7S. ISbA D & W 1988 | Bulletin of the British Museum (Natural History) Geology series Vol 41 1987 British Museum (Natural History) London 1987 Dates of publication of the parts Nol . : : 2 : é : : ? : : 29 January 1987 Nod : : : d : : : : : : 30 April 1987 INOS . , é : : : : : : : : 30 July 1987 No4 . : : é : : : : : : : 29 October 1987 ISSN 0007-1471 Printed in Great Britain by Henry Ling Ltd., at the Dorset Press, Dorchester, Dorset No 1 No 2 No 3 No 4 Contents Geology Volume 41 The Downtonian ostracoderm Sclerodus Agassiz (Osteostraci: Tremataspididae). P. L. Forey Lower Turonian (Cretaceous) ammonites from south-east Nigeria. P. M. P. Zaborski The Arenig Series in South Wales: Stratigraphy and Palaeontology I. The Arenig Series in South Wales. R.A. Fortey & R.M. Owens . II. Appendix. Acritarchs and Chitinozoa from the Arenig Series of south-west Wales. S. G. Molyneux Miocene geology and palaconologyo of Ad Dabiiah Saudi Arabia. Compiled by P. J. Whybrow Summary, with Table of vertebrate fauna. P. J. Whybrow Miocene stratigraphy, geology and flora (Algae) of eastern Saudi Arabia and the Ad Dabtiyah vertebrate locality. P. J. Whybrow, H. A. McClure & G. F. Elliott. The phyletic position of the Ad Dabtiyah hominoid. P. J. Andrews & L. Martin : : ’ Mastodons from the Miocene of Saudi Arabia. A. W. Gentry . Rhinoceroses from the Miocene of Saudi Arabia. A. W. Gentry . Ruminants from the Miocene of Saudi Arabia. A. W. Gentry . Miocene Suidae from Ad ae eastern Saudi Arabia. M. Pickford ; ; : 5 : A delphinoid ear bone from the Dam Formation (Miocene) of Saudi Arabia. F.C. Whitmore, jr . Early Miocene fish from eastern Saudi Arabia. P. H. Greenwood Index Page 309 365 367 371 383 395, 409 433 441 447 451 455 7 a os i, oe. i. .") t. . " z aa "4 5 re 3 Pas PA Pa aN in oie yeh ‘grad 4 hry wie a ae : : Be Aen sit » val ay A gee sity 4 ar -— 35 ~ _— 3 ’ =~ > a ree | ; “ oe re a E i , 1 ee = Sige ii ¥ : Bulletin of the E | ‘Br British Museum (Natural History) , Py “ i Bie a PK 1s; 1 h ' mH The Downtonian ostracoderm clerodus Agassiz Isteostraci: Tremataspididae) - a. | Be orey : . oe asa at ot ee i; 4 fe: eras tp aa '_ ae yr . bas fa id oa “ a ‘ x SS aoe . i.e Geology series Vol 41 Nol 29 January 1987 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum’s resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England. World List abbreviation: Bull. Br. Mus. nat. Hist. (Geol.) © Trustees of the British Museum (Natural History), 1987 The Geology Series is edited in the Museum’s Department of Paiaeontology Keeper of Palaeontology: DrL.R.M.Cocks Editor of the Bulletin: Dr M. K. Howarth Assistant Editor: Mr D. L. F. Sealy ca ISS \ece 29 JAN 1987 Ae ee ere es F) Ss . | f j + Fi % camer cf fi pe 2 ~ Cy a ae SD sae # Yin’ r a, , H ISBN 0 565 07015 0 ISSN 0007-1471 Geology series British Museum (Natural History) Vol 41 No 1 pp 1-30 Cromwell Road London SW7 5BD Issued 29 January 1987 The Downtonian ostracoderm Sclerodus Agassiz. (Osteostraci: Tremataspididae) P. L. Forey ,. Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 SBD Contents Synopsis 1 Introduction 1 Historical review D, Materials and methods 4 Abbreviations used in figures : : ; : : : : ; . 4 Systematic description ; : : : ; : ; : 5 : : 4 Genus Sclerodus Agassiz . 4 Relationships of Sclerodus . 17 26 Conclusions : ; , ; : 5 ‘ 3 ; ; ; : Acknowledgements. : ‘ : : , ; : 4 : é : 27 Appendix . : , ; ‘ ; : . , : ; F 5 ; 27 References . : ; ; : ; , ‘ F : : 4 . F 28 Index . : ; ; : : : ; é 5 ; H : ; ; 29 Synopsis Various aspects of the morphology of the osteostracan Sclerodus Agassiz are described and discussed in the light of new specimens in an attempt to reconcile four different morphological interpretations. It is concluded that Sclerodus has normal osteostracan sensory fields, that the lateral line system may be represented as a series of pits, and that the margin of the cephalothoracic shield is penetrated by four fenestrae which may possibly have served a stabilizing hydrodynamic function. Relationships of Sclerodus are discussed with a review of osteostracan classification leading to discussion of computer-generated trees. A phylogeny is favoured which treats ateleaspids as a paraphyletic group and tremataspids as a monophyletic derived group. Introduction Sclerodus pustuliferus Agassiz is a small and rather unusual osteostracan restricted to the Downtonian of the Anglo-Welsh basin. Remains of Sclerodus are very common in the Ludlow Bone Bed and immediately overlying rocks, where it is easily recognized by its distinctive ornament of small, closely-packed hemispherical tubercles. However, the abundance of its remains is matched by our ignorance of its morphology. Only the dorsal half of the cephalic armour is known, and even here knowledge is incomplete. There are very few reasonably intact specimens and regrettably one of the best (Stensid 1932: pl. 52, fig. 2) has now been lost. The fragments most commonly found are pieces of the main part of the shield and portions of the so-called ‘cornua’. Each ‘cornu’ bears a marginal row of elongate tubercles, so that fragments of them were initially confused with jaws and teeth. More complete material enabled Lankester (1870) to confirm a suggestion made by Harley (in Murchison 1867) that Sclerodus pustuliferus is, in fact, an osteostracan and to provide the first restoration (Lankester 1870: fig. 31). Since that time three further restorations have been attempted (Stensid 1932, Denison 1951a, Janvier 1975). No two of these agree on interpretation of structures which are obvious in other osteostracans, such as the presence or absence of cornua, pectoral sinuses or one or more lateral sensory fields. Bull. Br. Mus. nat. Hist. (Geol.) 41 (1): 1-30 Issued 29 January, 1987 2 P. L. FOREY These discrepancies, combined with the fact that Sclerodus shows some interesting special- izations along the rim of the cephalothoracic shield, make this monotypic genus an interesting subject for study. The main motivation derives from sporadic collecting in the Ludlow Bone Bed at Forge Bridge, Downton Castle estate, Shropshire by Dr W. Graham-Smith of Boars Hill, Oxfordshire. Dr Graham-Smith has collected many fragments of Sclerodus, some of which show unusual and previously undescribed pits surrounding the orbits. One specimen, BMNH P.58694, described and illustrated here (Fig. 5, p. 11), was particularly helpful in the reinterpre- tation of existing material. I am grateful to Dr Graham-Smith for the donation of several specimens. It has allowed me to redescribe and update our knowledge of Sclerodus and to offer comments on some of the more unusual aspects of this genus. Historical review Sclerodus pustuliferus was first described by Agassiz (in Murchison 1839: 606; pl. 4) from figures, sent to him by Murchison, of four fragmentary specimens. On the basis of these figures Agassiz likened the fragments to the grinding teeth of the “‘bradyodont’ Psammodus, and because of the rough pustulated surface he coined the name Sclerodus (rough-tooth) pustu- liferus. On the same occasion Agassiz described seven further specimens as jaws and teeth under the names Plectrodus mirabilis and Plectrodus pleiopristis. M’Coy (1853) reinterpreted nearly all the specimens figured by Murchison (1839: pl. 4), including those referred to species of the genera Sclerodus and Plectrodus, as being the remains of the crustacean Pterygotus. Additionally, he could see no reason to recognize separate species and united them all under the name Pterygotus pustuliferus (= Plectrodus mirabilis + P. pleiopristis + Sclerodus pustuliferus). It should be noted that M’Coy (1853: 13) apparently did not see the original material referred to those species because it had by then been ‘lost’ (see below). Murchison (1853) replied testily, saying that he, and Messrs Salter and J. Sowerby, who had prepared the drawings sent to Agassiz, disagreed with the crustacean interpretation and main- tained the identity of these remains as fish jaws and teeth. Egerton (1857) followed by describ- ing more material from Ludlow as jaws of Plectrodus mirabilis. Thus, while authors disagreed over whether there were one or more species, almost all agreed that they were fishes and not crustaceans. This was confirmed by Harley (1861: 544, footnote) who had sectioned specimens and found them to be made of bone and dentine. Harley further suggested that they were the posterior spines (cornua) of cephalaspid fishes rather than fish jaws and teeth. Murchison, while acknowledging Harley’s opinions (1867: pl. 35, legend), remained convinced that they were jaws and ankylosed teeth (1867: 241). Lankester (1870: 58) supported Harley’s view by describing tolerably complete head shields based on new material collected by Dr Grindrod and Mr Lightbody from the Downton Castle Sandstone of Ludford Lane, Ludlow. Lankester regarded Plectrodus mirabilis and P. pleiopristis as junior synonyms of S. pustuliferus and considered Sclerodus as a subgenus of Auchenaspis Egerton. He named this subgenus Eukeraspis, but he gave no reason why he dropped the name Sclerodus. Eukeraspis was associated with Thyestes (Auchenaspis) because Lankester believed that in both the shield was composed of a semicircular cephalic porton and an abdominal portion formed by separate paired plates. The abdominal portion was unknown for Eukeraspis, but Lankester predicted its presence, adding (1870: 59) *... this is a question which inquiry with the hammer may soon decide . . .. Such inquiry has failed to find the abdominal division, but from Lankester’s time Sclerodus has been closely associated with Auchenaspis (Thyestes). Nevertheless, Lankester (1870: fig. 31) did provide the first restoration of the cephalic shield. Woodward (1891) agreed with Lankester over the restoration, association with Thyestes, and renaming Eukeraspis. But Woodward considered Eukeraspis to form a distinct genus (syn. Sclerodus, Plectrodus). Thus, to this point in the confused history of Sclerodus there had been debate about whether the Ludlow Bone Bed material represented one, two or three species; whether it belonged to fishes or to crustaceans; whether it represented jaws and teeth or part of the cephalic shield; DOWNTONIAN OSTRACODERM SCLERODUS 3 and finally, whether the generic name should be changed to Eukeraspis. By the turn of the century the consensus seemed to be that there was one species, that it was a cephalaspid fish closely related to Thyestes and that the original material described by Agassiz represented the denticulated cornua and should go under the name of Eukeraspis. Woodward (1917) considerably clarified the situation by following through his earlier suspi- cion (1891: 195) and a suggestion by Priem (1910: 5) that Plectrodus mirabilis and P. pleiopristis represented the dentigerous jaws of ischnacanthid acanthodians. This is the current status of Plectrodus (Denison 1979: 41). Thus, Sclerodus pustuliferus remains the only cephalaspid material described by Agassiz in Murchison’s The Silurian System. The change of generic name to Eukeraspis is unnecessary (see also Stensio 1932: 175, footnote). Subsequent work*on Sclerodus is chiefly that of Stensid (1932), who has provided the most complete description, Denison (195la, b) and Janvier (1975). These authors differ in their interpretations of the ‘cornua’ and sensory fields and their ideas are discussed in the relevant descriptive sections below. A summary of the differing ideas of the morphology is provided in Fig. 1. Lankester 1870 Stensid 1932 lateral field cornua Janvier 1975 lateral fields rim of shield Fig. 1 Sclerodus pustuliferus Agassiz. Four morphological interpretations. 4 P. L. FOREY As described above, the early history of the study of Sclerodus was somewhat tangled. Unfortunately, the history of the original material is equally problematical. When Agassiz described Sclerodus and Plectrodus he did so from drawings. Murchison (1853), in his reply to M’Coy (see above), records that he had given the specimens, collected and mounted on cards by Rev. R. W. Evans, to the Geological Society of London. But he also records that the specimens could no longer be found (see also Murchison 1867: 133, footnote). So, sometime between 1839 and 1853 some of the specimens illustrated in plate 4 of The Silurian System had gone astray. It appears (Jeannet 1928) that some of the material was incorporated in the Musée dHistoire Naturelle de Neuchatel and later into the Institut de Geologie de /Université de Neuchatel where Agassiz taught, 1832-1846. It is possible that Agassiz, who visited England in 1840, or Joseph Dinkel, Agassiz’ artist based at the Geological Society for several years, took some specimens to Neuchatel with the intention of describing them more fully. But by then Agassiz was preoccupied with his glacial studies; in any event the descriptions were not forth- coming. Woodward (1917: 74) found three of the original specimens in Neuchatel in 1898 while Jeannet (1928: 106) records five of the original 11 specimens referred to Sclerodus and Plec- trodus. There is only one of the four original specimens of Sclerodus pustuliferus and this is listed as number 3 and represents that illustrated in Murchison’s The Silurian System (1839: pl. 4, figs 60-62). This was quite correctly chosen as the lectotype by Stensi6 (1932). Material and methods The material studied belongs to the British Museum (Natural History) (BMNH); the British Geological Survey, Keyworth (BGS); and the Department of Geology, University of Birming- ham (BU). The specimens are referred to by register number prefixed by their respective institutional abbreviations. Most of the material is from the Ludlow Bone Bed and was studied directly. Rubber latex casts were helpful in the study of the material from the Downton Castle Sandstone. Histological sections were made from isolated fragments found in the Ludlow Bone Bed. Abbreviations used in figures a.p grooves housing anterior and posterior 0.a foramen for occipital artery semicircular canals oes groove for oesophagus a.pit anterior pit orn ornament cf circumnasal fossa O.r olfactory recess d canal leading to dorsal sensory field p.d pineal duct d.a groove for dorsal aorta p.f prebranchial fossa d.s.f dorsal sensory field p.o pineal opening h.v groove for lateral head vein (jugular p.pit posterior pit vein) prof foramen for profundus nerve 1.C.a foramen for internal carotid artery V.C vestibular chamber Ls.f lateral sensory field V.S superficial vein issuing from head vein m.f marginal fenestra IV, V,, er re m.pit middle pit VII, IX n.c foramina for nerves 1-4 s.e.1. canals leading to lateral sensory fields n.d nasohypophysial duct (see p. 12) n.o nasohypophysial opening (o) orbit Systematic description Family TREMATASPIDIDAE Woodward, 1891 Genus SCLERODUS Agassiz, 1839 1839 Sclerodus Agassiz (in Murchison): 606. 1870 Auchenaspis Egerton (in part); Lankester: 58 (subgenus Eukeraspis). 1887 Eukeraspis Lankester; Zittel: 150. 1891 Eukeraspis Lankester (in part); Woodward: 193 (not Plectrodus). DOWNTONIAN OSTRACODERM SCLERODUS 5 DiAGNosIs (emended). Tremataspids in which the circumnasal fossa is deep and elliptical, with a smooth floor: cephalothoracic shield perforated along lateral margin by four fenestrations, the largest lying anteriorly; thereafter each decreasing in size posteriorly: sensory lines absent, but perhaps represented by three pairs of pits lying close to the orbit and circumnasal fossa: abdominal region of the shield ossified along lateral margin only, leaving central part naked or perhaps covered with scales: ornament developed as regular hemispherical tubercles which are particularly large over the swellings immediately in front of and behind the pineal recess: margin of shield bearing a regular row of enlarged tubercles: histology of exoskeleton very simple, represented only by basal layer and overlying spongy bone with no circumareal canals. TYPE AND ONLY SPECIES. Sclerodus pustuliferus Agassiz. Sclerodus pustuliferus Agassiz 1839 Sclerodus pustuliferus Agassiz (in Murchison): 606; pl. 4, figs 27-32, 60-62. 1854 Sclerodus pustuliferus Agassiz; Murchison: pl. 35, figs 9-12. 1870 Auchenaspis (Eukeraspis) pustulifera (Agassiz); Lankester: 58, figs 31, 32; pl. 31, figs 9-14. 1932 Sclerodus pustuliferus Agassiz; Stensi6: 177, fig. 62; pl. 52, figs 1, 2; pl. 53, figs 1-5; pl. 56, fig. 1. 195la_ Sclerodus pustuliferus Agassiz; Denison: 185. 1975 Sclerodus pustuliferus Agassiz; Janvier: figs 2B, 5. DiaGnosis. As for genus; the only species. LECTOTYPE. Fragment of cornu: Institut de Géologie, Université de Neuchatel number 3. Ludlow Bone Bed, Downtonian; Ludlow, Shropshire. Selected Stensio (1932: 177). MATERIAL. Fifty-five specimens were examined in this study, as detailed in Appendix, p. 27. The material comes from the Ludlow Bone Bed, Downton Castle Sandstone and Temeside Shales/Lower Red Downton Sandstone of Shropshire, Herefordshire and Staffordshire. DESCRIPTION. The general shape of the shield is seen in BMNH P.9756, on which the resto- ration in Fig. 2 is based. The cephalic portion is strongly vaulted at the level of the orbits but the rim of the shield and so-called cornua are shallow. There are few specimens which show the cephalic portion attached to the so-called cornua; more usually broken cephalic shields and isolated ‘cornua’ are found. One complete specimen (BMNH P.9756, Fig. 2) shows a total length of 45-5 mm, of which the cephalic portion is 21 mm long. Using the proportion of cephalic to ‘cornw’ length of this specimen one can estimate that the largest specimen (BGS GSM 5150) must have been about 85mm in total length (snout to posterior level of ‘cornua’). The greatest width occurs two-thirds of the way back and the outline of the head plus ‘cornua’ resembles that of Dartmuthia or Tremataspis. The orbits are placed close together and this means that the pineal area is confined to a narrow longitudinal strip. No dermal pineal plate has been found and the extreme narrowness of the space left between the orbits may imply its absence. The pineal opening lies slightly below the surface where it opens at the end of a short duct within the skeleton (Figs 3, 5B, C). The duct is slightly asymmetrical which no doubt reflects the asymmetry of the underlying habenular recess as described by Janvier (1977) for Belonaspis puella (Wangsj6). The anterior and posterior borders of the orbital area are raised into prominent ridges which bear ornament tubercles larger than those covering most of the shield. The anterior ridge runs into a crest surrounding the nasohypophysial opening, while the posterior ridge is continuous with a shallow ridge defining the dorsal sensory field (Fig. 3). The nasohypophysial opening is contained within the floor of a deép, well-defined depression—the circumnasal fossa (Stensio 1932: fossa circumnasalis, Stensid 1927; antorbital fossa, Lankester 1870). The nasohypophysial opening is slit-like and immediately surrounded by a narrow ridge of bone. The form of the circumnasal fossa is very similar to that seen in thyestidians (sensu Janvier 1981b) and espe- cially to Tremataspis (Janvier 1985b: fig. 34A). The nasohypophysial opening is an elongate slit similar to that seen in thyestidians. But it should be noted that a similarly-shaped opening is also seen in more plesiomorphic kiaeraspi- dians (Janvier 1981b). Such a shape implies that the hypophysial and nasal divisions are of P. L. FOREY ai m.f p. pit d.s.f tA oa LAL ANA A Re a a Fig. 2 Sclerodus pustuliferus Agassiz. Restoration of the cephalothoracic shield. Proportions based on BMNH P.9756. Scale bar in mm intervals. equal size, in contrast to some other osteostracans where there is marked inequality between these openings (Janvier 1985a: fig. 59). Several specimens show parts of the endocranial cavity, orbits and vestibular chambers, but all are poorly preserved so that only isolated details can be described. For the most part these details agree with those described for other osteostracans. The olfactory sac was housed within a deep recess (Fig. 5D) which forms the undersurface of the ridge between the orbits and the circumnasal fossa. More posteriorly, the grooves housing the anterior and posterior semi- circular canals, flanked by a groove for a large jugular vein, can be seen in BGS GSM 5150 (Fig. 4A). DOWNTONIAN OSTRACODERM SCLERODUS ‘s[BA.1OVUI WU UT 1eq BTPOS ‘666SE HNING ‘A ‘660LT'd HNWA ‘V ‘Sb2ze [eyIq10 pur [eourd ‘essoy [eseuUINOIIO Jo sIejap SuImoyYs suauTIoads Om} Jo ssuIMPI ‘ZISSeBY Sniafijnisnd snpo.aj2§ € ‘314 P. L. FOREY 7p) > DOWNTONIAN OSTRACODERM SCLERODUS ‘S[BAIOIUT WU UI Seq a[BIg “Ploy AOsUas [e19}e] 0} SUIpeay sfeuULS MOYs 0} PTarys arpeyden Jo jyey IYSLI JO MIA [SLOP : GPITS WS) S)D4 ‘a(S ‘39 ‘ES 1d :ZE6T) QISUAIS Aq payeA}sNI[! UoUTOAds aItUS :joNp jeiskydodAyoseu pal[y-xtyeu Jo joodse [e.QUOA puv SsBUI]IMS SB 1OOY [e1Iq1o BUIMOYsS JoquIeYo [eIyoULIgOTeIO Jo Jed [e1]UBO JO MOIA [BIIUPA :96L6'd HNWA ‘OD ‘soSpl [PIYOURIGIOJUI SUIMOYsS JaquIvYyo [eIyoURIqo[eIO Jo JOOI pue AjIABO UIeIG JO MAIA [eSIOP :69PTT WSD SOM “A ‘Worses sejnqysea pue s}Iqio jo 1009 ‘AYIABD UTBIQ JO MIA [BSIOP :ESTS WSD $94 ‘V ‘suoumoads ino ur paasosaid se parys otfeydad jo soinjoni}s [BUIOJUT MOYS 0} SSUIMBIP PION] eIOUIeD ‘zIssedY snuafijnisnd snpoi2}9g§ “SI 4Sa0 10 P. L. FOREY On the ventral surface the matrix infilling of the brain cavity suggests that the hypophysial duct is very long, reaching well back below the orbits and notched at the level of the anterior ends of the orbits by the entry of the internal carotid arteries (Figs 4B, C). The entry of the carotid arteries is asymmetrical, a fact which Janvier (1981b: 39) attributes to the constriction in this area caused by the proximity of the anterior cardinal veins. The material of Sclerodus is not good enough to comment on this suggestion. The floor of the orbit can be seen in BMNH P.9756 (Fig. 4C). Here it can be seen that the orbits of either side come into very close proximity with each other and may even meet, as in Tremataspis and Oeselaspis (Janvier 1985b: fig. 20). But there does not appear to be any medial recess of the posteroventral myodome (sensu Janvier 198l1a; myodome of Stensio 1927: fig. 28) such as is developed in most osteostracans (Janvier 1985a). The posterior wall of the orbit is perforated by at least two foramina which lie close together and may be confluent (Fig. 5C). They are very unequal in size, the more dorsal being the smaller. They probably gave exit to the trochlearis above and the profundus below. The oculomotor may have entered the orbit through a small foramen lying near the floor and the medial wall. The floor and part of the rear wall of the orbit is perforated by a large foramen (Fig. 5C: V,). I assume this is the trigeminal foramen through which V, and possibly VI passed. It probably also marks the place where the head vein entered the orbit, since there is no separate lateral foramen as in Norselaspis (Janvier 1981b: fig. 14A) or Belonaspis (Janvier 1977: fig. 7A). The path taken by the facial nerve marks the ventral surface of the orbit as a ridge (Fig. 4C) running anterolateraly immediately beneath the floor of the orbit. Beneath the occipital region there is a triangular depression which pierces the postbranchial wall. It is best seen in BMNH P.9756 (Fig. 4C) and was labelled by Stensi6 (1932: pl. 53, fig. 5) as the aortic groove. This is almost certainly correct, but the depression is of more complicated shape than implied by Stensio. The depression (Fig. 4C) shows a deep groove which swings to the right as it passes posteriorly. This is typical for osteostracans and carried the dorsal aorta. The position of the issuing occipital artery may be indicated on the specimen illustrated (Fig. 4C, 0.a). On the left side there is a shallower, shorter groove which appears to swing to the left. There are at least two interpretations of this groove: it could have housed the base of the subclavian artery (Janvier 1981b), or perhaps accommodated the oesophagus (Janvier 1984). The orientation of the gill chambers is of the ‘oligobranchiate’ type as defined by Stensio (1958). Only three branchial chambers are obvious on the specimens available (Fig. 4B) but there may well have been more, crowded posteriorly. Sensory canal system. The sensory lines of Sclerodus are thought to have lain entirely superficial to the exoskeleton (Denison 1951b: 214) or to have been absent altogether (Stensio 1932: 179). Certainly, no pit lines, grooves or pores mark the surface, a fact which Stensio related to the absence of the superficial layer of the dermal skeleton. There are, however, other structures which may reasonably be interpreted as evidence of the sensory line system. In specimens showing the orbital and nasohypophysial region there are often cup-shaped pits visible (Figs 3, 5, 6). Three pairs are consistently present (Fig. 3). The anterior is found at the level of the nasohypophysial opening and lies on or just outside the rim delimiting the circum- nasal fossa. The middle pit lies close to the orbital margin, roughly level with the middle of the orbit. The posterior pit lies just behind the postorbital swelling close to the edge of the dorsal sensory field. One or more of these pits may be seen in several specimens (BMNH 35999, 45949b, P.9756, P.27099, P.48704, P.58694, BGS GSM 5151 and BU 1992). These pits seem to have been overlooked by earlier investigators, since one or more are present in specimens used by Stensio and Lankester. It is possible that those authors considered the pits to be preser- vational artifacts, as many specimens show breaks in the exoskeleton. Two more recently discovered specimens (BMNH P.48704 and especially P.58694), which show the pits particu- larly clearly, demonstrate that they are not artifacts. In both specimens the rim of the pits is perfectly regular and the smooth lining is pierced by one or more minute foramina (Fig. 5). The undersurface of the left anterior pit of BMNH P.58694 (Fig. 5D) shows that the foramina pierce the base of a longitudinal groove upon the visceral surface. DOWNTONIAN OSTRACODERM SCLERODUS 11 Fig.5 Sclerodus pustuliferus Agassiz. BMNH P.58694, a specimen showing circumnasal fossa, orbits and pits: since its discovery the specimen has unfortunately been broken. A, specimen as originally collected—dorsal view, anterior towards top. B, drawing of remaining parts of specimen. C, anterior view of posterior wall of orbits. D, undersurface of fragment showing edge of naso- hypophysial opening and foramina piercing the floor of the anterior pit. Scale bars in mm inter- vals; note larger scale of C. 12 P. L. FOREY From the best preserved specimens it can be seen that each pit is ovoid to nearly circular in outline. The anterior and middle pits are of equal size and their longest diameter is about 75% as long as the nasohypophysial opening. The posterior pit is slightly smaller, being less than half the length of the nasophypophysial opening. In BMNH 45949b there is a further pit-like depression (Fig. 6A) located midway between the median line and the margin of the shield and at the transverse level of the posterior pit. This has been observed only on a single specimen, on one side only, and the borders of the pit are rather irregular; it thus may well be an artifact of preservation and it is not included in the restoration (Fig. 2). Dr Janvier (personal communication) suggested that these pits may be the location of very large tubercles which have become lost during preservation. Longitudinal rows of enlarged dentine-capped tubercles are known in cladistically more derived thyestidians and, furthermore, Denison (1951b: fig. 35b) records that in Thyestes verrucosus Eichwald each tubercle is under- lain by a deep cavity. This idea is attractive and it would certainly be good evidence for associating Sclerodus with thyestidians. However, I consider that the pits are real surface structures for three reasons. No specimen of Sclerodus shows any sign of tubercles other than the small general surface tubercles; the rim of each pit is perfectly smooth and it tips into the cavity without break; and the floor is perfectly smooth and pierced by foramina. Functional interpretation of these pits is hampered by the rarity of comparable structures in other fishes. No other osteostracan appears to have such pits, but comparison raises two possibilities. These pits may be an unusual development of the lateral sensory fields or they may be parts of the cephalic lateral line system. The first possibility seems unlikely because lateral fields, with their canal innervation, are present as in other osteostracans (see below). Also, the pits are wholly contained within the exoskeleton and therefore unlike sensory fields, vacuities passing completely through the exoskeleton and filled with small tesserae. The second possibility is more plausible. The floor of a pit (Fig. 5D) is pierced by a foramen of a size suggesting that nerves supplying neuromasts passed through. Furthermore, the pits are dis- posed in positions that, in thyestidians (sensu Janvier 1981b), would lie along the infraorbital line which lies close to the orbital margin and turns medially anterior to the nasohypophysial opening. I am therefore inclined to the view that these pits represent an unusual development of the sensory line system of Sclerodus. Sensory fields and related s.e.l. (sinus expansion of the labyrinth) canals. There have been three different interpretations of the sensory fields of Sclerodus (Fig. 1). Lankester (1870: fig. 31) recognized only the dorsal field which he described as the postorbital valley. Stensio (1932: fig. 62) identified both dorsal and lateral sensory fields, while Janvier (1975: fig. 5 //) restored a dorsal plus subdivided lateral fields. This investigation agrees with Stensi6’s results and suggests that Sclerodus possessed the usual osteostracan complement of single paired lateral fields plus a median dorsal field (Fig. 2). In no specimen are they clearly seen. The dorsal field is particularly poorly known. The anterior end is seen in BU 1992 where the margin is described by a low semicircular ridge and may (BMNH P.27099) be notched by the posterior sensory pit (Fig. 3A). The posterior limit of the dorsal sensory field remains unknown. Two specimens (BGS GSM 5150, and that figured by Stensid, 1932: pl. 52, fig. 2) show paired canals leading from the vestibular region to the dorsal field area. Stensid (1927: fig. 27A, des) has restored these canals for Kiaeraspis and Janvier (1977: fig. 9A, c.c.s.d.) for Belonaspis puella, and there is nothing to suggest conditions in Sclerodus were any different. Evidence for the presence of lateral sensory fields is provided by breaks in the exoskeleton and traces of the canals (s.e.1. canals) which lead to them. The canals may be seen most clearly in BGS GSM 5149 (Fig. 4D) among available material and they were also recorded by Stensio (1932: pl. 52, fig. 2). The partial counterpart of the specimen illustrated by Stensio is BMNH 45949b and is also illustrated by the author (1932: pl. 52, fig. 1); it can be seen that by superimposing the two illustrations the s.e.l. canals run to just within the inner margin of the space labelled as the lateral sensory field. The pattern of the canals is different in the two specimens. In the specimen illustrated by DOWNTONIAN OSTRACODERM SCLERODUS 13 Stensio the first canal is double and branches close to the lateral field; this is very similar to the pattern in thyestidians, kiaeraspidians and benneviaspidians (sensu Janvier 1981b). But in BGS GSM 5149 (Fig. 4D) the branching of the first canal occurs midway between the level of the orbit and the lateral field area, a condition which Janvier (1985a) ranks as plesiomorphic for osteostracans. Since only two specimens of Sclerodus show evidence of the s.e.l. canals it is unwise to speculate on the significance of one or the other pattern, particularly since variation is known within other thyestidian taxa (Denison 1951a). It is, however, worth remarking that the facial nerve appears to run in front of the first canal. Stensio illustrates five main canals and by comparing the relationship between the canals and the lateral fenestrations it appears that the posterior two are not seen in BGS GSM 5149, probably as a preservational defect. It also appears as though the s.e.l. canals radiate regularly from the otic region: that is, they are not branched into two distinct groups as they are in Oeselapis and Tremataspis where there are two separate sensory areas. A number of other specimens (BMNH 45949b, P.9756, P.41095, BU 1992) show evidence of the lateral sensory field as depressions or irregularly-shaped vacuities in the exoskeleton (Fig. 7A). These specimens show that the sensory field stretched from the level of the first marginal fenestration to the third. The lateral border is quite distinct but the inner margin is somewhat irregular. The size and extent of the lateral field is similar to that seen in Dartmuthia, Saare- maspis and Thyestes. The ‘cornua’. Lankester (1870) and Stensio (1932) both considered that the shield of Sclerodus continued posterolaterally on either side as long cornua. Stensid believed that the ‘cornw’ bordered a pectoral fenestra containing a fin. A countertheory (Denison 1951la, Janvier 1975) suggests that the so-called ‘cornu’ is really only the lateral margin of the cephalothoracic shield, that there were no pectoral fenestrae containing fins and that the area between the ‘cornua’ was occupied by an unarmoured abdomen. A number of observations suggest to me that the latter theory is correct. The ‘cornu’ is highly asymmetrical in cross section such that the ventral surface is flat, or nearly so, the dorsal and mesial surfaces are concave and the mesial edge is considerably deeper than the lateral edge. The cross-sectional shape looks like the sectioned edge of the cephalic shield. The lateral edge bears a single row of well-developed tubercles. The dorsal and ventral surfaces bear regular small tubercles. But tubercles are absent from the mesial surface which is instead perfectly smooth. These observations contrast with the cross-sectional aspect of true cornua as seen in most cornuate osteostracans. There, the shape is roughly symmetrical and is flattened, the surfaces are all convex to a greater or less degree, and the ornament continues on to the mesial surface and is usually developed as a series of enlarged tubercles. The medial wall of the ‘cornu’ sweeps anteromedially to merge with the postbranchial wall. If a pectoral fin were present there should be some sign of insertion as seen in Boreaspis (Janvier 1977) or Benneviaspis (Janvier 1985a). But in two specimens (BMNH P.45315, BGS GSM 5149) showing this area clearly the bone is perfectly smooth. One final observation is that the ‘cornua’ of Sclerodus are solid structures (Stensid 1932: pl. 56, fig. 1). Large cornua, such as are seen in cephalaspids and scolenaspids, are penetrated by several large canals thought to have contained various blood vessels (Wangsj6 1952: fig. 17). Thus, as restored, I believe that the area of the body between the ‘cornua’ was naked and suggest that the exoskeleton was coextensive with the endoskeleton. In both Tremataspis and Oeselaspis the endoskeleton of the cephalic portion curves posterolaterally to line the edge of the shield: the development of the endoskeleton is particularly extensive in Didymaspis (Janvier 1985b: fig. 19). Lateral fenestrations. The lateral margin of the cephalic shield is marked with fenestrations. Lankester (1870: 58) considered that these fenestrations were cells within the exoskeleton and were therefore roofed and floored by bone. Stensid (1927: 240) originally interpreted them as remnants of a much subdivided lateral sensory field, but subsequently changed his mind. Stensi6 (1932) and Denison (1951a) considered they were true holes passing through the shield 14 P. L. FOREY Fig. 6 Sclerodus pustuliferus Agassiz. Outline drawings of three specimens to show relative sizes and positions of nasohypophysial opening, dorsal and lateral sensory fields, pits and marginal fenes- trations. A, BMNH 45949b, latex cast. B, BU 1992, latex cast. C, BMNH P.9756. Sensory fields— heavy stipple: area of shield—light stipple: sensory pits—black. Scale bars in mm intervals. from top to bottom. Examination of specimens used in this study suggests this interpretation to be correct. Fenestrae are seen clearly in many specimens, and where the lateral margin of the shield is complete it is obvious that there are four such fenestrae (BU 1992, BGS GSM 89283, GSM 21469, GSM 5149. GSM 5150, BMNH P.9752, P.9756, P.9757, P.9758, P.49015, 45949b). DOWNTONIAN OSTRACODERM SCLERODUS 15 Lankester (1870: fig. 31) showed six fenestrations, but none of the specimens used by him or any of the specimens used here show so many; the anterior two he showed are not present in any specimen. The most anterior fenestra is located at the transverse level of the naso- hypophysial opening and is elongate. The second, third and fourth become progressively smaller and more equidimensional (Figs. 2, 4D, 6, 7). The apparent regularity prompted an attempt to express the area of the posterior three fenestrae relative to the first (most anterior), which is always the largest in the series. The results obtained were very variable and this is probably because different specimens have been broken at different horizontal levels through the thickness of the shield. As an average of eight of the best preserved specimens, the area of the second fenestra is 75% of the first, the area of the third 56% and the fourth 46% of the first. That the fenestrae passed right through the shield is not immediately obvious from the specimens available. The majority are preserved in dorsal view, in which it can be clearly seen that the ornament tips into the posterior walls of at least the anterior three fenestrae. Unfor- tunately the few specimens showing the ventral aspect are broken so it is not clear that the fenestrae reappear on the ventral surface. However, in these specimens the matrix infilling of the fenestrae stands well proud, implying considerable depth (P.49015, P.9756). The most direct evidence is provided by BMNH P.3247 (Fig. 7B). In this specimen the anterior end is broken through the last fenestra and it shows the walls of the fenestra passing without interruption from one surface to the other. The posterior wall of each fenestration slopes anteroventrally and it appears that the slope is greatest within the anteriormost fenestra and becomes progressively more upright in more posterior fenestrae until the rear wall of the fourth fenestra is nearly vertical. Distortions of individual specimens preclude any attempt to measure precise angles. The anterior wall of each fenestra passes nearly vertically or only slightly anteroventrally through the shield. Several specimens (BMNH 45949b, P.9758, P.49015) show that the posterior wall of each fenestra, except perhaps the last, is ornamented with fine tubercles, considerably smaller than those covering the adjacent part of the shield. In BMNH 45949b (Fig. 7A) there is a clear line of division between the fine tuberculations lining the fenestra and the shield surface, suggesting that there may have been a small separate plate forming the rear wall of the fenestra, but this observation could not be confirmed on any other specimen. Despite this uncertainty the exis- tence of an ornamented lining reinforces the view that they are true fenestrations rather than depressions or ‘cells’ within the structure of the bone. The regularity of these fenestrae suggests that they were functionally important but it is difficult to be certain what this function may have been. There are no other osteostracans with such fenestrae, nor indeed are there many other animals showing such structures. The most obvious modern analogues are the marginal lunules in some clypeasteroid echinoids (sand dollars), the structure, evolution and possible functions of which have been discussed by Smith & Ghiold (1982). It is not possible to stretch comparison between lunulate echinoids and Sclerodus too far. There are quite considerable differences: unlike the fenestrations of Sclerodus the lunules of sand dollars are of roughly equal size, and in life they are partly filled with a thick epidermis containing spines and pedicellariae. Smith & Ghiold (1982) review the various hypotheses of echinoid lunule function. They are careful to point out that there may be a difference between the function of the anal lunule and the marginal lunules, which would be more comparable to those in Sclerodus. For the marginal lunules some seven hypotheses have been suggested (Smith & Ghiold 1982: 244-246). From their discussion those suggesting involvement with food gathering may be ignored. The most likely hypotheses for Sclerodus are hydrodynamic and, perhaps, assistance in burial, because these functions simply rely on lunule space and no associated epidermal structures. Furthermore the definite anteroposterior gra- dient in fenestra size (not seen incidentally in sand dollars) might suggest a hydrodynamic function. Experiments on sand dollar tests have been carried out in wind tunnels and flume tanks (Telford 1981, 1983). It must be emphasized that these experiments treat the sand dollar simply as a geometrical shape, a fact which critics of hydrodynamic theories are quick to point out. But, accepting these parameters, the results indicate that the overall shape of a flat undersurface “S[@AIOUI WU UL SIE J[VIS “UOT}eIJSOUIJ JSOWIOTIO}SOd Jo [[eM 1vd1 SUIMOYS NUIOS, UdYO1g JO MOIA JOLIOUR ‘/p7E'd HNWA ‘A 9914} JOMIO}Ue oY} JseO] 1e UT [[eM Jeol Po]USWeUIO YIM SUOT}I}SAUI] [e19}e] MOYS 0} (Y9 “3I4) G6P6Shb HNING JO 28pe oy) JO 1S¥d xo}R] JO SUIMVIP “Y ‘ZISSedY Snuafijnisnd snpowajIg LL “BLY }°S*| DOWNTONIAN OSTRACODERM SCLERODUS 7) and convex upper surface creates lift in a water current which is counteracted by the lunules. Some experiments (Telford 1983) suggest that the critical velocity (the water speed at which the sand dollar is lifted) might be increased by some 30%, implying that lunules could have a significant stabilizing effect; calculations made by Telford (1981: 619) suggest that moderate- sized sand dollars (7cm diameter) may well experience critical velocities in many modern shallow water environments. There is little point in trying to perform similar experiments on a Sclerodus model since the shield is only part of the animal. But it is possible that the fenestrae provided some similar stabilizing function appropriate to a fish presumed to have lived in littoral waters and presum- ably subject to varying water speeds. The sectioned shape of Sclerodus is certainly that which would create lift if left uncompensated. Relationships of Sclerodus In this paper Sclerodus is placed in the family Tremataspididae. This requires some explanation in view of the very different hypotheses of relationships of Tremataspis and allies (Westoll 1945; Denison 1951a; Halstead 1982, 1985; Halstead Tarlo 1967; Janvier 1981a, 1985a, b, c) and the fact that Sclerodus is often placed in its own monotypic family. The assignment of Sclerodus to a distinct family (Sclerodidae Berg 1940, Sclerodontidae Fowler 1947) is no more than a recognition of its uniqueness, which cannot be denied, but does not imply much about relationship. Stensid (1932: 176) suggested it to be closely related to Thyestes and Didymaspis because the facial nerve (called V, by StensiO) runs anterior to the first s.e.l. canal. This character is now known to be more widely distributed (Procephalaspis, Witaaspis, Oeselaspis, Tremataspis) but may still be significant, depending on whether one rates these taxa as constituting a monophyletic (Janvier 1981b), polyphyletic (Denison 1951a) or paraphyletic group (Halstead Tarlo 1967). Denison (195la: fig. 31) regarded tremataspids (Didymaspis, Tremataspis, Saaremaspis, Dartmuthia) as an ancestral group from which at least four different lineages of osteostracans evolved. Denison (1951a: 180) acknowledged that his group Tremataspidae was not necessarily a natural assemblage but that it only emphasised ‘... the convergence towards a central ancestral type’. He arrived at his conception of the ancestral type by determining polarity of several different character transformations using stratigraphical occurrence as the arbiter. Thus, he observed that the majority of Ludlovian osteostracans have small lateral fields, relatively few s.e.l. canals and a short prepineal region; the converse conditions would be derived. He admit- ted that the stratigraphical occurrence did not resolve whether the primitive osteostracan shield was long or short, or whether paired fins were primitively present or absent. But he decided that because Tremataspis showed the primitive condition of lateral fields, s.e.l. canals and prepineal length, then a long shield and absence of paired fins must also be primitive. Denison’s is the most clearly reasoned advocacy of the stratigraphical argument and the primitiveness of tremataspids, a view shared by Westoll (1945, 1985) and Halstead Tarlo (1967). Sclerodus shares many of these ‘primitive’ features such as a long carapace, short prepineal region, no paired fins or cornua and relatively short lateral sensory fields. But Sclerodus cannot be classified with tremataspids purely on the basis of ‘primitive’ features, since on these terms it would only mean that Sclerodus looked something like the ancestral osteostracan. Janvier (1985a, b, c) has criticized this stratigraphical approach to character phylogeny in osteostracans by pointing out that forms such as Ateleaspis, with paired fins and large lateral fields, and Procephalaspis, with cornua and paired fins, occur contemporaneously with or even earlier than Tremataspis. Halstead (1985) introduced another line of argument by claiming that Tremataspis shows a primitive geometry of the cephalic shield since, in gross outline, it resembles a cyathaspid heterostracan. If this doubtful reasoning is to mean anything then, presumably, its import lies in character distribution. Thus, if it could be shown that the Tremataspis/Cyathaspis-shaped shield was widely distributed amongst primitive members of the jawless fish groups then there might be some justification in assuming it to be a generalized feature. However, irrespective of 18 P. L. FOREY which proposed phylogeny of jawless fishes one accepts (Forey 1984: fig. 3) the ‘primitive’ nature of the Tremataspis/Cyathaspis geometry cannot be justified on grounds of commonality. If stratigraphy and commonality fail us then we are left with congruence of character distribution as the overriding criterion of choice: this has been the line of argument adopted by Janvier (198la, b; 1985a, c). He concludes, like Stensio, that Tremataspis and traditionally- accepted related genera are derived osteostracans. Janvier’s approach is cladistic classification and he has attempted to determine plesiomorphic and apomorphic states, and then to check these against congruence. Janvier (1985a) suggests that non-cornuate genera such as Ateleaspis and Aceraspis are primitive because they exhibit micromery on the undersurface of the head, broad-based pectoral fins not flanked by cornua, and two dorsal fins. These features are generalized, based on outgroup comparison. Using this assumption Janvier’s phylogeny of osteostracans (1985a: fig. 69) rates tremataspids as derived cornuate forms which have second- arily lost pectoral fins, reduced the number of s.e.l. canals and developed an elongate carapace. Furthermore, Janvier considers that the sister-group of tremataspids is Thyestes, with forms such as Witaaspis, Auchenaspis salteri and Procephalaspis as progressively more plesiomorphic forms. He refers to this entire assemblage as thyestidians. Janvier’s thyestidians include tremataspids as well as forms which Denison (1951la: fig. 31) regards as ancestral to the Ateleaspidae (non-cornuate osteostracans with paired fins) and Cephalaspidae (including benneviaspidians). Janvier’s classification with respect to tremataspids agrees with Stensio (1958), and is almost the antithesis of that of Denison (and also Halstead Tarlo 1967). In reaction to some recent criticism (Westoll 1985, Halstead 1985) Janvier (1985c: fig. 36) translated Denison’s (1951a) tree into a cladogram and detailed some 14 incongruous character distributions which result. I was interested to see what might happen if some of the data presented by Janvier (1985a, c) were subjected to cladistic computer analysis using PAUP (Phylogenetic Analysis using Parsimony) version 2.2, a program prepared by Dr David Swofford, which is designed to select the most parsimonious tree or trees which can be rooted to follow the fate of different character transformations. I chose to look at 16 taxa with respect to 27 characters. Some of them were higher taxa (cephalaspidians, kiaeraspidians, scolenaspidians, benneviaspidians, tremataspids sensu stricto), and it was therefore assumed that these groups are monophyletic. This may, of course, be disputed but Janvier (1985a) has discussed the arguments and I find his reasoning sound. More importantly, the advocates of competing theories also accept these groups (Denison 1951a, Halstead Tarlo 1967). Groups about which there is argument include the Ateleaspidae (Ateleaspis, Aceraspis, Hirella, Hemiteleaspis and Hemicyclaspis) and osteostracans traditionally associated with Tremataspis (Auchenaspis', Witaaspis, Thyestes, Didymaspis). Denison (1951a) and Ritchie (1967: 79) regarded the Ateleaspidae as monophyletic and derived from Witaaspis or the tremataspid Saaremaspis. Janvier, however, regards ateleaspid genera as primitive osteostracans forming a paraphyletic assemblage, with some being more nearly related to cornuate osteostracans than to other ateleaspids. In other words, for this analysis I have chosen to designate separate genera in those areas where classifications are substantially different. Another problem area is character designation. As Janvier (1985a) implies in his classification there are several well-defined groups of osteostracans, but there is a problem of identifying characters to link groups other than those which are general to osteostracans. Thus, there may be polychotomies within osteostracan classification simply because there are no identifiable characters to resolve the issue further. This is a problem distinct from conflicting character distribution. The final difficulty stems from the uncertainty of distinguishing polarity of charac- ter state transformations. This is, of course, the source of most disagreements between conflict- ing classifications (see above) and is particularly difficult to resolve in an extinct group such as osteostracans. " For many years Auchenaspis and Thyestes have been regarded as synonyms (Woodward 1891: 195). Janvier (1985a: 122), however, retains Auchenaspis salteri Egerton and Auchenaspis egertoni Lankester as distinct from Thyestes verrucosus Eichwald. Janvier recognizes several synapomorphies of Thyestes verrucosus and Tremataspididae not present in species of Auchenaspis. DOWNTONIAN OSTRACODERM SCLERODUS 19 The most obvious features which can be compared amongst osteostracans are size, shape and complexity of the dorsal and lateral sensory fields, the canals leading to them and their relationship to cranial nerves. Additionally, there is variation in the development of the cornua and the trunk shield. Understandably, classifications have used these features. But since non- osteostracans do not have sensory fields or related canals, and the cornua are not easily compared with the skeletal outgrowths in other groups (e.g. spinals of placoderms or the cornual plates of pteraspidiform heterostracans), the polarity of many features associated with these structures is not resolvable by outgroup comparison. The computer program built the ‘tree’, paying no regard to the entered polarity even though the data had been scored, in large part, in agreement with Janvier’s assessment of primitive (0) or derived (1). The derived characters used in the program were: hs 2 \O oo Pectoral fins present, as evinced by sinus and/or area of attachment. Presence of pectoral fins in Didymaspis after Janvier (1985a). Dorsal field separated from pineal plate. Converse condition—pineal plate contacting dorsal field—regarded as plesiomorphic within osteostracans. There are some intragroup exceptions where, for instance, nearly all members show one condition (e.g. benneviaspi- dians show pineal contacting the dorsal field) with one species (Benneviaspis holtedahli Stensi6) showing the derived condition. . Tesserae on undersurface of oralobranchial chamber. The plesiomorphic condition is micromery where there is a shagreen of minute scales exemplified in, for instance, Atel- easpis. Like Janvier (1984, 1985a) I feel confident about the polarity of this character since micromery (covering of small, equal-sized and regular-shaped units) is widespread amongst agnathan groups and primitive gnathostomes. The tesserate condition is, on the other hand, regarded as derived and is exemplified by Saaremaspis (Janvier 1985b: fig. 16) or Hemicyclaspis murchisoni (Egerton) (Stensio 1932: pl. 7, fig. 3). Here the covering of the oralobranchial chamber is made up of irregularly-sized and irregularly-shaped units. . Pineal plate equidimensional or longer than broad. There are some intragroup exceptions which must be regarded as secondary reversals. For example, tremataspids generally show the derived condition but Timanaspis is exceptional. . Pineal plate absent. There are some intragroup exceptions; for instance, amongst cephal- aspidians, which generally have a well-developed plate, this has been secondarily lost in Hildenaspis and Mimetaspis. . Orthobranchiate condition. . Pattern of branching of the first canal leading to the lateral sensory field. There are three conditions (Janvier 1985a: 107) but the polarity of transformation is by no means clear. For this reason the character is scored quite arbitrarily here: 0 = branching near lateral field, 1 = branching midway between eye and lateral field, 2 = branching near orbit. . Abdomen with scale-covered ventrolateral crest. . Cornual process. The development of the cornual process is regarded as a derived condi- tion. There are problems with identifying a cornual process in kiaeraspidians but I follow Janvier (1981b) in believing the cornual process to be primitive for that group. . Long abdominal division of cephalothoracic shield (more than two segments incorporated into the shield). . Facial nerve running alongside or in front of first canal to the lateral sensory field. . Abdominal part of shield closed ventrally (may be secondarily reduced in extent in some, e.g. Nectaspis). . Branchial nerves penetrating gill chamber laterally. The condition of this character is only known sporadically throughout osteostracans. . Extrabranchial divisions large. . Opening of endolymphatic duct lying outside dorsal sensory field. The converse condition is considered plesiomorphic because it is more widely distributed amongst osteostracans. Some benneviaspidians and also Didymaspis have openings on the edge of the sensory field; these are considered to show the plesiomorphic condition. 20 P. L. FOREY 16. Lateral sensory fields not extending greatly beyond level of nasohypophysial opening. 17. Lateral sensory fields reaching posteriorly well beyond level of dorsal sensory field. 18. Lateral sensory fields posteriorly expanded. 19. Supraoral fields with denticles. Condition is only known sporadically throughout osteostracans. 20. Anterior dorsal fin or fin scale absent. This character is regarded as unquestionably derived. Most primitive members of agnathan groups have two dorsal fins. 21. Solid rim to the shield. 22. Infraorbital line stopping short of lateral sensory field. 23. Posteroventral ‘myodome’ absent. See Janvier (1985a: 77) for discussion. 24. Infraorbital line running close to circumnasal fossa. 25. Paired fins constricted at base or separated from trunk scales. 26. Horizontal perforated lamina within the sensory canals of the middle layer of exoskeleton. 27. Enameloid layer. The data matrix, as given in Table 1, includes the characters used by advocates of opposing hypotheses. The difference is that Janvier would choose Ateleaspis as the root of the tree whereas Denison, Westoll and Halstead would favour tremataspids. So the computer program was run twice using a different root. On each occasion there were 60 equally parsimonious trees, this being a reflection of the relatively poor quality of the data (approximately 30% homoplasy, and some possible dichotomies unsupported by characters—see below). The con- sensus trees (the common element of the 60 most parsimonious trees) are shown in Fig. 8 where the root is fixed at Ateleaspis on the left and tremataspids on the right. The first observation is that the computer-generated tree, based on parsimony and using tremataspids as ancestor, is not at all like the tree advocated by Halstead Tarlo (1967) as illustrated in Fig. 9. Halstead Tarlo’s tree is less highly resolved and, potentially, there may be considerably more dissimilarity between the two solutions presented in this figure: the major areas of difference may, however, be briefly mentioned. Halstead Tarlo’s tree ranks cephalaspi- dians, kiaeraspidians and benneviaspidians as a trichotomy and as the most derived osteostracans. The consensus tree ranks these as successively more plesiomorphic sister-groups Table 1 Character data matrix for 16 osteostracan taxa. For explanation of characters see text. A, Ateleaspis; B, Aceraspis; C, Hirella; D, Hemiteleaspis; E, cephalaspidians; F, kiaeraspidians; G, scolena- spidians; H, benneviaspidians; I, tremataspids; J, Procephalaspis; K, Auchenaspis salteri; L, A. egertoni; M, Witaaspis; N, Thyestes; O, Didymaspis; P, Hemicyclaspis. Characters Taxon 12354 SOS SOI BIA MISC N IMSS 202122384526) 27 A ro? Oo oo ? it OG? @O0O OO POO OOOO? OO OD B lQ? OOOO i tO OO OO OOOO OOOO LO? OO O O Cc i Oil Oo ?7il tO OO OC OOO OO O lt OP? O i O D D LO tO? © ? 1O OO O1OO OO OO O t © 20 I O O E i Ot@®oo2iio@ogqoroeoogigdooit@ogooid@d ds F a PEO? 10 OTS a ORO) 10) 210 OR lO ae Oe 0 G LOQkh OO @® i it Ogi om @@Mi it Oi t ti O@ Oil @ OY H VO Se sO tas 0 Ht iho Ov 1 OO O M i @ 2 wat © @ t © @ I OCW 2 LO) 20 MOLIOD SO 1 Oe LS £0) RO A Sees On J bY 0% 1,40) 0 0 20) 1 oh OF A On ORIOL Os Oe 1 ees) K LO: 1. 0:20) O70) Ae 1a ZO OO ail ORO Oiler ee L 1 O f 0 OO OO 1 Ly We OR OR LO nO iO ela 1s re eee) M bolt, 2 Oy 10M 10) ty Oe A A Oe I OY 20) Oy esl as? so iia eal N 1 te ah tO) O70) a 1S AP Oe I ee Ol Oi Te Ts i ie le ot ool | O ft tO ff 20? 2 1 te Oy 0 LOK OF 0) le Ie) 50)5 a P tok o@o@gdit® ooo O OO OOOO Oil ®O PO il @© DOWNTONIAN OSTRACODERM SCLERODUS Dil Ateleaspis as ancestor C *tremataspids’ as ancestor =12 P -22, -20 E -23,-14 -16, -11 G -24 21 H -4,-15,-26, -27 3 fi 1,8 12 O 20, 22 é AF 23,14 K~ 11,16, 26 i 24 L 415,27 -M Fig. 8 Consensus tree produced from the 60 most parsimonious alternative solutions. Topography of the trees is identical. Left-the root placed at Ateleaspis. Right-the root placed at tremataspids. Only those characters treated as synapomorphies are shown. Taxa A-P as in Table 1. Numbers refer to characters detailed in text. Particular taxa under consideration are enclosed within stippled area. to the Ateleaspidae. Thyestes occupies a very different position, being a derived taxon and the sister-group to cephalaspidians + kiaeraspidians + benneviaspidians in the Halstead Tarlo tree, while in the consensus tree it is the sister-group to all other osteostracans with the exception of tremataspids. Thus, whatever else the Halstead Tarlo tree might contain, it does not approach a parsimonious solution given the data used here (Fig. 9). In Fig. 8 only the characters used once (synapomorphies) are applied to the diagrams. Those against the ‘tremataspids as ancestor’ tree are largely indicated as negative features, but this is a consequence of the way in which the characters were coded in the first place. Perusing this list we may note that some characters (11, 15, 16, 24, 26) are only found in osteostracans and it is therefore difficult to establish polarity. But one prediction of fixing the root at tremataspids is the deduction that the primitive osteostracan developed a complex horizontal lamina within the exoskeleton which was subsequently lost by most osteostracans. The alternative assumption 22 P. L. FOREY © 6 © iS . iS 2 y Ry g ~ Ss oS ~ © 5S @ oO RG Q ) g © 2 ® No 6 Q S o _‘ ~ K } Q 8 @ o ® @ Pm) 3 ~ M2 2 2) @ 4 bd 2 CQ) 2 ® & 2 ® Nd S Ae) o 5 ® © i ee eM Oe ® ‘ LS © RY o Sar 1Oan, VS ees aS uae ei aes after Halstead Tarlo (1967) © © ® % < > oO ~? 2 > a ® > > x Q ~ " > Q ~ -) S 6 o % Q ~ ©) Q % 2 “ 2 2 2 ” SN! ° we a & a ® 2 & @ e ° &) aS o we 2 o ~ oS g o Y < Ss S CO o < A 9 a2 Se) (2) zt zy zy vt v consensus tree Fig. 9 Tree of osteostracan taxa produced by Halstead Tarlo (1967) compared with computer consensus tree rooted at tremataspids. (Ateleaspis as ancestor) rates this character as an acquisition within a small group of osteostracans and so appears a more plausible alternative. My main reasons for preferring the ‘Ateleaspis tree’ centre on the behaviour of characters 3, 12 and 20. I believe that micromery is a primitive condition and that the tesserate (3) condition is derived. The micromeric undersurface of Ateleaspis and Aceraspis may be associated with the very small trunk scales in these forms which is probably also a plesiomorphic feature (Janvier 1985a: 106). I also believe that the ventral enclosure of the abdominal region (12) is a derived DOWNTONIAN OSTRACODERM SCLERODUS 23 condition. And, finally, I consider that the presence of two dorsal fins (20) is a plesiomorphic feature, this being found in lampreys, some placoderms and the more primitive acanthodians. Many osteostracans have enlarged dorsal scales in place of one or both dorsal fins. Janvier (1984) uses the presence of such scales as a character. For instance, in the solution given in Fig. 10, Janvier would suggest that a character linking Hemicyclaspis and cladistically more derived taxa would be the presence of a modified scale in place of the anterior dorsal fin. This is perfectly acceptable and would be one way of resolving what is otherwise a trichotomy between some ateleaspidian genera (Fig. 8). It should also be noted that the presence of paired fins is here regarded as plesiomorphic for osteostracans, based on a higher level phylogeny which ranks gnathostomes and osteostracans as sister-groups (Janvier 1981c, Forey 1984). In Fig. 10 one of the ‘Ateleaspis trees’ is given in full and all characters (except no. 7— branching of s.e.l. canals) are included. It can be seen that the node linking Hemicyclaspis and cladistically more derived osteostracans is not supported by any characters, and that linking Hemiteleaspis and more derived osteostracans is only supported by character 7 which is very difficult to evaluate. These areas of uncertainty give rise to many of the alternative trees and are probably better depicted as a polychotomy incorporating Hemiteleaspis, Hemicyclaspis, Hirella and cladistically more derived taxa. The inclusion of additional characters (e.g. inferred modifi- cation of the anterior dorsal fin) may partially resolve this polychotomy. Another area in which alternative trees varied concerns benneviaspidians, scolenaspidians and cladistically more derived osteostracans. The solution shown in Fig. 10 suggests bennevi- aspidians and scolenaspidians to be sister-groups, based on the common possession of posteriorly-expanded lateral sensory fields (18). The alternative solution ranks benneviaspidians as the sister-group to cladistically more derived forms, with scolenaspidians as the plesiomor- phic sister-group. This is the solution preferred by Janvier (1985a), who bases it on the fact that in benneviaspidians and cladistically more derived forms the first s.e.l. canal branches near the lateral sensory field. The trichotomy shown in the consensus tree (Fig. 8) is therefore the result of conflicting character distributions which may only be resolved by discovering more characters. Character 7—the branching pattern of the first s.e.l. canal—was entered as a multistate character, but the resulting transformations implied by the computer tree were very ambiguous. The primitive condition was fixed with reference to the condition in Aceraspis (see also Janvier 1985a), in which the canal branches midway between the eye and the lateral sensory field. A transformation of this presumed general condition in Hemiteleaspis and cladistically more derived forms shows branching very near the orbit (best exemplified in cephalaspidians). An even more restricted grouping—scolenaspidians, benneviaspidians and their sister-group— show transformation to a canal which branches near to the lateral field. Scolenaspidians show a reversal to ‘primitive’ conditions. None of the conditions of the branching pattern is coextensive with any of the groups specified in Fig. 10. This character might therefore need re-examination in the light of the classification proposed here. In view of the fact that it is sometimes difficult to be certain whether the canal branches midway between the eye and the lateral field or whether it is nearer one than the other, and of the fact that there may be variation within a single species (p. 13), this character is not considered further even though it has traditionally been used in classi- fications of osteostracans. Characters 5, 6, 13, 17 are treated as parallelisms. The behaviour of character 6 (orthobranchiate condition) is perhaps interesting. This is a character, used by Stensio (1958), which might suggest that kiaeraspidians and benneviaspidians are sister-groups (see alternative in Janvier 1981b: fig. 43), but none of the computer-generated trees suggested this grouping. The last area of uncertainty within the computer-generated tree concerns tremataspids, Witaaspis and Thyestes. The solution shown here (Fig. 10) ranks tremataspids and Witaaspis as sister-groups based on the secondary loss of the cornual processes (9). The alternative solution places tremataspids and Thyestes as sister-groups because of the common possession of a denticulated supraoral field (11). Neither character is clear cut: it is sometimes very difficult to be certain whether cornual processes are short or absent altogether, and the condition of the P. L. FOREY Fig. 10 One of the 60 most parsimonious trees rooted at Ateleaspis with all characters and charac- ter transformations shown. Synapomorphies designated with ‘prime dot’. Other characters are parallelisms or reversals (minus signs). Character 7 omitted. See text for list and discussion of characters. DOWNTONIAN OSTRACODERM SCLERODUS 25 TREMATASPIDIDAE Fig. 11 Systematic position of Sclerodus inserted in a phylogeny of thyestidians produced by Janvier (1985c: fig. 40) with synapomorphies specified by that author—his numbering is used here. The synapomorphies are as follows: l-infraorbital sensory canal running close to orbit and circumnasal fossa, 2—canal for facial nerve not running in orbit, 3—medial recess of posteroventral myodome absent, 4—pineal plate narrow or short, 5—horizontal lamina devel- oped within exoskeleton, 6—abdominal division long or very long, 7-openings of endolym- phatic ducts outside dorsal sensory field, 8— pineal plate very short, 9-dorsal sensory field separated from pineal recess, 10- nasohypophysial opening short or very short, 11—abdominal division very long, 12-superficial enameloid layer developed, 13—longitudinal rows of enlarged tubercles developed, 14- supraoral field triangular with denticles, 15— absence of paired fins, 16—circumnasal fossa deep and elliptical, 17—cosmine’ forming a con- tinuous layer, 18—dorsal sensory field very short, 19-circumnasal fossa very short, 20- dorsal sensory field extremely short, 21—lateral sensory field divided into two parts. See text for discussion. supraoral field is poorly known in many osteostracans. Janvier (1985c: fig. 40) prefers the second solution; he suggests that, in addition to the denticulated supraoral field, Thyestes and tremataspids show a slightly longer abdominal division and a shorter nasohypophysial opening. I find these characters difficult to evaluate, but they could be one way of resolving a trichotomy shown in the consensus tree. The important conclusion to be drawn is that, despite the differences between the computer-generated tree and Janvier’s classification (1985a: fig. 69), both firmly support thyestidians as a group. The implication for discussion about the interrelationships of Sclerodus is that I feel entitled to regard Tremataspididae as a monophyletic taxon to which additional taxa can be added in pectinate fashion as specified by Janvier (1985a, c); see Fig. 11. There are, as Janvier freely admits, some problems with this classification; the greatest is perhaps incomplete knowledge of 26 P. L. FOREY morphology in certain forms. But given these constraints, Sclerodus, which is particularly poorly known, can be placed within the Tremataspididae as the sister-group of Dartmuthia, or of Saaremaspis, Tremataspis and Oeselaspis. These conflicting solutions are shown as a tri- chotomy in Fig. 11. With respect to the cladogram of thyestidians produced by Janvier (Fig. 11), Sclerodus agrees in showing the synapomorphies numbered 2, 3, 9, 15, 16. Characters numbered 4 and 8 refer to the shape of the pineal plate, or the pineal recess when the plate has never been found (as in Sclerodus). These two characters are really differing degrees of development of the same feature and Sclerodus would appear to match that specified under character 8 most closely, exemplified by Witaaspis and Thyestes. Character 1—medial course of the infraorbital line—depends on interpretation of the pits within the shield (see p. 12). Characters numbered 6 and 11 refer to progressive lengthening of the abdominal shield which is also seen in kiaeraspidians. Sclerodus certainly shows a long abdominal division but this is only developed laterally. Characters 7, 18, 20 refer to the size and shape of the dorsal sensory field, while character 14 concerns the supraoral field. These structures are unknown or too poorly known in Sclerodus to assess their status. Character 13 (longitudinal rows of enlarged tubercles), absent in Sclerodus (see p. 13), must be considered as a reversal. Character 21 (divided lateral sensory fields) is known to be absent; but the status of this character as a synapomorphy must be questioned since it is present in kiaeraspidians. Characters 5, 12, 17 refer to details of histology, an aspect in which Sclerodus appears unique in several respects. There is one further feature of Sclerodus which recalls conditions in thyestidians. The lateral head vein (Fig. 4A) runs well outside the vestibular chamber and makes a broad medial sweep at the level of the orbit. This feature has not been considered in the above discussion on osteostracan classification because this part of the anatomy remains poorly known in most species. Conclusions The comparative information available for Sclerodus suggests that it is a member of the Trema- taspididae, which is here accepted as a monophyletic group. The Appendix (opposite) lists the stratigraphical occurrence of the Sclerodus specimens used in this study. In addition to this Sclerodus has been reported from the Ludlow Bone Bed of Brook House, Llangibby, Gwent; Downton Castle Sandstone of Beech Hill Farm, Usk, Mon- mouth; also Turners Hill (Temeside Beds), south Staffordshire (Ball 1951). Thus, Sclerodus is restricted to the Downtonian of the Anglo-Welsh depositional basin. It is also apparent that, even allowing for collecting bias at long-known and well-collected sites such as the Ludlow Bone Bed exposure at Ludford Lane, Ludlow, Shropshire, the majority of the specimens are found in the Ludlow Bone Bed and in the lower part of the Downton Castle Sandstone. Very few have been found in the overlying Lower Red Downton Group or its equivalent, the Temeside Shale. Even the listing of the specimens from Wallop Hall as coming from the Temeside Shale may have to be revised, to place them in the Downton Castle Sandstone (Dr J. B. Richardson, personal communication). This, and the evidence of the associated fauna and sedimentological features, indicate that Sclerodus was a marine fish becoming rare with the onset of brackish water conditions (Allen & Tarlo 1963). The vertebrates most commonly associated with Sclerodus are acanthodians, Cyathaspis banski and thelodonts. In terms of Turner’s (1973) thelodont faunas Sclerodus would be part of the upper part of the Thelodus parvidens fauna. Osteostracan congeners are Auchenaspis (Thyestes) and Hemicyclaspis, which are found in all but the lowermost levels of the Downton Castle Sandstone, and Didymaspis which appears in the overlying Lower Red Downton Group. Thus, Sclerodus is the earliest osteostracan to appear in the Anglo-Wesh basin and is one of the truly marine osteostracans. It appears to be the ecological equivalent of the Wenlock and Ludlovian thyestidians from Estonia which Marss & Einasto (1978) suggest occupied shallow lagoonal waters shoreward of sand-belt facies. Janvier (1985c: 211) suggested that Tremataspis and other derived tremataspids inhabiting these Baltic waters may have been burrowing forms. DOWNTONIAN OSTRACODERM SCLERODUS 27 These osteostracans have rather convex ventral surfaces such that the cross-sectional profile would be elliptical. I do not think that Sclerodus showed such a convexity; rather it is possible that it was able to submerge itself beneath the loose surface sand, and that the marginal fenestrae may have helped in this activity. Acknowledgements I would like to thank Drs Colin Patterson, British Museum (Natural History), and Philippe Janvier, Institut de Paleontologie du Muséum National d’Histoire Naturelle, for encouragement and reading the manuscript, and Dr L. B. Halstead, Reading University, for bringing to my attention the fact that Dr W. Graham-Smith had recovered some interesting fragments of Sclerodus. My especial thanks are due to Dr Graham-Smith for donating his material. Dr C. J. Humphries and Mr A. Paterson, British Museum (Natural History), both provided considerable assistance with manipulation of the computer program: their help is gratefully acknowledged. Finally, I would like to thank Dr A. B. Smith, British Museum (Natural History), for his discussions of echinoid hydrodynamics and for his understanding following the inadvertent sacrifice of one specimen of a sand dollar. Appendix Material examined in the course of this study is listed below. The specimens are of very different quality and a mere listing of numbers might be misleading, so they are divided into three categories denoting different parts. Within each category the specimens are arranged stratigraphically, beginning with Ludlow Bone Bed, then overlying Downton Castle Sandstone, then Temeside Shale/Lower Red Downton or presumed equivalent. Cephalic shields including details of orbit, nasohypophysial opening, brain etc. : Ludlow Bone Bed, Ludford Lane, Ludlow, Shropshire—BMNH P.48704, BGS GSM 89284 (Lankester 1870: pl. 13, figs 10, 10a) and counterpart GSM 5151. Ludlow Bone Bed, Forge Bridge, Downton Castle estate, Shropshire—BMNH P.58694. Downton Castle Sandstone, Ludford Lane, Ludlow, Shropshire—BM NH 45949b (Lankester 1870: pl. 13, fig. 14; Stensio 1932: pl. 52, fig. 1), 45949e, P.9756 (Stensio 1932: pl. 53, fig. 5), BGS GSM 5150 (Lankester 1870: pl. 13, fig. 12), GSM 89283. Downton Castle Sandstone, Kington, Hereford & Worcester—BMNH P.9752, P.31857, BGS GSM 89285, BU 1992. Downton Castle Sandstone, Onibury (Norton), Shropshire—BMNH 35999, P.27099, BGS GSM 5149, 21469, 21470. Temeside Shales, Wallop Hall, Shropshire—BMNH P.49015. Portions of cephalic shields only showing marginal fenestrae : Downton Castle Sandstone, Ludford Lane, Ludlow, Shropshire—BMNH 45949f, 45962, P.9757, P.9758. Downton Castle Sandstone, Kington, Hereford & Worcester—BMNH P.5044, P.25403, BGS GSM 57541, GSM 89298. Downton Castle Sandstone, Onibury (Norton), Shropshire—BGS GSM 89296. Downton Castle Sandstone, Shobdon, Hereford & Worcester—BMNH P.25401. Downton Castle Sandstone, Presteigne, Powys—BMNH P.31745. ‘Cornua’: Ludlow Bone Bed, Ludford Lane, Ludlow, Shropshire—BMNH 45970b, c, P.3247 (Stensi6 1932: pl. 53, fig. 1), P.7360, P.25204, P.32255. Ludlow Bone Bed, Clun, Shropshire—BMNH P.39559, P.39562, P.39572-6, P.49016. Ludlow Bone Bed, Rushall, Woolhope, Hereford & Worcester—BMNH P.53119. Downton Castle Bone Bed, Lucton, Hereford & Worcester—BMNH P.8927. Downton Castle Sandstone, Ludford Lane, Ludlow, Shropshire—BMNH 45949 (Lankester 1870: pl. 13, fig. 11), 45949c, d, 45973 (Stensid 1932: pl. 56, fig. 1), P.9897 (Stensid 1932: pl. 53, fig. 3), P.25203 (Stensid 1932: pl. 53, fig. 2). Downton Castle Sandstone, Kington, Hereford & Worcester—BMNH P.25402, BGS GSM 5152. Downton Castle Sandstone, Downton Bridge, Shropshire—BMNH 45970, 45970a. Downton Castle Sandstone, Onibury (Norton), Shropshire—BMNH P.9897 (Stensio 1932: pl. 53, fig. 3). Temeside Shales, Wallop Hall, Shropshire—BMNH P.48954. Temeside Shales, Baggeridge Colliery, south Staffordshire—P.17383-4. 28 P. L. FOREY References Agassiz, J. L. R. 1839. See Murchison. Allen, J. R. L. & Tarlo, L. B. 1963. The Downtonian and Dittonian facies of the Welsh Borderland. Geol. Mag., Hertford, 100: 129-155, 4 figs. Ball, H. W. 1951. The Silurian and Devonian rocks of Turner’s Hill and Gornal, South Staffordshire. Proc. Geol. Ass., London, 62: 225-236, 3 figs. Berg, L. S. 1940. Classification of fishes, both Recent and fossil. (1st edn.) Trudy zool. Inst. Leningr., 5 (2): 87-517 [In Russian; compl. Engl. transl. ]. Denison, R. H. 1951a. Evolution and classification of the Osteostraci. Fieldiana, Geol., Chicago, 11: 156-196, 12 figs. —— 1951b. The exoskeleton of early Osteostraci. Fieldiana, Geol., Chicago, 11: 197-218, 6 figs. — 1979. Acanthodii. In Schultze, H. P. (ed.), Handbook of Paleoichthyology, 5. vi + 62 pp., 35 figs. Stuttgart. Egerton, P. M. G. 1857. Palichthyological Notes, 9. On some fish-remains from the neighbourhood of Ludlow. Q. JI geol. Soc. Lond., 13: 282-291. Forey, P. L. 1984. Yet more reflections on agnathan—gnathostome relationships. J. Vert. Paleont., Norman, Ok., 4: 330-343, 5 figs. Fowler, H. W. 1947. New taxonomic names of fish-like vertebrates. Notul. Nat., Philadelphia, 187: 1-16. Halstead, L. B. 1982. Evolutionary trends and the phylogeny of the Agnatha. In Joysey, K. A. & Friday, A. E. (eds), Problems of Phylogenetic Reconstruction: 159-196, 6 figs. London. — 1985. Discussion [on The Environment of Osteostracans]. Phil. Trans. R. Soc., London, (B) 309: 270. Halstead Tarlo, L. B. 1967. Agnatha. In Harland, W. B. et al. (eds), The Fossil Record: 629-636, 1 fig. London. Harley, J. 1861. On the Ludlow Bone Bed and its crustacean remains. Q. JI geol. Soc. Lond., 17: 542-552, pl. 17. Janvier, P. 1975. Spécialisations précoces et characteéres primitifs du systeme circulatoire des ostéostraces. Colloques int. Cent. natn. Rech. scient., Paris, 218: 15—30, 5 figs, pl. 1. 1977. Contribution a la connaissance de la systématique et de l’anatomie du genre Boreaspis (Agnatha, Cephalaspidomorphi, Osteostraci) du Dévonien inferieur du Spitsberg. Annls Paléont., Paris, 63: 1-32, 14 figs. — [198la.] Les Osteostraci de la Formation de Wood Bay (Devonien inferieur, Spitsberg) et le probleme des relations phylogenetiques entre Agnathes et Gnathostomes. 2 vols mimeogr. These, Sci. Nat., Univ. Pierre et Marie Curie, Paris. — 1981b. Norselaspis glacialis n.g., n.sp. et les relations phylogénétiques entre les Kiaeraspidiens (Osteostraci) du Dévonien inférieur du Spitsberg. Palaeovertebrata, Montpellier, 11: 19-131, 42 figs, pls 1-3. — 198lc. The phylogeny of the Craniata, with particular reference to the significance of fossil “agna- thans’. J. Vert. Paleont., Norman, Ok., 1: 121-159, 17 figs. —— 1984. The relationships of the Osteostraci and Galeaspida. J. Vert. Paleont., Norman, Ok., 4: 344-358, 8 figs. —— 1985a. Les cephalaspides du Spitsberg. 244 pp., 119 figs, 10 pls. Paris. —— 1985b-c. Les thyestidiens (Osteostraci) du Silurien de Saaremaa (Estonie). Premiere partie: Morphol- ogie et Anatomie. Annls Paléont., Paris, 71: 83-147, 35 figs. (1985b). Deuxiéme partie: Analyse phylo- génétique, répartition stratigraphique, remarques sur les genre Auchenaspis, Timanaspis, Tyriaspis, Didymaspis, Sclerodus et Tannuaspis. Loc. cit. 187-216, 8 figs. (1985c). Jeannet, A. 1928. Les poissons fossiles originaux conserves a l'Institut de Geologie de Université de Neuchatel. Bull. Soc. neuchatel. Sci. nat., 52: 102-124. Lankester, E. R. 1870. In Powrie, J. & Lankester, E. R., A monograph of the fishes of the Old Red Sandstone of Britain. Part 1 (concluded}—The Cephalaspidae: 33-62, 24 figs, pls 6-14. Palaeontogr. Soc. (Monogr.), London. M’Coy, F. 1853. On the supposed fish remains figured on Plate 4 of the ‘Silurian System’. Proc. geol. Soc. Lond., 9: 12-15. Marss, T. & Einasto, R. 1978. [Distribution of vertebrates in deposits of various facies in the North Baltic Silurian.] Eesti NSV Tead. Akad. Toim., Tallinn, 27: 16-22 [In Russian, Estonian and Engl. abstracts]. Murchison, R. I. 1839. The Silurian System. xxxii + 768 pp., 37 pls + 1 map. London. 1853. On some of the remains in the bone-bed of the upper Ludlow rock. Proc. geol. Soc. Lond., 9: 16-17. —— 1854-67. Siluria. (1st edn.) xv + 523 pp., 37 pls. (1854). (4th edn.) xvii + 566 pp., 41 pls, 1 map col. (1867). London. DOWNTONIAN OSTRACODERM SCLERODUS 29 Priem, F. 1910. Sur les Poissons et autres fossiles du Silurien supérieur du Portugal. Comungoes Comm. Trab. Serv. geol. Port., Lisbon, 8: 1-10, 2 pls. Ritchie, A. 1967. Ateleaspis tessellata Traquair, a non-cornuate cephalaspid from the Upper Silurian of Scotland. Zool. J. Linn. Soc., London, 47: 69-81, 3 figs, pls 1-4. Smith, A. B. & Ghiold, J. 1982. Roles for holes in sand dollars (Echinoidea): a review of lunule function and evolution. Palaeobiol., Ithaca, N.Y., 8: 242-253, 8 figs. Stensio, E. A. 1927. The Downtonian and Devonian vertebrates of Spitzbergen. Part 1. Family Cepha- ~laspidae. Skr. Svalbard Ishavet, Oslo, 12: 1-391, 103 figs, pls 1-112. — 1932. The Cephalaspids of Great Britain. 220 pp., 70 figs, 66 pls. London. — 1958. Les Cyclostomes fossiles ou Ostracodermes. In Grasse, P. P. (ed.), Traite de Zoologie 13 (1): 173-425, 110 figs. Paris. Telford, M. 1981. A hydrodynamic interpretation of sand dollar morphology. Bull. mar. Sci., Coral Gables, 31: 605-622, 11 figs. — 1983. An experimental analysis of lunule function in the sand dollar Mellita quinquiesperforata. Mar. Biol. Berlin, 76: 125—134, 6 figs, 2 tables. Turner, S. 1973. Siluro-Devonian thelodonts from the Welsh Borderland. J. geol. Soc. Lond., 129: 557-584, 14 figs, pls 1-2. Wangsjo, G. 1952. The Downtonian and Devonian vertebrates of Spitsbergen. IX. Morphologic and systematic studies of the Spitsbergen cephalaspids. Results of Th. Vogt’s expedition 1928 and the English-Norwegian—Swedish expedition 1939. Skr. norsk Polarinst., Oslo, 97: 1-611, 108 figs, 1 table, 118 pls. Westoll, T. S. 1945. A new cephalaspid fish from the Downtonian of Scotland, with notes on the structure and classification of ostracoderms. Trans. R. Soc. Edinb., 61: 341-357, 7 figs, 1 pl. — 1985. Discussion [on The Environment of Osteostracans]. Phil. Trans. R. Soc., London, (B) 309: 271. Woodward, A. S. 1891. Catalogue of fossil fishes in the British Museum (Natural History). xliv + 567 pp., 58 figs, pls 1-16. London. — 1917. Note on Plectrodus, the jaw of an Upper Silurian fish. Geol. Mag., London, (6) 4: 74-75. Zittel, K. 1887-90. Handbuch der Palaeontologie, 3 Vertebrata (Pisces, Amphibia, Reptilia, Aves). 900 pp., 718 figs. Munich. Index The page numbers of the main references are in bold type. An asterisk (*) denotes a figure. Acanthodii 23, 26 dorsal aorta 9*, 10 Aceraspis 18, 20-1, 23-4 dorsal fin 18, 23 anterior cardinal vein 10, 20 Downton Castle Sandstone 2, 4-5, 26-7 Ateleaspidae 1, 18, 23 Ateleaspis 17—20, 22-4 endolymphatic duct 19, 25 Auchenaspis 2, 4, 18, 26 Eukeraspis 2-4; see Auchenaspis egertoni 18, 20, 24 exoskeleton 5, 10, 12, 20-1, 25 (Eukeraspis) pustulifera 5 salteri 18, 20, 24-5 facial nerve 8*, 9*, 10, 19, 25 fenestrations 1, 5, 6*, 13, 14*, 15 Belonaspis puella 5, 12 enestrations benneviaspidians 13, 18-24 Benneviaspis 13 holtedahli 19 glossopharyngeal 8*, 9* Boreaspis 13 habenular recess 5 head vein, see jugular vein cephalaspidians 13, 18-24 Hemicyclaspis 18, 20, 22-4 cephalothoracic shield 1, 2, 5, 6* murchisoni 19 circumnasal fossa 5-6, 6*, 10, 11*, 20, 25 Hemiteleaspis 18, 20-1, 23-4 clypeasteroid echinoids (sand dollars) 15 Hildenaspis 19 cornua 1, 3, 3*, 5, 13, 17-19, 23 Hirella 18, 20-1, 23-4 hydrodynamic function 1, 15, 17 Dartmuthia 5, 13, 25-6 Didymaspis 13, 17-21, 24, 26 internal carotid artery 8*, 9*, 10 30 P. L. FOREY jugular vein 6, 8*, 9*, 10, 26 postbranchial wall 10 prebranchial fossa 8* kiaeraspidians 5, 13, 18-24, 26 Procephalaspis 17-18, 20, 24-5 Kiaeraspis 12 profundus 10, 11* Psammodus 2 lateral line 1, 12, 20, 25-6 Pterygotus pustuliferus 2 Lower Red Downton 5, 26-7 Ludlow Bone Bed 1-2, 4-5, 26—7 Saaremaxpiei Gel aton sc sand dollars 15 Sclerodidae (Sclerodontidae) 17 scolenaspidians 13, 18, 20, 23-4 s.e.]. canals 9*, 13, 17, 19, 23 semicircular canal, anterior 6, 8* micromery 19, 22 Mimetaspis 19 myodome 10, 20, 25 nasohypophysial duct 9*, 10 : Gas posterior 6 opening 5, 6*, 11*, 12, 15, 20, 25 Saigon fade 312 region dorsal 5, 6*, 10, 12, 19-20, 25-6 ann - lateral 1, 3*, 6*, 10, 12, 19-20, 25-6 occipital artery 9*, 10 ; gee ; superficial vein 8 Oeselaspis 10, 13, 17, 25-6 es supraoral field 20, 23, 25—6 oesophagus 9*, 10 hies 24 olfactory sac 6 synapomorphies 24, 25-6 orbit 5-6, 6*, 8*, 9*, 10, 23, 25 Temeside Shales 5, 26-7 PAUP computer program 18 tesserae 19, 27 pectoral fin 3*, 13, 17—20, 23, 25 Thyestes 2-3, 13, 17-18, 20-6 pectoral sinus 1, 19 verrucosus 12, 18 Petromyzontiformes 23 thyestidians 5, 12-13, 25-6 pineal area 5, 25-6 Timanaspis 19, 25 duct 11* tree, computer generated 1, 25 opening 5, 6*, 11* consensus 20, 23, 25 plate 5, 25-6 Tremataspididae 1, 4, 17-24, 26 pits 1, 10, 26 Tremataspis 5, 10. 13, 17-18, 25-6 anterior 6*, 11*, 12 trigeminal nerve 8*, 10 middle 6*, 11*, 12 trochlearis 10, 11* posterior 6*, 11* Placodermi 19, 23 Plectrodus 3-4 mirabilis 2-3 pleiopristis 2-3 Witaaspis 17-18, 20, 23-6 vestibular chamber 6, 8*, 12, 26 — en Accepted for publication 19 December 1985 \) ng ; - 75 ath ee. Bulletin British Museum (Natural History) . Geology Series BACK NUMBERS lume 1 3 The pterobranch Rhabdopleura in the English Eocene. H. D. Thomas & A. G. Davies. 1949. Pp. 1-24, 3 plates, 4 figs. £2.50 A reconsideration of the Galley Hill skeleton. K. P. Oakley & M. F. Ashley Montagu. 1949. Pp. 25-48, 1 _ plate, 4 figs. with _ The vertebrate faunas of the Lower Old Red Sandstone of the Welsh Borders—Pteraspis leathensis White: A Dittonian Zone-Fossil. E. I. White. 1950. Pp. 49-90, 1 plate, 27 figs. £8.00 A new Tithonian Ammonoid fauna from Kurdistan, Northern Iraq. L. F. Spath. 1950. Pp. 93-146, 5 plates. £5.00 Some Jurassic and Cretaceous crabs (Prosoponidae). T. H. Withers. 1951. Pp. 171-192, 3 plates, 14 figs. £3.20 A new Trochiliscus (Charophyta) from the Downtonian of Podolia, eastern Europe. W. N. Croft. 1952. Pp. 187-220, 2 plates, 7 figs. £3.00 Cretaceous and Tertiary Foraminifera from the Middle East. T. F. Grimsdale, Pp. 221-248, 6 plates, 3 figs. : £3.00 . 10 Cyclopygid trilobites from Girvan, and a note on Bohemilla. W. F. Whittard. 1952. Pp. 305-324, 2 plates. ; £1.80 ume 2 A coniferous petrified forest in Patagonia. M. G. Calder. 1953. Pp. 97-138, 7 plates, 7 figs. £4.50 Some Upper Cretaceous and Eocene fruits from Egypt. M. E. J. Chandler (with appendices by M. Y. Hassan & M. I. Youssef). Pp. 147-187, 7 plates, 1 fig. £4.00 The Carboniferous flora of Peru. W. J. Jongmans. 1954. Pp. 189-224, 11 plates. £3.80 Further contributions to the solution of the Piltdown problem. J. S. Weiner, W. E. Le Gros Clark, K. P. Oakley & others. 1955. Pp. 225-287, 5 plates. £5.60 7 The Schizaeaceae of the south of England in early Tertiary times. M. E. J. Chandler. 1955. Pp. 291-314, 7 plates, 2 figs. £2.90 e3- _ The structure, evolution and nomenclature of the ostracod hinge. P. C. Sylvester-Bradley. 1956. Pp. 1-21, 4 plates, 2 figs. £2.20 Eocene mollusca from Nigeria: a revision. F. E. Eames. 1957. Pp. 23-70, 6 plates. £4.60 The structure of some leaves and fructifications of the Glossopteris flora of Tanganyika. D. D. Pant. 1958. Pp. 125-175, 4 plates, 21 figs. £6.30 Lidgettonia, a new type of fertile Glossopteris. H. H. Thomas. 1958. Pp. 177-189, 2 plates, 2 figs. with The faunal succession in the Caradoc series of south Shropshire. W. T. Dean. 1958. Pp. 191-231, 36 plates, 4 figs. £5.40 A new labyrinthodont (Paracyclotosaurus) from the Upper Trias of New South Wales. D. M. S. Watson. 1958. Pp. 233-263, 5 plates, 16 figs. with An early Pleistocene mammalian fauna from Bethlehem. D. A. Hooijer. 1958. Pp. 265-292, 4 plates. £7.00 The Upper Permian flora of England. H. M. M. Stoneley. 1958. Pp. 293-337, 5 plates, 16 figs. with - Blue-Green algae from the Middle Devonian of Rhynie, Aberdeenshire. W. N. Croft & E. A. George. 1959. Pp. 339-353, 4 plates. £7.00 Titles to be published in Volume 41 The Downtonian ostracoderm Sclerodus Agassiz (Osteostraci: Tremataspididae). By P. Li Perey: Lower Turonian (Cretaceous) ammonites from south-east Nigeria. By P. M. P. Zaborski. The Arenig Series in South Wales: Stratigraphy and Palaeontology. I. The Arenig Series in South Wales. By R. A. Fortey and R. M. Owens II. Appendix. Acritarchs and Chitonozoa from the Arenig Series of South-west Wales. By S. G. Molyneux. Miocene geology and palaeontology of Ad Dabtiyah, Saudi Arabia. Compiled by P. J. Whybrow. Typeset by Santype International Ltd., Salisbury, Wilts. Printed in Great Britain at the University Printing House, Oxford cite : } ritish Museum (Natural History) Me Pay # at ee r Turonian (Cretaceous) ammonites south-east Nigeria i Lo 2a gy Vol41 No2 30 April 1987 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum’s resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 S5BD, England. World List abbreviation: Bull. Br. Mus. nat. Hist. (Geol.) © Trustees of the British Museum (Natural History), 1987 The Geology Series is edited in the Museum’s Department of Palaeontology Keeper of Palaeontology: Dr L.R.M. Cocks Editor of the Bulletin: Dr M. K. Howarth Assistant Editor: Mr D.L. F. Sealy ISBN 0 565 08015 6 ISSN 0007-1471 Geology series British Museum (Natural History) Vol 41 No 2 pp 31-66 Cromwell Road London SW7 5BD Issued 30 April 1987 , i Lower Turonian (Cretaceous) ammonites from south- east Nigeria P. M. P. Zaborski Department Br Oeolony, University of Maiduguri, P.M.B. 1069, Maiduguri, Nigeria | - Contents PALAEON SWiNOOSS.. cosuaadooabasogaspnegsene coats Hosen epeOtem eC eee CER ET Concer onenran Satara ner eee 31 ISTIFOCICITION . cons caenatos cabonsopdd coocormordo dct nd aabreMercane teemoc cua ane Cen oneme 3) SHAISMAUICCESTTIOUON ao poo0'e odoadema cous soot. Socoe cod saaen a SDane ees ssomemcs te sete 33 ata vaDesmocenaticaeyZittel paereccsarecrctierstets.-isyereaiols eictel sie ese 5 clovesoieiant ohne retonn a egasversice 33 Subfamuilygseuzosinae S pathy asasaaseeeees eer eccentric seers 33 GenussrachydesmocenasiSpathwecrnesceneeceerccecreataeecceeeter mecca scene. 33 Pachydesmoceras denisonianum (Stoliczka) ..............000.c00eccee eee eee 34 Family Acanthoceratidae GrossOuvre .............0000ccceceec eee eee eect eens eeeeeues 36 Subfamily Euomphaloceratinae Cooper ..................c00eeccecceee eee eee sees 36 GenusiKamentmocerassReyment seer... -sese ees sees ecieeccnisssiceeeen ssn se 36 Mamenunocerasicliescnitl(SOlger)ee-eeeeeeeteo-coe eee eee encore eee 36 Kamerunoceras puebloense (Cobban & Scott) ............... cece cece eee ees 36 StbfamilyMammitimae Huyattier.... 22 sgpearannscn wcivaxecscser «cabs aneceone s+: 37 GenusiWatinocerasiWiaTtenis sss. -. saeeete eens sees eee ee eerie =: 37 Watinoceras aff. amudariense (Arkhanguelsky) .....................-0000000- 37 WORT OOCETCDT OF nae baaecG uaecee ecco Somers eae cee ter ceiete ta oto eee are 38 Genus Mammites Laube & Bruder ...................ccec cece ce eee eee e ee nne cece 40 Miammites nodosoides|(SmMluten))-cenecesk ee o--2--ceeineseccesce cence nse: 40 VEGTAINILES WIS Das. Re Os et ten eee asses Sanat ooeeisame aeieee oye ees 41 amilyaVascoceratidac Do wmille, - ee. 1, Aces cta se eiene acne csiceeanenneneseennen ss 41 Subfamily Vascoceratinae Douvillé ............ 0.0.00 cece cece ec e eee t eee en eens 41 GenuspiagestasPenvingUlene nea serra ceeeaes snes ace es. racecar 41 ETL CSTAMED ES TR EMTs ee fore Py HA ais ice ose Sin din asaln's able atreints v's oedema areas ac 41 Genus NeopaychitessKOssmat senaaee cases sa: ees eee nclaseeec «oe seems e 43 Neoptychites cephalotus (Courtiller) .................0eeeeeeeeceee cece eeeeeees 43 Subfamily Pseudotissotiinae Hyatt ............... 0. cece cece eee e teen eee e eee e eee e ees 45 Genusmiomasires Penvinguilerel erase ecueee a saeeeeeeeaae eee tere 45 dihomasitesigongilensis| (WOOGS))neaesaesese cee tarceen eter tere cee: 47 MRTOMASTLESIKOULADICUS) (IKG1ETs)) erepstrsrsteter=sovavar-roreieieiatovalore olotelelete oie \ctetotereiosToevarel ete ver=ere 48 GenuissVanightocends Reyiments eres acl see ierieienel = sleteieie sie ale elena eta ec eeeleinicielielees 49 WinightocerasiwalistsReyment erranae acer serene coerce ace eee ace arer 51 Wrightoceras cf. munieri (Pervinquiere) ...........-....+eecene eee e cette eee ees 51 antlya€ olopocenatidaeiey att cn. sen serieees ee oct sgece tence eenecteewien cnees ae 52 GenusyHloplitoides VonIKOenen te accre cece oeeee eee ee eeeeeceeeeccee casceeee eae sy) JHODITOMES [AGP SO. WON co0ccoce ooodndenonauaavcepuassveccceunaeneone 53 Genus Herrickiceras Cobban & Hook ..................:ceeeecee eect eee eteeees 56 UCIT CHICELASUAS Ds Pe es SPAM. Go sie ols tin nical tin apse heere Fda nele al ee? 56 Stratigraphical conclusions ......:.............+. cian ae SUR Soon can an Panerai 56 AG RMOWGCISTTASTUS «. ocinab doo bagdanen de HOROa ROE pone era ae aera na Serr nap onan nnecn ice 60 IRGISICMGES oo ncoocandsascdsodad sede DOSER ADEE CEO OSe ACROICr? JURE Aeon AC ISE a ia steer ei Rar roe 60 ITGIENS sco dodoaddoud Gop Abb AHe6 Sb GORD NS CORE GCE eee OT aes eee ne aera arena 64 Synopsis Within most of Nigeria’s Benue Trough uppermost Cenomanian and Lower Turonian strata are domi- nated by vascoceratid ammonite faunas of Tethyan affinities. Such assemblages range as far south as the Ezillo region in south-east Nigeria where Nigericeras, Paravascoceras, Fagesia, Thomasites and Wrighto- Bull. Br. Mus. nat. Hist. (Geol.) 41 (2): 31-66 Issued 30 April, 1987 32 P. M. P. ZABORSKI ceras occur. Barely 60km south, however, at Lokpanta, the Lower Turonian contains faunas more easily correlated with those of the western interior of the United States and north-west Europe. The basal Turonian here is mainly characterized by Watinoceras spp., while the upper part of the Lower Turonian contains Pachydesmoceras, Mammites nodosoides (Schltiter), Kamerunoceras, Fagesia, Neoptychites, Her- rickiceras? and Hoplitoides latefundatus sp. nov. This last form appears to provide an evolutionary link between Wrightoceras and typical Hoplitoides. The absence of the vascoceratid-rich faunas at Lokpanta is probably because of palaeoenvironmental factors. Introduction Uppermost Cenomanian and Turonian sediments are among the most widely distributed and richly fossiliferous parts of the Cretaceous system in Nigeria. During the early phases of systematic palaeontological work in Nigeria diverse collections of ammonites were attributed to the Lower Turonian (Reyment 1954, 1954a, 1955, 1957, Barber 1957). In recent years, NIGERIA 200 km, Ashaka» Pindiga. Enugu calc. siltst. with *Abakaliki Okigwe- a shales/Ists with ) 1 . Allocrioceras? IN ‘Calabar Paravascosceras? {Als Themen as mudsts/ssts with Wrightoceras Ezillow9 _ gastropods, bivalves as mudsts with 62 Amika Esbices 26% Mortoniceras cS Ndieze eal: unfossiliferous mudsts/ssts W. Aboine River ee Abakaliki = can Paravascoceras?, E ae Fagesia, Thomasites, SECTION Lokpanta EZE- meh Wrightoceras = L. TUR. c nh == <== to 9° aaa =) ld Enugu - CEN.? Allocrioceras?, \w) =, a Enuane Pt. Harcourt Ezilloella R | i Okigwe expressway | road | T E i ‘______-« Mammites Sear re c R —E > — <= @ = BS sS= & E N a N i a (ae — Watihoceras A ° — = auna M N = a a 31 U 2 ; \ A rh ee Metengonoceras? 9 oe Ber 5 — N A ' U : 33 N? | P Xs LP G \ Pp Yo to R | E non-ammonitiferous ALBIAN fo — R black shales ae oe A U SSS Elobiceras, (S Pp a eee Mortoniceras E 500m (LOkigwe E GENERALIZED SECTION ABAKALIKI TO EZILLO Fig. 1 Maps showing fossil localities and sections described in the text. NIGERIAN LOWER TURONIAN AMMONITES 33 however, much of this material has been reassigned to the Upper Cenomanian and Middle Turonian. As a consequence, the faunal character of the Lower Turonian and its correlation within and outside Nigeria are in need of clarification. Material from the Lower Turonian of south-eastern Nigeria described here makes a considerable contribution to this. The present faunas come from two areas (see Fig. 1). On the western outskirts of Ezillo (Eze-ilo), adjacent to the Enugu—Abakaliki highway, a disused pit exposes gently-dipping, micaceous, calcite-cemented siltstones yielding Thomasites gongilensis (Woods), T. koulabicus (Kler), Wrightoceras wallsi Reyment, W. cf. munieri (Pervinquiére), Fagesia sp. and Paravasco- ceras? sp. Immediately south of Lokpanta three cuttings are present in a distance of less than 1km on the Enugu—Port Harcourt expressway. In the most northerly of these cuttings 10-12 m of black, weathering sandy yellow, shales with large calcareous nodules contain an ammonite fauna of Pachydesmoceras denisonianum (Stoliczka), Kamerunoceras puebloense (Cobban & Scott), Mammites nodosoides (Schliter), Neoptychites cephalotus (Courtiller), Fagesia levis Renz, Hoplitoides latefundatus sp. nov. and Herrickiceras? sp. Although ammonites occur here in their hundreds, larger specimens are, almost without exception, so badly crushed as to be unidenti- fiable. It is therefore usually the juvenile and middle whorls alone that are suitable for descrip- tion. Directly south, a second cutting displays about 25m of the closely similar underlying shales. In the lower 5m of this sequence there are bands of dark grey, weathering sandy yellow, calcareous nodules whose surfaces are studded with the impressions of Watinoceras aff. amu- dariense (Arkhanguelsky), W. sp., Kamerunoceras cf. eschii (Solger), Mammites? sp. and Neopty- chites cephalotus. Only tiny ammonites are preserved complete, the larger forms being recognized by fragments. The most southerly cutting exposes over 40 m of shales, silty in places, which include thin calcareous horizons formed from coalesced nodules. These beds are less fossiliferous but from their upper part have yielded several poorly preserved impressions of Metengonoceras? (Fig. 9, p. 39). The regional dip in these cuttings averages a little over 30° north-west, but rises abruptly to 70° west at the northern end of the last-described exposure. This fact suggests the proximity of faulting, and indeed minor faults can be observed within these cuttings. The shales, silts and calcareous beds of late Cenomanian? to early Turonian age in most parts of southern Nigeria are conventionally referred to as the Eze-Aku Formation. This lithostratigraphic unit, formalized by Simpson (1954), is, however, not readily distinguishable from superjacent beds and is recognized primarily on the basis of its age. Though this is contrary to accepted stratigraphical practice, it is outside the scope of the present work to adopt other than a traditional approach here. A fuller discussion can be found in Petters & Ekweozor (1982). Systematic descriptions Repositories. Register numbers prefixed by the letter C are of specimens in the Department of Palaeontology, British Museum (Natural History), London. Those prefixed by the abbreviation UIN are of specimens in the Department of Geology, University of Ilorin, Nigeria. Dimensions (in mm). D, diameter; Wb, whorl breadth; Wh, whorl height; U, umbilical diam- eter. Figures in parentheses are dimensions as a percentage of the total diameter. N = number of ribs in last whorl. Superfamily DESMOCERATACEAE Zittel, 1895 Family DESMOCERATIDAE Zittel, 1895 Subfamily PUZOSIINAE Spath, 1922 Genus PACH YDESMOCERAS Spath, 1922 TYPE SPECIES. Ammonites denisonianus Stoliczka, 1895; by original designation. 34 P. M. P. ZABORSKI Pachydesmoceras denisonianum (Stoliczka, 1865) Figs 2-4 1865 Ammonites denisonianus Stoliczka: 133 (pars); pl. 66, fig. 2 (only); pl. 66a (non pl. 65, fig. 4; pl. 66, fig. 1). 1898 Puzosia Denisoniana (Stoliczka) Kossmat: 121; pl. 14, figs 5a, 5b; pl. 15, figs 5a, 5b. 1898 Desmoceras Kamerunense von Koenen: 55; pl. 7, figs 1—3. 1899 Puzosia alimanestianui Popovici-Hatzeg: 14; pl. 1. 1904 Puzosia Denisoniana (Stoliczka); Solger: 103; pl. 3, figs 1a, 1b; text-fig. 5S. 21904 Puzosia Denisoni (Stoliczka); Douvillé: 237; pl. 29, figs 1-4; pl. 30, figs 1a, 1b. 1907 Desmoceras (Puzosia) Denisonianum (Stoliczka) Boule, Lemoine & Thévenin: 21; pl. 5, figs 3-5. 21912 Puzosia denisoniana (Stoliczka); Zimmermann: 542; pl. 26. 1914 Puzosia denisoniana (Stoliczka); Yabe: 72; pl. 7 1922 Pachydesmoceras denisonianum (Stoliczka) Spath: 127 1954 Pachydesmoceras denisonianum (Stoliczka); Matsumoto: 100 (with synonymy). 1955 Pachydesmoceras kamerunense (von Koenen); Reyment: 19. 1958 Pachydesmoceras denisonianum (Stoliczka); Reyment: 54. 1961 Pachydesmoceras denisoni (Stoliczka); Collignon: 39; pl. 8, figs la, 1b. 1965a Pachydesmoceras denisoni (Stoliczka); Collignon: 22; pl. 422, fig. 1752. MATERIAL AND OCCURRENCE. Fourteen specimens (C.83511, C.85290—2, C.90292-301) from the Eze-Aku Formation (Lower Turonian, Mammites nodosoides Zone), Lokpanta, south-east Nigeria. DIMENSIONS. D Wb Wh U C.90298 36 12-5 (35) 17 (47) 9 (25) C.90297 50 15 (30) 22 (44) 13 (26) C.90294 60 PBaeS) 28 (47) 17 (28) C.90292 65 25 (38) 29 (47) 18 (28) C.90293 71 27 (38) 35 (49) 19 (27) DESCRIPTION. The shell is moderately evolute, moderately compressed and has a broadly rounded venter. The earliest whorls are smooth but at diameters of 15—20 mm faint, narrow ribs appear in the ventral area. These ribs gradually extend to the umbilical shoulder, there being about eight in each whorl. At diameters of 40-50mm minor ribs appear, confined to the outer part of the flank and the venter. They gradually strengthen and in the middle stages may outnumber the major ribs by as many as ten to one. In the later stages the major ribs dominate and at diameters in excess of 150 mm there are only one or two minor ribs intervening. REMARKS. This material is exactly comparable with that described from Cameroun by Solger (1904: 103; pl. 3, figs 1a, 1b; text-fig. 5) as Pachydesmoceras denisonianum (Stoliczka). Reyment (1955: 17) at first included Solger’s specimens in P. kamerunense (von Koenen), differentiating it from P. denisonianum on the basis of minor ribbing details and degree of inflation. Later, however, Reyment (1958: 54) referred them back to the latter species. There do not seem to be any significant differences between the Nigerian/Camerounian material and P. denisonianum and they are treated here as conspecific. Puzosia alimanestianui Popovici-Hatzeg (1899: 14; pl. 1) is another probable synonym. The inner whorls in the holotype of P. hourcqi Collignon (1961: 42; pl. 11, fig. 1) are poorly preserved but its later ornament resembles that in the present material and it may also be conspecific. P. denisonianum and indistinguishable forms have a long stratigraphical range, at least from Cenomanian to Coniacian. P. pachydiscoides Matsumoto (1954: 101; pl. 9, figs 2a, 2b) is said to Figs 2-4 Pachydesmoceras denisonianum (Stoliczka). Eze-Aku Formation (Lower Turonian, Mam- mites nodosoides Zone), Lokpanta, south-east Nigeria. Fig. 2, C.85290, x 1. Fig. 3, C.90301, x 1. Fig. 4a, b, C.90297, x 1. Figs 5-6 Kamerunoceras cf. eschii (Solger). Eze-Aku Formation (basal Turonian), Lokpanta, south- east Nigeria. Fig. 5a, b, C.90370, x 1-5. Fig. 6, C.90371, x 1. Both latex casts. NIGERIAN LOWER TURONIAN AMMONITES 35 36 P. M. P. ZABORSKI have higher whorls and more projected ribs than P. denisonanum. P. maroccanum Collignon (1966: 26; pl. 12, fig. 4) has broader whorls and develops strong ribbing earlier. In P.(?) linderi (Grossouvre 1894: 188; pl. 24, fig. 4; Collignon 1961: 41; pl. 10, figs 1, 1a; 1965: 8; pl. 379, fig. 1640) all ribs are of equal strength in the earlier whorls. P. radaodyi Collignon (1964: 58; pl. 333, fig. 1498) and P. rarecostatum Collignon (1961: 40; pl. 9, figs 1, 1a) both have stronger major ribbing which dominates the ornament even in the middle growth stages. Superfamily ACANTHOCERATACEAE Grossouvre, 1894 Family ACANTHOCERATIDAE Grossouvre, 1894 Subfamily EUOMPHALOCERATINAE Cooper, 1978 Genus KAMERUNOCERAS Reyment, 1954a TYPE SPECIES. Acanthoceras eschii Solger, 1904; by original destination. Kamerunoceras cf. eschii (Solger, 1904) Figs 5-6 cf. 1904 Acanthoceras eschii Solger: 124; pl. 4, figs 1-4. non 1954a Kamerunoceras eschii (Solger) Reyment: 251; pl. 3, fig. 5; pl. 5, figs 3, 6; text-figs 2a, 2b (= Kamerunoceras seitzi (Riedel)). non 1955 Kamerunoceras eschii (Solger); Reyment: 59 (= Kamerunoceras seitzi (Riedel)). cf. 1958 Kamerunoceras eschii (Solger); Reyment: 55; pl. 1, figs 1a, 1b; pl. 2, figs 1a, 1b. cf. 1979 Kamerunoceras eschii (Solger); Kennedy & Wright: 1166, 1175-1176; pl. 1, figs 4-9. MATERIAL AND OCCURRENCE. Two specimens (C.90370—-1) from the Eze-Aku Formation (basal Turonian), Lokpanta, south-east Nigeria. DESCRIPTION. The smaller of these two specimens (C.90370, Fig. Sa, b) has a diameter of about 25 mm. It is evolute, its whorl breadth and whorl height being approximately equal. There are 16-17 ribs per whorl which are prominent only on the flanks where each bears bullate umbil- ical and inner ventrolateral tubercles. The ribs are effaced upon the venter where the ornament consists of clavate outer ventrolateral and siphonal tubercles, the latter situated a little adoral of the former. There are no minor ribs. The larger specimen (C.90371, Fig. 6) has a whorl breadth of some 20mm. Again ribbing dominates the flank ornament while tubercles are more prominent upon the venter. There are strong, moderately spinose inner ventrolateral and weaker outer ventrolateral tubercles fusing to form a bituberculate structure. Along the median line are strong, rounded siphonal tubercles. Rib spacing is irregular. Minor ribs are absent. REMARKS. Of the multitude of species referred to Kamerunoceras (see list in Kennedy & Wright, 1979), the present material is closest to the type species K. eschii (Solger 1904: 124; pl. 4, figs 1-4; Reyment 1958: 55; pl. 1, fig. 1; pl. 2, fig. 1; Kennedy & Wright 1979: 1175-1176; pl. 1, figs 4-9). Unfortunately, this is a poorly understood form, known only from the imperfectly pre- served type specimen from southern Cameroun. It agrees, however, in its rather broad whorls, prominent flank ribs and comparatively massive ventrolateral tuberculation. Kamerunoceras puebloense (Cobban & Scott, 1972) Figs 7-8 1972 Kanabiceras puebloense Cobban & Scott: 73; pl. 15, figs 8, 9; pl. 37, figs 1-8; pl. 38, fig. 1. 1979 Kamerunoceras puebloense (Cobban & Scott) Kennedy & Wright: 1170. MATERIAL AND OCCURRENCE. Four specimens (C.83518, C.90342-4) from the Eze-Aku Forma- tion (Lower Turonian, Mammites nodosoides Zone), Lokpanta, south-east Nigeria. Two further specimens (C.85299—300) from the same horizon and locality may also belong here. DESCRIPTION. The shell is evolute with pentagonal whorl sections, which are a little higher than broad in the early stages, but slightly broader than high later on. Between diameters of 15 and NIGERIAN LOWER TURONIAN AMMONITES 37) 25mm there are numerous ribs of markedly uneven development. These are projected forwards on the flank, backwards over the ventrolateral shoulder and form distinct chevrons over the venter. Major ribs bear prominent and spinose umbilical, inner and outer ventrolateral tuber- cles and more subdued but still well-defined siphonal tubercles. There are numerous inter- calated ribs lacking umbilical tubercles but sometimes bearing subdued ventrolateral and siphonal tubercles; others show virtually no tubercle development at all. The venter is crossed by deep, chevron-shaped constrictions. At diameters of 50-70mm the ribbing becomes more rectiradiate and regular, constrictions disappear and the major ribs come to dominate the ornament. Each bears highly spinose umbilical, inner and outer ventrolateral tubercles and a slightly lower siphonal tubercle. Intercalated ribs, dying out in the mid-flank region, generally alternate with the major ribs. They usually show exactly the same style and strength of ventral tuberculation as do the latter though the inner ventrolateral tubercles may be effaced. At these diameters there occur only a very few additional weak, fold-like ribs lacking tubercles altogether. REMARKS. Cobban & Scott (1972: 73) first described Kamerunoceras puebloense from the Lower Turonian (Mammites nodosoides Zone) of Colorado. They referred it to Kanabiceras Reeside, a genus shown by Kennedy et al. (1981: 55) and Wright & Kennedy (1981: 54-55) to be a synonym of Euomphaloceras Spath. In fact, as pointed out by Wright & Kennedy (1981: 56), this species combines juvenile ornamentation similar to that in late members of Euomphaloceras with the evolute coiling and adult ornament characteristic of Kamerunoceras. Thus the early constriction-bearing whorls with their much multiplied, chevron-forming ventral ribs and tubercles resemble those in the late Cenomanian Euomphaloceras euomphalum (Sharpe) (see Kennedy 1971: 91; pl. 43, fig. 1; pl. 59, figs 1-5; Wright & Kennedy 1981: pl. 11, figs 1-8) and E. septemseriatum (Cragin) (see, for example, Cobban & Scott 1972: pl. 12, figs 5-27; Wright & Kennedy 1981: pl. 12, figs 1-8; pl. 13, figs 1-6; pl. 14, figs S—9). Conversely, the evolute coiling and adult whorls with their lesser secondary ornament and more rectiradiate ribbing show closer similarities to Kamerunoceras (see Kennedy & Wright 1979 for a review of this genus). Since this material is clearly trending towards Kamerunoceras it is included therein, following the suggestions of Cooper (1978: 110), Kennedy & Wright (1979: 1170) and Wright & Kennedy (1981: 56). Although K. puebloense forms a clear link between its genus and the presumably ancestral Euomphaloceras, the more typical Kamerunoceras, K. cf. eschii, occurs below it at Lokpanta. As in Nigeria, K. puebloense is found in the upper part of the Lower Turonian in Colorado, though a similar form has been described from the very late Cenomanian of southern England (Wright & Kennedy 1981: 56; pl. 14, figs 3, 11). The later whorls of K. schindewolfi (Collignon 1965: 31; pl. 389, fig. 1665) resemble those in the present material in their spinose tubercles and persistent minor ribbing. There are, however, mid-lateral tubercles in K. schindewolfi. Subfamily MAMMITINAE Hyatt, 1900 Genus WA TINOCERAS Warren, 1930 TYPE SPECIES. Watinoceras reesidei Warren, 1930 (= Acanthoceras amudariense Arkhanguelsky, 1916); by monotypy. Watinoceras aff. amudariense (Arkhanguelsky, 1916) Figs 10-12 MATERIAL AND OCCURRENCE. Three specimens (C.90366-—8) from the Eze-Aku Formation (basal Turonian), Lokpanta, south-east Nigeria. DIMENSIONS. D Wb Wh U N C.90367 13 4-5 (35) 5-6 (43) 3-8 (29) 37) C.90366 9 = 3-8 (42) 3-1 (34) 42 DESCRIPTION. The material at hand is of small size, the largest specimen having a diameter of only 13mm. The shell is moderately evolute, compressed, with flattened flanks and, at first, a 38 P. M. P. ZABORSKI rounded venter, the whorl section becoming more pentagonal in shape later. The whorls are smooth up to diameters of 3-4mm, when ribs appear, at first confined to the outer part of the flank and the venter. The whorls thereafter bear dense, rounded ribs mostly arising in pairs at the umbilical shoulder but sometimes bifurcating some distance down the flank. Intercalated ribs also occur. The ribs are projected over the venter but are interrupted by a narrow ventral sulcus. There are weak ventrolateral tubercles but umbilical tubercles become noticeable only at the largest diameters seen. One specimen (C.90368, Fig. 12), 11mm in diameter, shows somewhat coarser ribs than the other two on its outer whorl. They form prominent chevrons upon the venter where they are again interrupted by a narrow sulcus. REMARKS. Owing to their extremely small size, it is not possible to assign these specimens definite specific status. They most closely resemble the inner whorls of Watinoceras amudariense (Arkhanguelsky 1916: 48; pl. 7, figs 8-13; Wright & Kennedy 1981: 51; pl. 10, figs 6, 14; text-figs 19N, 19Q (with synonymy)), of which W. reesidei Warren (1930: 67; pl. 3, fig. 2; pl. 4, figs 9-12; see also, for example, Cobban & Gryc 1961: 186; pl. 38, figs 46-49) is the main synonym. The Nigerian material has a similar rib density and style of ornament, and such range as it shows in these features falls within that exhibited by figured specimens of this species. Thus the most coarsely ribbed Nigerian variant (C.90368) is similar to the example figured by Arkhanguelsky (1916: pl. 7, fig. 9). The main difference shown by the present material is its rather smaller umbilical diameter, 29-34% of the overall diameter against a consistent figure of about 40% in W. amudariense. In this respect the former is closer to specimens from Tarfaya figured by Collignon (1966: pl. 19, figs 14, 15) as W. sp. aff. reesidei Warren. Watinoceras sp. Figs 13-17 MATERIAL AND OCCURRENCE. Seven specimens (C.90361—Sa, b, C.90369) from the Eze-Aku Formation (basal Turonian), Lokpanta, south-east Nigeria. DIMENSIONS. D Wb Wh U N C.90365a 9 33} (7) 333} (637) 2:8 (31) 33 C.90365b 9:5 — 4-3 (45) 2:6 (27) 25 C.90369 12:5 6:5 (52) 4:5 (36) 3-6 (29) 27 C.90363 14 — 6 (43) 4 (29) 26 DESCRIPTION. This material is again of small size, the largest specimen having a diameter of 14 mm. The shell is moderately evolute. Its whorl height and whorl breadth are approximately equal at a diameter of about 9 mm but the latter increases much more rapidly so that the whorls develop a markedly depressed, pentagonal shape later on. The shell is smooth until Figs 7-8 Kamerunoceras puebloense (Cobban & Scott). Eze-Aku Formation (Lower Turonian, Mammites nodosoides Zone), Lokpanta, south-east Nigeria. Fig. 7a, b, C.90342, x1. Fig. 8a-—c, C.90344, x 1. Fig. 9 Metengonoceras? sp. Eze-Aku Formation (Upper Cenomanian?), Lokpanta, south-east Nigeria. C.90373, x 1. This specimen shows pseudoceratitic sutures. Figs 10-12 Watinoceras aff. amudariense (Arkhanguelsky). Eze-Aku Formation (basal Turonian), Lokpanta, south-east Nigeria. Fig. 10a, b, C.90367, x 1-5. Fig. 11, C.90366, x 1:5. Fig. 12, C.90368, x 1-5. All latex casts. Figs 13-17 Watinoceras sp. Eze-Aku Formation (basal Turonian), Lokpanta, south-east Nigeria. Fig. 13a, b, C.90369, x 1-5. Fig. 14a, b, C.90362, x 1-5. Fig. 15a, b, C.90365a, x2. Fig. 16, C.90365b, x 1-5. Fig. 17a, b, C.90363, x 1-5. All latex casts. Figs 18-20 Mammites nodosoides (Schliter). Eze-Aku Formation (Lower Turonian, Mammites nodo- soides Zone), Lokpanta, south-east Nigeria. Fig. 18a, b, C.90336, x 1. Fig. 19a, b, C.90337, x 1. Fig. 20a, b, C.90341, x 1-5. See also Fig. 22a, b. Fig. 21a, b Mammites? sp. Eze-Aku Formation (basal Turonian), Lokpanta, south-east Nigeria. C.90360, x 1-5. Latex cast. NIGERIAN LOWER TURONIAN AMMONITES 39 40 P. M. P. ZABORSKI diameters as large as 4 mm, when fairly dense, rounded ribs appear. Prominent tubercles arise at a very early ontogenetic stage, the outer ventrolateral tubercles developing first, followed by the inner ventrolateral and umbilical tubercles. By a diameter of 10 mm all the tubercle rows are prominent and spinose. As growth proceeds rib density is reduced, the ribs themselves becoming higher and sharper. They mostly arise in pairs from the umbilical tubercles, though some intercalatories are also present. The ribs are projected forwards over the ventrolateral shoulders but are interrupted along the median line by an increasingly broad ventral sulcus. REMARKS. The main characteristics of this form are its great whorl breadth and strong orna- ment. No previously described Watinoceras shows such a depressed whorl section at compara- ble diameters. W. devonense Wright & Kennedy (1981: 52; pl. 10, figs 7, 10, 12?, 13, 16) and W. coloradoense praecursor Wright & Kennedy (1981: 52; pl. 10, figs 4, 8, 9, 11, 15, 17, 18; text-figs 19G, 19K) show a rather similar style of ventral ribbing but are more compressed and the described material is of much larger size, precluding detailed comparison. The closely related genus Benueites Reyment, 1954 generally shows a less regular ornament than Watinoceras. The inner whorls of one species, B. trinidadensis Renz, however, may exhibit a similar style of ribbing and tuberculation to the present material (see Renz 1982: pl. 28, figs 14a, 14b) but they are again much more compressed. The Nigerian specimens may well represent a new species but without knowledge of their outer whorls this matter cannot be decided and it would be premature to describe them as such. Their initial whorls are similar to those in the contempora- neous material referred above to Watinoceras aff. amudariense. The latter, however, retains its fine ribbing and compressed whorls later into ontogeny. Reyment (1971), impressed by the fact that Benueites is frequently to be found represented by finely and coarsely ornamented forms lying side by side, considered that a novel form of ornamental dimorphism characterized the genus. Cooper (1978) doubted this view, as did Renz (1982: 91) who noted that the more coarsely ribbed members in Venezuela occupy a lower stratigraphical position. Cooper (1978: 120-122) himself proposed that Watinoceras amudariense was the microconch of W. color- adoense (Henderson), a large, coarsely decorated form (see, for example, Cobban & Scott 1972: 76; pl. 27, figs 11-19; pl. 28, figs 1-3, 5—9; text-figs 35-37; Wright & Kennedy 1981: 53, text-figs 18C-F). This suggestion was in turn doubted by Renz (1982: 93), and by Wright & Kennedy (1981: text-fig. 18G) who figured a large, densely ribbed Watinoceras which they suggested might represent the macroconch of W. amudariense. The nature of any dimorphism shown in W atinoceras therefore remains uncertain. It is of interest, however, to note the likeness of the earliest whorls in the two Nigerian forms described here, but whether this similarity has any special significance is difficult to ascertain, especially without the knowledge of their adult whorls. Genus MAMMITES Laube & Bruder, 1887 TYPE SPECIES. Ammonites nodosoides Schliiter, 1871; by monotypy. Mammites nodosoides (Schliter, 1871) Figs 18-20, 22 1829 Ammonites nodosoides Schlotheim [MS]; von Buch: 424 (nom. nud.). 1871 Ammonites nodosoides Schliiter: 19; pl. 8, figs 1—4. 1887 Mammites nodosoides Schlotheim sp. Laube & Bruder: 229; pl. 25, figs 1a, 1b. 1903 Schlutericeras nodosoides (Schliiter) Hyatt: 111. 1907 Mammites nodosoides (Schlotheim); Pervinquiére: 309; pl. 18, figs 1a, 1b. 1907 Mammites nodosoides var. afra Pervinquiére: 310; pl. 18, figs 2, 3; text-fig. 18. 1972 Mammites nodosoides (Schlotheim); Cobban & Scott: 78 (with synonymy). 1981 Mammites nodosoides (Schliiter); Wright & Kennedy: 75; pl. 17, fig. 3; pl. 19, fig. 3; pl. 20, fig. 4; pl. 22, fig. 4; pl. 23, figs 1-3; pl. 24, figs 2, 3; text-figs 19B, 23, 24 (with synonymy). 1982 Mammites nodosoides (Schlotheim); Renz: 89; pl. 27, figs 1-10. MATERIAL AND OCCURRENCE. Seventy-two specimens (C.83517, C.85293-8, C.90327-41, UIN 486.1—50) from the Eze-Aku Formation (Lower Turonian, Mammites nodosoides Zone), Lok- panta, south-east Nigeria. NIGERIAN LOWER TURONIAN AMMONITES 41 REMARKS. This important species occurs abundantly in the upper part of the Lower Turonian at Lokpanta. Variation in the material is slight. The development of the outer ventrolateral tubercles is a little inconsistent. In the later stages they become highly clavate and sometimes fuse with the inner ventrolateral tubercles. Whorl height is greater than whorl breadth in the early and middle stages, but in the largest specimens collected the whorls become rather less compressed and here may resemble those in Mammites wingi Morrow (see Cobban & Scott 1972: 79; pl. 26, figs 1-4, 9, 10; pls 31-33; text-fig. 38; Wright & Kennedy 1981: 79; pl. 25, fig. 2; pl. 26, fig. 1; text-figs 25, 27). The inner whorls of M. wingi differ, however, in their more delicate ornament while rib density is greater in the adult stages. Mammites? sp. Fig. 21a, b MATERIAL AND OCCURRENCE. A single specimen (C.90360) from the Eze-Aku Formation (basal Turonian), Lokpanta, south-east Nigeria. REMARKS. This small fragment, probably referable to the genus Mammites, shows a rather depressed whorl section. There are broad, rounded ribs, convex on the flanks and bent for- wards over the venter, carrying strong umbilical and pointed inner and outer ventrolateral tubercles. Intercalated ribs, bearing outer ventrolateral tubercles alone, are present upon the venter. Such meagre material is impossible to identify to species level, and even the generic assign- ment is questionable. It does, however, resemble the inner whorls of the basal Turonian Mammites dixeyi Reyment (1955: 50; pl. 9, fig. 4; pl. 10, fig. 3; pl. 11, figs 2a, 2b; text-figs 20, 21) from Nigeria. A Mammites from Trinidad (Reyment 1972: 365; fig. 8, 4a, 4b), probably of a somewhat younger age, is also similar, as is M. nodosoidesappelatus Etayo-Serna (1979: 85; pl. 13, fig. 1) from Colombia. Family VASCOCERATIDAE Douville, 1912 (nom. correct. & transl. Spath, 1925; ex Vascoceratinés) Subfamily VASCOCERATINAE Douville, 1912 Genus FAGESIA Pervinquiere, 1907 TYPE SPECIES. Olcostephanus superstes Kossmat, 1897; by original designation. Fagesia levis Renz, 1982 Figs 23-4, 27-8 1982 Fagesia levis Renz: 78; pl. 22, figs 20a, 20b; pl. 23, figs 1-3; text-figs 53, 59a—c. MATERIAL AND OCCURRENCE. Twenty-nine specimens (C.85281—3, C.90319—20, UIN 487.1—24) from the Eze-Aku Formation (Lower Turonian, Mammites nodosoides Zone), Lokpanta, south- east Nigeria. A further six specimens (C.90321-—6) from the same horizon and locality probably also belong here (Figs 25-6). DESCRIPTION. Six small specimens (C.90321—6) show what are probably the juvenile whorls of this species. The shell at this stage is evolute and the whorls only slightly depressed with a rounded venter. There are strong, rounded ribs mostly springing in pairs from rather spinose umbilical turbercles and bending forwards over the venter. Intercalated ribs also occur. The ribs fade early in ontogeny and between diameters of about 45 and 60 mm they practically disappear, being represented thereafter by faint, broad folds, convex over the venter. The shell now becomes globular and evolute with 10-11 strong, rounded tubercles at the umbilical shoulder in each whorl. At diameters around 100 mm these tubercles become broader, flatter and less distinct and finally fade out altogether. The maximum diameter attained is about 200 mm. The suture is typical of the genus with highly elongated, much incised elements. 42 P. M. P. ZABORSKI NIGERIAN LOWER TURONIAN AMMONITES 43 REMARKS. The inner whorls in Fagesia characteristically show strong ribs arising in twos or threes from umbilical tubercles. The ornament is generally lost later in ontogeny, the stage at which this occurs being a major factor in distinguishing between species. F. superstes (Kossmat 1897: 26; pl. 6, fig. 1; Pervinquiére 1907: 322; pl. 20, figs 1-4) shows persistently strong ribbing and umbilical tuberculation, while F. thevestensis (Peron 1896: 23; pl. 7, figs 2, 3; Pervinquiére 1907: 325; pl. 20, figs 5, 6) and the similar F. boucheroni (Coquand) (see Kennedy & Wright 1979a: 669, text-figs 1A, 1B) lose their ribbing a little earlier. F. peroni Pervinquiére (1907: 329; pl. 20, figs 7, 8), on the other hand, loses first its ribbing, then its tuberculation, at a very early ontogenetic stage. The present material is intermediate between F. peroni and F. thevestensis in these respects. It conforms closely with the globular, evolute, early Turonian F. levis Renz (1982: 78; pl. 22, fig. 20; pl. 23, figs 1-3; text-figs 53, 59a—c) from Venezuela. Ontogenetic development in this species is very similar to that in the Nigerian specimens, though juvenile F. levis (see Renz 1982: pl. 23, figs 3a, 3b) are rather less densely ribbed than the presumed early whorls in the present Nigerian material (Figs 25-6). At this growth stage the latter is closer to the Colombian F. zanelli Etayo-Serna (1979: 89; pl. 13, fig. 11; pl. 14, fig. 5). In this form, however, the ribbing is said to strengthen during ontogeny with prominent narrow ribs persisting up to diameters of at least 50 mm (see Etayo-Serna 1979: pl. 14, fig. 5). Venezuelan specimens occurring with F. levis and referred to F. aff. superstes by Renz (1982: 78; pl. 22, figs 19a, 19b; pl. 23, figs 4a, 4b) differ only in having slightly stronger, more persistent ribbing. The Nigerian material includes forms such as these (Fig. 24); they are probably variants of F. levis. F. bomba (Eck 1909: 179; pl. 17, figs 1, 2), the similar F. involuta Barber (1957: 27; pl. 9, fig. 3; pl. 29, figs 6, 7) and F. simplex Barber (1957: 27; pl. 8, fig. 1; pl. 29, figs 4, 5) all lose their ribbing early, but the first two species are markedly more involute than the present material, while F. simplex has a much simpler suture pattern. F. lenticularis Freund & Raab (1969: 36-42; pl. 6, figs 3-7; pl. 7, figs 1-3; pl. 8, figs 1, 2; text-figs 7h—-k, 8a—i, 9a—c) and its varieties have a peculiar, eccentric mode of coiling. F. rudra (Stoliczka 1865: 122; pl. 60, fig. 1; Kennedy & Wright 1979a: 666; pl. 82, figs 1, 2) lacks umbilical tubercles. F. pachydiscoides Spath (see Wright & Kennedy 1981: 97, text-fig. 37) is more compressed. In addition, its umbilical tuber- cles persist to a very late stage, as is the case in F. catinus (Mantell) (see Wright & Kennedy, 1981: 88; pl. 26, fig. 2; text-figs 31-36, for review and synonymy) where they increase in strength during ontogeny. The ribbing also persists longer in F. catinus (see Powell 1963: 320; pl. 33, fig. 2; pl. 34, figs 1-5). Genus NEOPTYCHITES Kossmat, 1895 TYPE SPECIES. Ammonites telinga Stoliczka, 1865 (=A. cephalotus Courtiller, 1860); by original designation. Neoptychites cephalotus (Courtiller, 1860) Figs 31-2 1860 Ammonites cephalotus Courtiller: 248; pl. 2, figs 1-4. 1865 Ammonites xetra Stoliczka: 124; pl. 61, figs 1, 2. 1865 Ammonites telinga Stoliczka: 125; pl. 62, figs 1, 2. 1895 Neoptychites xetra (Stoliczka) Kossmat: 72. 1895 Neoptychites telinga (Stoliczka) Kossmat: 71; pl. 7, fig. 1. Fig. 22a, b Mammites nodosoides (Schliiter). Eze-Aku Formation (Lower Turonian, Mammites nodo- soides Zone), Lokpanta, south-east Nigeria. C.90327, x 1. See also Figs 18-20. Figs 23-24 Fagesia levis Renz. Eze-Aku Formation (Lower Turonian, Mammites nodosoides Zone), Lokpanta, south-east Nigeria. Fig. 23, C.90319, x1. Lateral view of specimen shown in Fig. 27. Fig. 24, C.90320, x 1. A variant with abnormally persistent ribbing. See also Figs 25—26(?), 27—28. Figs 25-26 Fagesia levis Renz?. Eze-Aku Formation (Lower Turonian, Mammites nodosoides Zone), Lokpanta, south-east Nigeria. The presumed early whorls. Fig. 25a, b, C.90321, x1. Fig. 26, C.90322, x 1. 44 P. M. P. ZABORSKI Figs 27-28 Fagesia levis Renz. Eze-Aku Formation (Lower Turonian, Mammites nodosoides Zone), Lokpanta, south-east Nigeria. Fig. 27, C.90319, x 1. Ventral view of specimen shown in Fig. 23. Fig. 28a, b, C.85282, x 1. See also Figs 23-24, 25-26(?). Fig. 29a, b Thomasites gongilensis (Woods). Eze-Aku Formation (uppermost Cenomanian or lower- most Turonian), Ezillo, south-east Nigeria. C.90354, x 1. See also Figs 34-35. Fig. 30a, b Herrickiceras? sp. Eze-Aku Formation (Lower Turonian, Mammites nodosoides Zone), Lokpanta, south-east Nigeria. C.85287, x 1. NIGERIAN LOWER TURONIAN AMMONITES 45 1907 Neoptychites cephalotus (Courtiller) Pervinquiére: 393; pl. 27, figs 1-4; text-fig. 152. 1907 Neoptychites gourguechoni Pervinquiére: 400; pl. 27, figs 8, 9; text-figs 155, 156. 1979a Neoptychites cephalotus (Courtiller); Kennedy & Wright: 670; pl. 82, figs 3-5; pl. 83, figs 1-3; pl. 84, fig. 3; pl. 85, figs 1-5; pl. 86, figs 5, 6; text-fig. 2 (with synonymy). 21979 Franciscoites suarezi Etayo-Serna: 87; pl. 13, fig. 2; text-figs 8X, Y, wu. 1982 Neoptychites aff. crassus Solger; Renz: 88; pl. 26, figs 16a, 16b. 1982 Neoptychites aff. telingaeformis discrepans Solger; Renz: 88; pl. 26, figs 17a, 17b. 21982 Neoptychites transitorius Renz: 87; pl. 26, figs 15, 18; text-figs 65d, 66A, 66a—d. MATERIAL AND OCCURRENCE. At least five specimens, four (C.83512, C.85289, C.90317—8) from the Eze-Aku Formation (Lower Turonian, Mammites nodosoides Zone), Lokpanta, south-east Nigeria; the other (C.90359) from a slightly lower (basal Turonian) horizon at Lokpanta. REMARKS. This material is identical with Neoptychites cephalotus (Courtiller). The smallest specimen (C.90359), with a diameter of 21mm, is moderately compressed, shows a sharply rounded venter and has the characteristic collared constrictions of the juveniles in this species (see, for example, Solger 1904: pl. 3, fig. 4; Pervinquiére 1907: pl. 37, figs 3a, 3b; Riedel 1932: pl. 26, fig. 5; Reyment 1972: fig. 7, 1-3; Renz 1982: pl. 26, figs 17a, 17b; Cobban & Hook 1983: pl. 3, figs 9-11). Franciscoites suarezi Etayo-Serna (1979: 87; pl. 13, fig. 2), known only from juveniles, is also very similar and seems to be a synonym. The largest Nigerian specimen (C.85289) has a diameter of some 200mm. It is adult and shows the distinctive constricted aperture of this species. Its whorls are smooth and triangular, the venter being narrowly rounded and raised up slightly along the median line on the early part of the body chamber. Kennedy & Wright (1979a: 670-680) have discussed the genus Neoptychites at length. They treated N. xetriformis Pervinquiere as a distinct species but noted its association with N. cephalotus in the Touraine area of France and raised the possibility of the two being dimorphs. Cobban & Hook (1983: 14-15), working with large collections from New Mexico, considered N. xetriformis to be a synonym of N. cephalotus. They regarded the latter species as highly variable, including both small, stout, ribbed forms of N. xetriformis character, as well as larger, slender, more weakly ornamented individuals having the form of N. cephalotus. They, too, suggested dimorphism but were disturbed by the lack of a clearly bimodal size pattern within the population. The Nigerian material is too sparse to contribute greatly to this problem and a conservative approach is followed here. It may be relevant, however, that along with the material from the Mammites nodosoides Zone at Lokpanta there comes a rather poorly pre- served specimen (C.85288) of 90mm diameter which shows broad whorls, a rounded venter and distinct, broad ribs. This individual is close to N. xetriformis, adding to the evidence for its coexistence with N. cephalotus. Kennedy & Wright (1979a: 680-681) showed that in France Neoptychites occurs some way up in the Turonian, above beds with Mammites of the nodosoides group. They regarded it as being approximately contemporaneous elsewhere in the world except for the western interior of the United States (Colorado) where it appears in the basal Turonian Zone of Watinoceras coloradoense (see Cobban & Scott 1972; Kauffman et al. 1978). The Nigerian material confirms this early occurrence of the genus, it being found below beds with Mammites nodosoides. Subfamily PPEUDOTISSOTIINAE Hyatt, 1903 Genus THOMASITES Pervinquiere, 1907 TyPE SpEcIES. Pachydiscus rollandi Peron, 1889; by original designation. REMARKS. The close similarity between Thomasites and Gombeoceras Reyment has been remarked upon by several authors. Basse (1940: 457) included Vascoceras gongilense Woods, the type species of Gombeoceras, under synonymy in Thomasites. Reyment (1954: 151), in proposing the genus Gombeoceras, differentiated it from Thomasites by its more evolute shell, less triangular whorls, weaker umbilical tuberculation and non-constricted aperture. Freund & Raab (1969: 42-43), however, considered the morphological range in Thomasites to be wide enough to include Gombeoceras. They noted that the Nigerian material of Gombeoceras figured 46 ee M. P. ZABORSKI NIGERIAN LOWER TURONIAN AMMONITES 47 by Reyment (1954, 1954a, 1955) and Barber (1957) consisted entirely of phragmocones, the nature of the adult aperture being unknown. In fact, an undescribed, entire, adult specimen of G. gongilense (C.47561) from the Numan area of north-east Nigeria does indeed show the moderately constricted aperture characteristic of adult Thomasites rollandi (see Pervinquiére 1907: pl. 22, figs 4a, 4b). The Nigerian specimen has a diameter of 130mm; in T. rollandi the constricted aperture appears at diameters of 110-130mm (Pervinquiére 1907: 343). Reyment (1979) sought to maintain the separate status of Gombeoceras, stressing that Thomasites lacks a median or ventrolateral keel at any ontogenetic stage. Wright & Kennedy (1981: 99), however, pointed out that the siphonal ornament is highly variable in both Thomasites and Gombeoceras, and they could find no character to distinguish between the two. In view of their clear simi- larity, this latter view is followed here, Gombeoceras being treated as a synonym of Thomasites. Thomasites gongilensis (Woods, 1911) Figs 29, 34-5 1911 Vascoceras gongilense Woods: 282; pl. 21, fig. 7; pl. 22, fig. 1. 1954 Gombeoceras gongilense (Woods) Reyment: 151; pl. 2, fig. 1; pl. 3, fig. 6; text-fig. 1. 1954a Gombeoceras subtenue Reyment: 261; pl. 4, fig. 4; text-fig. 3f. 1955 Gombeoceras gongilense (Woods); Reyment: 63; pl. 14, fig. 5; pl. 21, fig. 4. 1957 Gombeoceras gongilense (Woods); Barber: 79; pl. 17, figs 1-6; pl. 18, figs 1-4; pl. 19, figs 1-6; pl. 20, fig. 3; pl. 37, figs 1-20. 1965 Gombeoceras gongilense (Woods); Reyment: pl. 2, fig. 3; pl. 3, figs 16, 19. 1976 Gombeoceras gongilensis (Woods); Offodile: 69; pl. 12, fig. 3. 1976 Gombeoceras compressum Barber; Offodile: 69; pl. 12, fig. 4. 1976 Gombeoceras gongilense (Woods); Offodile & Reyment: 58, text-figs 31, 32. 1976 Gombeoceras compressum Barber; Offodile & Reyment: 59, text-fig. 33. 1981 Thomasites gongilensis (Woods) Wright & Kennedy: 100; pl. 24, fig. 1; pl. 25, fig. 1. MATERIAL AND OCCURRENCE. Two specimens (C.90353-4) from the Eze-Aku Formation (uppermost Cenomanian or lowermost Turonian), Ezillo, south-east Nigeria. A further speci- men (C.90355) from the same horizon and locality may also belong here. REMARKS. Barber (1957) demonstrated very clearly the wide degree of variation shown by populations of Thomasites gongilensis in north-east Nigeria. He suspected, but was unable to prove, that this variability was partly stratigraphical and geographical as well as individual in nature. He therefore separated his diverse morphotypes into a number of subspecies. Reyment (in Offodile & Reyment 1976: 53), however, preferred to regard these forms as separate species. In as much as they occur side by side or within a few metres of section in both the middle Benue Valley and north-eastern regions of Nigeria (Barber 1957, Offodile & Reyment 1976, Wozny & Kogbe 1983) and that they tend to grade into one another, it is probably more correct to regard them as mere varieties of T. gongilensis. Wright & Kennedy (1981: 100) suspected that a similar situation might prove to exist amongst the Tunisian populations of Thomasites described by Pervinquiere (1907); all these forms may be varieties of T. rollandi (Peron). Two varieties of T. gongilensis have been identified from Ezillo. The first (C.90354, Fig. 29a, b) is compressed, involute and with flattened flanks and a rounded venter. There are feeble ribs on the outer parts of the flank and venter; the latter also bears weak ventrolateral and siphonal tubercles. In all these respects, as in suture pattern, this specimen conforms closely with the variety T. gongilensis var. compressus (see Barber 1957: 41; pl. 19, figs 2a, 2b, Sa, Sb; pl. 33, figs 15, 16), the most compressed and one of the most feebly ornamented varieties of this species. The Tunisian form named T. meslei by Pervinquiére (1907: 345; pl. 22, figs 8, 9) is very Figs 31-32 Neoptychites cephalotus (Courtiller). Fig. 31a, b, C.85289, x 0-65. Eze-Aku Formation (Lower Turonian, Mammites nodosoides Zone), Lokpanta, south-east Nigeria. Fig. 32, C.90359, x 1. Eze-Aku Formation (basal Turonian), Lokpanta, south-east Nigeria; latex cast. Fig. 33a—c Thomasites koulabicus (Kler). Eze-Aku Formation (uppermost Cenomanian or lower- most Turonian), Ezillo, south-east Nigeria. C.90352, x 1. 48 P. M. P. ZABORSKI close to T. gongilensis var. compressus. Wright & Kennedy (1981: 100) noted the morphological overlap between the Tunisian and Nigerian populations of Thomasites but because the former was poorly preserved were unable to decide whether T. gongilensis should be brought into synonymy under T. rollandi. The second specimen (C.90353, Fig. 34a, b) is moderately evolute and moderately com- pressed, though slightly crushed laterally. Its venter is broadly rounded at first but shows a low siphonal ridge later in ontogeny. There are low, rounded ribs upon the inner two-thirds of the flanks. The ventral ornament is weak at first but at its adoral end the specimen shows strong ventrolateral tubercles and clavate swellings upon the siphonal ridge. Two varieties of T. gongilensis resemble this specimen in their later whorls. T. gongilensis var. inflatus (Barber 1957: 43; pl. 18, figs 3, 4; pl. 33, figs 18-20) shows weakened ornament and a broad keel in its later stages but is more inflated. Rather closer is T. gongilensis var. tectiformis (Barber 1957: 41; pl. 17, figs 1-4; pl. 19, fig. 6; pl. 33, figs 4-6; Fig. 35) which, while displaying angular ventrolateral shoulders in its early whorls, develops a more rounded whorl section towards adulthood, when its ornament weakens though the keel persists. A third Ezillo specimen (C.90355) has the overall dimensions of T. gongilensis var. costatus (Barber 1957: 41; pl. 18, fig. 1; pl. 19, figs 1, 3; pl. 37, figs 9, 10) but is too poorly preserved for certain identification. In Nigeria, Thomasites has previously been reported from only one locality south of the Benue River, a section in the Konshisha River near Oturkpo (Reyment 1955: 63, 98; Fig. 46, p. 58). The present records, therefore, considerably extend its southerly geographical range. Thomasites koulabicus (Kler, 1909) Fig. 33a—c 1909 Pseudotissotia koulabica Kler: 157; pl. 6, figs 1-3; pl. 7, figs 1, 2; pl. 8, figs 1, 2. 1954a Gombeoceras koulabicum (Kler) Reyment: 261. 1958 Thomasites koulabicus (Kler) Tsagareli, Glazunova, Luppov & Mikhailov (in Orlov): 124; pl. 61, figs 3a, 3b; text-fig. 99b. 1966 Koulabiceras koulabicum (Kler) Atabekyan: 77. 1969 Gombeoceras (Ferganites) koulabicum (Kler); Stankievich & Pojarkova: 94; pl. 2, figs 3a, 3b; pl. 3, figs la, 1b. 21969 Gombeoceras (Ferganites) kanicum Stankievich & Pojarkova: 95; pl. 3, figs 2a, 2b. 1969 Gombeoceras (Ferganites) kleri (Luppov MS) Stankievich & Pojarkova: 96; pl. 4, figs 1-3. 1981 Thomasites koulabicus (Kler); Wright & Kennedy: 100. MATERIAL AND OCCURRENCE. A single specimen (C.90352) from the Eze-Aku Formation (uppermost Cenomanian or lowermost Turonian), Ezillo, south-east Nigeria. DESCRIPTION. This specimen is moderately evolute. At first its whorls are rather inflated and the venter arched, but the body chamber becomes compressed with flatter flanks and a broadly rounded venter. The ornament of the phragmocone is very coarse; massive umbilical bullae give off pairs of strong, rounded ribs, while additional intercalated ribs arise some distance down the flanks. There are prominent ventrolateral tubercles, and clavate siphonal tubercles which tend to fuse so producing an intermittent keel. Upon the body chamber the ornament weakens, the ribs becoming flatter and lower, though the umbilical tubercles remain prominent. REMARKS. As Tsargareli et al. (in Orlov 1958: 124) and Wright & Kennedy (1981: 99) have pointed out, Pseudotissotia koulabica Kler (1909: 157; pls 6-8), previously unknown outside Soviet Central Asia, is a Thomasites. It represents the most coarsely ornamented member of the genus yet known. Reyment (1954a: 261) and Barber (1957: 39, 45) were of the opinion that the morphological range of T. gongilensis overlaps that of T. koulabicus. Wright & Kennedy (1981: 99-100) preferred to maintain the two as distinct species. Since the ornament in T. koulabicus is consistently stronger than that in even the most coarsely decorated variety of T. gongilensis, T. gongilensis var. crassicostatus (Barber 1957: 45; pl. 18, figs 2a, 2b; pl. 33, figs 7, 8), and since the adult body chamber appears to undergo wholesale compression at adulthood (see Figs 33a, c; Kler 1909: pl. 8, fig. 1; Stankievich & Pojarkova 1969: pl. 3, figs 1a, 1b) rather than merely NIGERIAN LOWER TURONIAN AMMONITES 49 showing a constricted aperture like that in T. gongilensis, Wright & Kennedy’s view is followed here. The size at which compression of the body chamber takes place in T. koulabicus is, however, variable. In the present material it takes place at an overall diameter of some 85mm (phragmocone diameter 55mm) when the ornament becomes effaced, indicating adulthood. In specimens figured by Kler (1909: pl. 7, figs 1, 2; pl. 8, fig. 1) these characters do not appear until phragmocone diameters of 90-95mm and certain specimens (see Kler 1909: 157-158) reach overall diameters of nearly 150mm. Perhaps this size difference reflects dimorphism. The genus Koulabiceras Atabekyan, 1966 (type species Pseudotissotia koulabica Kler) is a subjective synonym of Thomasites. Gombeoceras (Ferganites) Stankievich & Pojarkova (1969: 94) shares this type species and is therefore an objective synonym of Koulabiceras. Stankievich & Pojarkova (1969) referred the following species to G. (Ferganites): G. (F.) koulabicum (Kler), G. (F.) kleri Stankievich & Pojarkova (1969: 96; pl. 4, figs 1-3) and G. (F.) kanicum Stankievich & Pojarkova (1969: 95; pl. 3, figs 2a, 2b). The last of these has rather broad whorls, especially in its later growth stages. All three are nevertheless similar and, bearing in mind the wide degree of variation shown by species of Thomasites, all three probably belong in T. koulabicus (see also Wright & Kennedy 1981: 100). Stankievich & Pojarkova (1969: 91—93; pl. 1, figs 2, 3; pl. 2, figs 1, 2) described several additional Thomasites from the Soviet Union as T. cf. glo- bosotuberculatus Pervinquiere, T. cf. jordani Pervinquiére and T.(?) inflatus sp. nov. These show a consistently strong ornament which is only rarely approached by the Tunisian material described by Pervinquiére (1907). All the central Asiatic forms come from a similar strati- graphical horizon (Stankievich & Pojarkova 1969: 87—88) and one might speculate that, just as in the cases of the Tunisian and north-east Nigerian populations of Thomasites, they represent a complex of variants all belonging to the single coarsely ornamented species T. koulabicus. Genus WRIGHTOCERAS Reyment, 1954 TYPE SPECIES. Bauchioceras (Wrightoceras) wallsi Reyment, 1954; by original designation. REMARKS. Members of this genus were first described from the Lower Turonian of Tunisia as Hoplitoides munieri Pervinquiére (1907: 217; pl. 10, figs 1, 2) and H. mirabilis Pervinquiére (1907: 218; pl. 10, fig. 3). Pervinquiere (1907: 215-216) was aware that these forms were untypical of Hoplitoides von Koenen in that they exhibit tabulate venters not in the juvenile stages alone but throughout ontogeny. Rather than refer them to a new genus, however, he preferred to emend the diagnosis of Hoplitoides (see Solger 1904) and recognized within it two groups, those with tabulate venters throughout growth and those with narrowly rounded or sharpened venters in their later stages. He regarded the former group as ancestral to the latter. Kummel & Decker (1954) described similar, broad-ventered material from northern Mexico as ‘Hoplitoides’ cf. ‘H. munieri Pervinquiére but indicated that it did not truly belong in this genus and should be included in a new taxon (see also Benavides-Caceres 1956: 476). Reyment (1954), working primarily with material from northern Nigeria, proposed Wrightoceras, originally as a subgenus of Bauchioceras Reyment, 1954, in which he included B. (W.) wallsi as type species along with Hoplitoides munieri and H. mirabilis. Later, Reyment (1955) treated both Bauchio- ceras and Wrightoceras as subgenera of Pseudotissotia Peron, Wrightoceras being distinguished mainly by its lack of a siphonal keel. Barber (1957) followed this procedure. Kennedy, Cooper & Wright (1979) re-examined Ammonites gallienni d’Orbigny, the type species of Pseudotissotia, and suggested that Bauchioceras be treated as a strict synonym of Pseudotissotia. Wrightoceras, however, was maintained as a separate genus altogether since it shows at most a feeble siphonal keel in the juvenile stages only and possesses comparatively weak, impersistent ornament. At present, the following can be referred to Wrightoceras: W. wallsi Reyment, W. munieri (Pervinquiére), of which W. mirabilis (Pervinquiére) may be a synonym, W. inca (Benavides- Caceres), W. llarenai (Karrenberg), W. submunieri Wiedmann and W. reymenti Collignon & Roman. Pseudotissotia gagnieri Faraud (1951: 149; pl. 5, fig. 1) may also belong here but develops an untypical rounded venter. The Colombian genus Imlayiceras Leanza, 1967 (type species Imlayiceras washbournei Leanza 1967: 198; pl. 4, figs 1-4; pl. 6, figs 1, 4-6) differs from Wrightoceras only in the reported presence of faint constrictions on the early whorls. The later P. M. P. ZABORSKI 50 NIGERIAN LOWER TURONIAN AMMONITES 51 stages in the two genera are, however, indistinguishable and the Colombian material is perhaps better included in Wrightoceras also. The relative stratigraphical positions of Wrightoceras and Hoplitoides are discussed below (p. 53). Wrightoceras wallsi Rement, 1954 Figs 36-7 1954 Bauchioceras (Wrightoceras) wallsi Reyment: 160; pl. 2, fig. 4; pl. 3, figs 3, 3a. 1955 Pseudotissotia (Wrightoceras) wallsi (Reyment) Reyment: 71; pl. 24, fig. 1; text-figs 32a, 32c, 32d. 1957. Pseudotissotia (Wrightoceras) wallsi (Reyment); Barber: 53; pl. 24, figs 1, 2; pl. 34, figs 5, 13. 1965 Pseudotissotia (Wrightoceras) wallsi (Reyment); Reyment: pl. 4, figs 14-17. MATERIAL AND OCCURRENCE. Two specimens (C.90349-50) from the Eze-Aku Formation (Lower Turonian), Ezillo, south-east Nigeria. REMARKS. Barber (1957: 51) found populations of Wrightoceras wallsi to show a variable degree of compression. The two present specimens represent relatively slim variants and tend towards the material referred below to W. cf. munieri (Pervinquiére). The smaller of the two (C.90349, Fig. 36a, b), with a diameter of 70mm, has at first a slightly concave venter bordered by ventrolateral keels, becoming more tabulate later. It is exactly comparable with an individual of W. wallsi (C.47617, Fig. 37) from Kanawa near Gombe in north-east Nigeria at an equivalent size. The larger of the two specimens (C.90350) has a diameter of 240mm and, apart from having a slightly more rounded adult venter, is very close to the large form from Deba-Habe near Gombe (C.47421) figured by Reyment (1955: pl. 24, fig. 1). Material from northern Mexico described by Kummel & Decker (1954: 317; pl. 33, figs 1, 2; text-figs 7, 10) as ‘Hoplitoides’ cf. “‘H. munieri Pervinquiére has a rather broader and more sulcate venter than Pervinquiére’s species, as these authors pointed out. It is close to W. wallsi but shows weak ribs on the flanks which sometimes cross the venter. Wrightoceras cf. munieri (Pervinquiere, 1907) Figs 38, 40 cf. 1907 Hoplitoides munieri Pervinquiére: 217; pl. 10, figs 1, 2. cf. 1907 Hoplitoides mirabilis Pervinquiére: 218; pl. 10, fig. 3. cf. 1969 Hoplitoides cf. H. mirabilis Pervinquiére; Freund & Raab 65; pl. 10, figs 1, 2; text-figs 13i-1. cf. 1982 Hoplitoides munieri Pervinquiére; Renz: 100; pl. 31, figs 3, 4, 6, 11. cf. 1982 Hoplitoides cf. munieri Pervinquiére; Renz: 100; pl. 31, fig. 5. MATERIAL AND OCCURRENCE. Four specimens (C.90345-—8) from the Eze-Aku Formation (Lower Turonian), Ezillo, south-east Nigeria. DESCRIPTION. These forms are compressed, involute and smooth at all observed growth stages. The early whorls are a little inflated in the mid-flank region and have a moderately broad, sulcate venter up to a diameter of about 60mm. The flanks later become more flattened and the venter tabulate. At diameters in excess of 70mm the ventrolateral shoulders become rounded. The suture is not displayed. Figs 34-35 Thomasites gongilensis (Woods). Fig. 34a, b, C.90353, x 1. Eze-Aku Formation (upper- most Cenomanian or lowermost Turonian), Ezillo, south-east Nigeria. Fig. 35, C.47555, x 1. Dukul Formation (uppermost Cenomanian or lowermost Turonian), near Numan, north-east Nigeria. See also Fig. 29a, b. Figs 36-37 Wrightoceras wallsi Reyment. Fig. 36a, b, C.90349, x 1. Eze-Aku Formation (Lower Turonian), Ezillo, south-east Nigeria. Fig. 37, C.47617, x 1. Pindiga Formation (Lower Turonian), Kanawa, north-east Nigeria. Fig. 38a, b = Wrightoceras cf. munieri (Pervinquiére). Eze-Aku Formation (Lower Turonian), Ezillo, south-east Nigeria. C.90347, x 1. See also Fig. 40. Fig. 39a, b Hoplitoides latefundatus sp. nov. Eze-Aku Formation (Lower Turonian, Mammites nodo- soides Zone), Lokpanta, south-east Nigeria. Paratype C.85284, x 1. See also Figs 41-44. 52 P. M. P. ZABORSKI 2cm Fig. 40 Whorl section in Wrightoceras cf. / munieri (Pervinquiére). Based on specimen /, C.90345. See also Fig. 38a, b. REMARKS. These specimens differ from the material referred above to Wrightoceras wallsi in their rather narrower venters which tend to become more distinctly rounded on the later whorls. The differences are, however, more of degree than kind and the two forms tend to grade into one another in the Ezillo population. The closest previously described species is W. munieri (Pervinquiére) which is distinguished from its contemporary W. mirabilis (Pervinquiére) on sutural grounds alone. These two have virtually identical gross morphologies and, as suggested by Reyment (1954: 157), Benavides-Caceres (1956: 476) and Barber (1957: 53), they may be synonyms. Comparable material occurs in the Lower Turonian of the Negev, the Hoplitoides cf. mirabilis of Freund & Raab (1969: 65; pl. 10, figs 1, 2), and in Venezuela, the H. munieri and H. cf. munieri of Renz (1982: 100; pl. 31, figs 3-6, 11). Poorly preserved specimens from northern Mexico described by Bose (1920: 225; pl. 19, figs 1-3) also seem to be closely related. Wrightoceras inca (Benavides-Caceres 1956: 475; pl. 63, figs 6-11) is another similar form, said to be distinguished from W. munieri by faint, falciform ribs on the inner whorls. W. submunieri Wiedmann (see Wiedmann 1975: figs 6A—C; 1979: pl. 8, fig. 1; Wiedmann & Kauffman 1978: pl. 8, fig. 2) has a broader venter, more inflated inner flanks and strong ribbing on the early whorls; it resembles W. Ilarenai (Karrenberg 1935: 143; pl. 31, fig. 14; pl. 33, fig. 14). W.(?) gagnieri (Faraud 1951: 149; pl. 5, fig. 1) develops a rounded venter with flank ribbing terminating in ventrolateral tubercles. W. reymenti Collignon & Roman (in Amard et al. 1981: 57; pl. 9, figs 7a, 7b) has a very wide umbilicus, highly inflated whorls and a very broad, tabulate venter. Family COILOPOCERATIDAE Hyatt, 1903 Genus HOPLITOIDES von Koenen, 1898 TYPE SPECIES. Hoplitoides latesellatus von Koenen, 1898 (=Neoptychites ingens von Koenen, 1897); by original designation. REMARKS. The genus Hoplitoides includes involute, slender ammonites in which the venter is sulcate or flattened in the initial whorls and becomes sharp or narrowly rounded later on. The early whorls may be ribbed, with ventrolateral tubercles and umbilical bullae developed in some species. The suture possesses a very wide lateral lobe. The taxonomic history and occurrence of Hoplitoides has been reviewed by Cobban & Hook (1980: 5—6). Pervinquiére (1907: 216) suggested that its ancestors were to be found amongst forms now included in NIGERIAN LOWER TURONIAN AMMONITES 53 Wrightoceras which retain broad venters until adulthood. Reyment (1954: 157-158) at first doubted this view but subsequently (Reyment 1954a: 261; 1955: text-fig. 31) indicated just such a scheme, deriving Hoplitoides from Wrightoceras. Cobban & Hook (1980: 6) alternatively suggested that the ancestral form may have been Choffaticeras Pervinquiére. The Nigerian material described here suggests that Hoplitoides was indeed derived from Wrightoceras, by way of forms such as those referred above to W. cf. munieri. These show a very broadly rounded venter in their adult stages but the juvenile whorls are close to those in slender W. wallsi. Certain forms belonging in this group show a suture pattern intermediate between those characteristic of Wrightoceras and Hoplitoides (see Pervinquiére 1907: text-fig. 84; Freund & Raab 1969: text-figs 13j, 131). As Pervinquiére (1907: 218-219) noted, such types differ from the relatively simple pattern typical of Wrightoceras not in their basic construction, but in the accentuation of their subdivisions. The Nigerian material of W. cf. munieri seems to be of very early Turonian age (see pp. 58-9). The upper part of the Lower Turonian at Lokpanta yields apparently more advanced transitional forms described below as Hoplitoides latefundatus sp. nov. These retain truncated venters until relatively large diameters of 65—75mm but show a suture pattern like that in other Hoplitoides with a deep, wide lateral lobe even at a very early growth stage. The main occurrence of Hoplitoides in Nigeria is at Wadatta near Makurdi, where a thin limestone contains H. ingens (von Koenen), H. gibbosulus (von Koenen), H. cf. wohltmanni (von Koenen), H. koeneni Solger and H. crassicostatus Reyment. Reyment (1955) originally assigned this fauna an early Turonian age but later (Reyment 1978: 2) revised this to early Middle Turonian. In the western interior of the United States Hoplitoides occurs only in New Mexico. The earliest recorded forms come from the lower part of the Zone of Collignoniceras woollgari (Mantell) of early Middle Turonian age. This material includes specimens comparable to H. wohltmanni (see Cobban & Hook 1979: 19; pl. 4, figs 3, 4; text-figs 10, 11; 1980: 7; pl. 1, figs 3, 4; text-figs 4, 5; 1981) but retaining truncated venters to a large diameter, and others close to H. koeneni (see Cobban & Hook 1979: 19; pl. 4, figs 1, 2; 1980: 6; pl. 1, figs 1, 2; pl. 3, figs 4, 5), again showing persistent truncation of the venter. H. sandovalensis Cobban & Hook (1980: 8; pl. 2; pl. 3, figs 6-8, 12-16; pl. 4; pl. 11, fig. 1; pl. 18, figs 4-6; text-figs 6, 7), a species developing a sharp venter at a very early ontogenetic stage, appears in the overlying Zone of Prionocyclus hyatti (Stanton). Elsewhere in the world, specimens close to Wrightoceras munieri (see Freund & Raab 1969: 65; pl. 10, figs 1, 2) occur in ‘Zone 6’ in Israel, towards the top of these authors’ Lower Turonian. In Venezuela similar forms (Renz 1982: pl. 31, figs 3, 5) are amongst the earliest Turonian ammonites known, occurring in Renz’ (1982: 72-73) “Assemblage 1’ and ‘Assemblage 2 of early Turonian age. Wrightoceras characterizes the basal Turonian of the Algerian Sahara, Hoplitoides occurring some distance above (Amard et al. 1981: 43-45) and these genera have a similar stratigraphical distribution in Brazil (Bengtson 1983: 44-47). Wiedmann (1960, 1979) describes three Spanish sections containing Wrightoceras and Hoplitoides. In two of these, at Puentedai and Picofrentes (see Wiedmann 1979: 191-193, text-fig. 15; 207—210, text-fig. 24), the former appears below the latter. At the third, Las Fuentes, however, Hoplitoides is listed as occurring, rather incongruously, very close to the bottom of the Turonian (Wiedmann 1979: 205). This last record notwithstanding, the available evidence indicates that Wrightoceras appears very early in the Turonian. Hoplitoides, on the other hand, is a younger genus, being most common in the Middle Turonian. Indeed, Kennedy & Wright (1984: 288, 290) found no convincing evidence for its occurrence any earlier. The Nigerian Lower Turonian, however, contains the first members of this genus, as may also be the case in Venezuela. Hoplitoides latefundatus sp. nov. Figs 39, 41-4 Compare: 1979 Hoplitoides cf. H. wohltmanni (von Koenen); Cobban & Hook: 20; pl. 4, figs 3, 4; text-figs 10, 11. 1980 Hoplitoides cf. H. wohltmanni (von Koenen); Cobban & Hook: 7; pl. 1, figs 3, 4; text-figs 4, 5. 54 P. M. P. ZABORSKI 1981 Hoplitoides wohltmanni (von Koenen); Cobban & Hook: 30; pl. 5. 1982 Hoplitoides mirabilis Pervinquiere; Renz: 99; pl. 30, figs 6, 7; pl. 31, fig. 10. Ho.ortyPe. C.90302 (Fig. 42a, b), from the Eze-Aku Formation (Lower Turonian, Mammites nodosoides Zone), Lokpanta, south-east Nigeria. PARATYPES. At least fourteen specimens (C.83515, C.85284—5, C.90303-13) from the same horizon and locality as the holotype. Name. From the unusually persistent truncation of the venter. Di1aGnosis. A smooth Hoplitoides retaining a truncated venter until relatively large diameters of 65-75 mm. Venter thereafter sharply rounded. DESCRIPTION. The juvenile whorls are compressed and almost invariably smooth, though rare specimens display faint, broad, prorsiradiate ribs until diameters as large as 45mm. The juve- nile venter is narrow and markedly sulcate until diameters of up to 60mm, thereafter becoming tabulate and finally, at diameters of 65—75mm, sharply rounded. The flanks at this stage are usually smooth, though very weak, fold-like ribs may be present in some specimens. The sutures are displayed only in a juvenile specimen (C.85284) of 25mm diameter; the lateral lobe is broad and deep, the first lateral saddle elongate and rather deeply incised. Along with this material a mass of large, smooth, involute ammonites were collected. These are all badly crushed and cannot be adequately described. Almost certainly, many represent the outer whorls of this species, indicating maximum diameters in excess of 200 mm. REMARKS. This material resembles certain previously-described Hoplitoides in its lack of pro- nounced ornament at any growth stage. H. ingens (von Koenen) includes one variety, H. ingens var. laevis Solger (1904: 145; pl. 5, fig. 9), which is virtually smooth and H. wohltmanni (von Koenen 1897: pl. 1, fig. 2; pl. 2, figs 3, 9; 1898: 11; pl. 2, figs 1, 4, 7; Solger 1904: 133; pl. 5, fig. 7) possesses at most only a feeble decoration on its inner whorls. Both of these species, however, lose their truncated venters much earlier in ontogeny, at diameters around 20mm in H. ingens (see Solger 1904: 141, 144, 145) and 30mm in H. wohltmanni (see Solger 1904: 136). The material from the early Middle Turonian of New Mexico described as H. cf. H. wohltmanni by Cobban & Hook (1979, 1980), on the other hand, retains a truncated venter until diameters of at least 65mm, much longer than in H. wohltmanni, as these authors remarked. It closely resembles the present specimens and, though a little younger, it may be conspecific. Another large Hoplitoides from New Mexico (Cobban & Hook 1981) does not develop a rounded venter until a diameter of nearly 70mm and also seems to be closely related. Specimens from Vene- zuela referred by Renz (1982: 99; pl. 30, figs 6, 7; pl. 31, fig. 10) to Hoplitoides mirabilis Pervinquiere are further smooth examples showing a comparable ontogenetic development. Pervinquiere’s species differs in retaining its truncated venter into adulthood and, as indicated above, is better placed in Wrightoceras. Egyptian material described by Douvillé (1928: 30; pl. 6, figs 2a, 2b) shows a broad venter at diameters of 60mm or more and may be related to H. latefundatus. H. gibbosulus (von Koenen) and its varieties (see Solger 1904: pl. 4, fig. 10; text-figs 44-48; Reyment 1955: pl. 17; fig, 85 pl. 18) figs Le spl 19) tig, 4o pl. 215 tig. 32 ply 22) ties Zea crassicostatus Reyment (1955: pl. 17, figs 3, 4) differ in being strongly ornamented species. H. koeneni Solger (1904: pl. 4, figs 8, 9; Reyment 1955: pl. 17, fig. 7; pl. 22, fig. 4) possesses strong ribs on its inner whorls. H. lacabagnae Etayo-Serna (1979: 91; pl. 13, fig. 9) and the similar H. lagiraldae Etayo-Serna (1979: 92, pl. 13, fig. 14) differ from H. latefundatus in that both show pronounced sickle-shaped ribs; in this ornament they resemble the ‘H. aff. mirabilis ribbed variety of Renz (1982: 91; pl. 13, figs 7-9). H. hernanmojicae Etayo-Serna (1979: 90; pl. 13, figs 4, 8) is a poorly defined species, but is known to lose its truncated venter early in ontogeny. Figs 41-44 Hoplitoides latefundatus sp. nov. Eze-Aku Formation (Lower Turonian, Mammites nodo- soides Zone), Lokpanta, south-east Nigeria. Fig. 41a, b, paratype C.90304, x 1. Fig. 42a, b, holo- type C.90302, x 1. Fig. 43a, b, paratype C.90303, x 1. Fig. 44a, b, paratype C.90309, x 1. See also Fig. 39a, b. 5 NIGERIAN LOWER TURONIAN AMMONITES 56 P. M. P. ZABORSKI Genus HERRICKICERAS Cobban & Hook, 1980 TYPE SPECIES. Placenticeras costatum Herrick & Johnson, 1900; by monotypy. Herrickiceras? sp. Fig. 30a, b MATERIAL AND OCCURRENCE. A single specimen (C.85287) from the Eze-Aku Formation (Lower Turonian, Mammites nodosoides Zone), Lokpanta, south-east Nigeria. REMARKS. Cobban & Hook (1980: 22) proposed the genus Herrickiceras for the single species Placenticeras costatum Herrick & Johnson (1900: 214; pl. 28, figs 2, 3; Cobban & Hook 1980: 22; pl. 19, figs 10-18; text-fig. 16) from the Middle Turonian of New Mexico. It is characterized by a highly involute, compressed shell with a fairly broad, sulcate, bicarinate venter. There are sinuous ribs which are best developed on the outer part of the flank where they are projected forwards and expanded into clavate ventrolateral tubercles. The Nigerian material, admittedly meagre, shows all these features and may be best referred here. It occurs in the upper part of the Lower Turonian. Herrickiceras is otherwise known only from the Rio Puerco Valley area of New Mexico where it occupies a higher stratigraphical position, above the basal Middle Turonian Collignoniceras woollgari woollgari Subzone of the western interior. Stratigraphical conclusions Owing to the wealth of their ammonite faunas, the Nigerian and Camerounian Turonian have been the subject of considerable biostratigraphic attention (see Fig. 45). Reyment (1954) pro- posed their first comprehensive, three-fold, subdivision into: REYMENT: REYMENT (1956, 1965) BARBER (1957) | (1954, 1955) ; VW u YU R P O Youngest Zone of Zone of E N Beds Romaniceras uchauxiense | Romaniceras uchauxiense | Ria [ N +- Zone of Intermediate Hoplitoides ingens Zone of PRESENT WORK | Beds =Zone of Kamerunoceras eschii 5 Kamerunoceras eschii) N.E. NIGERIA LOKPANTA AREA Ww E i =| R Zone of Pseudotissotia 5 if (Bauchioceras) _nigeriensis y Zone of R R Mammites nodosoides oO Zone of Pseudotissotia q Oldest (Wrightoceras) wallsi Zone. of t A Beds (=Zone of Pachyvascoceras | Paravascoceras costatum 7 nN costatum) © | Zone of Pseudotissotia Zone of m nigeriensis beds with A (upper extent uncertain) } Vascoceras bulbosum N Watinoceras Cc E Zone of uN p O Paravascoceras costatum Ba EN aa A Zone of N | Vascoceras bulbosum Fig. 45 Table showing previous and present proposals for a biostratigraphic subdivision of the Turonian stage in Nigeria and Cameroun. The present work deals only with the Lower Turonian. NIGERIAN LOWER TURONIAN AMMONITES 57 1. ‘The Oldest Beds’, containing the rich vascoceratid faunas characteristic of north-east Nigeria (later described in detail by Barber, 1957) but also represented in less abundance at Ezillo in south-east Nigeria; 2. ‘The Intermediate Beds’, with Hoplitoides, Mammites, Kamerunoceras, Benueites, Choffati- ceras and Neoptychites; and 3. ‘The Youngest Beds’, with Hoplitoides and ‘Romaniceras’, occurring in south-west Cameroun only. Reyment (1955) retained this scheme with little modification though Watinoceras was added to the fauna of the ‘Intermediate Beds’. Not recognizing a Middle Turonian substage, he regarded the “Oldest Beds’ and ‘Intermediate Beds’ as Lower Turonian and the ‘Youngest Beds’ as Upper Turonian. Barber (1957) was able to establish a more detailed zonation of the ‘Oldest Beds’ at Pindiga in north-east Nigeria, identifying a basal Zone of Vascoceras bulbosum (Reyment), a middle Zone of Paravascoceras costatum (Reyment) and an upper Zone of Pseudotissotia (Bauchioceras) nigeriensis (Woods). Wozny & Kogbe (1983) found a similar zonation to be applicable at Ashaka Quarry, some 100km north of Pindiga. In later schemes, Reyment (1956, 1965) termed his “Oldest Beds’ the “Zone of Pachyvascoceras costatum Reyment (subsequently the “Zone of Pseudotissotia (Wrightoceras) walls’), his ‘Intermediate Beds’ the ‘Zone of Kameru- noceras eschii (subsequently the “Zone of Hoplitoides ingens’) and his ‘Youngest beds’ the ‘Zone of Romaniceras uchauxiense Collignon’, remarking on the ill-defined nature of this last subdivision. Recent work has shown that the fossiliferous beds in north-east Nigeria described by Barber (1957) are not, in fact, Turonian throughout, but are Upper Cenomanian in their lower part. Barber (1957: 61) himself suspected this, having made a special note of the occurrence of Metengonoceras dumbli (Cragin) in the lowest limestone bed exposed at Pindiga. This species is confined to the Cenomanian elsewhere (see, for example, Cobban & Scott 1972; Kennedy, Juignet & Hancock 1981). At Ashaka Quarry both M. dumbli and Euomphaloceras septemseria- tum (Cragin), a reliable late Cenomanian guide fossil, occur in the zone of Vascoceras bulbosum in the lower part of the limestone-shale sequence exposed there (Wozny & Kogbe 1983). E. septemseriatum is also known from the middle Benue Valley region of Nigeria where it occurs directly below beds with Paravascoceras, Vascoceras, Thomasites, Pseudotissotia, Wrightoceras and Neoptychites (Offodile & Reyment 1976). Though the Upper Cenomanian is, therefore, undoubtedly present in central and north-east Nigeria, the location of its boundary with the Turonian is problematical. The overwhelmingly Tethyan nature of the vascoceratid faunas occurring here precludes ready comparison with the Boreal faunas of areas such as north-west Europe and the western interior of the United States where comprehensive zonal schemes are available and the position of the boundary can be fixed. Thomasites gongilensis, however, occurs occasionally in the Upper Cenomanian of the Sergipe Basin, Brazil (Bengtson. 1983), and Wright & Kennedy (1981) also recovered examples from the uppermost Cenomanian of southern England. Since this species abounds in the Zone of Paravascoceras costatum in north- east Nigeria, Hancock & Kennedy (1981) suggested that the Cenomanian—Turonian boundary might best be placed at the base of the zone of Pseudotissotia nigeriensis, thus leaving only this last part of the sequence within the Turonian. As far as southern Nigeria is concerned, Reyment (1978) removed his ‘Intermediate Beds’ or ‘Zone of Hoplitoides ingens’ fauna from the Lower Turonian, preferring to regard it as early Middle Turonian in age. As a result, only the small fauna from Ezillo listed by Reyment (1955: 98) was assigned to the Lower Turonian. Being composed almost entirely of indigenous species, this fauna defies easy correlation. The extent and faunal character of the Lower Turonian in Nigeria has, therefore, become something of a problem. The present material is thus of great value in reassessing this part of the Nigerian Cretaceous and in suggesting correlation outside the country. In the western interior of the United States and southern England two zones can be recog- nized in the Lower Turonian (see Cobban & Scott 1972, Kauffman et al. 1978, Cooper 1978, Wright & Kennedy 1981): a basal Zone of Watinoceras coloradoense, characterized above all by species of Watinoceras; and an upper Zone of Mammites nodosoides from which comes the 58 P. M. P. ZABORSKI bulk of the described European Lower Turonian ammonites. A similar subdivision seems to be possible in the Lokpanta area of Nigeria (Fig. 45, p. 56). Here, beds dominated by Watinoceras spp. can probably be correlated with the zone of W. coloradoense. Lying above are beds clearly corresponding with the zone of M. nodosoides. They contain the nominal species, Fagesia and Kamerunoceras puebloense, all characteristic of this zone in the United States western interior (see Cobban & Scott 1972). Below the Turonian beds at Lokpanta, shales have yielded a possible Metengonoceras which would indicate the presence of Upper Cenomanian sediments. If confirmed, this occurrence would form an interesting geographical link between beds of this age in north-east Nigeria, the middle Benue Valley and the Calabar region in the extreme south-east of Nigeria (see above, Zaborski 1985). The fauna from Ezillo described here is more difficult to date precisely. It contains both Thomasites and Wrightoceras, though the latter occurs a little above the former in north-east Nigeria (Barber 1957, Wozny & Kogbe 1983). Since the entire Ezillo fauna was collected loose from a section several metres thick, however, it is possible that these genera are strati- graphically separated here also. Clearly the affinities of the Ezillo population are with the vascoceratid forms occurring as much as 600km north in Nigeria, having virtually nothing in common with those at Lokpanta, barely 60km distant (see Fig. 46). In north-east Nigeria Thomasites gongilensis occurs with Paravascoceras, another Ezillo faunal element, in beds which, according to Hancock & Kennedy (1981), lie very close to, and possibly just below, the Cenomanian—Turonian boundary. Wrightoceras wallsi, on the other hand, is confined to the upper part of the fossiliferous sequences at Pindiga and Ashaka and is of early Turonian age. The ammonite-bearing siltstone at Ezillo described here is underlain by a sequence of shales and oyster-rich limestones exposed on the eastern fringe of Ezillo town in the Western Aboine River (Fig. 1, p. 32). From here Reyment (1955: 65) recorded Ezilloella ezilloensis Reyment, a species probably occurring alongside Euomphaloceras septemseriatum in the Upper Cenoma- nian of the middle Benue Valley (Offdile & Reyment 1976). In addition, Offodile & Reyment (1976: 43) reported a possible specimen of Allocrioceras annulatum (Shumard) from the Western sedimentary rocks a (Cretaceous to Recent) 200 km ate numerous localities in upper Benue Trough and southern Chad Basin Met, Eum, Psd, Nig, Pav, Vas, Pam, Fag. Thm, Pst, Wri, Eot Ezillo All?, Mam, Psd, Nig, Ezl, Pay, Fag, Thm, Wri middle Benue Trough vie oe ea Eum, Ezl, Pav, Vas ne Bk Thm, Npt, Pst, Wri e Konshisha River Lokpanta Eokpanta e Psd, Thm, Pst, Wri Pac, Met?, Wat, Mam, Kam, Fag, Wn Npt, Hop, Her? °. b: ay pS Calabar Psc, Moi > “C oCcEAN Fig. 46 Map of Nigeria showing distribution of late Cenomanian and early Turonian ammonite genera. All, Allocrioceras; Pac, Pachydesmoceras; Met, Metengonoceras; Psc, Pseudocalycoceras; Eum, Euomphaloceras; Kam, Kamerunoceras; Moi, Metoicoceras; Wat, Watinoceras; Mam, Mam- mites; Psd, Pseudaspidoceras; Nig, Nigericeras; Ezl, Ezilloella; Pav, Paravascoceras; Vas, Vasco- ceras; Pam, Paramammites, Fag, Fagesia; Thm, Thomasites; Npt, Neoptychites; Pst, Pseudotissotia; Wri, Wrightoceras; Eot, Eotissotia; Hop, Hoplitoides; Her, Herrickiceras. Data from Reyment (1954, 1954a, 1955), Barber (1957), Offodile & Reyment (1976), Wozny & Kogbe (1983), Zaborski (1985) and original. NIGERIAN LOWER TURONIAN AMMONITES 59 Aboine River, another late Cenomanian species of the western interior and elsewhere (Cobban & Scott 1972, Wright & Kennedy 1981: 112). The present Ezillo fauna may therefore be regarded as occurring very close to the Cenomanian—Turonian boundary but almost certainly extending some distance above it. Its direct correlation with the Lokpanta assemblages is, however, not yet possible. Although the fragment of Mammites? described from the W atinoceras-bearing beds at Lokpanta bears some similarity to M. dixeyi, an Ezillo species, there is no useful correspondence between the faunas in these two areas. Only the relatively cosmopolitan genus Fagesia otherwise occurs in common. In the Sergipe Basin, Brazil, Wrightoceras, Thomasites and Paravascoceras occur within a fauna regarded by Bengtson (1983) as basal Turonian (‘Turonian 1’). Since, however, Euompha- loceras septemseriatum is present also, Bengtson (1983: 43-44) admitted the possibility that the base of the Turonian might be better placed at the bottom of the succeeding (‘Turonian 2’) faunal assemblage. This part of the Brazilian Turonian contains Watinoceras amudariense, W. spp., Neoptychites cephalotus and Pachydesmoceras among others, with Fagesia spp. present in the middle part and Mammites nodosoides extending throughout its middle and upper parts. This sequence appears to correlate with both the Watinoceras and M. nodosoides-bearing beds at Lokpanta, although the presence of Coilopoceras, Hoplitoides ingens and H. gibbosulus suggests that the Brazilian beds extend somewhat higher. On the Brazilian evidence, the Ezillo fauna could represent an horizon immediately below the Watinoceras-bearing beds at Lok- panta, its absence there being explained by non-exposure of the relevant beds. Alternatively, since uppermost Cenomanian and basal Turonian faunas appear to be present at both Ezillo and Lokpanta, palaeoenvironmental factors would be responsible. During the late Cenomanian and early Turonian, the Benue Valley and north-eastern part of Nigeria were occupied by an arm of the Tethys extending across the Sahara (see review in Reyment 1980). By late Cenomanian times this epeiric seaway had already flooded the whole of this area. Ammonites such as Metengonoceras dumbli and Euomphaloceras septemseriatum seem to have been introduced from the widening Atlantic Ocean in the south, the former penetrating as far as Damergou in southern Niger (Schneegans 1943, Greigert & Pougnet 1967: 128). The vascoceratids seem to have been introduced from the north. This vast, shallow seaway, in which similar environmental conditions prevailed over great distances, was overwhelmingly populated by vascoceratid ammonites which have few counterparts in southern Nigeria. The area around Ezillo seems to have been close to the southern limit of this faunal province during the late Cenomanian and early Turonian. The most southerly part of Nigeria was, at this time, prob- ably an area of deeper water, subject to greater influence from open oceanic circulation. Its ammonite faunas show their closest affinities with those of the western interior of the United States. Lower Turonian biofacies show similar variations in Iberia, where mammitids dominate the northern part of Spain, vascoceratids characterizing contemporaneous beds to the south (see Wiedmann 1979: text-fig. 6). Similarly, according to Young & Powell (1978), the Tethyan faunas of Mexico and trans-Pecos Texas are absent from central and northern Texas as a result of environmental factors. However, perhaps the situation most closely similar to that in Nigeria occurs in Morocco (see Wiedmann et al. 1982, Einsele & Wiedmann 1982). The Lower Turonian of the Atlas— Meseta Basin in the north of the country is represented by a limestone facies containing familiar Tethyan genera such as Vascoceras, Paravascoceras, Nigericeras and Thomasites. In the west-coastal Tarfaya Basin, on the other hand, a series of laminated bituminous marls with limestone bands and nodules accumulated during the late Cenomanian to early Coniacian. This area lacks both the Turonian vascoceratids and the late Cenomanian Neolobites, a widely distributed genus in Tethyan faunas. Instead, the Upper Cenomanian and Turonian are domi- nated by forms of boreal affinities including Metoicoceras, Watinoceras, Mammites, Benueites and Collignoniceras (see Collignon 1966). Einsele & Wiedmann (1982) interpreted these faunal differences environmentally, believing the ‘black shale’ facies of the Tarfaya Basin to have formed in deeper, cooler waters subject to upwelling from the Atlantic Ocean. There is an obvious parallel between the Lower Turonian of Morocco and of Nigeria in both lithofacies and biofacies, but how far the comparison can be taken is, as yet, uncertain. Petters (1978) 60 P. M. P. ZABORSKI accounted for Cretaceous black shales in the Benue Trough by envisaging a high influx of organic matter into a relatively shallow seaway. During the later part of the Turonian the sea retreated from the interior of Nigeria (Reyment 1980) and beds of this age are less easy to identify. Offodile & Reyment (1976), however, recorded Middle Turonian bivalves along with Collignoniceras, a diagnostic Middle Turonian ammonite (see Kauffman et al. 1978, Cobban & Hook 1979, Kennedy, Wright & Hancock 1980), from Nkalagu in south-east Nigeria. In addition, Reyment (1978) came to regard his ‘Zone of Hoplitoides ingens’ fauna from Wadatta as early Middle Turonian. This fauna is dominated by Hoplitoides and Benueites, which occur with Mammites, Kamerunoceras and Coilopoceras. Beds of similar age in western Cameroun also contain Watinoceras, Neoptychites and Choffaticeras. In Venezuela, Renz (1982) suggested Benueites to be characteristic of the Middle and Upper Turonian and the genus also occupies a relatively high position in the Turonian of the Tarfaya Basin (Collignon 1966). Hoplitoides is confined to the Middle Turon- ian of the western interior, into which Watinoceras persists (Cobban & Hook 1979, 1980). Hoplitoides does appear somewhat earlier in Nigeria but these are apparently transitional forms with persistently broad venters. Kennedy & Wright (1984) doubted that Coilopoceras occurs anywhere earlier than the Middle Turonian. It is probable, therefore, that Reyment (1978) is correct in his Middle Turonian dating of the Wadatta fauna. Coilopoceras (= Glebosoceras Reyment, 1954) has been recorded from several localities in Nigeria south of the Benue River by Reyment (1954, 1955, 1957) and from the top of the Odukpani Formation near Calabar (Zaborski 1985). In view of Kennedy & Wright’s (1984) findings concerning the age of this genus, it is probable that it occurs in Nigeria somewhat higher than the Lower Turonian position previously suggested. Since diagnostic genera of the Upper Turonian such as Pri- onocyclus and Subprionocyclus (see Kauffman et al. 1978, Wright & Kennedy 1981, Hancock & Kennedy 1981) are unproven in Nigeria, it is not yet possible to identify this substage here with any certainty. Offodile & Reyment (1976: 46) believed the Upper Turonian to be absent in the Nkalagu region, Coniacian beds directly overlying the Middle Turonian. Acknowledgements Thanks are due to Dr M.K. Howarth, Dr H.G. Owen and Mr D. Phillips for assistance in many ways. Photographs were provided by the British Museum (Natural History) Photographic Unit. Field work was completed with the help of a University of Ilorin Senate Research Grant. References Amard, B., Collignon, M. & Roman, J. 1981. Etude stratigraphique et paléontologique du Crétacé sup- érieure et Paleocene du Tinrhert-W et Tademait-E (Sahara Algérien). Docums Lab. Geol. Fac. Sci. Lyon (H.S.) 6: 15-173, pls 1-17. Arkhanguelsky, A. D. 1916. [The Upper Cretaceous molluscs of Turkestan. Part 1.] Trudy geol. Kom., St Petersburg, 152: 1-57, pls 1-8. [In Russian]. Atabekyan, A. A. 1966. [New genus Koulabiceras from the Turonian of the eastern parts of Central Asia. ] Izv. Akad. Nauk armyan. SSR, Erevan, 19: 75—78. [In Russian ]. Barber, W. 1957. Lower Turonian ammonites from north-eastern Nigeria. Bull. geol. Surv. Nigeria, Kaduna, 26: 1—86, pls 1-35. Basse, E. 1940. Les céphalopodes crétacés des massifs cOtiers syriens. Pt 1. Notes Mem. Ht.-Comm. Syrie Liban, Paris, 3: 412-490, pls 1-9. Benavides-Caceres, V.E. 1956. Cretaceous system in northern Peru. Bull. Am. Mus. nat. Hist., New York, 108: 353-494, pls 31-66. Bengtson, P. 1983. The Cenomanian—Coniacian of the Sergipe Basin, Brazil. Fossils Strata, Oslo, 12: 1-78, 1 map. Bose, E. 1920. On a new ammonite fauna of the Lower Turonian of Mexico. Bull. Univ. Tex. Bur. econ. Geol. Technol., Austin, 1856: 179-252, pls 12-20. Boule, M., Lemoine, P. & Thevenin, A. 1907. Paléontologie de Madagascar. III—Céphalopodes crétacés des environs de Diégo-Suarez. Annls Paleont., Paris, 2: 1-56, pls 1-8. Buch, L. von 1829. Sur la distribution des ammonites en familles. Annls Sci. nat., Paris, 18: 417-426. NIGERIAN LOWER TURONIAN AMMONITES 61 Cobban, W. A. & Gryc, G. 1961. Ammonites from the Seabee Formation (Cretaceous) of northern Alaska. J. Paleont., Tulsa, 35: 176-190, pls 37, 38. & Hook, S. C. 1979. Collignoniceras woollgari woollgari (Mantell) ammonite fauna from the Upper Cretaceous of Western Interior, United States. Mem. Inst. Min. Technol. New Mex., Socorro, 37: 1-51, pls 1-12. ——-—— 1980. The Upper Cretaceous ammonite family Coilopoceratidae Hyatt in the Western Interior of the United States. Prof. Pap. U.S. geol. Surv., Washington, 1192: 1—28, pls 1-21. —— —— 1981. An unusually large specimen of the Turonian ammonite Hoplitoides von Koenen from New Mexico. Circ. New Mex. St. Bur. Mines Miner. Resour., Socorro, 180: 30—34, pl. 5. —— —— 1983. Mid-Cretaceous (Turonian) ammonite fauna from Fence Lake area of west-central New Mexico. Mem. Inst. Min. Technol. New Mex., Socorro, 41: 1—S0, pls 1-14. — & Scott, G. R. 1972. Stratigraphy and ammonite fauna of the Graneros Shale and Greenhorn Limestone near Pueblo, Colorado. Prof. Pap. U.S. geol. Surv., Washington, 645: 1-108, pls 1-41. Collignon, M. 1961. Ammonites néocrétacées du Menabe (Madagascar). VII. Les Desmoceratidae. Annls geol. Serv. Mines Madagascar, Tananarive, 31: 1-115, pls 1-32. 1964-65. Atlas des fossiles caracteristiques de Madagascar (Ammonites). XI (Cénomanien). xi + 152 pp., pls 318-375 (1964). XII (Turonien). iv + 82 pp., pls 376-413 (1965). XIII (Coniacien). vil + 88 pp., pls 414454 (1965a). Service géologique, Tananarive. —— 1966. Les céphalopodes crétacés du bassin cotier de Tarfaya. Notes Mem. Serv. géol. Maroc, Rabat, 175 (2): 7-148, pls 1-35. Cooper, M. R. 1978. Uppermost Cenomanian—basal Turonian ammonites from Salinas, Angola. Ann. S. Afr. Mus., Cape Town, 75: 51-152. Courtiller, M. A. 1860. Description de trois nouvelles espéces d’ammonites du terrain créetacé des environs du Saumur. Mem. Soc. imp. Agric. Sci. Angers, 3: 246-252, 3 pls. Douville, H. 1904. Paléontologie, mollusques fossiles. In: Morgan, J. de, Mission scientifique en Perse 3 (4): 191—380, pls 25—SO. Paris. — 1912. Evolution et classification des Pulchelliidés. Bull. Soc. géol. Fr., Paris, (4) 11: 285-320. —— 1928. Les ammonites de la Craie supérieure en Egypte et au Sinai. Mem. Acad. Sci. Inst. Fr., Paris, (2) 60: 1-44, pls 1-7. Eck, O. 1909. Bemerkungen tiber drei neue Ammoniten aus der oberen egyptischen Kreide. Sher. Ges. naturf. Freunde Berl., 1909 (3): 179-191. Einsele, G. & Wiedmann, J. 1982. Turonian black shales in the Moroccan coastal basins: first upwelling in the Atlantic Ocean?. In: Rad, U. von, Hinz, K., Sarntheim, M. & Siebold, E. (eds), Geology of the northwest African continental margin: 396-414. Berlin. Etayo-Serna, F. 1979. Zonation of the Cretaceous of central Colombia by ammonites. Publnes geol. esp. Ingeominas, Bogota, 2: 1-188, pls 1-15. (D.Phil. Dissertation, Univ. Calif., Berkeley, 1975). Faraud, M. 1951. La famille des Tissotiidae dans le Turonien du Gard. Bull. Soc. geol. Fr., Paris, (6) 1: 148-157, pl. Sa. Freund, R. & Raab, M. 1969. Lower Turonian ammonites from Israel. Spec. Pap. Palaeont., London, 4: 1-83, pls 1-10. Greigert, J. & Pougnet, R. 1967. Essai de description des formations géologiques de la republique du Niger. Mem. Bur. Rech. géol. minier., Paris, 48: 1-238. Grossouvre, A. de 1894. Recherches sur la Craie Supérieure. 2, Paléontologie. Les ammonites de la Craie Supérieure. 264 pp., atlas of 39 pls. Mem. Serv. Carte det. géol. Fr., Paris. Hancock, J. M. & Kennedy, W. J. 1981. Upper Cretaceous ammonite stratigraphy: some current prob- lems. In: House, M. R. & Senior, J. R. (eds), The Ammonoidea: 531-553. London (Academic Press, for the Systematics Association). Herrick, C. L. & Johnson, D. W. 1900. The geology of the Albuquerque sheet. Bull. scient. Labs Denison Univ., Granville, Ohio, 11: 175—239, pls 27-32. Hyatt, A. 1900. Cephalopoda. In: Zittel, K. A. von (transl. Eastman, C. R.), Textbook of Palaeontology, 1: 502-604. London. —— 1903. Pseudoceratites of the Cretaceous. Monogr. U.S. geol. Surv., Washington, 44: 1-351, pls 1-47. Karrenberg, H. 1935. Ammonitenfaunen aus der Nordspanischen Oberkreide. Palaeontographica, Stutt- gart, (A) 82: 125-161, pls 30-33. Kauffman, E. G., Cobban, W. A. & Eicher, D. L. 1978. Albian through Lower Coniacian strata, bio- stratigraphy and principal events, Western Interior, United States. Annis Mus. Hist. nat. Nice 4 (XXIII): 1-52, pls 1-17. Kennedy, W. J. 1971. Cenomanian ammonites from southern England. Spec. Pap. Palaeont., London, 8: 1-133, pls 1-64. 62 P. M. P. ZABORSKI ——, Cooper, M. R. & Wright, C. W. 1979. On Ammonites gallienni d’Orbigny, 1850. Bull. geol. Instn Univ. Uppsala, (n.s.) 8: 5-15. , Juignet, P. & Hancock, J. M. 1981. Upper Cenomanian ammonites from Anjou and the Vendée, western France. Palaeontology, London, 24: 25—84, pls 3-17. —— & Wright, C. W. 1979. On Kamerunoceras Reyment, 1954 (Cretaceous Ammonoidea). J. Paleont., Tulsa, 53: 1165-1178, 4 pls. ——-—— 1979a. Vascoceratid ammonites from the type Turonian. Palaeontology, London, 22: 665-683, pls 82-86. —— —— 1984. The Cretaceous ammonite Ammonites requienianus d’Orbigny, 1841. Palaeontology, London, 27: 281-293, pls 35—37. —., & Hancock, J. M. 1980. Collignoniceratid ammonites from the mid-Turonian of England and northern France. Palaeontology, London, 23: 557-603, pls 62-77. Kler, M. O. 1909. [Neoceratites of eastern Bukhara.] Trudy geol. miner. Muz., St Petersburg, 2 (for 1908): 157-174, pls 6-8. [In Russian]. Koenen, A. von 1897-98. Ueber Fossilien der unteren Kreide am Ufer des Mungo in Kamerun. Abh. K. Ges. Wiss. Gottingen, (N.R.) 1: 1-48, 4 pls (1897). Nachtrag. loc. cit.: 51-65, 3 pls (1898). Kossmat, F. 1895-98. Untersuchungen tber die Stidindischen Kreideformation. Beitr. Palaont. Geol. Ost.-Ung., Vienna and Leipzig, 9: 97-203, pls 15-25 (1895); 11: 1-46, pls 1-8 (1897); 12: 89-152, pls 14-19 (1898). Kummel, B. & Decker, J. M. 1954. Lower Turonian ammonites from Texas and Mexico. J. Paleont., Tulsa, 28: 310-319, pls 30-33. Laube, G. C. & Bruder, G. 1887. Ammoniten der bohmischen Kreide. Palaeontographica, Stuttgart, 33: 217-239, pls 23-29. Leanza, A. F. 1967. Algunos ammonites nuevos o poco conocidos del Turoniano del Colombia y Vene- zuela. Acta geol. lilloana, Tucuman, 9: 189-229, pls 1—7. Matsumoto, T. 1954. Family Puzosiidae from Hokkaido and Saghalien. Mem. Fac. Sci. Kyushu Univ., Fukuoka, (D) 5: 69-118, pls 9-32. Offodile, M. E. 1976. The geology of the Middle Benue, Nigeria. Spec. Vol. palaeont. Inst. Univ. Uppsala, 4: 1-166, pls 1—20. —— & Reyment, R. A. 1976. Stratigraphy of the Keana—Awe area of the middle Benue region of Nigeria. Bull. geol. Instn Univ. Uppsala, (n.s.) 7: 37—66, figs 1-41. Orlov, Y. A. (ed.) 1958. [Cephalopoda 2. Ammonoidea (Ceratitida and Ammonitida), Endoconchlia, Conioconchia.] Osnovy Paleontologii, 6. 333 pp., 71 pls. Moscow. [In Russian]. Peron, A. 1889-93. Description des invertebres fossiles des terrains crétacés de la région sud des Hauts- Plateaux de la Tunisie recuellis en 1885 et 1886 par M. Phillipe Thomas. In: Exploration scientifique de la Tunisie. 405 pp., Atlas 35 pls. Paris. 1896. Les ammonites du Crétacé supérieure de Algérie. Mem. Soc. geol. Fr. Paleont., Paris, 17: 1-88, pls 1-18. Pervinquiere, L. 1907. Carte géologique de la Tunisie. Etudes de paléontologie tunisienne, 1. Céphalopodes des terrains secondaires. 438 pp., 27 pls. Paris. Petters, S. W. 1978. Mid-Cretaceous paleoenvironments and biostratigraphy of the Benue Trough, Nigeria. Bull. geol. Soc. Am., New York, 89: 151-154. & Ekweozor, C. M. 1982. Petroleum geology of Benue Trough and southeastern Chad Basin, Nigeria. Bull. Am. Ass. Petrol. Geol., Tulsa, 66 (8): 1141-1149. Popovici-Hatzeg, V. 1899. Contribution a l’etude de la fauna du Crétacé supérieur de Roumanie. Mem. Soc. geéol. Fr. Paleont., Paris, 20: 1—20, pls 1, 2. Powell, J. D. 1963. Cenomanian—Turonian (Cretaceous) ammonites from Trans-Pecos, Texas and north- eastern Chihuahua, Mexico. J. Paleont., Tulsa, 37: 309-322, pls 31-34. Renz, O. 1982. The Cretaceous ammonites of Venezuela. 132 pp., 40 pls. Basle. Reyment, R. A. 1954. New Turonian (Cretaceous) ammonite genera from Nigeria. Colon. Geol. Miner. Resour., London, 4: 149-164, pls 1-4. —— 1954a. Some new Cretaceous ammonites from Nigeria. Colon. Geol. Miner. Resour., London, 4: 248-270, pls 1-5. —— 1955. The Cretaceous Ammonoidea of southern Nigeria and the southern Cameroons. Bull. geol. Surv. Nigeria, Kaduna, 25: 1-112, pls 1—25. 1956. On the stratigraphy and palaeontology of the Cretaceous of Nigeria and the Cameroons, British West Africa. Geol. For. Stockh. Forh., 78: 17-96. — 1957. Uber einige wirbellose Fossilien aus Nigerien und Kamerun, Westafrika. Palaeontographica, Stuttgart, (A) 109: 41—70, pls 7-11. NIGERIAN LOWER TURONIAN AMMONITES 63 —— 1958. Ubersichtliche erganzung von F. Solger’s “Die Fossilien der Mungokreide in Kamerun und ihre geologische Bedeutung” (1904). Stockh. Contr. Geol., (n.s.) 2: 51-72, pls 1-7. —— 1965. Aspects of the geology of Nigeria. 145 pp., 18 pls. Ibadan. — 1971. Vermuteter Dimorphismus bei der Ammonitengattung Benueites. Bull. geol. Instn Univ. Uppsala, (n.s.) 3: 1-18, pls 1-10. — 1972. Some Lower Turonian ammonites from Trinidad and Colombia. Geol. For Stockh. Forh., 94: 357-368. —— 1978. The mid-Cretaceous of the Nigerian coastal basin. Annls Mus. Hist. nat. Nice, 4 (XX): 1-13. —— 1979. Variation and ontogeny in Bauchioceras and Gombeoceras. Bull. geol. Instn Univ. Uppsala, (n.s.) 8: 89-111. 1980. Biogeography of the Saharan Cretaceous and Paleocene epicontinental transgressions. Cret. Res., London, 1: 299-327. Riedel, L. 1932. Die Oberkreide von Mungofluss in Kamerun und ihre Fauna. Beitr. geol. Erforsch. dt. Schutzgeb., Berlin, 16: 1-154, pls 1-32. Schliiter, C. 1871. Die Cephalopoden der oberen deutschen Kreide. Palaeontographica, Stuttgart, 21: 1-24, pls 1-18. Schneegans, D. 1943. Invertébres du Crétacé supérieur du Damergou (Territoire du Niger). Bull. Dir. Mines Afr. occid. fr., Dakar, 7: 87-150, 8 pls. Simpson, A. 1954. The Nigerian coalfield. The geology of parts of Onitsha, Owerri and Benue Provinces. Bull. geol. Surv. Nigeria, Kaduna, 24: 1-85, pls 1-5. Solger, F. 1904. Die Fossilien der Mungokreide in Kamerun und ihre geologische Bedeutung, mit besonderer Berticksichtigung der Ammonitiden. In: Esch, E., Solger, F., Oppenheim, M. & Jakel, O., Beitrage zur Geologie von Kamerun: 83-242, pls 3—S. Stuttgart. Spath, L. F. 1922. On the Senonian ammonite fauna from Pondoland. Trans. R. Soc. S. Afr., Cape Town, 10: 113-147, pls 5-9. —— 1925. On Upper Albian Ammonoidea from Portugese East Africa. With an appendix on Upper Cretaceous ammonites from Maputoland. Ann. Transv. Mus., Pretoria, 11: 179-200, pls 28-37. Stankievich, E. S. & Pojarkova, Z. N. 1969. [Vascoceratids from the Turonian of southern Kirgisia and the Tadzhiksian depression.] In: Barkhatova, N. N. (ed.), [Continental formations of eastern regions of Soviet Central Asia and Kazakhstan]: 86—113, pls 1-10. Leningrad. [In Russian]. Stoliczka, F. 1865. The fossil Cephalopoda of the Cretaceous rocks of southern India. Mem. geol. Surv. India Palaeont. indica, Calcutta, (3) 6-9: 107—154, pls 55-80. Warren, P. S. 1930. New species of fossils from Smoky River and Dunvegan Formations, Alberta. Rep. Res. Coun. Alberta, Edmonton, 21: 57-58, pls 3-7. Wiedmann, J. 1960. Le Crétacé supérieur de Espagne et du Portugal et ses céphalopodes. In: Colloque sur le Crétacé Supérieur Frangais (Dijon 1959). C.r. Congr. Socs sav. Paris Sect. Sci., 1959: 709-764, 8 pls. —— 1975. El Cretacico superior del Picofrentes (Soria), Cadenas Celtibéricas (Espana). Boln geol. min., Madrid, 86 (3): 252-261. —— 1979. Itineraire géologique a travers le Crétacé moyen des chaines vascogotiques et celtibériques (Espagne du nord). Cuad. Geol. iber., Madrid, 5: 127-214, pls 1-12. ——, Butt, A. & Einsele, G. 1982. Cretaceous stratigraphy, environment and subsidence history at the Meeecan continental margin. In: Rad, U. von, Hinz, K., Sarntheim, M. & Siebold, E. (eds), Geology of the northwest African continental margin: 366-395. Berlin. & Kauffman, E. G. 1978. Mid-Cretaceous biostratigraphy of northern Spain. Annis Mus. Hist. nat. Nice, 4 (III): 1-34, pls 1-12. Woods, H. 1911. The palaeontology of the Upper Cretaceous deposits of Northern Nigeria. Jn: Falconer, J. D., The geology and geography of Northern Nigeria: 273-286. London. Wozny, E. & Kogbe, C. A. 1983. Further evidence of marine Cenomanian, Lower Turonian and Maas- trichtian in the Upper Benue Basin of Nigeria (west Africa). Cret. Res., London, 4: 95-99. Wright, C. W. & Kennedy, W. J. 1981. The Ammonoidea of the Plenus Marls and the Middle Chalk. 148 pp., 32 pls. Palaeontogr. Soc. (Monogr.), London. Yabe, H. 1914. Ein neuer Ammonitenfund aus der Trigoniasandstein-Gruppe von Provinz Tosa. Sci. Rep. Tohoku Univ., Sendai, (2) 1: 71-74, pl. 12. Young, K. & Powell, J. D. 1978. Late Albian—Turonian correlations in Texas and Mexico. Annls Mus. Hist. nat. Nice, 4 (XXV): 1-36, pls 1-9. Zaborski, P. M. P. 1985. Upper Cretaceous ammonites from the Calabar region, south-east Nigeria. Bull. Br. Mus. nat. Hist., London, (Geol.) 39 (1): 1-72, 66 figs. Zimmerman, E. 1912. Puzosia Kauffi n. sp., Puzosia Denisoniana Stol. in der Oberen Kreide Nord- 64 P. M. P. ZABORSKI deutschlands und die Loben der bischer bekannten Puzosia-Arten. Jb. preuss. geol. Landsanst. Berg- Akad., Berlin, 33 (1): 533-556, pls 25, 26. Zittel, K. A. von 1895. Grundziige der Palaeontologie (Palaeozoologie). vii + 972 pp. Munich and Leipzig. Index New taxonomic names and the page numbers of the principal references are in bold type. An asterisk (*) denotes a figure. Abakaliki 32* Colombia 41, 43, 49, 51 Abakaliki-Enugu highway 32*, 33 Colorado 37, 45 Aboine, see Western Aboine River Coniacian 34, 59-60 Acanthoceras amudariense 37 eschii 36 Acanthocerataceae 36 a ee 1 COSINE SEB LE 36 Desmoceras Kamerunense 34 Albian 32 ee (Puzosia) Denisonianum 34 Allocrioceras 32*, 58 Desmocerataceae, Desmoceratidae 33-6 pee 58 dimorphism 40, 45, 49 Ammonites cephalotus 43 denisonianus 33-4 Egypt 54 gallienni 49 England, southern 57 dosoides 40 Enugu 32* nodosoides telinga 43 Enugu—Port Harcourt expressway 32*, 33 xetra 43 Enugu—Okigwe road 32* Ashaka 32*, 57-8 Eotissotia 58* Asu River Group 32* Euomphaloceras 37, 58* Atlantic Ocean 59 euomphalum 37 Atlas-Meseta Basin 59 septemseriatum 37, 57-9 Awgu 32* Euomphaloceratinae 36 Europe 58; see also England, France, Iberia, Bauchioceras 49 Spain nigeriensis, see Pseudotissotia north-west 57 (Wrightoceras) wallsi 49, 51 Eze-Aku Formation 32*, 34, 36-8, 40-1, 45, 47-8, Benue River 32*, 48, 58*, 60 _ 51, 54, 56 Trough 58*, 59 Ezillo 32*, 33, 47-8, 51—2, 57, 58*, 59 Valley 47, 57-9 Ezilloella 32*, 58* Benueites 40, 57, 59-60 ezilloensis 58 trinidadensis 40 Boreal faunas 57, 59 Fagesia 32*, 41-3, 58*, 59 Brazil 53, 59; see also Sergipe Basin bomba 43 British Museum (Natural History) 33, 60 boucheroni 43 catinus 43 Calabar 32*, 58*, 60 involuta 43 Cameroun 34, 36, 60 lenticularis 43 Cenomanian 32*, 34, 57 levis 33, 41-3, 42*, 44* late 33, 37, 57, 59 pachydiscoides 43 Upper 32*, 33, 56*, 57-9 peroni 43 uppermost 32, 47-8, 57, 59 rudra 43 Cenomanian—Turonian boundary 58-9 simplex 43 Chad Basin 58* superstes 43 Choffaticeras 53, 57, 60 thevestensis 43 Coilopoceras 59-60 zanelli 43 Coilopoceratidae 52-6 sp. 33 Collignoniceras 59-60 Ferganites, see Gombeoceras woollgari, Zone of 53 France 45; see also Touraine woollgari woollgari, Subzone of 56 Franciscoites suarezi 45 NIGERIAN LOWER TURONIAN AMMONITES 65 Glebosoceras 60 Gombe 51 Gombeoceras 45, 47 compressum 47 gongilense 47-9 koulabicum 48 subtenue 47 (Ferganites) 49 kanicum 48-9 kleri 48-9 koulabicum 48-9 Herrickiceras 56, 58* Herrickiceras? 33 sp. 44*, 56 Hoplitoides 49, 51, 52-5, 57*, 58*, 60 crassiornatus 53—4 gibbosulus 53-4, 59 hernanmojicae 54 ingens 53-4, 59; Zone of 56*, 57, 60 var. levis 54 koeneni 53-4 lacabagnae 54 lagiraldae 54 latefundatus 33, 50*, 53—5* latesellatus 52 mirabilis 49, 51-2, 54 munieri 49, 51—2; cf. munieri 49, 51-2 sandovalensis 53 wohltmanni 53-4 Howarth, Dr M. K. 60 Iberia 59 Ilorin, University of 33, 60 Imlayiceras 49 washbournei 49 Intermediate Beds 56*, 57 Israel 53; see also Negev Kanabiceras 37 puebloense 36 Kanawa 51 Kamerunoceras 36-7, 57, 58*, 60 eschii 36; Zone of 56*, 57; cf. eschii 33, 35*, 36, 37 puebloense 33, 36-7, 39*, 58 schindewolfi 37 seitzi 36 Konshisha River 48, 58* Koulabiceras 49 koulabicum 48 Las Fuentes 53 Lokpanta 32*-4, 36-8, 40-1, 45, 53-4, 56, 58*, 59 Makurdi 53 Mammites 40-1, 57, 58*—60 dixeyi 41, 59 nodosoides 33, 39*, 40-1, 42*, 45, 59: fauna 32*: Zone of 34, 36-7, 40-1, 45, 54, 56*, 57-8 nodosoidesappelatus 40 wingl Mammites? 41, 59 sp. 33, 39*, 41 Mammitinae 37-41 Metengonoceras 32*, 33, 39*, 57, 58* dumbli 57, 59 Metoicoceras 58*, 59 Mexico 49, 51—2, 59 Morocco 59; see also Atlas—Meseta Basin, Tarfaya Basin Mortoniceras 32* Ndieze 32* Negev 52 Neolobites 59 Neoptychites 43, 45, 57, 58*, 60 cephalotus 33, 43, 45, 46*, 59 crassus 45 ingens 52 telinga 43 telingaeformis discrepans 45 transitorius 45 xetra 43 xetriformis 45 New Mexico 53-4, 56 Niger 59 Niger River 32*, 58* Nigeria, north-east 47, 49, 51, 57-9; see also Ashaka, Deba-Habe, Gombe, Kanawa, Numan, Pindiga central 57 Nigericeras 58*, 59 Nkalagu 32*, 60 Numan 47 Odukpani Formation 60 Okigwe 32* Olcostephanus superstes 41 Oldest Beds 56*, 57 Oturkpo 48 Owen, Dr H. G. 60 Pachydesmoceras 33-6, 58*, 59 denisoni 34 denisonianum 33, 34-6, 35* hourcqui 34 kamerunense 34 linderi 36 maroccanum 36 pachydiscoides 34 radaodyi 36 rareccstatum 36 Pachyvascoceras costatum, Zone of 56*, 57 Paramammites 58* Paravascoceras 32*—3, 57, 58*, 59 costatum, Zone of 56*, 57 66 P. M. P. ZABORSKI Phillips, D. 60 Picofrentes 53 Pindiga 32*, 57-8 Placenticeras costatum 56 Prionocyclus 60 hyatti, Zone of 53 Pseudaspidoceras 58* Pseudocalycoceras 58* Pseudotissotia 49, 57—8* gagnieri 49 koulabica 48-9 nigeriensis, Zone of 56*, 57 (Wrightoceras) wallsi 51; Zone of 56*, 57 Pseudotissotiinae 45—52 Puentedai 53 Puzosia alimanestianui 34 Denisoni, denisoniana 34 Romaniceras 57 uchauxiense, Zone of 56*, 57 Sahara 59 Algerian 53 Schlutericeras nodosoides 40 Sergipe Basin 57, 59 Soviet Central Asia 48 Soviet Union 49 Spain 53, 59; see also Las Fuentes, Picofrentes, Puentedai Subprionocyclus 60 Tarfaya Basin 38, 59-60 Tethyan faunas 57, 59 Texas 59 Thomasites 32*, 45—9, 57, 58*, 59 cf. globosotuberculatus 49 gongilensis 33, 44*, 47-8, 50*, 57-8 var. costatus 48 var. compressus 47-8 var. crassiornatus 48 var. inflatus 48 var. tectiformis 48 inflatus 49 cf. jordani 49 koulabicus 33, 46*, 48-9 meslei 47 rollandi 47 = Touraine 45 Tunisia 47-9 Turonian 32, 57—60 basal 36-8, 41, 45, 59 early 33, 53, 58-9 Lower 32*, 33-4, 36-7, 40-1, 45, 51-4, 56*, 57, 59-60 lowermost 47-8 Middle 33, 53-4, 56-7, 60 Upper 56*, 57, 60 United States, see western interior Vascoceras 57, 58*, 59 bulbosum, Zone of 56*, 57 gongilense 45, 47 Vascoceratidae 41—52 Vascoceratinae 41—5 Venezuela 40, 43, 52-4, 60 Wadatta 53, 60 W atinoceras 37—40, 57—8*, 59-60; beds with 56*; fauna 32* amudariense 38, 40, 59; aff. amudariense 33, 37-8, 39*40 coloradoense 40; Zone of 45, 57-8 praecursor 40 devonense 40 reesidei 37-8; sp. aff. reesidei 38 sp. 33, 38-40 Western Aboine River 32*, 58—9 western interior, United States 45, 53, 57—60; see also Colorado, New Mexico Wrightoceras 32*, 49-52, 53—4, 57, 58*, 59 inca 49, 52 llarenai 49, 52 mirabilis 49, 52 munieri 49, 52-3; cf. munieri 33, 50*, 51—2*, 53 reymenti 49, 52 submunieri 49, 52 wallsi 33, 49, 50*, 51-3, 58 Wrightoceras? gagnieri 52 Youngest Beds 56*, 57 Printed in Great Britain at the University Printing House, Oxford Accepted for publication 18 November 1985 Bulletin of the British Museum (Natural History) a Geology Series Most earlier Geology Bulletins are still in print. A full list of available titles can be obtained from Publications Sales aa (address inside front cover). , - Vol. 29 No.1 Aspects of mid-Cretaceous stratigraphical micropalaeontology. D. J. Carter & M. B. Hart. 1977. Pp. ___— 1-135, 4 plates, 53 figs. £14.25 _ Vol. 29 No.2 The Macrosemiidae, a Mesozoic family of holostean fishes. A. W. H. Bartram. 1977. Pp. 137-234, 4 plates, 53 figs. £10.00 Vol. 29 No.3 The stratigraphy and ammonite fauna of the Upper Lias of Northamptonshire. M. K. Howarth. 1978. Pp. 235-288, 9 plates, 5 figs. £6.00 .29 No.4 Fossil Bovidae (Mammalia) of Olduvai Gorge, Tanzania, Part I. A. W. Gentry & A. Gentry. 1978. Pp. ) 89-446, 41 plates, 34 figs. £17.50 . Vol. 30 No.1 Fossil Bovidae (Mammalia) of Olduvai Gorge, Tanzania. Part II. A. W. Gentry & A. Gentry. 1978. Pp. ‘1-83, 3 figs. £7.50 Vol. 30 No. 2 A revision of the Miocene Hominoidea of East Africa. P. J. Andrews. 1978. Pp. 85-224, 7 plates, 29 figs. £15.30 Vol. 30 No.3 Early Ordovician (Arenig) stratigraphy and faunas of the Carmarthen district, south-west Wales. R. A. ot M. Owens. 1978. Pp. 225-296, 11 plates, 12 figs. £9.60 lol. 30 No. 4 Macroscopic inclusions of fluid in British fluorites from the mineral collection of the British Museum (Natural History). A. H. Rankin. 1978. Pp. 297-307, coloured frontispiece, 9 plates (7 coloured), 4 figs. £12.00 Il. 31 No. 1 Foraminifera of the Togopi Formation, eastern Sabah, Malaysia. J. E. Whittaker & R. L. Hodgkinson. 979. Pp. 1—120, 10 plates, 71 figs. £14.00 Vol. 31 No.2 Cretaceous faunas from Zululand and Natal, South Africa. The ammonite family Gaudryceratidae. _ W.J. Kennedy & H.C. Klinger. 1979. Pp. 121-173. £6.25 31 No.3 Benthic community organization in the Ludlow Series of the Welsh Borderland. R. Watkins. 1979. Pp. 75-279. mn A229 31 No.4 The ammonites of the English Chalk Rock (Upper Turonian). C. W. Wright. 1979. Pp. 281-330. £6.50 | 32. No.1 Miscellanea: Observations on Cycloclypeus—Provenance of Sivapithecus—Iranian Silurian —b achiopods—New English condylarths—Miocene sharks’ teeth—East African isopod—The Singa skull— : ‘boniferous insects. 1979. Pp. 1-90. £10.50 No.2 Palaeoenvironments and correlations of the Carboniferous rocks in west Fermanagh, Ireland. C. H. C. inton & T. R. Mason. 1979. Pp. 91-108, 6 figs, folded map. £4.00 ; 32 No.3 The Ordovician trilobite faunas of the Builth-Llandrindod Inlier, central Wales. Part III. C. P. Hughes. " s 979. Pp. 109-181, 177 figs. £10.00 Vol. 32 No.4 The stratigraphy and brachiopods of the upper part of the type Caradoc of south Salop. J. M. Hurst. 1979. Pp. 183-304, 557 figs. £18.50 Vol. -33.No.1 An account of the Ordovician rocks of the Shelve Inlier in west Salop and part of north Powys. W. F. ittard, F. R. S. (Compiled by W. T. Dean). 1979. Pp. 1-69, 38 figs, frontispiece, coloured map, folded, in pocket. £10.00 ap available separately £1.00 No.2 Miscellanea: Lower Carboniferous microproblematicum—Miniature trilobite—Pleistocene bird ins—English Eocene H yracotherium—Salenia trisuranalis—Antarctic brachiopods—Diphyphyllum and Murchi- son’s Russian corals—Lebanese amber Neuroptera. 1980. Pp. 71-164. £12.00 33 No.3 The Caradoc faunal associations of the area between Bala and Dinas Mawddwy, north Wales. M. G. ockley. 1980. Pp. 165-235, 105 figs. £9.00 I. 33 No.4 Fossil insects from the Bembridge Marls, Palaeogene of the Isle of Wight, southern England. E. A. jarzembowski. 1980. Pp. 237-293, 77 figs. £7.50 Vol. 33 No.5 The Yorkshire Jurassic fern Phlebopteris braunii (Goeppert) and its reference to Matonia R.Br. T. M. es Harris. 1980. Pp. 295-311, 20 figs. £2.75 \ 1. 34 No.1 Relative dating of the fossil hominids of Europe. K. P. Oakley. 1980. Pp. 1-63, 6 figs, 17 tables. £8.00 Vol. 34. No.2 Origin, evolution and systematics of the dwarf Acanthoceratid Protacanthoceras Spath, 1923 (Cretaceous Ammonoidea). C. W. Wright & W. J. Kennedy. 1980. Pp. 65-107, 61 figs. £6.25 y. ol. 34 No.3 Ashgill Brachiopoda from the Glyn Ceiriog District, north Wales. N. Hiller. 1980. Pp. 109-216, 408 ¥ figs £14.75 Vol} 34 No.4 Miscellanea: Upper Palaeozoic Athyrididae brachiopods—New British Cretaceous Epitoniidae— _ Microproblematicum Prethocoprolithus—Glabellar structure of asaphid trilobites—New Lower Ordovician bivalve _ family—Cretaceous brachiopods—T upus diluculum sp. nov.—Revision of Plummerita. 1980. Pp. 217-297. £11.00 a Vol. 35 No.1 Lower Ordovician Brachiopoda from mid and south-west Wales. M. G. Lockley & A. Williams. 1981. Pp. 1-78, 263 figs, 3 tables. £10.80 — Vol. 35 No.2 The fossil alga Girvanella Nicholson & Etheridge. H. M. C. Danielli. 1981. Pp. 79-107, 8 figs, 3 tables. £4.20 Vol. 35 No. 3 Centenary Miscellanea: Budleigh Salterton brachiopods—Oswald’s Turkish algae—J. A. Moy- Thomas—Burials, bodies and beheadings—Nucleolites clunicularis—Phanerotinus cristatus—Fossil record of teleosts—Neanderthal dating—Hippoporidra edax. 1981. Pp. 109-252. £20.00 Vol. 35 No.4 The English Upper Jurassic Plesiosauroidea (Reptilia) and a review of the phylogeny and classification of the Plesiosauria. D. S. Brown. 1981. Pp. 253-347, 44 figs. £13.00 Vol. 36 No. 1 Middle Cambrian trilobites from the Sosink Formation, Derik—Mardin district, south-eastern Turkey. W. T. Dean. 1982. Pp. 1-41, 68 figs. £5.80 Vol. 36 No. 2. Miscellanea: Dinantian terebratulids—New microfossils—Neseuretus—Archaeocidaris whatleyensis— Carboniferous dasyclad—Nanjinoporella—Toarcian bryozoans—Drybrook Sandstone plants—British fossil bintoniellids—Uraloporella. 1982. Pp. 43-155. £19.80 Vol. 36 No. 3. The Ordovician Graptolites of Spitsbergen. R. A. Cooper & R. A. Fortey. 1982. Pp. 157-302, 6 plates, 83 figs, 2 tables. £20.50 Vol. 36 No.4 Campanian and Maastrichtian sphenodiscid ammonites from southern Nigeria. P. M. P. Zaborski. 1982. Pp. 303-332, 36 figs. £4.00 Vol. 37 No. 1 Taxonomy of the arthrodire Phlyctaenius from the Lower or Middle Devonian of Campbellton, New Brunswick, Canada. V. T. Young. 1983. Pp. 1-35, 18 figs. £5.00 Vol. 37 No. 2. Ailsacrinus gen. nov., an aberrant millericrinid from the Middle Jurassic of Britain. P. D. Taylor. 1983. Pp. 37-77, 48 figs, 1 table. £5.90 Vol. 37 No. 3. Miscellanea: Permian Glossopteris in Turkey—Wealden Theriosuchus—Wealden conifer—Permian plants of Saudi Arabia—Carboniferous Edrioasteroidea—British cicadas—Dittonian cephalaspids. 1983. Pp. 79-171. £13.50 Vol. 37 No. 4 The relationships of the palaeoniscid fishes, a review based on new specimens of Mimia and Moy- thomasia from the Upper Devonian of Western Australia. B. G. Gardiner. 1984. Pp. 173-428, 145 figs, 4 plates. 0 565 00967 2. £39.00 Vol. 38 No. 1 New tertiary pycnodonts from the Tilemsi valley, Republic of Mali. A. E. Longbottom. 1984. Pp. 1-26, 29 figs, 3 tables. 0 565 07000 2. £3.90 Vol. 38 No. 2 Silicified brachiopods from the Viséan of County Fermanagh, Ireland. (III) Rhynchonellids, Spiriferids and Terebratulids. C. H. C. Brunton. 1984. Pp. 27-130, 213 figs. 0 565 07001 0. £16.20 Vol. 38 No. 3. The Llandovery Series of the Type Area. L. R. M. Cocks, N. H. Woodcock, R. B. Rickards, J. T. Temple & P. D. Lane. 1984. Pp. 131-182, 70 figs. 0 565 07004 5. £7.80 Vol. 38 No. 4 Lower Ordovician Brachiopoda from the Tourmakeady Limestone, Co. Mayo, Ireland. A. Williams & G. B. Curry. 1985. Pp. 183-269, 214 figs. 0 565 07003 7. £14.50 Vol. 38 No. 5 Miscellanea: Productacean growth and shell shape—Jurassic alga Palaeosiphonium—Upper Ordovi- cian brachiopods and trilobites—Lower Devonian Osteostraci from Podolia—Hipparion from Diavata—Preparation and study of Singa skull—Carboniferous and Permian bryozoa—Lower Eocene trionychid—Montsech fossil insects. 1985. Pp. 271-412. 0 565 07004 S. £24.00 Vol. 39 No. 1 Upper Cretaceous ammonites from the Calabar region, south-east Nigeria. P. M. P. Zaborski. 1985. Pp. 1-72, 66 figs. 0 565 07006 1. £11.00 Vol. 39 No. 2 Cenomanian and Turonian ammonites from the Novo Redondo area, Angola. M. K. Howarth. 1985. Pp. 73-105. 33 figs. 0 565 07006 1. £5.60 Vol. 39 No. 3. The systematics and palaeogeography of the Lower Jurassic insects of Dorset, England. P. E. S. Whalley. 1985. Pp. 107-189, 87 figs, 2 tables. 0 565 07008 8. £14.00 Vol. 39 No.4 Mammals from the Bartonian (middle/late Eocene) of the Hampshire Basin, southern England. J. J. Hooker. 1986. Pp. 191-478, 71 figs, 39 tables. 0 565 07009 6. £49.50 Vol. 40 No. 1 The Ordovician graptolites of the Shelve District, Sinan I. Strachan. 1986. Pp. 1-58, 38 figs. 0 565 07010 X. £9.00 Vol. 40 No. 2. The Cretaceous echinoid Boletechinus, with notes on the phylogeny of the Glyphocyphidae and Tem- nopleuridae. D. N. Lewis. 1986. Pp. 59-90, 11 figs, 7 tables. 0 565 07011 8. £5.60 Vol. 40 No. 3. The trilobite fauna of the Raheen Formation (upper Caradoc), Co. Waterford, Ireland. A. W. Owen, R. P. Tripp & S. F. Morris. 1986. Pp. 91-122, 88 figs. 0 565 07012 6. £5.60 Vol. 40 No. 4 Miscellanea I: Lower Turonian cirripede—Indian coleoid Naefia—Cretaceous—Recent Craniidae— Lectotypes of Girvan trilobites—Brachiopods from Provence—Lower Cretaceous cheilostomes. 1986. Pp. 125-222. 0 565 07013 4. ‘ £19.00 Vol. 40 No. 5 Miscellanea II: New material of Kimmerosaurus—Edgehills Sandstone plants—Lithogeochemistry of Mendip rocks—Specimens previously recorded as teuthids—Carboniferous lycopsid Anabathra—Meyenodendron, new. Alaskan lepidodendrid. 1986. Pp. 225—297. 0 565 07014 2. £13.00 Vol. 41 No. 1 The Downtonian ostracoderm Sclerodus Agassiz (Osteostraci: Tremataspididae). P. L. Forey. 1987. Pp. 1-30. 11 figs. 0 565 07015 0. £5.50 Vol. 41 No. 2 Lower Turonian (Cretaceous) ammonites from south-east Nigeria. P. M. P. Zaborski. 1987. Pp. 31-66. 46 figs. 0 565 07016 9. £6.50 Vol. 41 No. 3. The Arenig Series in South Wales: Stratigraphy and Palaeontology. I. The Arenig Series in South Wales. R. A. Fortey & R. M. Owens. II. Appendix. Acritarchs and Chitinozoa from the Arenig Series of South-west Wales. S. G. Molyneux. 1987. 0 565 07017 7. In press. Vol. 41 No. 4 Miocene geology and palaeontology of Ad Dabtiyah, Saudi Arabia- Compiled by P. J. Whybrow. 1987. 0 565 07019 3. In press. Vol. 42 No. 1 Cenomanian and Lower Turonian echinoderms from Wilmington, south-east Devon. A. B. Smith, C. R. C. Paul, A. S. Gale & S. K. Donovan. 1987. 0 565 07018 5S. - In press F Bulletin of the British Museum (Natural History) : 4 ue Series in South Wales: atigraphy and Palaeontology B arenig Series in South Wales : . Fortey & R. M. Owens \ 4 opendix. Re isnchs and tinozoa from the Arenig Series south-west Wales 3M olyneux mt | Vol 41 No3 30 July 1987 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum’s resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 S5BD, England. World List abbreviation: Bull. Br. Mus. nat. Hist. (Geol.) © British Museum (Natural History), 1987 The Geology Series is edited in the Museum’s Department of Palaeontology Keeper of Palaeontology: Dr L.R.M. Cocks Editor of the Bulletin: Dr M. K. Howarth Assistant Editor: Mr D. L. F. Sealy ISBN 0 565 07017 7 ISSN 0007-1471 Geology series British Museum (Natural History) Vol 41 No 3 pp 67-364 Cromwell Road London SW7 5BD Issued 30 July 1987 The Arenig Series in South Wales: Stratigraphy and Palaeontology I. The Arenig Series in South Wales by R. A. Fortey and R. M. Owens with a preliminary note on the chordates by R. P. S. Jefferies II. Appendix. Acritarchs and Chitinozoa from the Arenig Series of South-west Wales by S. G. Molyneux ISSUED 30 JUL 1987 The Arenig Series in South Wales R. A. Fortey Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 5BD R. M. Owens Department of Geology, National Museum of Wales, Cardiff CF1 3NP with a preliminary note on the chordates by R. P. S. Jefferies Contents SIVAN OPS Sea car acepstey sistas sie cisracane “ise evevereio ce are ere arainapepele stehc en sacye eter ire Meeguisieteseuure erecta ela a MNO GUCH ON ese nee oe ace nee ie ateverat tee mine ete cede eee ae rere site PU ISTOGIGAIUS UL CY most ccckesctrays ere ares = Nera anise Ch moe cles eT ieee Cetera Pees Geology and lithostratigraphy of the Whitland district ..................... Biostratigraphy, ereee a corte ase teien vice oacee mente. eee ae deen ee StapesmnechesbuitisnvATenige jane eer aereace ates ancien enone cece nee ihevArenie—Wlanvirn) boundary s.--2.-sss--osseoe cess scce neces see cere cece Correlationtin' the BritishvIsles’ “2a. -ree- ++. ossceece- seer ceeeeeee ee eee: MinteLnational COmelation aa ccwanene ee oe OO eee 191 Genusilliaenopsis;Saltery Gere. seer eee ae oooh peer rere 193 Famulyalilaenidaesrawle:&aCordagereee eee eee ee eee eee eeeeer eee eee eee eer reer 199 Genus Ectillaenus,Salter scosysconeee acoee ee PRE ee en oo eee 199 Ramuily sirimnucleidaewHawleyéa\ Corda geeeeeeer eee eeeee eee ee EReeePe ee etree eee 203 Subfamily sirinucleinae Hawlerséai@ordalee see eee eee eee ee eee ereee eee 204 GenuseBengamiaaWwihittatdseereer renee Eee eeeeee ee ECE EEC eRe eRe eee rer Eee 204 GenusiHurcalithussnovn perrece cere eee ree ae eee eee eee cents 207 GenusyStapelayellayWihittardeererer eee eee eee ere ee eee rer Eee eee eerre 211 Genus: GiyMNnOstOmix NOV. 6 Nee ans retes epie testa em EL le ee ele ee 215 Family Dionididae/Gurich ii. cecses. scence eee ies <- epee eee eee hee 218 Genus DionidesBarrande: 7. csncnerans ononiysan cinta ree eee eee 219 Genus Dionidella Erantli Sab iibyleeee-eeeeen a eee Eee EEE EEE eee eee 221 HannlyiRaphiophoridacyAn¢e lint -eeeeeerter eee ee ee eeeeeE Erne ee Eee eererer 223 Genus. Am pyr Dalaran © coon ciosasssoes cys coetste are sores oe teclee eo nles ee eee 223 Genus@GnemidopygeWihittardeeepees ae eee eee eee eee ee eee ere eerie 227 Ramuily7Alsataspididaemliunnen -a-.cesasaeeeeece cee een cee rece ec Ee eee Lee EEr ree 230 Genus Seleneceme Clark... hc sastask secret Soe eee toon ee eee eee ae eee 230 Family Cheirunidae Salter Jc. c.nceye cso ceccnce eee ee ee OER OEE ee 231 SubfamilysBccoptochilinaelBaneieeereeeee eee eee eee Cee EEE eee eee eer 231 GenuspPlacoparinaaWwinittardae eee e eee eee eee eer eer Eee eee ece eee renee 231 RamilysbliomendacsRaymond eeppaeeeh eee eeeEE Eee en ee eee eeeEereE eee ene 231 SubfamilysPlacopaninacs lupe mereree rere ee er eeee ree eee Eee eee ener ee eere 73) GenuspRlacopariastiawlersa Cordaueereetceeeer eer nett eer ee cnc err ence rr rrte 231 HamilysEnerinunidaesAngelinges-necee eee eE eee nee Eee ee eee eC ee Ee eEe eee eeeeee 234 Subfamily Dindymeninae Henningsmoen .....................eeeeeeeeeeeeeees 234 GenussDindymeneiawlersai@ordayea-eeeee eee eee eee Renee ee 234 RamuilyiCalymenidaeyMilnesEdwardsie-----peeeeee ee ree ee eee eee Eee eee eee eee 237 SubfamilysReedocaliymeninac) lupe seeee seer e eee ee eee Eee eee een eeeee eee 237 Genus Neseuretus HICKS a aceccacesacee eae tee aa eeceeee reece reece 237 Subfamilyi@olpoconyphinacsHupeleeeeseeee eee eee eee Eee eeeer eee 240 Genus Colpocoryphe Novak in Perner .....................00eeecceseeeecees 240 Sarai hy IDyalvanevatireare WORGCIES jc coeccodoonsonagecccconsasaasnaovccancoccoosasene: 243 Subfamuily7ZeliszkellinaesDclomeereeereh ee ee eee eee eee eee ECE eee eee eee rate 243 GenusiOrmathopssDelo nic aececerss ota eee ate See Ee Cee 243 Family Odontopleuridae Burmeister .................. ee eeeeeene eee e eee e eee eeees 248 Genus)Selenopeliis Hawlerca Cord aaerece peer ee tree rere ee eee eee errr eee eerie 248 Systematic descriptions: Graptoloids .................eeeeeee eee e teeter teens eens 251 Family Dichograptidaes Bap worthieeee reece eeeee eee ELE eee eee eee eeeee 251 Subfamily Dichograptinae Lapworth ....................eeee eee cece eee e ee eees 251 Genus’ Tetragraptus Saltetgcsss5 A eccce ee e eee ee eee Cee eee eee 251 Subgenusmiletiagnapius|Salteiaasere eee Eee ee eeeEEee eee eeCe Ee EE EEE Ere eee eee 251 Genus DidiymograpiuseMiCOVmerre eee eeeer eee Cree EGE eee eee EEE eet ee EEE ee 254 Subgenus) idymograpitiss MC oye er eeeee ere eee Eee eee eeeere 254 Subgenus Didymograptellus Cooper & Fortey ...............--00e-eeeees 259 ARENIG IN SOUTH WALES 71 Subgenus Expansograptus Bouéek & Pribyl ..........................0- 260 Genus Azygograptus Nicholson & Lapworth ...................00000e0000ee 275 Genus Pseudotrigonograptus Mu & Lee ........... 20. c cece cence eee 278 Subfamily Sigmagraptinae Cooper & Fortey ...................00cc cee cee eee 278 GenusPACKOGrapluUusU Za sacar neiecccs wees reac aee clone Sule ee Reet aE ayn 278 HamilyiGlossograptidaezapworth) sass -eeeeeeceee eee eae seer e eee 281 GenusiGlossograptusiEMmmMOnSmercerreeere eee ee cee ee eee eeee eee eee 281 Family Diplograptidae Lapworth 32 .2n.-e0- «sae o-oo eee e+ seeeeeeeeins sees ee 282 GenusiGhyptograpius Lapworth a.-ceeeeer reece eee ee eee cere ree 282 INCKNOWLEGBEMENtS ie. Cees chetayconioct ani denorsie Deis MAE na ae ne ct Cee 285 The chordates—a preliminary note. By R. P. S. Jefferies ........................... 285 INGISWGINGES, = boat boa pe eractes cutie ote ohio Hae oe aiaGiae aaa SO aORPcec epcicaie corti aoe 290 INCI. gan ho Aeecseu teeta er aS Ones OS Gory Ari Nene aA cine aa tare onaGn aire So acinar ee 303 Synopsis This paper completes the first detailed description of the Arenig Series in south Wales, the fullest develop- ment of the Series in Britain. The fossil faunas are richer in species, and more differentiated strati- graphically, than has been supposed. The British Arenig is comparable in development to that in Scandinavia and North America, and affords the fullest section yet known at the edge of the Ordovician Gondwanan continent. The sequence is best developed in south Wales, between Carmarthen and St David’s. Based on the stratigraphy of this area we propose three stages, in ascending order the Moridu- nian, Whitlandian and Fennian, which can be correlated internationally, although only the Fennian includes enough cosmopolitan graptolites to place the correlation unequivocally. The threefold division is recognizable throughout Wales, and has refined the correlation considerably. Important stratigraphical conclusions are: (i) The type Arenig section at Arennig Fawr certainly includes the Moridunian; Whitlandian and earlier Fennian equivalents are not yet proven there. The base of the Moridunian may be as old as the Tetragraptus approximatus Biozone of the Pacific graptolite province. (ii) Conversely, on Anglesey only the later Arenig is present. (iii) There is a full Arenig succession on the western Ll¥n Peninsula. (iv) The Shelve Inlier is consistently developed in shallow-water facies compared with elsewhere in Wales, and may be somewhat incomplete at the base. (v) The ‘Tetragraptus Beds’ (now Penmaen Dewi Formation) of the coastal sections at St David's, which have been referred to the late Arenig Didymograptus hirundo Biozone, are demonstrated to be of mid-Arenig (Whitlandian) age; Fennian has only been proved on Ramsey Island. (vi) The Abercastle Formation, and part of what have been referred to as the Brunel Beds, which have been assumed to be of early Arenig age like the Ogof Hén Formation, are Whitlandian in age. In the Carmarthen—Whitland area the Arenig succession is thick—a cumulative thickness of more than 1300 m—compared with its equivalents on the coast at St David’s. Refinement of the biostratigraphy has led to the recognition of several new formations for the Whitlandian and Fennian rocks, which are described here: Blaencediw Formation, Colomendy Formation, Cwmfelin Boeth Formation, Pontyfenni Formation and Llanfallteg Formation in ascending order. A new map of the Whitland area records the outcrop of these formations, and their correlation throughout south Wales is discussed. The Afon Ffin- nant Formation, an eastward equivalent of the Blaencediw Formation, is also proposed. The fauna of the Whitlandian includes two new species, and the remainder, originally described by Salter and by Hicks, are here given their correct stratigraphical placing. The Fennian, however, includes a major new addition to the British trilobite fauna, with some twenty new species, many of them cyclo- pygids. This is the earliest known occurrence of the cyclopygid biofacies in Britain, and includes a suite of genera otherwise mostly known from younger rocks. Large-eyed pelagic forms are accompanied by a characteristic group of blind or nearly blind species, which we term the atheloptic assemblage. The combination of pelagic and atheloptic trilobites typified ‘oceanic’ sites with sediment accumulation prob- ably below 300m. The following new taxa are described: the subfamily Pricyclopyginae; the genera Circulocrania, Fur- calithus, Gymnostomix; the subgenus Bohemilla (Fenniops); and the following species: Segmentagnostus whitlandensis, Shumardia (S.) gadwensis, S. (Conophrys) crossi, Bohemilla (Fenniops) sabulon, Degamella evansi, Microparia (M.) porrecta, M. (M.) teretis, Novakella copei, Sagavia glans, Pricyclopyge dolabra, Circulocrania orbissima, Bergamia rushtoni, Furcalithus radix, Stapeleyella abyfrons, Dionide levigena, 72 R. A. FORTEY & R. M. OWENS Ampyx linleyoides, Dindymene saron and Colpocoryphe taylorum; also the subspecies Cyclopyge grandis brevirhachis and Pricylopyge binodosa eur ycephala. Seven trilobite assemblage biozones are proposed for use in local correlation; some are traceable throughout Wales and England. From the base of the Arenig these are: Merlinia selwynii and M. rhyakos (Moridunian); Furcalithus radix and Gymnostomix gibbsii (Whitlandian); Stapeleyella abyfrons, Bergamia rushtoni and Dionide levigena (Fennian). Of these, only the last-named seem to correspond with the hirundo Biozone as recognized in the Lake District and north Wales. This attests to the disproportionate extent of the old extensus Biozone. A good, continuous fossiliferous section between the Arenig and Llanvirn is described from the Llanfall- teg railway cutting, which may be a candidate for a stratotype for the base of the Llanvirn, and the distribution of fossils through this section is documented. In this confacial section the Llanvirn boundary is marked by the appearance of abundant ‘tuning fork’ graptolites, but many species of both graptolites and trilobites cross the boundary. The name D. artus Biozone is proposed as a replacement for the D. ‘bifidus’ Biozone of the early Llanvirn. Introduction The Arenig Series has been relatively neglected in its type development in Britain compared with the Llandeilo, Caradoc or Ashgill. This is partly because the Arenig rocks have been regarded as poorly fossiliferous—there are few of the prolific localities which yielded classic faunas to the nineteenth century monographers—and because the twofold division into Didy- mograptus extensus and D. hirundo Biozones obscured their stratigraphical diversity. After studying fossiliferous and stratigraphically differentiated Arenig faunas in Spitsbergen and elsewhere we realised that either the Welsh Arenig Series must be incompletely developed, or its true stratigraphical sequence had not been recognized. In this work we show that the latter is the case. The Arenig in the type area of north Wales is indeed incomplete; it apparently comprises only the lower one-third of the Series. The fullest sequence has proved to be in south Wales, outcropping over a stretch of country running from Carmarthen to the coast at St David’s. We studied the lower part of the sequence in the Carmarthen area, the results of which were published by Fortey & Owens (1978). The present work completes the account of the trilobite and graptolite faunas through the mid- and late Arenig of south Wales, and summarizes the biostratigraphy of the Arenig Series throughout Wales. Faunas of articulate brachiopods have been recovered from the M. selwynii, F. radix and S. abyfrons Biozones, and will be described elsewhere, and a first account of the acritarch flora is given by Molyneux (1987). A preliminary note on the chordates, by R. P. S. Jefferies, is appended to the present paper. We can now recognize seven broadly conceived biozones within the Series, and a major threefold division applicable throughout Britain, which we propose as stages. A preliminary account of the major divisions was given by Fortey (in Whittington et al. 1984). The Welsh Arenig is now known to be comparable in development to that of Scandinavia, China and North America. The number of trilobite species known has been trebled as a result of recog- nizing the true extent of the Series; the late Arenig fauna is almost entirely new, and the stratigraphical placing of the middle Arenig faunas has previously been incorrect. It remains true that, although rich in species, the fossils are relatively sparse—the specimens figured here are the product of nearly ten years’ collecting. Nonetheless, it would not have been possible to unravel the complexities of the Arenig of south Wales without the help of the trilobite faunas. If this work has a general relevance, it is that macrofossils can still prove indispensible in solving broad-scale geological problems in areas of structural complexity. Historical survey The term ‘Arenig’ was introduced by Sedgwick (1852) for strata cropping out in the Arennig Fawr district, Gwynedd. Its first application in south Wales was by Hicks & Salter (1867) in the St David’s district, Arenig strata there having previously being regarded (e.g. by Murchison (1839), and on Old Series one-inch Geological Survey maps) as Upper Cambrian, Llandeilo or ARENIG IN SOUTH WALES 73 a Anglesey XY : { Arennig Fawr ay, if if Shelve Inlier Aberystwythp motisey —_ _ Fig.3 Islandg a a David’s s ee Whitlandn; 7 Carmarthen s Fig.2 °o Haverfordwest Cardiff fo} Fig. 1 Outcrops of Arenig age in Wales and the Welsh Borderland referred to in the text. Caradoc. Subsequently, some strata, since shown to be of Arenig age (see Whittard 1960), were considered to belong to the Tremadoc (the Ogof Hén Formation: Hicks & Salter (1867), Hicks (1873); the Carmarthen Formation: Crosfield & Skeat (1896).) True Tremadoc strata have, however, recently been found in the Llangynog area near Carmarthen (Cope et al. 1978). Arenig strata in south Wales were described in the last century by Hicks (1873, 1875) in the St David’s area, by Roberts (1893) in the Whitland area, where he termed them ‘Tetragraptus- beds’ (which name, or its variant ‘Tetragraptus Shales’ came to be widely applied to the Arenig in south Wales in the ensuing years) and by Crosfield & Skeat (1896) in the Carmarthen area. 74 R. A. FORTEY & R. M. OWENS Roberts’ work was taken up by a local amateur, D. C. Evans, whose description (1906) of the Whitland—St Clears area forms the basis of the current paper. In the words of O. T. Jones in Stephens (1941) ‘It was recognised as a first-rate, masterly, piece of work that no one could pick to pieces’. In the same year Cantrill & Thomas (1906) described a small area around Llangy- nog. At the beginning of the present century the entire Arenig outcrop from the Whitland area in the west to Llanarthney in the east was mapped on the scale of 6in: 1 mile by the Geological Survey, and was described in three memoirs (Strahan et al. 1907, 1909, 1914). For the Whitland area the Survey mappers drew much upon Evans’ work. Over the following years interest in the St David’s-Haverfordwest area was renewed, and Arenig and contiguous strata were described in a series of papers, including those of Pringle (1911), Thomas & Jones (1912), Cox & Jones (1914), Thomas & Cox (1924), Cox et al. (1930), Cox (1930), Pringle (1930), and Williams (1934). In these, several local stratigraphical terms, especially for more arenaceous developments in the Arenig, e.g. Abercastle Beds, Porth Gain Beds, Brunel Beds, were proposed. After this period, apart from brief mention in W. D. Evans’ (1945) paper on the Prescelly Hills, the Arenig of south Wales attracted little attention until Bates (1969) described the early Arenig Ogof Hén Formation on Ramsey Island. More recently Fortey & Owens (1978) revised the early Arenig stratigraphy of the Carmarthen area, and Cope (1980) described similar strata in the Llangynog area a short distance to the west. The first zonation of the British Arenig was by Elles (1904), and was based upon graptolites from various sections in Wales (Afon Seiont, Caernarfon, Menai Straits, Llyn, Arennig Fawr, St David’s, and Haverfordwest-—Carmarthen). She defined two new zones, Didymograptus extensus and D. hirundo, and included Lapworth’s (1880) D. bifidus zone. The last-named was incorpo- rated in the Arenig because Elles at that time did not accept Hicks’ (1881) Llanvirn Series. Elles & Wood (1914) added a Dichograptus Zone below extensus in the Skiddaw Slates, but this has not been generally recognized, and Bulman (1958) could find no stratigraphical evidence for its existence. Later, Elles (1933) subdivided the extensus zone into four subzones, based upon the Skiddaw Slates succession, which in ascending order are: Upper Subzone of Tetragraptus (reclined), Didymograptus deflexus, D. nitidus and Isograptus gibberulus. Of these, the first- named has not been subsequently recognized (see Jackson 1962). The other three are now commonly accorded the status of zones (e.g. Skevington 1976; Cooper & Fortey 1982), which practice is followed here. A basal Arenig zone, the T. approximatus Zone, has been widely recognized throughout the world, but was not until recently (Stone & Rushton 1983) identified in the British Isles, where it has been found in the Girvan succession. General reviews of the trilobites were given by Stubblefield (1939) and Thomas et al. (1984), and the total fauna of the Arenig of Wales by Bates (1969). Individual elements of the fauna have been described over almost 150 years: trilobites by Murchison (1839), Salter in Murchison (1859), Salter (1866a, b, 1867), Hicks (1873, 1875), Crosfield & Skeat (1896), Reed (1931), Whittard (1955-67), Bates (1969), Fortey & Owens (1978); brachiopods by Murchison (1839), Davidson (1868, 1869), Bates (1969); graptolites by Hopkinson (1872, 1873), Hopkinson & Lapworth (1875), Bulman (1928, 1934), Fortey & Owens (1978), Jenkins (1982), Zalasziewicz (19845); crinoids by Hicks (1873), Ramsbottom (1961), Bates (1968b), Donovan (1984); asteroid by Hicks (1873), Spencer (1918, 1950); parablastoid by Paul & Cope (1982); bivalves by Hicks (1873), Carter (1971); hyolithid by Hicks (1873). None of these works, however, give any hint of the diversity of the faunas which are described herein. Geology and lithostratigraphy of the Whitland district In our previous paper (Fortey & Owens 1978) we described the geology of the Carmarthen district, in which the early Arenig (Moridunian) is best developed. The Moridunian rocks and the underlying Tremadoc (Cope et al. 1978) are there brought to the surface in the centre of a large anticline. However, the higher parts of the Arenig Series are not well exposed (p. 97). Westwards, early Arenig rocks are again brought to the surface in the structurally complex Llangynog inlier, but the higher parts of the Series are apparently faulted out. Between this area and St Clears, only late Arenig (Fennian) rocks are exposed. Yet another anticline brings ARENIG IN SOUTH WALES WS earlier Arenig rocks to the surface in the area immediately to the north of Whitland. It is here that the succession of middle and upper Arenig rocks is most fully exposed, and in which the faunas are generally well preserved. It proved to be the critical area for an understanding of the biostratigraphical divisions of the Arenig Series, and for this reason a detailed account of the geology of the area is given here. Our new geological maps (Figs 2, 3) show the distribution of Arenig rocks over a 30km tract from east of Whitland to Treffgarne, north of Haverfordwest. The original account of the geology of the Whitland district was by D. C. Evans (1906), and the general structure he deduced has been confirmed by subsequent mapping. However, he conflated two coarse units within the Arenig Series, which we now recognize as the Blaencediw and Cwmfelin Boeth Formations, and referred the whole succession to the ‘Tetragraptus Beds’ without further stratigraphical refinement. Cantrill & Thomas (in Strahan et al. 1909) largely followed Evans’ interpretations, although they did separate two units of ‘grits’, and their fossil collections have proved most useful. Since then, the area has been virtually unstudied. Structure The best exposures are to be found in an area of high ground lying about 3km north of Whitland. The country is relatively hilly compared with the green and fertile agricultural land flanking the wide Taf valley around Whitland, because the anticline brings to the surface a series of coarse turbidites which take much of the high ground. Physiography closely follows the lithology and structure of the rocks. The coarse units are exposed in two main areas: (1) in the vicinity of Whitland Abbey, on either side of the Afon Gronw, and westwards past Pen-cil- post to Cwmfelin-Boeth, and (2) in a broad upland area running more or less east-west to the south of Llwyn-derw, and extending past the farms known as Blaenweneirch and Blaencediw. The latter comprise the oldest beds in the area (Blaencediw Formation); at the centre of the anticlinal area dips are generally low. This area of coarse-grained rocks has, however, been thrust over younger beds to the south of Blaenweneirch, where a tract of soft mudstones in the Pontyfenni Formation produces a broad depression. A second anticlinal area is centred on an east-west axis just to the north of Whitland Abbey. The coarse turbidites here are much younger than those to the north, but they were equated by Evans (1906). Because they are underlain by shales (Whitland Abbey Member herein) in the neighbourhood of Whitland Abbey, Evans (1906) erroneously concluded that these shales were the oldest beds in the area (lowest shales’). The late Arenig mudstones of the Pontyfenni Formation appear everywhere to overlie normally the higher turbidite formation, and form a drape around the edge of the anticlinal structure. Over much of the area there are small scale reverse faults, indicating movement from the north. Faulting of this kind appears to predominate in this area compared with the Carmar- then district, over which repeated strong folding was the norm. Repetition by reverse faults of this kind probably accounts for the broad area of outcrop of the upper turbidites along the eastern side of the Afon Gronw. Dips increase progressively both north and south from the centre of the anticline. The generally northward dips along the southern area of the map indicate that the contact with the Llanvirn there is likely to be inverted and/or a thrust contact. The northern limit of the structure, where there is faulting against the Llanvirn, is probably a major normal fault to the north of Blaencediw, although there are sections further to the west where there is continuous passage between the Arenig and Llanvirn. Finally, there are a number of important north-south faults; the course of the Nant Colomendy clearly follows one of these, and boulders of vein quartz are common in the stream bed. Distortion of fossils is generally slight, even in the soft mudstones, except at certain horizons in the Llanfallteg Formation. None of the argillaceous units is cleaved in this area, although correlative rocks in the Haverfordwest district can be strongly cleaved. Lithostratigraphy The elucidation of the stratigraphy of this district has entailed the discrimination of a number of lithostratigraphic units, which appear on the maps. With the exceptions of the Rhyd Henllan 76 R. A. FORTEY & R. M. OWENS LITHOSTRATIGRAPHICAL UNITS D.artus Biozone (may include some Llanfallteg Fm.) Alluvium and river terrace Whitland Abbey a Member Oe nets Geological boundary (drift) == Castelldraenog = Member a - - - Geological boundary (solid) < Rhyd Henllan a N Member 3 - — Probable fault [ _]Blaencediw Fm o 05 1.0 a ; m - ! Fig. 2 Geological map of the Whitland area, based on the ARENIG IN SOUTH WALES 77 22 3 LY Castelldraenog ov; aenweneirch Sangin ~ Epo) 0'= 0-0 = = swmfelin : oo f° ————————— i) ef =e 5 >Boeth aac: ‘Cefnmeuri © 0 0 0 ¢ o = = ¢ e & 7oS 1S _—— o PO ° ° ¥ cs Qo 0 ~ 56 P8309 0,09,0 29; SR . 89.6 os °, Pant-y- Z : SO °,Pant-y-grug : fe © 50 2 2 00 2 A= < A Peeks oo = authors’ mapping and Geological Survey six-inch sheets. 78 R. A. FORTEY & R. M. OWENS Member and Cwmfelin Boeth Formation they can be traced for considerable distances across country, in some cases right to the coast. In no case is the exposure of any unit complete; thickness estimates are invariably the minimum that we can be sure about, and especially in the case of the poorly exposed late Arenig the real thickness may be considerably more than these minimum estimates. The stratigraphical column is summarized in Fig. 4. Blaencediw Formation. This is exposed in the centre of the anticline in the area around Blaen- - cediw and Blaenweneirch; it exceeds 80 m in thickness. The Blaencediw Formation forms the local base of the Arenig in the area, but we know from the fossils that it is younger than the Carmarthen Formation exposed to the east, and it is of Whitlandian age. Older beds are not seen, and it is possible that equivalents of the Carmarthen Formation were lacking in this area as they may be further to the west, in the area north of Haverfordwest. The type development of the Formation is in the old quarry (locality 31), 0-3 km NW of Blaencediw farm. The Blaencediw Formation consists of closely set, poorly graded turbidites and channelled mass flow deposits occasionally reaching +m or so in thickness, separated by gritty shales or silts, with sporadic seams of black shale. Dendroid graptolites are particularly abundant at some horizons, and the type locality of this Formation yielded the well-preserved specimens of Callograptus species figured by Bulman (1928) (his ‘Blaenweneirch’ locality). When fresh the turbidites are black, and characteristically include clear, subrounded quartz clasts, which show ‘wave’ extinction in thin section. Pebbles of fine-grained igneous rocks and the occasional shale, fine sandstone or albitic felspar clasts have also been noted. These are considered to have been derived from a Precambrian source. When weathered, the siltier interbeds take on a yellow, buff, to almost white colour (possibly this appearance led to earlier descriptions of the beds as ‘ashes’). The coarse beds often yield good moulds of single valves of brachiopods, and a rich brachiopod fauna has been obtained from several localities. The grits have been quarried in a number of places. At the eastern edge of the outcrop a new quarry (locality 30) was opened in 1976 in the side of the track leading to Blaencediw farm. There are old quarries on the western side of the Nant-Cediw near Blaenweneirch. The Blaen- cediw Formation forms the high ground south of Felin Henllan Amgoed, and there are several old quarries here. The best fossil collecting is probably in the large quarry at the top of the hill (locality 39), which is reached by following the track over the hillside beyond Rhyd-henllan. The middle Arenig (Whitlandian) age of the Blaencediw Formation is proved by the occurrence of the trilobite Ogyginus hybridus and a characteristic assemblage of articulate brachiopod species. An abundance of dendroid graptolites, including Callograptus and Den- drograptus, is also typical of the Whitlandian across south Wales, although particular species may have longer ranges, and are difficult to characterize. The Blaencediw Formation probably represents relatively immature sediments, possibly derived from an uplifted area in the Haverfordwest district. We presume that the articulate brachiopods were derived from shallow water, and their disarticulated valves were emplaced along with the turbidites and mass flow deposits. Large inarticulate brachiopods, including a Lingulella as much as 5cm long, are found between the turbidites, as are the dendroid grapto- lites, and these presumably inhabited the offshore environment. From fossil evidence, the Blaencediw Formation is considered the equivalent of the Afon Ffinnant Formation of the Carmarthen district, which overlies the Carmarthen Formation there. Because the exposure is continuous in the east, while we cannot see the base of the Blaencediw Formation in the type area, we have chosen to make the formal definition of the base of the Whitlandian in the area east of Carmarthen, rather than in the Whitland district. Colomendy Formation. The Colomendy Formation includes a considerable thickness of mud- stones and shales comprising the rest of the mid-Arenig. In the Whitland area it can be divided into three members which are distinguished below. RHYD-HENLLAN MEMBER. This overlies the Blaencediw Formation and is exposed to either side of the main outcrop of the latter. It is probably about 150m thick. Passage between the two formations is poorly exposed in the east side of the Nant Cediw valley, and on the valley side ya Lianvirn JG ee, BU Arenig UJ LITHOSTRATIGRAPHICAL UNITS Pest — Llanvirn Ordovician Dark shales with D. artus Tuffs and interbedded shales Lianfallteg Fm. Pontyfenni Fm. ok Spire Be presumed Arenig | Colomendy Fm., —j Rhyd Honilan Member Blaencediw Fm. Tremadoc? shales with lingulid brachiopods Treffgarne Bridge Fm | (Cambrian: Merioneth Series) 3234] Treffgarne Volcanic Fm. #\ sediments Fennian pene ie = es Alluvium and river terrace - Geological boundary (drift) - Geological boundary (solid) —-— Probable fault fe} 0.5 1,0 Fig. 3. Geological map of the Arenig and early Llanvirn outcrops west of Whitland, based on the authors’ mapping and Geological Survey six-inch sheets. ‘Lfandlasiiof oe > ARENIG IN SOUTH WALES 79 south of Felin-Henllan-Amgoed. The Rhyd-Henllan Member consists of sandy or silty shales, which characteristically break out in large, flaggy slabs a centimetre or two thick. Their appear- ance is very like that of much of the Mytton Flags of Shropshire. By comparison with most of the Arenig rocks fossils are not hard to find, and certain bedding planes are covered with pygidia of Ogyginus hybridus (Salter). Poorly preserved specimens of the trinucleid Gym- nostomix gibbsii are frequent, and there are occasional specimens of articulate brachiopods of Blaencediw type, and the asaphid trilobite Bohemopyge scutatrix (Salter). The best (and type) exposure of the Member is in the lane leading to Felin Henllan Amgoed, a locality known to Salter, and in the banks of the lane leading downhill from Llywn-derw farm towards Rhyd-Henllan, where a good continuous section is seen. Elsewhere, exposures are confined to small scrapings in the sides of tracks, as on the hillside east of Rhyd-Henllan and near Blaenweneirch, or in the stream beds of the Nant Colomendy or Nant-Cediw. The fauna is typical of the Middle Arenig (Whitlandian). CASTELLDRAENOG MEMBER. This is about 150m thick, and consists of a series of highly fissile steel-grey shales. It is the only part of the Arenig Series in the Whitland district in which the rocks can be split to give completely flat bedding surfaces which are often slightly lustrous (a ‘slaty appearance, although there is no cleavage). Very weathered surfaces are often coloured yellow with a fluffy deposit of alum. This lithology is identical with that of the Arenig in Pwlluog, Whitesand Bay (part of the Penmaen Dewi Formation of Hughes et al., 1982). There is no doubt that the Castelldraenog Member is also the temporal equivalent of this part of the Penmaen Dewi Formation, but for reasons explained later (p. 96) it is probable that the latter includes much else besides; hence we consider it preferable to introduce a local member name in the Whitland district. The Castelldraenog Member is exposed extensively around the only outcrop area to the south of Castelldraenog Farm. The type locality is by the footpath leading down to the Nant-Cediw from Blaencediw and Castelldraenog farms. The member is also exposed along the Nant Colomendy and its small tributaries. Dips are generally low, and the beds form small bluffs among the trees. Fossils are very rare; we have only recovered some fragments of the same dendroid graptol- ites which are abundant at certain horizons in Pwlluog, and Tetragraptus sp. However, the stratigraphical position is not in doubt because characteristic Whitlandian trilobites, including Bohemopyge scutatrix, occur below in the Rhyd Henllan Member and above in the Whitland Abbey Member. Even on the coast, where exposure of the Penmaen Dewi Formation is complete and accessible, fossil horizons are scarce, and more searching is required to find the appropriate horizons inland. WHITLAND ABBEY MEMBER. This is extensively exposed around the ruin of Whitland Abbey, and in the lanes and streams around Castelldraenog farm. The gradation from the fissile shales of the Castelldraenog Member can be seen along the lane running from Blaencediw to Castell- draenog. The member is at least 200m thick, and consists of a monotonous series of mud- stones, often poorly fissile compared with the underlying member. When unweathered the mudstones are sooty black, but they are rarely encountered in this condition, usually weather- ing to a blotchy, rusty colour, and breaking into crumbling fragments. In the centre of the southern part of the anticline, the mudstones can be examined in an old quarry opposite the Abbey ruins, and in the lane running from the Abbey to Pass-by, and in the dingle to either side of this lane. They are exposed again along the track on the west side of the Gronw about $km north of the Abbey Farm. Here the shales and mudstones can be seen dipping under the Cwmfelin Boeth turbidites, a fact which led Evans (1906), who confused the Cwmfelin Boeth and Blaencediw formations, to regard the Whitland Abbey Member as the oldest beds in the area (‘lowest shales’). The Whitland Abbey shales are exposed in the northern outcrop around Castelldraenog farm, in the farmyard and the lane leading down to the farm, and in the small dingle behind the farm. The mudstones are also extensively exposed in the stream bed of the upper reaches of the Nant Colomendy, and in the banks of the valley. Fossils are very difficult to collect, but the occurrence of Bohemopyge scutatrix, Cyclopyge R. A. FORTEY & R. M. OWENS 80 “SUOT}BUIOJ UQH{ JOSO pue usyJIeUIVD 10J (8/6 ]) SUEME 2 Ad}104 puke ‘}x9} 998 saIZO[OYR] JO UONdIOsOp 104 “‘S9TBAA YINOS UI jsSvd 0} JSAM WOIJ SUOIVUIIOJ JO SUOT}RI9II09 pojsas3ns pue AydeisneNsoyy Busy fp “Sty sajeys pue S}IS DOpewal] = — = & Ajtwiojuoosip qn >a 9 = 2) }]6oueysho uv | = aW fe jneyeiog W 4MqQ!d } ; qw | S]PM1j; WMD | qw ’eqqy uagas jou we uayyeWwieD Ww yueUUIy4 Ud}Y imaq-|ede9g D ° Cc ® © UD | w4 Apuswojo9 qw Aeqqy a= Slime) (pao ' i =| uljayWmMo ; : Bee Me ~ —S SS ~ S x io) S we A “SS wy Aww ~ “os Nw juUasA}UOY He - w Ng VN &) ~ y “a b 7 >> a SS ce ~ —- ~ ———— a ee See ————— wi = {pauiewuen usyyeWwsIeD PUCIVIUM isvai NVINNGIYOW w4 NVIGNVILIHM ee! ul sajyeys ; SS uPipuelyIUM i 1 — ! fe ! 7, a —NVIGNYTLIHM PUBITIUM ——Jimeqg-jedeg ul NVINN3SZ —7— —- = s- - —- —_ —_ — — sbe\4 ejnbulq = tes =]u2aH 4060 w4 SSS s| Me) OS i jAMeW Jaqy SOIUBDIOA pue sjuawipas UsIAURI JeyyINy Aq ulejsaao s.piaeq 3S LSAM ) : / ' | ARENIG IN SOUTH WALES 81 grandis grandis and Gymnostomix gibbsii in the Abbey exposures show that the age is Whit- landian. A small, deflexed Didymograptus species has been collected at several localities. Cwmfelin Boeth Formation. The base of the Cwmfelin Boeth Formation has been taken at the first coarse turbidite above the Whitland Abbey Member, and this is also taken as the base of the Fennian Stage. It is exposed by the track along the western side of the Gronw Valley (locality 22), 1 km NNE of The Manor at Whitland Abbey. The Formation probably exceeds 100m in thickness, but it is hard to be more precise because we are uncertain about how much fault repetition there may be, and it is clear that the turbidites thicken and thin locally. A thick series of such beds is present behind Whitland Abbey Farm and Manor, where they form a conspicuous feature, described by D. C. Evans as a ‘peculiar tump’. Over much of the outcrop single turbidite beds are about 1 m or less in thickness, and, unlike the Blaencediw Formation, they are usually well graded, with rounded clasts of rhyolite and rhyolitic tuffs, and mudstone flakes and plagioclase crystals. Unfossiliferous black shales a few cm thick serve to separate the turbidites. The Cwmfelin Boeth Formation is exposed around a broad area of outcrop in the southern anticline, where it takes the high ground between outcrops of the Whitland Abbey Member and the Pontyfenni Formation. It appears at intervals on the valley sides of the Afon Gronw; on the western side there are several old quarries in the woods, and $km south of Whitland Abbey Farm. The Formation is exposed at several points along the road running along the ridge north of Regwm to Pen-glog. West of the Nant Colomendy the Formation is exposed in an old quarry on the hillside south-west of Aberdeuddwr, and near the abandoned farm of Pen-cil- post. The best exposures are in the neighbourhood of Cwmfelin Boeth, in a quarry adjacent to the Henllan Amgoed road (locality 35), in several quarries on the hill to the west, and in a small stream south of Bryngwelltyn farm where the upward passage into the Pontyfenni Formation can be examined. Fossils are numerous in the upper part of the Formation within the turbidites, and they are probably derived from a relatively shallow source. Brachiopods are accompanied by the trilobite Asaphellus, the only horizon from which this genus has been recovered in this area. Pontyfenni Formation. With the disappearance of the Cwmfelin Boeth turbidites the succession continues upwards into a considerable thickness of black mudstones and shales of the Ponty- fenni Formation. The base may be formally defined above the last thick turbidite bed in the stream section (locality 37), near Cwmfelin Boeth, mentioned above. The whole thickness of the Pontyfenni Formation is nowhere exposed, but it occupies a very broad area of outcrop surrounding the anticlinal area and extending eastwards to Carmarthen and westwards to Llandissilio and beyond. We can only prove about 100m of section, but the total thickness must be several times this figure: 300m would probably be a conservative estimate. Parts of the succession are undoubtedly repeated by numerous faults and small folds, and there is local inversion near the southern limit of its area of outcrop. Because of the poor exposure, no attempt has been made to subdivide the formation on lithological grounds, although it certain- ly contains two successive faunas. The Pontyfenni Formation consists of shales and blocky mudstones. When fresh these are dark, sooty grey to black in colour, and the mudstones are poorly fissile, breaking conchoidally rather than along the bedding. In the weathered condition the mudstones become almost white, producing a blotchy and crumbly rock from which it is difficult to extract fossils. A weathering pattern of concentric rings is typically developed, and exposed surfaces are often dark choco- late brown or purplish, rather than vermilion like the mudstones of the succeeding Llanfallteg _ Formation. Fossils are generally sparse, but when obtained are often remarkably well pre- served, sometimes in full or partial relief (e.g. Figs 57, 101). The lower part of the Pontyfenni Formation includes shales which are more fissile than the typical mudstones, with silty and micaceous partings and numerous graptolites on some bedding planes. A common feature of many of the Pontyfenni outcrops is the presence of black, siliceous nodules, usually no bigger than a walnut, but occasionally 30cm across. These split with the greatest difficulty, but may contain very well preserved fossils. 82 R. A. FORTEY & R. M. OWENS Unlike some of the Whitlandian units, the Pontyfenni Formation is remarkably uniform lithologically—it can be traced from Carmarthen with some breaks to the St David’s area. In the Whitland area the best (and type) exposure is at Pontyfenni on the old course of the A40 east of Whitland. Beyond the eastern edge of the map (Fig. 2) it is again well exposed in a cutting by the main road just west of Banc-y-felin, and at several places, including the railway cutting, near Pwll-trap. It is exposed on both sides of the Afon Fenni: in an old quarry (locality 24) in the lane leading to Llwyn-crwn farm, and in the farmyard, about 1-5km north of Pontyfenni; also on the western side in the streams and trackways running down from the high ground near Pen-glog and Pant-y-grug farms. In the area to the north of Whitland Abbey, an important fossiliferous outcrop is in the lane running from Pen-y-parc to Pass-by, which is in the lower part of the Formation not far above the last turbidite. Outcrops of Pontyfenni Formation in close proximity to some of the older formations prove the north-south faults marked on the map (Fig. 2). Below Gelli-diogyn farm, for example, a track running down to the stream has several exposures, while the track on the opposite side of the valley displays the Castelldraenog shales. Similarly, the courses of the Nant Colomendy and the stream west of the road from Henllan Amgoed to Llwyn Derw follow the outcrop of the soft Pontyfenni mud- stones, which are exposed in the stream bed, passing to either side of the Whitlandian outcrop with its coarse grits. Outcrops are sporadic over the main outcrop area to the west of the Gronw. There are several small exposures on the western bank of the Nant Cwmfelin Boeth, and quite an extensive quarry behind Sarn-las farm. In the north of the area the Pontyfenni Formation is exposed at several places along the road to Cwmmiles, in a quarry below the village, and along the disused railway track. The same railway cutting exposes the upper part of the Formation to the north of Llanfallteg. Pontyfenni Formation mudstones again reach the surface in what we suppose is a subsidiary anticline in the neighbourhood of Llandissilio (Fig. 3). Exposure is poor, but typical mudstones are seen about 1 km SW of Llandissilio in the lane side (locality 53) where the road crosses the Afon Rhydybennau, and west of Llandissilio in an old quarry (locality 54) 500m SSW of Brechfa on the B4313. The fauna of the Pontyfenni Formation is rich, but nowhere are trilobites common. In small outcrops one is most likely to find one of the characteristic cyclopygids. Small, lentil-shaped ostracodes are also typical. On faunal grounds we can divide the Pontyfenni Formation into two, but there is no question of lithological division with the poor outcrop available. Llanfallteg Formation. The highest mudstones of the Pontyfenni Formation grade upwards into the Llanfallteg Formation, the highest lithostratigraphic unit described in this paper, which includes within it the Arenig—Llanvirn boundary. The contact with the Pontyfenni Formation is nowhere well exposed, but high Pontyfenni mudstones of typical lithology, with occasional biserial graptolites, underlie the Llanfallteg Formation in the type section along the Llanfallteg railway cutting (Fig. 8). This is much the best exposure of the Formation, where at least 60m of section are exposed, but the complete thickness of the Formation must be somewhat greater, and certainly exceeds 100m. The Llanfallteg Formation consists of light grey mudstones and shales, much lighter in colour than those of the Pontyfenni Formation. They weather white or yellowish, but do not become as crumbly as the weathered Pontyfenni beds. Weathered surfaces often develop a characteristic vermilion staining. Much of the Llanfallteg Formation is more highly fissile than the underlying mudstones, but the shales are very soft and trilobites and other fossils take on any distortion to which the rocks have been subjected. Partial stipes of Acrograptus acutidens are probably the commonest fossils on weathered pieces of Llanfallteg shales. The siliceous nodules found in the Pontyfenni Formation are uncommon in the uppermost Arenig. It is important to emphasize that there is no lithological change at all across the Arenig— Llanvirn boundary. The junction is well exposed in the small quarry at the southern end of the Llanfallteg railway section. As discussed below, the recognition of the Llanvirn boundary is governed by the sudden appearance of many ‘tuning-fork’ graptolites in the section (which are absent below), but this invasion is not reflected in any lithological change. For this reason there ARENIG IN SOUTH WALES 83 is no possibility of dividing the Llanvirn portion of the Llanfallteg Formation into a separate member. Similar shales and mudstones continue up into the Llanvirn for some tens of metres—all are highly fossiliferous. Shales of this horizon are well exposed around Rhyd-y-wrdach as well as in the Llanfallteg section. The beds become more blocky and less fossiliferous higher in the Llanvirn, and there is apparently a gradual transition into the typical black, micaceous fissile shales, mapped as bifidus beds by the British Geological Survey, that occupy a broad outcrop to the north and south of the Whitland area. The Llanfallteg Formation is an easily recognizable unit which takes up a broad stretch of ground in the neighbourhood of the type locality, where dips are low, and is repeated by minor folds and faults. We have not yet proved it in the Carmarthen area. Westwards it extends as far as the Scolton railway cutting, near Haverfordwest (locality 55), which has yielded a number of the earliest Llanvirn specimens figured in this paper. Biostratigraphy Elles’ (1904, 1933) graptolite zonation is difficult to apply to much of the Arenig in south Wales, chiefly because the lowest two-thirds is in essentially non-graptolitic facies. Thus the correlation of our new stages (Fig. 11, p. 100) with the graptolite zones is tentative, but it may be possible to refine it in due course through Dr A. W. A. Rushton’s current work on the Skiddaw Slates Group. In the Arenig of south Wales, a more precise and satisfactory biozonation can be achieved using trilobites, and we here propose seven assemblages biozones. It is likely that it will eventually prove possible to create a finer biozonation, for some of our biozones are broadly based. As with much of the later Ordovician, trinucleids have proved to be some of the most sensitive biozonal indices, and we have used these for four of our biozones. Because appropriate trinucleids are lacking in the Moridunian and the later Fennian, we have resorted to asaphids in the former and a dionidid in the latter. The seven biozones are now described in ascending order. Merlinia selwynii Biozone This is well developed in the Ogof Hén Formation and in the lower Carmarthen Formation (Pibwr Member) and well seen at localities 1, 5 and 63. The eponymous asaphid occurs throughout most of the biozone, but is absent from the basal part, which instead has Merlinia murchisoniae; the two species overlap in range in the Henllan Ash and in the Pibwr Member. Other typical trilobites include Myttonia fearnsidesi, Ampyx cetsarum, Neseuretus ramseyensis and N. murchisoni. Of brachiopods, Paralenorthis alata and Monorthis menapiae are typical. Merlinia rhyakos Biozone To date, this biozone has only been proved in the Carmarthen area, where it occurs in the upper part of the Carmarthen Formation (Cwmffrwd and Cwm yr Abbey members) and the lowest 40m of the succeeding Afon Ffinnant Formation. Here, in an olenid biofacies, it is accompanied by Bienvillia praecalva in the lower part and by Hypermecaspis venerabilis and Porterfieldia punctata in the upper. Its base is defined in Allt Pen-y-Coed, locality SE of Fortey & Owens 1978, and the lower part is well seen in this section; the upper part is well displayed in Nant y glasdwr (Fortey & Owens 1978 localities 3A, B) and in Cwm yr Abbey (Fortey & Owens 1978, locality 16). Furcalithus radix Biozone The base of this biozone is drawn at locality 16J, Cwm yr Abbey. It appears to incorporate only a small thickness of strata, with the fauna restricted to Furcalithus radix, Ogyginus hybridus and Azygograptus eivionicus. It has only been identified with certainty at Cwm yr 84 R. A. FORTEY & R. M. OWENS Abbey and Afon Ffinnant, but is probably represented in the arenaceous facies of the Blaen- cediw and Abercastle formations further west, and in a locality near Mathry (R. Kennedy, personal communication). Gymnostomix gibbsii Biozone A suitable section to define the base of this biozone precisely has not been found, but it presumably lies within the lower half of the Afon Ffinnant and Penmaen Dewi formations, and close to the junction of the Blaencediw and Colomendy formations. It includes a considerable thickness of strata, and is one of the most readily recognized in the field by the presence of the distinctive Gymnostomix gibbsii, which is usually associated with Ogyginus hybridus. The fauna of the biozone is restricted, and also includes Cnemidopyge salteri, Shumardia (S.) gadwensis and Bohemopyge scutatrix; the most diverse and abundant fauna has been collected at Pwlluog, Whitesand Bay. Azygograptus hicksii is locally abundant in the Afon Ffinnant and Penmaen Dewi formations, and dendroid graptolites (especially Callograptus) are frequent in the Blaen- cediw Formation and in the lower part of the Penmaen Dewi Formation. Stapeleyella abyfrons Biozone The characteristic fauna of this biozone, dominated by Stapeleyella abyfrons and Segmen- tagnostus whitlandensis is found at several localities in the basal Pontyfenni Formation, imme- diately overlying the Cwmfelin Boeth Formation, for example at Pen-y-parc (locality 38). The turbidites of the latter, which intervene between this fauna and that of the top of the G. gibbsii Biozone, are arbitrarily included in the Stapeleyella abyfrons Biozone. Bergamia rushtoni Biozone We estimate that some two-thirds of the Pontyfenni Formation is included in this biozone; in our present state of knowledge it is not possible to give a finer biozonation, although the presence of different trinucleids (e.g. Stapeleyella aff. abyfrons) in the lower part suggests that this may eventually be possible. The type development is at Pontyfenni (locality 23), but because of the nature of the exposures it is not possible to define its base satisfactorily. Included in this biozone are the diverse cyclopygid and atheloptic trilobite assemblages of the Pontyfenni Formation which includes over half those described herein; for many groups it is their first known appearance in the geological record. Graptolites include D. hirundo, D. uniformis lepidus, Tetragraptus reclinatus and Pseudotrigonograptus ensiformis. Their presence makes this biozone widely correlateable outside Britain. Dionide levigena Biozone This incorporates the Arenig part of the Llanfallteg Formation, and its type development is on the Llanfallteg railway cutting (locality 52). Passage downwards into strata containing a B. rushtoni Biozone fauna can nowhere be observed. Many of the trilobites, including the tri- nucleid Stapeleyella inconstans, pass upwards into the Llanvirn. The same applies to the grap- tolites; biserials such as ‘Glyptograptus’ austrodentatus, ‘G. dentatus and Acrograptus acutidens cross the boundary, which is marked by the influx of abundant pendent didymograptids. Stages in the British Arenig Moridunian Stage Nowhere is there exposed a continuous passage between the Tremadoc and Arenig, and a suitable international stratotype for the base of the Arenig is lacking; the formal base of the Moridunian has still to be defined. It will be necessary to resort to artificial excavations to expose a section which will bear comparison with those outside Britain. There are two possible sites in Wales for such an excavation: around the type area of Arennig Fawr in north Wales, ARENIG IN SOUTH WALES 85 Table 1 Stratigraphical distribution of trilobites and graptolites in the Arenig and basal Llanvirn in south Wales. Key to biozones: 1—selwynii; 2—rhyakos; 3—radix; 4—gibbsii; 5—abyfrons; 6—rushtoni; 7—levigena; 8—artus. Moridunian Whitlandian Fennian Lower —-~_—_ ae ee ~__——_—- Llanvirn TRILOBITES page 1 2 3 4 5 6 iT 8 Merlinia murchisoniaeft ........... = dk = i _ att AMPYGGELSATUMT, ....500000. 0.000. - + _ = = SS) ree = i Neseuretus ramseyensis ........... 238 + = = = = Merlinia selwyniit ................ — + = = = = = bs i= Myttonia cf. fearnsidesit .......... - + = = os Se = Neseuretus murchisoni ............ 239 + = = x = = = A Bienvillia praecalvat .............. ~ + + = = te ye hs Merlinia rhyakos{ ..............-.. ~ _ + Hypermecaspis venerabilist ....... ~ — + = = ok Spawn, 4 re (GYCLIOPVG!E? SPP cides eee ceecee ees — _ + = = Sy ale fs Porterfieldia punctatat ........... ~ + + — o@e tye. DE = WAMGGOUENUS TAGIX. «22. c0cisccewcceces 208 = = ae a Lip ge Be = Bergamia sp. A ..........--00ee000- 207 _ ~ + aa Ogyginus hybridus ................ 143 - - + + SS oo ui BIMSPMINGL CLS Fister cree te iesciees eee reis 148 = — = dt a RS ae Segmentagnostus hirundo ......... 116 - - = a Se - Shumardia (S.) gadwensis ......... 121 _ — = a = Bi = id Leioshumardia sp. A ............-. 126 + sai eS oe Microparia(M.) boia ............. 172 aE a eee at Bohemopyge scutatrix ............ 136 = — = + ares = Cnemidopyge salteri .............. 228 — = = ae oa rt = Furcalithus sedgwicki ............. 209 + ay ee es Gymnostomix gibbsii .............. 216 — — = + |) aa es ty Cyclopyge grandis grandis ........ Sul — = = a eh i Segmentagnostus whitlandensis ... 116 — — — = ofl 1 a See rz Asaphellus whittardi .............. 1132 — — = = oe ea eee ibs Stapeleyella abyfrons ............. 213 — - — = sin SEO Es = Dionidella? sp. indet. 1 ........... 222 — — — = ££ = = Arthrorhachis sp. indet. ........... 114 = — _ ~ Roos — Dindymene saron ...............++. 235 — — = = ie cle ghee = Shumardia (S.) sp. A ..........264. 123 — = _ ss a 9 = = Placoparia (P.) cambriensis ....... 232 — - _ = a fo dt +t Ellipsotaphrus monophthalmus ...__ 189 — ~ = = ah) ite st a Stapeleyella aff. abyfrons ......... 215 = _ _ _ = ? = = Leiagnostus cf. erraticus .......... 112 - - — _ = 6 = = Shumardia (Conophrys) crossi .... 123 — - - - = + = = Girvanopyge sp. indet. ............ Lay) - = = = = a SSeS ie Bohemilla (Fenniops) sabulon ..... 129 = = = = Sat iaelae = Cyclopyge grandis brevirhachis ... 154 — — ~— = = + = zs Cyclopyge cf. umbonata ........... 156 - = = = See = = Degamella evansi .................. IS - - = = a AS Microparia (M.) broeggeri ........ 164 _ - — = = 4 = = V0? S\o) iG (a 173 = ~ = = a: EQ = M. (Heterocyclopyge) sp. indet. .. 174 - _ - — = 2 = = Prospectatrix cf. superciliata ..... 176 - - - - = £ = = GEL GIANS 25. 52 :cd20c0ocnee eda e 177 _ - — — = ££ = = Pricyclopyge binodosa BHVGEDNGIA socccoceociacceasceas 181 - _ _ = = + = = ML OREDT oie 002 (5 o:0\e nisin aiaiaeesisiein es 184 — — — = SE Ras? = Circulocrania orbissima ........... 187 = = _ _ = at = = Psilacella cf. doveri................ 190 ~ — _ = =, pe 86 R. A. FORTEY & R. M. OWENS Table 1 Continued Moridunian Whitlandian Fennian Lower saat tiniest retin Zs Llanvirn TRILOBITES page 1 2 3 4 5 6 7 8 Barrandia sp. indet. ............... 193 — — - = = 5 Soe = Illaenopsis harrisoni .............- 194 - — _ = ee a Bergamia rushtoni...............-. 205 = — _ = =e = Ampyx linleyoides ................. 223 — = = = = a as Ectillaenus ?bergaminus .......... 202 — - — = = 4 es Dionidella? sp. indet.2 ........... 223 — _ - = eee eg Ee BE Colpocoryphe taylorum ........... 241 — — = = ery ee =e Ormathops nicholsoni ............. 244 — — — = Se = us Corrugatagnostus cf. refragor .... 113 = - - = ee = Segmentagnostus scoltonensis ...._ 118 _ — — _ = £4 4 at Microparia (M.) porrecta ......... 168 — — = = = ae a ae Selenopeltis buchi macrophthalma ................. 250 — _ — - = + 4 st Dionide levigena................... 220 _ — — — = = &£ au Placopaninaspaerrree eee eee eeer 231 — — _ = SS = = a Dindymene didymograpti? ........ 237 _ — aL Bn Microparia(M.) teretis ........... 170 _ = = = = = aL at Novakella copéi ................+.: 174 — — — = = = + au Pricyclopyge binodosa binodosa .._ (181) — = — = =.) a Barrandia homfrayi ............... 191 ~ — = = = are a Stapeleyella inconstans ........... Di _ = — = = = ae a Ampyx linleyensis ................. 226 - _ - — = = + au Seleneceme acuticaudata .......... 230 - — _ _ = = + ae Ectillaenus perovalis .............. 119 - - = = ee a Ormathops llanvirnensis ........... 247 - — — = = = + du Cyclopyge kossleri ...............- 155 _ _ = — =e at Gastropolus obtusicaudatus ....... 161 = — — - = = = = Microparia (M.) aff. broeggeri .... 168 a = = = = =te= st Selenopeltis buchi buchi ........... 249 _ = — = 2 Se a Dionide turnbulli ...............-.. 219 ~ - — — a ait GRAPTOLITES Phyllograptus cf. densust ......... - + = = = eae ee = P. cf. angustifoliust ............... = + = = = = ie Pi whe zygograptus eivionicus .......... 276 — _ + 2 = -. oe = VARI CKS Il gees cee Rn CORES 275 + — | = = Tetragraptus (T.) serra ........... (251) _ = = ae = = = = Didymograptus (Expansograptus) goldschmidti .................4.. 272 - — = + wae ae D. (Expansograptus) sparsus ...... 267 — - = = Sc se Tetragraptus (T.) reclinatus HECUNGLUS lo tetroscecargtasye ete: 252 - — - = a = T. (T.) reclinatus abbreviatus ..... 253 — — ap = = T. (T.) bigsbyi askerensis ......... 252 _ = = = = a = D. (Expansograptus?) uniformis [Git oUesecasepoonveaessoaanadaes 270 — — — = = ee = D.(Didymograptellus) sp. ......... 260 — — = = = he = Pseudotrigonograptus ensiformis . 278 _ _ = = = nines =. Pseudisograptus stellus* .......... - — = = = = a = Isograptus caduceus subsp.* ...... = = = = = = a me = D.(Expansograptus) hirundo ..... 260 _ — = = ee ? au ‘Glyptograptus dentatus .......... 282 _ _~ aR Or + ARENIG IN SOUTH WALES 87 Table 1 Continued Moridunian Whitlandian Fennian Lower fs ee = phe ns Llanvirn GRAPTOLITES page I 2 3 4 5 6 7 Glossograptus acanthus ........... 281 fe a ‘Glyptograptus austrodentatus ... 284 Jt a Acrograptus acutidens ............ 278 - = = = Sp a ot an PAPAS TOMA Fels -seicbeistaievesoiets ole ese are oss 280 4 Didymograptus (D.) artus ......... 258 4 IDN(D))iSpINUlOSUS «......20002226-5+-- 255 — = = = == St + Trilobites and graptolites marked with a dagger were described by Fortey & Owens (1978). * Graptolites marked with an asterisk were described by Jenkins (1982). and in the Carmarthen—Llangynog area of south Wales. Both, however, have attendant prob- lems. At Arennig the position of the boundary coincides with the site of emplacement of an intrusion (Zalasiewicz 1984a), but this complication probably does not apply in the Carnedd Iago area, a short distance to the north. In the Llangynog district exposure is poor, and the area is structurally complex (Owens et al. 1982: fig. 1). We are currently investigating the possibilities of the few available sites in Wales, and recent discoveries in the Lake District are promising. The Moridunian incorporates the Merlinia selwynii and M. rhyakos biozones (Figs 4, 5). Its most complete development is to be found in the Carmarthen area, although the best section for the lowest part is the type section of the Ogof Hén Formation on Ramsey Island (locality 63). The overlying Carmarthen Formation crops out only in the Carmarthen area, the only district in Wales where the upper part of the Stage has been proved. Faunally, the stage is characterized by the trilobites Merlinia, Neseuretus, Myttonia and Ampyx and by the brachio- pods Paralenorthis and Monorthis. The oldest beds are of particular interest because they have yielded some of the earliest crinoids (Ramseyocrinus cambrensis), parablastoids (Blastoidocrinus antecedens) and asteroids (Petraster ramseyensis), as well as those of various groups of bivalves, currently being investigated by Dr J. C. W. Cope. Whitlandian Stage The standard section for the base of the stage is below the bridge in Cwm yr Abbey (SN 5002 1985) in a continuous stream exposure of the Afon Ffinnant Formation, 40m above the base. It coincides with the Merlinia rhyakos—Furcalithus radix Biozone boundary. The fauna below the TRILOBITE | GRAPTOLITE STAGES BIOZONES |BIOZONES abyfrons rhyakos MORIDUNIAN selwynii ? earlier biozone in Lake District Fig. 5 Summary of stages and biozones in the British Arenig. hirundo gibberulus sesso ocS ? ———<—— — — deflexus ?approximatus 88 R. A. FORTEY & R. M. OWENS metres Fig. 6 Map showing proposed stratotype for the base of the Whitlandian Stage at Cwm yr Abbey, east of Carmarthen (see Fortey & Owens 1978: 231, fig. 3). boundary includes M. rhyakos and abundant Porterfieldia punctata and that above it Ogyginus hybridus, F. radix and orthoconic nautiloids. The higher part of the Whitlandian, which incorporates the Gymnostomix gibbsii Biozone, is represented in the Carmarthen area by the upper part of the Afon Ffinnant Formation which contains abundant G. gibbsii and O. hybridus, but the most complete and fossiliferous section is at Pwlluog, Whitesand Bay, in the Penmaen Dewi Formation. The trilobite fauna of this stage in its type development is characterized by the trinucleid genera Furcalithus and Gymnostomix, which are restricted to it, the asaphids Ogyginus hybridus and Bohemopyge scutatrix and the shumardiid Shumardia gadwensis. The graptolite fauna is small, and is dominated by Azygo- graptus. Fennian Stage The base is defined in the lane exposure opposite the bridge leading to Cwm-banau, north of Whitland (SN 2123 1862), at the base of the Cwmfelin Boeth Formation. Strata on either side of the boundary have yielded few fossils in the immediate vicinity, but the Whitland Abbey Member of the Colomendy Formation has yielded a typical G. gibbsii Biozone fauna at nearby Whitland Abbey. The Cwmfelin Boeth Formation carries a distinctive (allochthonous) fauna of articulate brachiopods, Asaphellus and Cyclopyge grandis grandis. Apart from the turbidites of the Cwmfelin Boeth Formation, much of the stage in its type development in the Whitland area is represented by sediments of offshore origin which carry a mixed graptolite—trilobite fauna. ARENIG IN SOUTH WALES 89 100 metres Coed Ysgubor-fawr Oo ce) Fig. 7 Map showing proposed stratotype for the base of the Fennian Stage, northeast of Whitland Abbey (see Fig. 2). The trilobites are dominated by cyclopygids, especially Pricyclopyge, and by the atheloptic assemblage of blind or small-eyed benthonic forms described below. The trinucleids Bergamia and Stapeleyella are typical of the stage, although both extend upwards into the Llanvirn. The Fennian includes at least part of the J. gibberulus and all the Didymograptus hirundo graptolite biozones (see Fig. 11, p. 100). Jackson (1962) recognized the D. hirundo Biozone in the Lake District more from the abundance of biserial graptoloids than from the presence of the eponymous species. As Jenkins (1982) has observed, the ranges of D. hirundo and I. gibberulus overlap, and both can be found below the range of biserials. On Jackson’s criterion, only the highest part of the Fennian (Dionide levigena Biozone) in the Llanfallteg Formation of south Wales would conform to his usage of the Didymograptus hirundo Biozone. The range of D. hirundo certainly extends down into the Bergamia rushtoni trilobite Biozone, and it is this fauna which includes identical trilobite species to those at Randel Crag in the Lake District. It is 90 R. A. FORTEY & R. M. OWENS likely that the Dionide levigena and B. rushtoni Biozones together are equivalent to all the Didymograptus hirundo and at least part of the I. gibberulus Biozones, and in Jackson’s (1962) usage the zonal boundaries would approximately coincide. Further refinement of detailed graptolite taxonomy from the Lake District, as opposed to the gross aspect of the fauna, will help to place these correlations on a firmer basis. The Arenig—Llanvirn boundary The base of the Llanvirn Series defines the top of the Fennian. The type locality of the Llanvirn is on the coast at Abereiddi Bay, north of St David’s, SW Dyfed (Hicks 1881, Hughes et al. 1982, Whittington & Rickards in Whittington et al. 1984) where the boundary between the Arenig and Llanvirn is placed with some question between the Penmaen Dewi Formation and the Aber Mawr Shale Formation. The problem here is that the boundary interval is not well exposed and lacks fossils, and prolonged search in temporary sections apparently close to the Llanvirn boundary as Caer-Rhys Farm (at SM 7954 3031 and SM 7957 3050) in 1974 and 1977 failed to provide any. The early Llanvirn has a rich fauna, and has traditionally been distinguished regionally from the underlying Arenig by the appearance of tuning-fork graptolites. The best section for the boundary interval that we have been able to find is at Llanfallteg, near Whitland (locality 52), where the exposure is continuous, fossiliferous, and with little distortion (Fortey & Owens in Rushton et al. 1979). Here, the boundary lies within the Llanfallteg Formation, and there is every indication that sedimentation was continuous. As might be expected in such circum- stances there is no sharp faunal break at the Llanvirn boundary (Fig. 9), which has to be arbitrarily taken at the appearance of the first tuning-fork graptolites. Many other species appear for the first time lower in the section, and continue upwards into the Llanvirn. This kind of gradation is to be expected in a truly continuous succession and we do not regard it as a disadvantage. It would have been possible to take a lower horizon as the base of the Llanvirn, e.g. at the appearance of the first abundant biserial graptolite fauna, but this would have placed it at or near the base of the Llanfallteg Formation, and at a lithological change, which seems inadvisable especially as it also takes the boundary away from the horizon traditionally used. We know of one other Welsh section in which there is continuous exposure across the Arenig—Llanvirn boundary, that on the Afon Seiont, Caernarfon (Elles 1904). Fossils are rarer here, and because it is much more distant from the type area of the Llanvirn, we do not consider that the section is appropriate as a potential stratotype for the Arenig—Llanvirn boundary. Name of basal Llanvirn biozone The earliest biozone of the British Llanvirn has been termed the Didymograptus bifidus Biozone. D. bifidus (Hall) is a species from the Quebec Group, and it occurs in rocks of Arenig, not Llanvirn, age. It has recently been shown (e.g. Cooper & Fortey 1982, Jenkins 1983) that Hall’s D. bifidus is not conspecific with the British Llanvirn forms identified with D. bifidus by Elles & Wood (1901). Obviously a new name is needed for the British early Llanvirn Biozone; on recent stratigraphical columns it has tended to appear as the D. ‘bifidus’ Biozone. Jenkins (in Hughes et al. 1982) utilized ‘Glyptograptus’ dentatus for this purpose. This is not an apposite choice, because we show here (p. 282) that the range of “G.’ dentatus extends well down into the Arenig. We suggest that the Biozone of Didymograptus artus is a suitable replacement. D. artus is widely recorded from this horizon, and is probably one of the more distinctive species in the variable plexus of Llanvirn pendent didymograptids. Are stages necessary ? Chlupaé¢, Fliigel & Jaeger (1981) have presented a case that the ‘series’ of the British Ordovician are not greatly different in duration from stages in other parts of the geological column, and that the division of the British ‘series’ into stages (in the Caradoc, for example) represents an 91 ARENIG IN SOUTH WALES 16 0 10 20 30 40 50 metres i I Fig. 8 Map of Llanfallteg railway cutting (disused), showing position of localities spanning the Arenig—Llanvirn boundary. 92 R. A. FORTEY & R. M. OWENS Fig. 9 Range chart of trilobites and graptolites across the Arenig—Llanvirn boundary at Llan- fallteg cutting. Asterisks indicate species not recorded from the early Llanvirn here, but at nearby localities, and arrows Fennian species from outcrops north of the measured section. euozolg euabiag| ‘GQ euozoig snwe’q NYIANV11 distance above or below boundary (in metres) —> TRILOBITES Corrugatagnostus cf. refragor Segmentagnostus scoltonensis Microparia (M.) teretis Microparia (M.) aff. broeggeri Novakella copei Pricyclopyge binodosa binodosa Ellipsotaphrus monophthalmus Barrandia homfrayi Ectillaenus perovalis Stapeleyella inconstans Dionide levigena Ampyx linleyensis Seleneceme acuticaudata Placoparina sp. Placoparia (P.) cambriensis Dindymene ?didymograpti Ormathops Ilanvirnensis Selenopeltis buchi macrophthalma GRAPTOLITES Didymograptus (D.) artus Didymograptus (D.) spinulosus Acrograptus acutidens Acrograptus? sp. A Glossograptus acanthus “Glyptograptus” dentatus “Glyptograptus’ austrodentatus * * i * SS ] = * * a = a = = unnecessarily fine subdivision, which was not considered useful in the international strati- graphical arena. We need to justify the introduction of stages within the Arenig Series here. 1. If the Ordovician (including Tremadoc) is taken as lasting between 65 (Harland et al. 1982) and 78m.y. (McKerrow et al. 1985), we may estimate the duration of our proposed stages as follows. We assume that the Ordovician series are equal in duration, except that the Llanvirn and Llandeilo taken together are regarded as no longer than any other series. The Llandeilo has long been recognized as comparatively short, and our assumption is justified by the numbers of graptolitic divisions which can be recognized within a series. We regard the Harland et al. estimated duration of 10m.y. for the Llandeilo, which consists of only one biozone, as improbable. On our estimation the Arenig would have lasted between 13 m.y. and 15m.y., according to which radiometric age estimate is followed. Our three stages, if about equal in duration, would each have the order of magnitude of 4 to Sm.y. Stages in the Jurassic are about 6m.y. on average. There is thus little difference in estimated duration between our Arenig stages and those of the Jurassic, which are often accepted as exemplifying what a stage should be. The stage divisions of the Caradoc would still be considerably more finely drawn | : | | ARENIG IN SOUTH WALES 93 (less than 2 m.y. if the chronology of Harland et al. is followed) and the criticisms of Chlupaé et al. (1981) might be more persuasive at this stratigraphical level. 2. Stage names exist in Scandinavia (and Estonia) for the Arenig interval. Why not apply these names to the British series? First, as we discuss below, the criteria for correlation with the Scandinavian succession are few and uncertain. Second, Britain probably lay at the edge of the Gondwanan continent (Cocks & Fortey 1982), well removed and at a latitude different from the Scandinavian craton. Faunal provinciality was at a maximum in the Arenig. In the words of Chlupaé et al. (1981: 78) ‘separate and different sets of stages for a single system are deemed acceptable for distinct palaeobiogeographic regions as long as sufficiently precise interregional correlations cannot be achieved.’ The British regional stages should form the standard for the Gondwanan sector. 3. Stages defined on graptolitic rocks are based on the sequence of faunas in Victoria, Australia (Thomas 1960). They have been widely applied in North America and New Zealand, and are of correlative value in China, arctic U.S.S.R., South America and elsewhere (Cooper & Fortey 1982). Four stages (Bendigonian (part), Chewtonian, Castlemainian and Yapeenian) approx- imate to the Arenig Series, a division on about the same scale as that proposed here. 4. The threefold division used here has already proved its worth in refining our understanding of sedimentation and correlation of Arenig sequences within the British Isles. It approximates to the informal threefold division of the Arenig in the Pacific graptolite realm (‘lower’, ‘middle’ and ‘upper’) suggested by Cooper & Fortey (1982: 168). There are still many difficulties in the way of detailed correlation, not least the endemicity of the faunas, but this correspondence does at least permit informal discussion of worldwide events during the Arenig, with reference to the type development of the Series. In summary, proposition of Arenig stages is justified because they are divisions broad enough to be commensurate with stages elsewhere in the geological column, and comparable with other subdivisions already usefully employed through the same time interval. Separate stage divisions for the type area are particularly applicable to the Arenig where correlation problems are at their most acute; our divisions may provide a standard for Gondwana. Finally, they have already proved useful nationally, and there is reason to suppose that they will correspond to a broad-scale, and practical, threefold division of the Arenig on the international scale. Correlation in the British Isles 1. St David’s—Haverfordwest Cox (e.g. 1916: 283, 1930: 279) stressed the twofold division of the Arenig in Pembrokeshire into a lower arenaceous and an upper argillaceous division. It is now evident (see p. 94) that the arenaceous sediments in this area, as elsewhere, are not coeval, and the succession is therefore more complex than has been supposed. Moridunian. The Ogof Hén Formation at its type development rests discomfortably upon ‘Lingula Flags’ (Ogof Velvet Formation of Kokelaar et al. 1985). The faunas from the earlier sandy/silty sediments are rich and well known (see Bates 1969), dominated by Paralenorthis alata (Sowerby), Monorthis menapiae (Davidson), Merlinia murchisoniae (Murchison) and Neseuretus ramseyensis Hicks, and are broadly equivalent to those of the Bolahaul Member at Carmarthen. The Ogof Hén Formation also crops out along the north side of the St David’s anticline near Carnedd-lleithr and Llan-verran, and was formerly exposed near Tremaenhir, NW of Solva (Hicks 1873). Other arenaceous strata in the neighbourhood, formerly equated with the Ramsey Island succession, are now known to be of early Whitlandian age, and belong | to the Abercastle Formation (see below). Later Moridunian strata have nowhere been proved in this area, unless the olenid recorded from the Roch borehole (Dr A. W. A. Rushton, verbal communication 1984) is conspecific with one of those from the Carmarthen Formation. 94 R. A. FORTEY & R. M. OWENS However, it is possible that some of the Arenig sequence exposed between Aber-mawr and Pwllderi on the west side of Pen Caer is of Moridunian age. Cox (1930: 281) described 200 feet of flaggy sandstones and micaceous sandy shales as ‘Abercastle and Porth Gain Beds’ which are apparently underlain by flags, quartzites and dark shales variously named the Cerrig Gwynion Quartzite, Trwyn Llwyd Quartizite, Aberbach Quartzites, Porth Duggan Flags and Porth Duggan shales. He noted (1930: 280) that they included lithological types peculiar to this region of Pembrokeshire in the Arenig. They occur in three fault-bounded outcrops, and some horizons have yielded Callograptus and lingulacean brachiopods, but no diagnostic fauna. Since the Abercastle Formation is of early Whitlandian age, a large proportion of these sedi- ments may, by inference, be of Moridunian age; if they are, it seems that in this area compara- tively shallow-water conditions persisted throughout the Moridunian. However, further work is needed before their age and correlation can be placed on a firmer footing. Whitlandian. Cox (1916) divided the arenaceous sediments underlying the Tetragraptus Shales in the Abereiddi-Abercastle area into the Abercastle Beds below and the Porth Gain Beds above. The Abercastle Beds, a series of cleaved, sandy blue-grey mudstones, are exposed in the cores of anticlines at Abercastle and at Pwll Llong, about 1 km to the west. Nowhere is the base seen. In thin section the rocks show numerous phenocrysts of plagioclase felspar and rounded lava fragments derived from the erosion of a calc-alkaline magma similar io that of the Treffgarne Volcanic Formation. The Porth Gain Beds have a larger outcrop, being exposed at various points along the coast from Porth Egr in the west to Aber Yw, a short distance east of Abercastle. They comprise micaceous sandy mudstones overlain by a coarse felspathic grit. They can be observed to pass upwards into the Tetragraptus Shales near the jetty on the east side of Porth Gain (locality 59). In most outcrops, they are faulted at their base against Lingula Flags, although sandy mudstones like those at Porth Gain can be seen passing down into the Abercastle Beds at Abercastle (Cox 1916: 287). Cox (1916: 288) was of the opinion that these are almost certainly the correlatives of the sediments at Porth Gain, an opinion with which we concur. We regard this succession as comprising the Abercastle Formation, and consider Cox’s Porth Gain Beds to be a member at the top. The presence of Ogyginus hybridus (in the past usually referred to ‘O. selwyniv’) in the lower part of this formation indicates its Whitlandian age and its equivalence to the Blaencediw Formation of the Haverfordwest and Whitland areas; an orthid brachiopod which occurs in the Porth Gain Member at Porth Gain also occurs in the Blaen- cediw Formation and in the lower part of the Rhyd Henllan Member of the Colomendy Formation at Whitland. Although they have afforded no fossils, the siltstones and sandstones on the northern part of Trwynhwrddyn are lithologically like the Abercastle Formation, with which we here tentatively equate them, rather than with the Ogof Hén Formation as has been the practice in the past. The nature of the junction with the underlying Lingula Flags here is controversial; the presence of a thin conglomerate suggests an unconformable base for the Arenig, although the whole area of the boundary is much disturbed by faulting. A further arenaceous division, the Brunel Beds, was described by Thomas & Cox (1924) from the Brunel cuttings on the east side of the Treffgarne Gorge. These overlie a succession of extrusive igneous rocks, the Treffgarne Volcanic Formation. At the northern end of the Brunel cuttings the Brunel Beds have yielded a fauna of dendroid and extensiform graptolites. Correla- tive strata in Triffleton Quarry a short distance to the east have yielded a richer fauna including Ogyginus hybridus which shows them to be equivalent to the Blaencediw Formation in which we here include them. Some of the rocks described as Brunel Beds by Williams (1934) from the east end of the St David’s anticline may also be equivalent (but see above, under Moridunian). The succession at the southern end of the Brunel cuttings is apparently separated by a fault from those at the northern end, and the lowest sediments seen, at the extreme southern end, are cream-coloured shales lithologically somewhat resembling the Treffgarne Bridge Beds of Upper Cambrian (Merioneth Series) age. They have yielded a fauna of lingulacean brachiopods and occasional bivalves, the latter including Glyptarca, a genus not hitherto recognized from pre- Arenig strata and known from the Ogof Hén Formation on Ramsey Island (Dr J. C. W. Cope, ‘Soe PUL PUL[ZUY UI sUOT}eUIIOJ BtUsIY Jo UONFIAI0D = OT “SI a) oN ueliqued ejppiw 10 xejdwod eucw UR |IqQWeDeld J.opewes, .opewes, Wy Y@Aj@A JOBO .86e14 ejn6uy7, .86e)4 ejn6uly, SO}18S YeUud|OW 2opewes, RAMA s0qWweW JOqUoW wai ie iD yey Boueishd HIV Ineyejog ie queusAyy 3 sojeys = Hy Jequew rs smqid ° i] s0qUeW 3 s0qwonm ysy ueyjueH | PAIjjWAD Sie ee wy 9UeII0A sequen euseBbsjosy uees jou eseq Aoqqy 4A wad é é we e1seqseqy mipeoueeig mipeouee|g n aa = | uo} AW (equew Te os: Zz sourkKma GW wed yWod ue||ueH-PpAyY) uejwen-pAuy | (as) 3 ye pue _snides6oyed, w4 Apuewojod g = z uemped-A-juen uA sojeys “4equew 3 a 4e. ue s6e 3 AY i) i JO epis "Mm 44ePp Pp ; ub BoueespyjeyseD a o (e) = Bulpnjouy ‘230 5 a riJ uo sejeys n ‘ce 1mog ueewued S6el) ‘soyzysenb seqUOW 3 (equen Vz z% yO uojsseoons lAeqay pueny Aeqqy pueiiuM) ae ev wy Apuewojoa g Aled & A Velealiis: os wa Cy y1s: € yjeog UN1E;WMD u o ail /el meee § ---4J-----4 O[Asoyuey uempesd-A-jue = peo-A-juen wg s|é so eps “3 seyon peou £ uo sejeys Uepliad w4 5 5 38 sejeys juuesAyuog = z juuesAyUu0g 5 FE eI dnoi5 auesjoqn OH ee eee feorsaas my] Bouueyuen mud SOAR rc rc p> z snje'q wy eje4us snyse ‘a snue'a < 4A soleus imew seqy UA sejeys UIA sojeys Po] =z YuMV4 GNV1SI LS3M- ‘ ‘ GNVILIHM | NSHLYVAYY Mebtmleeded || eo SINNSYV aATaHS AASWVY S clave ts GyuOsYySsAVH 96 R. A. FORTEY & R. M. OWENS personal communication 1984). Their precise age remains equivocal, but has implications for that of the Treffgarne Volcanic Group, which may be of early Arenig or Tremadoc age. Its resemblance to the Tremadoc Rhobell Volcanic Group of Gwynedd favours the latter (Dr R. E. Bevins, personal communication 1984). Rounded lava fragments of Treffgarne type are common in the Blaencediw Formation at Triffleton. Sediments closely resembling the Rhyd Henllan Member of the Colomendy Formation overlie the Blaencediw Formation in the country adjacent to Triffleton Quarry, but have so far yielded no fauna. The Tetragraptus Shales of the St David’s peninsula were named the Penmaen Dewi Shale Formation by Jenkins in an unpublished Ph.D. thesis (University of Cambridge 1979) and by Hughes et al. (1982) in an excursion guide. The latter (1982: 54) inferred that the type section was at Pwlluog, Whitesand Bay. However, although they are most fossiliferous there, owing in part to favourable cleavage—bedding relationship, they are in faulted contact with the under- lying (presumed) Abercastle Formation; this circumstance renders it an unsuitable section in which to define the base. We here define the base of the formation (whose name we here shorten to Penmaen Dewi) at the jetty on the east side of Porth Gain (loc. 59), where its passage downwards into the underlying Porth Gain Member of the Abercastle Formation is well exposed. The only richly fossiliferous section is at Pwlluog. Beds near the base, in Hicks’ (1875) ‘Lower Arenig’, have yielded Bergamia? sp., O. cf. hybridus, an orthid brachiopod and dendroid graptolites, suggesting that this horizon is not far above the Abercastle Formation. Higher horizons, as for example at the north end of the section (Hicks’ (1875) ‘Middle Arenig’), have afforded Gymnostomix gibbsii, Furcalithus sedgwicki, Bohemopyge scutatrix, Cyclopyge grandis grandis, Cnemidopyge salteri and Shumardia gadwensis, the richest known assemblage of the G. gibbsii Biozone in south Wales. These faunas clearly show that the Penmaen Dewi Formation is equivalent to the Colomendy Formation at Whitland and most of the Afon Ffinnant Formation (see below) at Carmarthen. Fennian. Because they tend to be well cleaved, most of the higher beds of the Penmaen Dewi Formation on the mainland have yielded no fossils, and their assumed Fennian age cannot be proved, only inferred on the assumption that the succession is complete. At Pwllderi, Pen Caer, shales lithologically like the Pontyfenni Formation are said to have yielded extensiform didy- mograptids (Cox 1930: 284), although we have not seen any material, and failed to collect any further. Similar shales in the Spittal district, Haverfordwest, are badly cleaved, and have yielded no diagnostic fauna. On Ramsey Island, the Road Uchaf Formation (Kokelaar et al. 1985) has yielded a rich early Fennian graptolite fauna including Isograptus caduceus (Salter) sspp., Pseudisograptus stellus (Hopkinson) and numerous dendroids at Road Uchaf (Jenkins 1982), whilst the Arenig part of the overlying Aber Mawr Formation (of Kokelaar et al. 1985, not of Hughes et al. 1982), the pencil slates at Aber Mawr, lithologically like the Pontyfenni Formation, have yielded abun- dant extensiform didymograptids, Degamella evansi and Ormathops nicholsoni, indicative of the Bergamia rushtoni Biozone. Attenuated equivalents of the Arenig part of the Llanfallteg Forma- tion have yielded no fossils on Ramsey Island (Aber Mawr), and nor have the equivalent strata on the mainland. 2. The Carmarthen area The lower, more arenaceous part of the Bolahaul Member of the Ogof Hén Formation con- tains the earliest Moridunian fauna in the area, and with Merlinia murchisoniae, Neseuretus ramseyensis and Paralenorthis alata compares closely with the Ogof Hén fauna from Ramsey Island. The higher, more muddy part of the member has M. murchisoniae, Neseuretus murchi- soni and Ampyx cetsarum, but very few articulate brachiopods. The succeeding Pibwr Member of the Carmarthen Formation is dominated by Merlinia selwynii, with rare A. cetsarum, M. murchisoniae, Myttonia cf. fearnsidesi and Pseudophyllograptus spp. These faunas are closely matched in the Henllan Ash (see below). The remainder of the Carmarthen Formation has a characteristic olenid fauna, not yet identified elsewhere. The Carmarthen Formation is con- formably overlain by a thick sequence of alternating shales and turbidites comprising what we here name the Afon Ffinnant Formation. The base is defined at the base of the first turbidite ARENIG IN SOUTH WALES 97 below the bridge in Cwm yr Abbey (SN 5002 1983). The lowest beds are of Moridunian age, but the bulk of the formation belongs to the Whitlandian (Furcalithus radix and Gymnostomix gibbsii Biozones). Ogyginus hybridus occurs throughout, accompanied by F. radix at the base and by G. gibbsii for most of the rest of the sequence. The turbidites contain polydeformed clasts of Precambrian quartz and schist, indicating erosion of a Precambrian basement, pre- sumably to the south. Probable higher Whitlandian strata with Bohemopyge scutatrix occur in Capel-Dewi stream (Survey locality Carm. 40NW W4A7) in a fault-bounded outcrop. These Whitlandian faunas correspond closely with those of the Colomendy and Penmaen Dewi formations further west. Fennian faunas which include Ormathops nicholsoni, Illaenopsis harri- soni and Pricyclopyge binodosa eurycephala have also been recorded from Capel-Dewi stream, and evidently broadly correspond with the Pontyfenni Formation, and the presence of Asa- phellus whittardi near the old adit (locality 20E) suggests early Fennian, by analogy with its occurrence in the Whitland district. Elsewhere in the Carmarthen district, proved Fennian strata occur only in isolated exposures in the neighbourhood of Bancyfelin (e.g. locality 21). 3. North Wales The stratigraphy and graptolite fauna of the type Arenig area has been recently revised by Zalasiewicz (1984a), who employed Lynas’ (1973) nomenclature from the neighbouring Mig- neint. He recognized three successive graptolite faunas, a lower one in the Llyfnant Member characterized by Didymograptus aff. simulans, a middle one in the Henllan Ash Member with Azygograptus cf. eivionicus and an upper one from the top of the Henllan Ash with Didy- mograptus cf. praenuntius and Tetragraptus reclinatus. He noted that these faunas offered no firm correlation with any of the deflexus, nitidus or gibberulus Biozones, but showed that there was no clear evidence for the hirundo Biozone, as previously claimed (e.g. by Fearnsides (1905), Lynas (1973)). Whittington (1966) described the limited trilobite fauna of the Henllan Ash; of this, Merlinia murchisoniae, Neseuretus murchisoni and Ampyx cetsarum occur also in the higher part of the Bolahaul Member in south Wales, whilst M. murchisoniae and A. cetsarum are common to the Pibwr Member. These indicate a Moridunian age for the Henllan Ash, and it is likely that most, if not all, of the Whitlandian and Fennian are absent from the type area. Work in progress by A. Beckly has proved the presence of all three Arenig stages in the — western Llyn (Fig. 10, p. 95), and much of the Arenig is probably also present in the Bangor— Caernarfon area; Elles (1904) described the Afon Seiont sequence which apparently has repre- sentatives of the gibberulus and hirundo Biozones, and Jenkins (1982) demonstrated the presence of the gibberulus Biozone on the Menai Strait. Correlation of the largely arenaceous sequence on Anglesey (Bates 1972) with the rest of the Welsh Arenig is equivocal. Williams (1974: 12) noted the common occurrence of Paralenorthis proava in the Carmel Formation and the Henllan Ash, but Dr M. G. Bassett has examined specimens from the latter in the collec- tions in the National Museum of Wales and considers them to be closer to P. alata, and not to belong to P. proava. This means that there are no species in common, for the trilobites, Neseuretus monensis, Ogyginus? sp. (Ogygiocaris selwynii of Bates 1968a), Ampyx sp. and Anna- mitella (= Monella) perplexa are all endemic. A ‘late Arenig’ age was implied for the succeeding Treiorwerth Formation by analogy of its brachiopod fauna with that of the Summerford _ Group, Newfoundland (Neuman & Bates 1978: 577). Assuming this to approximate to the _ Fennian, the Carmel Formation may by implication be Whitlandian, or could fall also within _ the Fennian. In the adjacent Bangor area Beckly has found new asaphid and Neseuretus species _ in arenaceous deposits underlying shales with Azygograptus eivionicus, which are of Whit- | landian age. It seems that shallow-water onshore conditions prevailed in much of the Whit- _ landian and Fennian in the Bangor—Anglesey area, although the offshore mudstones with _ isograptids on the Menai Strait of gibberulus Biozone age suggests that the pattern is more _ complex, and is likely also to involve structural factors. | 4. Shelve | Before our work in south Wales, the richest known Arenig shelly faunas from England and | Wales originated from the Mytton and Tankerville Flags, and the trilobites, brachiopods and 98 R. A. FORTEY & R. M. OWENS graptolites have been described by Whittard (1955-67), Williams (1974) and Strachan (1986) respectively. The basal transgressive Stiperstones Quartzite has yielded only Neseuretus gran- dior, which we consider to be a synonym of N. ramseyensis which occurs in the Ogof Hén Formation; this being the case, the two formations would appear to be approximately coeval. The lowest third of the Mytton Flags contains trinucleids, asaphids and calymenids. The two former are represented by endemic species, although related species of the same genera occur in the Henllan Ash and Ogof Hén and lower Carmarthen formations. Of the calymenids, Neseu- retus parvifrons occurs also in the Henllan Ash, whilst N. murchisoni is common to that formation and the upper part of the Bolahaul Member in the Carmarthen district, which implies that this part of the Mytton Flags equates with the Moridunian. The presence of Cyclopyge grandis grandis about 300m above the base of the Mytton Flags suggests a Whit- landian or early Fennian age, although the characteristic Whitlandian asaphid and trinucleid species are absent. The diverse trinucleid fauna from the top third of the Mytton Flags has no counterpart in Wales, but occurring as it does above the level with C. grandis grandis must be either late Whitlandian or early Fennian in age. The apparent presence of N. parvifrons with this fauna suggests that this species persisted longer at Shelve, perhaps because rather shallow water conditions continued. The fauna of the Tankerville Flags and Shelve Church Beds includes the trilobites Asaphellus whittardi, Placoparia cambriensis, Pricyclopyge binodosa and Selenopeltis buchi macrophthalma and the graptolite Didymograptus (Expansograptus) hirundo, which occur in the Fennian in south Wales. There is no indication at Shelve of the late Fennian fauna found in the Llanfallteg Formation; this fauna and its containing sediments apparently represent a regressive phase (see also p. 105), and the on-shelf position of the Shelve inlier might mean that it did not reach this area. It is likely, therefore, that the latest Arenig is absent from Shelve, and it is possible that there are also gaps within the Mytton Flags sequence, as might be expected in a relatively onshore shallow-water succession. 5. The Lake District The Skiddaw Group, a series of flaggy sandstones, silty mudstones and slates, is generally rather poorly fossiliferous. The supposed Tremadoc age of certain beds at Barf has long been discounted (Molyneux & Rushton 1985: 123), but definitive late Tremadoc trilobites and acri- tarchs have recently been found in the River Calder section at the western edge of the outcrop. A late Lancefieldian graptolite fauna identified from the area north of Skiddaw (Rushton 1985b) suggests that it might be possible to find a continuous Tremadoc—Arenig passage in the Skiddaw Slates, but despite past claims to the contrary, T. approximatus has not been found (Stone & Rushton 1983). The Hope Beck Slates, hitherto thought unfossiliferous, have recently yielded Didymograptus cf. vacillans Tullberg (Molyneux & Rushton 1985), indicative of the deflexus Biozone, which is also the age of the succeeding lower part of the Loweswater Flags; the upper Loweswater Flags contain a nitidus Biozone fauna. The Hope Beck Slates and Loweswater Flags thus broadly equate with the Moridunian and Whitlandian stages, but have not afforded any of the characteristic trilobites. The report of ‘Trinucleus gibbsii from the Ullswater inlier cannot be substantiated, firstly because the specimen(s) was apparently mislaid (Postlethwaite & Goodchild 1886: 468), and secondly because only the Llanvirn is represented there. Most, if not all, of the trilobites from the Skiddaw Slates appear to have originated from the gibberulus and hirundo Biozones, well represented in the Kirk Stile Slates. Here there occur cyclopygids and members of the Fennian atheloptic fauna (e.g. Ormathops nicholsoni, Illaenopsis harrisoni and Selenopeltis), also found in the Pontyfenni Formation in south Wales. It seems likely that offshore conditions persisted throughout much of the Arenig in the Lake District, and the rich isograptid fauna (Jenkins 1982) of the Kirk Stile Slates, a sure indicator of more oceanic conditions, occurs only in the Road Uchaf Formation, Ramsey Island in south Wales. The top of the Skiddaw Slate Group (or Eycott Group of Wadge, 1978) is of Llanvirn age, and the artus and murchisoni Biozones are present (Jackson 1978). Whether there is a continuous Arenig—Llanvirn passage is not yet known. ARENIG IN SOUTH WALES 99 International correlation The faunas of the Arenig Series in Wales are dominated by endemic forms, and international correlation is correspondingly difficult (Fortey in Whittington et al. 1984). Assuming that the position of Wales at the edge of an early Ordovician Gondwanan continent is correct, there are virtually no other known faunas occupying a similar palaeogeographical position with which a direct comparison is possible. The Welsh Arenig is in this sense unique: most of the comparable faunas are known from younger rocks in England or Bohemia; a number of the brachiopods are known from ‘island’ faunas, the stratigraphical positions of which are themselves in dispute. Almost all the trilobite genera are making their earliest appearance in the Welsh Arenig, but a few (e.g. Microparia, Ormathops) are known from even earlier occurrences in the Montagne Noire, southern France. Correlation is based on a few more widespread species, and the inferences are particularly indirect for the earlier part of the Series. In general, though, it is apparent that our complete biostratigraphical sequence shows that the type Arenig is compar- able in development with other fully developed sequences of Arenig age in Scandinavia, Spits- bergen and the western United States. Moridunian The base of the Moridunian Stage (not yet formally defined) is also the base of the Arenig Series; its international correlation is difficult. The Moridunian of south Wales (Fortey & Owens 1978) includes a trilobite fauna with species endemic to Wales, and the occurrence in the Carmarthen Formation of Pseudophyllograptus densus, a long-ranging species in Scandinavia, is not diagnostic. The faunas from the Henllan Ash in the type area of Arennig Fawr (Whittington 1966, Zalasiewicz 1984a) are the same endemic suite as the Moridunian of south Wales, and the graptolites recorded by Zalasiewicz (1984a, b) are not stratigraphically unequivocal. Correlation of the early Moridunian in north Wales with the deflexus Biozone of the Lake District is based on the presence at Arennig Fawr of a didymograptid of deflexus type (Zalasiewicz 1984a), a typical morphology for that horizon in the Lake District (Jackson 1962), but since on a world scale graptolites of the same general type have a long range through the Arenig (Cooper & Fortey 1982) this is not entirely satisfactory. The earliest described Arenig graptolite fauna from the Lake District (Jackson 1979, who states it is ‘slightly older than the deflexus’ Biozone) may correlate with the T. approximatus Biozone of Scandinavia, because Jackson (1979: 29) records a distinctive species otherwise only known from that horizon in Scandinavia. Rushton (1985b) has recently described a late Lancefieldian graptolite from below, and in stratigraphical continuity with, Jackson’s fauna, and so the evidence for a complete Tremadoc to early Arenig sequence in the Lake District appears to be increasing, notwith- standing that earlier reports of T. approximatus itself from the Lake District are unconfirmed or erroneous (Stone & Rushton 1983). In Sweden, T. approximatus occurs within the phyllograp- toides Biozone, low in the Hunneberg Series (Tornquist 1904: 6); the base of the Moridunian may, therefore, eventually prove to equate with the base of the Hunnebergian if the former is defined within the Lake District sequence. It is usually stated that equivalents of the T. approximatus Biozone are missing from Wales (Skevington 1969). The T. approximatus fauna is widespread in the ‘Pacific’ graptolite province, and has been suggested as a widely correlatable base for the Arenig Series (Skevington 1963, 1968). However, the possibility cannot be excluded that equivalents of at least part of the T. approximatus Biozone are present in the basal Arenig in Wales; the described graptolites are not sufficiently diagnostic to disprove this (Zalasiewicz 1984a). In addition, we note: 1. Faunas with Merlinia selwynii are underlain in south Wales by the Ogof Hén Formation with M. murchisoniae. It is accordingly possible that these basal beds are referable to an earlier stratigraphical horizon, and hence possibly as early as the range Biozone of T. approximatus. 2. Although there was a regression at the end of the Tremadoc, and the Arenig is transgressive at its base, there is little physical evidence for a break in the Arennig Fawr area (although R. A. FORTEY & R. M. OWENS 100 SIJIWIS-A “9 snjeiydesgipse} “Ss snjeiaaiqge snyeuijoas “1 SNX|/JOsJa4 “SIUdIY 3dA} 94} JO UOT}R[IIION [BUOT}eUIN}UT snjewsxoiddeg SIWIOSIIN 0) ueiuisse snxajjap _Snoi9ans “yy, opunsty “49 snjyejuapouirs ueeule A snjejuaposjsne snjiajuoo “vy Biaqouuny snoi4yeg snsuap uabulig snjebuoja niyosisnbue opunsiy euojsawiy eyequiy, snjejuap sauozolg 9}1]0}de16 elwayog sauozolq 9ay10}de1b6 BUIYD “MS VOIGAWYV N IS9}110}deu6 | Avous | NAGSMS éesnyewsixosddesg snsoolns snpijiqojoid snpisig snjdesBbos; snjejnoejual SauOoZOIq 9a}1oy}desH QOUIAOIdg DIDeY IT ‘3tq snxajjap snpiqiu snjnsaqqib opunity S9UO0ZOIG 9}1}0}de416 JOUISIQ sye] uAMAaS soyeAys suoijhqge fuoqlysn4s euabiaal peuijap jaA jou (S8IEM 'S) seuozolg OUgojla | PLOW ueiun UeIPURL UM > me) m z .) S| ueIUUS NYIANVT1 ARENIG IN SOUTH WALES 101 elsewhere in north and south Wales it is an unconformity), the boundary between the Trema- doc and Arenig being marked only by a thin, sandy seam. 3. The distribution of the T. approximatus plexus of species with H-shaped rhabdosomes may have been related to palaeolatitude and bathymetry. Its known stations are from sites which have been described as lying near the palaeoequator or at temperate latitudes in the earlier Ordovician: North America (Quebec, Newfoundland, Idaho, Nevada, British Columbia, Yukon, etc.), Australia, New Zealand, arctic Siberia, eastern China, cordilleran Argentina, Scotland and Scandinavia. It is possible that the approximatus group failed to penetrate very high latitudes near the Ordovician pole, and particularly more inshore facies, and this might explain its absence from the English Lake District, Bohemia, France, and other areas in boreal Gondwana. If this is the case its absence may be related to palaeolatitude and water tem- perature. The higher part of the Moridunian (Biozone of M. rhyakos) presumably correlates approx- imately with some or all of the higher part of the Bendigonian Stage of Australia, and possibly with the T. fruticosus Biozone of North America (Cooper & Fortey 1982: fig. 2), but detailed correlation is impossible without graptolitic evidence. Whitlandian There is little evidence bearing on the international correlation of the Whitlandian, other than its intermediate position between the Moridunian and Fennian. It is broadly the equivalent of the D. nitidus Biozone of the Lake District. None of its characteristic trilobite fauna has been recorded outside Wales, although it is widespread and in places abundant there. Once again this is probably the result of the unique development of biofacies in Wales at this time, because some of the characteristic forms (Ogyginus, Cnemidopyge, Cyclopyge) are more widespread in the Llanvirn or later. Nor do the small numbers of graptolites so far recovered help much, because Azygograptus hicksii is unknown outside Wales, Tetragraptus serra has a long range, and D. (Expansograptus) goldschmidti, sensu Kraft is otherwise only known from the Klabava Formation of Bohemia. Based on the faunas below and above, we may suggest a general correlation with the Chewtonian (and probably early Castlemanian Cal) of the Australian sequence, and the D. ‘protobifidus’ and D. bifidus Biozones of North America, (?all) Billingenian of Sweden, and ‘Middle Arenig’ in the usage of Cooper & Fortey (1982: 168). Fennian The correlation of the later Arenig is somewhat easier than in the case of the Moridunian or Whitlandian because of the occurrence of more widespread, and stratigraphically reliable, graptolite species; it includes the equivalents of the J. gibberulus and D. hirundo Biozones. Isograptid species from Wales (Jenkins 1982) are all from Fennian localities; based on the work of Cooper (1973), these isograptids are the basis of correlation with the upper Castlemainian (Ca3) to Yapeenian of the Australian sequence, the Isograptus Biozone of North America (Berry 1960), and the ‘upper Arenig’ of the Pacific Province in general (Cooper & Fortey 1982: 168). Other graptolites confirm this. For example, Pseudotrigonograptus (quadriserial) has been recovered from the Pontyfenni Formation, and is known in many localities (Australia, New Zealand, U.S.A., Spitsbergen, Tamir, SE China) from Castlemainian (Ca2—3) and Yapeenian. So the general correlation of the Fennian is not in doubt. What is more difficult is the precise correlation of its limits, and of the biozones within it. Because the D. levigena and B. rushtoni trilobite biozones equate with all of the hirundo and at least part of the gibberulus graptolite biozones (see above), it is assumed that the early Fennian is the equivalent of the earliest part of the Isograptus Biozone with I. victoriae victoriae (Ca2 of the Australian sequence), and that the boundary between the Whitlandian and Fennian approximates to that between the informally- named ‘Middle’ and ‘Upper’ Arenig of the Pacific graptolite province (Cooper & Fortey 1982). This provisional conclusion may have to be modified when more graptolites are discovered. 102 R. A. FORTEY & R. M. OWENS The appearance of Llanvirn pendent didymograptids stratigraphically shortly after the appearance of true (diplograptid) biserials in the latest Arenig appears to be a widespread phenomenon; it happens in Scandinavia (Skevington 1965) and SE China (Mu et al. 1979), and so far as can be judged from the other graptolite species the appearance of these pendent didymograptids is coeval. Bohemia Bouéek (1973), and, in several papers, Kraft (1971, 1972, 1973, 1974) have discussed the grapto- lites and correlation of the Arenig of Bohemia. The Klabava Formation there is divided into several biozones; the oldest of these, the v-similis Biozone, was correlated by Bou¢éek with the balticus Biozone of Scandinavia and the deflexus Biozone of the Lake District. The grounds for this are obscure, other than the occurrence of numerous deflexed graptolites at this horizon. The species listed by Bouéek (1973: 138) are D. (Corymbograptus) v-similis, D. (C.) uniformis and D. (Expansograptus) densus. Of these, the first and last named are so far endemic to Bohemia; D. uniformis is known otherwise from the Lake District, from the D. nitidus or I. gibberulus Biozones. We have identified D. goldschmidti, sensu Kraft 1977, from the Whitlandian and earliest Fennian in south Wales. What evidence there is from specific comparisons tends to suggest that the early part of the Klabava Formation is mid- to early late Arenig in our usage. The trilobites from the middle to upper part of the Klabava Formation (‘Euloma Shales’) include Microparia broeggeri and an Illaenopsis, which compare with the fauna from the Pontyfenni Formation in south Wales of late Arenig age. Bouéek’s youngest graptolite biozone (T. reclinatus abbreviatus Biozone) includes further endemic graptolites, but also “Phyllograptus’ (presumably Pseudophyllograptus) angustifolius, Azygograptus suecicus, and Tetragraptus pseu- dobigsbyi, all of late Arenig age elsewhere. Boucek also mentions an unchanging dendroid fauna running throughout the Bohemian succession. It seems to us probable that the Arenig succession in Bohemia is incomplete. Synoptic history of the Arenig in south Wales The regression at the end of the Tremadoc brought shallow-water Neseuretus facies across the entire south Welsh region. Whether or not there was an unconformity on the Tremadoc beneath is not yet certain, but it is certain that a previously accepted view of uplift during Tremadoc times is incorrect, because rich Tremadoc faunas are present in the Llangynog area (Owens et al. 1982), and these are of a more oceanic character than those of Shropshire. The Tremadoc history further to the west will depend on a reassessment of the age of the so-called ‘Lingula Flags’ there. The subsequent transgression during the Moridunian was gradual, recorded in two upward-deepening successions at Carmarthen and St David’s, in which flaser- bedded silts and sands with numerous trace fossils grade upwards into mudstones (Bates 1969, Fortey & Owens 1978). Locally, as at Llangynog, Precambrian rhyolitic islands were onlapped during the transgressive phase, and the fringing seas afforded suitable habitats for bivalves and echinoderms (Paul & Cope 1982) as weil as more widespread trilobites and brachiopods. Basal beds in these areas may include coarse rhyolite-pebble conglomerates. The transgression pro- ceeded from west to east, because it is probable that the earliest part of the Moridunian is absent in Shropshire. The shallow-water and presumably diachronous Neseuretus facies was succeeded by the raphiophorid biofacies (Fortey & Owens 1978) representing soft, muddy sea floor conditions which supported great numbers of Merlinia selwynii, rarer Ampyx cetsarum, and infaunal bivalves, but little else. Rocks deposited in this environment were thick and widespread around Carmarthen, but their westward equivalents are probably much thinner shales. Open ocean connections were sufficient for the appearance of a few graptolites (Pseudophyllograptus) at this time at one horizon, but a fully oceanic suite of species is not yet known. In the later Moridunian a positive fault-bounded block (or blocks) was becoming established in the Haverfordwest area. This may have contributed to turbidites in the Carmarthen area. ARENIG IN SOUTH WALES 103 FENNIAN Haverfordwest SW. St. David’s Whitland Carmarthen Shropshire NE uniform pelagic sedimentation thin development . P In west Atheloptic benthic assemblage WHITLANDIAN Whitland Carmarthen Shropshire dendrold’ ‘gardens_ In Sey interbeds gravity. flow SSS AfonFfinnant Fm, with allochthonous turbidites and mudstones “brachiopods ‘with A t apes Fm, ave MORIDUNIAN may have been land at this time St. David's Probable barrier to Ramsey Island oceanic circulation Carmarthen area Shropshire Olenid basin with restricted circulation leseuretus facies Raphiophorid facles Fig. 12 East—west profiles from Shropshire across south Wales, to summarize distribution of sedi- mentary environments and associated faunas for the three stages of the Arenig. 104 R. A. FORTEY & R. M. OWENS The effect of this seaward barrier combined with the continuing transgression was to produce a stagnant basin with restricted oceanic circulation in the Carmarthen region. The low bottom- water oxygenation was well suited to the olenid trilobites (olenid biofacies), which gradually become the dominant element of the faunas in the upper part of the Carmarthen Formation (Cwmffrwd and Cwm yr Abbey Members). The westward equivalents of these deposits are not certainly identified. It is clear that the area of thick sediment fill lay to the east, and this may have been associated with an extensional basin. The early Whitlandian is a dominantly clastic facies with fine silts, and volcaniclastic rocks including fragments much like the Treffgarne volcanics, forming a distinctive lithofacies extend- ing from Abercastle on the coast to an area north of Haverfordwest. Similar rocks in the Whitland district contain clasts of Precambrian rhyolites and quartz. ‘Island’ faunas were extensively developed, with a rich fauna of articulate and inarticulate brachiopods, together with cystoids in the shallower sites, often now found as allochthonous material in mass flow deposits and turbidites which comprise the thick sequence of the Blaencediw Formation north of Whitland. The early Whitlandian marks the end of the restricted olenid basin in the east, with the complete disappearance of the characteristic trilobite family within a few metres of the base. During the Whitlandian, conditions suitable for flourishing ‘gardens’ of dendroid grapto- lites were widespread: they appear within the well-laminated siltstones separating allochthon- ous brachiopod-rich beds in the Haverfordwest area, at Blaeweneirch near Whitland, and at St David’s; the last two localities furnished many of the specimens monographed by Bulman (1927-34). The eastward equivalents of the Blaencediw Formation are to be found in the Afon Ffinnant Formation—a thick sequence of turbidites and interbedded shales. Again, the main sediment fill appears to have been in the Carmarthen—Whitland area rather than to the west, where evidence of turbidites is confined to fine distal turbiditic horizons usually on the scale of a few mm, forming light-coloured stripes within the Penmaen Dewi Formation. The character- istic trilobite fauna with Ogyginus, Bohemopyge, Shumardia gadwensis, Furcalithus sedgwicki and Gymnostomix gibbsii spreads across all south Wales during the Whitlandian, but is nowhere abundant. Cyclopygids, especially Cyclopyge grandis grandis, appear in numbers for the first time in the south Welsh Arenig, and graptoloids, Azygograptus, Expansograptus and Tetragraptus, are locally numerous—the first suggestion of a more ‘oceanic’ character in the faunas. The late Arenig, Fennian Stage, begins with a local turbiditic interval in the Whitland area (Cwmfelin Boeth Formation), which yields the youngest of the three faunas of articulate brach- iopods (here allochthonous) in south Wales. Its equivalents to east and west have not been identified. But the succeeding Pontyfenni Formation represents the most uniform, and most oceanic, conditions over the whole area. Dark mudstones with siliceous nodules are character- istic, and this lithofacies with its accompanying fauna outcrops from east of Carmarthen to the extreme west of the area on Ramsey Island. The trilobite fauna is distinctive, a combination of large-eyed pelagic species dominated by the Cyclopygidae, with blind, or nearly blind, benthic forms which we have termed (p. 106) the atheloptic assemblage. We believe that this formation accumulated beneath a water depth of at least 200m and probably more. Its oceanic character is indicated also by the occurrence of cosmopolitan graptolites of the genera Pseudotrigono- graptus and Isograptus (Jenkins 1982) which do not penetrate into shallower shelf seas (Fortey 1984). The trilobite fauna is rich in species but sparse in numbers of individuals, although some compensation for this is the high proportion of articulated specimens, often moulted arrange- ments of parts, which attests to the quiet sea-floor conditions that pertained during the Fennian. Trilobites are accompanied by a varied fauna of carpoid chordates, including the genera Cothurnocystis, Balanocystites, Anatifopsis, Guichenocarpos, Reticulocarpos and Lagy- nocystis (see R. P. S. Jefferies herein, p. 285), many hyolithids, and rare, soft-bodied coel- enterates. This peculiar facies has its earlier counterpart in the lower Arenig of the Montagne Noire, and appears in the Llanvirn of Bohemia and Shropshire, then (Fortey 1984) attaining a wider spread later in the Ordovician with the Llandeilo—Caradoc transgression. The Pontyfenni Formation is its only known occurrence in the late Arenig. Several of its characteristic species have been found also in the Lake District. ARENIG IN SOUTH WALES 105 Upward passage into the Llanfallteg Formation is marked by a lithological change to grey-coloured, very soft and fissile shales, and with this change the atheloptic trilobite assemblage disappears, although the Cyclopygidae do not. We attribute the faunal change to a worldwide regression at the end of the Arenig, which reduced sea level enough to permit the establishment of a trilobite fauna including species with normally-developed eyes, of which Barrandia is the most numerous. There is no change in lithology, and no drastic faunal change, across the Arenig—Llanvirn boundary, which is best exposed along the Llanfallteg railway cutting. Pendent didymograptids do appear suddenly, however, and this is the criterion for recognition of the boundary, together with the first appearances of a few trilobite species. The Llanfallteg Formation is also widely developed, from east of Carmarthen to Aber Mawr on Ramsey Island. As usual, it is thinner to the west. Volcanic activity was prevalent in the early Llanvirn, as at Fishguard, and volcanogenic sediments may partly account for the lithological change near the Arenig—Llanvirn boundary. Shortly above our studied sections there are numerous tuffaceous ‘chinastone’ horizons. The ensuing Llanvirn transgression is accompanied by a reversion to deep-water sedimentation, with black shales, mudstones and turbidites which once more comprise the dominant lithology from Carmarthen to the coast. In summary, the Arenig Series in south Wales records an upward-deepening and then ulti- mately regressive sequence much complicated by local uplift and turbidite emplacement. It includes the best-developed deeper water fauna on the edge of the Gondwanan continent at this time, in a marginal but ensialic basin. Sedimentation is particularly thick and complete in the Carmarthen—Whitland area, thinner near the coast. Although the east-west transect is now described, little is known about what happens to the north, where the stratigraphy of the Sealyham Group and the shales of the Maenclochog area have yet to provide the fossils crucial to unravelling the stratigraphy. The cyclopygid biofacies and atheloptic trilobite assemblage The Arenig of south Wales includes the earliest known development of the cyclopygid biofacies. The Cyclopygidae were a group of trilobites with hypertrophied eyes, wide axial development, and other morphological features consistent with a pelagic mode of life (Fortey 1974, 1985b). They were not, however, epipelagic like their homoeomorphs Carolinites and Opipeuter, both of which penetrated into inshore sediments in the earlier Ordovician around the palaeoequator. Cyclopygids are rare or absent in inshore facies, for example, the Neseuretus biofacies in the early Ordovician (Fortey & Morris 1982). This is shown by their rarity in the Mytton Flags in Shropshire compared with the contemporary rocks in south Wales. Diverse and numerous cyclopygids are associated with peripheral continental sites with free access to the open ocean. For the Arenig around Ordovician Gondwana the appropriate sites are really only described on its eastern margin in south and north Wales and the Lake District. There is also some evidence of a like development on the opposite side of T6rnquist’s Ocean (Tjernvik 1956, Cocks & Fortey 1982) where the cyclopygid biofacies is developed on Bornholm, and encountered in boreholes off the southern edge of the Scandinavian platform; it may be present this early at the eastern edge of Kazakhstania (Apollonov 1975). By the Llanvirn there was a general shoreward onstep of exterior biofacies accompanying the Llanvirn transgression, and this introduced the cyclopygid biofacies into Bohemia (Sarka Formation), Thuringia (Richter & Richter 1954), Bulgaria (Spassow 1958) and the Hope Shales, Shropshire (Whittard 1961a). The cyclopygids were probably mesopelagic (Fortey 1985b; see also p. 180), occupying the depth zone of about 200-700 m and living about the upper part of the continental slope and within deep marginal basins with free access to the open ocean. They constituted a natural community, but one which was unrelated to the sea bottom conditions where the sediments in which they are found were accumulating. They can be found with more than one assemblage of benthic genera; this is shown in south Wales by the different associations with which they are found in the Pontyfenni Formation (with blind or small-eyed species) and in the Llanfallteg Formation (with normal-eyed trilobites including Barrandia and Ectillaenus). For this reason it 106 R. A. FORTEY & R. M. OWENS is difficult to apply the term ‘community’ to the total assemblage, including the cyclopygids, as it is recovered in the field. We refer to the faunal associations in which cyclopygids are a dominant element by the broad term ‘cyclopygid biofacies’, recognizing that this includes a blend of the pelagic with the benthic elements, and as such is a superimposition of more than one community in the strict sense. It is important to emphasize that the recognition of the cyclopygid biofacies depends on the family being an important fraction of the total trilobite fauna—30% or more, as in the Pontyfenni Formation. The recovery of a single specimen in a large non-cyclopygid fauna does no more than suggest that it may be worthwhile searching for the cyclopygid biofacies in adjacent areas. As well as the cyclopygids, the pelagic fauna certainly included the peculiar trilobite Bohemilla—a genus also making its first known appearance with the cyclopygid biofacies in south Wales—and the enigmatic Girvanopyge (= Cremastoglottos). The odontopleurid Seleno- peltis is more difficult to interpret. It is found with the cyclopygid biofacies, but also extends (Bruton & Henry 1978) into areas in France and Spain where cyclopygids are rare or unknown. Nowhere is it common. It has large eyes, but otherwise its morphology is unlike that of a typical pelagic trilobite, being relatively broad and with a narrow axis. Its wide tolerance of facies changes might suggest that it alone was epipelagic or epiplanktonic, the alternative being that it was nektobenthic with a wide range of tolerance. It is interesting that Whittington & Hughes (1972) used Selenopeltis as an ‘index fossil’ of the high latitude Selenopeltis province, and if the animal indeed had epipelagic habits one would expect it to be a good indicator of temperature (and hence broadly latitudinal) constraints. There is an interesting association, particularly in the Fennian, of the mesopelagic trilobite fauna with a benthic fauna of blind or nearly blind trilobites. These include Ormathops nichol- soni, Illaenopsis harrisoni, Dindymene saron, Colpocoryphe taylorum, Ampyx linleyoides and Bergamia rushtoni. This is evidently different from what Fortey & Owens (1978) termed the ~ raphiophorid community, in which normal-eyed asaphids predominate, and which replaces the Neseuretus fauna in deepening-upward sequences. Equally it is distinct from the olenid biofacies—not one olenid has been found above the Whitlandian. If the olenid biofacies is a response to low oxygen tension (and relatively deep water) to which the olenids were specifi- cally adapted, then this other benthic association is different. In support of normal oxygenation is the fact that the small-eyed fauna is associated with other benthic organisms—bivalves, carpoid chordates and hyolithids—none of which occurs in the Carmarthen Formation with the olenids. The associated lithologies are dark mudstones with siliceous nodules, and a similar lithology is also associated with Llanvirn occurrences of the cyclopygid biofacies. More than half the trilobites recovered are more or less articulated, often in moult arrangements, which testifies to relatively quiet bottom conditions. Certain beds show much evidence of burrowing by soft-bodied animals, and we presume that the sea bottom was an unconsolidated mud. We term this distinctive type of benthic trilobite association the atheloptic (Gr. ‘shrunken-eyed’) assemblage. It probably represents a genuine community, but without further documentation of its occurrence elsewhere we prefer the noncommital term ‘assemblage’ for the moment. The depth at which the atheloptic assemblage lived can be estimated from two lines of reasoning. First, the geological occurrence is in an exterior site, well removed geographically from the shelf faunas of the Mytton Flags. Within the column it lies far above the transgressive base of the Arenig, and includes no elements of the Neseuretus community and none of the asaphids abundant in what we (Fortey & Owens 1978) termed the raphiophorid community; the only genus in common is Ampyx, a large genus containing groups of species that may be only distantly related (A. linleyoides from the Fennian is not closely related to A. cetsarum from the Moridunian). If the Arenig represents a deepening-upwards sequence in south Wales, as we propose, then the atheloptic assemblage lies at the deepest end of the gradient. The second approach derives from the morphology of the trilobites themselves by compari- son with modern analogues. Clarkson (1967) pointed out that crustaceans which are blind or with atrophied visual organs tend to be commoner below about 600 m water depth. In oceanic waters sunlight can penetrate at very low intensities to 600-650 m (Boxshall 1981: 152), but in more turbid epicontinental seas to less than half this depth. Since the cyclopygids may have occupied the depth zone of 200m and below, it seems reasonable to infer a depth of 300m or more for the atheloptic assemblage. ees ty LEVEE Cee | ais ey a Se b Fig. 13 Atheloptic assemblage of blind, or nearly blind, benthic trilobites (af), together with free-swimming, large-eyed mesopelagic forms (g-i). a, Illaenopsis; b, Shumardia; c, Bergamia; d, Colpocoryphe taylorum; e, Ormathops nicholsoni; f, Ampyx; g, Pricyclopyge; h, Bohemilla (Fenniops); i, Degamella. The regression near the Arenig—Llanvirn boundary presumably reduced the water depth sufficiently for normal-eyed trilobites to thrive during the deposition of the Llanfallteg Forma- tion. Since the cyclopygids are also numerous there the depth could not have been less than about 200m. The change during the regression need not have been very great if the atheloptic assemblage lived just below, and the normal-eyed assemblage just above, the critical depth of light penetration. Because they were inhabitants of the water column well below the surface and somewhat insulated from the constraints of surface temperature conditions, the cyclopygids are not confined to the Gondwanan plate in the early Ordovician, as are the shallow shelf genera such as Neseuretus. They were capable of living in the appropriate sites at former temperate latitudes around Baltica and Kazakhstania, as indicated above. They did not penetrate into tropical latitudes at that time, however. One of us (RAF) has made an intensive search in the appropriate lithology in the Cow Head Group of western Newfoundland, which accumulated off the shelf of the North American continent during the early Ordovician. No fossils of cyclopygids were discovered. The oldest occurrences in North America are Middle Ordovician: the earliest of these is probably from the Normanskill Shale (Niobe? huberi Roy, 1929 is likely to be a Degamella sp.), while typical cyclopygid biofacies are known from the Ashgill of Quebec (Cooper & Kindle 1936) and the Whitehouse Formation in the Girvan district of Scotland. The early history of the group is associated with higher latitudes in the early Ordovician, and with a distribution mutually exclusive from that of the tropical pelagic genera Carolinites and Opi- peuter. The oldest known cyclopygid is Prospectatrix Fortey 1981, from the British Tremadoc, but it is a rare fossil, and its occurrence there is not equivalent to the cyclopygid biofacies of the later Ordovician. Cyclopygids are also known from the latest Tremadoc-earliest Arenig of the 108 R. A. FORTEY & R. M. OWENS Montagne Noire, southern France (Thoral 1935), and it is perhaps here that the earliest cyclopygid biofacies may be sought. The presence in the Montagne Noire of Illaenopsis—a characteristic trilobite of the atheloptic assemblage—and of the same genera of carpoid chord- ates as those from the Pontyfenni Formation indicates that the combination benthic and pelagic faunas typical of the cyclopygid biofacies may have been established before the end of the Tremadoc. Raymond (1925), Weir (1959) and Kobayashi & Hamada (1971) emphasized supposed migra- tions of cyclopygids, basing their evidence on stratigraphical and geographical distributional data. We prefer to regard distribution as the result of the discovery of the appropriate biofacies, with the distribution of the trilobites themselves being effectively simultaneous within the biofacies belt. This is shown, for example, by the similarity or identity of species of the family from China (Zhou 1977), Kazakhstan (Koroleva 1982) and Bohemia (Marek 1961), almost at the extremes of its geographical range, and we suspect that critical revision will show that individual species may be dispersed far more widely than the present taxonomy allows. Dis- covery of the cyclopygid biofacies depends on the preservation of the appropriate former geographical site. In many areas the appropriate exterior site is simply not preserved: for example, there are relatively few places where sufficiently exterior Arenig sites have survived the effects of tectonism, and south Wales is one of the few. Transgressive periods introduced the cyclopygid biofacies onto peripheral shelf regions where it had a higher chance of preservation, as during the Llanvirn and in the later Ordovician (Fortey 1984). Palaeogeographic affinities of the Arenig faunas The position of Wales on the western edge of an early Ordovician Gondwana has been claimed on evidence derived both from faunas (e.g. Cocks & Fortey 1982) and from multidisciplinary studies (such as Ziegler et al. 1979). The Arenig faunas of south Wales confirm this from several aspects. 1. Inshore faunas with the trilobite Neseuretus are only found at the regressive intervals, e.g. in south Wales at the base of the Arenig, but the Neseuretus biofacies is found through most of the sequence in Shropshire, indicating its persistence in shallower clastic facies there. This fauna is to be regarded as one of the more reliable indicators of the former extent of Gondwana (Fortey & Morris 1982) and may be found over a wide area of Armorica, Iberia, northern Africa, South America, eastern Newfoundland, the Middle East, the Himalaya and ultimately Yunnan and other parts of China. Individual species of Neseuretus probably range widely over this area. Most of the records, however, are from Llanvirn or Llandeilo rocks. Wales is the one site peripheral enough to retain a record of this fauna in the earlier part of the Arenig. Wherever the appropriate transgressive facies occurs—in the Fennian of Anglesey, for example—so too does Neseuretus, accompanied by a restricted fauna of brachiopods. In south Wales the appropriate conditions are found at the base of the Arenig sequence (Fortey & Owens 1978) and do not reappear, for the only subsequent record we have of Neseuretus is as a rare occurrence in a siltstone of Whitlandian age. 2. The majority of our trilobite genera, even those occurring in more exterior biofacies, are also confined to Gondwanan occurrences, mostly in younger rocks elsewhere in the appropriate palaeoecological settings. The exception is the olenid biofacies, which includes genera of world- wide distribution and independent of ‘provincial’ or palaeogeographic boundaries (Hypermecaspis, Bienvillia, Porterfieldia). The olenid biofacies was peculiarly adapted to low- oxygen sea floor conditions, and can be found wherever the right conditions pertained. For the rest of the fauna, apart from the endemic trinucleids (which are largely confined to Britain), we list the following genera which are restricted to other Gondwanan occurrences: Corrugat- agnostus, Illaenopsis, Barrandia, Ectillaenus, Ormathops, Colpocoryphe, Dindymene, Merlinia, Ogyginus, Selenopeltis, Placoparia, Bohemilla, Girvanopyge and Bohemopyge. The Cyclo- pygidae, including another ten genera, were also confined to peripheral Gondwana sites and along the poleward edge of Baltica until late in the Ordovician, when they appeared en masse in peripheral sites at the edge of the former North American continent. ARENIG IN SOUTH WALES 109 3. We noted before (Fortey & Owens 1978: 45) that the deeper water facies also included genera which could be found in exterior facies at the edge of other palaeocontinents—in this case outside Gondwana. This is in contrast to the platform Gondwana faunas, such as those described recently by Henry (1980) and Hammann (1983), which consist entirely of Gondwanan endemic genera. Excluding the specialized olenid biofacies this list includes the following genera from the Arenig of south Wales: Leiagnostus, Arthrorhachis, Segmentagnostus, Cnemidopyge, Shumardia, Leioshumardia, Asaphellus, Dionide, Ampyx, Seleneceme. Of these Ampyx, Shumardia and Arthrorhachis are cosmopolitan. Dionide, Seleneceme and Leioshumardia are so far recorded from peripheral sites at the edge of Ordovician North America. The first-named is known from numerous stations in Britain, Bohemia, Armorica etc., while Leioshumardia has only previously been described from western Newfoundland (Whittington 1965). Leiagnostus and Cnemidopyge are known otherwise from Baltica. The fact that close comparisons can be made between species in marginal facies on different continents already by the Arenig—the time usually accepted as one of maximum ‘provinciality—demonstrates the possibility of faunal interchange between exterior facies in advance of major provincial merging. 4. Graptoloids include a number of forms that are confined to peri-Gondwanan graptolitic facies. Azygograptus in the Arenig is one such. It appears as the first graptolite in upward- deepening successions and may have been adapted to epiplanktonic life in less ‘oceanic’ sites. Other Gondwanan forms are: Didymograptus (Expansograptus) uniformis lepidus, D. (E.) sparsus, D. hirundo s.s. and Acrograptus acutidens, and probably the Llanvirn pendents. But other graptolites, like some of the exterior facies trilobites, appear to have been independent of palaeogeography. These belong to what Fortey (1984) termed the ‘isograptid biofacies’, a more exterior suite of species regarded as living at greater depth in the water column. Examples listed from south Wales include Pseudotrigonograptus ensiformis, “Glyptograptus’ austrodentatus, “G. dentatus, Glossograptus acanthus, Tetragraptus serra, T. bigsbyi and Pseudisograptus spp. These bridge the so-called “Atlantic’ and ‘Pacific’ graptolite provinces, and are correspondingly impor- tant for correlation purposes. They can be regarded either as precocious invaders signalling the broadly uniform worldwide graptolite faunas of the Caradoc, or as members of a continuously more uniform oceanic graptolite fauna, a view which we favour from the palaeogeographic and facies evidence. The four lines of evidence are consistent with the inferred position for Wales during the Arenig. Most of the trilobite genera we have found in south Wales only became widespread over Gondwana in the Llanvirn or tater, as the Llanvirn transgression moved shelfwards. A few genera (Asaphellus, Prospectatrix) had persisted from the Tremadoc. There were no genera in common with the cratonic, low latitude faunas of North America or Australia, and very few in common with Baltica, which we regard as having been at temperate latitudes. All the evidence we have is consistent with the view that Britain must have lain at high latitudes in the earlier Ordovician, while the facies distribution proves the marginal position of south Wales relative to that huge area of Gondwana over which Grés Armoricain clastic facies were accumulating. Fossil localities The localities referred to in the text, and on the maps (Figs 1, 2, 4-6) are listed below, and follow on from the list of Fortey & Owens (1978: 240) which gives details of localities 1-15 and 16A-—G. Apart from locality 16, they are arranged geographically from east to west; old Geo- logical Survey localities from which we did not collect are quoted in the text under their original survey designations. 16. Cwm yr Abbey (Afon Ffinnant Formation): 16H 50m at 355° from the road bridge (SN 5002 1983); 16J 60m at 355° from the road bridge (SN 5002 1985); 16K 85m at 356° from the road bridge (SN 5002 1986); 16L 90 m at 337° from the road bridge (tributary to main stream) (SN 4449 1986). 17. Allt Cwm-arbont (Afon Ffinnant Formation): 17A 85m at 108° from Arbont Cottage (SN 5251 1883); 17B 55m at 73° from Arbont Cottage (SN 5238 1891). 110 18. iI), 20. 45. 46. R. A. FORTEY & R. M. OWENS Afon Ffinnant (Afon Ffinnant Formation): 18A 700m at 174° from Pont ar Ffinnant (SN 5099 1942); 18B 642m at 170° from Pont ar Ffinnant (SN 5100 1949): 18C 410m at 156° from Pont ar Ffinnant (SN 5106 1973); 18D 160m at 124° from Pont ar Ffinnant (SN 5103 2003); 18E 150m at 124° from Pont ar Ffinnant (SN 5102 2003); 18F 50m at 136° from Pont ar Ffinnant (SN 5093 2007). Exposure on south side of B4300, 140m west of Pont ar Ffinnant (SN 5078 2006) (Afon Ffinnant Formation) Stream section at Capel-Dewi (Afon Ffinnant? (20F) and Pontyfenni formations (20 A—E)): 20A 137m at 181° from road bridge (SN 4705 2120); 20B 176m at 186° from road bridge (SN 4703 2017); 20C 178m at 186° from road bridge (SN 4703 2017); 20D 188m at 187° from road bridge (SN 4702 2016); 20E 195m at 189° from road bridge (SN 4701 2015); 20F 205m at 190° from road bridge (SN 4701 2014). . Road cutting north side of A40, 100m at 148° from Castell-y-waun, 0:-6km SW of Bancyfelin (SN 3191 1756) (Pontyfenni Formation). . Stream exposure 447m at 133° from Sabulon (SN 2490 1628) (Pontyfenni Formation, ?B. rushtoni Biozone). . Old quarry at Pontyfenni, on north side of old A40 (SN 2379 1690 to SN 2381 1693) (Pontyfenni Formation, B. rushtoni Biozone). . Old quarry 117m at 112° from Llwyn-crwn (SN 2399 1795) (Pontyfenni Formation, B. rushtoni Biozone). . Stream exposure 190m at 256° from Pant-y-grug (SN 2253 1842) (Pontyfenni Formation, B. rushtoni Biozone). . Temporary exposure in farmyard at Regwm (SN 2166 1767) (Pontyfenni Formation, S. abyfrons Biozone). . Old quarry 100m west of Whitland Abbey (SN 2070 1817) (Colomendy Formation, Whitland Abbey Member, G. gibbsii Biozone). . Stream-bed in Allt y Clyn, 800m at 327° from Whitland Abbey (SN 2040 1883) (Colomendy Forma- tion, Whitland Abbey Member, G. gibbsii Biozone). . Stream-bed, Nant Colomendy, 350m at 214° from Pant-gwyn (SN 2074 1949) (Pontyfenni Formation, 2B. rushtoni Biozone). . Trackside quarry 190m at 256° from Blaencediw (SN 2056 2050) (Blaencediw Formation). . Old quarry 330m at 288° from Blaencediw (SN 2043 2065) (Blaencediw Formation). . Exposures on west side of lane NE of Gellidiogyn (Pontyfenni Formation): 32A 140m at 49° from Gellidiogyn (SN 2010 2119); 32B 160m at 39° from Gellidiogyn (SN 2010 2123). . East side of valley, 370m at 41° from Gellidiogyn (SN 2023 2138) (Colomendy Formation, Castell- draenog Member). . Exposures at Castelldraenog: 34A west side of farmyard, 70m at 234° from Castelldraenog (SN 2080 2143) (Colomendy Formation, Whitland Abbey Member); 34B west side of track, 150m at 3° from Castelldraenog (SN 2084 2160) (?Pontyfenni Formation). . Old quarry on west side of valley at Cwmfelin Boeth (SN 1908 1924) (Cwmfelin Boeth Formation). . Stream exposure 150m at 140° from Bryngwelltyn, Cwmfelin Boeth (SN 1940 1940) (Cwmfelin Boeth Formation). . Stream exposure 100m at 136° from Bryngwelltyn, Cwmfelin Boeth (SN 1937 1945) (Pontyfenni Formation, S. abyfrons Biozone). . Roadside section at Pen-y-parc (SN 1981 1951 to SN 1988 1954) (Pontyfenni Formation, S. abyfrons Biozone). . Old quarry on hillside 420 m at 89° from Felin-Henllan-Amgoed (SN 1931 1972) (Blaencediw Formation). . Exposure at Pant, 690m at 152° from Blaen-lliwe (SN 1892 1779) (Pontyfenni Formation). . Lane cutting by old quarry 520m at 154° from Blaen-lliwe (SN 1884 1795) (Pontyfenni Formation). . Exposure in hedge bank 190m at 108° from Blaen-lliwe (SN 1879 1834) (Pontyfenni Formation). . Old quarry on hillside 430m at 236° from chapel at Cwmfelin Boeth (SN 1887 1882) (Cwmfelin Boeth Formation). . Stream bed exposure immediately south of bridge at Rhyd-caer-Emlyn, Henllan Amgoed (SN 1847 2015) (Pontyfenni Formation). Old quarry on east side of valley, 130m at 166° from Felin-Henllan-Amgoed (SN 1892 2058) (Colomendy Formation, Rhyd Henllan Member). Exposures in lane leading to Felin-Henllan-Amgoed (Colomendy Formation, Rhyd Henllan Member): 46A 50m at 238° from Felin-Henllan-Amgoed (SN 1884 2066); 46B 90m at 288° from Felin-Henllan-Amgoed (SN 1880 2072). ARENIG IN SOUTH WALES 111 47. Laneside exposures north of Llwyn-derw (Colomendy Formation, Rhyd Henllan Member): 47A 40m at 55° from Llwyn-derw (SN 1888 2084); 47B 90 m at 38° from Llwyn-derw (SN 1890 2088). 48. Laneside exposures immediately west of Sarn-las (Pontyfenni Formation): 48A 100m at 302° from Sarn-las (SN 1735 1878); 48B 70m at 248° from Sarn-las (SN 1738 1871). 49. Temporary exposure at Hendref, Henllan Amgoed (SN 1772 2042) (Pontyfenni Formation). 50. Exposure at Cefn-maen-llwyd, Rhyd-y-wrach (SN 1679 1976) (Llanfallteg Formation, D. artus Biozone). 51. Old roadside quarry 200m at 212° from Cwmmiles (SN 1608 2224) (Pontyfenni Formation). 52. Old railway cutting north of Llanfallteg (Pontyfenni (52Z) and Llanfallteg (52, D. artus Biozone, 52A-Y, D. levigena Biozone) formations): 52 small quarry at south end of cutting (SN 1571 2013): basal 3m of Llanvirn. 52A—S east side of track, south of core of anticline (SN 1571 2014 to SN 1575 2019), stratigraphically below the base of the Llanvirn as follows: 52A, 0-1-5m; 52B, 1-5—3m; 52C, 3-4 m; 52D, 5m; 52E, 6m; 52F, 7-9 m; 52G, 9-11 m; 52H, 11-13 m; 521, 13-15 m; 52J, 15-17 m; 52K, 17-19 m; 52L, 19-21m; 52M, 21-23m; 52N, 23-25m; 52P, 25-27m; 52Q, 27-29m; 52R, 29-31 m; 52S, 31-33 m. Positions shown on Fig. 8, p. 91. 52T, core of anticline, 33m below base of Llanvirn (SN 1575 2020). 52U, immediately north of core of anticline, 31-33 m below base of Llanvirn (SN 1576 2021). 52V (SN 1579 2024), 52W (SN 1583 2028), 52X (SN 1587 2037), S52Y (SN 1588 2039): separated by structural complexities from southern part of section, and distance below Llanvirn boundary unknown. 52Z by entrance to narrow track leading from path of railway (SN 1591 2057). 53. Section on north side of road, 340m at 124° from Brynaeron, Llandissilio (SN 1295 2101) (Pontyfenni Formation). 54. Old quarry 520m at 154° from Brechfa, Llanycefn (SN 0994 2168) (Pontyfenni Formation). 55. Long Plantation railway cutting, Scolton (SM 9916 2152 to SM 9930 2153) (Llanfallteg Formation, D. artus Biozone). 56. Triffleton Quarry (SM 9774 2428) (Blaencediw Formation). 57. Brunel cuttings, east side of Treffgarne Gorge (SM 9602 2448 to SM 9603 2470) (Blaencediw Forma- tion at north end; ?Tremadoc at south end). 58. Old quarry at Abercastle (SM 8531 3355) (Abercastle Formation). 59. Exposure on east side of Porth-gain (SM 8143 3262) Gunction of Abercastle Formation (Porth Gain Member) and Penmaen Dewi Formation). 60. Small quarry 500m at 356° Lleithyr, north of St David’s (Ogof Hén Formation). 61. Exposures in Pwlluog, north end of Whitesand Bay (61A, B: southern inlet; 61C, E: northern inlet, beach level; 61D, F: northern inlet, cliff top; 61G: north side of Craig y Creigwr) (all Penmaen Dewi Formation) (61A SM 7328 2738, 61B SM 7322 2747, 61C SM 7320 2746, 61D SM 7322 2753, 61E SM 7312 2758, 61F SM 7317 2760, 61G SM 7302 2759). 62. Ramsey Island: Aber Mawr, beach level (SM 7005 2425) (Pontyfenni Formation equivalent). 63. Ramsey Island: Cliff top at Ogof Hén (SM 7080 2520) (Ogof Hén Formation). Collections and repositories The trilobites and graptolites described herein were largely collected by the authors over the ten year period from 1974 to 1983. They are supplemented by existing museum collections, especially those of the British Geological Survey, and by specimens kindly presented by numer- ous individuals, listed in the acknowledgements. Figured and cited specimens are housed in the following institutions, except where stated otherwise: British Museum (Natural History) (register number prefixes I, Q, E, H and It), British Geological Survey, Keyworth (BGS and/or GSM), National Museum of Wales (NMW) and Sedgwick Museum, Cambridge (SM). Systematic descriptions: Trilobites by R. A. Fortey and R. M. Owens Trilobites are described family by family in the same order as in the Treatise on Invertebrate Paleontology (Harrington et al. 1959), with the exception that the Nileidae are classified with the Cyclopygidae in the superfamily Cyclopygacea, as proposed by Fortey (1981). Terminology generally also follows the Treatise except that the glabella is understood to include the occipital ily R. A. FORTEY & R. M. OWENS ring. The use of the term baccula (-ae) for swellings at the base of the glabella follows Opik (1967) and Fortey (1975). We use the term axial shield (Henningsmoen 1957) for moulted exoskeletons consisting of cranidium + thorax and pygidium. Family AGNOSTIDAE Salter 1864 Genus LEIAGNOSTUS Jackel 1909 TYPE SPECIES. Leiagnostus erraticus Jaekel 1909, by monotypy. Leiagnostus cf. erraticus Jaekel 1909 (Figs 14a, b) MATERIAL. Imperfect dorsal exoskeleton, BGS Pr557 + Pr581; pygidia: It.19560, NMW 84.17G.30. STRATIGRAPHICAL RANGE. Upper Arenig, Fennian, zone of Bergamia rushtoni, Pontyfenni For- mation. LocaLitigs. Pontyfenni (loc. 23), and Cwmmiles (loc. 51). Discussion. So featureless an agnostid does not require description of the general morphology. One pygidium (Fig. 14b) shows details of the pygidial axial musculature. The axis appears to be subfusiform, with a terminal tubercle at about two-thirds pygidial length; the median tubercle is far forward and is flanked by two pairs of muscle impressions; two further pairs lie behind these. The articulating half ring is well shown on the large pygidium of Fig. 14a; it is short (sag.) with a narrow ring furrow, which is not deepened into pits at its lateral extremities. Using Opik’s (1967: 71-72) categories of articulating devices, that of Leiagnostus appears to be of the basic (peronopsid) kind. The articulating device of Metagnostidae (Geragnostidae of authors, see Fortey 1980) is of the glyptagnostid kind: for example, Geragnostus (Tjernvik 1956: pl. 1, fig. 10), Arthrorhachis (Fortey 1980: pl. 2, fig. 15) and Galbagnostus (Whittington 1965: pl. 4, fig. 7). Hence Leiagnostus is not likely to be an effaced metagnostid if this criterion is of phylogenetic importance. Effacement itself is highly polyphyletic in the Agnostida, and Howell’s (1935) family Leiagnostidae based on this aspect of morphology is clearly unsatisfactory. That metag- nostids did become almost as effaced as Leiagnostus is shown by highly effaced Geragnostus from Bohemia termed Neptunagnostella by Pek (1977); the articulating ring remains long (sag.) Fig. 14 Leiagnostus cf. erraticus Jaekel 1909. Pontyfenni Formation, late Arenig (Fennian), Bergamia rushtoni Biozone. a, large pygidium, loc. 23, x 8, It.19560; b, cephalon and pygidium, the latter showing muscle impressions on internal exoskeletal surface, loc. 23, x 8, GSM Pr581. | ARENIG IN SOUTH WALES 113 in these forms (Pek 1977: pl. 3, fig. 3). So on the strength of the articulating device we place Leiagnostus in the Agnostidae, following Opik’s (1967) classification. The pygidial musculature is not particularly helpful because indications of four pairs of impressions are to be found on both Agnostidae (e.g. Westergard 1946: pl. 16, fig. 16) and Metagnostidae (e.g. Pek 1977: pl. 3, fig. 4). The type species of Leiagnostus, L. erraticus Jaekel, is known from the holotype, a complete enrolled specimen from geschiebe material probably originating from the uppermost Arenig or early Llanvirn of Sweden. This specimen was re-illustrated by Neben & Krueger (1971: pl. 11, figs 37-38). The cephalon lacks a border, which is wide on the pygidium, itself only about 3mm long. Our large pygidium is twice this size, and differs from the type of L. erraticus in being almost parallel-sided with the maximum width behind the mid-length. A smaller pygidium from Wales (Fig. 14b), which is only a little larger than the type, resembles it in general proportions and in having an oval outline, so it is likely that small changes in pygidial outline accompanied continued growth. The pygidial tubercle does not show up on Jaekel’s original; this may be because it is covered with cuticle in this region, and the tubercle may be mostly a thinning of the exoskeleton. The Welsh specimens are crushed, and largely show parietal surfaces on which the tubercle is visible, so this difference may be less than it seems, but caution dictates that we should qualify our determination. The development of the tubercle is variable in other species; in L. bohemicus from the Llanvirn of Bohemia it varies from a prominent bulge to a minute, faint pustule (Pek 1977: pl. 5, figs 1-5). It is in a more posterior position than in L. cf. erraticus—at a fifth to a quarter of pygidial length (excluding half ring), but at one seventh or less of pygidial length on the Fennian specimens. The same distinction applies to L. franconicus Sdzuy 1955 from the Tremadoc of Germany, and to L. cf. turgidulus from the Tremadoc of south Wales (Fortey & Owens, in Owens et al. 1982: pl. 1, fig. b). L. turgidulus Harrington & Leanza 1957 from the Tremadoc of Argentina lacks a defined mid-axial tubercle but the postaxial tubercle is clearly visible, indicating that the axis in this species extended much further backwards than on L. cf. erraticus. The pygidial border on L. peltatus Tjernvik 1956 from the early Arenig of Sweden is extremely narrow. L. alimbeticus Balashova 1961 is known from cephalic parts only, and cannot be properly evaluated in a genus in which pygidial details are essential in discriminating species. L. foulonensis Howell 1935 appears to show a cephalic border as well as a broad pygidial one, and is therefore excluded from Leiagnostus. Family METAGNOSTIDAE Jaekel 1909 (= Geragnostidae Howell 1935; Trinodidae Howell 1935; Arthrorhachinae Raymond 1913). Genus CORRUGA TAGNOSTUS Kobayashi 1939a TYPE SPECIES. Agnostus morea Salter 1864 (see Whittard 1955: 11). Corrugatagnostus cf. refragor Pek 1969 (Figs 15a—c) ef. 1969 Corrugatagnostus refragor Pek: 383; pl. 1, fig. 1. cf. 1977 Corrugatagnostus refragor Pek; Pek: 29-30; pl. 7, figs 4, 5; text-fig. 9. _ MATeRIAL. Headshields, It.19561, It.19563; pygidium, It.19562. STRATIGRAPHICAL RANGE. Upper Arenig, Fennian; biozones of Bergamia rushtoni and Dionide | levigena. Loca ities. Type locality of Pontyfenni Formation, and Llanfallteg Formation, type section. | Discussion. This species is known from sparse and rather poorly preserved material which does | not permit a full description. In its glabellar structure it is a typical Corrugatagnostus— essentially a metagnostid en grand tenu (Fortey 1980: text-fig. 4). The scrobiculae are, however, | rather weakly developed by comparison with the type species (Whittard 1955: text-fig. 2; Pek _ 1977: 27-29; pl. 5, figs 6-8; pl. 6, figs 1-7; pl. 9, fig. 1; Pek & Prokop 1984) and with some 114 R. A. FORTEY & R. M. OWENS ee ry ees § Fig. 15 Corrugatagnostus cf. refragor Pek 1977. Late Arenig (Fennian). a, latex cast from cephalic shield showing pattern of scrobiculae, Pontyfenni Formation, B. rushtoni Biozone, loc. 23, x 10, It.19561; b, latex cast from incomplete pygidium, Llanfallteg Formation, late Arenig, Biozone of Dionide levigena, type section, loc. 52, x 10, It.19562; c, latex cast from incomplete cephalon, Pontyfenni Formation, locality as Fig. 15b, x 12, It.19563. others: C. sol Whittard 1955, C. convergens Weir 1959. C. fortis (Novak 1883; see Pek 1977: 31; pl. 7, figs 1, 2; pl. 8, fig. 7) is weakly scrobiculate, but with a strongly zonate cephalon, and with an anterior glabellar furrow which does not arch forwards medially. There are four species with scrobiculae weakly developed as in our species: C. refragor Pek 1969, C. chekiangensis Sheng 1964, C. transitus Lu 1975 and C. jiangshanensis Lu 1964 (see also Lu et al. 1976: pl. 9, fig. 3). Of these, C. jiangshanensis lacks the second transglabellar furrow, which indicates that it is a scrobiculate Segmentagnostus rather than a true Corrugatagnostus. The same probably applies to C. chekiangensis Sheng. C. transitus is known from a pygidium only, and this has more, but weaker, scrobiculae than in our species, and the terminal piece of the pygidium contracts in width from the preceding axial ring. C. refragor Pek is from the Llanvirn Sarka Formation of Bohemia, and is very close to our species; apart from similar density of scrobiculae, the anterior transglabellar furrow is arched gently forwards and the glabellar tubercle does not strongly protrude into the frontal glabellar lobe as it does on C. morea. Such differences as there are may well be because of preservation, the Bohemian species being well-preserved in relief. For example, the Fennian cephala are wider than long, where the type material of C. refragor has cephalic length almost equal to width, but if crushing ‘opened out’ the fixed cheeks this distinction may not be important. However, the cephalic border furrows are broader antero- laterally on the Welsh specimens, as if they were incipiently zonate. This difference is important enough for us to introduce the qualification into the determination. Genus ARTHRORHACHIS Hawle & Corda 1847 TYPE SPECIES. Arthrorhachis tarda Hawle & Corda 1847, by monotypy. REMARKS. We follow Fortey (1980: 25-29) in using the generic name Arthrorhachis for metag- nostids lacking transglabellar furrows, and having a short pygidial axis with the terminal piece shorter than the postaxial field. These will have been referred to Geragnostus or Trinodus previously. Arthrorhachis sp. indet. (Figs 16a—d) MATERIAL. Cephalic shields: It.18569, NMW 33.189.G23; pygidia: It.19566—7. STRATIGRAPHICAL RANGE. Upper Arenig, Fennian, biozone of Bergamia rushtoni and probably also that of Stapeleyella abyfrons. Loca.itigs. Pontyfenni Formation, type locality; and Llwyn-crwn, loc. 24; Pen-y-parc roadside section, loc. 38. ARENIG IN SOUTH WALES eS) Fig. 16 9 Arthrorhachis sp. indet. Late Arenig (Fennian). a, imperfect cephalon and thoracic segment, S. abyfrons Biozone, loc. 38, x 4, It.19564; b, cephalon, B. rushtoni Biozone, loc. 24, x 4, NNW 33.189.G23; c, latex cast from thorax and pygidium, B. rushtoni Biozone, loc. 23, x 6, It.19566: d, latex cast from pygidium, loc. 23, x 6, It.19567. Discussion. The Fennian species from south Wales is known by imperfectly preserved material, and adds little to our understanding of the genus as a whole. A full description of the morphology of Arthrorhachis has been provided by Kielan (1960) and Fortey (1980), and comparative remarks only are given here. The Fennian species differs from most other Arthro- rhachis spp., including the type species, in that the pygidial axis has a terminal piece which is wider than the preceding axial ring; in this respect it is more like some (but not all) species of Geragnostus. This character alone serves to distinguish our species from the following Arthro- rhachis from the early Ordovician: A. saltaensis (Harrington & Leanza 1957) from Argentina; A. danica (Poulsen 1965) (and its subspecies in Fortey, 1980), A. erratica (Jaekel 1909), A. mobergi (Tjernvik 1956), A. elliptifrons (Tjernvik 1956) and A. lentiformis (Angelin 1854) from Scandinavia and Spitsbergen; A. hebetatus (Dean 1973b) from Turkey; Howell’s (1935; see also Capera et al. 1978) three species from the early Arenig of the Montagne Noire, southern France (A. chinianensis, A. abruptus and A. corpulentus, probably all variants of one species); and A. cf. mobergi (Tjernvik) Chang & Fan 1960 (also Lu et al. 1965: pl. 2, figs 13-14) from China. Trinodus valmeyensis Ross 1958, from the western United States, has an advanced glabellar tubercle and would now be referred to Galbagnostus Whittington. The only early Ordovician species with the same pygidial axial structure is A. hupehensis Lu 1975 (especially his pl. 1, fig. 13) from the late Arenig (or earliest Llanvirn) of SW China, although Lu’s pl. 14, fig. 14 shows a more usual Arthrorhachis pygidial axis. The Welsh and Chinese specimens are unlikely to be conspecific, because the pygidial border of the latter is consistently much wider—about one-sixth (sag.) pygidial length as compared with less than one-tenth. One much later (Caradoc) species with a similar pygidial axial structure is A. pragensis Piibyl & Vanék 1968 (see Pek, 1977: 24-25; pl. 1, figs 9, 10; Pek & Prokop 1984), but in this species the axis (excluding half ring) is much shorter than the postaxial field, where they are nearly equal on our specimens. Although the median pygidial tubercle is not well preserved on the Welsh pygidia, it is clearly a narrower, less tumid structure than that of A. pragensis. The stratigraphically early cephalon from the Stapeleyella abyfrons Biozone has a wider border than those from the Bergamia rushtoni Biozone, but with so few specimens we can say nothing about the intraspecific variation. Cephalic features are generally similar throughout the genus. In summary, it is probable that we have a new Arthrorhachis species in the Pontyfenni Formation, but the material is inadequate to name it as such; hence we place it under open nomenclature. 116 R. A. FORTEY & R. M. OWENS Genus SEGMENTAGNOSTUS Pek 1977 TYPE SPECIES. Agnostus caducus Barrande 1872, by original designation; Llandeilo, Bohemia (Whittard 1955, Pek 1977, Pek & Prokop 1984). DIAGNosISs. Metagnostids with one deep transglabellar furrow, shaped like an inverted ‘v’;; pygidium like that of Geragnostus. DISCUSSION. This genus was discussed at length by Fortey (1980: 27), who noted that the glabellar structure differed fundamentally from that of Arthrorhachis and Geragnostus with regard to the incision of the transglabellar furrow relative to cephalic musculature. The inverted ‘v transglabellar furrow is regarded as the defining character; if a second pair of glabellar furrows is present at all they appear as small indentations in the sides of the glabella. Pek (1977) included Agnostus frici Holub 1908 within his concept of the genus, on which a second pair of furrows is strongly developed, essentially in the arrangement characteristic of Corrugat- agnostus, and this species is probably better regarded as a non-scrobiculate member of that genus. Fortey (1980) listed five species besides the type species assignable to Segmentagnostus as we understand it: from Britain, France and Argentina. To these may be added: Geragnostus merus Zhou (in Lu et al. 1976) from Jiangxi, China; that specimen of Geragnostus sinensis Sheng 1974 on his pl. 1, fig. 1b from Yunnan; and Geragnostus scoltonensis Whittard 1966, from Wales, and S. stubblefieldi Rushton & Hughes 1981 from the Great Paxton borehole. The distribution of the genus appears to be peri-Gondwanan, matching that of Neseuretus, for example, although Segmentagnostus is found in deeper water facies. However, the presence of a probable Segment- agnostus in western Newfoundland (Fortey 1982) suggests that the genus could yet prove to be more widespread in the appropriate facies. Segmentagnostus hirundo (Hicks 1875) (Fig. 17a) 1875 Agnostus hirundo Hicks: 176; pl. x, fig. 10. 1914 Agnostus maccoyi Salter; Thomas in Strahan et al.: 18. 1939b Geragnostus (Micragnostus) hirundo (Hicks) Kobayashi: 579. 1955 Geragnostus hirundo (Hicks); Whittard: 7-8 (pars); pl. 1, fig. 4, non figs 1-3. Ho.otypee. Pygidium, SM A15265, from the Whitlandian of ‘Whitesands Bay’ (presumably Pwlluog), St David’s, Dyfed. The counterpart of this specimen is 1.709 in the British Museum (Natural History). STRATIGRAPHICAL RANGE. Whitlandian, Bergamia gibbsii Biozone. OTHER OCCURRENCE. Rhyd Henllan Member of Colomendy Formation, loc. 47. MATERIAL. Cephalic shield, It.19596; pygidium BGS Pr1850. Discussion. Whittard (1955) refigured the holotype, a poorly preserved pygidium. He believed the type material was of late Arenig age, which we now know to be incorrect, and we here assign Whittard’s supposed S. hirundo from Shropshire to S. scoltonensis (see below). There is nothing to add to Whittard’s description of the type specimen. We have collected an extremely poorly preserved cephalic shield (It.19596) from Pwlluog which shows that the species is prob- ably correctly referred to Segmentagnostus. A well-preserved, incomplete pygidium from the Whitland area (Fig. 17a) gives a better idea of pygidial morphology than the type. The pygidial marginal spines are minute compared with S. scoltonensis as interpreted here (cf. Whittard 1955: pl. 1, fig. 2), but like that species the mid-part of the first axial ring is hardly defined, in this respect contrasting with S. whitlandensis sp. nov. Segmentagnostus whitlandensis sp. nov. (Figs 17d—g) 1914 Agnostus hirundo Salter; Thomas in Strahan et al.: 19. ARENIG IN SOUTH WALES i7/ HoLoryPe. Cephalic shield, It.19569. PARATYPES. Pygidia: It.19570-2; BGS Pr1756—8, Pr1766. TYPE LOCALITY AND HORIZON. Loc. 38; Late Arenig, Fennian, Pontyfenni Formation, Stapeley- ella abyfrons Biozone. Known only from the type locality. Name. After the Whitland district. DIAGNOSIS. Segmentagnostus with granulate surface sculpture; median glabellar tubercle not immediately behind transglabellar furrow; terminal piece on pygidial axis equal to or exceeding length of postaxial field; first pygidial axial ring well-defined medially. DESCRIPTION. Cephalon with maximum width near rear, this equal to sag. length. Glabella occupies two-thirds cephalic length, and two-fifths maximum width, expanding in width gently forwards to anterolateral corners. Axial furrows deep, somewhat wider posteriorly. Trans- glabellar furrow only slightly less deep, forming a deep inverted ‘v’ and dividing the glabella into two lobes, the frontal lobe slightly more than half the length of the posterior lobe. The median glabellar tubercle is prominent, and at a small distance behind the transglabellar furrow, almost opposite its outer ends. Posteromedially the glabella is extended into a median acumination between the basal lobes, which are of the inflated triangular form usual in metag- nostids. Border defined by a very deep furrow which is widest anterolaterally; border comprises well under 10% cephalic length. The best preserved pygidia include external moulds, showing the granulation which we presume covered the whole dorsal surface. The pygidial axis is of the usual Segmentagnostus Fig. 17 a, Segmentagnostus hirundo (Hicks 1875). Late Arenig (Fennian); pygidium, Rhyd Henllan Member, loc. 47, Whitlandian, gibbsii Biozone, x 12, GSM Pr1850. b—c, Segmentagnostus scolton- ensis (Whittard 1966); b, cephalon, Llanfallteg Formation, type section, Fennian Dionide levigena Biozone, x 15, It.19597; c, cast from incomplete cephalon, Pontyfenni Formation, loc. 23, Bergamia rushtoni Biozone, x 8, It.19598. d—g, Segmentagnostus whitlandensis sp. nov., late Arenig (Fennian, Stapeleyella abyfrons Biozone), Pontyfenni Formation, loc. 38. d, latex cast from external mould of incomplete pygidium showing granulation, x 12, It.19570; e, holotype, cephalon, internal mould showing tubercle behind transglabellar furrow, x 12, It.19569; f, latex cast from well-preserved pygidium, x 20,GSM Pr1758; g, pygidium, x 12, GSM Pr1766. 118 R. A. FORTEY & R. M. OWENS form, tapering past the two axial rings, and expanding gently around the terminal piece. The structure of the axial rings is generally the same as in S. mccoyii as described by Hughes (1969) except that the first axial ring is well defined medially; the prominent median tubercle is on the second ring only, and slightly impinges on the terminal piece. There is also a minute tubercle at the tip of the axis. Terminal piece long, equal to or exceeding in length the postaxial field. Borders relatively narrow, as on cephalic shield, with very small marginal spines. Discussion. There are two differences between the pygidia of this species and those of S. hirundo and S. scoltonensis which are not likely to be accountable as differences in size or preservation. The pygidial border on scoltonensis is almost twice as wide posterolaterally, and the first axial ring narrows markedly abaxially and is effaced, whereas on S. whitlandensis the median lobe is quite well defined there and only slightly narrower (sag.) than those on either side. The second distinction also applies to S. mccoyii (Salter) (see Hughes, 1969: pl. 1, fig. 3). The granulate surface is at present unique to S. whitlandensis among described Segmen- tagnostus, but preservation may not be adequate to record it on other species. S. whitlandensis differs from all other post-Arenig Segmentagnostus, including the type species (Whittard 1955: text-fig. 2a; Pek 1977: pl. 1, fig. 7) in having the glabellar tubercle distinctly posterior to the transglabellar furrow. Segmentagnostus changes little throughout its long history: the earliest species S$. neumanni (Harrington & Leanza 1957) from the Lower Tremadoc of Argentina is generally like the youngest, S. merus (Zhou 1976), from the Upper Ordovician of China. Harrington & Leanza (1957: 69) state that the glabellar tubercle on their species “does not reach the transglabellar furrow’ (but see their fig. 13, 10), and if this is so it is the only species besides S. whitlandensis with this character. S. neumanni differs from the Welsh species because the transglabellar furrow is a very weak ‘inverted v’, the posterior glabellar lobe is longer, and the pygidium carries long \ marginal spines. Segmentagnostus scoltonensis (Whittard 1966) (Figs 17b, c) 1955 Geragnostus hirundo (Hicks); Whittard: 7—8 (pars); pl. 1, figs 1-3, non fig. 4. 1966 Geragnostus scoltonensis Whittard: 266-267; pl. 46, figs 3—S. HovorypPe. Distorted complete exoskeleton, SM A44493. TYPE LOCALITY AND HORIZON. Llanfallteg Formation, early Llanvirn part, D. artus Biozone; Scolton Railway cutting, Dyfed. OTHER OCCURRENCES. Llanfallteg Formation, latest Arenig part; Tankerville Flags of Shrop- shire (Upper Arenig); Pontyfenni Formation (Upper Arenig, B. rushtoni Biozone), loc. 23; ?Stapeley Volcanics, Llanvirn, Shropshire. FIGURED MATERIAL. Cephalic shields: It.19597-8. Discussion. The holotype of this species is considerably elongated (sag.) by distortion (Whittard 1966: pl. 46, fig. 4) and its morphology is not easy to interpret. Our view of this species is therefore inevitably influenced by stratigraphical considerations. When Whittard (1955) described S. ‘hirundo’ (Hicks) from the Tankerville Flags, he assumed that the type of that species from St David’s was from the late Arenig (“D. hirundo Zone’) also. We now know it to be much older (Whitlandian). On the other hand the type of S. scoltonensis is from the Llanfallteg Formation, ‘and there are other similarities between the Llanfallteg Formation fauna and that of the Tankerville Flags in Shropshire. Whittard (1966) stressed the absence of a second pair of glabellar furrows (posterior to the transglabellar furrow) as the important specific character of S. scoltonensis, but the same is true of his (1955) so-called hirundo from the Tankerville Flags. These specimens also show the glabellar tubercle immediately behind the transglabellar furrow, which is an important difference from the S. whitlandensis cephalon figured here. Broad pygidial borders, compared with S. mccoyii (Salter) or S. stubblefieldi Rushton & Hughes 1981, are present on the type of scoltonensis and the Segmentagnostus from | } | | | | ARENIG IN SOUTH WALES 119 the Tankerville Flags. Hence we consider it likely that the Scolton and Tankerville specimens are conspecific, the name scoltonensis should be applied to them, and that the species ranges across the Arenig—Llanvirn boundary. We figure here cephalic shields from the upper part of the Pontyfenni Formation, of similar age to the Tankerville scoltonensis, and from the succeeding Dionide levigena Biozone of the Llanfallteg Formation. Both show a slightly shorter frontal glabellar lobe when compared with Whittard’s (1955: pl. 1, fig. 1) cephalic shield, but since this proportion shows a certain range of variation in Segmentagnostus (Hughes 1969: table 1), and is altered by distortion, this is not regarded as important. The larger cephalic shield (Fig. 17c) shows a faint indication of a second pair of glabellar furrows. Whittard (1966: pl. 46, fig. 5) also figured a cephalic shield from the Stapeley volcanics, which would imply a stratigraphical range into the later Llanvirn, although it is as well to be cautious about this until a pygidium is discovered from the later horizon. S. scoltonensis is distinguished from S. stubblefieldi Rushton & Hughes 1981, from the Llan- virn of the Great Paxton Borehole, by the latter having a pygidial axis which tapers uniformly past the first two rings to an almost rectangular terminal piece; a prominent median ridge on the pygidial axis of S. stubblefieldi produces a distinct tripartition of the anterior axial ring. The type of S. scoltonensis, poorly preserved though it is, is clearly like the Tankerville specimens in these characters. S. scoltonensis is very like S. hirundo from the middle Arenig; if our attribution of the Tankerville specimens to the former is correct the marginal pygidial spines are more prominent on S. scoltonensis, and the first two pygidial axial rings account for a greater proportion (sag.) of the pygidial axis. More material of both species needs to be discovered to assess the limits of variation. Family SHUMARDIIDAE Lake 1907 Genus SHUMARDIA Billings 1862 TYPE SPECIES. Shumardia granulosa Billings 1862, by monotypy. DiscussION. Several generic and subgeneric names have been proposed for Shumardia-like trilobites recently, and there is the question of whether the genus Conophrys Callaway (type species C. salopiensis Callaway 1877) should be revived. Fortey (1980) reviewed Shumardia, sensu lato and concluded that the described species did not unequivocally divide into two— Shumardia and Conophrys—in spite of the differences between their type species. Fortey & Rushton (1980) redescribed the shumardiid Acanthopleurella Groom, which is distinguished from Shumardia by having only four thoracic segments, two of which are macropleural, and in having a long (sag.) occipital ring, genal spines, and short, stubby pygidial axis. They also illustrated what is probably the best-preserved specimen of Conophrys salopiensis yet dis- covered. Fortey (1980) did not consider Kweichowilla Chang 1964 (type species K. minuta Chang 1964), and Piibyl & Vanék (1980: 14) have recently proposed the subgenus Shumardia (Shumardella) with S. bohemica Marek 1964 as type species. So there are now four possible supraspecific taxa to be considered: Shumardia, Conophrys, Kweichowilla and Shumardella. The type species of Shumardia was revised by Whittington (1965); we note that its glabella is well defined anteriorly, with large anterolateral lobes. Whittington considered that it lacks a macropleural thoracic segment, and the long (sag.) pygidium is without borders laterally. Conophrys salopiensis from Shropshire and Wales has been identified with the species S. pusilla (Sars) (Stubblefield 1926), but is a distinct form (P. Whitworth, personal communication 1972). The glabella is defined anteriorly, with rather small anterolateral lobes; genal spines are lacking (Fortey & Rushton 1980: fig. 17); the fourth thoracic segment is macropleural; the pygidium is transverse with the axis extending nearly to the border, which is narrow but distinct. Kweichowilla has not been formally diagnosed, although the type species has been illustrated three times (Chang et al. 1964; Lu & Zhang 1974: pl. 55, figs 2, 3; Yin & Li 1978: pl. 162, figs 10, 11). It is distinguished by its relatively huge, drop-like glabella lobes, which extend far back, and by the broad glabellar tongue which extends to the anterior cranidial margin. The pygidium attributed to K. minuta is elongate and 120 R. A. FORTEY & R. M. OWENS subtriangular like that of the type species of Shumardia. The type species of Shumardella is known from the cranidium alone, a fragmentary pygidium figured by Marek (1964) being only doubtfully associated. On S. bohemica the furrows defining the anterolateral glabellar lobes are narrow, and the lobes do not bulge outwards in the manner of most shumardiids, so that the glabellar outline is nearly parabolic. The posterior borders on the fixed cheeks are relatively wide (exsag.) and not depressed below the rest of the cheek. On the basis of the type species alone the four taxa appear easily definable, but as Fortey (1980) observed, when other species are taken into account the distinctions are less clear-cut. If the genera were phylogenetic units one might expect pygidial characters to be consistent with cephalic ones, but this is not so. For example, several species with Shumardia-like cranidia have transverse, Conophrys-like pygidia (S. minaretta Fortey 1980), others with Kweichowilla-like cranidia have transverse pygidia unlike the type species of that genus (e.g. S. forbesi Stait & Laurie 1983). The classification of such species as S. curta Stubblefield 1927 and S. ctenata Robison & Pantoja-Alor 1968, which have free pleural tips on the pygidium, is also unclear. There may be an argument for employing subdivisions within the large genus Shumardia as Dean (1973a) suggested, and a compromise may be to use subgeneric categories based on the type species; species for which no pygidium is known, or with ambiguous combinations of characters, may be referred to Shumardia, sensu lato. Using this approach, the following diagnoses may be proposed, with some of the previously described Shumardia species listed in Fortey (1980) indicated with their subgeneric placing. Shumardia (Shumardia): Glabella defined anteriorly inside cephalic margin; anterolateral glabel- lar lobes large, swollen, but not extending far back; macropleural thoracic segment lacking; pygidium elongate-triangular, axis not extending near margin, and well-defined borders lacking. Shumardia granulosa Billings 1862, S. dicksoni Moberg & Segerberg 1906, S. lacrima Koroleva 1964, S. tarimuensis Zhang 1983 and S. gadwensis sp. nov. (p. 121). Shumardia (Conophrys): Glabella defined anteriorly inside cephalic margin; anterolateral glabel- lar lobes small to moderate sized, not greatly inflated; macropleural thoracic segment present (where thorax known); pygidium transversely oval or semicircular, axis extending to border; narrow pygidial borders present, or elevated marginal rim. C. salopiensis Callaway 1877, C. pusilla (Sars 1835), C. nericiensis Wiman 1905 (but with short pygidial axis), C. oelandica Moberg 1901, ?C. bottnica Wiman 1902, C. changshanensis Lu (in Lu et al. 1976), C. keguqin- ensis Xiang & Zhang 1984. Shumardia (Kweichowilla): Glabella continued forward as broad tongue to cranidial margin, often more or less effaced; glabellar lobes extending backwards, large, drop-like; pygidium transverse, or triangular. K. minuta Chang 1964, K. hongyaensis Lu & Zhang 1974, K. lacrimosa Dean 1973a, K. acuticaudata Fortey 1980, K. matchensis Legg 1976, S. sagittula Whittington 1965, K. forbesi Stait & Laurie 1983. Shumardia (Shumardella): Cranidium narrow (tr.), glabella parabolic to clavate, not reaching cranidial margin, with anterolateral lobes not greatly projecting, defined by narrow glabellar furrows. S. bohemica Marek 1964, S. polonica Kielan 1960, S. scotica Reed 1903, S. extensa Weir 1959, S. tenacis Zhou in Lu et al. 1976. These are all late Ordovician species; the earlier Ordovician S. phalloides Fortey 1980 is generally similar, but the glabella has a narrow tongue. The genus Eoshumardia Hupé 1953 is only described from its type species from the Upper Cambrian, Shumardia orientalis Mansuy (1916: pl. 1, figs 28a—e). Mansuy’s cranidium on fig. 28a shows a prominent occipital spine; the anterolateral glabellar lobes are well developed as in Shumardia (Shumardia); there may be bacculae in the axial furrows (but from the illustrations it is impossible to tell whether these structures may simply be uncleaned matrix); the posterior cranidial border is wide (exsag.). The pygidium assigned by Mansuy is somewhat elongate (sag.) in the manner of Shumardia (Shumardia), but with the axis long and narrow and extending almost to the margin. Eoshumardia is closest to Shumardia (Shumardia) of our usage, but its validity will depend on the redescription of the type species. The genus Koldinioidia Kobayashi ARENIG IN SOUTH WALES 121 1931 has been used by Robison & Pantoja-Alor (1968) and Shergold (1975) to apply to shumardiids with tapering glabella, and glabellar lobes hardly defined. Zhou & Zhang (1984) have redescribed the type species (K. typicalis), which is not at all like the species subsequently assigned to Koldinioidia; the cranidium is similar to that of Eoshumardia. The pygidium figured by Zhou & Zhang is transverse, like that of Conophrys. Subgenus SHUMARDIA Billings 1862 Shumardia (Shumardia) gadwensis sp. nov. (Figs 18a—1, 20) Ho.ortypee. Dorsal exoskeleton, It.19573. PARATYPES. Dorsal exoskeletons: It.19574—S, It.19578; cranidia: It.19576-7, NMW 84.17G.35; pygidium: It.19579. TYPE LOCALITY AND HORIZON. East side of Nant-y-Gadwen, Llanfaelrhys, Llyn Peninsula; Middle Arenig (Whitlandian), Biozone of Gymnostomix gibbsii, unnamed mudstone formation. Name. After the type locality. OTHER OCCURRENCES. Also known from the Whitlandian mudstones beside track of Dwyrhos Farm, west of Aberdaron, Ll¥n Peninsula. In south Wales, S. gadwensis occurs in the Whit- landian of Pwlluog, north of Whitesand Bay, St David’s, Dyfed, in the Penmaen Dewi Forma- tion, Biozone of Gymnostomix gibbsii, and from slightly higher in the section just below the igneous intrusion (loc. 61G). From the Afon Ffinnant Formation, S. gadwensis has been reco- vered from the Cwm Arbont locality (Fortey & Owens 1978: fig. 3), near top of section, and from Afon Ffinnant loc. 16K. DIAGNosIs. Shumardia (Shumardia) with fixed cheeks of transverse width at posterior margin less than that of occipital ring. Pygidium with four axial rings, two deep pairs of pleural furrows, and without border. Exoskeletal surface not granulate. DESCRIPTION. None of the material of this species is perfectly preserved, but we have several articulated specimens, and taken together the material is adequate to produce the reconstruc- tion shown on Fig. 20, p. 125. It is a useful guide fossil for the Whitlandian, and the only representative of Shumardia (Shumardia) from undoubted Arenig rocks. For these reasons it seems advisable to name it. The entire specimens are either somewhat stretched or compressed and it is not possible to give a length/width ratio; however the cephalon, thorax and pygidium are about equal in length (sag.). Transverse convexity is considerable, and as usual in shumardi- ids this is because the cheeks and thoracic pleurae are sharply downturned laterally. On many specimens the lateral edges have collapsed, often in a somewhat lopsided fashion (Fig. 18a). On the least distorted specimens the cranidium is twice as wide as long. The width of the fixed cheek at its widest, at the posterior margin, is a little less than the transverse width of the occipital ring, but the cheeks appear relatively narrower when crushed. Glabella tapers only slightly forwards to drop-like glabellar lobes which are fairly prominent, slightly inflated, and protrude into the side axial furrows. The transverse width of the lobes is a little less than the width of the mid-part of the glabella between them. The front of the glabella is distinctly defined at an obtuse point well inside the margin. The preglabellar furrow is deep on all but the small cranidium in Fig. 18e (on which the front of the glabella can still be seen), which may be attributable to preservation as the occipital furrow is also faint on this specimen. Smaller specimens, including the complete exoskeleton in Fig. 18c, show a small occipital spine which we have not seen on larger cranidia; loss of this spine is presumed to be a feature of later ontogeny. Posterior border furrow deep adaxially, but fading rapidly laterally, as it curves a little forwards. No specimen shows details of the free cheeks. The holotype shows a small baccula on the left side of the glabella; no other specimen is well enough preserved to show this feature. 122 R. A. FORTEY & R. M. OWENS EES ie Fig. 18 a-i, Shumardia (Shumardia) gadwensis sp. nov. Middle Arenig (Whitlandian, G. gibbsii Biozone); a, b, cranidium, internal mould but well-preserved, with original relief preserved on right side, dorsal and anterior views, Afon Ffinnant Formation, loc. 17, x 12, It.19576; c, cast from holotype, somewhat compressed small dorsal exoskeleton, mudstones in Nant-y-Gadwen, Llyn, north Wales, x 12, It.19573; d, poorly-preserved but not distorted dorsal exoskelton, showing thoracic segments, Penmaen Dewi Formation, near intrusion, north of Whitesand Bay, x 6, It.19575; e, g, pygidium (It.19579) and cranidium (It.19577) on same slab, internal mould and latex cast from counterpart showing undistorted proportions of pygidium; note occipital spine on small cranidium, loc. as Fig. 18a, x 15; f, h complete exoskeleton, slightly distorted by tectonic extension, dorsal and lateral views, locality as holotype, Fig. 18c, x 14, It.19574; 1, small exoskeleton, internal mould, Dwyrhos Quarry, Llyn, north Wales, x 20, It.19578. j-k, Shumardia (Shumardia) sp. A. Upper Arenig (Fennian); j, pygidium, Pontyfenni Formation, (S. abyfrons Biozone), loc. 38, x 14, It.19582; k, cranidium, same locality, x 12, It.19580. 1, Leioshumardia sp. A; dorsal view, Middle Arenig, Whitlandian, Afon Ffinnant Formation, loc. 17, x 14, It.19583. The smaller complete specimen apparently shows five thoracic segments. The larger (Fig. 18f) one is not well preserved but shows a ring-like structure behind the occipital ring which could be interpreted as evidence of a sixth segment. A poorly preserved but otherwise undistorted specimen from Whitesand Bay (Fig. 18d) shows six thoracic segments also, as does a small specimen (Fig. 18i) from Dwyrhos Quarry, Llyn. None is macropleural. The axis tapers back- wards after the third segment. Pleural furrows are short, not extending onto the downturned, faceted pleural tips. Axial taper continues backwards on the pygidial axis: axial furrows enclose an angle of about 30°. Pygidium triangular, slightly wider than long, without border; axis extends to two-thirds pygidial length. There are certainly four axial rings; a faint fifth may be present on the flanks of the rounded terminal piece. There are two pairs of deep pleural furrows, which curve backwards but do not extend to the pygidial margin. The external mould (Fig. 18g) of the ARENIG IN SOUTH WALES 123 best-preserved specimen shows a much shorter third pair of pleural furrows and a suggestion of a fourth pair. Some of the material is well enough preserved to suggest that the exterior surface of the whole exoskeleton was probably smooth, and assuredly not granulate like S. (S.) granu- losa Billings. Discussion. Whittington (1965) has given a full description of the type species Shumardia (Shumardia) granulosa Billings 1862, from the Shumardia Limestone, Quebec. It is generally simiiar to S. gadwensis, particularly with regard to pygidial morphology. Whittington’s pl. 16, fig. 12 clearly shows the bacculae adjacent to the base of the glabella which are also present on the holotype of S. gadwensis. There are several good specific differences: the glabella of S. granulosa is proportionately narrower, width of the occipital ring being less than that of the fixed cheeks; the anterior glabellar lobes are wider (tr.) and more inflated; the pygidium has a flattened postaxial border (this is not true of Whittington’s pl. 16, figs 5-9, however); there is a granulate surface sculpture. S. dicksoni Moberg & Segerberg (1906: pl. 4, figs 17-22) is less similar to S. gadwensis; if the cranidium is correctly assigned the glabella has small antero- lateral lobes, a rounded front, and a second pair of glabellar furrows; the pygidial axis is much shorter than that of S. gadwensis. S. lacrima Koroleva 1964, from the Middle Ordovician of Kazakhstan, has a cranidium very like that of S. gadwensis, although the fixed cheeks are narrower and the anterolateral glabellar lobes protrude further into the axial furrows. The pygidium is distinctive: it has five or six pairs of short pleural furrows. S. (S.) gadwensis, dicksoni and granulosa together constitute the concept of Shumardia (Shumardia) advocated here. Shumardia (Shumardia) sp. A (Figs 18), k) MATERIAL. Cranidia: It.19580—1; pygidia: It.19582-3. STRATIGRAPHICAL RANGE. Upper Arenig (Fennian), Biozone of Stapeleyella abyfrons; ?Biozone of Bergamia rushtoni. Loca.ities. Pontyfenni Formation, Pen-y-parc section (loc. 38); Gelli Diogyn (loc. 32a). Discussion. There is a second species of Shumardia (Shumardia) from a higher stratigraphical level than S. (S). gadwensis. It appears to be another new species, but the material is far from sufficient to name it formally. The cranidium differs from that of S. gadwensis in having even narrower fixed cheeks and in having a depressed area in front of the glabella (like the type species, S. granulosa). The pygidium is highly effaced, and in this it differs from all other Shumardia (Shumardia) spp. Although the axis is long, only one ring is defined, and that incompletely. This is not an artefact of preservation, because both specimens show it, and the axial furrows are of usual depth. Subgenus CONOPHRYS Callaway 1877 TYPE SPECIES. Conophrys salopiensis Callaway 1877; Shineton Shales, Shropshire. Shumardia (Conophrys) crossi sp. nov. (Figs 19a—h, 20) Hovorype. Exoskeleton with right side of cranidium damaged, It.19584. PARATYPES. Imperfect exoskeletons: It.19585-6, NMW 84.17G.36; cranidia: It.19587—8, BGS JP4883-4; pygidia: It.19589-90, NMW 84.17G.37-38. TYPE LOCALITY AND HORIZON. Loc. 23; Pontyfenni Formation type locality, Upper Arenig (Fennian), biozone of Bergamia rushtoni. NAME. For Mr F. Cross, who collected the holotype, and many other specimens figured in this work. 124 R. A. FORTEY & R. M. OWENS OTHER OCCURRENCES. S. (C.) crossi has only been found in the Pontyfenni Formation of south Wales, biozone of Bergamia rushtoni. It is known otherwise from west of Banc-y-felin, at Rushmoor (Survey loc. 38SW EAS), at Bron-y-Gaer (Survey loc. 38SW EA1), and at Sarn-las (loc. 48). DiaGnosis. Shumardia (Conophrys) with broad (tr.) anterior glabellar lobe, which is hardly pointed and defined by faint preglabellar furrows; bacculae present in axial furrows. Six tho- racic segments, no macropleural segment. Pygidial axis exceptionally broad for subgenus, occupying half pygidial width at first ring. DESCRIPTION. The holotype (Figs 19a—c) is almost undistorted and shows the original convexity. Exoskeleton almost twice as long as wide. This specimen demonstrates that the glabella was turned down with the cephalic edge, but still terminates inside the margin. Other cranidia are all slightly crushed. That in Fig. 19h apparently shows the glabella continuing to the margin as a tongue in the manner of Kweichowilla (right side), but this is an artefact of preservation. Glabella occupies just under half cranidial width at occipital ring. Axial furrows are deep posteriorly but shallow abruptly around the anterior glabellar lobe, so that they appear effaced on the more poorly preserved material; the course of the preglabellar furrows can be discerned on internal moulds (Fig. 19d). There are distinct bacculae which constrict the base of the glabella a little. The anterolateral glabellar lobes protrude outwards, but not far beyond the level of the lateral margins of the occipital ring; the furrows defining them are narrow and do not extend more than one-third across the median glabellar lobe. On flattened cranidia the position of the anterolateral lobes is displaced relatively backwards because of the splaying out Bi a ee Fig. 19 Shumardia (Conophrys) crossi sp. nov. Upper Arenig, Fennian (Bergamia rushtoni Biozone), loc. 23. a-c, holotype, incomplete dorsal exoskeleton, dorsal, oblique and lateral views, x 8, It.19584; d, e, cranidium and one thoracic segment, dorsal and lateral views, x 9, It.19587; f, latex cast from natural mould of larger pygidium, x 10, It.19589; g, well-preserved pygidium, x 12, It.19590; h, crushed cranidium, x 8, It.19588. Fe Eg ARENIG IN SOUTH WALES 125 of the cranidial margin. The front of the glabella is almost bluntly rounded about the mid-line, or with a faint suggestion of a point (Fig. 19d). Posterior border furrow deep, not reaching cranidial margin; border widens laterally. Thorax with axis nearly three times as wide as thoracic pleurae; we have found no evidence of a macropleural segment. Structure of pleurae typically shumardiid, with depressed front margin fitting beneath preceding pleura, transverse pleural furrows near back of pleura, and deep, downturned facet. Axis hardly tapers backwards. Pygidia in relief are two-thirds to three-quarters as long as wide (long (sag.) half-ring is excluded); flattened examples may appear a little longer as the postaxial field is extended. The axis is very wide (tr.}—half pygidial width at first rimg—and transversely convex, with only gentle backward taper to a broadly rounded termination at about 0-7 of pygidial length (again half-ring excluded). Four axial rings are defined by furrows which are distinctly fainter medially (the fourth may be hard to discern); the first three rings are of equal length (sag.). The three pairs of pleural furrows are progressively shorter backwards, the third pair being very short. In the relief material the postaxial field slopes down to a narrow (sag.) flattened border, which becomes narrower away from the mid-line and does not extend to the anterolateral margins. This may be obliterated on crushed material. The holotype shows the extent of the doublure where it is impressed on the dorsal surface; wide on the mid-line, and narrowing rather abruptly to the second pair of pleural furrows. Fig. 20 Reconstructions of Shumardia (Shumardia) gadwensis sp. nov. (left) and S. crossi sp. nov. (right); x 18 approx. 126 R. A. FORTEY & R. M. OWENS Well-preserved material shows a distinctive surface sculpture of irregular anastomosing ridges; around the front margin of the cranidium there are close-set raised lines parallel to the margin. One pygidium (Fig. 19g) shows lines of small tubercles on the hind margin of the axial rings, and a similar sculpture was probably present on thoracic rings. DIscussION. S. (C.) crossi is a distinctive species of Conophrys; its wide pygidial axis dis- tinguishes it from all others given in the generic discussion above. Indeed the relatively long pygidium, with a flattened border, is perhaps more like Shumardia (Shumardia), although it does not have the typical triangular pygidial outline, and the development of the anterolateral glabellar lobes is more like that of most Conophrys spp. Tremadoc species of Conophrys have a macropleural segment, for example the type species C. salopiensis (Fortey & Rushton 1980: figs 16, 17). Its loss in the Arenig may not be important. It is possible that the presence of bacculae may prove a more reliable character to distinguish Shumardia (Shumardia) which this species shares with S. (S.) granulosa and S. (S.) gadwensis, but there are many shumardiids described which are too imperfect to be sure of this detail. A Conophrys of similar age is S. (C.) minaretta Fortey from the Arenig of Spitsbergen (Fortey 1980: pl. 3, figs 1-10); it has an acute front of the glabella, wide cephalic axial furrows, and the pygidial axis, apart from having the usual, relatively narrow proportions of Conophrys, also has five axial rings. Genus LEIOSHUMARDIA Whittington 1965 TYPE SPECIES. Leioshumardia minima Whittington 1965, by original designation. Leioshumardia sp. A (Fig. 181) MATERIAL. Cranidium, It.19583 OccuRRENCE. Arbont, in mudstones with Shumardia (Shumardia) gadwensis, near top of local section. See Fortey & Owens, 1978: fig. 3, loc.17A. STRATIGRAPHICAL RANGE. Middle Arenig (Whitlandian), biozone uncertain, possibly G. gibbsii. Discussion. Leioshumardia has hitherto been known only from the type species, itself represent- ed by two cranidia. Our single specimen is clearly a different species, but cannot be named formally. Its occurrence in south Wales is nonetheless of interest in showing how off-shelf genera may be widely distributed as early as the Arenig. We have already noted similar species of Shumardia (Shumardia) and Hypermecaspis in eastern North America and south Wales. The Welsh specimen differs from the type species in having a triangular, rather than barrel-shaped glabella, and a weakly defined occipital furrow. The glabella terminates at a point close to, but inside, the cranidial margin, as it does on L. minima. There is a superficial similarity between L. sp. A and Clelandia, especially C. reliqua Rushton & Tripp 1979. The sutures on Clelandia are not marginal as they are in Leioshumardia, and the structure of the cranidial posterior border is of the usual ptychoparioid type. On both characters the Arenig species is like Leioshumardia and unlike Clelandia. Family REMOPLEURIDIDAE Hawle & Corda 1847 Genus GIRVANOPYGE Kobayashi 1960 1961a Cremastoglottos Whittard: 187 1976 Gamops Snajdr: 232 1983 Nanlingia Wei & Zhou: 217. TYPE SPECIES. Lichapyge? problematica Reed 1906, by original designation. Discussion. The peculiar genus Cremastoglottos was described from the Hope Shales on the basis of cranidia only (Whittard 1961a). A year previously Kobayashi had proposed Girva- nopyge for some pygidia from the Whitehouse Beds of Girvan figured by Reed (1906). There was no way of knowing that these pygidia belonged to the same genus of trilobites as the ARENIG IN SOUTH WALES 127 Cremastoglottos cranidia until Marek (1977) described a complete specimen from Bohemia. The cranidium and pygidium are both so distinctive that there is no doubt that Cremastoglottos is a junior synonym of Girvanopyge. Fortey (1981) proposed that Girvanopyge (there called Cremastoglottos) was allied with the remopleuridids rather than the ellipsotaphrids, and this family placing is adopted here. Marek (1977) noted that the genus Gamops Snajdr 1976 was congeneric with Cremastoglottos, with which we concur. Finally, Wei & Zhou (1983) described Nanlingia from pygidia only from a cyclopygid biofacies in east China, but without mentioning any of the foregoing. This is a Girvanopyge species, closely similar to the type species. An additional species ascribed to Cremastoglottos was proposed by Hoérbinger & Vanék (1983), who recognized Lichapyge? problematica Reed 1906 as a probable Cremastoglottos, but without mentioning Kobayashi’s genus, of which they may have been unaware. Additionally, a speci- men from Germany attributed to Cyclopyge by Jentsch & Stein (1961) may prove referable to Girvanopyge. Now that the somewhat lengthy synonymy has been recognised it is clear that Girvanopyge is another widespread, circum-Gondwanan genus confined to the cyclopygid biofacies appropri- ate to its pelagic habits (Fortey 1981: 609). Like many of the cyclopygid genera, it ranges from Arenig to Ashgill with apparently little change. SPECIES INCLUDED. G. problematica (Reed 1906), G. occipitalis (Whittard 1961a), G. aff. occipitalis (Marek 1977), G. mrazeki (Snajdr 1976), G. caudata (Wei & Zhou 1983), G. barrandei (Hérbinger & Vanék 1983) and G. sp. indet. (below). Girvanopyge sp. indet. (Figs 21a—c) STRATIGRAPHICAL RANGE. Upper Arenig, Fennian, biozones of Bergamia rushtoni and Dionide levigena. Loca.itigs. Pontyfenni Formation type locality, and Llwyn-crwn, loc. 24. Llanfallteg Forma- tion, type section, 18 m below Arenig—Llanvirn boundary. MATERIAL. Pygidia: It.19592—3, NMW 84.17G.120; incomplete cranidia: NMW 84.12G 41a, b, It.19594; thoracic segment: It.19595. Discussion. The poor material of this species is enough to show that it is a new species of Girvanopyge, but not adequate to name it. The cranidia are not complete, but clearly show the course of the axial furrow as interpreted by Fortey (1981), with a sharp outward bend in front of the occipital region. Of the two pygidia the smaller example (Fig. 21c) retains one spinose thoracic segment, as do several of the pygidia figured by previous authors (e.g. Wei & Zhou 1983: pl. 72, fig. 9). This specimen is from the Llanfallteg Formation; the posterior pygidial Fig. 21. Girvanopyge sp. indet. a, latex cast of pygidium showing median acumination, Upper Arenig, Fennian (B. rushtoni Biozone), loc. 24, x 34, It.19592; b, incomplete, crushed cranidium, horizon as last, loc. 23, x 6, NMW 84.12G.41; c, flattened pygidium and one segment, oblique illumination, D. levigena Biozone, Llanfallteg Formation, loc. 52P, x 3, It.19595. 128 R. A. FORTEY & R. M. OWENS margin appears to be evenly rounded about the mid-line. Two stratigraphically earlier pygidia from the Pontyfenni Formation are better preserved and on these specimens the border is distinctly acuminate, a feature which distinguishes it from all the other Girvanopyge species listed above. There is some doubt whether the Pontyfenni and Llanfallteg specimens are con- specific, or whether the difference in the pygidial border is a reflection of a difference in preservation. However, both differ from other Girvanopyge spp. in having weakly-defined pleural and interpleural furrows which are only present adaxially; on all the later species the pleural and interpleural furrows are equally well developed, and are directed backwards almost to the pygidial margin. The short, triangular pygidial axis and postaxial ridge on Girvanopyge sp. indet. are still distinctive enough to place the generic assignment on a firm basis. A spinose thoracic segment from the Pontyfenni Formation is unlike that of any cyclopygid, and is probably one of the posterior segments of Girvanopyge sp. indet. (cf. Marek, 1977). Family BOHEMILLIDAE Barrande 1872 Genus BOHEMILLA Barrande 1872 TYPE SPECIES. Bohemilla stupenda Barrande 1872. Subgenus FENNIOPS nov. TYPE SPECIES. Bohemilla (Fenniops) sabulon sp. nov. DIAGNOsIS. Subgenus of Bohemilla with small palpebral lobes, symmetrically disposed about the 3P glabellar furrows; wide (tr.) frontal glabellar lobe, equal in width to occipital ring; post- ocular fixed cheeks wider (tr.) than in Bohemilla (Bohemilla). Name. After the Afon Fenni, near Whitland. Discussion. Species which have been attributed to Bohemilla fall into two morphological groups, one of which is recognized here as the new subgenus Fenniops. The type species of Bohemilla (Bohemilla), B. stupenda Barrande, has been revised by Whittard (1952) and Marek (1966). The front of the glabella contracts in width in front of the 2P glabellar furrows, coincident with the large palpebral lobes that lie alongside the forward part of the glabella; glabellar furrows are long (tr.), median glabellar lobe less than one-third glabellar width. The relict postocular fixed cheeks are very narrow (tr.), and widen forwards only slightly. Other Bohemilla species are very similar in cephalic construction: B. scotica Reed 1914, B. pragensis Marek 1966 and B. tridens Rushton & Hughes 1981. A second group of species (‘Gen. indet.’ of Whittard, 1952) has shorter palpebral lobes, and a broad frontal glabellar lobe gently rounded about the mid-line. The postocular fixed cheeks are relatively wide (tr.) behind the eyes in this group. This is the basis of the new subgenus Fenniops. As well as the type species, which is new, two species are included: B. praecedens Kloucek 1916 and B. klouceki Marek 1966. The structure of Fenniops is presumed to be the more primitive, because the wider fixed cheeks are more comparable with those of other ptychoparioid trilobites, and the occipital ring and glabellar furrows are not as specialized as they are in Bohemilla (Bohemilla). We believe subgeneric status for Fenniops is appropriate, because the Bohemilla and Fenniops morphol- ogies have overlapping stratigraphical ranges, and do not intergrade. Bohemilla (Bohemilla) extends from the Llanvirn to the Ashgill, Bohemilla (Fenniops) from the Arenig to the Llandeilo. However, their shared peculiarities are such that they surely have a common ancestor, which subgeneric classification implies. The earlier members of both B. (Bohemilla) (Rushton & Hughes 1981: pl. 5, fig. 16) and B. (Fenniops) (Fig. 22d herein) have a cranidial anterior border, and short thoracic pleurae. B. (Fenniops) sabulon is the oldest described bohemillid (although A. W. A. Rushton informs us that he has a still earlier example from the Late Tremadoc), and its glabellar structure is likely to be primitive for the group. It certainly lacks the peculiarities of B. (B.) stupenda, which had suggested to Fortey (1974) that Bohemilla might be related to Opipeuter, a view he subse- quently rejected (Fortey 1981). Marek (1966), Fortey (1974) and Rushton & Hughes (1981) all ) : | ARENIG IN SOUTH WALES 129 accepted Bohemilla as a remopleuridacean. The glabella of B. (Fenniops) sabulon is consistent with this interpretation because it shows a bulge in width in front of the occipital ring, a typical remopleuridacean feature. The free cheek of B. (Fenniops) klouceki is very like that of some kainellids (e.g. Pseudokainella keideli; see Harrington & Leanza, 1957). On the other hand Fortey (1981) regarded the glabellar furrows of Psilacella and Bohemilla as homologous, and hence both ellipsotaphrines and Bohemilla as derived from a common ptychoparioid ancestor. We have suggested below (p. 187) a different interpretation of the glabellar furrows of ellip- sotaphrines, one which would allow them to be included within the Cyclopygidae (from which they were excluded by Fortey, 1981). There is no feature of Bohemilla which indicates that it should be included within the Cyclopygacea. Deciding the affinities of these specialized pelagic trilobites is particularly difficult in the absence of stratigraphical and morphological interme- diates connecting them with known groups, and much depends on the interpretation of features which may have been profoundly modified in response to this mode of life. However, this probably does not apply to the mid-glabellar bulge on Bohemilla (Fenniops) sabulon, and taken with the form of the free cheek, the spinose pygidium and the narrow anterior cranidial border, this tends to favour a remopleuridacean origin for Bohemilla. The early species also shows a small interocular cheek (Fig. 22g), a character present on many early Apatokephalus-like remo- pleuridids. If Bohemilla is a remopleuridacean its origin is presumably independent of that of Opipeuter Fortey 1974 and Girvanopyge Kobayashi 1960 (= Cremastoglottos Whittard 1961a), two more extraordinary pelagic trilobites with supposed remopleuridacean origins (Fortey 1981). Pelagic morphology arose on five different occasions in the Ordovician; our view on how this may have happened is summarized in Fig. 62 (p. 188), a modification of Fortey, 1981: text-fig. 4. Bohemilla (Fenniops) sabulon sp. nov. (Figs 22a—g, 23) Hov.oryPe. Cranidium, with three thoracic segments detached, It.15939. Fig. 22b, c. PARATYPES. Cranidium and partial thorax, It.15940; cranidia: It.15942—5; incomplete exoskele- ton NMW 84.12G.31. TYPE LOCALITY. Type section of Pontyfenni Formation, loc. 23. STRATIGRAPHICAL RANGE. Found only at the type locality; upper Arenig, Fennian, biozone of Bergamia rushtoni. Name. Sabulon is a farm near the type locality. DiaGnosis. Bohemilla (Fenniops) with palpebral lobes about same length as occipital ring; distinct cranidial anterior border. Back margin of postocular fixed cheek subtends an angle of 40°—S0° to sagittal line. DESCRIPTION. Cranidia which are preserved in relief have a rather gentle transverse convexity; the fixed cheeks are disposed horizontally, while the palpebral lobes are elevated. Glabella excluding occipital ring of length equal to width between anterolateral corners; occipital ring about one-quarter length of preoccipital glabella. Course of axial furrows is sinuous: occipital ring tapers forwards, glabella then expands in width past 1P lobe to a maximum at 2P, tapers again past 3P lobe as far as 3P furrow, finally expanding in width once more towards the anterolateral corners of the glabella. The line connecting the anterolateral corner of the glabella with the posterolateral edge of the occipital ring is parallel to the sagittal line, and tangential to the median glabellar ‘bulge’. The occipital furrow is narrow, curved rather evenly backwards, very slightly shallower medially. Glabellar furrows narrow, slit-like, extending one-third of the way across the glabellar. 1P curves a little forwards and inwards and has a hooked inner end; 2P is transverse. This makes the 1P lobe wider (exsag.) inwards, while the 2P lobe is widest at its outer end. The 3P furrow is only half as long as the other two, and slopes slightly backwards. Frontal lobe of the glabella, that area in front of the 3P glabellar furrows, is twice as wide (tr.) as long (sag.), and slopes down peripherally into a very narrow but sharply defined preglabellar R. A. FORTEY & R. M. OWENS Fig. 22 Bohemilla (Fenniops) sabulon subgen. et sp. nov. Upper Arenig, Fennian, Bergamia rushtoni Biozone, loc. 23. a, incomplete cranidium and three thoracic segments, x 6, It.15940; b, c, holotype part and counterpart, showing thoracic pleurae, x 6, It.15939; d, cranidium, showing border well, It.15945, x 6; e, f, g, incomplete exoskeleton, NMW 84.12G.31; e, x 4; f, latex cast from pygidium under high contrast illumination, showing displaced seventh thoracic segment, x 12; g, detail of glabella, x 8. furrow. The specimen shown in Fig. 22d has a slight median dimple. Anterior cranidial border narrow, rim-like, distinctly wider at mid-line. It just curves around the anterolateral corner of the glabella, narrowing into a ‘gutter’ running to the palpebral lobe. The occipital ring carries a sagittal ridge which extends into a tiny tubercle at the posterior margin; the specimen in Fig. 22d shows a pair of posterior tubercles developed in a position comparable with the occipital spines of B. (Bohemilla) tridens Rushton & Hughes 1981. Palpebral lobe gently curved, sited adjacent to the glabella and symmetrically disposed about the 3P glabellar furrow; a faint, narrow rim is defined, outlining a crescentic interocular area. The palpebral lobe has a length (exsag.) slightly less than that of the occipital ring. Flat, triangular postocular cheek with maximum transverse width between 0-3 and 0-45 of that of adjacent glabella. Posterior margin inclined forwards at an angle between 40° and 50° to sagittal line. Narrow, bevelled posterior border, defined by shallow furrow. Suture runs inwards and forwards from lateral tip of fixed — cheek to palpebral lobe, in front of which it follows the ‘gutter’ to curve round the front margin of the cranidium. Cranidial surface probably minutely granulose. We have not found the free ARENIG IN SOUTH WALES 131 Fig. 23. Provisional reconstruction of Bohemilla (Fenniops) sabulon subgen. et sp. nov. It is based on the assumption that NMW 84.12G.31 shows the full complement of thoracic segments, and that the free cheek was like that of B. (Fenniops) klouceki and other bohemillids. cheek. Thoracic segments had minute triangular pleurae (Fig. 22c) like those figured by Rushton & Hughes (1981) on B. tridens, with prominent articulating boss adjacent to the axial furrows. Structure of axial rings much like that of occipital ring, with sagittal ridge and faint indication of a pair of tubercles along posterior margin. The most complete specimen (Fig. 22e) shows seven thoracic segments, but the pygidium is detached from the thorax and it is possible that there were additional segments; if this were so it is likely that there would be traces of them on the same bedding plane, since the specimen is otherwise articulated. We have used the seven thoracic segments in our tentative reconstruction, Fig. 23. The pygidium, the first associ- ated with a bohemillid, is not well preserved, but shows a short axis without rings, and clearly a pair of long posterior border spines; there was probably a second, shorter pair outside these but they are not so clearly shown. This suggests a pygidial structure comparable with a remopleuridid such as Robergiella. Discussion. The two species of B. (Fenniops) which require discussion are both from the Bohemian Ordovician: B. (Fenniops) praecedens Klouéek from the Llanvirn and B. (Fenniops) klouceki Marek from the Llandeilo. Both have been revised by Marek (1966). Whittard (1952: pl. 33, figs 13-16) illustrated four cranidia under open nomenclature, which Marek subse- quently referred to B. (Fenniops) klouceki. B. (Fenniops) praecedens is quite coarsely tuberculate, especially at the inner ends of the glabellar furrows. The glabella expands forwards such that the frontal lobe is the widest part. B. (Fenniops) sabulon is more similar to B. (Fenniops) klouceki. The median glabellar expansion is more developed on the latter, such that the width of the glabella behind the palpebral lobes exceeds that across the frontal glabellar lobe. The cranidial anterior border of klouceki has apparently become incorporated within the frontal glabellar lobe: Marek (1966: 150) records a ‘shallow inframarginal furrow’ which may corre- spond with the preglabellar furrow on sabulon. Three specimens figured by Marek (1966: pl. 1, fig. 7; pl. 2, figs 3, 8) show that the palpebral lobes on klouceki are slightly longer than the occipital ring, and hence larger than on sabulon. The 1p glabellar furrow on klouceki is much more hooked at its inner end. Two, perhaps three, of Whittard’s cranidia show the |p furrows united across the glabella. This may be an artefact of preservation, as his figs 13, 14 show obvious crushing, but if so it is puzzling that the 2p furrows are not so conjoined. In any case, the glabellar shape of Whittard’s figs 13-15 is more like that of klouceki than like sabulon. Family ASAPHIDAE Burmeister 1843 Fortey & Owens (1978: 260) discussed in detail the classification and differentiation of certain Arenig and related asaphids, and demonstrated a succession of Merlinia species in the Moridu- nian. Previously it had been assumed that most British Arenig asaphids belonged to ‘Ogygia’ or 132 R. A. FORTEY & R. M. OWENS ‘Ogygiocaris’ selwynii; this assumption dates from a statement by Thomas (in Strahan et al. 1907: 7) that ‘specimens of O. marginata have since been submitted to Mr P. Lake, who compared them with Salter’s types of hybridus from Henllan Amgoed. He is of the opinion that O. marginata and A. hybridus are identical, but that both must be referred to O. selwynii Salt.’. Whittard (1964: 232) upheld this synonymy. Such an assessment was perfectly reasonable, given the indifferent preservation of the types of hybridus and of much of the material from other localities (only recently has our collecting afforded well-preserved specimens). However, it has led to the erroneous correlation in particular of Arenig arenaceous deposits such as those described herein as the Ogof Hén, Abercastle and Blaencediw formations. The correct determi- nation of the asaphids is critical to a proper understanding of the stratigraphy. It is possible to identify three successive asaphid faunas in the Arenig of south Wales. Merlinia characterizes the Moridunian (see Fortey & Owens 1978), Ogyginus and Bohemopyge the Whitlandian and Asaphellus the Fennian. It should be noted that in other areas these genera are found outside these ranges. For instance Merlinia major (Salter 1866a) ranges throughout the Mytton Flags Formation in the Shelve inlier, which presumably incorporates strata of Moridunian, Whitlandian and Fennian age (see also p. 98). Elsewhere in Britain Merlinia has also been recorded from the Llanvirn of the Great Paxton borehole, Cambridge- shire (Rushton & Hughes 1981). Ogyginus is unknown from the British Fennian, but is common in the succeeding Llanvirn and Llandeilo (Whittard 1964, Hughes 1979), and Asaphellus occurs also in the Tremadoc. Subfamily ISOTELINAE Angelin 1854 Genus ASAPHELLUS Callaway 1877 (Synonyms: Asaphelloides Kobayashi 1937; Asaphoon Hutchison & Ingham 1967; Hemigyraspis Raymond 1910; Megalaspidella Kobayashi 1937; Plesiomegalaspis Thoral 1946). TYPE SPECIES. Asaphus homfrayi Salter 1866, by original designation. DIAGNOsIS. Subisopygous asaphids with flat to slightly concave cephalic and pygidial borders. Glabella subparallel sided, or with slight constriction at level of eyes; axial furrows often effaced but frontal glabellar lobe always clearly demarcated from border. Eyes small to medium-sized at, or slightly in advance of, cephalic mid-length. Preocular facial sutures sub- parallel and close to glabella, curving strongly adaxially near cephalic margin to meet at mid-line at highly obtuse point; postocular sutures strongly divergent and distally curving strongly backwards and becoming slightly recurved before cutting posterior cephalic margin. Hypostoma with elongate (sag.) oval outline, medially rounded, slightly concave or with small point. Pygidial axis well defined; pleural fields smooth to moderately furrowed, and up to eight pairs of ribs. Inner margin of pygidial doublure subparallel to pygidial margin. REMARKS. Arenig and Llanvirn asaphid species sharing the characters listed above have been placed in several different genera, in part because it has been fashionable to consider Asaphellus as a ‘Tremadoc’ genus and most of the others as ‘Arenig’. We believe that such a division is artificial, and when compared one with the other, the features by which they have been discriminated (e.g. size and position of eye, degree of taper of glabella, whether the posterior margin of the hypostoma is weakly concave or slightly pointed, definition of pygidial pleural ribs) can hardly be considered of higher than specific rank. We therefore follow the similar arguments of Pillet et al. (in Courtessole et al. 1985: 39) in placing such species in Asaphellus, and regarding the other names (listed above) as subjective synonyms. Asaphoon was based upon four tiny specimens which we think are likely to be meraspides or small holaspides of Asaphellus. Asaphellus whittardi (Bates 1969) (Figs 24a—h) 1914 Ogygia sp.; Cantrill in Strahan et al.: 15. 1964 Ogygiocaris murchisoniae (Murchison); Whittard: 238 (pars); pl. 37, figs 12, 13; pl. 38, figs 1-4 [non pl. 38, figs 5-11, = Merlinia murchisoniae]. 1969 Megalaspidella(?) whittardi Bates: 20, 22. 1978 Megalaspidella whittardi Bates; Fortey & Owens: 280. ARENIG IN SOUTH WALES 133 Ho.otyPe. BGS GSM85363, internal mould of cranidium (Whittard 1964: pl. 38, fig. 3). TYPE HORIZON AND LOCALITY. Fennian, D. hirundo Biozone, Tankerville Flags Formation; Bergam Quarry, Shelve inlier, Shropshire. MATERIAL. Numerous specimens have been recovered from the type locality (see Whittard 1964: 240); in south Wales the following material is known. From Fennian, S. abyfrons Biozone, Cwmfelin Boeth Formation: BGS TCC928/929, incomplete cephalon with seven attached thoracic segments from Survey loc. Carm. 37SW E26, Whitland Abbey; cranidia It.18910, It.18911, free cheek It.18912, pygidia It.18913-15, NMW 84.17G.la, b, 2a, b, all from locality 36, Cwmfelin Boeth. From presumed early Fennian: cranidium It.18916, from locality 20E, Capel-Dewi. DIAGNOSIS. Asaphellus with moderately inflated glabella, weakly laterally constricted with four pairs of weakly impressed furrows; thoracic and pygidial axes well defined, the latter with 10 rings; pygidial pleurae with 6—7 pairs of well-defined ribs; interpleural furrows weak and shallow; broad, concave pygidial border. DESCRIPTION. Glabella, including occipital ring, ranges from about 0-8 times as wide as long on large specimens to 0-6 times on smaller ones. From the posterolateral corners it first widens slightly to achieve its greatest posterior width at about half the distance to the posterior end of the palpebral lobe. It then narrows as far as a point opposite the posterior end of the palpebral lobe and then widens again to achieve its greatest anterior width (tr.), which is about the same as its greatest posterior width, just anterior to 3P furrows. A small sagittal tubercle is present opposite 1P lobes. In sagittal profile it is almost flat, the frontal lobe curving down gently and merging insensibly with the weakly concave preglabellar area, which is 0-15 times length (sag.) of glabella on large specimens, 0:25 on smaller ones. In transverse profile it is gently and evenly curved. Four pairs of weakly impressed muscle areas: 1P originates close to the axial furrow opposite anterior end of palpebral lobe, and directed obliquely forwards; 2P subparallel, a short distance behind 1P; 3P and 4P close together and of similar length and depth, opposite posterior part of palpebral lobe and directed nearly transversely. Lateral parts of occipital furrow indicated by weak, transversely elongated depressions similar to 3P and 4P and directed weakly obliquely backwards. Axial furrow shallow, deepest at its posterior extremity and in the stretch opposite the palpebral lobe. Palpebral lobe ranges from being 0-2—0-3 of length (sag.) of glabella plus occipital ring on small specimens to 0-12 on large ones, its outer margin elevated almost to height of sagittal region of glabella. Eye apparently narrow and crescentic. Preocular sutures more or less paral- lel, converging opposite frontal lobe of glabella in an even curve, meeting to form an obtuse point sagittally. Preocular section of fixed cheek markedly narrower than preglabellar area. Postocular suture defines a broad, subtriangular cheek. Pleuroccipital furrow wide and deep, defining narrow posterior border. Genal spine broad-based, with a broad, shallow median furrow which runs into a very weak lateral border furrow, which like preglabellar furrow is indicated only by a change of slope. Cephalic doublure broad, extending inwards almost as far as outer edge of eye, with fine, subparallel terrace lines (Fig. 24b; Whittard 1964: pl. 38, fig. 2). Reduced in breadth at hypostomal suture. Hypostoma unknown. Thorax with eight segments, axis narrow, well defined and scarcely narrowing backwards. Pleurae with rather shallow, oblique pleural furrows which are terminated laterally beyond the fulcrum by the posterior edge of the facet. Ends of pleurae falcate. Whittard (1964: 240) noted a panderian protuberance, succeeded posteriorly by an apparent panderian opening, on each pleura (Whittard 1964: pl. 38, fig. 4). Pygidium subparabolic. Axis long, narrow, well defined and tapering evenly backwards and terminating at the inner edge of the border. Ten rings present, separated by broad ring furrows that deepen laterally, producing the effect of paired furrows. Pleural areas gently convex with six or seven pairs of ribs with deep pleural furrows, the depth accentuated on internal moulds; ribs flat-topped, with weak interpleural furrows extending along their length. Both pleural and interpleural furrows terminate at inner edge of border. Border concave and broad, rising steeply 134 R. A. FORTEY & R. M. OWENS Fig. 24 Asaphellus whittardi (Bates 1969). Upper Arenig, Fennian Stage, S. abyfrons Biozone. a, cranidium, internal mould, x 2, It.18916, loc. 20E, Capel-Dewi; b, incomplete cephalon with seven attached thoracic segments, x 1:5, BGS TCC928, Cwmfelin Boeth Formation, Whitland Abbey. c-h, Cwmfelin Boeth Formation, loc. 36, Cwmfelin Boeth: c, cranidium, internal mould, x 1-25, It.18910; d, free cheek, latex cast of external mould, x 2, It.18912; e, incomplete pygidium, internal mould, x 1-5, It.18913; f, incomplete cranidium, internal mould, x 2, It.18911; g, incomplete pygidial doublure, internal mould, x 1-5, It.18914; h, incomplete pygidium, internal mould, x 2:5, It.18915. ARENIG IN SOUTH WALES 135 up towards pleural lobes. Pygidial doublure broad, ventrally convex, well seen on a fragmen- tary specimen from Cwmfelin Boeth (Fig. 24g). REMARKS. Whittard (1964: 238) conflated A. whittardi with the Moridunian species Merlinia murchisoniae, but Bates (1969: 20) demonstrated the distinctness of Whittard’s Tankerville Flags material. We have identified this species in the basal Fennian in the Whitland and Carmarthen areas and the description above includes both these and the Shelve specimens. The presence of A. whittardi in the early Fennian implies that the Tankerville Flags may be of early Fennian age. It is one of the few Tankerville species so far identified in south Wales; others such as Pricyclopyge binodosa eurycephala are longer-ranging. A. whittardi is distinctive among Asaphellus species in having well-defined pygidial axial rings and pleural furrows. In overall morphology it closely resembles A. graffi (Thoral 1946), type species of Megalaspis (Plesiomegalaspis) from the ‘lower middle Arenig’ of Cabriéres, Montagne Noire, but the latter differs in having a less inflated glabella, the eye proportionately further from the lateral margin, a longer preglabellar area, a broader pygidial axis and ill-defined axial rings and pleural ribs (except in small specimens). Thoral recognized several varieties of graffi as well as other asaphid species from the same locality. The name graffi was applied to large specimens in which the pygidial pleural furrows are more or less effaced; ‘varieties’ such as lata and major we suspect to be intraspecific variants. Plesiomegalaspis? convexilimbata Thoral (1946: 72; pl. 9, fig. 3; pl. 13, fig. 2; pl. 15, fig. 6) and P. angustirhachis Thoral (1946: 71; pl. 16, fig. 2) are based upon smaller specimens with better-defined pleural ribs; Megalaspis mucronata Thoral (1946: 56; pl. 7, fig. 2) and M. striatula Thoral (1946: 58; pl. 7, figs 3(?), 5) are based on even smaller ones. These appear to us simply immature A. graffi; Gigout (1951: 282; pl. 2, figs 1-5) figured similar small specimens from the Arenig between Casablanca and Mazagan (El Jadida), Morocco, as Plesiomegalaspis graffi. In none of these are the pygidial axial rings and pleural ribs as well defined as in A. whittardi, but a pygidium figured by Thoral (1946: 89; pl. 15, fig. 2) as Ogygiocaris? inflexicostata is very similar, differing only in having a narrower border, relatively broader pleural areas and more (9) pygidial pleural ribs; if not conspecific it must be very closely related. A new Asaphellus species discovered recently in sandstones of possible Whitlandian age in the Bangor area by A. Beckly (personal communication 1984) is clearly distinct from A. whit- tardi, although it generally resembles A. graffi. A further Asaphellus has been recovered from siltstones of controversial (Tremadoc or Arenig) age near Carmarthen (Cope et al. 1978: 196; Owens & Fortey 1982: 253). This specimen (NMW 78.1G.1) is also similar to A. graffi, but has a smaller eye and prominent terrace lines on the thoracic axis; the pygidium is unknown. Asaphellus lugneensis Pillet, Courtessole & Vizcaino (in Courtessole et al. 1985: pl. 4, figs 1-12; pl. 5, figs 1-8) from the Arenig Gres du Foulon Formation, Montagne Noire, is also similar to A. whittardi, but is distinguished by its shorter (sag.) preglabellar field, more distinct lateral border furrow and more effaced pygidial axial rings and pleural ribs. Subfamily NIOBINAE Jaanusson 1959 Genus BOHEMOPYGE Piibyl 1950 TYPE SPECIES. Ogygia discreta Barrande 1872, by monotypy. DiAGNnosis. Niobine asaphids which may attain a large size. Differing from Niobella in relatively narrow cephalic and pygidial axis, broader preglabellar field, and wider (tr.) postocular cheeks. Differing from Gog in having almost triangular to transverse pygidium with relatively well defined border, and without scalloped inner edge of pygidial doublure. Differing from Niobina in having a notched hypostoma. Discussion. Bohemopyge was hitherto known from B. discreta from the Llanvirn of Bohemia. It is one of a group of early Ordovician niobines differing from Ogygiocaris and Ogygiocarella in having a well-defined occipital ring. As the diagnosis indicates, it is very close to the Tremadoc genus Niobina, and the only important difference relates to the unforked hypostoma of the 136 R. A. FORTEY & R. M. OWENS latter, a primitive character of general occurrence in early asaphids. The relatively complete development of interpleural furrows on Niobina davidis is scarcely a generic character. Jaanus- son (in Harrington et al. 1959: O350) states that the pygidium of Bohemopyge lacks a border. This is a puzzling assertion, because the original of Barrande (1872: pl. 7, fig. 23) clearly shows one, as does the specimen of Novak & Perner (1918: pl. 3, figs 1, 2; re-illustrated by Horny & Bastl, 1970: pl. 5, fig. 6). Presumably the statement was based on Perner’s drawing which does not well represent the border. The known material of B. discreta is all small; assigning B. scutatrix (Salter) to the genus means that some large trilobites are included, and some of the differences they show from B. discreta may be ontogenetic. Short genal spines are present on B. discreta, being absent from B. scutatrix, but it is possible that they were reduced later in ontogeny. In any case, the presence or absence of genal spines is not considered a generic character. When undistorted, large pygidia of B. scutatrix have a triangular outline, which is different from the posteriorly truncate to gently rounded outline of Niobella. The cephalon of B. scutatrix is very like that of the type species of Gog, G. catillus Fortey 1975 from the Arenig of Spitsbergen, but the pygidium of that genus has a distinctively scalloped dorsal surface along the paradoublural line which might indicate its closest relatives are to be found in Ogygiocaris rather than Bohemopyge or Niobina. Pribyl & Vanék (1980) erected a genus Araiocaris based on Ogygiocaris araiorhachis Harrington & Leanza 1957 from the Arenig of Argentina. This form is also similar to Bohemopyge; it has more segments in the pygidial axis and the occipital structure may be more like that of Ogygiocaris. Ptibyl & Vanék considered differences from Ogygiocaris in their original discussion, but did not discriminate Araiocaris from Bohemopyge or any other asaphid. There are thus five trilobite generic names within this closely-knit group, which is more than the morphological variation really allows: Niobella, Niobina, Bohemopyge, Araiocaris and Gog. For the moment we retain Bohemopyge as the closest match for the Welsh species, while recording that it may be possible to include Bohemopyge within an enlarged concept of Niobina if one allows the same kind of hypostomal variation within this genus as in Niobe and Niobella. Bohemopyge scutatrix (Salter 1859) (Figs 25—28) 1859 Ogygia peltata Salter in Murchison: 54 (nom. nud.). 1859 Ogygia scutatrix Salter in Murchison: 52; Fossils (9) 1. 1866a Ogygia peltata Salter; Salter: 135-136; pl. 17, figs 8-10. non 1866a Ogygia scutatrix Salter; Salter: 133-134; pl. 17, figs 11-13. 1866b Ogygia peltata Salter; Salter in Ramsay; 313: pl. 12, fig. 8. 1867 Ogygia peltata Salter; Salter: 177-178; pl. 25*, figs 1-4. 1867 Ogygia bullina Salter: 178; pl. 25*, fig. 5. 1875 Ogygia peltata Salt.; Hicks: 176. 1928 Ogygia selwyni (Salter); Matley: 491. 1931 Niobe (Niobe) peltata (Salter) Reed: 446. 1946 Ogygia scutatrix Salter; Lake: 336. 1964 Ogyginus peltatus (Salter) Whittard: 246. 1984 Gog peltata (Salter); Fortey in Whittington et al.: 21. NOMENCLATURE. This species has usually been referred to Ogygia peltata Salter, 1866. However, Lake (1946) pointed out that the single specimen figured by Salter in 1859 as O. scutatrix, which is from Whitesand Bay, Dyfed, is the same as that subsequently used by Salter (1866a: pl. 17, fig. 8) as the type of O. peltata. In his text Salter (1866a: 133) definitely excluded the originals of his pl. 17, figs 9, 10 from his ‘peltata’ there, so there is little choice but to conclude that O. peltata is an objective synonym of O. scutatrix. Salter had evidently intended peltata to apply to the Arenig species from south Wales, and scutatrix to the Tremadoc species from north Wales, now known as Niobina davidis Lake, 1946. But we are obliged to follow the Rules of Zoological Nomenclature in using the name scutatrix. Ho.otyPe. BGS GSM7618. Incomplete dorsal exoskeleton from the old slate quarry north of Whitesand Bay; Penmaen Dewi Formation, Whitlandian. Original of Salter, 1859. Fig. 25a. Fig. 25 Bohemopyge scutatrix (Salter 1859). Middle Arenig, Whitlandian (G. gibbsii Biozone), Penmaen Dewi Formation, slate quarry at Pwlluog, north of Whitesand Bay, St David’s, Dyfed. a, holotype, imperfect dorsal exoskeleton, original of Salter, 1859, x 1, GSM 7618; b, small imperfect axial shield with dislocation of right side over thorax, type of Ogygia bullina Salter 1867, x 2, SM A16728; c, external mould of undistorted but flattened large pygidium and partial thorax, para- doublural line on left, x 2, SM A44344: d, small axial shield, flattened but otherwise undistorted, x 1,SM A44343a. 138 R. A. FORTEY & R. M. OWENS FIGURED MATERIAL. Dorsal exoskeletons and axial shields in varying degrees of completeness: BGS GSM7616; SM A16728—30, A33438, A44340, A44343-4; 1.14284; NMW 27.110.G251; hypostoma: BGS GSM 12873; cranidium: NMW 27.110.G253. DIAGNosIs. Bohemopyge with triangular pygidium on which interpleural furrows are present, but do not extend to axial furrows. LocaLitiEs. This is a widespread guide fossil to Whitlandian rocks. Most of the specimens in Museum collections derive from the type locality, north of Whitesand Bay in the Penmaen Dewi Formation. In south Wales, it has also been recovered from the Whitland Abbey Member of the Colomendy Formation (loc. 27, NMW 84.17G.39), from the Rhyd Henllan Member of the Colomendy Formation (loc. 47A, NMW 84.17G.40) and from the equivalent of the Whit- land Abbey Member in the stream immediately west of Capel-Dewi, east of Carmarthen (loc. 20F). In north Wales, it is confined to the Llyn Peninsula, where it has been recovered from two localities: mudstones in east side of Nant-y-Gadwen, and mudstones and shales in the track south of Dwyrhos Farm, Aberdaron. STRATIGRAPHICAL RANGE. Whitlandian (M. Arenig), Zone of Gymnostomix gibbsii. DESCRIPTION. Salter’s specimens from north of Whitesand Bay are large, flattened examples, with the indifferent preservation usual from that locality. Relatively well preserved specimens from north Wales are mostly smaller, some in relief, and such differences as there are from the type material can be attributed to the style of preservation. Whole exoskeletons are about 1-7 times as long as wide, with the cephalon, thorax and pygidium of equal length (sag.). Convexity is low (sag., tr.), the whole axis gently convex with a gentle downward slope on the pleural regions. Cephalon 0-6 times as long as wide. Cranidium twice as wide at posterior margin as at palpebral lobes. Glabella (including occipital ring) up to twice as long as wide at mid-length, but often somewhat less. Glabellar shape is a little variable, which may be partly attributable to preservation. The least distorted example (Fig. 28b) is fusiform with a truncately rounded anterior lobe; axial furrows are subparallel, with a slight expansion in front of, and a slight taper before, the palpebral lobes. Axial furrows are well defined, except at the bacculae, where they almost disappear. Occipital ring invariably well defined by deep occipital furrow, which is somewhat backward-curved medially. Some examples (Fig. 27b) have a glabella which is more obviously truncate at the front. The holotype of Ogygia bullina Salter, 1867, has an apparently subcircular frontal glabellar lobe. This is a small, ill-preserved specimen, and the right fixed cheek (and the thoracic pleurae behind) has been displaced towards the axis, thereby obscuring the right posterior part of the glabella. It is this displacement which produces an apparently narrow axis and distorts the shape of the glabella. All other features are those of scutatrix, and we regard this specimen as an unusually preserved example of that species, and bullina as a subjective synonym of scutatrix. Glabellar furrows are obscure in available preservation but what can be seen on the best-preserved glabella (Fig. 28b) are of typical niobine pattern (compare Fortey 1975: pl. 2, fig. 1; Horny and Bastl 1970: pl. 5, fig. 6). Palpebral lobes close to glabella, anterior ends almost reaching axial furrow, small, only about one-sixth glabellar length. The transverse line connecting the anterior limits of the palpebral lobes is at cephalic mid-length. Sutures diverge at 80° to sag. line behind the eyes, and only curve backward at the outer one-third of the fixed cheeks to cut the posterior margin at a right angle. Anterior branches of facial sutures diverge at a lesser angle in front of the palpebral lobes (30°—40°) before curving round the anterior cranidial margin to meet on the mid-line in an obtuse point. The preglabellar field so defined is about one-sixth glabellar length, and is flat. The wide postocular cheeks are about the same width (tr.) as the occipital ring; posterior border widens (exsag.) laterally. Salter originally (1859, 1866a) inferred a genal spine on the free cheek; later (1867) he portrayed a rounded genal angle, which we believe is correct. No specimens are well enough preserved to show the eye. Doublure (Fig. 28a) very wide beneath the free cheek and narrowing around the preglabellar furrow as is usual in niobines, carrying about twelve terrace lines parallel to its margin. Some specimens (Fig. 27a) show the trace of the median sutures. One ARENIG IN SOUTH WALES 139 Fig. 26 Bohemopyge scutatrix (Salter 1859). Middle Arenig, Whitlandian (G. gibbsii Biozone), Penmaen Dewi Formation, slate quarry at Pwlluog, north of Whitesand Bay, St David’s, Dyfed. a, large, imperfect axial shield showing doublure on pygidium, x 1, GSM 7616; b, hypostoma, x 1, GSM 12873. quite well preserved hypostoma (Fig. 26b) is typical of niobines, being very like that of Niobe emarginula (Tjernvik 1956: pl. 4, fig. 15). The median notch is hardly developed, but it is certain that there was no median acumination as in Niobina. The apparently rather transverse attitude of the maculae may be an artefact of flattening. Neither the hypostome nor the dorsal exoskeletal surface shows consistent evidence of sculp- ture. Niobines frequently have fine exoskeletal lines and ridges, and their apparent absence on most specimens may be a matter of their lack of preservation in argillaceous rocks. One of the north Wales relief specimens has indications of such fine ridges (Fig. 28d). Thorax with axis only slightly tapering backwards, width less than that of pleurae. Salter abandoned a distinction between forms with relatively wide axis and those with a narrower axis 140 R. A. FORTEY & R. M. OWENS Fig. 27 Bohemopyge scutatrix (Salter 1859). Middle Arenig, Whitlandian (G. gibbsii Biozone), Penmaen Dewi Formation, slate quarry at Pwlluog, north of Whitesand Bay, St David’s, Dyfed. a, external mould of imperfect dorsal exoskeleton, showing median suture and slightly truncate glabella, original of Salter 1867: pl. 25*, fig. 1 (as Ogygia peltata), x 2, SM A16729; b, glabella and part of thoracic axis, x 1,SM A33438. (allegedly female and male), recognizing the variability in this character. The largest specimens have axial width more closely approaching pleural width. Deep pleural furrows almost bisect the pleurae and run nearly to their tips. Tips of pleurae of anterior segment are bluntly rounded; posterior segments acquire short but distinct spines. Pygidium when undistorted always has a distinctly triangular outline, but a little transverse extension will destroy this (Fig. 28a). Pygidial length/width ratios fall in the range 0-51 to 0-75 for less distorted material; the smallest ratio belongs to the largest specimen. The narrow axis always occupies less than one-third, and on some specimens less than one-quarter anterior pygidial width, and continues gentle backwards thoracic taper to rounded tip at 0-8 of pygidial length. Eight axial rings are discernible, occasionally a faint ninth, and the less flattened material shows that the terminal piece was distinctly elevated above the border. Eight pairs of pleural furrows, with a short ninth pair adjacent to the terminal piece, deep and narrow and running to border. Interpleural furrows are much weaker, and correspond to the outer parts of the pleural furrows only; on some flattened specimens they may be almost obliterated. Trans- verse extension (Fig. 28a) tends to overdeepen them. Border not sharply defined, especially laterally, and may have been gently sloping behind the axis rather than horizontal. Doublure of slightly less width than that of pygidial axis, concave inner margin over which tips of pleural furrows pass, and with about 12 terrace lines. Axial half rings appear to increase in length progressively (sag.) forwards from pygidial half ring. ARENIG IN SOUTH WALES 141 Fig. 28 Bohemopyge scutatrix (Salter 1859). Middle Arenig, Whitlandian. a, small exoskeleton in partial relief, slightly extended transversely, which exaggerates pygidial interpeural furrows and increases resemblance to Niobina, Nant-y-Gadwen, Llanfaelrhys, Aberdaron, Llyn, north Wales, Whitlandian mudstones, probably G. gibbsii Biozone, x 1, NMW 27.110.G251; b, well-preserved incomplete cranidium in relief, same locality, x 2, NMW 27.110.G253; c, small incomplete dorsal exoskeleton, in relief, showing slightly acuminate pygidial margin, mudstones below Dwyrhos Farm, Aberdaron, Llyn, probably G. gibbsii Biozone, x 2, 114284; d, incomplete axial shield, flattened but undistorted otherwise (compare Fig. 28c), G. gibbsii Biozone, Penmaen Dewi Formation, slate quarry at Pwlluog, north of Whitesand Bay, St David’s, Dyfed, x 0:5, SM A44340; e, large pygidium retaining posteromedian acumination, but more transverse than Fig. 28d, same locality, original of Salter 1867: pl. 25*, fig. 4, x 0:5, SM A16730. 142 R. A. FORTEY & R. M. OWENS Discussion. As discussed above, B. scutatrix is morphologically intermediate between Niobina davidis and Bohemopyge discreta. Apart from the difference in the hypostoma, which is pointed in N. davidis, the Tremadoc form has consistently longer interpleural furrows on the pygidium, which extend to the axial furrows. The pygidial doublure is also wider (Lake 1946: pl. 47, fig. 2). The type species of Bohemopyge apparently lacks pygidial interpleural furrows entirely (Barrande 1872: pl. 7, fig. 23) and is more transverse; the pygidial border widens backwards slightly but is not pointed medially. Cranidia of B. scutatrix and B. discreta are nearly identical; better-defined glabellar furrows on the latter (Horny & Bastl 1970: pl. 5, fig. 6) are probably because of the superior preservation of Bohemian material. The presence of genal spines on B. discreta is not necessarily of importance, because the known material is small, and asaphid genal spines often become shorter during ontogeny (e.g. Fortey 1980: 260). Whittard (1964) placed scutatrix (‘peltata’) in Ogyginus. The cranidium, however, is so typi- cally niobine, especially with regard to the good definition of the occipital ring and the sharply defined preglabellar field, that such an assignment is likely to be incorrect. Subfamily OGYGIOCARIDINAE Raymond 1937 Genus OGYGINUS Raymond 1912 TYPE SPECIES. Asaphus corndensis Murchison 1839, by original designation. DIAGNOSIS. Preocular suture intramarginal but may run very close to margin in some individ- uals (Hughes 1979: 126); glabella flask-shaped, widest (trans.) across frontal lobe; up to three pairs of weak muscle impressions; occipital ring ill-defined; anterior border narrow (sag.), flat or weakly convex; cephalic border furrow crosses onto preocular cheek only in early species, where it is weak, but typically terminates at the preocular facial suture; hypostome with entire posterior margin, which is acuminate; thoracic segments typically with Z-shaped or zetoidal axial furrows; pygidial axis up to 11 rings; pleural areas with 7-9 pairs of pleural ribs, pleural furrows typically deep and broad, interpleural furrows when present are weak. REMARKS. Hughes (1979: 122) followed the diagnosis of Whittard (1964: 245), but we have found it necessary to modify it here in order to include Arenig species and to exclude those characters common to many other asaphids. Hughes (op. cit.) gave the stratigraphical range of the genus from the Lower Llanvirn to the upper part of the Lower Llandeilo and noted its occurrence in Wales, Shropshire and Brittany, regarding records from elsewhere as either very tentative or invalid. Several Arenig species are attributable to Ogyginus; they are O. hybridus (Salter 1866a) and O. sp. indet. (see below) from south Wales, O. armoricanus (Tromelin & Lebesconte 1876) from the Grés Armoricain supérieur, Brittany (Henry 1971: 66, pl. 1, figs 1-11; 1980: 37, pl. 1, figs 4, 5, 7), O. terranovicus Dean (in Dean & Martin 1978: 15; pl. 5, figs 4-9; pl. 6, figs 1-7; pl. 7, figs 1—4) from the Wabana Group, Bell Island, eastern Newfoundland, O. planus (Thoral 1946: 86; pl. 11, fig. 6; pl. 15, fig. 7) from the ‘lower middle Arenig’, Montagne Noire, and O. orbensis Pillet, Courtessole & Vizcaino (in Courtessole et al. 1985: 44; pl. 6, figs 1-8) from the Gres de la Cluse de Orb and Grés du Foulon formations, Montagne Noire. The earliest of these are O. planus, O. orbensis and O. hybridus; Courtessole et al. (1985) gave the age of the two former as lower Arenig, but by analogy with the stratigraphical occurrence of O. hybridus in the Whit- landian, these species may also be of a similar age; this can only be resolved when the correlation between Wales and the Montagne Noire becomes better understood. The oldest O. terranovicus occur with graptolites indicating an horizon near the top of the D. extensus Biozone (Dean in Dean & Martin 1978: 3), suggesting equivalence to the later Whitlandian, whilst O. armoricanus is probably of a similar age (Henry 1980: 223). At Cwm yr Abbey, in the stratotype section for the base of the Whitlandian, O. hybridus occurs only a short distance above the youngest Merlinia rhyakos (Fortey & Owens 1978), and the ranges of the two species probably overlap in the basal few metres of the Whitlandian. The name hybridus is more apposite than Salter can ever have imagined, for this species has characters typical of both Ogyginus and Merlinia; of the latter, the presence of the cephalic border furrow on the pre- ) | | ; ARENIG IN SOUTH WALES 143 ocular fixed cheek, the rather elongated hypostome and the weak pygidial pleural furrows may be noted in particular. Ogyginus characters include the flask-shaped glabella and the very narrow anterior border; assignment to Ogyginus is based upon the acquisition of the latter features. It seems likely that Ogyginus has its origins in Merlinia, and in addition it may be noted that some M. rhyakos (e.g. Fortey & Owens 1978: pl. 5, fig. 4) show a tendency towards developing zetoidal thoracic axial furrows, a feature that Whittard (1964: 245, 246) stressed as an important attribute of Ogyginus; also immature M. selwynii pygidia (Fortey & Owens 1978: pl. 8, fig. 4; pl. 9, fig. 4) are strikingly similar in shape and proportions to those of adult O. hybridus. Ogyginus hybridus (Salter 1866a) (Figs 29-32) 1866a Asaphus? (Basilicus) hybridus Salter: 153; pl. 23, figs 8, 9. 1906 Ogygia marginata, var.; Evans: 608. 1906 Ogygia Selwynii Salt.; Evans: 608. 1907 Asaphus? (Basilicus) hybridus Salter; Thomas in Strahan et al.: 7. 1907 Ogygia marginata? Crosfield & Skeat; Thomas in Strahan et al.: 15. 1914 Ogygia selwyni (Salter); Cantrill & Thomas in Strahan et al.: 14 (table). 1914 Ogygia selwyni (Salt.); Thomas in Strahan et al.: 18. 1914 Asaphus (Basilicus?) hybridus Salt. [Ogygia selwyni Salt.]; Thomas in Strahan et al.: 18. 1931 Basilicus? hybridus (Salter); Reed: 451. 1964 Ogygiocaris selwyni (Salter) Whittard: 232; pl. 35, fig. 10. 1982 Ogyginus cf. hybridus (Salter); Owens & Fortey: 251, 256, 257. LecToTyPE (selected Whittard 1964: 237). BGS GSd3151, internal mould of pygidium. TYPE STRATUM AND LOCALITY. Salter (1866a: 154) stated this to be ‘Llandeilo Flags?, Henllan Amgoed, Carmarthenshire’, but in the plate caption (1866a: pl. 23, figs 8, 9) modified the horizon to ‘Caradoc Shales’. Since only Arenig strata crop out in the environs of Hennlan Amgoed, it may safely be assumed that both these alternatives are incorrect. Thomas (in Strahan et al. 1914: 18) considered that exposures in the lane from Llwyn-derw to Felin Henllan Amgoed (loc. 46, Colomendy Formation, Rhyd Henllan Member) were probably those from which the type specimens of this species were obtained. The lithology of the matrix of Salter’s specimens is certainly that of the Rhyd Henllan Member, but it is equally possible that they originated from other exposures in the vicinity of Llwyn-derw (e.g. locs 47A, B) which have also yielded the species. Thus while we are confident this is the general area from which the types originated, we would hesitate to regard Thomas’ suggestion as definitive. OCCURRENCE. Type stratum: locs 46A, B, 47A, B, Rhyd Henllan. Blaencediw Formation: loc. 39, Rhyd Henllan; loc. 56, Triffleton. Abercastle Formation: loc. 58, Abercastle. Afon Ffinnant Formation: locs 16H-L, Cwm yr Abbey; locs 17A, B, Cwm Arbont; locs 18A—F, Afon Ffin- nant; loc. 19, Ffinnant road section. DIAGNOsIS. Ogyginus with short length of cephalic border furrow on preocular fixed cheek, effaced on larger specimens; hypostome with small tongue on posterior margin; thoracic and pygidial pleural areas very broad, so that anterior width of pygidial axis is markedly less than distance between axial furrow and inner edge of facet; pygidial pleural ribs weakly defined, more accentuated on smaller specimens, where interpleural furrows are also present; dorsal exoskeleton smooth, terrace lines restricted to cephalic and pygidial margins and lateral parts of thoracic pleurae. DESCRIPTION. This species may reach a large size, and a thorax plus pygidium from loc. 16L (Fig. 30b) has a sagittal length of almost 80mm, and the complete specimen would have been perhaps 120mm long. The glabella expands forwards so that the transverse distance across the frontal lobe is on average about 80% of the sagittal length; it is narrowest at mid-length, opposite the palpebral lobe. It is only weakly inflated both in sagittal and transverse profiles, and the axial furrows are only weakly defined, especially posterior of the palpebral lobe. Any degree of crushing renders the glabellar muscle areas difficult to see, as in much of our material. mee “ pie em ‘ SS. aces OO Be Nl! 2 Fig. 29 Ogyginus hybridus (Salter 1866a). Middle Arenig; a, d, f, h, i, k from Whitlandian, Colomendy Formation, Rhyd Henllan Member, Henllan Amgoed, locs 46A (i), 46B (h), 47A (a, f, k), 47B (d); b, c, e, g, J, 1,m from Whitlandian, Afon Ffinnant Formation; Cwm yr Abbey, loc. 16L (g); Afon Ffinnant, locs 18C (b, e, j, 1, m) and 18E (c). a, axial shield, x 2, It.18917; b, meraspid pygidium, latex cast of external mould, x 20, It.18918; c, small cranidium, x 2:5, It.18919; d, incomplete thorax with elongated pygidium, x 1-5, It.18920; e, meraspid, latex cast from external mould, x 10, It.18921; f, cranidium showing weak furrow on preocular cheek, x 3, It.18922; g, small cranidium, x 5, NMW 84.10G.2; h, pygidium with parts of five thoracic segments, x 2, BGS HT518; i, small pygidium, latex cast from external mould, x 2, BGS Pr1846; j, free cheek, x 2, It.18923; k, pygidium, latex cast from external mould, x 2:5, It.18924; 1, small pygidium, x 4, It.18925; m, incomplete large cranidium showing muscle impressions, x 2, It.18926. ARENIG IN SOUTH WALES 145 Fig. 30 Ogyginus hybridus (Salter 1866a). Middle Arenig, Whitlandian Stage, Afon Ffinnant Forma- tion; a—d, f, g, i from Cwm yr Abbey, locs 16K (a, g) and 16L (others); e, h from Afon Ffinnant, loc. 18E. a, thorax and pygidium, latex cast from external mould, x 1-25, NMW 84.17G.1b; b, large thorax and pygidium, x 1, NMW 84.10G.6a; c, cranidium, latex cast from external mould, x 2:5, NMW 84.10G.1a; d, incomplete free cheek, x 2, NUW 84.10G.3a; e, disarticulated thorax with broad pygidium, x 2, It.18927; f, incomplete large pygidium, x 1, NMW 84.10G.7a; g, pygidium, latex cast from external mould, x 2, NMW 84.17G.180a; h, small thorax and pygidium, x 2:5, It.18928; i, hypostoma, x 2, NMW 84.10G.10. 146 R. A. FORTEY & R. M. OWENS They are proportionately deeper on large specimens (Fig. 29m) which show probably four pairs; none of our material shows the basal (1p) pair well, but transverse 2p is well seen in Fig. 29m. The same specimen shows a forwardly oblique 3p anterior to the palpebral lobe, and in front of this there is a suggestion of a weak 4p. The palpebral lobe is placed at approximately mid-length and close to the axial furrow, although is proportionately further forward on large cranidia. On the latter, the postocular suture describes a more sharply backward oblique course than on smaller ones, and on crossing the pleuroccipital furrow it turns inwards (well seen on some free cheeks, e.g. Fig. 29j). The preocular suture runs close to, and diverges forwards slightly from, the frontal lobe of the glabella. Many specimens have a short, shallow length of the cephalic border furrow traversing the preocular fixed cheek (e.g. Fig. 29f), but this is absent from all specimens from locs 18C and 18E (e.g. Figs 29c, m) which however correspond in other details with specimens in which it is present. The pleuroccipital furrow is broad and marked, terminated laterally at the genal angle. It delimits a narrow posterior border. Lateral border furrow broad and rather shallow, not well seen on our material. Genal spine short, broad-based. Doublure rather narrow, of similar width to the border, narrowing at hypostomal suture. Hypostome broadly triangular, very large specimens being as long (sag.) as wide (tr.), smaller ones 85-90% as wide as long. This is in contrast to those of O. corndensis which are wider than long, even in smaller specimens (e.g. Hughes 1979: 131, figs 58, 59) and also of smaller specimens of O. armoricanus which are as wide as long (e.g. Henry 1980: pl. 1, fig. 7a). O. hybridus also differs in having rather long (exsag.) anterior wings. Maculae prominent, lying in front of narrow, crescentic posterior body. Posterior margins acuminate, like Merlina rhyakos (Fortey & Owens 1978: pl. 5, figs 1, 2). Fine terrace lines laterally run subparallel to the margin, and are transverse on the sagittal part of the anterior body. Thorax of typical Ogyginus type, with zetoidal axial furrows and spindle-shaped pleural furrows. Pleurae very broad, so that distance from axial furrow to the inner end of the fulcrum is always markedly greater than the transverse width of the axis. Pygidium with a long, narrow axis with up to 12 rings. These are defined by very shallow, transverse ring furrows which are deepest laterally and almost effaced sagittally. Axis weakly convex, and hardly elevated above adjacent pleural areas, and terminates at inner edge of doublure. Pleural areas gently convex and very broad, so that distance from axial furrow to inner edge of fulcrum is almost invariably in excess of transverse width of anterior end of axis, although in a few specimens (e.g. Fig. 29d) the values are almost equal. Seven to eight pairs of weakly-defined ribs, almost effaced on larger specimens where traces of only more anterior pleural furrows are discernible; interpleural furrows can be seen only on small specimens (Figs 29b, 1). Abaxial parts of pleural areas slope down quite steeply (seen on specimens preserved in relief) to narrow pygidial border. Pygidial doublure narrow, about same width as border. Dorsal exoskeletons smooth, apart from terrace lines around margins of cephalon and pygidium and at lateral extremities anterior and poste- rior pleural bands of thoracic pleurae. The collection from loc. 18C includes several meraspides, mostly pygidia which show well- defined axial ring furrows and deep, parallel pleural and interpleural furrows on the pleural areas. The only cephalon is ill-preserved, but an isolated free cheek (Fig. 31b) shows a pro- portionately longer genal spine and larger eye than on mature specimens. REMARKS. Intrinsic factors such as the subtle differences upon which many asaphid species are defined, and extrinsic ones including type of sediment and distortion from different directions conspire to make the specimens described here as Ogyginus hybridus particularly intractable to interpret. The most obvious differences are seen in the pygidia. Some (e.g. Fig. 30e) are decidedly broader than are others (e.g. Fig. 30h). In an attempt to express this objectively we measured the anterior end of the pygidial axis and plotted this against the distance from the axial furrow to the inner end of the articulating facet; the result showed that two groups could just be differentiated (Fig. 32). Wider specimens are nearly all from the mudstones of the Afon Ffinnant Formation, less wide ones from silty horizons within it, or from similar lithologies in ARENIG IN SOUTH WALES 147 be ie | re en] ea AD ee % gi i Fig. 31 Ogyginus hybridus (Salter 1866a). Middle Arenig, Whitlandian Stage, Afon Ffinnant Forma- tion; a, c-f from Ffinnant road cutting, loc. 19; b from Afon Ffinnant, loc. 18C; g from Allt Cwm-arbont, loc. 17A; h from Penmaen Dewi Formation, probably lower part, Whitesand Bay. a, cranidium, x 2, It.18929; b, small free cheek with long genal spine, x 5, It.19004; c, d, pygidia, respectively x 3, It.18930, x 2, It.18931; e, free cheek, x 1-5, It.18932; f, hypostoma, x 3, It.18933; g, pygidium, x 2, It.18934; h, tectonically distorted complete specimen, note tiny genal spine on left side, x 2, SM A44355. the Colomendy, Blaencediw and Abercastle formations, although wider forms do occur also in most of these. Comparatively few cephalic remains are known; on cranidia the only marked variation occurs within the Afon Ffinnant mudstones; specimens from locs 18C and 18E (Figs 29c, m) lack the border furrow on the preocular fixed cheek, whilst all others have it. Free cheeks mostly have a short, broad-based genal spine, but those from loc. 19 (Fig. 3le) have a proportionately longer one, whilst specimens from the Penmaen Dewi Formation (Fig. 31h) have a minute genal spine. Unfortunately these cephalic and pygidial differences are not distributed in a consistent way throughout our samples, and it is not possible at this stage to suggest what, if any, taxonomic significance they might have, or whether they are a product of intraspecific variation, perhaps exacerbated by effects of preservation. We suspect, however, that there might be two sub- species; the types of hybridus (Whittard 1964: pl. 35, fig. 10) belong to the comparatively narrower morph. 148 R. A. FORTEY & R. M. OWENS mm 14 al 13 12 11 % 10 7 S S & & 9 ; 6 5 A) 5 Rs £8 Sy 7 3 oid i = 7 K e Cwm yr Abbey 3s S 4 Ffinnant 36 ey a RK & Ffinnant road x + Cwm-arbont 5 Pe A Oo Rhyd Henllan eNO Ce vy Blaenweneirch 4 bd 4 = lectotype (0) 1 2 3 4 5 6 7 8 9 10 11 #12 13 #14 15 16 17 %18 mm distance to facet Fig. 32 Plot of anterior axial width against distance from axial furrow to inner end of articulating facet on pygidia of Ogyginus hybridus (Salter), to show distribution of ‘wider’ and ‘longer’ forms from various localities. Thoral (1946: 86; pl. 11, fig. 6; pl. 15, fig. 7) described Ogygiocaris? plana from the ‘lower middle Arenig’ of Cabriéres, Montagne Noire. His three syntypes are large specimens, each with the pygidium and part of the thorax preserved, and they compare exactly with our broader forms of O. hybridus, and there can be no doubt that they also belong to Ogyginus. Thoral had no cephalic material of O. planus but Pillet, Courtessole & Vizcaino (in Courtessole et al. 1985: 43) figured as Ogygiocaris? plana cranidia (1985: pl. 2, figs 1-3) and free cheeks (pl. 2, figs 4, 5) from different localities in the Montagne Noire in the upper part of the Grés et Schistes de La Maurerie and the Gres de la Cluse de l’Orb formations. The cranidia differ from O. hybridus in having a longer preglabellar field and the frontal lobe of the glabella being poorly defined; no specimen has a section of the cephalic border furrow on the preocular fixed cheek. On the assumption that these specimens are conspecific with Thoral’s, O. plana is evidently distinct from, but clearly closely related to, O. hybridus. O. orbensis is distinguished from O. hybridus by a longer (sag.) preglabellar field, the preocular facial suture being close to the glabella, no cephalic border furrow on the preocular fixed cheek and a wider base to the genal spine. The pygidium is proportionately longer than in most O. hybridus. Ogyginus sp. indet. (Figs 33a, b) MATERIAL. It.18935, cephalon with 7 thoracic segments from loc. 47B, Rhyd Henllan. NUW 33.189.G152, pygidium and 8 segments from ‘Henllan Amgoed’. Both from Colomendy Forma- tion, Rhyd Henllan Member. ARENIG IN SOUTH WALES 149 Fig. 33. Ogyginus sp. indet. Middle Arenig, Whitlandian Stage, Colomendy Formation, Rhyd Henllan Member. a, cephalon with seven thoracic segments, Rhyd Henllan, loc. 47B, x 1-5, It.18935; b, thorax and pygidium, Henllan Amgoed, x 0-8, NMW 33.189.G152. DESCRIPTION. The cephalon is badly preserved, but shows a very long, broad-based genal spline that extends backwards as far as the sixth thoracic segment; the eye is forwardly placed, close to the glabella, which apparently has a weak constriction at mid length. The thoracic axial furrows have the characteristic zetoidal configuration of Ogyginus. The thoracic segments have the inner parts of the pleurae markedly narrower (tr.) than the axis. The pygidium is 0-6 times as long (sag.) as wide (tr.), with the axis tapering quite rapidly backwards; the pygidial doublure is rather broad. The anterior part of the glabella and the thoracic axial rings carry prominent terrace lines. REMARKS. The blade-like genal spine, narrow inner sections of the pleurae, broad pygidial doublure and presence of terrace lines on the axial regions readily distinguish these specimens from O. hybridus, which occurs at the same locality. A new species is evidently represented, but the present material is insufficient to name it. Superfamily CYCLOPYGACEA Raymond 1925 We follow here the definition of Cyclopygacea given by Fortey (1981), and include the Nileidae in this superfamily as well as the Cyclopygidae. Family CYCLOPYGIDAE Raymond 1925 Subfamily CYCLOPYGINAE Raymond 1925 DiaGnosis. Cyclopygids with five to seven thoracic segments, no nodes on third axial ring. Cranidium with an arcuate to parabolic outline, in some genera extended into a ‘nose’ anteri- orly; glabellar furrows variably expressed, with a maximum of three pairs in Novakella. Pygidium with or without border; axis usually wide and relatively short, and may be more or less effaced. GENERA INCLUDED. Cyclopyge Hawle & Corda 1847; Microparia Hawle & Corda 1847; Micro- paria (Heterocyclopyge) Marek 1961; Microparia (Quadratapyge) Zhou 1977; Aspidaeglina Holub 1911; Sagavia Koroleva 1967; Degamella Marek 1961; Novakella Whittard 1961a; Xenocyclopyge Lu 1962; Gastropolus Whittard 1966. Discussion. Since the Treatise (Harrington et al. 1959) the number of cyclopygid genera has trebled. The Arenig of south Wales includes the earliest known diverse cyclopygid assemblage, particularly rich in the Fennian, where they are accompanied by two other, morphologically ‘}X9} UI passnosip viouad Jo sorydiowodeiny “410M Sty} Ul PastAes sv aepiskdojokeD Ajrumey oy) uryiIM sdrysuoneyel Jo weisopelyD pe “BIYy Japiog jeipibAd poulyap—|jom soka paiydosjiadAy sjuawbes 9198104} 2-9 syaouo aalj pasny x1417e0190dSO01d aeueb jo uoljonpal Bijaqe|O o1oqeied sajoiaqn} jeixe paiied Ssixe jeip!1bAd y10YUS sixe jeipibAd seynBueisy Ajaynoe 1SO] SMOJIN}y seyJeqe|b pooe}yjyo Ssaj 10 a1OW 19P410q jeipibAd wnipibAd j$O JuUgWaDe}Jo [exe sjuawbaes 9198104} G ie abAdojoAdt4g SAoyIiswwy sdosA{ydwahs (36A4d0/949019}9H) \ (e14edossiy) e1edosoly einebes e1sedosoipy abAdosaAI0uUs8 xX ejpawebaq e/JaYeAON abAdojoAD (Gee ee oa aes at Te Tae sees ee ot) AVNIDAGOTINAD Ls ay Nin Ad079A01Nd —— ARENIG IN SOUTH WALES 151 convergent but unrelated, trilobites Girvanopyge and Bohemilla. It is clear that many of the cyclopygid genera—Microparia, Cyclopyge, Degamella, Novakella, Gastropolus—are extraordi- narily conservative. Species from the Arenig of south Wales are very similar to ones from the Ashgill of Scotland, Bohemia and Kazakhstan. This may reflect the stability of the off-shelf habitat in which they lived, relatively free from the perturbations that influenced the shelf faunas. The genera appear to be quite well defined by the later Arenig, and presumably the earlier history of the family must be sought in the early Arenig and Tremadoc of such areas as the Montagne Noire. Kobayashi & Hamada (1971) made a start on recognizing subfamilies within the Cyclopygidae with the discrimination of the Ellipsotaphrinae (p. 187). Here we separate another monophyletic subfamily, the Pricyclopyginae, to include Pricyclopyge, Symphysops and ? Emmrichops. The remaining genera are accommodated within the Cyclo- pyginae. The subfamily includes cyclopygids with mostly elongate (sag.) cranidia with a para- bolic outline and the maximum width at or near the posterior margin. Palpebral rims are narrow, gutter-like; anterior fusion of the eyes occurs in several genera. Thoracic segment number is constant within genera through the Arenig—Ashgill interval, but we know that primitively cyclopygids may have had seven segments, and loss of one or two segments presum- ably happened in the later Tremadoc or early Arenig. For example, Cyclopyge azaisi Thoral, 1935, from the ?latest Tremadoc of the Montagne Noire is essentially a Degamella with seven thoracic segments. The most primitive cyclopygid is Prospectatrix Fortey, 1981, from the Tremadoc of Britain, which also has seven segments. As is usual with primitive forms its assignment to a subfamily, when all are based on advanced characters, is difficult; because it lacks the thoracic structure of the Pricyclopyginae, it may be arbitrarily placed in the Cyclo- pyginae. A series of vincular notches have been recognized on the free cheeks of Cyclopyge and Degamella (Fig. 40b); this may prove to be a uniting character of Cyclopyginae, or it may be generally distributed in the family. The hypostome is known in Cyclopyge, Microparia and Degamella: it is wide (tr.) and short (sag.), and quite different from that cautiously assigned by Whittard (1961a) to Pricyclopyge. Genus CYCLOPYGE Hawle & Corda 1847 TYPE SPECIES. Egle rediviva Barrande 1846. Discussion. Fortey & Owens (1978) redescribed Aeglina grandis Salter, 1859, and attributed it to Microparia. Since then, the discovery of new material has added more information. The assignment of grandis to Microparia was incorrect; well-preserved cranidia show the basal pair of glabellar furrows characteristic of Cyclopyge, together with a typical pygidial structure. We follow the diagnosis of Marek (1961) here. Cyclopyge grandis grandis (Salter 1859) (Figs 35a-f, 36) Synonymy as in Fortey & Owens (1978: 256, 258) under Microparia grandis but their pl. 4, fig. 9 is excluded. TYPE INFORMATION. See Fortey & Owens 1978: 258. FIGURED MATERIAL. Axial shields: NMW 27.110.G259, G362, I.711; cranidium: It.19600; pygidium: It.19601. STRATIGRAPHICAL OCCURRENCE. Middle to basal Upper Arenig (Whitlandian—lowest Fennian). Frequent in Penmaen Dewi Formation at Pwlluog; Whitland Abbey Formation, loc. 27; Cwmfelin Boeth Formation, in turbidites and interbedded shales, locs 36, 43; unnamed Arenig mudstones on east side of Nant-y-Gadwen, near Llanfaelrhys, Llyn Peninsula, north Wales, and equivalent beds at Dwyrhos Farm, near Aberdaron; one pygidium from the Mytton Flags, Shropshire. The specimen assigned here by Fortey & Owens (1978: pl. 4, fig. 9) from the Carmarthen Formation is now referred to ?Cyclopyge sp. (see Table 1, p. 85). 152 R. A. FORTEY & R. M. OWENS Discussion. When we first redescribed this species we based our interpretation on poorly preserved type and topotype material, and a well-preserved incomplete dorsal exoskeleton from the Carmarthen Formation (Fortey & Owens 1978: pl. 4, fig. 9) which we assumed was conspecific. This assumption was wrong. Well-preserved specimens from Whitlandian mud- stones in north Wales compare exactly with the poor material from the type locality, but differ from the specimen from the Carmarthen Formation, which now has to be excluded from the species. The cranidia show the broadly truncate-oval form, long palpebral rims and spine-like relict fixed cheeks typical of Cyclopyge, as well as the basal pair of glabellar furrows, which are subdued on flattened material. A specimen in full relief from the Cwm-felin-Boeth Formation (Fig. 35b) also shows a sagittal ridge, running forwards from the glabellar furrows to the glabellar tubercle. Pygidia were correctly assigned previously; the new material includes well- preserved examples showing four or five pairs of extremely narrow furrows probably represent- ing interpleural boundaries. These do not survive distortion or flattening. The border is narrow, and well defined along its length, with a concomitantly narrow doublure. The axis tapers evenly backwards, and one axial ring is invariably deeply defined; the second ring is only faintly indicated adjacent to the axial furrows, and rarely defined. Length of the axis exceeds the length of the postaxial field to the border furrow (see discussion of Cyclopyge grandis brevirhachis below). This is possibly the oldest known species of Cyclopyge; both those attributed to this genus by Tjernvik (1956) from the early Ordovician of Sweden lack the typical glabellar structure, and were referred to Heterocyclopyge by Lu (1975: 175). A species from the Montagne Noire described by Pillet & Courtessole (1985) as Incisopyge? theroni is closely similar to C. grandis grandis, and is described as originating from the Lower Arenig; it is certainly better referred to Cyclopyge. The type species from the Caradoc of Bohemia was revised by Marek (1961: 19-21; in Horny & Bastl 1970: pl. 7, fig. 3), and is like C. grandis grandis in most features of the exoskeleton, again showing how conservative was cyclopygid morphology. It is best distin- guished by its short pygidial axis; on Marek’s reconstruction the terminal piece is shown as hardly longer than the one defined axial ring. There are also one or two pairs of pygidial pleural furrows. On the cranidium, the axial furrows on C. rediviva take a much sharper adaxial bend inside the relict postocular cheek. Other species of Cyclopyge differ from C. grandis grandis in pygidial details: pygidial axes are either shorter, as in C. rediviva, or if as long have more defined axial rings and furrowed pleural fields, as in C. kossleri (Kloucek 1916). Several species (C. rotundata Lu 1975; see Han 1978) has the eyes fused into a single unit—a derived character in several cyclopygid genera. The eyes of C. grandis were separate. There is variation in the preservation of the raised circular areas (probably representing a thinning of the cuticle) Fig. 35 a-f, Cyclopyge grandis grandis (Salter 1859). a, cranidium and thorax retaining some relief, Nant-y-Gadwen, Llanfaelrhys, Llyn Peninsula, Whitlandian (probably G. gibbsii Biozone), x 5, NMW 27.110.G362; b, c, uncrushed internal mould of incomplete cranidium, dorsal and lateral views, Showing basal pair of muscle impressions and sagittal ridge, Cwmfelin Boeth Formation, early late Arenig, Fennian, S. abyfrons Biozone, loc. 36, x 5, It.19600; d, three thoracic segments and pygidium, near original relief, locality as Fig. 35a, x 5, NMW 27.110.G259; e, entirely undistorted pygidium, horizon and locality as Fig. 35b, x S, It.19601; f, thorax and pygidium from type locality, showing indication of second axial ring, Whitlandian (G. gibbsii Biozone), old quarry in Pwlluog, north of Whitesand Bay, St David’s, Dyfed, x 4, 1711. g—h, j-o, Cyclopyge grandis brevirhachis subsp. nov., Upper Arenig, Fennian. g, small cranidium and 3 thoracic segments, most transverse specimen, B. rushtoni Biozone, loc. 24, x 5, NMW 21.306.G32; h, mould of pygidium, full relief, stratigraphical and morphological intermediate between C. grandis grandis and C. grandis brevirhachis, S. abyfrons Biozone, loc. 36, Cwmfelin Boeth Formation, x 4, It.19603; j, thorax and pygidium, B. rushtoni Biozone, loc. 23, Pontyfenni Formation, x 5, It.19602; k, pygidium, horizon and locality as Fig. 35), x 6, It.15906; 1, holotype incomplete dorsal shield with free cheeks displaced backwards, B. rushtoni Biozone, loc. 24, x 4, It.15931; m, cranidium, cast of external mould, horizon and locality as holotype, x 5, It.18578; n, pygidium, horizon and locality as holotype, x 4, It.15904; 0, cranidium, horizon and locality as holotype, x 5, It.19607. p, Cyclopyge cf. C. umbonata (Angelin). Axial shield, Upper Arenig, Fennian, Biozone of B. rushtoni, loc. 23, x 5, It.15908. Id3 ARENIG IN SOUTH WALES 154 R. A. FORTEY & R. M. OWENS which lie behind the inner ends of the glabellar furrows, which are obliterated during any compression. Cyclopyge grandis grandis is a useful guide fossil to the Whitlandian; the earliest Fennian specimens from the Cwmfelin Boeth Formation (Fig. 35e) cannot be distinguished from those from the earlier horizons. Later Fennian specimens all have the characters of C. grandis brevirhachis, which is discussed below. We noted previously that Whittard’s (1961a) Cyclopygid C was to be referred to C. grandis grandis, and indeed provides one of the faunal links between the Mytton Flags and the Arenig of south Wales. Cyclopyge grandis brevirhachis subsp. nov. (Figs 35g—o, 36) Ho ortype. Incomplete dorsal exoskeleton, It.15931. PARATYPE MATERIAL. Thorax and pygidium, It.19602; pygidia, It.15904, It.15906, It.15907, NMW 84.17G.171; cranidia: It.18677—8, It.19607, NMW 21.306.G32, 84.17G.172; cranidium with free cheek, NMW 84.12G.29. TYPE LOCALITY AND HORIZON. Pontyfenni Formation, loc. 24; Upper Arenig, Fennian, B. rush- toni Biozone. STRATIGRAPHICAL RANGE. Upper Arenig, Fennian, B. rushtoni Biozone; locs 23, 24. Name. ‘Short axis’. DraGnosis. A subspecies of Cyclopyge grandis with length of pygidial axis equal to or less than length of postaxial field, measured from the tip of the axis to the border furrow on the sagittal line. DISCUSSION. This subspecies is established for populations of Cyclopyge grandis from the Ponty- fenni Formation which differ from the the nominate form in having a relatively long pygidium, and hence the pygidial axis occupying a proportionately shorter fraction of the pygidial length. This can be expressed as the ratio of the length of the axis, measured from the furrow defining the half ring to the furrow defining the tip, to the postaxial field from the tip of the axis to the middle of the border furrow. The holotype of C. grandis grandis is poorly preserved and the tip of the axis and the border furrow are obscured, so that it is not possible to obtain a reliable measurement of this proportion. However, better preserved, flattened but otherwise almost undistorted pygidia are known from the type locality at Pwlluog, St David’s (one of these is shown on Fig. 35f), and five of these have given ratios of between 1-20 and 1-33 with a mean of 1:28. From the good material in full or partial relief at Nant-y-Gadwen, Llyn, we obtained ratios between 1:20 and 1-60 with a mean of 1-30. All available material of C. g. brevirhachis yields ratios 0-85—1-0, mean 0-95. Pygidia of C. grandis grandis are typically nearly semicircular if undistorted; width/length ratios of C. g. brevirhachis are in the range 1-6—1-8. This proportion is readily altered by distortion, however. On pygidia of brevirhachis the width of the terminal piece on the axis is equal to or less than its length; the terminal piece on grandis grandis is wider than long, but again axial or transverse distortion obscures this difference. Where the pygidial doublure can be observed on C. grandis brevirhachis it is wider than on C. grandis grandis, which has the merest strip underlying about half the border. If the proportions of the thorax and pygidium of brevirhachis from the type locality of the Pontyfenni Formation are not greatly altered, as seems likely, then its thorax is also proportionately longer than that of C. grandis grandis: width only slightly more than length, whereas undistorted C. grandis grandis have length about two-thirds width. We can find no differences between the two subspecies on the cranidia. Free cheek with border, as C. grandis grandis (Fortey & Owens 1978: pl. 4, fig. 8). In an undistorted condition pygidia of the two subspecies can be distinguished quite easily but the differences can be obscured by distortion—particularly transverse extension, which reduces the measurable proportional distinctions. Specimens from the upper part of the range of grandis grandis in the basal Fennian Cwmfelin Boeth Formation are not different from those ARENIG IN SOUTH WALES 155 Fig. 36 Comparative reconstructions of (left) C. grandis grandis (Salter), to replace that in Fortey & Owens, 1978, and (right) C. grandis brevirhachis subsp. nov., both about x 3. in the Whitlandian, although one specimen (Fig. 35h) has the axis hardly greater in length than the postaxial field, and might be regarded as transitional in this character. The slender pygidial axis with elongate terminal piece is not known in other Cyclopyge species; in those with axes as short, the terminal piece is usually rather bluntly rounded. Cyclopyge alia and C. festa from Kazakhstan (Koroleva 1967) are species with short pygidial axes which are difficult to assess from the available illustrations; both are from the Upper Ordovician and are unlikely to be related to C. grandis brevirhachis. Both also appear to be similar to C. marginata Hawle & Corda (Marek 1961: pl. 1, figs 11, 12), which has a wider, blunt-tipped pygidial axis compared with the new subspecies. Cyclopyge kossleri Kloucek 1916 (Figs 37a, b) 1916 Aeglina kossleri Holub (in litt.); Klouéek: 8. 1961 Cyclopyge kossleri (Klouéek 1916) Marek: 25-26; pl. 1, figs 14-17; text-fig. 7. MATERIAL. One incomplete dorsal shield, SM A44533. LOCALITY AND HORIZON. Llanfallteg Formation, earliest Llanvirn, Scolton Railway Cutting. Discussion. This is the first record of the species in Britain. It has been fully described from the Llanvirn Sarka Formation of Bohemia by Marek (1961). Type material is in full relief, whereas the specimen from the Llanfallteg Formation is slightly flattened, and the cranidium is also distorted. The latter, however, shows the characteristic pair of furrows which prove assignment to Cyclopyge. The pygidium of C. kossleri is distinctive: it has a long and well-segmented axis showing three axial rings and a small triangular terminal piece, and the pleural fields carry two pairs of furrows. It is, in fact, similar to transitory pygidia of other cyclopygids (e.g. C. umbon- ata bohemica; see Marek, 1961: pl. 2, fig. 5), but is too large to be such, and is associated with the full complement of thoracic segments. C. kossleri probably arose by neotenic displacement of immature characters into the adult. The specimen from south Wales clearly shows the same pygidial features, and the determination can be made with confidence. 156 R. A. FORTEY & R. M. OWENS a BREE see pt bea Re a Fig. 37 Cyclopyge kossleri Klouéek, 1916. Early Llanvirn, D. artus Biozone, Llanfallteg Formation, Scolton cutting, x 6, SM A44533. a, cranidium and external mould of incomplete thorax and pygidium; b, counterpart of latter. Cyclopyge cf. umbonata (Angelin 1854) (Fig. 35p) cf. 1854 Corynexochus ? umbonatus Angelin: 60; pl. 33, fig. 10. cf. 1961 Cyclopyge umbonata umbonata (Angelin) Marek: pl. 1, fig. 10. cf. 1971 Cyclopyge umbonata (Angelin 1854); Neben & Krueger: pl. 13, fig. 19. MATERIAL. One nearly complete axial shield, It.15908. LOCALITY AND HORIZON. Pontyfenni Formation, type locality at Pontyfenni; Upper Arenig, biozone of Bergamia rushtoni. DISCUSSION. One articulated specimen from the late Arenig falls outside the range of variation of C. grandis, and in particular of C. g. brevirhachis from the same horizon. The pygidial border is very wide, about one-quarter total pygidial length; on C. grandis, sensu lato, there is some variation in this character, but it is never wider than one-sixth pygidial length, and usually one-eighth or less. The pygidial axis is also relatively wide, and bluntly rounded, when com- pared with that of C. grandis. The pygidium of C. umbonata, Angelin’s original specimen and presumably the holotype by monotypy, was illustrated by Marek (1961); its origination from the ‘Orthoceras Limestone’ of Sweden suggests it is not very different in age from the Welsh specimen (see Neben & Krueger 1973). Neben & Krueger (1971) attributed a cranidium, but from a glacial erratic. C. umbonata also has a wider pygidial border than other Cyclopyge spp., and is close to our specimen on most other characters. The identification has to be qualified, however. The Welsh specimen shows a faint second axial ring on the pygidium, whereas the type of umbonata has apparently one strong ring—more like C. grandis in this respect. The cranidium attributed by Neben & Krueger was probably longer than that of our specimen. It is impossible to assess the impor- tance or otherwise of such small differences until more is known about variation in C. umbon- ata from the type area, hence our determination is qualified. Another similar species is C. stigmata Poulsen 1965 from Bornholm, also from the later Arenig. However, Poulsen mentions a third pygidial axial ring, and the pygidial border narrows postaxially. Genus DEGAMELLA Marek 1961 TYPE SPECIES. Aeglina princeps Barrande 1872, original designation. DiaGnosis. Large cyclopygid trilobites having elongated cranidium, occupying as much as half total length, which is produced into an anterior ‘nose’. Eyes of medium to large size with ARENIG IN SOUTH WALES 15)7/ laterally-directed field of view. Three pairs of cephalic muscle impressions may or may not be visible, and never deeply incised as in Novakella. Thorax of six or seven segments without dorsal organs on third segment. Pygidium sub-semicircular; axis long, usually effaced pos- teriorly; border ill-defined; pygidial doublure wide. Discussion. Marek (1961: 45, footnote) and Whittard (1966: 287) stated that Degamella was likely to be a junior synonym of Novakella Whittard 1961a. Here we retain them as separate genera, and distinguish both from Microparia also at the generic level. The south Wales faunas show that the typical morphologies of Degamella, with extended glabella, and Novakella, with incised slit-like glabellar furrows, are already distinct in the Arenig, and indeed they continue with little change into the Ashgill of the Whitehouse Group, Scotland (J. K. Ingham, personal communication). Such long independent histories of two distinctive groups, each with several species, is stratigraphical evidence for two separate monophyletic taxa, for which generic status is appropriate. Degamella and Novakella share a number of characters which indicate that they are more closely related one to another than to Microparia: six thoracic segments; long cranidium; wide pygidial doublure. The morphology of Microparia was also established by the Arenig, with a similarly long subsequent history, and its generic status seems to be equally justified. Five thoracic segments appears to be a stable character in Microparia. Microparia nudus Whittard 1961a, from the Llanvirn of Shropshire, has six thoracic segments, and propor- tions more like those of Degamella princeps than typical Microparia species; hence we would refer it to Degamella, as defined here. We allow seven thoracic segments in Degamella because a specimen which has been attributed (possibly incorrectly) to Aeglina azaizi Thoral, 1935 (Fig. 38e) appears to be a typical Degamella except in that it retains the primitive segment number. Degamella evansi sp. nov. (Figs 38—40) Hovorype. Axial shield, It.15909. PARATYPE MATERIAL. Axial shields|s NMW _ 33.189.G133, It.15932, NMW _ 33.189.G3, 84.17G.41-3; exoskeleton with displaced cheeks, It.19707; free cheeks, It.15910—1, It.19610; cranidium, It.15912; hypostoma, It.19611; pygidia, It.19612, NMW 84.17G.44. TYPE LOCALITY. Pontyfenni Formation, loc. 23; Fennian, B. rushtoni Biozone. STRATIGRAPHICAL RANGE. Upper Arenig, Fennian, B. rushtoni Biozone, loc. 23; and Ramsey Island, Aber Mawr (? B. rushtoni Biozone), loc. 62. Name. After D. C. Evans, pioneer geologist of the Whitland District. DiaGnosis. Degamella with cranidium longer (sag.) than width at posterior margin; eyes larger than in D. princeps. Narrow palpebral rims extend very far back; no posterolateral fixed cheeks defined. Pygidial doublure of equal width along its length. DESCRIPTION. The species is known from three well-preserved axial shields, which are 2} times longer than wide, an elongate form characteristically associated with an actively swimming mode of life (Fortey 1985b). The cranidium accounts for half this length, and the length of the thorax is equal to, or slightly less than, that of the pygidium. The axial shields are somewhat flattened, but the cranidium illustrated in Fig. 40c is preserved in the ‘nodular’ fashion in which much of the original convexity survives; if this is so then the transverse vaulting is lower than it is in D. princeps. Convexity across thorax and pygidium is always low in Degamella. Cranidium longer than wide, the ‘nose’ broadly rounded rather than acuminate; cranidium expands in width (tr.) slightly forwards to a maximum at about one-third cranidial length, tapering more rapidly forwards. There is no indication of glabellar furrows or ‘muscle insertion areas’, but the preservation is likely to be inadequate to preserve them. Palpebral gutters only just fall short of posterior margin; anteriorly they run outwards alongside the glabella as it expands in width, and shortly in front of this, at about half cranidial length, they become ventral enough to be concealed beneath the frontal glabellar lobe. Fused free cheeks are 158 R. A. FORTEY & R. M. OWENS Fig. 38 a-—d, Degamella evansi sp. nov. Upper Arenig, Fennian, B. rushtoni Biozone, Pontyfenni Formation, loc. 23. a, axial shield, sixth thoracic segment not clearly visible on this specimen which has cranidium slightly displaced relative to thorax, presumably during moulting, x 2, NMW 33.189.G133; b, latex cast from natural mould of small dorsal shield, only four thoracic segments visible (possibly indicating meraspide condition if not displaced), x 4, It.15932; c, holotype, large damaged axial shield, x 2, It.15909; d, latex cast from external mould of hypostoma, x 5, It.19611.e, Degamella azaisi (Thoral)?, latex cast from well-preserved shield for comparison with D. evansi, early Arenig, Montagne Noire, x 2, It.8158. undoubtedly to be referred here, both on grounds of size and because they are closely similar to those of D. princeps (Marek 1961: text-fig. 18). They show the broad doublural platform which lay ventrally beneath the glabellar ‘nose’; but the eyes are large, taking up virtually all the dorsal librigenal surface, except for a very narrow border, and they presumably occupied the entire area abutting the palpebral rims. Hypostome (Fig. 38d) is extremely transverse, almost four times wider than long, densely covered with terrace ridges, and with a prominent posterior border but a less distinct anterior border. It is like that attributed to Degamella cf. princeps by Whittard (1966: pl. 49, fig. 12), and is reasonably assigned to D. evansi. Six thoracic segments of usual cyclopygid form, with adaxial articulation on the first segment, and somewhat further from axis posteriorly. Axis occupies about two-thirds width at ARENIG IN SOUTH WALES 159 ome Fig. 39 Reconstruction of axial shield of Dega- mella evansi sp. nov., x 2 approx. first segment, but less than half on sixth segment. Most of the axial taper occurs on the posterior three segments. Prominent articulating facets, steeply downsloping on first segment. The holotype shows the narrow thoracic doublure, which apparently narrows to almost nothing on the first segment. The half rings are raised, rim-like, and longer (sag.) anteriorly, indicating that the bulk of flexure during enrollment occurred on the anterior segment. Pygidium five-eighths as long as wide, with pygidial axis occupying about three-eighths pygidial width at anterior margin. Axial furrows continue backward taper of posterior part of thorax, enclosing an angle of 30°—40°; tip of axis effaced, but it certainly extends to nearly two-thirds pygidial length. One ring furrow is defined, with traces of a second. Broad, but ill-defined, border hardly present postaxially. Wide doublure extends back beneath border, not Narrowing conspicuously towards anterolateral edges, and carrying 8-10 terrace ridges which run slightly oblique to the posterior pygidial margin. Discussion. Considering the difference in age, this species is remarkably like the type species, D. princeps princeps (Barrande 1872: pl. 14, figs 3-8), from the Llandeilo Dobrotiva Formation of Bohemia (see Marek 1961: 46-48; pl. 4, figs 1-7; text-figs 17, 18; in Horny & Bastl 1970: pl. 7, fig. 8). Thorax and pygidium are identical. The cephalic impressions shown on Marek’s (1961: text-fig. 17) reconstruction are visible only with exceptional preservation—many specimens from the Dobrotiva Formation do not show them—and the fact that none of the Arenig specimens preserve them is of no importance. The distinguishing characters of the new species are entirely cephalic: in D. princeps princeps the axial furrows curve adaxially away from the facial sutures as the posterior margin is approached, before becoming effaced, thereby defining a long (exsag.), triangular fixed cheek; on D. evansi the axial/palpebral furrow continues along the edge of the cranidium almost to the posterior margin. The eyes on D. princeps are remark- ably small for a cyclopygid, and there is concomitantly a relatively large border on the free cheek; this is not the case in D. evansi, on which the eye occupies most of the available area as 160 R. A. FORTEY & R. M. OWENS ’ Fig. 40 Degamella evansi sp. nov. Upper Arenig, Fennian, B. rushtoni Biozone, Pontyfenni Formation, loc. 23. a, internal mould of conjoined free cheeks from ventral side, narrow, long eyes preserved at lateral edges, x 2-2, It.15910; b, external mould of conjoined free cheeks, vincular noches shown on right, x 3, It.19610; c,d, x 2, It.15912, dorsal and lateral views of cranidium; e, detail of eye showing its large size as compared with D. princeps, x 6, It.15911. it does on other cyclopygids. The close similarity between D. evansi and D. princeps might imply that the smaller eye of the latter is in this case a derived character, the larger eye being primitive. The occurrence of the small-eyed form is itself surprising, because the family as a whole maximizes visual area. Kloucek (1916) distinguished the subspecies Degamella princeps praecedens from a strati- graphically earlier horizon than D. princeps princeps, in the Llanvirn Sarka Formation of Bohemia. The holotype, a fragmentary cephalon, was illustrated by Marek (1961: pl. 4, figs 8, 9), and considered by him to be unique. In the collections of the British Museum (Natural History) there is a nearly complete exoskeleton of a Degamella (1.15277), presented by Kloucek, and stated to be from D1 of Sarka and therefore likely to represent the subspecies praecedens. This exhibits no differences from D. princeps princeps, and the specimen differs from D. evansi in the same characters; in view of this the status of the subspecies praecedens is regarded as doubtful. D. gigantea (Barrande 1872) from the Ashgill of Bohemia has even smaller eyes than D. princeps. Degamella nuda (Whittard 1961a) from the Llanvirn of Shropshire differs from D. ARENIG IN SOUTH WALES 161 princeps and D. evansi in that the cranidium of the mature holaspis (1961la: pl. 24, fig. 5) is much less than half the total axial length. The pygidium is like that of D. evansi externally, but the doublure widens noticeably backwards; small cranidia of D. nuda are quite transverse. D. nuda, like D. evansi, has large eyes. Whittard identified a series of vincular notches in the cheek doublure of D. nuda, as did Marek (1961) on D. princeps, and Han (1978) on Cyclopyge. These notches may prove to be another character linking the subfamily Cyclopyginae as defined here. If the large eyes of D. evansi and D. nuda are primitive it would confirm Marek’s view that Microparia and Degamella are closely related. Genus GASTROPOLUS Whittard 1966 1974 Lisogoraspis Apollonov: 76. TYPE SPECIES. Gastropolus brevicaudatum Whittard 1966, see below. D1AGnosis. Cyclopygid trilobites with six thoracic segments, and cephalon more than twice as long (sag.) as the pygidium. Cranidium resembling that of Microparia but with large, triangular remnant fixigenal areas. Pygidium exceptionally transverse, with an elliptical outline; border and axis well defined. DISCUSSION. The genus Gastropolus was originally proposed by Whittard (1966) on the basis of two specimens of a distinctive pygidium, together with a pygidium with six thoracic segments attached, all from the Hope Shales (Llanvirn) of Shropshire; he was not able to associate cephalic parts. We are now able to diagnose the genus more completely because there is a complete, but poorly-preserved, specimen from the Lake District (Fig. 41a), which allows us to associate relatively well-preserved cranidia and pygidia from south Wales. Mr R. Kennedy has also kindly shown us an entire specimen he has collected from the Llanfallteg Formation. Gastropolus is a distinctive cyclopygid, especially because of its excessively wide pygidium, which is like that of no other trilobite. The cranidium is more typically cyclopygine, but the comparatively large remnant fixed cheeks form a pair of wide triangular areas at the posterior cephalic margin, which distinguishes the cranidium from those of Microparia, Degamella and Sagavia. Apollonov (1974) described Lisogoraspis from the Upper Ordovician of Kazakhstan, again from pygidia. These are similar in all respects to that of Gastropolus, and Lisogoraspis is accordingly listed as a subjective synonym of Gastropolus. Whittard proposed the species G. brevicaudatum as the type species but we regard it as being the same as the poorly known species obtusicaudata Hicks from Llanvirn quarry. In its cranidial morphology Gastropolus is clearly a cyclopygine, related to Microparia, Sagavia and Degamella, which also have triangular fixigenal remnants, although not as wide (tr.). The presence of six thoracic segments, and a well-defined pygidial border and axis, incline us to classify it closer to Degamella and Sagavia than to Microparia. It is a homoeomorph of Emmrichops, another transverse cyclopygid, but which we regard as a probable pricyclopygine (below). Microparia itself includes one species, M. porrecta sp. nov., which has a pygidium as transverse as that of Gastropolus. Presumably there was a particular niche in the bathypelagic habitat to which these forms were adapted. Unlike other cyclopygids there is a poor ‘fit’ between cephalon and pygidium and it is not obvious whether Gastropolus was capable of enrollment. It is interesting that the hypostoma of Degamella is also remarkably wide (Fig. 38d). If that of Gastropolus were similar it is possible that the wide pygidium engaged with the wide hypostoma rather than the cephalic doublure during enrollment. Gastropolus obtusicaudatus (Hicks 1875) (Figs 41—43) 1875 Aeglina obtusicaudata Hicks: 185; pl. 10, fig. 3. 1885 Group A, no. 2; Postlethwaite: 74; pl. 3, fig. 16. 1886 Aeglina obtusicaudata Hicks; Postlethwaite & Goodchild: 463; pl. 8, fig. 16. 196la Cyclopyge obtusicaudata Hicks; Whittard: 179. 1966 Gastropolus brevicaudatum Whittard: 294; pl. 50, figs 10-12. 162 R. A. FORTEY & R. M. OWENS Fig. 41 Gastropolus obtusicaudatus (Hicks 1875). a, cast of poorly preserved axial shield, original of Postlethwaite 1885: pl. 3, fig. 16 (as drawn by Goodchild). Outerside, Lake District, x 1, SM A45444; b, holotype, thorax and pygidium in poor preservation, original of Hicks 1875: pl. 10, fig. 3, Llanvirn quarry, early Llanvirn, presumed to be D. artus Biozone, x 1,SM A45154. Hovotype. Thorax and pygidium, SM A45154, from the early Llanvirn of Llanvirn quarry. MATERIAL. Pygidia: It.19615—6; cranidia: It.19613—4. DIAGNOSIS. See that of genus. STRATIGRAPHICAL RANGE. The species occurs in the Llanfallteg Formation, early Llanvirn part, in south Wales at Cefn-maen-llWyd farmyard. Hope Shales, Shropshire, early Llanvirn; Skiddaw Slates at Outerside, probably also from the earliest Llanvirn. DESCRIPTION. Entire exoskeleton slightly less than twice as long as wide, the cephalon account- ing for half the total length. Cranidium wider than long, maximum width at posterior margin between 0:7 and 0-9 of sagittal length. The glabella has a nearly circular outline; the axial furrows converge inwards and backwards quite sharply, but fade out before reaching the posterior cranidial margin. The fixed cheeks so defined have a maximum width about three- quarters length (exsag.) and are slightly inflated. The sutural margins of the remnant fixed cheeks are gently convex-outwards. Palpebral rims are apparently lacking. Otherwise, the cranidium is almost featureless; there are faint indications of a few terrace lines on the frontal glabellar lobe. We presume that large eyes lay along the cephalic margin as in other cyclo- pygids. Thorax with six segments. Apart from being relatively wide (tr.) it is of usual cyclopygid construction, with the axis hardly tapering over the first three segments, thereafter tapering Fig. 42 Gastropolus obtusicaudatus (Hicks 1875). Early Llanvirn, D. artus Biozone, Llanfallteg Formation, loc. 50. a, large cranidium with partially damaged frontal glabellar lobe, x 2, It.19613; b, cranidium, x 3, It.19614;c, pygidium, x 3, It.19615. ARENIG IN SOUTH WALES 163 gently backwards into the pygidial axis. On the sixth segment the width of the axis is only slightly greater than the width of the pleurae. Pygidium three times as wide as long, with gently convex pleural fields, and a well-defined flat to gently convex border. Axis occupies one-third or less of pygidial width at anterior margin, and this width exceeds its length; posterior taper very gentle to broadly rounded tip at border furrow. Although the half-ring is prominent no axial rings are defined. Posterior border widest posterolaterally. Doublure coincident with border, carrying very sparse terrace ridges. Discussion. Although the Lake District specimen is not well preserved it is clearly identical to the species from the Llanfallteg Formation and from Shropshire. The latter was originally described with the specific name brevicaudatum Whittard. Whittard (1966) described the thorax of the articulated specimen from the Hope Shales as ‘almost barrel-shaped’. We regard this observation as mistaken, because the barrel-like appearance of Whittard’s pl. 50, fig. 10 is produced by partial breakage of the thoracic axis at its anterior end. Comparison of all this material with the type of obtusicaudatus Hicks is more difficult because of the imperfect preser- vation of the unique specimen from Llanvirn quarry. It is, however, certainly a Gastropolus, showing the six segments, transverse pygidium and short pygidial axis typical of the genus. The pygidial pleural fields are slightly less transversely extended but it is doubtful whether much importance can be attached to this in view of the degree of distortion often found in Llanvirn quarry specimens. Since almost all the species originally described by Hicks from Llanvirn quarry have also been recognized in the early Llanvirn of the Llanfallteg Formation it is likely that this applies to the Gastropolus as well. We have one fragment of a thorax and pygidium from Llanfallteg which appears to be more similar to Hicks’ specimen. We therefore feel obliged to use the Hicks name for all these early Llanvirn Gastropolus specimens, and G. brevicaudatum Whittard accordingly becomes a subjective junior synonym of obtusicaudatus Hicks. Gastropolus mirabilis (Apollonov) from the upper Ordovician of Kazakhstan is known from pygidia which appear to be indistinguishable from those of G. obtusicaudatus. Apollonov (1974: pl. 3, figs 10, 11) figured a cephalon as ‘Cyclopygidae gen.’ which is of the appropriate kind to belong with the Gastropolus pygidia, and which has even wider triangular fixed cheeks than G. brevicaudatum. His ‘gen. et sp. indet 2’ (1974: pl. 21, fig. 11) is a transitory pygidium with 3 thoracic segments which is likely to be that of Gastropolus. Fig. 43 Reconstruction of axial shield of Gastro- polus obtusicaudatus (Hicks), x 3 approx. 164 R. A. FORTEY & R. M. OWENS Genus MICROPARIA Hawle & Corda 1847 TYPE SPECIES. Microparia speciosa Hawle & Corda 1847, by original designation. DiAGnosis. Middle-sized cyclopygids with five thoracic segments. Cranidium with parabolic outline, lacking incised glabellar furrows. Front of glabella never extended into a long ‘nose’, and cephalic doublure correspondingly short (sag.) Eyes extend almost whole length of crani- dium and may become fused anteriorly. Pygidium usually with weakly-defined border best developed posterolaterally; doublure narrow to wide. Pygidial axis wide and short, not convex (tr.), either poorly defined, with up to one or two rings weakly indicated (subgenus Microparia), or clearly defined with three axial rings and a minute triangular terminal piece (subgenus Heterocyclopyge). DIscussIon. Our concept of Microparia differs in some respects from that of Marek (1961). We exclude from the genus those large cyclopygids with six thoracic segments and a long glabellar ‘nose’ which are placed in Degamella or Novakella herein. On the other hand we include Heterocyclopyge Marek 1961, as a subgenus of Microparia. Marek originally discriminated Heterocyclopyge from Microparia on the basis of its ‘almost subtetragonal’ pygidium, with the pygidial axis distinctly defined. The close relationship between Microparia and Heterocyclopyge is demonstrated by M. broeggeri from south Wales, described below. In that species there is variation in the definition of the pygidial axis, which is posteriorly effaced in the majority of specimens, and typical of Microparia, but in others is defined almost to the tip (Fig. 44d); the pygidium is long (sag.) compared with many Microparia spp. Variation within M. broeggeri bridges the characters of Microparia and Heterocyclopyge rather well. The two morphologies appear to separate in the Llanvirn and later. It may be that M. broeggeri is one of the rare examples where an ancestral species can be identified. In any case it proves that Microparia and Heterocyclopyge are more closely related to one another than to Degamella or Novakella, and subgeneric status for Heterocyclopyge is used to express this relationship. Zhou (1977) proposed a subgenus Microparia (Quadratapyge) for cyclopygids with five thoracic segments, well-defined, short pygidial axis, and very wide flattened border. This pygi- dial morphology is more different from that of Microparia (Microparia) than that of Microparia (Heterocyclopyge), and if further species of Quadratapyge can be recognized it may warrant full generic status. Subgenus MICROPARIA Hawle & Corda 1847 DiaGnosis. Microparia with pygidial axis only clearly defined anteriorly, if at all. Microparia (Microparia) broeggeri (Holub 1912) (Figs 44a-e, 45a-—e, 48) 1912 Aeglina broggeri Holub: 7; pl. 1, fig. 8. 1961 Microparia (Microparia) broggeri (Holub) Marek: 45; pl. 3, fig. 16. TYPE LOCALITY. Klabava Formation, Arenig, Bohemia. FIGURED MATERIAL. Axial shields: It.18551, It.18586, It.15933; pygidia: It.19617, It.15934—5; free cheek: It.19626. OTHER MATERIAL. It.18588—99, It.18617, It.15936, NMW 84.17G.45-S2. STRATIGRAPHICAL RANGE. Upper Arenig, Fennian. Widely distributed through the upper part of the Pontyfenni Formation, Bergamia rushtoni Biozone, Llwyn-crwn, Pontyfenni, etc.; north Wales, west side of Nant-y-Gadwen, Llyn. DEscRIPTION. Axial shields are approximately twice as long as wide; on the best preserved dorsal exoskeleton the cranidium accounts for 0-43 of total length, and the thorax 0-25 of total length, with the pygidium 0-32 of total length. In other specimens the cranidium can occupy 0:40 of the dorsal length; the thorax is invariably shorter than the pygidium. All specimens have suffered some degree of flattening; because Microparia spp. do not have much convexity, ARENIG IN SOUTH WALES Fig. 44 a-e, Microparia (Microparia) broeggeri (Holub 1912). Upper Arenig, Fennian. a, axial shield, Pontyfenni Formation, B. rushtoni Biozone, loc. 24, x 3, It.18551; b, cast from dorsal exoskeleton, comparing with Holub’s original, locality as Fig. 44a, fifth thoracic segment partly concealed beneath cephalon, x 3, It.18586; c, cast from axial shield in presumed moult arrangement, loc. 24, x 4, It.15934; d, cast from axial shield of slightly distorted specimen, B. rushtoni Biozone, loc. 23, x 4, It.15933; e, axial shield of largest specimen, with more elongate cranidium, x 3, It.19617; f, M. (Microparia) aff. broeggeri, stratigraphically young specimen with slightly better definition of pygidial axis, Llanfallteg Formation, loc. 52E, Dionide levigena Biozone, x 4, It.19624. 166 R. A. FORTEY & R. M. OWENS except at the front of the cranidium, it is thought that the proportions have not been altered. No doubt the free cheeks hung down almost vertically from the sides of the cranidium as they do in other species. Cranidium with the broadly arcuate outline typical of the genus, and virtually featureless; widest posteriorly; length/width ratios somewhat variable, ranging between 1:3 and 0-75 with a mean of 1:0. Palpebral rims do not extend very far back; posterolaterally on the cranidium there are slightly inflated subtriangular areas presumably representing relict fixed cheeks. Free cheeks consisting of large oval eyes connected by united doublure, not closely approaching mid-line in front, with a small section of border posteriorly. Cephalic surface smooth except for fine lines close to, and parallel to, the margin of the cranidium. One specimen (Fig. 45b) shows the cephalic doublure with the hypostoma in place. Thorax of five segments, parallel-sided or expanding in width slightly backwards; axial taper slight over first three segments, more rapid over posterior two. Half-rings about one-third sagittal length of axial rings. Pleurae bluntly spinose, point of articulation remaining quite close to axis. Dorsal surface of axial rings carry about six raised lines near posterior margins, which are also present on the posterior part of the pleurae. Pygidium generally about two-thirds as long as wide, but highly variable (length/width ratio in range 0-60 to 0-83); outline broadly arcuate, larger examples showing tendency to become triangular (Fig. 45d). Pygidial axis occupies 0:44-0:48 of pygidial width at anterior margin, where it is defined by a prominent rim-like half-ring, tapering backwards quite rapidly, axial furrows enclosing an angle of about 60°. On most examples the axis becomes effaced pos- teriorly, but on the large specimen (Fig. 44a) it is traceable to beyond half pygidial length and almost to its tip. One axial ring is discernible on all but the most completely flattened speci- mens, of similar length (sag.) to the axial rings on the thorax, and defined either by a complete ring furrow or by a pair of depressions representing its outer ends. Faint indications of a second ring on many specimens. Narrow border, present only laterally, present as a rather poorly defined flattened area on most specimens, but can appear as a narrow, gently convex rim on some flattened material. Pygidial doublure widens backwards to a maximum on mid- line of about one-quarter pygidial length, and on exterior surface carries 7-10 terrace ridges, which can appear on dorsal surface of composite moulds. Apart from anterior furrow pleural fields are smooth, or with two or three pairs of obscure ridges indicating segmentation. Discussion. The holotype and only known Bohemian specimen of Microparia broeggeri 1s lost, according to Marek (1961). We have to rely on Holub’s (1912: pl. 1, fig. 8) figure to characterize the species. This does, however, clearly show features which can be matched on the numerous specimens from the Pontyfenni Formation: the pygidium slightly longer than the thorax with a deeply rounded outline; effaced pygidial axis with only the first ring defined by a pair of lateral impressions. Holub’s figure also shows a slight increase in the transverse width of the thorax backwards, which can be matched on the specimen shown in Fig. 45c. Given the otherwise conservative morphology of Microparia these shared similarities are likely to be of specific significance, and Holub’s name is accordingly used for our new material from south Wales. It is like the type species, M. speciosa (e.g Marek 1961: pl. 3, figs 5-10; in Horny & Bastl 1970: pl. 7, fig. 1; Kielan 1960: pl. 10, fig. 6), of Ashgill age, which is distinguished mainly by its relatively transverse pygidium, with a wider axis and relatively effaced pygidial axis. The other Bohemian species revised by Marek (1961) include M. brachycephala (Kloucek 1916), a species in which the maximum cranidial width exceeds the sagittal length in dorsal view, and the pygidial doublure is broader than in M. broeggeri, and M. prantli Marek 1961, which has thorax and pygidium equal in length, and the cranidium is produced into a distinct ‘nose’. The closest species is probably M. plasi Marek 1961, from the Dobrativa Formation (Llandeilo), which is very similar to M. broeggeri in the relative lengths of cephalon, thorax and pygidium, degree of effacement, and width of pygidial doublure. Marek (1961: 44) notes the presence of three faint pairs of cephalic muscle impressions; their absence on the Arenig material from Wales could well be a matter of preservation. The distinguishing character is in the shape of the cranidium, which is widest not at the posterior margin, as in M. broeggeri, but ARENIG IN SOUTH WALES 167 Fig. 45 a—e, Microparia (Microparia) broeggeri (Holub 1912). Upper Arenig, Fennian, B. rushtoni Biozone, Pontyfenni Formation, loc. 24. a, small thorax and pygidium, most transverse specimen, x 6, It.19625; b, fused cheeks and hypostoma, x 3, It.19626; c, small specimen comparing with Holub’s type specimen, x 4, It.15931; d, large pygidium showing tendency to elongate-triangular shape, x 3, It.15934; e, cast of pygidium, loc. 24, x 6, It.15935. f, Microparia (Microparia) boia (Hicks 1875). Middle Arenig, Whitlandian, G. gibbsii Biozone, Pwlluog, north of Whitesand Bay, St David's, Dyfed. Holotype, axial shield with cranidium displaced over thorax, x 4,SM A16731. some distance in front, both palpebral lobes and axial furrows being bowed outwards, rather in the manner of Pricyclopyge. Rushton & Hughes (1981) suggested that M. plasi might prove to be a junior synonym of M. (Heterocyclopyge) shelvensis Whittard 1961, from the Llanvirn of Shropshire, attributing the better definition of the pygidial axis and its rings on the holotype of that species to ‘frontal compression’ in the preservation. In our view this is not probable; the good definition of the pygidial axis is the critical character of Heterocyclopyge (see above), to which shelvensis should be referred. There are three specimens of M. (H.) shelvensis in the collections of the British Museum (Natural History) and the axial furrows are deep on all of them. The form described as M. cf. shelvensis from the Great Paxton borehole has an effaced pygidial axis, and also shows a forward expansion on the cranidium (Rushton & Hughes 1981: pl. 3, fig. 7), which in our view renders it indistinguishable from Microparia (Microparia) plasi, except perhaps for a slightly more rapid taper on the pygidial axis. Of Whittard’s (1940: pl. 5, figs 6-8) original figured material all except the subsequently designated holotype of M. shelvensis are more like M. plasi. Microparia (Microparia) major (Salter 1853) is a well-preserved, unique specimen of a thorax and pygidium from a probable late Llandeilo locality in Anglesey. It is virtually without a pygidial border, and has a very narrow doublure, and hence is quite different from M. (M.) broeggeri; it may prove to be a senior synonym of M. (M.) lusca Marek 1961. As discussed above, the definition of the pygidial axis on M. (M.) broeggeri is somewhat variable, and the species forms a link between Microparia (Microparia) and M. (Heterocyclopyge). The stratigraphically youngest specimen we have recovered is from the 168 R. A. FORTEY & R. M. OWENS Llanfallteg Formation (Fig. 44f) and on this specimen the axis is visible almost to its tip. It is almost intermediate between M. broeggeri and M. shelvensis, although the weak development of the border is like the Arenig species. We refer to it as M. aff. broeggeri. Microparia (Microparia) porrecta sp. nov. (Figs 46a—d, 47) Hovortype. Pygidium, BGS Pr1940. PARATYPES. Cranidium with three thoracic segments It.19630; cranidium It.19631. TYPE LOCALITY AND HORIZON. ‘Roadside 200 yd E by S of Llan, ? mile NE of Llanfallteg railway station’ (Geol. Survey loc. WA9); Llanfallteg Formation, uppermost Arenig part. STRATIGRAPHICAL RANGE. The paratypes are from the earliest Llanvirn part of the Llanfallteg Formation at Cefn-maen-Llwyd farmyard, and from the Upper Arenig (Fennian, Biozone of Bergamia rushtoni) of the type locality of the Pontyfenni Formation, loc. 23, respectively. DiAGNosIs. Microparia with cranidium 1-5 times wider than long, and palpebral rims not extending far back; pygidium more than 2:5 times wider than long, axis not defined, and with broad doublure. NAME. ‘Extended’. DESCRIPTION. No complete specimen of this species is known, but the association of the very Fig. 46 Microparia (Microparia) porrecta sp. nov. a, cranidium and three thoracic segments, Llanfall- teg Formation, early Llanvirn, D. artus Biozone, loc. 50, x 4, It.19630; b, cranidium, preserved with natural relief, Pontyfenni Formation, upper Arenig, Fennian, B. rushtoni Biozone, loc. 23, x 6, It.19631; c, d, pygidium with 2 thoracic segment tips, Llanfallteg Formation, latest Arenig, D. levigena Biozone, Geological Survey locality WA9 (Strahan et al. 1914: 20), holotype BGS Pr1940, dorsal, x 8,and posterior, x 6, views respectively. ARENIG IN SOUTH WALES 169 holotype because it differs from that of any other cyclopygid. The cranidium is wider than any Microparia species but does not differ in structure. Maximum cranidial width is attained some distance in front of the posterior cranidial margin, this being one and a half times the sagittal length, or slightly more. The specimen from the Pontyfenni is preserved in relief, showing very gentle transverse convexity, but a steep downward slope around the edge of the forward cranidial margin. Palpebral rims are comparatively well defined anteriorly, where they are narrow and tucked beneath the frontal part of the median cephalic lobe; backwards they are wider but become fainter, being hardly discernable behind the point of maximum cranidial width. However, the pleuroccipital furrows are faintly indicated at the posterior cranidial margin, indicating that the glabella there occupies about three-fifths the maximum cranidial width. Only the first three thoracic segments are preserved, but we have based the reconstruction (Fig. 47) on the assumption that there were five segments as in all Microparia. The axis is very wide; on the anterior segment the pleurae are only about one-third the axial width (tr.). Otherwise the thoracic structure is as conservative as in other cyclopygids. The first thoracic ring shows well the characteristic sagittal elongation of the first articulating half ring. The transverse pygidium has a gently arcuate posterior outline, and wide, weak border only developed laterally. There is no trace of the axis, but some indication of its extent is given by the inward extension of the narrow (exsag.) facets, which show that it occupied about half the pygidial width. Doublure wide, occupying more than a third of the sagittal length, and slightly wider laterally, carrying 7—8 terrace lines. Discussion. The cranidium from the Pontyfenni Formation is so like that from the Llanfallteg Formation that it is likely the one species ranges from the Fennian into the earliest Llanvirn. It is such a distinctive form that the possibility of its generic separation from Microparia was considered. However, it is best regarded as a Microparia in which transverse extension has reached an extreme development, while it retains the structures of the cranidium and pygidium seen on other members of the genus. It is a homoeomorph of Emmrichops and Gastropolus, but in our view is closely related to neither of these genera. The width of the cranidium and pygidium distinguishes M. porrecta from any other species of Microparia. Fig. 47. Reconstruction of axial shield of Micro- paria (Microparia) porrecta sp. nov., x 3 approx. 170 ; R. A. FORTEY & R. M. OWENS Microparia (Microparia) teretis sp. nov. (Figs 48, 49) Ho.oryPe. Axial shield, BGS Pr1991. PARATYPES. Incomplete dorsal exoskeleton SM A44523; cranidium It.19634; pygidia It.19632— 3, NMW 84.17G.54-55; free cheek It.19706; thorax and pygidium NMW 84.17G.53. TYPE LOCALITY AND HORIZON. Llanfallteg Formation, Rhyd-y-wrdach, near Llanfallteg; earliest Llanvirn. STRATIGRAPHICAL RANGE. Llanfallteg Formation, spanning the Arenig—Llanvirn boundary. Type section, 20m below boundary (52L); earliest Llanvirn of locs 50, 52; and Scolton railway cutting, loc. 55. DIAGNOsIs. Microparia with broadly oval exoskeleton less than twice as long as wide; thorax and pygidium about equal in length, and cranidium not greatly longer. Cranidium broad (tr.) for genus, poorly defined palpebral rims. Pygidium almost twice as wide as long with relatively wide border which 1s distinctly defined laterally. NAME. ‘Smoothed off’. DESCRIPTION. The holotype preserves what is probably the original convexity: low transversely, with a gently swollen cranidium, narrowly but steeply down-turned around its anterior edge. The exoskeleton is less elongate than Microparia broeggeri, width across the mid-part of the thorax more than half sagittal length. Cranidial outline is very broadly and gently rounded about mid-line, cranidial length only about two-thirds width (length/width ratio in the range 0:62-0:75). Maximum cranidial width achieved shortly in front of posterior cranidial margin. Otherwise dorsal cranidial surface is devoid of features; specimens do not show palpebral rims. Fig. 48 Reconstructions of dorsal exoskeleton of Microparia (Microparia) broeggeri (Holub), left, and axial shield of M.(Microparia) teretis sp. nov., right, both x 3 approx. ARENIG IN SOUTH WALES 171 We presume that a gently curved eye belongs here, which has a short postocular border like most Microparia spp. Facial sutures evenly curved except for a short sigmoidal curve at posterior margin. Glabellar tubercle at two-thirds cranidial length. Thorax about two-thirds cephalic length. Broad axis about three-fifths cranidial width at first r taper. Pygidium of about the same length (sag.) as the gentle posterio = me ent, with very segm e Fig. 49 Microparia (Microparia) teretis sp. nov. Llanfallteg Formation. a, holotype, axial shield, early Llanvirn, D. artus Biozone, loc. 50, x 3, BGS Pr1991; b, c, pygidium with one thoracic segment, internal mould and latex cast from counterpart external mould to show sculpture, loc. 50, x 6, It.19632; d, cranidium and thorax, slightly distorted, loc. 55, x 6, SM A44523; e, small pygidium, loc. 52E, latest Arenig, D. levigena Biozone, x 6, It.19633. 172 R. A. FORTEY & R. M. OWENS thorax, and almost twice as wide as long. Axis poorly defined, but rapidly tapering, as in most Microparia spp. Three axial segments are faintly indicated on the holotype. The border is relatively well defined and gently convex anterolaterally, wide, and widening backwards until it becomes obscure postaxially. Doublure also widens backwards to midline, at its widest about one-quarter of pygidial length. Dorsal surface is almost without sculpture; there are a few terrace lines on the thoracic axial rings, and fine, subtransverse terrace ridges on the posterior part of the pygidium. The pygidial doublure carries about eight stronger terrace ridges. Discussion. This Microparia is distinguished from most others, including the type species, by its very broad and gently rounded cranidium lacking distinct palpebral rims. Of the Bohemian species revised by Marek (1961) it is closest to the Llandeilo M. brachycephala (Klouéek 1916), which is itself distinguished from the type species, M. speciosa, by a relatively broad cranidium. Marek’s figures and measurements show that the length/width ratio of the cranidium of M. brachycephala is in only just less than 1, whereas the comparable ratio in M. teretis is less than 0-8. The pygidial doublure of M. brachycephala is wider, and the palpebral rims are better defined. M. laevis Whittard, 196la, has an even wider cranidium than M. teretis, and the pygidial border is narrow and distinctly flattened. M. porrecta sp. nov. (p. 168) also has a wider cranidium, and the pygidium is altogether more transverse. Pricyclopyge wattisoni Hughes 1979, from the Lower Llandeilo of Builth, is very like Microparia brachycephala, and it would be difficult to refer it to any genus other than Microparia unless the six-segment thorax (Hughes 1979: fig. 5) could be proved to belong. The pygidium illustrated by Hughes (1979: fig. 8) is more transverse than that of large M. teretis, and with narrow lateral borders, but it resembles the smaller examples of our species more closely (Fig. 49e). Its axis is much narrower, however, and distinctly defined anteriorly. It is interesting to note some resemblance between M. teretis and the small specimens (not the large holotype) of “M. nudus Whittard, 1961a, here assigned to Degamella. The degree 5 meraspis (Whittard 196la: pl. 24, fig. 9) has a very wide cranidium, like M. teretis. There is of course no question of our large specimens being meraspides. Even the small nudus specimens are distinguished by their wide pygidial doublures and gently tapering pygidial axes. Microparia (Microparia) boia (Hicks 1875) (Fig. 45f) 1875 Aeglina Boia Hicks: 185; pl. 10, figs 9, 9a. 1939b Gallagnostoides boia (Hicks) Kobayashi: 580. 196la Cyclopyge boia (Hicks) Whittard: 179. HooryPe. Crushed and slightly distorted axial shield, SM A16731. TYPE LOCALITY. Whitlandian of the old quarry north of Whitesand Bay, St David’s. STRATIGRAPHICAL RANGE. Only known from the type locality, Whitlandian, biozone of Gym- nostomix gibbsii. OTHER MATERIAL. Axial shield, GSM 8689. DISCUSSION. This species is known from imperfect material from the type locality; attempts to recover it from the mudstone facies of the Whitlandian have not been successful. It is of some interest as the earliest Microparia. The holotype has the cranidium displaced backwards over the front of the thorax. Even so traces of three and possibly four thoracic segments can be seen. Kobayashi (1939b) presumably based his determination of the specimen as representing a new agnostid genus on Hicks’ schematic figure. The species is generally like M. broeggeri (Holub) (p. 164), and Hicks’ name would, of course, have priority over Holub’s if it were conspecific. Several features suggest that this is not the case. There is no trace of the pygidial axial furrows on M. boia. Since the half-ring and anterior pygidial pleural furrows are discernible, one would expect the axial furrows to be preserved also if they had been defined. There appears to be no ARENIG IN SOUTH WALES 173 lateral pygidial border, and the doublure, clearly shown on the holotype, is narrower. Although M. boia is probably distinct from M. broeggeri, we cannot give a satisfactory diagnosis of the species on the basis of the material known at present. Microparia? sp. indet. 1 (Fig. 50) MATERIAL. Incomplete axial shield It.19635, from the type locality of the Pontyfenni Forma- tion, loc. 23. STRATIGRAPHICAL RANGE. Fennian, Upper Arenig, biozone of Bergamia rushtoni. DiscussION. This species is known from only one specimen, but is obviously distinct from Microparia broeggeri (Holub) from the same horizon. The cranidium is wider than long, and the front of the glabella is broadly rounded, almost semicircular, carrying a conspicuous sculpture of incised lines running parallel to the cranidial margin. The axial furrows take an inward bend at about cranidial mid-length, thereby defining a long (exsag.), narrow fixed cheek, but they fade out well before the posterior cranidial margin. The thorax shows five segments of usual Microparia type. The pygidium is incomplete, but what there is shows that the axis was probably not defined dorsally, and that there was a distinct border, at least laterally. The broadly rounded frontal glabellar lobe, and the inward curve of the axial furrows, is very different from M. broeggeri, and indeed other Microparia species. M. teretis also has a wide cranidium, but it is featureless, without sculpture or distinct furrows. Later species of Degamella tend to have axial furrows which take an inward bend, and this is associated with smaller eyes than is usual for cyclopygids. However Microparia sp. indet. 1 is not related to Degamella, because it lacks a cranidial ‘nose’ and has five thoracic segments. It is certainly a new species, and may even be generically distinct from Microparia, but we have too little material to name it formally. Fig. 50 Microparia? sp. indet. 1. Pontyfenni For- mation, Upper Arenig, Fennian, Bergamia rush- toni Biozone, loc. 23, imperfect axial shield, x 5, Tt.19635S. Fig. 51 Microparia (Heterocyclopyge?) sp. indet. Pontyfenni Formation, Fennian, B. rushtoni Biozone, loc. 23, x 6, BGS Pr579. 174 R. A. FORTEY & R. M. OWENS Subgenus HETEROCYCLOPYGE Marek 1961 TYPE SPECIES. Cyclopyge pachycephala Hawle & Corda 1847. DIAGNOSIS. Microparia species with a well-defined pygidial axis. Microparia (Heterocyclopyge?) sp. indet. (Fig. 51) MATERIAL. Pygidium with three displaced thoracic segments, BGS Pr579. STRATIGRAPHICAL RANGE. Pontyfenni Formation, loc. 23; Upper Arenig, Fennian, B. rushtoni Biozone. Discussion. This single specimen cannot be formally named. The pygidium is at once distin- guished from that of Microparia (Microparia) broeggeri (Holub) by its narrow axis and deep axial furrows, except at the tip. It is thus closest to Heterocyclopyge as defined here. However it also differs from Microparia (Heterocyclopyge) pachycephala (Hawle & Corda) and M. (H.) shelvensis Whittard in that the axial rings are not defined, so that even its inclusion in Hetero- cyclopyge is uncertain. The doublure appears to be narrow, and the border ill-defined. See Circulocrania orbissima gen. et sp. nov., pp. 186-7. Genus NOVAKELLA Whittard 1961la (= Incisopyge Pillet & Courtessole 1985: 213) TYPE SPECIES. Aeglina bergeroni Novak, in Novak & Perner 1918, by original designation; authorship discussed in Whittard, 1961a: 170, 1966: 285. Discussion. We recognize Novakella as a genus distinguished from Degamella by its deep, slit-like glabellar furrows, which are visible even on flattened material. Other dorsal furrows are also generally deep, for example axial rings and pleural furrows on the pygidium. On the type species the glabellar furrows are perhaps less incised than they are in N. copei sp. nov. or N. incisa Whittard, but they are quite clearly of the same form (Marek 1961: pl. 4, fig. 11). The genus Incisopyge Pillet & Courtessole 1985, with N. incisa as type species, is diagnosed in the same way as Novakella herein, of which it is considered a subjective synonym. Novakella copei sp. nov. (Figs 52a—d) 1885 Aeglina sp.; Postlethwaite: pl. 3, fig. 15. 1886 Aeglina sp.; Postlethwaite & Goodchild: 463: pl. 8, fig. 15. Hovorype. Axial shield It.15937. PARATYPES. Cranidium It.19637; pygidium It.19638; axial shield GSM 32833. TYPE LOCALITY AND HORIZON. Llanfallteg Formation, earliest Llanvirn of loc. 50. STRATIGRAPHICAL RANGE. From the Llanfallteg Formation, the species occurs in the type section below the Llanvirn boundary, and at the type locality just above the boundary. A specimen from the Lake District recorded as ‘Skiddaw is presumably of Upper Arenig (Fennian) age there. Name. For Dr J. C. W. Cope, who has assisted the authors in countless ways during the research for this paper. DiaGnosis. Novakella with relatively short, bluntly rounded cranidium; pygidium triangular, border ill-defined, with about four pairs of pleural furrows, and rapidly tapering axis. DESCRIPTION. Exoskeleton twice as long as wide, the thorax and pygidium being subequal in length and each about 0-6 that of cranidium. Specimens are flattened, and characters such as the depth of furrows may have been affected by this, but flattening will not have affected the ARENIG IN SOUTH WALES WS) Fig. 52 Novakella copei sp. nov. a, holotype, axial shield, Llanfallteg Formation, loc. 50, early Llanvirn, D. artus Biozone, x 2, It.15937; b, pygidium showing doublure, locality as Fig. 52a, x 2, It.19638; c, latex cast from fragmentary cranidium, Llanfallteg Formation, loc. 52V, latest Arenig, Fennian, Dionide levigena Biozone, x 2, It.19637; d, small poorly preserved axial shield, Skiddaw Slates of ‘Skiddaw’, original of Postlethwaite 1885: pl. 3, fig. 15, x 4, GSM 32833. broadly rounded outline of the front of the cranidium, which is almost semicircular. Maximum cranidial width is at its posterior margin and this just exceeds the sag. length. Axial furrows weakly defined posteriorly, such that the short, triangular postocular cheeks are hardly distin- guishable from the glabella. Three pairs of glabellar furrows of usual form for Novakella, with 2P longest and 3P most forward-inclined. Narrow, gutter-like palpebral rims extend backwards as far as the outer ends of 2P. Free cheeks not known. Six thoracic segments, pleurae widening progressively backwards. Articulation on first segment close to axial furrow and further removed therefrom on posterior segments. Pleural furrows distinctly defined, nearly reaching tips of pleurae. Pleural terminations are distinctly truncate. Pygidium broadly triangular, almost twice as wide as long. At least in the flattened preser- vation the border is not conspicuously flattened, but may have been more so in full relief. Four pairs of rather broad pleural furrows are defined, with indications of a weak fifth pair. Axis sharply conical, axial furrows including an angle of 40°—45°, and extending to two-thirds pygidial length; five (a faint sixth) axial rings are defined, which become progressively shorter (sag.) posteriorly; small terminal piece forms an almost equilateral triangle. Surface apparently lacking any sculpture. 176 R. A. FORTEY & R. M. OWENS Discussion. Novakella is a very rare genus, and the holotype of the new species is one of the best preserved specimens known. N. copei is easily distinguished from the type species, N. bergeroni, from the Llanvirn Sarka Formation of Bohemia and the Hope Shales, Shropshire (Marek 1961: 50-52; pl. 4, figs 10-15; Whittard 1961a: pl. 23, fig. 5), which has an extended glabellar ‘nose’ making the cranidium longer than wide, and a semicircular, rather effaced pygidium. Novakella incisa Whittard, 1961a (:170; pl. 23, fig. 6) is known from a single speci- men from the Hope Shales; compared with the holotype of N. copei the glabella is less broadly rounded anteriorly and clearly defined posteriorly; the 3P glabellar furrows are more strongly forward-directed; the pygidium has a narrow, more gently tapering axis and six pairs of clearly-defined pleural furrows. However, it should be noted that Whittard’s holotype is smaller than ours, and it is likely that the relative depth of the furrows, for example, is a function of size; our smaller cranidium is more like N. incisa in its cephalic furrows (Fig. 52c). The difference in cranidial outline, and the shape and width of the pygidial axis, may be better specific characters. A smaller specimen of N. copei from the Lake District supports this; it is poorly preserved, but it shows that the pygidial axis was like that of copei rather than incisa at this size, and the glabellar front is broadly rounded. Genus PROSPECTATRIX Fortey 1981 TYPE SPECIES. Cyclopyge genatenta Stubblefield 1927, by original designation. DIAGNOsIS. See Fortey 1981: 611. Prospectatrix cf. superciliata (Dean 1973b) (Fig. 53) cf. 1973b Pricyclopyge superciliata Dean: 314-316; pl. 6, figs 2, 4, 6, 8, 9, 14. TYPE LOCALITY AND HORIZON. Sobova Formation, Taurus Mountains, Turkey; late Arenig. OCCURRENCE IN WALES. Pontyfenni Formation, Upper Arenig, Fennian, biozone of Bergamia rushtoni; from type locality of Pontyfenni Formation and Capel-Dewi, loc. 20E. MATERIAL. Cephalon NMW 84.17G.56; free cheek NMW 85.7G.16. Discussion. A single specimen of a cephalon from the Pontyfenni Formation is preserved in partial relief, with the forward part of the median cephalic lobe partly filled by a siliceous nodule; it is the posterior boundary of this nodule which produces the appearance of a trans- glabellar furrow. As its irregularity shows, this is an artefact of preservation. Prospectatrix is the most primitive cyclopygid known, with long (exsag.) relict fixed cheeks. When redescribing the type species from the Tremadoc of Shropshire and Wales, Fortey (1981: 612) suggested that the species superciliata Dean might also be referred to the same genus, which is confirmed by the new specimen from Wales. The course of the axial furrow is shown on the right-hand side of the specimen, the eye having been partly pushed over the cranidial margin. The Welsh Fig. 53 Prospectatrix cf. superciliata (Dean 1973). Cephalon. Apparent transglabellar furrow is an artefact of preservation, as forward part of glabella is preserved within a nodule; course of axial furrow shown on right with fixigena over which the eye has been displaced. Upper Arenig, Fennian, B. rushtoni Biozone, Pontyfenni Formation, loc. 23, x 6, NMW 84.17G.56. ARENIG IN SOUTH WALES 7 specimen is very like the holotype figured by Dean (1973: pl. 6, figs 2, 6, 8) from the Sobova Formation. P. superciliata is distinguished from P. genatenta in having fixed cheeks only about half (tr.) as wide, and in the broadly rounded frontal glabellar lobe. The Welsh specimen could be considered conspecific with the Turkish species on these criteria. A small difference is that our cephalon shows a sculpture of fine lines around the front of the glabella, as in many other cyclopygids; because the Turkish material is exfoliated we cannot say whether the same kind of sculpture was present. Accordingly, we are obliged to qualify our determination of P. supercili- ata. Dean (1973: 343) gives the type locality of P. superciliata an early Arenig age assignment, and if this is correct there is a significant age difference between the Turkish and Welsh occurrences. Genus SAGAVIA Koroleva 1967 TYPE SPECIES. Sagavia felix Koroleva 1967, by original designation. DIscussION. The type species of Sagavia has been reported from the upper Middle Ordovician of Kazakhstan (Koroleva 1967, 1982) and Uzbekistan (Abdullaev 1972); another species, S. elongata, was described by Petrunina (1975) from southern Tien Shan. Sagavia is distinguished from Microparia by its well-defined pygidial axis, and relatively broad and well-defined pygidial border; the cranidium is parallel-sided and subrectangular rather than broadly arcuate in outline. Heterocyclopyge Marek 1961 also has a well-defined pygidial axis, but in that genus the cranidium is Microparia-like and there is no distinct pygidial border. Two well-preserved cyclopygid exoskeletons (lacking free cheeks) from the Arenig of south Wales are immediately distinguishable from Microparia and Degamella species in the Fenni Formation by their wide and well-defined pygidial borders. This species differs from previously described Sagavia species in several details, however: the pygidial axis is not so clearly defined posteriorly, and the cranidium is elongate-arcuate in outline, more like that of Degamella. There are five thoracic segments, like Sagavia, while Degamella has six. With this combination of characters its generic assignment is difficult. Variation in the shape of the cephalic lobe is known within genera in the cyclopygids, as in Pricyclopyge, and it seems reasonable, therefore, to emphasize the broad pygidial border in the generic placing. On this interpretation, the structure of the pygidial axis is then presumably primitive, indicating a common ancestry of Sagavia and Microparia or Degamella. Sagavia glans sp. nov. (Figs 54a—c, 55) Ho.ortyre. Well-preserved exoskeleton lacking free cheeks, BGS Pr574-5. PARATYPES. Dorsal exoskeleton It.19640; thorax and pygidium, It.19641; pygidium, It.19642. TYPE LOCALITY AND HORIZON. Pontyfenni Formation, type locality at Pontyfenni. STRATIGRAPHICAL RANGE. Upper Arenig, Fennian. Name. ‘A bullet or projectile’, referring to the shape of the cranidium. DiaGcnosis. A Sagavia species with cranidial outline more deeply arcuate than other species of the genus; pygidial axis not defined posteriorly; pygidial border widening backwards. DESCRIPTION. Like most of the specimens from the type locality of the Pontyfenni Formation the holotype has escaped much distortion, and other specimens have similar proportions. The elongate-oval outline of the exoskeleton was thus presumably original—the exoskeleton is three times as long as wide. Thorax slightly shorter than pygidium (sag.), which attains nearly three-quarters of cephalic length. Cranidium 1-3 times as long as wide, widest at rear end, whole outline deeply arcuate. Palpebral rims narrow adjacent to anterior two-thirds of cranidium. Relict fixed cheeks form long, narrow (exsag.) triangular areas at base of glabella, poorly defined on inner edges. Gla- bella without muscle insertion areas—or preservation may be inadequate to show them—and 178 R. A. FORTEY & R. M. OWENS a b i ate Fig. 54 Sagavia glans sp. nov. Upper Arenig, Fennian, B. rushtoni Biozone, Pontyfenni Formation, loc. 23. a, holotype, axial shield, x 3, BGS Pr574; b,c, slightly imperfect dorsal exoskeleton, x 4, and enlargement of thorax and pygidium taken from latex mould of counterpart to show sculpture of raised lines, x 7, It.19640. smooth apart from scattered raised lines around front margin running parallel to cranidial outline. Free cheek consisting largely of eye which extends along much of the cranidial margin (Fig. 54b). Thorax with five segments, parallel-sided, but with axis tapering backwards, so that the pleurae correspondingly increase in width; pleurae on the first segment are exceedingly short and nearly triangular, whereas at hind margin of fifth segment the axis only comprises half the thoracic width. As in other cyclopygids the articulating half ring on the first segment is rela- tively long (sag.) compared with the posterior segments. Transverse raised lines run across axial rings. Pygidium % as long as wide, with axis slightly more than one-third transverse width at anterior margin. Axis continues taper of posterior part of thorax (axial furrows enclosing about 40°). On the holotype, only the first ring is clearly defined, the second ring faintly so, and behind this the axis is hard to distinguish from the pleural fields. Internal mould of paratype (Fig. 54b) has better-defined axis (except tip) showing three axial rings. Border distinctly defined, almost flat, comprising one-fifth (sag.) of pygidial length, narrower anteriorly, and widest on midline. Dorsal surface sculpture of raised lines tending to form a network on pleural flanks, not extending over axis, and parallel to pygidial margin on border. Doublure extended beyond border furrow, carrying 8-10 terrace ridges. There are obscure indications of perhaps two pleural furrows. Discussion. The general problems of classifying this species have been discussed above. Sagavia felix Koroleva 1967, S. modica Koroleva 1967, S. elongata Petrunina 1975, S. novakellaformis Koroleva 1982 and S. heterocyclopygeformis Koroleva 1982 all have pygidial axes which are ARENIG IN SOUTH WALES 179 rd Fig. 55 Reconstruction of Sagavia glans sp. nov., x 3 approx. well-defined around their tips, and their pygidial borders are widest posterolaterally rather than postaxially; none appear to have the surface sculpture of raised lines noted on our species. The south Wales species is considerably older than the Russian ones, which come from the Caradoc or Ashgill. Subfamily PRICYCLOPYGINAE nov. DiAGNnosis. Cyclopygidae having six thoracic segments, the third axial ring carrying a pair of hollow bulbs. Cranidium round to ovoid, with maximum width some distance in front of posterior margin. Glabellar furrows feebly developed, up to two pairs of transverse impressions. Pygidium wide, triangular to transversely elliptical, with a deeply defined border. Pygidial axis convex (tr.) and well defined, not rapidly tapering, falling somewhat short of border. GENERA INCLUDED. Pricyclopyge Richter & Richter 1954; Symphysops Raymond 1925; if the association of cephalon and pygidium of Emmrichops Marek, 1961, given by Hughes (1979) is correct this genus also probably belongs here (for further discussion see under Gastropolus, p. 161); Circulocrania n. gen. (p. 186) may also prove to be a pricyclopygine. Discussion. This subfamily is erected to include cyclopygids with the apomorphic character of a pair of hollow nodes or bulbs on the third axial ring of the thorax. The structure and possible function of these is discussed further below. Whatever their function they are unusual structures which are unlikely to have evolved more than once, and they are regarded as good evidence for the common ancestry of cyclopygids having them. The thorax has been described in Pricyclo- pyge and Symphysops, and the bulbs are present on thoraces of all the known species. The pricyclopygines are also characterized by having a cranidium which expands in width forwards at first, so that the cranidial outline is rounded, or elongate-elliptical rather than parabolic as in cyclopygines. Palpebral rims on pricyclopygines are flatter and broader than on cyclopygines. The pygidial borders are characteristically rather flat, and distinctly defined by a deep border furrow which does not get shallower over the mid-line, as it does on many cyclopygines. The 180 R. A. FORTEY & R. M. OWENS thorax is not known for either Emmrichops or Circulocrania n. gen., but a pygidium attributed to the former, and the cranidial outline of the latter, suggest that both may belong in this subfamily rather than Cyclopyginae. Like almost all other cyclopygids, Pricyclopyge and Symphysops show the trend towards anterior fusion of the eyes. It seems unlikely that the specialized organs on the third thoracic segment could have arisen more than once in separate cyclopygid groups: hence Symphysops Raymond and Amicus Koroleva, both of which have these organs, are likely to belong to the same monophyletic group as Pricyclopyge. Symphysops has well-developed glabellar muscle impressions compared with Pricyclopyge, and the glabella produced forwards into a ‘nose’ of varying proportions, and the glabellar tubercle is advanced. The type species of Amicus, A. montanus Koroleva 1967, is closely similar to Symphysops, and it may prove to be a junior synonym of that genus. As discussed by Whittard (1961a), the poorly known Aspidaeglina Holub probably also belongs within this group. THORACIC ORGANS OF Pricyclopyge. The third thoracic segment of Pricyclopyge, as mentioned above, includes on the axis a pair of bulbs or ‘hollow nodes’ (Whittard 1961la: 174). These hollow, inflated structures (Fig. 59a) appear to be formed entirely within the cuticle: internal moulds often show the lower surface of the structure, external moulds the exterior, convex surface, while some specimens show a convex mould of the hollow interior. Ruedemann (1934) made a comparison between these hollow nodes and the luminous organs of certain free- swimming crustacea. Since cyclopygids almost certainly had pelagic habits (Fortey 1985b and p. 105) this comparison is of considerable interest. Much new information is available about the luminous organs of living crustaceans (Herring, 1978, reviews the literature). Many such organs are internal structures and would be expected to leave little trace on the exoskeleton. But some, especially superficial photophores on decapod crustacea, appear to form similar hollow struc- tures (Dennell 1940: fig. 30), sometimes with cuticular lenses on the exterior surfaces. True cuticle does not extend on the inner side of such structures, however, and the largest we can find, on Sergestes (or Sergia) challengeri, are not more than 170 wm long, whereas the structures on Pricyclopyge attain a length of 600 um. Crustaceans often have many such organs, rather than a single pair. In support of the photophore explanation is the suggestion that ‘those species possessing photophores have larger eyes than those lacking these organs’ (Dennell 1940: 376), which obviously applies to Pricyclopyge as compared with the typical trilobite. Herring (1978: 230) notes that ‘cuticular (superficial) photophores are generally common in mesopelagic species, though absent in the bathypelagic realm’. Pricyclopyge is absent from inshore sediments, such as the Mytton Flags, or the Armorican—Iberian region during the Arenig, but is regularly found in deeper facies in Wales. During the Llanvirn transgression, as the deeper facies migrated shelfwards, Pricyclopyge appeared in the Hope Shales, or in the Synclinal d’Ancenis at the southern edge of the Armorican Massif (Henry 1980). It is reasonable to deduce mesopelagic habits from its occurrence. The evidence for a cuticular photophore function for the hollow nodes on Pricyclopyge is incomplete and partly circumstantial, but is consistent with its inferred life habits deduced from other morphological features, and with field occurrence. If correct, it would imply that Pricy- clopyge probably lived below 200m and above 700m in the water column. The majority of such photophores in living crustaceans are ventrally directed; if this were true for Pricyclopyge it would have swum on its back. Genus PRICYCLOPYGE Richter & Richter 1954 TYPE SPECIES. Aeglina prisca Barrande, 1872, by original designation. Discussion. The genus Pricyclopyge includes cyclopygids with six thoracic segments, the third segment carrying the peculiar pair of organs discussed in detail above. The median cephalic lobe is generally subcircular, and the front margin is gently rounded about the mid-line; the pygidium has a triangular outline, with well-defined borders and a relatively long axis. ARENIG IN SOUTH WALES 181 Horbinger & Vanék (1985) distinguished a subgenus Bicyclopyge (type species Aeglina bino- dosa Salter) on the basis of lack of extended thoracic pleural spines on the posterior thoracic segments; we regard this character as significant only at the specific or subspecific level. Pricyclopyge binodosa eurycephala subsp. nov. (Figs 56, 57) 1985a_Pricyclopyge binodosa (early form) Fortey: 23; fig. 5A. Ho orype. Perfectly preserved cephalon, It.15918. PARATYPES. The series includes axial shields: It.15913—4, It.15928, It.18525, It.18531, It.18516, It.18527, NMW 84.17G.57-60, 84.12G.9a, 84.12G.20; cranidia and cephala: It.15924-6, It.15921, [t.18518—-9, It.18548, NMW 84.17G.61-63; pygidia (+thorax): It.15920, It.15927, It.18523; free cheeks: It.15915, It.18539, NMW 84.17G.64-66. TYPE LOCALITY AND HORIZON. Pontyfenni Formation, type locality, loc. 23; Fennian (B. rushtoni Biozone). STRATIGRAPHICAL RANGE. The subspecies is widespread through the Pontyfenni Formation, Upper Arenig, Fennian, biozone of Bergamia rushtoni: locs 20C—E, 21, 23, 25, 29, 32A, 37, 40, 44, 48, 53, 54. Stratigraphically early forms have been found in the S. abyfrons Biozone, loc. 38. Transitional material with P. binodosa binodosa occurs in the Llanfallteg Formation. DiAGnosis. A stratigraphically early subspecies of P. binodosa with nearly circular cranidium in which maximum cranidial width is attained behind cranidial mid-length; entire cephalon with greatest width (tr.) further forward than in P. binodosa binodosa. NAME. From the Greek—‘wide head’. Discussion. A full description of P. binodosa binodosa has been given by Whittard (1961a), and because almost all the details are the same as in the present form, they need not be described here. Rushton & Hughes (1981) have also discussed P. binodosa binodosa and figured some small growth stages. It is a highly conservative form, ranging from the Arenig to the Llandeilo. Two subspecies have previously been distinguished from P. binodosa binodosa: P. binodosa prisca (Barrande 1872) and P. b. longicephala (Kloucek 1916). The former was regarded as a synonym of P. binodosa binodosa (Salter 1859) by Marek (1961), who observed that the preser- vation of the cephalic muscle impressions was highly variable (features used by Whittard, 1961a, to discriminate P. b. prisca); spines on the sixth thoracic pleurae of P. b. prisca were ‘preserved only rarely in several specimens’ (Marek 1961: 32), and Marek regarded their appar- ent absence on all British specimens as a matter of preservation. Rushton & Hughes (1981: pl. 2, fig. 21) figured a beautifully preserved thorax of P. binodosa binodosa on which the non- spinose pleural terminations of the sixth thoracic segment are undeniable. Hence we concur with their view and that of Whittard (1966: 287) that prisca is a valid subspecies characterized by spinose posterior segment(s), while agreeing with Marek (1961) that any supposed cephalic differences between prisca and binodosa are probably the result of preservation—their cephala are identical. It remains a possibility that both subspecies prisca and binodosa are present in the Llanvirn of Bohemia (Horbinger & Vanék 1985). No specimen from the Arenig of south Wales has an extended sixth thoracic segment, and the preservation is adequate to be certain on this point in at least six specimens. The differences between P. binodosa binodosa and P. binodosa eurycephala are on the cephalon only, and of course the same differences apply between P. binodosa prisca and P. b. eurycephala. On the cranidium of the new subspecies the axial furrows swing outwards rather sharply very shortly in front of the posterior cephalic margin. The angle enclosed by these furrows is typically in the range 65°-80° in the population of P. binodosa eurycephala, com- pared with 45°-65° in P. b. binodosa. This results in a more rapid attainment of maximum cranidial width, which was quantified in some detail by Fortey (1985a): the population from loc. 23 shows a normal distribution in this character, clearly distinct from P. binodosa binodosa. R. A. FORTEY & R. M. OWENS Fig. 56 Pricyclopyge binodosa eurycephala subsp. nov. Upper Arenig, Fennian. a, b, cranidium and left cheek in dorsal and anterior views, Pontyfenni Formation, B. rushtoni Biozone, loc. 23, x 4, It.15921; c, small axial shield showing cranidial sculpture, horizon and locality as Fig. 56a, x 3-5, It.15928; d, large flattened cranidium and front part of thorax, the former showing the glabellar tubercle, horizon and locality as Fig. 56a, x 4, It.18519; e, moult arrangement of axial shield, specimen showing exceptional subparallel posterior part of glabella, horizon and locality as Fig. S6a, x 3-5, It.15913; f, g, internal mould of pygidium in full relief, dorsal and posterior views, strati- graphically earliest example, S. abyfrons Biozone, Cwmfelin Boeth Formation, Geological Survey locality Carm. 37SW EA14, lane SW of Pass-by, x 6, BGS TCC959; h, axial shield, horizon and locality as Fig. 56a, x 4, It.15916; j, widest cranidium, horizon and locality as Fig. 56a, x 4, NMW 84.17G.61; k, stratigraphically early cranidium with very posterior maximum width of median cephalic lobe (crinkle cleavage does not distort the proportions), S. abyfrons Biozone, loc. 38, x 4, 1t.19644. ARENIG IN SOUTH WALES 183 Fig. 57 Pricyclopyge binodosa eurycephala subsp. nov. Upper Arenig, Fennian, B. rushtoni Biozone, Pontyfenni Formation, loc. 23. a, b, holotype, perfectly preserved cephalon, dorsal and lateral views, x 6, [t.15918; c, pygidium in full relief, x 3, It.19643. There is a stratigraphical intergradation, but the two forms are distinct between the rushtoni Biozone and the Llanvirn. The maximum cranidial width of the median cephalic lobe in P. binodosa eurycephala is behind cranidial mid-length, as measured along the sagittal line. On P. b. binodosa maximum width is attained at, or in front of, cranidial mid-length. The glabellar tubercle, and the circular depressions to either side of it, are variably preserved, and no taxonomic importance is attached to their absence on certain specimens. Palpebral rims are always clearly defined. One specimen (Fig. 56e) shows the glabella only gently expanding in width at first near its posterior margin but still attaining maximum cranidial width behind cranidial mid-length. On the entire undistorted cephalon (Fig. 57a) maximum cephalic width of P. b. eurycephala is far forwards in comparison with the perfectly preserved cephala of P. b. prisca illustrated by Marek (1961: pl. 1, fig. 20; pl. 2, fig. 1), on which maximum width is near the back end of the cephalon in dorsal orientation. Specimens of this quality are rare, however. These subtle differences are worth taxonomic recognition, because they are of stratigraphical importance: P. b. eurycephala is apparently confined to the Fennian, where Pricyclopyge is one of the commoner fossils. aa ae) * Fp 4 ar Fig. 58 Pricyclopyge binodosa binodosa (Salter 1859), for comparison with P. binodosa eurycephala subsp. nov. a, axial shield, slightly distorted, showing forward expansion of median glabellar lobe, Llanvirn above Llanfallteg Formation, St Clears, Dyfed, x 2, NMW 33.189.G49; b, transition cephalon between P. binodosa binodosa and P. b. eurycephala from Llanfallteg Formation, loc. 50, maximum width of median cephalic lobe at half cranidial length, x 3-5, It.19708; c, undistorted thorax and pygidium, Llanfallteg Formation, D. levigena Biozone, loc. 52K, x 3, It.19709. 184 R. A. FORTEY & R. M. OWENS The change between P. binodosa eurycephala and P. binodosa binodosa is probably entirely gradual (Fortey 1985a and Fig. 58b). In this connection it is interesting to note that P. binodosa longicephala shows the same changes taken further (Marek 1961: pl. 1, fig. 21); on this sub- species cranidial maximum width has become very anterior, while maximum cephalic width is at the posterior cephalic margin. All these changes are related to a change in the course of the axial furrows in front of the thorax, which gradually more and more come to continue (on the cephalon) the forward expansion in width of the thoracic axis, by reducing the sharp outward bend in the course of the axial furrows in the earlier subspecies. Rushton & Hughes (1981) showed that the glabellar expansion of small growth stages of P. binodosa binodosa was gradual, and unlike that of the adult. Hence the change from eurycephala—binodosa/prisca— longicephala seems to be an example of heterochrony. The same glabellar structure is also present on the primitive cyclopygid Prospectatrix (Fortey 1981), and on Pricyclopyge supercili- ata Dean 1973b, from Turkey. Pricyclopyge wattisoni Hughes 1979, which has adaxially effaced pygidial borders, may perhaps be better referred to Microparia. Pricyclopyge ? campestris Koroleva 1967 is represent- ed only by a poorly preserved exoskeleton, but this clearly shows gentle anterior divergence of the axial furrows on the cranidium in the manner of P. binodosa longicephala. P. synophthalma (Koucek 1916) has the eyes confluent anteriorly, which we have not observed on any Welsh specimen; P. obscura Marek 1961 is discussed below; P. sichuanensis Li 1978 has an extended sixth thoracic segment, and may prove to be a synonym of P. bindosa prisca. Pricyclopyge dolabra sp. nov. (Figs 59a—d) Ho ortype. Nearly complete dorsal exoskeleton, It.19662. PARATYPES. Cephalon, BGS JP3525; cranidium, It.19660; pygidium, It.18517. TYPE LOCALITY AND HORIZON. Upper Arenig (Fennian, biozone of Bergamia rushtoni); Ponty- fenni Formation, type locality. STRATIGRAPHICAL RANGE. Fennian (biozone of Bergamia rushtoni) Pontyfenni Formation at type locality and Capel-Dewi (Geological Survey loc. Carm. 40NW WA9). DIAGNosIs. Pricyclopyge with long cranidium, length equal to, or slightly exceeding, maximum transverse width of median cephalic lobe, which is posterior to cranidial mid-length. Glabellar tubercle at mid-length. Palpebral rims relatively narrow. Anterior glabellar tongue broad (tr.). Thoracic axis wide, weakly tapering. Pygidium longer than that of P. binodosa. Name. ‘A grubbing mattock’, which the cranidium somewhat resembles. Discussion. Most of the general features of this species are present on Pricyclopyge binodosa eurycephala, and do not require description. The holotype is a large specimen, 34cm long, preserved in some relief. Specimens of P. binodosa eurycephala are up to 3cm long, and have the same broadly rounded cranidium as smaller examples of the subspecies. P. dolabra has a relatively much longer cranidium, with a broadly truncate outline at the front. There are no intermediates between this species and P. binodosa eurycephala, and it is clearly a distinct species, although rare compared with the binodosa group. Like P. binodosa eurycephala in the Pontyfenni Formation the cranidium expands in width rather rapidly to a maximum behind glabellar mid-length; on the small cranidia this is attained further back than on the holotype. The lateral outline of the cranidium is a rather gentle outward-bowed curve; this is reflected also in the outline of the eye, which is gently curved, and apparently less bulbous than that of P. b. eurycephala. A median cephalic tubercle is preserved on one example at cranidial mid- length. The thorax of the holotype only preserves five segments: the pygidium is a little displaced, and presumably the whole is a moulted exoskeleton. A sixth segment is likely to have been present, particularly in view of the discrepancy in width between the axis of the fifth segment ARENIG IN SOUTH WALES 185 Fig. 59 Pricyclopyge dolabra sp. nov. Upper Arenig, Fennian, B. rushtoni Biozone. a, holotype, exoskeleton with pygidium a little displaced and eyes poorly preserved, loc. 23, x 2, It.19662; b, latex cast from cephalon, Capel-Dewi, x 6, GSM 3525; c, cranidium, locality as holotype, x 6, It.19660; d, cast from mould of pygidium, locality as holotype, x 6, It.18517. and the anterior part of the pygidial axis, but whether or not this segment was macropleural is unknown. The very narrow (tr.) first thoracic pleurae are probably related to the large size of the holotype, and largest P. binodosa eurycephala are little different. However, the posterior taper of the thoracic axis on P. dolabra is much less, so that at the fifth segment the pleurae are only just over half as wide as the axis; on P. binodosa eurycephala on the largest individual the pleurae are two-thirds as wide as the axis here, and relatively even wider on smaller specimens. The pygidium is incompletely preserved on the holotype, but shows a long, tumid axis with one ring defined, and three more indicated along the axial furrow. In view of the relatively wide proportions of the thoracic axis it seems reasonable to assign the pygidium shown in Fig. 59d to the species; it is generally more elongate, convex (tr.) and with narrower pleural regions than the pygidium of P. binodosa eurycephala. Pygidia of the latter are half as long as wide, or less, whereas the pygidium assigned to P. dolabra is two-thirds as long as wide. P. dolabra is clearly distinct from the P. binodosa group on most characters. Marek (1961) described P. obscura from the Llanvirn Sarka Formation of Bohemia, a species in which the length of the cranidium exceeds the maximum width of the median cephalic lobe, and with narrow, gutter-like palpebral rims. P. obscura is known only from the cranidium, although specimens attributed to P. binodosa by Horbinger & Vanék (1985) may belong here. The cranidium has a broadly rounded anterior outline rather than a truncate one as in P. dolabra. Marek states that the maximum width of the median cephalic lobe is at half cranidial length, while it is always posterior to this in P. dolabra; less important is the pair of occipital muscle impressions on P. obscura, because such impressions vary widely with preservation in cyclo- pygids, and particularly Pricyclopyge. 186 R. A. FORTEY & R. M. OWENS Genus CIRCULOCRANIA nov. TYPE SPECIES. Circulocrania orbissima sp. nov. DIAGNOsIs. Cranidium flat, nearly circular. Front margin downturned into a wide (tr.) but short tongue. NAME. ‘Circular head’. Discussion. This very odd cyclopygid was confused at first with flattened Pricyclopyge binodosa eurycephala, which similarly has a broad cranidium. However, it soon became clear that the lack of palpebral rims was an original feature: these are easily prepared on Pricyclopyge. Moreover, the flat surface of the cranidium was also original; whenever Pricyclopyge is flat- tened, it being a genus with a certain amount of transverse relief, the flattening is apparent as strong cracks on the surface (e.g. Fig. 56d). Furthermore, Pricyclopyge has a long anterior tongue which accommodated the dorsoventrally deep eyes. The holotype of C. orbissima (Figs 61a, c, d) is preserved in nearly full relief, with a short tongue, and has almost no other features. A few transverse raised lines are somewhat like those on Pricyclopyge, where they are forward- bowed in the occipital region. There is a median tubercle at about two-thirds cranidial length, with circular impressions to either side. Circulocrania is without question a new cyclopygid, but we hesistated before naming it on the cranidium alone. Dr J. K. Ingham has informed us that a similar kind of circular cranidium occurs in the Whitehouse Formation (Ashgill) of Girvan, so it seems this is another long- ranging cyclopygid morphology. It is considered preferable to name formally what is, after all, a distinctive if featureless cranidium, rather than add it to our long list of undetermined cyclo- pygids. There are three candidates for its post-cephalic parts: cyclopygid thorax and pygidium 1 and 2 (Figs 60a, b), or possibly the pygidium and incomplete thorax described as Microparia (Heterocyclopyge?) sp. indet. (p. 174; Fig. 51). But there is no evidence to assign one or the other to the cranidium. Fig. 60 Undetermined probable cyclopygid thoraces and pygidia, Upper Arenig, Fennian, B. rushtoni Biozone, loc. 23.a, x 3, It.19667; b, x 6, It.19668. (Not described, but see above). ARENIG IN SOUTH WALES 187 Circulocrania orbissima sp. nov. (Figs 61a—e) Ho.oryPe. Cranidium, It.19663. PARATYPES. Cranidia: It.19664-5, NMW 84.17G.167-8. TYPE LOCALITY AND HORIZON. Upper Arenig, Fennian, biozone of Bergamia rushtoni. So far recovered only from Pontyfenni, loc. 23. DIAGNOSIS. See genus. NAME. ‘Most circular’. DESCRIPTION. The cranidium departs slightly from the circular, being a little wider than long. The posterior margin is gently backward-curved, the anterior margin deeply rounded about the midline and downturned. The holotype shows a faint sagittal ridge. The eyes presumably ran around the margin of the cranidium as far as the downturned tongue, but may have continued to be united anteriorly as in other cyclopygids. Fig. 61 Circulocrania orbissima gen. et sp. nov. Upper Arenig, Fennian, B. rushtoni Biozone. a, c, d, holotype, internal mould of cranidium, dorsal, oblique lateral and anterior views, x 3, It.19663; b, cranidium, x 3, It.19664; e, enlargement of latex cast from holotype in oblique light, to show posterior terrace lines and glabellar tubercle with faint circular impressions to either side, x 4, It.19663. Subfamily ELLIPSOTAPHRINAE Kobayashi & Hamada 1971 Fortey (1981) pointed out the difficulty in reconciling ellipsotaphrine morphology with that of other cyclopygaceans, especially with regard to their deeply incised lateral glabellar furrows, and favoured a separate ptychoparioid origin for them, possibly in common with Bohemilla (see p. 129). On the other hand the structure of the rest of the exoskeleton, including the thorax, and the fact that the pygidium is almost identical to those of some species of Cyclopyge, argues for a common ancestry of ellipsotaphrines and other cyclopygids. The problem may be resolved if the ellipsotaphrines are paedomorphic, with the lateral glabellar furrows originating as larval features and finding expression in the adult. Such an explanation would be compatible also with the pygidial structure, which is like that of immature cyclopygids illustrated by Marek (1961). Psilacella pulchra Zhou 1977 has three pairs of glabellar furrows which are comparable with those of Novakella. Ellipsotaphrus itself is more of a problem, as it apparently shows an 188 R. A. FORTEY & R. M. OWENS UPPER TREMAD ARENIG LLANVIRN |LLANDEIL CYCLOPYGINAE PRICYCLOPY GINAE ELLIPSOTAPHRINAE NISIWHO NVNVMGNODS-Id4d NILEIDAE TELEPHINIDAE NIDIWO TWWINOLWVNOA PROETIDA BOHEMILLIDAE --" Robergia zt Girvanopyge NIDIWO v m D | (0) fe) z iS) = > z > z OPIPEUTERIDAE — = =_—_— eadeplinejdoweay Remopleuridiella NIDIYO TIWId¥OLVNOSA Fig. 62 Theory of relationships of Ordovician pelagic trilobites, showing polyphyletic derivation of pelagic morphology. occipital ring, and the absence of a defined occipital ring is a derived character of the Cyclo- pygacea sensu Fortey, 1981. This problem might be resolved if the basal transglabellar furrow on Ellipsotaphrus is not truly occipital, that is, if it is homologous with the basal pair of glabellar furrows on other cyclopygaceans. Support for this is to be found in three species of Ellipsotaphrus, E. infaustus (Barrande 1852), E. pumilio Whittard 1952 and E. zhongguoensis Zhou 1977. On these there are distinct furrows near the posterolateral cranidial margins, which probably represent the pleuroccipital furrows. These are behind the first transglabellar furrow, and on E. zhongguoensis (Zhou 1977: pl. 69, fig. 4) extend adaxially behind it. This may be taken as evidence that the occipital segment is incorporated within the unfurrowed area behind the first transglabellar furrow, which would then, of course, no longer be occipital. Then the three sets of glabellar furrows would be homologous with those of other cyclopygaceans. The transglabellar first two furrows are still a considerable departure from usual cyclopygacean morphology. Similar furrows are developed in other trilobites with probable pelagic habits—in Irvingella, for example. This discussion well illustrates the problems of phylogenetic argument ARENIG IN SOUTH WALES 189 when the majority of the characters available are the result of specialized adaptation. We prefer the interpretation of Ellipsotaphrus as an aberrant cyclopygacean rather than a homoeomorph of the group, but the problem will not be completely solved until primitive ellisotaphrines are discovered. Genus ELLIPSOTAPHRUS Whittard 1952 TyPE SPECIES. Aeglina monophthalma Klouéek 1916, by original designation. Ellipsotaphrus monophthalmus (Klouéek 1916) (Figs 63a-e) 1916 Aeglina monophthalma Klouéek: 13; pl. 1, figs 4-6. 1940 Phylacops monophthalmus (Kloucek) Whittard: 137; pl. 6, figs 1-3. 1952 Ellipsotaphrus monophthalmus (Klouéek) Whittard: 312; pl. 32, figs 10-16. 1954 Cyclopyge (Ellipsotaphrus) monophthalmus (Klouéek) Richter & Richter: 11-12. 1959 Ellipsotaphrus monophthalmus (Klouéek); Richter & Richter, in Harrington et al: 0363. 1961 Ellipsotaphrus monophthalmus (Klouéek); Marek: 60-61; pl. 6, figs 13-17; text-fig. 24. 1961a Ellipsotaphrus monophthalmus (Kloucek); Whittard: 169; pl. 23, figs 3, 4. 1970 Ellipsotaphrus monophthalmus (Kloucéek); Marek, in Horny & Bastl: pl. 7, fig. 6. 1983 Ellipsotaphrus whittardi Horbinger & Vanék: 304; pl. 1, figs 3, 4. TYPE LOCALITY AND HORIZON. From the Dobrotiva Formation (Llandeilo) of Bohemia. OCCURRENCE IN SOUTH WALES. Fennian, Upper Arenig, biozones of Stapeleyella abyfrons to Dionide levigena. Shales interbedded with Cwmfelin Boeth turbidites, loc. 36. Pontyfenni For- mation loc. 23 (type locality), loc. 62, and Bancyfelin. Llanfallteg Formation type section, loc. 52, and early Llanvirn part of Llanfallteg Formation (Whittard 1952: pl. 32, fig. 16). MATERIAL. Complete exoskeleton, It.19669; thorax and pygidium, It.19670; cranidia and cephala It.19671—-2, NMW 84.17G.67-68, 85.1G.47; pygidium BGS HT335. Fig. 63 Ellipsotaphrus monophthalmus (Klouéek 1916). a, entire exoskeleton, Upper Arenig, Fennian, D. levigena Biozone, Llanfallteg Formation, loc. 52R, x 8, It.19669; b, cranidium, Pontyfenni Formation, B. rushtoni Biozone, loc. 23, x 11, It.19671; c, three thoracic segments and pygidium, Cwmfelin Boeth Formation, loc. 36, S. abyfrons Biozone, x 12, It.19670; d, e, cephalon, dorsal and anterior views, loc. 62, x 8, NMW 85.1G.47. 190 R. A. FORTEY & R. M. OWENS Discussion. This species has been fully described by Marek (1961) and further description is unnecessary. The complete specimen from the Llanfallteg Formation (Fig. 63a) is one of the best specimens known. The type material of the species is very much younger than our material, but we can detect no differences between our specimens and a complete Bohemian specimen (Marek 1961: pl. 6, figs 15-17). E. monophthalmus is thus one of the longest-ranging trilobite species, and is further evidence of the conservatism of cyclopygids. One might have anticipated that the early forms would have had eyes separate, rather than fused, in line with other cyclopygids. The eye is not well preserved on the complete specimen, but seems to form a single unit, and a specimen (Fig. 63d, e) from Ramsey Island certainly has a united eye. The oldest specimen is from an early Fennian horizon, and is a partial thorax and pygidium. Without a cranidium it is not possible to be certain of its identity with E. monophthalmus, but the pygidium is so similar to that from the type locality that there is no good reason to qualify the determination. Horbinger & Vanék (1983) have separated Whittard’s (1961) material of E. monophthalmus as a distinct species, E. whittardi, based on the fact that the cranidium of Whittard (196la: pl. 23, fig. 3) has a slight median accumination, and the second transglabellar furrow is straight rather than kinked backwards medially. In fact Whittard’s other specimen (1961a: pl. 23, fig. 4), which they also assign to whittardi, has a median acumination no more developed than many speci- mens of E. monophthalmus (cf. Horny & Bastl 1970: pl. 7, fig. 6) from the type area. Nor is the supposed difference in the transglabellar furrow reliable, because it is much influenced by preservation: crushing or transverse extension both produce a straight, deep furrow. Our specimens from the Pontyfenni Formation have transverse glabellar furrows (and show an acuminate glabellar front) whereas the Llanfallteg complete specimen is in all respects a typical monophthalmus. Hence it seems likely to us that there is a single species in Britain and Bohemia, and we include whittardi in the synonymy of monophthalmus. Genus PSILACELLA Whittard 1952 TYPE SPECIES. Psilacella trirugata Whittard 1952, by original designation. Psilacella doveri (Etheridge 1876) (Figs 64a, b) 1876 Niobe doveri Etheridge, in Ward: 110-111; pl. 12, fig. 2 (note that the names for figs 2 and 3 on the plate are transposed). 1885 Niobe doveri Etheridge; Postlethwaite: 71; pl. 2, fig. 13 (drawing by Goodchild). 1886 Niobe doveri Etheridge; Postlethwaite & Goodchild: 461; pl. 8, fig. 13. Ho.otype. Incomplete dorsal exoskeleton, SM A40455. TYPE LOCALITY AND HORIZON. Etheridge (1876) states that the specimen was recovered from the scree at Randel Crag in the Lake District; specimens from here are from Jackson’s gibberulus Zone and probably equate with the Fennian in south Wales. Discussion. The holotype is a poor flattened exoskeleton. Nonetheless it does show the con- spicuous lateral glabellar furrows which distinguish Psilacella from other cyclopygids. The cranidium is twice as long as Whittard’s specimens of P. trirugata and the third pair of glabellar furrows are not discernible; this may be a specific difference, an ontogenetic difference, or the result of poor preservation. Six thoracic segments are apparently of usual cyclopygid form. The pygidium has two pairs of deep pleural furrows and the pygidial axis has four (possibly five) rings, and extends to the border, which is widest on the midline, and different from that of P. trirugata which has a border which becomes narrower and fainter posteriorly, and a short axis. Psilacella is otherwise known from China (Zhou 1977); P. hunanensis has two pairs of strong pygidial pleural furrows like P. doveri, but on other characters is more like the type species. From south Wales we have one external mould of an incomplete pygidium (Fig. 64b) from the Pontyfenni Formation which is much smaller than that on the holotype of P. doveri and comparable with those of other described species. Like P. pulchra Zhou 1977, both pleural and ARENIG IN SOUTH WALES 191 Fig. 64 a, Psilacella doveri (Etheridge 1876). Holotype, poorly preserved axial shield, Randel Crag, Lake District, late Arenig, ‘gibberulus Zone’, x 2, SM A40455; b, P. cf. doveri, cast from pygidium (plus partial thoracic segment) of small example, Upper Arenig, Fennian, Pontyfenni Formation, loc. 32A, x 10, It.19673. interpleural furrows are developed; the pygidial axis, probably with four rings, falls short of a moderately well defined convex border, shorter (sag.) on the midline than on the holotype of P. doveri. Like the type species there is a well-defined anterior segment. Whether this pygidium is that of P. doveri at a smaller stage of growth is uncertain, although it is certainly like that of Psilacella. Perhaps the difference in the relative lengths of the pygidial axis makes conspecificity unlikely. For this reason we record this pygidium as Psilacella cf. doveri. Family NILEIDAE Angelin 1854 Genus BARRANDIA M‘Coy 1849 TYPE SPECIES. Barrandia cordai M‘Coy 1849, by monotypy. D1aGnosis. See Whittard (19616: 221-222) and Hughes (1979: 154). Barrandia homfrayi Hicks 1875 (Figs 65a—d) (For synonymy see Whittard (19616: 222) and Hughes (1979: 156)). Ho.otyPe. Imperfect dorsal exoskeleton, SM A15627, Llanvirn quarry. OCCURRENCE IN SOUTH WALES. Earliest occurrence in south Wales is in the Llanfallteg Forma- tion, type section, 20m below Arenig—Llanvirn boundary. The species continues and is common into the Llanvirn part of the Llanfallteg Formation, at Rhyd-y-wrach, Scolton and elsewhere, and continues into the dark ‘bifidus’ beds where it is widespread. We have recovered specimens from St Clear’s and from Bifidus Shales exposed by the A 48 just east of Carmarthen. MatERIAL. Figured: axial shield It.19674; cheeks and hypostoma It.19675; pygidia It.19676—7. Other material: It.19678-9, NMW 84.17G.69-75. Discussion. The type species of Barrandia has been fully described by Hughes (1979), and Whittard (1961b) gave an exhaustive account of B. homfrayi itself. Fig. 65a shows what is evidently a degree 6 meraspis. It is relatively large: as in Illaenopsis the final complement of thoracic segments may be achieved at a much later stage in ontogeny 192 R. A. FORTEY & R. M. OWENS eS, Se —s ie ae ost sie ee ay Fig. 65 a—d, Barrandia homfrayi Hicks 1875. Llanfallteg Formation. a, axial shield of large degree 6 meraspis moult arrangement showing pygidial segmentation, early Llanvirn, D. artus Biozone, loc. 50, x 6, It.19674; b, cast from external mould of pygidium to show details of fine surface sculpture, loc. 50, x 6, It.19676; c, fused free cheeks and hypostoma from ventral side, loc. 50, x 6, It.19675; d, small pygidium showing two axial rings, internal mould, loc. 50, x 6, It.19677. e, Barrandia sp. indet., pygidium, upper Arenig, Pontyfenni Formation, B. rushtoni Biozone, Banc-y-felin, compare sculp- ture with Fig. 65b, x 4, GSM HT338. ARENIG IN SOUTH WALES 193 than in other trilobites. Hughes (1979) and Fortey (1975) noted the retention of ‘parathoracic’ segments in comparatively large nileid pygidia. In the specimen of B. homfrayi two segments are retained, and they are ‘thoracic’ even to the extent of showing articulating half-rings, yet they are equally clearly still a part of the pygidium. The retention or release of segments was of less functional importance in nileids than in some other families, and this may account for the variability in the number of segments (7-9) within the group. There may be seven or eight segments within Nileus itself. B. bianularis Whittard, 19616, differs from homfrayi in having two pygidial axial rings, albeit rather weakly defined; other differences between bianularis and homfrayi cited by Whittard, such as a supposedly shorter pygidial axis, cannot be confirmed from measurements on a population of homfrayi. In view of the propensity of Barrandia to retain segments in the pygidium this axial distinction is not convincing, and it is possible that bianularis is merely an intraspecific variant. However, all the larger (pygidium 1 cm long or more) specimens in our collections from the Llanfallteg Formation are like typical homfrayi. Barrandia sp. indet. (Fig. 65e) MATERIAL. Pygidium, BGS HT338-9. OCCURRENCE. Pontyfenni Formation, Bancyfelin Railway cutting, Geol. Survey loc. WA3. STRATIGRAPHICAL RANGE. Upper Arenig, Fennian, probably biozone of Bergamia rushtoni. DIscCuSSION. This single example of a pygidium is undoubtedly distinct from B. homfrayi, which it underlies stratigraphically. The outline of the pygidium is oval rather than sub-semicircular as it is in B. homfrayi, and with a width/length ratio of 1-5 it is more transverse than many specimens of homfrayi. The axis is much less than half pygidial length. The simplest distinction is the surface sculpture: the dorsal surface of homfrayi is covered with terrace ridges of the utmost fineness, which on the Fennian form are much stronger and only about half as dense. The closest comparison among described species is probably with Barrandia cf. cordai from the Llandeilo, figured by Hughes (1979: fig. 126), which also has a strong sculptural pattern, although denser than Barrandia sp. indet. Furthermore, the axis on Hughes’ specimen is much longer (sag.). So the Arenig species is probably a new one, but it cannot be formally named on the basis of one specimen. Genus ILLAENOPSIS Salter 1866a TYPE SPECIES. [llaenopsis thomsoni Salter 1866a, by monotypy. Discussion. Whittard (1961b) redescribed the type species of I/laenopsis, and noted that Salter’s genus was probably the senior synonym of Eurymetopus Postlethwaite & Goodchild 1886. The type species of the latter is E. harrisoni Postlethwaite & Goodchild 1886, and this is the species of which we have much new material from the Pontyfenni Formation. Like many of the species there its eyes are extremely reduced, the palpebral lobes being present as tiny flexures near the front of the cranidium. Whittard (1961b: 220) interpreted J. thomsoni thus: ‘... it may have been a blind trilobite like so many from the Hope Shales, or one in which the eyes had almost atrophied. Examination of Whittard’s material has convinced us that I. thomsoni had minute palpebral lobes like I. harrisoni (e.g. top right on his pl. 31, fig. 5). This being so, there are no significant differences between I/laenopsis and Procephalops Whittard 1967, the type species of which, P. hopense, was also collected from the Hope Shales. In fact, we consider that P. hopense is probably the same species as I. thomsoni, such small differences as there are being because of the smaller size of the former and because of preservational variation. Whittard did not consider Rokycania Pribyl & Vanék 1965, based on Barrandeia (sic) primula Holub 1912 from the Arenig Klabava Formation of Bohemia, when he erected Procephalops. The type of cephalon of R. primula (Horny & Bastl 1970: pl. 5, fig. 9) shows the advanced, tiny eyes, and well-defined anteriorly expanding glabella typical of Illaenopsis; a nearly complete specimen of R. primula as illustrated here (Fig. 66e). Hence we would regard Rokycania also as a synonym of I/laenopsis. Furthermore, Courtessole & Pillet (1976), apparently unaware of 194 R. A. FORTEY & R. M. OWENS both Rokycania and Procephalops, erected another new generic name, Pseudobarrandia, on the same type species, Barrandeia (sic) primula (which they erroneously record as prima) as Rokyca- nia. Thus Pseudobarrandia is an objective synonym of Rokycania, and joins the list of subjective synonyms of I/laenopsis. To summarize, Illaenopsis is probably the senior synonym of no less than four other generic names, a remarkable number for so a rare trilobite: Eurymetopus, Rokycania, Procephalops and Pseudobarrandia. It is also very like two other genera, both from the Tremadoc: Psi- locephalinella Kobayashi 1951 (senior synonym of Borthaspis Stubblefield 1951, with the same type species, Psilocephalus innotatus Salter 1866a) and Borthaspidella Rasetti 1954. Both have small eyes in a forward position; Psilocephalinella is somewhat more effaced than I/laenopsis, and with the palpebral lobes further from the glabella and slightly further back. Borthaspidella has a narrower glabella more abruptly expanding anteriorly, and wider, acute postocular fixed cheeks. These are perhaps rather subtle differences to be of generic significance. Illaenopsis as understood here includes the following species: I. thomsoni Salter 1866a; I. primula (Holub 1912); I. harrisoni (Postlethwaite & Goodchild 1886); I. griffei (Courtessole & Pillet 1976). I. stenorhachis Harrington & Leanza 1957, which was used in the Treatise (Harrington et al. 1959) to illustrate the characteristics of the genus, has a narrow rhachis and wide frontal glabellar lobe, and should perhaps be referred to Borthaspidella, although its convex postocular sutures are like those of Illaenopsis. Illaenopsis harrisoni (Postlethwaite & Goodchild 1886) (Figs 66a—d, 67, 68, 69) 21876 Asaphus sp.; Etheridge, in Ward: 111-112; pl. 12, fig. 1. 1885 sp. C no. 1; Postlethwaite: 77—78; pl. 3, fig. 21. 21886 Unnamed trilobite; Postlethwaite & Goodchild: 456-457; pl. 6, fig. 1. 1886 Eurymetopus cumbrianus Postlethwaite & Goodchild: 459-460; pl. 7, fig. 10. 1886 Eurymetopus harrisoni Postlethwaite & Goodchild: 460-461; pl. 8, fig. 21. 1897 Eurymetopus harrisoni Postlethwaite & Goodchild; Postlethwaite: 13; and cover illustration. 1961b Eurymetopus harrisoni Postlethwaite & Goodchild; Whittard: 219-220. Ho.ortype. Incomplete dorsal exoskeleton, Fitz Park Museum, Keswick, C1. The only original specimen. TYPE LOCALITY. Randel Crag, Lake District; upper part of the Skiddaw Slates. FIGURED MATERIAL. Incomplete exoskeletons: [5331, It.19680—-1; cranidia: It.19685, It.19688, It.19690; hypostoma: It.19684; free cheek: It.19682; pygidia: BGS Pr620, It.19689, It.19683, It.19686-7. ADDITIONAL MATERIAL. NMW 84.12G.32, 84.17G.76—80. OCCURRENCE IN SOUTH WALES. Upper Arenig, Fennian, Pontyfenni Formation at its type locality and at Llwyn-crwn, locs 23, 24, and loc. 20D; B. rushtoni Biozone. DESCRIPTION. Postlethwaite & Goodchild (1886) named two species from two specimens from the Lake District—Eurymetopus cumbrianus and E. harrisoni—which we regard as belonging to a single taxon. Because the holotype of harrisoni is much the more complete and better preserved, we select this name to apply to the species. I. harrisoni grew to a considerable size; it is much the largest Fennian trilobite, fragments from the Lake District and south Wales indicating a maximum length of about 20cms. Changes in proportion occur throughout ontogeny; larger specimens from the Pontyfenni Formation are like the holotype, but we have a suite of specimens spanning a wide size range. Most of the available material is flattened, but a cranidium in full relief from the Pontyfenni Formation (Fig. 68a) shows considerable transverse convexity, with the cheeks sloping down from a convex glabella, and a steep forward slope on the frontal glabellar lobe. Entire exoskele- tons are nearly twice as long as wide; if the small axial shield with imperfect thorax shown on Fig. 67a is undistorted it suggests that smaller specimens were exactly twice as long as wide. ARENIG IN SOUTH WALES 195 d Fig. 66 a-—d, Illaenopsis harrisoni (Postlethwaite & Goodchild 1886). Upper Arenig. a, poorly pre- served, incomplete large axial shield, holotype of Eurymetopus cumbrianus Postlethwaite & Good- child, late Arenig, Randel Crag, Bassenthwaite, Lake District, x 0-5, 15331; b, holotype, imperfect axial shield, Randel Crag, Bassenthwaite, Lake District, probably I. gibberulus Biozone, x 1, Fitz Park Museum, Keswick; c, imperfectly preserved axial shield with minute palpebral lobe at top right, Pontyfenni Formation, B. rushtoni Biozone, loc. 23, x 1, It.19680; d, detail of surface sculpture of fixed cheek, Pontyfenni Formation, B. rushtoni Biozone, loc. 24, about x 10, latex cast taken from It.19690. e, I. primula (Holub 1912), latex cast from a partly articulated axial shield from the Klabava Formation, x 1, Rokycany Museum, Bohemia. 196 R. A. FORTEY & R. M. OWENS Fig. 67 Illaenopsis harrisoni (Postlethwaite & Goodchild 1886). Upper Arenig, Fennian, B. rushtoni Biozone, Pontyfenni Formation. a, axial shield with damaged axial area, loc. 23, x 1-5, It.19681; b, large pygidium showing sculpture, loc. 24, x 1, GSM Pr620; c, free cheek and doublure, loc. 24, note how fine lines on cheek turn into base of ocular region, x 3, It.19682; d, incomplete large thorax and pygidium, showing wide pleurae, loc. 24, x 1, It.19683; e, cast from incomplete hypostoma, loc. 24, x 2, It.19684. ARENIG IN SOUTH WALES 197 The thorax accounts for about 40% of the length, and the length of the cephalon is slightly greater than the length of the pygidium. The cranidium has its maximum width at the posterior margin, this exceeding the sagittal length. On small examples in relief the width/length ratio is about 1-25; large examples are more transverse with the same ratio up to 1:7. The glabella occupies half, or slightly less than half, the width at the occipital margin. Axial furrows are deep posteriorly, parallel on the smallest examples, diverging slightly forwards on cranidia of length about 1-5cm, and more strongly so on larger specimens. Regardless of size, the axial furrows take a sharp outward bend at the level of the palpebral lobes, becoming shallower in the process. They do not quite reach the cranidial margin, but may appear to do so on some flattened specimens. Width of frontal glabellar lobe is 1-4-1-5 times width of glabella at posterior margin on cranidia 1 cm long or larger. No glabellar muscle impressions defined. Small specimen (Fig. 68a) shows a prominent median tubercle at glabellar mid-length. Internal moulds show the typical cyclo- pygacean articulating pits in the axial furrows at the posterior cranidial margin. Minute pal- pebral lobes are about one-tenth sagittal cranidial length, and only gently arcuate-outwards, hence easily overlooked on poorly preserved material. Postocular facial sutures are gently convex-outwards, meeting the posterior margin at a right angle, often rounded. Free cheeks yoked together by doublure, as usual in advanced nileids. They probably hung down at a steep angle to create a broad upward arch about the mid-line, because on the flattened specimen they are twisted outwards relative to the doublure. Genal angle very broadly rounded, the profile of the cheek almost forming a semicircle. The small specimen with the cheek in place (Fig. 68e) shows that at this size there was certainly an eye. We have not found evidence of the visual surface on large specimens, and it is possible that it was lacking (note how the terrace ridges take an inward turn into the ocular area on Fig. 67c, an unusual feature). Doublure widens medially, with about fifteen very strong terrace ridges. Hypostome is typically nileid, subsquare in outline, with the middle body weakly defined, tapering backwards, and maculae at its posterolateral edges. The posteromedian acumination 1s not prominent. Terrace lines are strong, as on doublure, transverse over middle body, but curving parallel to the lateral edges of the borders. The hypostoma is generally less transverse than that of Nileus itself, but very like that of Poronileus (Fortey 1975: pl. 15, fig. 7) except for the less prominent median acumination. Eight thoracic segments of similar length (sag., exsag.) along the thorax, and with gentle axial taper; at third segment width of axis is 1:3—1-4 times the width of the pleura, but on very large specimens pleural width may equal or even exceed axial width. Articulation of usual nileid form: adaxial on first segment, and progessively removed from the axis posteriorly. Pleural furrows deep. Pygidium twice as wide as long, axis at front occupying less than one-third pygidial width and tapering gently posteriorly to rounded tip at three-quarters pygidial length. One axial ring is usually defined across the axis, the secondly feebly so at most, and there are additional indications of three (?four) axial segments as faint depressions alongside the axial furrows. Apart from a prominent half-rib, the pleural fields are unfurrowed on larger pygidia. On relatively small ones, about 1 cm across, one pair of pleural and interpleural furrows are defined (Fig. 69). The largest example is more than 2cm across, and is the largest transitory pygidium of any trilobite! It is relatively common for nileids to retain a single segment defined in the pygidium, but not normally to a diameter of 1cm or more, and so this may be regarded as a neotenic feature associated with the relatively great size of I. harrisoni. A convex border is usually defined laterally, and never extends behind the axis. Doublure with inner outline closely following the posterior pygidial margin, its width at mid length (measured normal to the pygidial margin) being 0-5—0-6 of that of the pygidial axis anteriorly. It also carries terrace lines, but finer and denser than those on the cephalic doublure. An immature pygidium (Fig. 68d) is assigned to J. harrisoni, with a long, narrow axis and a few faint pleural furrows. The whole exoskeletal surface except the thoracic axis is covered with a dense, fine surface sculpture of lines. These are not terrace ridges; they are more like finely incised grooves. On well-preserved surfaces (Fig. 66d) they are themselves interspersed with other lines of the 198 R. A. FORTEY & R. M. OWENS Fig. 68 Jllaenopsis harrisoni (Postlethwaite & Goodchild 1886). Upper Arenig, Fennian, B. rushtoni Biozone, Pontyfenni Formation, loc. 23. a, b, well-preserved smaller cranidium, in relief, dorsal and oblique lateral views, x 6, It.19685; c, latex cast of smaller pygidium, x 4, It.19686; d, presumed transistory pygidium with indications of four segments, x 6, It.19687; e, small cephalon in relief in nodule, showing eye on left (the apparent occipital ring is an artefact produced by the first thoracic segment), x 6, It.19688. utmost fineness. The lines are bowed forwards over the glabella and pygidial axis. On the free cheeks they run almost parallel with the margin except where curving inwards in the ocular region; on the pygidial pleural fields they are more or less transverse. Discussion. This species is very close to I. thomsoni (Salter), as redescribed in Whittard (19615), particularly incorporating Procephalops hopense Whittard, 1967, into that species, as we sug- gested above. The lectotype (Whittard 1961b: pl. 31, fig. 3) shows four clearly defined axial rings on the pygidium, which we cannot match on any specimen of I. harrisoni. We have observed the same feature on other specimens of I. thomsoni (e.g. BGS GSM292) from the Hope Shales, and the well-defined rings seem to be associated with a small amount of transverse extension; in any case they are not visible on other specimens (Whittard 1961b: pl. 31, fig. 4; 1966: pl. 50, fig. 7) which are little distorted. All the articulated specimens of J. thomsoni have wider thoracic pleurae, compared with I. harrisoni of similar size, such that the exoskeleton is more broadly oval, width only about two-thirds the length. Such wide thoracic pleurae can be matched on only the largest specimens of I. harrisoni. The holotype of Procephalops hopense is more elongate, but even here the axis is narrow, and there may have been a certain amount of axial extension on this specimen. The surface sculpture of J. thomsoni and I. harrisoni is similar. The most consistent difference relates to the width of the pygidial doublure, which is wider on I. thomsoni, two-thirds or more the maximum pygidial axial width. The facial sutures behind the palpebral lobes on J. thomsoni diverge outwards at a higher angle than they do on I. harrisoni cranidia in relief, but since flattening would presumably alter this angle little reliance can be ARENIG IN SOUTH WALES 199 Fig. 69 Illaenopsis harrisoni (Postlethwaite & Goodchild 1886). Upper Arenig, Fennian, B. rushtoni Biozone, Pontyfenni Formation, loc. 23. Transistory pygidium, with one unreleased segment, x 3, It.19689. This species has the largest transitory pygidium of any trilobite. placed on it; again, very large I. harrisoni cranidia are little different in this character. In general, specimens of J. thomsoni about 6cm long most closely resemble those specimens of I. harrisoni which would have been perhaps 12cm or more long, and there may be no more difference between the species than a different rate of ontogenetic development. This is sup- ported by the fact that ‘giant’ transitory pygidia of I. harrisoni are known (Fig. 69) at a size when all thomsoni are mature. Undistorted material in relief of J. thomsoni is needed before the status of J. harrisoni can be clarified further. I. primula (Holub 1912) (see Kraft 1972; Horny & Bastl 1970; Fig. 66e herein) from the Arenig of Bohemia is also similar to I. harrisoni, but the fixed cheeks are probably narrower, and the eye larger, such that the palpebral lobe is about one-sixth cephalic length. The articu- lated specimen (Fig. 66e) of IJ. primula apparently shows only seven thoracic segments, but the cranidium is slightly displaced, and there may have been eight as in other IIlaenopsis. In any case this would not constitute a generic difference. I. griffei Courtessole & Pillet 1976, from the Upper Tremadoc of Montagne Noire also has larger palpebral lobes than I[. harrisoni, and the glabella hardly expands outwards at the level of the eyes. Family ILLAENIDAE Hawle & Corda 1847 We here accept the arguments of Lane & Thomas (1983) in not recognizing subfamilial group- ings in the Illaenidae at this stage. Genus ECTILLAENUS Salter 1867 TYPE SPECIES. Illaenus perovalis Murchison 1839, from the Llanvirn of the Shelve inlier, by original designation. DiaGnosis. Illaenid with more or less straight cephalic axial furrows; eye tiny, or absent; postocular facial suture incurved at its posterior end; hypostoma triangulate with anterior margin nearly transverse, or curved gently forwards medially; anterior wings long and rather narrow; thorax of ten segments, axis well defined; pygidium approximately as long (sag.) as cephalon, doublure extending for c. 30%-—50% of pygidial length (sag.) REMARKS. The hypostoma of Ectillaenus (e.g. Snajdr 1957: pl. 2, fig. 10; pl. 3, fig. 4; pl. 6, figs 2, 3; Rabano & Gutierrez-Marco 1984: pl. 3, figs 3, 4) contrasts strongly with that of I/laenus (e.g. Whittington 1965: pl. 52, figs 7-13; Fortey 1980: pl. 10, fig. 11) in being proportionately shorter and wider, with much narrower anterior wings and in lacking an indentation on the anterior margin. In these respects it more closely resembles that of certain styginids—e.g. Bumastus barriensis Murchison 1839 (see Lane & Thomas, in Thomas 1978: pl. 2, figs 5, 11a, b). Ectillaenus perovalis (Murchison 1839) (Figs 70a—1) 1839 Illaenus? perovalis Murchison: 661; pl. 23, figs 7a, b. 1867 Illaenus (Ectillaenus) perovalis Murchison; Salter: 211; pl. 26, figs 5S—8. 1875 Illaenus Hughesii Hicks: 184; pl. 9, fig. 7. 200 R. A. FORTEY & R. M. OWENS 1909 IJllaenus perovalis Murch.; Cantrill, in Strahan et al.: 33. 1914 IJllaenus perovalis Murch.; Thomas, in Strahan et al.: 27, 28. 1961b Ectillaenus perovalis (Murchison) Whittard: 211; pl. 2, figs 6-13; pl. 30, figs 1, 2 (with further synonymy). 1961b Ectillaenus hughesi (Hicks) Whittard: 214; pl. 30, figs 3—7 (with further synonymy). 1984 Ectillaenus perovalis (Murchison); Rabano & Gutiérrez-Marco: 231. 1984 Ectillaenus hughesi (Hicks); Snajdr: 21. LeEcTOTYPE. Selected Whittard (1961b: 213). BGS GSM6844, complete internal mould from fine-grained tuff interbedded with Hope Shales, Hope Mill, Shelve inlier; figured Whittard 1961b: pl. 29, figs 6, 7 (specimen lost—Whittard 1961b: 213). MATERIAL. Relatively common in Llanfallteg Formation at Llanfallteg; from D. levigena Biozone recorded from locs 52C, 52D, 52H, 521, 52N, 52P, 52Q, 52W and 52X, and from D. artus Biozone recorded from loc. 52. Also found in Llanfallteg Formation, D. artus Biozone at loc. 50, Cefn-maen-llWyd (It.18938-9, BGS Pr1996), and loc. 55, Scolton (NMW 84.17G.125), quarry 119m SE of Cefn-farchen, south of Rhyd-y-wrach, Pem. 24SE E/21 (BGS Pg96). Else- where in D. artus Biozone: old quarry at Dan-yr-Allt, 2:7km SW of Bancyfelin, Carm. 45NW E/4 (BGS JP4953), Llanvirn quarry (type locality of E. hughesii) and in Hope Shales Forma- tion, Shelve inlier (see Whittard 1961b: 213, 215). Diacnosis. Ectillaenus which lacks eyes. Cephalon with sculpture of closely-set terrace lines interspersed with rows of small puncta; entire pygidium with sculpture of puncta: small ones on pygidial axis, larger ones on pleural areas and the largest around posterior ends of axis. Pygidial doublure extending for almost 50% of pygidial length. REMARKS. This species was described fully by Whittard (1961b: 211). He considered it to be separate from the contemporaneous E. hughesii, which he claimed is to be distinguished on proportions, on relative abundance and on the restriction of E. perovalis to few localities, with E. hughesii being ubiquitous in the lower part of the Hope Shales. The latter cannot be claimed in any way as biological differences, and there are no morphological features to back them up. In particular, the sculpture, which is clearly distinct in different Ectillaenus species, is identical in perovalis and hughesii. The only one of Whittard’s criteria which could possibly have any validity is that of proportion. He claimed that length—breadth ratios could be used to dis- tinguish the species, stating (1961b: 214), for instance, that in pygidia width : length ratios of 10:9 and 10: 6 are typical for E. perovalis and E. hughesii respectively. However, on foregoing pages (1961b: 212, 213) he noted that in perovalis the uncrushed holotype (sic) has pygidial proportions of 10: 7 whilst in partly crushed specimens they are 10: 9. He thus seems to have placed greater stress on crushed specimens than on uncrushed ones in his arguments. Our specimens give pygidial ratios of between 10: 5-5 and 10: 7-5. All the material is to a greater or lesser degree crushed, but the least compressed specimens give ratios of between 10:6 and 10:7, which seem to represent the true proportional range. There can be no doubt that the wide range in ratios is to be accounted for by extrinsic agencies of crushing and compression. We thus consider that only one species is represented both in Shelve and in south Wales. Whittard (1961b: 214) included the Bohemian species E. benignensis (Novak & Perner 1918) (Llandeilo, Dobrotiva Formation) and, with a mark of interrogation, E. sarkaensis (Novak & Perner 1918) (Llanvirn, Sarka Formation) in synonymy with E. hughesi, thereby extending the stratigraphical and geographical range of the latter considerably. Snajdr (1984) has, however, conclusively demonstrated that E. benignensis is a distinct species, with distinct sculpture and a narrower pygidial doublure (see Snajdr 1984: pl. 1, figs 4-6). E. sarkaensis (Snajdr 1957: pl. 3, figs 1-7; pl. 5, fig. 11) seems also to be distinct, with a much narrower pygidial doublure and more effaced cephalon and pygidium. Whittard (1961b) described two further species, E. cunicularis and E. bergaminus from the Shelve Church Beds and Tankerville Flags respectively. The former was based upon a crani- dium and a distorted pygidium, and he claimed (1961b: 210) that it differed from E. bergaminus (based on well-preserved complete specimens and isolated exoskeletal parts) in having less dense terrace lines which are more regular, a narrower cranidium and shorter pygidium in which the axis is well defined and an equilateral triangle in outline (compared with a biconcave ARENIG IN SOUTH WALES 201 oiins | MMC oi See a [yt ere Fig. 70 Ectillaenus perovalis (Murchison 1839). Llanfallteg Formation, lower Llanvirn, D. artus Biozone (c-f) or upper Arenig, Fennian Stage, D. levigena Biozone (a, b, g-i). a, cephalon with three thoracic segments, loc. 52N, Llanfallteg, x 5, It.18936; b, pygidium showing doublure, loc. 52], Llanfallteg, x 2-5, It.18937; c, axial shield, latex cast from external mould, loc. 50, Cefn-maen-llwyd, x 2, It.18938; d, small axial shield, Cefn-farchen, x 4:5, BGS Pg96; e, complete specimen with cephalon damaged, to show rostral plate, Rhyd-y-wrach, x 1-5, NMW 33.189.G127; f, pygidium showing form of axis and doublure, loc. 50, Cefn-maen-llwyd, x 4, It.18939; g, large pygidium with part of thorax, showing sculpture, loc. 521, Llanfallteg, x 1-5, It.18940; h, laterally compressed axial shield, latex cast from external mould, loc. 52P, Llanfallteg, x 1-5, It.18941; i, complete specimen, preserving original proportions and relief, loc. 52W, Llanfallteg, x 2, It.18942. 202 R. A. FORTEY & R. M. OWENS axis, frequently unenclosed posteriorly in E. bergaminus). The Shelve Church Beds are now known to be of Fennian age, not Lower Arenig as Whittard supposed. Thus these species are more or less contemporaneous. A further topotype almost complete E. bergaminus (NMW 72.5G.1a, b; Fig. 71a) has the pygidium partially exfoliated to expose a triangular, well-defined axis, and the cranidium has a sculpture of terrace lines like those on E. cunicularis. The paratype pygidium of E. cunicularis is a crushed and distorted external mould on which the axis seems to be accentuated. There seems little reason to retain cunicularis and bergaminus as separate species; we recommend that the latter name is used, since it is based upon more complete material. E. bergaminus differs from E. perovalis in having a minute eye (Whittard 1961b: pl. 28, fig. 11) and in having a sculpture of terrace lines with only subordinate puncta on the cranidium and pygidium, the puncta on the latter concentrated on the posterior part. Rare Ectillaenus in the Fennian in south Wales, probably belonging to E. bergaminus, are described below. é Ectillaenus giganteus (Burmeister 1843) from the Llanvirn and Llandeilo of Brittany and the Iberian peninsula broadly resembles E. perovalis, but is immediately distinguished in possessing an eye and a much shorter (sag.) pygidial doublure (cf. Rabano & Gutiérrez-Marco 1984: pl. 1, figs 1, 7 and specimens shown in Fig. 70, p. 201). Ectillaenus ?bergaminus Whittard 1961b (Figs 71b—d) MATERIAL. Badly preserved complete specimen lacking free cheeks NMW 85.7G.1la, b, and cranidium It.18975, from Fennian, B. rushtoni Biozone, Capel-Dewi, locs 20D, E; pygidium It.18976 from same horizon, Pontyfenni, loc. 23. REMARKS. None of the material is sufficiently well preserved to assign it to E. bergaminus with confidence, but such morphological details as can be seen suggest that in all probability it belongs to this species. These include the fine terrace lines on the cranidium and pygidium (those on the latter can only be detected in the external mould NMW 85.7G.1b in very low Fig. 71 a, Ectillaenus bergaminus Whittard 1961b. Axial shield, upper Arenig, Fennian Stage, Tanker- ville Flags, Bergam quarry, Shelve inlier, x 2, NMW 72.5G. 1a. b—d, E. ?bergaminus Whittard 1961b. b, badly preserved axial shield, Fennian Stage, loc. 20E, Capel-Dewi, x 1-5, NMW 85.7G.1la; c, cranidium, loc. 20D, Capel-Dewi, x 5, It.18975; d, pygidium, latex cast from external mould, loc. 23, Pontyfenni, x 6:5, It.18976. ARENIG IN SOUTH WALES 203 angle light) and those on the lateral parts of the thoracic pleurae (cf. Whittard 1961b: pl. 29, fig. 2; Fig. 71b-d). The small pygidium from Pontyfenni has the pygidial axis more or less effaced, and is possibly specifically distinct from the Capel-Dewi specimens. Family TRINUCLEIDAE Hawle & Corda 1847 Comparatively few trinucleids have been reported from the Arenig, and most are from Wales and the Welsh borders, although Hanchungolithus species occur in SE Ireland and southern France (Hughes et al. 1975: 553). In recent years they have also been documented from China (Lu 1975 [several “Llanvirn’ species herein are really Upper Arenig, Glyptograptus sinodentatus minor Biozone age: Zhou Zhiyi, personal communication 1984]; Yin & Lee 1978). All these occurrences are in shelf facies in the peri-Gondwanan area, attesting to the probable origin and early radiation of the family in this region. Most known hitherto in the Welsh area are from the Mytton Flags Formation, Shelve inlier, from which Whittard (1955, 1966) described species of Myttonia, Bergamia, Incaia, Lordshillia and Cochliorrhoe. The last-named has subsequently been synonymized with Bergamia, and the Incaia species transferred to Anebolithus (Hughes et al. 1975: 560). Elsewhere, the old species gibbsii and sedgwicki from Pwlluog, Whitesand Bay were placed by Whittard (1955) in Bergamia (the latter with a mark of interrogation); Whit- tington (1966) described M yttonia fearnsidesi from the Henllan Ash and Fortey & Owens (1978) M. cf. fearnsidesi from the Carmarthen Formation. The present study has revealed more trinucleids from the Whitlandian and Fennian, but no more from the Moridunian. Combining the information from these with that from already known material, it is now possible to reassess the early history of the Trinucleidae. Hughes et al. (1975: 583, fig. 120) gave concurrent ranges in the Arenig for Hanchungolithus, Myttonia, Anebolithus, Lordshillia and Bergamia, with only the latter extending into the upper half of the series (the presence of Hanchungolithus being inferred, for it is known also from the Llanvirn). Much of their information derived from the Mytton Flags species, but we are now able to offer greater stratigraphical precision. Moridunian strata in south and north Wales have yielded only Myttonia (although Han- chungolithus has recently been reported from the Llyn Peninsula: A. Beckly, personal commu- nication 1984). We presume that the lower part of the Mytton Flags Formation with Myttonia and Anebolithus is of (?mid-) Moridunian age (see also p. 98). Whittard (1966: 303) stated, and indicated on his table 2 (1961b: 209) that Myttonia multiplex ranged throughout the Mytton Flags, although in describing this species (1966: 273) he quoted examples only from the lowest Mytton Flags, and we have not seen specimens in substantiation of his higher stratigraphical records. Aside from this somewhat ambiguous information, all other Myttonia records are from the Moridunian, of which stage the genus appears to be characteristic. The earliest Whitlandian is characterized by Furcalithus gen. nov. (p. 207), which in the mid-Whitlandian is joined by early Bergamia species, and the distinctive Gymnostomix gibbsii. The first Stapeleyella occur in the early Fennian. We assume that the higher Mytton Flags (the horizon 2000ft above the base, of Whittard 1966) with Bergamia, Lordshillia and Anebolithus are of early Fennian age. It is thus apparent that there is a succession of trinucleids throughout the Arenig, with three distinct kinds—Myttonia, Anebolithus and Hanchungolithus—already present by the mid- Moridunian. Hughes et al. (1975: 553) proposed that all other trinucleids are likely to be derived ultimately from the Hanchungolithinae which they regarded (1975: 585) as the ‘ances- tral trinucleid group’. We consider this unlikely, although we would admit that all presumably were derived from a single ancestral group in the Tremadoc. The morphology of Hanchungo- lithus, and of other members of the Hanchungolithinae as recognized by Hughes et al., except Myttonia, is broadly similar to that of early marrolithines, in particular Protolloydolithus. Indeed, it is only necessary to add E, and F pits to Hanchungolithus to derive a Protolloydolithus-like form; so we would concur with the Hanchungolithus—Protolloydolithus link they proposed (Hughes et al. 1975: 583, fig. 20). We are less convinced, however, by their contention that Myttonia is a late hanchungolithine. In his numerical taxonomic study of trinucleids, Temple (1980: 227) placed Myttonia cf. fearnsidesi Whittington (of Fortey & Owens 1978) with the Trinucleinae, an association with which we agree, especially since Furcalithus, 204 R. A. FORTEY & R. M. OWENS unknown to Hughes et al., is morphologically and stratigraphically intermediate between Myt- tonia and Bergamia. Hughes et al. placed Myttonia in the Hanchungolithinae on account of such characters as the marginal girder and large number of irregular pits, but on the lower lamella of M. fearnsidesi Whittington (1966: 494) describes ‘a broad, smooth, gently convex area about which the fringe is flexed, .. . suggestive of a weakly developed girder’. A similar feature is present in M. cf. fearnsidesi (e.g. Fortey & Owens 1978: pl. 11, figs 4, 6). If an ‘incipient girder’ such as this were to develop into a true girder, the fringe structure thereby produced would not be very different from that of Furcalithus radix, given further organization of the I series of pits. Despite the presence of a large number of irregular pits in Myttonia, these are fewer and much larger than in Hanchungolithus, but comparable in size and number to early trinucleines such as F. radix. We thus propose to transfer Myttonia to the Trinculeinae, even though it necessitates admitting into the subfamily a genus in which there is only rudimen- tary radial alignment of pits at best (see modified diagnosis below). Hughes et al. (1975: 583, fig. 120) placed Anebolithus between Myttonia and Lordshillia, presumably lending weight to the possession of a marginal girder. Such a placement is to us untenable, since it seems that Anebolithus developed a very simple fringe early on, and is much more difficult to link with Myttonia than are Furcalithus or early Bergamia. It is likely that Anebolithus represents a small, independent Arenig—Llanvirn lineage upon which some Berga- mia (e.g. B. rushtoni sp. nov. (p. 205); B. artemis Rushton & Hughes) with secondary simplifica- tion of the fringe, and reduction in the number of radii, converge morphologically. So far only two species of Anebolithus have been described, A. simplicior (Whittard) and A. sp. of Hughes (1971), the second represented by but one specimen. Whittard (1966) recorded A. simplicior rom horizons 600ft and 2000 ft above the base of the Mytton Flags. From the lower horizon there appears only to be a single specimen (Whittard 1955: pl. 3, fig. 7), most of the fringe of which is not preserved, and it may not be conspecific with A. simplicior. In fringe, genal, sculptural and pygidial characters the Whitlandian Gymnostomix gen. nov. (p. 215) closely resembles Anebolithus simplicior, and it is likely that the genera are closely related. Furcalithus is apparently restricted to the lower part of the Whitlandian. The pits are essen- tially radially aligned, with two E arcs and three or four I arcs, with some remnant irregularity apparent in the latter. It appears as if a Whitlandian radiation from Furcalithus gave rise on the one hand to Bergamia, through reduction in number of I arcs and development of deep radial sulci and increased organization of the posterolateral I pits, and on the other to Stapeleyella with its characteristic interradii on the E arcs. The earliest Stapeleyella, S. abyfrons, has a highly irregular arrangement of adventitious pits, giving rise to an incipient ‘Y’ arrangement of inter- radial ridges. Meraspides of S. inconstans figured by Whittard (1955: pl. 4, figs 11, 12) have a fringe with a large number of long, narrow radial sulci very like that mature Furcalithus? sp. (Fig. 76) from the Whitlandian of Ll¥n (although the latter lack inter-radii in the E series), adding further weight for an origin of Stapeleyella in the Bergamia—Furcalithus complex. We agree with Hughes et al. (1975: 560) that Lordshillia, presumably derived from Bergamia in the late Whitlandian—early Fennian, almost certainly gave rise to Trinucleus. Subfamily TRINUCLEINAE Hawle & Corda 1847 DIAGNOSIS. Fringe with no F pits, other than posterior fossula. Pits essentially radially aligned except in Myttonia, which has incipient alignment only. Glabella with prominent pseudofrontal lobe with three pairs of lateral glabellar furrows. Pits on upper lamella generally regularly arranged. Usually no occipital spine. REMARKS. The above diagnosis is essentially that of Hughes et al. (1975: 555), modified to include Myttonia (see also discussion above). Genus BERGAMIA Whittard 1955 TYPE SPECIES. Bergamia rhodesi Whittard 1955, Fennian Stage, Tankerville Flags; Shelve inlier, Shropshire; by original designation. D1AGNosIs. See Hughes et al. (1975: 558). ARENIG IN SOUTH WALES 205 Bergamia rushtoni sp. nov. (Figs 72a—k) 1906 Trinucleus sp.; Evans: 612 (list). HoLotyPe. NMW 21.306.G4a, b, dorsal exoskeleton with associated, displaced lower lamella, from Fennian, Pontyfenni Formation; loc. 23, Pontyfenni. PARATYPES. Numerous specimens from the type locality in BM(NH), NMW and BGS from the type locality; also from type horizon, loc. 20D, Capel-Dewi; BGS TCC759, a badly preserved cranidium from Rushmoor near Bancyfelin (Geol. Survey loc. Carm. 38SW EA5) probably also belongs to this species. STRATIGRAPHICAL RANGE. Fennian, B. rushtoni Biozone. DIAGNosIs. Typically 15 or 16, rarely up to 18 radii per half-arc, in deep radial sulci in adult specimens, but sulci of E and I series separated in some immature ones; radii widely spaced in front of glabella, interradial areas considerably wider than radii, the latter becoming more crowded posterolaterally; E, and I, complete; E, close to E,, and commences normally at R2 or R3; I, begins at R6 or R7; auxiliary pits commonly absent, but may rarely be developed in one or two interradii; coarse reticulation on glabella and genal lobe. Name. For Dr A. W. A. Rushton, for the continued help he has given in this project. DESCRIPTION. Fringe rather narrow and of constant width. Typically 15 or 16, and rarely up to 18 rows of pits per half-arc, in deep radial sulci in adult specimens (e.g. Fig. 72e), but E and I series separated in immature specimens (e.g. Fig. 72)). Radii widely separated in front of gla- bella, but become more crowded laterally; thus the interradial areas are about twice as wide as the sulci frontally, but are only as wide laterally. Most specimens have 15 or 16 radii per half-arc, in rare cases up to 18. E,; complete, E, normally commencing at R2 or R3, but may be delayed until R9. Auxiliary pits rarely developed, and their position is random, for example €, > has been observed in iii and xii. I, complete, I, developed from RS-R7 up to R15. A typical fringe formula is: E, 0-16, E, 3-15, I, 0-16, I, 7-15. Glabella with coarse reticulate sculpture and with swollen pseudofrontal lobe which over- hangs the fringe sagittally to obscure its inner portion in dorsal view. Three pairs of lateral glabellar furrows of which 1P and 2P are both deep and of similar depth; 3P weak and small, forming small indentation on the flank of the pseudofrontal lobe. Axial furrow narrow and deep at anterior end of pseudofrontal lobe, but backwards broadens where the posterior part of the glabella narrows, whilst the inner flank of the cheek runs more or less exsagittally. Occipital ring narrow and upturned, differentiated from the glabella by a broad, shallow occipital furrow that deepens into small pit close behind the 1P furrow. Genal lobe with a coarse reticulate sculpture, coarsest on the inner posterior part, finer towards the leading edges. There is a suggestion of a faint, backwardly oblique eye ridge on some specimens (e.g. Fig. 72b), a character typically absent in Bergamia. Pleuroccipital furrow broad and deep, and behind it is a very narrow, upturned posterior cephalic border. Genal spine long and grooved, extending considerably further backwards than the pygidium. Thorax of six segments, in general structure like those of other Bergamia species. The first four maintain an almost constant width (tr.), the fifth and sixth being progressively narrower (tr.). Pygidium with three or four narrow axial rings, the axis extending backwards almost to the posterior margin where it impinges upon the downturned pygidial border. Pleural area with two weakly-defined ribs that extend to the well-marked inner edge of the border. REMARKS. The type species, B. rhodesi Whittard (1955: 32; pl. 3, figs 8-13) differs from B. rushtoni in having only auxiliary pits (e,_,) in RO, with E, and E, commencing at R1, a greater number of radii (18-19) and I, developed only in R1-—RS, and a decidedly finer sculpture. B. inquilina (Whittard 1966: 278; pl. 48, figs S—10) and B. matura (Whittard 1966: 280; pl. 48, figs 11, 12) (possibly conspecific, see Hughes et al. 1975: 558) are both similar to B. rushtoni in 206 R. A. FORTEY & R. M. OWENS Fig. 72 Bergamia rushtoni sp. nov. Upper Arenig, Fennian, B. rushtoni Biozone; all except i from Pontyfenni Formation, loc. 23, Pontyfenni. a, small entire specimen, latex cast, x 5, It.18943; b, axial shield with well-developed radial sulci, and traces of ocular ridges, x 3-5, NMW 84.11G.1a; c, axial shield with radial sulci only developed laterally, x 4-5, It.18944; d, axial shield with some anterolateral pits of E and I series not in sulci, x 5, It.18945; e, holotype, axial shield with associated displaced lower lamella, x 3-5, NMW 21.306.G4a; f, large axial shield with well-developed radial sulci frontally, x 4, NMW 84.11G.2a; g, incomplete lower lamella, x 6, It.18946; h, ditto, x 7, It.18947; i, ditto, loc. 24, Llwyn-crwn, x 8, It.18948; j, small entire specimen, latex cast, x 4, NMW 77.9G.35a; k, disarranged axial shield with some pits of E and I series separated anterolaterally, x 45, NMW 84.11G.3a. number of radii and distribution of pits in E,, I, and I,, but there are small differences in E,; in B. inquilina its development is irregular, and it commences at R3, RS, R7 or R8, extending to R14 or R15 (Whittard 1966: 279), and in B. matura ranges from RO or R1 to R17 (Whittard 1966: 280). Both B. inquilina and B. matura have a comparatively large number of auxiliary pits ARENIG IN SOUTH WALES 207 in e, and e,, and it is this feature more than any other that distinguishes them from B. rushtoni, in which auxiliary pits are but rarely developed (see above). It is likely that B. rushtoni has evolved from B. inquilina, B. matura or a similar form by the total or almost total loss of auxiliary pits, and a similar reduction has taken place in B. rhodesi. B. inquilina and B. matura are of late Whitlandian or earliest Fennian age, and thus occupy an appropriate stratigraphical position to have given rise to B. rushtoni in the Pontyfenni Formation and B. rhodesi in the Tankerville Flags. A further stage in the simplification of the fringe is seen in the Llanvirn species B. artemis Rushton & Hughes (1981: 630; pl. 2, figs 1-14, 16, 17) from the Great Paxton Borehole; here E, and E, (with 12-15 pits per half-arc) converge posterolaterally to form twin pits at about R10, and become a single arc in the last three radii. I, is the only inner arc developed. The dorsal exoskeleton of B. rushtoni shows a striking similarity to Anebolithus simplicior (Whittard) (Hughes et al. 1975: pl. 1, figs 17, 19), particularly in its cephalic sculpture and in the presence of deep radial sulci on the fringe. Without a lower lamella it might be easy to confuse the two, but they can be readily distinguished by the spacing of the sulci in front of the glabella. The interradial areas are narrower than the sulci in A. simplicior (Hughes et al. 1975: pl. 1, fig. 17), but much wider in B. rushtoni (e.g. Fig. 72e). Bergamia sp. A (Fig. 73) MATERIAL. Incomplete external mould of cranidium, NMW 85.9G.1, from Whitlandian, Blaen- cediw Formation, Blaencediw Quarry (loc. 29). DESCRIPTION. The pseudofrontal lobe appears to be rounded and somewhat swollen; traces of a reticulate sculpture can be seen on the genal lobes. A half-arc has 24 radii in deep, rather narrow radial sulci; three pits can be seen in one of these, but the position of the girder is unknown. REMARKS. The disposition of pits in deep radial sulci suggests that this specimen is a Bergamia. A second similar fragmentary cephalon has been recovered from the same horizon at loc. 39, and specimens that are possibly conspecific have been found low in the section at Pwlluog (loc. 61A). The latter include a very badly preserved, cleaved cranidium which shows distinct radial sulci like the present specimen, and which clearly differentiate it from G. gibbsii and F. sedg- wicki that occur higher in the section at Pwlluog. , eee bi s coat ; cee Sa t jy Oe gaig ght : 5g ae: : ESSA Ms Fig. 73 Bergamia sp. A. Middle Arenig, Whit- landian, ?F. radix Biozone, Blaencediw Forma- tion, loc. 31, Blaencediw. Incomplete cranidium, latex cast showing radial sulci and traces of reticulation on genal lobe, x 5, NMW 85.9G.1. Genus FURCALITHUS nov. TYPE SPECIES. Furcalithus radix sp. nov. Name. Latin furca, ‘a pitchfork’, which some radii with closely associated interradii in type species resemble, plus suffix -lithus. DIAGNosis. Trinucleine with radial sulci only very weak, if present; E,, E, complete; I,, I, complete, I, complete or extending round most of fringe, I; may also be present; 4—?8 inter- radii present in I series. Pygidium with axis reaching, but not impinging on, narrow posterior border. 208 R. A. FORTEY & R. M. OWENS REMARKS. The general relationships between Furcalithus and other contemporaneous tri- nucleines are discussed above. Main differences from Myttonia are: presence of girder and radially aligned pits around entire fringe (apart from proterolateral corners); from contempora- neous Bergamia species, Furcalithus differs in the lack of deep sulci, presence of more I arcs, presence of interradii in the I series and a narrower pygidial border onto which the axis does not extend. However, the Llandeilo Bergamia species B. prima and B. whittardi (Hughes 1971: 140 and 146 respectively) have more pits in the I series than do earlier ones, and in this respect resemble Furcalithus. However, interradii are absent and the pygidium has the typically broad border of Bergamia; the similarity of the fringe is presumably the result of convergence. Furcalithus radix sp. nov. (Figs 74a—e, 75) Hovorype. [t.18949, dorsal exoskeleton lacking lower lamella and genal spines. TYPE LOCALITY AND HORIZON. Whitlandian Stage, Furcalithus radix Biozone; about 70m above base of Afon Ffinnant Formation, Cwm yr Abbey, loc. 16K. PARATYPES. It.18950, NMW 84.10G.9a, b—14a, b, all from basal 100m of Afon Ffinnant Forma- tion, Cwm yr Abbey, loc. 16L. STRATIGRAPHICAL RANGE. Whitlandian, F. radix Biozone, basal 100m of Afon Ffinnant Formation. Fig. 74 Furcalithus radix gen. et sp. nov. Middle Arenig, Whitlandian Stage, F. radix Biozone, Afon Ffinnant Formation. a, holotype, axial shield, loc. 16K, Cwm yr Abbey, x 5, It.18949; b, incomplete axial shield, loc. 16L, Cwm yr Abbey, x 6, NMW 84.10G.11a; c, partial lower lamella, loc. 18E, Afon Ffinnant, x 5, It.18950; d, lower lamella with genal spine, locality as Fig. 74b, x 4, NUW 84.10G.12a; e, cranidium with five thoracic segments, locality as Fig. 74b, x 5, NMW 84.10G.13b. ARENIG IN SOUTH WALES 209 Fig. 75 Half-fringe of Furcalithus radix gen. et sp. nov., showing distribution of pits, based on Fig. 74e. DiAGNnosis. Radii per half-arc 20-22, E,, E, complete, with paired pits, close together; I, I,_; complete; four or five interradii in I series, typically at or near radii 4, 6, 14 and 17; weak flexure on genal lobe. Name. ‘A root’; the earliest member of the Furcalithus—Bergamia-Stapeleyella stock. DESCRIPTION. Glabella with deep 1P, 2P and 3P furrows. Glabellar node about half-way along glabella. Axial furrows broad and shallow, with anterior fossula represented by a weak linear feature. A short distance abaxially from 1P furrow, and slightly posterior to it, there is a small, shallow pit. There is a second similar pit a short distance behind the 1P furrow (well seen on Fig. 74b); this we interpret as the occipital pit, although some trinucleines possess two small pits in addition to the occipital pit close to the base of the glabella (Hughes et al. 1975: 541). Occipital furrow broad and shallow; occipital ring a narrow, backwardly-curving band, without spine. At posterior end of axial furrow is a weak, elongate alar lobe. Genal lobe traversed by a weak flexure which near the genal angle is expressed as a narrow ridge which runs onto the posterior cephalic border. A small posterior fossula is situated just adaxial to this ridge in the posterior border furrow. The flexure has the appearance of being accentuated in front of the glabella, but this seems in large part to be a result of compression. The genal lobes and glabella have a fine reticulate sculpture, that of the leading edges of the former being somewhat finer. The axial furrows are devoid of such sculpture. Fringe strongly flexed at the girder, the position of which on the dorsal surface is indicated by a smooth band (see especially Fig. 74e); this is not a raised ridge, or ‘list’ (cf. Hughes et al. 1975: 550). Twenty to twenty-two radii per half-arc, not in sulci, although these appear to be present on internal moulds (Fig. 74a). E, and E, complete, close together. E, pits seen in radii 9, 16 and 17 on one specimen (Fig. 74e). I, and I, _, extend at least as far as radius 17 or 18, posterior to which the arrangement becomes more or less chaotic. On the best-preserved fringe (Fig. 74e) interradii 4, 6, 14 and 17 are well seen, but position of the interradii does not appear to be consistent, although there is insufficient material to be able to analyse this statistically. Certain interradii converge in an outwards direction with adjacent radii to produce a ‘fork’. Thorax of six segments, of the normal trinucleine structure. Pygidium with narrow border, weakly reflexed on either side of the axis, which just impinges onto it. Six or seven axial rings, which become progressively more weakly defined backwards. Pleural fields with five pairs of ribs, more clearly identified on external rather than on internal moulds. Furcalithus sedgwicki (Salter 1866b) (Figs 76a—d) 1866b Trinucleus Sedgwicki Salter: 319; pl. 12, fig. 9. 1881 Trinucleus Sedgwicki Salter; Salter & Etheridge: 516; pl. 12, fig. 9 (copy of Salter, 1866). non 1906 Trinucleus sedgwickii Salt.; Evans: 609 (list) [= Stapeleyella abyfrons]. non 1914 Trinucleus sedgwicki Salt.; Thomas, in Strahan et al.: 14 [= Myttonia cf. fearnsidesi]. 1955 Bergamia? sedgwicki (Salter) Whittard: 33; pl. 4, figs 3, 4. 1960 ?Bergamia sedgwicki (Salter); Whittard: 182 (list). 1971 Bergamia? sedgwicki (Salter); Hughes: 145. 1975 Bergamia? sedgwicki (Salter); Hughes et al.: 558. 1980 Trinucleus sedgwicki Salter; Temple: 224. 210 R. A. FORTEY & R. M. OWENS LECTOTYPE (here selected). GSM 49670, cephalon. TYPE LOCALITY AND HORIZON. Whitlandian Stage, Gymnostomix gibbsii Biozone; south side of St David’s Head (presumably Pwlluog, vicinity of locs 61D and F), Dyfed. MATERIAL. Cranidia and cephala from type stratum, Pwlluog, locs 61C—F. STRATIGRAPHICAL RANGE. Whitlandian, G. gibbsii Biozone. DIAGNosIs. Approximate distribution of pits as follows: 17-19 radii per half-arc, E,, E, com- plete; I,,, I, 0-16 or 17, I, (?)6—16; interradii in, il? ii—vi, xii—xv; i, at least xii, xiii, xiv. Glabella and genal lobes apparently smooth. DESCRIPTION. The small amount of material available of this species is all indifferently preserved and variously distorted. The proportions of the glabellar and genal lobes are like F. radix. The fringe differs in that the individual pits are larger, there are fewer radii, a greater number of interradii, and no I, in the I series; there are fewer irregular pits on the posterolateral corners. REMARKS. Whittard (1955: 34) remarked on the difference in the I series of pits between F. sedgwicki and the type species of Bergamia, B. rhodesi, and only placed sedgwicki in Bergamia with considerable doubt. These same characters of the I series are very similar to those of F. radix, and are noted above. Whittard (1955: 31) suggested that trinucleids from the Whitlandian of Dwyrhos, Aberdaron and Nant-y-Gadwen, Llyn Peninsula, may belong to Myttonia, but noted that the external pits in the upper lamella appear to occur in sulci. We have obtained and examined further material Fig. 76 Furcalithus sedgwicki (Salter 1866b). Middle Arenig, Whitlandian Stage, G. gibbsii Biozone, Penmaen Dewi Formation, Pwlluog, Whitesand Bay. a, cephalon, latex cast from external mould, x 3, In.38217; b, cranidium, x 4, BGS GSM49672 (original of Whittard 1955: pl. 4, fig. 4); c, cephalon, latex cast from external mould, loc. 61F, x 4, It.18953; d, lectotype cephalon, latex cast from external mould, x 3, BGS GSM49670 (original of Whittard 1955: pl. 4, fig. 3). ARENIG IN SOUTH WALES PAI Fig. 77 Furcalithus? sp. Middle Arenig, Whilandian Stage, Llyn Peninsula, Gwynedd; Dwyrhos, Aberdaron (a, b, e) or east side of Nant-y-Gadwen, Rhiw (c, d). a, small cranidium with associated thoracic segments, x 10, NMW 27.110.G276; b, complete specimen, cephalon damaged (note broad genal spine), x 4, 114282; c, incomplete lower lamella, x 4, It.18951; d, incomplete cranidium, latex cast from external mould, x 8, It.18952; e, cephalon, showing long, broad genal spines, x 4, NMUW 27.110.G275. (Figs 77a—e) from these localities. The general aspect of the fringe is not unlike F. sedgwicki, but we can confirm the presence of distinct radial sulci; the poor preservation of the available specimens means that it is not possible to be certain whether or not there are interradii in the I series. There are arguments for placing this material either in Bergamia (distinct radial sulci) or in Furcalithus (the probability of a large number of I arcs); because it bears closest resemblance to F. sedgwicki, we place it with question in Furcalithus, but definite assignment must await better material. Genus STAPELEYELLA Whittard 1955 TYPE SPECIES. Stapeleyella inconstans Whittard, 1955; from Lower Llanvirn, D. artus Biozone, topmost Hope Shales Formation, Shelve inlier: by original designation. DIAGNos!s. E, , and sometimes E; present, and all generally complete; some pits of E, may be present. I,, I, complete, very close together frontally. Pits on upper lamella in radial sulci, many interradial ridges bifurcating external to E, to produce intercalated short sulci with E, or E, and E, and frontally E,, typically forming a series of ‘Y’s, but arrangement may be highly irregular. Pygidium like Bergamia. REMARKS. We have modified the diagnosis of Stapeleyella given by Hughes et al. (1975: 559) in order to accommodate S. abyfrons sp. nov. (see below). This extends the range of the genus at least as far back as the early Fennian. Its possible relationships with Furcalithus and Bergamia are discussed above. 212 R. A. FORTEY & R. M. OWENS Stapeleyella inconstans Whittard 1955 (Figs 78a-f) 1955 Stapeleyella inconstans Whittard: 36; pl. 4, figs 7-13; pl. 5, figs 1-6 (with earlier synonymy). 1975 Stapeleyella inconstans Whittard; Hughes et al.: 559; pl. 2, figs 29-31. 1980 Stapeleyella inconstans Whittard; Temple: 221. LECTOTYPE (selected Temple 1980: 221). GSM 92971, complete dorsal exoskeleton on slab with two individuals (figd Whittard 1955: pl. 4, fig. 7 (left hand specimen); Hughes et al. 1975: pl. 2, fig. 29 (upper specimen)). TYPE LOCALITY AND HORIZON. Topmost Hope Shales Formation, path W of Brithdir, 1-6km ENE of Old Church Stoke, Powys (Shelve inlier). OCCURRENCE IN SOUTH WALES. Arenig, highest Fennian Stage, Llanfallteg Formation, Llanfall- teg (locs 52A—C, N, R, T—X). Llanvirn, artus Biozone, Llanfallteg (loc. 52), Llanilwyd, St Clears, Cefn-maen-llWyd, Rhyd-y-wrach (loc. 50), Llandissilio (loc. 53), Long Plantation railway cutting, Scolton (loc. 55), railway cutting WNW of Clarbeston Road station. DIAGNosIs. Stapeleyella with E,-E, normally complete, but E, sometimes absent from more anterior radii. E, may be developed laterally. Interradii with e, and e, (but only exceptionally with e,), these pits normally persisting to interradii 9 or 10, occasionally extending as far as 16. Interradial ridges enclosing interradii give rise to characteristic ‘Y’s on anterior part of fringe. I,, I, normally complete, although in some only I, present frontally. 1, and I, may appear as early as third and eighth radii respectively, but can appear in more posterior radii. Sculpture reticulate, coarse on glabella, finer on genal lobes. Fig. 78 Stapeleyella inconstans Whittard 1955. Llanfallteg Formation, Upper Arenig, Fennian Stage, D. levigena Biozone (a, f) or lower Llanvirn, D. artus Biozone (others). a, moult arrangement showing lower lamella, loc. 52U, Llanfallteg, x 3, It.18954; b, incomplete lower lamella, loc. 52, Llanfallteg, x 3, It.18955; c, group of specimens, loc. 55, Scolton, x 3-5, It.18956; d, incomplete lower lamella, Cl6g-y-fran, Whitland (Geological Survey loc. 37SE WA1), x 2:5, BGS TCC866); e, cephalon with attached thoracic segments, latex cast from external mould, loc. 55, Scolton, x 5, It.18957; f, pygidium, loc. 52B, Llanfallteg, x 4, It.18958. ARENIG IN SOUTH WALES 213 REMARKS. Whittard (1955: 36) described this species in detail, and noted the variable nature of details on the fringe. The material from south Wales falls within the range of variation described by Whittard, although in this material most specimens apparently have only I, present anteriorly, whilst Whittard (1955: 37) noted that only occasionally I, (i.e. I, in current terminology) is delayed until the eighth radius, with the norm being the presence of complete I, and I, arcs. However, I, and I, tend to be very close together frontally, so any distortion, which is widespread in our samples, could easily make a close pair of pits appear as one. Nevertheless, several specimens (e.g. Fig. 78b) definitely have only I, anteriorly. In the E series, most specimens have complete E,—E, arcs, although some have only E, and E,. Those with only E, and E, frontally occur in samples which include specimens with E,_, frontally. Our material is neither plentiful enough nor sufficiently well preserved for the identification of any successive changes in fringe morphology, if indeed such occur. S. inconstans is uncom- mon in the Arenig part of the Llanfallteg Formation, but becomes abundant in the basal Llanvirn, and slabs have been found at Scolton with large numbers of specimens. Two further species from the Lower Llanvirn of England and Wales have been ascribed, one with question, to Stapeleyella. Whittard (1955: 40) distinguished S. murchisoni (Salter), possibly from the Arenig, but more likely from the Lower Llanvirn, of the Shelve inlier, from S. inconstans on: ‘the more elongated glabella, on the sloping outer face of the cheek-lobe and on the absence of deep second and third glabellar furrows, of a pseudofrontal lobe, of a median tubercle, of a reticulated orna- ment on the glabella, and of a scrobiculate ornament on the cheek-lobes; the fringe is also more simple because the inner series of pits shows no more than two rows (?) and few auxiliary pits appear in the outer series’. The type material of S. murchisoni is variously distorted, and we suspect that many of these ‘differences’ are due to the vagaries of preservation. However, the details of the fringe do seem to be distinct. The type material of S.? etheridgei (Hicks) from the Lower Llanvirn of Llanvirn Quarry is badly preserved and distorted. It was placed by Whittard (1955: 34) in Bergamia with doubt, and transferred tentatively to Stapeleyella by Hughes et al. (1975: 560). Since S.? etheridgei occurs at exactly the same horizon as S. inconstans elsewhere in south Wales, it seems likely that it belongs to the same species. However, without extra better-preserved specimens from Llanvirn Quarry we regard it best for the time being to restrict the name S.? etheridgei to the type material. Stapeleyella abyfrons sp. nov. (Figs 79a—h) 1906 Trinucleus Sedgwickii Salt.; Evans: 609 (list). 1914 Trinucleus sp.; Thomas, in Strahan et al.: 19 (list). 1914 Trinucleus gibbsi Salt.; Thomas, in Strahan et al.: 19 (list). HOLoryPeE. It.18962, dorsal exoskeleton lacking lower lamella. TYPE LOCALITY AND HORIZON. Fennian, S. abyfrons Biozone, Pontyfenni Formation; Pen-y-parc (loc. 38). PARATYPES. Numerous specimens (e.g. It.18959-61, It.18964—-5, NMW 84.17G.140-159) from type locality, several specimens (It.18963, NMW 84.17G.161—3) from Regwm (loc. 26), and one specimen from entrance to former Llangan vicarage. STRATIGRAPHICAL RANGE. Fennian, S. abyfrons Biozone. DiaGnosis. Radii 16-17. E,_, complete, with e, , or e, only in interradii 0, iv, v or 1x or close thereto, but distribution highly variable, and can occur up to 12th interradius. Resultant crudely “Y’-shaped interradial ridges irregularly disposed, and fewer than in S. inconstans. I, 1, complete. Some I, pits laterally paired. Very fine reticulate sculpture on glabella and genal lobes. 214 R. A. FORTEY & R. M. OWENS Fig. 79 Stapeleyella abyfrons sp. nov. Upper Arenig, Fennian Stage, Pontyfenni Formation, S. abyfrons Biozone; loc. 38, Pen-y-parc (except e). a, cranidium, x 6, It.18959; b, small complete specimen, latex cast from external mould showing bacculae, x 3-5, It.18960; c, lower lamella, x 3-5, It.18961; d, latex cast from holotype, external mould of almost complete axial shield, x 2, It.18962; e, small cranidium with attached thoracic segments, loc. 26, Regwm, x 8, It.18963; f, small pygidium, latex cast from external mould, x 10, BGS Pr1765; g, lower lamella, x 6, It.18964; h, lower lamella, x 6, It.18965. Name. In reference to lack of ordered Y-shaped interradial ridges. DESCRIPTION. Glabella clavate, with deep 1P and 2P furrows, and weak 3P on pseudofrontal lobe. Small occipital pit a short distance behind 1P furrow, running into axial furrow close to it. Occipital furrow very shallow medially, occipital ring rather narrow, without spine. Broad, shallow axial furrow narrows and deepens anteriorly in vicinity of pseudofrontal lobe. Promi- nent, small, ovate alar lobe at posterior end of axial furrow. Genal lobes evenly rounded; those and glabella with very fine reticulate sculpture. Fringe narrow, with 16-17 radii, strongly flexed at girder. E, , complete. e, , or e, only in some interradii, concentrated towards the anterior; a typical distribution is in 0, iv, v and ix, but there is considerable variation on this pattern. These give rise to a few irregularly disposed, commonly asymmetrical ‘Y’-shaped interradial ridges. I,, 1, complete; some pits in I, laterally paired, the components of each pair being very close together. Pits near posterolateral corner of fringe irregularly disposed. Pygidium and thorax like those of S. inconstans except that the pygidial border does not broaden so much towards the posterior and the pleural furrows are a little deeper. REMARKS. S. abyfrons is the earliest Stapeleyella so far recorded. The principal distinctive feature which differentiates it from S. inconstans is the comparative disorganization of the e pits to produce a small number of rather irregularly disposed, ill-formed ‘Y’-shaped interradial sulci; there are fewer pits in both the E and I series. The presence of two arcs of E pits and two of I makes S. abyfrons intermediate in this feature between contemporaneous Bergamia with E,_, and I, and S. inconstans. The differentiation of Stapeleyella from Bergamia presumably took place in the Whitlandian, and the lower Llanvirn specimen that Hughes et al. (1975: 558; pl. 2, fig. 28) suggested was an ‘intermediate’ stage between Bergamia and Stapeleyella from north Wales is evidently not on this part of the phyletic lineage. It seems more likely to be a different Stapeleyella species. ARENIG IN SOUTH WALES 215 Stapeleyella aff. abyfrons sp. nov. (Figs 80a, b) 1909 Trinucleus sp.; Thomas, in Strahan et al.: 18. MATERIAL. Disarranged dorsal exoskeleton, and several cephalic fragments. HORIZON AND LOCALITY. Fennian, Pontyfenni Formation, ?B. rushtoni Biozone; cutting on A40, Castell-y-waun, west of Bancyfelin (loc. 21). DESCRIPTION. Glabella and genal lobes with fine reticulate sculpture, occipital ring without spine. Sixteen radii per half-arc. E,, E, complete, very close together anteriorly, more widely separated posteriorly. A few interradii with e, and e, present towards anterior. I, and I,, like the E series, very close together anteriorly, and more widely separated posteriorly. Both appar- ently complete; I, in fifteenth and possibly sixteenth radius. REMARKS. The presence of interradii in the E series invites comparison with such species as Bergamia rhodesi Whittard, B. matura Whittard and B. inquilina (Whittard), although there are apparently fewer than in any of these species. Unlike those, there is an (apparently) complete I, (note that Whittard 1955, 1966 used a different I series terminology from that used here) and some I,. The presence of incomplete E, and E,, together with (apparently) complete I, and I,, immediately invites comparison with Stapeleyella abyfrons, although there are fewer radii, larger pits and greater regularity, with only a few interradii in the E series and no paired I,. Although there are virtually no characteristic Y-shaped interradial ridges, we on balance prefer to associate these specimens with Stapeleyella, for other fringe characters are more similar to b Fig. 80 Stapeleyella aff. abyfrons sp. nov. Upper Arenig, Fennian Stage, Pontyfenni Formation, ?B. rushtoni Biozone, loc. 21, Castell-y-waun. a, complete specimen, x 5, It.18966; b, latex cast from external mould of same specimen, ~ 5. Genus GYMNOSTOMIX nov. TYPE SPECIES. Trinucleus gibbsii (Salter in Murchison 1859). Name. Greek yvyvoc bare, plus ctovvé, a beam, in reference to the marginal girder. DIAGNOSIS. Fringe narrow with marginal girder. I, and I, complete, I, present anteriorly and anterolaterally, a few I, anterolaterally. Prominent crest on genal lobe, extending across anterior end of axial furrow and merging with front of glabella. Pygidium like Bergamia. REMARKS. Whittard (1955: 34) placed the type species of Gymnostomix in Bergamia, and sug- gested that it might be the same as B. rhodesi. He listed a number of syntypes and gave a fringe 216 R. A. FORTEY & R. M. OWENS formula which included both E and I arcs, but admitted that it might be unreliable because of poor preservation of the syntypes. He was clearly mistaken in identifying E and I arcs, for G. gibbsii has a marginal girder (well seen in Figs 81c, d), which feature immediately distinguishes it from Bergamia. Instead G. gibbsii shows a greater or lesser degree of resemblance to other trinucleid genera with a marginal girder, Anebolithus, Incaia and Famatinolithus. From all it is distinguished by the presence of prominent genal ridges; in addition it differs from Anebolithus (Hughes et al. 1975: pl. 1, figs 16-19) in possessing some pits in I, and I, and lacking deep radial sulci. Famatinolithus (Hughes et al. 1975: pl. 1, figs 13-15) differs in having a broad marginal rim, no genal prolongations and a preglabellar field. Incaia (Hughes & Wright 1970: pls 127, 128; Hughes et al. 1975: pl. 2, figs 20-22) has a similar fringe, but differs in having prominent lateral eye tubercles and a different pygidial structure. Immature G. gibbsii (Fig. 81g), however, have distinct eye ridges, in contrast with, for example, similar-sized Stapeleyella inconstans (Whittard 1955: pl. 4, fig. 9) and Bergamia whittardi (Hughes 1971: pl. 8, figs 4, 6). Of the three genera discussed above, Gymnostomix shows closest resemblance to Anebolithus, the earliest representative of which occurs in the Moridunian, low in the Mytton Flags Forma- tion, Shelve inlier, and it seems likely that its origins lie in this genus. Gymnostomix gibbsii (Salter in Murchison 1859) (Figs 81a—j) 1859 Trinucleus Gibbsii Salter, in Murchison: 53; Fossils (9), fig. 7. 1866b Trinucleus Gibbsii Salter; Salter: 319; pl. 12, fig. 10 [copy, Salter in Murchison, 1859]. 1867 Trinucleus Gibbsii Salter; Murchison: 51; Fossils (10), fig. 7 [copy, Salter in Murchison, 1859]. 1872 Trinucleus Gibbsii Salter; Murchison: 51; Fossils (10), fig. 7 [copy, Salter in Murchison, 1859]. 1873 Trinucleus Gibbsii Salter; Salter: 22 [with figure]. 1881 Trinucleus Gibbsii Salter; Salter & Etheridge: 380 (list), 516; pl. 12, fig. 10 [copy, Salter in Murchison, 1859]. non 1884 Trinucleus Gibbsii Salter; La Touche: 56; pl. 2, fig. 33 [= Stapeleyella inconstans Whittard 1955]. 1906 Trinucleus sp.; Evans: 608 (list). ?non 1913 ?Trinucleus Gibbsii Salter; Postlethwaite: 14 (list). non 1914 Trinucleus gibbsi Salter; Thomas, in Strahan et al.: 19 (list). [= Stapeleyella abyfrons sp. nov. |. non 1932 Trinucleus gibbsi Salter (= T. etheridgei Hicks); Matley: 262 [= Bergamia? sp. of Hughes et al. 1975: 558; see discussion of S. abyfrons]. 1955 Bergamia gibbsi (Salter) Whittard: 33; pl. 4, figs 1, 2, S. 1960 Bergamia gibbsi (Salter); Whittard: 182 (list). 1971 Bergamia gibbsi (Murchison); Hughes: 145. 1975 Bergamia gibbsi (Salter in Murchison); Hughes ef al.: 558 (list). 1982 Bergamia gibbsii (Salter in Murchison); Owens & Fortey, in Bevins & Roach: 76 (list). 1982 Bergamia gibbsii (Salter in Murchison); Owens & Fortey: 257 (list). LECTOTYPE (selected Temple 1980: 221). GSM 23037, external mould of cephalon and thorax. TYPE LOCALITY AND HORIZON. Whitlandian, Penmaen Dewi Formation; old slate quarry south of St David’s head (presumed to be our loc. 61F). OTHER MATERIAL. Numerous specimens from Penmaen Dewi Formation, Pwlluog (loc. 61E, F); Colomendy Formation, Rhyd Henllan (locs 47A, B) and Whitland Abbey members (loc 27), and Afon Ffinnant Formation, loc. 18C. Outside south Wales, G. gibbsii has been found in the Whitlandian of Parwyd, Llyn Penin- sula (A. Beckly, personal communication, February 1984). Postlethwaite’s (1913) doubtful record from the Lake District seems to have been spurious, for no trinucleids are known from the Arenig part of the Skiddaw Slates Group; the only records appear to have been from the Llanvirn part in the Cross Fell inlier (Dr A. W. A. Rushton, personal communication, February 1984). ARENIG IN SOUTH WALES DY pas TE Ps ey JERE en aha: Ss Ree ” heise Sie. alia ces : OM RA Pi fe ee ie ER Be oF i g. x whe, i # ¥ ete Fig. 81 Gymnostomix gibbsii (Salter in Murchison 1859). Middle Arenig, Whitlandian Stage, G. gibbsii Biozone, Penmaen Dewi Formation, Pwlluog, Whitesand Bay (a-f), or Afon Ffinnant Formation, loc. 18C, Afon Ffinnant (g-j). a, lectotype cephalon and thorax, latex cast from external mould, old slate quarry (probably loc. 61F), x 4, BGS GSM23037 (original of Whittard 1955: pl. 4, fig. 1); b, complete specimen, latex cast from external mould, x 3-5, BM(NH) 46404; c, latex cast of ventral side of cephalon, showing marginal girder, loc. 61F, x 3, It.18967; d, cephalon, internal mould showing marginal girder, x 4, 11370; e, cranidium, loc. 61F, x 5, It.18968; f, cranidium, x 4-5, BGS GSM23034 (original of Whittard 1955: pl. 4, fig. 2); g, meraspid cranidium with eye ridges, x 15, It.18969; h, pygidium, x 8, It.18970; i, incomplete cranidium, latex cast of external mould, x 10, It.18971; j, incomplete cranidium, x 10, It.18972. STRATIGRAPHICAL RANGE. Whitlandian, middle and upper parts. DIAGNOsIs. Fringe with I, and I, complete, I, present anteriorly and anterolaterally, a few I, anterolaterally. Sculpture of coarse reticulate pattern on glabella and posterior genal lobes; finer on anterior part of genal lobes. 218 R. A. FORTEY & R. M. OWENS DESCRIPTION. Glabella clavate, expanding rapidly forwards. Deep 1P and 2P furrows, 3P probably present. Small glabellar node. Small, ovate, alar lobe at posterior end of axial furrow, occipital ring narrow, backwardly curving, and without spine. Genal lobe with prominent crest running from posterolateral corner to merge with front of glabella and enclosing anterior end of axial furrow. In the broad posterior border furrow, just inside the crest, there is a prominent posterior fossula. On glabella and genal lobe posterior to crest there is a coarse reticulate sculpture; the leading edge of the genal lobe, anterior to the ridge, has a very fine reticulate sculpture. This is commonly obscured when preservation is poor (Fig. 81a). A similar differen- tiation of genal lobe sculpture is seen in Anebolithus simplicior (e.g. Hughes et al. 1975: pl. 1, figs 17, 19). An immature cranidium (Fig. 81g) has prominent eye ridges, but no trace of these is present in the adult. Fringe narrow with narrow, upturned margin and rather narrow marginal girder. Number of radii in range 18-21. I, and I, complete in very shallow radial sulci (accentuated by certain directions of distortion). I, commonly present in anterior radii, typically between 2 and 6. Occasionally I, seen in posterior radii (e.g. Fig. 81a). 1; rarely present in radii 2-4. Thorax of typical trinucleine structure; pygidium similar to that of Bergamia, with broad margin which widens towards the sagittal line. Axis impinges onto margin, and contains 4—5 axial rings. REMARKS. Gymnostomix gibbsii is readily recognized and distinguished from other trinucleids by the presence of the distinctive genal ridges, which make it distinctive in the field and a valuable guide fossil for Whitlandian strata. Whittard (1955: 34) claimed that the ridges ‘were probably caused by the compression of the lateral cheek-lobes over their leading edges’ and stated that they were not present on all specimens, quoting GSM 49673 as substantiating this. We have examined this specimen, which is a slab containing in the region of thirty cranidia, nearly all of which clearly show the ridges. One specimen near the centre appears not to have them, but this is preserved in such a way that the anterior parts of the cheek lobes are compressed under the posterior parts. This would seem to be the only specimen upon which Whittard could have based his contention. We have found G. gibbsii in several different lithologies and kinds of preservation, and all specimens have the genal ridges. Moreover, no specimens of another trinucleid, Furcalithus sedgwicki, which occurs in association with G. gibbsii at the type locality, have any suggestion of such ridges, which might be expected if they were a preservational feature. We are thus in no doubt that the ridges are original. Bergamia? sp. of Bates (1968a: 184; pl. 13, figs 3, 4, 9, 13) from the ‘bifidus’ Beds of Anglesey also has crests on the genal lobes, but they are more rounded than in G. gibbsii, meet the glabella further back and do not cross the axial furrow. Because of preservation, Bates was unable to detect the position of the girder. It is possible that this material represents a second Gymnostomix species, although on balance an assignment to Bergamia seems more appropriate. Family DIONIDIDAE Giirich 1907 Snajdr (1981) proposed two dionidid genera, Dionidepyga (type species Dionide jubata Raymond 1925) and Dionideina (type species Dionide prima Klouéek 1916) which are of Llanvirn— Llandeilo and Llanvirn age respectively. Whilst he pointed out the difference one from the other, he compared neither with Dionide. Many of the characters listed for each genus (e.g. details of lateral glabellar furrows, presence of external girder) are common to both, and to Dionide, and cannot therefore be regarded as diagnostic. For Dionidepyga Snajdr noted that the pits on the fringe were not differentiated as to size; however, one of the cranidia figured by him (1981: pl. 4, fig. 3) shows clearly large pits just inside the border, very similar to those seen, for example, in the type species of Dionide, D. formosa (Whittington 1952: pl. 1, figs 1-3) from the mid-Caradoc of Bohemia and in D. magnifica (Owen & Bruton 1980: pl. 6, figs 1, 3, 6, 7) from the late Caradoc of the Oslo district. Snajdr’s (1981: 281) statement that a median glabellar spine is present in D. jubata cannot be substantiated: all figured specimens have only a small median node. We can see no justification for retaining Dionidepyga as a separate genus, and here regard it as a junior subjective synonym of Dionide. Whilst many characters of Dionideina ARENIG IN SOUTH WALES 219 are similar to Dionide, it differs in having alar lobes, a broad band of large pits inside the cephalic border and a reduced number of pygidial axial rings (c. 10) and pleural ribs (8—9). This combination of characters suggests that Dionideina should probably be regarded as being distinct from Dionide. In possessing alar lobes it is similar to Dionidella Prantl & Pribyl 1949a, but differs from it in lacking well-defined 2p glabellar furrows, in having large pits close to the cephalic border and a contrasting structure of the pygidial pleural ribs (see Ronit Iie fol il fig. 1). A further dionidid, Dionide (Paradionide) Chang & Fan 1960, type species D. (P.) anmenensis from the Lower Llanvirn of west Kansu, China, is distinctive in having a small, circular glabella, nine thoracic segments of similar length (sag., exsag.) and a very narrow axis. We agree with Owen & Bruton (1980: 21) in considering Paradionide a separate genus. Four dionidids are present in the Arenig—early Llanvirn of south Wales. The two later species can be assigned to Dionide, but the two earlier ones are more difficult to assign, and are placed with question in Dionidella (see p. 221). It is possible that these species belong to an independent genus, but the material is insufficient for its discrimination. Genus DIONIDE Barrande 1847 [ = Polytomurus Hawle & Corda 1847, Trigyrops Kobayashi 1940, Dionidepyga Snajdr 1981.] TYPE SPECIES. Dione formosa Barrande 1846, from the Zahorany Formation (mid Caradoc) of Bohemia; by original designation. Dionide turnbulli Whittington 1952 (Figs 82a, b) 1958 Dionide turnbulli Whittington; Whittard: 96; pl. 13, figs 1-8 (with earlier synonymy). Ho.otyre. SM A16715a, b; cephalon. TYPE LOCALITY AND HORIZON. Llanvirn, artus Biozone, Llanfallteg Formation; Long Plantation Cutting, Scolton (loc. 55). STRATIGRAPHICAL RANGE. Llanvirn, artus Biozone. D1aGnosis. Genal lobes smooth and connected by a smooth ‘preglabellar field’ in front of glabella; strong genal caecum; fringe crossed by anostomosing ridges; single row of larger pits immediately inside cephalic border. Pygidium with c. 18 axial rings, pleural areas with c. 12 ribs with deep interpleural furrows. OCCURRENCE. Outside the type locality recorded from Hope Shale and Stapeley Volcanic formations, Shelve inlier (see Whittard 1958: 98). Fig. 82 Dionide turnbulli Whittington 1952. Lower Llanvirn, D. artus Biozone, Llanfallteg Formation. a, latex cast of incomplete axial shield, old quarry north of Clarbeston, x 4, It.18973; b, incomplete topotype cranidium, loc. 55, Scolton, x 9, It.18974. 220 R. A. FORTEY & R. M. OWENS REMARKS. This species has been adequately described by Whittington (1952: 8), with supple- mentary remarks by Whittard (1958: 96). Extra topotypic specimens are figured here for comparison with D. levigena sp. nov., described below. Dionide levigena sp. nov. (Figs 83a-e, 84) Ho.otyPe. Complete specimen with anterior part of thorax telescoped beneath cephalon, NMW 85.26G.1. TYPE LOCALITY AND HORIZON. Lower Llanvirn, D. artus Biozone, Llanfallteg Formation; Cefn- maen-llwyd, loc. 50. PARATYPES. Complete specimen with damaged cephalon, BGS TCC454, stream 320m NW of Gelli, near Llangynog (Geol. Survey loc. Carm. 45NE W44); partially complete dorsal exoskele- tons (It.19006—7, NMW 84.17G.134, 136, 138), cephalon (NMW 84.17G.139), cranidia (It.19005, It.19008-9, NMW 84.17G.18, 19, 135), and lower lamella (NMW 84.17G.137) from D. levigena Biozone, Llanfallteg Formation, Llanfallteg, locs 52E, P-T, W. D1aGnosis. Cephalon with narrow, well-defined border; genal lobes smooth, not connected by smooth preglabellar field in front of glabella; fringe with numerous small, equal-sized pits, no anastomosing ridges; pygidium transverse, with nine axial rings and eight unfurrowed pygidial pleural ribs. Upper Arenig, Fennian Stage, D. levigena Biozone. a, holotype, complete specimen with anterior part of thorax telescoped beneath cephalon, latex cast from external mould, Cefn-maen-llwyd, loc. 50, x 6, NMW 85.26G.1; b, incomplete cephalon, loc. 52E, Llanfallteg, x 7, It.19005; c, complete specimen, cephalon damaged but showing pygidium well, stream 320m NW of Gelli, near Llangy- nog (Geological Survey loc. Carm. 45NE W4A4), x 6, BGS TCC454; d, incomplete cranidium showing median node on glabella, loc. 52W, Llanfallteg, x 10-5, It.19008; e, incomplete cranidium showing detail of pitted fringe, latex cast from external mould, loc. 52T, Llanfallteg, x 7, It.19009. ARENIG IN SOUTH WALES 221 Name. Latin: levis, ‘smooth’ and gena, ‘cheek’; reference to the smooth genal lobes. DESCRIPTION. Cephalon with narrow, well-defined border. Glabella plus occipital ring occupies about three-quarters cephalic length. Small median glabellar node present, no lateral furrows but short, deep basal furrows present. Occipital ring narrow (sag., exsag.), extending trans- versely as far as outer margin of basal glabellar furrow. Pleuroccipital furrow narrow but distinct, defining broad posterior cephalic border. Inner part of cheek smooth, weakly inflated. Fringe with numerous, small equal-sized pits which extend as far as frontal lobe of glabella. Anastomosing ridges absent. Genal angle prolonged into broad-based spine of unknown length. Thorax presumed to be of six segments, as in other Dionide species, but in the holotype (Fig. 83a) it is telescoped beneath the cephalon, and only parts of five segments can be seen. The other complete specimen (Fig. 83c) has only four, so this is probably a large meraspid. Axis narrow (tr.), the axial furrows forming a zetoidal pattern, so that each ring is distinctly wider (tr.) anteriorly than posteriorly. Pleurae very broad, bluntly truncated distally and with shallow, narrow pleural furrows that extend for most of their length, adaxially lying close to the anterior margin of the pleura, abaxially about half-way along (exsag.). Pygidium subparabolic, slender axis with nine narrow well-defined rings. Pleural areas with eight pairs of ribs with rather narrow, shallow pleural furrows that curve gently backwards adaxially; distally, all except the first turn more strongly posteriorly. No interpleural furrows. REMARKS. This species is distinct from most other Dionide species in having smooth genal lobes and a rather short, transverse pygidium with comparatively small numbers of axial rings and pleural ribs without interpleural furrows. These characters invite comparison with the dionidid Trinucleoides, although the latter has a rather different glabella with a deep, pit-like 2P furrow and a prominent spine on the frontal lobe. The sum of characters of D. levigena is closer to Dionide than to Trinucleoides, so we prefer to classify it with the former, at least until further material becomes available. Fig. 84 Reconstruction of Dionide levigena sp. nov., x 6 approx. Genus DIONIDELLA Prantl & Pribyl 1949a Type species. Dionidella incisa Prantl & Piibyl 1949a; Llanvirn, Sarka Formation, Sarka, Prague; by original designation. DiaGnosis. Cephalic anterior and lateral border furrows ill-defined, with cheeks and anterior and lateral cephalic borders pitted, except for small, smooth alar lobes; glabella with deep 1P furrows and small, pit-like 2P; thorax relatively longer and narrower than Dionide; pygidium 2D R. A. FORTEY & R. M. OWENS with 11 axial rings and 10 pairs of pleural ribs, with posterior pleural bands terminating well short of the margin and the anterior pleural bands abaxially tapering to a point, extending close to margin. REMARKS. Whittard (1958: 95) contended that acceptance of Dionidella as a distinct genus depended upon the ‘unique arrangement of the glabellar furrows and on the preglabellar field’. The preglabellar field does not seem to us to be a significant feature, but the glabellar morph- ology and the disposal of the pygidial pleural ribs, together with the combination of characters listed above, can be regarded as diagnostic. Rare dionidids from the Pontyfenni Formation have cephala whose morphology approaches that of D. incisa more closely than any Dionide species, although they lack the deep, pit-like 2P. Because of this latter circumstance, we assign these specimens to Dionidella with question. Dionidella? sp. indet. 1 (Figs 85a, b) 1914 Dionide sp.; Thomas, in Strahan et al.: 19. MATERIAL. Cranidium, BGS Pr1735-—6 (counterparts). HoRIZON AND LOCALITY. Fennian, S. abyfrons Biozone, Pontyfenni Formation; Pen-y-parc, loc. 38. DESCRIPTION. Cephalon sub-semicircular, without well-defined border, which is represented by only a slight change in slope. Glabella 0-6 of length (sag.) of cephalon, with small basal lobes, but these are not well seen on the specimen. Occipital ring much narrower (tr.) than glabella. Small, smooth triangular alar lobes extend forwards almost as far as basal glabellar lobes, their forward edges defined by low ridges. Anastomosing ridges with interspersed puncta radiate from glabella and extend up to cephalic margin. Posterior cephalic border broad and flat, narrowing slightly laterally, where the narrow posterior border furrow is deflected weakly backwards. It dies out before reaching the lateral margin, at position of lateral border furrow. Fig. 85 a, b, Dionidella? sp. indet. 1. Upper Arenig, Fennian Stage, S. abyfrons Biozone, Pontyfenni Formation, loc. 38, Peny-y-parc. a, cranidium, x 11, BGS Pr1735; b, incomplete cranidium showing fine pits and anastomosing ridges, x 8, It.18977. c, d, Dionidella? sp. indet. 2. Upper Arenig, Fennian Stage, B. rushtoni Biozone. c, cranidium, loc. 23, Pontyfenni, x 6, It.18978; d, pygidium, latex cast from external mould, loc. 24, Llwyn-crwn, x 6, It.18979. ARENIG IN SOUTH WALES 223 REMARKS. Apart from lacking 2P furrows, this cranidium differs from those of D. incisa in being proportionally shorter and wider, and in having prominent anastomosing ridges. So far as we are aware, this is the earliest recorded dionidid. Dionidella? sp. indet. 2 (Figs 85c, d) MATERIAL. Cranidia: It.18978 from Pontyfenni, loc. 23 and NMW 84.17G.165a, b from Blaen- lliwe, loc. 42. Pygidium: It.18979 from Llwyn-crwn, loc. 24. All from Fennian, B. rushtoni Biozone, Pontyfenni Formation. DESCRIPTION. Cranidium with semicircular outline, anterior and lateral borders defined only by a change in slope. Glabella 0-6 of length of cephalon (sag.), with basal lobes defined by shallow furrows. No alar lobes, and entire cheeks and preglabellar area covered by anastomosing ridges with interspersed puncta; these are weaker than in D.? sp. indet. 1. Pleuroccipital furrows markedly arched forwards, defining broad, smooth posterior cephalic border. Pygidium of subparabolic outline with narrow axis extending for most of pygidial length and comprising 14 rings. Pleural areas with 14 pairs of ribs which do not extend to margin. Interpleural furrows only apparent in their short, distal abaxial parts. Marginal area of pygidium smooth. REMARKS. We assume that the above material belongs to one species, since all occurrences are at approximately the same horizon. The principal difference from D.? sp. indet. 1 is in the cephalic outline (which may be in part the result of tectonic influences), the lack of alar lobes and weaker anastomosing ridges. The pygidium differs from that of D. incisa in having more axial rings and pleural ribs, and in not having the anterior pleural bands extending on to the pygidial border region. Family RAPHIOPHORIDAE Angelin 1854 Genus AMPYX Dalman 1827 TYPE SPECIES. Ampyx nasutus Dalman 1827; see Whittington, 1950. Ampyx linleyoides sp. nov. (Figs 86a—-e, 87d, 88a) 1906 Ampyx cf. salteri; Evans: 612. Hotorype. Exoskeleton lacking free cheeks, It.15946. TYPE LOCALITY AND HORIZON. Pontyfenni Formation, loc. 23; Upper Arenig, Fennian, B. rush- toni Biozone. STRATIGRAPHICAL RANGE. Upper Arenig, Fennian, B. rushtoni Biozone, also at loc. 53. FIGURED PARATYPES. Axial shields: It.19694, NMW 84.17G.81-2; cranidium: It.19693; pygidium: NMW 84.17G.83. OTHER PARATYPE MATERIAL. Axial shields: It.18579, NMW 33.189.G102; thorax and pygidium: It.19696; thorax: It.18553; pygidia: It.18559, NMW 84.17G.84. DIAGNOSIS. Ampyx with stout, long frontal spine with T-shaped cross section. Pygidium with axial rings poorly defined; border not clearly marked from pleural fields; raised lines on posterior border curving up onto periphery of pleural fields. Punctate surface sculpture. Name. Distinct from linleyensis. DIsCusSION. This species is extremely like A. linleyensis Whittard 1955, which he exhaustively described. Whittard described ‘long’ and ‘wide’ morphs of linleyensis; no taxonomic importance is attached to this, and the ‘long’ forms (such as in Fig. 86a) may include those which have undergone a small amount of tectonic extension. The holotype of linleyoides is undistorted. 224 R. A. FORTEY & R. M. OWENS Fig. 86 a-—e, Ampyx linleyoides sp. nov. Upper Arenig, Fennian, B. rushtoni Biozone, Pontyfenni Formation, loc. 23. a, axial shield slightly extended by distortion, x 2, NMW 84.17G.81; b, holotype, incomplete axial shield, x 3, It.15946; c, small axial shield, x 6, NMW 84.17G.82; d, cranidium preserving frontal spine showing carinate form, x 2:5, It.19693; e, axial shield, x 4, It.19694. f, Ampyx aff. linleyoides, same horizon and locality, cast from incomplete axial shield, x 4, It.19697. This specimen, and others as well preserved, show fine-scale dorsal punctation, faint on the pygidium, which is not recorded on linleyensis, but Rushton & Hughes (1981) record this kind of sculpture on A. cf. linleyensis from the Great Paxton Borehole. The frontal spine is long on A. linleyoides and in flattened condition has a characteristic carinate appearance. Examination of A. linleyensis shows that the spine on this species had the shape in cross-section of an inverted T, and that the consequent dorsal carina can, on some specimens (Whittard 1955: pl. 2, fig. 1), continue onto the front of the glabella. The specimen shown in Fig. 86d shows that the same spine shape was likely in A. linleyoides. One small specimen in relief (Fig. 86f) shows a frontal spine which is square in cross section; this specimen apparently also has wider fixed cheeks than linleyoides, and we are cautious about referring it to the same species; it is recorded as A. aff. linleyoides. ARENIG IN SOUTH WALES 225 The specific characters which permit recognition of linleyoides are on the pygidium. The pygidial border on linleyensis is extremely short and vertical, and clearly defined by an abrupt change in slope at the edge of the pleural fields. In A. linleyoides the border is not so defined, is wider, and carries some 8-10 raised lines which extend onto the lateral parts of the pleural fields. Flattening serves to exaggerate the differences; on A. linleyensis the border crushes down and the margin of the pleural fields may become a little elevated as a rim, while on A. linleyoides the border opens backwards to display the raised lines on the border clearly. We have examined more than thirty specimens of linleyensis, and of these only three show the raised lines extending onto the pygidial pleural fields. Other specific differences may be affected by preservation. The pygidial axis of linleyensis is more convex (tr.) posteriorly, standing well above the pleural fields; the tip of the pygidial axis on linleyoides is low and often effaced to such a degree that it is difficult to distinguish from the border. The expression of this difference obviously depends on the degree of flattening. The axis on linleyensis is narrower: its width at the anterior ring is half its length or even less, on undistorted material; on linleyoides we have width/length ratios between 0-6 and 0-7 in undistorted material, but measurement of axial length is difficult because of the posterior effacement. Most well-preserved pygidia of linleyensis have a distinctly concave-sided axis (Whittard 1955: pl. 2, figs 1-3); linleyoides has a straight- sided pygidial axis, even when it is distorted (Fig. 87d). On almost all specimens of linleyensis from the Stapeley Volcanics in Shropshire the axial rings on the pygidium are clearly-defined across the mid-part of the axis (the furrows are often kinked backwards medially). Six rings are clearly visible and up to fifteen have been observed. Definition of axial rings on linleyoides is poor; only two or three are clearly defined, the remainder up to a maximum of ten being very indistinct. This may have some connection with preservation in the dark mudstones of the Pontyfenni Formation, because linleyensis preserved in a similar lithology also appear to have less clearly defined axial rings (Fig. 87b), and we attach less importance to this than a compari- son of the types alone would suggest. A crushed linleyoides pygidium on which the ring furrows are deepened is shown on Fig. 87d. Probably more useful is the greater backward curvature of the anterior pygidial pleural furrows on linleyensis. The transverse line connecting the most posterior parts of these furrows cuts the third or fourth axial ring on linleyensis, but on linleyoides it cuts the second ring or the furrow behind it. The pygidial differences have been discussed at length because they are rather subtle. The specimen of linleyensis most like linleyoides which we have been able to find is figured in Fig. 87c; although the axial ring development is like the Arenig form it can still be distinguished on the structure of the border and the pleural furrows. The two species are, however, very closely related. A. linleyoides belongs to Ampyx, sensu stricto, as defined by Fortey (1975). Whittard (1955) distinguished linleyensis from other British species, and the same distinctions apply to linleyoides. As noted above, the species referred by Whittard (1955) to Ampyx salteri Hicks, from the Mytton Flags, is certainly not that species, which we here assign to Cnemidopyge. Whittard’s salteri (Fig. 87a) is another Ampyx, but not the same as either linleyensis or linley- oides. Its pygidium is relatively small, less transverse, and without backward-concave pleural furrows; the frontal spine on the cranidium is short, and with a circular cross section. Several other Ampyx species require discrimination from linleyoides. The type species, A. nasutus (Whittington 1950: pl. 74, figs 3-9) is generally very similar in proportions; but it has a tubular frontal spine, more obviously concave facial sutures, and an emarginate pygidial border. Because of their bowed pygidial pleural furrows A. nasutus, linleyensis and linleyoides should be included within any restricted concept of Ampyx. Of the numerous raphiophorids from the Arenig of Spitsbergen only one, A. spongiosus Fortey 1975, resembles A. linleyoides in overall proportions and surface sculpture. A. spongiosus has a rather slender frontal spine with a circular cross section, and the anterior pleural furrows on the pygidium (Fortey 1975: pl. 22, fig. 9) are not curved as they are on linleyoides, linleyensis and nasutus. Fortey (1975) noted the resemblance between A. spongiosus and A. volborthi Schmidt (sensu Skjeseth, 1952), which has a pygidial pleural structure like that of spongiosus, but with a rather sharply downturned border like linleyensis. Ampyx abnormis Yi, 1957 (see Lu 1975: pl. 39, figs 5-11; pl. 40, figs 1-7) apparently ranges from the late Arenig to Llandeilo in China. Lu notes that the pygidial 226 R. A. FORTEY & R. M. OWENS iE Fis i le x Ai ze vm i E 4 % pee Diag © jy ee th : i Cc %, —— ~ & Fig. 87 a, Ampyx cf. reyesi Benedetto & Malanca 1975 (Whittard’s A. ‘salteri’). Cast from external mould of axial shield for comparison with A. linleyoides sp. nov., ‘Lowermost Mytton Flags, head of Mytton Batch’, Shropshire, x 3, GSM 92940 (counterpart of Whittard, 1955: pl. 1, fig. 19). b, c, Ampyx linleyensis Whittard. b, axial shield preserved in mudstone for comparison with A. linleyoides in similar preservation, early Llanvirn, 300m east by south of Wernddu, 4 mile west of Llanllwch, Dyfed, x 4, GSM HT354; c, latest Arenig, Fennian, D. levigena Biozone, Llanfallteg Formation, loc. 52Q, x 3, It.19698. d, Ampyx linleyoides sp. nov. Flattened pygidium with overdeepened furrows, for comparison with A. linleyensis, Fennian (B. rushtoni Biozone), Pontyfenni Formation, loc. 23, x 4, NMW 84.17G.83. structure is like that of A. spongiosus rather than A. nasutus (and A. linleyoides). Of the specimens from the Arenig horizon illustrated by Lu that on his pl. 39, figs 10, 11 has a slender, unfluted frontal spine, and other cranidia show a preglabellar field, and apparently do not have incised glabellar furrows. Another Arenig form from China, A. yii Lu, would now be referred to Rhombampyx Fortey, 1975. Several South American species have also been described. One of these, A. reyesi Benedetto & Malanca 1975, from the ‘upper Arenig or Lower Llanvirn’ of Jujuy Province, Argentina, is similar to A. linleyoides in most respects, but like other species mentioned above the pygidial pleural furrows are nearly straight, and the frontal spine is described as having a circular cross section. Pribyl & Vanék (1980) described A. pallens from the Llanvirn of Bolivia, but without reference to reyesi, which it so strongly resembles that it may well prove its junior synonym. In any case A. reyesi/pallens provides by far the closest comparison with Whittard’s Ampyx (‘salteri’) from the Mytton Flags. Compare, for example, the pygidium of Fig. 87a with Pribyl & Vanék, 1980: pl. 19, fig. 1. For this reason we have designated the Shropshire species Ampyx cf. reyesi (Fig. 90). Similarity of raphiophorids between Britain and South America is given further support by the resemblance between the Ampyx? sp. figured by Harrington & Leanza (1957: fig. 116.4) from Argentina and A. linleyensis. Ampyx linleyensis Whittard 1955 (Figs 87b, c, 88c) (For synonymy see Whittard 1955: 18). Ho.LoryPe. Axial shield, GSM 92943. TYPE LOCALITY AND HORIZON. Stapeley Volcanic Group, Tasgar Quarry, Shelve district, Shrop- shire; Llanvirn, D. ‘bifidus’ Zone. ARENIG IN SOUTH WALES 227 OCCURRENCE IN SOUTH WALES. Llanfallteg Formation, type section at Llanfallteg Railway cutting from 20m below Arenig—Llanvirn boundary to highest beds in Llanvirn exposed; and at Rhyd-y-wrach farm and Scolton railway cutting. The species extends into the overlying black Llanvirn shales. Latest Arenig (Dionide levigena Biozone) to early Llanvirn. FIGURED MATERIAL. It.19698, GSM HT354. ADDITIONAL MATERIAL. NMW 33.189.G134, 33.189.G18, It.19699. Discussion. This species has been fully discussed under A. linleyoides sp. nov. above (pp. 224-5), which it resembles closely. Specimens from the Llanfallteg Formation have the charac- ters of linleyensis whether they are above or below the Arenig/Llanvirn boundary. Rushton & Hughes (1981) have described A. cf. linleyensis from the Great Paxton borehole. These speci- mens are more like linleyensis than like linleyoides, and they could well be accommodated within the range of variation of the former. We have mentioned above the possible occurrence of linleyensis in Argentina. Fig. 88 Comparative reconstructions of axial shields of British Arenig raphiophorid species. a, Ampyx linleyoides sp. nov., Fennian, B. rushtoni Biozone; b, Cnemidopyge salteri Hicks, Whitlandian, G. gibbsii Biozone; c, pygidium of A. linleyensis Whittard, latest Fennian and early Llanvirn. All x 3 approx. See also Fig. 90. Genus CNEMIDOPYGE Whittard 1955 TYPE SPECIES. Trinucleus nudus Murchison 1839, by original designation. DiaGnosis. The diagnosis of Hughes (1969) is followed here, except that we admit species with the pygidium smaller than the cephalon. 228 R. A. FORTEY & R. M. OWENS Cnemidopyge salteri (Salter 1873) (Figs 88b, 89a—c) 1873 Ampyx Salteri Hicks MS; Salter: 22. 1875 Ampyx salteri Hicks: 182; pl. X, figs 7, 8. non 1884 Ampyx salteri; La Touche: 56; pl. 2, fig. 34. non 1940 Ampyx salteri Hicks; Whittard: 161; pl. S, fig. 8. pars 1955 Ampyx salteri Hicks; Whittard: 15-18; pl. 1, fig. 15, non figs 16-21. non 1966 Ampyx aff. salteri Hicks; Whittington: pl. 2, figs 1-3, 6. 1969 Ampyx salteri Hicks; Hughes: 63 (compared with Cnemidopyge). pars 1978 Ampyx salteri Hicks; Fortey & Owens: 255; non fig. 8a. LectoryPE (herein selected). Exoskeleton lacking free cheeks, original of Hicks, 1875: pl. 10, fig. 8; BM(NH) 1352. TYPE LOCALITY AND HORIZON. ‘Middle’ Arenig of Hicks (1875), Penmaen Dewi Formation, the old quarry in Pwlluog. It is associated here with Bohemopyge scutatrix, Shumardia (Shumardia) gadwensis, Gymnostomix gibbsii etc. Whitlandian (biozone of Gymnostomix gibbsii). OTHER LOCALITIES. South of Dwyrhos Farm, Aberdaron, Llyn Peninsula; Whitlandian mud- stones. FIGURED MATERIAL. Dorsal exoskeletons lacking free cheeks: 114278, In.48526. ADDITIONAL MATERIAL. I730, 114279, SM A15592. DIAGNosIS. Cnemidopyge with relatively small and transverse pygidium having only three pairs of strong pleural furrows. DESCRIPTION. Available material of this species is all imperfect—flattened from the type locality, or somewhat distorted from north Wales. Partly because of this it has been much confused with other raphiophorids, such as Ampyx cetsarum. Axial shields show that the thorax is about as long (sag.) as the glabella, and both are longer than the pygidium. The length of the latter is exaggerated in flattened material because the borders are opened out in this preservation; less distorted material (Fig. 89b) indicates that the pygidial borders were declined nearly vertically. Cranidium with maximum width at posterior margin, this more than twice sag. length (excluding spine). Glabella is crushed on our material; it expands forwards rather more rapidly than is the case in many Ampyx species, for example to a maximum at about its mid-length, this being 0-8 of its sag. length. Although muscle insertions areas are obscured by crushing the larger specimens show clear evidence of two pairs of impressions within the glabella, which isolate a lateral glabella lobe, and which are consistent with the crushing of a cranidium of the kind illustrated by Hughes (1969: pl. 2, fig. 11). Frontal spine with circular cross section, slender, length not exceeding that of glabella. Occipital ring and border both well defined, border furrow a little convex-forwards. So far as can be judged the facial suture was also convex-forwards over much of its length, except where kinked backwards near genal angle. Free cheeks not known. Thorax with six segments, maximum transverse width at back end of second, first segment slightly macropleural (exsag.). Thorax tapers backwards, as it does in small, but not in large, holaspides of C. nuda; axis hardly tapers. The notable thoracic feature is the strong, forward arching on the pleural furrows of the early segments, particularly the first; it is progressively less marked on the second to fourth segments. Hughes (1969) figures triangular extensions of the axial rings towards the crests of the arched furrows, and some indication of the same feature in a crushed condition is shown by Fig. 89b. Pygidium transverse; as preserved flattened, its width is more than twice length, but with the borders turned downwards in original orientation it may have been three times as wide as long. Axis tapers gently to border (furrows enclose angle of about 30°); the back end of the axis is apparently not defined. Only three or four narrow (sag.) axial rings are defined. There are three pairs of deep pleural furrows, which slope gently backwards. Border wide, not medially emar- ginate. Apart from raised lines on pygidial border, surface sculpture is not preserved. ARENIG IN SOUTH WALES 229 Fig. 89 a—c, Cnemidopyge salteri (Salter 1873). Middle Arenig, Whitlandian, G. gibbsii Biozone. a, lectotype, flattened axial shield, Penmaen Dewi Formation, old state quarry, Pwlluog, north of Whitesand Bay, St David’s, Dyfed, x 2, 1352 (original of Hicks, 1875: pl. 10, fig. 8); b, cast from slightly distorted axial shield but preserving natural convexity, Whitlandian mudstones on track below Dwyrhos Farm, Aberdaron, Llyn Peninsula, north Wales, x 3, 114278; c, cast from axial shield, somewhat flattened but otherwise undistorted, horizon and locality as lectotype, x 3, In.48526. d, Ampyx cetsarum Fortey & Owens 1978, latex cast from entire specimen, Moridunian, lower part of M. selwynii Biozone, Ogof Hén Formation, Bolahaul Member, Llangynog, x 2, It.19700; this form has been confused with Cnemidopyge salteri in the past, but note distinctive pygidial structure. Discussion. Hughes (1969) suggested that Ampyx salteri Hicks might be an early representative of the genus Cnemidopyge. Hughes cited the well-furrowed pygidium as evidence for this assignment; the strong forward arching on the anterior thoracic pleural furrows and the struc- ture of the glabella are also consistent with the type species, C. nuda. Hughes revised the British species of Cnemidopyge, which are from the Llanvirn or later; all of them have larger pygidia with at least twice as many pleural furrows as in C. salteri. It was to be anticipated that the earlier species of the genus would have a smaller pygidium like those of other raphiophorids and we see no reason to exclude salteri from Cnemidopyge on the basis of this character alone. Raphiophorids from the early Arenig (Moridunian) in Wales which have been compared with C. salteri are all referable to Ampyx cetsarum Fortey & Owens 1978. There is now no possibility of confusing the two: A. cetsarum has a gently-forward expanding glabella, a charac- 230 R. A. FORTEY & R. M. OWENS Fig. 90 Reconstruction of axial shield of Ampyx cf. reyesi Benedetto & Malanca 1975. Moridu- nian, Shelve Inlier, Shropshire. x 3 approx. teristic occipital structure, and almost all furrows are suppressed on the pygidium. Our original material of this species was imperfect. Here (Fig. 89a) we illustrate a well-preserved, complete specimen, which was collected after our original work. Whittard (1955) identified C. salteri from the Shelve Inlier, from specimens in the Mytton Flags. Fig. 87a shows a cast from the external mould of the best preserved of Whittard’s specimens. On the basis of this identification, Fortey (1975) assigned salteri to Ampyx, sensu stricto. The Mytton form is, however, clearly different from salteri: it lacks the characteristic thoracic structure, the pygidium is relatively long (sag.) and the pygidial pleural furrows are not distinct. It is here referred to Ampyx cf. reyesi Benedetto & Malanca (Fig. 90). Fortey & Owens (1978: fig. 8a) used Whittard’s concept as the basis for their reconstruction of salteri, which is now seen to be incorrect; a new version is presented here as Fig. 88b. Family ALSATASPIDIDAE Turner 1940 Genus SELENECEME Clark 1924 TYPE SPECIES. Seleneceme propinqua Clark 1924, by monotypy. DIAGNOSIS. See Whittard (1960: 117-118); comments in Fortey & Shergold (1984: 352). Seleneceme acuticaudata (Hicks 1875) (Fig. 91) (For synonymy see Whittard 1960: 118). Hovotype. SM A15628, Llanvirn quarry. OCCURRENCE IN SOUTH WALES. S. acuticaudata appears in the Llanfallteg Formation, 18m below the Arenig/Llanvirn boundary, in the type section (loc. 52k), and persists into the early Llanvirn, ranging from the Dionide levigena to D. artus Biozones. It occurs in the Scolton railway cutting (loc. 55), Llanfallteg Formation, and widely in the D. artus Biozone. FIGURED SPECIMEN. It.19701. ADDITIONAL MATERIAL. NMW 33.89.G160, 84.17G.85-9. ARENIG IN SOUTH WALES 231 Fig. 91 Seleneceme acuticaudata (Hicks 1875). Uppermost Arenig, Fennian Stage, D. levigena Biozone, loc. 52a, Llanfallteg, cranidium and partial thorax, x 3, It.19701. Fig. 92. Placoparina sp. Latex cast from external mould, basal Llanvirn, loc. 52, Llanfallteg, x 7, NMW 85.67G. 1b. DISCUSSION. This species has been fully described by Whittard (1960), and the present material adds nothing to his description; it is figured as a matter of record. The first occurrence of the species is below the Arenig/Llanvirn boundary and hence earlier than in Shropshire. The glabellar tubercle is very subdued on our material, but we regard this as a matter of preser- vation only. Family CHEIRURIDAE Hawle & Corda 1847 Subfamily ECCOPTOCHILINAE Lane 1971 Genus PLACOPARINA Whittard 1940 TYPE SPECIES. Cryphaeus Sedwickii M‘Coy 1849, from presumed Llandeilo; Builth Wells district, Powys. By original designation. Placoparina sp. (Fig. 92) MATERIAL. A small, complete internal mould with counterpart external mould, NMW 85.67G.1a, b, from the basal Llanvirn, Llanfallteg Formation; loc. 52, Llanfallteg. A larger, badly preserved ‘ghosted’ complete specimen, NMW 84.17G.173, from same horizon, loc. 52V. REMARKS. The smaller, better preserved specimen is disarticulated behind the cephalon, the second and the seventh thoracic segments; the eighth thoracic segment is attached to the pygidium, and we interpret the specimen as a degree 7 meraspis. Although not well preserved, its general morphology is consistent with that of Placoparina (compare Whittard 1958: pl. 15, figs 6, 9). The larger specimen is less informative, but appears to possess the characteristic pleural spines on the pygidium. The specimens are not well enough preserved to tell whether they belong to P. sedgwicki (M‘Coy). Family PLIOMERIDAE Raymond 1913 Subfamily PLACOPARIINAE Hupeée 1953 Genus PLACOPARIA Hawle & Corda 1847 TYPE SPECIES. Trilobites zippei Boeck 1828, from the Llandeilo, Dobrotiva Formation; Prague district, Czechoslovakia. By original designation. REMARKS. Three subgenera of Placoparia have been described: Placoparia, Coplacoparia Hammann 1971a and Hawleia Prantl & Snajdr 1957. Of these Placoparia has been recognized in the Arenig of Britain, the Llanvirn of Britain, Brittany, Spain, Portugal and Bohemia and the Llandeilo of Bohemia; Coplacoparia from the Llandeilo of Spain and Brittany, and Hawleia from the Caradoc and Ashgill of Bohemia and Poland respectively. Hammann (1971a), Henry & Clarkson (1975) and Romano (1976) have studied in detail the evolution and distribution of DBP R. A. FORTEY & R. M. OWENS Placoparia species, recognizing in particular the progressive development of coaptative struc- tures. The only species with known Arenig representatives is P. (P.) cambriensis, a few speci- mens of which Whittard (1958: 108, as P. zippei; 1966: 284, as P. barrandei) recorded from the Tankerville Flags and Shelve Church Beds in the Shelve inlier. We have identified further specimens from the Pontyfenni Formation in south Wales. Subgenus PLACOPARIA Hawle & Corda 1847 Placoparia (Placoparia) cambriensis Hicks 1875 (Figs 93a—m) 1875 Placoparia cambriensis Hicks: 186; pl. 9, figs 1, 2. 1906 Placoparia cambrensis (sic) Hicks; Evans: 614, 617 (lists). 1909 Cheirurus sp.; Thomas, in Strahan et al.: 13. 1909 Sao sp.; Cantrill, in Strahan et al.: 33. 1958 Placoparia zippei (Boeck); Whittard: 104 pars; pl. 16, figs 6-10, text-fig. 6a, d—h [non fig. 6b, c, =P. zippei]. (With full synonymy). 1958 Placoparia sp.; Whittard: 108; pl. 16, fig. 11. 1966 Placoparia barrandei Prantl & Snajdr; Whittard: 283. 1967 Placoparia cambriensis Hicks; Dean in Whittard: 309. 1971a Placoparia (Placoparia) cambriensis Hicks; Hammann: 57; pl. 1, figs 3-9; pl. 3, fig. 27; text-fig. 2. 1971b Placoparia cambriensis Hicks; Hammann: 266, 270. 1974 Placoparia (Placoparia) cambriensis Hicks; Hammann: 114; pl. 10, figs 172-174. 1975 Placoparia (Placoparia) cambriensis Hicks; Henry & Clarkson: 88; pl. 1, figs 1-4; pl. 3, figs 1-3; text-fig. 3A. 1976 Placoparia (Placoparia) cambriensis cambriensis Hicks; Romano: 15; pl. 1a, b. 1976 Placoparia (Placoparia) cambriensis armoricensis Romano: 15. 1980 Placoparia (Placoparia) cambriensis Hicks; Henry: 51; pl. 5, figs 1—3; pl. 6, fig. 4; text-fig. 16A. 1984 Placoparia (Placoparia) cambriensis Hicks; Rabano: 10; pl. 1, figs 1-15 (with further synonymy for Iberian material). LecTOTYPE (selected Hammann 1971a: 58). BGS GSM35263, complete specimen from Llanvirn, D. artus Biozone; Llanvirn quarry, Abereiddi, Dyfed. Figured Hicks 1875: pl. 9, fig. 2; Whittard 1940: pl. 5, fig. 3; refigured here as Fig. 93m. MATERIAL. From Fennian, Pontyfenni Formation, B. rushtoni Biozone: NMW §84.11G.5a, cephalon with two attached thoracic segments; It.18980, ill-preserved complete specimen; It.19010, NMW 84.12G.10a, b, ill-preserved enrolled specimens; NMW 84.11G.4a, pygidium; all from loc. 23, Pontyfenni. BGS Pr616, ill-preserved complete specimen from loc. 24, Llwyn- crwn; BGS JP3533/34, incomplete cranidium from loc. 20A (Carm. 40NW WAS), Capel-Dewi. From S. abyfrons Biozone: BGS Pr1751, small incomplete cephalon and pygidium; BGS Pr1794, cranidium; NMW 84.17G.181, cephalon and part of thorax; all from loc. 38, Pen-y- parc. From Aber Mawr Formation, Pencil Slates: NMW 84.17G.182 enrolled specimen from loc. 62, Aber Mawr, Ramsey Island. From Tankerville Flags: BGS GSM92936, cranidium from Bergam quarry, Shelve inlier. From Shelve Church Beds: BGS GSM102436-7, cranidia from Wood House Gravels, Shelve inlier and BGS GSM92937, cranidium from Shelve. Numerous specimens, sometimes occurring in ‘graveyards’, from Fennian, D. levigena Biozone, Llanfallteg Formation: locs 52C, 52M, 52P-—S, 52U, 52W, Llanfallteg. Also common in basal Llanvirn, D. artus Biozone, Llanfallteg Formation: loc. 50, Cefn-maen-llWyd; loc. 52, Llanfallteg; loc. 55, Scolton. Elsewhere in south Wales, the species has been recorded from the Llanvirn, D. artus Biozone: old quarry north of Clarbeston (SN 0465 2210), It.18981; stream section SW of Cefn-maen- llwyd (Pem. 24SE EA12), BGS Pr2013; Bodau farmyard, Rhyd-y-wrach (Pem. 24SE EA13), BGS Pr2018/19; St Clears area, NMUW 33.189.G38, 33.189.G82, 33.189.G88; south of Castell-gorfod (Carm. 38NW W2A8); stream section between Ty-rhos and Gors, SW of Llangynog (Carm. 45NE W222), BGS JP3483/84. In other parts of Britain, the species occurs commonly at the same horizon in the Hope Shales Formation (see Whittard 1958: 108) and occasionally in the Skiddaw Slates Group, “oe A k x 5 aE agi Nite ‘ | m hy, Fig. 93. Placoparia (Placoparia) cambriensis Hicks 1875. Upper Arenig, Fennian Stage, Pontyfenni Formation, S. abyfrons Biozone, loc. 38, Pen-y-parc (c, f); B. rushtoni Biozone, loc. 23, Pontyfenni (a, b, d); Fennian Stage, Tankerville Flags, Bergam Quarry, Shelve inlier (g) and Shelve Church Beds, cart track north of Wood House, Gravels, Shelve inlier (e, h); Fennian Stage, D. levigena Biozone, Llanfallteg Formation, loc. 52P, Llanfallteg (j); lower Llanvirn, D. artus Biozone, Llanfallteg Formation (i, k—m). a, axial shield, x 3, It.18980; b, small pygidium, x 10, NMW 84.11G.4a; c, cranidium with associated free cheek, x 4, BGS Pr1794; d, cephalon with two attached thoracic segments, x 3:5, NMW 84.11G.5; e, cranidium, latex cast from external mould, x 5, BGS GSM102437; f, small incomplete axial shield, x 6:5, BGS Pr1751; g, cranidium, latex cast from external mould, x 5, BGS GSM92936 (counterpart of original of Whittard, 1958: pl. 14, fig. 10); h, cranidium, x 6, BGS GSM102436; i, cranidium with associated free cheek, old quarry north of Clarbeston, x 3-5, It.18981; j, ventral side of anterior end of axial shield with hypostoma, x 5-5, It.18982; k, pygidium with two attached thoracic segments, latex cast from external mould, loc. 50, Cefn-maen-llwyd, x 8, It.18983; 1, enrolled specimen, latex cast from external mould showing cephalon and pygidium (the pygidial pleural spines have been elongated by distortion), stream 275 m south of Rhyd-y-wrach (Geological Survey loc. Pem. 24SE E/12), x 5, BGS Pr2013; m, lectotype, complete axial shield, Llanvirn quarry, x 2, BGS GSM35263 (original of Hicks 1875: pl. 9, fig. 2). 234 R. A. FORTEY & R. M. OWENS where it has been reported from Outerside (SM A15622-3) and Whiteside (BGS GSM35239: syntype of Ormathops nicholsoni, see p. 244) in the Lake District, and at Ellergill (SM A15624) in the Cross Fell inlier. The numerous foreign occurrences are listed by Hammann (1971b), Henry & Clarkson (1975), Romano (1976) and Henry (1980). D1AGnosis. Glabella widens only slightly towards anterior; axial furrows straight; preglabellar furrow distinct, anteriorly convex; cheek with coarse pitted sculpture, with larger pits towards outer edge; pygidium with three or four axial rings and a terminal piece; pygidial spines short. REMARKS. This species has been described by Hammann (1971a: 58), and the diagnosis above is based upon his (our translation from the original German). Because Whittard (1966: 284) considered the Shelve Church Beds to be of extensus Biozone age, the range of P. cambriensis has been assumed to extend down into the lower- part of the Arenig. It is shown above (p. 98) that these are of likely Fennian age, and the known range of the species therefore extends no further back than the early Fennian. Whittard (1958: 108; pl. 16, fig. 11) figured a pygidium from the Shelve Church Beds as Placoparia sp., but noted that it showed no apparent differ- ences from P. cambriensis; we here include it in that species. Placoparia occurs infrequently in the lower two biozones of the Fennian in south Wales. The material is fragmentary and mostly not well preserved, but as far as can be seen corresponds in all essentials to P. cambriensis from the D. levigena Biozone and the Lower Llanvirn. Romano (1976) recognized two subspecies of P. (P.) cambriensis, one (cambriensis) with four rings on the pygidial axis, the other (armoricensis) with three, the two of mutually exclusive distribution in the Llanvirn: the former in Britain, north Portugal and Bohemia, the latter in Brittany and Spain. Rabano (1984) identified P. (P.) cambriensis (at one locality in association with P. (P.) c. armoricensis) in the Montes de Toledo, Spain, and argued on morphological and distributional grounds that there was no basis for two subspecies. In the light of Rabano’s data, we consider the pygidial difference to be more likely the result of individual variation, or of preservational differences, than of taxonomic significance. Family ENCRINURIDAE Angelin 1854 Subfamily DINDYMENINAE Henningsmoen 1959 Genus DINDYMENE Hawle & Corda 1847 TYPE SPECIES. Dindymene fridericiaugusti Hawle & Corda 1847, Ashgill Kraluy Dvir Forma- tion; Prague district, Czechoslovakia. Subsequently designated by Barrande, 1852: 816. DiAGnosis. Glabella inflated, clavate, expanding forwards; lateral glabellar furrows, when developed, short and shallow, being confined to vertical face of glabella adjacent to axial furrow; facial suture close to cephalic margin, running along border furrow; thorax of 10 or 11 segments; pygidium with 5—12 axial rings and two, or rarely three, pairs of pleural ribs. REMARKS. The above diagnosis is modified from those of Henningsmoen (in Harrington et al. 1959: 0448) and Kielan (1960: 146) to include Cornovica Whittard. The latter was proposed for the Llanvirn species C. didymograpti Whittard 1960, from the Hope Shales Formation, Shelve inlier, and Whittard laid stress upon the cephalic sculpture, the glabella not projecting beyond the cephalic border, the presence of glabellar furrows in addition to 1P and 11 rather than 10 thoracic segments as particulars which distinguish it from Dindymene species. The sculpture, however, is closely similar to that of certain Dindymene species (e.g. D. longicaudata—compare with Whittard 1960: pl. 17, figs 8-10 and Kielan 1960: pl. 30, fig. 2). The amount of glabellar projection is only a matter of degree, and can be affected by distortion; all Whittard’s Corno- vica didymograpti with the glabella well preserved (Whittard 1960: pl. 17, figs 8-10) are some- what distorted, and their appearance is very similar to that of distorted Dindymene hughesiae (Ingham 1974: pl. 18, fig. 12). The only real differences, therefore, are the presence of 2P and 3P lateral glabellar furrows and 11 thoracic segments. ARENIG IN SOUTH WALES 235 The bulk of described Dindymene species are of Ashgill age, and the genus has only scant representatives in the Caradoc and Llandeilo. The Llanvirn Cornovica didymograpti and the Arenig species described below are considerably earlier than most ‘typical’ Dindymene, and they might be expected to retain what are evidently primitive characters for the genus—2P and 3P glabellar furrows and 11 thoracic segments. The presence of three pairs of pleural ribs rather than the customary two in Dindymene longicaudata and an undescribed Dindymene from the Upper Whitehouse Beds, Girvan (Ingham 1974: 85) presumably bears testimony to earlier species with 11 thoracic segments. On the basis of these arguments we believe that Cornovica is best regarded as a subjective synonym of Dindymene. Whittard (1960: 123) claimed that Cornovica occupied an intermediate position between Dindymene and the Llanvirn genus Plasiaspis Prantl & Pribyl, because the latter has 12 thoracic segments and deep lateral glabellar furrows. However, the structure of the glabella and anterior border area of Plasiaspis (see Horny & Bastl 1970: pl. 15, fig. 2) is quite different from those of both of the others, although its general morphology suggests that its placement in the Dindymeninae is correct. More recently Strusz (1980: 6) excluded both Cornovica and Plasi- aspis from the Dindymeninae because, among other characters, both had a rostral plate; he claimed that the rostral and connective sutures are ankylosed in Dindymene, ‘of levisellid pattern’. Kielan (1960: 143) stated that the ventral cephalic sutures are incompletely known, but were probably of levisellid type. We can find no reference to a description of the ventral cephalic morphology in Dindymene, but a specimen of D. longicaudata figured by Kielan (1960: pl. 28, fig. 5) has the front part of the glabella missing and exposing what appears to be a rostral plate, which compares closely with that of Cornovica didymograpti figured by Whittard (1960: pl. 17, figs 1, 2). We would thus dispute that the sutures are ankylosed in Dindymene, or at least in D. longicaudata. There is therefore no reason on this basis to exclude Cornovica or Plasiaspis from the Dindymeninae. Even if they are found to be ankylosed in some Dindymene, and such might be expected to occur in late members of the lineage, all other characteristics point to a close association—to the extent that we have synonymized Cornovica with Dindymene. Strusz’ (1980: 6) claim that Cornovica and Plasiaspis are ‘probably cybelinids’ cannot therefore be upheld, nor do we see any reason to elevate the Dindymeninae to family status. Prosopiscus Salter (in Salter & Blandford 1865), included with question in the Dindymeninae by Hennings- moen (in Harrington et al. 1959: 0449), has recently been shown by Fortey & Shergold (1984: 357) to have affinities with the Phacopina. Kielan (1960: 144) added the Ashgill genus Eodindy- mene Kielan, which is like Dindymene in all respects apart from having the facial suture crossing the cheek. We are doubtful if this one character warrants generic distinction, and suggest that separation at subgeneric level at the most would be more appropriate. Dindymenella Lu et al. 1976 has lateral glabella furrows slightly longer than in Dindymene didymograpti, but not as long as in Plasiaspis. The pygidial pleurae are broad (tr.) and in contact for most of their length. On the basis of the figured material (Lu et al. 1976: pl. 14, figs 9-11) we are uncertain whether Dindymenella is a dindymenine, or belongs to another encrinurid subfamily. Dindymene saron sp. nov. (Figs 94a-1, 95) Ho .orype. It.18984, cephalon with eight thoracic segments. TYPE LOCALITY AND HORIZON. Fennian Stage, S. abyfrons Biozone, Pontyfenni Formation; loc. 38, Pen-y-parc. PARATYPES. From type locality: BGS Pri755, It.18989, NMW 84.17G.160, thoraces with pygidia; It.18986, incomplete cranidium. From B. rushtoni Biozone; loc. 23, Pontyfenni: It.18985, It.18988, complete specimens; It.18991, cephalon and thorax; It.18990, cephalon with thoracic segments; It.18987, cephalon. DiAGNnosis. Dindymene with three pairs of short lateral glabellar furrows; small, prominent 1P lobe; glabella with sculpture of fine granules, cheeks reticulate with tiny granules; thorax of 11 segments with long pleural spines; pygidium with 5—6 axial rings and two pairs of long pleural spines. 236 R. A. FORTEY & R. M. OWENS Fig. 94 Dindymene saron sp. nov. Upper Arenig, Fennian Stage, Pontyfenni Formation, B. rushtoni Biozone, loc. 23, Pontyfenni (b, d, e, h, i) or S. abyfrons Biozone, loc. 38, Pen-y-parc (a, c, f, g). a, holotype, cephalon with seven thoracic segments, x 5, It.18984; b, small axial shield, latex cast from external mould, x 10, It.18985; c, incomplete cephalon showing sculpture of cheek, x 5, It.18986; d, cranidium, latex cast from external mould, showing sculpture on glabella, x 12-5, It.18987; e, small axial shield, x 6, It.18988; f, thorax and pygidium, x 8, It.18989; g, small thorax and pygidium, latex cast from external mould, x 12-5, BGS Pr1755; h, cephalon with two thoracic segments, x 4-5, It.18990; i, small cephalon and thorax, x 11, It.18991. Fig. 95 Reconstruction of Dindymene saron sp. nov., x 8 approx. Facial suture not indicated (but see Whittard 1960: pl. 17, fig. 10). ARENIG IN SOUTH WALES D3, Name. Greek oapoy, a broom, alluding to the long thoracic pleural spines. DESCRIPTION. Glabella clavate, extending as far as, but not overhanging, anterior border, which seems to be defined by a shallow anterior border furrow (Fig. 94a); of three pairs of short lateral glabellar furrows, 1P deepest, which defines a small, ovate and rather prominent 1P lobe depressed below level of remainder of glabellar surface. Sculpture of fine, dense granules. Cheek coarsely reticulate; interspersed fine granules only seen in external moulds (Fig. 94d). Genal spine not well preserved in our material, but on the holotype appears to extend straight outwards and backwards from the genal angle. Facial suture not seen, but probably it ran outside the genal spine as on D. didymograpti. Thorax of 11 segments, whose pleural ribs are extended into long spines. A small specimen (Fig. 94b) appears to have only 10 thoracic segments, and this is presumably a late meraspis; it appears to have a finely granulose sculpture on the pleurae. On only one specimen (Fig. 94g) is the pygidial axis at all well preserved. It tapers backwards rapidly to a point, and has five rings and a tiny terminal piece. A larger specimen shows perhaps six axial rings. There are two pairs of long spinose pleurae, the anterior longer than the posterior pair. REMARKS. D. saron is the earliest known Dindymene. It differs from D. didymograpti (Whittard 1960) in details of the cephalic sculpture, particularly in the coarser reticulation of the cheek and the more numerous, finer granules on the glabella. The thoracic and pygidial pleural spines are much longer, and there are fewer (five or six as opposed to seven) axial rings. Rare and ill-preserved specimens of a Dindymene have been found in the Fennian part of the Llanfallteg Formation at locs 52V and 52W, Llanfallteg; these are presumably referable to D. didymograpti, although it is possible that they belong to what appears to be another species that occurs in the D. ‘bifidus’ Beds at Penarfynydd, Llyn (Figs 96a, b). This has a reticulate sculpture on the cheek, with only occasional small granules; the glabella is apparently smooth, and has only sparse granules. There are 10 thoracic segments. This combination of features distinguishes it from both D. didymograpti and D. saron, suggesting the presence of a third early Dindymene in the Arenig—Llanvirn of Wales. “S Fig. 96 Dindymene cf. didymograpti (Whittard } 1960). Lower Llanvirn, D. artus Biozone, Pen- arfynydd, Rhiw, Llyn Peninsula, Gwynedd. a, complete axial shield, x 65, NMW 27.110.G795; b, cranidium with four thoracic segments, showing sculpture on cheek, x 4, NMW 27.110.G648. Family CALYMENIDAE Milne Edwards 1840 Subfamily REEDOCALYMENINAE Hupé 1955 Genus NESEURETUS Hicks 1873 TYPE SPECIES. Neseuretus ramseyensis Hicks 1873, Moridunian, Ogof Hén Formation; Ramsey Island. Subsequently designated by Vogdes (1925: 106). REMARKS. We follow here the arguments of Whittard (1960), Whittington (1966) and Henry (1980) in placing Neseuretus in the Calymenidae, not the Homalonotidae, and accept Henry’s (1980) and Hammann’s (1983) assignment to the subfamily Reedocalymeninae. 238 R. A. FORTEY & R. M. OWENS Several Neseuretus species have been recovered from the British Arenig, and these have been described by Whittard (1960), Whittington (1966) and Bates (1968a, 1969), mostly from the lower part of the series, and are listed by Fortey & Morris (1982: 69, 70). Some of these are likely to be synonyms—for instance N. brevisulcus Whittard 1960 is distinguished from N. parvifrons (M‘Coy 1851) principally upon the degree of development of the weak anterior border furrow, a feature that Whittington (1966: 503) noted was highly variable. The distinc- tion seems to us to be accounted for by a combination of this variation and preservation; both species occur together in the lower third of the Mytton Flags. Bates (1969: 23) included N. grandior with question in the synonymy of N. ramseyensis. The differences between the two (only pygidia are known of the former) are 10 axial rings and 8-9 pleurae in the former, with corresponding figures of 8 or 9 and 7 or 8 in the latter. This hardly seems to us a specific difference, and we here include grandior in ramseyensis. Cyclopyge (see above) and Neseuretus are the only trilobites having species common to the Mytton Flags and the Arenig of north and south Wales; thus N. parvifrons occurs also in the Henllan Ash and N. murchisoni (rarely) in that formation and in the Ogof Hén Formation, Bolahaul Member. Our (Fortey & Owens 1978: 233 and 237, fig. 6) records of N. parvifrons in south Wales were erroneous; the material is all referable to N. murchisoni. Specimens resem- bling N. complanatus occur rarely at loc. 19. It may be assumed that those parts of the Mytton Flags (the lower third) with N. murchisoni and N. parvifrons equate with the Moridunian. Records of N. murchisoni from higher in the Mytton Flags (upper third), on the basis of Whittard’s 1960: pl. 29, fig. 12, do not belong to this species, and his figured specimen more closely resembles N. parvifrons, which he records (1966: 304) as occurring throughout the formation—in contradiction to his earlier (1960: 144) statement that it is apparently restricted to the lower third. It is possible that the higher records refer to forms like those on his 1960: pl. 20, fig. 12, and not to parvifrons, sensu stricto. Hammann (1983: 57) has shown how Llanvirn—Llandeilo Neseuretus species in Spain favour particular facies, and the same apparently holds true for British Arenig species. Thus Neseu- retus ramseyensis occurs in micaceous silty mudstones or in arenaceous sediments; N. parvifrons in micaceous silts of the Mytton Flags and sandy mudstones or muddy felspathic sandstones in the Henllan Ash; and N. murchisoni in the muddy sediments of the Bolahaul Member and within the Mytton Flags. Neseuretus ramseyensis Hicks 1873 (Figs 97a—g) 1873 Neseuretus ramseyensis Hicks: 44-45; pl. 3, figs 7-10, 16-22. 1960 Neseuretus grandior Whittard: 141; pl. 20, figs 1, 2. 1960 Neseuretus parvifrons (Salter); Whittard: 145 pars (reference to N.? elongatus only). 1960 Neseuretus murchisoni (Salter); Whittard: 148 pars; pl. 21, figs 1, 2 only. 1969 Neseuretus ramseyensis Hicks; Bates: 23; pl. 8, figs 3, 4, 6-12; pl. 9, figs 1-3, 6 (With full synonymy). 21975 Neseuretus grandior Whittard; Struve: 279, fig. 28. LECTOTYPE (selected Whittard 1960: expl. to pl. 21). BGS GSM10166, cranidium from Ogof Hén Formation, Ramsey Island. OccuRRENCE. At type locality; in lowermost, silty part of Bolahaul Member, Llangynog inlier and Star Cottage, Carmarthen (loc. 8, Fortey & Owens 1978); Stiperstones Quartzite, Shelve inlier. DiaGnosis. Neseuretus with narrow, upturned anterior border, distinct preglabellar furrow; lateral glabellar furrows weak on larger specimens; pygidium with 8-10 axial rings, 7—9 pleural ribs, axis not reaching posterior margin. REMARKS. Bates (1969: 23) redescribed this species, but noted that distortion makes accurate description difficult. The cephalon closely resembles that of N. murchisoni, and these species can really only be discriminated upon pygidial differences, N. ramseyensis typically having 8 ARENIG IN SOUTH WALES 239 Fig. 97 Neseuretus ramseyensis Hicks 1873. Lower Arenig, Moridunian Stage, Ogof Hén Formation, Bay Ogof Hén, Ramsey Island (a, b, f, g), Bolahaul Member, Llangynog (c, d) and Star Cottage, Carmarthen (Fortey & Owens 1978: loc. 8B) (e). a, cephalon and thorax, latex cast from external mould, x 2:5, NMW 29.308.G240b; b, pygidium showing eight axial rings, x 3, NMW 29.308.G63; c, large pygidium, x 1:5, NMW 78.8G.41; d, small pygidium, relatively undistorted, latex cast from external mould, x 5-5, NMW 78.8G.42; e, small distorted pygidium (compare with Fig. 97d), x 4, NMW 75.45G.235a; f, small cranidium, x 6, NMW 27.110.G762; g, small pygidium with four attached thoracic segments, x 2, NMW 85.68G.1. (sometimes 9) axial rings and 7 pairs of pleural ribs, compared with 5 and 4 in N. murchisoni. Much of the material of N. ramseyensis is of a comparatively large size, and it might be considered that the larger numbers of rings and ribs are a function of this. However, specimens of comparable size show the same differences, so the presence of two taxa, albeit closely related, seems to be real. Struve (1975: 219, fig. 28) figured an incomplete external mould of a Neseuretus pygidium in an erratic quartzite pebble from Hessen, Germany, as N. grandior Whittard (= N. ramseyensis, see above). Although there is general resemblance to Whittard’s specimens, we only include this specimen with question in ramseyensis because the pleural areas are proportionately much narrower (although this may be an artefact of preservation), and insufficient details can be discerned for a definitive identification. Nevertheless, its occurrence suggests the presence of shallow-water Neseuretus-bearing quartzites on the southern margin of Tornquist’s Sea, be they of Arenig or later age. Neseuretus murchisoni (Salter 1865) (Figs 98a—e) 1960 Neseuretus murchisoni (Salter); Whittard pars: 148; pl. 20, figs 6-11, 13-15 [non fig. 12, =?N. parvifrons ]. (With full synonymy). 1966 Neseuretus murchisoni (Salter); Whittington: 503; pl. 4, figs 14, 15, 17-19. 1978 Neseuretus parvifrons (M‘Coy); Fortey & Owens: 233, 237. LecToryPE (selected Shirley 1931: 14). BGS GSM35256, cranidium from Mytton Flags; Lord’s Hill, Snailbeach. Figured Shirley (1931: pl. 1, fig. 5) and Whittard (1960: pl. 20, fig. 6). OCCURRENCE. In south Wales the species is recorded from mudstones of the Bolahaul Member, Ogof Hén Formation, locs 9 and 12 of Fortey & Owens (1978: 241), and from temporary exposures (1981) along Carmarthen bypass, south of Pensarn, Carmarthen. It occurs rarely in 240 R. A. FORTEY & R. M. OWENS Fig. 98 Neseuretus murchisoni (Salter 1865). Lower Arenig, Moridunian Stage, Ogof Hén Formation, Bolahaul Member, temporary exposures (1981) on Carmarthen bypass, south of Pensarn (a, e), temporary exposure at water treatment plant, Penddaulwyn Fawr, Capel-Dewi (Fortey & Owens 1978: loc. 12) (d) and tipped material at Wennallt, Allt Cystanog, Carmarthen (Fortey & Owens 1978: loc. 9) (b, c). a, undistorted cranidium, preserving original relief, x 3-5, NMW 85.17G.1a; b, cranidium, latex cast from external mould, x 4, NMW 75.45G.230; c, thorax and pygidium, latex cast from external mould, x 3-5, NMW 75.45G.228; d, cranidium showing large palpebral lobe, x 4:5, NMW 76.3G.19; e, thorax and pygidium, x 2, NMW 85.17G.2a. the Carnedd Iago Formation, Henllan Ash Member, Arennig Fawr district (Whittington 1966: 503), and commonly in the lowest one-third of the Mytton Flags, Shelve inlier. DIAGNOsIS. Cranidium like N. ramseyensis; pygidium with axis anteriorly about one-third anterior breadth (tr.), and extending to posterior margin, composed of five rings and a terminal piece; four pairs of pleural ribs. REMARKS. This species has been fully described by Whittard (1960: 148), and well-preserved specimens from the Bolahaul Member are illustrated here for comparison. Neseuretus cf. complanatus Whittard 1960 (Figs 99a—d) MATERIAL. One incomplete cephalon and two incomplete pygidia from the Whitlandian, Afon Ffinnant Formation, loc. 19. REMARKS. The cephalon resembles N. complanatus from the Mytton Flags in having a flat lateral profile (compare Fig. 99b with Whittard 1960: pl. 20, fig. 5), but we are uncertain whether this is at least in part produced by compression both in our specimen and in Whit- tard’s. Otherwise it is closely comparable with that of N. parvifrons except that in the type of complanatus the palpebral lobe is apparently larger and in it and our specimen the glabella is proportionately longer and narrower. Subfamily COLPOCORYPHINAE Hupé 1955 Genus COLPOCORYPHE Novak, in Novak & Perner 1918 TYPE SPECIES. Calymene Arago Rouault 1849 from the Llanvirn—Llandeilo of la Couyere, Ille-et- Villaine, Brittany, France. By original designation. ARENIG IN SOUTH WALES 241 Fig. 99 Neseuretus cf. complanatus Whittard 1960. Middle Arenig, Whitlandian Stage, Afon Ffinnant Formation, loc. 19, Ffinnant road cutting. a—c, cranidium, internal mould, lateral view of same and latex cast from external mould, x 3, It.18992; d, incomplete pygidium, x 3, It.18993. Colpocoryphe taylorum sp. nov. (Figs 100a—h) Ho.otyPe. NMW 84.11G.6a, complete specimen. PARATYPES. It.18994—S, cranidia; It.18996, NMW 77.9G.36, 84.12G.26, 28, 84.17G.128-133, thoraces with pygidia; NMW 84.12G.25a, b, 84.12G.27, pygidia. All loc. 23. TYPE LOCALITY AND HORIZON. Fennian Stage, B. rushtoni Biozone, Pontyfenni Formation; loc. 23, Pontyfenni. OTHER MATERIAL. A badly-preserved complete specimen from loc. 22, Sabulon, probably belongs to this species. D1AGNosIs. Colpocoryphe with subquadrate glabella with truncated frontal lobe; palpebral lobe small, distant from axial furrow; pygidium with one incised pleural and interpleural furrow; axis with 12 rings. Name. For Mr C. T. and Mrs I. Taylor, who presented the authors with a number of important specimens. DESCRIPTION. Glabella subquadrate, frontal lobe truncated, more or less transverse. 1P furrow deep, running obliquely backwards at about 45° to exsagittal line drawn through its axial end. Abaxially there is a suggestion of a bifurcation. 2P a little over half way along glabella from posterior; it is somewhat shorter and shallower than 1P, and is directed backwards at a smaller angle. 3P short, weak, rather less than halfway from 2P to anterior margin of glabella, and directed slightly backwards, transverse or slightly forwards: this variation is probably due in part to degree of compression. One specimen shows a small 4P, a short distance in front of 3P, of similar depth and length, and directed slightly forwards. Occipital furrow approximately transverse, curved weakly forwards laterally. Occipital ring as wide as, or marginally wider (tr.) than glabella. Axial furrow more or less straight for most of its length, curving inwards towards occipital furrow round 1P lobe. Preglabellar furrow shallower than axial, confluent with anterior border furrow. Inner part of fixed cheek broad; palpebral lobe situated distant from axial furrow with its margin slightly elevated above level of adjacent cheek. It is small, about an eighth of the sagittal length of the glabella on larger specimens. The only specimen with the fixed cheek preserved (Fig. 100c) shows a small eye, making this one of the few eyed species in the Fennian atheloptic assemblage. Pre- and postocular branches of facial suture both outwardly convex, the postocular branch being considerably longer than the preocular. Anterior cranidial border clearly defined and backwardly convex in dorsal view; lateral cephalic border hardly defined, being confluent with remainder of free cheek. Posterior border furrow deep and broad, defining abaxially widening posterior cephalic border. 242 R. A. FORTEY & R. M. OWENS Fig. 100 Colpocoryphe taylorum sp. nov. Upper Arenig, Fennian Stage, B. rushtoni Biozone, Ponty- fenni Formation, loc. 23, Pontyfenni. a, cranidium showing fine granulose sculpture and small palpebral lobe, x 4, It.18994; b, small thorax, x 10, NMW 77.9G.36; c, holotype, almost complete axial shield, x 2, NMW 84.11G.6a; d, f, small pygidium, latex cast from external mould showing granulose sculpture, and internal mould, x 10, NMW 84.12G.25a, b; e, large pygidium, x 3, NUW 84.12G.27; g, small cranidium, x 5, It.18995; h, thorax and pygidium, showing rounded lateral extremities of thoracic pleurae, x 3, It.18996. Thorax of thirteen segments, which give a typically concave lateral profile which may be consistent with a burrowing mode of life for this species of the kind suggested by Hammann (1983: 29, text-fig. 11). Pleurae with narrow anterior and broad posterior bands, and flexed steeply downwards at fulcrum. Pygidium with distinct border furrow typical for Colpocoryphe, which tends to be less pro- nounced on smaller specimens than on larger ones. Pleural area with one pleural and one interpleural furrow, but otherwise smooth. Axis broad, with 11-12 rings which become progres- sively narrower (sag.) and less well defined towards posterior. It tapers backwards only slightly as far as a position opposite border furrow, after which it tapers rapidly to a blunt end which merges insensibly with posterior border before reaching posterior margin. Sculpture on cranidium and pygidium of numerous, very fine granules. REMARKS. The most distinctive feature of this species is the small palpebral lobe, situated distant from the axial furrow. This, together with the subquadrate glabella, immediately dis- tinguishes it from the other Arenig species C. thorali Dean and C. maynardensis Courtessole, Pillet & Vizciano, both from the Montagne Noire. Both have a much larger eye which is situated close to the axial furrow, as well as more furrows in the pleural fields (see e.g. Dean 1966: pl. 11, figs 1, 2, 10; pl. 12, figs 1, 4, 6-9, 11; Courtessole et al. 1983: pl. 3, figs 5, 7, 8; pl. 4, figs 2, 3, 6, 10, 11, 13, 14, 16, 18; pl. 5, figs 3, 6, 7). Courtessole et al. (1983: 17, 22) considered two species, with concurrent stratigraphical ranges, to be represented within Dean’s (1966) material of C. thorali. Some they retained in Colpocoryphe as a new species, C. deani, with the ARENIG IN SOUTH WALES 243 remainder placed in Salterocoryphe thorali. They did not make clear their precise reasons for this division, nor did they fully justify their transferrence of thorali to Salterocoryphe. On the basis of the illustrated material we follow Dean, and consider one species only to be present, and to belong to Colpocoryphe rather than the closely related Salterocoryphe; in particular it lacks the distinctively shaped glabella and ribbed pygidial border of the latter genus. Of Llanvirn species, C. thorali conjugens Hammann from Spain has a larger eye, close to the axial furrow (e.g. Hammann 1983: pl. 12, figs 112a, b), but like this subspecies, C. taylorum has a subquadrate glabella and one pleural and one interpleural furrow on the pygidium (cf. Figs 100d, e, f and Hammann 1983: pl. 12, figs 111b, 113a), whilst the Llandeilo C. rouaulti Henry from Brittany (Henry 1970: pl. B) and Spain (Hammann 1983: pl. 13) has a proportionately longer, narrower glabella with a more rounded frontal margin, the eye closer to the axial furrow and only one pleural furrow on the pygidial pleural field. Compared with all other Colpocoryphe species, C. taylorum is the only one found in sediments of comparatively deep water origin, and to this may be related the much reduced eye. Family DALMANITIDAE Vogdes 1890 Subfamily ZELISZKELLINAE Delo 1935 Genus ORMATHOPS Delo 1935 Type species. Dalmanites atavus Barrande 1872, from the Llanvirn Sarka Formation, Osek, near Rokycany, Bohemia. By original designation. DiaGnosis. Zeliszkelline with broad, flat anterior border; facial suture distant from anterior glabellar margin; 3p straight or sigmoidal; eye forwardly placed, ranging from being moder- ately large to being reduced to a few lenses, or may be absent; lenses, apart from those on upper horizontal row, are all the same size, and lens-packing is never entirely regular; deep palpebral furrow present in most species; small genal spine present in immature specimens. Pygidium triangular to subparabolic, axis with 7-10 rings, pleural areas with 4-6 pairs of ribs; hypostome with straight lateral margins, or slightly constricted at mid-length. OccurRRENCE. Arenig—Llanvirn of British Isles, Montagne Noire, France, Anti-Atlas, Morocco, and Bohemia. REMARKS. Ormathops is the earliest known member of the Phacopina and is the earliest trilobite with schizochroal eyes (Clarkson 1971). Unlike all other Phacopina, the eye lenses of Ormathops are the same size, and the packing arrangement is never entirely regular, as has been demon- strated by Clarkson (1971). The earliest Ormathops is O. borni Dean 1966 from the D. extensus Biozone of the Montagne Noire. Destombes (1972) recognized this species from a similar horizon in the Anti-Atlas, Morocco, but placed it in Pterygometopus because all of the lateral glabellar furrows are deeply incised, the eyes are large and because of the outline and configu- ration of the palpebral lobes, hypostome and pygidium. Now that more Ormathops species are known these characters cannot be used to exclude O. borni from the genus. In particular, Clarkson (1971) has shown that O. borni and the type species O. atava both have a distinctive kind of primitive schizochroal eye. The Moroccan O. borni has a consistently larger eye than does the material from the Montagne Noire, and it may be a different species. On the other hand, Clarkson (1971) described a wide variation in the size of the eye and number of lenses in O. atava which might be the result of intraspecific variation; the same may apply to O. borni. O. atava, of Llanvirn age, has somewhat smaller eyes than does O. borni, but Fennian species from south Wales are either blind, as in the case of O. nicholsoni, or have the eye reduced to only a dozen or so lenses, as in O. Ilanvirnensis. Thus eye reduction or loss occurs in the earliest Phacopina, and anticipates the same sequence which is repeated in several different lineages of later taxa, especially in the late Devonian. In both the Fennian and the late Devonian, this eye reduction and loss is found in atheloptic assemblages. There might be a temptation to separate off the reduced-eyed and blind Ormathops species as subgenera. We would caution against this, however, because it is likely to produce a facile 244 R. A. FORTEY & R. M. OWENS artificial classification that might obscure true relationships. Eye reduction and loss probably occurred iteratively within Ormathops, with several short-lived independent lineages developing reduced eyes or eye loss as they invaded deeper-water offshore habitats. Ormathops nicholsoni (Salter 1866) (Figs 101a—m, 102) 1866 Phacops Nicholsoni Salter, in Harkness & Nicholson: 486 pars, figs c, d (lectotype only; other syntype probably Placoparia). 1875 Phacops Nicholsoni Salter; Hicks: 187. 1876 Phacops Nicholsoni Salter; Ward: 106. 1881 Phacops Nicholsoni Salter; Etheridge: 113. 1885 Phacops nicholsoni Salter; Postlethwaite: pl. i, figs 2-4. 1886 Phacops nicholsoni Salter; Postlethwaite & Goodchild: 457; pl. 1, fig. 2. 1905 Phacops Nicholsoni Salter; Reed: 176. 1935 Phacops nicholsoni Salter; Delo: 403. 1940 ?Calyptaulax nicholsoni (Salter) Delo: 9, 12. 1958 Ormathops nicholsoni (Salter) Struve: 183, 187. 1960 Ormathops nicholsoni (Salter); Whittard: 128 pars; pl. 16, fig. 6 [non fig. 7, = O. llanvirnensis]. 1966 Ormathops nicholsoni (Salter); Dean: 297. 1971 Ormathops nicholsoni (Salter); Clarkson: 52. LecToTyPe. From Salter’s two syntypes, Whittard (1960: 129) selected a moult arrangement of cephalon and thorax plus pygidium (Tullie House Museum, Carlisle) from the Skiddaw Slates Group (probably Fennian) of Whiteside, Cumbria. However, Goodchild (in Postlethwaite 1885, plate explanations) stated that his figure 2, which is the same specimen, was ‘drawn from the original type specimen figured by Mr Salter’; selection of the lectotype therefore dates from Goodchild, not Whittard. MATERIAL. Besides the lectotype, Salter (in Harkness & Nicholson 1866: 486) mentioned a further specimen, now BGS GSM352339, an ill-preserved incomplete thorax and pygidium from Whiteside. On examination, we consider that this specimen is referable to Placoparia cam- briensis and not to O. nicholsoni. Another ill-preserved Lake District specimen labelled nichol- soni, GSM 35240 from Skiddaw, is also more likely to be a Placoparia. In south Wales abundant well-preserved specimens have been collected from the Fennian, B. rushtoni Biozone, Pontyfenni Formation, at Pontyfenni, loc. 23, and a few also from Capel-Dewi, loc. 20D, and the Aber Mawr Formation at Aber Mawr, Ramsey Island, loc. 62. D1AGNosis. Ormathops with preocular suture close to anterolateral corner of glabella; without eyes, but with deep palpebral furrow on anterior end of cheek; 3P furrow deep, sigmoidal; pygidium broadly triangulate with 9 axial rings and 6 pairs of pleural ribs with deep, narrow interpleural furrows; sculpture of fine pits on cheek, and very fine puncta on glabella and cephalic border; hypostoma with straight, backwardly converging lateral margins. DescriPTION. Glabella expanding rapidly forwards in front of 2P lobe so that width (tr.) across basal lobes is 50%-—60% of that across frontal lobe. 1P—3P furrows all of similar depth, and all extend for about the same distance inwards, extending more than halfway towards sagittal line. 1P furrow directed slightly backwards, and deepens and widens adaxially, and bifurcates dis- tally into two short branches, the posterior of which runs approximately exsagittally. 2P furrow more or less transverse; 2P lobe slightly longer (exsag.) than 1P; 3P furrow sigmoidally curved, and directed strongly obliquely backwards. Occipital furrow arched weakly forwards sagittally, and of approximately equal depth along its length (tr.). Occipital ring widens a little sagittally, and has a small median node. Cheek weakly convex. Short, prominent palpebral furrow present in a forward position, running parallel to the margin, its adaxial end opposite 3P furrow, close to axial furrow. It is proportionately further back in small specimens (e.g. Fig. 101i). A short, raised palpebral lobe is present in front of palpebral furrow, but the eye is apparently absent. We have failed to detect any lenses, but such are easily seen on material of O. Ilanvirnensis in which the preservation is similar, or in some cases worse. Facial suture runs obliquely across cephalic margin from a nitidus > hi- rundo has been quoted, but D. extensus, sensu stricto, is an unlikely ancestor of nitidus because of differences in the proximal end structure. D. hirundo and D. nitidus are more alike—for example in the arch beneath the proximal end, and in the form of the distal thecae—and it is possible to derive hirundo from nitidus by continued growth of both thecae and stipes. Didymograptus (Expansograptus) sparsus Hopkinson 1875 (Figs 121a, b, 122-3) 1875 Didymograptus sparsus Hopkinson, in Hopkinson & Lapworth: 643; pl. 33, figs 2a—d. 1875 Didymograptus pennatulus Hall; Hopkinson, in Hopkinson & Lapworth: 643-644; pl. 33, fig. 3a-d. 1901 Didymograptus sparsus Hopkinson; Elles & Wood: 17-18, fig. 10; pl. 1, figs 6a, b. 21909 Didymograptus cf. sparsus Hopkinson; Cantrill, in Strahan et al.: 20. LECTOTYPE. Specimen with proximal end, on type slab, SM A16947. Selected Elles & Wood, 1901: explanation of plate 1. TYPE LOCALITY AND HORIZON. Upper Arenig, Fennian, probably Bergamia rushtoni Biozone; Road Uchaf, Ramsey Island. OTHER OCCURRENCES. ?South Wales, Llwyn-crwn, loc. 24; B. rushtoni Biozone (stipe fragments only; may be end variant of D. cf. goldschmidti). Lake District, Outerside; at Arenig/Llanvirn boundary. MATERIAL. On type slab, also SM A16948, at least 5 proximal ends and numerous distal fragments; numerous specimens on slabs SM A16946, A16951, A16958; Fitz Park Museum, Keswick, Harrison Coll., for Outerside specimen. Fig. 120 Arenig Didymograptus (Expansograptus) species. a, Didymograptus (Expansograptus?) uni- formis uniformis Elles & Wood 1901, holotype, late Arenig, Bassenthwaite Sand Beds, Cumbria, x 2, Q8; b, Didymograptus (Expansograptus?) uniformis cf. lepidus Ni 1979, Upper Arenig, Fennian, B. rushtoni Biozone, loc. 23, x 5, Q5180: differs from typical lepidus in its more declined stipes; c, D. (Expansograptus) hirundo, Upper Arenig, Lake District, x 1:5, Fitz Park Museum, Keswick, Harrison Coll.; d, D. (Expansograptus) nitidus (Hall 1865), type loc. of Hall, Lévis, Quebec, x 3, GSC 514e; e, D. (Expansograptus) uniformis lepidus Ni 1979, Upper Arenig, Fennian, B. rushtoni Biozone, loc. 23, x 2, Q5092. 268 R. A. FORTEY & R. M. OWENS Fig. 121 Didymograptus (Expansograptus) sparsus Hopkinson 1875. a, near Arenig/Llanvirn bound- ary, Outerside, Cumbria, large undistorted specimen, x 1, Fitz Park Museum, Keswick, Harrison Coll.; b, Rhoad Uchaf, Ramsey Island, Dyfed, Fennian, probably B. rushtoni Biozone, slab with distorted material, x 3, SM A16951: pointer at centre bottom indicates inferred direction of extension. (Specimen at bottom is that used by Hopkinson (1875: pl. 33, fig. 3a) as illustrative of D. pennatulus; all such specimens are here regarded as tectonically altered D. sparsus). Di1aGnosis. Large Expansograptus, proximal part declined, but only for two or three thecae, thereafter horizontal. Sicula large, stipes exceed 2mm width at th 1 and achieve maximum width of about 2:8 mm (range of variation probably considerable) at about th 10, which then remains constant. Thecae very widely spaced, between 8 and 9 per 10mm in mature part of stipe (th 10-th 15 5-Smm); apertures very flared, and deeply cut back to more than one-third stipe width, t = 1mm, ¢ = 20°-30°. Discussion. The slab including the type specimen has obviously suffered a certain measure of distortion. The type specimen lies at a low oblique angle to the direction of maximum exten- sion, and the stipes have been slightly thinned as a result and the thecal spacing increased. Many of the other stipes on the type slab are close to the direction of extension: the very low thecal spacing recorded by Elles & Wood (1901) of 7 in 10mm seems to be based on such specimens; on others preserved more or less normal to the long axis of the strain ellipsoid spacing is reduced to 10 or 11 in 10mm. We regard almost all the specimens on the type slab as belonging to the single species sparsus, and include also a specimen from the same horizon figured by Hopkinson (1875: pl. 33, fig. 3a) as D. pennatulus Hall (which is one of the tectoni- cally compressed specimens). The only exception (Hopkinson 1875: pl. 33, fig. 3e) is a very broad stipe that may belong to D. hirundo. Fig. 122 Didymograptus (Expansograptus) spar- sus Hopkinson. Proximal end of specimen = shown in Fig. 121a, flattened but otherwise not distorted, to show small sicula ‘tooth’ and low NVA, angle between ventral walls of th 11 and th 17, apex of sicula resorbed, x 5. ARENIG IN SOUTH WALES 269 th 10 20 30 40 50 Fig. 123 Stipe expansion diagrams for D. (Expansograptus) sparsus Hopkinson; a, wide stipe (called pennatulus by Hopkinson) from type locality; b, from holotype, tectonically extended (Elles & Wood 1901: pl. 1, fig. 6a); c, undistorted specimen Fig. 121a herein, lying about midway between these two extremes. Knowledge of D. sparsus in an undistorted condition has derived from a specimen from the Lake District, figured here as Fig. 121a. The specimen has a final stipe width of 2:8 mm, and its stipe expansion diagram lies between the wider and narrower specimens on the type slab (Fig. 123). The distal thecal spacing is between 8 and 9 thecae in 10mm, which is still low for an Expansograptus. It is not possible to be sure exactly how much distortion accounts for the variation in distal stipe width on the type slab. Extreme values are 2:2 mm parallel to extension and just over 4mm normal to it. The type specimen, 2.5mm wide, is nearly parallel to exten- sion, and certainly thinned by tectonism. On the other hand some specimens nearly parallel to extension are 3mm wide, and were originally wider. There was presumably a good deal of variation in final stipe width (as in other extensiforms), embracing the values 2:6 to 3:3mm and maybe more. Taking the extreme values produced by distortion on the type slabs the formula original width = ,/compressed width x extended width gives values close to that of the Outerside specimen. The important specific character is not the stipe width, but the widely-spaced thecae with their flared apertures, and the wide initial stipe width with relatively low stipe expansion (Fig. 123). The species certainly grew very large: the Lake District specimen has one stipe 12 cm long without a termination. The proximal end of the holotype was figured by Elles & Wood; the sicula exceeds 3mm in length, but has possibly been tectonically altered. However, all the type series show a similarly large sicula regardless of distortion type. Not much can be inferred about mode of development from the type series, but the proximal end profile is much like that of D. nitidus (Fig. 118b). The Lake District specimen (Fig. 122) is not much better; the top of the sicula has been resorbed to form a low hump. The free ventral walls of th 1' and th 1? are as long as 1-6mm and enclose a small acute angle. The asymmetrical sicular aperture forms a ‘tooth’ about 0-5 mm long. Stipe width at th 1 about 2mm; smaller values on the type slab are because of distortion. Didymograptus (Expansograptus) patulus (Hall) has similarly flared thecae, but (Cooper & Fortey 1982: fig. 44) narrower proximal stipe width and hence a steeper proximal growth gradient, and more closely spaced thecae. D. (E.) sparsus is one of the more distinctive Expan- sograptus, but it is not yet recorded outside the United Kingdom. 270 R. A. FORTEY & R. M. OWENS Didymograptus (Expansograptus?) uniformis uniformis Elles & Wood 1901 (Fig. 120a) 1901 Didymograptus uniformis Elles & Wood: 12-13, text-fig. 6; pl. 1, fig. 4. Hovorype. Well-preserved complete rhabdosome, Q8. The only original specimen. TYPE LOCALITY AND HORIZON. The holotype bears a label recording its occurrence in the Bassenthwaite Sand Beds; this is a late Arenig (probably gibberulus to hirundo Biozone) horizon. DiaGnosis. Deflexed Expansograptus? with distal stipes almost horizontal; deflexed portion of thabdosome protrudes about 6mm above distal stipes, and involves some 10-12 thecae on either side of the sicula. Development not known, acute angle between ventral walls of th 1! and th 1*. Gradual increase in stipe width from 0-8 mm at th 1 to 1-1 mm at th 10; subsequent increase extremely gentle, to a maximum of 1-4mm. Distal stipes appear nearly uniform, with 11-12 thecae in 10mm (th 10-th 15 4-5 mm), 0 about 30°, t = 0-Smm, ¢ = 60°. Discussion. The type specimen of uniformis is the only one available, and our concept of the form must perforce be based upon it. Fortunately, it is well preserved. It is, however, difficult to be certain that there is no distortion, for example, with regard to the rather acutely declined proximal end. The fact that the two stipes are identical suggests that distortion is not serious, and a fragmentary stipe lying at right angles to the holotype has similar thecal proportions, although without a proximal end there is no way of being sure that it belongs to uniformis. We also have specimens with similarly declined proximal ends from south Wales referred to the subspecies lepidus (below). There is little to add to Elles & Wood’s description apart from a photograph of the type specimen. The sicula is not clearly shown and does not exceed 1:5 mm, but Elles & Wood’s text-fig. 6 shows 1-6mm and it is conceivable that the specimen was more complete when first drawn. There is an apparent contradiction between their diagnosis and description, the former stating that the maximum stipe width is 1-6mm, the latter that it is 1-3mm. The maximum width is in fact 1-45mm measured to the tip of a small denticle visible on some distal thecae. There is, of course, no way of knowing what intraspecific variation there may have been within D. uniformis uniformis; for example, it is not likely that the very slight declination of the distal stipes is of specific significance, and if it is like other dichograptids the distal stipe width would be somewhat variable. The closest species is D. nitidus (Hall), redescribed herein, from which it differs in having about twice as many thecae involved in the prominent declined part of the rhabdosome, thinner distal stipes (not regarded as important), and a slow rate of stipe expansion that soon levels off. Elles & Wood also described D. cf. uniformis (1901: pl. 1, fig. 3) from a slab from Raulnay in the Lake District. This slab includes some 8 specimens with proximal ends, showing structure like that of D. nitidus, but similar to uniformis in stipe width; the general habit, such as the projection of the declined part above the distal stipes, is like D. nitidus. The population is apparently an intermediate between uniformis and nitidus, but the development is like the latter. Populations from the Pontyfenni Formation in south Wales differ from uniformis uniformis consistently in a single character only, and these are described below as a different subspecies. As mentioned below, both may be related to Llanvirn species attributed to Corymbograptus rather than to Expansograptus. The straight- sided inverted V of the proximal end distinguishes the uniformis group from the early Arenig Corymbograptus of v-fractus type. Didymograptus (Expansograptus?) uniformis lepidus Ni 1979 (Figs 120e, 124a—c, 125) 1906 Didymograptus nitidus (Hall); Evans: 611. 1906 Didymograptus extensus (?); Evans: 612. 1909 Didymograptus cf. uniformis Elles & Wood; Cantrill, in Strahan et al.: 12, 20. 1951 Didymograptus simulans Elles & Wood; Gigout: 277-278, text-fig. 57. 1979 Didymograptus lepidus Ni, in Mu et al.: 99; pl. 34, figs 16, 17; pl. 35, figs 1—S. ARENIG IN SOUTH WALES 271 TYPE MATERIAL. In Nanjing Institute of Palaeontology. Late Arenig of south-west China, Didy- mograptus nexus Biozone (equivalent to ‘Glyptograptus’ sinodentatus Biozone). OCCURRENCE IN SOUTH WALES. Bergamia rushtoni Biozone; loc. 21, Castell-y-waun; loc. 23, Pontyfenni; and Survey locality Carm. 37SE W/4, Nant-yr-allwyn. MATERIAL. NMW 84.12G.5-7, 84.12G.18, 84.12G.21, 84.12G.36, 84.17G.96-100; Q5092-103, Q5182-5. D1AGNosis. Subspecies of D. uniformis differing from the type in its closely spaced thecae, 14-16 (217) in 10mm in mature parts of stipe (th 10-15: 3-0 to 3-6mm). Proximal end declined, but proportions variable, involving 6-15 thecae to either side of sicula; distal stipes very gently declined or horizontal. Sicula 1-2-1-4mm long, development uncertain; acute angle between ventral walls of th 1’ and th 1*. Gentle increase in stipe width through first 20 thecae or so, distally uniform, but final stipe width varies between 1-20mm and 1-6mm on different speci- mens. On distal stipes 6 varies between 30° and 55°, @ between 45° and 60°. DiscussION. The very close thecal spacing is exceptional in an extensiform, as noted by Ni (in Mu et al. 1979: 99). Ni also provided data on stipe width at various thecal numbers which is consistent with growth of our material (Fig. 125). Ni does not mention a stipe width greater than 1 mm, but her pl. 34, fig. 17 is 1-2 mm wide distally and hence within the range of variation of our material. The material figured by Ni has weakly declined proximal portions, projecting only about 2mm above the distal stipes, like our specimen shown in Fig. 120e. However, the Welsh material does include specimens, such as NMW 84.12G.7, with steeply declined proximal ends perhaps 4mm above the distal stipes, and hence much more like the type of uniformis uniformis. The initial angle of declination varies between 100° and 130°, the lower angles being on those specimens more like uniformis uniformis. This is a wider range than that given by Ni, but assuredly within a population from a single locality at Pontyfenni. Given the similarity of the material identified as lepidus with the type of uniformis, for example in the acute angle enclosed by the free ventral walls of th 1’ and th 17, and in view of our lack of knowledge of intraspecific variation of uniformis at the type locality, it is appropriate to regard the Fennian form as a subspecies of uniformis distinguished by its dense thecal spacing, which is a consistent character. D. uniformis lepidus grew quite large: incomplete rhabdosomes with proximal ends show that whole colonies were more than 16cm long. aa b Cc Fig. 124 a-—c, Didymograptus (Expansograptus?) uniformis lepidus Ni 1979. Upper Arenig, Fennian, B. rushtoni Biozone, Pontyfenni Formation, loc. 23; a, stipes partly chloritized, x 2, Q5182; b, proximal end, x 10, from Q5092; c, showing sicula ‘tooth’, x 10, Q5099. d, Didymograptus (Expansograptus ?) simulans Elles & Wood 1901, proximal end for comparison with lepidus, showing how distal part of sicula aligns itself with one stipe; type slab, Skiddaw Slates, Barf, Cumbria, probably D. nitidus Biozone, x 10,SM A17698. Dae R. A. FORTEY & R. M. OWENS th 10 20 30 40 50 60 Fig. 125 Stipe expansion diagrams for population of D. (Expansograptus?) uniformis lepidus Ni 1979 from loc. 23, showing range in distal stipe width. Despite some quite well preserved material the proximal end development is not clear. The sicula is about 1-3 mm long, narrow, with an oblique aperture that forms a very short ‘tooth’ in profile between the stipes (0-2 mm long). Th 1! begins high on the sicula; free ventral walls of th 11 and th 1? are 0:5-O-7mm long, and enclose an acute angle (45°-70° on measured material). The possibility of artus development cannot be excluded. Proximally declined large specimens of this species resemble narrow-stiped versions of such Llanvirn Corymbograptus species as C. retrofiexus, which is reported by Boucéek (1973: 54) to have artus development. The stipe expansion diagrams show a good deal of distal variation in stipe width; as usual in dichograp- tids this is not an important character. Wider stipes are those associated with higher 0 and lower ¢ (narrower stipes vice versa) for the geometrical reasons described by Fortey (1983). It seems possible that the wide specimens are the result of continued thecal growth in older colonies. Of our specimens, 80% have distal stipe widths 1-3—1-5mm and thecal spacing is 15 in 10mm. We have two specimens (one shown in Fig. 120b) which have a longer declined proximal portion of the rhabdosome. These may be no more than population variants: they are named here as D. (Expansograptus?) uniformis cf. lepidus. D. simulans Elles & Wood 1901 is superficially similar to D. uniformis lepidus in growth habit, but has a distinctively different proximal end structure (Fig. 124d). Didymograptus (Expansograptus) goldschmidti Monsen 1937, sensu Kraft 1977 (Figs 126-7) 1937 Didymograptus goldschmidti Monsen: 117-118; pl. 1, figs 22, 39, 45. 1973 Expansograptus extensus (Hall 1865); Bouéek: 37; pl. 6, fig. 5. 1977 Expansograptus goldschmidti (Monsen) Kraft: 15-16; pl. 8, figs 1-3. 21979 Didymograptus alatus Chen; Mu et al.: 107-108; pl. 37, figs 2-4. 21979 Didymograptus patulentis Chen; Mu et al.: 105—106; pl. 36, figs 5, 20, 21; pl. 37, fig. 1. TYPE LOCALITY AND HORIZON. The holotype is K 0666, in the Palaeontological Museum, Oslo. From the biozone of Phyllograptus densus, Lower Didymograptus Shales; Ensjo, Norway. OCCURRENCE IN SOUTH WALES. Rare in the later Whitlandian, biozone of Gymnostomix gibbsii; common in the Fennian, biozones of Stapleyella abyfrons and Bergamia rushtoni. Localities: 24, 25, 38, 40, 48 in Pontyfenni Fm.; 28, 34 in Whitland Abbey Member of Colomendy Fm. MATERIAL. Q5150—2, Q5790-7; NMW 33.189.G204.1—2, 33.189.Gla, 33.189.G128, 33.189.G135, 84.17G.174-179. DIAGNOSIS. See Kraft, 1977: 15. Discussion. A group of deflexed forms is included here, which have proved particularly difficult to determine specifically. These all have a small deflexed proximal portion involving 3—5 thecae tO ~j Ww ARENIG IN SOUTH WALES Fig. 126 Didymograptus (Expansograptus) goldschmidti Monsen 1937, sensu Kraft 1977. a, Upper Arenig, Fennian, B. rushtoni Biozone, loc. 24, thin-stripe morph, x 3, Q5790; b, probably B. rushtoni Biozone, loc. 40, latex cast from relief specimen showing isograptid development in obverse view, x 8, Q5791; c, same horizon, loc. 25, x 5, Q5792; d, S. abyfrons Biozone, loc. 38, x 3, Q5793. to either side, distally nearly horizontal to slightly reclined. Sicula exceeds 2mm in length (2:1-2:-4mm), forming a distinct tooth between th 1! and th 17, the ventral walls of which enclose an acute angle. Thecae are distinctly flared, trumpet-like, with the apertures cutting back to up to half the stipe width. Specimens in relief show that th 1’ originates high on the sicula, and that there was isograptid development. There is a great deal of variation in other characters. The Whitlandian specimens are small (about 3cm across); some Fennian specimens are also small, but others grew to at least three times that length. Some specimens (Fig. 127a) have distal apertural denticles (d) 1mm apart (corresponding to spacing of 9-10 in 10mm); others have thecal spacing of 12-13 in 10mm. Stipe width at th 1 is 1-1-1-Smm;; distal stipe width is usually stable after about the fourth theca, and may be as little as 1-4mm, but other specimens are up to 2.0mm wide; stipe fragments from Pen-y-parc in the early Fennian probably belong here and are up to 2:‘5mm wide. @ prox 20°—30°, distal 45°—60°; @ is acute, 60°-80°. If this is a single species, it is a highly variable one. Those specimens with widely spaced thecae are like D. sparsus, but with narrower stipes. The large sicula and dependent proximal thecae distinguish the species from D. nitidus (Hall). Kraft (1977) figured some specimens from the Arenig of Bohemia which include some examples identical to our material except for a R. A. FORTEY & R. M. OWENS 274 ZSTSO ‘€ x ‘sadns pourpoos APYBYs YIM ‘9 ‘1S 1SO “E x ‘WLOJ podiys-r9x91Y “q “8ZTO'6ST'EE MINN “€ X “(OYsLI ye poyeounsy adijs ‘wu 7g) uourtoads aR] ‘e {PZ ‘9O] ‘QuUOZOIg !uoJYsn4 ‘g ‘URIUUD4 “BTUZTY roddQ “LL61 ery nsuas ‘LE6] UssuO|W Upuuyospjob (snidvabosupdx 7) snidvibowdpiq LZ ‘st q 2) POTN EIEN ow 57 ARENIG IN SOUTH WALES DAS slightly less declined proximal part—which we know to be somewhat variable in Expanso- graptus. Kraft referred these to D. goldschmidti Monsen, and this identification is followed here. However, it is noted that Monsen (1937: 117) did not describe such acute thecal apertures, and she quotes a sicula length as long as 3mm. For this reason the identification of D. goldschmidti is qualified, until a revision of the type material is available. A broader specimen was figured by Kraft as Expansograptus cf. cinnereus Monsen. Kraft’s description of goldschmidti quotes 9-10 thecae per 10mm, but if it is correct to assign specimens such as our Fig. 126a to the same species the total range would be 9-13. Some of the variation in this species may be accounted for by continued growth of the rhabdosome, although this ought not to apply to thecal spacing. Some small specimens are also narrower; continued growth at the distal end of the stipes may have been accompanied also by small growth increments on proximal thecae. Large rhabdosomes clearly show a decline in stipe width at the distal, growing tip (Fig. 127a). However, since we have only found the larger specimens in the Fennian it is possible that there was also a phylogenetic change in the direction of large colonies, from small in the Whitlandian to large in the Fennian. Some of the species from the later Arenig of China described in Mu et al. (1979) are probably conspecific with the present form, especially those referred to Didymograptus patulentis Chen (Mu et al. 1979: pl. 36, fig. 20; pl. 37, fig. 1) and D. alatus Chen. The former compares with the larger specimens from Llwyn-crwn, for example. There are nomenclatorial problems here also. D. patulentis was proposed to distinguish some late Arenig forms from D. patulus (Hall), under which name they had been described from Sweden by Tornquist (1901). Cooper & Fortey (1982) suggested that Tornquist’s material might be referred to Xiphograptus, on the basis that Tornquist described a virgellar spine on one of his specimens. Examination of Tornquist’s types in Lund University in 1985 failed to substantiate the presence of this spine, and it does not seem to be present on the Chinese material illustrated by Mu et al. (1979). D. alatus Chen (in Mu et al. 1979: pl. 37, figs 2-4) has slightly reclined stipes, like the south Wales specimen shown in Fig. 126a. Clearly the group of didymograptids having a small declined proximal part, large sicula, dependent th 1’ and th 17, and flared thecae, requires comparative nomenclatorial revision, starting with the earliest named taxa. For this reason we employ Monsen’s name here, while acknowledging that identical forms to the south Wales specimens occur in Bohemia and China. Genus AZYGOGRAPTUS Nicholson & Lapworth, in Nicholson 1875 TYPE SPECIES. A. lapworthi Nicholson 1875, by monotypy. DiaGnosis. See Bulman 1970: V116. Azygograptus hicksii (Hopkinson 1875) (Figs 128b, c) 1875 Tetragraptus Hicksii Hopkinson, in Hopkinson & Lapworth: 651; pl. 33, figs 12a—d. 1902 Azygograptus Hicksii (Hopkinson) Elles & Wood: 94-95, text-fig. 55; pl. 13, figs 2a—c. Lectotype. SM A17831. Selected from Hopkinson’s syntypes by Elles & Wood (plate ex- planation) as ‘type’; original of Hopkinson, in Hopkinson & Lapworth 1875: pl. 33, fig. 12c, d. TYPE LOCALITY AND HORIZON. Penmaen Dewi Formation, Pwlluog, Whitesand Bay, near St David’s, Dyfed; Whitlandian, Gymnostomix gibbsii Biozone. OTHER OCCURRENCE. Afon Ffinnant Formation, on Afon Ffinnant, east of Carmarthen, isolated shale outcrop between locs 18C and 18D; Whitlandian, ?G. gibbsii Biozone. MATERIAL. SM A17379-94; Q5172-3. DIAGNOSIS. Azygograptus with more widely spaced thecae than any other species of the genus; distance between thecal apertures d = 1:6-2:8mm. Stipe originates near base of sicula, which carries prominent spine on opposite side. Distal stipe width 0-9-1-5 mm. 276 R. A. FORTEY & R. M. OWENS Fig. 128 Arenig Azygograptus species, all x 3. a, A. eivionicus Elles 1922, Afon Ffinnant Formation, Middle Arenig, Whitlandian, ?F. radix Biozone, loc. 18E, Q5175; b, A. hicksii (Hopkinson 1875), Penmaen Dewi Formation, Middle Arenig, Whitlandian, C. gibbsii Biozone, Pwlluog, north of Whitesand Bay, St David’s, Dyfed, rhabdosome showing spine on sicula, SM A17393; c, A. hicksii (Hopkinson), same horizon and locality, SM A17392. Discussion. Elles & Wood (1902) gave a good general description of this species, which is known from a type population of more than 20 specimens. It has wide stipes with very spaced thecae, and this alone distinguishes it from other Azygograptus. This feature does not seem likely to be the result of distortion; other species of trilobites and graptolites from near the type locality are not distorted, and the species occurs also in the Carmarthen district where there is a different tectonic setting. Sicula 1-5—2:1 mm long, th 1 probably originating quite low down on sicula and curving away from it at once, 0:2-0:-3mm above sicula aperture. Well-preserved proximal ends, such as on SM A17392-3 (two specimens; Fig. 128b, c) show a prominent spine curving away from the sicula on the opposite side to the stipe; this is regarded as a likely specific character, although it is not visible on poorly preserved specimens (or ones, like the type, which have been covered in Canada balsam). Width at th 1, 0-6—0-9 mm; attaining mature stipe width by th 3 to th 9 with little change therafter. The majority of specimens in the type population (Fig. 130) are 1:2-1-4mm wide. Overall rhabdosome shape varies from nearly straight to quite strongly curved, and no taxonomic importance is attached to this. A. hicksii is known from the later Whitlandian. A. eivionicus from the early Whitlandian, with thinner stipes and more crowded thecae, is distinguished below. A population attributed to hicksii from Afon Ffinnant, which is probably stratigraphically between the two, has interme- diate characters (Fig. 130) with regard to stipe width and sicula size. It is likely that there is a temporal gradation between the two species. These Ffinnant specimens are referred to A. hicksii because they have the very loose thecal spacing noted in the type population. Azygograptus eivionicus Elles 1922 (Figs 128a, 129) 1915 Azygograptus lapworthi; Nicholas: 113. 1922 Azygograptus eivionicus Elles: 299-301. 21979 Azygograptus eivionicus Elles; Mu et al.: 110; pl. 38, figs 11, 12. Hovotype. Original of Elles 1922: fig. 1, which is referred to by her as ‘type’ on the figure explanation; SM A17372. Strachan (1971: 19) referred to Elles’ originals as ‘syntypes’. ARENIG IN SOUTH WALES Didi : 5 Sent ee ae Fig. 129 Azygograptus eivionicus Elles 1922. Middle Arenig, Whitlandian, ?F. radix Biozone, Afon Ffinnant Formation, loc. 18E, pyritized material; a, b, small specimen with well-preserved proximal end, x 10 in reflected light, and sicula (under alcohol) showing sicula spine on anti-stipe side, x 20, Q5174; c, fully grown rhabdosome, x 5, Q5176. TYPE LOCALITY AND HORIZON. Nant, south of Llanengan, Llyn Peninsula, north Wales; before late Arenig, but exact horizon unknown. OCCURRENCE IN SOUTH WALES. Afon Ffinnant Formation, abundant at locs 18D and 18E, middle Arenig, Whitlandian, F. radix Biozone. Also Blaencediw Formation at loc. 31, Blaen- cediw. MATERIAL. Q5174-9; NMW 84.17G.105-—7. DIAGNOSIS. Azygograptus with stipes almost straight to strongly curved. Sicula 1-1—-1-3 mm long with slender apertural process. Stipe originates at 0-2mm from sicula aperture, and makes an acute angle with it. Stipe width at th 1 0-S—0-6mm; distal stipe width 0-6—0-9mm with a mean between 0-7 and 0-8 mm; distance between thecal apertures d = 1-3-1:5mm. Discussion. The salient features are given in the diagnosis. As noted above there is a gradation between this form and A. hicksii, which has wider stipes (Fig. 128) and more spaced thecae: it is like an A. eivionicus in which the thecae have become larger. Like A. hicksii, well-preserved eivionicus show a delicate sicula spine, which curves outwards almost at right angles to the long axis of the sicula. A. lapworthi Nicholson 1875 is best distinguished by the comparatively high origin of the stipe on the sicula, with which it subtends a right angle. A. suecicus Moberg 1892 from the late Arenig is altogether thinner and more gracile. Specimens from the Lake District referred to this species by Elles & Wood are probably more correctly assigned to A. eivionicus. hicksii Fig. 130 Histograms of mature stipe width, w, for Whitlandian Azygograptus species; A. hicksii populations do not overlap with those of A. w eivionicus, but a transitional population ‘trans’ mm has been found on Afon Ffinnant. 278 R. A. FORTEY & R. M. OWENS There they occur with a D. nitidus Biozone assemblage. Supposed A. suecicus from China described by Mu et al. (1979) also appear to compare more closely with A. eivionicus, which they record from the same horizon. There is a good deal of variation in the curvature of the stipes in the south Wales population: some specimens are hardly curved, others turn through a right angle. The species’ occurrence in the Afon Ffinnant Formation is as a monospecific ‘graveyard’. It is the first graptoloid to appear in the succession there, and may have been capable of living shorewards from other graptoloid species. Genus PSEUDO TRIGONOGRAPTUS Mu & Lee 1958 TYPE SPECIES. Graptolithus ensiformis Hall 1865; see Cooper & Fortey 1982: 247. Pseudotrigonograptus ensiformis (Hall 1865) (Fig. 135a) (For synonymy see Rickards (1973) and additional comments in Cooper & Fortey (1982)). TYPE LOCALITY AND HORIZON. Lévis, Quebec, from latest Arenig. OCCURRENCE IN SOUTH WALES. Fennian, Bergamia rushtoni Biozone, Pontyfenni Formation; loc. 23, Pontyfenni. MATERIAL. Q5104; NMW 84.17G.108-9. DISCUSSION. This species was given a full description by Rickards (1973) and by Cooper & Fortey (1982), and further description is not necessary. The specimens from the Fennian are relatively well preserved, but in the usual mode for this genus, in which the apertures are not visible; the specimen breaks along the septum to give a straight-sided appearance (Fortey 1971). The specimen illustrated is the most complete we have discovered, exceeding 45cm in length, but other fragments are from even larger individuals. The illustrated specimen is the ‘narrow form associated with the shorter diameter of the more or less rectangular cross section of the rhabdosome (Cooper & Fortey 1982: text-fig. 53d); in these the cross-sectional width is slightly less than 3mm. Other specimens show the ‘wide’ section, with a transverse width of up to 4mm. The fact that these specimens break with a 180° angle between opposite thecal series shows that the form of Pseudotrigonograptus here is the quadriserial scandent type, rather than the triserial type described from relief material from Spitsbergen by Fortey (1971), in which the angle between adjacent thecal series is 120°. Suggestions that the triserial and quadriserial forms might be placed in different genera (e.g. Cooper 1979) do not seem well advised, as virtually all other characters are closely similar, including the unique stipe construction. P. ensiformis is an almost ubiquitous species in more ‘oceanic’ graptoloid biofacies and is one of the few distinctive species to span the so-called Atlantic and Pacific provinces in the late Arenig. It is present in black shale facies, but not cratonic graptolite facies, in North America, as in Texas, New York, British Columbia, Newfoundland and Quebec. It is distributed in Taimyr, parts of China, Australia (Victoria) and New Zealand. Its occurrence also in south Wales and the Lake District is therefore of considerable importance for correlation. The first appearance of the quadriserial form is in the later Arenig. Subfamily SIGMAGRAPTINAE Cooper & Fortey 1982 Genus ACROGRAPTUS Tsai 1969 TYPE SPECIES. Didymograptus affinis Nicholson 1869, by original designation. Acrograptus acutidens (Elles & Wood 1901) (Figs 131-133) 1875 Didymograptus affinis Nicholson; Hopkinson, in Hopkinson & Lapworth: 645; pl. 33, figs 6b, c (?6a). 1901 Didymograptus acutidens Lapworth MS; Elles & Wood: 25-26, text-fig. 15a—c; pl. 2, figs 3a—d. ARENIG IN SOUTH WALES 279 non 1904 Didymograptus acutidens Lapworth; Ruedemann: 683-684; pl. 13, fig. 15. 1909 Didymograptus acutidens Lapw.; Cantrill, in Strahan et al.: 30 (listed). 21931 Didymograptus acutidens Lapworth MS, Elles & Wood; Bulman: 30-31; pl. 2, fig. 13. 1934 Didymograptus acutidens Lapworth MS em. Elles & Wood; Hsu: 33; pl. 2, fig. 3. non 1947 Didymograptus acutidens Lapworth; Ruedemann: 324; pl. 55, fig. 1; pl. 56, fig. 18. 1971 Didymograptus acutidens Elles & Wood; Strachan: 14. LECTOTYPE (selected Strachan 1971: 14). SM A16985; original of Elles & Wood 1901: pl. 2, fig. 3a. TYPE LOCALITY AND HORIZON. Porth Hayog (= Porth Llauog), Ramsey Island; Llanvirn. 8 OCCURRENCE IN SOUTH WALES. A. acutidens is abundant in the Llanfallteg Formation, appear- ing in the uppermost Arenig (Dionide levigena Biozone) and continuing into the early Llanvirn. We have recovered it from all our Llanfallteg Formation localities. D1aGnosis. Slender, gently declined (130°-150°) Acrograptus, with stipes that increase very gradually from 0-4-0:6mm to a maximum of 1:5—2:0mm. Thecal shape characteristic, with acute @, and apertural margin usually making an angle of slightly less than 90° to dorsal stipe margin to give acutely toothed appearance. Distal thecal spacing variable with 10-13 (?14) thecae per 10mm; th 10-15 4:3-4-5mm; t = 0-5mm, 6 = 30° or less. MATERIAL. Q5156-60, Q5186; NMW 84.17G.110—-115. DIscussION. The species has a slender sicula 1:3—1-5mm long, and very thin proximal parts of both stipes (Fig. 131b) which originate slightly asymmetrically to either side of the sicula; this suggests that acutidens is correctly referred to Acrograptus. Distal stipes abound in the Llanfall- teg Formation, but proximal ends are rare. From the slow rate of distal stipe width increase it is likely that the species had stipes at least 15cm long to attain maximum width. Elles & Wood (1901) gave a good general description of the species, to which a little can be added. All the Fig. 131 Acrograptus acutidens (Elles & Wood 1901). Llanfallteg Formation, both x 3; a, two distal stipes, early Llanvirn, D. artus Biozone, loc. 50, Q5186; b, well-preserved proximal end showing slow increase in stipe width, same locality, Q5156. 280 R. A. FORTEY & R. M. OWENS —> continues th 10 20 30 Fig. 132 Stipe expansion diagram for best-preserved Acrograptus acutidens (Elles & Wood). material we have found indicates that a declined proximal part is characteristic of the species, and we would exclude such specimens as Ruedemann’s (1947: pl. 55, fig. 1) from Deepkill with immediately horizontal stipes; these are from an earlier Arenig horizon, probably equivalent to the D. nitidus Biozone of Britain. Many distal stipes are flexed or bent, and it is probable that they were somewhat flexible. None of the proximal ends are well enough preserved for the development to be reliably discovered; a projecting portion of the sicula is characteristically sigmagraptine. The stipe expansion diagram (Fig. 132) is a gentle curve from a narrow origin; the stipes continue to expand in width distally, but at a very low rate, about 0-1 mm per 15-20 thecae. Proximal thecae are about 1mm long, and nearly rectimarginate, with a short lip. Distal thecae have a distinctive appearance, as noted by Elles & Wood: Q5160 is in relief, and shows some details of this thecal structure. The common canal is 0-4mm wide; thecae are as long as 1-7mm, and inclined at a low angle of 20° proximally, and curving very gently to 30° at the aperture. The apertural margin is cut back steeply, often so that the angle between it and the dorsal wall of the stipe is slightly less than 90°; @ is therefore acute, as low as 30°. There is an apertural denticle. The stipe appearance then is characteristically sharp-toothed, as implied by the specific name. Note that other species can acquire this appearance with tectonic short- ening, but there is no question of this with the material from the Whitland area, which is often well preserved. Distal thecal spacing appears to be very variable, usually 12-13 thecae in 10mm, but examples of distal stipes have been found with spacing as low as 10 or as high as 14 per 10mm. This is probably a consequence of small variations in 6 at low inclinations, since thecal spacing d = t sin 0 (Fortey 1983), and for constant t and low @ any variation in sin @ will produce a relatively large variation in d. A. acutidens appears below the Llanvirn boundary, but is abundant in the early Llanvirn also. Cantrill & Thomas (in Strahan et al. 1909: 25) record the species in what they believed to be the top of the D. ‘bifidus’ shales. It is one of the few species distinctive enough to be recognized from distal stipe fragments. Fae Fig. 133. Proximal end of Acrograptus acutidens, x 10, based on Q5156. Acrograptus ? sp. a of Skevington, 1965 (Fig. 134) 1965 Didymograptus n. sp. a aff. D. gracilis Tornquist, 1890; Skevington: 21-22, fig. 24 (gives earlier synonymy). 1965 Didymograptus cf. n. sp. a aff. D. gracilis Tornquist; Skevington: 22. ORIGINAL LOCALITY AND HORIZON. Halludden Borehole, Oland; at, or immediately below, Arenig/Llanvirn boundary (Skevington 1965: fig. 73). ARENIG IN SOUTH WALES 281 Fig. 134 Acrograptus? sp. a of Skevington, 1965. Llanfallteg Formation, latest Arenig, D. levigena Biozone, loc 52H, x 5, Q5172. OCCURRENCE IN SOUTH WALES. Fennian, Dionide levigena Biozone, Llanfallteg Formation; loc. 52W. MATERIAL. Q5172. Discussion. Skevington (1965) named his ‘nov. sp. a’ for Holm’s ‘mutation’ of Didymograptus gracilis Tornquist. He noted that true gracilis had an asymmetrical proximal end, which we would now regard as sigmagraptine (Cooper & Fortey 1982: fig. 66f). In ‘sp. a’ the origin of the thecae is almost symmetrical to either side of the sicula, the apertural lip of which projects as a minute tooth. From the Llanfallteg Formation—and from the same stratigraphical position as Skevington’s examples—we have a gracile specimen with exactly the same proximal end. This is more complete than was the material from Oland, from which it differs only in having the sicula 0-1 mm longer, and the stipe width also being 0:1 mm wider. The sicula is 0:9mm long and 0:25mm wide at its base. The stipes are declined at about 140° initially but become more horizontal after th 2. Thecae have the very low inclination invariable in gracile species, @ is a high acute angle and there appears to be a minute apertural denticle. Distance between thecal apertures varies between 0-8 and 1:0mm (equivalent to 10-12 thecae in 10mm). Stipe width at th 1 is 0-3 mm, increasing to no more than 0-4 mm at th 17. Extremely thin species such as this are easy to overlook. However, the difference in proximal end structure from other Acrograptus spp. indicates that this A.? sp. a is likely to prove a distinctive late Arenig form. Because of the difference from the typical sigmagraptine, it is included with question in Acrograptus. None of the Acrograptus species described by Bouéek (1973) from Bohemia is as thin—the closest is apparently A. lipoldi, but this is reported to be 1-0mm wide at 15mm from the sicula, where our specimen is still less than 0-4mm. It is tempting to name it as a new species but the scarcity of material makes this inadvisable at the moment. Accordingly, the open nomenclature coined by Skevington (1965) is used again here. Family GLOSSOGRAPTIDAE Lapworth 1873 Genus GLOSSOGRAPTUS Emmons 1855 TYPE SPECIES. G. ciliatus Emmons, by monotypy (Emmons 1855: 108). D1AGNnosIs. See Bulman 1970: 122. Glossograptus acanthus Elles & Wood 1908 (Fig. 135b) 1908 Glossograptus acanthus Elles & Wood: 314, figs 208a, b; pl. 33, figs 4a—c. 1935 Glossograptus acanthus Elles & Wood; Harris & Thomas: 302-303, fig. 3 (13-16). 1960 Glossograptus acanthus Elles & Wood; Berry: 70-71. 1960 Glossograptus acanthus Elles & Wood; Turner: 89; pl. 7, fig. 8. 1964 Glossograptus acanthus Elles & Wood; Obut & Sobolevskya: 71; pl. 15, figs 1, 2. 1971 Glossograptus acanthus Elles & Wood; Skevington & Archer: 76 (listed). 1974 Glossograptus acanthus Elles & Wood; Tsai: 105, text-fig. 38; pl. 11, figs 25a, b. 1979 Glossograptus acanthus Elles & Wood; Cooper: 81, text-fig. 65; pl. 15, fig. k. 282 R. A. FORTEY & R. M. OWENS HotoryPe. According to Strachan (1971: 27), Elles & Wood (1908) indicated their pl. 33, fig. 4a as type specimen, SM A17441. TYPE LOCALITY AND HORIZON. ‘Sruffaunduff, 4 mile W of Summit of Bencraff, Connemara’, Republic of Ireland. Originally described by Elles & Wood as Arenig, the fauna from here has been re-evaluated (Dewey, Rickards & Skevington 1970: 30-31) as of early Llanvirn age. DIAGNOsIS. See Elles & Wood, 1908: 314. OCCURRENCE IN SOUTH WALES. Elles & Wood recorded the species from the early Llanvirn of St David’s. Our new occurrence is from 15m below the Arenig/Llanvirn boundary in the Llanfall- teg section, latest Arenig, Dionide levigena Biozone. MATERIAL. Q5171. Discussion. The species is represented in our collections by a well-preserved specimen from the latest Arenig. The specimen retains sufficient relief to see the monopleural arrangement of the thecal series, showing that it is a typical Glossograptus. There is nothing to add to the recent descriptions given in the synonymy. Our specimen has a maximum diameter of 6mm; some seven pairs of apertural spines up to 2mm long curve downwards in the proximal part; lengthwise distance between apertural spines in the mature part of the rhabdosome is 0-8— 1:0mm. The species is of considerable stratigraphical importance, being very widespread: Australia, New Zealand, Texas, Ireland, Kazakhstan, Taimyr, South America and Wales. It appears to be confined to the uppermost Arenig and early Llanvirn. Family DIPLOGRAPTIDAE Lapworth 1873 Genus GLYPTOGRAPTUS Lapworth 1873 TYPE SPECIES. Diplograptus tamariscus Nicholson 1868; original designation of Lapworth (1873: table of classification). Discussion. The concept of Glyptograptus has to be based on the type species, G. tamariscus. Glyptograptus has been defined on the basis of its thecal form, as have other biserials; modern work would emphasize instead the mode of development at the proximal end. On this criterion it is doubtful whether either of the species described here would be referred to Glyptograptus s.s. ‘G. dentatus has steeply upward curved proximal thecae, in this respect resembling the type species of Undulograptus Boucek (U. paradoxus, well illustrated from relief material by Boucek, 1973). ‘G. austrodentatus, on the other hand, has proximal thecae with rather low inclination, and probably a different proximal structure. It may not be congeneric with ‘G. dentatus. Such generic reassignments are now being carried out in the context of a revision of these genera as a whole (C. E. Mitchell, personal communication 1985), using isolated or relief material. Here the usual name of Glyptograptus is used in quotations to indicate the provisional classification. ‘Glyptograptus’ dentatus (Brongniart 1828) (Figs 135d, e, g, j, 136) (For synonymy see Bulman (1963: 673) and Skevington (1965: 55)). NEOTYPE AND HORIZON. Geological Survey of Canada, 943, proposed Bulman 1963: 672, 675, text-fig. 4a; from what is probably the early Llanvirn (?P. tentaculatus Biozone), Point Lévis, Quebec. OCCURRENCE IN SOUTH WALES. Widely recorded in the Didymograptus artus (D. ‘bifidus’, auctt.) Biozone of the early Llanvirn, in which it is generally distributed. Here we add the common occurrence of the species in the uppermost Arenig biozone of Dionide levigena (throughout loc. 52) in the Llanfallteg Formation, and rare occurrence in the Fennian, biozone of Bergamia rushtoni; type locality of Pontyfenni Formation, loc. 23. The species ranges through the top two biozones of the Arenig and into the early Llanvirn. MATERIAL. Q5105, Q5808-15. ARENIG IN SOUTH WALES 283 Fig. 135 Scandent graptoloids from the late Arenig to earliest Llanvirn of south Wales. a, Pseudotri- gonograptus ensiformis (Hall 1865), Upper Arenig, Fennian, B. rushtoni Biozone, Pontyfenni Forma- tion, loc. 23, x 3, Q5104. b, Glossograptus acanthus Elles & Wood 1908, latest Arenig, Fennian, D. levigena Biozone, Llanfallteg Formation, loc. 52J, x 2, Q5171. c, f, h, k, ‘Glyptograptus’ austro- dentatus Harris & Keble 1932, Llanfallteg Formation; c, large specimen widening slowly to more than 2mm, late Arenig, D. levigena Biozone, loc. 52S, x 6, Q5802; f, well-preserved rhabdosome, early Llanvirn, artus Biozone, loc. 50, x 10, Q5803; h, detail of proximal part of wide morph in relief, showing median septum and th 1', late Arenig, D. levigena Biozone, loc. 52L, x 8, Q5804; k, well-preserved form, slowly increasing in width, locality and horizon as last, x 5, Q5805. d, e, g, j, ‘Glyptograptus’ dentatus (Brongniart 1828); d, narrow morph from early horizon, Upper Arenig, Fennian, B. rushtoni Biozone, loc. 23, x 6, Q5808; e, wide morph from early horizon as Fig. 135d, x 5, Q5809; g, early Llanvirn, D. artus Biozone, loc. 50, x 6, Q5810; j, D. levigena Biozone, loc. 52P, < S,O58iil- 284 R. A. FORTEY & R. M. OWENS 136 i Fig. 136 ‘Glyptograptus’ dentatus (Brongniart 1828). Well-preserved small rhabdosome, x 5, and detail of its proximal end, about x 20, early Llanvirn, D. artus Biozone, loc. 50, Llanfallteg Formation, Q5810. Fig. 137 ‘Glyptograptus’ austrodentatus Harris & Keble 1932. Typical specimen, Llanfallteg Forma- tion, early Llanvirn, D. artus Biozone, loc. 50, x 10, Q5803. DIAGNOsIS. See Bulman, 1963: 673-675. Discussion. Bulman (1963) revised G. dentatus, and further description of isolated material was given by Skevington (1965). Elles & Wood (1907: fig. 174; pl. 31, figs 4b-d) have already illustrated material from the Llanvirn part of the Llanfallteg Formation (misspelled ‘Llanfanteg’ on their p. 254). This is often well preserved, in full or partial relief. The species is distinguished from G. austrodentatus particularly by the sharp upward growth of th 1! and th 17, with a stout virgellar spine. Specimens conforming to Bulman’s sensu stricto usage of dentatus are numerous in the latest Arenig and earliest Llanvirn. The earliest biserials in Wales, and probably as old as anywhere, are from the Bergamia rushtoni Biozone. These include (Fig. 135e) a specimen which has a maximum width of 2:5mm, and which also tapers distally. Bulman cites 2-2mm as a maximum for G. dentatus. Tectonic ‘extension’ is not usually evident in the Pontyfenni Forma- tion and it is believed that the wide stipe was originally so. The same horizon has also yielded a narrow specimen which widens rather rapidly to 1-6-1-7mm (Fig. 135d), a width which is retained throughout the rest of the rhabdosome. This form is intermediate between typical dentatus and what Bulman (1963) called G. shelvensis (with distal width 1-3-1-5mm) from the Shelve Church Beds, Shropshire. We have other specimens (Fig. 135j) from beds as high as Llanvirn which are only slightly wider than shelvensis, and it seems possible that shelvensis may not prove distinct from dentatus. Here a broad view of the latter species is taken to include both wide and narrow morphs. In any case the choice of G. dentatus as a zonal fossil for the early Llanvirn (cf. Jenkins, in Hughes et al. 1982: 53) is inappropriate because it extends so far below the usual usage of that interval. ‘Glyptograptus austrodentatus Harris & Keble 1932 (Figs 135c, f, h, k, 137) (For synonymy see Bulman (1963: 679) and Mu et al. (1979: 134)). LECTOTYPE (selected Bulman 1963: 679). Geol. Surv. Victoria 31365. TYPE LOCALITY AND HORIZON. Basal Darriwilian (equivalent to early Llanvirn); Victoria, Australia. ARENIG IN SOUTH WALES 285 OCCURRENCE IN SOUTH WALES. Llanfallteg Formation, from latest Arenig (Dionide levigena Biozone) to early Llanvirn, throughout loc. 52; also locs 50, 55. MATERIAL. Q5802-7; NMW 84.17G.116—7. DIAGNOsIs. See Bulman, 1963: 679, and comments below. DISCUSSION. This species is differentiated from G. dentatus primarily by its streptoblastic devel- opment, which results in a conspicuously truncated proximal end, and with the first pair of thecae not growing strongly upwards. This feature is shown well on specimens from south Wales in partial or full relief (Fig. 135h). The virgellar spine is more slender than on G. dentatus. Specimens such as that in Fig. 135f belong within the type population as described by Bulman (1963: text-fig. 6). Bulman described a number of ‘varieties’, which have no formal taxonomic status; Skevington (1965: 56-58) chose to elevate one of these to formal subspecific rank. Mu (in Mu et al. 1979) added one further subspecies and additionally proposed many new species, all from the G. austrodentatus Biozone. Legg (1976) described G. situlus from Western Australia, which also appears to have an austrodentatus-like proximal end. We have specimens from the Llanfallteg Formation (Fig. 135h) which widen rather rapidly to 2.2mm. This exceeds the width allotted to any of Bulman’s ‘varieties’, but is apparently the same as Mu’s species G. robustus (Mu, in Mu et al. 1979: 135; pl. 47, figs 14, 16, 18) and possibly G. austrodentatus major Mu 1979. While it is interesting that such similar specimens occur in both Wales and China I am reluctant to use Mu’s names for such specimens, because there is gradation between them and more ‘typical’ forms. Such wide forms tend to be flattened distally, and flattening too may have contributed to increase the transverse dimension; Mu’s specimens are also clearly flattened. Skevington (1965: 58) mentions the occurrence of broad and narrow morphs of G. austro- dentatus, but attaches no taxonomic importance to this feature. It seems probable that there was a good deal of intraspecific variation in G. austrodentatus, especially with regard to stipe width. Bulman (1963: 680) also noted a wide range of variation in distal thecal spacing, and measuring the distal 3 or 4 thecae and correcting for ‘th in 10mm’ on the Llanfallteg specimens there is a range from 13 to 16 in 10mm, with a mean between 14 and 15. Whatever the final taxonomic status of the several ‘varieties’ of, and co-occurring ‘species’ with, G. austrodentatus it is significant for correlation that in China, Britain, and the Canning Basin, Western Australia (Legg 1978) Glyptograptus of austrodentatus type underlie pendent didymograptids, including D. artus. Acknowledgements We could not have completed this work without the generous help of friends and colleagues in collecting specimens or providing us with information. We thank especially Dr J. C. W. Cope, Mr F. Cross, Mr E. K. Jones, Mr C. T. and Mrs I. Taylor and Dr R. P. S. Jefferies for giving us specimens they found for study. We also acknowledge the help of Dr R. E. Bevins, Dr R. A. Cooper, Dr C. J. Jenkins, Mr R. Kennedy, Dr M. G. Bassett, Mr S. F. Morris and Dr A. W. A. Rushton. Dr Rushton and Dr D. Price helped us greatly in locating specimens in the British Geological Survey and Sedgwick Museum, respec- tively. Dr W. Wimbledon and the Nature Conservancy Council are thanked for their help in exposing the Llanfallteg railway section. We record our debt to the photographic unit of the British Museum (Natural History) and Mrs Kathi Bryant for assistance with photography, and to Mrs Lin Norton for drafting diagrams. Miss Paula Westall, Mrs Edna Richards and Mrs Beryl Chant typed various parts of the manuscript. The chordates—a preliminary note by R. P. S. Jefferies Eight genera, and eight or perhaps nine species, of primitive chordates are now known from the Upper Arenig of the Whitland area. This is remarkable, seeing that none at all was known from the Welsh Lower Ordovician until May 1979 when Dr R. M. Owens found a specimen of Cothurnocystis sp. (Figs 139a, b) in the Upper Arenig Pontyfenni Formation, in the disused 286 R. A. FORTEY & R. M. OWENS Fig. 138 a, b, Mitrocystella sp. a, dorsal and b, ventral aspect. Locality: Pontyfenni. B. rushtoni Biozone, Pontyfenni Formation, Upper Arenig. E63158a, b. Fig. 139 a, b, Cothurnocystis sp. a, dorsal and b, ventral aspect. Note the gill slits as a prominent series of openings on the left side of the head. Locality: Llwyn-crwn. B. rushtoni Biozone, Pontyfenni Formation, Upper Arenig. NMW 84.17G.119a, b. quarry at Llwyn-crwn. The purpose of this note is to record these zoologically important animals, in the hope that further search will discover more of them and thus allow this part of the fauna to be properly described. All the chordates found are primitive forms which retain a calcite skeleton of echinoderm type (‘calcichordates’). General accounts of these animals are given in Jefferies (1981a, 1986). To avoid confusion it must be mentioned that many workers regard them as echinoderms, notably ARENIG IN SOUTH WALES 287 Ubaghs in the Treatise of Invertebrate Paleontology (1967), and sometimes they are referred to as carpoid echinoderms. There has recently been some published controversy on their affinities (Philip 1979, Chauvel 1981, Ubaghs 1981, Jefferies 1981b, Jollie 1982). The calcichordates, in cladistic terms, are a paraphyletic group and so the word ‘calcichord- ate’ should only be used informally. They are co-extensive with the ‘Stylophora’ as used in the Treatise and can be divided into two major kinds—the ‘Cornuta’ and the ‘Mitrata’. Both groupings are likewise paraphyletic and the names are placed in inverted commas for that reason. Satisfactory classification of the calcichordates, however, depends on placing them in the stem-groups of recent groups, as in the list below. The term ‘stem-group’, and also ‘crown- group’, have a precise meaning as explained in Jefferies (1979, 1986). The cornutes are stem-group chordates. The mitrates, on the other hand, are primitive crown-group chordates since all known mitrates are primitive members of the extant chordate subphyla—whether stem-group acraniates, stem-group tunicates or stem-group vertebrates. All calcichordates are divided anatomically into a head and a tail. The tail was almost certainly locomotory (Jefferies 1984) and served to pull the head rearwards across the sea floor, either in the sediment or on its surface according to the species. Cornutes had external gill slits, on the left side of the head only, and were primitively asymmetrical in other respects also. Mitrates are deduced to have had internal gill slits, on left and right sides of the head, and were fairly symmetrical in external outline, though with large asymmetries inside the head. The forms are all found in the Pontyfenni Formation near Whitland and are as follows: (1) Cornutes (a) Cothurnocystis sp. Localities: Pen-y-parc, Llwyn-crwn, S. abyfrons to B. rushtoni Bio- zones. Fig. 139a, b. A new species. It is similar to Cothurnocystis fellinensis Ubaghs (1969) from the Lower Arenig of the Montagne Noire but lacks a median plate (y) dorsal to the tail Insertion. (b) Reticulocarpos sp. 1. Locality: Pen-y-parc, S. abyfrons Biozone. Fig. 141la, b. A new species, not identical with either of the two described forms, i.e. Reticulocarpos hanusi Jefferies & Prokop from the Llanvirn of Bohemia or R. pissotensis Chauvel & Nion from the Llandeilo of Normandy. (c) Reticulocarpos ?sp. 2. Locality: Pontyfenni, B. rushtoni Biozone. Fig. 146. This is much smaller than R. sp. 1 and is either a juvenile of that species or belongs to a different species. (2) Mitrates (a) Lagynocystis sp. Locality: Pontyfenni, B. rushtoni Biozone. Fig. 145. A stem-group acra- niate. Probably a new species since, unlike the only described species L. pyramidalis (Barrande) from the Llanvirn of Bohemia, it seems to lack dorsal spikes in the foretail and has only three spikes on the styloid. (b) Balanocystites sp. Locality: Pontyfenni, B. rushtoni Biozone. Fig. 142a, b. A stem-group tunicate. It is characteristic of the genus that the two enlarged ventral plates (g on the right and j on the left) encroach only a short distance onto the dorsal surface. (c) Guichenocarpos sp. Locality: Pontyfenni, B. rushtoni Biozone. Fig. 144. A stem-group tunicate. This genus resembles Balanocystites but differs in having a tall flange, U-shaped in plan with the U open anteriorly, on the dorsal surface. (d) Anatifopsis sp. Locality: Pontyfenni, B. rushtoni Biozone. Fig. 143. A stem-group tunicate. This genus resembles Balanocystites but the ventral plates g and j extend further onto the dorsal surface. (Anatifopsis Barrande 1868 is identical with Anatiferocystis Chauvel 1941 and is often regarded as a crustacean.) (e) Mitrocystites sp. Locality: Pen-y-parc, S. abyfrons Biozone. Fig. 140. A stem-group verte- brate. Only one specimen is known and it is too badly dissociated to be compared with particular described species. (f) Mitrocystella sp. Locality: Pontyfenni, B. rushtoni Biozone. Fig. 138a, b. A stem-group vertebrate. This is a new species. It differs from Mitrocystella incipiens (Barrande), from the Llandeilo of Brittany and Bohemia, in its small size; and from Mitrocystella barrandei Jaekel, from the Llanvirn of Bohemia, in having cuesta-shaped ribs on the posterior part of the ventral 288 R. A. FORTEY & R. M. OWENS Fig. 140 Mitrocystites sp. Hind-tail ossicles are visible near the left of the picture and head plates elsewhere. Badly dissociated, but mainly ventral in aspect. Locality: Pen-y-parc. S. abyfrons Biozone, Pontyfenni Formation, Upper Arenig. E63146a, b. Fig. 141 a, b, Reticulocarpos sp. 1. a, dorsal and b, ventral aspect. Locality: Pen-y-parc. S. abyfrons Biozone, Pontyfenni Formation, Upper Arenig. E29927. surface instead of being smooth. It differs from both these species in having a concave dorsal surface. A similar chordate fauna is known from other argillaceous occurrences in the earlier part of the Ordovician at the western edge of what is considered to have been a Gondwanan continent. Thus the present fauna is much like that of the Llanvirn Sarka Formation of Bohemia, as exposed at the famous localities of Sarka and Osek. At these localities, as near Whitland, the ARENIG IN SOUTH WALES 289 145 Fig. 142 a, b, Balanocystites sp. a, dorsal and b, ventral aspect; bar = 3mm. Locality: Pontyfenni. B. rushtoni Biozone, Pontyfenni Formation, Upper Arenig. NMW 78.8G.43a, b. Fig. 143 Anatifopsis sp. Dorsal aspect. Locality: Pontyfenni. B. rushtoni Biozone, Pontyfenni Forma- tion, Upper Arenig. E63160. Fig. 144 Guichenocarpos sp. Dorsal aspect. Locality: Pontyfenni. B. rushtoni Biozone, Pontyfenni Formation, Upper Arenig. E63155. Fig. 145 Lagynocystis sp. Dorsal aspect of foretail and posterior part of head. Locality: Pontyfenni. B. rushtoni Biozone, Pontyfenni Formation, Upper Arenig. E63159b. Fig. 146 Reticulocarpos sp. ?2. Dorsal aspect. Locality: Pontyfenni. B. rushtoni Biozone, Pontyfenni Formation, Upper Arenig. E29928. 290 R. A. FORTEY & R. M. OWENS genera Cothurnocystis, Reticulocarpos, Lagynocystis, Balanocystites, Guichenocarpos, Anati- fopsis, Mitrocystites and Mitrocystella occur, though some, or perhaps all, of the species are different. It can also be compared with the Traveusot Formation (Schistes a Calyménes) of the Llandeilo of Brittany in which most of the same genera recur, although Mitrocystites is rare there (Chauvel 1981) and Cothurnocystis and Reticulocarpos are absent. The Lower Arenig of the Montagne Noire also contains calcichordates (Ubaghs 1969, Thoral 1935), though the only genera known to be shared with the Whitland occurrence are Cothurnocystis and Guicheno- carpos (Anatifopsis escandei Thoral 1935 is a Guichenocarpos according to my observations). It is unfortunate that the south Welsh chordates have never been found in siliceous nodules like those which contain the best-preserved specimens in Brittany, Bohemia and the Montagne Noire. As to facies, it seems that advanced cornutes, such as Reticulocarpos, and most mitrates, preferred a soft muddy sea bottom in which sessile benthos was often rather rare. The south Welsh occurrences show that chordates can be found in shaly Ordovician rocks if especially searched for. The paucity of known occurrences world-wide is probably not caused by original absence. Rather is it the case that palaeontologists have not been interested in these animals, despite their zoological significance, and therefore did not know them when they saw them. References Abdullaey, R. N. 1972. In Abdullaev, R. N., Ogienko, L. V. & Semenova, V. S., [New early and middle Ordovician trilobites of Sibera and central Asia.] In Zanina, I. E. (ed.), [New species of fossil plants and invertebrates of USSR]: 237-242, pls 55, 56. Moscow, Nauka. [In Russian]. Angelin, N. P. 1854, 1878. Palaeontologia Scandinavica. 1, Crustacea formationis transitionis. Fasc. 2: IX, 21-92, pls 25-41 (1854). Lund. 2nd edn, Lindstr6m, G. (ed). Fasc. 1 and 2. VIII + 96 pp., 42 pls (1878). Stockholm. Apollonov, M. K. 1974. Ashgill’skie trilobity Kazakhstana. 136 pp., 21 pls. Alma-Ata, Nauka. [In Russian]. — 1975. Ordovician trilobite assemblages of Kazakhstan. Fossils Strata, Oslo, 4: 375-380. Balashova, E. A. 1961. Tremadoc trilobites of the Aktubinsk district. Trudy geol. Inst. Leningr., 18: 102-145, pls 1-4. Barrande, J. 1846. Notice preliminaire sur le systeme Silurien et les trilobites de Boheme. vi + 97 pp. Leipzig. 1847. Uber das Hypostoma und Epistoma, zwei analoge, aber verschiedene organe der Trilobiten. Neues Jb. Miner. Geogn. Geol. Petrefakt., Stuttgart, 1847: 385-399, pl. 8. 1852-72. Systeme Silurien du centre de la Bohéme. lére Partie: Recherches Paléontologiques, 1 (Crustacés: Trilobites). xxx + 935 pp., Atlas 51 pls (1852). 1 supplement (Trilobites, Crustacés divers et Poissons). xxx + 647 pp., 35 pls (1872). Prague & Paris. Barrois, C. 1892. Mémoire sur la distribution des Graptolites en France. Annls Soc. geéol. N., Lille, 20: 75-191. Bassett, M. G. (ed.) 1982. Geological excursions in Dyfed, south-west Wales. 327 pp. Cardiff, Natl Mus. Wales. Bates, D. E. B. 1968a. The Lower Palaeozoic brachiopod and trilobite faunas of Anglesey. Bull. Br. Mus. nat. Hist., London, (Geol.) 16 (4): 127-199, pls 1-14. 1968b. On ‘Dendrocrinus’ cambriensis Hicks, the earliest known crinoid. Palaeontology, London, 11: 406—409, pl. 76. —— 1969. Some early Arenig brachiopods and trilobites from Wales. Bull. Br. Mus. nat. Hist., London, (Geol.) 18 (1): 1-28, pls 1-9. — 1972. The Stratigraphy of the Ordovician rocks of Anglesey. Geol. J., Liverpool, 8: 29-58. Beck, H. H. 1839. In Murchison, R. I., The Silurian System: q.v. Benedetto, J. L. & Malanca, S. 1975. Los trilobitos ordovicicos de Los Colorados (Departamento de Tambaya, Provincia de Jujuy). Actas Congr. argent. Palaeont. Bioestratigr., Tucuman, 1: 149-173, 2 pls. Berry, W. B. N. 1960. Graptolite Faunas of the Marathon Region, West Texas. Publs Bur. econ. Geol. Univ. Tex., Austin, 6005. 129 pp., 20 pls. Bevins, R. E. & Roach, R. A. 1982. Ordovician igneous activity in south-west Dyfed. In Bassett 1982: 65-80 (q.v.). ARENIG IN SOUTH WALES 291 Billings, E. 1861-65. Palaeozoic Fossils, 1. 426 pp., 401 figs. Montreal, Canada geol. Surv. Boeck, C. 1828. Notitser til Laeren om Trilobiterne. Magazin Naturv. Christ., 8: 11—44, 1 pl. Bouéek, B. 1932. Piispévek k poznani éeskych Didymograpti. Cas. narod. Mus., Prague, 106: 119-133. —— 1956. The graptolite and dendroid fauna of the Klabava Beds. Sb. Ustred. Ust. Geol., Prague, 22: 1-105, pls 1-6. — 1973. Lower Ordovician Graptolites of Bohemia. 185 pp., 24 pls. Prague. [In English, Czech summary ]. —— & Pribyl, A. 1953. Taxonomie a kmenovy vyoj nékterych ordovickych graptolitu. Rozpr. Ceske Akad. Ved. Uméeni, Prague, (2) 61 (20): 1-18 (for 1951). 1953a. Taxonomy and phylogeny of some Ordovician graptolites. Bull. int. Acad. Tcheque Sci., Prague, 52 (20): 1-17. Boxshall, G. 1981. Community structure and resource partitioning—the plankton. In Forey, P. L. (ed.), The Evolving Biosphere: 143-156. London & Cambridge. Braithwaite, L. F. 1976. Graptolites from the Lower Ordovician Pogonip Group of Western Utah. Spec. Pap. geol. Soc. Am., New York, 166. 106 pp., 21 pls. Brongniart, A. 1828-38. Histoire des végétaux fossiles, . . . (&c.), 1. xil + 488 pp., 171 pls. Paris. Bruton, D. L. 1968. A revision of the Odontopleuridae (Trilobita) from the Palaeozoic of Bohemia. Skr. norske Vitensk-Akad. mat.-nat. KI., Oslo, (n.s.) 25. 73 pp., 11 pls. & Henry, J. L. 1978. Selenopeltis [Trilobita] from Brittany and its distribution in the Ordovician. Geobios, Lyon, 11: 893-907. Bulman, O. M. B. 1927-67. A Monograph of British dendroid graptolites. Part 1: 1-28, pls 1-2 (1927); Part 2: i-xxxii, 29-64, pls 3-6 (1928); Part 3: xxxiti-Ix, 65-92, pls 7-10 (1932); Part 4: Ixi-Ixiv, 93-97 (1967). Palaeontogr. Soc. (Monogr.), London. —— 1931. South American graptolites with special reference to the Nordenskiold collection. Ark. Zool., Stockholm, 22: 1-111, 12 pls. —— 1936. On the Graptolites prepared by Holm. 7. Ark. Zool., Stockholm, 28A: 1—107, 4 pls. — 1958. The sequence of graptolite faunas. Palaeontology, London, 1: 159-173. —— 1963. On Glyptograptus dentatus (Brongniart) and some allied species. Palaeontology, London, 6: 665-689. — 1970. Graptolithina (2nd edn). In Teichert, C. (ed.), Treatise on Invertebrate Paleontology, V. Xxxil + 163 pp. Lawrence, Kansas. Burmeister, H. 1843. Die Organisation der Trilobiten, aus ihren lebenden Verwandten entwickelt ; nebst einer systematischen Ubersicht aller zeither beschriebenen Arten. 147 pp., 6 pls. Berlin. Callaway, C. 1877. On a new area of Upper Cambrian rocks in South Shropshire, with description of a new fauna. Q. JI geol. Soc. Lond., 33: 652-672, pl. 24. Cantrill, T. C. & Thomas, H. H. 1906. On the igneous and associated sedimentary rocks of Llangynog (Carmarthenshire). Q. JI geol. Soc. Lond., 62: 223-252, pls 23—26. Capera, J. C., Courtessole, R. & Pillet, J. 1978. Contribution a l'étude de l’Ordovicien inférieur de la Montagne Noire. Biostratigraphie et révision des Agnostida. Annls Soc. geol. N., Lille, 98: 67—88, pls 5-7. Carter, R. M. 1971. Revision of Arenig bivalvia from Ramsey Island, Pembrokeshire. Palaeontology, London, 14: 250-261, pls 38, 39. Chang Went’ang & Fan Chiasung 1960. [Class Trilobita of the Ordovician and the Silurian Periods of the Ch’i-Lien Mountains]. In Yin Tsanhsun (ed.), [Geological gazetteer of the Chi-Lien Mountains 4 (1, Palaeontological description . . .)]: 83-147, 10 pls. Beijing, Science Press. [In Chinese: Engl. transl. by National Technical Information Service, U.S. Dept of Commerce, pp. 246-430]. —— et al. 1964. In: [Atlas of Palaeozoic fossils from North Guizhou]: pls 5—10 (no text). Nanjing. [In Chinese]. ——,, see also Zhang. Chauvel, J. 1981. Etude critique de quelques echinoderms stylophores du Massif Armoricain. Bull. Soc. geol. miner. Bretagne, Rennes, (C) 13: 67-101, pl. 13. Chlupaé, I., Fligel, H. & Jaeger, H. 1981. Series or Stages within Palaeozoic Systems? Newsl. Stratigr., Leiden, 10: 78-91. Clark, T. H. 1924. The paleontology of the Beekmantown Series of Lévis, Quebec. Bull. Am. Paleont., Ithaca, 10: 1-119, 9 pls. Clarkson, E. N. K. 1967. Environmental significance of eye-reduction in trilobites and Recent arthropods. Mar. Geol., Amsterdam, 5: 367-375. — 1971. On the early schizochroal eyes of Ormathops (Trilobita, Zeliszkellinae). Mem. Bur. Rech. geol. minier., Paris, 73: 51-63, 1 pl. 292 R. A. FORTEY & R. M. OWENS Cocks, L. R. M. & Fortey, R. A. 1982. Faunal evidence for oceanic separations in the Palaeozoic of Britain. J. Geol. Soc. Lond., 139: 465-478. Cooper, G. A. & Kindle, C. H. 1936. New brachiopods and trilobites from the Upper Ordovician of Percé, Quebec. J. Paleont., Menasha, 10: 348-372, pls 51—S3. Cooper, R. A. 1973. Taxonomy and evolution of Isograptus Moberg in Australasia. Palaeontology, London, 16: 45-115. —— 1979. Ordovician geology and graptolite faunas of the Aorangi Mine area, north west Nelson, New Zealand. Palaeont. Bull. Wellington, 47: 1-127, 19 pls. —— & Fortey, R. A. 1982. The Ordovician graptolites of Spitsbergen. Bull. Br. Mus. nat. Hist., London, (Geol.) 36 (3): 157-302, pls 1-6. — ——.,, 1983. Development of the graptoloid rhabdosome. Alcheringa, Adelaide, 7: 201-221. Cope, J. C. W. 1980. Early history of the southern margin of the Tywi anticline in the Carmarthen area, South Wales. pp. 527-532 in Harris, A. L., Holland, C. H. & Leake, B. E., The Caledonides of the British Isles—reviewed. Spec. Publ. geol. Soc. Lond., 8. xii + 768 pp. ——,, Fortey, R. A. & Owens, R. M. 1978. Newly discovered Tremadoc rocks in the Carmarthen district, South Wales. Geol. Mag., Cambridge, 115: 195-198, 1 pl. Courtessole, R., Marek, L., Pillet, J.. Ubaghs, G. & Vizciano, D. 1983. Calymenina, Echinodermata et Hyolitha de ’Ordovicien inférieur de la Montagne Noire (France méridionale). Mem. Soc. Etud. scient. Aude, Carcassonne. 62 pp., 13 pls. —— & Pillet, J. 1976. Contribution a l’étude des trilobites de ’Ordovicien inférieur de la Montagne Noire: Eulominae et Nileidae. Annls Soc. geol. N., Lille, 95: 251-272, 3 figs, pls 24-27. See Pillet & Courtessole 1985. , , Vizcaino, D. & Eschard, R. 1985. Etude biostratigraphique et sedimentologique des Forma- tions arenacées de l’Arenigien du Saint Chinianais oriental (Hérault) versant sud de la Montagne Noire (France méridionale). Mem. Soc. Etud. scient. Aude, Carcassonne. 99 pp. Cox, A. H. 1916. The geology of the district between Abereiddy and Abercastle (Pembrokeshire). Q. J] geol. Soc. Lond., 71: 273-342, pls 22-26. 1930. Preliminary note on the geological structure of Pen Caer and Strumble Head, Pembrokeshire. Proc. Geol. Ass., London, 41: 274-289. , Green, J. F. N., Jones, O. T. & Pringle, J. 1930. The geology of the St. David’s district, Pembroke- shire. Proc. Geol. Ass., London, 41: 241-273, pls 15, 16. —— & Jones, O. T. 1914. The geology of the district between Abereiddy and Pencaer, Pembrokeshire. Rep. Br. Ass. Adumt Sci., Birmingham, 1913: 484-485. Crosfield, M. C. & Skeat, G. G. 1896. On the geology of the neighbourhood of Carmarthen. Q. JI geol. Soc. Lond., 52: 523-541, pls 25, 26. Dalman, J. W. 1827. Om palaeoderna eller de sa kallande Trilobiterna. K. svenska Vetensk Akad. Handl., Stockholm, 1826: 113-152, 226-294, pls i—vi. Davidson, T. 1868. On the earliest forms of Brachiopoda hitherto discovered in the British Palaeozoic rocks. Geol. Mag., London, (1) 5: 303-316, pls 15, 16. — 1869. A Monograph of the British fossil Brachiopoda. Part VII, The Silurian Brachiopoda. 3: 169-248, pls 23-37. Palaeontogr. Soc. (Monogr.), London. Dean, W. T. 1964. The status of the Ordovician trilobite genera Prionocheilus and Polyeres. Geol. Mag., Hertford, 101: 95-96. 1966. The Lower Ordovician stratigraphy and trilobites of the Landeyran Valley and the neighbour- ing district of the Montagne Noire, south-western France. Bull. Br. Mus. nat. Hist., London, (Geol.) 12 (6): 245-353, 21 pls. —— 1967. See Whittard 1955-67. — 1973a. Ordovician trilobites from the Keele Range, Northwestern Yukon Territory. Bull. geol. Surv. Can., Ottawa, 223. 39 pp., 5 pls. 1973b. The Lower Palaeozoic stratigraphy and faunas of the Taurus Mountains near Beysehir, Turkey, III. The trilobites of the Sobova Formation (Lower Ordovician). Bull. Br. Mus. nat. Hist., London, (Geol.) 24: 279-348, 12 pls. —— & Martin, F. 1978. Lower Ordovician acritarchs and trilobites from Bell Island, eastern Newfound- land. Bull. geol. Surv. Can., Ottawa, 284. 35 pp., 7 pls. Delo, D. M. 1935. A revision of the phacopid trilobites. J. Paleont, Menasha, 9: 402-420, 45 figs. — 1940. Phacopid trilobites of North America. Spec. Pap. geol. Soc. Am., New York, 29. 135 pp., 13 pls. Dennell, R. 1940. On the structure of the photophores of some decapod crustacea. ‘Discovery Rep., Cambridge, 20: 307-382. ARENIG IN SOUTH WALES 293 Destombes, J. 1972. Les trilobites du sous-ordre des Phacopina de l’Ordovicien de l’Anti-Atlas (Maroc). Notes Mem. Serv. Mines Carte géol. Maroc, Rabat, 240: 1-113, pls 1-16. Dewey, J. F., Rickards, R. B. & Skevington, D. 1970. New light on the age of the Dalradian deformation and metamorphism in western Ireland. Norsk geol. Tidsskr., Oslo, 50: 19-44. Donovan, S. K. 1984. Ramseyocrinus and Ristnacrinus from the Ordovician of Britain. Palaeontology, London, 27: 623-634. Elles, G. L. 1898. The Graptolite-fauna of the Skiddaw Slates. Q. JI geol. Soc. Lond., 54: 463-539. —— 1904. Some graptolite zones in the Arenig rocks of Wales. Geol. Mag., London, (5) 1: 199-211. —— 1922. Sedgwick Museum Notes. A new Azygograptus from North Wales. Geol. Mag., London, 59: 299-301. 1933. The Lower Ordovician graptolite faunas with special reference to the Skiddaw Slates. Summ. Progr. geol. Surv. Lond., 1932: 94-111. & Wood, E. M. R. 1901-18. A Monograph of British Graptolites. m + clxxi + 539 pp., 52 pls. Palaeontogr. Soc. (Monogr.), London. See Strachan, 1971. Emmons, E. 1855. American Geology, containing a statement of the principles of the Science ... &c. 1. 251 pp., 18 pls. Albany, N.Y. Etheridge, R. 1876. See Ward. — 1881. On the analysis and distribution of the British Palaeozoic fossils. Proc. geol. Soc. Lond., 1880-81: 51-235 (in Q. JI geol. Soc. Lond. 37). Evans, D. C. 1906. The Ordovician rocks of western Caermarthenshire. Q. JI geol. Soc. Lond., 62: 597-643, pl. 46. Eyans, W. D. 1945. The geology of the Prescelly Hills, north Pembrokeshire. Q. JI geol. Soc. Lond., 101: 89-110, pls 3, 4. Fearnsides, W. G. 1905. On the geology of Arenig Fawr and Moel Llyfnant. Q. JI geol. Soc. Lond., 61: 608-640, pl. 41. Fortey, R. A. 1971. Tristichograptus, a triserial graptolite from the Lower Ordovician of Spitsbergen. Palaeontology, London, 14: 188-199. 1974. A new pelagic trilobite from the Ordovician of Spitsbergen, Ireland and Utah. Palaeontology, London, 17: 111-124, pls 13, 14. —— 1975-80. The Ordovician trilobites of Spitsbergen. II. Asaphidae, Nileidae, Raphiophoridae and Telephinidae of the Valhallfonna Formation. Skr. norsk Polarinst, Oslo, 162: 1-125, pls 1-41 (1975). III. Remaining trilobites of the Valhallfonna Formation. Loc. cit., 171: 1-113, pls 1-25 (1980). — 1981. Prospectatrix genatenta (Stubblefield) and the trilobite superfamily Cyclopygacea. Geol. Mag., Cambridge, 118: 603-614. —— 1982. In Fortey, R. A., Landing, E. & Skevington, D., Cambrian—Ordovician boundary sections in the Cow Head Group, Western Newfoundland. In Bassett, M. G. & Dean, W. T. (eds), The Cambrian— Ordovician boundary . . . &c: 95-129. Cardiff, Natl Mus. Wales. 1983. Geometrical constraints in the construction of graptolite stipes. Paleobiol., Menlo Park, 9: 116-125. —— 1984. Global earlier Ordovician transgressions and regressions and their biological implications. In Bruton, D. L. (ed.), Aspects of the Ordovician System: 37—S0. Oslo. — 1985Sa. Gradualism and punctuated equilibria as competing or complementary theories. Spec. Pap. Palaeont., London, 32: 17—28. 1985b. Pelagic trilobites as an example of deducing the life habits of extinct arthropods. Trans. R. Soc. Edinb., 76: 219-230. — & Morris, S. F. 1982. The Ordovician trilobites Neseuretus from Saudi Arabia, and the palaeogeog- raphy of the Neseuretus fauna related to Gondwanaland in the earlier Ordovician. Bull. Br. Mus. nat. Hist., London, (Geol.) 36: 63-75. & Owens, R. M. 1978. Early Ordovician (Arenig) stratigraphy and faunas of the Carmarthen district, south-west Wales. Bull. Br. Mus. nat. Hist., London, (Geol.) 30 (3): 225—294, pls 1-11. —— & Rushton, A. W. A. 1980. Acanthopleurella Groom, 1902: origin and life habits of a miniature trilobite. Bull. Br. Mus. nat. Hist., London, (Geol.) 33: 79-89. —— & Shergold, J. H. 1984. Early Ordovician trilobites, Nora Formation, Central Australia. Palaeont- ology, London, 27: 315-366, pls 38-46. Gigout, M. 1951. Etudes géologiques sur la Méséta Marocaine Occidentale (arriére Pays de Casablanca, Mazagan et Safi). Trav. Inst. scient. cherif., Tangier, 3: 1-507, 18 pls. Giirich, G. 1907. Versuch einer Neueinteilung der Trilobiten. Zentbl. Miner. Geol. Palaont., Stuttgart, 1907: 129-133. 294 R. A. FORTEY & R. M. OWENS Hall, J. 1858. Descriptions of Canadian graptolites. Rep. geol. Surv. Can., Toronto, 1857: 111-145, pls 1-8. 1865. Graptolites of the Quebec Group. Figures and descriptions of Canadian organic remains, Dec. II. 151 pp. Montreal, Canada Geol. Surv. Hammann, W. 1971a. Die Placopariinae (Trilobita, Cheirurina; Ordovizium). Senckenberg. leth., Frank- furt a.M., 52: 53-75, pls 1-3. 1971b. Stratigraphische Einteilung des spanischen Ordoviziums nach Dalmanitacea und Cheirurina (Trilobita). Mem. Bur. Rech. géol. minier., Paris, 73: 265-272, 1 pl. —— 1974. Phacopina und Cheirurina (Trilobita) aus dem Ordovizium Spaniens. Senckenberg. leth., Frankfurt a.M., 55: 1-151, 12 pls. —— 1983. Calymenacea (Trilobita) aus dem Ordovizium von Spanien; ihre Biostratigraphie, Okologie und systematik. Abh. senckenb. naturforsch. Ges., Frankfurt-a.M., 542: 1-177, 25 pls. Han Nai-ren 1978. Panderian organs of Cyclopyge rotundata Lu (Trilobita). Acta palaeont. sin., Beijing, 17: 351-356, 2 pls [In Chinese and English]. Harkness, R. & Nicholson, H. A. 1866. Additional observations on the Geology of the Lake Country. With a note on two species of trilobites by J. W. Salter esq., F.G.S. Q. JI geol. Soc. Lond., 22: 480-88, 4 woodcuts. Harland, W. B., Cox, A. V., Llewellyn, P. G. et al. 1982. A geologic time scale. 131 pp. Cambridge. Harrington, H. J. & Leanza, A. F. 1957. Ordovician trilobites of Argentina. Spec. Publ. Dept. Geol. Univ. Kans., Lawrence, 1: 1-276, 140 figs. et al. 1959. Trilobita. In Moore, R. C. (ed.), Treatise on Invertebrate Paleontology, O (Arthropoda 1): O38—O540. Lawrence, Kansas. Harris, W. J. & Keble, R. A. 1932. Victorian graptolite zones, with correlations and description of species. Proc. R. Soc. Vict., Melbourne, 44: 25—48. & Thomas, D. G. 1935. Victorian Graptolites (New Series). Part III. Proc. R. Soc. Vict., Melbourne, 47: 288-313. Hawle, I. & Corda, A. J. C. 1847. Prodrom einer Monographie der bohmischen Trilobiten. 176 pp., 7 pls. Prague. Henningsmoen, G. 1957. The trilobite family Olenidae. Skr. norske Vidensk-Akad. mat.-nat. Kl., Oslo, 1957 (1): 1-303, 31 pls. —— 1959. See Harrington et al. Henry, J.-L. 1970. Quelques Calymenacea (Trilobites) de ’Ordovicien de Bretagne. Annls Paléont., Paris, (Invert.) 56: 1-27, pls A-C. —— 1971. Les trilobites Asaphidae et Eohomalonotidae du Grés Amoricain supérieur (Arénigien) de Youest de la France. Mem. Bur. Rech. géol. minier., Paris, 73: 65—77, 2 pls. —— 1980. Trilobites ordoviciens du Massif Armoricain. Mem. Soc. geéol. miner. Bretagne, Rennes, 22: 1-250, 48 pls. —— & Clarkson, E. N. K. 1975. Enrollment and coaptations in some species of the Ordovician trilobite genus Placoparia. Fossils Strata, Oslo, 4: 87-95, 3 pls. Herring, P. J. 1978. Bioluminescence in action. 570 pp. London. Hicks, H. 1873. On the Tremadoc rocks in the neighbourhood of St. David’s, South Wales, and their fossil contents. Q. JI geol. Soc. Lond., 29: 39-52, pls 3—S. —— 1875. On the succession of the ancient rocks in the vicinity of St. David’s, Pembrokeshire, with special reference to those of the Arenig and Llandeilo Groups, and their fossil contents. Q. JI geol. Soc. Lond., 31: 167-195, pls 8-11. 1881. The classification of the Eozoic and Lower Palaeozoic rocks of the British Isles. Pop. Sci. Rev., London, (n.s.) 5: 289-308, 3 pls. —— & Salter, J. W. 1867. Second report on the “Menevian Group” and the other formations at St. David’s, Pembrokeshire. Rep. Br. Ass. Advmt Sci., Nottingham, 1866: 182-186. Hoeck, H. 1912. In Steinmann, G. & Hoeck, H., Das Silur und Cambrium des Hochlandes von Bolivia und ihre Fauna. Neues Jb. Miner. Geol. Palaont. BeilBd, Stuttgart, 34: 176-252, pls 7-14. Holub, K. 1908. Prispevek ku poznani fauny pasma Dd,,. Rozpr. Ceske Akad., Prague, 17: 1-19, 1 pl. — 1911. Nova fauna spodniho siluru v okoi Rokycan. Rozpr. Ceske Akad., Prague, 20 (15): 1-18, pls 1, 2. —— 1912. Doplnky ku fauné Eulomoveho horizontu v okoli Rokycan. Rozpr. Ceske Akad., Prague, (2) 21 (33): 1-12, 1 plate. Hopkinson, J. 1872. On Callograptus radicans, a new dendroid graptolite. Ann. Mag. nat. Hist., London, (4) 10: 232-237, pl. 10. —— 1873. The graptolites of the Arenig rocks of St. David’s, South Wales. Abstr. Proc. Lpool geol. Soc., 14: 36-42. ARENIG IN SOUTH WALES 295 — & Lapworth, C. 1875. Descriptions of the graptolites of the Arenig and Llandeilo rocks of St. David’s. Q. JI geol. Soc. Lond., 31: 631-672, pls 33-37. Horbinger, F. & Vanék, J. 1983. New Ordovician Ellipsotaphridae and Remopleuridae (Trilobita). Cas. Miner. Geol., Prague, 28 (3): 303-306, pls 1, 2. 1985. New cyclopygid trilobites from the Ordovician of Bohemia. Cas. Miner. Geol., Prague, 30: 59-64, 2 pls. Horny, R. & Bastl, F. 1970. Trilobita. Type specimens of fossils in the National Museum Prague, 1. 354 pp., 20 pls. Prague. Howe, M. P. A. 1983. Measurement of thecal spacing in graptolites. Geol. Mag., Cambridge, 120: 635-638. Howell, B. F. 1935. Cambrian and Ordovician trilobites from Herault, southern France. J. Paleont., Menasha, 9: 222-238, pl. 23. Hsii S. C. 1934. The graptolites of the Lower Yangtze Valley. Monogr. natn Res. Inst. Geol. Shanghai, (A) 4. 106 pp, 7 pls. Hughes, C. P. 1969-79. The Ordovician trilobite faunas of the Builth—Llandrindod inlier, central Wales. I. Bull. Br. Mus. nat. Hist., London, (Geol.) 18: 41-103, 4 pls (1969). II. Loc. cit. 20: 115-182, pls 1-16 (1971). III. Loc. cit. 32: 109-181 (1979). ——, Ingham, J. K. & Addison, R. 1975. The morphology, classification, and evolution of the Trinucleidae (Trilobita). Phil. Trans. R. Soc., London, (B) 272: 537-604, pls 1-10. ——., Jenkins, C. J. & Rickards, R. B. 1982. Abereiddi Bay and the adjacent coast. In Bassett 1982: 51-63 (q.V.). — & Wright, A. J. 1970. The trilobites Incaia Whittard 1955 and Anebolithus gen. nov. Palaeontology, London, 13: 677-690, pls 127, 128. Hupe, P. 1953. Classe de trilobites. In Piveteau, J. (ed.), Traite de paleontologie, 3: 44-246. Paris. —— 1955. Classification des trilobites. Annls Paléont, Paris, 41: 91-325, figs 93-247. Hutchison, R. & Ingham, J. K. 1967. New trilobites from the Tremadoc Series of Shropshire. Palaeont- ology, London, 10: 47-—S9, pl. 8. Ingham, J. K. 1974. The Upper Ordovician trilobites from the Cautley and Dent districts of Westmorland and Yorkshire. Part 2: 59-87, pls 10-18 Palaeontogr. Soc. (Monogr.), London. Jaanusson, V. 1959. See Harrington et al. Jackson, D. E. 1962. Graptolite zones in the Skiddaw Group in Cumberland, England. J. Paleont., Tulsa, 36: 300-313. —— 1978. The Skiddaw Group. In Moseley, F. (ed.), The Geology of the Lake District: 79-98, pls 6-9. Leeds, Yorkshire Geol. Soc. (Occ. Publ. 3). — 1979. A new assessment of the stratigraphy of the Skiddaw Group along the northern edge of the main Skiddaw inlier. Proc. Cumberld geol. Soc., Cleator Moor, 4: 21-31. Jaekel, O. 1909. Uber die Agnostiden. Z. dt. geol. Ges., Berlin, 61: 380-401. Jefferies, R. P. S. 1979. The origin of chordates—a methodological essay. In House, M. R. (ed.), The origin of major invertebrate groups: 443-477. London (Systs Assoc. spec. Vol. 12). — 198la. Fossil evidence on the origin of the chordates and echinoderms. In: Ranzi, L. (ed.), Origine dei grandi phyla dei metazoa: 487-561. Atti Conv. Lincei, Rome, 49. 565 pp. — 1981b. In defence of the calcichordates. Zool. J. Linn. Soc., London, 73: 351-396. — 1984. Locomotion, shape, ornament, and external ontogeny in some mitrate calcichordates. J. Vert. Paleont., Norman, Ok., 4: 292-319. 1986. The ancestry of the vertebrates. 384 pp. London, British Museum (Natural History). Jenkins, C. J. (1979). See p. 96. —— 1982. Isograptus gibberulus (Nicholson) and the isograptids of the Arenig Series (Ordovician) of England and Wales. Proc. Yorks. Geol. Soc., Leeds, 44: 219-248, pls 16, 17. — 1983. Ordovician graptolites from the Great Paxton borehole, Cambridgeshire. Palaeontology, London, 26: 641-653. Jentsch, S. von & Stein, V. 1961. Neue fossilfunde im Ordovizium des Ebbe-Sattels. Palaont. Z., Stuttgart, 35: 200-208. Jollie, M. 1982. What are the ‘Calcichordata’? and the larger question of the origin of chordates. Zool. J. Linn. Soc., London, 75: 167-188. Kielan, Z. 1960. Upper Ordovician trilobites from Poland and some related forms from Bohemia and Scandinavia. Palaeont. pol., Warsaw, 11. vi + 198 pp, 36 pls. Klouéek, C. 1916. O vrstvach D-d,,, jich trilobitech a nalezistich. Rospr. Ceske Akad., Prague, 25 (2): 1-21, 1 pl. Kobayashi, T. 1931. Studies on the stratigraphy and palaeontology of the Cambro-Ordovician faunas of Hualienchai and Niuhsintai, south Manchuria. Jap. J. Geol. Geogr., Tokyo, 8: 139-189, pls 16—22. 296 R. A. FORTEY & R. M. OWENS —— 1937. The Cambro-Ordovician shelly faunas of South America. J. Fac. Sci. Tokyo Univ., (2 Geol.) 4 (4): 369-522, pls 1-8. — 1939a. On the Agnostids, 1. J. Fac. Sci. Tokyo Univ. (2 Geol) 5: 66-198. —— 1939b. Supplementary notes on the Agnostida. J. geol. Soc. Japan, Tokyo, 46: 577-580. —— 1940. Note on the Dionideidae. Jap. J. Geol. Geogr., Tokyo, 17: 203-308. —— 1951. On the Ordovician trilobites in Central China. J. Fac. Sci. Tokyo Univ., (2 Geol.) 8: 1-87, pls 1-S. 1960. The Cambro-Ordovician Formations and Faunas of South Korea, VI, Palaeontology V. J. Fac. Sci. Tokyo Univ., (2 Geol.) 12: 217-275, pls 12-14. & Hamada, T. 1971. Contributions to the geology and palaeontology of Southeast Asia. 78. A cyclopygid-bearing Ordovician faunule discovered in Malaysia with a note on the Cyclopygidae. Geol. Palaeont. S.E. Asia, Tokyo, 8: 1-18, 2 pls. Kokelaar, B. P., Bevins, R. E. & Roach, R. A. 1985. Submarine silicic volcanism and associated sedimen- tary and tectonic processes, Ramsey Island, S.W. Wales. J. geol Soc. Lond., 142: 591-614. Koroleva, M. N. 1964. [New Middle Ordovician shumardiid trilobites from northern Kazakhstan. ] Paleont. Zh., Moscow, 1964 (1): 71-75, pl. 3 [In Russian; Engl. transl. available: see Int. Geol. Rev., Washington, 7 (8): 1408]. —— 1967. Kazakhstanskie Trilobity semejstva Cyclopygidae. Paleont. Zh., Moscow, 1967 (1): 79-91, pl. 10. [In Russian: Engl. transl. Paleont. J., Washington, 1 (1): 74-85]. — 1982. [Ordovician trilobites of Northeastern Kazakhstan.] 164 pp., 26 pls. Moscow, Nedra (KazIMS). [In Russian]. Kraft, J. 1971-72. Nove nalezy graptolitu v klabavskych vrstrach (arenig) barrandienskeho ordoviku. Zpr. Muz. Zapadoéeskeho Kraje, Plzen, (Priroda) 12: 47—S1 (1971). Loc. cit. 13: 39-42 (1972). —— 1973. New graptolites from the Klabava Formation (Arenig) of the Barrandian Ordovician. Cas. Miner. Geol., Prague, 18: 25—30. — 1974. [On the graptolite fauna from the Klabava Formation (Arenig) of the Ordovician of the Barrandian (Part 1—Myto and Volduchy localities).] Zpr. Muz. Zapadoceskeho Kraje, Plzen, (Prtiroda) 16: 53—59 [in Czech and English]. —— 1977. Graptolites from the Klabava Formation (Arenigian) of the Ordovician of Bohemia. Folia Mus. Rerum nat. Bohemiae occident., Plzen, (Geol.) 6: 1-31, 12 pls. Lake, P. 1906-46. A Monograph of the British Cambrian trilobites. 350 pp., 47 pls. Palaeontogr. Soc. (Monogr.), London. Lane, P. D. 1971. British Cheiruridae (Trilobita). 95 pp., 16 pls. Palaeontogr. Soc. (Monogr.), London. — & Thomas, A. T. 1983. A review of the trilobite suborder Scutelluina. Jn Briggs, D. E. G. & Lane, P. D., Trilobites and other early arthropods: papers in honour of Professor H. B. Whittington, F.R.S. Spec. Pap. Palaeont., London, 30: 141—160. Lapworth, C. 1873. On an improved classification of the Rhabdophora. Geol. Mag, London, (1) 10: 500-504, 555-560. — 1875. See Hopkinson & Lapworth 1875. — 1880. On the geological distribution of the Rhabdophora. Part III. Results. Ann. Mag. nat. Hist., London, (5) 6: 16—29, 185-207. La Touche, J. D. 1884. A handbook of the geology of Shropshire. 99 pp., 22 pls. London & Shrewsbury. Legg, D. P. 1976. Ordovician trilobites and graptolites from the Canning Basin, Western Australia. Geologica Palaeont., Marburg, 10: 1—58, 10 pls. — 1978. Ordovician biostratigraphy of the Canning Basin, Western Australia. Alcheringa, Adelaide, 2: 321-334. Li Shanji 1978. In: [Fossils of South-west China. Szechuan region], 1. 617 pp., 185 pls. Beiying. [In Chinese]. Lu Yanhao 1962. Trilobita. In Wang Yu (ed.), [A Handbook of index fossils of the Yangtze region]. 188 pp., 98 pls. Beijing. [In Chinese]. — 1964. [A handbook of the index fossils from South China]. Beijing (Nanjing Inst. Geol. Palaeont., Acad. Sin.) [In Chinese; not seen]. — 1975. Ordovician trilobite faunas of Central and Southwestern China. Palaeont. sin., Beijing, 152: 1-463, 50 pls. [In Chinese and English]. —, Chu Chaoling, Chien Yiyuan, Zhou Zhiyi, Chen Junyuan, Liu Gengwu, Yu Wen, Chen Xu & Xu Hankui 1976. [Ordovician biostratigraphy and palaeozoogeography of China]. Mem. Inst. Palaeont., Nanjing, 7: 1-83, 14 pls. [In Chinese]. — & Zhang Wentan 1974. [Ordovician trilobites]. In: [A handbook of stratigraphy and palaeontology in South-west China]: 124-136, pls 49—S6. Beijing. [In Chinese]. ARENIG IN SOUTH WALES 297 et al. 1965. [Fossils of each group of China: Chinese trilobites.] 2 vols. 766 pp., 134 pls. Beijing. [In Chinese]. Lynas, B. D. T. 1973. The Cambrian and Ordovician rocks of the Migneint area, North Wales. J. geol. Soc. Lond., 129: 481-503. M‘Coy, F. 1849. On the classification of some British fossil Crustacea, with notices of new forms in the University collection at Cambridge. Ann. Mag. nat. Hist., London, (2) 4: 161-174, 330-335, 392-414, 15 figs. —— 1851. In Sedgwick, A. & M‘Coy, F., A synopsis of the classification of the British Palaeozoic rocks .. . (&c.), pt 1. 184 pp. Cambridge. McKerrow, W. S., Lambert, R. St J. & Cocks, L. R. M. 1985. The Ordovician, Silurian and Devonian. In Snelling, N. J. (ed.), Geochronology and the geological record. Mem. geol. Soc. Lond. 10: 73-80. Mansuy, H. 1916. Faunes Cambriennes de |’extréme-orient meéridionale. Mem. Serv. geol. Indoch., Hanoi, 5: 1-48, pls 1-7. Marek, L. 1961. The trilobite family Cyclopygidae Raymond in the Ordovician of Bohemia. Rozpr. ustred. Ust. geol., Prague, 28. 85 pp., 6 pls. —— 1964. Shumardia Billings, 1862 a Staurocephalus Barrande, 1846 v ceském ordoviku. Cas. narod Mus., Prague, 133: 153-154, pl. 1. — 1966. Nadéeléd Bohemillacea. Cas. narod. Mus., Prague, 135: 145-153. — 1977. Celed Ellipsotaphridae Kobayashi et Hamada 1970 (Trilobita). Cas. narod. Mus., Prague, 143 (for 1974): 69-71. Matley, C. A. 1928. The Pre-Cambrian complex and associated rocks of south-western Lleyn (Carnarvonshire), with a chapter on the petrology of the complex by E. Greenly. Q. JI geol. Soc. Lond., 84: 440-504, pls 26-32. —— 1932. The geology of the country around Mynydd Rhiw and Sarn, south-western Lleyn, Carnarvon- shire. Q. JI geol. Soc. Lond., 88: 238-273, pl. 14. Milne Edwards, H. 1840. Histoire naturelle des Crustaces, comprenant l’anatomie, la physiologie et la classification de ces animaux. 3. 638 pp. Paris. Moberg, J. C. 1892. Om nagra nya graptoliter fran skanes undre graptolit-skiffer. Geol. For. Stockh. Forh., 14: 339-349, pl. 8. — 1901. Nya bidrag till utredning af fragan om gransen mellan undersilur och kambrium. Geol. For. Stockh. Forh., 22: 523-540, pl. 14. —— & Segerberg, C. O. 1906. Bidrag till Kannedomen om Ceratopygeregionen med sarskild nansyn till dess utveckling 1 Fagelsangstrakten. Acta Univ. lund., (N.F.) 2: 1-113, pls 1-7. Molyneux, S. G. 1987. Appendix. Acritarchs and Chitinozoa from the Arenig Series of south-west Wales. Bull. Br. Mus. nat. Hist., London, (Geol.) 41 (4): 309-364. — & Rushton, A. W. A. 1985. Discovery of Tremadoc rocks in the Lake District. Proc. Yorks Geol. Soc., Leeds, 45: 123-127. Monsen, A. 1937. Die graptolithenfauna in Unteren Didymograptusschiefer (Phyllograptusschiefer) Nor- wegens. Norsk geol. Tiddskr., Oslo, 16: 57—267. Mu A. T., Ge M. Y., Chen X., Ni Y. N. & Lin Y. K. 1979. Lower Ordovician graptolites of Southwest China. Palaeont. sin., Beijing, 156B: 1-192, pls 1-48. [Chinese, English summary ]. — & Li Jijin 1958. Scandent graptolites from the Ningkuo Shale of Kiangshan—Changshan area, Western Chekiang. Acta palaeont. sin., Beijing, 6: 391—428. [In Chinese and English]. Murchison, R. I. 1839. The Silurian System, founded on geological researches in the counties of Salop, Hereford, Radnor, Montgomery, Caermarthen, Brecon, Pembroke, Monmouth, Worcester, Gloucester and Stafford ; with descriptions of the coalfields and overlying formations. xxxii + 768 pp., 37 pls. London. —— 1859-72. Siluria. The history of the oldest known rocks containing organic remains, with a brief description of the distribution of gold over the earth. 3rd edn (1859), xx + 592 pp., 41 pls. 4th edn (1867), xvill + 566 pp., 41 pls. Sth edn (1872), xviii + 566 pp., 41 pls. London. Neben, W. & Krueger, H. H. 1971-73. Fossilien ordovicischer [und silurischer] Geschiebe. Staringia, Pinneberg, 1. 5 pp., 50 pls (1971). Loc. cit., 2. 10 pp., pls 51-109 (1973). Neuman, R. B. & Bates, D. E. B. 1978. Reassessment of Arenig and Llanvirn age (early Ordovician) brachiopods from Anglesey, north-west Wales. Palaeontology, London, 21: 571—613, pls 63-68. Nicholas, T. C. 1915. The geology of the St Tudwal’s Peninsula, Carnarvonshire. Q. JI geol. Soc. Lond., 71: 83-143. Nicholson, H. A. 1868. On the graptolites of the Coniston Flags; with notes on the British species of the genus Graptolites. Q. JI geol. Soc. Lond., 24: 521-545, pls 19-20. —— 1869. On some new species of graptolites. Ann. Mag. nat. Hist., London, (4) 4: 231-242, pl. 11. — 1870. On the British species of Didymograptus. Ann. Mag. nat Hist., London, (4) 5: 337-357, pl. 7. 298 R. A. FORTEY & R. M. OWENS —— 1875. On a new genus and some new species of graptolites from the Skiddaw Slates. Ann. Mag. nat. Hist., London, (4) 16: 269-273. Novak, O. 1883. Zur Kenntniss der béhmischen Trilobiten. Beitr. Palaont. Geol. Ost.-Ung., Vienna, 3: 26-63, pls 8-12. —— & Perner, J. 1918. Die Trilobiten der Zone D-d,, von Prag und Umgebung. Palaeontogr. Bohem., Prague, 9: 1-55, pls 1-4. Obut, A. M. & Sobolevskaya, R. F. 1964. Graptolity Ordovika Taimyra [Ordovician graptolites of Taimir (N. Siberia)]. 92 pp., 16 pls. Moscow, Akad. Nauk SSSR (Sibirsk. otdel. Inst. Geol. Geofiz.) [In Russian ]. Opik, A. A. 1967. The Mindyallan fauna of north-western Queensland. Bull. Bur. Miner. Resour. Geol. Geophys. Aust., Melbourne, 74; 1—404, pls 1—S1. Owen, A. W. & Bruton, D. L. 1980. Late Caradoc-early Ashgill trilobites of the central Oslo Region, Norway. 63 pp., 10 pls. Oslo, Paleont. Mus. Owens, R. M. & Fortey, R. A. 1982. Arenig rocks of the Carmarthen—Llanarthney district. In Bassett 1982: 249-258 (q.v.). 3 » Cope, J. C. W., Rushton, A. W. A. & Bassett, M. G. 1982. Tremadoc faunas from the Carmarthen district, South Wales. Geol. Mag., Cambridge, 119: 1—38, pls 1-8. Paul, C. R. C. & Cope, J. C. W. 1982. A parablastoid from the Arenig of South Wales. Palaeontology, London, 25: 499-507, pl. 51. Pek, I. 1969. Corrugatagnostus refragor sp. n. (Trilobita) from the Llanvirnian of the Barrandian Region. Vest ustred. Ust. geol., Prague, 44: 383-384, 1 pl. —— 1977. Agnostid trilobites of the central Bohemian Ordovician. Sb. geol. Véd. Praha, (P) 1977 (19): 7-44, pls 1-12. —— & Prokop, R. J. 1984. New finds of the Agnostid trilobites from the Ordovician of the Prague area (Czechoslovakia). Cas. narod. Mus., Prague, 153: 17-20, 1 pl. [English, Czech summary]. Perner, J. 1895. Etudes sur les Graptolites de Bohéme II. Monographie des Graptolites de étage D. Palaeontogr. Bohem., Prague, 3: 1-31, pls 4-8. Petrunina, Z. E. 1975. In Repina, L. N. et al., [Stratigraphy and fauna of the Lower Palaeozoic of the northern submontane belt of Turkestan and Alai Ridges (Southern Tien-Shan)]. Trudy Inst. Geol. Geofiz. sib. Otd., Novosibirsk, 278: 1-248, 48 pls. [In Russian]. Philip, G. M. 1979. Carpoids—echinoderms or chordates? Biol. Rev., Cambridge, 54: 439-471. Pillet, J. & Courtessole, R. 1985. Contribution a l’étude des trilobites de Ordovician inférieur de la Montagne Noire: Cyclopygidae et Isocolidae. Annls Soc. geol. N., Lille, 104: 209-218. Postlethwaite, J. 1885. Trilobites of the Skiddaw Slates. Trans. Cumberland Ass. Advanc. Sci., Keswick, 10: 71-100, pls 1-4. —— 1897. The Geology of the English Lake District with notes on the minerals. 78 pp. Keswick. —— 1913. Mines and Mining in the (English) Lake District, 3rd edn. x + 164 pp., 15 pls. Whitehaven. — & Goodchild, J. G. 1886. On some trilobites from the Skiddaw Slates. Proc. Geol. Ass., London, 9: 455-469. Poulsen, V. 1965. An early Ordovician trilobite fauna from Bornholm. Meddr dansk geol. Foren., Copen- hagen, 16: 49-113, pls 1-9. Prantl, F. & Pribyl, A. 1949a. O novych nebo malo znamych trilobitech éeskeého ordoviku. Rozpr. ceske Akad. Ved. Umeéni, Prague, (2) 58 (8): 1—22, pls 1-3. 1949b. A study of the superfamily Odontopleuracea nov. superfam. (trilobites). Rozpr. st. geol. Ust., Prague, 12: 119-221, 11 pls. & Snajdr, M. 1957. A study on the genus Placoparia Hawle and Corda (Trilobitae). Sb. Ustred. Ust. Geol., Prague, 23: 497-521, 2 pls. Pyibyl, A. 1950. Nova pojmenovani pro nékolik homonymickych jmen éeskych i cizich Trilobitovych rodu [Re-naming some homonymic names of Bohemian and foreign trilobite genera]. Sb. st. geol. Ust. Csl. Repub., Prague, 17: 173-200 (with Russian and English summaries). & Vanék, J. 1965. Neue Trilobiten des bohmischen Ordoviziums. Vést. ustred. Ust. geol., Prague, 40: 277-282, 2 pls. —— —— 1968. Einige Trilobiten aus dem béhmischen Ordovizium. Vést. ustred. Ust. geol., Prague, 43: 191-197, pls 1, 2. 1973. Einige Bemerkungen zu den Vertretern von Selenopeltis Hawle & Corda, 1847. Cas. Miner. geol., Prague, 18: 63-70, 4 pls. 1980. Ordovician trilobites of Bolivia. Rozpr. esl. Akad. Ved., Prague, 90: 1—90, 26 pls. Pringle, J. 1911. Note on the ‘Lower Tremadoc rocks of St. David’s, Pembrokeshire. Geol. Mag., London, (5) 8: 556-559. ARENIG IN SOUTH WALES 299 —— 1930. The geology of Ramsey Island. Proc. geol. Ass., London, 41: 1-31, pls 1-3. Rabano, I. 1984. Nuevas observaciones sobra Placoparia (Placoparia) cambriensis Hicks, 1875 (Trilobita, Cheirurina) en el Llanvirn de la Zona Centroibérica. Coloquios Cated. Palaeont. Fac. Cienc. Univ. Madrid, 39: 7—16, 1 pl. — & Gutierrez-Marco, J. C. 1984. Revision de género Ectillaenus Salter, 1867 (Trilobita, Illaenina) en el Ordovicico de la Peninsula Ibérica. Boln R. Soc. esp. Hist. nat., Madrid, (Geol.) 81: 225-246. Ramsbottom, W. H. C. 1961. The British Ordovician Crinoidea. 37 pp., 8 pls. Palaeontrogr. Soc. (Monogr.), London. Rasetti, F. 1954. Early Ordovician trilobite faunules from Quebec and Newfoundland. J. Paleont., Tulsa, 28: 581-587. Raymond, P. E. 1910. Notes on Ordovician trilobites. II. Asaphidae from the Beekmantown. Ann. Carneg. Mus., Pittsburgh, 7 (1): 35-45, pl. 14. — 1912. Notes on parallelism among the Asaphidae. Trans. R. Soc. Can., Ottawa, (ser. 3, sect. 4) 5: 111—120, pls 1-3. —— 1913. Some changes in the names of genera of trilobites. Ottawa Nat., 26 (11): 137-142. — 1914. The Succession of Faunas at Lévis, P.Q. Am. J. Sci., New Haven, (4) 38: 523-530. —— 1925. Some trilobites of the Lower Middle Ordovician of eastern North America. Bull. Mus. comp. Zool. Harv., Cambridge, Mass., 67: 1-181, pls 1-10. —— 1937. Upper Cambrian and Lower Ordovician Trilobita and Ostracoda from Vermont. Bull. geol. Soc. Am., New York, 48: 1079-1146, pls 1—4. Reed, F. R. C. 1903-31. The Lower Palaeozoic trilobites of the Girvan district, Ayrshire. 186 pp., 20 pls (1903-06). Supplement, 56 pp., 8 pls (1914). Supplement no. 2, 64 pp., 4 pls (1931). Palaeontogr. Soc. (Monogr.), London. —— 1905. The classification of the Phacopidae. Geol. Mag., London, (5) 2: 172-178, 224-228. 1931. A review of the British species of the Asaphidae. Ann. Mag. nat. Hist., London, (10) 7: 441-472. Richter, R. & Richter, E. 1954. Die Trilobiten des Ebbe-Sattels und zu vergleichende Arten (Ordoviczium, Gotlandium (Devon)). Abh. senckenb. naturforsch. Ges., Frankfurt-a.M., 488: 1—76, 6 pls. Rickards, R. B. 1973. The Arenig graptolite genus Pseudotrigonograptus Mu and Lee. Acta geol. pol., Warsaw, 23: 597-604. Roberts, T. 1893. Notes on the geology of the district west of Caemarthen. Q. JI geol. Soc. Lond., 49: 166-170. Robison, R. A. & Pantoja-Alor, J. 1968. Tremadocian trilobites from the Nochixtlan Region, Oaxaca, Mexico. J. Paleont., Tulsa, 42: 767-800, pls 97-104. Romano, M. 1976. The trilobite genus Placoparia from the Ordovician of the Valongo area, north Portugal. Geol. Mag., Cambridge, 113: 11—28, 1 pl. Ross, R. J. 1958. Trilobites in a pillow-lava of the Ordovician Valmy Formation, Nevada. J. Paleont., Tulsa, 32: 559-570, pls 83-84. —— & Berry, W. B. N. 1963. Ordovician graptolites of the Basin Ranges in California, Nevada, Utah and Idaho. Bull. U.S. geol. Surv., Washington, 1134. 177 pp., 13 pls. Rouault, M. 1849. Memoire 1° sur la composition du test des Trilobites; 2° sur les changements de formes dus a ces causes accidentelles, ce qui a pu permettre de confondre des espéces différentes. Bull. Soc. geol. Fr., Paris, (2) 6: 67-89, pls 1, 2. Roy, S. K. 1929. Contributions to Paleontology. Publs Field Mus. nat. Hist., Chicago, (Geol.) 4: 203—220, pls 32-40. Ruedemann, R. 1904. Graptolites of New York, Part 1. Graptolites of the lower beds. Mem. N.Y. St. Mus. nat. Hist., Albany, 7: 457-803. 1934. Paleozoic Plankton of North America. Mem. geol. Soc. Am., Washington, 2: 1-106, pls 1—26. —— 1947. Graptolites of North America. Mem. geol. Soc. Am., Washington, 19: 1-652, 92 pls. Rushton, A. W. A. 1985a. The type material of Didymograptus hirundo Salter. Geol. Mag., Cambridge, 122: 197-198. —— 1985b. A Lancefieldian graptolite from the Lake District. Geol. Mag., Cambridge, 122: 329-333. ——., Fortey, R. A. & Owens, R. M. 1979. Excavation of two candidate sections for stratotypes in Wales by the Nature Conservancy Council. Earth Sci. Conserv., Newbury, 16: 1-5, pl. 1. —— & Hughes, C. P. 1981. The Ordovician trilobite fauna of the Great Paxton Borehole, Cambridge- shire. Geol. Mag., Cambridge, 118: 623-646, 6 pls. —— & Tripp, R. P. 1979. A fossiliferous lower Canadian (Tremadoc) boulder from the Benan Conglomer- ate of the Girvan district. Scott. J. Geol., Edinburgh, 15: 321-327, 1 pl. Salter, J. W. 1853. On the lowest fossiliferous beds of North Wales. Rep. Br. Ass. Adumt Sci., Belfast, 1852: 56-58. 300 R. A. FORTEY & R. M. OWENS — 1859. See Murchison. —— 1863. Note on the Skiddaw Slate fossils. Q. JI geol. Soc. Lond., 19: 79-84, 135-140. —— 1864. Trilobites (chiefly Silurian). Figures and descriptions illustrative of British organic remains, dec. 11. 10 pls. Mem. geol. Surv. U.K., London. —— 1865S. In Salter, J. W. & Blandford, H. G., Palaeontology of Niti in the northern Himalaya. 112 pp., 23 pls. Calcutta. —— 1866-67. A Monograph of the British trilobites from the Cambrian, Silurian and Devonian forma- tions. Part 3: 129-176, pls 15-25 (1866a); Part 4: 177-214, pls 25*-30 (1867). Palaeontogr. Soc. (Monogr.), London. —— 1866b. On the fossils of North Wales. In Ramsay, A. C., The Geology of North Wales. Mem. geol. Surv. U.K., London, 3: 239-381, pls 1-26. See Salter & Etheridge 1881. 1873. A catalogue of the collection of Cambrian and Silurian fossils contained in the geological museum of the University of Cambridge. xlvi + 204 pp. Cambridge. & Etheridge, R. 1881. On the fossils of North Wales. In Ramsay, A. C., The Geology of North Wales. Mem. geol. Surv. U.K., London, 3 (2nd edn): 331-611, pls 1-26. See Salter 1866b. Sars, M. 1835. Ueber einige neue oder unvollstandig bekannte trilobiten. Isis, Jena, Leipzig, 28 (4): col. 333-343, pl. 9. Sdzuy, K. 1955. Die Fauna der Leimitz-Schiefer (Tremadoc). Abh. senckenb. naturforsch. Ges., Frankfurt a.M., 492: 1-74, pls 1-8. Sedgwick, A. 1852. On the classification and nomenclature of the Lower Palaeozoic rocks of England and Wales. Q. JI geol. Soc. Lond., 8: 136-168. Sheng Xinfu 1974. [Ordovician trilobites from western Yunnan and their stratigraphical significance]. In: Sheng Xinfu, [Subdivision and correlation of the Ordovician System in China]: 96-143, 9 pls. Beying. [In Chinese]. Shergold, J. H. 1975. Late Cambrian and early Ordovician trilobites from the Burke River structural belt, western Queensland. Bull. Bur. Miner. Resour. Geol. Geophys. Aust., Melbourne, 153: 1-251. Shirley, J. 1931. A redescription of the known British Ordovician species of Calymene (s.l.). Mem. Proc. Manchr lit. phil. Soc., 75: 1—35, pls i, 11. Skevington, D. 1963. A correlation of Ordovician graptolite-bearing sequences. Geol. For. Stockh. Forh., 85: 298-319. —— 1965. Graptolites from the Ontikan Limestones (Ordovician) of Oland, Sweden. II. Graptoloidea and Graptovermida. Bull. geol. Instn Univ. Upsala, 43: 1-74. —— 1968. British and North American lower Ordovician correlation: discussion. Bull. geol. Soc. Am., New York, 79: 1259-1264. —— 1969. The classification of the Ordovician System in Wales. In Wood, A., The Pre-Cambrian and Lower Palaeozoic rocks of Wales: 161-179. Cardiff. —— 1970. The Lower Llanvirn graptolite fauna from the Skiddaw Slates, Westmorland. Proc. Yorks geol. Soc., Leeds, 37: 395-444, 4 figs. —— 1973. Graptolite fauna of the Great Paxton Borehole, Huntingdonshire. Bull. geol. Surv. Gt Br., London, 43: 41-57. —— 1976. British Ordovician graptolite zones and inter-regional correlation. In Kaljo, D. & Koren, T. (eds), Graptolites and Stratigraphy: 171-178. Tallinn, Estonian Acad. Sci., Geol. Inst. —— & Archer, J. B. 1971. A review of the Ordovician graptolite faunas of the West of Ireland. Ir. Nat. J., Belfast, 17: 70-78. Skjeseth, S. 1952. On the Lower Didymograptus Zone (3B) at Ringsaker, and contemporaneous deposits in Scandinavia. Norsk geol. Tidsskr., Oslo, 30: 138-182, pls 1-5. Skwarko, S. 1967. Some Ordovician graptolites from the Canning Basin, Western Australia, 1. On the structure of Didymograptus artus Elles & Wood. Bull. Bur. Miner. Resour. Geol. Geophys. Aust., Mel- bourne, 92: 171-190. Snajdr, M. 1957. Classification of the family Illaenidae (Hawle et Corda) in the Lower Palaeozoic of Bohemia. Sb. ustred. Ust. Geol., Prague, 23: 125—284. — 1976. New finds of trilobites from the Dobrotiva Formation (Llandeilan) in the Barrandian. Vést. ustred. Ust. geol., Prague, 51: 231—237. 1981. On some rare Bohemian Trinucleina (Trilobita). Vést. ustred. Ust. geol., Prague, 56: 279-285, 4 pls. —— 1984. On the supposed presence of Ectillaenus hughesi (Hicks) and Ormathops nicholsoni (Salter) in Ordovician of Bohemia. Cas. narod. Mus., Prague, 153: 21-24, 1 pl. (in Czech, English summary). Spassow, H. 1958. Ere Paléozoique. In Tzankov, V. (ed.), Les Fossiles de Bulgarie, 1. 90 pp., 16 pls. Sofia, Naukite. [In Bulgarian; Russian and French résumés ]. ARENIG IN SOUTH WALES 301 Spencer, W. K. 1918. A Monograph of the British Palaeozoic Asterozoa. Part 3: 109-168, pls 6-13. Palaeontrogr. Soc. (Monogr.), London. —— 1950. Asterozoa and the study of Palaeozoic faunas. Geol. Mag., Hertford, 87: 393-408. Spjeldnaes, N. 1953. The Middle Ordovician of the Oslo Region, Norway. 3. Graptolites dating the beds below the Middle Ordovician. Norsk Geol. Tidsskr., Oslo, 35: 171-184. Stait, B. & Laurie, J. 1983. In Burrett, C., Stait, B. & Laurie, J., Trilobites and microfossils from the Middle Ordovician of Suprise Bay, Southern Tasmania, Australia. Mem. australas. Palaeontol., Sydney, 1: 177-193. Stephens, J. O. 1941. David Cledlyn Evans, Hon. M.Sc. (Wales), F.G.S. (Schoolmaster, Geologist, Antiquary) 1858-1940. Carmarthen Antiq., 1: 11—20. Stone, P. & Rushton, A. W. A. 1983. Graptolite faunas from the Ballantrae ophiolite complex and their structural implications. Scott. J. Geol., Edinburgh, 19: 297-310. Strachan, I. 1971. A synoptic supplement to “A monograph of British graptolites by Miss G. L. Elles and Miss E. M. R. Wood”. 130 pp. Palaeontogr. Soc. (Monogr.), London. — 1986. The Ordovician graptolites of the Shelve District, Shropshire. Bull. Br. Mus. nat. Hist., London, (Geol.) 40 (1): 1-58. — & Khashogji, M. S. 1984. The type specimen of Didymograptus murchisoni. Lethaia, Oslo, 17: 223-231. Strahan, A. et al. 1907-14. The geology of the South Wales Coalfield. Part VII. The country around Ammanford. Mem. geol. Surv. U.K., London, sheet 230. viii + 246 pp. (1907). Part X. The country around Carmarthen. Loc. cit., sheet 229. viii + 177 pp. (1909). Part XI. The country around Haverford- west. Loc. cit., sheet 228. viii + 262 pp. (1914). Strusz, D. L. 1980. The Encrinuridae and related trilobite families, with a description of Silurian species from southeastern Australia. Palaeontographica, Stuttgart, (A) 168: 1-168, 6 pls. Struve, W. 1958. Beitrage zur Kenntnis der Phacopacea (Trilobita), 1: Die Zeliszkellinae. Senckenberg. leth., Frankfurt a.M., 39: 165-219, pls 1+4. —— 1975. Die altesten Fossilien Hessens. Natur Mus., Frankf. 105: 262-282. Stubblefield, C. J. 1926. Notes on the development of a trilobite, Shumardia pusilla (Sars). J. Linn. Soc., London, (Zool.) 36: 345-372, pls 14-16. —— 1927. Trilobita. Pp. 127-143, pl. 4 in Stubblefield, C.J. & Bulman, O.M.B. 1927. The Shineton Shales of the Wrekin district: with notes on their development in other parts of Shropshire and Herefordshire. Q. JI geol. Soc. Lond., 83: 96-146, pls 3—S. —— 1939. Some aspects of the distribution and migration of trilobites in the British Lower Palaeozoic faunas. Geol. Mag., London, 76: 49-72. —— 1951. New names for the trilobite genera Menevia Lake and Psilocephalus Salter. Geol. Mag., Hertford, 88: 213-214. Temple, J. T. 1980. A numerical taxonomic study of Trinucleidae (Trilobita) from the British Isles. Trans. R. Soc. Edinb., (Earth Sci.) 71: 213-233. Thomas, A. T. 1978. British Wenlock trilobites. Part 1: 1-56, pls 1-14. Palaeontogr. Soc. (Monogr.), London. , Owens, R. M. & Rushton, A. W. A. 1984. Trilobites in British stratigraphy. Spec. Rep. geol. Soc. Lond., 16. 78 pp. Thomas, D. E. 1960. The zonal distribution of Australian graptolites. J. Proc. R. Soc. N.S.W., Sydney, 94: 1-58. Thomas, H. H. & Cox, A. H. 1924. The volcanic series of Trefgarn, Rock and Ambleston. Q. JI geol. Soc. Lond., 80: 520-548, pl. 40. —— & Jones, O. T. 1912. The Pre-Cambrian and Cambrian rocks of Pembrokeshire. Q. JI geol. Soc. Lond., 68: 374400, pl. 40. Thoral, M. 1935. Contribution a l’etude paleontologique de |'Ordovicien inferieur de la Montagne Noire et revision sommaire de la faune Cambrienne de la Montagne Noire. 362 pp., 35 pls. Montpellier. 1946. Cycles géologiques et formations noduliféres de la Montagne Noire. Nouv. Archs Mus. Hist. nat. Lyon, 1: 1-103, pls 1-16. Tjernvik, T. 1956. On the early Ordovician of Sweden. Stratigraphy and fauna. Bull. geol. Instn Univ. Upsala, 36: 107-284, 11 pls. Tornquist, S. L. 1890. Urdersokning ofver Siljansomradets Graptoliter. Acta Univ. lund., 26 (4): 1-33, 2 pls. —— 1901-04. Researches into the graptolites of the lower zones of the Scanian Vestrogothian Phyllo- Tetragraptus beds. I. Acta Univ. lund., 37: 1-26 (1901). II. loc. cit. 40: 1-29 (1904). Tromelin, G. de & Lebesconte, P. 1876. Essai d’un catalogue raisonné des fossiles siluriens des départ- ments de Maine-et-Loire, de la Loire-Inferieure et du Morbihan, avec des observations sur les terrains 302 R. A. FORTEY & R. M. OWENS paléozoiques de louest de la France. C. r. Ass. fr. Avanc. Sci., Paris, 4° session (Nantes, 1875): 601-661. Tsai, A. T. 1969. [A new Ordovician genus Acrograptus.] Paleont. Zh., Moscow, 1969: 142-143. [In Russian: Engl. transl. Paleont. J.. Washington, 3 (1): 133-134]. —— 1974. [Early Ordovician graptolites of Kazakhstan.] 115 pp., 11 pls. Moscow, Akad. Nauk [In Russian ]. Turner, F. E. 1940. Alsataspis bakeri, a new Lower Ordovician trilobite. J. Palaeont., Tulsa, 14: 516-518. Turner, J. C. M. 1960. Faunas graptoliticas de America der Sur. Revta Asoc. geol. argent., Buenos Aires, 14: 5—180, pls 1-9. Ubaghs, G. 1967. Stylophora. In Moore, R. C. (ed.), Treatise on Invertebrate Paleontology, S (Echinodermata 1| (2)): S495—S565. Lawrence, Kansas. —— 1969. Les échinodermes carpoides de |’'Ordovicien inferieur de la Montagne Noire (France). 112 pp., 17 pls. Paris, C.N.R.S. (Cah. Paléont.) —— 1981. Reflexions sur le nature et la fonction de l’appendice articulé des carpoides Stylophora (Echinodermata). Annls Paleont., Paris, (Invert.) 67: 33-48, pls 1-17. Van Ingen, G. 1901. The Siluric fauna near Batesville, Arkansas. Columbia Univ. Sch. Mines Q., New York, 23: 34-74. Vodges, 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., Washington, 63: 1-177. — 1925. Palaeozoic Crustacea. Part II. An alphabetical list of the genera and subgenera of the Tri- lobita. Trans. S. Diego Soc. nat. Hist., 4: 89-115. Wadge, A. J. 1978. Classification and stratigraphical relationships of the Lower Ordovician rocks. In Moseley, F. (ed.). The Geology of the Lake District: 68-78. Leeds, Yorkshire Geol. Soc. (Occ. Publ. 3). Ward, J. C. 1876. The geology of the northern part of the English Lake District. With an appendix on new species of fossils, by R. Etheridge, F. R. S. xii + 132 pp., 13 pls. Mem. geol. Surv. U.K., London. Wei Xiu & Zhou Zhi-yi 1983. Trilobites. In Qui Hong-An et al., Palaeontological Atlas of East China, 1 (Early Palaeozoic). 657 pp., 176 pls. Beijing (Geol. Publ. House). Weir, J. A. 1959. Ashgillian trilobites from Co. Clare, Ireland. Palaeontology, London, 1: 369-383, pls 62-63. Westergard, A. H. 1946. Agnostidea of the Middle Cambrian of Sweden. Sver. geol. Unders. Afh., Stock- holm, (C) 477. 140 pp., 16 pls. Whittard, W. F. 1940. The Ordovician trilobite fauna of the Shelve-Corndon district, west Shropshire, Part I. Agnostidae, Raphiophoridae, Cheiruridae. Ann. Mag. nat. Hist., London, (11) 5: 153-172, pls 5, 6. — 1952. Cyclopygid trilobites from Girvan and a note on Bohemilla. Bull. Br. Mus. nat. Hist., London, (Geol.) 10: 305-324, pls 32-33. — 1955-67. The Ordovician trilobites of the Shelve inlier, west Shropshire. Part 1: 1-40, pls 1-4 (1955); Part 2: 41-70, pls 5-9 (1956); Part 3: 71-116, pls 10-15 (1958); Part 4: 117-162, pls 16-21 (1960); Part 5: 163-196, pls 22-25 (1961a); Part 6: 197-228, pls 26-33 (1961b); Part 7: 229-264, pls 34-45 (1964); Part 8: 265-306, pls 46-50 (1966); Part 9: 307-352 (1967). Palaeontogr. Soc. (Monogr.), London. —— 1960. Angleterre, Pays de Galles, Ecosse. Ordovicien. Lexique stratigraphique international, 1 (3aIV). 296 pp. Paris. Whittington, H. B. 1950. Sixteen Ordovician genotype trilobites. J. Paleont., Tulsa, 24: 531-565, pls 68-75. — 1952. The trilobite family Dionididae. J. Paleont., Tulsa, 26: 1-11, pls 1, 2. — 1965S. Trilobites of the Ordovician Table Head Formation, western Newfoundland. Bull. Mus. comp. Zool. Harv., Cambridge, Mass., 132: 275—442, 68 pls. — 1966. Trilobites of the Henllan Ash, Arenig Series, Merioneth. Bull. Br. Mus. nat. Hist., London, (Geol.) 11 (10): 489-505, pls 1-5. ——, Dean, W. T., Fortey, R. A., Rickards, R. B., Rushton, A. W. A. & Wright, A. D. 1984. Definition of the Tremadoc Series and the series of the Ordovician system in Britain. Geol. Mag., Cambridge, 121: 17-33. & Hughes, C. P. 1972. Ordovician geography and faunal provinces deduced from trilobite distribu- tion. Phil. Trans. R. Soc., London, (B) 263: 235-278. Williams, A. 1974. Ordovician Brachiopoda from the Shelve district, Shropshire. Bull. Br. Mus. nat. Hist., London, (Geol. Suppl.) 11. 163 pp., 28 pls. Williams, T. G. 1934. The Pre-Cambrian and Lower Palaeozoic rocks of the eastern end of the St. David’s Pre-Cambrian area, Pembrokeshire. Q. JI. geol. Soc. Lond., 90: 32-75, pl. 2. Wiman, C. 1902. Studien tiber das Nordbaltische Silurgebiet 1. Olenellussandstein, Obolussandstein und Ceratopygeschiefer. Bull. geol. Instn Univ. Upsala, 6: 12-74, pls 1-4. — 1905. Ein Shumardia-schiefer bei Lanna in Nerike. Ark. Zool., Stockholm, 2: 1-20, pls 1, 2. ARENIG IN SOUTH WALES 303 Xiang Liwen & Zhang Tairong 1984. Tremadocian trilobites from western part of northern Tianshan, Xinjiang. Acta palaeont. sin., Beijing, 23: 399-410, pls 1-3. Yi Yongen 1957. The Caradocian trilobite fauna from the Yangtze Gorges. Acta palaeont. sin., Beijing, 5: 545-559, 5 pls. Yin Gongsheng & Li Shanji 1978. Trilobita. In: [Fossils of South-west China: Guizhou], 1 (Cambrian— Devonian periods): 385-594, pls 144-192. Beijing. [In Chinese]. Zalasiewicz, J. 1984a. A re-examination of the type Arenig Series. Geol. J., Liverpool, 19: 105-124. —— 1984b. Dichograptid synrhabdosomes from the Arenig of Britain. Palaeontology, London, 27: 425— 429. Zhang Tairong 1983. Trilobita. In: [Atlas of Palaeontology of North-west China: Precambrian—Lower Palaeozoic]: 134-213, pls 54-79. Beijing. [In Chinese]. —,, see also Chang Zhou Tienmei 1977. Cyclopygidae. In: [Palaeontological Atlas of Central and Southern China], 1 (Early Palaeozoic): 229-232. Beijing. [In Chinese]. Zhou Zhi-Yi 1976. See Lu et al. 1976. —— & Zhang Jinlin 1984. Uppermost Cambrian and lowest Ordovician trilobites of north and northeast China. In: Papers for the symposium on the Cambrian—Ordovician and Ordovician—Silurian boundaries, Nanjing, China. October, 1983: 25-30, pls 1—3. Nanjing. Ziegler, A. M., Scotese, C. R., McKerrow, W. S., Johnson, M. E. & Bambach, R. K. 1979. Paleozoic palaeogeography. Ann. Rev. Earth Planet. Sci., Palo Alto, 7: 437—S02. Index New taxonomic names and the page numbers of the principal references are in bold type. Aber Mawr Formation 96 Arthrorhachis danica 115 Abercastle Formation 71, 74, 94, 96 Arthrorhachis elliptifrons 115 Acrograptus acutidens 82, 278 Arthrorhachis hebetatus 115 Acrograptus lipoldi 281 Arthrorhachis hupehensis 115 Acrograptus? sp. a 280 Arthrorhachis lentiformis 115 Afon Ffinnant Formation 71, 78, 96, 104 Arthrorhachis pragensis 115 Afon Seiont, Caernarfon 90, 97 Arthrorhachis saltaensis 115 Agnostus hirundo 116 Arthrorhachis tarda 114 Amicus 180 Arthrorhachis sp. indet. 114 Ampyx abnormis 225 Asaphellus 132 Ampyx cetsarum 83, 96-7, 102, 228-9 Asaphellus graffi 135 Ampyx cf. reyesi 226, 230 Asaphellus lugneensis 135 Ampyx linleyensis 223, 225, 226, 227 Asaphellus whittardi 97-8, 132 Ampyx linleyoides 106, 223, 227 Asaphoon 132 Ampyx nasutus 223, 225 Asaphus corndensis 142 Ampyx pallens 226 Aspidaeglina 149, 180 Ampyx reyesi 226 atheloptic trilobite assemblage 71, 104, 106 Ampyx salteri 228 Azygograptus eivionicus 83, 276 Ampyx spongiosus 225 Azygograptus hicksti 84, 275 Ampyx volborthi 225 Azygograptus lapworthi 277 Ampyx yii 226 Azygograptus suecicus 277 Anatifopsis 287 Anebolithus 203-4 Balanocystites 287; sp. 287 Anebolithus simplicior 204, 207 Baltica 108 Anglesey 71, 97, 108 Barrandia bianularis 193 apertural angle 251 Barrandia cordai 191 Araiocaris 136 Barrandia cf. cordai 193 Arenig-Llanvirn boundary 82, 90 Barrandia homfrayi 191 Arennig Fawr 71-2, 97, 99 Barrandia sp. indet. 193 Arthrorhachis abruptus 115 Bendigonian Stage 101 Arthrorhachis chinianensis 115 Bergamia artemis 207 Arthrorhachis corpulentus 115 Bergamia gibbsi 216 304 R. A. FORTEY & R. M. OWENS Bergamia inquilina 205 Bergamia matura 205 Bergamia rhodesi 204-5, 210 Bergamia rushtoni 205; Biozone 84 Bergamia sp. A 207 Bicyclopyge 181 Bienvillia praecalva 83 Blaencediw Formation 71, 75, 78, 94 Bohemia 102 Bohemilla 128 Bohemilla (Bohemilla) stupenda 128 Bohemilla (Bohemilla) tridens 130 Bohemilla (Fenniops) klouceki 131 Bohemilla (Fenniops) praecedens 131 Bohemilla (Fenniops) sabulon 128, 129 Bohemopyge 135 Bohemopyge discreta 142 Bohemopyge scutatrix 79, 84, 88, 96-7, 136 Bolahaul Member 96 Borthaspidella 194 Borthaspis 194 Brunel Beds 71, 74, 94 Bumastus barriensis 199 Carmarthen 96, 102; Formation 78, 87, 96, 104 Carmel Formation 97 Castelldraenog Member 79 Castlemainian stage 101 Circulocrania 179 Circulocrania orbissima 186, 187 Clelandia 126 Clelandia reliqua 126 Cnemidopyge salteri 84, 96, 227, 228 Cochliorrhoe 203 Colomendy Formation 71,78, 96 Colpocoryphe deani 242 Colpocoryphe maynardensis 242 Colpocoryphe taylorum 106, 241 Colpocoryphe thorali 242 Colpocoryphe thorali conjugens 243 Conophrys salopiensis 119, 126 Coplacoparia 231 Cornovica 234 Cornovia didymograpti 234 Corrugatagnostus chekiangensis 114 Corrugatagnostus convergens 114 Corrugatagnostus fortis 114 Corrugatagnostus Jiangshanensis 114 Corrugatagnostus cf. refragor 113 Corrugatagnostus sol 114 Corrugatagnostus transitus 114 Corymbograptus retroflexus 272 Cothurnocystis sp. 287 Cremastoglottos 126 Cwmfelin Boeth Formation 71, 81, 104 Cyclopyge 149 Cyclopyge alia 155 Cyclopyge festa 155 Cyclopyge genatenta 176 Cyclopyge grandis brevirhachis 154 Cyclopyge grandis grandis 96, 151 Cyclopyge kossleri 155 Cyclopyge marginata 155 Cyclopyge pachycephala 174 Cyclopyge stigmata 156 Cyclopyge umbonata 156 Cyclopyge cf. umbonata 156 cyclopygid biofacies 105, 106 cystoids 104 Degamella 149 Degamella azaisi 158 Degamella evansi 96, 157 Degamella gigantea 160 Degamella nuda 160-1 Degamella princeps praecedens 160 Degamella princeps princeps 159 dendroid graptolites 78, 84, 96, 104, 251 Dichograptus Zone 74 Didymograptellus 254 Didymograptus ‘bifidus’ Biozone 72, 90 Didymograptus (Didymograptellus) sp. 260 Didymograptus (Didymograptus) artus 258; Biozone 72, 90 Didymograptus (Didymograptus) spinulosus 255-6, 258 Didymograptus (Expansograptus) goldschmidti 272 Didymograptus (Expansograptus) hirundo 98, 260; Biozone 89, 101 Didymograptus (Expansograptus) nitidus 264, 273; Biozone 102 Didymograptus (Expansograptus) patulus 269 Didymograptus (Expansograptus) sparsus 267 Didymograptus (Expansograptus?) uniformis lep- idus 270 Didymograptus (Expansograptus?) uniformis unt- formis 270 Didymograptus abnormis 263 Didymograptus alatus 275 Didymograptus barrandei 257 Didymograptus bifidus 260 Didymograptus bifidus incertus 257 Didymograptus changningensis 259 Didymograptus deflexus Biozone 99 Didymograptus gracilis 281 Didymograptus oligotheca 257 Didymograptus patulentis 275 Didymograptus pennatulus 267 Didymograptus pluto 254-S Didymograptus protobifidoides 260 Didymograptus simulans 272 Didymograptus stabilis 257, 259 Dindymene cf. didymograpti 237 Dindymene fridericiaugusti 234 Dindymene hughesiae 234 Dindymene longicaudata 234 Dindymene saron 106, 235 ARENIG IN SOUTH WALES 305 Dindymenella 235 Dionide (Paradionide) 219 Dionide formosa 218 Dionide jubata 218 Dionide \evigena 220; Biozone 84 Dionide magnifica 218 Dionide prima 218 Dionide turnbulli 219 Dionideina 218 Dionidella incisa 221 Dionidella? sp. indet. 1 222 Dionidella? sp. indet. 2 223 Dionidepyga 218 Ectillaenus benignensis 200 Ectillaenus bergaminus 200; ? bergaminus 202 Ectillaenus cunicularis 200 Ectillaenus giganteus 202 Ectillaenus hughesii 200 Ectillaenus perovalis 199 Ectillaenus sarkaensis 200 Ellipsotaphrus infaustus 188 Ellipsotaphrus monophthalmus 189 Ellipsotaphrus pumilio 188 Ellipsotaphrus whittardi 190 Ellipsotaphrus zhongguoensis 188 Emmrichops 169, 179 Eoshumardia 120 Eurymetopus 193 Eurymetopus cumbrianus 194 Eurymetopus harrisoni 194 Extensograptus 260 eye reduction 106 Famatinolithus 216 Fennian Stage 88, 96, 98, 101, 104 Fenniops 128 Furcalithus 203 Furcalithus radix 83, 207, 208; Biozone 83 Furcalithus sedgwicki 96, 209 Gallagnostoides 172 Gamops 126 Gastropolus 149 Gastropolus brevicaudatum 161 Gastropolus mirabilis 163 Gastropolus obtusicaudatus 161 Geragnostus merus 116 Girvanopyge barrandei 126 Girvanopyge caudata 126 Girvanopyge mrazeki 126 Girvanopyge occipitalis 126 Girvanopyge sp. 126 Glossograptus acanthus 281 Glyptograptus 282 Glyptograptus austrodentatus 284 Glyptograptus austrodentatus major 285 ‘Glyptograptus’ dentatus 282 Glyptograptus robustus 285 Glyptograptus shelvensis 284 Glyptograptus situlus 285 Gondwana 93, 99, 105, 108 Gres Armoricain 109 Guichenocarpos sp. 287 Gymnostomix 215 Gymnostomix gibbsii 79, 81, 84, 96, 216; Biozone , 84 Hanchungolithus 203 Hawleia 231 Hemigyraspis 132 Henllan Ash 97 Hope Beck Slates 98 Hypermecaspis venerabilis 83 Illaenopsis griffei 194, 199 Illaenopsis harrisoni 97-8, 106, 194 Illaenopsis primula 195, 199 Illaenopsis stenorhachis 194 Illaenopsis thomsoni 193, 198 Illaenus Hughesii 199 Illaenus perovalis 199 Incaia 203 Incisopyge 174 Incisopyge? theroni 152 Isograptus gibberulus Biozone 89, 101 Klabava Formation 102 Kweichowilla 119 Kweichowilla minuta 119 Lagynocystis pyramidalis 287 Lagynocystis sp. 287 Lake District 98 Leiagnostus alimbeticus 113 Leiagnostus bohemicus 113 Leiagnostus erraticus 112 Leiagnostus cf. erraticus 112 Leiagnostus foulonensis 113 Leiagnostus franconicus 113 Leiagnostus peltatus 113 Leiagnostus turgidulus 113 Leioshumardia minima 126 Leioshumardia sp. A 126 Lichapyge? problematica 126 Lisogoraspis 161 Llanfallteg Formation 71, 82, 105 Llyfnant Member 97 Llyn Peninsula 71, 97 Lordshillia 203 Loweswater Flags 98 Megalaspidella 132 Megalaspidella whittardi 132 Megalaspis mucronata 135 Megalaspis striatula 135 Merlinia murchisoniae 83, 93, 96-7, 135 Merlinia rhyakos Biozone 83 306 R. A. FORTEY & R. M. OWENS Merlinia selwynii 99, 102; Biozone 83 mesopelagic 105, 180 Microparia 149 Microparia (Heterocyclopyge) 149 Microparia (Heterocyclopyge) shelvensis 167 Microparia (Heterocyclopyge?) sp. indet. 174 Microparia (Microparia) boia 172 Microparia (Microparia) broeggeri 102, 164, 173 Microparia (Microparia) lusca 167 Microparia (Microparia) porrecta 168 Microparia (Microparia) teretis 170 Microparia (Quadratapyge) 149, 164 Microparia brachycephala 166, 172 Microparia laevis 172 Microparia plasi 166 Microparia speciosa 164, 166 Microparia? sp. indet. 1 173 Mitrocystella barrandei 287 Mitrocystella incipiens 287 Mitrocystella sp. 287 Mitrocystites sp. 287 Monorthis menapiae 83 Montagne Noire 99, 108, 290 Moridunian 74, 84, 93, 98-9, 102 Mytton Flags 98 Myttonia fearnsidesi 83 Nanlingia 126 Neptunagnostella 112 Neseuretus biofacies 102, 108 Neseuretus brevisulcus 238 Neseuretus complanatus 238 Neseuretus cf. complanatus 240 Neseuretus grandior 238 Neseuretus monensis 97 Neseuretus murchisoni 96-7, 238 Neseuretus ramseyensis 83, 93, 96, 237, 238 Nileus 193, 197 Niobe doveri 190 Niobe emarginula 139 Niobe? huberi 107 Niobina davidis 136, 142 Novakella 149 Novakella bergeroni 176 Novakella copei 174 Novakella incisa 176 Ogof Hén Formation 93, 96, 71, 74 Ogof Velvet Formation 93 Ogygia bullina 136 Ogygia discreta 135 Ogygia peltata 136 Ogyginus 142 Ogyginus armoricanus 142 Ogyginus hybridus 79, 83-4, 88, 94, 97, 142, 143 Ogyginus orbensis 142 Ogyginus planus 142, 148 Ogyginus terranovicus 142 Ogyginus sp. indet. 148 Ogygiocaris araiorhachis 136 olenid biofacies 104, 108 Ormathops alata 247 Ormathops atava 243 Ormathops barroisi 246 Ormathops borni 243 Ormathops llanvirnensis 243-4, 247 Ormathops nicholsoni 96-8, 106, 244 Paralenorthis alata 83, 96 Paralenorthis proava 97 Penmaen Dewi Formation 71, 73, 96, 104 Phacops Nicholsoni 244 photophores 180 Phylacops 189 Pibwr Member 96 Placoparia (Placoparia) armoricensis 234 Placoparia (Placoparia) cambriensis 232 Placoparina sp. 231 Plasiaspis 235 Plesiomegalapsis 132 Plesiomegalaspis angustirhachis 135 Plesiomegalaspis? convexilimbata 135 Pontyfenni Formation 71, 81, 184 Poronileus 197 Porterfieldia punctata 83 Porth Gain Beds 94 Porth Gain Formation 74 Pricyclopyge 179 Pricyclopyge binodosa binodosa 181 Pricyclopyge binodosa eurycephala 97, 181 Pricyclopyge binodosa longicephala 181 Pricyclopyge binodosa prisca 181 Pricyclopyge? campestris 184 Pricyclopyge dolabra 184 Pricyclopyge obscura 184-5 Pricyclopyge sichuanensis 184 Pricyclopyge synophthalma 184 Pricyclopyge wattisoni 172, 184 Pricyclopyginae subfam. nov. 179 Procephalops 193 Procephalops hopense 198 Prosopiscus 235 Prospectatrix 151 Prospectatrix genatenta 177 Prospectatrix cf. superciliata 176 Protolloydolithus 203 Pseudobarrandia 194 Pseudophyllograptus densus 99 Pseudotrigonograptus ensiformis 278 Psilacella doveri 190 Psilacella hunanensis 190 Psilacella pulchra 187, 190 Psilacella trirugata 190 Psilocephalinella 194 Pterygometopus 243 Ramsey Island 71, 74, 93, 105 raphiophorid biofacies 102 ARENIG IN SOUTH WALES 307 Reticulocarpos sp. 287 Rhyd-Henllan Member 78 Road Uchaf Formation 96 Rokycania 193 Sagavia 149 Sagavia elongata 178 Sagavia felix 177-8 Sagavia glans 177 Sagavia heterocyclopygeformis 178 Sagavia modica 178 Sagavia novakellaformis 178 Scolton 83 Segmentagnostus hirundo 116 Segmentagnostus mccoyii 118 Segmentagnostus merus 118 Segmentagnostus neumanni 118 Segmentagnostus scoltonensis 116, 118 Segmentagnostus stubblefieldi 116 Segmentagnostus whitlandensis 84, 116 Seleneceme acuticaudata 230 Seleneceme propinqua 230 Selenopeltis buchi buchi 249 Selenopeltis buchi macrophthalma 98, 250 Selenopeltis inermis 249 Shelve 97 Shelve Church Beds 98 Shelve Inlier 71 Shumardia (Conophrys) 120 Shumardia (Conophrys) bottnica 120 Shumardia (Conophrys) changshanensis 120 Shumardia (Conophrys) crossi 123, 125 Shumardia (Conophrys) keguqinensis 120 Shumardia (Conophrys) nericiensis 120 Shumardia (Conophrys) oelandica 120 Shumardia (Conophrys) pusilla 120 Shumardia (Conophrys) salopiensis 120 Shumardia (Kweichowilla) 120 Shumardia (Kweichowilla) acuticaudata 120 Shumardia (Kweichowilla) forbesi 120 Shumardia (Kweichowilla) hongyaensis 120 Shumardia (Kweichowilla) lacrimosa 120 Shumardia (Kweichowilla) matchensis 120 Shumardia (Kweichowilla) minuta 120 Shumardia (Kweichowilla) sagittula 120 Shumardia (Shumardella) 119-20 Shumardia (Shumardella) bohemica 119-20 Shumardia (Shumardella) extensa 120 Shumardia (Shumardella) phalloides 120 Shumardia (Shumardella) polonica 120 Shumardia (Shumardella) scotica 120 Shumardia (Shumardella) tenacis 120 Shumardia (Shumardia) 120 Shumardia (Shumardia) dicksoni 120, 123 Shumardia (Shumardia) gadwensis 84, 88, 96, 120, 121, 125 Shumardia (Shumardia) granulosa 119-20, 123 Shumardia (Shumardia) lacrima 120, 123 Shumardia (Shumardia) tarimuensis 120 Shumardia (Shumardia) sp. A 123 Shumardia orientalis 120 Skiddaw Group 98 St David’s 73, 93, 102 Stapeleyella 204 Stapeleyella abyfrons 84, 204, 213; Biozone 84 Stapeleyella aff. abyfrons 215 Stapeleyella etheridgei 213 Stapeleyella inconstans 211, 212, 214 Stapeleyella murchisoni 213 Stiperstones Quartzite 98 Symphysops 179 Tankerville Flags 98 Tetragraptus approximatus Biozone 74, 99 Tetragraptus bigsbyi askerensis 252 Tetragraptus bigsbyi bigsbyi 252 Tetragraptus pseudobigsbyi 253 Tetragraptus (Tetragraptus) reclinatus abbreviatus 253 Tetragraptus (Tetragraptus) reclinatus reclinatus 252 Tetragraptus reclinatus toernquisti 253 Tetragraptus serra 101, 251 ‘Tetragraptus Shales’ 73 thecal inclination 251 Tornquist’s Ocean 105 Traveusot Formation 290 Treiorwerth Formation 97 Trinucleoides 221 Trinucleus gibbsii 215 Trinucleus Sedgwicki 209 Trwynhwrddyn 94 Undulograptus 282 Whitland 74; Abbey Member 79 Whitlandian Stage 87, 94, 98, 101, 104 Xenocyclopyge 149 Xiphograptus 275 Accepted for publication 11 November 1985 i , — 7 , \ — a “ae a => ‘ rT , = ‘ e ; % ‘ b. ey F nl \ 7 ti i Mie” isr “ rua mtn 4 W pays = Tee df ~ Aan f “i . - Appendix. Acritarchs and Chitinozoa from the Arenig Series of south-west Wales S. G. Molyneux British Geological Survey, Keyworth, Nottingham NG12 5GG. Contents SHiMODIS voasocoooogooscaogc0csscaso0 cncqag asad anon qDONSADGODOOADaDObAOAOHAGEDONA 310 LiMPOCINCHIOM. socacassadanoadassondadononodoqbusanddaGaKHEoAaGanadoDadeSoBooadadoBAd 310 PLAY ANG LO Bymmerererersterererotatele ee lclesereleloteteleeCetoTelevelateteleievaictelelsieteteialecctetelelercieteieietaicieistactereleieleicieieleisicis 313 SAMO sogesaosoonacccnb000on 0s 0ng09daaaGog0 ND NBOODOORAdOODODDODROAOnOBORGDG 313 BOS tral bi reap lay eters Geer aR RABE AGES eS BAR Ano cacumonacubnrncncerceuetichosucnsncascadaaneeccecncas 363 Synopsis Thirty-two samples were collected from the Arenig Series of south-west Wales to assess the occurrence and stratigraphical distribution of microfossils. All the samples yielded acritarchs and 6 also yielded chitinozoa. The microfossils are generally poorly preserved and rare, many taxa being represented by single specimens. The species recorded from 14 samples are grouped into 7 assemblages (Microfossil Assemblages I-VII). Assemblages I-IV are of Moridunian (lower Arenig) age. Assemblage I occurs in the Allt Cystanog Member of the Ogof Hén Formation, assemblages II and III are present in the Cwmffrwd Member of the Carmarthen Formation, and assemblage IV occurs in the Cwm yr Abbey Member of the same formation and at the base of the overlying Afon Ffinnant Formation. Assemblage V is of Whitlandian (middle Arenig) age, occurring in the Whitland Abbey Member of the Colomendy Formation. Assemblages VI and VII are both of Fennian (upper Arenig) age, and occur in the Pontyfenni Formation. Biozones are not formally defined, but published and unpublished sources suggest that at least some of the microfossil assemblages from south Wales are comparable with those from Arenig rocks of northern England, north Wales and western Europe. One new acritarch species Stellechinatum papulessum is described and one new combination Stellechina- tum uncinatum (Downie) is proposed. Several acritarch and one chitinozoan species are described under open nomenclature. Introduction Fortey & Owens (1978, 1987) have demonstrated the presence of a complete and fossiliferous Arenig succession in south-west Wales, extending from east of Carmarthen to Ramsey Island ACRITARCHS AND CHITINOZOA 311 Ramsey Island armarthen ame Haverfordwest Cardiff ) 40 km S Fig. 1 Outcrop of Arenig rocks in south-west Wales (stippled). (Fig. 1). Acritarchs and chitinozoa have been recorded from Arenig rocks in the Carmarthen— Whitland area; in this paper, their stratigraphical distribution is described. Lower Arenig lithostratigraphy in the Carmarthen area is as follows (Fortey & Owens 1978, 1987): minimum thickness Afon Ffinnant Formation ? Cwm yr Abbey Member 45m Carmarthen Formation jcxni Member c. 70m Pibwr Member 85m mee ee Member 50m Oeok Hem Faunation Allt Cystanog Member 25m The Allt Cystanog Member of the Ogof Hén Formation comprises conglomerates, sand- stones and siltstones. Its contact with the underlying Tremadoc rocks is nowhere exposed but it becomes finer upwards, passing through a transition into the micaceous mudstones and shales of the Bolahaul Member. The lowest 5m of the Pibwr Member, comprising the lowest beds of the Carmarthen Formation, are transitional in character with the Bolahaul Member. Above that, the Pibwr Member comprises black, well-bedded mudstones. Above the Pibwr Member, the Carmarthen Formation is divided into the Cwmffrwd Member, consisting of turbidites and shales, and the Cwm yr Abbey Member, comprising grey, poorly bedded mudstones. The latter are overlain by turbidites of the Afon Ffinnant Formation. 3A S. G. MOLYNEUX The Afon Ffinnant Formation is considered to be equivalent to the Blaencediw Formation of the Whitland area, where the middle and upper Arenig sequence is as follows (Fortey & Owens 1987): minimum thickness Llanfallteg Formation (in part) 100m Pontyfenni Formation 300m Cwmfelin Boeth Formation 100m Whitland Abbey Member 200 m Colomendy Formation | Castle Member c. 150m Rhyd Henllan Member c. 150m Blaencediw Formation 80m The Blaencediw Formation consists of poorly graded turbidites and channelled mass flow deposits, gritty shales and siltstones, and occasional black shales. Its base is not seen. The Colomendy Formation is divided into the sandy and silty shales of the Rhyd Henllan Member, the grey, fissile shales of the Castelldraenog Member and the black, poorly fissile shales of the Whitland Abbey Member. The latter are overlain by the well graded turbidites and black shales of the Cwmfelin Boeth Formation. The Pontyfenni Formation consists of black or dark grey shales and poorly fissile mudstones, passing upwards into the light grey mudstones and shales of the Llanfallteg Formation. The base of the Llanvirn Series lies within the latter. Fortey & Owens (1987) define seven trilobite assemblage biozones in the Arenig Series of south-west Wales. The Merlinia selwynii Biozone is well developed in the Bolahaul and Pibwr Members. The Merlinia rhyakos Biozone occurs in the Cwmffrwd and Cwm yr Abbey Members and in the lowest 40m of the Afon Ffinnant Formation. The Furcalithus radix Biozone is restricted to the Afon Ffinnant Formation and probably also to the Blaencediw Formation. The base of the succeeding Gymnostomix gibbsii Biozone is presumed to lie within the lower half of the Afon Ffinnant Formation and close to the boundary of the Blaencediw and Colo- mendy Formations. The Stapeleyella abyfrons Biozone is represented by faunas from several localities in the basal Pontyfenni Formation but the base of the Biozone is arbitrarily taken at the base of the Cwmfelin Boeth Formation. The upper two-thirds of the Pontyfenni Formation is included in the Bergamia rushtoni Biozone and the Arenig part of the Llanfallteg Formation in the Dionide levigena Biozone. (See Fig. 4.) On the basis of the trilobite faunas, three major divisions of the Series have been recognized. The base of the lower Arenig Moridunian Stage, incorporating the M. selwynii and M. rhyakos Biozones, has still to be defined. The base of the succeeding Whitlandian Stage is placed 40m above the base of the Afon Ffinnant Formation and coincides with the base of the F. radix Biozone. The base of the upper Arenig Fennian Stage is defined at the base of the Cwmfelin Boeth Formation and is arbitrarily correlated with the base of the S. abyfrons Biozone. The base of the Llanvirn Series, defining the top of the Fennian, is taken at the first appearance of pendent didymograptids in the type section of the Llanfallteg Formation. The Arenig rocks of south-west Wales were deposited at the edge of the Gondwanan conti- nent (Fortey & Owens 1984). The Ogof Hén Formation comprises shallow water sediments, deposited during the initial phase of the Arenig transgression and containing the Neseuretus Community, an association of inshore trilobites (Fortey & Owens 1978). In the Pibwr Member, this association is replaced by the Raphiophorid Community, suggesting deeper water, which in turn is replaced in the Cwmffrwd and Cwm yr Abbey Members by the Olenid Community, indicating deep, oxygen-deficient conditions. It is believed that the Carmarthen area was the site of a stagnant basin with restricted oceanic circulation, separated from the open ocean by a positive, fault-bounded block or blocks in the Haverfordwest district. The turbidites of the Afon Ffinnant and Blaencediw Formations mark the end of the restricted Olenid basin, the abun- dance of dendroid graptolites at certain horizons in the Blaencediw Formation suggesting quiet, shallow, oxygenated conditions. In the later Whitlandian, the trilobite and graptolite faunas provide evidence for an open oceanic environment. This environment persisted into the ACRITARCHS AND CHITINOZOA 313 Fennian, predominantly a time of mud deposition throughout south Wales with local turbidite sedimentation represented by the Cwmfelin Boeth Formation in the Whitland area. The faunal evidence suggests that the Pontyfenni Formation may have been deposited at a depth of 300m or more, and the Llanfallteg Formation at shallower depth but probably more than 200m. There is every indication that sedimentation was continuous throughout the Arenig and across the Arenig—Llanvirn boundary. Palynology Thirty-two samples have yielded acritarchs and six have also yielded chitinozoa. Abundance and diversity are generally low and preservation is poor; much of the material is heavily carbonized, opaque and brittle. Several specimens are distorted by the internal growth of crystals, probably of pyrite. Sampling Full details of sample localities (Figs 2, 3) are given on pp. 359-60. Ogof Hén Formation. Five samples were collected from this formation, one (MPA 20074) from the top of the Allt Cystanog Member and four (MPA 20075—6, 20079-80) from the Bolahaul Member. All yielded microfossils. Carmarthen Formation. Four samples (MPA 20077-8, 20086—7) were collected from the Cwmffrwd Member and nine (MPA 20081—5, 20088—90, 20103) from the Cwm yr Abbey Member. All yielded acritarchs. No samples were collected from the Pibwr Member. Afon Ffinnant Formation. Two samples (MPA 20104—5S) from the base of this formation yielded acritarchs. Colomendy Formation. Five samples (MPA 20094-8), all yielding acritarchs, were collected from the Whitland Abbey Member. No samples were collected from the Rhyd Henllan and Castell- draenog Members nor from the underlying Blaencediw Formation. Pontyfenni Formation. Four samples (MPA 20099-102) were collected immediately above the base and three (MPA 20091-3) from about the middle of this formation. All yielded micro- fossils. No samples from the Cwmfelin Boeth or Llanfallteg Formations were examined. Biostratigraphy The definition of acritarch biozones is unjustified because sampling and recording of species is incomplete, but the microfossils can be grouped into seven assemblages (I—VII). The strati- graphical position of these assemblages is shown in Fig. 4 and the occurrence of microfossil taxa in Fig. 5. Microfossil Assemblage I. This is present in MPA 20074 and includes the acritarchs Acantho- diacrodium aff. spinum Rasul, ?Coryphidium minutum Cramer & Diez, Micrhystridium aff. acu- minosum Cramer & Diez, Polygonium sp. A, ?Uncinisphaera? sp. D, Veryhachium minutum Downie, ?Vogtlandia flosmaris (Deunff) Dean & Martin and species of Micrhystridium, Petein- osphaeridium and Stelliferidium. One chitinozoan and several scolecodonts were recorded but have not been determined. Microfossil Assemblage II. This occurs in MPA 20077 from the middle of the Cwmffrwd Member in Nantycaws dingle (Owens & Fortey 1982; Fig. 2C herein) and is dominated by acanthomorphitic acritarchs. Species present include Acanthodiacrodium aff. angustum (Downie) Combaz, Baltisphaerosum? sp., Cymatiogalea? sp., ?Polygonium sp. A, Stelliferidium sp., Uncinisphaera? sp. D, Uncinisphaera? sp. E and the ‘Veryhachium trispinosum group. The presence of Uncinisphaera? sp. E and ‘V. trispinosum’ and the absence of M. aff. acuminosum distinguishes this assemblage from Microfossil Assemblage I. 314 S. G. MOLYNEUX | rll ‘AA <; | vl Ait i Wi : lone = ten amit engi T LOU aang Te” ye af " i q bad , ~ e Mil An f nu Wilt 2 a hy, am m pee et il Sau PT mem C muill J} HI nh Mh, yy) Mes AN TTT Mi) AAA WANA Wit i) il Hunan ih Hi VA | AAAI H il —MPA 20074 MPA 20075 MPA 20076 185 metres Aig, MPA 20079 MPA 20080 MPA 20105 MPA 20086 MPA 20087 MPA 20088 MPA 20080 Nant-y- glasdwr- fich @ Gwynion = ge Dale N, <2, Gi, ¢ AN ) OTA ES a 2 MPA 20077 3 ODDHO zs Cwmffrwd 3 seoge MPA 20078 re ANNA a o S sseée SALLE Se, oN SP 450 Fig. 2 Sample localities east of Carmarthen. Insets B—D drawn to same scale as A. Microfossil Assemblage III. One sample (MPA 20087) from the top of the Cwmffrwd Member in Nant y Glasdwr yielded a poor microflora in which a number of acritarch taxa are represent- ed by single specimens. Species present include Acanthodiacrodium sp. A, Barakella sp. A, Coryphidium? sp. A, Peteinosphaeridium sp., Stelliferidium sp. and cf. Uncinisphaera’ sp. D. Microfossil Assemblage IV. This occurs in samples from the Cwm yr Abbey Member (MPA 20084, 20103) and the base of the Afon Ffinnant Formation (MPA 20104), comprising a rich ACRITARCHS AND CHITINOZOA 315 .Boeth > . o ° --) =a _:Whitland eo Pe [| 3 ie) ~ 5 a = ° os v ’ f Afon oe T 195 = toe O metres 300 MPA 20102 MPA 20101 lesooussA SS << (e] A40 (ola [ase to 8A Sy 195 =| He MPA zoge3 x vT MPA 20100 g 2 MPA aon MPA 20099 Ey 8 4 to} < a ou = a C MPA 20091 = Whitland Abbey MPA 200696 we Cwmfelin ‘MPA’ 20097 r Boeth Penlan MPA 20098 Fig. 3. Sample localities in the Whitland area. Insets B and C drawn to same scale as A. and diverse assemblage that is dominated by acanthomorphitic acritarchs. Micrhystridium aff. henryi Paris & Deunff, M. cf. inconspicuum aremoricanum Paris & Deunff and M. aff. nann- acanthum Deflandre distinguish this assemblage from others. Other important taxa are Poly- gonium sp. B, Solisphaeridium sp. B and Uncinisphaera? sp. F. Microfossil Assemblage V. This is present in the Whitland Abbey Member (MPA 20098) and is distinguished by the dominance of small acanthomorphitic acritarchs, including Micrhystridium spp. A-D. Microfossil Assemblage VI. This occurs immediately above the base of the Pontyfenni Forma- tion (MPA 20099-102). It comprises a diverse microflora distinguished from other assemblages by the presence of Coryphidium bohemicum Vavrdova, ?Frankea hamata Burmann, Ortho- sphaeridium sp., Stellechinatum unicinatum (Downie) comb. nov., ?Striatotheca mutua Burmann, ‘SBOIE PUL[UYA\ PUL UoYyJIeUIIeD 94) UI sode[quiasse [IssojOINIW pue sajdwes jo uONNQIsIp jeorydeisyeNs sy pp “BI yL00z Wd Bourisho ily 94002-S2002 vVdW 08002-64002 VWdW 4aW Ineye|og HUAMIAaS 82002-22002 vdW SQW PAsWMD 28002-98002 VdW 06002-88002 Vd ; $800z-1800z2 WdW 44 Asqqy 1A wM9 CARMARTHEN FM] OGOF HEN FM. €0l0e VdW SOLOc-vOLOe VWdW ! 0) * ‘W4 MIGSZONAVI1d ‘Wa LNVNNIS3 NOSV = 2 c 2 oO Z = 3 iS 4QW UeUeH PAY] (5 — > oO ie) eI = = 4IQW Boueeipyjayseg w usqqib 5 © QW gs (4qW Aeqqy puelyiyM)) ws 8600¢-r6002 Vd Aegay pueUM S) ‘Ws AGNAWO109 ‘W4 HL308 NITSSAWMO suossAge c0102-66002 VWdW ¢ by ‘Wd INNSSALNOd = £6002-16002 vd ‘W4 INNSSALNOd juoqysns ie snyie -g snjie gq peulyep) you NYIANYV 17 yyIM seyeyus yum saleys ebejqwuossy sojdwes Aydesibiyesjysouyiy ebejquessy sejdwes Aydesbiyesysoyyiy [ISSOJOJOIW 1WSSOJOJDIW s9uozolg sabeys PUeIIIUM uayyewsed aygojiy S8113S 316 ACRITARCHS AND CHITINOZOA 317 2S. rarirrugulata (Cramer, Kanes, Diez & Christopher) Eisenack, Cramer & Diez, and Uncini- sphaera? spp. A and B. Rare chitinozoa and scolecodonts are present in MPA 20099. Microfossil Assemblage VII. Three samples (MPA 20091-3) from the type section of the Ponty- fenni Formation yielded diverse and abundant microfossils, including the acritarchs Cory- phidium bohemicum Vavrdova, Dasydorus cirritus? Playford & Martin, Orthosphaeridium ternatum (Burmann) Eisenack, Cramer & Diez, Solisphaeridium sp. A, Stellechinatum papulessum sp. nov, Stelliferidium aff. fimbrium (Rasul) Rasul and Uncinisphaera? sp. C. This assemblage also has a more abundant and diverse chitinozoan fauna than any other, comprising Belonechi- tina spp., Conochitina cf. chydaea Jenkins, Lagenochitina sp. A and L. cylindrica? Eisenack. Discussion The samples from the Bolahaul Member yielded rare, small and simple acanthomorphitic acritarchs whose preservation is too poor for identification. The ‘Veryhachium trispinosum’ group may be a useful biostratigraphical marker in the Mori- dunian. In south-west Wales it appears in Microfossil Assemblage II from the Cwmffrwd Member of the Carmarthen Formation, and in eastern Newfoundland its earliest recorded occurrence is approximately 500m above the base of the Arenig (Dean & Martin 1978). However, Martin (1982) showed that it also occurred in the Tremadoc, which suggests that the taxonomy and biostratigraphy of forms placed in the group will have to be revised and clarified before its significance can be appreciated. Reworked Cambrian and Tremadoc acritarchs, including Cymatiogalea bellicosa Deunff, Timofeevia lancarae (Cramer & Diez) Vanguestaine, Vulcanisphaera britannica Rasul and V. turbata? Martin, occur at the top of the Cwm yr Abbey Member in MPA 20103 and imme- diately above the base of the Pontyfenni Formation in MPA 20099. Comparison with assemblages from other areas Arenig acritarchs are known from other areas of England and Wales, principally through the unpublished work of Booth (1979). Samples collected in north Wales, from the Afon Seiont at Caernarfon, the Menai Straits Inlier and Garth Point at Bangor, are considered to be of Fennian age (Dr R. A. Fortey, personal communication). The acritarch floras contained a number of taxa not recorded from south-west Wales, but also included Coryphidium bohemi- cum, Striatotheca rarirrugulata, Frankea hamata and Uncinisphaera? sp. B, species that are present in Microfossil Assemblage VI. Booth also reported C. bohemicum, S. rarirrugulata, Uncinisphaera? sp., Orthosphaeridium ternatum and Frankea sartbernardensis (Martin) Burmann from Outerside in the Lake District, where Jackson (1978: 92) recorded graptolites of the Didymograptus hirundo Biozone. The Outerside assemblage is again similar to Microfossil Assemblage VI although O. ternatum has only been recorded from the higher part of the Pontyfenni Formation, in Microfossil Assemblage VII. Acritarchs of inferred Arenig age have been described from elsewhere in north-west England. Turner & Wadge (1979) have published an account of ‘mid’ Arenig acritarchs from the south- western Lake District, recording C. bohemicum, S. rarirrugulata and F. hamata. This assemblage is best compared with Microfossil Assemblage VI and is probably of Fennian age. A similar assemblage was recorded by Molyneux (1979) from the Lady Port Banded ‘Group’ on the Isle of Man. Lister (in Arthurton & Wadge 1981: 6-11) has reported early Ordovician acritarchs from the Cross Fell Inlier where he recognized four assemblages, the older two being of probable late Arenig age. There is little similarity between Lister’s assemblages and those from south-west Wales but many key taxa were described subsequent to Lister’s investigations in the late 1960s. His material needs to be re-examined before any useful comparisons can be made. Nothing has been published previously on acritarchs from rocks of known Moridunian or Whitlandian age in Britain, but assemblages from the Glen Dhoo Flags and Lonan Flags on the Isle of Man are of probable latest Tremadoc or earliest Arenig age (Molyneux 1979). Differences between these assemblages and those from the Moridunian of south-west Wales hinder detailed comparison, but the ‘Veryhachium trispinosum’ group has not been recorded 318 S. G. MOLYNEUX al 91 MPA 2010 MPA 20077 MPA 20093 MPA 20092 MPA 200 MPA 20101 MPA 20100 MPA 20099 MPA 20098 MPA 20104 MPA 20103 MPA 20084 MPA 20087 MPA 20074 Belonechitina spp. ; Conochitina cf. chydaea Lagenochitina cylindrica? Lagenochitina sp. A ?Adorfia prolongata Dasydorus cirritus? Orthosphaeridium ternatum Solisphaeridium sp. A : Stellechinatum papulessum 8 2 Stelliferidium aff. fimbrium 1 Uncinisphaera? sp. C Coryphidium bohemicum 1 Stellechinatum uncinatum ‘Veryhachium trispinosum’ group 3 4 Cymatiogalea bellicosa \(r) ?Frankea hamata : : 1 Orthosphaeridium sp. 1 ?Striatotheca mutua 5 1 ?Striatotheca rarirrugulata 1 Timofeevia lancarae 1(r) Uncinisphaera? sp. A 3 Uncinisphaera? sp. B 7 Vulcanisphaera britannica 1(r) Vulcanisphaera turbata? 1(r) 1(r) Nothooidium? spp. 1 1 Micrhystridium sp. A Micrhystridium sp.B Micrhystridium sp. C Micrhystridium sp. D Micrh. aff. henryi 1 4 Micrh. cf. inconspicuum aremoricanum 5). Ie 3 Micrh. aff. nannacanthum 2 11 Polygonium sp. B 6 Solisphaeridium sp. B 4 2 Striatotheca sp. 1 Uncinisphaera? sp. F 4 Acanthodiacrodium sp. A Barakella sp. A Coryphidium? sp. A Uncinisphaera? sp. D Acanthodiacrodium aff. angustum Baltisphaerosum? sp. Cymatiogalea? sp Uncinisphaera? sp. E Acanthodiacrodium aff. spinum ?Coryphidium minutum Micrh. aff. acuminosum Polygonium sp. A ?Vogtlandia flosmaris DH 2 @ Oe Ke Or Wk WON = >) st woo w i) fos _ uw WAH Geiss wore Q en Oe Microfossil Assemblages VII vI Vv IV MOL) Jat I _! <=! Fig. 5 Distribution and abundance of acritarch and chitinozoa species: (r) indicates that the species is probably reworked from the Tremadoc or Cambrian. from either the Manx assemblages or those from the Ogof Hén Formation. Furthermore, specimens of Coryphidium in the Glen Dhoo and Lonan Flags, referred previously to C. bohemicum, are probably not the same as Fennian specimens, having long, slender and flexible rather than short, conical and capitate processes. Acritarchs of reported Arenig age are also known from Europe, north Africa, North America and Australia (Martin 1982). In a number of cases there is no independent evidence for an Arenig age, the age of the assemblages being inferred from their composition and comparisons with existing data. However, Rauscher (1973) has described acritarchs from the undivided Didymograptus extensus Biozone of the Montagne Noire, France, recording Coryphidium bohe- ACRITARCHS AND CHITINOZOA 319 micum, Striatotheca rarirrugulata and specimens of the ‘V. trispinosum’ group. Comparison with south-west Wales suggests that the assemblage is probably of early Fennian age, implying that it is from the upper part of the D. extensus Biozone. Vavrdova (1965, 1966, 1972, 1973, 1976) has described acritarchs from the Klabava Shales in the Rokycany district, ‘U Starého hraduw’ [‘at the old castle’] south-east of Klabava, Bohemia, where they occur in the Tetragraptus cf. pseudobigsbyi Biozone, a biostratigraphical unit that replaced the T. reclinatus abbreviatus Biozone (see Martin 1982: 35). The Tetragraptus reclin- atus abbreviatus Biozone was regarded as being approximately equivalent to the Isograptus gibberulus Biosubzone of the British succession (Cooper & Fortey 1982: fig. 2), implying correl- ation with the Fennian. The acritarch assemblage ‘U Stareho hradw’ includes C. bohemicum and the ‘V. trispinosum’ group as well as a number of taxa not recorded from south-west Wales. Similarity between the two areas is apparently limited. Arenig acritarchs from the upper part of the Bell Island Group and the overlying Wabana Group of Bell Island, eastern Newfoundland, have been recorded by Martin (in Dean & Martin 1978). Graptolites from approximately 20m above the base of the Wabana Group are reported to indicate the upper part of the D. extensus Zone, implying a Fennian age. The acritarch assemblages have little in common with those from south-west Wales, containing thirty species of which six are recorded in this paper. Acritarchs from equivalents of the Arenig Series in the Baltic region of the U.S.S.R. have been described by Timofeev (1959). The Baltic assemblages are unlike those from south-west Wales, where none of Timofeev’s species have been recognized. The differences might arise from separation of the two areas across climatic zones in the Arenig, faunal evidence placing south- west Wales at high latitudes and the Baltic at temperate latitudes (Cocks & Fortey 1982). No Arenig chitinozoa have hitherto been described from the British Isles, although Lister (in Arthurton & Wadge 1981) has reported chitinozoa from rocks of probable Arenig age in the Cross Fell Inlier. Post-Arenig chitinozoa have been described by Jenkins (1967) from the Hope Shales of Lower Llanvirn age in the Shelve Inlier of Shropshire. The specimens of Conochitina ef. chydaea and Lagenochitina cylindrica? that are present in Microfossil Assemblage VII resem- ble species recorded by Jenkins, but there is otherwise little in common with his Lower Llan- virn faunas. Many of the species that are characteristic of Llanvirn assemblages, notably species of Siphonochitina, are absent from the Pontyfenni Formation, as are other taxa such as Cyatho- chitina campanulaeformis and species of Rhabdochitina which range upwards from the Llanvirn. In contrast, the most common form in the Pontyfenni Formation, Lagenochitina sp. A, is apparently absent from the Hope Shales. Arenig chitinozoa have been described from Quebec (Achab 1982), Spitsbergen (Bockelie 1980), Belgium (Martin 1969a), France (Rauscher 1968, 1973), north Africa (Benoit & Tau- gourdeau 1961), south-west Europe (Paris 1981), Australia (Combaz & Peniguel 1972) and Sweden (Grahn 1980). Further work, including scanning electron microscopy, is needed before comparisons can be made between these faunas and the chitinozoa from south-west Wales. Very little comparison can be made with other areas, probably reflecting inadequate sam- pling for Arenig acritarchs and chitinozoa. Even this account is based on so few samples and such incomplete coverage that it should be regarded as preliminary. Martin (1982) notes that information from the Lower Arenig graptolite biozones of Tetragraptus approximatus and Didymograptus defiexus is sparse. It may be significant that many of the previously described assemblages are more similar to Microfossil Assemblage VI than any other, the earlier assemblages from south-west Wales occupying an interval that has not been sampled elsewhere. Systematic descriptions: Acritarchs Figured specimens are deposited in the Palaeontological Collections of the British Geological Survey, Keyworth, and are registered in the series MPK 4870-4978. Figure explanations include the specimen’s register number (e.g. MPK 4971), details of the sample and an England Finder co-ordinate (e.g. K34/0) to locate the specimen on the slide. The co-ordinates were obtained on a Zeiss photomicroscope bearing the number 66303. Acritarch genera and species are described alphabetically. 320 S. G. MOLYNEUX Open nomenclature. The genus name followed by ‘sp. A’, &c., is used when the species seems to be new but cannot be formally described on the available material. I use ‘sp.’ alone to indicate the material cannot be assigned to an existing species owing to poor preservation, or because the potentially diagnostic characters carry low taxonomic weight, requiring population study (for which there is insufficient material) to diagnose a new species, as against a variety of an existing species (as with Striatotheca sp., p. 346). The use of ‘? in different positions follows the convention of the British Geological Survey Notes for Authors (Dhonau 1982: 22). Genus ACANTHODIACRODIUM Timofeev, 1958 TYPE SPECIES. Acanthodiacrodium dentiferum Timofeev 1958. Acanthodiacrodium aff. angustum (Downie 1958) Combaz 1967 Fig. 6A, B aff. 1958 Diornatosphaera angusta Downie: 345-346; pl. 17, figs 7, 8; text-fig. 3e. aff. 1962 Lophodiacrodium angustum (Downie) Deflandre & Deflandre-Rigaud: 194. aff. 1967 Acanthodiacrodium angustum (Downie) Combaz: 15; pl. 3, figs 67—72. MATERIAL. One specimen. OCCURRENCE. Cwmffrwd Member: MPA 20077. DESCRIPTION. The vesicle is ellipsoidal and opaque but more or less intact. About 20 short, rounded, densely crowded cones are present at each pole on the long axis of the vesicle. DIMENSIONS. Vesicle diameter 37 x 30 um; cone length less than 1 wm. REMARKS. This specimen resembles the Tremadoc species Acanthodiacrodium angustum and morphologically similar species such as Lophodiacrodium filiforme (Timofeev) Deflandre & Deflandre-Rigaud. Poor preservation does not allow a positive identification. 6a . 6b Fig. 6A, B_ Acanthodiacrodium aff. angustum (Downie 1958) Combaz 1967, high and low focus; MPK 4870, sample MPA 20077, Cwmffrwd Member; slide 2, J20/1, x 1200. Acanthodiacrodium aff. spinum Rasul 1979 Fig. 7A, B aff. 1979 Acanthodiacrodium spinum Rasul: 66-67; pl. 3, figs 1-7. MATERIAL. One specimen. OCCURRENCE. Allt Cystanog Member: MPA 20074. DESCRIPTION. The specimen is opaque but otherwise has suffered little damage. The vesicle is ellipsoidal with a slight equatorial constriction. The processes are short, flexible and tapering, and have evexate or capitate distal terminations. They may be solid or hollow. Approximately 25-30 processes are present at each pole. ACRITARCHS AND CHITINOZOA 321 7a 7b 10 Fig. 7A, B_- Acanthodiacrodium aff. spinum Rasul 1979, high and low focus; MPK 4871, sample MPA 20074, Allt Cystanog Member; slide 2, N63/1, x 1200. Fig. 8 Acanthodiacrodium sp. A; MPK 4872, sample MPA 20087, Cwmffrwd Member; slide 2, E24/0, x 1200. Fig.9 Baltisphaerosum ? sp.; MPK 4873, sample MPA 20077, Cwmffrwd Member; slide 2, U24/2, x 1200. Fig. 10 ?Adorfia prolongata Burmann 1970; MPK 4874, sample MPA 20093, Pontyfenni Formation; slide 1, P22/0, x 480. See Fig. 12. Fig. 11 Barakella sp. A; MPK 4875, sample MPA 20087, Cwmffrwd Member; slide 1, E30/3, x 1200. See Fig. 13. 322 S. G. MOLYNEUX DIMENSIONS. Vesicle diameter 24 x 19 wm; process length 4 um. REMARKS. The gross morphology and dimensions of this specimen resemble those of Acantho- diacrodium spinum, described by Rasul (1979) from the Clonograptus tenellus Zone and Brachio- pod Beds of the Tremadocian Shineton Shales. A. spinum, however, has hollow, acuminate processes and a finely striate vesicle. Acanthodiacrodium sp. A Fig. 8 MATERIAL. One specimen. OCCURRENCE. Cwmffrwd Member: MPA 20087. DESCRIPTION. The specimen is dark brown to grey and more or less intact. The vesicle is roughly hexagonal, but is drawn out along one axis. The processes are concentrated at the poles of the long axis, seven at one pole and four at the other. The processes are stout, conical, and together with the vesicle are covered by robust, hollow cones or hairs with solid tips. The narrow equatorial zone may be striate, the rather indistinct striae being parallel to the long axis of the vesicle. DIMENSIONS. Vesicle diameter 40 x 33 um; process length 12 um. REMARKS. The robust ornament distinguishes this specimen from most other species of Acanthodiacrodium. A. achrasi Martin, 1972, is similar but is smaller and has a finer ornament. Genus ADORFIA Burmann, 1970 TYPE SPECIES. Adorfia firma Burmann 1970. ? Adorfia prolongata Burmann 1970 Figs 10, 12 21970 Adorfia prolongata Burmann: 295; pl. 5, figs 1, 2, 5. 21978 Adorfia prolongata Burmann; Dean & Martin: 7; pl. 2, figs 6, 9; pl. 3, fig. 27. MATERIAL. Two specimens. OccuRRENCE. Pontyfenni Formation: MPA 20091, MPA 20093. DESCRIPTION (based on one specimen from MPA 20093). The vesicle is subpolygonal to quad- rate in outline. The process bases coalesce in part to mask the vesicle outline. Process stems are stout and cylindrical or slightly tapered. The processes divide distally by dichotomy, with up to five orders of division. The terminal branches on each process are long and recurved, and are apparently capitate. Nine processes are present. DIMENSIONS. Vesicle diameter: 26 x 28 wm and 28 x 344m Process length (overall): c. 20-25 um Process length (stem): c. 13-16 um Process width (base): 4-5 ym REMARKS. The specimen from MPA 20093 apparently has capitate process terminations, a character diagnostic of the genus Adorfia Burmann. The dimensions of the vesicle and processes Fig. 12 ?Adorfia prolongata Burmann_ 1970, detail of processes; MPK 4874. Bar represents ee ee ee oe 30 um. See Fig. 10. ACRITARCHS AND CHITINOZOA B28 are consistent with those of A. prolongata, and the number of processes on the specimen falls within the range of variation recorded by Martin (in Dean & Martin 1978). Poor preservation precludes positive identification of the specimen. Genus BALTISPHAEROSUM Turner, 1984 TYPE SPECIES. Baltisphaerosum christoferi (Kjellstrom 1976) Turner 1984. Baltisphaerosum? sp. Fig. 9 MATERIAL. One specimen. OCCURRENCE. Cwmffrwd Member: MPA 20077. DESCRIPTION. The specimen is hemispherical with a smooth or finely granulate wall. Its shape suggests that it is one half of a spherical vesicle which has split equatorially. About 30 short, slender, hollow and evexate processes, which are plugged at the base and do not communicate with the interior of the vesicle, are present. DIMENSIONS. Vesicle diameter 40 x 20 um; process length 7 um. REMARKS. The shape of the specimen suggests excystment by means of an equatorial split. Simple, hollow, proximally plugged processes and this type of excystment are diagnostic of Baltisphaerosum. Assignment to Baltisphaerosum is tentative, however, because this taxon is poorly recorded in the Arenig of south-west Wales, and also because it is difficult to be certain that the splitting is not the result of accidental damage. The specimen has much shorter processes than other known species of Baltisphaerosum. Genus BARAKELLA Cramer & Diez, 1977 Type species. Barakella fortunata Cramer & Diez 1977. Barakella sp. A Figs 11, 13 MATERIAL. One specimen. OCCURRENCE. Cwmffrwd Member: MPA 20087. DESCRIPTION. The vesicle is rectangular and bears four processes, one at each corner. The processes are short and stout, with rounded distal terminations. The vesicle and processes are covered by short hairs or grana, and one of the two shorter sides of the vesicle has an area of short anastomosing hairs midway along its length. DIMENSIONS. Vesicle diameter 36 x 24 um; process length 8 um. REMARKS. The area of short anastomosing hairs is diagnostic of the genus but the ornament distinguishes this specimen from other species of Barakella. Fig. 13 Barakella sp. A, detail of surface orna- ment and the structure midway along one of the two shorter sides; MPK 4875. Bar rep- resents 30 wm. See Fig. 11. 324 S. G. MOLYNEUX Genus COR YPHIDIUM Vavrdova, 1972 TYPE SPECIES. Coryphidium bohemicum Vavrdova 1972. Coryphidium bohemicum Vavrdova 1972 Figs 14-18, ? 20 1972 Coryphidium bohemicum Vavrdova: 84-85; pl. 1, figs 1, 2; text-fig. 4. MATERIAL. Nine specimens. OccuRRENCE. Pontyfenni Formation: MPA 20092-20100, 20102. DESCRIPTION. The vesicle is quadrate with straight or concave sides and broadly rounded corners. Striations on the vesicle wall are more or less parallel to the sides of the vesicle. There are about 30 processes, concentrated at the corners of the vesicle, which have short, stout, conical stems and apparently hollow capitate distal terminations; the conical stems may be solid or hollow. DIMENSIONS. Vesicle diameter: range 18-26 um; mean 21 ym. Process length: range 1-5—2:5 um; mean 2 pm. Process width: less than 1 um. REMARKS. The vesicles of the specimens from the Pontyfenni Formation conform to the diag- nosis of Coryphidium bohemicum, but the processes are slightly shorter than those originally described by Vavrdova (1972). The distal process terminations of the Pontyfenni specimens are unlike any of the examples illustrated by Vavrdova, but as the processes of C. bohemicum are reported in the diagnosis to be distally heteromorphic, capitate terminations are not necessarily inconsistent with the determination. Specimens with predominantly capitate terminations on short, stout, conical stems appear to be characteristic of the Fennian; Booth (1979) illustrates a number of examples from the Fennian of north Wales, and Turner & Wadge (1979) illustrate three poorly preserved specimens with apparently similar processes from rocks of probable Fennian age in the Lake District. The specimens from the Pontyfenni Formation have fewer processes than the type material. ?Coryphidium minutum Cramer & Diez 1976 Fig. 19 21976 Coryphidium minutum Cramer & Diez: 205; pl. 23, figs 7, 10; text-fig. 2: 7. MATERIAL. One specimen. OCCURRENCE. Allt Cystanog Member: MPA 20074. DESCRIPTION. The specimen is split at one end but is otherwise undamaged. The vesicle is quadrate with more or less straight sides and broadly rounded corners. The vesicle wall is apparently smooth. The processes are more prominent at the corners of the vesicle but are not restricted to that position. They are relatively short, slender and bifid, and may be hollow. It is not clear whether the processes communicate with the interior of the vesicle. About 40 pro- cesses are present. DIMENSIONS. Vesicle diameter 28 x 20 um; process length 2-5 ym. REMARKS. The specimen resembles Coryphidium minutum as illustrated by Cramer & Diez (1976) from rocks of alleged Upper Arenig age in Morocco. According to its diagnosis, C. minutum has slightly shorter processes with simple or capitate distal terminations. Coryphidium? sp. A Figs 21-22 MATERIAL. One specimen. OCCURRENCE. Cwmffrwd Member: MPA 20087. ACRITARCHS AND CHITINOZOA 325 21 Figs 14-18 Coryphidium bohemicum Vavrdova 1972. All Pontyfenni Formation; x 1200. Fig. 14, MPK 4876, sample MPA 20100; slide 1, W64/2. Fig. 15, MPK 4877, sample MPA 20102; slide 1, Q66/0. Fig. 16, MPK 4878, sample MPA 20102; slide 1, X74/4. Fig. 17, MPK 4879, sample MPA 20102; slide 1, MS52/0. Fig. 18, MPK 4880, sample MPA 20092; slide 1, Q62/4. Fig. 19 ?Coryphidium minutum Cramer & Diez 1976; MPK 4881, sample MPA 20074, Allt Cystanog Member; slide 1, J49/3, x 1200. Fig. 20 Coryphidium bohemicum? Vavrdova 1972; MPK 4882, sample MPA 20093, Pontyfenni Forma- tion; slide 1, P30/0, x 1200. Fig. 21 Coryphidium? sp. A; MPK 4883, sample MPA 20087, Cwmffrwd Member; slide 1, H37/0, x 1200. See Fig. 22. DESCRIPTION. The vesicle is quadrate with straight sides and broadly rounded corners. The vesicle wall is apparently smooth. The processes, which are concentrated at the corners of the vesicle, are stout and have elaborate distal terminations that bifurcate to the second order. The terminal branches of the processes may be capitate. Sixteen processes are present. Excystment may be by means of a straight split that occurs along one side of the vesicle. 326 S. G. MOLYNEUX Fig. 22. Coryphidium? sp. A; MPK 4883. Bar represents 30 ym. See Fig. 21. DIMENSIONS. Vesicle diameter: 32 x 25 um. Process length: 7 um. Process width: 1-5 wm. REMARKS. The vesicle has been distorted by crystal growth in the internal cavity but otherwise preservation is fair. The shape of the vesicle and the concentration of processes at the corners are characteristic of Coryphidium, but the specimen has longer and more elaborately branching processes than known species of that genus. The specimen may also differ from Coryphidium in its excystment mechanism. The split along one side is interpreted as a means of excystment whereas in the type species, C. bohemicum, it is reported to be by means of a large opening of irregular shape, usually oval or polygonal (Vavrdova 1972; Martin in Dean & Martin 1978). The genus Tetraniveum Vavrdova, 1976, has a similar vesicle shape and process arrangement, but the processes are simple. Genus CYMATIOGALEA Deunff, 1961 TYPE SPECIES. Cymatiogalea margaritata Deunff 1961. Cymatiogalea bellicosa Deunff 1961 Fig. 25A, B 1961 Cymatiogalea bellicosa Deunff: 42; pl. 1, fig. 13. 1961 Cymatiogalea pudica Deunff: 42; pl. 1, fig. 4. 1964 Cymatiogalea bellicosa Deunff; Deunff: 122; pl. 1, figs 10-12, 16, 19-20. MATERIAL. One specimen. OCCURRENCE. Pontyfenni Formation: MPA 20099. DESCRIPTION. The vesicle is hemispherical, with a large polar opening (macropyle). The pro- cesses are stout and cylindrical, dividing distally into several short branches, all of which are of the first order, arising from a common point on the stem. The processes show some variation in length, the longer ones being situated opposite the macropyle while those nearer the opening are much shorter. The processes support a veil. Figs 23, 24 Cymatiogalea? sp.; sample MPA 20077, Cwmffrwd Member; bar represents 30 pum. Fig. 23, MPK 4885; slide 2, N24/0. See Fig. 26. Fig. 24, MPK 4886; slide 2, V25/4. See Fig. 27. 23 ACRITARCHS AND CHITINOZOA 327 DIMENSIONS. Vesicle diameter: 26 x 36 um; process length 8 um opposite macropyle, decreasing to 3 wm near macropyle. REMARKS. Cymatiogalea bellicosa is widespread in rocks of Tremadoc age. In Britain, it occurs in the lower part of the Shineton Shales, of early Tremadoc age, in Shropshire (Rasul 1974, 1979). Its presence in the Pontyfenni Formation indicates probable reworking. Cymatiogalea? sp. Figs 23-24, 26-27 MATERIAL. Three specimens. OCCURRENCE. Cwmffrwd Member: MPA 20077. DESCRIPTION. The vesicles are subspherical, and that of one specimen (Figs 24, 27) may be divided into polygonal fields; this specimen may also have a macropyle. The processes are short, hollow, and cylindrical, and are plugged at the base so that the process interiors do not communicate with the vesicle cavity. They are usually divided distally into four or five fila- ments, but on one specimen (Figs 23, 26) they have more elaborately branched distal termina- tions which divide to the second order. DIMENSIONS. Vesicle diameter: range 24-36 wm; mean 29 pm. Process length: range 3—5 wm; mean 4 um. 25a 25b 27 Fig. 25A, B_ Cymatiogalea bellicosa Deunff 1961; MPK 4884, sample MPA 20099, Pontyfenni Forma- tion; slide 1, $34/3, x 1200. Figs 26,27 Cymatiogalea? sp.; x 1200. Fig. 26, MPK 4885. See Fig. 23. Fig. 27, MPK 4886. See Fig. 24. 328 S. G. MOLYNEUX REMARKS. Assignment of these three specimens to Cymatiogalea is based on the apparent presence of a macropyle and polygonal fields on one specimen. These characters are diagnostic of the genus according to the emended diagnosis given by Deunff et al. (1974). The determi- nation is tentative because of poor preservation. Genus DASYDORUS Playford & Martin, 1984 TYPE SPECIES. Dasydorus cirritus Playford & Martin 1984. Dasydorus cirritus? Playford & Martin 1984 Figs 28-43 21984 Dasydorus cirritus Playford & Martin: 198, fig. 6A—C. MATERIAL. Eighteen specimens. OccuRRENCE. Pontyfenni Formation: MPA 20091-3. DESCRIPTION. The vesicle is subtriangular or egg-shaped. One end, usually the narrower, is smooth and is acutely rounded or is drawn out into a short apical protuberance. The rest of the vesicle is covered by numerous randomly distributed, short, stiff and evexate or capitate hairs. Excystment may have been by means of a longitudinal split, either alone or in combination with loss of the smooth apical region. DIMENSIONS. Vesicle length: range 38-58 um; mean 49 pm. Vesicle width: range 32—48 um; mean 37 um. Length of hairs: less than 2—2-5 wm. REMARKS. The specimens are very similar to those described and figured by Playford & Martin (1984), but the smooth apex is more acutely rounded and in some cases is developed into a short protuberance. These albeit slight morphological differences, and the difference in preser- vation, disallow a confident identification until more is known about the morphology and occurrence of the species. The excystment mechanism of Dasydorus is unknown but four specimens from the Ponty- fenni Formation (Figs 33-36, 41-43) provide some evidence. Each shows a longitudinal split, accompanied on one specimen by loss of the smooth apical region. Given their state of preser- vation, it is difficult to eliminate incidental damage as its cause, but the consistent appearance of the split suggests that it may be a true excystment opening. Playford & Martin (1984) note that the genus Pirea Vavrdova differs from Dasydorus by possessing a distinct apical process. The short apical protuberance on some of the Welsh specimens resembles this process, suggesting a possible relationship between the two genera. Genus FRANKEA Burmann, 1970 TYPE SPECIES. Frankea hamata Burmann 1970. ?Frankea hamata Burmann 1970 Fig. 44 21970 Frankea hamata Burmann: 290-291; pl. 2, figs 7, 9, 10. MATERIAL. One damaged specimen. OccuRRENCE. Pontyfenni Formation: MPA 20102. DESCRIPTION. The vesicle is broken but was probably triangular. Two processes are present, situated at two corners of the triangle; the third corner is broken. The processes are short and divide distally. One process divides into two long, recurved filaments. DIMENSIONS. Vesicle diameter 24 x 20 um; process length 4 um. 31a 32 310" Figs 28-32 Dasydorus cirritus? Playford & Martin 1984. All Pontyfenni Formation; x 1200. Fig. 28A, B, high and low focus; MPK 4915, sample MPA 20091; slide 1, J28/1. See Fig. 39. Fig. 29A, B, high and low focus; MPK 4916, sample MPA 20092; slide 2, Q34/0. See Fig. 37. Fig. 30A, B, specimen with protuberance resembling apical horn, high and low focus; MPK 4917, sample MPA 20093; slide 1, J31/4. See Fig. 40. Fig. 31A, B, high and low focus; MPK 4918, sample MPA 20092; slide 1, W36/0. Fig. 32, MPK 4919, sample MPA 20091; slide 1, D35/1. See Fig. 38. 330 S. G. MOLYNEUX Figs 33-36 Excystment mechanism of Dasydorus cirritus? Playford & Martin 1984. All sample MPA 20092, Pontyfenni Formation; x 1200. Fig. 33, specimen with longitudinal split; MPK 4920; slide 1, E24/0. Fig. 34, specimen with partial longitudinal split developing at the antapex; MPK 4921; slide 1, F56/2. See Fig. 41. Fig. 35, specimen with partial longitudinal split restricted to one side of the vesicle and also exhibiting loss of the apical region; MPK 4922; slide 1, K24/1. See Fig. 43. Fig. 36, specimen with partial longitudinal split developing at the antapex; MPK 4923; slide 2, K34/2. See Fig. 42. REMARKS. Identification of the specimen is tentative because of damage to the vesicle. Even so, the probable shape of the vesicle and the distal terminations of the processes are characteristic of Frankea hamata, and the determination is probably correct. The processes are shorter than those of the type specimen (Burmann 1970). Genus MICRHYSTRIDIUM Deflandre, 1937 TYPE SPECIES. Micrhystridium inconspicuum Deflandre 1937. Micrhystridium aff. acuminosum Cramer & Diez 1977 Figs 45—47, 70 aff. 1977 Micrhystridium acuminosum Cramer & Diez: 347; pl. 1, figs 3, 4, 10; text-fig. 3: 3. ACRITARCHS AND CHITINOZOA 331 Hayy © nz Aye Sv wl TD, Seles Figs 37-40 Dasydorus cirritus? Playford & Martin 1984. Bar represents 30 um. Fig. 37, MPK 4916. See Fig. 29. Fig. 38, MPK 4919. See Fig. 32. Fig. 39, MPK 4915. See Fig. 28. Fig. 40, MPK 4917. See Fig. 30. MATERIAL. Three specimens. OCCURRENCE. Allt Cystanog Member: MPA 20074. DESCRIPTION. The vesicle is small and subspherical. The outline of the vesicle is partially masked by the process bases. The processes are numerous, relatively short and conical, extend- ing distally into acuminate, needle-like tips. Figs 41-43 Excystment mechanism of Dasydorus cirritus? Playford & Martin 1984. Bar represents 30 um. Fig. 41, MPK 4921. See Fig. 34. Fig. 42, MPK 4923. See Fig. 36. Fig. 43, MPK 4922. See Fig. 35. 332 S. G. MOLYNEUX 51 Fig. 44. ?Frankea hamata Burmann 1970; MPK 4887, sample MPA 20102, Pontyfenni Formation; slide 1, R54/3, x 1200. Figs 45-47 Micrhystridium aff. acuminosum Cramer & Diez 1977; sample MPA 20074, Alit Cystanog Member; x 1200. Fig. 45, MPK 4888; slide 2, H70/2. See Fig. 70. Fig. 46, MPK 4889; slide 1, S53/3. Fig. 47, MPK 4890; slide 1, H71/0. Figs 48-50 Micrhystridium cf. inconspicuum aremoricanum Paris & Deunff 1970; x 1200. Fig. 48, MPK 4891, sample MPA 20103, Cwm yr Abbey Member; slide 2, L57/3. Fig. 49, MPK 4892, sample MPA 20103, Cwm yr Abbey Member; slide 2, D74/2. Fig. 50, MPK 4893, sample MPA 20104, Afon Ffinnant Formation; slide 2, L51/2. See Fig. 74. Figs 51-53 Micrhystridium aff. henryi Paris & Deunff 1970; sample MPA 20103, Cwm yr Abbey Member; x 1200. Fig. 51, MPK 4894; slide 2, H74/2. Fig. 52, MPK 4895; slide 2, V70/1. See Fig. 73. Fig. 53, MPK 4896; slide 2, R53/1. DIMENSIONS. Vesicle diameter range 10—22 wm, mean 18 wm; process length 3 um. REMARKS. The processes on each of the three specimens resemble those of Micrhystridium acuminosum and suggest an affinity with that species, although M. acuminosum has a larger vesicle and longer, stouter processes. ACRITARCHS AND CHITINOZOA 333 Micrhystridium aff. henryi Paris & Deunff 1970 Figs 51—53, 73 aff. 1970 Micrhystridium henryi Paris & Deunff: 31—32; pl. 2, figs 2, 10, 14, 15, 18; pl. 3, fig. 7. MATERIAL. Five specimens. OCCURRENCE. Cwm yr Abbey Member: MPA 20103. Afon Ffinnant Formation: MPA 20104. DESCRIPTION. The vesicle is subspherical with about 30-40 processes. The sides of the vesicle between the process bases are straight or slightly curved; if curved they may be either concave or convex. The processes taper from narrow bases to acuminate distal terminations. DIMENSIONS. Vesicle diameter range 20—26 wm, mean 23 um; process length 4-5 ym. REMARKS. The small, subspherical vesicles and numerous, relatively short processes suggest a relationship between these specimens and Micrhystridium henryi. M. henryi has shorter and more numerous processes, their bases coalescing to mask the outline of the vesicle. Specimens from the Fennian of north Wales, referred by Booth (1979) to M. henryi, are closer to this material from south-west Wales than to the type material, but have slightly shorter processes. Micrhystridium cf. inconspicuum aremoricanum Paris & Deunff 1970 Figs 48-50, 74 cf. 1970 Micrhystridium inconspicuum aremoricanum Paris & Deunff: 32; pl. 2, fig. 20. MATERIAL. Twenty-one specimens. OCCURRENCE. Cwm yr Abbey Member: MPA 20084, MPA 20103. Afon Ffinnant Formation: MPA 20104. DESCRIPTION. The vesicle is small and subspherical, bearing about 30 processes. The processes are simple, relatively short, acuminate and narrowly conical. Their bases tend to coalesce, masking the outline of the vesicle. DIMENSIONS. Vesicle diameter range 14-24 um, mean 19m; process length range 3-5—8 ym, mean 5 ym. REMARKS. The material from south-west Wales is very similar to the type material of Micrhys- tridium inconspicuum aremoricanum and to specimens recorded by Booth (1979), but the pro- cesses of the specimens from south-west Wales are commonly about a quarter of the vesicle diameter in length, rarely a third or more. The original diagnosis states that the processes are about a third of the vesicle diameter, and Booth also records the length as being approximately a third. The difference is slight but may distinguish the present specimens from the type and Booth’s material. The type material of M. inconspicuum aremoricanum was recorded from the base of the Andouillé Formation north of Rennes (Paris 1981: 19), considered to be of Llanvirn age (Babin et al. 1974: 365). Booth’s material came from the Fennian of north Wales and the Llanvirn of the Welsh Borderland and Lake District. In south-west Wales, M. cf. inconspicuum aremorica- num appears in the upper Moridunian, but has not been recorded from the Whitlandian or Fennian. Micrhystridium aff. nannacanthum Deflandre 1945 Figs 54-57, 71 aff. 1942 Micrhystridium nannacanthum Deflandre: 476; fig. 13 (nomen nudum). aff. 1945 Micrhystridium nannacanthum Deflandre: 66; pl. 3, figs 5—7. MATERIAL. Thirteen specimens. OCCURRENCE. Cwm yr Abbey Member: MPA 20084, MPA 20103. DESCRIPTION. The vesicle is small and subspherical, bearing about 30 short, hair-like processes. The processes are slender, parallel-sided, possibly solid and are evexate or capitate. 334 S. G. MOLYNEUX DIMENSIONS. Vesicle diameter range 12—22 ~m, mean 16 um; process length less than 2 wm. REMARKS. According to the diagnosis (Deflandre 1945), Micrhystridium nannacanthum has short spines that do not exceed 1 wm in length. The specimens from south-west Wales are distinct, in that they have slightly longer processes, some having distinctive rounded or capitate distal terminations. They also have larger vesicles than the type material. A specimen illustrated by Lister (1970: pl. 10, fig. 11) from the late Silurian of Shropshire also differs from the present material, having a smaller vesicle with shorter, stouter and fewer processes. Booth (1979) has recorded several specimens of ?M. nannacanthum from the Fennian of north Wales and the Llanvirn of the Lake District. His specimens resemble these from the Cwm yr Abbey Member but he notes that his have numerous, short, evenly distributed, conical spines which have either blunt or acuminate distal terminations. In south-west Wales, M. aff. nannacanthum has only been recorded from the upper Moridunian. Micrhystridium sp. A Figs 58-59, 66 MATERIAL. Six specimens. OcCURRENCE. Whitland Abbey Member: MPA 20098. DESCRIPTION. The vesicle is small and subspherical. Its outline is largely masked by the pro- cesses, the bases of which tend to coalesce. The processes are short, stout, cylindrical or slightly tapered, with either evexate or acuminate distal terminations. The sides of the vesicle, where visible between the process bases, are straight or concave. About 20 processes are present. DIMENSIONS. Vesicle diameter range 10-16 wm, mean 12 um; process length less than 2-2-5 ym. REMARKS. Micrhystridium sp. A may be distinguished from M. aff. nannacanthum by its smaller size and longer, relatively stout processes and from Micrhystridium spp. B—D by its cylindrical rather than conical processes. Micrhystridium sp. B Figs 60-61, 67 MATERIAL. Six specimens. OcCuURRENCE. Whitland Abbey Member: MPA 20098. DESCRIPTION. The vesicle is small and subspherical. Its outline is masked by the process bases, which tend to coalesce. The processes are numerous, short, conical and acuminate. DIMENSIONS. Vesicle diameter range 9-15 ym, mean 12 ym; process length less than 2 wm. REMARKS. It is difficult to determine the number of processes present on each specimen because of poor preservation, but there are probably more than thirty. Micrhystridium sp. C Figs 62-63, 68 MATERIAL. Seven specimens. OccuRRENCE. Whitland Abbey Member: MPA 20098. DESCRIPTION. The vesicle is small and subspherical. The processes are numerous, slender, tapering and acuminate. Their bases tend to mask the outline of the vesicle, but where visible the sides are convex. DIMENSIONS. Vesicle diameter range 8-16 um, mean 13 wm; process length range 3—5 um, mean 3-5 um. REMARKS. The exact number of processes is uncertain because of poor preservation, but is probably 25 or more. Micrhystridium sp. C may be distinguished from Micrhystridium sp. B by its longer processes. ACRITARCHS AND CHITINOZOA 335 54 € | 4 ed 58 59 60 61 62 63 64 : 65 Figs 54-57 Micrhystridium aff. nannacanthum Deflandre 1945; sample MPA 20084, Cwm yr Abbey Member; x 1200. Fig. 54, MPK 4897; slide 2, T55/3. See Fig. 71. Fig. 55, MPK 4898; slide 2, T64/2. Fig. 56, MPK 4899; slide 2, W54/0. Fig. 57, MPK 4900; slide 2, X66/4. Figs 58,59 Micrhystridium sp. A; sample MPA 20098, Whitland Abbey Member; x 1200. Fig. 58, MPK 4901; slide 1, W62/2. See Fig. 66. Fig. 59, MPK 4902; slide 1, B57/2. Figs 60,61 Micrhystridium sp. B; sample MPA 20098, Whitland Abbey Member; x 1200. Fig. 60, MPK 4903; slide 1, N60/4. See Fig. 67. Fig. 61, MPK 4904; slide 1, 52/2. Figs 62,63 Micrhystridium sp. C; sample MPA 20098, Whitland Abbey Member; x 1200. Fig. 62, MPK 4905; slide 1, S55/1. See Fig. 68. Fig. 63, MPK 4906; slide 1, R60/1. Fig. 64 Micrhystridium sp. D; MPK 4907, sample MPA 20098, Whitland Abbey Member; slide 1, R53/2, x 1200. See Fig. 69. Fig. 65 Micrhystridium sp.; MPK 4908, sample MPA 20098, Whitland Abbey Member; slide 1, P64/3, x 1200. See Fig. 72. Micrhystridium sp. D Figs 64, 69 MATERIAL. Two specimens. OccuRRENCE. Whitland Abbey Member: MPA 20098. DESCRIPTION. The vesicle is small and subspherical, its sides masked by the process bases which tend to coalesce. The processes are numerous, short, conical and acuminate. DIMENSIONS. Vesicle diameter range 13—16 wm, mean 15 um; process length range 2—3 ym, mean 2:5 um. 336 S. G. MOLYNEUX REMARKS. The exact number of processes is uncertain owing to poor preservation, but one specimen has at least 40. Micrhystridium sp. D may be distinguished from Micrhystridium sp. B by its longer, more numerous processes, and from Micrhystridium sp. C by its shorter, more numerous processes. Micrhystridium sp. Figs 65, 72 MATERIAL. One specimen. OCCURRENCE. Whitland Abbey Member: MPA 20098. DESCRIPTION. The specimen has a small, subspherical vesicle with convex sides. Eight processes are present. They are long, slender, possibly solid and evexate, tapering slightly from narrow bases that have an angular contact with the vesicle wall. The vesicle diameter is 14 x 12 wm and process length is approximately 7 ym. Genus NOTHOOIDIUM Loeblich & Tappan, 1976 TYPE SPECIES. Nothooidium mordidum (Cramer, Allam, Kanes & Diez 1974) Loeblich & Tappan 1976. Nothooidium? spp. Figs 75-76 MATERIAL. Two specimens. OccurRRENCE. Afon Ffinnant Formation: MPA 20104. Pontyfenni Formation: MPA 20101. 67 68 ee 69 70 73 74 Figs 66-74 Micrhystridium spp., comparative illustrations of specimens recorded from the Arenig Series in south-west Wales. Bar represents 30 um. Fig. 66, Micrhystridium sp. A; MPK 4901. See Fig. 58. Fig. 67, Micrhystridium sp. B; MPK 4903. See Fig. 60. Fig. 68, Micrhystridium sp. C; MPK 4905. See Fig. 62. Fig. 69, Micrhystridium sp. D; MPK 4907. See Fig. 64. Fig. 70, Micrhystridium aff. acuminosum Cramer & Diez 1977; MPK 4888. See Fig. 45. Fig. 71, Micrhystridium aff. nannacanthum Deflandre 1945; MPK 4897. See Fig. 54. Fig. 72, Micrhystridium sp.; MPK 4908. See Fig. 65. Fig. 73, Micrhystri- dium aff. henryi Paris & Deunff 1970; MPK 4895S. See Fig. 52. Fig. 74, Micrhystridium cf. inconspicuum aremoricanum Paris & Deunff 1970; MPK 4893. See Fig. SO. ACRITARCHS AND CHITINOZOA 337 DESCRIPTIONS. The specimen from the Afon Ffinnant Formation (Fig. 75) has an elongate, pear-shaped vesicle, one end of which is concave suggesting the presence of an opening. The ornament consists of grana and short, conical, evexate and possibly solid processes. The vesicle of this specimen is 48 ym long and 36 um wide. The specimen from the Pontyfenni Formation (Fig. 76) has a similar elongate, pear-shaped vesicle, one end of which is concave, and an ornament of cones and rods with evexate distal terminations. The vesicle of this specimen is 45 um long and 40 wm wide. REMARKS. The two specimens are very similar and may represent the same species, although the processes on the specimen from the Pontyfenni Formation are longer and more slender than those on the Afon Ffinnant specimen. Determination of both specimens as Nothooidium is tentative because the poor preservation makes it difficult to demonstrate that the truncated ends of the vesicles are cyclopylomes. N. mordidum is smaller and has an ornament that consists of flat-crested verrucae. Ooidium sp. 2 of Cramer & Diez (1977: pl. 6, fig. 20) is very similar but is only illustrated and not described. Genus ORTHOSPHAERIDIUM Eisenack, 1968 TYPE SPECIES. Orthosphaeridium rectangulare (Eisenack 1963) Eisenack 1968. Orthosphaeridium ternatum (Burmann 1970) Eisenack, Cramer & Diez 1976 Fig. 77A, B 1970 Baltisphaera ternata Burmann: 306; pl. 7, fig. 1; pl. 9, fig. 4. 1976 Orthosphaeridium ternatum (Burmann) Eisenack, Cramer & Diez: 529. MATERIAL. Four specimens. OCCURRENCE. Pontyfenni Formation: MPA 20091, MPA 20093. DESCRIPTION. The vesicle is subspherical with three long, slender, acuminate processes which are arranged at c. 120° intervals around the circumference. They are constricted slightly towards the base. Both vesicle and processes bear an ornament of short hairs or cones. DIMENSIONS. Vesicle diameter range 48-60 wm, mean 52 um; process length up to 76 wm. REMARKS. The specimens are poorly preserved but are readily determined as Orthosphaeridium ternatum. O. procerum (Burmann) Eisenack et al. 1976 has a more asymmetrical arrangement of processes. Orthosphaeridium sp. Fig. 79 MATERIAL. One specimen. OccuRRENCE. Pontyfenni Formation: MPA 20101. DESCRIPTION. The vesicle is subspherical with four processes, all broken. The process stems are stout and are constricted towards the base. Both vesicle and processes have an ornament of short, robust hairs. DIMENSIONS. Vesicle diameter 50 x 43 um. REMARKS. This specimen resembles Orthosphaeridium quadrinatum (Burmann) Eisenack, Cramer & Diez 1976, but its poor preservation precludes a positive determination. Genus POLYGONIUM Vavrdova, 1966 TYPE SPECIES. Polygonium gracile Vavrdova 1966. MOLYNEUX S. G. 338 ACRITARCHS AND CHITINOZOA 339 Polygonium sp. A Figs 80-81 MATERIAL. Two specimens, plus several fragments. OccurRRENCE. Allt Cystanog Member: MPA 20074. DESCRIPTION. The outline of the vesicle is subpolygonal. The sides of the vesicle are straight or concave; occasionally they may be masked by the wide process bases. The processes are long, tapering and evexate. The proximal half of each process is relatively stout and thick-walled, while the distal half is thinner-walled and flexible. The difference between the two halves is quite distinct, the change taking place abruptly. The distal half of the process may break away leaving a straight or V-shaped distal edge. About 25 processes are present on one specimen. DIMENSIONS. Vesicle diameter range 26-30 um, mean 28 um; process length c. 18 um. REMARKS. The description is based on one specimen from the Allt Cystanog Member (MPA 20074); a number of fragments in the same sample have similar features. One specimen may also be present in the Cwmffrwd Member (MPA 20077) but the preservation is too poor to be certain. Polygonium sp. B Fig. 78 MATERIAL. Six specimens. OCCURRENCE. Cwm yr Abbey Member: MPA 20084. DESCRIPTION. The vesicle is large and subspherical with a polygonal outline. The sides of the vesicle are straight or concave; occasionally they may be masked by the broad process bases. The processes are long, stout, flexible, tapering and acuminate; about 20 are present. DIMENSIONS. Vesicle diameter range 24-36 um, mean 31 wm; process length range 14-16 um, mean 15 ym. REMARKS. Most specimens are broken and their preservation is poor. The size and polygonal outline of the vesicle and the relatively long, stout processes are distinctive. Genus SOLISPHAERIDIUM Staplin, Jansonius & Pocock, 1965 TYPE SPECIES. Solisphaeridium stimuliferum (Deflandre 1938) Staplin, Jansonius & Pocock 1965S. Solisphaeridium sp. A Figs 82-83, 86 MATERIAL. Four specimens. OCCURRENCE. Pontyfenni Formation: MPA 20091. DESCRIPTION. The vesicle is subspherical with convex sides. The processes are long, slender, smooth, evexate, moderately flexible, solid and slightly tapering. They number more than 40 on each specimen. Figs 75, 76 Nothooidium sp.; x 1200, Fig. 75, MPK 4909, sample MPA 20104, Afon Ffinnant Forma- tion; slide 2, G54/0. Fig. 76, MPK 4910, sample MPA 20101, Pontyfenni Formation; slide 1, G28/0. Fig. 77A, B_ Orthosphaeridium ternatum (Burmann 1970) Eisenack, Cramer & Diez 1976; A x 1200, B x 480; MPK 4911, sample MPA 20091, Pontyfenni Formation; slide 1, X41/0. Fig. 78 Polygonium sp. B; MPK 4912, sample MPA 20084, Cwm yr Abbey Member; slide 2, Y30/4, x 1200. Fig. 79 Orthosphaeridium sp.; MPK 4913, sample MPA 20101, Pontyfenni Formation; slide 1, S26/2, x 1200. Fig. 80 Polygonium sp. A; MPK 4914, sample MPA 20074, Allt Cystanog Member; slide 2, U42/4, x 1200. See Fig. 81. Fig. 81 Polygonium sp. A; MPK 4914. See Fig. 80. Bar represents 30 wm. Figs 82, 83, 86 Solisphaeridium sp. A; sample MPA 20091, Pontyfenni Formation; x 1200. Fig. 82A, B, high and low focus; MPK 4924; slide 1, U27/1. Fig. 83, MPK 4925; slide 1, P43/4. Fig. 86, MPK 4928; slide 1, Q28/3. Figs 84, 85, 87-89 Solisphaeridium sp. B; Cwm yr Abbey Member; x 1200. Figs 84, 85, 87, sample MPA 20103. Fig. 84, MPK 4926; slide 2, N58/2. Fig. 85, MPK 4927; slide 2, Q53/3. Fig. 87, MPK 4929; slide 2, Q75/3. Figs 88, 89, sample MPA 20084. Fig. 88, MPK 4930; slide 2, X59/4. Fig. 89, MPK 4931, slide 2, X51/4. ACRITARCHS AND CHITINOZOA 341 DIMENSIONS. Vesicle diameter range 34-43 wm, mean 37 um; process length range 11—20 um, mean 14 um. REMARKS. Superficially, the specimens resemble two species of Solisphaeridium described by Cramer & Diez (1977) from the Upper Arenig of Morocco. S. solare has a smaller vesicle and relatively long, stout processes, while S. solidispinosum is distinguished by its numerous, acu- minate, conical processes. Solisphaeridium sp. B Figs 84-85, 87-89 MATERIAL. Six specimens. OCCURRENCE. Cwm yr Abbey Member: MPA 20084, MPA 20103. DESCRIPTION. The vesicle is subspherical and has a granulate wall. The sides of the vesicle between the process bases are convex, rarely concave. Numerous short, slender, tapering, acuminate and possibly solid processes are present. DIMENSIONS. Vesicle diameter range 17—24 um, mean 21 «um; process length range 2-8 wm, mean 5 um. REMARKS. The short processes and the nature of the ornament on the vesicle distinguish these specimens from the type and other species of Solisphaeridium. The specimens from MPA 20084 have about 15-20 processes, their lengths being more or less equal to a third of the vesicle diameter. Those from MPA 20103 differ slightly, in that they have more than 30 processes with an average length of less than a fifth of the vesicle diameter. Since relatively few specimens have been recorded and their preservation is poor, these differences are not used to separate the specimens from the two samples. The specimens bear a strong resemblance to Baltisphaeridium cinctum (Timofeev) Rauscher (in Rauscher 1973: 71; pl. 2, figs 7a—b) but have smaller vesicles. Genus STELLECHINA TUM Turner, 1984 TYPE SPECIES. Stellechinatum celestum (Martin 19695) Turner 1984. Stellechinatum papulessum sp. nov. Figs 90-92 DiAGnosis. A species of Stellechinatum with eight to ten processes, which are wide and conical at the base but cylindrical distally. The processes, and possibly the vesicle, bear an ornament of very small grana. HovotyrPe. MPK 4932 (Figs 90-1): MPA 20091, Pontyfenni Formation. OTHER MATERIAL. Twelve specimens, paratypes. Fig. 90 Stellechinatum papulessum sp. nov., holo- type; MPK 4932, sample MPA 20091, Ponty- fenni Formation; slide 1, O039/3. Bar represents 30 um. See Fig. 91. 342 S. G. MOLYNEUX OCCURRENCE. Pontyfenni Formation: MPA 20091-3. NAME. ‘Small pimple or pustule’, referring to the fine ornament. DESCRIPTION. The outline of the vesicle is variable. On some specimens, the sides of the vesicle are masked by the wide, coalescing process bases; on others, the sides of the vesicle are clearly visible, resulting in a polygonal outline. The processes are wide and conical at the base, becoming more cylindrical distally. They taper to an evexate distal termination. A slight con- striction is rarely present at the base of the distal, cylindrical part of the process. The average process length is just over half of the vesicle diameter. DIMENSIONS. Vesicle diameter: range 24-36 wm; mean 29 um. Process length: range 14-21 wm; mean 17 pm. Process width (at base): range 6-11 wm; mean 9 um. REMARKS. Several specimens are distorted by the growth of crystals within the vesicle cavity. The type species, Stellechinatum celestum (Martin) Turner, has a similar outline, similar vesicle dimensions and approximately the same number of processes with wide bases. It is distinguished from S. papulessum sp. nov. by its relatively long, acuminate processes and its ornament of slender spines on the processes and vesicle. S. helosum Turner 1984 has compara- ble vesicle dimensions and a similar ornament but has longer and more numerous acuminate processes with narrower bases. S. brachyscolum Turner 1984 has more numerous processes with an ornament of long spines. Goniosphaeridium splendens (Paris & Deunff) Turner 1984 has a similar gross morphology, but has a smooth wall and more numerous processes. Stellechinatum uncinatum (Downie 1958) comb. nov. Figs 93-94 1958 Hystrichosphaeridium longispinosum var. uncinatum Downie: 337; text-fig. 2a. 1965 Baltisphaeridium longispinosum vat. uncinatum (Downie) Downie & Sarjeant: 92. 1965 Baltisphaeridium uncinatum (Downie) Martin: 425-426; text-fig. 1. 1970 Micrhystridium uncinatum (Downie) Cramer: 107-108; pl. 6, figs 97-98, 101; text-fig. 29d (pars). non 1971 Goniosphaeridium uncinatum (Martin) Kjellstrom: 27—28; fig. 18 [err. cit. pro Goniosphaeridium uncinatum (Downie) Kjellstrom]. MATERIAL. Seven specimens. OCCURRENCE. Pontyfenni Formation: MPA 20099, MPA 20101, MPA 20093. DESCRIPTION. The vesicle is polygonal or subpolygonal in outline, with more or less straight sides. There are about 15 processes, which are slender, tapering, acuminate and may be solid. They have relatively narrow bases, bear an ornament of short spines, and are just over half the vesicle diameter in average length. DIMENSIONS. Vesicle diameter: range 22-32 wm; mean 29 um. Process length: range 15—18 wm; mean 17 ym. Process width (at base): 2:5—3 um. REMARKS. The holotype of Hystrichosphaeridium longispinosum var. uncinatum Downie (1958: fig..2a) has a polygonal vesicle with several tapering, acuminate processes, the latter bearing short lateral hairs. The species is morphologically comparable with the genus Stellechinatum Turner and is here recombined with it. Turner (1984) describes Stellechinatum as having wide process bases. In contrast, on the holotype of S. uncinatum (Downie) comb. nov. they are relatively narrow. Other species of Stellechinatum, notably S. brachyscolum Turner 1984 and S. helosum Turner 1984, have process bases of variable width, the narrower bases being comparable with those of S. uncinatum (Downie) comb. nov. S. helosum is distinguished from S. uncinatum (Downie) comb. nov. by its ornament of grana rather than of spines on the processes. S. brachyscolum has more processes than S. uncinatum, their conical bases coalescing to form the vesicle outline. ACRITARCHS AND CHITINOZOA 343 Figs 91-92 Stellechinatum papulessum sp. nov.; sample MPA 20091, Pontyfenni Formation; x 1200. Fig. 91A, B, holotype, high and low focus; MPK 4932; slide 1, 039/3. See Fig. 90. Fig. 92A, B, high and low focus; MPK 4933; slide 1, N41/1. Figs 93-94 Stellechinatum uncinatum (Downie 1958) comb. nov.; sample MPA 20099, Pontyfenni For- mation; x 1200. Fig. 93, MPK 4934; slide 1, R30/2. Fig. 94, MPK 4977; slide 1, M35/3. 344 S. G. MOLYNEUX Three of the specimens illustrated by Cramer (1970: pl. 6, figs 97-98, 101) are unlike S. uncinatum (Downie) comb. nov., two having subspherical vesicles and relatively long processes, the third fewer processes that, together with the vesicle, are covered by an ornament of rela- tively coarse spines. Only the text-figure (Cramer 1970: text-fig. 29d) bears some resemblance to the holotype and to the material illustrated herein. The specimen illustrated by Kjellstr6m (1971: fig. 18), from the Middle Ordovician of Gotland, has more numerous processes with broad, conical bases and is comparable with Stellechinatum brachyscolum from the Caradoc of England. Genus STELLIFERIDIUM Deunff, Gorka & Rauscher, 1974 TYPE SPECIES. Stelliferidium striatulum (Vavrdova 1966) Deunff, Gorka & Rauscher 1974. Stelliferidium aff. fimbrium (Rasul 1974) Rasul 1979 Figs 95, 97 aff. 1974 Priscogalea fimbria Rasul: 47; pl. 3, figs 1-2. aff. 1979 Stelliferidium fimbrium (Rasul) Rasul: 69. MATERIAL. Two specimens. OccCURRENCE. Pontyfenni Formation: MPA 20091-2. DESCRIPTION. The vesicle is subspherical and granulate, and has a large polar opening (macropyle). About 20—25 processes are present; they are hollow and cylindrical but it is not clear if they communicate with the vesicle cavity. Distally, they divide into three or four long, recurved branches that lie at a tangent to the surface of the vesicle. Striations radiate from the base of each process across the surface of the vesicle. DIMENSIONS. Vesicle diameter range 26-28 wm, mean 27 wm; process length range 5—8 ym. REMARKS. The specimens resemble the Tremadoc species Stelliferidium fimbrium (Rasul) Rasul but are smaller and may have more robust distal terminations. Genus STRIATOTHECA Burmann, 1970 TYPE SPECIES. Striatotheca principalis Burmann 1970. ? Striatotheca mutua Burmann 1970 Fig. 96A, B 21970 Striatotheca mutua Burmann: 301; pl. 11, fig. 2. 21978 Striatotheca mutua Burmann: Dean & Martin; 9; pl. 2, figs 12, 13; pl. 3, fig. 18. MATERIAL. One specimen. OccuURRENCE. Pontyfenni Formation: MPA 20099. DESCRIPTION. The vesicle is elongate and pentagonal in outline. Two processes are present but the total number may have been five or six; they are relatively short, conical and evexate. Striations extend from the base of each process and are more or less parallel to the five sides of the vesicle. DIMENSIONS. Vesicle diameter 38 x 26 um; process length 18 um. REMARKS. The holotype (Burmann 1970: pl. 11, fig. 2) has a more or less rectangular outline with three processes at each pole on the long axis. The diagnosis also refers to the vesicle as being four-sided. The processes on the holotype are longer than those on the present specimen from the Pontyfenni Formation. Martin (in Dean & Martin 1978) describes the vesicle of S. mutua as being polyhedral, and although she remarks that the outline is quadrangular, one specimen (Dean & Martin 1978: pl. 3, fig. 18) may be interpreted as having five or six sides. Martin’s specimens have dimensions similar to that from the Pontyfenni Formation, although the latter has relatively shorter processes. ACRITARCHS AND CHITINOZOA 345 Figs 95, 97 Stelliferidium aff. fimbrium (Rasul 1974) Rasul 1979; Pontyfenni Formation; x 1200. Fig. 95A, B, high and low focus; MPK 4978, sample MPA 20092; slide 1, L34/3. Fig. 97, MPK 4936, sample MPA 20091; slide 1, V62/0. Fig. 96A, B_ ?Striatotheca mutua Burmann 1970, high and low focus; MPK 4935, sample MPA 20099, Pontyfenni Formation; slide 1, M31/4, x 1200. Fig. 98 ?Striatotheca rarirrugulata (Cramer, Kanes, Diez & Christopher 1974) Eisenack, Cramer & Diez 1976; MPK 4937, sample MPA 20102, Pontyfenni Formation; slide 1, Q70/4, x 1200. Fig. 99 Striatotheca sp.; MPK 4938, sample MPA 20103, Cwm yr Abbey Member; slide 2, H67/0, x 1200. 346 S. G. MOLYNEUX ? Striatotheca rarirrugulata (Cramer, Kanes, Diez & Christopher 1974) Eisenack, Cramer & Diez 1976 Fig. 98 21974 Rugulidium rarirrugulatum Cramer, Kanes, Diez & Christopher: 61; pl. 25, figs 19, 21, 23; pl. 26, fig. 24. 21976 Striatotheca rarirrugulata (Cramer, Kanes, Diez & Christopher) Eisenack, Cramer & Diez: 775— 776. MATERIAL. One damaged specimen. OccuRRENCE. Pontyfenni Formation: MPA 20102. DESCRIPTION. The vesicle is four-sided, broken and striate; there are 4-6 widely spaced stri- ations across its width. Two processes, both broken, are present at two of the corners of the vesicle. DIMENSIONS. Vesicle diameter 21 x 20 um; process length more than 11 um. REMARKS. Although the specimen is too badly damaged for a positive identification, its size, shape and the nature of its striate ornament suggest it is most probably a specimen of S. rarirrugulata. Striatotheca sp. Fig. 99 MATERIAL. One specimen. OCCURRENCE. Cwm yr Abbey Member: MPA 20103. DESCRIPTION. The vesicle is four sided, folded and closely and relatively coarsely striate; there are about two striations per wm across the width of the vesicle. Four processes, at the corners of the vesicle and lying in the same plane, are long and conical, tapering to an acuminate tip from a broad base. The striations extend a little way onto the base of each process but the process stems are smooth. DIMENSIONS. Vesicle diameter: 22 x 14 um. Process length: 25 um. Process width (at base): 4 um. REMARKS. The specimen is distinguished from other species of Striatotheca by the combination of small size and relatively long, broad-based processes. Striatotheca frequens Burmann, 1970, and S. principalis Burmann, 1970, are both larger and have relatively short processes. S. principalis parva Burmann, 1970, is of similar size but also has relatively short processes. Genus TIMOFEEVIA Vanguestaine, 1978 TYPE SPECIES. Timofeevia lancarae (Cramer & Diez 1972) Vanguestaine 1978. Timofeevia lancarae (Cramer & Diez 1972) Vanguestaine 1978 Fig. 127 1972 Multiplicisphaeridium lancarae Cramer & Diez: 42; pl. 1, figs 1-4, 6, 8; text-fig. 1. 1978 Timofeevia lancarae (Cramer & Diez) Vanguestaine: 272. MATERIAL. One specimen. OCCURRENCE. Pontyfenni Formation: MPA 20099. DESCRIPTION. The vesicle is subspherical and its surface is divided into a number of polygonal fields by dark folds or protuberances. Twenty-five processes are visible, situated at the angles of the polygonal fields. They have hollow, cylindrical stems and divide distally to the fourth order. The terminal branches are fine filaments which, in some cases, link adjacent processes. ACRITARCHS AND CHITINOZOA 347 DIMENSIONS. Vesicle diameter 32 x 28 um; process length 12 um. REMARKS. The elaborate distal process terminations of Timofeevia lancarae are diagnostic. The species has been recorded from the early Middle Cambrian to earliest Tremadoc (?) in Spain (Cramer & Diez 1972; Fombella 1978), and from the Middle and Upper Cambrian of Random Island, eastern Newfoundland (Martin & Dean 1981). Its presence in the Pontyfenni Formation is probably on account of reworking from rocks of Cambrian or, less likely, Tremadoc age. Genus UNCINISPHAERA Wicander, 1974 TYPE SPECIES. Uncinisphaera lappa Wicander 1974. REMARKS. A number of species recorded from the Arenig succession of south-west Wales have similar features and are tentatively referred to the genus Uncinisphaera Wicander, 1974. The diagnostic characteristics of Uncinisphaera are a granulate wall, spherical vesicle and ornament- ed processes: at least two of these features are present on each of the taxa considered here. It is possible that all the species described herein are related to each other, but unlikely that they are closely related to the Devonian species Uncinisphaera lappa Wicander 1974 and U. acantha Wicander & Wood 1981. Uncinisphaera is distinguished from Stellechinatum by its spherical or subspherical rather than polygonal vesicle. Uncinisphaera ? sp. A Figs 100-101, 119-120 MATERIAL. Three specimens. OccuURRENCE. Pontyfenni Formation: MPA 20099. DESCRIPTION. The vesicle is large, subspherical and, although poorly preserved, appears to have a granulate wall. There are about 40—SO slender, acuminate, moderately flexible and slightly tapering processes. They are relatively short, their average length being about one quarter of the vesicle diameter, possibly solid and bear an ornament of short lateral hairs. DIMENSIONS. Vesicle diameter range 33—48 wm, mean 40 um; process length range 9-12 um. REMARKS. The apparently granulate wall and ornamented processes would justify assignment Uncinisphaera, but the determination is tentative because of poor preservation. U. lappa is smaller and has fewer, stouter, more conical processes. U. acantha is smaller and has fewer, relatively long processes. For comparisons with other taxa see the remarks under the following species of Uncinisphaera. Uncinisphaera ? sp. B Figs 102, 104, 121-122 MATERIAL. Seven specimens. OCCURRENCE. Pontyfenni Formation: MPA 20099. DESCRIPTION. The vesicle is large and subspherical, with convex or concave sides between the process bases. There are about 50 relatively short, stiff or slightly flexible and evexate processes, which bear an ornament of short hairs; they average about one-ninth of the vesicle diameter in length. DIMENSIONS. Vesicle diameter range 46-66 um, mean 54m; process length range 5—10 um, mean 6 um. REMARKS. Uncinisphaera ? sp. B is distinguished from Uncinisphaera ? sp. A by its larger size and shorter processes. The two taxa occur together in the Pontyfenni Formation, but no gradation occurs between them. 348 Ss. G. MOLYNEUX ACRITARCHS AND CHITINOZOA 349 Uncinisphaera ? sp. C Figs 103, 105, 123 MATERIAL. Nine specimens. OCCURRENCE. Pontyfenni Formation: MPA 20091. DESCRIPTION. The vesicle is subspherical with straight or convex sides. There are about 20 slender, simple, stiff, tapering and evexate processes, with an ornament of fine hairs or grana. One specimen apparently has an equatorial split. DIMENSIONS. Vesicle diameter range 34-46 um, mean 41 ym; process length range 4-10 um, mean 8 pm. REMARKS. Uncinisphaera ? sp. A has longer processes and a more prominent ornament. Uncinisphaera ? sp. B is larger and also has a more prominent ornament on the processes. U. lappa and U. acantha have smaller vesicles. For other comparisons see the remarks under Uncinisphaera ? sp. D. Uncinisphaera ? sp. D Figs 106-109, 124 MATERIAL. Nine specimens. OCCURRENCE. Cwmffrwd Member: MPA 20077, MPA 20087(?). Allt Cystanog Member: MPA 20074(?). DESCRIPTION. The vesicle is large and subspherical with convex sides. The few, short, slender processes are relatively stiff, tapering and evexate. The number of processes is variable, between seven and seventeen; the lower number may reflect poor preservation. DIMENSIONS. Vesicle diameter range 39—44 um, mean 41 um; process length range 5—8 wm, mean 6 um. REMARKS. Uncinisphaera ? sp. D is very similar to Uncinisphaera ? sp. C, but may be distin- guished by its fewer, shorter processes. The differences are very slight, however, and it is possible that the two forms are conspecific. One specimen from the Allt Cystanog Member has shorter processes than the specimens of Uncinisphaera ? sp. D from the Cwmffrwd Member, but is otherwise similar. The presence of the species in the upper part of the Cwmffrwd Member (MPA 20087) is questionable because it is based on the occurrence of a few, poorly preserved specimens. Uncinisphaera ? sp. E Figs 110-111, 125 MATERIAL. Two specimens. OCCURRENCE. Cwmffrwd Member: MPA 20077. DESCRIPTION. The vesicle is subspherical to subpolygonal. The processes are relatively short, conical and acuminate, and bear short lateral hairs. About 30 processes, about one third of the vesicle diameter in length, are present. DIMENSIONS. Vesicle diameter range 30—32 wm, mean 30-5 um; process length 8-10 um. REMARKS. Uncinisphaera ? sp. A is larger and has more numerous and more slender processes. Figs 100, 101 Uncinisphaera? sp. A; sample MPA 20099, Pontyfenni Formation; x 1200. Fig. 100, MPK 4939; slide 1, M31/2. See Fig. 119. Fig. 101, MPK 4940; slide 1, N38/0. See Fig. 120. Figs 102, 104 Uncinisphaera? sp. B; sample MPA 20099, Pontyfenni Formation; x 1200. Fig. 102A, B, MPK 4941; slide 1, M24/0. See Fig. 122. Fig. 104, MPK 4943, slide 1, R27/0. See Fig. 121. Figs 103, 105 Uncinisphaera? sp. C; sample MPA 20091, Pontyfenni Formation; x 1200. Fig. 103, MPK 4942; slide 1, X42/1. Fig. 105A, B (B phase contrast), MPK 4944; slide 1, Q43/0. See Fig. 123. 350 S. G. MOLYNEUX ad 110b Figs 106-108 Uncinisphaera? sp. D; sample MPA 20077, Cwmffrwd Member; x 1200. Fig. 106, MPK 4945; slide 2, M32/1. See Fig. 124. Fig. 107, MPK 4946; slide 2, R38/3. Fig. 108, MPK 4947; slide 2, L25/1. Fig. 109 cf. Uncinisphaera? sp. D; MPK 4948, sample MPA 20074, Allt Cystanog Member; slide 1, F22/0, x 1200. Figs 110-111 Uncinisphaera? sp. E; sample MPA 20077, Cwmffrwd Member; x 1200. Fig. 110A, B (B phase contrast), MPK 4949; slide 2, J42/0. See Fig. 125. Fig. 111, MPK 4950; slide 2, P29/1. ACRITARCHS AND CHITINOZOA 351 112a Figs 112-115 Uncinisphaera? sp. F; sample MPA 20103, Cwm yr Abbey Member; x 1200. Fig. 112A, B, high and low focus, MPK 4951; slide 2, G21/2. See Fig. 126. Fig. 113, MPK 4952; slide 2, J56/0. Fig. 114, MPK 4953; slide 2, JS8/4. Fig. 115, MPK 4954; slide 2, S64/0. Figs 116-118 Uncinisphaera? spp.; Cwm yr Abbey Member; x 1200. Fig. 116, MPK 4955, sample MPA 20084; slide 2, K44/0. Fig. 117, MPK 4956, sample MPA 20103; slide 1, E33/0. Fig. 118, MPK 4957, sample MPA 20084; slide 2, G35/3. 352 S. G. MOLYNEUX 120 Figs 119-120 Uncinisphaera? sp. A. Bar represents 30 um. Fig. 119, MPK 4939. See Fig. 100. Fig. 120, MPK 4940. See Fig. 101. Uncinisphaera ? sp. F Figs 112-115, 126 MATERIAL. Four specimens. OCCURRENCE. Cwm yr Abbey Member: MPA 20103. DESCRIPTION. The vesicle is subspherical and its sides are either masked by the process bases or concave. About 30 conical, relatively short, acuminate and granulate processes, with an average length of one quarter of the vesicle diameter, are present. DIMENSIONS. Vesicle diameter range 28—35 wm, mean 33 um; process length c. 8 um. REMARKS. Preservation is generally poor and three out of the four specimens are broken. The species may be readily distinguished by its broad conical processes which tend to mask the vesicle sides. Uncinisphaera ? spp. Figs 116-118 REMARKS. Three poorly preserved specimens from the Cwm yr Abbey Member are included in the genus. 121 Figs 121-122 Uncinisphaera? sp. B. Bar represents 30 um. Fig. 121, MPK 4943. See Fig. 104. Fig. 122, MPK 4941. See Fig. 102. ACRITARCHS AND CHITINOZOA 353 123 124 Fig. 123. Uncinisphaera? sp. C; MPK 4944. Bar represents 30 pm. See Fig. 105. Fig. 124 Uncinisphaera? sp. D; MPK 4945. Bar represents 30 yum. See Fig. 106. Two specimens from MPA 20084 have subspherical vesicles with straight or convex sides. The processes, numbering 30—40, are relatively short, slender and moderately flexible, and bear a sparse ornament of fine hairs. The vesicle diameters are 38 x 39 wm (Fig. 116) and 48 x 37 um (Fig. 118), and the average process length is 64m. The specimens may represent the same species but this cannot be confirmed because of their poor preservation. One specimen (Fig. 117), from MPA 20103, has a large subspherical vesicle with straight or convex sides and a granulate surface. The processes, of which there are at least 21, are long, stout and moderately flexible, and are covered in relatively long hairs; their distal terminations are evexate or consist of two short, recurved hairs. The vesicle diameter is about 50 um, and process length 16 um. Genus VERYHACHIUM Deunff 1954 ex Downie 1959 TYPE SPECIES. Veryhachium trisulcum (Deunff 1951) Deunff 1959 ex Downie 1959. ‘Veryhachium trispinosum’ group Figs 129, 131, 133 REMARKS. Three-spined species of Veryhachium are recorded from the Cwmffrwd Member of the Carmarthen Formation and throughout the succeeding part of the Arenig Series (Fig. 5, p. 318). They are most common in the Fennian. 125 126 Fig. 125 Uncinisphaera? sp. E; MPK 4949. Bar represents 30 ym. See Fig. 110. Fig. 126 Uncinisphaera? sp. F; MPK 4951. Bar represents 30 um. See Fig. 112. 354 S. G. MOLYNEUX 133 ACRITARCHS AND CHITINOZOA 355 Genus VOGTLANDIA Burmann, 1970 TYPE SPECIES. Vogtlandia ramificata Burmann 1970. ?Vogtlandia flosmaris (Deunff 1977) Dean & Martin 1978 Fig. 128 21977 Evittia flosmaris Deunff: 143; pl. 1, fig. 18; pl. 2, figs 7, 9, 11, 14. 21978 Vogtlandia coalita Dean & Martin: 9-10; pl. 2, figs 3, 7. 21978 Vogtlandia flosmaris (Deunff) Dean & Martin: 19. 21982 Vogtlandia flosmaris (Deunff) Dean & Martin: Martin; pl. 1, fig. 3. MATERIAL. One specimen. OccuRRENCE. Allt Cystanog Member: MPA 20074. DESCRIPTION. The vesicle is subpolygonal in outline but the vesicle sides are masked to some extent by the wide process bases. The processes, of which there are 12, have stout stems and divide distally to the fourth order, terminating in long, slender filaments. DIMENSIONS. Vesicle diameter: 24 um x 24 um. Process length (overall): 14 um. Process length from proximal contact to first bifurcation: 8 wm. Process length from first bifurcation to distal termination: 6 ym. REMARKS. The specimen has relatively long distal filaments like those on the specimens illus- trated by Deunff (1977) and Martin (1982), but there is no evidence they are intertwined (cf. Dean & Martin 1978: 10; pl. 2, figs 3, 7). This difference might be explained by the poor preservation of the specimen from the Allt Cystanog Member. Genus VULCANISPHAERA Deunff, 1961 TYPE SPECIES. Vulcanisphaera africana Deunff 1961. Vulcanisphaera britannica Rasul 1976 Fig. 130 1976 Vulanisphaera britannica Rasul: 482-484; pl. 1, figs 2, 7-9, 13-16; text-fig. 1: 4, 5. MATERIAL. One specimen. OccuRRENCE. Pontyfenni Formation: MPA 20099. DESCRIPTION. The vesicle is subpolygonal and bears at least twenty processes, which have stout stems dividing distally into three or four branches. All the branches arise from a common point and are acuminate and possibly acicular. DIMENSIONS. Vesicle diameter 32 x 28 xm; process length c. 12 um. Fig. 127 Timofeevia lancarae (Cramer & Diez 1972) Vanguestaine 1978; MPK 4958, sample MPA 20099, Pontyfenni Formation; slide 1, N34/4, x 1200. Figs 129, 131, 133. ‘Veryhachium trispinosum group; x 1200. Fig. 129, MPK 4960, sample MPA 20103, Cwm yr Abbey Member; slide 2, D74/4. Figs 131, 133, sample MPA 20091, Pontyfenni Formation. Fig. 131, MPK 4962; slide 1, U29/2. Fig. 133, MPK 4964; slide 1, 046/0. Fig. 128 ?Vogtlandia flosmaris (Deunff 1977) Dean & Martin 1978; MPK 4959, sample MPA 20074, Allt Cystanog Member; slide 2, K51/4, x 1200. Fig. 130 Vulcanisphaera britannica Rasul 1976; MPK 4961, sample MPA 20099, Pontyfenni Formation; slide 1, G26/2, x 1200. Figs 132, 134 Vulcanisphaera turbata? Martin in Martin & Dean 1981; x 1200. Fig. 132, MPK 4963, sample MPA 20099, Pontyfenni Formation; slide 1, L37/3. Fig. 134, MPK 4965, sample MPA 20103, Cwm yr Abbey Member; slide 1, L38/2. 356 S. G. MOLYNEUX REMARKS. The specimen is most like Vulcanisphaera britannica, forma 2 of Rasul (1976), record- ed from the Shineton Shales (of Tremadoc age), Shropshire. It indicates probable reworking of Tremadoc forms into the lower part of the Pontyfenni Formation. Vulcanisphaera turbata? Martin 1981 Figs 132, 134 21981 Vulcanisphaera turbata Martin in Martin & Dean: 23-24; pl. 1, figs 2-4; text-fig. 6. MATERIAL. Two specimens. OCCURRENCE. Cwm yr Abbey Member: MPA 20103. Pontyfenni Formation: MPA 20099. DESCRIPTION. The vesicle is subspherical. The processes are rarely single, more usually grouped into clusters of two, three or four. Filamentous threads arise laterally from some of the pro- cesses. A number of polygonal fields are delimited by dark folds or protuberances on the surface of the vesicle. DIMENSIONS. Vesicle diameters 40 x 36 um and 48 x 45 um; process length 8 x 10 um. REMARKS. The specimens resemble V. turbata, and also show some similarity to Vulcanisphaera cirrita Rasul, 1976, and V. africana Deunff, 1961, as understood by Martin (in Martin & Dean, 1981). The determination is tentative on account of poor preservation. V. turbata was described from late Middle Cambrian and Upper Cambrian rocks of Random Island, eastern Newfound- land, and its presence may indicate the possible reworking of Cambrian forms into the Arenig Series of south-west Wales. Systematic descriptions: Chitinozoa The descriptive terminology used in this section is that of Laufeld (1974). The general remarks on p. 319-20 apply here also. Genus BELONECHITINA Jansonius, 1964 TYPE SPECIES. Conochitina micracantha robusta Eisenack 1959. Belonechitina spp. Figs 135-137, 139 Eight specimens from the Pontyfenni Formation (MPA 20091) are assigned to this genus. There are two distinct forms, distinguished by size, but it is possible that they represent a single species. DESCRIPTIONS. Four specimens have cylindro-conical vesicles which are poorly divided into a chamber and a neck by weakly-developed flexures and shoulders. The chamber is subconical, the basal edge rounded and the base flat to slightly convex. The neck is cylindrical and slightly narrower than the chamber. The vesicle bears an ornament of short hairs and grana which are most prominent on the basal edge. Dimensions of the four specimens are as follows: vesicle length 76-84 ym, mean 81 wm; chamber width 61—72 ym, mean 66 um; neck width 34-53 um, mean 44 yum. (See Figs 137, 139). The other four specimens also have cylindro-conical vesicles, but the neck is not differen- tiated from the chamber, or is only poorly differentiated by a weakly-developed flexure. The basal edge is rounded and the base is flat or concave. The neck of all four specimens is broken orally. An ornament of fine hairs is present on the vesicle. Dimensions of these specimens are: vesicle length 118-152 um, mean 133 wm; chamber width 57-74 wm, mean 65 um; neck width 29-48 um, mean 34 pm. (See Figs 135, 136). REMARKS. Belonechitina henryi Paris 1981 has a longer vesicle than either form from the Pontyfenni Formation. B. micracantha typica (Eisenack 1965) has similar dimensions, is vari- able in shape and has an ornament well developed on the basal edge but not on the flanks. All the specimens from the Pontyfenni Formation resemble B. micracantha typica, but the orna- ment may be more widely distributed over the vesicle surface than in that species. ACRITARCHS AND CHITINOZOA 357 ‘a x 135 137a 138a d Se 3 sb 139a « Figs 135-137, 139 Belonechitina spp.; Pontyfenni Formation; x 480 (Fig. 139B x 1200). Fig. 135, MPK 4966, sample MPA 20092; slide 2, L21/0. Fig. 136, MPK 4967, sample MPA 20091; slide 1, K22/4. Fig. 137A, B, high and low focus, MPK 4968, sample MPA 20091; slide 2, H25/3. Fig. 139A, B (B detail of surface ornament x 1200), MPK 4970, sample MPA 20092; slide 2, J22/1. Fig. 138A, B_ Lagenochitina cylindrica? Eisenack 1931, high and low focus; MPK 4969, sample MPA 20091, Pontyfenni Formation; slide 1, U25/3, x 480. Fig. 140 Conochitina cf. chydaea Jenkins 1967; MPK 4971, sample MPA 20092, Pontyfenni Formation; slide 1, P19/0, x 1200. 139b Genus CONOCHITINA Eisenack, 1931 TYPE SPECIES. Conochitina claviformis Eisenack 1931. Conochitina cf. chydaea Jenkins 1967 Fig. 140 cf. 1967 Conochitina chydaea Jenkins: 453-454; pl. 70, figs 4-8. MATERIAL. Six specimens. OccuRRENCE. Pontyfenni Formation: MPA 20091. DESCRIPTION. The vesicle is cylindro-conical, the base flat or slightly convex and the basal edge rounded. Flexures and shoulders are not developed. The neck is poorly developed and may flare aborally of the aperture. 358 S. G. MOLYNEUX DIMENSIONS. Vesicle length: range 93-148 wm; mean 111 pm. Width (chamber): range 44-57 um; mean 53 um. Width (neck): range 30-38 wm; mean 35 ym. Width (aperture): range 23-42 wm: mean 33 pm. REMARKS. Conochitina chydaea is variable in morphology. The specimens recorded from the Pontyfenni Formation bear some resemblance to the type material (Jenkins 1967) from the Llanvirn of Shropshire, but are smaller. Genus LAGENOCHITINA Eisenack, 1931 TYPE SPECIES. Lagenochitina baltica Eisenack 1931. Lagenochitina cylindrica? Eisenack 1931 Fig. 138A, B 21931 Lagenochitina cylindrica Eisenack: 81; pl. 2, figs 18, 19. 21967 Lagenochitina cylindrica Eisenack ; Jenkins: 463; pl. 74, figs 1-3. MATERIAL. One specimen. OccCURRENCE. Pontyfenni Formation: MPA 20091. DESCRIPTION. The vesicle has a conspicuous flexure and shoulder that differentiate the chamber from the neck. The chamber is subcylindrical with a flat base and a rounded basal edge. The neck is cylindrical. DIMENSIONS. Vesicle length 112 um; length of chamber 65 um; length of neck 48 um; width of chamber 48 um; width of neck 36 um. REMARKS. The specimen is small for Lagenochitina cylindrica but resembles those illustrated by Jenkins (1967) from the Llanvirn of Shropshire. Lagenochitina sp. A Figs 141-143 MATERIAL. Thirty specimens. OCCURRENCE. Pontyfenni Formation: MPA 20091 -3. DESCRIPTION. The chamber is subcylindrical to subconical and rarely sphaeroidal. The flexure is usually more conspicuous than the shoulder, the latter being broadly rounded. The basal edge may or may not be present; when present it is convex and rounded to broadly rounded. The base is flat or convex. The neck is long and cylindrical, widening slightly near the aperture. The length of the neck is usually a little over half the total length of the vesicle. The aperture is fringed by short spines but the vesicle is otherwise smooth. DIMENSIONS. Vesicle length: range 110-162 um; mean 141 ym. Length of neck: range 53-99 wm; mean 76 um. Length of chamber: range 49-84 ym; mean 66 um. Width of neck: range 29-42 um; mean 36 um. Width of chamber: range 44-68 wm; mean 59 um. Width of aperture: range 29-42 wm; mean 38 um. REMARKS. Lagenochitina sp. A is the most commonly occurring chitinozoan in the Pontyfenni Formation near Maesyrwyn. It is unlike other species of Lagenochitina, being distinguished by its relatively long neck, fringed aperture and the shape of its chamber. Lagenochitina obeligis Paris 1981 is larger, has a relatively shorter neck, an ovoidal chamber and a more convex base, and its aperture is not fringed. L. esthonica Eisenack 1955 has a very characteristic shape that distinguishes it from the present Lagenochitina sp. A. L. shelvensis ACRITARCHS AND CHITINOZOA 359 _ 4 441 442 Figs 141-143 Lagenochitina sp. A; Pontyfenni Formation; x 480 (Fig. 143 x 1200). Fig. 141, MPK 4976, sample MPA 20091; slide 2, T31/2. Fig. 142, MPK 4975, sample MPA 20092; slide 1, P40/3. Fig. 143, detail of aperture x 1200, MPK 4974, sample MPA 20091; slide 2, G32/3. Jenkins 1967 and Lagenochitina sp. of Achab (1982) are both larger than the present form, and both have relatively shorter necks. Achab’s Lagenochitina sp. also has a more or less quad- rangular chamber. Acknowledgements The author is grateful to Dr R. A. Fortey, Dr. A. W. A. Rushton and Dr. P. M. Allen for discussion and support, and to Dr G. A. Booth for access to his unpublished data. Mrs J. Lines typed most of the manuscript. This paper is published with the permission of the Director, British Geological Survey (N.E.R.C.). Sample localities 1. Ogof Hen Formation: Allt Cystanog Member. MPA 20074. 140 m at 112° from Star Cottage (SN 4305 1979). Loc. 4b of Owens & Fortey (1982). 2. Ogof Hén Formation: Bolahaul Member. MPA 20075-6. 10m at 227° from Star Cottage (SN 4291 1981). Loc. 8 Fortey & Owens (1978). MPA 20079. Roman Road, Pensarn. 105m at 140° from chapel at Pensarn (SN 4141 1911). Loc. 6 of Fortey & Owens (1978). MPA 20080. Roman Road, Pensarn. 80m at 149° from chapel at Pensarn (SN 4137 1914). Loc. 6 of Fortey & Owens (1978). 3. Carmarthen Formation: Cwmffrwd Member. MPA 20077-8. Nantycaws dingle. 410m at 233° from Ty-cerig (SN 4510 1844). Loc. 7 of Owens & Fortey (1982). MPA 20086. Nant y Glasdwr. 50m at 130° from Nant-y-Glasdwr-fach (SN 4239 1756). Loc. 3E of Fortey & Owens (1978). MPA 20087. Nant y Glasdwr. 95m at 131° from Nant-y-Glasdwr-fach (SN 4242 1753). Loc. 3D of Fortey & Owens (1978). 4. Carmarthen Formation: Cwm yr Abbey Member. MPA 20081. Nant y Glasdwr. 40m at 169° from Gwynion Dale (SN 4272 1748). Loc. 3A of Fortey & Owens (1978). MPA 20082. Nant y Glasdwr. 55m at 218° from Gwynion Dale (SN 4267 1748). Loc. 3A of Fortey & Owens (1978). MPA 20083. Nant y Glasdwr. 75m at 228° from Gwynion Dale (SN 4265 1747). Loc. 3A of Fortey & Owens (1978). 360 S. G. MOLYNEUX MPA 20084. Nant y Glasdwr. 100m at 235° from Gwynion Dale (SN 4263 1746). Loc. 3A of Fortey & Owens (1978). MPA 2008S. Nant y Glasdwr. 120m at 238° from Gwynion Dale (SN 4261 1746). Loc. 3A of Fortey & Owens (1978). MPA 20088. Nant y Glasdwr. 160m at 252° from Gwynion Dale (SN 4256 1747). Near loc. 3B of Fortey & Owens (1978). MPA 20089. Nant y Glasdwr. 165m at 251° from Gwynion Dale (SN 4255 1747). Near loc. 3B of Fortey & Owens (1978). MPA 20090. Nant y Glasdwr. 185m at 257° from Gwynion Dale (SN 4253 1748). Near loc. 3B of Fortey & Owens (1978). MPA 20103. Cwm yr Abbey. 400m at 322° from Abbey Farm (SN 5002 1978). Loc. 16 of Fortey & Owens (1978). 5. Afon Ffinnant Formation. MPA 20104-5. Cwm yr Abbey. 400m at 322° from Abbey Farm (SN 5001 1979). Loc. 16 of Fortey & Owens (1978, 1987). 6. Colomendy Formation: Whitland Abbey Member. MPA 20094. 340m at 290° from Whitland Abbey (SN 2060 1822). Near loc. 27 of Fortey & Owens (1987). MPA 20095. 210m at 287° from Whitland Abbey (SN 2072 1818). Near loc. 27 of Fortey & Owens (1987). MPA 20096. 210m at 285° from Whitland Abbey (SN 2072 1818). Near loc. 27 of Fortey & Owens (1987). MPA 20097. 210m at 284° from Whitland Abbey (SN 2072 1817). Near loc. 27 of Fortey & Owens (1987). MPA 20098. 200m at 278° from Whitland Abbey (SN 2072 1815). Near loc. 27 of Fortey & Owens (1987). 7. Pontyfenni Formation. MPA 20091. 560 m at 260° from Maesyrwyn (SN 2379 1694). Loc. 23 of Fortey & Owens (1987). MPA 20092. 540 m at 262° from Maesyrwyn (SN 2381 1696). Loc. 23 of Fortey & Owens (1987). MPA 20093. 520 m at 264° from Maesyrwyn (SN 2383 1699). Loc. 23 of Fortey & Owens (1987). MPA 20099-100. 400m at 353° from Penlan (SN 1984 1948). Loc. 38 of Fortey & Owens (1987). MPA 20101—2. 400m at 356° from Penlan (SN 1986 1949). Loc. 38 of Fortey & Owens (1987). References Achab, A. 1982. Chitinozoaires de l’Arenig supérieur (Zone D) de la Formation de Lévis, Québec, Canada. Can. J. Earth Sci., Ottawa, 19 (6): 1295-1307. Arthurton, R. S. & Wadge, A. J. 1981. Geology of the country around Penrith. Mem. Geol. Surv. G. B., London, Sheet 24. 177 pp. Babin, C., Arnaud, A., Blaise, J., Cavet, P., Chauvel, J. J., Deunff, J., Henry, J.-L., Lardeux, H., Melou, M., Nion, J., Paris, F., Quete, Y. & Robardet, M. 1976. The Ordovician of the Armorican Massif (France). In Bassett, M. G. (ed.), The Ordovician System: proceedings of a Palaeontological Association symposium, Birmingham, September 1974: 359-385. Cardiff, Univ. Wales & Natl Mus. Wales. Benoit, A. & Taugourdeau, P. 1961. Sur quelques chitinozoaires de l’Ordovicien du Sahara. Revue Inst. fr. Petrole, Paris, 26: 1403-1421. Bockelie, T. G. 1980. Early Ordovician chitinozoa from Spitsbergen. Palynol., Dallas, 4: 1-14. Booth, G. A. (1979). Lower Ordovician acritarchs from successions in England and North Wales. Ph.D. thesis, Univ. Sheffield (unpubl.). Burmann, G. 1970. Weitere organische Mikrofossilien aus dem unteren Ordovizium. Palaont. Abh. Berl., (B) 3: 289-325. Cocks, L. R. M. & Fortey, R. A. 1982. Faunal evidence for oceanic separations in the Palaeozoic of Britain. J. geol. Soc. Lond., 139: 465-478. Combaz, A. 1967. Un microbios du Trémadocien dans un sondage d’Hassi—Messaoud. Act. Soc. linn. Bordeaux, (B) 104 (29): 1-26. & Peniguel, G. 1972. Etude palynostratigraphique de l’Ordovicien dans quelques sondages du Bassin de Canning (Australie occidentale). Bull. Cent. Rech. Pau, 6 (1): 121-167. Cooper, R. A. & Fortey, R. A. 1982. The Ordovician graptolites of Spitsbergen. Bull. Br. Mus. nat. Hist., London, (Geol.) 36 (3): 157-302, pls 1-6. ACRITARCHS AND CHITINOZOA 361 Cramer, F. H. 1970. Distribution of selected Silurian acritarchs. Revta esp. Micropaleont., Madrid, num. extraord. 203 pp., pls 1—23. ——,, Allam, B., Kanes, W. H. & Diez, M. del C. R. 1974. Upper Arenigian to lower Llanvirnian acritarchs from the subsurface of the Tadla Basin in Morocco. Palaeontographica, Stuttgart, (B) 145: 182-190. —— & Diez, M. del C. R. 1972. Acritarchs from the Upper Middle Cambrian Oville Formation of Leon, northwestern Spain. Revta esp Micropaleont, Madrid, num. extraord. (XXX Aniv. E. N. Adaro): 39-50. —— —— 1976. Seven new late Arenigian species of the Acritarch Genus Coryphidium Vavrdova, 1972. Palaont. Z., Stuttgart, 50: 201-208. 1977. Late Arenigian (Ordovician) acritarchs from Cis-Saharan Morocco. Micropaleontology, New York, 23 (3): 339-360. ——., Kanes, W. H., Diez, M. del C. R. & Christopher, R. A. 1974. Early Ordovician acritarchs from the Tadla Basin of Morocco. Palaeontographica, Stuttgart, (B) 146: 57-64. Dean, W. T. & Martin, F. 1978. Lower Ordovician acritarchs and trilobites from Bell Island, eastern Newfoundland. Bull. geol. Surv. Can., Ottawa, 284. 35 pp., 7 pls. Deflandre, G. 1937. Microfossiles des silex crétacés. II. Flagellés incertae sedis. Hystrichosphaeridés. Sarcodinés. Organismes divers. Annls Paleont., Paris, 26: 51-103 (3-55), pls 8-15 (11-18). —— 1938. Microplancton des mers jurassiques conserve dans les marnes de Villers-sur-Mer (Calvados). Etude preliminaire et considérations générales. Trav. Stn zool. Wimereux, Paris, 13: 147-211, pls 5-11. — 1942. Sur les Hystrichospheres des calcaires Siluriens de la Montagne Noire. C. r. hebd. Seanc. Acad. Sci., Paris, 215: 475-476. —— 1945. Microfossiles des calcaires Siluriens de la Montagne Noire. Annls Paleont., Paris, 31: 41-75. —— & Deflandre-Rigaud, M. 1962. Nomenclature et systematiques des Hystrichosphéres (sens. lat.). Observations et rectifications. Revue Micropaleont., Paris, 4: 190-196. Deunff, J. 1951. Sur la présence de microorganismes (Hystrichospheres) dans les schistes ordoviciens du Finistére. C. r. hebd. Seanc. Acad. Sci., Paris, 233: 321-323. — 1954. Veryhachium, genre nouveau d’Hystrichosphéres du Primaire. C. r. somm. Seanc. Soc. geéol. Fr., Paris, 13: 305-307. — 1959. Microorganismes planctoniques du Primaire armoricain. I. Ordovicien du Veryhac’h (Presquile de Crozon). Bull. Soc. geol. miner. Bretagne, Rennes, (n.s.) 1958 (2): 1—41. — 1961. Un microplancton a Hystrichosphéres dans le Trémadoc du Sahara. Revue Micropaleont., Paris, 4: 37-52, pls 1-3. —— 1964. Systématique du microplancton fossile a acritarches. Révision de deux genres de l’Ordovicien inférieur. Revue Micropaleont., Paris, 7: 119-124. — 1977. Un microplancton a acritarches dans les schistes llanvirniens de l’anti-Atlas (Zagore—Maroc). Notes Mem. Serv. Mines Carte geéol. Maroc, Rabat, 38 (268): 141-151. —, Gorka, H. & Rauscher, R. 1974. Observations nouvelles et précisions sur les acritarches a large ouverture polaire du Paleozoique inférieur. Geobios, Lyon, 7: 5-18. Dhonau, T. J. 1982. Notes for Authors. iv + 35 pp. London, Institute of Geological Sciences. Downie, C. 1958. An assemblage of microplankton from the Shineton Shales (Tremadocian). Proc. Yorks. geol. Soc., Hull, 31: 331-350. — 1959. Hystrichospheres from the Silurian Wenlock Shale of England. Palaeontology, London, 2: 56-71. —— & Sarjeant, W. A. S. 1965. Bibliography and index of fossil dinoflagellates and acritarchs. Mem. geol. Soc. Am., Washington, 94: 1-180. Eisenack, A. 1931. Neue Mikrofossilien des baltischen Silurs I. Palaont. Z., Berlin, 13: 74-118. — 1955. Neue Chitinozoen aus dem Silur des Baltikums und dem Devon der Eifel. Senckenberg. leth., Frankfurt a.M., 36: 311-319. —— 1959. Neotypen baltischer Silur — Chitinozoen und neue Arten. Neues Jb. Geol. Palaont. Abh., Stuttgart, 108: 1—20. —— 1963. Mitteilungen zur Biologie der Hystrichospharen und tiber neue Arten. Neues Jb. Geol. Palaont. Abh., Stuttgart, 118: 207-216. —— 1965. Die Mikrofauna der Ostseekalke. I. Chitinozoen, Hystrichospharen. Neues Jb. Geol. Palaont. Abh., Stuttgart, 123: 115-148. — 1968. Mikrofossilien eines Geschiebes der Borkholmer Stufe, baltisches Ordovizium F,. Mitt. geol. StInst. Hamb., 37: 81-94. , Cramer, F. H. & Diez, M. del C. R. 1976. Katalog der fossilen Dinoflagellaten, Hystrichospharen und verwandten Mikrofossilien, 4 (Acritarcha 2). xxiv + 863 pp. Stuttgart. Fombella, M. A. 1978. Acritarcos de la Formacion Oville, edad Cambrico Medio-Tremadoc, Provincia de Leon, Espana. Palinologia, Leon, nam. extraord. 1: 245-261. 362 S. G. MOLYNEUX Fortey, R. A. & Owens, R. M. 1978. Early Ordovician (Arenig) stratigraphy and faunas of the Carmarthen district, south-west Wales. Bull. Br. Mus. nat. Hist., London, (Geol.) 30 (3): 225-294, pls 1-11. 1984. A synopsis of the Arenig Series in South Wales. Proc. Geol. Ass., London, 95 (4): 389-390. —— —— 1987. The Arenig Series in South Wales. Bull. Br. Mus. nat. Hist., London, (Geol.) 41 (3): 69-307. Grahn, Y. 1980. Early Ordovician chitinozoa from Oland. Sver. geol. Unders. Afh., Uppsala, (C) 775: 1-41. Jackson, D. E. 1978. The Skiddaw Group. In Moseley, F. (ed.), The Geology of the Lake District: 79-98, pls 6-9. Leeds, Yorkshire Geol. Soc. (Occas. publ. 3). Jansonius, J. 1964. Morphology and classification of some Chitinozoa. Bull. Can. Petrol. Geol., Calgary, 12: 901-918. Jenkins, W. A. M. 1967. Ordovician Chitinozoa from Shropshire. Palaeontology, London, 10 (3): 436-488. Kjellstrom, G. 1971. Middle Ordovician microplankton from the Grotlingbo Borehole No. 1 in Gotland, Sweden. Sver. geol. Unders. Afh., Stockholm, (C) 669: 1-35. —— 1976. Lower Viruan (Middle Ordovician) microplankton from the Ek6n Borehole No. 1 in Ostergotland, Sweden. Sver. geol. Unders. Afh., Stockholm, (C) 724: 1-43. Laufeld, S. 1974. Silurian Chitinozoa from Gotland. Fossils Strata, Oslo, 5: 1-130. Lister, T. R. 1970. The acritarchs and chitinozoa from the Wenlock and Ludlow Series of the Ludlow and Millichope areas, Shropshire, (1): 1-100, pls 1-13. Palaeontogr. Soc. (Monogr.), London. Loeblich, A. R. & Tappan, H. 1976. Some new and revised organic-walled phytoplankton microfossil genera. J. Paleont., Tulsa, 50 (2): 301—308. Martin, F. 1965. Les acritarches de Sart-Bernard (Ordovicien Belge). Bull. Soc. belge Geol. Paleéont. Hydrol., Brussels, 74: 423-444. —— 1969a. Chitinozoaires de l’Arenig supérieur—Llanvirn inférieur en Condroz (Belgique). Revue Micro- paleont., Paris, 12 (2): 99-106. 1969b (for 1968). Les acritarches de l’Ordovicien et du Silurien belges. Determination et valeur stratigraphique. Mem. Inst. r. Sci. nat. Belg., Brussels, 160: 1—75. 1972. Les acritarches de lOrdovicien inférieur de la Montage Noire (Hérault, France). Bull. Inst. r. Sci. nat. Belg., Brussels, (Sci. Terre) 48 (10): 1-61. 1982. Some aspects of late Cambrian and early Ordovician acritarchs. In Bassett, M. G. & Dean, W. T. (eds), The Cambrian—Ordovician boundary: sections, fossil distributions and correlations: 29-40. Cardiff, Natl Mus. Wales (Geol. Ser. 3). —— & Dean, W. T. 1981. Middle and Upper Cambrian and Lower Ordovician acritarchs from Random Island, eastern Newfoundland. Bull. geol. Surv. Can., Ottawa, 343: 1-43. Molyneux, S. G. 1979. New evidence for the age of the Manx Group, Isle of Man. In Harris, A. L., Holland, C. H. & Leake, B. E. (eds), The Caledonides of the British Isles—reviewed. Spec. Publs geol. Soc. Lond., 8: 415-421. Owens, R. M. & Fortey, R. A. 1982. Arenig rocks of the Carmarthen—Llanarthney district. In Bassett, M. G. (ed.), Geological excursions in Dyfed, south-west Wales: 249-258. Cardiff, Natl Mus. Wales. Paris, F. 1981. Les chitinozoaires dans le paléozoique du sud-ouest de ’Europe. Mem. Soc. geol. miner. Bretagne, Rennes, 26. 412 pp., 41 pls. & Deunff, J. 1970. Le paléoplancton llanvirnien de la Roche-au-Merle (commune de Vieux-Vy-sur- Couesnon, Ille-et-Vilaine). Bull. Soc. geol. miner. Bretagne, Rennes, (C) 2 (1): 25-43. Playford, G. & Martin, F. 1984. Ordovician acritarchs from the Canning Basin, Western Australia. Alcheringa, Sydney, 8: 187-223. Rasul, S. M. 1974. The Lower Palaeozoic acritarchs Priscogalea and Cymatiogalea. Palaeontology, London, 17 (1): 41-63. — 1976. New species of the genus Vulcanisphaera (Acritarcha) from the Tremadocian of England. Micropaleontology, New York, 22 (4): 479-484. — 1979. Acritarch zonation of the Tremadoc Series of the Shineton Shales, Wrekin, Shropshire, England. Palynol., Dallas, 3: 53-72. Rauscher, R. 1968. Chitinozoaires de Il’Arenig de la Montagne Noire (France). Revue Micropaléont., Paris, 11 (1): 51-60. —— 1973. Recherches micropaléontologiques et stratigraphiques dans lOrdovicien et le Silurien en France. Etude des acritarches, des chitinozoaires et des spores. Sciences geol. Inst. Geol. Strasbourg, (Mem.) 38. 224 pp., 12 pls. Staplin, F. L., Jansonius, J. & Pocock, S. A. J. 1965. Evaluation of some acritarchous hystrichosphere genera. Neues Jb. Geol. Palaont. Abh., Stuttgart, 123 (2): 167-201, pls 18-20. Timofeey, B. V. 1958. Uber das Alter sdchsischer Grauwacken. Mikropalaophytologische Untersuchungen von Proben aus der Weesensteiner und Lausitzer Grauwacke. Geologie, Berlin, 7: 826-845. ACRITARCHS AND CHITINOZOA 363 — 1959. Drevneyshaya flora Pribaltiki i ee stratigraficheskoe znachenie. [The most ancient flora of the Prebaltic and its stratigraphical significance]. Trudy vses. neft. nauchno-issled. geol.-razv. Inst., Lenin- grad, 129: 1-319, 25 pls (many of them multiple). [In Russian]. Turner, R. E. 1984. Acritarchs from the type area of the Ordovician Caradoc Series, Shropshire, England. Palaeontographica, Stuttgart, (B) 190: 87-157. Soc., Hull, 42 (3): 405—414. & Wadge, A. J. 1979. Acritarch dating of Arenig volcanism in the Lake District. Proc. Yorks. geol. Vanguestaine, M. 1978. Criteres palynostratigraphiques conduisant a la reconnaissance d’un pli couché Revinien dans le sondage de Grand-Halleux. Annls Soc. geol. Belgiq., Li¢ge, 100: 249-276. Vavrdova, M. 1965. Ordovician acritarchs from Central Bohemia. Vést. usted. Ust. geol., Prague, 40: SDI SE —— 1966. Palaeozoic microplankton from Central Bohemia. Cas. Miner. Geol., Prague, 11: 409-414. — 1972. Acritarchs from Klabava Shales (Arenig). Vést. ustred. Ust. geol., Prague, 47: 79-86. — 1973. New acritarchs from Bohemian Arenig (Ordovician). Vést. ustred. Ust. geol., Prague, 48: 285-289. 1976. Excystment mechanism of Early Paleozoic acritarchs. Cas. Miner. Geol., Prague, 21: 55-63. Wicander, E. R. 1974. Upper Devonian-—lower Mississippian acritarchs and prasinophycean algae from Ohio, U.S.A. Palaeontographica, Stuttgart, (B) 148: 9-43. —— & Wood, G. D. 1981. Systematics and biostratigraphy of the organic-walled microphytoplankton from the Middle Devonian (Givetian) Silica Formation, Ohio, U.S.A. Contr. Ser. Am. Ass. stratigr. Palynologists, Dallas, 8. 137 pp., 17 pls. Index Most species names are not included in this Index. Species are dealt with in alphabetical order, and are listed on pages 309-10. Abbey Farm 360 acanthomorphitic acritarchs 313, 315, 317 Afon Cothi 314 Afon Fenni 315 Afon Ffinnant Formation 310-14, 316, 332-3, 336-7, 339, 360 Afon Gronw 315 Afon Seiont 317 Afon Taf 315 Afon Tywi 314 Allt Cystanog Member 310-11, 313, 316, 320-1, 324-5, 331-2, 339, 349, 355, 359 Andouillé Formation 333 Arenig Series 310-13, 316-19, 323-4, 336, 341, 347, 353, 356 Australia 318-19 Baltic 319 Bangor 317 Belgium 319 Bell Island 319; Group 319 Bergamia rushtoni Biozone 312, 316 Blaencediw Formation 312-13, 316 Bohemia 319 Bolahaul Member 311-13, 316-17, 359 Brachiopod Beds 322 Britain 319, 327; see England British Geological Survey 319-20 Caernarfon 317 Cambrian 317-18, 347, 356 Caradoc Series 344 Cardiff 314 Carmarthen 310-12, 314, 316; Formation 310-11, 313, 316-17, 353, 359 Castelldraenog Member 312-13, 316 Chitinozoa 319, 356-9 Clonograptus tenellus Zone 322 Colomendy Formation 310, 312-13, 316, 360 Cross Fell Inlier 317, 319 Cwmfelin Boeth 315; Formation 312-13, 316 Cwmfirwd 314; Member 310-14, 316-17, 320-7, 339, 349—S0, 353, 359 Cwm yr Abbey 314, 360; Member 310-14, 316-17, 332-5, 339-41, 345-6, 351-2, 355-6, 359 dendroid graptolites 312 Devonian 347 Didymograptus deflexus Biosubzone 319 extensus Biozone 318-19 hirundo Biozone 317 Dionide levigena Biozone 312, 316 England 310, 317, 344 Europe 310, 318-19 Fennian Stage 310, 312-13, 316-19, 324, 333-4, 353 France 318-19 Furcalithus radix Biozone 312, 316 Garth Point 317 Glen Dhoo Flags 317-18 Gondwanan continent 312 Gotland 344 364 S. G. MOLYNEUX Gwynion Dale 314, 359-60 Gymnostomix gibbsii Biozone 312, 316 Haverfordwest 311—12 Hope Shales 319 Iscoed 314 Isle of Man 317-18 Tsograptus gibberulus Biosubzone 319 Klabava 319; Shales 319 Lady Port Banded ‘Group’ 317 Lake District 317, 324, 333-4 Llanfallteg Formation 312-13, 316 Llanvirn Series 312-13, 316, 319, 333-4, 358 Lonan Flags 317-18 Maesyrwyn 358, 360 Menai Straits Inlier 317 Merlinia rhyakos Biozone 312, 316 selwynii Biozone 312, 316 Microfossil Assemblages: I 310, 313, 316, 318 If 310, 313, 316-18 III 310, 314, 316, 318 IV 310, 314-16, 318 V 310, 315-16, 318 VI 310, 315-19 VII 310, 316-19 Montagne Noire 318 Moridunian 310, 312, 316-17, 333-4 Morocco 324, 341 Nantycaws 314; dingle 313, 359 Nant Colomendy 315 Nant Cwmfelin Boeth 315 Nant Cwmffrwd 314 Nant y Glasdwr 314, 359-60 Nant-y-Glasdwr-fach 314, 359 Nant Pibwr 314 Neseuretus Community 312 Newfoundland 317, 319, 347, 356 North Africa 318-19 North America 318 Ogof Hén Formation 310-13, 316, 318, 359 Olenid Community 312 Ordovician 317, 344 Outerside 317 pendent didymograptids 312 Penlan 315, 360 Pensarn 314, 359 Pibwr Member 311-13, 316 Pont-y-Fenni 315 Pontyfenni Formation 310, 312-13, 315-17, 319, 321-2, 324-30, 332, 336-7, 339-47, 349, 355-60 Quebec 319 Ramsey Island 310-11 Random Island 347, 356 Raphiophorid Community 312 Rennes 333 reworking 317-18, 327, 347, 356 Rhyd Henllan Member 312-13, 316 Rokycany 319 scolecodonts 313, 317 shales with D. artus 316 Shelve Inlier 319 Shineton Shales 322, 327, 356 Shropshire 319, 327, 334, 358 Silurian 334 Spain 347 Spitsbergen 319 Stapeleyella abyfrons Biozone 312, 316 Star Cottage 314, 359 Stellechinatum papulessum sp. nov. 341-2 Sweden 319 Tetragraptus Biozones 319 Tremadoc 317-18, 320, 322, 327, 344, 347, 356 Ty-cerig 359 US:S.R 319 Wabana Group 319 Wales 310-13, 317, 319, 323-4, 328, 333-4, 336, 347, 356 Welsh Borderland 333 Whitland 311, 313, 315-6; Abbey 315, 360; Member 310, 312-13, 315-16, 334-6, 360 Whitlandian 310, 312, 316-17, 333 Accepted for publication 14 February 1986 Bulletin of the British Museum (Natural History) Geology Series Most earlier Geology Bulletins are still in print. A full list of available titles can be obtained from Publications Sales (address inside front cover). Vol. 29 No.1 Aspects of mid-Cretaceous stratigraphical micropalaeontology. D. J. Carter & M. B. Hart. 1977. Pp. 1-135, 4 plates, 53 figs. £14.25 Vol. 29 No. 2 The Macrosemiidae, a Mesozoic family of holostean fishes. A. W. H. Bartram. 1977. Pp. 137-234, 4 plates, 53 figs. £10.00 Vol. 29 No.3 The stratigraphy and ammonite fauna of the Upper Lias of Northamptonshire. M. K. Howarth. 1978. Pp. 235-288, 9 plates, 5 figs. £6.00 | Vol. 29 No.4 Fossil Bovidae (Mammalia) of Olduvai Gorge, Tanzania, Part I. A. W. Gentry & A. Gentry. 1978. Pp. 289-446, 41 plates, 34 figs. £17.50 Vol. 30 No. 1 Fossil Bovidae (Mammalia) of Olduvai Gorge, Tanzania. Part II. A. W. Gentry & A. Gentry. 1978. Pp. 1-83, 3 figs. £7.50 | Vol. 30 No.2 A revision of the Miocene Hominoidea of East Africa. P. J. Andrews. 1978. Pp. 85-224, 7 plates, 29 figs. £15.30 Vol. 30 No.3 Early Ordovician (Arenig) stratigraphy and faunas of the Carmarthen district, south-west Wales. R. A. Fortey & R. M. Owens. 1978. Pp. 225-296, 11 plates, 12 figs. £9.60 | Vol. 30 No.4 Macroscopic inclusions of fluid in British fluorites from the mineral collection of the British Museum (Natural History). A. H. Rankin. 1978. Pp. 297-307, coloured frontispiece, 9 plates (7 coloured), 4 figs. £12.00 Vol. 31 No.1 Foraminifera of the Togopi Formation, eastern Sabah, Malaysia. J. E. Whittaker & R. L. Hodgkinson. 1979. Pp. 1-120, 10 plates, 71 figs. £14.00 Vol. 31 No.2 Cretaceous faunas from Zululand and Natal, South Africa. The ammonite family Gaudryceratidae. W. J. Kennedy & H. C. Klinger. 1979. Pp. 121-173. £6.25 Vol. 31 No.3 Benthic community organization in the Ludlow Series of the Welsh Borderland. R. Watkins. 1979. Pp. 175-279. £12.25 Vol. 31 No.4 The ammonites of the English Chalk Rock (Upper Turonian). C. W. Wright. 1979. Pp. 281-330. £6.50 Vol. 32 No.1 Miscellanea: Observations on Cycloclypeus—Provenance of Sivapithecus—Iranian Silurian brachiopods—New English condylarths—Miocene sharks’ teeth—East African isopod—The Singa skull— Carboniferous insects. 1979. Pp. 1-90. - £10.50 Vol. 32 No. 2 Palaeoenvironments and correlations of the Carboniferous rocks in west Fermanagh, Ireland. C. H. C. Brunton & T. R. Mason. 1979. Pp. 91-108, 6 figs, folded map. £4.00 Vol. 32 No.3 The Ordovician trilobite faunas of the Builth-Llandrindod Inlier, central Wales. Part III. C. P. Hughes. 1979. Pp. 109-181, 177 figs. £10.00 Vol. 32 No.4 The stratigraphy and brachiopods of the upper part of the type Caradoc of south Salop. J. M. Hurst. 1979. Pp. 183-304, 557 figs. £18.50 |Vol. 33 No.1 An account of the Ordovician rocks of the Shelve Inlier in west Salop and part of north Powys. W. F. Whittard, F. R. S. (Compiled by W. T. Dean). 1979. Pp. 1-69, 38 figs, frontispiece, coloured map, folded, in pocket. £10.00 | Map available separately £1.00 Vol. 33 No.2 Miscellanea: Lower Carboniferous microproblematicum—Miniature trilobite—Pleistocene bird remains—English Eocene Hyracotherium—Salenia trisuranalis—Antarctic brachiopods—Diphyphyllum and Murchi- | son’s Russian corals—Lebanese amber Neuroptera. 1980. Pp. 71-164. £12.00 |Vol. 33 No. 3. The Caradoc faunal associations of the area between Bala and Dinas Mawddwy, north Wales. M. G. | Lockley. 1980. Pp. 165-235, 105 figs. £9.00 | Vol. 33 No.4 Fossil insects from the Bembridge Marls, Palaeogene of the Isle of Wight, southern England. E. A. | Jarzembowski. 1980. Pp. 237-293, 77 figs. £7.50 |Vol. 33 No.5 The Yorkshire Jurassic fern Phlebopteris braunii (Goeppert) and its reference to Matonia R.Br. T. M. Harris. 1980. Pp. 295-311, 20 figs. £2.75 ‘Vol. 34 No.1 Relative dating of the fossil hominids of Europe. K. P. Oakley. 1980. Pp. 1-63, 6 figs, 17 tables. £8.00 ‘Vol. 34 No.2 Origin, evolution and systematics of the dwarf Acanthoceratid Protacanthoceras Spath, 1923 | (Cretaceous Ammonoidea). C. W. Wright & W. J. Kennedy. 1980. Pp. 65-107, 61 figs. £6.25 ‘Vol. 34 No.3 Ashgill Brachiopoda from the Glyn Ceiriog District, north Wales. N. Hiller. 1980. Pp. 109-216, 408 figs. £14.75 |Vol. 34 No.4 Miscellanea: Upper Palaeozoic Athyrididae brachiopods—New British Cretaceous Epitoniidae— | Microproblematicum Prethocoprolithus—Glabellar structure of asaphid trilobites—New Lower Ordovician bivalve | family—Cretaceous brachiopods—T upus diluculum sp. nov—Revision of Plummerita. 1980. Pp. 217-297. £11.00 |Vol. 35 No. 1 Lower Ordovician Brachiopoda from mid and south-west Wales. M. G. Lockley & A. Williams. 1981. | Pp. 1-78, 263 figs, 3 tables. £10.80 Vol. 35 No.2 The fossil alga Girvanella Nicholson & Etheridge. H. M. C. Danielli. 1981. Pp. 79-107, 8 figs, 3 tables. £4.20 | Vol. 35 No.3 Centenary Miscellanea: Budleigh Salterton brachiopods—Oswald’s Turkish algae—J. A. Moy- Thomas—Burials, bodies and beheadings—Nucleolites clunicularis—Phanerotinus cristatus—Fossil record of teleosts—Neanderthal dating—Hippoporidra edax. 1981. Pp. 109-252. £20.00 Vol. 35 No.4 The English Upper Jurassic Plesiosauroidea (Reptilia) and a review of the phylogeny and classification of the Plesiosauria. D. S. Brown. 1981. Pp. 253-347, 44 figs. £13.00 Vol. 36 No. 1 Middle Cambrian trilobites from the Sosink Formation, Derik—Mardin district, south-eastern Turkey. W. T. Dean. 1982. Pp. 1-41, 68 figs. £5.80 Vol. 36 No.2 Miscellanea: Dinantian terebratulids—New microfossils—Neseuretus—Archaeocidaris whatleyensis— Carboniferous dasyclad—Nanjinoporella—Toarcian bryozoans—Drybrook Sandstone plants—British fossil — bintoniellids—Uraloporella. 1982. Pp. 43-155. £19.80 Vol. 36 No.3 The Ordovician Graptolites of Spitsbergen. R. A. Cooper & R. A. Fortey. 1982. Pp. 157-302, 6 plates, 83 figs, 2 tables. £20.50 Vol. 36 No. 4 Campanian and Maastrichtian sphenodiscid ammonites from southern Nigeria. P. M. P. Zaborski. 1982. Pp. 303-332, 36 figs. £4.00 Vol. 37 No. 1 Taxonomy of the arthrodire Phlyctaenius from the Lower or Middle Devonian of Campbellton, New Brunswick, Canada. V. T. Young. 1983. Pp. 1-35, 18 figs. £5.00 Vol. 37 No. 2 Ailsacrinus gen. nov., an aberrant millericrinid from the Middle Jurassic of Britain. P. D. Taylor. 1983. Pp. 37-77, 48 figs, 1 table. £5.90 Vol. 37 No. 3. Miscellanea: Permian Glossopteris in Turkey—Wealden Theriosuchus—Wealden conifer—Permian plants of Saudi Arabia—Carboniferous Edrioasteroidea—British cicadas—Dittonian cephalaspids. 1983. Pp. 79-171. £13.50 Vol. 37 No. 4 The relationships of the palaeoniscid fishes, a review based on new specimens of Mimia and Moy- — thomasia from the Upper Devonian of Western Australia. B. G. Gardiner. 1984. Pp. 173-428, 145 figs, 4 plates. 0 565 00967 2. £39.00 Vol. 38 No. 1 New tertiary pycnodonts from the Tilemsi valley, Republic of Mali. A. E. Longbottom. 1984. Pp. 1-26, 29 figs, 3 tables. 0 565 07000 2. £3.90 Vol. 38 No. 2 Silicified brachiopods from the Viséan of County Fermanagh, Ireland. (III) Rhynchonellids, Spiriferids and Terebratulids. C. H. C. Brunton. 1984. Pp. 27—130, 213 figs. 0 565 07001 0. £16.20 Vol. 38 No.3. The Llandovery Series of the Type Area. L. R. M. Cocks, N. H. Woodcock, R. B. Rickards, J. T. Temple & P. D. Lane. 1984. Pp. 131-182, 70 figs. 0 565 07004 5. £7.80 | Vol. 38 No. 4 Lower Ordovician Brachiopoda from the Tourmakeady Limestone, Co. Mayo, Ireland. A. Williams & — G. B. Curry. 1985. Pp. 183-269, 214 figs. 0 565 07003 7. £14.50 Vol. 38 No.5 Miscellanea: Productacean growth and shell shape—Jurassic alga Palaeosiphonium—Upper Ordovi- cian brachiopods and trilobites—Lower Devonian Osteostraci from Podolia—Hipparion from Diavata—Preparation and study of Singa skull—Carboniferous and Permian bryozoa—Lower Eocene trionychid—Montsech fossil insects. 1985. Pp. 271-412. 0 565 07004 5. £24.00 © Vol. 39 No. 1 Upper Cretaceous ammonites from the Calabar region, south-east Nigeria. P. M. P. Zaborski. 1985. _ Pp. 1-72, 66 figs. 0 565 07006 1. £11.00 Vol. 39 No. 2 Cenomanian and Turonian ammonites from the Novo Redondo area, Angola. M. K. Howarth. 1985. Pp. 73-105. 33 figs. 0 565 07006 1. £5.60 Vol. 39 No.3 The systematics and palaeogeography of the Lower Jurassic insects of Dorset, England. P. E. S. Whalley. 1985. Pp. 107—189, 87 figs, 2 tables. 0 565 07008 8. £14.00 Vol. 39 No.4 Mammals from the Bartonian (middle/late Eocene) of the Hampshire Basin, southern England. J. J. Hooker. 1986. Pp. 191-478, 71 figs, 39 tables. 0 565 07009 6. £49.50 Vol. 40 No. 1 The Ordovician graptolites of the Shelve District, Shropshire. I. Strachan. 1986. Pp. 1-58, 38 figs. 0 565 07010 X. £9.00 — Vol. 40 No. 2. The Cretaceous echinoid Boletechinus, with notes on the phylogeny of the Glyphocyphidae and Tem- nopleuridae. D. N. Lewis. 1986. Pp. 59-90, 11 figs, 7 tables. 0 565 07011 8. £5.60 Vol. 40 No. 3 The trilobite fauna of the Raheen Formation (upper Caradoc), Co. Waterford, Ireland. A. W. Owen, - R.P. Tripp & S. F. Morris. 1986. Pp. 91-122, 88 figs. 0 565 07012 6. £5.60 Vol. 40 No.4 Miscellanea I: Lower Turonian cirripede—Indian coleoid Naefia—Cretaceous—Recent Crantidae— Lectotypes of Girvan trilobites—Brachiopods from Provence—Lower Cretaceous cheilostomes. 1986. Pp. 125-222. | 0 565 07013 4. £19.00 Vol. 40 No. 5 Miscellanea II: New material of Kimmerosaurus—Edgehills Sandstone plants—Lithogeochemistry of © Mendip rocks—Specimens previously recorded as teuthids—Carboniferous lycopsid Anabathra—Meyenodendron, new Alaskan lepidodendrid. 1986. Pp. 225-297. 0 565 07014 2. £13.00 Vol. 41 No. 1 The Downtonian ostracoderm Sclerodus Agassiz (Osteostraci: Tremataspididae). P. L. Forey. 1987. Pp. 1-30. 11 figs. 0 565 07015 0. £5.50 Vol. 41 No. 2. Lower Turonian (Cretaceous) ammonites from south-east Nigeria. P. M. P. Zaborski. 1987. Pp. 31-66. 46 figs. 0 565 07016 9. £6.50 | Vol. 41 No.3 The Arenig Series in South Wales: Stratigraphy and Palaeontology. I. The Arenig Series in South Wales. R. A. Fortey & R. M. Owens. II. Appendix. Acritarchs and Chitinozoa from the Arenig Series of South-west | ! Wales. S. G. Molyneux. 1987. Pp. 67-364. 289 figs. 0 565 07017 7. In press... Vol. 41 No.4 Miocene geology and palaeontology of Ad Dabtiyah, Saudi Arabia. Compiled by P. J. Whybrow. 1987. 0 565 07019 3. In press. | Vol. 42 No. 1 Cenomanian and Lower Turonian echinoderms from Wilmington, south-east Devon. A. B. Smith, C. R. C. Paul, A. S. Gale & S. K. Donovan. 1987. 0 565 07018 S. In | = ea Bulletin of the British Museum (Natural History) Miocene geology and palaeontology of Ad Dabttyah, Saudi Arabia Compiled by P. J. Whybrow ee lO Uh eee —— oe , m Geology Vol41 No4 29 October 1987 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum’s resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History) Cromwell Road, London SW7 5BD, England. W orld List abbreviation: Bull. Br. Mus. nat. Hist. (Geol.) © British Museum (Natural History), 1987 The Geology Series is edited in the Museum’s Department of Facer, Keeper of Palaeontology: Dr L. R. M. Cocks ae aa Editor of the Bulletin: Dr M. K. Howarth Assistant Editor: Mr D. L. F. Sealy ISBN 0 565 07019 3 ISSN 0007-1471 Geology series British Museum (Natural History) Vol 41 No 4 pp 365-457 Cromwell Road London SW7 5BD Issued 29 October 1987 Miocene geology and palaeontology of Ad D en Saudi Arabia H NAIU CCEAITE! Compiled by P. J. Whybrow |. i? NA Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 S Bb, PALAE( Contents Summary, with Table of vertebrate fauna. By P. J. Whybrow ................... Miocene stratigraphy, geology and flora (Algae) of eastern Saudi Arabia and the Ad Dabtiyah vertebrate locality. By P. J. Whybrow, H. A. McClure & Gra LEST) rpm reese Tare or ares eens cic mraials «ioe Dae eek « Actclofe cronies aicGiale wae elas vad The phyletic position of the Ad Dabtiyah hominoid. By P. J. Andrews & Le INIENTUIN caudecadss aeaode Gepr aAeOn Gch aT Ete Mera eric acre a ttn eerie ie ae eer Mastodons from the Miocene of Saudi Arabia. By A. W. Gentry ............... Rhinoceroses from the Miocene of Saudi Arabia. By A. W. Gentry ............. Ruminants from the Miocene of Saudi Arabia. By A. W. Gentry ................ Miocene Suidae from Ad Dabtiyah, eastern Saudi Arabia. By M. Pickford .... A delphinoid ear bone from the Dam Formation (Miocene) of Saudi Arabia. yale Caw Ditin Ofer eee veaen as ceils te oe ieee ener eve Foigdieed Vieseemad wneisee Early Miocene fish from eastern Saudi Arabia. By P. H. Greenwood ........... HINGES. ododaoooosddoccs stauco0 ccCod Gu qHGRORMEG ar SrcGn end CONROE ae ae aaeerCr erie ree cy > 4 JULUGI ly ‘T[TY 24} JO JOO] oy} sossed ‘WyBII 0} Yo] ‘“Iopiog 1e}eGD-uLIqeIy IpNeg oy} :1e1VH WoO G16] Ul poydesFoj0Yq ‘s[issoy Usd0I punoy ysy AQIIYd “f IS “H E€6T UE Q10Y4M Apeoo] YL TAEM NQY We | “BY Tax aoe © Ss = < abe = Summary P. J. Whybrow Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 5BD The importance of the eastern Saudi Arabian Miocene mammal-bearing localities lies in their intermediate palaeogeographical position between the eastern African Miocene sites and those of Turkey and Pakistan, and their proximity to the shores of the contracting Tethys epiconti- nental sea. The eastern Arabian deposits consist of three units. These are, from oldest to youngest, the continental Hadrukh Formation, which unconformably overlies Eocene rocks, the marine Dam Formation and the continental Hofuf Formation. The Hadrukh can always be distinguished from the Hofuf when the intervening marine Dam is present, but where continental equivalents of the Dam occur, such as at Ad Dabtiyah, they are difficult to separate from the underlying Hadrukh. In the Mesopotamian region of the Middle East the ancient marine connection between the Mediterranean and the Indian Ocean—the Tethys—had been lost at the time of deposition of the Dam Formation. The break in the marine sequence, indicated in part by the continental Hadrukh and its equivalent chronostratigraphic units in the region, suggests the presence of a land connection between Saudi Arabia and southwestern Asia by at the latest mid-Burdigalian times, at about 18 Ma (Adams, Gentry & Whybrow 1983; Whybrow 1984). Keller & Barron (1983) report worldwide low sea level between 20 and 18 Ma, while Barry, Johnson, Raza & Jacobs (1985) believe an Africa to southern Asia land connection must have existed before 18 Ma and possibly even before 20 Ma as suggested by Whybrow et al. (1982). The Ad Dabtiyah fauna, described here, represents part of a fauna of cosmopolitan distribu- tion in Africa, Europe and Asia during a time equivalent to the mid-Orleanian of Europe, 17-19 Ma. The Asian mastodon Gomphotherium cooperi, previously known only from the basal Miocene deposits at Dera Bugti, Pakistan, is present. Of the two rhinoceros species found, both are early and primitive members of their genera; one is not unlike the European Dicerorhinus sansaniensis. Even so, the poor ruminant and suid fauna, together with the mastodon and rhinoceroses, does not on balance suggest any discrete palaeobiogeographical affinities for the Ad Dabtiyah fauna. Part of its African element is a new genus of hominoid which is interpreted as the sister group of the great ape and man clade and is more closely related to the African members of that clade than to their Eurasian representatives. The palaeoenvironment at Ad Dabtiyah appears to have been a tropical ponded area of fresh water with centropomid and cyprinid fishes. Logs, probably of palm, are found encrusted with cyanophyte algal material of fresh-water origin, and large stromatolitic bioherms occur. The ruminants and rhinoceroses suggest a woodland habitat near to the Ad Dabtiyah depositional area. From the marine Dam Formation near its type locality at Jabal Lidam, the first cetacean fossil to be reported from Saudi Arabia is described (p. 447). Acknowledgements Many people have contributed to Miocene geological and palaeontological studies in Arabia. In the State of Qatar, Dr Omar Abdel Rahman, University of Qatar Scientific and Applied Research Centre, provided invaluable assistance; Professor M. A. Bassiouni, Ain Shams University, Cairo, was always a helpful and informative colleague; Dr Darwish M. Al-Far, Director of the Qatar National Museum, was most generous; M. Abd al-Hadi al Mari provided local knowledge of southern Qatar; and the Qatar Petroleum Producing Authority (Onshore) provided logistic support during the initial studies. Bull. Br. Mus. nat. Hist. (Geol.) 41 (4): 367-369 Issued 29 October 1987 368 P. J. WHYBROW In the United Arab Emirates, His Excellency Dr Shibeeb M. Al-Marzouqi, Secretary General of the United Arab Emirates University, generously provided field work facilities; and Dr M. Y. Hassan and Dr Abdul Rahmin Hamdam of U.A.E. University, and Dr Obeid E] Hakeem, Director of the U.A.E. Natural History Museum, all gave valuable assistance. Appendix. Table of eastern Saudi Arabian Miocene vertebrates Hofuf Formation—vertebrate fauna from Al Jadidah From Sen & Thomas 1979; Thomas et al. 1978; Thomas 1983. Rodentia Sciuridae: Atlantoxerus sp. Ctenodactylidae: Metasayimys intermedius Sen & Thomas 1979 Carnivora Hyaenidae: Percrocuta sp. Proboscidea Gomphotheriidae: Gomphotherium angustidens Perissodactyla Rhinocerotidae: Dicerorhinus cf. primaevus Artiodactyla Suidae: cf. Lopholistriodon Giraffidae: Palaeotragus sp. Bovidae: Pachytragus ligabuei Thomas 1983 Caprotragoides aff. potwaricus Protragocerus sp. cf. Homoiodorcas? Indeterminate Pisces, Chelonia and Crocodilia. Dam Formation (continental equivalents)}— vertebrate fauna from Ad Dabtiyah Hominoidea Heliopithecus leakeyi Andrews & Martin 1987 (herein) Proboscidea Gomphotheriidae: Gomphotherium cooperi Perissodactyla Rhinocerotidae: Dicerorhinus sp. aff. sansaniensis Brachypotherium sp. Artiodactyla Suidae: Listriodon cf. lockharti or L. cf. akatikubas ? Kenyasus sp. Tragulidae: Dorcatherium sp. Dorcatherium, large: sp. Giraffoidea: Canthumeryx sp. Bovidae: Eotragus sp. Bovid species 2. Bovid species 3. Chelonia Crocodilia: cf. Crocodylus pigotti Osteichthyes: Cyprinidae Acanthoptergii: ? Centropomidae Vertebrates from the As Sarar (Al Sarrar) locality Provisional list from Thomas et al. 1982. Insectivora: Erinaceidae ? Primates gen. et sp. indet. Lagomorpha: Ochotonidae Rodentia Cricetidae Ctenodactylidae: Metasayimys cf. intermedius Gerbillidae Pedetidae: Megapedetes cf. pentadactylus cf. Protalactaga Thryonomyidae: Paraphiomys sp. Carnivora Viverridae: Viverra sp. Mustelidae: cf. Martes Mionictis sp. Felidae: Pseudaelurus turnauensis Amphicyonidae: Amphicyon sp. Proboscidea Deinothertidae: cf. Deinotherium Gomphotheriidae: Gomphotherium sp. ? Amebelodontinae Sirenia indet. Hyracoidea Saghatheriinae: Pachyhyrax aff. championi Perissodactyla Rhinocerotidae: Aceratherium sp. Dicerorhinus sp. Artiodactyla Suidae: Listriodon sp. gen. et sp. indet.; giant species. Tragulidae: Dorcatherium cf. libiensis Bovidae gen. et sp. indet. Aves Threskiornithidae Ciconiidae: Mycteria cinereus ? Mycteria sp. Scolopacidae: Charadriinae indet. spp. unidentified. Crocodilia Crocodylidae: Crocodylus cf. pigotti Chelonia Pelomedusidae: cf. Schweboemys SUMMARY Centropomidae: Lates sp. Sphyraenidae: Sphyraena sp. Sparidae indet. 369 aff. Stereogenys Selachii Trionychidae: aff. Cycloderma Hemigaleidae: Hemipristis serra Carettochelyidae Carcharhinidae: Testudinidae: Geochelone sp. Carcharhinus aff. priscus Serpentes Carcharhinus aff. plumbeus Scolecophidia Galeocerdo cf. aduncus Boidea: Scoliodon sp. Python sp. Negaprion eur ybathrodon Eryx/Gongylophis spp. Sphyraenidae: Sphyraena sp. Colubridae Dasyatidae: Dasyatis sp. Elapidae: Naja/Palaeonaja spp. Myliobatidae: Viperidae M yliobatis sp. Squamata Aetobatus arcuatus Sauria: Lacertidae Amphisbaenia: Amphisbaenidae Rhinopteridae: Rhinoptera Amphibia Bufonoidea Hadrukh Formation—vertebrates from Jabal Ranoidea Midra ash-Shamali Pisces Mormyridae: Hyperopisus sp. From Whybrow et al. 1982. Cyprinidae: Rodentia Barbus sp. Zapodidae: Arabosminthus quadratus Daams Labeo sp. 1982 Clariidae: Cricetidae: Shamalina tuberculata Daams 1982 Heterobranchus sp. Artiodactyla Clarias sp. Bovidae: cf. Oioceros sp. References Adams, C. G., Gentry, A. W. & Whybrow, P. J. 1983. Dating the Tethyan terminal event. Utrecht Micropaleont. Bull., 30: 273-298. Barry, J. C., Johnson, N. M., Raza, S. M. & Jacobs, L. L. 1985. Neogene mammalian faunal change in southern Asia: Correlations with climatic, tectonic, and eustatic events. Geology, Boulder, Col., 13: 637-640. Keller, G. & Barron, J. A. 1983. Paleoceanographic implications of Miocene deep-sea hiatuses. Bull. geol. Soc. Am., New York, 94: 590-613. Sen, S. & Thomas, H. 1979. Découverte de rongeurs dans le Miocene moyen de la Formation Hofuf (Province du Hasa, Arabia Saoudite). C.r. somm. Seanc Soc. geol. Fr., Paris, 1979 (1): 34-37. Thomas, H. 1983. Les Bovidae (Artiodactyla, Mammalia) du Miocéne moyen de la Formation Hofuf (Province du Hasa, Arabie Saoudite). Palaeovertebrata, Montpellier, 13 (5): 157—206. , sen, S., Khan, M., Battail, B. & Ligabue, G. 1982. The Lower Miocene fauna of Al-Sarrar (Eastern Province, Saudi Arabia). ATLAL, JI Saudi Arab. Archaeol., Jeddah, 5: 109-136, pls 110-116. , Taquet, P., Ligabue, G. & Del’Agnola, C. 1978. Découverte d’un gisement de vertébrés dans les depots continentaux du Miocéne moyen du Hasa (Arabia saoudite). C.r. somm. Séanc. Soc. géol. Fr., Paris, 1978 (2): 69-72. Whybrow, P. J. 1984. Geological and faunal evidence from Arabia for mammal ‘migrations’ between Asia and Africa during the early Miocene. Cour. ForschInst. Senckenberg., Frankfurt a.M., 69: 189-198. —, Collinson, M. E., Daams, R., Gentry, A. W. & McClure, H. A. 1982. Geology, fauna (Bovidae, Rodentia) and flora from the early Miocene of eastern Saudi Arabia. Tertiary Res., Leiden, 4: 105-120. se ’ Miocene stratigraphy, geology and flora (Algae) of eastern Saudi Arabia and the Ad Dabtiyah vertebrate locality P. J. Whybrow Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 SBD H. A. McClure Arabian American Oil Company, Dhahran, Saudi Arabia G. F. Elliott Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 5BD Synopsis Studies of the Miocene deposits in eastern Saudi Arabia are briefly reviewed. The stratigraphical suc- cession is explained and the geological details of the vertebrate-bearing, non-marine deposits at Ad Dabtiyah presented. The stratigraphical position of the fossiliferous beds is believed to lie near the boundary of the continental sequence of the Hadrukh Formation and continental equivalents of the Dam Formation, the beds themselves locally close to, and the lateral equivalent of, the basal deposits of the Burdigalian marine carbonates of the Dam Formation. The fresh-water depositional environment at Ad Dabtiyah contains many bones of terrestrial vertebrates, found in close association with several large stromatolitic bioherms. These, and similar encrustations also of a fresh-water origin, are associated with in situ logs, probably of palm trees. Overall, both the stratigraphical position of the Ad Dabtiyah deposits and their contained fauna suggest an age of about 17-19 Ma; middle Orleanian (European land-mammal age equivalent); early Burdigalian (marine chronology). Introduction Until the 1930s, almost nothing was known of the geology of the central part of the Arabian Peninsula. Observations on its general geology had been carried out by Philby (1933) during his explorations of Saudi Arabia, and his fossil collections were presented to the British Museum (Natural History) by King Abdul Aziz ibn-Saud. Cox (1933) studied the Tertiary fossils Philby had found at a hill called Qarn Abu Wayil (Fig. 2) and identified the oyster Ostrea latimarginata, and natural casts of Mytilus, Anomia, Cardium, Clementia, Anadara and other molluscs previously known from the Lower Fars rocks of Iran. South of Qarn Abu Wayil at Jaub Anbak (Fig. 2), Philby had noted the marine beds were overlain by a considerable thickness of red sandstones which Cox suggested might also be of Miocene age, but equally might be equivalent to the Pliocene Bakhtiyari deposits of Iran (Cox, in Philby 1933: 386-387). Cox’s work therefore provided the first evidence of the presence of Miocene rocks in Arabia. Mapping and surface collecting by geologists of the Arabian American Oil Company, ARAMCO, started in the mid-1930s and revealed in detail the extent of Miocene deposits in eastern Saudi Arabia. Three formations were formally designated (Steineke et al. 1958), together with the lithological details at their type localities. Later, the first published evidence that vertebrate fossils occurred in the region was given by Powers et al. (1966: D97) who recorded ‘vertebrate fragments’ in their lists of the Miocene biota. In 1974 a collection of Miocene crocodile, turtle, antelope, rhinoceros and proboscidean remains was presented to the British Museum (Natural History), representing the first known Bull. Br. Mus. nat. Hist. (Geol.) 41 (4): 371-382 Issued 29 October 1987 3 P. J. WHYBROW, H. A. MCCLURE & G. F. ELLIOTT ARABIAN Ad Dabtiyah Y ¥ e J. Midra ash-Shamali ® e S BAHRAIN \ oe P ‘X ¢q NE Uray’irah e x Al Jadidah SAUDI ARABIA EG “oe Anbak & ed xX { UAE Fig. 2 Eastern Saudi Arabia, the State of Qatar and part of the United Arab Emirates showing localities referred to in the text. The most western extent of the Miocene deposits in Saudi Arabia and their location in south-western Qatar is indicated. vertebrate palaeofauna from Arabia (BM(NH) 1975: 18). In the same year staff of the Palaeon- tology Department added more material by collecting from Miocene continental deposits at Ad Dabtiyah (Fig. 2). They also collected vertebrate-bearing rocks near a hill called Jabal Midra ash-Shamali, c. 6km north-west of Dhahran (Fig. 2) where Tleel (1973) had discovered ‘artio- dactyl remains. Chemical breakdown of these rocks yielded rodent, lagomorph and bovid teeth and preliminary reports on the faunas from both localities were published (Hamilton et al. 1978; Andrews et al. 1978). Further work on the Ash-Shamali material showed a new genus of fruits of aquatic plants, Midravalva arabica (Collinson 1982), to be present, together with new rodents, Arabosminthus quadratus and Shamalina tuberculata (Daams, in Whybrow et al. 1982: 111-116). GEOLOGY & FLORA 373 In 1978, H. Thomas of the Museum National d’Histoire Naturelle, Paris, and colleagues excavated vertebrate fossils, Percrocuta, Gomphotherium angustidens, Dicerorhinus, Lopholistrio- don, Pachytragus, Protragocerus, Caprotragoides and a new rodent, Metasayimys intermedius, from red-coloured sandstones at Al Jadidah (Fig. 2); see Thomas et al. (1978), Sen & Thomas (1979), Thomas (1983). Later, they collected from the As Sarar region (Fig. 2) in collaboration with the Saudi Arabian Department of Antiquities. Abundant vertebrate remains were found including two gomphothere species, a deinothere, several carnivores and cricetid, ctenodactylid, dipodid, gerbellid, pedetid and phiomorph rodents (Thomas et al. 1982). These discoveries of terrestrial vertebrates in eastern Arabia have now bridged the palaeo- geographical gap between the better-known Miocene faunas of Africa and those of south- western Asia. In addition, interpretations of the eastern Arabian palaeoenvironments have been made (Whybrow & McClure 1981; Thomas et al. 1982; Whybrow et al. 1982) and, because of the proximity of the fossil localities to the contracting Tethys epicontinental seaway, there have been suggestions concerning the location of a Neogene land connection between Arabia and south-western Asia (Adams et al. 1983; Rogl & Steininger 1983; Whybrow 1984; Thomas 1985). The collected papers in this issue of the Bulletin, with the exceptions of the descriptions of a delphinoid ear bone and a gomphothere tooth from other localities, are the results of studies on the Ad Dabtiyah fauna and flora collected and donated in 1974. Stratigraphy and the age of the deposits Towards the end of the middle Eocene, widespread emergence of the eastern Arabian shelf coincided with continued uplift and a slight north-easterly tilting of the Arabian plate, events that began in the late Cretaceous and continue today as a consequence of the movement of the Arabian plate against the more stable south-western Asian plate. Red Sea rifting was also a contemporaneous consequence of this plate activity (Schmidt et al. 1983; Sellwood & Nether- wood 1984). Since that time mainly continental deposition, with the exception of a marine transgression from the Indian Ocean, has prevailed in eastern Arabia. Rocks of Oligocene age have not been recognized in the region; Miocene deposits unconformably overlie rocks of Ypresian or Lutetian age. Where the rocks formed by the marine transgression (represented by the Dam Formation) occur, the continental Neogene has been divided into units. From the oldest, these are the Hadrukh Formation (c. 20-120m thick), succeeded by the Dam Formation itself (c. 30-100 m thick), and the Hofuf Formation (c. 30-100 m thick). Towards the western interior, where the marine marker beds of the Dam intercalate with the continental deposits, become thin and eventually disappear, the eastern divisions of Hadrukh, Dam and Hofuf no longer apply. The undifferentiated deposits are treated as a single un-named unit—Tertiary continental sandstone, marl and limestone, marked “Tsm’ in Figs 3 and 5 (Steineke et al. 1958). In such deposits a gomphothere M* + M? (M.42946) was found in the 1930s; see Gentry, p. 401 in this issue. The age of the Hadrukh is important and is currently controversial (see Whybrow 1984) as it contains a new cricetid rodent, Shamalina tuberculata, whose descendant relatives appear to be present in the Miocene Lower Siwaliks of Pakistan (Daams, in Whybrow et al. 1982; E. H. Lindsay, personal communication 1985). The Hadrukh is undoubtedly coeval with the Ghar Formation of Kuwait and southern Iraq and, in the neighbouring part of Iran, 215 m (700 ft) of sandstone (subsurface section from a drill hole) is said to be the eastern wedge-end of the Ghar Formation and called the Ahwaz Sandstone (James & Wynd 1965: 2229). Adams et al. (1983: 278) indicated that at least part of the Ahwaz Sandstone is of Late Oligocene (Chattian) age, but Murris (1980: 614) suggested an Early Miocene, Aquitanian, age for these deposits, which he called the Ahwaz delta formed from eroded Saudi Arabian pre-Neogene rocks. Thin beds near to the top of the Hadrukh contain poorly-preserved marine molluscs and the oyster O. latimarginata which indicates a Burdigalian (marine chronostratigraphy) age. These beds crop out in a small area near the modern coastline; their lithology has not been described and 374 P. J. WHYBROW, H. A. MCCLURE & G. F. ELLIOTT “R b ee » i s ~” ats a . : < s 7. ee , e S ewe i : ss Ps > : es at ip ee. ad oe 3 P- ri) * Py ae 4 7 » * 3 oe J 1 2 ‘ 3 mS hire p “ ; , - Y "i oo 7m, 4 ‘é 4 >» Fig. 6 Similar view to Fig. 4 (photo taken 1979) showing the in situ ‘logs’. The drainage channel, shown in Fig. 4, indicates little erosion in over 40 years. Tape = 1 metre. P. J. WHYBROW, H. A. MCCLURE & G. F. ELLIOTT 378 "YIPIM UI WO €7Z POINSBOUL ¢ “A}I[BIO] 9}¥.1G9]19A JY} JO JSB9-YIIOU JY} O} PUNO] SUIIOYOIG INTTOVUIONS OBIeT LL “BI GEOLOGY & FLORA 379 Fig. 8 Photomicrograph of a thin section of calcareous crust (V.60434c) found covering the skull of a rhinoceros. Banded algal growth of radial cyanophyte filaments are at top right. x 45. The vertebrate site lies at or near the contact of the Hadrukh Formation with other contin- ental sediments coeval with and laterally equivalent to the basal parts of the marine Dam sequence. The basal Dam marker horizon—the “Button bed’ with Fibularia damensis—is absent at the locality, but is found about 10km to its east. The south-western limit of the Dam sea in Fig. 9 Algal crust showing conspicuous charophyte oogonia, left and right. Same sample as Fig. 8. x 45. 380 P. J. WHYBROW, H. A. MCCLURE & G. F. ELLIOTT Fig. 10 Algal crust with probable traces of chironomid larval tubes, centre. Same sample as Fig. 8. x 45. the area appears to have been near Kashm Khizami (Fig. 3), where marine fossils are associated with beach boulder conglomerates. A schematic cross section of the area indicating the relationship of the stratigraphy to the vertebrate site is shown in Fig. 5. The main excavation was carried out in hard, unbedded White N9 (United States Geological Survey Rock-Color Chart 1980) sandstones. These were well sorted with fine-grained (about 280 um), angular to subrounded, micrite supported quartz clasts. Occasionally, rounded micri- tic pebbles were present. Voids in the sandstones were sometimes filled with sparry calcite or, rarely, a form of manganese oxide known as wad. The excavation (about 22m x 17m, 50- 80cm in depth) produced scores of isolated bones, mainly lower jaws, teeth, limb bones, pectoral and pelvic elements, mostly of rhinoceros. None was preferentially orientated. Verte- brae and ribs were rare and, although none of the bones showed sign of depositional transport, rhinoceros mandibles had been broken and their anterior parts were missing. Except for the dryopithecine maxilla (see Andrews & Martin, this issue, p. 383), no other cranial bones were excavated. About 5m stratigraphically higher than the main excavation, and 110m to its west, the sediments exposed on a ridge trending north-east showed a change in lithology. They consisted of an unsupported conglomerate formed of pebbles and cobbles of micritic limestone. In these sediments a proboscidean scapula and incomplete but uncrushed fish skulls were found (see Greenwood, this issue, p. 451). On top of this ridge were five in situ fallen logs encrusted with stromatolite (Fig. 6). Three measured 3:0m, 5:-3m and 7:2m in length and all, including the encrustation, were about 1:5m in diameter. At one end of each of these logs stromatolitic crusts, 2m in diameter, suggested the position of the bole of the fallen tree. The microstructure of a large amount of silicified wood found in this area resembled palm wood. At the same level and 11m east of the logs, a bioherm had been fractured to reveal a fragmented rhinoceros skull encrusted with a 3-cm layer of algal material. Surface collecting in the area of the conglomerate produced proboscidean, giraffoid and tragulid remains. The conglomerate facies continued to the east and on the northern flank of a parallel ridge, many large stromatolitic bioherms were present (Fig. 7). GEOLOGY & FLORA 381 The bioherms and the crusts are largely of cyanophyte (myxophyte) algal origin. In thin section the rock shows marked banding with differential growth, and in places a ragged radial structure survives from the original microscopic thread-algae (Fig. 8). The rock shows intrinsic evidence of freshwater origin with embedded charophyte oogonia and debris, and what are probably chironomid larval tubes. All are poorly preserved and filled with sparry calcite (Figs 9, 10). The depositional environment Immediately prior to or at about the time of the deposition of basal transgressive marine sediments of the Dam Formation coming from the east, the environment at Ad Dabtiyah appears to have been a fluvial regime, transporting sandy carbonate muds, with laterally discontinuous conglomerates suggestive of channel sediments. Remains of terrestrial mammals, freshwater fish, turtle and crocodile (cf. Crocodylus pigotti, see Buffetaut, 1984) occur in these sediments. Subsequently, clastic deposition ceased and ponded fresh water was present; this was perhaps as a lake high in dissolved carbonates and deep enough to allow continuous growth of large stromatolitic bioherms and thick stromatolite crusts on hard substrates. References Adams, C. G., Gentry, A. W. & Whybrow, P. J. 1983. Dating the Tethyan terminal event. Utrecht Micropaleont. Bull., 30: 273-298. Andrews, P. J., Hamilton, W. R. & Whybrow, P. J. 1978. Dryopithecines from the Miocene of Saudi Arabia. Nature, Lond., 274: 249-251. British Museum (Natural History) 1975. Report on the British Museum (Natural History) 1972-1974. vii + 173 pp. London; Trustees of the B.M. (N.H.). Buffetaut, E. 1984. On the occurrence of Crocodylus pigotti in the Miocene of Saudi Arabia, with remarks on the origin of the Nile crocodile. Neues Jb. Geol. Palaont. Mh., Stuttgart, 9: 513-520. Collinson, M. E. 1982. A reassessment of fossil Potamogetoneae fruits with description of new material from Saudi Arabia. Tertiary Res., Leiden, 4: 83-104. Cox, L. R. 1933. Stratigraphy and palaeontology. In Philby 1933: 383-387 (q.v.). Hamilton, W. R., Whybrow, P. J. & McClure, H. A. 1978. Fauna of fossil mammals from the Miocene of Saudi Arabia. Nature, Lond., 274: 248-249. James, G. A. & Wynd, J. G. 1965. Stratigraphic nomenclature of Iranian Oil Consortium agreement area. Bull. Am. Ass. Petrol. Geol., Tulsa, 49 (12): 2182-2245. Kier, P. 1972. Tertiary and Mesozoic echinoids of Saudi Arabia. Smithson. Contr. Paleobiol., Washington, 18: 1-242. Murris, R. J. 1980. Middle East: Stratigraphic evolution and oil habitat. Bull. Am. Ass. Petrol. Geol., Tulsa, 64 (5): 597-618. Philby, H. St J. 1933. The Empty Quarter. 433 pp. London. Powers, R. W., Ramirez, L. F., Redmond, C. D. & Elberg, E. L., jr 1966. Sedimentary geology of Saudi Arabia. Prof. Pap. U.S. geol. Surv., Washington, 560 (D): 1-127. Rogl, F. von & Steininger, F. F. 1983. Vom Zerfall der Tethys zu Mediterran und Parathethys. Annin naturh. Mus. Wien, 85 (A): 135-163. Schmidt, D. L., Hadley, D. G. & Brown, G. F. 1983. Middle Tertiary continental rift and evolution of the Red Sea in southwestern Saudi Arabia. Open File Rep. U.S. geol. Surv., Denver, 83-0641. 60 pp., map. Sellwood, B. W. & Netherwood, R. E. 1984. Facies evolution in the Gulf of Suez area: sedimentation history as an indicator of rift initiation and development. Mod. Geol., New York, 9: 43-69. Sen, S. & Thomas, H. 1979. Découverte de rongeurs dans le Miocéne moyen de la Formation Hofuf (Province du Hasa, Arabia Saoudite). C. r. somm. Seéanc. Soc. géol. Fr., Paris, 1979 (1): 34-37. Shipman, P., Walker, A., Van Couvering, J. A., Hooker, P. J. & Miller, J. A. 1981. The Fort Ternan hominoid site, Kenya: geology, age, taphonomy and palaeoecology. J. hum. Evol., London, 10: 49-72. Steineke, M., Harriss, T. F., Parsons, K. R. & Berg, E. L. 1958. Geological map of the western Persian Gulf quadrangle, Kingdom of Saudi Arabia. Miscellaneous geological investigations Map I-208A. Reston, Va.; US. geol. Survey. Thomas, H. 1983. Les Bovidae (Artiodactyla, Mammalia) du Miocéne moyen de la Formation Hofuf (Province du Hasa, Arabie Saoudite). Palaeovertebrata, Montpellier, 13 (5): 157—206. 382 P. J. WHYBROW, H. A. MCCLURE & G. F. ELLIOTT 1985. The Early and Middle Miocene Land Connection of the Afro-Arabian Plate and Asia: A major event for Hominoid Dispersal? In: Delson, E. (ed.), Ancestors: The Hard Evidence: 42-50. New York. , Sen, S., Khan, M., Battail, B. & Ligabue, G. 1982. The Lower Miocene fauna of Al-Sarrar (Eastern Province, Saudi Arabia). ATLAL, JI Saudi Arab. Archaeol., Jeddah, 5: 109-136, pls 110-116. , Taquet, P., Ligabue, G. & Del’Agnola, C. 1978. Découverte d’un gisement de vértebrés dans les dep6ts continentaux du Miocene moyen du Hasa (Arabia saoudite). C. r. somm. Seanc. Soc. geol. Fr., Paris, 1978 (2): 69-72. Tleel, J. W. 1973. Surface geology of Dammam Dome, Eastern Province, Saudi Arabia. Bull. Am. Ass. Petrol. Geol., Tulsa, 57 (3): 558-576. Whybrow, P. J. 1984. Geological and faunal evidence from Arabia for mammal ‘migrations’ between Asia and Africa during the early Miocene. Cour. ForschInst. Senckenberg., Frankfurt a.M., 69: 189-198. & Bassiouni, M. A. 1986. The Arabian Miocene: rocks, fossils, primates and problems. Proc. int. Congr. Primatol. (10th: 1984: Nairobi), Cambridge, 1: 85-89. , Collinson, M. E., Daams, R., Gentry, A. W. & McClure, H. A. 1982. Geology, fauna (Bovidae, Rodentia) and flora from the early Miocene of eastern Saudi Arabia. Tertiary Res., Leiden, 4: 105-120. & McClure, H. A. 1981. Fossil mangrove roots and palaeoenvironments of the Miocene of the eastern Arabian peninsula. Palaeogeogr. Palaeoclimat. Palaeoecol., Amsterdam, 32: 213-225. The phyletic position of the Ad Dabtiyah hominoid P. J. Andrews Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 5BD L. Martin Department of Anthropology, State University of New York, Stony Brook, New York 11794, U.S.A. Synopsis The hominoid maxilla and four isolated teeth from Ad Dabtiyah, Saudi Arabia, are assigned here to a new genus and species Heliopithecus leakeyi. It shares numerous primitive characters with Proconsul (for example, molar cingula and premolar cusp heteromorphy), and a few advanced characters with Kenya- pithecus (for example premolar enlargement and molar enamel thickening). The latter characters are also characteristic of the great ape and human clade, and for this reason it is grouped with that clade, but it is a more primitive member than Kenyapithecus because it retains more primitive characters. Molar enamel is intermediate in thickness and all pattern 3, the first hominoid so far described to have this combination, and this places it intermediate between gibbons and Proconsul which have thin pattern 3 enamel, and Kenyapithecus and the ancestral great ape and human morphotype which have thick pattern 3 enamel. Introduction We describe here a new genus and species of hominoid primate from continental equivalents of the basal deposits of the marine Dam Formation near Ad Dabtiyah, Saudi Arabia: see Whybrow et al. (this issue, p. 371). The hominoid specimens from these early middle Miocene deposits were first described by Andrews et al. (1978), and the associated fauna and geology were described by Hamilton et al. (1978). The hominoids were not named, but were considered to be intermediate in morphology between the early Miocene species of Proconsul from east Africa and later Miocene species of Ramapithecus and Sivapithecus from Eurasia. Comparison was made with the type specimen of what was originally called Sivapithecus africanus (Le Gros Clark & Leakey 1950) and subsequently Kenyapithecus africanus (Leakey 1967), but the tax- onomic position and provenance of this species was too uncertain itself for this comparison to be particularly helpful. Much progress has been made recently which has improved our ability to determine the phylogenetic status of the hominoid from Ad Dabtiyah. Both the hominoid clade and its constituent clades have now been better defined (Harrison 1982, Andrews 1985, Martin 1986). In particular, work on the structure and thickness of molar enamel has clarified its significance in hominoid evolution (Martin 1983, 1985), and comparisons with Spanish and Hungarian material, all assigned to Dryopithecus (Martin & Andrews 1982), have expanded our knowledge of this middle Miocene genus. Some new material is also available for Kenyapithecus (Pickford 1982, Ishida et al. 1984). The Ad Dabtiyah material is considered to resemble both Dryopithecus and Kenyapithecus in derived characters and to be linked with them in the great ape and human clade. Proconsul, by contrast, cannot be shown to share any derived characters with this clade although it does appear to have some hominoid synapomorphies (Andrews 1985, Fleagle 1986). It is now clear that the Ad Dabtiyah hominoid shares only primitive characters with Proconsul and is not therefore uniquely related to it; it would also appear that the characters it shares with Kenyapithecus and Dryopithecus are synapomorphies of the great ape and human clade and are also not indicative of special relationship. On the contrary, both Kenyapithecus and Dryopithecus share derived characters with the great apes and humans not present in the Bull. Br. Mus. nat. Hist. (Geol.) 41 (4): 383-393 Issued 29 October 1987 384 P. J. ANDREWS & L. MARTIN Ad Dabtiyah specimens. For these reasons, we have decided to name a new genus and species for this material while recognizing its taxonomic relationship with Kenyapithecus and the great apes and man. Systematics Superfamily HOMINIODEA Simpson 1931 Genus HELIOPITHECUS gen. nov. D1aGnosis. A genus of hominoid with the enlarged premolars characteristic of the great ape and human clade; the P* is elongated with a massive buccal cusp and with great buccal flare, and the P* is also elongated but without the buccal flare; the premolars are large relative to M’, both in length and breadth, and are comparable to Kenyapithecus in this respect; they differ from this genus in the greater cusp heteromorphy of the premolars and the greater cingulum development on the upper molars; the tooth enamel is all pattern 3 and is interme- diate in thickness, that is thicker than in Proconsul but thinner than in Kenyapithecus; the teeth wear with the dentine separation pattern. Name. Greek, Helios, the sun, and pithekos, an ape. TYPE SPECIES. Heliopithecus leakeyi sp. nov. Heliopithecus leakeyi sp. nov. Hototype. M.35145, a slightly crushed maxilla from the left side with the crowns of P* to M? and the lingual alveolar margins of I? and C. The specimen is housed in the Department of Palaeontology, British Museum (Natural History). Type LOCALITY. Ad Dabtiyah, Saudi Arabia: 4km south-east of the salt flat named Ad Dabtiyah, 26° 27’ 02” N, 48° 35’ 24” E. PARATYPE. M.35146, isolated upper third molar. REFERRED MATERIAL. Three isolated teeth, M.35147-9. DiAGnosis. As for genus. Name. In honour of Louis Leakey, who did so much to add to our knowledge of hominoid evolution. DESCRIPTION. The descriptions of this fossil hominoid can be added to in three ways from the previous descriptions (Andrews et al. 1978): variability within the sample; the significance of premolar enlargement; and changes in enamel thickness. Measurements, see Table 1. Table 1 Measurements of the teeth of dHeliopithecus leakeyi. md = mesiodistal length, bl = buccolingual breadth; all measurements in millimetres. Crown height Specimen md bl bl/md Crown module buccal lingual P? M.35145 ed 11-6 150-1 OT 8-4 4-0 P* M.35145 7-0 11-4 162-8 92. 5:8 5-4 M.35149 5:3 937) 183-0 WS SP7/ 4-2 M! M.35145 8-8 10-5 119-3 9-7 = = M? M.35145 9:5 11-9 125-3 10-7 = = M? M.35146 10-4 12:9 124-0 11:7 = = dP* M.35147 6:9 8:3 120-3 7-6 — = dC M.35148 5-9 49 83-1 = 5-0 = AD DABTIYAH HOMINOID 385 Fig. 11 Occlusal view of the type specimen of Heliopithecus leakeyi (M.35145). Below right are three of the isolated teeth, from left to right, M.35147, right dP*; M.35148, right dC; and M.35149, right P*. Sample variability It was originally suggested (Andrews et al. 1978) that the isolated P* (M.35149) might belong to a separate species from the maxilla M.35145. This is no longer considered likely on the basis of metrical dimensions in comparison with other closely related taxa. For instance, in the genus Kenyapithecus we would now combine K. wickeri from Fort Ternan and K. africanus from Maboko (Andrews & Walker 1976, Pickford 1982) into a single species (Greenfield 1979), and recognize this species as being distinct from the Asian genera Sivapithecus and Ramapithecus which themselves have now been grouped together (Greenfield 1980, Andrews & Cronin 1982). The metrical and morphological differences in the premolars of the wickeri and africanus specimens are very similar to those seen in the two specimens from Ad Dabtiyah: for instance the buccolingual variation of the P* is 10-5 to 12-0mm in the African specimens and 9-7 to 11-4mm in the Ad Dabtiyah specimens (Fig. 12). Similarly, the ranges in P* buccolingual dimensions for Proconsul africanus (from Rusinga only) is 8-5 to 99mm and for Proconsul nyanzae is 9-6 to 11:7mm (Andrews 1978), both similar to the range seen in the Ad Dabtiyah specimens; greater if the Rusinga africanus is combined in a single species with the Rusinga nyanzae. In comparison with these there is no good reason for not including all the Ad Dabtiyah specimens in the one species Heliopithecus leakeyi. 386 P. J. ANDREWS & L. MARTIN 15 a Gs bl P. major @ 23 12 M.16649EY oo O a®) ae My. 35145 tt Ronyanzae on. O O z 10; M. 35149 wre oe P. africanus 9 e | & e 81 Dendropithecus oe © NM, | . AN © Rangwapithecus Xk 5 le 5 6 7 gu malls SemEo Fig. 12 Size variation of the upper fourth premolar. The two specimens from Ad Dabtiyah (M.35145 & M.35149) are identified by closed stars and the type specimen of Kenyapithecus africanus (M.16649) by an open star. The sample ranges of three species of Proconsul and two species of Rangwapithecus are shown for comparison. AD DABTIYAH HOMINOID 387 Premolar enlargement The main characteristic of Heliopithecus leakeyi is the great enlargement of the premolars. The P? is larger than the P* and both are nearly as large as the first molar in cross-sectional area (Andrews et al. 1978). Fig. 13 shows this to be an important feature: the extant great apes have a relatively larger P* than do most Miocene hominoids, and Heliopithecus and Kenyapithecus are both within the great ape range and outside the range of other Miocene hominoids. For the P*, on the other hand, they differ from both living and fossil apes. They both have unusually large P*s, and in this they resemble the palate from Moroto, Uganda, which has previously been incorrectly referred to Proconsul major (Pilbeam 1969, Andrews 1978), but which would seem on this evidence to belong either to Kenyapithecus or Heliopithecus. There is evidence here, therefore, both for the relationship of these two genera based on premolar morphology, and for their relationship with the extant great apes; and it may be that the Moroto palate also belongs with this group. (See Note added in proof, p. 391.) The premolars and molars of Heliopithecus generally have low rounded cusps, the exception being the buccal cusp of P*, which is more than twice the height of the lingual cusp. In this latter feature it differs from Kenyapithecus and Dryopithecus but resembles the Moroto palate, and it would seem likely to be a primitive retention which has been lost in Kenyapithecus and P3 %& (Moroto) oe %& Kenyapithecus africanus H Heliopithecus leakeyi Sivapithecus Dryopithecus Proconsul gorilla chimpanzee orang —utan OY 3 (Moroto) M1 #& Kenyapithecus africanus M# Heliopithecus leakeyi Sivapithecus Dryopithecus Proconsul gorilla chimpanzee orang —utan 70 80 90 100 110 CROWN MODULE RATIO Fig. 13 Size ranges of the third and fourth premolars. The horizontal axis shows the premolar/ molar size ratios calculated for the crown modules (length + breadth/2). The total ranges for a number of living and fossil taxa are shown (Proconsul = 3 species; Sivapithecus = 2 species) for comparison with Heliopithecus leakeyi. 388 P. J. ANDREWS & L. MARTIN Dryopithecus. Other characters of the molars show this same combination, particularly the presence of upper molar and premolar cingula, which have been lost in later Miocene fossil hominoids and in the extant great apes. The type specimen of Kenyapithecus africanus has only a slight lingual and mesial cingulum on its P* and M’, but material described more recently by Pickford (1982) from Mayiwa includes some upper teeth with at least as great a cingulum development as on the Ad Dabtiyah specimens. Our reason for interpreting premolar cusp heteromorphy and presence of cingula as primi- tive in this instance is the widespread occurrence of these characters in earlier Miocene and Oligocene hominoids and catarrhines. Neither character is present throughout the living catarrhines, and on this basis their absence would appear more likely to be primitive for this group. In this case, however, we feel that the fossil evidence can add to the evidence of living forms and suggest the alternative interpretation. Cusp heteromorphy and cingula are ubi- quitous among early catarrhines like Propliopithecus (including Aegyptopithecus), Dendropi- thecus, Micropithecus and Limnopithecus; they are also present on early Miocene hominoids like Proconsul and Rangwapithecus, and they are now seen to be present in Heliopithecus and Kenyapithecus of the early middle Miocene. It is not until later in the middle Miocene that hominoids lacking these characters first appear, such as Dryopithecus and Sivapithecus, and we consider these hominoids to be derived in this respect. These morphological changes can be put into phylogenetic perspective as follows: the primi- tive hominoid condition is considered to include premolars that were small relative to molar size, were mesiodistally compressed and had heteromorphic cusps; upper molars and the fourth premolar had low rounded cusps and well-developed lingual and mesial cingula. Kenyapithecus and Dryopithecus differ from this condition in the enlargement of the premolars, loss of pre- molar cusp heteromorphy, and the partial reduction of the cingulum, which is not developed on most specimens. Heliopithecus is intermediate in these characters, retaining a greater degree of cusp heteromorphy and cingulum development than seen in Kenyapithecus and Dryopithecus but linked with them through premolar enlargement. Enamel structure and thickness It has been possible to examine the enamel of one of the specimens from Ad Dabtiyah. The isolated M? (M.35146) is naturally fractured, and the fractured surface has been exploited to study enamel thickness and enamel microstructure. The naturally fractured face passing through the paracone revealed a nearly ideal plane of section which minimized obliquity (Martin 1983, 1985). The slightly ragged fracture was flattened by diamond polishing to facilitate enamel thickness measurements and to produce a relief-free surface for back scattered (high energy) electron imaging. The plane in which enamel thickness was measured is shown in Fig. 14. Although this does not correspond exactly with the buccolingual plane of section through the mesial cusps recommended by Martin (1983), it is clear that it approximates to a section passing through the maximum diameter of the dentine horns and should produce results little affected by obliquity of section and therefore comparable with those from sectioned teeth. The enamel thickness was measured for a number of linear dimensions which have been used previously (Martin 1983). Linear enamel thickness over the tip of the paracone of the M? is 1:0mm, and lateral enamel thickness on the buccal cusp is 0-92mm (average 0-96mm). The breadth of the tooth across the cervix is 8-1 mm, approximately the size of a chimpanzee M?, and comparable mean dimensions for chimpanzee enamel thickness are 0-5 mm at the tip of the paracone and 0-7 mm laterally. The enamel of Heliopithecus leakeyi is thus considerably thicker in absolute terms than that of the chimpanzee. An attempt was made to scale enamel thickness by comparing enamel and dentine areas (Martin 1983, 1985). The area of enamel visible in section (as shown in Fig. 14) was measured, and this was then divided by the length of the enamel dentine junction in the same section; this approximates to the dimension c/e of Martin (1983, 1985) for the whole tooth. This dimension, which is called the Average Enamel Thickness, was then scaled for body size using the area of AD DABTIYAH HOMINOID 389 Fig. 14 Sections across the mesial face of the crown of the isolated M?* (M.35146). The buccal half of the crown was broken when discovered and has been polished and etched to expose the prism structure and the enamel—dentine junction. 390 P. J. ANDREWS & L. MARTIN dentine in the same section (Martin 1983) as the estimator of body size. This gives the scaled dimension called the Relative Enamel Thickness, which in the case of the Ad Dabtiyah M2? has a value of 17-35. This compares with values of 8-90—-11-30 for thin enamel (for instance in the chimpanzee and gorilla); 11-31-14-64 for intermediate/thin enamel; 14-65-17-49 for interme- diate/thick enamel (as seen in the orang-utan); and 17-50—26-20 for thick enamel, which is seen in Homo and Sivapithecus. The enamel of Heliopithecus falls at the top end of the range of the intermediate/thick category, and although obliquity of section may have slightly increased the apparent enamel thickness, it is quite clear that Heliopithecus leakeyi has significantly thicker enamel than in chimpanzees, gorillas or gibbons, once size has been taken into account. Although the errors inherent in this estimation are recognized, we are confident that they have been reduced to a minimum and that H. leakeyi has enamel which is of intermediate thickness (as defined by Martin, 1985). This result is significant in the light of the ancestral conditions for hominoid enamel deter- mined by Martin (1985). The ancestral hominoid is thought to have had thin enamel, with thickened enamel as a derived character of the great ape and human clade. The presence of thickened enamel in Heliopithecus therefore represents a shared derived character with the great ape and human clade. Of the living members of this clade, only the orang-utan has enamel of intermediate thickness, and this might appear to be a point of resemblance to H. leakeyi in simple thickness terms. However, the enamel in the orang-utan is intermediate in thickness as a result of secondary reduction from thick enamel, while the enamel microstructure of H. leakeyi shows no such reduction, with the enamel being formed at a fast, pattern 3, rate throughout the enamel thickness. In Pongo the outer 20% of the enamel is formed at a reduced rate, as measured from prism cross-striation repeat intervals, but this is not the case in H. leakeyi. The enamel in H. leakeyi is of intermediate thickness in relation to the time available to develop enamel, and is not due to secondary reduction. As such it could represent an early stage in the evolution of thick enamel in the common ancestor of the great ape and human clade. It is interesting to note that this is the first evidence for intermediate-thickness enamel, all of which is fast-formed pattern 3 enamel, that has been seen in any hominoid species, these conditions having previously been predicted solely on the basis of end conditions of change (Martin 1983, 1985). Phylogenetic interpretation The new pieces of evidence presented here, from examination of the enamel and the reinterpre- tation of premolar and molar morphology, are consistent in their placement of Heliopithecus leakeyi in hominoid phylogeny. Premolar enlargement with retention of what are interpreted as ancestral characters, such as retention of molar cingula and premolar cusp heteromorphy, place Heliopithecus as an intermediate between the hominoid ancestral pattern and the great ape and human ancestral pattern which is shared also by Kenyapithecus and Dryopithecus. In other words, it is more closely related to the great apes and man than are the gibbons but less closely than are Dryopithecus and Kenyapithecus. The evidence from the enamel shows the same thing: the ancestral hominoid pattern is thin pattern 3 enamel such as is present in gibbons, while the ancestral great ape and man pattern is thick pattern 3 enamel which is retained unchanged in modern and fossil man; the intermediate thickness of enamel (all pattern 3) in Heliopithecus shows that it lacks the full development of this character, and our interpretation is that it is the sister group to the great ape and man clade, with some of its characters developed but not others. These relationships are shown in Fig. 15. This shows Proconsul as the sister group to all other hominoids, living and fossil, and branching off before the divergence of the gibbons. After the gibbon divergence, first Heliopithecus and then Kenyapithecus diverged, so that both are successively sister groups to the living great apes and humans. The position of Dryopithecus with respect to Kenyapithecus is not certain. The basal split of the great ape and human clade is shown as that separating the orang-utan from the African apes and man (Andrews & Cronin 1982, Ward & Pilbeam 1983, Martin 1983), with the orang-utan joined with Sivapithecus. The remaining divergence is that between the African apes and man. AD DABTIYAH HOMINOID 391 Si h “Pith Cu 5 G Fig. 15 Cladogram showing the proposed relationships of Heliopithecus leakeyi. Note added in proof Since this paper was submitted for publication in September 1984 new specimens have been found and named from East Africa (Leakey & Leakey 1986). We have not yet had the opportunity of making direct comparisons of Heliopithecus leakeyi with this new material, but the published descriptions and examination of casts indicates that there is a strong similarity between them. The new material is from the site of Kalodirr west of Lake Turkana. It has been named Afropithecus turkanensis by R. E. and M. G. Leakey (1986) and the type specimen consists of a relatively complete skull with a number of unusual and rather baboon-like characters of the facial skeleton. In addition there are less complete specimens from the same site and from Buluk, east of Lake Turkana, which had been described in an earlier paper (Leakey & Walker 1985). The much less complete specimen from Ad Dabtiyah described here is not so well preserved as the East African material, but the parts that are preserved in common show a high degree of similarity. Many characters of the Afropithecus specimens which are shared with other early Miocene and earlier fossil anthropoids would appear to represent primitive retentions for the Homi- noidea. This applies to the wide interorbital distance, the massive glabellar region, the narrow and lightly built supraorbital tori not linking across the glabella, the oval-shaped nose, the nasal floor morphology, the single infra-orbital foramen, the relatively large lateral incisors compared with medial incisors, the heteromorphic premolars, and molars retaining distinct lingual cingula. These last two characters are seen to be present also in Heliopithecus, and the two genera also share the distinctive premolar enlargement described here. For example, the P?/M! crown module ratio for Afropithecus is just over 100%, which is at the limits of the 392 P. J. ANDREWS & L. MARTIN gorilla range, and the P*/M? ratio is 96%. In both cases the greatest similarities are with the Miocene genera Heliopithecus and Kenyapithecus together with the Moroto palate, which we have subsequently suggested represents a second species of Heliopithecus (Andrews, Martin & Whybrow 1987). In terms of size, the Afropithecus specimens appear to group with the Moroto palate from Uganda and are considerably larger than Heliopithecus from Saudi Arabia. Because of this, there is little doubt about the species differentiation between the Saudi Arabian and African material, but it is unclear whether the generic distinction is justified. Without changing the main text of the present paper, we would like to place on record our doubts about the generic distinction. Additional material from Saudi Arabia providing data on the face of Heliopithecus, or information on the enamel structure and thickness of the teeth of Afropithecus, would either confirm or remove these doubts. Acknowledgements We are grateful to Peter Whybrow and Terry Harrison for comments on the text. Alan Boyde provided encouragement, support and SEM facilities for the enamel microstructure work. L.M. was supported by an MRC Research Training Fellowship and P.J.A. acknowledges funds from the British Council and the Wenner Gren Foundation for the work in Spain. References Andrews, P. J. 1978. A revision of the Miocene Hominoidea of East Africa. Bull. Br. Mus. nat. Hist., London, (Geol.) 30: 85-224. — 1985. Family group systematics and evolution among catarrhine primates. In: Delson, E. (ed.), Ancestors: The Hard Evidence: 14-32. New York. & Cronin, J. 1982. The relationships of Sivapithecus and Ramapithecus and the evolution of the orang-utan. Nature, Lond., 297: 541-546. , Hamilton, W. R. & Whybrow, P. J. 1978. Dryopithecines from the Miocene of Saudi Arabia. Nature, Lond., 274: 249-251. —, Martin, L. & Whybrow, P. J. 1987. Earliest known member of the great ape and human clade. Am. J. Phys. Anthrop., New York, 72: 174-175. — & Walker, A. C. 1976. The primate and other fauna from Fort Ternan, Kenya. In: Isaac, G. & McCown, E. R., (eds), Human Origins: 279-304. Menlo Park. Fleagle, J. E. 1986. The fossil record of early catarrhine evolution. In: Wood, B. A., Martin, L. B. & Andrews, P. J. (eds), Major Topics in Primate and Human Evolution: 130-149. Cambridge. Greenfield, L. O. 1979. On the adaptive pattern of ‘Ramapithecus’. Am. J. phys. Anthrop., Philadelphia, 50: 527-548. —— 1980. A late divergence hypothesis. Am. J. phys. Anthrop., New York, 52: 351-365. Hamilton, W. R., Whybrow, P. J. & McClure, H. A. 1978. Fauna of fossil mammals from the Miocene of Saudi Arabia. Nature, Lond., 274: 248-249. Harrison, T. (1982.) Small-bodied apes from the Miocene of East Africa. 647 pp., 103 figs. Ph.D. thesis, Univ. London (unpubl). Ishida, H., Pickford, M., Nakaya, H. & Nakano, Y. 1984. Fossil anthropoids from Nachola and Samburu Hills, Samburu District, Kenya. Afr. Stud. Monogr., Kyoto, (suppl.) 2: 73-85. Leakey, L. S. B. 1967. An early Miocene member of Hominidae. Nature, Lond., 213: 155-163. Leakey, R. E. & Leakey, M. B. 1986. A new Miocene hominoid from Kenya. Nature, Lond., 324: 143-146. —— & Walker, A. C. 1985. New higher primates from the early Miocene of Buluk, Kenya,'Nature, Lond., 318: 173-175. Le Gros Clark, W. E. & Leakey, L. S. B. 1950. Diagnoses of East African Miocene Hominoidea. Q. JI geol. Soc. Lond., 105: 260-262. Martin, L. B. (1983.) The Relationships of the Later Miocene Hominoidea. 450 pp., 50 figs. Ph.D. thesis, Univ. London (unpubl.). —— 1985. Significance of enamel thickness in hominoid evolution. Nature, Lond., 314: 260-263. —— 1986. Relationships among extant and extinct great apes and humans. In: Wood, B. A., Martin, L. B. & Andrews, P. J. (eds), Major Topics in Primate and Human Evolution: 161-187. Cambridge. —— & Andrews, P. J. 1982. New ideas on the relationships of the Miocene hominoids. Primate Eye, Cambridge, 18: 4~7. AD DABTIYAH HOMINOID 393 Pickford, M. 1982. New higher primate fossils from the middle Miocene deposits at Majiwa and Kaloma, Western Kenya. Am. J. phys. Anthrop., New York, 58: 1-19. Pilbeam, D. R. 1969. Tertiary Pongidae of East Africa: evolutionary relationships and taxonomy. Bull. Peabody Mus. nat. Hist., New Haven, 31: 1-185. Simpson, G. G. 1931. A new classification of mammals. Bull. Am. Mus. nat. Hist., New York, 59: 259-293. Ward, S. C. & Pilbeam, D. R. 1983. Maxillofacial morphology of Miocene hominoids from Africa and Indo-Pakistan. In: Ciochon, R. L. & Corruccini, R. S. (eds), New Interpretations of Ape and Human Ancestry: 211-238. New York. we FES PEGs | we Lh Mastodons from the Miocene of Saudi Arabia A. W. Gentry Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 S5BD Synopsis Mastodon dental remains from Ad Dabtiyah, Saudi Arabia, are assigned to Gomphotherium cooperi, a species hitherto known only from the basal Miocene of Dera Bugti, Pakistan. The Ad Dabtiyah popu- lation would best fit a late Lower Miocene or earliest Middle Miocene date—equivalent to the middle Orleanian of Europe and coming between Rusinga and Maboko in east Africa. A right M* + M? from 60km south of Ad Dabtiyah is tentatively included in G. cooperi. Introduction All but one (M.42946) of the Proboscidea described in this paper come from Ad Dabtiyah, Saudi Arabia, and were collected by P. J. Whybrow, H. A. McClure and the late W. R. Hamilton in 1974. The locality is situated at 26° 27'02” N, 48° 35’ 24” E (Hamilton et al. 1978; see also Whybrow et al., this issue, p. 375) where the fossils occur in continental deposits thought to be laterally equivalent to the nearby extreme limits of the marine Miocene Dam Formation. Register numbers of specimens refer to the collection of the British Museum (Natural History), London. Measurements (Table 2) are given in millimetres. In mastodons the pretrite is the lingual half of an upper molar loph and the labial half of a lower molar lophid. The posttrite is the remaining half of each loph or lophid. Pretrites become worn in advance of the corresponding posttrites. Systematics Order PROBOSCIDEA Illiger, 1811 Family GOMPHOTHERIIDAE Hay, 1922 Genus GOMPHOTHERIUM Burmeister, 1837 Gomphotherium cooperi (Osborn, 1932) Figs 16-22 1932 Trilophodon cooperi Osborn: 3; figs 1—2. MATERIAL. Measurements are given in Table 2. M.42940 Conjoined left M, and back of M,. The rear lophid of M, has begun to wear. Figs 16, 22. M.42941 Left lower tusk. Fig. 17. M.42942 Much of right mandible with M,. M,, present in life, is now missing. All lophids of M; are in wear. Only the back of the rostrum is present. Fig. 18. M.42943 Right M®. The front loph has begun wear. Figs 19, 22. M.42944 Back of left M,. Lophids not yet in wear. M.42945 Left M,. Wear has barely started on the second lophid; incomplete posteriorly. Figs 20, 22. M.42946 Conjoined right M* and M?’. M? is heavily worn and the front two lophs of M? are worn. Rear of M? is damaged and incomplete. Figs 21-22. The locality for this specimen is about 10 km west of Jabal Urayirah, an area about 60 km south of Ad Dabtiyah, where undifferentiated deposits are thought to include equivalents of the Dam and Hofuf Formations (Steineke et al. 1958). The specimen was collected 40-50 years ago by oil company geologists. It is questionably included in G. cooperi. M.42947 Lateral part of proximal left ulna from Ad Dabtiyah. It can be assumed to be conspecific with the teeth M.42940-5. M.42948 Partial right scapula from Ad Dabtiyah. Again conspecificity with the teeth can be assumed. Bull. Br. Mus. nat. Hist. (Geol.) 41 (4): 395-407 Issued 29 October 1987 396 A. W. GENTRY Fig. 16 Gomphotherium cooperi. Occlusal view of left M, and back of M,, M.42940, from Ad Dabtiyah. Anterior side to the left. x 0-75. DESCRIPTION. The teeth all come from one or more species of bunodont, trilophodont mastodon. Except for M.42942 they are preserved only as crowns which has made identifica- tion as uppers or lowers less secure in some cases. The M3 s have four lophids plus a posterior cingulum; the M? M.42943 has three lophs and that of M.42946 has four. The third and fourth lophs and lophids show signs of chevroning (forwardly-directed indenting in their centres). The cones and conules of the lophs and lophids are moderate- to high-crowned. In M.42940 and M.42943 they give the appearance of growing out of a basal shelf and the lower parts of their sides are not closely pressed to one another. Small cingular tubercles may be visible between cones. Besides the main (outside) cone in each half loph or lophid there is one additional conelet budded off towards the longitudinal median line of the tooth. M.42943 shows an irregularly surfaced swelling on the rear of the posttrite of the anterior loph, which could be taken as a rudimentary posterior conelet. The lower tusk M.42941 is peg-like, has a slight twist, a concave upper surface and convex lower surface in cross section, and longitudinal grooves and striations on its lower surface. The latter feature is reminiscent of hippopotamus canines. No enamel band is visible along its outer surfaces. The length, as preserved, is 345mm and its mid-length diameters 45-3 x 28 mm. The two postcranial bones, M.42947-8, will not be considered further. Table 2 Measurements in mm of molars of Gomphotherium cooperi (Osborn) from Saudi Arabia. Width Maximum across Maximum Specimen length 1st loph(-id) width M, M.42940 136 c. 60:0 64-2 M, M.42942 147 64:3 70-6 M, M.42945 GIST c. 69-0 c. 76:1 M? M.42943 148 c. 63-0 66:8 M? M.42946 c. 147 76:1 c. 78-0 M? M.42946 c. 108 = 67-0 3977 MASTODONS °¢-0 x ‘do} 94} 0} [esp ‘VJa] 9Y) 0} Opts [eIpauUt ‘Ysn} Jo s[PprIu sso1Oe UOT}Oas SSIOASUBI} : JOSUT YSN} UIRS JO MIA [P19}R| : MO[IG “YeANQEC PY WO ‘THETP'W ‘YSN} JOMOT Yo] JO MOIA [esIOP :9AOQy ‘1ad000 wniuayjoyduoy se pice Re GREE LI ‘314 398 A. W. GENTRY Background to comparisons Work by Tobien and Tassy since 1970 has greatly improved our grasp of mastodon evolution. Even if not found to be correct in all details their various proposals do at least add up to a comprehensible framework (see Tassy 1983a, b and references; Tobien 1973). It seems that the following groups of Neogene (effectively post-Egyptian Fayum) Old World mastodons can be recognized. 1. Zygodont or ridge-toothed mastodons of Family Mammutidae, Eozygodon Tassy & Pick- ford (1983) in Africa and Zygolophodon Vacek in Eurasia, the latter surviving until the later Pliocene. The remaining mastodons mentioned below are all bunodont. 2. Gomphotherium, an early trilophodont mastodon of Family Gomphotheriidae. The Euro- pean type species, G. angustidens (Cuvier), has been known since Cuvier’s time and G. cooperi and G. browni (Osborn) come from the basal Miocene Nari Formation (its upper part) at Dera Bugti, and from the Middle Miocene Chinji Formation, respectively, of Pakistan. Raza & Meyer (1984: 45) place Bugti in the Chitarwata Formation. 3. Shovel-tusked trilophodont mastodons belonging to Platybelodon Borissiak, Protanancus Arambourg and Archaeobelodon Tassy 1983b, best known from Asia and Africa. This group is put in the subfamily Amebelodontinae of the Gomphotheriidae, named after the North Amer- ican type genus. 4. The persistently trilophodont Choerolophodon Schlesinger, which develops a crowded and irregular pattern of cones and conules on its molars. This genus also is accorded subfamily rank within the Gomphotheriidae. 5. The tetralophodont mastodons Tetralophodon Falconer & Cautley, mainly from the Hip- parion faunas of Europe, and Paratetralophodon Tassy 1983a from the Dhok Pathan Forma- tion of the Siwaliks. They are put into subfamily Gomphotheriinae. 6. More advanced relatives of Tetralophodon, comprising Stegolophodon Schlesinger and Stegotetrabelodon Petrocchi which could in their turn be close to Stegodon Falconer, Pri- melephas Maglio and later elephants. The differences in molar teeth of these four genera, present though they are, are outweighed by the similarities. Their family affiliation has long been variably interpreted. The Arabian mastodon remains are clearly trilophodont and bunodont and the central ques- tions are whether they belong to Gomphotherium or the Amebelodontinae and how far these groups can be separated anyway. Until recently Gomphotherium angustidens was thought to have lived in Europe from the middle of the Orleanian (late Lower Miocene) as at Artenay, France, until the Vallesian (early Upper Miocene). The name of G. angustidens has been used as a blanket and convenient designation for any trilophodont, bunodont mastodon and Osborn (1936: 340; fig. 299) selected as lectotype an M, from Simorre, a middle or late Astaracian locality. However Tobien (1973: 255) drew attention to finds of shovel-like mastodon tusks in western Europe and Tassy (1983b: 462) founded the genus Archaeobelodon for them. Tassy has also emphasized that at Sansan, France (early Astaracian) and in Africa and southern Asia the most abundant early mastodons appear to be amebelodontines. Gomphotherium is held to have strong upper tusks of oval or triangular cross section, twisting outwards and downwards, and with a broad enamel band along the lateral to ventral surface. The lower tusks are short with a rounded oval cross section (Tobien 1973: fig. 3 nos 1-3) and concave dorsal surface. They are set in a long rostrum and Tobien refers to them as peg-tusks. Amebelodont lower tusks are more flattened and can reach enormous dimensions, as in the Asiatic Platybelodon (see Osborn 1936: figs 426, 437). Internally they may acquire a medulla of close-packed dentine tubercles instead of laminated dentine (Tassy 1983a: pl. 1, fig. 2; Gaziry 1976: pl. 3, fig. 2). Much taxonomic weight has been put on these tubercles, which are presum- ably a mechanical response to either pronounced flattening or increased size (as in Schlesinger 1917: pl. 34, fig. 2) of lower tusks. However they are not yet present in early amebelodonts (Tassy 1983a: 126-127) so cannot help to define the subfamily. Upper tusks of amebelodonts 399 MASTODONS ‘ST: 0 xX ‘yednged py Wo ‘7p6zP'W ‘fW YIM 2[QIpueU 7YBII JO SMOIA [eSIOP pur [e1a}e 7] 14ado09 wnisayjoyduoy 400 A. W. GENTRY Fig. 19 Gomphotherium cooperi. Occlusal view of right M*, M.42943, from Ad Dabtiyah. Anterior side of the right. x 0-75. are alleged to be small; this is obviously true in relation to their own lower tusks, but less so in relation to Gomphotherium upper tusks (Osborn 1936; compare figs 416 A2 and 436). Another feature of amebelodonts is the narrowness of their molars and Tobien (1973: fig. 9) demon- strates this as between Asiatic Platybelodon and European and North American Gom- photherium. Also considered helpful is the development of posterior conules on the posttrites which may give rise to a more fully trefoil pattern as on the pretrites. Tassy (1979: 267; 1983b: 466) referred Protanancus macinnesi Arambourg of the lower Middle Miocene of Maboko, Kenya, to the Amebelodontinae. Its molars are certainly nar- rower than in European Gomphotherium and sometimes there is an indication of posterior conules on the posttrites (MacInnes 1942: pl. 4, fig. 2). Moreover somewhat widened lower tusks are known from Maboko. Tassy (1979: 267) took the mastodon of the Lower Miocene of Fig. 20 Gomphotherium cooperi. Occlusal view of left M,, M.42945, from Ad Dabtiyah. Anterior side to the left. x 0-75. 401 MASTODONS ‘$L-0 X ‘JYBII 9Y} 0} Opis JOLIOWUY ‘YeANged PY Jo yINOs Wy C9 WO “ 9b67r W ‘-W Pur ZW 14311 JO MIA [esn[d9Q ‘14ad009 wnidayjoyduoy 17 ‘314 402 A. W. GENTRY Rusinga, Kenya, as also an amebelodontine; the flattened lower tusk from Loperot (Maglio 1969) may be another earlier record than Maboko, although Pickford (1981: 90) has some doubts. Tassy (1983a: 116) put into Protanancus the common mastodon of the Chinji Forma- tion, P. chinjiensis (Pilgrim), which is somewhat advanced on P. macinnesi in, for example, the more obvious posterior conules of the posttrites (Tassy 1983a: fig. 10). The narrow molars and high cones of Protanancus and Platybelodon are definitely different from Gomphotherium, either G. angustidens of Europe or the more rarely preserved G. browni (Osborn 1936: fig. 416) which is contemporaneous and sympatric with Protanancus chinjiensis. As regards posterior conules of the posttrites, these can sometimes be seen in Gomphotherium, as on the front lophid of M, of the G. browni holotype. The relationships of Gomphotherium cooperi are problematical and it could be a junior synonym of G. inopinatus (Borissiak & Beliaeva) of Kazakhstan (see Osborn 1936: fig. 224). The holotype of G. cooperi is a mandible, M.12181 (Osborn 1936: fig. 222). Its M3, and others from Bugti, have length/width proportions closer to the Maboko and Rusinga amebelodontines than to European G. angustidens. However, three Bugti M*s, M.12185, M.12190 and the smaller M.12180 are as wide as in Gomphotherium. Tassy (1983a: 259) surprisingly assigned a Bugti M, to Choerolophodon. The tooth (cast M.11050) is longer than Choerolophodon M,s from Maboko and his idea may be that some shortening of M,; occurred in the earliest Choerolophodon. In western Europe the distinction between the amebelodontine Archaeobelodon and Gom- photherium can be a matter of some difficulty, especially where isolated cheek teeth are involved (Tassy 1983b: 463). It is also noticeable that the amebelodont incisors recorded by Tobien (1973: fig. 15, nos 4-8) come from Sansan and La Grive, France, localities considerably post- dating the arrival of Proboscidea in Europe. If the amebelodontine did evolve from earlier peg-tusked gomphotheres (Tobien 1973: fig. 3, nos 1-3), this would be out of line with Tassy’s (1979: 265) view of the plesiomorphy of flattened incisors, but Tassy himself (1983b: 465) affirms that Archaeobelodon was present in Europe well before the time level of Sansan. The less extreme widening of the tusks of Archaeobelodon than in the Asian Platybelodon suggests the possibility of regional or continental variation in this character. The documentation of the all-important Gomphotherium remains from En Péjouan (Tassy 1983b: 463) is needed to estab- lish that G. angustidens is indeed a species additional to and continentally sympatric with the western European amebelodontine. The rest of this paper is mainly concerned with molar teeth, and those coming from Europe and used in comparisons will be referred to in the traditional manner as Gomphotherium or G. angustidens. Comparisons Against this background comparisons can be made between the Ad Dabtiyah mastodons and other relevant material. M.42940. The M, is the only tooth worth considering in detail. It differs from M,s of Sansan by being smaller, less robust, lower-crowned, and the lophids being more separated at their bases. This last character gives the appearance of the lophids having grown from a basal shelf and is reminiscent of Zygolophodon. It is similar to the Rusinga M, M.15300 (MacInnes 1942: pl. 6, fig. 8) but lacks the irregular cingular tubercles evident in M.15300. It also seems to have less sign of swellings (? incipient conelets) around the lophids. It is not so narrow (Fig. 23, smallest reading for X) and is widest across the third instead of the second lophid. It is also very like the Bugti M3s but shorter than M.12183, with less chevroning of the third lophid than in M.11050, and perhaps with smaller posterior conules of the pretrite trefoils, ie. smaller conules between the lophids. It has one less lophid than M,s of Maboko Protanancus macinnesi and less or no chevroning of the fourth lophid. It has a better developed fourth lophid and less chevroning than Maboko Choerolophodon M.15542 (MacInnes 1942: pl. 6, fig. 7). MASTODONS 403 Fig. 22 Gomphotherium cooperi. Labial views of molars. A—-C, from Ad Dabtiyah. A, left M;, M.42940. B, left M,, M.42945. C, right M?, M.42943. D, right M* from 60km south of Ad Dabtiyah, M.42946. It looks very like the Artenay M, of Ginsburg & Antunes (1966: pl. 3, fig. 2) but is possibly lower-crowned. At first sight it also looks narrower but this is not borne out by Ginsburg & Antunes’ measurements of length x breadth (179 x 82mm) nor by my own measurements from their picture. M.42941. The dorsal concavity of the tusk is very evident but the width is not sufficient for it to fit an amebelodontine (Tobien 1973: fig. 15). M.42942. The M; in the mandible is better preserved and slightly more worn than the very similar M.42940. Its lophids look less as if they are growing out of a shelf. It is more evenly wide along its length instead of having the third lophid noticeably the widest. The posterior lobes of the pretrite trefoils, consituting the conules which lie centrally between the lophids, are small as in M.42940 and no larger than in Rusinga M.15300. These conules appear to be smaller than in M;s from Bugti. The third lophid is less chevroned than in Bugti M.11050. Its length/breadth proportion looks similar to the Artenay M, of G. angustidens (see Gins- burg & Antunes 1966: pl. 3, fig. 2), but the latter looks as if it has less marked anterior conules of the pretrite trefoils and a more obvious shelf from which grow the lophids. 404 A. W. GENTRY M.42943. This M° differs from Sansan examples, e.g. 32534, by being smaller, narrower, lower-crowned, with three instead of four lophs and with less of an anteromedial cingulum. It also has only one, not two, conelets budded off towards the median line on each posttrite loph. It is higher-crowned than the Rusinga M* M.15318 (MacInnes 1942: pl. 5, fig. 3) and has less of a cingular shelf and no obvious tubercles decorating the shelf labially and lingually. It is narrower than the Bugti M*s M.12185 and M.12190, and shows stronger development of the posterior lobes of the pretrite trefoils on lophs 2 and 3. It differs from the Maboko Protanancus by having three instead of four lophs somewhat more widely spaced, being lower-crowned and showing less exaggerated anterior lobes on the pretrite trefoils of lophs 2 and 3. M.42944. Little can be noted about this back of a left M,. M.42945. This M3 looks more advanced than M.42940 in that it is larger and higher- crowned. It differs from Sansan M,3s by being slightly lower-crowned and less robust and having more trace of a basal shelf between the front and second lophid row. Its fourth lophid is less developed than in most of the M3; illustrated by Bergounioux et al. (1953) from the Lisbon ‘middle Helvetian Vb’ faunas, thought to be of late Orleanian age. It is also lower-crowned than some of these M,s. It would fit better with the small number of M,s illustrated from the earlier ‘upper Burdigalian IVb’ faunas of middle Orleanian age (Bergounioux et al. 1953: figs 125, 143, 147, 148, 266). The IVb fauna, later called the R2 fauna, is stratigraphically positioned between the start of N7 and somewhere within N8 of the Blow marine planktonic foraminiferan scale (Van Couvering & Berggren 1977: 299). All the non- zygodont Portuguese mastodons were accepted as G. angustidens by Tobien (1973: 207). It is higher-crowned and wider than Rusinga M.15300. It is difficult to judge whether, when complete behind the fourth lophid, it would have been as long as the Maboko Protanancus, i.e. longer than in Rusinga M.15300. The third lophid may be less chevroned than in Maboko or Chinji Protanancus. The front two lophs of M.42945 are very similar to the same part of the M, of Gom- photherium browni (AMNH 19417, BM(NH) cast M.15035). Possibly the posterior lobe of the posttrite of the first lophid is less marked than in G. browni. The tooth is longer and more robust than in G. cooperi M.12181, but has about the same crown height. M.42946. The M? is as long as M.42943 and wider as well. There appear to have been four lophs and the second posttrite had a posterior conule. There were probably two conelets budded off medianwards on each posttrite loph, although preservation and wear impose some uncertainty about this. At the back of the M? the talon has been incorporated into the third pretrite wear facet. The size of the talon in earlier wear would have been about as in the M’ on the Bugti palates M.12178-9, the normal Gomphotherium size. Its relatively great width is like the Artenay M° of Ginsburg & Antunes (1966: pl. 4, figs 2, 3), but its fourth loph is more advanced. It is very like the Sansan M* 32534, except for the absence of a medianmost lobe on the first pretrite trefoil of the latter which may be fortuitous. Compared with Maboko Protanancus macinnesi, the M* on M.42946 is wider but otherwise very similar. Possibly the wings of the pretrite trefoils are better developed in P. macinnesi. It is similar to the M°? of G. browni (cast M.15035; Osborn 1936: fig. 416) but the chevron effect on the third loph may be more pronounced and the lobes of the pretrite trefoils have less of an appearance of being gathered into the line of the lophs. A median longitudinal groove, incipient on M.15035, is not apparent on M.42946. Compared with the Bugti M*s M.12185 and M.12190 it is more advanced in its fourth lobe and higher cones. Again the posterior lobe of the pretrite trefoil on the first loph of M.42946 is better developed than in the Bugti teeth. The Bugti teeth could be foreshadowing the condition of G. browni. Coming from a different locality and with a definitely advanced morphology, M.42946 may be a different species from the other Arabian mastodon teeth but for the present it need not be separated from them. MASTODONS 405 Conclusions There is evidently a range of variation among the Arabian mastodon teeth. M.42940 is small like Rusinga and Bugti M;s but less narrow than in the single Rusinga example. M.42945 is larger and higher-crowned. All three Ad Dabtiyah M,s are wider than in the Maboko Prota- nancus and could most probably belong to a Gomphotherium, an attribution compatible with the peg-like lower tusk, M.42941. The M* M.42943, however, is narrower than in the Sansan, Artenay or Bugti gomphotheres, and hence appears more akin to an amebelodont. On balance it seems best not to split these teeth into different species but to take them all as one species of Gomphotherium. G. browni is poorly known and it differs in that its molars appear too large and advanced and the lower incisors have a more rounded cross section Maximum 80 width 70 60 Maximum length 130 140 150 160 170 180 mm Fig. 23 Width/length proportion for some mastodon M,s. X = Ad Dabtiyah Gomphotherium cooperi; a = Artenay G. angustidens; m = Maboko Protanancus macinnesi; r = Rusinga amebelo- dontine; s = Sansan G. angustidens; u = Dera Bugti G. cooperi. Upper diagonal line is that along which width is 50% of length; the lower line is 40%. Maximum a width 80 70 eo Maximum length 130 140 150 160 AZO Fig. 24. Width/length proportion for some mastodon M?*s. Symbols and diagonal lines as in Fig. 23. 406 A. W. GENTRY (Tassy 1983a: fig. 25A). G. angustidens, at least as represented in Sansan and later localities, is also more advanced. Probably the best designation for the Arabian species is G. cooperi. Bugti specimens included in this species show a wide range of variation between narrow M,s and wide M?s, but the unerupted M, of the holotype mandible M.12181 could not be improved upon as a match for a species embracing M.42940 and M.42945. Size alone is not a good guide for correlation since European G. angustidens has a consider- ably greater size range than shown on Figs 23—24 for the small BM(NH) sample from Sansan. It would not be reliable to take the small size of the Arabian teeth on these graphs as indicating a pre-Astaracian time of occurrence. Nor is the appearance that the lophs or lophids are growing up from a basal shelf or plate a satisfactory character to use for correlations. Such a shelf is seen in M.42940 and some other early mastodon teeth, but it may also be found in later mastodons, e.g. M.7228, a cast M,3 of “Mastodon pyrenaicus’ Lartet identified by Tassy (1977: 1391) as an Astaracian occurrence of Tetralophodon longirostris (Kaup). Hence the similar structure of M.42940 cannot be regarded as a primitive character indicating an early time level. We are left with the number and height of lophids and lophs, and these suggest that the Ad Dabtiyah specimens lived at a period before the Maboko level in Africa or Sansan in Europe. They are probably not as old as Rusinga in east Africa, and in the European sequence they would best fit a middle Orleanian time level. They are probably younger than the Dera Bugti G. cooperi. M.42946, from 60 km south of Ad Dabtiyah, looks like a Gomphotherium as advanced as that at Sansan and could come from a higher stratigraphical level than at Ad Dabtiyah. It does not appear to be evolving towards G. browni. The Arabian mastodon teeth have not improved understanding of the relations between Gomphotherium and amebelodontines, although the dorsal concavity of the tusk M.42941 sup- ports the idea that they were closely linked. References Bergounioux, F. M., Zbyszewski, G. & Crouzel, F. 1953. Les Mastodontes Miocénes du Portugal. Mem. Servs geol. Port., Lisbon, (NS) 1: 1-139. Gaziry, A. W. 1976. Jungtertiare Mastodonten aus Anatolien (Turkei). Geol. Jb., Hannover, (B) 22: 3-143. Ginsburg, L. & Antunes, M. T. 1966. Considerations sur les mastodontes du Burdigalien de Lisbonne et des Sables de l’Orléanais (France). Revta Fac. Ciénc. Univ. Lisb,, (C) 14 (2): 135-150. Hamilton, W. R., Whybrow, P. J. & McClure, H. A. 1978. Fauna of fossil mammals from the Miocene of Saudi Arabia. Nature, Lond., 274: 248-249. MaclInnes, D. G. 1942. Miocene and post-Miocene Proboscidea from east Africa. Trans. zool. Soc. Lond., 25 (2): 33-106. Maglio, V. J. 1969. A shovel-tusked gomphothere from the Miocene of Kenya. Breviora, Cambridge, Mass., 310: 1-10. Osborn, H. F. 1932. Trilophodon cooperi sp. nov., of Dera Bugti, Baluchistan. Am. Mus. Novit., New York, 585: 1-6. 1936. Proboscidea, I. 802 pp. New York: Am. Mus. Nat. Hist. Pickford, M. 1981. Preliminary Miocene mammalian biostratigraphy for western Kenya. J. hum. Evol., London, 10: 73-97. Raza, S. M. & Meyer, G. E. 1984. Early Miocene geology and paleontology of the Bugti Hills, Pakistan. Mem. geol. Surv. Pakistan, Quetta, 11: 43-63. Schlesinger, G. 1917. Die Mastodonten des K.K. naturhistorischen Hofmuseums. Denkschr. K.K. naturh. Hofmus., Vienna, 1 (Geol.-pal. 1): 1-230. Steineke, M., Harriss, T. F., Parsons, K. R. & Berg, E. L. 1958. Geological map of the western Persian Gulf quadrangle, Kingdom of Saudi Arabia. Miscellaneous geological investigations Map 1—208A. Reston, Va.; U.S. geol. Survey. Tassy, P. 1977. Les mastodontes miocénes du Bassin aquitain: une mise au point taxonomique. C.r. hebd. Seanc. Acad. Sci., Paris, (D) 284: 1389-1392. 1979. Les proboscidiens (Mammalia) du Miocéne d’Afrique orientale: résultats préliminaires. Bull. Soc. geol. Fr., Paris, (7) 21: 265-269. MASTODONS 407 1983a. Les Elephantoidea Mioceénes du Plateau du Potwar, Groupe de Siwalik, Pakistan. Annls Paleéont., Paris, 69: 99-136, 235-297, 317-354. 1983b. Le mastodonte a dents étroites, le grade trilophodonte et la radiation initiale des Ameébélo- dontidae. In: Buffetaut, E.. Mazin, J. M. & Salmon, E. (eds), Actes du symposium paleéontologique G. Cuvier: 459-473. Montbeliard. & Pickford, M. 1983. Un nouveau mastodonte zygolophodonte (Proboscidea, Mammalia) dans le Miocéne inférieur d’Afrique orientale: systematique et paleoenvironnement. Géobios, Lyon, 16: 53-77. Tobien, H. 1973. On the evolution of mastodonts (Proboscidea, Mammalia), Part 1: The bunodont trilophodont groups. Notizbl. hess. Landesamt. Bodenforsch. Wiesbaden, 101: 202-276. Van Couvering, J. A. & Berggren, W. A. 1977. Biostratigraphical basis of the Neogene time scale. In: Kauffman, E. G. & Hazel, J. E. (eds), Concepts and methods of biostratigraphy: 283-306. Stroudsburg, Penn. r Rhinoceroses from the Miocene of Saudi Arabia A. W. Gentry Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 SBD Synopsis Dental and postcranial remains of the rhinoceroses from the Miocene of Ad Dabtiyah, Saudi Arabia, are compared with Oligocene and Miocene Old World representatives of Ronzotherium, Diceratherium, Brachypotherium, Aceratherium, Dicerorhinus, Paradiceros, Rhinoceros and Hispanotherium. From this it is concluded that two species are present, Dicerorhinus aff. sansaniensis and Brachypotherium sp. Both are primitive, and by implication early, members of their genera. Attention is drawn to the absence of Aceratherium and its occurrence in Africa is questioned. The rhinoceroses suggest a woodland habitat at Ad Dabtiyah and a geological age early in the Middle Miocene, not later than the Orleanian in European terms. Introduction The fossils described below were collected in 1974 from continental deposits, thought to be the lateral equivalent of the Dam Formation, at Ad Dabtiyah, Saudi Arabia. This locality is situated at 26° 27'02” N, 48° 35’ 24” E (Hamilton et al. 1978; see also Whybrow et al., this issue, p. 375). Register numbers of individual specimens refer to the collection of the British Museum (Natural History), London. Nomenclature for rhinoceros teeth is shown in Fig. 26, p. 410. Systematics Order PERISSODACTYLA Owen, 1848 Family RHINOCEROTIDAE Owen, 1845 Genus DICERORHINUS Gloger, 1841 Dicerorhinus sp. aff. sansaniensis (Lartet, 1851) Figs. 25, 27-32, 33A, 34-37, 38A, 39A, B, 40, 42-43, 44A, B MATERIAL. Measurements in mm. M.36890 Anterior part of conjoined nasal bones. Fig. 25. M.36891—2 Left M’ and M?, early middle wear, possibly from one individual, occlusal lengths 41-8 and 51-6. Figs 27B, C. M.36893 Left P*, middle wear, occlusal length 37-7. Fig. 27A. M.36894 Right upper molar, probably M7’, early middle wear, occlusal length 49-3. M.36895 Right M7’, middle wear, occlusal length 40-4. Figs 28B, 33A. M.36896 Left M?, middle wear, occlusal length 40-5. M.36897 Right M®, early wear, occlusal length 41-1. Fig. 28A. M.36898 Right P*, middle wear, occlusal length 32-4. Fig. 29B. M.36899 Left P*, middle wear, occlusal length 30-6. Fig. 29A. M.36900 Left P?, anterolabial parts missing, late middle wear. M.36901 Right P?, middle wear, occlusal length 27-3. Fig. 29C. M.35012 Left deciduous P", late wear, occlusal length 24-1. M.36902 Left mandible with labial side of P,, P,-P,, much of M,, much of the labial side of M,. Early middle wear. The front premolar is identified as P, and not P, because M,, itself identified by being more worn than P,, is the fifth tooth from the front of the cheek tooth sequence. Occlusal lengths: P, 19-4, P, 25-4, P, 31-7, P, 36-8, P,P, 113-4, M,c. 43-0. Fig. 30. M.36903 Left mandible with P, to Mg, early middle wear. Occlusal lengths: Pa SOP, IP Slee Wily SID; M, 42:6, M, 41:2, M,—-M, 124-0. Fig. 31. Bull. Br. Mus. nat. Hist. (Geol.) 41 (4): 409-432 Issued 29 October 1987 A. W. GENTRY 410 Fig. 25 ectoloph Parastyle paracone mesostyle metacone crochet metaloph protoloph hypocone cingulum flange protocone postfossette medisinus hypocone antecrochet Fig. 26 Dorsal (above) and lateral (below) views of Dicerorhinus nasal, M.36890 from Ad Dabtiyah. Anterior side to the left. x 0-5. protoconid hypoconid Paraconid hypolophid entoconid metaconid metalophid Nomenclature in rhinoceros upper molar (left) and lower molar (right). Anterior side to right and labial side to top. RHINOCEROSES 411 Fig. 27 Occlusal views of teeth of Dicerorhinus from Ad Dabtiyah. Anterior sides to left. A, left P*, M.36893. B, left M', M.36891. C, left M?, M.36892. M.36904 Back of right mandible with M,—M, and part of M,, early middle wear. Occlusal lengths: M, 45-0, M, 42-7. Possibly the same individual as the last specimen. M.36309 Right P,, middle wear, occlusal length 32-8. Possibly belongs to mandible M.36904. M.36905 Most of the crown of a right P,, early middle wear, occlusal length 29-5. M.36906 Most of the crown of a right P,, middle wear, occlusal length 25-5. Fig. 32. M.36907 Right I,, little worn. Fig. 34 (top). M.36908a, b_ Paired I,s, about half worn by comparison with last specimen. Fig. 34 (bottom). M.35076 Two pieces of mandibular symphysis with alveoli for I, and I,. Fig. 35. M.35075 Ventral part of right scapula, doubtfully rhinocerotid. M.36909 Right humerus, complete and undistorted. Length from top of lateral tuberosity to base of medial condyle 460; length from top of articular head to base of medial condyle 410; least transverse width of shaft 64. Fig. 36. M.36910 Left humerus, less complete and crushed anteroposteriorly in proximal part. M.36912 Right ulna, complete. Overall length 440. Fig. 37. M.36913 Right metacarpal IV complete but partially shattered. Overall length 155; transverse width in middle of shaft 32. Fig. 39A. M.36914 Proximal left metacarpal III. M.36915-6 Two right scaphoids. Fig. 38A. M.36917-8 Parts of left and right unciforms. M.36919 Right magnum. M.36782 Partial right magnum. Fig. 28 Occlusal views of teeth of Dicerorhinus from Ad Dabtiyah. Anterior sides to right. A, right M3, M.36897. B, right M?, M.36895. 412 A. W. GENTRY M.36783 Left femur, complete. Length from top of great trochanter to base of lateral trochlear ridge 520. Fig. 40. M.36784 Left tibia, complete. Length from centre of medial facet proximally to posterior tip of medial facet distally 371; transverse width in middle of shaft 60. Fig. 42. M.36785 Left astragalus, damaged mediodistally. Fig. 43. M.36786-7 Trochleae of one left and one right astragali, both slightly smaller than preceding specimen. M.36788 Left calcaneum. M.36789 Medial side of left cuboid. M.36299 Left metatarsal III, proximal end damaged anteromedially. Overall length 169; transverse width in middle of shaft 41-4. Fig. 39B. M.35077 First phalanx of median digit. Fig. 44A. M.36911 Second phalanx of median digit. Fig. 44B. Many other more fragmentary bones have been left unregistered. DESCRIPTION. The nasal fragment, M.36890, has a small protuberance for a horn base, but its dorsal surface is not very rugose. The tip of the anterior end is missing but it is clear that the portion of bone in front of the horn base is neither long nor at all downturned. The cheek teeth are brachyodont and the premolar row relatively long. The upper molars show a hypocone flange spreading up from the cingulum and meeting a corresponding but smaller flange from the ectoloph to close the postfossette, no lingual cingula, a prominent paracone rib, no mesostyle, the merest trace of constriction of the protocone and no antechro- chet, a small or very small crochet on M! and M? and a moderate-sized one on the M?, and a straight ectometaloph on M?. However, the M? ectometaloph would become more curved in late wear. The upper premolars have only an occasional trace of the lingual cingulum, no fusion between hypocone and protocone, hypocone somewhat narrowly connected with more labial cusps to make a metaloph, and only a poor metacone rib on the labial wall. The lower molars have fewer distinctive characters; they show no anterolabial or posterolabial cingula, and the vertical indentation centrally on the labial wall is weak. The lower premolars show poor anterolingual cingula, a moderate depth of the ventral indentation centrally on the labial wall, an anterolabial concavity on the wall of P,, and a large P, or persistent dP,. The central vertical indentation on the labial wall of P, is quite narrow behind the sharp-angled proto- conid, and the effect is accentuated by the labial flange developed from the protoconid. This may be an individual or a species character. It is possible to split the adult upper teeth into two groups: M.36891—4 on the one hand and M.36895—901 on the other. The second group consists of smaller-sized teeth which are also small in comparison with the rhinocerotid mandibular and postcranial remains at Ad Dabtiyah. The premolars and molars are less wide than in M.36891—3, although M.36894 may be in an intermediate state. The upper premolars of this second group have a hypocone less completely bound in with the metaloph, and in fact in the P? it is altogether isolated—probably an individual variation. The metacone rib is also probably slightly larger giving a less flat or concave appearance. Fig. 29 Occlusal views of teeth of Dicerorhinus from Ad Dabtiyah. A, left P?, M.36899, anterior side to left. B, right P*, M.36898, anterior side to right. C, right P?, M.36901, anterior side to right. 413 RHINOCEROSES €£-0 x ‘YeANged PY Woy snurysosao1g JO ‘Z7O69E W ‘A1QiPUU Io] JO MATA [eINVT OF “BLY 414 A. W. GENTRY Fig. 31 Occlusal view of P;—M,; in left mandible of Dicerorhinus M.36903. Anterior side to the left. x 0-5. The pair of Is in later wear have retained more of their roots than the unworn one and these roots are more or less straight. The unworn incisor may have a more curved course of its crown and root but this is very uncertain. Each lower jaw has preserved nearly all of its vertical ramus. The angle of the jaws projects or bulges a little posteriorly. The lower edges of the horizontal rami are very slightly convex in outline, but only M.36902 is visible as far forwards as diastema level. The back of the symphysis in M.36902 is about level with the back of P,. The postcranial bones are further considered on p. 423, following the comparisons of cranial and dental remains. GROUPS USED IN COMPARISONS. In order to identify and understand these remains, comparisons were made with fossils and illustrations of the following Eurasian and African rhinoceroses: 1. Ronzotherium Aymard, mainly from the illustrations of Brunet (1979), Heissig (1969) and Roman (1912). This was the earliest rhinoceros in Europe at localities like Ronzon and survived until the end of the Oligocene. 2. Diceratherium pleuroceros (Duvernoy), a small rhinoceros from the Upper Oligocene and basal Miocene of Europe. It has two horns side by side at the front of its nasals. Upper cheek teeth in middle wear are illustrated in Piveteau (1958: 440, fig. 77), lowers in Roman (1912: pl. 6, figs 4-6). 3. The hornless rhinoceros Aceratherium Kaup, mainly as illustrated in Guérin (1980), Bonis (1973) and Heissig (1969). It is first known in Europe in the middle of the Oligocene. The small Agenian ‘Dicerorhinus tagicus’ (Roman), possibly conspecific with Protaceratherium minutum (Cuvier) as used by Abel (1910: pl. 2, fig. 8), has crochets on its upper molars, at least in earlier wear, and no fifth metacarpal (Roman 1924). It could be an offshoot of Oligocene Aceratherium. It also occurs in the basal Miocene of Dera Bugti, Pakistan (Cooper 1934: 602) and near the Oligocene/Miocene boundary in Russia (Borissiak 1938a). Larger and smaller sized Acera- therium continued into the Upper Miocene (Vallesian). 4. The large, short-legged and usually hornless Brachypotherium is known from the Agenian onwards and replaced Ronzotherium. It may have originated from a form near Diceratherium Marsh, judged by traces of paired horns in some early examples, or one close to Aceratherium, judged by the difficulties with generic classification of the early species lemanense Pomel (Bonis 1973: 124). It is illustrated in Depéret & Douxami (1902), Hooijer (1966), Viret (1929) and Roman (1912). 5. Miocene species of the horned rhinoceros Dicerorhinus, as in Guérin (1980) and Hooijer (1966). [Dicerorhinus has been validated in preference to Didermocerus by the International Fig. 32 Occlusal view of left P,, M.36906, of Dicerorhinus from Ad Dabtiyah. Anterior side Oonm to left RHINOCEROSES 415 Fig. 33. Occlusal and labial views of labial sides of rhinoceros upper molars from Ad Dabtiyah. Anterior sides to right. A, Dicerorhinus, a right tooth, M.36895 (occlusal and reversed labial views). B, Brachypotherium, a left tooth, M.36300 (labial and reversed occlusal views). Commission of Zoological Nomenclature (1977).] Dicerorhinus sansaniensis (Lartet) and the smaller D. steinheimensis Jager are best known from the Middle Miocene of Europe, D. leakeyi Hooijer from the Lower Miocene of east Africa, D. primaevus Arambourg from the Upper Miocene of north Africa and D. schleiermacheri (Kaup) from the Upper Miocene (Vallesian and Turolian) of Europe. Diceros pachygnathus (Wagner) from the Turolian of Pikermi, Greece, lacks lower incisors but is otherwise very similar to Dicerorhinus schleiermacheri. Dicerorhinus sansaniensis is known back to the Orleanian or late Lower Miocene in Europe (Heizmann et al. 1980: 7; Guerin 1980: 201). Finally it may be mentioned that ‘“Aceratherium’ abeli Cooper (1934: 596) from Dera Bugti appears to be a Dicerorhinus, as already noted by Heissig (1972: 27). 6. The African horned rhinoceros Paradiceros mukirii Hooijer (1968) from the Middle Miocene of Fort Ternan, Kenya (Shipman et al. 1981), related to Diceros. 7. A rather incompletely known group of Miocene horned rhinoceroses held to be related to the Pliocene and Pleistocene Elasmotherium Fischer of Asia and centred on Hispanotherium Crusafont & Villalta, within which Ginsburg & Antunes (1979) would also include the Asiatic Beliajevina Heissig and Caementodon Heissig. Hispanotherium appeared for only a limited duration in Spain and Portugal (Antunes 1979: 20) and is known at what is probably a later horizon in Turkey (Heissig 1976). The African Chilotheridium pattersoni Hooijer (1971) and possibly the Chilotherium Ringstrom of the Chinese Hipparion faunas (Ringstr6m 1924), as well as the much earlier Chilotherium blanfordi (Lydekker), the commonest true rhinoceros at Dera Bugti, could also belong here. All these rhinoceroses were hypsodont and had upper molars with constricted protocones and strong antecrochets; the (? primitive) protocone/hypocone fusion on upper premolars persisted until the start of the late Miocene. There are, however, some differences among them. In particular the Turkish Hispanotherium has a reduced man- dibular symphysis and no enlarged lower incisors (Heissig 1976: 33, fig. 2), whereas the Bugti Chilotherium shares a very wide symphysis (Cooper 1915: figs 4, 5) with the Chilotherium of the Chinese Hipparion faunas. The Chinese Chilotherium has no horns, the state of C. blanfordi is unknown, Chilotheridium possessed a nasal horn and pneumatized frontals (Hooijer 1971: pls 2, 4) and Beliajevina Borissiak had what must have been a horn base towards the back of the nasals (Borissiak 1938b: 8). In addition to the foregoing references, see also Heissig (1972, 1974), Antunes (1972) and Antunes et al. (1972). 416 A. W. GENTRY Fig. 34 Incisors of Dicerorhinus in medial view to show differences in root curvature. Right I, , M.36907 (above), left I,, M.36908b (below). x 0-5. 8. Rhinoceros browni, first described by Colbert (1934) under the generic name Gaindatherium and figured by him and by Heissig (1972). It is known from the Chinji Formation and other pre-Hipparion localities of the Siwaliks Group in Pakistan, where it predates other rhinoceroses like Aceratherium and Brachypotherium (Guérin in Pilbeam et al. 1979: 36; Barry et al. 1982: 113-4). The taxonomy and history of Oligocene and earlier Miocene rhinoceroses is confused. Many generic names have been used besides those so far mentioned while multitudes of species-level names have been founded and used in differing combinations with the generic names. Comparative material in the British Museum (Natural History) comprised mainly fossils and casts from the Upper Miocene of Eppelsheim, Germany and the Lower and Middle Miocene of some French localities, the Lower Miocene of Jebel Zelten, Libya, the Lower Miocene of some Kenyan localities and the basal Miocene of Dera Bugti, Pakistan. COMPARISONS. In Ronzotherium upper molars it is mainly the massive lingual and posterolabial cingula, the stronger indication of a mesostyle, the posteriorly open postfossette and the curved ectometaloph of M? which differ from the Arabian teeth. Strong cingula also occur on Ronzo- therium lower molars and premolars. The upper premolars differ by having strong cingula as in the molars, a stronger metacone rib and fusion between the protocone and hypocone. Radinsky (1967: 5) and Heissig (1969: 15) agree that primitively rhinocerotid P*s and P*s would have had a protocone linked or almost linked by a protoloph to the ectoloph. The hypocone was definitely linked with the protocone- protoloph but only more weakly via the metaloph to the ectoloph. This is the condition found in Ronzotherium, whereas in other rhinoceroses the hypocone is part of a metaloph and linked only weakly or in later wear, if at all, to the protocone. They thus look much more like molars. Fig. 35 Anterior view of mandibular symphysis of Dicerorhinus, M.35076. Notice alveoli for I,s as well as I, s. Natural size. RHINOCEROSES 417 The change has been carried less far in Aceratherium in which a narrow protocone—hypocone link may be present (e.g. A. cf. platyodon Mermier, of Roman & Viret 1934: pl. 8, fig. 1). In Hispanotherium (Heissig 1976: pl. 1, figs 14, 15) this same link has survived undiminished from its ancestral Ronzotherium-like condition (or has been strengthened anew, helped perhaps by antecrochet growth from the protoloph, from an Aceratherium-like condition). Against this background the Arabian P?® is interesting in that the hypocone is separated both from the 36 Fig. 36 Anterior view of right humerus of Dicerorhinus, M.36909. x 0:33. Fig. 37 Medial view of right ulna of Dicerorhinus, M.36912. x 0-33. 418 A. W. GENTRY protocone and from a metaloph growing towards it from the ectoloph. It has, however, already developed a posterolabial flange (perhaps a raising of the old posterior cingulum) which is closing off the postfossette posteriorly. The P, or persistent dP, on M.36902 is almost as large as in a Ronzotherium filholi from Bournouncle I, France (Heissig 1969: fig. 15C; table 14), although in other Ronzotherium the P, may be only a single-rooted peg or altogether absent. The upper premolars of Diceratherium pleuroceros do not have the primitive Ronzotherium- like fusion of protocone and hypocone but the teeth are otherwise little advanced. They differ from the Arabian species by being smaller, the M* having a curved ectometaloph (only avail- able from the aged specimen 28845, a cast of the holotype skull), and anterolabial and antero- lingual cingula being present on the lower molars. “Dicerorhinus’ tagicus has stronger cingula on the premolars and molars, stronger crochets and mesostyles, some sign of antecrochets on the upper molars and only a peg-like P,. The only figured specimen also lacks a nasal horn. Aceratherium differs very distinctly by its tendency to have a reduced paracone rib and by the antecrochets and constricted protocones on the upper molars, by the curved ectometaloph and localized posterolabial cingulum on M?, and by an internal cingulum and narrow protocone-— hypocone fusion on the upper premolars. The lower molars have a stronger labial indentation between the metalophid and hypolophid, and small but definite anterolabial and anterolingual cingula. P, is smaller. Brachypotherium has rather primitive teeth but has nonetheless developed some special- izations of its own. The large size, wide and evenly flat or slightly concave ectoloph surface behind the rather insignificant paracone rib, persistence of internal cingula on its upper cheek teeth and of external cingula on its upper and lower molars are all different from the Arabian form. Orleanian Brachypotherium already had a smaller P, (Mayet 1908: pl. 2, fig. 2). The P? of ‘Rhinoceros (Diceratherium) asphaltense’ Depéret & Douxami (1902: pl. 2, fig. 1), which Bonis (1973: 123), following Schlosser (1904: 443), includes in B. lemanense, is one of the few in which the hypocone is not linked by a metaloph to the ectoloph. Other cases are found in some Ronzotherium P*s, e.g. that shown by Heissig (1969: fig. 13). The distinctive Hispanotherium has hypsodont upper molars, often with abundant cement; the paracone rib is probably weaker than in the Arabian specimens, the protocone strongly constricted and an antechrochet is present. On the upper premolars the protocone and hypo- cone are fused and the metacone rib is strong. The lower molars have more of a labial 20 mm Fig. 38 Right scaphoids of rhinoceros from Ad Dabtiyah in medial view (above) and dorsal view (below). Anterior sides to left. A, Dicerorhinus, M.36916. B, Brachypotherium, M.36302. RHINOCEROSES 419 indentation between metalophid and hypolophid. As regards reduction of the anterior pre- molars, even P, has diminished in size (Heissig 1976: 33, fig. 2). Paradiceros mukirii was described as an early relative of the living African black rhinoceros, Diceros bicornis, apparently about twice as old as the well-known Turolian Diceros pachygnathus. Paradiceros differs most notably from the Arabian species by the absence of lower incisors. Other differences are its shorter premolar row without a P,, stronger antero- lingual cingula on the lower premolars, and probably the larger crochets of the upper molars. It is not clear why Rhinoceros browni need be generically separated from Dicerorhinus as used here (cf. Groves 1983: 310). Alleged differences are that it has no sign of a horn base on the frontals, the top of the nasals are less bent downwards and the anterior border of the orbit lies above the middle of M’. The first two characters resemble later Rhinoceros but could as easily fit a female Dicerorhinus. As to the third, the front of the orbit lies above P* in adult modern R. unicornis, above M', perhaps even its back half, in D. sumatrensis, and above the M!/M7? boundary in Miocene Dicerorhinus. Here again the state of R. browni could fit Dicero- rhinus as easily as Rhinoceros. The union of the posttympanic and postglenoid processes beneath the external auditory meatus, mentioned by Colbert (1934), is like Rhinoceros and Miocene Dicerorhinus but unlike D. sumatrensis. Modern R. unicornis and sondaicus have a longer P? and P? than D. sumatrensis (Guérin 1980: table 5) but there is no foreshadowing of this in R. browni. Rhinoceros browni differs from the Arabian species in having a smaller crochet and curved ectometaloph on M2, a shorter premolar row, a more prominent metacone rib on the upper premolars and probably a stronger mesostyle rib on the upper molars. Crochet size on M' and M? of R. browni must be variable according to the illustrations of Colbert (1934: fig. 4) and Heissig (1972: pl. 1, figs 7, 8). P; was reckoned by Colbert (1934: 9) to be absent in R. browni and although Heissig (1972, pl. 2, fig. 3) figured one, it was nevertheless smaller than in the Arabian rhinoceros. The anterolabial wall of P3; is very slightly concave in the Arabian species—more so than in many later rhinoceroses but like R. browni in Heissig (1972: pl. 2, fig. 3). Fig. 39 Metapodials of rhinoceros from Ad Dabtiyah in anterior view. A, Dicerorhinus, right meta- carpal IV, M.36913. B, Dicerorhinus, left metatarsal III, M.36299. C, Brachypotherium, left metatar- sal III, M.36308. A. W. GENTRY 420 'C£-0 X “ESLOEW ‘Snulys0.1a91q JO INUIT] Ia] JO MOIA JOLIQUY Op “SI RHINOCEROSES 421 Dicerorhinus shows fewer differences in its teeth from the Arabian species than genera already mentioned and some species-level comparisons are needed. Dicerorhinus primaevus Arambourg (1959: 56) comes from the Upper Miocene of Bou Hanifia (= Oued el Hammam), Algeria. It has larger teeth than the Arabian species and larger crochets on M! and M?. Unfortunately premolars are known only from the milk dentition. Dicerorhinus abeli differs from the Arabian species by having larger crochets on its upper molars, a smaller P,, a lingual cingulum at the medisinus entrance on the upper premolars, a stronger metacone rib on the upper premolars, antero- or posterolabial cingula on the lower molars and a shallow anterolabial concavity on the wall of P,. D. leakeyi appears to show fewer differences: a less straight ectometaloph in M* (Hooijer 1966: pl. 7, fig. 5), a more prominent metacone rib on the upper premolars, more fusion between hypocone and protocone in later wear on the upper premolars and deeper labial grooves on the lower premolars. P, is present in one of the two specimens but is not quite as large as in M.36902. D. sansaniensis, known to me only from illustrations, differs by the probably stronger fusion between hypocone and protocone of the upper premolars in later wear. It also has better lingual cingula on both its upper molars and premolars and larger labial grooves on its lower molars. One mandible (Guérin 1980: pl. 9C) has a P, almost as large as in the Arabian specimen, but the holotype (Guérin 1980: pl. 6) has a smaller P,. D. steinheimensis is a smaller species. According to Guérin (1980: table 47), the smaller Arabian upper teeth (M.36895—901) would match D. steinheimensis in size and the larger ones (M.36891—4) D. sansaniensis. I shall not, however, split the Arabian material at species level. D. schleiermacheri is a later form. Its upper molars contrast with the Arabian form in their weaker paracone rib but stronger mesostyle rib. The closure of the postfossette by flanges from hypocone and ectoloph is also more apparent. Its upper premolars show fusion between the protocone and hypocone in middle and late wear and a stronger metacone rib. The lower molars have more of a central indentation on their labial walls and the P, is smaller or absent. The lower incisors from Arabia, M.36907-8, are smaller than the large ones assigned to Aceratherium in the Eppelsheim collection but larger than BM(NH) 21490 in the same collec- tion assigned to D. schleiermacheri. The more or less straight roots of the more worn pair make them more akin to Dicerorhinus according to Hooijer (1966: pl. 4, figs 2-5). Guerin (1980: 218), following Heissig (1972), points out that incisors of Dicerorhinus differ from those of Acera- therium in possessing a neck. The worn pair from Saudi Arabia do not have such a neck whereas the unworn one does. It may also be noted that the mandible M.35076 shows alveoli for two small I,s, thereby agreeing with D. leakeyi (Hooijer 1966: 123) and D. schleiermacheri (M.2781). The posterior projection of the angle of the lower jaws differs from Aceratherium, in which the back of the vertical ramus descends in more of a straight line. The Arabian mandibles resemble two casts of Dicerorhinus schleiermacheri from Eppelsheim, Dicerorhinus sansaniensis (Guerin 1980: pls 5, 6), D. leakeyi (Hooyer 1966: pl. 2, fig. 4) and Brachypotherium (Mayet 1908: pl. 2, figs 1, 2; Roman 1912: pl. 8, figs 1, 3). The lower edge of the horizontal ramus in M.36902 does not curve upwards anteriorly so much as in Diceratherium pleuroceros (Roman 1912: pl. 6, fig. 4) or as in Aceratherium or Dicerorhinus schleiermacheri. Like D. sansaniensis and D. leakeyi, in Brachypotherium it looks curved but less so than in Aceratherium. Heissig (1972: 21) gives a forwardly-directed mandibular symphysis as a character of the subfamily Rhinocerotinae. The nasal fragment M.36890, with its small horn base, is unlike Aceratherium and most Brachypotherium which are hornless. It is also unlike the twin-horned Diceratherium or some early Brachypotherium with vestiges (?) of paired horns (Osborn 1900: 253, figs 12B, D; Dietrich 1931: 210, figs 10, 11). The smallness and absence of surface rugosity could suggest that the bone is from a female, juvenile, primitive or geologically old animal. The absence of down- turning anteriorly is unlike Dicerorhinus sansaniensis or D. leakeyi but like D. schleiermacheri and Rhinoceros browni. Whether or not there was a frontal horn in the Arabian species is not known. A. W. GENTRY 422 "b-0 X “SOE9EW ‘wniuayjoddyonag;, JO INUII] 1J9] JO MOIA [e1O}R[OIOIUY Tp “SIA RHINOCEROSES 423 Fig. 42 Anterior view of left tibia of Diceror- hinus, M.36784. x 0-33. POSTCRANIAL BONES. The ventral part of a damaged right scapula, M.35075, could belong to a rhinoceros but the glenoid facet is very narrow transversely. It is probably too small to fit a mastodon. The well-preserved right humerus, M.36909, is about the size of those of Dicerorhinus leakeyi, schleiermacheri and orientalis (Schlosser) listed in Hooijer (1966: 160, table 28). It is about as wide at the proximal as at the distal end. The olecranon fossa is deep at the distal end. Compared with the humeri of Pikermi rhinoceroses, the shaft looks longer between the base of the deltoid crest and the supinator ridge distally. The humerus of Ronzotherium filholi (Osborn), as figured by Brunet (1979: pl. 20b, c), is a slighter bone and that part of the distal end lateral to the condyles is narrower. It agrees well with M.2783, the cast of a specimen from Eppelsheim labelled as D. schleiermacheri. The right ulna, M.36912, is again complete; even the top of the olecranon is present in its entirety. The process at the top of the olecranon projects strongly medially and the mid-shaft diameter from front to back is considerably less than in Pikermi examples of Diceros pachygnathus. 424 A. W. GENTRY Fig. 43 Anterior view of left astragalus of Dicerorhinus, M.36785. x 0-5. The right metacarpal IV, M.36913, is as curved and less thickset than M.18814 from Rusinga (perhaps belonging to D. leakeyi; Hooier 1966: pl. 12, figs 2, 3); the latter is itself less thickset than Pikermi examples probably belonging to Diceros pachygnathus. M.36913 has no facet for articulation with a metacarpal V. The two right scaphoids, M.36915-—6, are less thickset than examples from Pikermi probably belonging to Diceros pachygnathus. There is only poor development of a downwards projection posteroventrally and this may be a resemblance to Dicerorhinus rather than to Aceratherium (see Bonis 1973: fig. 36). The complete left femur, M.36783, is about the size of Pikermi examples but more gracile especially at the distal end. The third trochanter is smaller and higher on the shaft and in lateral view the top of the great trochanter slopes downwards anteriorly. These are additional differences from the Pikermi bones and the downward slope of the great trochanter may be linked with the third trochanter appearing to be higher on the shaft. The Arabian femur is not quite so long as a cast of an Eppelsheim femur (M.1283, labelled D. schleiermacheri), on which the great trochanter has a similar slope and the third trochanter is about the same size as the Arabian one. The left tibia, M.36784, is the size of the Dicerorhinus leakeyi and D. orientalis listed in Hooyer (1966: 171, table 40) but a bit longer than the D. schleiermacheri. It is about the size of a rhinoceros tibia 27458 from Sansan, but less gracile as shown by the wider distal part of the shaft and distal articular surface. Compared with Pikermi rhinoceros tibiae it is slightly more gracile, the tibial tuberosity at the proximal end is less massive in proximal view and in anterior view there is less of a deep groove at the proximal end. The left astragalus, M.36785, matches M.2786 from Eppelsheim which is labelled as D. schleiermacheri and is unlike 1290 and M.2785 labelled as Aceratherium incisivuum from the same locality. The agreement with M.2786 lies in the prominent overhang of the lateral parts of the proximal trochleae and in the top edge of the front of the cuboid facet being widely separated from the base of the more proximal trochlear facets on the anterior surface, but it must be stated that Pikermi examples of astragali of D. schleiermacheri or Diceros pachygnathus do not match the Eppelsheim bone in these respects. M.36785 is taller than the Pikermi astragali and its proximal lateral trochlea is less bulbous. M.36785 is larger than most of the east African Dicerorhinus and Aceratherium listed by Hooijer (1966: table 42) but is not as large as the only one listed as definitely Dicerorhinus—that which belongs to the associated skeleton no. 2. The left calcaneum, M.36788, has its tuber less thick, front to back, than in Pikermi examples. The left metatarsal III, M.36299, is more slender than Pikermi ones. The first and second phalanges of the median digit, M.35077 and M.36911, by comparison with Hooijer (1966: pl. 10, figs 4-7) are seen to match Dicerorhinus rather than Brachy- potherium. There is also another, more damaged, median second phalanx and a number of phalanges of the side toes. RHINOCEROSES 425 20 mm Fig. 44 Rhinoceros bones from Ad Dabtiyah. A, Dicerorhinus, median 1st phalanx in anterior view, M.35077. B, Dicerorhinus, median 2nd phalanx in anterior view, M.36911. C, Brachypotherium, left astragalus in anterior view, M.36306. Discussion. Early Dicerorhinus species and Rhinoceros browni emerge as close to the Arabian material. The Dicerorhinus species with which comparisons have been made come from contin- entally-separated localities and the one which differs least could be D. sansaniensis. It has to be noted, however, that the characters of the latter were available to me only from plates in Guerin (1980) and Filhol (1891). Many characters in this assessment would also probably turn out to be variable in large samples in spite of their long-standing use in rhinoceros taxonomy. Furthermore, on dental morphology alone the Arabian species could only be distinguished from Diceratherium pleuroceros by its greater size. The Arabian species can be named as Dicerorhinus sp. aff. sansaniensis, but this may reflect no more than the attainment of a similar evolutionary level of tooth morphology. In particular it need not imply zoogeographical relationship with Europe. A skull from Ad Dabtiyah would be needed to make a reliable identification. As far as geological age is concerned, Sansan itself is of Astaracian age but D. sansaniensis is known back to the Orleanian as noted previously. In Africa D. leakeyi is best known from the time range 20—-16-:5 Ma, which probably corresponds to the Orleanian in Europe, and Hoower (1978: 374) regards Alengerr, Kenya (12-14 Ma?) as the latest reasonable record. The preorbital part of its skull is longer than in D. sansaniensis (Hooijer 1966: pl. 1; Guérin 1980: pls 5, 6) which would fit with the suggestion of an earlier age than Sansan. The Dera Bugti D. abeli could well date from the basal Miocene (Eames 1950, Khan 1968), 1.e. have an age equivalent to the Agenian in Europe. Some of the other Dera Bugti mammals support this assessment, for example the mandibular piece (M.12339) of ‘Amphicyon’ shahbazi (Pilgrim), which looks like Pseudocyonopsis Kuss, a genus extinct in Europe after the Agenian (Springhorn 1977: 37). All this suggests that the rather primitive Dicerorhinus of Ad Dabtiyah could be of an age equivalent to the Orleanian land mammal age in Europe. The rather large size of the dP' (M.35012) could also support an early date for Ad Dabtiyah. It appears to be about 10% longer relative to M? than in D. primaevus or Astaracian D. sansaniensis (Guérin 1980: 233), and thereby in closer agreement with D. leakeyi and D. abeli. NUMERICAL APPROACHES. As an alternative to the above comparisons a matrix was drawn up of 25 cranial and dental character differences in 18 taxa of Oligocene and Miocene rhinoceroses. The characters used were: SKULL: . Paired horn bases present or absent at front of nasals. . Horn bases present or absent posteriorly on nasals. . Lower incisors directed forwards or upwards. . Lower edge of mandible convex or straight. . Premolar row short or long compared with molar row. . Cheek teeth higher- or lower-crowned. UPPER MOLARS: . Paracone rib prominent or weak. . Trace of mesostyle absent or present. CN DANhWN 426 A. W. GENTRY 9. Protocone partially constricted or not. 10. Hypocone partially constricted or not. 11. Antecrochet present or absent. 12. Crochet present or absent. 13. Hypocone flange closing postfossette or not. 14. M? ectometaloph straight or curved. 15. Lingual cingula absent or present. UPPER PREMOLARS: _ 16. Lingual cingula absent or present. 17. P? and P* with hypocone and protocone unfused or fused. 18. Metacone rib absent or present on labial wall. 19. Crenulations present or absent on front of metaloph. LOWER MOLARS: 20. Antero- or posterolabial cingula absent or present. 21. Central labial indentation deep or shallow. LOWER PREMOLARS: 22. Anterolingual cingula absent or present. 23. Concavity on anterolabial wall of P, absent or present. 24. Central labial indentation shallow or deep. 25. P, absent or present. In the above list the first mentioned alternative was considered advanced. The state of all characters was ascertainable in the Arabian material, but not invariably in the other taxa. Every character state existed in more than one taxon so there were no unique occurrences. Two dendrograms were constructed from this matrix. The first was a phenetic dendrogram (Gentry 1974: 184, based on a method of Corbet & Hanks), for which one counts the differences of the taxa from one another, standardizes the totals as percentages of the number of characters compared, and associates on the dendrogram those forms which are least different. The second dendrogram associates forms sharing the largest numbers of supposedly advanced similarities. Neither dendrogram allows for parallel evolution or displays the contri- bution of individual characters but each has a percentage scale of difference or similarity. There were difficulties in constructing these dendrograms. Once a character difference had been spotted between two taxa it was often hard to assign other taxa to one or other state and adoption of intermediate categories was not always satisfactory. Secondly, consistency of assessment was hard when using photographic illustrations, despite the excellent reproduction in many older publications. And finally, on the second dendrogram some character polarities were in doubt. As noted earlier, several early Brachypotherium specimens have traces of bipar- tite horn bases and this could imply descent of hornless rhinoceroses from ancestors with Diceratherium-like horns. Again, if Radinsky (1966: 636) is right that procumbency of I,s is part of the initial family-level specialization of rhinocerotids, then the more upright I,s of Acera- therium should be counted as secondary despite their unspecialized appearance. Alternatively Aceratherium could be removed from the Rhinocerotidae. The following comments can be made on the two dendrograms (Fig. 45). 1. The Arabian rhinoceros forms part of a grouping of Dicerorhinus species and Rhinoceros browni on both dendrograms and within that grouping it associates with early or primitive Dicerorhinus rather than with D. schleiermacheri. 2. The phenetic dendrogram presents compfehensible major groupings of the rhinoceroses despite the limitations on its construction. One sees on it the three clusters of (A) the mid- Tertiary hornless rhinoceroses plus the primitive Ronzotherium and Diceratherium; (B) horned rhinoceroses centred upon Dicerorhinus; (C) the Hispanotherium group. The last two groups are also recognizable on the second dendrogram but here advanced Brachypotherium and Acera- therium attach themselves to Hispanotherium presumably because of parallel advances. The closeness of many of the horizontal linking lines on the chart suggests that repeats of the same exercise would be unlikely to produce the same result. 3. The association of Chilotherium, Chilotheridium and Hispanotherium on both dendrograms supports the view of their similarities taken above. If they should be a natural group, the early Miocene irruption of Hispanotherium into Iberia may have come from an African origin close RHINOCEROSES 427 Ronzaotherium Drceerathertum pleuroceros Aceratherrum (Oligocene) Aceratherirum (later spp.) “DicerorAtinus” tagrcous Brachypotheritum (Agenian) Brachypotherium (later spp.) Paradiceros mukirit DicerorhAinus schlerermacherit Diceros pachygnathus DircerorhAinus sansaniensis DicerorhAinus leakeyi RAtnoceraos browns Drcerarhkinus abel Arabian DrcerorAtnus CArtlotherium blantfords CAtlotheridium pattersani Hispanotherium Ronzothertum Diceratherium pleuraceras "“DiceroréAinus” tagreus Aceratheritum (Oligocene) RAtnoceras browns Drcerorhinus leakeyi DicerorArinus Sansaniensis Arabian DrcerorAtnus DicerorAinus abeli Dicerorhinus schletiermacherst ——— Diceros pachygnathus Brachypothertum (Agenian) Paraditceros mukrrigd Brachypotheritum (later spp.) Aceératheritum (later spp.) CAtLlotherium blanfordi CAtLlotheridium pattersoni HispanothEeritium SSS Ee ee Fig. 45 Above, phenetic dendrogram of relationships among some Oligocene and Miocene rhino- ceroses. The scale runs from 10% nearest the names to 60% nearest to the base of the tree. Below, advanced characters dendrogram for the same rhinoceroses. The scale runs from 0% nearest the base of the tree to 60% nearest the names. 428 A. W. GENTRY to Chilotheridium instead of from east to west dispersal along the northern side of Tethys (cf. Antunes 1979). Genus BRACHYPOTHERIUM Roger, 1904 Brachypotherium sp. Figs 33B, 38B, 39C, 41?, 44C MATERIAL. Measurements in mm. M.36300 Tooth fragments including a reassembled labial wall of a left upper molar in middle wear. Occlusal length 56:5. Fig. 33B. M.36301 Right radius preserved over its complete length but without anterior parts. Overall length 308. M.36302 Right scaphoid. Fig. 38B. M.36303 Right metacarpal III, much damaged. Overall length c. 148. M.36305 Left femur without posteroproximal part and with damage medially on patellar groove. Overall length 432. Fig. 41. M.36306 Left astragalus, much damaged. Dimensions (after Guérin 1980: fig. 22): height 76, anteropos- terior dimension (DAP int) 51, breadth (DT) 95. Fig. 44C. M.36307 Right cuboid. M.36308 Left metatarsal III, proximal posterior part damaged. Overall length 133; transverse width in middle of shaft 48-5. Fig. 39C. DescrIPTION. The large size of the labial wall of an upper molar (M.36300), its flatness and the small size of the paracone rib in comparison with the large flat area leave no doubt of the generic identity being with Brachypotherium. The right radius M.36301 is a little shorter than rhinocerotid radii from Pikermi belonging to either Diceros or Dicerorhinus, but its shaft width is about the same. Its proportions match closely a radius from Rusinga, M.18908, identified by Hooijer (1966: 148; pl. 9, fig. 1) as Brachypotherium heinzelini Hooijer. It is too short to match the complete ulna M.36912. The scaphoid M.36302 is less high in side view than the other two from Ad Dabtiyah. Judged from the anterior facet at its distal end, the metacarpal III, M.36303, looks less wide than a corresponding metacarpal III, M.18813, from Rusinga identified by Hooijer (1966: pl. 10, fig. 2) as Brachypotherium. The femur, M.36305, is appreciably smaller than the Dicerorhinus femur. Its third trochanter is smaller, it lacks a vertical ridge running down from its great trochanter anterolaterally, and the lateral and medial condyles are less widely separated in ventral view. The presence of any third trochanter at all shows that it must belong to a perissodactyl. If not of a rhinoceros it could perhaps be a chalicothere. Although not very like the femur of the middle Miocene Chalicotherium grande (Blainville) figured by Zapfe (1979: 184, fig. 107), it is like the North American Moropus Marsh of earlier geological age (Coombs 1978: fig. 13B). However, a right femur of a Brachypotherium (Mayet 1908: pl. 2, fig. 3) also looks as if it lacks the vertical ridge, so the present bone is tentatively placed in that genus. The measurements of the astragalus M.36306 show that it is low and wide, as befits a Brachypotherium astragalus. In anterior view the lateral part of the proximal trochleae has quite an overhang. The process low on the medial side also projects well transversely. The top edge of the front of the cuboid facet is widely separated from the base of the more proximal trochlear facets on the anterior surface. The ventral facet for the navicular has a very concave profile in anterior view. The right cuboid, M.36307, is less deep than that in the associated skeleton of Dicerorhinus leakeyi figured by Hooijer (1966: pl. 13, fig. 4) but not quite as shallow as the Brachypotherium figured by Guerin (1980: fig. 48G). The left metatarsal III, M.36308, is shorter and has a wider shaft than the left metatarsal III referred to Dicerorhinus. It is longer than the upper Miocene examples of Brachypotherium measured by Guerin (1980: 342), but Hooijer (1966: table 17) quoted Brachypotherium measure- ments which match it. RHINOCEROSES 429 Table 3. Brachypotherium species from Europe and Turkey: proportions of astragali (height x 100/ breadth) and third metatarsals (breadth x 100/length). Bone Proportion Likely age Source Astragalus 89% Agenian Measured from Bonis (1973: fig. 32). 171% Late Orleanian Ginsburg & Bulot (1984: 358). T4% Late Orleanian BM(NH) M.7760 from Thenay, France. 716% Middle Ataracian Steinheim (Hooyer 1966: 148, quoting Roger). 712% Middle Astaracian BM(NH) 33529 from Villefranche d’Astarac. 712% Late ‘Astaracian’ Heissig (1976: 88). 64% Late Astaracian Guerin (1980: 311). Metatarsal III 31% Agenian Mean of three readings from Repelin (1917: 35, 36); Viret (1929: 267). 37% Orleanian B. stehlini (Hooijer 1966: 147). 37% Middle Astaracian B. brachypus (Lartet) (Hooijer 1966: 147, quoting Roger). 37% Late Astaracian Heissig (1976: 89). 41% Late Astaracian Guerin (1980: 342). Discussion. A distinctive feature of Brachypotherium is the progressive shortening and widening of its limb bones. If this evolved in a straightforward manner during the course of the Miocene it ought to be possible to see how far along the line the Arabian species fits in. Unfortunately there is a shortage of specimens with recorded measurements, but some data for European and Turkish astragali and metatarsals is assembled in Table 3. The Arabian species has these astragalus and metatarsal ratios at 80% and 36% respectively, which show that it is fairly primitive for a Brachypotherium and is likely to date from the Lower or lowest Middle Miocene. One would not expect it to postdate the early Astaracian. However, it should be noted that a Rusinga astragalus measured by Hooijer (1966: 148) has a ratio of 70% yet would probably date in European terms from the earlier Orleanian. Palaeoecology of Arabian rhinoceroses Guerin (1980: 380) and many others have commented on the ‘hippopotamid’ aspect of Brachy- potherium as manifested by its large skull, barrel-like body and short, stubby limbs. The usual and reasonable conclusion from this is that it was an animal of aquatic habitats. A wooded environment is also often mentioned, which would be a point of difference from Hippopotamus amphibius. However, Webb (1983: 289) quotes taphonomic evidence that the north American Teleoceras Hatcher, very like Brachypotherium, ‘lived in the water but grazed on adjacent dense grasslands’ exactly as does the modern hippopotamus (Kingdon 1979: 250)}—an amazing paral- lelism for two mammals in different orders and with different digestive strategies. Miocene Dicerorhinus stands close to the ancestry of the extant horned rhinoceroses and what is known of their ecology can be used as a guide to that of the fossil form. The somewhat specialized African rhinoceroses live in lightly wooded areas, preferably with thickets (Diceros bicornis) or in more open environments with grasses (Ceratotherium simum); both species need access to water (Kingdon 1979: 80-119). Rhinoceros unicornis was in historic times an inhabit- ant of the grassed and wooded Indian alluvial plains and it feeds mainly on grasses (Laurie 1982: table 2). Dicerorhinus sumatrensis inhabits densely wooded areas but prefers their margins and disturbed areas. It can ascend and descend steep slopes with agility and, a century or two ago, may have been found in hillier country than the sympatric R. sondaicus (Groves & Kurt 1972; van Strien 1975: 37). One can therefore conclude that the Arabian Dicerorhinus is likely to have fed by browsing in wooded habitats with easily available water. 430 A. W. GENTRY Zoogeography of Arabian rhinoceroses If the Arabian Brachypotherium really is of Lower Miocene age, as implied by the proportions of its astragalus and metatarsal, it could be conspecific with B. snowi (Fourtau 1918) from Moghara, Egypt. The occurrence of Brachypotherium at this period as far east as Ad Dabtiyah would be interesting inasmuch as Guerin (in Pilbeam et al. 1979: 36) does not find it in the Siwaliks sequence until after Hipparion is present. However, Heissig (1972: 103) claimed to have identified Brachypotherium as far back as in the Kamlial Formation. A comment may also be made here on the question of the occurrence of Aceratherium in Africa. The holotype skull of Turkanatherium acutirostratus Deraniyagala 1951 from the Middle Miocene of Moruorot, Kenya, was subsequently referred to Aceratherium by Aram- bourg (1959: 74). This fossil was the first described from Africa adequate to sustain an identifi- cation as either Aceratherium or something else. Its upper cheek teeth are large and wide by comparison with European Aceratherium. They do show moderate or strong antecrochets, but in this, as in the two preceding characters, they agree well with the later Miocene Brachy- potherium lewisi Hooyer & Patterson from Lothagam (Hooijer & Patterson 1972: 2). It is possible that Deraniyagala’s skull is the same species or lineage as the short metapodials, phalanges and other elements found in east Africa and referred to Brachypotherium, e.g. those of Hooyer (1966: pl. 10, figs 1-3, 6-8). The skull material of Aceratherium campbelli Hamilton (1973: table 3; pls 1, 3) from Zelten, Libya, may also have been incorrectly placed at generic level. Its teeth were large and appear to have been wide; it looks very like the Moruorot skull. The high occiput and concave profile of the cranial roof in both skulls is unlike European Aceratherium (Mermier 1896: pl. 2, fig. 2; Guérin 1980: pl. 3) but can be matched from within Brachypotherium (Mayet 1980: pl. 2, fig. 1). African Brachypotherium, as in Europe identified by its short limb extremities, may have been an entirely separate development and have evolved Aceratherium-like antecrochets on its upper molars. This idea needs further investigation. If it were correct, the absence of Aceratherium at Ad Dabtiyah could be held to align the fauna with Africa rather than with Europe. Unfortunately the concurrent absence of Chilotheridium could be held to indicate the reverse. Hence the rhinoceroses reported in this paper remain zoogeo- graphically inconclusive. References Abel, O. 1910. Kritische Untersuchungen tber die palaogenen Rhinocerotiden Europas. Abh. K.-K. geol. Reichsanst., Vienna, 20 (3): 1-52. Antunes, M. T. 1972. Notes sur la géologie et la paleontologie du Miocéne de Lisbonne. XI—Un nouveau rhinocerotide, Chilotherium ibericus n. sp. Bolm Mus. Lab. miner. geol. Univ. Lisb. 13 (1): 25-33. 1979. ‘Hispanotherium fauna’ in Iberian Middle Miocene, its importance and paleogeographical meaning. Annls geol. Pays hell., Athens, (HS) 1: 19-26. ——., Viret, J. & Zbyszewski, G. 1972. Notes sur la géologie et la paléontologie du Miocene de Lisbonne. X—-Une conference de J. Viret sur l’Hispanotherium (Rhinocerotidae). Quelques données com- plémentaires; autochtonie et endémisme. Bolm Mus. Lab. miner. geol. Univ. Lisb. 13 (1): 5—23. Arambourg, C. 1959. Vertébrés continentaux du Miocéne supérieur de l'Afrique du Nord. Publs Serv. Carte geol. Alger., Algiers, (n.s., Mem. Paléont.) 4: 1-159. Barry, J. C., Lindsay, E. H. & Jacobs, L. L. 1982. A biostratigraphic zonation of the Middle and Upper Siwaliks of the Potwar Plateau of northern Pakistan. Palaeogeogr. Palaeoclimat. Palaeoecol., Amster- dam, 37: 95-130. Bonis, L. de 1973. Contribution a l’étude des mammiféres de l’Aquitanien de lAgenais. Rongeurs— carnivores—perissodactyles. Mem. Mus. natn. Hist. nat. Paris, (C) 28: 1-192. Borissiak, A. A. 1938a. Contribution to the phylogeny of Dicerorhinae. Dokl. Akad. Nauk SSSR, Lenin- grad, 19 (9): 767-770. — 1938b. A new Dicerorhinus from the Middle Miocene of north Caucasus. Trudy paleont. Inst., Moscow, 8 (2): 1-68. Brunet, M. 1979. Les grandes mammiferes chefs de file de l' immigration Oligocene et le probleme de la limite Eocene-Oligocene en Europe. 281 pp. Paris. Colbert, E. H. 1934. A new rhinoceros from the Siwalik beds of India. Am. Mus. Novit., New York, 749: 1-13. RHINOCEROSES 431 Coombs, M. C. 1978. Reevaluation of early Miocene North American Moropus (Perissodactyla, Chalico- theriidae, Schizotheriinae). Bull. Carnegie Mus. nat. Hist., Pittsburgh, 4: 1-62. Cooper, C. F. 1915. New genera and species of mammals from the Miocene deposits of Baluchistan. Ann. Mag. nat. Hist., London, (8) 16: 404-410. 1934. The extinct rhinoceroses of Baluchistan. Phil. Trans. R. Soc., London, (B) 223: 569-616. Deperet, C. & Douxami, H. 1902. Les vertébrés oligocenes de Pyrimont-Challonges (Savoie). Abh. schweiz. palaont. Ges., Geneva, 29: 1—90. Deraniyagala, P. E. P. 1951. A hornless rhinoceros from the Mio-Pliocene deposits of East Africa. Spolia zeylan., Colombo, 26: 133-135. Dietrich, W. O. 1931. Neue Nashornreste aus Schwaben (Diaceratherium tomerdingensis n.g. n.sp.). Z. Saugetierk, Berlin, 6 (5): 203-220. Eames, F. E. 1950. On the age of the fauna of the Bugti bone beds, Baluchistan. Geol. Mag., London, 87: 53-56. Filhol, H. 1891. Etudes sur les mammiferes fossiles de Sansan. Annls Sci. geol. Paris, 21: 1-319. [Also issued as Bibltque Ec. ht. Etud., Paris Sect. Sci. nat., 37 (1): 1-319.] Fourtau, R. 1918. Contribution a l'étude des vertebres miocénes de |’Egypte. 122 pp. Cairo. Gentry, A. W. 1974. A new genus and species of Pliocene boselaphine (Bovidae, Mammalia) from South Africa. Ann. S. Afr. Mus., Cape Town, 65: 145-188. Ginsburg, L. & Antunes, M. T. 1979. Les rhinocérotidés du Miocene inférieur et moyen de Lisbonne (Portugal). Succession stratigraphique et incidences paléogéographiques. C.r. hebd. Seance. Acad. Sci., Paris, 288 (D): 493-495. — & Bulot, C. 1984. Les Rhinocerotidae (Perissodactyla, Mammalia) du Miocéne de Bézian a _la Romieu (Gers). Bull. Mus. natn. Hist. nat. Paris, (4) 6 (C): 353-377. Groves, C. P. 1983. Phylogeny of the living species of Rhinoceros. Z. Zool. Syst. EvolForsch., Frankfurt a.M., 21: 293-313. — & Kurt, F. 1972. Dicerorhinus sumatrensis. Mammalian Spec., New York, 21: 1-6. Guerin, C. 1980. Les rhinoceros (Mammalia, Perissodactyla) du Miocéne terminal au Pleistocéne supérieur en Europe occidentale; comparaison avec les espéces actuelles. Docums Lab. Geol. Fac. Sci. Lyon, 79 (1): 1-421. Hamilton, W. R. 1973. North African Lower Miocene rhinoceroses. Bul. Br. Mus. nat. Hist., London, (Geol.) 24: 349-395. —, Whybrow, P. J. & McClure, H. A. 1978. Fauna of fossil mammals from the Miocene of Saudi Arabia. Nature, Lond., 274: 248-249. Heissig, K. 1969. Die Rhinocerotidae (Mammalia) aus der oberoligozanen Spaltenftillung von Gaimers- heim bei Ingolstadt in Bayern und ihre phylogenetische Stellung. Abh. bayer. Akad. Wiss., Munich, (NF) 138: 1-133. — 1972. Paladontologische und geologische Untersuchungen im Tertiar von Pakistan. 5. Rhinocerotidae (Mamm.) aus den unteren und mittleren Siwalik-Schichten. Abh. bayer. Akad. Wiss., Munich, (NF) 152: 1-112. 1974. Neue Elasmotherini (Rhinocerotidae, Mammalia) aus dem Obermiozan Anatoliens. Mitt. bayer. St. Palaont. Hist. Geol., Munich, 14: 21-35. 1976. Rhinocerotidae (Mammalia) aus der Anchitherium-Fauna Anatoliens. Geol. Jb., Hannover, (B) 19: 3-121. Heizmann, E. P. J., Ginsburg, L. & Bulot, C. 1980. Prosansanosmilus peregrinus, ein neuer machairodon- tider Felide aus dem Miocan Deutschlands und Frankreichs. Stutt. Beitr. Naturk., (B) 58: 1—27. Hooijer, D. A. 1966. Fossil mammals of Africa, No. 21. Miocene rhinoceroses of East Africa. Bull. Br. Mus. nat. Hist., London, (Geol) 13 (2): 117-190. 1968. A rhinoceros from the Late Miocene of Fort Ternan, Kenya. Zool. Meded. Leiden, 43 (6): 77-92. 1971. A new rhinoceros from the Late Miocene of Loperot, Turkana District, Kenya. Bull. Mus. comp. Zool. Harv., Cambridge, Mass., 142: 339-392. 1978. Chapter 19, Rhinocerotidae. In: Maglio, V. J. & Cooke, H. B. S. (eds), Evolution of African Mammals: 371-378. Cambridge, Mass. (Harvard U.P.). & Patterson, B. 1972. Rhinoceroses from the Pliocene of northwestern Kenya. Bull. Mus. comp. Zool. Harv., Cambridge, Mass., 144: 1-26. International Commission on Zoological Nomenclature 1977. Opinion 1080. Didermocerus Brookes, 1828 (Mammalia) suppressed under the plenary powers. Bull. zool. Nom., London, 34 (1): 21—24. Khan, M. H. 1968. The dating and correlation of the Nari and Gaj Formations. Geol. Bull. Punjab Univ., Lahore, 7: 57-65. 432 A. W. GENTRY Kingdon, J. 1979. East African mammals, an atlas of evolution in Africa, 11 B. 436 pp. London. Lartet, E. 1851. Notice sur la colline de Sansan. Annu. Dep. Gers, 1851: 3—45, 1 pl. Laurie, A. 1982. Behavioural ecology of the greater one-horned rhinoceros (Rhinoceros unicornis). J. Zool., Lond., 196: 307-341. Mayet, L. 1908. Etude des mammiféres miocénes des Sables de l’Orleanais et des faluns de la Touraine. Annls Univ. Lyon, (N.S.) 1 (24): 1-336. Mermier, E. 1896. Etude complémentaire sur I’ Acerotherium platyodon. Annls. Soc. linn. Lyon, 43: 225-240. Osborn, H. F. 1900. Phylogeny of the rhinoceroses of Europe. Bull. Am. Mus. nat. Hist., New York, 13: 229-267. Pilbeam, D. R., Behrensmeyer, A. K., Barry, J. C. & Shah, S. M. I. 1979. Miocene sediments and faunas of Pakistan. Postilla, New Haven, Conn., 179: 1-45. Piveteau, J. (ed.) 1958. Mammiféres. Evolution. Traite de Paleontologie 6 (2). 962 pp. Paris. Radinsky, L. B. 1966. The families of the Rhinocerotoidea (Mammalia, Perissodactyla). J. Mammal., Lawrence, 47: 631-639. — 1967. A review of the rhinocerotoid family Hyracodontidae (Perissodactyla). Bull. Am. Mus. nat. Hist., New York, 136: 1-45. Repelin, J. 1917. Les rhinocérotidés de ’Aquitanien supérieur de lAgenais (Laugnac). Annls Mus. Hist. nat. Marseille, 16: 1-45. Ringstrom, T. 1924. Nashorner der Hipparion-Fauna Nord-Chinas. Palaeont. sin., Pekin, (C) 1 (4): 1-156. Roman, F. 1912. Les rhinocéridés de l’Oligocéne d’Europe. Archs Mus. Hist. nat. Lyon, (7) 11 (2): 1-92. 1924. Contribution a l’étude de la faune de mammiferes des Littorinenkalk (Oligocene supérieur) du Bassin de Mayence. Les Rhinocéros. Trav. Lab. Geol. Univ. Lyon, (7) 6: 1-54. & Viret, J. 1934. La faune de mammiferes du Burdigalien de la Romieu (Gers). Mem. Soc. géol. Fr., Paris, (N.S.) 9 (21): 1-67. Schlosser, M. 1904. [Review of Depéret & Douxami 1902]. Neues Jb. Miner. Geol., Stuttgart, 1904 (1): 443-446. Shipman, P., Walker, A., Van Couvering, J. A., Hooker, P. J. & Miller, J. A. 1981. The Fort Ternan hominoid site, Kenya: geology, age, taphonomy and palaeoecology. J. hum. Evol., London, 10: 49-72. Springhorn, R. 1977. Revision der alttertiaren Europaischen Amphicyonidae (Carnivora, Mammalia). Palaeontographica, Stuttgart, (A) 158: 26-113. Strien, N. J. van 1975. Dicerorhinus sumatrensis (Fischer), the Sumatran or two-horned Asiatic rhinoceros. Meded. ned. Comm. int. NatBerscherm., Amsterdam, 22: 1-82. Viret, J. 1929. Les faunes de mammiferes de l’Oligocéne supérieur de la Limagne bourbonnaise. Annls Univ. Lyon, (N.S.) 47: 1-328. Webb, S. D. 1983. The rise and fall of the late Miocene ungulate fauna in north America. In: Nitecki, M. H. (ed.), Coevolution: 267-306. Chicago. Zapfe, H. 1979. Chalicotherium grande (BLAINV.) aus der miozanen Spaltenfiillung von Neudorf an der March (Dévinska Nova Ves), Tschechoslowakei. Neue Denkschr. naturh. Mus. Wien, 2: 1-282. Ruminants from the Miocene of Saudi Arabia A. W. Gentry Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 SBD Synopsis Six ruminant species have been identified from the Miocene of Ad Dabtiyah, Saudi Arabia. These com- prise two tragulids belonging to Dorcatherium, the giraffoid Canthumeryx, and Eotragus and two other bovid species. The tragulid teeth, known only from the smaller of the two species, are well worn. This fauna seems to date from the early Middle Miocene, which would be the later Orleanian or earliest Astaracian of the European scale. Introduction The ruminant fossils to be described in this paper come from Miocene continental deposits thought to be laterally equivalent to the Dam Formation, at Ad Dabtiyah, Saudi Arabia. This locality is situated at 26° 27'02” N, 48° 35’ 24” E. Further details are given in Hamilton et al. (1978) and Whybrow et al. (this issue, p. 375). Register numbers of individual specimens refer to the collection of the British Museum (Natural History), London. Systematics Order ARTIODACTYLA Owen, 1848 Infraorder TRAGULINA Flower, 1883 Family TRAGULIDAE Milne Edwards, 1864 Genus DORCATHERIUM Kaup, 1833 Dorcatherium sp. Figs 46, 48D Some tragulid remains from Ad Dabtiyah are about the size of Dorcatherium pigotti Whitworth (1958: 9) of the east African early Miocene and Dorcatherium sp. of Colbert (1935: 311) of the Siwaliks Group, Pakistan. They are smaller than D. libiensis Hamilton (1973: 80; pl. 1, fig. 1) from Gebel Zelten, Libya. The first is part of a fragmentary right mandible with P, and P,, M.34278 (Fig. 46C), in which P, and the back of P, are well worn. The occlusal lengths of P, and P, are 6:5 and 8-5 mm respectively. Two pieces of right mandibles (Fig. 46D, E) also show M, (M.34279a) and M;s (both M.34279a and b) in late wear. What can be discerned of the occlusal pattern on the M,s does not look like a pecoran. The central cavities have wide posterior openings to the outside and the labial lobes are insufficiently narrow and pointed. They lack basal pillars and are thus unlike D. libiensis. The occlusal lengths of the M,s are 12-8 and 13-2mm. Three right upper molars and two right upper molars, M.30131 (Fig. 46A) and M.30132 respectively, belong to a tragulid of similar size to the above pieces. They are very well worn. They have a cingulum around the lingual lobes, especially the anterior one, as is usual in tragulids. The posterolabial rounding of the anterior lingual lobe (protocone) also looks tragu- lid rather than pecoran. The rib between parastyle and mesostyle is less deflected anteriorly than in the Siwaliks Dorcatherium of Colbert (1935: figs 138, 141, 143), which perhaps indicates that the Arabian fossils are geologically older. The occlusal lengths of M! and M? in M.30131 are 8-3 and 9:7 mm; the two upper molars in M.30132 measure 8-7 mm (front one) and 9-1 mm (rear one). Bull. Br. Mus. nat. Hist. (Geol.) 41 (4): 433-439 Issued 29 October 1987 434 A. W. GENTRY Fig. 46 Teeth of Dorcatherium from Ad Dabtiyah. Anterior side to right (except B). A, occlusal view of right M' *, M.30131. B, occlusal view of left dP*, M.30133. C, occlusal and lateral views of right P;, M.34278. D, occlusal view of right M, in mandible, M.34279b. E, lateral view of another right mandible with M, _,, M.34279a. M.30133 is a left dP* in late middle wear with an occlusal length of 9-3 mm (Fig. 46B). It has a notably large parastyle and a cingulum on its lingual lobes. The strong parastyle and the convergent front and back walls of the tooth suggest it is a dP* and not a molar. Thomas et al. (1982: 127) have reported a similar-sized Dorcatherium from Al-Sarrar. A right and a left astragalus (M.35265; M.35266, Fig. 48D), a left naviculocuboid plus ectocuneiform (M.35267) and a proximal left metatarsal (M.35268) have tragulid morphology and are about the size of D. pigotti. Dorcatherium, \arger sp. Fig. 48A—C A further left astragalus, M.35269 (Fig. 48C), is larger than those assigned to the above Dor- catherium sp. and matches the size of D. chappuisi Arambourg of the east African early Miocene (see Whitworth 1958: 4). Much of a left tibia including the distal end, M.35079 (Fig. 48A, B), could also belong here. It is larger than the tibia of the living Hyemoschus aquaticus and the distal articular facet is longer anteroposteriorly and less wide transversely. These proportions are nearer to those of Suidae and therefore presumably nearer to the ancestral state. The total length of the tibia would have been in the region of 150mm. Infraorder PECORA Linnaeus, 1758 Superfamily GIRAFFOIDEA Simpson, 1931 Genus CANTHUMERYX Hamilton, 1973 Canthumeryx sp. Fig. 48E A much damaged but practically complete giraffoid left metatarsal, M.34277, is about 320— 330mm long and about 36mm wide across its distal condyles. It is close in size to the RUMINANTS 435 metatarsal of Canthumeryx sirtensis Hamilton from Moruorot, Kenya (Arambourg 1947: pl. 22, fig. 5; Hamilton 1978: 178) but about 25%-30% longer and thereby presumably more advanced. It is too gracile to belong to a Palaeomeryx such as that known from Sansan, France. A left unciform, M.35078 (Fig. 48E), is also of an appropriate size to belong to Canthumeryx. Compared with modern Bovidae it shows more resemblance to Tragelaphini than to Alcela- phini, in that the back of the facet for the cuneiform is less deeply excavated towards the ventral edge of the bone, and also in that there are separate dorsal and ventral facets on the posterior part of the medial side. It differs from examples of extant giraffids, cervids and bovids in having a prominent downturned flange posteriorly. Thomas et al. (1982: 124) recorded a probable Canthumeryx from the Al-Sarrar locality. They illustrated an upper molar (1982: pl. 116, fig. 5) on which the rather bulky condition of styles and anterior rib are presumably owing to late wear. They draw attention to the difficulty in identifying giraffoids from isolated upper teeth. Among their other remains was a metacarpal with a length and distal condylar width of 360mm and 38mm respectively. According to Pickford (1981: 96) Canthumeryx belongs to ‘Set IP faunas in east Africa, dating from 18-5— 16.5 Ma and possibly to “Set II’ ones at 16:-5—14-5 Ma. It is zoogeographically interesting that the mandible with M,_, of Progiraffa exigua from Dera Bugti (Pilgrim 1911: pl. 1, fig. 1) looks as if it could be congeneric with Canthumeryx (as Hamilton had independently recognized— Patterson 1981: 462). This slightly later faunal element contrasts with the antique aspect of most of the Bugti mammals and perhaps came from a higher level; Pilgrim (1912: 2) and Pascoe (1964: 1656) refer to bones in the deposit being scattered through most of its thousand- foot (305m) thickness. Canthumeryx precedes and may be ancestral to Giraffokeryx Pilgrim, and also to pre-Hipparion representatives of Palaeotragus Gaudry which Hamilton (1978: 200) believed to be not congeneric with the type species P. roueni. Canthumeryx is unknown from Europe. Superfamily BOVOIDEA Simpson, 1931 Family BOVIDAE Gray, 1821 Genus EOTRAGUS Pilgrim, 1939 Eotragus sp. Figs 47A, 48F This is represented by M.30134, a left P,, P, and M, in early wear and with occlusal lengths 9-3, 11-1 and 12:0mm respectively (Fig. 47A). The central parts of the M, are missing. The size and proportions of the teeth would be similar to Protragocerus Depéret, e.g. P. labidotus Gentry (1970: 247) of Fort Ternan, except that they are lower-crowned and the basal pillar on the molar is smaller, both of which make them like Eotragus as illustrated in Thenius (1952). A trace of an anterior cingulum on the molar is like both Eotragus and P. gluten (Pilgrim) of the Siwaliks. The premolars match Eotragus and Protragocerus in that the anterolabial wall is not turned to lie in a transverse plane, in the diagonal alignment of the metaconid, and in the weak differentiation of a paraconid. The molar is about the size of a bovid lower molar from Maboko, Kenya (Whitworth 1958: Fig. 10a—c), but the labial lobes may be more narrowly pointed. The teeth are a little larger and higher-crowned than in the bovid-like pecoran Walangania africanus (Whitworth 1958) from the east African early Miocene (Hamilton 1973: 146), and the molar has a smaller basal pillar, no diagonal fold on the rear wall of the protoconid (Palaeomeryx fold), a weaker metastylid and a weaker anterior rib on its lingual wall. A probable bovid from Arrisdrift, Namibia (Hendey 1978: fig. 12) is smaller and has quite a strong paraconid onits P,. Eotragus is the most likely identity for M.30134. This genus is known in Europe from the Orleanian and Astaracian (MN 4-7 of Mein 1975, 1979). It also appears to be represented at Gebel Zelten, Libya by horn cores M.26688 and M.26689 (Hamilton 1973: 127; pl. 13, fig. 1), and comes in at Maboko in the east African succession of Miocene faunas with a cranial roof 436 A. W. GENTRY M.15544 (Gentry 1970: 303; Thomas 1979: 296). It may be present at the later Fort Ternan locality (Gentry 1970: 261; pl. 15, figs 6, 7). Pickford (1981: 96) has Maboko as one of his ‘Set III’ faunas with a probable age back to 16-5 Ma. A partial right horn core of a bovid, M.34270, has only the anterior part of its base pre- served, along with the supraorbital pit and part of the frontal. Its insertion angle is inclined backwards at about the same angle as in Eotragus. The supraorbital pit is narrow. Not enough is left of the horn core above the pedicel for definite identification as Eotragus. The greater part of a left astragalus, M.34271 (Fig. 48F), rather larger than in Walangania, could also belong to the same species as M.30134. Bovidae, Genera indet. Bovid species 2 Fig. 47B The back part of an unworn right M,, M.34272, looks like a bovid and comes from a larger species than M.30134. It is like M.42005, the back of an M, identified as cf. Oioceros sp. and coming from Jabal Midra ash-Shamali in the Hadrukh Formation (Whybrow et al. 1982: 110). It is smaller than M.42005 but about as hypsodont. Further, the entostylid part of the middle lobe on the lingual side of the tooth is at a markedly lower level than the entoconid constitut- ing the rest of this lobe, a character unlike other bovids but probably present in M.42005. Bovid species 3 Figs 47C-F, 48G-K The back of a right M,, M.34273 (Fig. 47C), is in early middle wear and may belong to a species smaller than M.30134 and about the size of the bovid from the Dam Formation at Al-Sarrar (Thomas et al. 1982: 126; pl. 116, figs 6, 7). It is about as high-crowned as Eotragus. Fig. 47 Bovid teeth from Ad Dabtiyah. Anterior side to right (except E). A, Eotragus sp., occlusal and medial views of left P,-M,, M.30134. B, bovid sp. 2, lateral-occlusal view of back part of right M,, M.34272; a indicates low entostylid behind entoconid. C-F, bovid sp. 3. C, occlusal and lateral views of back part of right M,, M.34273. D, occlusal view of right lower molar, M.34274. E, occlusal view of left P*, M.34275. F, occlusal view of right P*, M.34276. RUMINANTS 437 The hypoconulid, or the main part of the rear lobe, has a flat lingual wall. A flange, present at its posterior end, does not loop round anteriorly to meet the entostylid or entoconulid and thus enclose a central cavity. This makes the specimen unlike Walangania but more like most Eotragus and the Arrisdrift, Namibia (Hendy 1978) species. But its condition is not like the early Eotragus which at Artenay, France (Ginsburg & Heintz 1968; pl. 1, figs 4a, b) has a completely enclosed central cavity. The entostylid of M.34273 is less large than in the Arrisdrift specimen. The central cavity on the second lobe does not open to the exterior posteriorly. Thomas et al. (1982: 126) suggested that the Al-Sarrar bovid was congeneric if not conspe- cific with the Arrisdrift species. However, it can be seen from their figure (1982: pl. 116, fig. 6) that P;_, are about as long as in the Arrisdrift specimen whereas M,_, is considerably longer. From my own measurements the premolar row length would have been around 75% of that of the molar row length in the Arrisdrift specimen whereas on the two illustrated Al-Sarrar specimens the same ratio is around 61%. It is thus unlikely that they could be conspecific. A right lower molar in middle wear with an occlusal length of 9-5 mm, M.34274 (Fig. 47D), could also belong to the same species. It is smaller than the M, of M.30134 and is close in size to the molars of Gazella from Gebel Zelten (Hamilton 1973: 128; pl. 13, figs 2, 3). It has transversely long labial lobes, a tiny basal pillar, and may be higher-crowned than M.30134. A left P* in middle wear, M.34275 (Fig. 47E), has an occlusal length of 7-1 mm and a right P* in early wear, M.34276 (Fig. 47F), has an occlusal length of c. 7-8mm. They are low-crowned but smaller than Eotragus. The second can be seen to differ from Eotragus by its stronger metastyle and stronger rib in front of it, and by a narrower lingual part of the tooth. The last feature is presumably primitive. A proximal right tibia, M.35080 (Fig. 48G, H), smaller than Protragocerus labidotus (Gentry Fig. 48 Ruminant postcranial bones from Ad Dabtiyah; cross-hatching indicates areas of broken bone. A-C, Dorcatherium, larger sp. A, medial view of distal left tibia, M.35079. B, articular surface of same, anterior side towards top. C, anterior view of left astragalus, M.35269. D, Dorcatherium sp., anterior view of left astragalus, M.35266. E, Canthumeryx sp., lateral view of left unciform, M.35078. F, ?Eotragus sp., anterior view of left astragalus, M.34271. G—K, bovid species 3. G, lateral view of proximal right tibia, M.35080. H, articular surface of same, anterior side towards right. I, articular surface of distal left tibia, M.35070, anterior side towards top. J, medial view of partial left calcaneum, M.35071. K, anterior view of distal right humerus, M.35074. 438 A. W. GENTRY 1970: pl. 11, fig. 2) or Walangania, may also be of this species. It shows boselaphine or possibly primitive characters like P. labidotus in that there is only a shallow hollowing between the two flanges in the centre of the top articular surface, there is little differentiation of a tubercle and medial hollow in the area in front of the flanges, and the lateral edge of the lateral facet is not upwardly turned. A distal left tibia, M.35070 (Fig. 481), has an anterior as well as a posterior facet for the fibula. Three partial calcanea, M.35071—3 (Fig. 48J), look pecoran but come from animals smaller than Walangania. A distal right humerus, M.35074 (Fig. 48K), could also be of the same species. Either or both of two first phalanges, M.35081—2, also belong to a species smaller than Walangania africanus. The central groove on the proximal articular surface passes completely to the front edge of the bone. In this they look fully pecoran, i.e. more advanced over the tragulid condition than in some BM(NH) examples of early Miocene ‘Dremotherium’ from Allier, France. Of course it is conceivable that in early pecorans the front and back legs might be different for such a character as this. The Ad Dabtiyah phalanges have at best only indistinct facets at the back of the proximal articular surfaces for the sesamoid bones. Discussion of ruminant fauna The tragulid teeth from Ad Dabtiyah are well worn but appear to belong to Dorcatherium, first known from the early Miocene (c. 19-20 Ma) of east Africa and probably contemporaneously in Europe. It may also be noted that Dorcatherium occurs in the basal Miocene of the upper part of the Nari Formation at Dera Bugti, Pakistan; a cast M.11080 of M,—-M, of ‘Gelocus ?gajensis’ Pilgrim (1912: pl. 25, fig. 5) is very close or identical to the east African D. chappuisi. A larger species of Dorcatherium is represented by postcranial bones. Canthumeryx has been recorded, but no remains definitely identifiable as Cervidae. The main bovid species from Ad Dabtiyah resembles an Eotragus; the second and larger species is like the primitive cf. Oiceros sp. of the Hadrukh Formation (Whybrow et al. 1982: 109); a third probable bovid is smaller than the Eotragus and is possibly the same species as that described by Thomas et al. (1982: 126) from Al-Sarrar. Ruminants like these suggest that the fauna dates from before the period of Fort Ternan and is about equivalent to Maboko in the east African Miocene. It would thus be from the early Middle Miocene and would be expected to correspond to later Orleanian or early Astaracian in European terms. This is a more precise temporal placing than was possible with the rhino- ceroses (Gentry, this issue, pp. 425, 429) and is later than suggested by the mastodon (Gentry, this issue, p. 406). Thomas et al. (1978: 71) and Thomas (1983) record Protragocerus, Gazella and Caprotragoides potwaricus (Pilgrim) from the Hofuf Formation overlying the Dam Forma- tion. Such an assemblage would match Fort Ternan and kindred sites and gives support for the earlier age suggested for the underlying fauna herein described. Palaeoecologically it may be noted that Dorcatherium is an important constituent of the Kenyan early Miocene communities held by Evans et al. (1981: 116) to have inhabited forests. Modern tragulids too are inhabitants of forests. With Canthumeryx the difficulty would be to choose between okapi (forest) and giraffe (savanna) as analogues for habitat choice. The appearance of any bovids at all in a Miocene fauna can be held to herald less closed habitats, but only one of the species shows possible beginnings of hypsodonty which could, but need not, be correlated with grazing. Maboko is the African locality with the most similar list of rumi- nant taxa and here the deduced habitat was woodland (Evans et al. 1981: 112). The ruminants, as well as the mastodon and rhinoceroses described elsewhere in this issue, largely fail to indicate zoogeographical relationships for the Ad Dabtiyah fauna. Canthumeryx, being unknown in Europe, would have suggested an African affinity but for the Manchar Formation, Pakistan (Raza et al. 1984: 591) and Bugti occurrences of the probably congeneric Progiraffa exigua. Detailed study of original material of Gomphotherium, Dicerorhinus, Dor- catherium and perhaps Eotragus in various museum collections might, or might not, allow one to discriminate morphologically between allied species living on different continents during the Miocene, but hitherto there has been little consideration of this question. RUMINANTS 439 References Arambourg, C. 1947. Contribution a l’etude géologique et paléontologique du bassin du lac Rodolphe et de la basse vallée de ’Omo. 2, Paléontologie. Mission scient. Omo 1932-1933, Paris, I Geol. Anthrop. (3): 232-562, 40 pls. Colbert, E. H. 1935. Siwalik mammals in the American Museum of Natural History. Trans. Am. phil. Soc., Philadelphia, (N.S.) 26: 1-401. Evans, E. M. N., Van Couvering, J. A. H. & Andrews, P. J. 1981. Palaeoecology of Miocene sites in western Kenya. J. hum. Evol., London, 10: 99-116. Gentry, A. W. 1970. The Bovidae (Mammalia) of the Fort Ternan fossil fauna. In: Leakey, L. S. B. & Savage, R. J. G. (eds), Fossil Vertebrates of Africa 2: 243-323. London & New York. Ginsburg, L. & Heintz, E. 1968. La plus ancienne antilope d’Europe, Eotragus artenensis du Burdigalien d’Artenay. Bull. Mus. natn. Hist. nat. Paris, (2) 40: 837-842. Hamilton, W. R. 1973. The Lower Miocene ruminants of Gebel Zelten, Libya. Bull. Br. Mus. nat. Hist., London, (Geol.) 21 (3): 73-150. — 1978. Fossil giraffes from the Miocene of Africa and a revision of the phylogeny of the Giraffoidea. Phil. Trans. R. Soc., London, (B) 283: 165-229. , Whybrow, P. J. & McClure, H. A. 1978. Fauna of fossil mammals from the Miocene of Saudi Arabia. Nature, Lond., 274: 248-249. Hendey, Q. B. 1978. Preliminary report on the Miocene vertebrates from Arrisdrift, South West Africa. Ann. S. Afr. Mus., Cape Town, 76: 1-41. Mein, P. 1975. Proposition de biozonation du Neogene Mediterranéen a partir des mammiféres. Trab. Neogeno-Cuaternario, Madrid, 4: 112. 1979. Rapport d’activiteé du groupe de travail vertébrés mise a jour de la biostratigraphie du Neogéne basée sur les mammifeéres. Annls geol. Pays hell., Athens, (Tome hors série) 3: 1367-1372. Pascoe, E. H. 1964. A manual of the geology of India and Burma (3rd edn), 3: 1345-2130. Calcutta (Misc. Publs Geol. Surv. India 4). Patterson, C. 1981. Methods of paleobiogeography. In: Nelson, G. & Rosen, D. E. (eds), Vicariance biogeography, a critique: 446-489. New York. Pickford, M. 1981. Preliminary Miocene mammalian biostratigraphy for western Kenya. J. hum. Evol., London, 10: 73-97. Pilgrim, G. E. 1911. The fossil Giraffidae of India. Mem. geol. Surv. India Palaeont. indica, Calcutta, (NS) 4 (1): 1-29. 1912. The vertebrate fauna of the Gaj Series in the Bugti Hills and the Punjab. Mem. geol. Surv. India Palaeont. indica, Calcutta, (NS) 4 (2): 1-83. Raza, S. M., Barry, J. C., Meyer, G. E. & Martin, L. 1984. Preliminary report on the geology and vertebrate fauna of the Miocene Manchar Formation, Sind, Pakistan. J. Vert. Paleont., Norman, Ok., 4: 584-599. Thenius, E. 1952. Die Boviden des steirischen Tertiars. Sher. Ost. Akad. Wiss., Vienna, (I) 161: 409-439. Thomas, H. 1979. Les bovidés miocénes des rifts est-africains: implications paléobiogéographiques. Bull. Soc. géol. Fr., Paris, (7) 21: 295-299. 1983. Les Bovidae (Artiodactyla, Mammalia) du Miocene moyen de la Formation Hofuf (Province du Hasa, Arabie Saoudite). Palaeovertebrata, Montpellier, 13 (5): 157—206. , Sen, S., Khan, M., Battail, B. & Ligabue, G. 1982. The lower Miocene fauna of Al-Sarrar (Eastern province, Saudi Arabia). ATLAL, JI Saudi Arab. Archaeol., Jeddah, 5: 109-136, pls 110-116. ——., Taquet, P., Ligabue, G. & Del’Agnola, C. 1978. Découverte d’un gisement de vertébrés dans les depots continentaux du Miocéne moyen du Hasa (Arabie Saoudite). C. r. somm. Seanc. Soc. geol. Fr., Paris, 1978 (2): 69-72. Whitworth, T. 1958. Miocene ruminants of East Africa. Fossil Mammals Afr., London, 15: 1—5O0. Whybrow, P. J., Collinson, M. E., Daams, R., Gentry, A. W. & McClure, H. A. 1982. Geology, fauna (Bovidae, Rodentia) and flora from the early Miocene of eastern Saudi Arabia. Tertiary Res., Leiden, 4: 105-120. : i : "ig! nt Se aa = == , | (MON. Cah ‘ j ‘ = ; ‘on. Lime <5 + 4 ! oeé-4i : - : ' a Sa: : WEE A) Pia = rr , nS Os ii ro ae be iF "wr. \ rh, oa Miocene Suidae from Ad Dabtiyah, eastern Saudi Arabia M. Pickford Department of Palaeontology, Johannes-Gutenberg-Universitat, Postfach 3980, Saarstrasse 21, D-6500 Mainz, BRD Synopsis Two of four suid teeth from Ad Dabtiyah, Saudi Arabia, are ascribed to Listriodon; the others are fragmentary. A Middle Miocene (Mein Zone 4a or 4b) age is suggested. The wider significance of the assemblage is discussed. Introduction Four suid teeth from Ad Dabtiyah, Saudi Arabia, have been examined and assigned to two taxa, one of which is a reliable marker for the Middle Miocene period in Europe, India and Africa. The best preserved specimens, an M? and an M,, belong to a bunodont, thick- enamelled listriodont similar to Listriodon lockharti (Pomel) from early Middle Miocene strata of Europe, and to Listriodon akatikubas Wilkinson from sites of similar age in Kenya. The remaining two specimens are fragmentary, one of them being an incompletely formed tooth germ. Both, however, are bunodont and recall the genera Conohyus Pilgrim and Kenyasus Pickford, but the material is too incomplete for the purposes of specific identification. A specimen from another collection from Saudi Arabia suggests affinities with Kenyasus, a rela- tively common genus in the Kisingiri sites of Rusinga, Karungu and Uyoma, dated about 17-8 + 0-2 Ma (Drake et al., in prep.). Viewed as an assemblage, the four teeth would not be out of place in early Middle Miocene strata of East Africa and Europe, perhaps 16-17 Ma old. The listriodont teeth in particular appear to represent a period of evolution prior to Listriodon splendens von Meyer of Europe and L. pentapotamiae (Falconer) of late Middle Miocene strata of Pakistan. Correlation of the Ad Dabtiyah site with Mein Zone 4a or 4b, and Maboko (Kenya), is indicated. This leads us to postulate that the Ad Dabtiyah site preserves faunal elements which lived just after the closure of the Tethys at the beginning of the Langhian Stage, an event which not only provided ‘dryshod’ access for Eurasian faunas to Africa and vice versa, but also resulted in major global climatic changes recorded in western Kenya, the Tethys, and as far afield as Japan. Zoogeo- graphic boundaries shifted latitude during this event, which resulted in the widespread estab- lishment of African faunal elements in southern Eurasia and a marked influx of Eurasian faunal elements into Africa. Listriodon was one of these widespread genera, which makes it a useful biostratigraphic marker for this and immediately subsequent periods. Systematics Subfamily LISTRIODONTINAE Simpson, 1945 Genus LISTRIODON Von Meyer, 1846 TYPE SPECIES. Listriodon splendens von Meyer, 1846 Listriodon cf. lockharti (Pomel, 1848) Figs 49-50 MATERIAL. Right M, M.42949; right M? M.42950. DESCRIPTION. M.42949 is a right M, lacking the anterior buccal cusp. The cusp morphology and simple talonid of this tooth indicate that it belongs to the genus Listriodon, although its Bull. Br. Mus. nat. Hist. (Geol.) 41 (4): 441-446 Issued 29 October 1987 442 M. PICKFORD enamel is somewhat thicker and the cusps more bunodont than in L. splendens and L. pentapo- tamiae. The median accessory cusp is large but closely joined to the cross lophs which are well formed, but low. The ‘furchen’ or grooves are well marked considering the wear stage of the tooth and its listriodont affinities. The talonid is simple and strongly joined to the posterior loph via the posterior accessory cusp. There are cingular remnants in the ends of the buccal valleys. The overall morphology of this tooth resembles that seen in bunodont listriodonts from Europe and Africa. The length and estimated breadth of M.42949 at the base of the crown are 32-8 and 19-5mm respectively. M.42950 is a right M® lacking only the roots and small fragments of enamel at the cervix level. Its enamel is thicker than is usually found in L. splendens and the tooth is more bunodont and lower-crowned, while the lophs are less well formed than in L. pentapotamiae. The ‘furchen’ are not clear, because they have apparently been eradicated by wear which has advanced to the stage where dentine is exposed on the two lingual cusps. The talon is simple and the lingual flare is marked. The length and breadth of M.42950 are 25-9 and 22:0 mm respectively. DISCUSSION. These two teeth probably belong to the same taxon, a bunodont listriodont close to Listriodon lockharti (Pomel, 1848) or L. akatikubas Wilkinson 1976. Teeth like these have Figs 49-50 M.42949 and M.42950, respectively lower and upper third molars of a bunodont Listrio- don. Ad Dabtiyah, Saudi Arabia. Fig. 51 M.42951, fragment of lower molar of a bunodont suid, possibly Kenyasus. Ad Dabtiyah, Saudi Arabia. Fig. 52 M.42952, fragment of unerupted and incompletely formed lower molar of a bunodont suid. Ad Dabtiyah, Saudi Arabia. SUIDAE 443 never been reliably recorded from Lower Miocene sites in Kenya, despite the reports of Wilkin- son (1976). Indeed, the subfamily Listriodontinae is unknown in Kenyan Lower Miocene deposits. The genus Bunolistriodon is a synonym of Listriodon (see Leinders 1975). In any case the material identified as Bunolistriodon by Arambourg (1963) and Wilkinson (1976) belongs to the subfamily Kubanochoerinae Gabunia, 1960 (Pickford 1985). The Ad Dabtiyah listriodont teeth do not resemble the species ?L. akatidogus Wilkinson, which is probably not a listriodont suid, but possibly a tayassuid (Pickford 1985). The only African species which resembles the Ad Dabtiyah sample is Listriodon akatikubas, which is known from Maboko, Nyakach, Kirimun and Fort Ternan, all Middle Miocene sites in Kenya. From the European viewpoint, the Ad Dabtiyah listriodont teeth resemble those of L. lockharti known from a number of lower middle Miocene sites in southern Europe such as La Romieu (France), and from Pasalar (Turkey). The teeth are considerably more lophodont than the most lophodont Orleanian suids of southern Europe, such as Palaeochoerus giganteus described by Golpe-Posse (1972). The Saudi Arabian teeth are considerably more bunolopho- dont than teeth assigned to L. splendens, which is characteristic of late Vindobonian localities of Europe. As far as is known, the genus Listriodon has at most only minimal overlap in time with Hipparion, its stratigraphical range being nearly totally limited to the Middle Miocene. From the Asian viewpoint the Saudi listriodonts compare with some of the teeth identified as Listriodon guptai from the Sind in Pakistan (Pilgrim 1926), although much of this material is more likely to represent L. pentapotamiae, a very lophodont form from Chinji levels in the Potwar Plateau. Listriodon affinis from Bugti is inadequately known, and the holotype may well represent a kubanochoere, (Pickford, in prep.). In any case it does not match the Saudi specimens. It therefore seems that in Asia Listriodon is confined to Middle Miocene deposits as it is in Africa and Europe. Taking everything into consideration, I consider it likely that the Ad Dabtiyah listriodonts indicate an age close to Mein Zone MN 4a or 4b or perhaps a little later (Mein 1977, 1979, 1985). This would correspond to middle Orleanian deposits. Subfamily ? KUBANOCHOERINAE Gabunia, 1958 Genus indet. Figs 51-52 MATERIAL. M.42951, less than half the crown of a left lower molar; M.42952, distal half of an incompletely formed left lower molar. DESCRIPTION. M.42951 is less than half of the crown of a left lower molar. The enamel is relatively thick, the cusps appear to be rounded and the ‘furchem’ are relatively shallow. There is a beaded cingular remnant in the buccal end of the median valley and the median accessory cusplet is small. These features suggest that we might be dealing with the genus Kenyasus Pickford (1985) or Conohyus Pilgrim. It is unlikely that the tooth belongs to Hyotherium von Meyer, although in view of the fragmentary nature of the specimen there must be room for doubt. The somewhat open labial notch is wider than is usually the case in Conohyus, which possibly tilts the balance in favour of this specimen representing a kubanochoerine such as Kenyasus. M.42592 is the distal half of an incompletely formed lower molar. Considering that it was not fully formed at the time of death of the individual, it is difficult to make a valuable statement about this tooth. It may perhaps represent the same taxon as the previously described frag- ment. DIscussION. I have seen a small kubanochoerine tooth in another collection from Saudi Arabia, which seems to be close in morphology to the Lower Miocene Kenyasus rusingensis. It is likely that this tooth and M.42591—2 described above belong to a single taxon. If these suppositions are correct, then it would follow that we are examining a suid which is usually Lower Miocene in age, although it extends up to lower Middle Miocene sites in Kenya such as Nachola. 444 M. PICKFORD In view of the fragmentary nature of these two specimens, I weigh the listriodont teeth as far more valuable from the point of view of biostratigraphy. Nevertheless, the presence of Kenyasus at Ad Dabtiyah, if it is eventually sustained, would not invalidate an age estimate of lower Middle Miocene, but would suggest that it was very early in that period. Kubanochoeres seem to have evolved in Africa, spreading into Eurasia during the Middle Miocene, where they have been recorded from Turkey, Georgia, China and India (Pickford 1985). It is possible that Ad Dabtiyah records the first of these emigrant kubanochoeres. Palaeozoogeography Ad Dabtiyah has yielded only a small sample of suids, but they are exceptionally interesting from the point of view of zoogeography. Suids often seem to be in the vanguard of emigrations. Along with gomphotheres, they seem to comprise a sort of ‘chef de file’, appearing in new regions in advance of many other taxa. They are therefore generally good indicators for changes in environment or for the opening up of migration pathways. At Ad Dabtiyah, it is possible that a crossing of the ways has been sampled, with the genus Listriodon making its way into Africa and the genus Kenyasus emigrating to southern Eurasia. I would be happier, however, if we had more definitive samples of the bunodont suid from the site before fully accepting this. Although we need not postulate the existence of “‘dryshod’ access across the Tethys to account for the migration of suids and proboscideans, it would appear that the Tethys had indeed closed by the time the Ad Dabtiyah sediments accumulated (Adams et al. 1983, Whybrow 1984, Thomas 1985). Closure of the Tethys would surely have had marked effects on regional and perhaps global climates as a result of changes in circulation of the Atlantic and Indian Oceans. It was at about the time of the postulated closure that widespread ‘heating’ events occurred in the western Tethys (Anon. 1984), Japan (the Kurosedani event, Karyu et al. 1984) and western Kenya (Pickford & Senut, in prep.). In west Kenya the regional climate, as inferred from fossil terrestrial gastropods, was humid and hot during the Lower Miocene (at Koru, Songhor and Rusinga) but changed dramatically by the time the Maboko sediments accumulated. At Maboko, the gastropods suggest that the region was a semi-desert with gallery forests fringing the rivers. Somewhat later, at Fort Ternan, cooler, wetter conditions were re-established. The Kurosedani event in Japan is characterized by the establishment of tropical to sub- tropical mangroves and associated mollusc assemblages in many parts of Japan, at least 1000km north of their closest living occurrences. Sea temperatures in Japan rose by 10°C (summer temperatures) or 20°C (winter temperatures). This event has been dated between 15-16 Ma (Karyu et al. 1984). It is tempting to ascribe these ‘heating’ events at the beginning of Langhian times, observed in three widely separated areas, to a single cause, occurring as they do at about the same time (as far as current datings indicate). If this is so, then a single major geological event such as closure of the Tethys might represent the fundamental root cause of such a global effect. Confirmatory evidence is afforded by the faunas, which underwent marked changes in many parts of the Old World at about the same time. The faunas of western Kenya underwent a major turnover between Rusinga (Lower Miocene) and Maboko (Middle Miocene) (Pickford 1981). West European faunas underwent comparable changes, which have been utilized in defining the boundary between Lower and Middle Miocene faunas in that part of the world (Mein 1979). Although the Asian evidence is not so well dated, it seems that similar changes in fauna may have occurred at the beginning of the Middle Miocene. For these widespread and apparently synchronous faunal changes to have occurred, it seems likely that tropical and sub-tropical zoogeographic conditions extended appreciably further north during the Middle Miocene than they did before. Even the suid evidence on its own supports this contention because at present, and at times during the past, suids have shown marked latitudinal stratification, with well-defined Palaearctic elements clearly distinguishable from their more southerly tropical counterparts (Pickford 1985, and in prep.). At Ad Dabtiyah SUIDAE 445 and penecontemporary sites we have what I consider reasonable evidence that the Ethiopian zoogeographic realm had spread northwards at the expense of the Palaearctic realm, and that it incorporated Saudi Arabia and much of southern Eurasia. In this respect the evidence of the suids from Ad Dabtiyah is most intriguing: clearly, however, better samples are required. In particular, the reports of giant kubanochoeres at Al-Sarrar (Thomas et al. 1982) may provide further support for the hypothesis that zoogeographic boundaries shifted northwards during the Middle Miocene. In conclusion, on the basis of the suids from Ad Dabtiyah, I see no reason to consider the locality as being Lower Miocene in age, unless one either is prepared to re-arrange the upper boundary of the Lower Miocene period to include Mein Zone 4a, or wishes to postulate that the genus Listriodon appeared substantially earlier at Ad Dabtiyah than it did in Europe, Africa or Asia. Acknowledgements I wish to thank Peter Whybrow for asking me to examine these interesting and significant specimens from Saudi Arabia. I would also like to thank Herbert Thomas for discussions about other fossil sites in Arabia. References Adams, C. G., Gentry, A. W. & Whybrow, P. J. 1983. Dating the terminal Tethyan event. Utrecht Micropalaeont. Bull., 30: 273-298. Anon., 1984. Compilation charts Neogene faunal and floral changes. Paleobiol. cont., Montpellier, 14 (2): 485-493. Arambourg, C. 1963. Le genre Bunolistriodon Arambourg, 1933. Bull. Soc. géol. Fr., Paris, (7) 5: 903-911. Drake, R., Van Couvering, J., Pickford, M. & Curtis, G. (in prep.). K-Ar geochronology of early Miocene volcanic strata and associated vertebrate and early homonoid fossil localities: Rusinga and Mfwangano Islands, Uyoma Peninsula and Karungu, western Kenya. Gabunia, L. 1958. O cherepe rogatoi iskopaemoi svinia iz srednego miotsena Caucasia. Dokl. Akad. Nauk SSSR, Moscow, 118 (6): 1187-1190. 1960. Kubanochoerinae, nouvelle sous-famille de porcs du miocene moyen du Caucase. Vertebr. palasiat., Peking, 4 (2): 87-97. Golpe-Posse, J. M. 1972. Suiformes del Terciario Espanol y sus yacimientos. Paleont. Evoluc., Sabadell, 2: 1-97. Karyu, T., Itoigawa, J. & Yamanoi, T. 1984. On the middle Miocene palaeoenvironment of Japan with special reference to the ancient mangrove swamps. In Whyte, R. O. (ed.), The Evolution of the East Asian Environment 1: 388-396. Hong Kong (Cent. Asian Stud.). Lienders, J. 1975. Sur les affinités des Listriodontinae bunodontes de Europe et de l’Afrique. Bull. Mus. natn Hist. nat. Paris, 341: 197-201. Mein, P. 1977. Adopted subdivision and correlation charts. In Alberdi, M. T. & Aguirre, E. (eds), Round- table on Mastostratigraphy of the west Mediterranean Neogene. Trab. Neogeno-Cuaternario, Madrid, 7: 21-23, tab. 1-3. 1979. Rapport d’activiteé du groupe de travail verteébrés mise a jour de la biostratigraphie du Neogéne basée sur les mammiféres. Annls géol. Pays hell., Athens, (Tome hors série) 3: 1367-1372. 1985. A new direct correlation between marine and continental scales in Rhodanian Miocene. Abstr. Congr. reg. Comm. Mediterr. Neogene Stratigr., Budapest, 8th: 377-379. (Hungarian Geol. Surv.). Pickford, M. 1981. Preliminary Miocene mammalian biostratigraphy for western Kenya. J. hum. Evol., London, 10: 73-97. 1985. A revision of the Miocene Suidae and Tayassuidae (Artiodactyla, Mammalia) of Africa. Tert. Res. spec. Pap., Leiden &c., 7: 1-83. (in prep.). A revision of the Suidae of the Indian Subcontinent. — & Senut, B. (in press). Habitat and Locomotion in middle Miocene Cercopithecoids. In: Gautier- Hion, A., Bourliére, F., Gautier, J.-P. & Kingdon, J. (eds), Evolutionary Biology of the African Guenons. Cambridge (C.U.P.). Pilgrim, G. 1926. The fossil Suidae of India. Mem. geol. Surv. India Palaeont. indica, Calcutta, 4: 1-68. Pomel, A. 1848. Observations paléontologiques sur les hippopotames et les cochons. Archs Sci. phys. nat., Genéve, 8: 155-162. 446 M. PICKFORD Thomas, H. 1985. The Early and Middle Miocene Land Connection of the Afro-Arabian Plate and Asia: A major event for Hominoid Dispersal? In Delson, E. (ed.), Ancestors: The Hard Evidence: 42-50. New York. ——., Sen, S., Khan, M., Battail, B. & Ligabue, G. 1982. The lower Miocene fauna of Al-Sarrar (Eastern Province, Saudi Arabia). ATLAL, JI Saudi Arab. Archaeol., Jeddah, 5: 109-136, pls 110-116. Whybrow, P. 1984. Geological and faunal evidence from Arabia for mammal ‘migrations’ between Asia and Africa during the early Miocene. Cour. ForschInst. Senckenberg, Frankfurt a.M., 69: 189-198. Wilkinson, A. 1976. The Lower Miocene Suidae of Africa. In Savage, R. J. G. & Coryndon, S. C. (eds), Fossil Vertebrates of Africa 4: 173-282. London & New York. A delphinoid ear bone from the Dam Formation (Miocene) of Saudi Arabia F. C. Whitmore, jr United States Geological Survey, National Museum of Natural History, Washington, D.C. 20560, U.S.A. Synopsis A right periotic of a dolphin-like cetacean was found in near-shore marine rocks of Burdigalian age in eastern Saudi Arabia. Its resemblance to ear bones of primitive Delphinidae suggests that a primitive delphinid, or an advanced kentriodontid, lived in the Tethys epicontinental sea in Burdigalian times. Introduction A single delphinoid.ear bone was collected by P. J. Whybrow in 1979 from the type locality of the Dam Formation at Jabal Lidam, 26° 21’ 42” N, 49° 27’ 42” E, eastern Saudi Arabia. It is the first cetacean fossil to be reported from Saudi Arabia. Fragmented ribs of sirenians were also found at the locality. Systematic palaeontology Order CETACEA Brisson, 1762 Suborder ODONTOCETI Flower, 1867 Superfamily DELPHINOIDEA Flower, 1864 Delphinoidea, gen. et sp. indet. Fig. 53 MATERIAL. A right periotic, M.42836. DESCRIPTION. Among the holdings of the United States National Museum of Natural History (USNM 258859), a periotic of the living species Sousa chinensis Osbeck 1765 from the Gulf of Siam, is most similar to the fossil. The pars cochlearis is similarly prominent (transverse measurement from apex of pars cochlearis to posterior border of anterior process = 20-2 mm in M.42836 and 21:2mm in USNM 258859), although in the fossil, its dorsal side is slightly more bulbous than in S. chinensis. In both the fossil and S. chinensis, the foramen singulare and the internal aperture of the Fallopian aqueduct are included in the depression of the internal auditory meatus (Fig. 53a); this depression is slit-like, is widest in the area of the internal auditory meatus, and is orientated at an angle of about 45° to the transverse axis of the pars cochlearis. It is separated from the triangular hollow surrounding the opening of the endolym- phatic duct by a keel on its lateral margin (see Kasuya 1973: 34, fig. 65). In both the fossil and S. chinensis, the anterior process is only slightly elongated, is turned only slightly in a medial direction, and has a rectangular end (Fig. 53b); the posterior process, seen laterally (Fig. 53c), is in the same plane as the anterior process; its articular surface in both specimens bears fine grooves where it was attached to the tympanic (Fig. 53b). The fossil periotic is 28-75mm long; its pars cochlearis is 16-2mm long at the base. The ratio between these measurements reflects the shortness of the anterior and posterior processes. In both specimens, the superior process is divided into lateral and dorsal planes by a longitudinal keel. Barnes (1978) raised the Kentriodontinae of Slijper 1936, a subfamily of the Delphinidae, to family rank as the Kentriodontidae, a family of middle and late Miocene delphinoid cetaceans which are more primitive than the Delphinidae. The Arabian specimen differs from the periotics of species of Kentriodontidae in having derived characters typical of Delphinidae. It differs from Kentriodon, Liolithax kernensis Kellogg 1931, and Delphinodon dividum True 1912 in Bull. Br. Mus. nat. Hist. (Geol.) 41 (4): 447-450 Issued 29 October 1987 448 F. C. WHITMORE “ZS p.p. Fig. 53 Delphinoidea, genus and species indeterminate. Right periotic, BM(NH) M.42836. a, cere- bral (dorsal) view; b, ventral view; c, lateral view. x 2. Key: a.p., anterior process; aq.F., opening of Fallopian aqueduct; d.e., opening of endolymphatic duct; f.s., foramen singulare; m.a.i., internal auditory meatus; p.c., pars cochlearis; p.p., posterior process; s.p., superior process. From Dam Formation, Jabal Lidam, Saudi Arabia. Burdigalian. having the anterior and posterior processes in the same horizontal plane, and in having a flat superior process, grooved articular surface of the posterior process, and relatively short anterior process. From Liolithax pappus (Kellogg 1955) it differs in having a relatively short anterior process and relatively larger pars cochlearis. It differs from Kentriodon, Lophocetus calvertensis (Harlan 1842), and Delphinodon dividum in having a slit-like depression for the internal auditory meatus, which also contains the foramen singulare and the internal aperture of the Fallopian aqueduct. The Arabian specimen was also compared with a periotic from Lee Creek, North Carolina (USNM 183001), probably from the Pungo River Formation (early and middle Miocene), that was identified by L. G. Barnes (oral communication, September 17, 1975) as identical to the periotics accompanying a skull of Pithanodelphis from the late Miocene of California. In the shape of its anterior and posterior processes, this Pithanodelphis specimen closely resembles the Arabian specimen and the periotics of Delphinidae; however, its pars cochlearis is smaller than that of the Arabian periotic. The Arabian specimen also closely resembles a periotic (USNM 317874) of an undescribed ondontocete from the Pungo River Formation of North Carolina. The major difference between the two specimens is that the periotic from North Carolina has a smaller pars cochlearis. ; Although they are smaller, two unidentified periotics (UCMP 88582 and UCMP 88583) from the San Diego Formation (Pliocene) of California resemble the Arabian specimen in the slit-like internal auditory meatus, in the shape of the pars cochlearis and anterior process, and in the articular surface of the posterior process being in the same plane as the anterior process (see Barnes 1973: fig. 2g—4). DELPHINOID EAR BONE 449 Discussion The single cetacean periotic from the Dam Formation possesses characters typical of primitive Delphinidae, and possibly of advanced Kentriodontidae. Its morphology is closest to that of the living genus Sousa, the humpbacked dolphin, now living in coastal waters and the mouths of rivers from the South China Sea west through the Straits of Malacca, the Bay of Bengal, and the Arabian and Red Seas to the Suez Canal, and in waters off South Africa and west Africa (Hershkovitz 1966: 18-25). However, definite generic assignment can be made only in the basis of the skull. The only conclusion that can be reached from study of a single ear bone is that a possible close relative of Sousa is present in the Dam Formation. Sousa itself has never been reported as a fossil. Remington Kellogg, in a letter to D. W. Rice (December 4, 1962: Smithsonian Institution Archives, Record Unit 88, Box 6), pointed out the resemblance between the periotics of the Miocene genus Kentriodon and the living New World freshwater porpoise Sotalia. He wrote: The fresh water porpoises of the genus Sotalia all possess periotic bones with similar characteristics. If you will refer to the following article... [Kellogg 1927] .. . you will find illustrations . . . [pl. 3, figs 2-4]... of this fossil porpoise which resemble those of Sotalia rather closely. The configuration of the cerebral surface and the shape of the internal acoustic meatus is similar in both. Kentriodon which was present in the Miocene period in the Chesapeake embayment may have been an antecedent of the fresh water porpoise Sotalia. ... Dr Fraser and I have not come as yet to any final conclusion as to the family allocation of Sotalia. Relatively few specimens have been received by museums. On the basis of present information it would appear that Sotalia may possibly be somewhat closely related to Steno and Sousa, but in my opinion this allocation should be deferred until more adequate information is available. The family Kentriodontidae, as defined by Barnes (1978), reflects a middle to late Miocene delphinoid radiation of animals that were more primitive than, but in part contemporaneous with, members of the more advanced family Delphinidae. True (1912) favourably compared Delphinodon dividum True 1912 with Pithanodelphis Abel 1905 of the Miocene, and with living Steno and Sotalia. Barnes (1978) placed Delphinodon and Pithanodelphis with Kentriodon in the Kentriodontidae; I include Sousa with Steno and Sotalia as structurally primitive living Delphinidae. The periotic from the Dam Formation is delphinoid in that the articular surface for the bulla on the posterior process is in the same horizontal plane as the ventral side of the anterior process and, concomitant with this, the superior process is low and flat. The Arabian periotic has a longer, straighter anterior process than do those of advanced Delphinidae, whose anterior process is directed medially and is partly appressed against the anterior side of the pars cochlearis. This combination of characters is probably a morphological stage between the relatively primitive periotics of the Kentriodontidae and the derived condition in the Delphinidae. If the Arabian periotic is accepted as representing a species in the Delphinidae, this record extends the range of the Delphinidae farther back in time than has previously been reported. Barnes (1976: 330, tab. 4; fig. 2) reported a late Miocene species of Delphinidae sensu stricto, known from a complete skull from California. This specimen is at least ten million years old, but even this is at least five million years younger than the specimen from the Dam Formation. On the slim evidence that we have, familial assignment of the Arabian periotic to the Delphin- idae or to the Kentriodontidae must await collection of more material. Tethyan Distribution of Miocene Delphinoidea In Burdigalian time the area that is now eastern Saudi Arabia was separated from the ancestral Mediterranean by an evaporite realm that formed a land bridge between Asia and Africa (Steininger et al. 1985). Earlier in the Miocene the land area was occupied by a strait allowing access by its marine fauna to the western Tethys. The marine mammal fauna represented by the specimen from the Dam Formation could, therefore, have been related to forms from farther 450 F. C. WHITMORE west in the Tethys Sea. Unfortunately, only one penecontemporary delphinoid, ‘Delphinus’ vanzelleri Fourtau 1918, is known from the Mediterranean Tethys. This species, represented only by a partial jaw from the Lower Miocene Moghara Formation of Egypt, is probably generically unidentifiable (Barnes & Mitchell 1978) and cannot be compared to the Saudi Arabian specimen. Five genera of Delphinoidea are known from the late Miocene (Sarmatian) of the Caucasus. Three of these, Leptodelphis Kirpichnikov 1954, Sarmatodelphis Kirpichnikov 1954, and Micro- phocaena Kudrin & Tatarinov 1965, have been placed in the Kentriodontidae by Barnes (1978). The other two, Anacharsis Bogachev 1956 and Imerodelphis Mchedlidze 1959, are tentatively assigned to the Delphinidae. These genera have not been identified elsewhere, and it is possible that they were endemic to Paratethys. The resemblance, pointed out above, of the Saudi Arabian specimen to isolated periotics from North Carolina and California may indicate distribution of related Delphinoidea throughout the Tethys in Miocene time. Conclusions A primitive delphinoid, similar to and perhaps related to Sousa, was present in the ancestral Arabian Gulf in Burdigalian time. Confirmation of the taxonomic position of this cetacean must await collection of more material from the Dam Formation. Despite this taxonomic uncertainty, the morphology of the periotic makes it clear that we have here another bit of evidence of the radiation of the earliest modernized dolphins. Similar bones (unfortunately usually unaccompanied by skulls) from the middle Miocene to lower Pliocene of North Carolina may indicate a Tethyan distribution of related primitive delphinoids. A continuous range of these forms from the Tethys to the west coast of North America would have been possible because of the existence of the Panama seaway (see Whitmore & Stewart 1965). Acknowledgements I thank Peter J. Whybrow of the British Museum (Natural History) for making the specimen available for study. Cary T. Madden, then of the U.S. Geological Survey, encouraged discussion of the Miocene vertebrate faunas: of Saudi Arabia among the specialists working in different parts of the Kingdom. I am grateful to Lawrence F. Barnes and Samuel A. McLeod for reviewing the manuscript. The photographs were taken by Robert H. McKinney of the U.S. Geological Survey. References Barnes, L. G. 1973. Pliocene cetaceans of the San Diego Formation, San Diego, California. In Ross, A. & Dowlen, R. J. (eds), Studies in the Geology and Geologic Hazards of the Greater San Diego Area, California: 37-42. San Diego, Calif. 1976. Outline of eastern North Pacific fossil cetacean assemblages. Syst. Zool., Washington, 25: 321-343. 1978. A review of Lophocetus and Liolithax and their relationships to the delphinoid family Ken- triodontidae (Cetacea: Odontoceti). Sci. Bull. nat. Hist. Mus. Los Ang. Cty, 28: 1-35. & Mitchell, E. D. 1978. Cetacea. In Maglio, V. J. & Cooke, H. B. S. (eds), Evolution of African Mammals: 582-602. Cambridge, Mass. (Harvard U.P.). Hershkovitz, P. 1966. Catalog of living whales. Bull. U.S. natn. Mus., Washington, 246: 1—259. Kasuya, T. 1973. Systematic consideration of recent toothed whales based on the morphology of tympano-periotic bone. Scient. Rep. Whales Res. Inst., Tokyo, 25: 1-103. Kellogg, R. 1927. Kentriodon pernix, a Miocene porpoise from Maryland. Proc. U.S. natn. Mus., Washing- ton, 69 (19): 1-55, 14 pls. Steininger, F. F., Rabeder, G. & Rogl, F. 1985. Land mammal distribution in the Mediterranean Neogene: a consequence of geokinematic and climatic events. In Stanley, D. J. & Wezel, F.-C. (eds), Geological Evolution of the Mediterranean Basin: 559-571. New York, &c. True, F. W. 1912. Description of a new fossil porpoise of the genus Delphinodon from the Miocene formation of Maryland. J. Acad. nat. Sci. Philad., 15 (2): 165-194. Whitmore, F. C. & Stewart, R. H. 1965. Miocene mammals and Central American seaways. Science, N.Y., 148: 180-185. Early Miocene fish from eastern Saudi Arabia P. H. Greenwood, F.R.S. Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Synopsis Fish remains collected from early Miocene (Burdigalian marine chronology, middle Orleanian European land-mammal age equivalent, c. 17-19 Ma) continental deposits at Ad Dabtiyah, Saudi Arabia, can be referred to the family Cyprinidae and, possibly, to the Centropomidae. Teeth resembling those found in certain members of the family Labridae are also present. Introduction The fish remains were recovered from a continental white, pebbly, calcareous grit thought to be laterally equivalent to the basal deposits of the nearby marine Dam Formation (see Whybrow, McClure & Elliott, this issue, p. 371). In a preliminary report on the vertebrate fauna from the Saudi Arabian Miocene (Hamilton et al. 1978: table 1), the fish described here were erroneously stated to have been collected from the Jabal Midra ash-Shamali locality. Register numbers refer to the collections of the Department of Palaeontology, British Museum (Natural History), London. Systematic description Subclass OSTEICHTHYES Huxley, 1880 Superorder OSTARIOPHYSI Sagemehl, 1885 Family CYPRINIDAE Bonaparte, 1837 Fig. 54 The Cyprinidae is a family of essentially freshwater species, a very few of which have a limited tolerance of brackish water. The family is represented by a single, well-preserved pharyngeal tooth, P.61555. Judged from the nature of its basal attachment area and from the unworn occlusal surface, it was probably an unattached replacement tooth. The ornate cusp pattern on the occlusal surface is most unusual (Fig. 54) and cannot be matched with that from any extant species represented in the Zoological collections of the British Museum (Natural History), nor with any species which has been described in the literature. Thus it is impossible to suggest an infrafamilial taxon to which the fossil might be allied. Gayet (in Thomas et al. 1982: 116) identified cyprinid pharyngeal teeth from Lower Miocene deposits of the Al-Sarrar region (Eastern Province, Saudi Arabia) as derived from fishes refer- able to the genera Barbus and Labeo, with certain of those from the former taxon closely resembling median row teeth of Barbus bynni, an extant Nilotic species. The presence of Barbus and Labeo in Miocene deposits of Saudi Arabia supports, according to Gayet, Menon’s (1964) suggestion that these fishes ‘. . . entered Africa (from Asia) as late as Plio-Pleistocene times’. The presence of a presumed Barbus from the Upper Miocene of Tunisia (Greenwood 1972), however, would appear to weaken that argument. None of the cyprinid fossil evidence available so far would seem to indicate, as Gayet suggested, either the routes or the dispersal pattern of the family, if indeed there was dispersal on this scale. Those questions are more likely to be elucidated by evidence ultimately obtained from a detailed phylogenetic analysis of living taxa. For the moment it would seem preferable only to use the fossil record as indicative of cyprinids occurring in both Africa and Saudi Arabia during the Miocene. Bull. Br. Mus. nat. Hist. (Geol.) 41 (4): 451-453 Issued 29 October 1987 452 P. H. GREENWOOD Siete | if tau \ Fig. 54 Lower pharyngeal tooth from a cyprinid fish in lateral (left) and anterior (right) views. Bar =1mm. Early Miocene, Ad Dabtiyah, —— Saudi Arabia. Superorder ACANTHOPTERYGII Gouan, 1770 Group PERCOMORPHA Bleeker, 1859 Family ?>CENTROPOMIDAE Poey, 1865 An extensively damaged and incomplete neurocranium, P.61556, is referred tentatively to the Centropomidae, on the basis of its overall proportions taken in combination with the shape and proportions of individual skull bones. The skull is elongate and slender and is noticeably narrow across the otico-occipital region. The supraoccipital is long-based and protracted anteriorly, thereby entirely and widely separat- ing the elongate parietals, and also, but more narrowly, the frontals over the greater part of their medial faces. The exoccipital is an elongate and expansive bone with the foramen for the occipitospinal nerve situated immediately above the buttress extending upwards from the exoccipital facet. In all these features the morphology of the fossil skull resembles that in extant Lates species (the African and Indo-Pacific, predominantly freshwater, representatives of the family) more closely than it does the neurocranial form in any Centropomus species (the American, and predominantly marine—brackish water representatives) or in Psammoperca waigiensis Cuvier & Valenciennes, a coastal marine centropomid of the Indo-Pacific area; see Greenwood (1976) for details and figures. It differs, however, in the great forward extension of the blade-like anterior supraoccipital prolongation. In no extant centropomid species does the supraoccipital separate more than the posterior third to quarter of the frontals. Likewise, in none of the serranid or percichthyids (the other presumed basal percoids) I have examined does the supraoccipital extend forward beyond that level. The bones present and recognizable in the fossil are: the left and right frontals in part (those portions medial to the frontoparietal crests), the supraoccipital (missing its crest and the ventral portion of its posterior median extension), the parietals in part (the bone medial to the crest), the greater part of the right epioccipital, virtually the entire right exoccipital, the posterior region of the right prootic (but lacking its ventral margin), part of the dermethmoid, and the proximal part of the right lateral ethmoid. Based on comparisons with skulls from Lates calcarifer (Bloch) and Lates niloticus (L.), it is estimated that the fossil skull, when entire, measured c. 10cm from the anterior tip of the vomer to the ventral rim of the basioccipital condyle. It would thus be derived from a fish of about 30-35 cm standard length. EARLY MIOCENE FISH 453 Part of a right dentary, P.61557 (that region of the bone immediately anterior to its separa- tion into an ascending coronoid and a horizontal lower limb), is also thought to be derived from a centropomid fish. In its gross morphology, and in having small tooth scars densely packed into a broad alveolar surface, this dentary closely resembles that found in extant Lates species. It is probably from a fish of about the same size as that from which the skull was derived. When compared with the dentary in specimens of extant Lates species of comparable size, the fossil is more robust and the alveolar surface is, relatively, somewhat wider. The distal end of a proximal dorsal fin radial (pterygiophore), P.61558, compares closely with the third pterygiophore in the anterior dorsal fin of Lates. It is difficult to estimate the size of the individual from which the fossil came, but certainly it would not have been of more than 35—40 cm standard length. A fragment from the proximal end of a spinous fin ray, P.61559, could also be derived from a centropomid fish within the size range 30-35 cm standard length. Many extant species of the Centropomidae are known to be euryhaline; a few species are apparently confined to marine habitats, and some of the African species are apparently con- fined to fresh waters, with a number physiographically restricted to lakes. Remains of a Lates species are recorded by Gayet (in Thomas et al. 1982: 116) from the Lower Miocene of the Al-Sarrar region, Eastern Province, Saudi Arabia. Material of indeterminate origin Two small teeth, P.61560, with molariform crowns, cannot be referred with certainty to any taxon. In their gross morphology they resemble a type of pharyngeal tooth frequently found in members of the Labridae, a family whose living members are exclusively marine. Superficially these teeth do bear some resemblance to those from the parasphenoidal and basihyal tooth plates in members of the mormyrid genus Hyperopisus. However, on closer examination of the gross morphology, overall proportions and details of wear pattern, they are quite unlike those found in extant Hyperopisus specimens. One tooth is squat, its maximum height (2mm) is equal to the greatest diameter of its subcircular crown. The occlusal surface is slightly convex, with a well-defined wear facet occu- pying about one-third of its area; the entire surface slopes gently into the plane of its greatest diameter. The other tooth is more elongate, its height measured from the crown to the lowest point of the neck preserved is 4.5mm, and the maximum and minimum crown diameters are 2-‘0mm and 1-5mm respectively. Almost the entire occlusal surface is worn, the wear plane sloping to one side across the narrow diameter of the crown. The third specimen of indeterminable origin, P.61561, is a small fragment of bone which resembles the head region of a percoid pleural rib. References Greenwood, P. H. 1972. Fish fossils from the late Miocene of Tunisia. Notes Serv. geol. Tunis, 37 (Trav. Géol. Tunisienne 6): 41-72. 1976. A review of the family Centropomidae (Pisces, Perciformes). Bull. Br. Mus. nat. Hist., London, (Zool.) 29: 1-81. Hamilton, W. R., Whybrow, P. J. & McClure, H. A. 1978. Fauna of fossil mammals from the Miocene of Saudi Arabia. Nature, Lond., 274: 248-249. Menon, A. G. K. 1964. Monograph of the cyprinid fishes of the genus Garra Hamilton. Mem. Indian Mus., Calcutta, 14: 173-260. Thomas, H., Sen, S., Khan, M., Battail, B. & Ligabue, G. 1982. The Lower Miocene fauna of Al-Sarrar (Eastern Province, Saudi Arabia). ATLAL, JI Saudi Arab. Archaeol., Jeddah, 5: 109-136, pls 110-116. *) mt Ae ‘ be 4 | 1 ~ = **Gp J ‘ > » es | S| 2 a) a 4 ~ ~ if Eo 2 : \ =< 5 te -, Zips o - : rT) ‘ J wit - : ¥ ay 6S iT — = ; 7 tr : j oe is A ; an i _—__ ie > - ery se wa é aa ? > ppg _ vy # m4 Fe > - xb 2 ea m Oo z ise Bb op PS r TP vary all ie aL Ate 3 oo Tee. Index Aceratherium 409, 414, 416-18, 421, 424, 426, 430 ‘Aceratherium’ abeli 415 Aceratherium campbelli 430 incisivum 424 sp. 368 Aegyptopithecus 388 Aetobatus arcuatus 369 Afropithecus 392 turkanensis 391 Ahwaz Sandstone 373 Alengerr 425 Al Jadidah 368, 373, 375 Allier 438 Al-Sarrar, see As Sarrar ‘Amphicyon’ shahbazi 425 Anacharsis 450 Anadara 371 Anomia 371 Arabosminthus quadratus 369, 372 Archaeobelodon 398, 402 Arrisdrift 435, 437 Artenay 398, 403-5, 437 As Sarrar (Al-Sarrar) 368, 434-8, 445, 451, 453 Atlantoxerus 368 Barbus 451 bynni 451 sp. 369 Beliajevina 415 bioherms 380-1 Bou Hanifa 421 Brachypotherium 409, 414, 416, 418, 421, 424, 426, 428-30 heinzelini 428 lemanense 418 lewisi 430 snowi 430 Bugti, see Dera Bugti Bunolistriodon 443 Caementodon 415 Canthumer yx 433-4, 438 sirtensis 435 Caprotragoides 368, 373 potwaricus 438; aff. potwaricus 368 Carcharhinus aff. plumbeus 369 aff. priscus 369 Cardium 371 Centropomus 452 Ceratotherium simum 429 Chalicotherium grande 428 Chilotheridium 426, 428, 430 pattersoni 415 Chilotherium 426 blandfordi 415 Chinji 443; Formation 398, 402, 404, 416 Choerolophodon 398, 402 Clarias sp. 369 Clementia 371 Conohyus 441, 443 Crocodylus pigotti 381; cf. pigotti 368 Dam Formation 367-8, 373-5, 377, 381, 395, 433, 436, 438, 447, 449-51 Dasyatus sp. 369 Deinotherium, cf. 368 Delphinodon dividum 447-9 ‘Delphinus’ vanzelleri 450 Dendropithecus 388 Dera Bugti (Bugti) 367, 395, 398, 403-6, 414, 416, 425, 435, 438, 443 Diceratherium 409, 426 asphaltense 418 pleuroceros 414, 418, 421, 425 Dicerorhinus 368, 373, 409, 419, 421, 424-6, 428, 438 abeli 421, 425 leakeyi 415, 421, 423-5, 428 orientalis 423-4 primaevus 421, 425; cf. primaevus 368 sansaniensis 415, 421, 425; aff. sansaniensis 368, 409 schleiermachi 415, 421, 423-4, 426 steinheimensis 415, 421 sumatrensis 419, 429 ‘Dicerorhinus tagicus’ 414, 418 Diceros 428 bicornis 419, 429 pachygnathus 415, 419, 423-4 Didermocerus 414 Dorcatherium 433-4, 438 chappuisi 434, 438 libiensis 433; cf. libiensis 368 pigotti 433-4 ‘Dremotherium’ 438 Dryopithecus 383, 387-8, 390 Echinocyamus 374 Elasmotherium 415 En Péjouan 402 Eotragus 433, 435-7 Eozygodon 398 Eppelsheim 416, 421, 423-4 Eryx/Gongylophis sp. 369 Fibularia damensis 374, 379 Fort Ternan 385, 415, 436, 438 Gaindatherium 416 Galeocerda cf. aduncus 369 Gazella 437-8 Gebel Zelten, see Jebel Zelten 456 ‘Gelocus ?gajensis’ 438 Geochelone sp. 369 Ghar Formation 373 giraffoid 380 Giraffokeryx 435 Gomphotherium 398, 400, 402, 405, 438 angustidens 368, 373, 398, 402-4, 406 browni 398, 402, 404-5 cooperi 367, 395, 402, 404 inopinatus 402 sp. 368 Gongylophis, see Eryx Hadrukh Formation 367, 369, 371, 373-4, 377, 379, 436 Heliopithecus leakeyi gen. et sp. nov. 368, 383-4 Hemipristis serra 369 Heterobranchus sp. 369 Hipparion 443; faunas 398, 415 Hippopotamus amphibius 429 Hispanotherium 409, 415, 417-18, 426 Hofuf Formation 367-8, 373, 375, 377, 438 Homo 390 Homoiodorcas?, cf. 368 Hyemoschus aquaticus 434 Hyotherium 443 Hyperopisus 453; sp. 369 Imerodelphis 450 Jabal Lidam 447 Jabal Midra ash-Shamali 369, 372, 436, 451 Japan 441, 444 Jaub Anbak 371 Jebel Zelten (Gebel Zelten) 416, 430, 433, 435, 437 Kalodirr 391 Kamlial Formation 430 Karungu 441 Kentriodon 447-9 Kenyapithecus 383, 387-8, 390, 392 africanus 383, 388 Kenyasus rusingensis 443 Kirimun 443 Kisingiri 441 Koru 444 La Grive 402 La Romieu 443 Labeo 451 Lake Turkana 391 Lates 453 calcarifer 452 niloticus 452 sp. 369 Leptodelphis 450 Limnopithecus 388 Liolithax kernensis 447 pappus 447 INDEX Lisbon 404 Listriodon 443-5 affinis 443 akatikubus 441-2 guptai 443 lockharti 441-2 pentapotamiae 441-3 splendens 442 sp. 368 Lophocetus calvertensis 448 Lopholistriodon 368, 373 Lothagam 430 Lower Siwaliks 373 Maboko 400, 402, 404, 406, 435-6, 438, 443-4 Majiwa 388 Manchar Formation 438 Martes, cf. 368 ‘Mastodon pyrenaicus’ 406 Megapedetes cf. pentadactylus 368 Metasayimys intermedius 368, 373 Micropithecus 388 Microphocaena 450 Midravalva arabica 372 Mionictis sp. 368 Moghara 430; Formation 450 Moropus 428 Moroto 387, 392 Moruorot 430, 435 Mycteria cinereus 368 Myliobatis sp. 369 Mytilus 371 Nachola 443 Naja/Palaeonaja spp. 369 Nari Formation 398, 438 Negaprion eurybathrodon 369 Nyakach 443 Oiceros sp., cf. 369, 436 Ostrea latimarginata 371, 373 Pachyhyrax aff. championi 368 Pachytragus 368, 373 ligabuei 368 Palaeochoerus giganteus 443 Palaeomeryx 435 Palaeonaja, see Naja Palaeotragus 435 roueni 435 sp. 368 Paradiceros 409 mukirii 415, 419 Paraphiomys sp. 368 Paratetralophodon 398 Pasalar 443 Percrocuta 368, 373 Pikermi 415, 423-4, 428 Pithanodelphis 448-9 Platybelodon 398, 400, 402 Pongo 390 Primelephas 398 proboscidean 380 Proconsul 383, 388 africanus 385 major 387 nyanzae 385 Progiraffa exigua 438, 453 Propliopithecus 387 Protaceratherium minutum 414 Protalactaga, cf. 368 Protanancus 398, 402, 404-5 chinjiensis 402 macinnesi 400, 402, 404 Protragocerus 368, 373, 435, 438 gluten 435 labidotus 435, 437-8 Psammoperca waigiensis 452 Pseudaelurus turnauensis 368 Pseudocyonopsis 425 Pungo River Formation, U.S.A. 448 Python sp. 369 Qarn Abu Wayil 371 Qatar 374-5 Ramapithecus 383 Rangwapithecus 388 Rhinoceros 409, 419 browni 416, 419, 421, 425-6 ‘Rhinoceros (Diceratherium) asphaltense’ 418 Rhinoceros sondaicus 419, 429 unicornis 419 rhinoceros 380 Rhinoptera 369 Ronzon 414 Ronzotherium 409, 414, 416, 418, 426 filholi 418, 423 Rusinga 385, 402, 404-6, 424, 428-9, 441 San Diego Formation, U.S.A. 448 INDEX Sansan 398, 402, 404-5, 424-5, 435 Sarmatodelphis 450 Schweboemys, cf. 369 Scoliodon sp. 369 Shamalina tuberculata 369, 372-3 silicified wood 380 Simorre 398 Sind 443 Sivapithecus 383, 388, 390 Siwaliks Group 373, 416, 430, 433, 435 Songhor 444 Sotalia 449 Sousa 449-50 chinensis 447 Sphyraena sp. 369 Stegodon 398 Stegolophodon 398 Steno 449 Sterogenys, aff. 369 stromatolitic bioherms 380-1 Teleoceras 429 Ternan; see Fort Ternan Tethys 367, 373, 428, 441, 444, 447, 449-50 Tetralophodon 398 longirostris 406 tragulid 380 Trilophodon cooperi 395 Turkana, Lake 391 Turkanatherium acutirostratus 430 Uyoma 441 Viverra sp. 368 Walangania 436-8 africanus 435, 438 wood 380 5 Zelten, see Jebel Zelten | PALAL Zygolophodon 398, 402 457 Accepted for publication 3 January 1986 = Sug ; —, ee al ek ' 4 ' ~ 5 : yo ye Se Te i 7h Oud Aion ee ‘, ; _ —— a | = : ; (ie iad { t = — * i is ¥, S i a —- | 4 fe wt —_ ~ alll ry ‘. s 9 ; > =pten & + cash 4s Oe 4 pain 7; eis } » rt oo —— 1 oo 2. q PR) oteoay U id 7 4 ~ j o a R q Wie io ¢ = ‘ es ~ | _ _ = - —— = ; Ls! X A = vs g @ ‘qs _ : ' ‘ * ~ ie oe — ope a in i i { 2 i a Pa, : £ yn : yi L. 7) - i Bulletin of the British Museum (Natural History) Geology Series Most earlier Geology Bulletins are still in print. A full list of available titles can be obtained from Publications Sales (address inside front cover). Vol. 29 No. 1 Aspects of mid-Cretaceous stratigraphical micropalaeontology. D. J. Carter & M. B. Hart. 1977. Pp. 1-135, 4 plates, 53 figs. £14.25 Vol. 29 No.2 The Macrosemiidae, a Mesozoic family of holostean fishes. A. W. H. Bartram. 1977. Pp. 137-234, 4 plates, 53 figs. £10.00 Vol. 29 No.3 The stratigraphy and ammonite fauna of the Upper Lias of Northamptonshire. M. K. Howarth. 1978. Pp. 235-288, 9 plates, 5 figs. £6.00 Vol. 29 No. 4 Fossil Bovidae (Mammalia) of Olduvai Gorge, Tanzania, Part I. A. W. Gentry & A. Gentry. 1978. Pp. 289-446, 41 plates, 34 figs. £17.50 Vol. 30 No. 1 Fossil Bovidae (Mammalia) of Olduvai Gorge, Tanzania. Part II. A. W. Gentry & A. Gentry. 1978. Pp. 1-83, 3 figs. £7.50 Vol. 30 No.2 A revision of the Miocene Hominoidea of East Africa. P. J. Andrews. 1978. Pp. 85-224, 7 plates, 29 figs. £15.30 Vol. 30 No. 3. Early Ordovician (Arenig) stratigraphy and faunas of the Carmarthen district, south-west Wales. R. A. Fortey & R. M. Owens. 1978. Pp. 225-296, 11 plates, 12 figs. £9.60 Vol. 30 No. 4 Macroscopic inclusions of fluid in British fluorites from the mineral collection of the British Museum (Natural History). A. H. Rankin. 1978. Pp. 297-307, coloured frontispiece, 9 plates (7 coloured), 4 figs. £12.00 Vol. 31 No. 1 Foraminifera of the Togopi Formation, eastern Sabah, Malaysia. J. E. Whittaker & R. L. Hodgkinson. 1979, Pp. 1-120, 10 plates, 71 figs. £14.00 Vol. 31 No.2 Cretaceous faunas from Zululand and Natal, South Africa. The ammonite family Gaudryceratidae. W. J. Kennedy & H. C. Klinger. 1979. Pp. 121-173. £6.25 Vol. 31 No. 3 Benthic community organization in the Ludlow Series of the Welsh Borderland. R. Watkins. 1979. Pp. 175-279. - £12.25 Vol. 31 No.4 The ammonites of the English Chalk Rock (Upper Turonian). C. W. Wright. 1979. Pp. 281-330. £6.50 Vol. 32 No.1 Miscellanea: Observations on Cycloclypeus—Provenance of Sivapithecus—Iranian Silurian brachiopods—New English condylarths—Miocene sharks’ teeth—East African isopod—The Singa skull— Carboniferous insects. 1979. Pp. 1-90. £10.50 Vol. 32 No.2 Palaeoenvironments and correlations of the Carboniferous rocks in west Fermanagh, Ireland. C. H. C. Brunton & T. R. Mason. 1979. Pp. 91-108, 6 figs, folded map. £4.00 Vol. 32 No.3 The Ordovician trilobite faunas of the Builth—Llandrindod Inlier, central Wales. Part III. C. P. Hughes. 1979. Pp. 109-181, 177 figs. £10.00 Vol. 32 No.4 The stratigraphy and brachiopods of the upper part of the type Caradoc of south Salop. J. M. Hurst. 1979. Pp. 183-304, 557 figs. £18.50 Vol. 33 No.1 An account of the Ordovician rocks of the Shelve Inlier in west Salop and part of north Powys. W. F. Whittard, F. R. S. (Compiled by W. T. Dean). 1979. Pp. 1-69, 38 figs, frontispiece, coloured map, folded, in pocket. (Map available separately for £1.00) £10.00 Vol. 33 No.2 Miscellanea: Lower Carboniferous microproblematicum—Miniature trilobite—Pleistocene bird remains—English Eocene Hyracotherium—Salenia trisuranalis—Antarctic brachiopods—Diphyphyllum and Murchi- son’s Russian corals—Lebanese amber Neuroptera. 1980. Pp. 71-164. £12.00 Vol. 33 No.3 The Caradoc faunal associations of the area between Bala and Dinas Mawddwy, north Wales. M. G. Lockley. 1980. Pp. 165-235, 105 figs. £9.00 Vol. 33 No. 4 Fossil insects from the Bembridge Marls, Palaeogene of the Isle of Wight, southern England. E. A. Jarzembowski. 1980. Pp. 237-293, 77 figs. £7.50 Vol. 33 No.5 The Yorkshire Jurassic fern Phlebopteris braunii (Goeppert) and its reference to Matonia R.Br. T. M. Harris. 1980. Pp. 295-311, 20 figs. £2.75 Vol. 34 No.1 Relative dating of the fossil hominids of Europe. K. P. Oakley. 1980. Pp. 1-63, 6 figs, 17 tables. £8.00 Vol. 34 No.2 Origin, evolution and systematics of the dwarf Acanthoceratid Protacanthoceras Spath, 1923 (Cretaceous Ammonoidea). C. W. Wright & W. J. Kennedy. 1980. Pp. 65-107, 61 figs. £6.25 Vol. 34 No.3 Ashgill Brachiopoda from the Glyn Ceiriog District, north Wales. N. Hiller. 1980. Pp. 109-216, 408 fi £14.75 gs. Vol. 34 No.4 Miscellanea: Upper Palaeozoic Athyrididae brachiopods—New British Cretaceous Epitontidae— Microproblematicum Prethocoprolithus—Glabellar ‘structure of asaphid trilobites—New Lower Ordovician bivalve family—Cretaceous brachiopods—Tupus diluculum sp. nov—Revision of Plummerita. 1980. Pp. 217-297. £11.00 Vol. 35 No. 1 Lower Ordovician Brachiopoda from mid and south-west Wales. M. G. Lockley & A. Williams. 1981. Pp. 1-78, 263 figs, 3 tables. £10.80 Vol. 35 No.2 The fossil alga Girvanella Nicholson & Etheridge. H. M. C. Danielli. 1981. Pp. 79-107, 8 figs, 3 tables. £4.20 Vol. 35 No. 3 Centenary Miscellanea; Budleigh Salterton brachiopods—Oswald’s Turkish algae—J. A. Moy- Thomas—Burials, bodies and beheadings—Nucleolites clunicularis—Phanerotinus cristatus—Fossil record of teleosts—Neanderthal dating—Hippoporidra edax. 1981. Pp. 109-252. £20.00 Vol. 35 No.4 The English Upper Jurassic Plesiosauroidea (Reptilia) and a review of the phylogeny and classification of the Plesiosauria. D. S. Brown. 1981. Pp. 253-347, 44 figs. £13.00 Vol. 36 No. 1 Middle Cambrian trilobites from the Sosink Formation, Derik—Mardin district, south-eastern Turkey. W. T. Dean. 1982. Pp. 1-41, 68 figs. £5.80 Vol. 36 No.2 Miscellanea: Dinantian terebratulids—New microfossils—Neseuretus—Archaeocidaris whatleyensis— Carboniferous dasyclad—Nanjinoporella—Toarcian bryozoans—Drybrook Sandstone plants—British fossil bintoniellids—Uraloporella. 1982. Pp. 43-155. £19.80 Vol. 36 No. 3 The Ordovician Graptolites of Spitsbergen. R. A. Cooper & R. A. Fortey. 1982. Pp. 157-302, 6 plates, 83 figs, 2 tables. £20.50 Vol. 36 No. 4 Campanian and Maastrichtian sphenodiscid ammonites from southern Nigeria. P. M. P. Zaborski. 1982. Pp. 303-332, 36 figs. ; £4.00 Vol. 37 No. 1 Taxonomy of the arthrodire Phlyctaenius from the Lower or Middle Devonian of Campbellton, New Brunswick, Canada. V. T. Young. 1983. Pp. 1-35, 18 figs. £5.00 Vol. 37 No. 2 Ailsacrinus gen. nov., an aberrant millericrinid from the Middle Jurassic of Britain. P. D. Taylor. 1983. Pp. 37-77, 48 figs, 1 table. £5.90 Vol. 37 No.3 Miscellanea: Permian Glossopteris in Turkey—Wealden Theriosuchus—Wealden conifer—Permian plants of Saudi Arabia—Carboniferous Edrioasteroidea—British cicadas—Dittonian cephalaspids. 1983. Pp. 79-171. £13.50 Vol. 37 No.4 The relationships of the palaeoniscid fishes, g review ‘based on new specimens of Mimia and Moy- thomasia from the Upper Devonian of Western Australia. “B. = Gardiner. 1984. Pp. 173-428, 145 figs, 4 plates. 0 565 00967 2. £39.00 Vol. 38 No. 1 New tertiary pycnodents from the Tilemsi valley, Renchie of Mali. A. E. Longbottom. 1984. Pp. 1-26, 29 figs, 3 tables. 0 565 07000 2. £3.90 Vol. 38-No. 2 Silicified brachiopods from the Viséan of County Fermanagh, Ireland. (III) Rhynchonellids, Spiriferids and Terebratulids. C. H. C. Brunton. 1984. Pp. 27-130, 213 figs. 0 565 07001 0. £16.20 — Vol. 38 No.3 The Llandovery Series of the Type Area. L. R. M. Cocks, N. H. Woodcock, R. B. Rickards, J.T. Temple & P. D. Lane. 1984. Pp. 131-182, 70 figs. 0 565 07004 5. £7.80 Vol. 38 No.4 Lower Ordovician Brachiopoda from the Tourmakeady Limestone, Co. Mayo, Ireland. A. Williams & G. B. Curry. 1985. Pp. 183-269, 214 figs. 0 565 07003 7. £14.50 Vol. 38 No.5 Miscellanea: Productacean growth and shell shape—Jurassic alga Palaeosiphonium—Upper Ordovi- cian brachiopods and trilobites—Lower Devonian Osteostraci from Podolia—Hipparion from Diavata—Preparation and study of Singa skull—Carboniferous and Permian bryozoa—Lower Eocene trionychid—Montsech fossil insects. 1985. Pp. 271-412. 0 565 07004 5. £24.00 Vol. 39 No. 1 Upper Cretaceous ammonites from the Calabar region, south-east Nigeria. P. M. P. Zaborski. 1985. Pp. 1-72, 66 figs. 0 565 07006 1. £11.00 Vol. 39 No. 2 Cenomanian and Turonian ammonites from the Novo Redondo area, Angola. M. K. Howarth. 1985. Pp. 73-105. 33 figs. 0 565 07006 1. £5.60 Vol. 39 No.3. The systematics and palaeogeography of the Lower Jurassic insects of Dorset, England. P. BE. S. Whalley. 1985. Pp. 107-189, 87 figs, 2 tables. 0 565 07008 8. £14.00 Vol. 39 No.4 Mammals from the Bartonian (middle/late Eocene) of the Hampshire Basin, southern England. J. J. Hooker. 1986. Pp. 191-478, 71 figs, 39 tables. 0 565 07009 6. £49.50 Vol. 40 No. 1 The Ordovician graptolites of the Shelve District, Shropshire. I. Strachan. 1986. Pp. 1-58, 38 figs. 0565 07010 X. £9.00 si Vol. 40 No. 2 The Cretaceous echinoid Boletechinus, with notes on the phylogeny of the Glyphocyphidae and Tem- ~ nopleuridae. D. N. Lewis. 1986. Pp. 59-90, 11 figs, 7 tables. 0 565 07011 8. £5.60 =. | Vol. 40 No. 3. The trilobite fauna of the Raheen Formation (upper Caradoc), Co. Waterford, Ireland. A. W. Owen, a R. P. Tripp & S. F. Morris. 1986. Pp. 91-122, 88 figs. 0 565 07012 6. £5.60 ' | Vol. 40 No.4 Miscellanea I: Lower Turonian cirripede—Indian coleoid Naefia—Cretaceous—Recent Craniidae— ~ Lectotypes of Girvan trilobites—Brachiopods from Provence—Lower Cretaceous cheilostomes. 1986. Pp. 125-222. 0 565 07013 4. £19.00 ' Vol. 40 No. 5 Miscellanea II: New material of Kimmerosaurus—Edgehills Sandstone plants—Lithogeochemistry of Mi Mendip rocks—Specimens previously recorded as teuthids—Carboniferous lycopsid Anabathra—Meyenodendron, — | new Alaskan lepidodendrid. 1986. Pp. 225-297. 0 565 07014 2. £13.00 Vol. 41 No. 1 The Downtonian ostracoderm Sclerodus Agassiz (Osteostraci: Tremataspididae). P. L. Forey. 1987. Ppa 1-30. 11 figs. 0 565 07015 0. £5.50 Vol. 41 No. 2. Lower Turonian (Cretaceous) ammonites from south-east Nigeria. P. M. P. Zaborski. 1987. Pp. 31-66. — 46 figs. 0 565 07016 9. £6.50 — Vol. 41 No.3 The Arenig Series in South Wales: Stratigraphy and Palaeontology. I. The Arenig Series in South — Wales. R. A. Fortey & R. M. Owens. II. Appendix. Acritarchs and Chitinozoa from the Arenig Series of South-wes Wales. S. G. Molyneux. 1987. Pp. 67-364. 289 figs. 0 565 07017 7. £69. a Vol. 41 No.4 Miocene geology and palaeontology of Ad Dabtiyah, Saudi Arabia. Compiled by P. J. Whybrow. 1987. _ Pp. 365-457. 54 figs. 0 565 07019 3. £18.00 Vol. 42 No. 1 Cenomanian and Lower Turonian echinoderms from Wilmington, south-east Devon. A. B. Smith, C.R. C. Paul, A. S. Gale & S. K. Donovan. 1987. 0 565 07018 5. In ce my | Pee ee age vencavaeuevearenny cnet OR hee wee =