= : SITISISHISES eas Srigiecates sata nerneset te keneesr seme ped peep hecens ede Deaene at eke teat setzs, SSS SSS =: 3s 2 eir. SiriricisiseiSs> sizszerssiopse wai a : resnssaers wisi richgisieieinietetereie set leesesce =: eee ewer So soe neens, Sea Se athe ans PE SeSerNES MED enEE TOE aeoubrcear meee segere ares ereres seicccie serene pe Teese tone Taree. aes 95s Te RTI ee eee eT Se tote Teter este sees, Sood Foietet store to: 5 Te TASS TST WoT eat lie Se A TST a Fa an a Fein ee ee i = oe, de 4 ? en eee pe ars 0 2 : a - _ a ‘ - — ae he / 9 - = ¢ ; a i BULLETIN OF THE BRITISH MUSEUM INA TURAL HISTORY) GEOLOGY, : /s \ VOL. Vo [27 we0 | 1960-1962 Neen, PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) LONDON: 1960-1962 DATES 30) PU BLT GAT TON BO: Say EE eee AGS No. 1. 18th November : : : : - 1960 No. 2 17th January : : . ° « 1961 No. 3 17th March 0 : C 0 0 - I961 No. 4. 21st March - A 5 f 6 » I961 No. 5. 14th July : : 5 5 5 - I961 No. 6. 14th July 5 : 0 4 5 - 1961 No. 7. 6th October . : f 5 - 1961 No. 8. 28th November 3 f 5 : - I9Q61 No. 9. 17th April 5 2 . 0 - 1962 PRINTED IN GREAT BRITAIN AT THE BARTHOLOMEW PRESS DORKING BY ADLARD AND SON, LTD. No. CONTENTS 967 | GEOLOGY VOLUME V Fossil Denticipitid Fishes from East Africa. P. H. GREENWOOD Post-Ypresian Plant Remains from the Isle of Wight and the Selsey Peninsula, Sussex. M. E. J. CHANDLER Notes on Eocene Tarsioids and a Revision of some Necrolemurinae. E. L. Stvons Fractured Chert Specimens from the Lower Pleistocene Bethlehem Beds, Israel. J. D. CLARK Flora of the Lower Headon Beds of Hampshire and the Isle of Wight. M. E. J. CHANDLER The Form and Structure of Ctenozamites cycadea. T. M. Harris The Old Red Sandstone of Brown Clee Hill and the Adjacent Area. I. Stratigraphy. H. W. Batt & D. L. DINELEY II. Palaeontology. E. I. WHITE The Ordovician Trilobite Faunas of South Shropshire, II. W. T. DEAN . The Morphology of Botryopteris antiqua. H. S. HoLpEN Index to Volume V PAGE 73s gli 159 175 243 311 359 381 1 FOSSIL DENTICIPITID FISHES FROM EAST AFRICA P. H. GREENWOOD BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 5 No. 1 i LONDON : 1960 FOSSIL DENTICIPITID FISHES FROM EAST (APRICA BY PETER HUMPHRY GREENWOOD, Department of Zoology, British Museum (Natural History). Pp. 1-11; Pls. 1-3; 2 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 5 No. 1 LONDON: 1960 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted im 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical Series. Parts will appear at irregular intervals as they become yeady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 5, No. i of the Geological (Palae- ontological) series. © Trustees of the British Museum, 1960 PRINTED BY ORDER OF DHE DPRUSTEES OF THE BRITISH MUSEUM Issued November, 1960 Price Seven Shillings and Sixpence FOSSIL DENTICIPITID FISHES FROM EAST AFRICA By P. H. GREENWOOD SYNOPSIS Recently, a new and distinctive family of small and primitive isospondylous fishes (Denticipi- tidae) was described from a living population in West Africa. Fossil representatives of this family have now been found in presumed Tertiary deposits near Mahenge, Tanganyika Territory. The living and fossil fishes are closely related but generically distinct ; a new and monotypic genus (Palaeodenticeps tanganikae) is created for the fossils. Other fish remains, predominantly Cichlidae, from the same deposits are described and discussed. INTRODUCTION SOME months ago I was asked to study a collection of fossil fishes from Mahenge in the Singida district of Tanganyika Territory. The age of the deposits is unknown but judging from some of the fishes present, it is probably of middle to late Tertiary age. This conclusion is in accordance with geological evidence from the sediments in which the specimens were found (Dr. G. Mannard, im litt.). Amongst the fishes were a number of small specimens whose identity was a complete puzzle except that they were clearly Isospondyli of a clupeoid type. Various characters especially the denticulate cranial bones appeared to be unique and precluded the identification of the fishes with any known family of living or extinct Isospondyli. This dilemma was solved recently by the publication of a paper describing a new family, Denticipitidae, of extant but primitive clupeoid fishes from rivers in western Nigeria (Clausen, 1959). The degree of resemblance between these fishes and the fossils is so remarkably close, that what differences there are can only be considered of generic value. MATERIAL The fossils are in the form of imprints left in lacustrine shales which have preserved the finest details of bone form and ornamentation. Eight specimens are referable to the Denticipitidae, the majority being entire or almost entire skeletons still naturally articulated. Only one specimen is dis- articulated and dispersed, although all have the posterior part of the skull somewhat crushed. The registration numbers given throughout are those of the Department of Palaeontology, British Museum (Nat. Hist.). (1) Entire skeleton, well preserved. Standard length, 28 mm., total length (measured from snout tip to the posterior tip of the upper caudal lobe) about 32 mm. Holotype of the species: P. 42610. GEOE.|5,) I. 1§ 4 FOSSIL DENTICIPITID FISHES FROM EAST AFRICA (2) Entire skeleton, well preserved. S.L., 28 mm., T.L. 32 mm. Although the finer osteological detail is less obvious in this specimen, there is a very clear imprint of the swim-bladder ; also, the structure of the caudal fin skeleton is clearer than in (i =P. 426m (3) Distorted and incomplete skeleton. S.L. about 28mm. P. 42612. (4) A well-preserved and almost complete skeleton lacking the anterior region of the skull and part of the anal fin. S.L. about 28 mm. P. 426713. (5) and (6) Two specimens, one on each face of a piece of shale ; both are incom- plete skeletons rather poorly preserved. P. 42614 and P. 42615. (7) A very badly preserved but almost entire skeleton of a fish about 25 mm. SL -P42616:; (8) Part of the trunk and caudal skeleton (including the dorsal and anal fins) from a fish about 25 mm. S.L.; moderately preserved. P. 42617. The descriptions which follow are based mainly on specimens (1) and (2) but some additional data were obtained from (3) and (4). The study and interpretation of this material was greatly facilitated by the use of ‘ Silastomer’’ casts which gave a positive mould of the negative impression. From these casts it is possible to deter- mine the extent and distribution of the finest cranial denticulations, as well as many other delicate structures in the cranial and axial skeleton. I am greatly indebted to Messrs. M. J. Meade and H. K. Ward of the Department of Palaeontology for their skilful preparation of the casts. THE BRAMIEY DEN TLClPILE DAE The most outstanding characteristic of the Denticipitidae is, as the name implies, the presence of numerous extra-oral denticles on most dermal skull bones, on the dorsal part of the secondary pectoral girdle and on the scales around the anterior end of the lateral line (see Clausen, 1959 and Text-fig. 1). As far as skull denticula- tion is concerned, almost complete similarity exists between the fossils and the living species Denticeps clupeoides. | Compression of the hinder end of the skull in the fossils makes it impossible to determine the dorsal elements of the pectoral girdle series. Consequently, this region could not be studied. The few and slight differences which do occur in the skull denticulation of the two forms are principally: (i) the presence in the fossils of denticles on the sub- operculum and (ii) fewer tooth rows on the lateral face of the maxilla and mandible, there being one less row on each bone. However, since these points can be checked in only two specimens and because nothing is known about the variability of such characters in living fishes, the taxonomic value of these differences cannot be assessed. Despite careful search I am unable to find denticles on any scale impression, but, as the impressions are weak this difference also should not be given undue importance. More significant differences noted in the fossils are: fewer vertebrae (31 or 32 cf. 40 in Denticeps) and lateral line scales (32 or 33 cf. 37 or 38) dorsal fin arising above the origin of the anal (and not more posteriorly as in Denticeps) ; more principal branched caudal rays (20 cf. 18). These differences are slight when compared with the otherwise great similarity FOSSIL DENTICIPITID FISHES FROM EAST AFRICA 5 that exists between every verifiable osteological character in the two forms. Un- fortunately, the nature of the impressions precludes the study of median structures in the cranium (especially the relation between the supraoccipitals and parietals, and the shape of the nasal and parietal bones) nor is it possible to investigate the nature of the auditory fenestra. The large posttemporal foramen in Denticeps does appear to be developed in the fossils and is visible despite compression and . distortion. Clausen (1959) lays particular emphasis on the unique opercular lateral line system of Denticeps. The operculum carries from five to seven major sensory canals derived Soar fe wal G2 Bay des? Capea g 4 BT es Fic 1. Denticeps clupeoides Clausen Drawing of the syncranium and associated structures showing the superficial bones and the distribution of the dermal denticles ; prepared from an alizarinspecimen. In fresh specimens the denticles on the operculum and on the scales are more numerous. (After Clausen, 1959.) For abbreviations, see Text-fig. 2. from the posttemporal-extrascapular and infraorbital-preopercular branches of the head. The canals are superficial, uncovered by bone and run obliquely across the bone in a postero-ventral direction. Each canal is fringed by rows of teeth. Naturally, such delicate structures are not well-preserved in the fossils. Nevertheless I believe that in two specimens the course of these canals is faintly visible. In these specimens there are two large toothed areas on the operculum, one situated dorsally and the other, larger area, occupying the antero-ventral angle of the bone (see PI. 2). In both areas the teeth are arranged in orderly and approximately parallel lines running obliquely downwards and backwards; the lower group slopes less steeply 6 FOSSIL DENTICIPITID FISHES FROM EAST AFRICA than the upper. The position and distribution of these teeth correspond to those fringing the canals in Denticeps. Both Denticeps and the fossils have virtually identical caudal skeletons; as Clausen remarks, the caudal fin must be described as hemi-homocercal. A peculiar feature of the caudal is the presence of a comb-like group of three or four short, unbranched and detached rays which immediately precede the upper and lower caudal lobes. Laterally, the rays are supported by a band of narrow scales. The fossils show this character very clearly (see Pl. 3). The one difference I can find in the caudal skeleton is that only the last two vertebrae are curved upwards in the fossils whereas in Denticeps the last four to six vertebrae are involved. Finer details in the caudal osteology of the fossils are not easily discernible, but from observations on three specimens I should doubt whether a distinct urostyle is developed ; the last vertebra is short and compact, and the ultimate uroneural does not extend beyond its posterior limits. Clausen (1959) mentions a urostyle in Denticeps but from his photograph of the caudal skeleton in an alizarin preparation, the situation would appear identical with that of the fossils. One other familial character may be noted here, namely the postcleithrum. In Denticeps there are upper and lower postcleithra, both resembling dermal body scales and lying mesially to the shoulder girdle ; the upper postcleithrum is pierced by the main lateral line tube. It is impossible to give a detailed description of the postcleithrum in the fossils because it is largely obscured by the cleithrum ; only the ventral part is visible as a long, slender and postero-ventrally directed rod lying mesially to the cleithrum. No fine details are preserved and any dorsal relations with the lateral line system are obliterated by the crushed posterior part of the skull and upper girdle elements. Considering the differences known to exist between the fossil and living Dentici- pidae, it seems reasonable to recognize the former as a separate genus. Certainly, when equated with the differentiae used in classifying present-day African fishes, these differences cannot be considered to have more than generic value. Genus PALAEODENTICEPS nov. Diacnosis. A member of the family Denticipitidae, differing from the extant genus Denticeps in having fewer vertebrae (31 or 32 cf. 40) and lateral line scales (32 or 33 cf. 37 or 38) ; origin of dorsal fin above or slightly anterior to first anal fin ray, suboperculum with numerous small denticles, caudal fin with 20 (cf. 18) principal branched rays. TYPE SPECIES. P. tanganikae sp. nov. Palaeodenticeps tanganikae sp. nov. (Pl. 1-3; Text-fig. 2) Diacnosis. As for the genus. HoLotyPe. A specimen about 28 mm. standard length. P. 42610. LocaLity. Lacustrine shales near the western margin of the Iramba Plateau, Singida district, Tanganyika Territory, FOSSIL DENTICIPITID FISHES FROM EAST AFRICA 7 AcE. Uncertain but probably middle to late Tertiary. DESCRIPTION. In general appearance, Palaeodenticeps bears a striking resemblance to Denticeps. This overall similarity extends to a number of osteological characters (see above). The scales along the dorsal mid-line form a clearly demarcated ridge and there is a distinct line of keeled scutes along the belly ; about twenty scutes, extending from the level of the posterior preopercular margin to a short distance from the anal fin, form this ventral keel. The first ten pairs of ribs are intimately united with the scutes but the last three pairs are apparently free ventrally. Fic. 2. Palaeodenticeps tanganikae sp. nov. Outline drawing of the syncranium, from the holotype (x 10). P. 42610. BRST., branchiostegal rays; CL., cleithrum; COB., circumorbitals; EX., extra- scapula; F., frontal; IOP., interoperculum ; M., maxilla; N., nasal; OP., operculum P., parietal; PC., postcleithrum; PM., premaxilla; POP., preoperculum; PS., patasphenoid ; PT., posttemporal; SC., supracleithrum ; SOB., supraorbital ; SOP., suboperculum ; ?UH., urohyhal (?) As in Denticeps, the postero-ventral angle of the preoperculum is divided and forms a double projection fringed with denticles. Slightly posterior to the base of the division, there is a round depression which resembles an opening to the lateral line tube. No such pore is mentioned or figured in the description of Denticeps. The divided angle of the preoperculum is not markedly protracted in Palaeodenticeps ; 8 FOSSIL DENTICIPITID FISHES FROM EAST AFRICA in Denticeps it forms a double spine extending to a point below the posterior margin of the operculum (cf. Text-fig. 1 and Plate 2). Possibly, the spinous portion has been broken off in the fossils. As mentioned above, the suboperculum carries scattered denticles. The intero- perculum, like that of Denticeps, is toothed and pointed posteriorly ; it appears to be stouter in Palaeodenticeps than in the extant genus. There are four or five, broad and stout branchiostegal rays ; it is impossible to ascertain whether, as in Denticeps, the median pair is toothed. No obvious differences could be detected in the circumorbital series of the two genera ; the conspicuous, comb-like supraobital of Denticeps does seem to be present in the fossils but this point needs confirmation. In general, the lateral aspects of the pectoral girdle are alike in both genera but it is still unknown whether, as in Denticeps, there are two rows of pectoral radialia in Palaeodenticeps. The fin itself occupies a similar position, but the tip extends relatively more posteriorly in Palaeodenticeps where it reaches to the level of the ventral fin origin ; there are nine pectoral rays in this genus. The ventral fins are poorly preserved in all the specimens and it is impossible to count the number of rays, except in so far as there are at least five and probably not more than eight. The lateral line canal of the flanks is complete and extends from behind the cleithrum to the extremity of the caudal peduncle. It follows a gently curved course, situated midway between the dorsal and ventral body outlines. There are 32 or 33 scales in the lateral line series (counted in two specimens). Apart from the total number of vertebrae and the number of up-turned vertebrae involved in the caudal fin skeleton (see above) there is little difference between the axial skeletons of the two genera. Faint traces of epineural and epicentral bones can be detected in two specimens, but the evidence for the presence of epipleurals is less definite. Dorso-spinal bones (i.e. elements apparently homologous with the dorsal pterygiophores but not bearing fin rays) are present above each of at least the first twelve vertebrae (that is, two more than in Denticeps). The anal fin is long and low, comprising 24 or 25 branched rays ; in the two speci- mens with entire anals there are three unbranched, unsegmented spine-like rays preceding the branched rays, thus giving a total fin ray count of 27 and 28. The dorsal fin is short and comprises eight rays (counted in three specimens) ; the first ray may be unbranched but apart from this possibility and its greater length, it is in no way different from the others. In some specimens, the branched parts of the dorsal rays appear to be very pliable and almost filamentous (judging from the way in which the distal ends of the rays have been thrown into curves). Other specimens seem to have more rigid tips to the dorsal rays, but this may merely be due to the vagaries of preservation. One specimen (P. 42611) is particularly well preserved with little distortion or disarticulation. From this fish I have made a number of measurements, tabulated below. The fixed points used to obtain these measurements are, as far as possible, those used by Clausen (1959) to describe Denticeps clupeoides. Exceptions are marked with an asterisk (taken from the anterior tip of the premaxilla and not the mid-point of the upper jaw between the premaxillae) ; those marked with a dagger FOSSIL DENTICIPITID FISHES FROM EAST AFRICA 9 are taken from the lateral end of the premaxilla and thus accord with Clausen’s practice. mm. Of Sy lle Standard length : : . : : ¢ : = 745100) =- tLength of head é 3 : F ‘ ‘ ; . 6:0 21°5 Depth of body : . . . : - 9:0 B20 {Length of snout I-3 4°7 Length of orbit 3:0 10°7 Postorbital head length 3:0 10'7 *Snout to 1st dorsal ray 17°0 60°8 ;Snout to 1st anal ray 17°5 62:5 {Snout to pectoral insertion 6:0 21-4 jSnout to pelvic insertion . 13°0 46°5 Pectoral to pelvic insertions 7:0 25°0 Length of dorsal fin at base 3:0 10°7 Length of anal fin at base 8-0 28-6 _ Length 5:0 17°8 Caudal peduncle : esr tee abe ECOLOGY Dr. G. W. Mannard of McGill University, who is studying the Singida deposits, has kindly informed me of his preliminary conclusions. He writes “ These sedi- ments are lacustrine shales which fill a depression in an area of granite near the western margin of the Iramba Plateau of central Tanganyika. The structure and heavy mineral content of the shales suggest that they are crater lake deposits over a kimberlite pipe’’. If Palaeodenticeps tanganikae inhabited crater lakes, then its ecology differed considerably from the extant Denticeps clupeoides. This species is fluviatile and occurs in “‘ fast-running, medium-sized streams of low mineral content ”’ (Clausen, 1959). However, the ecological evidence for the fossils is too scanty to be considered conclusive, especially since the evidence furnished by associated fishes (see below) neither confirms nor contradicts it. DISCUSSION The Denticipitidae, both living and fossil, provide an intriguing puzzle. Clausen (1959) considers the relationships of the family with various isospondy- lous families and finds characters in common with the Elopidae, Albulidae and Clupeidae (all of the suborder Clupeoidei) as well as with Osteoglossidae (Osteo- glossoidei). In addition, there are characters not found in any of these families or either suborder, as, for example, the denticulate dermal bones of the head. Indeed, the only other fishes with such extensive extra-oral cephalic dentition are those of the genus Atherion (Atherinidae, Mugiloidei, see Rofen, 1958), which are phyletically far removed from the Denticipitidae. As Dr. Clausen is still studying the anatomy of Denticeps I am unwilling to carry this discussion further. It must suffice to say that wherever the relevant characters are preserved in the fossils I have been able to confirm most of his observations ; from Clausen’s 10 FOSSIL DENTICIPITID FISHES FROM EAST AFRICA published account and my own observations I would suspect that the Denticipitidae represent a group of subordinal status within the Isospondyli. The known distribution of the family, with the fossil genus in the east (Tanganyika) and the extant Denticeps in the west (Nigeria), is unusual; no other family of African freshwater fishes is restricted to such widely isolated areas. Further collecting may show that the living forms have a wider distribution and thus narrow the present geographical gap. Nevertheless the Denticipitidae do provide additional support for the concept of a once widespread, possibly pan-African, and relatively uniform ichthyofauna during the late Tertiary and early Quaternary periods. Evidence already available (Greenwood, 1951, 1959) shows that this was the condition over a relatively restricted area in East Africa where today there are at least three distinct faunal units. THE FISH-FAUNA ASSOCIATED WITH PALAEODENTICEPS At least two other families are present in the Singida collection. One of these is unidentifiable from the impressions preserved. The fishes represented are moder- ately large ; one almost complete vertebral column is from a fish about Io cm. long. Other unidentifiable material consists of isolated vertebrae, ribs and fin rays. None of these specimens is as well preserved as the Palaeodenticeps specimens or those of the other and identifiable family, the Cichlidae. Thirteen specimens, all negative impressions and all referable to the same genus can be identified as cichlids (P. 42602-09 and P. 42618-22). Unfortunately, no entire skeletons are preserved, but from various fragments it has been possible to reconstruct several of the more important taxonomic characters. The Cichlidae are widespread in Africa, and represented by numerous genera and species notorious for their complex and often difficult taxonomy. The really trenchant generic characters are found in the cranial osteology and in the pharyn- geal dentition. None of these characters is preserved in the fossils so it has been necessary to rely on secondary characters such as scale types, number of vertebrae, fin ray counts and the oral dentition. In a few specimens there are clear impressions of relatively stout and unicuspid teeth on the premaxilla and dentary ; the teeth are not numerous in either jaw. Where scale impressions are present it is obvious that the scales are ctenoid. Taken together, these characters virtually eliminate the possibility of these fishes being Tilapia, one of the genera likely to be present in this region. The low number of vertebrae (about 28) and the number of dorsal fin rays (15 spines -+ 9 branched rays) also reduced the possibility of these specimens being Tilapia. The sum of these characters, on the other hand, strongly suggests the genus Haplochromis, an ubiquitous genus distributed throughout Africa. The same characters, together with the general facies of the few almost entire specimens suggest close affinities with the extant H. bloyeti (Sauv.) and H. wingati (Blgr.), especially the former. The resemblance is not exact, but because of the difficulties in differentiat- ing species of living Haplochromis I do not consider that any advantage would acrue from describing the fossils as a distinct species. Haplochromus bloyeti (which HOSSLL DENTICIPLLID PISHES FROM EAST APRICA II occurs in the streams and rivers of present-day Tanganyika) has a wide ecological tolerance, as have most of the living generalized and fluviatile species. It would not be unreasonable to assume the same characteristics for the earlier species, which cannot therefore provide any critical information on the ecological conditions at the time preceding their death. The earliest record of fossil Cichlidae in Africa is from Miocene deposits on Rusinga Island in Lake Victoria (Greenwood, 1951). Fishes from these beds share some characters with those from Singida, notably the small size, similar fin ray counts anda low number of vertebrae. Since the teeth and scales of the Rusinga fishes were not well preserved no generic identification could be made with certainty. Although I suggested that these fishes might be Tilapia (Greenwood, 1951), further material examined by Dr. Ethelwynn Trewavas indicates that some could be referred to Haplochromis or even to Pelmatochronis on the characters preserved (Trewavas, 1957, unpublished report). The identity of the Singida cichlids is certainly more definite. ACKNOWLEDGMENTS It is with great pleasure that I thank Dr. G. W. Mannard and Dr. H. S. Clausen who have discussed with me their unpublished researches on, respectively, the deposits and the living fishes ; my thanks are also due to Dr. E. I. White, Keeper of Palaeon- tology, who read and criticized the manuscript of this paper. REFERENCES CLAUSEN, H. S. 1959. Denticipitidae, a new family of primitive isospondylous teleosts from west African freshwater. Vidensk. Medd. naturh. Foren. Kbh., 121 : 141-151, pls. 1, 2. GREENWOOD, P. H. 1951. Fish remains from Miocene deposits of Rusinga Island and Kavirondo Province, Kenya. Ann. Mag. Nat. Hist., London (12) 4: 1192-1201, pl. 21. 1959. Quaternary fish-fossils. In Exploration du Parc National Albert, Mission J. de Heinzelin de Braucort (1950), 4 (1) : 1-80, pls. 1-14. Institut des Parcs Nationaux du Congo Belge. Bruxelles. ' RoFen, R. R. 1958. The Marine Fishes of Rennell Island. In The Natural History of Rennell Island, British Solomon Islands, 1 : 149-220, pls. 1-11. (Ed. Torben Wolff). Copenhagen. (ENE ANID ID, a Palaeodenticeps tanganikae sp. nov. Cast of the holotype. x 5. With respect to plates 2 and 3 the lighting is reversed. Bull. B.M. (N.H.) Geol. 5, 1 IL ANTES, 3 PLATE 2 Palaeodenticeps tanganikae sp. noy. Cast of the holotype. x To, Bull. B.M. (N.H.) Geol. 5, 1 PIL AN IPIE, PAT EIS Palaeodenticeps tanganikae sp. nov. Cast of the holotype. ™ to. Bull. B.M. (N.H.) Geol. 5. 1 IP IE AD IE; 3} ~ il ee ee — — — a “ — = PRINTED IN ADLARD AND | os BARTHOLOMEW PRESS, DO POST-YPRESIAN PLANT REMAINS FROM THE ISLE OF WIGHT AND THE SELSEY PENINSULA, SUSSEX M. E. J. CHANDLER 4 BULLETIN OF _ THE BRITISH MUSEUM (NATURAL HISTORY) ; GEOLOGY Vol. 5 No. 2 4 LONDON: 1961 4g POST-YPRESIAN PLANT REMAINS FROM THE Die Or WIGHT AND THE . SELSEY PEINTINSULA, SUSSEX BY MARJORIE E. J. CHANDLER Pp. 13-41 ; Pls. 4-11 BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 5 No. 2 LONDON : 1961 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 5, No. 2 of the Geological (Palaeontological) sertes. © Trustees of the British Museum, 1961 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued January, 1961 Price Twenty Shillings POST-YPRESIAN PLANT REMAINS FROM THE ISEE, OF WIGHT “AND. THE SELSEY PENINSULA, SUSSEX By M. E. J. CHANDLER SYNOPSIS The larger geographical features which are the background of post-Ypresian plant life in the Hampshire Basin are indicated. Marine dated deposits in the east of the Basin are seen as related to continental deposits in the west. In the marine beds there is a scarcity of plants, partly no doubt owing to the distance from the land surface on which they grew, partly to the narrow outcrop of the strata. Two marine areas, Whitecliff Bay, Isle of Wight, and Selsey, Sussex, are considered. The former has yielded only one determinable plant up to date, a Limnocarpus in Cuisian Beds. The Selsey area with its variable beds and tropical or subtropical fauna has yielded a few species in part at least of Poltavian type. Reference is made to published descriptions of the deposits by Fisher, Reid, Wrigley & Davis and Curry. INTRODUCTION In studying the fossil plants of the Hampshire Basin it is helpful to have in mind some general picture of the larger geographical features of the period. In post- Ypresian times evidence of various kinds indicates a continental area to the west with a river system which debouched into a sea lying to the east. Tidal swamps occurred at various points as denoted by beds of Acrostichum lanzeanum. They have been found on the present-day coast at Studland and Arne (Dorset), Bourne- mouth and Hordle (Hampshire). The massive freshwater Lower Bagshot Beds of Dorset and Middle Bagshot of Dorset and west Hampshire (to use Gardner’s familiar terms for the Dorset Pipe-clay Series and Bournemouth Freshwater Beds respectively) and the succeeding brackish estuarine beds with coastal shingle banks of the Bournemouth Marine Series were deposited in the continental region while contemporary marine beds were being formed further to the east. For suggested correlation of the deposits of the two areas by Wrigley & Davis and by Arkell see Chandler (1g60a, Table on p. 8). The marine beds are now exposed at the eastern end of the Isle of Wight and in the coast section on both sides of the Selsey peninsula. As might be expected from its position midway between Bournemouth and Sussex the Isle of Wight displays non-marine beds at its western end, well seen at Alum Bay but with a little known flora, and marine beds at its eastern end visible in Whitecliff Bay. Accounts of the fossil plants of the continental western area await publication in a number of mono- graphs and bulletins of the British Museum (Natural History). A consideration of the plant remains of the marine eastern area follows below. GEOL, 5, 2. 2 16 POST-YPRESIAN PLANT REMAINS PLANT REMAINS IN THE MARINE BEDS Although the marine beds were laid down in the sea, or at least at the seaward end of an estuary with marine currents, the land was near enough for plant remains to drift from it to the localities under consideration before they became water-logged to sink and be buried among the marine detritus. Yet it was sufficiently remote for such remains to be scarce. Moreover land mollusca appear so far to be absent altogether while according to Reid (1897 : 8) evidence of only one land mammal has been found. The sparse plant remains include scanty evidence of a true land flora (Arvaucarites, Laurocalyx, Wetherellia) together with estuarine or riparian plants (Limnocarpus, Caricoidea) and Nipa from tidal mud swamps. Although so little of the contemporary flora has survived, that little is important because the beds are well dated by marine organisms including Nummulites while the range of the strata, Cuisian to Auversian, is wide. In the Isle of Wight younger beds are also exposed but these up to the present have not yielded determinable plant remains although sifting may yet reveal some. It must be borne in mind that the habitats of the plants in both the continental and marine areas are in the same land mass, so that the dated plants of the marine beds represent the impoverished relics of arich contemporary flora only vaguely dated where found in the continental beds. Possibly in time to come plants from the dated marine area may help in the more exact correlation of the continental plant beds, for although macrofossils are at present few, pollen may yet contribute to this end and additional plant débris of a larger kind may reward persistent search. Hence to stimulate interest in collecting plants from the marine area is most desirable although the limited outcrop of succes- sive horizons at Selsey makes such collecting an arduous task calling for long-term effort. A. A Solitary Plant, Limnocarpus, in Cuisian Beds at Whitecliff Bay, Isle of Wight The Whitecliff Bay section was described in some detail by Fisher (1862 : 65). He numbered successive beds of the Bracklesham Series there with Roman numerals. The section was again examined by Wrigley & Davis (1937 : 205) who used Fisher’s Roman numerals on which they superimposed their own Arabic numbering of subdivisions. These Bracklesham Beds include Cuisian, Lutetian and Auversian horizons. At horizon 1 of Fisher Bed IV a solitary endocarp of ?Limnocarpus forbest (Heer) was found by A. G. Davis. It is the only identifiable plant yet found and is the earliest record of a species which, under the name Limmnocarpus headonensis (Gardner) is abundant in the Bournemouth Marine and later Eocene and Oligocene beds. Fisher Bed IV yields a fauna of Cuisian age including Nummulites planulatus. Wrigley & Davis suggested that the Lower Bagshot Beds and the succeeding lowest Bracklesham, Fisher Beds I to III which they regarded as azoic, may be non-marine (1937 : 205, 219). This opinion Mr. Curry writes (personal note) is doubtful because there are moulds of probably marine mollusca in Bed I, and a band of well-rolled flints at its base. It is necessary here to place on record that Mr. Curry has kindly read this manuscript and made many helpful criticisms together with most generous POST-YPRESIAN PLANT REMAINS 17 and valuable contributions from his great knowledge of the Bracklesham localities. Wrigley & Davis placed Fisher Bed IV at the top of the Cuisian although they stated that the precise upper limits of this stage are obscure. They suggested that Fisher Bed III, where lignitic seams occur (1937 : 219) may yield plants as some seeds have been found in these seams. Unfortunately these seeds have not been traced. B. Plant-bearing Beds of the Selsey Peninsula (Cmsian to Auversian) The Bracklesham Series at Selsey is very variable and yields a prolific fauna said to be of distinctly more tropical aspect than that of the London Clay. According to Gardner (1882 : 472) the animals are those of a tropical sea which extended south over the Paris-Belgian Basin with southern limits just south of Paris and western perhaps near Evreux. Reid (1897) reiterates the variability of the series with its fauna of tropical and subtropical aspect. The classic account of the beds by Osmond Fisher is in the same paper as his description of the Whitecliff Bay sequence and there is some comparison and equating (albeit erroneous, Mr. Curry states) of the two localities. Only the horizons which outcrop along the shore between the Bracklesham Bay Hotel and Selsey Bill and again on the east side of the peninsula were known to Fisher who recorded that the lowest beds (i.e. those north-west of the Hotel towards Chichester Harbour) were covered at that time (Fisher, 1862: 76). The successive horizons which he did examine were denoted by Arabic numerals still in use but he thought, incorrectly it appears, that these beds spanned approximately the same period as the Bracklesham Beds of the Whitecliff Bay section. In 1937 Wrigley & Davis demonstrated that the lower part of the Bracklesham Beds of the Selsey area (those, that is, unknown to Fisher) were of Cuisian age. They succeed the fine-grained marine sediments of the London Clay and were named the Cakeham Beds by Wrigley & Davis who divided them into a Series A below formed of grey quartzose sands equivalent to the Lower Bagshot of Whitecliff Bay and a Series B above formed of sandy clay with a Cuisian fauna including Nummulites planulatus and equivalent to Fisher Bed IV of Whitecliff Bay. Fisher himself (1862 : 93) had equated his Bed 1 of Bracklesham with Bed IV of Whitecliff Bay. A recent Excursion Handbook by Curry & Wisden (1958) is of great practical help to anyone studying the Section and gives an excellent up-to-date summary of the successive beds and their positions in the foreshore. It is unnecessary to repeat here details already published by the various workers except as far as may be necessary to explain the age of the plants. As at Whitecliff Bay the Bracklesham Beds include deposits of Cuisian, Lutetian and Auversian age. The exposures are between tide marks so that collecting is most satisfactory at low water of spring tides as the greatest area is then uncovered (Curry & Wisden, 1958 : 13) but at any time the section or parts of it are liable to be obscured by beach sand toavarying degree. Hence the valueof constant observation spread over many years which can best be undertaken by local collectors. By such efforts the brief plant list may yet be considerably augmented. In particular it is important that plants 7m situ should be obtained and Mr, Curry, in searching for other fossils, has 18 POST-YPRESIAN PLANT REMAINS already made a start by the discovery of plants in washed residues from carefully specified beds. PLANTS OF THE SELSEY AREA AND THEIR HORIZONS Whereas long-drifted wood and twigs are not uncommon, Selsey has so far yielded few genera, species, or even (in most cases) individual examples of a particular plant. The majority of finds have been picked up loose on the beach, probably derived from beds near the spots where they were found e.g. loose fruits of Nipa from the Palate Bed (see p. 34). A few specimens have been collected in situ. In order to show clearly the horizons which have yielded plants, all found up-to-date are listed in the following Table. The first column shows horizon and locality if known, and the finders of the specimens. The second column gives the names of the plants with brief notes and the B.M. (N.H.) registration numbers of extant material. CUISIAN Cakeham A Beds. Opposite West Wittering Beacon. Nipa burtint (Brongniart) Large car- C. Reid (1897 : 6), Gardner (letter to Reid un- bonaceous “skins ’’ filled with grey dated) quartzose sand (see Wrigley & Davis, 1937 : 215, footnote). Friable, col- lapse on drying (see p. 33). A. G. Davis Wetherellia sp. (? W. dixont (Carruth- ers)) (V.35718). Large carbonaceous coccus mineral impregnated in coarse grey quartzose sandstone crowded with lignitic fragments (see p. 37). Cakeham B Beds (Marine Phase). East Wittering. D. Curry Scirpus lakensis Chandler (V.29213). Abraded fruit (see p. 33). LIGNITIC PHASE BETWEEN CUISIAN AND LUTETIAN (ex lit. A. G. Davis) Equated by Davis with Fisher Bed V of Whitecliff Bay, Isle of Wight. Laminated clays and lignite. In situ. East Wittering. A. G. Davis Grey clay with leaves. In situ. East Wittering. A. G. Davis Grey clay, 10-20 yds. east of Bracklesham Lane. In situ. M. Goodchild LUTETIAN Fisher Bed 2 (Turritella Bed). Bracklesham. D. Curry Fisher Bed 4 (Palate Bed). In situ (ex lit. D. Curry) and loose on shore. D. Curry, S. D. & L. T. Garner. Also one fruit recorded C. Reid (1897 : 6) loose near channel from Earnley and Turritella Beds, presumed therefore to come from Fisher Bed 4 From washings. Avaucarites selseyensis n. sp. (V.41918, holotype) and slides of cuticle (V. 41918a—-0). Detached cone-scale on back of same block (see pp. 23, 26). Araucarites selseyensis (V.41919 and slides V.41919a-l). Twig and coun- terpart embedded in clay (see p. 23). Wetherellia dixoni (Carruthers) (V. 41923). Fruit (see p. 37). . ?Wetherellia variabilis Bowerbank (V. 41886). Small coccus (see p. 39). Nipa burtint (Brongniart) (V.29214, V.29215, V.32593, V.40260). Fruits (see p. 34). POST-YPRESIAN PLANT REMAINS 19 Probably Fisher Bed 6. Bracklesham Bay. Ferru- ginous cemented sand block with lignite, glauco- nite and Nummulites laevigatus. 5S. D. Garner Near Fisher Bed 7, loose on shore. Bracklesham Bay. D. Curry Loose on shore, probably from Fisher Bed 7 or 8 certainly Lutetian. D. Curry Fisher Bed 9 (Ostvea teneva Bed). In situ. D. Curry ca. LUTETIAN-AUVERSIAN BOUNDARY Fisher Bed 11 or 12. Washings of sandy marine clay. D. Curry AUVERSIAN Fisher Bed 21 or 22. Opposite Medmery Farm (Medmeney of Dixon, Medmerry of Reid, Med- mery of Fisher, Wrigley & Davis) at low water in beds with Nummulites variolarius. In situ. A. G. Davis. (Equated by Davis with Fisher XVII of Whitecliff Bay, Isle of Wight on field label and in Wrigley & Davis, 1937 : 216) Fisher Bed 21 (Hard Bed). From washings, Selsey. D. Curry Fisher Bed 21. Selsey. D. Curry FAIRLY LOW IN THE SERIES, TOWARDS THE MIDDLE OF BRACKLESHAM Bay (Gardner, 1886 : 4) Dixon (1850 : 84, pl. 9, fig. 3; 1878 : 162) Gardner (1884, pl. 13, figs. 1, 2, 8) Dixon (1850 : 84, pl. 9, fig. 4; 1878 : 163) Gardner (1884 ; pl. 13, fig. 9) . . J. J. B. Ogle. Bracklesham Bay, labelled Pinites Dixoni Bowerbank Horizon ? Small rolled clay pebble on shore. Bracklesham Bay. E. M. Venables. Thought by finder to be Lutetian Nipa burtint (V.33732). Minute, im- mature fruit and unrecognizable poorly preserved seeds (see p. 34). Laurocalyx sp. (V.41885). Cupule (see Pp. 35): Nipa burtini (V.35722, V.35743). In- cluding very large carbonaceous fruits (see p. 35). Carpolithus curryi mn.sp. (V.41887). Pyritized imperfect fruit (see p. 39). Limnocarpus forbest (Heer) (L. headon- ensis (Gardner)) (V.41870—73). En- docarps (see p. 28). Limnocarpus(?) enoymis nmn.sp. (V. 41874-83). Endocarps (see p. 29). Caricoidea obscura Chandler (V.41884). Fruit (see p. 33). Avaucarites selseyensis? (slide V.41921). Twig, totally decayed (see p. 26). Carpolithus sp. (V.41888). Valve of endocarp (see p. 40). Family Potamogetonaceae, Genus? (V. 41922). Rhizome (see p. 30). Pinus dixoni (Bowerbank). Holotype (decayed) formerly in Botany Dept. B.M. (N.H.). Cones. (see p. 22). Pinus dixom. Formerly in Sedgwick Museum, Cambridge (now decayed) (see p. 22). Pinus bowerbanki (Carruthers). Holo- type decayed. Cones. Formerly Gardner Coll. decayed (see p. 22). Pinus bowerbanki (V.3313). Cone (see p. 23). Araucarites selseyensis (V.41920a-4). Twigs now represented by slides (see p. 23). Family Potamogetonaceae, Genus? (V. 41889-41917). Water plant, buds and vegetative fragments (see p. 31). 20 POST-YPRESIAN PLANT REMAINS Horizon? (Widely distributed loose on beach) Loose on shore. Bracklesham Bay. Dixon (1850: . Wetherellia dixoni (Carruthers). Car- 84, pl. 9, fig. 2) bonaceous fruits variously distorted (see p. 36). V.40509, holotype. Five-carpelled fruit. Loose. Thought to be Auversian. E. M. Venables . V.33825-26. Two fruits, one six- carpelled. Loose. Thought to be Lutetian. Mrs. J.G. Turner . V.33827. Seven-carpelled fruit with spine-bases and casts of hollow spines. Loose. Thought to be Auversian (Brook Bed). Mrs. . V.29219. An imperfect fruit. J. G. Turner V.29218, V.29220. Two perfect fruits showing casts of hollow spines. Loose. Picked up nearer the Witterings than Thor- . V.32591, V.33828. ney, site of Campanile Bed. Thought to be Cuisian possibly or from beds equivalent to Fisher V of Whitecliff Bay. A. G. Davis Loose on shore. N. A. Pye . : : , . V.40259. About too fruits scattered over a restricted area of Recent sand probably within the span of Fisher Bed 4 to Fisher Bed 7. (Information from D. Curry.) Much water-worn, very variable in size, laterally com- pressed. Apart from the records in the above list there is drifted and teredo-bored wood which lies outside the scope of this paper. It may, however, be noted in passing that coniferous wood and palm wood have both been recorded from Bracklesham Bay. Palm wood from Tertiary beds is also known from Bognor and Worthing (Carruthers in Dixon, 1878 : 162-167). Of the Selsey plants Avaucarites, Limnocarpus, Caricoidea obscura, Nipa, Wether- ella and Lauraceae are typical constituents of the Poltavian flora (cf. Chandler, 1960: 11). Pinus occurs in London Clay, Hengistbury and Barton Beds. Limno- carpus forbesi (= L. headonensis) is highly characteristic of beds above the Bourne- mouth Freshwater Beds in Hampshire and the Isle of Wight possibly because it is an estuarine plant. Its absence from the older beds of the Bournemouth area may be due to the lack of estuarine conditions rather than to the non-existence of the species at an earlier period. A record by Gardner (1886 : 400) has now been eluci- dated. He mentioned that in the higher beds the surface was dotted with Posidomia “a marine monocotyledonous plant identical with a species now inhabiting the Mediterranean’. Of this plant no figures were given and few words of description, but a recent discovery (p. 30) appears to indicate the presence of a marine mono- cotyledon of the family Potamogetonaceae although the generic affinity is not yet clear, POST-YPRESIAN PLANT REMAINS 21 1. WHITECLIFF BAY, ISLE OF WIGHT (CUISIAN BEDS) ANGIOSPERMAE Class MONOCOTYLEDONES Family POTAMOGETONACEAE Genus LIMNOCARPUS C. Reid emend. Reid & Chandler, 1926 : 68 ? Limnocarpus forbesi (Heer) (PE 7, figs. 20; 27) For synonyms and history of this species see pp. 28, 29. DESCRIPTION. Endocarp (V.33813): A subglobose internal cast broken at the style so that the hilar end of the enclosed seed projected through the matrix which was a cast of the locule. Germination valve not represented, its form nevertheless clearly defined by the smooth edges of the locule-cast. It arose at the attachment but its exact limits at the upper end were difficult to distinguish because the locule- cast had split in the plane of symmetry at the apex. Maximum breadth of valve, 0-4 mm.; length about 1-3 mm. The internal processes on the carpel wall were partly broken so that the corresponding pits on the cast are not preserved, but sufficient remained to show their original position beneath the exposed hilar end of the seed close to the ventral margin towards which they are directed obliquely downwards. Surface of cast smooth, cells obscure apparently about 0-016 mm. in diameter. External surface of endocarp not preserved. Length of endocarp, slightly incomplete, 1-7 mm. ; breadth, 1-3 mm. ; thickness, 1-15 mm. Seed: Solitary, curved, pendulous. Testa thin, surface smooth, shining, cells indistinguishable, hilum indicated by a small truncation on the upper end of the seed. REMARKS AND AFFINITIES. One specimen belonging to Potamogetonaceae and to the extinct genus Limnocarpus. The characters of the external surface are not known, so that whether or not it was rugose cannot be seen, hence the determination is regarded as provisional. The form and size agree with L. forbesi (formerly L. headon- ensis). It closely resembles specimens from Bed K, Bartonian of Barton cliff, Hampshire which are similar smooth locule-casts comparable in every respect but in that case the specific identity is confirmed by the presence of abundant rugose carbonaceous endocarps in the adjacent Lower Headon Beds. The older L. cooperi Chandler from the Oldhaven Beds (Chandler, 1960a : 93, pl. 9, figs. 8-10) is larger than L. forbesi. Its surface, unlike the marked rugose surface of L. forbes? when complete is only slightly rugose, 22 POST-YPRESIAN PLANT REMAINS 2. SELSEY, SUSSEX GYMNOSPERMAE Order CONIFERALES Family ABIETINEAE Genus PINUS Linnaeus Pinus dixoni (Bowerbank) Gardner 1850 Pinites dixoni Bowerbank (in part) in Dixon, p. 84, pl. 9, fig. 3. 1878 Pinites dixoni Bowerbank : Carruthers 7m Dixon, p. 162. 1884 Pinus dixoni (Bowerbank) Gardner, p. 66, pl. 13, figs. I, 2, 5, 8; text-fig. 27. 1960 Pinus dixoni (Bowerbank) : Chandler, pp. 202, 220, pl. 29, figs. 12, 13; pl. 30, figs. 14, 15; pl. 33, figs. 75-79. The original cone from Bracklesham was said by Gardner to be preserved in the Botany Department of the British Museum although in a very imperfect state. It has now decayed. He reproduced Dixon’s figure in his text-fig. 27 (1884: 67). This cone was said to be 5 in. (ca. 126 mm.) long by 13 in. (ca. 38 mm.) broad in the upper part. It agreed closely in size with a cone from the Barton Beds (now decayed) figured and described by Chandler (1960 : 220, pl. 33, figs. 75-79). Asmaller cone in the Sedgwick Museum (85 by 40 mm.) figured by Gardner (1884, pl. 13, figs. 1, 2) appears to have been almost identical with a beautiful specimen from Hengistbury (75 by 35 mm.) taken by Chandler as the neotype of the species (1960 : 202, pl. 29, figs. 12, 13; pl. 30, figs. 14, 15) (V.36352). The exact horizon of these cones is not recorded (see Table, p. 19). Gardner’s cone (1884, pl. 13, fig. 8) was decayed when examined by C. Reid. Pinus bowerbanki (Carruthers) Gardner (Pl. 9, fig. 50) 1850 Pinites dixont Bowerbank (in part) 7m Dixon, p. 84, pl. 9, fig. 4. 1878 Pinites bowerbanki Carruthers in Dixon, p. 163. 1884 Pinus bowerbanki (Carruthers) Gardner, p. 68, pl. 13, fig. 9; text-fig. 28. 1960 Pinus bowerbanki (Carruthers) : Chandler, p. 219, pl. 32, figs. 73, 74. The species was based by Carruthers on very imperfect material and was therefore ill-defined. It appears that the apophyses were not preserved as is obvious from the figure reproduced by Gardner from the original one (1884, text-fig. 28). A second cone (whereabouts not known, probably decayed) was collected by Gardner himself and figured as P. bowerbanki (Gardner, 1884, pl. 13, fig. 9). It was about 100 mm. long, 50 mm. broad with apophyses of the scales having a recurved umbo and marked transverse carina. A cone from Barton, imperfect length, 85 mm. ; breadth between 20 and 40 mm., probably belongs to the same species. It was described by Chandler (1960: 219, pl. 32, figs. 73, 74). Whether these two are really identical with the original mutilated cone cannot now be decided, but since POST-YPRESIAN PLANT REMAINS 23 there were two Pinus species at Bracklesham, it seems best to retain the name P. bowerbanki for these larger cones with sharply carinated apophyses and recurved umbos rather than to introduce another specific name. A cone (V.3313) purchased from J. J. B. Ogle, 1891, labelled Pinites Dixont Bowerbank, Bracklesham Bay appears to belong to this species. It is imperfect below, 130 mm. long as preserved, with maximum breadth (80 mm. from apex) of about 65mm. The apophyses are markedly elongate transversely, the largest about 22 by 5 mm.; they are sharply carinate with conspicuous recurved umbo. Family ARAUCARINEAE Genus ARAUCARITES Presl. Araucarites selseyensis n. sp. (Pl. 4, figs. 1-5; Pl. 5; Pl. 6) 21850 Lycopodites squamatus Brongn. : Dixon, p. 84, pl. 9, fig. 1. Di1Acnosis. Some twigs with broad short spreading leaves, others with narrow acicular leaves. Leaves amphistomatous, the dorsal stomatal bands being restricted in the broader leaves to the lower end and to a small patch or line of stomata (some- times one stoma only) separated from the main band by a distinct area of ordinary epidermal cells. Sometimes in acicular leaves dorsal stomata absent in the upper half of the leaf, sometimes represented by a narrow line of stomata extending almost to leaf tip. Ventral stomatal bands in all types of leaf extending throughout the length. Many stomata on both surfaces transversely oriented in the broader leaves but some longitudinal or oblique. On the acicular leaves all stomatal bands narrow, the majority of their stomata being longitudinally aligned. Ordinary epidermal cells with numerous small oval pits, commonly rectangular, sometimes slightly narrowed towards their extremities. Cone-scale subquadrangular, broadest at about two-thirds of the length from the base, apex with long parallel-sided process having rounded distal end. Hototyre. A twig with broad leaves embedded in small clay block (about 45 to 50 mm. in diameter) on back of which is detached cone-scale. Brit. Mus. (N.H.), No. V.41918. DESCRIPTION. Twigs: Varying in habit with broad spreading leaves as in the holo- type (PI. 4, fig. 1) or narrow acicular leaves as in Pl. 5, fig. 7. A twig showing both types of leaves was macerated in order to obtain leaves for cuticle preparation (cf. Pl. 6, fig. 13 andslides V.41920a-7). Leaves in all twigs falcate, spirally arranged, sharply pointed, decurrent, normally four-sided, the sides or facets often unequal. Broader leaves often with a wide angle over the midrib on the ventral surface, while some of the lateral leaves may be distorted. In some leaves distortion in growth so great as to produce a bifacial effect. V.41918 is about 25 mm. long with leaves which may measure about 5 mm. from the tip to the point where they spring away from the twig on the dorsal side. V.41919, represented by counterparts, has a more slender appearance and is about 35 mm. long with leaves about 6 or 7 mm. measured 24 POST-YPRESIAN PLANT REMAINS as in V.41918 above. The length and narrowness of the leaves on this twig produce much narrower facets but a few broader more spreading leaves were exposed at the base of the twig after the figure in Pl. 5, fig. 7 had been taken. Carbonaceous substance was preserved in all specimens together with impressions in clay but as it was much cracked cuticle preparation was difficult. In particular the dorsal cuticle was cracked and brittle and broke up when mounted permanently after preliminary examination. Yet enough fragments remain to demonstrate some of the outstanding features. All leaves amphistomatous, the stomata being concentrated in well- defined stomatal bands. Upper surface : Stomatal bands always extend from base to apex but do not merge at the apex in the examples seen. They narrow upwards so that whereas near the base there may be as many as nine longitudinal lines of stomata, near the apex these are reduced to one or two lines per band. Sometimes many stomata in a line are contiguous, eight to ten such having been counted (slides V.41918 a, b, V.41920c). Adjacent lines may also be contiguous but auxiliary cells are never shared either in the same line or in adjacent lines. The stomatal bands are inevitably narrower in the long narrow leaves of V.41919 (cf. also slides V.41919a, V.41920b). Some of these narrower leaves show only two lines of stomata, or even one, in a band ; others show three or four lines always reduced to one towards the apex. On the decurrent basal flange (V.41918c, e) there may be two or three lines of oblique stomata above, but one only at the extreme base. In this decurrent region in the broad leaves the cells appear somewhat larger than in the free part of the ventral surface. On the whole in the broad leaves transversely oriented stomata predominate (PI. 6, figs. 14, 16, 17) although oblique and longitudinally oriented ones also occur. The asymmetry of a leaf may however result in one stomatal band having transverse stomata and the other oblique (slides V.41918a-c, e, V.41920c). Again the band on the broader of two facets in an asymmetrically developed leaf may have more lines of stomata than that on the narrower facet. In the acicular leaves stomata are more frequently longitudinally oriented but transverse and oblique ones may occur. They vary in size, usually being larger on the broader leaves. The outer pore, where measured, ranges from 0-027 to 0:04 mm. There are from four to six auxiliary cells arranged end to end around the pore and these cells are normally differentiated into polar and lateral. There is also commonly a regular outer ring of concentric cells. In the broader leaves the breadth of the auxiliary cells combined with their transparency and the manner in which they sweep around the stomata make the apparati con- spicuous as light coloured oval to subcircular patches in the cuticle. Occasionally ordinary epidermal cells are so arranged as to produce three concentric rings in parts of astomatal apparatus. Where the guard cells abut on the auxiliaries a pair of thick scales is present or there may be a ring of cuticle (Pl. 6, figs. 15, 17). The ordinary epidermal cells of the stomatal bands are often quite irregular in shape and frequently equiaxial. However several transversely elongate and aligned cells commonly occur between adjacent stomata in the same longitudinal row as well as elsewhere some- times. The three bands of ordinary epidermal cells outside the stomatal bands naturally tend to be wider in the broader than in the acicular leaves, the cells them- selves also being broader. In the marginal bands of the broad leaves the two POST-YPRESIAN PLANT REMAINS 25 outermost lines of cells are aligned obliquely. In one leaf examined six to ten rows of cells were visible in the marginal bands ; in another there were twenty to twenty- six rows. The median band of ordinary epidermal cells also varies greatly in width, twenty to thirty-eight rows of cells having been counted. These epidermal cells are commonly rectangular in broad leaves and in all leaves are straight-sided not sinuous. They become shorter towards the junction with the twig and may here be equiaxial or even transversely elongate and aligned. In acicular leaves they tend to be relatively narrow and are sometimes narrowed towards their extremities (Pl. 5, fig. 8; cf. also slides V.41919a, k, V.41920g). Small oval or subcircular pits irregular in size and distribution are scattered over the cell walls in all leaves (PI. 6, figs. 15, 19). Lower surface: The cuticle is somewhat denser with narrower and smaller cells (slides V.41918f, V.41920a) this character being most marked in acicular leaves (Pl. 5, figs. 9, 10; cf. slides V.4rg19¢, V.419200, g). The stomatal bands, unlike those of the upper surface are almost restricted to the base of the leaf in the broader leaves, sometimes extending for a short distance only onto the free outstanding part (slide V.41918f) in other cases occurring only on the decurrent part (slides V.41918a, b). Hence they always terminate a long way below the corresponding ventral stomatal bands (PI. 4, fig. 2). Sometimes, however, a few stomata (three were seen in slide V.41918f but the tip of this leaf later became disorganized on permanent mounting,) or a short line of stomata are present nearer the apex separated from the main stomatal band by a distinct area of ordinary epidermal cells. In slide V.41918a a single isolated stoma occurred in this position (Pl. 4, fig. 2). It is not possible on the amount of material seen to state that these separated patches of stomata are always present, further evidence is needed. In the acicular leaves there may be a long stomata-free tip (Pl. 5, fig. 10) but an asymmetrically developed leaf in slide V.41920a (PI. 5, fig. 9) shows a single line of stomata at the distal end of the leaf surrounded by dense narrow epidermal cells. In this latter case it has not been possible to ascertain whether this line is separated from the main stomatal band below as only the upper half of the leaf was preserved. In the broad leaves seen the stomatal bands may have seven to nine longitudinal lines of stomata at the base the number decreasing upwards. Invariably individual stomata and lines of stomata are more widely spaced than on the upper surface. Frequently two lines have four or six lines of ordinary epidermal cells between them (V.41918), k, Pl. 6, figs. 18, 19) which form conspicuous “‘ lanes ’’’ in the stomatal band. In these “ lanes ”’ the cells are usually rectangular and longitudinally aligned. A few stomata may occur quite close to the margin of a leaf on the narrow decurrent basal region in the middle of ordinary epidermal cells. In the acicular leaves two lines of stomata are common but four to five lines have been seen (V.41919f, 7). Stomata in the broad leaves are oblique or transverse, sometimes longitudinal (V.41918c, d, k) but transverse and longitudinal stomata occur in a single leaf. In the acicular leaves the stomata are commonly oblique or longitudinal and frequently contiguous. As on the ventral surface scales or a ring of cuticle are present at the junction of guard cells and auxiliaries and there is the same stomatal structure but on the whole the stomata tend to be smaller. Pits are conspicuous on the walls of the ordinary epidermal cells in both types of leaf. At least two fragments from a decurrent leaf base showed 26 POST-YPRESIAN PLANT REMAINS pits with two concentric rings like bordered pits. In the broad leaves the median band of ordinary epidermal cells displays a variable number of cell rows, eight being seen in one leaf, twenty in another. These cells are most commonly rectangular. In the acicular leaves they are long and narrow with a more noticeable tendency to narrow towards their ends. A stout twig with acicular curved leaves is figured in Pl. 5, fig. 12. The leaves were angled on both surfaces, the angle on the upper surface being inconspicuous, that on the lower clearly marked as were the lateral angles also. Unfortunately this twig has completely decayed and a mere scrap of cuticle showing the ordinary epidermal cells of the median band was obtained (V.41921). While therefore it probably belonged to the same species as the other twigs this awaits confirmation from further material from horizon Fisher 21 or 22. It is highly probable, as suggested by Gardner (1883 : 59) that Lycopodites squama- tus Brongn. (Dixon, 1850, pl. 9, fig. 1) was also an Avaucarites twig, but as the relation- ship of the new material either to Brongniart’s species or Dixon’s specimen cannot be established this specific name has not been adopted. Female cone-scale (Pl. 4, figs. 3-5): The detached cone-scale (V.41918) is sub- quadrangular in outline, broadest at about three-quarters of the length from the base where there are two lateral angles of about go°. Below this line of greatest breadth the scale is contracted to the basal attachment. Above it the sides converge more rapidly to the apex. At the apex there is an attenuated subparallel-sided process with rounded distal end and slight median ridge on its upper surface. The external apophysis of the scale lies above the line of greatest breadth. It can be seen on the impression to the left in Pl. 4, figs 3, 4 where exposed by the flaking away of the carbonaceous substance. It shows as a marked facet separated by a distinct ridge from the conspicuous longitudinally striate limb of the scale. A less prominent transverse ridge on the upper surface of the carbonaceous scale has flutings, rounded upwards, which are visible on the right in PI. 4, figs. 4, 5 where the actual substance of the scale is preserved. This substance has cracked on drying along the flutings. Possibly this fluted ridge represents a ligule. Below the ridge (or ligule) the upper surface shows two or three rows of large rounded closely contiguous pits producing a regular pattern for a short distance. Further down the limb there are rectangular cells (larger than those in an Avaucarites scale from the Upper Headon of Colwell awaiting description). They give rise to longitudinal striations coarser than the ones in a corresponding position on the lower surface of the scale. In addition there are more conspicuous longitudinal rounded ridges and grooves giving rise to flutings due to thick resin ducts in the wall of the scale. The ducts lie opposite to and alternating with the apices of the rounded flutings of the ridge and are from 0-2 to 0-4 mm. apart. On drying the resin swelled, rupturing the thin tissues which originally covered the ducts so that solidified resin was thereby exposed. Two small matrix-filled apertures are seen on the upper surface just below the base of the terminal process (PI. 4, fig. 5) which appear to lead into canals. A longitudinal section at right angles to the breadth exposed by the rupture, through drying, of the carbonaceous scale revealed a small cavity indicated superficially by a slight median convexity. It extended only for a short distance down the scale. A second smaller cavity with a trans- POST-YPRESIAN PLANT REMAINS 27 versely striate lining layer lay behind it in the thickness of the scale. The significance of these cavities is not clear. No seed seen. Length of scale including apical process, 16 mm. ; maximum breadth, 11 mm. ; length from tip of spine to line of greatest breadth, 7 mm.; length of limb below line of greatest breadth, 9 mm. ; length of spine, 4 mm. REMARKS AND AFFINITIES. The association of female cone-scales similar to the one described above with foliage of Avaucarites type at Bembridge (PI. 4, fig. 6), Selsey and Célas (Alais basin, France) points to a connexion between scales and foliage confirmed by actual organic continuity in the Célas material described by Marion as Doliostrobus (1888 : 1-20, pls. 1, 2). The scale from Bracklesham Bay has therefore been referred with the associated foliage to Avaucarites selseyensis. Despite differences of habit shown by the foliage the cuticle structure appears to indicate that the twigs belong to a single species. This is distinguished in the best preserved broad leaves from the more or less contemporary Avaucarites in the Bournemouth Beds by the restriction of the stomata on the dorsal surface to a stomatal band at the base of the leaf with sometimes a small area, separated from the main band, nearer the leaf tip as described above. In the Bournemouth species the dorsal stomatal bands extend throughout the length of the leaf ending only a short distance below those of the ventral surface. A. selseyensis is also distinguished by the more regular rectangular ordinary epidermal cells particularly in the broader leaves, with their clearly defined irregularly scattered pits. Also a larger proportion of the stomata are transversely oriented. A fragment of cuticle from Bournemouth has been described by Bandulska (1923 : 248, pl. 20, figs. 13, 14) and more awaits publi- cation in a monograph of the Bournemouth Flora. The distribution of the dorsal stomata also distinguishes A. selseyensis from A. gurnardi Florin from the Bembridge Beds, Isle of Wight (én Reid & Chandler, 1926: 48) where again the dorsal stomatal bands extend almost throughout the length of the leaf. However the ordinary epidermal cells of A. gurnardi more nearly resemble those of A. selseyensis in being rectangular-ended and conspicuously pitted. An Australian species of probable Oligocene age described by Cookson & Duigan (1951 : 428, pl. 3, figs. 19-24) and actually referred to the genus Avaucaria itself as A. lignitica shows a somewhat similar distribution of stomata which the authors point out is found in seven living species of Avaucaria of the section Eutacta. In this case the characters of the female cone, seed and male cones confirm the relationship. As regards the scale of A. selseyensis it appears quite distinct from the Bembridge scales (Pl. 4, fig. 6, also Reid & Chandler, 1926: 52, pl. 2, figs. 17, 18) and from a Colwell Upper Headon scale mentioned above on account of its narrower, more attenuated form, possibly also on account of the long apical process which may however be incomplete in the Bembridge and Colwell material. The cone-scales of the Bournemouth Avaucarites are not known. There is a close resemblance between the scale of A. selseyensis and those of Marion’s Doliostrobus Sternbergii. Both have a similar cuneate form with steep-sided apex and long apical process. Doliostrobus is known to show great variation in scale size in accordance with the position in the cone e.g. one scale was 17 mm. long, 12 mm. broad and the distance from the tip of the spine to the line of maximum breadth was 10 mm. The largest scale depicted 28 POST-YPRESIAN PLANT REMAINS was about 18-20 mm. long (including the spine), the smallest 8 mm. A consistent difference lies in the sharply pointed tip of the spine of Doliostrobus as against the rounded tip of A. selseyensis. No leaf cuticle structure is known in the case of Doliostrobus. The cuticle, form of foliage, and scale with long process in the Selsey material indicate Araucarian affinities, but more information is needed as regards the cone- scales and seed structure before it can be referred to the living genus Avaucaria. Hence the reference to the conventional genus Araucarites as A. selseyensis. ANGIOSPERMAE Class MONOCOTYLEDONES Family POTAMOGETONACEAE Genus LIMNOCARPUS C. Reid emend. Reid & Chandler, 1926 : 68 Limnocarpus forbesi (Heer) (Pl. 7, figs. 22-24) 1862 Cyperites Forbesii Heer, p. 373, pl. 18, figs. 20, 21. 1888 Cyperites forbesti Heer: Gardner, p. 422, pl. 3, fig. 16. 1888 Carpolithes headonensis Gardner, pp. 420, 423, pl. 3, fig. 30. 1898 Limnocarpus headonensis (Gardner) C. Reid, p. 464, text-fig. p. 465. 1925 Limnocarpus headonensis (Gardner) : Chandler, p. 13, pl. 1, figs. 4a—c, text-fig. 3. 1926 Limnocarpus headonensis (Gardner) : Reid & Chandler, p. 70, pl. 4, figs. 1-3. 1960 Limnocarpus headonensis (Gardner) : Chandler, pp. 205, 222, pl. 30, figs. 16, 17; pl. 33 figs. 84-86. DESCRIPTION. Endocarp: Subovoid with large deep pit on each broad surface at about one-third of the length of the endocarp from the apex (PI. 7, figs. 22, 24). Style long, patent, slightly reflexed at apex when perfect (Pl. 7, fig. 23), terminal on the ventral margin. Short stalk persistent. Germination valve narrow-ovate arising near the stalk but not reaching the style base, with three strong longitudinal ridges one median and two marginal. Lateral ridges only slightly tubercled. Broad surfaces rugose and pitted, the pits formed by the merging of adjacent rugosities. Length of endocarp, 1-75 mm. including style (about 0-25 mm.) ; breadth, 1-25-13 mm.; thickness, about 1 mm. A smaller abraded endocarp measured: length, I-5mm.; breadth, 1 mm. ; thickness, 0-75 mm. REMARKS. The species is represented by three figured and five unfigured (V.41873) specimens. They agree with the species hitherto referred to L. headonensis (Gardner). It occurs in the Lower Headon of Hordle and the Bembridge Beds as well as in the Hengistbury and Barton Beds. There are unpublished records in the Lower and Upper Headon of Colwell Bay, Isle of Wight and in the Bournemouth Marine Beds. The occurrence of the endocarps in enormous quantities in estuarine beds e.g. in the Limnocarpus Band at Hordle, indicates a gregarious habit probably in brackish lagoons or quiet backwaters. Reid & Chandler (1926: 71) expressed the opinion that fruits named Cyperites forbest by Heer (1862 : 373, pl. 18, figs. 20, 21) were POST-YPRESIAN PLANT REMAINS 29 probably crushed specimens of L. headonensis (as it was then named) but as no extant material of C. forbesi was known and Heer’s figures and description were inadequate the suggestion could not be confirmed. Recently a hand specimen (No. 6712) labelled Cyperites forbesi Heer full of crushed Limnocarpus has been seen in the Geological Survey Museum where Heer’s specimens from Hamstead are pre- served. It is believed to be Heer’s figured specimen (1862, pl. 18, figs. 20, 21) and is so labelled with a query. Other fruits (No. 76921) similarly labelled C. forbes: (Burdett Coutts collection) confirm this relationship. Hence under the rules of nomenclature Limnocarpus headonensis must now be referred to L. forbes: (Heer) since the genus Cyferites was first created by Lindley & Hutton for Cyperaceae leaves and was subsequently used by Heer for all remains of Cyperaceae of unknown generic affinity. Limnocarpus is definitely referred to Potamogetonaceae and is therefore retained as the generic name of this species. Limnocarpus? enormis n. sp. (Pl. 7, figs. 25-33) Diacnosis. Endocarp ovoid with large scattered tubercles. Style long straight, ventral margin with conspicuous rounded ridge. Internal condyle short associated with a small inconspicuous comma-shaped depression on each side of the ventral edge at about one-third of the length from the apex. Keel with three longitudinal tubercled ridges having also a few scattered tubercles on the surface between the ridges. Size very variable. Length (including style) about 2-5-3-5 mm. ; breadth, 1-75-25 mm.; thickness, 1-5-2 mm. HoLotyPe. V.41874. DeEscrIPTION. Endocarp: Ovoid (frequently somewhat flattened in fossilization) with long patent style more or less terminal on the ventral margin and slightly forwardly directed. Ventral margin with a conspicuous longitudinal rounded ridge. Broad surfaces with large scattered rounded tubercles conspicuous in unworn specimens (Pl. 7, figs. 25, 30) but scarcely seen in worn ones (Pl. 7, figs. 26, 28). Keel (germination valve) ovate, broad, arising near the basal attachment but not extending to the base of the style having three conspicuous longitudinal ridges, one median and two marginal bearing a few low rounded tubercles. A few scattered tubercles also lie between the ridges (Pl. 7, figs. 27, 32, 33). Lateral pits differing from those of Limnocarpus forbest in being very inconspicuous, long and narrow, comma-shaped close to the ventral ridge arising not far below the style base and directed towards the ridge narrowing downwards. Sometimes they are indicated only by a line of close-set tubercles externally (Pl. 7, figs. 29, 30). Internally these pits are associated with only a small process on the locule wall (PI. 7, fig. 31). Exter- nal surface of obscure equiaxial cells. Seed not preserved but from the character of the locule it can have been only slightly emarginate near the hilum around the small process and with a large lower and small upper limb. Locule lining of small equiaxial cells with large cavities. Dimensions very variable. The following are examples : I) Length (including 1 mm. long style), 3-5 mm.; breadth, 2 mm.; thickness, GEOL. 5, 2. 3 30 POST-YPRESIAN PLANT REMAINS 2mm. Valve or keel missing extending to 0-3 mm. from style base. 2) Length (including style, I mm.), 3:25 mm.; breadth,-2 mm.; thickness, 1-55 mm. Keel missing, endocarp collapsed below. Lateral pits begin at approximately 0-5 mm. from style base and extend for about 1 mm. (Pl. 7, fig. 26). 3) Length (including style, 0-5 mm.), 25 mm. ; breadth, 1:75 mm. ; thickness, 1-6mm. Keel preserved. 4) Length (style missing), 2-5 mm. ; breadth, 1-75 mm.; thickness, 1:75 mm. Keel missing. 5) Detached keels ca. 2-2-5 mm. long, 1:2 mm. broad. 6) Length of style in broken endocarp, I mm. Length of endocarp incomplete. Breadth, 2:-5 mm. Foramen indicated by flanking tubercles (PI. 7, fig. 30). REMARKS. Represented by seven endocarps (some imperfect), two detached keels and three fragments. They appear to differ from the genus Potamogeton and from Limnocarpus forbesi in the less curved seed-cavity and from Ruppia in the gently curved ventral margin which is unlike the gibbous margin of that genus and in the keel arising near the peduncle. Although the resemblance is greatest to Limnocarpus the fruits are referred to it only doubtfully as there may be other genera as yet unknown to the writer which are more comparable. From all the fossil species of Potamogetonaceae previously described it is distinguished by its large size and the closeness of the comma-shaped inconspicuous lateral pits to the ventral margin. There is no evidence as to whether the low tubercles scattered over the surface were originally spines. Genus? (PL. 8, fig. 38) DESCRIPTION. Rhizome: Repeatedly branched with close-set nodes and long narrow transversely elongate and oriented leaf-scars which do not completely encircle the stem but appear to embrace about half its circumference. One leaf- scar underlies each node but it is not clear whether there are additional scars between the nodes. The scars show traces of several bundle fibres. The branches arise in the axils of the leaves. One branch a (PI. 8, fig. 38) is seen on the right at the base (as preserved) and extends for about 28 mm. before it is broken at the edge of the stone. At its point of origin it is very small but it broadens rapidly upwards to 6 mm. in diameter. Its exposed upper surface shows at least six leaf-scars and the cut transverse end of another branch on the right at b (PI. 8, fig. 38) close to its origin. The main stem is visible for a short distance only as it is broken above the second node. It bears another branch c (PI. 8, fig. 38) on the left in the axil of a conspicuous leaf-scar. About 6 mm. of this branch is preserved before it reaches the edge of the stone where it is about 5 mm. broad. A small circular funnel at one node may indicate the point of emergence of another branch now broken. No roots have been seen but these, if present would lie on the lower surface of the rhizome and be concealed by matrix. The leaf scars project, each internode gradually increasing in diameter upwards so that the rhizome has a distinctive jointed appear- ance. Surface covered by a thick shining somewhat dimpled epidermis which is formed of small square or rectangular cells about 0:008-0-009 mm. in diameter. The POST-YPRESIAN PLANT REMAINS 31 slightly thickened lateral walls of these cells give a fine longitudinally striate appear- ance and on drying the epidermis and rhizome tend to crack along the striae. Inside the stem there are traces of large air-cells or cavities varying much in size, the largest being about 0-027 mm. in diameter. Subglobular pyrites casts sometimes occupy their cavities and are conspicuous when the carbonaceous walls have decayed. Leaf : A detached parallel-sided fragment of an elongate leaf / (Pl. 8, fig. 38) lies athwart the lower end of the specimen but almost certainly belongs to it as indicated by its identical epidermal structure. The tip (seen on left) is blunt and rounded. The width is 2 mm. but only about 13 mm. of the length are preserved. There are traces of several parallel longitudinal nerves. REMARKS. This unique specimen, found by Curry in Fisher Bed 21 at Selsey, suggests at a first glance rhizomes radiating from a centre although closer examination shows repeated branching at successive levels. The jointed stem with large leaf- scars is highly characteristic of certain marine genera of Potamogetonaceae such as Posidomia and Cymadocea. Gardner (1886 : 400) reported that the surface of one of the Bracklesham Beds was “ dotted over ”’ with “ Posidonia’”’ but added that “ the rhizomes radiate from a centre, whilst in the French and other European fossil species they are long and branching”’. It appears highly probable that V.41922 is a rhizome of the supposed “ Posidonia’”’. The living Posidonia has close-set leaf- scars and a jointed rhizome thickly clothed with shredded remains of leaves reduced to bast fibres. Cymadocea shows similar features. The generic relationship cannot be regarded as finally established and the possibility that these rhizomes may belong to one of the species of the extinct Limnocarpus found in the Bracklesham Beds of this area cannot be excluded. It does at least seem clear that it represents a genus of Potamogetonaceae with jointed rhizomes. Whether there is a connexion with the branches and buds of Genus? (see Pl. 7, figs. 35-37; Pl. 8, figs. 39-48) is again purely a matter for speculation on evidence at present available. Family POTAMOGETONACEAE? Genus? (Pl. 7, figs. 35-37; Pl. 8, figs. 39-48) Description. Buds or bulbils preserved in two different ways giving different appearances. Some, seen from the side, are oval or oboval in outline, originally ovoid but compressed in fossilization, truncated at the base by a large slightly sunk attach- ment scar now ellipsoidal originally subcircular. From the edges of the scar can be traced overlapping three or four stiff rather pointed bud-scales or bracts with irreg- ularly toothed upper margins. The scales are represented by casts of their inner surfaces with traces only of organic substance. They are seen in V.41895-06, V.41902, V.41907, V.41909 (cf. Pl. 8, figs. 43, 45). Others, the most numerous, are usually larger and inflated, subglobular and they show very clearly the basal sunk circular attachment scar with punctations marking points of entry of a number of fibro-vascular bundles. These clearly represent internal casts of the external scales, GEOL. 5, 2. 3§ 32 POST-YPRESIAN PLANT REMAINS the growing point and inner scales having presumably decayed, their place being taken by a mud-cast of the globular cavity thus formed. Such mud-casts are pre- served in V 41889-94, V.41900 (cf. Pl. 7, figs. 35-37, Pl. 8, figs. 39-45). In V.41890 (Pl. 7, fig. 36) and V.41897, scales still remain adhering to the mud, the former specimen also showing the internal mud-casts last described still adhering to the scales. In V.41902 near the basal scar thin-walled areoles, about 0-038 mm. in diameter are visible in small patches, each areole being occupied by minute equiaxial convex rounded cells, also seen in V.41895. These cell patches represent organic tissues still adhering to the casts or their impressions. Obscure longitudinal striation is seen in V.41894 and V.41895 suggesting veins although no actual fibres are pre- served. Certain of the buds show an asymmetric development as in V.41891 and V.41897 (Pl. 7, fig. 37; Pl. 8, fig. 46) probably indicating that they have begun to grow out laterally. A further stage of growth is seen in V.418q9 (PI. 8, fig. 48). Buds arise at the nodes of a puckered or flanged stem fragment (V.41890, Pl. 7, fig. 36) but sometimes from a flatter more ribbon-like stem (V.41892; and PI. 8, fig. 39). They may be solitary and alternate or several may be associated at one level. The stems retain remains of carbonaceous tissues and show the impressions of thin-walled cells about 0-038 mm. in diameter like those on the bracts (V.41889, V.41892). There is also obscure evidence of flat ribbon-like finely longitudinally striate leaves always very imperfect. One arises from a node in V.4188qQ (cf. Pl. 7, fig. 35). Occasionally when the globular internal casts are broken from the stems a deep cup- like hollow is exposed with circular cavity at its base. It is surrounded by puckered remains of scales (V.41901). In V.41889 basal scales are well seen with: the bud-cast in profile projecting from between them. This cast also has fine concentric puckerings on its surface suggesting that the surrounding bracts (now gone) hada diaphanous thin lining. Dimensions vary considerably according to the stage of development. Examples of the subglobular casts are: 1) Length, 3 mm. ; breadth, 2-5 mm. 2) Length, 2-75 mm.; breadth, 3:25 mm. 3) Length, 1-6 mm. ; breadth, 1-7 mm. 4) Length, 1-6 mm.; breadth, 2:16 mm. Diameter of basal scar, I'25-2 mm. ‘Transverse diameter of dorsiventral flattened detached casts, 2°5 by 25 mm.; 2-4 by 2-4 mm. Examples of the laterally flattened ovoid buds are: 1) Length, 2:16 mm.; breadth) a-remm. 2) Length, 3-2 ami; "breadth 2mm. 3) Length, 3 mm.; breadth, 2 mm. REMARKS AND AFFINITIES. Twenty-nine specimens (V.41889—-V.41917) show obscure fragments of stem and leaf with buds (or bulbils) connected with propagation or the formation of lateral branches. All were derived from the small rolled clay pebble which yielded an Avaucarites twig with cuticle (see p. 19). The evidence for a water plant lies in the character of the large thin-walled areoles or cells. The possible indications of parallel longitudinal nerves suggest Monocotyledones. The buds with circular attachment scar could belong to Cyperaceae, Gramineae or Pota- mogetonaceae, but scales with denticulate margins appear to exclude relationship with the two first-named families so far as these could be examined. The specimen is referred tentatively to Potamogetonaceae but the genus cannot at present be determined. There is as yet no evidence to connect these fragments with Limno- carpus which also occurs in the Bracklesham Beds of the Selsey area. POST-YPRESIAN PLANT REMAINS 33 Family CyPERACEAE Genus SCIRPUS (Linnaeus) Scirpus lakensis Chandler 1960 Scirypus lakensis Chandler, p. 206, pl. 30, fig. 26. The apical end of a characteristic fruit showing the contraction at the style. Only the inner yellow shining translucent integument is preserved but it shows clearly the characteristic longitudinal ridges and transversely elongate hexagonal cells. Genus CARICOIDEA Chandler, 1957 : 86 Caricoidea obscura Chandler (PID7, fig. 34) 1960 Caricoidea obscura Chandler, pp. 207, 223, pl. 30, figs. 27-33, 342; pl. 33, figs. 98-105. Description. A typical fruit, subovoid in outline, truncated below by a large foramen which marks the attachment, rounded at the stylar end. The surface is much abraded but shows the corky wall formed of small equiaxial cells. Length of fruit, 1-75 mm.; breadth, 1-25 mm.; breadth of (flattened) basal truncation, 0-5 mm. REMARKS. The holotype of this species is from Cliff End, near Mudeford. The species also occurs in the Lower Bagshot and Bournemouth Freshwater and Marine Beds. Family NIPACEAE Genus NIPA Thunberg Nipa burtini (Brongniart) (Pl. 9, fig. 49; Pl. 10, figs. 51-53) 1894 Nipadites Burtini (Brongniart) : Rendle, pp. 147-150, pl. 6. 1897 Nipadites Burtini (Brongniart) : C. Reid, pp. 6, 7, fig. 11. 1960a Nipa burtini (Brongniart) : Chandler, p. 133, pl. 14, figs. 4-9. Undoubted large drifted fruits of Nipa burtini “ about the size of a coconut ” were obtained by Clement Reid from Cakeham Beds A and many had been seen earlier by Gardner. One specimen figured by Reid (1897, fig. rz) had previously been figured by Rendle, natural size (1894, pl. 6). When efforts were made to preserve it, it began to break up but remains showing the grey quartzose matrix were formerly in the British Museum (Natural History). As in the case of the Boscombe N7pfa fruits the preservation made it difficult to keep them for long. The carbonized sand-filled hulls were liable to crumble on being taken from the damp beach while the carbon films cracked and peeled, hence it may be assumed that such 34 POST-YPRESIAN PLANT REMAINS specimens would not drift intact far from the beds in which they lay. The published figures show a longitudinally angled fruit, incomplete at the base, umbonate at the apex, typical of N. burtini, and about 13 cm. long, 11 cm. broad. An undated letter from Gardner to Clement Reid states “the Bracklesham fruits [Nipas] are always the same size and nearly as large as the Belgian ones”. Gardner added that it was years since he had found one at Bracklesham and he thought that the beds had since been covered up. Curry records (1958:15) that the Cakeham Sandstone with fruits of Nipa has not recently been seen. Two smaller very differently preserved pyritized Nipa fruits found loose on the shore by S. Garner are figured (Pl. 10, figs. 51, 52). They show marked differences in size and development and resemble the Sheppey specimens in their preserva- tion. Mr. Curry (ex. lit.) reports that they are probably from the Palate Bed (Fisher 4) which has fairly frequently been exposed in recent years, he himself having found similar pyritized fruits im situ at this horizon as well as loose on the shore (V.32593) in the neighbourhood of this bed. V.29214 is a small flattened oboval immature fruit, slightly imperfect at the extreme base. Its surface is longi- tudinally angled and smooth and evidently represents the epicarp. Obscure, ill-preserved, oblong, longitudinally aligned cells are seen. Length of fruit (as preserved), 5-8 cm.; maximum diameter, 3-6 cm.; diameter at right angles to this, 1:5 cm. V.40260 is a fragment, 6-7 cm. long, 5 cm. broad, of a much better developed fruit showing part of the fibrous sarcocarp; when complete this fruit must have been at least 10 cm. long and 6 cm. broad but the breadth is imperfect and the apex much broken. The outer layers are worn so that at the lower end part of the margin of the foramen for germination is visible. Impressions of the stout longitudinal fibres of the sarcocarp are seen all over the surface but the parenchyma cells between them are obscure. There are no details of seed structure on the ill-preserved concave inner side of this fragment but its form and the size of the concavity on the reverse side suggest that it once contained a well-developed seed. A third fruit incomplete at base and apex, 38 mm. broad was found by L. Garner (V. 29215) in the same sort of position. Reid (1897 : 6) describes the finding of a single Nzpa close to the Turritella Beds and to the channel that runs from Earnley (i.e. near the outcrop of the Palate Bed). Although this fruit is no longer extant and its mode of preservation is not described the locality suggests derivation from the Palate Bed especially as the carbonaceous sand-filled shells from Cakeham would be unlikely to survive transport for such a distance. One extremely small immature much ribbed fruit (V.33732) is almost certainly of Lutetian age (communication from Mr. Curry) as the loose block of matrix in which it was found is full of Nummulites laevigatus which only occurs abundantly in Fisher Bed 6, a bed of which only small areas have been seen while its lower part is de- calcified and weathered and buried under Pleistocene and Holocene mud. The lower end of the small fruit is buried in the matrix, the length must have been about 10 mm.; maximum breadth, 4-5 mm. The surface at the upper exposed end shows a film-like layer of rounded equiaxial cells about 0-028—0-038 mm. in diameter exactly comparable with cells seen at the upper end of living Nzfa both in mature POST-YPRESIAN PLANT REMAINS 35 and immature specimens. Lower down the fruit, Recent and fossil, the superficial cells become somewhat elongate and longitudinally aligned giving rise to longi- tudinal striations. A collection of at least seven very perishable fruits (V.35743) was made by Mr. Curry in an area of about 50 square yards on the surface of the Eocene clay eastwards of Bracklesham. The condition of the specimens was such that they could not have travelled far and Mr. Curry considered that they were probably from Fisher Bed 7 or 8 and must certainly in any case be Lutetian. The largest specimen (V.35722) is shown in Pl. 9, fig. 49 ; it was about 19 or 20 cm. long and about 11 cm. broad. Another long narrow fruit, slightly imperfect at the apex, was about 15 by 6cm. Other specimens were mere fragments of the broadest part of the fruits they represented. A noticeable feature of this batch of specimens is that one side only is preserved ; on the reverse side is matrix with, or without, embedded fragments of fruit. Some specimens are overlain by an incrustation of pyritized sand which may mask them almost completely. The matrix in all cases is a somewhat glauconitic quartz sand bound by ferruginous cement and the whole specimen is brittle. . The carbonaceous substance of the fruits shows, when fresh, the striate surface and long narrow cells of the smooth epicarp in the middle and lower regions, and underneath this are the coarse fibres (conspicuous) and parenchyma (obscure and much pyritized) of the sarcocarp. These tissues are compressed and have become very thin. They quickly disintegrate by cracking into rectangular fragments. Class DICOTYLEDONES Family LAURACEAE Genus LAUROCALYX Reid & Chandler, 1933 : 219 Laurocalysx sp. (Pl. ro, figs. 54, 55) DESCRIPTION. Cupule (V.41885): Broadly obconical (now somewhat flattened in fossilization), rather abruptly narrowed into the thick peduncle, smooth rimmed without sepals. Thick walled. Surface wrinkled and puckered, conspicuous longi- tudinal wrinkles occur in the lower half and are continued on to the stalk, while equally marked transverse wrinkles are seen at the rim of the cupule and parallel with its margin. Fine irregular convex more or less equiaxial puckerings of varying shape are found all over the surface. Surface cells very obscure, but apparently originally convex and equiaxial, about 0-013—0-018 mm. in diameter. These cells are now almost obliterated by the stresses due to crushing which may have produced narrow parallel bands variously oriented and having close parallel striae at right angles to the length. Some of the rounded convexities of the surface suggest the presence beneath them of secreting glands. No berry was preserved but the scar of attachment of the fallen berry is seen in the base as a dull unpolished area, 36 POST-YPRESIAN PLANT REMAINS presumably it was originally circular, now, through compression, 10 by 4 mm. in diameter. Length of cupule including stalk, 18 mm.; transverse diameter as crushed, 20 by 6 mm. (probably originally about 15 or 16 by the same). REMARKS AND AFFINITIES. It is impossible to tell the generic relationship of this empty cupule but clearly it belongs to the family Lauraceae. It is most probably derived from the Lutetian horizon (Fisher Bed 7) near to which Mr. Curry found it. Family EUPHORBIACEAE? Genus WETHERELLIA Bowerbank emend. Reid & Chandler, 1933 : 251 Wetherellia dixoni (Carruthers) (Pl. ro, figs. 56-63 ; Pl. 11, figs. 64, 65) 1850 Cucumites variabilis Dixon (non Bowerbank), p. 84, pl. 9, fig. 2. 1850 ‘“‘ Undetermined fossils ’’? Dixon, pl. 12, fig. 20. 1878 Carpolithes Dixon Carruthers in Dixon, p. 164, pl. 9, fig. 2; pl. 12, fig. 20. 1954 Wetherellia dixoni (Carruthers) Chandler, p. 173. EMENDED DIAGNOsIS. Fruit a loculicidal and septicidal syncarpous capsule, originally subglobular but somewhat dorsiventrally depressed, beset externally with hollow spines. Carpels so far seen four to at least eight, locules and septa marked by longitudinal external ribs or angles. Seeds usually shorter in proportion to their breadth than those of W. variabilis Bowerbank. Length of fruit (frequently distorted), 4-16 mm.; breadth, 13-20 mm. Seeds (much distorted), length, 3°7-7 mm.; breadth, 1-4—4-75 mm. ; thickness, I-I°5 mm. Hototype. A five-carpelled syncarpous subglobular capsule falling into segments along loculicidal and septicidal planes of dehiscence. Surface showing spine-bases. Seeds shorter in proportion to the breadth than in W. variabilis. Brit. Mus. (N.H.), No. 40509. DESCRIPTION. Originally spiny with spine-bases or internal casts of hollow spines alone remaining (V.29218, V.29220, V.33827 and 40509). Syncarpous with from four to at least eight radially arranged carpels (possibly more). Carpels often unequally developed at the opposite ends of a transverse diameter giving the fruits a lop-sided arrangement (PI. 11, fig. 64). Subglobular but usually somewhat dorsi- ventrally compressed, the compression much increased by fossilization, occasionally laterally compressed (40259). Longitudinal ribs or angles correspond with the locules and septa, so that there are at least eight to sixteen longitudinal ribs. Dehiscence both septicidal and loculicidal (Pl. ro, figs. 56, 58). Axis fibrous giving off coarse branching fibres which pass more or less transversely across the septal walls of the cocci to the exterior of the fruit. These fibres, and others in the thick- ness of the pericarp apparently terminated in external spines (Pl. 10, figs. 57, 59) which were hollow as indicated by thin shining pyrites internal casts (V.33827) and especially abundant in the furrows between the ribs and angles. Locules extending almost from base to apex of the fruit, and from the axis almost to the periphery, so much compressed tangentially as to be flat except for a narrow inflated area POST-YPRESIAN PLANT REMAINS 37 occupied by the seed. The seeds are solitary in the locules, pendulous, suspended by long arched funicles (about 2 mm. in length) from a point on the axis about one-sixth to two-sevenths of the length of the fruit from the apex. They lie some- what obliquely in the locules so that the distal end is nearer to the axis than the proximal end. Locule lining formed of cells, 0-or mm. in diameter, aligned so as to produce a general appearance of radial striation. The walls between the locules are thick, formed of coarse equiaxial angular cells, but the structure is often obscured by carbonization which causes the fractured surfaces to appear vitreous and struc- tureless. Dimensions of several fruits, all more or less dorsiventrally compressed or otherwise distorted : 1) Length (holotype), 7 mm. along axis ; diameter, 13-5 mm. 2) (V.33825) Length, 4 mm.; diameter, 15 mm. Dimensions of some laterally crushed fruits which therefore show maximum length (V.40259): 1) Length, I6 mm.; diameter, 20 mm. 2) Length, 11 mm.; diameter, 17 mm. 3) Length, I4 mm.; diameter, 14 mm. Some small abortive fruits (V.40259) with minute flat immature seeds measured about 10-12 mm. in both length and diameter. See also V.33827 and V.41923 with abortive seeds. Seed : Oval or elongate-oval, anatropous, narrowed towards the hilum, chalaza at the broader extremity. Testa thin, formed of a single layer of equiaxial poly- gonal or quadrangular cells, 0-025 mm. in diameter. Dimensions of much distorted seeds: 1) Length, 4:7 mm.; breadth, 1-4 mm. 2) Length, 3:3 mm.; breadth, 33 mm. 3) Length, 5-5 mm.; breadth, 4 mm.; thickness, 1 mm. 4) Length, 7mm.; breadth, 4-75 mm.; thickness, 1-5 mm. REMARKS AND AFFINITIES. The most abundant plant fossil of the Selsey shore represented in the British Museum (Natural History) alone by well over a hundred specimens. The horizon or horizons from which these fruits are derived is unknown except in the case of V.41923 from Bracklesham Beds, foreshore at Bracklesham at low-tide mark about twenty yards east of Bracklesham Lane end (probably equi- valent to the top of Fisher Bed V at Whitecliff Bay—Cuisian?). They are carbon- aceous and unlike W. variabilis relatively thin-walled and very light in weight and therefore capable of being drifted along the shore by tides and currents. The generic position of these fruits is clear. The genus Wetherellia was placed tentatively in the family Euphorbiaceae for reasons given by Chandler (1954 : 175). The Bracklesham fruits appear to belong to a distinctive species, for up to the present there is no evidence of external spines in W. variabilis, the locules of which vary in number from two to five (four to eight at least in W. dixon), the walls appear thicker and its fruits are therefore less liable to distortion, the external angles and ribs are less marked, the seeds relatively longer in proportion to their breadth, while the dehiscence is commonly septicidal first and later loculicidal. Wetherellia sp. (? W. dixoni) (Pl. 11, figs. 66-68) DEscrRIPTION. Fruit: Originally two-loculed, ellipsoid or subovoid but now represented by a single one-loculed semiellipsoid coccus which has resulted from 38 POST-YPRESIAN PLANT REMAINS the septicidal splitting of the fruit. Ventral surface flat, dorsal rounded. A median longitudinal suture on both faces indicates the plane of the locule but loculicidal dehiscence has been prevented by the infiltration of mineral cement. Remains of the axial fibres are seen along the suture line on the septicidal inner surface. From the axis one distinct fibre is given off on its right side which is directed radially for a short distance, it then forks, the lower fork quickly disappearing into the thickness of the wall, the other curving upwards and giving off subsidiary branches which also penetrate the wall. Carpel thick and woody with indications of coarse equiaxial parenchymatous cells. Locule lining close-textured seen in section but its surface cells not visible. External surface of coccus slightly rugose owing to a few scattered spine or hair-bases forming short protuberances with a central hollow now filled with matrix or calcite. Length of coccus (slightly broken at apex), 21 mm. Estimated complete length about 30 mm.; breadth, 24 mm.; thickness in plane of locule, 10 mm. Seed not seen. REMARKS. The specimen is a mineral impregnated carbonaceous coccus together with its concave external mould preserved in a coarse-grained grey sandstone block full of lignite fragments from Cakeham Beds A. Owing to the coarseness of the matrix and the mineralization of the specimen fine details of structure are some- what obscure, but there is indubitable evidence of hair-bases both on the coccus and its cast. The large size of the specimen is a striking feature. The coccus fits the cast exactly so that no shrinking has occurred probably owing to infiltration of mineral matter into the carbonaceous substance. The complete length of about 30 mm. is in excess of a rare maximum of 26 mm. in Wetherellia variabilis Bower- bank (20 mm. is the published maximum) and 16 (allowing for distortion) or possibly 20 mm. in W. dixon. It is necessary in making comparison to remember that W. variabilis from the London Clay is always much pyritized, the cells being distended by infiltration of pyrites, hence the measurements given represent the maximum known for that species. On the other hand W. dixont, is carbonaceous, relatively unpyritized and usually much crushed and may easily have shrunk to half or one- third of its original dimensions. On these grounds the Cakeham fossil more resembles W. dixom than W. variabilis. The poorly preserved evidence of external spines also recalls W. dixomi. Against this relationship are two facts, first, that the specimen was clearly two-loculed when perfect and this character is seen in specimens of W. variabilis but not as yet in W. dixoni ; second, that as in W. variab- ilis the conspicuous or primary splitting has been septicidal, whereas in W. dixont it is frequently loculicidal. It is possible to account for this latter fact by early infiltration of calcite into a primary incipient loculicidal fissure which on crystallizing acted as a cement or in the case of W. variabilis from the London Clay by infiltration of pyrites, but if so it is curious that the cementing should have occurred in the loculicidal planes rather than between the closely adpressed septicidal surfaces. While awaiting fresh evidence it seems advisable to refer the specimen to Wetherellia sp. (W. dixoni?) bearing in mind the possibility that a distinct species may be represented, or that W. variabilis and W. dixon will prove to be inseparable given a sufficiently large range of material showing varying types of preservation. POST-YPRESIAN PLANT REMAINS 39 ? Wetherellia variabilis Bowerbank (Pi rx, figs 69) 1840 Wetherellia variabilis Bowerbank (pars), p. 84, pl. 12, figs. 1-5, 8-40. 1933 Wetherellia variabilis Bowerbank : Reid & Chandler, p. 251, pl. 9, figs. 7-22. A small coccus, flattened on the inner septicidal surface, angled on the outer face, along the plane of loculicidal splitting, and elliptical in outline was washed by Mr. Curry from the Twrritella Bed (see Table, p. 18). Loculicidal splitting has started. Typical coarse angular parenchyma is seen in the thickness of the wall. Length, 16 mm.; breadth, 6 mm. ; dorsiventral thickness, 4 mm. The specimen from the Lower Lutetian had collapsed on drying but was repaired by Mr. Curry. The flat ventral surface of the coccus suggests a two-carpelled fruit. The walls appear thick and there are no traces of external spines, but the specimen is in poor condition and they would probably not show even if present. While clearly a coccus of Wetherellia the specific relationship is doubtful but the characters recall W. variabilis rather than W. dixont. INCERTAE SEDIS Carpolithus curryi n. sp. (Pl iis ties. 70,71) Diacnosis. Fruit probably inferior with accrescent perianth, gamosepalous below, divided? at the apex. A syncarpous, three-loculed, loculicidal capsule, obovoid, many-seeded with axile placentation. Seeds probably in two rows in each locule borne horizontally, pointed at the hilar end, rounded at the distal end, curved? Length of fruit, 10-5 mm. ; estimated transverse diameter, 7 mm. HoLotyPe. V.41887. DeEscripTIon. Fruit: Probably inferior with accrescent perianth, gamosepalous below, possibly dividing into free broad perianth segments at the apex of the fruit and if so segments alternating with the locules and reaching to the extreme tip of the fruit. Surface much abraded, cell structure obscure. Fruit a syncarpous capsule, approximately obovoid with three slightly unequally developed loculicidal locules indicated by lobes with a median longitudinal line of dehiscence. Surface of fruit finely striate transversely, formed of small transversely aligned cells. Placentation axile on projecting placentas. Shallow transverse furrows on the external surface of a perfect carpel indicate numerous horizontal seeds piled one upon the other probably in two rows. These impressed outlines have a slight sig- moidal curve upwards towards the plane of dehiscence. The seeds must have been narrow and pointed at the hilar end, rounded at the distal end and flattened on their contiguous surfaces, and the form suggests a curved cavity orembryo. Length of fruit, 10°5 mm. ; estimated transverse diameter, 7 mm. 40 POST-YPRESIAN PLANT REMAINS REMARKS. One pyritized fruit from Fisher Bed 9 with embedded sand and glauconite. It shows the perianth on one side, imperfect at the apex but suggesting a free lobe alternating with two of the locules. The upper end of one locule with seed impressions is well displayed and the ventral side of another is exposed by the loss of the dorsal wall revealing projecting axile placentas. The relationship has not been discovered but the fruit would be recognizable if found again so that a specific name, Carpolithus curry1, in honour of the finder has been given. It is possible that it may belong to the Monocotyledones or to Theaceae but without further evidence of seed characters no closer determination is possible. Carpolithus sp. (Pl. rn, figs. 725 7) One valve of a bisymmetric endocarp, much abraded. Obovate in outline prob- ably with an organ at the narrow obliquely truncate end but the internal surface is too decayed to show details of the cells and their direction. Externally the margin at the broad end shows a few coarse teeth and the lateral margins show radial striations due to the alignment of obscure ill-preserved cells. A raised obovate area occupies most of the external surface but leaves a broad rim on the right. The raised area is ornamented by a rounded sigmoidal ridge which becomes thin and sharp near the base. Cell structure is obscure but a few equiaxial pits about 0-038 mm. are visible. Length of valve, 3-5 mm.; breadth, 2-8 mm.; thickness (i.e. of half fruit), 1mm. The relationship has not been traced. REFERENCES BANDULSKA, H. 1923. A Preliminary Paper on the Cuticular Structure of certain Dicotyle- donous and Coniferous Leaves from the Middle Eocene Flora of Bournemouth. J. Linn. Soc. Lond. (Bot.), 46 : 241-269, pls. 20, 21. BOWERBANK, J. S. 1840. A History of the Fossil Fruits and Seeds of the London Clay. 144 pp., 17 pls. London. CHANDLER, M. E. J. 1925. The Upper Eocene Flora of Hordle, Hants.,1. 32 pp., 4 pls. [Mon. Palaeont. Soc., London.] 1954. Some Upper Cretaceous and Eocene Fruits from Egypt. Bull. Brit. Mus. (Nat. Hist.) Geol., London, 2 : 147-187, pls. 10-16. 1957. The Oligocene Flora of the Bovey Tracey Lake Basin, Devonshire. Bull. Brit. Mus. (Nat. Hist.) Geol., London, 3 : 71-123, pls. 11-17. 1960. Plant Remains of the Hengistbury and Barton EOS Bull. Brit. Mus. (Nat. Hist.) Geol., London, 4 : 119-238, pls. 29-35. 1960a. The Lowey Tertiary Floras of Southern England, 1. Palaeocene Flovas. London Clay Flova (Supplement). Brit. Mus. (Nat. Hist.), London. Cookson, I. C. & Duican, S. L. 1951. Tertiary Araucariaceae from South-eastern Australia, with notes on living species. Aust. J. Sci. Res., Melbourne (B) 4 : 415-449, pls. 1-6. Curry, D. & Wispon, D. E. 1958. Geologists Association Guides. No. 14: Geology of the Southampton Avea including the Coast-Sections at Barton, Hants, and Bracklesham, Sussex. 16 pp., 2 text-figs. London. Dixon, F. 1850. The Geology and Fossils of the Tertiary and Cretaceous Formations of Sussex. xvi + 422 + xvipp., 40 pls. London. POST-YPRESIAN PLANT REMAINS 41 Dixon, F. 1878. The Geology and Fossils of the Tertiary and Cretaceous Formations of Sussex. xxiv + 469 pp., 28 pls. London. FIsHER, O. 1862. On the Bracklesham Beds of the Isle of Wight Basin. Quayt. J. Geol. Soc. Lond., 18 : 65-94, text-figs. 1-4. GARDNER, J. S. 1882. Suggestions for a revised Classification of the British Eocenes. Geol. Mag., Lond., 9 : 466-472. 1883-86. A Monograph of the British Eocene Flora, 2: Gymnospermae. 159 pp., 27 pls. {Mon. Palaeont. Soc., London. } —— 1886. Report on the Fossil Plants of the Tertiary and Secondary Beds of the United Kingdom. Rep. Brit. Ass., London, 1885 : 396-404, pls. 1-3. 1888. Report on the Higher Eocene Beds of the Isle of Wight. Rep. Brit. Ass., London, 1887 : 414-423, pls. 3-5. HEER, O. 1862. On Certain Fossil Plants from the Hemstead Beds of the Isle of Wight. Quart. J. Geol. Soc., Lond., 18 : 369-377, pl. 18. Marion, A. -F. 1888. Doliostrobus Sternbergii. Nouveau genre de coniféres fossiles tertiaires. Ann. Sci. géol., Paris, 20 : 1-20, pls. 1, 2. Rep, C. 1897. The Geology of the Country around Bognor. (Explanation of Sheet 332.) 12 pp., 14 figs. [Mem. Geol. Surv. U.K.] 1898. On Limnocarpus, a new Genus of Fossil Plants from the Tertiary Deposits of Hampshire. /. Linn. Soc. Lond. (Bot.), 33 : 464-466, illust. Reip, E. M. & CHANDLER, M. E. J. 1926. The Bembridge Flova. Catalogue of Cainozoic Plants in the Department of Geology, I. viii + 206 pp., 12 pls. Brit. Mus. (Nat. Hist.), London. 1933. The Flora of the London Clay. viii + 561 pp., 33 pls. Brit. Mus. (Nat. Hist.), London. RENDLE, A. B. 1894. Revision of the Genus Nipadites Bowerbank. J. Linn. Soc. Lond. (Bot.), 30 : 143-154, pls. 6, 7. WRIGLEY, A. & Davis, A. G. 1937. The Occurrence of Nummulites planulatus in England, with a Revised Correlation of the Strata containing it. Pvoc. Geol. Ass., Lond., 48 : 203-228, pls. 17, 18. GEOL. 5, 2. 4 PLATE % Araucarites selseyensis n. sp. Fic. 1. Holotype. Twig showing broad spreading sometimes laterally flattened leaves. xX 2°7. (V.41918.) Fic. 2. Part ot the ventral and dorsal cuticle of a leaf from the above. The dorsal cuticle shows the tip free of stomata except for a solitary stoma separated by a belt of ordinary epider- mal cells from the main left stomatal band. The leat has been completed by a diagram based on slide V.41918a when perfect. v1 and v2 the two halves of the ventral cuticle split longitud- inally show stomatal bands, sb, extending to the leaf tip. di and dz are the main dorsal sto- matal bands. x50. Fic. 3. Conescale, upper surface with apica lprocess buried in matrix, on back of holotype. x 255. Fic. 4. Same after process, p, was fully exposed. Impression of lower surface here exposed by collapse of scale substance. X 2°5. Fic. 5. Distal end of same. x 6:5. From laminated clays and lignite equal to Fisher Bed V of Whitecliff Bay (undated lignitic phase between Cuisian and Lutetian); East Wittering. Doliostrobus sp. (Avaucarites gurnardi Florin) Fic. 6. Scale for comparison. Apical process broken or buried, note greater breadth of scale. x2. (V.17514.) Bembridge Beds; Gurnard Bay, Isle of Wight. Bull. B.M. (N.H.) Geol. 5, 2 PASH: PLA TEMS Avraucarites selseyensis n. sp. Fic. 7. Twig, one of a pair of counterparts showing acicular leaves. 2°7. (V.41919.) Fic. 8. Part of ventral cuticle of a leat from the above near leaf tip showing narrow stomatal band, sb; m, the median band of ordinary epidermal cells. x50. (V.419194.) From grey clay (undated lignitic phase between Cuisian and Lutetian) ; East Wittering. Fic. 9. Tip of a leaf from a twig in Pl. 6, fig. 13. J, is the sharp lateral margin ; m, obscure median angle on dorsal surface flanked by unequal facets. Thinner ventral cuticle with broader stomatal band is seen at v, v. Denser dorsal cuticle at d, d. A line of stomata, ds, is seen on one dorsal facet. There is no evidence to show whether it was continuous with the dorsal stomatal band below. x50. (V.41920a.) Fic. 10. Distal end of leaf cuticle from same twig ; v, v, ventral cuticle with stomatal band ; 1, lateral angle of leaf between ventral and dorsal, d, cuticles; m, median angle of dorsal surface flanked by ordinary epidermal cells without stomata. 50. (V.41920b.) Fic. 11. Ventral cuticle of a broad leaf from same twig near junction, 7, with twig. It shows the broad median band, m, of ordinary epidermal cells and one broad stomatal band. Still attached is part of the dorsal cuticle, d. Stomata are present on this surface only below the level of the free leaf tip. x50. (V.41920¢.) Horizon? In rolled clay pebble thought to be Lutetian ; Bracklesham Bay. Fic. 12. A larger twig with spreading acicular leaves probably this species. (Now disinte- grated.) 1:8. Fisher Bed 21 or 22, opposite Medmerry Farm in beds with Nummulites vario- layius (Auversian). Bull. B.M. (N.H.) Geol. 5, 2 PLATE 5 alii ear f sl "ee all ( A = * ah 7 PLATE 6 Araucarites selseyensis n. sp. Fic. 13. Twig from which cuticles of broad and acicular leaves have been prepared. (Now disintegrated and represented by slides V.41920a-7.) Horizon? x2. From rolled block thought to be Lutetian ; Bracklesham Bay. Fic. 14. Right side of ventral cuticle near junction with twig, 7; s, stomatal band. x50. (V.41918D.) Fic. 15. Part of same to show stomata. The ordinary epidermal cells of the median band, m, show pits as rounded shadows. X100. (V.419184.) Fic. 16. Part of left ventral stomatal band of another leaf. Distortion in growth causes the stomata to be obliquely oriented. 100 approx. (V.41918c.) Fic. 17. Part of right stomatal band of same leaf with less obliquely distorted stomata. X 200, approx. Fic. 18. Dorsal cuticle from decurrent basal part of leaf showing broad “ lanes ”’ of ordinary epidermal cells between lines of stomata. x 100. (V.41918c.) Fic. 19. Part of same cuticle showing details of stomata and pits in epidermal cells. 200 approx. From laminated clays and lignite (undated lignitic phase between Cuisian and Lutetian) ; East Wittering. Bull. B.M. (N.H.) Geol. 5, 2 PLATE 6 poe avy , ve re. Nee A, 1c 6b. t= as = as - > K | te 7 ‘ » —_ POL nre * v an - o r. | 7 4 a} ' 2 | - *» PLATE 7 Limnocarpus forbesi (Heer) Fic. 20. Locule cast, side. s, projecting hilar end of seed. Germination valve on left. x15. (V.33813-) Fic. 21. Same, dorsiventral view, showing gaping aperture left by the germination valve. X15. From Cuisian (Lower Bracklesham) ; Whitecliff Bay, Isle of Wight. Fic. 22. Carbonaceous endocarp, side, showing st, style (imperfect) ; a, attachment and f, the conspicuous lateral foramen. x15. (V.41870.) Fic. 23. Another showing complete recurved style, st. Valveisinshadow atv. x15. (V.41871.) Fig. 24. Asmallerendocarp. x15. (V.41872.) From Fisher Bed 11 or 12 (Lutetian-Auversian boundary) ; Bracklesham Bay. Limnocarpus ? enoymis n. sp. Fic. 25. Holotype. Large endocarp with long style and tubercled surface. Gaping aperture, k, for fallen valve is in shadow. Long lateral foramen, f, is obscurely seen. 15. (V.41874.) Fic. 26. Another somewhat distorted endocarp, a, attachment. Aperture for valve at k, elongate lateral foramen at f. x15. (V.41875.) Fic. 27. A smaller endocarp, dorsilateral, with broad ridged valve. x15. (V.41876.) Fic. 28. Same, ventral, showing ridge and long style. x15. Fic. 29. Another with broken style, ventral. The lateral foramina flanking the ventral ridge at f are partly concealed by associated tubercles. 15. (V.41877.) Fic. 30. A very large endocarp, imperfect below. Lateral foramina concealed by tubercles are atf. X15. (V.41878.) Fic. 31. Interior of one half showing oblique ridge of condyle, c, corresponding with external foramina. Broken below. x15. (V.41879.) Fic. 32. A detached valve, side. x15. (V.41880.) Fic. 33. Another, dorsal. x15. (V.41881.) From Fisher Bed 11 or 12 (Lutetian-Auversian boundary) ; Bracklesham Bay. Caricoidea obscura Chandler Fic. 34. Fruit, side showing base truncated by hilar opening. x15. (V.41884.) Fisher Bed 11 or 12 (Lutetian-Auversian boundary) ; Bracklesham Bay. POTAMOGETONACEAE? Genus? Fic. 35. Stem lying in matrix with leaf at node. Also internal cast of bud, b. br, a ring of bracts from which a bud has fallen. x6°5. (V.41889.) Fic. 36. Much puckered stem bearing clay internal casts of buds at b, bb. Basal scars are clear and an embracing bract is seen above bb. X6°5. (V.41890.) Fic. 37. Internal cast of a large bud with basal scar on right below. Two obliquely oriented stem fragments are seen above and below the bud. 6:5. (V.41891.) Horizon? Rolled block with Avaucarites twig (V.41920) thought to be Lutetian ; Bracklesham Bay. PEATE Bull. B.M. (N.H.) Geol. 5, 2 “a PLATE 8 POTAMOGETONACEAE Genus ? Fic. 38. Branched jointed rhizome, a, b, c, indicate branches. x3. (V.41922.) Fisher Bed 21 (Auversian) ; Selsey. POTAMOGETONACEAE? Genus? Fic. 39. Striate stem showing two nodes with buds, 8, of varying size, some still attached. Another stem with a bud lies on the extreme left of clay fragment. Persistent scales, sc, once surrounded a bud (now gone). x6*5. (V.41892.) Fic. 40. Internal cast of detached bud showing large circular attachment scar. Impressions of overlapping scales are seen on sides beyond the scar. x6°5. (V.41893.) Fic. 41. Another bud cast tilted to show basal scar. Fine longitudinal striae diverge from scar over sides. X6°5. (V.41894.) Fic. 42. Opposite end of same showing collapsed condition at apex. X 65. Fic. 43. Another bud cast, side. Striate lateral surfaces are seen. Attachment scar is ob- scurely visible in profile at base. 15. (V.41895.) Fic. 44. A distorted bud showing rounded attachment scar and at the margins the serrate edges of scales in profile. 15. (V.41896.) Fic. 45. Same, opposite side showing serrate edge of a scale lying obliquely. x15. Fic. 46. External impression of a bud, side still surrounded by carbonaceous scales. A frag- ment of longitudinally striate stem is seen below. 6°5. (V.41897.) Fic. 47. Internal cast of a bud, 8, side, lying in a hollow in clay still flanked at base by car- bonaceous scales, s. Xx 6°5. (V.41898.) Fic. 48. Bud which appears to have begun to elongate. x15. (V.41899.) Horizon? Rolled block with Avaucamnites twig (V.41920) thought to be Lutetian ; Bracklesham Bay. PEATEs Bull. B.M. (N.H.) Geol. 5, 2 Py mye Ls lies io. PLATE 9 Nipa burtini (Brongniart) Fic. 49. Large carbonaceous fruit, broken transversely and partly encrusted above with pyrites. Attached by pyrites cement at w is teredo-bored wood. x 0°8. (V.35722.) Lutetian, probably from Fisher Bed 7 or 8; Bracklesham Bay. Pinus bowerbanki (Carruthers) Fic. 50. Cone, imperfect below, showing the narrow apophyses with marked transverse carina. X 0'9. (V.3313.) Bracklesham Bay. Bull. B.M. (N.H.) Geol. 5, 2 PLATE 9 PLATE tro Nipa burtini (Brongniart) Fic. 51. Flattened immature pyritized fruit. x1. (V.29214.) Fic. 52. Larger imperfect fruit incomplete at base and apex, showing in profile aperture for germination. Fibres of sarcocarp show on sides. XI. (V.40260.) From Fisher Bed 4 (Palate Bed) (Lutetian) ; Bracklesham Bay. Fic. 53. Minute immature fruit with exocarp preserved embedded in glauconitic ferruginous sand with abundant remains of Nummulites laevigatus. X2°7. (V.33732-) Lutetian, most probably from Fisher Bed 6; Bracklesham Bay. Laurocalyx sp. Fic. 54. Cupule, slightly distorted so that the rim of the cup is exposed all round. x 6:5. (V.41885.) Fic. 55. Same, opposite side. x 6:5. Loose on shore near Fisher Bed 7 (Lutetian) ; Bracklesham Bay. Wetherellia dixoni (Carruthers) Fic. 56. Holotype, fruit, apex. x2. (40509.) Fic. 57. Same, base, showing casts of spine bases as light coloured spots. x2. Fic. 58. Same fractured longitudinally. On left of axis is a septicidal surface with fibres from the axis; on the right a locule with impression of distorted pendulous seed. x 2°8. Fic. 59. Same showing oblique dorsiventral compression. Two locules are exposed. Passage of funicles from axis to seeds is shown. Spine bases in thickness of wall appear white through partial decay of pyrites casts. x 6°5. Fic. 60. A seven-carpelled fruit with marked alternating ribs and furrows. Incipient splitting is seen at the base. x2°8. (V.33827.) Fic. 61. Same apex. Spines represented by pyrites casts of their hollow bases worn down to surface level. x 2:8. Fic. 62. Same side, showing dorsiventral crushing. 2°7. Fic. 63. Same, split loculicidally. In right locule a minute immature pendulous seed is visible. There is a developed but distorted seed in the left locule. x 2:7. Loose on shore, Bracklesham Bay. PLATE 10 Bull. B.M. (N.H.) Geol. 5, 2 Ga uy sie PLATE 11 Wetherellia dixon (Carruthers) Fic. 64. Base of six-carpelled fruit, asymmetrically developed and obliquely compressed. Shows incipient splitting. 2:8. (V.33825.) Fic. 65. Same, apex. xX 2:8. Loose on shore, Bracklesham Bay. Wetherellia sp. (2 W. dixon (Carruthers)) Fic. 66. External cast of a coccus. The apex is truncated obliquely by the fracturing of the stone, casts of spines show as small white circular scars. X2°5. (V.35718.) Fic. 67. The carbonaceous coccus from the above (similarly truncated), external surface. The longitudinal angle marks the plane of loculicidal dehiscence. x 1-7. Fic. 68. Same, septicidal surface, showing axial fibres with a branch diverging over the right septicidal surface and itself branching again. x 1I°7. From Cakeham Beds A (Cuisian) ; opposite West Wittering Beacon. Wethevellia variabilis Bowerbank Fic. 69. Much broken coccus, inner septicidal surface showing median line of loculicidal dehiscence. 2°8. (V.41886.) Fisher Bed 2 (Turvitella Bed) (Lutetian) ; Bracklesham. Carpolithus curvyt n. sp. Fic. 70. Battered pyritized fruit, side. The carpel facing the camera has lost its outer wall so that the placentas, p, are exposed, showing attachment scars of numerous seeds. Carpel seen in profile on the left shows impressed through the walls the horizontally attached and oriented seeds within. Carpel on right irregularly broken. x 6:5. (V.41887.) Fic. 71. Same, apex. Locule on left in Fig. 70 is at /. Placentas in broken locule are at p. A perianth segment, s, is seen extending to the apex of the fruit over the third imperfect locule. x 6°8. From Fisher Bed 9 (Ostvea teneva Bed) (Lutetian) ; Bracklesham Bay. Carpolithus sp. Fic. 72. Valve of a fruit, decaying external surface. x6°5. (V.41888.) Fic. 73. Diminished outline drawing to elucidate features now obscured through decay. From Fisher Bed 21 (Hard Bed) (Auversian) ; Selsey. PEATE it Bull. B.M. (N.H.) Geol. 5, 2 NOTES ON EOCENE TARSIOIDS AND A REVISION OF SOME NECROLEMURINAE E. L. SIMONS BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 5 No. 3 LONDON : 1961 NOTES ON EOCENE TARSIOIDS AND A REVISION OF SOME NECROLEMURINAE BY ELWYN L. SIMONS Department of Human Anatomy, Oxford Pp. 43-69; Pls. 12-14; 3 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 5, No. 3 LONDON: 1961 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), «stituted in 1949, 1s issued in five series, corresponding to the Departments of the Museum, and an Historical senes. Parts will appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 5, No. 3 of the Geological (Palaeontological) series. © Trustees of the British Museum, 1961 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued March, 1961 Price Twelve Shillings INOTES:ON EOCENE TARSIOIDS AND A REVISION OF SOME NECROLEMURINAE By E. L. SIMONS CONTENTS Page Page I. Necrolemur antiquus, AN EOCENE IV. REVISION OF Nannopithex AND TARSIOID PRIMATE . ; . 40 ITS RELATIONSHIP TO OTHER : NECROLEMURINES . : . 61 II. ASSIGNMENT OF Pseudolovis TO THE NECROLEMURINAE 7 asa V. CONCLUSIONS ‘ é : . 67 III. THE ANTERIOR TEETH OF WES USSU) VUBISIE TE SIEIENSS : 102 Microchoerus AND Tarsius . . 57 VII. REFERENCES j ; : . 68 SYNOPSIS In this paper the problems of the relationship of European Eocene Necrolemurinae to each other and to contemporaneous and later Primates are considered. A re- affirmation of tarsioid affinities is made for this group, and it is ranked together with Tarsiinae in the family Tarsiidae. Evidence substantiating resemblances in dental formulae, tooth cusp patterns, auditory construction, and cranial osteology between tarsiines and necrolemurines is presented. Pseudoloris is assigned to the Necrolemurinae and the taxonomy of Nannopithex is revised. In conclusion, a number of dental similarities between necrolemurines, omomyids and North American anaptomorphids (as defined by Gazin, 1958) are evaluated. ABBREVIATIONS A.M.N.H., American Museum of Natural History, New York B.M.N.H., British Museum (Natural History) G.P.I.H., Geological and Paleontological Institute, University of Halle-Wittenburg Montauban, Natural History Museum, Montauban, France M.C.Z., Museum of Comparative Zoology, Harvard University Paris, National Museum of Natural History, Paris P.U., Princeton University GEOL. 5, 3. 6 40 NOTES ON EOCENE TARSIOIDS I. NECROLEMUR ANTIQUUS FILHOL, AN EOCENE TARSIOID PRIMATE The Eocene necrolemurine primate Necrolemur antiquus! from the Quercy phos- phorites and possibly of early Ludian age, has long been thought to be a typical tarsioid primate. Indeed, this form has been one of the primary bases for the hypothesis that many of the small Eocene Primates of Europe and North America can correctly be called tarsioid, in that its characters link it and thereby some other Eocene Primates, in particular the necrolemurines, with the living Tarsius. Gregory (1915 : 30) has given some of the reasons for this placement of the group as has Simpson (1940 : 198) for Pseudoloris, here assigned to the necrolemurines, and the case for affinities with Tarsius has been most recently expressed by Gazin (1958 : 99). In view of Hiirzeler’s conclusion (1948) that Necrolemur is a lemuroid primate the investigations presented in this paper were undertaken at the suggestion of Sir Wilfrid Le Gros Clark. The lemuroid interpretation. Hiirzeler (1946, 1948) stated that the structure of the skull of Necrolemur (principally the auditory region) as seen in Montauban 9, is lemuroid, and that this primate exhibited no special resemblance to Tarsius that could justify calling it a tarsioid. This opinion, which was in part suggested by the material examined by him, considerably affected subsequent thinking on primate taxonomy and it is advisable to consider here in some detail the construction and placement of the tympanic ring of Necrolemur together with the significance of this structure in the reconstruction of the phylogeny of Primates. Hill’s comments (1955 : 293) are fairly typical of the reaction to Hiirzeler’s observations. “ Hiirzeler has dissected the bulla of one specimen and finds, surprisingly, that the interior does not resemble that of Tarsius, but agrees rather with that in Malagasy lemurs and of Adapis. He disproves the conclusion of Stehlin that the annulus tympanicus (ectotympanic) is fused with the bulla, for he declares it to be quite freely suspended in the tympanic cavity. Moreover, it is not a simple ring, such as occurs in the fossil Malagasy types, but expanded into a plate-like form, attached below and laterally to the inner end of the osseous meatus.”’ Since Hiirzeler’s suggestion that Necrolemur is lemuroid there has been a tendency to maintain that a distinction between lemuroid and tarsioid types cannot be made among early Tertiary primate species. This view was expressed by Simpson (1955 : 438) as follows : “The majority of Eocene genera have at one time or another been con- sidered ‘ tarsioid.’ Re-study of some, especially those represented by skulls or other relatively good material has resulted in their transfer to the ‘ lemu- roids.’... Nannopithex, Necrolemur and Miucrochoerus (see Hiirzeler, 1948) 1Determinations on the cranial anatomy of Necrolemuy antiquus are facilitated by the relatively large number of skulls of this species which are available for study in the following European and North American museum collections: Basel number Q.H. 470, Montauban 9, Paris Museum (five skulls), British Museum M3747 and M4490. Museum of Comparative Zoology, Harvard University, M.C.Z. 8879, and at Princeton University, P.U. 11465. Ihave studied all these specimens excepting Montauban 9. Stehlin (1916) also mentioned two skulls in the collections of the Faculty of Sciences at Marseilles. NOTES ON EOCENE TARSIOIDS 47 are examples particularly to the point. Still the mass of ‘ Anaptomorphidae ’ (probably not a natural family even after recent removals) has continued to be labelled ‘tarsioid.’ The fact is that there is no convincing evidence that any early primate is more ‘ tarsioid’ than ‘lemuroid’ in natural affinities.” Simpson (following Hiirzeler’s assignment of the necrolemurines to the Lemuroidea) was correct in suggesting that the term tarsioid was no longer meaningful. Necro- lemur and its allies had been thought to have more resemblance to Tarszus than any other European or North American early Tertiary primate stock. It is possible to question on various grounds whether the Necrolemurinae should be called tarsioids, even if one believes that the evidence for their close association with the ancestral line of Tarsius is sound. It is not possible, however, to assign the necrolemurines to the Lemuriformes as Piveteau (1957 : 64~71) has done while leaving all the much less Tarsius-like North American “ anaptomorphs ”’ ranked with the Tarsiiformes. No other early Tertiary primate is more convincingly tarsioid than Necrolemur. ” MORPHOLOGY Ectotympanic.—Hiirzeler’s view is most clearly stated (1948 : 28) in a section beginning, “ Das Verhalten des Annulus tympanicus von Necrolemur entspricht prinzipiell jenem von Adapis...”. The accompanying Plates 12 and 13, of the annulus tympanicus of Necrolemur, B.M.N.H. M4490 (from Mouillac, France) and Plate 14 of Adapis parisiensis, B.M.N.H. M1345 (from Caylux, France), are sufficient to show that these two forms have little if any significant similarity in the position and relationships of the ectotympanic element. Stehlin (1916 : 1352) also stressed this difference remarking : “‘ Das Verhalten des Annulus tympanicus ist somit bei Necrolemur ein ganz anderes als bei Adapis’”’. In Adapis the annulus lies within the bulla, it is free, and there is no bony external auditory meatus whatever in specimens examined by me. In these regards the annulus tympanicus of Adapis corresponds closely to the situation of this bone in the great majority, if not all, of the living members of the Lemuriformes, e.g. Daubentonia and Lemur. It would, perhaps, be tedious to recite how often the free, ring-like tympanic has been given as a character of the Lemuriformes ; van der Klaauw (1931 : 12) listed about thirty page-references to this effect for various living and fossil lemuroids (including Adapis and Notharctus). In some living prosimians, other than Tarsius, such as Perodicticus and Lemur, a slightly developed bony external meatus has been reported, but it is not truly tubular (see van der Klaauw, 1931 : 155-156). Unlike any living lemur or loris, Necrolemur has a remarkably long bony external meatus (typically longer than that of Tarsius) which is fused to the bulla internally. The latter feature can clearly be determined by examination of the internal posterolateral part of the ventral bulla wall of Necrolemur in B.M.N.H. M4490. The conclusion that the structure applied and fused to the ventrolateral bulla wall (see Plates 12 and 13) is the inner end of an expanded and tubular annulus seems inescapable. It is not very difficult to reconcile this observation with the figures and discussion in Hiirzeler (1948). I interpret the scalloped areas seen in the region of the anterointernal margin of the ventral side of the annulus of Montauban 48 NOTES ON EOCENE TARSIOIDS FIG. 1a Cross sectional diagram of auditory bulla of Necrolemur_antiquus (after Hiirzeler, 1948, fig.29.) FIG, 1b Diagram ‘of cross section of auditor bulla of Necrolemur_antiquus (Brit. Mus. Nat. Hist. no. M 4490. FIG. 1c Cross sectional diagram of the auditory bulla of Tarsius spectrum (Princeton Univ.. ost. coll. no. 375.) P-Petroso/ B8-Bulla Cc-Corotid cana! At-Annulus tympanicus AM-Annulus Membrane Mae-Ext. auditory meatus Sg-Sguomosal S- strut joining tympanic ring ond bulla wall NOTES ON EOCENE TARSIOIDS 49 g (see his figure 27) as the bases of the struts which connect the expanded ring and bulla wall in this area. In Necrolemur the annulus consists of a thin lamina of bone partly fused to and partly composing the ventral wall of the bulla in the region of the internal end of the external bony meatus, turning up from the internal surface of the bulla along its mesiad margin. This internal free margin of the ring (the direct support of the tympanum) is, at intervals, further secured to the ventral bulla wall by struts (six are visible in B.M.N.H. M4490, four in M.C.Z. 8879 and one in P.U. 1165), most of which run parallel to the main axis of the bony external meatus. These struts do not occur in Tarsius or in any other primate described to date. Nevertheless, Necrolemur exhibits a real resemblance to Yarsius in the fusion of the ring with the bulla wall and in the lengthening of the bony meatus. In Tarsius spectrum a ridge projects from the inside of the ventral bulla wall running around the circumference of the ear drum upon which the tympanum is stretched. This, as in Necrolemur, is the internal free margin of the ring. The main difference from Necrolemur being that this ridge or margin is inclined more laterally. The conformation of the annulus tympanicus suggested for Necrolemur is indicated in Text-fig. 1, which compares a cross-section of the auditory bulla given in Hiirzeler (1948, fig. 29) with that of B.M.N.H. M-44090, and of Tarsius spectrum. As can be seen in Plate 13, it is possible to trace continuous bone from the left hand margin of the removed part of the bulla wall (at A) to the remaining fragment of the external bony meatus (at B). Evidently the plate-like area in the centre of this figure corresponds to the part of Montauban 9g referred to by Hiirzeler as the “‘ Annulus- Membran ”’. This structure is clearly ossified and would appear to be more reasonably considered as part of the expanded annulus. There is clearly a continuum of bone between the internal free margin of the annulus and osseous meatus. Possibly the expanded and fused ring in the ancestry of both Tarsius and Necrolemur arose initially by a spreading of the ossification centre of the ectotympanic through the annulus membrane and outward to form a tubular meatus. Whatever the origin, the end result is very similar in these two Primates. Hill (1955 : 9) has summed up the conventional view as to the significance of an expanded tympanic ring for primate taxonomy as follows :— “The tympanic bone [in ‘haplorhine’ primates] is very variable, but the specialized lemurine condition, where the annulus is enclosed within the bulla, never occurs. In tarsioids and catarrhines the ring is produced outwards in the form of a bony tube (osseus meatus), but in Platyrrhini the lorisoid condition persists, i.e. with the annulus exposed and contributing to the formation of the wall of the bulla (van Kampen, 1905) ”’. Hirzeler (1948 : 29) cites Lorenz von Liburnau (1905) as a source for statements about a tubular ectotympanic in Megaladapis but the latter’s comments are vague and capable of more than one interpretation (van der Klaauw, 1931 : 33). Of further interest in this regard is van der Klaauw’s discussion of the false external auditory meatus, or meatus spurius, formed when the glenoid region is shifted posteriorly and abuts against the mastoid (or post-tympanic process) thus constituting a ventral bridge between the bulla and external auditory opening. Such a squamosal- mastoid arch is sometimes present in Nycticebus and similar occurrences may account 50 NOTES ON EOCENE TARSIOIDS for some of the reports of expanded ectotympanics in Lorisiformes. The derivation of the bony external meatus in prosimians can be of several different kinds. In tupaiaoids (Dendrogale) an entotympanic meatus, in some lemurs and lorises a partial contribution from the squamosal (meatus spurius), and a true or tubular bony meatus, which in Tarsius, cercopithecoids and hominoids is derived from the ectotympanic. Comparison of forms such as Loris or Nycticebus with Tarsius or Necrolemur at once shows how different these two types of progressive prosimians are in this region. Auditory bullae of several well-preserved specimens of Megaladapis in the British Museum (Natural History) and in the American Museum of Natural History show that this genus exhibits a partial meatus spurius, but most sutures between the bones involved are closed. The arcuate bony attachment for the tympanum in Megaladapis (A.M.N.H. 30024) lies high on the dorsal face of a rather large bullar cavity (near the promentorium) and from it a sheet of bone fans out dorsolaterally, apparently an ossified annulus membrane. In complete contrast with Necrolemur the internal free margin of the ectotympanic extends through the spurius meatus surrounding it. However at the margin of the auditory aperture lies a partially separate, rugose bony rim. This rim is presumably the outer end of an ectotympanic tube (see Lorenz, 1905 : 463). Whatever the exact relationships with surrounding bones are, and these are obscured by closure of sutures, the ectotympanic of Megaladapis has no resemblance to that of Necrolemur. In fact, the appearance of the auditory region seems about as different in these two genera as could be expected for two mammals referable to the same order. Nor does the morphology of this part of the skull suggest any special affinity between them. Hiirzeler’s view, that the annulus and osseous meatus in Megaladapis and Necrolemur are similar, even if tenable, would hardly provide relevant information for the assignment of the latter to the Lemuroidea. Rather, it emphasizes an aberrant feature of Megaladapis, a form which (together with a few little-known allied genera) has long been known to exhibit many characters divergent from the remainder of Malagasy Primates (Hill 1953 : 655). On the basis of a single feature, the con- formation of the tympanic annulus, it cannot, of course, be asserted that similarities in its make-up relate Necrolemur and Tarsius, any more than one can say that a tubular construction of this bone relates Necrolemur and Megaladapis. If it were not for the many other characters shared by Necrolemur and Tarsius (see Table I, p. 52) ectotympanic resemblances between any two of the three could be ascribed to parallelism. Two conclusions suggested by the relationships discussed above are : 1) that Megaladapis in the region of the external auditory meatus diverges from both living Lemuroidea and from Necrolemur and 2) that Necrolemur exhibits a tubular osseous meatus not unlike that to be expected for an Eocene member of the particular stock from which Tarsius arose (Text-fig. I). Compared with living and post-Eocene Primates the structure of the ectotympanic of Necrolemur is more “ advanced’”’ than in any other known prosimian except Tarsius. Of course, demonstration that the centre of ossification of the external auditory meatus in Necrolemur lies in an embryonic horseshoe-like ectotympanic is not possible for this fossil species. Such a ring-like structure does precede the NOTES ON EOCENE TARSIOIDS 51 tubular conformation in Tarsius. Consequently in the latter, this manner of development justifies reference to the tubular structure as part of the ectotympanic. Although developmental history is not known for Necrolemur, there is certainly nothing in the conformation of the annulus of the known adults to prove that it did not arise much as in Tarsius. The only other fossil prosimian known to have a tubular bony meatus is Plesia- dapis. A skull of this primate recently described by Russell (1960 : 312-314) found near Cernay, in French continental late Paleocene beds, shows that this form has an extremely long osseous auditory meatus. Considering the very great differen- ces between Necrolemur and Plesiadapis in the remainder of their cranial anatomy, the conclusion that this feature in common is due to parallelism seems almost unavoidable. Moreover, instead of standing out from the skull base, as in Necro- lemur and Tarsius, the osseous meatus of Plesiadapis is sunk between the adjoining bones, as in Galeopithecus. Entotympanic and entocarotid circulation.—Hiirzeler (1948 : 29) was unable to find a suture between the petrosal and bulla in Necrolemur and implied that this ranked it definitely with Lemuroidea and not Tarsioidea. However, van der Klaauw (1931 : 285) indicates that the bulla arises in the latter group in the same manner as in lemurs. The presence of a separate centre of ossification for the bulla in tupaiaoids is now well established, but in the remaining Prosimii it apparently always arises from the petrosal. Consequently, the absence of a suture between bulla and petrosal does not distinguish lemuriform from tarsiiform Primates, and is equivocal here. It is, perhaps, not accurate to speak of prosimians, other than tupaiaoids, as having an entotympanic, but van der Klaauw (1931) considers that the entotympanic, when continuous with the bulla has simply lost its independence as an ossification centre. Gregory (1915 : 430) was able to deduce from the positions of the entrance and exit of the carotid artery on the bulla that Necrolemur resembled Tarsius and not other prosimians in entocarotid circulation. MHiirzeler’s illustrations (1948, figs. 30, 31) confirm Gregory’s suggestion of tarsioid affinities, by showing that his posited course for the entocarotid was correct and that the stapedial branch is smaller than the promentory artery, as in Tarsius. In Tupaia and those lemurs in which the carotid enters the bulla the reverse is the case (see Saban, 1957, fig. 36). There are, of course, the expected differences (between the Eocene and the living tarsiids) in the exact route of the entocarotid artery but such distinctions in Necrolemur do not approximate the Lemuroidea to any appreciable degree. The tarsioid characters of Necrolemur.—Table I lists those features shared by necrolemurines and Tarsius, particularly those not occurring or seen only sporadically in Recent and fossil Lemuriformes and Lorisiformes. The limitations of tabular expression prevent detailed citing of sources (and partial exceptions); other students have, and may again assess some of these characters differently. Although opinions vary as to individual features it does not seem possible to question any longer that the complex, or nexus, of shared osteological details in necrolemurines and tarsiines justifies their close taxonomic association. Particularly to the point in this regard are the remarks of Gregory (1920 : 193) on the amount of variability GEOL. 5, 3. 6§ 52 NOTES ON EOCENE TARSIOIDS in morphology allowable within a single mammalian family. Many cranial similarities between Tarsius and Necrolemur can be attributed to parallelism, but to explain all those listed (and others not amenable to tabular presentation) in this manner would be to assert the impossibility of distinguishing between real evolutionary affinities and parallelism. Such an agnostic attitude does not seem justified in this case. Moreover, in mammalian palaeontology it is almost always easier to produce an argument favouring separation of two given groups than it is to prove their valid TABLE I. A COMPARISON OF LEMUROID AND TARSIOID CHARACTERS SS IE ET ES ES SE RA ET TE SE AS SE TI Leal CHARACTER LEMUROIDS TARSIOIDS LEMURIFORMES LorRISIFORMES TARSIIFORMES (Lemuridae (Tarsiinae, and Adapidae) and Necrolemurinae) . Lower canine shorter than upper incisiform longer than upper iS) Zygomatic arches Postorbital opening (area) Brain case (transversely) Muzzle (length) Muzzle (width) Contact between jugal and lacrymal Ectotympanic (position) Ectotympanic (shape) Elongation of caleaneum and astragalus 11. Inflation of bullae 12. Bony canal of promontory artery 13. Median lacerate foramen 14. Carotid foramen COD WOR B Y 15. Septum between tympanic cavity and hyptotympanic sinus 16. Tibio-fibula 17. Palate 18. Tooth rows 19. Upper canine 20. Molar hypocone 21. Posterior nares (shape) 22. Position, posterior nares 23. Pterygoid alae 24. Bullae (position) 25. Mastoid region 26. Direction of foramen magnum 27. Posterior palatine torus 28. Contact of external pterygoid alae with bulla 29. Interfrontal suture or incisiform typically stout variable (flaring) (flaring) large very Narrow rather large expanded to expanded long typically broad typically occurs long to medium variable often occurs in bulla at margin of bulla annular annular but broad none some typically inflated small, in bulla variably present at posterolateral little inflation not ossified not in bulla present not present angle of bulla incomplete complete does not occur* broad anteriorly parallel (U-shaped) much larger than ant. incisor variable but often large broad anterior to M3 long anteroposteriorly well separated little or not inflated largely backward absent touching typically remains open in adults slender (close to skull) almost closed or small much expanded very short very narrow does not occur extends out of bulla tubular some much inflation large, in bulla not present on ventromedial face of bulla incomplete may occur in both subfamilies narrow anteriorly converge (V-shaped) smaller than ant. incisor often small narrow well behind M$ short anteroposteriorly approximated anteriorly inflated largely downward present broadly overlapping fused in adults * Characters following 15 are typical of both Lemuriformes and Lorisiformes. NOTES ON EOCENE TARSIOIDS 53 association. Fortunately, in the case of Necrolemur, preservation of almost every part of the skull allows for more precise judgements of affinity than is usually possible for a fossil primate. In summary, that Necrolemur stands close to Tarsius taxonomically is most strong- ly suggested by essential agreement in the following areas : entocarotid relationships within the bulla ; shape and position of ectotympanic ; position of the pterygoid wings and posterior nares ; presence of a posterior palatine torus ; likeness of relative sizes of teeth, tooth patterns, and arrangement of tooth rows ; fusion of frontals, and many lesser features (Table I) which differentiate them from living and fossil lemuriform and lorisiform prosimians. The post-cramial skeleton in necrolemurines.—Most features of the necrole- murine post-cranial skeleton reported to date need further confirmation. Schlosser (1907) assigned to Microchoerus edwards: and Necrolemur antiquus (without asso- ciation) several isolated limb-bones from the Quercy deposits. The calcanea figured by him resemble those of Primates, but there is always the possibility that the fused tibio-fibulae that he referred to Necrolemur actually belong to a rodent or insectivore. A fusion of this sort can occur in either order, and members of both occur in the Quercy phosphorites. The only valid association of post-cranial and cranial material in necrolemurines is reported by Weigelt (1933 : 138). His specimen includes most of the bones of a left hind limb, found together with a skull and mandibles here assigned to Nannopithex (p. 61). Weigelt claimed that the tibia and fibula were fused, but, on examination of the specimen, I can see no evidence for this. A small splinter of bone adjacent to the left tibia was regarded by him as a part of the fibula, but it does not contact the tibia. Also, the surface of the tibia is much worn and broken, has probably lost its original surface, and does not show any elevation for fibular attachment. Remaining parts of the pelvis and femur are not particularly like those of Tarsus, but the calcaneum, although broken, does show some elonga- tion. If tibio-fibular fusion occurs in necrolemurines it has yet to be convincingly demonstrated. Moreover, it seems unlikely that the specializations of the hind limb of Tarsius should be expected to have been attained by any Eocene Primates. About all that can be said in defence of the possibility of such fusion is that in hopping prosimians like galagos and the tarsier the foramen magnum is directed almost downward, and since it also has this position in Necrolemur some such speciali- zation of the hind limb might be expected. Relationships——The construction of the ectotympanic and the entocarotid circulation of Necrolemur are more like those of Tarsius than of any other known prosimian. Preservation of the bony canals of the promontory and stapedial arteries within the bulla of Necrolemur (Hiirzeler, 1948, figs. 30, 31), shows that this form here resembles Tarsius, tupaioids, and Catarrhine Primates, but not lemurs or lorises (See Le Gros Clark, 1959 : 151). In consideration of the many other similarities between these two genera, it seems advisable to rank the Necrolemurinae as a subfamily of the Tarsiidae. At present the case for associating the North American Anaptomorphidae (s.s.) with Tarsius is not nearly as sound, and rests largely on dental resemblances between necrolemurines and anaptomorphids. Only one partial skull (of Tetonius) is known for any of the numerous genera and species of North 54 NOTES ON EOCENE TARSIOIDS American “‘ tarsioids’”’, and in it much of the basicranium is missing. Therefore, it is not particularly sound to refer to any of these New World species as tarsioids, although some of them may eventually prove to be so. At present, only the Necrole- murinae are demonstrably tarsioid. For an early Tertiary primate Necrolemur is a progressive or advanced form, and in some features, such as the more complicated tooth patterns and more inflated mastoid region, less structurally primitive than Tarsius. Much of the overall morphology of the skull in Necrolemur is about as divergent from that of Eocene lemurs like Adapis and Pronycticebus as is that of the Recent tarsier. This pro- gressiveness of necrolemurines, among Eocene prosimians, requires emphasis here because others have erroneously reported degrees of primitiveness not actually to be seen in known early Tertiary Primates, i.e. the supposed presence of three upper incisors and posterior palatine fenestrae in Pseudoloris. These two characters are exhibited by some Insectivora but not in Primates. Had they really occurred in Pseudoloris they would represent interesting morphological relics of the insectivore grade in primate ancestry, but actually Pseudoloris (and other necrolemurines as well) is as devoid of such features as are Recent Prosimii. In Primates antiquity is not always synonymous with primitiveness. II. ASSIGNMENT OF PSEUDOLORIS TO THE NECROLEMURINAE In 1890, Filhol described a small primate from the late Eocene phosphorites of Quercy, France, which he called Necrolemur parvulus. Later Stehlin (1916 : 1397) proposed a separate genus for this species, Pseudoloris, and pointed out a number of distinctions in the dentition separating it from Necrolemur. The validity of the genus Pseudoloris has been accepted in subsequent publications. Teilhard reviewed the species in 1921, basing his comments largely on new and more complete specimens. He disagreed with Stehlin’s suggestion of an affinity between this Eocene primate and the living lorises, indicated in the generic name, and proposed that Pseudoloris not only had tarsioid affinities, but in some respects is more like the living Tarsius than any other Eocene primate. Simpson (1940: 198) also considered the small Quercy primate particularly Tarsius-like and established for the one species Pseudoloris parvulus a separate subfamily, Pseudolorisinae, which he tentatively assigned to the Anaptomorphidae (s./.) while remarking that “ this genus stands considerably closer to Tarsius than does any other known from the Paleocene or Eocene, and should perhaps be placed definitely in the Tarsiidae’”’. Simpson, how- ever, did not list any characteristics of the subfamily. In the course of recent studies on European Eocene Primates the writer has concluded that it is not possible to place Nannopithex and Pseudoloris in separate subfamilies. Since the former is clearly associated with Necrolemur in dental and cranial morphology and, through Necrolemur, with Microchoerus, it is advisable to formalize this association of four genera at the subfamily level. This subfamily has most often been called Necrole- murinae but the term Microchoerinae is also available. Due to greater currency and early establishment the former name is to be preferred. Discussion.—Simpson did not specify why Pseudoloris should not be included in the Necrolemurinae, Teilhard having previously pointed out that among European NOTES ON EOCENE TARSIOIDS 55 Primates, its closest resemblances are to species of Necrolemur. However, Simpson (1940 : 199, footnote) observed that Necrolemur may have been derived from a paromomyine, and Pseudoloris from an omomyine source, and so indicated a separa- tion in their lineages going back to the Paleocene. It is unlikely that Necrolemur could be derived from a paromomyine stem, because the anterior lower dentition is less reduced in number of teeth than in paromomyines (unless one assumes that the anterior enlarged tooth in Necrolemur is an incisor, see page 58). Even assuming that the teeth in question in Necrolemur are incisors, then the corresponding teeth in Pseudoloris are also most likely to be the same. In this regard these two genera stand together in the way in which they differ from, or, as is less probable, resemble the paromomyines. In a point for point comparison of tooth shape and pattern in Microchoerus, Necrolemur, Pseudoloris and Tarsius it is evident that there is a greater similarity between species of the latter two groups, but this agreement may equally well be due to a relative lack of specialization in tooth patterns of both, as to a phyletic connection between them not involving the other genera. Hiirzeler’s illustration (1948, fig. 4) shows that Nannopithex (at least in the characters of premolars and molars) approaches the dentition of Tarsius almost as closely as does that of Pseudoloris. Moreover, Nannopithex is certainly nearly related to Necrolemur, so that the Tarsiinae are linked by two separate types of intermediate dental morphologies with the Necrolemurinae. That Pseudoloris has one or more small incisors anterior to the large lower tooth, here considered the canine (and thereby differed from other necrolemurines) has been repeated in the literature on this primate, for instance by Hill (1955 : 313). Nevertheless, no teeth or alveoli have ever been observed in this position. The erroneous view that Pseudoloris possessed lower incisors seems to derive from Teil- hard (1921 : 6). However, in spite of suggesting that three upper incisors indicate the existence of lower incisors, Teilhard’s final statement was neutral, being only (?) for lower incisors in the dental formula he gave. Specimens of Pseudoloris now in the Basel and Paris collections indicate that the alveolar border of the man- dible (between the large anterior lower tooth and the symphyseal surface) is too thin to have supported a more anterior tooth. Moreover, this narrow band of bone never contains evidence of alveoli. The partial facial region of Pseudoloris (Montauban C) from which Teilhard concluded that there are three upper incisors in this small primate, and the only specimen of the species which preserves alveoli here indicates that the existence of an T' should be strongly questioned. Considering the highly eroded surface of the premaxilla the tiny spot, visible only on the right side, which Teilhard interpreted as the alveolus of an I? could equally well be the result of post-mortem damage. For instance, erosion of the original alveolar margin may have exposed a matrix filled cavity in the bone. If an I? does exist in Pseudoloris it constitutes a more primitive condition than occurs elsewhere among all Primates, with the possible exception of Nannopithex raabi in which the evidence on this point is equivocal. Furthermore, if the presence of upper incisors calls for lower ones as well, then this assumption (to mention just a few cases) would apply equally to species of Mzcro- 56 NOTES ON EOCENE TARSIOIDS choerus, Nannopithex, and Necrolemur. Although the latter have two pairs of incisors above, most of the included species in these genera do not have teeth anterior to the enlarged lower tooth. The retention of P, in necrolemurines although P4 is lost need not be anomalous because (unlike most Primates) members of this sub- family have rather small upper canines, which do not necessitate the formation of a diastema below for their reception. Mucrochoerus ornatus and Pseudoloris parvulus are at opposite extremes in this subfamily as regards dental patterns, but they are linked by a series of morphologically (not sequentially) intermediate species. If, as some authors have preferred, these genera are divided into three subfamilies, Microchoerinae—Necrolemurinae—Pseudolorisinae, the number of subfamilies approxi- mates to the number of genera involved and completely conceals their essential similarities. The proliferation of dental cuspules and crenulations observed in the late Eocene species of Necrolemur and Microchoerus, but lacking in Necrolemur zitteli, and the species of Nannopithex and Pseudoloris should not be given too much taxonomic weight. Complex molar patterns in Microchoerus erinaceus and M. ornatus particular- ly, were considered by some early authors as indicative of a separate subfamily or even family for Mzcrochoerus. Nevertheless, such changes might be due to the spread of very few or even single gene alleles which, as the temporal succession of species in the European Eocene suggests, might have appeared and been selected for rapidly. Apart from this, necrolemurine species, including Pseudoloris parvulus, are so similar in general conformation of details of tooth structure, dental formula, shape of the horzontal ramus, jaw angle, depth of the mandible, and position of the primary mental foramen, that it is hard to avoid the conclusion that they belong to a distinct taxon of closely related species and genera. Teilhard (1921) described and figured two large reniform fenestrae in the posterior palatal region of Pseudoloris. Such fenestrations do not occur in Tarsius, nor in other Primates, although some insectivores have openings here. However, in Necrolemur there is often an area just in front of the posterior palatine foramina where the palate is broken away. Tarsius also has a thin posterior palate, and in a much smaller and more delicate primate like Pseudoloris the probability that this part of the palatine should have been lost during fossilization or subsequent pre- paration is appreciable. Microscopic examination of the edges of these fenestrae shows no smooth margin but everywhere the rim of these holes is fractured and broken. Evidently these openings are not natural features but artifacts of prepara- tion. They can no longer be considered significant diagnostic features of the genus. Agreements between facial regions of the skull of Pseudoloris (Montauban C), of Microchoerus (Sedgwick Museum, Cambridge), and the more complete skulls of Nannopithex (G.P.I.H. 4236) and of Necrolemur include the following relatively diagnostic tarsioid features : I) comparatively large, anteriorly directed orbits ; 2) reduction of snout size and length; 3) narrowing of the interorbital septum ; 4) convergence of the tooth rows anteriorly ; and in the latter two genera, at least, 5) fusion of frontals in the mid-line. These shared cranial features corroborate dental resemblances discussed on page 61. To conclude, the similarities between Pseudoloris and remaining necrolemurines NOTES ON EOCENE TARSIOIDS 57 are so great that it should be assigned to this subfamily. It may be noted that at one time or another one or more species of all the genera here included in the Necrolemurinae (Necrolemur, Microchoerus, Nannopithex, Pseudoloris) have been assigned to the genus Necrolemur. Ill. THE ANTERIOR TEETH OF MICROCHOERUS AND TARSIUS Interpretation of dental formulae in Primates.—For two principal reasons identifica- tion of the number and the kind of teeth in early Tertiary Primates is of basic importance in reconstructing primate evolutionary history. First, teeth are the most commonly preserved parts of these animals, and, second, once a tooth is lost from the series it cannot be reproduced again as such. From this latter it follows that species postulated to lie on or near the line of ancestry of a given form must have the same or a greater number of each kind of tooth than does a supposed descendant. An example is provided by the Paleocene and Eocene Plestadapis. Latest known species of this genus have lost canines, first and second premolars (above and below) and a pair of lower incisors,—eight teeth in all. Since all living Primates, with the exception of the Madagascan Aye-Aye, Daubentonia, possess at least some of these teeth, Plesiadapis is thus eliminated from the near vicinity of any of their ancestral lines (and from that of Daubentoma on other, and very con- vincing, grounds ; Simpson 1935). Huxley (1876) termed such forms (which cannot have given rise to later stages of a given group) zutercalary types. From these, he distinguished linear forms, which in their structure admit of the possibility of being ancestral to subsequent groups, even when the lack of intermediate connecting links makes absolute demonstration of such ancestor-descendant relationships impossible (Le Gros Clark, 1959: 48). Although Huxley’s terms have not received general acceptance they serve to emphasize the essential problem to be considered here,— whether the dental conformation in necrolemurines and tarsiines indicate intercalary or linear relationships between the two subfamilies. A number of procedures, derived from comparative studies of all known forms govern the identification of teeth in Primates. Upper incisors in Primates, as in all mammals are always situated in the premaxilla. The upper canine, unless lost, is the most anterior maxillary tooth. When present, the lower canine is situated in front of the upper and shears against its anterior margin. Thus, the lower caniniform tooth in lemurs and lorises can be determined as the P, (despite a canine-like appearance), since its anterior edge shears against the posterior face of the upper canine. In these Prosimii the true canine has become incisiform and inclined forward, forming part of the tooth-comb apparatus. Apparently P, cannot shift forward so as to shear against the anterior face or blade of the upper canine and this makes it possible to distinguish P, from the canine in all cases. Identification of the anterior enlarged lower tooth in necrolemurines is not so simple, and the different assignments given this tooth have complicated discussion of possible relationships between the anterior lower teeth of necrolemurines and tarsiines. For many years there have existed two conflicting interpretations of lower dental formula in Necrolemurinae, one affirmed by Stehlin (1916) for Necrolemur 0.1.4.3., favoured by Hiirzeler (1948) and by Hill (1955) and another presented by 58 NOTES ON EOCENE TARSIOIDS Gregory (I9I5), I.1.3.3., and avowed by Simpson (1937, and later). What has been obscured by this difference of opinion is that necrolemurines typically lack a tooth in front of the enlarged tooth and behind it have seven teeth. Consequently, no matter which of the alternative dental formulae is correct, all but one species of necrolemurines differ from, or resemble, Tarsius in the same way. This is not re- flected by the taxonomic key for the group presented by Hill (1955 : 292) in which the procumbent anterior tooth is regarded as a canine in Necrolemur and Nanno- pithex and as an incisor in Microchoerus. Differences between anterior lower dental formulae in necrolemurines and Tarsius require elucidation in view of the cranial resemblance of both subfamilies. More- over, the upper dental formulae of all species of both groups are apparently the same (2.1.3.3.), aS are the sizes of the teeth relative to each other (which may be more important). For example, all tarsiids have an I? that is larger and longer than the upper canine. This is the reverse of the condition in all notharctines, adapines, living lemurs and lorises, in which the upper canine is much larger and longer than I?. A definite assignment as to type of the anterior tooth in the lower dental series of necrolemurines remains difficult. If it is unlikely, as Simpson (1937 : 146) remarked, that a lower canine can occlude in the position it does here, then it would appear that the lower enlarged tooth in all these European Eocene tarsioids, in- cluding Pseudoloris, is an incisor and that the lower canine is the reduced and premolariform tooth immediately behind it. Alternatively, the canine may be supposed to have been lost and four lower premolars retained. The presence of a small P, consistently in Tezdlhardina, situated somewhat to the outside of the midline of the tooth row (as in Necrolemurinae) seems significant here (in view of the other dental resemblances between Tezlhardina and necrolemurines noted by Hiirzeler, 1948). In Teilhardina the lower dental formula is 2.1.4.3. Because of this, it is no longer possible to say that there are no known tarsioids of the early Tertiary which retain four premolars (Simpson 1937 : 146). As Hiirzeler noted the Necrolemurinae, on the basis of dental evidence, probably derive from an Old World omomyid like Teilhardina and represent a phylum in which progressive reduction of the lower incisors took place. The criticism can be made that the assumption of total loss of lower incisors in typical necrolemurines has no known parallel among other Primates. This is correct, but the loss of all but a single lower pair of incisors is more common in Primates than has been previously stressed. Apart from Tarsius and one species of Microchoerus this condition also obtains in Parapithecus, in the adapine genera Caenopithecus and Protoadapis, perhaps in the notharctine Pelycodus, and in indrisines including Indri and Propithecus. Such rather broad distribution of lower incisor reduction among both living and fossil tarsioids and lemuroids indicates that parallelistic tendencies for tooth loss here are strong among Primates, and make the total loss of lower incisors in late necrolemurines appear more plausable than the assumption that the large lower anterior tooth is not a canine but an incisor, as in plesiadapids and Phenacolemur. 1The most anterior upper incisor in necrolemurines is here arbitrarily considered 12. NOTES ON EOCENE TARSIOIDS 59 Should the enlarged anterior tooth be an incisor in Necrolemurinae then all members of this group stand distinctly apart from Tarsius, and it could not have descended from a species even distantly resembling necrolemurines in this regard. However, an alternative interpretation is suggested by specimens of Microchoerus erinaceus and M. edwards: in the British Museum (Natural History) which is much more in agreement with the tarsier dental formula. The British Museum specimens.—The lower dental formula of Tarsius is generally assumed to be 1.1.3.3. From front to back in the mandible these teeth are identified as, first, a small, pointed incisor, followed by a large and long canine, three premolars (increasing in size posteriorly), and three simply constructed molars, of which M, is the largest. In necrolemurines, on the other hand, the lower dental formula (assuming the anterior tooth is a C) is written 0.1.4.3. However, an exception to this formula apparently occurs in one necrolemurine species. Two specimens of Mzcro- choerus edwardsi in the British Museum show a small alveolus anterior to the enlarged procumbent tooth, giving a lower dental formula of 1.1.4.3. This is further confirmed by a specimen of the same species in the Paris Museum (identified as Bach, lot 1893-11). Other necrolemurines do not show this small anterior alveolus. However such a possible identification of the lower dental formula of Microchoerus edwardsi (because that of Tarsius could be derived from it) indicates the need for a more certain assignment of the large procumbent lower tooth in necrolemurines. Some clarification of this point is provided by the following specimen of Microchoerus in the British Museum (Natural History). The holotype of Muicrochoerus erinaceus Wood, B.M.N.H. 25229 (Text-fig. 2) includes the entire upper dentition (on one side or the other) and an associated mandible with P;-M;. The anterior half of this jaw (with all teeth) was part of the original find, by Wood (1846). Cooper (1910) noted that the anterior portion of this mandible had been lost. However, the series of specimens of Micro- choerus erinaceus in the British Museum (Natural History) from the late Eocene deposits near Hordle, Hampshire and from various localities in the Isle of Wight is extensive and includes several isolated lower canines, and a number of anterior parts of mandibles with premolars. These specimens of Microchoerus erinaceus are also supplemented by a number of jaws and a fragmentary facial region from Hordle in the Sedgwick Museum, Cambridge. The occlusal diagram shown here (Text- fig. 2) is based on the holotype, supplemented by the others mentioned above and by Wood’s figure. The holotype has well-preserved wear facets on the three anterior upper teeth, I?, I?, and C. These surfaces of wear indicate that the tip of the lower canine could be extended anteriorly to occlude against I?, presumably when the animal was nibbling, but that when the jaws were shut its apex wore against I° and its base against the upper canine (see Text fig 13b). In Tarsius, the lower canine (which no one seems to doubt is a canine) can likewise occlude against these same three upper teeth. In Tarsius there is not the same degree of procumbency as in Mzicrochoerus, so that wear on these three upper teeth in the living species, takes a rather different course. However, relative sizes of the teeth involved, and their general position in the jaw is similar in both genera, and upper dental formulae the same. 60 NOTES ON EOCENE TARSIOIDS FIG.2a Upper right I*-C: C-1°: Cross hatched—wear surface sloping postero-lingually made by lower canine when teeth are fully occluded. I?: Cross hatched(fine) internal wear surface made by lower canine when nibbling \° Dotted lines-external wear surface made when lower canines are swung laterally during nibbling. FIG.2s Left anterior dentition (fully occluded ). i E ae Fic. 2. Microchoerus evinaceus. Wood. Diagrams of anterior wear pattern and occlusal relationships. (B.M.N.H. 25229). NOTES ON EOCENE TARSIOIDS 61 A reasonable assumption would be that a generalized necrolemurine, with a lower dental formula identical to that of Mucrochoerus edwards: could have given rise to that of Tarsius.1 All that would be needed to achieve this is loss of P, (already non-functional in necrolemurines) and reduction of the degree of procum- bency of the anterior lower teeth. If, in spite of the foregoing analysis, anyone still prefers the idea that the enlarged anterior tooth below (in this subfamily) is an incisor, then it would be equally logical to conclude that the large second tooth in the lower dentition of Tarsius is also an incisor. It is known that the P, in Primates can become caniniform, but apparently no one has yet suggested that incisors might do this also. One reservation regarding lower dental formulae in this group should be made. Because of the great delicacy of the anterior alveolar border of the mandibles in these small mammals, and consequent breakage, it is difficult to be certain that they could not have had teeth anterior to the lower canine. Nevertheless, I have not found alveoli in this position except in Microchoerus edwardsi. Dental formulae in remaining necrolemurines may be written as follows :— Microchoerus erinaceus Pseudoloris parvulus BAe Qe ss ia tae ey 0.1.4.3. On4.3: Necrolemur antiquus Nannopithex raabr PeT 235 Bete (Ok Ale Se 0.14.3: 0.1.4. Be As will be seen from the above formulae it is fairly certain that all four species have the same number and kind of teeth, identical with the living Tarsius above, but differing from the latter in lacking the single pair of lower incisors, and in possessing a vestigial P,. IV. REVISION OF NANNOPITHEX AND ITS RELATIONSHIP TO OTHER NECROLEMURINES Taxonomic position. An examination of the type of “ Psuedoloris abderhaldini ’”’ Weigelt (1933) indicates conclusively that it is conspecific with “ Necrolemur”’ vaabi Heller (1930). The later species is referable to Nannopithex. Weigelt was perplexed by the recovery of a skull different from Necrolemur (G.P.1.H. 4236) associated with lower jaws that were remarkably like those of ‘‘ Necrolemur’’ raabi. These mandibles were illustrated by Weigelt (1933, pl. 4, figs. 2A, B), who quite correctly observed that this upper dentition could not belong to a species of Necro- lemur. He therefore described the specimen as a new species of Pseudoloris. This assignment, although still not correct, was an improvement. In the vicinity of this \Microchoerus edwardsi appears to be disqualified as a direct forerunner of the tarsiers by possession of complexities in cheek tooth patterns not seen in Tarsius. 62 NOTES ON EOCENE TARSIOIDS specimen were also collected elements of a left hind limb of a primate. These Weigelt believed could be assigned with high probability to the same individual as the skull (G.P.I.H. 4236). As such, they comprise the oldest partial skeleton of a tarsioid. Published observations on the Geiseltal species of Nannopithex are somewhat confusing because of the early misidentification of the generic position of the Brown Coal species and it is necessary to outline the successive views held regarding it and closely allied forms. Simpson (1940 : 199) questioned the taxonomic placement of the “ two”’ species and first suggested that evidence for a specific separation of “ Pseudoloris abderhaldini”’ and “ Necrolemur’’ raabi is dubious. In an attempt to separate the two supposed species, Weigelt went to great lengths to set out some consistent differences in their mandibular dentitions. These dis- tinctions are not convincing. It appears that the individuals called “‘ Necrolemur ”’ vaabt are in most respects slightly larger and have a deeper horizontal ramus of the mandible than has “ Pseudoloris abderhaldim”’. There does not seem to be any distinction between the two “ types ”’ that could not be due to sexual dimorphism or to population variation within one species. The partial skeleton of “ Pseudoloris abderhaldim ’’ was recovered from the Leonhardt Mine locality, while all the individuals of ‘‘ Necrolemur”’ raabi appear to be from the Cecilie Mine. It is unlikely that the fossil-bearing strata in the two mines are of exactly the same age, but no appreciable time interval can be assumed. Not knowing the upper dentition, Heller (1930) had some basis for placing “ Necrolemur”’ raabi in the genus Necrolemur, even though later evidence proves this determination to be wrong. His assignment was plausible at that time because of the strong similarity in lower tooth structure between the German form and some of the less specialized necrolemurines such as ‘‘ Necrolemur”’ filholi. In defence of Weigelt’s placement of the species he described, it can be said that the upper dentition of ‘‘ Pseudoloris abderhaldini’’ is superficially like that of Pseudoloris parvulus. This is mainly because, in dental patterns, they are both of a rather generalized tarsioid type. However, this similarity is not profound, and the lower dentitions of the two species are quite different. Hiirzeler (1948) tentatively equates Necrolemur filholi Chantre & Gaillard (1897) with Nannopithex pollicaris Stehlin (1912). He appears to hesitate somewhat on the question of whether or not to retain the genus Nannopithex for this form, but the identity of the two species seems to be beyond question. Although Nannopithex filholi has much simpler cusp patterns than Necrolemur antiquus the two are linked morphologically by Necrolemur zitteli, as is convincingly demonstrated by Hurzeler (1948). Hiirzeler’s elucidation of the dental characters of Nannopithex filholi has provided the key to the correct assignment of the Brown Coal necrolemurine. Comparisons of the two species indicate that they are congeneric. The evidence (given below) for a specific distinction between them is not extensive, but at present it seems better to retain vaabi for the Geiselthal form. Both species of Nannopithex appear to be of early Middle Eocene age. Faunal correlations indicate that species of this genus occur in Lower or Middle Lutetian deposits. At present, specimens of Nannopithex NOTES ON EOCENE TARSIOIDS 63 have been recovered near Egerkingen in Switzerland, at Buchsweiler in Alsace, and in the Geisel valley near Halle, Germany. Nannopithex does not occur in the Quercy Fic. 3. A. Pseudolovis parvulus, based onmandibles in the Paris Museum, x 8 approx. B. Nannopithex vaabi, based on specimens at the Geological and Paleonto- logical Institute of Halle University, particularly G.P.I.H. 4254 and 4255 x 8 approx. phosphorites of Bartonian and Ludian age. Its absence at this later period reinforces the idea based on morphology that a species of this genus could have given rise to Necrolemur, 64 NOTES ON EOCENE TARSIOIDS Genus NANNOPITHEX Stehlin, 1916 Type SpeciES.—Nannopithex policaris Stehlin, (=Necrolemur filholt Chantre & Gaillard). 21-85 (Ol4) 3. GENERIC CHARACTERS..—+ Upper MM differs from that of Pseudoloris Oni: An, 3: in presence of nannopithex-fold running from posterior crest of protocone toward hypocone. Hypocone of M1 comparatively smaller than in Pseudoloris ; external median indentation between paracone and metacone of M!° less pronounced, and metacones of M1’ relatively smaller and lower than in Pseudolorts ; all teeth with some crenulation of enamel suggestive of teeth of Necrolemur but unlike those of Pseudoloris. P,, compared to lower molars, much larger, and molar cusps more rounded and less high than in Pseudoloris ; molar paraconids more distinct than in Pseudoloris ; M, entoconid and hypoconulid not as high or as sharply set off as in Pseudoloris. (Known species slightly smaller than Tarsius spectrum or Necrolemur zitte, somewhat larger than Pseudoloris parvulus.) Discussion.—Considering that Pseudoloris and Nannopithex possess very simple and unspecialized dental patterns the differences between them are marked, being about as great as those which separate either from the living Tarsius. However, both show distinct affinities with Necrolemur and with Microchoerus as well as with Tarsius. Insofar as dental characters go (and osteological features do not contradict them) no basis whatever remains for placing these five genera in several different families and sub-families, as has been done in the past. The skull of Nannopithex raabi from Halle retains the entire upper dental series on one side or the other (except the canines) and this indicates the dental formula given above. Weigelt (1933 : 129) suggested an upper dental formula of 2.?.3.3. for the Geiseltal species, but his discussion, figures, and (much better) the specimen itself, indicate another possibility, 2.1.4.3., (Weigelt, 1933, pl. 5, figs. 1, 4, 5) for the upper dentition of Nannopithex raabi (G.P.I.H. 4236). This photograph shows that the most anterior tooth preserved in position is much smaller than that just posterior to it which is clearly P?. Although there is some breakage in the specimen, microscopic examination confirms that the seven teeth of the upper left side are an unbroken series. The four anterior teeth make up a sequence increasing in size posteriorly, and (considering what is known of the anterior upper teeth in Micro- choerus and Necrolemur, and of these teeth in European Eocene Primates generally) they can hardly be other than the series P!™. Such an assumption is particularly necessary, if, as Hiirzeler has suggested, a species of Nannopithex is to be considered ancestral to Necrolemur antiquus. Otherwise, it is required to assume that in the line leading through Nannopithex to Necrolemur the upper canine was first reduced (from a primitive large size, presumed characteristic of ancestral Primates) to a size smaller than the P? and subsequent to this re-acquired a size larger than that which the P? has in Pseudoloris, Microchoerus and (as is more important) in Necrolemur. It is more likely, therefore, that the canines have been lost from this specimen. Weigelt is apparently correct in his identification of the two loose teeth found in the region of the anterior extremity of this skull, The larger and most anterior tooth NOTES ON EOCENE TARSIOIDS 65 he considered I?, a smaller one I’. These teeth resemble 17? of Microchoerus. Should the dental series be complete in G.P.I.H. 4236 then the smaller tooth anterior to the P? would have to be considered a canine, and the D. F. would be 2.1.3.3. above, as in Muicrochoerus and Necrolemur. Even so, this is a different formula from that given by Weigelt which is one tooth short by his own account. The presence of a small upper P? would correlate well with the dental formula expected for a necrolemurine ancestor. NANNOPITHEX RAABI (Heller) (Text-figure 3) 1930. Necrolemuy vaabi Heller, p. 35, pl. 5, figs. 5, 6. 1933. Pseudoloris abderhaldini Weigelt, p. 128, pl. 3, fig. 5; pl. 4, figs. 1-3; pl. 5, figs. 1-5; pl. 9, figs. 4, 5 ; text-figs. 1-3. MATERIAL.—G.P.I.H. 4254-4257, mandibles, and G.P.I.H. 4236, skull, with associated mandibles and hind limb. HoRIZON AND LOCALITY.—Gesieltal Brown Coals, Middle Eocene; near Halle am Saale, Germany,—from Leonhardt and Cecilie mines. DrAGNOosIs.—Size : about as in Nannopithex filholi or slightly larger. Protocones of P*4 somewhat more distinctly set off than in Nannopithex filholi, P® slightly longer antero-posteriorly than in the latter species ; nannopithex-fold on M2 less distinct than in Nannopithex filholi. Mandible : M* hypoconulid slightly more distinct than in most specimens of Nannopithex filholi but enamel crenulations not as pro- nounced as in latter. Discussion.—The lower teeth of Nannopithex raabi approach more closely the dentition of Nannofpithex filholi from Buchsweiler, in Alsace at Basel, Bchs. 647, than they do the specimen from Egerkingen, Basel Eh. 601,—particularly in the conformation of the M, talonid. Otherwise, except for the characters mentioned in the diagnosis above, the dentitions of the various specimens of Nannopithex illu- strated by Hiirzeler (1948) are not very distinct from the Geiseltal specimens. Pending a better understanding of faunal correlations and, as long as so few really complete individuals are known, it seems better to retain two species for the genus Nannopithex. Perhaps the greatest significance of the Geiseltal finds of Nannopithex is the information supplied regarding early primate anatomy by the comparatively complete skull, mandibles, and hind limb found there. These remains make up the oldest known partial skeleton of an Old World primate. As a consequence of their significance, these remains were discussed in some detail by Weigelt, and will not be repeated here. There are, however, a few further comments of value to be made about the specimen. Although the skull is crushed, it is evident that the orbits in this primate were very large. Most of the posterior wall of the orbit can be seen on the left side and it stretches from the lateral extremity of the skull to the mid-line (Weigelt, 1933, pl. 3, fig. 1). On the left side of the dorsal surface of the frontal the supraorbital margin and the juncture with the zygomatic arch is preserved, This region is very similar 66 NOTES ON EOCENE TARSIOIDS in appearance to the corresponding area of Necrolemur skulls and, as in the latter, the interfrontal suture appears to be closed. Most of the characters of the basicranium are indeterminate, but in the region of the left external auditory meatus are numerous fragments of cancellous bone, which are presumably remnants of air cells of an inflated mastoid region, as in Necrolemur. The appearance of the de- tached back of the palate also agrees with that of Necrolemur antiquus. As already mentioned (p. 53) the hind limb figured by Weigelt provides no evidence that this primate had distal fusion of tibia and fibula, which is commonly supposed to be the case in Necrolemurinae. Simpson (1940 : 196) expressed the opinion that there is a general resemblance between the hind limb of the Geiseltal tarsioid and that of Hemzacodon, insofar as he was able to determine without seeing the originals. Examination indicates that, as he then suggested, most of the features of difference from Hemiacodon apparent in the Halle specimen seem to be the result of crushing. MEASUREMENTS (in. mm.) NANNOPITHEX RAABI (Specimens at the Geological and Paleontological Institute of the Martin-Luther University of Halle-Wittenberg, Halle-am-Saale, Germany). G.P.I.H. 4236, Leonhardt Mine, holotype of “ Pseudoloris abderhaldini ”’ Anteroposterior Transverse Maxilla : diameter diameter Pt Tear 0-7 Re 18 I'l Ps 7 2:0 ipo 15 2°3 M? I°9 2°5 M? 18 CAG | M3 I°4 2:2, Anteroposterior diameter p! through M?,—10: mm. G.P.I.H. 4255, Cecilie Mine. Mandible :— Anteroposterior diameter C through M,,—11-7 mm. Anteroposterior diameter C through P,,—6-8 mm. Anteroposterior diameter M, through M;,—4-9 mm. NANNOPITHEX FILHOLI (Chantre & Gaillard) Discussion.—Inasmuch as the studies undertaken by the writer are in agreement with Hiirzeler’s tentative equation of Nannopithex pollicaris and Necrolemur filholi this species takes the name Nannopithex filholi, as discussed above. Hiirzeler (1948) figured and reviewed amply all known materials of this species. Consequently, it will not be re-diagnosed here. One possible difference between Stehlin’s type species and the mandibles described by Chantre & Gaillard is that the P, of the former NOTES ON EOCENE TARSIOIDS 67 specimen does not have the roots as well separated as in typical Nannopithex. However, this character is variable in individual specimens of Necrolemur, and may not have much taxonomic value. V. CONCLUSIONS The findings reported here substantiate the position that tarsioids can be dis- tinguished from lemuroids as far back as the Middle Eocene. Neither the Middle Eocene Nannopithex, nor the better-known genus Necrolemur provide any anato- mical basis for a horizontal grouping of the Necrolemurinae with contemporary Eocene lemuroid Primates. Consequently, the Necrolemurinae have been ranked under the Tarsiidae. Tarsioid claims for any other early Tertiary primate stocks are hard to support on the basis of present evidence. Some or all of the forms now ranked in the Omomyidae and Anaptomorphidae, may eventually prove to be so (with the advent of new and better specimens), but demonstrating tarsioid characters—ultimately derived from the living genus—becomes increasingly difficult with greater antiquity. Because of this it is possible that the direct ancestry of the Necrolemurinae might not have been definable as tarsioids much before the early Eocene. Probably Teilhardina, an Omomys-like primate from the Belgian early Eocene (Sparnacian), bears a relationship to the stock from which the necrolemurines arose, and it in turn suggests an omomyid derivation for these tarsioids. A complex of characters including large anterior lower teeth, a tendency toward reduction of tooth 4 ; : number, frequent occurrence of a large P| (non-molarized), nannopithex-fold and a number of other features suggests that the anaptomorphids (s.s.) are also a closely related stock. The current narrowly drawn higher categories of Paleocene and Eocene primates tend to obscure broad synthetic similarities among a number of stocks, but without better material there is little basis for grouping together any of such currently proposed early Tertiary primate families as Adapidae, Omomyidae, Anaptomorphidae, Paromomyidae (including Phenacolemurinae), and Tarsiidae (including Necrolemurinae). VI. ACKNOWLEDGEMENTS I should like to thank Dr. K. P. Oakley of the British Museum (Natural History), Dr. J- P. Lehman of the National Museum of Natural History, in Paris, Dr: A. S. Romer of the Museum of Comparative Zoology at Harvard University, Dr. G.. L. Jepsen of Princeton University, and Dr. G. Krumbiegel of the Geological and Paleontological Institute of the Martin-Luther University of Halle-Wittenberg for the opportunity to study and report on necrolemurine specimens in their respective collections. Thanks are also due to Sir Wilfrid Le Gros Clark, Dr. Bryan Patterson of Harvard University, and many others with whom this subject has been discussed. Figures included in this paper were prepared by Miss Christine Court of Oxford University, Mr. Chester Tarka of the American Museum of Natural History and Miss Ellen Cole of Philadelphia. I am indebted to the Wenner-Gren Foundation of New York for a grant which facilitated early completion of this paper. 68 NOTES ON EOCENE TARSIOIDS VII. REFERENCES CHANTRE, E, & GAILLARD, C. 1897. Sur la faune du gisement sidérolithique éocéne de Lissieu (Rhone). C.R. Acad. Sci. Paris, 125 : 986-987. Criark, W. E. LE Gros, 1959. The Antecedents of Man. 374 pp., Edinburgh. CoopER, C. FORSTER. 1910. Microchoerus erinaceus (Wood). Ann. Mag. Nat. Hist., London (8) 6: 39-43, pl. 3. FILHOL, H. 1873. Sur un nouveau genre de Lémurien fossile recemment découvert dans les gisements de phosphate de chaux du Quercy. C.R. Acad. Sci. Paris, 77 : 1111-1112. 1890. Description ‘d’une nouvelle espéce de Lémurien fossile (Necrolemur parvulus). Bull. Soc. philom. Paris (8) 2 : 39-40. Gazin, C. L. 1958. A‘review of the Middle and Upper Eocene Primates of North America. Smithson. Misc. Coll., Washington, 136 : 1-112, pls. 1-14. Grecory, W. K. 1915. I. On the relationship of the Eocene lemur Notharctus to the Adapidae and to other Primates. II. On the classification and phylogeny of the Lemuroidea. Bull. Geol. Soc. Amer., New York, 26 : 419-446. — 1920. On the structure and relations of Notharctus, an American Eocene primate. Mem. Amer. Mus. Nat. Hist., New York (n.s.) 3 : 49-243, pl. 27. HELLER, F. 1930. Die Saugetier fauna der mitteleozinen Braunkohle des Geiseltals bei Halle. Jb. hallesch. Verb. Erf. mitteldtsch Bodensch., Halle (N.F.) 9 : 13-41. Hirt, W. C. O. 1953. Primates. Comparative Anatomy and Taxonomy, I. Strepsirhint. xxiv + 798 pp., 34 pls. Edinburgh. 1955. Primates. Comparative Anatomy and Taxonomy, II. Haplorhini: Tarsioidea. 347 pp., 14 pls. Edinburgh. HUrRzELER, J. 1946. Zur Charakteristik, systematischen Stellung, Phylogenese und Ver- breitung der Necrolemuriden aus dem europidischen Eocaen. Ecl. geol. Helv., Basel, 39: 352-361. — 1948. Zur Stammesgeschichte der Necrolemuriden. Abh. schweiz. paldont. Ges., Zurich, 66 : 13-46, 41 figs. Kampen, P. N. vAN. 1905. Die Tympanalgegend des Saugetierschadels. Gegenbaur Jb., Leipzig, 34 : 321-722, 96 figs. Kvaavuw, C. J. VAN DER. 1931. The auditory bulla in some fossil mammals. Bull. Amer. Mus. Nat. Hist., New York, 62 : 1-352, 18 figs. Lorenz, L. von LIBURNAU. 1905. Megaladapis edwardsi G. Grandidier. Denkschr. Akad. Wiss. Wien, 77 : 451-490, pl. 6. PIVETEAU, J. 1957. Primates, Paléontologie humaine. Tvaité de Paléont., Paris, 7 : 1-675, 8 pls. RussELt, D. E. 1960. Le crane de Plesiadapis. Note préliminaire. Bull. Soc. géol. Fr., Paris (7) 1: 312-315. SaBaN, R. 1957. Affinités du genre Tupaia (contd.). Ann. Paléont., Paris, 43 : 57-100, 15 figs. ScHLOSSER, M. 1907. Beitrag zur Osteologie und systematischen Stellung der Gattung Necrolemur, sowie zur Stammesgeschichte der Primaten tiberhaupt. N. Jb. Min. Geol. Paléont., Stuttgart, Festband (1907) : 197-226, pl. ro. Simpson, G. G. 1935. The Tiffany fauna, upper Paleocene, III. Primates, Carnivora, Condylarthra and Amblypoda. Amer. Mus. Novit., New York, 817 : 1-28, 14 figs. 1937. The Fort Union of the Crazy Mountain Field, Montana, and its mammalian faunas. Bull. U.S. Nat. Mus., Washington, 169 : 1-287, pl. ro. 1940. Studies in the earliest Primates. Bull. Amer. Mus. Nat. Hist., New York, 77: 185-212, 8 figs. — 1955. The Phenacolemuridae, new family of early Primates, Bull, Amer, Mus, Nat, Hist., New York, 105, 5 : 415-441, pls. 1-6, NOTES ON EOCENE TARSIOIDS 69 STEHLIN, H. G. 1912. Die Saugetiere des schweizerischen Eocaens, Abh. schweiz. paldont. Ges., Zurich, 38 : 1165-1298, 46 figs. —— 1916. Die Sdugetiere des schweizerischen Eocaens. 5 D> z 20 30 40 50 60 70 80 90 100 110 DEGREES w) z Ww = g a. 15 KAFU RIVER GRAVELS : TOTAL 61 ies O 10 c w o5 > z 20 30 40 50 60 70 80 90 100 110 DEGREES w z = Gee KALOMO RIVER PRE-CHELLES ACHEUL: TOTAL 180 a w 5 10 a WW a 5 = =) z 20 30 40 50 60 70 80 90 100 110 DEGREES Ww 72 : = KALAMBO FALLS ACHEULIAN FLAKE TOOLS: TOTAL 115 wl 15 w © 10 ax ra a2 5 D> z 60 DEGREES ILA WIE 22 Frequency distribution : lengths and breadths. B.M. Bull. (N.H.) Geol. 5, 4 IPIL NAME, AA FREQUENCY DISTRIBUTION: LENGTHS AND BREADTHS 7X05 LENGTHS OF FLAKES AND PSEUDO-FLAKES: BETHLEHEM. TOTAL: 75 a) | ' | z | 2 2 A & SG 7 B © (OM 1 is 12 13 1G tA Te Ie 26 tp ae UNIT 24,” = z LENGTHS OF FLAKE TOOLS : KALAMBO FALLS ACHEULIAN G OWNS Ws LJ a Ww re fo) 20 BREADTHS OF FLAKES AND PSEUDO -FLAKES : BETHLEHEM. w TOTAL : 75 =z 10 LJ a = a 2 Zz 20 BREADTHS OF FLAKE TOOLS: KALAMBO FALLS ACHEULIAN TOTAL : 115 j 9 10 11 12 13 14 15 16 17 18 19 20 UNIT 2/44” JPG AN AL 1D, 28} Length/breadth patterns and ratios B.M., Bull. (N.H.) Geol. 5, 4 PLATE 23 LENGTH / BREADTH PATTERNS AND RATIOS 15- 14 7 55 2 "i BETHLEHEM FLAKES eG e 5:4 AND PSEUDO-FLAKES: TOTAL 75 / y Za 12 BREADTH C.M.S ho ON @ nu WwW i} I-25 3s) 4°55) (65557 78 (9) 10 11 12) 13°5514 15) 16 177 18) 19) 20 C.M.S. LENGTH Si poe Lee 5:4 g 7 Ve KALAMBO FALLS ACHEULIAN =, ya a i FLAKE TOOLS: TOTAL 115 “A ohn ia 7 7 a 10 y, y iy Ses 9 Zi mee 4. Wa Y . < 8 Ye ig Bie 2s = 7 Z = a Ba Z Oo = a cae tz oe a = OE . Wj 6 ris 5:2 ao 7 go 5 < @ Za oo e_ ae 4 Bee eee - a ae 3 atte ats 511 8 9 10 11 12 13 14 15 16 17 18 19 20 C.M.S. LENGTH GEOL. 5, 4. ry) 23 MARI96I Se se ADLARD AND SON, LIMITE pte ; _ BARTHOLOMEW PRESS, DORK 4 wt . Died ; ae E ‘ : sal at r 3 ‘i Mt * TA s 7 4 : r \- j . Sigh ex Y Ls i ae) a - aor, ta : 2 ages ‘seis 5 Re ; ; rr yy » ¥ = eS : Ee! i bid : san w FLORA OF THE LOWER HEADON BEDS OF HAMPSHIRE AND THE ISLE OF WIGHT M. E. J. CHANDLER BULLETIN OF _ THE BRITISH MUSEUM (NATURAL HISTORY) _ GEOLOGY Vol. 5 No. 5 LONDON : 1961 FLORA OF THE LOWER HEADON BEDS OF HAMPSHIRE AND THE ISLE OF WIGHT BY MARJORIE E. J. CHANDLER Pp. 91-158 ; Pls. 24-30; 4 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 5 No. 5 LONDON: 1961 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 5, No. 5 of the Geological (Palaeontological) serves. © Trustees of the British Museum, 1961 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued July, 1961 Price Thirty Shillings ERORA, OF THE LOWER HEADON BEDS OF HAMPSHIRE AND THE ISLE OF WIGHT By M. E. J. CHANDLER SYNOPSES The type section of the Lower Headon at Hordle, Hampshire, is briefly described and reference is made to the earlier work on the beds and their flora. The most varied collection of land plants comes from Bed 10 (the Leaf Bed) at Hordle especially east of Beckton Bunny and below Hordle House but gregarious water plants occur at a number of horizons. The mode of preservation of the fruits and seeds is noted. Twenty-seven species new to the flora are added and fifteen genera while some former erroneous generic records have been corrected. Four species formerly described as Carpolithus can now be placed in their systematic position. A few sparse records from the Lower Headon of the Isle of Wight are included. INTRODUCTION It is over thirty years since the Lower Headon flora of Hordle was studied during which time much new knowledge both of living and fossil material has been gained. The flora has been enriched by more collecting from the type locality at Hordle ; there is also sparse material from other sites. Research on older Eocene floras of the South coast has shed fresh light on the relationship of some earlier erroneously determined genera and species so that mistakes have now been corrected. In view of the above considerations the revision of the Lower Headon flora is overdue and is attempted in the following pages. As always, the author is deeply indebted to Dr. K. I. M. Chesters and Mr. F. M. Wonnacott for their invaluable help in the preparation of this paper. EXPOSURES OF THE LOWER HEADON BEDS Beds of Lower Headon age are exposed both on the mainland and in the Isle of Wight: on the mainland at Hordle cliff, near Milford-on-Sea, Hampshire and on the island on its north-western shores. The cliffs at Hordle, unlike most mainland sections at the present day, are still unobstructed by man-made obstacles to col- lecting although it is clear from the local anxiety about the present rapidity of coast erosion that this state of affairs cannot persist much longer. In the past the condi- tion of the cliff has varied from year to year with major changes extending over long periods. Hence sometimes collecting has been carried out readily, while at other times the productive beds have been inaccessible if high up, or completely masked by scree when low down. In the Isle of Wight a few plants only have yet been found for no rich horizons or pockets with remains of land plants like those of Hordle Bed 10 have so far been discovered. GEOL. 5, 5 12 94 FLORA OF THE LOWER HEADON BEDS rt. HORDLE Towards the end of the marine episode represented by the Barton Beds of Barton cliffs recession of the sea was again taking place. As a result the truly marine Barton clays pass gradually through transition beds (the Long Mead End Beds) into the estuarine and fluviatile deposits of the Lower Headon. The land surface must now have extended much further east than during the deposition of the Barton Beds but appears still to have been drained by a now much lengthened river flowing from the west. The variable clays, current bedded sands and lignitic seams of the Lower Headon were laid down by the river which at certain horizons left remains of the plants growing upon its banks in the sediments. These plants are repre- sented by fruits, seeds and battered leaves. At some stages the only plants pre- served were the monotonous remains of gregarious genera and species of the aquatic or subaquatic vegetation, e.g. Chava, Limnocarpus, Stratiotes and Brasenia, which may be presumed to have grown in brackish lagoons and delta channels in view of their association with estuarine shells. The most detailed accounts of the Hordle cliff section were those by the Mar- chioness of Hastings (1852: 191) and Tawney & Keeping (1883: 566). During the years 1920-25 the author carried out continuous observation of the section and, as a result, the strata were tabulated so as to show the connexion between these earlier records and the beds then exposed (Chandler, 1925: 3). With the exception of Chara, Sequoia, Limnocarpus, Stratiotes, Brasenia, Rhamnospermum and rare Caricoidea and Spirematospermum the plant remains have all been derived from two or three stations at a single horizon named by Tawney & Keeping the Leaf Bed or Bed Io. Because of the action of the sea and the fall of talus, the chance of collecting from a good exposure has been known to change even during a single tide. Thus E. M. Reid related in 1920 how on one occasion she and Clement Reid found abundant remains of Acrostichum lanzaeanum exposed in the cliff base and foreshore below Hordle House (now Hordle House School). By the next day the exposure had been completely hidden by a cliff fall. Bed 10 is very variable in character. It consists of sands, clays, lignites and cement stone of limited extent, all of which may yield plants sometimes sparsely disseminated, sometimes concentrated in lenticles or pockets. Its first appearance in the west is at the top of the cliff near Beckton Bunny. Passing eastwards it dips gradually until it reaches beach level a short distance to the east of Hordle House, where it forms the base of the cliff and adjacent foreshore and is about 8 feet thick. Between 1926 and 1931 a magnifi- cent section was exposed in this position by the scouring of the talus from the cliff foot and the shingle from the beach. The constant cutting of the sea into the bed produced a succession of fresh surfaces. Grey tenacious well-bedded clay formed the cliff base and upper foreshore with some interbedded sands. A hard mudstone or cement stone about two inches thick extended laterally for some distance in the foreshore about two feet above the underlying Mammal Bed (green clays with pockets of Paludina also well exposed at this time in the lower part of the foreshore). Soft strata immediately overlying the mudstone yielded fragmentary leaf-remains disposed irregularly without reference to bedding planes, the leaves being mangled, twisted and decayed. The carbonaceous films of leaf tissue disintegrated immedi- FLORA OF THE LOWER HEADON BEDS 95 ately on exposure to dry air, no cuticle having been preserved as in the leaves from Bournemouth or the Hordle Sequoia twigs. Hence only impressions are found in old collections and those are liable to be very imperfect on account of the soft matrix. By splitting blocks a number of specimens of Salvinia were found. These became hard on drying as often happens when soft strata have been exposed to the action of salt water. In the hardened state the reddish matrix resembles that of the palm leaf (Nipa) impressions in the Sedgwick Museum, Cambridge. It therefore seems probable that the palms came from this part of the section. The plants collected from Bed 10 below Hordle House were chiefly representative of a limited freshwater or subaquatic plant-formation. The most abundant genera were Salvinia, Brasenia, Stratiotes and Limnocarpus. The few fruits of Spzvemato- spermum now referred to a new species, S. headonense, were from this site. The occasional patches of Acrostichum and (probably from here) of Nipa indicate brackish conditions and tropical coastal marshes. The matted occurrence of the Acrostichum pinnules here as in the Dorset Pipe-clay Series suggests that it grew where it was preserved, the fronds falling continually into the mud by which they were buried. Land plants were sparse below Hordle House but a few were found by Clement Reid. A small stoned Rubus and Sambucus were abundant. There were also Icacinaceae and Zanthoxylum and a seed of Cucurbitaceae. It is a handicap to further research that at present tons of talus from the overlying Headon and Pleistocene beds obscure this part of the cliff while the foreshore is deeply buried under beach gravel. It may be many years, if ever, before collecting is again possible below Hordle House. Concentrations of plants have also been seen from time to time in Bed Io at other parts of the cliff section. A band rich in seeds and twigs of Sequoia couttsiae was long exposed east of Taddiford (= Long Mead End) and was accessible a few feet above the beach. It was made conspicuous both by the purplish tinge of its clays and sands and by a bright yellow efflorescence on the surface due to pyrites. There were also plant pockets high up in the cliff between Beckton Bunny and Taddiford but it was necessary to exercise the greatest caution in approaching them owing to the softness of the sands and clays which produce treacherous deep mud flows. Certainly since 1920 and until recent years a most productive lenticle was available near the cliff top a few yards east of Beckton Bunny. Here the richest matrix was coarse purplish sands underlying a few feet of well laminated tenacious purplish-grey clay beneath the Pleistocene gravels. This bed, from six to twelve inches thick, rapidly thinned and finally died out eastwards. In a westerly direc- tion it was cut out altogether by the sloping sides of the stream at Beckton Bunny. The bed yielded remains of trees, shrubs and woody climbers all of which may be presumed to have grown upon the river banks. Endocarps of Mastixioideae, represented especially by two genera, were particularly abundant and could be picked out of the weathered surface in handfuls. The loose sandy matrix was readily sifted on the spot without boiling thus reducing the bulk to be carried home for examination. Water plants occurred more sparsely than land plants. Towards Taddiford Bed ro has always been obliterated by a contemporary stream channel cut transverse to the cliff face. The channel is rendered conspicuous by 96 FLORA OF THE LOWER HEADON BEDS the immense amount of black carbonaceous detritus and masses of broken woody fragments. It did not yield a sufficient quantity of fruits and seeds to be worth working extensively but two good endocarps of Natsiatum eocenicum were picked up on rain-washed slopes of talus and there were also Sequoia seeds and much abraded fruiting heads of Protoaltingia hantonensis formerly referred to Liquidambar sp. (Chandler, 1925 : 25, pl. 4, fig. 1). The Lower Headon fruits and seeds are carbonaceous entities. When first extracted from the cliff they appear as unshrunken and undistorted beautifully preserved specimens scarcely distinguishable except by colour and differences of specific character from their living counterparts. Thanks to the soft carbonaceous and uncrushed condition they are easily dissected so that they afford invaluable evidence about their internal characters. Natural fractures and dissections of such material give readily available information for comparison with broken surfaces in living material. There is always considerable shrinkage of carbonaceous material on drying e.g. large endocarps of Mastixicarpum (Mastixioideae) may lose from half to three-quarters of their length in this way. Dried specimens also tend to become distorted and the coarse sandy matrix causes pitting of both external and internal surfaces in some specimens while sand grains may have penetrated along cracks, natural canals and planes of weakness like the edges of valves, filling locule- and seed-cavities and often pock-marking the lining membranes. Testas inside woody endocarps, or tegmens in the case of seeds, are commonly present as tough readily detachable skins. As so frequently happens in carbonaceous material pyrites disseminated throughout the tissues renders them liable to disruption and decay. The softness of the carbonaceous substance may also give rise to crumbling although in both cases deterioration may be delayed by treatment with paraffin wax. Hence the importance of adequate photographic records of the specimens showing range of variation and characteristic features. The first account of the Hordle flora (Chandler, 1925, 1926) was based on collec- tions by C. & E. M. Reid and by M. E. J. Chandler. A few specimens in the Sedg- wick Museum, Cambridge and in the British Museum (Natural History) were also examined. Since that time, in 1939 and 1940, the slope of the cliff near Beckton Bunny made Bed Io particularly easy to work. Every visit produced valuable material, much being new, and some twenty-seven additions to the published flora were then made. In June 1940 national defence restrictions made further field work impossible and when the coast was once again open in 1945 the cliff was in an unfavourable condition for collecting. The following are additions to the flora: Stvatiotes hantonensis Chandler, Cari- coidea obscura Chandler, Scleria hordwellensis n. sp., Myrica boveyana (Heer), Carpinus boveyanus (Heer), Moroidea hordwellensis n. sp., Becktonia hantonensis n. sp., Hantsia glabra n.sp., Lauraceae Genus ?, Foliquidambar hordwellensis n. sp., Leguminosae (two unidentified genera), Palaeobursera lakensis Chandler, Iodes ? hordwellensis n. sp., Icacinicarya transversalis n.sp, Icacinicarya becktonensis n.sp., Frangula hord- wellensis n. sp., Meliosma sp., Actinidia sp., Cleyera ? stigmosa (Ludwig), Anneslea ? costata n. sp., Eurya becktonensis n. sp., Thymeliaceae Genus? Microdiptera parva Chandler, Cornus quadrilocularis Chandler, Olea headonensis n. sp., ? Acanthus sp. FLORA OF THE LOWER HEADON BEDS 07 Not only have new species been found since the original monograph was published but fresh knowledge both of Recent and fossil material has enabled certain mistakes to be corrected. Just as initially the Hordle flora helped in the understanding of other Tertiary floras examined later, so, too, these other older floras have in their turn shed new light on that of Hordle. Thus it is now possible to place in their proper botanical positions over a dozen species which were before incorrectly named. The following corrections may be noted: Salvinia hantoniensis (Chandler, 1925 : Io, pl. 1, fig. 1a—-d ; text-fig. 1) has been referred by Shaparenko (1956) to S. mildeana Goeppert. Cladium minimum (Chandler, 1925: 14, pl. 1, fig. 5a, 5) is now referred to the form-genus Caricoidea denoting relationship with Caricoideae rather than with Cladium itself. Limnocarpus headonensis (Chandler, 1925: 13, pl. I, fig. 4a-c; text-fig. 3) becomes L. forbesi (Heer) for reasons given by Chandler 1961a. Campylo- spermum hordwellense (Chandler, 1925: 16, pl. 1, fig. 6a-c; text-fig. 4) formerly referred to Araceae, belongs to the Theaceae, section Taonabeae, although no identical living genus has yet been found. Nuphar ovatum (Chandler, 1925: 22, pl. 3, fig. 3a, b) thanks to the work of Docturowsky, Nikitin and Korzhinsky becomes Aldvovandra ovata (Droseraceae) (see Reid & Chandler, 1926: 111-114, pl. 6, figs. 24-29). Menispermum obliquatum (Chandler, 1925: 24, pl. 3, fig. 9a, 5) is trans- ferred to the fossil genus Palaeosinomenium (Chandler, 1961: 159). Rhamno- spermum bilobatum (Chandler, 1925 : 30; 1926, pl. 5, fig. Ia-c; text-fig. 13) cannot be retained in Rhamnaceae although its relationship is obscure. It has been found now at horizons ranging from the Lower Bagshot to the Hamstead Beds and there is a new and unpublished record from the London Clay. Rhamnaceae Genus ? sp. 2 (Chandler, 1925: 31; 19206, pl. 5, fig. 2) is now referred to Frangula hordwel- lensis. Corydalis pulchra (Chandler, 1925: 25, pl. 3, fig. 10a, b) must be removed from Corydalis and the family Papaveraceae and placed in Caryophyllaceae under a new generic name Hanisia for reasons given on p. 114. Liquidambar sp. (Chandler, 1925: 25, pl. 4, fig. 1) is actually Protoaltingia hantonensis common in the older Tertiary Beds of the Dorset and Hampshire coast section as shown by the abraded remains of the fruits with antero-posterior locules and both loculicidal and septicidal dehiscence. Zanthoxylum ornatum (Chandler, 1925: 27, pl. 4, fig. 4a, 0; text-fig. 10) is removed from the living genus and placed in a form-genus Rutaspermum Chandler. Actimidia crassisperma (Chandler, 1926: 34, pl. 6, fig. 2; text-fig. 15) is referred to a new genus Hordwellia of the family Theaceae, section Taonabeae. Carpolithus sp. 2 (Chandler, 1926: 44, pl. 7, fig. 12a-d; text-fig. 28) belongs to the Mastixioideae and probably to Mastixia. It resembles Retinomastixia Kirchheimer in the numerous resin-cavities in the wall and is named Mastixia? glandulosa. Eomastixia bilocularis (Chandler 1926 : 37, pl. 6, fig. 6a-e; text-fig. 20) has to be placed in Zenker’s species EF. vugosa. Ericaceae Genus ? (Chandler, 1926 : 37, pl. 6, fig. 7a, b) is Epacridaceae and is described under a form-genus Epacridicarpum. Symplocoides glandulosa (Chandler, 1926: 41, pl. 7, fig. 5a, 0; text-fig. 26) and Carpolithus sp. 3 (Chandler, 1926: 45, pl. 8, fig. Ia-e) both belong to Cornaceae and are now referred to Dunstania, a genus from the London Clay which, according to Kirchheimer should itself be placed in Cornus. Carpolithus sp. 4 (Chandler, 1926 : 45, pl. 8, fig. 2a, b) is referred to Cornus quadrilocularis Chandler, Oleaceae Genus ? 98 FLORA OF THE LOWER HEADON BEDS (Chandler, 1926: 42, pl. 7, fig. 7; text-fig. 27) appears to be Olea and is named O. headonensis. Omphalodes platycarpa (Chandler, 1926: 42, pl. 7, fig. 8a, b) needs confirmation from additional material but no alternative suggestion as to its rela- tionship can at present be made. Ovites sp. (Chandler, 1926: 47, pl. 8, fig. 6): better evidence as to structure and some in regard to the seeds is needed. The fruit spike shows some of the characters of Carpolithus gardneri Chandler (see p. 154) ; its relationship still awaits discovery. Atviplex sp. (Chandler, 1925: 21, pl. 3, fig. 2) must be deleted. Further study reveals that it is an accidental Recent inclusion. Carpolithus sp. 5 (Chandler, 1926: 46, pl. 8, fig. 3a-d) is referred to Mastixioideae Genus ? The invaluable detailed studies by Kirchheimer of well-preserved Brown Coal material confirms the reference of Spivematospermum to the Zingiberaceae but a consideration of his published evidence necessitates placing the Lower Headon fruits and seeds in a distinct species now named S. headonense (p. 108). The full up-to-date list of Lower Headon plants omitting the Charophyta (described by Reid & Groves in 1921) is given below. All come from Bed 10, just east of Beckton Bunny unless otherwise stated in the systematic description. The numbers by which the beds are designated are those used by Tawney & Keeping (1883) and repeated by Chandler (1925: 3). Bed ro has been referred to by C. Reid and by Chandler as the Leaf Bed, Bed 29 as the Chava Bed and Bed 31 as the Limnocarpus Band of the Unio Bed. In several cases where no new material is available and there is nothing fresh to add to previously published work the registration number in the British Museum (Natural History) is added opposite the name. Where material is no longer extant this is indicated by the word Decayed. Erdtmann (1960) gave a brief account with figures of three new pollen genera from the Chara Bed of Hordle cliff (erroneously said to be in Berkshire). As how- ever a larger work on pollen from the Headon Beds is in preparation by Miss J. Pallot it is left to her to discuss the significance of these determinations. The importance of the discovery of a number of the Hordle plants at other older horizons will be discussed later when work on Eocene floras is surveyed. It is sufficient now to reaffirm that the flora is of tropical [Poltavian] type as indicated by the presence of Acrostichum, Nipa, Burseraceae, Icacinaceae and Mastixioideae. Associated animal remains also indicate a warm climate, among them are Tvionyx, Emys and Crocodilus hastingsiae. There is a rich but little known fish and mammalian element. Among the former Lepidosteus scales and Myliobates are frequent (see Marchioness of Hastings, 1852; Tawney & Keeping, 1883). There are forty-three plant families, at least sixty-seven different genera and eighty-seven named or at least distinctive species. 2. COLWELL Little collecting has been carried out in the Lower Headon of the Isle of Wight. A few plants only were obtained in the coast section between Colwell and Totland from the Limnocarpus Band, a readily recognizable horizon in spite of the fact that it is only a few inches thick. This distinctive band was formerly well exposed at FLORA OF THE LOWER HEADON BEDS List of Lower Headon Plants C, also at Colwell Bay, Isle of Wight. D, also at Downton, Hordle. Family Filicales Polypodiaceae Salviniaceae Gymnospermae Taxodineae Abietineae . Angiospermae Monocotyledones Potamogetonaceae Hydrocharitaceae Cyperaceae Nipaceae . Araceae Zingiberaceae Dicotyledones Myricaceae Betulaceae Moraceae Caryophyllaceae Nymphaeaceae Menispermaceae Lauraceae Droseraceae Hamamelidaceae Rosaceae Leguminosae ? . Rutaceae Burseraceae Anacardiaceae Spondieae Genus and species Acrostichum lanzaeanum (Visiani). Salvinia mildeana Goeppert. Sequoia couttsiae Heer. Pinus sp. (Decayed.) Potamogeton pygmaeus Chandler (C). Limnocarpus forbesi (Heer) (C, D). Stvatiotes headonensis Chandler (C). Stratiotes hantonensis Chandler. Caricoidea minima (Chandler). Caricoidea obscura Chandler. Sclevia hoydwellensis n. sp. Genus ? ? Nipa sp. (leaves probably of N. burtini). Genus ? (V.20036). Spivematospermum headonense n. sp. Myrica boveyana (Heer). Carpinus boveyanus (Heer). Chlorvophora bicavinata Chandler. (Decayed.) Broussonetia yvugosa Chandler. Moroidea hordwellensis n. sp. Becktonia hantonensis n. gen. & sp. Hantsia pulchra (Chandler). Hantsia glabra n. sp. Brasenia ovula (Brongniart) (C). Brasema spinosa Chandler (includes B. antiqua Chandler). Brasema oblonga Chandler. Palaeosinomenium obliquatum (Chandler). Genus ? Aldrovanda ovata (Chandler) (C). Eoliquidambar hordwellensis n. gen. & sp. Protoaltingia hantonensis Chandler. Rubus acutiformis Chandler. Genus ? Genus ? Zanthoxylum hordwellense n. sp. Zanthoxylum compressum Chandler. Phellodendron costatum Chandler. Rutaspermum ornatum (Chandler). Palaeoburseva lakensis Chandler. Genus ? 99 Ioo FLORA OF THE LOWER HEADON BEDS Family Icacinaceae Rhamnaceae Sabiaceae Vitaceae . Dilleniaceae Theaceae Theae . Taonabeae Thymeliaceae Lythraceae Cornaceae Mastixioideae Cornoideae Ericaceae Andromedeae Rhododendroideae . Epacridaceae Ebenaceae Symplocaceae . Styracaceae Oleaceae . Boraginaceae Acanthaceae Caprifoliaceae . Cucurbitaceae . Family ? Genus and species Natsiatum eocenicum Chandler. ? Iodes sp. (or ? Natsiatum sp.). Iodes ? hordwellensis n. sp. Icacinicarya transversalis n. sp. Icacinicarya becktonensis n. sp. Frangula hordwellensis n. sp. Meliosma sp. Parthenocissus hordwellensis n. sp. Vitis uncinata Chandler (V.20065). Ampelopsis votundata Chandler. Tetrastigma lobata Chandler. (Decayed.) Actimidia sp. Gordonia minima Chandler. Gordonia truncata Chandler. Eurya becktonensis n. sp. Campylospermum hordwellense Chandler. Cleyera ? stigmosa (Ludwig). Anneslea ? costata n. sp. Hordwellia crassisperma (Chandler). Genus ? Microdiptera parva Chandler. Eomastixia vugosa (Zenker). Mastixicarpum crassum Chandler. Mastixia ? glandulosa n. sp. Genus ? Cornus quadvilocularis Chandler. Dunstania glandulosa (Chandler). Genus ? Genus ? (Decayed.) Epacridicarpum headonense Chandler. Diospyros headonensis Chandler. Symplocos headonensis Chandler. Symplocos sp. Styvax elegans Chandler. Olea headonensis 0. sp. Omphalodes platycarpa Chandler. ? Acanthus sp. Sambucus parvulus Chandler. Cucurbitospermum reidit Nn. sp. Rhamnospermum bilobatum Chandler. Carpolithus fibrosus n. sp. Carpolithus apocyniformis n. sp. Carpolithus sp. 8. Carpolithus sp. 6. (V.20107). Carpolithus sp. cf. C. gardneri Carpolithus sp. 1. Carpolithus spp. and tubers, rhizomes (V.20111 and V.40182-83 ‘‘ Nelumbium buchii Ettingshausen ’’), placentas, FLORA OF THE LOWER HEADON BEDS IOI Paddy’s Gap, Milford-on-Sea on the mainland. Wherever seen it is a conspicuous feature owing to a compacted mass of white shells—Paludina, Melamia and frag- ments of Unio—which show up conspicuously the black dots due to crowded endo- carps of Limnocarpus together with seeds of Brasenia and Stratiotes. The Limno- carpus Band is part of the Hordle Bed 31 of Tawney & Keeping. A few specimens labelled Limnocarpus Band were collected at Colwell by James Groves with whom Clement Reid worked in the field while in retirement at Milford- on-Sea. A few others merely labelled Colwell Bay were in the Reid Collection. No distinctive Upper Headon fruits are among them and it therefore seems probable that these also were from the Limnocarpus Band. A section given on pp. 132-133 of the Geological Survey Memoir on the Isle of Wight (1889) records a number of seams with lignite and seeds but experience suggests that these seeds were the usual Brasenia, Stratiotes and Limnocarpus. Nevertheless there is always a possibility that some rich pocket like those of Bed 10, Hordle, may someday reward persistent search. So far, however, despite the vigilance of Groves and Colenutt no such has been discovered in the Isle of Wight. The short list of plants from the Lower Headon of Colwell is as follows : Potamogeton pygmaeus Chandler Limnocarpus forbesi (Gardner) Stratiotes headonensis Chandler all are water plants. Brasenia ovula (Brongniart) Aldrovanda ovata (Chandler) PTERIDOPHYTA Order FILICALES Family PoLYPODIACEAE Genus ACROSTICHUM Linnaeus Acrostichum lanzaeanum (Visiani) (Bi 245 tie.2) 1858. Fortisia lanzaeana Visiani, p. 431, pl. 1, fig. 8, pl. 2, figs. I, 5. 1879. Chrysodium lanzaeanum (Vis.): Gardner & Ettingshausen, p. 26, pl. 2, figs, 3, 4. 1886. Chrysodium lanzaeanum (Vis.): Gardner, p. 402, pl. 3, fig. 5. 1925. Acrostichum (Chrysodium) lanzaeanum (Vis.): Chandler, p. to. 1926. Acrostichum lanzaeanum (Vis.): Reid & Chandler, p. 33, pl. 1, figs. 4, 5. Acrostichum lanzaeanum (Vis.): Chandler (in press), pl. 5, figs. 4, 5. Dorset Pipe-clay Series. Reference to pinnules from Hordle was made by Chandler in 1925 and the first description, based on material in the Sedgwick Museum, Cambridge, was published by Reid & Chandler in 1926. In the British Museum there are seven further speci- mens (41420(5), V.42050-51) one of which shows no nervation so is doubtful and one showing the branching of the pinna. All are preserved in a carbonaceous grey clay and appear to come from Bed 10 below Hordle House, 102 FLORA OF THE LOWER HEADON BEDS Two fragments collected by H. Eliot Walton (V.42057—-58) one of which showed also a seed of Spivematospermum (V.42057a) were alleged to come from the Barton Beds and a similar specimen, also collected by Walton, is in the Geological Survey Museum (7927). Mr. Walton was an amateur with no scientific training and the specimens may have been found loose on the foreshore. In view of the continuity of the Barton and Hordle sections the origin of this material requires confirmation by the finding of fresh pinnules at recognized Barton horizons. There is, however, no reason why the fern should not occur in the more estuarine Barton Beds for it persists from the Dorset Pipe-clay Series up to the Bembridge Series. A frond is shown (PI. 24, fig. 2) for comparison with Hordle pinnules figured by Reid & Chandler, 1926 (see above). Family SALVINIACEAE Genus SALVINIA Adanson Salvinia mildeana Goeppert (El 245 shee) 1886. Salvinia sp. Gardner, p. 395. 1925. Salvinia hantoniensis Chandler, p. 10, pl. 1, fig. 1a—-d; text-fig. 1. 1956. Salvinia mildeana Goeppert : Shaparenko, p. 30, pl. 1, figs. 1-3; text-fig. 9. Where earlier references to S. mildeana are given in full. In a detailed study of the genus Salvinia, both living and fossil, Shaparenko (1956) has shown that the Hordle species should be referred to S. mildeana, a widespread form in the Tertiary of Europe from the Eocene to the Upper Miocene as shown by the list of fossil occurrences which he gives on p. 30. Material from Bed 10 below Hordle House is well represented in the British Museum. V.20010 and V.36341—46 include the specimens figured by Chandler (1925) and two unfigured specimens. V.36343 (Chandler, 1925, pl. 1, fig. Ic) also shows the counterpart of V.36342 (Chand- ler, 1925, pl. 1, fig. 1) while V.36341 (Chandler, 1925, pl. 1, fig. 1d) shows the counter- part of V.36344 (Chandler, 1925, pl. 1, fig. Ia). V.36346 includes fourteen speci- mens from a single ironstone block found im situ two feet above the mammal bed in the foreshore. There are also nine blocks (V.43698), and one specimen (V.43699) in grey clay matrix. GYMNOSPERMAE Family PoDOCARPACEAE Genus PODOCARPUS (L’Hérit.) ? Podocarpus sp. 1883. Podocarpus sp. Gardner, p. 49, pl. 2, figs. 13, 14. In a discussion under the heading Podocarpus eocenica Unger Gardner refers to two small coriaceous leaves from Hordle (spelt Hordwell) which he figures (V.15117— 18). Another incomplete detached leaf impression is in the British Museum FLORA OF THE LOWER HEADON BEDS 103 (V.15119). Gardner comments of these leaves “if referable to Podocarpus ’’ [they] ‘“‘would probably belong to another species ’’ [i.e. not to P. eocenica from Bourne- mouth which he was describing]. As the material is so doubtful and ill-preserved it is ignored in the plant list. The preservation suggests “ Bed 10’ below Hordle House. Family TAXODINEAE Genus SEQUOIA Endlicher Sequoia couttsiae Heer The Hordle material is discussed and illustrated in detail in a forthcoming mono- graph on the Flora of the Dorset Pipe-clay Series (Chandler, in press) in which Sequoia couttsiae from all known Tertiary sites in the South of England is investi- gated. Hordle material is registered under the numbers V.20011-18, V.42052-53. It is particularly common in Bed 10 between Hordle House and Long Mead End and just east of Beckton Bunny ; it also occurs in the contemporary channel west of Long Mead End and is rare (V.20017) in the Chava Bed east of Long Mead End (Beds 27-29 of Tawney & Keeping). ANGIOSPERMAE Class MONOCOTYLEDONES Family POTAMOGETONACEAE Genus POTAMOGETON Linnaeus Potamogeton pygmaeus Chandler (Pl. 24, figs. 3-6) 1925. Potamogeton pygmaeus Chandler, p. 13, pl. I, fig. 3. 1926. Potamogeton pygmaeus Chandler: Reid & Chandler, p. 66, pl. 3, figs. 20-22. The range has now been extended backwards into the Bournemouth Beds as will be described in a later monograph. Specimens have also been found in the Limno- carpus Band, Lower Headon, Colwell Bay, Isle of Wight (Pl. 24, figs. 3, 4). The holotype (V.20019) is in poor condition so a fruit impression (with counterpart V.17537) from the Bembridge Beds has been chosen as the neotype. It shows style and spines beautifully preserved (Reid & Chandler, 1926, pl. 3, fig. 20). There are about fourteen other specimens from Hordle which show spine bases, some having the valve preserved (V.20020—20a, V.42056). They are from Bed io east of Beckton Bunny. P. I. Dorofeev considers (ex lit) that the nearest living affinity is with the North American Potamogeton spirillus Tuckerm. 104 FLORA OF THE LOWER HEADON BEDS Genus LIMNOCARPUS Reid emend. Reid & Chandler, 1926 : 68 Limnocarpus forbesi (Heer) (Pl. 24, figs. 7-11) For synonyms and references see Chandler, 19614, p. 28. The species formerly designated L. headonensis (Gardner) must now be referred to L. forbesi (Heer) for reasons explained by Chandler (1961a@: 28). Endocarps are extremely abundant at certain horizons in the Lower Headon Beds at Hordle, notably in the Crocodile Beds between Hordle House and Long Mead End (Bed 15 of Tawney & Keeping) and in the Limnocarpus Band formerly peculiarly well dis- played at Paddy’s Gap, Milford-on-Sea. Figured specimens still extant are V.20021 (Chandler, 1925, pl. 1, fig. 4a) and V.200226 (Chandler, 1925, pl. 1, fig. 4b). The originals of Chandler, 1925, pl. I, fig. 4c and text-fig. 3 have decayed but V.20023 still shows five endocarps split longitudinally through the funicle and placenta as in the text-figure. Other specimens from Hordle are V.20022a, c, V.20024-27 (including four paired endocarps), V.20028-30, V.31789, V.42059-60. Endocarps (V.42062—-63) were also found in a bomb crater at 50 ft. O.D. on the east side of the Dane stream, Downton, Hordle. The crater showed about 5 ft. of white micaceous sand like the Umio sands of Paddy’s Gap, Hordle cliff, and below were Io ft. of greenish clay with a band of Limnocarpus 6 ft. down. At Colwell Bay, Isle of Wight, Groves collected numerous specimens (V.42064-67) in the Limnocarpus Band (exact site unspecified) and Chandler obtained others (V.42068) at the same horizon below Warden Battery. Family HyDROCHARITACEAE Genus STRATIOTES Linnaeus Stratiotes headonensis Chandler (Pl. 24, fig. 17) 1923. Stvatiotes headonensis Chandler, p. 117, pl. 5, figs. 1-3, 24-26; pl. 6, fig. 23. 1925. Stvatiotes headonensis Chandler: Chandler, p. 14. 1960. Stvatiotes headonensis Chandler: Chandler, p. 223, pl. 33, figs. 96, 97. A full description of the species was given in 1923. For convenience of com- parison with S. hantonensis a diagnosis is given: Seed oblong, ovate or subtrigonal, somewhat flattened, hooked at the base. Keel often from one-third to one-half the breadth of the seed. Collar small, conspicuous, usually warty. Testa thick, woody with marked longitudinal ridges. Micropyle basal or subbasal, usually oblique. Hilum basilateral, raphe marginal or submarginal almost to the apex. Cells of internal surface of keel much contorted. Length of seed from about 3°25 to 65 mm.; breadth about 2°25 to 4 mm. Remarks. The following were figured by Chandler, 1923, pl. 5, fig. 1 (V.40068), pl. 5, fig. 2 (V.40067a), pl. 5, fig. 24 (V. 40066). The species is very abundant in FLORA OF THE LOWER HEADON BEDS 105 Bed 10, 49438, V.13567, V.16538, V.20031-32, V.31788, V.40067-68a, V.42073, V.42078 (precise position in the cliff unspecified), V.42069 (east of Beckton Bunny), V.42070 (below Hordle House). Also very abundant in Bed 31 at Paddy’s Gap (V.42066, V.42071). It occurs sparsely in Bed 29 between Hordle House and Milford Corner (V.42072). Horizon in Lower Headon Hordle unspecified in V.16537-384. It is equally abundant at Colwell Bay, V.42076 (Chandler, 1923, pl. 5, fig. 3), V.42077, V.42079-81, all presumed Lower Headon, and V.42074 Limnocarpus Band, site unspecified, V.42075 same band below Warden Battery. The oldest record of the species up to date is at Horizon L. of the Barton Series, Barton cliff (Chandler, 1960 : 223). Stratiotes hantonensis Chandler (Pl. 24, figs. 12-16) 1960. Stratiotes hantonensis Chandler, pp. 205, 222, pl. 30, figs. 18-25; pl. 33, figs. 87-95. The description given in 1960 applies equally to the seeds from the Lower Headon Beds. Typical measurements of Hordle material are 2-6 by 1-6 mm.; 2°8 by 1:2 mm.; 3 by 1:6mm.; 3 by 1-5 mm. The thickness of the seeds is about 1:3 mm. This small species was overlooked when S. headonensis was described in 1923, its tiny seeds being merely regarded as small immature examples of S. headonensis, only about sixteen specimens (V.42082—86a) were found. Now that S. hantonensis has been collected in large numbers in older beds ranging from Bournemouth through the Hengistbury Beds, Highcliff Sands and Barton Beds its distinctive characters are well known and it was no longer confused with S. headonensis when the Hordle seeds were re-examined. The range of the two species therefore overlaps at the western end of the Hordle section. So far only gatherings of S. headonensis from Bed ro close to Beckton Bunny have yielded S. hantonensis. The absence from older deposits of typical S. headonensis with its broad flat keel, coarse pitting, pronounced ridges and larger size in association with S. hantonensis confirms that the latter is a distinct species. Family CYPERACEAE Genus CARICOIDEA Chandler, 1957 : 86 Caricoidea minima (Chandler) (Pl. 24, figs. 18-21) 1925. Cladium minimum Chandler, p. 14, pl. 1, fig. 5a, b. Caricoidea minima (Chandler) Chandler (in press), pl. 7, fig. 1. Diacnosis. Fruit sub-obovoid sometimes slightly flattened laterally, rather abruptly narrowed to the apex, truncate at the base. Fruit wall thick, epidermis of digitate rectangular cells. Length about 2-75 to 3 mm.; maximum diameter, 2 to 25 mm. Endocarp oburceolate with short truncate basal neck, rounded or 106 FLORA OF THE LOWER HEADON BEDS pointed apex. Wall of two layers the inner columnar in section, formed inside and out of conspicuous equiaxial cells. Length, 1-25 to 1-65 mm.; breadth, 1 to 1°55 mm. Seed with testa of transversely elongate cells. HoLotyPe. Brit. Mus. (N.H.), No. V.20033. DESCRIPTION. Fruit: Sub-obovoid, sometimes slightly flattened laterally, truncate at the base which has a short canal with large foramen marking the attach- ment. Somewhat pointed at the apex, not conspicuously attenuated or mucro- nate but abruptly narrowed. Epidermis shining of digitate cells usually more or less rectangular and somewhat longitudinally elongate and aligned, their width about 0-012 to 0-016 mm. Wall thick, spongy, formed of a mass of equiaxial cells about 0-oI mm. in diameter. Length of fruit, 2°75 to 3 mm. ; maximum diameter, 2 to 2°5 mm. Endocarp : More or less oburceolate but contracted rather abruptly as a rule to form a short truncate basal neck pierced by a canal, rounded or pointed at the apex, sometimes with two or three external strands of longitudinal fibres. Wall thick, compact tending in fossilization to separate into two layers, an outer compact close-textured layer about 0-15 mm. thick and an inner layer, columnar in section, about 0-037 mm. thick formed superficially inside and out of equiaxial cells about 0-016 mm. in diameter. Length of endocarp about 1:25 to 1-65 mm.; breadth, I to 1°55 mm. Seed: Represented by testa formed of transversely elongate cells arranged in rows. REMARKS. Detached endocarps (V. 20033, V.20033@) were described as Cladium nunimum before fruits had been seen (Chandler, 1925: 14, pl. 1, fig. 5a, 6). Fruits from Bed 10 (V.42087-88, V.42091) and additional endocarps (V.42089-g0) have now been found. Four crushed endocarps have also been recognized from Bed 29 in the lignitic bands (V.20047b). While relationship is certainly with the Caricoideae the fruit characters show that these specimens differ from Cladium itself which has a thin spongy fruit wall and an epidermis of large rectangular smooth-walled cells. The fossils are therefore referred to the form-genus Caricoidea pending the discovery of further evidence. A fruit has now been found in the Dorset Pipe-clay Series at Arne. Caricoidea obscura Chandler (Pl. 24, figs. 22-24) 1960. Cayvicoidea obscuva Chandler, pp. 207, 223, pl. 30, figs. 27-33; pl. 33, figs. 98-105. Caricoidea obscura Chandler : Chandler (in press), pl. 6, figs. 20-33. 1961a. Cavicoidea obscura Chandler : Chandler, p. 33, pl. 7, fig. 34. DescripTIon. Fruit: As described from Hengistbury (p. 207 quoted above). The single fruit seen from Hordle has a somewhat three-sided basal truncation and is itself obscurely three-sided one side being broader and flatter than the other two. Epicarp although poorly preserved shows the finely digitate cell walls about 0-016 mm. broad. Beneath the epicarp the equiaxial cells of the spongy tissue are about 0-OI mm. in diameter and light brown in colour. Length about 1-75 mm. ; breadth, I°-5 mm. FLORA OF THE LOWER HEADON BEDS 107 Endocarp : Also as described. One or more longitudinal ridges arise at the base and gradually die out on the body of the endocarp but these are not always present. Surface smooth externally and the spongy wall, about 0:07 mm. thick, is formed of equiaxial cells. Dimensions of several endocarps are: I by 08 mm.; 0-9 by 0-75 mm.; 1°15 by 0°85 mm.; I by 0-7 mm.; 0-9 by 0-g mm. Seed : Poorly preserved represented only by much decayed adherent testa formed of transversely elongate cells which produce striae about 0-012 mm. apart. Remarks. About a dozen endocarps were found (V.42095) in addition to the figured material. The species is based on well preserved fruits and endocarps from the Highcliff Sands near Mudeford and the Dorset Pipe-clay Series and also awaits description from the Bournemouth Freshwater Beds. Specimens have been described from the Upper Hengistbury and Lower Barton Beds (Chandler, 1960). Genus SCLERIA Berg. Scleria hordwellensis n. sp. (Pl. 24, figs. 25, 26) Diacnosis. Fruit subglobular, mucronate at the apex, flattened at the base and seated on a rounded triangular thick disc with transversely puckered surface. Surface of fruit smooth but with slight longitudinal fluting on the sides. Surface of both fruit and disc of equiaxial cells. Length of fruit with mucro and disc, 1-75 mm.; breadth, 1°75 mm. HoLotypre. Brit. Mus. (N.H.), No. V.42096. DEscRIPTION. Fruit: Subglobular with short thick apical mucro. Base some- what flattened pierced by an aperture but without a narrow neck like that of Cladium and Mapania. Circumference of fruit rounded not triangular. Sides slightly grooved or fluted longitudinally, seated when found on a thick rounded triangular disc which still adheres on one side but is elsewhere now broken. Its removal has shown a rounded triangular scar at the base of the fruit itself about 1-3 mm. in diameter. The angles of the disc originally extended half way up the fruit. Its surface is rippled or puckered transversely. Epidermis of fruit and disc formed of equiaxial cells, many about 0-008 mm. in diameter. Inside the disc as shown along the broken edge are equiaxial cells, 0-016 mm. in diameter. Length of fruit with disc and mucro, I-75 mm.; breadth, 1°75 mm. REMARKS AND AFFINITIES. One fruit which suggests relationship to Scleria of the type S. /aevis. The genus occurs in all tropical and subtropical regions with a single species in Atlantic North America. Family CYPERACEAE, Genus ? (Pl. 24, fig. 27) A fruit {V.42097) too poorly preserved for generic determination. Suboval in outline, shortly stipitate with acuminate apical style. Probably originally inflated but somewhat compressed in fossilization. Surface with regular equiaxial cells about 0-025 mm. in diameter with raised walls (these possibly due to differential decay). Length of fruit, 1-45 mm.; breadth, I mm. GEOL. 5,5 13 108 FLORA OF THE LOWER HEADON BEDS Family ZINGIBERACEAE Genus SPIREMATOSPERMUM Chandler, 1925 : 17 Spirematospermum headonense n. sp. (Pl. 24, figs. 28-30; Pl. 25, figs. 31, 32) 1925. Spivematospermum wetzlevi (Heer): Chandler, p. 17, pl. 1, fig. 8a-c; text-fig. 5. Diacnosis. Fruit: Bisymmetric with rounded upper surface and sharply angled lower margin hence obovoid in transverse section. Estimated length probably less than 25 mm.; maximum transverse diameter, 7 to 9g mm. Thickness of fruit wall, 0-228 mm. (maximum). Number of seeds in fruit probably not more than twelve arranged in one or two rows on parietal placentas in the plane of symmetry. Locules one? or two? Length of seeds, 6-25 to 7:25 mm.; maximum breadth, 15 to 2:25 mm. All tissues of seeds spirally arranged in relation to the seed cavity. Hototype. Brit. Mus. (N.H.), No. V.42008. DESCRIPTION. Fruit: Inferior, bisymmetric, many-seeded, the maximum number of seeds estimated to be not more than twelve, eight actually counted in the holotype. Truncate lanceolate, the apex bearing a flat elliptical superior perianth disc (V.42100 Chandler, 1925, text-fig. 5 (4)). Base narrowing gradually into a thin curved stalk (V. 42098-99). Upper margin rounded, lower sharply angled in the plane of symmetry the angle being continued along the stalk, so that clearly it is an original feature; hence in transverse section the fruit is obovoid. External surface ornamented with longitudinal ribs due to fibres within the wall. Whole surface covered by square or hexagonal cells from 0-027 to 0-055 mm. broad, arranged more or less regularly in longitudinal rows. Cell walls thin, possibly the cells were much inflated in life and have collapsed. Thickness of fruit wall now 0:055 to 0-228 mm. at most. Even supposing that the wall has been thinned by collapse of the cells on drying the maximum original thickness can scarcely have exceeded 0:35 mm. (contrast S. wetzler1 (Heer) from the German Brown Coal 3 to 6 mm. thick). Number of locules cannot be verified in the absence of new and better material, although it should be borne in mind that the supposed septum may have been inner layers of fruit wall separated by maceration. Seeds arranged in one or two rows on parietal placentas in the plane of symmetry. (S. wetzleri one- loculed with three parietal placentas and six rows of seeds.) Although no fruits are complete it is estimated that not more than a dozen seeds can have been enclosed by these small fruits. Eight are preserved in the holotype (numerous seeds in excess of twelve are shown in the figures of S. wetzlevi). Dehiscence by irregular fracture or dissolution of the fruit wall. Estimated length of fruit not less than 20 ormore than25mm. Maximum diameter,7togmm. (S. wetzleri 25 to loo mm. in length. Maximum diameter, 30 mm.) Seed: (V.20037-38a, V.42101) Anatropous ; varying in shape through mutual pressure in the pod, hence frequently somewhat facetted, elongate, often sub- cylindrical, apex rounded or pointed, base truncate directed towards the apex of the fruit. Base with a funnel-shaped depression which encloses the hilum, a lateral opening leading into the lateral raphe canal which traverses the thick testa to the FLORA OF THE LOWER HEADON BEDS 109 subapical chalaza (Chandler, 1925, text-fig. 5 (1)). At the base of the depression is the circular micropylar opening into the seed cavity. In well preserved specimens it is closed by a thin circular plug with central external mucro. The raphe opens into a small apical cavity at the extremity of the flask-shaped seed cavity and the chalaza is situated where the two cavities are adjacent. The testa is thick and in all layers the cells are arranged spirally around the seed cavity, an arrangement apparent in weathered or dissected specimens but not seen in a freshly sectioned surface or thin section. The epidermis is smooth and shining externally with spiral striae, deeper striae alternating with four or five slighter ones. The cells which form this surface are very long and narrow as described by Kirchheimer for S. wetzlevi but in sections of the seed they produce a coarsely columnar effect. Within the epidermis in longitudinal sections of the seed two closely fused thick layers are distinguishable. The outer of these, about o-Ir mm. thick in one seed measured, shows small equiaxial cells (secreting ? or endosperm cells ?). They represent the transverse diameters of spirally arranged cells (in S. wetzleyt this outer layer has tangentially elongate cells as seen in section). The inner layer is 0-275 to 0:33 mm. thick and the cells, in section equiaxial, are larger than those of the outer layer. A thin cuticle lines the seed cavity which is formed of hexagonal cells, 0-or8 mm. in diameter. The spiral arrangement of the inner layers is clearly seen impressed on the lining of the seed cavity. Length of seeds, 6:25 to 7-25 mm.; diameter, 1°5 to 2:25 mm. (S. weftzleri length, 5 to 10-2 mm. ; diameter up to 5 mm.). REMARKS. Fruits and seeds were formerly obtainable in Bed 10 below Hordle House and it appears fairly certain that the holotype was from this site where Mr. & Mrs. Clement Reid were in the habit of collecting. Seeds were rare in Beds 27-29 (V.20038). A seed embedded in a mass of Acrostichum (V.42057a) was alleged to be from Barton but see comment on p. 102. In 1936 and again in 1957 the structure of Spivematospermum wetzlert (Heer) from the German Brown Coal was fully described by Kirchheimer. Much help in inter- preting details of structure in the Headon material was obtained from his work. The existing condition of Hordle cliff makes it impossible to find new material for further study and this is unfortunate as only three fruits (all imperfect) and a number of isolated seeds are known. All the fruits show consistent characters which now seem to distinguish them from the typical Brown Coal S. wetzler1. These distinctive characters are indicated in the above description by introducing the corresponding features of S. wetzleri in brackets. There seems to be no doubt that the bisymmetry of the Hordle fossils is original and the fruits were never round in transverse section. The almost papery thin walls are unlikely through shrinkage on drying to have been derived from walls 3 to 6 mm. thick like those of S. wetzlevi. The number of locules cannot finally be settled at present. However apart from this doubtful character the shape, size and thickness of the German fruits, the larger number of seeds, and the size of these seeds seem sufficient ground for separating the specimens from Hordle as a distinct species, S. headonense. Kirchheimer is in agreement that the genus Sfivematospermum is an extinct member of the family Zingiberaceae. 110 FLORA OF THE LOWER HEADON BEDS The figures and description of Spivematosbermum from the Bovey Tracey Lignite, Devon, give insufficient evidence as to the species which has therefore been retained in S. wetzleri as originally described. The small size of the isolated seeds from Bembridge suggests that these should possibly be redesignated S. headonense although they fall within the range of size of the Oligocene S. wetzlert. Class DICOTYLEDONES Family MyRrIcACcEAE Genus MYRICA Linnaeus Myrica boveyana (Heer) pars (Pl. 25, figs. 33, 34) 1862. Cayrpolithes boveyanes Heer, p. 1077, pl. 70, ? figs. 7-14 (in part). 1957. Myrica boveyana (Heer) pars: Chandler, p. 90, pl. 12, figs. 45-48. 1961. Myrica boveyana (Heer) pars: Chandler, p. 328, pl. 33, figs. 1-4. DEscrIPTION. LEndocarp: One-loculed, bisymmetric, splitting into equal valves. Valves more or less oval in outline, somewhat inflated. External surface so much abraded as to show little structure but a few obscure rather irregular longitudinal furrows for fibres. Locule suburceolate, not nearly as deep as broad, narrowing above into the apical stylar canal. Wall pierced at the base by a short straight canal for the funicle. Marginal sutures wide, smooth, flat, close-textured, broadest at about a quarter of the length from the apex where they may be 0-6 to 0-7 mm. wide. These wide sutures tend to produce a slight marginal flange on the outer surface which appears to break readily owing to its thinness as shown in one specimen. At the base of the endocarp the suture is narrower and only about 0-2 mm. wide. Seed not preserved. Length of endocarp, 2°75 mm. ; maximum breadth, 2°5 mm. REMARKS. Two valves from different endocarps. Although so abraded and sand-pitted that the cell structure is obscure, they appear to be indistinguishable from M. boveyana (Heer) from the Oligocene Lake Basin of Bovey Tracey, Devon, and from the London Clay of Nursling, Southampton, judging by size, proportions and locule shape. Family BETULACEAE Genus CARPINUS Linnaeus Carpinus boveyanus (Heer) (Pl. 25, fig. 35) 1862. Carpolithes boveyanus Heer, p. 1077, pl. 70, figs. 7-14 (in part). 1957. Carpinus boveyanus (Heer) pars: Chandler, p. 92, pl. 13, figs. 52-67. DEscRiPTION. Fruit: Pointed-ovate in outline, bisymmetric, much compressed with large basal scar of attachment of which half is on each face. Surface covered FLORA OF THE LOWER HEADON BEDS III by an accrescent calyx. Longitudinal furrows arise at the margin of the scar ; three can be seen on one face, four on the other but owing to abrasion they become obscure in places, some being entirely obliterated. The vascular strands they once contained are not preserved. External surface finely striae, the striae formed by elongate cells about 0-009 to 0-012 mm. broad. Beneath this surface and partly exposed by abrasion are equiaxial cells about 0:025 mm. in diameter. Length of fruit, 2°8 mm.; breadth, 2:25 mm.; thickness, 0-75 mm. Diameter of basal scar, I mm. ; height measured on one surface, 0°25 mm. REMARKS AND AFFINITIES. One fruit with accrescent calyx preserved but no bract. Comparison with fruits from the Bournemouth Marine Beds at Southbourne, the Highcliff Sands (both awaiting publication) and the Bovey Tracey Beds, Devon, indicate that a single species occurs at all these localities. Family MorACEAE Genus BROUSSONETIA L’Hérit. Broussonetia rugosa Chandler (Pl. 25, figs. 36, 37) 1925. Broussonetia rugosa Chandler, p- 21, pl. 3, fig. ra, dD. Diacnosis. Endocarp suboval in outline with knob-like projection carrying stylar canal and funicular opening, bisymmetric, somewhat compressed at right angles to plane of symmetry. Surface finely pitted, ornamented with elongate tubercles with sinuous outlines. Thick inner part of wall columnar. Dimensions : 1-75 by I:4mm.; 1°25 by I mm. DESCRIPTION. Endocarp: Suboval in outline, with knoblike projection at one end, bisymmetric, somewhat compressed at right angles to the plane of symmetry with broad surfaces gently convex. Stylar canal and funicular canal adjacent on the projection. Endocarp splitting marginally for germination, having a single pendulous seed. External surface ornamented with elongate tubercles having irregular sometimes sinuous outlines. The tubercles radiate from the stylar pro- jection. Whole surface finely pitted. Thickness of wall about o-I mm., seen in section to be formed of an outer thin compact layer and an inner thick layer of columnar cells. Diameters of two endocarps in plane of symmetry: 1°75 by 1°4 mm. and 1-25 by I mm. Remarks. In addition to the two figured specimens there are three fragments (V.20039a). The projection at one end of the endocarp is seen in living material of Broussonetia and Morus to taper into the funicle but in the fossils the funicle itself is gone and only the worn down process remains. The fossils resemble Ficus in size and form but no rugose species of Ficus has been seen whereas in Broussonetia all the characters described are present, 112 FLORA OF THE LOWER HEADON BEDS Section MOROIDEAE Genus MOROIDEA Chandler, 1957 : 95 Moroidea hordwellensis n. sp. (Pl. 25, fig. 38; Text-fig. 1) DiaGnosis. Fruit more symmetric than M. boveyana, with broad flat stylar projection and canal. Dimensions in plane of symmetry: 1-7 by 1:5 mm. Hototyre. Brit. Mus. (N.H.), No. V.42105. DESCRIPTION. Fruit: Bisymmetric, subcircular in outline, cuneate in trans- verse section, the narrow edge crested for nearly half the circumference and the opposite edge rounded. Style and funicle close together each marked by a con- Fic. 1. Moroidea hordwellensis n. sp. Longitudinal section through fruit and seed. (st) stylar prominence, (f/) placenta, (ch) chalaza on closely adherent seed within fruit. X2I approx. spicuous marginal prominence. The stylar prominence is broad, flattened terminal on the crested margin and pierced by a flat canal o-I mm. broad ; the attachment or funicular prominence is recurved away from that of the style, subapical, terminal on the rounded margin. Pericarp oI mm. thick along the crest, 0-05 mm. thick on the rounded margin, hard, woody formed of compact parenchyma, the cells on the external surface measuring about 0-012 to 0-016 mm. in diameter and in sections of the wall showing a radial alignment. Locule lining of approximately hexagonal cells about 0-025 mm. broad and 0-05 mm. long, their longer axes diverging from the placental region giving rise to obscure divergent striae. Dehiscence occurs along the crested margin. Seed solitary, pendulous from the subapical placenta which immediately underlies the funicular prominence, the funicle being short and thick. Dimensions of fruit in plane of symmetry, I-7 by 1-5 mm. ; thickness at right angles to plane of symmetry, 1mm. Length of funicular prominence, 0:17 mm. Seed : Campylotropous, chalaza conspicuous, subcircular, 1:2 mm. in maximum diameter contiguous with the hilum but on the far side of it from the style as the FLORA OF THE LOWER HEADON BEDS 113 seed is seen lying within the fruit. Testa brittle, only 0-025 mm. thick tending to adhere to the locule, formed of small equiaxial cells 0-012 to 0-016 mm. in diameter near the hilum where best preserved. Inner surface of testa with elongate cells about 0-025 mm. broad, having cell walls represented by a double line. REMARKS AND AFFINITIES. One fruit slightly broken near the crested margin and, on one surface, near the attachment. The broken surface was subsequently removed and the remains of a seed exposed. Only the hilar end of the seed was preserved but it showed the perfect chalaza scar. Form and structure indicate Moraceae and probably the section Moroideae (Engler) or the Artocarpoideae in both of which the seeds may be campylotropous and the chalaza in the position described. On the whole the Moroideae appear to bear the closer resemblance to the fossil. A species distinguished by a narrower stylar prominence and canal and greater asymmetry was described from Bovey Tracey under the name Moroidea boveyana Chandler (1957 : 95, pl. 13, fig. 74). Genus BECKTONIA nov. DiacGnosis. One-loculed one-seeded endocarps referable to Moraceae, roundly triangular at apex the angles alternating with more or less flattened sides, rounded below, with oblique apical foramen and placenta subjacent to it. Length, 3:3 to 4°25 mm. ; breadth, 2-4 to 3:5 mm. (compressed). Type SpPEcIES. Becktonia hantonensis n. sp. Becktonia hantonensis n. sp. (Pl. 25, fig. 39) Diacnosis. That of the genus. Hototyre. Brit. Mus. (N.H.), No. V.42106. DeEscripTION. Endocarp: Subovoid, roundly triangular above, rounded below, the rounded apical angles dying out gradually but subsidiary intermediate ridges occurring in the lower part. Foramen for funicle subapical oblique. Surface rather rough formed of fine equiaxial cells 0-007 to 0-008 mm. in diameter. Wall about 0-15 mm. thick, close textured in section, carbonized so that cell structure is obscure but a radial or columnar alignment can be detected. Length, 4:25 mm. ; breadth (reduced by contraction), 2 mm. REMARKS AND AFFINITIES. An endocarp bursting below. The specimen is identical with endocarps from the Bournemouth Freshwater Beds awaiting publica- tion. These endocarps range in length from 3:3 to 3-5 mm.; breadth, 2-4 to 2°75 mm. The single seeded locule with hard endocarp, locule lining with digitate cells (seen in the Bournemouth specimens), oblique funicular aperture and pendulous anatropous seed (also seen in Bournemouth specimens) indicate the family Moraceae, 114 FLORA OF THE LOWER HEADON BEDS Order CENTROSPERMAE Family CARYOPHYLLACEAE Genus HANTSIA Chandler, 1960 : 209 Hantsia pulchra (Chandler) (Pl. 25, fig. 40; Text-fig. 2) 1925. Corydalis pulchva Chandler, p. 25, pl. 3, fig. 10a, b. 1929. Corydalis pulchrva Chandler: Reid & Chandler, pl. 1, fig. 18. 1960. Hanisia pulchva (Chandler) Chandler, p. 209, pl. 31, fig. 39. DESCRIPTION. Seed: Obliquely or transversely oboval or subcircular in outline, much compressed laterally, bisymmetric about a plane passing through hilum and micropyle. Seed cavity undivided by any partition between the limbs of the large h Fic. 2. Hantsia pulchva (Chandler). Longitudinal section through a seed showing large embryo (inferred) and large hilar opening (#) between limbs. X15. curved embryo of which the radicular limb is longer and narrower than the other. The embryo may have occupied the whole seed cavity or have been surrounded by only a thin film of endosperm as is suggested by the surface contours (cf. Pl. 25, fig. 40). Hiluma large oval or subcircular gaping aperture between the tips of the limbs with smooth raised rim and cells with straight walls. Testa formed of two coats, the outer 0-1 to 0-15 mm. thick, black and shining, closely and conspicuously tubercled, the tubercles being dome-shaped and subcircular in outline except around the margin and over the micropylar limb where they tend to be elongate parallel with the margin. Each tubercle arises from a single “cell’’ or area with finely toothed outline, the teeth measuring about 0-02 mm. in breadth, the diameter of the tubercled areas being about 0:07 to o-I mm. The teeth are frequently but not invariably difficult to see in surface view but are clear on broken edges where they produce a columnar effect on account of the rounded ridges they form on the sides of the cells. The testa readily cracks along these toothed outlines. The tubercled areas sometimes break down and then appear to be formed of small equiaxial cells 0-016 to 0-025 mm. in diameter. An inner black coat, 0-016 mm. thick where measured, may become partly separated in fossilization from the outer FLORA OF THE LOWER HEADON BEDS 115 coat. Its surface is close textured formed of equiaxial cells about 0-015 mm. in diameter. Tegmen light brown, semitranslucent, of straight-sided angular cells about 0-012 to 0-015 mm. in diameter, rough. Dimensions of typical seeds are: 1) Length along axis between limbs, 2 mm.; breadth across both limbs, 2°5 mm. 2) Length, 2:25 mm. ; breadth, 1-75 mm. REMARKS. The species was originally referred to the genus Corydalis but fuller information now available suggests that the affinities need revision. Fine teeth around the tubercled surface cells (or areas) were originally overlooked but have now been clearly seen. No such digitate cells occur in Corydalis where the walls of the tubercled areas are straight and smooth. Again in Corydalis the tegmen is thin and diaphanous adhering closely to the outer coat and is formed of large irregular cells which vary in size from 0-016 by 0:05 mm. to 0-012 by 0:08 mm. This in no way resembles the tegmen of the fossil with its small uniform equiaxial cells. Further search was therefore made in the order Centrospermae which has large curved embryos occupying the circumference of the seed. Among the Chenopodiaceae Beta shows superficially similar cells but these are smooth walled not digitate ; its tegmen is similar to that of the fossil. Portulacaceae, with digitate cells, is excluded by the form of the seed. In Caryophyllaceae the tubercles are always finely digitate while in Saponaria they may be similar in appearance and distribution. Again the limbs of the curved seed may be unequal as in the fossil but there is no conspicuous hilar rim projecting between them although the tegmen is formed of small equiaxial cells. In Silene a hilar rim occurs but shows distinct digitate cells on its surface and the testa cells are radially aligned the limbs of the seed being more or less symmetrical. No living genus seen shows all the characters described but as they do occur within the family Caryophyllaceae the nearest relationship is pres- sumed to lie here. The generic name indicates the geographical origin of all specimens to date. In addition to the figured material there are a number of broken or incomplete seeds (V.42108-09) some of which show important characters. Hantsia glabra n. sp. (Pl. 25, figs. 41-43) Diacnosis. Resembling H. pulchra except in the smooth surface, the digitate cells being flat or scarcely convex, not markedly tubercled. Dimensions of best developed seed: 2-3 by 2 mm. Hototyre. Brit. Mus. (N.H.), No. V.42rrr. DESCRIPTION. Seed: Subcircular or transversely oboval in outline, much com- pressed laterally, bisymmetric about a plane passing through hilum and micropyle, embryo curved with unequal limbs the micropylar limb longer and narrower than the other. Seed cavity not divided by a partition between the limbs. Hilum a large oval or subcircular gaping aperture between the limbs or almost terminal on the narrower limb surrounded by smooth upstanding rim as in Hantsia pulchra. Testa black and shining formed of cells which are subcircular or oval in outline with finely toothed margins having from about nine to fifteen small teeth, The 116 FLORA OF THE LOWER HEADON BEDS surface of these cells is flat or very slightly convex not tubercled as in H. pulchra. The testa has a marked tendency to break along the toothed outline of the cells. Surface of cells finely granular. Size of cells varying from about 0-075 to o-I mm. in diameter but on the tip of the micropylar limb they are much smaller and some at least are straight-sided not toothed. Along the circumference of the seed they tend to be elongate and narrower. These toothed cells show clearly on the inner surface of the testa which is about 0:05 mm. thick. There are traces of a thin light brown inner coat (tegmen?) not clearly seen. The manner in which the cells of the outer coat break up suggests that they may be formed of groups of smaller equiaxial cells which perhaps give rise to the granular appearance. Greatest diameter of seed measured across two limbs, 2-3 mm.; greatest diameter at right angles to this in plane of symmetry, 2 mm.; thickness at right angles to plane of symmetry, 0‘-¢ mm. Hilar aperture including thickness of wall, o-r by 0-15 mm. A second seed is distorted so does not give correct measurements but may have been somewhat smaller. REMARKS. Obviously related to Hantsia pulchra which it closely resembles in everything but its smooth surface and in the apparent absence of an inner coat of the testa formed of small equiaxial cells. The possibility that both may belong to a single species with some smooth and some tubercled seeds as in Caryophyllaceae and Fumariaceae cannot be excluded, but the difference in the fossils is so pro- nounced that pending further discovery as to closer relationship it appears best to regard them as distinct. In this as in H. pulchra a depressed area between the limbs on the broad surfaces suggests a large embryo, probably with little albumen, having a conspicuous large radicle and broad cotyledonary limb. Family NYMPHAEACEAE Genus BRASENIA Schreber Brasenia ovula (Brongniart) 1926. Bvrasenia ovula (Brongn.): Reid & Chandler, p. 99, pl. 6, figs. 15-18 (see for earlier references). 1957. Brasenia ovula (Brongn.): Chandler, p. 96, pl. 13, fig. 75. 1960. Byrasenia ovula (Brongn.): Chandler, pp. 210, 224, pl. 31, figs. 40, 41; pl. 33, figs. 106, 107. Seeds abound in the Lower Headon both at Hordle and Colwell Bay, certain horizons being blackened by their shining testas. At Hordle they are especially abundant in Beds 9 and 10 below Hordle House (V.20045-46, V.20048, V.42112-14 embryotegas) and just east of Beckton Bunny (V.42115), Bed. 29 (V.20047) and Bed 31 (V.20044). Site in Bed ro unspecified (V.31790). At Colwell Bay a number of seeds collected by Clement Reid are presumed to come from the Lower Headon (V.42116-18). Those by J. Groves (V.42119) are from the Limnocarpus Band and those by Chandler (V.42120) from this same horizon below Warden Battery. FLORA OF THE LOWER HEADON BEDS 117 Brasenia spinosa Chandler 1925. Brasenia spinosa Chandler, p. 22, pl. 3, fig. 4; text-fig. 7. 1925. Byrasenia antiqua Chandler, p. 22, pl. 3, fig. 5a, b. Diacnosis. Seed globular or ovoid, raphe ridge inconspicuous. Surface with faint longitudinal ridges and slender spiney tubercles or broad, short or wart-like tubercles especially abundant over the chalazal end of the seed. Surface of inter- locking digitate cells. Embryotega circular or elliptical with central rimmed micropyle and oblong or square convex surface cells with finely toothed margins. Tegmen semitranslucent of equiaxial angular frequently hexagonal thin-walled cells. Length of seed, 1 to 2-6 mm. ; breadth, 0:8 to 2 mm., exceptionally 2-5 mm. Hototyre. Brit. Mus. (N.H.), No. V.20041a@ originally regarded as a distinct species named Brasemia antiqua. The original and only seed of B. spinosa now decayed. DeEscripTIoN. Seed: Globular or ovoid, anatropous, with inconspicuous longitudinal raphe ridge, hilum obscure where the ridge terminates against the embryotega. Surface with faint rounded longitudinal ridges, sometimes orna- mented with slender spines or tubercles, sometimes with shorter, thicker, blunt or wart-like tubercles. The tubercles are usually most abundant over the chalazal end of the seed. The rounded ridges frequently correspond with the rows of cells which form the testa. The cells which may be about 0-05 to o-r mm. in diameter are digitate some having as many as ten digitations, others only six to eight ; these may be arranged in three or four pairs. Each tubercle or spine arises from the centre of a cell although not all cells bear tubercles. In section the testa may be about 0°125 to 0:15 mm. thick and shows radial columns with fluted walls, the flutings corresponding with the surface digitations. The testa tends to break along the outlines of these cells. Embryotega with a large rimmed median micropyle. Its cells are square or oblong, much inflated with finely toothed margins. Some of the embryotega cells are as much as 0:05 mm. in diameter. The tegmen is semi- translucent, formed of angular frequently hexagonal thin-walled cells about 0-075 mm. in diameter. The seeds vary greatly in size. Typical measurements are: 1) Length, 1-25 mm. ; breadth, 1-25 mm. 2) Length, 1-25 mm. ; breadth too crushed to measure. 3) Length, 1-5 mm.; breadth, 1-25mm. 4) Length, 1 mm.; breadth, o‘8 mm. 5) Length, 1:2 mm.; breadth, 1 mm. 6) Length, 2-6 mm.; breadth, 2°5 mm. Specimens from the Upper Headon of Colwell Bay range in size from: length, I-25 to 2-5 mm.; breadth, I to 2 mm. Remarks. After examining a large number of specimens from the Upper Headon of Colwell Bay, it appears that Brasenia antiqua must be regarded as identical with B. spinosa, the two types with spiney tubercles and short thick tubercles respectively appear to grade into one another so that they cannot be separated. The variety with thick tubercles was fairly abundant at Hordle, the thin spined variety was scarce. Its small size as well as the large tubercles appear to distinguish this species from the living B. schreberi and from all other fossil forms 118 FLORA OF THE LOWER HEADON BEDS so far seen. B. spinosa is known from Bed 10 below Hordle House (V.20041, Chandler, 1925, pl. 3, fig. 5a; V.20041a, Chandler, 1925, pl. 3, fig. 5b; V.43737) and Bed to east of Beckton Bunny (V.42121, a). Brasenia oblonga Chandler (Pl. 25, figs. 44-46) 1925. Brasenia oblonga Chandler, p. 23, pl. 3, fig. 6. Diacnosis. Seed elongate, rounded cylindrical, somewhat truncate at the apex, raphe forming a longitudinal ridge, aperture for embryotega large. Surface tubercled especially at base and along raphe ridge. Also with obscure longitudinal ridges. Length, 1°75 mm.; breadth, 0-9 to I mm. Hototype. Brit. Mus. (N.H.), No. V.20043. DEscRIPTION. No additions to this are needed. REMARKS. Both specimens had lost the embryotega. Family MENISPERMACEAE Section CoccuLEAE Diels Subsection CoccuLInEAE Diels Genus PALAEOSINOMENIUM Chandler, 1961 : 159 Palaeosinomenium obliquatum (Chandler) (PL. 25, figs. 47-50) 1925. Menispermum obliquatum Chandler, p. 24, pl. 3, fig. 9a, b. DiaGnosis. Endocarp curved somewhat obliquely, ventral margin slightly concave. Radially aligned ridges conspicuous outside the horse-shoe shaped ridge and on the marginal flange. Foramen between the limbs elongate, its lower end well above a line drawn between the extremities of the two limbs of the endocarp. Diameter about 3 by 4 mm. Hototyre. Brit. Mus. (N.H.), No. V.20049 (now shattered). Remarks. Although the holotype is now much shattered, the characters are also shown by the fresh but imperfect material here illustrated. In the light of fuller experience the species is transferred to the fossil genus Palaeosino- menium which is defined as having a more oblique endocarp and a more elongate foramen than the living Sinomenium and Menispermum. P. obliquatum is smaller than P. venablesi (Chandler, 1961: 159, pl. 16, figs. 9-13) and somewhat smaller than P. pulchrum (Chandler, 1961: 329, pl. 33, figs. 5-7) which it very closely resembles but it also has a relatively narrower marginal flange than P. pulchrum, FLORA OF THE LOWER HEADON BEDS 119 Family LAURACEAE Genus? (Pl. 25, figs. 51, 52) DEscRIPTION. Berry: One-carpelled, probably originally subglobose, com- pressed dorsiventrally in fossilization, seated on a shallow cupule or swollen flattened peduncle, about 3 mm. in diameter, the scar of attachment on the berry about 1°8 mm. in diameter. Epicarp shining, 0-025 to 0-05 mm. thick, with scattered circular depressions, formed of a layer of equiaxial cells which are slightly convex super- ficially and about 0-016 mm. in diameter. Mesocarp (heavily impregnated with amorphous pyrites as in the London Clay fossils) varying considerably in thickness owing to uneven contraction on drying or crushing, probably averaging originally about 1°5 to 2 mm., its structure obscure. Endocarp 0-15 mm. thick, pointed at the apex, rounded conical, formed of columnar cells in section which are about 0-016 mm. broad. Testa and chalaza not seen but seed probably conforming to the outline of the endocarp the chalaza occupying the base of the rounded conical locule. Transverse diameter of berry (as crushed), 5-6 mm. ; height (reduced by compression), 3°5 mm.; diameter of endocarp (not crushed), 3-3 mm. ; height of endocarp, 2°6 mm. REMARKS AND AFFINITIES. Whereas the relationship to Lauraceae is clear, without more and better material no satisfactory generic determination is possible. Clearly the specimen belongs to a small fruited genus with thick endocarp and small cupule (or swollen flattened peduncle). Cinnamomum, Litsea, and Lindera appear to be possible relationships. The circular depressions of the epicarp occur in Cinnamomum. C. Reid records leaves of Cimnamomum but without figures or descriptions. His specimens are not extant (Reid & Groves, 1921 : 178). Family DROSERACEAE Genus ALDROVANDA Monti Aldrovanda ovata (Chandler) (Pl. 26, figs. 53-55) 1925. Nuphar ovatum Chandler, p. 22, pl. 3, fig. 3a, b. 1926. Aldrovanda ovata (Chandler) Reid & Chandler, p. 113, pl. 6, figs. 24-26. Hototyre. Brit. Mus. (N.H.), No. V.17485. Remarks. The discovery of the true relationship of these seeds by Docturowsky following upon work on living Aldrovanda by Nikitin and Korzhinsky is reported by Reid & Chandler (1926: 112). Specimens (V.17485-87) from Hordle (precise locality and horizon not specified) were figured by Chandler (1925) and by Reid & Chandler (1926). Also from an unspecified horizon are V.20053a,b. V.42130 is from Bed ro east of Beckton Bunny and V.20052 from Bed 31. The same species characterized by its long neck, conspicuous apical mucro, marked raphe ridge and 120 FLORA OF THE LOWER HEADON BEDS relatively smooth surface, with a thick testa of two coats together measuring nearly a quarter of the diameter of the seed occurs at Colwell Bay in beds presumed to have been Lower Headon (V.42124, V. 42127) and in the Limnocarpus Band (V.42128-29). Some seeds both from Hordle and Colwell show the small cir- cular plug closing the narrow neck. It has a central short thick mucro associated with the micropyle from which small convex equiaxial cells diverge over its outer surface increasing in size towards the circumference of the plug. Hilum at the margin of the plug. Dimensions of a number of seeds are as follows: 1) Length of seed, 1°75 mm.; breadth, 1:35 mm.; length of neck, 0-25 mm. 2) Length, 1-76 mm. ; breadth, 1-4 mm.; length of neck, 0-3 mm. 3) Length, 1-9 mm.; breadth, 1-25 mm. ; length of neck, 0-25 mm. 4) Length, 2 mm.; breadth, 1:25 mm. ; length of neck, 0:25 mm. 5) Length, 1°75 mm.; breadth, 1:25 mm.; length of neck, 0:25 mm. 6) Length, 2 mm.; breadth, 1:25 mm.; length of neck, 0-3 mm. The average length of seeds is 1°83 mm.; average breadth, 1:28 mm. i.e. relatively longer in proportion to the breadth than the specimens from Hordle measured by Reid & Chandler (1926 : 113). Family HAMAMELIDACEAE Genus EOLIQUIDAMBAR nov. Diacnosis. Agglomerate fruiting head with deep areoles. Carpels splitting into two equal valves which are not produced into awns. Type SpEcIES.—Eoliquidambar hordwellensis n. sp. Eoliquidambar hordwellensis n. sp. (Pl. 26, figs. 56-59) Diacnosis. That of the genus. HoLotype. Brit. Mus. (N.H.), No. V.421314. DESCRIPTION. Fruiting Head: Pedunculate, woody agglomerate, ovoid, formed of about eighteen to twenty-five deeply hollowed polygonal areoles each holding a single fruit. Wall between areoles thick, formed of woody fibres radiating from the axis of the fruit. Surface of these walls covered by a shining wrinkled epidermis, the wrinkles parallel with the margins of the areoles. Each wall shows a conspicuous median furrow at the surface, the furrows indicating the plane of fusion between the walls of adjacent areoles. Length of the ripe fruiting head, 17-5 mm. ; trans- verse diameter, 12°5 mm. Diameter of areoles at surface about 7°5 to 8 mm. Fruitlets ; Capsular, probably two-loculed, splitting at the apex into two valves which are not produced into awns. At the surface of the head the valves are crescentic in outline and are covered by a light coloured shining wrinkled epidermis formed of oblong cells, o-or mm. in greatest diameter, aligned parallel with the crescents at the middle of the valves and at right angles to them at the margins. Endocarp and seeds not seen. Length of crescents, 3°5 to 4:5 mm.; breadth, I°5 mm. FLORA OF THE LOWER HEADON BEDS 121 REMARKS. In addition to the holotype there are immature fruiting heads: V.42131 is from Bed 10 between Beckton Bunny and Long Mead End and V.42132 from Bed ro east of Beckton Bunny. The heads even in the immature condition are easily distinguishable from Protoaltingia hantonensis by the single pair of crescentic valves in each areole whereas in Protoaltingia even in young heads four knob-like valves are seen. The fibrous agglomerate character and the cap- sular fruits in the areoles indicate the family Hamamelidaceae section Altingieae. That the absence of awns at the apex of the fruits is an original feature and not due to abrasion is clearly indicated by the well preserved unbroken continuous epidermis covering the external margins of the valves. All living species of Liguidambar seen have long persistent styles which form gradually attenuated awns to the valves on dehiscence of the fruit while the margins of the areoles have a complex lobed structure quite unlike the smooth continuous surface of epidermis seen in the fossil. At the same time the mode of dehiscence of the fossil fruitlets into two valves is like that of Liqguidambar. In this genus dehiscence is septicidal and the endo- carps are persistent. The mode of dehiscence is not shown in the fossil and there is no evidence whether or not the endocarps are shed. In Altingia and Protoaltingia although the styles do not form awns dehiscence is in two planes, septicidal and loculicidal, hence the apex of the fruit is four-partite giving rise to the four knob- like valves mentioned above. It appears, therefore, that the fossil does not agree in all respects with Altingia, Protoaltingia or Liquidambar and it has been referred to a new genus Eoliquidambar. Genus PROTOALTINGIA Chandler Protoaltingia hantonensis Chandler (Pl. 26, fig. 60) 1925. Liquidambar sp. Chandler, p. 25, pl. 4, fig. 1. Protoaltingia hantonensis Chandler (in press), pl. 9, figs. 18-24; pl. 28, figs. 6-8. DeEscRIPTION. Fruiting head: Always so much abraded that the complete head is not preserved so far as is known at Hordle, but the specimen (V.20054) originally figured appears to have been ovoid although the long axis was erroneously oriented transversely. Agglomerate, with numerous funnel-like alveolae showing fibrous but somewhat cavernous walls the bases only remaining. These alveolae hold the worn down persistent fruitlets, one in each alveole. Largest fragment now preserved about 5 to 7 mm. in diameter. Fruitlets : Capsular, two-loculed with antero-posterior locules formed of trans- verse fibres, elongate as seen in transverse section at right angles to the septum. Septicidal as shown by the curvature and separation of the walls at the septum. Evidence of loculicidal dehiscence almost destroyed by the wearing away of the distal ends of the fruit walls but indicated in one or two alveolae by the thinning and curvature of the walls in the plane at right angles to the septum. REMARKS. With the help of well-preserved material from the Bournemouth Beds these fragmentary remains can be better understood than in 1925 when they 122 FLORA OF THE LOWER HEADON BEDS were erroneously referred to Liquidambar itself through lack of understanding of the characters of the fossil combined with lack of knowledge of the living genus. The reference to Hamamelidaceae was correct. The account of the Bournemouth fruits is in the press. The Hordle specimens despite their poor condition agree with them in the characters seen. The elongate form of the fruitlets at right angles to the septum in transverse section allies them with Protoaltingia hantonensis and not with Eoliquidambar hordwellensis (p. 120). In addition to the figured fruit (now more broken) and V.20056 from the contemporary channel west of Long Mead End, there are also a few fragmentary heads from Bed Io, east of Beckton Bunny (V.20055, V.42133). Family RosACEAE Genus RUBUS (Tourn.) L. Rubus acutiformis Chandler 1925. Rubus acutiformis Chandler, p. 26, pl. 4, fig. 2. Rubus acutifoymis Chandler: Chandler (in press), pl. 4, figs. 36-40. There is nothing to add to the descriptions already published except that the endocarps occur in the greatest abundance in Bed Io at the cliff base below Hordle House (V.20057—58, V.42134) and more rarely east of Beckton Bunny (V.42135). Specimens from Studland (in press), Bournemouth Freshwater Beds and Cliff End (awaiting publication) appear to be identical but the Hordle specimens are much better preserved than those from other localities. Family LEGUMINOSAE? Genus? (Pl. 26, fig. 61) DESCRIPTION. Seed: Smooth broadly subovoid or subglobular having a very slight and inconspicuous bisymmetry, probably splitting down one side in the plane of symmetry where a finished suture can be seen. Hilum marked by an obscure projection on the margin. Surface of fine equiaxial cells, o-or mm. in diameter, the testa compact 0-2 to 0-4 mm. thick, becoming thinner towards the hilum, columnar in section the fine columns, about 0-or mm. broad and apparently formed of short cells placed accurately end to end. Length of seed, 4-5 mm.; breadth, 4 mm.; thickness rather less than breadth but as the seed has burst it cannot be measured accurately. The seed shows no outstanding readily recognizable charact- ers, but the general appearance and testa structure suggest Leguminosae. Genus? (Pl. 26, fig. 62) DESCRIPTION. Seed: Oval in outline, much compressed, compression increased in fossilization so that the opposite walls are now in contact, having a small marginal FLORA OF THE LOWER HEADON BEDS 123 circular hilum, 0-25 mm. in diameter, situated near the end of one of the long margins. This hilar scar is surrounded by a somewhat rough tumescent area on both surfaces. Elsewhere the surface of the seed is smooth and shining, one side showing an obscure curved groove sub-parallel with the circumference, its extremities arising at the hilar region. Testa superficially pitted, pits 0-004 to 0-005 mm. in diameter, some- what obscured over much of the surface by tension fissures and striae. Thickness of testa 0-15 to 0-25 mm., formed of radially aligned cells having an obscure columnar appearance. Inner surface of convex equiaxial cells 0-I mm. in diameter. Length of seed, 5-5 mm.; breadth, 3-75 mm.; thickness, I mm. REMARKS AND AFFINITIES. One seed broken obliquely into two fragments, with a third small fragment missing on the dorsal margin. The form of the seed and character of the hilum suggest possible relationship with Leguminosae, but the nearer affinities have not been found. There is no distinct inner coat inside the colum- nar coat as in Acacia or Albizzia nor is the testa conspicuously columnar as in these and many other Leguminosae seeds. Family RUTACEAE Genus ZANTHOXYLUM Linnaeus Zanthoxylum hordwellense n. sp. (Pl. 26, figs. 63, 64) 1925. Zanthoxylon cf. ailanthoides Sieb. & Zucc.: Chandler, p. 26, pl. 4, fig. 3a, b; text-fig. 9. DiAGnosis. Seed with very slightly convex ventral margin, sides less inflated than living Z. ailanthoides. Surface with concentric wrinkles but few or no con- necting ridges. Orifice to raphe canal ventral not basiventral. Length of seed, 2:25 to 4:5 mm.; breadth, 2:25 to 3-5 mm.; thickness, 1-85 to 2°75 mm. HoLotyre. Brit. Mus. (N.H.), No. V.20059. DESCRIPTION. Seed: Semi-anatropous, gibbous in outline having the ventral margin slightly convex, the dorsal rounded. Somewhat laterally flattened especially on each side of the ventral margin which is occupied by an _ elongate- triangular depressed hilar scar. Chalaza basal on the ventral side of the longest axis of the seed, marked by a large circular internal aperture (V.20059a) ; raphe canal prominent with a conspicuous orifice at the lower end of the hilar scar (PI. 26, fig. 64) ; micropyle terminal on the ventral margin at the opposite end of the scar to the raphe orifice ; epidermis thin, shining, black, ornamented with a few concentric wrinkles and fine polygonal equiaxial pits ; within is a thick coat formed of equiaxial cells. Dimensions of seeds: 1) Length, 3-5 mm.; breadth, 3 mm. ; thickness, 2°75 mm. 2) Length, 4 mm.; breadth, 3 mm.; thickness, 1-85 mm. ; length of hilar scar, 3 mm. 3) Length, 3:25 mm.; breadth, 2-5 mm.; thickness, 2°25 mmi.; length of hilar scar, 2-5 mm. 4) Length, 4-5 mm.; breadth, 3:5 mm. ; thickness, 2:25 mm. Other small seeds varied in length or breadth from 2-25 to 275mm. A typical scar measured 0:65 by o-Ig mm. REMARKS AND AFFINITIES. Seeds fairly abundant, although very liable to decay, GEOL. 5, 5 14 124 FLORA OF THE LOWER HEADON BEDS occurring in Bed 10 below Hordle House (V.20059-60) and east of Beckton Bunny (V.42136). Usually the thin black epidermis has been destroyed and the thick inner integument is heavily impregnated with pyrites. It was pointed out (Chandler, 1925 : 27) that there is a close resemblance to Zanthoxylum (Fagara) atlanthoides Sieb. & Zucc. from which the fossil differs in the rarity or absence of transverse ridges connecting the concentric wrinkles, in its lesser degree of inflation, in its less convex ventral margin and its ventral orifice to the raphe canal as contrasted with the basiventral orifice of the living. Additional fossils since found merely increase the known variation in size of the species which further experience suggests deserves a distinct name. It is therefore called Z. hordwellense. Zanthoxylum compressum Chandler (Pl. 26, figs. 65-67 ; Text-fig. 3) 1925. Zanthoxylon compressum Chandler, p. 28, pl. 4, fig. 5. Diacnosis. Seed rounded subquadrangular in outline, broadest dorsiventrally. Hilum very short, concave. Length of seed parallel with hilar scar, 3:25 to 3:5 Fic. 3. Zanthoxylum compressum Chandler. Longitudinal section through a seed to show relative positions of hilar scar (h), hilar opening (ho), raphe (v) and chalaza (ch). x 12 approx. mm.; maximum diameter, 3°75 to 4 mm. approximately at right angles to scar. HoLotyre. Brit. Mus. (N.H.), No. V.20061. DESCRIPTION. (Slightly emended) Seed: Semi-anatropous, rounded subquad- rangular in outline, broadest dorsiventrally, much compressed. Ventral margin short, concave, occupied by the sunk subtriangular hilar scar. Opposite margin rounded, the internal chalaza being situated where this rounded margin gradually merges into the lower long lateral margin of the seed (as revealed by a break in the testa of the second specimen). Raphe, connecting the lower broad end of the hilar scar with the chalaza, a long canal which traverses the testa within the long lower FLORA OF THE LOWER HEADON BEDS 125 lateral margin of the seed. Outer surface black and shining ornamented with finely crenulate pits like those on the middle black coat of some living species. A second seed (V.42137) was abraded and rough superficially but retained a tiny fragment of a black shining superficial layer with slightly convex, equiaxial angular cells about 0-05 mm. in diameter. These cells are separated by fine hair-like straight raised walls. Tegmen thin, of equiaxial cells 0-016 to 0-025 mm. in diameter. Length of holotype (parallel with hilar scar), 3-5 mm.; breadth (maximum dia- meter), 4 mm. Length of second abraded seed, 3:25 mm.; breadth (maximum diameter), 3°75 mm.; thickness, 1-5 mm. In this specimen the hilar scar was considerably shorter than in the first seed. Thickness of wall where broken at chalaza, 0-3 mm. RemMARKs. Two seeds clearly related to Zanthoxylum distinguished specifically by the broad subquadrangular rounded form and very short hilum. Genus PHELLODENDRON Rupr. Phellodendron costatum Chandler 1925. Phellodendyon costatum Chandler, p. 28, pl. 4, fig. 6a—c. Phellodendron costatum Chandler: Chandler (in press), pl. ro, fig. 1. Diacnosis. Seed sometimes with knob-like subapical constriction. More inflated than any living species. Longitudinal ribs unusually well marked. Length, 3 to 3°5 mm., breadth, 1-8 to 2:25 mm. ; thickness 1-8 to 2 mm. HoLtotype. Brit. Mus. (N.H.), No. V.20062. EMENDED DESCRIPTION. Seed: Semi-anatropous, subovoid, bisymmetric, with a more or less straight ventral margin and a rounded dorsal margin in the plane of symmetry ; apex curved slightly towards the ventral side, sometimes with a knob due to an inconspicucus subapical constriction. Hilum strap-shaped, extending almost the whole length of the ventral margin, micropyle sunk in a depression at the apical end of the hilar scar i.e. on the knob when present ; raphe canal opening at a foramen at the base of the hilar scar, short, curved; chalaza a large circular internal aperture at the end of the major axis of the seed away from the micropyle. Germination by splitting in the plane of symmetry ; surface sculpture of numerous acute longitudinal ribs (about twenty-four were counted in one specimen) frequently connected by transverse spurs ; finer structure of small cells equiaxial or transversely elongate at right angles to the ribs. Testa in section 0-3 mm. thick at the middle of the broad surface, formed of square cells about 0-02 to 0-025 mm. in diameter arranged in radial rows. Length of seed, 3 to 3-5 mm.; breadth, 1-8 to 2:25 mm. ; thickness, 1:8 to 2 mm. REMARKS AND AFFINITIES. In addition to the holotype there are four other seeds (V.42138-41). The species is now also known from the Dorset Pipe-clay Series at Arne (in press) and the Bournemouth Freshwater Beds at Sandbanks (awaiting publication). The form, character of hilum, raphe and chalaza and testa structure all relate these specimens to Rutaceae. Although the form and ornamentation in general and the strap-shaped hilar scar indicate Phellodendron as the nearest living 126 FLORA OF THE LOWER HEADON BEDS genus, the fossil seeds are clearly distinguished from any living or other fossil species by the greater degree of inflation, somewhat smaller size and more marked longi- tudinal ribs of the surface. On account of this last feature the name P. costatum was given. Genus RUTASPERMUM Chandler, 1957 : 102 Rutaspermum ornatum (Chandler) (Pl. 26, fig. 68 ; pl. 27, figs. 69-71) 1925. Zanthoxylon ornatum Chandler, p. 27, pl. 4, fig. 4a, b; text-fig. ro. DiaGnosis. Seed broadly semilunate in outline with straight ventral margin slightly beaked at the micropyle, hilar depression up to three-quarters of the length of the seed. External surface ornamented with sinuous longitudinal ribs which occasionally branch and anastomoze, sometimes connected by short transverse ribs. Testa conspicuously pitted, smooth around the hilar depression. Length of seeds, 4 to 7:25 mm.; breadth, 3:5 to 4 mm.; thickness, 2-5 to 3 mm. NEoTYPE. Brit. Mus. (N.H.), No. V.42142. (Holotype decayed.) DESCRIPTION. Seed: Semi-anatropous, broadly semilunate in outline, inflated. Ventral margin more or less straight, slightly beaked above, occupied in the upper part by the elongate narrowly triangular hilar depression which extends from the apex for nearly three-quarters of the length in some specimens, two-thirds of the length in others and usually well below the middle. Chalaza marked internally by a large orifice situated at the extremity of the longest axis remote from the apical beak. External surface ornamented with sinuous longitudinal ribs which occasion- ally branch and anastomoze or are sometimes connected by short transverse ribs. Surface formed of equiaxial or elongate pits, 0-025 to 0-03 mm. in diameter. Testa about 0°3 to 0-4 mm. thick on the dorsal side, thicker on the ventral side. Com- pletely surrounding the hilar depression is a smooth somewhat flattened area. A few seeds show traces of a thin black shining pitted epidermis but this is almost invariably destroyed. Length of seeds, 4 to 7:25 mm. (frequently about 4°5 mm.) ; breadth, 3:5 to 4 mm.; thickness, 2:5 to 3 mm. Typical measurements are: 1) Length, 5:5 mm.; breadth, 4 mm.; thickness, 2°5 mm. 2) Length, 4-5 mm. ; breadth, 3-5 mm.; thickness, 2-5 mm. 3) Length, 5-5 mm.; breadth, 4 mm. ; thickness, 3 mm. 4) Length, 6 mm.; breadth, 4-5 mm.; thickness, 3 mm. 5) Length, 5 mm.; breadth, 3:5 mm.; thickness, 3 mm. REMARKS AND AFFINITIES. Numerous seeds (cf. V.42146) almost invariably preserved in a soft pyrites mud which is peculiarly liable to rapid decay on exposure to air. At first these seeds were referred to the genus Zanthoxylum but it is now clear that no living representative of Zanthoxylum has such clearly defined regular superficial ornamentation, hence they are now relegated to the form-genus Rutaspermum Chandler, no comparable living genus having been found. The species is distinguished clearly from all others from the south coast deposits by the combined characters of size and smooth unornamented area immediately surrounding the hilar scar. FLORA OF THE LOWER HEADON BEDS 127 Family BURSERACEAE Genus PALAEOBURSERA Chandler, 1961 : 202 Palaeobursera lakensis Chandler (Pl. 27, figs. 74, 75) Palaeobursera lakensis Chandler (in press), pl. to, figs. 20-26. DESCRIPTION. Pyvene: Pointed-ovate in outline, subtriangular in transverse section, dorsal surface convex, ovate; ventrilateral surfaces meeting to form a median ventral longitudinal angle especially clear in the upper two-thirds. Angle pierced about 2 mm. from the apex (i.e. below the middle) by a curved transverse aperture, concave towards the apex of the fruit, for the ingress of the funicle to the placenta. Germination valve occupying the upper two-thirds of the dorsal surface. Wall formed superficially of equiaxial pits about 0-03 to 0-05 mm. in diameter over most of the surface but becoming narrower and elongate along the ventral angle where they may be 0-016 mm. broad and 0:03 to 0:05 mm. long. Thickness of wall on dorsal side o-I mm. Small cells, about 0-or mm. in diameter can be seen forming the walls of the large external pits. Length of endocarp, 3-5 mm. ; breadth, 2°75 mm.; thickness, I mm. REMARKS AND AFFINITIES. A perfect specimen (V.42147) and three other fragments (V.42148), one the rounded end of a larger pyrene. A comparison with pyrenes from the Dorset Pipe-clay Series of Lake (in press) and the Bournemouth Freshwater Beds at Sandbanks (awaiting publication) indicates that the Hordle specimens belong to the same species. It must be noted that the best preserved Lower Headon pyrene is a broad example, uncrushed (as is often the case in the Hordle fossils) and therefore more inflated. Its surface is well preserved unlike the older specimens which are both abraded and sand pitted, but identical pits have been seen in a pyrene from Sandbanks and in other fragments. The Hordle pyrene is therefore referred to P. lakensis Chandler. Family ANACARDIACEAE Section SPONDIEAE Genus? (Pl. 27, figs. 72, 73) DEscRIPTION. Part of a thin-walled fruit with obscure cell structure and rough external surface. Two compartments are occupied by bolster-shaped closely adpressed bodies which may be resin sacs as in Dracontocarya glandulosa from the Dorset Pipe-clay Series (Chandler (in press), pl. 12, figs. 10-21). These bodies appear to have been partially released by maceration so that they resemble lines of parietal seeds. The exposed surface is ornamented by a network of fine ridges forming irregular polygonal areas about 0-025 to 0:05 mm, in diameter, The fruit 128 FLORA OF THE LOWER HEADON BEDS fragment is 2°5 by 3 by 1-5 mm. The sacs vary, those measured were respectively : 0-4 by 0:25 mm.; I:2 by 0-6 mm. at one end and 0-3 mm. at the other end ; 0-6 by o-'2mm.; I by 0:3 mm. REMARKS. No true locules or seeds have been seen nor is the form of the perfect fruit known. While the presence of the resin sacs suggests relationship with Spondieae, the fragment cannot be identified with Dvacontocarya glandulosa to which it bears a general resemblance as the endocarp in that species is much more fibrous and its sacs are larger (about 0-912 to I-g mm. long). Family ICACINACEAE Genus NATSIATUM Buch.-Ham. Natsiatum eocenicum Chandler 1925. Natsiatum eocenicum Chandler, p. 29, pl. 4, fig. 7a-d; text-fig. 11. Natsiatum eocenicum Chandler: Chandler (in press), pl. 13, figs. 6-16. The endocarps have been redescribed (see above) in an account of the species from the Dorset Pipe-clay Series and some fresh Hordle material illustrated for comparison. V.40614 (the neotype), V.42151-52 (fragments) are from Bed 10 east of Beckton Bunny. V.40615-16 from Bed 10 below Hordle House. The original holotype, V.42149, (Chandler, 1925, pl. 4, fig. 7b-d) is now much broken and decayed as is V.42150 (Chandler, 1925, pl. 4, fig. 72). Both were from the contemporary stream channel west of Long Mead End. ? Iodes sp. (or ? Natsiatum sp.) (Pl. 27, figs. 76, 77) 1925. odes sp., Chandler, p. 29, pl. 4, fig. 10; text-fig. 12. DescriPTIon. Endocarp : Incomplete, only the funicular margin being preserved. Longitudinal section narrow-ovate, the lateral funicular canal gradually broadening at the apex; placenta subapical on the funicular margin marked externally by a pair of short horn-like processes which flank it. Walls thick and woody. External surface ornamented by a few conspicuous ridges which branch and anastomoze to form a very coarse network. Locule lining thin, semitranslucent, formed of inflated equiaxial cells. Length of fragment, 9°75 mm.; incomplete breadth, 4 mm.; thickness, 5-75 mm. REMARKS. V.20063, Bed 10 below Hordle House. Relationship to Icacinaceae indicated by the character of the funicular margin. The form of the longitudinal section recalls Natsiatum herpeticum Ham. and species of Iodes. The convex cells of the locule lining recall Natstatwm rather than Iodes but the relationship must remain uncertain without better material, FLORA OF THE LOWER HEADON BEDS 129 Genus IODES Blume Iodes ? hordwellensis n. sp. (Pl. 27, figs. 78, 79; Text-fig. 4) Diacnosis. Endocarp broader than long, much inflated. Ridges of surface ornamentation somewhat rounded, not acute like those of Natsiatum eocenicum. Length, 6-5 mm.; breadth, 8-5 mm.; thickness about 5 mm. Hototype. Brit. Mus. (N.H.), No. V.42153. ' Description. Endocarp: Similar superficially to that of Natsiatwm eocenicum but broader than long and much more inflated. Bisymmetric about a plane passing through funicle, attachment and style. Funicular margin having a rounded rib which broadens conspicuously over the subapical placenta. Opposite margin with a f p st a Fic. 4. odes ? hordwellensis n. sp. Diagram to show endocarp in longitudinal section. (st) style, (p) placenta, (f) funicle, (a2) attachment. 4 approx. sharper ridge. Style apical, marked by a small knob. External surface with a network of rounded ridges less prominent and acute than in N. eocenicum, surface formed of fine equiaxial cells, finer than those of N. eocenicum. A small fragment (V.42154) of a second specimen? shows a similar curvature. On its inner surface there are obscure digitate cells covered by a thin integument bearing small papillae. Length of endocarp (possibly slightly reduced by distortion), 6-5 mm.; breadth, 8-5 mm. ; thickness (possibly slightly increased by distortion), 5 mm. REMARKS AND AFFINITIES. One endocarp and a fragment. The species is clearly distinguished by its broad transverse form and less sharp ridges from Natsiatum eocenicum. The appearance and papillae of the locule lining of the second specimen suggest relationship with Jodes, but the transversely broad form has not been seen in this genus, Hence the determination is regarded as provisional. 130 FLORA OF THE LOWER HEADON BEDS Genus ICACINICARYA Reid & Chandler, 1933 : 344 Icacinicarya transversalis n. sp. (Pl. 27, figs. 80-82) Diaenosis. Endocarp transversely-oval in outline. Style asymmetric on the broad apical margin, attachment near middle of broad basal margin. External surface with network of bold sharp ridges. Locule tubercled. Length of uncrushed valve, 10:25 mm.; breadth, 14 mm. Hototyre. Brit. Mus. (N.H.), No. V.23420. DESCRIPTION. Endocarp: Woody, transversely-oval in outline, somewhat compressed, bisymmetric about a plane which includes the two major axes and the funicle and style, dehiscing in this plane into symmetrical halves. Attachment near the middle of the broad basal margin, style asymmetric on the broad apical margin, so that a straight line drawn between these two organs divides the endocarp un- equally. Funicle marginal within the thickness of the wall in the larger of the two unequal divisions described, placenta apical adjacent to the stylar canal, marked externally by a slight thickening of the margin ; seed pendulous, solitary. Endo- carp from 0:3 to 0°35 mm. thick, hard and compact, formed of cells, 0-016 mm. in diameter arranged in a somewhat columnar manner as seen in sections of the wall, the ‘‘columns’’ in parts being somewhat oblique. External surface boldly ornamented by a network of sharp ridges, some of these given off from the margin with the funicle are longer, straighter and more conspicuous than the others. Locule tubercled, the tubercles formed by a coat of convex cells. Length of speci- men, 10:25 mm.; breadth, 14 mm. Length of a distorted endocarp, probably referable to the same species, imperfect ; breadth, Io mm. Seed: Represented by shrivelled and half-decayed testa, formed of equiaxial cells, somewhat irregular in shape and alignment, about 0-025 mm. in diameter. REMARKS AND AFFINITIES. A valve (V.23420) from Bed 10 below Hordle House having a small fragment missing on the funicular margin. Also a much distorted endocarp (V.23419). It was imperfect at the apex, with better preserved orna- mentation than that of V.23420 but probably belonged to the same species. This imperfect specimen was dissected after it had been photographed ; it then showed the remains of the seed and the tubercles of the locule lining, Form and structure indicate relationship with Icacinaceae. The broad oval form also seen in the smaller endocarps of Jodes? hordwellensis has not so far been found among living representatives of the family but occurs in fossil forms from the London Clay e.g. Faboidea crassicutis (Reid & Chandler, 1933 : 341, pl. 16, figs. 3-10) and Icacinicarya reticulata (Chandler, 1961 : 225, pl. 22, figs. 25-34). Icacinicarya becktonensis n. sp. (Pl. 27, figs. 83, 84) Dracnosis. Endocarp bisymmetric, subobovoid, funicular margin only slightly convex, opposite margin almost semicircular in plane of symmetry. Funicular FLORA OF THE LOWER HEADON BEDS 131 canal broad. Surface coarsely but very obscurely reticulate. Length of endocarp (slightly imperfect at style), 4 mm. ; breadth in plane of symmetry, 3 mm. ; thick- hess at right angles to plane of symmetry, 2°5 mm. Hototyre. Brit. Mus. (N.H.), No. V.42155. DescripTIon. Endocarp: subobovoid, bisymmetric, lateral funicle in plane of symmetry broad, incurving abruptly at the apex to the placenta. Margin carrying funicle only slightly convex, opposite margin almost semicircular in plane of sym- metry. No evidence is preserved of short projecting horns or canals flanking the style at the apex, probably because the apex and outer walls of the funicular canal are abraded. Surface showing a few coarse rather obscure reticulations. Small coarsely digitate cells can be seen in patches. Wall (as preserved) 0-228 mm. thick. Seed : Represented only by a few patches of testa or by impressions on a film of pyrites, formed of equiaxial angular cells about 0-013 mm. in diameter. A large circular scar with rimmed edge originally lay within the broken apex of the endo- carp and clearly represented the hilum of the seed. Dimensions of seed somewhat smaller than those of the endocarp, too incomplete to measure. RemArKS. The specimen was originally perfect apart from its abraded condition which exposed the funicular canal. It subsequently collapsed showing the wall in section almost entirely replaced by amorphous pyrites except for a film of carbon- aceous substance lining it outside and in. The species resembles Icacimicarya pygmaea (Chandler, 1960 : 226, pl. 34, figs. 118-121) in its somewhat gibbous outline with slightly convex funicular margin and highly convex opposite margin, but is more obovoid in shape than that species and considerably larger. There is no evidence of short apical canals flanking the style as in J. pygmaea. It is also distin- guished from the subglobular endocarps of J. bavtonensis (Chandler, 1960 : 227, pl. 34, fig. 122) by its larger size and subobovoid form. Family RHAMNACEAE Genus RHAMNUS Linnaeus Sub-genus FRANGULA Tourn. ex Hall. Frangula hordwellensis n. sp. (Pl. 27, figs. 85, 86; pl. 28, fig. 87) 1925. Rhamnaceae ? Genus ? sp. 2, Chandler, p. 31. 1926. Rhamnaceae Genus ? sp. 2, Chandler, pl. 5, fig. 2. DiaGnosis. Pyrene and seed together 3°5 to 5:25 mm. long; 3°6 to 3:75 mm. broad. Oboval or suboboval in outline with emarginate distal end. HoLotyrPe. Brit. Mus. (N.H), No. V.42156. DESCRIPTION. Fruit: A globose berry 6-25 mm. long, 4 mm. broad in the one specimen found with obscure structure having two pyrenes (or cocci) preserved. Each pyrene formed of a single seed closely embraced by the endocarp except at 132 FLORA OF THE LOWER HEADON BEDS the extreme proximal end where it projects beyond the carpellary coat ; outline of pyrene oboval or suboboval but slightly asymmetric because one margin is a little longer than the other, conspicuously emarginate at the distal end, gently convex with a narrow median longitudinal concavity on one surface and an obtuse but clearly defined median longitudinal angle on the other; formed superficially of concave equiaxial cells, 0-016 to 0:025 mm. in diameter, beneath which are several layers of fibres that diverge obliquely upward from the median angle on the one face, and sweep transversely across the gently convex face. Dimensions of two pyrenes from different fruits: 1) Length, 3-5 mm.; breadth, 3-75 mm. 2) Length, 5:25 mm.; breadth, 3-6 mm. Seed: Anatropous with marginal raphe; agreeing in shape with the closely investing endocarp but projecting at its proximal end as described. Hilar end with a deep elongate-oval hollow surrounded by a rim formed by an upstanding arillus or projection of the testa. Within the hollow is a median ridge parallel with the broad face of the seed. The raphe enters the testa at one end of the hollow and can be traced along the shorter lateral margin of the seed ; the micropyle is at the opposite end of the hollow. Testa black and shining, 0-2 mm. thick around the basal hollow, only about o-1 mm. thick at the middle of the lateral margin, formed superficially of digitate cells, about 0-025 mm. in diameter, markedly columnar in section, the columns about 0-025 mm. broad, having uneven sinuous walls owing to the variation in diameter of the short cells, arranged end to end, of which they are formed. Seed cavity with a shallow median longitudinal furrowed angle (underlying the external furrow of the pyrene) from which the cells of the internal surface appear to diverge. Tegmen thin, semitranslucent, readily detachable, formed of elongate cells which produce parallel striae about 0-008 mm. apart near the lateral margins, elsewhere of equiaxial cells, 0-012 to 0-016 mm. in diameter. Chalaza a conspicuous thickened scar with branching fibres at the end of the tegmen which immediately underlies the emargination of the seed. Dimensions of seeds approximately those of the pyrenes. REMARKS AND AFFINITIES. One fruit in a much decayed condition with two seeds preserved was originally described but is now completely decayed. One larger perfect detached seed and part of another (V.42157) have since been found which have added considerably to the knowledge of the structure recorded in 1925. The endocarp is preserved only in patches on these seeds, the distintegrating processes of fossilization having separated testa and endocarp in a way difficult to effect in living material without maceration. It is now clear that the relationship is with the sub-genus Frangula of the genus Rhamnus. Comparison has been made with R. frangula Linn. and R. purshiana DC. The more slender form and emarginate apex distinguish the fossil from either of these living species, but the general resemblance especially to R. purshiana is very close. R. purshiana is a native of the Rocky Mountains. The sub-genus Frangula is largely American occurring in both the east and west of North America and in South America also. Species are found in Europe, Central Asia, North Africa, the Canaries and the Azores, the mountains of South Europe, the Caucasus and Persia and in Japan, FLORA OF THE LOWER HEADON BEDS 133 Family SABIACEAE Genus MELIOSMA Blume Meliosma sp. (Pl. 28, figs. 88, 89) DEscrRIPTION. Endocarp: One loculed, one seeded, ovoid, bisymmetric in a plane through the attachment and major axis, each broad face with a median broad rounded ridge which extends from the apex to about the middle. Attachment mark- ed by a conspicuous circular aperture. Character of funicular canal not seen. Surface of endocarp showing equiaxial pits about 0-016 mm. in diameter ; thickness of wall about 0-15 mm. of similar cells no definite alignment detected. Lining of endocarp of elongate cells with finely digitate walls their length rather obscure, breadth about o‘or mm. Testa thin, semitranslucent, formed of inflated longitudin- ally elongate cells about o-or mm. broad, length obscure. Maximum diameter in plane of symmetry, 2°75 mm. ; diameter at right angles to this in plane of symmetry, 2mm.; diameter at right angles to plane of symmetry, 1:25 mm. REMARKS. Two endocarps now shattered but represented by remains with cell structure. They closely resemble material awaiting publication from the Bourne- mouth Beds and are similar in their small size. The highly characteristic digitate cells lining the locule confirm the ascription based on form and structure to the genus Meliosma. Family VITACEAE Genus PARTHENOCISSUS Planchon Parthenocissus hordwellensis n. sp. (Pl. 28, figs. go-95) 1926. Parthenocissus sp., Chandler, p. 33, pl. 6, fig. 1a—c. Diacnosis. Apex of seed flattened or scarcely furrowed, raphe somewhat sunk above the chalaza ; chalaza oval, median, sunk. Ventral infolds more than three- quarters of the length of the seed, diverging upwards. Length of ripe seeds, 3:5 to 5 mm.; breadth, 2 to 3-5 mm.; thickness about 2 mm. HoLotyre. Brit. Mus. (N.H.), No. V.20066. DEscRIPTION. Seed: Obovate in outline, smooth, narrowed gradually to the sti- pitate base, apex slightly flattened or scarcely furrowed. Dorsal surface convex with a median oval slightly sunk chalaza passing abruptly into the narrow somewhat sunk raphe. Ventral angle about 90° when uncrushed, deep lateral depressions extend almost from the base for more than three-quarters of the length and diverge upwards. Surface pitted, pits separated from adjacent ones by their own diameter. Length of seed, 3:5 to 5 mm. ; breadth, 2 to 3:5 mm. ; thickness (uncrushed), 1°5 to 2mm. A typical seed measured 4 by 2 by 2 mm. Smaller seeds about 2°75 by 1-6 by 1-5 mm, appear to be immature, 134 FLORA OF THE LOWER HEADON BEDS REMARKS AND AFFINITIES. The seeds being brittle are often broken. Several specimens (V.42161-64) including an immature seed have been found and a few fragments (V.20067). The species is closely comparable with the living P. henryi Hems. A single seed of P. hordwellensis (V.42237) was found in the Barton Beds, Horizon A3, after the paper dealing with Barton plants was in proof. It is an internal cast in pyrites, 4-5 mm. long, 2°5 mm. broad. The ventral surface shows the infolds clearly but the chalaza on the dorsal side is obscured by a coarse crystalline efflorescence. The seed is shown for comparison with the Hordle material in Pl. 28, figs. 94, 95. Genus AMPELOPSIS (L. C. Rich. in) Michx. Ampelopsis rotundata Chandler 1926. Ampelopsis votundata Chandler, p. 33, pl. 5, fig. 5a—c. Diacnosis. Seed roundly obovate in outline, much inflated, apex rounded, base acutely pointed. Dorsal face rounded with raised convex oval chalaza and prominent raphe. Ventral face convex with short divergent deep wide infolds. Length of largest seed, 3-5 mm.; breadth, 3 mm. Length of smallest, 2-5 mm. ; breadth, 2 mm. Ho.totypre. Brit. Mus. (N.H.), No. V.20068. REMARKS. Other material (V.20068a) a small immature seed and three in- complete specimens. All the seeds are from Bed 10 near Long Mead End. Genus TETRASTIGMA Planchon Tetrastigma lobata Chandler (Pl. 28, figs. 96, 97) 1925. Tetvastigma lobata Chandler, p. 32. 1926. Tetvastigma lobata Chandler, pl. 5, fig. 3a-c. ? Tetrastigma lobata Chandler ; Chandler (in press), pl. 15, figs. 35-38. DiaGnosis. Seed deeply grooved at apex, base tapering or broad and slightly stipitate. Median chalaza oval sunk, raphe narrow sunk. Both surfaces lobed, lobes of ventral surface on both sides and on raphe ridge. Length of seed, 5:5 to gmm.; breadth, 4 to Io mm. in well developed seeds. HoLotypPe. Completely decayed as is all other material. See Chandler, 1926, pl. 5, fig. 30, c. DESCRIPTION. Seed: Obovate, roundly oblong or subcircular in outline, much compressed dorsiventrally, apex deeply grooved, base tapering or broad and slightly stipitate ; dorsal face with small sunk median oval chalaza and narrow sunk raphe ; surface deeply lobed, lobes radiating from the chalaza. Ventral face with longi- tudinal infolds from which rounded lobes diverge both over the sides and the raphe ridge. Surface pitted, testa columnar in section, Length of seed, 5:5 to 6°5 or FLORA OF THE LOWER HEADON BEDS 135 even 9 mm.; breadth, 4 to 10 mm. An exceptionally small seed was 3 mm. long, 3:25 mm. broad. A typical seed was 6:5 mm. long, 5 mm. broad, 3 mm. thick, but the measurements were made after drying so that the seed had contracted to perhaps two-thirds of its original size. REMARKS. Three seeds of which two were figured in 1926 and one in the present volume. The mode of preservation was such that the species was peculiarly liable to decay. The relationship to living Tetrastigma was discussed by Chandler (1925 : 32) and a similar seed referred to ? T. lobata was described from the Dorset Pipe-clay Series of Lake (in press). Kirchheimer describes a comparable species of Tetra- stigma under the name T. chandleri from the Middle to Upper Oligocene of Saxony (Kirchheimer, 1957 : 324, 602, fig. 197a-d) which he states is relatively narrower in proportion to the length. Family DILLENIACEAE Genus ACTINIDIA Lindl. Actinidia sp. (Pl. 28, fig. 98) DESCRIPTION. Seed: A fragment showing the rounded apex but unfortunately neither the raphe, micropyle or hilum are preserved. The testa has coarse hexagonal external pits, 0°05 to 0-I mm. in diameter, delimited by rounded walls with median suture line along which it tends to break. Internally there are corresponding con- vexities characteristic of Actinidia. The fragment is 1-5 mm. long and its maximum breadth is r mm. REMARKS. The seed in spite of its imperfect condition appears to be identical with better preserved material from Bournemouth awaiting publication in which all the distinctive characteristics of Actinidia are shown. Family THEACEAE Section THEAE Genus GORDONIA Ellis Gordonia minima Chandler (Pl. 28, figs. g9q-102) 1926. Gordonia minima Chandler, p. 34, pl. 6, fig. 3a, b; text-fig. 16. DiacGnosis. Capsule pointed-ovoid, scar of attachment small so as scarcely to truncate the base. Sutures of five valves forming thin longitudinal ridges. Length of capsule, 8-5 to 10 mm.; breadth, 4 to 5-75 mm. HototypPe. Brit. Mus. (N.H.), No. V.20070. DEscriPTION. Frut: A five-loculed woody capsule with five loculicidal thick- walled valves which remain attached to one another and to the receptacle at the base ; capsule pointed-ovoid the apex much attenuated, scar of attachment small, 136 FLORA OF THE LOWER HEADON BEDS remains of receptacle persistent, prominent, circular or pentagonal. The sutures of the valves form thin longitudinal external ridges. Valves triangular in cross section, before dehiscence united throughout their length by thin septa to a five- angled columella. On dehiscence the septa break away at the base along a curved line leaving the valves separated from the columella. Length of fruit, 8:5 to ro mm.; breadth, 4 to 5°75 mm. REMARKS AND AFFINITIES. Numerous specimens in addition to those figured, almost always incomplete above, and some detached valves (cf. V.20071, V.20071c, V.20072, V.42166) only two virtually perfect fruits having been seen. The characters are those of Gordonia but no such small living species is known, the nearest in size being G. lasianthus Linn. (length, 13-5 mm.) from swamps of Mexico and Florida. The genus also occurs in China, the Himalayas, Ceylon, Further India and the East Indies. Gordonia truncata Chandler 1926. Gordonia truncata Chandler, p. 35, pl. 6, fig. 4a, b; text-fig. 17. Diacnosis. Fruit ovoid not as conspicuously attenuated to the apex as is G. minima. Base truncate by the large scar of attachment which is much broader than in G. minima. Sutures of valves smooth externally not ridged. Length of fruit, 9g mm. ; diameter (compressed), 3°5 by 5:5 mm. Hototyre. Brit. Mus. (N.H.), No. V.20073. REMARKS. Seven additional fruit bases (V.42167) have now been recognized by the great breadth of the scar of attachment. For a discussion of the relationship of the species see Chandler (1926 : 35). Section TAONABEAE Genus EURYA Thunb. Eurya becktonensis n. sp. (Pl. 28, fig. 103) Diacnosis. Seed transversely oval in outline with reniform cavity (inferred from form and pitting) the micropylar limb apparently broader at its end than the other. Hilum marginal somewhat nearer to the micropylar than to the chalazal limb. Testa thickened laterally near the hilum by a conspicuous patch of coarsely pitted tissue. Dimensions of seed, 2 by 1:25 mm. Hototyre. Brit. Mus. (N.H.), No. V.42168. DESCRIPTION. Seed: Bisymmetric, transversely oval in outline (dorsal margin slightly imperfect), inflated. Cavity (as inferred from form and pitting) reniform the micropylar limb being slightly broader at its extremity than the other. Hilum an elongate-oval depression on the broad margin between the limbs. Micropyle near one end of the same margin terminating the broad limb and indicated super- ficially by a point around which the surface pits are small and radially aligned. FLORA OF THE LOWER HEADON BEDS 137 Around the hilum (seen on one well preserved surface of the seed) is a thick, coarsely pitted, triangular patch of tissue the inner point of the triangle lying nearer to the chalaza than to the micropylar limb. Pits on the thickened area mostly equiaxial and from 0-057 to 0-I mm. in diameter. Whole surface of testa deeply and coarsely pitted, the rows of pits lying parallel with the dorsal margin of the seed and with the curvature of the seed cavity. Over the middle part of the curve there are six or seven very regular rows of pits the largest about o-1 by 0-076 mm. in diameter. Many of the pits are elongate radially. Over the chalazal limb the pits become long and narrow and on its ventral side their walls are scarcely prominent perhaps because only in this part of the testa is the outer wall of the pit preserved obscuring the cavity usually exposed elsewhere by the collapse and abrasion of this wall. Maximum diameter of seed in plane of symmetry, 2 mm. ; diameter at right angles to it, 1:25 mm. ; thickness at right angles to plane of symmetry, 0°75 mm. REMARKS. One seed only of which the details of structure are clearly preserved on one surface but obscured by the poorer preservation on the other. The size and form of the pitted seed with its reniform cavity place it in the Taona- beae a section of the Theaceae comprising trees and shrubs. A similar reniform cavity has been seen in Eurya and Adinandra but the fossil more resembles the thinner-walled Ewrya than the latter genus. In Eurya, however, there appears to be considerable variation in the form of the seed cavity in different species which suggests that there may be some confusion among the living herbarium material of which unfortunately only a very limited amount could be dissected. Thus EF. japonica R. D. Oldham No. 92 Nagasaki with seeds about 1-2 by 1°5 and I by 1°5 mm. in diameter has a reniform seed cavity closely comparable with that of the fossil whereas E. japonica var. nitida Dyer from West Borneo, Mondi 267, has a hooked cavity with one long and one short limb and is unlike the fossil. E. japonica ranges from Japan, China and the East Indies to the Indian Archipelago while allied species (section Pyoteuvya) are found in the Sandwich Islands, Malaya and Ceylon. The seed has been named Eurya becktonensis. Genus CAMPYLOSPERMUM Chandler, 1925 : 16 Campylospermum hordwellense Chandler (Pl. 28, figs. 104-107) 1925. Campylospermum hordwellense Chandler, p. 16, pl. 1, fig. 6a—c; text-fig. 4. Diacnosis. Fruit with five to eight (or more?) radially arranged seeds in a single row pendulous from an axile apical placenta. Length of fruit about 4 to 5 mm. Seeds bisymmetric, flattened, oblong or oboval in outline, splitting in the plane of symmetry with U-shaped cavity having very unequal limbs. Condyle between limbs occupied by long raphe canal from about one-quarter to one-half the length of the seed. Testa formed externally of coarse inflated cells, internally compact of several layers of small rectangular cells. Seed about 4 mm. long, 1-75 to 2 mm. broad. Hototyre. Brit. Mus. (N.H.), No. V.20034 (much decayed), Neotype V.20035. 138 FLORA OF THE LOWER HEADON BEDS DEscCRIPTION. Fruit: Globular or subglobular, apex incomplete, base with short stalk (erroneously regarded as style in 1925), lateral walls rarely preserved but showing rounded secreting cells, inner thin layers more frequently seen with remains of longitudinal fibres branching upwards at acute angles. Plurilocular, commonly three-loculed with a single row of five to eight (or more?) pendulous seeds arranged radially around a fibrous axis. Placentation axile, apical. Length of fruit at least 4 mm. ; breadth, 3-5 to 5 mm. (without lateral walls). Seed; Bisymmetric approximately, oblong or oval in outline, laterally flattened by mutual pressure of adjacent seeds in berry, splitting for germination in plane of symmetry, the equal valves showing the U-shaped cavity with U opening upwards with very unequal limbs. The longer micropylar limb lies nearest to the axis of the fruit, the shorter limb carries the subterminal chalaza on its inner angle. Limbs separated by a narrow condyle, hilum marginal at the end of the condyle indicated externally by a small aperture surrounded by radial ridges (cf. Chandler, 1925, pl. 1, fig. 6a—c and text-fig. 4). The hilar aperture is situated at about one-third of the length of the seed from its upper margin on the side sloping away from the fruit axis. Raphe canal wide lying within the condyle and extending for about one-quarter to one-half of the length of the seed. Micropyle terminal on the longer limb. Testa about 0-25 to 0-4 mm. thick its outer part (about 0-03 to o-I mm. thick) is a layer of large cells much inflated superficially when unabraded or not collapsed. These cells vary in diameter from 0:03 mm. to at least 0-15 mm. ; they diverge from the chalazal region over the broad surfaces of the seed (Pl. 28, fig. 106). On the broad surfaces overlying the chalaza is a conspicuous thickening of much inflated large equiaxial cells like those in a similar position in Eurya becktonensis (p. 136). Inner part of testa compact formed of several layers of small rectangular cells about 0-016 to 0-025 mm. in diameter as in Cleyera ? stigmosa (p. 139). Seed cavity finely striate transversely. Length of seed about 4 mm.; breadth about 1°75 to 2mm. REMARKS AND AFFINITIES. The species is very abundant but perfect fruits are scarce owing to the decay and collapse of the lateral walls which allow the seeds to fall free. Examples of four good fruits are in V.42171. Isolated seeds are common (V.20035a, V.42172). These specimens were originally referred to the Araceae (Chandler, 1925) but at a later stage the testa structure appeared to con- nect them with Cleyera ? stigmosa then thought to belong to Myrtaceae on account of its U-shaped seed cavity (Chandler, 1961: 80, 108, 335, pl. 8, figs. 8, 9; pl. 11, fig. 18; pl. 34, figs. 16-18). Recent work has however necessitated transferring both fossils to the section Taonabeae of the family Theaceae (Chandler, 1960 : 211) where alone the U-shaped cavity is combined with the distinctive testa structure described. No living genus exactly comparable with Campylospermum has been seen but there is a close general resemblance to species of Tervnstroemia and Anneslea with a single ring of seeds. So far as the structure in these genera could be observed the testa appears to be smooth and without the large inflated cells described above. It must be borne in mind, however, that very little living material is available for dissection. Hence pending further information the name Campylospermum is retained. FLORA OF THE LOWER HEADON BEDS 139 Genus CLEYERA DC. Cleyera ? stigmosa (Ludwig) (Pl. 28, figs. 108-112) 1960. Cleyeva? variabilis (Chandler) Chandler, pp. 212, 213, 229, pl. 31, figs. 48-56; pl. 34, fig. 145. (see also for earlier references). 1960. Eurya stigmosa (Ludwig) Mai, p. 79, pl. 4, figs. 8-17. DESCRIPTION. Several unmistakable seeds, much decayed or abraded and heavily pyritized show all the specific characters of this widely distributed Eocene species. Two seeds show remains of an outer thin striate coat which somewhat conceals the surface pits normally exposed ; it is formed of radially elongate cells, 0-013 to 0-016 mm. broad, which are often elongate-hexagonal in surface view. The descriptions of the species already published are applicable in every detail. There is the same U-shaped cavity and internal condyle, the walls show the same structure superficially, in section they are of the same thickness and are made up of the same small quadrangular cells. The tegmen is transversely striate, formed of equiaxial cells about 0-012 mm. in diameter. Diameter of three seeds respectively : I-I by 1:2 mm.; 1°05 by 1-I mm.; 1°5 by 1:3 mm. RemMARKS. Fourteen seeds (V.42176) and three figured ones. Also known from the Hengistbury and Barton Beds (Chandler, 1960), the Woolwich and Oldhaven Beds, the London Clay (Chandler, 1961) and the Bournemouth Beds, Highcliff Sands and freshwater Bracklesham Beds of Alum Bay (awaiting publication). Mai (1960) described this species from German localities identifying it with Eurya stigmosa (Ludwig) after comparison with original material. Genus ANNESLEA Wall. Anneslea ? costata n. sp. (Pl. 29, figs. 113-115) DiaGnosis. Seed oboval to narrow-oboval in outline, much compressed having a median slightly curved ridge with rounded inner end over the condyle and a curved rounded ridge over the curved seed cavity beyond which is a marginal rim. Testa fine-celled. Length of seed, 2-2 mm.; breadth, 1-3 to 2 mm. HototyPe. Brit. Mus.(N.H.), No. V.42177. DESCRIPTION. Seed: Somewhat compressed, oboval to narrow-oboval in out- line, curved so that the cavity is U-shaped, the limbs being separated by a condyle which is rounded at the inner end. At the base of the condyle is the marginal hilar aperture leading into the narrow median raphe canal within the condyle. Micro- pyle possibly represented by a small pit near the hilum at the end of one limb of the cavity. Externally a median slightly curved ridge with rounded inner end overlies the condyle and is separated by a narrow U-shaped furrow from a rounded U-shaped ridge which surrounds it and corresponds with the seed cavity. Outside again is a marginal rim (PI. 20, fig. 113). Wall of condyle formed of coarse cells. GEOL. 5,5 15 140 FLORA OF THE LOWER HEADON BEDS The external wall of the testa of fine cells the outlines of which are obscure. Seed cavity lined by cells transverse to its length, about 0-012 to 0-016 mm. broad. Length of seed, 2:2 mm.; breadth, 1-3 to 2 mm.; maximum breadth of condyle, 0-5 mm. REMARKS AND AFFINITIES. Two seeds, one imperfect which was subsequently treated with nitric acid and dissected to show the condyle (V.42178). The character of the U-shaped cavity and raphe canal in condyle indicate the family Theaceae. The seed with its unpitted surface is unlike Eurya, Adinandra or Cleyera. The fine texture of the testa is more suggestive of Amneslea or Ternstvoemia. Bearing in mind that few species of these genera are available for study and that other genera have not been seen at all the species is referred tentatively only to Anneslea. The rounded ridges of the surface appear to be distinctive so the name Anmneslea? costata has been given. The genus Azmneslea is Indomalayan. Ternstroemia belongs to Asia, Africa and South America. Genus HORDWELLIA Chandler, 1960 : 228 Hordwellia crassisperma (Chandler) 1926. Actinidia crassisperyma Chandler, p. 34, pl. 6, fig. 2; text-fig. 15. 1960. Hordwellia crassisperma (Chandler) Chandler, p. 229, pl. 34, figs. 140-144. Hordwellia crassisperma (Chandler) Chandler (in press), pl. 17, figs. 8-61. The species was referred to Theaceae by Chandler (1960) and has been redescribed and fully illustrated in a monograph on the Dorset Pipe-clay Series (now in press) where range of variation and geological distribution are shown. In addition to the holotype and seeds there illustrated, other seeds from Hordle are V.20069a and V.42179. Family THYMELIACEAE Genus ? An obovoid seed slightly imperfect at the pointed micropylar end, originally with crustaceous shining tegmen preserved, now represented by the white sandy cast and pulverized remains of the carbonaceous tegmen (V.42180) found in Bed Io east of Long Mead End. Length of cast about 5 mm.; breadth, 2°75 mm. The tegmen was columnar in section with straight columns 0-008 to 0-or mm. broad. A few broken but typical tegmen fragments of other seeds were also seen (V.42181) from Bed 10 east of Beckton Bunny. The surface was shining, formed of equiaxial cells flat or slightly convex superficially, some showing the tiny central lumen, the appearance varying greatly with the degree of abrasion. In some instances the whole surface appears minutely pitted. Most of the fragments show sub-spiral striation or furrowing caused by contraction and curling on drying. Although the material is too incomplete for more definite determination the characters seen indicate Thymeliaceae. FLORA OF THE LOWER HEADON BEDS 141 Family LyTHRACEAE Genus MICRODIPTERA Chandler, 1957 : 107 Microdiptera parva Chandler (PLZ; tigss IO. 17) 1957. Microdiptera parva Chandler, p. 107, pl. 15, figs. 133-149 ; text-fig. 2. DEscRIPTION. Seed: Subquandrangular in outline, much compressed. Raphe prominent median longitudinal on the ventral surface. Wings relatively thick, not concave on the ventral surface but breaking away from the seed body on both sides of the raphe. Dorsal surface with elliptical operculum or germination valve in the lower half of the seed, 0°5 by 0°35 mm. in diameter, ornamented with quad- rangular cells in longitudinal rows about 0:03 mm. in diameter. Surface somewhat worn on the dorsal side over the wings, the external coat being abraded so that the underlying striate surface is exposed. The striae swerve around the operculum and across the wings. On the less abraded ventral surface there is a coat of variously oriented convex cells. Length of seed along axis, 0-9 mm.; total breadth across body and wings, I mm. REMARKS. One seed (V.42182) has been compared with similar seeds from Bovey Tracey and Cliff End, Mudeford. It can be matched exactly in material from these deposits and has therefore been referred to M. parva. Family CORNACEAE The family is represented at Hordle by six genera, four of which belong to the Mastixioideae (Eomastixia, Mastixicarpum, Mastixia? and Genus?) and two to the Cornoideae (Cornus and its sub-genus Dunstamia). Four are extinct. Section MASTIXIOIDEAE Genus EOMASTIXIA Chandler, 1926 : 37 Eomastixia rugosa (Zenker) 1960 Eomastixia yvugosa (Zenker) : Chandler, p. 234. Eomastixia vugosa (Zenker) : Chandler (in press), pl]. 18, figs. 4-41; pl. 19, figs. 1-7. For references see Chandler Dorset Pipe-clay Flora (in press) where the species is dealt with in detail and its variations and geological range are noted. One of the most characteristic fossils of Bed 10, east of Beckton Bunny. In addition to the Hordle material recorded in the Dorset Pipe-clay monograph it is represented by the following specimens: V.20079a (Chandler, 1926, pl. 6, fig. 6a), V.20079) (Chandler, 1926, pl. 6, fig. 6), V.20079¢ (Chandler, 1926, pl. 6, fig. 6d), V.20079d (Chandler, 1926, pl. 6, fig. 6e), V.20081 and V.42183 are three-loculed endocarps, V.42184 a four-loculed endocarp (rare), V.20080, V.20082, V.31792, V.42185 numerous endocarps. The original of Chandler, 1926, pl. 6, fig. 6c has decayed. GEOL. 5, 5 15§ 142 FLORA OF THE LOWER HEADON BEDS In Germany the species ranges from the Lower to the Middle Oligocene (Kirch- heimer, 1957 : 260) whereas in Southern England, so far as present evidence goes, it is confined to the Eocene above the London Clay. Genus MASTIXICARPUM Chandler, 1925 : 36 Mastixicarpum crassum Chandler (Pl. 30, fig. 161) 1960 Mastixicarpum crassum Chandler: Chandler, p. 215. Mastixicarpum crassum Chandler : Chandler (in press), pl. 19, figs. 14-23 ; pl. 20, figs. 1-4. For references and range of variation of the species as well as its geological distribu- tion see the Dorset Pipe-clay Flora (in press). Like Eomastixia rugosa it is among the commonest of the Hordle plants at Bed Io, east of Beckton Bunny and has also been found in an abraded state in rainwash from the contemporary stream channel west of Long Mead End (V.20077). In addition to the Hordle material to which reference is made in that monograph and above there are the following specimens: V.20075-76, V.20078, V.31791, V.42186. An endocarp (V.42235) alleged to come from the Barton Clay of Barton was found in the Museum after the publication of the Bulletin on the Plant Remains of the Hengistbury and Barton Beds (Chandler, 1960). It is shown here for comparison and in order to complete the records (cf. Pl. 30, fig. 161). It has been fractured transversely and in spite of its poor condition shows the single locule with large longitudinal dorsal infold along the length of which splitting had started. Genus MASTIXIA Blume Mastixia ? glandulosa n. sp. (Pl. 29, figs. 118-121) 1926. Carpolithus sp.2, Chandler, p. 44, pl. 7, fig. 12a-d; text-fig. 28. Diacnosis. Endocarp fusiform with a longitudinal external groove along which it splits, normally one-loculed, locule C-shaped in transverse section owing to an internal longitudinal infold which corresponds with the external groove. Wall riddled with innumerable resin ducts close to its outer surface. Length of endocarp, 47-75 to 13°5 mm. ; breadth about 3°5 to 5 mm. Hototyre. Brit. Mus. (N.H.), No. V.42187. DESCRIPTION. Fruit: Not seen. Endocarp : Fusiform sometimes more acutely pointed at one end than at the other, having a median external groove corresponding with a deep longitudinal infold which projects into the single locule causing it to be C-shaped in transverse section. Splitting tends to occur along the groove. External surface where best preserved showing a network of superficial fibres ; when somewhat abraded as is commonly the case, smooth with transverse striae. Walls thick with three regions as seen in transverse section: a thin outer region which tends to peel away on drying; a FLORA OF THE LOWER HEADON BEDS 143 thick middle region riddled with innumerable elongate, oval, or irregularly angular lacunae now containing yellow solidified resin ; a thin inner region lining the locule. Locule transversely striate. No germination valve could be detected. Length of endocarp, 7°75 to 13°5 mm. ; breadth, 3-5to5mm. Three typical endocarps showed the following measurements respectively : 1) Length, 10:5 mm. ; transverse diameter, 4 by 3°55 mm. 2) Length, 9°5 mm.; transverse diameter, 5 mm. 3) Length, 7-75 mm. ; transverse diameter 3°5 mm. REMARKS AND AFFINITIES. These peculiarly fragile endocarps are liable to shatter as they dry owing to the expansion of the resin in their walls and the contraction on drying of the carbonaceous tissues. Originally described as Carpo- lithus sp. 2 (Chandler, 1926 : 44) and believed to be two-loculed, the two-loculed effect is now known to be due to the close contact of the infold with the ventral wall of the locule caused by compression and partial collapse. At the same time the possibility of two-loculed specimens sometimes occurring cannot be wholly excluded. The single locule with longitudinal infold suggests alliance with Mastixia but the evidence of a germination valve is lacking probably because of the ready collapse of the resin riddled walls which makes a valve very difficult to detect. So far these abundant resin cavities have not been seen in Mastixza hence the ascription to the genus is regarded as doubtful but it must be borne in mind that only a few species of living Mastixia have been examined in section and that in ungerminated and unmacerated Mastixia it is almost impossible to detect the valve although longitudinal splitting down the infold is a fairly common feature. Under the name Retinomastixia schulter (Kirchheimer, 1937: 915, fig. 9) de- scribed an endocarp from the Middle Oligocene Brown Coal of Germany which he separated from Mastixia because of its numerous secreting cavities and the absence of an external longitudinal furrow associated with splitting. The absence of the furrow also distinguishes Retinomastixia from the Hordle fossils although this character can be obscured completely in compressed Hordle material. However Retinomastixia appears to differ also in the position of the resin layer which is nearer to the inner surface of the endocarp than in Mastixia? glandulosa so that there is a thicker woody layer outside it. Retimomastixia is, moreover, described as having a dorsal germination valve. In addition to the material figured from Hordle there are also V.42191 (three endocarps and a few fragments), V.z0100, V.20102, V.42192 (various fragments with resin). V.zo101 originally figured as Carpolithus sp. 2 is now in many fragments. ?Section MASTIXIOIDEAE Genus? (Pl. 29, figs. 122-124) 1926. Carpolithus sp. 5, Chandler, p. 46, pl. 8, fig. 3a—d. The median infold on each dorsal germination valve of this three-loculed fruit (V.20106a) suggests that the specimen may belong to Mastixioideae. Previously no suggestion as to its botanical position could be made. The subglobular shape is 144 FLORA OF THE LOWER HEADON BEDS one character which distinguishes it from any other species referred to Mastixioideae and the wide external grooving of the dorsal infolds on the valves is another. Also there appears to be no tendency to split along these infolds so far as can be seen. Section CORNOIDEAE Genus CORNUS (Tourn.) L. Cornus quadrilocularis Chandler (Pl. 29, figs. 129-133) 1926. Carpolithus sp. 4, Chandler, p. 45, pl. 8, fig. 2a, b. Cornus quadrilocularis Chandler (in press), pl. 20, figs. 23-26. DESCRIPTION (emended). Endocarp : Subglobular with shallow apical depression but without central canal. Rounded or somewhat pointed at the base, four-loculed each locule having a large dorsal germination valve which separates from the apex downwards but remains attached at the base where it breaks irregularly. External surface finely pitted, walls thick, woody, of regular more or less equiaxial slightly sinuous cells. Locule lining of convex cells with finely digitate walls. Seeds solitary, the testa adhering closely to the locule wall, probably pendulous. Length of original endocarp (V.20106), 3:25 mm. ; transverse diameter, 3°5 mm. Length of V.42193, 3 mm. ; transverse diameter, 3-8 mm. Seed: Represented only by collapsed remains of testa showing transversely aligned cells which give rise to transverse striae about 0-008 mm. apart. Remarks. The fossil was originally described as Carpolithus sp. 4. The finding of a second endocarp made it possible to correct certain errors in the original des- cription. Moreover these specimens are identical with endocarps from the Dorset Pipe-clay Series at Lake which show clearly that the seed was pendulous and confirm the observations recorded above based on Hordle material. A tendency to slight lateral compression is seen in V.42193 as in the Lake endocarps. For further remarks on relationship see Chandler (in press). Sub-genus DUNSTANIA Reid & Chandler, 1933 : 459 Dunstania glandulosa (Chandler) (Pl. 29, figs. 125-128) 1926. Symplocoides glandulosa Chandler, p. 41, pl. 7, fig. 5a, 6; text-fig. 26. 1926. Carpolithes sp. 3, Chandler, p. 45, pl. 8, fig. 1a-e. DiaGnosis. Pointed sub-fusiform endocarps about 5 to 9°5 mm. long, two- to four-loculed with small apical depression and ovoid secreting cavities in endocarp wall and to a less degree in septa. Ho.otype. Brit. Mus. (N.H.), No. V.z0090 (now much broken), Neotype V.42194. DeEscriIPTION. Endocarp: Pointed sub-fusiform, or pointed-ovoid, slightly compressed laterally in some instances, with a small truncation at the apex due to FLORA OF THE LOWER HEADON BEDS 145 an obconical or basin-shaped depression which may or may not have crenulate margins. Base usually sharply pointed. Locules two to four some of which may be abortive, they communicate with the apical depression. Each locule with a large dorsal germination valve extending almost throughout its length, with smooth sutures except at the base where they break irregularly from the lower part of the endocarp wall. Wall thick formed of cells with sinuous outlines. In section three regions can be distinguished: a thin compact dark outer region, a thick middle region more spongy in texture with numerous ovoid secreting cavities, a thin com- pact inner layer. Secreting cavities, diameter about o-r to 1 mm. with smooth shining walls lined by cells with sinuous outlines often about 0-05 mm. in diameter. The cavities occur in the septa but to a much less extent than in the outer walls of the endocarp, usually smaller in the septa. Locule lining smooth, transversely striate, with sinuous-walled cells. Dimensions of several endocarps: 1) Length, 9°5 mm. ; breadth, 4:25 mm. 2) Length, 8-75 mm.; breadth, 3-5 mm. 3) Length, 5°5 mm.; breadth, 4 mm. 4) Length, 9 mm.; breadth, 8 mm. 5) Length, 7 mm.; breadth, 2:25 by 3:25 mm. Thickness of dorsal wall where measured, 1°75 to 2°5 mm. Seed : Solitary, represented only by remains of testa adhering to the locule (N.B. the body described as a seed in 1926, p. 45 was undoubtedly a locule cast). Testa structure obscured owing to crumpling. REMARKS. When originally described under the names Symplocoides glandulosa and Carpolithes sp. 3 the structure of these fruits was only partly understood and was therefore wrongly interpreted. The truncate end was then regarded as the base. Carpolithes sp. 3 is a well developed endocarp in which the septa has been accidentally pierced so that the three locules have a false appearance of being in communication with one another. Three additional endocarps without septal break were, however, referred tentatively to the same species. In the holotype the base was broken and abraded so that three locule cavities were exposed simula- ting the apex of an endocarp like Symplocos. The examination of hundreds of pyritized fruits of Dunstania multilocularis from the London Clay illuminated the structure of the smaller carbonaceous Hordle specimens and it became clear that they and Dunstama were closely related. The Hordle endocarps are much smaller than either of the London Clay species, also relatively longer and narrower with more acutely pointed base while so far no more than four locules have been seen. Reid & Chandler referred Dunstania to Cornaceae for reasons given (1933: 461). Kirch- heimer has since confirmed the relationship noting that in the sub-genus Bothrocaryum of Cornus there is a deep depression at the apex of the endocarp and that in the sub- genus Macrocarpum and in Cornus Volkensi (sub-genus Afrocrania) there are secreting cavities. He stresses the fact that the occurrence of more than two locules is not uncommon in living Cornus (Kirchheimer, 1948 : 85; 1957 : 443). The fossils combine the presence of an apical pit as in Bothrocaryum with that of secreting cavities as in Macrocarpum. In the latter sub-genus only, a small apical pit sometimes occurs e.g. in Cornus mas. The fossils may therefore represent a distinct sub-genus of Covnus for which the name Dunstanmia may conveniently be retained at least for the present, The specific name D. glandulosa is dictated by 146 FLORA OF THE LOWER HEADON BEDS its earlier use. In addition to the specimens figured or mentioned above there are the followiug three-loculed endocarps : V.20103-04c (V.20103 being figured Chandler, 1926, pl. 8, fig. ra-e), V.20104d a two-loculed endocarp, V.z0104e a four-loculed endocarp and V.42196 several endocarps, some abraded and some sectioned. Family ERICACEAE Section ANDROMEDEAE Genus? sp. 2 (Pl. 30, fig. 138) 1926. Ericaceae Genus ? sp. 2, Chandler, p. 38, pl. 6, fig. 8a—c. DESCRIPTION. Fyvuwit: A superior loculicidal capsule, subcircular in outline, somewhat depressed from apex to base, not stipitate but seated upon a persistent perianth disc equal to the fruit in diameter. Surface of fruit with shallow longitud- inal concavities over the septa and slight rounded ridges between these concavities overlying the locules. Apex slightly incurved but having a prominent broad circular style base. External surface finely but irregularly pitted and rugose, formed of equiaxial cells. Carpel wall thick. Placentation axile from large basal placentas in the inner angles of the locules. Seeds numerous sunk partially in shallow depressions of the locule wall. Transverse diameter of fruit originally described (now decayed), 1 mm.; height, 0-9 mm. Transverse diameter of a larger fruit, 1-5 mm.; height, 0-7 mm. Thickness of perianth disc about o-I mm. Seed : Rounded-oblong, anatropous, ornamented with a raised network of elong- ate longitudinally aligned meshes crossed by fine transverse striae, the meshes following the curvature of the seed at the rounded distal end. Length of seed, 0°33 mm. ; breadth, 0:16 mm. REMARKS. The seeds were beautifully shown in the first described specimen, preserved in shining pyrites which displayed the reticulate pattern very clearly. V.42198 has not dehisced but shows well the external characters of the fruit. V.42199 originally measuring I by 0-75 mm. broke longitudinally later and showed three seeds in one locule and their mode of attachment. Reference to the section Andromedeae of Ericaceae was suggested in 1926 and no closer suggestion as to affinity can yet be made. The specimens although somewhat resembling Epacridt- carpum mudense are distinguished by the depressed form, persistent perianth and several-seeded locules. Family EPACRIDACEAE Genus EPACRIDICARPUM Chandler, 1960 : 214 Epacridicarpum headonense Chandler (Pl. 29, figs. 134-137) 1926. Andromedeae Genus ? sp. 1, Chandler, p. 37, pl. 6, fig. 7a, b. 1960. Epacridicarpum headonense Chandler, p. 234, pl. 34, figs. 146, 147, DESCRIPTION. Fruit: A sub-hemispherical or oblate-sphaeroidal five-loculed loculicidal capsule. Upper surface somewhat flattened, lower convex, sometimes FLORA OF THE LOWER HEADON BEDS 147 slightly stipitate. Surface ornamented with unequal irregular pits. Walls thick the outer layers about 0-05 to 0-15 mm. thick formed of several layers of rounded equiaxial cells many being about 0-025 mm. in diameter; inner layers somewhat thinner of narrow elongate longitudinally aligned cells closely compacted. Locule lining of elongate convex cells which diverge obliquely from subapical axile placentae. These cells vary considerably in length and breadth, many are 0-009 mm. broad and 0:05 to o'r mm. long. Length of fruit, 0°55 to 1-25 mm.; diameter I to 1°75 mm. (Note in Chandler, 1960, p. 234 the diameter of the species was incorrectly given). Seed : Pendulous, solitary. Testa too crumpled to show cell structure. REMARKS AND AFFINITIES. The capsule figured by Chandler (1926, pl. 6, fig. 76) is now represented by two only of its valves. V.20083a which was figured by Chandler (1926, pl. 6, fig. 7a) and again in this present work and V.20084a. In addition there is the beautifully preserved fruit now fractured longitudinally (V.42197, see Pl. 20, figs. 136, 137). V.20084b shows two valves of another fruit and V.20084c, d are valves of doubtful identity. The characters are those of Ericales. Loculicidal dehiscence and slightly sunk apex seen in Clethraceae, Pyrolaceae, Ericaceae and Epacridaceae. Clethraceae are three-loculed. Pyrolaceae and Ericaceae of this type have much larger capsules with thin walls and usually with numerous seeds. The Stypheliae of the family Epacridaceae appear most to resemble the fossil. The locules are single-seeded. Although the dry multilocular “‘stone’’ does not appear to dehisce there can be little doubt that maceration prior to germination causes splitting for the release of the seeds. Although the Epacridaceae are chiefly Australian they also grow in India, South America and the sandy heaths of the Malay Peninsula. The fossil species also occurs in the Lower Bartonian and the Highcliff Sands at Cliff End near Mudeford. Family EBENACEAE Genus DIOSPYROS Linnaeus Diospyros headonensis nom. nov. 1926. Diospyros antiqua Chandler, p. 38, pl. 7, fig. 2; text-figs. 21, 22. Diospyros headonensis Chandler : Chandler (in press), pl. 20, figs. 28, 29 ; pl. 21, figs. 1-5. Remarks. For additional details see Chandler (in press) where a specimen from the Lower Bagshot, Lake, Dorset, with a young fruit was described and illustrated in addition to further Hordle material. Respective registered numbers are also shown there. It may now be added that the calyx is occasionally five-partite (one of the calyces in V.42200) and the range of variation of size is greater than hitherto recorded (cf. V.42201). Several small specimens from Hordle, probably immature, are only about 2 mm. long as against the normal 4-5 to 6:5 mm. In a few cases the sepals have incurled edges giving a deceptive narrow-pointed form. V.20086d shows detached sepals. The name Diospyros antiqua is pre-occupied by Watelet (1866 : 207) for a leaf impression from the Eocene of Belleu, Paris Basin, France and the Hordle fruits have therefore been given the new name Diospyvos headonensis, 148 FLORA OF THE LOWER HEADON BEDS Family SYMPLOCACEAE Genus SYMPLOCOS Jacquin Symplocos headonensis Chandler (Pl. 30, figs. 139, 140) 1926. Symplocos headonensis Chandler, p. 40, pl. 7, fig. 3a-c; text-fig. 24. 1957. Symplocos headonensis Chandler: Chandler, p. 117, pl. 17, figs. 187, 188. 1960 ?Symplocos headonensis Chandler: Chandler, p. 215, pl. 31, figs. 58, 59. Symplocos headonensis Chandler : Chandler (in press), pl. 21, figs. 19, ? 20. Diacnosis. Endocarp truncate subglobular, ovoid or urceolate, three-to-five- loculed ; external surface smooth formed of fine equiaxial cells. Length, 5 to6mm. ; breadth, 3:5 to 5:5 mm. Neotyre. Brit. Mus. (N.H.), No. V.42202. REMARKS. The smooth external surface of the endocarp is formed of equiaxial cells 0-016 to 0-025 mm. in diameter and the external carpel wall may be as much as 0-4 to 0-45 mm. thick at the middle. For relationship to living species see Chandler (1926: 40). The originals of Chandler, 1926, pl. 7, fig. 3a, c are now decayed while that of fig. 3b (V.20087) is in poor condition. Hence the selection of a neotype from fresh material. Numerous endocarps are still extant (V.42204). Symplocos sp. 1926. Symplocos sp. 2, Chandler, p. 40, pl. 7, fig. 4; text-fig. 25. Symplocos sp. Chandler (in press), pl. 21, figs. 21, 22. Although no further material has been found at Hordle, a specimen is known from the Dorset Pipe-clay Series. The small apical depression and smooth external surface appear to be constant features but the evidence does not seem sufficiently distinctive for specific diagnosis. The original endocarp from Hordle (V.20088) is now much broken but shows remains of two seeds extracted from it (V.200890). Family STYRACACEAE Genus STYRAX (Tourn.) L. Styrax elegans Chandler 1926. Styvax elegans Chandler, p. 41, pl. 7, fig. 6a-f. Styvax elegans Chandler: Chandler (in press), pl. 21, figs. 24-28. The holotype (V.20091) was illustrated by Chandler (1926, pl. 7, fig. 6a—c ; in press, pl. 21, figs. 26, 27). The specimen represented in 1926, pl. 7, fig. 6d is now decayed. V.20092 was illustrated in 1926 in pl. 7, fig. 6¢ to show the internal nerves of the testa. Broken endocarps somewhat decayed are preserved in V.20093-94, V.42205— 06. The latter shows a seed with one concave surface owing to pressure of a second seed in the fruit during growth. The discovery of the species in the Dorset Pipe- clay Series at Lake may be noted. FLORA OF THE LOWER HEADON BEDS 149 Family OLEACEAE Genus OLEA (Tourn.) L. Olea headonensis n. sp. (Pl. 30, figs. 141-143) 1926. Oleaceae Genus ?, Chandler, p. 42, pl. 7, fig. 7; text-fig. 27a~-c. Diacnosis. Endocarp subovoid but with slight bisymmetry, two-loculed, one locule being abortive, dehiscing loculicidally. External surface smooth or slightly rugose. Length of endocarp, 3:25 to 5-5 mm.; breadth in plane of symmetry, 2:25 to 4:25 mm.; thickness, 2 to 3-5 mm. Hototype. Brit. Mus. (N.H.), No. V.42207. DEscRIPTION. Endocarp : Woody, subovoid but slightly bisymmetric, angled at the apical end in the plane of symmetry, dehiscing into equal valves loculicidally, two-loculed, one locule being abortive, having one developed pendulous seed in the mature locule. Attachment scar large, basal, leading into a short oblique fibro- vascular canal (PI. 30, figs. 142, 143 and Chandler, 1926, text-fig. 27b) which lies in the plane of dehiscence. External surface smooth, or only slightly rugose, usually without fibres although occasional fibre impressions are seen. Locule lining transversely striate and puckered, shining, with underlying transverse fibres. Smooth suture for dehiscence showing fine fibres variously aligned. Dimensions of four endocarps: 1) Length, 4°75 mm.; breadth, 3°75 mm.; thickness, 3 mm. ; thickness of wall, o-°8 mm. 2) Length, 3-75 mm.; breadth, 2:75 mm.; thickness, 25mm. 3) Length, 5-5 mm.; breadth, 4:25 mm.; thickness, 3-5 mm. 4) Length, 3°25 mm.; breadth, 2:25 mm.; thickness, 2 mm. Seed : Much crumpled as preserved. Testa formed of double-walled rectangular cells frequently measuring o'r by 0:05 mm. The seed shows remains of dark- coloured broad flat fibre bands (now much shattered). REMARKS AND AFFINITIES. Perfect endocarps are V.42208, V.42210. There are also an endocarp split loculicidally (V.42211) and numerous fragments (V.42212). V.42209 was figured by Chandler, (1926, pl. 7, fig. 7; text-fig. 27a-c). Form, locules, dehiscence and funicle, also the structure of endocarp and seed suggest Olacineae, family Oleaceae. No Recent Olea has been seen without external fibres associated with marked rugosities. Their absence in the fossil is no doubt due to abrasion as in some species of Olea they are readlly removable and traces of them have been seen in O. headonensis. Endocarps similiar in form and size are found in the living genus Olea which today occurs in the Mediterranean, South Africa, the East Indies, Polynesia and Australia. Family BORAGINACEAE Genus OMPHALODES Tourn. ex Moench. Omphalodes platycarpa Chandler 1926. Omphalodes platycarpa Chandler, p. 42, pl. 7, fig. 8a, b. HoLotypPe. Brit. Mus. (N.H.), No. V.20095. REMARKS, The original of Chandler, 1926, pl. 7, fig. 8b is decayed. V.20095a 150 FLORA OF THE LOWER HEADON BEDS is a carpel partially dissected to show the inner surface (figured Chandler, 1926, pl. 7, fig. 8a). Additional material is needed to confirm the determination. Family ACANTHACEAE Genus ACANTAUS Linnaeus ? Acanthus sp. (Pl. 30, figs. 144, 145) DESCRIPTION. Seed: Much compressed, more or less quadrangular in outline, the angles being somewhat rounded. Two of the adjacent sides of the quadrangle are of approximately equal length and are appreciably shorter than the other two. A large elliptical marginal scar is seen at the middle of one of the shorter sides, this side being concave. The other short side is slightly convex. Unfortunately the angle where the two longer sides meet is broken. One broad surface of the seed is more or less flat, the other is concave near the hilar scar but has a long narrow curved facet extending from the angle where the two shorter sides meet to that where the two longer sides meet. It is probably caused by pressure of an adjacent seed or by contact with the fruit wall during growth. Surface ornamented all over with close- set puckerings many of which appear as shining black convex areas with toothed outlines. These areas may be circular, elliptical or irregular and are about 0:05 to O-I mm. in diameter. In general the surface is rough owing to fine equiaxial pits but a few of the black convex puckerings do not show the pitted tissue seen elsewhere. Along a line between the broken angle of the seed and the hilum the puckerings become somewhat elongate. Maximum diameter of seed, 4 mm. ; diameter at right angles to this, 3 mm.; thickness at right angles to the broad surfaces, I to I°5 mm. REMARKS. The form and puckered surface suggest Acanthaceae. The fossil most resembles a few seeds in the Reid Collection found as accidental impurities in a packet of Acacia catechu from an Indian Collection of seeds presented by the Keeper of Kew Herbarium. These were provisionally identified as Acanthus sp. Family CAPRIFOLIACEAE Genus SAMBUCUS (Tourn.) L. Sambucus parvulus Chandler 1926. Sambucus parvulus Chandler, p. 43, pl. 7, fig. 9a—c. Sambucus parvulus Chandler: Chandler (in press), pl. 23, figs. I-10. DiaGnosis. Seed rr to 1-5 mm. long. Surface with about six to ten sinuous transverse ridges with interspaces as broad as or broader than the ridges. Hototyre. Brit. Mus. (N.H.), No. V.20006. DEscCRIPTION. The original description has been emended in minor respects and now stands as follows: Seed exceptionally small and thin, obovate or elongate- obovate, compressed, often concavo-convex, anatropous, micropyle and hilum FLORA OF THE LOWER HEADON BEDS 151 terminal on the ventral face ; raphe ventral median longitudinal marked externally by an obscure longitudinal angle. Surface ornamented with from six to ten sinuous, sometimes interrupted transverse ridges with interspaces as broad as or broader than the ridges. Testa formed externally of equiaxial cells about 0-012 to 0-016 mm. in diameter. Length of seed, 1-r to 1-5 mm. ; breadth, 0-6 to I mm. RemARKsS. Although the majority of the seeds were found in Bed 10 east of Beckton Bunny, a few occurred in Bed 10 below Hordle House. The well repre- sented seeds are readily distinguished by their small size and delicate appearance from any living Sambucus seed. The wide spacing and relative coarseness and fewness of the transverse ridges is a distinctive feature. A single seed was found in the Dorset Pipe-clay Series at Lake. Although primarily a temperate genus of the Northern Hemisphere it also occurs in the East Indies, Australia, Tasmania and South America. In addition to specimens figured there are numerous seeds (V.42214). Family CUCURBITACEAE Genus CUCURBITOSPERMUM Chesters, 1957 : 56 Cucurbitospermum reidii n. sp. (Pl. 30, fig. 146) 1926. Cucurbitaceae Genus ?, Chandler, p. 43, pl. 7, fig. 10a, b. Diacnosis. Seed subtriangular in outline, elongate and narrow at the base, truncate at the apex, with a raised somewhat rugose median area pointed at both ends and a flat smooth marginal area which disappears at the apex, ends abruptly at the base and corresponds with an internal suture. Length of valve, 8-5 mm. ; breadth, 4:5 mm. HoLtotyre. Brit. Mus. (N.H.), No. V.20008. DEscRIPTION. Seed: Subtriangular in outline, elongate and narrow at the base, truncate at the apex, flattened, with a raised ornamental median area which is pointed at both ends. The ornamentation is of elongate rugosites aligned more or less at right angles to the margin of the raised area. There is a broad smooth flat marginal rim which does not extend around base or apex ; at the apex its termina- tion causes the truncate appearance. At the base also it ends abruptly. Hence at both ends of the seed the pointed extremities of the raised median area project beyond the rim. An internal hollow corresponds to the raised area and a marginal internal suture to the external smooth rim (Chandler, 1926, pl. 7, fig. roa, 6). The broad suture indicates that the specimen is one valve of a bisymmetric seed which has split marginally in the plane of symmetry, the other valve is missing. Length of valve, 8-5 mm.; breadth, 4:5 mm. REMARKS AND AFFINITIES. A single seed from Bed Io below Hordle House. Both apex and base of this seed are very slightly imperfect, and the specimen is abraded so that the micropyle and raphe are obscure but form and structure indicate Cucurbit- aceae. The seed is of a distinctive form which has not been matched precisely among the living genera that could be examined. Hence the reference to the form- 152 FLORA OF THE LOWER HEADON BEDS genus Cucurbitosbermum. A similar raised median area occurs in Sicydium Schlechtd. a tropical American genus, otherwise not closely similar, and in Lagenaria Ser. the median area is of the same shape but is not conspicuously raised and rugose as in the fossil. Family ? Genus RHAMNOSPERMUM Chandler, 1925 : 30 Rhamnospermum bilobatum Chandler 1960. Rhamnospermum bilobatum Chandler : Chandler, pp. 216, 236, pl. 32, figs. 60, 61 ; pl. 35, figs. 155, 150. Rhamnospermum bilobatum Chandler: Chandler (in press), pl. 23, figs. 18-38, pl. 24, figs. 1-9. See also for earlier references. The species occurs abundantly at Hordle in Bed io, east of Beckton Bunny (V.42216) and below Hordle House (V.42215) and is rare in Bed 29 (V.20047a). In addition to the above there are V.20064a (Chandler, 1926, pl. 5, fig. 1a), V.20064 (Chandler, 1926, pl. 5, fig. 15; Chandler (in press), pl. 23, fig. 18) and V.20064c (Chandler, 1926, pl. 5, fig. 1c ; Chandler (in press) pl. 23, fig. 19). For range of species and its variations, also remarks on affinities see Chandler, Flora of the Dorset Pipe-clay Series (in press). Carpolithus fibrosus n. sp. 1926. Carpolithus sp. 7, Chandler, p. 46, pl. 8, fig. 5a, b; text-fig. 30. DiaGnosis. Fruit syncarpous, two to six-carpelled, seated on a low conical re- ceptacle with free persistent sepals, obovoid or ovoid, splitting loculicidally. Endo- carp thick and woody formed of stout longitudinal fibres. Length of fruit, 4-25 to 6mm. ; breadth, 3:25 to 4:5 mm. Seeds linear, anatropous, about 2 to 3 mm. long with hyaline integuments. Ho.otype. Brit. Mus. (N.H.), No. V.20108. DEscrIPTION. Fruit: Syncarpous, woody, ovoid or obovoid, superior, attached to a low conical receptacle which bears at its margin a small calyx of persistent free sepals. (Both calyx and receptacle often broken away as in holotype, see Chandler, 1926, pl. 8, fig. 5a.) Two to six-carpelled, splitting loculicidally especially at the apex but the smooth surfaces of the septa rarely seen except towards the base. Interior of fruit occupied almost completely by longitudinally aligned agglomerated woody fibres (endocarp?) among which slender pendulous seeds lie in elongate narrow cavities. Whether there is more than one seed in a cavity has not been determined. External surface rough, exocarp about 0-2 mm. thick formed of shining, longitudinally aligned cells of irregular shape about 0-03 mm. in diameter. There are a few embedded longitudinal fibres in this coat, one such delimiting the margin of each carpel. Internally each carpel has a conspicuous median longi- tudinal ridge formed of woody fibres especially noticeable at the apex where the fibres are somewhat contorted and appear to form placentas, Length of fruit, 4:25 to6mm.; diameter, 3:25 to 4-5 mm, FLORA OF THE LOWER HEADON BEDS 153 Seed : Linear, compressed or inflated, anatropous. Hilar surface flat, oval ; chalaza a black oval thickened scar 0°35 by 0-25 mm. in diameter at the opposite end to the hilum situated therefore at the end of the seed near the base of the locule. Integuments thin. The coat usually preserved (tegmen?) is hyaline formed of longitudinally elongate parallel-sided cells 0-012 mm. broad, square or facetted at the ends. Outside this a rough coat is sometimes seen with similiar cells and there is sometimes a coat of equiaxial cells about 0-01 to 0-016 mm. in diameter (Chandler, 1926, text-fig. 30a) which may be part of the testa, or may belong to the locule wall. Sometimes the hyaline integument bears internal impressions of large rectangular cells 0-016 to 0:025 mm. in diameter. Transversely elongate cells arranged in longitudinal rows have also been seen internally (Chandler, 1926, text-fig. 300). Length of seed, 2 to 3 mm. REMARKS. Fruits common but often imperfect owing to the tendency to break longitudinally again and again. V.z010Q¢ is an imperfect fruit with remains of receptacle still attached. V.20109 and V.20109¢ are longitudinally fractured fruits. V.42219 has the pericarp abraded so that the fibre mass is exposed with some seeds showing through it and V.42220 is a similar specimen. Woody interior is well shown in V.42218. In V.42217 half of a longitudinally fractured fruit shows seeds one with a black oval chalazal scar. Other specimens or segments of fruits are V.20109, V.42221-23. The original of Chandler, 1926, pl. 8, fig. 5b is decayed. The botanical position has not yet been determined but this readily recognizable species appears to be worthy of a specific name especially as it has now been found also in the Bournemouth Beds. There is a certain resemblance between this and the specimen from the London Clay called Tvochodendron? pauciseminum (Reid & Chandler, 1933 : 155, pl. 3, figs. 29-33) but that was thought to be septicidal. If further London Clay material could be found a comparison should be made, observa- tions checked and the relationship reconsidered. The size is similar as are the coats of the slender seed. Carpolithus apocyniformis n. sp. (Pl. 30, figs. 147, 148) Diaenosis. Detached valves of a loculicidal multilocular capsule, narrow-ovate in outline, slightly convex on the external surface which is finely furrowed and longitudinally striate, slightly concave on the inner surface which has a median longitudinal septal ridge throughout. Length of valves about 4:25 to 6°5 mm. ; breadth about 1°5 to 3 mm. Hototyre. Brit. Mus. (N.H.), No. V.42224. DescripTION. Valve: Apparently part of a multilocular capsule now always detached, narrow-ovate in outline, pointed or rounded at the distal end (usually broken), truncate at the proximal end where broken from the fruit. External surface slightly convex, black and rather rough owing to short irregular longitudinal furrows, cell structure obscure but some cells with black shining contents can be seen. Inner surface slightly concave, smooth finely striate longitudinally with low median longitudinal ridge throughout its length which appears to be the remains of a 154 FLORA OF THE LOWER HEADON BEDS septum. Beneath the epicarp is a light brown layer formed of cells about 0-012 mm. long, 0-016 mm. broad. The valve is solid as seen in transverse section. Length of valve about 4-25 to 6:5 mm.; breadth, 1-5 to 3 mm. REMARKS, These specimens are not uncommon and two have occasionally been seen adhering by their lateral margins. The solid section excludes all possibility that they are seeds in spite of a superficial resemblance to seeds of Apocynaceae. The adherent specimens, the constant presence of the median ridge and the trans- verse break at one end suggests that they are detached valves of a dehisced capsule but no suggestion as to relationship can at present be made. The distal end of the holotype is imperfect but the form of the unbroken valve tip has been seen on a broken specimen with other imperfect valves in V.42226. There is also an almost perfect tip in V.42227 and in the figured specimen V.42225 (somewhat pyritized). Further material will be described in due course when the British Museum mono- graph on the Bournemouth Beds is published. Carpolithus sp. (Pl. 30, figs. 149-153) 1926. Carpolithus sp. 8, Chandler, p. 47; text-fig. 31. DESCRIPTION. Ovary: Superior, conical with flat pentagonal basal perianth disc showing remains of peduncle. Carpels normally three (rarely four), syncarpous with three short apical styles and three locules. Epicarp thin, loosely fitting, shining, formed of square or hexagonal cells, carpel wall within spongy in transverse section. Locule lining of transversely elongate cells arranged in vertical rows, longitudinally puckered the puckerings sometimes branching and anastomozing. Seeds not seen. Length of ovary about 3 mm. ; breadth, 1-75 to 2-5 mm. REMARKS. These specimens seem to be immature fruits or ovaries rather than ripe fruits. They are often collapsed and distorted. The relationship has not been discovered. The following are three-loculed: V.20110, V.20110a (Pl. 30, fig. 151), V.201100 (Pl. 30, fig. 150, shows perianth disc clearly), V.2or10c (Pl. 30, fig. 149), V.20110d (PI. 30, fig. 153). V.2z0r10e (PI. 30, fig. 152) is a rare four-loculed specimen. Numerous fruits are registered under V.42228. ? Carpolithus gardneri Chandler 1926. Ovites sp. Chandler, p. 47, pl. 8, fig. 6. V.15520 the finest specimen was originally described as Ovites sp. There are in addition nineteen other specimens V.421a-s. All are now much decayed. Some are merely detached fruits. One internal cast shows fine transverse striations. As noted in 1926 there is a resemblance to a species from the Reading Beds now described as Carpolithus gardneri (Chandler, 1961 : 85, pl. 9, figs. 1-5 ; text-fig. 4) both in the branching racemose habit and structure of the pods but C. gardneri appears to have more slender fruits and it is doubtful whether the two are identical or whether there is sufficient evidence to connect the Hordle species with Orites FLORA OF THE LOWER HEADON BEDS 155 in spite of a similarity of habit. The specimens are preserved in a sandy whitish- grey clay. The locality in Hordle cliff is almost certainly Bed 10 as the matrix agrees with that of “ Nelumbium buchi Ett.’’ (Chandler, 1926 : 48). Carpolithus sp. (Pl. 30, figs. 154-156) 1926. Carpolithus sp. 1, Chandler, p. 44, pl. 7, fig. 114, b. This six-lobed solid organ is still unidentified. Each of the six lobes has two transverse constrictions and new material shows remains of a shining epidermis of equiaxial cells. The comparable Recent object found as described in 1926 as an impurity in a packet of seed of Zelkova Keaki from the Yokohama Nursery Company is figured in PI. 30, fig. 157. Length of specimens, 3 to 6 mm. ; transverse diameter, Ito3°5mm. V.422209 (see Pl. 30, figs. 155, 156) and V.42230-31 are extant but the specimen figured in 1926 is now decayed. Carpolithus sp. (Pl. 30, fig. 158) DESCRIPTION. Seed: Small, bisymmetric, curved, reniform represented by one valve only. Hilum near the middle of the concave margin. Other organs obscure but as one limb is narrower than the other the micropyle probably lay at its tip. Hilum a conspicuous aperture leading directly to the seed cavity. As seen in section the wall of the chalazal limb somewhat overlaps the wall of the micropylar limb and is recurved lying closely upon itself and passing into the tegmen. Testa surface of equiaxial cells superficially which diverge from the hilum, are about 0-012 to 0-016 mm. in diameter, and are so thick-walled that the small cavity appears as a central pit or depression. In section the testa is columnar and about 0:05 mm. thick. Lining of testa obscure but apparently formed of digitate cells. Maximum diameter of seed about I-15 mm.; diameter through the hilum at right angles to this about o-9mm. The seed has split in the plane of symmetry as for germination. Its systematic position is not known. Carpolithus sp. (Pl. 30, fig. 159) DEscRIPTION. Lvuit: Subglobular, syncarpous, four-loculed, locules unequally developed, opening by dorsal valves which gape at the apex and sides but break irregularly below. No central canal seen but there appears to be a thickened scar on the median line of the ventral wall of each locule at about one-quarter of the length from the apex which may mark the attachment of a solitary seed. Walls woody, septa formed internally of equiaxial cells about 0-016 mm. in diameter, but superficially of elongate cells parallel with the surfaces, the length of these cells being about 0:05 mm. and their breadth about 0-008 mm. Cells of the locules elongate, 156 FLORA OF THE LOWER HEADON BEDS about 0:007 mm. broad and of variable length, frequently 0-05 mm. These cells are arranged in groups on the whole obliquely aligned and diverging from the sub- apical placenta (?) and below this point upwards from the median ventral line of the locule. Seed: Agreeing with the locule in shape, with large thickened basal chalaza ; testa of longitudinally elongate cells, o-or mm. broad. Remarks. A single fruit of unknown affinity now dissected. Gall or Tuber? (Pl. 30, fig. 160) DescriPTION. A subglobular hollow body, burst at one end (now fractured into two pieces) with scaly surface having irregular rugosities separated by deep cracks sometimes arranged concentrically around a more or less smooth area. There are a few small circular apertures seen especially on the smooth areas. Wall finely columnar in section and about 0-05 mm. thick. Within are remains of a yellow semi-translucent skin of chitinous appearance in which no cell walls can be seen. Dimensions, 2°5 by 2 mm. Remarks. A gall or tuber? which closely resembles similar organisms from the Dorset Pipe-clay Series (Chandler, (in press), pl. 27, figs. 26-29) and from the London Clay (Chandler, 1961 : 324, pl. 32, fig. 33). Identity not determined. REFERENCES CHANDLER, M. E. J. 1923. The Geological History of the Genus Stvatiotes: An Account of the Evolutionary Changes which have occurred within the Genus during Tertiary and Quaternary Times. Quart. J. Geol. Soc. Lond., 79 : 117-138, pls. 5, 6. 1925. The Upper Eocene Flova of Hordle, Hants., 1. 32 pp., 4 pls. [Mon. Palaeont. Soc., London.] 1926. The Upper Eocene Flora of Hordle, Hants., 2. vii-+ 20 pp., 4 pls. [Mon. Palaeont. Soc., London. ] 1957. The Oligocene Flora of the Bovey Tracey Lake Basin, Devonshire. Bull. Brit. Mus. (Nat. Hist.) Geol., London, 3 : 71-123, pls. 11-17. 1960. Plant Remains of the Hengistbury and Barton Beds. Bull. Brit. Mus. (Nat. Hist.) Geol., London, 4 : 191-238, pls. 29-35. 1961. The Lower Tertiary Flovas of Southern England. 1. Palaeocene Flovas. London Clay Flova (Supplement). xi + 354 pp., atlas 34 pls. Brit. Mus. (Nat. Hist.), London. 1961a. Post-Ypresian Plant Remains from the Isle of Wight and the Selsey Peninsula, Sussex. Bull. Brit. Mus. (Nat. Hist.) Geol., London, 5: 13-41, pls. 4-11. —— The Lowey Tertiary Flovas of Southern England. 2. Flova of the Pipe-clay Series of Dorset (Lower Bagshot.) {In press]. Cuesters, K. I. M. 1957. The Miocene Flora of Rusinga Island, Lake Victoria, Kenya. Palaeontographica, Stuttgart, 101, B: 30—-71. pls. 19-21. ERDTMAN, G. 1960. On Three New Genera from the Lower Headon Beds, Berkshire. Bot. Notisey, Lund, 113 : 46-48, pls. 1, 2. GaRDNER, J. S. 1883-86. A. Monograph of the British Eocene Flora, 2: Gymnospermae. 159 pp., 27 pls. (Mon. Palaeont. Soc., London.|} 1886. Report on the Fossil Plants of the Tertiary and Secondary Beds of the United Kingdom. Rep. Brit. Ass., London, 1885 : 396-404, pls. 1-3. FLORA OF THE LOWER HEADON BEDS 157 GARDNER, J. S. & ETTINGSHAUSEN, C. von 1879-82. A Monograph of the British Eocene Flova,1: Filices. 86 pp., 13 pls. [Mon. Palaeont. Soc., London.] Hastines, Marchioness of 1852. Description géologique des falaises d’Hordle, sur la cote du Hampshire, en Angleterre. Bull. Soc. géol. Fy., Paris (2) 9 : 191-203. HeEER, O. 1862. On the Fossil Flora of Bovey Tracey. Philos. Trans., London, 152 : 1039— 1086, pls. 55-71. KIRCHHEIMER, F. 1936. Beitrage zur Kenntnis der Tertiarflora. Friichte und Samen aus dem Deutschen Tertiar. ~Palaeontographica, Stuttgart, 82, B: 73-141, pls. 7-13. 1937. Palaobotanische Beitrage zur Kenntnis des Alters deutscher Braunkohlenschichten, 11. Die braunkohlefiihrenden Tone von Sieburg (Rheinland) und Kamenz (Oberlausitz). Braunkohle, Halle, 36 : 893-899, 915-919, 925-931, figs. I-26. 1948. Uber die Fachverhiltnisse der Friichte von Cornus L. und verwandter Gattungen. Planta, Berlin, 36 : 85-102, figs. 1-5. 1957. Die Laubgewdchse der Braunkohlenzeit. ix + 783 pp., 55 pls. Halle (Saale). Mat, D. H. 1960. Uber neue Friichte und Samen aus dem deutschen Tertiar. Paldont, Z., Berlin, 34: 73-90, pls. 3-6. Reip, C. & Groves, J. 1921. The Charophyta of the Lower Headon Beds of Hordle [Hord- well] Cliffs, South Hampshire. Quart. J. Geol. Soc. Lond., 77 : 175-192, pls. 4-6. — & StTRAHAN, A. 1889. The Geology of the Isle of Wight. 2nd edition. xiv + 349 pp., 5 pls. [Mem. Geol. Surv. U.K.] Rerp, E. M. & CHANDLER, M. E. J. 1926. The Bembridge Flova. Catalogue of Cainozoic Plants in the Department of Geology, I. viii + 206 pp., 12 pls. Brit. Mus. (Nat. Hist.), London. 1929. Palaeobotany or Vegetable palaeontology. pp. 7oc—96a, pl. 1. Encyclo- paedia Britannica, London. 14th edition. 1933. Ihe Flova of the London Clay. viii + 561 pp., 33 pls. Brit. Mus. (Nat. Hist.), London. SHAPARENKO, K. K. 1956. “‘Salviniaceae’”’. In Palaeobotanica, 2. Edited by A. L. Takhtajan. Acta Inst. Bot. Komarov. Acad. Sct. URSS, Moscow & Leningrad (8) : 14-44, pls. 1-3. TAWNEY, E. B. & KreEpine, H. 1883. On the Section at Hordwell Cliffs, from the top of the Lower Headon to the base of the Upper Bagshot Sands. Quart. J. Geol. Soc. Lond., 39: 566-574, fig. I. VisIANI, R. 1858. Piante fossili della Dalmazia. Mem. R. Ist. veneto, Venezia, 7: 423-455, pls. 15-20. WATELET, A. 1866. Descriptions des Plantes fossiles du bassin de Paris. 264 pp., atlas 60 pls. Paris. EXPLANATION OF PLATES Unless specifically stated otherwise all specimens figured are from Hordle. PLATE 24 Salvinia mildeana Goeppert Fic. 1. Part of the leaf (figured Chandler, 1925, pl. 1, fig. 1a) to show arrangement of papillae between secondary nerves. 4. (V.36344.) Acrostichum lanzaeanum (Visiani) Fic. 2. Pinnule fragment. x2. (V.42058.) Alleged to come from Barton clay, Barton. For comparison with Hordle material previously figured. Potamogeton pygmaeus Chandler Fics. 3, 4. Abraded endocarps. In fig. 3 keel tip broken. 15. (V.42054-55.) Limno- carpus Band, Colwell Bay, Isle of Wight. Figs. 5,6. Endocarps from Hordle for comparison. Fig. 5 shows gaping keel, fig. 6 the style. x15. Decayed. Limnocarpus forbesi (Heer) Fic. 7. Endocarp, side, with style and lateral foramen; slightly imperfect below. x15. (V.42064.) Limnocarpus Band, Colwell Bay, Isle of Wight. Fic. 8. Another, similar aspect, tip of keel broken. X15. (V.42062.) Lower Headon, bomb crater, Downton, Hordle. Fic. 9. Endocarp, ventral, with style and stalk. x15. (V.42065.) Fic. 10. Endocarp, dorsilateral, tilted to show tip of keel. x15. (V.42066.) Fic. 11. Endocarp, dorsilateral, sand filled cavity exposed by lossofkeel. 15. (V.42061.) Figs. 9-11 from Limnocarpus Band, Colwell Bay, Isle of Wight. Stratiotes hantonensis Chandler Fics. 12,13. Inner surface of two valves from sigmoidal seeds showing short raphe. 6:5. (V.42084, V.42086.) Fic. 14. Seed somewhat dorsiventrally crushed. 6:5. (V.42085.) Fies. 15, 16. Two seeds showing sigmoidal and hooked forms. x6-5. (V.42083, V.420862.) Stratiotes headonensis Chandler Fic. 17. Typical seed for comparison with the above. x6. (V.200314a.) Caricoidea minima (Chandler) Fic. 18. Holotype. Endocarp tilted to show basal aperture and a longitudinal rib. X15. (V.20033.) Fic. 19. Fruit, side. X15. (V.42087.) Fic. 20. Another endocarp, side. X15. (Decayed.) Fic. 21. Larger fruit. x15. (V.42088.) Caricoidea obscura Chandler Fic. 22. Fruit, side. X15. (V.42092.) Fics. 23, 24. Twoendocarps. x15. (V.42093-94.) Scleria hordwellensis n. sp. Fic. 25. Holotype. Fruit, apex, calyx can be seen projecting around the circumference. xX6:5. (V.42006.) Fic. 26. Same, side, at the base is seen thick calyx with transverse puckerings, (s) style. X15. CYPERACEAE, Genus ? Fic. 27. Abraded fruit, side, (s) style. 15. (V.42097.) Spivematospermum headonense n. sp. Fic. 28. Holotype. Fruit broken at stylar end so that seeds are exposed. XI approx. (V.42098.) Fic. 29. Same with near wall removed to expose more of the seeds. 1-3. Fic. 30. Another fruit broken at stylar end from which seeds have fallen. 1 approx. (V.42099.) Bull. B.M. (N.H.) Geol. 5, 5 PLATE 24 ihe 5 Dy iJ : id. ‘hy 16§ is oa i PLATE 25 Spivematospermum headonense n. sp. Ime, Bue Seale Se, (N25 ¥7/,)) Fic. 32. Seed. xg. (V.42057a.) In matted mass of Acvrostichum lanzaeanum said to come from Barton clay, Barton cliff. Myrica boveyana (Heer) Fic. 33. Endocarp abraded surface cracked along basal attachment and placenta. x 6:5. (V.42102.) Fic. 34. Valve of another, interior showing locule and broad flat marginal suture. X15. (V.42103.) Carpinus boveyanus (Heer) Fic. 35. Abraded fruit, (b) basal attachment scar. Longitudinal furrows due to calyx fibres are obscurely seen. X15. (V.42104.) Broussonetia yugosa Chandler Fic. 36. Holotype. Endocarp slightly broken. 15. (V.20039.) Fic. 37. Another endocarp. X15. (V. 20039b.) Moyvoidea hordwellensis n. sp. Fic. 38. Holotype. Fruit, side, (st) style; (f) funicle; (ch) position of chalaza within. X15. (V.42105.) Becktonia hantonensis n. gen. & sp. Fic. 39. Holotype. Endocarp, side (bursting below), (a) subapical foramen for funicle. xX 6:5. (V.42106.) Hantsia pulchva (Chandler) Fic. 40. Seed, side showing curvature, (#) rimmed hilum. X15. (V.42107.) Hantsia glabra n. sp. Fic. 41. Holotype. Seed, side, smooth surface with finely toothed cells, (#) rimmed hilum. 5 (V4 2a) Fic. 42. Same, opposite side. Aperture of hilum (h) turned towards camera. X15. Fic. 43. Slightly distorted seed, () hilar rim (slightly incomplete). 15. (V.42I1TI0.) Brasenia oblonga Chandler Fic. 44. Seed, side, longitudinal raphe on left. Tubercles wellshown. X15. (V.200434.) Fic. 45. Same, hilar end looking on to aperture from which embryotega has fallen. x15. Fic. 46. Holotype. Uncrushed seed. X15. V.20043. Palaeosinomenium obliquatum (Chandler) Fic. 47. Holotype. Endocarp (s) stylar limb, elongate foramen (f) seen near stylar limb. X6:5. (V.20049.) Fics. 48, 49. Much broken valves of a second endocarp, (st) stylar limb. 6-5. (V.20050.) Fic. 50. Stylar limb of another endocarp showing ornamentation of marginal flange. X15. (V.42122.) LAURACEAE, Genus ? Fic. 51. Berry, apex (crushed dorsiventally) 6:5. (V. 42123.) Fic. 52. Same in longitudinal section showing conical endocarp with columnar wall, meso- carp represented by light coloured pyrites. 6:5. Bull. B.M. (N.H.) Geol. 5, 5 PLATE 25 PLATE 26 Aldvovanda ovata (Chandler) Fic. 53. Seed showing long hilar neck, mucro at distal end, and longitudinal raphe ridge (on right). x15. (V.42126.) Fic. 54. Another with somewhat abraded surface. 15. (V.42124.) Fic. 55. Seed in longitudinal section showing prismatic cells within testa. x15. (V.42125.) Colwell (presumed Lower Headon). Eoliquidambar hordwellensis n. gen. & sp. Fic. 56. Holotype. Mature fruiting head, base, showing areoles. 2-6. (V.421314a.) Fic. 57. Same, different aspect, showing a fruit (f) with two crescentic valves in an areole. X 2:6. Fic. 58. Three areoles of same showing wrinkled shining epidermis over the valves of enclosed fruits, the evidence that they bore no awns. X6°5. Fic. 59. Immature fruiting head with areoles enclosing two-valved fruit apices. X6'5. (V.42131.) Protoaltingia hantonensis Chandler Fic. 60. Much abraded fruiting head. x6. (V.20054.) LEGUMINOSAE, Genus ? - Fic. 61. Crushed seed. x6:5. (V.43696.) Fic. 62. Another genus. Hilum (h) associated with tumescent area. X15. (V.43697.) Zanthoxylum hordwellense n. sp. Fic. 63. Holotype. Seed, side. Long hilar scar on right, (ch) position of chalaza. X15. (V.20059.) Fic. 64. Same, ventral view showing narrow-triangular hilar scar with orifice for raphe at its lower end. X15. Zanthoxylum compressum Chandler Fic. 65. Holotype. Seed, side, short hilar scar on right. 6:5. (V.20061.) Fic. 66. Small area of same showing characteristic surface ornamentation. X15. Fic. 67. Another seed with abraded surface and shorter hilar scar (right). (ch) indicates position of chalaza. 6:5. (V.42137.) Rutaspermum ornatum (Chandler) Fic. 68. Seed, broken on dorsal side, showing cell structure and surface ornamentation, also smooth area around hilar scar (right). (ch) position of chalaza. X15. (V.42145.) Bull. B.M. (N.H.) Geol. 5, 5 PLATE 26 iH l) sy if o ed PLATE 27 Rutaspermum ornatum (Chandler) Fic. 69. Neotype. Seed, hilar scar to left. 6:5. (V.42142.) Fic. 70. Another slightly tilted to show hilar scar on right, smooth area around it clearly seen. X6°5. (V.42143.) Fic. 71. Smaller seed, hilar scar to left. x6:5. (V.42144.) ANACARDIACEAE SPONDIEAE, Genus ? Fic. 72. Fragment of fruit showing sac-like, closely packed, secreting cavities exposed by fracture and abrasion. X15. (Decayed.) Fic. 73. Another fragment of same. X15. Palaeoburseva lakensis Chandler Fic. 74. Endocarp, ventral, (broken on left) showing cell structure. (pf) Aperture leading to placenta. X15. (V.42147.) Fic. 75. Same, dorsal. X15. ? Todes sp. (or ? Natsiatum sp.) Fic. 76. Funicular margin of an endocarp, (st) style. Funicular canal exposed below by a break in its outer wall. x4. (V.20063.) Fic. 77. Same, interior, (p) placenta. x4. Iodes ? hordwellensis n. sp. Fic. 78. Holotype. Endocarp, apex, funicle on left. 2-8. (V.42153.) Fie. 79. Same, side, (st) style, funicle on left. x15. Icacinicarya transversalis n. sp. Fic. 80. Holotype. Valve of endocarp, exterior (slightly imperfect), funicle on right, (sé) style. x2°8. (V.23420.) Fic. 81. Same, interior, (st) style, funicle on left. 2-8. Fic. 82. Another distorted endocarp, broken at style (st). 2:8. (V.23419.) Icacinicarya becktonensis n. sp. Fic. 83. Holotype. Endocarp, side, abraded at apex, funicle on right. 6-5. (V.42155.) Fic. 84. Same, apex, abrasion has exposed circular hilar scar (s) on seed within and part of funicle to right of it. 6:5. Frangula hordwellensis n. sp. Fic. 85. Holotype. Pyrene, side, showing median longitudinal angle. (p) Fragment of outer pitted coat. Hilum at base. x65. (V.42156.) Fic. 86. Same, opposite, gently convex surface. X06°5. Bull. B.M. (N.H.) Geol. 5, 5 PLATE 27 | | PLATE 28 Frangula hordwellensis n. sp. Fic. 87. Seed in figs. 85, 86 fractured longitudinally to show internal surface and (h) hilar depression; (v7) marginal raphe; (m) micropyle; (ch) chalaza; (vd) ridge within hilar depression ; (¢) projecting testa forming rim around this depression. X15. Meliosma sp. Fic. 88. Endocarp. (a) Attachment and funicular aperture. 15:5. (V.42158.) Fic. 89. Another somewhat broader specimen. 15:5. (V. 42159.) Parthenocissus hordwellensis n. sp. Fic. 90. Seed, dorsal, with chalazal scar. 6:5. (V.42161.) Fic. 91. Same, ventral, showing long infolds. 6:5. Fig. 92. Another seed, dorsilateral. Chalaza on left. 6:5. (V.42161.) Fic. 93. Same, ventrilateral. Raphe ridge on left, facet with long infold seen. x 6:5. Fic. 94. Poorly preserved seed, dorsilateral. 6:5. (V.42237.) Fic. 95. Same, ventral, with infolds. x 6:5. Figs. 94, 95 from Horizon A3, Barton Beds, Highcliff for comparison. Tetrastigma lobata Chandler Fic. 96. Seed, ventral, showing lobing around infolds and raphe ridge. 6-5. (Decayed.) Fic. 97. Same, dorsal, showing sunk chalaza and radial lobes and furrows diverging from it O35) Actinidia sp. Fic. 98. Apical rounded end of a seed with pitted surface. X15. (V.42165.) Gordonia minima Chandler ~ Fic. 99. Fruit, lower end (upper part broken) tilted to show receptacle. 6-5. (¥V.20070@.) Fic. 100. Another (apex broken) tilted to show interior with broken central columella and bases of three valves below, lower end of two valves at v, v. 6:5. (Decayed.) Fic. 101. Fruit broken transversely showing five locules and septa. Valves triangular in section, columella at centre. 6:5. (V.200710.) Fic. 102. Fruit broken above, sectioned longitudinally showing remains of columella and of two valves. X6:5. (V.20071a.) Eurya becktonensis n. sp. Fic. 103. Holotype. Seed slightly broken dorsally. (p) Coarsely pitted thickening around hilum. X15. (V.42168.) Campylospermum hordwellense Chandler Fic. 104. Fruit, side, with attachment and fruit wall remaining below. 6:5. (V.20035.) Fic. 105. Same (as seen from opposite side i.e. in longitudinal section) showing axis and a seed on each side of it. 6:5. Fic. 106. Detached seed from another fruit showing beautifully preserved cell structure of testa ; (2) hilum ; (m) micropyle. Irregular thickening of coarse cells is well seen in hilar region. X15. (V.42170.) Fic. 107. Neotype. Valve of seed showing curved cavity, (m) micropyle. Raphe in wall between limbs opening into chalaza at tip of shorter limb. x15. (V.42169.) Cleyera ? stigmosa (Ludwig) Fic. 108. Seed, pitted surface poorly preserved. X15. (V.42173.) Fic. 109. Same broken irregularly longitudinally shows condyle between limbs of curved locule. X15. Fic. 110. Inner surface of fragment removed by fracture showing same features. X15. Fig. 111. Another. x15. (V.42174.) Fic. 112, Smallerseed. X15. (V.42175.) PEATES28 Bull. B.M. (N.H.) Geol. 5, 5 at ,) ia a a fi! m rv) . fr : sl 4 ff wi : | “pe pie 7 ry PLATE 29 Anneslea ? costata n. sp. Fics. 113, 114. Holotype. Seed, opposite surfaces, (2) hilum. X15. (V.42177.) Fic. 115. Another broken seed dissected to show condyle between curved limbs. x15. (V.42178.) Microdiptera parva Chandler Fic. 116. Seed, dorsal with operculum partly detached but embedded in white matrix. X15. (V.42182.) Fic. 117. Same, ventral, showing median longitudinal body, lateral wings breaking away from it. XI5. Mastixia ? glandulosa n. sp. Fic. 118. Endocarp, dorsal surface with external longitudinal furrow. x6:5. (V.42188.) Fic. 119. Holotype. Endocarp, ventral surface. 6:5. (V.42187.) Fic. 120. Laterally compressed endocarp with ventral wall removed so as to expose inner surface (z) of dorsal infold projecting into locule. 6-5. (V.42189.) Fic. 121. Another with part of external wall removed showing solidified resin secretions (light coloured) occupying cavities in middle layer of wall. 6:5. (V.42190.) MASTIXIOIDEAE, Genus ? Fic. 122. Endocarp, apex with two germination valves removed, the third which remains shows an infold at (1). x2 approx. (V.20106a.) Fic. 123. Same, side, one locule exposed by removal of a valve. 2 approx. Fic. 124. Upper part of the valve, inner surface showing infold. x2 approx. Dunstania glandulosa (Chandler) Fic. 125. Holotype. Endocarp, broken at base. 4. (V.20090.) Fic. 126. Same, sectioned longitudinally showing two locules and resin cavities in walls. x 4. Fic. 127. Neotype. Endocarp somewhat abraded, shows apex truncated by small depres- sion. X15. (V.42194.) Fic. 128. Broader abraded endocarp with resin cavities partly exposed superficially. x15. (V.42195.) Cornus quadrilocularis Chandler Fic. 129. Endocarp, side, apex excavated by small depression. 6:5. (V.42193.) Fic. 130. Same, opposite side with a valve partly removed showing locule, the ventral wall of which is broken at apex. 6:5. Fic. 131. Another endocarp, side, germination valve removed showing locule and apical depression. X06-5. (V.20106.) Fic. 132. Apex of same, three valves have been removed, one remains on left. 6:5. Fic. 133. One of valves removed, inner face. X6'5. Epacridicarpum headonense Chandler Fic. 134. Fruit, base. X15. (V.42197.) Fic. 135. Same, apex. XI5. Fic. 136. Valve of another, inner surface showing septum and two locules. X15. (V.20083.) Fic. 137. Same, side view, axis to left. X15. Rew e29 Bull. B.M. (N.H.) Geol. 5, 5 | i > ) we % 7 a =: 1 ee b ; FAD a PAA Eso ERICACEAE, Genus ? sp. 2. Fic. 138. Fruit, apex showing incipient splitting into valves. X15. (V.42108.) Symplocos headonensis Chandler Fic. 139. Neotype. Endocarp, apex showing depression with four apertures to locules and a central aperture to axial canal. 6:5. (V. 42202.) Fic. 140. Endocarp fractured longitudinally showing two locules (filled with white matrix communicating with apical depression). Central canal exposed in upper part of endocarp. At apex behind it is the apical aperture of third locule. 6:5. (V.42203.) Olea headonensis n. sp. Fic. 141. Small endocarp. (a) Attachment. 6:5. (V.42208.) Fic. 142. Holotype. Valve of endocarp which has split loculicidally. LLocule with short oblique funicular canal (f). 6:5. (V.42207.) Fic. 143. Same, opposite valve, incomplete above, (f) funicular canal. 6:5. ? Acanthus sp. Fic. 144. Seed, side, (z) hilum. x6:5. (V.42213.) Fic. 145. Same, opposite surface showing characteristic puckering. X15. Cucurbitospermum reidii n. sp. Fic. 146. Holotype. Valve of seed, exterior. 4. (V.20008.) Carpolithus apocyniformis n. sp. Fic. 147. Holotype. Detached valve of capsule imperfect at distal end. Interior with median ridge representing remains of septum. %15°5. (V.42224.) Fic. 148. Valve of another capsule, outer surface. 15:5. (V.42225.) Carpolithus sp. Fic. 149. Fruit with three styles. 6. (V.20110C.) Fic. 150. Fruit with three styles. 6. (V.20110b.) Fic. 151. Fruit with three styles broken on one side. x6. (V. 201104.) Fic. 152. Four-loculed fruit sectioned transversely. x6. (V.201102.) Fic. 153. Three-loculed fruit sectioned transversely. x6 (V.201I0d.) Carpolithus sp. Fic. 154. Six-lobed organ, lobes with slight transverse constrictions. 6:5. (Decayed.) Figs. 155, 156. Another, opposite ends. 6:5. (V.42229,.) Fic. 157. Recent specimen for comparison. Accidental impurity in seed packet from Yokohama Nursery Co. (unidentified). 6-5. Carpolithus sp. Fic, 158. Valve of reniform seed, interior. Hilum between limbs. x15. (V.42232.) Carpolithus sp. Fic. 159. Imperfect four-loculed endocarp. One valve faces camera but it is broken on left of median line exposing locule beneath it. Another abortive locule on left is not clear in figure ; (/) is a third locule. The fourth is behind and out of sight. 6:5. (V.42233.) Tuber ? Fic. 160. Organism with rough, scaley, cracking surface. 6:5. (V.42234.) Mastixicarpum crassum Chandler Fic. 161. Endocarp, dorsal, breaking longitudinally along median infold. The transverse fracture showed C-shaped locule. 2-4. (V.42235.) For comparison with Hordle material, Alleged to come from Barton clay, Barton cliff, PLATE 30 Bull. B.M. (N.H.) Geol. 5, 5 152 «4 " ‘a = waa, ; aoe ae THE FORM AND STRUCTURE OF CTENOZAMITES CYCADEA THOMAS. MAXWELL HARRIS BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 5 No. 6 LONDON : 1961 abe es) ih te ; miohy ise 5 THE FORM AND STRUCTURE OF CTENOZAMITES CYCADEA (BERGER) SCHENK BY THOMAS MAXWELL HARRIS, FE.RS. (Professor of Botany in the University of Reading) 2 0.JULI94! Pp. 159-173; Plates 31, 32; 2 Text-figures | BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 5 No. 6 LONDON : 1961 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, 1s issued in five series, corresponding to the Departments of the Museum, and an Historical senes. Paris will appear at trregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 5, No. 6 of the Geological (Palaeontological) sertes. © Trustees of the British Museum, 1961 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued July, 1961 Price Seven Shillings and Sixpence fake FORM AND STRUCTURE, OF CTENOZAMITES CYCADEA (BERGER) SCHENK By T. M. HARRIS SYNOPSIS This paper deals with a single magnificent specimen of the Lower Liassic leaf which has been called Ctenopteris cycadea (Brongn.). It is nearly complete and shows that the leaf is larger than had been supposed and also proves that the rachis forks. This forking brings Ctenozamites close to Ptilozamites and possibly also to Odontopteris. The specimen has also the merit of being partially petrified and gives some indication about its internal tissues. INTRODUCTION THE specimen (V.36330) bears the label “‘ Plant from the Flatstones Nodules, Black Marl, Black Ven, Charmouth, Dorset. Collected by J. F. Jackson, 27-9-58.’’ The rock is from the Flatstones of bed 83 of the Lower Lias succession of Dorset (Lang & Spath, 1926), and contains the well-known ammonites Asteroceras obtusum (J. Sowerby) and Pyomicroceras planicostata (J. Sowerby) which fix its age as the Obtusum Subzone, the lowest division of the Obtusum Zone. The specimen was in a large concretion which fell from the cliff and when the boulder was split it was shattered. Most of the pieces were reassembled, including some of the Counterpart, which, together with the Part, was used in the drawing in Text-fig. I. The flora of the British Lower Liassic is small and the floras of the Middle and Upper Liassic are even smaller. Land plants occur in its extremely limited fresh- water facies (see Kendall, 1949 for West Scotland, Lewarne & Pallot, 1957 and Harris, 1957 for South Wales). In its widespread marine facies, plants, apart from driftwood, seem to be very rare, but a moderate number have been collected over the last century and most of these are magnificent specimens. The flora was de- scribed by Seward in 1904 and since then some additional species have been collected and others have been revised. The flora is as follows (the list may include some from the Rhaetic) : “ Carpolithes sp.”” Seward, 1904. “>? Avaucarites sp.”” Seward, 1904, p. 20. Cycadolepis sp. (as Hippuntes Buckman and indeterminable according to Seward, 1904). Equisetites sp. (as E. muenstert Seward, 1904). Cycadopteris anglica Gothan = Thinnfeldia rhomboidalis of Seward, under revision by Townrow & Hancock. Ctenozamites cycadea (Berger) of this paper = Ctenopteris cycadea of Seward. The previously known specimens are in a soft shale from Lyme Regis. GEOL. 5, 6. 17 162 THE FORM AND STRUCTURE OF CTENOZAMITES CYCADEA Cycadites vectangularis Brauns, see Seward, 1904. Otozamites bechei Brongn. (as O. obtusus L. & H. in Seward, 1904; see also Harris, 1960). Cycadeoidea (Yatesia) gracilis, Seward, 1904. (A Bucklandia.) Cycadeoida pygmaea L. & H. (specimen not seen, see Seward, 1904). Pagiophyllum peregrinum. See Seward, 1904, Kendall, 1948. Strobilites elongata L. & H. (specimen not seen ; see Seward, 1904). Pagiophyllum sewardi Kendall, 1948. In addition there are in the British Museum a leaf resembling Stenopteris, a forked leaf or branch with parallel veined pinnae (or leaves) and some alga-like filaments. In the Geological Survey Museum there are also : “ Zamites megaphyllos Phillips ’’ (a large, simple leaf). “ Cupressinites liasinus’”’ (possibly distinct from Pagiophyllum peregrinum). Several of these fossils, and also the fossil wood, might yield information beyond what has been published. The specimens come from numerous quarries and coastal exposures in the SW. of England and by far the richest region is around Lyme Regis, Dorset, which has provided more than half the specimens. Unfortunately the specimens, apart from those collected by Mr. Jackson, are imperfectly localized being merely labelled “Lyme Regis’. Most of them are in a soft black shale, presumably exposed between the tide marks, but the exact locality is lost to memory and the specimens could be from one or more zones. I spent a few days searching at Lyme Regis and merely found a good deal of drift wood, some of it with structure preserved but no leaf or twig. There were no cuticle fragments in the blocks I macerated. DESCRIPTION Form. The two pieces fitted together are 55 cm. long, but the apex is lost and so is the base of the rachis. It is possible to estimate the length of the missing top from the taper of the half rachises, which is about Imm.in5cm. As the width at the top is 7 mm. the missing part might be 35 cm. long, but rather less if the rachis tapered more quickly in the upper and more leafy part and this would seem likely. The petiole increases in width downwards and is about 2 cm. wide at the point where the rock is broken. There is nothing to suggest how much of the original leaf is missing. Thus the leaf, apart from the missing base, may have been 80 or go cm. long. The lower pinnae are severely abraded and some must have dis- appeared altogether, but those between the forks of the rachis which may have been protected are better preserved. The most complete is 11 cm. long, but the pinnae below the fork have distinctly larger pinnules and were perhaps rather longer. It will be noticed that the pinnae of the right fork are alternate and those of the left nearly opposite. The lowest basiscopic pinnules are borne just below the point of attachment of the pinna, that is, on the main rachis, but the plane of cleavage has not followed the specimen perfectly and only a few are seen. These first pinnules arise on the upper surface of the rachis and the later pinnules arise near the top of THE FORM AND STRUCTURE OF CTENOZAMITES CYCADEA 163 Fic. 1. A, silhouette of the whole specimen, V.36330. There is a gap between the top 4 cm. and the rest. The top is drawn from part and counterpart. Reduced to 1/3 natural size. B, tissues of vascular bundle of rachis in horizontal section, some bundle sheath cells enclose one or two scalariform tracheids, V.36330a. xX 200. C, surface of rachis in horizontal section, showing from the top downwards : cuticle, epidermal cell contents, stone cells flanked by narrow cells, V.36330a. x 200, 164. THE FORM AND STRUCTURE OF CTENOZAMITES CYCADEA the pinna rachis. The pinnules are almost flat but with slightly depressed margins and in none is the margin rolled back. The pinnules have obtuse apices, in none was it mucronate or denticulate. Their veins are clearly seen where the plane of cleavage has passed through them. They diverge slightly and some fork ; main- taining a concentration of about 20 per cm. Anatomy of lamina. The fossil is preserved in an unusual way, intermediate between a compression and a petrefaction. The lamina is only slightly compressed and when buried in the marine calcareous mud seems to have undergone two changes. The first is that most of the cells became filled with a brown or black bituminous substance (there seems too much of it to represent the original cell contents), and second that the spaces between the cells were filled with calcite. The fossil as exposed is a light brown colour when dry. The lamina cannot be picked off the rock, a hard limestone, but chips can be treated with dilute acid when pieces of the lamina, especially its epidermis are freed. They were studied in three ways: 1. The cuticles were prepared. 2. The internal tissues were teased apart after decalcification. 3. The internal tissues were studied in sections. The cuticles were rather difficult to prepare because nitric acid attacks the sub- stance of the cuticle almost as rapidly as the mesophyll cells. Good results were obtained as follows : 1. The decalcified epidermis was soaked in moderately strong nitric acid for 10-60 min. 2. It was washed in dilute ammonia. 3. It was left in glycerine on a slide for several hours, This seems to toughen it. 4. The mesophyll cells were carefully teased away from its inner side. Even after this treatment the cell contents remain and the preparation is of the epidermis rather than purely of the cuticle. The isolation of internal tissues by teasing was only moderately successful. The palisade mesophyll cells are very conspicuous as little dark cylinders representing their contents ; their walls are not seen in teased preparations. The upper cuticle is thick and shows the outlines of polygonal epidermal cells. These cells are almost uniform, the veins being scarcely distinguishable. The outlines of the cells are marked by broad but by no means prominent ridges. The sides of the cells are straight but interrupted by a few pits. The surface walls are flat and the surface may be finely granular or the granules may be elongated and this appearance seems to pass gradually into a system of fine parallel striations orientated longitudinally. Only one trichome base was seen on the upper side (Text-fig. 2F) ; it is like those along the veins on the lower side and consists of a small rounded cell on top of the ordinary epidermal cells. The lower cuticle is also thick. The cells are like those of the upper side, but smaller and the veins are clearly shown by about five rows of elongated rectangular cells and no stomata ; the wide intervenal areas have irregular or isodiametric cells and numerous scattered stomata. The guard cells are mostly longitudinally orientated or more or less oblique, but seldom transverse, They are at the bottom THE FORM AND STRUCTURE OF CTENOZAMITES CYCADEA Vek 165 Fic. 2. Tissues of lamina A, astoma seen from within. A few mesophyll cells remain ; the epidermis at a deeper level of focus is shown by broken lines, V.36330b. x 800. B, imaginary T.S. through A, the vertical distances being determined by focusing. Cuticle in black, other walls outlined, x 800. C, horizontal section through middle of mesophyll showing a vein with sheath and tracheids (right) and the sheath along (left), V.36330c. x 200. D, horizontal section through upper surface of lamina showing from right to left, cuticle, epidermal cell walls, epidermal cell contents, outer layer of palisade, deeper layer of palisade mesophyll, V.36330d. x 200. E, lower cuticle from outside. The content is still present in most epidermal cells. Faint surface striations are shown by dotted lines and the guard cell aperture at a deeper focus by a pair of broken lines, V.36330b. x 800. F, upper cuticle showing a complex trichome base, V.36330e. X 200, 166 THE FORM AND STRUCTURE OF CEENOZAMITES CYCADEA of a polygonal or rectangular pit with a thickly cutinized wall formed by an irregular ring of about six subsidiary cells. The mouth of the pit is covered by a strongly raised dome (almost a hemisphere) of thin cuticle. The top of this dome is pierced by a hole which may be round or oval and is often more or less transversely elongated. Encircling cells may occur but are inconstant. The guard cell surface is thinly cutinized but there is a slightly thickened band by the aperture. The aperture itself is arched, the middle part being more deeply sunken than the ends. The outline of the poles of the guard cells can just be traced, but nothing remains of any lignine thickenings originally present in the guard cells. The epidermis is scarcely compressed and it was possible to recognize differences of level of the stomata by focusing. The top of the dome is about 8 pu above the general surface of the cuticle and the middle part of the stomatal aperture is about 8 uw below this surface. Trichome bases are frequent on the veins and consist of one or occasionally two small oval cells on top of the ordinary cells. Sometimes a ring is visible on the surface of one of these cells but more often the whole cell surface would seem to have been covered by the trichome. No hypodermal cells occur, but fibres along the veins come close to the upper epidermis. ; The mesophyll was studied successfully both by maceration and horizontal sections. The palisade mesophyll is well developed and consists of at least two, probably more, layers of closely packed cylindrical cells 20 ~ wide and 50 yw long. In the middle region of the lamina, at the sides of the veins, the mesophyll cells are oval and orientated transversely. Their walls appear to be thin. In the lower part of the leaf the mesophyll cells are of irregular shapes, but next the lower epidermis they are oval again. There is a gap above each stoma and the mesophyll cells here tend to be rather elongated, about five forming a ring. The veins were seen in sections. They have a sheath of elongated cells with massive dark contents. These cells appear to have transverse ends, but it is difficult to distinguish between a transverse break and a true end. The walls as far as they were seen are only moderately thick and not obviously pitted. This sheath is in contact with the spongy mesophyll and encloses various narrow elongated cells, which have little or no bituminous filling but are merely preserved as poorly seen walls. Most of these elongated cells are narrow and show no obvious features, they presumably are fibres and perhaps sieve tubes. In addition a few small tracheids were seen. They have scalariform thickening. The leaf contains no resin cavities in the mesophyll. There is no sign of any secretory duct in the veins, but unless such a duct had obvious contents it might be hard to recognize in horizontal section. Anatomy of rachis. The rachises are preserved in the same way as the lamina, but underwent far more compression before they were calcified. Thus the main rachis or petiole which is 2 cm. wide is barely I mm. thick in the fossil. Most of its substance has been lost, but a few small chips were obtained when pieces covered by the matrix were exposed. As with the lamina, the plane of cleavage passes through the rachis substance rather than outside it, so each fragment has one THE FORM AND STRUCTURE OF CTENOZAMITES CYCADEA 167 epidermis and more or less of the internal tissues. The broken surface showed elongated cells but nothing else clearly. Good pieces of cutinized epidermis together with isolated cells from the interior were obtained by soaking chips in dilute acid. These represent the various thick- walled tissues seen in the sections, apart from the tracheids which did not survive this treatment. Sections were again cut in a nearly horizontal plane but efforts to obtain more transverse ones gave no useful result. The epidermis is probably more thickly cutinized than that of the lamina and it was easy to obtain a good preparation, though the dark contents of the epidermal cells usually remains. The cells are elongated and have straight, conspicuous walls and the surface is longitudinally striated. No trichomes were seen but stomata are rather frequent and well developed. The subsidiary cell group is elongated, the guard cells longitudinal and the pore in the dome is rather large and also longi- tudinally orientated. A well developed sclerotic layer occurs underneath the epidermis. This layer has at least three sorts of cells. The most conspicuous are thick-walled cells looking like elongated stone cells and with conspicuously pitted walls. Then there are strands of narrow fibre-like cells without much contents. Finally there are short thin-walled cells full of black contents which are regarded as assimilatory parenchyma. Some of these are just under a stoma where they are arranged in a ring to form a substomatal air cavity. In the cells like stone cells the dark content is slightly contracted and forms spikes which are evidently the filling of large pits. The cell wall is not very clearly seen but the pits show that it was thick. The ends of these cells may be more or less transverse, or they may be obtusely pointed. This sclerotic layer is by no means thick, in the slightly oblique section it appears about five times as broad as the epidermis so that it can be inferred that it is about five times as thick. The tissue inside the sclerotic layer which must have formed most of the massive petiole shows several kinds of cells, but most of it is disorganized to a brown humus- like substance. It may well be that this substance was originally parenchyma but the outlines of its cells are not clearly seen. This contrasts with the lamina where the mesophyll cells were filled with dark bituminous matter and very clearly seen, Most of the cells clearly preserved in the interior of the petiole are elongated ones with dark contents and just like those forming the bundle sheath in the lamina. In the horizontal section these cells form numerous longitudinal files and while some of them probably form bundle sheaths, they are so numerous that most are probably scattered through the tissues. Numerous small vascular bundles were recognized by the presence of scalariform tracheids (Text-fig. 2) but as in the lamina these tracheids were poorly preserved. No massive strands of tracheids occur. In a few bundles there are possibly tracheids transitional between scalariform and multi- seriate pitted, that is to say they seem to have thick walls marked with longitudinal rows of small oval pits. The appearance was, however, obscure and it is possible that the cells concerned are of different nature. The bundles include other elongated cells but nothing could be learnt about the arrangement of the cells in these bundles, 168 THE FORM AND STRUCTURE OF CTENOZAMITES CYCADEA DISCUSSION 1. Determination. This specimen, apart from being more complete than usual, agrees in every respect with typical leaves of Ctenozamites cycadea and with no other species. C. cycadea was already known from this locality and if only a small part from the top had been collected it would have been an ordinary specimen. Indeed if no more than one or two pinnules showing their venation had been available, as in Berger’s holotype, it could still have been determined with confidence. The most similar fossils are other species of Ctenozamites, particularly the Oolitic C. leckenbyi where, however, the pinnules are usually longer and acuminate and Piilo- zamites nilssont of the Rhaetic (see Harris, 1932), where the leaf is simply pinnate but small fragments look similar. The cuticle of the present specimen looks just like that of a French specimen figured by Mme. Corsin (1950). It also agrees with the upper cuticles (which alone could be prepared) of the English Liassic leaf, Brit. Mus. (N.H.) no. 40, 674, and with the specimen figured by Saporta (1873, pl. 40, fig. 2) of which Dr. Archangelsky showed me a preparation. The cuticle thus fully confirms the identification. To judge from its pinnules the present specimen was a leaf of average size. Fragments with smaller pinnules have been figured as for example by Schenk (1887) and these might belong to smaller leaves but there are also fragments with dis- tinctly larger pinnules, as for example, Brit. Mus. (N.H.) no. 40, 674, figured by Seward (1904, pl. 3, fig. 1). The detached pinnule which Mme. Corsin (1950) deter- mined as Ctenopteris cf. sarrani may be an unusually large form of Ctenozamites cycadea (with which it is associated) and if so may represent a leaf twice as large as the present specimen. 2. Habitat and age. The present specimen and the two others known from England were found near Lyme Regis, Dorset, in marine Lower Lias. The limestone there is well provided with ammonites, but it also contains very occasional leaves and pieces of wood of land plants. The French specimens figured by Saporta (1873) seem to have a similar origin and so apparently do the German ones of Germar (1847) and Salfeld (1912). This occurrence in marine rocks may be significant even though it also occurs in rocks of fresh water origin (Corsin, 1950). It would be interesting to know more about the provenance of the numerous other specimens described. If this species does prove exceptionally common in marine rocks, it would suggest that the plant grew near the sea, perhaps near the strand of a tidal river, where it would have a special chance of being preserved at the sea bottom. The same may be true of Otozamites bechet (O. obtusus). The present specimen is well dated, being in a matrix full of ammonites, and many others are in rocks zoned by ammonites as Lower Liassic age, as in several parts of Germany (see Salfeld, 1907, 1909). It also occurs in localities regarded as Rhaetic, but some of these belong to the ‘‘ Grenzschichten ”’ which were formerly regarded as Rhaetic, but shown by Gothan (1914) to be Lower Liassic, THE FORM AND STRUCTURE OF CTENOZAMITES CYCADEA t69 The age of others may be unsettled, but I am not aware of any specimen of C. cycadea which occurs in beds with the Rhaetic Pteria contorta. The question is worth investigation because this should be a valuable zone fossil. Specimens agreeing with C. cycadea have been found in the Lower Oolite of Roseberry Topping, Yorkshire. If these are correctly determined the species would extend through the Lias. 3. Nomenclature. Select list of references to Ctenozamites cycadea : 1832 Odontopteris cycadea Berger, p. 23, pl. 3, figs. 2, 3. 1836 Filicites cycadea (Berger) Brongniart, p. 387, pl. 129, figs. 2, 3. 1873 Ctenopteris cycadea (Brongniart) Saporta, p. 355, pl. 40, figs. 2-5, pl. 41, fig. I, 2. 1882 Ctenopteris cycadea (Brongn.) : Staub, p. 249 and plate. 1886 Ctenozamites bergeri (Goeppert) Nathorst, p. 122 (Name). 1887 Ctenozamites cycadea (Brongn.) Schenk, p. 5, pl. 3, fig. r1-16a, pl. 4, fig. 18; pl. 6, fig. 30; pl. 7, fig. 36; pl. 8, fig. 43; pl. 9, fig. 54. (Includes a forked specimen.) 1904 Ctenopteris cycadea (Brongn.) : Seward, p. 36, pl. 3; fig. 1, 1a; probably text-fig. 2. 1950 Ctenopteris cycadea (Brongn.): Corsin, p. 258, pl. 11, fig. 3; pl. 12, figs. 6-11, text-figs. 6, 8. (Perhaps large pinnules and cuticles described as C. cf. savvani should also be included in C. cycadea.) The name used for this leaf by most authors is Ctenopteris cycadea (Brongn.), but both generic name and authority are wrong. The first specimens described, and fortunately reasonably well figured, were those called Odontopteris cycadea by Berger (1832) from the Rhaetic (or Lower Lias) of SW. Germany. Some have suggested that “‘ Filicites agardhiana’’ of Brongniart (1824) belonged to it but both Nathorst (1909) and Antevs (1919) who examined the Swedish specimen agree that it does not, but very probably is a bad specimen of Dictyophyllum nilssom. Possibly the reason why the species has been attributed to Brongniart rather than to Berger is that the first part of Brongniart’s Histoire appeared in 1828, and the fact that Liv. 10 with Filicites cycadea only appeared after Berger’s work is overlooked. Thus Salfeld (1909) gives the F. cycadea reference as 1828, but this is a mistake. Brongniart cites Berger’s reference and makes it quite plain that he took the specific name from Berger. It is not likely that Goeppert added to this confusion by chang- ing the specific name to Bergeri because everyone seems to have known that it was a later substitution. The name Bergeri has been dropped for many years. The generic name Ctenopteris was first used for this fossil by Saporta in 1873. He took the name from a manuscript of Brongniart and published it as Brongniart’s genus, though Saporta would seem to be the responsible author. The question who is the author of Ctenopteris is, however, immaterial because the name is not valid, being an homonym, Cienopteris having been used by Blume in 1828 as a genus of Recent ferns (allied to Grammitis). I am indebted to Dr. R. E. Holttum for pointing this out. Nathorst (1886) instituted the new name Ctenozamites to replace Ctenopteris, not on the grounds that it was a homonym, of which he was unaware, but because he considered it unsuitable. He was convinced that the plant was no fern but some 170 THE FORM AND STRUCTURE OF CTENOZAMITES CYCADEA kind of Cycad and allied to Ptilozamites. Few authors accepted this but continued to use Ctenopteris on the grounds of priority, or because they thought it a fern. However, Schenk (1887) used the combination Ctenozamites cycadea. Nathorst gave no diagnosis nor adequate description but he made it clear that the name was to be used for Odontopteris bergeri (which he recognized as the same as Ctenopteris cycadea) and for C. leckenbyi from the Oolite of Yorkshire. 4. Morphology. The main new fact contributed here is that the rachis forks. Although the specimen has been damaged by the cliff fall or when collected, the continuity of the two branches is fortunately still visible. This forking might have been recognized long ago because one of the specimens figured by Schenk (1887, pl. 3, fig. 13) shows forking, though at a point much nearer the leaf apex. Possibly forking is variable, or, as in Ptilozamites nilssoni, some leaves may fork twice. 5. Anatomy. The new facts here contributed about the anatomy of the lamina and rachis do give some help for comparison, but not much. Their main value lies perhaps in that they prove that plants with structure preserved are to be found in limestone nodules near Lyme Regis, and the information we have should be sufficient to identify another and less compressed petiole when one is found. Search should be made. Apart from its stoma the lamina is unspecialized. It lacks, for example, the strongly developed transfusion parenchyma of many Recent Cycadales and also their hypodermis, but these facts are not ones to which comparative anatomists give weight. The stoma is more useful ; it is haplocheilic as in most Gymnosperms and this distinguishes it sharply from all Bennettitales. The general arrangement of the subsidiary cells and the arched shape of the guard cells agrees closely with the Cycadales but less closely with Conifers or Ginkgo or such few Pteridosperms as we know. The dome of cuticle over the stomatal pit is a peculiar feature shared by a few fossils, Ptilozamites spp. and certain species of Pseudoctenis and not seen in any Recent Cycad though Cycads do not differ fundamentally. The rachis, as far as it is known, may be very like that of a Cycad. The subepidermal sclerotic tissue may be a similar mixture of thick-walled fibres and assimilatory cells but the thick-walled fibres of Cycads are at least in several genera, longer and less freely pitted. In having small vascular bundles with scalariform tracheids it agrees with Cycads. These features are by no means exclusive of Cycadales, for they occur, for example, in the petiole of Medullosa. 6. Systematic position. The four best known species of Ctenozamites make a compact group and it is reasonable to regard them as forming a natural genus and to discuss them together. These four are : Ctenozamites cycadea (Berger) Schenk. Ctenozamites leckenby1 (Leckenby) Nathorst (see Harris, 1943). Ctenozamites sarrani (Zeiller) nov. comb. for Ctenopteris sarrani Zeiller, 1903 : 53, pls. 6-8. Ctenozamites wolfiana (Gothan) nov. comb. for Ctenopteris wolfiana Gothan, 1914: 142, pl. 26, figs. 2, 2a, 5, pl. 37, figs. 5, 6. A dozen other species have been THE FORM AND STRUCTURE OF CTENOZAMITES CYCADEA 171 described but are either poorly known or probably belong to other genera and they are not considered here. C. cycadea and its allies were at first regarded as species of Odontopteris or allied to that genus and some kind of fern, and later by Nathorst as Cycads allied to Piilozamites. It is now generally regarded as a Gymnosperm and by different authors placed more or less close to the Cycadales. Nathorst was chiefly impressed by the leathery texture of C. cycadea and the close resemblance between a pinna fragment of C. cycadea and a piece of the rachis of Ptilozamites nilssom. To these we can add the very close resemblance of their cuticles, particularly their stomata, and the fact that the main rachis in both forks. The only important difference is that the leaf is once pinnate in Ptilozamites but bipinnate in Ctenozamites. We have no evidence to show whether the rachis forks in the other three species of Ctenoza- mites, but in Ptilozamites we are almost certain that it is simple in some species. While the difference between a simply pinnate and a bipinnate leaf is so obvious that it makes a basis for generic grouping of fossil leaves that is too useful to neglect, one may suspect that the plants could all have belonged to one natural genus. We know nothing about the reproductive organs or other parts of the plant bearing Ctenozamites leaves, but we have probably the pollen producing organs of Ptilozamites nilssont. These were described by Harris (1932) under the incorrect determination “ Hydropteridangium ” or Hydropterangium marsilioides Halle and renamed Harrisia marsitloides by Lundblad (1950) and then as Harrisotheciwm marsilioides by Lundblad (1961). It was attributed to P. nilsson: on the evidence of association and agreement in stomatal structure. This fructification is like a loosely organized cone, but its morphology is obscure. The main axis branches in all planes, the branches themselves branch and end in 2-valved capsules. Each valve contains a row of elongated microsporangia with pollen grains, each with two nearly opposite air sacs. If we may accept that Harrisothecium belongs to Ptilozamites its value in classification is negative in that it provides evidence against including the plant in the Cycadales and indeed in any other established group. The capsules look like those of Williamsoma or Cycadeoidea but they differ completely in pollen, and stomata, and in the mode of branching. Ctenozamites (‘‘ Ctenopteris ’’) has been classified as possibly Cycadalean, possibly a Pteridosperm, a Cycadophyte or just a Gymnosperm incertae sedis. There is very little reason for preferring one to another, except that if the fructification is like Harrisothecitum the Cycadales would be excluded and so would the Pteridosperms unless we extend the meaning of that extensive class still further. Whether it is a “ Cycadophyte ”’ is a matter of definition. The term is for many no more than a convenient abbreviation for the phrase “‘ A Mesozoic Gymnosperm with a pinnately constructed leaf and any sort of reproductive organ”’. It includes Cycadales and Bennettitales with totally different reproductive organs and stomata. If ‘‘ Cycado- phyte ”’ is used in this way, Ctenozamites would be placed there readily enough. For other authors “ Cycadophyte ”’ is supposed to be some kind of large family of plants with a relationship that is real even if rather obscure. This is near to its original meaning and if such meaning is held, Ctenozamites should be left zncertae sedis, and indeed the term Cycadophyte should be abandoned. 172 THE FORM AND STRUCTURE OF CTENOZAMITES CYCADEA REFERENCES ANTEVS, E. t1g19. Die liassische Flora des Hoérsandsteins. K. svenska VetenskAkad. Handl., Stockholm, 59, 8: 1-71, pls. 1-6. Bercer, H. A. C. 1832. Die Versteinerungen der Fische und Pflanzen im Sandsteine der Coburger Gegend. 30pp., 4 pls. Coburg. BRONGNIART, AD. 1836. Histoive des Végétaux fossiles. Livr. 11 : 369-416; Livr. 10, pl. 129. Paris & Amsterdam. Corsin, MME Pave. 1950. Découverte de sédiments rhétiens dans le Boulonnais. Etude préliminaire de leur flore et de leur faune. Ann. Soc. géol. Nord, Lille, 70 : 243-272, pls. II-I3. GerMar, E. F. 1847. Jn Dunker, W. Ueber die in dem Lias bei Halberstadt vorkommenden versteinerungen. Palaeontographica, Stuttgart, 1 : 107-125, pls. 12-17. Gotuan, W. 1914. Die unter-liassische (rhatische) Flora der Umgegend von Niirnberg. Abh.. natur. Ges. Nuynberg, 19 : 91-186, pls. 17-39. Harris, T. M. 1932. The Fossil Flora of Scoresby Sound, East Greenland, II. Description of Seed Plants Incervtae Sedis together with a discussion of certain Cycadophyte cuticles. Medd. Gronland, K@benhavn, 85, 3 : 1-114, pls. 1-9. 1932a. The Fossil Flora of Scoresby Sound, East Greenland, III. Caytoniales and Bennettitales. Medd. Gronland, Kobenhavn, 85, 5 : 1-133, pls. 1-19. —— 1943. Notes on the Jurassic Flora of Yorkshire, 4-6. Ann. Mag. Nat. Hist., London (11) 10 : 505-522, 6 figs. —— 1957. A Liasso—Rhaetic flora in South Wales. Proc. Roy. Soc., London (B) 147 : 289- 308. 1960. Otozamites bechei Brongniart from the Rhaetic of Co. Antrim and its comparison with British specimens. [In the Press.] KENDALL, M. W. 1948. On Six Species of Pagiophyllum from the Jurassic of Yorkshire and southern England. Ann. Mag. Nat. Hist., London (12) 1 : 73-108, 12 figs. 1949. On a new Conifer from the Scottish Lias. Ann. Mag. Nat. Hist., London (12) 2 : 299-307, 3 figs. Lane, W. D. & Spatu, L. F. 1926. The Black Marl of Black Ven and Stonebarrow in the Lias of the Dorset Coast. Quart. J. Geol. Soc. Lond., 82 : 144-179, pls. 8-11. LEWARNE, G. & Pattot, J. M. 1957. Mesozoic Plants from fissures in the Carboniferous Limestone of South Wales. Ann. Mag. Nat. Hist., London (12) 10: 72-79, 3 figs. LUNDBLAD, B. 1950. Studies in the Rhaeto-Liassic Floras of Sweden. K. svenska Vetensk- Akad. Handl., Stockholm (4) 1: 1-82, pls. 1-13. 1961. Harrisothecium nomen novum. Taxon, Utrecht, 10, 23-24. NatnHorst, A. G. 1878-1886. Om floran i Skanes kolf6rande bildningar, I. Floran vid Bjuf. Sverig. geol. Unders. Afh., Stockholm (C) 27, 33, 85 : 1-131, pls. 1-26. 1909. Uber die Gattung Nilssona Brongn. mit besonderer beriicksichtigung schwedischer arten. K. svenska VetenskAkad. Handl., Stockholm, 43, 12 : 1-40, pls. 1-8. SALFELD,H. 1907. Fossile Land-Pflanzen der Rat und Jura Formation Stidwestdeutschlands. Palaeontographica, Stuttgart, 54 : 164-203, pls. 14-22. 1909. Beitrage zur Kenntnis jurassischer Pflanzenreste aus Norddeutschland. Palaeonto- graphica, Stuttgart, 56 : 1-36, pls. 1-6. Saporta, G. DE. 1873. Plantes Juvassiques, I. Algues, Equisétacées, Chavacées, Fougéres. In Paléontologie frangaise ou description des fossiles de la France (D’Orbigny). 2e Serie, Végétaux. 506 pp., atlas, pls. 1-60. Paris. ScHENK, A. 1887. Fossile Pflanzen aus der Albourskette. Bibl. Bot., Abh. Gesamtgeb. Bot., Cassel, 6: 1-12, pls. 1-9. SEwarD, A.C. 1904. The Jurassic Flova, II. Liassic and Oolitic Floras of England. Cata- logue of the Mesozoic Plants in the Department of Geology, British Museum (Natural History), 4. xvi-+ 192 pp., 13 pls. London. THE FORM AND STRUCTURE OF CTENOZAMITES CYCADEA 173 Straus, M. 1882. Ctenopteris cycadea, Brngt. in der fossilen Flora Ungarns. Fdéldt. Kézl., Budapest, 12 : 249-256, pl. 1. ZEILLER, R. 1903. Flore fossile des Gites de Charbon du Tonkin. Etudes des Gites Minévaux de la France. viii + 328 pp., atlas (1902), 56 pls. PLATE 31 Ctenozamites cycadea (Berger) Upper part of the specimen. Counterpart of what is shown in Text-fig.1,A. x 1. V.36330. Bull. B.M. (N.H.) Geol. 5, 6 PLATE 31 + eo PLATE 32 Ctenozamites cycadea (Berger) Fic. 1. The rachis dichotomy. I. Fic. 2. The lower part of the rachis showing stumps of two pinnae. X TI. PLATE 32 Bull. B.M. (N.H.) Geol. 5, 6 ae re fit = .RD AND SON, L THE OLD RED SANDSTONE OF BROWN CLEE HILL AND THE ADJACENT AREA I. STRATIGRAPHY H. W. BALL anp D. L. DINELEY II. PALAEONTOLOGY E. I. WHITE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY | Vol. 5 No. 7 LONDON : 1961 THE OLD RED SANDSTONE OF BROWN CLEE HILL AND THE ADJACENT AREA I. STRATIGRAPHY BY HAROLD WILLIAM BALL AND DAVID LAWRENCE DINELEY Pp. 175-242 ; 2 Text-figures, 1 table, 1 map II. PALAEONTOLOGY BY ERROL IVOR WHITE, E.R. Pp. 243-310 ; Pls. 33-48 ; 61 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 5 No. 7 LONDON: 1961 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series, corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 5, No. 7 of the Geological (Palaeontological) series. © Trustees of the British Museum, 1961 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued October, 1961. Price Seventy Shillings fas OLD RED: SANDSTONE OF BROWN, CLEE, HILL AND THE ADJACENT AREA This paper is dedicated to the late William Wickham King PART la. SERATIGRAPLY By HAROLD WILLIAM BALL & DAVID LAWRENCE DINELEY CONTENTS Page I. INTRODUCTION . é 6 é : : : 5 TEG AS) (a) Geographical Setting : : ; : e : tS (6) Historical Review . : 5 : , b a 207h0) (c) Stratigraphical Guceescion : 9 : : ‘ . 180 II. Lower O_p RED SANDSTONE : , : ‘ : : . 180 STRATIGRAPHY 0 2 . ‘ : 0 4 ; LOO (a) Downton Series : : c . : : . I81 (i) Grey Downton Foumation : : : : : aS2 (ii) Red Downton Formation . é : 3 . 5 tS (b) Ditton Series . 5 : : 4 5 Hehe) (i) Main “ Psamnmmosteus ne Limestone : . 189 (ii) Lower group of marls, sandstones and Eorasroues! a eLOr (iii) Upper group of marls, silts, sandstones and cornstones. 193 (c) Clee Series’: j F A : : ; F SELOS) FAUNA AND FLORA . : : 0 5 6 3 é LOO (a) Vertebrates . : : é : - 199 (6) Succession of the Vertebrate Fauna : 0 . ZO (i) Downton Series, Red Downton Bosmation : . eZOL (ii) Ditton Series . : a . 2 2 : eZO2 (c) Invertebrates . : : 6 : 3 ; 5 5) | ZO (d) Plants . : ; : : : : 2 z04 III. Upper OLD RED SANDSTONE ¢ : : é : : ZOOS STRATIGRAPHY , ‘ : F ; ‘ ; . 205 (a) Farlow Sandstone Series : : : : : . 205 (i) Yellow Farlow Sandstone Bormation : A 5 - 205 (ii) Grey Farlow Sandstone Formation . : 5 a Z00 FAUNA . : : . : . : 0 5 AO IV. Lower CARBONIFEROUS : : : : : é : Zoo) (a) Basal Conglomerate A ‘ 5 : : : = 208 V. STRUCTURE : : F . ; : é ; : 7209 (a) Folding . : é < : 5 : : ; 5 Xoo) (6) Faulting 5 : ; : “ : : : 5 PKS) (c) Unconformities : ; 0 c : 5 BQuP (d) Age of the Movements . . ‘ 5 2S VI. ConDITIONS OF DEPOSITION AND PALAEOGEOGRAPHY 5 : a Bt VII. List oF Fossizt Loca.ities, FAUNA AND FLORA : A ; 2tO VIII. ACKNOWLEDGEMENTS . 5 : . é : d : 239 IX. REFERENCES 5 : > “ : - 5 2 240 GEOL 5. 7. 18 178 THE OLD RED SANDSTONE OF BROWN CLEE HILL SYNOPSIS The stratigraphical and faunal succession of the Old Red Sandstone, which forms the greater part of the area about Brown Clee Hill, Shropshire, is described. The Lower Old Red Sandstone forms a conformable sequence which is subdivided into the Downton, Ditton and Clee Series. The base of the Downton Series (Downtonian) is demarcated by the Ludlow Bone Bed, and the Series shows a gradation from predominantly buff and grey sediments (Grey Downton Formation) to purple and red beds (Red Downton Formation) which include local developments of sandstones near the top. The upper part of the Downton Series falls within the zone of Tvaquairaspis symondsi. The Ditton Series (Dittonian) commences with the main “‘ Psammosteus’’ Lime- stone ; it is subdivided into a lower group composed largely of red and purple marls with sand- stones and cornstones, broadly equating with the zones of Ptevaspis (Simoptevaspis) leathensis and Ptevaspis (Belgicaspis) crouchi, and an upper group in which the marls are largely replaced by silts, with an increase in the proportion of sandstones and cornstones, and a change to pre- dominantly red-brown, brown-buff and grey-green coloration. The upper group corresponds to the zone of Ptevaspis (Cymripteraspis) leacht. The Clee Series (Breconian) consists largely of unfossiliferous grey, buff, brown and purple sandstones and silts, the base of the Series being formed by the lower Abdon Limestone. This sequence is unconformably overlain by the Upper Old Red Sandstone comprising the Farlow Sandstone Series (Farlovian), which is subdivided into a lower (Yellow) and an upper (Grey) Formation. A further unconformity occurs between the Farlow Sandstone Series and the basal conglomerate of the Carboniferous. An account is also given of the structure and geological history of the area. The fauna is discussed and five new genera, one new subgenus and fifteen new species are described and figured. I. INTRODUCTION (a) Geographical Setting DomINATING the area to be described is Brown Clee Hill which has an “ hour-glass ”’ shape and is elongated in a N.-S. direction. The hill is largely composed of beds of Breconian age folded into a syncline, the axis of which extends NNE.—-SSW. thus diverging slightly from the hill’s topographical axis. A thin series of Coal Measures unconformably overlies the Breconian and is in turn capped by the remnants of dolerite sheets! at Abdon Burf and Clee Burf, the northern and southern eminences respectively, the former at 1,790 ft. O. D. being the highest point in the Midlands. Brown Clee Hill is connected by relatively high saddles with Titterstone Clee Hill to the south and Weston Hill to the south-west, whilst west and north extends the high level platform of Ditton Priors. This platform is formed by the Ditton Series and bounded by the prominent escarpment of the sandstones associated with the ‘“‘ Psammosteus’’ Limestones. Below the escarpment to the west lies Corvedale, incised into the less resistant strata of the Downton Series and drained to the south-west by the River Corve ; to the south-east a smaller complementary valley has been cut by the Ledwyche Brook. 1 A contact between the top of the dolerite and the base of the overlying carbonaceous shales, exposed in the old quarry at Clee Burf, showed the shales to be baked to a depth of 2 in., supporting Marshall’s (1942) interpretation of the intrusive form of the Clee dolerites as opposed to Pocock’s (1931) extrusion hypothesis. THE OLD RED SANDSTONE OF BROWN CLEE HILL 179 To the east, between Brown Clee and the Coal Measures of the Wyre Forest, a broad outcrop of Ditton Series gives rise to more undulating topography, the cornstone and sandstone horizons characteristic of this Series forming numerous minor ridges and escarpments. The southern margin of the area is demarcated by the scarp slope of the Cornbrook Sandstone which forms the north flank of Titterstone Clee Hill, and by the ridge of the Farlow Sandstone Series and the overlying Carboniferous Limestone extending eastwards from Farlow. The drainage of the area to the east of Brown Clee Hill centres upon the River Rea which flows to the south-south-east, breaching the Farlow ridge at Prescott and ultimately joining the River Teme, as does its western neighbour the Corve. (b) Historical Review Despite the relatively numerous references to the geology of the area, most of our knowledge is based upon the researches of two authors. The first of these was Murchison, who gave the fullest account in his epic Silurian System (1839), subse- quent versions of which appeared in the several editions of his Siluria. It was almost a century later that the other major contributions appeared by W. Wickham King (1925, 1934) who established the first detailed succession based on lithological characteristics. More recently, White (1950a ; White & Toombs, 1948) has suggested a modifica- tion of this succession to conform with the results yielded by his valuable researches into the ostracoderm faunas of the Old Red Sandstone of the Welsh Borderland and South Wales. Other contributions to the vertebrate palaeontology of the area have been those of Egerton (1862), Stensid (1932, 1948), A. S. Woodward (1934) and Wills (1935). The area was first systematically mapped by Robert and Romley Wright during the early part of the last century, as noted by Murchison (1839 : 112) who also appended a map and sections to the Silurian System. These were followed by the Geological Survey during the period 1850-55, the results being published in the Old Series I-in. scale maps, sheets 55 NW. and NE., and 61 SW. and SE. The eastern extremity of the area is included on the New Series r-in. scale map, sheet 167, and some interesting conclusions regarding the subdivision of the Lower Old Red Sandstone are included in the Memoir on the sheet (Whitehead & Pocock, 1947). The present account, based on mapping on a scale of 6 in. to a mile and in some instances 25 in. to a mile, originated as two independent pieces of research, one author working upwards from the Downtonian, the other working downward from the Farlovian, both meeting in the common ground of the Dittonian. Thus one of us (D. L. D.) is entirely responsible for observations on the succession ranging upwards from the base of the Downton Series to the ‘‘ Psammosteus ’’ Limestones, the other (H. W. B.) for the succession upwards from the base of the Clee Series. The description of the Ditton Series above the “‘ Psammosteus ’’ Limestones is a combined account. A brief summary of these researches has been published in an excursion guide to the area (Ball & Dineley, 1952). 180 THE OLD RED SANDSTONE OF BROWN CLEE HILL (c) Stratigraphical Succession LowErR CARBONIFEROUS Basal Conglomerate 6-40 ft. unconformity Upper OLD RED SANDSTONE Farlow Sandstone Series 230-510 ft. Grey Farlow Sandstone Formation 40-300 ft. Yellow Farlow Sandstone Formation 190-210 ft. unconformity Lower OLD RED SANDSTONE Clee Series ooo ft. Ditton Series 1,200-1,450 ft. Upper group of marls, silts, sandstones and corn- stones 400-550 ft. Lower group of marls, sandstones and cornstones 800-900 ft. Main “ Psammosteus ’’ Limestone o-16 ft. Downton Series : 1,150-1,250 ft. Red Downton Formation 1,070—I1,100 ft. Grey Downton Formation (with Ludlow Bone Bed at base) 80-150 ft. Ii LOWER OLD RED) SANDSDONE STRATIGRAPHY The delimitation of the subdivisions of the Lower Old Red Sandstone sequence in the Anglo-Welsh area has long been a source of controversy, and also its junction with the Silurian (see White, 1950a). As originally defined by Lapworth (1879-80), the Downtonian comprised the Upper Ludlow Shales, the Ludlow Bone Bed and the Downton Castle Sandstone. Subsequently, the term was used by several authors to embrace differing groups of strata. However, the most generally accepted classifi- cation of the Old Red Sandstone has been that of W. Wickham King who divided it into lithological stages. Utilizing Lapworth’s term “ Downtonian”’ as defined by Peach & Horne (1899 : 568), King applied it to a succession of purple marls, shales and sandstones overlying the Temeside Shales (1917 : 97, 98). He later (1925) ex- tended it to embrace a great thickness of predominantly marly beds, and referred to it in his table of succession (p. 383) as the “‘ Downtonian or Anaspida marls’’, though it is not clear from his table whether he regards the Ludlow Bone Bed as forming the base of the “ so-called Old Red Sandstones’”’. Subsequently in the same paper a further name, “‘ Downtonian Series ’’, was introduced for the same beds. Similarly, the succeeding strata were referred to in the table of succession as the “‘ Dittonian or Ptevaspis Cornstones ”’ and later as the “‘ Dittonian Series ’’, the term “‘ Dittonian ’’ having been introduced by King in two earlier papers (19214, 0). The junction between the Downtonian and Dittonian was designated as the base of the “‘ Cephalaspis Sandstone—cornstone’’, occurring up to 300 ft. above the ‘‘ Psammosteus’’ Limestones. Above the Dittonian, “‘ Brownstones’”’ were tentatively recorded, with the “ Farlow Sandstones ”’ unconformably overlying the Dittonian and the Brownstones. King (1934) retained the terms ‘“‘ Downtonian ’’ and “ Dittonian ’’, discarding the THE OLD RED SANDSTONE OF BROWN CLEE HILL 181 alternative names, although in his table of succession (p. 527) he referred to the “ Downtonian Marls’’. Both the Downtonian and Dittonian were regarded as the upper part of the Silurian, the Ludlow Bone Bed being included with the Upper Ludlow. A more recent summary of the Old Red Sandstone succession based largely on King appears in Wills (1948), which notes the ‘‘ Brownstones ”’ as unconformably overlying the Dittonian, though this is considerably emended in a later edition (Wills, 1950). The practice of the Geological Survey regarding the subdivision of the Old Red Sandstone has been varied and is summarized by White (1950a@ : fig. 2). Most recently, and with special reference to this area, Whitehead & Pocock (1947) placed the Downtonian-Dittonian boundary at the base of the “‘ Psammosteus ’’ Limestones, retaining the Downtonian with the Silurian. The beds below the “ Psammosteus ”’ Limestones were given the stratigraphical name ‘“‘ Downton Series’’, and those above and including the limestones the name “ Ditton Series ’’. From his researches into the fish faunas of the Lower Old Red Sandstones of the Anglo-Welsh area, White (1950a ; White & Toombs, 1948) arrived at similar con- clusions, though differing from Whitehead & Pocock on one very important point, namely, the inclusion of the Downton Series with the Old Red Sandstone. He re- garded the Dittonian as being marked by the appearance of Ptevaspis, approximately at the level of the “‘ Psammosteus ’’ Limestones. In Corvedale, one of us (D. L. D.) has found that there is one main limestone horizon which, though showing some lateral variation, can be traced through- out the area. Therefore, with regard to drawing a well-defined marker horizon, the acceptance of the “‘ Psammosteus’’ Limestones as the Downtonian-Dittonian boundary, as proposed by Whitehead & Pocock (1947), is an obvious solution. . However, since Ptevaspis (Simopteraspis) leathensis White has now been found to occur below the limestones, the boundary does not coincide with that of White’s faunal zonation. Nevertheless, it has been amply demonstrated in many fields of palaeontology that faunal assemblages and maxima are more reliable indices than the appearance and disappearance of individual species. Moreover, for this area, the correlation between the faunal and stratigraphical horizons is suggested as being sufficiently close for them to be regarded as broadly contemporaneous. In the present work the stratigraphical names (i.e. time-rock units) ‘‘ Downton Series”’ and “ Ditton Series’’ are retained, whilst the ‘“‘ Downtonian’’ and “ Dittonian ’”’ are regarded as denoting the equivalent epochs (time units) within the Lower Old Red Sandstone. The highest part of the Lower Old Red Sandstone sequence occurring in this area is formed by the Clee Series, representing the Breconian epoch. (a) Downton Series The Downton Series comprises a sequence of 1,150 to 1,250 ft. of predominantly red and purple marls, with thin lenticular sandstones and pellet beds distributed irregularly throughout its middle and upper parts. The Ludlow Bone Bed is here accepted as the base of the Series, whilst the succeeding Ditton Series is regarded as commencing with the main “‘ Psammosteus’’’ Limestone (see below). The Down- ton Series is subdivided into two groups, the lower Grey Downton Formation, 182 THE OLD RED SANDSTONE OF BROWN CLEE HILL corresponding to the Temeside Group of Robertson (1927), and the succeeding Red Downton Formation, equivalent to the ‘‘ Red Downtonian beds ”’ of Whitehead & Pocock (1947 : 4). (i) Grey Downton Formation This formation consists of olive, grey and buff flags and shales with occasional purple marls, and is transitional between the shales of the Upper Ludlovian and the marls of the Red Downton Formation. In the Ludlow area the subdivision of this facies into Ludlow Bone Bed, Downton Castle Sandstones and Temeside Beds can be clearly demonstrated (Elles & Slater, 1906), but it is not possible to maintain this subdivision with certainty in Corvedale where the exposures are few and poor. The Ludlow Bone Bed in Corvedale has been mapped by Robertson (1927). No new exposures have been found, whilst some of those examined by Robertson no longer exist and there is no certainty that they all form part of a constant, single stratum ; indeed, Elles & Slater and Robertson recorded more than one bone bed horizon between Ludlow and Much Wenlock. It has not been possible to distinguish the Downton Bone Bed. Whilst there is no sharp division of the Grey Downton Formation into the members recognized at Ludlow or Much Wenlock, the same general sequence occurs. There is a progressive transition from grey calcareous to red facies : Lithologies Ludlow Corvedale Much Wenlock Grey, olive, purple and chocolate mott- 120 ft. é Generally : 50 ft. led shales and mudstones less than go ft. Thin-bedded pale sandstones. Massive yellow sandstones. Downton Bone 30-50 ft. 5 30-40 ft. : 30 ft. Bed (local). Olive sandy shales. Lud- low Bone Bed Despite the lack of exposures, it is apparent that thick sandstones are few and impersistent and that the facies change occurs within a maximum thickness of 170 ft. in the south, the thickness decreasing northwards, at least as far as Shipton. From the area around Norton to Diddlebury, the formation decreases from at least 150 ft. to go ft. in thickness. Near Hungerford and Broadstone, an estimated roo ft. of olive shales and chocolate-purple mudstones lie below dark red marls, some of the beds being extremely micaceous. Nothing can be added to previous accounts of the palaeontology. (ii) Red Downton Formation The red marls and occasional impersistent sandstones constituting this formation outcrop in the low land of Corvedale, the Ledwyche Valley and to the north of Monkhopton and Morville. There is a continuous upward transition and the fossils are rare and confined to highly localized concentrations. In view of this, the limits of the formation are based largely upon lithology. The junction of the Red Downton Formation with the Grey is drawn where the first thick red marls appear and is thus somewhat arbitrarily fixed between exposures in western Corvedale. As White (19504 : 54) has suggested, the lowest parts of the red facies may well be THE OLD RED SANDSTONE OF BROWN CLEE HILL 183 coeval in places with part of the grey elsewhere. The main “ Psammosteus ’’ Lime- stone (see below) appears to form a mappable horizon about 1,100 ft. above the base of the Red Downton Formation. The marls greatly resemble those of the Trias and are usually extremely fine- grained, red, chocolate or bluish purple, with frequent green-stained, ramifying bands, spots and joint-planes. Petrographically they are identical with those of the Red Marl Group in the Cardiff district described by Heard & Davies (1924 : 502). Usually there is little trace of bedding or lamination, but the cuboidal jointing on weathered surfaces, characteristic of Triassic marls, commonly occurs. Some horizons, how- ever, include finely laminated, micaceous siltstones. Also associated with the marls are thin bands of small, closely packed, calcareous concretions (‘‘ race ’’), which sometimes take the form of thin, irregular, vertical rods, reaching 4 in. or so in height and 4 in. in diameter. Thin, lenticular sandstones and pellet beds are irregularly distributed throughout the marl sequence, especially near the base and locally in the middle and upper parts of the formation. In particular, two sandstone developments, the “‘ Holdgate coarse sandstone’”’ and the “ Ischnacanthus Sandstone ’’, were defined by King (1925 : 384; 1934 : 527) as “‘stages”’ of his Downtonian, occurring 815 ft. and about 1,230 ft. respectively above the Ludlow Bone Bed. Subdivision based upon lithology is, however, impracticable in these variable sediments and the scarcity of fossils makes palaeontological zonation equally difficult. Nevertheless, the lowest sandstones of the formation may be equivalent to the “‘ Thyestes (Auchenaspis) Sandstone’ of Ledbury, whilst those at the top are indistinguishable from the Dittonian sandstones. The sandstones are lenticular and vary considerably in colour and grade. They consist largely of subangular or angular quartz grains, with varying amounts of mica, feldspar, accessory heavy minerals and clay minerals with ferruginous or calcareous cements, the latter often showing lustre-mottling. Pellets and flakes of marl are frequently included, sometimes in such profusion that the sandstone grades into a pellet rock. With an increase in calcium carbonate the pellet rocks may in turn grade into a cornstone. Near the base of the “ Psammo- steus ’’ Limestones escarpment several very coarse grits are known, but no pebble beds of the type described by Heard & Davies (1924 : 495) have been found. Opinions as to what constitutes a “cornstone’”’ vary widely (see Allen, 1960), but as stated by Murchison (1839 : 55, footnote 2), ‘‘In the country ... of the Old Red Sandstone, the name of ‘ Cornstone’ is restricted to the coarse, sandy, conglomerate-like masses, and is never applied to the large concretions of purer limestone”’ ; and again a little further on (p. 180), “ while at intermediate places [the cornstones] consist of marl, limestone, and sandstone, irregularly con- creted, and have the aspect of a conglomerate. In the last-mentioned form alone, they constitute the cornstones of the inhabitants ...”’. Thus, throughout this paper the term will be employed in this context. Cornstones are most conspicuous in the uppermost Red Downton Formation and the Ditton Series. Generally they are from 1 to 4 ft. thick, though reaching as much as 14 ft. in the Ditton Series. They vary widely in lithology, ranging from relatively pure to gritty and marly limestones, and may be red, purple, buff, grey-green, green 184 THE OLD RED SANDSTONE OF BROWN CLEE HILL or variegated in colour, according to the form and amount of the iron salts present. They are largely composed of fragments of muddy limestone or calcareous marl, commonly rounded or sub-rounded, though occasionally sub-angular, ranging in size from a few millimetres up to several centimetres. Some cornstones consist mainly Fic. r. Current directions as indicated by dips recorded from foreset beds in sandstone and conglomeratic cornstone units within the upper part of the Downton Series and the lowest 400 ft. of the Ditton Series, measured on the northern and western slopes of Brown Clee Hull. of one type of pellet or of water-sorted “ race ’’, whilst others also incorporate pebbles of vein quartz, cherts, jasper, and pellets of vegetable carbon. The matrices consist of crystalline and muddy calcite, marl and sand in varying proportions and, when weathered, the beds may assume a nodular or rubbly appearance. Fragmentary THE OLD RED SANDSTONE OF BROWN CLEE HILL 185 fish remains frequently occur, sometimes in considerable local concentrations. The calcareous bodies ascribed to algae by Heard & Davies (1924 : 505) in the Cardiff district, have not been found in the present area. Many of these beds rest upon highly irregular erosion surfaces, sculptured into the marls to a depth of several inches. Cross-bedding, in which the foreset beds are generally less than 1 ft. thick, is common and some of the thicker cornstones show successively overlapping, short, irregular, cross-bedded lenticles between 1 and 4 ft. thick. A few strikingly cross- bedded cornstones appear as “‘ wash-outs ”’, cutting locally through the underlying beds to a depth of 3 to 4 ft. From the measurement of some 300 separate foreset bed dips within the uppermost Downtonian and the Dittonian strata, it would seem that the dominant direction of transport was from the north-west (Text-fig. 1), though in any one section the directions may be reversed several times. Oscillation ripple-marks are common in many sandstones, the crest intervals being usually between 3 and 5 in., though smaller ones are known and shallow marks with crest intervals of 9 in. or more occur upon the upper surface of some cornstones. Current ripple-marks with crest intervals of about 23 in. have also been found. Small desiccation cracks are present below a number of sandstones, whilst in the Lye Brook and on Titterstone Clee Hill very large ones occur. Vertical sandstone- filled fissures, some 2 ft. deep and 4 in. wide, similar to those noted by Dixon (1921 : 29) in Pembrokeshire, penetrate the upper parts of cornstones at Aston Eyre and near The Alders (Text-fig. 2). In the Fishmore Brick Pit, Ludlow, in beds which are probably low in the Red Downton Formation, several cross-bedded cornstones are exposed, each about 18 in. thick and 15 ft. long and consisting of poorly sorted, rounded marl pellets and “race’’. They rest upon sharp, irregular marl surfaces and appear to represent gravel-filled runnels cut into the marl shortly after its deposition. Desiccation cracks and rain-pit markings occur in a sandstone band, and near the top of the brick-pit section a conspicuous sandstone wedges out sharply, showing near its termination small circular contortions of the bedding, possibly resulting from the slumping of the sediment in a channel incised in the marl (Text-fig. 2). Near the base of the formation in western Corvedale, scattered, thin, lenticular, red and purple sandstones may represent the “‘ Ledbury Beds’ of Worcestershire, but have not yielded any diagnostic fossils. King appears to have established the presence of beds yielding Thyestes (Auchenaspis) egertoni (Lankester) to the north of Morville (Whitehead & Pocock, 1947 : 22). Six feet of closely alternating purple marls and fine ferruginous sandstones, temporarily exposed at Little Thonglands, near Holdgate, greatly resembled the “ Ledbury Beds’ of Ledbury (Piper, 1898), and pale red and green micaceous sandstones 400 yds. south-east of Brockton have yielded acanthodian spines. King’s stage I. 4, ““ Holdgate Sandstones ’’, was stated by him (1925 : 384; 1934: 527) to occur at 815 to 830 ft. above the Ludlow Bone Bed. Subsequently, the term “ Holdgate Sandstones Group ”’ was retained for the sandstones near the middle of the Red Downton Formation (Ball & Dineley, 1952). However, there is little to distinguish these sandstones from others within the formation and it is perhaps best to use the term in its purely local connotation, i.e. only for the sandstones in the STRUCTURES in sandstones and conglomeratic corn stones . Lenticular units, slum- ping of sand down sides of small channels . “ Spring-pits ”’ . Ripple-markings, asym- metric c : - Ripple-markings, sym- metric . Desiccation-cracks . ‘“ Wash-outs ”’, scoured and filled channels in sandstone-cornstone units . “ Rolling ” or “ wavy ”’ bedding in _ sand- stones ; undersurface may show tendency towards “ flow-cast- ing” . Cross-bedding, foreset beds at Low angles to surface of unit: units with little or no discordance between them - Undersurface, possibly of relief, with non- linear features . Sandstone with “ sand- stone dykes’”’ pene- trating bed below . Cross-bedding, foreset beds at high angles to surface of unit. Units may be very ir- regular in thickness, foreset dips differ in LOcALITIES where structures were noted (List not exhaustive) Ludlow 513754, Stoke St. Mil- borough 574811. Tugford 565872. Bouldon 549841, 548840, Hayton 525815, Monkhopton 631926. Chetton 665901, Monkhopton 633922, Neenton 651871, Silving- ton 607798, Stanton Lacy 509802, 524778. Bouldon 541840, Ludlow 513754, Morville 672928, Tugford 565869. Bouldon 552854, Hopton Cangeford 530805, Monkhopton 633943, Tug- ford 555863, Upton Cressett 648932. Bouldon 549850, Ditton Priors 606906, Monkhopton 634928, Neenton 653895, Stoke St. Milborough 512816, Upton Cres- sett 661922. Ditton Priors 586887, Hopton Cangeford 549799, Monkhopton 625926, Silvington 623801, Stan- ton Lacy 509802. Ditton Priors 612904, Middleton Priors 622907, Morville 672929, Stanton Lacy 519778, Tugford 555862. Aston Eyre 643932, Stoke St. Mil- borough 562802. Bouldon 546848, Ditton Priors 610905, Holdgate 574889, Hopton Cangeford 530805, Monkhopton 631927, Morville 671928, Stanton Lacy 510801, Tugford 568877. successive units, base marked by erosion surface with sharp relief ; flutings and scour-marks may show preferred orient- ation Fic. 2. Sedimentary features found in the Downtonian and Dittonian sandstones and conglomeratic cornstones of the Clee Hills area. Argillaceous beds left unstippled, vertical scale in feet. — THE OLD RED SANDSTONE OF BROWN CLEE HILL 187 vicinity of Holdgate village, where the principal sandstone horizon is about 650 ft. above the Ludlow Bone Bed, Sandstone lenses are recorded at many localities throughout central and northern Corvedale and in the country near Acton Round, at a few localities in southern Corvedale, and in the Ledwyche Valley. They are variable in grade and colour and many are very similar to the sandstone from “I. 5 ”’ at Eastham brickworks, near Tenbury, Worcestershire, described by Walder (1941 : 245). The individual lenses are rarely thicker than 15 ft., and are separated by varying thicknesses of marl. These beds are best developed near Acton Round where relatively coarse sand- stones, intercalated with red and green marls, form a distinct feature. They are probably also present north and east of Morville where much of the solid geology is obscured by drift. Between Acton Round and Morville the sandstones are distri- buted throughout some 400 ft. of strata, the highest occurring 280 ft. below the main ‘‘ Psammosteus’”’ Limestone at Monkhopton. The Beaconhill Brook section shows the typically lenticular nature of the sandstones, many resting upon erosion surfaces cut into green marl. The individual beds are thin and seldom more than 50 yds. long, their basal layers usually containing water-rolled flakes of marl. Strong cross-bedding is general and suggests a predominantly northern derivation. A few local concentrations of fish fragments occur and include Kallostrakon and acanthodian remains. Westwards to Skimblescott the sandstones decrease in number, whilst between Skimblescott and Stanton Long only a few very thin friable sandstones have been found. At Holdgate a marked N.S. scarp is formed by a coarse sandstone or group of sandstones, about 18 ft. thick, the top bed being some 650 ft. above the Ludlow Bone Bed. Further sandstones are developed at this horizon at Baucott, Culmington and Stanton Lacy. In the Ledwyche Valley dark, mottled sandstones occur near Fishmore Hall, Warthill and Brook House. The uppermost roo ft. of the Red Downton Formation include further coarse beds, localized in groups near the foot of the ‘‘ Psammosteus ’’ Limestones escarp- ment. Sandstones and cornstones here alternate with marls and are so similar to the Dittonian beds that they may be regarded as locally developed precursors to those of the Ditton Series. These beds are well exposed in Lye Brook, Morville, where some 100 ft. of strata occur below the limestone, the lowest conspicuous bed being a cross-bedded, white, pellety grit yielding Tvaquaivaspis symondsi (Lankester). Above this, lenticular groups of thin cornstones and sandstones are exposed and occupy about 15% of the succession. Several prominent sandstone and cornstone bands can be traced along Meadowley Hill and Aston Hill, some of which have yielded Pteraspis (Simopteraspis) leathensis White; but, with the exception of Lye Brook 4, their exact relationship to the main “‘ Psammosteus’’ Limestone is not certain since some of them are undoubtedly displaced by large-scale hill-creep and may be Dit- tonian. The cornstone near Yewtree Dingle (121)! yielding Pteraspis (Simopteraspis) leathensis White and other fossils, mentioned by White (1950 : 74, 75), is placed by King just below the ‘‘ Psammosteus”” Limestone and by Pocock (in lit.) 50 ft. above “a Psammosteus Limestone ’’. However, each is probably referring to a different limestone, 1 The numbers in parentheses after localities correspond with those given in the List of Fossil Localities and on the map 188 THE OLD RED SANDSTONE OF BROWN CLEE HILL and it seems likely that the lower is the main limestone (see below). A coarse white grit similar to that in Lye Brook outcrops about 100 ft. below the only ‘“‘ Psam- mosteus’’’ Limestone at Stapely Dingle, but the sections in Hudwick and Sudford Dingles, Monkhopton, show only a few thin impersistent sandstones in the red marls below the main “‘ Psammosteus ’’ Limestone. A thin red cornstone 25 ft. below the main limestone in Foxhole Coppice, Monkhopton (31), contains Kallostrakon and acanthodian remains. Even in closely adjacent sections on the escarpment between Monkhopton and Tugford the successions differ, comparable sequences occurring in two only. The fossiliferous cornstone and underlying marl and “ race ’’ outcropping in Earnstrey Brook (25. See Wills, 1935) are exactly like the beds at the foot of the section near Little Oxenbold (51), two miles to the north-east and both horizons yield a rich fauna. Exposures between the two localities show red and green marls with occa- sional coarser beds. The sporadic outcrop of cornstones suggests restricted accumu- lation, perhaps in a stream channel. At Bouldon rapidly alternating red and green shales and thin-bedded argillaceous sandstones and cornstones lie below the main ‘‘ Psammosteus’’ Limestone. South and east of Bouldon only a few scattered, lenticular, barren sandstones and corn- stones have been found near the top of the Red Downton Formation. (b) Ditton Series The Ditton Series comprises a sequence of closely alternating marls, silts, lenticular sandstones and cornstones, delimited by the main “‘ Psammosteus’’ Limestone at the base and the lower Abdon Limestone above. However, as is pointed out below (p. 198), it may be necessary at some future date to redefine the top of the Ditton Series, since the uppermost 150 ft. may be coeval with the lower part of the Senni Beds (i.e. Breconian) of the Black Mountains, as defined by Croft on the basis of their flora. The Series can be broadly subdivided into two subsidiary groups, though owing to the lack of a mappable marker horizon the boundary between them is arbi- trary. The lower group is lithologically similar to the uppermost Downton Series and principally bright red and green in colour. The upper group is more varied and like the lower part of the succeeding Clee Series, chocolate, buff-grey and orange tints being common, whilst many of the beds are strikingly coarse-grained. Thus, the following lithological subdivisions of the Ditton Series are recognized in the Brown Clee Hill area : (iii) Upper group of marls, silts, sandstones and cornstones 400-550 ft. Ditton Series (ii) Lower group of marls, sandstones and cornstones 800-900 ft. (i) Main “‘ Psammosteus ’’ Limestone o-16 ft. Total 1,200-1,450 ft. During the mapping of the area, local names were erected for convenience and applied to the lower and upper groups of the Ditton Series ; ‘‘ Bouldon Beds ”’ for the lower THE OLD RED SANDSTONE OF BROWN CLEE HILL 189 and ‘‘ Wheathill Beds’”’ for the upper. It is not proposed that these should be retained. However, the term “ Wheathill Beds ”’ was incorporated by Wills (1950 : 24, 30) in a classification of the Old Red Sandstone, though assigned by him to the Breconian owing to the alleged presence of “‘ Rhinoptevaspis dunensis’’. This was due to uncertainty over the status of Ptevaspis (Cymripteraspis) leachi originally described by White (1938) and hitherto recognized as a variety of Pteraspis (Rhin- opteraspis) dunensis (s.s.), but which must now be treated as a separate species (White, 1950a@ : 56) and sub-genus (White, 1960 : 8). Thus, Pteraspis (Rhinopteraspis) dunensts (s.s.) has not as yet been recorded in the Clee Hills area. The Ditton Series is best exposed in the several stream courses which drain west- wards across the Ditton platform, especially where they cut into the ““ Psammosteus ”’ Limestones escarpment. But even here the sequences are discontinuous, the upper part of the lower group and the upper group being poorly exposed. Hence it has proved impracticable to define a type section. The lowest beds of the lower group include a high proportion of sandstones and cornstones which give rise to a prominent escarpment, the so-called “‘ Psammosteus ’’ Limestones escarpment. This well-marked feature extends westwards from Underton, in the north-west of the area, along the eastern edge of Corvedale, to Tar Grove on the south-western margin of the Ledwyche Valley. The upper beds of the lower group include a larger amount of marls and form the broadly undulating surface which constitutes the greater part of the Ditton platform. The upper group incorporates a higher proportion of coarser beds which form the steeper footslopes of Brown Clee Hill and Weston Hill. To the north-west of Brown Clee Hill the Ditton Series has a thickness of 1,200 ft., comprising approximately 800 ft. of the lower group and 400 ft. of the upper. However, to the south-east of the hill the Series thickens to 1,450 ft., consisting of some goo ft. of the lower group, which also includes more coarser beds in its upper part, and 550 ft. of the upper group. Around the head and eastern margins of the Ledwyche Valley, the “ Psammosteus’’ Limestones escarpment merges with the generally steeper slopes resulting from the inclusion of a higher proportion of coarser beds in the lower part of the Series. Between Brown Clee Hill and the Coal Measures of the Wyre Forest, delimited in the north by the ‘‘ Psammosteus ’’ Limestones escarpment and in the south by the northern slopes of Titterstone Clee Hill and the Farlow-Walton ridge, the Ditton Series forms a broad area of rolling country. Extending north-eastwards around Brown Clee Hill, the outcrops of the subdivisions of the Series expand rapidly, the lower group in general outcropping north of a line from Cleobury North to Deuxhill, the upper group largely occupying the area to the south of this line. (i) Main “‘ Psammosteus ”’ Limestone Since the earliest studies of the Anglo-Welsh Old Red Sandstone were made, the presence of a group of widely distributed, distinctive limestones, the ‘‘ Psammosteus ”’ Limestones, lying immediately above a thick marl division, has been known. Several workers have used them as marker horizons, Whitehead & Pocock (1947 : 7, 8) de- fining them as demarcating the base of the Ditton Series in the north-east of the area. 190 THE OLD RED SANDSTONE OF BROWN CLEE HILL McCullough (1870 : 35) noted that one apparently continuous limestone could be traced for some distance in the Pontrilas district. More recently White (1946, 1950a) has discussed the derivation, validity and continued use of the name “‘ Psammosteus Limestones ”’ and has suggested that they may be diachronous. The general charac- teristics of their lithology have been described (recent notes have been added by White, 1946; Dineley, 1951 and im Ball & Dineley, 1952), but a detailed petro- graphical study is still greatly needed. Fleet (1m King, 1925 : 385) demonstrated that they are relatively pure limestones and that their inorganic nature seems to be sub- stantiated by the complete absence of fossils. The distribution of these limestones throughout the Brown Clee Hill area has been mapped with two aims in view, to ascertain (a) their reliability as marker horizons, and (6) whether they are diachronous. More than one limestone is exposed in several sections on the “‘ Psammosteus ’’ Limestones escarpment, but the principal horizon is usually 5 ft. or more thick, reaching 16 ft. at The Hope. The other limestones are generally much thinner, impersistent and nodular, and also more argillaceous and less compact. Where the position of the main limestone was un- certain it has proved possible to locate it with a ‘‘ Megger ”’ earth-tester, following the technique used by D. W. Gossage whilst mapping the ‘‘ Psammosteus ’’ Lime- stones near Cleobury Mortimer (Dineley & Gossage, 1959 : 227-229). Elsewhere small excavations have been made to determine the position of the limestone, and at only a few points has no limestone been detected. The “ Psammosteus’’ Lime- stone outcrop given on the map is here regarded as the main limestone. Two limestones are recorded in a small inlier in Borle Brook, near Criddon Bridge (Pocock & Whitehead, 1947 : 22) and between Underton and Lye Brook, the upper being only poorly developed and continuing westwards to Moor Dingle, rising from 400 ft. O.D. at Underton to 480 ft. at The Lye. The lower limestone outcrops at a number of points along the escarpment between Meadowly and Moor Dingle and attains a thickness of 10 ft. near Monkhopton, but in Hudwick Dingle it wedges out and reappears within a few yards. Both nodular and massive limestone occurs in Sudford Dingle. From Foxhole Coppice to Oxenbold Coppice there is only one lime- stone, infrequently exposed, but detected by the “‘ Megger ’’ earth-tester and main- taining a constant level at about 700 ft. O.D. No limestone has been detected in Oxenbold Coppice, but from its southern end a thick limestone can be followed southwards between 500 and 600 ft. O.D. into the Ledwyche Valley, rising north- eastwards to 650 ft. near Gibbridge where it is faulted down go ft. to the east. A little further to the east, it is downthrown some 150 ft. to the south-east by the Brown Clee Fault, but rises to nearly 800 ft. at Langley, its easternmost outcrop in the valley. South-westwards from Stantongate the limestone reaches an elevation of goo ft. O.D. on the western flank of Titterstone Clee Hill, then descends to 700 ft. at Bitterly, a mile to the south. Only one limestone appears to be persistent throughout the Brown Clee Hill area. It is irregular in thickness and locally absent, like a sheet with holes and thin patches. This bed, referred to as the main “‘ Psammosteus’’ Limestone, is here regarded as an important marker horizon, defining the base of the Ditton Series. Thinner, imper- sistent limestones are locally, but not extensively, developed. THE OLD RED SANDSTONE OF BROWN CLEE HILL 191 (ii) Lower Group of Marls, Sandstones and Cornstones The best exposures in the lower group of the Ditton Series occur in its lower part, particularly along the streams cutting through the “‘ Psammosteus’’ Limestones escarpment. One of the most continuous sections occurs in the stream draining from Cockshutford to Tugford, where approximately 60% of the exposed strata consists of marls, though the true proportion of these is probably nearer to King’s estimate of 70% (1934 : 529). The marls in the lower part of the group are pre- dominantly red and purple in colour, the latter generally being highly micaceous, with numerous layers and lenticles of green marl. Silty marls also occur, particularly in the upper part of the group, and tend to be duller and slightly browner. Both types frequently contain ramifying pipes and veins of purple marl, and erosion sur- faces are common where marls are overlain by sandstones. The marls also incorporate lenticles of ‘‘ race’ and variegated, concretionary limestones. Near the top of the group two thin bands of limestone, lithologically identical with the Abdon Limestones, occur near New Earnstrey Park. The sandstones of the lower group of the Ditton Series are almost invariably fine- grained, though as pointed out by King (1934 : 528) they are generally coarser and much less micaceous than those of the Downton Series. The sandstones are princi- pally red- and purple-brown in colour, with many green bands, and are frequently calcareous and hard, sometimes incorporating sandy, calcareous nodules. They are characteristically current-bedded and frequently show ripple-marked bedding planes, whilst desiccation-cracks, (?) animal tracks, and other trace fossils have also been recorded. Sole markings occur in many of the sandstones and are clearly of more than one type. Most conspicuous are the impressions of erosion surfaces, with small washouts and sandstone piping, but possible post-depositional structures also occur. A few thin bands of coarse grit occur, particularly in the “ Psammosteus’’ Lime- stones escarpment. Associated with the sandstone bands are lenses of cornstones, commonly composed of pellets of red-brown marl, and red, grey and yellow-buff marly limestones set in a marly, silty or occasionally sandy, calcareous matrix. The size of the pellets varies widely, typically between 0-5 and 1-5 cm. in length, though ranging up to 8-0 cm. Some of the cornstones incorporate angular or sub-angular pebbles of vein quartz, and many include fragments and scales of ostracoderms. Lenses of green cornstones also occur, usually less calcareous than the preceding cornstones, being composed of pellets of green and olive marl, sometimes calcareous, in a green marl or sand matrix. Such cornstones are generally found associated with green sandstones and frequently incorporate small quartz pebbles and carbonized plant remains. Passing up the sequence, the lower group of the Ditton Series becomes duller and more silty, and the relative proportion of marls increases. King’s ‘‘Cephalaspis Sandstones ”’, present in the Trimpley inlier, cannot be distinguished in the Brown Clee area, and no marker horizons have been detected in the Ditton Series above the main ““ Psammosteus’’ Limestone. Thus, though both the upper and lower groups have distinctive faunas and in general the upper group is marked by an increase in the amount of sandstones and cornstones, the junction between them is arbitrary. GEOL. 5. 7. 19 192 THE OLD RED SANDSTONE OF BROWN CLEE HILL The best section in the lower part of the lower group occurs in Hudwick Dingle where sandstones and cornstones constitute about 52% of the measured succession of 220 ft., a higher proportion than in other nearby sections. The junction of marls with overlying sandstones is here usually sharp, frequently uneven, sometimes consisting of an erosion surface beneath which the marls are stained green to a depth of 2 to 3 ft. Cornstones are developed at the base of many sandstones, which are as a whole much less micaceous than the Downtonian sand- stones. Cross-bedding is common and oscillation ripple-marks are found in a number of the higher beds. No desiccation-cracks have yet been found in the succession in Hudwick Dingle, but in the lower part bands of “‘ race’ and thin argillaceous lime- stones occur. A thick cornstone immediately above the main “‘ Psammosteus ”’ Limestone (37) has yielded a large fauna, including Tvaquairaspis symondsi and Pteraspis (Simopteraspis) leathensis. The latter also occurs at an horizon some 8o ft. higher (39), whilst Pteraspis (Pteraspis) rostrata var. indet. is found at 175 ft. above the limestone (40). Coarse beds constitute about 43% of the strike section, vertical thickness roo ft., occurring above the main “‘ Psammosteus’’ Limestone in Borle Brook between Upton Cressett and Wallsbatch, and lying in the zone of P. (S.) leathensis. Several thin sandstones and cornstones are exposed in Rea Brook, which are disrupted by small faults and thrown into small sharp folds, in the vicinity of Middleton Priors. The outcrop of thick sandstones and cornstones in the lower part of the Ditton Series on the western side of Brown Clee Hill, gives rise to well-defined local features, particularly at Bouldon and Sutton Hill. Stream sections along the “ Psammosteus ”’ Limestones escarpment between 700 and 800 ft. O.D. expose sequences of rapidly alternating marls, sandstones and cornstones, dipping south-eastwards at 5 to Io degrees, the marls forming about 60% of the total. Many of the cornstones are extremely coarse containing pellets up to 10 cm. in length. A good section outcrops in Kidnall Gutter, Tugford, where the main ‘‘ Psammosteus’’ Limestone is par- ticularly well exposed. The Kidnall Gutter section is noteworthy for having yielded P. (P.) rostrata var. indet. from a cornstone 15 ft. above the main “‘ Psammosteus ”’ Limestone (44) with P. (S.) leathensis occurring a further 15 ft. above (45). The stream bifurcates in its upper reaches and the contrast between the two resultant sections illustrates the rapidity of lateral change within the beds. This is also well shown in Oak Dingle (75), Tugford, where a green cornstone-sandstone, estimated to lie about 125 ft. above the main “‘ Psammosteus ’’ Limestone, has yielded Pteraspis (Pteraspis) rostrata var. trimpleyensis, Traquaivaspis symondsi and Tesseraspis. Several small structures resembling spring-pits (Shrock, 1948 : 136) occur in a sand- stone at the foot of the section. The sections in Clee and Abdon Brooks are poor, the scattered exposures largely consisting of red and green marls with a few green sandstones and cornstones. Old quarries on the escarpment near Bouldon expose sections in buff and olive shales and sandstones and extremely coarse cornstones. A cornstone exposed in Clee Brook at Bouldon (13) about 50 ft. above the main ‘‘ Psammosteus’”’ Limestone, has yielded Pteraspis (Pteraspis) rostrata var. trimpleyensis, whilst a very coarse corn- stone outcropping in the Old Forge Quarry (77), Bouldon, is possibly the same as THE OLD RED SANDSTONE OF BROWN CLEE HILL 193 that occurring 24 ft. above the main “ Psammosteus’’ Limestone in Strand Brook, three-quarters of a mile to the south. The quarry at Hoptongate (36) shows an exceptional sequence of 8 ft. of green, argillaceous cornstone in thin, successive, cross-bedded lenticles, many of which have truncated upper surfaces. The fine section in Tar Grove Quarry (105), east of Whitbatch and about 175 ft. above the main ‘“‘ Psammosteus ’’ Limestone, has yielded P. (P.) rostrata var. trimpleyensis and T. symondsi from the same stratum, this being the highest record of the latter. The middle and upper parts of the lower group are poorly and intermittently exposed in stream sections to the west of Brown Clee Hill, and in Ledwyche Brook and Newton Dingle near whose confluence thick sandstones and cornstones occur. The beds are sparsely fossiliferous, but have yielded P. (B.) croucht, P. (P.) rostrata, P. (P.) dairydinglensis and arctolepids. Bright red and green marls with thin cross- bedded sandstones and cornstones of the lower group outcrop in the “ saddle’ be- tween Bromdon and Titterstone Clee Hill and extend eastwards into the Silvington Valley where they have yielded P. (B.) crouchi. There appears to be considerable faulting in this region and the exact thickness of the lower part of the Ditton Series here is uncertain. The lower group of the Ditton Series equates with the leathensis and croucht Zones. (iii) Upper Group of Marls, Silts, Sandstones and Cornstones The transition from the lower group to the upper is marked by a sharp break in fauna, though further intensive collecting may prove that the faunal dichotomy is not so distinct as that implied by present evidence. Furthermore, there is an increase in the proportion of sandstones and cornstones, and also in the coarseness of grade, marls being replaced to a large extent by silts and silty shales, whilst the upper beds become red-brown brown-buff and grey-green in colour. The trend in increasing coarseness of grade upwards in the succession, shown by the Downton Series and the lower group of the Ditton Series, is continued in the upper group of the latter. Thus, in the lower part of the upper group, marls and silts predominate over sandstones and cornstones. However, the proportion of the latter increases in the upper part and yellow- and brown-buff silts become increasingly common. The marls of the upper group are predominantly chocolate and green in colour ; the silts are red-brown, buff, grey and green, frequently laminated, and often in- corporate sandy, calcareous concretions, nodules of argillaceous limestone and purple “ piping ’”’. The sandstones are largely fine-grained or silty and predominantly red-brown, grey, chocolate and pale-buff or green in colour. Some of the sandstones are calcareous, hard and thickly bedded, whilst others are laminated and silty, incorporating scattered marl pellets and sandy, calcareous concretions. Current- bedding is common and many of the sandstones are markedly lenticular. Cornstones are most common in the upper group and show considerable variation in lithology. They frequently have a yellow-buff hue, though they are also red, green and grey, being composed of pellets of yellow-, orange- and chocolate-buff, red-brown, grey and purple calcareous marl and argillaceous limestone. The matrices vary from argil- laceous or sandy limestone, which has sometimes become recrystallized, to marl, 194 THE OLD RED SANDSTONE OF BROWN CLEE HILL when the boundaries of individual pellets are frequently difficult to distinguish. The pellets are largely sub-rounded to sub-angular and average I to 2 cm. in size, though ranging up to 7 cm. The cornstones are commonly current-bedded and well sorted, the pellets in any given band being of approximately the same size. Some consist of rapidly alternating well- and poorly-sorted lenses, the latter usually being much more marly than the former. The green cornstones are frequently very argillaceous and rubbly, and incorporate much carbonaceous plant debris. A few thin, grey, concretionary limestones occur which are lithologically similar to the “ Psammosteus ’’ Limestones. The outcrop of the upper group of the Ditton Series occupies the broad and undu- lating area to the east of Brown Clee Hill, approximately between Upper Overton and the Farlow-Walton ridge, and forms the footslopes of Brown Clee Hill and Weston Hill. Exposures in the group are few and poor, being either short, discon- tinuous sections along stream courses and road cuttings, or small isolated quarries. The best section occurs along Bensons Brook, Titterstone Clee Hill, which lies outside the area being described ; but within the area, the most complete section is exposed in Newton Dingle on the southern slopes of Brown Clee Hill. In Newton Dingle, in the area of The Gore, a succession of marls, laminated siltstones and cornstones out- crops, two of the cornstones having yielded Pteraspis (Cymripteraspis) leacht White and, in one instance, Protaspis (Europrotaspis) crenulata sp. nov., whilst plant fragments occur at several horizons in laminated siltstones and sandstones. North of The Gore, the sequence is less well exposed but includes purple, green and buff, silty sandstones and siltstones, with a few bands of rubbly cornstone. The thick bands of cornstone and sandstone, characteristic of the group, give rise to prominent features such as The Thrift, the col which links Brown Clee Hill with Weston Hill, and the succession of small escarpments in the Wheathill-Silvington area. The scarp slopes face north-west and an outcrop near the top of the slope occurring 1,200 yds. north-east of Silvington (93) reveals a massive cornstone, some 14 ft. thick, dipping to the south-east. Sections in sandstones and cornstones also occur in road cuttings at Clee St. Margaret, Stoke St. Milbrough, Wheathill and north of Farlow Bank. Cornstones at several localities have yielded the characteristic fauna of Pteraspis (Cymripteraspis) leachi and Protaspis (Europrotaspis) crenulata, with numerous arctolepid fragments, and a notable locality (8) near to Besom Farm has yielded Benneviaspis salopiensis sp. nov., the youngest cephalaspid yet recorded in this area (see p. 282; also Ball & Dineley im Whittard, 1953: 249, 250). Many of the laminated siltstones have yielded plant fragments, though these are poorly preserved and largely indeterminable. However, the late W. N. Croft was able to recognize Prototaxites, Pachytheca, leaf-like patches of Nematothallus and slender branched axes, cf. Cooksonia, from an assemblage collected from a small outcrop in silty sandstones at Dodshill Bank (23), half a mile to the south of Besom Farm. The top of the upper group is exposed in a small quarry 200 yds. north of Walton (x14), where the lower Abdon Limestone, which forms the base of the Clee Series, is interbedded with some 15 ft. of sandstones and cornstones. The sandstones include grey and chocolate-grey, massive, hard bands and also buff, silty beds ; whilst the THE OLD RED SANDSTONE OF BROWN CLEE HILL 195 cornstones are red- and green-grey in colour and very gritty, that immediately below the Abdon Limestone having yielded to King a fragment of an arctolepid and an acanthodian scale, the highest record of ostracoderms in the area. The upper group of the Ditton Series equates with the zone of Pleraspis (Cymrip- teraspis) leachi White. (c) Clee Series The base of the Clee Series is here defined as the bottom of the lower of two major limestone horizons which, with the intervening strata, form a prominent feature around the north-western slopes of Brown Clee Hill. This lower limestone horizon is a fairly consistent and mappable feature, and below it the beds are, in general, more typically Dittonian in aspect than those above which rapidly assume the characteristic Clee Series lithology. The upper limestone is also relatively persistent and lies some 200 ft. above the lower. Since both limestones are especially well developed in the area of Abdon where they have been extensively quarried in the past, King has called them (7m Jit.) the Abdon Limestones, and this name will be retained here. Both horizons show rapid lateral and vertical variations in lithology, from massive to nodular limestones, and calcareous marls somewhat like the “race’’ of the Downton and Ditton Series. The massive limestones are generally dark brown and brown-purple in colour when fresh and often include a few angular grains of quartz averaging 0-2 mm. in size, with rarer grains of plagioclase feldspar, microcline, tourmaline, zircon, garnet, muscovite, cherts and “ felsite’’. With an increase in arenaceous material the limestone passes into calcareous and frequently lustre- mottled sandstone. Variations of the massive limestone facies also occur. Much of it appears to have been deposited as a relatively homogeneous calcareous mud showing in thin section signs of movement in a plastic state. Another common variation consists of angular, brown, grey and brown-purple limestone fragments ranging up to 10 cm. in size, set in a like matrix, possibly originating from the pene-contempor- aneous breaking up of thin sheets of calcareous mud with rapid re-cementation. The third variation appears to be concretionary in origin (cf. the ‘‘ Psammosteus ’’ Lime- stones) and sometimes distinctive problematical structures (p. 198) are found asso- ciated with it. With an increase of marl and silt the concretionary limestones pass into the nodular facies and ultimately into a calcareous marl, both commonly variegated. Concomitant with rapid changes in facies, the limestones vary in thickness. The greatest development appears to be along the lower slopes of Brown Clee Hill above Abdon, where the lower limestone is at least ro ft. thick, Murchison (1833 : 473) recording a thickness of 18 to 20 ft. The outcrop is marked by a succession of small quarries (Murchison, 1839 : 179, 180) now largely overgrown, which extend round the foot of the hill to a point near Hillside Farm, where the limestone is very nodular and marly, and interbedded with lenses of calcareous marl and red-brown calcareous sandstone, or white, lustre-mottled sandstone. To the south of Abdon Liberty the outcrop of the lower horizon is marked by a number of smaller workings which be- 196 THE OLD RED SANDSTONE OF BROWN CLEE HILL come more sporadic towards Cockshutford, where the limestone is frequently re- placed by calcareous sandstones, often white and lustre-mottled. South of Nordy- bank it appears to be intermittently replaced by calcareous sandstones as far as Pel Beggar, where the limestone becomes re-established in its nodular facies and can be traced eastwards to the south of Bockleton Court. To the south-west of Pel Beggar, a small outlier of the lower limestone caps Weston Hill. At Bockleton Court both horizons are displaced to the south by the Brown Clee Fault, but owing to solifluction the outcrop of the lower is difficult to trace east of the fault, though it is in part marked by overgrown workings and outcrops in the track leading to Short- wood. Nowhere beyond this point along the eastern side of the hill has the lower limestone been detected. Thus the base of the Clee Series on the eastern slopes of Brown Clee Hill has been extrapolated and is wholly conjectural. The upper limestone, whilst not reaching the thickness of the lower at its maximum, undergoes similar variations in thickness and lithology, and is much more persistent. At Abdon Liberty and along the northern slopes of the hill, the upper limestone forms a prominent rib above the quarries in the lower, and has itself been quarried on a small scale. The trace of the outcrop becomes less pronounced on the north- eastern slopes of the hill, where the limestone becomes more nodular and marly, as can be seen in the old workings at Cleobury North Liberty. South of Abdon Liberty the outcrop is marked by further small workings, whilst between Cockshutford and The Sands the upper limestone has been quarried on a scale commensurate with that of the lower limestone workings on the northern slopes of the hill. The best exposure in the Abdon Limestones occurs in the upper horizon at Cockshutford where it has been exposed by recent quarrying. The upper limestone outcrops in small workings to the north of Bockleton Court where it is displaced to the south by the Brown Clee Fault. East of Bockleton Court it forms a prominent feature which extends round the south-eastern slopes of the hill to Banbury, having been extensively quarried near The Toot. Like the lower limestone, the upper has not been detected on the eastern side of Brown Clee Hill between Banbury and Cleobury North Liberty. The strata between the Abdon Limestones consists largely of sandstones and silty shales, predominantly red-brown at the lower levels, becoming increasingly buff and grey in colour upwards, which with the limestones form a distinct feature at the foot of the northern and western slopes of Brown Clee Hill. The sandstones are composed dominantly of quartz, both igneous and metamorphic, the grains being largely sub- angular to angular but also sub-rounded and occasionally rounded. The grains range in size from 0-or mm. (fine silt) to o-r mm. (coarse sand) but average between o-I and 0-4 mm. (fine to medium sand) ; most of the sandstones being moderately well sorted. Also incorporated are frequent grains of plagioclase feldspar in the albite- oligoclase range, myrmekite, perthite and microcline, and fragments of a variety of rock types including “ felsite ’’, basalt, volcanic glasses, quartz schist, cherts and calcareous marls. The feldspars are largely fresh but some are undergoing sericiti- zation. Accessory minerals include worn zircons, white mica, chlorite, tourmaline, garnet and epidote. Cementation is largely by calcite or ferruginous mud. In some of the calcareous sandstones, the component grains do not form a mesh and were probably deposited by lime-rich waters as a sandy, calcareous mud which has under- THE OLD RED SANDSTONE OF BROWN CLEE HILL 197 gone subsequent recrystallization. Some of the quartz grains have been strained and fractured during recrystallization, whilst others have been partially replaced by calcite. Intercalated with the sandstones and silts are lenses of massive and nodular limestones identical with the Abdon Limestones, and variegated calcareous marls. Succeeding the Abdon Limestones are some 700 ft. of strata consisting largely of sandstones and silts (estimated 85 to 90%) and forming the greater part of the Clee Series. The sandstones are commonly current-bedded and predominantly grey, buff, brown and purple in colour, and though similar in composition to those between the limestones they tend to be coarser, with grain size averaging 0-4 to 0-75 mm. The succession as a whole becomes coarser upwards, with the appearance in the upper part of coarse sandstones and grits. A large number of the sandstones incorporate pellets and pebbles of variegated calcareous marls, many of which cannot be matched with the marl bands intercalated with the sandstones. The pellets are typically between I and 6 cm. in size, but range up to 16 cm., and occur throughout the succession. In contrast, pebbles of vein quartz, yellow and liver-coloured quartzite and cherts are especially common in the upper part of the Clee Series. A derived Silurian polyzoan, Fistulipora sp., was found as a pebble in the lower part of the Clee Series at the head of Batch Gutter, 320 yds. north-west of The Toot, but it has not proved possible to assign it to an horizon within the Silurian. Though predominantly sub-angular, some rounded pebbles do occur, and these are usually less than 3 cm. in size. However, like the constituent grains of the sandstones, the pebbles increase in size upwards in the succession, one pebble of 22 cm. having been found in the upper part of the Clee Series. A few bands are entirely composed of pellets and fragments of grey and green calcareous marl with small pebbles of vein quartz, and are comparable with the corn- stones of the Downton and Ditton Series. Intercalated with the sandstones are occasional lenses of silty marl, usually purple and green in colour, and variegated, calcareous, nodular marls, some of which grade into nodular limestones similar to the nodular facies of the Abdon Limestones. Rapid lateral facies variation and lensing-out of beds is characteristic of the group as a whole, though many of the strongly developed sandstones in the upper part of the Clee Series are more persistent and form a succession of well-marked features along the western slopes of Brown Clee Hill. The Clee Series is best exposed along the streams draining the western and south- western slopes of Brown Clee Hill. The most continuous section occurs along the stream course extending north-westwards from Clee Burf to Cockshutford, whilst other relatively good, but more broken sections, can be seen along the streams flowing westwards from Five Springs and south-westwards from Warren Cottage. Numerous small exposures occur on the northern, western and southern slopes of the hill, but are almost entirely lacking on the east. The full development of the Clee Series is not seen since it is sharply truncated by the plane of the Coal Measures unconformity, but its maximum thickness as exposed on the western slopes of Brown Clee Hill is calculated to be goo ft. A further small outcrop of the Clee Series occurs as a downfaulted wedge at Walton, in the south-east of the district. A limestone some 3 to 4 ft. thick outcrops in a small 198 THE OLD RED SANDSTONE OR BROWN CELEB HITE quarry 200 yds. to the north of Walton, (114). It is identical with the massive facies of the Abdon Limestones and containing, in addition, the same distinctive proble- matical markings. The line of outcrop of a similar limestone was revealed by deep ploughing in fields to the south of Walton. Since the latter is calculated to be some 200 ft. above the first limestone it seems probable that they represent the Abdon Limestones. The associated sandstones and silts show a typical Clee Series lithology. The Clee Series has, as yet, yielded no indigenous fauna or flora. However, as has been stated above, certain problematical markings appear to be confined to the Clee Series, occurring in the Abdon Limestones, in some of the sandstones between them, and continuing at least into the middle of the Series. The markings are identical with those first noted by Croft (Pocock, Brammall & Croft, 1940 : 55) as “ pepper-pot ”’ markings in the Senni Beds and subsequently recorded by him as “ distinctive problematical structures ”’ confined to these beds and occurring through- out the Black Mountains and neighbouring areas (Croft, 1953 : 431). Despite this absence of diagnostic fauna or flora from the Clee Series, there are strong strati- graphical grounds for correlating them with the Senni Beds of South Wales. Both show similar gradational changes from the Dittonian strata which they overlie con- formably (Croft, 1953 : 429), with an increase in the proportion of sandstones and change in colour from predominantly reds and purples to greens and greys. More- over King has long recognized the close similarity between the Abdon Limestones and the Fynnon Limestones of the Black Mountains. In both areas there are two major limestone horizons separated by some 200 ft. of strata, and all are lithologically alike (King im Uit.). It should be noted that the base of the Clee as defined above does not correspond with that of the Senni Beds as defined by Croft (7m lit.), which is at the bottom of a prominent group of sandstones containing a typical Senni Bed flora (Croft & Lang, 1942) extending a short distance below the lower Fynnon Limestone. As yet, the equivalent strata occurring below the Abdon Limestones have not yielded a flora, but if the correlation between the Abdon and Fynnon Limestones can be substan- tiated and they are not diachronous, it will be necessary to re-define the top of the Ditton Series. Similarly it has been impossible to draw an upper limit to the Clee Series corresponding with the Senni Beds-Brownstones boundary in the Brecon Beacons. Thus it is expedient to retain for the Shropshire Breconian strata the local name of Clee Series until a more exact correlation is possible. The presence of true Brownstones, namely, strata succeeding the Senni Beds and underlying the Farlovian (Croft, 1953 : 429), cannot be proved in the Shropshire area. But since Croft gives a thickness of 850 ft. for the Senni Beds in Breconshire, and in view of the general correlation between the Clee Series and the Senni Beds, the Brownstones may well be represented by the highest beds of the Clee Series on Brown Clee Hill. The ‘‘ Brownstones ”’ previously recorded in the area (King, 1925 : 383, 388; Edmunds & Oakley, 1947 : 27; Whitehead & Pocock, 1947 : 23, 24; Pocock & Whitehead, 1948 : 66; Wills, 1948, 1950) have proved to be largely of Upper Dittonian age and, in part, lower Breconian. White (1956 : 5) has suggested that the Senni Beds are probably of Middle to Upper Siegenian age, which is there- fore true for the Clee Series. THE OLD RED SANDSTONE OF BROWN CLEE HILL 199 FAUNA AND FLORA The Downton and Ditton Series are largely barren, though there are a few rela- tively fossiliferous horizons, particularly the cornstones and sandstones associated with the ‘‘ Psammosteus’’ Limestones. The fossils are frequently concentrated into lenticles, sometimes forming veritable bone beds. The major part of the fauna consists of ostracoderms, invertebrates being rare, especially in the Ditton Series. The associated flora is poorly preserved. With the possible exception of the prob- lematical structures they contain, the Clee Series appear to be entirely barren. (a) Vertebrates The Lower Old Red Sandstone of the Clee Hills areahas long been known for its ostra- coderm fauna, notable collections having been made in the last century by G. Cocking, Rev. J. F. Crouch, Dr. J. Harley, Rev. T. T. Lewis, R. Lightbody, Dr. T. Lloyd, A. Marston and H. Salwey ; and more recently by Dr. E. I. White, H. A. Toombs, Prof. L. J. Wills, and especially the late W. Wickham King. The localities yielding ostracoderm remains are recorded on the map and are listed below. Many of the classic localities which yielded rich faunas in the past now appear to be com- pletely barren, e.g. Whitbach, near Ludlow, which was re-excavated for the 1948 International Geological Congress Excursion C.16, and yielded not a single fragment. The number of forms present in the fauna is large, and the writers are greatly in- debted to Dr. E. I. White for the faunal determinations and for his description of the new forms which it contains. The delicate and laborious preparation of the material has been undertaken by Mr. H. A. Toombs utilizing the acetic acid technique (Toombs & Rixon, 1959). This has revealed incidentally that some of the cornstones incorporate a rich and diversified micro-fauna. Mode of Occurrence of the Vertebrate Fauna In his description of the remarkable faunule at Wayne Herbert, Herefordshire, White (1935) was among the first to appreciate the necessity of carefully evaluating the mode of occurrence of ostracoderm faunas. His later studies also emphasized the peculiar nature of the distribution and composition of these faunas, and he suggested (1950a@ : 57, 58) that the vertebrate faunas (faunules) are spasmodic introductions from a plexus of streams emptying into a common basin. This is substantiated by the occurrences in the Clee Hills area, thus : I. the ostracoderm remains are usually concentrated into small lenticles ; 2. the fossils most frequently occur in cornstones which rest on erosion surfaces and show signs of rapid deposition ; 3. most of the remains are fragmentary, complete and undamaged carapaces being very rare ; 4. the remains are frequently concentrated into residues of fragments of a similar size, buoyancy or weight; 5. the composition of each faunule tends to be distinctive and unique. As noted above, the cornstones are generally the most fossiliferous, the sandstones being less so, whilst at only two localities have ostracoderms been found in marls. 200 THE OLD RED SANDSTONE, OF BROWN CLEE HILL The largest and most complete remains are frequently found in the bottom few inches of a cornstone, though they are often distorted and damaged by compression and compaction. The base of a conglomeratic, coarse, white grit at Lye Brook, which rests upon an erosion surface, has yielded the following fragments of Tvaqairaspis symondst : Dorsal shields. é . 4Telatively complete. 4 large fragments. Dorsal spines. 9 . 26 fragments. Ventral shields . 5 . 13 relatively complete. 17 large fragments. Lateral plates . ¢ . 13 relatively complete. Orbital plates . : . 12 relatively complete. Rostrum . : : : I incomplete. ? Branchio-cornual plate . 3 incomplete. ? Pineal : : . . I incomplete. Scales (flank) . é 8 2h, A number of small fragments were also incorporated, as well as Tessevaspis, Kallo- stvakon, and acanthodian spines. In the finer sediments above, only very rare small fragments occur. At least 26 individuals are represented, but only 8 dorsal shields and I rostrum are incorporated. It is probable that this selective concentration has been effected by current action. At Netchwood Common (62), 36 rostra of Ptevaspis (Belgicaspis) croucht were collected from about 8 sq. ft. of a cornstone 8 to Io in. thick, though an even more remarkable concentration of plates has been recorded by White (1938 : 102) in Carmarthenshire. Wills (1948 : 28) and White (1950 : 73) commented upon a cornstone slab from Morville which incorporated 30 dorsal and 3 ventral shields of Ptervaspis (Simopter- aspis) leathensis. The lithology of the cornstones suggests rapid deposition and supports White’s hypothesis of a fluviatile deposition rather than Wills’ suggestion of accumulation in a pool. White further notes that a more or less uniform orientation of plates may occur on a bedding plane, which also suggests transport in, and by, continuously moving water currents. Most of the larger pteraspid fragments are of dorsal or ventral shields ; branchial and other paired plates may occur intact or broken, yet still attached to the shield, and where detached they are usually imperfectly preserved. The fragments usually show little sign of post-mortem attrition, and where abrasion is present it appears to have occurred during life on those parts of the plates which would be expected to come into contact with the river bed. No complete specimens as recorded by White (1935 : 383) at Wayne Herbert, Herefordshire, have been found. There, the evidence suggested essentially quiet deposition, possibly in a pool. Ostracoderm remains occurring in the finer sediments of the Clee Hills area are rare and usually fragmentary, the fossiliferous marls at Earnstrey Brook (25) and Little Oxenbold (51) being exceptional. , A remarkable feature of the traquairaspid and pteraspid faunules is that each is composed of individuals which are at approximately the same stage of development though they may show some degree of variation in form. Moreover, only rarely is THE OLD RED SANDSTONE OF BROWN CLEE HILL 201 one species of Ptevaspis found in association with another, which may imply that the species occupied different ecological zones (see p. 252). Utilizing the pteraspid faunas, White (1950, 1956) has erected a zonal sequence which has enabled the correlation of the Lower Old Red Sandstone of the Anglo- Welsh area with that of the continent. The present research has facilitated the more accurate definition of the limits of these zones within the type area, and the estab- lishment of characteristic zonal assemblages. However, in view of the range within the type area of P. vostvata and also of its confused taxonomic status (see p. 247), White’s zone of crouchi and rostrata is here referred to as the zone of croucht. More- over, it should be noted that throughout this work where reference is made to Pteraspis (Pteraspis) rostrata Agassiz, this does not indicate that the specimen re- ferred to is rostrata s.s., but instead, vostvata var. indet. (b) Succession of the Vertebrate Fauna (i) Downton Series, Red Downton Formation Zone of Tvaquairaspis symondsi (Lankester) Of the small number of fossiliferous localities found in the Red Downton Forma- tion, the great majority occur within the top hundred feet, being associated with the sandstones forming the lower slopes of the ““ Psammosteus ’’ Limestones escarpment. White (1946 : 214 ; 1950: 55; White & Toombs, 1948 : 5) has shown that the upper part of the Downtonian can be subdivided into two zones, a lower with Traquatraspis pococki and an upper with 7. symondsi. However, since the two species occur to- gether near Clifton-on-Teme, and 7. symondsi has been found roo ft. below, and T. pococki 17 ft. below what is considered to be the same limestone horizon at Gardners Bank (Dineley & Gossage, 1959 : 232), it is evident that the two forms overlap. Their mutual exclusiveness within the zone of overlap (with the exception of Clifton-on- Teme) may indicate, as in the pteraspids, that they occupied different environments. Only T. symondsi has been found in the Brown Clee Hill area. It occurs at five localities, the lowest, Lye Brook (55), being 125 ft. below the main “ Psammosteus ”’ Limestone, which broadly accords with White’s (1950a@ : 55) suggestion of a thickness of up to 150 ft. for the zone of T. symondsi. Below this level the sandstones and cornstones are largely replaced by marls, and it seems probable that the zone of T. pococki is represented by unfossiliferous strata. The fauna associated with the zone of Tvaquairaspis symondsi has been recorded by White (1950a : 56) as comprising Anglaspis, Corvaspis, Tesseraspis, Didymaspis, Cephalaspis, Ischnacanthus, Onchus, Onychodus, Kallostrakon, Thelodus, to which must be added a single anaspid scale from Targrove Dingle 6 (104. See also Woodward, 1948). Most of the elements of the fauna occur in the pococki Zone, or in the succeed- ing zone of Pleraspis (Simopteraspis) leathensis. However, both Corvaspis and Anglaspis appear in the symondsi Zone and, near the top, P. (S.) leathensis and Poraspis, in association at New Inn 1 (71) ; whilst Kallostrakon does not appear to extend beyond it. Thus, the faunal assemblage characteristic of the symondsi Zone is T. symondst, Corvaspis, Anglaspis, Tesseraspis, Ischnacanthus wickhami (see pp. 262-264) and Kallostrakon. 202 THE OLD RED SANDSTONE OF BROWN CLEE HILL (ii) Ditton Series Zone of Pteraspis (Sumopteraspis) leathensis White The base of the Ditton Series is marked by a distinctive lithological feature, the “ Psammosteus ’’ Limestones, and an important faunal change, the replacement of Traquaivaspis by Pteraspis (White, 1950a : 56). Although P. (S.) leathensis occurs below the “‘ Psammosteus’’ Limestones, and T. symondsi just persists into the bottom of the zone of P. (B.) croucm, as can be seen from Table 1, T. symondsi is largely concentrated below the limestones and is rare above them (see also White, 19500: 74). A further important form occurring with P. (S.) leathensis is Pteraspis (Pteraspis) rostrata var. trimpleyensis White, which appears about the middle of the leathensis Zone and persists into the lower part of the succeeding zone of P. (B.) crouchi; moreover P. (P.) vostvata var. indet. occurs near the base of the leathensis Zone at Kidnall Gutter, 1 (44). The earliest arctolepid (cf. Kujdanowiaspis sp.) in the area is recorded from the bottom of the deathensis Zone, at Criddon Bridge (109). In addition to these, the fauna associated with the Jeathensis Zone is relatively large (White, 1950a), consisting of 7. symonds1, Corvaspis, Anglaspis, Poraspis, Tesseraspis, ““ Ischnacanthus’’, Gomphodus, Onchus, Cephalaspis, and acanthodian and thelodont scales. Several of these forms occur in the preceding and succeeding zones, but the assemblage which may be regarded as typical of the Jeathensis Zone is P. (S.) leathensis, P. (P.) rostrata var. trimpleyensis, Tesseraspis, Poraspis, and Anglaspis. The highest locality from which P. (S.) leathensis is recorded is about 110 ft. above the main ‘“‘ Psammosteus’’ Limestone (16. Clapgate Cottage Quarry), whilst the lowest occurrence of P. (B.) crouchi is about 170 ft. above the limestone (98. Sudford Dingle 2). Thus, there is a gap of some 60 ft. between the two, and pending further information regarding the range of these species, the /eathensis Zone is here regarded as extending 140 ft. above the main “‘ Psammosteus’’ Limestone. The provisional determination of P. (S.) leathensis for small pteraspids from Hopton Brook given by Dineley & Gossage (1959 : 232), has now been verified by Dr. White. However, though recorded as occurring 250 ft. above the “‘ Psammosteus’’ Limestone, it is now evident that owing to faulting their position relative to the “‘ Psammosteus ”’ Limestone is uncertain. Zone of Ptervaspis (Belgicaspis) croucht Lankester The croucht Zone is poorly fossiliferous, including much barren marls and silts. However, its onset is marked by a further important faunal change in that some of the heterostracans which constitute such a significant element in the faunas of the preceding zones persist only into the lowest part of the zone, being replaced by new forms at a slightly higher level. The former include P. (P.) rostrata var. trimpleyensis as well as P. (P.) rostrata var. indet., T. symonds1, which occurs at the very base, ? Corvaspis, and Poraspis (though it should be noted that Poraspis sericea probably extends well into the crouchi Zone in Herefordshire and Monmouthshire). Although Anglaspis is largely confined to the preceding zones, its occurrence at Clee THE OLD RED SANDSTONE OF BROWN CLEE HILL 203 St. Margaret (17) demonstrates its persistence into the upper part of the crouchi Zone, and Tesseraspis is noteworthy for also extending into the middle of the zone. New forms which appear in the zone include Wezgeltaspis and P. (P.) rostrata var. waynensis, in the lower part, Pteraspis (Pteraspis) datrydinglensis sp. nov. and P. (B.) croucht var. heightingtonensis in the middle, whilst arctolepids, notably Kujdanowiaspis anglica, become more common in the upper part of the zone. P. (P.) dairydinglensis also occurs at the same level in the Cleobury Mortimer area (Dineley & Gossage, 1959 : 232. ... anew and distinctive species of Pteraspis ’’), and it is now evident that the specimens from beds 600 to 850 ft. above the base of the Ditton Series recorded as P. vostvata (Dineley & Gossage, 1959) are indeed P. (P.) dairy- dinglensis. Cephalaspis fragments occur throughout the zone, which includes near its base the distinctive cephalaspid fauna of Whitbatch (Stensi6, 1932), and also acanthodian and thelodont scales. The diagnostic fauna of the cvoucht Zone may therefore be regarded as consisting of P. (B.) croucht, P. (P.) rostrata, P. (P.) datry- dinglensis, Tessevaspis, ? Weigeltaspis, and Kujdanowiaspis. This is a much restricted fauna compared with that occurring with P. (B.) crouchi in Herefordshire and Mon- mouthshire (White, 19504 : 56). As defined above, the arbitrary base of the crouchi Zone is here regarded as being 140 ft. above the main “‘ Psammosteus’’ Limestone. The highest horizon yielding P. (B.) croucht in the type area is about 800 ft. above the main “ Psammosteus ”’ Limestone, although it has been recorded from a level of goo ft. in the Cleobury Mortimer area (Dineley & Gossage, 1959 : 232). The lowest horizon yielding a fauna of the succeeding Jeachi Zone, in the Clee Hills, also occurs about goo ft. above the main “‘ Psammosteus’’ Limestone. This discrepancy between the two areas may be accounted for by the increase in thickness of the Ditton Series, south-eastwards from Brown Clee Hill to Cleobury Mortimer. Thus, the upper limit of the crouchi Zone is arbitrarily placed at 850 ft. above the main “‘ Psammosteus ’’ Limestone, though it may occur at a somewhat lower level in the north-west of the area and higher in the south-east. Zone of Ptevaspis (Cymripteraspis) leachi White Like P. (P.) rostrata and P. (B.) crouchi, P. (C.) leachi occurs widely in the Anglo- Welsh area and also on the Continent. In the Clee Hills area it is commonly found in association with Pyotaspis (Europrotaspis) crenulata sp. nov., and a diversified arctolepid fauna consisting largely of new forms, but including Kujdanow1aspis anglica. A rich fauna, with several new genera and species, has been obtained from Besom Farm Quarry (8), but, ike many of the upper Dittonian localities to the east of Brown Clee, it is not possible to estimate its height above the main “‘ Psammosteus”’ Limestone. Besom Farm Quarry is also notable for having yielded Benneviaspis salopiensis sp. nov., the youngest cephalaspid in this area. As in the lower zones, acanthodian and thelodont scales form the major part of the micro-fauna, the former persisting to the very top of the zone where they were collected by King, with a fragment of an arctolepid, from a pellety sandstone imme- diately underlying the lower Abdon Limestone at Walton Quarry (114) estimated 204 THE OLD RED SANDSTONE OF BROWN CLEE HILL to be about 1,450 ft. above the main “‘ Psammosteus’’ Limestone. The highest locality in the area yielding Pteraspis is The Thrift (106), some 1,300 ft. above the main ‘‘ Psammosteus’’ Limestone. In all, the assemblage characteristic of the leacht Zone comprises Ptervaspis (C.) leachi, Protaspis (Europrotaspis) crenulata sp. nov., Benneviaspis salopiensis sp. nov., Kujdanowtaspis anglica, K. willsi sp. nov., Wheathillaspis wickhamkingi gen. et. sp. nov., Prescottaspis dineleyi gen. et sp. nov., Overtonaspis billballi gen. et sp. nov., Onchus wheathillensis sp. nov., Nodonchus bambusifer gen. et sp. nov., and Ischnacanthus (?) anglicus sp. nov. As shown above, the lowest records of P. (C.) leachi are from about goo ft. above the main “ Psammosteus ’’ Limestone (Bockleton Brook and Upper Overton Quarry), but the junction of the leachi Zone with the underlying crowchi Zone is drawn arbi- trarily some 500 ft. below. Upwards, the zone is regarded as extending to the base of the Clee Series, though its upper part appears to be poorly fossiliferous. (c) Invertebrates A varied invertebrate fauna is known from the Lower Old Red Sandstone of the Anglo-Welsh area, but only a few localities yielding invertebrates have been recorded in the Clee Hills district. The occurrence of Mollusca is of particular interest, since they may imply the establishment of brackish water conditions. Lamellibranchs are known from the lowest part of the Downtonian, and Modiolopsis sp. occurs just above the main “‘ Psammosteus’’’ Limestone at Criddon Bridge (19). The highest horizon yielding lamellibranchs in the Clee Hills is that recorded by the Geological Survey (Eyles, 1953 : 26), about 150 ft. above the ‘‘ Psammosteus ’’ Limestone, near Ditton Priors. An occurrence of especial interest is the first record of Spzvorbis from the Old Red Sandstone. A number of specimens were found attached to the interior of the dorsal disc of the holotype of Protaspis (Europrotaspis) crenulata (p. 269 Text-fig. 16; PI. 43, fig. 3). Like the Spirorbis of the Spivorbis limestones which occur in the Upper Carboniferous of the Wyre Forest, they imply the establishment of more saline conditions, though lacustrine rather than marine. Fragments of eurypterids are scattered throughout the lower part of the succession, though they are usually poorly preserved. However, densely packed bands of euryp- terid debris incorporating better preserved specimens occur in a green, silty clay at Leath Bank. Crustacean tracks were recorded by Roberts (1863) from a fine sand- stone at Bouldon, and further tracks have recently been found there and at Tugford. (d) Plants Plant remains occur throughout the Downton-Ditton succession, especially the latter, and are usually associated with green, grey or buff shales, siltstones and marly cornstones, and frequently with traces of (?) malachite. A similar association of plant remains with copper minerals was noted by Dixon (1921 : 30, 31) in Pembrokeshire, and he suggested that the precipitation of the malachite may have been effected by the decomposing organic matter, as in the German Kupferschiefer. THE OLD RED SANDSTONE OF BROWN CLEE HILL 205 The preservation of the plant fragments is almost invariably poor, with the excep- tion of Pachytheca which is preserved in both calcareous and carbonized states, and occurs widely throughout the succession. Among the better preserved material, it has been possible to determine Prototaxites, Nematothallus,? Cooksonia, and Pachytheca. Ill. UPPER OLD RED SANDSTONE STRATIGRAPHY (a) Farlow Sandstone Series The name ‘‘ Farlow Sandstones ”’ was first applied by King (1925 : 383, 386) to beds outcropping at Farlow and to the south of Titterstone Clee Hill, which contained a typical Upper Old Red Sandstone fauna. However, as has been shown above, the beds outcropping at Clee Burf and ascribed to the Farlow Sandstones (p. 386) are actually Clee Series. King had subsequently arrived at a similar conclusion (im lit.), including the beds within his stage III or “‘ Brownstones’’. In a later paper King (1934) introduced the term “ Farlovian ”’ using it with both a stratigraphical and a time connotation. As Croft (1953 : 430) has pointed out, King subsequently extended it to embrace all Upper Old Red Sandstone beds of the Anglo-Welsh area which un- conformably overlie lower horizons. Here, however, it is proposed to retain ‘‘ Farlo- vian’”’ purely as a time division and to re-introduce the stratigraphical name “ Farlow Sandstones ”’ for the beds of Farlovian age at Farlow and the neighbouring areas. The total area of outcrop of the Farlow Sandstone Series is very small, the greater part forming the prominent ridge extending east-north-eastwards from Farlow. The remainder consists of small areas outcropping to the south and east of Titterstone Clee Hill. It has been found convenient to subdivide the Farlow Sandstone Series on the basis of marked lithological contrast into an upper Grey Farlow Sandstone Formation and a lower Yellow Farlow Sandstone Formation. (i) Yellow Farlow Sandstone Formation As the name implies, the Yellow Farlow Sandstone Formation consists almost entirely of yellow, yellow-buff and sometimes yellow-orange friable sandstones, siltstones and grits. These are largely composed of angular to sub-rounded and sometimes rounded grains of igneous and metamorphic quartz, metamorphic quartzite and quartz-schist, “‘ felsite’’, cherts in variety, with rarer grains of muscovite, fresh plagioclase feldspars, tourmaline and garnet (see also Fleet, 1925 : 507). The grain size ranges widely, from silt (o-or mm.) to grit (3-0 mm.) but averages between 0-02 and 0-8 mm. The grains are poorly cemented, largely by ferruginous and clay cements. Iron staining and thin limonitic banding also occur. The sandstones tend to be massive and characteristically current-bedded, indi- vidual lenses frequently showing rapid alternation of fine and gritty layers which vary widely in degree of sorting. Some of the coarser lenticles are very poorly sorted and it is in these that the most rounded grains commonly occur. The finer sandstones 206 THE OL DT REDE SANDS ®ONE OP eB ROWAN Cie Ei riya and siltstones are frequently flaggy, with micaceous bedding planes. A few layers of yellow-buff silty marl, rarely exceeding 2 in. in thickness, occur as partings between the sandstone. In contrast, pellets of buff and grey silty marl are commonly incor- porated in the sandstones throughout the formation, and also pebbles, largely of vein quartz and to a lesser extent grey and purple quartzites and cherts, especially jasper. The pebbles are mostly sub-angular to angular though a few are rounded, being between 1 and 4 cm. in size. Some of the more conglomeratic lenses also incorporate angular and subangular fragments of sandstone, obviously derived from beds but lately consolidated. The basal 10 ft. of the formation is markedly con- glomeratic and contains pebbles up to 14 cm. in size. The Yellow Farlow Sandstone Formation (which for brevity will hereafter be referred to as the Yellow Formation) is relatively well exposed at Farlow, where its scarp face forms the steep feature of Farlow Bank. An excavation made on the lower slopes of the feature showed the basal conglomerate of the formation to be resting on red-brown silty marls of the Ditton Series. The detailed nature of the contact was obscured by the highly weathered condition of the marls, but there appeared to be little, if any, angular unconformity. The thickness of the Yellow Formation at Farlow is estimated to be about 190 ft., and the dip here, as in the formation along the whole of the Farlow-Walton ridge, averages 20 degrees to the south-south-east. Eastwards from Farlow, the ridge becomes less steep and exposures are few and poor as far as Oreton, two-thirds of a mile from Farlow, where the Yellow Formation is exposed in a road section along the northern foot of the feature. Although it has not been possible to determine the full thickness of the formation at this point, it appears to be in excess of 110 ft. Still further to the east the feature is much reduced, being breached by the River Rea at Prescott. Here, however, the Yellow Formation again gives rise to a strong feature forming a small river-cliff on the east bank of the Rea, and is well exposed along the road cutting. The thickness of the formation at this point is about 210 ft. West of Farlow exposures are again few and poor, and the Cornbrook Sandstone oversteps the Carboniferous Limestone and Grey Farlow Sandstone Formation on to the Yellow Formation. However, the Farlow Sandstones again outcrop at Cleeton, 24 miles to the south-west. (ii) Grey Farlow Sandstone Formation The top of the Yellow Formation appears to have been marked by a sudden change in the conditions of deposition, since there is a sharp transition from yellow sandstone and grits to a succession of sandstones, cornstones, rubbly limestone and marls, predominantly grey in colour. The Grey Farlow Sandstone Formation is poorly exposed and such outcrops as occur are mostly of the more resistant sandstones and rubbly limestones. Though the sandstones are largely grey or shades of grey in colour, green, brown, buff and purple bands also occur. Like those of the underlying Yellow Formation they consist predominantly of sub-angular to sub-rounded quartz grains, with grains of quartzite, feldspars and cherts. Similarly, the grain size ranges be- tween silt and grit, but most of the sandstones are fine to medium (i.e. 0-I to 0-5 mm.). THE OLD RED SANDSTONE OF BROWN CLEE HILL 207 Many of the sandstones are calcareous, often showing lustre-mottling, and they commonly incorporate small sub-angular and rounded pebbles of vein quartz, cherts, jasper, quartzite and pellets of grey and buff marl. Unlike the homogeneous and relatively massive appearance of the Yellow Formation, the Grey Formation is a far less uniform and more thinly bedded series, individual beds rarely exceeding 4 ft. in thickness. Current bedding is common and many of the finer sandstones are laminated, showing ripple marking and occasionally micaceous bedding planes. Frequent conglomeratic cornstones occur and are usually grey or green in colour. They are composed largely of fragments of silty and calcareous marl, silty sandstones, and small pebbles like those in the sandstones, set in a matrix of sandy calcareous marl. As in the cornstones of the Lower Old Red Sandstone, fragments of fish are sometimes incorporated. Also characteristic of the Grey Formation are the lenticular “ rubbly ”’ limestone bands, ranging in thickness from a few inches up to 4 ft., which are grey, purple-grey and green-grey in colour. They are composed of fragments of sandy and marly lime- stone and calcareous sandstone set in a limestone matrix. Sub-rounded, rounded and sub-angular sand grains are commonly incorporated, with occasional pebbles of vein quartz, quartzite and cherts. In thin section, many of the limestones are seen to have a clotted or breccioid structure, having undergone a varying degree of recrystal- lization, with large zones or centres of finely-crystalline calcite set in a coarsely crystalline matrix. Like the conglomeratic facies of the Abdon Limestones it seems probable that the limestone originated from the breaking up of sheets and crusts of muddy and sandy limestones, with pene-contemporaneous re-cementation by muddy limestones. In several instances secondary precipitation has occurred around many of the fragments. The rubbly limestones are frequently associated with white lustre- mottled sandstones, a further parallel with the Abdon Limestones. Though the marl pellets which form such a conspicuous element in the composition of the cornstones and many of the sandstones are mainly grey and shades of grey in colour, most of the marls seen to outcrop are red with subordinate green bands. A well-marked red marl, some 6 ft. thick, outcrops immediately above the Yellow Formation, forming the basal member of the Grey Formation. However, owing to the lack of exposures it has not proved possible to assess the total proportion of marls present in the latter. To the south of Prescott the Grey Formation is estimated to be 275 to 300 ft. thick and is overlain by the basal conglomerate of the Carboniferous Limestone. At Oreton the thickness is reduced to roo ft., and at Farlow only 40 ft. is present. The significance of this great thinning of the Grey Formation is reflected in the topo- graphy and will be discussed below. At Prescott, the Farlow Sandstone Series gives rise to two well-marked escarpments, one formed by the Yellow Formation and the other by the upper part of the Grey Formation, capped by the basal conglomerate of the Carboniferous. Owing to the thinning of the Grey Formation these merge to form a single feature at Oreton, which continues westwards to Farlow. FAUNA Though poorly fossiliferous, the Farlow Sandstone Series has long been known to GEOL. 5. 7. 20 208 THE OLD RED SANDSTONE OF BROWN CLEE HILL contain a characteristic Upper Old Red Sandstone fish fauna (Egerton, 1862 : 103- 105; Traquair, 1894-1906 : 128, 129). That collected from the Yellow Formation is largely fragmentary, with a few rare complete plates and scales. With the excep- tion of the original finds described by Egerton, all are in the form of moulds, and despite the fact that most are preserved in medium to coarse, and even gritty sandstones, the detail of the ornamentation is mostly good. In some instances the vascular cavities have been preserved as a delicate filigree of limonitic mud. Apart from the classic locality at Farlow, the exact position of which is now unknown, only two new localities in the Yellow Formation have yielded fossils, one at Farlow (29), the other at Prescott (82). The latter has proved to be the more prolific, the fragments being concentrated in a small lens of gritty sandstone, obviously repre- senting a winnowed deposit. But here, as at Farlow, other plates occur scattered randomly in finer sandstones. Forms represented include Bothriolepis sp., B. macro- cephala, Holoptychius sp., Pseudosauripterus anglicus gen. nov., Eusthenopteron farloviensis sp. nov., and indeterminate crossopterygian and dipnoan remains. The Grey Formation is even less fossiliferous than the underlying Yellow Forma- tion, having yielded only a small fauna of mostly indeterminable bone fragments, but including Holoptychius sp. The fragments come from the stream section at, 710 yds. (112) and 800 yds. (113) south-south-east of Prescott, in grey, calcareous pebbly sandstones and cornstones, and are much abraded. IV. LOWER CARBONIFEROUS (a) Basal Conglomerate Though seemingly outside the scope of the present paper, a brief account of the conglomerate which occurs between the base of the Carboniferous Limestone and the top of the Grey Farlow Sandstone Formation forms a necessary corollary to a description of the latter. The conglomerate has been variously regarded as constitu- ting the uppermost part of the Upper Old Red Sandstone ; as forming the basal bed of the Carboniferous ; and as a transitional bed between the two. The conglomerate is buff-brown and orange in colour and is largely composed of angular and sub-angular quartz grains, ranging in size between o-I and 3:0 mm., the coarser grades predominating. It is as a whole friable, being poorly cemented by ferruginous and calcareous cements. Current bedding is characteristic and the lenses of fine-grade material are usually better cemented. The constituent pebbles range upwards from 0-3 to I5 cm. in size and are mostly sub-angular to sub-rounded. By far the greater proportion are of white and pink vein quartz, with numerous pebbles of purple, pink, yellow and grey quartzites, cherts in variety, jasper and “ lydite ’’. The conglomerate forms a prominent crest along the greater part of the Farlow- Walton ridge, and is exposed at several points. In the road section of Farlow Bank it is some 6 ft. thick, but owing to slipping its relationship with the underlying Grey Farlow Sandstones and the overlying Carboniferous Limestone is not clear. East- wards from Farlow, the conglomerate begins to increase in thickness, being some 12 ft. thick at ‘‘ The Rough ’’, one-third of a mile due east of Farlow Church. Here, the top of the conglomerate becomes progressively more calcareous upwards, passing THE OLD RED SANDSTONE OF BROWN CLEE HILL 209 into thin-bedded, sandy, pebbly limestones containing numerous crinoid ossicles and characteristic Carboniferous Limestone brachiopods, which are succeeded by typical limestone shales. The conglomerate is not well exposed at Oreton but appears to be at least 12 ft. thick. The greatest development of the conglomerate is at Prescott, where it is some 35 to 4o ft. thick and gives rise to a small but well-marked escarpment. But neither here nor elsewhere is there a reliable exposure of the contact of the base of the con- glomerate with Grey Farlow Sandstones. However, since as will be shown below, the conglomerate truncates the Grey Farlow Sandstone Formation, and in view of the passage from it into the Carboniferous Limestone, the conglomerate can be regarded as forming the basal bed of the Carboniferous. The limestone shales into which the conglomerate passes upwards have been assigned by Vaughan (1905 : 252- 254) to his Cleistopora zone. Thus the conglomerate may well represent the lower part of K. V. STRUCTURE (a) Folding The principal tectonic feature of the area is the folding of the strata along Cale- donoid axes to form three major, asymmetrical but relatively simple structural units, the Brown Clee and Titterstone Clee Synclines and the interposing Ledwyche Anticline. Brown Clee Syncline. The north-western margin of the Brown Clee Syncline is delimited by the Church Stretton fault zone. Its broad north-western flank com- prises Ordovician strata dipping south-eastwards from the Stretton Hills, succeeded by the Silurian in Apedale, Wenlock Edge and Hopedale, and in Corvedale by the Downton and Ditton Series dipping at 5 to 10 degrees. The south-eastern flank of the syncline is, in contrast, very narrow and the strata are in general more steeply inclined ; whilst its core is formed by the Clee Series of Brown Clee Hill. The axis of the syncline extends east-north-eastwards from Bromfield, to Hopton Cangeford ; there it swings north-eastwards to Warren Cottage, near which it can be accurately determined in a stream section, and thence to Clee Burf, up to which point it plunges steadily to the north-east. Between Clee Burf and Cleobury North Liberty the direc- tion of plunge becomes reversed and the axis extends east-north-eastwards nearer to the topographic axis of the hill, and it is also paralleled by the plane of the Brown Clee Fault. From Cleobury North Liberty the axis reverts to a north-easterly direc- tion, extending through Morville to Linley and plunging to the south-west. Titterstone Clee Syncline. This has previously been referred to as the “‘ Clee Hill Syncline ’”’ (Whitehead & Pocock, 1947 : 120), but to avoid possible confusion it is here specifically related to Titterstone Clee Hill. It is a more complex structure than the Brown Clee Syncline since in the area of Titterstone Clee Hill it incorporates several subsidiary, but well-defined folds. However, only the north-eastern extremity of the syncline occurs in the area being described and here the subsidiary folding is not evident. Like the Brown Clee Syncline, the Titterstone Clee Syncline is asym- metrical, with a broad north-western flank dipping at Io to 20 degrees to the south- 210 THE OLD RED SANDSTONE OF BROWN CLEE Hire east, increasing towards the Farlow-Walton ridge where dips of up to 65 degrees occur in the Avonian Limestones. It is possible that these latter dips are, at least in part, a consequence of the collapse and slipping of the limestones into the sub- terranean stream course which drains from Farlow eastwards to the Factory Cottages, along the southern margin of the ridge (Morris & Roberts, 1862 : 96). The few dips which have been obtained from the succeeding Cornbrook Sandstone indicate that they are inclined at a very much shallower angle, averaging 10 degrees. The south- eastern flank of the syncline is narrow and in the present area consists of Ditton Series dipping moderately steeply, averaging 30 to 40 degrees, to the north-west. The axis of the Titterstone Clee Syncline is difficult to determine with accuracy owing to the cover of the Cornbrook Sandstone and the Coal Measures, but in general it diverges slightly from that of the Brown Clee Syncline, extending from Knowbury east-north-eastwards to Catherton Marshes where it swings to the north-east, con- tinues to Bagginswood, and becomes coincident with the Titterstone Clee Fault. Ledwyche Anticline. This is an asymmetrical structure separating the Brown Clee and Titterstone Clee Synclines, the north-western flank being narrow and more steeply inclined than the broader south-eastern flank. Its axis is broadly parallel to that of the Brown Clee Syncline and extends east-north-eastwards from Middleton to Langley, and then north-north-eastwards to Coldgreen, where it swings round to assume a north-eastwards course to Underton. In the area of Coldgreen there is some evidence that the fold axis is associated with local faulting. The axis of the Ledwyche Anticline plunges steadily to the north-east, though the amount of plunge decreases between Langley and Underton, north of which it appears to die out. To the south-west the Brown Clee and Titterstone Clee Synclines, with the inter- posing Ledwyche Anticline, merge into the structural complex of Silurian strata extending between Ludlow and Presteign and delimited on the west by the Church Stretton fault zone. The relatively more complex structure of the Titterstone Clee Syncline is reflected in its south-westward prolongation through Caynham into the more tightly folded area extending from Richard’s Castle, through Leinthall Earls to Presteign, accompanied by a well-defined fault-belt. It has also been suggested that the Titterstone Clee Syncline itself forms the south-western prolongation of the Stafford Syncline (Whitehead & Pocock, 1947 : 122). The Ledwyche Anticline expands south-westwards into the broad anticlinal core in Wenlock Shales at Wig- more Rolls, whilst the Brown Clee Syncline noses out in the area of Adforton. How- ever, to the south-west of the Silurian structural complex extending from Ludlow to Kington, the broad fold of the Brown Clee Syncline appears to have a southern counterpart in the Mynydd Eppynt. To the north-east of the present area the folds become much reduced and lost beneath the Coal Measures of the Wyre Forest. (b) Faulting There is considerable evidence of faulting in the area, but except where such well- defined horizons as the ‘“‘ Psammosteus ’’, Abdon and Avonian Limestones are affected, the extent, direction and downthrow of the faults are difficult to determine. The majority of the faults are orientated broadly NE.-SW., parallel to the main fold THE OLD RED SANDSTONE OF BROWN CLEE HILL 211 axes, with a minor series approximately at right angles to them. In the few instances where the hade and throw can be determined, the faults appear to be normal. The area is affected by two major faults, the Brown Clee Fault and the Titterstone Clee Fault, both associated with the two major fold axes. Brown Clee Fault. Nowhere has the plane of the Brown Clee Fault been seen to outcrop, but its effects can first be traced near Gibbridge in the Ledwyche Valley, where it has a Caledonoid trend and downthrows the main ‘‘ Psammosteus’’’ Lime- stone some 150 ft. to the south-east. North of Gibbridge the fault appears to follow a gently sinuous but broadly N.-S. course, faulting out a prominent sandstone/ cornstone feature at Bank House, and downthrowing the Abdon Limestones some 200 ft. to the east near Bockleton Court, where the fault begins to swing round again to the north-east. The Coal Measures of Brown Clee Hill are poorly exposed and it is difficult to trace the fault across them, but it upthrows a narrow strip of Coal Measures to the north-west of Clee Burf, and downthrows the south-eastern portion of the Abdon Burf outcrop. At Hillside, the Brown Clee Fault again displaces the Abdon Limestones, though to a lesser degree than at Bockleton Court, namely some 75 ft. to the east. Over the southern part of Brown Clee Hill the course of the fault appears generally to coincide with the axis of the Brown Clee Syncline. The effects of the fault can next be detected in the stream section 400 yds. south-east of Middleton Priors, and again at Moore Dingle, three-quarters of a mile south-west of Aston Eyre, where the “ Psammosteus ’’ Limestones are downthrown some 150 ft. to the east. Titterstone Clee Fault. The present area includes only the north-eastern part of the Titterstone Clee Fault, which extends north-eastwards from Hope Bagot on the southern slopes of Titterstone Clee Hill, in which area it displaces the outcrop of the Avonian Limestones between Gorstly Rough and the Novers, and throws Coal Mea- sures against Lower Old Red Sandstone to the south and south-west of Clee Hill. At Catherton Common the Cornbrook Sandstone is thrown against both Downtonian and Dittonian strata, and though it has not proved possible to measure the amount of downthrow, it would appear to be at least 1,000 ft. to the north-west. North- eastwards from Catherton Marshes the line of the fault coincides broadly with the axis of the Titterstone Clee Syncline, and is defined to the east of the River Rea by vertical and disturbed strata in exposures half a mile north of Detton Hall. Here the fault appears to bifurcate, the main branch extending north-north-eastwards between Prescott and Walton, throwing the Farlow Sandstone Series against the Clee Series, and continuing as the Deuxhill Fault (Whitehead & Pocock, 1947 : 131). The subsidiary branch extends at first east-south-eastwards and can be traced in the stream section to the south of Walton where the beds are contorted and disturbed ; it then gradually swings northwards through Bagginswood, where it is possibly joined by the Brock Hall or Billingsley Fault (Whitehead & Pocock, 1947 : 130). From Bagginswood it continues north-north-eastwards to rejoin the main branch south-west of Chorley. It seems probable that the Titterstone Clee Fault links up to the south-west with the fault extending between Caynham Camp and Leinthall Earls. Faulting exerts considerable influence on the Farlow-Walton Ridge which, as noted by Murchison (1839 : 121), is broken by a series of dip-faults into a number of 212 THE OLD RED SANDSTONE OF BROWN CLEE HILL blocks which are slightly displaced relative to each other. The amount of downthrow appears to be small, not exceeding 100 ft., and is in most instances to the west. It is along the line of one of the faults that the River Rea cuts across the ridge at Prescott. The escarpment of the ‘‘ Psammosteus ’’ Limestones also shows a number of small displacements owing to minor faulting, as do the Abdon Limestones at Abdon Liberty and Nordybank, again the amount of downthrow usually not exceeding roo ft. (c) Unconformities The sequence from the Upper Ludlow Series to the top of the Clee Series is per- fectly conformable and there is no evidence to support the existence of the uncon- formity postulated by King (1934 : 529 ; 1m lit. ; see also Wills, 1948 : 29, 30) to occur at the base of his “‘ Brownstones ”’ or “ Division III’. It is evident from King’s field-maps (deposited in the Department of Geology, University of Birmingham) that on the western side of Brown Clee Hill his “‘ Brownstones ”’ are broadly equi- valent to the Clee Series of the present classification. However, to the east of the hill, King includes a large area of what can now be shown to be the Ditton Series within his ‘‘ Brownstones’’. In postulating an intra~-Lower Old Red Sandstone unconformity, King was obviously influenced by the alleged unconformity claimed by Miss D. M. Williams (1926 : 219-225) to occur beneath the Brownstones in the Gower area. But this too has been disproved (George, 1939 : I-5 ; Croft, 1953 : 429), there being an unbroken sequence from Downtonian to Breconian in South Wales as in the Welsh Borderland. As is proved by their faunas, a large time-gap exists between the Farlow Sandstone Series and the underlying Lower Old Red Sandstone, yet the angular disconformity between them appears to be very slight. At Farlow the Farlow Sandstones rest upon high Ditton Series, but owing to a divergence in strike, the Dittonian striking NE.-SW. and the Farlovian a little south of east, the Farlow Sandstones rest upon successively higher beds eastwards to Prescott. Between Prescott and Walton it is probable that the Farlow Sandstones overstep onto the Clee Series though the junction is buried beneath the Farlovian cover, the latter having been eroded from the upthrown wedge of Clee Series at Walton. King (1925 : 386) states that beds which he regards as being Farlow Sandstones show much greater angular discon- formity on the southern Flanks of Titterstone Clee Hill, where they rest across de- nuded fold axes in Downtonian and Dittonian strata. A further break occurs at the top of the Farlow Sandstone Series, although these have previously been suggested to have a perfectly conformable relationship with the succeeding Carboniferous Limestone (King, 1925 : 389 ; Dixon, 7m Watts, 1925 : 395). Again, it is difficult to detect any angular disconformity between the two series except at Farlow, where it may be more apparent than real owing to slipping. However, the magnitude of the break is demonstrated by the relationship of the basal Carboniferous Conglomerate to the underlying Grey Farlow Sandstone Formation. South of Prescott the conglomerate rests upon 275 to 300 ft. of the Grey Farlovian ; three-quarters of a mile to the west at Oreton the Grey Farlovian has become reduced to roo ft., whilst at Farlow a further three-quarters of a mile to the THE OLD RED SANDSTONE OF BROWN CLEE HILL 213 west it is only 40 ft. thick. Owing to lack of exposures it has not proved possible to ascertain whether the basal conglomerate oversteps onto the underlying Yellow Farlovian to the south-west of Farlow, but south of Silvington an attentuated Avonian sequence and the Grey Farlovian (if present) are overstepped by the Cornbrook Sandstone. The relationship of the Cornbrook Sandstone to the underlying strata between Cleeton and Factory Cottages, and also on the western flanks of Titterstone Clee Hill, has led one of the writers (H. W. B.) to arrive independently at the same con- clusions as George (1956 : 308, 309) regarding its unconformable nature. The Corn- brook Sandstone is itself unconformably overlain by Coal Measures which can be correlated with the Kinlet Beds (Ammanian) of the Wyre Forest Coalfield. The latter beds rest upon Ditton Series along the south-eastern margin of the area and in small outliers near Middleton Scriven and Chorley, whilst the Coal Measures of Brown Clee Hill which are of the same age (Whitehead & Pocock, 1947 : 45-47) rest upon Clee Series. As has been pointed out by Whitehead & Pocock (1947 : 44), in each of these areas, as at Titterstone Clee, the base of the Coal Measures is marked by a distinctive white or grey, hard sandstone which may in part be conglomeratic. A further unconformity occurs between the Kinlet Beds and the succeeding Highley Beds (Morganian) which overstep onto the Ditton and Downton Series in the north- east of the area. (d) Age of the Movements The principal folding movements affecting the area were those of the Caledonian orogeny which uplifted the Lower Old Red Sandstone of the whole of the Anglo- Welsh region. These movements were of post-Breconian, pre-Farlovian date, but since the upper limit of the age of the Breconian is unknown, though it may extend into the Middle Old Red Sandstone (Croft, 1953 : 431), more precise dating with reference to the continental Devonian stages is impossible. Jones (1956 : 334) has shown that much faulting accompanied the Caledonian folding in South Wales, but it has not proved possible to distinguish any undoubted Caledonian faulting in the Clee Hills, though the alignment of the major faults with the folds indicates some degree of interrelationship. This is particularly apparent in the association of the Brown Clee and Titterstone Clee Synclines and Faults, and in the former instance there is some indication that the Coal Measures are less affected by the Brown Clee Fault than the underlying Clee Series. However, even this can only be taken as evidence of pre-Kinlet Beds faulting with subsequent post-Kinlet Beds rejuvenation. Whatever the age of the faulting, since it appears to be normal, it occurred during periods of tension interspersed between phases of major compression. A characteristic feature of the folds, and also of the Neen Savage Anticline and the Cleobury Mortimer Syncline to the south (Dineley & Gossage, 1959 : 233), is their asymmetry ; the synclines having broad, north-western flanks and narrow, more- steeply dipping, south-eastern flanks, the anticlines showing the reverse. It has not been possible to determine whether this is an original feature of the Caledonian folding movements, but even if this was so the asymmetry was further accentuated 214 THE OLD RED SANDSTONE OF BROWN CLEE HILL by Carboniferous movements (see also Squirrell & Tucker, 1960 : 167-170). This is demonstrated by the Cornbrook Sandstone which in the area of Cleeton on the north- western flank of the Titterstone Clee Syncline dips at about ro degrees to the south- east, whilst near to Clee Hill, on the southern flank, it dips to the north-west at 60 to 70 degrees. A further feature of the folds is the gradual swing in strike from the Armoricanoid trend of the Cleobury Mortimer Syncline (Squirrell & Tucker, 1960) to the typical Caledonoid direction of the Brown Clee Syncline, the folds seemingly having rotated about Titterstone Clee Hill. Moreover, due north of the hill the axes of the Ledwyche Anticline and the Brown Clee Syncline have been deflected north- wards, as has the Brown Clee Fault ; whilst along the same line the dolerite intru- sions of both Brown Clee and Titterstone Clee occur. It is probable that all these features, plus the more complex form of the Titterstone Clee Syncline itself, are reflections of the influence of pre-Caledonian structures at depth. The Caledonian orogeny was succeeded by a period of epeirogenic movements, marked by unconformities between the Farlow Sandstone Series and the Carbonif- erous Limestone, and between the latter and the Cornbrook Sandstone. A more profound phase of movement involving some degree of folding followed the depo- sition of the Cornbrook Sandstone which George (1956 : 307-309) has suggested to be most probably of Namurian age, and preceding the “ Middle ’’ Coal Measures or Ammanian (= “ Yorkian’’. Whitehead & Pocock, 1947 : 40). This was succeeded by early Morganian (= “ Staffordian ’’. Whitehead & Pocock, 1947 : 121) movements, regarded by Wills (1956 : 59) as representing the First Malvernian Movements, and evidenced by the marked overstep of the Kinlet Beds (Ammanian) by the Highley Beds (Morganian) onto the Downton and Ditton Series in the north-east of the area. The only direct evidence of post-Morganian movement in the area is provided by the Deuxhill Fault which throws Highley Beds against Lower Old Red Sandstone ; and it has been suggested (Whitehead & Pocock, 1947 : 122) that the Titterstone Clee Syncline, and presumably the Brown Clee Syncline, owe their present elevation to post-Triassic movements. However, as is shown by the configuration of the Coal Measure basins of Brown Clee and Titterstone Clee, post-Caledonian movements continued folding and faulting along pre-existing axes. VI. CONDITIONS OF DEPOSITION AND PALAEOGEOGRAPHY It is necessary to consider the history of the Clee Hills area in the broader context of that of the Anglo-Welsh region. The major tectonic units of the Anglo-Welsh region as defined by Wills (1951) comprise the Welsh geosyncline (Jones, 1938) and the Midland kratogenic block, the boundary between the two broadly delimited by the great line of disturbance extending from Pembrokeshire, through Church Stretton, and possibly northwards into the Lake District (Jones, 1927). By the end of Silurian times the rate of subsidence in the geosyncline had begun to slow down, and the presence of layers of phosphatic pebbles or “‘ bone-beds”’ in the sediments of the marginal zones indicates pauses in subsidence, or even periods of emergence of areas of the Midland block above the shallow shelf sea which covered it (Lawson, 1954, 1955 ; Squirrell & Tucker, 1960). THE OLD RED SANDSTONE OF BROWN CLEE HILL 215 The base of the Old Red Sandstone is marked by the Ludlow Bone Bed (White, 1950a : 63), a remarkable deposit which has been recorded from a number of localities ranging over a wide area of the Welsh Borderland, and also South Wales (Walmsley, 1959) and the Midlands (King & Lewis, 1912; Whitehead & Pocock, 1948 ; Ball, 1951). It appears to be a remanié deposit (Whittard, 1952 : 176, 177) resulting from a prolonged pause in subsidence over the whole area, which marks the inception of the main phase of the Caledonian orogeny, and was accompanied by folding in South Wales. Although the actual bone-beds are lenticular, the Ludlow Bone Bed forms a stratigraphical unit which is probably broadly synchronous over the whole of the area (cf. Denison, 1956 : 389). The conditions of deposition of the Ludlow Bone Bed indicate shallow waters with active bottom currents, and alternation of marine and brackish environments (see Wills, 1948 : 25; Whittard, 1952 ; Denison, 1956). Other thin, lenticular bone-beds occur in the lower part of the overlying Grey Downton Formation. In the Ludlow area, the Ludlow Bone Bed is succeeded by the Downton Castle Sandstone which marks the inception of the great delta plain which formed the major palaeogeographic feature of the Lower Old Red Sandstone. The Downton Castle Sandstone comprises thick, current-bedded sandstones interbedded with laminated siltstone and shales. Despite the evidence adduced by Denison (1956 : 390, 391) for the marine origin of the Downton Castle Sandstone, the lithology, fauna and flora of the sandstones are strongly indicative of fresh-water or brackish deltaic conditions ; whilst the siltstones and shales are brackish, with rare marine incursions. Similar conditions obtain in the South Staffordshire area (Ball, 1951), where there is no need to postulate a lateral facies change to account for the presence of Hemuicyclaspis murchisoni, as suggested by Schmidt (1959 : 28). But in the geosynclinal zone to the west of Ludlow (Stamp, 1919; Earp, 1938, 1940; Holland, 1959) the equivalent beds are finer grained, and much more brackish in character, with marine bands near the base and they are considerably thicker. Thus it appears that the sandy-delta facies were confined to the north-western area of the Midland block, whilst the very late stages of geosynclinal subsidence were continuing in the Kerry-Knighton areas, where a brackish-water lagoon environment obtained. The sediments would there- fore possibly have been derived from areas lying to the north and possibly the east. Thus the Grey Downton Formation thins north-eastwards from Ludlow (p. 182), and in South Staffordshire the sedimentary structures in the equivalent of the Downton Castle Sandstone indicate a north-easterly origin (Ball, 1951 : 232). The presence of thin, highly localized bone-bed lenticles in the lower part of the succession throughout the area, suggests further brief pauses in subsidence. In the upper part of the Grey Downton Formation the conditions were largely brackish-water lagoonal, with rare marine incursions, though plant fragments occur disseminated throughout the shales. However, the transition from predominantly grey to red sediments which takes place at the top of the Grey Downton Formation, marks a profound change in the geographic and climatic conditions. Some of the views regarding the conditions under which the Old Red Sandstone was deposited have been summarized by Jones (1956 : 336-338) and, as he points out, many of the early opinions favoured an arid climate. However, many of the Downtonian and 2106 THE OLD ERED SANDSTONE VOD VB ROMN Oi Eis Eni Dittonian marls are strikingly similar to the superficial deposits accumulating at present in parts of Africa with a tropical monsoonal climate. Indeed, red and green colours in sediments are no longer regarded as indicating deposition under arid con- ditions, but rather of intense weathering under hot, humid conditions (Krynine, 1949). Thus, the dominant red coloration of the Downton and Ditton Series, as well as their lithology, implies a profound alteration in the climatic conditions prevailing over the land-mass from which they were derived. In the Downton Series marls predominate and suggest an ample supply of lateritic clays resulting from tropical weathering, and deposition under quiet conditions. The marls are probably fluviatile, many being highly micaceous, though some of the clay and silt particles may have been wind-borne. The “ race’ and nodular limestones appear to be penecontemporaneous with the marls, and may have originated like the Kunkar deposits of modern tropical deltas. There is ample evidence of the fluviatile origin of the pellet rocks and of the sandstone bands and lenses, many of which occur in channels cut into the underlying beds. Current ripple-marks are common, and to a lesser extent desiccation-cracks and rain-prints. The pellet beds probably originated from the erosion of sun-baked clay crusts by “ flash ’’ floods, with rapid disintegration of the crusts, attrition of the fragments and re-deposition. Brackish and brackish-marine incursions are marked by bands containing Lingula and molluscs, whilst a trilobite has been recorded from the lower part of the Red Downton Series at Ledbury (Piper, 1898). In the upper part of the succession sand- stones become more common, e.g. the Holdgate Sandstones, as well as cornstones incorporating fragments of fish. The amount of plant debris also increases, whilst bands with invertebrate fossils are rare. The overall impression is one of a broad delta-plain, similar to the sub-aerial part of the present Colorado delta (Sykes, 1935). This covers an area of some 3,250 Square miles, and all but the finest material is deposited by braided and shifting streams, periodically flushed into coalescing flood sheets. Sediment is then rapidly deposited over wide areas, and the mud plains and flats are left dotted with playas and ox-bow lakes. Elsewhere, mud banks and old levées are eroded and current-raked. Re- working of the sediments is common, and deposition takes place not only in stream channels but occasionally over wide continuous areas. Plant debris is accumulated and macerated by the floods, and is often thinly spread over the flats. Though deposition is very near to sea level, sedimentation is active enough to prevent the advance of the sea over the delta, except in rare incursions (cf. Denison, 1956 : 392, 393): At its maximum during the Downtonian and Dittonian, the delta probably ex- tended over the greater part of South and Central Wales, its western limits being formed by the uprising Caledonian Mountains, and across the Midland kratogenic block at least as far as North London, its southern margin being formed by the great geosyncline extending from Southern Ireland, through the Bristol Channel, into France and Germany (Simpson, I95I : 60, 61). Owing to the lack of fossils in the Old Red Sandstone outcrops to the west of the Church Stretton fault zone, it is impossible to correlate them with the sequence in the type area. Moreover, the ‘‘ Psammosteus ’’ Limestones have not been detected THE OLD RED SANDSTONE OF BROWN CLEE HILL 217 west of the Clee Hills area. However, in the Knighton district, Holland (1959 : 472) gives a thickness of approximately 2,000 ft. for his ‘“‘ Red Downtonian ”’, which consists of red and green “‘ marls’”’ and sandstone bands. He also points out that cornstones, typical of the Dittonian, are absent, implying a correlation wholly with the Red Downton Formation of the Clee Hills. In the Clun Forest area, Earp (1938 : 127) records a thickness of 1,000 ft.+ for his “‘ Red Downtonian Beds ’’, but again, it is impossible to correlate them with the type sequence. If these western outcrops are of Downtonian age, the thickness of 2,000 ft. in the Knighton area, compared with 1,200 ft. in the Clee Hills, implies that the rate of subsidence and sedimentation continued to be greater in the basin to the west of the Church Stretton fault zone than on the kratogenic block to the east. The “‘ Psammosteus’’ Limestones, which extend over the greater part of the Anglo-Welsh Old Red Sandstone outcrop from Shropshire to Pembrokeshire, mark a widespread change in the conditions obtaining over the delta, probably effected by slight variations in geography and climate. The limestones probably originated in a series of lagoons with restricted circulation, and though the “‘ Psammosteus”’ Limestones “ phase’ was of short duration, it appears to have been broadly syn- chronous over the whole of the area. As has been shown above, the “‘ Psammosteus ”’ Limestones have not yet been found west of the Church Stretton fault zone, and there is no information regarding their extent to the east of its present outcrop. However, during a re-examination of the Streatham Common Boring (Whitaker 1889 : 224-229) instigated by Mr. J. D. D. Smith of the Geological Survey,! frag- ments of fish from the core at a depth of about 1,225 ft. were examined by Mr. H. A. Toombs and determined as Tvaquaivaspis symondsi, Tesseraspis tessellata and (?) Corvaspis. These forms indicate a high Downton or low Ditton horizon, but broadly about the level of the ‘‘ Psammosteus’’ Limestones. The presence at this level of beds of typical Old Red Sandstone facies in South London raises the interesting point of their relationship with the vertebrate-bearing marine strata of ““... a late Downtonian or post- Downtonian ...”’ age (Straw, 1933 : 139), occurring in the Little Missenden Boring, some 30 miles to the north-west. The “‘ Psammosteus’”’ Limestones usher in a change of conditions between the Downton and Ditton Series; the sediments become coarser and thick cornstone lenses are developed. Fluviatile conditions are well established, with marked lenticularity of the beds, current bedding, ripple-marks, desiccation-cracks and washouts. These changes probably reflect an increase in the tempo of the uplift of the Caledonian mountains with attendant increase in gradient and rainfall. Asso- ciated with this, fish and plant remains become more abundant, particularly in the lower part of the Ditton Series. Erosion surfaces are common, some being deeply incised, suggesting intermittent “sheet ”’ or “ flash ”’ flooding, with swiftly flowing but heavily laden streams, which were braided and shifting. Sudden floods flushing out the streams of a hilly region and debouching onto the delta would doubtless engulf living fish, as well as dead organisms and plant debris. The cornstones of the Ditton Series, which are the richest source of fossils, are formed by the pene- ce 1 The authors are indebted to the Director of the Geological Survey for permission to publish this information, 218 THE OLD RED SANDSTONE OF BROWN CLEE HILL contemporaneous erosion of layers of muddy limestone and calcareous mud resulting from such floods. The pellets may be well consolidated or plastic, and vary from angular to rounded, the fish remains being incorporated as fragments. The corn- stones are markedly lenticular and frequently well sorted. Nodular limestones and “‘race’’ similar to those of the Downton Series also occur. The presence of rare bands of lamellibranchs in the lower part of the Series and Spivorbis in the upper indicates brief periods of increased salinity, though in the case of the latter implying playa lacustrine, rather than marine conditions. The Ditton Series are not known to outcrop west of the Church Stretton fault zone, but an extrapolation of their thickness from the Clee Hills implies their continuity into Central Wales. However, it is not possible to determine whether sedimentation continued to be greater in the geosynclinal region or on the margin of the kratogenic block. Towards the end of Lower Dittonian times there appears to have been a slowing down in the uplift of the Caledonian mountains which is reflected in the sediments becoming finer, with silts greatly predominating over sandstones and cornstones. Thereafter, the increasing coarseness of the strata, persisting into the uppermost Clee Series, reflects renewed and accelerated uplift of the mountains culminating in the main phase of the Caledonian orogeny. In the area of the Clee Hills, the end of the Dittonian is marked by the re-establishment of conditions similar to those under which the ‘‘ Psammosteus ’”’ Limestones originated, with the formation of the lower of the Abdon Limestones. A further period of normal conditions ensued before the recurrence of the calcareous lagoon environment and the deposition of the upper Abdon Limestone. The appearance and structure of some of the limestones in the Clee Series suggests that they may have been formed in a manner similar to recent surface limestones, or “‘ calcrete ’’ (see Du Toit, 1954 : 445-447). The coarse sandstones and grits of the Clee Series with predominantly sub-angular pebbles, imply increased gradients and the closer proximity of mountain footslopes. The composition of the sandstones indicates that an area of metamorphic and igneous rocks was being subjected to erosion, though some of the pebbles of quartzite and jasper may have been secondarily derived. The source of the Upper Silurian limestone pebbles in the Clee Series poses a problem. It seems likely that the nearby Silurian outcrops of the Welsh Borderland and West Midlands were blanketed by a thick cover of Downtonian and Dittonian rocks by the beginning of Breconian times. However, the pebbles may have been derived from outcrops which lay to the north- east. The Breconian delta probably extended over the whole of the Anglo-Welsh area, though its eastern limits are unknown. It is not possible to estimate the original thickness of the Clee Series, but it may well have compared with that of the Breconian strata in the Brecon Beacons, namely 2,100 ft.+ (Croft, 1953 : 429). Between Breconian and Farlovian times the Anglo-Welsh Caledonian orogeny reached its culmination and was succeeded by a period of profound erosion and plan- ation, the whole of the Breconian and the upper part of the Dittonian strata having been removed from the area between Brown Clee Hill and Walton prior to the de- position of the Farlow Sandstones. The Yellow Farlow Sandstones appear to mark the re-establishment of deltaic conditions, with ample evidence of very vigorous stream action. The climate appears to have been relatively arid, but a sudden THE OLD RED SANDSTONE OF BROWN CLEE HILL 219 change in climatic conditions ushered in the deposition of the Grey Farlow Sand- stones, with a return to monsoonal conditions similar to those of the Dittonian, and with broadly similar types of sediments. In the Clee Hills area the Farlovian strata appears to be wholly fluviatile with an extremely sparse freshwater fish fauna, and in the Grey Farlow Sandstones a few carbonaceous plant remains. The Farlovian delta extended over a wide area, Farlovian strata outcrop in the Forest of Dean, Bristol and the Mendips, and are well represented in South Wales where in the upper part of the succession they interdigitate with marine Devonian beds. More- over, Farlovian deposits are known from boreholes in the North and East Midlands, as well as the North London area where they are overlain by marine Devonian. Thus the Midland block persisted as an area of relative stability undergoing subsidence, its southern margin being delimited by approximately the same line as that of the Dittonian, extending from Southern Ireland, through the Bristol Channel and the London area, into France. Its northern extent is equivocal but may have persisted into North England (Wills, 1951 : pls. 3-5). In the area of the Clee Hills, the Upper Old Red Sandstone was terminated by a period of uplift, erosion and planation prior to the deposition of the basal Carboni- ferous conglomerate and the advance of the sea. VII. LIST OF LOCALITIES, FAUNA AND FLORA In addition to the authors’ localities, this list incorporates what may be termed “classic ’’ localities and many fossiliferous horizons discovered by the late W. W. King. Where no faunal list is given, the material requires further preparation. The national grid reference and 23 inch new series O.S. sheet number are given for each locality, though in the old series the prefix ‘“SO”’ is replaced by ‘‘ 32”’, Le. SO/68 new series = 32/68 old series. Wherever possible the position of each locality is given relative to the main “‘ Psammosteus ’’ Limestone, for which the abbreviation “P”’. L. is used throughout. 1. ABDON BRIDGE SO/58 1,050 yds. E. 6° S. of Tugford Church (5668/8696). Cornstone outcropping in bank of stream. Ditton Series, lower group ; c. 140 ft. above “ P’’. L. Tesseraspis sp. 2. ASTON Hitt Woop 1 SO/69 500 yds. W. 42°S. of Morville Church (6663/9358). Loose blocks in field below wood, not 7 situ. Stratigraphical position uncertain. Pteraspis (Simopteraspis) leathensis White. 3. Aston Hitt Woop 2 SO/69 430 yds. S. 28° W. of Morville Church (6679/9355). Red cornstone exposed in bank. Stratigraphical position uncertain. Pteraspis sp. Poraspis sp. 220 THE OLD RED SANDSTONE OF BROWN CLEE HILL 4. BACHE MILL $0/58 170 yds. W. 39° N. of Bache Chapel (5013/8614). Ludlow Bone Bed. 5. BatcH Brook SO/68 1,530 yds. S. 6° E. of Wheathill Church (6232/8076). Cornstone outcropping in southern bank of stream, a little distance below new footbridge. Ditton Series, lower group ; c. 800 ft. above “ P’”’. L. Pieraspis (Belgicaspis) croucht var. heightingtonensis White. Arthrodire. Thelodont scales. Acanthodian scales. 6. BaucoTT, SANDY LANE $O/58 510 yds. N. 31° W. of Broncroft (5430/8710). Green and red sandstones in old quarry on north side of lane. Downton Series, Red Downton Formation. Indet. spine. 7. BEACONHILL BROOK I SO/69 1,200 yds. N. 41° E. of Monkhopton Church (6330/9426). Purple and green micaceous sandstone in cliff in southern bank of stream, 430 yds. upstream from roadbridge. Downton Series, Red Downton Formation. 8. BESomM FARM QUARRY SO/68 475 yds. E. 29° S. of Besom Farm (6076/8094). Cornstone exposed at top of old quarry, on southern side. Ditton Series, upper group. Pteraspis (Cymripteraspis) leachi White. Protaspis (Europrotaspis) crenulata sp. nov. Benneviaspis salopiensis sp. nov. Kujdanowiaspis anglica (Traquair). Kujdanowraspis willst sp. nov. Wheathillaspis wickhamkingi gen. et sp. nov. Ischnacanthus anglicus sp. nov. Onchus wheathillensis sp. nov. Onchus besomensis sp. nov. Nodonchus bambusifer gen. et sp. nov. Acanthodian tooth (Plectrodus-type). Thelodont scales. Acanthodian scales. Spirorbis sp. Pachytheca sp. TO? 2036 12. 13: I4. ads THE OLD RED SANDSTONE OF BROWN CLEE HILL . BITTERLEY WorKS RAILWAY CUTTING 590 yds. S. 34° E. of Bitterley Church (5740/7685). Red sandstone in southern face of cutting, about 20 yds. from east end. Ditton Series, lower group ; immediately overlying “ P”’. L. Anglaspis sp. Biucks HousE 150 yds. SW. of Blucks House (6794/8957). Sandy cornstone exposed in stream. Ditton Series, lower group. Ptevaspis sp. Cephalaspis cf. langi Stensio. Cephalaspis cf. flettt Stensio. Cephalaspis cf. agassizi Lankester. BOCKLETON BROOK 1,030 yds. S. 13° E. of Stoke St. Milborough Church (5686/8136). Cornstone exposed in stream, 30 yds. NE. of roadbridge. Ditton Series, upper group ; c. goo ft. above “ P”’. L. Protaspis (Europrotaspis) crenulata sp. nov. BouLDON, CLEE BROOK 300 yds. S. 33° E. of Bouldon Church (5484/8500). Cornstone exposed in Clee Brook. Downton Series, Red Downton Formation. BouLpon ForpD 420 yds. E. 25° S. of Bouldon Church (5496/8499). Sandstone and cornstone exposed in stream. Ditton Series, lower group ; c. 50 ft. above “‘ P’”’. L. Pteraspis (Pteraspis) rostrata var. trimpleyensis White. Cephalaspis bouldonensis sp. nov. BROCKTON 160 yds. N. 34° E. from crossroads at Brockton (5789/9389). On east side of road, in bank north of gate leading into farmyard, and in farmyard. Ludlow Bone Bed. BroMDON DINGLE BrRooK 400 yds. S. of Cleedownton (5813/8041). Green sandy shale parting in sandstone-cornstone bluff, 30 yds. downstream from footbridge. Ditton Series, lower group. Pteraspis sp. Poraspis sp. Thelodont scales. Acanthodian scales. Pachytheca sp. SO/68 SO/58 SO/58 SO/58 S0/59 SO/58 222 THE OLD RED SANDSTONE OF BROWN CLEE HILL 16. CLAPGATE COTTAGE QUARRY 1,030 yds. E. 35° N. of Tugford Church (5650/8756). Cornstone in old quarry, 40 yds. from road. Ditton Series, lower group ; c. 110 ft. above “ P’”’. L. Pteraspis cf. (Simopteraspis) leathensis White. Traquairaspis symondsi (Lankester). Anglaspis sp. Tesseraspi1s sp. ? Cephalaspid. Acanthodian scales, spines and jaws. Thelodont scales. 17. CLEE ST. MARGARET 630 yds. N. 38° W. of Clee St. Margaret Church (5609/8480). Green sandstone and cornstone in eastern bank of stream section. Ditton Series, lower group ; c. 675 ft. above “ P”’. L. Pteraspis (Belgicaspis) crouchi var. heightingtonensis White. Anglaspis sp. Cephalaspid. Arctolepid. 18. CoLD WESTON QUARRY 490 yds. W. 8° N. of Cold Weston Church (5474/8300). Old quarry. Ditton Series, lower group ; c. 850 ft. above “ P’’. L. ? Arctolepid. I9. CRIDDON BRIDGE 1,620 yds. E. 36° S. of Upton Cresset Church (6681/9159). Red silty sandstone outcropping in stream. Ditton Series, lower group ; 5 ft. above “ P’”’. L. cf. Kujdanowraspis sp. Modtolopsis sp. 20. DAIRY DINGLE 1,425 yds. E. 3° S. of Neenton Church (6503/8767). Cornstone exposed in stream. Ditton Series, lower group ; c. 800 ft. above “ P’’. L. Pteraspis (Pteraspis) dairydinglensis sp. nov. Acanthodian scales. Thelodont scales. Cephalaspid. Indet. spine. Pachytheca sp. SO/58 SO/58 SO/58 SO/69 SO/68 w2i. 22. 23. 24. 25. 26. THE OLD RED SANDSTONE OF BROWN CLEE HILL DERRINGTON REA BRIDGE 500 yds. SE. of Derrington (6099/9052). Cornstone exposed in stream section, 20 yds. downstream from bridge. Ditton Series, lower group ; c. 500 ft. above “‘ P’”’. L. Pieraspis (Pteraspis) dairydinglensis sp. nov. Thelodont scales. Acanthodian scales. Cephalaspid. DEVIL’s MOUTHPIECE QUARRY 1,850 yds. E. 43° N. of Upper Hayton Church (5303/8210). Sandstone and cornstone in old quarry. Ditton Series, lower group ; c. 400 ft. above “ P’”’. L. DODSHILL BANK 630 yds. S. 23° W. of the Three Horseshoes Inn (5978/8138). Grey and buff silty sandstone in small quarry. Ditton Series, ? upper group. Prototaxites sp. Pachytheca sp. Nematothallus sp. cf. Cooksonta sp. Down Quarry, THE 350 yds. S. 10° E. of The Down (6374/8157). Cornstone exposed in old quarry, largely overgrown. Ditton Series, upper group. Pteraspis ? (Cymripteraspis) sp. Protaspis (Europrotaspis) crenulata sp. nov. EARNSTREY Brook (LEATH 3) 570 yds. N. 11° E. of Earnstrey Hall (5752/8889). Massive cornstone forming waterfall, and grey marl immediately underlying it (see also Wills, 1935 : 427; White, 1946 : 210). Downton Series, Red Downton Formation ; c. 70 ft. below “ P.’’. L. Traquairaspis symondsi (Lankester). Anglaspis maccullought A. S. Woodward. Corvaspis kingt A. S. Woodward. Tesseraspis tessellata Wills. Cephalaspis sp. Onchus sp. Pachytheca sp. EARNSTREY HALL I 510 yds, E. 7° N. of Earnstrey Hall (5787/8825). Red sandstone exposed in stream, 70 yds. downstream from road. Ditton Series, lower group ; c. 400 ft. above “‘ P’”’. L. Pteraspis (Belgicaspis) croucht Lankester. GEOL. 5. 7. 223 SO/69 SO/58 SO/58 SO/68 SO/58 SO/58 21 224 27 28. 29. 30. 31. ga 33- THE OLD RED SANDSTONE OF BROWN CLEE HILL EARNSTREY HALL 2 600 yds. E. 6° S. of Earnstrey Hall (5794/8826). Sandstone exposed immediately below bridge. Ditton Series, lower group ; c. 420 ft. above “ P’’. L. Pieraspis cf. (Belgicaspis) crouchi Lankester. FARLOW Brook BRIDGE 325 yds. N. 32° W. of Farlow Church (6380/8088). Cornstone exposed in bank on south side of lane. Ditton Series, upper group. Pteraspis ? (Cymripterasprs) leacht White. Protaspis (Europrotaspis) crenulata sp. nov. Arctolepid. “ Tchthyodorulite.”’ FARLOW, OLD LANE 160 yds. E. 35° N. of Farlow Church (6408/8071). Sandstones exposed on south side of cutting, about half-way down old lane. Yellow Farlow Sandstone Series. Crossopterygian indet. Fox COVERT 720 yds. E. 20° N. of Wheathill Church (6283/8237). Cornstone poorly exposed in stream. Ditton Series, upper group. FOXHOLE CoPPICE, MONKHOPTON gto yds. E. 25° S. of Monkhopton Church (6180/9308). Cornstone forming waterfall. Downton Series, Red Downton Formation ; 25 ft. below “ P”’. L. Kallostrakon sp. Acanthodian scales, jaws and spines. FOXHOLE COPPICE, NEWTON DINGLE 1,650 yds. W. 9° S. of the Three Horseshoes Inn (5849/8168). Red and green cornstone and sandstone exposed in cliff above east bank of stream. Ditton Series, lower group ; c. 750 ft. above “ P’”’. L. GREAT NORTHWOOD 60 yds. ESE. of Great Northwood (6798/8410). Old quarry. Ditton Series, upper part of lower group. Pteraspis cf. (Belgicaspis) croucht Lankester. SO/58 SO/68 SO/68 SO/68 SO/69 SO/58 SO/68 THE OLD RED SANDSTONE OF BROWN CLEE HILL 34. HEATH QUARRY 230 yds. W. 15° N. of Heath Church (5551/8570). Cornstone exposed at western end of old water-filled quarry. Ditton Series, lower group ; c. 300 ft. above “ P’’. L. Pteraspis (Belgicaspis) crouchi Lankester. ? Weigeltaspis or Tesseraspis sp. 35. Hitts, THE ? 1,230 yds. E. 20° S. of Downton Hall (5386/7895) ; a classic locality, the exact position of which is uncertain. Old quarry. Ditton Series, lower group. Pteraspis (Belgicaspis) croucht Lankester. Cephalaspis sp. Acanthodian. 36. HoPTONGATE 2,040 yds. W. 2° N. of Hopton Cangeford Church (5299/8047). Old quarry. Ditton Series, lower group ; c. ? 700 ft. above “ P’”’. L. Pteraspis sp. Kujdanowiaspis anglica (Traquair). 37. Hupwick DINGLE I 980 yds. S. 39° E. of Monkhopton Church (6314/9274). Thick cornstone immediately overlying the “ P’’. L. at the conflu- ence of Hudwick and Stapeley Dingles. Ditton Series, lower group ; immediately overlying “ P’’. L. Pteraspis ? (Simopteraspis) leathensis White. Traquatraspis symondsi (Lankester). Anglaspis sp. Tesservaspis sp. Corvaspis kingt A. S. Woodward. Ischnacanthus wickhami sp. nov. Acanthodian scales (cf. Gomphodus sp. and Nostolepis sp.) and spine. Thelodont scales. Cephalaspid. Pachytheca sp. 38. Hupwick DINGLE 2 1,700 yds. S. 28° E. of Monkhopton Church (6330/9206). Pavement of cornstone exposed a few yards below confluence of streams. Ditton Series, lower group ; c. 250 ft. above “ P’’. L. Pteraspts sp. 225 SO/58 SO/57 SO/58 SO/69 SO/69 226 THE OLD RED SANDSTONE OF BROWN CLEE HILL ' 39. Hupwick DINGLE 3 SO/69 1,310 yds. S. 34° E. of Monkhopton Church (6323/9243). Purple and green cornstone and sandstone in west bank of stream. Ditton Series, lower group ; c. 80 ft. above “ P’”’. L. Pteraspis (Simopteraspis) leathensis White. 40. Hupwick DINGLE 4 SO/69 1,510 yds. S. 32° E. of Monkhopton Church (6329/9224). Grey-green cornstone outcropping Io ft. above base of thick corn- stone-sandstone series, forming rapids and waterfall. Ditton Series, lower group ; c. 175 ft. above “ P’’. L. Pteraspis (Pteraspis) rostrata (Agassiz) var. indet. Cephalaspid. Thelodont scales. Acanthodian. Pachytheca sp. - 41. Hupwick DINGLE 5 SO/69 2,190 yds. S. 29° E. of Monkhopton Church (6353/9168). Red and green cornstone exposed in stream, 175 yds. upstream from footbridge. Ditton Series, lower group; c. 400 ft. above “ P”’. L. Ptevaspis sp. Acanthodian scales. Pachytheca sp. 42. JUBILEE BROOK I S$O/57 500 yds. W. 35° N. of Upper Ledwyche (5503/7948). Green sandstone exposed in stream. Ditton Series, lower group ; c. 250 ft. above “ P”’. L. Pieraspis (Pteraspis) rostrata var. waynensis White. Poraspis sp. ? Corvaspis sp. Acanthodian spine. 43. JUBILEE BRooK 2 SO/57 460 yds. W. 22° N. of Upper Ledwyche (5501/7938). Stream section. Ditton Series, lower group. 44. KIDNALL GUTTER I SO/58 870 yds. S. 15° W. of Tugford Church (5551/8628). Red cornstone exposed in stream. Ditton Series, lower group; 15 ft. above “ P”’. L. Pteraspis (Pteraspis) rostrata (Agassiz) var. indet. Cephalaspid. Thelodont scales. Acanthodian scales. THE OLD RED SANDSTONE OF BROWN CLEE HILL 227 45. KIDNALL GUTTER IA 875 yds. S. 13° W. of Tugford Church (5554/8627). Cornstone exposed in stream. Ditton Series, lower group ; c. 30 ft. above “ P’’. L. Pieraspis (Simopteraspis) leathensis White. ? Traquatraspis sp. Acanthodian spines. 46. KIDNALL GUTTER 2 SO/58 880 yds. S. 12° W. of Tugford Church (5556/8626). Cornstone exposed in stream. Ditton Series, lower group ; c. 75 ft. above “ P’’. L. Pteraspis (Pteraspis) rostrata cf. var. trimpleyensis White. 47. KIDNALL GUTTER 3 SO/58 890 yds. S. 4° W. of Tugford Church (5567/8625). Cornstone exposed in stream. Ditton Series, lower group ; c. 180 ft. above “ P”’. L. Pieraspis (Belgicaspis) croucht Lankester. Pteraspis ? sp. nov. Cephalaspid. Thelodont scales. Acanthodian scales and spines. 48. KIDNALL GUTTER 4 SO/58 goo yds. S. 2° W. of Tugford Church (5570/8625). Grey, brown and red mottled sandstone with cornstone lenses, exposed in stream section. Ditton Series, lower group ; c. 200 ft. above “ P’’. L. Pteraspis sp. ? Tesseraspis sp. Thelodont scales. Acanthodian scales. Pachytheca sp. 49. LEATH I SO/58 30 yds. E. of Leath Cottage (5846/8980). Deep gulley running parallel with road, on north side of Leath Bank (see also White 1950a : 70, 71). ? Ditton Series, lower group. Pteraspis (Simopteraspis) leathensis White. Cephalaspid. Pachytheca sp. 228 THE OLD RED SANDSTONE OF BROWN CLEE HILL 50. LEDWYCHE BROOK SO/58 1,480 yds. E. 9° S. of Hopton Cangeford Church (5620/8078). Cornstone exposed in west bank of stream, 180 yds. downstream from ford. Ditton Series, lower group ; c. 60 ft. above “ P”’. L. Pievaspis (Pteraspis) rostrata var. trimpleyensis White. Cephalaspid. Acanthodian scales and spine. Thelodont scales. 51. LITTLE OXENBOLD SO/59 540 yds. SE. of Little Oxenbold (5906/9117). Purple and green cornstone and marls exposed in stream section. Downton Series, Red Downton Formation ; 50 ft. below “‘ P’”’. L. Traquairaspis symondsi (Lankester). Anglaspis sp. Corvaspis sp. Tesseraspis sp. Acanthodian. 52. LowE FARM QUARRY SO/68 150 yds. S. 19° W. of Lowe Farm (6310/8034). Cornstone exposed in small overgrown quarry. Ditton Series, ? upper group. Pteraspis sp. ? Arctolepid. 53. LowER INGARDINE FoRD SO/68 780 yds. S. 35° E. of Wheathill Church (6250/8150). Very coarse cornstone exposed in stream. Ditton Series, upper group. Pteraspis (Cymripteraspis) leacht White. Protaspis (Europrotaspis) crenulata sp. nov. cf. Kujdanowaspis anghca (Traquair). Arctolepid. Cephalaspid. Thelodont scales. 54. LYDEHOLE SO/58 1,150 yds. N. 38° E. of Upper Hayton Church (5244/8179). Grey-buff cornstone exposed in stream section. Ditton Series, lower group ; c. 130 ft. above ‘‘ P’”’. L. Pteraspis sp. THE OLD RED SANDSTONE OF BROWN CLEE HILL 229 55. Lye Brook 1 SO/69 1,050 yds. S. 19° E. of Morville Church (6728/9298). Cornstone exposed in stream, forming waterfall. Downton Series, Red Downton Formation ; 125 ft. below “ P”’. L. Traquaivaspis symondsi (Lankester). Cephalaspid. Acanthodian spines. 56. LYE Brook 2 SO/69 1,140 yds. S. 11° E. of Morville Church (6717/9287). Block of green cornstone in stream, uncertain whether 7m situ. Stratigraphical position uncertain, but the locality is below “‘ P’’. L. Pteraspis (Simopteraspis) leathensis White. Acanthodian spine. 57. LYE Brook 3 SO/69 1,100 yds. S. 15° E. of Morville Church (6723/9290). Cornstone outcropping about 14 ft. above waterfall formed by mud- stone and fine calcareous sandstones. Downton Series, Red Downton Formation ; c. go ft. below “ P’”’. L. Traquaivaspis symondsi (Lankester). Thelodont scales. Acanthodian scales. Cephalaspid scales. 58. LYE Brook 4 SO/69 1,180 yds. S. 4° E. of Morville Church (6703/9281). Cornstone outcropping in stream, forming waterfall. Ditton Series, lower group ; 75 ft. above “ P’’. L. Pteraspis (Simopteraspis) leathensis White. Poraspis sp. 59. MIDDLETON PRIORS SO/69 210 yds. N. 9° W. of Middleton Priors School (6235/9049). Cornstone exposed in stream section, 20 yds. above bridge. Ditton Series, lower group ; c. 500 ft. above “ P’”’. L. Pteraspis (Belgicaspis) croucht Lankester. Acanthodian scales. Thelodont scales. 60. Mitt FARM BRIDGE SO/68 210 yds. E. 30° S. of Mill Farm (6408/8127). Cornstone exposed in cutting on east side of road. Ditton Series, upper group. Pteraspis ? (Cymripteraspis) leachi White. Thelodont scales. 230 6I. 62. 63. 64. 65. 66. 67. 68. THE OLD RED SANDSTONE OF BROWN CLEE HILL MuUNSLOW 820 yds. E. 12° N. of Munslow Church (5286/8787). Ludlow Bone Bed. NETCHWOOD COMMON 2,260 yds. S. 13° W. of Monkhopton Church (6202/9110). Cornstone exposed in stream section. Ditton Series, lower group ; c. 400 ft. above “ P’’. L. Pieraspis (Belgicaspis) croucht Lankester. New Buivpinecs A 510 yds. S. 38° E. of New Buildings (5796/9898). Grey-green cornstone exposed at top of gully. Ditton Series, lower group. New Buitpinecs B 500 yds. E. 22° S. of New Buildings (5808/8918). Large blocks of grey-green cornstone in field. Ditton Series, lower group ; blocks not 7m situ. Pteraspis cf. (Belgicaspis) croucht Lankester. New Buiipincs C 475 yds. E. of New Buildings (5810/8934). Grey cornstone outcropping half-way up escarpment. Ditton Series, lower group. New Buitpines D 290 yds. S. 37° E. of New Buildings (5781/8914). Red sandstone exposed in stream near foot of escarpment. Downton Series, Red Downton Formation ; 110 ft. below “ P’”’. L. Indet. spine. New Buirpines E/2 590 yds. S. 17° E. of New Buildings (5782/8882). Red and green sandstone and cornstone exposed at top of stream section. Ditton Series, lower group ; c. 110 ft. above “ P’’. L. Pteraspis (Pteraspis) rostrata (Agassiz). New Bulitpincs E/3 320 yds. S. 34° E. of New Buildings (5784/8912). Cornstone exposed in stream section. Downton Series, Red Downton Formation ; c. 75 ft. below “ P.”’. L. Traquairaspis symondsi (Lankester). Anglaspis sp. Acanthodian spines. SO/38 SO/69 SO/58 SO/58 SO/58 SO/58 SO/58 SO/58 THE OLD RED SANDSTONE OF BROWN CLEE HILL 231 69. NEw BuiLpincs E/3A SO/58 430 yds. S. 35° E. of New Buildings (5789/8904). Cornstone exposed in stream section. Ditton Series, lower group ; 2 ft. above “ P’”’. L. Pteraspis sp. 70. NEw BuILpInGs F SO/58 570 yds. E. 27° S. of New Buildings (5817/8910). Green sandy cornstone exposed in stream section. Ditton Series, lower group ; c. 175 ft. above “ P”’. L. 71. NEw INN I SO/58 1,440 yds. W. 43° S. of Upper Hayton Church (5092/8012). Cornstone forming waterfall in stream, 640 yds. above road. Downton Series, Red Downton Formation ; 16 ft. below “ P’’. L. Pteraspis (Simopterasprs) leathensis White. Traquairaspis symondsi (Lankester). Poraspis sp. Tesseraspis sp. Cephalaspid. Acanthodian spines and teeth. Thelodont scales. Pachytheca sp. 72. NEw INN 2 $O/58 1,260 yds. S. 44° W. of Upper Hayton Church (5099/8012). Cornstone forming waterfall in stream, 700 yds. above road. Ditton Series, lower group ; c. 75 ft. above “ P”’. L. Pteraspis (Simopteraspis) leathensis White. Poraspis sp. Anglaspis sp. Acanthodian scales, spines and teeth. Thelodont scales. Pachytheca sp. 73. NEWTON DINGLE I SO/58 520 yds. W. 40° N. of the Three Horseshoes Inn (5963/8221). Cornstone exposed in eastern bank of stream. Ditton Series, upper group ; c. goo ft. above “ P’”’. L. Pteraspis (Cymripteraspis) sp. Kujdanowiaspis anglica (Traquair). 232 THE OLD RED SANDSTONE OF BROWN CLEE HILL 74. NEWTON DINGLE 2 SO/58 700 yds. N. 9° W. of the Three Horseshoes Inn (5990/8254). Cornstone exposed in stream section immediately below lane. Ditton Series, upper group; c. 950 ft. above “P”’.L. ~ Ptevaspis ? (Cymripteraspis) leacht White. Protaspis (Europrotaspis) crenulata sp. nov. Arctolepid. Acanthodian scales and spine. Thelodont scales. 75. OAK DINGLE SO/58 g20 yds. E. 4° N. of Tugford Church (5656/8712). Red sandstone exposed in stream, 100 yds. upstream from road. Ditton Series, lower group ; c. 125 ft. above “ P”’. L. Pteraspis (Pteraspis) rostrata var. trimpleyensis White. Traquairaspis symondsi (Lankester.) Tesseraspis sp. Cephalaspis sp. Acanthodian spine. Thelodont scale. 76. OLDFIELD, THE LOBBY SO/68 140 yds. S. of The Lobby (6710/8850). Green sandy cornstone exposed in stream. Ditton Series, lower group. Pteraspis (Belgicaspis) croucht Lankester. 77. OLD FORGE QUARRY SO/58 330 yds. S. 35° E. of Bouldon Church (5488/8489). Sandstones and cornstones in old quarry. Ditton Series, lower group. 78. PARK BARN QUARRY SO/58 850 yds. E. 29° N. of Tugford Church (5639/8744). Red cornstone exposed in old quarry, 70 yds. from road. Ditton Series, lower group ; c. 80 ft. above “ P”’. L. Ptevaspis sp. Tessevaspts sp. Cephalaspid scales. Thelodont scales. Acanthodian scales and spines. 79. PARK DINGLE SO/69 2,460 yds. S. 35° E. of Monkhopton Church (6386/9158). Green sandstone and shales with pellet bands, exposed in stream section and forming waterfall. Ditton Series, lower group. THE OLD RED SANDSTONE OF BROWN CLEE HILL 233 ‘80. PARLOUR COPPICE SO/69 660 yds. W. 7° N. of Upton Cresset Church (6500/9253). Cornstone exposed in stream section. Ditton Series, lower group ; c. 400 ft. above “ P”’. L. Tesseraspis sp. ‘81. PostON OLD QUARRY SO/58 1,300 yds. N. 17° W. of Hopton Cangeford Church (5452/8154). Cornstone and sandstone in old quarry. Ditton Series, lower group ; c. 800 ft. above “ P”’. L. Pteraspis sp. Arctolepids. 82. PRESCOTT CORNER SO/68 t10 yds. N. of Prescott Farm (6636/8118). Sandstones exposed in south side of bank at sharp bend in road, 5 ft. above road surface. Yellow Farlow Sandstone Series. Bothriolepis sp. Holoptychius sp. Pseudosauripterus anglicus (A. S. Woodward). ? Eusthenopteron farloviensis sp. nov. Indet. Crossopterygil. 83. PrEscoTT, OLD LANE SO/68 175 yds. S. 15° W. of Prescott Farm (6632/8092). Pellety sandstones exposed below hedge on east side of lane. Grey Farlow Sandstone Series. 84. PRESCOTT, REASIDE SO/68 300 yds. N. 35° W. of Prescott Farm (6622/8129). Cornstone exposed in east bank of river. Ditton Series, upper group; ? c. 1,100 ft. above “ P”’. L. Pteraspis (Cymripteraspts) leacht White. Protaspis (Europrotaspis) crenulata sp. nov. Kujdanowiaspis cf. anglica (Traquair). Prescottaspis dineleyi gen. et sp. nov. Thelodont scales. Pachytheca sp. 85. REA BribGE, DiTToN PRIoRS SO/68 1,650 yds. E. 28° N. of Ditton Priors Church (6216/8985). Loose blocks of cornstone in field. Ditton Series, lower group ; not 7m situ. Pteraspis (Pteraspis) rostrata (Agassiz) var. indet. 86. 87. 88. 89. go. gl. g2. 93- THE OLD RED SANDSTONE OF BROWN CLEE HILL REA Brook, NEw House FARM 160 yds. W. of New House Farm (6375/8810). Cornstone exposed in banks of stream. Ditton Series, lower group ; ? c. 650 ft. above “ P’”’. L. Kujdanowiaspis anglica (Traquair). SHIPTON 160 yds. E. 42° S. of Shipton Church (5630/9176). Exposure in sides of lane. Ludlow Bone Bed. SHORTWOOD 1,300 yds. W. 30° S. of Downton Hall (5179/7870). Old quarry. Ditton Series, lower group. SILVINGTON CHURCH Cornstone exposed on west side of road near to churchyard gate (6215/7984). Ditton Series. SILVINGTON, DRAINAGE GULLY 550 yds. N. 30° W. of Cleeton Court (6094/8002). Cornstone exposed in drainage gully. Ditton Series, lower group; ? c. 575 ft. above “ P”’. L. Pteraspis cf. (Belgicaspis) crouch Lankester. SILVINGTON, OLD QUARRY I 175 yds. N. 6° E. of Silvington Church (6216/8002). Cornstone exposed in old quarry. Ditton Series. Pteraspts sp. SILVINGTON, OLD QUARRY 2 270 yds. NE. of Silvington Church (6231/8003). Cornstone exposed in old quarry. Ditton Series. cf. Climatius sp. SILVINGTON, UPPER QUARRY 1,630 yds. S. 20° E. of Wheathill Church (6272/8076). Massive cornstone, 14 ft. thick, exposed in old quarry. Ditton Series, upper group. Pteraspis ? (Cymripteraspis) leachi White. SO/68 $0/59 50/57 SO/67 SO/68 SO/68 SO/68 SO/68 THE OLD RED SANDSTONE OF BROWN CLEE HILL 94. SILVINGTON, WATERFALL 95. 06. 97. 08. 99- 1,250 yds. W. 5° N. of Silvington Church (6096/7994). Cornstone exposed in north bank of stream, 5 yds. upstream from waterfall. Ditton Series, lower group ; ? c. 600 ft. above “ P”’. L. Pteraspis (Belgicaspis) croucht var. heightingtonensis White. ? Tesseraspis sp. Kujdanowiaspis anglica (Traquair). Arctolepid. Cephalaspid. Acanthodian. STAPELEY DINGLE ggo yds. SE. of Monkhopton Church (6320/9279). Red and green cornstone immediately overlying “ P ’’. L. and form- ing first waterfall above the confluence with Hudwick Dingle. Ditton Series, lower group ; immediately overlying “ P”’. L. STOTTESDON BROOK 1,150 yds. S. of Stottesdon Church (6724/8184). Cornstone exposed in stream. Ditton Series, upper group ; ? c. 1,100 ft. above “ P’”’. L. Pteraspis ? (Cymripteraspis) leachi White. Protaspis ? (Europrotaspis) crenulata sp. nov. ? Kujdanowtaspis sp. Thelodont scale. SUDFORD DINGLE I 780 yds. S. 9° E. of Monkhopton Church (6268/9273). Grey-green sandy cornstone exposed in stream. Ditton Series, lower group ; 3-4 ft. above “ P”’. L. Traquairaspis symondsi (Lankester). SUDFORD DINGLE 2 1,250 yds. S. 11° W. of Monkhopton Church (6235/9230). Pavement of cornstone at confluence of streams. Ditton Series, lower group ; c. 170 ft. above “ P’’. L. Pteraspis (Belgicaspis) crouchi Lankester. TARGROVE DINGLE I gto yds. E. 25° N. of Whitbatch (5251/7780). Red cornstone exposed in stream section, 140 yds. below drive to Downton Hall. Ditton Series, lower group ; 18 ft. above “ P”’. L. Anglaspis sp. SO/67 SO/69 SO/68 SO/69 SO/69 50/57 236 roo. Iot. 102. TOS. 104. 105. THE OLD RED SANDSTONE OF BROWN CLEE HILL TARGROVE DINGLE 2 930 yds. E. 30° N. of Whitbatch (5249/7795). Red and green cornstone and sandstone exposed in stream, 40 yds. below drive to Downton Hall. Ditton Series, lower group. Ptevaspis sp. TARGROVE DINGLE 3 goo yds. E. 36° N. of Whitbach (5243/7803). Red and green cornstone forming waterfall near top of western tributary, 60 yds. above drive to Downton Hall. Ditton Series, lower group ; c. 250 ft. above “ P’”’. L. Pteraspis (Belgicaspis) croucht Lankester. TARGROVE DINGLE 4 920 yds. E. 23° N. of Whitbatch (5253/7787). Cornstone forming waterfall, 160 yds. below drive to Downton Hall. Downton Series, Red Downton Formation ; 6 ft. below “‘ P’”’. L. TARGROVE DINGLE 5 g20 yds. E. 28° N. of Whitbatch (5250/7793). Cornstone forming waterfall, 78 yds. below drive to Downton Hall. Ditton Series, lower group. TARGROVE DINGLE 6 1,030 yds. E. 13° N. of Whitbatch (5266/7775). Red calcareous mudstone outcropping in stream section 330 yds. below drive. Downton Series, Red Downton Formation ; c. 100 ft. below “ P’’. L. Kallostrakon sp. “ Onychodus.”’ Acanthodian. Anaspid. TARGROVE QUARRY 1,000 yds. E. 31° N. of Whitbatch (5254/7799). Old quarry on north side of drive to Downton Hall. Ditton Series, lower group; c. 175 ft. above “ P”’. L. Pteraspis (Pteraspis) rostrata var. trimpleyensis White. Traquairaspis symondsi (Lankester). Acanthodian scales. Thelodont scales. SO/57 SO/57 SO/57 SO/57 50/57 S0/57 THE OLD RED SANDSTONE OF BROWN CLEE HILL 237 to6. THRIFT, THE SO/58 goo yds. S.27° E. of Clee St. Margaret Church (5682/8363). Cornstone exposed in west bank of road, 200 yds. south of Burnt House. Ditton Series, upper group ; c. 1,300 ft. above “ P’”’. L. Pteraspis 2? (Cymripteraspis) leachi White. 107. UPPER LEDWYCHE, OLD QUARRY SO/57 600 yds. W. 11° N. of Upper Ledwyche (5488/7933). Cornstone outcropping in old quarry. Ditton Series, lower group ; ? c, 200 ft. above “ P’”’. L. Pteraspis sp. 108. UPPER OVERTON QUARRY SO/68 too yds. N. 30° W. of Upper Overton (6647/8688). Cornstones in quarry, now overgrown and in cutting on south side of quarry. Ditton Series, upper group ; ? c. goo ft. above “ P’’. L. Pteraspis ? (Cymripteraspis) leacht White. Protaspis (Europrotaspis) crenulata sp. nov. Overtonaspis billballi gen. et sp. nov. ? Kujdanowiaspis sp. ? Wheathillaspis sp. Cephalaspid. Thelodont scales. Acanthodian scales. 109. UPTON CRESSET QUARRY SO/69 150 yds. N. of Upton Cresset Church (6560/9259). Purple cornstone exposed in old quarry. Ditton Series, lower group ; c. 400 ft. above “ P”’. L. Pteraspis cf. (Pteraspis) dairydinglensis sp. nov. Cephalaspid scales and fragments. Acanthodian scales and spine. Thelodont scales. 110. UpToN LoDGE QUARRY SO/69 1,360 yds. S. 12° E. of Upton Cresset Church (6586/9124). Cornstone exposed in old quarry. Ditton Series, lower group. Indet. spine. I11. WALLSBATCH QUARRY SO/68 1,300 yds. E. 35° N. of Faintree Hall (6720/8965). Cornstone exposed in old quarry. Ditton Series, lower group. 238 THE OLD RED SANDSTONE OF BROWN CLEE HILL I12. WALTON BROOK I SO/68 710 yds. S. 31° E. of Prescott Farm (6670/8051). Sandy cornstone exposed in stream. Grey Farlow Sandstone Series. Holoptychius sp. 113. WALTON BROOK 2 SO/68 800 yds. S. 37° E. of Prescott Farm (6680/8048). Grey calcareous sandy pebbly cornstone exposed in stream. Grey Farlow Sandstone Series. 114. WALTON QUARRY SO/68 175 yds. N. 15° W. of Walton (6757/8138). Pellety sandstone at foot of quarry face. Ditton Series, upper group; immediately beneath lower Abdon Limestone ; c. 1,450 ft. above “ P’”’. L. Arctolepid. Acanthodian scale. 115. WHITBATCH LANE SO/57 200 yds. E. 12° N. of Whitbatch (5193/7759). Exposure in roadside at sharp angle in Jane. Downton Series, Red Downton Formation. Cephalaspis sp. 116. WHITBATCH QUARRY SO/57 300 yds. NE. of Whitbatch (5195/7776). The fauna listed probably comes from more than one locality, but mainly from old quarry. Ditton Series, lower group ; c. 200 ft. above “ P”’. L. Pteraspis (Pteraspis) rostrata (Agassiz). Cephalaspis fletti Stensi6. Cephalaspis whitbatchensis Stensio. Cephalaspis whiter Stensio. Cephalaspis lankestert Stensio. Cephalaspis acutirostris Stensi6. Cephalaspis ? sp. nov. 117. WILDERNESS SO/58 500 yds. E. 35° S. of Ashfield (5911/8917). Cornstone exposed in stream, 40 yds. south of roadbridge. Ditton Series, lower group; c. 500 ft. above “ P’”’. L. THE OLD RED SANDSTONE OF BROWN CLEE HILL 239 118. WINTERBURN BRIDGE SO/68 950 yds. NW. of Upper Overton (6592/8740). Cornstone exposed below bridge. Ditton Series, ? top of lower group. Pteraspis sp. Cephalaspid. Acanthodian scales. Thelodont scales. 119. WitcHcoT-PosTon DINGLES SO/58 1,120 yds. N. 39° W. of Hopton Cangeford Church (5421/8124). Sandstone and cornstone exposed in scar on east bank, near con- fluence of Witchcot and Poston Dingles. Ditton Series, lower group ; c. 800 ft. above “ P”’. L. Pteraspis sp. 120. WOODHOUSE QUARRY SO/57 ? 1,400 yds. W. 16° N. of Downton Hall (5157/7967). Cornstones exposed in old quarry. Ditton Series, lower group. 121. YEWTREE DINGLE SO/69 1,200 yds. W. 22°S. of Morville Church (6596/9348). Green calcareous sandstone and cornstone, exposed amongst roots of tree at head of dingle. Ditton Series, lower group ; c. 50 ft. above ‘‘ P’”’. L. Pteraspis (Simopteraspis) leathensis White. VIII. ACKNOWLEDGEMENTS It is with great pleasure that we tender our thanks to friends and colleagues at the Universities of Birmingham and Exeter, King’s College, London, and the British Museum (Natural History), for assistance in a variety of fields during the many years we have been engaged on this project, which was commenced during the tenure of research grants from the D.S.I.R. We wish to acknowledge our especial indebtedness to Prof. L. J. Wills, Dr. E. I. White, Mr. H. A. Toombs and the late W. N. Croft for their unfailing help and encouragement. Finally, we should like to pay tribute to the pioneer work of the late William Wickham King, who for over forty years devoted his leisure time to the study of the Anglo-Welsh Old Red Sandstone, and who so generously put at the disposal of all subsequent workers a great wealth of unpublished information. It is to his memory that we re- spectfully wish to dedicate this paper. GEOL. 5. 7. 22 240 THE OLD RED SANDSTONE OF BROWN CLEE HILL IX. REFERENCES ALLEN, J. R. L. 1960. Cornstone. Geol. Mag. Lond., 97 : 43-48. Batt, H. W. 1951. The Silurian and Devonian rocks of Turner’s Hill and Gornal, South Staffordshire. Proc. Geol. Ass. Lond., 62 : 225-236, figs. 1-3. Bat, H. W. & DINELEY, D. L. 1952. Notes on the Old Red Sandstone of the Clee Hills. Pvoc. Geol. Ass. Lond., 63 : 207-214, fig. 1. Crort, W. N. 1953. Breconian: a stage name of the Old Red Sandstone. Geol. Mag. Lond., 90 : 429-432. Crort, W. N. & Lance, W. H. 1942. The Lower Devonian Flora of the Senni Beds of Monmouth- shire and Breconshire. Philos. Tvans., London (B) 231 : 131-163, pls. 9-11. Denison, R. 1956. A review of the habitat of the earliest vertebrates. Fieldiana, Chicago, 11 > 359-457. Dinetey, D. L. 1951. The northern part of the Lower Old Red Sandstone Outcrop of the Welsh Borderland. Tvans. Woolhope Nat. Fld Cl., Hereford, 33 : 127-147, figs. 1-3. 1953. Notes on the genus Corvaspis. Proc. Roy. Soc. Edinb. (B) 65 : 166-181, pls. 1, 2. DineLey, D. L. & GossaGE, D. W. 1959. The Old Red Sandstone of the Cleobury Mortimer Area, Shropshire. Pvoc. Geol. Ass. Lond., 70 : 221-238, figs. 1-8. Dixon, E. E. L. 1921. The Geology of the South Wales Coalfield. Part XIII. The country around Pembroke and Tenby. Mem. Geol. Surv. Gt. Britain, London (n.s.). Sheets 244 and 245 : 220 pp., 5 pls. . Du Tort, A. L. 1954. The Geology of South Africa. 3rd ed. xiv + 611 pp., 41 pls., 1 map. Edinburgh and London. Earp, J. R. 1938. The higher Silurian rocks of the Kerry District, Montgomeryshire. Quart. J. Geol. Soc. Lond., 94 : 125-160, pls. 12, 13. 1940. The Geology of the South-Western Part of the Clun Forest. Quart. J. Geol. Soc. Lond., 95 : 1-12, pl. 1. EpmunpDs, F. H. & OaKLeEy, K. P. 1947. British Regional Geology : The Central England District. 2nded. 80pp. Dep. Sci. Industr. Res., Geol. Surv. Mus., London. EceErton, P. M. DE G. 1862. Description of a new Pterichthys. Quart. J. Geol. Soc. Lond., 18 : 103-108, pl. 3. ELLEs, G. L. & SLATER, I. L. 1906. The highest Silurian rocks of the Ludlow District. Quart. J. Geol. Soc. Lond., 62 : 195-221, pl. 22. Eyes, V. A. 1953. District Report: Wales and West Midlands District. Summ. Progr. Geol. Suv. Gt. Britain, London, 1952 : 23-26. FLEET, W. F. 1925. The chief heavy detrital minerals in the rocks of the English Midlands. Geol. Mag. Lond., 62 : 98-128, pls. 4, 5. GrorGE, T. N. 1939. The Cefn Bryn Shales of Gower. Geol. Mag. Lond., 76 : 1-6. 1956. The Namurian Usk Anticline. Proc. Geol. Ass. Lond., 66 : 297-314, 4 figs. Gross, W. 1950. Die palaontologische und stratigraphische Bedeutung der Wirbeltierfaunen des Old Reds und der marinen altpalaozoischen Schichten. ) 289 Kujdanowtiaspis anglica (Traquair) : : : . 289 Kujdanowtaspis willsi sp. nov. . : c = 290 Genus Overtonaspis nov. . : 4 F 6 5 20)5} Overtonaspis billballi sp.nov. . 6 2 0 Zo! Genus Prescottaspis nov. . : ° : : = 294) Prescotiaspis dineleyi sp. nov. . . ° : 297) Genus Wheathillaspis nov. é ¢ . 297 Wheathillaspis wickhamhingi sp. nov.. : 3 - 300 Undetermined Arctolepids : : 0 : : - 300 III. FAuNA OF THE FARLOW SANDSTONE SERIES : 2 2 a - Or Class GNATHOSTOMATA. ‘ ; : ; : ; ESOL Order Antiarchi é 6 a é . 4 . 301 Genus Bothriolepis Eichwald 2 : : : . . 301 Bothriolepis macrocephala Egerton . : é - 301 Order Crossopterygii : é : : : : 0 3 302 Family Holoptychiidae c ; : 3 : : - 302 Genus Pseudosauripterus nov. . : . 302 Pseudosauripterus anglicus (A. S. Woodward) 4 =) 805 Family Rhizodontidae . ; : : : : - 305 Genus Eusthenopteron Wir rece : 4 5 - 305 Eusthenopteron farvloviensis sp. nNOv. . é ; - 305 Order Dipnoi : . 5 ; é : : ¢ ee S07 IV. ACKNOWLEDGEMENTS . 2 : : ° - ° c SOF; V. REFERENCES 6 ; ; : : 0 4 : c . 308 THE OLD RED SANDSTONE OF BROWN CLEE HILL 245 I. FAUNA OF THE UPPER PART OF THE RED DOWNTON FORMATION AND THE LOWER GROUP OF THE DITTON SERIES UNDER this heading are grouped all the elements found in the zones of Traquairaspis symondst, Pteraspis leathensis and P. croucht, since, as Drs. Ball and Dineley have explained in Part Iof the paper (pp: 201-202) most elements of the T. symondsi Zone pass, although in greatly reduced numbers, into the succeeding zone and, from the systematic viewpoint, may therefore be conveniently dealt with together. Throughout the text, the heights given are relative to the main “‘ Psammosteus ”’ Limestone. The numbers in parentheses refer to registered speciemens in the British Museum (Natural History) collections. Class AGNATHA Order HETEROSTRACI Genus KALLOSTRAKON Lankester, 1870 Kallostrakon sp. LOCALITIES AND MATERIAL. Fragments probably of this genus have been found at Foxhole Coppice, Monkhopton (25 ft. below “ Psammosteus ’’ Limestone) and at Targrove Dingle 6 (100 ft. below “‘ Psammosteus ’’ Limestone). Genus TESSERASPIS Wills, 1935 Tesseraspis tessellata Wills Tesseraspis sp. LOCALITIES AND MATERIAL. Except at the type locality, Earnstrey Brook (Leath 3: 70 ft. below “ Psammosteus”’ Limestone), the material is extremely fragmentary, although widely distributed. Specimens have been recorded from the following localities : Abdon Bridge (140 ft. above ‘“‘ Psammosteus’’ Limestone) ; Clapgate Cottage Quarry (110 ft. above ‘“ Psammosteus’’ Limestone) ; ? Heath Quarry (300 ft. above ‘‘ Psammosteus”’ Limestone) ; Hudwick Dingle 1 (immediately above “ Psammosteus’’ Limestone); ? Kidnall Gutter 4 (200 ft. above “‘ Psam- mosteus’’ Limestone) ; Little Oxenbold (50 ft. below ‘‘ Psammosteus’”’ Lime- stone) ; New Inn 1 (16 ft. below ‘“‘ Psammosteus’’ Limestone) ; Oak Dingle (125 ft. above ‘‘ Psammosteus’’ Limestone) ; Park Barn Quarry (80 ft. above “ Psammosteus’”’ Limestone) ; Parlour Coppice (400 ft. above ‘“‘ Psammosteus ”’ Limestone) ; ? Silvington, Waterfall (600 ft. above ‘‘ Psammosteus’”’ Lime- stone). The range is from 50 ft. below the “‘ Psammosteus’”’ Limestone to 400 ft. (? 600 ft.) above it, the higher records being based on uncertain fragments, possibly in some cases of Wezgeltaspis. 246 THE OLD RED SANDSTONE OF BROWN CLEE HILL Genus CORVASPIS A. S. Woodward, 1934 Corvaspis kingi A. S. Woodward LOCALITIES AND MATERIAL. Apart from the type-locality, Earnstrey Brook (Leath 3: 70 ft. below “ Psammosteus’’ Limestone), the material is very frag- mentary. Single fragments of this or a related species have been found at Hudwick Dingle 1 (immediately above ‘‘ Psammosteus’’ Limestone) and Little Oxenbold (50 ft. below ‘‘ Psammosteus’””’ Limestone). A doubtful specimen is recorded from Jubilee Brook (250 ft. above ‘‘ Psammosteus ’’ Limestone). The range of this species is from 70 ft. below to ? 250 ft. above the ““ Psammosteus ”’ Limestone. Genus ANGLASPIS Jaekel, 1927 Anglaspis macculloughi (A. S. Woodward) LOCALITIES AND MATERIAL. The only extensive material comes from Earnstrey Brook (Leath 3: 70 ft. below ‘‘ Psammosteus’”’ Limestone), where it is very fine and plentiful, but rare and fragmentary specimens from the following places are almost certainly conspecific: Clapgate Cottage Quarry (110 ft. above ‘‘ Psammos- teus’’ Limestone) ; Clee St. Margaret (675 ft. above ‘‘ Psammosteus ’”’ Limestone) ; Hudwick Dingle 1 (immediately above “ Psammosteus’’ Limestone) ; Little Oxen- bold (50 ft. below “‘ Psammosteus’”’ Limestone) ; Targrove Dingle 1 (18 ft. above “ Psammosteus’”’ Limestone) ; New Buildings E/3 (75 ft. below “ Psammosteus ”’ Limestone) ; New Inn 2 (75 ft. above “‘ Psammosteus ’”’ Limestone). The range of this species in this area is therefore from 75 ft. below to 675 ft above the ‘‘ Psammosteus ’’ Limestone. Genus PORASPIS Kiaer, 1930 Poraspis sp. LOCALITIES AND MATERIAL. All the specimens are small rare fragments of shields. Specimens are recorded from Aston Hill Wood 2 (relationship to ‘‘ Psammosteus ”’ Limestone unknown) ; Bromdon Dingle Brook (Lower Dittonian) ; Jubilee Brook 1 (250 ft. above “‘ Psammosteus’’ Limestone) ; Lye Brook 4 (75 ft. above ‘‘ Psam- mosteus’’ Limestone); New Inn 1 (16 ft. below “ Psammosteus’’ Limestone) ; New Inn 2 (75 ft. above “‘ Psammosteus ’’ Limestone). Therefore the local range is from 16 ft. below to 250 ft. above the “‘ Psammosteus ”’ Limestone. Genus TRAQUAIRASPIS Kiaer, 1932 Traquairaspis symondsi (Lankester) LOCALITIES AND MATERIAL. Apart from the classical locality of Earnstrey Brook (Leath 3: 70 ft. below ‘“‘ Psammosteus’’ Limestone), about 30 good fragments and a fine dorsal disk (P.31146) from Lye Brook 1 (125 ft. below ‘‘ Psammosteus ’’ Lime- THE OLD RED SANDSTONE OF BROWN CLEE HILL 247 stone) ; a good ventral plate (P.34345) from Lye Brook 3 (go ft. below “‘ Psam- mosteus ’’ Limestone) and another (P.30026) from Little Oxenbold (50 ft. below “ Psammosteus ’’ Limestone), the material consists of small rare fragments: Clap- gate Cottage Quarry (110 ft. above ‘‘ Psammosteus ”’ Limestone) ; Hudwick Dingle 1 (immediately above ‘‘ Psammosteus’’ Limestone) ; Kidnall Gutter 14 (30 ft. above “ Psammosteus’”’ Limestone) ; Lye Brook 1 (125 ft. below “‘ Psammosteus ’’ Lime- stone) ; New Buildings E/3 (75 ft. below “‘ Psammosteus’”’ Limestone) ; New Inn 1 (16 ft. below ‘“‘ Psammosteus’”’ Limestone) ; Oak Dingle (125 ft. above “‘ Psam- mosteus’’ Limestone) ; Sudford Dingle 1 (3 to 4 ft. above “‘ Psammosteus’”’ Lime- stone) ; Targrove Quarry (175 ft. above ‘‘ Psammosteus’”’ Limestone). The total range in the area is from 125 ft. below to 175 ft. above the “ Psam- mosteus’’ Limestone, but apart from that from Hudwick Dingle 1 and Sudford Dingle 1, where the fossiliferous beds lie just over the “‘ Psammosteus ’”’ Limestone, the total material from above the ‘‘ Psammosteus’”’ Limestone consists of ten very small fragments ; three from Clapgate Cottage Quarry (110 ft. above) ; one from Oak Dingle (125 ft. above) and six from Targrove Quarry (175 ft. above). Genus PTERASPIS Kner, 1847 Subgenus PTERASPIS typicum Pteraspis (Pteraspis) rostrata (Agassiz) LOCALITIES AND MATERIAL. The surviving type-specimen (Agassiz, 1935 : 148, pl. 1, fig. 6) is a small unsatisfactory specimen, 5:2 cm. long as preserved, largely an internal cast. Only the base of the snout is preserved as an impression, and this and the form of the cornua establish the species, but not the variety. The speci- men came from Whitbatch, Shropshire, a locality in which several quarries were worked from time to time (about 200 ft. above the “‘ Psammosteus ’’ Limestone), and from this area material of Pteraspis is rare and Cephalaspis is relatively common. The Pteraspis material comprises ventral disks 6-7 cm. long (P.43448-51) and an even larger fragment (P.43453), and the posterior end of a shield; but none is sufficient to indicate the particular form, even if such indirect argument were in the circumstances acceptable. Only nine or ten other localities in the Clee area have produced rostra to establish the occurrence of the species, and in two cases these are unusually broad. Doubt- less much of the material identified merely as Pteraspis sp. belongs to P. rostrata. Specimens assigned to undetermined forms in addition to those from Whitbach are : Hudwick Dingle 4. Three rostra and a fragment (175 ft. above “‘ Psammosteus ”’ Limestone). Kidnall Gutter 1. A rostrum (P.34328) and a ridge-scale (P.34327). The rostrum is exceptionally broad. (15 ft. above ‘“‘ Psammosteus ’’ Limestone.) New Buildings E/2. One very broad rostrum (P.34336) which is not, however, very different from a diseased specimen from Guildings Brook (P.16472). (i110 ft. above ‘‘ Psammosteus ’”’ Limestone.) Rea Bridge, Ditton Priors. Good rostrum (P.31640). (Not 2” situ.) 248 THE OLD RED SANDSTONE OF BROWN CLEE HILL Pteraspis (Pteraspis) rostrata var. waynensis White (Text-fig. 1) Jubilee Brook 1. Two rostra with part of orbitals, etc. attached (P.30053, P.30056). The inner anterior borders of the orbitals are concave and the medial extensions meet the pineal plate. The second specimen is 2-7 cm. in length with a maximum breadth of 2-4 cm. (Text-fig. 1). (250 ft. above “ Psammosteus ’’ Limestone.) Fic. 1. Pteraspis (Ptevaspis) rostvata var. waynensis White. Rostrum with part of left orbital plate. Jubilee Brook 1. P.30056. x 1°5. Fic. 2. Ptevaspis (Ptevaspis) rostrata var. trimpleyensis White. Anterior part of dorsal shield, in impression. Bouldon Ford. P.29955. x 1°5. Pteraspis (Pteraspis) rostrata var. trimpleyensis White (Pl; 34, fig, 52 Pl 35, tgs. 1, 20 lext-fig.2) Bouldon Ford. External impression of part of dorsal shield with abnormal sensory canals (P.29955; Text-fig. 2) and other fragments. (50 ft. above “ Psammosteus ’’ Limestone.) Oak Dingle. Right half of small dorsal disk (P.33573-74; Pl. 35, fig. 2), rostrum with parts of orbitals (P.33572) and other fragments. (125 ft. above “ Psam- mosteus ’’ Limestone.) Ledwyche Brook. Anterior part of dorsal shield with rather short and broad rostrum, 2 x 1°8 cm. (P.33766; Pl. 35, fig. 1). (50 ft. above ‘‘ Psammosteus ”’ Limestone.) Targrove Quarry. Greater part of old dorsal shield (45963) and several frag- ments almost certainly of this form. (175 ft. above ‘‘ Psammosteus’”’ Lime- stone.) THE OLD RED SANDSTONE OF BROWN CLEE HILL 249 REMARKS. The ranges shown are as follows : Pteraspis (P.) rostrata var. indet. . . 15-200 ft. above “ Psammosteus ”’ Limestone. Pteraspis (P.) rostrata var. waynensis . 250 ft. above “ Psammosteus”’ Limestone. Pteraspis (P.) rostrata var. trimpleyensis . 50-175 ft. above “ Psammosteus ” Limestone. Pteraspis (Pteraspis) dairydinglensis sp. nov. (Pl. 33; Pl. 34, figs. 1-4; Pl. 35, figs. 4,5; Pls. 37-41; Text-figs. 3-5) Diacnosis. A typical Pteraspis with dorsal shield (without spine) attaining a length of 10 cm., strongly vaulted, with maximum breadth of disk when flattened about four-fifths of its total length; anterior margin of disk with wide, shallow median indentation and usually rounded antero-lateral corners ; posterior margins long and gently concave when uncrushed; spinal socket about one-third of total length of disk and two-thirds of short compressed triangular spine. Rostrum very short with somewhat angular tip, its length being about one-quarter of the total length of the dorsal shield (without spine) and four-fifths of the distance between the orbital openings. Pre-oral field very short. Pineal plate much wider than long, moderately large and meeting medial extensions of orbital plates except superficially in some old specimens. Branchial plates wide and longitudinally keeled, the flat or slightly concave upper and lower surfaces divided by an acute angle of about 65°. Cornual plates rather narrow, reaching level of centre of ossi- fication of dorsal disk, with concave posterior margin or short hook when abraded. Ventral disk of usual ovoid shape, much vaulted with rounded or briefly flattened anterior margin and obtuse posterior angle partly overlapped by first ventral ridge- scale. Ridges of ornamentation + 60 per cm., gently rounded in section when fresh. Ornamentation on scales consisting of longitudinal ridges divided by chevron- shaped grooves at intervals slightly increasing to rear, rarely with anterior ridges parallel with front margin. HoLotyPe. Imperfect dorsal shield (P.29936; Text-fig. 3). MATERIAL. The material is extensive but although mostly fragmentary is well preserved. In addition to the holotype it includes half-a-dozen imperfect dorsal shields, five ventral disks, a dozen rostra, over 50 other substantial fragments of various parts including a number of oral and circum-oral plates, and more than 100 isolated scales. LocaLiTiEs. Bulk of material from Dairy Dingle, East of Neenton ; also from Derrington Rea Bridge, North of Ditton Priors, and a few pieces and scales, almost certainly of this species, from Upton Cresset. The most complete shields come from the Hopton Brook section near Hopton Wafers, and a snout from Hazeley Brook West, outside the area studied. 250 THE OLD RED SANDSTONE OF BROWN CLEE HILL The position of the Clee localities are: Dairy Dingle (800 ft. above ‘‘ Psam- mosteus’’ Limestone); Derrington Rea Bridge (500 ft. above “‘ Psammosteus”’ Limestone) ; Upton Cresset (400 ft. above ““ Psammosteus ” Limestone). DescripTion. The dorsal shield (P.42366, Pl. 33, fig. 2) from Hopton Brook is by far the most complete specimen and lacks only the left anterior region. Like all the other dorsal shields it is partly flattened and so appears considerably broader Ptevaspis (Ptevaspis) dairydinglensis sp. nov. Fic. 3. Imperfect dorsal shield. The holotype. Dairy Dingle. P.29936. xX 1:2. Fic. 4. Dorsal spine. Derrington Rea Bridge. P.42005. x 1-6. than in life. The region around the socket of the dorsal spine is naturally thick and thus reinforced does not crush so readily and still retains much of its original shape, whereas the rostrum is flattened. As preserved the total length, including the socket, is 9:3 cm., the breadth 6-2 cm. flat at the maximum (just in front of the branchial opening) and about 7-2 cm. over the curve. THE OLD RED SANDSTONE OF BROWN CLEE HILL 251 In this specimen the rostrum is short and is a little less than one-quarter of the length of the shield. It usually appears to be very broad owing to flattening (PI. 34, figs. 1, 3; Text-fig. 6), and the measurements in various specimens are : Max. breadth Length Over curve Specimen Locality (cm.) (cm.) P. 42366 : Hopton Brook é 23 3°4* P. 42361 - A . 2°5 3°2 P.42360 3 hy 2:0 3:2 P.32249 . Derrington 1-9 2:6 P.32248 , 2°5 3°71 P.29997 : Dairy Dingle 2:0 3:0* * Double half breadth. The real form of this plate, which is quite high, is only seen in P. 42360 of which the maximum breadth when viewed directly is 2-7 cm., 0-5 cm. less than over the curve (Pl. 33, fig. I). Isolated rostra or fragments of them are exceptionally common, especially the solid tips, for there is usually a large space above the pre-oral field occupied in life by cartilage. The undersurface of the snout is very like that of P. rostrata (Pl. 34, figs. Ia, 5), only much shorter, especially in the younger specimens, in one of which (Pl. 34, fig. 2) a very marked median spur is preserved. There are twelve isolated oral or tooth-plates excellently preserved (P.43709-20 ; Pl. 37). As in P. rostrata (White, 1935: 408, text-figs. 41-47) they are flat and wide proximally, high and narrow distally. The median oral plate (or plates) is very short and cut off from the post-oral covers by those on either side (Pl. 37, fig. 1). It consists chiefly of the large head, resembling a triangular arrow-head, which greatly overhangs the smooth inner surface. The neighbouring tooth-plates are longer and curved, but the plates in- crease in length and become straighter towards the corners of the mouth, and at the same time the heads become smaller. The outer or lower surface is ornamented, so are all the sides of the heads (cf. Stensid, 1958 : 265). The plates overlapped one another decreasingly outwards, the area of overlap being wide in the middle of the series, but at the sides the lateral face on the outermost plate has an almost vertical contact with the fixed lateral oral plate. All but the short median plate were overlapped proximally by the post-oral covers, and the inner surface, by which each plate was doubtless attached to a cartilaginous framework, is smooth except for a few pores. The ornamentation of the heads of these plates consists of numerous slightly raised triangles, none of which shows any sign of wear. On the outer surface the dentine ridges are short and transverse, with occasionally a few longitudinal ridges along the overlapped margin. Ventrally the ornament sometimes shows signs of having been rubbed by movement against the post-oral covers (Pl. 37, fig. 10). The smooth basal layer of the inner surface is thin and shows a number of openings for vascular canals. A careful re-examination of the oral plates of P. rostrata toombs: (White, 1935 : 408, 252 THE OLD RED SANDSTONE OF BROWN CLEE HILL text-figs. 41-48) shows that they are in fact similar in shape and disposition to those described here. Neither in that form nor in P. dairydinglensis are there any signs of the false upper dentures with which Stensi6 (1958 : 350, text-figs. I9gIA, B, etc.) has invested the pteraspids, and their supposed presence is against all the evidence. In the former species where the plates are all preserved in place as moulds such a feature would certainly be apparent and in the case of the new species, where very delicate plates are preserved as complete entities there is no sign of them; more- over the tips of the dental plates show no signs whatever of wear, as would be bound to appear if they worked against an upper set. They must have had some degree of movement or the mouth could have opened but relatively little, possibly no more than the areas of overlap indicate ; and since these are unilateral towards the out- side, with the largest in the middle and below, the movement must have been towards forming a scoop, that is with the centre pushed forwards relatively to the sides. Pteraspis probably fed on organic matter off the bottom in muddy waters, as already suggested (White, 1935: 412), but the vaulted ventral and dorsal plates clearly indicate that they were free swimming, as the hypocercal tail also shows (Kermack 1943), although not very efficiently so. It is possible that the more strongly vaulted forms such as P. dairydinglensis were also surface-feeders, as Harris (1936) has suggested in the case of the more evenly balanced Anaspids. Such forms, which usually appear to be wider and flatter than the others owing to crushing, always have short snouts and therefore small pre-oral fields, and since these are sensory areas connected with food-finding, it is probable that their restriction in size does go with surface-feeding where feeding would be more nearly mechanical or based partly on sight, and where a relatively broader mouth would be useful. The long, broad-snouted pteraspids, like P. (Cymripteraspis) leacht and P. (Rhinopteraspis) dunensis, usually large animals with even smaller pre-oral fields were possibly also surface—or even mid-water feeders living in stronger currents, sometimes under estu- arine conditions, but P. (Belgicaspis) croucht with a narrow typically rounded snout and large pre-oral field was probably again a bottom-feeder in cloudy water, as its frequent association with P. (P.) rostrata indicates. It is perhaps not without significance that only once has P. (P.) datrydinglensis been found with another species, P. (B.) crouchi, and it is possible that the appearance of broad, long-snouted forms in the higher beds is connected with changed environmental conditions. It seems likely that the extreme rarity of pteraspids in the Scotch Lower Old Red Sandstone is due to the lacustrine conditions there prevalent, for these would be less suitable for free-swimming forms like pteraspids with their ability, albeit limited, to live in current-dominated areas, such as rivers and estuaries, than for the wholly bottom-dwelling cephalaspids, whose flat underside and dorsal eyes clearly show their usual method of living on the bottom, preferably of quiet waters. The dorsal disk has the usual outline, heart-shaped in front and pointed behind. The anterior indentation is wide to accommodate the short, wide pineal plate, and the front corners are rounded. The median length (including spinal socket) of the most complete specimen (PI. 33, fig. 1) is 6-9 cm., the maximum breadth over the curve, 6cm. The hinder part of the plate is much vaulted (Pl. 33, fig. 3) and the hinder margin on each side of the spine was in life nearly straight. THE OLD RED SANDSTONE OF BROWN CLEE HILL 253 The dorsal spine was clearly short and deep (Text-fig. 4; P.32243) but although the specimen figured has little missing, the tip is rubbed and may have been some- what more pointed. The orbitals are rather short but wide, with their medial extensions usually in full contact with the very wide and short pineal plate (PI. 33, figs. 1, 2; Pl. 34, figs. if 3 4). Fic. 5. Ptevaspis (Pteraspis) dairydinglensis sp. nov. Imperfect dorsal shield. Hopton Brook 4. P.42364. x 1:2. The branchial opening is at about two-thirds of the distance from the anterior margin to the posterior outer corner of the disk, so that the branchial plates are relatively long (Pl. 33, fig. 1). The plates have a very marked keel dividing the almost flat upper and lower surfaces (Pl. 33, figs. 4, 4a) which are separated by about 65°. The cornual plate is narrow but increases in width to the rear, ending in a slight flat hook, which is usually rubbed off (Pl. 33, figs. 2, 4, 5). The ventral disks (Pl. 35, figs. 4, 5) show no special features. They were un- doubtedly very convex in life, like the dorsal shield, and consequently appear short and broad when flattened. The transverse front margin is very short and the posterior point blunt with its two sides slightly concave. Owing to the nature of the matrix, which disintegrates in acetic acid, a very large 254 THE OLD RED SANDSTONE OF BROWN CLEE HILL number of isolated scales were obtained (Pl. 39, figs. 9-25; Pls. 40, 41). Their variation in shape, size and function conforms to the usual pattern in Pteraspis (see White, 1935 : 414, 410, pl. 27, text-figs. 51-65 ; 1950: 81, text-figs. 10-13) with the several kinds of ridge-scales, flank-scales, double flank-scales and so on. One point not previously observed is the piercing of scales for the continuation of the sensory canals from the carapace on to the hinder part of the body; this can be seen in some seventeen specimens (PI. 39, figs. 9-25). The canal entered the scale in the anterior overlapped area and left it in the short overlapping margin on the inner surface some distance from the hinder margin. There was no external opening through the scale. This explains why indications of the lateral lines are so rare, even in the complete specimens with tails, which are preserved as external impres- sions. In these only one shows substantial parts of this system on the squamation, parts of both left and right dorsal lines appearing as infillings of the canals (White, 1935 : 42, text-fig. 64). There were three pairs of lateral lines, as the dorsal and ventral disks show, a dorsal line on each side of the dorsal ridge-scales, the main dorsal lateral line running low on the upper surface just above the level of the cornua, and a ventro-lateral line just below the level of the cornua and branchial plates. Of the extent of these lines we only know that the dorsal line ran almost to the caudal constriction at least. It is difficult to ascertain to which of these lines the isolated scales belong, in particular, as between the two lateral canals, but the smaller scales, especially the more oblique ones, probably belong to the dorsal line. Nor can one be certain of the distribution of all the ridge-scales. The long scales with little anterior overlap are almost certainly dorsal ridge-scales from the separated anterior series, while the more compressed scale (No. 32) is from the pre-caudal row. The broad flattish scales (P.43621I-23) are ventrals, and so are the two very different simply orna- mented scales (Pl. 40, figs. 10, 11) characteristic of the anterior ventral ridge- series. The predominantly diamond-shaped body-scales show a remarkable range in form, and include the usual double scales (Pl. 40, figs. 14-27; Pl. 41). The ornamentation of the scales is typical of the genus, but varied in detail. In general it resembles that of P. (P.) rostrata trimpleyensis, without anterior ridges parallel with the front margin, and with the chevron-grooves more or less evenly distributed on the body of the scales, except on the ridge-scales. On the anterior ventral ridge-scale they are entirely absent, and on the others the intervals separating them increase rapidly towards the rear. There are, however, a number of flank scales in which the anterior ridges appear in varying numbers and in some they are as numerous as in those of the typical scales of P. (P.) rostrata toombsi (White, 1935, pl. 27). The area of overlap is variable but usually narrow. A curious feature of these scales in general is the frequency with which the orna- mentation is carried over on to the underside at the free posterior end. It suggests that there were considerable pockets of epidermis and that the scales did not lie flat but were raised like spent pine-cones. This is also shown in the American P. carmani (Denison, 1960 : 396, fig. 141J) and is probably common to all species. Like the scales and oral plates, a number of circum-oral plates have been isolated from the matrix, all but one from Dairy Dingle (Pl. 38; Pl. 39, figs. 1-8). THE OLD RED SANDSTONE OF BROWN CLEE HILL 255 There are five examples of anterior lateral plates (Pl. 38, figs. 4, 5). They are distinguished at once by the branched sensory canal system which consists of a large, short canal near one side with a single smaller branch nearly at right angles to it. The main canal is a segment of the ventral longitudinal canal which runs from the ventral disk through the lateral plates on to the dorsal side of the snout. The plates are wider than long, the smaller one figured being irregularly squarish, the larger almost L-shaped. These two are clearly from opposite sides of the animal, and both show an overlapped projection to one side. The smaller lateral canal opens towards the curved longer margin which is feather-edged with the ornament invading the undersurface. Clearly this must be the free side along the ventral disk and thus, the overlapped projection must be for the fixed lateral oral plate which itself is overlapped along the rest of the contact. There is no branch in this plate to the orbital behind the orbit, so that the infraorbital sensory canal must have been dis- continuous. Four out of seven posterior lateral plates are figured (Pl. 39, figs. 1-4). They are in general considerably longer than wide with the sides roughly parallel. Their distinguishing feature is the single sensory canal, which runs down their length, usually forming a broad ridge on the undersurface. Like all these circum-oral plates, the shape is varied and they tend to fuse to the sides of the shield. One such plate shows that the basal and polygonal layers were completely fused to the shield, but that the outer layer was free (cf. White, 1938, text-figs. 2-4). The nine fixed lateral oral plates (Pl. 38, figs. 1-3) are roughly triangular, some- times with the apex truncated, in which case there will have been a smaller tri- angular plate such as in P. (P.) rostrata toombsi (White, 1935, text-figs. 43, 45). The single canal crosses the plate but probably missed the small anterior plate when separate. The contacts of the larger plate are deep and nearly vertical. This plate sometimes fuses to the oral margin of the rostrum forming a marked “angle préorogoniale ” (Pl. 34, fig. 1a) such as is said to characterize certain of Stensi6’s (1958, text-fig. 139, etc.) forms. The post-oral covers (Pl. 39, figs. 5-8) are amazingly variable in shape, and the number also must have been variable in order to accommodate such odd shapes, most of which show overlapping sides. All are marked by concentric ornamentation and the absence of sensory canals. It is interesting to note that none of the lateral plates nor the post-oral covers shows any sign of internal ornamentation other than occasionally at the margin, as in some scales. This is in contrast to the condition described by Stensi6 (1958 : 265, text-figs. I40B, I9gIA, B) on which he bases the presence of his ‘‘ fente post- orogoniale ”’, Remarks. P. (P.) dairydinglensts in all its known features is just an exaggerated, vaulted, short-snouted edition of P. (P.) rostrata from which it was undoubtedly derived. Isolated fragments of plates or isolated scales would often be difficult or impossible to identify with one species or the other, and this provides a warning in accepting past identifications based on such material. The vaulted shape, as noted above, suggests that this species may have been a freer swimmer than P. (P.) rostrata and a surface rather than a bottom-feeder. GEOL. 5. 7. 23 256 THE OLD RED SANDSTONE OF BROWN CLEE HILE The range of P. (P.) dairydinglensis is 400 to 800 ft. above the “‘ Psammosteus ”’ Limestone ; approximating to the upper part of the croucht Zone. Subgenus BELGICASPIS Zych, 1931 Pteraspis (Belgicaspis) crouchi Lankester (Pl. 36, figs. 2-11) LOCALITIES AND MATERIAL. Batch Brook. Small imperfect dorsal disk, 3-1 cm. long and about 2 cm. broad, with large pineal indentation (P.33799). Near var. heightingtonensis. (800 ft. above ‘“‘ Psammosteus ’’ Limestone.) Clee St. Margaret. Numerous pieces, chiefly fragments. One narrow dorsal disk 4:2 cm. long and 2-4 cm. broad (P.32171). Near var. heightingtonensis, but sides not so straight as in holotype. There are seven remarkable rostra, all short and triangular (e.g. P.29833-35, P.32164-66, P.32179; Pl. 36, figs. 6-8). (675 ft. above “ Psammosteus ’’ Limestone.) Earnstrey Hall 1. Several fragments including good small normal rostrum and a dorsal spine (P.27067—68) and a normal small rostrum and one broader on same slab (Pl. 36, figs. 4, 5), the measurements being 1:3 x 0:35 cm. and I'5 X 0-6 cm. respectively. (400 ft. above “ Psammosteus ’’ limestone.) Earnstrey Hall 2. Fragment probably of this species. (420 ft. above “ Psam- mosteus ’’ Limestone.) Great Northwood. Fragment probably of this species. (Upper part of lower group.) Heath Quarry. Small typical rostrum; a very imperfect dorsal disk over 5:5 cm. long and 4-0 cm. wide (P.29847) ; a good branchial (PI. 36, fig. 2) about 4:5 cm. long when complete with the undersurface having a maximum breadth of 0-9 cm.; and other fragments. (300 ft. above “‘ Psammosteus’’ Lime- stone.) The Hills, Downton. Isolated typical rostrum, and dorsal and ventral disks. (Position uncertain.) Kidnall Gutter 3. A typical branchial (P.43222) and other fragments. (About 180 ft. above “‘ Psammosteus ”’ Limestone.) Middleton Priors. Six rostra, one ventral disk 7-6 cm. long, etc. (500 ft. above “ Psammosteus ”’ Limestone.) New Buildings B. One large dorsal disk 5-5 cm. long and about 4 cm. wide. (Loose block.) Oldfield, the Lobby. Rostrum etc. (Position uncertain.) Silvington, Drainage Gully. Typical pieces of disk. (575 ft. above ‘‘ Psam- mosteus ’’ Limestone.) Silvington, Waterfall. Very numerous specimens. The dorsal disks range in size from 5:5 X 3°6 cm. (P.29877) through 4:0 xX 2°6 cm. (P.29045) down to 2:8 x 2 cm. (P.29881), approaching var. heightingtonensis in form. The rostra show a remarkable range in size and shape (Pl. 36, figs. 9-11), the largest THE OLD RED SANDSTONE OF BROWN CLEE HILL 257 (P.29043) is normal in shape, the middle sizes are triangular (P.29832, P.29041, P.29830) and the smallest are a little longer than wide (P.29823-24, P.29831, P.29042). (About 600 ft. above “ Psammosteus ’’ Limestone.) Sudford Dingle 2. Small narrow rostrum (P.34332) and other fragments. (About 170 ft. above “‘ Psammosteus ’’ Limestone.) Targrove Dingle 3. Part of large narrow rostrum (P.43023), etc. (About 250 ft. above “‘ Psammosteus ’’ Limestone.) Remarks. The known range of this species is very different from that of P. (P.) rostrata, starting at about 170 ft. above the ‘“‘ Psammosteus’’ Limestone it reaches 800 ft., so that in this area, the two species only overlap at the very bottom of its range ; but in fact they have never been taken together in the Clee district, although they certainly occur together elsewhere. The most interesting feature of this species is the occurrence of very short rostra in several localities. They undoubtedly belong to P. (B.) crouch and are usually associated with others of normal shape. They all show the single cross-over of the ridges of the ornamentation from the upper to the lower surface typical of this species, which clearly distinguishes them from rostra of nearly similar shape of P. (P.) rostrata. In the latter the ridges form a Z-pattern along the sides (cf. Pl. 35, fig. r and Pl. 36, figs. 6-11). These rostra cannot represent growth-stages, for some of the short and long rostra belonged to animals of comparable size (Pl. 36, figs. 4, 5), and in any case, one rostrum could not develop from another except by complete resorption, of which I have seen no evidence. It is also unlikely that they represent sexual dimorphism, although a similar phenomenon is known in P. (P.) rostrata, for there is con- siderable variation in the form of the short snouts and there may be gradations between the two types. The explanation of the phenomenon is probably quite different. P.(B.) crouchi isa very specialized isolated form, particularly in respect of its snout. When typically de- veloped it is unlike that of any other species of Ptevaspis. The rostrum of Pteraspis in the earlier forms, was short and broad, and among such were the ancestors of P. (B.) croucht. It is probable that the spasmodic occurrence of these unusual stout and broad snouts both in this species and in P. (P.) rostrata is due to a degree of ‘atavism’ (de Beer, 1951: 104), possibly brought about by genetic drift in small, relatively isolated populations. The short snouts have not been found below 400 ft. above the ‘‘ Psammosteus ”’ Limestone, but the amount of material in the lower beds is not significant. Subgenus SIMOPTERASPIS White, 1950 This subgenus was founded on P. Jeathensis White (1950 : 76) and also was held to include P. gosseleti Leriche, P. primaeva Kiaer and P. vogti Kiaer, all small short- snouted forms. Of these only in P. leathensis (Pl. 35, fig. 3a) and P. primaeva is the important pre-oral region known and thus the relationships of P. leathensis to P. gosseleti which Leriche (1924: 149, footnote) has made the type a subgenus of 258 THE OLD RED SANDSTONE OF BROWN CLEE HILL Protopteraspis on the supposition that the spine was not separated from the dorsal disk, is uncertain and they are for the time being retained in the subgenera of which they are the respective types. Pteraspis (Simopteraspis) leathensis White (Pl. 35, figs. 3, 34) LOCALITIES AND MATERIAL. This species is recorded from eleven localities. Aston Hill Wood 1. External impression of whole dorsal shield and spine (P.25253). (Loose blocks from unknown level.) Clapgate Cottage Quarry. Fragment possibly referable to this species (P.37776). (110 ft. above ‘‘ Psammosteus ’’ Limestone.) Hudwick Dingle 1. Poor fragments, probably of this species. (Immediately overlying “‘ Psammosteus ’’ Limestone.) Hudwick Dingle 3. External impression of anterior half of large dorsal shield (P.42834-35), ventral disk (P.34042), etc. (80 ft. above “ Psammosteus ”’ Limestone.) Kidnall Gutter 14. Typical fragments including part of rostrum (P.43109) and disk (P.43110). (30 ft. above “‘ Psammosteus ’’ Limestone.) Leath 1. The type locality. (Level unknown, but ?Ditton Series, lower group). Lye Brook 2. Fragments probably of this species. (Position uncertain.) Lye Brook 4. Imperfect dorsal shield (P.26927), etc. (75 ft. above “‘ Psam- mosteus ”’ Limestone.) New Inn i. Over 40 specimens finely preserved, mostly fragmentary, but in- cluding fine dorsal shields (P.34120, Pl. 35, fig. 3; P.34076), dorsal spine (P.34256) and ventral disks (P.34851). (16 ft. below ‘‘ Psammosteus”’ Lime- stone.) New Inn z. A dozen finely preserved typical fragments and scales (P.34518-27, P.43014, (75 ft. above “ Psammosteus ’’ Limestone.) Yewtree Dingle. Characteristic ventral disks (e.g. P.26937-38); rostrum (P.26942) and other fragments. (50 ft. above “ Psammosteus’’ Limestone.) RemarRKS. P. (S.) leathensis ranges from 16 ft. below to 110 ft. above the “ Psam- mosteus ’’ Limestone. Order ANASPIDA A single scale (P.28665) is recorded from Targrove Dingle 6 (100 ft. below “‘ Psam- mosteus ’’ Limestone). Order THELODONTI Thelodont scales have been noted from about 20 localities in the lower group, ranging from Lye Brook 3 (g0 ft. below the ‘‘ Psammosteus ’’ Limestone) to Batch Brook and Dairy Dingle (800 ft. above). Almost invariably they are accompanied by Acanthodian scales. It is probable that their presence is universal, and the THE OLD RED SANDSTONE OF BROWN CLEE HILL 259 restricting factor is the ease with which the matrix yields to acid treatment. At the moment insufficient material has been recovered to make a close study worth while. Order OSTEOSTRACI Genus CEPHALASPIS Agassiz, 1835 Cephalaspis acutirostris Stensi6 Locatity. Whitbatch Quarry (about 200 ft. above “‘ Psammosteus’’ Lime- stone). Cephalaspis cf. agassizi (Lankester) Locatity. Bluck’s House (Ditton Series, lower group), centre of shield (P. 33752). Cephalaspis cf. fletti Stensi6 Locatity. Bluck’s House (Ditton Series, lower group), centre of shield (P.33749- 50). Cephalaspis cf. langi Stensid Locatity. Bluck’s House (Ditton Series, lower group), a small badly preserved head shield (P.33751) measuring 3-7 cm. across base of cornua, with rounded rostral margin. Cephalaspis lankesteri Stensié Locatity. Whitbatch Quarry (about 200 ft. above “ Psammosteus ’’ Limestone). Cephalaspis whitbatchensis Stensié ’ LocaLity. Whitbatch Quarry (about 200 ft. above “‘ Psammosteus”’ Lime- stone). Cephalaspis whitei Stensid Locatity. Whitbatch Quarry (about 200 ft. above “ Psammosteus ’’ Limestone). Cephalaspis bouldonensis sp. noy. (Pl. 36, fig. 1; Text-fig. 6) Diacnosis. A Cephalaspis of moderate size, with maximum breadth of cephalic shield across top of pectoral sinuses, approx. 7 cm. when complete, and probably about one and one-third times median length. Rostral margin broad and rounded, the shield being semicircular. Length of cornua not known ; pectoral sinuses wide 260 THE OLD RED SANDSTONE OF BROWN CLEE HILL and shallow. Interzonal part short, broad and high. Orbital openings small in middle of length of cephalic shield. Dorsal sensory field, bottle-shaped, two and one-third times as long as broad at the maximum. Lateral sensory fields reaching backwards on to cornua. Ho.otyPe. Imperfect cephalic shield in part counterpart (P.43055). LocaLity. Bouldon Ford (about 50 ft. above “ Psammosteus ’’ Limestone). Fic. 6. Cephalaspis bouldonensis sp. nov. Imperfect cephalic shield partly restored fromcounterpart. (a) rightside view. See also Pl. 36, fig.1. Theholotype. Bouldon Ford. P.43054-55. X 1°5. DESCRIPTION. This specimen is in many respects badly preserved, but sufficient is present to show its unique character. The exoskeleton is thin and in poor condi- tion, the greater part of the cornua and the middle of the interzonal part are missing, the right margin defective and the whole somewhat distorted. The shape of the shield is remarkable, for the front margin is completely rounded, and indeed the whole forms a semicircle with the centre just behind the posterior end of the dorsal sensory field. The maximum breadth across the pectoral sinuses (making allowance THE OLD RED SANDSTONE OF BROWN CLEE HILL 2601 for the defective right side) is 7 cm. and the length, as preserved, is 4:5 cm.; the height at the back of the shield without the spine (which was certainly present) is I*3 cm. with a straight facial profile. The specimen is probably not very much crushed. The orbits lie approximately in the middle of the known length of the shield measured to the hinder lateral margin of the interzonal part. The sinuses are very wide without an angle and shallow. The dorsal sensory field is narrow in front and then quickly broadens, reaching its maximum breadth near the rounded posterior end. The normal lateral sensory fields can be clearly made out and they run onto the cornua as far as these are preserved. Nothing can be seen of the ornamentation. REMARKS. This is one of the most distinctive of the English species, by reason of its completely semicircular shape. C. langi Stensi6 (1932: 134), the only other completely rounded form, is not only much smaller, but its eyes are further forward and the interzonal part relatively wider. INDETERMINABLE CEPHALASPIDS Other remains of cephalaspids are widely spread although rare and fragmentary. Pieces of the cephalic shields or scales occur in the following localities : Clee St. Margaret (675 ft. above ‘“‘ Psammosteus ’’ Limestone). Dairy Dingle (800 ft. above ‘“‘ Psammosteus ’’ Limestone). Derrington Rea Bridge (500 ft. above ‘‘ Psammosteus ”’ Limestone). Earnstrey Brook (70 ft. below “‘ Psammosteus ”’ Limestone). The Hills (Ditton Series, lower group). Hudwick Dingle I (immediately over “‘ Psammosteus ’’ Limestone). Hudwick Dingle 4 (175 ft. above ‘‘ Psammosteus’” Limestone). Kidnall Gutter 1 (15 ft. above ‘‘ Psammosteus ’’ Limestone). Kidnall Gutter 3 (180 ft. above “‘ Psammosteus ’’ Limestone). Leath 1 (? Ditton Series, lower group). Ledwyche Brook (60 ft. above ‘‘ Psammosteus ’’ Limestone). Lye Brook 1 (125 ft. below ‘“‘ Psammosteus ”’ Limestone). Lye Brook 3 (go ft. below ‘“‘ Psammosteus ’’ Limestone). New Inn 1 (16 ft. below ‘“‘ Psammosteus ’’ Limestone). Oak Dingle (125 ft. above ‘“‘ Psammosteus ’’ Limestone). Park Barn Quarry (80 ft. above “‘ Psammosteus ’’ Limestone). Silvington, Waterfall (600 ft. above ‘‘ Psammosteus ”’ Limestone). Upton Cresset Quarry (400 ft. above “‘ Psammosteus ’’ Limestone). The range is therefore from 125 ft. below to 800 ft. above the ““ Psammosteus ”’ Limestone. Class GNATHOSTOMATA Order ACANTHODII Acanthodian scales are closely associated with thelodont scales (q.v., p. 258) in 262 THE OLD RED SANDSTONE OF BROWN CLEE HILL their known distribution, and the same remarks apply regarding their probable universality. Fin-spines of several types (Onchus, Climatius, etc.) are also fairly widespread but few in number, about 35 distributed over 15 localities. All are either very frag- mentary or small. Most are probably referable to acanthodians. At both the New Inn localities partly nodular spines of the type ascribed to “ Climatius’”’ by Gross (1947, pl. 27) have been found (P.43008-12, P.34284) and from New Inn 2 comes also some of ‘‘ Onchus’”’ type (P.43003-07). In only four localities have teeth or pieces of jawbone been found. From New Inn 2 (75 ft. above the ‘‘ Psammosteus ’’ Limestone) come several curious “‘ hedge- hog ”’ multiple tooth spirals (P.42991-99) similar to those described and figured by Gross (1957; 5, pl. 3, figs. 1-7) under the name “ Gomphodus’’ from the Upper Ludlow “ Beyrichienkalk’’’, only much larger, up to 6 mm. across without the spines ; and from New Inn 1 (16 ft. below the “‘ Psammosteus ”’ Limestone) comes an even larger example, 7 mm. across (P.34287). Still larger tooth spirals, I cm. across (P.29709-12) have been collected at Onen, Monmouthshire in association with Traquairaspis symondst. Genus ISCHNACANTHUS Powrie, 1864 Two fragments of acanthodian jawbone have been collected in the area, one from Foxhole Coppice, Monkhopton, and the other from Hudwick Dingle 1. About the former nothing further can usefully be said, but the latter (P.29725; Pl. 42, figs. 3, 3a) is a fragment of the dentigerous margin of the lower jaw of the type usually referred to ‘‘ Ischnacanthus kingi”’ and is recorded under that genus by Dineley (1953 : 167). The name “ [schnacanthus kingi”’ has long been familiar to workers on the Lower Old Red Sandstone of the Welsh Borders, and a practice has grown up of labelling thus all fragments of big acanthodian jaws, and occasionally the term has appeared in print in faunal lists (King, 1934: 540; White, 1945: 212); but no such species has been described, and it is now clear that more than one species is involved. The first mention of it occurs in 1917 (King & Lewis, 1917 : 97) where it is stated in a footnote that it was “since described in Geol. Mag., dec. 6, vol. IV (1917), p. 74’. This reference is to a brief note by A. S. Woodward on “ Plectrodus”’, on the second page of which he thanks ‘“‘ Mr. Wickham King for the opportunity of examining the microscopical structure of an unusually large specimen, probably of a new species, which he has recently found in the Downtonian of Baggeridge, S. Staffordshire’. No specific name is mentioned here and “ [schnacanthus kingt”’ remains a nomen nudum. However, there are many interesting specimens so labelled in Wickham King’s collection and they and others subsequently collected are described below as belonging to two species. Ischnacanthus kingi sp. nov. (Pl. 42, fig. 4; Text-fig. 7) Diacnosis. An IJschnacanthus of very large size with massive jaws greatly THE OLD RED SANDSTONE OF BROWN CLEE HILL 263 thickened along the upper and lower margins. Teeth acuminate and generally round in section; the smaller teeth relatively large, about one-half or two-thirds the size of the larger teeth. Horotype. Imperfect lower margin of palato-quadrate (P.12202) from Bag- geridge (zone I.2 of Wickham King). LOCALITIES AND MATERIAL. The holotype and a second specimen from Baggeridge (zone I.2 of Wickham King) are all that are definitely assigned to this species. The palato-quadrate margin of a small fish from the Ludlow Bone-bed at Brockton, Shropshire (P.33879) and a fragment of a small jaw-bone from Man Brook 2 (P.17395) are rather similar. None is known from the Clee area. DESCRIPTION AND REMARKS. Clearly this species must be based on the specimen from Baggeridge (P.12202) to which King first applied the name and on which Fic. 7. Ischnacanthus kingi sp. nov. Part of dentition of upper jaw. Baggeridge Colliery, South Staffordshire. The holotype. P.12202. 4:5. Woodward (1917) wrote his note. It is a very large bone that formed the lower border of the palato-quadrate and measures Io cm. in length, although lacking a substantial part of both ends. It is 0-8 cm. in depth and shows the lingual side. Unfortunately the tooth-bearing surface is very imperfect and the dentition is repre- sented by the remains of two large teeth, the broken bases of five others and a number of small teeth or their bases (Text-fig. 7). The most complete tooth is about 0-5 cm. high, acuminate and nearly round in section. The sides show fine vertical ridges. The bases of some of the other large teeth are more oval and between them are the remains of much smaller sharply pointed round teeth, most of which line the vertical flange on the outer side, now broken away. None has a pulp-cavity but in cross-section they show a finely reticular pattern with numerous openings of canaliculi on the surface, exactly like the jaw-bone itself from which they arise gradually. They are apparently out- growths of the bone-substance, like arthrodire teeth, and not true teeth ankylosed to the jaw. The second specimen from Baggeridge Colliery is an almost complete right lower jaw of a much smaller fish (Pl. 42, fig. 4). It is just under 7 cm. long and 1-8 cm. at its deepest. Watson’s (1937: 79, pl. 9, figs. 3a, 3b) description of the same bone in the very much smaller J. gracilis applies very closely except that this bone shows the outer instead of the inner side of the jaw. The thickened lower margin is deep laterally and so is the upper dentigerous margin. The latter is twisted and almost 264 THE OLD RED SANDSTONE OF BROWN CLEE HILL vertical at the articular end, thus forming a groove for the passage of powerful jaw muscles which were accommodated in the hollow between the two ridges, but in this specimen the thin wall connecting them has been lost. Beyond the point where the two ridges meet there is a long anterior beak-like prolongation, triangular in section and hollow, the “ anterior ossification ’’ of Watson, which is externally continuous with the rest of the jaw. The tooth-bearing area is very short, but not quite as short as shown in the figure, since a part behind, as far as the vertical crack, has been broken away. The teeth are of two sizes, as before, and are part of the vertical outer flange. They are narrow and acuminate, but show signs of wear at the tips. The species is not easily defined in distinction from J. gracilis, apart from its great size, since so little is known of it, but the difference between the large and smaller teeth seems to be less here than in the type-species, although the degree of wear may be misleading. Fic. 8. Ischnacanthus wickhami sp. nov. Median spiral tooth-roll. Man Brook 7, Trimpley. PEI 397-es 5 Ischnacanthus wickhami sp. nov. (Pl. 42, figs. 1-3; Text-fig. 8) Diacnosis. An Ischnacanthus of great size, with the teeth laterally compressed and sometimes much elongated. Hototypre. Part of right palato-quadrate margin with teeth (P.24625) from Gardener’s Bank, Cleobury Mortimer. LOCALITIES AND MATERIAL. The only Clee locality is Hudwick Dingle 1. The great mass of the material is from Man Brook 6 and 7 near Trimpley (King’s zone 1.6). Only single specimens have been obtained from comparable strata elsewhere, the holotype from Gardener’s Bank ; P.17409 from near Abberley, Worcestershire, and P.23817-18 from Hay Castle, Brecon. The species would appear to be rather later in range than J. kingi. Description. All the specimens of jaws labelled “ Jschnacanthus kingi”’ by Wickham King collected from other localities than Baggeridge Colliery (and pre- sumably subsequently to his collection from that locality) show a very different type of dentition in which the teeth are laterally compressed and sometimes elongated into long shearing teeth. This is best seen in the holotype, a fragment from the right side, from Gardener’s Bank, (not collected by King) which has_ been removed completely from the matrix by acetic acid (Pl. 42, figs. 2, 2a). The smaller LHE OLD RED SANDSTONE OF BROWN CLEE HILL 205 teeth are very irregular and include the shearing teeth and, as in the preceding species, form part of the upturned lateral margin of the smooth horizontal palato- quadrate shelf which increases in width as it goes forward. As Watson (1937 : 80) states of J. gracilis, the series of smaller teeth is “ interrupted by the large teeth which are rooted on the surface of the jaw and rise from it with their outer surfaces in complete continuity with the labial flange’’. Most of the specimens of this species are jaw fragments of various sizes and were collected by King at Man Brook near Trimpley. They are very imperfect and many show a remarkable degree of wear, the teeth often being reduced to mere stumps (Pl. 42, fig. 1), so it is clear that they were not replaceable and the life of the fish depended on their durability. Among the remains of the lower jaw is the only specimen from the Clee district, part of the oral margin of the left ramus from a relatively small fish (Pl. 42, figs. 3, 3a). All the smaller teeth on the labial flange are laterally compressed and variable in size; while of the two larger teeth, one is much wider than the other. The oral shelf widens rapidly forwards and bears an irregular series of denticles internally to the main teeth. The bone shows on its outer side the marked oblique groove noted in J. kingi above. Associated with the jaw-bones at Man Brook is a large symphysial spiral (Text- fig. 8) closely resembling that on the lower jaw in Watson’s (1937, text-fig. 11) restoration, and half a dozen smaller ones, as well as a large curved anterior dorsal spine (P.14444) about 9°5 cm. long and parts of another (P.16555). Smooth convex scales are commonly associated with the other parts of the fish. REMARKS. As already noted these specimens are very like the comparable parts of J. gracilis described by Watson (1937). The teeth in their various forms closely resemble those on the minute jaws from the Upper Ludlow “ Beyrichienkalk ”’ figured by Gross (1957: 13, pl. 5, figs. 2-9) as ‘‘ Nostolepis”’, a genus based on scales supposedly of coelolepid origin (Pander, 1856 : 68, pl. 6, fig. 7), but the asso- ciated symphysial spirals are very different. The rounded form of some of the teeth described by Gross (1947 : 148, pl. 9, fig. 3), so apparent in those from Man Brook, seems to be due to wear. Order ARTHRODIRA Rare specimens that have been referred to arthrodires have been identified from six or seven localities. They are all small fragments of plates or spines that cannot be assigned to a particular genus, although most are clearly arctolepid, with three exceptions which are all referable to Kujdanowtaspis anglica. The localities for the indeterminable remains are : Batch Brook (800 ft. above “‘ Psammosteus ’’ Limestone). Clee St. Margaret (675 ft. above “‘ Psammosteus ’’ Limestone). ? Cold Weston Quarry (850 ft. above ‘“‘ Psammosteus ’’ Limestone). Criddon Bridge (5 ft. above “‘ Psammosteus ’’ Limestone). Poston Old Quarry (800 ft. above ““ Psammosteus ”’ Limestone). 266 THE OLD RED SANDSTONE OF BROWN CLEE HILL Fics. 9-14. Kujdanowiaspis anglica (Traquair). Outlines of six skull-roofs to show range in variation of component plates. All nat. size. (9) P.27550, Silvington Waterfall. (10) No. 42147, Cradley, Herefordshire. (11) Geol. Surv. No. 4740, “‘ Here- ford’’ (probably Heightington). (12) No. 38032, Heightington, Worcestershire. (13) P.29417, Cradley (cast of holotype in Royal Scottish Museum, No. 1926. 57-2.) (14) Geol. Surv. No. 4417, Heightington, Worcestershire. THE OLD RED SANDSTONE OF BROWN CLEE HILL 267 Genus KUJDANOWIASPIS Stensio, 1942 Kujdanowiaspis anglica (Traquair) (Pl. 45, fig. 7; Text-figs. 9-15) LOCALITIES AND MATERIAL. Three specimens only may be certainly attributed to this species: the external impression of a head-shield from Silvington, Water- fall (600 ft. above ‘‘ Psammosteus’”’ Limestone) ; the internal impression of the left anterior quarter of the ventral armour from Hoptongate (700 ft. above “ Psam- mosteus’’ Limestone) and an anterior ventro-lateral plate from Rea Brook, New House Farm (650 ft. above “ Psammosteus ’”’ Limestone). DESCRIPTION. The impression of the skull (P.27550; Text-fig. 9) is finely pre- served and on it may be traced the outlines of the component bones and the course of the sensory canals. It may be compared with the sketches of the other known skull-roofs of this species (Text-figs. 10-14) which show the great variation in the shape of the individual plates (see also Text-figs. 42, 43). The second specimen, from Hoptongate (P.27097; Text-fig. 15) shows, mostly as internal impressions, the left anterior ventro-lateral plate, the anterior median ventral plate and part of the inter-lateral plate with coracoid process of a young fish. The whole is only 2:8 cm. long, and the tubercles of the ornamentation are correspondingly fine. It is very much as in Stensid’s (1944: 60, text-fig. 178) restoration of a Podolian form. The last, from Rea Brook, New House Farm, Neenton, is a very well preserved triangular piece consisting of part of the right anterior ventro-lateral plate with the spinal plate and a very small fragment of the inter-lateral plate (P.29821 ; Pl. 45, fig. 7). It measures 3-7 cm. across and 2:8 cm. from the anterior margin to the Ibe CO. / Fic.15. Kujdanowiaspis anglica (Traquair). Left anterior quarter of the ventral armour of a small fish, largely an internal cast. AMV, anterior median ventral plate. Co, coracoid. IL, interlateral plate. LAVL, left anterior ventro-lateral plate. Hopton- Sate wk 270072 268 THE OLD RED SANDSTONE OF BROWN CLEE HILL pectoral sinus. Apart from the pectoral sinus and the contact with the spinal plate, the margins are missing, but the spinal plate is complete except for the free portion. In no important feature does it differ from specimens from elsewhere, but the ornamentation, developed by acid, is very clear. On the AVL plate the rounded tubercles tend to run in lines very roughly parallel with the margins. Most are somewhat rubbed but it can be seen that they were rather small and low with crinkled bases. The tubercles along the free lateral margin of the spinal plate are larger and taller, and their size and arrangement in numerous parallel longitudinal rows mark them off from the ornamentation on the undersurface. REMARKS. All these specimens may be considered typical of this as yet in- differently described species, which is widely spread, although generally rare, in corresponding strata in Herefordshire. Arthrodires range throughout the whole of the lower group of the Ditton Series, extending into the middle of the upper group, i.e. from 5 ft. to 1,100 ft. above the “ Psammosteus”’ Limestone; and K. anglica is known from 600 ft. to 1,100 ft. above the ‘“‘ Psammosteus ’’ Limestone. it, BAUNA OF DHE UPPER GROUP OB DHE DITLON Si RES The most striking feature of the fauna of these upper beds is the high incidence of new forms, ten out of twelve named species, four of them here referred to new genera. On the other hand Kwjdanowiaspis anglica is a rather unexpected sur- vival from the lower beds, where it is widely distributed along the Welsh Marches, although rather rare ; and Pteraspis (Cymripteraspis) leachi provides a useful means of correlation with continental strata of Lower Siegenian age (Schmidt, 1960 : 140). Class AGNATHA Order HETEROSTRACI Genus PROTASPIS Bryant, 1933 Subgenus EUROPROTASPIS nov. Denison (1953 : 318) has re-defined the genus Prvotaspis in the light of new in- formation, but his diagnosis is based largely on the American species. He notes, however, that in P. arnelli, the only well-known form from Europe, that the position of the branchial opening and the form of the cornual plate are different, but never- theless states that these differences ‘“‘ are hardly adequate to characterize a new genus ”’ in spite of the possibility that the Podolian species ‘‘ may have acquired its Protaspis-like characteristics independently ”’ (1953 : 326). These features seem to be worth at least subgeneric distinction, especially as there are other differences in P. crenulata, particularly in the pre-oral field and in the squamation. I therefore propose a new subgenus, Europrotaspis, to include the new species from Shropshire and P. arnelli from Podolia. The difference between the pre-oral fields is indeed most marked, for in the English species it is large and flat (Text-fig. 17), whereas in the specimen referred to P. bucheri (Denison, 1953, text-fig. 78) it is virtually non-existent. THE OLD RED SANDSTONE OF BROWN CLEE HILL 269 There are two most puzzling features about the scales of P. crenulata (all isolated ; there are no scales recorded of P. arnelli) which seem to separate it widely from the American species. The more marked is the very large size of the associated ridge- Fic. 16. Pvrotaspis (Europrotaspis) crenulata sp. nov. Imperfect dorsal shield. The holotype. Besom Farm Quarry. P.28801. Slightly enlarged. By., branchial plate ; By.D, branchial duct ; By.O, branchial opening ; Co.C, cornual contact surface on branchial plate; Co.P, cornual plate; Do, dorsal disk; Or, orbital plate; Ov.A, orbital aperture; Ro, rostrum; Sp, shells of Spivorbis attached to undersurface of dorsal disk. scales, which, relative to the largest plates known, are much greater than those figured in the tailed specimens of the American species, although these are large compared with those recorded in other pteraspids (Bryant, 1934, pls. 22, 23, 24, fig. 1; Denison, 1953, text-fig. 74). Even more surprising is the fact that a number 270 THE OLD RED SANDSTONE-OF BROWN CLEE HILL of them are strongly asymmetrical, with a marked keel and almost flat underside, which clearly suggest a lateral series—an entirely novel feature in pteraspids unless Zych’s (1931, figs. 41, 49) restorations of Podolian species of Pteraspis are to be taken seriously. The existence of a lateral series of ridge-scales is supported by the shape of the branchial plates which corresponds to that of the scales, for there is an acute longitudinal keel separating the flat undersurface from the convex upper surface. Whether these characters occur in all the European forms is not yet known, but the shape of the branchial plates in the Podolian P. arnelli as restored by Brotzen (1936, text-fig. 3) suggests that they do occur in that species at least. There are two other points in Denison’s (1953) diagnosis that may be called into question : the first is the length of the rostrum, which is stated to be “‘ very short to moderately long’”’. Presumably the “‘ moderately long’ refers to P. priscillae, but the reference of this species to Protaspis seems very doubtful, for none of the critical hinder part of the shield is known. More important is the supposed shape of the tail, which Denison describes as being short and “nearly symmetrical ”’. The irregularity of the tail that he figures (Denison, 1953, text-fig. 74), not only in shape but also in the arrangement of the marginal scales, proclaims it to have been malformed (compare that of ‘‘ Malamia’’, the second specimen of the living coela- canth—see White, 1953 : 114, text-fig. 3). The diminution in size of the scales on the tail of Pteraspis is perfectly regular, like the shape itself, and there are small fulcral scales on both margins of the lobe. Either the tail of this specimen was injured when young and has grown irregularly, or it has been bitten off and partly regenerated. Moreover, there is reason to believe that the specimen is figured upside down, for the scaling at the end of the rows, as illustrated, runs upwards and backwards instead of the normal downwards and backwards, and in Pteraspis it is the dorsal series of ridge-scales that tend to be long and narrow and the ventral series that are short and deep (the depth of the “‘ upper ”’ series as figured by Denison is clearly seen in his figure 744, although not in his restoration). When a laterally compressed tail is twisted at right angles to a rigid body, it is not always easy to see which way it has gone. Diacnosis. Europrotaspis is defined as a subgenus of Protaspis in which the corual plate is reduced to a short narrow element lying between the dorsal disk and the end of the elongated branchial plate ; the branchial opening is large, terminal in position and partly directed upwards; the snout short with a well-developed wide pre-oral field having a slight posterior median ridge; ridge-scales probably extremely large relative to the body-size and arranged, in addition to the usual symmetrical dorsal and ventral series, in very asymmetrical lateral series following the branchial plates, which have an acute longitudinal keel and a flat undersurface. SpEcIES. P. crenulata sp. nov., the subgenotype ; P. arnelli Brotzen and possibly P. wiheriensis Brotzen and P. rotunda Gross. Protaspis (Europrotaspis) crenulata sp. nov. (Pls. 43, 44; Text-figs. 16-33) Diacnosis. An Ewuroprotaspis with rostrum rounded and median length of THE OLD RED SANDSTONE OF BROWN CLEE HILL 271 upper surface contained about four and a half times in that of dorsal shield ; dorsal disk with maximum breadth somewhat less than median length and narrowing only moderately towards anterior end. Dentine ridges of superficial ‘‘ ornamenta- tion” coarse, average 5-6 per mm., when unworn A-shaped with relatively coarse lateral denticulations continued on to the ridges, giving a crenulate appearance. Hototyre. An imperfect dorsal shield (P.28801). MaTERIAL. In addition to the holotype, about 70 specimens, including a dorsal disk ; one almost complete and two fragmentary ventral disks ; three incomplete branchial plates ; 34 isolated scales and about 30 other fragments. Fic. 17. Pyotaspis (Europrotaspis) crenulata sp. nov. A, B, C, three fragments of different rostra, ventral view, with median longitudinal section of centre fragment, showing the pre-oral field (PF). Besom Farm Quarry. P.33832, P.28802, P.28825 respectively. X 2:25. LocatitiEs. The most important is the old quarry + mile SE. of Besom Farm, Wheathill, Shropshire ; another is Upper Overton Quarry, nearly 5 miles to the north-east. Exposures at Prescott Reaside and Farlow Brook Bridge have also yielded some good specimens. Scales and fragments have been found at three or four other places in the neighbourhood. Description. The holotype is an eroded and primarily imperfect dorsal shield (Text-fig. 16). The right side of the rostrum and most of the nght orbital plate were lost before fossilization, and subsequently almost the whole of the dorsal disk and the dorsal side of the right branchial have been removed by weathering, the specimen being on a loose surface block. The left branchial plate with the minute cornual is very well preserved, the posterior half being uncrushed (as was presumably the whole specimen) and preserved in the round. An interesting feature of this specimen is the occurrence of a number of small spiral shells attached to the inner surface of the dorsal disk (Pl. 43, fig. 3; Text- fig. 16, Sp). These are apparently the remains of Spirorbis sp. and indicate that GEOL. 5. 7. 24 272 THE OLD RED SANDSTONE OF BROWN CLEE HILL the carapace had lain exposed on the bottom for some considerable time after the decay of the soft parts and before being covered over. This perhaps accounts for the fractured state of the rostrum. The length of the fossil from the broken tip of the rostrum to the end of the left branchial plate is 9-2 cm. and the maximum breadth, at about 3-5 cm. in front of the hinder end of the branchial, is 7 cm. I9 2 { D e >!) ray —. — & » nin) — | ot Ban ae. D .— e = S> 3) ero ry Se Sez = KE S KE N &® \ YX & ¥: Protaspis (Europrotaspis) crenulata sp. nov. Fic. 18. Imperfect, flattened dorsal disk of adult, possibly aged animal, showing base of dorsal spine and small scales (Sc.) attached to left hinder margin. Upper Overton Quarry. PE 20 405 G5: Fic. 19. Imperfect part-grown dorsal disk in impression with indications of sensory canals. Besom Farm Quarry. P.26313. X 1°'5. Probably little has been lost of the rostrum in length, for fragmentary specimens of this plate indicate that it was very much rounded in front (Text-fig. 17). If so, its length was contained about four and a half times in the total length of the shield, omitting the dorsal spine. The orbital plates were small with very small orbits, but the mesial processes were broad and certainly met the pineal plate, which is unknown, completely separat- ing the rostrum from the dorsal disk. THE OLD RED SANDSTONE OF BROWN CLEE HILL 273 The anterior margin of the disk is shallowly but widely indented with rounded corners, and the posterior margins are sigmoidally excavated (Pl. 43, fig. 1; Text- fig. 18). In this fully grown specimen a number of very small scales are fused to the hinder margin, the largest being immediately behind the cornual plate and fitting on to a short emargination of the disk. These small scales are probably no more than a manifestation of old age, which is often shown in compression, fusion Fic. 20. Pyrotaspis (Europrotaspis) crenulata sp. nov. Restoration of dorsal shield in outline, with enlarged cross-sections through branchial plate, etc., at A, B, and C. x 1 approx. By, branchial plate; Co, cornual plate; Do, dorsal disk; Ve, ventral disk. and other irregularities commonly seen along the margins of the disks of pteraspids. They have nothing to do with the definite structures which Denison calls “‘ posterior scale-like process of cornual plate ”’. The sides of the dorsal disk are moderately convex, particularly in the hinder quarter as the cornual plate is approached, for which there is an emargination 274 THE OLD RED SANDSTONE OF BROWN CLEE HILL varying in definition with age, being most marked in the oldest specimens (Text- fig. 16), scarcely perceptible in the half-grown (Text-fig. 19), and altogether absent in the young which have the usual rounded outlines (P.2888r). The socket for the dorsal spine is relatively short in the adult, occupying about one-fifth of the median length of the dorsal disk. Of the spine itself only the basal portion is known (Pl. 43, fig. 1; Text-fig. 18; P.29076), but to judge from these fragments it was clearly stout and low, coming off the disk at about 30°, and probably short. a4 IN \\i\ | Uh L(y TF ts v . 1) t ‘ x - . , ' ” a | 1 ra | oy ' , 1 tar oe ni. - roy Oe lat © ae PLATE 37 Pteraspis (Pteraspis) daivydinglensis sp. nov. Oral plates. Dairy Dingle. x 4. Fic. 1. Median plate. P.43709. Fics. 2, 3. Near-median plates, Fig. 2 slightly imperfect basally. P.43710-11. Fics. 4-7. Plates progressively more lateral in position. P.43712-15. Fics. 8-11. Plates with heads broken away. Figs. 8, 11 are near-laterals with little or no overlapped areas; Fig. 9, a left near-median plate with wide lateral and ventral overlapped areas; Fig. 10 shows overlapped area at base only (O), and ornamented surface rubbed by movement against post-oral cover (W). P.43716—109. a, c, aboral views; 0, d, oral views. ALAMOS Si7/ Bull. B.M. (N.H.) Geol. 5,7 ORAL EEAIES OF BIBRASEIS . yer = : I * ; seg 7 - 7 : oe a | = 7 = 2 es iy : i ; i j é ets as vw ; 1) ri : : ¥ PLATE 38 Ptevaspis (Ptevaspis) daivydinglensis sp. nov. Fics. 1-3. Fixed lateral oral plates. P.43721-23. Fics. 4, 5. Anterior lateral plates. P.43730-31. (a) outer; (6) inner faces. All from Dairy Dingle. PEE so Bull. B.M. (N.H.) Geol. 5,7 PTERASPIS PLATE 39 Pteraspis (Pteraspis) dairydinglensis sp. nov. Fics. 1-4. Posterior lateral plates. P.43700-03. x 6. Fics. 5-8. Post-oral covers. P.43704-07. x 6. Fics. 9-25. Lateral line scales. P.43601-17. x 4, except Fig. 14, x 8. Figs. 11, 13, 15, 19, 20, 24, 25 show inner surfaces. All from Dairy Dingle. LAINE 3h) Bull. B.M. (N.H.) Geol. 5,7 PTERASPIS ie ad L YI I 4 ny - is "7 t Ap z ha ; (Cah Pi } 1 hf — 4 u iN 5s i y - i: - 2 ‘ i= 40 hs 4 Vr | F re ‘ 7 ine — 1 i ¢ ‘ | ‘ -) ; o ‘ =. ; ‘ : ; 1 ea ‘ n i i Pt ch oe pty ts ” ey ) eM eee ee ee 3 To " ; ; i. “1 be, ; _ _ we wt =, i oi i 7 : ; Tee ces | Ot) aaa cb omg) LAT ve ni Spd us Peak el) wi: iinet co umes i vii y 7 i ar oe P a - Mas) Vf : Na) UM go ar |), nee Mesh Te ee | a - I PLATE 40 Pteraspis (Pteraspis) dairydinglensis sp. nov. Fics. 1-3. Dorsal ridge-scales. (Fig. 1, inner surface.) P.43618-20. Fic. 4. Imperfect asymmetrical ridge-scale. P.43621. Fics. 5, 6. Imperfect ventral ridge-scales showing signs of wear. P.43622, P.43624. Fics. 7-9. Posterior ridge-scales. P.43625-27. Fics. 10, 11. Anterior ventral ridge-scales. P.43629-30. Fic. 12. Anterior ventral ridge-scale showing ornament on undersurface. P.43631. Fic. 13. Anterior ridge-scale of tail. P.43632. Fics. 14-27. Flank-scales. Figs. 14, 20, 22, 26 show undersurfaces, the last two with ornamentation. P.43633-37, P.43668-69, P.43639-40, P. 43671, P.43643-46. All from Dairy Dingle. x 4. PLATE 40 Bull. B.M. (N.H.) Geol. 5, 7 PTERASPIS PLATE 41 Ptevaspis (Pterapis) dairydinglensis sp. nov. Fics. 1-22, 24. Flank-scales. Figs. 2, 17show undersurfaces. Dairy Dingle. P.43641-42, P.43647-56, P.43658, P.43660, P.43667, P.43676, P. 43678, P.43682-83, P.43686, P. 43689, P. 43693, P. 43692. Fics. 23, 25-27. Caudalscales. Fig. 27 shows the undersurface. Dairy Dingle. P.43690, P. 43695, P. 43604, P.43696. Fic. 28. Irregular scale of unknown position. Dairy Dingle. P.43698. Fics. 29-31. Flank-scales. Derrington Rea Bridge. P.44395, P.44398, P.44396. Fics. 32-37. Various ridge-scales. Upton Cresset. P.42873, P.42872, P.42874, P.42877, P. 42876, P. 42881. Fics. 38-44. Various flank-scales. Upton Cresset. P.42897, P.42893, P.42894, P.42901, P.42900, P.42904, P.42884. All xX 4. PLATE 41 Bull. B.M. (N.H.) Geol. 5,7 PTERASPIS ~~ We aan 9 ih 8 = = na vd / ig thes ‘ oi < . ae. _ - 2 ov ae 2, 7 ; a p ‘ ‘ Mu srl | i] J 7 ae <_< i < _ ae : 2 - ee oP ta F - a 7 : : a : ° 7 | ; 7 | a ate 2% . E er. * a = | é a se ee = - ; .. << ig? je ; . = tf j , 7 Dre Ny { | il Y / . = 5 f i j : ; i i . a . ¢ ae =| as 7 => ; f : : Je. Ti 2 ke ; } ; : Zs t = = na 4 i) 7 4 pp: eae oghin@eic ‘ Ss eed j ‘ xe 2 ae on : 7 he = ~, if ’ md - i ar = a i | Wh 2 - ib : Md ee ite i TOD ee i ih Pee ees ¢ : q Said te SPs t ( Ye a tbe zy yerebi Sar, 2 et Las i oS i REE GRE Cte SM Stal te = fel ; , ; oe hed is 4 : a ne = s hae i on + Wes iry tne arent rs ah gt en # ny ees, irene an | 7 dae or * yi be , Loe ware ne ”s ay sas Lt eee : i ‘ 1 , , Ta gph 0 i 4 : tees a eee a ean) Ln lek ia ai, Ps a my +9 ; . Bo 1.5 . aa a rabid la? a oe ef ss a. i Aran eis ia Ra: % a: Sed le cre 28 Fia. 1. P.17451. Fic. 2. holotype. FIG. 3. xX 2°5. FIG. 4. PLATE 42 Ischnacanthus wickhami sp. nov. Part of upper jaw with worn dentition. Man Brook 4, Trimpley, Worcestershire. x 15. Part of upper jaw. a, Oral view. Gardener’s Bank, Cleobury Mortimer. The P.24625. X I°5. Part of upper margin of lower jaw. a, Oral view. Hudwick Dingle 1. P.29725. Ischnacanthus kingi sp. nov. Right lower jaw. Baggeridge Colliery, South Staffordshire. P.15362. X 1°5. Bull. B.M. (N.H.) Geol. 5, 7 PLATE 42 ISCHNACANTHUS PLATE 43 Protaspis (Europrotaspis) crenulata sp. nov. Fic. 1. Imperfect dorsal disk. Upper Overton Quarry. P.29415. x 2. Fic. 2. Posterior end of left branchial plate with part of dorsal disk attached. Part of holotype. Besom Farm Quarry. P.28801. x 2. Fic. 3. Part of holotype showing specimens of Spivobis sp. which were attached to inner surface of dorsal disk. Besom Farm Quarry. P.28801. x 7. Fic. 4. Impression of part of dorsal disk showing discontinuous ornamentation. Besom Farm Quarry. P.28879. X 3. Fic. 5. Flank-scale. Besom Farm Quarry. P.29068. x 8. Bull. B.M. (N.H.) Geol. 5, 7 ILMIVE, 413) PROTASPIS = “i 1 i tL i h. i Nii % ‘ 4 nb ‘ 1” : i 7 My mi a | hy i ort 0 ¥ Ni i .. Air hi ifs 4) 4 pi iy ae, ne a PLATE 44 Protaspis (Europrotaspis) crenulata sp. nov. Fic. 1. Imperfect ventral disk, largely in impression. Besom Farm Quarry. P.26311. Fic. 2. Dorsal (?) ridge-scale. Newton Dingle 2, Loughton. P.29386. x 7. (?) Ventro-lateral ridge-scale. Besom Farm Quarry. P.29077. x 8. FIG. 3. Fic. 4. Part of Fig. 3 further enlarged to show details of ornamentation. x 20. Fic. 5. Slightly worn ornamentation on dorsal disk. Upper Overton Quarry. P.31642. x 20. Fic. 7. Part of left anterior border of ventral disk showing irregularity of ornamentation. Besom Farm Quarry. P.29479. *X 5. Protaspis (Europrotaspis) aynelli Brotzen Fic. 6. Ornament of dorsal disk. Stage IIIb. Kujdanow, Podolia. (N.H.M. Stockholm PE LO6>)i< ZO: Bull. B.M. (N.H.) Geol. 5, 7 PLATE 44 ‘ 5B We? 7 Res z ae ae ~s ES page SE PROTASPIS 1 1.4 ib ee ste oi tals i tf sori s ae 7 is ome * et) a) ae mw ug or A ’ I i / ay i = h wr 4 \ ‘ Me i } i. A bf a ih iat - ae | a ca es ae Soe hin ites eal a Wiad , Vy o “i ee 4 on! : , i t ( j t T ni t oD ‘ ¥ n rt f j a ' wae Peet eee PLATE 45 Fic. 1. Ptervaspis (Cymripteraspis) leachti White. Flank-scale. Swanlake Bay, Pembroke- shire. P.29329. xX 8. Fic. 2. Pteraspis (Cymripteraspis) leacht White. Flank-scale. Prescott Reaside. P.29261. x 8. Fic. 3. Tooth of Acanthodian. Besom Farm Quarry. P.28903. x 5. Fic. 4. Ischnacanthus (?) anglicus sp. nov. Median fin-spine with restored dorsal view. (a) Right side view. The holotype. Besom Farm Quarry. P.29082. x 3. Fic. 5. Onchus wheathillensis sp. nov. Median fin-spine. The holotype. Besom Farm Quarry. P.29083. x 3. Fic. 6. Overtonaspis billballi gen. et sp. nov. Right anterior lateral plate. The holotype. Upper Overton Quarry. P.29272. xX I°5. Fic. 7. Kujdanowiaspis anglica (Traquair). Imperfect right anterior ventro-lateral plate. New House Farm, Neenton. P.29821. x 1°5. PEE] 45 Bull. B.M. (N.H.) Geol. 5, 7 PTERASPIS, ACANTHODIANS and ARTHRODIRES ef : PLATE 46 Fic. 1. Kujdanowtaspis willsi sp. nov. Left spinal plate, dorsal view. The holotype. Besom Farm Quarry. P.28913. xX 5. Fic. 2. Kujdanowtaspis willsi sp. nov. Left central plate, referred to this species. Besom Farm Quarry. P.28889. xX 3. Fic. 3. Prescottaspis dineleyi gen. et sp. nov. Median scute. Prescott Reaside. P.29267. S< Go Fic. 4. ? Prescottaspis dineleyi gen. et sp. nov. Left posterior ventro-lateral plate without surface ornamentation, possibly of this species. Prescott Reaside. P.29259. xX 3. Fic. 5. Wheathillaspis wickhamkingi gen. et sp. nov. Imperfect anterior ventro-lateral and spinal plates. The holotype. Besom Farm Quarry. P.28908. xX 4. Fic. 6. Lateral scale of an arctolepid. Upper Overton Quarry. P.29756. x 3. PLATE 46 Bull. B.M. (N.H.) Geol. 5, 7 ARTHRODIRES ath I ‘il | Ty 2 =.) = ie} ; wir aay Ae ra g Recs f a eee Pd ia ty ae ” é , ar : os ea : i . net ’ A ith ; Fi 4 f i * te, : { Sar | ow @ i if 7 * f : oe ‘Va ni ie a. i uu ae Bay he ni he r y we ry) | im a 4 ee nn a i HA sath ane) Pe mae me ui 1 i vy ‘ i , b , / i. eh Cx: . TVR : Pon on 2 eA er gh. del Lice if ara nu ibd A od 9 Ze oe san) plete Lo ened Dd * Wee i ic By * pia MG? aren gh Es he. : PLATE 47 Pseudosauripterus anglicus (A. S. Woodward) Fic. 1. Cast of right entopterygoid, oral view. Church Quarry, Farlow. P.27176. x 1°5. Fics. 2-4. External impressions of scales of paratype block, with light reversed to give positive effect. Church Quarry, Farlow. P.200a. x 2. Fic. 5. External impressions of scales on counterpart of type-block, with lighting reversed. Scale on right is the underside of that figured by A. S. Woodward, 1891, pl. 16, fig. 6. Church Quarry, Farlow. P.7601. x 2. Bull. B.M. (N.H.) Geol. 5,7 PICATE 4 PSEUDOSAURIPTERUS PLATE 48 Sauripteris taylori Hall Fic. 1. External impression of scale with lighting reversed to give positive effect. The cracks are of post-mortem origin. Blossburg Red Beds, Cassadaga Stage, lower Upper Devonian, near Blossburg, Tioga County, Pennsylvania. P.43518. xX 2. Fics. 2, 3. Scales from same slab as Fig. 1. Nat. size. Bull. B.M. (N.H.) Geol. 5, 7 PLATE 48 SAURIPTERIS THE ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, Il W. T. DEAN BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 5 No. 8 LONDON: 1961 THE ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE? jl BY WILLIAM THORNTON DEAN Pp. 311-358 ; Pls. 49-55 BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 5 No. 8 LONDON: 1961 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), tnstituted in 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calender ‘year. This paper 1s Vol. 5, No. 8 of the Geological (Palaeontological) series. © Trustees of the British Museum, 1961 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued November 1961 Price Twenty five shillings THE ORDOVICIAN TRILOBITE FAUNAS OF SOUTH -SHRORSHIRE: Il By WILLIAM THORNTON DEAN Page SYSTEMATIC DESCRIPTIONS . : : : : , : : es k4! Family Cheiruridae Salter é : é 5 BA Subfamily Acanthoparyphinae Whittington & Evitt : : 6 Buz Acanthoparypha stubblefieldi (Bancroft) : : : Se 3r4 Subfamily Cyrtometopinae Opik : ; : j : 4 BG Pseudosphaerexochus sp. . ; : . ¢ : hs SHG) Family Encrinuridae Angelin . ; ; : ; 5 : S77 Subfamily Encrinurinae Angelin : ; . ; : 6 Si Encyinurus sp. (? nov.) c : : : : ‘ ESET. Subfamily Cybelinae Holliday . ; : : : : BES Atractopyge sp. : : : : 0 : : 5) 318 Family Dalmanitidae Reed . 0 : 6 : é : 5 B20) Subfamily Acastinae Delo : ; : 5 BA) Kloucekia (Phacopidina) harnagensis (Bancroft) ; : > 32 Kloucekia (Phacopidina) Tinie (M‘ mo ; : : 2324 Family Pterygometopidae Reed d : : ‘ 0 328) Subfamily Pterygometopinae Reed . : : : : . 328 Calyptaulax actonensis sp. nov. . : : ‘ : 5 328) Subfamily Chasmopsinae Pillet : ; ¢ . ; > 335 Chasmops extensa (Boeck) . : : : : ‘ = 331 Chasmops salopiensis sp.nov. . : c : 2 a 337 Family Homalonotidae Chapman . : : i ‘ ‘ 0 SiSK) Platycoryphe dentata sp. nov. : : : : 2 . 340 Platycoryphe ? sp : 0 : : : - 344 Brongniartella ether (M‘ Can) , : F : 40 Brongmartella cavadociana sp. nov. ‘ ; : F - 349 Brongniartella minor (Salter) 3 : : o | Sn Brongniartella minor subcarinata subsp. nov. : c - 352 Brongmiartella edgelli (Salter) 4 : j 5 : SoS Brongniartella sp. ‘ ; ; i ; : : - 354 REFERENCES . : . . ¢ : : : ; . = 355 SYNOPSIS The present paper is the second of four describing and figuring the trilobites of the Caradoc Series in south Shropshire. The families dealt with include the proparian Cheiruridae, Encrin- uridae, Dalmanitidae and Pterygometopidae, and the gonatoparian Homalonotidae. These comprise nine genera and subgenera, containing seventeen species and subspecies, five of them new. The type-species of Phacopidina Bancroft is redescribed, the genus being removed from the subfamily Calmoniinae and assigned to the subfamily Acastinae as a subgenus of Kloucekia Delo. GEOL. 5. 8. 29 314 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II SYSTEMATIC DESCRIPTIONS Family CHEIRURIDAE Salter, 1864 Subfamily ACANTHOPARYPHINAE Whittington & Evitt, 1953 Genus ACANTHOPARYPHA Whittington & Evitt, 1953 TYPE SPECIES. Acanthoparypha perforata by original designation of Whittington & Evitt (1953 : 72). Acanthoparypha stubblefieldi (Bancroft) (Pl. 49, figs. 1, 3-6, 11) 1949. Nieszkowskia stubblefieldi Bancroft, p. 311, pl. 10, figs. 26, 27. 1958. Nieszhowskia stubblefieldi Bancroft : Dean, pp. 201, 219. The only figure of the cranidium of the species published by Bancroft (1949, pl. 10, fig. 26) shows an incomplete and somewhat distorted internal mould which has apparently been slightly compressed laterally and sheared dextrally. It is necessary, therefore, to redescribe the species and to figure topotype material showing the true glabellar form. Bancroft’s photograph does not give a true impression of the pro- portions of the glabella, and also exaggerates the angle at which the glabellar furrows are directed backwards. DESCRIPTION. In plan the glabellar outline is subrectangular, with the frontal lobe only slightly convex forwards and the anterolateral margins and angles be- coming respectively slightly convergent and rounded. The glabella is moderately convex both longitudinally and transversely, its length being estimated as slightly greater than the maximum breadth, measured across the second and third glabellar lobes. The anterior border is imperfectly preserved, but appears to be small and narrow (sag.). There are three pairs of glabellar lobes, those of the third pair being slightly the largest. The frontal glabellar lobe is short (sag.), less than one-fifth of the length of the glabella, and broad ; frontally it is almost transversely straight, but the anterolateral margins are broadly rounded. Thefirst glabellar furrowsare deep, shallowing adaxially and curving inwards and backwards from the axial furrows for almost one-third the breadth of the glabella. The first glabellar lobes are subrectangular in plan and of uniform length (exsag.). The second glabellar furrows are deep, parallel to the first pair of furrows but turning slightly backwards at their proximal ends ; in one specimen (PI. 49, figs. I, 4) the right-hand furrow of the second pair does not quite attain the axial furrow, but there is insufficient material to ascertain whether this is an exceptionalcase. The second glabellar lobes are of similar shape to those of the first pair, and of only slightly smaller size. The third glabellar furrows are deep, more or less parallel to the other pairs of furrows. They extend inwards from the axial furrows at an angle of about 65~70 degrees, but towards their proximal ends they turn back sharply through about 30 degrees and extend towards, though they do not reach, the occipital furrow. The proximal ends of the three pairs of glabellar furrows are almost longitudinally in line, forming a median body of uniform breadth. ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 315 The third glabellar lobes are larger than the other pairs and tend towards the so-called “cat’s ear’ outline (see Pl. 49, fig. 6), whilst the gently convex posterior margins give a sinuous appearance to the moderately deep occipital furrow. The occipital ring curves forwards slightly towards the axial furrows and is broader (sag.) than the posterior border ; only the proximal parts of the latter are known and they are transversely straight. The fixigenae, librigenae and facial sutures have not been found preserved. The hypostoma and thorax are unknown. Only one specimen of the pygidium is available and, although incomplete, is estimated to have been slightly wider than long. The axis is roughly triangular with three axial rings and occupies one-third of the frontal breadth of the pygidium. First axial ring is narrow (sag.), curved sharply backwards distally where there are poorly- defined axial furrows. Beyond the axial furrows the pleurae of the first pair become suddenly and markedly wider (exsag.) and are produced posteriorly as a pair of large spines. The anterior edge of each pleura runs transversely straight from the axial furrow and then turns sharply backwards through a right angle to form the lateral margin of the pygidium. The posterior edges of the same pleurae almost coalesce with the second pleurae, being separated from them by only shallow interpleural furrows which extend posteriorly, parallel to the lateral margin until just beyond the tip of the axis ; they then diverge backwards markedly and finally intersect the lateral margins so as to form long, stout pleural spines. The second pleurae are nearly continuous with the second axial ring, beyond the distal ends of which they are directed almost straight backwards as a pair of large, parallel spines, the bases of which meet proximally. Between their point of contact and the second axial ring is situated the third axial ring. This is not differentiated from its corresponding pleurae which converge sharply backwards and inwards, and meet to form a sub- triangular area with a central depression. The surface of the glabella is covered with large tubercles, interspersed with smaller granules. Similar, though more closely crowded tubercles ornament the axial rings of the pygidium, whilst on the side-lobes they are much more dispersed. All the furrows on both cranidium and pygidium are smooth. HorRIZON AND Locatity. Acanthoparypha stubblefieldi is known as yet from only one locality in south Shropshire, the old cartway section near the south-eastern corner of Smeathen Wood, Horderley, 250 yards west-south-west of Wood House. The horizon is in the Smeathen Wood Beds, Harnagian Stage, zone of Reuscholithus veuschi. LECTOTYPE, here selected. BM. In. 42084a, 6 (PI. 40, figs. 5, 11). PARATYPE. BM. In. 52085 (PI. 49, fig. 3). FIGURED SPECIMENS. BM. In. 50498 (Pl. 49, fig. 1); BM. In. 50499 (PI. 49, fig. 6). Discussion. The genus Acanthoparypha, although close to Nieszkowskia Schmidt, was separated by Whittington & Evitt so as to include only those Nieszkowskia-like forms which lack the glabellar spine such as is found in NV. cephaloceros (Nieszkowski) the type species. They were unable to differentiate between the two genera on the basis of the pygidia. Acanthoparypha perforata (Whittington & Evitt, 1953 : 73, pl. 13, figs. 2, 4-6 ; pls. 316 ORDOVICIAN TRILOBITE FAUNAS OF SOULH SHROPSHIRE, Li 14-16; pl. 17, figs. 1-6), from the Edinburg Limestone of Virginia, is easily dis- tinguished from A. stubblefieldi by its strongly convex glabella, the sides of which are more convergent anteriorly, and by the proportionately longer pygidium, with four long, subequal spines and only one fully-developed ring furrow, behind which is situated a deep, mesial pit. The cephalon of Acanthoparypha chiropyga (Whittington & Evitt, 1953 : 77, pl. 28, figs. 31-44; pl. 29; pl. 30, figs. 1-28, from the Lincolnshire Limestone of Virginia, possesses less coarsely-developed tuberculation than that of A. perforata. In this respect, and also in having better-defined glabellar furrows, it approaches A. stubblefieldi but can be separated from the latter by means of its more tapered glabellar outline and by its pygidium, only the first axial ring of which is known, but which has four pleural projections of almost equal size, and carries several large tubercles on its upper surface. Both these North American species are lower Middle Ordovician in age, and occur respectively at horizons which are generally contem- poraneous with, and slightly earlier than, the Shropshire specimens. Cheirurus variolaris was founded by Linnarsson (1869 : 60, pl. 1, fig. 6) on an in- complete pygidium but later Schmidt (1881 : 183, pl. 9, figs. 1-8) illustrated the species more fully and assigned it to Nieszkowskia, apparently correctly, as it has a glabellar spine. The trilobite which Schmidt (1881, pl. 11, figs. 25, 26) described as Cheirurus (Nieszkowskia) variolaris var. mutica lacks a glabellar spine and is appar- ently a true Acanthoparypha. The glabellar furrows and surface tuberculation re- semble those of A. stubblefieldi but the glabella is more transversely convex and tapered frontally than the Shropshire species. Acanthoparypha mutica is from Stage C i of Gostilizy, and is therefore of generally similar age to A. stubblefieldt. Subfamily CyRTOMETOPINAE Opik, 1937 Genus PSEUDOSPHAEREXOCHUS Schmidt, 1881 TYPE SPECIES. Sphaerexochus hemicranium Kutorga, 1854 by subsequent desig- nation of Reed (1896 : 119). Pseudosphaerexochus sp. (Pl. 49, figs. 2, 7, 8) 1958. Pseudosphaerexochus sp., Dean, pp. 213, 225. Only one specimen is available, a large but somewhat distorted cephalon approxi- mately 60 mm. wide. The glabella is large and inflated, with a basal breadth about one-third that of the cephalon ; anteriorly it expands until, at about its mid-point, it attains a maximum breadth equal to half that of the cephalon. In front of the second glabellar lobes the glabella narrows, but details are obscured by crushing. There are three pairs of glabellar furrows. The first two pairs are moderately deep at their intersection with the axial furrows, but become shallower as they curve inwards and backwards, each furrow extending adaxially for about one-quarter of the breadth of the glabella. The first and second glabellar lobes are of similar form, ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 317 uniformly long (exsag.), and apparently parallel to each other. The third glabellar furrows are deep, directed only slightly backwards near the axial furrows; they then turn fairly sharply backwards and run towards, though they do not reach, the occipital furrow, at the same time becoming noticeably shallower. They delimit large, subangular basal glabellar lobes, each of which occupies almost one-third of the basal breadth of the glabella. The axial furrows are deep and narrow, though their narrowness may be exaggerated by crushing. The occipital furrow is narrow (sag.) and transversely straight, becoming deeper towards the axial furrows. The pleuroccipital furrow is deep near the axial furrows but shallows distally ; towards the lateral margins it meets the lateral border furrows, demarcating a broad (é7.) lateral border which is continuous with the posterior border. The latter is narrow (exsag.) at the axial furrows but widens towards the genal angles which form broad, bluntly-pointed fixigenal spines. Only the left palpebral lobe is preserved intact, situated opposite the second glabellar lobes and about mid-way between the axial furrow and the lateral margin. It is small and short (exsag.), less than the length of the second glabellar lobes, with a poorly-defined palpebral furrow. The fixigenae are large, tumid and their surface is continuous with that of the librigenae. The two together are covered with conspicuous shallow pits extending as far as the lateral border furrow which separates them from the smooth lateral border, itself continuous with the distal portions of the smooth posterior border. As far as can be ascertained the surface of the glabella is smooth. HorIZON AND Locality. Onmia Beds, Onnian Stage, zone of Onnia superba, at the so-called ‘‘ Cliff Section ”’ in the north bank of the River Onny, 720 yards west- south-west of Wistanstow Church. FIGURED SPECIMEN. BM. In. 50660. Discussion. Pseudosphaerexochus is figured here for the first time from Shropshire. In the Ordovician of England and Wales the genus is generally more characteristic of the Ashgill Series, though it is recorded from the Onnian and Pusgillian Stages of the Cross Fell Inlier (Dean, 1959a). Insufficient material is available for a detailed comparison, but the Onny specimen is larger than any of those from Cross Fell. Family ENCRINURIDAE Angelin, 1854 Subfamily ENCRINURINAE Angelin, 1854 Genus ENCRINURUS Emmrich, 1844 TYPE SPECIES. Entomostracites punctatus Wahlenberg, 1821 by original designa- tion of Emmrich (1844 : 16). Encrinurus sp. (? nov.) (Pl. 49, figs. 10, 12) 1958. Encrinurus sp., Dean, pp. 209, 223. The single specimen available is a pygidium, strongly convex transversely and preserved as an internal mould, In outline it is almost triangular, the maximum 318 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II breadth, which is attained frontally, being approximately equal to the length. The axis occupies one-third of the frontal breadth and is delimited laterally by deep axial furrows which converge gently rearwards ; it does not extend to the tip of the pygi- dium. There are at least sixteen or seventeen axial rings preserved, with room for a few additional rings as the axis at and near the terminal piece is slightly abraded. At least the first nine axial rings are entire, but owing to the state of preservation it is difficult to say whether the subsequent rings are obsolete mesially or whether this feature is due to abrasion. Traces of small, median tubercles are visible on the second, third, fifth, ninth, twelfth and thirteenth axial rings. Each side-lobe possesses seven well-defined ribs, with an eighth less well defined, separated from each other by conspicuous interpleural furrows which are only slightly narrower (exsag.) than the ribs themselves. The ribs of the first pair are transversely straight and horizontal as far as the fulcrum, beyond which they turn down and sharply back. Subsequent ribs turn backwards even more markedly, and the seventh and eighth pairs run almost straight back parallel to each other. As far as can be seen, all the ribs end in small, free points. Beyond the tip of the axis is a narrow, postaxial ridge which ex- tends to the tip of the pygidium and is separated from the eighth pair of ribs by weak, shallow furrows. HORIZON AND LOCALITY. The middle subdivision, the Dalmanella unguis Beds or Zone, of the Marshbrookian Stage, forming the upper half of the disused Marshwood Quarry, about half a mile south of Marshbrook Station. FIGURED SPECIMEN. BM. In. 49312. Discussion. No other Encrinurus of comparable age or form has been described or figured from the Anglo-Welsh Caradoc Series, and the Shropshire species may be new. Encrinurus trentonensis Walcott (1879 : 68; Wilson, 1947: 45, pl. 8, fig. 5) has a pygidium of generally similar type but carrying at least twenty-three axial rings and nine or ten pleural ribs. This species is from the Lower Trenton of Ontario and therefore probably rather earlier than the Shropshire specimen. Subfamily CyBELINAE Holliday, 1942 Genus ATRACTOPYGE Hawle & Corda, 1847 TYPE SPECIES. Calymene verrucosa Dalman, 1827 by original designation of Hawle & Corda (1847 : 90). Atractopyge sp. (Pl. 49, fig. 13) 1854. Cybele verrucosa Dalman: Salter & Aveline, p. 67. 1958. Altvactopyge sp., Dean, p. 211. Only one specimen has so far been collected from south Shropshire, a damaged pygidium with axis almost complete, and retaining part of the side-lobes. The maximum length, excluding the articulating half-ring, is 15 mm. The length of the axis is three times the breadth, and its sides are parallel for about two-thirds of the length but then converge to meet at the pointed axial tip. There are at least twelve ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 319 axial rings, though the posterior portion of the axis is too badly damaged for more to be distinguished. The first four axial rings pass laterally without interruption into the ribs of the side-lobes, axial furrows being present merely as ill-defined, shallow, longitudinal depressions ; only to the rear of the fourth axial ring is the axis defined laterally by deep axial furrows which curve gently back to the pointed tip. The first two, and perhaps three, axial rings are entire across the median line, but all subse- quent rings are obsolete medially for about one-third of their breadth(¢r.), so that they appear to be defined by two series of transverse notches which run adaxially from just inside the axial furrows and are separated from each other by a median band which extends, apparently, to the tip of the axis. This band is mostly smooth, but the fifth to seventh, and perhaps also the fourth, axial rings bear traces of paired median tubercles. The ring furrows beyond the sixth become gradually more reduced. The first four axial rings are produced laterally to form the ribs of the side-lobes and extend transversely for a short distance from the axis; they then curve sharply backwards and at the same time converge slightly so that the fourth pair of ribs embraces the sides of the axis and terminates in a pair of free points which almost meet just beyond its tip. The remaining ribs, from first to third, run parallel to the fourth pair and terminate in a series of free points which are not in line but are stepped adaxially rearwards from first to fourth. Each rib of the fourth pair carries a poorly-defined tubercle opposite the seventh, tenth and twelfth ring furrows, and similar tubercles are visible on the other ribs, though the preservation is too poor to ascertain their exact distribution. HoRIZON AND LOCALITY. The disused quarry by the east side of the road, almost opposite Church Farm, Acton Scott, in the Acton Scott Beds (s.s.), representing part of the middle third of the Actonian Stage. FIGURED SPECIMEN. BM. In. 49319. Discussion. The name “ Cybele verrucosa’’ has been applied indiscriminately for many years to specimens of Atractopyge from a large number of horizons and localities in Britain, and it seems probable that several different species will prove to be involved when they are better known. The true Atvactopyge verrucosa was first described by Dalman (1827) from the Red Tvetaspis Shales of Billingen, Vastergét- land, but the original illustrations were too poor for detailed comparison. More recently Henningsmoen (in Moore, 1959, fig. 349, Ia, 6) has refigured the species from type material ; his illustration of the pygidium shows there are about eleven axial rings behind the first four entire axial ring furrows, and that the tips of the pleural ribs end almost in line with each other. This last feature marks a conspicuous difference from the Shropshire specimen, which probably also has fewer axial rings. The species described by M‘Coy (2m Sedgwick & M‘Coy, 1851 : 156, pl. 16, figs. 1-5) as Zethus atractopyge has often been regarded as a synonym of Atractopyge verrucosa, and is certainly congeneric with the latter. Atvactopyge atractopyge was founded on a number of syntypes in the Sedgwick Museum numbered A. 41875, A. 41903 and A. 41904a, 6, deriving respectively from “ Alt yr Anker, Meifod ’’ (now Gallt-yr-Ancr), “ the schists of Ravenstone Dale’ (presumably the Cautley district), and “the slate of Coniston’’. Of these the pygidium from near Meifod (M‘Coy, fig. 4) is here chosen as lectotype, the other two being either incom- 320 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II plete or distorted. The hill of Gallt-yr-Ancr, near Meifod, Montgomeryshire, is made up largely of Caradoc strata belonging to the Soudleyan and Longvillian Stages. M‘Coy’s species is in need of redescription from topotype material and its exact type locality is, unfortunately, not known, but King (1928 : 698) has recorded “‘ Cybele verrucosa ’’ as occurring commonly at one locality on Gallt-yr-Ancr, as well as several others in the district, in company with Pterygometopus (now Estoniops) jukest (Salter), a species which, so far as is known, is restricted to the Upper Longvillian Substage and is regarded here as a subjective synonym of Estoniops alifrons (M‘Coy, 1851). Two pygidia assigned to Atvactopyge have been figured from the Derfel Limestone of North Wales by Whittington (7m Whittington & Williams, 1955, pl. 40, figs. 109, 110). One of these (fig. 109) is broadly comparable with the Shropshire specimen, though the preservation is too poor for detailed comparison, but the other (fig. 110) appears to be a typical Pavacybeloides of the givvanensis (Reed) species-group. As has been noted elsewhere (Dean, 1958 : 212), the association of Atractopyge, Illaenus, Platylichas and Nicolella in the Derfel Limestone is matched at a much later date in south Shropshire by part of the fauna of the Actonian Stage. In west Shropshire Whittard (1960 : 126, pl. 17, figs. 12-16) has recorded the Scot- tish species Atractopyge michelli (Reed) from the Llandeilo Series ; the Acton Scott form can readily be separated from this species by means of its narrower, proportion- ately longer pygidial axis. Atractopyge killochanensis (Tripp, 1954 : 678, pl. 4, figs. 1-5), though broadly similar to the east Shropshire species may be distinguished by the proportionately longer axis behind the first four axial rings, and by the free points of the pleural ribs which are markedly convergent backwards. Family DALMANITIDAE Reed, 1905 Subfamily ACASTINAE Delo, 1935 Genus KLOUCEKIA Delo, 1935 ? 1956. Dreyfussina Choubert e al., p. 394. Type SPECIES. Phacops phillips Barrande, 1846 by original designation of Delo (1935 : 408). Subgenus Phacopidina Bancroft, 1949 Phacopidina was first proposed by Bancroft (1949 : 310) as a genus of the family Phacopidae. Both Pillet (1954) and Hupé (1955 : 261) placed it in the subfamily Acastinae, using the latter as a subdivision of the Phacopidae, but since then Struve («2 Moore, 1959 : 0.486) has preferred to assign the genus to the subfamily Calmoni- inae, within the family Calmoniidae. Struve qualified this transfer by claiming that Phacopidina is closer to what he called “‘ the Calmoniinae trend ”’ on account of the direction of the second glabellar furrows, the pairs of glabellar lobes of equal size, and the shape of the genal spines. He pointed out that Bancroft’s illustration of the type-species was indifferent, but his own diagram of a, presumably, mature indi- vidual of Phacopidina harnagensis Bancroft (Struve im Moore, 1959, fig. 385, 4a, 0) ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 321 is in turn somewhat misleading. The first glabellar furrows are too deeply impressed, the second furrows diverge forwards too strongly and are shown to reach the axial furrows, the genal spines are too short, and the facial suture is omitted. As will be described later, features approaching some of these are to be found in immature specimens of Kloucekia (P.) harnagensis, but in a typical adult individual the first and second glabellar lobes are fused distally, the second furrows being slightly curved, convex forwards, and not reaching the axial furrows. These characters are con- sidered here to more in keeping with those of the subfamily Acastinae. The Barrande Collection of Bohemian trilobites at the British Museum (Natural History) contains several specimens of Kloucekia phillips: (Barrande), and examina- tion of these has shown that the cephala of Kloucekia and Phacopidina are indis- tinguishable. One specimen, In. 42327, has had the genal angle developed to show the existence there of an extremely small fixigenal spine, smaller even than that of Phacopidina, but situated in an analogous position. The only difference between Kloucekia and Phacopidina appears to lie in the latter’s possession of a mucronate pygidium, and it has been found convenient to separate the two subgenerically on this basis. Recently the name Dveyfussina was proposed (Choubert et al., 1956 : 394) as a subgenus of Phacopidina, though no adequate diagnosis or description was given. The type-species of Dreyfussina was originally described by Dreyfuss (1948 : 56, pl. 5, figs. 1-5, pl. 9, figs. 6, 7,9) as Dalmania exophtalma from the Caradoc of the Montagne Noire in southern France. His photograph of the cephalon is indistinct and difficult to match with his line drawings in the same paper, which show the second glabellar furrows diverging forwards to cut the axial furrows. In all other respects Dreyfussina resembles a typical Kloucekia and the two may well be synonymous. TYPE SPECIES. Phacopidina harnagensis by original designation of Bancroft (1949 : 310). Kloucekia (Phacopidina) harnagensis (Bancroft) (Pl. 49, figs. 9, 14; PI. 50, figs. 1-5) 1949. Phacopidina harnagensis Bancroft, p. 310, pl. Io, figs. 24, 25. 1958. Phacopidina harnagensis Bancroft: Dean, p. 201. 1959. Phacopidina harnagensis Bancroft : Struve in Moore, p. O.486, fig. 385, 4a, b. DESCRIPTION. The cephalic outline is sub-semicircular, bluntly pointed frontally ; the maximum breadth is almost twice the median length and is measured approxi- mately at the line of the pleuroccipital furrow. The glabella is pentagonal in outline, expanding anteriorly to attain its maximum breadth across the frontal glabellar lobe, the apex of which is bluntly pointed ; the median length is slightly more than the breadth. The axial furrows are almost straight, diverging forwards at about 30 degrees ; they are moderately deep posteriorly, becoming shallowest anteriorly where poorly-defined hypostomal pits are sited just forwards of the first glabellar furrows. There are three pairs of glabellar lobes and furrows. The first glabellar lobes are roughly triangular in shape, slightly larger than the other glabellar lobes. The first glabellar furrows are almost obsolete ; distally they cut the axial furrows, whence 322 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II they run adaxially and sharply backwards until they end longitudinally in line with the other glabellar furrows, leaving a smooth, median band. The second glabellar lobes are smaller, becoming shorter (exsag.) distally. They are bounded anteriorly by the second glabellar furrows which are faintly impressed, slightly curved, convex side forwards, set almost at right angles to the sagittal line, and which do not attain the axial furrows. The third glabellar furrows are deep, with apodemes, directed adaxially and slightly rearwards from the axial furrows, each extending across about one-third of the breadth of the glabella. The third glabellar lobes are subtriangular in outline, and in the internal mould they appear to constrict adaxially owing to the presence of apodemes in the occipital and third glabellar furrows. The frontal glabellar lobe is transversely lozenge-like in outline, with median length about half the breadth ; the straight, anterolateral margins converge frontally at 135 degrees, meeting at the bluntly-pointed apex. The anterior border of the cephalon is a thin, raised rim which becomes obsolete laterally, just beyond the line of the axial furrows, and is separated from the glabella by a narrow (sag.), moderately deep furrow con- taining the anterior branches of the facial suture. The lateral border furrows are represented by shallow depressions running parallel to the lateral margins and ex- tending from the axial furrows almost to the outer ends of the pleuroccipital furrow. The distal parts of the occipital furrow are transversely straight and deep with apodemes, but the median third is much shallower and set forwards. The occipital ring is transversely convex, longest medially, becoming shorter distally where it curves forwards slightly to form a pair of small occipital lobes. The pleuroccipital furrow is set slightly to the rear of the distal ends of the occipital furrow ; it is deep and broad (exsag.) adaxially, and extends distally in an almost straight line, becom- ing shallower towards, but not reaching, the lateral margins. The posterior border is narrow (exsag.) immediately outside the axial furrows, but becomes markedly wider (exsag.) abaxially where the blunt genal angles are produced rearwards to form a pair of small, thin genal spines. The latter are situated a short distance adaxially from the lateral margins and are seldom preserved intact. The eyes are crescentic in plan, and slightly divergent backwards. Their anterior ends are sited just outside the axial furrows, opposite the anterior halves of the first glabellar lobes, and they extend back and slightly outwards until they are level with the posterior halves of the second glabellar lobes. The ocular surfaces are schizochroal and that of the lecto- type, though damaged, has a total of about 115 facets on the right eye, arranged in twenty-three vertical rows. The palpebral lobes are broad (év.) and horizontal, de- limited adaxially by broad (i.), shallow palpebral furrows, and poorly defined frontally where they extend beyond the eye as far as the axial furrows. The fixigenae decline gently from the palpebral furrows to the axial furrows, but more steeply abaxially from behind the eyes to the lateral margin. The anterior branches of the facial suture are slightly convergent forwards, where they cut the axial furrows just in front of the hypostomal pits ; they then follow the outline of the frontal glabellar lobe and meet frontally at its apex. The posterior branches curve gently forwards from the rear ends of the eyes and then slightly rearwards, cutting the lateral margins approximately in line with the third glabellar furrows. A few immature cranidia have been found, one of which is illustrated (PI. 50, ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 323 fig. 4). They differ from those of typical adult individuals in the following respects : the glabella is slightly narrower across the third glabellar lobes, so that the axial furrows appear more divergent forwards ; the glabellar lobes decrease noticeably in size from front to rear; the second glabellar furrows are deep and transversely straight, intersecting the axial furrows; and the first glabellar furrows are better defined than in the adult form. The hypostoma is unknown. The thorax consists of eleven segments and is divided into three longitudinal lobes of approximately equal breadth by well-developed axial furrows. Each axial ring is moderately convex transversely and is curved forwards both medially and distally. The pleurae extend abaxially and horizontally for about half their length (¢7.) as far as the fulcra, where they turn downwards through almost a right angle, at the same time curving first rearwards and then forwards slightly towards the tips, which are truncated, convergent forwards. The anterolateral part of each pleural tip is bluntly rounded, but the posterolateral part is almost a right angle, produced distally to form a small, projecting process (see Pl. 50, fig. 1) which probably functioned as a “‘stop”’ during enrolment. The proximal half of the upper surface of each pleura is horizontal, but beyond the fulcrum the distal half forms a surface which intersects the proximal half obliquely so as to form a slight ridge extending from the anterior border near the fulcrum to the posterior border just inside the pleural tip. Each pleura carries a pleural furrow which commences at the intersection of the axial furrow and the anterior border of the pleura; the furrow then curves gently and abaxially rearwards, and finally forwards in the direction of, though without reaching, the pleural tip. The proximal half of the furrow is markedly deep, but beyond the oblique ridge near the fulcrum, as described above, it becomes markedly shallower for the remainder of its length (¢.). The pygidium is broadly subparabolic in outline, its anterior margin only moder- ately covex, and length, excluding the terminal spine, roughly half the maximum breadth. The axis is delimited laterally by pronounced axial furrows, does not reach the posterior margin, and there are seven axial rings, of which the first four are the most clearly defined. Each side-lobe has a pronounced facet situated anterolaterally, and five ribs separated from each other by well-impressed pleural furrows and carry- ing fainter rib furrows. All these furrows die out laterally, leaving a smooth border. The margins of the pygidium are entire and pass posteriorly into a conspicuous, smooth spine, in length about one-third of the pygidium and rising from the axis at an angle of about 30 degrees. Behind the tip of the axis is a small postaxial ridge which merges into the base of the terminal spine. In general the dorsal surface of the entire exoskeleton appears to be smooth but occasional well-preserved specimens show that the test, with the exception of the furrows and terminal spine, may be covered with fine, closely-set granules. HORIZON AND LOCALITIES. The type-specimens are from the section in the disused cartway by the south-eastern corner of Smeathen Wood, 250 yards west-south-west of Wood House, five-sixths of a mile south of Horderley Station. The horizon is in the lowest part of the Harnagian Stage, the zone of Reuscholithus reuschi. At this locality the species is fairly abundant and is associated with R. reuschi Bancroft, 324 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II Smeathenia smeathenensis (Bancroft) and Smeathenella harnagensis Bancroft. Kloucekia (Phacopidina) harnagensis occurs also, but in less abundance, at the field exposure 550 yards north-west of Woolston House, Woolston. No other localities are yet known. LECTOTYPE, here selected. BM. In. 42089 (Pl. 50, figs. I, 5). PARATYPE. BM. In. 42088 (PI. 50, fig. 3). FIGURED SPECIMENS. BM. In. 40131 (Pl. 40, figs. 9, 14); BM. In. 50576 (Pl. 50, fig. 2); BM. In. 52211 (PI. 50, fig. 4). Discussion. It has been found more convenient to group this section with that of Kloucekia (Phacopidina) apiculata (M‘Coy) later in this paper (see p. 326). Kloucekia (Phacopidina) apiculata (M‘Coy) (Pl. 50, figs. 6-10, 12) 1847. Phacops apiculatus MS. Sedgwick, p. 149. Nomen nudum. 1851. Poyrtlockia ? apiculata M‘Coy ex Salter MS., in Sedgwick & M‘Coy, p. 162, pl. 1G, figs. 18, 19 only. 1852. Phacops (Phacops) apiculatus Salter, p. iii. 1853. Phacops (Acaste) apiculatus Salter : Salter, p. 9. 1859. Phacops apiculatus Salter : Murchison, p. 75, fig. 3. 1864. Phacops (Acaste) apiculatus Salter: Salter, p. 28, pl. 1, figs. 36-38. 1867. Phacops apiculatus Salter : Murchison, p. 69, fig. 3. 1924. Dalmanitina apiculata (Salter) : McLearn, p. 166. 1939. “‘ Acaste’’ apiculata (Salter) : Stubblefield, p. 56. 1945. Dalmanitina apiculata (Salter) : Bancroft, p. 195. 1947. Scotiella apiculata (Salter) : Harper, p. 169. pl. 6, figs. 6, 9. 1949. Phacopidina apiculata (Salter) : Bancroft, p. 310. 1958. Phacopidina apiculata (Salter) : Dean, pp. 206-207, 221-222. 1959. Phacopidina apiculata (M‘Coy) : Dean, p. 146. 1959a. Phacopidina apiculata (M‘Coy) : Dean, pp. 212, 214, 220. DESCRIPTION. The cephalic outline is almost semicircular, its length about two- thirds of the maximum breadth which is measured just in front of the pleuroccipital furrow. The glabella is pentagonal in plan, expanded anteriorly and bluntly pointed in front ; its length is about equal to, or slightly greater than, its maximum breadth. There are three pairs of glabellar furrows and glabellar lobes. The first glabellar furrows are shallow, in some cases almost obsolete, particularly adaxially, and run almost straight inwards and rearwards from the axial furrows. The first glabellar lobes are subtriangular in form, whilst the second glabellar lobes are subrectangular. The two pairs of lobes become fused distally, but proximally they are separated by shallow, slightly curved second glabellar furrows, convex forwards and set at right angles to the median line of the glabella. The third glabellar furrows, deep and with apodemes, are straight, directed inwards and rearwards from the axial furrows for about one-third of the glabellar breadth. The third glabellar lobes are subtriangular in outline, not unlike the slightly larger first glabellar lobes. All the glabellar lobes and the frontal glabellar lobe are moderately convex in line with one another. The axial furrows are almost straight, becoming slightly convex laterally opposite the first glabellar lobes; they are shallow medially, deepening to front and rear, and ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 325 inconspicuous hypostomal pits occur opposite the distal ends of the first glabellar furrows. The glabella is bounded frontally by a narrow (sag.) groove, representing the preglabellar furrow and field, and immediately in front of this rises the anterior border, uniformly narrow (sag.) and parallel to the margin of the frontal glabellar lobe. The occipital furrow is curved forwards and shallow medially, deepening laterally where a pair of apodemes is situated just inwards from the axial furrows. The occipital ring is transversely convex, longest medially but shortening (exsag.) distally where it curves forwards slightly so as to form a pair of occipital lobes. The pleuroccipital furrow is deep and moderately wide (exsag.) near the axial furrows but becomes shallower distally and finally dies out before reaching the lateral margins. The posterior border is narrow by the axial furrows but widens (exsag.) distally and terminates posterolaterally in a pair of almost right-angled genal angles which are produced into a pair of small, sharp fixigenal spines directed backwards and slightly outwards (see Pl. 50, fig. 6). The largest-known fixigenal spine, about 2-5 mm. long, belongs to a cephalon Ir mm. in median length. The eyes are crescentic in plan with prominent palpebral lobes which commence at the axial furrows, immediately behind the distal ends of the first glabellar furrows, and run rearwards, curving first abaxially and then adaxially until they terminate opposite the frontal half of the third glabellar lobes and a short distance outside the axial furrows. Moderately-deep palpebral furrows run parallel to the margins of the palpebral lobes and die out as they skirt the rear ends of the eyes. The eye surfaces are schizochroal and, in the case of a cephalon 8 mm. long, each contains at least thirty-two vertical rows of up to eight facets. From the palpebral lobes the proximal surface of the fixigenae slopes steeply down to the axial furrows ; the distal surface is rather less steeply declined to the poorly-defined marginal furrows, situated just inwards from the lateral margins. The anterior branches of the facial suture are at first virtually coincident with the axial furrows, then skirt the frontal glabellar lobe and finally coalesce at the apex of the latter. The posterior branches of the suture are sigmoidal, curving gently forwards from the ends of the eyes and then back- wards to cut the lateral margins opposite the mid-point of the third glabellar lobes. The hypostoma is unknown. No syntype thorax was available to Harper (1947) in his redescription, and only one, moderately well-preserved Shropshire specimen has been found (PI. 50, fig. 12). This contains eleven thoracic segments and appears to agree in all essentials with the thorax of Kloucekia (Phacopidina) harnagensis already described (see p. 323), although full details of the pleural points cannot be distinguished. The pygidium, excluding terminal spine, is subparabolic in outline, with frontal margin moderately convex forwards. Frontally the axis occupies between one-quarter and one-third of the maximum breadth and it is bounded laterally by deep axial furrows which converge gently rearwards. A maximum of eight axial rings is present but beyond the fifth ring these are often difficult to distinguish, particularly in internal moulds. Each side-lobe is convex with an anterolateral facet, and there are five deep pleural furrows. The ribs so formed carry faintly-impressed rib furrows and these, together with the pleural furrows, do not reach the lateral margin, extending only as far as the smooth border which also defines the extent of the doublure. At the tip of the 326 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II pygidium the border is produced to form a thorn-like terminal spine, directed up- wards and backwards (see Pl. 50, fig. 8) and equal to almost one-third of the length of the pygidium proper. The axis extends only to the smooth border but is con- nected to the base of the terminal spine by a low postaxial ridge. The surface of the exoskeleton is apparently smooth in most individuals, but occasional specimens show trances of fine granulation, at least on the surface of the cephalon. HoRIZoN AND Loca itis. In south Shropshire Kloucekia (Phacopidina) apiculata is first encountered in the lowest subdivision of the Lower Longvillian Substage, the zone of Dalmanella horderleyensis. It then ranges upwards through the remainder of the Lower Longvillian, though not in large numbers. In the succeeding Upper Longvillian Substage the species is more abundant, although, like the remainder of the fauna, it is generally most common in calcareous bands within the large thickness of flags and sandy shales constituting the Alternata Limestone and Lower Cheney Longville Flags. In an earlier paper (Dean, 1958 : 223) K. (P.) apiculata was recorded from the Marshbrookian Stage, but the evidence for this was fragmentary and un- satisfactory, and it is still not known for certain whether the species occurs outside the limits of the Longvillian Stage of which it is so characteristic. The species is to be found at numerous points along the Longvillian outcrop throughout the Caradoc District but the following may particularly be noted. Lower Longvillian Substage: Long Lane Quarry, two-thirds of a mile south-west of Cheney Longville ; the large road-side quarry 280 yards south-east of Glenburrell Farm, Horderley ; various exposures near the top of Rookery Wood, just south of the River Onny three-quarters of a mile south-east of Horderley ; the quarry by the east side of the road 590 yards east-north-east of Cwm Head Farm, one and a half miles north-east of Horderley. Upper Longvillian Substage : various exposures along Longville Lane, from 400 to 440 yards north-west of the “‘ Earthwork ”’ at Cheney Longville ; the old quarry below the north side of the track at the north-eastern corner of Burrells Coppice, one mile south-east of Horderley Station; various exposures by the north side of the Bishops Castle road to the east of New House, just east of Horderley; the stream-bed by the west side of the Horderley- Marshbrook road, 560 yards north-east of Crosspipes ; Soudley Quarry, near Soudley Pool, between Hope Bowdler and Ticklerton. LECTOTYPE, selected by Harper (1947 : 170). Sedgwick Mus. A. 10962, a cephalon originally figured by M‘Coy (im Sedgwick & M‘Coy, 1851, pl. 1G, fig. 18) from “Caradoc Beds” near Glyn Ceiriog, Montgomeryshire, and probably of Longvillian age. FIGURED SPECIMENS. BM. In. 32253 (PI. 50, fig. 6) ; BM. In. 51259 (PI. 49, figs. 7, 8); BM. In. 51260 (PI. 50, fig. 9) ; BM. In. 50558 (PI. 40, fig. 12). Discussion. Kloucekia (Phacopidina) apiculata is one of a number of fossil species, including also Brongniartella bisulcata and B. rudis (see p. 345), the authorship of which has usually been attributed to Salter. Salter had, in fact, prepared descriptions of some new species which were to be published as an appendix to Sedgwick & M‘Coy’s great work on the British Palaeozoic rocks and fossils. Publication of the appendix was delayed until July, 1852, and in the meantime M‘Coy had published, in May, 1851, his own longer and more detailed descriptions, using Salter’s manuscript ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 327 specific names but frequently changing the genera to which they were assigned. There is no question of M‘Coy’s having used Salter’s descriptions, and Sedgwick (zm Sedgwick & M‘Coy, 1851 : xill—xiv) stated that ““With the exception of the Appendix to the Second Fasciculus (which is from the pen of my friend and former fellow- labourer Mr. Salter) all the descriptions of the following works are due, exclusively, to Professor M’Coy ... ’’. M’Coy must therefore be considered the author of the species in question. Salter’s descriptions in the Appendix (1852) refer to the plates published in the preceding year and M‘Coy’s Portlockia ? apiculata was described by Salter as Phacops (Phacops) apiculatus sp. nov. The lectotype chosen by Harper is one of the specimens figured by M‘Coy in 1851. The specific name as used for the first time by Sedgwick (1847 : 149) was invalid. Although no other forms of Kloucekia have been found in Shropshire between K. (P.) harnagensis in the lower Harnagian Stage and K. (P.) apiculata in the Lower Longvillian Substage, nevertheless the two species appear to be closely related. They may be distinguished by the fact that K. (P.) harvnagensis has a slightly shorter pygidium, and that the eyes of K. (P.) apiculata are larger and less divergent back- wards. A feature common to immature individuals of both species is the lateral extension of the second glabellar furrows, which are transversely straight, to cut the axial furrows. The only other described British species of Kloucekia, originally Scotiella major (Harper, 1947 : 171, pl. 6, fig. 11), was separated by means of its short cephalon, the length of which is equal to less than half the breadth. K. (P.) major was described from the Lower Longvillian Substage of Caernarvonshire. Both K. (P.) apiculata and K. (P.) major were assigned to the more typically Silurian genus Scotiella by Harper (1947 : 169), but recently Struve (7 Moore, 1959 : 0.486) has maintained Phacopidina as a separate genus. The type species of Scotzella (Delo, 1935 : 409; 1940: 33) is S. logani (Hall) from the Silurian of Nova Scotia, and differs from typical species of Kloucekia in the following respects: the first glabellar furrows are obsolete ; the second glabellar furrows of adult individuals cut the axial furrows; the anterior branches of the facial suture converge frontally from the eyes to the frontal lobe ; fixigenal spines are totally absent and the genal angles are rounded. The trilobite described by Barrande (1852 : 556, pl. 26, figs. 44, 45; pl. 27, figs. 12-14) as Dalmanites solitaria, from Stage D of Bohemia, has been placed doubtfully in Phacopidina by Struve (im Moore, 1959 : 0.486, fig. 385). Struve’s illustration shows the second glabellar furrows definitely cutting the axial furrows, though this is not apparent from Barrande’s original figures. D. solitaria possesses amucronate pygidium, and examination of a topotype specimen, BM. In. 42336, from Gross Kuchel, Czechoslovakia, suggests there is little doubt that the species is a typical Kloucekia (Phacopidina) not very different from the Anglo- Welsh species. Phacops cf. apiculatus was recorded by King (1m Wedd e¢ al., 1929 : 61) from the Ashgill Series of the Oswestry district, but the specimen in question, GSM. WI. 218, is a small pygidium of Phillipsinella parabola (Barrande), and it is extremely doubt- ful whether a true Kloucekia (Phacopidina) apiculata has been found higher than the Caradoc Series. GEOL. 5. 8. 30 328 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II Family PTERYGOMETOPIDAE Reed, 1905 Subfamily PTERYGOMETOPINAE Reed, 1905 Genus CALYPTAULAX Cooper, 1930 The genus Homalops was introduced in 1884 by Remelé (p. 200), though only briefly diagnosed, and was figured by him in the following year (1885 : 25, fig. 3). Remelé’s species Homalops altumi was redescribed later by Wiman (1907 : 130, 131, pl. 8, figs. 7-10) who considered Homalops to be a subgenus of Phacops. In the meantime Reed (1905 : 226-228) had included Homalops in his subfamily Pterygo- metopinae but had considered it to be too indifferently defined to be retained, and quoted as reference only Remelé’s 1884 paper. Later Kozlowski (1923 : 29) rejected the genus on the grounds that it had never been adequately described or figured, and was apparently unaware of either Remelé’s 1885 work or that of Wiman. In a later classification of the Phacopidae Reed (1927 : 352, 353) omitted all mention of Homalops. Wiman’s more than adequate illustrations show that Homalops Remelé is identical with the later described Calyptaulax Cooper, 1930. It is clear that the name cannot, therefore, be rejected on the grounds of insufficient diagnosis, and certain Swedish material has, in fact, been referred to Homalops by Jaanusson (1953 : 102), but Homalops Remelé, 1884 is a junior homonym of both Homalops Burmeister, 1840 and Homalops Motschoulsky, 1850, and is now regarded as a subjective synonym of Calyptaulax. TyPE spEcIES. Calyptaulax glabella by original designation of Cooper (1930 : 387). Calyptaulax actonensis sp. nov. (PIN 50; figs. 11,135 74); PI 51) figs: 1, 2) 1958. Calyptaulax sp. Dean, pp. 211, 212, 224. DiaGnosis. Cranidium typical of the genus. Pygidium broader than long, with six interpleural furrows, up to twelve axial ring furrows, and terminating in a blunt, upturned point. DeEscrIPTION. The glabella is depressed, roughly pentagonal in outline, its length about equal to the maximum breadth which is measured across the anterolateral margins of the first glabellar lobes. The frontal glabellar lobe is bluntly pointed frontally, alate, and extends abaxially to the anterolateral tips of the first glabellar lobes where it meets the axial furrows. There are three pairs of glabellar lobes which diminish in size from front to rear. The first glabellar lobes are long, each almost an isosceles triangle in outline; they are delimited frontally by moderately deep first glabellar furrows which are slightly sigmoidal and widely divergent forwards. The second glabellar lobes are small, only about one-quarter the length of the first pair, and are directed abaxially rearwards parallel to the second glabellar furrows which are shallow, curved, convex forwards, directed slightly rearwards but not reaching the axial furrows, the distal third of the second glabellar lobes being fused with the first pair of lobes. The third glabellar furrows are deep, diverging posteriorly and quickly bifurcating so as to form two branches which isolate the third glabellar ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 329 lobes ; the latter, slightly smaller than the second lobes, are thus present only as detached tubercles at the distal ends of the basal glabellar segment. The axial furrows are shallow, diverging forwards at 50-55 degrees for most of their length but finally becoming slightly convergent forwards around the anterior end of the first glabellar lobes. The anterior border has not been found preserved, but there appears to be a shallow preglabellar furrow bordering the anterior margin of the frontal glabellar lobe. The occipital furrow is shallow medially, deepening distally where it is flexed backwards and then abaxially so as to accommodate the posterior margins of the third glabellar lobes. The occipital ring is transversely convex, long medially becoming shorter (exsag.) distally where it curves forwards slightly to form a pair of ill-defined occipital lobes. Immediately outside the axial furrows the pleur- occipital furrow is moderately deep and broad (exsag.), delimiting the narrow (exsag.) posterior border which becomes still narrower distally. Neither the furrow nor the border has been found preserved beyond the line of the palpebral lobes. The fixigenae are broad and flat, bounded abaxially by well-defined palpebral furrows ; the latter appear narrow on the external mould but broader on the internal mould and extend from near the anterior ends of the first glabellar lobes to the pleuroccipital furrow. The palpebral lobes are broad, flat, level with the fixigenae and crescentic in plan. The ocular surfaces of the eyes have not been found. The anterior branches of the facial suture appear to cut the axial furrows immediately in front of the palpebral lobes ; thence they follow the anterolateral margins of the frontal glabellar lobe, converging at its apex. The posterior branches of the suture, together with the outermost parts of the fixigenae, have not been found preserved. The hypostoma and thorax are unknown. The pygidium is subtriangular in outline, its median length equal to two-thirds of the maximum breadth which is measured across the almost straight anterior margin ; it tapers to a blunt, slightly upturned point, the extreme tip of which has not been seen intact. The axis occupies just less than one-third of the frontal breadth of the pygidium and its sides converge rearwards for about two-thirds of their length, after which they continue almost parallel to the tip of the axis. Owing to the state of preservation, in a sandstone, of the holotype pygidium only eight axial rings are visible but other specimens, found in a fine-grained argillaceous matrix, have eleven or twelve rings. The axial rings themselves are slightly sigmoidal in plan, curving back both medially and distally. Beyond the tip of the axis is a low, smooth, postaxial ridge which runs into the terminal point. On each side- lobe of the pygidium there are six well-defined interpleural furrows ; these decrease in depth from first to sixth and extend about two-thirds of the distance from the axial furrow to the lateral margin, apparently ending in line with the inner margin of the doublure. An equal number of rib-furrows is present, each commencing about one-third of the distance from the axial furrow to the lateral margin and then turn- ing rearwards markedly until it intersects the margin. The first pair of rib-furrows is deeply defined, but the remainder become progressively fainter towards the tip of the pygidium. A well-developed facet is present anterolaterally on each side-lobe, immediately in front of the first rib-furrow. Fragmentary evidence suggests that the surface of the test of both glabella and 330 ORDOVICIAN DRILOBITE PAUNAS! OF SOUDH SHROPSHI REs in pygidium, excepting the furrows, is covered with fine, evenly-distributed granules. HORIZONS AND LOCALITIES. The type-locality is the site of the old quarry, now filled in, at Quarry Field, Gretton, half a mile east of Cardington. The horizon is in decalcified sandstones belonging to the middle third of the Actonian Stage. The species has also been found in the disused quarries 80 yards south-east and 250 yards east of St. Margaret’s Church, Acton Scott. At one time Calypiaulax actonensis was thought to occur only in the middle Actonian of south Shropshire (Dean, 1958 : 211), but whilst it is still true that the species is most abundant at this level, it has now been found in the lowest Actonian strata at the large exposure in the south bank of the River Onny about 85 yards west of its junction with Batch Gutter stream, and also in the topmost Actonian and basal Onnian beds exposed in the north bank of the Onny about 30 yards east of the same river/stream junction. Specimens are rare at these latter localities and have not been found higher in the succession. Hororyver. BM) In. 40770 (Pl. 50, figs. 11, 24): PARATYPES. BM. In. 48492 (PI. 50, fig. 13) ; BM. In. 49318 (Pl. 51, fig. 1); BM. In. 49768 (Pl. 51, fig. 2). Discussion. Calyptaulax is a predominantly North American trilobite genus and is figured here for the first time from Shropshire. C. actonensis bears some resemblance to C. altumi as figured by Wiman (1907, pl. 8, figs. 7-10) but differs in having slightly wider fixigenae and a longer pygidium. I am indebted to Dr. Valdar Jaanusson of Uppsala University for the opportunity of comparing my Shropshire specimens with material collected from a South Bothnian drift boulder and determined by Professor Wiman as C. altwmi (Remelé). Dr. Jaanusson kindly informs me these specimens are derived from strata belonging to the zone of Pleurograptus linearis, that is to say, beds younger than those of south Shropshire which belong to the zone of Dicrano- graptus clingant The cranidium of the Norwegian species Calyptaulax norvegicus (Stormer, 1945 : 417, pl. 4, figs. 2, 3) is virtually indistinguishable from that of C. actonensis, but the pygidium differs in having a narrower, more uniformly tapered axis with slightly longer (sag.) axial rings. In outline the pygidium of the new species resembles that figured by Jaanusson (1953: 103, fig. 4) as Homalops cf. altwm: Remelé from the Slandrom Limestone of Dalarne, a horizon belonging to the zone of Pleurograptus linearis but containing, nevertheless, a fauna resembling in many respects that of the Actonian Stage in South Shropshire. The Swedish pygidium is incomplete but appears to have fewer axial rings and, probably, rib-furrows. Under the name “ Subgenus A ”’ Struve (1m Moore, 1959 : 0.491, 492) has recently separated from Calyptaulax (s.s.) a cranidium of Calyptaulax altumi from a Pleisto- cene erratic in Sweden, together with the pygidium recorded from the Slandrom Limestone by Jaanusson as Homalops cf. aliwmi and noted above. The latter pygi- dium compares closely with that of C. actonensis, a species which, in turn, possesses a cranidium indistinguishable, even subgenerically in the writer’s opinion, from the type-species of Calyptaulax. According to Struve’s illustration the cranidium of “Subgenus A ”’ possesses palpebral lobes which become markedly narrow frontally and second glabellar lobes which are present only as detached tubercles. These features may possibly merit subgeneric separation but they have not yet been fully OREDIOWG TAING Wh elie ONS teh Bs EyANUINEATS OF SiO iH eS FIROm SiH DRIES ih 33n investigated and are not, in fact, present on the specimens of C. altwmz lent by Dr. Jaanusson, which are best grouped with Calyptaulax (s.s.). Calyptaulax compressa (Cooper, 1930, pl. 5, figs. 7, 8) has a shorter glabella and smaller fixigenae than C. actonensis, whilst C. glabella (Cooper, 1930, pl. 5, figs. 9-11) has only nine well- defined axial rings on the pygidium. C. schucherti Troedsson (1929, pl. 19, figs. 17-20) from the Cape Calhoun Formation of Greenland may be distinguished from the Shropshire species by its shorter frontal lobe and smaller fixigenae and first glabellar lobes. Subfamily CHASMOPSINAE Pillet, 1954 Genus CHASMOPS M‘Coy, 1849 TYPE SPECIES. Calymene odim Eichwald, 1840 by original designation of M‘Coy (1849 : 403). Chasmops extensa (Boeck) (Bis Er digsy 310 sl 52, figsax, 35°45 017) 1838. Tvrilobites extensus Boeck, p. 139. 1839. Asaphus Powisit Murchison, p. 661, pl. 23, fig. 9a and b only. 1848. Dalmania affinis Salter, p. 337, pl. 5, fig. 5. 1849. Phacops truncato-caudatus var. 8, affinis Salter, p. 7. 1851. Odontochile truncato-caudata (Portlock) M‘Coy in Sedgwick & M‘Coy, p. 162, pl. 1G, figs. 20, ? 21. 1853. Phacops conophthalmus (Boeck) : Salter, pp. 7, 11. 1854. Phacops conophthalmus (Boeck) : Murchison, pl. 4, figs. 11, 12. 1854. Phacops conophthalmus (Boeck) : Salter & Aveline, p. 67. 1854. Phacops truncato-caudatus (Portlock) : Salter & Aveline, p. 67. 1859. Phacops conophthalmus (Boeck) : Murchison, pl. 4, figs. 11, 12. 1864. Phacops (Chasmops) macroura Sjogren : Salter, p. 37, pl. 4, figs. 18-23. 1864. Phacops (Chasmops) conophthalmus Boeck? (pars), Salter, p. 40, pl. 4, fig. 24. 1867. Phacops conophthalmus (Boeck) ; Murchison, pl. 4, figs. 11, 12. 1872. Phacops conophthalmus (Boeck) : Murchison, pl. 4, figs. 11, 12. 1884. Phacops conicophthalmus (Boeck) : La Touche, p. 57, pl. 3, figs. 65, 65a. 1884. Phacops macroura Sjogren: La Touche, p. 57, pl. 3, fig. 66. 1938. Chasmops cf. macroura (Sjogren) : Stubblefield in Pocock ef al., pp. 90, 255. 1940. Chasmops extensa (Boeck) : Stormer, p. 138, pl. 4, figs. 7-11. 1945. Chasmops macroura (Sj6gren) : Bancroft, p. 183. 1953. Chasmops extensa (Boeck) : Stormer, pp. 65, 68-69, 87, 130, 134. 1958. Chasmops sp., Dean, pp. 222-224. 1959a. Chasmops extensa (Boeck) : Dean, p. 221. DEscriPTiIon. The cephalon is longitudinally and transversely convex with median length almost three-fifths of the maximum breadth. The glabella is large, almost subpentagonal in outline, strongly convex, its length about three-quarters of the maximum breadth measured across the frontal glabellar lobe. There are three pairs of lateral glabellar lobes. The first glabellar furrows diverge forwards at about 110 degrees ; at their mid-points they are slightly concave forwards but abaxially they curve sharply backwards through almost a right angle, becoming deeper and finally cutting the axial furrows, The first glabellar lobes are extremely large by comparison 332 ORDOVICIAN DRIVPOBIDE PAUNAS (Ob SOUDEH SHOPS Hi RE al with the other pairs of lobes ; they are roughly triangular in plan, expanding enor- mously in length (exsag.) from just in front of the second glabellar lobes and indenting the frontal glabellar lobe posterolaterally. Behind the first glabellar lobes a pair of deep furrows extends adaxially, finally bifurcating to form the second and third glabellar furrows which appear then as notches delimiting, both frontally and pos- teriorly, the small tubercle-like projections which serve as second glabellar lobes. Third glabellar lobes are represented by the slightly curved-forwards, posterolateral extremities of a basal glabellar segment which extends abaxially almost to the posterior ends of the first glabellar lobes. When preserved as internal moulds the third glabellar lobes appear as almost detached, rounded, knob-like projections. The frontal glabellar lobe is large, strongly convex, and broadest frontally where the breadth is about twice the length. Occasionally, cephala of Chasmops extensa pre- served as internal moulds exhibit a V-shaped group of tubercles, directed backwards and extending from the preglabellar furrow to a position opposite the frontal portion of the first glabellar lobes. This structure was noted in specimens from Shropshire by Salter (1864 : 39, pl. 4, fig. 20), and a similar feature has been figured by Harrington (n Moore, 1959, fig. 69E) for Chasmops odimi (Eichwald) from Estonia and termed “auxiliary impressions ’’’, the exact function of which is not yet clear. Schmidt (1881, pl. 2, figs. 2, 3, 10; pl. 5, figs. 1, 9) has figured analogous structures in other species of Chasmops. The frontal margin of the frontal glabellar lobe is indented medially by a deep, broad (sag.) preglabellar furrow which curves gently forwards and then slightly backwards abaxially, at the same time becoming shallower and almost dying-out where it meets the axial furrows. The latter diverge forwards at an average angle of about 40 degrees but are laterally convex in part so as to accom- modate the distal margins of the first glabellar lobes ; they are generally deep and narrow but become markedly shallower medially, just to the rear of centre of the first glabellar lobes. The preglabellar field is convex, becoming thicker medially, coincident with the concave margin of the frontal glabellar lobe ; distally it narrows (exsag.) slightly, and then maintains a uniform breadth along the anterior margin of the frontal glabellar lobe, finally curving around the tips of the latter to meet the axial furrows. The anterior border furrow is narrow (sag.) frontally, becoming broader and shallow abaxially where it curves sharply backwards to intersect the axial furrows. The anterior border is uniformly convex forwards in plan and has a flattened upper surface declined frontally ; the breadth (sag.) is generally uniform but at the line of the axial furrows it expands to form V-shaped projections extending into the anterior parts of the axial furrows themselves. Beyond the axial furrows the border narrows (exsag.) again and is separated from the fixigenae by broad (év.), moderately-deep lateral furrows which extend rearwards to meet the distal ends of the pleuroccipital furrow. The occipital ring is long (sag.), transversely convex but flattened longitudinally, curving forwards distally to form a pair of large occipital lobes which are poorly defined laterally by the ends of the axial furrows. The occipital ring is bordered frontally by a deep, uniformly broad (sag.) occipital furrow. The pleuroccipital furrow is deep and broad (exsag.), curving rearwards slightly distally where it intersects the shallower lateral border furrows without reaching the lateral margins, The posterior border is narrow (exsag.) immediately ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 333 outside the axial furrows, becoming still narrower mid-way to the lateral margins and then expanding rapidly towards the genal angles which are produced backwards to form large fixigenal spines. The latter are about as long as the cephalon (sag.) and their upper surface is flattened, declined distally at almost a right angle ; the lateral margins are longitudinally straight and the proximal margins run abaxially backwards to meet them and form pointed tips. The smaller, proximal halves of the fixigenae decline steeply from the palpebral lobes to the axial furrows; the distal halves are plump, quadrant shaped, and steeply declined to the lateral border furrows. The schizochroal eyes are semicircular in plan, standing higher than the median lobe of the glabella; they are highest medially, becoming lower both anteriorly and posteriorly, and the visual surface is steeply declined abaxially. The eyes are not generally preserved whole, but the left eye of a cephalon, In. 52166, with a median length of 32 mm. possesses 350 facets arranged in twenty-nine vertical rows containing from four to sixteen facets. Immediately below each eye there is a broad, pronounced eye platform, as long as and parallel to the visual surface. The palpebral lobes are broad, strongly convex distally in plan, steeply declined adaxially and delimited by deep, broad palpebral furrows. Each palpebral furrow follows an outwardly-directed, V-shaped course above the apex of the eye, curving around both ends of the eye to join the eye-platform below. The anterior branches of the facial suture run almost straight forwards from the eyes, cutting first the eye- platforms near their anterior ends, crossing the fixigenae, and cutting the axial furrows mid-way between the distal ends of the first glabellar furrows and the lateral extremities of the frontal glabellar lobe; they follow the axial furrows forwards and then curve sharply adaxially, first cutting the preglabellar field and finally coin- ciding with the anterior border furrow until they meet frontally. The posterior branches of the facial suture are placed in shallow grooves suggestive of those found in the other phacopid genera Pterygometopus and Estoniops, and run abaxially from the rear ends of the eyes, at first almost straight, cutting the eye-platforms ; they then curve gently forwards and, finally, more markedly rearwards, meeting the lateral margins at a sharply acute angle. The surface of the glabella and, to a lesser degree, of the occipital ring is covered with scattered, coarse tubercles each of which, on magnification (see Pl. 52, fig. 6), is seen to comprise a cluster, generally about fifteen to twenty-five, of smaller granules. Granules of similar size are thinly distributed between the tubercles, and become still sparser across the proximal ends of the first glabellar lobes. The surface of the preglabellar field is conspicuously distinct from that of the glabella, being covered with closely-set, uniformly distributed granules (Pl. 52, fig. 1). A similar ornamentation is found on the test of the anterior border (not preserved in Pl. 52, fig. 1) and extends posterolaterally along the lateral borders. The proximal surface of the librigenae is similar to that of the glabella, but the tubercles are less well defined. The test of the posterior border is finely granulate, particularly the anterior half which appears almost smooth. Abaxially all the granules disappear and the test of the genal spines is quite smooth. The deepest parts of all the cephalic furrows, together with the proximal parts of the eye-platforms, are also smooth. The hypostoma, rarely found unbroken, is roughly triangular in plan, its maximum 334 ORDOVICIAN TRILOBILTE FAUNAS OF SOUTH SHROPSHIRE, SLL breadth, measured across the anterior wings, approximately equal to, or slightly less than, the median length. The anterior border, barely differentiated from the median body, is rounded, convex forwards, and continuous in line with the anterior wings which form conspicuous lateral projections and are twisted through almost a right-angle near their tips. A short distance behind the anterior wings is a pair of ill-defined projections which probably represent attenuated posterior wings. The lateral border is narrow, passing into the posterior border which forms a flattened, bluntly pointed, spatulate extension occupying almost one-quarter of the median length of the hypostoma. The median lobe is moderately tumid, about three- quarters of the total length, narrowing rearwards where it is delimited by the trans- verse median furrow. The last-named is almost obsolete medially but deepens laterally where it runs forwards to meet the lateral border just to the rear of the anterior wings. Immediately behind, and parallel to, the transverse median furrow is a narrow posterior lobe, poorly defined medially but expanding laterally and for- wards to give rise to a pair of oblique maculae sited opposite and just to the rear of the posterior wings. A shallow, posterior transverse furrow, also parallel to the transverse median furrow, separates the posterior lobe from the posterior border and becomes deeper laterally and forwards. A few better-preserved specimens show traces of coarse tubercles on the ventral surface of the median lobe. The thorax is known so far from only a few incomplete segments, none of which has the tips preserved. Each axial ring is of similar size to the occipital ring, being transversely broad and convex but longitudinally flat ; the anterior and posterior margins are, for the most part, transversely parallel except near the deep axial fur- rows where the distal portions curve forwards slightly to form a pair of small lobes. The articulating half-ring is large, about two-thirds the length (sag.) of the axial ring, from which it is separated by a deep, narrow (sag.) articulating furrow. Each pleura is of uniform breadth (exsag.), equal to that of the axial ring, as far as the fulcrum, beyond which point its form is not known. A deep, narrow (exsag.) pleural furrow commences at the junction of the anterior margin with the axial furrow, then curves gently and abaxially back as far as the fulcrum where it begins to curve forwards again, its subsequent course being unknown. The pygidium is large, elongate, subparabolic in outline, its anterior margin only gently convex forwards. Strongly arched transversely, and gently declined pos- teriorly, it terminates in a bluntly-pointed, slightly upturned tip. Frontally the axis occupies about one-third of the projected breadth, is straight-sided, and tapers uniformly rearwards, extending for about five-sixths of the length of the pygidium. The axial furrows are, for the most part, well defined but become obsolete near the end of the axis which passes into a transversely convex postaxial ridge extending to the tip of the pygidium. The largest available specimen, about 36 mm. long, has sixteen axial rings which do not reach the tip of the axis. The ring-furrows are generally well defined, becoming less so after about the twelfth, and some of the first few axial rings have their posterior margins indented medially, revealing what would be the corresponding articulating half-ring on the unfused segments of the thoracic axis (see Pl. 51, figs. 5, 7). Each side-lobe carries up to sixteen ribs which become progressively less transverse in direction from front to rear until the ORDOVICIAN ST Ri OBIE WAUINAS JO (SOUDH SEH ROP SM TRI, Il 335 final rib is parallel to the sagittal line. The ribs are separated by deep, narrow (exsag.) pleural furrows which are gently sigmoidal in plan and become almost obsolete at what is probably the inner margin of the doublure, so that an almost smooth border results. The hindmost pleural furrows are fainter than the rest and the final pair encroaches upon the sides of the postaxial ridge which may, in conse- quence, have a slightly waist-like appearance. Each rib is divided into two almost equal bands by a faint interpleural furrow which commences at the junction of the posterior margin of the rib with the axial furrow (see Pl. 52, fig. 7) and does not extend so far distally as do the adjacent pleural furrows. The pygidial margin is entire, arched upwards strongly at the tip. The surface of each axial ring is covered with fine, evenly-distributed granules at and near the sagittal line. Similar granules cover the surface of the posterior band of each rib but tend to die away on the anterior band (see PI. 52, fig. 7). The border and postaxial ridge are also finely gran- ulate but the axial furrows and articulating furrow are quite smooth. HORIZONS AND LOCALITIES. The earliest known specimens of Chasmops extensa in south Shropshire have been collected from the upper portion of the Upper Long- villian Substage, the Lower Cheney Longville Flags. At this horizon C. extensa, though not abundant, is found in association with Kyaerina typa Bancroft, Dolerorthis duftonensis (Reed) and Kloucekia (Phacopidina) apiculata (M‘Coy). It may be found at a number of localities along the River Onny Valley, about a mile south-east of Horderley Station, and along Longville Lane, between 35 and 120 yards north-west of the Earthwork at Cheney Longville. The species is more abundant in the Marshbrookian Stage, or Upper Cheney Longville Flags, particularly the lowest two-thirds. At this horizon it is to be found throughout much of the Caradoc Area, the most prolific localities including: the Onny Valley east of Burrells Coppice, south-east of Horderley ; the exposure 250 feet north-north-east of Woolston House, Woolston ; numerous exposures in the vicinity of Marshbrook, including the well-known Marsh Wood quarry; and the stream-section 550 yards south-south-west of Common Farm, Wallsbank. Chasmops extensa occurs sporadically throughout the mudstones of the Actonian Stage in the Onny Valley, north-east of Cheney Longville, certain levels being occasionally marked by an abundance of individuals. Some of these may attain a large size and frequently have the test preserved. Northwards from the Onny the species is to be found at many places where the Actonian Stage crops out, but generally tends to be less abundant in the arenaceous rocks of the Stage. Localities include : the south bank of the River Onny about 80 yards west of its junction with Batch Gutter stream ; the upper reaches of Batch Gutter itself, half a mile north- west of Wistanstow ; various quarries in the vicinity of Acton Scott village; ex- posures around the village of Plaish, 5 miles north-east of Church Stretton ; and the old quarry (now filled in) at Gretton, half a mile east of Cardington. The last two of these are in the arenaceous development of the Acton Scott Beds, probably representing the middle part of the Actonian Stage. The stratigraphically youngest specimens of Chasmops extensa in south Shropshire are known from the lowest zone of the Onnian Stage in the Onny Valley, in association with Onnia? cobboldi (Bancroft). A typical locality is in the north 336 ORDOVICIAN TRILOBITE BAUNAS OF SOULE SHROPSHIRE, ii bank of the river 35 yards east of its junction with the stream of Batch Gutter. FIGURED SPECIMENS. BM. In. 49126 (PI. 51, fig. 6) ; BM. In. 49127 (PI. 51, figs. 4; 8; 10); Plo 52, figs: 1,16): BM any 40128 Pls, tess; On ele 52 pecan) etal In. 50546 (Pl. 51, fig. 7); BM. In. 50547 (PI. 51, fig. 3); BM. In. 50548 (Pl. 52, fig. 4); BM. In. 50549 (PI. 52, fig. 3). Discussion. Chasmops extensa was first introduced as a species by Boeck (1838) but the name was allowed to lapse from general usage until revived more recently by Stermer (1940 : 138) who has figured the lectotype from Gaasgen, near Oslo. In southern Norway the species is characteristic of the Upper Chasmops Shale (4by) and the Upper Chasmops Limestone (4bd), a vertical range broadly comparable with that found in south Shropshire. It has been suggested by Stormer that Chasmops macroura, from Sweden, may be synonymous with C. extensa, but the problem is not yet resolved owing to lack of detailed information on the type specimen of C. macroura. The Swedish species was first described and figured by Angelin (1854 : 9, pl. 7, figs. 3, 4) but he attributed the name to Sjogren who has since been generally accepted as its author. Howell (1951 : 299) has pointed out that no evidence exists to show that Sjégren’s was anything more than a manuscript name and Angelin must accordingly be considered the author. The material from the Lower Cheney Longville Flags of Shropshire agrees well with Stermer’s illustrations of the lectotype of C. extensa, and the Shropshire specimens from this horizon are only slightly larger. Specimens from the Actonian Stage of the Onny Valley frequently attain much larger dimensions but this is not considered to be more than a local variation, perhaps the result of a more favourable environment. Warburg (1925 : 402) considered Chasmops maxima (Schmidt, 1881 : 112) to be identical with C. macroura, but according to Stermer (1945 : 421) the former may justifiably be regarded as a distinct species on account of the longer, deeper axial furrows of the pygidium. The specimens, including the pygidium, of C. maxima illustrated by Stormer (1945, pl. 2, figs. 1-5) show such a strong resemblance to C. extensa that the former may well be no more than a subspecies or variety of the latter, if not actually conspecific with it. Salter (1864: 41) recorded what he called “‘ Phacops (Chasmops) conophthalmus (sic) Boeck’ from the Caradoc Series of ‘‘ Acton Scott ’’. This species, the lecto- type of which has been refigured by Stormer (1940, pl. 3, figs. 1-6), is more correctly called Chasmops conicophthalma (Sars & Boeck), and is characteristic of the Lower Chasmops Shale (4ba) and the Lower Chasmops Limestone (4b/) in southern Norway. One of the most conspicuous differences between this species and C. extensa is the former’s possession of a short, rounded pygidium, contrasting with the elongated pygidium of the latter species. Specimens of Chasmops with somewhat similar, short pygidia have been collected from the Lower Longvillian Substage of the Cross Fell Inlier and North Wales, but as far as is known no genuine C. conicophthalma has yet been recovered from the Caradoc of Great Britain. Chasmops amphora was founded by Salter (1853: 12; 1864: 42, pl. 4, fig. 16) on a large pygidium from the Craig (= Crugg or Grug) Limestone of the Llandilo district of South Wales, a horizon stated by Williams (1953 : 195) to be Longvillian to Marsh- ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 337 brookian in age. The associated fauna includes Atractopyge, Illaenus, Leptaena, Nicolella and Onniella ?, an assemblage which suggests that the limestone, at least in part, might well prove to include strata as high as Actonian in age. At all events, C. amphora appears to be generally contemporaneous with C. extensa as found in Shropshire. An undoubted cephalon of C. amphorva has apparently never been figured, but cranidia of Chasmops recently obtained from the Craig Limestone are remarkably close to those of C. extensa and it is therefore possible that the two species may occur in association, though the pygidium of the latter species has yet to be found in South Wales. The pygidium of C. amphora is easily recognized by its tumid form which is more strongly arched, both longitudinally and transversely, than that of C. extensa, by its more rounded tip, and by the pleural furrows which, on the holo- type at least, appear to be pitted and which are more strongly defined both at and across the otherwise smooth border. Chasmops salopiensis sp. nov. (Pl. 52, figs. 2?, 5) Diacnosis. Glabella of distinctive shape, triangular in plan, narrowing back- wards markedly to small occipital ring. First glabellar lobes small. Frontal glabellar lobe large, long, about two-thirds median length of glabella. Description. The cephalon is roughly semicircular in outline, its median length less than half the maximum breadth. The glabella is moderately convex, both longitudinally and transversely, broader than long in the ratio of about 3:2, and roughly triangular in outline. It is bounded abaxially by almost straight axial furrows which diverge forwards at about 55 degrees; these become shallower medially, where they are constricted slightly opposite the first and second glabellar furrows. The frontal glabellar lobe is large, alate, subtriangular in outline, about twice as broad as long, and occupying more than half the median length of the glabella. The first glabellar furrows are deep, almost straight, and diverge anteriorly at 125 degrees; the first glabellar lobes, each roughly the shape of an isosceles triangle, occupy one-third of the length of the glabella, and are connected to the central lobe of the glabella by narrow “‘necks”’. The second glabellar furrows are short (év.), deep, and converge anteriorly at about 70 degrees; second glabellar lobes are sited at the distal ends of the second furrows and are represented by a pair of small tubercles, one at either end of a poorly-defined glabellar segment. The third glabellar furrows are barely visible small notches, and the third glabellar lobes are formed at the abaxial extremities of a small, basal glabellar segment which is slightly more conspicuous than that immediately in front. The occipital furrow is transversely straight, of moderate depth medially but becoming deeper abaxially where it ends in a pair of deep apodemal pits. The occipital ring is incompletely known, but is small, transversely convex, and apparently of small size. The posterior median lobe of the glabella is conspicuously narrow, equal to only one-sixth of the maximum breadth of the frontal glabellar lobe. The preglabellar field, though not conspicuous, is best developed medially, becoming narrower (exsag.) abaxially where it follows the outline of the frontal glabellar lobe until it is truncated by the axial 338 ORDOVICIAN TRILOBIDE FAUNAS (OF (SOUR TSH ROP SORE mln furrows. Frontally the preglabellar field is delimited by a shallow, narrow (exsag.) anterior border furrow which runs parallel to the preglabellar furrow until it inter- sects the axial furrows. The anterior border is narrow (sag.), reflexed so as to pro- duce a bluntly-rounded cross-section, and broadens (exsag.) distally where it passes without interruption into the lateral border. The latter is delimited by deep, broad lateral border furrows which run parallel to the lateral margin from the axial furrows. The posterior border furrows are narrow (exsag.) and moderately deep towards the axial furrows, but they widen abaxially and finally intersect the lateral border furrows. Near the axial furrows the posterior border at first narrows (exsag.) and then becomes parallel-sided, but at the posterolateral angles it increases in size abruptly and is produced rearwards to form a pair of large, tapering fixigenal spines. The upper surface of each spine is steeply declined outwards, its abaxial margin con- tinuous with that of the lateral borders, and the length of each spine is at least equal to that of the cranidium at the sagittal line. The proximal parts of the fixigenae are narrow and declined steeply from the palpebral lobes to the axial furrows; they are surmounted by conspicuous, reni- form, palpebral lobes set high above the level of the glabella and separated from the fixigenae by broad, well-defined, palpebral furrows. The schizochroal eyes are large, their visual surfaces steeply declined distally, and they extend from points in line with the adaxial ends of the first glabellar furrows until they are opposite the third glabellar lobes. The visual surface of the right eye of the holotype is partly preserved, though slightly displaced from its normal position. At least fifteen vertical rows of facets are visible as far as the mid-point of the palpebral lobe, so that about thirty rows must originally have been present. The number of facets in each row, as far as can be judged, varies from five to thirteen. The anterior branches of the facial suture run almost straight forwards from the eyes to cut the axial furrows, diverge forwards around the distal ends of the frontal glabella lobe, and finally converge to meet frontally, their course beyond the axial furrows almost coinciding with that of the anterior border furrow. The posterior branches run outwards from the eyes, arching first forwards and then rearwards until they cut the lateral border at an acute angle roughly in line with the third glabellar lobes. Immediately below each eye, and equal in length to the latter, is a well-defined, level, eye platform from which the main convex part of the librigena is steeply declined as far as the lateral border furrow. The surface of the test of the glabella and fixigenae is only partly known but most of it, at least, is ornamented with what appear to be large dispersed tubercles. As in the case of Chasmops extensa (see p. 333) each of these is seen, on magnification, to consist of a group of closely-packed, smaller granules, others of which are dispersed over the intervening spaces. The surface of the preglabellar field, anterior border and lateral border is granulate, but that of the various cephalic furrows and the genal spines is smooth. The hypostoma and thorax are not known. One fairly complete pygidium of Chasmops has been found in association with the holotype cephalon of C. salopiensis and may belong to the species. It is subparabolic in outline and, in spite of some crushing, is apparently strongly convex transversely. ORDOVICIAN LRIVOBILE PAUNAS OF SOUDH SHROPSHIRE, If 339 and about as long as broad. The axis, which carries sixteen axial rings, is bounded laterally by deep axial furrows along the whole of its length. In plan the axis tapers gently, ends in a bluntly rounded tip, and does not extend the whole length of the pygidium, being followed posteriorly by a small postaxial ridge. Each side-lobe carries sixteen ribs, the hindmost of which are only poorly defined. The ribs are separated by moderately deep pleural furrows which extend uniformly abaxially as far as the inner margin of the doublure, when they become suddenly shallower and fainter, at the same time curving back to the lateral margins. The pairs of ribs become progressively less divergent rearwards until the final, poorly-defined pair runs almost parallel to, and delimits, the postaxial ridge. The test of the pygidium is imperfectly known, but near the lateral margin appears to be finely granulate. HoRIZON AND LOCALITY. The specimens described were collected from a 3-inch band containing fragmentary trilobites and brachiopods in the north bank of the River Onny, about 30 yards east of its junction with the stream from Batch Gutter. The associated fauna includes Platylichas, Remopleurides, Tretaspis, Onniella and Sampo. The horizon is in the topmost part of the Actonian Stage and the species has not yet been found elsewhere. HoLotyPe. BM. In. 50551 (Pl. 52, fig. 5). Discussion. Chasmops salopiensis is probably most closely related to C. extensa but its cephalon may be distinguished by its longer and proportionately larger frontal glabellar lobe, the smaller first glabellar lobes, the narrower posterior portion of the glabella with its smaller occipital ring, and by the straighter axial furrows. The pygidium which is tentatively assigned to the new species appears to have a narrower axis than that of C. extensa, but the specimen is not well preserved and so has not been included as type-material. Family HOMALONOTIDAE Chapman, 1890 Genus PLATYCORYPHE Foerste, 1919 Some of the south Shropshire trilobites now assigned to Platycoryphe have been referred previously to Eohomalonotus Reed, 1918. The latter genus, together with Calymenella Bergeron, 1890, has been discussed by Prantl & Pribyl (1948) and Sdzuy (1957). The former authors regarded Calymenella as a subgenus of Eohoma- lontus, whilst the latter author believed the reverse to be true on the grounds that Calymenella was the earlier described genus. Sdzuy’s views have since been reiterated (tm Moore, 1959 : 0.484). The evidence for separating Calymenella and Eohomalonotus seems to be tenuous and Whittard (1960 : 162) regards them as synonymous, but both are typically genera differing markedly from the south Shropshire forms ; certain of the latter are, in turn, remarkably close to the type-species of the North American genus Platycoryphe. Recently this name has been placed by Henningsmoen (tn Moore, 1959 : 0.524) with the trilobites whose Order and Family are doubtful, but Whittard (1961 : 163) has used it to include certain Ordovician Homalonotidae from the Shelve Inlier, west Shropshire, and it is regarded here as being best grouped with 340 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II Calymenella and related genera. Platycoryphe is recorded from such widely-scattered regions as Bohemia, Shropshire and eastern North America, all the known occurrences being of Middle or Upper Ordovician age. Platycoryphe dentata sp. nov. (Pl. 53, figs. I-7, 9, 13; Pl. 55, figs. 3, 10) 1854. Homalonotus rudis M‘Coy : Salter & Aveline, p. 65. 1865. Homalonotus vudis M‘Coy : Salter, pl. 10, fig. 11. 1865a. Homalonotus vudis M‘Coy : Salter, pl. to, fig. 13. 1873. Homalonotus rudis M‘Coy : Salter, p. 53. 1918. Homalonotus biserrvatus (pars). Reed, p. 273. MS. name only. 1938. ‘“‘ Homalonotus’’ cf. bohemicus Barrande : Stubblefield im Pocock e# al., p. 255. 1939. Eohomalonotus? cf. bohemicus (Barrande) Stubblefield, p. 55. 1958. Eohomalonotus sp., Dean, p. 218. Diacnosis. Platycoryphe with three well-defined pairs of glabellar lobes. Distinct paraglabellar areas present. Tumid pygidium with six or seven axial rings and five pairs of pleural ribs. Each rib terminates laterally in two tooth-like projections, the anterior of which is the larger, separated from each other by the interpleural furrow. DeEscrIPTIon. The cranidium is moderately convex transversely, less convex longitudinally, with a maximum breadth generally about equal to its basal breadth, but two, probably dimorphous, forms of the species are present and in the second of these the glabella is conspicuously narrower (see PI. 53, figs. 3, 9). Such a feature may possibly be of sexual significance. The glabella is gently convex, subtrapezoidal in outline, with three pairs of well-defined glabellar lobes. The frontal glabellar lobe is subquadrate in plan with a slightly convex anterior margin and occupies between one-quarter and one-third of the median length of the glabella. The first glabellar furrows are shallow, gently curved, convex forwards, and extend adaxially and slightly rearwards from the axial furrows for about one-quarter of the glabellar breadth. The first glabellar lobes are subquadrate, equal to about one-sixth of the length of the glabella ; they are bounded to the rear by the second glabellar furrows which are parallel to, but deeper than, the first pair. The second glabellar lobes are of similar length to the first pair but project beyond them so that the out- line of the glabella is broken on either side by a “‘step”’ at the second glabellar furrows. The third glabellar furrows are deep and, near the axial furrows, almost parallel to the second glabellar furrows; they then curve adaxially and sharply backwards, extending almost half the length of the third glabellar lobes. The three pairs of glabellar furrows terminate longitudinally in line, parallel to the median line. The third glabellar lobes are about one and a half times as long as the first and second glabellar lobes ; they are almost of “‘ cat’s ear’ form and their abaxial margins are slightly indented. The axial furrows are broad and moderately shallow, becoming still shallower opposite the third glabellar lobes ; their rear halves are subparallel as far as the second furrows, in front of which, beyond the break in the glabellar outline described above, they become more strongly convergent forwards. The ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 341 frontal glabellar lobe and the anterior portions of the fixigenae are bounded frontally by a broad (sag.), moderately-deep, transverse furrow, in front of which rises the anterior border, becoming thicker medially where its apex is bluntly pointed or gently rounded. The occipital furrow is narrow (sag.), moderately deep and trans- versely straight except for its median third where it becomes shallower and slightly convex forwards. The occipital ring is of almost uniform length (sag.), equal to about one-sixth that of the glabella, and carries a weakly-developed median node. The pleuroccipital furrow is broad (exsag.) and well defined, follows a gently sigmoidal course abaxially from the axial furrows and finally flexes more strongly towards its outer ends. The posterior border is of uniform breadth (exsag.) and runs parallel to the posterior border furrow as far as the rounded genal angles. The anterior portions of the fixigenae are equal to or rather more than half the corrresponding glabellar breadth ; the posterior portions contain well-developed paraglabellar areas which are semicircular in outline and abaxially convex, the breadth of each being about equal to one-quarter of the basal glabellar breadth. Each paraglabellar area is bounded adaxially by the axial furrow and abaxially by a semicircular furrow which meets the axial furrow frontally in a broad, pit-like depression opposite the third glabellar furrow. The palpebral lobes, though imperfectly known, are apparently crescentic in plan, moderately convex abaxially, extending backwards from points level with the first glabellar lobes as far as the line of the third glabellar furrows ; their apices are levei with the front half of the second glabellar lobes. They are defined adaxially by shallow palpebral furrows which extend from just outside the axial furrows, opposite the mid-points of the first glabellar lobes, until level with the middle of the paraglabellar areas ; the furrows then turn through slightly more than a right angle and extend rearwards until they intersect the paraglabellar areas ; the furrows then turn through slightly more than a right angle and extend backwards until they intersect the paraglabellar furrows. The anterior branches of the facial suture converge forwards parallel to the axial furrows as far as the anterior margin, where they turn adaxially and meet frontally. The posterior branches extend straight abaxially from the eyes almost half-way to the lateral margins, but then curve back- wards through about 60 degrees and finally cut the lateral margins, apparently just in front of the rounded genal angles. The librigenae are not known with certainty, but a large specimen which probably belongs to the species is figured (Pl. 53, fig. 13). This is incompletely preserved but there is a suggestion of an eye-platform, outside of which the librigena is strongly convex and steeply declined towards the thickened border, from which it is separated by a broad, shallow, lateral border furrow. The hypostoma and thorax are unknown. The pygidium is tumid, wider than long in the ratio 5 : 3 or thereabouts, with rounded anterior and posterior margins which are strongly convex to front and rear respectively. Frontally the axis occupies about one-third of the maximum breadthof the pygidium, and extendsalmost to the tipof the latter. Theaxis is bounded laterally by deep, narrow, axial furrows which converge backwards strongly for about two thirds of their length and then become slightly less convergent as far as the tip; there they join with a shallow but well-defined postaxial groove which separates the axis from the tip of the pygidium. Nearly all the available specimens 342 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II have six axial rings, the sixth ring-furrow being shallow medially and obsolete distally. One small specimen (In. 54875) is exceptional in possessing seven axial rings, both the sixth and seventh ring furrows being obsolete distally. The terminal piece of the axis is swollen so as to produce a large, upwardly-directed, tubercle-like projection which is conspicuous even in a crushed specimen. In all the pygidia ex- amined, each side-lobe has five ribs which diminish slightly in size from first to fifth and are separated from each other by deep, narrow (exsag.), pleural furrows. Each rib carries a faintly-impressed interpleural furrow which divides the rib into an anterior and a posterior band, the former being about twice as wide (exsag.) as the latter. The interpleural furrows are, for the most part, barely visible but each becomes suddenly and markedly deeper near the lateral margin which it intersects. The resulting notches, situated asymmetrically between adjacent pleural furrows, have the effect of producing a series of ten, tooth-like projections, alternately large and small in size, along either margin of the pygidium. There is sometimes a faint suggestion of a further pair of projections to the rear of the fifth pleural furrow. With the exception of the furrows, the whole of the surface of the dorsal exo- skeleton is covered by small, closely-grouped granules. These are particularly con- spicuous on the outer surface of the test but less so on internal moulds. HORIZONS AND LOCALITIES. The strata from which practically all the known spe- cimens of Platycoryphe dentata have been collected form the Harknessella subquadrata Zone (known also as Beds or Limestone), the topmost subdivision of the Costonian Stage in the northern Caradoc Area. Only an occasional isolated specimen has been found in the succeeding Harnagian Stage of Coundmoor Brook (PI. 55, fig. 3) ; this resembles the Costonian material in practically all respects but shows indications of eight axial rings, the seventh ring-furrow being obsolete distally and the eighth furrow still less well defined. The tip of this pygidium appears more prominent than usual owing to crushing which has emphasized the region immediately behind the axis. No specimens have been found in the Costonian of the southern Caradoc Area and only one individual, described later (p. 344) as P. cf. dentata, is known from rocks of, presumably, Harnagian age in the Horderley District. Costonian localities include: the disused quarry in Black Dick’s Coppice, Even- wood ; the section exposed in Bullhill Gutter, Evenwood ; disused quarries in the north-western and north-eastern parts of Round Nursery, south of Harnage Grange ; and the old quarry by the south-eastern side of Coundmoor Brook, 1,500 yards south-west of Harnage Farm, Harnage. The Harnagian locality, comprising shales of the Reuscholithus reuschi Zone, is in the south bank of Coundmoor Brook, 1,340 yards south-west of Harnage Farm. HoLotyPe. GSM. WM. 1335 (PI. 53, fig. 1). PARATYPES. GSM. WM. 1371 (PI. 53, fig. 3) ; BM. In. 54875 (Pl. 53, figs. 2, 4) ; BM. In. 54876 (Pl. 53, fig. 5); GSM. WM. 1338 (PI. 53, fig. 6). OTHER FIGURED SPECIMENS. BM. In. 54584 (PI. 53, fig. 7) ; BM. In 49310 (PI. 53, fig. 9); BM. In. 54879 (Pl. 53, fig. 13); BM. In. 54913 (Pl. 53, fig. 3); BMoin. 54880 (Pl. 54, fig. 5); BM. In. 54325 (Pl. 55, fig. 3); GSM. 35516 (Pl. 55, fig. ro). Discussion. The manuscript name Homalonotus biserratus was applied by Reed (1918 : 273) to specimens, from the Hoar Edge Grits (of Costonian age) near Harnage, ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 343 which had earlier been thought by Salter (1865a; pl. to, fig. 13) to belong to Homalonotus rudis. The specific name used by Reed referred to the saw-like margins of the pygidium and he almost certainly had in mind the new species. He appears, however, to have used the name to include all the homalonotids of the Hoar Edge Grits and another specimen, a cranidium Sedgwick Mus. A. 43966 (PI. 55, fig. 2), labelled by him as belonging to the same manuscript species is now known to be a typical Brongniartella and elsewhere in this paper is made one of the paratypes of B. cavadociana sp. nov. Platycoryphe dentata is close to the Czech species P. bohemica (Barrande, 1852 : 580, pl. 34, figs. 40-42; 1872: 37, pl. I, figs. 6, 7) from the Drabov Quartzites, of Llandeilo age. This species was re-examined by Prantl & Pribyl (1948 : 7) and placed in Eohomalonotus. The figures of the Bohemian form suggest some variation in the proportions of the glabella, but none shows any trace of paraglabellar areas such as are strongly developed in the Shropshire species. The pygidium of P. bohemica bears a general resemblance to that of P. dentata and possesses a similar number of both axial rings (six or seven) and pleural ribs (five or six), the sixth and seventh axial rings becoming obsolete distally. The terminal piece of the pygidial axis in most figured examples of P. bohemica lacks the conspicuously tumid form found in the Shropshire species, though one specimen (Barrande, 1852, pl. 34, fig. 41) shows a weak development of this feature. The lateral margins of the pygidium are smooth, contrasting strongly with the serrated margins of P. dentata. The type-species of Platycoryphe was first described by Foerste (1910 : 81-82, pl. 2, fig. 7) as Calymene platycephala and was founded by him on a single cranidium. In many respects the latter resembles that of P. dentata but differs in having an “ un- stepped ” glabellar outline, narrower fixigenae frontally, and a smaller, narrower (¢.), anterior border. There is a possible trace of a paraglabellar area on the right fixigena but the preservation of both fixigenae is too poor to be certain. A fragmentary pygidium tentatively assigned to P. platycephala (Foerste, 1910, pl. 4, fig. 21) has a narrower axis than that of P. dentata, with at least six axial rings. The pleural lobes have a larger number of ribs, six pairs, with a suggestion of a seventh pair ; there is no trace of serrations along the lateral margins. The specimens of P. platy- cephala originated from the Saltillo Limestone of Tennessee, a horizon which Foerste (1910 : 19) described as the stratal equivalent of the Hermitage Limestone, of middle Trenton age. The North American species is thus somewhat later in age than the south Shropshire specimens. Platycoryphe dubia (Savage, 1913 : 60, pl. 2, figs. 8, 9; Foerste, 1919 : 395, pl. 19, figs. 15a, b) from the Girardeau Limestone (Richmondian) of Illinois differs markedly from all the above species. The axial furrows are uniformly convergent forwards, the basal glabellar lobes are smaller, bounded by straight third glabellar furrows, and paraglabellar areas are not developed. The front of the pygidium is only slightly convex forwards and the pleural furrows do not reach the margins. The species may only tentatively be retained in Platycoryphe. Recently Whittard (1961: 164, pl. 22, figs. 8-19) has redescribed and figured Platycoryphe vulcani (Murchison) from the Weston Beds (high Llanvirn Series) of the Shelve Inlier, west Shropshire. This species is conspicuously different from GEOL. 5. 8. 31 344 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II P. dentata, possessing a more trapezoidal glabellar outline with less well-developed glabellar furrows, and lacking paraglabellar areas and serrated pygidial margins. Some of the specimens figured by Whittard are difficult to distinguish from Brongniartella, and a reappraisal of the distribution of Platycoryphe in Britain must await the redescription of the type-species. Two specimens are figured in this paper as Platycoryphe cf. dentata ; although they probably belong to that species, some doubt persists owing to the state of preservation. Specimen BM. In. 51726 (Pl. 53, fig. 8) is a flattened, internal mould of an incom- plete cranidium, the anterior border and palpebral lobes of which are missing. The glabellar outline is elongated and subtrapezoidal, and the axial furrows are almost straight. The second glabellar lobes are longitudinally in line with the first and third pairs of lobes and lack the distally projecting outline of those of P. dentata, but such modification might be the result of crushing. Despite the absence of the palpebral lobes the position of the eyes seems to be as in P. dentata, and the three pairs of glabellar furrows are of similar type, though somewhat distorted. BM. 51726 is from the Neptunian dyke in the north-eastern corner of the disused quarry on the south side of Hazler Hill, east of Church Stretton. The country rock forms part of the Eastern Uriconian (Pre-Cambrian) and the fissure infilling contains a fauna of early Harnagian age, equivalent to the zone of Reuscholithus veusch, though the zonal trilobite has not been found there. Cranidium GSM. 35524 (PI. 53, fig. 10) was described by Salter (1865@ : 108) as Homalonotus edgell1, but in the same work (pl. Io, fig. 10) he figured it as Homalonotus bisulcatus. Later, Reed (1918 : 272) considered the specimen to belong to the same group as Homalonotus bisulcatus and H. ascriptus Reed. In the shape of the narrow glabella and the medially thickened anterior border the specimen, a small internal mould, matches the narrow form of Platycoryphe dentata, but the glabellar furrows found in that species are completely absent. This may be due to the state of pre- servation and it seems unlikely that the specimen is specifically distinct. It is labelled merely as having been obtained from the “‘ ? Caradoc ”’ of “‘ Horderley’’, but the state of preservation is characteristic of the Smeathen Wood Beds which form an outcrop at and near the River Onny Valley, south-east of Horderley. These strata belong to both the Reuscholithus reuschi and Salterolithus caractaci Zones of the Harnagian Stage, but there is no indication as to which of these horizons provided the specimens. Platycoryphe ? sp. (Pl. 53, fig. 11 ; Pl. 55, fig. 14) 1958. Eohomalonotus sp. (? nov.), Dean, pp. 206, 221. DESCRIPTION. Two isolated specimens which may belong to Platycoryphe have been collected from the Lower Longvillian rocks of the Onny Valley, strata much higher in the succession than any others in Shropshire known to contain the genus. One, an incomplete cranidium (Pl. 53, fig. 11), is poorly preserved as an internal mould but can be seen to possess a glabella superficially similar to that of Platycoryphe ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 345 dentata, though the frontal glabellar lobe is broader and more rectangular in outline. There are three pairs of straight glabellar furrows, directed adaxially and moderately rearwards from the axial furrows and becoming progressively deeper from first to third. They resemble those of Platycoryphe dubia (Savage) as figured by Foerste (x919, pl. 19, fig. 15a) and, like that species, the Longvillian form has smaller basal glabellar lobes than P. dentata and comparable forms. The anterior border is longer (sag.) and transversely straighter than that of P. dentata but the preservation is too poor to warrant a detailed comparison. The other specimen, a small, convex pygidium (PI. 55, fig. 14) is also somewhat indifferently preserved as an internal mould. The frontal margin is strongly arched forwards as in Platycoryphe dentata, and the axial furrows also become less convergent opposite the posterior third of the axis. There appear to be seven axial rings, the rearmost two of which are less well defined than the rest. The terminal piece is poorly preserved and a comparison with that of P. dentata is not possible. The pleural lobes possess at least five pairs of ribs, a number comparable with that found in P. dentata, but the pleural furrows are straighter and directed backwards even more strongly than in that species. HoRIZONS AND Loca.ities. Cranidium BM. In. 50511 (PI. 53, fig. 11) is from the middle third, the zone of Dalmanella indica and D. lepta, of the Lower Longvillian Substage exposed just west of the south-western corner of Rookery Wood, three- quarters of a mile south-east of Horderley. Pygidium BM. In. 50816 is from the lowest third, the zone of Dalmanella horder- leyensis, of the Lower Longvillian. The specimen was collected from the western end of the quarry by the north side of the Bishops Castle road, 250 yards south-east of Glenburrell Farm, Horderley. Genus BRONGNIARTELLA Reed, 1918 TYPE SPECIES. Homalonotus bisulcatus M‘Coy, 1851 ex Salter MS. by original designation of Reed (1918 : 322). The authorshop of Brongniartella bisulcata, like that of Kloucekia (Phacopidina) apiculata (see p. 326) has usually been attributed to Salter though the species was, in fact, described first by M’Coy in 1851, Salter’s description not being published until the following year. In M‘Coy’s original work (7m Sedgwick & M‘Coy, 1851, pl. IG) figs. 24-29 were said to be Homalonotus bisulcatus, whilst figs. 30 and 31 were varieties of the species, but according to Salter (1852) figs. 29-31 were “ variety £ minor ’’, the specimens deriving from Bala and Westmorland. Of the specimens figured by M‘Coy, the original of his pl. 1G, fig. 24, was mis- takenly stated, in the explanation of the plate, to have come from Pwllheli, though his text (p. 168) gave the locality correctly as “‘ Wittingslow, near Acton Scott, Shropshire’’. In the Appendix to M‘Coy’s work Salter stated that the originals of figs. 24, 26 and 27 came from Wittingslow (now Whittingslow, a farm three-quarters of a mile south-west of Marshbrook) and those of figs. 25 and 28 from Tremadoc (2 Pwllheli). In attempting to clarify the species Reed (1918 : 270) stated erroneously that the originals of figs. 24, 26 and 27 were housed in the British Museum (Natural 346 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II History). In fact they are in the Geological Survey and Museum, though the counter- part of fig. 24 has now been traced in an old collection at the British Museum, where it is numbered In. 54914. Reed also stated, in my opinion correctly, that the species is undoubtedly established on Shropshire specimens and the original of M‘Coy’s fig. 24 is now chosen as lectotype. The present description is founded on the Shrop- shire syntypes, supplemented by further material from the type-area. Brongniartella bisulcata (M‘Coy) Pl. 54, figs. 2, 2, 45°38, 9: PlNs5e figsien, 45/7) 1847. Homalonotus bisulcatus MS., Sedgwick, p. 149. Nomen nudum. 1851. Homalonotus bisulcatus (Salter in Appendix), M‘Coy im Sedgwick & M‘Coy, p. 168. 1852. Homalonotus bisulcatus M‘Coy : Salter, p. v. 1854. Homalonotus bisulcatus M‘Coy : Salter 1m Morris, p. 109. 1859. Homalonotus bisulcatus M‘Coy : Murchison, p. 74, fig. 2. 1865. Homalonotus bisulcatus M'Coy : Salter, pl. xvi, figs. 1-8. 1865a. Homalonotus bisulcatus M‘Coy : Salter, p. 105, pl. to, figs. 3, 4, 5, 9. 1867. Homalonotus bisulcatus M‘Coy : Murchison, p. 69, fig. 2. 1884. Homalonotus bisulcatus M‘Coy : La Touche, p. 57, pl. 3, fig. 62. 1918. Homalonotus (Brongniartella) bisulcata M‘Coy: Reed, p. 322. 1930. Homalonotus bisulcatus M‘Coy : Whitcomb, p. 348. 1947. Brongniartella bisulcata (M‘Coy) : Harper, p. 166. 1948. Brongniartella bisulcata (M‘Coy) : Prantl & Pribyl, p. 15. 1949. Brongniartella bisulcata (M‘Coy) : Bancroft, p. 309. 1958. Bvrongniartella bisulcata (M‘Coy) : Dean, pp. 207-210, 222-223. 1959a. Brongniartella bisulcata (M‘Coy) : Dean, pp. 207, 220. DEscRIPTION. The cephalon is sub-semicircular in outline, blunted frontally. The glabella is trapezoidal in plan, gently convex both longitudinally and transversely, and stands slightly above the level of the remainder of the cephalon ; it is slightly longer than broad, narrowing forwards slightly so that the breadth of the frontal lobe is about three-quarters of the basal breadth. The axial furrows are shallow with poorly-defined distal margins ; for about half the length of the glabella they con- verge gently forwards from the occipital furrow, then suddenly converge more sharply, constricting the glabellar outline, and finally continue forwards parallel to their former course but set slightly closer together. Hypostomal pits are apparently absent. There are no glabellar furrows and the glabella is generally quite smooth though an occasional specimen, for example the lectotype, shows a suspicion of a low, longitudinal, median ridge. The anterior margin of the frontal glabellar lobe is weakly convex forwards, separated by an ill-defined furrow from the anterior border: the two last coalesce to form a broad (sag.), smooth, frontal brim which passes laterally without division into the anterior portions of the fixigenae. The occipital furrow is shallow but well defined, transversely straight for most of its length (év.) but arched forwards medially, though this feature is less conspicuous in immature cranidia. The lectotype has part of the test preserved, showing that the occipital furrow is much less well defined on the external than on the internal surface. The occipital ring is of moderate, uniform length (sag.), generally about one-eighth that of the glabella ; in plan it resembles a flattened trapezoid owing to the fact that the axial furrows diverge rearwards markedly on meeting the occipital ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 347 furrow. The pleuroccipital furrow is set slightly to the rear of the occipital furrow ; although shallow near the axial furrows it deepens abaxially at first but then becomes shallower again, at the same time curving gently forwards, and finally dies out with- out reaching the lateral margins. The posterior border, where it meets the axial furrows, is of the same length (exsag.) as the occipital ring, but becomes longer distally and terminates in bluntly-rounded genal angles. The fixigenae are large, each with a frontal breadth about, or rather less than, half that of the frontal glabellar lobe. Between the eyes and the sides of the glabella the fixigenae are level, but to the rear of the eyes they are turned down strongly, through about 70 degrees, towards the genal angles. The eyes are small, situated opposite the mid-point of the glabella, with small, smooth, palpebral lobes which are gently declined adaxially and pass without interruption into the fixigenae, palpebral furrows being absent. From the eyes the anterior branches of the facial suture extend forwards, roughly parallel, until opposite the front of the glabella, but then become more convergent, sweeping forwards and adaxially until they meet frontally, the suture so formed being above and just inside the naterior margin. The posterior branches of the facial suture run straight backwards from the eyes for only a short distance, then arch abaxially for about half their length, and finally curve back near the lateral margins which they cut acutely at, or immediately in front of, the genal angles. Only the more robust, triangular, hindmost portions of the librigenae are usually found. Each carries a broad, smooth, level eye-platform immediately beneath the visual surface of the eye, which has not been found preserved. Below each eye- platform the librigena is convex, steeply declined abaxially, and separated from the narrow, thickened lateral border by a shallow, broad, lateral border furrow which dies out towards the genal angle and so does not meet the pleuroccipital furrow. The hypostoma has not previously been described or figured. The overall length and breadth are about equal, and the transverse convexity is greater than that longitudinally. The median body is moderately convex, slightly elongated, divided by a broad (sag.), shallow, median furrow into anterior and posterior lobes, the latter about one-quarter the length (sag.) of the former. The anterior border is narrow (sag.) medially, widening distally to form large anterior wings, the anterolateral margins of which are broad and obliquely truncated. The border is delimited by a broad (sag.), shallow furrow which curves backwards distally, meeting the median furrow, here slightly deeper, just behind the anterior wings. The lateral margins are deeply indented by large lateral notches which produce their maximum con- striction of the outline almost opposite the mid-point of the hypostoma. Behind the lateral notches the lateral margins expand distally into rounded shoulders which form flanges directed ventrally at a moderate angle. The posterior margin of the hypostoma is forked, terminating in a pair of broad points separated by a deep, curved notch. Each point carries a small indentation, poorly preserved but perhaps representing a macula. The posterior border is separated from the posterior lobe of the median body by a broad (exsag.), shallow furrow which runs parallel to the median furrow, finally becoming deeper and terminating level with, and a short distance inside, the shoulders. No complete thorax is yet known from south Shropshire, but one containing ten 348 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, Il thoracic segments is figured here (Pl. 55, figs. 1, 4). The thoracic axis is broad, apparently more than twice the breadth of the side-lobes, though only slightly more than one-third of the total breadth in flattened specimens. Each axial ring is gently convex, its breadth (tv.) about eight to nine times the uniform length (sag.), and is separated from the articulating half-ring by a broad (sag.), shallow, articulating furrow. The axial furrow marking the distal end of each axial ring is oblique and runs abaxially from front to rear, its intersection with the posterior margin of the segment being marked by an articulating socket into which fits a corresponding articulating process on the subsequent segment. Outside the axial furrows the pleurae are of uniform breadth (exsag.), curving first slightly rearwards and then more strongly forwards to the large, bluntly-rounded tips. Each pleura carries a conspicuous pleural furrow which commences at the axial furrow as an abaxial con- tinuation of the articulating furrow, and then arches gently back abaxially for almost half the length (t.) of the pleura, when it becomes suddenly narrower (exsag.) and slot-like; it then curves gently forwards towards the anterior margin and ends sharply without reaching either the margin or the pleural tip. The pygidium is parabolic in outline, broader than long in the ratio of about 4: 3, its anterior margin arched gently forwards. It is plump, convex both transversely and longitudinally, and the upper surface is moderately declined from front to rear. The axis is subtriangular in outline, bounded by smooth, shallow axial furrows ; frontally it occupies slightly less than half the maximum breadth, a proportion falling to a little less than one-third in flattened specimens. The number of axial rings ranges from nine to twelve, the most common number present being ten or eleven. The articulating furrow is deep but the remaining ring-furrows are less well defined and become fainter backwards ; they are not aligned with the rib-furrows of the side-lobes and do not quite attain the axial furrows. There are generally seven or eight ribs present on each side-lobe. The rib-furrows of the first pair are deep and narrow (exsag.) and those of the second pair are only slightly less strongly developed, but the remainder are markedly shallower. All these furrows become much atten- uated at a narrow, poorly-defined border which they Cross to reach the lateral margin. The border is best developed posterolaterally and at the tip of the pygidium where, in some specimens, it is turned up slightly to form a flange-like rim, though this feature may have been exaggerated by crushing. HORIZONS AND LOCALITIES. Brongniartella bisulcata is one of the most abundant and characteristic of Shropshire trilobites and may be found at numerous localities along the Ordovician outcrop. It makes its appearance in the Alternata Limestone, the basal subdivision of the Upper Longvillian Substage, and it was from this horizon that the lectotype was originally collected. The species continues through the overlying Lower Cheney Longville Flags, the upper subdivision of the Upper Longvillian Substage, but it is in the succeeding Upper Cheney Longville Flags, representing the Marshbrookian Stage, that the greatest abundance is attained. No undoubted records from later strata are known, but rare fragmentary traces from the arenaceous development of the Actonian Stage suggests that the species may eventually prove to have persisted in south Shropshire after the close of the Marsh- brookian in places where the environment was favourable. ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 349 Lectotype. GSM. 7626 (Pl. 54, fig. 1). The counterpart of this specimen is in the British Museum (Nat. Hist.), where it is numbered In. 54914. FIGURED SPECIMENS. BM. In. 48952 (PI. 54, figs. 2, 4); BM. In. 50791 (Pl. 54, fig. 8); BM. In. 49557 (Pl. 54, fig. 9) ; GSM. 35519 (Pl. 55, figs. 1, 4); BM. In. 48598 (Pl. 55, fig. 7). Discussion. For convenience all the species of Brongniartella described in this paper will be discussed together at the end of the section. Brongniartella caradociana sp. nov. (Pl. 53, fig. 12; Pl. 54, figs. 3,5, 7, 10; Pl. 55, fig. 2) 1918. Homalonotus biserratus Reed (pars), p. 273. MS. name only. 1958. Brongniartella aff. bisulcata (M‘Coy) : Dean, p. 218. Diacnosis. Large species of Brongniartella distinguished by conspicuously broad glabella, narrowing frontally, large, scoop-like anterior border and broad fixigenae. Paraglabellar areas developed on immature individuals. DESCRIPTION. The cephalon has not yet been found completely preserved. The cranidium is depressed, with median length about three-fifths of the maximum breadth. The glabella is gently convex, both longitudinally and transversely, its length equal to the basal breadth. The outline of the glabella in the adult trilobite is slightly more irregular than is usual for the genus. Shallow axial furrows, contain- ing poorly-defined hypostomal pits frontally, run almost straight forwards for about half the length of the glabella, converging only very slightly; they then become strongly convergent for a short distance, about one-seventh of the median length of the glabella, finally becoming less convergent as far as the anterolateral angles of the frontal glabellar lobe, the breadth of which is two-thirds that across the base of the glabella. In general there are no glabellar furrows present, but one specimen (Pl. 53, fig. 12) shows traces of shallow depressions which do not cut the axial furrows, but which meet mesially to form what appears as a single, broad (sag.), transglabellar depression or furrow. It is not clear whether this represents a pair of third glabellar furrows or whether it is the result of crushing. The frontal lobe of the glabella is slightly convex forwards, standing only slightly higher than the anterior border ; the latter is broad (sag.) and gently inclined anteriorly so that a slightly concave, scoop-like brim results, the frontal margin of which is arched forwards mesially. The distal portions of the occipital furrow are deep and transversely straight but the median third is markedly shallower and, in the largest-known specimen, convex forwards, though straight in small individuals. The occipital ring is of almost uniform length (exsag.) which increases slightly mesially. In small individuals the lines of the axial furrows diverge rearwards from the frontal edge of the occipital furrow, cutting the outer ends of the occipital ring obliquely. The occipital ring of the holotype, a large and presumably mature specimen, passes abaxially into the pos- terior border, only faint traces of axial furrows being present. The posterior border furrow is transversely straight, moderately deep, and is set slightly to the rear of the occipital furrow. The posterior border is not yet completely known, but the 350 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II inner parts, at least, are transversely straight and of uniform breadth (exsag.). The genal angles have not yet been found intact. The fixigenae are large and conspicu- ously broad ; frontally their individual breadth equals more than half that of the frontal glabellar lobe in the holotype, a proportion rising to more than two-thirds in smaller specimens. The hindmost portions of the fixigenae of immature specimens carry conspicuous, slightly-raised, subcircular, paraglabellar areas, each defined posteriorly and adaxially by the posterior border furrow and axial furrow respectively, and frontally and abaxially by a continuous, shallow, arched furrow. In smaller individuals the paraglabellar areas indent the axial furrows which then appear slightly concave abaxially, but in the largest specimen they are scarcely visible. The eyes are small, situated forwards of centre of the glabella. Small palpebral lobes are continuous with the upper surface of the fixigenae and gently declined adaxially ; no palpebral furrows have been seen. The anterior branches of the facial suture arch forwards and adaxially from the eyes, meeting the anterior border in line with points mid-way between the palpebral lobes and the frontal glabellar lobe, and then converge frontally. The posterior branches curve abaxially backwards from the eyes to the as yet unknown genal angles. The librigenae and hypostoma are unknown. Only one specimen of the thorax has been discovered, a poorly-preserved indi- vidual (Pl. 54, fig. 3) with twelve, or possibly thirteen, segments. The thoracic axis is broad, occupying about half the total breadth. Each axial ring is of uniform length (sag.), bounded distally by shallow axial furrows which converge forwards markedly. Each pleura is of fairly uniform breadth (exsag.) and directed slightly rearwards abaxially ; it carries a pronounced pleural furrow which commences at the line of the axial furrow, almost cutting the frontal margin, and then runs distally and gently backwards finally pursuing an almost median course towards the tip. It is not clear from the available material whether the furrow actually reaches the tip, nor is the shape of the latter known. A single pygidium, small and possibly immature, which may probably be referred to the new species has been collected. Though incompletely preserved, it is broader than long, approximately in the ratio 9: 7, and semielliptical in outline. The axis occupies more than one-third of the frontal breadth, is moderately tapered towards the bluntly-rounded tip, and occupies about nine-tenths of the overall length, excluding the articulating half-ring. Seven axial rings are visible, becoming less well defined posteriorly, but they extend for only three-quarters of the length of the axis. The pleural lobes carry five pairs of ribs, defined by pleural furrows which appear to extend to the lateral margins. The first axial ring furrow is deep and cuts the straight, moderately-deep, axial furrows in line with the equally deep first pair of pleural furrows. The remaining axial ring furrows do not quite intersect the axial furrows and are not in line with the corresponding pleural furrows. The second pleural furrows are slightly shallower than the first, and the remaining pleural furrows become progressively shallower. The surface of the test in all the materal available is covered with fine pitting visible only under the microscope. HoRIZON AND LOCALITIES, All the known specimens of Brongniartella caradociana ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 351 are from the limestones and calcareous sandstones forming the upper portion of the Costonian Stage, the zone of Harknessella subquadrata, in the district around Even- wood and Harnage Grange, near the northern end of the outcrop of the Caradoc Series in the type-area. Important localities are: the old quarry by the south- eastern side of the brook, 1,500 yards south-west of Harnage Farm ; the section in Bullhill Gutter, Evenwood ; various old quarries in Round Nursery, south-south- east of Harnage Grange; and Black Dick’s Quarry, Evenwood. Hototyre. BM. In. 51485 (Pl. 53, fig. 12). PARATYPES. BM. In. 50769 (PI. 54, fig. 10) ; BM. In. 54871 (PI. 54, fig. 7); BM. In 54880 (Pl. 54, fig. 5) ; BM. In. 54913 (PI. 54, fig. 3). Sedgwick Museum A. 43866 (Pl. 55, fig. 2). Brongniartella minor (Salter) (Pl. 54, fig. 6; Pl. 55, fig. 11) 1851. Homalonotus bisulcatus (Salter in Appendix), M‘Coy in Sedgwick & M‘Coy, p. 168 (pars), pl. 1G, fig. 30 (? also fig. 29). 1852. Homalonotus bisulcatus var. 8 minor, Salter, p. v. 1873. Homalonotus bisulcatus M‘Coy : Salter, p. 53. 1891. Homalonotus bisulcatus M’Coy : Woods, p. 145. 1918. Homalonotus bisulcatus M‘Coy var. B minor Salter, Reed, p. 270. 1947. Brongniartella parva Harper, p. 165, pl. 6, figs. 7, 8. Several specimens were described and figured by M‘Coy (em Sedgwick & M‘Coy, 1851 : 168, pl. 1G, figs. 24-31) as Homalonotus bisulcatus. Of them the original of fig. 29 was unlocalized, whilst the originals of figs. 30 and 31 were stated in the text to be “‘ supposed varieties ’’ from ‘‘ Ravenstone Dale ’’, although in the explanation of the plate the latter locality was modified to ‘‘ Ravenstone Dale, Maes Meillion ”’. In his supplement to the foregoing work Salter (1852:v) applied the name Homalonotus bisulcatus var. 6 minor to the originals of figs. 29-31, both 29 and 30 being stated to derive from Bala and 31 from Westmorland. Mr. A. G. Brighton kindly informs me that the original of fig. 29 cannot now be traced in the Sedgwick Museum, and has made available the other syntypes, numbered A. 10574 (fig. 30) and A. 10576 (fig. 31). A. 10574, which may be referred to the genus Brongniartella, is a pygidium said to be from the “‘ Bala Limestone’”’ of Maes Meillion, near Bala, Merioneth. It is associated, on the same hand-specimen, with Kloucekia apiculata (M‘Coy), Flexi- calymene cf. cavactaci (Salter), Sampo sp. (? oepiki Whittington) and ? Paucicrura sowerbi1 Cave & Dean ; the specimen is therefore probably of Lower Longvillian age. Owing to the state of preservation of the surface of the type pygidium it is not easy to determine the exact number of axial rings and pleural ribs present, but there appear to be eight and seven respectively. These numbers are the same as those described by Harper (1947 : 166) for Brongniartella parva, a species which possesses a similar pygidial outline and was described from the Lower Longvillian strata of Ynys Galed, Caernarvonshire. Pygidium A. 10574 is selected here as the lectotype of Brongniartella minor, of which species Brongniartella parva Harper is now regarded as a subjective synonym. 352 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II The other syntype of Homalonotus bisulcatus var. 8 minor is a pygidium numbered A. 10576. The locality, given as “‘ Ravenstone Dale’”’, together with the state of preservation, suggests that it was obtained from the Upper Ordovician strata of the Cautley district in north-west Yorkshire, between Sedbergh, Yorkshire, and Raven- stonedale, Westmorland. Homalonotus sedgwicki was described from the same neighbourhood by Salter (1865@ : 107) and founded by him on cranidia only; the species is in need of revision but may prove to be conspecific with the specimen figured by M‘Coy. Should this be the case, the name Homalonotus sedgwicki would be available for both specimens. In south Shropshire a few cranidia have been collected which agree in size and proportions with the holotype of Brongniartella parva Harper. All the specimens are preserved as internal moulds found in weathered lenses of sandy limestone within the Horderley Sandstone ; they are apparently undistorted and show no trace of the median ridge found in B. minor subcarinata subsp. nov. (see later), a form which possibly also occurs, though rarely, in the Horderley Sandstone. HORIZONS AND LOCALITIES. The earliest-known specimen is from the Lower Hor- derley Sandstone, Upper Soudleyan Stage, zone of Broeggerolithus soudleyensis, at the small, disused, road-side quarry 100 yards south-east of Glenburrell Farm, Horderley. The youngest specimen was obtained from the topmost beds of the Horderley Sandstone, the Bancroftina typa Zone of the Lower Longvillian Substage, at the disused quarry by the west side of New House, Horderley. Harper’s type material of Brongniartella parva was collected from North Welsh strata containing Dalmanella indica Whittington together with other typical Lower Longvillian species. FIGURED SPECIMENS. BM. In. 52315 (PI. 54, fig. 6) ; BM. In. 51529 (PI. 55, fig. 11). Brongniartella minor subcarinata subsp. nov. (Pl. 55, figs. 6, 8, ? 13) 1958. Bvrongniartella sp., Dean, pp. 203, 220. 1961. Bvrongniartella aff. minor (Salter) : Dean & Dineley, p. 374, pl. 20, figs. 3, 7. Diacnosis. Subspecies of Brongniartella minor characterized by presence of median ridge on glabella, and constriction of glabella outline frontally. Pygidium with seven axial rings and five pleural ribs. DESCRIPTION. Several specimens have been found, particularly in the Lower Soudleyan strata of south Shropshire, of a Brongniartella showing an obvious re- semblance to B. minor (Salter) but differing consistently from that species in a few respects. There is generally a slight constriction of the glabellar outline at about its mid-point, a feature seen also in Brongniartella bisulcata and B. caradociana (see earlier) ; its apparent absence from the North Welsh specimens may, however, be due to their being mechanically deformed and is probably not of major significance. The obvious difference in the glabella of the Shropshire subspecies is the presence of a pronounced longitudinal ridge coincident with the sagittal line over the median half of the glabellar length. Such a structure has not been observed in Brongniartella minor, perhaps owing to deformation ; on the other hand, Shropshire cranidia figured ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 353 in this paper as Brongniartella minoy are apparently undistorted but show no sign of a median carina. The pygidium of the new subspecies is of generally similar outline to that of B. munor, but the number of axial rings and pleural ribs is consistently smaller, being respectively seven and five, compared with eight and seven. HORIZON AND LOCALITIES. The new subspecies has been found in moderate abun- dance in the higher Glenburrell Beds, representing the Lower Soudleyan Stage, zone of Broeggerolithus broeggert. Localities include: the stream-section at the extreme north-eastern corner of Smeathen Wood, Horderley ; the exposure behind the farm- house, Glenburrell Farm, Horderley ; and the cartway 50 yards north-east of Glen- burrell Farm. A further specimen referred tentatively to the same subspecies (Pl. 55, fig. 13) was collected from the Horderley Sandstone, Lower Longvillian Substage, zone of Dalmanella indica and D. lepta, at the old quarry by the west side of the path through Longville Plantation, 720 yards north-west of the ““Earthwork”’ at Cheney Longville. No stratigraphically intermediate specimens have yet been found, and this suggests that the subspecies may have been subject to ecological control, showing a pre- ference for the mudstone environment of the Glenburrell Beds rather than the, presumably, shallower conditions which gave rise to the Horderley Sandstone. The same subspecies has been collected from Habberley Brook, north-west of The Long- mynd, where the higher beds of the so-called Pontesford Shales have a lithology and fauna similar to that of the Lower Soudleyan of the Onny Valley. Ho.otyre. BM. In. 50781 (PI. 55, fig. 6). PARATYPE. BM. In. 51182 (PI. 55, fig. 8). OTHER SPECIMEN?. BM. In. 51726 (PI. 55, fig. 13). Brongniartella edgelli (Salter) (PI. 55, figs. 9, 12) 1865a. Homalonotus edgelli Salter, p. 108, pl. ro, fig. 11 only. 1918. Homalonotus edgelli Salter : Reed, p. 270. The type and only known specimen is an incomplete, damaged pygidium, pre- served as an internal mould, having an estimated projected length of 9 mm. and a frontal breadth of about 11 mm. It is strongly convex transversely and highest frontally, its dorsal surface moderately declined rearwards. The anterior margin is convex forwards medially, and terminates laterally in a pair of large facets. The lateral margins are relatively straight for the genus and converge from the antero- lateral facets at roughly 60 degrees. The tip of the pygidium is missing but the border there appears to have possessed a small lateral flange. The narrow pygidial axis occupies about one-third of the frontal breadth and tapers gently backwards, delimited laterally by deep, straight, axial furrows. The surface of the axis is damaged but, excluding the articulating half-ring, there are five axial rings in just over half the length. The side-lobes are convex, narrowing backwards and declined steeply towards the lateral margins; each has a well-developed facet and carries seven 354 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II pleural furrows, the first two of which are deep, the others becoming shallower and extending almost as far as the tip of the axis. All the pleural furrows are well defined as far as the inner margin of the doublure, but can scarcely be traced to the lateral margins owing to the poor state of preservation. HORIZON AND LOCALITY. The specimen is a limonitic internal mould embedded in a fragment of green-brown, fine-grained sandstone. The matrix is somewhat inde- terminate, but is suggestive of that found in the Lower Cheney Longville Flags. The label states merely ‘“‘ Acton Scott ’’, but as this name was often used in old collections to denote the whole of the Acton Scott-Marshbrook district it is of little value. A reliable appraisal of the stratigraphical position of Brongniartella edgelli must there- fore await the collection of further material. HOLoTyPeE. GSM. 35525. Brongniartella sp. (Pl. 55, fig. 5) One specimen has been collected, an incomplete pygidium, which is apparently distinct from the other forms of Brongniartella in south Shropshire and may repre- sent a new species. The pygidium, preserved as an internal mould, is almost semi- circular in outline, only moderately convex, with median length about two-thirds of the maximum breadth. The axis, occupying just over three-quarters of the median length and, frontally, rather more than one-third of the maximum breadth, tapers sharply rearwards. Laterally it is defined by straight, shallow, axial furrows, and it passes into the border with no development of a postaxial ridge. Excluding the articulating half-ring there are seven axial rings, extending almost to the tip of the axis. Each pleural lobe is moderately declined abaxially, with traces of up to seven ribs. There is evidence of small anterolateral facets and although the first two pleural furrows are deep the remainder become rapidly shallower, with the hindmost barely visible; they end abaxially at a narrow border which becomes broader (sag.) at and near the tip of the pygidium. HorRIZON AND LOCALITY. The pygidium was collected from a lenticular bed of weathered, sandy limestone occurring in massive sandstones of the Horderley Sand- stone at the south-eastern corner of Rookery Wood, three-quarters of a mile south- east of Horderley. The horizon is in the Lower Longvillian Substage, zone of Dalmanella indica and D. lepta. FIGURED SPECIMEN. BM. In. 51520. Discussion. Brongniartella caradociana is of particular interest as it is the earliest species of the genus known from south Shropshire, appearing suddenly in the top- most Costonian strata within a geographically restricted area of the county. The holotype is also the largest form of Brongniartella yet known in the Caradoc District prior to B. bisulcata, which does not appear until much later, at the base of the Upper Longvillian Substage. All the stratigraphically intervening species are of small size. Our incomplete knowledge of the pygidium B. caradociana makes comparison with other species difficult, but the cephalon is easily distinguished from that of B. bisu- Icata by the broad glabella and large size of the fixigenae. Small individuals provision- ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 355 ally assigned to B. cavadociana possess distinctive paraglabellar areas which have not yet been seen in any of the other Shropshire material of Brongniartella. The name Homalonotus rudis has sometimes been applied to specimens of Brongniartella now known to belong to B. cavadociana. H. rudis was first described and figured by M‘Coy (in Sedgwick & M‘Coy, 1851 : 168, pl. IE, figs. 20, 20a) from the Capel Garmon district, Denbighshire, and the syntypes are in the Sedgwick Museum, where they are numbered A. 10564 and A. 10565. Both are incomplete, distorted pygidia from which it is virtually impossible to obtain an accurate impression of the species, and it is suggested that the specific name should no longer be used until well-preserved topotype material has been adequately figured and described. The eastern North American species Brongniartella trentonensis (Simpson) has been shown by Whitcomb (1930) to be remarkably similar to B. bisulcata, but the glabella of the former is narrower and conspicuously more carinate than that of the latter, whilst the pygidium of B. trentonensis, although possessing a comparable number of axial rings and pleural ribs, is proportionately broader and has an axis which is more strongly convergent posteriorly. The characteristic species of the genus during the Lower Longvillian is B. minor, a form which may have developed from the closely similar b. minor subcarinata. The latter may be separated on the basis of the carinate glabella and the smaller number of axial rings and pleural ribs, and its presence in the Lower Longvillian is somewhat doubtful ; it is certainly more characteristic of the Lower Soudleyan and has not yet been found coexisting with B. minor. The relationship between the generally small species of Brongniartella existing in south Shropshire during the deposition of the Soudleyan and Lower Longvillian rocks, and the large B. bisulcata of the Upper Longvillian is still not clear. Certain other shelly fossils, for example Kloucekia and Bancroftina, exhibit a trend towards a larger size of individual above the boundary separating Lower and Upper Long- villian in south Shropshire, perhaps as the result of a change in environment, but such an explanation seems inadequate to explain the appearance of B. bisulcata, and it is more likely that the latter is an immigrant form, perhaps from elsewhere in the Anglo-Welsh area. REFERENCES ANGELIN, N. P. 1854. Palaeontologia Scandinavica. ix + 92 pp., 42 pls. Lund. Bancrort, B. B. 1945. The brachiopod zonal indices of the Stages Costonian to Onnian in Britain. J. Paleont., Tulsa, 19: 181-252, pls. 22-38. 1949. Upper Ordovician trilobites of zonal value in south-east Shropshire. Proc. Roy. Soc., London (B), 136 : 291-315, pls. 9-11. (Edited by A. Lamont.) BARRANDE, J. 1852. Systéme silurien du centre de la Bohéme. 1éve partie. Recherches paléonto- logiques, I. xxx + 935 pp., Atlas, 51 pls. Prague & Paris. 1872. Systéme silurien du centre de la Bohéme. léve partie. Récherches paléontologiques, I (supplement). xxx + 647 pp., atlas, 35 pls. Prague & Paris. Boeck, C. P. B. 1838. Uebersicht dey bisher in Norwegen gefundenen Formen dey Trilobiten- Familie. In Keirwau, B. M., “ Gaea Norvegica”’ 1 : 138-145. Christiania. CHOUBERT, G., HupE, P., VAN Leckwijcx, W. & SuTER, G. 1956. Sur l’Age caradocien des quartzites du pays des Sokhret (Maroc hercynien central). C.R. Acad. Sci., Paris, 242 (3) : 392-395. 356 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II Cooper, G. A. 1930. Upper Ordovician and Lower Devonian stratigraphy and paleontology of Percé, Quebec, Pt. II]. Amey. J. Sci., New Haven, 20 : 365-392, pls. 4, 5. Daiman, J. W. 1827. Om Palaeaderna eller de sa kallade Trilobiterna. K. svenska Vetensk. Akad Handl., Stockholm, 1 : 226-294, pls. 1-6. Dean, W. T. 1958. The faunal succession in the Caradoc Series of south Shropshire. Bull. Brit. Mus. (Nat. Hist.), Geol., London, 3 : 191-231, pls. 24-26. 1959. Duftonia, a new trilobite genus from the Ordovician of England and Wales. Palae- ontology, London, 2 : 143-149, pl. 19. 1959a. The stratigraphy of the Caradoc Series in the Cross Fell Inlier. Pvoc. Yorks. Geol. Soc., Hull, 32 : 185-228, 6 figs. DEAN, W. T. & DINELEY, D. L. 1961. The Ordovician and associated Pre-Cambrian rocks of the Pontesford district, Shropshire. Geol. Mag., Lond., 98: 367-376, pl. 20. Deto, D. M. 1935. A revision of the Phacopid trilobites. J. Paleont., Tulsa, 9 : 402-420, 45 figs. 1940. Phacopid trilobites of North America. Geol. Soc. Amer. Spec. Papers, Baltimore, 29 : vii + 135, 13 pls. DreyFuss, M. 1948. Contributional’étude géologique et paléontologique del’Ordovicien Supérieur de la Montagne Noire. Mém. Soc. géol. Fy., Paris, 27, Mém. 58 : 1-62, pls. 3-11. EmmricH, H. F. 1844. Zur naturgeschichte dev Trilobiten. Realschul-Programm. Meiningen. ForrsteE, A. F. 1910. Preliminary notes on Cincinnatian and Lexington fossils of Ohio, Indiana, Kentucky and Tennessee. Bull. Sci. Lab. Denison Univ., Granville, 16 : 15-87, pls. 1-6. tg1g. Silurian fossils from Ohio, with notes on related species from other horizons. Ohio Jour. Sci., Columbus, 19 : 367-404, pls. 16-19. Harper, J.C. 1947. The Caradoc fauna of Ynys Galed, Caernarvonshire. Ann. Mag. Nat. Hist., London, (11) 14: 153-175, pls. 6, 7. Hawte, I. & Corpa, A. J. C. 1847. Prodrom einer Monographie der béhmischen Trilobiten. Abh. K. Bohm. Ges. Wiss., Prague, 5 : 117-292, 7 pls. Howe tt, B. F. 1951. The Vogdes Collection of trilobites. Tvans. San Diego Soc. Nat. Hist. 11 : 257-328, 13 pls. Hurt, P. 1955. Classification des trilobites. Aun. Paléont., Paris, 41 : 91-325, figs. 93-247. Jaanusson, V. 1953. Uber die Fauna des oberordovizischen Slandrom-Kalksteines im Siljan- Gebiet, Dalarna. Geol. Féren. Stockh. Férh., 75 : 97-105, 4 figs. Kine, W. B. R. 1928. The geology of the district around Meifod (Montgomeryshire). Quart. J. Geol. Soc. Lond., 84 : 671-700, pl. 52. Koztowsk1, R. 1923. Faune dévonienne de Bolivie. Ann. Paléont., Paris, 12 : 1-112, pls. I-10. La ToucueE, J. D. 1884. A Handbook of the Geology of Shropshire. 91 pp., 22 pls. London & Shrewsbury. Linnarsson, J. G. O. 1869. Om Vestergotlands cambriska och siluriska aflagringar. K. svenska. Veiensk-A kad. Handl., Stockholm, 8 : 1-89, 2 pls. M‘Coy, F. 1849. On the classification of some British fossil Crustacea, with notices of some new forms in the University collection at Cambridge. Ann. Mag. Nat. Hist., London, (2) 4: 161-179, 330-335, 392-414. McLearn, F. H. 1924. Palaeontology of the Silurian rocks of Arisaig, Nova Scotia. Mem. Geol. Surv. Canada, Ottawa, 137 : 1-179, 30 pls. Moore, R. C. 1959. Treatise on Invertebrate Paleontology. Part O. Arthropoda I. xix + 560 pp., 415 figs. Lawrence & Meriden. Morris, J. 1854. A Catalogue of British fossils : comprising the geneva and species hitherto des- cribed ; with veferences to theiv geological distribution and to the localities in which they have been found. 2nd edit. viii + 372 pp. London. Murcuison, R. I. 1839. The Silurian System founded on geological vesearches in the counties of Salop, Hereford, Radnor, Montgomery, Caermarthen, Brecon, Pembroke, Monmouth, Gloucester, Worcester and Stafford; with descriptions of the coal-fields and overlying formations. Xxxii + 768 pp., 40 pls. London. 1854. Siluria. xv + 523 pp., 37 pls. London. 1859. Siluvia. 3rd edit. xix + 592 pp., 41 pls. London. ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, II 357 Murcuison, R. I. 1867. Siluvia. 4thedit. xvii + 566 pp., 41 pls. London. 1872. Siluria. 5thedit. xvii + 560 pp., 41 pls. London. Pirtet, J. 1954. La classification des Phacopacea (Trilobites). Bull. Soc. géol. Fr., Paris, (6) 3 : 817-839, pls. 1-3. Pocock, R. W., WHITEHEAD, T. H., WEDD, C. B. & Ropertson, T. 1839. Shrewsbury District including the Hanwood Coalfield (One-Inch Geological Sheet 152 New Series). Mem. Geol. Surv. Gt. Britain, London. xii + 297 pp., 8 pls. Prantl, F. & Prispyt, A. 1948. Classification of the Bohemian Homalonotidae (Trilobitae). Bull. internat. Acad. tchéque Sci., Prague, 48, 9 : 1-24, 2 pls. ReeEp, F. R. C. 1896. Notes on the evolution of the genus Cheivurus. Geol. Mag., Lond., (4) 3: 117-123, 161-167. 1905. The classification of the Phacopidae. Geol. Mag., Lond., (5) 2: 172-178, 224-228. 1918. Notes on the genus Homalonotus. Geol. Mag., Lond., (6) 5 : 236-276, 314-327. 1927. Recent work on the Phacopidae. Geol. Mag., Lond., 64 : 308-322, 337-353. ReEMELf, A. 1884. “‘ A new phacopid genus Homalops’’ (Vortrag in Deutscher Geologischer Gessellschaft den 5. Marz 1884). Z. dtsch. geol. Ges., Berlin, 36 : 200. 1885. Katalog dey beim international Geologen-Congress zu Berlin ausgestellten Geschiebe- sammlung. 30 pp. Berlin. SALTER, J. W. 1848. In Pui.ips, J. & SALTER, J. W., Palaeontological Appendix to Professor John Phillips’ Memoir on the Malvern Hills, compared with the Palaeozoic Districts of Abberley, &c. Mem. Geol. Surv. U.K., London, 2 : 331-386, pls. 4-30. 1849. Figures and descriptions illustrative of British organic remains. Dec. 2. Mem. Geol. Surv. U.K., London, pls. I-10. 1852. A systematic description of Palaeozoic fossils in the Geological Museum of the University of Cambridge. Appendix A. Description of a few species from Wales and Westmoreland referred to in the foregoing work. Appendix to Sedgwick & M‘Coy, 1851: 1—Vv111. 1853. Figures and descriptions illustrative of British organic remains. Dec. 7. Mem. Geol. Surv. U.K., London : 1-78, pls. 1-10. 1864. A Monograph of the British Trilobites from the Cambrian, Silurian, and Devonian formations, 1. 1-80, pls. 1-6. Palaeontogr. Soc., [Monogr.], London. 1865. On the fossils of North Wales. Jn A. C. Ramsay, The geology of North Wales, PP. 239- 363, pls. 1-26. — 1865a. A Monograph of the British Trilobites from the Cambrian, Silurian and Devonian formations, 2. 81-128, pls. 7-14. Palaeontogr. Soc., [Monogr.], London. 1873. A catalogue of the collection of Cambrian and Silurian fossils contained in the geological museum of the University of Cambridge. xlviil + 204 pp. Cambridge. SALTER, J. W. & AVELINE, W. T. 1854. On the Caradoc Sandstone of Shropshire. Quart. J. Geol. Soc. Lond., 10 : 62-73. SavaGE, T. E. 1913. Stratigraphy and paleontology of the Alexandrian Series in Illinois and Missouri. Part I. Bull. Geol. Surv. Illinois, Urbana, 23 : 67-160, pls. 3-9. ScumipT, F. 1881. Revision der Ostbaltischen Silurischen Trilobiten. I. Phacopiden, Cheiru- riden und Encrinuriden. Mém. Acad. Sct. St. Pétersb., 30 : i-iv, 1-237, pls. 1-16. Spzuy, K. 1957. Bemerkungen zur Familie Homalonotidae (mit der Beschreibung einer neuen Art von Calymenella). Senckenbergiana, Frankfurt a. M., 38 : 275-290, 1 pl. SEpGwick, A. 1847. On the classification of the fossiliferous slates of North Wales, Cumberland, Westmoreland and Lancashire. Quart. J. Geol. Soc. Lond., 3 : 133-164. SEDGWIcK, A. & M‘Coy, F. 1851-55. A synopsis of the classification of the British Palaeozoic vocks, with a systematic description of the British Palaeozoic fossils in the geological museum of the University of Cambridge. 1851 :i-iv, 1-184, 11 pls.; 1852: i-x, 185-406, pls. 1H, I, K, L, 2A, B; 1855 : i-xcvili, 4—7-662, pls. 2 C, D, 3 A-I, K. London & Cambridge. STORMER, L. 1940. Early descriptions of Norwegian trilobites. The type specimens of C. Boeck, M. Sars and M. Esmark. Norsk. geol. Tidssky., Oslo, 20 : 113-151, pls. 1-3. 358 ORDOVICIAN TRILOBITE FAUNAS OF SOUTH SHROPSHIRE, ii STORMER, L. 1945. Remarks on the Tyvetaspis (ITvinucleus) Shales of Hadeland, with des- cription of trilobite faunas. Norsk. geol. Tidssky., Oslo, 25 : 379-426, pls. 1-4. 1953. The Middle Ordovician of the Oslo Region, Norway. 1. Introduction to strati- graphy. Norsk. geol. Tidssky., Bergen, 31 : 37-141, pls. 1-6. STUBBLEFIELD, C. J. 1939. Some aspects of the distribution and migration of trilobites in the British Lower Palaeozoic faunas. Geol. Mag., Lond., 76 : 49-72. Tripp, R. P. 1954. Caradocian trilobites from mudstones at Craighead Quarry, near Girvan, Ayrshire. Tvans. Roy. Soc. Edinb., 62 : 655-693, 4 pls. TROEDsson, G. T. 1929. On the Middle and Upper Ordovician faunas of Northern Greenland. Pt. Il. Med. Gronland. Kjobenhavn., 72 : 1-197, 56 pls. Wa tcott, C. D. 1879. Descriptions of new species of fossils from the Chazy and Trenton Lime- stones. 31st Ann. Rept. New York State Mus. Nat. Hist., Albany : 68-71. WarsurG, E. 1925. The trilobites of the Leptaena Limestone in Dalarne, with a discussion of the zoological position and the classification of the Trilobita. Bull. Geol. Inst. Univ. Uppsala, 17 ; i-vill, 1-446, pls. I-11. WEDD, C. B., Situ, B., Kine, W. B. R. & Wray, D. A. 1929. The country around Oswestry. Expln. Sheet 137. Mem. Geol. Surv. Engl. & Wales, London. xix + 234 pp., 4 pls. Wuitcoms, L. 1930. New information on Homalonotus trentonensis. Bull. Geol. Soc. Amer., New York, 41 : 341-350, pls. 5, 6. WuittarpD, W. F. 1960. The Ordovician trilobites of the Shelve Inlier, West Shropshire, 4 : 117-162, pls. 16-21. Palaeontogr. Soc. [Monogr.|, London. 1961. The Ordovician trilobites of the Shelve Inlier, West Shropshire, 5 : 163-196, pls. 22-25. Palaeontogr. Soc. [Monogy.], London. WuittineTon, H. B. & Evitt, W. R. 1953. Silicified Middle Ordovician trilobites. Mem. Geol. Soc. Amer., New York, 59 : 1-137, 33 pls. Wuittincton, H. B. & Witttiams, A. 1955. The fauna of the Derfel Limestone of the Arenig District, North Wales. Philos. Tvans., London (B), 238 : 397-430, pls. 38-40. WiLiiamMs, A. 1953. The geology of the Llandeilo District, Carmarthenshire. Quart. J. Geol. Soc. Lond., 108 : 177-208, pl. 9. Witson, A. E. 1947. Trilobita of the Ottawa Formation of the Ottawa-St. Lawrence Lowland. Bull. Geol. Surv. Canada, Ottawa, 9 : i-v, 1-86, Io pls. Wiman, C. 1907. Studien iiber das nordbaltische Silurgebiet. II. Bull. Geol. Inst. Univ. Uppsala, 8 : 73-168, pls. 5-8. Woops, H. 1891. Catalogue of the Type Fossils in the Woodwardian Museum, Cambridge. xvi + 180 pp. Cambridge. EXPLANATION OF PLATES Specimen numbers with prefixes BM. In., GSM. and A. are housed respectively in the British Museum (Natural History), London, the Geological Survey & Museum, London, and the Sedg- wick Museum, Cambridge. All specimens were whitened with ammonium chloride before being photographed. GEOL. 5. 8. 32 PLATE 409 Acanthoparypha stubblefieldi (Bancroft) . ; : = op. 3 Harnagian Stage, zone of Reuscholithus reuschi, section in old cartway near southern end of Smeathen Wood, Horderley. Fics. 1, 4. Incomplete, undistorted cranidium preserved as internal mould, showing outline and convexity of glabella. BM. In. 50499. 2:1. Fic. 3. Paratype, internal mould of pygidium. BM. In. 42085. x2. Figs. 5, 11. Lectotype, internal mould of incomplete, slightly distorted cranidium. BM. In. 42084. X 2:25. Fic. 6. Internal mould of fragment of glabella showing true outline posterolaterally. BM. In. 50498. 2:25. Pseudosphaerexochus sp. . é : : Sune OFC Bo) Onnian Stage, zone of Onnia superba, north bank of River Onny, 720 yards west-south-west of Wistanstow Church. Fics. 2, 7, 8. Almost complete distorted cephalon preserved as internal mould. BM. In. 50660. X2. Kloucekia (Phacopidina) harnagensis (Bancroft) . ; e320 Horizon and locality as for Fig. 1. Fics. 9, 14. Latex cast from almost complete external mould of dorsal exoskeleton, showing thoracic segments, pygidium, and upturned terminal spine. BM. In. 49131. X1°5. Encrinurus sp. (2 nov.) . F : F DS Le Marshbrookian Stage, zone of Dalmanella unguis, Marshwood Quarry, 4 mile south of Marsh- brook Station. Fics. 10, 12. Internal mould of pygidium. BM. In. 49312. 2:5. Atvactopyge sp. . 4 9 p. 318 Middle Actonian Stage, disused quarry by east side of road, opposite Church Farm, Acton Scott. Fic. 13. Latex cast from external mould. BM. In. 49319. x2. Figs. 3, 9, 10, 12-14, by Mr. J. Brown, remainder by the writer. Bull. B.M. (N.H.) Geol. 5, 8 PLATE 49 ACANTHOPARYPHA, ATRACTOPYGE, ENCRINURUS, KLOUCEKIA (PHACOPIDINA), PSEUDOSPHAEREXOCHUS { { “s r - ¥ : Q ; of j hay, F ~~ - = j ww ‘ 7 i = i ul i be a ry eu ry - 4 ' r @ j i % 799 0 » Pen ee Ce ree PLATE 50 Kloucekia (Phacopidina) hayvnagensis (Bancroft) . : ) ADAser Harnagian Stage, zone of Reuscholithus veuschi, section in old cartway near southern end of Smeathen Wood, Horderley. Figs. 1, 5. Lectotype B.M. In. 42089. x2°3. Fig. 2. Cranidium showing visual surface of eye. BM. In. 50576. x 4. Fic. 3. Paratype pygidium. BM. 42088. 3:5. Fic. 4. Internal mould of immature cranidium, showing deep first and second glabellar furrows. BM. In. 52211. x5. Kloucekia (Phacopidina) apiculata (M‘Coy) : : > jos san Upper Longvillian Substage, zone of Kjaevina bipartita. Soudley Quarry, by Soudley Pool, near Hope Bowdler. Fics. 6, 10. Internal mould of cephalon showing palpebral furrow and fixigenal spine. BM. In. 54864. x2. Fic. 12. Internal mould of whole individual with terminal spine preserved. BM. In. 50558. X15. Horizon as for Fig. 6. Bed of stream by west side of Horderley/Marshbrook road, 560 yards north-east of Crosspipes. Fics. 7, 8. Pygidium showing upturned terminal spine. BM. In. 51259. 2:2. Fic. 9. Latex cast from external mould of cephalon showing glabellar furrows and palpebral lobe. BM. In. 51260. x 3:3. Calyptaulax actonensis sp. nov. : ‘ 5 fe S45) Middle Actonian Stage, site of old quarry, Quarry Field, Gretton, near Obaeee Fics. 11, 14. Holotype pygidium, an internal mould. BM. 49771. 11, X2; 14, X2°5 Horizon as for Fig. 11. Probably from the north side of the Bishops Castle road, south-west of Wistanstow. Fic. 13. Internal mould of pygidium. BM. In. 48492. 2:2. Figs. 2, 12, 14, by Mr. E. W. Seaville, remainder by the writer. Bull. B.M. (N.H.) Geol. 5, 8 CALYPTAULAX, KLOUCEKIA (PHACOPIDINA) 0 eles ed 4 oe PIC IIE IS, 57 Calyptaulax actonensis sp. nov. ; : > jah 328) Middle Actonian Stage, site of old quarry, Quarry Field, Gretton, near cardiac Fic. 1. Paratype cranidium, an internal mould. BM. In. 49318. X2:3. Fic. 2. Latex cast from external mould of incomplete paratype cranidium. BM. In. 49768. 3. Chasmops extensa (Boeck) . : ; : 5. jh Reke Upper Longvillian Substage, zone of Kjaevina typa, old quarry below path at north-eastern corner of Burrells Coppice, south-east of Horderley. Fic. 3. Internal mould of incomplete cranidium. BM. In. 50547. 1-4. Fic. 7. Internal mould of pygidium. BM. In. 50546. X1I-1. Middle Actonian Stage, south bank of River Onny, 80 yards west of junction with Batch Gutter. Fics. 4, 8, 10. Almost complete large cephalon with part of test preserved. BM. In. 49127. XI. Fics. 5,9. Pygidium with test preserved. BM. In. 49128. x1. Fic. 6. Hypostoma preserved as internal mould. BM. In. 49126. xT. Figs. 2, 3 and 7 by Mr. E. W. Seavill, remainder by Mr. J. Brown. Bull. B.M. (N.H.) Geol. 5, 8 PLATE 51 CALYPTAULAX, CHASMOPS PLATE 52 Chasmops extensa (Boeck) . : : ‘ TPs. 33 Marshbrookian Stage, zone of Omnniella veuschi, road-fork 75 yards north-north-east of Woolston House, Woolston. Fic. 4. Incomplete cranidium. BM. In. 50548. x2. Middle Actonian Stage, south bank of River Onny 80 yards west of junction with Batch Gutter. Fic. 1. Enlargement to show surface of glabella near frontal rim. BM. In. 49127. x5. Fic. 6. Surface of glabellar lobes showing constitution of the large tubercles. BM. In. 49127. x 6. Fic. 7. Enlargement to show surface granulation of pygidium. BM. In. 49128. 5:5. Onnian Stage, zone of Onnia ? cobboldi, south bank of River Onny 45 yards east of junction with Batch Gutter. Fic. 3. Incomplete cranidium. BM. In. 50549. x1°8. Chasmops salopiensis sp. nov. : ; EP A337. Uppermost Actonian Stage, north bank of River Onny 30 yards east of f jumeaee with Batch Gutter. Fic. 2. Internal mould of pygidium tentatively assigned to the species. BM. In. 50552. x2. Fic. 5. Holotype, an incomplete cephalon with test partially preserved. BM. In. 50551. x2. Figs. 1, 6 and 7 by Mr. J. Brown, 2 and 5 by Mr. E. W. Seavill, 3 and 4 by the writer. Bull. B.M. (N.H.) Geol. 5, 8 as" } ' PAE Ss? ee CHAS MOPS ! i ri PLATE 53 Platycoryphe dentata sp. nov. : : c 7 P3400 Costonian Stage, Harknessella subquadvata Beds, old quarry near eastern end of Black Dick’s Coppice, Evenwood. Fic. 1. Holotype cranidium. GSM. WM 1335. X1I°5. Fic. 3. Paratype cranidium. GSM. WM 1371. x2. Fic. 6. Paratype pygidium. GSM. WM 1338. x2. Horizon as for Fig. 1. Small quarry in north-western part of Round Nursery, south of Harnage Grange. Fics. 2, 4. Paratype pygidium illustrating swollen terminal piece of the axis. BM. In. 54875. x2°8. Fic. 5. Large paratype cranidium showing paraglabellar areas. BM. In. 54876. 2:2. Fic. 7. Pygidium preserved as flattened external mould. BM. In. 54874. 2:2. Fic. 13. Librigena preserved as internal mould. BM. In. 54879. x18. Horizon as for Fig. 1. Old quarry by south-eastern side of brook, 1,500 yards south-west of Harnage Farm, Harnage. Fic. 9. Internal mould of incomplete cranidium with narrow form of glabella. BM. In. 49310. XI°5. Platycoryphe cf. dentata sp. nov. : : ¢ . Pp. 344 Harnagian Stage, zone of Reuscholithus veuschi, neptunian dyke in eastern corner of quarry on south side of Hazler Hill, east of Church Stretton. Fic. 8. Internal mould of incomplete cranidium. BM. In. 51726. x 2:2. Horizon probably as for Fig. 8. Locality shown only as “ Horderley ’’, but see p. 344. Fic. 10. Internal mould of small cranidium, figured by Salter (1865, pl. 10, fig. 10) as Homalonotus bisulcatus, but referred by him in the same paper (p. 108) to ? Homalonotus edgells. GSM. 35524. X3°4. Platycoryphe? sp. : ; : : - Pp. 344 Lower Longvillian Substage, zone of Dalmanella indica and D. lepta. Exposure west of south- western corner of Rookery Wood, south-east of Horderley. Fic. 11. Incomplete cranidium preserved as internal mould in decalcified sandstone. BM. In. 50511. X2. Brongniartella cavadociana sp.nov. . : : . Pp. 349 Horizon and locality as for Fig. 9. Fic. 12. Holotype cranidium preserved as an internal mould. BM. In. 51485. x1-75. Photographs 1, 3, 6 and 9 by Mr. J. Brown; remainder by the writer. Bull BM. (NH) Geol. 5, 8 . PLATE 53 BRONGNIARTELLA, PLATYCORYPHE i= f PLATE 54 Brongniartella bisulcata (M‘Coy) s : ‘ 5 PSS Upper Longvillian Substage, zone of Kjaevina bipartita (=Alternata Limestone). Localized merely as “‘ Whittingslow ’’, south of Marshbrook. Fic. 1. Lectotype cranidium with part of test preserved. GSM. 7626 (counterpart = BM. In. 54914). X15. Marshbrookian Stage, zone unknown, Marshbrook. Fics. 2, 4. Cephalon with librigenae in position. BM. In. 48592. x1. Marshbrookian Stage, zone of Dalmanella watts1, Marshwood Quarry, 4 mile south of Marsh- brook Station. Fic. 8. Flattened cranidium showing exaggerated tape of glabellar outline. BM. In. 50791. S108, Horizon as for Fig. 1. Soudley Quarry, near Soudley Pool, south-east of Hope Bowdler. Fic. 9. Internal mould of almost complete hypostoma. BM. In. 49557. 1-8. Brongniartella cavadociana sp.nov. . j : 7) P. 349 Costonian Stage, Harknessella subquadvata Beds, section in Bullhill Gutter, Evenwood. Fic. 3. Latex cast from external mould of incomplete thorax. BM. In. 54913. 2-4. Horizon as for Fig. 3. Old quarry in north-western part of Round Nursery, south-south-east of Harnage Grange. Fic. 5. Internal mould of incomplete thoracic segment. BM. In. 54880. x2. Horizon and locality as for Fig. 5. Fic. 7. Latex cast from external mould of paratype pygidium. BM. In. 54871. x4. Horizon as for Fig. 3. Old quarry in Black Dick’s Coppice, Evenwood. Fic. 10. Incomplete paratype cranidium with part of test intact, showing paraglabellar areas. BM. In. 50769. x3. Brongniartella minor (Salter) ; : : Pe Pag5r Lower Longvillian Substage, zone of Bancroftina typa, old quarry by west side of New House, south-east of Horderley. Fic. 6. Incomplete cranidium preserved as internal mould in decalcified sandstone. BM. In. 52315. x18. Photographs by the writer. Bull. B.M. (N.H.) Geol. 5, 8 PLATE 54 BRONGNIARTELLA rd ‘te ie ry! 7 oa ; ‘ a 0 . ss if f Ai = PLATE 55 Brongniartella bisulcata (M‘Coy) j . : - Pp. 346 Marshbrookian Stage, localized only as ‘‘ Marshbrook ’’. Figs. 1, 4. Pygidium with ten attached thoracic segments. GSM. 35519. x2. Horizon as for Fig. 1. ‘‘ Near Horderley.”’ Fic. 7. Internal mould of pygidium. BM. In. 48598. x1-1. Brongniartella cavadociana sp.nov. . ‘ ; 5 oh SY) Costonian Stage, probably zone of Harknessella subquadraia, ““ near Harnage Grange ”’. Fic. 2. Incomplete paratype cranidium with part of test preserved. Sedg. Mus. A. 43866. 15. Platycoryphe dentata sp. nov. : : ‘ EP As40 Harnagian Stage, zone of Reuscholithus veuschi, south bank of Coundmoor Brook, 1,340 yards south-west of Harnage Farm, Harnage. Fic. 3. Depressed pygidium with part of granulated test preserved, but dentate margin absent. BM. In. 54235. X 3:4. Horizon and locality probably as for Fig. 2. Fic. to. Latex cast from external mould of incomplete pygidium, GSM. 35526. The internal mould of the same specimen, GSM. 35527, was figured by Salter (1865, pl. 10, fig. 13; 18654, pl. 16, fig. 11) as ? Homalonotus vudis. 1. Brongniartella sp. : : , : - Pp. 354 Lower Longvillian Substage, zone of Dalmanella indica and D. lepta, south-eastern corner of Rookery Wood, south-east of Horderley. Fic. 5. Nearly complete pygidium preserved as internal mould. BM. In. 51520. 1:8. Brongniartella minor subcarinata subsp. nov. 0 é ED! 352 Soudleyan Stage, zone of Broeggevolithus broeggeri, stream-section at north-eastern corner of Smeathen Wood, Horderley. Fic. 6. Holotype, almost complete cranidium preserved as internal mould. BM. In. 5078r. K 2:5. Horizon as for Fig. 6. Section behind north-western corner of Glenburrell Farm, Horderley. Fic. 8. Internal mould of pygidium. BM. In. 51182. x2. Horizon as for Fig. 5. SmalJl quarry by west side of path through Longville Plantation, c. 60 yards from north-eastern corner. Fic. 13 2. Small, incomplete pygidium, preserved as internal mould and only tentatively referred to the subspecies. BM. In. 51726. 2:5. Brongniartella edgellt (Salter) : : ; - Pp. 353 Localized merely as “ Acton Scott’”’, but probably from the Cheney Longville Flags of the Marshbrook district. Fics. 9, 12. Holotype pygidium, an internal mould. GSM. 35525. x 2:5. Brongniartella minor (Salter) : : F 2 oh Site Soudleyan Stage, zone of Broeggerolithus soudleyensis, quarry by north side of road 100 yards south-east of Glenburrell Farm, Horderley. Fic. 11. Internal mould of incomplete cranidium. BM. In. 51529. x2. Platycoryphe? sp. ; : : ‘ passa Lower Longvillian Substage, zone of Dalmanella horderleyensis, west end of quarry by north side of road, 250 yards south-east of Glenburrell Farm, Horderley. Fic. 14. Internal mould of complete pygidium. BM. In. 50816. x2°5. Fig. 2 by Mr. J. Brown, Fig. 6 by Mr. E. W. Seavill, remainder by the writer. RISA Eee) Bull. B.M. (N.H.) Geol. 5, 8 BRONGNIARTELLA, PLATYCORYPHE es 73 7 oa : ee 2 | ry ’ a ae = acl flat THE MORPHOLOGY OF BOTRYOPTERIS ANTIQUA H. S. HOLDEN BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 5 No. 9 | LONDON : 1962 THE MORPHOLOGY OF BOTRYOPTERIS ANTIQUA BY HENRY SMITH HOLDEN Pp. 359-380 ; Plates 56-60 ; 12 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 5 No. 9 LONDON: 1962 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series, corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 5, No. 9g of the Geological (Palaeontological) series. © Trustees of the British Museum, 1962 LeMRaIEINP ARIS ID) 15S COR IDID IRL (OH ANISND APIS AMIDISS; Oy THE BRITISH MUSEUM Issued April, 1962 Price Fifteen Shillings THE MORPHOLOGY OF BOTRYOPTERIS ANTIQUA’ KIDSTON By HENRY SMITH HOLDEN SYNOPSIS An account is given of the morphology of Botvyopteris antiqua, a small fern from the Lower Carboniferous. It discusses the work of previous investigators and re-interprets this in the light of the discovery of additional material. The species is diagnosed and a lectotype chosen. INTRODUCTION In 1908 the late Robert Kidston published a relatively short description of a minute fossil fern from the Lower Carboniferous of Burntisland, Fifeshire, which he named Botryopteris antiqua. Although his account of the morphology of this plant was, in some respects, incomplete it provided adequate data for its ready recognition by subsequent investigators. Petioles from the Culm of Autun, France, closely resembling those described by Kidston had been known for some time and specimens from this locality were investigated by Pelourde (1910) who, in his paper, states that Kidston regarded the French specimens as being conspecific with those described by him. This view has been generally accepted (Bertrand, 1912 ; Scott, 1920 ; Corsin, 1937) although we have no adequate knowledge of the stems to which the Autun petioles belonged. It is a point of interest in this connection that, whilst with very rare exceptions, the petioles from Burntisland are oval in transverse section those from Autun that have been studied by the author consistently show a well-marked adaxial concavity (Pl. 60, fig. 4) a characteristic also recorded in the specimens from Autun examined by Corsin (1937 : 223, 225). This difference is one of the criteria used in the specific separation of the Cheilanthoid ferns and it suggests the possibility that the French specimens may well belong to a different, though closely allied, species. Among the more important characters established by Kidston was the fact that the xylem of the stem stele consisted entirely of tracheids without any admixture of parenchyma. He was unable, however, to distinguish in the stem any groups of protoxylem. He also described the main features of the petiole and leaf trace. In the latter he interpreted as protoxylem a band of small tracheids with scalariform thickening normally occupying the more pointed adaxial side of the trace but in some examples, such as that in his fig. 8, forming a narrow band down both sides of the metaxylem and even extending to the abaxial margin. The thickenings of the tracheids of the petiolar metaxylem he described as porose. He also figured a speci- men (figs. 11, 12) which he interpreted as being an example of petiolar dichotomy. Three years after the publication of Kidston’s paper Margaret Benson, who had obtained further specimens from Burntisland, gave a fuller description of the plant and, among other things, suggested a different interpretation of some of the sections GEOL. 5,9 33 362 THE MORPHOLOGY OF BOTRYOPTERIS ANTIQUA figured by Kidston (Benson, 1911). Benson’s material enabled her to trace the connection between the vascular supply of the petiole and that of the stem and to show that the protoxylem in the petiole was initially mesarch in position. She regarded the plant as “‘ heterophyllous ’’ and considered that the stem bore petioles of two kinds. One of these, which she termed the monarch type, had a single adaxial pro- toxylem group and was accompanied by a uninerved sheathing organ which she interpreted “‘ as of the nature of an aphlebia’’. The second type of lateral structure arising from the stem, which she regarded as a diarch petiole, possessed two adaxial protoxylem groups and was not accompanied by an aphlebia-like structure. She stated that the protoxylem in both types of petiole became “‘ aggregated into one or two adaxial grooves ”’ and that the diarch traces appeared to be preceded and followed by monarch ones. The smaller elements and protoxylem according to her show all intermediate types of thickening from porose to reticulate. The sections studied by both Kidston and Benson were prepared by grinding down thin slices of the rock containing the fossil until they were transparent enough to be examined by transmitted light. The drawback to sections prepared in this way, especially in dealing with incomplete and frequently fragmentary specimens, is that it is often impossible to be certain of the exact relationships between the separate parts such as stem and petiole and even to be certain that they belong to the same plant. With the advent of modern “ peel ’’ techniques some of these difficulties have disappeared and the most recent investigation of Botryopteris antiqua by Surange (1952) using such techniques has added materially to our knowledge of the plant. He has established the fact that the tip of the young petiole was circinately coiled (Pl. 59, fig. 3) and has suggested a new interpretation of the morphology differing radically from that put forward by Benson. Surange interprets the structures termed diarch petioles by Benson as dorsiventral stems from which radial stems, possessing an abundant system of adventitious roots, arise and on which true petioles (i.e. Benson’s monarch petioles) are developed in spiral succession. The development of a radially organized stem from the diarch axis is initiated by the separation from it of a lateral trace. This is smaller than the parent axis and carries with it one of the two protoxylem groups. The radial stem, which is characterized by the smaller calibre of its tracheids, develops on the adaxial side of the branch near its base and is adnate to it. Surange states that, in the material examined by him, the structure he terms a dorsiventral stem “‘ invariably gives rise to a radial stem and to no other structure ’’. He also states that the tracheids of the petiolar protoxylem show spiral thickenings. Prior to the publication of Surange’s paper a number of specimens of Botryopteris antiqua from Burntisland had been collected with the intention of making a fuller study of its morphology but, in view of the differing interpretations resulting from the work of the previous investigators to which reference has already been made, it was considered desirable to examine also the sections upon which they had based their views. For this purpose Professor Walton, Head of the Botany Department of Glasgow University, kindly loaned to the Museum not only the slides figured by Kidston but also the whole of the slides of Botryopteris antiqua in the Kidston Collection. In addition to these he also loaned slides containing sections of the petiole of B. antiqua THE MORPHOLOGY OF BOTRYOPTERIS ANTIQUA 363 from the departmental Figured Slide Collection which he had collected from a locality in Dunbartonshire. Professor Jane, of the Royal Holloway College, was equally helpful with regard to the slides of Botryopteris antiqua from the Benson Collection as was Dr. Hamshaw Thomas with regard to those studied by Surange. The grateful thanks of the author are due to all these fellow botanists for their ready co-operation. In addition to the Kidston, Benson and Surange slides several from the Gordon Collection were kindly looked out and sent over to the Museum by Dr. O. Bradley of the Department of Geology, King’s College, London, and Mr. A. G. Long sent me his sections of the specimen of B. antigua to which reference is made in Surange’s paper. These slides together with those in the Oliver and Scott Collections and that of the author form the basis of the present account but, in addition, a large number of “peels ’’ have been made from blocks of Lower Carboniferous material from that part of the Hemingway Collection which is now in the British Museum (Natural History), Department of Palaeontology. The majority of these “ peels ’’ have been prepared by C. H. Shute, Scientific Assistant in the Department. The study of the considerable number of specimens indicated above, whilst it has not provided an answer to all the morphological problems raised by this Lower Carboniferous fern, has enabled a fuller and more detailed account of its morphology to be given than had been possible previously. DESCRIPTION The Stem A typical plant of Botryopteris antiqua possesses a radially organized stem develop- ing abundant adventitious roots. These roots are most numerous at and near the base of the stem but are also produced quite close to the apex (e.g. Pl. 56, figs. I, 3). The stem bears closely set spirally arranged petioles which, in many specimens, show a two-fifths divergence (Pl. 56, fig. 1; Text-figs. 1-4) though this is by no means constant. Its centre is occupied by a stele which has the shape of a shallow dome. The stelar xylem is composed entirely of tracheids which are smaller in diameter than those of the petioles (Pl. 56, fig. r) and which show predominantly scalariform pitting though this type of pitting may be replaced by a reticulate type with transversely elongated pits at and near the periphery. Near the stem apex where the differentiation of the tracheids is not completed thin-walled cells occur scattered among the xylem elements which are already lignified (Pl. 59, fig. 6; Text-fig. 5). The protoxylem occupies a mesarch position and consists of single minute tracheids or groups of these distributed near the outer part of the stele. The determination of the nature of the protoxylem thickenings proved to be a matter of some difficulty but its spiral character is clearly shown in one of Surange’s slides (B 78) and is illustrated in Pl. 58, figs. 1, 2. Neither the phloem and its associated parenchyma nor the inner cortex are preserved in most specimens so that there is a gap between the stelar xylem and the tissues forming the middle and outer cortex (Pl. 56, fig. 1; Text-figs. 1-4). There are a few sections, however, in which some or all of these tissues are present and two of these from the Gordon Collection are illustrated in Pl. 56, fig. 2 and Pl. 57, fig. 1. In these sections there are indications that the phloem consisted of a 304 THE MORPHOLOGY OF BOTRYOPTERIS ANTIQUA more or less continuous ring of large sieve tubes and phloem parenchyma similar to that occurring in Ankyropteris corrugata and many other Palaeozoic ferns. The bulk of the inner cortex is composed of small-celled parenchyma but, scattered among this, there are a number of groups of thick-walled cells with dark contracted contents (Text-fig. 6). The middle cortex consists of larger, somewhat thick-walled, cells many of which, like the similar groups in the inner cortex, possess dark contents (Pl. 56, fig. 2). These thick-walled cells merge into a well-defined parenchymatous ec i Fics. 1-4. Camera lucida drawings of four transverse sections showing a two-fifths spiral sequence of monarch petiole traces. Figs. 1-3 show three stages in the emission of a re- duced lateral outgrowth (vo). The petiolar bundles are numbered in the order of their development. All x15. B.M. (N.H.) Holden Coll., slides H.92, H.153, H.177, H.192. outer cortex of cells larger than those of the inner cortex enclosed by the small-celled epidermis. The surface of the stem, especially at and near the apex, bears numerous hairs. These are pointed and predominantly uniseriate though occasionally they possess a stouter multicellular base (Text-figs. 7a—g). They are present also on the basal parts of the petioles but do not appear to extend on to the more distal parts of THE MORPHOLOGY OF BOTRYOPTERIS ANTIQUA 365 these. The actual stem apex is roofed in by a closely set mantle of these hairs which acted, presumably, as a protective cover for the delicate underlying tissues. Two specimens of the stem in which the stele shows an equal dichotomy have been noted. The better example of this condition is that shown in Slide No. 549c and figured by Kidston (1908, fig. 6). Fics. 5,6. Fig. 5, Transverse section of part of the stem stele near its apex showing the incompletely differentiated metaxylem and an isolated mesarch protoxylem tracheid. B.M. (N.H.) Oliver Coll. 1247. x 80. Fig. 6, A small group of dark sclerized cells from the inner cortex of the stem. Gordon Coll., King’s College. 80. ale Fic. 7. Hairs from the stem and the base of the petiole. a—d, Kidston Coll., Glasgow University, slide 549c¢ ; e-g, Hemingway Coll., B.M. (N.H.) No. V.35555 (sections 35, 52 and 98 respectively). All * 135 approx. The Petiole As has been stated already, the petioles arise from the stems in a close spiral and usually show an approximately two-fifths divergence (Text-figs. 1-4). In the great majority of specimens the xylem and cortex are the only tissues preserved and even these may show some degree of breakdown. In the case of the xylem the protoxylem is the only part affected but the cortical tissues may become disorganized and, partly or almost wholly, be replaced by aggregates of calcite. These have crushed and ruptured the cortical cells and produced a pseudo-cellular structure in which the sphaerules of calcite have become surrounded by the broken down cell remnants (Pl. 57, fig. 6). This condition is not uncommon in petrified plant remains from Burntisland and was noted and figured by Gordon in some of the specimens of Diplolabis examined by him (Gordon, 1911). The protoxylem of the petiole trace is initially mesarch with a single protoxylem 366 THE MORPHOLOGY OF BOTRYOPTERIS ANTIQUA group situated close to the adaxial surface. This mesarch position of the protoxylem persists for a short time after the separation of the trace from the stem stele but endarchy is usually attained by the time the petiole is free from the stem. The petiole trace is ovoid in transverse section with the spirally thickened protoxylem tracheids forming a distinct adaxial ridge (Pl. 56, fig. 1). Quite commonly, however, the protoxylem tracheids show partial or more or less complete disintegration and either are not preserved at all or are represented by dissociated fragments of the spiral thickening. An example of this latter condition is shown in Pl. 58, fig. 1, from a longitudinal section in the Figured Slide Collection of the Botany Department, Glasgow University. It is this variation in the degree of preservation of the protoxy- lem which probably accounts for the differing descriptions of Kidston (1908), Benson (1911) and Surange (1952). It seems evident that the small tracheids described by Kidston as protoxylem are in fact metaxylem elements and that the median groove described by Benson as situated in the centre of the protoxylem group is due to the loss of the spirally thickened tracheids forming the apex of the protoxylem ridge. The occurrence of this groove is relatively infrequent and is, in no sense, a characteris- tic feature of the trace. I am thus in agreement with the description of Surange who correctly describes the character and position of the protoxylem. The narrower metaxylem tracheids show scalariform thickening. Most of the metaxylem, how- ever, the tracheids of which are much larger in diameter than those of the stem, possess two or more rows of transversely elongated pits (PI. 58, fig. 1). Occasionally, traces of the phloem and its associated parenchyma are preserved but these are usually too small in amount to yield much information. The best preserved example so far discovered is that illustrated in PI. 58, fig. 4, in which a strip of sieve tubes and phloem parenchyma occurs on the abaxial side of the vascular bundle. The endo- dermis consists of small, tangentially flattened thin-walled cells devoid of thickening and which may have dark contents. The inner cortex consists of parenchymatous cells, some of which may have been secretory in character. The latter are distributed singly through the rest of the parenchyma and are distinguished by their brown contents (Pl. 58, fig. 5). They bear a close resemblance to the mucilage-secreting cells commonly occurring in the cortex of many existing ferns. The cortical cells diminish in size towards the exterior and, in mature petioles, these smaller cells develop thickened walls and form a well-defined sclerenchymatous outer zone. The small-celled epidermis shows no specially distinctive features, apart from the uni- seriate surface hairs developed at and near the bases of the petioles. The branching of the petiole follows no set pattern and the examples figured by Surange (1952) adequately illustrate its general character. The branches undergo progressive diminu- tion in size and may undergo, occasionally, equal bifurcation near their junction with a branch of a higher order, as in the example shown in PI. 60, fig. 5. Associated with the vascular bundles which supply the petioles are much smaller monarch traces supplying the organs which Benson termed aphlebiae and which she described as uninerved sheathing organs. According to her, the monarch petiolar bundles “‘ apparently never come off (from the stem stele), pass through the cortex and enter the petiole without being accompanied by aphlebiae’’. Surange, although describing and figuring one of these small monarch bundles as being given off from THE MORPHOLOGY OF BOTRYOPTERIS ANTIQUA 307 the stem stele immediately following the emission of the first petiole trace (Surange, 1952, text-fig. 1, cf) does not record or figure them as constantly accompanying subsequent petiole traces. The sections studied during the present investigation do not support the view that they accompanied all, or indeed the majority, of the petiole traces. Where they are present, their relation to these traces is very variable and the two examples figured here show a wide divergence in their point of origin from the stem stele. The first, illustrated in Pl. 56, fig. 1 and Text-figs. 1-3, arises quite inde- pendently as a small endarch trace between the first and second of the petiole traces figured and is approximately equidistant between them. It ultimately supplies a small outgrowth of the stem which is ovoid in transverse section with the flatter side adaxial. The two examples illustrated by Benson (1011, text-figs. 1a, 6) appear to supply similar outgrowths. Both are from thick and somewhat poorly preserved sections but that shown in her text-fig. 1a clearly shows the beginning of the ovoid shape in transverse section which is assumed when the “ aphlebia’’ becomes free from the stem cortex. The second example (Benson’s text-fig. 1b) appears to be more closely associated with an outgoing petiole and has an irregular shred of cortical tissue linking it with this. As in her text-fig. ra, there are indications that, in this example also, the “ aphlebia ’’ was ovoid and not winged in transverse section. The second example figured in the present paper is one in which the vascular bundle destined for the petiole and that destined for the small monarch trace arise simultaneously from the stem stele and are initially completely fused. As the petiolar trace passes into the stem cortex the small bundle becomes independent and diverges from the larger one though it accompanies it into the petiole (Pl. 56, figs. I, 2). Whether this bundle should be interpreted as an abortive “ aphlebia ’’ or a precocious branch of the petiole is a matter of opinion. It should be pointed out, however, that a complete series exists ranging at the one extreme from a strand arising from the stem stele without any obvious connection with that supplying the nearest petiole to that just described in which the linkage between the two is close and prolonged. Whatever the interpretation of the morphological nature of the structures supplied by these strands may be, they differ widely from the type of structure to which Solms-Laubach (1891) applied the descriptive name “‘ Aphlebia’’. This was applied by him to a distinctive and anomalous type of pinna or pinnule possessing a reduced and deeply dissected lamina which occurred on the rachises of certain fossil ferns and pteridosperms. These were compared by him to similar pinnae present in Hemitelia capensis and other existing ferns. Subsequent authors have applied the name to the scale-like structures occurring on the stems of several of the zygopterid ferns and to the somewhat similar pinnules borne on their petioles (Potonié, 1903 ; Seward, 1910; Bertrand, 1911 ; Gordon, 1911 ; Scott, 1920). This extended use of the name “‘ Aphlebia ’’ does not appear to serve any really useful purpose and it would seem preferable to limit it to the type of structure to which Solms-Laubach originally applied it. The Diarch Axes In addition to the monarch petioles, there were other, generally stouter, axes possessing two adaxial protoxylem groups, which were obviously parts of the same GEOL, 5,9 33$ 368 THE MORPHOLOGY OF BOTRYOPTERIS ANTIQUA plant. These, as already stated, were regarded by Benson (1911) as diarch petioles but have been interpreted provisionally by Surange (1952) as dorsiventral stems from which the radially organized plants arose. We owe the detailed description of the development of these radially organized plants from the diarch axes to Surange and, apart from minor details, there is little to add to his account of the specimens he studied. A typical specimen, cut obliquely and showing two stages in the forma- tion of a radial stem, is illustrated in Pl. 56, figs. 3, 4. It is evident that Benson’s material was inadequate for a similarly detailed study and this led her to misinterpret the sections figured in her paper as “ successive transverse sections through a stem stele ’’ (Benson, r1g1I, pl. 81, figs. I-5). Re-examination of the relevant slides in the Benson Collection leaves little doubt that they represent part of a diarch axis from which a radial stem is developing. The section figured by Kidston (1908, fig. 11) and described by him as that of a petiole showing an early stage in petiolar dichotomy is also probably one of a diarch axis showing an early stage in the development of a radial stem. Whilst, however, both the Benson and Kidston specimens described above fit without difficulty into Surange’s interpretation of the behaviour of the diarch axis this interpretation is unsatisfactory when applied to the specimen figured by Benson (r911, pl. 82, figs. 11-13, 11a—13a@). In the sections from which this series of figures has been made there is no indication that the diarch axis has divided into two monarch daughter axes, one of which is destined to be the source of a radial stem. The figured sections are interpreted by Benson as “a series through a stem bearing at first a diarch trace and later roots ’’’. She also states that “ the series shows diagrammatic- ally the change in character of the stem as it passes from a purely leaf-bearing zone into a root-bearing zone’’. The sections upon which this interpretation is based are the second, third and fourth of a series of five slides numbered 417-1 to 417°5, figs. I1 and 11a showing the stage present in 417-4, figs. 12 and 12a that present in 417°3 and figs. 13 and 13a that present in 417-2. Of the two remaining slides in this series 417'5 shows no substantial change from that in 417-4 and 417-1 has been shattered during grinding. A comparison of Benson’s pl. 81, fig. 2 with her fig. 11a shows that these are essentially similar and it is suggested that both represent an early stage in the development of a radial stem. Whereas, however, the plant illustrated in her figs. 1-5 fits in with Surange’s account of the behaviour of the diarch axis that illustrated in the 417 series clearly differs since the radial stem originates on one side of this axis without involving its division into two monarch daughter axes. In addition to the Benson Collection specimen, there is a relatively short series of “peels ’’ in the author’s collection which presents some difficulties of interpretation when considered in connection with Surange’s views. This series consists of ten transverse sections numbered H.96 and H.103 to H.111 and ten longitudinal sections through the base of the specimen numbered H.112 to H.121. A typical transverse section shows a well-developed diarch axis linked with the stem by its cortical tissues (Pl. 59, fig. 1). The stele of this axis unites with that of the stem at a lower level and is well shown in slide number H.115 (Pl. 59, fig. 2). This slide, owing to the curvature of the specimen, shows the axis and the stem partly in transverse and partly in longitudinal section. The stem stele is clearly radial in type but the series THE MORPHOLOGY OF BOTRYOPTERIS ANTIQUA 309 is unfortunately too short to yield any information as to its origin. On the opposite side of the stele to that at which fusion with the diarch axis has occurred there is a large developing lateral trace occupying almost two-thirds of the area of the stele. The stem cortex consists almost entirely of thick-walled cells and only towards the lower transverse sections is there any indication of root development though at this level two root traces are coming off in close juxtaposition. Whether or not this speci- men is one of a stem from which two successive diarch axes have been produced must be an open question but if this interpretation is accepted it provides some support for the view put forward by Benson as to the plant being heterophyllous. It certainly differs to a marked extent, both in the nature of its cortical tissues and in its stelar characters, from a typical radially organized stem. Fic. 8. Camera lucida outline of a diarch petiole in transverse section which has given off a lateral strand (/s) and which in its turn has given off a daughter stele (ds). Fig. 8a, Daughter stele, x 150. B.M. (N.H.) Oliver Coll. 1272. An additional and different type of behaviour is also shown by some diarch axes which produce lateral branches having no apparent connection with rad‘al stem formation. Benson (1911 : 1048) refers to a specimen of this kind (Benson C llection, Series 390) and figures an early stage in the formation of such a lateral branch (pl. 83, fig. 24). These branches are initiated by the division of one of the two protoxylem groups of the diarch axis following which the outer protoxylem group, together with its associated metaxylem, separates off to form the branch trace. Three of the stages in the formation of such a branch are illustrated in Pl. 57, figs. 4-6, from slides in the Oliver Collection. In an additional section at a stage similar to that in Pl. 57, fig. 6, the lateral trace has, in its turn, developed a small, somewhat irregular group of tracheids on its abaxial side which suggests that it may be possibly an early stage in the formation of a stem (Text-figs. 8, 8a). Any firm view as to its significance, however, is impossible on the evidence of an isolated section and it is figured here primarily to record its existence. The specimens studied by Surange (1952) gave no clue to the origin of the diarch axes and it was clearly important to ascertain this, if possible, before any assessment 370 THE MORPHOLOGY OF BODRY OPDRERTS “AN DLOUWA could be made of their morphological status. Before attempting this, however, it seemed desirable that a review of the evidence upon which Surange had based his interpretation should be made. He, himself, realized that he had to rely on the criteria provided by comparative anatomy and that the acceptance of his view presented certain difficulties. Some of these, such as the pseudocollateral vascular structure, the possession of tracheids of wide diameter similar to those of the monarch petioles and differing from those of the radial stem and the fact that the axes do not appear to develop either petioles or a root system, were noted by him. In support of his view, however, he points out that among the Hymenophyllaceae the stems of certain species such as Hymenophyllum sericeum possess a pseudocollateral type of vascular structure closely resembling the condition present in the diarch axes of Botryopteris antiqua. It is doubtful, however, how much weight can be given to this comparison since Hymenophyllum sericeum is one of a series of filmy ferns in which there is a progressive reduction in the amount of vascular tissue this consisting, in the ultimate members of the series such as H. tunbridgense, of a minute pseudocollateral monarch strand (Boodle, 1900 ; Tansley, 1908 ; Ogura, 1938). There does not seem to be any evidence that Botryopteris antigua was a species in which reduction of the vascular system had occurred and it is suggested that a study of the anatomy of the stem and leaf stolons characterizing certain living ferns would provide a closer analogy with the diarch axes of B. antiqua than that used by Surange. Among the ferns developing stem stolons those which have been most extensively studied are various species of Nephrolepis (e.g. N. exaltata, N. cordifolia, N. neglecta) (Lachmann, 1885, 1889 ; Sperlich, 1906 ; Sahni, 1916) but Bower (1923, 1926) also records their occurrence in Plagiogyria pycnophylla and Thyrsopteris elegans. In all these ferns the symmetry of the vascular supply of the stolon is radial in type and, initially at least, is a simplified version of that of the stem of the parent plant. The ferns in which flagelliform leaf stolons are developed are among those which have been investigated by Kupper (1906) in his comprehensive study of the formation of sporophytic buds on the leaves of ferns. Among these there is a group of related species of small fern belonging to the genus Asplenium namely A. obtusilobum Hk., A. lauterbachti Christ and A. mannii Hk. The first of these three species in addition to the normal leaves produces a series of transitional forms showing progressive reduc- tion in the number of pinnae (Text-figs. 9a, b) the ultimate members being flagelliform stolons in which no pinnae at all are produced. These stolons and the modified leaves in which the pinnae are reduced develop sporophytic buds at intervals and are apparently of unlimited growth. Material of A. obtusilobum, a species from the New Hebrides, and of A. lauterbachi1, from New Guinea, was not available for study but through the good offices of Mr. N. C. Chase of Umtali, Southern Rhodesia, a good supply of A. mannii was obtained. The specimens were collected from the North Range of the Vamba Mountains where it grows as an epiphyte on the boles and branches of forest trees. These plants, in addition to the normal leaves, bear flagelli- form stolons (Hooker, 1854) on which sporophytic buds in various stages of develop- ment are produced, the longest of these in the material sent by Chase measuring 35 cm. and bearing ten attached plantlets (Text-fig. gc). It is hoped to make a THE MORPHOLOGY OF BOTRYOPTERIS ANTIQUA 371 detailed investigation of this species later but, in connection with the present paper, attention has been concentrated on the morphology of the bud-bearing stolons. The vascular supply of these stolons consists of a somewhat strap-shaped xylem strand surrounded by phloem and phloem ee Gin with a protoxylem group occupying each extremity of the trace (Text-fig. roa). When a sporophytic bud is to be formed ¥ Fic. 9. a, b, Two leaves of Asplenium obtusilobum showing progressive pinna reduction and the formation of sporophytic buds. Aftey Kupper, 1906. c, Part of a leaf stolon of Asplenium mannii showing three developing plantlets. All nat. size. one of the protoxylem groups accompanied by a small amount of metaxylem is detached from the rest of the leaf trace and becomes linked with the vascular supply of the developing sporophyte (Text-figs. 10), c). This condition appears to furnish a close parallel to the behaviour of the vascular bundle in the diarch axes of Botryopteris antiqua and suggests that, applying the criteria of comparative anatomy, these 372 THE MORPHOLOGY OF BOTRYOPTERIS ANTIQUA would be more accurately interpreted as specialized petioles rather than as dorsiventral stems. Whilst, however, on the criteria provided by comparative anatomy, there seem to be reasonable grounds for interpreting the diarch axes as petioles rather than stems it was considered that a search for specimens showing their origin was Fic. 10. Transverse sections of a leaf stolon of Asplenium mannii showing the develop- ment of the connection between the vascular bundle of the stolon and that of the sporophytic bud. x 170. essential and several of tbe slides in both the Oliver and Gordon Collections yielded useful information in this connection. Two of the slides from the Oliver Collection (Pl. 57, figs. 2, 3) are from consecutive sections numbered 1263 and 1264 and show two successive stages in the departure of a typical diarch strand. This diarch strand and that illustrated in Text-fig. 11a arise from stems which appear to be radial in THE MORPHOLOGY OF BOTRYOPTERIS ANTIQUA 373 type. That these are, in fact, successive stages in its departure is shown by an examination of the next slide in the series (No. 1265). The stem and petioles in this slide are badly disintegrated but the xylem of the stem stele is intact and in it the differentiation of the xylem is still incomplete. It is interesting to note that this stem does not show a close succession of spirally developed monarch strands. In the section shown in Text-figs. 11a, 6, however, a much smaller double vascular bundle has come off from the stem stele almost at right angles to that from which the diarch ie NA a A f \ { \ 1 i} ‘ t \ ’ \ / AS } . 4 x 7 . - ~ -_- 7s ewe 7 Fic. 11. a, Camera lucida outline of an oblique stem section showing an attached diarch axis and a small double vascular strand (dvs); b, the double vascular strand more highly magnified. B.M. (N.H.) Oliver Coll. 1274. a, X30; b, X1I50. strand has arisen. This smaller bundle appears to be the vascular supply of a monarch petiole and is composed of two unequal endarch xylem groups (Text-fig. 11d). It bears a close resemblance to those strands already described in which the main petiolar strand is accompanied by a smaller accessory strand. Of the three slides from the Gordon Collection which are figured, two (Pl. 57, fig. r and Pl. 58, fig. 3) show a large diarch strand, with two mesarch protoxylem groups, lying in close proximity to the radially organized stem stele. In the third (Pl. 56, fig. 2) a similar 374 THE MORPHOLOGY OF BOTRYOPTERIS ANGIOUA large diarch strand with mesarch protoxylems is dividing into a larger and a smaller bundle though these are still united. They show a pronounced difference from those cases in which a monarch petiolar strand is associated with a much smaller one. It is suggested that the dividing mesarch strand in this Gordon specimen may represent an unusually early division of the diarch axis prior to the formation of a radial stem from the smaller bundle. Like the specimens from the Oliver Collection, none of the Gordon specimens shows any indication of a close succession of spirally arranged monarch petioles. It would seem fair to sum up the evidence as to the nature of the diarch axes by stating that the data provided by comparative anatomy and the evidence yielded by newer discoveries among the specimens of Botryopteris antiqua itself favour the interpretation of the diarch axes as petioles and that their development leads to a pronounced modification in the emission of the monarch petiole traces though there is no convincing evidence that the monarch and diarch petioles were developed in alternating series. The Root System The adventitious root system is mainly composed of slender diarch roots which show a close measure of structural agreement with those of existing leptosporangiate ferns. They possess a separate cortex from their point of origin with the stem stele and, in this region, the xylem is bulkier than it is in the more distal parts thus forming an expanded base of attachment (Pl. 56, fig. 2). The phloem and its associated parenchyma and most of the parenchymatous inner cortex are seldom preserved. In the few cases where it has survived petrifaction the inner cortex has, scattered through it, an irregular ring of cells with dark contents similar in type to those present in the inner cortex of the petiole (Pl. 56, fig. 3). The endodermis, also like that of the petiole, is formed of small tangentially flattened cells which often have dark contents. The outer cortex consists of thick-walled cells, smaller than those of the inner cortex, these forming a zone several cells in depth (Pl. 56, figs. 3, 4). Asa rule, the surface layer of the roots has been eroded but even when this has survived no evidence has been found of the existence of root hairs. In addition to the roots of the type just described, there are others which appear to be most abundant towards the stem apex and which show pronounced morphological differences from these. The evidence for the existence of this second type of root was initially somewhat unsatisfactory as their morphology was obscured by partial collapse and distortion. Happily, among the plants from a block in the Hemingway Collection a stem occurred which had developed a number of uncrushed roots thus enabling their salient features to be determined. There appears to be no doubt that these organs are roots since, like those already described, they possess their own cortex from the time of their inception and their xylem consists of a small diarch plate with an exarch protoxylem group at each end (PI. 60, figs. 2, 3). The stele is relatively small and is surrounded by a wide cortex composed entirely of parenchyma, the cells of which have dense finely granular contents especially in the immature region near the apex (PI. 60, figs. 1, 2). The surface layer is formed of dark flattened cells in which no detailed structure can be discerned (PI. 60, figs. 1-3). LEE MORPHOLOGY OF BOFRYOPTERTIS ANTLIOUA 375 Any attempt to indicate the function of these roots must be somewhat speculative although their origin near the upper end of the stem and their morphology both suggest the possibility that they may be aerial and perhaps assimilatory in character. Sporangia and Spores Our knowledge of the sporangia and spores of Botryopteris antiqua is the least satisfactory part of the information available about this Lower Carboniferous plant though one can assign these to it with some degree of probability. The most satis- factory evidence yielded so far is that produced by Surange though even this is by no means conclusive. Commonly associated with the stems and petioles are sporangia with a multiseriate annulus, some of which contain well-preserved and quite distinc- tive spores. These sporangia are apparently borne on slender, branching axes each of which is traversed by a delicate vascular strand and the epidermis of which is characterized by dark cell contents. These have contracted away from the cell walls in many instances giving them a readily recognized appearance, especially in surface view (Pl. 60, fig. 6). The slender branches are closely associated with small branches of Botryopteris antiqua petioles which also often have epidermal cells with dark contents (PI. 60, fig. 7). The spores contained in the sporangia are triangular in plan and are somewhat flat with trilete rays extending about three-quarter$ of the way towards the margin of each radius. They are smooth-walled and devoid of any ornamentation. The radial extremities are rounded and the interradial margins are concave. In proximal view the spores show a narrow raised margin but when seen from the distal side they show three large equal lobes surrounding a central triangular depression which leads to a flat, partly enclosed, central area (Text-figs. 12a-c). They are identical with those filling the sporangia which Scott (1910) figured and described and which he regarded as possibly those of B. antiqua. Careful and detailed descriptions of palaeozoic spores are a relatively recent feature of palaeobotanical studies and already palynology has produced a considerable specialized literature with regard to them. The earliest reference to spores with a triangular contour which has been traced so far is one by Renault (1896). His figure 17 shows a sporangium termed by him Hymenophyllites 8 in which most of the spores are spherical but in which there are also some spores which are triangular (tvigones). In Renault’s view (1896 : 20) it is probable that this appearance is due to a partial desiccation of the spores. Recent investigations, however, have shown that subtriangular types of spore occur quite commonly among palaeozoic spore assemblages. Among these Kosanke (1950, pl. 3, fig. 4) illustrates a spore of this kind to which he has given the name Granulosporites concavus. This has rounded corners, is smooth-walled and has strongly concave margins between the radii. Mamay (1950) also lists a species, Botryopteris allinoensts, which, like Kosanke’s material, is from the Pennsylvanian of Illinois and in which the spores are described as “ axially flattened, triangular, with concave sides’ (Mamay, 1950, pl. 2, fig. 9). These spores are from more recent geological strata than that in which Botryopteris antiqua occurs but spore assemblages from parts of the Scottish Lower Carboniferous have recently been investigated by Butterworth & Williams (1958) and by Love 376 THE MORPHOLOGY OF BOLRYOPLERTS VANE LOUA (1960). Both these authors have described isolated spores which appear to show a close resemblance to those associated with B. antiqua. These were assigned by Butterworth & Williams to the genus Granulosporites as G. politus but by Love to Leiotriletes politus. Love very kindly loaned to the Department of Palaeontology a “single spore ”’ slide of his Lezotviletes politus for comparison with the spores present in the sporangia from the Calciferous Sandstone and this has established beyond reasonable doubt that this spore and the spores from the petrified sporangia are identical. a b ms mf Fic. 12. a, Proximal and 6 distal appearance of trilete spores ; c, section of spore along plane A, B, of 12a. a, Surange Coll. B.154; 6, Surange Coll. B.7o The evidence for the assigning of these sporangia and spores to Botryopteris antiqua is obviously not conclusive and it is hoped that proof of their relationship will be furnished when suitable specimens are available. Neither Kidston, Benson nor Surange have attempted a formal diagnosis of Botryopteris antiqua and the following is suggested as covering the known data : Botryopteris antiqua Kidston D:aGNosIs. Small palaeozoic fern. Stem radially organized, protostelic with wholly tracheidal xylem, narrow scalariform tracheids forming bulk of stele, often wider tracheids with two or more rows of transversely elongated pits developed peripherally; ribs MOk PE OWOGC YIOR- BOT RY OPE hl Se AlNeh OWA 377 protoxylem spirally thickened ; mesarch : cortex divided into parenchymatous inner and outer zones with sclerenchymatous middle zone ; epidermis, especially at and near stem apex, developing numerous unbranched uniseriate hairs forming a protec- tive mantle over the apex and extending on to bases of petioles. Petioles oval in transverse section, monodesmic with initially mesarch protoxylem becoming endarch ; petioles of two kinds, one possessing a single adaxial protoxylem group, the other possessing two adaxial protoxylem groups; monarch petioles developed in close spiral succession, branching repeatedly, sometimes accompanied by small endarch subsidiary vascular bundle; diarch petioles often developing sporophytic buds giving rise to radially organized daughter plants and, when present, largely replacing monarch petioles. Roots diarch with separate cortex throughout. majority with sclerenchymatous outer cortex. A second type of root with wide, wholly parenchy- matous cortex developed at and near stem apex. Sporangia provisionally assigned to B. antiqua having a multiseriate annulus and containing subtriangular, smooth-walled and trilete spores with rounded angles and concave interradii. LecToTypE.—A transverse section of a dichotomizing stem figured by Kidston (1908, fig. 6) bearing roots and a petiole. Kidston Coll., Glasgow University, slide 549c. DISCUSSION It is almost inevitable that the study of the morphology of any fossil plant is one which shows a progressive series of stages largely dependent on the number and quality of preservation of the specimens available to the investigator and, to a less extent, on the evolution of techniques which enable a better use to be made of those at his disposal. The series of morphological studies of Botryopteris antiqua including the present paper clearly exemplify this. Although the plant was first recognized and briefly described by Kidston the chief value of his paper was that it provided a basis for further study. Benson’s work appreciably amplified our knowledge though the techniques then available led her to misinterpret some of the plant’s morphological characteristics. Surange, applying newer techniques, has been able to correct certain of these misinterpretations and to make appreciable additions to our knowledge though some of his tentative interpretations have oversimplified the morphological problems involved. These three investigators have provided, however, a reasonably secure foundation for further work and it is considered that the present account simply marks an additional stage in the elucidation of the morphology of this palaeozoic fern. The most important feature about which additional information is desirable is definite proof that the associated sporangia are those of Botryopteris antigua. More detailed knowledge of the behaviour of the diarch petioles would be welcome also. Their preservation mainly as comparatively short, isolated segments fails to give a clear idea of their complete morphology. This may be due to their being shed from the parent plant when mature or by their being parted from it by decay of the parts connecting one with the other. The radially organized plants to which they give rise vary considerably in size but this may well be due to the stage of 378 THE MORPHOLOGY OF BOTRYOPTERIS ANDIOUWA development they have reached. They are obviously not terminal in position but the minute stele figured in Pl. 59, fig. 7, which, though fully differentiated, only measures 0:3 mm. in diameter probably developed from the distal part of the parent axis. Quite small petioles with steles having two symmetrically placed adaxial protoxylem groups are not uncommon but whether they represent the initial stage in the branching of a monarch petiole or the distal parts of a diarch petiole cannot be determined from isolated sections. Botryopteris antiqua seems to fall naturally into place taxonomically among the European Coal Measure species of the genus though its petiolar anatomy is somewhat simpler. It shares with them a homogeneous protostele with mesarch protoxylem groups, the apex of which is protected by predominantly uniseriate unbranched hairs and a monodesmic petiolar vascular supply which retains the form of a tan- gentially compressed cylinder. It also shares with Botryopteris hirsuta the specialized type of petiole on which sporophytic plantlets are developed. Whether the genus Botryopteris as at present constituted is a natural one is perhaps open to some doubt. B. forensis and some of the American species such as B. americana (Delevoryas & Morgan, 1954) and B. tvisecta (Mamay & Andrews, 1950) have much in common and show considerable differences from the European Coal Measure species and from B. antiqua. The best solution, if all are to be retained in the same genus, would seem to be the creation of separate subgenera for the two groups. SUMMARY Botryopteris antiqua is a minute fern occurring in the Scottish Lower Carboniferous. The stem, which is radially organized, possesses a stele which has the shape of a shallow dome. It is protostelic with xylem which is wholly tracheidal and in which the protoxylem is mesarch. The xylem is enclosed by a continuous sheath of phloem and by a cortex differentiated into parenchymatous inner and outer zones and a sclerotic middle zone. The stem apex is overarched by a closely set system of un- branched and predominantly uniseriate hairs which extend on to the proximal parts of the petioles. These are of two kinds, monarch with a single adaxial protoxylem and diarch with two adaxial protoxylem groups. The petiolar vascular supply is monodesmic and has the form of a tangentially flattened cylinder with the protoxylem forming either one or two adaxial ridges. The monarch petioles branch freely and are developed in spiral succession. The diarch petioles are generally larger than the monarch ones and branch less frequently. They often produce sporophytic buds and their development in- terrupts the spiral development of the monarch ones. The root system consists of diarch roots of two kinds, the majority having an outer zone of cortical sclerenchyma. Ina second type of root the cortex is wholly parenchy- matous. Sporangia which are frequently associated with the stems and petioles have a multiseriate annulus and contain smooth-walled subtriangular trilete spores with concave interradii. LE MORPHOLOGY O8 BOT RY OPTERITS “AN i LOW A 379 ACKNOWLEDGEMENTS In addition to those whose help has been acknowledged in the body of the paper, the author is also greatly indebted to Dr. E. I. White, Keeper of the Department of Palaeontology, for facilities for studying the palaeobotanical collections in his charge, to Mr. F. M. Wonnacott for his expert help in many ways and to the staff of the Photographic Department of the Museum for their care and co-operation in the production of the negatives and prints from which the plates illustrating this paper have been made. REFERENCES Benson, M. tg911. New observations on Botryopteris antiqua, Kidston. Ann. Bot., London, 25 : 1045-1057, pls. 81-83. BERTRAND, P. t1o11. L’étude anatomique des Fougéres anciennes et les problemes quelle souléve. Progr. Rei. Bot., Jena, 4: 182-302, 59 figs. Boopte, L. A. 1900. Comparative Anatomy of the Hymenophyllaceae, Schizaeaceae and Gleich- eniaceae, I. Ann. Bot., London, 14: 455-496, pls. 25-27. Bower, F.O. 1923. The Ferns (Filicales), I. Analytical Examination of the Criteria of Comparison. 359 pp., 309 figs. Cambridge. 1926. The Ferns (Filicales), II. The Euspovangiatae and other relatively primitive Ferns. 344 pp., figs. 310-580. Cambridge. Butterworth, M. A. & Wirttams, R. W. 1958. The Small Spore Floras of Coals in the Lime- stone Coal Group and Upper Limestone Group of the Lower Carboniferous of Scotland. Trans. Roy. Soc. Edinb., 63 : 353-392, pls. 1-4. Corsin, P. 1937. Contribution a l'étude des Fougeéves anciennes du Groupe des Inversicatenales. xl + 247 pp., 42 pls. Lille. DELEVoryYAS, T. & MorGan, J. 1954. Observations on Petiolar Branching and Foliage of an American Botryopteris. Amer. Midl. Nat., Notre Dame, Ind., 52 : 374-387, 22 figs. Gorpon, W. T. 1911. On the Structure and Affinities of Diplolabis vomeri (Solms). Trans. Roy. Soc. Edinb., 47 : 711-736, pls. 1-4. Hooker, W. J. 1854. A Century of Ferns. viipp., 100 pls. London. Kipston, R. 1908. On a new species of Dineuvon and of Botryopteris from Pettycur, Fife. Trans. Roy. Soc. Edinb., 46 : 361-364, I pl. KosankE, R. M. 1950. Pennsylvanian Spores of Illinois and their use in Correlation. Bull. Til. Geol. Surv., Urbana, 74 : 1-128, pls. 1-16. Kupper, W. 1906. Uber Knospenbildung an Farnblattern. Flora, Jena, 96 : 337-408, 47 figs. LacHMANN, J. P. 1885. Recherches sur la morphologie et l’anatomie des Fougeres. C.R. Acad. Sct. Paris, 101 : 603-606. -— 1889. Contributions a 1’Histoire naturelle de la Racine des Fougeres. Ann. Soc. bot. Lyon, 1889 : 1-189, pls. I-5. Love, L. G. 1960. Assemblages of small spores from the Lower Oil-shales Group of Scotland. Proc. Roy. Soc. Edinb. (B) 67 : 99-126, pls. I, 2. Mamay, S. H. 1950. Some American Carboniferous Fern Fructifications. Ann. Mo. bot. Gdn., St. Louis, 37 : 409-476, pls. 1-9. Mamay, S. H. & ANDREws, H. N. 1950. A Contribution to our knowledge of the Anatomy of Botryopteris. Bull. Torrey bot. Cl., New York, 77 : 462-494, pls. 1-7. OcuRA, Y. 1938. Anatomie der Vegetationsorgane dey Pteridophyten. In Handbuch dev Pflan- zenanatomie (II) 7, 2. vili + 476 pp., 382 figs. Berlin. PELOURDE, F. 1910. Observations sur quelques végétaux fossiles de l’Autunois. Ann. Sci. Nat. Paris (9) 2 : 361-371, 9 figs. PotoniE£, H. 1903. Zur Physiologie und Morphologie der fossilen Farn-Aphlebien. Ber. dtsch. bot. Ges., Berlin, 21 : 152-165, pl. 8. 380 THE MORPHOLOGY OF BOTRYOPTERTS ANLIOUA RENAULT, B, 1896. Bassin Houilley et Permian d’Autun et d’Epinac, IV. Flore Fossile, 2. 578 pp., atlas, pls. 28-89. [Etudes Gites Min. France.] Paris. SAHNI, B. 1916. Vascular Anatomy of the Tubers of Nephrolepis. New Phytol., Cambridge, 15 : 72-80, 3 figs. Scott, D. H. 1910. Sporangia attributed to Botryopteris antiqua, Kidston. Ann. Bot., London, 24 : 819-820, I fig. —— 1920. Studies in Fossil Botany, 1. Plevidophyta. 3rd edit. xxiii + 434 pp., 190 figs. London. SEWARD, A.C. 1910. Fossil Plants, 2. xxi + 624 pp., 265 figs. Cambridge. Sorms Lauspacn, H. GRAF zu. 1891. Fossil Botany. English edit. xi + 401 pp., 49 figs. Oxford. SPERLICH, A. 1906. Erganzungen zur Morphologie und Anatomie der Auslaugfer von Nephrolepis. Flora, Jena, 96 : 451-473, pls. 3, 4. SURANGE, K. R. 1952. The Morphology of Botrvyopteris antiqua with some observations on Botryopteris vamosa. Palaeobotamst, Lucknow, 1: 420-434, pls. 1-4. TansLey, A. G. 1908. Lectures on the Evolution of the Filicinean Vascular System. New Phytologist Reprint No. 2: 141 pp., tor figs. Cambridge. ” PLATE 56 Botryopteris antiqua Kidston Fic. 1. Transverse section of a stem showing stages in the emission of petioles and root traces. Note the sclerized outer cortex of the roots and, at the right bottom corner of the figure, one of Benson’s “ aphlebiae’’. B.M. (N.H.) Holden Coll., No. JBI, Seis Fic. 2. Transverse section of stem in which the whole of the cortical tissues are preserved and which shows the emission of three roots and a dividing petiole trace with mesarch protoxy- lems. Gordon Coll., King’s College, London. x 12. Fic. 3. Oblique section of a dividing diarch petiole, the smaller branch of which is associated with the basal root-bearing part of a daughter sporophyte. Note the increase in the number of tracheids where the root steles join the parent stele. B.M. (N.H.) Holden Coll., No. H.152. x12. Fic. 4. Oblique section of the same plant as that shown in Fig. 3 but cut at'a lower level and in which the daughter stem is well shown. B.M. (N.H.) Holden Coll., No. H.162. x12. Bull. B.M. (N.H.) Geol. 5, 9 PLATE 56 BOTRYOPTERIS PLATE 57 Botryopteris antiqua Kidston Fic. 1. Transverse section of a stem with a large mesarch petiole trace with two protoxylem groups. The cortical tissues are well preserved and the small-celled epidermis is clearly shown on the lower side of the figure. Gordon Coll., King’s College, London. x to. Fics. 2, 3. Transverse sections showing two stages in the emission of a diarch petiole. Fig. 2, B.M. (N.H.), Oliver Coll., No. 1263 ; Fig. 3, Oliver Coll., No. 1264. x Io. Fics. 4-6. Sections showing three stages in the emission of a branch trace from that of a diarch petiole. B.M. (N.H.) Oliver Coll., Nos. 1261, 1262, 1238 respectively. All «15 approx. PIPANE 57 Bull. B.M. (N.H.) Geol. 5, 9 BOTRYOPTERIS PLATE 58 Botryopteris antiqua Kidston Fic. 1. Longitudinal section of part of the xylem of a petiole showing the types of tra- cheidal pitting and the disruption of the protoxylem. Glasgow Univ. Figd. Slide Coll., No. 714. X 150. Fic. 2. Longitudinal section through the protoxylem of a petiole showing an early stage in the breakdown of the spirally thickened tracheids. B.M. (N.H.) Holden Coll., No. H.162. x70. Fic. 3. Transverse section of stem giving off a root trace and with a large petiole trace with two protoxylem groups to the left of the figure. Gordon Coll., King’s College, London. x 16. Fic. 4. Transverse section of a petiole showing a strip of sieve tubes on the abaxial side of the stele. Gordon Coll., King’s College, London. x 35. Fic. 5. Part of a petiole in both transverse and longitudinal section showing the cortical secretory cells. Surange Coll., No. B. 34. X20. Fics. 6-8. Transverse sections of successive stages in the emission of a petiole trace and a subsidiary daughter trace. B.M. (N.H.) No. V.35555; Fig. 6, slide 73b; Fig. 7, slide 826; Fig. 8, slide 86b. All x20. PEAS Bull. B.M. (N.H.) Geol. 5, 9 WL HAN 7 BOTRYOPTERIS PLATE 59 Botryopteris antiqua Kidston Fic. 1. Transverse section of a stem with attached diarch petiole to the right and showing an early stage in the formation of a second petiole trace to the left. Note the character of the stem cortex. B.M. (N.H.) Holden Coll., No. H.1o1. x 16. Fic. 2. The same stem cut partly in transverse and partly in longitudinal section showing the junction of the petiole trace with the stem stele. B.M. (N.H.) Holden Coll., No. H.115. x16. Fic. 3. Section through a stem showing a circinately coiled petiole and, to the right, parts of surface hairs. The small ringed group of xylem is much more highly magnified in Fig. 4. Surange Coll., No. B.78. x 20. Fic. 4. Small group of tracheids showing a spiral protoxylem and a pitted metaxylem element. Surange Coll., No. B.78. x 700. Fic. 5. Transverse section of an immature root still embedded in the stem cortex. Note the diarch xylem plate and the cortical cells with dark contents. Kidston Coll., Glasgow University, No. 670¢. X 100. Fic. 6. Transverse section of a stem stele near its apex showing incomplete differentiation of the xylem. Surange Coll., No. B.34. x50. Fic. 7. Transverse section of the stele of a very small stem in which the differentiation of the xylem is complete. B.M. (N.H.) No. V.35555, slide 3b. X00. Bull. B.M. (N.H.) Geol. 5, 9 PLATE 59 BOTRYOPTERIS PLATE 60 Botryopteris antiqua Kidston Fics. 1-3. Successive stages in the differentiation of a second type of root. Note the dense contents of the cortical cells and the traces of surrounding hairs in Figs. 1, 2. B.M. (N.H.) No. V.35555- Fig. 1, slide 58b x100; Fig. 2, slide 65b x100; Fig. 3, slide 44b x 100. Fic. 4. Transverse section of a petiole from Autun showing the characteristic adaxial concavity. Kidston Coll., Glasgow University, No. 2046. X25. Fic. 5. Transverse section of a branching petiole giving off two approximately equal daughter traces. Kidston Coll., Glasgow University, No. 1471. X40. Fic. 6. Minute sporangium-bearing branches. Note the dark cell contents of the epidermal cells and the subtriangular spore in the sporangium. Surange Coll., No. B.70. X roo. Fic. 7. Tangential longitudinal section of a small branch giving off a still smaller branch. Note the dark epidermal cell contents. Surange Coll., No. B.33. x 8o. PLATE 60 Bull. B.M. (N.H.) Geol. 5, 9 BOTRYOPTERIS INDEX TO: VOLUME. V New taxonomic names and the page numbers of the principal references are printed in Clarendon type. An asterisk (*) indicates a figure. Abietineae 22, 99 Acacia 123 catechu 150 Acanthaceae 100, 150 Acanthodian scales 202, 203, 220-229, 232, 235-239, 261 spines 226, 227, 230-232, 285 tooth Pl. 45, fig. 3 Acanthodii 261, 284 Acanthus sp. 96, 100, 150; Pl. 30, figs 144, 145. Acrostichum 95, 98, 101 lanzaeanum 15, 94, 99, 101; PI. 24, fig. 2 Actinidia crassisperma 97, 140 sp. 96, 100, 135; Pl. 28, fig. 98 Adapidae 52, 67 Adapis 46, 47, 54 parisiensis 47; Pl. 14 Adinandra 137, 140 Afrocrania 145 Agnatha 245, 268 Albizzia 123 Albulidae 9 Aldrovanda 119 ovata 97, 99, 101, 119; Pl. 26, figs. 53-55 Althaspis 280 Altingia 121 Altingieae 121 Alum Bay 15 American Museum of Natural History 303 Ampelopsis votundata 100, 134 Anacardiaceae 99, 127 Genus ? 127; Pl. 27, figs. 72, 73 Anaptomorphidae 47, 53, 54, 67 Anaspida 236, 258 Andromedeae 100, 146 Genus sp. I 146 sp. 2 146; Pl. 30, fig. 138 Angiospermae 21, 28, 99, 103 Anglaspis 201, 202 maccullought 223, 246 Sp220 222225, 229) 230,231,235 Ankyropteris corrugata 364 Anneslea 138-140 costata 96, 100, 139, 140; PI. 29, figs. 113-115 Antiarchi 301 Apocynaceae 154 Araceae 97, 99, 138 Genus ? 99 Araucarineae 23 Ayaucarites 16, 20, 23, 26-28, 32 gurnardi 27; Pl. 4, fig. 6 selseyensis 18, 19, 23, 27, 28; Pl. 4, figs. 1-5, Pls. 5, 6 sp. 161 Archidiskodon planifrons 76 Arctolepids 222, 224, 228, 232, 233, 235, 238, 293, 300* ; Pl. 46, fig. 6 Arne, Dorset 15 Arthrodira 220, 265, 289 Artocarpoideae 113 Asplenium lauterbachii 370 mannit 370, 371* 372* obtusilobum 370, 371* Astevocevas obtusum 161 Atherinidae 9 Atherion 9 Atriplex sp. 98 Australopithecus 87 Bagshot Beds 15 BAKER, R. 308 BALL, H. W. 175-242 - Bate, D. M. A. 74 Becktonia 113 hantonensis 96, 99, 113; Pl. 25, fig. 39 Belgicaspis 256 croucht 178, 193, 200, 202, 203, 223-225, 227, 229, 230, 232, 234-236, 252, 256, 257, 280; Pl. 36, figs. 2-11 var. heightingtonensis 203, 220, 222, 235, 256 Bennettitales 170, 171 Benneviaspis 282 anglica 283 lankesteri 283 salopiensis 194, 203, 204, 220, 282, 283* Beta 115 Bethlehem 73; Pl. 15 Betulaceae 99, IIo Birmingham University 239 Boraginaceae 100, 149 Bothriolepis 30% hydrophila 301*, 302 385 INDEX Bothrioletis—contd. macrocephala 208, 301, 302 sp. 208, 233 Bothrocaryum 145 Botryopteris 378 americana 378 antiqua 361-378 forensis 378 hirsuta 378 illinoensis 375 trisecta 378 Bournemouth, Hants 15 Freshwater Beds 15 Marine Series 15 Bovey Tracey, Devon 110, 111 BRADLEY, O. 363 Brasenia 94, 95, 101, 116 antiqua 99, 117 oblonga 99, 118; Pl. 25, figs. 44-46 ovula 99, tot, 116 schveberi 117 spinosa 99, 117, 118 Broussonetia 111 rugosa 99, 111; Pl. 25, figs. 36, 37 Brown Clee Hill, Shropshire 175-310 Bryantolepis 298 Burseraceae 98, 99, 127 Caenopithecus 58 Cakeham Beds 17, 18, 33, 34, 38 Campanile Bed 20 Campylospermum 137, 138 hordwellense 97, 100, 187; Pl. 28, figs. 104-107 Caprifoliaceae 100, 150 Caricoidea 16, 33, 94, 97, 105 minima 99,105; Pl. 24, figs. 18-21 obscura 20, 33, 96, 106; PI. 7, fig. 34; Pl. 24, figs. 22-24 Caricoideae 97, 106 Carpinus 110 boveyanus 96, 99, 110; PI. 25, fig. 35 Carpolithus apocyniformis 100, 153; Pl. 30, figs. 147, 148 boveyanus 110 curryi 19, 39, 40; Pl. 11, figs. 70, 71 dixoni 36 fibrosus 100, 152 gardneri 97, 100, 154 headonensis 28 Sp.I 100, 155 Sp.2 97, 142, 143 SP-3 97, 144, 145 Sp.4 97, 144 sp.5 98 sp.6 100 Sp.7 152 sp.8 100, 154 spp. 19, 40, 100, 154, 155; PI. 11, figs. 72, 73; Pl. 30, figs. 149-159 Caryophyllaceae 97, 99, 114-116 Caton-TuHompson, G. 73 Centrospermae 114, 115 Cephalaspids 222, 225-229, 231, 232, 235, 237, 239, 261, 284 Cephalaspis 201-203, 247, 259 acutivostris 238, 259 agassizi 221, 259 bouldonensis 221, 259, 260* ; Pl. 36, fig. 1 fletti 221, 238, 259 langi 221, 259, 261 lankesteri 238, 259 SP. 223, 225, 232, 238, 284 whitbatchensis 238, 259 whitet 238, 259 CHANDLER, M. E. J. 13-41, 91-158 Chara 94 Bed 98 Charmouth, Dorset 161 Charophyta 98 CuasE, N. C. 370 Cheilanthoid ferns 361 Chenopodiaceae 115 CHEsSTERS, K. I. M. 93 Chlovophora bicarinata 99 Chrysodium lanzaeanum tor Cichlidae 10, 11 Cinnamomum 119 Cladium 97, 106, 107 minimum 97, 105, 106 CLarRK, J. D. 71-90 CLARK, W. E. LE Gros 46, 67 CLAUSEN, H. S. 11 Clethraceae 147 Cleyera 139, 140 stigmosa 96, 100, 138, 189; PI. 28, figs. 108-112 variabilis 139 Climatius 262, 288 Sp. 234 Clupeidae 9 Clupeoidei 9 Cocculineae 118 Colwell, I.o.W. 98 Coniferae 170 Coniferales 22 Coniferous wood 20 Cooksonia 194, 205 Sp. 223 Cornaceae 97, 100, 141, 145 Cornoideae 100, 141, 144 Cornus 97, 141, 144, 145 mas 145 quadvilocularis 96, 97, 100, 144; Pl. 29, figs. 129-133 volkensit 145 Corvaspis 201, 202, 217, 240 kingi 223, 225, 246 sp. 226, 228 Corydalis 97, 115 pulchra 97, 114 Crocodilus hastingsiae 98 INDEX 386 CrortT, W. N. 239 Crossopterygii 233, 302 Crustacean tracks 204 Ctenacanthus 289 Ctenopteris 169, 170 cycadea 161, 169, 170 leckenbyi 170 sarvant 168, 169 wolfiana 170 Ctenozamites 161, 169-171 bergeri 169 cycadea 161-173 ; Pls. 31, 32 leckenbyi 170 savvani 170 wolfiana 170 Cucumites variabilis 36 Cucurbitaceae 95, 100, 151 Genus ? I51 Cucurbitospermum 151, 152 reidii 100, 151; Pl. 30, fig. 146 Cupressinites liasinus 162 Curry, D. 16-20, 31, 34-36, 39 Cycadales 170, 171 Cycadeoidea 171 gracilis 162 pygmaea 162 Cycadites vectangularis 162 Cycadolepis sp. 101 Cycadophyte 171 Cycadopteris anglica 101 Cymadocea 31 Cymripteraspis 279, 280 leachi 178, 189, 194, 195, 203, 204, 220, 224, 228, 229, 232-235, 237, 252, 268, 276, 279, 280, 281*, 289; Pl. 45, figs. 1, 2 Cyperaceae 29, 32, 33, 99, 105 Genus ? 99, 107; Pl. 24, fig. 27 Cyperites 29 forbesti 28, 29 Daubentonia 47, 57 Davis, A. G. 16, 19, 20 Dendrogale 50 Denticeps 4-10 clupeoides 4, 5*, 8, 9 Denticipitid fishes 1-11 Denticipitidae 3, 4, 9, 10 Dicotyledones 35, 99, 110 Dictyophyllum nilssoni 169 Didymaspis 201 Dilleniaceae 100, 135 DINELEY, D. L. 175-242 Diospyros 147 / antiqua 147 headonensis 100, 147 Diplolabis 365 Dipnoi 307 Ditton Series 188, 202, 245 Doliostrobus sp. Pl. 4, fig. 6. sternbergii 27 ¢ DoROFEEV, P. I. 103 Dorset Pipe-clay Series 15 Downton Series 181, 201 Dracontocarya glandulosa 127, 128 Droseraceae 97, 99, 119 Dunstania, 97, 141, 144, 145 glandulosa 100, 144, 145; Pl. 29, figs. 125-128 multilocularis 145 Ebenaceae 100, 147 Elephant Pit, Bethlehem 87; Pls. 16, 17 Elopidae 9 Emys 98 Eoliquidambar 120, 121 hordwellensis 96, 99, 120, 122; Pl. 26, figs. 56-59 Eomastixia 141 bilocularis 97 vugosa 97, 100, 141, 142 Epacridaceae 97, 100, 146, 147 Epacridicarpum 97, 146 headonense 100, 146,; Pl. 20, figs. 134-137 mudense 146 Equisetites muensteri 161 sp. 161 Ericaceae 100, 146, 147 Genus ? 97 Genus ? sp.2 146 Ericales 147 Euphorbiaceae 36, 37 Europrotaspis 268, 270 arnelli 268-270, 275, 279; Pl. 44, fig. 6 crenulata 194, 203, 204, 220, 221, 223, 224, 228, 232, 233, 235, 237, 268, 269*, 270, 271*— 278*, 279, 280, 289; Pl. 43; PI. 44, figs. 1-5, 7 Eurya 136, 137, 140 becktonensis 96, 100, 136, 137, 138; Pl. 28, fig. 103 japonica 137 var. nitida 137 stigmosa 139 Eurypterids 204 Eusthenodon 305 Eusthenopteron 305, 307 dalgleisiensis 307 farloviensis 208, 233, 305, 306* foordi 306*, 307 sdive-soderberghi 307 SP. 307 tvaquaiyl 307 wengukovi 307 Exeter University 239 Faboidea crassicutis 130 Fagara 124 Farlow Sandstone Series 205, 301 Ficus 111 Filicales 99 Filicites agardhiana 169 cycadea 169 387 INDEX Fisher Beds 18, 19, 31, 34, 36, 37, 40 Fistulipora sp. 197 Forlisia lanzaeana 101 Frangula 132 hordwellensis 96, 97, 100, 181; Pl. 27, figs. 85, 86; Pl. 28, fig. 87 Fumariaceae 116 Galeopithecus 51 Gall 156; Pl. 30, fig. 160 GARDNER, E. W. 73, 74 GaRn_r, L. T. 18, 34 GaRNER, S. D. 19, 34 Geological Survey Museum 29, 102, 162, 217 Ginkgo 170 Glasgow University 362, 365, 366 Glyptolepis 302, 303 leptopterus 303, 305 Gnathostomata 261, 284, 301 Gomphodus 202, 262 GoopDcHILD, M. 18 Gordonia 135, 136 lastanthus 136 minima 100, 185, 136; Pl. 28, figs. 99-102 tvuncata 100, 136 Gramineae 32 Grammitis 169 Granulosporites concavus 375 politus 376 GREENWOOD, P. H. 1-11 GROVES, J. Ior GuRR, P. 308 Gymnospermae 22, 99, 102 Halle University 63, 66 Hamamelidaceae 99, 120-122 Hampshire Basin 15 Hantsia 97, 114 glabra 96, 99, 115; Pl. 25, figs. 41-43 pulchva 99, 114*, 115, 116; Pl. 25, fig. 4o Haplochromis to, 11 bloyett 10 wingali 10 Harris, T. M. 159-173 Harrisia marsilioides 171 Harvisothecium 17% Hemiacodon 66 Hemicyclaspis murchisoni 215 Hemitelia capensis 367 Heterostraci 245, 268 Hipparion 76 Hippurites 161 Ho pen, H. S. 359-380 Holoptychiidae 302 Holoptychius 303 sp. 208, 233, 238 Ho.ittvum, R. E. 169 Hordle, Hants. 15, 93, 94 Hordwellia 97, 140 crassisperma too, 140 Hydrocharitaceae 99, 104 Hydvopteridangium 171 marsilioides 17% Hymenophyllites 375 Hymenophyllum sericeum 370 tunbridgense 370 Icacinaceae 95, 98, 100, 128, 130 Icacinicarya 130 bartonensis 131 becktonensis 96, 100, 1380; PI. 27, figs. 83, 84 pygmaea 131 veticulata 130 transversalis 96, 100, 180; Pl. 27, figs. 80-82 Ichthyodorulite 224 Ichthyostegopsis wimani 305 Indri 58 Insectivora 54 Todes 129 hordwellensis 96, roo, 129*, 130; Pl. 27, figs. 78, 79 sp. 100, 128; Pl. 27, figs. 76, 77 Ischnacanthus 201, 202, 262, 285 anglicus 204, 220, 285; Pl. 45, fig. 4 gracilis 263-265, 286 Kingi 262, 263*, 264, 265; Pl. 42, fig. 4 wickhami 201, 225, 264* ; Pl. 42, figs. 1-3 Isle of Wight 15, 93 Isospondyli 3, 10 Israel 71-90 Jackson, J. F. 161, 162 Jepson, G. L. 67 Kafuan Culture (Africa) 83 Kallostrakon 187, 188, 200, 201, 245 SP. 224, 236, 245 Kew Herbarium 150 King’s College, London 239, 363, 365 KRUMBIEGEL, G. 67 Kujdanowiaspis 203, 267, 293*, 297, 298, 300 anglica 203, 204, 220, 225, 228, 231, 233-235, 265, 266*, 267*, 268, 289, 291*, 292, 301; Pl. 45, fig. 7 Spi 202)1222)1235,237 willsi 204, 220, 290, 291*, 292 ; Pl. 46, figs. 1, 2 Lagenaria 152 Lamellibranchs 218 Lanarkia 284 Lauraceae 20, 35, 36, 96, 99 Genus ? 99, 119; Pl. 25, figs. 51, 52 Laurocalyx 16, 35 sp. 19, 85; Pl. 10, figs. 54, 55 Leguminosae 96, 99, 122, 123 Genus ? 99, 122; Pl. 26, figs. 61, 62 LreumMaAN, J. P. 67 Leiotriletes politus 376 Lemuridae 52 Lemuriformes 47, 51, 52 Lemuroidea 47, 51 INDEX 388 Lepidosteus 98 Leptobos 76 Liassic floras (British) 161 Limnocarpus 15, 16, 20, 21, 28-32, 94, 95, IOI, 104 cooperi 21 enormis 19, 29; Pl. 7, figs. 25-33 forbesi 16, 19, 20, 21, 28, 29, 30, 97, 99, Ior, 104; Pl. 7, figs. 20-24; Pl. 24, figs. 7-11 headonensis 16, 19-21, 28, 29, 97, 104 Lindera 119 Liquidambar 121, 122 sp. 96, 97, 121 Litsea 119 London Clay 17 Lone, A. G. 363 Loris 50 Lorisiformes 50-52 Lower Carboniferous 208 Lycopodites squamatus 23, 26 Lyme Regis, Dorset 161, 162, 168, 170 Lythraceae 100, 141 Macrocarpum 145 Mahenge, Tanganyika 3 Malania 270 MANNARD, G. W. 3, 9, II Mapania 107 Mastixia 97, 141-143 glandulosa 97, 100, 142, 143; Pl. 20, figs. 118— 121 Mastixicarpum 96, 141, 142 crassum 100, 142; Pl. 30, fig. 161 Mastixioideae 95-98, 100, 141, 143, 144 Genus ? 98, 100, 141, 143; Pl. 20, figs. 122-124 MEADE, M. J. 4 Medullosa 170 Megaladapis 49, 50 Megalichthys 305 Melania 101 Meliosma sp. 96, 100, 133; Pl. 28, figs. 88, 89 Menispermaceae 99, 118 Menispermum obliquatum 97, 118 Microchoerinae 54, 56 Microchoerus 45, 46, 54-59, 64, 65 edwardsi 53, 59, 61 evinaceus 56, 59, 60*, 61 ornatus 56 Microdiptera 14% parva 96, 100, 141; Pl. 20, figs. 116, 117 Modiolopsis sp. 204, 222 Monocotyledones 21, 28, 32, 40, 99, 103 Moraceae 99, III, 113 Moroidea boveyana 112, 113 hordwellensis 96, 99, 112* ; Pl. 25, fig. 38 Moroideae 112, 113 Morus 111 Mugiloidei 9 Myliobates 98 Myrica boveyana 96, 99, 110; Pl. 25, figs. 33, 34 Myricaceae 99, 110 Myrtaceae 138 Nannopithex 45, 46, 53-58, 61-64, 65, 67 filholi 62, 65, 66 pollicaris 62, 64, 66 vaabi 61, 63*—65, 66 Natsiatum 128 eocenicum 96, 100, 128, 129 sp. 100, 128; Pl. 27, figs. 76, 77 Necrolemur 46, 47, 49-58, 61-67 antiquus 45, 46, 48*, 53, 61, 62, 64, 66; Pls. 12S filholi 62, 64 parvulus 54 vaabi 61, 62 zitteli 56, 62, 64 Necrolemurinae 45, 47, 52-59, 66, 67 Nelumbium buchii too, 155 Nematothallus 194, 205 Sp. 223 Nephrolepis cordifolia 370 exaltata 370 neglecta 370 Nigeria 10 Nipa 16, 18, 20, 33, 34, 95, 98 burtini 18, 19, 88, 34, 99; Pl. 9, fig. 49; Pl. 10, figs. 51-53 sp. 99 Nipaceae 33, 99 Nipadites burtini 33 Nodocosta 289 Nodonchus 288 bambusifer 204, 220, 287*, 288 Nostolepis 265 Notharctus 47 Nummulites 16 laevigatus 19, 34 planulatus 16, 17 variolarius 19 Nuphar ovatum 97, 119 Nycticebus 49, 50 Nymphaeaceae 99, I16 OAKLEY, K. P. 67, 73 Odontopteris 161, 171% bergeri 170 cycadea 169 OGrEy Ja). Lo; 23 Olacineae 149 Old Red Sandstone 175-310 Clee Series 195 Ditton Series 188, 202 Downton Series 181, 201 Farlow Sandstone Series 205 Olea 98, 149 headonensis 96, 98, 100, 149 ; Pl. 30, figs. 141- 143 Oleaceae 100, 149 Genus ? 97, 149 Omomyidae 67 389 INDEX Omomys 67 Omphalodes 149 platycarpa 98, 100, 149 Onchus 201, 202, 262, 286, 288 besomensis 220, 286, 287* maior 286 overathensis 286 SP. 223 wheathillensis 204, 220, 286; Pl. 45, fig. 5 Onychodus 201, 236 Orites sp. 98, 154 Osteodontokeratic culture 87 Osteoglossidae 9 Osteoglossoidei 9 Osteostraci 259, 282 Ostrea teneva Bed 19 Otozamites bechei 162, 168 obtusus 168 Overtonaspis 298, 301 billballi 204, 237, 293*, 294; Pl. 45, fig. 6 Pachytheca 194, 205, 220 SP. 221-223, 225-227, 231, 233 Pagiophyllum peregrinum 162 sewardi 162 Palaeobursera lakensis 96, 99, 127; Pl. 27, figs. 74, 75 Palaeodenticeps 6, 7, 8, 10 tanganikae 3, 6, 7*, 9; Pls. 1-3 Palaeosinomenium 97, 118 obliquatum 99, 118; Pl’ 25, figs. 47-50 pulchrum 118 Palate Bed 18 Palestine Archaeological Museum 87 PAaLLot, J. 98 Palm wood 20 Paludina 94, tor Papaveraceae 97 Parapithecus 58 Paris Museum 63 Paromomyidae 67 Parthenocissus henryi 134 hordwellensis 100, 133, 134; Pl. 28, figs. 90-95 Sp. 133 PATTERSON, B. 67 Pelmatochromis 11 Pelycodus 58 Perodicticus 47 Phaneropleuron 301, 307 Phellodendyon costatum 125, 126 Phenacolemur 58 Phenacolemurinae 67 Phlyctaenaspis 2098 Pinites bowerbanki 22 dixoni 19, 22, 2 Pinus 20, 22, 23 bowerbanki 19, 22, 23; Pl. 9, fig. 50 dixoni 22 Sp. 99 Plagiogyria pycnophylla 370 Platyrrhini 49 Plectvodus 220, 262, 284 Plesiadapis 51, 57 Podocarpus eocenica 102, 103 sp. 102 Polypodiaceae 99, 101 Poraspis 201, 202, 246 SeviCeA 202 SP. 219, 221, 226, 229, 231, 246 Portulacaceae 115 Posidonia 20, 31 Potamogeton 30, 103 pygmaeus 99, 101, 103; Pl. 24, figs. 3-6 spivillus 103 Potamogetonaceae 19-21, 28, 29, 31, 32, 99, 103 Genus? 30, 31; Pl. 7, figs. 35-37; Pl. 8, figs. 38-48 Prescottaspis 294, 300 dineleyi 204, 233, 296*, 297; Pl. 46, figs. 3, 4 Promicrocevas planicostata 161 Pronycticebus 54 Propithecus 58 Prosimii 57 Prosphymaspis 208 Protaspis 268 bucheri 268 (Euryprotaspis) arnelli 268-270, 275, 279; Pl. 44, fig. 6 crenulata 203, 204, 220, 221, 223, 224, 228, 232, 233, 235, 237, 268, 269*, 270, 280, 289; Pl. 43, Pl. 44, figs. 1-5, 7 Proteurya 137 Protoadapis 58 Protoaltingia hantonensis 96, 97, 99, 121, 122; Pl. 26, fig. 60 — Protoptevaspis 258 Prototaxites 194, 205, 223 Pseudoctenis 170 Pseudoloris 45, 46, 54-58, 61, 64 abderhaldini 61, 62, 65, 66 parvulus 50, 61, 62, 63*, 64 Pseudolorisinae 54, 56 Pseudopteraspis 280 Pseudosauripterus 302, 303, 305 anglicus 208, 233, 303*, 304*, 305; Pl. 47 Pteraspis (Belgicaspis) crouchi 178, 193, 200, 202, 203, 223-225, 227, 220, 230, 232, 234-236, 252, 256, 257, 280; Pl. 36, figs. 2-11 var. heightingtonensis 203, 220, 222, 235, 256 cayman 254 (Cymripteraspis) leachi 178, 189, 194, 195, 203, 204, 220, 224, 228, 229, 232-235, 237, 252, 268, 276, 279, 280, 281*, 289; Pl. 45, figs. 1, 2 Sp. 223, 231 gosseleti 257 primaeva 257 (Ptevaspis) dairydinglensis 193, 203, 222, 223, 249, 250*, 252, 253*, 255, 256; Pl. 33; Pl. 34, figs. 1-4; Pl. 35, figs. 4,5; Pls, 37-41 INDEX 390 Pteraspis—contd, vostvata 192, 193, 201-203, 226, 230, 233, 238, 247, 249, 251, 252, 255 var. toombsi 276, 251, 254, 255 var. trimpleyensis 192, 193, 202, 221, 227, 228, 232, 236, 248*, 249, 254; Pl 34, fig. 5; PIn35, tgs. 0, 2 var. waynensis 203, 226, 248*, 240 (Rhinopteraspis) dunensis 189, 252, 279, 280, 281*, 282 var. leachi 280 (Simopteraspis) leathensis 178, 187, 192, 200- 202, 219, 222, 225-227, 229, 231, 239, 257, 258, 277; Pl. 35, fig. 3 Sp. 219, 221, 225-228, 232-234, 237, 239, 247 Pteria contorta 169 Pteridophyta 1or Ptilozamites 161, 170, 171 nilssont 170, 171 Pye, N. A. 20 Pyrolaceae 147 REID, C. 94, 95, 101 Rep, E. M. 94 Retinomastixia 97, 143 schultei 143 Rhamnaceae 97, 100, 131 Genus ? sp.2 97, 131 Rhamnospermum 94, 152 bilobatum 97, 100, 152 Rhamnus 131, 132 frangula 132 purshiana 132 Rhinopteraspis 280 dunensis 189, 252, 279, 280, 281*, 282 Rhizodontidae 303-305 Rhododendroideae 100 Romer, A. S. 67 Rosaceae 99, 122 Roseberry Topping, Yorkshire 169 Royal Holloway College 363 Rubus 95, 122 aculiformis 99, 122 Ruppia 30 Rusinga Island 11 Rutaceae 99, 123, 125 Rutaspermum 97, 126 ornatum 99, 126; PI. 26, fig. 068; Pl. 27, figs. 69-71 Sabiaceae 100, 133 Salvinia 95, 102 hantonensis 97, 102 mildeana 97, 99, 102; Pl. 24, fig. 1 sp. 102 Salviniaceae 99, 102 Sambucus 95, 150, 151 pavvulus 100, 150 Saponaria 115 Sauripteris taylori 303; Pl. 48 Sauripterus anglicus 301, 302, 303* SCHAEFFER, B. 303 Scirpus lakensis 18, 33 Sclevia hordwellensis 96, 99, 107; Pl. 24, figs. 25, 26 laevis 107 Sedgwick Museum, Cambridge 19, 56, 50, 95, 96, 101 Selsey, Sussex 15, 17, 18, 22 Sequoia 94, 95, 103 couttsiae 95, 99, 103 SHAW, S. H. 75, 87-89 SHUTE, C. H. 363 Sicydium 152 Silene 115 Simons, E. L. 43-69 Simopteraspis 257 leathensis 178, 187, 192, 200-202, 219, 222, 225-227, 229, 231, 239, 257, 208, 277; Pl. 35, fig. 3 Sinanthropus 83 Sinomenium 118 SMITH, J. D. D. 217 Spivematospermum 94, 95, 98, 102, 108-110 headonense 95, 98, 99, 108, to9, 110; Pl. 24, figs. 28-30; Pl. 25, figs. 31, 32 wetzleyt 108-110 Spirorbis 204, 218, 220, 269*, 271; Pl. 43, fig. 3 Spondieae 99, 128 Genus ? 99, 127; Pl. 27, figs. 72, 73 STEKELIS, M. 74 Stenopteris 162 Strvatiotes 94, 95, LOI, 104 hantonensis 96, 99, 104, 105; Pl. 24 figs. 12-16 headonensis 99, 101, 104, 105; Pl. 24, fig. 17 Strobilites elongata 162 Studland, Dorset 15 Stypheliae 147 Styracaceae 100, 148 Styvax elegans 100, 148 Symplocaceae 100, 148 Symplocoides glandulosa 97, 144, 145 Symplocos 145, 148 headonensts 100, 148; Pl. 30, figs. 139, 140 sp. 100, 148 sp.2, 148 Tanganyika 10, 11 Taonabeae 97, 100, 136-138 Tarsiidae 53, 54, 07 Tarsiiformes 47, 52 Tarsiinae 52 Tarsioidea 51 Tarsius 45, 40, 49-59, O61, 64 spectrum 48*, 49, 64 Taurus 47 Taxodineae 99, 103 Teilhardina 58. 67 Ternstroemia 138, 140 391 INDEX Tessevaspis 192, 200-203, 245 Vitis uncinata 100 tessellata 217, 223, 245 SP. 219, 222, 225, 227, 228, 231-233, 235, 236, WALTON, J. 362 245 Warp, H. K. 4 Tetonius 53 Watson, D. M. S. 308 Tetrastigma chandleri 135 Weigeltaspis 203, 225, 245 lobata 100, 134, 135; Pl. 28, figs. 96, 97 Wellcome Trust 73 Theaceae 40, 97, 100, 135, 137, 138, 140 Wenner-Gren Foundation, N.Y. 67 Theae 100, 135 Wetherellia 16, 20, 36, 37, 39 Thelodont scales 202, 203, 220-223, 225-229, dixoni 18, 20, 36, 37, 38, 39 ; Pl. 10, figs. 56-63 ; 231-233, 235-237, 239, 261 Pl. 11, figs. 64-68 Thelodonti 258, 284 sp. 18, 37; Pl. 11, figs. 66-68 Thelodus 201, 284 variabilis 18, 36-39; Pl. 11, fig. 69 Thinnfeldia rhomboidalis 161 Wheathillaspis 297 Tuomas, H. HaMSHAW 363 wickhamkingi 204, 220, 299*, 300; Pl. 46, Thyestes (Auchenaspis) egertont 185 fig. 5 Thymelaeaceae 96, 140 SP. 237 Genus ? 100, 140 WuiteE, E. I. 11, 239, 243-310, 379 Thyrsopteris elegans 370 Whitecliff Bay, I.o.W. 15-17, 21 Tilapia 10, 11 WICKHAM KineG, W. 239 Toomss, H. A. 239, 284, 308 Williamsaspis 294, 295 Traquaivaspis 202, 227, 246 Williamsonia 17% pococki 201 Wits, L. J. 239, 292 symondst 187, 192, 193, 200-202, 217, 222, Wonnacort, F. M. 93, 308, 379 223, 225, 228-232, 235, 236, 239, 246, 262 TREWAVAS, E. 11 Yatesia gracilis 162 Trionyx 98 Tristichopterus 307 Zamites megaphyllos 162 Trochodendron pauciseminum 153 Zanthoxylon ailanthoides 123, 124 Tuber 156; Pl. 30, fig. 160 compressum 124 Tupaia 51 ornatum 126 TuRNER, Mrs. J. G. 20 Zanthoxylum 95, 123, 125, 126 Turritella Beds 18, 34, 39 compressum 99, 124* ; Pl. 26, figs. 65-67 Berane ae 99, 123, 124; Pl. 26, figs. 63, 64 Belge Zelkova keakt 155 VENABLES, E. M. 19, 20 Zingiberaceae 98, 99, 108, 109 Vitaceae 100, 133 Zygopterid ferns 367 | PRINTED IN GREAT BRIT, af <—? ae i ~ {~ iss: Pi]SpSIETRESIIS THES Saedesgeessciseess ests =e soa eae corerceees Peters cesar sed