by exhibiting a prominent anterior border furrow and part of the anterior border on the cranidium, and much deeper, curving caecae on fixed cheeks, by apparently lacking a clearly differentiated posterior border and posterior border furrow, and by having a thorax with a relatively broader (tr.) axis, with more pro- nounced taper posteriorly. In addition, it differs from M. oepiki (dimorph A) in having a much broader (sag. and exsag.) preglabellar area, and a more acutely taper- ing axis and fewer, broader pleural ribs on the pygidium. M. oepiki (dimorph B) has a similar broad preglabellar area, but even it differs in showing a pair of small tubercles, and the pygidium is completely different. Ampyxina biloba Tschugaeva from the Kopalin and Karakan horizons (Llanvirn to Lower Llandeilo) of Kazakhstan, one of the species assigned by Koroleva (1959) to Ampyxinella, is described by Tschugaeva (1958, p. 35) as having only five thoracic segments. Apart from this feature, it bears close similarities to M. oepiki, and may have been representative of the stock from which the Australian Caradoc genus evolved. Whittington and Hughes (1972, p. 273) have suggested that MalonguUia should be grouped with the Endymioniidae rather than the Raphiophoridae. The species of Endymionia Billings, 1 865 from the Table Head Formation and correlatives (Llanvirn) of Newfoundland and Quebec (Whittington 1965), differ fundamentally from M. oepiki in lacking eye tubercles and a differentiated basal median lobe, and in having seven thoracic segments. However, a number of unusual features are shared by the two genera and suggest a close relationship. For instance, the presence of a pair of small tubercles on the preglabellar field and the single, median tubercle on the glabella. Such similarities probably merely stress the close relationships between the two families. Although showing resemblances to species of Dionide Barrande, 1847 (see Whit- tington 1952; Whittard 1958), M. oepiki (dimorph B) differs basically in lacking a bilaminar pitted fringe, and in having a small basal median lobe on the glabella and definite eye lobes on the cheeks. The thorax is similar to that of Dionide, but for the markedly zigzag-shaped axial furrows. Pleural ribs on the anterior part of the pygidium of M. oepiki (dimorph B) have a slight concave forward curvature, in contrast to the typical convex forward curvature of species of Dionide. MalonguUia oepiki Webby, Moors, and McLean, 1970 Plate 30, fig. 12; Plate 31, figs. 4-12; Plate 32, figs. 1-6 1970 MalonguUia oepikiWthhy, Moors, and McLean, p. 882, pi. 125, figs. 1-12 (dimorph A only). Additional material of dimorph A. Eleven specimens (SUP 20908a-b, 26947-26949, 27900, 27902a-b, 27911, 27913-27914) from the Malongulli Formation at Copper Mine Creek, and six specimens (SUP 27904, 27905c, 27906-27909) from the Malongulli Formation near Mirrabooka homestead, north of Cheeseman’s Creek Post Office. Supplementary description of dimorph A (PI. 30, fig. 12; PI. 31, fig. 12; PI. 32, WEBBY: ORDOVICIAN TRILOBITES 233 figs. 1-4). Compression seems to have caused some specimens to break and flex along lines between lateral glabellar furrows, and to give accentuated differentiation of frontal and postero-lateral lobes of the glabella (Webby et al. 1970, pi. 125, figs. 3, 9-10). In other, less crushed, specimens the frontal lobe is less clearly separated from the postero-lateral lobes (Webby et al. 1970, pi. 125, figs. 2, 4, 6-8). Lateral glabellar furrows Ip and 2p appear to form a connected, deep, curved, slit-like, posterior impression. Ip directed anteromedially between postero-lateral and basal median lobes, and 2p, extending antero-laterally from oblique intersection with Ip, as short, deep, slot between posterior part of frontal and postero-lateral lobes; broad, shallow transglabellar furrow directed medially from junction of Ip and 2p, isolating small, basal median lobe from frontal lobe. Anteriorly, just inside axial furrow, and just behind mid-length (exsag.) of eye tubercle, short, almost transverse, less deeply impressed lateral glabellar furrow 3p. Tiny, rounded apodemal pits developed at lateral ends of occipital furrow of paratype SUP 7920. Exsagittally elongated, ovate, gently raised eye tubercle situated on fixed cheek next to axial furrow, and well inside facial suture. Well-preserved specimen, SUP 27909, exhibits fine granulation on dorsal surface of fixed cheek and glabella, in- cluding inner side of eye tubercle, but not outer side which may have been occupied by visual area. No lenses seen. Lateral pit placed near extremity of posterior border furrow, just inside sharply rounded genal angle. Facial suture runs forward and inward from genal angle towards anterior margin, presumably meeting its counterpart in smooth curve between narrow anterior border and preglabellar field; posteriorly, it intersects posterior margin just behind lateral pit. Anterior and lateral border narrow, continuous, delimited by prominent anterior and lateral border furrow ; border furrow apparently in continuity with a less strongly indented dorsal, longitudinal furrow on genal spine (PI. 32, fig. 3). Narrow doublure connects free cheeks anteriorly; no median or connective sutures crossing it; hori- zontal anteriorly and antero-laterally, divided into relatively broad, flattened outer part, and narrow, sharp, ridge-like, inner part; anteriorly, doublure widens (sag. and exsag.) slightly towards mid-line; posteriorly, beneath prolongation of free cheek, it expands to form depressed, triangular area flanked by straight, lateral and posterior, and curved antero-median ridges (PI. 30, fig. 12; PI. 32, fig. 2); no doublure along inner part of posterior border. Long, slender genal spines have fairly persistent dorsal and ventral, longitudinal furrows, giving an hour-glass shaped cross-section ; ventral longitudinal furrow dies out proximally; not continuous into depressed, triangular area beneath genal prolongation ; obliquely aligned rows of fine granules cover surface of genal spines. Apodemal pits of thorax gently crescent-shaped (concave forward) to transverse slots, set well inside axial furrow; anterior pair on macropleural segment more obliquely placed. On pygidium, six to seven pairs of transverse (slightly crescent- shaped), slot-like apodemal pits occur, decreasing in size posteriorly. One pair of small, oval muscle scars developed on anterior side of second pair of apodemal pits (SUP 27904) ; from constriction of first axial ring, this pair of scars may have been linked across mid-line. Second, smaller pair of isolated muscle scars occurs in front of third pair of apodemal pits. Pattern of fine granules along raised terrace lines of posterior border (SUP 27904). 234 PALAEONTOLOGY, VOLUME 17 Material of dimorph B. Seven specimens (SUP 19945, 20915, 26940-26941, 26942a, 26943-26944) from the Malongulli Formation at Copper Mine Creek, two specimens (SUP 19913, 20901) from the same horizon at Trilobite Hill, and six specimens (SUP 26933-26934, 26936-26938, 27905a) from the Malon- gulli Formation near Mirrabooka homestead, north of Cheeseman’s Creek Post Office. Comparative description of dimorph B (PI. 31, figs. 4-11; PI. 32, figs. 4-6). Exoskeleton ranges from 4-3 to 41 mm in length (sag.) and from 5-8 to 41 mm in width (measured across cephalon). Subrounded outline, with long, gently curved genal spines, pro- jecting well beyond rear of pygidium. Cephalon sub-semicircular. Thorax about two-thirds length of cephalon. Pygidium subequal or slightly longer than cephalon in adult forms. Glabella subquadrate in outline; narrowest across occipital ring; prominent broad, convex frontal lobe expands rapidly forwards to almost three times width posteriorly; flanked in posterior part by pair of elongated, suboval postero-lateral lobes, seemingly accentuated by crushing, and behind by short, narrow, basal median lobe. Probably because of crushing, lateral glabellar furrows appear as narrow, deep troughs, but may have originally been wider muscle areas; lateral glabellar furrows as in M. oepiki (dimorph A). Median tubercle not clearly developed in larger forms, but present on elevated part of frontal lobe of one small specimen (PI. 31, fig. 11). Axial furrows moderately well formed, in continuity anteriorly with pre- glabellar furrow. No anterior pits seen. Occipital ring narrow (sag. and exsag.), gently convex backward, with steeply inclined posterior margin and gentle forward slope into broad, shallow occipital furrow. Fixed cheeks subtriangular, gently convex; eye tubercles ovate, situated on fixed cheek close to axial furrow, opposite lateral glabellar furrow 3p. Two gently diverg- ing caecae extend outward from posterior portion of eye tubercle towards genal angle; less conspicuous anastomosing ridges developed on fixed cheek antero- laterally. Moderately broad (sag. and exsag.), flattened preglabellar field exhibits pair of small tubercles just in front of preglabellar furrow. Inner part of posterior border transverse and relatively broad (exsag.), with flattened dorsal surface; EXPLANATION OF PLATE 31 Figs. 1-3. Parkesolithus dictyotos sp. nov., from the Malongulli Formation. 1, from near Mirrabooka homestead, north of Cheeseman’s Creek; 2-3, from Copper Mine Creek. 1, view of three pygidia (one inverted) and part of lower lamella of fringe, paratypes SUP 26906a-d, x 2. 2, ventral view of latex cast of lower lamella, paratype SUP 26924, x 2-5. 3, ventral view of latex impression of part of lower lamella, paratype SUP 26926, x 4. Figs. 4-11. Malongullia oepiki Webby, Moors, and McLean 1970 (dimorph B) from the Malongulli For- mation. 4-9, from Copper Mine Creek; 10-11, from near Mirrabooka homestead, north of Cheese- man’s Creek. 4, dorsal view of latex impression of entire exoskeleton lacking free cheeks, SUP 26940, X 2. 5, ventral view of part of cephalic doublure, SUP 26942a, x 4. 6, dorsal view of doublure linking free cheeks and genal spine, SUP 19945, x2-5. 7-8, dorsal and enlarged dorsal views of latex impres- sion of part of right free cheek and genal spine, SUP 19945. 7, x 4; 8, x 10. 9, enlarged view of part of genal spine showing pattern of micro-ornamentation, SUP 19945, x8. 10, view of dorsal surface of part of pygidium, and undersurface of cranidium, SUP 26936, x 3-5. 1 1, dorsal view of latex cast of small holaspid, SUP 26934, x 7. Fig. 12. Malongullia oepiki Webby, Moors, and McLean 1970 (dimorph A) from the Malongulli Forma- tion at Copper Mine Creek. Dorsal view of typical, almost complete exoskeleton, SUP 27902a, x 4. PLATE 31 WEBBY, Parkesolithus, Malongullia 236 PALAEONTOLOGY, VOLUME 17 expanding and backwardly curved distally. Posterior border furrow dies out approaching axial furrow; lateral pit situated at distal end of furrow, just inside sharply rounded genal angle. Doublure more or less horizontal, widest (sag. and exsag.) anteriorly across mid- line; outer margin evenly curved; inner margin straighter, more transverse across mid-line; divided into broad, gently convex to flattened outer part, and narrow, sharply ridged inner part, both exhibiting more or less continuous, fine terrace lines running parallel to margins. Anterior and lateral border furrow prominent and con- tinuous around margin, but appears to die out in prolongation ; not clearly shown to be in continuity with dorsal longitudinal furrow of genal spine; dorsal and ventral longitudinal furrows fairly persistent along spine as in M. oepiki (dimorph A). Sinuous rows of raised lines borne on prolongation of free cheek; gradation into inclined rows at about 45° to exsagittal line on genal spine (PI. 31, figs. 7-9); rows become progressively more clearly granulated distally, and tend to be more regular and close spaced on outer side of spine; a V-shaped pattern of rows may occur across a longitudinal furrow (PI. 32, fig. 4); in some parts of genal spine, rows of granules appear more like rows of short spines. Regular rows of fine granules also occur along raised, dorso-posterior margin of posterior border, and on most elevated part of frontal lobe, in vicinity of median tubercle. Thorax of six segments, sub-rectangular, with zigzag-shaped axial furrow; usually about three-and-one-half times wider than long, but approximately four times in small holaspid (PI. 31, fig. 1 1); anterior segment longer than succeeding ones. Each axial ring as in M. oepiki (dimorph A), with prominent, small antero-lateral lobes; some antero-lateral lobes exhibit longitudinal creases which seem to have formed by compression. Articulating half ring, in front of sharply grooved articulat- ing furrow, gently convex forwards, about two-thirds length (sag.) of axial ring. Transverse rows of fine granules run along raised, dorso-posterior part of anterior three axial rings. Anterior pleurae slope backwards and outwards from fulcrum to pointed, postero-lateral tips; succeeding pleurae usually exhibit rather blunter ter- minations. Pleural furrows as in M. oepiki (dimorph A). Pygidium sub-semicircular, with straight anterior margin and sharply rounded antero-lateral corners; two to two-and-one-half times wider than long; four times wider than long in small holaspid (PI. 31, fig. 11). Axis convex, tapering gradually posteriorly, with rounded end just inside posterior border, except in small specimens which have axis extending back to border. Usually from 16 to 20 axial rings, the last being barely recognizable; small specimens exhibit only 10-11. Up to seven pairs of linked, oval muscle scars on ring furrows of anterior half of pygidial axis; pattern of joined pairs of muscle scars medially, and lateral slopes to axial furrows distally gives this part of axis a trilobed form. Up to nine pairs of isolated, oval muscle scars are developed on posterior ring furrows, becoming progressively smaller to rear. Pleural field horizontal, usually with eight-ten gently convex to flattened ribs, the first three-four with slight forward concavity, the remainder directed progressively more strongly backwards to rear; pleural furrows more strongly im- pressed towards axial furrow, weakening distally; occasionally impersistent, rather faint interpleural furrows developed on pleural ribs, especially anteriorly. Irregular hummocky areas on posterior part of pleural field and immediately behind axis WEBBY: ORDOVICIAN TRILOBITES 237 perhaps suggestive of pygidial caecae (PI. 32, figs. 5-6). Broad, flattened postero- lateral border; probably originally, prior to compression, inclined backwards and outwards. Backward deflection of inner edge of posterior border across mid-line; fine terrace lines on border, running subparallel to postero-lateral margin. A summary of the differences between dimorphs A and B has been given in a preceding section (p. 214). Family cheiruridae Hawle and Corda, 1847 Genus sphaerocoryphe Angelin, 1854 Type species. S. dentata Angelin, 1854; subsequent designation of Vogdes 1890. Sphaerocoryphe exserta sp. nov. Plate 33, figs. 1-9 Material. Holotype (SUP 27915) and five paratypes (SUP 27916-27920) from Ordovician limestone at Billabong Creek. All specimens are silicified. Description. Large, bulbous, elevated frontal lobe occupying virtually entire length of glabella, rising sharply just in front of occipital ring; very short, small, rather inconspicuous, gently convex, triangular basal lobes {Ip) isolated by weak axial furrows, more or less transverse gently depressed occipital furrow, and smooth, backwardly and inwardly curving Ip furrows. Occipital furrow passes laterally into very deep, prominent apodemal pits. Frontal lobe raised about four or five times higher than dorsal crest of convex occipital ring above adjacent posterior border. Frontal lobe occupies about three-quarters of palpebral width; length (sag.) varies from 51 to 8-5 mm, and width from 6 to 9 mm. Width across basal lobes about one-half of palpebral width. Fixed cheek incomplete, apparently originally subtriangular in shape, sloping out and back from narrow (exsag.), raised, crescentic, almost transversely aligned, palpebral lobe, situated just in front of intersection of axial and Ip furrows; pro- minent, sharp palpebral furrow along rear side of lobe, leading outwards and down- wards into lateral border furrow. Flanked in front by posterior branch of facial suture which extends antero-laterally on to lateral border and then deflected back- wards. Lateral border furrow prominent but virtually dying out near genal angle; posterior border furrow deeply impressed, but also dying out toward genal angle. Posterior border convex, flat on inner part, becoming gently convex outwards, widening (exsag.) slightly distally, and continuous into strong, gently arched genal spine. Estimated overall width across genal spines about 23 mm, i.e. about twice width between fulcra of either side at posterior border. Prominent, raised, straight, horizontal, posterior flange, and fulcral socket extends outward from axial furrow only to about two-thirds total length (tr.) of posterior border. Narrow doublure beneath occipital ring extends about one-third of sagittal length forwards; not represented beneath posterior border, but expands outward from fulcrum on to genal spines and over short, stubby downward and antero-laterally directed lateral spine on lateral border. Surface of cephalon exhibits fine granulation. Hypostome and thoracic segments have not yet been found. 238 PALAEONTOLOGY, VOLUME 17 Pygidium excluding genal spines, approximately subtriangular in outline; width, measured between fulcra on anterior edge, just less than twice sagittal length. First segment appears like normal thoracic segment with pair of curving backwardly and downwardly directed pleural spines (anterior pleural spines). Anterior margin straight and horizontal with prominent anterior, articulating flange. Axial furrows weakly impressed, subparallel anteriorly, but tapering rapidly together in posterior half of pygidium. Four convex axial rings, and small terminal piece; first two similar in size, second two decreasing considerably in size; ring furrows deepen and broaden laterally into large apodemal pits, though pits of first ring furrow not so prominent. Articulating half ring convex, extending forward to about one-half sagittal length of axial ring in front of first ring furrow. First axial ring arched convexly forwards partly exposing fused, incipient second articulating half ring and second ring furrow. Pleural fields narrow posteriorly into moderately well-defined, open V-shaped posterior border behind axis; weak pleural furrow between first and second segment, between bases of small, anterior pleural spine and very large, postero-lateral pleural spine, aligned on second segment and directed outwards and backwards. Doublure extends beneath posterior and lateral borders, but narrows toward antero-lateral angles; deflected on to anterior and postero-lateral pleural spines and into additional pair of short, pointed posteriorly directed border spines (PI. 33, figs. 5-6). Doublure also forms short, forward and upward, pointed, tongue-like extension of inner margin medially (PI. 33, fig. 7). On undersurface, four pairs of large tear-drop shaped apodemes, broadest and deepest near axial furrows; posterior pair smaller in size. Surface of pygidium finely granulate. Remarks. Poor and fragmentary preservation of the bulk of the described European and North American material makes comparisons with the undistorted, silicified New South Wales specimens difficult. Some fifteen European Caradoc-Ashgill species (six from various horizons at Girvan, Scotland) and six North American ‘Chazy-Trenton’ species of Sphaerocoryphe have been recorded (Lane 1971; Shaw 1968). Dean (1971) also indicated the presence of an unnamed species in the ‘Ash- giir of Quebec. Only the silicified material of S. goodnovi Raymond from the Chazyan of New York (Shaw 1968) can be adequately compared, and it differs markedly from S. exserta. The posterior part of the glabella is less constricted, basal lobes { Ip) are more prominent, and the bulbous frontal lobe has less dorsal elevation. There is marked narrowing from occipital ring (sag.) to inner part of the posterior border (exsag.), less well differentiated axial rings and apodemal pits on the pygidium, and a much more coarsely granulated exoskeleton (Shaw 1968). Among European forms, S. pemphis Lane (1971) from the Balclatchie Beds (Middle Caradoc) of Girvan, Scotland, and S. thomsoni (Reed) from the Drummuck Group (Ashgill) of Girvan are both readily distinguished by having two cephalic lateral spines, more prominent basal lobes {Ip), and a much less conspicuous posterior pygidial border (Lane 1971). S. glohiceps (Portlock) from the ‘Caradoc’ of Desert- creat. Northern Ireland also has much more prominent basal lobes {Jp), a more elongated glabella, and apparently less conspicuous pleural areas between diverging pleural spines on the pygidium (Lane 1971). Both S. psiles Tripp from the Craighead Mudstones (Middle-Upper Caradoc) of Girvan and S. akimbo Tripp from the WEBBY: ORDOVICIAN TRILOBITES 239 Upper Stinchar Limestone (Lower Caradoc) of Girvan have like S. exserta only one pair of cephalic lateral spines. However, S. psiles has more prominent basal lobes {Ip), a longer, more slender posterior part of glabella, and a weakly defined pygidial axis (Tripp 1954). S. akimbo has more slender, curved genal spines, and a more coarsely granulate frontal lobe and pygidium (Tripp 1 967). S. punctata (Angelin) from the Boda Limestone (Ashgill) of Sweden, the Chair of Kildare, Ireland, and probably from Keisley, northern England, has a less dorsally elevated and con- stricted posterior part of the frontal lobe, more conspicuous basal lobes {Ip), more prominently pitted fixed cheeks, and a tuberculate frontal lobe (Warburg 1925; Dean 1971). S. erratica Mannil (1958), from the Pirgu (Fic) horizon of Estonia bears the closest resemblance of the Baltic species, but also has a more coarsely granulate frontal lobe. The proximity of the bulbous frontal lobe to the occipital ring which distinguishes S. exserta from most other species of Spbaerocoryplie recalls the genus Hemi- sphaerocoryphe. The type species, H. pseudohemicraniwn (Nieszkowski) from the Johvi (Di) horizon of Estonia, however, has a more squat frontal lobe, more con- spicuous basal lobes {Ip), a relatively much wider (tr.) occipital ring, and it lacks lateral spines (Opik 1937). Another species, H. granulata (Angelin) from the Kulls- berg and other equivalent limestones of Sweden, has more closely comparable crani- dial proportions, but also lacks lateral cephalic spines and has a tuberculate glabellar ornamentation (Warburg 1925). Furthermore, comparisons with the pygidium of Hemisphaerocoryphe cannot be made, as it remains unknown. Family encrinuridae Angelin, 1854 Genus encrinuraspis Webby, Moors, and McLean, 1970 Type species. E. optimus Webby, Moors, and McLean, 1970. Encrinuraspis optimus Webby, Moors, and McLean, 1970 Plate 32, figs. 7-10 1970 Encrinuraspis optimus Webby, Moors, and McLean, p. 884, pi. 126, figs. 1-16. Material. Five specimens (SUP 18907, 20906-20907, 20912b, 28911b) from the Malongulli Formation at Copper Mine Creek. A specimen (SUP 25902) also occurs in the Malongulli Formation of the Cheeseman’s Creek area. Supplementary description. Preglabellar furrow separates frontal lobe of glabella from preglabellar field; commences at axial furrow between prominent anterior pit and lateral glabellar furrow 3p, and curves evenly forward and inward to near mid- line where it is deflected downward and slightly forward into small, V-shaped medial extension, and upward and backward into short, slot-like median longitudinal furrow on anterior margin of frontal lobe. Preglabellar field cut by anterior branch of facial suture, which crosses axial furrow just in front of anterior pit, and curves evenly across preglabellar field, drawing progressively closer to preglabellar furrow medially, almost cutting tip of V-shaped median extension (PI. 32, fig. 7). Rostral plate not seen but presumed to be very narrow, as in Encrinuroides. Preglabellar area on free cheeks forms enlarged, gently convex, subrectangular lobe, steeply 240 PALAEONTOLOGY, VOLUME 17 declined, bounded dorsally by gently curving anterior branch of facial suture, medi- ally by connective suture, and laterally by broad, deep axial furrow. Anterior border furrow separates preglabellar area from anterior border, but dies out about two- thirds distance to connective suture (PI; 32, fig. 10). Anterior border narrows towards connective suture ; may be differentiated medially from preglabellar area in absence of anterior border furrow by lack of tubercles. Doublure appears to form as rolled extension of anterior and lateral borders. Apodemes prominent below lateral ends of occipital furrow; weakly developed beneath lateral glabellar furrows Ip and 2p. Eye lobe raised, reniform in outline, with steeply backward and inwardly sloping palpebral lobe clearly delimited from gently convex, pitted, L-shaped area of fixed cheek by arcuate palpebral furrow. Broad basal part of eye lobe on free cheek con- sists of steeply inclined eye socle surmounted by convex, crescent- to oval-shaped visual surface composed of numerous, small hexagonal eye facets, occupying some- what more than one-half total height of eye lobe and slightly more than three- quarters of length (exsag.) of eye lobe (PI. 32, figs. 8-9). Both anterior and lateral borders and preglabellar area of free cheek show fine granulation on external surfaces; preglabellar area in addition exhibits tubercles, and sometimes granules may appear on individual tubercles or a centrally placed pit; inside lateral border of free cheek, excluding eye socle, there may be pits as well as granules. Remarks. To the original diagnosis of the genus Encrinuraspis given by Webby, Moors, and McLean (1970, p. 883) should be added the important distinguishing features of the moderately large, but not inflated or bulbous, frontal lobe, and the anterior branch of the facial suture obliquely intersecting the preglabellar field, the EXPLANATION OF PLATE 32 Figs. 1-3. Malongullia oepiki Webby, Moors, and McLean 1970 (dimorph A) from the Malongulli Forma- tion. 1-2, from Copper Mine Creek. 3, from near Mirrabooka homestead, north of Cheeseman’s Creek. 1, ventral view of doublure linking free cheeks, SUP 27902b, x 6. 2, dorsal view of right free cheek, showing doublure and part of genal spine, SUP 27913, x 8. 3, dorsal view of latex cast of part of cephalon and thorax, SUP 27906, x 8. Fig. 4. Malongullia oepiki Webby, Moors, and McLean 1970 (dimorphs A and B) from near Mirrabooka homestead, north of Cheeseman’s Creek. View of cephalic doublure of dimorph A, SUP 27905c, and of part of genal spine of dimorph B, SUP 27905a, x 9. Figs. 5-6. Malongullia oepiki Webby, Moors, and McLean 1970 (dimorph B) from near Mirrabooka homestead, north of Cheeseman’s Creek. 5, dorsal view of internal mould of incomplete pygidium, SUP 26938, X 4. 6, dorsal view of internal mould of part of large pygidium, SUP 26937, x 3. Note small pygidium of similar dimorph at bottom left of figure. Figs. 7-10. Encrinuraspis optimus Webby, Moors, and McLean 1970 from Malongulli Formation of Copper Mine Creek. 7, dorsal view of internal mould of part of cephalon, SUP 20907, x 6. 8, oblique lateral view of internal mould of left free cheek, SUP 28911, x 10. 9, detailed view of visual surface of eye shown in Fig. 8, x 20. 10, view of latex impression of cranidium (ventral side) and free cheeks of SUP 18907, x6. Figs. 11-12. Amphilichas encyrtos sp. nov., from the Quondong Formation, Bowan Park Group, at Quon- dong. Dorsal and lateral views of cranidium, holotype SUP 21914. 11, x 4; 12, x 3. Fig. 13. Triarthrus sp. from Cheesemans Creek Formation near Keenan’s Bridge. Dorsal view of latex impression of part of lateral border on left free cheek, thorax and pygidium, SUP 37999, x 7. PLATE 32 4 WEBBY, Malongullia, Encriiniraspis, Amphilichas, Triarthms 242 PALAEONTOLOGY, VOLUME 17 preglabellar area of the cranidium thus being divided on either side of the mid-line into inwardly narrowing areas. In the type species of Encrinuroides, E. sexcostatus (Salter) from the Shoalshook Limestone (Middle Ashgill) of Pembrokeshire, and in E. autochthon Tripp from the confinis Flags (Lower Caradoc) of Kirkdominae, Girvan, the anterior branch of the facial suture runs more or less parallel to the preglabellar furrow, forming a narrow, entire, band-like preglabellar area of equal width laterally and medially, and the frontal lobe, especially in the type species, is markedly bulbous (Whittington 1950; Tripp 1962fl). Encrinuroides zhenxiongensis Sheng (1964) from the Yentsin Formation (Caradoc- Lower Ashgill) of Zhenxiong, north-east Yunnan, bears very close resemblances to Encrinuraspis optimus, and should be considered congeneric with it. It characteristi- cally has the moderate sized, non-bulbous, frontal lobe, sharply rounded genal angles, large compact eyes, and anterior branches of the facial suture running obliquely across the preglabellar field, isolating narrow, wedge-like areas of the preglabellar field on the cranidium to either side of the mid-line as in Encrinuraspis. At the species level, it differs from E. optimus in having rather different cephalic and thoracic pro- portions, being wider across the posterior part of the glabella and axis of the thorax relative to the total cephalic and thoracic width, having slightly deeper and broader lateral glabellar furrows, giving stronger emphasis to lateral glabellar lobes, exhibit- ing a shorter (sag.) occipital ring, and fewer (only six or seven) pygidial ribs. Encrinuraspisl sp. A Plate 33, figs. 10-21 Material. Seven specimens (SUP 29916-29922) from the Ordovician limestone at Billabong Creek, and one specimen (SUP 29923) from the Quondong Formation at Quondong. Description. Fixed cheek highly inflated, L-shaped, between very deep axial furrow and equally deep posterior border furrow. Posterior border smooth, more or less trans- verse in dorsal outline, very gently convex outward from axial furrow. Only lateral part of occipital ring preserved; arches gently upwards above level of posterior border. Posterior branch of facial suture curves outward and downward towards lateral margin well inside genal angle; anterior branch curves forward, inward, and downward towards anterior end of axial furrow. Overall rounded outline EXPLANATION OF PLATE 33 Figs. 1 9. Sphaerocoryphe exserta sp. nov., from Ordovician limestone at Billabong Creek. 1-4, posterior, lateral, ventral, and postero-lateral views of incomplete cranidium, holotype SUP 27915, x3. 5-9, dorsal, ventral, anterior, lateral, and posterior views of pygidium; paratype SUP 27916, x 3. Figs. 10-21. Encrinuraspisl sp. A. 10-19, from Ordovician limestone at Billabong Creek; 20-21, from Quondong Formation, Bowan Park Group, near Quondong. 10-14, ventral, dorsal, lateral, posterior, and anterior views of hypostome, SUP 29916, x5. 15, lateral view of right free cheek, SUP 29920, x3-5. 16-18, lateral, interior, and ventral views of left free cheek, SUP 29919, x3-5. 19, enlarged lateral view of left free cheek shown in Figs. 16-18, x 6. 20-21, postero-dorsal and interior views of part of right fixed cheek, SUP 29923, x4. Figs. 22-23. Encrinuraspisl sp. B from ‘brachiopod’ unit of Daylesford Formation, Bowan Park Group, north-east of Quondong. Ventral and lateral views of left free cheek, SUP 29926, x 6. PLATE 33 WEBBY, Sphaerocoryphe, Encrinuraspisl 244 PALAEONTOLOGY, VOLUME 17 of eye lobe; defined by deep, curved palpebral furrow which is continuous across facial suture into well-defined furrow (socle furrow); large, raised, crescent-shaped palpebral lobe consists of lower area with row of tubercles running around edge of lobe immediately above palpebral furrow, and upper, smooth, almost conical area. Highest point of eye lobe on facial suture close to mid-point of palpebral lobe and visual surface. Inflated, L-shaped part of fixed cheek covered by moderately large tubercles. Character of free cheek very similar to that of E. optimus differing only in having large tubercles on surface, and more continuous anterior border furrow medially. Anterior border furrow continues towards connective suture, swinging upwards close to suture in direction of preglabellar furrow (PI. 33, fig. 19); axial furrow very broad, deep, separating inflated subrectangular-oval, tuberculate preglabellar area of free cheek from steeply inclined, enlarged area inside lateral border, surmounted by eye. On dorsal side preglabellar area cut sharply by curving anterior branch of facial suture. Anterior border widens postero-laterally into lateral border, usually with large tubercle situated on border approximately in line with anterior end of axial furrow; additional two moderately large tubercles on lateral border. Deep, broad lateral border furrow in continuity with anterior border furrow across anterior end of axial furrow, separating steeply inclined, inflated, tuberculate area of free cheek from lateral border; tubercles of varying sizes, including an especially large one on lower, antero-lateral part of inflated area. Well-defined smooth, convex eye socle presumably surmounted by visual surface, and separated from tuberculate area below by distinct socle furrow. Doublure beneath anterior and lateral borders more or less flattened, horizontal, as seen in side view, but with small downward deflection at antero-median tip (PI. 33, fig. 19); formed from rolled anterior and lateral borders widening slightly posteriorly and opposite anterior end of axial furrow. Doublure beneath lateral border rolled inwards and upwards, whereas on anterior border only projecting inwards and cut by hypostomal suture (PI. 33, fig. 17). Surface of anterior and lateral borders and doublure granulate. Hypostome with oval outline, excluding anterior wings; maximum width about three-quarters sagittal length; inflated middle body has pronounced median anterior lobe with steep, near vertical, upward slope just behind anterior border furrow; shallow furrows to either side of steep slope of median anterior lobe; extend back- ward and slightly outward, becoming weaker and finally dying out just in front of anterior wings. Anterior border has broad V-shaped outline, and sharply downwardly deflected rim, broadening sutural surface of contact with adjacent anterior border of free cheek and rostral plate; anterior border furrow deep and asymmetrically V-shaped in profile, running continuously forward to mid-line from adjacent to base of anterior wings. Lateral borders more gently declined and wider, bounded by more symmetrically open, V-shaped lateral border furrows; posterior border flatter and widest medially, but imperfectly preserved margin prevents determina- tion of original shape of border whether it was pointed or more broadly rounded. No macula seen. Undersurface exhibits no doublure. Anterior wings situated just in front of mid-length of middle body, with only triangular, outwardly directed bases preserved; upwardly inclined wings damaged; posterior wings not shown in material studied. Ventral surface of middle body exhibits fine granulation. WEBBY: ORDOVICIAN TRILOBITES 245 Remarks. Possibly less stress should be given to ornamentation in the diagnosis of Encrinuraspis (Webby, Moors, and McLean 1970, p. 883) in order to accommodate the silicified materal from Billabong Creek referred to Encrinuraspisl sp. A, and from the lower part of the Daylesford Formation near Quondong assigned to Encrinuraspisl sp. B (see below). Neither the Billabong Creek nor the Daylesford material can be included with forms like the type species of Encrinurus, E. punctatus (Wahlenberg), and E. macrourus Schmidt, from the Silurian of Gotland (Tripp 1962^), because they have free cheeks exhibiting an enlarged preglabellar area separated from a continuous, well-dilferentiated anterior border by a well-defined anterior border furrow. E. macrourus has poorly differentiated, more constricted, preglabellar and anterior border areas, divided by an ill-defined, shallow anterior border furrow, which gives the anterior border a marked, medially tapering wedge- shaped appearance (Tripp \962b, p. 469, pi. 66, fig. Ic). The tuberculated New South Wales forms could perhaps be regarded as representatives of the Encrinurus multi- segmentatus species group (Tripp 1957) but these characteristically have a more uniform, coarse tuberculation over most of the cephalon, including the anterior and lateral borders. Pitting, as seen on the area of the free cheek inside the lateral border furrow of Encrinuraspisl sp. B, is not shown in representatives of the E. multi- segmentatus species group. It therefore seems preferable tentatively to refer these tuberculate forms to Encrinuraspis. They seem to exhibit a similar relationship to the more finely ornamented Encrinuraspis in New South Wales, as representatives of the species of the Encrinurus multisegmentatus group have to Encrinuroides in British successions. Encrinuraspis! sp. B Plate 33, figs. 22-23; Plate 34, figs. 1 -2 Material. Two silicified specimens (SUP 29926-29927) from the ‘brachiopod’ unit of the Daylesford Formation, Bowan Park Group, north-east of Quondong. Comparative description. These specimens of free cheeks are distinguished from those of Encrinuraspis! sp. A by being smaller, having a row of moderate-sized tubercles along the dorso-lateral edge of the lateral border, and small, spinose tubercles on the ventro-lateral edge. The preglabellar area is also tuberculate, and the area be- tween the eye lobe and the lateral border furrow also exhibits two or three prominent tubercles together with many small pits. The eye lobe is slightly more elevated and rounded, with the visual surface occupying the upper half. On the lower half, the eye socle exhibits on its antero-lateral corner a slightly expanded, rounded projection covered with granules. The region of the free cheek directly below the eye socle is constricted but lacks a defined socle furrow; it has a crimped appearance with verti- cally elongated pits. The pitting on the area of the free cheek between the eye lobe and lateral border furrow resembles that seen in E. optimus. Family lichidae Hawle and Corda, 1847 Genus amphilichas Raymond, 1905 Type species. Platymetopus lineatus A.ngQ\m, 1854. Discussion. In addition to the two New South Wales species of Amphilichas described 246 PALAEONTOLOGY, VOLUME 17 herein, the genus has been recorded from the Gordon Limestone of Tasmania (Whittington 1966). Amphilichas has a widespread distribution in Caradoc-Ashgill times in North America, Europe, and Asia, some 44 species being recorded by Tripp (1958). The genus first appeared in the Chazyan (about Llandeilo) of New York (Shaw 1968), probably derived from the closely related genus Apatolichas Whittington (1963) from the Lower Head (Whiterock or, in European terms, Llan- virn) of western Newfoundland. By Lower Caradoc times, Amphilichas has spread to Europe, Asia, and probably to Australia. Amphilichas nasutus sp. nov. Plate 34. figs. 3-9 Material. Holotype (SUP 27921) and two paratypes (SUP 27930-27931) from Ordovician limestone at Billabong Creek. Description. Glabella including occipital ring of rather similar length (sag.) to maxi- mum width, narrowing only slightly backwards; gently convex transversely between eye lobes; very convex forwards especially along sagittal line, with slight overhang of anterior border. Lateral glabellar furrows 3p curve inward and swing evenly round into backwardly directed longitudinal furrows; about opposite mid-point of eye lobes, longitudinal furrows become obsolete, and fail to meet occipital furrow. Median, anterior lobe expands forward and comparatively sharply rounded anteriorly along mid-line ; as seen in lateral profile, gently arched dorsal surface of lobe flexed sharply downwards anteriorly into gently convex, near vertical, anterior surface of lobe; at level of eye lobes, median, anterior lobe slightly wider than convex, compo- site lateral lobes on either side of it ; median, anterior lobes narrowest at about glabel- lar mid-length. Occipital ring longer sagittally than laterally ; occipital furrow behind median, anterior lobe straight and transverse, but laterally, behind composite lateral lobes, curving slightly backward and outward. Axial furrows curve forward and in- ward into anterior border furrow, and backward, from just in front of eye lobes, toward lateral ends of occipital ring. Axial, occipital, and lateral glabellar 3p (in anterior two-thirds of its course) furrows impressed equally deeply. Eixed cheek approximately subtriangular, strongly convex exsagittally and flat- tened to gently convex transversely. Eye lobe appears to have been large, situated EXPLANATION OF PLATE 34 Figs. 1-2. Encrmuraspisl sp. B from ‘brachiopod’ unit of Daylesford Formation, Bowan Park Group, north-east of Quondong. Lateral and interior views of incomplete left free cheek, SUP 29927, x 6. Figs. 3-9. Amphilichas nasutus sp. nov., from Ordovician limestone at Billabong Creek. 3-8, dorsal, anterior, lateral, posterior, ventral, and oblique dorso-lateral views of cranidium, holotype, SUP 27921. 3-7, X 3-5; 8, x 4. 9, ventral view of lateral part of hypostome, paratype, SUP 27931, x 4. Figs. 10-21. Amphilichas encyrtos sp. nov., from Quondong Formation, Bowan Park Group, near Quon- dong. 10-13, ventral, dorsal, lateral, and posterior views of incomplete hypostome, paratype, SUP 27922, X 4. 14, ventral view of hypostome lacking lateral notches and shoulders, paratype, SUP 27923, x3-5. 15-17, ventral anterior and oblique antero-ventral views of incomplete hypostome, paratype, SUP 27924, x4. 18, dorsal view of posterior part of fragmentary large hypostome, paratype, SUP 27925, x2-5. 19-21, dorsal, anterior, and lateral views of broken, incomplete cranidium, paratype, SUP 37916, x 3-5. PLATE 34 WEBBY, Encriniiraspisi, Amphilichas 248 PALAEONTOLOGY, VOLUME 17 on highest part of cheek, well out from axial furrow; broad, elevated area not dif- ferentiated; well-marked palpebral furrow and differentiated palpebral lobe only seen on rear slope. Anterior branch of facial suture runs inward and forward, then descends subparallel to axial furrow, cuts across anterior border, and turns sharply to extend in smooth curve along anterior margin of anterior border. Anterior border narrow (sag. and exsag.), horizontal, with evenly curved anterior margin. Posterior branch of facial suture curves in arc across eye lobe, then backward so as to inter- sect posterior margin inside presumed projection of posterior border. Posterior border narrow (tr.); inner part adjacent to occipital ring, short (exsag.) and hori- zontal; lengthening (exsag.) at sharp, nodular flexure of fulcrum; distally, outwardly and slightly forwardly inclined, terminating rather bluntly. Doublure seen to extend about one-third of length forward beneath occipital ring. Prominent groove developed on posterior border just below sharp flexure, presumably for reception of outer part of pleura of first thoracic segment during enrolment. External surface of cranidium covered by tubercles of varying sizes, up to 0-5 mm in diameter at bases. Rostral plate unknown. Only lateral part of hypostome preserved. Large lateral notch (relatively slightly larger than in A. encyrlos sp. nov.) behind base of anterior wing, presumably through which antennule protruded, and in front of well-developed shoulder with bluntly pointed anterior tip. Faint, irregular, anastomosing lines of ventral surface of margin of lateral notch and shoulder. Doublure seen above shoulder on lateral border. Remarks. A. nasutus appears to have the closest resemblances to A. karakanensis var. disjunctus Tschugaeva (1958) from the Anderken horizon (Lower Caradoc) of Kazakhstan, and A. atavus Warburg from the lower part of the Kullsberg Limestone of Dalarne (Warburg 1925, 1939), and similarly exhibits longitudinal furrows on the glabella which fail to reach the occipital furrow. However, it differs from both the Kazakhstan and Swedish forms in having a cranidium with a more acutely pointed anterior portion of the median, anterior lobe, and a more coarsely tuberculated external surface. Amphilichas encyrtos sp. nov. Plate 32. figs. 11-12; Plate 34, figs. 10-21 Material. Holotype (SUP 21914) and nine paratypes (SUP 27922-27929, 37916) from the Quondong Formation, Bowan Park Group, near Quondong. Comparative description. Glabella including occipital ring slightly wider than long (sag.); subparallel-sided, narrowing slightly in front of eye lobes; convex forwards, and more gently convex transversely. Lateral glabellar furrow 3p curves inward and backward into longitudinal furrow which is continuous to occipital furrow; inner course of longitudinal furrow directed backward and slightly inward, but close to intersection with occipital furrow deflected outward. Median, anterior lobe expanding forwards, with even convexity from highest part of glabella opposite eye lobes; narrowest just behind glabellar mid-length. Median, anterior lobe sub- equal (tr.) to composite lateral lobes to either side, at level of eye lobes. Axial, longi- tudinal, and occipital furrows, especially the latter, very deeply impressed. Fixed cheek with prominent palpebral lobe lacking tubercles. Anterior border and anterior WEBBY; ORDOVICIAN TRILOBITES 249 border furrow slightly upwardly flexed medially (PI. 34, flg. 20); in dorsal outline, median part of anterior margin transverse (PI. 34, fig. 19). Tubercles of varying sizes, up to 0-5 mm in diameter, covering most of the surface of cranidium; includes large tubercles set at intersection of longitudinal furrows and occipital furrow. Hypostome flattened to weakly convex, wider than long, with subtrapezoidal middle body separated from very broad lateral and posterior borders by deep, lateral, and posterior border furrows. Middle body divided by pair of short, inwardly and slightly forwardly directed middle furrows into larger, anterior, and shorter, posterior lobes. Hypostomal suture gently convex forwards, with slight V-shaped outline, bordering anterior margin of middle body; anterior border lacking. Anterior wings not preserved; seem to have extended upward from bases at antero-lateral corners. Lateral and posterior borders very broad, especially laterally; expanded out into very prominent, rounded shoulder behind large, lateral notch. Broad, gentle sag in lateral border runs from opposite posterior lobe of middle body outwards on to shoulder. Extended, flattened posterior border deeply notched medially. Pitting covers much of external surface of middle body and borders; also crudely concentric pattern of anastomosing lines on posterior and lateral borders. Doublure extends inward over much of the lateral and posterior border areas; U-shaped, median scallop on surface of doublure in front of median notch of posterior border (PI. 34, figs. 11, 18); also downward, tongue-like deflection of inner edge of doublure medially, just behind posterior border furrow of ventral surface; with sharp, out- wardly curved (concave forward) furrows running close to inner edge to either side of it. Thus, A. encyrtos may be distinguished from A. nasutus by having a glabella with a less convex, more evenly rounded anterior part of median, anterior lobe, longitudinal furrows continuous to occipital furrow, with a prominent tubercle at each furrow intersection, medially flattened anterior margin, a non-tuberculate palpebral lobe, and a more rounded anterior edge of shoulder on hypostome. Acknowledgements. I thank Professor H. B. Whittington and Dr. C. P. Hughes for reading and criticizing the manuscript. Dr. K. S. W. Campbell kindly read the section dealing with ParkesoHthus. Most of the work was carried out during a period of study leave in 1971-1972 at the Sedgwick Museum, University of Cambridge, with the support of a Royal Society and Nuffield Foundation Commonwealth Bursary. Additional financial support came from a Sydney University Research Grant, and the Australian Research Grants Committee. Financial assistance from Sydney University towards publication costs is gratefully acknowledged. Much of the material was originally collected during joint studies of Ordovician succes- sions in New South Wales by myself and Associate Professor G. H. Packham. I thank him, and also Dr. V. Semeniuk, Dr. H. T. Moors, Messrs. L. Sherwin, M. Tuckson, and C. J. Jenkins for making additional specimens available for study. REFERENCES ANGELIN, N. p. 1854. Paloeontologia Scandinavica I. Crustacea formationis transitionis. Ease. 2, 21-92. Fund. BANCROFT, B. B. 1933. Correlation-tables of the Stages Costonian-Onnian in England and Wales. 4 pp. Printed privately. BARRANDE, J. 1847. Uber das Hypostome und Epistoma zwei analoge, aber verschiedene Organe der Trilobiten. Neues Jahrb. f. Min. 385-399. BILLINGS, E. 1865. Palaeozoic fossils. Vol. 1 . Containing descriptions and figures of new or little known species of organic remains from the Silurian rocks. Pp. 169-426. Geol. Surv. Canada. Montreal. 250 PALAEONTOLOGY, VOLUME 17 BURMEISTER, H. 1843. Die Organisation der Trilobiten. 147 pp. Berlin. CAMPBELL, K. s. w. 1967. Trilobites of the Henryhouse Formation (Silurian) in Oklahoma. Bull. Okla. Geol. Surv. 115, 1-68. and DURHAM, G. J. 1970. A new trinucleid from the Upper Ordovician of New South Wales. Palaeonto- logy, 13, 573-580. CHURKIN, M. 1963. Ordovician trilobites from graptolitic shale in Central Idaho. J. Paleont. 37, 421-428. CLARKSON, E. N. K. 1969. Dimorphism of the Eye in Weberides shunnerensis (King) (Trilobita). In wester- MANN, G. E. c. (ed.). Sexual Dimorphism in Fossil Metazoa and Taxonomic Implications. Int. Union Geol. Sci. (A), 1, 185-195. COOPER, B. N. 1953. Trilobites from the Lower Champlainian formations of the Appalachian Valley. Mem. geol. Soc. Amer. 55, 1-69. DEAN, w. T. 1962. The trilobites of the Caradoc Series in the Cross Fell Inlier of Northern England. Bull. Br. Mus. nat. Hist. (Geol.), 7 (3), 65-134. 1971. The trilobites of the Chair of Kildare Limestone (Upper Ordovician) of eastern Ireland. Part 1 . Palaeontogr. Soc., Monogr. 1-60. EMMRICH, H. F. 1844. Zw Naturgescliickte der Trilobiten. 28 pp. Meinigen. GiLBERT-TOMLiNSON, J. 1961. Ordovician fossils from Thangoo No. 1 and Thangoo No. 1a, Western Australia. Bur. Min. Res. Aust., Petroleum Search Sub. Act Publ. 14, 24-29. GREEN, J. 1832. A Monograph of the trilobites of North America. 93 pp. Philadelphia. HARRINGTON, H. J. and LEANZA, A. F. 1957. Ordovician trilobites of Argentina. Univ. Kansas, Dept. Geol., Spec. Publ. 1, 1-276. HAWLE, I. and CORDA, A. J. c. 1847. Prodrom einer Monographic die bohmischen Trilobiten. 176 pp. Prague. HU, CHUNG-HUNG. 1964. The ontogeny and dimorphism of Welleraspis lata Howell (Trilobita). J. Paleont. 38, 95-97. 1971. Ontogeny and sexual dimorphism of Lower Paleozoic Trilobita. Palaeontogr. Americana, 7(44), 29-155. HUGHES, c. p. 1969. The Ordovician trilobite faunas of the Builth-Llandrindod Inlier, Central Wales. Part I. Bull. Br. Mus. nat. Hist. (Geol.), 18 (3), 39-103. 1970. Statistical analysis and presentation of trinucleid (Trilobita) fringe data. Palaeontology, 13, 1-9. HUPE, p. 1953. Classe des Trilobites. In piveteau, j. (ed.), Traite de Paleontologie, 3, pp. 44-246. Paris. INGHAM, J. K. 1970. The Upper Ordovician trilobites from the Cautley and Dent districts of Westmorland and Yorkshire. Part 1. Palaeontogr. Soc., Monogr. 1-58. KOBAYASHi, T. and HAMADA, T. 1971. Silurian trilobites from the Langkawi Islands, West Malaysia, with notes on the Dalmanitidae and Raphiophoridae. Geology and Palaeontology of Southeast Asia, 9, 87- 1 34. KOROLEVA, M. N. 1959. Novye rody trilobitov iz srednego i verhknego ordovika Severnogo Kazahkstana. Dokl. Akad. Nauk. SSSR, 124(6), 1313-1316. LAMONT, A. 1935. The Drummuck Group, Girvan ; a stratigraphical revision, with descriptions of new fossils from the lower part of the group. Trans, geol. Soc. Glasgow, 19, 288-332. LANE, p. D. 1971. British Cheiruridae (Trilobita). Palaeontogr. Soc. Monogr. 1-95. MANNiL, R. M. 1958. Trilobity semejstv Cheiruridae i Encrinuridae iz Estonii. Geoloogia Inst. Uurim. 3, 165-212. MITCHELL, A. H. and READING, H. G. 1969. Continental margins, geosynclines and ocean-floor spreading. J. Geol. 77, 629-646. MOORE, R. c. (ed.). 1959. Treatise on Invertebrate Paleontology. Part O, Arthropoda. xix-|-560 pp. Univ. Kansas Press. MOORS, H. T. 1970. Ordovician graptolites from the Cliefden Caves area, Mandurama, N.S.W., with a re-appraisal of their stratigraphic signihcance. Proc. roy. Soc. Viet. 83, 253-287. NICHOLSON, H. A. and ETHERIDGE, R. 1879. A monograph of the Silurian fossils of Girvan in Ayrshire with special reference to those contained in the "Gray Collection . Vol. 1, Fasc. 2, i-vi, 136-233. Edinburgh and London. OPIK, A. A. 1937. Trilobiten aus Estland. Acta Comment. Univ. Tartu (A), 32 (3), 1-163. {Publ. Geol. Inst. Univ. Tartu, 52). 1958. The Cambrian Trilobite Redlichia: organization and generic concept. Bull. Bur. Min. Resour. Aust. 42, 1 50. WEBBY: ORDOVICIAN TRILOBITES 251 OPiK, A. A. 1963. Early Upper Cambrian Fossils from Queensland. Ibid. 64, 1-133. PACKHAM, G. H. 1967. The Occurrence of Shelly Ordovician Strata near Forbes, New South Wales. Amt. J. Sci. 30, 106-107. 1969. Ordovician System. The Geology of New South Wales. J. geol. Soc. Amt. 16, 76-103. and FALVEY, D. A. 1971. An hypothesis for the formation of marginal seas in the Western Pacific. Tectonophysics, II, 79-109. PALMER, A. R. 1972. Problems of Cambrian biogeography. 24th Int. geol. Congr., sect. 7, 310-315. RAYMOND, P. E. 1905. Note on the Names Amphioii, Harpina, and Platymetopm. Amer. J. Sci. (4), 19, 377-378. REED, F. R. c. 1896fl. The fauna of the Keisley Limestone, Part I. Q. Jl geol. Soc. Lond. 52, 407-437. 18966. Woodwardian Museum notes. Notes on the evolution of the genus Cheirurm. Geol. Mag. (4), 3, 117-123. 1904. The Lower Palaeozoic trilobites of the Girvan district, Ayrshire. Part 2. Palaeontogr. Soc., Monogr. 49-96. 1928. Notes on the Bronteidae [= Goldidae]. Ann. Mag. nat. Hist. (10), 1, 49-78. 1931. The Lower Palaeozoic trilobites of Girvan. Supplement no. 2. Palaeontogr. Soc., Monogr. 1-30. RICHTER, R. and RICHTER, E. 1955. Scutelluidae n.n. (Tril.) durch ‘Kleine Anderung’ eines Familien-Namens wegen Homonymie. Senck. leth. 36, 291-293. 1956. Grundlagen fiir die Beurteilung und Einteilung der Scutelluidae (Tril.). Senck. leth. 37, 79-124. ROBISON, R. A. 1972. Hypostoma of agnostid trilobites. Lethaea, 5, 239-248. ROSS, R. J. and shaw, f. c. 1972. Distribution of the Middle Ordovician Copenhagen Formation and its Trilobites in Nevada. U.S. geol. Surv., Prof. Pap. 749, 1-33. SCHEIBNER, E. 1972. Actualistic Models in Tectonic Mapping. 24th Int. geol. Congr., sect. 3, 405 -422. SELWOOD, E. B. 1965. Dechenellid trilobites from the British Middle Devonian. Bidl. Br. Mas. nat. Hist. (Geol.), 10, 319-333. and BURTON, c. J. 1969. Possible Dimorphism in certain Devonian Phacopids (Trilobita). In wester- MANN, G. E. c. (ed.). Sexual Dimorphism in Fossil Metazoa and Taxonomic Implications. Int. Union Geol. Sci. (A), 1, 196-200. SHAW, E. c. 1968. Early Middle Ordovician Chazy Trilobites of New York. N.Y. State Mm. Mem. 17, 1-163. SHENG, s. F. 1964. Upper Ordovician trilobite faunas of Szechuan-Kweichow with special discussion on the classification and boundaries of the Upper Ordovician. Acta Palaeont. Sinica, 12 (4), 537-571. SHERWiN, L. 1970. Age of the Billabong Creek Limestone. Quart. Notes geol. Surv. N.S.W. 1, 1-3. 1971. Stratigraphy of the Cheeseman’s Creek District, New South Wales. Rec. geol. Surv. N.S.W. 13, 199-237. SINCLAIR, G. w. 1949. The Ordovician trilobite Eobronteus. J. Paleont. 23, 45-56. SKJESETH, s. 1955. The Middle Ordovician of the Oslo region, Norway. 5. The trilobite family Styginidae. Norsk geol. Tidsskr. 35, 9-28. SKWARKO, s. K. 1962. Graptolites of Cobb River-Mount Arthur Area, North-West Nelson, New Zealand. Trans, roy. Soc. N.Z., Geol. 1 (15), 215-247. SMITH, R. E. 1966. The geology of Mandurama-Panuara. J. Proc. R. Soc. N.S.W. 98, 239-262. SNAJDR, M. 1960. Studie o celedi Scutelluidae (Trilobitae). Rozpravy, LJstfed. ustav. geol. 26, 1-263. STRACHAN, I. 1972. Correlation of British and Australian graptolite zones. In williams, a., strachan, l, BASSETT, D. A., DEAN, W. T., INGHAM, J. K., WRIGHT, A. D., and WHITTINGTON, H. B. A correlation of Ordo- vician rocks in the British Isles. Spec. Pap. geol. Soc. Lond. 3, 10-14. THORSLUND, p. 1940. On the Chasmops Series of Jemtland and Sodermanland (Tvaren). Sver. geol. Unders. (C), 436, 1-191. TRIPP, R. p. 1954. Caradocian Trilobites from Mudstones at Craighead Quarry, near Girvan, Ayrshire. Trans. R. Soc. Edinb. 62, 655-693. 1957. The trilobite Encrinurus multisegmentatus (Portlock) and allied Middle and Upper Ordovician species. Palaeontology, 1, 60-72. 1958. Stratigraphical and geographical distribution of the named species of the trilobite superfamily Lichacea. J. Paleont. 32, 574-582. D 252 PALAEONTOLOGY, VOLUME 17 TRIPP, R. p. 1962o. Trilobites from the confinis Flags (Ordovician) of the Girvan district, Ayrshire. Trans. R. Soc. Edin.(>5,\-A0. 1962(7. The Silurian trilobite Encrinurus punctatus (Wahlenburg) and allied species. Palaeontology, 5, 460-477. 1967. Trilobites from the Upper Stinchar Limestone (Ordovician) of the Girvan district, Ayrshire. Trans. R. Soc. Edin. 67, 43-93. TSCHUGAEVA, M. N. 1958. Trilobity ordovika Tschu-Ilijskihk gor. Ordovik Kazahkstana, III. Trad. geol. Inst. Akad. Nauk. SSSR, 9, 5-138. VALENTINE, J. w. 1971. Plate tectonics and shallow marine diversity and endemism, an actualistic model. Syst. Zool. 20, 253-264. VOGDES, A. w. 1890. A bibliography of Palaeozoic Crustacea from 1698 to 1889, including a list of North American species and a systematic arrangement of genera. Bull. U.S. geol. Surv. 63, 1-177. WARBURG, E. 1925. The trilobites of the Leptaena limestone in Dalarne. Bull. Geol. Instn Univ. Uppsala, 17, 1-446. 1939. The Swedish Ordovician and Lower Silurian Lichidae. K. Svenska Vetens. Hand!. (3), 17 (4), 1-162. WEBER, v. N. 1945. O rode Thysanopehis Corda i nekotoryhk Uralbskihk Bronteidae. Materialy VSEGEl, paleontologija i stratigrafija. Vyp. 4, 24-45. 1948. Monograhi po paleontologii SSSR. Tom 69, Trilobity Silurijskihk oflozhenij SSSR. Vyp. 1. Nizhnesilurjskie trilobity. Vses. Naucimo-issled. geol. inst. (VSEGEl). Minist. geol. SSSR, 1-113. WEBBY, B. D. 1969. Ordovician stromatoporoids from New South Wales. Palaeontology, 12, 637-662. 1971. The trilobite PUomerina Chugaeva from the Ordovician of New South Wales. Ibid. 14, 612-622. 1973. Remopleurides and other Upper Ordovician trilobites from New South Wales. Ibid. 16, 445-475. MOORS, H. T. and mclean, r. a. 1970. Malongullia and Encrinuraspis, New Ordovician trilobites from New South Wales, Australia. J. Paleont. 44, 881-887. WHiTTARD, w. F. 1955. The Ordovician trilobites of the Shelve Inlier, West Shropshire, Part I. Palaeontogr. Soc., Monogr. 1-40. 1958. The Ordovician trilobites of the Shelve Inlier, West Shropshire, Part III. Ibid. 71-116. 1961. The Ordovician trilobites of the Shelve Inlier, West Shropshire, Part V. Ibid. 163-196. WHITTINGTON, H. B. 1950. Sixteen Ordovician genotype trilobites. J. Paleont. 24, 531-565. 1952. The trilobite family Dionididae. Ibid. 26, 1-11. 1957. Ontogeny of Elliptocephala, Paradoxides, Sao, Blainia, and Triartbrus (Trilobita). Ibid. 31, 934-946. 1959. Silicified Middle Ordovician trilobites: Remopleurididae, Trinucleidae, Raphiophoridae, Endymioniidae. Bull. Mus. comp. Zool., Harv. 121, 371-496. 1963. Middle Ordovician trilobites from Lower Head, western Newfoundland. Ibid. 129, 1-118. 1965. Trilobites of the Ordovician Table Head Formation, western Newfoundland. Ibid. 132, 275-442. 1966. Phylogeny and distribution of Ordovician trilobites. J. Paleont. 40, 696-737. and hughes, c. p. 1972. Ordovician geography and faunal provinces deduced from trilobite distribu- tion. Phil. Trans. R. Soc. Lond. B, 263, 235-278. B. D. WEBBY Department of Geology and Geophysics University of Sydney N.S.W., 2006 Typescript received 3 January 1973 Australia CAPITOSAUROID LAB YRINTHODONTS FROM THE TRIAS OF ENGLAND by ROBERTA L. PATON Abstract. Amphibian remains from several ’Lower Keuper’ Sandstone localities in England, including the holo- type of Cyclotosaums stantonensis (Woodward), are examined. Differences in patterns of skull ornamentation in mastodonsaurs and parotosaur/cyclotosaurs are discussed and it is suggested that these are of taxonomic value. The specimens fall into three groups, two of which are species of Cyclotosaurus, the other group representing a primi- tive mastodonsaur. Basic differences between mastodonsaurs and parotosaur/cyclotosaurs are discussed. The complicated nomenclature of the English fossils is investigated and the valid names of the cyclotosaurs are shown to be Cyclotosawus leptognatlius (Owen), of which C. stantonensis (Woodward) is a subjective junior synonym, and C. pachygnathus (Owen). The valid name of the mastodonsaur is demonstrated to be Mastodonsaurus lavisi (Seeley). The phylogeny of the superfamily Capitosauroidea is discussed. The significance of the vertebrate remains in the stratigraphic correlation of the ‘Lower Keuper' Sandstone is considered; the degree of specialization of the labyrinthodonts suggests an early Ladinian (Middle Trias, Upper Muschelkalk) age for their horizon. The fossils forming the basis of this study are from Triassic rocks in Warwickshire (Coten End SP 288655, Cubbington SP 335685), Bromsgrove (SP 960700) in Wor- cestershire, Stanton (SK 126462) in Staffordshire, and near Sidmouth (SY 105864) in Devonshire. All the specimens from Warwickshire were collected in the nine- teenth century from quarries which have since been closed. Unfortunately there is no record of the actual collection of the specimens and therefore any association between them cannot be determined. They are all from horizons of approximately the same age which lie near the top of the ‘Lower Keuper’ Sandstone (Building Stones Formation of Warrington 1970) just below its junction with the Waterstones Formation. The first major work on the English Triassic labyrinthodonts was published in 1842 by Owen who described and figured five species of labyrinthodont : Laby- rinthodon jaegeri which was based upon two casts of portions o^ lower jaws, the originals having been lost; L. ventricosus based upon a single tooth; L. scutulatus based on a specimen showing a collection of bones of what is now recognized as a small lepidosaur; L. pachygnathus', and L. leptognatlius. Some specimens referred to the latter two species were also reptilian, but others considered here are definitely labyrinthodont. Miall (1874) described further discoveries from Warwickshire, assigning some specimens to a new genus and species, Diadetognathus varvicensis, and also com- menting upon Owen’s memoir pointing out the reptilian specimens and separating the two verified labyrinthodont species into different genera: Mastodonsaurus pachygnathus and Labyrinthodon leptognatlius. Since 1874 the fragmentary laby- rinthodonts in Warwick County Museum have not been studied. Seeley (1876) described a labyrinthodont lower jaw from Sidmouth and assigned it to a new species, L. lavisi. He defended Owen’s original designation of specimens into species of Labyrinthodon. Rather unfortunately he also refuted a suggestion [Palaeontology, Vol. 17, Part 2, 1974, pp. 253-289, pis. 35-36.] 254 PALAEONTOLOGY, VOLUME 17 that the labyrinthodonts were amphibian, maintaining that they were definitely reptilian showing resemblances to the Teleosauria and marine Chelonia. Wills (1915) redescribed the lower jaws including L. lavisi and material from Warwickshire in the British Museum (Nat. Hist.) and the Geological Survey. Frag- ments of English Triassic labyrinthodonts are also present in the Sedgwick Museum and in the collection of the Department of Geology, Birmingham University. The complete skull of Cyclotosaurus stantonensis (Woodward) is from approximately the same horizon as the Warwickshire specimens. As a group the ‘capitosaurs’ were among the earliest amphibians discovered; Mastodonsaurus "giganteus" {M. jaegeri) from the Lettenkohle of Gaildorf in Ger- many was the first labyrinthodont to be described (Jaeger 1828). Since then, many new genera and species have been discovered and anatomical studies have been made by Huene (1922, 1932), Nilsson (1943, 1944), Watson (1919, 1926, 1951, 1958, 1962), and Romer (1947). Watson (1962) discussed various evolutionary trends in the group, e.g. progressive closure of the otic notch; progressive chondrification especially noticeable in the neurocranium and limb bones; an increase in dorso- ventral flattening of the skull (he suggested that there were two Upper Triassic lineages, one high-skulled and the other low-skulled) ; and an increase in size. Welles and Cosgriff (1965) completely revised the group excluding Mastodonsaurus from the family Capitosauridae to which they assigned three genera : Parotosaurus, Paracyclotosaurus, and Cyclotosaurus with a much reduced number of species in the first and last genera. They commented briefly upon the English remains but decided that the only valid English capitosaur was Cyclotosaurus stantonensis. Recently, several new species have been described by Konzhukova (1965), Bona- parte (1963), Heyler (1969), Chowdhury (1970), Howie (1970), and Ortlam (1970). ‘Capitosaurs’ are known only from the basal Lower Trias to the Upper Keuper. They vary widely in size, the smallest being a skull 70 mm long from the Lower Trias of Queensland, and the largest being skulls of C. hemprichi (about 700 mm) and Mastodonsaurus jaegeri {M. giganteus of early authors) (about 800 mm). Undescribed fragments of skulls from East Africa in the collections of the British Museum (Nat. Hist.) and the Museum of Zoology, Cambridge University show that much larger skulls did occur; possibly as long as 1700 mm. The labyrinthodont collection in Warwick County Museum contains many specimens but only those few which can be identified with reasonable certainty are included here. The rest of the specimens are on the whole very small and extremely fragmentary. Considerable difficulty was experienced with those described as very few show similar areas of the skull and so detailed comparisons between many of them were impossible. Abbreviations preceding specimen numbers. Gz, Warwick County Museum; SM, Sedgwick Museum; R, British Museum (Natural History); GSM, Geological Survey Museum; BSp, uncatalogued specimens from the Geology Department, Birmingham University. Abbreviations. A angular C coronoid ant. pal. vac. anterior palatal vacuity elm. m. cleidomastoideus muscle ART articular attachment area BO basioccipital D dentary PATON: CAPITOSAUROID LABYRINTHODONTS 255 ECT ectopterygoid p.q.f. paraquadrate foramen EO exoccipital PRA prearticular E frontal PRE prefrontal J jugal PSP parasphenoid L lachrymal PT pterygoid l.l.g. lateral line groove Q quadrate MX maxilla q. boss quadrate boss N nasal QJ quadratojugal obi. r. crista obliqua SA surangular P parietal SQ squamosal PE postfrontal ST supratemporal pin. f. pineal foramen T tabular PL palatine V vomer PMX premaxilla vag.(X) f. foramen for vagus (tenth) PO postorbital nerve PP postparietal v.c.d. vena capitis dorsalis Cyclotosaurus leptognathus, specimen R 3174 Until now, this specimen has been known as Cyclotosaurus stantonensis. How- ever, several of the fragmentary skull specimens from Warwickshire described in the following section are, for reasons discussed below (p. 264), believed to be con- specific with C. stantonensis. One of these specimens, Gz 38, was first described and figured by Owen (1842) as Labyrinthodon leptognathus. Labyrinthodon is a junior synonym of Mastodonsaurus. The trivial name leptognathus antedates stantonensis and so the correct name for the species including R 3174, is Cyclotosaurus lepto- gnathus. This is explained below (p. 280). The specimen R 3174 is from the ‘Lower Keuper’ Sandstone of Stanton in Stafford- shire, its exact stratigraphical horizon being unknown. Its existence was first noted by Ward (1900) and it was more fully described by Woodward (1904) as Capitosaurus stantonensis. Zittel (1911) transferred the species to Cyclotosaurus and most authors since have agreed with this. Watson (1958), however, suggested that it should be called Procyclotosaurus on the basis of the deep skull, short exoccipital/pterygoid suture, and the retention of the crista obliqua on the pterygoid, and further suggested that it was a member of his high-skulled lineage not closely related to other cyclo- tosaurs. Welles and Cosgriff (1965) considered the division into low- and high- skulled forms invalid and returned the species to Cyclotosaurus, agreeing, however, that it seemed primitive and aberrant. The specimen is in two parts, the one previously illustrated and discussed being the part with the well-preserved occiput, some of the palate, and the ventral impres- sion of the cranial bones. The hitherto undescribed counterpart showed the under- side of the cranial bones, the dorsal surface being embedded in a thick block of hard sandstone. Only a few sutures on the posterior part of the skull could be traced with certainty, some of those indicated by Woodward (1904) being incorrect. Very little useful comparison with other species showing the dorsal skull roof was thus pos- sible. Preparation of this part of the specimen was carried out after the underside of the cranial bones had been embedded in Plaster of Paris ; and the specimen was completely cleared of the matrix, a coarse, hard, pale-coloured sandstone stained with haematite close to the bone. From this part, and the fragments of bone adhering to the counterpart, the 256 PALAEONTOLOGY, VOLUME 17 1cm. TEXT-FIG. 1. Specimen R 3174, skull roof as recently prepared. complete skull roof can be reconstructed. The specimen is virtually undistorted, the left side being more complete than the right, much of which adheres to the figured part of the specimen. The sutures on the skull are very clear and there can be no doubt about their positions. In fully adult labyrinthodonts, zones of intensive growth are visible as areas where the normal pits in the ornamentation are elongated into troughs (Bystrow 1935). In capitosaurs these zones are located anterior to the orbits at the prefrontal/ frontal/nasal sutures, and postero-lateral to the orbits at the quadratojugal/jugal/ squamosal sutures. In R 3174 the zones are present but not pronounced. This, and the open nature of the sutures, indicates that the skull is that of a young adult. Bystrow and Efremov (1940) studied age changes in Benthosuchus sushkini and from their data it can be predicted that with increasing age the skull would become larger EXPLANATION OF PLATE 35 Specimen R 3174, Cyclotosaurus leptogmithus, skull roof. PLATE 35 PATON, Cyclotosaurus 258 PALAEONTOLOGY, VOLUME 17 and broader ; the orbits relatively smaller and more laterally placed (thus making it more typically cyclotosaurid) ; the pineal foramen would be relatively further behind the orbits ; the occiput would be relatively shallower ; and the posterior edge of the skull roof more concave. The ornamentation consists of wide, deep pits separated by high, narrow ridges which nowhere exceed the diameter of the pits. The lateral line grooves are only obvious in two places : the lachrymal flexure of the suborbital groove ; and the supra- orbital groove on the postorbital, continuing into the jugal groove on the jugal and quadratojugal. These two areas of the grooves are those visible in all capitosaurs where the dorsal surface of the skull roof is preserved. The other grooves can be traced as slightly widened pits connected by lower ridges. Measurements (in mm) of specimen R 3174 (see diagram): B = 167 0 = 137 A = 32 K = 13 G = 31 R = = 61 L = 208 F = 110 D = 48 Y = 112 E = 16 H = = 30 M = 12 S = 72 N = 41 T = 29 P = 7 skull roof J = 21 I = 38 C = 67 OL 23 P = 32 palate Indices used by Welles and CosgrifT (1965): B:L = 80 H:B = 22 C:L = = 32 \ :C = 48 P:C = 10-5 T:C = = 43 S;L = 35 A:L = 15 A:OL = : 75 N :C = 61 K:C = 19 H PATON: CAPITOSAUROID LABYRINTHODONTS 259 When the skull is viewed in profile, it is noticeably ‘dish-faced’ and this is not due to crushing. The skull thus has a very crocodilian appearance as both orbits and, to a lesser extent, the nares are situated above the level of the surrounding areas of the skull. Most capitosaurs have this appearance but it is particularly pronounced in this specimen. Skull roof (text -figs. 1, 3a; Plate 35). The general plan of the skull roof is similar to that of typical members of the Capitosauridae, and this can be seen in the text-figures and plates. There are, however, one or two features of interest. An oval premaxillary foramen is present at the tip of the snout— this was probably present in all capitosaurs but can only be seen in well-preserved specimens. This foramen was probably connected with the anterior palatal vacuity and these may have been openings for a large intermaxillary gland which produced mucus as in modern Amphibia. The anterior commissure of the lateral line system can be seen running parallel with, and close to, the anterior edge of the snout on the premaxillae. No ornament is present here but the position of the commissure is marked by a series of nerve foramina. The jugal forms only a very small (5 mm) portion of the orbital border, being excluded mainly by the anterior expansion of the postorbital. The otic notches are impressions on both skull and counterpart and were obviously closed. They were pear shaped. The posterior edge of the left squamosal has a small, unorna- mented shelf which would be overlapped by the distal end of the tabular. Impressions indicate that the post-otic bar of the tabular was antero-posteriorly unexpanded unlike that of cyclotosaurs, and paroto- saurs with semi-closed otic notches generally. The palate and occiput are described from the already prepared part of the specimen which remained in the British Museum (Nat. Hist.). Some matrix adhering to the occiput tends to be misleading and the palate is incompletely exposed. Several features of interest can be seen. Palate (text-fig. 4). On the right side, the base of the last premaxillary tooth is visible. It and the adjacent socket appear to be larger than the following maxillary teeth. This unusual condition will also be noted in specimen Gz 38. Most of the parasphenoid can be seen in ventral view although the cultriform process is damaged. The narrow central portion of this is sharply keeled ventrally. The basal plate of the para- sphenoid is wide and short and the parasphenoid/pterygoid suture is short— these are both primitive features. The palatal ramus of the pterygoid is deeply concave ventrally and this surface is slightly sculp- tured as in Parotosaurus nasutus. On the left side the anterior face of the quadrate ramus of the pterygoid is exposed ventrally and shows a roughened area presumably for articulation with the hamate process of the prearticular. Occiput (text-figs. 2 and 3b). The processus lamellosus of the exoccipital can be seen and below this at the base of the dorsal process of the exoccipital, a small, medially directed processus basalis is visible separat- ing the partially ossified basioccipital from the foramen magnum. The crista basioccipitalis can be seen. The ridge for the attachment of the cleidomastoideus muscle (Howie 1970) is present on the ventral pro- cess of the tabular. The position of insertion of some of the occipital muscles is visible on the thickened upper edge of the triangular post-temporal fossa. The condyle faces postero-ventro-medially. Its surface is roughened and was covered in life by a cap of cartilage supported by the slight frill of smoother bone around the edge of the condyle. The stapes is not preserved. Posteriorly the quadrate ramus of the pterygoid forms a lateral vertical wing of bone. This has a slightly roughened surface and dorsally shows traces of a slight crista obliqua — not a strong one as previously thought. A large quadrate boss is present on the pterygoid and quadrate just above the inner part of the quadrate condyle. The paraquadrate foramen is very long and is situated in the quadrate/ quadratojugal suture. It is internally subdivided into several subforamina. Its long axis is parallel to the lateral occipital border formed by the quadratojugal. No braincase or otic ossifications other than those mentioned above could be seen. Primitive features q/R 3174: 1, The basal plate of the parasphenoid is wide and short. 2, The parasphenoid/ pterygoid suture is short. 3, The choanae are oval. 4, The snout is pointed. 5, The skull is high. Aberrant features ofR 3174: 1, The post-otic bar of the tabular is narrow. 2, The jugal is nearly excluded from the orbit. 3, The last premaxillary tooth and its adjacent socket are enlarged. TEXT-FIG. 2. Specimen R 3174, Cyclotosaurus leptognathus, occiput. pq.f B q.boss 1cm. TEXT-FIG. 3. Reconstruction of (a) skull roof and (b) occiput of Cyclotosaurus leptognathus. 1cm. EXPLANATION OF PLATE 36 Specimen R 3174, Cyclotosaurus leptognathus, dorsal surface of counterpart. PLATE 36 PATON, Cyclotosaurus 262 PALAEONTOLOGY, VOLUME 17 area 1cm, TEXT-FIG. 4. Reconstruction of palate of Cyclotosaurus leptognathus. Relationships ofK 3174. As the otic notch is closed, R 3174 is a cyclotosaur, albeit a primitive and aberrant one. It seems to me that there is a strong possibility that the genus Cyclotosaurus is in fact a polyphyletic group united by the common feature of the closed otic notch which distinguishes them from parotosaurs. However, at present there is insufficient evidence to separate the members of the group generically and it is therefore convenient to retain them all in the genus Cyclotosaurus. Thus R 3174 is not here placed in a separate genus as it was by Watson (1958). One of the features which distinguished his genus Procyclotosaurus was the presence of a strong crista obliqua on the pterygoid. As Welles and Cosgriff (1965, p. 42) point out, however, the crista obliqua is present in most capitosaurs, its absence being merely an accident of preservation. In any case, the crista obliqua in this specimen is only slightly developed. PATON: CAPITOSAUROID LABYRINTHODONTS 263 Fragmentary skull specimens Identification of skull fragments A visit to Ludwigsburg and Tubingen allowed study of the holotypes of Masto- donsaurus jaegeri Meyer (1844) from the Lettenkohle of Gaildorf, Cyclotosaurus posthumus Fraas (1913) from the Stubensandstein of PflFalfenhofen in Bavaria, C. mordax Fraas (1913) from the Stubensandstein of Pfaflfenhofen, and C. robustus (Quenstedt) Fraas (1889) from the Schilfsandstein north of Stuttgart. Other material of C. robustus and Mastodonsaurus jaegeri was examined, in particular a very well- preserved but fragmentary skull of the latter. The holotype and only skull of Cyclotosaurus posthumus is not complete as has been thought. The anterior part of the cranium and palate has been modelled in plaster. This area can be clearly seen in Fraas’s (1913) plates as the unornamented part of the snout and a corresponding area of the palate. Plaster is also responsible for the unusual drop-shaped orbits of Mastodonsaurus Jaegeri shown in all recon- structions. The holotype has been broken across the frontals just posterior to the prefrontal/frontal suture, and this break has been repaired with plaster which has been allowed to bulge into the orbits. These should be completely oval as are those of other specimens of the species. It should also be noted that the otic notches of the holo- type of M. jaegeri, and of all other known mastodonsaurs, are definitely wide open. Fraas (1889, pi. 1) is slightly misleading in this respect as Welles (1947, p. 264) noticed. As a result of the study of these specimens it is suggested that the genus Masto- donsaurus can be distinguished from the ParotosaurusI Cyclotosaurus lineage by the nature of the ornamentation and of the lateral line grooves. Different authors describe the ornamentation of labyrinthodonts in different terms. What might be coarse to one author is fine to another and thus it is virtually impossible to deduce from a description, diagram, or even from a photograph what the ornamentation is actually like. I have therefore confined my remarks to the admittedly limited number of specimens which I have seen myself. However, as these include the above holotypes and other specimens of these species, plus several other specimens of mastodonsaurs, parotosaurs, and cyclotosaurs, with casts of others, it is thought to be a reasonable conclusion. In all cases size has been taken into account. In Mastodonsaurus (using data from the holotype, another specimen and frag- ments of M. jaegeri', fragments of M. keuperinus', and fragments of Labyrinthodon jaegeri from Germany which is obviously a mastodonsaur) the ornamentation con- sists of shallow pits separated by wide ridges which vary in height. The depth of the pits varies but in all places the ridges form the most prominent part of the ornamenta- tion, often being as wide as the pits between them. This type of ornamentation appears to be present in juvenile specimens of Benthosuchus (Bystrow 1935, fig. 15). The lateral line grooves of Mastodonsaurus are very wide and shallow with respect to their width. This varies— they are slightly narrower posteriorly— but in all places the grooves are wider than the pits of the ornamentation. The ornamentation of Cyclotosaurus consists of deep pits separated by high, narrow ridges; the pits thus form the most prominent part of the ornamentation and are deep all over the skull. The ridges are never as wide as the pits. The lateral line grooves are also deep and narrow. Only close to the external nares do they exceed 264 PALAEONTOLOGY, VOLUME 17 the diameter of the pits, and then only by about one and a half times. Over the rest of the skull the grooves can only be distinguished from the pits by the fact that they are continuous. In places lower ridges do cross them and this may be why lateral line grooves are often incompletely figured in the skulls of cyclotosaurs. Specimens of large and small cyclotosaurs show this type of ornamentation and lateral line grooves, and that of parotosaurs, both adult and juvenile, is essentially the same. Very few lower jaws were available for examination but it seems that their orna- mentation is generally coarser than that on the corresponding skull roof. The lateral line grooves remain wide and shallow on the lower jaws of mastodonsaurs, and in cyclotosaurs they are narrow but more superficial than those of the skull roof. No differences in the ornamentation of parotosaurs and cyclotosaurs could be detected, but this would be expected as the two genera are very closely related. The differences in ornamentation of these from mastodonsaurs corroborates the view that the two lineages are not closely related. Possibly mastodonsaurs evolved from primitive benthosuchids retaining some juvenile features such as the ornamentation and the large orbits. Ornamentation is useful in the identification of some other labyrinthodonts, e.g. Peltobatrachus {?dLnc\\Qn 1959), Plagiosaurus 1937), Dvinosaurus { Amalitsky 1924); but these all possess very distinctive types of ornament. A survey of laby- rinthodonts with the more normal reticulate sculpture would be interesting in case such small differences as those found between mastodonsaurs and parotosaur/ cyclotosaurs are widespread. The ornamentation of Parotosaurus (Stenotosaurus) semiclausus will be discussed later. Differences in ornamentation corresponding to those described above can be found in most of the specimens from the English Trias. That of R 3174 is typical of the cyclotosaurid type. It is noteworthy that the specimens identifiable by other means, e.g. Gz 20, the interorbital plate; Gz 14, the otic notch region; all confirm these differences. The specimens fall into three groups: a species of mastodonsaur, and two species of cyclotosaur. The mastodonsaur specimens represent skull lengths of 400 to 600 mm. The skull had relatively large orbits, but not as large as those of Mastodonsaurus jaegeri. The species closest to the English one in relative orbital size seems to be M. cappelensis (see p. 266), of which one specimen was seen in Tubingen. The English species is probably later in age— M. cappelensis is from the Upper Bunter and is considered by Welles and Cosgriff (1965) to be a suitable ancestor for M. jaegeri. Possibly the English species is a little modified descendant of M. cappelensis or a closely related form. The material is thought to be sufficiently distinctive to allow a diagnosis of a new species of Mastodonsaurus— M . lavisi—io be made (see p. 282). One of the species of Cyclosaurus is C. leptognathus. The snout Gz 38 is very similar indeed to that of R 3174. One unusual feature in both is the enlargement of the last premaxillary tooth and its adjacent socket. Gz 6 is also very similar to the same area in R 3174, the paraquadrate foramen being relatively very large in both specimens. Gz 1 1 is included here as it seems to come from a deep-skulled cyclotosaur. The lower jaw Gz 35 is slender and is considered to be the most likely type of jaw to be associated with a skull such as R 3174. Thus Gz 38, Gz 6, Gz 11, and Gz 35 indicate the presence in the ‘Lower Keuper’ Sandstone of a small, deep-skulled cyclotosaur PATON: CAPITOSAUROID LABY RINTHODONTS 265 with a primitive anterior palatal structure, unusual last premaxillary teeth, slender snout, and large paraquadrate foramina. All these characters agree closely with those in R 3174 and the specimens are therefore thought to be conspecific. The other specimens attributed to the genus Cyclotosaurus apparently belong to a more typical cyclotosaur of moderate size with a shallow skull. Specimens Gz 14 and Gz 26 are typically cyclotosaurid with laterally elongated and distally expanded tabulars. They show differences in the shape of the tabular and in its relations to the surrounding bones, but these differences are only what would be expected from inter-individual variation. Wide variations in the positions of sutures and the rela- tive sizes of different bones have been noted in species known from many specimens, e.g. Benthosuchus sushkini (Bystrow and Efremov 1940); Biiettneria bakeri (Case 1932). In addition, Gz 13 is included here because of the shallowness of skull which it indicates, and Gz 36 because it too shows a shallow skull, and also because the anterior part of the palate shows a circular choana (a more advanced cyclotosaur character). GSM 27964 and BSp 2 are included here because they seem to be typical cyclotosaur lower jaws. The shape of the otic notch and its orientation, which varies in different species of cyclotosaurs, is very close to that of Cyclotosaurus posthumus and the tabulars appear to be similarly constructed in this species. No diagnosis is possible for these few fragments, but for ease of reference the species has been named C. pachygnathus (see p. 280). The fragmentary specimens are here described grouped in the above three species : 1 . Specimens attributed to Mastodonsaurus lavisi : Specimen Gz 20 (text-fig. 5). Identified by Miall (1874, pi. XXVI, fig. 1a) as the interorbital plate of Mastodonsaurus pachygnathus. TEXT-FIG. 5. Specimen Gz 20, dorsal view. 266 PALAEONTOLOGY, VOLUME 17 It is uncrushed. The reconstructed orbit is oval with the long axis almost antero-posterior and 57 mm long. The transverse axis is 31 mm. The lateral lines are very wide and deep and the ornamentation is of shallow pits separated by wide ridges. From Table 1 it can be seen that the only genus having a ratio of interorbital width: orbital length greater than 100% is Mastodonsaums. Because of this, and the nature of the ornamentation, the specimen is assigned to this genus and is considered to be closest to M. cappelensis. TABLE 1. Comparative ratios of orbit measurements m different species of eapitosauroids. Interorbital width: orbital length (%) Specimen Gz 20 121 Mastodonsaums cappelensis 123 M. acuminatus 207 M. jaegeri 286 Cyclotosaums robustus 81 C. leptognathus 75 C. posthumus 62 Interorbital width: orbital length (%) C. ebrachensis 62 C. hemprichi 45 Paracyclotosaums davidi 74 Parotosaums nasutus 74 P. hang lit on i 78 P. semiclausus 76 P. brookvalensis 85 Specimen Gz 9 (text-fig. 6). Identified by Miall (1874, pi. XXVI, fig. 2a, b, c) as a portion of the occipital border of Mastodonsaums sp. It shows the left otic notch and tabular horn. The posterior face of the tabular shows the ridge and trough for the dorsal attachment of the cleidomastoideus muscle used in raising the head (Howie 1970), TEXT-FIG. 6. Specimen Gz 9. a, dorsal view; b, ventral view; c, posterior view. PATON: CAPITOSAUROID LABYRINTHODONTS 267 the other end being attached to the dorsal process of the clavicle. Below the wall of squamosal forming the otic notch a small fragment of the quadrate ramus of the pterygoid has been displaced anteriorly for 18 mm. Although the specimen has been weathered, it is apparent that the otic notch was widely open, and the tabular horn resembles that of M. jaegeri. The ornamentation shows deep pits separated by low, wide ridges and for these reasons the specimen is assigned to the genus Mastodonsaurus. Lateral line grooves TEXT-FIG. 7. A, Specimen Gz 26, dorsal view; b, specimen Gz 1057, dorsal view. Orbital edge 1 cm. Posterior border otic notch Specimen Gz 1057 (text-fig. 7b). Identified by Miall (1874, pi. XXVI, fig. 1b) as the left postorbital of Mastodonsaurus sp. It was figured m conjunction with Gz 20, but there is no record of any association between the two. It is the left post- orbital with anteriorly, part of the jugal. The anastomosis of the supra- and infraorbital grooves is visible. The ornamentation is of deep pits separated by wide ridges. The orbit is large both relatively and abso- lutely, and the postorbital is unexpanded anteriorly. These features indicate that the specimen belongs to the genus Mastodonsaurus. Specimen BSp 1 (text-fig. 8). This is an unfigured specimen from the ‘Keuper’ Sandstone of Bromsgrove. It is part of the right upper jaw. The maxilla projects laterally as a wedge whose upper surface forms the base of a large and wide lateral line groove running in the jugal/maxillary suture. The ventral surface of the maxilla is separated from the palate by a step 8 mm deep. The external palatal element is the ectopterygoid. Internal to its tooth row, the bone is perforated by many small, posteriorly directed foramina which are probably for small blood vessels supplying the mucosa of the roof of the mouth which assisted in respiration. The extent of the other palatal bones is not great. The dorsal aspect of the specimen shows a small fragment of the dermal ornamentation 34 mm from the anterior end, indicating that only a small part of the bone is missing dorsally. E 268 PALAEONTOLOGY, VOLUME 17 1 cm. TEXT-FIG. 8. Specimen BSp 1. a, ventral view; B, dorsal view. The estimated skull length of the animal is 500-550 mm, and the large and wide lateral line groove, the fragment of ornamentation showing a wide pit, and the antero-posteriorly compressed teeth indicate that it was a species of Mastodonsaurus. Specimen SM 369 (text-fig. 13c). Identified by Wills (1907) as the cranial bones of a labyrinthodont (1 Mastodonsaurus). It fs part of the posterior edge of the skull. The ornamentation consists of shallow pits and wide ridges. The skull length is estimated at between 500 and 600 mm and the posterior border of the skull is only slightly concave. The ornamentation suggests that it is from a mastodonsaur. 2. Specimens attributed to Cyclotosaurus pachygnathus: Specimen Gz 14 (text-fig. 9). Figured, but not interpreted, by Owen (1842, pi. 46, figs. 6, 7) as Lahyrinthodon pachygnathus. Miall (1874) correctly identified it as the right otic notch area but confused the otic notch and the postero- ventral corner of the skull. He also referred it to his new genus Diadetognathus. The three bone masses on the specimen are obviously closely related and their correct positions in relation to one another can be restored (text-fig. 9b). The otic notch in this specimen was completely closed as a result of the postero-lateral growth of the tabular and it must therefore belong to the genus Cyclotosaurus. The ornamentation, of deep pits separated by high, narrow ridges, confirms this. Specimen Gz 36 (text-fig. 10). Identified by Owen (1842, pi. 43, figs. 9, 10) as the left maxilla and intermaxilla with the palatal plate of Lahyrinthodon pachygnathus. The ornamented part of the specimen is the right jugal. This is broken PATON: CAPITOSAUROID LABYRINTHODONTS 269 1 cm. A TEXT-FIG. 10. Specimen Gz 36, dorsal view. Edge of mterpterygoid vacuity Orbital edge Base of spina sublacrimalis Lateral line grooves Inte^ocking edge of suture with MX Edge of choana 1cm. 270 PALAEONTOLOGY, VOLUME 17 34 mm from the anterior end revealing part of the underlying palate with portions of the right inter- pterygoid vacuity and the subcircular choana. Between the choana and the interpterygoid vacuity is the base of a ridge which may be the spina sublacrimalis, a posteriorly directed spine with a broad base on the palatine close to the choana. This was figured by Bystrow and Efremov (1940) in Benthosiichus sushkini but is not known in other capitosaurs, probably because the dorsal surface of the palate is very rarely exposed. The palatal element has been displaced posteriorly and twisted slightly clockwise. The ornamentation consists of deep pits separated by high, narrow ridges. The animal apparently had a small orbit and these features suggest that it belonged to a cyclotosaur. It had a shallow skull. Specimen Gz 13 (text-fig. 11a, b). Identified by Miall (1874, pi. XXVII, fig. 1a, b) as the left postero-external angle of the skull of Masto- donsaurus pachygnatims. The ornamentation is of shallow pits and narrow ridges. Just ventral to the para- quadrate foramen and wholly on the quadratojugal, is a large oval projection. Its extreme lateral position shows that it cannot be the quadrate boss normally found close to the articulation. No similarly placed A Paraquadrate foramen QJ 1 cm. TEXT-FIG. 1 1. Specimen Gz 13. A, posterior view; b, lateral view; c, specimen Gz 1 1, lateral view. PATON; CAPITOSAUROID LABYRINTHODONTS 271 projection has been seen on other specimens and its position just above the lateral condyle must have inhibited the jaw mechanism. Possibly this lump is pathological in origin, the bone comprising it has an unusual surface texture not seen anywhere else, and the specimen’s general appearance suggests that it came from an old animal. The condylar surface is roughened and was covered by cartilage in life. The estimated skull length is about 450 mm and the occiput was very shallow. The nature of the ornamenta- tion suggests that it came from a cyclotosaur. Specimen Gz 26 (text-fig. 7a). Identified by Miall (1874, pi. XXVII, fig. 4a, b) as the epiotic and adjacent bones of IMastodonsaums. It is almost the complete right tabular with parts of the surrounding bones. The ornamentation is of deep pits and high narrow ridges. The tabular is distally expanded. Some parotosaurs have otic notches which are almost closed and in these the tabular has a lateral extension like that of Gz 26. However, direct com- parison shows that the extension in this specimen is relatively greater and therefore the otic notch was closed. Thus the specimen is from a cyclotosaur and the nature of the ornamentation confirms this. 3. Specimens attributed to Cyclotosaurus leptognathus: Specimen Gz 38 (text-fig. 12). Identified by Owen (1842, pi. 43, figs. 1, 2, 3) as Lahyrintliodon leptognathus and designated by Miall (1874) as type. It shows part of the snout from the posterior edge of the anterior palatal vacuity, to about the level of the anterior edge of the interpterygoid vacuities. Part of the right side is missing. The dorsal surface is flat and the specimen has been slightly crushed dorso-ventrally. The ornamentation is of small, deep pits and high ridges. There are two narrow and deep lateral line grooves. The premaxilla is only preserved posteriorly and bears one large, antero-posteriorly flattened tooth broken 6 mm from its base which is 5 x 4 mm. It is followed by a large socket. The circular maxillary teeth are much smaller (2 mm diameter anteriorly) and are arranged as follows (I = tooth present; X = tooth absent) : ^ ^ ^ , Premax. / Max. IX /IXIXIXIXIIIIXIXIXIXIXIXI A B TEXT-FIG. 12. Specimen Gz 38. a, dorsal view; b, ventral view. 272 PALAEONTOLOGY, VOLUME 17 Thus, apart from the group of four adjacent teeth, alternating tooth replacement occurs. All the teeth except (B) are broken, many showing the pulp cavity, (b) is 3-5 mm long and had probably just erupted as the adjacent teeth are equally long but broken. All the teeth except (a) are firmly fused to the bone of the upper jaw. (a) however, a fairly large tooth, may have been loosening in the jaw prior to shedding. Just lateral to the palatine tooth row are the crushed remains of two palatine tusks, the inner anterior one having been pressed down upon the outer one. The specimen belonged to an animal with a skull length of about 200 mm. Its shape shows that the snout was fairly pointed as in Cyclotosaurus " stantonensis', this shape being primitive for a cyclotosaur. It is assigned to the genus Cyclotosaurus, and its ornamenta- tion confirms this. Specimen Gz 6 (text-fig. 13a, b). Identified by Owen (1842, pi. 43, fig. 11) as the anterior frontal of Labyrinthodon pachygnathus. Miall (1874) reidentified it correctly as part of the occipital region. It is part of the left postero- ventral corner of the skull. The ornamentation is of shallow, narrow pits and high ridges. Traces of a narrow, shallow lateral line groove are present; it is difficult to distinguish from the ornamentation. The paraquadrate foramen is large and oval ; there are four small foramina in the outer rim and inside is a large dorsal foramen with TEXT-KiG. 13. Specimen Gz 6. a, lateral view; b, posterior view; c, specimen SM 369, dorsal view. PATON; CAPITOSAUROID LABYRINTHODONTS 273 two smaller ventral foramina and two or three accessory foramina. The quadrate/quadratojugal suture can be traced across the foramen. The estimated skull length from the specimen is 300-400 mm. The paraquadrate foramen is very large, a feature which also occurs in R 3174. The specimen is thought to have come from a cyclotosaur with a deep skull. Specimen Gz 1 1 (text-fig. 1 Ic). Identified by Miall (1874, pi. XXVII, fig. 2) as the right postero-external angle of the skull. However, he oriented it incorrectly and it is part of the left side just in front of the postero-ventral corner of the skull. The ornamentation is of fine pits and high ridges. There is a deep, narrow lateral line groove parallel to the lower edge and 22 mm above it. At the posterior edge of the specimen it appears to curve upwards following the broken edge which is probably the position of the squamosal/quadratojugal suture. Dorsally the bone thickens and the base of a downwardly projecting spur can be seen; this is presumably part of the dorsal support for the lateral wall of the braincase. The specimen indicates a skull length of about 250 mm. The nature of the ornamentation and lateral line groove suggest it belongs to a cyclotosaur and the skull appears to have been deep. Lower jaws The general strueture of capitosaur lower jaws is well known and need not be dis- eussed here (see Romer 1947; Nilsson 1943, 1944; Wills 1915; Howie 1970). The specimens of lower jaws all conform to this general pattern. Lower jaws from War- wickshire are as follows : Specimen Gz \5^ Mastodonsaunis pachygnathus Miall, 1874, no. 6, pi. XXVI, fig. 3a, b. The posterior end of a right lower jaw from just in front of the anterior end of the adductor fossa. Identified here as M. lavisi (text-figs. 14a, 16a). Specimen Gz 31 —Diadetognatlius varvicensis Miall, 1874, no. 9. Part of a right lower jaw showing the anterior part of the adductor fossa and the posterior part of the posterior Meckelian fossa. Identified here as Mastodonsciurus lavisi (text-fig. 14c). Specimen Gz 35— Diadetognathus varvicensis Miall, 1874, no. 8, pi. XXVII, fig. 3a, b. The posterior end of a left lower jaw showing the same area as Gz 15. Here referred to Cyclotosawus leptognathus (text -figs. 15a, 16c). Specimen GSM 21964— Mastodonsaurus Jaegeri Huxley, 1859; Wills, 1915, pi. 3. The posterior end of a left lower jaw from the middle of the adductor fossa. Here referred to Cyclotosaurus pachygnathus (text- figs. 15b, 17b). Specimen Gz 21 — Labyrinthodon leptognathus. Two small pieces of lower jaws, one the anterior part and the other a central part with a portion of the coronoid series. These are not considered further as the posterior part, only, of the lower jaw shows diagnostic features. Several other fragments of lower jaws are in Warwick County Museum, but these are mainly the undiagnostic anterior parts, some of which were figured by Owen (1842). The following specimens come from quarries at Bromsgrove: Specimen BSp 2—1 Diadetognathus Wills, 1915, no. A, pi. 2, figs. A, B, C, D. The posterior end of a right lower jaw from the anterior end of the adductor fossa. Identified here as Cyclotosaurus pachygnathus (text-figs. 15c, 17a). Specimen BSp 3— Labyrinthodon leptognathus Wills, 1915, no. B, pi. 2, figs. E, F, G. The anterior end of a left lower jaw. Not diagnostic. The following specimens come from Sidmouth : Specimen R 42\5— Labyrinthodon lavisi Seeley, 1876, pi. XIX, figs. 1, 2, 3. The posterior end of a right lower jaw from the posterior Meckelian fossa. Identified here as Mastodonsaurus lavisi (text-figs. 14b, 16b). 274 PALAEONTOLOGY, VOLUME 17 Groove for posterior lange of dentary Lateral line grooves TEXT-FIG. 14. A, specimen Gz 15, lateral view; b, specimen R 421 5, lateral view (anterior part of specimen omitted) ; c, specimen Gz 37, lateral view. Specimen R 33\^Mastodonsaurus. A portion of a right lower jaw at the level of the adductor fossa. Referred here to M. lavisi. Some specimens thus show the posterior ends of the lower jaws and can therefore be directly compared (Table 2). Specimens Gz 15 and Gz 35, although identical in size and area, show many differences as indicated in the following comparisons: 1, The specimens are of the same size but Gz 15 is much more massive than Gz 35. 2, The lateral line grooves are much wider and deeper in Gz 15. 3, The ornamentation is much coarser in Gz 15. 4, In Gz 15 there is a triangular area of ornamentation posterior to the lateral line grooves on the external side of the retroarticular process. 5, The base of the jaw is much wider and more convex in Gz 15. PATON: CAPITOSAUROID LAB Y RINTHODONTS 275 TABLE 2. Comparisons between lower jaws present in the material and others examined. B5p2 Gz 15 Gz 35 GSM 27964 R 4215 P. pronus P peabodyi M, jaeger i Size Massive Massive Slender Massive Massive Massive Slender ? Massive Lateral lines Shallow & wide Deep & wide Shallow 8i wide Shallow 8. wide Deep 8i wide Shallow 8i wide 7 Shallow 8i wide Ornament Coarse 8. shallow Fine & deep Fine 8. shallow Coarse 8< deep Fine 8< deep Coarse 8. deep 7 Coarse & shallow Area of ornament external to retroarticular process Absent Present Absent Absent Present Absent 0 Present Width of base of jaw Narrow Wide Narrow Narrow Wide Narrow 7 Wide Shape of base of jaw Slightly convex Slightly convex More convex More convex Slightly convex More convex Slightly convex More convex Shape of retroarticular process Rounded Angular Rounded Rounded Angular Rounded Rounded Rounded Shape of surface of retroarticular process Triangular + shelf Shelf Triangular Triangular Shelf Triangular + shelf Triangular Triangular + shelf Articulation 2 grooves Saddle 2 grooves Saddle Saddle Saddle ? 2 grooves ? Saddle Position of chorda tympani foramen High Low High High Low High High High Length of adductor fossa Long Long Short 7 Long Short Short 7 Length of tooth row Long Short Long Short Short Medium Medium 7 Surface of coronoid Ornament Ornament No ornament 7 Ornament . ? 7 7 276 A PALAEONTOLOGY, VOLUME 17 1 cm -~7i Lateral line grooves Latera grooves TEXT-FIG. 15. A, specimen Gz 35, lateral view; b, specimen GSM 27964, lateral view; c, specimen BSp 2, lateral view. PATON: CAPITOSAUROID LABYRINTHODONTS 111 6, The retroarticular process is more rounded and much narrower in Gz 35. 7, The retroarticular process has a definite dorsal surface in Gz 35 formed of a triangular hollow enclosed by ridges. No definite dorsal surface is present in Gz 15— the retroarticular process slopes smoothly from a high external edge down on to the inner side of the jaw. 8, The articulations are entirely different. In Gz 35 the articulation is limited posteriorly by a high post- condylar ridge but there is no precondylar ridge. The articulation itself consists of two parts: the lateral part is made up of a shallow trough bounded externally by the posterior extension of the lateral edge of the adductor fossa, and internally by a slightly higher ridge. This trough runs antero-posteriorly and widens anteriorly. The medial part consists of another trough at an angle to the lateral one. It is much deeper and has a V-shaped cross-section. It is bounded internally by the edge of the prearticular and laterally by the inner ridge. A possible explanation for this unusual structure is given later. In Gz 15 the articulation is limited by high pre- and postcondylar ridges. The articulation is not in two parts but shows a normal saddle shape and is limited laterally by a low ridge of surangular. Medially it slopes gently down to the angular/surangular/prearticular suture. The prearticular here forms a horizontal shelf which is probably the base of the hamate process. The articular surface is of cartilage finished bone. 9, Because of the differing articulations the chorda tympani foramen is much lower in position in Gz 15. 10, In Gz 15 the adductor fossa is 1 12-5 mm long; in Gz 35 it is 92 mm long. 1 1, The tooth row ends more posteriorly in Gz 35 and runs along the edge of the anterior part of the adductor fossa. 12, In Gz 15 the surface of the coronoid is ornamented by small foramina with little ridges radiating from them. In Gz 35 the coronoid has a fairly smooth surface. 13, A coronoid tooth is present in Gz 35; it is broken but fairly large. No coronoid tooth is present in Gz 15. The specimen R 4215 is almost identical with Gz 15 and the two must have be- longed to animals from the same species. R 331 and Gz 37 also appear to have come from this species. GSM 27964 and BSp 2 are very similar and they both resemble the lower jaw of Parotosaurus pronus. In many features, e.g. the structure of the retroarticular process and the nature of the lateral line grooves, they also resemble Gz 35. As described above, this has an unusual articular surface showing two troughs separated by a high ridge. Specimen BSp 2 has a similar articulation but no other jaws with this unusual structure have been seen. However, Welles and Cosgriff (1965) figure the lower jaw of P . peabodyi an articulation made up of two troughs but do not comment upon it. Comparison with Gz 15 and GSM 27964 has suggested a possible explanation. In Gz 15 the articular is well ossified, that of GSM 27964 is less so and those of Gz 35 and BSp 2 are poorly ossified. BSp 2 probably came from a young individual as the sutures are fairly open and the ornamentation shows juvenile characters. In all four jaws the articular/surangular and articular/prearticular sutures can be followed across the retroarticular process and the latter suture can be traced on the inner side of the cotylus forward to where the articular forms the thin sheet of Meckel’s cartilage lining the inner side of the jaw. However, in Gz 15 the articular covers the whole of the cotylus by overlapping the surangular labially in a thin sheet of cartilage bone. This sheet is broken laterally and the underlying surangular can be seen, its surface being covered with small ridges which held the articular in place. The articular thus forms the entire surface of the saddle-shaped articulation. In Gz 35 and BSp 2, with poorly ossified articulars, it seems likely that this sheet of articular overlapping the surangular is missing either because it was cartilaginous or because it has fallen away due to lack of ossification. Thus the normally unexposed surangular surface is visible— it has ridges corresponding with those seen in Gz 15 for securing the articular. With the hollowed-out articular on 278 PALAEONTOLOGY, VOLUME 17 SA- -PRA Depression in SA for ART Postcondylar_ ridge ART SA~ Retroarticular_ process P;'sw\ 1cm. TEXT-FIG. 16. A, specimen Gz 15, dorsal view; b, specimen R 4215, dorsal view (anterior part of specimen omitted); c, specimen Gz 35, dorsal view. the lingual side filled with cartilage, and the surface of the surangular covered with a sheet of cartilage, the articulation would have a normal saddle shape. Possibly this has also occurred in P. peabodyi. The retroarticular processes of Gz 15 and Gz 35 are very different, but GSM 27964 shows an intermediate structure although it is much closer to that of Gz 35 and BSp 2. That of Gz 15 has a high labial side which slopes steeply down lingually. The surface of that of Gz 35 is triangular in dorsal view with labial and lingual edges of equal height and with a deep triangular trough between them. Those of GSM 27964 and BSp 2 are also triangular in dorsal view but the labial side is much higher than the lingual side and there are traces of a lingual shelf present, more pronounced in GSM 27964. Thus three types of lower jaws are represented by these specimens. Gz 35, GSM 27964, and BSp 2 appear to be fairly closely related and the latter two are similar to P. pronus. As parotosaurs and cyclotosaurs are very closely related — possibly some early cyclotosaurs may have had open otic notches at an early growth stage — it seems reasonable to assume that the lower jaws would not show many differences. For this reason, and also because of the inconspicuous lateral line grooves, GSM PATON: CAPITOSAUROID LABYRINTHODONTS 279 27964 is thought to belong to a typical cyclotosaur with a skull length of about 450 mm. Gz 26 is a cyclotosaur tabular of comparable size. Specimen BSp 2 appears to belong to a smaller, juvenile individual of the same species. Gz 35 shows many slight differences and could belong to an aberrant cyclotosaur such as Cyclotosaurus leptognathus. Welles (1947) describes angulars assigned to the aberrant C. randalli which resemble Gz 35 in that the lower border is narrow with the sculpture ending close to the border, and also showing retroarticular processes with similar dorsal triangular hollows. Gz 15 and R 4215 are thought to belong to the genus Masto- donsaurus because: {a) the retroarticular processes are similar in shape to those of M. jaegeri, (b) the lateral line grooves are large and wide, and (c) in M. jaegeri, Gz 15 and R 4215, there are triangular areas of ornamentation on the external sides of the retroarticular processes. Capitosaur postcranial material, other than from the shoulder girdle, is usually scarce. Only one fragment which could be definitely assigned to a capitosaur was found, and this was specimen Gz 1050, the head of a lumbar rib— the position in the vertebral column being indicated by the angle between the capitulum and tuber- culum— from a very large animal. Shoulder girdle material is usually plentiful, e.g. at Tubingen most of the collection consists of clavicles and interclavicles. However, process TEXT-FIG. 17. A, specimen BSp 2, dorsal view; b, specimen GSM 27964, dorsal view. 280 PALAEONTOLOGY, VOLUME 17 only one specimen from England, in the Geology collection at Birmingham, shows the shoulder girdle in any completeness. It is not assignable as it is unassociated with skull material. The only other shoulder girdle material consists of a few frag- ments of the thoracic plates. Differences between mastodonsaurs and parotosaur/cyclotosaurs Mastodonsaurs Anterior palatal fenestrae paired Lower jaw tusks protrude through dorsal surface of snout Ornament coarse and shallow Lateral line grooves very wide Orbital length greater than 100% of interorbital width Orbits large and midway along skull Parotosaurlcyclotosaurs Anterior palatal fenestra single Lower jaw tusks do not protrude through dorsal surface of snout Ornament fine and deep Lateral line grooves narrow Orbital length less than 100% of interorbital width Orbits small and in posterior half of skull Mastodonsaurs can also be distinguished from advanced cyclotosaurs by the following characters: Mastodonsaurs Choanae elongated and oval Teeth relatively small, antero-posteriorly flattened and many in number Tusks very large Skulls high Advanced cyclotosaurs Choanae circular Teeth fairly small, circular, and fewer in number Tusks moderately large Skulls flattened NOMENCLATURE OE BRITISH CAPITOSAUROIDS Owen (1841) proposed the name Labyrinthodon as a substitute for Mastodonsaunis Jaeger, 1828, which he thought was inappropriate. Labyrinthodon is thus a junior synonym of Mastodonsaurus. The type species of Capitosaurus, C. arenaceus Munster, 1836, was based upon a specimen which lacks the diagnostic postorbital region with the otic notch and thus the genus is indeterminate. As it is from the Upper Trias, it seems likely that it is a cyclotosaur. The generic name Capitosaurus is therefore restricted to the unique undiagnostic holotype, but the family name Capitosauridae has been re- tained for reasons which Welles and Cosgriff (1965, p. 8) explain. It would be convenient to be able to use the trivial name pachygnathus in references to the more typical cyclotosaur represented by some of the specimens. Since Owen’s 1 842 paper is one of the earliest descriptions of labyrinthodont material, the species ^ pachygnathus' could only be upset and rendered synonymous if it was found to be conspecihc with Capitosaurus arenaceus Munster (1836, p. 580). This is unlikely as the specimens show non-homologous parts and also come from different horizons. No more material is likely to come from Warwick and it therefore seems safe to refer to the material as ffachygnathus' and convenient to designate specimen Gz 14 as the lectotype of Cyclotosaurus pachygnathus (Owen) 1842 (Gz 36, Gz 13, Gz 26, GSM 27964, and BSp 2 are paralectotypes). No diagnosis can be given for this species at present, but this action has been taken in order to stabilize the nomenclature and for ease of reference. The other cyclotosaur specimens are those which, with R 3174, are referred to one PATON: CAPITOSAUROID LABYRINTHODONTS 281 TEXT-FIG. 18. Comparative diagram of skulls of various capitosauroids reduced to same size: a, Paroto- saums peabodyi', from Welles and CosgrifF(1965); b, Parotosaurus pronus ,ixom Howie (1970); c, Eocycloto- saurus woschmidti\ from Ortlam (1970); d, Cyclotosaurus leptognathus', e, Cyclotosaurus post humus \ from Fraas (1913); f, Cyclotosaurus robustus; from Fraas (1889); g, Mastodonsaurus cappelensis\ from Wepfer (1923); h, Mastodonsaurus jaegeri', from Meyer (1844). species of a small, primitive, and slightly aberrant cyclotosaur. They comprize Gz 38, Gz 6, Gz 11, Gz 35, and R 3174. Diadetognathus Miall, 1874 antedates Cycloto- saurus Fraas, 1889. Thus, logically but rather unfortunately as Cyclotosaurus is such a well-known name, Cyclotosaurus is a junior synonym of Diadetognathus. However, Miall (1874) described four specimens (syntypes) as belonging to Diadeto- gnathus varvicensis without designating a holotype: Miah’s no. 8 = Gz 35; 9 = Gz 37 {Mastodonsaurus here); 10 = Gz 18— a fragment of a left lower jaw; 11 = Gz 974— a fragment of jaw, unidentihable as to position. Miall’s specimen no. 11, Gz 974 is designated lectotype. As this specimen has no diagnostic features, D. varvieensis is unrecognizable from a nomenclatorial point of view and so other specimens referred to it by Miall (1874) can be transferred to different genera and species if required. This procedure enables the name Cyclotosaurus to continue in general use and specimens Gz 38, 6, 11, and 35, and R 3174 are assigned to C. leptognathus (the trivial name antedates stantonensis), with Gz 38 as lectotype (Miall 1874, p. 430). It was originally intended to retain the name "pachygnathus' for the mastodonsaur species but Owen (1842) figured only cyclotosaur specimens as Labyrinthodon pachygnathus, so this was impossible. In 1876 Seeley proposed the name L. lavisi 282 PALAEONTOLOGY, VOLUME 17 for the portion of lower jaw found at Sidmouth, here referred to the mastodonsaur species. The trivial name must be retained, with the lower jaw R 4215 as holotype, as Mastodonsaurus lavisi. Cyclotosaurus leptognathus (new comb.) (Owen 1842) Lectotype. Gz 38. Part of snout. Paralectoiypes. Gz 6, Gz 11, Gz 35. Cranial material plus lower jaw. Referred specimen. R 3174— Holotype of Cyclotosaurus stanlonensis Locality of lectotype. Coten End near Warwick. Horizon. ‘Lower Keuper’ Sandstone. Top of the Building Stones Formation. Diagnosis. A small cyclotosaur with skull broad posteriorly (B:L = 80) and with slender snout (S:L = 35) and deep occiput (H:B = 22); orbits small, close together (A:OL = 75), separated by a shallow depression, situated slightly above general skull level, oval with long axis antero-posterior ; frontal and jugal enter orbital margin ; postorbital is anteriorly expanded; interpremaxillary foramen present; nares elongate, oval, lateral with long axis parallel to skull border, slightly elevated above surrounding skull surface; pineal foramen circular, close behind orbits; otic notch closed, post-otic bar of tabular narrow; supratemporal excluded from otic notch; posterior skull border moderately concave (index =19); lateral line grooves narrow; ornament fine and deep. Anterior palatal vacuity reniform; choanae elongate with long axis parallel to skull border; last premaxillary tooth large; pterygoid/parasphenoid suture short; pterygoid has facet for jaw articulation. Foramen magnum broad, separated from supraoccipital foramen by small pro- cessus lamellosus, with entry to basioccipital foramen constricted by small processus basalis; basioccipital partially ossified; crista basioccipitalis present; slight crista obliqua on quadrate ramus of pterygoid; large paraquadrate foramen situated in quadratojugal/jugal suture. Mastodonsaurus lavisi (new comb.) (Seeley 1876) Holotype. R 4215. Posterior part of a right lower jaw. Paratypes. Gz 20, Gz 9, Gz 1057, Gz 15, Gz 37, R 331, SM 369. Cranial material plus lower jaws. Locality of holotype. Cliffs of Picket Rock Cove near Sidmouth (La vis 1876). Localities of paratypes. Above plus quarries in Warwickshire. Horizon. ‘Lower Keuper’ Sandstone. Top of the Building Stones Formation in Warwickshire. Ten feet below the top of the Otter Sandstone and the base of the Keuper Marl in Devon. Diagnosis. A mastodonsaur with skull deep posteriorly; orbits relatively large (A:OL= 121), close together, oval with long axis antero-posterior; frontal forms large part of orbital border; pineal foramen oval with long axis antero-posterior, far behind posterior border of orbits; postorbital relatively small, unexpanded anteriorly; otic notch open with supratemporal excluded; posterior border of skull only slightly concave; lateral line grooves shallow and wide; ornament coarse and shallow. Lower jaw massive and well ossified; retroarticular process with high lateral edge shelving smoothly down medially; coronoid toothless with radiating ridged orna- ment; triangular area of ornamentation on external surface of retroarticular process. PATON: CAPITOSAUROID LAB YRINTHODONTS 283 PHYLOGENY OF THE CAPITOSAUROIDEA Romer (1966) included Mastodonsaurus in the Capitosauridae as did Bonaparte (1963) in his description of Promastodonsaurus bellmani. Bonaparte identified this species as a mastodonsaur because of; (i) the relative extension of the pterygoid/ parasphenoid suture; (ii) the relative broadening of the pterygoid/parasphenoid suture; (hi) the position of the mandibular articulation with respect to the condyles. The pterygoid/parasphenoid suture is not figured so (i) and (ii) cannot be checked. However, with regard to point (iii), Bonaparte states that Mastodonsaurus is the only genus of capitosaur with the condyles posterior to the mandibular articulation, but this is incorrect. Those of Parotosaurus peabodyi are posterior to the mandibular articulation as are those of P. birdi, Cyclotosaurus hemprichi, C. ebrachensis, and C. posthumus. The palate shows an unpaired anterior palatal vacuity and, anteriorly, swollen interpterygoid vacuities. The tabular is not the typical mastodonsaur shape but is similar to that of parotosaurs with semi-closed otic notches, being slightly expanded distally; and the orbits are in the posterior half of the skull. These points suggest that Promastodonsaurus bellmani is probably a parotosaur with a semi- closed otic notch and fairly deep skull. Unfortunately the orbits and ornament are not preserved. Most authors remove Mastodonsaurus from the Capitosauridae and place it in a family of its own of equal status to the Capitosauridae and Bentho- suchidae. It seems probable that the Mastodonsauridae were derived from primitive benthosuchids by the retention of some juvenile characters. Welles and CosgriflF (1965) include only three genera in the Capitosauridae: Parotosaurus, Cyclotosaurus, and Paracyclotosaurus. Romer (1947) suggested that "Capitosaurus" semiclausus (Swinton 1927) should be placed in a new genus because of the exclusion of the jugal from the orbit by an anterior extension of the post- orbital. Heyler (1969) has described a similar form, Stenotosaurus lehmani, with a completely closed otic notch. Welles and Cosgriff (1965) suggest that ''Parotosaurus' semiclausus is a good ancestor for Paracyclotosaurus davidi. Howie (1970) emphasizes their differences, however, and suggests that Parotosaurus pronus is a more suitable ancestor. In fact, Stenotosaurus semiclausus appears to be a young individual— the zones of intensive growth are small and the sutures are very open. The distance between the edges of the tabular and squamosal is about the same as the distance between the edges of the postparietal/supratemporal suture in the middle of the skull roof so it is probable that the otic notch was effectively closed. Both Stenoto- saurus semiclausus and S. lehmani lack the lachrymal flexure in the infraorbital lateral line groove. The ornamentation of S. semiclausus is also very different from that of cyclotosaurs and parotosaurs, consisting of very high, narrow ridges and deep, wide pits. The lateral lines are no wider than these pits. Unfortunately the ornamenta- tion is not described in S. lehmani. The distinctive sculpture, paired (or nearly so) anterior palatal vacuities, absence of a lachrymal flexure, and the exclusion of the jugal from the orbit seem sufficient grounds for retaining the genus Stenotosaurus and for placing it in a separate family, the Stenotosauridae, as Heyler (1969) has done. This family was probably derived from a form similar to Kestrosaurus dreyeri (included here in the family Stenotosauridae) in turn derived from a benthosuchid like Volgasaurus. Swinton (1927) noted the similarity of "Capitosaurus' semiclausus F 284 PALAEONTOLOGY, VOLUME 17 TEXT-FIG. 19. Family tree of the superfamily Capitosauroidea. to Kestrosaurus. Welles and Cosgriff (1965) suggested that Kestrosauriis has tremato- saur relationships and Heyler (1969) suggests a trematosaur relationship for Slenoto- sauriis. This resemblance seems superficial, however, and the slender snout, long vomerine plate, and paired anterior palatal vacuities indicated by these authors to show the similarity to trematosaurs can also be found in the benthosuchid Volgasaurus which seems a more reasonable ancestor. Text-fig. 19 shows a suggested family tree for the Capitosauroidea. PATON: CAPITOSAUROID LABYRINTHODONTS 285 STRATIGRAPHY Warrington (1970) has introduced formational names for the English ‘Keuper’ or ‘Lower Keuper’ Sand- stone. He recognizes a ‘Keuper’ Sandstone Group comprising the Conglomerate Formation and the Building Stones Formation. Succeeding this, the Waterstones Formation (sometimes confusingly included in the ‘Keuper’ Sandstone by earlier workers) is regarded as the basal unit of the ‘Keuper’ Marl Group. The Warwick and Bromsgrove specimens all come from approximately the same horizon (see Wills 1910, 1970; Walker 1969; Warrington 1970) in the upper part of the Building Stones Formation. R 3174 from Stanton is thought to come from near this horizon as well. The holotype of Mastodonsaums lavisi comes from the Otter Sandstone (Warrington 1971 —formerly known as the Upper Sandstone) of South Devon. Lavis (1876) said that it was collected from a level 10 feet below the base of the ‘Keuper’ Marl. Rhynchosaur remains belonging to a more primitive species of Rliyncliosaurus than that found in the Midlands have also been collected from this horizon and from the base of the Otter Sandstone (Huxley 1869; Metcalfe 1884; Walker, pers. comm.). Walker (1969, 1970) therefore suggests that the South Devon horizon is older than that of the Midlands. The great similarity between the lower jaws Gz 15 and R 4215 seems, however, to contradict this, indicating that the two horizons are of closely similar age. Cyclotosaurus is typically an Upper Triassic genus and most reported specimens come from the German Middle and Upper Keuper. Earlier specimens have been found : Cyclotosaurus randalli and " Rhadaloguathus' boweni are from the Holbrook Member of the Moenkopi Formation of Northern Arizona and Welles and Estes (1969) assign a lower Middle Triassic age to this assemblage which also contains an ilium of Arizonasaurus. This resembles a poposaurid ilium from Warwick (Walker 1969), the Warwick specimen being of a more advanced type, although less advanced than similar ilia from the Upper Tiias. This sug- gests that the English horizon is younger than the Holbrook Member. Ortlam (1970) has described a new genus and species, Eocyclotosaurus wosclmudti, from the Upper Bunter of South Germany. This species shows some similarities to Cyclotosaurus leptogiiatluis, e.g. general shape; shape and position of the pineal foramen; but the frontals do not project forwards in C. leptoguatlius, the nares are further forward, the orbits are larger, and the skull is broader generally in the latter species. Mastodonsaurus jaegeri is from the late Middle Trias (Lettenkohle) and the more primitive, presumably ancestral M. cappelensis is from the Upper Bunter of Germany. Krebs (1969) has recently argued that the upper part of the German Middle Bunter, and the German Upper Bunter are Anisian in age. He describes Cteiiosauriscus koeneni from the upper Middle Buntsandstein of Germany as a pseudosuchian and compares it with Hypselorachis mirabilis from the Manda Formation of Tanzania. The latter formation is probably Anisian in age and Krebs uses the close resemblance between the two species as evidence that the German strata are also Anisian. However, the Mands Formation itself has to be dated by roundabout methods so that Krebs’s suggestions cannot carry as much weight as they would otherwise do. The stratigraphical nomenclature of the British Trias is considerably confused as the marine Muschel- kalk which forms the greater part of the German Middle Trias is virtually absent in Britain. As there is no major unconformity in the middle of the British Trias, the time equivalent of the Muschelkalk must be contained in the British Bunter and Keuper, i.e. these cannot be directly equated with the German Bunter (Lower Trias) and Keuper (mostly Upper Trias). The Middle Trias is usually divided into two stages; the Anisian and the Ladinian. The limits of these Alpine stages are not precisely known within the German sequence. Most European authorities, however, consider that the Anisian/Ladinian boundary lies within the Middle Muschelkalk, or at its top (Rieber 1967). Thus the Anisian is thought to include the Lower, and part or all of the Middle Muschelkalk, and the Ladinian is believed to include the Upper Muschelkalk and the Lettenkohle. . Walker (1969) reviewed the reptile fauna of the ‘Lower Keuper’ Sandstone which comes from the same horizon at the top of the Building Stones Formation as the labyrinthodonts, and concludes that the most probable age for these beds is Early to Middle Ladinian (see Table 3) although Geiger and Hopping (1968) and more recent work by Warrington (1970) based on spore studies suggest that it is Middle Anisian or possibly even earlier. This latter result does seem very early when applied to the vertebrate evidence. The labyrinthodonts consist of a primitive mastodonsaur, a comparable German species being late Lower Trias, or possibly Anisian in age; a primitive and aberrant cyclotosaur which slightly resembles a German species from the late Lower Trias; and a more typical advanced cyclotosaur of a sort found only in Upper Triassic rocks, and resembling Cyclotosaurus posthumus from the Stubensandstein. Thus the labyrintho- dont evidence is rather anomalous, indicating either a late Lower Triassic or an Upper Triassic age. In 286 PALAEONTOLOGY, VOLUME 17 TABLE 3. Stratigraphical nomenclature of the Trias. TRADITIONAL BRITISH NOMENCLATURE (Hull. 1869) LITHOSTRATI GRAPHICAL NOMENCLATURE Central Midlands of England Germany ALPINE STAGES Keuper Marl Waterstones Building Stones C onglomerate Bunter Upper Mottled Sandstone Parva Formation Trent Formation* Edwalton Formation"* Harlequin Formation" Carlton Formation' Radcliffe Formation^ Waterstones Formation Building Stones Formation Conglomerate Formation Keuper Muschelkalk Bunter ^ Hardegsen discontormity Upper Mottled Sandstone Formation after Warrington, 1970- Ladinian Anisian Scythian Formations defined by Elliott. 1961. this case, greater weight is placed upon the occurrence of C. pachygnathus which is of a type limited so far to Upper Triassic rocks, than on the occurrence of Mastodonsaurus lavisi and Cyclotosaurus leptognathus whose close relatives have fairly long ranges. Primitive species tend to have much longer stratigraphic ranges than specialized species in the same group. Thus the earliest possible age for such a fauna seems to be late Anisian, and, taking into account the reptilian evidence, it seems much more likely that the upper part of the Building Stones Formation in Warwickshire and the Otter Sandstone in Devon is Early Ladinian (Middle Trias, Upper Muschelkalk) in age. CONCLUSIONS The remains of English Triassic capitosaurs are, but for the skull of Cyclotosaurus ^ stantonensis', very fragmentary. However, by comparison with closely allied forms with complete skulls, the remains can be placed in three groups: a primitive masto- donsaur; a primitive and aberrant cyclotosaur; and a more typical cyclotosaur. Much of the material is, however, undiagnostic and cannot be referred to any of the three species. Evidence from this and other sources shows that otic notch closure and development of the otic fenestra occurred independently many times as sug- gested by Welles and Cosgriff (1965). C. leptognathus shows an unusual method of otic notch closure as the tabular is unexpanded distally. Closure of the otic notch presumably resulted in the freeing of a larger area of the occiput leading to increased efficiency in raising the flattened skull and lowering the jaw. It seems very likely that PATON: CAPITOSAUROID LABYRINTHODONTS 287 some forms may have had open otic notches when young which closed with increas- ing age. Much of the material consists of lower jaws, very few of which are described in the literature on capitosaurs. Welles and Cosgriff (1965) state that lower jaws on their own are insufficiently diagnostic to allow identification of a species. Possibly this may be the case in closely related species but the jaws studied here fall into three different and distinct categories. All show similarities in general structure as would be expected in members of the same superfamily, but the mastodonsaur jaws differ greatly from those of the two cyclotosaur species, which in turn show substantial differences from each other. The discovery and description of more lower jaws of different species may therefore show that they are more distinctive than was pre- viously thought. Welles and Cosgriff (1965) also concluded that the only valid English capitosaur was C. ' stantonensis' . It is hoped that the evidence given here satisfactorily confirms the presence of three valid species of capitosauroid in the English Trias. The stratigraphical position of the horizon in which the labyrinthodonts are found can best be interpreted as Early Ladinian in age. This fits in well with the reptilian evidence, but spore studies at present indicate an earlier position for this horizon. Acknowledgements. I am very grateful to Dr. A. D. Walker and Professor T. S. Westoll for their constant help and advice throughout the period of this study. The work has been submitted to the University of Newcastle upon Tyne for the degree of Ph.D. I also wish to thank Mr. B. Playle, Dr. P. Cattermole, Dr. W. Allen, Miss J. Morris, and the Trustees of Warwick County Museum for allowing me to borrow and study their labyrinthodont collection. I am indebted to Dr. A. J. Charig, Dr. H. W. Ball, and the Trustees of the British Museum (Natural History) for permission to borrow, prepare, and study Cyclotosaums "stantonensis' . Financial support from Newcastle University for visits to museums in England and Germany is grate- fully acknowledged. REFERENCES AMALiTSKY, V. p. 1924. On the Dvinosauridae, a family of labyrinthodonts from the Permian of Russia. Ann. Mag. Nat. Hist. (9), 13, 50-64. BONAPARTE, J. F. 1963. Promastodonsawus hellmani, a capitosaurid from the Middle Trias of Argentina. Ameghiniana, 3, 67-78, 1 pi. [In Spanish.] BYSTROW, A. p. 1935. Morphologische Untersuchungen der Deckknochen des Schadels der Wirbeltiere. I Mitteilung. Schadel der Stegocephalen. Acta Zool. Stockh. 16, 65-141. and EFREMOV, j. a. 1940. Benthosuchus sushkini Efr., a labyrinthodont from the Eotriassic of the Sharzhenga River. Trudy paleozool. Inst. 10, 1-152. [Russian, English summary.] CASE, E. c. 1932. A collection of stegocephalians from Scurry County, Texas. Contr. Mas. Paleont. Univ. Mich. 4(1), 1-56, 7 pis. CHOWDHURY, T. R. 1970. A new capitosaurid from the Triassic Yerrapalli Formation of Central India. Journ. Geol. Soc. Ind. 11, 155-162. FRAAS, E. 1889. Die Labyrinthodonten der schwabischen Trias. Palaeontographica, 36, 1-158, 17 pis. 1913. Neue Labyrinthodonten aus der schwabischen Trias. Ibid. 60, 275-294, 7 pis. GEIGER, M. E. and HOPPING, c. A. 1968. Triassic stratigraphy of the Southern North Sea Basin. Phil. Trans. B, 254, 1-36. HEYLER, D. 1969. Un nouveau stegocephale du Trias inferieur des Vosges, Stenotosaurus lehmani. Ann. Paleont., Vertehres, 55, 73-80, 2 pis. 288 PALAEONTOLOGY, VOLUME 17 HOWIE, A. A. 1970. A new capitosaurid labyrinthodont from East Africa. Palaeontology, 13, 210-253, 1 pi. HUENE, F. VON. 1922. Beitrage zur Kenntnis der Organisation einiger Stegocephalen der schwabischen Trias. Acta Zool. Stockh. 3, 395-460, 2 pis. 1932. Ein neueartiger Stegocephalen-Fund aus der oberhessischen Buntsandstein. Palaont. Z. 14, 200-229, 2 pis. HU.XLEY, T. H. 1859. On a fragment of a lower jaw of a large labyrinthodont from Cubbington. In howell, H. H. 1899, Sclent. Mem. Huxley, 2, 269-270. 1869. On Hyperodapedon. Q. Jl. Geol. Soc. Land. 25, 138-152. JAEGER, G. F. 1828. Ueber die fossile Reptilien, welche in Wiirttemberg aufgefunden warden sind. Stuttgart. KONZHUKOVA, E. p. 1965. New parotosaurs from the Trias of the Tower Urals. Paleont. Zh. 1, 97-104. [In Russian.] KREBS, B. 1969. Ctenosaiiriscus koeneni (von Huene), die Pseudosuchia und die Buntsandstein Reptilien. Eclog. Geol. Helv. 62, 697-714, 2 pis. LAVis, H. J. j. 1876. On the Triassic strata exposed in the cliff sections near Sidmouth, and a note on the occurrence of an ossiferous zone containing bones of a Labvrinthodon. Q. Jl. Geol. Soc. Land. 32, 274-277. METCALFE, A. T. 1884. On further discoveries of vertebrate remains in the Triassic strata of the south coast of Devon between Budleigh Salterton and Sidmouth. Ibid. 40, 257-262. MEYER, H. and PLIENINGER, T. 1844. Beitrage zur Paldontologie Wiirttemberg' s, enthaltend die fossilen Wirbelthierreste aus den Triasgebilden mit besonderer Riicksicht auf die Labyrinthodonten des Keupers. Stuttgart. MiALL, L. c. 1874. On the remains of Labyrinthodonta from the Keuper Sandstone of Warwick. Q. Jl. Geol. Soc. Land. 30, 417-435, 3 pis. MUNSTER, G. VON. 1836. Letter on various fossils. Neues Jb. Min. Geog. Geol. 580-583. NILSSON, T. 1937. Ein Plagiosauride aus dem Rhat Schonens. Beitrage zur Kenntnis der Organisation der Stegocephalengruppe Brachyopoidei. Acta Univ. Lund. 34, 1-75. 1943. On the morphology of the lower jaw of Stegocephalia, with special reference to Eotriassic stegocephalians from Spitsbergen. Part I. K. svenska Vetensk. Akad. Handl. 20 (9), 1-46, 9 pis. 1944. Part II of above. Ibid. 21 (1), 1-70. ORTLAM, D. 1970. Ein neuer Capitosauride aus dem Oberen Buntsandstein des nordlicher Schwarzwaldes. Neues Jb. Geol. Palaont. Monatsh. 9, 558-568. OWEN, R. 1841. On the teeth of species of the genus Labyrinthodon (Mastodonsaurus) salamandroides and Phvtosaurus (?) of Jaeger from the Sandstone of Warwick and Leamington. Ann. Mag. Nat. Hist. 8, 58-61 . 1842. A description of the skeleton and teeth of five species of Labyrinthodon with remarks on the probable identity of the Chirotherium with this genus of extinct Batrachians. Trans. Geol. Soc. Land. 6, 515-543. PANCHEN, A. L. 1959. A new armoured amphibian from the Upper Permian of East Africa. Phil. Trans. B, 242, 207-281, 1 pi. RiEBER, H. 1967. liber die Grenze Anis-Ladin in den Siidalpen. Eclog. Geol. Helv. 60, 61 1-614. ROMER, A. s. 1947. Review of the Labyrinthodontia. Bull. Mus. Comp. Zool. Harv. 99, 1-352. 1966. Vertebrate Paleontology, 3rd edn. Chicago. SEELEY, H. G. 1876. On the posterior portion of a lower jaw of a Labyrinthodon (L. lavi.d) from the Trias of Sidmouth. Q. Jl. Geol. Soc. Land. 32, 278-284, 1 pi. swiNTON, w. E. 1927. A new species of Capitosaurus from the Trias of the Black Forest. Ann. Mag. Nat. Hist. (9), 20, 178-186, 1 pi. WALKER, A. D. 1969. The reptile fauna of the 'Lower Keuper' Sandstone. Geol. Mag. 106, 470-476. 1970. Written contribution to discussion. Q. Jl. Geol. Soc. Land. 126, 217-218. WARD, J. 1900. On the occurrence of labyrinthodont remains in the Keuper Sandstone of Stanton. Trans. N. Staffs. Lield Club, 34, 108-1 12, 2 pis. WARRINGTON, G. 1970. The stratigraphy and palaeontology of the 'Keuper' series. Q. Jl. Geol. Soc. Land. 126, 183-223, 6 pis. 1971. Palynology of the New Red Sandstone sequence of the South Devon coast. Proc. Ussher Soc. 2, 307-314. PATON: CAPITOSAUROID LABYRINTHODONTS 289 WATSON, D. M. s. 1919. The structure, evolution and origin of the Amphibia; the ‘orders’ Rhachitomi and Stereospondyli. Phil. Trans. B, 209, 1-73, 2 pis. 1926. The evolution and origin of the Amphibia. Ibid. 214, 189-257. 1951. Palaeontology and Modern Biology. New Haven, Yale University Press. 1958. A new labyrinthodont (Paraeyclotosaurus) from the Upper Trias of New South Wales. Bull. Br. Mus. Nat. Hist. 3, 233-264, 5 pis. 1962. The evolution of the labyrinthodonts. Phil. Trans. B, 245, 219-265. WELLES, s. P. 1947. Vertebrates from the Upper Moenkopi Formation of Northern Arizona. Univ. Calif. Pubis. Geol. Sci. 27, 241-289, 2 pis. and COSGRIFE, J. 1965. A revision of the labyrinthodont family Capitosauridae. Ibid. 54, 1-148, 1 pi. and ESTES, R. 1969. Hadrokkosaurus hradvi from the Upper Moenkopi Formation of Arizona. Ibid. 84, 1-56, 2 pis. WEPFER, E. 1923. Die Buntsandstein des badischen Schwarzwalds und seine Labyrinthodonten. Mon. Geol. Paleont. 5, 1-101. WILLS, L. J. 1907. Note on fossils from the Lower Keuper of Bromsgrove. Geol. Mag. 4, 28. 1910. The fossiliferous Lower Keuper rocks of Worcestershire. Proc. Geol. Assoc. 21, 249-331. 1915. The structure of the lower jaw of Triassic labyrinthodonts. Proc. Birm. Nat. Hist. Soc. 14, 1-16, 2 pis. 1970. The Triassic succession in the Central Midlands. Q. Jl. Geol. Soc. Lond. 126, 225-285. WOODWARD, A. s. 1904. On two new labyrinthodont skulls of the genera Capitosaurus and Aphaneramma. Proc. Zool. Soc. Lond. 2, 170-176, 2 pis. ziTTEL, K. A. VON. 1911. Grimdzuge der Paldontologie (Paldozoologie). H—Vertebrata. (2nd edn. revised by Broili, Koken, and Schlosser) Munich and Berlin. ROBERTA L. PATON Department of Geology Royal Scottish Museum Chambers Street Edinburgh, EHl IJF Revised typescript received 15 May 1973 A NEW GIANT PENGUIN FROM THE EOCENE OF AUSTRALIA by R. J. F. JENKINS Abstract. Pachydyptes simpsoni sp. nov., a 'giant’ fossil penguin, is described from fragmentary remains found at two levels in the Upper Eocene, Blanche Point Marls, South Australia, The remains from the lowest level imme- diately predate the first appearance of the index coccolith Isthmolithus recurvus Deflandre and are amongst the oldest well-dated penguin bones known. Comparison with the largest modern penguin Aptenodytes forsteri Grey suggests that the fossil species probably stood about 140 cm tall. Its coracoid was relatively short compared to this modern species, one cervical vertebra relatively longer, and the wing length similar proportionally. Penguins are considered to have evolved from flying birds (Simpson 1946, 1971u) and the short coracoid is interpreted as supporting this. The inferred relatively long neck may have been inherited from the volant ancestor of the penguins or represent an early specialization. The length of the wing reflects its function as a propulsive organ in water and reinforces the suggestion by Simpson that 'all the basic locomotory adaptations of the Spheniscidae were virtually complete in the late Eocene'. The need for precise stratigraphic data in analysing the palaeoecology of fossil penguins is emphasized, and so far as is possible, P. simpsoni and other Eocene fossil penguins are placed in a context of biostratigraphic, bio- geographic, and palaeoclimatic data. This suggests that the Eocene penguins from Australasia lived in seas with surface water temperatures of about 12-16 °C. They are interpreted as recording northward penetration of Antarctic populations, perhaps to a palaeolatitude of 45° S. In May 1968 several fossil penguin bones were discovered by Messrs. Bret Robinson and Harry Eames in the Blanche Point Marls (Reynolds 1953) at Blanche Point, 37 km SSW. of Adelaide, South Australia (text-fig. 1). The Blanche Point Marls are of Upper Eocene age (Lindsay 1969; Ludbrook 1969). Further bones were col- lected by Dr. M. R. Walter, Bureau of Mineral Resources, Canberra, and myself at a subsequent excavation made at the point of initial discovery. Two other bones collected at Blanche Point (apparently in association) in October 1932 by Mr. L. W. Parkin, now Director of the Australian Mineral Foundation, Frewville, Adelaide, were kindly made available by Emeritus Professor M. F. Glaessner, of the University of Adelaide. These bones also proved to be from a fossil penguin and are here considered conspecific with those discovered in 1968. Another small bone fragment was collected by a student during a teaching excur- sion to Blanche Point in June 1971 and is here tentatively referred to the same species. Twenty-six fossil penguin bones, all found isolated, have been documented or described previously from the Tertiary of Australia (Glaessner 1955; Simpson 1957, 1959, 1965, 1970). These bones are from the south-eastern part of the continent (mainland). The first fossil penguin remains identified from Australia, a humerus from Witton Bluff, 11 km north of Blanche Point (text-fig. 1), was reported by Finlayson (1938) who considered that the bone most closely resembled a humerus of Palaeeudyptes Huxley, 1859, from the New Zealand Tertiary. Glaessner (1955) recorded his own dis- covery of a fossil penguin tibiotarsus immediately south of Witton Bluff in the upper part of the Blanche Point Banded Marls Member (Reynolds 1953) of the Blanche Point Marls, and indicated that Finlayson’s bone was from the Transitional Marl [Palaeontology, Vol. 17, Part 2, 1974, pp. 291-310, pis. 37-.39.] 292 PALAEONTOLOGY, VOLUME 17 TEXT-FIG. 1. Locality maps and stratigraphic column measured from the Eocene section exposed in the sea-clilfs at Blanche Point. The known ranges of key planktonic index fossils in this section and the strati- graphic levels of the fossil penguin remains described in this paper are indicated. The Tortachilla Lime- stone varies in thickness and it and the Blanche Point Marls are virtually conformable. Member (Reynolds, op. cit.) of the Blanche Point Marls. These bones were studied by Simpson (1957) who identified the humerus as Palaeeudytes cf. antarcticm Huxley, 1859 and the tibiotarsus as P. cf. antarcticusl Both bones are presently referred by Simpson (1971a) to Palaeeudyptes sp. indet. Glaessner (1955) also recorded the occurrence of two Oligocene penguin bones, a damaged humerus and an incomplete small femur, from the Gambler Limestone near Mount Gambler, South Australia. Simpson (1957) described the humerus as Gen. et sp. indet. A and the femur as Gen. et sp. indet. B. Planktonic foraminiferal studies made on samples from the same or the probable locality of these bones suggest a dating within the Globigerina labiacrassata zone of Ludbrook and Lind- say (1969), in terms of local zonation, or approximately equivalent to zone P. 19/20 of Blow (1969, 1970), with respect to tropical biostratigraphy (Dr. B. McGowran, pers. comm.). Coccolith assemblages from the same samples, kindly examined by Mr. S. Shafik, were suggestive of a late Middle Oligocene age (late Rupelian). A new genus and species, Anthropodytes gilli, was erected by Simpson (1959) for a humerus which Gill (1959) reported from Miocene deposits exposed by the Glenelg River in western Victoria. Investigations which I have made on the planktonic fora- minifera in samples from the locality of this bone indicate a dating within the Globi- gerinoides trilobiis trilobus zone of Ludbrook and Lindsay (1969), or Lower Miocene. Simpson (1965, 1970) documented the occurrence of twenty fossil penguin bones JENKINS: GIANT EOCENE PENGUIN 293 from Beaumaris, Victoria, and one penguin bone from Spring Creek, near Minha- mite, Victoria. An approximately late Miocene age was ascribed to these materials (Simpson 1970). The approximately late Miocene deposit with which the penguin bones from Beaumaris are associated (the Black Rock Formation) rests discon- formably on sediments of Lower to Middle Miocene age (Gill 1957; Kenley 1967). The penguin remains are apparently from the bed of remanie phosphatic nodules which occurs at the base of the Black Rock Formation, and it seems probable that at least some of the bones may have been derived from the underlying pre-late Miocene rocks. The bone from near Minhamite, an incomplete humerus, was made the holotype of a new genus and species, Pseudaptenodytes macrai Simpson, 1970. The bones from Beaumaris were identified as P. macrai, IP. minor Simpson, 1970, and Spheniscidae gen. et sp. indet. None of the remains discussed above resemble the bones described in the present work. The materials dealt with here are catalogued in the palaeontological collec- tion of the South Australian Museum, Adelaide, and the registration numbers given in the text pertain to this collection. SYSTEMATIC PALAEONTOLOGY Order sphenisciformes Sharpe, 1891 Family spheniscidae Bonaparte, 1831 Genus pachydyptes Oliver, 1930 Type species. Pachydyptes ponderosus Oliver, 1930, by original designation. 1930 Pachydyptes Oliver, p. 85. 1946 Pachydyptes Oliver; Simpson, p. 41 (synonomy). 1952 Pachydyptes Oliver; Marples, p. 35. 1971a Pachydyptes Oliver; Simpson, p. 365 (revised diagnosis). Remarks. Three genera of closely similar, very large or ‘giant’ fossil penguins are currently recognized from the Early Tertiary, Palaeeudyptes Huxley, 1859, Anthro- pornis Wiman, 1905, and Pachydyptes Oliver, 1930. Bones referred to Palaeeudyptes are known from the Upper Eocene and Oligocene of the South Island of New Zealand, the Upper Eocene and (late Middle) Oligocene of southern, mainland Australia (Simpson 1971u), and from Seymour Island, off the Antarctic Peninsula, in sedi- ments dated tentatively on the basis of cetacean remains and their penguin fauna as being possibly late Eocene (Simpson 19716). Anthropornis is identified only from these same deposits on Seymour Island (Simpson 19716). Prior to the present study Pachydyptes was considered to include only the type species P. ponderosus from the Totara Limestone (Gage 1957) at Oamaru, Otago Province, southern New Zealand (Simpson 197 lu). The Totara Limestone (in the usage of Gage 1957 and Simpson 1971a) is dated by recent planktonic foraminifera and coccolith studies as extending from possibly the latest Kaiatan, through the Runangan and into the Whaingaroan, in terms of New Zealand marine stages, or Upper Eocene to Lower Oligocene (Jenkins 1971; Edwards 1971). Simpson (1971a, 19716) drew attention to the problem of the validity of the dis- tinctions made between the three above-mentioned genera. Simpson (1971a, foot- note p. 347) points out the close resemblance between Paehydyptes and Anthropornis 294 PALAEONTOLOGY, VOLUME 17 and suggests that the two are only tentatively distinct. Nevertheless, Simpson {\91\b) continues to distinguish all three genera but comments that ‘had the separa- tion not already been made, . . . [he] would hesitate to separate the genera now’. He indicates that their type-species are quite distinct and concludes ‘it appears that Pachydyptes, Anthropornis, and Palaeeudyptes are related genera and form a close morphological sequence in that order’; this order was ‘not phylogenetic as Pachy- dyptes and Palaeeudyptes are at least partly contemporaneous and Anthropornis may well be also’. The bones described in the present work most closely resemble their counterparts in Pachydyptes ponderosus. However, they are less robust than the bones of P. ponderosus and in this and several other respects approach bones referred to Anthro- pornis and Palaeeudyptes, thus lending further emphasis to the intergradational series evident between the early ‘giant’ penguins. Nevertheless, the bones of P. ponderosus and the new fossil species seem as similar as in any other two fossil penguin species assigned to a single genus, and in features such as the extremely large head of the humerus, the great width of the shaft of this bone, and the lengthened or expanded pectoralis secundus and pectoralis tertius muscle insertions, appear dis- tinct from other known fossil penguins. It thus seems justifiable to maintain the taxon Pachydyptes and to include the present species with it. Pachydyptes simpsoni sp. nov. Plates 37-39; text-fig. 2a Named for Professor George Gaylord Simpson, a most eminent palaeontologist and leading student of penguin evolution. Material. The holotype, P. 14157a-g, was discovered by Messrs. Robinson and Eames and consists of the following bones: {a) greater portion of left coracoid; {b) eroded head of right humerus; (c) broken head of left humerus; (d) right radius with the extremities damaged; (c) left carpometacarpus with the first metacarpal broken away; (/) left proximal phalange of second digit; (g) a damaged and distorted vertebra, possibly the twelfth. All of the bones occurred in a single 20-cm-thick stratum within an area of about half a square metre. Hence the assemblage almost certainly represents one individual. The paratype material, P. 14158a, b, collected by Mr. Parkin, consists of two bones: {a) badly broken proximal two-thirds of the right humerus; (h) proximal end of right radius. The specimen P.17913 collected in 1971 is a small fragment of a bone shaft and is interpreted as being broken from the proximal part of a rib. Preservation. The bones consist either of the natural skeletal material or are partly phosphatized. The matrix of the holotype, P.14157, consists of irregularly silicified, but otherwise poorly indurated marl. The matrix of P.14158 is a soft, glauconitic marl, a pale grey-white sediment including abundant green glauconite grains. EXPLANATION OF PLATE 37 Pachydyptes simpsoni sp. nov. Portions of bones which have been restored with plaster are outlined. Dashed lines (background) show a hypothetical restoration of the original outline of the bones. Impor- tant muscle attachments are outlined with dashed white lines. All figures natural size. 1, dorsal view of left coracoid of holotype, P. 141 57a. 2, dorsal view of articular surface of incomplete head of right humerus of holotype, P.141 57b. 3, dorsal view of incomplete proximal two-thirds of right humerus of paratype, P. 141 58a: ps, insertion of pectoralis .secundus; pt, insertion of pectoralis tertius; Id, insertion of latissimus dorsi. PLATE 37 JENKINS, Eocene penguin 296 PALAEONTOLOGY, VOLUME 17 Locality. The holotype is from the extreme tip of Blanche Point opposite Gull Rock. The fragment P.17913 is from about 20 m north of the foundations of 'Uncle Tom’s Cabin’ (locality indicated in map which Reynolds (1953) gives of the Maslin, Aldinga coast) 10 km ENE. of the tip of Blanche Point. Paratype P. 1 4 1 58 is from the vicinity of Blanche Point, but no additional information is given on its label. Stratigraphic Position. The holotype was collected 3-6 m below the top of the Blanche Point Banded Marls Member of the Blanche Point Marls. Specimen P.17913 was found loose on the Tortachilla Limestone (Reynolds 1953) which immediately underlies the Blanche Point Marls. The preservation of the specimen and study of the locality suggests that it weathered out of the Transitional Marl Member of the Blanche Point Marls, and is possibly from the glauconite-rich beds comprising most of the lower part of this member (text-fig. 1 ). Paratype P.14158 is also from the Blanche Point Marls, but its precise stratigraphic level is not recorded. Its richly glauconitic matrix is consistent with its having been collected from the lower part of the Transi- tional Marl Member. The matrix was kindly examined by Mr. S. Shafik and found to contain a sparse coccolith assemblage; the index coccolith Isthmolithus recurvus Deflandre was absent. Research by Mr. Shafik and myself shows that I. recurvus (or a closely allied form) first appears as an exceedingly rare ele- ment 0-3 m above the base of the Transitional Marl Member, and can be confidently identified as a rather frequent species 0-6 m above the base of the member; coccoliths in general are relatively rare in the lower 0-6 m of the member, and common or even abundant above. The combined sedimentological and cocco- lith evidence from the matrix of paratype P.14158 suggests that it is from the basal 0-45 m thick soft glau- conitic marl bed of the Transitional Marl Member. Age. The Blanche Point Marls comprise part of the Aldingan Stage in Australia (Ludbrook and Lindsay 1966). Glaessner (1951), Reynolds (1953), and Wade (1964) indicate that the planktonic foraminifera species Hantkenina ' alabamensis compressa Parr’ ( = H. primitiva Cushman and Jarvis according to Blow 1969) occurs in the Transitional Marl Member of the Blanche Point Marls. This record has been verified by obtaining rare specimens of the species from washings of sediment taken at 10 and 1-5 m above the base of the Transitional Marl Member at 'Uncle Tom’s Cabin’ near Blanche Point. As indicated above the fossil penguin P.14158 is apparently from the lowermost part of this member, below the first appear- ance of the coccolith I. recurvus. According to McGowran (1971) the H. primitiva interval is high in zone P. 1 5 of Blow ( 1 969) ; Gartner (1971) indicates the initial appearance of I. recurvus is also within zone P.15, that is early Upper Eocene in age. The age of fragment P.17813 is probably similar. Hornibrook and Edwards (1971) indicate that in New Zealand the time ranges of '//. alabamensis' (almost certainly = H. primitiva) and I. recurvus overlap in the late Kaiatan and early Runangan, approximately middle Upper Eocene, and that I. recurvus first appears in the late Kaiatan. Planktonic foraminiferal studies date the top part of the Blanche Point Banded Marls Member as also being Upper Eocene (Wade 1964; Lindsay 1969; Ludbrook and Lindsay 1969). This part of the section is stratigraphically higher than the local occurrence of H. primitiva and is shown by Lindsay (1969) to include Globigerapsis index (Finlay), but not Globigerina ampliapertura Bolli. This data suggests a tentative correlation with the upper-middle part of the Globigerina linaperta Zone in the New Zealand biostratigraphic scheme of Jenkins (1971), or with respect to the composite foraminiferal-coccolith bio- stratigraphic scheme of Hornibrook and Edwards (1971), an early Runangan age. The present penguin bones are similar in age to those from the Blanche Point Marls at Port Willunga and several partial skeletons from the Kaiatan or early Runangan at Dunedin, southern New Zealand; collectively these bones are the oldest well-dated remains of fossil penguins in the world. Description Coracoid of holotype, P.14 157a. Notable features of this bone are its broadly flared EXPLANATION OF PLATE 38 Pacliydyptes simpsoni sp. nov. Portions of bones which have been restored with plaster are outlined. All figures natural size. \a, b. The ?twelfth cervical vertebra of the holotype, P.14157, lu, right lateral view. 16, ventral view. 2, ventral view of incomplete proximal two-thirds of right humerus of paratype, P.141 58a. 3, ventral view of left coracoid of holotype, P.141 57a. PLATE 38 JENKINS, Eocene penguin 298 PALAEONTOLOGY, VOLUME 17 base and almond-shaped precoracoid foramen. The proportions of its shaft are comparable with those of the slightly larger coracoid of Pachydyptes ponderosus. However, the curvature of the proximal portion of the preaxial margin differs; the basal part of the preaxial margin lies at an angle of about 45° to the axis of the bone in the present fossil and at 40° in P. ponderosus (see illustration in Marples 1952, pi. 7, fig. 1). In its size and general form the present bone resembles an Oligocene, New Zealand coracoid referred by Simpson (1971u) to Palaeeudyptes sp. indet. (specimen illus- trated in Marples 1952, pi. 7, fig. 2) and a coracoid possibly belonging to Anthro- pornis nordenskjoeldii Wiman, 1905 (illustrated in Wiman 1905, pi. 7, figs. 3, 3a and in Simpson I97\b, fig. 4), from Seymour Island. It differs from both, however, in the greater relative width of its shaft measured at right angles to the middle of the glenoid surface, and in the shaft being wider at the level of the middle of the precoracoid foramen, particularly with respect to A. nordenskjoeldii. Humerus of paratype, P.14 158a. A bone of relatively massive proportions with a large head and a gently curved broad shaft. The insertion of pectoralis seeondus is slightly oblique relative to the axis of the bone. The latissimus dorsi attachment, indicated only by a slight lip of bone on the edge of a fracture, is widely separated from the insertion of peetoralis secundus. The pectoralis tertius insertion is large and faces almost dorsally ; it extends distally about two-thirds the length of the pectoralis secundus insertion and is widely separated from the latter. The preaxial tuberosity on the shaft at the proximal limit of attachment of brachialis internus is especially prominent; it is obtusely angulate with a concave contour on the proximal side. The curvature of the articular surface of the head of this bone is identical with that of the fragmentary humeri of the holotype, P. 14157b and c (PI. 37, figs. 2, 3). The dimensions of the bone are comparable with those of several humeri of P. ponderosus. However, the preaxial tuberosity on the shaft is much more prominent than in that species and the pectoralis tertius insertion is less extended distally and more widely separated from the pectoralis secundus attachment. The bone differs from humeri assigned to A. nordenskjoeldii in that the preaxial tuberosity on the shaft is again more prominent and the pectoralis secundus insertion extends further distally relative to the latissimus dorsi attachment. Radius of holotype, P.14157d. A notable character in this bone is the shape of the EXPLANATION OF PLATE 39 Pachydyptes simpsojti sp. nov. Portions of bones which have been restored with plaster are outlined. Dashed lines (background) show a hypothetical restoration of the original outline of the bones. All figures natural size. Ifl, h. Fragment presumed to have been broken from the proximal part of the ?seventh rib on the left side, specimen P.17913. \a. dorsal view; its apparent position relative to the shaft or blade (s), dorsal articulation (da), and proximal head (ph) of the rib is indicated. Ifi, ventral view. 2a, h. Proximal end of right radius of paratype, P.141 58b. 2a, dorsal view, position relative to shaft of bone indicated. 2h, ventral view. 2>a, h. Left proximal phalange of second digit of holotype, P. I4I57L 3a, dorsal view. 3fi, ventral view. 4a, h. Right radius of holotype, P.14157d. 4a, dorsal view. 4fi, ventral view. 5a, h. Left carpometacarpus of holotype, P.14157e. 5a, dorsal view, position of missing first metacarpal and damage to third metacarpal indicated. 5b, ventral view. PLATE 39 JENKINS, Eocene penguin 300 PALAEONTOLOGY, VOLUME-17 brachialis internus attachment which is hollowed to form a distinct notch making an angle of about 120° with the preaxial angle of the shaft. The bone resembles a New Zealand Oligocene radius belonging to a Palaeeudyptes sp. (illustrated in Marples 1952, pi. 5, fig. 2) and radii assigned to A. nordenskjoeldii. It differs from both in the shape of the brachialis internus attachment. In the radius of Palaeeudyptes this insertion is long and moderately concave, not notched, while in A. nordenskjoeldii it is flat or only slightly concave. The shape of the insertion approaches that in Palaeospheniscus robustus Ameghino, 1905, from the latest Oligo- cene or early Miocene of Patagonia. Radius of paratype, P.14 158b. This fragment from the proximal end of the right radius is essentially similar to the radius of the holotype except for slight differences in the proportions and an additional prominence at the extreme proximal, post- axial corner of the dorsal muscle attachment. This prominence has almost certainly been abraided away in the holotype specimen. The overall similarity between the two is strongly suggestive of conspecificity. Carpometacarpus of holotype, P.14157e. This bone has the third metacarpal separated from the second for the greater part of its length. The third metacarpal is slightly expanded at about two-thirds its length and near the distal end, and projects dis- tally beyond the second. The distally projecting third metacarpal distinguishes this bone from the carpo- metacarpus of P. ponderosus which has the ends of the second and third metacarpals nearly level. The third metacarpal projects distally in a specimen belonging to a Palaeeudyptes sp. from the New Zealand Oligocene (illustrated in Marples 1952, pi. 5, fig. 7) and carpometacarpi assigned to A. nordenskjoeldii. Proximal phalange of second digit of holotype, P.14157f. This phalange resembles one apparently belonging to A. nordenskjoeldii and a New Zealand Oligocene bone of a Palaeeudyptes sp. (illustrated in Marples 1952, pi. 6, fig. 4). Its length is 61-5% of the carpometacarpus. The equivalent value for the Palaeeudyptes sp. is 57 0% (Marples 1952) and ranges from 63-5% to 71-0% in six recent genera (Marples 1953). Cervical vertebra of holotype, P.14157g. Comparison with the extant Aptenodytes forsteri Gray, 1844, suggests that this is probably the twelfth cervical vertebra. There appears to be only a low neural crest, and the prezygapophyses seem relatively elongate. The broken base of a median ventral hypapophysis is preserved on the centrum. Fragment of ?seventh rib, P.17913. The spheniscine origin of this fragment is ad- duced from its resemblance to part of the proximal portion of the seventh rib of Aptenodytes forsteri. It is tentatively referred to the new species described here because of its robust dimensions and occurrence in probably the same stratum and locality as the paratype P.14158. Measurements. The measurements of the more complete bones described are con- tained in Table 1. JENKINS: GIANT EOCENE PENGUIN TABLE 1 . MEASUREMENTS The dimensions given are in millimetres and unless otherwise stated are as numbered and defined by Marples (1952, Table 9). Measurements placed in brackets are approximate because of deficiencies in the specimen. Coracoid of holotype, P.141 57a Dimensions: I 2 3 4 5 6 (171) (133) (85) 33 Humerus of paratype, P.141 58a (33) 15 Dimensions: 1 2 3 4 5 6 7 8 9 Radius of holotype, P. \A\51d — 39 40 19 16 (103) Dimensions: 1 2 3 4 5 6 116 (19) 23 23 10 10 Metacarpus of holotype, P. 141 57c Dimensions: 12 3 4 102 104 — 31 Phalange of second digit of holotype, P.14157/ 1 . Extreme length between parallels 64 2. Width i length from proximal end 18 3. Width I length from proximal end 18 301 Discussion. The present species resembles Pachydyptes ponderosus more than any other known fossil penguin, and, as argued above, the two can be considered con- generic. Pachydyptes simpsoni differs principally from P. ponderosus in having the proximal half of the preaxial margin of the coracoid more strongly concave, the preaxial tuberosity on the shaft of the humerus much more prominent, the pectoralis secundus and pectoralis tertius insertions on the humerus more widely separated, and the third metacarpal produced distally beyond the second, not nearly level with it. The coracoid and paratype humerus of P. simpsoni are smaller than in the type of P. ponderosus', the humerus seems closely similar in size to other bones referred to the latter species. Selected dimensions of the bones of P. simpsoni and an emperor penguin, Apteno- dytes forsteri, the largest extant species, are compared in Table 2. The values of the ratio of the dimensions of the cervical vertebra and coracoid of the two species average T4(4); for the three wing bones the values of the ratio of the dimensions average 1-4(6). The fossil was thus probably of the order of 1-45 times as large as the modern specimen. The skeleton of A. forsteri used for comparison belonged to an individual with an approximate standing height of 98 cm (feet flat, legs and neck extended, head horizontal). Hence the fossil possibly had a standing height of the order of 140 cm. Simpson (1946) reviewed the problem of estimating the standing height of fossil penguins. Estimates of the height of New Zealand fossil penguins are given by Simpson (1971a). He suggested an estimate of 150 cm for the standing height of P. ponderosus. Allowing for the fact that P. simpsoni was clearly a little smaller than P. ponderosus, the present estimate is in close agreement with that of Simpson’s. P. simpsoni and the largest species of Palaeeudyptes were probably about equal in size. 302 PALAEONTOLOGY, VOLUME 17 / TEXT-FIG. 2. Comparison between the wing of Pachy- dyptes simpsoni sp. nov. and the wing of the largest extant species of penguin, Aptenodytesforsteri Gray ; a, reconstructed wing of fossil species; h, wing of A . forsteri. JENKINS: GIANT EOCENE PENGUIN 303 The ratios of the dimensions of the bones of P. simpsoni and A.forsteri (Table 2) suggest several interesting features. The comparative ratios for the length of the centrum of the cervical vertebra and the longitudinal measurement on the coracoid indicate the P. simpsoni probably had a relatively longer neck and shorter coracoids than the emperor penguin. On the other hand, the comparative ratio for the thickness of the shaft of the coracoid is similar to the comparative ratios for the lengths of the individual wing bones. This suggests that though the coracoid was relatively short, it was otherwise adapted to provide an anchorage for the very heavy-boned wing. The comparative ratios for the wing bones indicate that these differ slightly in their relative proportions between P. simpsoni and A.forsteri, but suggest that the ratio of wing length to body length was probably quite similar in the two. TABLE 2. Comparison between dimensions of selected bones of Aptenodytes forsteri and Pachydyptes simpsoni. Measurements in millimetres. Twelfth cervical vertebra Aptenodytes forsteri A Pachydyptes simpsoni B Ratio B/A Width of centrum just behind postzygapophyses 11-5 15-8 1-37 External width at middle of neural arch Length of centrum from upper portion of posterior 16-3 22-7 1-39 articulation to centre of anterior articulation Coracoid Length from distal hp of fossa scapidaris to basal margin at about one-quarter width from preaxial 24-3 43 (impf.) 1-7(8) corner Thickness of shaft at right angles to middle of 132 143 108 glenoid surface Wing bones 14-9 23-2 1-56 Extreme length of radius 82 116 1-42 Extreme length of third metacarpal 68 105 1-54 Extreme length of proximal phalanx of second digit 45 64 1-42 Simpson (1946, 1971a) champions the hypothesis that penguins descended from flying birds, and discusses various features of the fossil species which support this. He considers that penguins probably have a common ancestry with the Procellarii- formes. The relative shortness of the coracoid in the present fossil is an additional feature hinting at a volant ancestry. In flying birds the wing is supportive as well as propulsive and hence the most important flight muscles, the pectoralis major and pectoralis minor are directed obliquely downwards ; the coracoids which sustain the tension of these muscles are also directed obliquely downwards, and because the bird’s body is sub-cylindrical, are necessarily rather short. In Aptenodytes the flipper is largely propulsive as the body is buoyed up by the water it displaces, and the ‘flight’ muscles, and hence the coracoids, are directed obliquely rearwards. This condition permits lengthening of the flight muscles for greater efficiency and corre- sponding elongation of the coracoids. Penguins use their wings in a flying motion to propel themselves rapidly through the water. The small size of the wing compared to flying birds is clearly an adapta- tion of the wing to enable it to function efficiently in a relatively dense medium. 304 PALAEONTOLOGY, VOLUME 17 The wings of P. simpsoni seem to have been similarly shortened to those of modern penguins (e.g. A.forsteri) and thus were well specialized for an aquatic life. Simpson (1971fl) similarly concludes that features of the wing of early penguins were highly specialized. In modern penguins the neck is secondarily shortened by an extraordinary develop- ment of the antero-posterior curvatures of the cervical column (Watson 1883). Watson suggested that these flexures are related to the erect stance of penguins on land. The shortening of the neck may also enhance the hydrodynamic profile of the bird while swimming and aid conservation of body heat in cold climates. The long neck inferred for P. simpsoni might have been advantageous for the capture of highly motile prey such as fish. P. simpsoni may have inherited its long neck from the volant ancestor of the penguins or it may represent a specialized adaptation. Additional fossil material is needed to shed light on this. PALAEOECOLOGY Stonehouse (1969) and Simpson (1971a, 1971^) have focused attention on the palaeoecology of fossil penguins, particularly their possible adaptation to certain climatic regimes. However, these studies lack stratigraphic precision. Modern studies of the Tertiary increasingly suggest rapid temporal changes in climate (e.g. Devereux 1967; Wolfe 1971) and thus it is considered that investigations of possible relationships between Tertiary organisms and climate can only be rationalized with the closest attention to all possible forms of chronostratigraphic control. Though the available data is to some extent incomplete or inferential, recent foraminiferal and coccolith studies (Lindsay 1967, 1969, 1970; McGowran, Lindsay and Harris 1971; Edwards 1971; Hornibrook and Edwards 1971; Jenkins 1971) now pro- vide a basis for a tentative stratigraphic framework to reference the Upper Eocene fossil penguins of Australia with their New Zealand counterparts (text-fig. 3). The Blanche Point Marls, in which P. simpsoni occurs, contain a marine fauna. The shelly elements of the fauna are not notably diverse; molluscs, particularly gastropods are the commonest species and are present at most levels. Bryozoans are relatively frequent in the lower half of the formation. Abundant sponge spicules are a characteristic feature of the Banded and Soft Marls Members and moulds or silicified remains of sponges are common in the Banded Member. Tropical ‘larger’ foraminifera (Nummulitidae, Alveolinidae and Discocyclinidae) are unknown in the Upper Eocene of south-eastern Australia, and hence on extra- tropical environment can be presumed. Occurrences of the pantropical benthonic foraminifera Halkyardia and Linderina in southern Australia (Ludbrook 1961; Lindsay 1969) are considered to indicate warm conditions (McGowran 1971). The pantropical planktonic genus Hantkenina occurs in a well-defined narrow interval (Lindsay 1969; here text-figs. 1 and 3) and is suggested by Jenkins (1968) to be a warm-water indicator. These genera all occur in the near vicinity of Blanche Point (and Witton Bluflf), but seemingly not at the actual stratigraphic horizons of the fossil penguin remains. The nautiloids Cimomia and Aturia are relatively frequent fossils at several levels in the Blanche Point Marls (Glaessner 1955; McGowran 1959). Aturia has been JENKINS: GIANT EOCENE PENGUIN 305 SERIES EPOCH RANGES OF KEY PLANKTONIC MICROFOSSILS COMMON TO AUSTRALIA AND NEW ZEALAND SOUTHERN AUSTRALIA ST. VINCENT BASIN OTWAY BASIN SOUTHERN NEW ZEALAND DUNEDIN BURNSIDE QUARRY NEW ZEALAND MARINE STAGES OAMARU I T I — Cs 17-0 17-0 GLEN AIRE PORT WILLUNGA BEDS Top of range of Halkyardia & Linde rmq_ 11 12 9-5 CLAYS I 17-8 I MCDONALD Q Ilimestone ^ Ij3_;5 1^4 DEBORAH VOLCANIC FORMATION iTT CHINAMANS GULLY BEDS BLANCHE POINT SOFT MARLS p BLANCHE POINT BANDED CASTLE COVE LIMESTONE BROWNS CREEK transitional MARL MEMBER = yis-s ? greensand 16-5 17'0 \(^1^ ? CLAYS BURNSIDE • 17-6 , lejr pp MUDSTONE 12-3- ^0- 13-3 11-5 MAHENO MARL I \J9;1_2F^*/ WAIAREKA VOLCANIC FORMATION TORTACHILLA LIMESTONE R.XFT Lice ’ll TEXT-FIG. 3. Correlation chart for late Eocene and early Oligocene Australasian sequences known to contain fossil penguins, and other sequences for which palaeotemperature data has been obtained. The ranges of the planktonic microfossils on which the correlation is largely based are shown. Inferred or known occurrences of fossil penguins are indicated ‘P’; each indication represents one individual. The numbers ‘12-3, 16 1’ record the oxygen isotope temperatures (in °C) given by Dorman (1966, 1968) and Devereux (1967); the measurements are chiefly from benthonic fossils; an asterisk denotes measure- ments from planktonic fossils; values bracketed are considered unreliable. The numbers '20-25' indicate possible ranges of palaeotemperature (in °C) estimated by Mandra (1971) from silico-flagellate assemb- lages in the Oamaru Diatomite. 306 PALAEONTOLOGY, VOLUME 17 suggested as possibly indicating warm water (Cockbain 1968). Its significance is not unequivocal, however, as live individuals of the modern Nautilus are occasion- ally stranded on southern Australian shores and the shells of this genus are known to float long distances after the death of the animal (Stenzel 1964). P. simpsoni occurs in precise stratigraphic contiguity with Cimomia, and close, but perhaps not exact contiguity with Aturia. Further studies are needed to elucidate the possible environmental significance of other molluscs occurring in the marls. Dorman (1966, 1968) and Devereux (1967) present numerous relatively recent results of oxygen isotope temperature measurements made on fossils from the Australian and New Zealand Cenozoic; items of their data of relevance to Upper Eocene and early Oligocene occurrences of fossil penguins in these regions are given in text-fig. 3. Temperature estimates made by Mandra (1971) from studies on silico- flagellate assemblages in diatomites of similar age in the Oamaru area, southern New Zealand, are also shown. The isotopic temperature measurements suggest a warm climate during the time of appearance of Hantkenina primitiva, and probable cooler conditions shortly before and soon after this event; the silicoflagellate data seem concordant with this and, as here interpreted (in the light of datings given by Edwards 1971), may indicate another similar climatic cycle later in the Upper Eocene. The same cycle is apparently evidenced by an ingression of the pantropical, larger foraminiferal genus Asterocyclina in calcareous facies at Oamaru. The dating of the 12-3 °C isotopic temperature from the Burnside Mudstone, Dunedin, New Zealand, corresponds closely with that of the fossil penguins from the Blanche Point Transitional Marls. The penguin bones from this level at both Blanche Point and Witton Bluff seem, on coccolith evidence, to be identical in age. Fossil penguins also occur in the Burnside Mudstone (Simpson 1971fl) but their precise stratigraphic level is apparently not recorded. As was noted earlier in this work, the penguin Pachydyptes ponderosus occurs in the Totara Limestone at Oamaru. Samples of matrix associated with remains of its bones from its type locality. Fortification Hill, and also another occurrence, ‘Oamaru’, were examined by Finlay (1952), and Asterocyclina was not found. As Asterocyclina evidently occurs in most of the lower part of the Totara Limestone (Hornibrook 1961), the bones may postdate its local occurrence. P. ponderosus also occurs in the limestones at the locality ‘Taylors Quarry’: these are dated by Jenkins (1971) as being early Oligocene. Thus this penguin occurs in the early Oligo- cene and probably the latest Upper Eocene. Low isotopic temperature values for other exposures at Oamaru nearby to the localities of its remains (12-3-14-4°C) and also from south-eastern Australia (9-5-12 °C) probably relate approximately to the time of its occurrence. From the above discussion it is apparent that Pachydyptes simpsoni lived in an extra-tropical environment. Fluctuating warm and somewhat cooler climatic con- ditions seem to have obtained during the time of its existence; its occurrence at two levels in the Blanche Point Marls and probably other occurrences of Upper Eocene and early Oligocene Australasian fossil penguins may be linked to times of relatively low water temperatures, with bottom waters perhaps 10-14°C. The mean annual surface-water temperatures implied by this data are probably of the order of 12-16°C, or temperate. JENKINS: GIANT EOCENE PENGUIN 307 The above findings contrast with the suggestion made by Stonehouse (1969) that the Eocene fossil penguins of New Zealand inhabited sub-tropical to tropical waters. Nowhere in the vicinity of the penguin remains are there faunas of comparable diversity to those of the tropics. Simpson (1971a) refutes the popular view that penguins are principally associated with the Antarctic and its cold surrounding seas and concludes that the fossil species probably inhabited waters which were cool to warm temperate. He maintains a com- parable opinion even with respect to the relatively diverse early Tertiary assemblage TEXT-FIG. 4, Approximate positioning of southern hemisphere lands in the Upper Eocene and localities of the fossil penguin genera known or suggested to occur at this time. The hypothetical pattern of oceanic, surface circulation is shown. 308 PALAEONTOLOGY, VOLUME 17 of fossil penguins from Seymour Island (Simpson \91\b). However, he concedes that with respect to middle latitude occurrences of fossil species there may have been ‘local or seasonal conditions of cooler water (and land temperature) when penguins actually occurred alive at the fossil localities’ (Simpson 1971a). While broadly con- curring with these conclusions, I suggest that, with respect to the late Eocene and early Oligocene at least, there were probably periods of cooler climate (of significant geological duration) when the large penguin species of those times penetrated north- wards into middle latitude regions. This implies that the early penguins lived pre- dominantly at high southern latitudes. Reconstructions of the Eocene and Oligocene positioning of southern hemisphere land areas (McKenzie and Sclater 1971 ; Griffiths and Varne 1972) are modified here to show the probable positioning of these lands in the early Upper Eocene (text- fig. 4). The known probable Upper Eocene occurrences of fossil penguins are shown on this reconstruction; they form a southern distribution with the most northerly (southern New Zealand) possibly about 45° S. The abundance of penguin bones and diversity of forms recognized from the little-explored rocks of far-southern Seymour Island (the genera Anthropornis, Palaeeudyptes, and Wimanornis Simpson, 1971, Archaeospheniscus Marples, 1952, Delphinornis Wiman, 1905, and several indeterminate forms, and perhaps seven species; Simpson 1971^) contrasts remark- ably with the relative rarity of penguin remains in the better-known Eocene rocks of Australasia and can hardly be due to vagaries of preservation or collection. It is also notable that the known Australian and New Zealand Eocene penguin localities are restricted to southerly located exposures. The simplest inference is that the Australasian localities represent the northern limits of distribution of more diverse and denser southern, presumably Antarctic populations of early penguins. Thus at the time of their earliest known occurrence, penguins were seemingly already adapted to life in the relatively cool high-latitude seas of the southern hemisphere. Acknowledgements. Dr. M. R. Walter, of the Bureau of Mineral Resources, Canberra, is sincerely thanked for his enthusiastic encouragement and the considerable help which he gave at many stages of this project. Professor G. G. Simpson, of the University of Arizona, offered valuable advice during the early prepara- tion of materials. I am grateful to Emeritus Professor M. F. Glaessner of the University of Adelaide, for kindly making available additional bones and offering helpful criticism of the manuscript; to Dr. B. McGowran, of the University of Adelaide, for foraminiferal datings, constructive discussion, and criticism of the text; and to Mr. S. Shafik of the University of Adelaide, for coccolith datings. A grant from the Burdett Research Fund of the South Australian Museum, Adelaide, helped finance this project. REFERENCES BLOW, w. H. 1969. Late Middle Eocene to Recent planktonic foraminiferal biostratigraphy. In p. bronni- MANN and H. H. RENZ (eds.). Proceedings of the First International Conference on Planktonic Micro- fossils, Geneva, 1967, 1, 199-421, pis. 1-54. E. J. Brill, Leiden. 1970. Validity of biostratigraphic correlations based on the Globigerinacera. Micropalaeontology, 16, 257-268. COCKBAIN, A. E. 1968. Distribution of the nantiloid Aturia in the Eocene. J. Paleont. 42, 1309-1310. DEVEREUX, I. 1967. Oxygen isotope palaeotemperature measurements on New Zealand Tertiary Fossils. N.Z. Jl Sci. 10, 988-1011. DORMAN, F. H. 1966. Australian Tertiary palaeotemperatures. J. Geol. 74, 49-61. — 1968. Some Australian oxygen isotope temperatures and a theory for a 30-million-year world tem- perature cycle. Ibid. 76, 297-313. JENKINS: GIANT EOCENE PENGUIN 309 EDWARDS, A. R. 1971, A calcareous nannoplankton zonation of the New Zealand Paleogene. In a. fari- NACi (ed.), Proceedings of the II Planktonic Conference, Roma 1970, 381-419. Edizioni Tecnoscienza, Rome. FINLAY, H. J. 1952. Microfaunal notes on matrices associated with fossil penguin bones. In b. j. marples. Early Tertiary penguins of New Zealand. Palaeont. Bull. N.Z. 20, 58-64. FiNLAYSON, H. H. 1938. On the occurrence of a fossil penguin in Miocene beds in South Australia. Trans. R. Soc. S. Aust. 62, 14-17, pi. 1. GAGE, M. 1957. The geology of Waitaki Subdivision. Bull. geol. Surv. N.Z. 55, 1-135. GARTNER, s., JUN. 1971. Calcareous nannofossils from the JOIDES Blake Plateau cores, and revision of Paleogene nannofossil zonation. Tulane Stud. Geol. Geophys. 8, 101-121, pis. 1-5. GILL, E. D. 1957. The stratigraphic occurrence and palaeoecology of some Australian Tertiary marsupials. Mem. natn. Mus. Viet. 21, 135-203, pis. 1-4. 1959. Provenance of fossil penguin from western Victoria. Proc. R. Soc. Viet. 71, 121-123. GLAESSNER, M. F. 1951. Three foraminiferal zones in the Tertiary of Australia. Geol. Mag. 88, 273-283. 1955. Pelagic fossils {Aturia, penguins, whales) from the Tertiary of South Australia. Rec. S. Aust. Mus. 11, 353-372, pis. 34-36. GRIFFITHS, J. R. and VARNE, R. 1972. Evolution of the Tasman Sea, Macquarie Ridge and Alpine Fault. Nature Physical Science, 235, 83-86. HORNiBROOK, N. DE B. 1961. Tertiary foraminifera from Oamaru District (N.Z.). Part 1 — Systematics and distribution. Palaeont. Bull. N.Z. 34 (1), 1 192, pis. 1-28. and EDWARDS, a. r. 1971. Integrated foraminiferal and calcareous nannoplankton datum levels in the New Zealand Cenozoic. In a. farinaci (ed.). Proceedings of the II Planktonic Conference, Roma 1970, 649-657. Edizioni Tecnoscienza, Rome. JENKINS, D. G. 1968. Planktonic foraminifera as indicators of New Zealand Tertiary paleotemperatures. Tuatara, 16, 32-37. 1971. New Zealand Cenozoic planktonic foraminifera. Palaeont. Bull. N.Z. 42, 1-278, pis. 1-23. KENLEY, p. R. 1967. Tertiary. In G. bell et al.. Geology of the Melbourne district, Victoria. Bull. geol. Surv. Viet. 59, 31-46, pi. 2. LINDSAY, J. M. 1967. Foraminifera and stratigraphy of the type section of Port Willunga Beds, Aldinga Bay, South Australia. Trans. R. Soc. S. Aust. 91, 93-110, pi. 1. 1969. Cainozoic foraminifera and stratigraphy of the Adelaide Plains Sub-Basin, South Australia. Bull. geol. Surv. S. Aust. 42, 1-60, pis. 1-2. 1970. Port Willunga Beds in the Port Noarlunga-Seaford area. Quart. Geol. Notes geol. Surv. S. Aust. 36, 4-10. LUDBROOK, N. H. 1961. Stratigraphy of the Murray Basin in South Australia. Bull. geol. Surv. S. Aust. 36, 1-96, pis. 1-8. 1969. Tertiary Period. In L. w. parkin (ed.). Handbook of South Australian Geology, 1 72-203, Govern- ment Printer, Adelaide. and LINDSAY, J. M. 1966. The Aldingan Stage. Quart, geol. Notes geol. Surv. S. Aust. 19, 1-2. 1969. Tertiary foraminiferal zones in South Australia. In p. bronnimann and h. h. renz (eds.). Proceedings of the First International Conference on Planktonic Microfossils, Geneva, 1967, 2, 366-374, pis. 1-2. E. J. Brill, Leiden. mcgowran, b. 1959. Tertiary nautiloids (Eutrephoceras and Cimomia) from South Australia. J. Paleont. 33, 435-448, pis. 64-66. 1971. Australia- Antarctic separation and the Eocene transgression in southern Australia. South Aust. Dept. Mines, Geol. Surv. rep. No. 4559, 1-8 (unpublished). LINDSAY, J. M. and HARRIS, w. K. 1971. Attempted reconciliation of Tertiary biostratigraphic systems. In H. WOPFNER and j. g. Douglas (eds.). The Otwav Basin of Southeastern Australia. Special Bulletin, Geol. Survs. S. Aust., Viet., 1971, 273-281. MCKENZIE, D. and SCLATER, J. G. 1971 . The evolution of the Indian Ocean since the Late Cretaceous. Geophys. J. 25, 437-528. MANDRA, Y. T. 1971. Upper Eocene silicoflagellates from New Zealand. Antarct. J. U.S. 6, 177-178. MARPLES, B. J. 1952. Early Tertiary penguins of New Zealand. Palaeont. Bull. N.Z. 20, 1-66, pis. 1-8. 1953. Fossil penguins from the mid-Tertiary of Seymour Island. Sclent. Rep. Falkld Isl. Depend. Surv. 5, 1-15, pis. 1-2. 310 PALAEONTOLOGY, VOLUME 17 OLIVER, w. R. B. 1930. New Zealand birds. Fine Arts, Wellington. REYNOLDS, M. A. 1953. The Cainozoic succession of Maslin and Aldinga Bays, South Australia. Trans. R. Soc. S. Aust. 76, 1 14-140. SIMPSON, G. G. 1946. Fossil Penguins. Bull. Am. Mus. nat. Hist. 87, 1-99. 1957. Australian fossil penguins with remarks on penguin evolution and distribution. Rec. S. Aust. Mus. 13, 51-70. 1959. A new fossil penguin from Australia. Proc. R. Soc. Viet. 71, 113-1 19. 1965. New record of a fossil penguin in Australia. Ibid. 79, 91-93. 1970. Miocene penguins from Victoria, Australia, and Chubut, Argentina. Mem. natn. Mus. Viet. 31, 17-23, pis. 1-4. 1971a. A review of the pre-Pliocene penguins of New Zealand. Bull. Am. Mus. nat. Hist. 144, 319-378. 1971b. Review of fossil penguins from Seymour Island. Proc. roy. Soc. Lond. B, 178, 357-387. STENZEL, H. B. 1964. Living Nautilus. In R. c. moore (ed.). Treatise on Invertebrate Palaeontology, Part K. Mollusca 3. Cephalopoda, pp. 59-93. Univ. Kansas Press and Geol. Soc. Am. STONEHOUSE, B. 1969. Environmental temperatures of Tertiary penguins. Science, 163, 673-675. wade, m. 1964. Application of the lineage concept to biostratigraphic zoning based on planktonic fora- minifera. Micropalaeontology, 10, 273-290, pis. 1-2. WATSON, M. 1883. Report on the anatomy of the Spheniscidae collected during the voyage of H.M.S. Challenger. In J. Murray (ed.). Report on the scientific results of the voyage of H.M.S. Challenger during the years 1873-1876, Zoology, vol. 7. Longmans, London, Edinburgh. WOLFE, J. A. 1971 . Tertiary climatic fluctuations and methods of analysis of Tertiary floras. Palaeogeography, Palaeoclimatol., Palaeoecol. 9, 27-57. wiMAN, c. 1905. Uber die alttertiaren Vertebraten der Seymourinsel. Wiss. Ergebn. schwed. Siidpolarexped. 3, 1-37, pis. 1-8. R. J. F. JENKINS Department of Geology and Mineralogy University of Adelaide Adelaide, S.A. 5001 Manuscript received 29 November 1972 Australia LOWER DEVONIAN (DITTONIAN) PLANTS FROM THE WELSH BORDERLAND by DIANNE EDWARDS and JOHN B. RICHARDSON Abstract. Fertile specimens of Zosterophyllum Penhallow, Salopella gen. nov. (assigned to the Rhyniaceae), and Prototaxites Dawson are described from the upper group of the Ditton Series (Lower Devonian, approximately late Gedinnian or lower Siegenian of Europe) of Shropshire. Spores have been isolated from Salopella gen. nov., oval and fusiform carbonaceous masses, and from the matrix. It has recently been suggested by Allen et al. (1968) that an investigation of the ranges of plant macro- and microfossils coupled with one on vertebrates might help to solve two major stratigraphic problems of the Lower Old Red Sandstone in the Anglo-Welsh region: first, that of correlation between South Wales and the Welsh Borderland and secondly, the precise delimitation of boundaries within the stages. Croft (1953) defined the Breconian of South Wales on the basis of its flora and fauna (the latter being represented by one fragment of Rhinopteraspis cornubica) and the Dittonian and Downtonian lithologically. In the Welsh Borderland, however, the stages have been delimited by vertebrates (White 1950, 1956; Ball and Dineley 1961). Except for Lang’s work on Downtonian floras of the region (Lang 1937), no detailed accounts of plant assemblages have been published, although ‘plant debris’ is often mentioned in stratigraphical and palaeontological papers (Allen and Tarlo 1963; Ball and Dineley 1961). Richardson and Lister (1969) have described spore assemblages from South Wales and the Welsh Borderland. This report is part of an investigation concerned with the collection and description of both macro- and microfloras throughout the Downtonian and Lower Devonian of the Anglo-Welsh region. The beds containing the plants described below are exposed in both banks of the stream in the upper part of Newton Dingle, in the area known as The Gore. Ball and Dineley (1961), consider the section found in Newton Dingle to be the most complete for the upper group of the Ditton Series in the Brown Clee Hill regions. The plant locality is just above their locality 74 (see text-fig. 1) and is therefore at about 290 m above the main Psammosteus limestone. In the upper of the two corn- stones (locality 74) Ball and Dineley found Althaspis leachil and Europrotaspis crenulata, typical of the A. leachi Zone. Their locality 73 has a more sparse fauna but is still probably in the A. leachi Zone. White (1956) tentatively correlated the A. leachi horizons with the lower Siegenian of Europe. DESCRIPTION OF THE PLANTS The plants were found in a coarse grey-green, sometimes micaceous sandstone. This is overlain by a cornstone containing very fragmentary remains. The fossils in the less micaceous areas are occasionally associated with malachite. Plant and [Palaeontology, Vol. 17, Part 2, 1974, pp. 311-324, pis. 40-41.] 312 PALAEONTOLOGY, VOLUME 17 TEXT-FIG. 1 . Location of plant-bearing horizon in Newton Dingle. Plants occur at locality 1. Positions of fish-bearing cornstone localities 73 and 74 after Ball and Dineley (1961), which occur in the Althaspis leaclii zone, are indicated. indeterminate animal fragments either completely cover the bedding planes or are of sporadic occurrence. Plant axes sometimes exhibit parallel alignment. The plants are preserved as iron-stained impressions or black compressions, which revealed little or no structure on maceration with Schulze’s solution. Oxidation of the spore masses was more successful. The axes range between 0-2 mm and 2-5 mm with some fragments as long as 9 0 cm. The narrowest ones (PI. 40, fig. 10), though parallel sided, are flexuous, but the majority are straight and show dichotomous or, more rarely, pseudomonopodial branching, the latter usually occurring in wider axes. Also in these wider types are central strands. No evidence of tracheids is present. Scattered in the matrix are coiled narrow axes (PI. 40, fig. 7) and less frequently, hook-like structures. One specimen has a small projection (?axillary tubercle) below a branching point. Perhaps the most exciting find at this locality is a single specimen and its counterpart (70. 14G. 3a and b) which look very similar to the Rhynie chert genus Rhynia. For reasons outlined below, it was decided to erect a new genus for this type of compression fossil. Also present are oval- and cigar-shaped spore masses. All specimens are housed in the Department of Geology, National Museum of Wales, CardilT. EDWARDS AND RICHARDSON. LOWER DEVONIAN PLANTS 313 Division tracheophyta Subdivision zosterophyllophytina Banks 1968 Family zosterophyllaceae Krausel 1938 Genus zosterophyllum Penhallow 1 892 Zosterophyllum sp. Plate 40, figs, 1, 4, 8, 9 Material. Two incomplete fructifications and their counterparts were found. Description. Specimen 70.14G.la originally consisted of an apparently single axis bearing sporangia, but removal of the matrix revealed a bifurcation in the base of the fertile region giving rise to an additional branch to the left of the original (PI. 40, fig. 1). Part of the counterpart (70.14G.lb) showing two basal sporangia was found (PI. 40, fig. 4). The fossil is preserved partly as a carbonaceous compression and partly as an impression stained with limonite. No structure is present. A drawing of the uncovered specimen, in which the numbers assigned to the sporangia corre- spond to those in Table 1, is given in text-fig. 2. The specimen is 35 0 mm long with the width of the fertile axis ranging between TO mm and 2-7 mm immediately below the dichotomy. The left-hand branch bears six widely spaced sporangia, alternately inserted, but in the right- hand one, five sporangia are attached to one side of the axis. The two remaining sporangia (2a and 4a) are almost completely covered by the axis. It seems probable that there is a two-rowed arrangement here also, but with the rows borne on the same side of the axis. This assumption is supported by the arrangement in specimen 70.14G.2a and b. Further uncovering is undesirable since the specimen is likely to break up. One sporangium occurs below the dichotomy. The sporangia are borne on distinct stalks which are attached to the fructification axis at an acute angle though sometimes almost at a right angle. They are curved distally so that the sporangia are held upright with their long axes parallel to that of the fructification axis. In almost all cases the spor- angia are seen in side view, with approximately one- half of one valve visible. This is because the valves are folded towards the axis. The majority of the sporangia are partially overlain by the fructification axis. This phenomenon was seen in Zosterophyllum fertile of Leclercq (1942) and in the new Scottish occurrence of that species (Edwards 1972). A narrow border (0- 1 mm wide) can be seen in some sporangia (e.g. 4 and 4a). The second specimen 70.14G.2 is more compact (PI. 40, figs. 8, 9). It is 14 mm long and bears six closely packed sporangia, which overlap each other and the TEXT-FIG. 2. Left'. Salopella allenii gen. et sp. nov. A composite drawing from part and counterpart (70.14G.3a and b), x3-3. Right: Zosterophyllum sp. A line-drawing of specimen 70.I4G.la with sporangia! numbers correspond- ing to those in Table 1 , x 4-5. 314 PALAEONTOLOGY, VOLUME 17 TABLE 1. Dimensions of sporangia and stalks in Zosterophyllum sp. from the Welsh Borderland. No. (in text-fig. 2b) Stalk (in mm) Sporangium (in mm) Max. length Width Height above Width (outer margin) stalk attachment Specimen 70.14G.1 1 2-0 0-6 L3 0-9 2 1-5 0-6 2-8 1-2* 3 20 0-7 2-8 1-7* 4 2-0 0-7 2-1 L7* 4a 2-5 5 >10 0-9 2-3 1-7* 6 L7* 7 2-0 0-8 8 20 0-9 9 1-5 0-9 20 0-8* 10 1-3 0-5 20 L2 11 1-5 0-6 2-1 0-9* 12 1-2 0-4 1-2 Specimen 70.14G.2 1 0-8 2-2 2-Ot numbering from 2 2-2 basal sporangium 3 2-1 2-Ot upwards 4 2-0 M 5 not uncovered 6 1-2 0-8 2-2 1-5 * Margin of value covered by axis. t Face view. fructification axis. Before development three sporangia could be seen. These were in side view as in 70. 14G.1. Removal of the axis in the basal region has revealed two further sporangia, this time in face view. These alternate with the others. A third sporangium in the upper region has not been uncovered. The uncovered sporangia are almost circular in outline and have a border c. OT mm wide. They are illustrated in Plate 40, fig. 9. Thus this specimen also has two rows of sporangia, alternately arranged, but not borne on opposite sides of the axis. This is possibly similar to the arrangement of sporangia in the right-hand branch in 70.14G.1. Discussion. Recent papers by one of us (Edwards 1969a, b, 1972) have illustrated the difficulties of identification of isolated compressed Zosterophyllum fructifications. The present specimens are no exceptions. Although the variation in sporangial arrangement has a parallel with Zosterophyllum llanoveranum (Croft and Lang 1942; Edwards 1969a) sporangial size, shape, and insertion are quite different. There is a superficial similarity with the plants described from the Breconian, called Z. cL fertile (Edwards 1969^). They have in common the bifurcation at the base of the fertile region and type of sporangial insertion and arrangement. The difference in sporangial shape could be due to preservation. The sporangia are borne closer to the axis in the new specimens. In this respect they are intermediate between Z. cf. fertile and Z. fertile, where the stalks are at right angles to the axis and curve through 90° distally (Leclercq 1942). Indeed, the sporangia in face view are very similar to those illustrated by Leclercq. The Welsh Borderland specimens are again slightly larger than Z. c.^. fertile and sporangium shape in side view is not identical. There is EDWARDS AND RICHARDSON: LOWER DEVONIAN PLANTS 315 no dichotomy in the fertile region in the Belgian plant. Leclercq found one specimen only, so there is no evidence for variation in sporangial arrangement. The new speci- mens are therefore thought possibly related to Z. fertile, but as differences exist should be left as Zosterophyllum sp. Schweitzer (personal communication, 1971) has been reinvestigating the Lower Devonian Rhenish flora. In addition to Zostero- phyllum rlienanum, he has found a new type of fructification, which he will publish as a new Zosterophyllum species. This has many characteristics in common with both the Welsh Borderland and Brecon Beacons fructifications. It differs in that the sporangia do not appear to be folded in side view, so that the dehiscence line is visible. Subdivision rhyniophytina Banks 1968 Family rhyniaceae Kidston and Lang 1920 Genus salopella gen nov. Type species. S. allenii sp. nov. Derivation of name. From Salop, an alternative name for the county of Shropshire. Diagnosis. Plant consisting of naked dichotomously branching axes, preserved as compression fossils with terminal, probably erect fusiform sporangia containing miospores (probably isospores). Anatomy unknown. Salopella allenii sp. nov. Plate 40. figs. 2, 3 ; Plate 4 1 , figs. 1 -3 Diagnosis. Plant at least 24 mm high with naked dichotomously branching axes up to 2-0 mm wide in the basal regions and 1 • 1 mm wide below the sporangia. Branching angles small (c. 15°). Dichotomous branching immediately below sporangia, which are long and narrow ; 2 0 mm at widest point in mid-region and up to 9 0 mm long. Spores azonate with thin exine variously covered with probable tapetal residue. Locality. Newton Dingle, Brown Clee Hill, Shropshire. Exposures in banks of stream just below North Lodge Farm, in area known as The Gore (SO 58 60048295). Plant horizon, grey-green micaceous sand- stone basal 5 cm with malachite flecks associated with grey-green siltstone. Overlain by brown sandstone with grey-green siltstone intercalation. Holotype. Specimens 70. 14G.3a and b. Department of Geology, National Museum of Wales, Cardiff. Derivation of name. After J. R. L. Allen who took us to the locality. Description of material. The generic and specific diagnoses are based on a single specimen and its counter- part. The axes are preserved as casts, but the sporangia are completely compressed. When the rock was split no sporangia were visible. Uncovering revealed two out of a probable four sporangia. (The two miss- ing ones were probably lost in a rock sliver on splitting the rock.) Considering 70. 14G.3a, the sporangium is erect, 9 0 mm long and tapering at base and apex. It attains a maximum width at its mid-point, where it is 2 0 mm wide. A dichotomy is seen 10 mm below the sporangium, but a line extends a further 3-5 mm beyond this along the centre of the axis. Because of the very narrow branching angle it is not known where the actual separation of the axes occurs. The axes above the dichotomy are considerably narrower than those below. The second sporangium was uncovered on the counterpart (70. 14G.3b). Here the sporangium is bent over because the axis immediately below it is curved. The sporangium again has a maximum width of 2 0 mm and is 7 0 mm long, but is incomplete at the apex because of irregular rock fracture. A reconstruc- tion of the specimen is given in text-fig. la. There is no anatomy preserved in the axes, but spores have H 316 PALAEONTOLOGY, VOLUME 17 been isolated from the sporangia (PI. 41, figs. 1-3). (The spores isolated from the sporangia 70.14G.3 are mounted on slides 70.14G.7 and 8.) These are simple and azonate; have a circular to subcircular amb; trilete mark indistinct but faint sutures are present and the exine is sometimes split along these rays. Exine thin, not exceeding 1 (nm and usually less, secondary compression folds common; exine laevigate but usually partially, or entirely, covered by irregular elements, possibly tapetal residue. Size range 23 (29-7) 37 ;ixm. Possibly curvatural thickenings are present but this is uncertain as preservation is poor. The thin nature of the wall, presence of possible tapetal material, and lack of features comparable to dispersed spores may be a reflection of immaturity. Discussion. It is immediately obvious that this plant with its naked dichotomously branching axes and terminal sporangia has affinities with members of the Rhyniaceae. Of the genera listed by Banks (1968) it most closely resembles Rhynia itself and pos- sibly Hedeia. The remarkable preservation of the Rhynie chert petrifaction flora allows almost complete anatomical descriptions of the plant present. Comparison with compres- sion fossils is therefore difficult, particularly when, as in this case, external shape is almost the only available diagnostic character. While sporangial shape and branch- ing are similar to those in Rhynia, the absence of any anatomical data makes it impossible to assign the plant to that genus with any certainty. The axes dimensions are intermediate between the two Rhynia species (Kidston and Lang 1917, 1920), while the sporangia are perhaps slightly closer to R. gwynne-vaughanii. However, no hemispherical protuberances nor adventitious branching have been seen. Spores have been previously described from two species of Rhynia, R. gwynne- vaughanii and R. major (A. A. Bhutta 1969, unpublished Ph.D. thesis. University of Wales) but no precise comparisons can be made because of the nature of the Salopella spore material. However, Bhutta’s description of the spores of the Rhynia species reveals that both show retusoid characters but whereas the spores of R. gwynne- vaughanii show ‘. . . Curvaturae present in most of the specimens’ in R. major spores ‘. . . Curvaturae have been noticed in only one specimen’. Also they appear to have a similar ornamentation to spores released by bulk maceration from the Newton Dingle sample which are here referred to the dispersed spore species Perotrilites microbaculatus Richardson and Lister 1969. The dimensions of the sporangia and axes of Hedeia corymbosa (Cookson 1935) are similar to the Dittonian ones, but branching is much more frequent. In an Australian specimen of similar height, four or five dichotomies occur, producing a much-branched structure with some of the axes terminating in sporangia while the remainder are sterile. The comparable Dittonian specimen has only two dicho- EXPLANATION OF PLATE 40 Figs. 1, 4, 8, 9. Zosterophylhmi sp. 1, fructification, x 3-3 (70. 14G.la). 4, fragment of counterpart illus- trated fig. 1, X 3-4 (70. 14G.lb). 8, 9, part and counterpart of a fragment of fructification. 8, x4-4 (70.14G.2a). 9, x 4-2 (70.14G.2b). Figs. 2, 3. Part and counterpart of fertile Salopella. 2, x2-5. 3, x2-8 (70.14G.3a and b). Figs. 5, 6. Compressed spore masses in plant debris (indicated by arrow on fig. 5). 5, x 3 0 (70.14G.4a). 6, X L6(70.14G.4b). Fig. 7. Isolated circinate tip, x 3-5 (70.14G.5). Fig. 10. Narrow sterile axes (indicated by arrow), x 1-8 (70.14G.6). Fig. 1 1. Debris with Prototaxites sp., showing longitudinal striations, x 2-9 (70. 14G.2a). PLATE 40 EDWARDS and RICHARDSON, Lower Devonian plants 318 PALAEONTOLOGY, VOLUME 17 tomies with a possible maximum of four sporangia, although, of course, the specimen could be part of a much larger branching system. The shapes of the sporangia in Cookson’s illustrations are not clearly seen and the sporangia are often incomplete. Thus it is considered unwise to place the new Dittonian fossils in either Rhynia or Hedeia. Furthermore, it is thought there is a need to erect a new genus to accommodate this type of organization in a compression fossil. The specimens named Rhynia major by Ishchenko (1969) might also be placed in this genus as there was no ana- tomical justification for her identification (Banks 1972). INCERTAE SEDIS ( ?Algac) Genus prototaxites (Dawson 1859) Longitudinally, but irregularly, furrowed axes, up to 35-0 mm wide are commonly found (PI. 40, fig. 11). On maceration, longitudinally running tubes are recovered. The fossils are therefore assigned to the genus Prototaxites, but specific identifica- tion is impossible. Sections through small fragments reveal large, but not small, tubes. The plant frequently occurs as irregularly shaped, heavily carbonized flakes, scattered among plant axes and animal fragments. Pachytheca, although usually common in the Welsh Borderland Lower Devonian, has not been recorded from this locality. DESCRIPTION OF MICROFLORA Description of spore masses and spores, occurring as compression fossils Some of the small fragments of carbonaceous material, abundant in the matrix, have a more definite outline and fusiform shape. Typical examples are illustrated in Plate 40, figs. 5, 6. When the black powdery material is cleared with Schulze’s solu- tion, spore masses are recovered but the majority are either fused together or too poorly preserved to be illustrated. EXPLANATION OF PLATE 41 All magnifications x 1000 except where stated otherwise. Instrument settings for Zeiss Photomicroscope No. 4709856. Figs. 1-3. Spores from sporangia of Salopella. 1, slide 70. 14G. 7/1701099 showing irregular, possibly tapetal, residue loosely adhering to a smooth exine. 2, slide 70. 14G. 7/1251025, two spores showing nature of possible tapetal residual. 3, slide 70. 14G. 8/1371 100, tetrad; all these spores possibly repre- sent an immature stage since exactly comparable forms have not been seen in dispersed spore assemblages. Figs. 4-10. Spores recovered from spore mass after bulk maceration. 4, slide 70. 14G. 9/1281049, spores similar to the dispersed spore species Archaeozonolriletes chulus var. cliulus showing curvatural thicken- ings. 5, slide 70. 14G. 10/1621069, cf. Ambitisporites dilutus. 6, slide 70.14G.il/1411151, cf. Ambiti- sporiles avitus partially covered by ‘tapetal’ residue. 7, slide 70. 14G. 9/1601093, cf. Archaeozonolrileles chulus with tapetal residue. 8, slide 70. 14G. 9/1921062, x750, tetrad cf. A. chulus showing equatorial crassitude (thickening). 9, slide 70. 14G. 9/1001045, tetrad of Perolrililes microbaculatus showing fine granulate scuplture. 10, slide 70. 14G. 12/1201 106, lateral view of specimen of /I. chulus showing rela- tively thin proximal surface. Fig. 1 1. Isolated spore found after bulk maceration, slide 70. 14G. 13/1 13 1086, x 500, cf. Ambitisporites dilutus. PLATE 41 EDWARDS and RICHARDSON, Lower Devonian spores 320 PALAEONTOLOGY, VOLUME 17 Two of these spore masses picked from the rock yielded essentially similar spores (preparations 70.14G.14 and 15) which can be assigned to the dispersed spore genus Ambitisporites Hoffmeister 1959 and are of considerable interest because they are similar to A. dilutus (Hoffmeister) one of the first, if not the first, trilete land plant spore to appear in the geological record. Hoffmeister’s spores are now regarded as middle Llandovery in age (based on a reassessment of the graptolites by Berry, see Gray and Boucot 1971). The size range of the Newton Dingle spores is 32-44 compared with 30-60 for Hoff- meister’s spores. Spore masses released by bulk maceration Both spore masses and spores were released by bulk maceration. Some of the best-preserved spores were from the spore masses which revealed spores of three types. The spores are assigned to the dispersed spore species Perotrilitesmicrobaculatus Richardson and Lister 1969 (preparations 70.14G.il and 9), spores with a thick exine and a microbaculate-granulate ‘perispore’ ; Arcliaeozonotriletes clndus (Cramer) 70.14G.16, size range 49-78 fxm, laevigate spores with a thin diaphanous proximal surface and a broad equatorial crassitude; and Ambitisporites dilutus 70.14G.17, size range 40-49 jjcm, laevigate spores with a narrow equatorial crassitude. It is possible that only one species is present and all the spores may come from one parent plant. Firstly, Archaeozonotriletes clmlus and Ambitisporites cf. dilutus both have an equatorial crassitude and may intergrade and, secondly, these spores may represent the inner bodies of the spores assigned to P. microbaculatus. The inner bodies of the latter although imperfectly seen through the ‘perispore’ are laevigate with thick distal walls and an equatorial crassitude similar to spores assigned to the Ambitisporites- Archaeozonotriletes chulus complex. The spore species mentioned above also form a prominent part of the dispersed spore assemblage which is dominated by laevigate spores. Species recorded in addition to those mentioned above are Retusotriletes warringtonii, R. cf. minor, Apiculiretusispora microconus, A. cf. spicula, Acinosporites salopiensis, and one specimen of Emphanisporites, E. cf. micrornatus (see Richardson and Lister 1969). The total assemblage is thus a very restricted one consisting of a total of nine species compared with twenty-seven species recorded from the middle Ditton Group. A further feature is the strikingly poor representation of the genus Emphanisporites. The species present, although clearly unrepresentative of the spore flora of the upper part of the Ditton Group, are typically Dittonian and show no specific differences with previously described Dittonian spores from lower horizons. It is noteworthy that the number of dispersed spore species identified is small compared with the number of species from a good siltstone or shale assemblage. This paucity of species and lack of variety is probably largely preservational and/or depositional due to the coarseness of the sediment and sorting factors. GENERAL DISCUSSION This flora is of interest because it is the first detailed record from the Dittonian rocks of the Welsh Borderland. It is intermediate in age between Lang’s Downtonian floras EDWARDS AND RICHARDSON: LOWER DEVONIAN PLANTS 321 (Lang 1937) and the very extensive Senni Bed floras of South Wales (Croft and Lang 1942; Edwards 1970a). Lang’s plants were found at various localities in Pembroke- shire and the Welsh Borderland ranging through King’s horizons 11-18. He de- scribed Prototaxites, Pachytheca, Nematothallus, Parka, two Cooksonia species, and cf. Zosterophyllum myretonianum, the last two genera having most significance for this account. The age of the rock in which the Zosterophyllum specimen (represented by an H-branch) was found is uncertain. Lang considered it to be from the 18 hori- zon, i.e. from the upper Downtonian of Caldy Island, Pembrokeshire, but it is pos- sible that it comes from younger rocks. Our own collections from Lower Dittonian localities have so far yielded Parka, Pachytheca, and Prototaxites. Dichotomous branching predominates in the presumed vascular axes and there is a noticeable increase in axis diameter in younger rocks. This new Dittonian flora, possibly lower Siegenian, contains Zosterophyllum and Rhynia {= Cooksonia) types of organization and therefore shows little advance on the Downtonian assemblages. In contrast, the Breconian floras of neighbouring South Wales are strikingly dif- ferent, including members of the Rhyniophytina, Zosterophyllophytina, Lycophytina, Trimerophytina, and Barinophytaceae. As it is not yet known to what extent the Newton Dingle assemblage is representative of the upper Dittonian flora of the region, a discussion of the evolutionary significance of these differences is perhaps premature. Unfortunately there are very few floras of identical age in other parts of the world which could be used for comparison. That described by Leclercq ( 1 942) from Belgium, containing Hostimella, Taeniocrada decheniana, and Zosterophyllum fertile is pos- sibly the only one. Streel (1967) gives its age as lowermost Siegenian. It is interesting that the types of plant present are similar to those in the Newton Dingle assemblage. The slightly younger lower Siegenian floras described by Stockmans (1940) are more extensive including Psilophyton, Drepanophycus, and Sporogonites species (Streel 1967). These, together with the Siegenian floras of Germany, have more in common with the Breconian floras of South Wales (Edwards 1970a). The floras from Canada, U.S.A., Australia, and most of the U.S.S.R. (Chaloner 1970; Petro- syan 1968) are also much younger. Earlier Gedinnian floras have been reported from Scotland, Czechoslovakia, and Spitzbergen. A rather restricted flora dominated by Zosterophyllum myretonianum (Lang 1927), Prototaxites species (Lang 1926), and Parka decipiens (Don and Hick- ling 1915) occurs in the Dundee formation in the Arbuthnott group of the Lower Old Red Sandstone (Dittonian) (Armstrong and Paterson 1970). More recently, two further species, Cooksonia caledonica (Edwards 1970Z)) and Zosterophyllum fertile (Edwards 1972) have been described. Thus, during the Gedinnian times in Scotland, two basic types of vascular plant organization have emerged, naked dichotomously branching axes with terminal sporangia and naked axes with lateral sporangia, aggregated into terminal spikes. A similar pattern is seen in the Gedinnian of Spitzbergen (H0eg 1942; Friend 1961) where the assemblage includes Pachytheca, Prototaxites, Taeniocrada{l) spitzhergensis, and sterile Zosterophyllum sp. In the Gedinnian of Czechoslovakia, however, only the Cooksonia type of organization is present (Obrhel, 1968). A further plant Taeniocrada decheniana Krausel and Weyland is recorded from 322 PALAEONTOLOGY, VOLUME 17 probable Gedinnian rocks in Germany (Schmidt and Teichmiiller 1954). Here some- what recurved sporangia are borne in terminal clusters and Banks (1968) places the plant with Cooksonia in the Rhyniophytina. As Chaloner (1970) points out, the sporangia may be regarded as terminal on overtopped lateral branches and may perhaps be considered as intermediate between the two basic types of organization. The preceding Downtonian floras are admirably summarized by Banks (1972) who emphasizes that the only vascular plants present at this time were members of the Rhyniophytina. In conclusion, therefore, on the limited evidence at present available, the upper- most Gedinnian-basal Siegenian floras have more in common with the earlier Gedinnian floras than with later Siegenian ones, and that soon after the beginning of the Siegenian great diversification occurred, resulting in the extensive and varied floras of later Siegenian times. Miospore assemblages So far, the volume of work on miospore assemblages is not great for the early part of their record in the Silurian and Lower Devonian. However, published data from the British Isles, Canada, western Europe, North Africa, and the U.S.S.R. is con- sistent in showing the same trends of development with closely similar spore assem- blages occurring in strata of the equivalent age from these different areas. The assemblages show the gradual increase in diversity of miospores from their first appearance in the Silurian (Llandovery, or possible Middle Ordovician if the records of StenozonotriletesY\vs\o{Q&y (Timofeev 1963a, b) prove to be trilete spores), through Wenlockian, Ludlovian, Downtonian, and Gedinnian. These assemblages are dominated by small simple spores or spores with a narrow equatorial crassitude but show a progressive increase in the variety of sculptural types present, the increase in diversity of the dispersed spore genus Emphanisporites and the temporary appear- ance of proximal inter-radial papillae in several genera. This latter feature appears in the Downtonian and reaches an acme in the Gedinnian, but rapidly declines in the Siegenian and does not reappear until the Carboniferous when it is a feature of some lycopod spores. Siegenian spore floras show an increase in size with the incoming of more robust- ribbed Emphanisporites, they also show the waning of tripapillate spores and an increasing diversity and mean size of spore species, and the genus Dibolisporites makes its first appearance. In the Anglo-Welsh area these changes take place in the Senni beds (lower Siegenian), e.g. the spores from the Storey Arms quarry show these characteristics. While the most striking change in the Lower Devonian sequence of spore assem- blages occurs higher in the succession in beds of Emsian age where in several areas, e.g. Germany (Lanninger 1968; Riegel 1968) and Canada (McGregor 1967) there is a sudden change to spore floras of Middle Devonian aspect, there also seems to be a change of lesser but still of considerable importance somewhere near the Gedin- nian and Siegenian boundary. However, since very little has been published on the Siegenian, and this stage represents a considerable period of time, it is too early to be precise. Nevertheless, the changes outlined above for the Anglo-Welsh area seem to reflect a general floral change for the period since it apparently occurs in several EDWARDS AND RICHARDSON: LOWER DEVONIAN PLANTS 323 areas, e.g. Canada (McGregor et al. 1970), Belgium (Streel 1967), and North Africa (Jardine and Yapaudjian 1968). Acknowledgements. We are particularly grateful to Professor J. R. L. Allen for leading us to this and to several other localities in the Welsh Borderland. Dianne Edwards was supported by a University of Wales Fellowship, and is indebted to Professor G. F. Asprey for research facilities at University College, Carditf. J. B. Richardson gratefully acknowledges financial support from the Natural Environment Research Council to provide a Research Assistant and a Zeiss photomicroscope. He is also indebted to Dr. N. loannides for the photography. REFERENCES ALLEN, J. R. L. and TARLO, L. B. 1963. The Downtonian and Dittonian Facies of the Welsh Borderland. Geol. Mag. 100, 129-155. HALSTEAD (TARLO), L. B. and TURNER, s. 1968. Dittonian ostracoderm fauna from the Brownstones of Wilderness Quarry, Mitcheldean, Gloucestershire. Proc. Geol. Soc. London. 1649, 49, 141-153. ARMSTRONG, M. and PATERSON, I. B. 1970. The Lower Old Red Sandstone of the Strathmore region. Inst. Geol. Sci. Kept. 70 (12), 1-23. BALL, H. w. and DiNELEY, D. L. 1961. The Old Red Sandstone of Brown Clee Hill and the adjacent area. Bull. Brit. Mus. (Nat. Hist.), Geol. 5, 177-242. BANKS, H. p. 1968. The early history of land plants. Pp. 73-107. In drake, e. (ed.). Evolution and Environ- ment. 1972. The stratigraphic occurrence of early land plants. Palaeontology, 15, 365-377, 1 pi. CHALONER, w. G. 1970. The rise of the first land plants. Biol. Rev. 45, 353-377, 3 pis. COOKSON, I. c. 1935. On plant remains from the Silurian of Victoria, Australia, that extend and connect floras hitherto described. Phil. Trans. Roy. Soc. London, 225B, 127-148, 2 pis. CROFT, w. N. 1953. Breconian; a stage name of the Old Red Sandstone. Geol. Mag. 90, 429-432. and LANG, w. H. 1942. The Lower Devonian flora of the Senni Beds of Monmouthshire and Brecon- shire. Phil. Trans. Roy. Soc. London, 231B, 131-163, 3 pis. DAWSON, J. w. 1859. On fossil plants from the Devonian rocks of Canada. Quart. J. Geol. Soc. London, 15, 477-488. DON, A. w. R. and hickling, g. 1915. On Parka decipiens. Ibid. 71, 648-666, 3 pis. EDWARDS, D. 1969fl. Further observations on Zosterophyllum llanoveranum from the Lower Devonian of South Wales. Amer. J. Bot. 56, 201-210, 3 pis. 19696. Zosterophyllum from the Lower Old Red Sandstone of south Wales. New Phytol. 68, 923- 931, 2 pis. 1970c/. Further observations on the Lower Devonian plant Gosslingia breconensis Heard. Phil. Trans. Roy. Soc. London. 258B, 225-243, 5 pis. 19706. Fertile Rhyniophytina from the Lower Devonian of Britain. Palaeontology, 13, 451-461, 4 pis. 1972. A Zosterophyllum fructification from the Lower Old Red Sandstone of Scotland. Rev. Palaeo- botan. Palynol. 14, 77-83, 1 pi. FRIEND, p. F. 1961. The Devonian stratigraphy of north and central Vestspitzbergen. Proc. York. Geol. Soc. 33, 77-118. GRAY, J. and BOUCOT, A. J. 1971. Early Silurian spore tetrads from New York : Earliest New World evidence for vascular plants? Science, 173, 918-921, 2 figs. H0EG, o. A. 1942. The Downtonian and Devonian flora of Spitzbergen. Norges. Svalbard Ishavs-Undersijrk, 83, 1-228, 62 pis. HOFFMEISTER, w. s. 1959. Lower Silurian plant spores from Libya. Micropaleontology, 5, 331-334, 1 pi. ISHCHENKO, T. A. 1969. The Cooksonia palaeoflora in the Skala horizon of Podolia and its stratigraphical significance. Geol. J. (Kiev), 29, 101-109, 10 figs. [In Russian.] JARDINE, s. and YAPAUDJIAN, L. 1968. Lithostratigraphie et palynologie au Devonien-Gothlandieu greseux du bassin de Polignac (Sahara). Rev. Inst. Er. Petrol. 23, 439-468. 324 PALAEONTOLOGY, VOLUME 17 KiDSTON, R. and lang, w. h. 1917. On Old Red Sandstone plants showing structure, from the Rhynie Chert bed, Aberdeenshire, Part I. Rhynia Gwynne-Vaughani, Kidston and Lang. Trans. Roy. Soc. Edinburgh, 51, 761-784, 10 pis. 1920. Part II. Additional notes on Rhynia Gwynne-Vaughani Kidston and Lang; with descriptions of Rhynia major, n.sp. and Hornea lignieri n.g., n.sp. Ibid. 52, 603-627, 10 pis. LANG, w. H. 1926. Contributions to the study of the Old Red Sandstone Flora of Scotland. V. On the identification of the large ‘stems’ in the Carmyllie Beds of the Lower Old Red Sandstone as Nemato- phyton. Ibid. 54, 792-799, 2 pis. 1927. VI. On Zosterophyllum myretonianum Penh., and some other plant remains from the Carmyllie Beds of the Lower Old Red Sandstone. Ibid. 55, 443-452, 2 pis. 1937. On the plant-remains from the Downtonian of England and Wales. Phil. Trans. Roy. Soc. London, 227B, 245-291, 7 pis. LANNiNGER, E. p. 1968. Sporcn-Gesellschaften aus dem Ems der S.W. Eifel. Palaeontographica (B), 122, 95-170, 7 pis. LECLERCQ, s. 1942. Quelques plantes fossiles receuillies dans le Devonien inferieur des environs de Non- ceveux (Bordure orientale du bassin de Dinant). Ann. Soc. Geol. Belg. Bull. 65, 193-211, 3 pis. MCGREGOR, D. c. 1967. Composition and range of some Devonian spore assemblages of Canada. Rev. Palaeobot. Palynol. 1, 173-183, 1 pi. SANDFORD, B. V. and NORRIS, A. w. 1970. Palynology and correlation of Devonian formations in the Moose River basin, northern Ontario. Proc. Geol. Ass. Canada, 22, 45-54, 2 pis., 3 figs. OBRHEL, J. 1968. Die Silur- und Devonflora des Barrandiums. Paldont. Abh. B, 2, 635-701, 5 pis. PENHALLOW, D. p. 1892. Additional notes on Devonian plants from Scotland. Can. Record Sci. 5, 1-16, 2 pis. PETROSYAN, N. M. 1968. Stratigraphic importance of the Devonian flora of the U.S.S.R. Pp. 579-586. In OSWALD, D. H. (ed.). Internal. Symposium on the Devonian System, Calgary, Canada, 1967, 1. RICHARDSON, J. B. and LISTER, T. R. 1969. Upper Silurian and Lower Devonian spore assemblages from the Welsh Borderland and south Wales. Palaeontology, 12, 201-252, 7 pis. RiEGEL, w. 1968. Die Mitteldevon-flora von Lindlar (Rheinland) 2. Sporae dispersae. Palaeontographica (B), 123, 76-96, 5 pis. SCHMIDT, w. and teichmuller, m. 1954. Pflanzen-Reste aus dem Gedinne des Hohen Venns. Geol. Jb. 69, 89-102, 2 pis. STOCKMANS, F. 1940. Vegetaux eodevoniens de la Belgique. Mem. Mus. roy. Hist. nat. Belg. 93, 1-90, 14 pis. streel, m. 1967. Associations de spores du Devonien inferieur Beige et leur signification stratigraphique. Ann. Soc. Geol. Belg. Bull. 90, 11-54, 5 pis. TIMOFEEV, B. V. 1963«. Ordovician and Silurian phytoplankton of the Siberian Platform. Dokladv Akad. Nauk SSSR, 149, 399-402. 19636. Phytoplankton and dispersed spores of the Ordovician, Silurian and Lower Devonian of the Baltic region, the Gory Swietokvzyskie and Podolia. Ibid. 150, 158-161. WHITE, E. I. 1950. The vertebrate faunas of the Lower Old Red Sandstone of the Welsh Borders. Bull. Brit. Mus. (Nat. Hist.) Geol. 1, 51-67. 1956. Preliminary note on the range of pteraspids in Western Europe. Bull. Inst. roy. Sci. nat. Belg. 32, 1-10. DIANNE EDWARDS Department of Botany University College Cardiff, CFl IXL JOHN B. RICHARDSON Department of Geology King’s College Strand London, WC2R 2LS Revised typescript received 15 May 1973 MEGALOSAURIDS FROM THE BAJOCIAN (MIDDLE JURASSIC) OF DORSET by MICHAEL WALDMAN Abstract. A new megalosaurid, Megalosaurus hesperis sp. nov., is described from the Upper Inferior Oolite (Bajo- cian) of Dorset ; the type material was previously assigned to M. hucklandi. The holotype of M. nethercomhensis von Huene, 1923, from the Middle Inferior Oolite of the same county, is redescribed; von Huene’s proposal of a new genus (Magnosaums 1932) for this species was unjustified. Sarcosawus andrewsi is transferred to the genus Megalo- saitrus. Among the many reptilian remains recovered from the Middle Jurassic strata of Dorset (Delair 1958, 1959, 1960, 1966) are two sets of megalosaurid bones. These were partially prepared by the author in 1964-1965, with the permission of the British Museum (Natural History) and Oxford University Museum. One of these specimens (R.332) has long been referred to Megalosaurus bucklandi by various authors, and it was only in 1964 that Walker first commented in print on the disparity in tooth-count between R.332 and M. bucklandi. Tooth-counts are a good diagnostic feature in the carnivorous saurischians, and the material is, therefore, redescribed and specifically separated from M. bucklandi. The second specimen (O.U.M. J. 12143) consists of a few rather ill-preserved post- cranial elements and an associated pair of dentaries. The latter had to be freed from matrix before any description could be undertaken. There is no doubt that the material represents a separate species of Megalosaurus, M. nethercombensis, as noted by von Huene. It may also belong to a juvenile individual. Since most of the Middle Jurassic sediments in this country are at least partially marine in origin these remains of terrestrial reptiles are significant in establishing that at least three species of Megalosaurus existed at that time in what is now the United Kingdom. Nomenclature. The generic name AUosaurus is used in preference to Antrodemus (Matthew and Brown 1922). Both names probably refer to the same animal, but this remains unproven. Abbreviations. O.U.M. or J. = Oxford University Museum. B.M.(N.H.) or R. = British Museum (Natural History). N.M.C. = National Museum of Natural Sciences, Ottawa, Canada. SYSTEMATIC DESCRIPTIONS Order saurischia Suborder theropoda Infraorder carnosauria Family megalosauridae Note. For recent discussions and diagnoses of the above taxa, reference may be made to Colbert (1964), Walker (1964), and Charig, Attridge and Crompton (1965). [Palaeontology, Vol. 17, Part 2, 1974. pp. 325-339, pis. 42-44.] 326 PALAEONTOLOGY, VOLUME 17 Genus Megalosaurus Buckland, 1824 Diagnosis. Twelve to thirteen teeth in maxilla and dentary, tooth carinae positioned anteriorly and posteriorly, not obliquely. Dentary straight, with symphysial facet. Vertebrae short; scapula large, with anterior expansion of middle part of blade; humerus stout; pubis with small distal thickening, extensive symphysis, no foot; ischium down-curved posteriorly ; femur massive, lesser trochanter placed well down shaft; tibia stout, 83% of femur (maxima of non-associated bones). (Diagnosis after Dr. A. D. Walker, personal communication, with minor additions by the author.) Type species. Megalosaurus bucklandi von Meyer, 1832. Megalosaurus liesperis sp. nov. Plates 42, 43 1883 Megalosaurus bucklandi von Meyer; Owen, p. 334, pi. XI. 1884 Megalosaurus bucklandi von Meyer; Owen, p. 166, pi. LXXXVII. 1890 Megalosaurus bucklandi von Meyer; Woodward and Sherborn, p. 249 {pars). 1926a Megalosaurus bucklandi von Meyer; von Huene, p. 37, item 14. 1932 Megalosaurus bucklandi von Meyer; von Huene, p. 220 (pars). 1934 Megalosaurus bucklandi von Meyer; Swinton, pp. 214-215 (pars). 1959 Megalosaurus bucklandi von Meyer; Delair, p. 78. 1960 Megalosaurus bucklandi von Meyer; Edmund, p. 130, fig. 43(k). 1964 "Megalosaurus bucklandi' von Meyer; Walker, p. 115. Diagnosis. A large megalosaurid, fifteen to eighteen maxillary teeth, seventeen or eighteen dentary teeth; only apical part of dentary teeth recurved. Holotype. B.M.(N.H.) R.332; a right maxilla, parts of both premaxillae, both dentaries, part of the right surangular, part of the ?vomer, and an isolated tooth (PI. 42, figs. 1-3). Derivation of specific name. Greek; hesperos— the West, western. Locality. There appears to be some confusion in the literature with regard to the exact provenance of the M. liesperis material. Owen (1883, 1884), and Mansel-Pleydell (1888) did not give exact locations in the Sherborne area, while Buckman (1893) gave the location as: ‘Redhole Lane, Sherborne,— (about 1 mile N. of the Abbey).’ Edward Cleminshaw, who originally obtained the material, is quoted by Richard- son (1916) as stating: ‘The site of the quarry in which the remains were found is very near the back of the houses on the north side of Cold Harbour Road, . . .’ It is evident from Richardson (1932) and the Directory of British Fossiliferous Localities that these locations do not coincide, and as Cleminshaw secured the material [although not its discoverer (Richardson 1916)] I have no reason to doubt his state- ment of the provenance of the specimen. Horizon. Upper Inferior Oolite, Parkinsonia parkinsoni Zone, Garantiana garantiana subzone; Bajocian, Middle Jurassic. (Stratigraphy according to Wilson et al., 1958.) EXPLANATION OF PLATE 42 Eigs. 1, 2, 3. Posterior, lateral, and anterior views of a tooth of Megalo.saurus liesperis. B.M.(N.H.) R.332. Natural size. Pig. 4. Lingual view of part of the left premaxilla of M. liesperis. B.M.(N.H.) R.332, xO-5. Pig. 5. Buccal view of right maxilla and partial premaxilla of M. liesperis. B.M.(N.H.) R.332, x 0-3. Arrows indicate premaxillary teeth. Pigs. 6, 7, 8. Buccal, posterior, and lingual views of the anterior part of the right surangular of M. liesperis. B,M.(N,H.) R.332, xO-5. Note. Numbers prefixed by ‘m’ indicate maxillary teeth, numbering from the anterior. PLATE 42 awpl » S ' «o> » g^. WALDMAN, Bajocian megalosaurids 328 PALAEONTOLOGY, VOLUME 17 Description. The right maxilla (PI. 42, fig. 5) is represented on two facing slabs and was first described by Owen (1883). The posterior and anterior margins of the maxilla are broken, but the ascending process is well preserved for much of its length. The maxillary-premaxillary junction is visible, although heavily plastered. Twelve maxil- lary teeth are preserved (Walker 1964) and comparison with Allosaurus tends to confirm Walker’s (1964) estimate of fifteen to eighteen teeth for the whole maxilla. These teeth differ little from those of Megalosaurus bucklandi. The maxilla was used by Edmund (1960) to illustrate the mode of tooth replacement in that species and by von Huene (1926a) in his ‘construction’ of the skull of M. bucklandi. The teeth are large and very deeply set in the maxilla, almost reaching the level of the ventral margin of the second antorbital fenestra. The serrations on the posterior carinae are all perpendicular to the long axis of each tooth, but obliquely set serra- tions are visible on some anterior carinae. The general shape of the maxilla is clearly seen in PI. 42, fig. 5. At the second ant- orbital fenestra the process is compressed, presumably for contact with the nasal and lachrymal elements (cf. Gilmore 1920). A pronounced vertical groove is present between maxilla and premaxilla, probably for the admission of the crown of a large dentary tooth (Walker 1964). Four premaxillary teeth are visible, two being preserved as crowns, one as a cross- section of its original position, and one as an imprint of a juvenile tooth. A cross- section of another ?extraneous tooth is also present immediately below the imprint. Walker (personal communication) has commented that there may have been five teeth in the premaxilla, as in Allosaurus, both genera having a similarly high maxillary tooth count. Only the postero-dorsal region remains of the right premaxilla, although the rest of the bone has been cast in plaster from the premaxillary impression on the opposing slab. The left premaxilla (PI. 42, fig. 4) is broken away posteriorly and only the first two alveoli are preserved, but even so there is a marked resemblance to the premaxilla of Allosaurus (Gilmore 1920). The lingual face for contact with the opposite premaxilla is flat, apart from a large shallow, neural groove which runs antero-ventrally. The first alveolus is empty, but the second reveals a large tooth lacking the crown, with a well-preserved replacement tooth lying lingually. Part of the imprint of the third premaxillary tooth is also present. Remains of interdental plates are preserved, separated from the sutural surface of the premaxilla by a deep blood-vessel groove and a bony parapet which curves in toward the second alveolus. Above the second interdental plate there is a circular foramen which runs dorsally. The antero-ventral face of the premaxilla is pitted with minute foramina close to the mid-line, and four larger foramina are present at a little distance from the latter. The nasal process of the premaxilla is broken off close to its base, but is seen to ascend postero-dorsally and to possess a concave posterior margin. Ventral to this con- cavity there is a single foramen, directed anteriorly. The lower jaw (PI. 43, figs. 1-3) is represented by paired dentaries and part of the right surangular. The right dentary, which is the more complete, contains thirteen alveoli, but the full complement is estimated to have been seventeen or eighteen. The dentaries are slenderly built by comparison with M. bucklandi and are narrow in cross-section relative to their length. The third alveolus is the largest and prob- ably bore the tooth received by the premaxillary-maxillary groove. The height of WALDMAN: BAJOCIAN MEGALOSAURIDS 329 the buccal wall reaches its maximum at the level of the third and fourth alveoli. Small, poorly preserved interdental plates are present, but the blood-vessel groove running along their bases is well preserved, and considerably deepened anterior to the fifth alveolus. A shallow longitudinal groove is present on the lingual face but is broken away at the level of the third alveolus. Below this groove there is a sharp- edged ridge, slightly convex dorsally, which extends from the level of the posterior margin of alveolus nine to that of alveolus six. At the emergence of the Meckelian canal from the dentary the ridge extends into the vacuity as a V-shaped fork, prob- ably for articulation with the anterior part of the splenial. The ventral prong of the fork extends posteriorly, although partially broken, and has longitudinal rugosities on the postero-dorsal surface. Five adult teeth are present in alternate alveoli beginning with the second, and all these alveoli bear successional teeth except alveolus ten. Replacement teeth are also present in alveoli three and five, the latter tooth being a hollow shell. In common with other carnosaurs all the adult teeth have hollowed out lingual faces where the successional teeth are already appearing (Lambe 1917). Crypts are present on the lingual walls of the alveoli where the successional teeth develop before erupting into the alveoli. Alveolus ten is an excellent example of such a crypt, being laid open to view from the buccal face. The left dentary is damaged, lacking much of the anterior and posterior regions and most of the buccal surface. The beginning of the forked articulation with the splenial is just visible and the interdental blood-vessel groove is well preserved. This groove deepens at the same point in each dentary, at the level of the centre of alveolus four. A similar situation was noted by Osborn (1912) in the dentary of Tyrannosaurus and may be a mechanism to ensure vascular protection near the front of the jaw. The alveoli show little of note, except that alveolus nine possesses three teeth in vertical succession, the smallest of which exhibits a remarkably well-preserved posterior carina, visible through a break in the posterior wall of the alveolus. The anterior wing of the right surangular (PI. 42, figs. 6-8) is a thin, lamellar ele- ment, but above the deep longitudinal groove for articulation with the posterior portion of the dentary it is somewhat thickened. The ventral margin of the sur- angular is sinuous but the dorsal rim is almost straight and the tip of the very thin anterior process to the dentary is broken off. No foramina of any kind are present in this wing of the surangular. During preparation of the block containing the right dentary, left premaxilla and right surangular, a portion of a median skull element was recovered (PI. 43, figs. 4-6). The bone is bilaterally symmetrical, laterally com- pressed, and is apparently the product of two fused elements. Two broad, thin flanges form a deep cleft, the flanges tapering rapidly into a flattened bar, with rem- nants of the cleft forming a long groove on one margin. The opposite edge of the bar is broken, but probably only a little more bone was present originally. Two feasible suggestions have been made concerning this element. The late Pro- fessor A. S. Romer and Dr. A. D. Walker (separate personal communications) have suggested the possibility of it being a part of the presphenoid, whereas Dr. D. A. Russell (personal communication) has proposed it to be a broken vomer. Neither of these elements is particularly well known in carnosaurs, but there is a resemblance to a vomer of a large tyrannosaurid in the N.M.C. collection {Dasplelosaurus torosus\ 330 PALAEONTOLOGY, VOLUME 17 N.M.C. 8506). In lateral view the specimen resembles the parasphenoid rostrum of Dromaeosaurus (Colbert and Russell 1969) but the very deeply cleft structure of the M. hesperis element is unlike that of Dromaeosaurus. Of the two suggestions put forward it seems more likely that the element is part of the vomer. Remarks. The material described must be specifically separated from M. bucklandi, as the tooth-count in both upper and lower jaws is widely different from that of the latter. Further comparisons are obviously impossible at the moment due to lack of material. From Table 1 it is evident that Allosaurus, Proceratosaurus, and ‘Zan- clodou' cambrensis closely approach M. hesperis in numbers of teeth. The form and disposition of the dentition and general jaw morphology of Proceratosaurus (Wood- ward 1910; von Huene 1926a, b) are sufficiently distinct, however, to preclude it from close relationship with either M. hesperis or Allosaurus. TABLE 1. A comparison of tooth counts in some theropods. No. of No. of No. of premaxillary teeth maxillary teeth dentary teeth Megalosaurus hesperis 4 + 15-18 17-18 Allosaurus 5 15-17 15-16 Megalosaurus bucklandi ? 12-13 12-13 Bust rep tospondylus 4 9 + ?13 Tyrannosaurus 4 12 14 Ceratosaurus 3 15 15 Albertosaurus (sensu Russell 1970) 4 13-15 14-16 Proceratosaurus 4-?5 18 ?18 ‘Zanclodon cambrensis ? 7 16-717 In occlusal view the dentaries of Allosaurus exhibit a marked curvature, becoming strongly convex toward the symphysis. As if to accentuate this curvature the anterior Carina of the anterior teeth is placed lingually and the posterior carina is displaced slightly towards the buccal face. The dentaries of M. hesperis show no indication of such longitudinal curvature, apart from a slight convexity to house the large second and third alveoli ; and the tooth carinae remain on the anterior and posterior margins. Although many carnosaurs do not possess ‘a definite symphysial area’ (Walker 1964) specimens of Allosaurus from the Upper Jurassic Morrison Formation, Cleveland (Quarry, Utah, U.S.A., in the N.M.C. collections have a definite facet at the sym- physis. This would have formed a very fiexible and elastic type of articulation, and is interesting in view of Gilmore’s (1920) statement that there is ‘. . . no indica- tion of a symphysial surface . . .’ at the anterior margin of the dentary. No such facet is present in the mandible of M. hesperis. EXPLANATION OF PLATE 43 Figs. 1, 2, 3. Lingual, buccal, and occlusal views of the right dentary of M. hesperis. B.M.(N.H.) R.332, xO-5. Figs. 4, 5, 6. Dorsal, lateral, and ventral views of the vomer of M. hesperis. B.M.(N.H.) R.332. Natural size. Numbers indicate dentary teeth, numbering from the anterior. PLATE 43 WALDMAN, Bajocian megalosaurids 332 PALAEONTOLOGY, VOLUME 17 TABLE 2. Comparison of certain features in the dentaries of some theropods. Abbreviations: a.c.. Anterior carina; p.c., Posterior carina. Shape of dentary in occlusal view Symphysial facet Position of tooth carinae Megalosaurus hesperis Straight Absent Posterior and anterior Allosaurus Well curved Present a.c. moved lingually on anterior teeth ; slight buccal displacement of p.c. Albertosaurus Straight Absent Posterior and anterior; ventral part of a.c. deflected lingually Megalosaurus bucklandi Straight Absent Posterior and anterior Megalosaurus Straight Absent Posterior and anterior nethercombensis Table 2 illustrates the apparently aberrant character of the lower jaw of Allosaums when compared with other theropods. Such comparison is limited to those genera which I have examined, and, therefore, Ceratosaurus nasicornis is omitted. It may be pertinent to note, however, that to judge from Gilmore (1920, pi. 17, fig. 2) the left dentary of Ceratosaurus is curved toward the symphysis in similar fashion to that of Allosaurus. Gilmore (1920) also stated that he could not distinguish between the teeth of Ceratosaurus and Allosaurus, and referred to ‘. . . the same placing of the carina . . .’. It may well be, therefore, that Ceratosaurus and Allosaurus are the only two large carnivorous dinosaurs from either North America or Europe which possessed curved dentaries and anterior teeth with lingual anterior carinae. The part of the surangular preserved in M. hesperis bears a close resemblance to that of Allosaurus (Gilmore 1920) and a tyrannosaurid, Daspletosaurus torosus (Russell 1970). The small foramen present in the mid-anterior part of the surangular of Allosaurus is not present in M. hesperis. Despite the large number of teeth in each genus I do not believe there to be a close relationship between M. hesperis and Allosaurus. I base this statement on the great mandibular curvature and tooth carina positions of the latter, which seem to deny relationship with any known British carnosaur. Although the systematics of Triassic ‘carnosaurs’ are in a state of confusion due to incomplete and possibly incorrectly associated material (for clarification see Walker 1964; Charig, Attridge and Crompton 1965) one particular specimen may be of relevance to the present study. This is the left dentary from the Rhaetic of Bridgend, Glamorganshire, South Wales, described as Zanclodon cambrensis by Newton (1899) and referred to " Plateosaurus' cloacinus Plieninger by von Huene (1908). As Walker (1964) pointed out, von Huene (1908) stated that the name Zan- clodon must be reserved for the original specimen described by Plieninger (1846), which lacked serrations on the teeth. The teeth of ‘Z.’ cambrensis are very different from those of other species of Plateosaurus and it is unlikely to belong to the latter genus. While it is not attempted to classify ‘Z.’ cambrensis generically, it should be noted that it bears a considerable resemblance to the lower jaw of M. hesperis. The jaw of ‘Z.’ cambrensis is straight and exhibits no true symphysis, although Newton (1899) mentioned ‘a slight flattening of the front part of the inner surface’. Newton (1899) stated that the number of alveoli was sixteen or seventeen (von Huene (1908) thought that there were sixteen) which is similar to that of M. hesperis. The teeth WALDMAN: BAJOCIAN MEGALOSAURIDS 333 appear slightly more blade-like, are a little longer-based than those of M. hesperis, and seem to lack the pillar-like appearance of the lower part of the teeth of the latter genus, although this may be due to differing degrees of tooth emergence. It seems that the teeth lie in line with the longitudinal axis of the jaw and that the carinae are positioned anteriorly and posteriorly as in megalosaurids and tyranno- saurids, but not as in Allosaurus. The resemblance between the dentaries of M. hesperis and ‘Z.’ cambrensis is marked, the only disparity, apart from size, being a very minor difference in tooth- shape. While such a resemblance based on a single, evolutionarily adaptable element is tenuous, it may not be dismissed as being entirely without significance. As skeletal elements closely similar to those of Megalosaurus bucklandi are now known from the Lower Lias (Lower Jurassic) of England (Newman 1968) it is reasonable to ex- pect the ancestors of similar species such as M. hesperis to be found in beds of similar or slightly greater age. It is not intended to enter upon the problem of whether a separate family is justified to include Allosaurus and its allies, or whether these genera should be placed in the Megalosauridae, but the unusual shape of the lower jaw and high number of teeth of that genus may be regarded by some as evidence in support of a separate family Allosauridae as proposed by von Huene (1932). M. hesperis does not appear to be closely related to Allosaurus. Walker (1964) gave a brief resume of the characters of the family Megalosauridae using Allosaurus as typical of that family. In the light of the present description it would seem that Allosaurus is aberrant, at least with regard to mandibular form and dentition and is, therefore, unsuitable for use in the diagnosis of the family Megalo- sauridae. Megalosaurus nethercombensis von Huene, 1923 Plate 44 1923 ''Megalosaurus' nethercombensis p. 450. 1926a Megalosaurus (subgen. b) nethercombensis von Huene; von Huene, p. 72. 19266 Megalosaurus nethercombensis von Huene; von Huene, p. 477. 1932 Magnosaurus nethercombensis von Huene; von Huene, p. 220. 1934 Megalosaurus nethercombensis von Huene; Swinton, pp. 214-215. 1959 Megalosaurus nethercombensis von Huene; Delair, p. 78. Diagnosis. A small megalosaurid. Dentary with curved dorsal and ventral margins. Twelve or thirteen dentary teeth, pillar-like in lateral view, recurved occlusally, both carinae serrated. Slender tibia, transverse diameter of head two-thirds of longi- tudinal diameter; cnemial crest projects anteriorly, small crista lateralis, astragalus with well-developed ascending process. Pubis rod-like, proximal median lamella. (Amended from von Huene (1962a, b) and Delair (1959).) Holotype. Oxford University Museum, J. 12143, J. 12143/1, paired dentaries; J. 12143/2, left tibia and fibula fragment ; J. 12143/3, right pubis; J. 12143/6/7, paired femora; J. 12143/8, caudal vertebra; J. 12143/9b, dorsal vertebra. Note. This material was collected by (or for) Mr. J. Parker of Oxford and described by von Huene (1923, 1926a, 6, and 1932). Von Huene had seen neither the femora nor one of the vertebrae, and the dentaries were 334 PALAEONTOLOGY, VOLUME 17 unprepared. In 1932 von Huene based a new genus Magnosawus on this specimen, but Swinton (1934) and Delair (1959) regarded the material as belonging to Megalosaums. Locality. Nethercomb, 1 mile north of Sherborne, Dorset. Horizon. Middle Inferior Oolite, Stephanoceras humphriesiannm zone and subzone; Bajocian, Middle Jurassic. (Stratigraphy according to Wilson et ai, 1958.) Description. The dentaries (PI. 44, figs. 1, 2, and 6), of which the right is the more completely preserved, give an impression of slenderness which is accentuated by their small size. The ventral margin of each is convex anteriorly, curving in a con- cave arc between the levels of the fifth and tenth alveoli. The dorsal margin follows a similar pattern giving the dentaries a ‘waisted’ appearance in lateral view. A shal- low groove runs anteriorly along the lingual face of the dentary from the splenial- dentary articulation to the anterior margin of the jaw. It traverses two small foramina at the level of the third and fourth alveoli, and beyond these becomes wider and deeper, bearing narrow longitudinal ridges. Apart from this structure the lingual wall of the dentary is smooth and flat. No true symphysis is present but minor rugosites occur lingually in front of the first alveolus, probably for membranous connection. On the buccal face of the dentary (PI. 44, fig. 2) a longitudinal groove runs anteriorly to the level of the sixth alveolus, 14 mm below the dorsal margin of the jaw. Anterior to this point the groove is replaced by a row of foramina which parallel the dorsal curve of the jaw and continue on to the ventral margin in the double row, ending in a deep, narrow foramen at the level of the fifth alveolus. There are twenty- three of these foramina on the buccal and ventral portions of the dentary, all of which measure about 2 mm in diameter. Two further foramina, spaced 29 mm apart, continue a slowly ascending trend posteriorly. The interdental plates are fairly well preserved and exhibit a longitudinal blood-vessel groove along their bases. The maximum thickness of the jaw occurs at the level of alveoli two and three, and from that point rearwards the buccal face is shallowly concave, the maximum arc being opposite alveolus seven. The thickness of the dentary increases beneath the longi- tudinal groove on the buccal face of the dentary. Eleven alveoli are known in the right dentary and by comparison with dentaries of Allosaurus, Megalosaums bucklandi, and some tyrannosaurids it seems likely that a further one or two teeth may have been present posteriorly, bringing the total to twelve or thirteen teeth. The teeth are present for the most part as broken stumps, although several of the replacement teeth are beautifully preserved. The teeth differ EXPLANATION OF PLATE 44 Figs. 1, 2, 6. Lingual, buccal, and occlusal views of the right dentary of Megalo.taurus netherconihensis. O.U.M. J. 12143/1, xO-5. Figs. 3, 7. Sagittal and lateral views of an anterior caudal vertebra of M. nethercombensi.s. O.U.M. J. 12143/8, xO-5. Figs. 4, 5. Paired femora of M. nethercomhensis. The femoral shafts are represented by internal calcareous cores. O.U.M. .1.12 143/6/7, xO-25. Figs. 8, 9. Sagittal and lateral views ofa posterior dorsal vertebra of M. nethercomhensis. O.U.M. J. 12 143/9b, xO-5. PLATE 44 WALDMAN, Bajocian megalosaurids 336 PALAEONTOLOGY, VOLUME 17 from those of M. bucklandi in their mode of implantation in the jaws. They stand vertically with the anterior and posterior margins remaining perpendicular to the long axis of the jaw for some way before assuming the usual recurved carnosaur shape. Von Huene’s (1926a, b) description of the teeth is misleading, stating that the teeth are thicker towards their anterior margins than those of M. bucklandi, and, by implication, that no anterior serrations are present. Now that the material has been prepared it is evident that these distinctions are invalid. Firstly, there is con- siderable variation in thickness of the anterior part of the tooth in both species, and secondly, both carinae are clearly serrated where preservation has allowed. This is visible in alveolus three of the right dentary (PI. 44, fig. 1), and in the replacement teeth. Successional teeth are visible in crypts on the lingual margin of many alveoli, and the tracing of a radiograph (text-fig. 1) shows the position of old tooth shells. TEXT-FIG. 1. A drawing made from a radiograph of the right dentary of Megalosaurus nethercombensis, O.U.M. J. 12143/1, to show the presence of successional teeth erupting through the shells of older teeth, xO-83. mature teeth, and replacement dentition. The first alveolus of the right dentary bears a tooth with a grossly curved base, which encroaches upon the second alveolus. This is definitely a growth phenomenon rather than a post-mortem distortion. The dentaries have been arranged in what is believed to be their original relation- ship (PI. 44, fig. 6). The result is a narrow jaw, comparable in symphysial angle to that of some tyrannosaurids, and probably also to British forms such as M. bucklandi and M. hesperis, but completely different from the lower jaws of Allosaurus. The postcranial material is very poorly preserved, and von Huene (1926a, b) has described and figured the left tibia-fibula (J. 1234/2) and the right pubis fragment (J. 1234/3). The femora (J. 12343/6/7; PI. 44, figs. 4, 5) are severely damaged, and represented mainly by cavity casts. Little may be said of their morphology, but they are evidently an associated pair. A single anterior caudal vertebra, J. 12343/8 is preserved (PI. 44, figs. 3, 7) but is missing the neural arch and one face of the centrum. The centrum is slightly oval dorso-ventrally in cross-section and has a facet at the postero-ventral margin for chevron articulation. Well-marked rugae are present on the ventral surface of the centrum close to the chevron facet. The second vertebra (J.12343/9b; PI. 44, figs. 8, 9) is more massive than the caudal and its centrum is almost circular in cross-section. The curvature of the centrum WALDMAN: BAJOCIAN MEGALOSAURIDS 337 toward the one remaining articular face is more marked than in the caudal vertebra, and the ventral rugae are more robust. The articular face is weakly concave as in the caudal vertebra, and only crushed fragments of the neural arch remain. This is probably a dorsal vertebra from a position immediately anterior to the sacrum. Remarks. It is difficult to comment upon the affiliation of M. nethercombensis from this fragmentary material, particularly as it may belong to a juvenile individual. M. nethercombensis is distinct from M. hesperis on the basis of the number of dentary teeth, even though there are certain resemblances in the form of the teeth and the shape of the dentary. There would appear to be no obstacle to the inclusion of the Nethercomb material (J. 12343) within the genus Megalosaurus, as suggested by von Huene (1923, 1926a, b), Swinton (1934), and Delair (1959). With regard to other so-called species within von Huene’s genus Magnosaurus (1932) the following points may be noted. Magnosaurus(l) lydekkeri von Huene (1932, p. 220) was based on a single tooth ascribed to Megalosaurus by Dawkins (in Huxley 1869) to Zanclodonl sp. by Lydekker (1888) and to "Zanclodoniiy by von Huene (1926a). It may be best regarded as belonging to a carnivorous dinosaur of indeterminable affinity. In 1908 Woodward described a tibia, B.M.(N.H.) R.3542, from the Lower Lias of Wilmcote, Warwickshire, as a megalosaurid. Von Huene (1908) referred the bone to Megalosaurus, and Andrews (1921) in a description of Sarcosaurus woodi mentioned that the Wilmcote tibia was probably referable to Sarcosaurus, even though no such element existed in his type material. Von Huene (1926a) referred the tibia to ‘'Megalosaurus'’ (subgen. a) sp. and in 1932 made the same bone the type of two new species in separate genera, Sarcosaurus andrewsi (1932, p. 51) and Magnosaurus woodwardi (1932, p. 219). There is no evidence to support the inclusion of the Wilmcote tibia within the genus Sarcosaurus, but there is a close resemblance between the former and the tibia of Megalosaurus nethercombensis, as noted by von Huene (1926a). I should prefer to transfer the species S. andrewsi to the genus Megalosaurus. Acknowledgements. I am indebted to the following for their advice and help; Dr. R, J. G. Savage and the late Professor W. F. Whittard (University of Bristol); Dr. E. I. White and Dr. A. J. Charig (B.M.N.H.); Dr. A. D. Walker (University of Newcastle upon Tyne); Dr. D. A. Russell (N.M.C.); and Mr. H. P. Powell (O.U.M.). I wish to thank Mr. E. W. Seavill, Mr. R. J. Godwin (University of Bristol), and the late Mr. J. Allan Cash for the photography, and Miss D. Ovenden (University of Bristol) for the radiography. I am especially grateful to the educational trusts which supported this research in 1964-1965; the C. H. Foyle Trust, the Gilbert Foyle Trust, and the Sir Richard Stapley Trust. Further work was undertaken during tenure of a National Research Council of Canada Post-Doctorate Fellowship from 1968-1970. REFERENCES ANDREWS, c. w. 1921. On some remains of a theropodous dinosaur from the Lower Lias of Barrow-on- Soar. Ann. Mag. nat. Hist. (9), 8, 570-576. BUCKLAND, w. 1824. Notice on the Megalosaurus, or great fossil lizard of Stonesfield. Trans. Geol. Soc. Loud. (2), 1, 390-396. BUCKMAN, s. s. 1893. The Bajocian of the Sherborne district; its relation to subjacent and superjacent strata. Q. Jl geol. Soc. Land. 49, 479-522. 338 PALAEONTOLOGY, VOLUME 17 CHARiG, A. J., ATTRiDGE, J. and CROMPTON, A. w. 1965. On the origin of the sauropods and the classifica- tion of the Saurischia. Proc. Linn. Soc. Loud. 176, 197-221. COLBERT, E. H. 1964. Relationships of the saurischian dinosaurs. Am. Mas. Novit. 2181, 1-24. and RUSSELL, d. a. 1969. The small Cretaceous dinosaur Dromaeosaurus. Ibid. 2380, 1-49. DELAIR, J. B. 1958. The Mesozoic reptiles of Dorset. Part 1. Proc. Dorset not. Hist, archaeol. Soc. for 1957, 79, 47-72. 1959. Op. cit. Part 2. Ibid, for 1958, 80, 52-90. 1960. Op. cit. Part. 3. Ibid, for 1959, 81, 57-85. 1966. New records of dinosaurs and other fossil reptiles from Dorset. Ibid, for 1965, 87, 57-66. EDMUND, A. G. 1960. Tooth replacement phenomena in the lower vertebrates. Contr. Life Sci. Div. R. Ont. Mus. 52, 1-190. GILMORE, c. w. 1920. Osteology of the carnivorous Dinosauria in the United States National Museum, with special reference to the genera Antrodemus (Allosaurus) and Ceratosaiirus. Bull. U.S. natn. Mus. no, 1-159. HUENE, E. VON. 1907-1908. Die Dinosaurier der europaischen Triasformation, mit Beriicksichtigung der aussereuropaischen Vorkmmnisse. Geol. paldont. Abh., Supplement-Band 1, i-xii, 1-419 pp. 1923. Carnivorous Saurischia in Europe since the Triassic. Bull. geol. Soc. Am. 34, 449-458. 1926a. The carnivorous Saurischia in the Jura and Cretaceous formations, principally in Europe. Reva Mus. La Plata, 29, 35-167. 19266. On several known and unknown reptiles of the Order Saurischia from England and Erance. Ann. Mag. nat. Hist. (9), 17, 473-489. 1932. Die fossile Reptil-Ordnung Saurischia, ihre Entwicklung und Geschichte. Monogr. Geol. Paldont. (1), 4, 1-361. HUXLEY, T. H. 1869. On the upper jaw of Megalosaurus. Q. Jl geol. Soc. Loud. 25, 311-314. LAMBE, L. M. 1917. The Cretaceous theropodous dinosaur Gorgosaurus. Mem. geol. Surv. Can. 100, 1-84. LYDEKKER, R. 1888. Catalogue of the fossil Reptilia and Amphibia in the British Museum {Natural History). Part 1. British Museum (Nat. Hist.), London. MANSEL-PLEYDELL, J. c. 1888. Fossil Reptiles of Dorset. Proc. Dorset nat. Hist, antiq. Fid Club, 9, 1-40-f appendix. MATTHEW, w. D. and BROWN, B. 1922. The family Deinodontidae, with notice of a new genus from the Cretaceous of Alberta. Bull. Am. Mus. nat. Hist. 46, 367-385. MEYER, H. VON. 1832. Palaeologica zur Geschichte der Erde und ihrer Geschopfe. Frankfurt am Main. NEWMAN, B. H. 1968. The Jurassic dinosaur Scelidosaurus harrisoni, Owen. Palaeontology, 11, 40-43. NEWTON, E. T. 1899. On the megalosauroid jaw from Rhaetic beds near Bridgend (Glamorganshire). Q. Jl geol. Soc. Lond. 55, 89-96. OSBORN, H. F. 1912. Crania of Tyrannosaurus and Allosaurus. Mem. Am. Mus. nat. Hist. (N.S.), 1, 1-30. OWEN, R. 1883. On the skull of Megalosaurus. Q. Jl geol. Soc. Lond. 39, 334-347. 1884. A history of British fossil reptiles. Cassell & Co. Ltd., London, 3, vii-|- 199 pp. PLIENINGER, T. 1846. Ueber ein neues Sauriergenus und die Einreihung der Saurier mit flachen, schnei- denden Zahnen in eine Familie. Jh. Ver. vateri. Naturk. Wurttemb. 2, 148-154 and 247-254. RICHARDSON, L. 1916. On the stratigraphical distribution of the Inferior Oolite vertebrates of the Cotteswold Hills and the Bath-Burton Bradstock District. Proc. Dorset nat. Hist, antiq. Fid Club, 37, 48-55. 1932. The Inferior Oolite and contiguous deposits of the Sherborne district, Dorset. Proc. Cottes- wold Nat. Fid Club, 24, 35-85. RUSSELL, D. A. 1970. Tyrannosaurs from the late Cretaceous of western Canada. Natl Mus. Can., Publ. Palaeontol. 1, viii-t- 1-34. swiNTON, w. E. 1934. The Dinosaurs. Thomas Murby & Co., London, xii-l-233 pp. WALKER, A. D. 1964. Triassic reptiles from the Elgin area: Ornithosuchus and the origin of carnosaurs. Phil. Trans. R. Soc. (B), 248, 744, 53-134. WILSON, V., WELCH, F. B. A., ROBBIE, J. A. and GREEN, G. w. 1958. Geology of the country around Bridport and Yeovil. (Explanation of Sheets 327 and 312.) Mem. geol. Surv. U.K. H.M.S.O. Lond. WOODWARD, A. s. 1908. Note on a megalosaurian tibia from the Lower Lias of Wilmcote, Warwickshire. Ann. Mag. nat. Hist. (8), 1, 257-265. WALDMAN: BAJOCIAN MEGALOSAURIDS 339 WOODWARD, A. s. 1910. On a skull of Megalosaurusirova the Great Oolite of Minchinhampton (Gloucester- shire). Q. Jl geol. Soc. Lond. 66, 111-115. and SHERBORN, c. D. 1890. A Catalogue of British Fossil Vertebrata. Dulau & Co., London. xxxv-|- 396 pp. Typescript received 11 January 1973 MICHAEL WALDMAN Stowe School Buckingham V, '4 1 lih ../. ' ■' ■ ' , i> 'Ml-*- V- . ■ ■ h .<«»< \:-,.f ■■ •- ■ -•> ■ ! 'S • V . . I '• **:, ill''' ■ .' .r LOWER CRETACEOUS SCLEROSPONGE FROM THE SLOVAKIAN TATRA MOUNTAINS by j6zEF KAi;MIERCZAK Abstract. The first fossil sclerosponge with well-preserved spicules is described as Murania lefeldi gen. et sp. nov. It comes from the Lower Cretaceous (Aptian) Muran Limestone of the Slovakian Tatra Mountains. The relation- ship of this sponge with recent sclerosponges is discussed and the affinities of the sclerosponges with the Palaeozoic stromatoporoids are briefly reviewed. The first fossil sclerosponges to be recognized as such have been found in thin sections of limestones donated to me by Dr. J. Lefeld (Institute of Geology, Polish Academy of Sciences). The samples come from the northern slope of Muran Moun- tain in the Lower Cretaceous Muran Limestone of the Sub-Tatric Succession of the Slovakian Tatra Mountains. The geological details and precise locality are given in Uhlig (1899) and Andrusov (1959). The most recent biostratigraphical investiga- tions on the Muran Limestone (Lefeld 1974) indicate an Upper Hauterivian to Lower Aptian age; the association of Orbitolina lenticularis Blumenbach with the sclerosponges is a good indication of an Aptian age for these specimens. Litho- logically the Muran Limestone comprises sparry calcirudites and coarse calcarenites. The sponges, together with a rich fauna, including green and red algae, (?)hydro- zoans, echinodermal and shell debris occur as strongly abraded clasts. The clasts were probably derived from the Aptian reefs (Urgonian facies) of the nearby High Tatric zone (Lefeld 1968). SYSTEMATIC DESCRIPTION Class SCLEROSPONGIAE Hartman and Goreau, 1970 Genus Murania gen. nov. Type species. Murania lefeldi sp. nov. Diagnosis. Encrusting sclerosponge with massive calcareous basal skeleton formed of closely spaced columns normal to the lower surface with irregularly polygonal cross-sections. The axial part of each column is occupied by upward radiating pri- marily siliceous spicules (styles). The spicules are embedded in amorphous carbonate. The outer part of each column is formed of fibrous calcite. The upper surface of the skeleton bears short, delieate calcareous processes. Derivation of name. From Muran Mountain (Slovakian Tatra Mts.). Murania lefeldi sp. nov. Plates 45, 46; text-fig. 1 Holotype. Thin sections Z.Pal.Pf.I/la, b; PI. 45, figs. 1-2; PI. 46, figs. 1-7. Paratype. Thin section Z.Pal.Pf.I/2. Derivation of name. The species is named in honour of Dr. Jerzy Lefeld who collected the material. [Palaeontology, Vol. 17, Part 2, 1974, pp. 341-347, pis, 45-46.] 342 PALAEONTOLOGY, VOLUME 17 Diagnosis. A species of Murania with about half the breadth of the columnar skeletal units occupied by style (?acanthostyle) spicules radiating upwards with mean length of 200 |(xm and mean width (at the head) of 30 /xm. Description. Sheet-like, encrusting, calcareous skeleton usually less than 4 mm thick, associated with the calcareous alga Lithocodium aggregatum Elliott (Codia- ceae). This alga encrusts and is encrusted by the sponge (PI. 45, figs. 1-2; PI. 46, figs. 1-3, 5; text-fig. 1). The rapid postmortal encrustation of the sponge skeleton by L. aggregatum was probably the main factor responsible for the preservation of the sponge surface in almost its original form. The skeleton is composed of distinct closely spaced columns with irregular polygonal cross-sections (PI. 46, fig. 7 ; text- fig. 1b). The mean column thickness is 300 |U,m (min. 250 ^m, max. 500 jixm). The boundaries between columns are usually distinct, and are sometimes separated by a narrow zone frequently filled with sediment or algal carbonate (text-fig. 1). Each column is composed of two zones: (1) the central spicular zone, and (2) the outer fibrous zone. The spicular zone occupies about half the column breadth and is composed of style type spicules which radiate upwards, more or less regularly from the axis. The length of the spicules is between 150-220 [xm (mean 200 pm); the width at the head is between 28-33 j^m (mean 30 pm). It is difficult to determine whether the spicules are smooth styles or acanthostyles, since only thin sections were avail- able. It is not clear whether the very fine notches visible on some of the spicules (PI. 46, fig. 4) form primary ornamentation or are marks of secondary corrosion on the smooth styles. Where distinct, the spicules are calcite pseudomorphs, pre- sumably after primary opal. However, many spicules have indistinct outlines due to imperfect replacement (PI. 46, fig. 4). The spicules are usually quite closely spaced, often almost touching, especially at their heads. The distribution of the spicules is most evident at the upper surface of the skeleton where, very often, their moulds are filled by encrusting algal carbonate, darker than the surrounding sponge skeleton (PI. 46, figs. 2-3, 5 ; text-fig. 1a). The upper surface of the least-corroded sponge frag- ments bear numerous irregular needle-like processes with an average length of 110 pm and width 12 pm forming a palisade around each spicule (text-fig. 1a). Deeper in the skeleton the spicules are embedded in amorphous and slightly floc- culent carbonate, the dark coloration of which could indicate a considerable ad- mixture of organic matter. The outer zone of each column is formed of densely packed, upward-radiating calcite fibres which are particularly evident in polarized light (PI. 46, fig. 6). Discussion. Murania lefeldi is the first fossil sclerosponge with well-preserved, though pseudomorphic, spicules in the basal calcareous skeleton. It differs from all described recent sclerosponges in that the skeleton is composed of columnar units. EXPLANATION OF PLATE 45 Figs. 1-2. Murania lefeldi gen. et sp. nov., x 15. Longitudinal sections showing sheet-like skeletons of the sponge interlayered and encrusted by codiacean calcareous algae Lithocodium aggregatum Elliott. The cross-section of the serpulid tubes overgrown by the sponge skeleton are visible in Fig. 1 (base and right). Note the columnar units composing the sponge skeleton. 1, thin section Z.Pal.Pf.l/lb. 2, thin section Z.Pal.Pf.I/Ta. PLATE 45 KAZMIERCZAK, Cretaceous sclerosponge 344 PALAEONTOLOGY, VOLUME 17 TEXT-FIG. 1. Murania lefeldi gen. et sp. nov. Diagrammatic details of the skeleton structure in (a) longi- tudinal and (b) transverse section, x 50. a, encrusting codiacean alga Lithocodium aggregation Elliott; sp, spicules (styles); am, central flocculent calcareous zone of columns;/, outer fibrous calcareous zone of columns. Note the calcareous processes on the top of the columns, the dark algal substance infilling the upper- most spicules (a) and the irregular zones (black) between the columns (a and b). A, based on thin section Z.Pal.Pf.I/la; b, based on thin section Z.Pal.Pf.I/lb. and the considerably thicker spicules are arranged in vertical zones. However, the massive encrusting skeleton and the numerous sharp processes surrounding the spicules on the upper surface resemble the condition in recent species of Hispi- dopetra, Stromatospongia, and Goreauiella described by Hartman (1969). The spicules in these recent forms are much thinner and are scattered within a compact skeleton. This skeleton is composed of typical sclerodermites gathered in trabeculae analogous to the fibrous columns of M. lefeldi. The irregularly polygonal cross-sections of the columns in M. lefeldi are similar to the irregularly pentagonal or, less frequently, hexagonal cross-sections of the skeleton elements of another recent sclerosponge, Ceratoporella niclwlsoni (Hickson). The walls of these elements can be considered equivalent to the outer columnar zone of M. lefeldi. However, the walls in C. nichol- soni surround hollow pits and the spicules (acanthostyles) are scattered vertically within the walls (Hickson 1911; Hartman and Goreau 1970), never forming regular EXPLANATION OF PLATE 46 Figs. 1-7. Murania lefeldi gen. et sp. nov. 1, longitudinal section of a sclerosponge upper surface with processes covered by codiacean alga Lithocodium aggregatum, x 40. 2, 3, longitudinal sections through the corroded surface of the sclerosponge showing pseudomorphs after siliceous styles infilled by a darker algal carbonate, x 40. 4, a few calcite pseudomorphs of styles (?acanthostyles) in longitudinal section of a column, x 150. 5, longitudinal section through a surface of a sclerosponge with partially corroded pseudomorphs of spicules infilled by dark algal carbonate, x 70. 6, longitudinal section of three col- lumns in polarized light showing their llocculent central zones and their fibrous fasciculate outer zones, X 40. 7, transverse section through the columnar units showing their irregularly polygonal outlines, x40. Figs. I 6, thin section Z.Pal.Pf.l/la. Fig. 7, thin section Z.Pal.Pf.l/lb. PLATE 46 KAZMIERCZAK, Cretaceous sclerosponge 346 PALAEONTOLOGY, VOLUME 17 clusters. In another recent sclerosponge Merlia normani Kirkpatrick, irregularly polygonal vertical elements are filled with soft tissue within which bunches of siliceous spicules can be found. Unlike M. lefeldi the spicules in M. normani are not incor- porated during growth of the sponge into the calcareous skeleton (Kirkpatrick 1911). Several Mesozoic forms generally described as Hydrozoa or Stromatoporoidea (Sphaeractinoidea) most probably belong to the Sclerospongiae and are similar to M. lefeldi. These are ^ Stromatopora' japoniea Yabe, Dehornella crustans Hudson, Dehornella choffati (Dehorne), Astroporina stellans Hudson, Astroporina orientalis Hudson, Parastromatopora jurensis Schnorf (Yabe and Sugiyama 1935; Hudson 1960; Schnorf 1960) from the Upper Jurassic, and Astroporina valangiensis Schnorf (Schnorf 1960; Hartman and Goreau 1970; Steam 1972) from the Lower Cretaceous. In all these forms, elongated clear areas, which radiate upwards, are visible in the axial zone of the fibrous vertical elements (pillars). These areas are similar in shape and size to the spicules of M. lefeldi and probably represent calcite pseudomorphs after siliceous styles. This assumption is based on the identical preservation of the spicules seen in many parts of the M. lefeldi skeleton. All these species differ from Murania lefeldi in that they have skeletons divided into a system of irregular vertical elements and much thinner horizontal tabules, with rather large interskeletal spaces. If the removal of some Mesozoic ‘Hydrozoa’ to Sclerospongiae seems to be justi- fied on the basis of traces of primary spiculation found in their calcareous skeletons, then the affinity of sclerosponges to the Palaeozoic Stromatoporoidea suggested recently by Hartman and Goreau (1966, 1970) and Steam (1972) is, in my opinion, open to question. The structural and microstructural changes in the evolution of the Palaeozoic Stromatoporoidea (Kazmierczak 1971) can be explained either by the poriferan or the coelenterate affinities of these organisms. In both cases it seems likely that the basally secreted skeleton performed the same supporting function. The main argument against sponge affinities of the stromatoporoids is the absence of spicules in their skeletons. The spicules described by Newell (1935) in the Penn- sylvanian ^ Parallelopora’’ mira Newell appear to be artefacts and, in the opinion of Fliigel and Fliigel-Kahler (1968) and Steam (1972) this form has nothing in common with the stromatoporoids. Even if it were accepted that all Palaeozoic stromatoporoids possessed, like Merlia normani, spicules not incorporated in the calcareous skeleton, then one would expect to find at least some signs of spicules preserved in the interskeletal spaces or in the surrounding sediment. The preserva- tion of distinct spicules in Murania lefeldi in strongly folded Lower Cretaceous sedi- ments contradicts Hartman and Goreau’s (1970) suggestion of quick, even pre-burial corrosion, and dissolution of the spicules. In their opinion this would be responsible for the absence of spicules in Palaeozoic stromatoporoids. The similarities between the exhalant canals of the sponges and the astrorhizae of the Stromatoporoidea which, according to some authors (Rosen 1869; Twitchell 1929; Hartman and Goreau 1966, 1970; Steam 1972) would be the crucial evidence for the poriferan affinities of these fossils, are superficial, and bear only on the ex- ternal appearance of these structures. It seems unlikely that the complex systems of astrorizal canals, always cut by tabulae or dissepiments, could be equivalent to the soft, tube-like exhalant canals of the Sclerospongiae or other sponges. These soft canals rarely leave any traces in the skeleton and those that remain are only very KAZMIERCZAK: CRETACEOUS SCLEROSPONGE 347 superficial impressions. The affinities of the Stromatoporoidea remain a geological enigma. It is likely that the fossil ‘Hydrozoa’ presently including the Palaeozoic Stromatoporoidea (with some marginal forms), Mesozoic Spongiomorphida and Sphaeractinoidea (with several enigmatic forms) are a heterogenic group with coelenterate, sponge (sclerosponge), algal, and other affinities. Acknowledgements. I am grateful to Dr. Jerzy Lefeld, Institute of Geology, Polish Academy Sciences, Warsaw for donating the rock samples containing the sclerosponges and for his comments on the geology of the fossil locality. The specimens are now in the Palaeozoological Institute at Warsaw. I am indebted to Dr. Roland Goldring for discussion and Miss Josephine Lewkowicz, also of Reading University, for helping with the translation. REFERENCES ANDRUSOV, D. 1959. Geologia ceskoslovenskych Karpat. Part 2, Bratislava. FLUGEL, E. and flOgel-kahler, e. 1968. Stromatoporoidea (Hydrozoa palaeozoica). In westphal, f. (ed.), Fossilium Catalogus I: Animalia, 115, 1/2. HARTMAN, w. D. 1969. Ncw genera and species of coralline sponges (Porifera) from Jamaica. Postilla, 137, 1-39. and GOREAU, t. f. 1966. Ceratoporella, a living sponge with stromatoporoid affinities. Amer. Zoologist. 6, 262. 1970. Jamaican coralline sponges: their morphology, ecology and fossil relatives. Svmp. Zool. Soc. Lond. 25, 205-243. HICKSON, s. J. 1911. On Ceratopora, the type of a new family of Alcyonaria. Proc. Rov. Soc. Lond. (B), 84, 195-200. HUDSON, R. G. S. 1960. The Tethyan Jurassic stromatoporoids Stromatoporina, Dehornella and Astroporina. Palaeontology, 2, 180-199. kaZmierczak, j. 1971. Morphogenesis and systematics of the Devonian Stromatoporoidea from the Holy Cross Mountains, Poland. Palaeont. Pol. 26, 1-150. KIRKPATRICK, R. 1911. On Merlia nonnani, a sponge with a siliceous and calcareous skeleton. Q. Jl micro- scop. Sci. 56, 657-702. LEFELD, J. 1968. Stratygrafia i paleogeografia dolnej kredy wierchowej Tatr (Stratigraphy and palaeo- geography of the High-Tatric Lower Cretaceous in the Tatra Mountains). Stadia Geol. Pol. 24, 5-115. 1974. Upper Jurassic and Lower Cretaceous stratigraphy and facies of the Sub-Tatric Succession of the Tatra Mts. Acta Geol. Pol. (in press). NEWELL, N. D. 1935. Some mid-Pennsylvanian invertebrates from Kansas and Oklahoma; II. Stroma- toporoidea, Anthozoa, and Gastropoda. J. Paleont. 9, 341-355. ROSEN, F. B. 1869. Uber die Natur der Stromatoporen und liber die Erhaltung der Hornfaser der Spongien im fossilen Zustand. Verb, russ.-kaiserl. mineral. Ges. St.-Petersh. (2), 4, 1-98. SCHNORF, A. 1960. Parastromatoporoidae nouveau du Jurassique superieur et du Valanginien inferieur du Jura. Eclog. geol. Helv. 53, 729-732. STEARN, c. w. 1972. The relationship of the stromatoporoids to the sclerosponges. Lethaia, 5, 369-388. TwiTCHELL, G. B. 1929. The structure and relationships of the true stromatoporoids. Amer. Midi. Nat. 11, 270-306. UHLiG, v. 1899. Geologie des Tatra Gebirges. Denkschr. Akad. IViss. Mat. -Nat. Kl. 68, 35-36. YABE, H. and SUGIYAMA, T. 1935. Jurassic stromatoporoids from Japan. Sci. Rep. Tohoku Univ. 2nd ser. (Geol.), 14, 135-192. j6zef kaZmierczak Institute of Palaeozoology Polish Academy of Sciences Al. Zwirki i Wigury 93 Typescript received 13 April 1973 02089 Warszawa, Poland K '‘if; ; ‘ft •) , :. ,,• :■» •. • » ::j *■>>-’ s'-t^ 1 Jlt/iS ' ■' '' ' ' 3(^'i ■ .'ll- ■jinf. ' I. ...5-- li- r V I ‘ < i, i ■.'*"' !* ■ l» ■'.. V ■• <■ • - i V ( ^ ' * ■■ * ' '^■i- V II... < TWO NEW SUBSPECIES OF PHACOPS RANA [TRILOBITA] FROM THE MIDDLE DEVONIAN OF NORTH-WEST AFRICA by CHRISTOPHER J. BURTON and NILES ELDREDGE Abstract. Phacops rana africanus subsp. nov. and P. rana tindoufensis subsp. nov. are described from the Middle Devonian of North-West Africa. They are considered to be close to P. ranamilleriStewaTt and P. rana crassituhercidata Stumm of North America. An origin for P. rana in the European and North African P. schloteimi (s.l.) group is postulated. Migration of forms between the two areas is necessary to explain the distribution pattern of P. rana, and routes between North-West Africa and North America are examined with reference to the contemporary posi- tions of the continents. During an investigation into the phacopids of North-West Africa by one of us (C. J. B.), two unusual forms were noted from the Middle Devonian of the Tindouf Basin of Spanish Sahara, northern Mauritania, and Morocco (see Map, text-fig. 1). These forms have been described in literature, and informally identified in collec- tions, as Phacops schlotheimi (s.l.) and P. fecundus degener Barrande. The present authors believe them to belong among the subspecies of P. rana (Green 1832). The two forms are close to P. rana milleri Stewart and P. rana crassituberculata Stumm of the Middle Devonian of North America, and two new subspecies, P. rana africanus and P. rana tindoufensis are erected for their reception. This identification of North-West African forms with a group not hitherto recog- nized outside North America and eastern Asia has raised a number of problems. Firstly, the problem of the origin of P. rana. In view of the close similarity with the P. schlotheimi (s.l.) group, especially in details of eye morphology, the authors believe that the latter and P. rana are closely related. This similarity, coupled with the new discoveries and the fact that there is no close relation to P. rana in the Lower Devonian of the Americas, has led us to explore the possibilities of a relationship between P. rana and European phacopid lineages. Secondly, a possible phacopid migration into North America raises the general problem of routes. The work is based entirely on museum material. French usage is followed for Arabic place names. NOTES ON MORPHOLOGICAL AND OTHER TERMS USED In this paper the terms used to describe the phacopid exoskeleton follow the usage of the Treatise on Invertebrate Palaeontology, Pt. O, Arthropoda 1 (R. C. Moore, ed. 1959), except in the following cases: Eye socle. This is the curb-like ridge supporting the visual surface of the eye, in the sense of Shaw and Ormiston (1964, p. 1002). Group. In this paper the "Phacops schlotheimi group’ refers to subspecies of P. schlotheimi and also to species morphologically close to it and demonstrably separate from other European and African species of the genus Phacops. The group in this sense is an informal device for assembling morphologically similar species and subspecies without any rigid taxonomic commitment (see Burton 1972 for morphological details and discussion). [Palaeontology, Vol. 17, Part 2, 1974, pp. 349-363, pis. 47-48.] 350 PALAEONTOLOGY, VOLUME 17 Intercalating ring. This is used instead of the term ‘preoccipital glabellar lobe’ of the Treatise (p. 125), since in most phacopids this character is a definite ring, not simply an area of the glabella. The intercalating ring corresponds to the ‘Zwischenring’ of Richter (1926, p. 126) and the ‘Anneau Intercalaire’ of Bar- rande (1852, p. 503). The intercalating furrow is that furrow anterior to the ring (glabellar furrow Ip). Interlensar sclera (Clarke 1889). That part of the visual surface which lies between the schizochroal lenses and is not covered by cornea. Palpebral lobe and fixigenal eye stem. The term palpebral lobe is reserved for the area above the visual surface. The term fixigenal eye stem is reserved for the discrete ridge separated from the palpebral lobe by the palpebral furrow and running posteriorly from it. Rear-eye ridge. A term used for the ridge which is a continuation, in a lateral direction parallel to the posterior cephalic margin, of the fixigenal eye stem. The ridge is bounded distally by the facial suture. The term rear-eye ridge is not a synonym of the term ‘postocular ridge’ used in the Treatise (p. 124). Abbreviations. The following abbreviations are used: AMNH— The American Museum of Natural History; BMNH— The British Museum (Natural History), London; ULL— L’Universite Libre de Lille, France; USNM— National Museum of Natural History (formerly United States National Museum); SUI— the State University of Iowa. STRATIGRAPHY AND PALAEOGEOGRAPH Y The material studied comes from parts of northern Spanish Sahara, north-western Mauritania, and southern Morocco (see Map, text-fig. 1). This area is occupied by the Palaeozoic Tindouf Basin, the Middle Devonian of which has been described TEXT-FIG. 1. The geography and Devonian outcrop of the Tindouf Basin and adjacent regions. BURTON AND ELDREDGE: SUBSPECIES OF PHACOPS RANA 351 TEXT-FIG. 2. Fossil localities and Middle Devonian outcrop on the southern flank of the Tindouf Basin. by Arden and Rehrig ( 1 964) and Sougy ( 1 964). Within this succession lies the Wernero- ceras Limestone, the only known source of the new subspecies. This limestone is a constant marker horizon throughout the Tindouf Basin and contains the goniatites Werneroceras crispiforme (Kayser) and Anarcestes {Anarcestes) lateseptatiis (Bey- rich), the presence of which suggests a late Eifelian to earliest Givetian age. The Werneroceras Limestone is subject to lateral facies changes, and varies from a blue- grey limestone to a calcareous marl, the latter called by Sougy (1964) the Marno- calcaires a Phacops fecundus. The main fossil locations lie around Ain Terguet (see Map, text-fig. 2), but to the south a form identical with P. rana tindoufensis has been reported from the same horizon at Aguelt Oudiate et Khyam (Sougy 1964). A further specimen was collected by Le Maitre from the area of Zagora in southern Morocco. No Devonian exists in this latter place but, from the evidence of the southern area, the specimen is likely to belong to the same horizon as the others, and is presumably from the nearest Devonian— that of Tafilalet. 352 PALAEONTOLOGY, VOLUME 17 These trilobite horizons of North-West Africa can be correlated with those of North America containing P. r. milleri and P. r. crassituberculata using the work of House (1962). House has suggested (1962, p. 254) that Werneroceras crispiforme (^ Cabrieroceras crispiforme) of Europe and North Africa is very closely compar- able with the American Werneroceras plebeiforme (Hall), both in morphology and restricted stratigraphical occurrence. The latter occurs in the Werneroeeras Bed at the top of the Union Springs Member of the lowest part of the Marcellus Formation. Therefore on this reasoning P. rana afrieanus and P. r. tindoufensis can be correlated stratigraphically with P. r. milleri and P. r. crassituberculata which first appear in the lower parts of the Marcellus Formation in the states of New York and Ohio. Migration of phacopid trilobites from the Old World to the New World was first hinted at by Hall and Clarke (1888, p. 24) who noted that P. rana appears to be more similar to European species of Phacops than to any other known North American species. Their conclusion has been substantiated by Eldredge (1972) and by the authors further on in this paper. Moreover, Eldredge (1972) has concluded that P. rana was derived from ‘European’ ancestors and was a migrant into the Hamilton (Middle Devonian) fauna of North America. The demonstration of the presence of P. rana in North-West Africa, while not sufficient in itself to indicate directions of migration, nevertheless tends to support this view. Further evidence that such a route was open is provided by Greenops (Greenops) boothi (Green 1837), the only member of the Asteropyginae known to occur in North America. The Asteropyginae are well represented throughout the Devonian of Europe, and in the Lower Eifelian of the Saoura Basin (Le Maitre 1952), the Pragian of Central Morocco (Alberti 1969), and the Lower and Middle Devonian of the Tindouf Basin (Sougy 1964). With a single doubtful exception the earliest occurrence of Greenops is in the Marcel- lus Formation of New York and adjacent states, coincident with the earliest occur- rence of the Hamilton fauna. There can be no question that Greenops is an immigrant trilobite. Pliaeops rana erassituberculata first appears at about the same time, and it is quite likely that the migration histories of the two species were similar. That a migration route existed between North-West Africa and North America is suggested by Sutton’s (1968) work on continental drifting and the proto- Atlantic, in which he suggests a close Devonian fit between Africa and South, Central, and Southern North America. Sougy (1962) has also linked the North-West African Palaeozoic Fold Belt with that of the Appalachians. It also seems likely that a shelf environment persisted right across the proto-Atlantic, since vagrant benthos such as phacopids (Clarkson 1966, p. 82) could only have migrated under such condi- tions. Furthermore, the absence of phacopids in the known Arctic Devonian faunas of North America (Ormiston 1967) taken together with the above arguments strongly supports a direct faunal connection along the route North-West Africa-southern North America-east-central North America. BURTON AND ELDREDGE: SUBSPECIES OE PHACOPS RANA 353 SYSTEMATIC PALAEONTOLOGY Family phacopidae Hawle and Corda, 1847 (nom. correct. Salter 1864 (pro Phacopides Hawle and Corda, 1847)) Subfamily phacopinae Hawle and Corda, 1847 (nom. transl. Reed 1905 (ex Phacopidae Hawle and Corda, 1847)) Genus phacops Emmrich, 1839 Phacops rana {GxQQn, 1832) Emended diagnosis (Eldredge 1972). Eyes large, bearing from 15 to 18 dorso-ventral files of lenses in normal adults. Trace of facial suture over ocular platform shallow. Genal angles gently rounded and near ventral cephalic margin. Glabella furrow Ip deeply incised, glabellar furrows 2p and 3p weakly developed or absent. Cephalon covered by low, rounded tubercles becoming transversely elongate at the anterior margin of the glabella, on the genae, and on the occipital lobe. Tubercles largest on central region of composite glabellar lobe and glabellar lobe Ip. Axis of thorax covered with transversely elongate tubercles. Tuberculation on pleura variably developed. Pygidium with from 7 to 1 1 axial rings and 6 or 7 pleura. Tubercles moderately elongate transversely on axis; tubercles cover pleura, becoming obsolescent on pygidial margin. Interpleural furrows generally obsolescent, anteriormost inter- pleural furrow occasionally present as shallow groove set off by parallel rows of tubercles. Pleural furrows rather shallow, pleura only moderately arched. Phacops rana africanus subspecies nov. Plate 47, figs. 8-9; Plate 48, figs. 1-4 1939 Phacops latiforns Bronn; Le Maitre, p. 203. 1964 Phacops schlotheimi Bronn; Arden and Rehrig, p. 1522. Deriv. nom. Africanus, of Africa, referring to the fact that this is the first subspecies of Phacops rana to be recognized in Africa. Localities. Tifariti area, Spanish Sahara; Gor Loutad, Spanish Sahara; Zagora, Tafilalet, Morocco. Horizon. Werneroceras Limestone, Upper Eifelian-Lower Givetian boundary. Middle Devonian. Material. 10 specimens, sample numbers USNM 174227 and USNM 174072, ‘Ain Terguet formation’ (American field usage) equivalent to Werneroceras Limestone, and Marno-calcaires a Phacops fecundus of Sougy (1964), Gor Loutad, Spanish Sahara; 3 specimens SUI, locality unknown, Spanish Sahara; 7 specimens, sample numbers In 56876-56878, 57166-57167 (Rod collection) and It 5821-5822 (Illing collection) BMNH, Tifariti area, Spanish Sahara; 3 specimens AMNH numbers 29130-29132 from the lowest Givetian (Unit 4b of Arden and Rehrig 1964) of the Gor Loutad region, Spanish Sahara (location at 10° 30' W., 26° 45' N.); 1 specimen ULL, Zagora, Tafilalet, Morocco. (Lor locations see Map, text- fig. 2.) Holotype. USNM number 174072, Werneroceras Limestone, Gor Loutad, Spanish Sahara. Diagnosis. Eyes small and with large eye socle. Intercalating ring very weakly de- veloped. Glabellar tubercles very large, not elongated anteriorly. Ornament of palpe- bral lobe and fixigenal eye stem sparse and large. Occipital ring ornament very sparse or missing. Genal ornament sparse. 354 PALAEONTOLOGY, VOLUME 17 Description. Large trilobites with cephalic lengths (sag.) ranging from 14-8 to 32-4 mm ; cephalic outline roughly semicircular, posterior margin curved only slightly an- teriorly, genal angles smoothly rounded and parabolic. In side view the glabella is inflated and rises vertically, or with a slight anterior bulge, in a smooth arc to a nearly flat summit level. The slightly arched top of the glabella then drops gently down to the intercalating ring, which is low and unobtrusive. The wide occipital ring projects strongly and has a vertical posterior face. The axial furrows diverge at angles varying from 55° to 68° with an average of 63°. The broad glabella is terminated at the rear by a wide intercalating furrow, and its anterior margin is a smooth curve. The 2p and 3p glabellar furrows are present, the latter being visible as a pair of short furrows on the natural cast. The glabellar ornament consists of numerous large hemispherical or flat-topped tubercles which are fairly widely separated at the posterior, but which become smaller and more closely packed anteriorly to form tesselations, on the anterior face of the glabella. The average diameter of the pos- terior tubercles is 1-5 mm for the range of cephalic lengths given. The intercalating ring is always low and narrow, the axial lobe is ornamented with a single large, laterally elongated tubercle, or occasionally two small tubercles. The occipital ring is wide, prominent, and is either smooth or has low randomly distributed central tuberculations. The tubercles when present are transversely elongated. The eyes are large and set high on the genae almost reaching the level of the top of the glabella. The eye socles are wide and steeply inclined. The visual surfaces are nearly vertical with always 18 dorso-ventral files of eye lenses, in these samples the eyes contain 70-80 lenses (Table 1) set flush with strong hexagonal rims of sclera. TABLE 1. Measurements and eye data for P. rana africanus subsp. nov. Wenwroceras Limestone, Givetian, Spanish Sahara. CL— total cephalic length; WBVS— width between visual surfaces; N LENS— total number of lenses on visual surface; # DV— number of dorso-ventral files. Measurements in centimetres. CL WBVS EYE FORMULA EYE N LENS # DV 1.73 _ 345 555 555 454 433 332 L 73 18 2.20 2.75 445 555 555 454 434 332 L 75 18 2.30 3.00 345 455 555 454 433 232 L 71 18 2.3 7 3.17 455 555 555 454 4 34 332 R 76 18 2.40 3. 10 455 555 555 554 433 332 L 76 18 2.47 3.52 455 555 555 554 434 332 L 77 18 2.91 4.46 455 565 655 554 444 332 L 80 18 3.24 4.71 455 565 655 554 444 332 L 80 18 3.50 4.50 455 555 655 554 444 332 L 79 18 3.90 4.90 455 555 555 554 444 332 L 78 18 There is a maximum of 6 lenses per dorso-ventral file. The palpebral lobes are crescent-shaped with large tubercles, and the fixigenal eye stems are rounded and bear a cluster of a few large tubercles on their anterior distal extremities. They are extended into small, low, rear-eye ridges ending against the facial sutures. The palpebral furrows are strongly accentuated. The genae are flexed strongly downwards and are usually smooth or possess at the most a single posterior row of tubercles. The posterior marginal ridge is prolonged into a wide, flat, lateral area slightly elevated above the fixigena. The vincular furrow possesses 7 cusp-like pro- BURTON AND ELDREDGE: SUBSPECIES OF PHACOPS RANA 355 jections on its inner wall at a point below the eye. The hypostome is unknown. The thorax carries little ornament, the axial segments having very low, randomly dis- tributed, central tuberculation together with a posterior row of tubercles. The pleurae are smooth. There is a constriction (ex-sag.) near the distal ends of the axial segments forming incipient nodes. The pygidium has 9-11 axial rings and 6-7 pleurae. It has no ornament. Measurement and eye data: see Table 1. Discussion. Phacops rana africanus is discussed below in conjunction with Phacops rana tindoufensis. Phacops rana tindoufensis subspecies nov. Plate 47, figs. 1-3 1964 Phacops (Phacops) fecimdus degener (Barrande); Sougy, p. 447, pi. 41, figs. 5, 5a. Deriv. nom. Tindoufensis, of the Tindouf Basin. Localities. USNM locality H-23, 4 miles south of the junction of Oued Ratmia and Oued Ain Terguet, about 1 mile west of Oued Ratmia, south-west Tindouf Basin, Spanish Sahara. Also region between Smara and Tifariti, 1 1° 15' W., 26° 39' N. Horizon. Shales interbedded with Werneroceras Limestone, Upper Eifelian-Lower Givetian boundary. Middle Devonian. Material 5 specimens, USNM number 174228, and 1 specimen USNM number 174073, from the shales of the Werneroceras Limestone, USNM locality H-23. 6 specimens, AMNH numbers 29133-29138, from the Smara-Tifariti region, from the same limestone as Phacops rana africanus but at a point where it is less silty. Unit 4b of Arden and Rehrig (1964). Holotype. USNM number 174073, from the Werneroceras Limestone at USNM locality H-23. Diagnosis. Eyes with most lenses protruding beyond interlensar sclera, except for top 2 to 3 lenses of each dorso-ventral file which are flush with sclera. Intercalating ring well developed but narrow. Glabellar tubercles elongated transversely close to anterior glabellar margin in some cases, otherwise merely flattened. Whole of dorsal exoskeleton richly ornamented. Description. Small trilobites with cephalic lengths (sag.) ranging from 9 0 to 17 0 mm; cephalic outline slightly wider than semicircular, posterior margin curved only slightly anteriorly, genal angles smoothly rounded but never parabolic. In side view the glabella is scarcely at all inflated and rises vertically, or slightly less than verti- cally, to a flat summit level. This flat surface then drops gently down to a pronounced intercalating ring. There is a wide occipital ring. The axial furrows diverge at angles approaching 65°. On the glabella 2p and 3p glabellar furrows are visible, but only faintly impressed. The glabellar ornament consists of numerous small to medium- sized rounded to conical tubercles, evenly distributed posteriorly, becoming smaller and more closely packed anteriorly, but never forming tesselations. In some cases those tubercles closest to the anterior glabellar margin become flattened and elon- gated transversely (Sougy 1964, pi. 41, fig. 5a). The average diameter of the posterior tubercles is 0-7 mm. The intercalating ring is pronounced but narrow, the axial lobe being ornamented with one or two equidimensional tubercles. The occipital ring 356 PALAEONTOLOGY, VOLUME 17 is wide and prominent and has numerous randomly distributed tubercles, all of which are flattened and elongated transversely. The eyes are rather short compared with the length of the cephalon. The eye socles are narrow and insignificant. The visual surfaces are tall and slightly less than vertical with 18 dorso-ventral files, a maximum of 9 lenses per dorso-ventral file, and an average of 109 lenses per eye (Table 2). The lenses are closely spaced and protrude beyond the interlensar sclera TABLE 2. Measurements and eye data for P. rana tindoufensis subsp. nov. Werneroceras Limestone, Givetian, Spanish Sahara. Abbreviations as in Table 1 (above). CL WBVS EYE FORMULA EYE N LENS # DV 1.70 2.00 567 787 877 776 655 432 R 107 1 8 0.90 1.10 566 676 767 665 655 332 L 97 1 8 1.35 1.65 567 777 777 676 655 442 L 105 1 8 1.00 1.20 678 898 888 877 767 543 R 1 24 1 8 1.2 0 1.40 568 787 878 776 655 442 R no 1 8 near the bottom of the dorso-ventral files, and are more widely spaced and flush with the sclera near the tops of the dorso-ventral files. The palpebral lobes are crescent-shaped with small tubercles, and the fixigenal eye stems are flattened and bear numerous, evenly distributed, small tubercles. They are extended into small, low rear-eye ridges ending against the facial sutures. The palpebral furrows are moderately accentuated. The genae are not strongly flexed downwards and are abundantly ornamented with a single row of posterior tubercles, and small, ran- domly distributed, tubercles on the rear halves of the genae, these tubercles fading out anteriorly. The posterior marginal ridge is weak and does not persist laterally beyond the rear of the eye. The hypostome is unknown. Flattened tubercles, elongated transversely, cover the axial rings of the thorax and pygidium. The posterior ramus of the pleura in the thorax and pygidium is covered densely with low, rounded tubercles. There is a constriction (ex-sag.) near the distal ends of the axial rings of the thorax forming incipient nodes. The pygidium has 9 or 10 axial rings and a terminal piece, and 7 pleurae. 1 pair of interpleural furrows are present. Measurements and eye data: see Table 2. Comparisons. The pair Phacops rana tindoufensis and P. r. africamis are close to the American pair P. r. milleri and P. r. crassituherculata, these 4 also forming a small complex distinct from all other subspecies of P. rana. Moreover, within this com- EXPLANATION OF PLATE 47 Figs. 1-3. Phacops rana tindoufensis suhsx>- nov. Holotype, LfSNM 174073. All x 2. Givetian, USNM loc. H-23, 1 1° 15' W., 26'’ 39' N., Spanish Sahara. Figs. 4-5. Phacops rana crassituherculata Stumm, AMNH 28898. 4, x2. 5, x 3. Lower Cazenovian, Ohio, U.S.A. Figs. 6-7. w///er/ Stewart, AMNH 28896. 6, x2. 7, x 3. Lower Cazenovian, Ohio, U.S.A. Figs. 8-9. Phacops rana africanus subsp. nov. Holotype, USNM 174072. 8, x 1-5. 9, x 1-5. Givetian, Gor Loutad, Spanish Sahara. PLATE 47 BURTON and ELDREDGE, Phacops rana ssp. 358 PALAEONTOLOGY, VOLUME 17 plex P. r. tindoufensis compares closely with P. r. milleri, and P. r. africanus with P. r. crassituberculata, although less closely. The reasons for this latter pairing will be made clear by the comparison between P. r. tindoufensis and P. r. africanus. This brings out critical differences which are also valid for differentiating between the pairs tindoufensis-miUeri and africanus-crassituberculata. The critical differences between P. r. tindoufensis and P. r. africanus ean be divided into those of major structural features and those of ornament. In P. r. tindoufensis (PI. 47, figs. 1-3) the genal angles are rounded but never parabolic, the intercalating ring is pronounced, the eyes large with a maximum of 9 lenses per dorso-ventral file and with lenses projecting beyond the interlensar sclera. The fixigenal eye stems are flattened. The ornament is generally rich, with glabellar tubercles being conical to rounded and of small to medium size, those along the anterior margin being trans- versely elongated. The palpebral lobes, fixigenal eye stems, occipital ring, and the genae bear numerous low tubercles. The thorax is also well ornamented. The corre- sponding features in P. r. africanus (PI. 47, figs. 8, 9; PI. 48, figs. 1-4) show clear differences, the genal angles being always parabolic, the intercalating ring low, the eyes smaller with a maximum of 6 lenses per dorso-ventral file, and the lenses set flush with the sclera. The fixigenal eye stems are rounded. The ornament is sparse with large hemispherical or flattened glabellar tubercles which are never transversely elongated along the anterior glabellar margin. The palpebral lobes and fixigenal eye stems have small numbers of large tubereles, and the occipital ring may be devoid of tubercles or have very few large ones. The genae never bear more than one row of tubercles, these being always on their posterior margins. This latter pattern is followed in the thoracic pleurae. P. r. tindoufensis and P. r. milleri (PI. 47, figs. 1-3, 6-7) show remarkably few differences. The eye of the latter is longer compared with the length of the cephalon than that of the former; and although the eye details are almost identical, only the topmost lenses of each dorso-ventral file are flush with the interlensar sclera in P. r. milleri, whereas in P. r. tindoufensis the upper two or three lenses in each file are flush with the sclera. The intercalating ring of P. r. milleri is marginally more pronounced than that of P. r. tindoufensis, and the distal portions of the thoracic axial rings of the latter (PI. 47, fig. 2) are always constricted into incipient nodes. This node formation is never seen in P. r. milleri. The only other differences are those of ornamentation, in P. r. milleri the flattened, transversely elongated tubercles occur well up the anterior slope of the anterior glabellar lobe, whereas those of P. r. tindoufensis are restricted to the most anterior part of the lobe or are not present at all (PI. 47, fig. 1). Also in the American subspecies the glabellar tubercles are more closely packed and sometimes, viewed from above, are more polygonal than rounded, whereas those of the African subspecies are slightly less crowded and always rounded. The amount of ornament is noticeably less, although of the same type and in the same places, on the genae, thorax, and pygidium of P. r. milleri, than that on P. r. tindoufensis. The comparison P. r. africanus-P. r. crassituberculata yields differences of some- what greater magnitude, but again more of degree than kind. The most noticeable difference lies in the size of the eye. That of P. r. africanus is relatively small and lies high on the gena, whereas that of P. r. crassituberculata is larger and occupies more BURTON AND ELDREDGE: SUBSPECIES OF PHACOPS RANA 359 of the gena (PL 47, figs. 5, 8). Correspondingly the former’s eye socle is larger and the latter’s small. In all other features the eyes are identical. Differences in size and distribution of ornament are marked, the American subspecies being the more richly ornamented. Elongated tubercles are common on the anterior portion of the anterior glabellar lobe of P. r. crassituberculata, but are not present on P. r. africaims. Furthermore, although the glabellar tubercles are arranged in much the same fashion and have the same shapes in the two subspecies, those of P. r. crassituberculata are considerably smaller than those of P. r. africanus. This size difference is again seen in the ornament of the palpebral lobe and fixigenal eye stem, that of the American subspecies being much smaller than that of the African subspecies. In contrast the occipital ring of P. r. africanus has either no ornament or a few low transversely elongated tubercles, whereas that of P. r. crassituberculata always has many, small, transversely elongated tubercles. Genal ornament is much the same size in both subspecies but is confined to the rear of the genae in P. r. africanus, occupying over half the genae in P. r. crassituberculata. The only other differences are seen in the thorax and pygidium. The distal ends of the thoracic axial rings of the African subspecies have a slight constriction which is unknown in the American subspecies, and the pygidium of the African subspecies has 9-11 axial rings to a maximum of 9 in the American subspecies. Although at first sight there might appear to be a wide gap between the two sub- species, this is illusory when details are considered. Eye details which in general indicate fundamental differences are identical, except for the actual size of the eye. The ornament of course is of discriminatory value but at the subspecific level. Other basic characters even of some subspecific value are identical. These characters in- clude (PI. 47, figs. 4, 5, 8, 9; PI. 48, figs. 1, 3, 4) intercalating ring form, cephalic outline and profile, genal angle shape, angle between axial furrows (average 63°), rear-eye ridges. Furthermore, the pairings P. r. milleri-P. r. tindoufensis and P. r. crassituberculata-P. r. africanus are the only admissible ones between American and African subspecies, since the two pairs have different morphological characteristics, as stated above. The milleri-tindoufensis pair share the strongly accentuated inter- calating ring, flattened fixigenal eye stems, distinct eye features, genal angle shape, and ornamental features which cannot be duplicated within the pair crassituberculaia- africanus. Comparisons with other North African species: the specimen illustrated by Sougy (1964, p. 447, pi. 41, fig. 5) and identified by him as P. (Phacops) fecundus degener belongs to P. rana tindoufensis. Further, P. r. tindoufensis bears a resemblance to P. menchikofd Fe Maitre from the Fower Eifelian of the Saoura Basin. However, P. menchikoffi possesses fewer lenses per dorso-ventral file, and appears to have small, sparsely scattered tubercles on the anterior glabellar lobe. P. speculator from the Eifelian of western Morocco has been compared by Alberti (1970) with P. rana milleri. However, although it is clearly allied to P. rana, it lacks the dense development of tubercles of the tindoufensis-milleri pair, and is closer to P. menchikoffi. It is apparent that P. rana (s.l.) is widely represented in North-West Africa, and it appears to one of us (C. J. B.) that there may be links between the North-West African representatives of this species and the P. schlotheimi (s.l.) group of the 360 PALAEONTOLOGY, VOLUME 17 Old World. P. schlotheimi (s.s.) (Burton 1969 for morphological details) bears only a general resemblance to either P. rana africanus or P. r. tindoufensis and cannot be considered directly ancestral. However, members of the P. schlotheimi (s.l.) group are known to exist in the French Pyrenees (Cavet and Fillet 1958, p. 21), Morocco (Richter 1943), and in the Saoura Basin of Algeria (Le Maitre 1952, p. 156). The Algerian form (PI. 48, figs. 5-6) is a new subspecies of P. schlotheimi of Lower Eifelian age and appears to have characters intermediate between P. schlotheimi s.s. and P. rana tindoufensis. However, the authors do not at this time intend to press this similarity any further, being content to maintain that there are sufficient North- West African representatives of the group of P. schlotheimi (s.l.) to have provided an ancestral complex to the P. rana group and that few other groups are thus situated. Discussion. The comparisons have shown that the African subspecies P. rana tindou- fensis and P. r. africanus are, respectively, close to the North American subspecies P. r. milleri and P. r. crassituberculata recently redescribed by Eldredge (1972). The differences separating the subspecies in each continent parallel one another to a remarkable degree, and the four subspecies appear to form a complex distinct from all the other subspecies of P. rana. The nature of this complex is not yet fully under- stood. Eldredge (1972) has shown that P. r. milleri and P. r. crassituberculata do not generally occur together in the same fauna. Their geographical distribution does overlap, especially in northern Ohio and southern Michigan where both are known from the Silica Shale fauna. However, they rarely occur in the same unit. P. r. cras- situberculata shows a marked ‘preference’ for relatively pure limestone, while P. r. milleri is usually found in calcareous shales. A complicating factor is that the two eye variants are occasionally found in association, and that all small (meraspid(?) and early holaspid) cephala from the Silica shale show w///cr/-type bulging lenses, indicating that the early ontogeny of the eye in both P. r. milleri and P. r. crassi- tuberculata was probably the same. It is tentatively concluded that the milleri and crassituberculata eye variants probably represent a stable situation in population genetics, where local populations are adapted to harder substrates and presumably cleaner water (P. r. crassituberculata) or softer, muddier substrates, hence more turbid water (P. r. milleri). Alternatively, the ontogeny of the eye might have been capable of responding to local conditions, i.e. possesses a broad ‘norm of reaction’. However, the precise nature of this relationship cannot be explained on the data available, but in view of the relative ease of differentiation of the variants Eldredge concludes that the best course is to continue to treat them as subspecies. In any case, this mode of interpopulation variation is peculiar to the above two subspecies of P. rana in North America. A similar situation appears to exist for P. r. africanus and P. r. tindoufensis, both EXPLANATION OF PLATE 48 Figs. 1-4. Phacops rana africanus subsp. nov. 1, In 57166, x 2. 2, In 56877, x 1. 3, In 57166, x 1. 4, In 57166, X 2. Givetian, Tifariti area, Spanish Sahara. Figs. 5-6. Phacops .schlotheimi ssp., ULL 256 d 49, a-b. 5, ULL 256 d 49 b, x 3-4. 6, ULL 256 d 49 a, X 3-2. Eifelian, Erg Djemel, Algeria. PLATE 48 BURTON and ELDREDGE, Phacops 362 PALAEONTOLOGY, VOLUME 17 forms being found at the same horizon, the Werneroceras limestone. This varies in lithology from a more or less pure limestone to a calcareous marl, and the two subspecies are found in different facies. P. r. africanus is found in the relatively pure limestones, silty limestones, and coarsely sandy calcareous beds, whereas P. r. tindoufensis is found in shales and marly limestones. There are obvious parallels with the situation cited above by Eldredge, although the two subspecies are not quite as close as the American ones and locality detail is by no means as precise, which limits the rigour of the comparison. However, it is beyond coincidence that a similar adaptional pairing should be found in African subspecies of P. rana which are individually closely similar to one or other of the American pair. This therefore suggests that the African subspecies are related in the same fashion as the American subspecies and that there was likely to have been communication between the two areas during the Middle Devonian. Acknowledgements. The thanks of one of us (C.J.B.) are due to Mile D. Le Maitre for permission to ex- amine material collected by her in Morocco, to Mile D. Brice of the Universite Libre de Lille for permis- sion to examine these and other specimens in her care, to Dr. W. T. Dean formerly of the British Museum (Natural History) for loan of material, to Professor T. N. George, University of Glasgow, to Dr. E. B. Selwood, University of Exeter, and to Drs. W. D. I. Rolfe and J. K. Ingham, Hunterian Museum, Uni- versity of Glasgow, for advice and discussion. The other author (N. E.) acknowledges with thanks the aid of E. J. Collier of the National Museum of Natural History, and H. Strimple of the State University of Iowa in arranging loans from their respective institutions. REFERENCES ALBERTI, G. K. B. 1969. Trilobiteii des jungeren Siluriums sowie des Unter- und Mitteldevons. I. Abh. senckenb. natwforsch. Ges. 520, 692 pp. 1970. Trilobiten des jungeren Siluriums sowie des Unter- und Mitteldevons. II. Ibid. 525, 233 pp. ARDEN, D. D., JNR. and REHRiG, w. A. 1964. Middle Devonian Stratigraphy of Northeastern Spanish Sahara. Bull. Amer. Ass. Petrol. Geol. 48, 1513-1525, 3 figs., 1 table. BARRANDE, J. 1852. Svsteme silurien du centre de la Bolieme: lere partie, Recherches paleontologiques. I. Crustaces; Trilobites. 935 pp., 51 pis. Prague-Paris. BURTON, c. J. 1969. Variation studies of some phacopid trilobites of Eurasia and North West Africa. Ph.D. Thesis, University of Exeter. 1972. Provincial affinities of Eifelian phacopids (Trilobita) of South West England and Brittany. Proc. Usslier Soc. 2 (5), 458-463. CAVET, p. and pillet, j. 1958. Les Trilobites des Calcairs a Polypiers Siliceux (Eifelien) du synclinal de Villefranche de Conflent (Pyrenees-Orientales). Bull. Soc. geol. Fr. (6), 8, 21-31, pi. Ill, 1 text-fig., 1 text-pl. CLARKE, J. M. 1889. The structure and development of the visual area in the trilobite Phacops rana Green. J. Morph. 2, 253-270, pi. 21. CLARKSON, E. N. K. 1966. The life attitude of the Silurian trilobite Phacops musheni Salter 1864. Scott. J. Geol. 2 (1), 76-83, figs. 1-3, pi. 1. ELDREDGE, N. 1972. Systeiuatics and evolution of Phacops rana (Green, 1832) and Phacops iowensis Delo, 1935 (Trilobita) from the Middle Devonian of North America. Bull. Amer. Mus. Nat. Hist. 147, 49 -1 1 1. HALL, J. and CLARKE, J. M. 1888. Palaeontology: Trilobites and other Crustacea. N. Y. Geol. Surv. (Albany), 7, 236 pp. HOUSE, M. R. 1962. Observations on the ammonoid succession of the North American Devonian. J. Paleont. 36, 247-284, 15 text-figs., pis. 43-48. BURTON AND ELDREDGE: SUBSPECIES OF PHACOPS RANA 363 LE MAiTRE, D. 1939. Observations sur la faune des gisements Devoniens du Tafilalet (Maroc). Bull. Soc. geol. Fr. (5), 9, 201-206. 1952. La faune du Devonien Inferieur et Moyen de la Saoura et des abords de I’Erg el Djemel. (Sud Oranais.) Mat. Carte geol. Algerie, L’’® ser.. Pal. 12. MOORE, R. c. (ed.). 1959. Treatise on Invertebrate Paleontology. O, Arthropoda 1. Univ. Kansas Press. ORMiSTON, A. R. 1967. Lower and Middle Devonian trilobites of the Canadian Arctic islands. Bull. Geol. Surv. Can. 153, 148 pp. RICHTER, R. and E. 1926. Die Trilobiten des Oberdevons. Abh. preuss. geol. Landesanst . N.F. 99, 1-314, 12 pis. 1943. Studien in Palaozoikum der Mittelmeer-Lander. 4A, Trilobiten aus dem Devon von Marokko mit einem Anhang iiber Arten des Rheinlands. Senckenbergiana, 26, (1-3), 116-199, 11 figs., 8 pis. SHAW, F. c. and ormiston, a. r. 1964. The eye socle of trilobites. J. Paleont. 38, 1001-1002, 1 fig. SOUGY, J. 1962. West African Fold Belt. Bull. geol. Soc. Amer. 13, 871-876, 1 fig. 1964. Fes formations paleozoiques du Zemmour noir (Mauritanie septentrionale). Ann. Fac. Sci. Dakar, 15, XII. 695 pp., 77 figs., 37 tables, 50 pis., 2 geol. maps. SUTTON, J. 1968. Development of the Continental framework of the Atlantic. Proc. Geol. Lond. 79, 275-303. C. J. BURTON Department of Geology The University Glasgow, G12 8QQ N. ELDREDGE The American Museum of Natural History Central Park West at 79th St. Final typescript received 13 April 1973 New York, N.Y. 10024, U.S.A. L 'V % (Ti . " '( V •■■«> ■' • >- -*> < ,v -■v:-< i,s„ -,i\ •'*'• ^ ' - I , i- ' ■ \ ' < m'*k. V ' - . .1 ‘iiW^ . ', ' - , . . ■•- - : ■ ■ '' 1- ■ . ■ I '.• , ■ , , ,7r ■^■ - •^•» ' ■ tS' S«- ./ ■ -I- . ^-^ ■;-• ■,,;. ,,v, J*' '' ' : ' , ,r-i. H •.-f.Vfe) ■ I I ,1 i / V *• PODOCARPUS FROM THE UPPER CRETACEOUS OF EASTERN ASIA AND ITS BEARING ON THE THEORY OF CONIFER EVOLUTION by V. A. KRASSILOV Abstract. Podocarpus tzagajanicus sp. nov. from the Uppermost Cretaceous (Tzagajan beds) of the Bureja River augments the Mesozoic record of the northern hemisphere Podocarpaceae. ‘Northern’ and ‘southern’ conifers grew side by side in Mesozoic and Tertiary forests. The distribution of conifers has been more deeply affected by climatic changes than by continental drift. In a review of the Cambridge symposium on the biogeographical aspects of con- tinental drift, Jardine and McKenzie (1972) quoted among selected examples Florin’s theory of conifer distribution. They claimed (p. 24) that ‘the history of the conifers provides another striking example of the action of drifting continents as agents of dispersal. Florin showed that from the Late Carboniferous (about 300 my) to the early Eocene (about 50 my) each of the conifer genera (with the exception of Arau- caria) had either a “Gondwana” distribution ... or “Laurasian” distribution. . . . The present disjunct distribution of Podocarpus and other Gondwana genera may be the product of the break-up of Gondwanaland, and the Tertiary spread of them into Indonesia and Southeast Asia may result from the northward drift of Australia.’ Such views on the history of conifers are widely accepted. Rudolf Florin, an outstanding palaeobotanist and a great authority on gymno- sperm taxonomy, came to his ideas of conifer evolution when describing Tertiary conifers from Chile (Florin 1 940). He held then that both lineages of conifers, northern and southern, had been perfectly separated through time and space. Paranocladus, Walkomiella, Buriadia, Araucariaceae, Podocarpaceae, and Athrotaxis constituted the main body of the southern group and the rest of the conifers the northern one. Twenty years later Florin (1963) reiterated his views. The only exception made was for the Araucariaceae (but not for the genus Araucaria) which had been recorded from several northern localities. The works of R. Krausel on fossil woods and of R. A. Couper on microfossils provided additional evidence in favour of Florin’s theory. Although several authors (Buchholz 1948; Ferguson 1967; Krassilov 1967, 1971) objected to this theory, it became fairly popular among botanists and earth scientists and was cited in many textbooks (e.g. Stebbins 1967). In recent years it has been used as a confirmation of continental drift. It is worth mentioning that Florin himself opposed the drift theory and relied upon ‘continental bridges’ as pathways of conifer distribution. The ‘southern’ palaeozoic conifer Walkomiella is hardly distinguishable from the ‘northern’ Lehachia. The early Triassic northern family Voltziaceae was represented in the southern hemisphere by the genus Voltziopsis. Other Mesozoic families such as Cycadocarpidiaceae and Cheirolepidiaceae were also distributed in both northern and (Palaeontology, Vol. 17, Part 2, 1974, pp. 365-370, pi. 49.] 366 PALAEONTOLOGY, VOLUME 17 southern continents. Cycadocarpidium has been recorded from the Upper Triassic of Argentina and the generic name Tomaxellia was recently proposed by Arch- angelsky (1968) for a southern cheirolepidiaceous conifer with pollen grains of Classopollis-type. He also suggested the cheirolepidiaceous athnities of Patagonian Jurassic conifer Pararaucaria, as well as Indostrobus from the Cretaceous of India. The ‘southern’ taxodiaceous genus Athrotaxis (or its nearest approach Atlirotaxites) has been repeatedly recorded from the Lower Cretaceous of Canada and U.S.S.R. (Bell 1956; Krassilov 1967). All the above-listed facts contradict the division of ancient conifers into Gond- wanian and Laurasian groups. But even more important is the evidence of the Laurasian distribution of Mesozoic Podocarpaceae. Pollen grains of the podo- carpaceous type are known from many northern localities. However, most of them have been recently attributed to artificial genera. I referred to Podocarpaceae several megafossils from the Lower Cretaceous of the Primorye (near Vladivostok, Far East of the U.S.S.R.). One of them, Podocarpus sujfunensis Krassilov, displays the characters of the Nageia section of the genus Podocarpus (Krassilov 1967). The leaves are 120 mm long and 2-2 mm broad, with numerous veins, amphistomatic. Stomata all over the surface in longitudinal files separated by several cell files, amphi- cyclic; the polar encircling cells shared by adjacent stomata or absent; subsidiary cells papillate. Another species, Podocarpus harrisii Krassilov has linear-lanceolate, shortly petiolate, single veined, hypostomatic leaves about 40 mm long and 4-6 mm wide. The abaxial epidermis with broad central stomatic band is occasionally divided into two or three parts by narrow and irregular nonstomatiferous zones. Stomata arranged in files, longitudinally orientated, amphicyclic; subsidiary cells papillate. The topography of the abaxial epidermis is rather unusual for conifers, with flat single-veined leaves. However, several living species of the subgenus Stacliycarpus show stomata over the vein. The division of stomatic bands into ‘Teilstreifen’ is also known among Stachycarpus species (Florin 1931). Bilaterally flattened leaves have been recorded from the Lower Cretaceous of Primorye under the name Paracnwpyle florinii Krassilov. Primorye certainly was not the only Laurasian territory where Podocarpaceae flourished during Mesozoic time. Gomolitzky (1962) described leaves with podo- carpaceous cuticle characters from the Jurassic of Central Asia. The fossil wood Podocarpoxylon triassicum has been found in the Keuper of Central Europe (Selmier and Vogellehner 1968). I suggested the podocarpaceous affinity of the Wealden species Tritaenia {Ahietites) linkii (Roem.) Magdefrau and Rudolf as evidenced by stomata organization and topography of abaxial epidermis with three ‘Teilstreifen’ (Krassilov 1967, 1971). Podocarpus tzagajanicus sp. nov. from the uppermost Cretaceous of the Amur- land augments the record of Mesozoic podocarps. As far as I know it is the only Late Cretaceous representative of the family. EXPLANATION OF PLATE 49 Figs. 1-8. Podocarpus tzagajanicus sp. nov. Upper Cretaceous, Bureja River. 1, 2, leaf fragments, x 1. .3, 4, abaxial cuticle, parts of stomatal band (slightly retouched), x 70. 5, 6, cells of adaxial epidermis, X 58 and 146. 7, 8, stomata, x 395. PLATE 49 KRASSILOV, Podocarpus 368 PALAEONTOLOGY, VOLUME 17 Genus podocarpus L’Herit Podocarpus tzagajanicus Krassilov sp. nov. Plate 49, figs. 1-8 Diagnosis. Leaves linear-lanceolate, flat, single-veined acuminate, with slightly thickened margins, 5-6 mm wide, hypostomatic. Abaxial epidermis with two sto- matic bands about 0-8 mm wide on either side of the midrib. Stomata well spaced, arranged in discontinuous rows, longitudinally orientated, amphicyclic, with 5-6 subsidiary cells. Stomatal pit elliptical, bordered with a ridge and overarching papillae. Cells of stomatal bands papillate. Marginal nonstomatiferous zones as wide as stomatal bands. Cells outside the stomatal bands without papillae. Anticlinal walls ridged, undulating to sinuous. Holotype. Specimen 575-126 and slide preparation 575-126a, Institute of Biology and Pedology, Far- Eastern Scientific Centre, Vladivostok; PI. 49, figs. 1, 3-8. Occurrence. Outcrop of Tzagajan clays near the mouth of Bureja River, tributary of the Amur. Age. Uppermost Cretaceous (Danian). Description. Three incomplete leaves have been collected from the light grey Tzagajan clays. They are fossilized as yellowish-brown incrustations with small pieces of cuticle. The largest leaf fragment is more than 60 mm long (the whole length was probably 80-90 mm), 6 mm wide, tapering towards the acuminate apex. The midrib is prominent, up to 0-9 mm broad, adaxially flat, abaxially appearing as a low ridge; the margins are microscopically even. The abaxial epidermis is divided into two stomatal bands and three nonstomatiferous zones, all of nearly equal width. The stomatal bands are not sunken, not sharply delimited; stomata forming discontinu- ous files, longitudinally orientated, rather evenly spaced; the subsidiary cells with large papillae overarching the stomatal pit. The papillae are dorsally united into prominent ridge. Guard cells invisible. Stomatal pit including the ridge 45-63 /xm long, 37-45 ju,m wide. The outlines of epidermal cells are rather indistinct within the stomatic bands. All cells are provided with round-elliptical papillae about 15 g.m in diameter. The cells of marginal zones are rectanguloid, arranged in files, about 54 j.im long, 22 ;u,m wide with undulating anticlinal walls. The leaf margins are bordered with narrow strips of elongated cells 13 ^uin wide. The cells of costal zone rectangular, up to 1 12 X 45 occasionally short and square or irregular, with more distinctly sinuous anticlinal walls. Remarks. These leaves are hardly distinguishable from the Fort Union specimens which have been identified by Brown (1962) as Amentotaxus camphelli (Gardner) Florin. However, the cuticle characters are different from those of Amentotaxus, as well as from other conifers except several living species of Podocarpus subgen. Stachycarpus, which have hypostomatic leaves. According to Florin (1931) the stomata of Stachycarpus are amphicyclic, with 4-6 subsidiary cells which are papil- late and heavily cutinized forming a ridge around the stomatal pit. The anticlinal KRASSILOV: PODOCARPUS AND CONIFER EVOLUTION 369 walls of epidermal cells more or less undulate. In contrast, Amentotaxus has mono- cyclic, comparatively frequent stomata with 4-10 subsidiary cells. The anticlinal walls are usually straight. Comparable leaves have been described from the Tertiary of Japan, North America (see Dilcher 1969), and Europe (e.g. Podocarpus kinkelini Madler 1939) but the cuticles of the latter are not known. Conclusions. We may conclude that there was no family of Mesozoic conifers with exclusively ‘Gondwanian’ or, ‘Laurasian’ distribution. Such genera as Cycado- carpidium, Araucarites, Athrotaxites, and Podocarpus successfully crossed the Tethys Sea long before Early Eocene time. The history of conifers has little bearing on the problem of continental drift. It seems that southern and northern conifers have been more effectively separated by equatorial temperature conditions than by water barriers. The probability of dispersal through the equatorial zone has been affected by the changing contrast between the tropical and extratropical climates. The equable Jurassic climate favoured the dispersion of conifers and the equatorial barrier was surmounted by the Araucariaceae, Podocarpaceae, Taxodiaceae, and less successfully by the more temperate Pinaceae. The Sequoia-Taxodium group of conifers appeared later when transequatorial migrations were barred by better- defined climatic zonation. They were confined to the northern hemisphere. The ‘northern’ and ‘southern’ conifers grew side by side during the Cretaceous and Tertiary periods, but the southern ones were gradually eliminated from Laurasia. They persisted on southern continents where the climatic conditions remained more or less comparable to those of the Jurassic period. REFERENCES ARCHANGELSKY, s. 1968. On the genus Tomaxetlia (Coniferae) from the Lower Cretaceous of Patagonia (Argentina) and its male and female cones. Journ. Litm. Soc. (Bor.), 61, 153-165. BELL, w. A. 1956. Lower Cretaceous floras of Western Canada. Geol. Surv. Canada Mem. 285, 1-153. BROWN, R. w. 1962. Paleocene flora of the Rocky Mountains and Great Plains. U.S. Geol. Surv. Prof. Paper, 375, 1-119. BUCHHOLZ, j. T. 1948. Generic and subgeneric distribution of the Coniferales. Bot. Gaz. 110, 80-91. DILCHER, D. L. 1969. Podocarpus from the Eocene of North America. Science, 164, 299-301. FERGUSON, D. K. 1967. On the phytogeography of conifers in the European Cenozoic. Palaeogeogr., Palaeo- climatoL, Palaeoecol. 3, 73-110. FLORIN, R. 1931. Untersuchungen zur Stammesgeschichte der Coniferales und Cordaitales. K. svenska VetenskAkad. Handl. 10, 1-588. 1940. The Tertiary fossil conifers of south Chile and their phytogeographical significance. Ibid. 3 ser. 18, 3-92. 1963. The distribution of conifer and taxad genera in time and space. Acta. Iiorti berg. 20, 121-312. GOMOLiTZKY, N. p. 1962. Podocarpopliyllum—a new genus of conifers from the Jurassic coal-bearing deposits of Angren, Central Asia. Bot. Journ. 47, 1020-1032. [In Russian.] JARDiNE, N. and MCKENZIE, D. 1972. Continental drift and evolution of organisms. Nature, 235, 20-24. KRASSILOV, v. A. 1 967. The Early Cretaceous flora of South Priniorye and its bearing on stratigraphy. Moscow, 364 pp. [In Russian.] 1971. Evolution and taxonomy of conifers (a critical review). Palaeont. Journ. 1, 7-20. [In Russian.] MADLER, K. 1939. Die pliozane Flora von Frankfurt am Main. Abhandl. Senckenberg. natiirforsch. Ges. 446, 1-202. 370 PALAEONTOLOGY, VOLUME 17 SELMIER, A. and VOGELLEHNER, D. 1968. Podocarpoxvloti triassicum n. sp., ein phylogenetisch bedeutsame ‘modernes’ Sekunderholz aus dem Keuper von Franken. Neues Jahrb. Geol. Paldont. Abhandl. 132, 70-86. STEBBINS, G. L. 1967. Variation and evolution in plants. New York-London. 643 pp. Typescript received 14 August 1972 V. A. KRASSILOV Institute of Biology and Pedology Vladivostok, U.S.S.R. LOWER CARBONIFEROUS CONODONT FAUNAS FROM NORTH-EAST DEVONSHIRE by s. c. MATTHEWS and J. m. thomas Abstract. Two distinct successions, the Bampton and the Westleigh, exist in the Lower Carboniferous of north- east Devonshire. The characters of the two are briefly described, mainly on the basis of field evidence. Conodonts reinforce earlier suggestions that these successions both belong in cull and cuIII of the Lower Carboniferous. The conglomeratic limestones in the Westleigh succession include some apparently shelf-derived clasts. The cono- donts indicate that these coarse limestones (with Gnathodus hilineatus) have received an admixture of distinctly earlier forms (ScaUognathus anchoralis, siphonodellids). If the reworked forms have come from shelf situations, as much as 800 m of Lower Carboniferous shelf-stratigraphy might have contributed conodonts to the Westleigh suc- cession. The Westleigh Limestone and the Hellefelder Kalk (Lower Carboniferous, Sauerland, Germany) are briefly compared. Lower Carboniferous outcrop runs across north Devon from Barnstaple in the west and ends in the east at the New Red Sandstone overstep (I.G.S. 1 : 63,360 Sheet 310, Tiverton). The Lower Carboniferous is set in a regional succession that is appar- ently conformable from Famennian (Goldring 1955, 1970 : lower part of the Pilton Beds) to Westphalian (Prentice 1960: Instow Beds in the west; Thomas in Webby and Thomas 1965: Holmingham Beds in the east). Near the eastern end of the out- crop belt, a change of facies is seen in the rocks that follow above the Pilton Beds. At Bampton (see text-fig. 1 for localities) the Doddiscombe Beds-Bampton Lime- stone Group succession is dominated by shales. Around Westleigh, limestones are dominant. We refer in this paper to a Bampton succession and a Westleigh succession. THE BAMPTON SUCCESSION At Bampton, the Pilton Beds are succeeded by the Doddiscombe Beds (a type sec- tion is available at Doddiscombe Quarry: National Grid Reference SS 985 233). These shales show a fine dark grey/black lamination. The dark colours survive weathering. Broken surfaces are distinctly rough, a characteristic attributable to the fact that the original fine clastic accumulate (silt grade or lower) received some secondary silicification. The Doddiscombe Beds have a thickness of approximately 8 m. There is a gradual passage upward into the softer, less finely laminated, pale- weathering shales of the Hayne Beech Beds. The Hayne Beech Beds are the lowest of the three units recognized in the Bamp- ton Limestone Group. Among its pale shales there are some minor developments of impure limestone. The unit is poorly exposed in the ground north and east of Bampton, where a topographic depression marks the run of its outcrop. The best available exposure is at Hayne Beech Farm (SS 992 229). Good exposures were available for a time during road-widening operations north-east of Hukeley Bridge (SS 972 235). A small exposure in the side of a lane near Hukeley Farm [Palaeontology, Vol. 17, Part 2, 1974, pp. 371-385, pis. 50-51.] TEXT-FIG. 1. Location maps, la, top right: regional location (at centre: location of area treated in lb), lb, geological map of the Bampton-Westleigh area (at right: area treated in Ic). Ic, location of exposures in the Holcombe Rogus-Westleigh neighbourhood. MATTHEWS AND THOMAS: CARBONIFEROUS CONODONTS 373 (SS 972 238) of uncertain stratigraphic position has produced some ostracodes (Maternella n. sp. aflf. circumcostata (Rabien I960)— A. Gooday, personal eom- munication 1973). The next unit in the Bampton Limestone Group, the Kersdown Chert, has cherts with interbedded shales and occasional fine-grained limestones. At the type oecur- rence, in Kersdown Quarry, Bampton (SS 964 222) 7 m stratigraphic thickness is exposed in a fractured anticlinal fold. Individual beds persist around this strue- ture. The cherts are slightly caleareous. The interbedded shales are slightly silicified and contain relics of radiolarians. The fine-grained limestones are usually about 0-5 m thick. A thicker, coarser limestone, with graded bedding, occurs near the top of the section exposed. The matrix of these limestones is permeated by secondary silica. Silica occurs also as nodules, and in such cases the loeal matrix is eompletely replaeed. Small corals, brachiopods, and a certain number of goniatites (Prentice and Thomas 1965, pp. 38-40: Merocanites cf. similis, Bollandites sp.) have been found in this unit. The Kersdown Chert outcrop corresponds with topographic highs in the area. At exposure the cherts are pale in colour and are usually well- jointed. They have a rough, porous texture due to weathering-out of calcareous material. The highest of the three Bampton Limestone Group units is the Bailey’s Beds (27 m thick). This is exposed in Bailey’s Quarry, south-east of Bampton (SS 960 218). The unit has dark impure limestones interbedded with dark shales. Only minor chert developments are present, although both the nodular and the dispersed mode of occurrence of secondary silica can be found. Individual limestone beds maintain their thickness and character throughout the extent of the exposures available. The top of the Bailey’s Beds (and the top of the Bampton Limestone Group) is taken at the top of the last thin limestone seen as the Bailey’s Beds pass upward into the unrelieved black shales of the Dowhills Beds. The shales near the top of the Bailey’s Beds have abundant large bivalves, many of them identifiable as Posidonia becheri. Michiganites hesteri (generic attribution as in Weyer 1972) has been found in this unit at a locality near Huntsham (ST 996 202: see Prentice and Thomas 1965, pp. 40- 41). Spirally striate goniatites, including forms referable to Neoglyphioceras spirale, occur in the dark shales immediately overlying the highest limestones, i.e. in the Dowhills Beds. The Bampton succession, therefore, has mainly fine-grained rocks, sometimes banded and with only relatively rare interruptions by eoarser elastics. There are oceasional developments of sole structures in the thin silicified limestones of the Kersdown Cherts unit, and some of these beds have a ripple cross-lamination (type C of Jopling and Walker 1968) near their tops. One example of grading has been noted above. Evidence of burrowing is also found. In the Bailey’s Beds the limestones are laterally persistent (so far as the available exposures show), their bases sharply defined. Some have parallel lamination throughout, others are struc- tureless internally, and in rare cases ripple cross-stratification is present. All are fine grained, and therefore none produces a convincing example of graded bedding. The Bampton succession is taken to represent the distal parts of a turbidite influx. The units of the Bampton suecession can be mapped eastward up to the New Red 374 PALAEONTOLOGY, VOLUME 17 Sandstone overstep near Ashbrittle. Southward from there, the Lower Carboni- ferous reappears; but it now has the characteristics of the limestone-dominant Westleigh succession. THE WESTLEIGH SUCCESSION The Westleigh Limestones are well shown in numerous quarries opened near Burles- combe. The stratigraphy in the immediate neighbourhood of the Pilton Beds- Westleigh Limestones boundary is not well exposed. On the other hand, at the top of the Westleigh Limestone it is clear that the carbonate-dominant sequence is fol- lowed by a succession of fine-grained rocks exactly comparable with the Dowhills Beds of the Bampton neighbourhood (Thomas in Webby and Thomas 1965). In Westleigh Quarry itself (ST 065 1 76) it is possible to separate a lower part (Lower Westleigh Limestone) of the succession, with relatively thin-bedded limestones and little shale, from an upper (Upper Westleigh Limestone), which has thicker limestones —often impressively coarse— interbedded with shales. Elsewhere in the area, but always in poorly exposed situations, the Upper Westleigh Limestone outcrops close to the Doddiscombe Beds. It has not yet been possible to unravel whatever strati- graphic relationships exist in such places. The Lower Westleigh Limestone is found only in the immediate neighbourhood of Westleigh. The regular, thin-bedded limestones, like the occasional intervening calcilutites, have finely disseminated secondary silica in their matrix. Graded beds are occasionally seen. Bioturbation is common. Near the top of the Lower West- leigh Limestones there are some coarse conglomerates whose clasts range up to a maximum seen in a slab that measures some 2 m long and 0T5 m thick. The clasts vary in shape from angular to moderately well rounded. The base of the Upper Westleigh Limestone is taken at the first occurrence of shales. These often have a fine colour-banding, and frequently show bioturbation. The majority of the shales are slightly calcareous. Some calcilutites (including the crenistria Bed, mentioned below) are also present. Certain laminae in the shales have flattened specimens of goniatites, orthocones, and large lamellibranchs {Posidonia hecheri). Occasionally, representatives of epibenthonic faunas are also present (athyrid and productid brachiopods, crinoid calices, and crinoid stem material, often found in a good state of articulation). These shales are interbedded with limestones whose thickness ranges between 0-5 cm and 10 m, with a mean around 0-6 m. The limestones are often con- glomeratic, with clasts whose dimensions may be of the order of several cm. Angular blocks of fine- and medium-grained limestone are common. Ooliths and pellets, and bryozoan, crinoid, coral, and brachiopod debris can be recognized. The whole fabric is matrix-supported (ooliths and crinoid debris in a fine calcareous matrix for example, or large clasts in an oolitic-crinoidal matrix). Each limestone has a sharp base. Sole marks are few. Graded bedding is seen in rare cases, always involving beds whose thicknesses are in the 1 5 to 20 cm range. The thicker units have thoroughly mixed grain sizes through the major part of a bed’s thickness, with finally a good grading (‘delayed graded bedding’ of Kuenen 1964) through some 5 cm of thickness leading into the shale above. Some beds, usually fine calcarenites with no large clasts, are laminated throughout (cf. Piper 1972), but most lack regular internal MATTHEWS AND THOMAS; CARBONIFEROUS CONODONTS 375 Structure and have clasts randomly distributed through all but the topmost part of the bed. Nodular developments of chert occur at various levels in the thick limestone beds. In the Upper Westleigh Limestone, trace-fossils occur in the limestone bands only in the form of simple burrows descending from shale above. There are no stratigraphically useful fossils in the Lower Westleigh Limestone. Posidonia becheri occurs through most of the Upper Westleigh Limestone, but in the uppermost parts of the succession it is apparently replaced by the smaller Caneyella membranacea. In a number of exposed sections Michiganites hesteri occurs near the base of the Upper Westleigh Limestone, and rare goniatites of the hudsoni-maximus group have been found in shales slightly higher in the succession. Higher still, speci- mens of Goniatites granosus, G. cf. granosus poststriatus, G. cf. sphaericostriatus, Sudeticeras cf. crenistriatus, and Neoglyphioceras spirale have been found, indicat- ing that the Upper Westleigh Limestone ranges into younger horizons than are represented in the Bampton Limestone Group (compare the record of Neoglyphio- eeras spirale from low in the Dowhills Beds at Bampton). It was suggested above that the Bampton Limestone Group represents a distal turbidite situation. The Lower Westleigh Limestone might have been produced by a similar set of processes, but delivering here a greater abundanee of fine carbonate debris. Some graded beds are found in the Lower Westleigh Limestone. Beds with bioturbation are very common, and concentrated borrowings may have obliterated the primary sedimentary characteristics of some beds. The Lower Westleigh rocks are usually silicified to such an extent that they cannot be broken down in 10% acetic acid. Beds which will break down produce a great deal of insoluble residue, mueh of it possibly originally delivered as volcanic ash. The thicker limestones of the Upper Westleigh Limestone have quite a different set of sedimentary character- istics, and their interbedded shales record the occasional establishment of epiben- thonic faunas. The Upper Westleigh Limestone, it seems, might record a proximal turbidite situation, which is not represented in the Bampton succession and which was still in existence at Westleigh after the last of the thin limestones had been deposited at Bampton. THE CONODONT FAUNAS The goniatites reported by Prentice and Thomas (1960, 1965) were sufficient to show that the Bampton Limestone Group and the Westleigh Limestones are of much the same age. It was hoped at the outset of the present projeet that eonodonts might provide for more detailed correlations between the two successions. However, the results so far have been disappointing in some ways. For example, the lower parts of the successions, which have produced relatively few macrofossils, have been almost equally unobliging as far as eonodonts are concerned. Also, in the higher parts of the Westleigh succession, where eonodonts are abundant, the results are eomplicated by frequent indications of reworking. There are, nevertheless, some useful finds to be recorded. And the evidence of reworking, at first sight confusing, can be made to yield information on the range of age of the rocks that contributed material to the Westleigh succession. The eonodonts listed below have been mounted on 32-cavity slides and the slides 376 PALAEONTOLOGY, VOLUME 17 deposited in the Geology Museum, University of Bristol. Five-figure numbers pre- fixed BU, cited at the end of the faunal lists, identify slides. Numbers with a suffix, as cited in the plate explanations, identify the particular cavities that contain in- dividual figured specimens. Pilton Beds (sample taken from a small lens of limestone set among slates in the courtyard of Doble Farm, National Grid Reference SS 0415 2205): Gnathodus cf. semiglaber, Siphonodella sp., Spathognathodus cf. stabdis (BU 22413). This fauna, though very small, is important in that palaeontological information is rare in the upper part of the Pilton Beds (see a review of present information on the Pilton Beds in Goldring 1970). The Doble Farm fauna may be correlated tenta- tively with the upper Siphonodella crenulata Zone of Germany (Voges 1959, 1960) which some might wish to identify with the lower part of the AmmonellipsitesSivdt (but see Matthews 1970). More abundant conodonts of these kinds have recently been found in the Cork Beds of south-west Ireland (Matthews and Naylor 1973). The Doble Farm sample, like the Irish ones, produced a relatively heavy acid residue dominated by skeletal debris (ostracode, gastropod, and crinoid material) preserved in iron oxide. Kersdown Chert (sample taken from a limestone high in the succession of siliceous rocks exposed in Kersdown Quarry, Bampton. Ref. SS 9635 2220): Gnathodus commutatus Iwmopimctatus, G. delicatus, G. semiglaber, G. texanus pseudosemiglaber, G. texanus texanus, G. sp. indet., G. sp., Mestognathus beckmanni, Pseudopolvgnathus sp. (BU 22418). EXPLANATION OF PLATE 50 Specimens dusted with ammonium chloride. All x 25. Figs. 1-22 from immediately below the crenistria Bed at Westleigh Quarry (see discussion of Westleigh Limestone). Figs. 1, 4. Genicukitus claviger. BU 22422/1 1, BU 22420/20. Figs. 2, 3. Apatognathus sp. BU llAlljS, BU 22422/4. Fig. 5. Polygnathus communis carina. BU 22422jl. Figs. 6, 10. Polygnathus communis communis. BU 22422/8, BU 22422/6. Fig. 7. Pscudopolygnathus triangulus triangulus. BU 22422/1. Figs. 8, 9. Scaliognathus anchoralis. BU 22422/3, BU 22422/2. Figs. 1 1, 12, 15. Siphonodella sp. BU 22422/14, BU 22422/15, BU 22422/17. Fig. 13. Polygnathus inornatus. BU 22422/16. Fig. 14. Siphonodella cooperi. BU 22422/13. Fig. 16. Siphonodella isosticha. BU 22422/12. Figs. 17, 18. Cavusgnathus unicornis. Lateral and upper views of BU 22422/10. Fig. 19. Gnathodus hilineatus. BU 22420/2. Figs. 20, 22. Gnathodus sp. BU 22420/1, BU 22420/4. Fig. 21. Gnathodus delicatus. BU 22422/18. Figs. 23-28 from the Kersdown Chert at a point high in the succession exposed in Kersdown Quarry (see discussion of Kersdown Chert). Figs. 23, 25, 26. Gnathodus texanus pseudosemiglaher. BU 22418/1, BU 22418/3, BU 22418/4. Fig. 24. Gnathodus texanus texanus. BU 22418/2. Figs. 27, 28. Mestognathus beckmanni. Upper and lateral views of BU 22418/5. PLATE 50 MATTHEWS and THOMAS, Carboniferous conodonts 378 PALAEONTOLOGY, VOLUME 17 A loose block of limestone found in Kersdown Quarry, and presumed to have come from much the same site in the stratigraphy as the sample above, yielded: Gnathodus commutatus commutatus, G. delicatus, G. texanus pseudosemiglaber, G. texanus texanus, Siphonodella sp. (BU 22417). Rare conodonts can be found in the cherty rocks of the quarry. They include nothing that does not appear in the two lists above. An interesting feature of these two lists is that they each have one conodont that is distinctly older than the others. It would be dangerous, however, to conclude from this that the Bampton succession, as well as the Westleigh, has reworked conodonts. The Kersdown Quarry samples have each produced only one ‘questionable’ specimen, and the possibility of con- tamination cannot be excluded. The question of reworking remains open. Kersdown Quarry has produced a specimen of Merocanites cf. similis and a Bollandites sp. (Prentice and Thomas 1965). The succession in the quarry probably equates with a high part of the German AmmoneUipsites-^X.\xiQ (see a discussion of conodonts and goniatites of this age in Weyer 1972). Neither the conodonts nor the goniatites give any indication of Goniatites-^ivdt horizons. Westleigh Limestone. The calcilutite bed with abundant specimens of Goniatites crenistria is a useful marker in the quarries around Burlescombe and Holcombe Rogus. At Westleigh Quarry itself, in the Middle Quarry (ST 065 176; see text-fig. Ic), a coarse limestone situated immediately below the crenistria Bed has produced abundant conodonts, including the following: Apatognathus sp., Cavusgnathus unicornis, Geniculatus claviger, Gnathodus bilineatus, G. commutatus commutatus, G. commutatus homopunctatus, G. delicatus, G. girtyi subspp., G. semiglaber, G. sp. EXPLANATION OF PLATE 51 Specimens dusted with ammonium chloride. All x 25. Figs. 1, 2, 3, 18, 19. Gnathodus sp. indet. BU 22422/27, BU 22422/28, BU 22420/8, BU 22422/29, BU 22422/30 respectively. Immediately below crenistria Bed, Westleigh. Figs. 4, 8, 9. Gnathodus sp. BU 22422/32, BU 22422/21, BU 22422/31 respectively. Immediately below crenistria Bed, Westleigh. Fig. 5. Gnathodus commutatus nodosus. BU 22442/5. From high in the Upper Westleigh Limestone (see discussion of higher horizons in the Westleigh Limestone), Whipcott Quarry east. Fig. 6. Gnathodus commutatus homopunctatus. BU 22422/26. Immediately below crenistria Bed, Westleigh. Fig. 7. Pseudopolygnathus triangulus pinnatus. BU 22430/20. Stout’s Cottage Quarry. Figs. 10, 11. Gnathodus commutatus commutatus. BU 22422/20, BU 22422/19 respectively. From imme- diately below crenistria Bed, Westleigh. Figs. 12U5, 20-24. Gnathodus bilineatus. 12, BU 22420/6. 13, BU 22434/2. 14, BU 22420/3. 15, BU 22434/3. 20, BU 22434/1. 21, BU 22434/4. 22, BU 22420/5. 23,22420/9. 24, BU 22420/4. Specimens in ligs. 13,15, 20, 2 1 high in the Upper Westleigh Limestone, Westleigh Lower Quarry ; others immediately below crenistria Bed, Westleigh. Figs. 16, 17, 28, 31. Gnathodus girtyi. BU 22422/23, 24, 22, 25 respectively. Immediately below crenistria Bed, Westleigh. Fig. 25. Siphonodella sp. BU 22413/14. Pilton Beds, Doble Farm. Fig. 26. Gnathodus c{. semiglaber. BU 22413/12. Pilton Beds, Doble Farm. Fig. 27. Spathognathodus cf. stabilis. BU 22413/13. Pilton Beds, Doble Farm. Figs. 29, 30. Gnathodus girtyi. BU 22434/5 and 6 respectively. High in the Upper Westleigh Limestone, Westleigh Lower Quarry. PLATE 51 MATTHEWS and THOMAS, Carboniferous conodonts 380 PALAEONTOLOGY, VOLUME 17 indet., G. cf. bilineatus, G. cf. semiglaber, Mestognathus beckmanni, Polygnathus communis communis, P. communis carina, P. inornatus, P. sp., Pseudopolygnatlius triangulus triangulus, Scaliognathus anchoralis, Siphonodella cooperi, S. isosticha, S. sp. (BU 22420, 22421, 22422). At East Whipcott Quarry (ST 074 187) the crenistria Bed is again available. The bed immediately below produced Doliognathus lotus, Gnathodus bilineatus, G. com- mutatus commutatus, Gnathodus cf. texanus, G. sp. indet. (BU 22423) and from a second sample Cavusgnathus naviculus, Gnathodus bilineatus, G. commutatus com- mutatus, G. commutatus homopunctatus, G. girtyi subsp., Mestognathus beckmanni, Siphonodella cf. obsoleta (BU 22424, 22425, 22426, 22427). At Fir Copse Quarry (ST 069 189), where no reference to the crenistria Bed as a marker is available, the coarse limestones again produced a mix of Gnathodus bilineatus and older forms: Geniculatus claviger, Gnathodus bilineatus, G. cf. bulbosus, G. commutatus commutatus, G. texanus pseudosemiglaber, G. sp. indet., Mesto- gnathusl sp.. Polygnathus communis carina, Scaliognathus anchoralis, Siphonodella sp., Spathognathodus scitulus{^Vi 22431). A second Fir Copse Quarry sample, collected from a coarse limestone situated immediately below an ash development, produced Gnathodus bilineatus, G. girtyi subsp., G. sp. indet., Mestognathus beckmanni, Siphonodella sp. (BU 22432). At Stout’s Cottage Quarry (ST 038 192), a sample taken from a point 7 m above the Michiganites hesteri occurrence (Prentice and Thomas 1965, p. 41) yielded the following: Gnathodus bilineatus, G. commutatus commutatus, G. commutatus homo- punctatus, Pseudopolygnatlius triangulus pinnatus, Pseudopolygnatlius cf. dentilineatus (BU 22430). These faunas (Westleigh, Whipcott, Fir Copse, and Stout’s Cottage Quarries) must be dated by reference to the youngest forms present. The occurrences of Gnatho- dus bilineatus, supported by the G. girtyi subspecies, G. commutatus commutatus and G. commutatus homopunctatus, suggest low GoniatitesSixdt horizons. It is satis- factory that this estimate of the age of these mixed faunas can be confirmed at West- leigh and Whipcott by occurrences of Goniatites crenistria, the zone fossil for cullla in the German orthochronology. Dr. W. H. C. Ramsbottom has kindly con- firmed the identification of G. erenistria and has observed that it is G. crenistria crenistria, and not G. crenistria schmidtianus (see Nicolaus 1963), that is involved. The crenistria Bed itself has few conodonts. On the other hand, it was discovered during acid preparation that lightly silicified ‘ghosts’ of ostracode valves are especi- ally abundant in this calcilutite. Higher horizons in the Westleigh Limestones produce faunas which include Gnathodus commutatus nodosus. A particularly good fauna has been obtained from the sec- tion on the upper bench (eastern side) of East Whipcott Quarry : Gnathodus bilineatus, G. commutatus commutatus, G. commutatus homopunctatus, G. commutatus nodosus, G. girtyi subspp. (BU 22442). This can be dated as being no older than cuIIIjS of Germany (cf. Meischner 1971u, fig. 2). Similar faunas occur in the highest stratigraphy available in the old Fower Quarry (see text-fig. Ic ) at Westleigh. None of them have given any sign of reworked material. The highest parts of the Westleigh stratigraphy obviously deserve further MATTHEWS AND THOMAS: CARBONIFEROUS CONODONTS 381 attention. Whipcott Quarry, abandoned when the sample listed above was taken, is now active again. When the section on the eastern side of the quarry is stable, it should be possible to integrate conodont, goniatite, trilobite, and brachiopod evidence there. REWORKED CONODONTS IN THE WESTLEIGH LIMESTONE Since the faunas collected from horizons about the crenistria Bed are obviously mixed, there would be little point in making counts of individuals— no indication of the proportions by number of any original, ‘natural’ associations of forms can have survived. Also, since any of the specimens might have travelled some distance before final burial, the faunal lists above do not necessarily stand as records of ‘basin-associated’ conodonts. The evidence of reworking can be used as a basis for comment on the source from which the conodonts were derived. It should be said first that the fact of redeposition is already clear in the lithological evidence. The limestones are of mixed character and include coarse ‘shelf’ debris (a particularly good example has been found in Stout’s Cottage Quarry, where one bed contains a clast, 10 x 10 x 10 cm approx., of the colonial coral Lithostrotion arachnoideum). There are also some signs of internal reworking— clasts of the red shaly interbeds are common, and at Westleigh, where the crenistria Bed is clearly recognizable, clasts of this same distinctive calcilutite occur 2 m higher in the succession, in a conglomeratic limestone. The conodonts, however, suggest that reworking might have been operating on a much greater stratigraphic scale than this. An estimate of the maximum range of stratigraphy that might have contributed conodont material to the Westleigh succession can be constructed as follows : 1. The mix of conodonts found in the coarse limestone immediately underlying the crenistria Bed at Westleigh includes Gnathodus bilineatus, a form which can be regarded as occurring at its proper position in the succession. 2. G. bilineatus may be taken to indicate approximate equivalence with the Hotwells Limestone (‘D Zone’) of the ‘Avonian’ shelf succession. Note that this is a somewhat uncertain matter: Rhodes, Austin, and Druce (1969, p. 46) observed that G. bilineatus is available near the base of D2 both on the North Crop of the South Wales Coalfield and in the Avon Gorge; but on p. 95 and fig. 51 of their paper they made no record of having found G. bilineatus in the Bristol stratigraphy (and, incidentally, the range- information given on p. 95 of Rhodes, Austin, and Druce 1969 would suggest that G. bilineatus is confined to the highest unit — the G. girtyi collinsoni Zone — of their zonal scheme, although elsewhere in their paper (pp. 44, 45) they observe that G. bilineatus is one of the characteristic forms in each of the two zones below — their Mestognathus beckmanni-Gnathodus bilineatus Zone and G. mononodosus Zone). The suggestion that G. bilineatus might occur in the Hotwells Limestone rests at present on no better basis than Rhodes, Austin, and Druce’s information. Eventu- ally, it may prove to occur somewhat lower. 3. The coarse limestone immediately below the crenistria Bed at Westleigh includes also such forms as Scaliognathus anchoralis, which is found in the Black Rock Lime- stone of the eastern Mendips, and siphonodellids such as can be found in the upper 382 PALAEONTOLOGY, VOLUME 17 part of the Lower Limestone Shale and the lower part of the Black Rock Limestone there (Butler 1973). The estimate of the range of stratigraphy that contributed conodonts to the Westleigh Limestone could therefore be set at 800 m— approximately the thick- ness that exists between the top of the Lower Limestone Shale and the base of the Hotwells Limestone in the Mendip Hills. This maximum estimate refers to the shelf situation. It is, of course, imaginable that sedimentary material due to be added to the Westleigh succession acquired some of its admixed conodonts from sources closer to the eventual point of final burial at Westleigh itself; but, because of the lithological evidence, shelf situations deserve examination as a possible source. Also, if one looks to the conodont evidence itself, one can observe that the admixed cono- donts encountered at Westleigh are in almost every case forms that are now known from the ^ \\on\diri’ — Doliognathus latus has not yet been discovered there, but Butler (1973) records a Doliognathus sp.— and certainly include no form (Devonian, say, or earliest Carboniferous) that is unlikely to have been available in the shelf succession. WESTLEIGH LIMESTONE AND HELLEFELDER KALK The Lower Carboniferous succession in the Sauerland region of Germany includes a limestone development, 100 m thick at maximum (Helmkampf 1969), which has been mapped for some 20 km along the strike. It is known as the Westenfelder Kohlenkalk (Helmkampf 1969) or the Hellefelder Kalk (Gauglitz 1967; Meischner \91\b). It may be compared with the Westleigh Limestone in several respects. The two occupy approximately the same range of Dinantian age. They are set among what are generally fine-grained and to some extent siliceous basinal sediments. The limestones can in each case reach coarseness above sand grade, and the ranges of bed thicknesses are comparable. In each case there are thin, reddish, shaly inter- beds, and chert, often nodular, is common near the bases of individual thick lime- stones. The Hellefelder Kalk (see Gauglitz 1967, fig. 10), like the Westleigh Limestone (see above), gives clear evidence of internal erosion, and in both cases reworked conodonts are found. The two developments of limestone are of the same facies- type, and are of much the same age; but they are in one respect entirely different. The Westleigh Limestone, it seems, acquired much of its substance from a shelf situated generally to the north. The Hellefelder Kalk was derived from southerly sources. Its carbonate sediment (‘pre-sorted’ according to Meischner \91\b) was recycled from reefs of Devonian age in the Attendorn neighbourhood. These reefs, originally grown near the shelf margin of late Middle and early Upper Devonian times, supplied carbonate sediment for delivery northward into the more extensive basin of early Carboniferous time. The Hellefelder Kalk includes among its examples of reworked conodonts Devonian as well as Lower Carboniferous forms. The West- leigh Limestone sediment, derived from an entirely different source but eventually assuming the same set of facies-characteristics, carried with it only Lower Carboni- ferous conodonts such as might be found in the Carboniferous Limestone. TEXT-FIG. 2. At left, association of English and German proposals on Visean goniatite zones, following Weyer 1972. At right, the Westleigh (cr. Bd. = crenistria Bed) and the Bampton succession, placed according to goniatite information (no attempt made to indicate thicknesses). At centre, ranges of conodonts (following Meischner’s 1971fl association of conodont ranges and goniatite zones); upper left, conodonts regarded as being approximately in place; lower right, conodonts regarded as reworked. 384 PALAEONTOLOGY, VOLUME 17 SYSTEMATIC NOTES The reworking must be assumed to have upset the original sets of variants of par- ticular forms. There is therefore no case for attempting a formal systematic discus- sion of the conodonts found. It is hoped that the photographic illustrations will be taken as sufficient justification of the names entered in the faunal lists above. The species listed are in most cases interpreted as by German authors (Bischoflf 1957; Voges 1959; Meischner 1971a) since this makes for convenient reference to German stratigraphic interpretations. Gnathodus texanus texanus and G. texanus pseudo- semiglaber are interpreted as in Thompson and Fellows (1970). One name used in the lists above deserves special discussion. Gnathodus sp. indet. Plate 51, figs. 1, 2, 3, 18, 19 Remarks. Gnathodids listed by this name have an outer platform which is of variable width and is ornamented by nodes. In slimmer forms there are relatively few nodes on the posterior part of the outer platform. The outer platform is highest near its anterior margin and is never as broad and flat as in G. bilineatus (Roundy), nor does its margin have the clear posterior-lateral angle common in that species. The outer platform of G. sp. indet. better resembles that of G. semiglaber Bischolf, but the inner side of the platform is more parapet-like than would be typical of G. semiglaber. This inner parapet is well developed beside an anterior part of the carina, but does not accompany the carina to the posterior end. In this respect G. sp. indet. again dilfers from G. bilineatus. In Boogaert (1967) G. delicatus Branson and Mehl is interpreted in a way that would admit specimens of the present kind (see Boogaert 1967, pi. 2, figs. 14, 15). However, the broadly developed forms of G. delicatus are normally expected to show a better-developed parapet than is present in Boogaert’s figured specimens, and often have a line of nodes accompanying the posterior course of the carina on its outer side. Certain of the slimmer forms of G. sp. indet. show some resemblance to G. typicus Cooper. Thompson and Fellows (1970), who have studied and refigured Cooper’s holotype, suggest that the gnathodid identified as G. typicus by Boogaert (1967, pi. 2, fig. 21) might better belong in their G. texanus pseudosemiglaber . It is likely that the same judgement would be made of certain of the specimens figured as G. typicus by Marks and Wensink (1970, pi. 4). If so, the validity of Marks and Wensink’s (1970, pp. 249-250) G. typicus Zone, proposed as following immediately above the anclwralis Zone, would come into doubt. Acknowledgements. The authors are grateful to Dr. W. H. C. Ramsbottom, Mr. M. Mitchell, and Dr. C. L. Forbes for their comments on some of the macrofossils. Mrs. B. McCafFery (Calgary) and Mrs. M. J. Rogers (Bristol) helped in preparation and discussion of the conodont faunas. The photographic illustrations are the work of Mr. E. W. Seavill. REFERENCES Hi.sc'HOFi', G. 1957. Die Conodonten-Stratigraphie des rhenoherzynischen Unterkarbons mit Beriick- sichtigung der WocklumeriuSiuk und der Dcvon/Karbon-Grenze. Abh. Itess. Landesamt. Bodenfor.sch. 19, 7 64, pis. 1-6. MATTHEWS AND THOMAS: CARBONIFEROUS CONODONTS 385 BOOGAERT, H. A. VAN A. 1967. Devonian and Lower Carboniferous conodonts of the Cantabrian Mountains (Spain) and their stratigraphic application. Proefschrift, University of Leiden, 129-192, pis. 1-3. BUTLER, M. 1973. Lower Carboniferous conodont faunas from the eastern Mendips (England). Palae- ontology, 16, 477-517, pis. 56-59. GAUGLITZ, R. 1967. Der Hellefelder Kalk (sogenannter ' Kohlenkalk') am Siidrand der Nuttlarer und Liiden- scheider Mulde im Zusammenhang mit einem alters-gleichen Konglomerat-Horizont am Nordostrand der Elsper Kulm-Mulde. University of Gottingen, 66 pp. GOLDRING, R. 1955. The Upper Devonian and Lower Carboniferous of the Pilton Beds in N. Devon. Senck. leth. 36, 27-48, pis. 1, 2. 1970. The stratigraphy about the Devonian-Carboniferous boundary in the Barnstaple area of north Devon, England. Congr. Avanc. Etud. Stratigr. carb. Sheffield 1967, II, 807-816. HELMKAMPF, K. 1969. Zur Sedimentpetrographie und Stratinomie des Westenfelder Kohlenkalks (Sauer- land). Fortschr. Geol. Rheinld (Vest/'. 16, 473-528, 7 pis. JOPLING, A. V. and WALKER, R. G. 1968. Morphology and origin of ripple-drift cross-lamination, with examples from the Pleistocene of Massachusetts. J. sedirn. Petrol. 38, 971-984. KUENEN, ph. H. 1964. Deep sea sands and ancient turbidites. In bouma, a. h. and brouwer, a. (eds.), Turbidites. pp. 3-33, Elsevier, London, Amsterdam, New York. MARKS, p. and WENSiNK, H. 1970. Conodonts and the age of the ‘Griotte' limestone formation in the Upper Aragon Valley (Huesca, Spain), I. Koninkl. Nederl. Akad. Wetens. Ser. B 73, 238-275, pis. 1-4. MATTHEWS, s. c. 1970. A ncw cephalopod fauna from the Lower Carboniferous of east Cornwall. Palae- ontology, 13, 112-131, pis. 25-28. and NAYLOR, D. 1973. Lower Carboniferous conodont faunas from southwest Ireland. Ibid. 16, 335- 350, pis. 35-38. MEISCHNER, D. 1971fl. Conodonten-Chronologie des deutschen Karbons. Congr. Avanc. Etud. Stratigr. carb. Sheffield 1967, III, 1169-1180. 1971b. Clastic sedimentation in the Variscan geosyncline east of the river Rhine, hr. Sedimentology of parts of Central Europe (Guide book, 8th International Sedimentological Congress), 9-43. NICOLAUS, H.-j. 1963. Zur Stratigraphie und Fauna der crenistria-Zont im Kulm des Rheinischen Schieferge- birges. Beih. Geol. Jb. 53, 1-246, pis. 1-18. PIPER, D. j. w. 1972. Turbidite origin of some laminated mudstones. Geol. Mag. 109, 1 15-126. PRENTICE, J. E. 1960. The Dinantian, Namurian and Lower Westphalian rocks of the region southwest of Barnstaple, north Devon. Q. Jl geol. Soc. Lond. 115 (for 1959), 261-289. and THOMAS, J. m. 1960. The Carboniferous goniatites of north Devon. Abstr. Proc. Conf. Geol. Geomorph. S. W. England, R. geol. Soc. Cornwall, Penzance, 6-9. 1965. Prolecanitina from the Carboniferous rocks of north Devon. Proc. Yorks, geol. Soc. 35, 33-46. RHODES, F. H. T., AUSTIN, R. L. and DRUCE, E. c. 1969. British Avonian (Carboniferous) conodont faunas, and their value in local and intercontinental correlation. Bull. Br. Mus. nat. Hist. (Geol.), 5, 4-313, 31 pis. THOMPSON, T. L. and FELLOWS, L. D. 1970. Stratigraphy and conodont biostratigraphy of Kinderhookian and Osagean rocks of southwestern Missouri and adjacent areas. Missouri Geol. Surv. Wat. Resour. Rept. Invest. 45, 3-263, pis. 1-9. voGES, A. 1959. Conodonten aus dem Untercarbon I und II (Gattendorfia- und Pericyclus-Stufe) des Sauerlandes. Paldont. Z. 33, 266-314, pis. 33-35. 1960. Die Bedeutung der Conodonten fur die Stratigraphie des Unterkarbons I und II (Gattendorfia- und Pericychis-Stufe) im Sauerland. Fortschr. Geol. Rheinld West. 3, 197-288. WEBBY, B. D. and THOMAS, J. M. 1965. Whitsun held meeting: Devonian of west Somerset and Carboniferous of northeast Devon. Proc. Geol. Assoc. 76, 179-193. WEYER, D. 1972. Trilobiten und Ammonoiden aus der Entogonites nasutus-Zone (Unterkarbon) des Biichenberg-Sattels (Elbingeroder Komplex, Harz). Teil I. Geologic, Jg. 21, 166-184. Typescript received 15 February 1973 S. C. MATTHEWS Geology Department University of Bristol Queen's Building, University Walk Bristol, BS8 ITR J. M. THOMAS Geology Department University of Exeter North Park Road Exeter, EX4 4QE I V <7 AN APPARENTLY HETEROSPOROUS PLANT FROM THE MIDDLE DEVONIAN OF NEW BRUNSWICK by HENRY N. ANDREWS, PATRICIA G. GENSEL, and WILLIAM H. FORBES Abstract. A new plant of Lower or Middle Devonian age is described from Chaleur Bay, New Brunswick, as a new genus and species, Clialeuria cirrosa. It consists of a main axis bearing closely spiralled monopodial branches which m turn bear spirally arranged dichotomous ultimate branchlets. Pairs of sporangia terminate some ultimate branch- lets. Macerations of sporangia yielded spores of two sizes (30-48 fxm and 60 156 (um) which also differ from one another in shape and ornamentation. Our evidence suggests that the sporangia contained predominantly large or predominantly small spores or a mixture of both. We suggest that the spores of Chaleuria are sufficiently distinct to be considered as megaspores and microspores and that the plant presents evidence of an apparently early stage in the evolution of heterospory. I T is our purpose here to describe well-preserved specimens of a compression fossil representing a new genus of vascular plants, Chaleuria cirrosa, from the Middle Devonian of New Brunswick, Canada. It presents, to the best of our knowledge, the earliest evidence of heterospory based on megafossil remains that has been reported to date. The rocks exposed along the shorelines of Gaspe Bay, Quebec, and the north and south shores of the lower Restigouche River (Quebec and New Brunswick respectively) have long been known to be rich in fossil plant remains. Lower, Middle, and Upper Devonian horizons all being represented. The cliffs in some places are being eroded rapidly, exposing new plant deposits each year and there are many miles of shoreline that merit frequent prospecting. The collecting trips that we have conducted in the past four years (1970-1973) have yielded several new plants and more complete data on some previously described ones. It is evident that the area holds great potential for increasing our knowledge of Lower and Middle Devonian plant life. LOCALITY AND STRATIGRAPHY A rather detailed notation on the locality from which our plant was collected is necessary for reasons that will be evident. In July 1972 we devoted several days to prospecting and digging along a few hundred yards of beach outcrop on the south side of the lower Restigouche River, as shown in the accompanying map (text-fig. 1). At a point one-quarter of a mile west from the crossing of the railway and high- way No. 11 at Dalhousie Junction there is a narrow cart road that descends to the beach in a westerly direction. From this point extending westward for about one- half mile there are several outcrops along the bank immediately above the normal high-tide mark. We obtained an abundance of fossil plant material here, the most [Palaeontology, Vol. 17, Part 2, 1974, pp. 387-408, pis. 52-57.] 388 PALAEONTOLOGY, VOLUME 17 TEXT-FIG. 1. Map of Dalhousie Junction area. New Brunswick, showing collection localities. Specimens of Chaleuria were obtained from Locality A. (From Oak Bay and Escuminac maps of Restigouche County, New Brunswick, Department of Natural Resources.) productive spots being shown as ‘A’, ‘B’, and ‘C’ on the map in text-fig. 1. Chaleuria was found at point ‘A’. In the Chaleur Bay area map (map 330A) that accompanies Alcock’s ‘Geology of Chaleur Bay’ (1935) this half-mile strip is included in the Bonaventure Formation of Carboniferous age. It has been recognized, however, that this is probably in- correct (personal communication from Hugo R. Greiner, University of New Bruns- wick, to William H. Forbes). The light brown to grey sandy shales of this narrow coastal strip are in striking contrast to the coarse red sandstone of the surrounding Bonaventure Formation. As to the fossils, in addition to Chaleuria, we found Psilo- phyton, Drepanophycus, Kaulangiophyton, and at least one other plant that probably represents a new genus. The assemblage indicates an age not younger than Middle Devonian. It may also be noted that the spores of Chaleuria are comparable to spores reported from both Lower and Middle Devonian strata in North America and Europe. In addition we have found in our macerates the spore genus Emphani- sporites which is known to be extremely abundant in the Lower Devonian and present in the Middle Devonian; ours are similar to E. rotatus, a common Lower Devonian spore. ANDREWS ET AL.: DEVONIAN HETEROSPOROUS PLANT 389 On the Escuminac Sheet map of Bonaventure County, Quebec (Canadian Geo- logical Survey map 286A, for 1936), there is a patch of the Middle Devonian Gaspe sandstone outcropping along the coast about three-quarters of a mile east of Dal- housie Junction and two other small patches between Pin-sec-Point and Peuplier Point. Since the particular area that we are concerned with (text-fig. 1) is, for the most part, densely covered with vegetation down to the beach it is not surprising that it was overlooked in Alcock’s account of 1935. Thus, on the basis of the lithology and the fossil plant content, we consider the strip as belonging to the Campbellton Formation of the Gaspe sandstone and not younger than Middle Devonian. At the point marked ‘A’ on our map the rock consists of a fine to medium-grained, grey to greenish-grey argillaceous sandstone; it is generally well sorted and small quantities of muscovite are present. The beds dip at an angle of about 40°, the base of the exposure being about 2 feet above the usual high-tide mark. We dug away sufficient soil and vegetation to expose several square yards of fresh rock and made what seemed to be an adequate collection through a thickness of about 18 inches. There was apparently much more fossil-bearing rock available below, and extend- ing upward, into the forested bank. Only two megafossils were present : Chaleiiria cirrosa in great abundance and a few scattered fragments that probably represent a species of Psilophyton. DESCRIPTION OF CHALEURIA CIRROSA, NEW GENUS AND SPECIES General morphology. Two representative specimens of Chaleuria cirrosa are shown in Plates 52 and 53. Judging from such specimens the plant consisted of a main axis about 10 mm in diameter which was upright and has not been observed to dichoto- mize. These two specimens are slightly curved and the axes are found in some blocks to be aligned in more or less parallel rows. There is a suggestion that they may have been borne by a rhizomatous stem as, for example, in Hyenia elegans as recently described by Schweitzer (1972). We can only report that no such organ has been found and we will therefore refer to specimens such as those shown in Plates 52 and 53 as composed of a main axis bearing primary branches. The main axis in most specimens is broken at both ends so that we do not know the maximum height that was attained. The greatest diameter of the main axis is a little over 1 cm, it shows little tapering, and it is our estimate that the plants attained a height of about 1 m. The main axis bears abundant primary branches in a very close spiral; numerous specimens were degaged to a greater or less degree and in all cases it was found that the primary branches are attached all around the main axis (text-fig. 2). The primary branches are either entirely sterile (presumably photosynthetic) or entirely fertile; in some specimens the fertile branches are found at the lower part of the main axis (PI. 53, arrow) but since the latter is not complete no more exact statement can be made concerning the relative position of fertile and sterile branches. The sterile branches bear secondary (ultimate) branchlets, also in a close spiral, and they dichotomize once or twice with the terminations tending to recurve rather strongly (PL 54, figs. 1, 2). A few specimens show secondary branches which are monopodial and bear third-order ultimate branchlets. 390 PALAEONTOLOGY, VOLUME 17 The fertile primary branches (PI. 54, fig. 6 and text-fig. 3) appear very similar to the sterile ones and likewise bear ultimate branchlets (PI. 54, figs. 3-5) that dichoto- mize once or twice, and occasionally even three times, with a pair of sporangia at each ultimate, recurved branch ending. The sporangia are ovoid, measuring approxi- mately 2 0 to 2-3 mm long and 0-6 to 0-7 mm broad. The primary branches, sterile and fertile, are 2 to 3 mm in diameter at their proxi- mal end and, like the main axis, the tips are lost; the longest ones are about 9 cm, but since the distal end is broken they may have been considerably longer. We have no positive evidence of dichotomy in the primary branches ; if this did take place it was quite rare. The ultimate branchlets are up to 8 mm long and their axes 0-5 to TO mm wide. The restoration drawings (text-figs. 2, 3) give our impression of a representative part of the plant as it appeared in life. It certainly had a much denser or ‘bushy’ appearance than is gained from an examination of a freshly exposed specimen. As has been explained for the restoration drawings of Pertica quadrifaria (Kasper and Andrews 1972), only a small portion of a three-dimensionally branching plant of this kind is initially exposed. Sporangia and Spores. The sporangia and their contained spores present the most significant aspect of our plant. Sporangia were degaged from fertile specimens which contained only Chaleuria. The first sampling included sporangia from several specimens while the second and third included sporangia from only one specimen. The sporangia selected for removal were attached to the plant and were removed with as little matrix as possible. All of the samples yielded essentially identical results. The rock fragments containing sporangia were treated with hydrochloric acid, washed, and then macerated in hydrofluoric acid. The macerate was washed and the liberated sporangia and spore masses were extracted with a pipette, cleared in a mixture of potassium chlorate and nitric acid, washed and treated with dilute ammonium hydroxide. They were washed again and mounted on slides with a 1 : 1 mixture of Turtox CMC-9 and CMC- 10 non-resinous mounting medium. Some sporangia and most spore masses were mounted on slides immediately after clearing and washing as the spore masses blackened when treated with the base. The sporangia were not obviously affected by the chemicals employed, but the spores showed vary- ing degrees of corrosion, some probably caused by fossilization and some due to maceration. The sporangia extracted from the macerate ranged from whole ones to small fragments. Many apparently intact sporangia broke apart during the clearing process. The spore masses varied from large fragments with pieces of carbonized EXPLANATION OF PLATE 52 Chaleuria cirrosa. Specimen showing erect main axis and closely spaced, spirally arranged primary branches. Arrow indicates fertile ultimate branchlet borne near base of a primary branch. This is enlarged in PI. 54, fig. 5. Type specimen No. 33258, x I . All specimen numbers are those of the Canadian Geological Survey. PLATE 52 ANDREWS et al., Chaleuria cirrosa 392 PALAEONTOLOGY, VOLUME 17 TExr-i iG. 2. Chaleuria ciiro.sa. Restoration ol a representative part of the plant showing the general habit ANDREWS ET AL.: DEVONIAN HETEROSPOROUS PLANT 393 TEXT-FIG. 3. Chaleuria cirrosa. Restoration of part of a plant showing the densely spiralled arrangement of the primary branches and portions of one fertile and one sterile primary branch in detail. material adhering to them, and resembling fragmentary sporangia, to nearly un- carbonized spore masses of varying sizes. After clearing with Schulze’s solution some spore masses were teased apart with a fine needle in order to better observe the spores. A few intact sporangia were also broken apart in the same manner; this proved difficult but a few sporangia were successfully teased apart and some of their spores spread out without being totally destroyed. As a supplement to the usual microscopic study, selected individual spores were also examined and photographed with a Zeiss photomicroscope using Nomarski interference-contrast optics, which allows examination of external exine features by reflected light. The sporangia. Plate 55, figs. 1 and 2 show sporangia that have discharged most of their spores before fossilization and reveal the form and wall structure quite well. The sporangia are fusiform to ovoid in shape and frequently curved (PI. 55, fig. 1). A few of them remained paired when removed from the macerate but they usually separated 394 PALAEONTOLOGY, VOLUME 17 while clearing. An unseparated pair, consisting of a whole sporangium and part of another one, is illustrated in Plate 55, fig. 2. The cells of the sporangium wall (PI. 55, figs. 1, 2) are polygonal and randomly arranged. Globules of what appears to be tapetal material are scattered over the wall cuticle and outline individual cells. Many of the sporangia are split open at the distal end or throughout the entire length, indicating that dehiscence occurred longitudinally, starting at the tip. No specialized annulus cells have been observed. Many sporangia or fragments of sporangia were recovered which retained their entire spore mass or varying portions of the original content ; in addition the macera- tion process left fragmentary spore masses from which the sporangium wall had more or less disintegrated. The entire spore masses usually did not clear satisfactorily, probably having been immature at the time of fossilization. Thus our information is based chiefly on partially filled sporangia and portions of spore masses where we are reasonably certain that the spores are mature. As the photographs reveal (Pis. 55-57) there is a remarkable variation in spore size, but for the most part they fall within two size ranges : some sporangia contain small spores 30 to 48 /xm in diameter (PI. 55, fig. 3) while others contain significantly larger spores ranging from 60 to 156 ixm in diameter (PI. 55, fig. 4). The sporangia containing predominantly small spores may include a few larger ones and the small spores may be present in sporangia with large spores; this kind of association was observed frequently enough to convince us that it is significant and not a matter of contamination. The small spores are morphologically similar to one another and they differ in both size and ornamentation from the larger ones. The larger (mega-) spores are more varied in both size and morphology than the smaller (micro-) spores; plotting the size distribution of 50 large spores (taken from several sporangia) reveals two peaks, one at 60 to 75 jxm and one at 120 to 130 jum. The two spore groups will be dealt with individually for a more detailed description. The large spores. The large spores (PI. 56, figs. 1-4), ranging from 60 to 156 ixm in diameter, are trilete and circular to subcircular in outline. The exine varies from 1 to 6 jum thick, the largest spores being thinner (1-3 ^ixm) than the smaller ones. Most of them possess distinct contact areas extending about two-thirds of the spore radius and delimited by curvaturae. The contact areas are either smooth or bear a reduced sculpture. The smaller spores have a very sinuous triradiate mark (PI. 56, figs. 1, 2) while some of the larger ones have slightly wavy to straight rays with vary- ing degrees of lip development (PI. 56, figs. 3, 4); the triradiate mark extends about one-half to two-thirds the spore radius and the rays are usually raised by folds 3 to 6 ixvn wide. The exine outside the contact areas on the large spores is sculptured with closely packed minute granules or rods (rarely cones) which are from 0-5 to 1-5 (um high and half as wide as high. EXPLANATION OF PLATE 53 CItaleuria cirrosa. Specimen showing general habit. Primary branch at arrow shows spirally arranged fertile ultimate branchlets. This is enlarged in PI. 54, tig. 6. No. 33259, x 1. PLATE 53 ANDREWS et al., Chaleuria cirrosa 396 PALAEONTOLOGY, VOLUME 17 The small spores (PI. 55, fig. 3; PL 56, figs. 5, 6; PI. 57, figs. 1, 2, 3, 5). These are 30 to 48 ju.m in diameter (50 specimens measured), are trilete, and subtriangular to triangular in outline. The exine is 1 to 2 ^m thick; it may be uniform or slightly thinner proximally. The triradiate mark is distinct and simple or bordered by narrow lips or folds 1 to 2 ;u.m wide; the latter may merge with the contact areas. The straight rays extend three-quarters to seven-eighths of the spore radius. The proximal face is smooth and often has distinct contact areas surrounded by folds or by a thicken- ing, either of which nearly coincides with the equatorial margin. A few spores observed in lateral compression exhibit very clearly a thickening delimiting the contact areas which can be interpreted as curvaturae perfectae. A small percentage of spores appear to lack curvaturae or distinct contact areas ; these often also have a uniformly thick exine and are more than 40 ^m in diameter. The exine outside the contact areas is sculptured with baculi, coni and rarely spinae; these elements are from 1-5 to 3 0 |U,m high and 0-5 to 1-5 /u.m wide at the base. The cones are broad-based, taper abruptly to fine points, and are occasionally biform, curved or anastomosing to form short ridges. The baculi are fine and straight. The sculptural elements are widely spaced, sometimes 2 to 3 [xm apart and the cones and baculi rarely occur together on the same spore. "Other' spores. Most of the small-spore masses include a few spores that are identical to the large type. In addition, two other kinds of larger bodies were observed. One is an apparent spore lacking a trilete mark but sculptured with anastomosing, wart- like elements tipped by small cones; one or two of these are constantly present in the small-spore masses. Another spore-like structure has been observed associated with both the small and large spores. These are large, thin-walled, transparent bodies (PI. 57, figs. 6, 7) that lack a trilete mark and several were observed with small spores inside; it is possible that these structures represent a persisting mother-cell wall. Dispersed spores. Although we are not primarily concerned with the palynology of the sediments a bulk maceration of several rock fragments was prep.ared chie% as an aid in interpreting the problematical spores found associated with our ‘small’ and ‘large’ types. For the most part we obtained only sporangial fragments and spores of Chaleuria. A few spores of Emphanisporites (cf. R. rotatus McGregor) were found and a reticulum fragment possibly referable to Brochotriletes. We feel EXPLANATrON OF PLATE 54 Figs. 1 -6. Chaleuria cirrosa. I , sterile ultimate branchlet with at least one dichotomy and strongly recurved tip. ‘B’ indicates primary branch; arrow points to first dichotomy of ultimate branchlet. No. 33258, x7. 2, sterile ultimate branchlet with one dichotomy visible. No. 33260, x 9. 3, fertile ultimate branch- let showing at least two divisions and a possible third one. Arrow points to a terminal pair of sporangia. No. 33261, x6-5. 4, two ultimate fertile branchlets; arrows indicate sporangial pairs. No. 33262, X 6-5. 5, fertile ultimate branchlet ; this is indicated by arrow in PI. 52. Lower arrow points to dichotomy of the branchlet; upper arrow indicates one sporangium of a pair. No. 33258, x 5. 6, enlargement of fertile primary branch shown by arrow in PI. 53. Several fertile ultimate branchlets bearing terminal pairs of sporangia are evident. No. 33259, x4. PLATE 54 ANDREWS et al., Chaleuria cirrosa 398 PALAEONTOLOGY, VOLUME 17 that this paucity of dispersed spores in the rock matrix strongly supports our sup- position that the ‘odd type’ spores contained in some sporangia are not contaminants. DIAGNOSTIC DATA CHALEURIA Andrews, Gensel, and Forbes, gen. nov. Diagnosis. Plants with erect main axis bearing closely spaced, spirally arranged pri- mary branches. Primary branches monopodial, either sterile or fertile, with closely spaced, spirally arranged dichotomous ultimate branchlets. Distal divisions of ulti- mate branchlets recurved ; those of fertile ones terminated by pairs of ovoid sporangia. Occasionally a third order of branching present. Sporangia ovoid, with morpho- logically distinct spores of two sizes. Chaleuria cirrosa Andrews, Gensel, and Forbes, sp. nov. Diagnosis. As in generic diagnosis. Plants with main axes up to 1 cm wide, probably up to 1 m tall in life. Sterile and fertile primary branches 2-3 mm wide at the base, up to 9 cm long. Ultimate branchlets 1-2 and rarely 3 times dichotomous, up to 8 mm high, individual axes 1 mm or less in width, tips strongly curved. Fertile ultimate branchlets with pairs of ovoid sporangia 2-2-5 mm long, 0-6-0-7 mm wide. Dehiscence longitudinal, no annulus. Sporangia with trilete spores of two sizes and distinct morphologies. Large spores 60-156 (xm in diameter, circular to subcircular in outline. Exine 1-6 /xm thick. Contact areas distinct to barely discernible, smooth or with reduced sculpture; where distinct, delimited by curvaturae perfectae. Distal exine sculptured with closely packed grana or baculi 0-5- 1-5 yum high. Triradiate mark i to | spore radius, straight to sinuous, simple or with lips. Small spores 30-48 /u,m in diameter, triangular to subtriangular in outline. Exine 1-2 thick, usually thinner proximally. Proximal hemisphere smooth, usually with contact areas delimited by folds or a thickening nearly coinciding with the equator. Distal hemisphere and equatorial margins sculptured with baculi or broad-based, some- times biform coni. Sculptural elements 2-3 /xm apart, 1-3 /xm high. Triradiate mark simple, with lips or with folds 1-2 /xm wide, rays extend | to | spore radius. Derivation of the name. The generic name Chaleuria is taken from the general locality, Chaleur Bay, New Brunswick. The specific epithet cirrosa is derived from the Latin word cirrosus meaning ‘full of curls’ in reference to the recurved tips of the ultimate branchlets. Deposition of specimens. The type specimens and other specimens and slides illus- trated will be deposited in the collections of the Canadian Geological Survey, Ottawa. EXPLANATION OF PLATE 55 Figs. 1-4. Chaleuria cirrosa. 1, a cleared sporangium showing cellular structure of wall, stalk (above), and split-open tip (below). No. 33264, x 45. 2, pair of sporangia, one only partially intact. No. 33264, X 45. 3, part of a mass of predominantly small spores. No. 33265, x 280. 4, part of a spore mass from a sporangium containing the large spores of variable size. Arrow indicates spores in the 60-80 {.im range. No. 33263, x280. PLATE 55 ANDREWS et al., Chaleun'a cirrosa 400 PALAEONTOLOGY, VOLUME 17 Holotype. Specimen No. 33258 — Plate 52, fig. 1. Faratypes. Specimen No. 33259— Plate 53, fig. 1. Slide No. 33264— Plate 55, figs. 1, 2. Slide No. 33265-Plate 56, figs. 5, 6; Plate 57, figs. 1,2; Plate 55, fig. 3. Locality. Beach outcrop three-quarters of a mile west of Dalhousie Junction, New Brunswick, Canada. Point ‘A’ on map in text-fig. 1. Horizon. Campbellton Formation, Gaspe Sandstone; probably Middle Devonian. DISCUSSION Comparison with other megafossils. There are few Devonian megafossils with which we are able to draw a very close comparison with Chaleuria. Hyenia elegans Krausel and Weyland, as described by Schweitzer in 1972, presents some comparable features. This lower Middle Devonian plant from the Rhineland consists of a rhizome about 4 cm in diameter with upright shoots borne all around the rhizome in a close phyl- lotaxy; the aerial shoots branch quite profusely developing a crown-like pattern which is unlike our plant. The ultimate fertile appendages seem nearly identical to those of Chaleuria but the sterile ones are not. Preserved spores are not reported in Schweitzer’s account. Arctophyton gracile, described by Schweitzer (1968) from the lower Middle Devonian of Spitzbergen, is quite closely comparable to Chaleuria in its general habit and in the morphology of the sporangiate appendages; the ultimate sterile appendages of Arctophyton are in part somewhat more profusely branched. It is unfortunate that the sporangial contents of this Arctic plant are not preserved. Comparison with dispersed spore genera. Large spores: the large Chaleuria spores fit the emended concept of the dispersed spore genus Apiculiretusispora Streel (Streel 1967), but are not identical with any of the presently described species. Chaleuria spores like the one illustrated in Plate 56, fig. 3 have some similarities with A. brandtii Streel (1964) and A. plieata (Allen) Streel (1967). A. brandtii differs, for the most part, in having a darkened apical area and simple trilete mark, and A. plieata is smaller and has a thinner exine. The large Chaleuria spores like those illustrated in Plate 56, figs. 1, 2 are strikingly similar to Cyclogranisporites flexuosus Playford (1962), from the Lower Carboni- ferous. Similar spores have been described from the Lower and Middle Devonian by Schultz (1968) and Lanninger (1968). Both C. flexuosus and the large Chaleuria spores referred to above are of similar size and all of them exhibit a raised sinuous EXPLANATION OF PLATE 56 Figs. 1 -6. Chaleuria cirrosa. All figures at x 640. Figs. 1 4. Large spores, showing variation in size and morphology. These are not from one sporangium but comparable spores have all been observed in one sporangium. 1, 2, large spores with distinct con- tact areas and sinuous triradiate rays. 1, No. 33267. 2, No. 34597. 3, large spore with less prominent contact areas and straighter rays. No. 33266. 4, thin-walled, probably corroded, large spore with barely discernible contact areas. Sporangium wall overlies spore. No. 33264. Figs. 5-6. Small spores. 5, proximal view showing curvaturae. 6, distal view of same spore. No. 33265. PLATE 56 ANDREWS et ai, Chaleuria cirrosa 402 PALAEONTOLOGY, VOLUME 17 triradiate mark, a distal sculpture of small elements (grana) and, for the most part, distinct contact areas which are either smooth or bear a reduced sculpture. A small difference between spores of C. flexuosus and the large Chaleuria spores is that in the former the contact areas are not always as distinct or curvaturate as in the large Chaleuria spores. A single spore described and illustrated by de Jersey (1966) from the Middle Devonian of Australia as Retusotriletes sp. cf. R. devonicus Naumova is quite similar to some of the large Chaleuria spores. The largest of the large Chaleuria spores which do not possess distinct contact areas or curvaturae agree with the concept of the genus Cyclogranisporites. Small spores : the small spores of Chaleuria resemble several dispersed spore genera but many of the spores observed fit most closely the concept of the genus Streeli- spora Richardson and Lister (1969) in possessing an equatorial crassitude which delimits contact areas, distal sculpturing of various kinds of elements and a smooth proximal face. S. newportensis (Chaloner and Streel) Richardson and Lister (1969) is most similar to the small Chaleuria spores but differs from them in possessing papillae and folds on the proximal facets, more robust-appearing cones and in lacking baculi. Streelispora is very close to, or possibly congeneric with, the genus Aneurospora Streel (1964). At present the nature of the curvaturae of the two separates them, the curvaturae of Streelispora corresponding to an equatorial crassitude and the curva- turae of Aneurospora being still interpreted by Streel (1972) as a denser, more rigid part of the exine. The small Chaleuria spores differ at present from Aneurospora therefore in the nature of their curvaturae and also in having a thinner proximal exine and a thicker distal one. The ornamentation of A. goensis, the species that is most closely comparable to the small Chaleuria spores, is also finer than that of the small Chaleuria spores. The small Chaleuria spores in which curvaturae are not apparent resemble in some aspects the genus Procoronospora Butterworth and Williams. P. dumosus (Staplin) Smith and Butterworth (1967) appears similar to the small Chaleuria spores but has an ornamentation of fairly long spines and a uniformly thick exine. A spore identified as cf. Procoronospora, from the Emsian of Ontario, Canada, in McGregor et al. (1970) is quite similar to some of the small Chaleuria spores but a description was not included. Another, cf. Procoronospora sp., is illustrated in McGregor and Owens (1966) from the Emsian of the Gaspe which also appears similar to the small Chaleuria spores. Most of the small Chaleuria spores differ from the concept of EXPLANATION OF PLATE 57 Figs. 1 -7. Chaleuria cirrosa. Figs. 1-4. Spores photographed with Nomarski interference-contrast optics. All at x 1600. 1, 2, small spores illustrated in PI. 56, figs. 5, 6. Proximal and distal focus respectively with sculpture visible especi- ally well along margins. 3, small type spore with corroded exine, but showing smooth proximal face and curvaturae. No. 34596. 4, large spore illustrated in PI. 56, fig. 1, showing part of contact areas. Figs. 5-7. Spores found within a single sporangium. No. 33263. All at x 280. 5, two small spores. 6, 7, large, problematical thin-walled bodies (spore mother cell wall?). Several small spores are present within in Fig. 7. PLATE 57 ANDREWS et al., Chaleuria cirrosa 404 PALAEONTOLOGY, VOLUME 17 Procoronospora in possessing curvaturae, and in not always exhibiting the charac- teristic absence of ornamentation at the radial margins. Finally, some of the small Chaleuria spores resemble Anapiculatisporites burnotensis Streel (1967) except that the latter lacks curvaturae. In summary, a comparison of possibly related megafossils and dispersed spores with Chaleuria cirrosa suggests that the latter is not younger than Middle Devonian in age. The evidence for heterospory in Chaleuria. Two sizes of morphologically distinguish- able spores have been described for Chaleuria that have been found in cleared and uncleared sporangia, in spore masses, and also have been removed from sporangia by teasing with a needle. Our evidence suggests that these two sizes of spores are either intermixed in a sporangium or occur separately in different sporangia. As noted above the only megafossils that were associated with our collections of Chaleuria cirrosa were a few sterile axes that might belong to a Psilophyton species. On this basis it is unlikely that the sporangia and spore masses that we have described belong to any other plant. Several of our specimens are fertile and in our initial investigation of the sporangia we degaged them from several specimens; some of the sporangia thus obtained were observed to be attached while others may have been buried in the matrix. Upon macerating them the results were obtained as described above. As a check we studied a second series of macerations in which all of the sporangia were obtained from a single specimen. It is not certain that all of these came from one plant but they did come from fertile branches of Chaleuria. A third, and seemingly conclusive, bit of evidence that all of the described material is cor- rectly attributed to the plant, lies in the sporangia themselves. Although there are some slight size differences they are all essentially identical in form and cell wall structure. Sussex (1966) has prepared a very informative discussion on ‘The origin and development of heterospory in vascular plants’ and it is therefore unnecessary for us to review the subject in detail; we will consider only such facets of the matter as seem pertinent to an understanding of Chaleuria. We would, of course, like to know what happened to the spores of Chaleuria when they germinated. Although this cannot be done directly, we do have what seems to be a closely comparable phenomenon in a living fern, Platyzoma microphyllum Robert Brown. The evolution of heterospory has taken place in numerous pteri- dophytic lines and at different times in the geologic past ; it appears in several groups in the late Devonian and Lower Carboniferous (Andrews 1961 ; Chaloner 1967 ; Sussex 1966) but the case of Platyzoma suggests that it is not a biological event that is con- fined to the distant past. P. microphyllum is a Queensland plant that has attracted the attention of several investigators during the past century. Although originally placed in the Gleicheniaceae by Brown it has more recently been assigned to a new subfamily (Platyzomatoideae) of Christensen’s Polypodiaceae by Tryon (1961, 1964). Of special interest in the present context are features of its sporangia. These vary in size and in the orientation of the annulus and the contained spores show a considerable size range (Thompson 1916, 1917); it seems pertinent to consider them in some detail. ANDREWS ET AL.\ DEVONIAN HETEROSPOROUS PLANT 405 In his 1917 account Thompson notes that ‘The large sporangia contained small numbers of large spores, the small sporangia larger numbers of small spores’ (p. 158). In two tables on pages 162 and 163 he shows that the sporangia that contain pre- dominantly small spores may contain a few of the large and/or intermediate size spores; likewise those sporangia that contain predominantly large spores may also contain a few of the small and/or intermediate ones. In her more recent study Tryon (1964) gives a range in size for the small spores of 71 to 101 [xm and she showed that these produce a filamentous gametophyte-bearing antheridia ; the large spores range from 163 to 183 fxm and produce larger, spatulate gametophytes that bear arche- gonia. In a personal communication she has informed us that she also observed spores intermediate between the small and large ones but information is not avail- able on the nature of the gametophytes they produce. The lycopods present another analogy that seems relevant and deserves at least a brief comment. In the Carboniferous members of the group the well-preserved cones of Lepidostrobus and Lepidocarpon show the most complete sequence that we have in the evolution of heterospory. Starting with a lower Mississippian species there is a reduction in the number of megaspores per sporangium, an increase in the size of the remaining megaspores, and in Lepidocarpon, where only one functional spore remains, the megasporangium is partially enclosed by the sporophyll to form a seed (Andrews 1961, pp. 231-232). In the living Selaginellas there is considerable variation, in several species, in the number and size of the megaspores per sporangium (Duerden 1929). For example, in Selaginella willdenowii Baker, Duerden reports a range of from one to as many as 42 megaspores in a sporangium. One of his summary comments seems to have a very direct bearing on what we have found in Chaleiiria : ‘The occurrence in Selaginella of megasporangia containing many com- paratively small spores, suggests a condition possibly not far advanced beyond the homosporous state, and, on the other hand, the sporangia with fewer, comparatively large spores, indicates an advance in the direction of the seed habit’ (p. 456). To the best of our knowledge Chaleuria is the oldest megafossil that is sufficiently well preserved to understand its gross morphology and in which the sporangia and spores present adequate evidence to indicate heterospory. A few other plants deserve discussion which seem to support our stand and which contribute to this very impor- tant phase in vascular plant evolution. One of the oldest genera in which hetero- spory is known is Barinophyton and in his review (Pettitt 1970) of the early evidences of heterospory Pettitt cited B. ricliardsoni (Pettitt 1965) from the Frasnian (lower Upper Devonian) of Maine and B. citrulliforme (Arnold 1939) from the Famennian (upper Upper Devonian) as being heterosporous. He also notes (1970, p. 404) that plants with a habit similar to Barinophyton have been reported from sediments older than the Upper Devonian but either the spores are not known or they have not been shown to be heterosporous. Pettitt also briefly mentions (1970, p. 405, pi. 4, fig. 2) an undescribed Emsian (upper Lower Devonian) plant from New Brunswick in which the spores range from 97 to 240 ixm. No other information is available and judging from his figure the plant is not closely related to Chaleuria. He offers the following comment on the New Brunswick plant in question: ‘Clearly, an evolu- tionary situation in heterospory which we would expect to find in a particular line would be an incipient condition presaging a later and more definite distinction 406 PALAEONTOLOGY, VOLUME 17 between microspore and megaspore’ (p. 405). We do not offer an opinion on the plant that Pettitt’s comment is based on but Chaleuria seems to fit very nicely this concept of incipient or primitive heterospory. Several other fossil plants present apparently comparable primitive stages in heterospory. Protopitys scotica Walton, according to Walton’s description (1957), bears sporangia with three kinds of spores ; in some sporangia the spores are about 82 nxm, others contain spores that are 147 ju.m, while a few have spores that are 98 /xm. In a later report on this plant Smith (1962a) concluded that the spores fall within two, rather than three, sizes but there is, nevertheless, a considerable range, from 75 to 355 |Lim. In another Lower Carboniferous plant {Staphylotheca kilpatrickensis) Smith (1962a) described two kinds of spores, one ranging from 42 to 57 jum and the other ranging from 78 to 100 /xm in diameter; these two, along with a third ‘type’ that is regarded as immature, occur in the same sporangium. He does not mention the possibility of heterospory. In a third lower Carboniferous plant, Alcicornopteris hallei Walton, Smith (\962b) described four different kinds of spores, the differences being more in gross morphology and exine structure than in size, with the overall range being 50 to 130 jtxm. These selected examples of both living and fossil plants, representing several diver- gent groups, afford a considerable amount of consistent evidence bearing on the evolution of heterospory. Apparently the segregation of two distinct types of spores, relative to size and sexual differentiation, evolved gradually. The smaller spores germinate to form male gametophytes and the larger ones to form female gameto- phytes and to date we have little information on the function of the intermediate types. Accordingly, we feel that it is highly probable that the same pattern applies to Chaleuria cirrosa, that the small and large spores produced male and female gameto- phytes respectively, and we suggest that some of the intermediate ones formed bi- sexual gametophytes. The classification of Chaleuria cirrosa. We are not able to fit Chaleuria satisfactorily into any existing scheme of classification. The problem is aggravated by a common one in palaeobotany, that is, lack of data comparable with those available in other plants. Several otherwise rather well-known Lower and Middle Devonian plants might bear comparison except they lack the preservation of the sporangiate organs that we do have in Chaleuria. Probably the closest comparison that we can make on the basis of general mor- phology is with Arctophyton gracile Schweitzer from the Middle Devonian of Spitz- bergen. Schweitzer refers to it as ‘. . . probably a fern or a “progymnosperm” related to Aneurophyton . It is our opinion that the ‘main stream’ of plant evolution leading to the Pterido- sperms runs in a general way from very simple plants such as Psilophyton through increasingly complex types such as Pertica, Rhacophyton, and Archaeopteris. We do not imply a straight-line series but rather suggest these genera as showing stages in the evolution of a laminate megaphyll, the cambium and heterospory. It is now clear that the early land plants of the Lower Devonian evolved into a rather diverse assemblage by Middle Devonian times. Some of the latter ran into ANDREWS ET AL.\ DEVONIAN HETEROSPOROUS PLANT 407 blind alleys or evolved into true ferns (Filicales and Marattiales), others led to the problematical and heterogeneous Coenopterids, and others led toward the Pro- gymnosperms, which we are just beginning to understand. On the basis of apparent heterospory we tentatively regard Chaleuria as a primitive member of the Pro- gymnosperms. Acknowledgements. For financial support of our research programme grateful acknowledgement is made to the National Science Foundation (Grant GB-3548X) and the University of Connecticut Research Foundation. Thanks are also due to Mary Hubbard, staff artist, for her care in rendering the restoration drawings. REFERENCES ALCOCK, F. J. 1935. Geology of Chaleur Bay Region. Mem. geol. Stirv. Can. 183, 146 pp. ANDREWS, H. N. 1961. Studies in Paleobotany. John Wiley and Sons, New York. 487 pp. ARNOLD, c. A. 1939. Observations on fossil plants from the Devonian of eastern North America. IV. Plant remains from the Catskill Delta deposits of northern Pennsylvania and southern New York. Contrib. Mils. Paleont. Univ. Mich. 5, 271-313, 10 pis. CHALONER, w. G. 1967. Spores and land-plant evolution. Rev. Palaeobot. Palynol. 1, 83-93. CHANDLER, M. E. J. 1964. The Lower Tertiary Floras of Southern England. IV. A summary and survey of findings in the light of recent botanical observations. Brit. Mus. Nat. Hist. 151 pp., 4 pis. DE JERSEY, N. J. 1966. Devonian spores from the Adavale Basin. Pubis, geol. Surv. Qd. (Palaeontological Papers No. 3), 334, 1-28, 10 pis. DUERDEN, H. 1929. Variations in megaspore number in Selaginella. Ann. Bot. 43, 451-457. KASPER, A. E. and ANDREWS, H. N. 1972. Pertica, a new genus of Devonian plants from northern Maine. Am. J. Bot. 59, 892-911. LANNiNGER, E. p. 1968. Sporen-Gesellschaften aus dem Ems der SW Eifel (Rheinisches Schiefergebirge). Palaeontographica. B 122, 95-170, pis. 20-26. MCGREGOR, D. c. and OWENS, B. 1966. Devonian spores of eastern and northern Canada. Pap. geol. Surv. Can. 66-30, 1-66, 29 pis. SANFORD, B. v. and NORRIS, A. w. 1970. Palynology and correlation of Devonian formations in the Moose River Basin, Northern Ontario. Proc. geol. /I55. Can. 22, 45-54, 2 pis. PLAYFORD, G. 1962. Lower Carboniferous Microfloras of Spitsbergen. Part I. Palaeontology, 5, 550-618, pis. 78-87. PETTiTT, JOHN M. 1965. Two heterosporous plants from the Upper Devonian of North America. Bull. Brit. Mus. (nat. Hist.) (Geol. & Miner.), 10, 83-92, 2 pis. 1970. Heterospory and the origin of the seed habit. Biol. Rev. 45, 401-415, 6 pis. RICHARDSON, J. B. and LISTER, T. R. 1969. Upper Silurian and Lower Devonian spore assemblages from the Welsh Borderland and South Wales. Palaeontology, 12, 201-252, pis. 37-43. SCHULTZ, G. 1968. Eine Unterdevonische mikroflora aus den Klerfer Schichten der Eifel (Rheinisches Schiefergebirge). Palaeontographica, B 123, 5-42, pis. 1-4. SCHWEITZER, H.-J. 1968. Pflanzenreste aus dem Devon Nord-Westspitsbergens. Ibid. B 123, 43-75, pis. 5- 1 5. 1972. Die Mitteldevon-Flora von Lindlar (Rheinland). 3. Filicinae-//vcm'a elegans Krausel and Weyland. Ibid. B 137, 154-175, pis. 30-39. SMITH, A. H. v. and butter worth, m. a. 1967. Miospores in the coal seams of the Carboniferous of Great Britain. Spec. Pap. Palaeont. 1, 324 pp., 27 pis. SMITH, D. L. 1962fl. Three fructifications from the Scottish Lower Carboniferous. Palaeontology, 5, 225-237, 2 pis. 19626. The spores of Alcicornopteris hallei Walton. Ann. Bot. N.S. 26, 261-211 , 1 pi. STREEL, M. 1964. Une association de spores du Givetian Inferieur de la Vesdre. Ann. Soc. geol. Belg. 87, 1-30, 2 pis. 1967. Assoeiations de spores du Devonien Inferieur Beige et leur signification stratigraphique. Ibid. 90, 11-54, 5 pis. 1972. Dispersed spores associated with Leclercquia complexa Banks, Bonamo and Grierson from the late Middle Devonian of eastern New York State (U.S.A.). Rev. Palaeobot. Palynol. 14, 205-215, 2 pis. 408 PALAEONTOLOGY. VOLUME 17 SUSSEX, I. M. 1966. The origins and development of heterospory in vascular plants. In e. g. cutter (ed.), Trends in Plant Morphogenesis. John Wiley and Sons, New York. Pp. 140-152. THOMPSON, J. M. 1916. The anatomy and affinity of Platyzoma microphvUum R. Br. Trans, row Soc. Edinb. 51, 631-656. 1917. A further contribution to the knowledge of Platyzoma microphyllum R. Br. Ibid. 52, 157-165. TRYON, A. F. 1961. Some new aspects of the fern Platyzoma microphyllum. Rhodora, 63, 91-102. 1964. Platyzoma— A. Queensland fern with incipient heterospory. Am. J. Bot. 51, 939-942. HALTON, J. 1957. On Protopitys (Goeppert): with a description of a fertile specimen "Protopitys scotica' sp. nov. from the Calciferous Sandstone Series of Dunbartonshire. Trans, roy. Soc. Edinb. 68, 333-340, 3 pis. HENRY N. ANDREWS PATRICIA G. GENSEL University of Connecticut Storrs, Conn., U.S.A. WILLIAM H. FORBES University of Maine Typescript received 30 March 1973 Presque Isle, Maine, U.S.A. THE SILURIAN TRILOBITE ONYCOPYGE WOODWARD by D. J. HOLLOWAY and k. s. w. Campbell Abstract. New material of Onycopyge liversidgei Woodward, from the Upper Silurian m the vicinity of the type locality at Quidong, New South Wales, permits the clarification of the morphology of the genus and its assignment to the Subfamily Deiphoninae. Another occurrence of the genus in Ludlovian rocks at Canberra is reported. Two types of pygidia at Quidong suggest that the species may be sexually dimorphic. Evidence is produced in support of the view that members of the Deiphoninae were benthonic. The trilobite Onycopyge liversidgei Woodward has been known only from a single specimen collected late last century by Professor A. Liversidge from a locality cited by Woodward (1880) as in ‘the Silurian rocks of Bombala, New South Wales’. However, there are no Silurian rocks at Bombala, the nearest ones being at Quidong, 22 km to the west. New material of this species has recently been discovered in that area from the Ludlovian Quidong Formation (Brunker et al. 1971), which consists of an unknown thickness of mudstones, siltstones, and limestones. None of the specimens are in the same matrix as the type specimen. Associated with the new material is a rich and varied assemblage of brachiopods, bivalves, gastropods, trilobites, and ostracods, of which the most abundant elements are Howellella nucula (Barrande), Atrypa sp., Molongia sp., Brachyprion sp., Salopina sp. aff. S. lunata (J. de C. Sowerby), Leptaena depressa (J. de C. Sowerby), Anodontopsis australis Etheridge, and Encrinurus sp. aff. E. mitchelli Foerste (Tassell 1972). The specimens are in a hard siltstone, much of which has been naturally etched to yield internal and external moulds. The material of Onycopyge is completely disarticu- lated and often broken, and no hypostome, rostral plate, or thoracic segments have been found. Librigenae are mostly separated from cranidia, although some in- complete cephalons have been found. Often the bulbous part of the glabella is crushed or flattened but otherwise there is little apparent distortion. The holotype (B.M. I 107) is a rather poorly preserved, almost complete individual with the exoskeleton of the thorax preserved. The thoracic axis has been destroyed, the pygidium badly broken, and the cephalon has been broken away to expose the external mould of the antero-ventral surface of the glabella and the ventral surface of the hypostome. How much of this Woodward understood is not clear. His reconstruction suggests that he mistook the hypostome for the posterior part of the glabella, and hence believed that the cephalon was preserved in an inverted position. His footnote (p. 97) suggests that he belatedly appreciated his error, but did not alter his text, which still refers to the glabella as ‘attenuated behind’ (although the posterior part of the glabella is not preserved), and the small fixigenal and librigenal spines as ‘directed forwards and upwards’ (although they are directed slightly forwards and downwards). [Palaeontology, Vol. 17, Part 2, 1974, pp. 409-421, pi. 58.] 410 PALAEONTOLOGY, VOLUME 17 TEXT-FIG. 1. Localities at which Onycopyge liversidgei Woodward was collected at Quidong, near Bombala, N.S.W. Locality 1, grid reference FV803181; locality 2, FV813163, Bombala Sheet 8724, Australian 1 : 100,000 Topographic Survey. Whittard (1934) gave a much more satisfactory account of the species based only on the holotype which by that time had been better prepared. He also had access to the counterpart of the cephalon which showed the forms of both glabella and genae. It is his reconstruction that appears in Figure 341(10) of the Treatise on Invertebrate Paleontology (Moore 1959). Because of the vague locality given for the holotype, the fact that it is preserved in a ‘splintery limestone’ whereas our new material was found in siltstone, and the absence of so many important morphological features in the holotype, we have been in some doubt about the assignment of our specimens to O. liversidgei. However, those features that are preserved are entirely comparable and we consider it better practice to adopt a conservative approach. HOLLOWAY AND CAMPBELL: ONYCOPYGE WOODWARD 411 SYSTEMATIC DESCRIPTION Family cheiruridae Hawle and Corda, 1847 Subfamily deiphoninae Raymond, 1913 Genus onycopyge Woodward, 1880 Type species. Onycopyge Liversidgei Woodward, 1880, p. 98, from the Ludlovian at Quidong, N.S.W. Diagnosis. Glabella in front of lateral glabellar lobe Ip greatly inflated, unfurrowed. Glabellar lobe Ip reduced to a slight swelling lying in the junction of the axial furrow and the broad depression behind the inflated part of the glabella. Axial furrow deep, directed antero-laterally adjacent to the postero-lateral edge of the inflated part of the glabella where apodemes Ip and 2p lie close together, but shallows rapidly oppo- site the palpebral lobe. Fixigena triangular, extended into a strong genal spine which at its base bears a stout, ventro-laterally directed spine on its ventral surface. Palpebral lobe inflated, with downturned lateral border. Librigena small, triangular, bearing a broad, ventro-laterally directed spine opposite y. Posterior border of cephalon with a narrow articulating flange originating opposite the axial furrow and extending almost to an exsagittal line through S. Hypostome deeply convex, with subtrapezoidal middle body. Thorax of nine segments; pleurae consisting of an inner, unfurrowed portion with anterior and posterior articulating flanges, and an outer, free, backwardly curved, spinose portion, the two being separated by a low, knob-like tubercle on the dorsal surface. Pygidium large, with 8-10 axial rings plus a terminal piece; first ring highly convex (sag. and tr.), more strongly curved anteriorly than the more posterior ones. First and second segments extended into strong, backwardly curved spines; postero-lateral border with a third pair of short, postero-ventrally directed spines; pleural region behind the second segment with variably developed pleural bands. Remarks. In the original description Woodward (1880) noted the similarity between Onycopyge and Deiphon. Whittard (1934) realized that it also had a close relation- ship to Sphaerocoryphe, although he did not consider the stratigraphic ranges of the genera when he suggested that Deiphon ‘could have evolved almost directly from either Onycopyge or Sphaerocoryphe’’ (p. 526). Opik (1937) first placed Onycopyge in the Subfamily Deiphoninae together with the other two genera, and since then several authors have followed this approach (Prantl and Pfibyl 1948; Hennings- moen, in Moore 1959). Lane (1971) removed Onycopyge from the Deiphoninae and in fact did not even regard it as a cheirurid. His reasons were threefold— his belief that the swollen part of the glabella comprises the frontal lobe only; the presence of a transverse furrow near the posterior border of each thoracic segment, which he maintained is not found in any cheirurid; and the larger number of axial rings in the pygidium. The first two of these reasons are based on misconceptions. As is shown below, the swollen part of the glabella does contain lobe 3p and at least part of 2p. The transverse furrow on the thoracic pleura is the one that separates the pleural band from the posterior articulating flange, and is present on most cheirurids, including members of the Deiphoninae. Lane (1971, pi. 13, figs. 1, 5, la, b, and 12) has illus- trated it for Sphaerocoryphe, and our investigations show that it is also present on Deiphon, albeit in a much reduced (tr.) form. In fact Onycopyge has both anterior and posterior articulatory flanges, features necessary for the type of articulation found in all cheirurids, in which there is no process-and-socket at the fulcrum but an anterior cavity is juxtaposed to a posterior cavity on the next segment. Although there is a process on the anterior edge of the segments of some genera (the flange process of Bergstrom 1973) this is really only the slightly expanded outer rim of the o 412 PALAEONTOLOGY, VOLUME 17 anterior cavity which, during enrolment, slips inside the rim of the cavity on the segment in front. This leaves only the number of rings in the pygidium to be considered. In his discussion of the taxonomy and phylogeny of the Cheiruridae, Lane placed con- siderable emphasis on the number of thoracic and pygidial segments, the family itself being diagnosed as having 9-12 thoracic segments and 1-4 pygidial rings. The Deiphoninae was defined as having 9 thoracic segments and 4 pygidial rings. Onycopyge is entirely comparable with members of this subfamily not only in gross morphology but also in such details as the pitting of the fixigenae, the shape of the librigenae, the presence of one secondary spine on the librigena and another on the fixigena, the segmental pattern of the glabella, the furrowing and articulation of the thoracic pleurae, and the development of large marginal spines from the first and second pygidial segments. It should also be noted that although Lane (p. 58) records only two pairs of pygidial spines on Sphaerocoryphe, Webby (1974) has described a species that has a third pair similar in position, shape, and orientation to that of Onycopyge, and Shaw (1968, pi. 13, fig. 23) figures similar spines on at least the early holaspids of S. goodnovi Raymond. Thus a very good case would have to be argued to exclude Onycopyge from the subfamily on the basis of the number of pygidial rings alone, especially as many other trilobite families are known to be very variable in this character (the Encrinuridae, for example) or to include an occasional genus with an unusual number of rings (the Cyclopygidae, for example). Lane (p. 58) suggested that Onycopyge may be related to the Encrinuridae, though no reasons for this assessment were offered. Woodward, in the original work, had noted a similarity to Staurocephalus, a genus that many authors include in the Encrinuridae. We are unable to detect any close similarities with any member of the Encrinuridae. Moreover, even using Lane’s criteria of number of thoracic and pygidial segments (which we cannot accept), it is difficult to see why Onycopyge is more closely related to the Encrinuridae than to the Cheiruridae. We conclude that the affinities of Onycopyge are with the Deiphoninae, and in fact see no reason why the genus should be excluded from this subfamily. This con- clusion requires that the definitions of the Cheiruridae and the Deiphoninae given by Lane (1971) be amended. Within the Deiphoninae, Onycopyge is more closely related to Sphaerocoryphe than to Deiphon. (We accept the view that Hemisphaerocoryphe Reed is probably a synonym of Sphaerocoryphe.) This judgement is based on similarities between Sphaerocoryphe and Onycopyge in the size of the genae, the secondary ventrally directed spines on the genae, the distance between the articulatory fulcra, the size and shape of the thoracic pleural spines, and the general pattern of the pygi- dium. Deiphon is unique in the degree to which the genae are reduced, the width of the articulatory flanges decreased, and the pygidium posterior to the first seg- ment modified. In view of the known stratigraphic ranges, these data imply the relationship shown below. The history represented by the dotted line remains unknown. HOLLOWAY AND CAMPBELL: ONYCOPYGE WOODWARD 413 Upper Silurian Onycopyge Middle Silurian Lower Silurian Upper Ordovician Onycopyge liversidgei Woodward, 1880 Plate 58, figs. 1-17; text-figs. 2a, b 1880 Onycopyge Liversidgei Woodward, p. 98. 1934 Onycopyge liversidgei Woodward, Whittard; p. 521, pi. 16, figs. 10, 11; text-fig. \b. Material. Holotype BM I 107. British Museum (Natural History) Collection. Other material 21668a-b, 28990a-b, 2899 la-b, 28992a-b, 28993a-f, 28994a-c, 28995a-b, 28996a-b, 28997a-b, 28998-28999, 29000a-d, 2900 la-b, 29002, 29003a-b, 29004a-b, 29005a-b, 29006a-b, 29007, 29008a-b, 29009-29012, 29014. Geology Department, Australian National University Collection. Description. Cephalon (excluding genal spines) with greatest width across the pos- terior border. Inflated part of glabella subspherical in dorsal view (for orientation see discussion below), rises vertically or sometimes slopes slightly backwards from the preoccipital depression, and steeply from the axial furrow just in front of apodeme Ip; anterior lobe extends beyond anterior border by about one-third its length. Anterior border furrow deeply impressed medially, giving a beak-like appearance to the anterior border in lateral view, but shallows laterally towards the axial furrow. Anterior border of cephalon vertical. Occipital ring about three-fifths the width (tr.) of the widest part of the glabella, strongly convex (tr. and sag.), sloping forwards to merge with the posterior part of the median glabellar lobe in the preoccipital depression (see discussion below), and curving downwards into apodemal pits at its antero-lateral borders. Occipital apodemal pits elongate, transverse; occipital furrow not clearly defined medially. Preoccipital depression very deep, arcuate, and U-shaped in section. Palpebral lobe elongated parallel to axial furrow, bounded posteriorly by a broad, shallow, rounded furrow originating about half-way between the axial furrow and e, deepening laterally and terminating in a rounded depression situated antero-laterally to the end of the posterior border furrow. Facial suture cuts the lateral margin behind the midpoint (exsag.) of the palpebral lobe and well in front of the posterior ventral spine, then swings back and on the ventral surface curves sharply around the base of the genal spine; m-e slightly sigmoid, rises sharply on to the palpebral lobe; e situated opposite apodeme 2p and below the downturned rim of the palpebral lobe; plane containing e-h-y slopes slightly downward anteriorly; y-(8 directed anteriorly, falling sharply away from the palpebral lobe; /S-a gently arcuate inwards, running subparallel to the lateral border so as to isolate a long, anterior, spine-like projection on the librigena, and meeting at a large angle with the rostral suture, which is approximately equal in width (tr.) to the occipital ring. Visual surface of eye crescentic, bounded below by a convex rim which, together with the downturned 414 PALAEONTOLOGY, VOLUME 17 5mm TEXT-FIG. 2. Reconstruction of the cephalon of Onycopyge liversidgei Woodward in dorsal view (a), and lateral view (b). EXPLANATION OF PLATE 58 Figs. 1-17. Onycopyge liversidgei Woodward, 1880. 1, latex cast of the broad type of pygidium, dorsal view, X 3 ; ANU 29006b. 2, 6, latex cast of the narrow type of pygidium, dorsal and dorso-lateral views, X 3; ANU 29003a. 3, 4, 8, latex cast of the narrow type of pygidium, ANU 29003b. 3, ventral view. The small projections on the doublure opposite the first pair of lateral spines are due to breakage of the original, x 3. 4, enlargement of part of 3 to show the surface ornament, x 9. 8, antero-ventral view showing the downturned ends on the anterior articulating flange, x 5. 5, 7, plaster replica of the holo- type, BM I 107. 5, dorsal view, x 1-5. 7, enlargement of part of 5 to show the anterior and posterior articulating flanges on the thoracic pleurae, x 5-5. 9, internal mould of the glabella, antero-ventral view, showing the rostral suture, x6-5; ANU 29011. 10, latex cast of the postero-lateral part of a cephalon, ventral view; free cheek in position but partly obscured; oc indicates the occipital apodeme; la, 2a indicate apodemes Ip and 2p respectively. Facial suture retouched, x4; ANU 29004a. 11, 13, 16, latex cast of the cranidium, ANU 28994c. 1 1, dorsal view, x 2. 13, antero-dorsal view, x 2. 16, enlargement to show the posterior articulating flange, palpebral lobes, and ornament on cheeks and glabella, x6-5. 12, latex cast of the cranidium, ventral view, x 3; ANU 29014. 14, latex cast of cheek, dorso-lateral view, showing the eye, x 3; ANU 29004b. 15, latex cast of the free cheek, dorsal view. The slender, spinose portion is directed anteriorly, x 6-5 ; ANU 28995b. 1 7, latex cast of the genal spine and part of the cheek, ventral view, showing the facial suture, x 3; ANU 28992a. PLATE 58 HOLLOWAY and CAMPBELL, Onycopvge Woodward 416 PALAEONTOLOGY, VOLUME 17 rim on the palpebral lobe, gives the impression that the visual surface is surrounded by a convex border. Posterior border strongly convex (exsag.), gently arched transversely, expanding slightly towards the genal angles, and lower than the surface of the librigena. Posterior articulating flange originates as a short (exsag.) ridge behind the axial furrow, expand- ing slightly until it is about one-half the length (exsag.) of the posterior border distally, and terminates with a hook about two-thirds the way along the posterior border. Posterior border furrow relatively long (exsag.), rounded in section, folded down- ward into the occipital apodemal pit proximally, expanding slightly in the vicinity of the base of the genal spine where it terminates, not meeting the lateral border furrow. Lateral border furrow very shallow, appears to terminate in the depression at the end of the post-ocular furrow. Lateral border strongly convex, high, increasing sharply in height at the base of the genal spine. Genal spine arcuate in both dorsal and lateral profiles. Glabella with pustules that are coarsest medially in the postero- dorsal region and above the anterior border; each major pustule with a secondary one at its apex. Genae ornamented with shallow pits in a triangular region bounded posteriorly by a low, smooth ridge adjacent to the posterior border furrow, and by fainter ridges parallel to the axial and palpebral furrows. Genal spines and posterior border densely and uniformly granulated. Hypostome (described from cast of holotype) with middle body longer (sag.) than wide, expanding anteriorly, and with a faint, arcuate middle furrow originating in the lateral furrow about three-quarters of the distance from the front of the middle body, delineating a crescentic posterior lobe. Maculae not observed. Neither anterior nor posterior border preserved on the holotype, but the narrow posterior furrow remains. Anterior wings poorly preserved, but lie opposite the anterior margin of the middle body. Surface of hypostome obliterated. Thorax (described from cast of holotype) with nine segments. Thoracic axis narrow, approximately one-sixth the total width of thorax, tapering slightly posteriorly. Axial rings not preserved. Inner portion of pleurae gently convex (exsag.), nearly horizontal, straight (tr.), parallel-sided, with very wide (tr.) anterior and posterior articulating flanges which are slightly downturned at their distal ends. Spinose portion of pleurae a little longer than articulated portion, flattened in section, gently tapered, directed downwards at about 30° to the plane of the inner portion, and curving uniformly postero-laterally. Surface of pleurae granulated, similarly to the genal spines. Pygidium, apart from the marginal spines, subpentagonal in outline; axis bullet- shaped in outline; axial furrow clear, slightly impressed at the ends of the axial rings, but transgressed by the ring furrows. First axial ring inclined slightly forwards, expanding slightly towards the axial furrow, and bounded behind by a sharply incised furrow; articulating half ring approximately as long (sag.) as the ring proper; articulating furrow clearly defined, well-rounded at its base. Second axial ring of similar dimension to the first, and with a strongly developed pseudo-half ring that is distinguishable across the entire width of the axis. Next four or five rings of approxi- mately the same sagittal length, but subsequent ones rapidly diminishing; pseudo- half rings weakly developed on the third, fourth, and fifth rings of some specimens. Anterior articulatory flange on the pleura of the first segment increasing gradually in length (exsag.) distally; an unfurrowed pleural band expands rapidly towards HOLLOWAY AND CAMPBELL: ONYCOPYGE WOODWARD 417 the lateral border where it is humped and flexed, and then produced into a long, posteriorly-curving spine ; in lateral view, spine slightly flexed ventrally at first, then reflexed dorsally towards its extremity. Second segment with pleural band much more posteriorly directed, exhibiting a similar humping across the border, and pro- duced into a spine somewhat larger than that of the first segment. Between the first and second pleural bands is a furrow with a faint transverse swelling composed of the posterior articulatory flange of the first pleura and the anterior flange of the second. First two marginal spines almost circular in cross-section and slightly con- stricted on the ventral side where they join the pygidium proper. The more posterior segments with rapidly decreasing and progressively more posteriorly curving pleural bands separated by furrows of similar dimensions; the last two or three segments without clearly distinguishable pleural bands. A pair of short postero-ventrally directed marginal spines present near the posterior extremity; these spines ovate in cross-section. Border clearly defined, well-rounded, and bearing a slight medial terminal hump. Doublure narrows markedly towards the antero-lateral corner of the pygidium, slightly flexed outwards between the first two pairs of spines, and deeply embayed and arched postero-medially. Doublure and dorsal surface, apart from furrows, finely granulated. Apodemes well developed on all except the most posterior segments and are antero-medially elongated. Remarks. As far as can be determined from the material available the species is quite constant in the characters of the cephalon, but there appear to be two distinct forms of pygidia— one with a narrow axis and narrow pleural region, the other relatively much broader. The observed differences are real, and cannot be explained in terms of distortion. The narrow form has a strongly arched axis which is subtriangular in cross- section. The furrows on the pleural regions of segments 3-8 are short, later ones being strongly arched backwards so that the intervening pleural bands are tapered at their distal ends. The last furrow that retains a linear form is directed postero- laterally. The spine on the second segment is of about the same diameter as that on the first segment. The posterior spines are rather robust and postero-ventrally directed. The pygidium on the holotype, although only poorly preserved as a mould of the ventral surface, appears to be of this form. In the broad form the axis is broader, more gently convex, and more rounded in cross-section. The axial rings are proportionately shorter (sag. and exsag.). The furrows on the pleural regions of the segments behind the second are longer, the pleural region being broader, and the difference in curvature between earlier ones and those on later segments is not as marked. The last distinguishable furrow curves more strongly backwards at its distal end than the corresponding furrow on the narrow pygidium. The spines on the second segment of the broad form are slightly stouter than those on the first segment, and diverge at a greater angle than the corre- sponding spines on the narrow form. Furthermore, one specimen of the broad type (ANU 29006) has 10 axial rings rather than the usual 8. One other specimen of this type has 8, but it is of a rather small holaspid, while the only other specimen avail- able, although larger, is incomplete. Only 8 rings can be counted on this specimen but there could be more present. 418 PALAEONTOLOGY, VOLUME 17 All specimens of the broad type of pygidium are from the one locality, while at the other locality only the narrow type occurs. However, one specimen of the narrow type was recovered from the former locality. Whether these differences are due to sexual dimorphism or to the presence of two subspecies is uncertain. From Bohemia, Alberti (1971) has also recognized broad and narrow forms of the pygidium of Cheirurus (Crotalocephalus) cf. pauper Barrande, which he interprets as sexual dimorphs. It may also be of significance that Deiplwn barrandei Whittard has two types of pygidium that can be readily distinguished by the form and orientation of the posterior pair of spines. The differences have been attributed to size by Whittard (1934) and Lane (1971). The figured specimens fall clearly into one type or the other, there being no mention of intermediates. They possibly indicate dimorphism. Recently, distorted specimens of a cranidium and two pygidia have been recovered from the Ludlovian Riverside Formation at Canberra. The pygidia are of interest in that, although they represent small holaspids and are poorly preserved, they seem to have only six axial rings plus a terminal piece. MORPHOLOGICAL NOTES In the above description there are two points that require some explanation. 1. The furrow forming the posterior margin to the inflated part of the glabella in the Deiphoninae was interpreted by Lane (1971, p. 58) as ‘a combined occipital and IS furrow’. The relationship between the first glabellar furrows and the occipital furrow is not clear in Deiphon and Onycopyge, in which the glabella posterior to the inflated portion has been greatly shortened relative to that of some species of Spliaero- coryphe. In Sphaerocoryphe glabellar lobes Ip are small, subtriangular, slightly raised structures, isolated by furrows Ip that join with the occipital furrow medially, the latter occupying only the posterior part of the broad furrow behind the inflated por- tion of the glabella. This is quite obvious in many species of this genus figured by Shaw (1968) and Lane (1971) and it would seem to imply that at least part of the furrow in question was a depressed part of lobe 2p. The ideal way to test this implication would be through an ontogenetic study, but no complete growth stages of a deiphonine are known. It seems clear from other cheirurids, however, that the part of the glabella between the inner ends of the isolated lobes Ip is part of lobe 2p, and this is confirmed by a study of Aeanthoparypha ehiropyga Whittington and Evitt (Whittington and Evitt, 1954, pi. 28, figs. 31-44). In the protaspis of that species the glabellar furrows are more or less complete, but during the meraspid stages the furrows become incomplete medially, furrows Ip become oblique and almost join the occipital furrow to isolate lobes Ip. Lobe 2p expands posteriorly and comes into contact medially with the occipital furrow, as can be demonstrated by the progressively more posterior movement of the paired tubercles associated with the glabellar lobes in successive ontogenetic stages. In the case of the Deiphoninae, some support for this point of view is obtained from the meraspid of S. goodnovi Raymond figured by Shaw (1968, pi. 13, fig. 18), which is very similar to the meraspids of Aeanthoparypha. Hence the median portion of the depression forming the posterior margin of the inflated part of the glabella is neither a true glabellar furrow nor an expanded occi- HOLLOWAY AND CAMPBELL: ONYCOPYGE WOODWARD 419 pital furrow, but an occipital furrow plus part of a median glabellar lobe. In view of the fact that furrow 2p is not developed, the extent to which lobe 2p is involved can- not be determined. Nor is it always possible to distinguish the extent of the occipital ring. Hence we prefer to use the term preoccipital depression for the depressed medial part of the glabella in front of the occipital ring and behind the inflated lobe. Although the structures of this region are less clear in Deiphon and Onycopyge, there are such obvious similarities between these genera and Sphaerocoryphe that the same termi- nology should be applied. 2. For purposes of description and functional interpretation, it is necessary to decide upon the relaxed, unrolled posture of these organisms. The nature of the thoracic articulations, the shape of the axial rings, and the length of the articulating half rings, suggest that the thoracic pleurae and the pleural regions of the pygidium would have been more or less coplanar. Using the same arguments with respect to the cephalon, it is clear that the relaxed position would have been with the plane containing the posterior margin of the occipital ring approximately vertical. In this orientation the visual field would have an appreciable anterior component. These comments apply to Deiphon, Sphaerocoryphe, and Onycopyge. We have used the above orientation in our description. MODE OF LIFE Previous authors (Staff and Reck 1911; Ruedemann 1934; Whittard 1934) have regarded Deiphon as planktonic, basing this interpretation on features such as spinosity and the swollen glabella which, they argued, could have contained a low- density fluid to aid flotation. Whittard (1934) extended the interpretation to Ony- copyge and Sphaerocoryphe. Recent plankton have acquired a variety of adaptations to aid flotation, some of which— such as storage of oil droplets in the protoplasm— we can have no knowledge of in fossil material. However, those adaptations involv- ing variation in skeletal morphology can be identified in fossils. These adaptations, which are usually aimed at the increase of the ratio, can be summarized as follows: weight (1) the development of fine spines; (2) the reduction of the thickness of the skeleton; (3) the small size of the organism ; (4) the development of globose chambers to contain low-density body fluids ; (5) the development of all-round vision. Many of these comments are applicable also to entirely nektonic crustaceans, though some of these reach a much larger size. The question now is to determine the extent to which the members of the Dei- phoninae exhibit these characteristics. Deiphon itself certainly has developed a highly spinose skeleton, but whereas recent pelagic crustaceans become spinose by the production of extensions of the normal body, in Deiphon the spines result from skeletal reduction. For a trilobite of any given length this produces a decrease of surface area, and hence it should be compensated for by a proportionately greater decrease in weight. Whittard (1934, p. 527) was of the opinion that this was achieved. 420 PALAEONTOLOGY, VOLUME 17 It is true that there would be a large decrease in the weight of the soft tissue and the exoskeleton of the cephalon, but the totally enclosed pleural spines of both the thorax and the pygidium are proportionately increased in length, and this has a tendency to increase the weight of the exoskeleton per unit volume. Further, the exoskeleton remains quite thick— it is 0-35 mm thick at the base of a genal spine of maximum diameter 2-5 mm, and reduces to only 0-30 mm at a maximum spine diameter of 1-5 mm. In size, Deiphon is certainly one of the smaller members of the Cheiruridae, but adults commonly attain a length of 25 mm whereas the majority of recent plankton are less than a centimetre in length and many do not exceed 2-3 mm (Raymont 1963). The animal certainly had a wide visual field in both vertical and horizontal senses. Of course the glabella is inflated, and may have contained a substance of low density. The bulk of this evidence is against the hypothesis that Deiphon was planktonic, and the remainder is inconclusive or cannot be checked. None of it specifically favours a nektonic mode of life either, though the total design does not favour the view that it was an active swimmer. In fact, if it could be shown that the glabella was inflated to contain an enlarged stomach rather than low-density material, the remainder of the morphology would be consistent with a predominantly benthonic mode of life. Some evidence on this point can be obtained from a study of the hypo- stome, which is considered below. If the above arguments with respect to Deiphon are correct, then Sphaerocoryphe and Onycopyge, both of which are generally larger in size and have relatively un- reduced skeletons, are even less likely to be planktonic or nektonic. In fact there is little reason to believe that their mode of life was much different from that of members of the Cheirurinae. The hypostome of Deiphon is a well-calcified structure, being as thick as the dorsal exoskeleton. In its normal position (i.e. with no gap at the hypostomal suture) the hypostome is inclined at a high angle to the horizontal plane (see orientation discussion above). This angle is so great that complete enrolment of the thorax and pygidium would be impossible; yet the animal could enrol, as is shown by the development of articulatory devices and normal articulating half rings on all seg- ments, and by the embayments in the anterior border of the cephalon on either side of the hypostomal suture that received the posterior spines of the pygidium (Lane 1971, pi. 12, fig. 6u, b). The deep furrow across the anterior half of the hypostome probably also fitted into the doublure of the pygidium. This implies that the hypo- stome was able to rotate up under the glabella during enrolment. The rotational axis must have been through a point at each extremity of the hypostomal suture, a feature well brought out by the figures referred to above. Rotation would have been achieved either by the attachment of muscles to the anterior wings or from the body of the hypostome to the anterior part of the glabella (see Eldredge 1971). The effect of this rotation would be to compress the oesophagus and probably the stomach when the animal was enrolled. An explanation has then to be offered for the function of the hypostome in the extended position. If the hypostome were capable of rotating while the animal was not enrolled, the system could act as a pump for sucking food into the stomach. This offers the possibility of an explana- tion for the expanded glabella. It also suggests a function for the ventrally directed HOLLOWAY AND CAMPBELL: ONYCOPYGE WOODWARD 421 secondary spines on the genae. These together with the genal spines would have allowed the cephalon to rest sufficiently clear of the sea bottom to permit the hypo- stome to rotate. Presumably mud would have been ingested. Sphaerocoryphe and Onycopyge had hypostomes of basically the same type as Deiphon, and presumably they functioned in the same way. Acknowledgements. We thank Dr. D. L. Strusz who first brought the occurrence of Onycopyge in Can- berra to our attention; Mr. C. B. Tassell who contributed stratigraphic information from Quidong and, with Professor D. A. Brown, gave assistance in collecting; Dr. B. D. Webby who allowed us the use of his work on Sphaerocoryphe from New South Wales prior to its publication; Mr. S. F. Morris who sent casts of the type specimen of O. liversidgei from the British Museum; and Mr. L. Seeuwen who assisted with the photography. REFERENCES ALBERTI, G. K. B. 1971. Sexual-Dimorphismus(?) bei Cheirurus (Crotcdocephalus) cf. pauper Barrande 1852 (Trilobita, Devon). Paldont. Z. 45, 167-172, pi. 18. BEGG, J. L. 1940. A note on the genera Staurocephalus and Sphaerocoryphe, with the description of a new species of Sphaerocoryphe. Geol. Mag. 77, 295-304, pi. 4. BERGSTROM, J. 1973. Organisation, life, and systematics of trilobites. Fossils and Strata, 2, 69 pp., pis. 1-5. BRUNKER, R. L., OFFENBERG, A. c. and WEST, J. L. 1971. (Compilers.) Monaro Sheet, Australian 1 : 500,000 Geological Series. Geological Survey N.S.W. DAVIS, c. c. 1955. The Marine and Freshwater Plankton. 562 pp. Michigan State University Press. ELDREDGE, N. 1971. Patterns of cephalic musculature in the Phacopina (Trilobita) and their phylogenetic significance. J. Paleont. 45, 52-67, pis. 13, 14. LANE, p. D. 1971. British Cheiruridae (Trilobita). Palaeontogr. Soc. [Monogr.], 95 pp., pis. 1-16. MOORE, R. c. 1959 (ed.). Treatise on Invertebrate Paleontology. Part O, Arthropoda I. i-xix, 560 pp. Univ. Kansas Press. OPIK, A. A. 1937. Trilobiten aus Estland. Acta Comment. Univ. tartu. (A), 32 (3), 1-163, pis. 1-26. PRANTL, F. and PRIBYL, A. 1948. Classification of some Bohemian Cheiruridae (Trilobitae). Sh. ndr. Mas. Praze, 3, Geol. (Paleont.), 1, 1-44, pis. 1-6. RAYMONT, J. E. G. 1963. Plankton and Productivity in the Oceans. 660 pp. Pergamon Press, Oxford. RUEDEMANN, R. 1934. Paleozoic Plankton of North America. Mem. geol. Soc. Am. 2, 141 pp., pis. 1-26. SHAW, F. c. 1968. Early Middle Ordovician Chazy Trilobites of New York. Mem. N.Y. St. Mus. nat. Hist. 17, 163 pp., pis. 1-24. STAFF, H. VON and RECK, H. 1911. Uber die Lebensweise der Trilobiten-eine entwicklungsmechanische Studie. Sher. Ges. naturf. Freunde Berl. (1911), 130-147. TASSELL, c. B. 1972. The Geology of Quidong, N.S. W. 95 pp., pis. 1-12. B.Sc.(Hons.) thesis, A.N.U. (unpubl ). WEBBY, B. D. 1974. Upper Ordovician trilobites from central N.S.W. Palaeontology, 17, 203-252. WHITTARD, w. F. 1934. A revision of the trilobite genera Deiphon and Onycopyge. Ann. Mag. nat. Hist. ser. 10, 14, 505-533, pis. 15, 16. WHITTINGTON, H. B. and EViTT, w. R. 1954. Silicified Middle Ordovician Trilobites. Mem. geol. Soc. Am. 59, 137 pp., pis. 1-33. WOODWARD, H. 1880. Description of a new genus of trilobites, Onycopyge liversidgei, from the Silurian of New South Wales. Geol. Mag. 7, 97-99. D. J. HOLLOWAY K. S. W. CAMPBELL Department of Geology Australian National University Canberra, A.C.T. Australia 2600 Revised manuscript received 20 August 1973 SELECTIVE EPIZOAN ENCRUSTATION OF SOME SILURIAN BRACHIOPODS FROM GOTLAND by J. M. HURST Abstract. In a large collection of brachiopods from the Ludlow of Gotland, epizoan attachment is related to sur- face shell ornament and rib angularity. 'Camarotoechia niiciila has lateral inhalant canals which are indicated by Cornulites encrusting that part of the host shell. The distribution of Spirorbis suggests that 'Camarotoechia micida and Homoeospira baylei lived with the pedicle valve up, whilst Ptychopleurella bouchardi lived with the anterior commissure up. Over 500 brachiopods were collected from one horizon in the lower Eke Beds (middle Ludlow) of Laubackar, Eastern Gotland, Sweden, about 1-25 kilometres east-north-east of the church of Lau. The Eke Beds are represented by 15 metres of argillaceous limestones alternating with arenaceous marlstone (Hede 1960). The collection studied from Laubackar came from an arenaceous marlstone. TABLE 1 . Composition of encrusting fauna on the different Brachiopods. BTachiopod species Number of brachiopod individuals Brachiopods encrusted by Spirorbis V /o 96 Brachiopods encrusted by Cornulites V Brachiopods encrusted by Berenicea V Brachiopods with no epizoans y Ptychopleurella bouchardi 41 /o 3 /o 8 /o 4 'Camarotoechia' nucula 91 15 22 11 62 Homoeospira baylei 120 30 21 2 57 Delthyris elevata 93 2 3 3 93 Spinatrypa sp. 57 15 6 10 73 Glassia obovata 37 8 8 0 84 The most abundant species are (with percentage occurrence following): Ptycho- pleurella bouchardi (Davidson) 9T%; "Camarotoechia' nucula (J. de C. Sowerby) 28-3 %; Homoeospira baylei (Davidson) 40-8%; Glassia obovata (J. de C. Sowerby) 4-2%; Delthyris elevata (Dalman) 9-8%; and Spinatrypa sp. 6-6%. These six species have an encrusting fauna consisting of: (1) a serpulid worm Spirorbis aff. lewisii (J. de C. Sowerby); (2) a bryozoan Berenicea aflf. consimilis (Lamx); and (3) Cornulites serpuliformis (Vine). Until recently the systematic position of the cornulitids was unknown. Originally they were thought to be tubicular annelids (Nicholson 1872). Fisher (1962) drew attention to the resemblance of some cornu- litids to the coelenterates, or some fusulines. However, on the basis of the similarity of some aspects of the shell structure of both cornulitids and molluscs Blind (1972) indicates that they should be placed with the primitive Mollusca. A commensal relationship between the host brachiopod and the encrusting fauna is suggested by the following observations. (1) Very few cases have been found of Berenieea, Spirorbis, or Cornulites crossing the hinge line or the anterior commissue [Palaeontology, Vol. 17, Part 2, 1974, pp. 423-429.] 424 PALAEONTOLOGY, VOLUME 17 TEXT-FIG. 1. \, 'Camarotoec/lia' nucula with Cornulites serpuliformis (arvowed), x 24. 2, Ptychoplewella boucliardi, encrusted by Spirorbis aff. lewisii, x 20. 3, Homoeospira baylei with Coniulites serpuliformis (arrowed) overhanging anterior commissure, x 24. 4, Spirorbis lewisii, x 24. of the brachiopods. Presumably if they did they would have had a detrimental effect, as the host animal would probably have been unable to open its valves. (2) The only disarticulated specimens in the collection were valves of Ptychoplewella, Spmatrypa, and Delthyris, which represented 12% of the fauna. In one case only was there any encrusting organism on the interior of the valves. This was a small cornulitid which no doubt grew here after the death of the brachiopod. It is concluded that the great majority of the encrusting fauna grew on living brachiopods. If this were not so, both the interior and exterior of diarticulated specimens would have been equally encrusted. (3) In the case of Cornulites, specimens at the valve edge can be seen to overhang the shell margin. This is interpreted as the site where water flowed into the mantle cavity. A similar case was reported by Trueman (1942) for Spirorbis on non- marine bivalves in the Carboniferous. The encrusting fauna (Table 1) is largely restricted to three brachiopod species: ” v> A ' HURST: SILURIAN BRACHIOPOD EPIZOA 425 BRACHIOPOD RIB TYPE Ptychopleurella bouchardi Homoeospira baylei Camarotoechia nucula WvW Delthyris elevata Glassia obovata s m oot h Spinatrypa sp. minute ribs TEXT-FIG. 2. Brachiopod rib-types. (1) Ptychopleurella is heavily encrusted by Spirorbis and little else. (2) 'Camarotoechia and Homoeospira are encrusted by roughly equal amounts of Cornulites and Spirorbis, plus a scattering of the bryozoan Berenicea (text-fig. 1). The remaining brachiopods Delthyris, Spinatrypa, and Glassia have a low percentage of epizoans. Spirorbis. Spirorbis is the most widespread encrusting faunal element. It is abundant on Ptychopleurella and fairly common on ‘’Camarotoechm and Homoeospira (Table 2). The explanation for this distribution may lie in the way which Spirorbis initially attaches itself to, and grows on, the host shell. Small spirorbids are always situated between ribs, though with growth large spirorbids may cover the ribs. Ptychopleurella has low ribs separated by wide shallow grooves, a situation affording the maximum possible protection for serpulid growth. The ribs in Homoeospira are stronger and in ‘'Camarotoechia' they are extremely angular (text-fig. 2), Ptychopleurella has more INHALANT CURRENT SULCUS TEXT-FIG. 3. Division oV Camarotoechia nucula. 426 PALAEONTOLOGY, VOLUME 17 spirorbids attached to it than Homoeospira, which in turn has more than "Cama- rotoechia' (Table 2). Also the spirorbids attached to Ptychopleurella are on the whole far larger than those attached to other species; the spirorbids attached to coarsely ribbed forms may have been prevented from reaching maturity. It would appear that Spirorbis prefers to grow in between the host’s ribs because of the protection afforded therein, but can develop fully only if the ribs are not angular. Glassia is a completely smooth brachiopod, and Spinatrypa has little or no rib development. Thus these shells afford less preferential sites for spirorbid encrustation, as does Delthyris, which has angular depressions surrounded by low rounded ribs. TABLE 2. Distribution of Spirorbis on Brachiopods. Brachiopod species Number of Spirorbis on pedicle valve Number of Spirorbis on brachial valve Total Spirorbis Ptychopleurella bouchardi 112 119 231 ’’Camarotoechm nucula 24 8 32 Homoeospira baylei 31 16 47 Delthyris elevata 4 0 4 Spinatrypa sp. 6 8 14 Glassia obovata 2 1 3 Richards (1972) asserted that the distribution of epizoans on late Ordovician brachiopods was controlled by the size of the brachiopod and its ornament. He also concluded that steeper ribs are a better deterrent to epizoans, than gentle ones. The present study shows that these generalizations also apply at this Gotland locality. Richards also noted that Platystrophia produced a mat-like external shell layer of fine spines, which he thought probably acted as another deterrent to epizoans. But Ptychopleurella, a member of the closely related subfamily Glyptorthinae, has strong concentric imbrication disposed as frills or drawn out as spines, which clearly did not act as a deterrent to Spirorbis. Consequently, Spirorbis distribution was probably mainly controlled by the size and angularity of the brachiopod ribs. Cornulites. Cornulites is by no means as widespread or abundant an encrusting faunal element as Spirorbis (Table 3). In 'Camarotoechia' and Homoeospira, the majority of the cornulitids are arranged in the grooves, parallel to the ribs of the host shell. The steep ribs of 'Camarotoechia' and Homoeospira carry, in general, larger Cornulites than do the smooth or broad-ribbed brachiopods. Ptychopleurella, Delthyris, and Spinatrypa are not sharply ribbed forms, but they are characterized by strong growth lines which are often spinose. The spinose shell morphology is probably a major factor in inhibiting Cornulites from encrusting the shell (Richards 1972). The low density of encrusting Cornulites on Glassia is probably due to the fact the smooth shell offered no protection. Consequently, the distribution of Cornulites appears to be limited to sharply ribbed forms, which do not have spinose growth lines. Spinose forms are not a suitable site for Cornulites encrustation. Cornulites encrustation is accompanied by some absorption of calcium carbonate from the host (Fisher 1962); consequently, the most advantageous place for a cornu- litid to grow was in between the ribs of coarsely ribbed shells, as, not only is this HURST: SILURIAN BRACHIOPOD EPIZOA 427 TABLE 3. Distribution of ConmUtes on Brachiopods. Brachiopod species Number of Cornulites on pedicle valve Number of Cornulites on brachial valve Total Cornulites PtychopIeureUci bouchardi 0 0 0 ''Camarotoechia inicitla 23 22 45 Homoeospira baylei 27 13 40 Delthyris elevata 2 0 2 Spinatrypa sp. 2 3 5 Glassia obovala 2 1 3 position well protected, but it will allow greater absorption of calcium carbonate from the host brachiopod. Berenicea. This byrozoan occurs in such low numbers that it provides no conclusions (Table 1). DISTRIBUTION AND ORIENTATION OF THE ENCRUSTING FAUNA The distribution and orientation of the epizoans may be used to throw light on the possible life positions and feeding habits of the various brachiopods. '’Camarotoechid' may be divided into three roughly equal areas, a central fold and sulcus, bounded by two flanks (text-fig. 3). The number of ConmUtes on these areas was counted on both pedicle and brachial valves. Seventeen and twenty-one speci- mens of ConmUtes occurred on the right and left flanks of the brachiopod respectively, whilst only seven occurred in the fold and sulcus region. This distribution is not random as a 97% significance was obtained from a modified Chi-square test (Rey- ment 1971). This is probably related to the filter feeding system of the brachiopod, with the majority of the cornulitids situated above the lateral inhalant canals (text- fig. 3). In recent rhynchonellid brachiopods, the region of the fold is the site of the exhalant canals (Thomson 1927; Rudwick 1960). Thus the preferential distribution of Cornulites on the lateral flanks of Silurian ‘'Camarotoechid conforms with the water movement seen in modern rhynchonellids. The orientation of the cornulitids, on the lateral flanks of "Camarotoechia\ is in accord with the findings of Schumann (1967). He observed commensal Cornulites preferentially attached to the lateral flanks of Mucrospirifer reidfordi, from which he concluded that the cornulitids bene- fited by feeding on the hosts inhalant currents. Hoare and Steller (1967) described a series of epifaunal elements on brachiopods in the Devonian Silica Formation of north-western Ohio. Examination of 3,105 specimens of different genera revealed that the epifaunal elements were distributed unequally in the same way as the Gotland collection. They also give a detailed description of epifaunal host relationships in a specimen of the Devonian brachiopod Paraspirifer brownockeri associated with a boring sponge, Cornulites, an articulate brachiopod and bryozoa. They note that Cornulites is orientated in an anterior direction, their direction of growth controlled by the grooves in the shell, and conclude that this relationship appears to be com- mensal with Cornulites benefiting from the current action of the host. The distribution of Cornulites on Homoeospira was tested, but was found to be a totally random one. p 428 PALAEONTOLOGY, VOLUME 17 TEXT-FIG. 4. Life position of "Camarotoechia nucula (a) and Homo- eospira baylei {b). Spirorbis has a totally different commensal association with the brachiopods, as it is not related to the inhalant or exhalant canal system. On "Camarotoechia and Homoeospira, Spirorbis prefers the pedicle valve. If Spirorbis preferred clearer water, to the ejected waste, it would attach itself to the parts of the brachiopod furthest away from the substrate, so that its feeding system is not polluted by mud. Con- sequently, Spirorbis encrustation on Homoeospira and ‘'Camarotoechia'’ may indicate the life position of these brachiopods. The heavily encrusted pedicle valve was prob- ably uppermost (text-fig. 4). By contrast the even distribution of Spirorbis on both pedicle and brachial valves of Ptychopleurella suggests that this brachiopod lived with both valves inclined at a high angle to the substrate (text-fig. 5). Ager (1961) found that Spirorbis had a general distribution on the Devonian spirifer Spinocyrtia iowensis. From this he concluded, as is the case here, that Spirorbis had ho relation to the feeding habits of the host. TEXT-FIG. 5. Life position of Ptychopleurella bouchardi. CONCLUSIONS 1 . Cornulites preferred to grow in the grooves of strongly ribbed brachiopods. 2. Spirorbis shows a preference for broad-ribbed brachiopods like Ptychopleurella. 3. ‘'Camarotoechia' has lateral inhalant canals which are indicated by Cornulites encrusting that part of the host shell. 4. The distribution of Spirorbis suggests that ''Camarotoechia' and Homoeospira lived with the pedicle valve up, whilst Ptychopleurella lived with the anterior commissure up. HURST: SILURIAN BRACHIOPOD EPIZOA 429 This concentration of encrusting fauna is of local occurrence. Other collections from the Eke Beds and other horizons in the Ludlow of Gotland, have a very low density of epizoans, when compared to this one. The high density of brachiopods, at this one locality, suggests that deposition was slow. In this environment, the brachiopods appear to have provided the only suitable substrate for encrustation by epizoans. Acknowledgements. I wish to express my sincere thanks to Mr. C. Richardson for help in collecting the fossils. I am also grateful to Dr. L. R. M. Cocks, Dr. W. S. McKerrow, Mr. F. Fursich, Mr. R. Watkins, and Mr. R. Sykes for reading the manuscript and suggesting improvements. The work was supported by a Natural Environment Research Council Grant. REFERENCES ACER, D. V. 1961. The Epifauna of a Devonian Spiriferid. Q. Jl geol. Soc. bond. 117, 1-10. BLIND, w. 1972. The systematic position of Cornulitids based on investigations of the structure of the shell. 24th Int. Geol. Congress. Section 7. Paleont. 5-7. FISHER, D. w. 1962. Other small conoidal shells. In moore, r. c. (ed.), Treatise on Invertebrate Paleontology, Part W, Miscellanea, W130-W138. Geol. Soc. Amer. Univ. Kansas Press. HEDE, J. E. 1960. The Silurian of Gotland. In regnell, y. and hede, j. e.. Guide to excursions Nos. A22 and Cl 7. Int. Geol. Congress. Norden, 1960. HOARE, R. D. and STELLER, D. L. 1967. A Devonian Brachiopod with Epifauna. Ohio J. Sci. 67, 291-297. NICHOLSON, H. A. 1872. On the genera Cornulites and Tentaculites and on a new genus Conchicolites. Am. J. Sci. 3, 202-206. REYMENT, R. A. 1971. Introduction to Quantitative Palaeoecology. Elsevier, 226 pp. RICHARDS, R. p. 1972. Autecology of Richmondian Brachiopods (Late Ordovician of Indiana and Ohio). J. Paleont. 46, 386-406. RUDWiCK, M. J. s. 1960. The feeding mechanisms of Spire-bearing fossil brachiopods. Geol. Mag. 97, 369-383. SCHUMANN, D. 1967. Die Lebensweise von Mucrospirifer Grabau, 1931 (Brachiopoda). Palaeogeogr., PalaeoclimatoL, Palaeoecol. 3, 381-392. THOMSON, J. A. 1927. Brachiopod Morphology and Genera (Recent and Tertiary). N.Z. Sci. and Art, Manual, 7. TRUEMAN, A. E. 1942. Supposed commensalism of Carboniferous Spirorbids and non-marine Llameli- branch. Geol. Mag. 79, 312-321. J. M. HURST Department of Geology Oxford, 0X1 3PR Revised typescript received 10 October 1973 SHORT COMMUNICATIONS AN IMPROVED METHOD OF MOUNTING PALAEONTOLOGICAL SPECIMENS FOR SEM EXAMINATION by E. M. FINCH Abstract. The technique involves the use of wax as the mounting medium for palaeontological specimens. A faster and more efficient mounting of specimens is achieved in addition to precise orientation and increased cleanliness of specimens. Following SEM examination, the specimens can be removed from the wax if necessary. The mounting of palaeontological specimens upon stubs prior to examination with the Scanning Electron Microscope (SEM) can be a lengthy and unsatisfactory exercise. When very small specimens are to be examined the need for a simple yet effective means of mounting is increased. Although the method to be described is primarily intended for use by the micropalaeontologist it is unlimited in its application. Current methods of mounting the smaller specimens for SEM observation require the use of an adhesive such as liquid glue, gum, or double-sided sticky tape. Having used these mountants, the author believes them to have disadvantages which limit their usefulness. Of the three, the tape is the least troublesome, although the exact orientation of specimens is difficult to achieve. The glues and gums are entirely unsatisfactory since both can be transferred on to the surface of the specimens during mounting. This leads to the obscuring of detail and undesirable ‘charging’ effects (PI. 59, fig. 2). In order to obtain adequate adhesion of the specimens to the stub it is necessary to use a large amount of glue. The excess glue is likely to flood into porous specimens such as planktonic foraminifera, thereby introducing artifacts and addi- tional problems of ‘charging’. The proposed, improved method employs the use of wax as the mounting medium (PI. 59, figs. 3, 4, and 5). METHOD The wax, entitled ‘Wax W’ is supplied in stick form by the manufacturers (Edwards High Vacuum). To begin with the surface of the stub must be coated with wax. The wax is softened in a bunsen flame and then applied to the stub surface. Gentle heating of the stub will cause the wax to flow evenly across the stub. The amount of wax that is put upon the stub is governed by the size of the specimen which is to be mounted. The larger specimens require more wax to secure them. As soon as the wax has cooled its surface may be washed with distilled water. The specimens can now be mounted upon the stub. This is usually carried out with the aid of an optical microscope. It is important [Palaeontology, Vol. 17, Part 2, 1974, pp. 431-434, pi. 59.] 432 PALAEONTOLOGY, VOLUME 17 that the microscope lamp has a heat-absorbing glass in place to prevent any soften- ing of the wax at this stage. Place the specimens upon the waxed surface in a position which is as near as possible to the required orientation. Remove the heat filter from the lamp and allow the wax to soften in the beam. When the wax has softened, orientate the specimens to their final positions and gently press them into the wax. The wax will not penetrate the specimens unless it has been considerably overheated. If this happens the specimens are likely to disappear into the wax. Such temperatures are not necessary to soften the wax in order to secure the specimens. All manipulations of the specimens are performed with a paint brush (Series ‘00’) and a fine sewing needle mounted in a suitable holder. If the specimens are too fragile to be pressed into the wax, a small indentation is made in the wax and the specimens are placed therein. In this case allow a longer time for the wax to soften. On sub- sequent hardening of the wax the specimens will be held firmly but safely in position. Alternatively, the softening of the wax can be allowed to proceed until the specimens settle into it of their own accord. As these steps are being carried out upon the stage of an optical microscope it is possible to see the depth to which the specimens become embedded in the wax. It is only necessary for the specimens to penetrate the surface of the wax for them to be completely secured (PI. 59, fig. 1 ). Once the wax has hardened the specimens may be cleaned and washed if necessary. It is always advisable to wash the stub with distilled water to remove any debris that has accumulated during the mounting procedure. Remove all the surplus wax from the edge of the stub and paint around the edge with ‘silver dag’. The stub may now be coated in the normal manner. The wax is unaffected by room temperature, vacuum pressures, ionic bombardment, or coating techniques. Once coated, the wax must not be allowed to soften. This technique of mounting specimens can also be used for the noncalcareous microplankton. In this case the stub must be covered with a thin layer of wax. Before the wax has hardened, one or more of the standard transmission electron microscope copper grids are pressed into it. Individual specimens can then be placed within the grid squares or alternatively, a drop of concentrate can be deposited across the grid surface. The specimens are secured by softening the wax as previously described (PI. 59, fig. 6). The main advantage in using the grids is that they enable re-location of specimens in a minimum of time. Following SEM examination the stubs may be stored for an indefinite period, preferably at a low temperature. If necessary, the specimens may be removed from the wax by cutting the wax into strips and removing it from the stub with a scalpel blade. The wax is dissolved by an organic aromatic solvent such as xylene, benzene, or toluene and the specimens can then be recovered. EXPLANATION OF PLATE 59 Fig. 1. Specimen mounted with wax demonstrates how only slight penetration of the wax is required to secure the specimen, x 100. Fig. 2. Internal view of specimen mounted with a liquid glue. Note obscur- ing of detail and ‘charging’ effects, x 160. Fig. 3. A similar specimen to that illustrated in Fig. 2, but mounted with wax. Note cleanliness of specimen, x 250. Fig. 4. Wax-mounted portion of outer test wall of a foraminifera. Both surface and pores are completely free of contaminant, x 1400. Fig. 5. Clearly revealed microcrystalline wall structure of foraminifer mounted with wax, x 1680. Fig. 6. Dinollagellate cyst mounted with wax. Only the tips of some of the processes are within the wax but this is sutlicient to retain the specimen, x 8000. PLATE 59 FINCH, mounting for S.E.M. 434 PALAEONTOLOGY, VOLUME 17 In the author’s experience there are no disadvantages associated with the use of this technique. By contrast, it enables a larger number of specimens to be mounted on any one stub. The actual number being limited only by the surface area of the stub. By following the correct procedure, the wax will not flood into the specimens thereby eliminating artefacts due to mounting materials. The charging effects associated with the use of other adhesives are overcome. With practice the technique provides a fast but safe way of mounting a wide variety of specimens. Acknowledgements. The author thanks the Chairman and Directors of the British Petroleum Company Limited for permission to publish this paper, also Dr. M. Muir (Imperial College, London), for organizing the EMCOM Symposium. E. M. FINCH British Petroleum Co. Ltd. Exploration and Production Research Division B.P. Research Centre Revised typescript received 24 June 1973 Sunbury-on-Thames A SPECIMEN LOCATION TECHNIQUE FOR SEM STREW MOUNTS by J. F. LAING Abstract. A technique is described for the preparation of strew mounts for SEM examination, which allows the position of any specimen to be recorded such that it can be re-examined at a later date. Simple strew mounts of palynological (or other micropalaeontological) material which are to be examined with a scanning electron microscope have previously suffered from the serious drawback that it is virtually impossible to relocate any particular specimen for later study. However, if a thin nickel grid, similar to the type used in transmission electron microscopy, is stuck on to the stub, it is possible to locate individual specimens by means of a series of co-ordinates. Text-fig. 1 is a scale diagram of the grid used. The grid was manufactured, according to the author’s specifications, by Smethurst High-Light Ltd. of Bolton, Lancs. lOOym BAB TEXT-FIG. 1. Scale diagram of grid, x 8. All grid bars 50 pm wide except where otherwise labelled. Triangles at the intersection of 100 pm bars, half circles at the intersection of 100 pm and 75 pm bars, and quarter circles at the intersection of 75 pm bars as shown. The co-ordinate reference system. ‘Eastings’ shown from left to right, unbracketed; ‘northings’ from bottom to top, in brackets. [Palaeontology, Vol. 17, Part 2, 1974, pp. 435-436.] 436 PALAEONTOLOGY, VOLUME 17 PREPARATION TECHNIQUE 1 . Clean the surface of the stub with a tissue soaked in ether in order to remove any grease. 2. Coat the surface of the stub with a suitable adhesive; I have found the artist’s material ‘acrylic polymer varnish’, manufactured by Reeves of Enfield, Middlesex, to be very suitable. A drop of the varnish is placed on the stub and spread evenly over the surface with a glass rod. It is then allowed to dry for about 10 minutes. The varnish is dry when it has changed from a milky colour to being transparent. (The stub should be covered as far as possible during preparation, in order to prevent dust and other foreign bodies from settling on to it.) 3. As soon as the varnish is dry, place the grid on to it; rub a glass rod over the grid applying a slight pressure. (It is advisable to place the grid on the stub as soon as the varnish is dry, otherwise it may not stick very well.) 4. After allowing the varnish to harden for about an hour, thoroughly mix the residue to be mounted with a little distilled water. Place a drop of this mixture on the area of the grid with a Pasteur pippette and gently stir the drop over the area of the grid with a glass rod. Allow to dry; it is advisable to stir occasionally during drying to prevent any floating material from being concentrated together at the end of drying. 5. Coat the stub (e.g. with gold-palladium) in the normal manner. It is now ready for examination. THE CO-ORDINATE REEERENCE SYSTEM (see text-fig. 1 ) Each square of the grid is examined systematically at a suitable magnification (I have found Ik to be a suitable working magnification). When a specimen has been found its position is noted in terms of ‘eastings’ and ‘northings’, in exactly the same manner as when giving a map reference. The triangle in the centre of the grid is taken to point ‘north’; the centre of the grid has the grid reference 300800. The ‘eastings’ run from 200 to 400, and the ‘northings’ from 700 to 900, the centre of each grid bar being the point where each unit of ten begins. Grid references are given in terms of the ‘eastings’ first, followed by the ‘northings’, for example point x on text-fig. 1 has the grid reference 273836 (the units at the end of the ‘eastings’ and ‘northings’ are estimated). The provision of some thicker grid bars, triangles, quarter circles, and half circles (see text-fig. 1) facilitates the location of any particular square by having to reduce magnification only a little, rather than to such an extent that the whole grid needs to be observed. This technique has been found to be successful in the examination of palynological residues, but there is no reason why it could not be used in the examination of other similar-sized material (e.g. coccoliths). Acknowledgements. This paper forms part of a more extensive study carried out whilst I was in receipt of a N.E.R.C. training award. 1 wish to thank Mr. N. F. Hughes and Mr. D. Newling for valuable discussion. J. F. LAING Department of Geology Manuscript received 25 May 1973 Sedgwick Museum, Cambridge AFFINITY OF DAYIACEAN BRACHIOPODS by J. G. JOHNSON Abstract. A recent proposal to assign the dayiacean brachiopods to the Athyridoidea is rejected. A new suborder Dayioidea is proposed. Copper (1973) has recently suggested a radical revision of the systematic place- ment of the dayiacean brachiopods, i.e. genera of the families Dayiidae, Anoplo- thecidae, Kayseriidae, and Leptocoeliidae. His studies centred particularly around the genera Bifida Davidson and Kayseria Davidson. Copper maintained that the dayiaceans belong not with the atrypids, but instead should be reassigned to the suborder Athyridina {sic). That the dayiaceans are distinct from atrypaceans there is no doubt; that is the basis for their separation at superfamily level. This being readily stipulated does not necessarily lead to the second conclusion— that dayiaceans are athyridids. To me there are as many, or more, reasons that show dayiaceans to be distinct from athyridids as from atrypaceans. Copper’s reasons for suggesting a dayiacean- athyridid alliance invite the closest scrutiny. Copper (1973, p. 118) noted that all dayiaceans are small brachiopods and this is true, but he also said, ‘the small size factor seems to hold generally true also for most athyridids’. The quoted statement is misleading; Cryptothyrella, Meristina, Meristella, and many species of Athyris itself are relatively large. A valid generaliza- tion is that athyridids comprise genera and species representing a wide range in size. Copper (1973, p. 118) noted that all dayiaceans lack a large pedicle opening and that they therefore were not primarily pedicle-supported brachiopods. This is true and it is surprising that Copper did not also point out that a large pedicle foramen is an athyridid hallmark. This attribute of the athyridid suborder is clearly noted in the Treatise diagnosis (Boucot et ai, in Moore 1965, p. H654). In page 1 19 Copper seems to say that athyridids lack or have a very small pedicle opening but his state- ment is ambiguous. Copper (1973, p. 1 19) noted that athyridids are identified by, among other things, ‘their relatively rounded shape’, but many dayiaceans are costate or plicate, as are atrypids. Copper regards ‘the radial rib structure’ of some dayiaceans as ‘superficial’. Why is this superficial? Because there are smooth and ribbed atrypids? Why then attempt to point to supposed exceptions by suggesting that Atrythyris Struve is ribbed? In fact Atrythyris, like Pradoia Comte, has radial lirae on an otherwise smooth surface. In his discussion of the superfamily Dayiacea Copper (1973, p. 120) reported, ‘Nearly all dayiaceans show a jugal saddle and accessory lamellae extending pos- teriorly from the jugal saddle. This feature is identical to that of the true athyridids (see Williams and Rowell 1965, p. H103).’ The truth is that no known dayiacean has a jugal saddle. The very Treatise figure to which Copper pointed (Williams and [Palaeontology, Vol. 17, Part 2, 1974, pp. 437-439.] 438 PALAEONTOLOGY, VOLUME 17 Rowell, in Moore 1965, p. H103, fig. 108) correctly shows that a jugal saddle is an anterior projection from a simple band-like jugum. Copper’s own illustrations on pp. 121, 122 (text-figs. 1, 2) show Bifida and Kayseria each without a jugal saddle. A similar lack of jugal saddle is evident in photographs of Coelospira Hall (Boucot and Johnson 1967, pi. 166, figs. 20, 21). Copper’s misunderstanding of jugal termi- nology is unfortunate, but no matter what names are applied to these structures it is impossible to agree with Copper’s contention that they are ‘identical’ in dayiaceans and athyridids. Leaving behind the question of jugal saddle, there remains Copper’s contention (1973, p. 120) that, ‘nearly all dayiaceans show . . . accessory lamellae’. One way to achieve this is by labelling; thus in Copper’s text-fig. 1 what is commonly called a jugal stem (Williams and Rowell, in Moore 1965, pp. H103, H146) is labelled an ‘abortive accessory lamella’, which is like saying the fossil would have lamel- lae if it had lamellae. In fact only Kayseria of all dayiacean genera has accessory lamellae. With these points in mind Copper’s conclusions (pp. 136, 137) may be evaluated. 1. It is contended by Copper that Bifida and Kayseria have ‘wedge-like crural bases instead of ball-like crural bases, as in atrypids’. It is impossible for atrypids to have ‘ball-like’ crural bases. Atrypid crural bases commonly appear subcircular in cross-section, but there is nothing in Copper’s sections (text-figs. 3, 5) of Bifida or Kayseria that suggests to me any fundamental difference. 2, 3, and 6. These points made by Copper are part of the admitted differences between dayiaceans and atrypaceans and are why these taxa have been regarded as distinct and separate superfamilies. They fail as reasons to include dayiaceans in the Athyridoidea. 4 and 5. Copper here points to two unique structures, one of Kayseria and one of Bifida, and says they are ‘absent in atrypids’. They also do not occur in any other dayiacean genus. 7 and 8. These points seem inconsequential to me. My own conclusion is that because of much work on atrypid morphology, pro- minent among which is Copper’s own, the group has grown while its recurring structures have become much better known. Copper has called, in essence, for a purification of the hierarchy to which the atrypid name shall be applied. To achieve this he has pointed to the contrasting features of dayiacean and atrypacean mor- phology. These are legion and are undeniable. However, Copper’s solution as to where the ousted superfamily should reside is ill-considered. As an alternative to inclusion of the dayiaceans in the Athyridoidea I propose a new suborder Dayioidea and suggest that its inherent morphologic structures are as characteristic and different as are the structures of the other spiriferid suborders, viz. Spiriferoidea, Retzioidea, Athyridoidea, and Atrypoidea, such that its recogni- tion is fully justified. I therefore append a diagnosis as follows: Suborder dayioidea Diagnosis. Spiriferida with costate or smooth, impunctate shells, lacking interareas JOHNSON: DAYIACEAN BRACHIOPODS 439 and having foramen within the delthyrium; spiralia directed ventrally, laterally, or planospiral parallel to the median plane, or not calcified ; jugum present or absent, jugal saddle not developed. REFERENCES BOUCOT, A. j. and Johnson, j. g. 1967. Species and distribution of Coelospira (Brachiopoda). J. Paleont. 41, 1226-1241. COPPER, p. 1973. Bifida and Kayseria (Brachiopoda) and their affinity. Palaeontology, 16, 1 17-138. MOORE. R. c. (ed.), 1965. Treatise on Invertebrate Paleontology, Part H, Brachiopoda, 927 pp. J. G. JOHNSON Department of Geology Oregon State University Corvallis, Oregon Manuscript received 4 September 1973 U.S.A. 97331 f I I i'l It 1 ' h>i:- . :¥ •*. '■'" V .i ■iS'h-ji :■ mh‘i •t. THE PALAEONTOLOGICAL ASSOCIATION The Association was founded in 1957 to further the study of palaeontology. It holds meetings and demonstrations as well as publishing Palaeontology and Special Papers in Palaeontology. Membership is open to individuals and to institutions on payment of the appropriate annual subscription : Institutional membership .... £10 00 (U.S. $26.00) Ordinary membership .... £5 00 (U.S. $13.00) Student membership .... £3-00 (U.S. $8.00) There is no admission fee. Institutional membership is only available by direct application, not through agents. Student members are persons receiving full-time instruction at educational institutions recognized by the Council. On hrst applying for membership, an application form should be obtained from the Membership Treasurer. Subscriptions cover one calendar year and are due each January; they should be sent to the Membership Treasurer, Dr. E. P. F. Rose, Department of Geology, Bedford College, Regent’s Park, London, NWl 4NS, England. PALAEONTOLOGY All members who join for 1974 will receive Volume 17, Parts 1-4. All back numbers are still in print and may be ordered from B. H. Blackwell, Broad Street, Oxford, 0X1 3BQ, England, at £5 per part (post free). A complete set. Volumes 1-16, consists of 63 parts and costs £315. SPECIAL PAPERS IN PALAEONTOLOGY The subscription rate is £8 (U.S. $22.00) for Institutional Members and £4 (U.S. $11.00) for Ordinary and Student Members. Subscriptions should be placed through the Membership Treasurer, Dr. E. P. F. Rose, Department of Geology, Bedford College, Regent’s Park, London, NWl 4NS, England. Ordinary and Student members only may obtain individual Special Papers from Dr. Rose at reduced rates. Non-members may obtain them at the stated prices from B. H. Blackwell, Broad Street, Oxford, 0X1 3BQ, England. COUNCIL 1974-1975 President : Professor C. H. Holland, Department of Geology, Trinity College, Dublin, 2, Ireland Vice-Presidents: Dr. R. Goldring, Department of Geology, The University, Reading, RG6 2AB Dr. W. D. 1. Rolfe, The Hunterian Museum, The University, Glasgow, G12 8QQ Treasurer : Dr, J, M. Hancock, Department of Geology, King’s College, Strand, London, WC2R 2LS Membership Treasurer: Dr, E. P. F. Rose, Department of Geology, Bedford College, Regent’s Park, London, NWl 4NS Secretary: Dr. C. T. Scrutton, Department of Geology, The University, Newcastle upon Tyne, NEl 7RU Editors Dr. J. D. Hudson, Department of Geology, The University, Leicester, LEI 7RH Dr. L. R. M. Cocks, Department of Palaeontology, British Museum (Natural History), Cromwell Road, London, SW7 5BD Dr. C. P. Hughes, Department of Geology, Sedgwick Museum, Downing Street, Cambridge, CB2 3EQ Dr. J. W. Murray, Department of Geology, The University, Bristol, BS8 ITR Other Members of Council Dr. M. G. Bassett, Cardiff Dr. M. C. Boulter, London Professor D. L. Dineley. Bristol Dr. J. K. Ingham, Glasgow Dr. J. E. Pollard, Manchester Dr. A. W. A. Rushton, London Dr. P. Wallace, London Dr. D. D. Bayliss, Llandudno Dr. C. H. C. Brunton, London Dr. J. A. E. B. Hubbard, London Dr. C. R. C. Paul, Liverpool Dr. P. F. Rawson, London Professor D. Skevington, Galway Overseas Representatives Australia: Professor Dorothy Hill. Department of Geology, University of Queensland, Brisbane Canada: Dr. B. S. Norford, Institute of Sedimentary and Petroleum Geology, 3303-33rd Street NW., Calgary, Alberta India: Professor M. R. Sahni, 98 The Mall, Lucknow (U.P.), India New Zealand: Dr. C. A. Fleming, New Zealand Geological Survey, P.O. Box 30368, Lower Hutt West Indies and Central America : Mr. John B. Saunders, Geological Laboratory, Texaco Trinidad, Inc., Pointe-a-Pierre, Trinidad, West Indies Western U.S. A. : Professor J. Wyatt Durham, Department of Paleontology, University of California, Berkeley 4, California Eastern U.S. A. : Professor J. W. Wells, Department of Geology, Cornell University, Ithaca, New York Palaeontology VOLUME 17 PART 2 CONTENTS Upper Ordovician Irilobites from central New South Wales B. D, WEBBY 203 Capitosauroid labyrinthodonts from the Trias of England ROBERTA L. PATON 253 A new giant penguin from the Eocene of Australia R. J. F. JENK.INS 291 Lower Devonian (Dittonian) plants from the Welsh Borderland DIANNE EDWARDS and JOHN B. RICHARDSON 311 Megalosaurids from the Bajocian (Middle Jurassic) of Dorset MICHAEL WALDMAN 325 Lower Cretaceous sclerosponge from the Slovakian Tatra Mountains JOZEF KAZMIERCZAK 341 Two new subspecies of Phacops rana [Trilobita] from the Middle Devonian of North-West Africa CHRISTOPHER J. BURTON and NILES ELDREDGE 349 Podocarpus from the Upper Cretaceous of Eastern Asia and its bearing on the theory of conifer evolution v. A. KRASSILOV 365 Lower Carboniferous conodont faunas from north-east Devonshire S. C. MATTHEWS and J. M, THOMAS 371 An apparently heterosporous plant from the Middle Devonian of New Brunswick HENRY N. ANDREWS, PATRICIA G. GENSEL, and WILLIAM H. FORBES 387 The Silurian trilobite Onycopyge Woodward D. J, HOLLOWAY and K. S. W. CAMPBELL 409 Selective epizoan encrustation of some Silurian Brachiopods from Gotland J. M. HURST 423 Short communications: An improved method of mounting palaeontological specimens for SEM examination E. M. FINCH 431 A specimen location technique for SEM strew mounts J. F. LAING 435 Affinity of Dayiacean brachiopods J. G. JOHNSON 437 Printed in Great Britain at the University Press, Oxford by Vivian Ridler, Printer to the University nES SMITHSONIAN INSTITUTION NOIlOiliSNI NVINOSHilWS S3iaVHail LIBRARIES SMITHSONL z; z , z: i- 2 ^ fr\ ^ tn ^ ^ ^ Z OT m 'CNOi^ ^ x^tvAst^ m '• \^'‘ ^ '•^,tvAsji)>' m Crt — CO ± CO — _ iSNl NVINOSHillMS SBiavyaiT LIBRARIES SMITHSONIAN INSTITUTION NOlifUliSNl NVINOSHill z ^ ^ 5 ’* ^ z . c/) z •vAs^s: _ - X MV z Vs > CO z CO (/) *'" Z CO nES SMITHSONIAN INSTITUTION NOIiniliSNI_ NVINOSHil/JS SBiavaail LIBRARIES SMITHSON!/ CO _ to \ CO — (/) to O z iSNi NViNOSHims saiavaan libraries smithsonian institution NoiiniiiSNi nvinoshih 70 > 70 m -yt, U1 m xi^vD^ ^ m CO ? CO — cn RIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHilWS S3iavaa CO 2 CO Z CO 03 PO > ^ PO X,<£0ms5i/ 2 2 > z CO *■ Z to Z IlSNI NVINOSHIIWS S3iyvaan libraries smithsonian institution NOliniliSNI nvinoshii CO P3 CO Z CO Q '\ _ x^JiiS^px O xc"! O XT' u'-^ XC!i!£2iP^ O Z "■ _l z _J 2 -I Z ''' RIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHillNS S3iavaan LIBRARIES SMITHSONI z o CO iSNi NviNOSHilws S3iavaan l 2 to Z 03 PO > 70 m -• \ w- " -xivAsti^ m xivos^ii^ CO Z CO z BRARIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHII ■, CO Z . CO z RIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHi^S S3iavaan LIBRARIES SMITHSONI liSNi NviNOSHims S3iavaan libraries smithsonian institution NoiiniiiSNi nvinoshih CO TITUTION NIOliniliSNI NVINOSHillMS S3iaVdail LIBRARIES SMITHSONIAN INSTITUTION NOliO Z r- , Z r- Z r- 2 c/5 CD LirCff&il'bX CD :d 5 i _^y — rn '■ ^ pri CO ■ _ Crt CO dVdan LIBRARIES SMITHSONIAN INSTITUTION NOliOiliSNI NVINOSHillAlS SaidVdail LI Z , ^ Z » CO 2 CO , 2 I A.. .. I Ma.. I I I riTUTlON NOliDiliSNI NVINOSHillMS SBIdVdail LIBRARIES SMITHSONIAN INSTITUTION NOliO P- c O Z _J dvaan libraries smithsonian institution NoiiniiiSNi nvinoshiiws saiavaan libr, Z r- ^ 2 r- 2 r- rx 9 ..m 2 ;;; 9 ~ TITUTION NOlinillSNI NVINOSHilWS SBldVdan libraries SMITHSONIAN INSTITUTION NOlif I CO O avaan*"LIBRARI ES”'SMITHS0NIAN institution NOlinillSNI NVINOSHilWS“'S3 iaVdan“LIBR CO Z CO — CO TITUTION NOlinillSNI NVINOSHIIWS S3ldVaan LIBRARIES SMITHSONIAN INSTITUTION NOIlfl P " x\, .. i I ° dvaan libraries smithsonian institution NoiiniiiSNi nvinoshiiws S3iavaan libr ■Ma,. 5 .. P I I I I I > \iVASiiX 2 -X 2 >■ S CO ' z CO Z CO 2 CO TITUTION NOlinillSNI NVINOSHIIINS S3ldVdan LIBRARIES SMITHSONIAN INSTITUTION NOUH \, (f) BV>^^5iS!5vU B RAR I ES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHIIWS S3ldVaail LIBR ' " Z r- Z r- 2 r-