1 1 BULLETIN OF (12 mm THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY VOL. XIII 1966-1967 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) LONDON: 1967 DATES OF PUBLICATION OF THE PARTS No. i. zoth June ...... 1966 No. 2. 3rd August ...... 1966 No. 3. 2ist September ..... 1966 No. 4. 4th October ...... 1966 No. 5. iyth November ..... 1966 No. 6. nth January ...... 1967 PRINTED IN GREAT BRITAIN BY ADLARD & SON LIMITED BARTHOLOMEW PRESS, DORKING CONTENTS GEOLOGY VOLUME XIII No. i. Cheilostomatous Polyzoa from the Upper Bracklesham Beds (Eocene) of Sussex. A. H. CHEETHAM i No. 2. Fossil Mammals of Africa No. 21. Miocene Rhinoceroses of East Africa. D. A. HOOIJER 117 No. 3. Thysanopeltidae (Trilobita) from the British Devonian. E. B. SELWOOD 191 No. 4. Exine Structure in some fossil and Recent Spores and Pollen as revealed by light and electron microscopy. J. M. PETTITT 221 No. 5. New Gymnosperms from the Tico Flora, Santa Cruz Province, Argen- tina. S. ARCHANGELSKY 259 No. 6. The British Silurian Cystoids. C. R. C. PAUL 297 Index to Volume XIII 357 CHEILOSTOMATOUS POLYZOA FROM I HE UPPER BRACKLESHAM BEDS (EOCENE) OF SUSSEX A. H. CHEETHAM BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 13 No. i LONDON: 1966 A? 10JUNW6 CHEILOSTOMATOUS POLYZOA FROM THEL V/y' <^> UPPER BRACKLESHAM BEDS (EOCENE) BY ALAN HERBERT CHEETHAM, Ph.D. Louisiana State University and University of Stockholm Pp. 1-115 ; 8 1 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 13 No. i LONDON: 1966 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 ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 13, No. i of the Geological (Palaeontological) series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. Trustees of the British Museum (Natural History) 1966 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 10 June, 1966 Price £2 2s. CHEILOSTOMATOUS POLYZOA FROM THE UPPER BRACKLESHAM BEDS (EOCENE) OF SUSSEX By A. H. CHEETHAM CONTENTS Page I. INTRODUCTION AND ACKNOWLEDGMENTS ..... 4 II. STRATIGRAPHY AND PALAEOGEOGRAPHY ...... 6 III. PRESERVATION OF MATERIAL AND TECHNIQUE OF STUDY . . ii IV. REPOSITORIES FOR MATERIAL . . . . . . .11 V. MORPHOLOGY OF UPPER BRACKLESHAM CHEILOSTOMATA ... 12 VI. CLASSIFICATION AND GENERAL ARRANGEMENT . . . .21 VII. SYSTEMATIC DESCRIPTIONS ........ 22 Suborder Anasca . . . . . . . . .22 Family Hincksinidae Canu & Bassler . . . . .22 Genus Ogivalina Canu & Bassler . . . . .22 Family Selenariidae Busk ....... 24 Genus Setosellina Calvet ...... 24 Family Lunulitidae Lagaaij ...... 28 Genus Lunulites Lamarck ...... 28 Family Onychocellidae Jullien . . . . . .31 Genus Onychocella Jullien . . . . . .31 Genus Smittipora Jullien ...... 33 Family Microporidae Gray. ...... 34 Genus Microporina Levinsen ...... 34 Genus Setosella Hincks ....... 37 Family Poricellariidae Harmer ...... 39 Genus Poricellaria d'Orbigny ...... 39 Family Steganoporellidae Hincks . . . . .41 Genus Labioporella Harmer . . . . . .41 Genus Gaudryanella Canu ...... 44 Family Aspidostomatidae Jullien ..... 46 Genus Entomaria Duvergier ...... 46 Family Farciminariidae Busk ...... 48 Genus Nellia Busk. ....... 48 Genus Vincularia Def ranee . . . . . -53 Suborder Acanthostega . ....... 58 Family Cribrilinidae Hincks . . . . . .58 Genus Cribrilaria Canu & Bassler ..... 58 Genus Membraniporella Smitt ...... 60 Suborder Ascophora . . . . . . . .62 Family Exechonellidae Harmer . . . . . .62 Genus Exechonella Duvergier . . . . . .62 Family Umbonulidae Canu ...... 64 Genus Hippopleurifera Canu & Bassler .... 64 Family Exochellidae Bassler ...... 65 Genus Escharoides Milne Edwards ..... 65 GEOL. 13, i. i 4 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Family Smittinidae Levinsen 68 Genus Smittoidea Osburn . . . . - . . 68 Family Escharellidae Levinsen . . , [ i C» ; . 71 Genus Escharella Gray . . . ._ . .; . . 71 Family Sertellidae Jullien . . . . % • . • . • 73 Genus Sertella Jullien . . . . . . -73 Family Schizoporellidae Jullien ...... 75 Genus Dakaria Jullien ....... 75 Genus Schizomavella Canu & Bassler . . . -77 Genus Escharina Milne Edwards ..... 78 Family Hippopodinidae Levinsen ..... 80 Genus Hippoporina Neviani ...... 80 Family Ditaxiporinidae Cheetham . . . . .81 Genus Caberoides Canu . . . . . . .81 Family Tubucellariidae Busk . . . . . .84 Genus Tubucella Canu & Bassler ..... 84 Family Adeonidae Hincks ....... 87 Genus Teichopora Gregory . . . . . .87 Genus Schizostomella Canu & Bassler . . . .91 Genus Adeonellopsis MacGillivray ..... 95 Family Celleporinidae Harmer ...... 99 Genus Celleporina Gray ...... 99 Family Mamilloporidae Canu & Bassler . . . . 101 Genus Kionidella Koschinsky . . . . . 101 Family Orbituliporidae Canu & Bassler .... 103 Genus Orbitulipora Stoliczka ...... 103 Genus Batopora Reuss ....... 106 VIII. REFERENCES .......... 108 SYNOPSIS Thirty-eight species distributed in 34 genera of Cheilostomata are identified in a collection of Polyzoa from the Upper Bracklesham Beds at Selsey, Sussex. All the species, including ten new ones and two left nomina aperta, are described, and several of the genera to which they belong are diagnosed. One species is re-named. The stratigraphical and geographical ranges of the Upper Bracklesham species are compiled from the literature and are used, together with the relative abundances of colonial growth forms, to interpret the stratigraphy and palaeoecology of the Upper Bracklesham Beds and the palaeo- geography of the Hampshire Basin in Late Eocene times. The morphology of the Upper Brackle- sham Cheilostomata, especially their zoarial habits and the linear dimensions of their zooecia and heterozooecia, are discussed. I. INTRODUCTION AND ACKNOWLEDGMENTS THE Eocene Polyzoan fauna of England contrasts so markedly in both abundance and diversity with that of the Continent that Vine (1889 : 156) at one time considered it doubtful that Polyzoa would ever be discovered in quantity in the British Lower Tertiaries. However, Gregory (1893), Davis (1928, 19290;, 1934, 19360, 19366, 1940, 1962), Burton (1929), and Thomas & Davis (1949) have since reported from the Eocene strata of both the London and Hampshire Basins a number of Polyzoa, including a total of 58 nominal species, with an additional 13 left nomina aperta, of Cheilostomata : 29 species from the London Clay, 29 species from the Lower CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 5 Bracklesham Beds, and 15 species from the Barton Beds. Only six species of Cheilostomata — Biselenaria off a Gregory, Cellepora petiolus Lonsdale, Heterocella sp., Lunulites transiens Gregory, Poricellaria alata d'Orbigny and Teichopora clavata Gregory (Lonsdale 1850, Gregory 1893, Davis 1962) — have been known to occur, heretofore, in the Upper Bracklesham Beds. On the other hand, more than 125 nominal species, a number of them doubtless synonyms, have been identified among the Cheilostomata from the Eocene beds of the Paris Basin (Buge 1946), most of them from the Lutetian and Auversian Stages. Such characteristic French Eocene genera as Caberoides Canu, Dakaria Jullien, Entomaria Duvergier, Gaudryanella Canu, Hippoporina Neviani, Kionidella Koschin- sky and Tubucella Canu & Bassler have not so far been reported in co-eval faunas in Britain. In November 1961 Mr. Dennis Curry presented to the British Museum (Natural History) an unsorted collection of Polyzoa from the Upper Bracklesham Beds at Selsey, Sussex. The task of sorting and studying this collection was undertaken in December 1961 and continued intermittently through 1964. Thirty-eight species belonging to 34 genera have been identified among the Cheilostomata, and a less diversified assemblage of Cyclostomata, not treated here, awaits investigation. This work fills a gap in the stratigraphical record of British Cheilostomata and permits fresh interpretation of the history of Anglo-Franco-Belgian Polyzoan faunas. The study has been complicated by the scattered, in many cases vague, descriptions and long synonymies of the Continental Eocene Cheilostomata and by some still controversial aspects of British-European stratigraphical correlation. The modern works of Curry (19580, 19586, 1962, 1962 MS.), Davis & Elliott (1957), and Wrigley & Davis (1937), which have especially promoted understanding the British Lower Tertiaries and their relation to those of the Continent, have proved indispensable to the compilation and interpretation of the stratigraphical ranges of the Upper Bracklesham species. Taxonomic distinctions between the Upper Bracklesham species are in general clear ; generic relationships are in some cases not so clear. Some of the species are almost exactly intermediate between genera, the more modern representatives of which are separated by definite morphological gaps ; others do not fit properly into any known genus. Still another source of difficulty is the confusion arising from mis-application of generic names, as in the case of Vincularia Defrance. As a taxonomic expedient, provisional assignment to named genera has been preferred to the institution of new, in most cases monotypical, genera. The main part of this study was carried out in the British Museum (Natural History) and at Louisiana State University. The biometrical analyses were com- pleted at the University of Stockholm. I am grateful to the Keeper of Zoology, British Museum (Natural History), for the use of research facilities and specimens in his care ; to the Keeper of Palaeon- tology for permission to continue working with the Curry Collection after my return to Louisiana State University, as well as to examine other specimens, including the H. Milne Edwards, J. W. Gregory, and A. G. Davis Collections, in his care ; to the 6 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Geology Department, Louisiana State University, for use of research facilities and for financial support for illustrations ; and to the Geological Institute, the University of Stockholm, for the use of research facilities. Dr. Anna B. Hastings and Dr. H. Dighton Thomas have been extraordinarily generous in guiding me through the many problems — taxonomic, nomenclatorial, and phylogenetic — which have arisen during the course of the study. It is impossible to express strongly enough my warm appreciation of their help and encouragement. Miss P. L. Cook has also given valuable advice on taxonomic problems as well as help in locating specimens and references. Others of the British Museum (Natural History) staff, including Miss S. A. Clark, Mr. R. F. Wise, and Mr. P. J. Hayward, have rendered competent technical assistance for which I am very grateful. Mrs. Martha M. Deboo, formerly of Louisiana State University, prepared the illustrations. Her assistance has contributed greatly and essentially to the comple- tion of this work. Mr. Dennis Curry has generously supplied additional material from the Upper Bracklesham Beds at Selsey and samples of the Barton Beds at Barton, Hants, and of the Sables de Fresville (Lutetian) at Gourbesville (Manche), France, from which specimens of Polyzoa were obtained for important comparisons. Dr. Sten Schager of the University of Stockholm provided Rumanian Eocene material which has proved equally important for comparisons. The National Science Foundation, U.S.A. (through a post-doctoral fellowship in 1961-62), Louisiana State University, and the Chancellor of the Swedish Universities (through a visiting professorship in 1964-65) have all made this research possible. II. STRATIGRAPHY AND PALAEOGEOGRAPHY The Curry Collection of Polyzoa was made from the well-known outcrop of the Bracklesham Beds in Bracklesham Bay, at Selsey, Sussex (National Grid, SZ 845926 ; Curry 19586 : 34). The exposures, on the foreshore, " are not so good as once they were and so much depends upon the chances of wind and tide that further detailed collecting there requires the complete leisure of a local resident " (Wrigley 1934 : i). Though the Bracklesham Beds at Selsey had been studied earlier by Dixon (1850), Fisher (1862) was the first to record the details of their stratigraphy. At Bracklesham Bay, Sussex, and at Whitecliff Bay, Isle of Wight, he recognized four fossil zones (A-D) which comprise alternating beds of glauconitic, shelly sand and sandy clay, " the clays being more prevalent in the highest member, and sands in the lower " (Fisher 1862 : 67). The three lower divisions (B-D) have since been assigned to the Lutetian and Cuisian Stages of the Continental Eocene on the evidence of their larger Foraminifera, Nummulites laevigatus (Bruguiere) and N. planulatus (Lamarck) (see Wrigley & Davis 1937, Curry 19580, 19586, 1962 MS.). The upper division has generally been regarded as Auversian (= Ledian = Lower Bartonian), owing to the occurrence in it of Nummulites variolarius (Lamarck), though there is other faunal evidence of its possibly closer affinity to the underlying Lutetian Stage (Curry 1962 MS.). The Upper Bracklesham Beds, Fisher's division A or the N. variolarius zone, has CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 7 four principal fossil beds (a-d, Fisher 1862 : 67), of which the lower three (b-d) are represented in the exposure in Bracklesham Bay (Fisher 1862 : 74 ; Curry 19580). The Polyzoan collection on which the present work is based came from (c), the highest bed but one, known as the Hard Bed or Fisher Bed 21, after the numbering scheme in Fisher's detailed section. This bed consists of " hard calcareous sand, with comminuted shelly matter and numerous Tellinae and other fossils " (Fisher 1862 : 74) ; it is overlain by a softer, clayey sand (b), rich in micro-fossils, termed the Clibs, and is underlain by clays, the Beloptera Bed and the Cypraea Bed (d), in descending order. The Hard Bed and the Beloptera Bed together constitute the Medmery Bed and with the overlying Clibs aggregate some 10 feet thick at Bracklesham Bay (Curry 19586 : 34). This group of beds occurs about half way up the Upper Brackle- sham sequence and is of similar stratigraphical position to the Tellina and N. vario- larius Beds of Whitecliff Bay, Isle of Wight (Curry 19586 : 34). In contrast to the Polyzoa of the Lower Bracklesham Beds, which are predomi- nantly encrusting forms (membraniporiform) adherent to the exteriors, and especially the interiors, of large molluscan shells, such as Venericor, Clavilithes, and Sycostoma (Davis 1934 : 205), the Cheilostomata of the Upper Bracklesham Beds include only a minority of encrusting types. Erect, arborescent (eschariform) and hollow, globular or discoidal (orbituliporiform ; see Morphology section below) zoaria occur in the greatest numbers ; erect, jointed (cellariiform) and membraniporiform zoaria are only slightly less frequent ; free, discoidal (lunulitiform) specimens are present in smaller numbers ; and fenestrate (reteporiform) zoaria are represented by only a few fragments (see Table I for the zoarial form and abundance of the species). The character of the Upper Bracklesham Cheilostomata, including both the excess of eschariform over membraniporiform zoaria and the relatively large diversity of species and genera, suggests accumulation in deeper, less turbulent water than that in which the Lower Bracklesham fauna lived (Davis 1934 : 205). The significant numbers of lunulitiform specimens attest to the granularity of the substrate, which was composed of shell fragments and Nummulitid and other larger foraminifers, as well as terrigenous detritus. Taken together, these data point to a biotope near the lower limit of effective wave action, perhaps at mid-sub-littoral depths, that is to say, 30-50 fathoms. The large cellariiform component could represent an epi-planktonic fauna essentially independent of bottom conditions. The growth requirements of the orbituliporiform species, making up so large a proportion of the Upper Brackle- sham fauna, are, unfortunately, unknown ; however, Lagaaij (1963 : 203-207) has presented convincing arguments that Fedora nodosa Silen, a Recent species with somewhat similar growth form, is attached, at least initially, to small lumps of mud on the substrate in deep water (see Morphology section below). The geographical connexions of the Upper Bracklesham Cheilostomata are more strongly Tethyan than those of any other division of the British Tertiary : 26 of the 38 species occur at least as far south-east as the Paris Basin (see Table II), whereas only 15 of the 29 Lower Bracklesham species and an even smaller proportion of the London Clay and Barton species do. Indeed, several of the Upper Bracklesham species, e.g. Escharina procumbens (Canu), Hippopleurifera canui nom. nov., and 8 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX TABLE I. — Zoarial form and abundance of Upper Bracklesham Cheilostomata O E C L M R T 1. Orbitulipora petiolus (Lonsdale) -. 149 — H9 2. Schizostomella curryi sp. nov. . . — 109 — — 109 3. Vincularia monstruosa (Canu) . . — — 94 — — — 94 4. Lunulites transiens Gregory . . — — 59 — — 59 5. Smittoidea variabilis (Canu) . 9 — 42 — 51 6. Adeonellopsis punctata (Canu) . . 45 — — — — 45 7. Entomaria dutempleana (d'Orbigny) . . 43* — — 43* — 43 8. Setosellina gregoryi sp. nov. ... — — 29* 29* — 29 9. Batopora glandiformis (Gregory) . . 27 — — — 27 10. Caber aides corniculatus sp. nov. . — 27 27 11. Escharoides aliferus (Reuss) . . 24 24 12. Teichopora clavata Gregory ... — 19 — — 19 13. Onychocella subpyriformis (d'Archiac) . 19 — J9 14. Microporina magnipora (Canu) . . . 17 — *7 15. Vincularia davisi sp. nov. ... — 15 — 15 16. Celleporina thomasi sp. nov. ... 12 — 12 17. Tubucella mamillaris (Milne Edwards) . 9 9 1 8. Sertella marginata (Reuss) ... 99 19. Schizostomella liancourti (Canu) . . 7 7 20. Nellia tenella (Lamarck) .... 6 6 21. Gaudryanella variabilis Canu ... — 6 — — — 6 22. Labioporellal dartevellei sp. nov. . . — 6 6 23. Cribrilaria parisiensis (Canu) ... — 5 — 5 24. Membraniporella radiata (Reuss) . 4 — 4 25. Escharella selseyensis sp. nov. ... 4 — 4 26. Ogivalina? dimorpha (Canu) ... — 3* — — 3* — 3 27. Kionidella hastingsae sp. nov. ... 2 — — — — 2 28. Schizomavella trigonostoma sp. nov. . . 2 2 29. Poricellaria alata d'Orbigny ... 2 — 2 30. Escharina procumbens (Canu) ... — 2 — — 2 31. Hippoporina globulosa (d'Orbigny) . . — 22 32. Adeonellopsis selseyensis sp. nov. . . — i — i 33. Dakaria beyrichi (Stoliczka) ... — i — i 34. Setosella fragilis Canu .... — i — — — — i 35. Smittipora? sp. ..... — i — — — i 36. Nellia ventricosa (Canu) .... — i — i 37. Exechonella sp. ..... — i — J 38. Hippopleurifera canui nom. nov. . — — — — i — i TOTAL 178 246 145 59 103 9 815 292* 88* 178* * Zoarial form uncertain. Abundance = number of zoaria and zoarial fragments. O = orbituliporiform ; E = eschariform ; C = cellariiform ; L = lunulitiform ; M = membrani- poriform ; R — reteporiform ; T = total. Nellia ventricosa (Canu), have not hitherto been reported north of the Aquitaine region of France. Most of the genera to which the Upper Bracklesham Cheilostomata belong, e.g. Adeonellopsis MacGillivray, Dakaria Jullien, Escharina Milne Edwards, Escharoides Milne Edwards, Hippopleurifera Canu & Bassler, Sertella Jullien, and CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 9 Smittoidea Osburn, to-day have rather wide latitudinal ranges but optima in the tropics or sub-tropics ; several, e.g. Labioporella Harmer, Poricellaria d'Orbigny, and Setosellina Calvet, are more distinctly tropical or sub-tropical ; and only a few, e.g. Escharella Gray and Microporina Levinsen, are more distinctly cool-water. The prevalence of warm-water genera seems, too, to be greater in the Upper Bracklesham fauna than in any other British Tertiary assemblage of Cheilostomata. TABLE II. — Distribution of Upper Bracklesham Cheilostomata Pre- Ypres. Ypres. Lutet. Auvers. Post- Bart. Bart. i. Ogivalinal dimorpha ?Pal. P P, A P — ?Lud. 2. Poricellaria alata ?Pal. P, A E Lud. 3- Teichopora clavata P P E, ?B, P E 4- Dakaria beyrichi P P B Olig. 5- Entomaria dutempleana B, P, A 6. Escharella selseyensis P 7- Gaudryanella variabilis P, A 8. Hippopleurifera canui A 9- Nellia ventricosa A 10. Schizostomella liancourti P ii. Setosella fragilis P 12. Escharina procumbens A ?A 13- Hippoporina globulosa P B, P I4. Labioporella? dartevellei P ?B 15- Vincularia monstruosa P, A ?E 16. Adeonellopsis punctata P, A — — ?Lud. 17- Escharoides aliferus . P, A — — Lud. 18. Microporina magnipora A — ?Lud. 19. Onychocella subpyriformis . E, P, A B, P ?E, B Lud. 20. Smittoidea variabilis . E, B, P B Lud. 21. Membraniporella radiata . P B Lud. 22. Tubucella mamillaris B, P, A B, P Mio. 23- Nellia tenella . E, P, A Rec. 24. Lunulites transiens . E E 25- Setosellina gregoryi . E, P E, P 26. Orbitulipora petiolus . E, B, P B Olig. 27. Cribrilaria parisiensis B, P B Olig. 28. Batopora glandiformis E 29. Sertella marginata Olig. 30- Adeonellopsis selseyensis 31. Caberoides corniculatus 32. Celleporina thomasi 33- Exechonella sp. 34- Kionidella hastingsae - KNOWN ONLY FROM THE UPPER BRACKLESHAM BEDS 35- Schizomavella trigonostoma 36. Schizostomella curryi 37- Smittipora? sp. 38. Vincularia davisi E = England ; B = Belgium ; P = Paris Basin ; A = Aquitaine region. Pal — Palaeocene Lud . = Ludian ; Olig. = Oligocene ; Mio. = Miocene ; Rec. = Recent. io CHEILOSTOMATOUS FOLYZOA FROM THE EOCENE OF SUSSEX Stratigraphically, the Upper Bracklesham species compare most closely with the Lutetian fauna of the Continent : 25 of the 38 species have been recorded no lower than that stage, and six species no higher. Resemblance to the British Lutetian is much less : only three species, Nellia tenella (Lamarck), Onychocella subpyriformis (d'Archiac), and Smittoidea variabilis (Canu), are common to the Upper and Lower Brackleshams (cf. Davis 1934 : 208). Resemblance to Lower Eocene faunas — those of the London Clay and the Continental Ypresian — is lower still. Six species, Batopora glandiformis (Gregory), Cribrilaria parisiensis (Canu), Lunulites transiens Gregory, Orbitulipora petiolus (Lonsdale), Sertella marginata (Reuss), and Setosellina gregoryi sp. nov., are restricted to Auversian and higher units. A total of nine species is so far known only from the Upper Bracklesham Beds. Despite their considerable resemblance to the French Lutetian fauna, the Upper Bracklesham Cheilostomata seem more likely to represent a late migrant, relict fauna related by descent to the strongly Tethys-influenced Parisian Middle Eocene fauna. Not only are six of the Upper Bracklesham species Auversian and younger guides as mentioned above, but also several of them are direct descendants of Continental Lutetian species : Lunulites transiens from L. urceolata Lamarck, Caber vides corniculatus from C. canaliculatus Canu, and Kionidella hastingsae from K. obliqueseriata Koschinsky. The origins of the Upper Bracklesham Cheilostomata have been traced as far as Senegal where Gorodiski & Balavoine (1962) encountered four of the species — Nellia. tenella, N. ventricosa, Labioporella? dartevellei and Ogivalina? dimorpha — in rocks of Palaeocene to Lutetian age. The character of the Upper Bracklesham fauna thus seems to have an important bearing on the palaeo- geographical relations between the London-Hampshire Basins and the Tethyan and boreal regions. Davis & Elliott (1957) adduced evidence that the London and Hampshire Basins were on the south-west margin of an extensive north-west European sea in Ypresian times and that, as the sea transgressed over Belgium and into England from the north-east, it brought a boreal marine fauna into juxtaposition with a tropical terrestrial flora. As a result, both the shoal-water, sandy London Clay (in places with pebble beds) of the Hampshire Basin and the deeper-water London Clay of the London Basin contain a mixture of drifted tropical plant remains with molluscan shells belonging to such boreal genera as Cyprina, Astarte, Pholadomya and Aporrhais (Davis & Elliott 1957). At the close of London Clay deposition, the sea regressed eastwards spreading the " Lower Bagshot Sands " across the Hampshire Basin (Davis & Elliott 1957, Curry 1962 MS.). The next transgressive phase, which began in the Cuisian and probably continued through the Lutetian and Auversian, resulted in prevalence of marine conditions in the Hampshire Basin where the Bracklesham Beds were deposited while the London Basin received sediments (the Bagshot Beds) which are at least in part deltaic (Curry 1962 MS.). The extreme western part of the Hampshire Basin, in the area around Bournemouth, retained non-marine conditions until Bartonian times. These relationships suggest that, whereas the London Clay transgression was essen- tially an extension of the North Sea, the Bracklesham Beds transgression was mainly, CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX n if not entirely, a Tethyan invasion through the Channel. This interpretation seems not out of harmony with the distribution of marine Lutetian, Auversian, and Bar- tonian deposits in western Europe (see Denizot 1957 : pis. 4, 5). Shoal-water conditions in the Hampshire Basin during the deposition of the Lower Bracklesham Beds may well have prevented the greater number of Tethyan Cheilostomata coming until Auversian times. The Polyzoa from a mid-Channel Tertiary outlier probably of Cuisian or Lutetian age (Curry 1962 : 194) include genera — Poricellaria, Vincularia, Lunulites — of a facies more " Upper Bracklesham " than " Lower Bracklesham " and could thus represent the vanguard of the Selsey fauna. III. PRESERVATION OF MATERIAL AND TECHNIQUE OF STUDY The Upper Bracklesham Polyzoa from Selsey are in general very well preserved. Fine tuberculation, striation, and crenulation of the surfaces of many structures have been retained in extraordinary detail (see Lunulites transiens, Text-fig. 7 ; Onychocella subpyriformis , Text-fig. 10 ; Entomaria dutempleana, Text-figs. 25-27 ; and Schizo- stomella curryi, Text -figs. 67-70). Avicularian cross-bars are usually intact, but oral spines have invariably been broken (e.g. in Hippopleurifera canui, Text-fig. 39). The effect of mechanical abrasion before interment is evident in species such as Smittipora? sp. (Text-fig. 12) and Exechonella sp. (Text-fig. 38) as well as in a few specimens of species in which the preservation is otherwise good (e.g. Microporina magnipora, Text-figs. 13, 14; and Gaudryanella variabilis, Text-figs. 23, 24). Orifices, opesiae, avicularia, ovicells, and frontal pores are free of matrix in nearly all specimens studied. Owing to its good preservation and to the very slight adherence of matrix to it, the material required no special preparation beyond the washing it had received before it was presented to the Museum.1 During the course of their investigation, surface structures such as pits, pores, and striations were put into higher relief by staining the specimens with ordinary, water-soluble, green food-colour. So treated, many otherwise obscure structures became observable under 90 X magnification with a stereo-microscope in reflected light. All the specimens were mounted on glass- or celluloid-covered, cardboard, micro- scope slides with gum-tragacanth. Illustrations were prepared with a camera-lucida attachment on the stereo- microscope at 90 x magnification. Measurements were made on the specimens, rather than on the drawings (see Morphology section below for dimensions measured) . IV. REPOSITORIES FOR MATERIAL British Museum (Natural History) All the specimens in the Curry Collection have been retained in the Department of Palaeontology, where each has been given a number bearing the prefix " D ". The holotypes of new species and all the specimens figured in this work, whether they are primary types or not, have been drawn from the Curry Collection. 1 Because the Tertiary formations exposed on the foreshore at Selsey are saturated with salt water, material collected from them requires several thorough rinses in fresh water to prevent deterioration of the fossils (Curry 19580). 12 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Geology Museum, Louisiana State University Where possible, non-figured paratypes of the new species have been obtained from the additional Upper Bracklesham material given to me by Mr. Curry and have been deposited in the Geology Museum, Louisiana State University. These specimens, assigned numbers prefixed " L.S.U.", have been included in the abundance data shown in Table I. V. MORPHOLOGY OF UPPER BRACKLESHAM CHEILOSTOMATA Zoarial Characters Growth habits of Cheilostomata have been analysed in detail by Stach (1936, 1937). To Stach's basic zoarial categories, Brown (1952 : 19, 32-35) added a number of forms which are fundamentally variations on the major growth themes. The main outlines of the Stach-Brown classification are followed here with few modifications. Membraniporiform, lunulitiform, cellariiform, eschariform, and reteporiform pat- terns, for the most part, are recognizable in the Upper Bracklesham species. Owing to the fragmentary condition of most of the specimens, however, it is not always possible to distinguish between : (r) membraniporiform zoaria which have become detached from their substrate and eschariform zoaria which are unilaminar (either originally or by separation of bilaminar zoaria along discrete basal walls), or (2) cellariiform zoaria which have lost their articulating ends and eschariform zoaria with cylindrical (" vinculariiform ") branches. Some species, e.g. Smittoidea variabilis, include two growth forms, even in the same zoarium ; attendant variations in shape, size, and, to some extent, structure of zooecia can be a source of confusion in species determination (Text-figs. 42-46). There is also in some species gradation between zoarial forms resulting from small differences in substratal or other factors in the micro-habitat ; in Setosellina gregoryi, for example, the zoarium is alternatively membraniporiform or lunulitiform depending upon whether it started on a large or a small piece of shell or other detritus. In most of the species, however, the zoarial form is not so variable. The celleporiform habit, essentially a variation of the membraniporiform one, seems not to be developed in Upper Bracklesham Cheilostomata with the possible exception of Hippoporina globulosa of which there are only two small fragments in the Curry Collection. It is not possible in the Upper Bracklesham species to discriminate vinculariiform1 and adeoniform zoaria from eschariform ones. In Tubucella mamillaris, for example, the zoarium begins from an encrusting base as a cylindrical (" vinculariiform ") stem, which, as it rises, soon gives way to lobate, flabellate (adeoniform) fronds, which, in turn, finally become broad and compressed (eschariform) branches (Text- fig. 62). Many of the Upper Bracklesham species, e.g. Adeonellopsis punctata, Schizostomella curryi, and Smittoidea variabilis, have this robust, dominating growth pattern, here characterized as eschariform. 1 The term is an unfortunate one, as the genus Vincularia invariably displays the cellariiform habit. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 13 Three species, Orbitulipora petiolus, Batopora glandiformis , and Kionidella hasting- sae, assume a colonial growth form different from any heretofore characterized by the terms listed above and for which the name orbituliporiform is here proposed. The zoarium in all three species has a distinctive axial hollow round which the erect, prismatic zooecia are grouped in one or more layers. Unlike the similar hollow in Conescharellina d'Orbigny, that of the three species mentioned is devoid of tubules presumably derived from heterozooecia (see Silen 1947 : 33). In Kionidella the hollow extends nearly the full length of the unilaminar zoarium, its blind (proximal) end formed by the basal wall of the ancestrula and its open (distal) end formed by the basal walls of the most distal circlet of zooecia (Koschinsky 1885 : pi. 7, figs. 5-10). In Batopora the hollow runs only part way up the axis of the multilaminar zoarium whose ancestrular region is apparently obscured by superposed zooecia ; the distal end of the hollow is the basal pore surrounded by superposed zooecia (Waters 1919 : 84, pi. 6, figs. 4, 6, 10). In Orbitulipora the hollow is restricted to the basal peduncle of the zoarium, round the upper end of which the zoarium expands in a compressed bilaminar frond (Text-fig. 80 ; Canu & Bassler 1931 : pi. 3, figs. 1-22). According to Canu & Bassler (1931 : 17), the ancestrula lies at the upper end of the peduncular tube. The frond of Orbitulipora may be either flabellate, with all growth distal from the smaller zooecia, or circular, with growth radial from the smaller zooecia which thus become central. Orbitulipora and Batopora have the distal ends of the zooecia oriented towards the proximal end of the zoarium as in Conescharellina, whereas Kionidella has the distal ends of the zooecia directed distally as in typical Cheilostomata. The mode of growth and the conditions for life of these and other orbituliporiform genera have not been established. Canu (1931 : 144-147) and Canu & Bassler (1931 : 19-21) supposed Orbitulipora to be free-swimming, the basal peduncle somehow lending locomotive force by a hydraulic process (presumably like that in the Cephalo- poda?). Silen (1947 : 9) stated, concerning a somewhat similar mode of life in Conescharellina conjectured by Canu & Bassler (19296 : 482, 498), " It must be emphasized that this represents entirely an assumption by these authors and that nobody has ever observed such a phenomenon. Nevertheless the mentioned authors describe the movements of these zoaria in very vivid terms as if they had actually witnessed them." Conescharellina either rests upon its flat base on the substrate (and is, thus, lunulitiform) or, more probably, is suspended by chitinous tubes from floating objects (Silen 1947 : 13). On the other hand, the orbituliporiform zoaria of Kionidella, Batopora, and Orbitulipora and the similar ones of Fedora nodosa Silen and F. edwardsi Jullien (Silen 1947 : 53-55), with their unfilled axial cavities could result from " colonial growth taking place on and around small lumps of hardened mud, which would explain the irregular shape of the cavity " (Lagaaij 1963 : 203, 204). The chitinous tubes which frequently occupy the axial hollows of Fedora nodosa, supposed by Silen (1947 : 13) to be connexions between the globular zoaria of that species (which Silen argued were the internodes of a cellariiform zoarium), have proved to be the perisarcs of an associated Hydroid (Lagaaij 1963 : 204-206). i4 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Zooecial Characters The Upper Bracklesham Cheilostomata occupy such a broad band on the taxonomic spectrum — only four genera, Nellia, Vincularia, Schizostomella, and Adeonellopsis, are represented by more than one species, and no genus by more than two — that their zooecial morphology is extremely diverse. Thus, it is not possible to make generalizations about zooecial characters which would apply to the whole assemblage, and the remarks that follow are intended only as amplifications of certain points in the more exhaustive treatments of zooecial morphology contained in the works of Harmer (1926, 1957), Brown (1952), Larwood (1962) and others. ZOOECIAL WALLS. Basal, lateral, distal, and proximal walls, communication pores, and pore chambers are visible in some of the Upper Bracklesham specimens and are described where observed. Walls associated with the frontal surface — including sub-frontal cryptocysts and frontal gymnocysts in the Anasca, super-frontal shields in the Acanthostega, and frontal walls in the Ascophora — are of first order significance in the classification of Cheilostomata and are described in as much detail in this work as the material permits. Descriptive terminology (e.g. perforate, marginally areolate) is preferred to a genetical one (e.g. tremocyst, pleurocyst). ORAL STRUCTURE. In fossil Cheilostomata the structure of the primary orifice is discernible only in the Acanthostega and Ascophora ; in the Anasca the orifice is lost during decomposition of the membranous frontal wall, though its form is indicated approximately by the size and shape of the opesia in species having an extensive cryptocyst, especially those belonging to the Microporidae, Poricellariidae, and related families. In the Ascophora the form of the primary orifice, which may be difficult to observe because of the development round it of a peristome with a secon- dary orifice usually more variable in form, is of paramount importance in classification at the generic level. In the Acanthostega " the shape of the primary orifice is not usually of diagnostic value " (Larwood 1962 : 35). In the present work, oral structure is described by referring to components (e.g. sinus, condyles) rather than to shape designations (e.g. hippoporine, cribriline) which are subject to personal interpretation. AVICULARIA AND viBRACULA. A variety of these structures is present in the Upper Bracklesham species : adventitious and vicarious avicularia and interzooecial vibracula (see Larwood 1962 : 41, for definitions of these terms) . Vicarious avicularia are certainly more common than adventitious ones in the Anasca and vice versa in the Acanthostega and Ascophora, but exceptions to the general rule occur among the Poricellariidae and Farciminariidae in the Anasca and among the Adeonidae in the Ascophora. The vibraculum-bearing species are Anascans which belong to closely related families, Selenariidae and Lunulitidae. In most of the Upper Bracklesham species, the avicularia or vibracula, whether adventitious, interzooecial, or vicarious, are well differentiated in form and structure from the normal zooecia. The avicularia often have pointed rostra and pivotal condyles or cross-bars. As noted by Larwood (1962 : 41) the absence of pointed CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 15 rostra or pivotal structures in fossils may be due to wear ; thus it is necessary to examine many specimens before it can be concluded that a species is characterized by avicularia with rounded rostra or without cross-bars. Most of the Upper Brackle- sham material is so well preserved, however, that details of avicularian structure can be ascertained from even a few zooecia (see, e.g. Hippopleurifem canui, Text-fig. 39 ; Escharina procumbens, Text-figs. 56, 57). In some major taxa, e.g. the Ascophoran family Adeonidae, avicularia are regularly without condyles or cross-bar, but Brown's (1952 : 191) suggestion that adventitious avicularia in the Anasca are always without such structures is incorrect (see Harmer 1926 : pi. 14, figs. 18-20, Nellia tenella ; pi. 23, figs. 6-8, Poricellaria ratoniensis). In four species of Upper Bracklesham Cheilostomata, belonging to widely separated families, vicarious avicularia2 are only feebly differentiated from the normal zooecia. The differentiation is least in Labioporella? dartevellei (Text-figs. 21, 22) where it consists in slight enlargement of the avicularian (?) zooecia. In the two species of Vincularia, V. monstruosa (Text-figs. 30, 31) and V. davisi (Text-figs. 33, 34), the avicularian (?) zooecia have undergone curvature on their proximo-distal axis in addition to slight enlargement. The avicularian zooecia of Schizostomella curry i (Text-figs. 68, 70) are enlarged in the oral region in just the same way as the avicularia (= B-zooecia) are in Steganoporella (see Harmer 1926 : 268). OVICELLS AND GONOECiA. The systematic importance of the ovicell, one of the basic tenets upon which Canu & Bassler (1917, 1920, 1927, 19296) established their classification of the Cheilostomata, has been seriously questioned by Silen (1944). Brown (1952), Lagaaij (1952) and Larwood (1962). The separation of five morpho- logical types — hyperstomial, entozooecial, peristomial, vestibular, and entotoichal (Brown 1952 : 36, 37) — is still a useful descriptive device in spite of the large amount of mutual inter-gradation. Hyperstomial ovicells are the most common type in the Upper Bracklesham Cheilostomata : 14 of the 38 species display them. Entozooecial ovicells occur in three species, and peristomial ovicells in one. None displays vestibular or ento- toichal ovicells, though Canu (1925 : 47) described the ovicell of Entomaria dutem- pleana as entotoichal evidently through mis-interpretation (see Systematics section below and Text-fig. 27) . Five species, all belonging to the family Adeonidae, possess specially modified zooecia (gonoecia = gonozooecia of Brown 1952 : 37, not Cheetham 1962) instead of ovicelled zooecia. The gonoecia show a progressive series of modifications in the Upper Bracklesham Adeonidae from the feebly differentiated dimorphs of the ordin- ary zooecia of Teichopora clavata (Text-figs. 65, 66), through the distally swollen, orally modified gonoecia of Schizostomella curry i (Text-figs. 67-69) and Adeonettopsis selseyensis (Text-fig. 73), to the generally swollen, orally and frontally modified gonoecia of Adeonellopsis punctata (Text-figs. 74, 75). The remaining 15 species of Upper Bracklesham Cheilostomata apparently lack specially modified fertile zooecia of any sort. 2 It is possible that these features are other forms of zooecial dimorphs in the four species concerned. 16 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Measurements Certain linear dimensions are usually measured in descriptive work on Cheilo- stomata (Canu & Bassler 1920, 1923 ; Osburn 1950, 1952 ; Brown 1952 ; Lagaaij 1952 ; Larwood 1962 ; Cheetham 1962, 19636). For the Upper Bracklesham species the following are recorded : Lz = zooecial length, the maximum distance between the zooecial proximal and distal margins projected to the proximo-distal axis of the frontal surface. lz = zooecial width, the maximum distance between the zooecial lateral margins measured perpendicular to the proximo-distal axis of the frontal surface. ho = oral or opesial length, measured in the same direction as the proximo-distal axis of the zooecial ^ntal surface and always in the plane of the orifice or opesia even wherethat plane is not parallel to the frontal surface. (For sinuate orifices,^ral length includes the sinus.) lo = oral or opesial width, measured perpendicular to oral or opesial length. (Oral dimensions are taken at the primary orifice where possible ; where measurements of secondary orifice are substituted, a notation to that effect is given in the Measurements section of the species description.) Lov = ovicell length, the maximum distance between the proximal and distal margins of the ovicell projected to the proximo-distal axis of its frontal surface. (Ovicell length is determined only for hyperstomial ovicells, owing to the lack of definite margins in entozooecial and peristomial ones. The full length, width, and oral dimensions of gonoecia are recorded for species in which those structures supplant ovicelled zooecia.) Lav or Lv = avicularian or vibracular length, the maximum distance between rostrum and opposite margin measured in the plane of the avicularian or vibracular upper surface. (Width and oral dimensions are recorded for vicarious avicularia only.) Zoarial dimensions are recorded for reteporiform and orbituliporiform species only. The arithmetic mean and standard deviation of each linear dimension of each species are calculated by the standard formulae, and the number of specimens measured and the observed range of the dimension are recorded. For brevity, the following scheme of presentation is used throughout the Systematics part of this paper : Lz (10) 0-500 (0-0500) mm. 0-45-0 -55 mm. Character (Number of Mean (Standard Observed range symbol specimens) deviation) All dimensions are in decimal fractions of a millimeter, the observed range expressed to the nearest o-oi mm., and the mean and standard deviation carried, respectively, one and two places farther. Measurements were taken with a 15 X micrometer ocular in an American Optical stereo-microscope fitted with 6x objective, and each CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 17 observation was recorded to the nearest whole micrometer division. Means, standard deviations, and observed ranges are calculated in micrometer units and then converted to millimeters by the factor 0-00855 mm. per micrometer unit. TAXONOMIC SIGNIFICANCE OF MEASUREMENTS. Assessment of the value of quantitative characters as a discriminatory tool in Cheilostome taxonomy has barely been started by studies of variation in a few species (e.g. the works of lilies 1953 and David & Mongereau 1961). The pessimistic feelings of most workers are probably typified by the remarks of Brown (1952 : 33, 34) that " no great importance is attached to the value of ... measurements as they are found to vary widely even in the same colony," or that " they are seldom to be used as the basis for differentiating species in the absence of other differences of a more positive character " (see also Larwood 1962 145, 46). Lack of evaluation has not kept some authors from founding species chiefly, or even solely (see Canu & Bassler 1919 : 89, Stegano- porella parvicella), on dimensional criteria. Examples of quantitative discrimination which has subsequently proved erroneous are provided by some of Lang's Cretaceous Cribrimorphs (revised by Larwood 1962) and some of Canu & Bassler's Tertiary species of Steganoporella (revised by Cheetham 19636). Knowledge of the magnitude of variability is crucial to the assessment of any dimensional character as a taxonomic criterion. Yet most authors who have described Cheilostomata have failed to include any meaningful measure of variability in their specific characterizations, despite the fact that they have given some form of measurement. They have generally followed Canu in giving either a single, " typical " value, perhaps obtained as the arithmetic mean of an unstated number of observations, or a range of values obtained as the extremes of an again unstated number of observations. Either alternative gives almost no information about the dispersion of the character in the sample at hand or in the population from which it was drawn (Simpson, Roe & Lewontin 1960 : 78-82). Variance and standard deviation are certainly the most widely used measures of variability in statistical work. From the standard deviation may be calculated, not only the standard error from which confidence limits for the population mean may be obtained, but also, a measure of relative variability, Pearson's coefficient of variation, from the formula T7 loos V == — = — . x By use of V's so obtained, it is possible to compare variation of a character in dif- ferent samples of the same species or in different species (Simpson, Roe & Lewontin 1960 : 90). For the Upper Bracklesham species, it is necessary to correct the V's, because of small sample size, by Haldane's method (Simpson, Roe & Lewontin 1960 : 101, 102) : F(4iV+i) 4N The values of Vc for the Upper Bracklesham species are presented in Table III. GEOL. 13, I. 2 i8 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX TABLE III. — Corrected coefficients of variability of linear dimensions of Upper Bracklesham Cheilostomata 1. Ogivalina? dimorpha 2. Setosellina gregoryi 3. Lunulites transiens 4. Onychocella subpyriformis 5. Smittipora? sp. . 6. Microporina magnipora 7. Setosella fragilis . 8. Poricellaria alata. g. Labioporellal dartevellei 10. ii. 12. 13- 14. 15- Gaudryanella varibilis . Entomaria dutempleana Nellia tenella Nellia ventricosa . Vincularia monstruosa Vincularia davisi . 1 6. Cribrilaria parisiensis . 17. Membraniporella radiata 1 8. Exechonella sp. . 19. Hippopleurifera canui . 20. Escharoides aliferus 21. Smittoidea variabilis 22. Escharella selseyensis 23. Sevtella marginata 24. Dakaria beyrichi . 25. Schizomavella trigonostoma 26. Escharina procumbens . 27. Hippoporina globulosa . 28. Caber aides corniculatus . Tubucella mamillaris 29. 30- Teichopora clavata Lz Iz ho lo Lov Lav or 13- 9 23- 2 »7' o 16- 5 — 9' 5 12- 7 10' 2 12 • 5 — IO •4 8- 2 4' 5 13- 4 9' 8 — 18 •9 IS' 8 10- 5 10' 2 II- 4 — 6 •3 ii • 7 6- 4 XX1 9 14- 9 — 6- 2 12- 8 4' 4 5' 8 — 2 • 9 7' 9 7' 4 ii • i — 6- 4 9' 4 13 •2 o- o — 18 •2 5' 6 ii • 8 9 '8 22 • 2 — 9' 7 J3' 7 10 o 14- 8 — 12- 9 10- o 23' o 9' 9 — 22 •2 4' 2 10- 8 10 '5 12- 2 — 23 •O 4' z 17- 8 5 •2 25- I — IO •2 8- 2 10- 9 9 8 6- 7 — a{26 •4 •cS 8- 9 6- 3 3 •3 10- 7 — 6 •8 6- o *3- 2 10 •8 T 9 5- 2 12 •8 4' 3 15- O IO •o 13- 2 6- 3 5 •2 4' o 6- 8 12 '3 10- 4 — 3' 3 3' 4 6 •i 12- 7 — 12- 0 12 • i 14 '3 4' 5 10- 6 22 •3 I5-2 f 12- 5 8 10 4 9' x {£ ^ 17 •8 IS' 9 I41 8 13 6 4' 9 6- 2 6- 6 6- 7 18 4 13- 6 9' 2 (21 •6 6- 8 14- 4 19 •6 7' 4 _ 12- 8 26- 8 13 •o 5' 2 15 •6 10- 8 23- 2 8 •5 9' 3 o- 0 12 •o 17- 0 13- 4 7 •3 4' 6 3' 6 3 •o 6- 6 10- 9 10 •4 7' 6 i- 5 7 •o 9' 9 13- 2 15 •i 18- 5 — 3- 3 5' 5 9 • i 12 • 7 5. 2 d/12 •9 V3 •I 9' i 3' i XI •5 12- 5 3' 6 d{i6 •4 •i 12- i 15' 9 19 •8 12- x 17 •6 9' 9 6' 7 8 •8 6- o 3' 3 9 •o 13- 9 9' 6 12 •7 7' 9 7' i 18 •2 13- 2 10- 4 13 •9 5' 4 6- (> 8 •8 20- 8 6' 5 6 •5 8- 9 10- 8 14 •8 II • 3 13' 4 20 •i 9" 5 ii • 6 18- 4 T4 3 12 •5 19- 9 ii- 4 o •o b Zooecia on uni- and bilaminar fragi 8-3 31. Schizostomella curryi .... 32. Schizostomella liancourti 33. Adeonellopsis selseyensis 34. Adeonellopsis punctata .... 35. Celleporina thomasi .... 36. Kionidella hastingsae .... 37. Orbitulipora petiolus .... 38. Batopora glandiformis .... a Vicarious and adventitious avicularia. c Polymorphic avicularia. d Dimorphic avicularia. Lz, Iz, ho and lo exclude modified zooecia. The significance of coefficients of variation can be judged by comparing F's for many species, a procedure which Simpson, Roe & Lewontin (1960 : 90) deemed CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 19 " generally valid if the variates are analogous and belong to the same category — for instance, if they are all linear dimensions of anatomical elements ".3 For mammals, linear dimensions yield " hundreds of V's, . . . the great majority " of which " lie between 4 and 10, and 5 and 6 are good average values " (Simpson, Roe & Lewontin 1960 : 91). By this standard, the Upper Bracklesham Polyzoa are extraordinarily variable. David & Mongereau (1961), in their study of Cellaria fistulosa auctt. non Linne, obtained variances and means from which V's of 6-9 for zooecial length and 9-0 for zooecial width may be calculated. These are lower than for most of the Upper Bracklesham species, though comparable to the ones for species having cellariiform zoaria. In order to broaden the comparison of variability of characters, it is possible to include in this analysis two American Early Tertiary assemblages of Cheilostomata (Cheetham 1962, 19636) for which x, s, and N of standard dimensions have been recorded. V's are calculated and corrected by the same formulae used for the Upper Bracklesham species. Then, for each linear dimension, the number of species in the combined fauna having Vc less than 10-0 is entered in Table IV. These species numbers are expressed as percentages of the total number of species showing the character, and for each percentage approximate 95% confidence limits are calculated by the formula (Sylvester-Bradley 1958 : 222). The value 10-0 was chosen as the limit between high V's and low ones, not just because Simpson, Roe & Lewontin considered values of 4-10 to be usual, but also because variability amounting to no more than 10% of the mean indicates a relatively small overlap between populations even where the difference between their means is fairly small. For example, if normal distribution of the variate is postulated for each of two populations A and B, 90% of A will be distinct from 90% of B even where the difference between their means (\XA — xn\) amounts to no more than 0-128 (XA + XB), provided that their V's are 10 or less and that their variances are approxi- mately equal.4 Canu & Bassler (1920 : 262, 263), as mentioned above, established two species of Steganoporella on a difference in mean zooecial width of this magnitude (\XA — XB\ = 0-12 mm. ; 0-128 (XA +##) = 0-077 mm.), but, unfortunately, the variation in that character was subsequently shown (Cheetham 19636 : 54-56) to be much too great, at V = 18-2, for specific separation. On the other hand, the dif- ference in mean zooecial lengths of two of the Upper Bracklesham species, Vincularia monstruosa and V. davisi, at 0-162 mm.,5 is greater than 0-128 (XA + XB) — 0-133 mm., 3 Reyment (1963 : 684) regarded as valid only comparisons of the same character. 4 Ninety per cent, joint non-overlap is regarded as the minimum requirement for separate taxonomic status by some systematists (e.g. Mayr, Linsley & Usinger 1953 : 145). However, as Sylvester-Bradley (1958 : 225-227) has pointed out, overlap in several uncorrelated characters taken simultaneously may be less than 10% even where any single one shows overlap of more than 10%. The expression o- 128 (£4 -\-~XB) may be derived from Mayr, Linsley & Usinger's (1953 : I46) coefficient of difference, by substituting SA = O-IOXA and SB = 0-10x3. 5 Even if the nearer limits of their 95% confidence intervals are used, the difference indicates 87-5% non-overlap. 20 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX and both V's are less than 10 ; thus the two species are approximately 90% distinct in this dimension (and, in addition, show other differences " of a more positive character "). TABLE IV. — Relative variability of linear dimensions of Cheilostomata (Based upon data from the Upper Bracklesham species and from Cheetham 1962, 19636) A. Relatively constant characters: 1. Length of vicarious avicularium . 2. Width of primary orifice ..... 3. Length of ovicell or gonoecium .... B. Relatively variable characters : 1. Length of zooecium ...... 2. Width of zooecium ...... 3. Length of primary orifice ..... 4. Length of opesia not separate from opesiules 5. Length of avicularium or vibraculum regardless of position ....... 6. Width of opesia not separate from opesiules 7. Length of adventitious avicularium . C. Characters for which evidence is insufficient to judge variability : 1. Length of opesia separate from opesiules 2. Width of opesia separate from opesiules 3. Length of inter zooecial vibraculum 4. Width of secondary orifice .... 5. Length of secondary orifice .... It must be emphasized that, in comparisons of the sort exemplified here, " the mean is a necessary statistic but it cannot be used on its own to test assumptions of difference in dimensions. For this one requires a statistic which is able to express the spread of the data " (Reyment 1963 : 684). Therefore, it is of the utmost importance that a measure of variability be attached to all data which are presented No. No. Spp. Conf. Spp. Vc 0-13-0-14 mm. lo (4) 0-081 (o-ono) mm., 0-07-0-09 mm. Lv (4) 0-086 (o-oooo) mm., 0-09 mm. Mature zooecia Lz (10) 0-302 (0-0280) mm., 0-26-0-36 mm. Iz (10) 0-250 (0-0310) mm., 0-21-0-30 mm. ho (10) 0-211 (0-0210) mm., 0-19-0-26 mm. lo (10) 0-129 (°'OI58) mm., 0-10-0-16 mm. Lv (10) 0-120 (0-0121) mm., 0-10-0-14 mm. REMARKS. Gregory (1893 : 235) described Biselenaria offa from the Barton Beds, Barton, Hants., on the basis of three specimens on slide 49759, Edwards Collection, illustrated on his plate 30 as " Fig. 4. Zoarium of type specimen : upper surface. Fig. 4«. Part of another specimen : under surface. Fig. 5. Upper surface of another zoarium." The specimens illustrated on his figs. 4 and 40 are conspecific with Cellepora petiolus Lonsdale, the type species of Orbitulipora (see description of that species below). The third specimen is conspecific with the material referred to by Canu, Burton, Dartevelle and Buge as " Biselenaria offa Gregory ". Unfor- tunately, as Gregory clearly designated as holotype one of the specimens of Orbituli- pora petiolus, there is no alternative to giving this species a new name. S. gregoryi is a guide species to Upper Eocene (Auversian and Bartonian) strata on both sides of the Channel. It has its nearest counterpart in the Recent Mediter- ranean species 5. capriensis (Waters) in which the zoarium is more conical with a more concave base, the zooecia are larger and arranged in clockwise spirals, and the peripheral vibracula are placed on the edge of the zoarium rather than on the basal surface. S. goesi (Silen), a Recent Caribbean and Gulf of Mexico species, encrusts larger grains than the other two species and thus never becomes fully lunulitiform. Lagaaij (1963 : 172) reported both clockwise and anti-clockwise spirals in 5. goesi. The coincidence of the smaller vibracular condyle and the direction of coiling of the zooecial rows noted by Lagaaij in 5. goesi is also distinguishable in 5. gregoryi. DISTRIBUTION. Eocene (Auversian, Bartonian) ; England, France. FIGS. 4-6. Fig. 4. Setosellina gregoryi sp. nov. 0.48346. Paratype. Basal view of a small, nearly complete zoarium. Fig. 5. Setosellina gregoryi sp. nov. 0.48345. Paratype. Basal view of a larger, fragmentary zoarium. Fig. 6. Setosellina gregoryi sp. nov. 0.48344. Holotype. Ancestrular region showing zooecia and vibracula. 28 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Family LUNULITIDAE Lagaaij Genus LUNULITES Lamarck Lunulites transiens Gregory (Text-figs. 7-9) 1850 Lunulites urceolata Lamarck ; Lonsdale : 159, pi. i, fig. 8. 1878 Lunulites urceolata Lamarck ; Lonsdale : 201, pi. i, fig. 8. 1893 Lunulites transiens Gregory : 233, pi. 29, figs. 13, 14 ; pi. 30, fig. i. 1933 Lunulites transiens Gregory ; Dartevelle : 106, 113. HOLOTYPE. 49724, Edwards Collection. Barton Clay ; Barton, Hants. Figured by Gregory (1893, pi. 29, fig. 14). FIGURED SPECIMENS. 0.48372 (Text-fig. 7), 0.48373 (Text-fig. 8), 0.48374 (Text-fig. 9). ADDITIONAL MATERIAL. Fifty-six specimens, 0.48375-0.48430. DIAGNOSIS. Lunulites with discoidal, slightly concavo-convex zoarium, basal surface divided into irregular, finely perforate, radial sectors ; new zooecial rows originating from vibracular rows ; mural rims of adjacent zooecia confluent, forming thick, high transverse ridge between zooecia of same row ; cryptocyst very narrow, shelf-like. DESCRIPTION. Zoarium free in adult stages, discoidal, probably circular in outline, the upper surface moderately convex, the basal side slightly concave. Ancestrula encrusts Nummulites or small shell fragment which usually shows at centre of basal surface. Zooecia arranged in radial lines emanating from central ancestrula or from central fragment of an older zoarium from which regeneration has occurred. Vibra- cula occur in similar radial lines alternating with zooecia, one at every zooecial corner, except at points of intercalation of new zooecial rows. New zooecial rows originate from vibracular rows, the first zooecium of the new row lacking vibracula at its proximo-lateral corners, but having a single one proximally. Basal surface of zoarium with irregular, radial, finely perforated sectors, increasing in number peri- pherally by intercalation, separated by deep furrows. Zooecia rhomboidal, slightly longer than wide, their mural rims confluent. Mural rim granular, broad and low laterally, but rising rapidly proximally to form a wide, thick, transverse wall separating cryptocyst of distal zooecium from opesia of proximal one. Cryptocyst a flat, imperforate, finely granular, very narrow shelf on proximal and lateral margins of opesia. Opesia oval, more broadly rounded proximally than distally, without differentiated indentations. Communication between zooecia by a single, large pore placed in middle of distal wall and a similar pore in each lateral wall. Vibracula interzooecial, slightly shorter and much narrower than zooecia, rhombic, symmetrical, those of a series increasing in size distally. Vibracular opesia sym- CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 29 If' 1.00mm FIGS. 7-9. Fig. 7. Lunulites transiens Gregory. 0.48372. Eight zooecia and four vibracula. Zooecium at upper right is primoserial, budded from a vibraculum. Fig. 8. Lunulites transiens Gregory. 0.48373. Basal aspect of a fragmentary zoarium emanat- ing from a small Nummulitid test. Fig. 9. Lunulites transiens Gregory. 0.48374. Basal aspect of a fragmentary zoarium regenerated from an older zoarial fragment. 30 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX metrical, oval, divided by a pair of stout condyles into a small distal portion and a much larger proximal one. Ovicell lacking. MEASUREMENTS : Ordinary zooecia Lz (5) 0-340 (0-0266) mm., 0-30-0-37 mm. Iz (5) 0-290 (0-0127) mm., 0-27-0-31 mm. ho (5) 0-238 (0-0305) mm., 0-20-0-27 mm. lo (5) 0-166 (0-0155) mm., 0-14-0-18 mm. Primoserial zooecia Lz (5) 0-398 (0-0239) mm., 0-36-0-42 mm. Iz (5) 0-260 (0-0076) mm., 0-26-0-27 mm- ho (5) 0-231 (0-0326) mm., 0-20-0-24 mm- lo (5) 0-137 (0-0086) mm., 0-13-0-14 mm. Vibracula Lv (5) 0-197 (0-0354) mm., 0-17-0-23 mm. Iv (5) 0-077 (0-0166) mm., 0-06-0-10 mm. Vibracula proximal to primoserial zooecia Lv (5) 0-327 (0-0245) mm., 0-30-0-36 mm. Iv (5) 0-162 (0-0270) mm., 0-14-0-20 mm. REMARKS. Canu (19076 : 83), in re-describing and illustrating L. urceolata Lamarck, has given several criteria for the separation of the French Lutetian species from L. transiens. Of these, the shape of the zoarium and the development of the cryptocyst seem to be the most readily applicable. The pattern of increase in number of zooecial rows exhibited by L. transiens, although distinctive, is by no means unique as Gregory (1893 : 233) implied. Among the American Tertiary species of Lunulites, L. distans Lonsdale and L. fenestrata (de Gregorio) show the same pattern. In L. jacksonensis (Canu & Bassler), L. ligulata (Canu & Bassler) and L. tintinabula (Canu & Bassler) increase is by bifurca- tion of zooecial rows ; in L. bouei Lea and L. truncata de Gregorio it is accomplished by bifurcation of vibracular rows with intercalation of zooecial rows between them. Among the European Tertiary species the L. transiens pattern seems to be much more widespread. Perhaps these patterns will make possible subgeneric partition of the very inclusive genus Lunulites. DISTRIBUTION. Eocene (Auversian, Bartonian) ; England. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 31 Family ONYCHOCELLIDAE Jullien Genus ONYCHOCELLA Jullien Onychocella subpyriformis (d'Archiac) (Text-figs. 10, n) 1846 Eschara subpyriformis d'Archiac : 195, pi. 5, figs. 21, 2ia. ?i85i Semieschara parisiensis d'Orbigny : 366. i86g« Membranipora angulosa (Reuss) ; Reuss : 253, pi. 29, figs. 9-11. 1891 Onychocella angulosa (Reuss) Waters : 9. 19076 Onychocella angulosa (Reuss) ; Canu : 21, pi. n, fig. n. 1910 Onychocella angulosa (Reuss) ; Canu : 840, 844. 1916 Onychocella angulosa (Reuss) ; Faura y Sans & Canu : 298. 1925 Onychocella angulosa var. parisiensis (d'Orbigny) ; Canu : 46. 1926 Onychocella angulosa var. parisiensis (d'Orbigny) ; Canu : 748, pi. 27, fig. 7. ?i929 Onychocella n. sp., Burton : 328. 1929^ Onychocella angulosa (Reuss) ; Canu & Bassler : 24. 1930 Onychocella parisiensis (d'Orbigny) ; Canu & Bassler : 22. !933 Onychocella parisiensis (d'Orbigny) ; Dartevelle : 64, 106, 113. 1934 Onychocella parisiensis (d'Orbigny) ; Davis : 220. J935 Onychocella parisiensis (d'Orbigny) ; Dartevelle : 112. 1936 Onychocella parisiensis (d'Orbigny) ; Dartevelle : 26. 1946 Onychocella angulosa var. parisiensis (d'Orbigny) ; Buge : 430. 1949 Onychocella angulosa (Reuss) ; Balavoine : 773. 1951 Onychocella angulosa (Reuss) ; Kyri : 71. 1956 Onychocella parisiensis (d'Orbigny) ; Balavoine : 321. 1957 Onychocella angulosa var. parisiensis (d'Orbigny) ; Balavoine : 191. 1960 Onychocella parisiensis (d'Orbigny) ; Balavoine: 246. ?i962 Onychocella angulosa (Reuss) ; Ghiurca, table i. ?ig63 Onychocella angulosa (Reuss) ; Malecki : 104, pi. 9, fig. 14. 1963 Onychocella angulosa (Reuss) ; Braga : 23, pi. 2, fig. 5. FIGURED SPECIMENS. 0.48431 (Text-fig. 10), 0.48432 (Text-fig, n). ADDITIONAL MATERIAL. Seventeen specimens, 0.48433-0.48449. DIAGNOSIS. Onychocella with zooecial opesia semi-circular, oval, or sub-pyriform and sub- terminal on cryptocyst, in some zooecia with distinct distal shelf ; opesiular indentations not differentiated ; avicularian opesia more broadly rounded distally than proximally. DESCRIPTION. Zoarium encrusting, but probably rising in unilaminar, erect, tubular fronds, the zooecia arranged in irregular, alternating rows. Additional rows originate from an avicularium inserted between rows of zooecia. Zooecia hexagonal, pentagonal, or rhomboidal, separated by a faint groove at the crest of the confluent mural rims. Zooecial length and width subequal, both ex- tremely variable. Mural rim thin, sharp, and finely granular in most zooecia, but thickened and raised in some ; gymnocyst lacking. Cryptocyst extensive, markedly concave, present all round the opesia, widest proximally. Surface of cryptocyst imperforate, finely granular, merging with mural rim. 32 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Opesia sub-terminal on the cryptocyst, extremely variable in size, semi-circular, oval, or sub-pyriform, the distal margin evenly rounded, the proximal margin straight, slightly convex, or slightly concave ; distinct distal shelf present in opesiae of zooecia with thickened mural rims. Opesiular indentations not differentiated from opesia. 0.500mm FIGS. 10-12. Fig. 10. Onychocella subpyriformis (d'Archiac). 0.48431. Four zooecia and an onychocellarian avicularium. Fig. n. Onychocella subpyriformis (d'Archiac). 0.48432. Four zooecia partly shaded to show details of the raised and thickened mural rim and the distal opesial shelf of the zooecium at lower centre. Fig. 12. Smittipora? sp. D . 48450. Fragment with worn and slightly broken zooecia. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 33 Avicularia vicarious, longer but narrower than zooecia, with mural rim and cryptocyst identical with those of zooecia. Inwardly directed processes of mural rim separate elongate rostrum from short proximal part.. Rostrum channelled, acutely pointed, directed sharply to either right or left, its axis diverging at angles of about 30° from main axis of avicularium. Avicularian opesia symmetrical, oval, more broadly rounded distally, sometimes pointed proximally, undivided by condyles or pivotal bar. Ovicell lacking, but zooecia with nearly terminal opesiae may be fertile ones. MEASUREMENTS : Ordinary zooecia Lz (10) 0-362 (0-0557) mm., 0-31-0-49 mm. Iz (10) 0-353 (°'°36i) mm., 0-30-0-39 mm. ho (10) 0-153 (°'OI53) mm., 0-13-0-18 mm. lo (10) 0-174 (0-0194) mm., 0-14-0-21 mm. Zooecia with distal opesial shelf ho (5) 0-128 (0-0264) mm., 0-10-0-17 mm. lo (5) 0-133 (0-0076) mm., 0-13-0-14 mm. Avicularia Lav (5) 0-506 (0-0305) mm., 0-46-0-55 mm. lav (5) 0-258 (0-0159) mm., 0-23-0-27 mm. ho (5) 0-171 (0-0121) mm., 0-15-0-18 mm. lo (5) 0-118 (0-0113) mm., 0-10-0-13 mm- REMARKS. Canu (1910 : 844) recognized that d'Archiac's species is the same as the Eocene species identified by most authors with 0. angulosa (Reuss), but, when he sorted the species out (1926 : 748), he used the unfortunate (and junior) name Onychocella parisiensis (d'Orbigny). The synonymy is further complicated by the existence of two Cretaceous species of Onychocella, 0. subpyriformis (Hagenow) and 0. parisiensis (d'Orbigny), whose names are homonyms of the Eocene species. 0. angulosa s.s., the Oligocene-Recent type species of Onychocella, is distinguished from 0. subpyriformis by its sub-central opesia and its less angulated avicularian rostrum (see Reuss 1848 : pi. n, fig. 10 ; Harmer 1926 : pi. 16, fig. 8). DISTRIBUTION. Eocene (Lutetian, Auversian) ; France. Eocene (Auversian, Bartonian) ; Belgium, Spain. PEocene (Bartonian) ; England. Eocene (Ludian) ; Italy, Rumania, ? Poland. Genus SM1TTIPORA Jullien Smittiporat sp. (Text-fig. 12) FIGURED SPECIMEN. 0.48450 (Text-fig. 12). GEOL. 13, I. 3 34 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX DESCRIPTION. Zoarium erect, cylindrical, slender, composed of 6 longitudinal rows of zooecia alternating in position. Zooecia separated by faint grooves at crest of confluent mural rims, rhomboidal to claviform ; length slightly less than twice width. Gymnocyst lacking. Mural rim worn, rather thick. Cryptocyst extends about two- thirds zooecial length from proximal margin ; surface slightly depressed, perceptibly concave, merging with mural rim peripherally, granular, imperf orate. Opesia terminal on cryptocyst, sub-circular, more broadly rounded proximally than distally, without differentiated indentations. Heterozooecia, ovicells not present. MEASUREMENTS : Lz (5) 0-446 (0-0496) mm., 0-38-0-50 mm. Iz (5) 0-253 (°-°i55) mm., 0-23-0-27 mm. ho (5) 0-135 (0-0153) mm., 0-11-0-14 mm- lo (5) 0-125 (°'OI77) mm., 0-10-0-14 mm- REMARKS. This species, represented solely by the fragment illustrated, resembles Smittipora midwayanica Canu & Bassler (1920 : 225, pi. 4, figs. 16-19) from the Palaeocene of Arkansas, U.S.A. It differs from the latter species in having fewer rows of larger zooecia with shorter opesiae. It also resembles Glauconome prismatica Hagenow and G. canalifera Hagenow, Cretaceous species often referred erroneously to Vincularia (see Berthelsen 1962 : 67-69). Siphonella cyUndrica Hagenow is another Cretaceous form with similar zooecial and zoarial morphology but with a hollow zoarium (see Voigt 1951 : 61, 62). The generic assignment of these so-called vincularian forms can be resolved better by instituting a new genus for them than by distorting the Farciminariid genus Vincularia to accommodate them (see discussion of that genus below) . Family MIGROPORIDAE Gray Genus MICROPORINA Levinsen 1909 Microporina Levinsen : 162. TYPE SPECIES (chosen by Bassler 1935). Cellularia articulata Fabricius 1824 : 27 (as Salicornaria borealis Busk 1855:254, pi. i, figs. 1-3). Recent; Greenland, 73° 20' N., 57° 20' W., 6-10 fms. DIAGNOSIS. Frontal wall membranous, without gymnocyst. Cryptocyst com- pletely developed, shallow, flat, not greatly depressed in region of opesiules, finely and evenly perforate. Opesia approximately co-extensive with orifice, terminal, semi-circular to elliptical, with straight, slightly concave, or slightly convex proximal lip. Opesiules simple, small, circular or elliptical, inconspicuous, paired, one placed on each side of opesia and just proximal to it. Some of lateral cryptocystal pores probably serve as auxiliary opesiules in some species. Avicularia common or rare (or absent?), interzooecial, small, with pivotal bar and pointed rostrum, developed CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 35 between zooecia of the same longitudinal row. Zooecia variable in size and shape, even in a single colony. Zoarium encrusting, erect-unilaminar or erect-jointed. REMARKS. This genus differs from Calpensia in having avicularia, from Micropora in lacking ovicells, and from both in having simple, not tubular opesiules. Kluge (1962 : 322) has done much to clarify the characters of the genus and its type species. Microporina magnipora (Canu) (Text-figs. 13, 14) 1914* Micropora magnipora Canu : 299, pi. 4, figs, i, 2. PI9&3 Steginoporella elegans (Milne Edwards) ; Malecki : in, pi. n, fig. 6. FIGURED SPECIMENS. 0.48451 (Text-fig. 13), 0.48452 (Text-fig. 14). ADDITIONAL MATERIAL. Fifteen specimens, D. 48453-0.48467. DIAGNOSIS. Unilaminar, encrusting to erect Microporina with large, rhomboidal zooecia ; opesiules not much larger than frontal pores ; avicularia lacking (?). DESCRIPTION. Zoarium unilaminar, encrusting or forming erect, flat or hollow- tubular masses, composed of zooecia arranged in regular, occasionally bifurcating, longitudinal rows, those in adjacent rows alternating in position. Zooecia separated by a very faint groove, rhomboidal, but becoming distorted or abortive in some parts of zoarium. Length almost twice width. Gymnocyst lacking. Mural rim finely granular, rounded, narrow distally, widening slightly proximally, raised round distal margin of opesia to form a half-collar. Interzooecial communication by a few large pores in lower half of distal and lateral walls. Cryptocyst complete, extending from proximal margin more than three-quarters of the zooecial length. Surface nearly flat, depressed well below level of mural rim, deepest at centre, evenly perforated by large, quincuncially arranged pores, of which the lateral ones probably served as auxiliary opesiules. Opesiules simple pores, occasionally with slightly raised margins, one placed in each of the two disto-lateral corners of the cryptocyst, just proximal to opesia. Opesia semi-circular, terminal, approximately co-extensive with orifice. Distal margin evenly rounded. Proximal margin nearly straight, but usually broken. Avicularia and ovicells not present. MEASUREMENTS : Lz (10) 0-742 (0-0450) mm., 0-66-0-81 mm. Iz (10) 0-398 (0-0497) mm., 0-34-0-50 mm. ho (6) 0-151 (0-0064) mm., 0-14-0-15 mm. lo (10) 0-180 (0-0102) mm., 0-17-0-20 mm. REMARKS. This species is very similar in both zooecial and zoarial characters to Poropeltarion lebanonense Cheetham (19636 : 50, pi. i, figs. 5, n) and P. newelli 36 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Cheetham (19636 : 51, pi. r, figs. 9, 10) from the Middle and Upper Eocene of Florida, U.S.A., but my suggestion (Cheetham 19636 : 50) that M. magnipora may belong to Poropeltarion has not been substantiated. The clearly differentiated opesiules of M. magnipora indicate that it should be assigned to Microporina despite the absence of avicularia. Dr. H. D. Thomas (personal communication, 1962) reports that a species of Microporina from the London Clay has avicularia only very occasionally. Whether M. magnipora proves to have avicularia or not, it forms a link with the slightly younger American Poropeltarion. DISTRIBUTION. Eocene (Lutetian) ; France. PEocene (Ludian) ; Poland. 13 14 1.00mm 0.500mm FIGS. 13-14. Fig. 13. Microporina magnipora (Canu). 0.48451. Four zooecia from a slightly irregular part of a zoarium, probably near a point of branching. Fig. 14. Microporina magnipora (Canu). 0.48452. Fragmentary zoarium showing more regularly arranged zooecia. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 37 Genus SETOSELLA Hincks 1877 Setosella Hincks : 529. TYPE SPECIES (by original designation). Membranipora vulnerata Busk 1860 : 124, pi. 25, fig. 3. Recent; British Isles. DIAGNOSIS. Frontal wall membranous, with reduced gymnocyst laterally and proximally. Cryptocyst completely developed, imperforate except for a pair of lateral, longitudinal, slit-like opesiules near lateral margins. Opesia small, semi- circular, terminal, co-extensive with orifice. Vibracula (missing in some species?) small, interzooecial, one at the distal end of each zooecium, the vibracular opesia reniform. Ovicell lacking, but fertile zooecia with swollen distal wall exhibiting a small depressed area and usually with a widened opesia. REMARKS. Jullien (1882 : 524, pi. 17, fig. 66) was the first to describe and illus- trate fertile zooecia in the type species of Setosella, but he interpreted them as having entotoichal (= " cassiform ") ovicells. This interpretation led Levinsen (1909 : 196) to establish a family, Setosellidae, to which later authors added Entomaria Duvergier (see discussion of the ovicell of that genus below) and other genera (see Bassler 1953 : 174). Material of 5. vulnerata in the Department of Zoology, British Museum (Natural History), e.g. 99.7.1.1490, Busk Collection, shows the structure of the gonoecium clearly : the " ovicell " is simply a post-oral swelling in the distal wall not separated internally from the zooecium, and the " pore " is simply a frontal depression on the distal part of the gonoecium. Absence of a distinct ovicell, as well as differences in zooecial structure, makes it necessary to remove Setosella from such genera as Lagarozoum, Entomaria, Crater o- pora, and Aspidostoma, for which the family name Aspidostomatidae Jullien is available (see Brown 1952 : 148-153). Setosella can be given temporary accommoda- tion in the Microporidae. Setosella fragilis Canu (Text-figs. 15, 1 6) 19076 Setosella fragilis Canu : 140, pi. 19, fig. 15. 1946 Setosella fragilis Canu; Buge : 433. FIGURED SPECIMEN. 0.48468 (Text-figs. 15, 16). DIAGNOSIS. Erect, cylindrical Setosella arising from encrusting base ; zooecia without gymnocyst ; vibracula lacking ; gonoecia unknown. DESCRIPTION. Zoarium erect, cylindrical, presumably arising from an encrusting base, the zooecia arranged in 8 longitudinal rows, those in adjacent rows alternating in position. Zooecia regularly rhomboidal, rounded distally, straight sided proximally, separated by a faint groove at the crest of the confluent mural rims. Length about one and a half times width. Gymnocyst lacking. Mural rim smooth, probably abraded in the specimen studied, narrow all round the zooecium, not raised distally. 38 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 0.500mm 18 0.500mm 16 1.00mm CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 39 Cryptocyst completely developed, extending from proximal margin more than three-quarters of the zooecial length. Surface somewhat undulating, convex and little depressed in proximal third, concave and moderately depressed in middle third, rising markedly in distal third ; imperf orate and smooth. Opesiules slit-like, unequal, one paralleling each lateral margin, removed slightly from mural rim. Longer opesiule, either right or left, markedly arcuate, extending from near proximal lip of opesia almost the whole length of the cryptocyst. Shorter opesiule less arcuate, extending from near proximal lip of opesia about half the length of the cryptocyst. Opesia semi-elliptical, terminal, approximately co-extensive with orifice. Distal margin evenly rounded. Proximal margin much more broadly rounded. Heterozooecia and gonoecia not observed. MEASUREMENTS : Lz (7) 0-757 (0-0209) mm., 0-70-0-81 mm. Iz (7) 0-486 (0-0340) mm., 0-44-0-54 mm. ho (7) 0-134 (°'°°95) mm., 0-13-0-15 mm. lo (7) 0-184 (°'OI97) mm., 0-15-0-20 mm. REMARKS. My earlier suggestion (Cheetham 19636 : 50) that this species might belong to Poropeltarion has proved erroneous. The imperf orate cryptocyst and slit-like opesiules exclude it from that genus. Although 5. fragilis lacks gonoecia and vibracula, its original placement (Canu 19076 : 140) in Setosella seems to be correct. The species is represented in the Upper Bracklesham collection by the single fragment illustrated. Canu's (19076 : pi. 19, fig. 15) restoration of the opesiules does not indicate their characteristically unequal lengths. DISTRIBUTION. Eocene (Lutetian) ; France. Family PORICELLARIIDAE Harmer Genus PORICELLARIA d'Orbigny 1854 Poricellaria d'Orbigny : 1106. ? 1 8696 Diplodidymia Reuss : 468. TYPE SPECIES (by original designation). Poricellaria alata d'Orbigny 1854 : 1106. Eocene (Lutetian) ; vicinity of Paris, France. DIAGNOSIS. Frontal wall membranous, gymnocyst well developed. Cryptocyst completely developed, perforate or imperforate, with single or multiple, slit-like opesiule lying along one side only. Opesia small, terminal, semi-circular, oblique, approximately co-extensive with orifice. Avicularium adventitious, small, placed FIGS. 15-18. Figs. 15, 16. Setosella fragilis Canu. 0.48468. Views of one zooecium and entire zoarial fragment. Fig. 17. Poricellaria alata d'Orbigny. 0.48469. Frontal view of an internode fragment showing four rows of zooecia. Fig. 18. Poricellaria alata d'Orbigny. 0.48470. View of reverse side of internode fragment with frontal surfaces hidden. 4o CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX on proximal gymnocyst. Ovicell lacking. Zoarium erect, jointed, giving off basal rootlets. Internodes composed of 4 longitudinal rows of zooecia, one row on each face, but with orifices turned so as to open on the same face. REMARKS. The Eocene species of Poricellaria, including P. alata, P. limanowskii (Canu) (19076 : 143, pi. 19, fig. 8) and P. vernoni Cheetham (19636 : 52, pi. i, figs. 7, 8), differ from the Oligocene-Recent species, including P. complicata (Reuss) (see Canu 19146 : 467, pi. 14, figs, n, 12) and P. ratoniensis (Waters) (see Harmer 1926 : 314, pi. 23, figs. 3-8), in having the frontal surfaces of all zooecia hidden when the zoarium is viewed from the reverse side (see P. alata, Text-fig. 18), in having the opesia less oblique to the zooecial axis, and in having the avicularium proportionately smaller. These differences suggest that Diplodidymia, of which P. complicata is the type species, should possibly be reinstated as a subgenus for the Neogene forms. Poricellaria alata d'Orbigny (Text-figs. 17, 18) 1854 Poricellaria alata d'Orbigny : 1106. ?i886 Cellularia diplodidymioides Meunier & Pergens : 3, pi. 2, fig. 3. 1891 Micropora articulata Waters : 14, pi. 2, figs. 5, 6. 19076 Diplodidymia alata (d'Orbigny) Canu : 143, pi. 19, fig. 7. ? 19076 Diplodidymia crassomuralis Canu : 144, pi. 19, fig. 6. ? 19076 Diplodidymia negrisi Canu : 144, pi. 19, figs. 9, 10. J933 Diplodidymia alata (d'Orbigny) ; Dartevelle : 74. 1935 Diplodidymia alata (d'Orbigny) ; Dartevelle : 112, text-fig, i. 1946 Diplodidymia alata (d'Orbigny) ; Buge : 431. ?i94& Diplodidymia crassomuralis Canu ; Buge : 432. ?i946 Diplodidymia negrisi Canu ; Buge : 432. I957 Poricellaria alata d'Orbigny ; Balavoine : 191. 1960 Poricellaria alata d'Orbigny ; Balavoine : 246. 1962 Poricellaria alata d'Orbigny ; Davis : 194. FIGURED SPECIMENS. 0.48469 (Text-fig. 17), 0.48470 (Text-fig. 18). DIAGNOSIS. Poricellaria with frontal surfaces of all 4 zooecial rows visible in obverse aspect, reverse side showing basal surfaces of 2 rows ; opesial bisectrices of all rows make angles of 30°-45° with main axis ; proximal-gymnocystal avicularium small, placed obliquely ; opesiules single or double, distal one circular, proximal one slit-like. DESCRIPTION. Zoarium erect, jointed. Internodes triangular in cross-section, with zooecia arranged in 4 longitudinal rows, all with frontal surface visible in obverse aspect, the zooecia of adjacent rows alternating in position. Proximal end of internode bluntly tapering, formed by 2 zooecia one of which is reduced and " mem- braniporoid ", the other normal but with a single, simple, proximal opening. Distal end not preserved. Zooecia elongate club-shaped, asymmetrical, curved towards mid-line of obverse side. Zooecial length more than twice width. Gymnocyst about half zooecial length, smooth, convex, limited to proximal end of zooecium. Mural rim thin, sharp, smooth. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 41 Cryptocyst completely developed, concave, deepest just proximal to opesia. Sur- face smooth, imperf orate except for opesiule. Opesiule single or double, the distal one circular, the proximal one slit-like, placed on same side of cryptocyst, nearer zoarial mid-line (as viewed frontally). Opesia semi-circular, oblique, its bisectrix making angles of 30°-45° with axis of internode. Proximal lip straight, as high as mural rim. Avicularium adventitious, small, placed on proximal gymnocyst, the chamber oriented longitudinally and opening just below cryptocyst. Rostrum rounded, short, directed obliquely distally and outwards. Pivotal condyles or bar lacking. Proximal portion with a small pit or perforation. Ovicell lacking. MEASUREMENTS : Lz (6) 0-453 (0-0276) mm., 0-43-0-51 mm. Iz (5) 0-185 (0-0167) mm-> 0-16-0-20 mm. ho (5) 0-068 (0-0086) mm., 0-06-0-08 mm. lo (4) 0-086 (o-oooo) mm., 0-09 mm. Lav (4) 0-075 (0-0128) mm., 0-06-0-09 mm- REMARKS. It is impossible to assess Dartevelle's interpretation (1935 : 112-114) of Meunier & Pergens's and Canu's species of Poricellaria as growth stages of P. alata until additional Lutetian and Montian material can be studied in detail. In the Recent P. ratoniensis the zooecia of the proximal internodes bear no resemblance to the normal, mature zooecia (see Harmer 1926, pi. 17, fig. 14), rather than being only slightly different as are P. limanowskii, P. crassomuralis, P. negrisi, P. diplodidy- mioides and P. alata. Preliminary study of specimens from the Sables de Fresville, Gourbesville (Manche), France (collected by Mr. Dennis Curry), suggests that P. limanowskii, at least, is a separate species from P. alata. Waters's (1891 : 15) description of the avicularia of P. alata as vibracula was apparently based on his interpretation of the avicularian proximal pits as vibracular pores. Similar pits occur on the avicularia of P. ratoniensis (see Harmer 1926, pi. 23, figs. 6-8). DISTRIBUTION. PPalaeocene (Montian) ; Belgium. Eocene (Lutetian) ; France. Eocene (Auversian) ; England. Eocene (Ludian) ; Italy. Family STEGANOPORELLIDAE Hincks Genus LABIOPORELLA Harmer Labioporella? dartevellei1 sp. nov. (Text-figs. 19-22) 1851 " Eschara fragilis (Michelin), non Defrance " ; d'Orbigny : 344. 19076 " Smittipora fragilis (d'Orbigny) " ; Canu : 80, pi. n, figs. 5-7. ?*933 Steganoporella fragilis Dartevelle : 74, pi. 2, fig. 2. 7 After the late Edmond Dartevelle. 42 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 1946 " Smittipora fragilis (d'Orbigny) " ; Buge : 431. 1949 " Smittipora fragilis (d'Orbigny) " Balavoine : 773. 1956 " Smittipora fragilis (d'Orbigny) J957 " Smittipora fragilis (d'Orbigny) 1960 " Smittipora fragilis (d'Orbigny) ?I9&2 " Smittipora fragilis (d'Orbigny) Balavoine : 320, 324. Balavoine : 191. Balavoine : 246. Gorodiski & Balavoine : 5. HOLOTYPE. 0.48471 (Text-figs. 19-21). PARATYPES. 0.48472 (Text-fig. 22), 0.48473-75 (3 specimens), L.S.U. 8033. DIAGNOSIS. Labioporella (?) with rudimentary zooecial dimorphism, the avi- cularian (?) zooecia differing only in size from the normal ones ; polypide tube short, wide, slightly eccentric ; mural rim smooth ; zoarium eschariform with edges bevelled owing to compression of zooecia of marginal rows. DESCRIPTION. Zoarium erect, bilaminar, compressed, with zooecia arranged in 2-7 longitudinal rows on each side, those in adjacent rows alternating in position. Number of zooecial rows increases distally by bifurcation. Lateral margins of zoarium with bevelled appearance produced by compression of zooecia of marginal rows. Zooecia rhomboidal to sub-rectangular, not markedly dimorphic, with evenly rounded distal borders, separated by a faint groove. Length about twice width. Mural rim smooth, thin and rather sharp distally, thicker and rounded proximally. Gymnocyst lacking. Cryptocyst broad, fully developed, flat and shallow proximally, concave and steeply descending distally, evenly perforated with numerous, small, quincuncially arranged pores. Distal portion reflected upwards, except at lateral extremities, to form short, wide, slightly eccentric polypide tube with fully developed basal portion. Opesia semi-circular, sub-terminal, with very narrow distal shelf. Opesiular indentations shallow, not well defined, unequal. Avicularian (?} zooecia slightly larger than ordinary ones, but not different in form. Ovicells lacking. MEASUREMENTS : Ordinary zooecia Lz (9) 0-577 (0-0317) mm., 0-53-0-65 mm. Iz (9) 0-264 (O'°304) mm., 0-20-0-30 mm. ho (6) 0-121 (0-0114) mm., 0-10-0-13 mm. lo (6) 0-131 (0-0279) mm., 0-10-0-17 mm. Avicularian (?} zooecia Lz (2) 0-761 (0-0605) mm., 0-72-0-80 mm. Iz (2) 0-325 (0-0242) mm., 0-31-0-34 mm. ho (i) 0-19 mm. lo (i) 0-23 mm. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 43 f> 20 1.00mm 0.500mm FIGS. 19-22. Figs. 19-21. Labioporellal dartevellei sp. nov. 0.48471. Holotype. Front and edge views of fragmentary zoarium and details of three ordinary zooecia. Edge view shows bevelling produced by compression of marginal zooecia. Fig. 22. Labioporellal dartevellei sp. nov. 0.48472. Paratype. Avicularian (?) zooecium. 44 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX REMARKS. This widespread Eocene species has been a nomenclatorial puzzle for at least a hundred years since d'Orbigny mis-identified it with Michelin's (1845 : 176) material which he considered to have been in turn mis-identified with Defrance's (1828 : 428) species. Whether Michelin and Defrance had the same species or not, both of them had a different one from the present species : their illustrations show cylindrical, tubular zoaria with zooecia having sub-central opesiae. Canu (19076 : 80, pi. II, figs. 5-7) re-described and, for the first time, illustrated d'Orbigny's material but, unfortunately, did not re-name it. Dartevelle (1933 : 74, pi. 2, fig. 2), who possibly had the same species, coined a new, but unfortunately homonymous, name for it. French Lutetian specimens (Canu, 19076 : pi. n, figs. 5-7) have many of the zooecia closed, especially along the zoarial margins, or with degenerated cryptocysts. Zooecia in the latter condition may represent fertile zooecia of the type noted by Cook (1964 : 51, 52) in various species of Steganoporella ; cryptocystal degeneration was not, however, noted by her in Labioporella. This species does not fit well in either Labioporella. or Steganoporella but seems closer to the former because of the asymmetrical polypide tube and the lack of a wide distal shelf in the avicularian (?) zooecia. DISTRIBUTION. Eocene (Lutetian) ; France, PSenegal. PEocene (Auversian) ; Belgium. Genus GAUDRYANELLA Canu 1900 Gaudryanella Canu : 380 (nomen nudum). 19076 Gaudryanella Canu : 141. TYPE SPECIES (by monotypy). Gaudryanella variabilis Canu 19076 : 142, pi. 19, figs. 17-20. Eocene (Lutetian) ; vicinity of Paris, France. DIAGNOSIS. Frontal wall membranous, gymnocyst lacking. Cryptocyst imper- forate, fully developed, markedly concave, descending steeply to basal wall distally so as to form a small, eccentric polypide tube. Opesia large, terminal, with irregular, unequal opesiular indentations. Zooecial dimorphism not apparent. Ovicell lacking. REMARKS. The " large, ovarian zooecia ", mentioned by Canu (19076 : 141) in the original diagnosis, may have been formed through wear or degeneration of the cryptocyst. The absence of cryptocystal pores and zooecial dimorphism in this genus distinguishes it from Steganoporella and Labioporella ; the lack of a gymnocyst distinguishes it from Siphonoporella. Gaudryanella variabilis Canu (Text-figs. 23, 24) 19076 Gaudryanella variabilis Canu : 142, pi. 19, figs. 17—20. 1925 Gaudryanella variabilis Canu ; Canu : 47. 1946 Steganoporella variabilis (Canu) Buge : 432. 1956 Gaudryanella variabilis Canu ; Balavoine : 320, 324. !957 Steganoporella variabilis (Canu) ; Balavoine : 191. 1960 Gaudryanella variabilis Canu ; Balavoine : 246. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 45 1.00mm FIGS. 23-27. Fig. 23. Gaudryanella variabilis Canu. 0.48476. Two complete zooecia and a partially developed or degenerated one. Fig. 24. Gaudryanella variabilis Canu. 0.48477. Fragmentary zoarium showing arrangement of zooecia. Fig. 25. Entomaria dutempleana (d'Orbigny). 0.48482. Three mature zooecia and an avicularium. Fig. 26. Entomaria dutempleana (d'Orbigny). 0.48483. A young zooecium with the spine bases not yet overgrown by conuli. Fig. 27. Entomaria dutempleana (d'Orbigny). 0.48484. Ovicelled zooecium and part of the distal zooecium. 46 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX FIGURED SPECIMENS. 0.48476 (Text-fig. 23), 0.48477 (Text-fig. 24). ADDITIONAL MATERIAL. Four specimens, D . 48478-0 . 48481. DIAGNOSIS. As for the genus (the only species). DESCRIPTION. Zoarium uni- or bilaminar, erect, composed of wide fronds with zooecia arranged in 8 or more longitudinal rows, those in adjacent rows alternating in position. Zooecia rhomboidal, rounded distally, straight proximally, separated by a very faint groove. Length and width subequal. Gymnocyst lacking. Mural rim thin, sharp, smooth, not distinctly raised. Cryptocyst concave, deeply depressed at its distal end where it extends to basal wall to form a small, short, eccentric polypide tube. Surface smooth, imperforate. Opesia large, asymmetrical, semi-circular, terminal. Opesiular indentations broad, shallow, unequal. Heterozooecia and ovicells lacking. MEASUREMENTS : Lz (10) 0-363 (0-0345) mm., 0-32-0-40 mm. Iz (10) 0-354 (°'°474) mm., 0-27-0-44 mm. ho (10) 0-131 (0-0128) mm., 0-11-0-14 mm. lo (10) 0-189 (0-0272) mm., 0-15-0-21 mm. REMARKS. The only suggestion of departure from the morphology of the ordinary zooecia in the Upper Bracklesham material is in a partially developed or degenerate zooecium (Text-fig. 23) in one fragment and several slightly reduced primoserial zooecia in another. None of these seems to fit the description given by Canu (19076 : 141) of " ovarian " zooecia. DISTRIBUTION. Eocene (Lutetian) ; France. Family ASPIDOSTOMATIDAE Jullien Genus ENTOMARIA Duvergier 1921 Entomaria Duvergier : 150 (introduced in faunal list). 1927 " Entomaria Canu (in Duvergier) " ; Canu & Lecointre : 45. TYPE SPECIES (by monotypy). Rhagasostoma spiniferum Canu 19146:469, pi. 15, fig. 7. Oligocene (Stampian) ; Gaas (Gironde), France. DIAGNOSIS. Frontal wall membranous, gymnocyst narrow but encircling zooe- cium. Cryptocyst extensive, imperforate, merging with mural rim. Distal margin of mural rim with spine bases, often overgrown by conule-like extensions of distal gymnocyst. Opesia with lateral opesiular indentations and proximal cryptocyst- flange (= false polypide tube) . Avicularia interzooecial, with opesiulated cryptocyst. Ovicell entozooecial. REMARKS. This genus, often erroneously attributed to Canu (see Canu & Lecointre 1927 : 45), has been considered congeneric with Lagarozoum (see Bassler 1953 : 174), CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 47 from which it differs in lacking distinct cryptocyst ridges and in having oral spines, though these are often overgrown by projections of the distal gymnocyst forming structures similar to the conuli of Lagarozoum (see Harmer 1926 : 325, pi. 22, fig. 6). The ovicell of Entomaria, a typical entozooecial chamber opening above the opesia and extending distally under the gymnocyst and proximal part of the cryptocyst of the next zooecium, has been misinterpreted as entotoichal (see Canu & Lecointre 1927 : 45)- Entomaria dutempleana (d'Orbigny) (Text-figs. 25-27) 1851 Semieschara dutempleana d'Orbigny : 366. 1900 Rhagasostoma dutempleanum (d'Orbigny) Canu : 428, pi. 7, figs. 22, 23. 19076 Rhagasostoma dutempleanum (d'Orbigny) ; Canu : 138, pi. 5, fig. 12. 1915 Rhagasostoma dutempleanum (d'Orbigny) ; Canu : 297. 1925 Entomaria dutempleana (d'Orbigny) Canu : 47. 1927 Entomaria dutempleana (d'Orbigny) ; Canu & Lecointre : 45. 1935 Entomaria dutempleana (d'Orbigny) ; Dartevelle : 115. 1937 Entomaria dutempleana (d'Orbigny) ; Dartevelle : no. 1946 Entomaria dutempleana (d'Orbigny) ; Buge : 433. 1957 Entomaria dutempleana (d'Orbigny) ; Balavoine : 191. 1960 Entomaria dutempleana (d'Orbigny) ; Balavoine : 246. FIGURED SPECIMENS. 0.48482 (Text-fig. 25), 0.48483 (Text-fig. 26), 0.48484 (Text-fig. 27). ADDITIONAL MATERIAL. Forty specimens, 0.48485-0.48524. DIAGNOSIS. Entomaria with small zooecia and narrow gymnocyst ; ovicell small, without large lateral slits ; avicularia initiate new zooecial rows. DESCRIPTION. Zoarium encrusting or erect, uni- or bilaminar, forming irregular, convoluted, hollow, tubular masses, the zooecia arranged more or less in longitudinal rows, those in adjacent rows alternating in position. Zooecia ideally rhomboidal, straight sided, but often distorted or abortive, separated by distinct, sometimes deep furrows. Length and width subequal. Gymnocyst very narrow, but present entirely round zooecium. Mural rim wide, rounded, irregularly crenulated or beaded ; distal margin with 4-6 large, hollow spine bases, the more distal ones larger than the others ; distal gymnocyst grows over all but two of the spines to form a series of irregular hummocks or conuli. Zooecia com- municate by simple pores placed near base of distal and lateral walls. Cryptocyst extensive, covering nearly three-quarters of zooecial length. Surface nearly flat, highest near centre, deepest along lateral margins, imperforate, coarsely tuberculate, not distinctly marked off from mural rim. Distal part, along proximal lip of opesia, slightly raised to form a very narrow flange or false polypide tube. Opesia semi-elliptical, much wider than high ; distal margin evenly and broadly rounded, provided with a very narrow distal shelf ; proximal margin scalloped, the broad cryptocystal flange limited on each side by a small, rounded opesiular indenta- tion. 48 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Avicularia interzooecial, initiating zooecial rows. Mural rim like that of zooecia but raised and pointed distally. Cryptocyst like that of zooecia but with one or two opesiules near proximal end. Opesia distal, oval. Spines and conuli absent. Ovicell entozooecial, globular, opening over the opesia, the surface a tuberculate, imperf orate awning projecting from the gymnocyst of the distal zooecium and discernible as a swelling under the proximal part of its gymnocyst. Ovicelled zooecia larger than ordinary ones, with larger opesia and without spines or conuli. MEASUREMENTS : Ordinary zooecia Lz (10) 0-494 (0-0623) mm., 0-41-0-62 mm. Iz (10) 0-423 (0-0417) mm., 0-37-0-51 mm. ho (10) 0-122 (0-0274) mm., 0-09-0-16 mm. lo (10) 0-208 (0-0202) mm., 0-18-0-25 mm. Ovicelled zooecia Lz (3) 0-596 (0-0623) mm., 0-53-0-65 mm. Iz (3) 0-445 (0-0534) mm., 0-40-0-50 mm. ho (3) 0-140 (0-0049) mm., 0-14-0-15 mm. lo (3) 0-242 (0-0178) mm., 0-22-0-26 mm. Avicularia Lav (7) 0-355 (0-0761) mm., 0-27-0-48 mm. lav (7) 0-227 (0-043°) mm., 0-17-0-29 mm. REMARKS. Canu's description of the ovicell of this species as entotoichal (1925 : 47) cannot be verified in the present material. Though the opening of the ovicell is distal to the opesia, it is not far removed, and the ovicell chamber is immersed in the distal zooecium and covered by its gymnocyst and the proximal part of its cryptocyst. DISTRIBUTION. Eocene (Lutetian) ; France, Belgium. Family FARCIMINARHDAE Busk Genus NELLIA Busk Nellia tenella (Lamarck) (Text-fig. 28) 1816 Cellaria tenella Lamarck : 135. 1851 Cellaria tenella Lamarck ; d'Orbigny : 28. 1851 Cellaria quadrilatera d'Orbigny : 29. 18520 Salicornaria dichotoma Busk : 367 ; non Schweigger, 1819. 18526 Nellia oculata Busk : 18, pi. 64, fig. 6 ; pi. 65, fig. 4. 1873 Nellia oculata Busk ; Smitt : 3, pi. i, figs. 53, 54. 1880 Nellia oculata Busk ; MdcGillivray : 51, pi. 49, figs. 5, $a, b. 1 88 1 Nellia oculata Busk ; Haswell : 36. 1883 Membranipora oculata (Busk) Waters : 434. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 49 1887 Farcimia oculata (Busk) Waters : 92. 1887 Nellia oculata Busk ; Hincks : 121. 1895 Farcimia oculata (Busk) ; MacGillivray : 50, pi. 6, figs. 6, 7. 1905 Nellia oculata Busk ; Thornely : no. 1905 Farcimia quadrilatera (d'Orbigny) Waters : 3. 1907 Nellia oculata Busk ; Thornely : 185. 19076 Farcimia tenella (Lamarck) Canu : 76, pi. 10, figs. 36, 37. 19076 Farcimia bituberculata Canu : 73, pi. 10, figs. 24-29. 1909 Farcimia oculata (Busk) ; Waters : 167. 1909 Nellia tenella (Lamarck) ; Levinsen : 120, pi. i, figs. 13 a-e. 1912 Farcimia oculata (Busk) ; Canu : 191, pi. 10, figs. 16, 19. 1913 Farcimia oculata (Busk) ; Waters : 489, pi. 67, figs. 8, 9. 1914 Nellia oculata Busk ; Osburn : 191. 1916 Farcimia bituberculata Canu ; Faura y Sans & Canu : 298. 1920 Nellia bifaciata Canu & Bassler : 197, pi. 32, figs. 15-19. 1920 Nellia oculata Busk ; Marcus : 5. 1921 Farcimia tenella (Lamarck) ; Duvergier : 8. 1921 Farcimia oculata (Busk) ; Robertson : 45. 1921 Nellia oculata Busk ; Marcus : 3. 1922 Nellia oculata Busk ; Marcus : 423. 1923 Nellia oculata Busk ; Canu & Bassler : 55, pi. 2, figs. 5-7. 1926 Nellia oculata Busk ; Harmer : 240, pi. 14, figs. 18-20. 1927 Nellia oculata Busk ; Osburn : 125. 1928 Nellia oculata Busk ; Canu & Bassler : 26. 19296 Nellia oculata Busk ; Canu & Bassler : 185, pi. 5, figs. 12, 13. 1932 Nellia oculata Busk ; Hastings : 410. 1934 Nellia tenella (Lamarck) ; Davis : 220. !935 Nellia oculata Busk ; Dartevelle : 115. J935 Nellia bituberculata (Canu) Dartevelle : 115. 1940 Nellia oculata Busk ; Osburn : 400. 1941 Nellia oculata Busk ; McGuirt : 66, pi. 2, fig. 5. 1941 Nellia oculata Busk ; Silen : 49. 1944 Nellia ocullata [sic] Busk ; Pokorny : 1-14. 1946 Farcimia tenella (Lamarck) ; Buge : 433. 1946 Farcimia bituberculata Canu ; Buge : 433. 1947 Nellia oculata Busk ; Osburn : 25. 1949 Farcimia bituberculata Canu ; Balavoine : 773. 1949 Nellia oculata Busk ; Vigneaux : 32, pi. i, figs. 10-12. 1950 Nellia oculata Busk ; Osburn : 119, pi. 13, fig. 4. 1956 Nellia tenella (Lamarck) ; Balavoine : 321. J957 Nellia bituberculata (Canu) ; Balavoine : 191. !957 Nellia tenella (Lamarck) ; Cheetham : 93. 1959 Nellia oculata Busk ; Lagaaij : 482, text-fig, i. 1960 Nellia tenella (Lamarck) ; Balavoine : 246. 1962 Nellia tenella (Lamarck) ; Cheetham : 326, pi. i, figs. 4, 5. ?I962 Nellia tenella (Lamarck) ; Gorodiski & Balavoine : 6. ?i962 Farcimia bituberculata Canu ; Gorodiski & Balavoine : 6. 1962 Nellia oculata Busk ; Ghiurca, table i. 1963 Nellia tenella (Lamarck) ; Braga : 27. 19636 Nellia tenella (Lamarck) ; Cheetham : 59, pi. i, fig. 14,- FIGURED SPECIMEN. 0.48525 (Text-fig. 28). ADDITIONAL MATERIAL. Five specimens, 0.48526-0.48530. GEOL. 13, I. 4 50 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX » m J 29 0.500mm CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 51 DIAGNOSIS. Nellia with relatively narrow zooecia of same width on all 4 faces of zoarium ; avicularia paired, placed in proximo-lateral corners of gymnocyst, small, sometimes vestigial ; proximal end of internode with I large and 2 smaller pores ; distal end with 2 groups of such pores. DESCRIPTION. Zoarium erect, delicate, articulated, dichotomously branching at nodes. Internodes long, slender, square in cross-section, the zooecia arranged in 4 longitudinal rows, those in adjacent rows alternating in position. Zooecia on all 4 faces subequal. Tips of articulating ends not preserved, but proximal portion consists of 2 normal and 2 greatly reduced zooecia. Zooecia elongate, club-shaped, separated by distinct grooves. Length more than twice width. Gymnocyst smooth, convex, extending about a third of the zooecial length from the proximal margin. Mural rim smooth, sharp, slightly raised. Cryptocyst a narrow crescent on proximal margin of opesia ; surface smooth, slightly concave, imperforate. Avicularia adventitious, small, paired except on zooecia just distal to reduced zooecia at proximal end of internode, placed in proximo-lateral corners of gymnocyst, the rostrum directed distally and slightly outwards. Pivotal structures lacking in material at hand. Proximal part of avicularian chamber sometimes with a small pore. Ovicell entozooecial, discernible as a slight swelling under gymnocyst of distal zooecium. Frontal depression lacking in material at hand. Opesia of ovicelled zooecium slightly enlarged. MEASUREMENTS : Ordinary zooecia Lz (8) 0-525 (0-0214) mm., 0-50-0-56 mm. Iz (8) 0-205 (0-0216) mm., 0-17-0-24 mm. ho (8) 0-358 (0-0367) mm., 0-32-0-41 mm. lo (8) 0-119 (0-0140) mm., 0-10-0-14 mm- Ovicelled zooecia Lz (i) 0-57 mm. ho (i) 0-42 mm. Iz (i) 0-26 mm. lo (i) 0-18 mm. Avicularia Lav (10) 0-067 (0-0150) mm., 0-04-0-09 mm. REMARKS. Although it is true, as Harmer (1926 : 242) has remarked, that the description given by Lamarck (1816 : 135) for Cellaria tenella is insufficient to separate it from other jointed species, the probability that Lamarck had this ubiquitous FIGS. 28-29. Fig. 28. Nellia tenella (Lamarck) . 0.48525. One face of broken internode showing four complete zooecia and a broken one. Proximal zooecium is reduced ; distal complete zooecium is ovicelled. Fig. 29. Nellia ventricosa (Canu). 0.48532. One face of broken internode showing three zooecia and a small portion of a fourth, distal one. 52 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX warm-water Cheilostomc in his collection is very great. Moreover, even if Harmer was correct in discarding C, tenella from its synonymy, the next oldest available name is C. quadrilatera d'Orbigny, not Nellia oculata Busk. Farcimia bituberculata Canu appears, from Canu's figures (19076, pi. 10, figs. 24-29), to be the same as Harmer's (1926 : 242) " typical form " of AT. tenella with vestigial avicularia. The type specimens of N. bifaciata Canu & Bassler, U.S. National Museum Cat. Nos. 63954, 63955, have zooecia of different widths on different fragments, not on different faces of the same fragment, and are thus conspecific with N. tenella. The Upper Bracklesham specimens differ from Recent N. tenella in having the functional avicularia without pivotal bar and the ovicell without the frontal depres- sion (see Harmer 1926, pi. 14, figs. 19, 20 ; Osburn 1950, pi. 13, fig. 4). American Eocene specimens have the depression on the ovicell and some of the avicularia vestigial (see Cheetham 1962, pi. i, figs. 4, 5). DISTRIBUTION. Eocene (Lutetian) ; France, Spain, PSenegal. Eocene-Oligocene (Claibornian, Jacksonian, Vicksburgian) ; Gulf Coast of U.S.A. Miocene ; Egypt, Jamaica, Australia. Recent ; tropical and warm-temperate zones of all oceans. Nellia ventricosa (Canu) (Text-fig. 29) 1914^ Quadricellaria ventricosa Canu : 298, pi. 4, figs. 6, 7. ?I962 Nellia ventricosa (Canu) Gorodiski & Balavoine : 6. FIGURED SPECIMEN. 0.48532 (Text-fig. 29). DIAGNOSIS. Nellia with short, stubby internodes, the proximal ends of which are formed by 2 opposing zooecia, each with a single, terminal opening ; gymnocyst and opesia of about equal length ; cryptocyst well developed ; avicularia large, paired, with chambers meeting on proximal gymnocyst and opening over opesiae of adjoining rows. DESCRIPTION. Zoarium erect, jointed, probably branching at nodes. Internodes short, square in cross-section, with zooecia arranged in longitudinal rows, one row of 3-4 zooecia on each face, the zooecia of adjacent rows alternating in position. Proximal end of internode bluntly tapering, formed by 2 opposing zooecia narrower than more distal ones. One simple, terminal opening occurs at proximal end of each of the 2 proximal zooecia. Distal end of internode not well preserved. Zooecia club-shaped, separated by furrows ; length nearly twice width. Gymno- cyst about half zooecial length, arched, smooth, developed proximally only, covered in all but the proximal zooecia by enlarged avicularian chambers. Cryptocyst imperf orate, smooth, well developed but narrow, horse-shoe shaped, limited to proximal and lateral margins ; not widened proximally. Surface of cryptocyst a smooth, horizontal shelf. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 53 Opesia oval, more broadly rounded proximally, with a narrow shelf, lower than the cryptocyst, developed distally. Avicularia adventitious, paired, with enlarged chambers meeting on mid-line of proximal gymnocyst and curving distally and outwards to open on disto-lateral margins of opesiae of adjoining rows. Rostrum short, pointed, directed proximally. Pivotal bar complete. Chamber with a distinct proximal pore. Ovicell unknown. MEASUREMENTS : Lz (3) 0-581 (0-0226) mm., 0-56-0-60 mm. lz (3) °*3I9 (0-0522) mm., 0-27-0-38 mm. h° (3) 0-359 (0-0171) mm., 0-34-0-38 mm. lo (3) 0-185 (°'°43°) mm., 0-14-0-23 mm. Lav (4) 0-090 (0-0086) mm., 0-09-0-10 mm. REMARKS. This species resembles N. appendiculata (Hincks) (Department of Zoology, British Museum (Natural History), 99.5.1.506, Hincks Collection) and N. tennis Harmer (holotype, Department of Zoology, B.M. (N.H.), 28.3.6.62, Siboga Collection), Recent Indo-Pacific species, in zooecial size and shape, zoarial articulation, and avicularian size, shape, and position. Moreover, it appears to agree with N. appendiculata in having short, stubby internodes. Neither of the Recent species, however, has the avicularian chambers so enlarged as to meet over the proximal gymnocyst as they do in N. ventricosa, and further in both of those species the cryptocyst widens proximally. Harmer (1926 : 246) deemed differences of this magnitude sufficient to separate the two Recent species, so N. ventricosa should probably be maintained as a separate species. N. midway anica Canu & Bassler (1920 : 197, pi. 4, figs. 10-15), from the Palaeocene of the Gulf Coast of the U.S.A., differs from N. ventricosa in having longer internodes, less-developed avicularian chambers, and no distal shelf. DISTRIBUTION. Eocene (Lutetian) ; France, PSenegal. Genus VINCULARIA Def ranee 1829 Vincularia Defrance : 214. 19076 Heterocella Canu : 70 (objective). TYPE SPECIES (by monotypy). Vincularia fragilis Defrance 1829 : 214, pi. 45, figs. 3, 3«, b. Eocene (Lutetian) ; vicinity of Paris, France. DIAGNOSIS. Zoarium erect, probably articulated, the internodes quadriserial, slender, with 3 openings in each proximal end and 6 in each distal end. Zooecia dimorphic or trimorphic, similar zooecia occurring in 2 adjacent series : (i) ordinary zooecia (= " c-zooecia " of Canu 19076 : 71) small, with symmetrical opesiae ; (2) avicularian (?) zooecia (= " D-zooecia " of Canu) usually larger, with asymmetrical opesiae, outwardly curving distal margins, and usually distinct distal cryptocystal shelves ; and (3) ovicelled zooecia (lacking in some species) large, with symmetrical opesiae surmounted by the entozooecial ovicells. Zooecia of all three types with 54 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX membranous frontal wall, lacking gymnocyst. Cryptocyst a narrow, imperf orate shelf round lateral and proximal margins of the large, oval opesia. Communication between zooecia of the same series by a large, median septule ; between zooecia of adjacent series by a septule connecting disto-lateral wall of one with proximo-lateral wall of the other. Small, single, adventitious avicularium present on mural rim of some species. REMARKS. Brown (1952 : 90, 91) remarked on the nomenclatorial inconsistency of declaring Vincularia an unrecognizable name while retaining its absolute junior synonym, Heterocella Canu (see Bassler 1935 : 125, 225). The genus Vincularia apparently ranges from Middle to Upper Eocene (Lutetian, Auversian, Bartonian) both in England and on the Continent. It is yet unknown from America or, with certainty, from strata of Oligocene age. Heterocella vicksbur- gica Canu & Bassler (1920 : 198, pi. 82, figs. 11-14), from the Oligocene of Alabama, U.S.A. ; Vincularia hians Reuss and V. haidingeri Reuss (Kyri 1951 : 74, 75), from the Eocene (Ludian) of Rumania ; Heterocella lerichei Canu (19070 : 512, pi. 12, fig. 2), from the Oligocene (Stampian) and Miocene (Aquitanian) of France ; and Heterocella pentagona Canu & Bassler (19296 : in, pi. 9, figs. 13-16), from the western Pacific, do not appear to have the avicularian (?) or " D " zooecia characteristic of the genus. Dartevelle (1942 : 149) stated that Glauconome tetragona Miinster is " an undoubted " Vincularia but gave no evidence to support his claim. The six species included in the following table are the only ones which can at present be placed in the genus with certainty. KEY TO SPECIES OF VINCULARIA (Species named in square brackets are not described here.) 1 Zooecia trimorphic ; ovicelled zooecia present ....... 2 Zooecia dimorphic ; ovicelled zooecia absent ....... 4 2 Ovicelled zooecia greatly enlarged ; avicularian zooecia separated by a thread ; adventitious avicularia on zooecia of all three types . . V. monstruosa (Canu) Ovicelled zooecia subequal to ordinary zooecia ; avicularian zooecia separated by a groove ... 3 3 Adventitious avicularia on ordinary and ovicelled zooecia only [V. polymorpha (Canu)] Adventitious avicularia on zooecia of all three types . [V. lediensis (Dartevelle)] 4 Adventitious avicularia present ; avicularian zooecia curved outwards strongly [V.fragilis Def ranee] Adventitious avicularia absent ; avicularian zooecia curved outwards only slightly 5 5 Zooecia of both types slender, with elongate cryptocyst ... V. davisi sp. nov. Zooecia of both types obese, with short cryptocyst . . [V. subsymmetrica (Canu)] i. Vincularia monstruosa (Canu) (Text-figs. 30-32) 19076 Heterocella monstruosa Canu : 71, pi. 10, figs, n, 12. 1946 Heterocella monstruosa Canu : Buge : 430. 1956 Vincularia monstruosa (Canu) Balavoine : 324. 1960 Vincularia monstruosa (Canu) ; Balavoine : 246. ?I9&2 Heterocella sp., Davis : 194. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 55 FIGURED SPECIMENS. 0.48533 (Text-fig. 30), 0.48534 (Text-fig. 31), 0.48535 (Text-fig. 32). ADDITIONAL MATERIAL. Ninety-one specimens, 0.48536-0.48626. DIAGNOSIS. Vincularia with trimorphic zooecia, the avicularian (?) ones curving away from each other markedly and separated by a thread ; ordinary and avicularian (?) zooecia both with very short cryptocysts ; adventitious avicularia at inner proximo-lateral angle of ordinary zooecia and at outer proximo-lateral angle of '--4 • '•--*"•, lit 33 34 1.00mm 1.00mm FIGS. 30-34. Fig. 30. Vincularia monstruosa (Canu). 0.48533. Internode fragment showing the two series of ordinary zooecia. Fig. 31. Vincularia monstruosa (Canu). D. 48534. Internode fragment showing the two series of avicularian (?) zooecia. Fig. 32. Vincularia monstruosa (Canu). 0.48535. Internode fragment showing a series of three avicularian (?) zooecia giving rise to a series of three ovicelled zooecia. Fig. 33 . Vincularia davisi sp. nov. 0.48627. Holotype. Internode fragment showing the two series of ordinary zooecia. The distal zooecium on the left has a pair of pores possibly associated with articulation. Fig. 34. Vincularia davisi sp. nov. 0.48628. Paratype. Inter- node fragment showing the two series of avicularian (?) zooecia. 56 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX avicularian (?) and ovicelled zooecia ; ovicelled zooecia greatly enlarged, with well- developed entozooecial ovicells. DESCRIPTION. Zoarium erect, probably articulated, composed of long, slender, curved internodes, tapering towards the base, lozenge-shaped in cross-section, the faces separated by obtuse angles having similar zooecia. Zooecia of adjacent rows alternate in position. Length of internode at least 7 zooecia. Proximal end of internode with 3 openings, 2 in avicularian (?) zooecia, I in ordinary zooecium. Distal end of internode with 6 openings, 2 in each of 2 avicularian (?) zooecia, i in each of 2 ordinary zooecia. Ordinary zooecia rounded rectangular, separated by their narrow, smooth, rounded mural rims. Length about twice width. Interior very shallow, with a large, circular septule centrally, 2 similar ones disto-laterally, and 2 smaller ones near middle of lateral walls. Cryptocyst smooth, imperforate, slightly concave, forming a narrow, crescent-shaped shelf proximally and laterally. A similar, but deeper shelf runs distally from central septule to distal mural rim, sometimes connecting with the proximal cryptocyst to form a closure. Opesia oval, more broadly rounded distally than proximally, without differentiated indentations. Avicularian (?) zooecia asymmetrically club-shaped, the longitudinal axes of those in one series curving strongly away from those of the other. Zooecia of both series separated by their narrow, smooth, rounded mural rims which are narrowest at the outer distal corner. Zooecial length twice width. Interior with same appearance and arrangement of pores as ordinary zooecia. Cryptocyst as in ordinary zooecia. Opesia as in ordinary zooecia but asymmetrical. Ovicelled zooecia occur at distal end of series of avicularian (?) zooecia, rounded rectangular, larger than zooecia of other types, separated by narrow, rounded, smooth mural rims. Length slightly greater than width. Cryptocyst greatly reduced. Interior as in ordinary zooecia, except at distal end where distal wall is concave and slightly elevated, forming an entozooecial ovicell which extends slightly under cryptocyst of distal zooecium. Avicularia adventitious, very small, single, placed proximo -laterally on mural rim with rounded rostrum directed distally and outwards. Ordinary zooecia with avicularium at inner proximal angle ; zooecia of other two types with avicularium at outer proximal angle. MEASUREMENTS : Ordinary zooecia Lz (9) 0-440 (0-0350) mm., 0-36-0-47 mm. Iz (10) 0-206 (0-0219) mm., 0-17-0-23 mm. ho (8) 0-346 (0-0330) mm., 0-27-0-38 mm. lo (10) 0-141 (0-0092) mm., 0-13-0-15 mm. Lav (9) 0-056 (0-0128) mm., 0-04-0-08 mm. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 57 Avicularian (?) zooecia Lz (10) 0-471 (0-0311) mm., 0-41-0-53 mm. Iz (9) 0-278 (0-0196) mm., 0-26-0-32 mm. ho (10) 0-362 (0-0214) mm., 0-34-0-41 mm. lo (9) 0-176 (0-0284) mm., 0-13-0-21 mm. Lav (6) 0-078 (0-0084) mm., 0-07-0-09 mm. Ovicdled zooecia Lz (3) 0-522 (0-1150) mm., 0-39-0-61 mm. Iz (2) 0-415 (0-0191) mm., 0-40-0-43 mm. h° (3) 0-504 (0-0964) mm., 0-39-0-56 mm. Lav (i) 0-09 mm. REMARKS. It is very likely that this abundant Upper Bracklesham species is the one which Davis (1962 : 194) identified from Selsey as Heterocella sp. DISTRIBUTION. Eocene (Lutetian) ; France. PEocene (Auversian) ; England. 2. Vincularia davisi8 sp. nov. (Text-figs. 33, 34) HOLOTYPE. 0.48627 (Text-fig. 33). PARATYPES. 0.48628 (Text-fig. 34), 0.48629-0.48640 (12 specimens), and L.S.U. 8034. DIAGNOSIS. Vincularia with dimorphic zooecia, the avicularian (?) ones curving away from each other only slightly so that they are not much different in form from the ordinary ones ; avicularian (?) zooecia separated by threads ; ordinary zooecia with very elongate proximal cryptocyst ; adventitious avicularia and ovicells lacking. DESCRIPTION. Zoarium erect, probably articulated, composed of long, very slender internodes, nearly square in cross-section ; similar zooecia occur in each of the two adjacent series, alternating in position. Articulating ends of internodes not preserved. Ordinary zooecia club-shaped, nearly three times as long as wide, separated by contiguous, broad, smooth, rounded mural rims. Interior shallow, with a large, circular, distal-median septule flanked on either side by a smaller one. Cryptocyst smooth, imperforate, slightly concave, forming a long shelf proximally and a very narrow one on each lateral margin. Distal wall without distinct shelf. Closures formed by proximal growth of a lamina from distal mural rim. Opesia oval, more broadly rounded distally than proximally, without opesiular indentations. Avicularian (?} zooecia club-shaped, only slightly asymmetrical, the longitudinal axes of those in one series curving just perceptibly away from those in the other. Zooecia of the two series separated by their contiguous, smooth, rounded mural rims. 8 After the late A. G. Davis. 58 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Zooecial length two and a half times width. Septules as in ordinary zooecia. Cryp- tocyst as in ordinary zooecia, but wider laterally and continuing round distal margin of opesia in a broad shelf. Opesia oval, nearly symmetrical. Adventitious avicularia and ovicell lacking. MEASUREMENTS : Ordinary zooecia Lz (7) 0-602 (0-0517) mm., 0-56-0-70 mm. Iz (7) 0-211 (0-0128) mm., 0-19-0-23 mm. ho (7) 0-372 (0-0119) mm., 0-36-0-39 mm. lo (7) 0-130 (0-0134) mm., 0-11-0-15 mm. Avicularian (?) zooecia Lz (6) 0-789 (0-0515) mm., 0-70-0-85 mm. Iz (7) 0-300 (0-0277) mm., 0-26-0-32 mm. ho (6) 0-370 (0-0467) mm., 0-31-0-43 mm. lo (7) 0-145 (0-0131) mm., 0-13-0-17 mm. REMARKS. This species is most closely related to the French Lutetian V. sub- symmetrica (Canu) which it resembles in having dimorphic zooecia with the avicul- arian (?) zooecia only slightly curved and in lacking adventitious avicularia. The zooecia of V. subsymmetrica contrast with those of V. davisi in being obese and in having short cryptocysts. Suborder ACANTHOSTEGA Family CRIBRILINIDAE Hincks Genus CRIBRILARIA Canu & Bassler Cribrilaria parisiensis (Canu) (Text-figs. 35, 36) 1913 " Cribrilina puncturata Busk " ; Canu : 148, pi. 4, fig. 3. 1926 Cribrilina parisiensis Canu : 751, pi. 27, fig. i ; pi. 29, fig. 6. 1929^ Cribrilina parisiensis Canu Canu & Bassler : 37. J933 Cribrilina parisiensis Canu Dartevelle : 106. 1936 Cribrilina parisiensis Canu Dartevelle : 29. 1954 Cribrilina parisiensis Canu Balavoine : 256. FIGURED SPECIMEN. 0.48641 (Text-figs. 35, 36). ADDITIONAL MATERIAL. Four specimens, D . 48642-0 . 48645 . DIAGNOSIS. Cribrilaria with paired adventitious avicularia and without distinct lumen pores ; ovicell without rim ; ovicelled zooecia without oral spines. DESCRIPTION. Zoarium encrusting, unilaminar, the zooecia arranged in irregular longitudinal rows increasing in number distally by bifurcation. Zooecia in adjacent rows alternate in position. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 59 Zooecia elliptical, separated by distinct grooves. Length about one and a half times width. Frontal shield convex, highest at sub-oral umbo, without definite gymnocyst. Costae 8-n excluding apertural bar, narrow, smooth, tapering towards centre of shield, without distinct lumen pores. Lacunae small, circular or elliptical, 6 or rarely 5 between adjacent costae. Apertural bar about half as wide again as costae, not tapering, strongly curved, with a projecting, central umbo which is concave on both distal and proximal sides. Orifice terminal on frontal surface, small, semi-circular, the proximal lip slightly concave. Distal margin with 4 erect, hollow spine bases (lacking in ovicelled zooecia) . Avicularia adventitious, small, paired, frontal, placed one on either side of orifice, the rostrum directed distally and outwards. Rostrum short, pointed. Complete pivotal bar sometimes preserved. 36 0.500mm FIGS. 35-37. Figs. 35, 36. Cribrilaria parisiensis (Canu). 0.48641. A non-ovicelled zooecium and two other zooecia, the proximal one ovicelled, in different parts of the same fragmentary zoarium. Fig. 37. Membraniporella radiata (Reuss). 0.48646. Three zooecia, the proximal one ovicelled. 60 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Ovicell hyperstomial, globular, nearly circular in frontal outline. Surface smooth, imperforate, without rim. Opening nearly perpendicular to zooecial orifice, small, semi-circular. MEASUREMENTS : Lz (8) 0-316 (0-0183) mm., 0-30-0-34 mm. Iz (8) 0-252 (0-0323) mm., 0-20-0-30 mm. ho (9) 0-057 (0-0060) mm., 0-05-0-07 mm. lo (9) 0-092 (0-0071) mm., 0-09-0-10 mm. Lav (7) 0-055 (0-0068) mm., 0-05-0-07 mm. Lov (6) 0-140 (0-0070) mm., 0-13-0-14 mm. REMARKS. This species does not fit easily in any known Cribrilinid genus. The frontal shield, including the absence of a gymnocyst, the orifice, and the ovicell are very much like those of Eschara radiata Moll, the type species of Cribrilaria, which has vicarious, rather than adventitious avicularia.9 The avicularia of C. parisiensis are quite like those of Lepralia gattyae Busk, the type species of Puellina, which has a different frontal shield. By analogy with other Acanthostega, e.g. Castanopora (see Larwood 1962 : 203-206), greater systematic weight is placed on the form of the frontal shield than on the position of the avicularia. C. parisiensis differs from the eastern European Upper Eocene species, C. calomorpha (Reuss), in number of costae and form of orifice (see Malecki 1963 : 115). DISTRIBUTION. Eocene (Auversian) ; Belgium, France. Eocene (Bartonian) ; Belgium. Oligocene (Stampian) ; France. Genus MEMBRANIPORELLA Smitt 1873 Membraniporella Smitt : 10. TYPE SPECIES (selected by Hincks 1877). Lepralia nitida Johnston 1838 : 277, pi. 34, fig. 7. Recent ; British Isles. DIAGNOSIS. Frontal shield composed of relatively few costae over-arching a well-developed membrane. Gymnocyst prominent proximally and laterally. Costae usually with conspicuous, open lumina. Lacunae slit-like and undivided or with few lateral fusions. Orifice semi-circular to sub-circular, sometimes with lateral condyles. Distal spines frequently present. Avicularia, where present, adventitious, placed usually on proximal part of gymnocyst near orifice of laterally adjacent zooecium. Ovicell hyperstomial, imperforate in the type species, elongate globular. Orifice of ovicelled zooecia slightly wider than that of non-ovicelled ones. REMARKS. Though the diagnosis given here is based in large part on the type species, L. nitida (holotype, Department of Zoology, British Museum (Natural His- tory), 47.9.16.66, Johnston Collection), it has been framed to embrace the extremes 9 C. calomorpha (Reuss) (Malecki 1963 : 115), from the Ludian of Poland and Rumania, also has paired adventitious avicularia ; C. simulator (Canu & Bassler 1920 : 298, pi. 41, fig. 21 ; pi. 84, fig. 14), from the Upper Eocene and Oligocene of the Gulf Coast of the U.S.A., has paired adventitious avicularia on the ovicelled, but not the ordinary zooecia. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 61 of variation shown by M. aragoi (Audouin) (see Harmer 1926, pi. 34, figs. 12-14) and M. compressa Canu & Bassler (1920, pi. 41, figs. 6-8). Such species as M. radiata (Reuss), described below, and M. planula Canu & Bassler (1920, pi. 5, figs. 8-10), M. ulrichi Canu & Bassler (1920, pi. 41, figs. 4, 5), M. monilifera Canu & Bassler (1920, pi. 41, figs. 9, 10), M. bioculata Canu & Bassler (1920, pi. 41, figs. 11-13), and M.? subagassizi Canu & Bassler (1920, pi. 84, figs. 8-13) seem to be much closer to the type species. Membraniporella radiata (Reuss) (Text-fig. 37) i86ga Celleporaria radiata Reuss : 292, pi. 30, fig. 9. 1885 Cribrilina chelys Koschinsky : 36. 1889 Cribrilina chelys Koschinsky ; Pergens : 70. 1891 Cribrilina chelys Koschinsky ; Waters : 16, pi. 2, fig. 10. 1898 Cribrilina chelys Koschinsky ; Neviani : 39, text-fig, i. 19076 Cribrilina chelys Koschinsky ; Canu : 145, pi. 20, fig. i. 1929^ Collarina radiata (Reuss) Canu & Bassler : 34, pi. 2, figs. 17, 18. 1933 Collarina radiata (Reuss) ; Dartevelle : 101. 1946 Cribrilina chelys Koschinsky ; Buge : 433. 1951 Collarina radiata (Reuss) ; Kyri : 74. 1962 Collarina radiata (Reuss) ; Ghiurca, table i. FIGURED SPECIMEN. 0.48646 (Text-fig. 37). ADDITIONAL MATERIAL. Three specimens, 0.48647-0.48649. DIAGNOSIS. Membraniporella with 5-7 costae in addition to the apertural bar and without lateral fusions ; costae with exposed, slit-like lumina ; orifice with lateral condyles ; avicularia adventitious, single, paired, or multiple on the lateral gymnocyst and, occasionally, vicarious as well ; ovicell large, coarsely perforate. DESCRIPTION. Zoarium erect, bilaminar, compressed, the zooecia arranged in irregular, alternating, longitudinal rows. Zooecia irregularly elliptical, nearly twice as long as wide, separated by furrows. Gymnocyst wide, especially proximally, strongly arched, smooth. Frontal shield relatively flat, small, barely twice the size of the orifice. Costae 5-7 (usually 7) excluding apertural bar, short, broad, triangular, smooth, each with a single, slit-like, exposed lumen. Adjacent costae without lateral fusions, separated by a single, slit-like lacuna. Median line of shield with thin, smooth, irregular ridge formed by fusion of ends of costae. Apertural bar formed by a pair of opposing costae slightly thicker than the others. Orifice sub-circular, less rounded proximally than distally, with a pair of stout, deeply placed lateral condyles close to proximal margin. Peristome short, thin, present distally only. Oral spines lacking. Avicularia adventitious, single, paired, or, rarely, multiple, placed on lateral gymnocyst proximal to orifice in such position as to be near orifice of zooecium of adjacent row. Rostrum umbo-like, rounded, directed upwards, outwards, and 62 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX slightly distally or proximally. Pivotal bar complete, with small, rounded knob on rostral side. Vicarious avicularia lacking in specimens at hand. Ovicell hyperstomial, larger than zooecia, globular, with coarsely perforate surface. Orifice of ovicelled zooecia semi-circular, wider than that of ordinary zooecia, with weaker condyles and better-developed distal peristome. MEASUREMENTS : Ordinary zooecia Lz (5) 0-575 (0-0237) mm., 0-56-0-62 mm. Iz (6) 0-304 (0-0438) mm., 0-23-0-34 mm. ho (5) 0-147 (0-0140) mm., 0-14-0-17 mm. lo (6) 0-144 (0-0183) mm., 0-13-0-17 mm. Lav (7) 0-171 (0-0086) mm., 0-16-0-19 mm- Ovicelled zooecia lo (5) 0-190 (0-0164) mm., 0-17-0-21 mm. Lov (5) 0-498 (0-0299) mm-> 0-44-0-51 mm. DISTRIBUTION. Eocene (Lutetian) ; France, Germany. Eocene (Auversian) ; Belgium. Eocene (Ludian) ; Poland, Italy, Rumania. Suborder ASCOPHORA Family EXEGHONELLIDAE Harmer Genus EXECHONELLA Duvergier 1924 Exechonella Duvergier : 18. TYPE SPECIES (by monotypy). Cydicopora! grandis Duvergier 1921 : 124, pi. 3, figs. 2, 3. Miocene (Aquitanian) ; Gironde, France. DIAGNOSIS. Frontal wall calcareous, over-arching a fully developed membrane. Frontal pores large, irregularly dispersed over surface, not filled with tissue. Peris- tome long or short, tubular, with small lateral denticles or none. Secondary orifice sub-circular. Avicularia, where present, adventitious, frontal, with pointed rostrum and pivotal bar. Ovicells usually lacking ; where present, small, developed on distal side of peristome. REMARKS. This genus is usually attributed to Canu & Bassler (1927 : 4), who selected as type species Hiantopora magna MacGillivray (see Harmer 1957 : 652). However irregular, Duvergier's introduction of the genus has priority, with C.? grandis the type species by monotypy. Exechonella sp. (Text-fig. 38) FIGURED SPECIMEN. 0.48650 (Text-fig. 38). CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 63 38 0.500mm FIGS. 38-39. Fig. 38. Exechonella sp. 0.48650. Three zooecia from the small zoarial fragment. Fig. 39. Hippopleurifera canui nom. nov. 0.48651. Two zooecia, the one on the right ovicelled, from the small zoarial fragment. 64 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX DESCRIPTION. Zoarium presumably encrusting, the basal surface uncalcified. Zooecia irregularly elliptical, separated by shallow, poorly denned grooves. Length and width subequal. Frontal wall strongly inflated, thick, smooth, with 12-15 large, circular, quincuncially arranged foramina. Peristome short, the secondary orifice terminal on frontal surface, sub-circular, slightly wider than long, without denticles. Avicularia and ovicells lacking in specimen at hand. MEASUREMENTS : Lz (4) 0-614 (0-0237) mm., 0-55-0-68 mm. Iz (4) 0-603 (0-0386) mm., 0-56-0-65 mm. ho (4) 0-141 (0-0164) mm., 0-12-0-15 mm. |second orifice lo (4) 0-214 (0-0209) mm-> 0-19-0-24 mm. J REMARKS. Only one fragment, badly worn and consisting of just 7 zooecia, recovered from the Upper Bracklesham material, is referable to Exechonella. Absence of positive characters, probably a consequence of its poor preservation, makes specific assignment impossible. Family UMBONULIDAE Canu Genus HIPPOPLEURIFERA Canu & Bassler Hippopleurifera canui10 nom. nov. (Text-fig. 39) 1873 non Hippothoa mucronata Smitt : 45, pi. 8, fig. 169. Petralia mucronata Canu : 301, pi. 4, fig. 5 (misspelt " micronata "). FIGURED SPECIMEN. 0.48651 (Text-fig. 39). DIAGNOSIS. Unilaminar, probably encrusting Hippopleurifera, with large, mucronate zooecia ; areolae in a single row ; orifice with feeble condyles and 6 distal spines on non-ovicelled and 4-6 on ovicelled zooecia ; avicularium single or paired, lateral sub-oral, with rostrum directed proximally and inwards. DESCRIPTION. Zoarium unilaminar, probably encrusting. Zooecia rhomboidal, separated by a shallow groove. Length nearly one and a half times width. Frontal wall thick, very convex, highest at mucro. Surface coarsely tuberculate centrally, areolate in a single row marginally, with inter-areolar costules nearly reaching centre. Orifice large, terminal, inclined distally, hidden by a large, projecting mucro proximally, rounded sub-quadrate, with a pair of feeble condyles near proximal corners of lateral margins. Distal margin with 6 large, hollow spine bases on both ovicelled and non-ovicelled zooecia, in material at hand. 10 After the late Ferdinand Canu. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 65 Avicularium adventitious, frontal, single in material at hand, placed near proximo- lateral margin of orifice with rostrum directed upwards, proximally, and slightly inwards. Rostrum rounded, pivotal bar complete. Ovicell hyperstomial, globular, elongate. Surface not preserved in specimen at hand. MEASUREMENTS : Lz (2) 0-821 (0-0242) mm., 0-80-0-84 mm. Iz (2) o-6n (0-0181) mm., 0-60-0-62 mm. ho (2) 0-222 (0-0121) mm., 0-21-0-23 mm. lo (2) 0-214 (0-0242) mm., 0-20-0-23 mm- REMARKS. Canu (19140 : 301) described the ovicell of this species as having crescents incompletely calcified on their edges, but in his figure (pi. 4, fig. 5) the surface of the ovicell appears smooth. The specimen illustrated by Canu has 4 spines on the ovicelled zooecia in contrast to the 6 (one overlapped by the floor of the ovicell) displayed by the Upper Bracklesham specimen. Canu characterized the species as having paired avicularia, but his figure shows only one zooecium with two avicularia. One of the zooecia in the Upper Bracklesham specimen may have a small remnant of a second avicularium. This species is represented in the Curry Collection by the single specimen of two almost complete and one fragmentary zooecia. DISTRIBUTION. Eocene (Lutetian) ; France. Family EXOCHELLIDAE Bassler Genus ESCHAROIDES Milne Edwards 1836 Escharoides Milne Edwards : 218. 1902 Peristomella Levinsen : 26. TYPE SPECIES (chosen by Norman 1903). Cellepom coccinea Abildgaard 1806 : 30, pi. 146, figs, i, 2. Recent ; North Sea, Heligoland. DIAGNOSIS. Frontal wall calcareous, granular to tuberculate, with a single row of marginal areolae separated by short, peripheral costules. Orifice sub-circular, deeply buried in peristome, steeply inclined distally, with a short, curved plate pro- jecting into the zooecium from its distal margin. Peristome not differentiated from frontal surface, sometimes with distal spines. Secondary orifice oval, usually with a median-proximal denticle, and sometimes with a pair of proximo-lateral ones as well. Avicularia adventitious, usually paired, placed on lateral margins of frontal near proximal part of peristome. Rostrum usually pointed, directed outwards and distally. Pivotal bar complete. Larger, vicarious avicularia, with swollen chamber margined with areolae, sometimes present. Ovicell hyperstomial, globular, with marginal areolae separated by costules, and, at least in some species, with finer perforations in radial lines between costules. Interzooecial communication by pore-chambers. Operculum a weakly chitinized valve. GEOL. 13, I. 5 66 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX REMARKS. Brown's suggestion (1952 : 298) that species of Escharoides with well- developed marginal areolae on the ovicells may have to be transferred to a new genus need not be followed, for specimens of the type species (e.g. Department of Zoology, British Museum (Natural History), 99.5.1.858, 91.8.7.16, 93.8.7.31, and 1950.11.6.4) show ovicells with distinct areolae. Most species of Escharoides have pointed, not spatulate avicularia (cf. Brown 1952 : 297). Escharoides aliferus (Reuss) (Text-figs. 40, 41) i86ga Eschara alifera Reuss : 274, pi. 33, fig. n. 1887 Eschara alifera Reuss ; Pergens : 7. 1889 Mucronella alifera (Reuss) Pergens : 71. 1891 Smittia coccinea var. alifera (Reuss) ; Waters : 21, pi. 3, fig. 7. 1908 Peristomella alifera (Reuss) Canu : 92, pi. 7, fig. 7. 1925 Peristomella alifera (Reuss) ; Canu : 47. 1946 Escharoides aliferus (Reuss) Buge : 434. 1949 Peristomella alifera (Reuss) ; Balavoine : 774. 1957 Escharoides aliferus (Reuss) ; Balavoine : 191. 1960 Romancheina gouetensis Balavoine : 247, pi. 6, figs. 11-14 '• pi- 7» nSs- I-4- 1962 Escharoides aliferus (Reuss) ; Ghiurca, table i. 1963 Escharoides aliferus (Reuss) ; Braga : 34. FIGURED SPECIMENS. 0.48652 (Text-fig. 40), 0.48653 (Text-fig. 41). ADDITIONAL MATERIAL. Twenty-two specimens, 0.48654-0.48675. DIAGNOSIS. Erect, bilaminar Escharoides with prominent mucro flanked by a smaller denticle on either side ; a single spine base at each lateral margin of secondary orifice ; avicularia paired, occupying swollen chambers at lateral margins of frontal ; ovicell elongate, irregularly perforated, areolated. DESCRIPTION. Zoarium erect, arborescent, arising from an encrusting base, composed of compressed, bilaminar fronds with more than 12 longitudinal rows of zooecia on each side ; the zooecia in adjacent rows alternate in position. Inter- zooecial communication by small pore-chambers in distal and disto-lateral walls. Zooecia irregularly rhombic, rounded distally, the lateral avicularia forming wing- like expansions on the sides. Zooecia separated by shallow, ill-defined depressions. Zooecial length almost one and a half times width. Frontal wall moderately thick, convex, highest at proximal lip of secondary orifice. Surface smooth except for FIGS. 40-47. Fig. 40. Escharoides aliferus (Reuss). 0.48652. Ovicelled zooecium. Fig. 41. Escharoides aliferus (Reuss). 0.48653. Two non-ovicelled zooecia. Fig. 42. Smittoidea variabilis (Canu). 0.48676. Two zooecia from a bilaminar fragment. The one on the right has an avicularium. Fig. 43. Smittoidea variabilis (Canu). 0.48677. A zooecium from another bilaminar fragment in a more advanced state of calcification. Fig. 44. Smittoidea variabilis (Canu). 0.48678. An ovicelled zooecium from a uni- laminar fragment. Fig. 45. Smittoidea variabilis (Canu). 0.48679. An ovicelled zooecium from another unilaminar fragment. The peristome is unusually developed. Figs. 46, 47. Smittoidea variabilis (Canu). 0.48679. A zooecium and the outline of an avicularium of another zooecium from the same unilaminar fragment. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 67 47 0.500mm 68 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX large, irregular tubercles along mid-line and prominent peripheral costules separating the areolae. Areolae large, sub-circular or elliptical, evenly spaced in a row of 18-25 round lateral and proximal margins. Additional, smaller areolae, 4-6, on line separating avicularian chamber from frontal surface. Orifice small, sub-circular, inclined distally, deeply buried in peristome. Distal plate well developed, with straight margin and concave surface. Peristome not differentiated from frontal surface, thin, with a small, hollow spine base at mid-point of each lateral margin. Secondary orifice inclined distally, roughly circular, with large, median mucro flanked on either side by a much smaller denticle. Avicularia adventitious, but nearly of interzooecial position, paired, occupying large, slightly swollen chambers on lateral margins of zooecia just proximal to orifice. Rostrum pointed, attenuated, directed transversely outwards. Pivotal bar complete. Ovicdl hyperstomial, globular, elongate, opening broadly into peristome, the lip of the ovicell forming a convex projection into secondary orifice. Surface convex, irregularly perforated with numerous small pores between radiating, costule-like ridges. Orifice of fertile zooecium not modified. MEASUREMENTS : Lz (7) 0-552 (0-0640) mm., 0-47-0-68 mm. Iz (6) 0-398 (0-0461) mm., 0-35-0-48 mm. ho (5) 0-123 (0-0167) mm., 0-10-0-14 mm. "\ second orifice lo (5) 0-176 (0-0076) mm., 0-17-0-19 mm. f Lav (7) 0-184 (0-0396) mm., 0-14-0-26 mm. Lov (2) 0-321 (0-0302) mm., 0-30-0-34 mm. REMARKS. Balavoine (1960 : 246) included this species in his list from the Lute- tian of Bois-Gouet (Loire-Atlantique) with the annotation that Canu (1908) had found it, but that it was missing from his (Balavoine's) material. At the same time he (1960 : 247) described Romancheina gouetensis as a new species similar to E. aliferus but having a uniformly perforate (" tremocystal ") frontal. His figures (1960 : pi. 6, figs. 11-14 > pl- 7» fig8- I-4) show, however, not tremopores but areolae in the double lateral rows characteristic of the frontal of E. aliferus where it is joined by the avi- cularian chambers. Thus there is little doubt that Balavoine's species is the one correctly identified by Canu with E. aliferus. The pore-chambers of the Upper Bracklesham specimens are very small and difficult to identify. DISTRIBUTION. Eocene (Lutetian) ; France. Eocene (Ludian) ; Italy, Hungary, Poland, Rumania. Family SMITTINIDAE Levinsen Genus SMITTOIDEA Osburn 1952 Smittoidea Osburn : 408. TYPE SPECIES (by original designation). Smittoidea prolifica Osburn 1952 1408, pl. 48, figs. 7, 8. Recent ; Californian coast, U.S.A. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 69 DIAGNOSIS. Frontal wall calcareous, smooth, granular, or tuberculate, with a single row of marginal areolae proximally and laterally. Orifice sub-circular to semi-circular with well-developed condyles and a median proximal lyrula. Peristome high distally and laterally, usually with a proximal notch ; spines, if present, evanes- cent. Avicularium adventitious, frontal, median sub-oral, placed within peristomial sinus or just proximal to it, rostrum directed longitudinally proximally. Ovicell hyperstomial, evenly and numerously perforated. REMARKS. Before Osburn's sweeping revision of the family Smittinidae (1952 : 390-440), almost all Tertiary Smittinids were placed in either Smittina or Porella, both of which had become ungainly (see Lagaaij 1952 : 97, for a discussion of the nomenclatorial difficulties attending these genera). Smittina, as now restricted, includes species having an evenly perforated frontal, a median sub-oral avicularium, the orifice with lyrula and condyles, and an evenly perforated ovicell. Revision of the Eocene and Oligocene species which have in the past been assigned to Smittina and Porella will be a major undertaking. The list of species assignable to Smittoidea may be started with the following : Smittia variabilis Canu (described below) from the Eocene of England, Belgium, and France ; Smittina angulata (Minister) (Darte- velle 1952 : 191) from the Oligocene of Germany ; Smittina orbavicularia Canu & Bassler (1920: 469, pi. 61, figs. 1-4) from the Eocene of the Gulf Coast of the U.S.A. ; and Smittina telum Canu & Bassler (1920 : 468, pi. 93, figs. 1-9) and Smittina reticu- loides Canu & Bassler (1920 : 467, pi. 96, figs. 1-9) from the Oligocene of the Gulf Coast, U.S.A. Smittoidea variabilis (Canu) (Text-figs. 42-47) 1908 Smittia (Porella) variabilis Canu : 97, pi. 8, figs. 1—7. 1929* Smittina variabilis (Canu) Canu & Bassler : 40. 1933 Smittina variabilis (Canu) ; Dartevelle : 107. ?i933 Palmicellaria lerichei Dartevelle : 82, pi. 3, figs. 3, 4. 1934 Smittina variabilis (Canu) ; Davis : 223, pi. 15, fig. 54. 1937 Smittina variabilis (Canu) ; Dartevelle : no. 1946 Smittina variabilis (Canu) ; Buge : 435. 1949 Porella variabilis (Canu) Balavoine : 774. 1956 Smittina variabilis (Canu) ; Balavoine : 322. 1957 Smittina variabilis (Canu) ; Balavoine : 192. ?ig63 Trigonopora monilifera (Milne Edwards); Malecki : 130, pi. 14, fig. i [non Eschara monilifera Milne Edwards]. FIGURED SPECIMENS. 0.48676 (Text-fig. 42), 0.48677 (Text-fig. 43), 0.48678 (Text-fig. 44), 0.48679 (Text-fig. 45), 0.48679 (Text-figs. 46, 47). ADDITIONAL MATERIAL. Forty-eight specimens, 0.48680-0.48727. DIAGNOSIS. Encrusting or erect, uni- or bilaminar Smittoidea with zooecia extremely variable in size and form depending on zoarial type ; on encrusting por- tions, zooecia short, with thick, smooth frontals, well-developed peristomes, and large sub-oral avicularium on protuberant umbo ; on erect portions, zooecia longer, with yo CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX thinner, coarsely tuberculate frontals, short peristomes, and smaller avicularia or none ; orifice large, semi-circular, with stout condyles and a rudimentary lyrula ; oral spines lacking ; ovicell very wide and finely perforate. DESCRIPTION. Zoarium encrusting, unilaminar, rising in uni- or bilaminar, erect fronds, the zooecia crudely aligned in longitudinal rows on unilaminar portions, regularly aligned in alternating series on bilaminar. Zooecia on unilaminar portions irregularly hexagonal, pentagonal, or tetragonal, separated by narrow, salient threads. Length nearly one and a half times width. Frontal wall very convex and thick, highest at sub-oral umbo. Surface smooth, almost hyaline, with 3-4 large areolae in a single row on each proximo-lateral margin. Interareolar costules feebly developed, peripheral. Sub-oral umbo large, salient, massive. Orifice deeply buried in peristome, large, semi-circular, the distal margin evenly rounded, the proximal margin slightly convex, forming an incipient lyrula. Condyles paired, stout, placed on lateral margins close to proximal lip. Peristome irregular, forming a pair of high, lateral lappets, coalescing with the umbo, and a lower distal collar. Secondary orifice variable in shape depending on relative development of umbo, lappets, and avicularium. Zooecia on bilaminar portions rectangular to rhomboidal, separated by narrow, raised threads. Length more than twice width. Frontal wall moderately convex and thick, highest near centre, except where avicularium is present. Surface smooth and almost hyaline at first but becoming coarsely tuberculate and porcellanous as calcification progresses. Areolae 5-7 along each proximo-lateral margin, without costules. Sub-oral umbo lacking, the avicularium, where present, occupying a swollen chamber covering most of the frontal surface. Orifice shallow, sub-terminal on frontal surface, large, semi-circular, the distal margin evenly rounded, the proximal margin convex, with a well-developed, broad, tapering, median lyrula. Condyles rudimentary. Peristome very short and thin, without distinct lateral lappets. Avicularium adventitious, usually lacking on bilaminar fronds, single, median, sub-oral, placed on umbo or in swollen frontal chamber, usually facing into peristome, but with rostrum directed frontally and proximally. Rostrum rounded ; pivotal bar complete. Ovicell hyperstomial, large, globular, wider than long, present on both uni- and bilaminar portions of zoarium. Surface evenly perforated with very small pores. Distal and lateral margins distinctly rimmed. Proximal margin obscured by elon- gated and inwardly bent peristomial lappets. MEASUREMENTS : Zooecia on unilaminar portions Lz (10) 0-447 (o'066i) mm., 0-31-0-56 mm. Iz (10) 0-325 (0-0397) mm., 0-23-0-36 mm. ho (10) 0-120 (0-0121) mm., 0-10-0-14 mm- lo (10) 0-146 (0-0130) mm., 0-13-0-17 mm. Lav (5) 0-109 (0-OI85) mm., 0-09-0-13 mm. Lov (3) 0-239 (0-0171) mm., 0-22-0-26 mm. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 71 Zooecia on bilaminar portions Lz (10) 0-634 (0-I°55) mm., 0-48-0-88 mm. Iz (10) 0-288 (0-0586) mm., 0-25-0-44 mm. ho (10) 0-122 (o-ono) mm., 0-09-0-13 mm. lo (ro) 0-143 (0-0107) mm., 0-13-0-16 mm. Lav (2) o-iii (0-0121) mm., 0-10-0-12 mm. Lov (4) 0-177 (0-0146) mm., 0-16-0-20 mm. REMARKS. There is so much difference in the appearance of the zooecia on the uni- and bilaminar fragments that they might easily be mistaken for distinct species. As in the Lower Bracklesham Beds (Davis 1934 : 223), unilaminar fragments account for the majority of Upper Bracklesham specimens. Although the illustrations of Palmicellaria lerichei Dartevelle (1933 : pi. 3, figs. 3, 4) are not completely clear, they resemble the general aspect of bilaminar fragments of S. variabilis. The bilaminar fragment described and illustrated by Malecki (1963 : 130, pi. 14, fig. i) as Trigonopora monilifera is almost certainly referable to S. variabilis. The globular hyperstomial ovicell (described as entozooecial) , the deeply notched peris- tome, and the median sub-oral avicularium are evident on the figure. Specimens from the Eocene of Rumania (supplied by Dr. Sten Schager) are identifiable with this species and similar to the one illustrated by Malecki. DISTRIBUTION. Eocene (Lutetian) ; England, Belgium, France. Eocene (Auver- sian) ; Belgium. Eocene (Ludian) ; Rumania, PPoland. Family ESCHARELLIDAE Levinsen Genus ESCHARELLA Gray Escharella selseyensis sp. nov. (Text-figs. 48-50) 1908 Smittia (Mucronella) hoernesi (Reuss) ; Canu : 96, pi. 7, fig. 16. 1946 Mucronella angustoecium Gregory ; Buge : 435. HOLOTYPE. 0.48728 (Text-fig. 48). PARATYPES. 0.48729 (Text-figs. 49, 50), 0.48730, and L.S.U. 8035. DIAGNOSIS. Unilaminar, probably encrusting Escharella with broad lyrula in primary orifice and peristome raised proximally to form a broad, lip-like mucro ; distal margin of orifice with 2 small, evanescent spines ; ovicell large, globular. DESCRIPTION. Zoarium unilaminar, presumably encrusting, with zooecia arranged in regular longitudinal rows, those in adjacent rows alternating in position. Zooecia elongate oval, more broadly rounded distally than proximally, separated by a narrow, raised thread. Length nearly twice width. Frontal wall moderately thick, strongly convex, highest near centre. Frontal surface coarsely granular, sometimes with a large, protruding umbo covering the greater part. Areolae large, 72 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 1.00mm CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 73 elliptical, in a single, evenly spaced row of 16-20 on lateral and proximal margins. Inter-areolar costules short, thin, peripheral. Orifice steeply inclined distally, deeply buried in peristome, semi-elliptical, with long axis transverse, the distal margin broadly rounded, the proximal margin straight or convex, with a broad, saddle-shaped, median lyrula. Peristome thin, granular, especially raised proximally to form a broad, lip-like mucro. Secondary orifice elliptical, with major axis transverse. Distal spines 2, small, evanescent. Ovicell hyperstomial, globular ; distal margin evenly rounded, marked by an irregular row of small areolae ; proximal margin with a slightly thickened, arcuate rim. Surface coarsely granular, impunctate. Heterozooecia lacking. MEASUREMENTS : Lz (8) 0-491 (0-0754) mm., 0-35-0-59 mm. Iz (5) 0-274 (°'°337) mm., 0-22-0-31 mm. ho (6) 0-081 (0-0105) mm., 0-07-0-09 mm. lo (5) 0-152 (0-0072) mm., 0-14-0-16 mm. Lov (5) 0-219 (0-OI3°) mm., 0-20-0-24 mm. REMARKS. The synonymy given by Canu (1908 : 96 ; 1913 : 149) for Lepralia hoernesi Reuss was considered to be heterogeneous by Canu & Bassler (19290 : 45). Canu's (19146) Stampian material is probably conspecific with Reuss's (1865) Oligocene specimens, and both are probably referable to Perigastrella. Canu's (1908) Lutetian material is identifiable with the Upper Bracklesham specimens for which the name Escharella selseyensis is here proposed. Mucronella angmtoecium Gregory, the Bartonian species (holotype, 49739, Edwards Collection, Barton, Hants) placed in synonymy by Canu (1908 : 96), differs from E. selseyensis in having larger zooecia with smaller, sub-circular orifices, and smaller ovicells. DISTRIBUTION. Eocene (Lutetian) ; France. Family SERTELLIDAE Jullien Genus SERTELLA Jullien Sertella marginata (Reuss) (Text-figs. 51-53) 1865 Retepora marginata Reuss : 661, pi. 10, figs. 6, 7. ?i866 Retepora marginata Reuss ; Reuss : 190. FIGS. 48-53. Fig. 48. Escharella selseyensis sp. nov. 0.48728. Holotype. Two zooecia, the proximal one ovicelled. Figs. 49, 50. Escharella selseyensis sp. nov. 0.48729. Paratype. Zooecium with well-developed frontal umbo, and oral outline of another zooecium. Fig. 51. Sertella marginata (Reuss). 0.48731. Dorsal aspect of a frag- mentary zoarium. An avicularium occurs just over the fenestrule. Fig. 52. Sertella marginata (Reuss). 0.48732. Ovicelled zooecium. Fig. 53. Sertella marginata (Reuss). 0.48733. Part of a zoarial fragment showing zooecia with oral and frontal avicularia and several with fenestral ones as well. 74 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX FIGURED SPECIMENS. 0.48731 (Text-fig. 51), 0.48732 (Text-fig. 52), 0.48733 (Text-fig. 53). ADDITIONAL MATERIAL. Six specimens, D . 48734-0 . 48739. DIAGNOSIS. Sertella with triserial trabeculae of about same width as the elliptical fenestrules ; zooecia frontally marginate with non-denticulate primary orifice and sinuate secondary orifice ; oral avicularium mucronate ; frontal avicularia rounded ; fenestral avicularium pointed ; dorsal avicularia spatulate. DESCRIPTION. Zoarium erect, unilaminar, reticulated, the fenestrules elliptical, quincuncially arranged, of about same width as branches. Zooecia disposed in 3 longitudinal rows on frontal face of branch (trabecula), those in adjacent rows alternating in position. Dorsal face finely tuberculate, divided into large, irregular polygons by narrow, slightly raised threads (vibices). Zooecia irregularly polygonal, separated by thick, low threads connecting orifices of adjoining rows. Zooecial length about one and a half times width. Frontal wall very thick, slightly convex. Surface smooth, imperforate. Orifice deeply buried in peristome, not visible in frontal aspect, small, elliptical, with long axis transverse and proximal lip straight. No condyles or denticles present. Peristome very long but almost completely immersed in thick frontal wall. Secondary orifice pyriform, with wide, deep proximal notch. A small, evanescent spine base occurs at mid-length on each lateral margin. Avicularia adventitious, frontal and dorsal, usually multiple and polymorphic : (1) Oral, single, of intermediate size, placed in proximal notch of secondary orifice, the rounded, non-mucronate rostrum projecting frontally ; cross-bar complete. (2) Ordinary-frontal, single or paired, very small, placed on frontal wall, removed from orifice, rounded, without pivotal structure. (3) Fenestral-frontal, single, large, placed on frontal wall of zooecia in neighbourhood of fenestrules, the pointed rostrum directed transversely outwards ; cross-bar complete. (4) Dorsal, large, widely scattered, the spatulate rostrum oriented more or less transversely to the axis of the trabecula ; cross-bar complete. Ovicell hyperstomial, elongate, globular, but flattened proximally. Surface smooth, with longitudinal fissure extending nearly whole length, narrow, often fusiform, the greatest width at mid-length or slightly distal to it. MEASUREMENTS : Length of fenestrule (6) 0-640 (0-1170) mm., 0-51-0-82 mm. Width of fenestrule (6) 0-294 (0-0655) mm., 0-23-0-42 mm. Lz (10) 0-307 (0-0195) mm., 0-27-0-32 mm. Iz (4) 0-175 (o-oiio) mm., 0-16-0-19 mm. ho (5) 0-056 (0-0098) mm., 0-04-0-07 mm. ~\ .,. i ,i( , ( r > secondary orifice lo (6) 0-054 (0-0070) mm., 0-04-0-06 mm. J Lov (6) 0-155 (0-0137) mm., 0-14-0-17 mm. Lav (oral) (5) 0-051 (0-0105) mm., 0-04-0-07 mm. Lav (frontal) (4) 0-030 (0-0049) mm., 0-02-0-03 mm. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 75 Lav (fenestral) (3) o-iii (0-0086) mm., 0-10-0-12 mm. Lav (dorsal) (2) 0-115 (0-0181) mm., 0-10-0-13 mm. REMARKS. This species is very much like the type species of Sertella, S. beaniana (King), from which it differs primarily in having trabeculae and fenestrules of about the same width, secondary orifice with sinus, fenestral avicularia pointed rather than spatulate, and dorsal avicularia large and spatulate. Although Reuss's figures (1865, pi. 10, figs. 6, 7) are not completely clear, the Upper Bracklesham specimens seem to agree with them. DISTRIBUTION. Oligocene (Stampian) ; Germany. PMiocene ; Germany. Family SCHIZOPORELLIDAE Jullien Genus DAKARIA Jullien 1903 Dakaria Jullien in Jullien & Calvet : 90. TYPE SPECIES (by original designation). Dakaria chevreuxi Jullien in Jullien £ Calvet 1903 : 90, pi. 10, fig. 6. Recent ; off Dakar, Senegal. DIAGNOSIS. Frontal wall calcareous, evenly perforated with numerous pores. Orifice with broad, rounded, proximal sinus. Avicularia lacking. Ovicells, if present, hidden from frontal view. REMARKS. This deceptively simple Schizoporellid genus is not yet completely understood. Harmer's (1957 : 1021) inclusion of Watersipora and Cribella in its synonymy is open to question. Dakaria beyrichi (Stoliczka) (Text-fig. 54) 1862 Cellaria beyrichi Stoliczka : 83, pi. i, fig. 10. 1908 Hippoporina beyrichi (Stoliczka) Canu : 83, pi. 6, fig. 2. 1935 Dakaria beyrichi (Stoliczka) Dartevelle : 115. 1946 Hippoporina beyrichi (Stoliczka) ; Buge : 434. 1952 Dakaria beyrichi (Stoliczka) ; Dartevelle : 190. 1956 Dakaria beyrichi (Stoliczka) ; Balavoine : 322. FIGURED SPECIMEN. 0.48740 (Text-fig. 54). DIAGNOSIS. Erect, cylindrical Dakaria with large, simple orifice lacking a peris- tome ; ovicell unknown. DESCRIPTION. Zoarium erect, cylindrical, the zooecia in 6-8 alternating, longi- tudinal rows. Zooecia irregularly rhombic, separated by a faint groove at the crest of a wide, low thread. Zooecial length slightly less than width. Frontal wall moderately thick, very convex, highest near centre. Surface smooth, evenly perforated with numerous, small, circular, quincuncially arranged pores. 76 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 57 0.500mm FIGS. 54-57. Fig. 54. Dakari beyrichi (Stoliczka). 0.48740. Two zooecia. Fig. 55. Schizomavella trigonostoma sp. nov. 0.48741. Holotype. Four zooecia, the one at upper left without an avicularium. Fig. 56. Escharina procumbens (Canu). 0.48742. Three ovicelled zooecia from, a unilaminar zoarial fragment. Fig. 57. Escharina procumbens (Canu). 0.48743. Three non-ovicelled zooecia from a bilaminar fragment. The upper zooecium has the right margin of its orifice broken. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 77 Orifice terminal on frontal surface, slightly inclined distally, the distal half slightly more than a semi-circle, the proximal sinus a slightly smaller semi-circle. Peristome, condyles lacking. Heterozooecia lacking. Ovicell unknown. MEASUREMENTS : Lz (4) 0-635 (0-0409) mm., 0-59-0-68 mm. Iz (4) 0-667 (0-0908) mm., 0-56-0-77 mm. ho (5) 0-180 (0-0337) mm., 0-13-0-22 mm. lo (5) 0-161 (0-0112) mm., 0-14-0-17 mm. DISTRIBUTION. Eocene (Ypresian, Lutetian) ; France. Eocene (Lutetian) ; Belgium. Oligocene (Lattorfian) ; Germany. Genus SCHIZOMAVELLA Canu & Bassler ?i8g3 Schismoporella Gregory : 243. 1917 Schizomavella Canu & Bassler : 40. 1917 Metroperiella Canu & Bassler : 40. TYPE SPECIES (by original designation). Lepralia auriculata Hassall 1842 : 411. Recent ; British Isles. DIAGNOSIS. Frontal wall calcareous, evenly perforated with numerous pores. Orifice oval, with distinct median-proximal sinus between small condyles; sinus usually rounded and shallow. Peristome low, usually thin. Avicularium adventi- tious, single, frontal, proximal to orifice, directed longitudinally proximally or trans- versely, sometimes placed on a sub-oral umbo. Ovicell hyperstomial, globular, perforate. REMARKS. Harmer (1957 : 1024-1027) considered Metroperiella a distinct genus by virtue of its large ovicell surrounding the orifice of the ovicelled zooecium, but fossil species are intermediate between the two extremes. Buge (1953 : 322) regarded Schismoporella as a genus distinguishable from Schizomavella in having the avicularium removed from the orifice. Unfortunately, no such regularity of position as Buge implied characterizes these Schizoporellids (see e.g., Canu & Bassler 1920 : pi. 46, figs. 4-17 ; Brown 1952 : 235-238). Moreover, the type species of Schismoporella, Lepralia schizogaster Reuss (1848 : 84, pi. 10, fig. 9), requires re-study before its generic affinities can be definitely established. If it proves to be congeneric with Schizomavella auriculata, then Schismoporella will, of course, be the correct name of the genus. Schizomavella trigonostoma sp. nov. (Text-fig. 55) HOLOTYPE. 0.48741 (Text-fig. 55). PARATYPE. L.S.U. 8036. 78 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX DIAGNOSIS. Erect, bilaminar Schizomavella with rounded-trigonal orifice, the large, V-shaped sinus limited by stout lateral condyles ; avicularium placed on proximo-lateral margin of orifice, the rostrum directed obliquely distally and out- wards. DESCRIPTION. Zoarium erect, bilaminar, compressed or cylindrical, the zooecia arranged in irregular longitudinal rows, those in adjacent rows alternating in position. Zooecia irregularly rhomboidal, separated by a narrow, salient thread. Zooecial length slightly greater than width. Frontal wall moderately thick, slightly convex, highest near centre. Surface granular, evenly perforated with numerous, small, circular, quincuncially arranged pores. Orifice entirely visible in frontal view, small, terminal on frontal wall, rounded trigonal, the distal margin evenly rounded, the proximal portion consisting of a very large, V-shaped sinus limited on each side by a stout, proximally directed condyle. Peristome thick but very low, widest proximally, smooth, flaring outwards. Avicularium adventitious, single, frontal, placed on proximo-lateral margin of orifice, partly buried in peristome, with rostrum directed obliquely distally and outwards. Rostrum rounded. Pivotal bar complete. Ovicell unknown. MEASUREMENTS : Lz (5) 0-477 (0>°584) mm., 0-38-0-52 mm. Iz (5) 0-397 (0-1013) mm., 0-33-0-57 mm. ho (6) 0-137 (0-0171) mm., 0-12-0-16 mm. lo (6) 0-107 (0-0°54) mm., o-io-o-n mm. Lav (5) 0-128 (0-0191) mm., 0-10-0-15 mm- REMARKS. The orifice of this species, with a much longer sinus and stronger condyles than are typical of Schizomavella, finds a parallel in 5. dubia Brown (1952 : 235, text-figs. 168, 169) from the Pliocene of New Zealand. The avicularium of S. trigonostoma, also atypical of Schizomavella, is similar in form, position, and orienta- tion to that of S. australis (Haswell) (see Harmer 1957 : 1031, pi. 66, fig. 9). Genus ESCHARINA Milne Edwards Escharina procumbens (Canu) (Text-figs. 56, 57) 1910 Schizoporella hoernesi var. procumbens Canu : 853, pi. 19, fig. 5. 1916 Schizoporella hoernesi (Reuss) ; Faura y Sans & Canu : 298. 1950 Schizoporella hoernesi (Reuss) ; Barroso : 179, text-fig. 7. FIGURED SPECIMENS. 0.48742 (Text-fig. 56), 0.48743 (Text-fig. 57). CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 79 DIAGNOSIS. Unilaminar or bilaminar, probably erect Escharina with somewhat elongate zooecia having coarsely perforate frontal and comparatively large orifice ; avicularitim large, single, transversely and outwardly directed, with raised, swollen chamber ; a grooved, protuberant boss is developed from a frontal pore on each side of the orifice. DESCRIPTION. Zoarium unilaminar, probably erect, tubular, sometimes becoming bilaminar, with zooecia arranged in longitudinal rows, those in adjacent rows alter- nating in position. Zooecia irregularly rectangular to rhomboidal, separated by a narrow, raised thread. Length nearly one and a half times width. Frontal wall very slightly convex, moderately thick. Frontal surface finely tuberculate, perforated with large, evenly spaced pores, those along the margins tending to be larger. Orifice large, semi-circular, completely visible in frontal view. Distal margin evenly rounded ; proximal margin nearly straight, interrupted medially by a narrow, shallow, rounded sinus. Peristome rudimentary. On each side of orifice occurs a tall, proximally grooved, protuberant boss originating from a frontal pore. Avicularium adventitious, single or absent, large, frontal, placed on lateral margin at about mid-length, its rostrum transversely outwardly directed. Chamber slightly raised and swollen, the exterior tuberculate but imperforate. Rostrum channelled, pointed, produced slightly beyond border of zooecium. Pivotal bar complete. Ovicell hyperstomial, recumbent, deeply immersed in distal zooecium, forming a small, hood-like swelling at distal margin of orifice. Surface imperforate, finely tuberculate, the proximal margin thickened, the distal margin merging with frontal wall of distal zooecium. MEASUREMENTS : Lz (7) 0-385 (0-0401) mm., 0-32-0-43 mm. Iz (7) 0-291 (0-0651) mm., 0-23-0-38 mm. ho (7) 0-127 (0-0104) mm., 0-11-0-14 mm. lo (8) 0-123 (o-oin) mm., o-io-o-n mm. Lav (3) 0-154 (0-0171) mm., 0-14-0-17 mm. Lov (2) 0-171 (o-oooo) mm., 0-17 mm. REMARKS. The Upper Bracklesham specimens agree best with the material from the Bartonian of Spain illustrated by Barroso (1950, text-fig. 7), from which they differ only in having well-developed lateral-oral bosses. The French specimens figured by Canu (1910 : pi. 19, fig. 5) have more numerous, smaller frontal pores and less-developed bosses. E. hoernesi (Reuss) has much more slender zooecia, smaller frontal pores, and much smaller avicularia in addition to lacking oral bosses entirely. Ovicells have not been noted heretofore in E. procumbens, though Reuss (18690 : pi. 33, figs. 6, 7) illustrated similar ones in E. hoernesi. DISTRIBUTION. Eocene (Lutetian, PAuversian) ; France. Eocene (Lutetian, Bartonian) ; Spain. 8o CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Family HIPPOPODINIDAE Levinsen Genus HIPPOPORINA Neviani Hippoporina globulosa (d'Orbigny) (Text-figs. 58, 59) 1851 Reptescharellina globulosa d'Orbigny : 453. 1908 Hippoporina globulosa (d'Orbigny) ; Canu : 82, pi. 6, fig. 9. 1933 Hippoporina globulosa (d'Orbigny) ; Dartevelle : 77. 1946 Hippoporina globulosa (d'Orbigny) ; Buge : 434. FIGURED SPECIMENS. 0.48744 (Text-fig. 58), 0.48745 (Text-fig. 59). DIAGNOSIS. Nodular, encrusting Hippoporina with paired, lateral-oral avicularia, the rostra directed inwards and distally ; ovicell large, globular, evenly perforated ; oral condyles stout, very close to proximal lip. DESCRIPTION. Zoarium unilaminar, encrusting, forming irregularly globular masses, with the zooecia arranged in longitudinal rows, those in adjacent rows alternating in position. * f * '^^^ '^/£^^W^-^ 58 0.500mm FIGS. 58-59. Fig. 58. Hippoporina globulosa (d'Orbigny). 0.48744. Two ovicelled zooecia, the proximal one incomplete. Fig. 59. Hippoporina globulosa (d'Orbigny). 0.48745. Ordinary zooecium. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 81 Zooecia irregularly rectangular, pentagonal, or rhomboidal, rounded distally, separated by narrow, sharp, raised threads. Zooecial length slightly greater than width. Frontal wall moderately thick, slightly convex, highest at proximal margin of peristome. Surface evenly perforated with numerous, closely spaced, quin- cuncially arranged, circular, funnel-shaped pores. Surface between pores ridged in a reticulate pattern, the intersections of the ridges often marked by small tubercles. Orifice entirely visible in frontal view, large, elliptical, with major axis longitudinal, sub-terminal, slightly inclined distally. Condyles stout, paired, directed slightly proximally, placed on lateral margins very near proximal lip. Peristome thick but low, smooth, flaring outwards all round orifice, enclosing a very narrow, crescent - shaped bare spot proximally and occasionally distally as well. Avicularia adventitious, frontal, paired, placed on lateral margins of peristome with rostra directed inwards and distally on to peristome. Rostrum pointed but not produced. Pivotal bar complete. Ovicell hyperstomial, large, globular, deeply sunk in distal zooecium, but with proximal margin arching above orifice of ovicelled zooecium. Surface evenly per- forated with numerous pores like those on frontal wall. Distal and proximal margins without rims. Orifice of ovicelled zooecia unmodified. MEASUREMENTS : Lz (4) 0-646 (0-1037) mm., 0-56-0-77 mm. Iz (4) 0-517 (0-0649) mm., 0-45-0-61 mm. ho (4) 0-156 (0-0108) mm., 0-14-0-17 mm. lo (4) 0-162 (0-0070) mm., 0-15-0-17 mm. Lav (3) 0-177 (0-0049) mm-> 0-17-0-18 mm. Lov (3) 0-399 (O'O^1) mm., 0-38-0-41 mm. REMARKS. D'Orbigny (1851) did not illustrate his material ; the holotype (Museum national d'Histoire naturelle, Paris, no. 9648) was figured by Canu (1908 : pi. 6, fig. 9). DISTRIBUTION. Eocene (Lutetian, Auversian) ; France. Eocene (Auversian) ; Belgium. Family DITAXIPORINIDAE Cheetham Genus CABEROIDES Canu 1908 Caber aides Canu : 87. TYPE SPECIES (by original designation). Caberoides canaliculatus Canu 1908 : 88, pi. n, figs. 11, 12. Eocene (Lutetian) ; vicinity of Paris, France. DIAGNOSIS. Frontal wall calcareous, convex, irregularly perforated with scattered pores of variable size. Orifice terminal on frontal surface, broadly sinuate between small, widely spaced condyles proximally. Peristome thin, short. Interzooecial communication by. simple pores. Vibracula adventitious, paired, with elongate chambers on frontal surface ; additional oblique, slit-like vibracula sometimes present GEOL. 13, I 6 82 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX on dorsal surface, following interzooecial sutures. Avicularium single, adventitious, placed on outer margin of zooecium, either above or below vibraculum, with short, pointed rostrum directed distally and outwards. Ovicell hyperstomial, elongate, globular, curved outwards from zooecial mid-line, ornamented with irregularly placed pores. Zoarium erect, jointed, branching at internodes ; internodes biserial, with orifices opening on one face only. REMARKS. Canu (1908 : 87) introduced this genus for two French Lutetian species, C. canaliculatus, the type, and C. grignonensis Canu (1908 : 88, pi. n, figs. 14, 15), both of which display dorsal vibracula in addition to the frontal ones and have the frontal avicularium placed above the outer vibraculum. C. corniculatus sp. nov., described below, lacks dorsal vibracula and has the avicularium below the outer vibraculum. The genus Caberoides, constituted of these three species, ranges from Lutetian to Auversian and is endemic to the Anglo-Paris Basin. I have earlier (Cheetham 19630; : 489) remarked upon the similarity of Caberoides and Ditaxiporina Stach and at that time erected the family Ditaxiporinidae for them. Caberoides corniculatus sp. nov. (Text-figs. 60, 61) HOLOTYPE. 0.48746 (Text-figs. 60, 61). PARATYPES. 0.48747-0.48771 (25 specimens), and L.S.U. 8037. DIAGNOSIS. Caberoides with vibracula on frontal side only, paired, the outer one elongate and horn-like ; avicularium placed on chamber of outer vibraculum and below its opening ; ovicell curved outwards markedly, its surface with numerous, irregular perforations and 2-3 larger fenestrae. DESCRIPTION. Zoarium erect, jointed, dichotomous at nodes. Internodes long, narrow, compressed, consisting of 10 or 12 zooecia arranged in 2 alternating series, with orifices all opening on one face. Proximal end of internode with single, simple, circular opening formed jointly by proximal ends of first zooecia of both series, one of which, either right or left, is slightly lower. Distal end of internode with 2 pores, one on mid-line of distal wall of last zooecium of each series. Zooecia rhomboidal, distorted on frontal side by vibracular chambers, separated on both frontal and dorsal sides by grooves. Communication between zooecia of the same series by a single, median pore in distal wall ; between zooecia of adjacent series by simple pores along mid-line of inner lateral wall. Zooecial length slightly greater than width. Frontal wall thin, irregularly convex, highest at proximal lip of orifice. Frontal surface smooth, perforated irregularly by pores sometimes forming short tubules. Basal surface gently convex, perforated in same way as frontal. Orifice terminal on frontal surface, large, inclined distally. Distal margin broadly and evenly rounded ; proximal margin broadly sinuate between a pair of small, widely spaced condyles. Peristome thin, short. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 83 60 61 0.500mm FIGS. 60, 61. Caberoides corniculatus sp. nov. 0.48746. Holotype. Frontal and dorsal views of a nearly complete internode, showing 10 zooecia of which three are ovicelled. 84 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Vibracula adventitious, paired, placed on lateral margins of frontal wall, opening beside orifice. Inner vibraculum with large, bulbous chamber, perforated like frontal wall on which it reposes, extending from proximal border of zooecium to orifice ; distal margin of vibraculum with crescent-shaped slit. Inner vibracula of adjacent series form an irregular, zig-zag ridge down mid-line of internode face. Outer vibraculum with long, horn-like chamber projecting outwards from lateral margin of zooecium ; vibracular slit straight, running down outer margin of chamber. Avicularium adventitious, single, small, placed on chamber of outer vibraculum below its opening. Rostrum short, pointed, directed distally and slightly outwards. Opening divided by pivotal bar at least in some specimens. Ovicell hyperstomial, elongate, globular, curved markedly outwards from longi- tudinal axis of zooecium, separated from distal zooecium by a furrow. Surface convex, smooth, perforated by irregularly spaced pores, 2-3 of which are commonly larger. Opening arched but very low, probably not closed by operculum, marked by a distinct ridge. MEASUREMENTS : Lz (10) 0-349 (0-0224) rnm., 0-32-0-37 mm. Iz (10) 0-321 (0-0340) mm., 0-26-0-37 mm- ho (10) 0-091 (0-0092) mm., 0-08-0-10 mm. lo (10) 0-098 (0-0073) mm., 0-09-0-10 mm. Lav (10) 0-093 (0-0063) mm., 0-09-0-10 mm. Lov (4) 0-297 (0-0043) mm., 0-29-0-30 mm. REMARKS. This species differs from both C. canaliculatus and C. grignonensis, not only in the position of the frontal avicularium and in lacking dorsal vibracula, but also in its more robust zooecia. Specimens of C. canaliculatus from the Sables de Fresville (Lutetian) at Gourbesville (Manche), France (supplied by Mr. Dennis Curry), are shorter (mean Lz = 0-30 mm. for 5 zooecia), and specimens of C. grignonensis from the same sample are narrower (mean Iz = 0-26 mm. for 5 zooecia). The ovicells of the Gourbesville specimens do not show the characteristic ornamentation of the ovicell described by Canu (1908 : 88, 89), i.e. channelled in C. canaliculatus and carinate in C. grignonensis. Family TUBUCELLARIIDAE Busk Genus TUBUCELLA Canu & Bassler 1917 Tubucella Canu & Bassler : 62. TYPE SPECIES (by original designation). Eschara mamillaris Milne Edwards 1836 : 336, pi. ii, fig. 10. Eocene (Lutetian) ; vicinity of Paris, France. DIAGNOSIS. Frontal wall calcareous, inflated, perforated with evenly spaced, quincuncially arranged pores. Orifice small, semi-circular, hidden completely by a tubular peristome equal in length to the frontal wall. Ascopore not much larger than frontal pores, simple, single, placed on frontal wall near peristomial suture. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 85 Scattered frontal avicularia present in some species. Ovicell peristomial, a small swelling near base of peristome of fertile zooecium ; fertile peristomes of different form from infertile ones. Zoarium erect, bilaminar, not jointed. REMARKS. This genus was established as a sub-genus of Tubucellaria d'Orbigny, but there is no tendency for the zoarium of Tubucella to be jointed, and avicularia, unknown in Tubucellaria, are present in many, though not the type, species of Tubucella. Tubucella mamillaris (Milne Edwards) (Text-figs. 62-64) 1836 Eschara mamillaris Milne Edwards : 336, pi. n, fig. 10. 190 7 a Tubucellaria mamillaris (Milne Edwards) Canu : 515. 1908 Tubucellaria mamillaris (Milne Edwards) ; Canu : 78, pi. 6, figs. 3-6. 1910 Tubucellaria mamillaris (Milne Edwards) ; Canu : 848. 1917 Tubucellaria (Tubucella} mamillaris (Milne Edwards) ; Canu & Bassler : 62. 19180 Tubucella mamillaris (Milne Edwards) Canu : 358. 1925 Tubucellaria mamillaris (Milne Edwards) ; Canu : 47. 19290 Tubucella mamillaris (Milne Edwards) ; Canu & Bassler : 46. 1933 Tubucellaria mamillaris (Milne Edwards) ; Dartevelle : 102. 1935 Tubucellaria mamillaris (Milne Edwards) ; Dartevelle : 116. 1946 Tubucellaria mamillaris (Milne Edwards) ; Buge : 436. 1949 Tubucella mamillaris (Milne Edwards) ; Balavoine 774. 1956 Tubucella mamillaris (Milne Edwards) ; Balavoine 324. 1957 Tubucella mamillaris (Milne Edwards) ; Balavoine 192. 1960 Tubucella mamillaris (Milne Edwards) ; Balavoine 246. FIGURED SPECIMENS. 0.48772 (Text-fig. 62), 0.48773 (Text-fig. 63), 0.48774 (Text-fig. 64). ADDITIONAL MATERIAL. Six specimens, D . 48775-0 . 48780. DIAGNOSIS. Tubucella with most of the peristome pitted, sessile, and bordered by large pores, and with the free portion longitudinally costate ; ascopore on a small frontal prominence separated from peristome by a single row of pores ; fertile per- stomes short, with a disto-lateral crown of pores ; avicularia lacking. DESCRIPTION. Zoarium erect, bilaminar, cylindrical to compressed, typically forming lobate or flabellate branches originating as a cylinder of 6-8 rows of zooecia and widening into flattened fronds with 12-15 rows of zooecia on each side. Zooecia arranged in longitudinal rows, those in adjacent rows alternating in position. Zooecia elongate, tubular, consisting of two regions of about equal length : a proximal frontal region bearing the ascopore, and a distal peristome bearing the secondary orifice. Zooecia separated by narrow, low threads. Frontal wall moder- ately thick, slightly convex, usually somewhat elevated immediately round the asco- pore. Surface evenly perforated with large, circular, quincuncially arranged pores. Ascopore slightly larger than frontal pores, tilted distally, placed very near peris- tomial suture, with only a single row of frontal pores between. 86 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Orifice small, semi-circular, with straight proximal lip, tilted distally, completely hidden from frontal view by peristome. Peristome sessile for most of its length, becoming free only at distal extremity. Sessile portion pitted and bordered on each side by a row of 5-8 circular pores, larger than the frontal pores. Free portion longitudinally costate or with longitudinal rows of small tubercles. Secondary orifice circular. Ovicell small, peristomial, formed as a globular swelling in distal wall of peristome, near its base, hidden completely from frontal view. Ovicelled zooecia with frontal wall like that of non-ovicelled ones, but with much shorter and wider sessile peristome 1.00mm 63 0.500mm FIGS. 62-64. Fig. 62. Tubucella, mamillaris (Milne Edwards). 0.48772. Outline of a zoarial fragment. Fig. 63. Tubucella mamillaris (Milne Edwards). 0.48773. Five ordinary zooecia. Fig. 64. Tubucella mamillaris (Milne Edwards). 0.48774. Two ovicelled zooecia and a small portion of a third, distal one. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 87 bordered laterally and distally by a crown of 12 large pores. Free peristome short, longitudinally costate, with a slightly widened secondary orifice. Avicularia lacking. MEASUREMENTS : Ordinary zooecia Lz (7) 0-812 (0-0780) mm., 0-67-0-90 mm. Iz (7) 0-211 (0-0269) mm., 0-17-0-24 mm. ho (7) o-ioo (0-0146) mm., 0-09-0-12 mm. ~\ lo (7) 0-106 (0-0190) mm., 0-09-0-13 mm. /secondaiy ormce Length of peristome (7) 0-478 (0-0726) mm., 0-38-0-52 mm. Ovicelled zooecia ho (5) 0-099 (00076) mm., 0-09-0-11 mm. ~\ i ( > secondary orifice lo (5) 0-144 (0-0112) mm., 0-14-0-16 mm. j Length of peristome (5) 0-333 (0-0200) mm., 0-30-0-35 mm. DISTRIBUTION. Eocene (Lutetian, Auversian) ; France, Belgium. Oligocene (Stampian) ; France. Miocene (Aquitanian) ; France. Family ADEONIDAE Hincks Genus TEICHOPORA Gregory 1893 Teichopora Gregory : 249. ? 19076 Poristoma Canu : 154. TYPE SPECIES (by monotypy). Teichopora clavata Gregory 1893 : 249, pi. 31, figs. 5-7. Eocene (Bartonian) ; Barton Clay, Barton, Hampshire. DIAGNOSIS. Frontal wall calcareous, smooth or granular, with single row of areolae continuing round distal margin. No ascopore or spiramen. Orifice sub- circular, with broad, U-shaped sinus proximally, without condyles. Peristome long enough to hide orifice, but immersed in thick frontal. Secondary orifice sub-circular, unmodified. Avicularia adventitious, small, single or paired, placed on proximo- lateral margins of orifice, becoming enclosed in thickening peristome so as to be nearly invisible exteriorly. Additional small, circular avicularia sometimes scattered over frontal. Gonoecia probably not of different form from zooecia. Zoarium erect, bilaminar, arborescent. REMARKS. This genus has been poorly understood because its oral structure has never been clarified. Gregory (1893 : 249) described the orifice of the type species as simple and sub-circular, but he did not differentiate between the primary and secondary orifices. Actually, the orifices in the holotype (49733, Edwards Collection) are so filled with quartz grains that the structure of the primary orifice is completely obscured, but a paratype on the same slide, however poorly preserved in other respects, has a well-preserved orifice with a sinus but without condyles. Canu & Caillot (1932 : 12), Canu (1926 : 456, 457), Davis (1934 : 228, 229), and Dartevelle (1933 : 108 ; 1936 : 29) all have confused Teichopora with Bracebridgia 88 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX MacGillivray. Mucronella pyriformis Busk, the Recent Indo-Pacific type species of Bracebridgia (lectotype, Department of Zoology, British Museum (Natural History), 87.12.9.615, Challenger Collection), differs from the European and British Eocene species in having a non-sinuate, semi-circular orifice with a short, broad lyrula occupying almost the entire proximal margin ; and vicarious, rather than adventi- tious avicularia. These differences are usually considered to be of generic magnitude in the Adeonidae ; therefore, the species exemplified by T, clavata and including at least Eschara syringopora Reuss in addition (see Waters 1891 : 20, pi. 3, figs. 2-4) must be excluded from Bracebridgia. Canu (19076 : 154) introduced the genus Poristoma (consistently misspelt " Poro- stoma " in later works) for Adeonidae having " an avicularium developed in the peristomie or on the peristome " (translation). He listed under this generic name one nomen nudum (Poristoma parisiensis) , one new species (P. incisa), and two species previously described (T. clavata Gregory and Eschara polymorpha Reuss). Selection of a type species for Poristoma, apparently an action that has not yet been taken, should be contingent upon a detailed study of P. incisa and E. polymorpha. Should T. clavata be chosen as type species, Poristoma would, of course, become an objective synonym of Teichopora. Another genus which might be confused with Teichopora is Meniscopora Gregory, the type species of which, M. bigibbera Gregory (1893 : 251, pi. 31, figs. 8, 9 ; holotype 49732, Edwards Collection) from the Lower Bracklesham Beds, differs from Teicho- pora in having oral condyles and the avicularium always outside the peristome. Teichopora clavata Gregory (Text-figs. 65, 66) 1893 Teichopora clavata Gregory : 249, pi. 31, figs. 5-7. 19076 Poristoma clavata (Gregory) Canu : 155, pi. 20, figs. 14, 15. 1925 Bracebridgia clavata (Gregory) Canu : 47. 1926 Bracebridgia gyrinus Canu : 756, pi. 28, figs. 3-6. ?i926 Bracebridgia grandis Canu : 757, pi. 28, figs, i, 2. 1929 Teichopora clavata Gregory ; Burton : 328. ?i92ga Bracebridgia grandis Canu ; Canu & Bassler : 49. I933 Bacebridgia [sic] gyrinus Canu ; Dartevelle : 108. ?J933 Bacebridgia [sic] grandis Canu ; Dartevelle : 108. 1933 Meniscopora clavata (Gregory) Dartevelle : 114. 1936 Bacebridgia [sic] gyrinus Canu ; Dartevelle : 29. ?i936 Bacebridgia [sic] grandis Canu ; Dartevelle : 29. 1946 Bracebridgia clavata (Gregory) ; Buge : 436. HOLOTYPE. 49733, Edwards Collection. Barton Clay ; Barton, Hants. Figured by Gregory (1893, pi. 31, fig. 5). FIGURED SPECIMENS. 0.48781 (Text-fig. 65), 0.48782 (Text-fig. 66). ADDITIONAL MATERIAL. Seventeen specimens, D . 48783-0 . 48799. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 70 66 FIGS. 65-70. Fig. 65. Teichopora clavata Gregory. 0.48781. Three zooecia with single avicularium. Fig. 66. Teichopora clavata Gregory. 0.48782. Four zooecia with paired avicularia. Fig. 67. Schizostomella curryi sp. nov. 0.48800. Holotype. Nearly complete gonoecium. Fig. 68. Schizostomella curryi sp. nov. 0.48801. Paratype. Four ordinary zooecia. Fig. 69. Schizostomella curryi sp. nov. 0.48802. Paratype. Broken gonoecium. Fig. 70. Schizostomella curryi sp. nov. 0.48803. Paratype. Two avicularian zooecia. 90 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX DIAGNOSIS. Teichopora with the oral avicularia either single or paired, but never meeting across the secondary orifice, the rostrum directed distally and slightly outwards. DESCRIPTION. Zoarium erect, arborescent, bilaminar, compressed, composed of longitudinal rows of zooecia, those in adjacent rows alternating in position. Zooecia club-shaped, separated by very faint furrows. Length more than twice width. Frontal wall thick, strongly convex, with a prominent median longitudinal ridge connecting secondary orifice with proximal margin of zooecium. Frontal surface finely granular, with a single row of 20-30 large areolae evenly spaced round entire margin, and one or two additional pits just proximal to peristome on heavily calcified zooecia. Orifice sub-circular to oval, the distal margin broadly rounded, the proximal margin curved to a slightly shorter radius to form a broad, median sinus. Peristome erect, not hiding distal areolae, very thick, long enough to hide orifice, but immersed in thickened frontal. Secondary orifice sub-circular, unmodified. Avicularia adventitious, single or paired, placed on peristome at proximo-lateral margin of orifice, becoming enclosed by peristome as calcification advances. Rostrum rounded, directed distally and slightly outwards. Pivotal structure lacking. Gonoecia, if present, not sufficiently differentiated from zooecia to be recognizable ; zooecia with paired avicularia and slightly enlarged secondary orifice may be gonoe- cia. MEASUREMENTS : Zooecia with single avicularium Lz (5) 0-734 (0-0229) mm., 0-71-0-76 mm. Iz (5) 0-277 (°*OI43) mm., 0-26-0-30 mm. ho (5) 0-132 (0-0115) mm., 0-12-0-14 mm. ~\ second orifice i / v /- / \ /^ oCL-LHJLU.cH V vJllllL/C lo (5) 0-126 (0-0153) mm., 0-11-0-14 mm- J Lav (5) 0-106 (0-0130) mm., 0-09-0-13 mm. Zooecia with paired avicularia Lz (5) 0-530 (0-0264) nun., 0-50-0-54 mm. Iz (5) 0-311 (0-0155) mm., 0-29-0-33 mm. ho (5) 0-156 (0-0127) mm., 0-14-0-17 mm. ~\ .- , f( r ; ' ( > secondary orifice lo (5) 0-169 (0-0195) mm., 0-14-0-19 mm. J Lav (5) 0-075 (0-0094) mm., 0-07-0-09 mm. REMARKS. Canu (1926 : 757) separated the specimens from the French Auversian that he had previously identified with T. clavata and gave them the new name B. grandis because of their larger zooecia and orifices. At the same time he (1926 : 756) described B. gyrinus as a new species differing from B. grandis " only in its smaller dimensions and not regularly elliptical orifice " (translation). B. gyrinus thus appears to be a synonym of T. clavata, and B. grandis probably is also. The holotype of T. clavata has its orifices filled with quartz grains, but the oral avicularium appears to be visible in one zooecium. The " gonoecia " described CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 91 and illustrated by Gregory (1893 : 249, pi. 31, fig. 6) in paratype 49757, Edwards Collection, seem to be frontally thickened ordinary zooecia from near the zoarial base. T. syringopora (Reuss) differs from T. clavata principally in having the oral avi- cularia always paired and their rostra directed transversely inwards so that they frequently meet across the proximal part of the secondary orifice making a false ascopore (see Waters 1891 : pi. 3, figs. 2, 3). DISTRIBUTION. Eocene (Ypresian, Lutetian, Bartonian) ; France. Eocene (Auversian) ; England, ? Belgium. Eocene (Bartonian) ; England. Genus SCHIZOSTOMELLA Canu & Bassler 1908 Schizostoma Canu : 69 (non Bronn 1834). 1927 Schizostomella Canu & Bassler : 20, 38. TYPE SPECIES (by original designation). Schizostoma crassum Canu 1908 : 70, pi. 8, figs. 6-8. Eocene (Lutetian) ; vicinity of Paris, France. DIAGNOSIS. Frontal wall calcareous, finely granular to coarsely tuberculate or with large, irregular gibbosities. Areolae in one row or two, the outer one continuing round distal margin. Orifice oval, with distinct median-proximal sinus, deep and usually narrow, sometimes limited by small condyles. Peristome thick, becoming immersed in thickened frontal. Secondary orifice oval, without sinus. Avicularia adventitious, frontal, single or paired, placed on lateral margins of zooecium, usually near orifice, sometimes on distal portion of gonoecium. Rostrum rounded, cross-bar or condyles lacking. Vicarious avicularia, modified from ordinary zooecia by oral enlargement, present in some species. Gonoecia larger than zooecia, with wide, elliptical orifice, single or multiple ascopore (rarely lacking), and swollen distal portion with imperforate, marginally areolate surface. REMARKS. This characteristic British and European Tertiary genus has not yet been found in the New World. The following species, in addition to the type and 5. curryi sp. nov. and 5. liancourti (Canu) described below, seem to have been referred correctly to Schizostomella : Schizostoma aviculiferum Canu (1908 : 71, pi. 8, fig. 12), Lutetian ; France. Schizostoma denticulatum Canu (1908 : 72, pi. 7, figs. 14-16), Ypresian-Lutetian ; France. Eschar ellina parnensis d'Orbigny (Canu 1908 : 74, pi. 8, figs. 14-16), Lutetian ; France. Schizoporella magnoaperta Gregory (1893:239, pi. 33, fig. 9), London Clay; Sheppey, Kent. Barton Beds ; Barton, Hants. Auversian ; France. Eschara socialis Busk (Lagaaij 1952 : 120, pi. 13, figs. 4, 5, 7, 8), Coralline Crag ; Suffolk. Pliocene ; Holland. Schizostoma gibbosum Canu (Buge 1957 : 296), Miocene and Pliocene ; France. 92 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Schizostoma helveticum Canu & Lecointre (Buge 1957 : 297), Miocene and Pliocene ; France. Eschara heteromorpha Reuss (Canu 19146 : 472, pi. 14, figs. 1-4, who erroneously referred it to Metrarabdotos), Oligocene ; France. i. Schizostomella curryi11 sp. nov. (Text-figs. 67-70) HOLOTYPE. 0.48800 (Text-fig. 67). PARATYPES. 0.48801 (Text-fig. 68), 0.48802 (Text-fig. 69), 0.48803 (Text-fig. 70) ; 0.48804-0.48905 (102 specimens) ; L.S.U. 8038. DIAGNOSIS. Schizostomella with relatively large zooecia having gibbosities and 24-28 areolae ; frontal avicularia usually paired, slightly removed from orifice ; vicarious avicularia little modified from zooecia ; gonoecia with a small distal avicularium and a single, median ascopore. DESCRIPTION. Zoarium erect, arborescent, branching, composed of compressed, bilaminar fronds, with zooecia arranged in as many as 12 longitudinal rows, those in adjacent rows alternating in position. Number of zooecial rows increases distally by bifurcation. Base of zoarium small, encrusting. Zooecia rhomboidal to club-shaped, very regularly arranged, not distorted round the gonoecia, separated by a faint groove. Communication between zooecia of the same series and adjacent series by simple pores placed in a single line of 14-16 near base of distal and disto-lateral walls. Zooecial length nearly twice width. Frontal surface finely granular, with larger gibbosities especially on central portion. Areolae small and circular proximally and laterally, smaller and slit-like distally, disposed in a single, evenly spaced row of 24-28 entirely round zooecial margin. Interareolar costules, where present, limited to periphery. Orifice nearly terminal on frontal surface, small, not inclined to frontal plane. Distal portion semi-circular, with evenly rounded, smooth margin. Proximal portion straight, but interrupted medially in a deep, V-shaped or linear sinus. Peristome thin and short, never standing much above frontal surface nor obscuring primary orifice from frontal view. Secondary orifice oval, larger than primary one. Avicularia dimorphic : small, adventitious and large, vicarious. Adventitious avicularia usually paired, rarely single or absent, placed on lateral corners of frontal wall over one or two areolae, having rounded rostra and lacking pivotal bar or condyles. Vicarious avicularia rare and sporadic, developed from zooecia by enlargement of orifice and oral region to greatly varying degrees. Avicularian orifice with broad distal shelf and greatly widened sinus flanked by small lateral condyles. Gonoecia slightly longer and broader than ordinary zooecia, with small, paired, adventitious avicularia developed proximally and laterally to orifice and a third, small avicularium placed medially on distal margin. Gonoecial orifice a very wide 11 After Mr, Dennis Curry. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 93 ellipse separated from the single, rounded ascopore. Oral region strongly raised, the post-oral portion with a distal marginal crown of enlarged areolae, but not other- wise ornamented. MEASUREMENTS : Ordinary zooecia Lz (10) 0-496 (0-0442) mm., 0-41-0-56 mm. Iz (10) 0-239 (0-0171) mm., 0-21-0-26 mm. ho (10) 0-090 (o-oroi) mm., 0-08-0-11 mm. lo (10) 0-088 (0-0107) mm-» 0-08-0-11 mm. Lav (9) 0-070 (0-0153) mm., 0-05-0-09 mm. Gonoecia Lz (5) 0-590 (0-0200) mm., 0-57-0-62 mm. Iz (5) 0-332 (0-0398) mm., 0-27-0-38 mm. ho (4) 0-051 (0-0099) mm-» 0-04-0-06 mm. lo (5) 0-188 (0-0326) mm., 0-17-0-21 mm. Lav (6) 0-067 (0-0219) mm., 0-05-0-10 mm. Avicularian zooecia Lz (3) 0-616 (0-0342) mm., 0-58-0-65 mm. Iz (3) 0-222 (0-0308) mm., 0-20-0-26 mm. ho (4) 0-165 (0-0214) mm., 0-14-0-19 mm. lo (4) 0-132 (0-0204) mm., 0-11-0-15 mm. REMARKS. The avicularian zooecia of this species are like the B-zooecia of Steganoporella in relation to the ordinary zooecia. It is, of course, possible that they represent a form of dimorphism different from avicularian, but the oral enlargement, markedly different from that of the gonoecia, is almost certainly a concomitant of opercular enlargement for presumably the same function as the development of avicularian mandibles. 5. curryi is very close to 5. gibbosa (Canu), a French Miocene and Pliocene species, from which it differs in having avicularian zooecia and in having the gonoecia with a distal avicularium and a single, median ascopore. 2. Schizostomella liancourti (Canu) (Text-figs. 71, 72) 1908 Schizostoma liancourti Canu : 72, pi. 22, figs. 10, n. 1946 Schizostomella liancourti (Canu) Buge : 438. FIGURED SPECIMENS. 0.48906 (Text-fig. 71), 0.48907 (Text- fig. 72). ADDITIONAL MATERIAL. Five specimens, 0.48908-0.48912. DIAGNOSIS. Schizostomella with relatively small zooecia having tuberculate frontal walls with 14-20 areolae ; secondary orifice elongate, constricted at middle ; 94 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX frontal avicularia small, usually single, placed on peristome ; vicarious avicularia lacking ; gonoecium (fide Canu 1908 : 72) without distal avicularium or ascopore. DESCRIPTION. Zoarium erect, arborescent, branching, composed of compressed, bilaminar fronds originating from a small, encrusting base. Zooecia arranged in 6-12 longitudinal rows on each side of frond, those in adjacent rows alternating in position. 72 FIGS. 71-75. Fig. 71. Schizostomella liancourti (Canu). 0.48906. Four heavily calcified zooecia. Fig. 72. Schizostomella liancourti (Canu). 0.48907. Four lightly calcified zooecia. Fig. 73. Adeonellopsis selseyensis sp. nov. 0.48913. Holotype. Two gonoecia (on the left) and two zooecia. Fig. 74. Adeonellopsis punctata (Canu). 0.48914. Four zooecia. Fig. 75. Adeonellopsis punctata (Canu). 0.48915. Gono- ecium. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 95 Zooecia club-shaped, very regularly arranged, separated by a shallow groove. Communication between zooecia by simple pores. Zooecial length about one and a half times width. Frontal wall thick, convex, becoming flat with age, highest at proximo-lateral corners of orifice. Frontal surface tuberculate in thin-walled zooecia, becoming finely granular as wall thickens. Areolae small, circular, nearly occluding as calcification progresses, slit-like round distal margin, disposed in a single, evenly spaced row of 14-20 entirely round zooecial margin. Interareolar costules lacking. Orifice sub-terminal on frontal surface, small, not inclined to frontal plane. Distal portion evenly rounded, smooth, semi-circular. Proximal portion nearly straight, interrupted medially in a deep but broad, U-shaped sinus. Peristome very thick, never standing much above frontal surface, but obscuring primary orifice from frontal view. Secondary orifice larger than primary one, constricted at middle so as nearly to form a spiramen. Avicularia monomorphic, adventitious, usually single, rarely paired or multiple, placed on peristome near proximo-lateral corners of orifice, or on lateral margins of frontal wall, over one or two areolae. Rostrum rounded. Pivotal structures lacking. Gonoecia lacking in material at hand. MEASUREMENTS : Lz (10) 0-355 (0-0418) mm., 0-30-0-43 mm. Iz (10) 0-216 (0-0335) mm., 0-17-0-26 mm. ho (9) 0-089 (0-0172) mm., 0-07-0-12 mm. \ .~ i / \ -> / ^ secondary orifice lo (n) 0-078 (0-0092) mm., 0-07-0-09 mm. J Lav (7) 0-077 (O'OJS1) mm., 0-05-0-09 mm. REMARKS. This species, hitherto known only from the French Lutetian, is easily distinguished from other species of Schizostomella by its constricted secondary orifice and its thick perstome on which the avicularia are placed. S. crassa (Canu) has a similar peristome, but its secondary orifice is different, and its frontal avicularium is well removed from the peristome. Canu (1908 : 72) characterized S. liancourti as always having paired or multiple avicularia, but his own illustrations (1908 : pi. 22, figs. 10, n) show zooecia with single avicularium. DISTRIBUTION. Eocene (Lutetian) ; France. Genus ADEONELLOPSIS MacGillivray Adeonellopsis selseyensis sp. nov. (Text-fig. 73) HOLOTYPE. 0.48913 (Text-fig. 73). DIAGNOSIS. Adeonellopsis with small, rhombic zooecia ; primary orifice visible in frontal aspect, proximal lip serrate ; ascopore simple or compound with as many as 96 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 5 perforations ; avicularia frontal, paired lateral and occasionally single distal as well ; gonoecia differ only slightly in size, shape, and oral structure from zooecia. DESCRIPTION. Zoarium erect, arborescent, composed of compressed, bilaminar fronds with more than 7 longitudinal rows of zooecia on each side, the zooecia in adjacent rows alternating in position. Zooecia rhombic, approximately equilateral, separated by distinct furrows. Length and width sub-equal. Frontal wall thick, markedly convex, highest round the proximal lip of the orifice. Frontal surface very finely granular, without tubercles or costules, margined by a single row of 18-20 small, circular areolae evenly spaced round the whole periphery. Ascopore placed just proximal to mid-length in a small, median pit, simple or compound, formed of a small circular disc with 2-5 minute perforations. Orifice removed slightly from distal end of frontal surface, small, semi-circular, surrounded by a thick peristome, but not hidden from frontal view. Distal margin evenly rounded, smooth ; proximal margin nearly straight, finely serrated. Peri- stome not elevated above frontal surface. Secondary orifice sub-circular. Avicularia adventitious, frontal, small, multiple, their rostra rounded or slightly attenuated ; pivotal structures lacking. One pair of avicularia present on each zooecium and gonoecium, one in each lateral corner ; rostra directed inwards and slightly distally or proximally. An additional unpaired avicularium with rounded rostrum present on a few zooecia and gonoecia, placed on the distal part of the peri- stome, either on the mid-line or slightly to one side. Vicarious avicularia unknown. Gonoecia slightly wider and longer than zooecia. Frontal wall like that of zooecia but more swollen round orifice. Ascopore, areolae, and avicularia similar to those of zooecia. Orifice slightly wider and shorter than that of zooecia, but with similar serration of proximal lip. MEASUREMENTS : Zooecia Lz (5) 0-282 (0-0264) mm., 0-26-0-33 mm. Iz (5) 0-231 (0-0148) mm., 0-22-0-26 mm. ho (5) 0-056 (0-0047) mm., 0-05-0-06 mm. \second Qrifice lo (5) 0-067 (°'°°38) mm., 0-06-0-07 mm- J Lav (5) 0-044 (0-0038) mm., 0-04-0-05 mm. Gonoecia Lz (4) 0-355 (o-ono) mm., 0-34-0-38 mm. Iz (4) 0-250 (0-0214) mm., 0-22-0-27 mm- ho (4) 0-043 (o-oooo) mm., 0-04 mm. "\ ; ( ; ( > secondary orifice lo (4) 0-081 (0-0049) mm-, 0-08-0-09 mm- J Lav (4) 0-047 (0-0049) mm-» 0-04-0-05 mm. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 97 Adeonellopsis punctata (Canu) (Text-figs. 74, 75) > 19076 Adeonella punctata Canu : 149, pi. 20, fig. 2. 1920 Cribricella punctata (Canu) Canu & Bassler : 564. 1946 Adeonella punctata Canu ; Buge : 436. 1960 Adeonella punctata Canu ; Balavoine : 246. 1963 " Adeonellopsis punctata (Reuss) " Malecki : 128, pi. 14, fig. 4. FIGURED SPECIMENS. 0.48914 (Text-fig. 74), 0.48915 (Text-fig. 75). ADDITIONAL MATERIAL. Forty-three specimens, 0.48916-0.48959. DIAGNOSIS. Adeonellopsis with large, rhomboidal zooecia ; primary orifice hidden from frontal view, the proximal lip smooth ; ascopore single and simple on zooecia, large and compound on gonoecia, with more than 20 perforations ; avicularium frontal, single, median, between orifice and ascopore, the rostrum directed distally ; gonoecia much larger than zooecia, the orifice greatly widened. DESCRIPTION. Zoarium erect, arborescent, branching, composed of compressed, bilaminar fronds, originating from a small, encrusting base, expanding rapidly to as many as 20 longitudinal rows of zooecia on each side. Zooecia of adjoining rows alternate in position ; new rows added distally by intercalation. Zooecia elongate, rhomboidal, variable in size and shape, but not distorted round gonoecia, separated by a furrow. Length nearly two-and-a-half times width. Frontal wall thick, only slightly convex, the area round the ascopore depressed. Surface finely granular, sometimes with small tubercles, margined entirely by a single, evenly spaced row of 20-30 small, circular areolae. Scattered areolae of a second, inner row sometimes present proximal to peristome and at proximal end of zooecium. Ascopore single, simple, circular, small, placed just proximal to mid- length in a very deep depression. Orifice just short of distal end of zooecium, small, semi-circular, hidden by peristome from frontal view. Distal margin evenly rounded ; proximal margin nearly straight, smooth. Peristome thick, elongate, but generally buried in the frontal so that only distal and lateral portions are raised, sometimes enough to form a hood-like projection over orifice. Secondary orifice about same size and shape as primary one. Avicularium adventitious, frontal, placed between orifice and ascopore, with ros- trum directed distally, either longitudinally or slightly obliquely. Rounded end of avicularium stops short of ascopore-pit, and rostrum does not quite reach peristome. Rostrum raised, pointed, slightly channelled. Pivotal structure lacking. Gonoecium slightly longer than and about twice as wide as zooecium. Frontal wall swollen all round orifice, with additional rows of areolae developed proximal to orifice. Ascopore greatly enlarged, compound, consisting of a flat, roughly circular disc perforated by 20 or more small pores. Orifice wider than that of zooecium, the proximal lip distinctly convex. Avicularium smaller than that of zooecium. GEOL. 13, I 7 g8 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX MEASUREMENTS : Zooecia Lz (10) 0-508 (0-0689) mm-, 0-43-0-62 mm. Iz (10) 0-210 (0-0198) mm., 0-19-0-24 mm. ho (10) 0-080 (0-0099) mm., 0-06-0-09 mm- I . ; ( > secondary orifice lo (10) 0-091 (0-0070) mm., 0-09-0-10 mm. j Lav (10) 0-160 (0-0285) mm., 0-13-0-22 mm. Gonoecia Lz (2) 0-573 (0-0363) mm., 0-55-0-60 mm. Iz (2) 0-423 (0-0665) mm., 0-38-0-47 mm. ho (2) 0-077 (0-0121) mm., 0-07-0-09 mm. ^1 > secondary orifice lo (2) 0-197 (0-0121) mm., 0-19-0-20 mm. J Lav (2) 0-124 (0-0181) mm., 0-11-0-14 mm. REMARKS. This species does not appear to be the same as the German Oligocene species reported by Reuss (1865 : 649, pi. 12, figs. I, 2) as Eschara coscinophora Reuss and placed in synonymy with A. punctata by Canu (19076 : 149). E. coscinophora from the Miocene of Austria is conspecific with E. imbricata Philippi according to Lagaaij (1952 : 120). Therefore, the synonymy given by Malecki (1963 : 128) is too inclusive. The generic assignment of A. punctata is difficult to make. The presence of an ascopore rather than a spiramen excludes it from Adeonella Busk (see Harmer 1957 : 802). Three genera of Adeonidae have well-developed ascopores : Adeona Lamouroux, Adeonellopsis MacGillivray, and Reptadeonella Busk. Harmer (1957 : 789-802, 814-818) differentiated these genera principally upon zoarial characters, at least some of which are not always evident in fragmentary fossil material. Cheet- ham & Sandberg (1964 : 1039) characterized Reptadeonella as having a simple, single or rarely double ascopore on both zooecia and gonoecia ; a sub-oral, median avicu- larium directed transversely or longitudinally or obliquely distally, its rostrum reaching the peristome ; occasionally additional frontal avicularia near the proximal margin of the zooecium ; and gonoecia with wide, short secondary orifice but not otherwise differentiated from the zooecia. The type species of Adeona, A. grisea Lamouroux (chosen by Gregory 1893), represented in the Department of Zoology, British Museum (Natural History) by specimens such as 99.7.1.2750 and 2753, Busk Collection, has a simple, single or double ascopore on both zooecia and gonoecia ; a single, frontal avicularium flanking the ascopore distally and directed transversely or obliquely distally, its rostrum not reaching the peristome ; and gonoecia larger than the zooecia, with short, wide orifices. Adeonellopsis seems to be much more variable than the other two ascopore-bearing Adeonid genera. The type species, A. foliacea MacGillivray (chosen by Canu & Bassler 1917), represented in the Department of Zoology, British Museum (Natural History) by specimen 97.5.1.709, has a compound ascopore on both zooecia and CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 99 gonoecia ; multiple sub-oral avicularia, both median and lateral, none with rostra reaching the peristome, on the zooecia ; supra-oral avicularia additionally on the gonoecia ; and gonoecia larger than the zooecia, with widened orifices. Because of its compound gonoecial ascopore, A. punctata seems to fit best in Adeonellopsis, even though it has a single sub-oral avicularium and a simple zooecial ascopore. In these respects it resembles Adeonellopsis arculifera (Canu & Bassler) and A. parvipuncta MacGillivray (see Harmer 1957 : pi. 56, figs. 13, 14, 16). DISTRIBUTION. Eocene (Lutetian) ; France. Eocene (Ludian) ; Poland. Family CELLEPORINIDAE Harmer Genus CELLEPORINA Gray Celleporina thomasi12 sp. nov. (Text-figs. 76, 77) HOLOTYPE. 0.48960 (Text-fig. 76). PARATYPES. 0.48961 (Text-fig. 77) ; 0.48962-0.48970 (9 specimens) ; L.S.U. 8039. DIAGNOSIS. Celleporina with median sub-oral avicularium on a prominent umbo and with lateral-oral bosses ; distal hood of ovicell forming a narrow rim within which is a row of small, marginal pores. DESCRIPTION. Zoarium encrusting, unilaminar, the zooecia arranged in irregular longitudinal rows, those in adjacent rows alternating in position. Portions of zoaria probably rose as unilaminar tubes from the substrate. Zooecia irregularly rhomboidal, separated by narrow, raised threads with a faint groove at the crest. Length almost twice width. Frontal wall thin, nearly flat except at sub-oral umbo. Surface tuberculate, radially striated, with large, circular areolae in a single marginal row of 4-6 on each side and usually an additional 2-3 proximally. Interareolar costules weak, peripheral. Sub-oral umbo, lacking on a few zooecia, usually massive. Orifice commonly partly obscured in frontal view by proximal umbo and lateral bosses, oval, the distal portion semi-circular, the proximal portion a deep, wide, rounded sinus between a pair of rudimentary, widely spaced, lateral condyles. Peristome an irregular distal ridge, not differentiated from frontal surface, thick, smooth, terminating on each side of the orifice in a tubercle which enlarges to form a boss with advancing calcification, and which connects with the sub-oral umbo. Avicularium adventitious, sometimes lacking, single, median, sub-oral, placed on the umbo, sometimes facing into the peristome, but with rostrum directed frontally, proximally, and usually slightly to one side. Rostrum rounded ; pivotal bar complete. 12 After Dr. H. Dighton Thomas. GEOL. 13, I. 7§ ioo CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 0.500mm 78 0.500mm 79 1.00mm FIGS. 76-79. Fig. 76. Celleporina thomasi sp. nov. 0.48960. Holotype. Three heavily calcified zooecia, all with oral avicularia and one with an ovicell. Fig. 77. Celleporina thomasi ^ sp. nov. 0.48961. Paratype. Two lightly calcified zooecia lacking avicularia. Fig. 78. Kionidella hastingsae sp. nov. 0.48971. Holotype. Four zooecia, three with ovicells. Fig. 79. Kionidella hastingsae sp. nov. 0.48972. Paratype. Basal view of a hollow zoarial fragment. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 101 Ovicell hyperstomial, globular, wider than long ; surface irregularly tuberculate, with a smooth, raised distal rim within which is a marginal row of small pores. Additional pores sometimes scattered over the surface. Proximal margin of ovicell generally visible between peristomial ridges which extend on to its surface. MEASUREMENTS : Lz (10) 0-397 (0-0513) mm., 0-36-0-50 mm. Iz (10) 0-221 (0-0224) mm., 0-19-0-26 mm. ho (10) 0-115 (0-0157) mm., 0-09-0-14 mm. lo (10) o-ioi (0-0054) mm., 0-09-0-11 mm. Lav (10) 0-092 (0-0079) mm., 0-09-0-10 mm. Lov (6) 0-167 (0>OI°5) mm., 0-15-0-18 mm. REMARKS. This species, the oldest referred to Celleporina, differs from Recent species of the genus in having the ovicell with numerous pores and without the hood-like extension of the distal rim, and in having the frontal avicularium median sub-oral rather than lateral. The lateral oral bosses may foreshadow the lateral columnar avicularia so characteristic of later species. Otherwise the Upper Brackle- sham species fits well in Celleporina. Family MAMILLOPORIDAE Canu & Bassler Genus KIONIDELLA Koschinsky 1885 Kionidella Koschinsky : 67. TYPE SPECIES (chosen by Canu & Bassler 19296). K. excelsa Koschinsky 1885 : 68, pi. 7, figs. 5-12. Eocene (Lutetian) ; Bavaria. DIAGNOSIS. Frontal wall calcareous, smooth or granular, without pores or areolae. Orifice elongate, with lateral condyles and a thin, smooth peristome. Avicularia adventitious, single or paired, lacking on some zooecia, placed on lateral margins of frontal, with long or short rostrum directed inwards. Zooecia erect, prismatic, with just orifice and small area of frontal wall showing on frontal surface, communicat- ing by simple pores placed low in zooecial walls. Zoarium hollow, tubular, uni- laminar, ancestrular end closed, and distal ends of zooecia directed towards distal margin of zoarium. Ovicell hyperstomial, large, not raised above zoarial surface. REMARKS. The presence of ovicells in Kionidella was first noted by Waters in the type species (1891 : 29, pi. 4, fig. 6). Canu (in Buge 1946 : 438) later misinterpreted the ovicell as intermediate between entozooecial and entotoichal. Waters's figure indicates, and the specimens of K. hastingsae sp. nov. described below substantiate, that the ovicell of this genus is hyperstomial but deeply immersed, similar to that of Mamillopora. Canu & Bassler (19296 : 477) placed Kionidella in synonymy with Discoflustrellaria, a genus founded by d'Orbigny (1853 : 508) for three species : D. dactylus d'Orbigny (1853 : 508, not figured) from the Lutetian of France ; D. clypeiformis d'Orbigny (1853 : 508, pi. 722, figs. 2-5) and D. doma d'Orbigny (1853 : 509, pi. 722, figs. 6-10), 102 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX both Lunulitidae from the Senonian of France. Canu (1900 : 378) indicated the type species of Discoflustrellaria to be D. clypeiformis, yet Canu & Bassler (19296 : 479) seem to have regarded D. dactylus as the type species when they stated, " D'Or- bigny's name Discoflustrellaria is the older, but the French author did not give a single figure of the genotype. The latter was figured only in 1908 by Canu." This is presumably a reference to Canu (1908 : 103, pi. 9, fig. 18) who described and illus- trated D. dactylus which Canu & Bassler (19296 : 478) referred to Kionidella. Thus Canu & Bassler seem to have been mistaken in considering Kionidella a synonym of Discoflustrellaria . Kionidella hastingsae13 sp. nov. (Text-figs. 78, 79) HOLOTYPE. 0.48971 (Text-fig. 78). PARATYPE. 0.48972 (Text-fig. 79). DIAGNOSIS. Kionidella with zooecia wide and rhombic in frontal aspect, the fron- tal wall finely pitted ; oral condyles slightly proximal to mid-length, and sinus as broad as distal part of orifice ; avicularia paired, lateral, relatively small, and with pivotal bar ; rostrum rounded ; ovicells large, pitted, the fertile zooecia with widened orifice. DESCRIPTION. Zoarium cylindrical, tubular, composed of a single layer of zooecia ; ancestrular end not preserved ; axial hollow circular in cross section, slightly larger in diameter than a zooecium. Zooecia in 14 very regular longitudinal rows round surface of cylinder. Zooecia vase-shaped, their long axes perpendicular to zoarial axis, with only orifices and small part of frontal wall, of rhombic shape, exposed at outer surface. Frontal surfaces of adjoining zooecia separated by distinct grooves. Frontal wall imperforate, convex, finely pitted. Interzooecial communication by a few simple pores placed near base of distal and lateral walls. Orifice elliptical, longer than wide, divided just proximal to middle by a pair of small, lateral, proximally inclined condyles, the sinus thus formed being as broad as the distal part of the orifice. Avicularia adventitious, usually paired, placed in lateral corners of frontal surface just distal to middle of orifice. Avicularian opesia sub-circular, divided by a pivotal bar. Rostrum rounded, but somewhat elongated, directed inwards and slightly distally over orifice. Ovicell hyperstomial, convex but not much raised above zoarial surface, semi- circular, of about same length as zooecium. Surface finely pitted, imperforate. Orifice of fertile zooecia widened. 13 After Dr. Anna B. Hastings. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 103 MEASUREMENTS : Ordinary zooecia Lz (5) 0-349 (°'°693) mm., 0-26-0-44 mm- Iz (5) 0-383 (0-0237) mm-> 0'34-°'39 mm- ho (5) 0-157 (0-0097) mm., 0-14-0-17 mm. lo (6) 0-117 (o-oioo) mm., 0-10-0-13 mm. Lav (6) 0-141 (0-0200) mm., 0-13-0-17 mm. Ovicelled zooecia lo (4) 0-180 (0-0070) mm., 0-17-0-19 mm. Lov (5) 0-209 (0-0215) mm., 0-17-0-22 mm. REMARKS. This species is most like K. obliqueseriata Koschinsky (1885 : 69, pi. 7, figs. 130, b) from the Lutetian of Bavaria. The two species differ primarily in the form of the avicularium : in K. obliqueseriata it is single and has a long, pointed rostrum. K. excelsa Koschinsky and K. dactylus (d'Orbigny) differ from K. hastingsae in the form of the avicularium and also in the shape of the zooecia and orifice. Family ORBITULIPORIDAE Canu & Bassler Genus ORBITULIPORA Stoliczka Orbitulipora petiolus (Lonsdale) (Text-fig. 80) 1850 Cellepora petiolus Lonsdale : 151. 1854 Cellepora petiolus Lonsdale ; Morris : 120. 1862 Orbitulipora haidingeri Stoliczka : 91, pi. 3, fig. 5. 1867 Orbitulipora petiolus (Lonsdale) Reuss : 217, pi. i, figs, i, 2. 1881 Orbitulipora petiolus (Lonsdale) ; Mourlon : 180, 191, 202. 1889 Orbitulipora petiolus (Lonsdale) ; Vine : 163, pi. 5, fig. 10. 1893 Biselenaria off a Gregory : 235, pi. 30, figs. 4, ^a (not fig. 5). 1893 Orbitulipora petiolus (Lonsdale) ; Gregory : 253, pi. 31, figs. 12-14. 19076 Orbitulipora petiolus (Lonsdale) ; Canu : 102. 1919 Orbitulipora petiolus (Lonsdale) ; Waters : 91. 1926 Orbitulipora petiolus (Lonsdale) ; Canu : 758, pi. 30, fig. 5. I92ga Orbitulipora petiolus (Lonsdale) ; Canu & Bassler : 49. 19296 Orbitulipora petiolus (Lonsdale) ; Davis : in. 1931 Orbitulipora petiolus (Lonsdale) ; Canu & Bassler : 16, pi. 3, figs. 1-22 ; pi. 4, figs. 1-4. *933 Orbitulipora petiolus (Lonsdale) ; Dartevelle : 108. 1936 Orbitulipora petiolus (Lonsdale) ; Dartevelle : 29. !939 Orbitulipora petiolus (Lonsdale) ; Franke : 64, pi. 3, figs. 2a-d. 1963 Orbitulipora petiolus (Lonsdale) ; Malecki : 137, pi. 15, fig. 5. FIGURED SPECIMEN. 0.48973 (Text-fig. 80). ADDITIONAL MATERIAL. One hundred and forty-eight specimens, 0.48974- 0.49073. 104 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 80 81 1.00mm FIGS. 80-8 1. Fig. 80. Orbitulipora petiolus (Lonsdale). 0.48973. Lateral view of discoidal zoarium showing well-developed basal peduncle and several ovicelled peripheral zooecia. Fig. 81. Batoporaglandiformis (Gregory). 0.49074. Lateral view of globular zoarium. Three ovicelled zooecia and three vicarious avicularia are visible. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 105 DIAGNOSIS. Orbitulipora with discoidal zoarium circular in lateral outline but for the basal peduncle. DESCRIPTION. Zoarium discoidal, flat on both sides, circular in lateral outline save for the basal peduncle. Zooecia arranged in two laminae, their basal walls in contact, those of each lamina in approximate annular arrangement, the largest zooecia at the periphery, the smallest at the centre of the face. Distal margins of zooecia directed towards centre of face. The central zooecia are often covered by randomly oriented, superposed zooecia. Peduncle equal to about four zooecia in size, the exterior irregularly wrinkled and pitted, the interior smooth and hollow, extending nearly to centre of zoarium. Zooecia irregularly rhombic, pentagonal, or hexagonal, separated by deep grooves. Frontal wall moderately thick, smooth or finely granular, slightly convex, without pores or areolae. Orifice central, taking up almost the whole frontal surface of zooecium, sub- circular, unmodified but buried in thick peristome. Ovicell hyperstomial, globular, smooth, with distinct distal rim. Orifice of fertile zooecia slightly wider than that of ordinary zooecia. Ovicells developed on zooecia of outer few annulae, on the sides towards the centre of the zoarial face, or on super- posed zooecia nearer the centre of the face. Heterozooecia unknown. MEASUREMENTS : Zoarium Zoarial height (5) 2-072 (0-5977) mm., 1-26-2-87 mm. Length of peduncle (5) 0-333 (0-1071) mm., 0-17-0-44 mm. Ordinary zooecia (at periphery] Lz (10) 0-345 (0-0380) mm., 0-27-0-38 mm. Iz (10) 0-322 (0-0421) mm., 0-24-0-38 mm. ho (10) 0-132 (0-0259) mm-> 0-09-0-16 mm. lo (to) 0-153 (0-0142) mm., 0-13-0-17 mm. Ovicelled zooecia Lov (6) 0-252 (0-0280) mm., 0-24-0-31 mm. lo (6) 0-185 (0-0476) mm., 0-17-0-21 mm. REMARKS. In their detailed study of this species, Canu & Bassler (1931 : 16-22) assumed that the centre of the disc-shaped zoarial frond is the ancestrular region, the actual ancestrula lying at the top of the peduncular tube covered by superposed zooecia. The ovicells of the Upper Bracklesham specimens are typical hyperstomial ones, not recumbent as Canu & Bassler (1931 : 17) described them. Avicularia and vibracula are lacking in the Upper Bracklesham specimens ; the structures illustrated by Canu & Bassler as avicularia (1931 : pi. 3, fig. 15) and vibracula (1931 : pi. 3, fig. 21 ) might be broken ovicells. 106 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX As mentioned above (Morphology section, Zoarial Characters), Canu (1931 : 144- 147) and Canu & Bassler (1931 : 19-21) supposed Orbitulipora to be free-swimming ; Waters (1919 : 90) had earlier suggested the more likely explanation of the peduncle of Orbitulipora as a structure for attachment to the substrate. Silen (1947 : 33) noted that the reversal of the normal proximal-distal relationship in the zoarium of Orbitulipora is similar to that in the Conescharellinidae. DISTRIBUTION. Eocene (Auversian) ; France, England. Eocene (Auversian, Bartonian) ; Belgium. Eocene (Ludian) ; Poland, ? Italy. Oligocene (Lattorfian) ; Belgium, Germany. Genus BATOPORA Reuss 1867 Batopora Reuss : 223. 1929^ Atactopora Canu & Bassler : 51, non Morren. 1931 Atactoporidra Canu & Bassler : 22. TYPE SPECIES (chosen by Waters 1919). Batopora stoliczkai Reuss 1867 : 223, pi. 2, figs. 2-4. Oligocene (Lattorfian) ; Germany. DIAGNOSIS. Frontal wall calcareous, granular or smooth, without pores or areolae. Orifice sub-circular to semi-circular, with nearly straight proximal margin. Condyles, sinus and ascopore lacking. Whole frontal wall built up round orifice as a kind of peristome. Zoarium spherical, ellipsoidal, discoidal, or conical ; hollow or solid ; the base with a special enlarged pore opening into an axial hollow which runs part way up the zoarium. Zooecia erect, arranged in several laminae, the superficial ones large, with or without ovicells, the deep ones small, without ovicells ; distal margins of zooecia directed apically. Ovicell hyperstomial, globular, smooth. REMARKS. The structure of Batopora was studied in detail by Waters (1919) ; the two German Oligocene species, B. stoliczkai Reuss and B. multiradiata Reuss, considered to be the same by him (1919 : 83, 84), have been re-illustrated by Franke (1939 : pi. 2, figs. 3«, b ; pi. 3, figs. ia, b) as separate species. Canu & Bassler (19290 : 51) established Atactopora, later re-named Atactoporidra, for Orbituliporids having hollow cylindrical to globular zoaria, differing from Batopora which supposedly has conical zoaria. This difference does not seem uniform in the species referred to the two genera, and, furthermore, the type species of Batopora has the Atactoporidra-type zoarium. Batopora glandiformis (Gregory) (Text-fig. 81) 1893 Heteropora glandiformis Gregory : 261, pi. 32, fig. n. 1934 Atactoporidra glandiformis (Gregory) Davis : 205. HOLOTYPE. 6.4511, Edwards Collection. Barton Beds; Barton, Hants. FIGURED SPECIMEN. 0.49074 (Text-fig. 81). CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 107 ADDITIONAL MATERIAL. Twenty-six specimens, 0.49075-0.49100. DIAGNOSIS. Batopora with nearly solid, usually spherical zoarium having an inconspicuous basal pit which extends only a short distance up axis ; superficial zooecia having large, semi-circular orifices and well-developed ovicells ; deep zooecia with much smaller, sub-circular orifices. DESCRIPTION. Zoarium globular, usually spherical, sometimes ellipsoidal or almost discoidal ; apical end rounded, basal end slightly more pointed, with a central, circular pit which is inconspicuous and only slightly larger than a zooecial orifice and surrounded by a thin, raised collar. Zooecia arranged irregularly, their distal margins directed towards the apical end of the zoarium. Deep zooecia with only their small orifices showing between larger superficial zooecia. Zooecia erect, their major axes perpendicular to zoarial surface, with only a small part of the frontal wall, surrounding the orifice, visible frontally. Superficial zooecia protuberant, irregularly polygonal, separated by deep depressions in which orifices of deep zooecia appear. Frontal wall thin, very convex, granular, without pores or areolae. Orifice of superficial zooecia semi-circular, broadly rounded distally, slightly rounded proximally. Orifice of deep zooecia smaller and sub-circular. Peristome, spines, condyles and sinus all lacking. Ovicell hyperstomial, small, globular, imperforate, finely granular, present on superficial zooecia only. Orifice of fertile zooecia slightly wider than that of ordinary superficial zooecia. Avicularia rare, vicarious, about half as large as superficial zooecia, with rounded rostrum and pivotal bar. MEASUREMENTS : Zoarium Zoarial length (4) 1-706 (0-3086) mm., 1-32-1-96 mm. Zoarial width (4) 1-511 (0-1812) mm., 1-28-1-71 mm. Ordinary superficial zooecia Lz (8) 0-245 (0-0438) mm., 0-19-0-31 mm. Iz (8) 0-239 (°'°333) mm., 0-17-0-27 mm. ho (8) 0-105 (0-0127) mm., 0-09-0-12 mm. lo (8) 0-105 (0-0203) mm., 0-09-0-13 mm. Deep zooecia ho (4) 0-083 (0-0118) mm., 0-07-0-10 mm. lo (4) 0-073 (0-0086) mm., 0-07-0-09 mm. Ovicelled zooecia Lov (3) 0-125 (0-0131) mm., 0-11-0-14 mm. lo (3) 0-120 (0-0086) mm., 0-11-0-13 mm. io8 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX Avicularia Lav (2) 0-128 (o-oooo) mm., 0-13 mm. REMARKS. Gregory's (1893) erroneous placing of this species in the Cyclostome genus Heteropora was first noted by Waters (1919 : 92) who considered the species a young stage of Orbitulipora petiolus. Davis (1934 : 205) recognized the true affinities of the species. DISTRIBUTION. Eocene (Bartonian) ; England. VIII. REFERENCES ABILDGAARD, P. C. 1806. In MULLER, O. F. Zoologica Danica sen Animalium Daniae et Norvegiae. 3rd edit., 4 : 1-46, Pis. 121-160. Havniae. ARCHIAC, E. J. A. D'. 1846. Description des fossiles recueillis par M. Thorent dans les couches a Nummulines des environs de Bayonne. Mem. Soc. geol. Fr., Paris (2) 2 : 189-217, pis. A-E. BALAVOINE, P. 1949. Note sur la faune lutetienne de Villers-sur-le-Roule et de Venables (Eure). Bull. Mus. natn. Hist, nat., Paris (2) 21 : 772-775. - 1954. Contribution a 1'etude des Bryozoaires du Bartonien du Bassin de Paris. C. r, somm. Seanc. Soc. geol. Fr., Paris, 12 : 255-257. - 1956. Quelques Bryozoaires eocenes du Bassin de Paris et du Cotentin de la collection Gustave-F. Dollfuss. Bull. Mus. natn. Hist, nat., Paris (2) 28 : 319-325. - 1957. Nouveaux gisements de Bryozoaires dans le Lutetien du Bassin de Paris. Bull. Mus. natn. Hist, nat., Paris (2) 29 : 190-192. - 1960. Bryozoaires du Lutetien de Bois-Gouet (Loire-Atlantique). Bull. Soc. geol. Fr., Paris (7) 1 : 245-251, pis. 6, ja. BARROSO, M. G. 1950. Briozoos terciarios de algunas localidades espanolas. Boln. R. Soc. esp. Hist, nat., Madrid, 47 : 171-191, pi. 9. BASSLER, R.S. 1935. Bryozoa. Fossilium Catalogus, I. Animalia. Pars 67. 229 pp. 's-Gravenhage. - 1953. In MOORE, R. C. Ed. Treatise on invertebrate paleontology, G. Bryozoa. i-xiii, Gi-253, 175 figs. Lawrence, Kansas. BERTHELSEN, O. 1962. Cheilostome Bryozoa in the Danian deposits of east Denmark. 2nd. Ser., 83, 290 pp., pis. 1-28. Geol. Surv. Denmark, Copenhagen. BRAGA, G. 1963. I Briozoi del Terziario veneto. Boll. Soc. paleont. ital., 2 : 16-56, pis. 2-5. BROWN, D. A. 1952. The Tertiary Cheilostomatous Polyzoa of New Zealand. xii + 4O5 pp., 296 figs. British Museum (Nat. Hist.), London. BUGE, E. 1946. Catalogue des Bryozoaires types et figures de la collection du Laboratoire de Paleontologie du Museum national d'Historie naturelle, pt. 4, Bryozoaires cheilostomes de rfiocene du Bassin de Paris figures par F. Canu (1907-1910). Bull. Mus. natn. Hist, nat., Paris (2) 18 : 428-439. - 1 953 . Le genre Schismoporella Gregory 1 893 (Bryozoa Cheilostomata) , caracteres generaux, composition specifique, et repartition. C. r. somm. Seanc. Soc. geol. Fr., Paris, 15 : 321-323. - 1957. Les Bryozoaires du Neogene de 1'ouest de la France et leur signification strati- graphique et paleobiologique. Mem. Mus. natn. Hist, nat., Paris (n.s.) 6 : 1-436, pis. 1-12. BURTON, E. ST. J. 1929. The horizons of Bryozoa (Polyzoa) in the Upper Eocene beds of Hampshire. Q. Jl Geol. Soc. Lond., 85 : 223-241. BUSK, G. i852a. In MACGILLIVRAY, J. Narrative of the voyage of H.M.S. " Rattlesnake " 1846-1 850. A n account of the Polyzoa and Sertularian Zoophytes, 1 : 343-385, pi. i . London. - 18526. Catalogue of marine Polyzoa in the collection of the British Museum, 1 : 1-54, pis. 1-68, 2 : 55-120, pis. 69-124. British Museum (Nat. Hist.), London, CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 109 BUSK, G. 1855. Zoophytology, Pt. i. Q. J. microsc. Sci., London, 3 : 253-256, pis. 3-4. - 1859. A monograph of the fossil Polyzoa of the Crag. 136 pp., 22 pis. Mon. Palaeont. Soc., London. - 1860. Zoophytology, Pt. 13. Q. J. microsc. Sci., London, 8 : 123-125, pis. 24-25. CALVET, L. 1906. Note preliminaire sur les Bryzoaires recueillis par les expeditions du " Travailleur " (1881-1882) et du " Talisman " (1883). Bull. Mus. Paris, 12 : 154-166. CANU, F. 1900. Revision des Bryozoaires du Cretace figures par d'Orbigny, II. Cheilostomata. Bull. Soc. geol. Fr., Paris (3) 28 : 334-463, pis. 4-7. -1902. Bryozoaires fossiles, I. Collection Compiche (Neocomien) ; II. Dutemple collection (Senonien superieur). Bull. Soc. geol. Fr., Paris (4) 2 : 10-14, 2 ngs- — igoja. Les Bryozoaires fossiles des terrains du sud-ouest de la France, I. Aquitanien. Bull. Soc. geol. Fr., Paris (4) 6 : 510-518, pis. 12-13. - 19076. Les Bryozoaires des terrains tertiaires des environs de Paris, I. Ann. Paleont., Paris, 2 : 57-88, pis. 1-4, 137-160, pis. 5-8. - 1908. Les Bryozoaires des terrains tertiaires des environs de Paris, II. Ann. Paleont., Paris, 3 : 65-104, pis. 9-10. - 1909. Les Bryozoaires fossiles des terrains du sud-ouest de la France, III. Burdigalien et IV. Helvetien. Bull. Soc. geol. Fr., Paris (4) 9 : 442-458, pis. 15-18. - 1910. Les Bryozoaires fossiles des terrains du sud-ouest de la France, V. Lutetien et VI. Bartonien. Bull. Soc. geol. Fr., Paris (4) 10 : 840-855, pis. 16-19. — 1912. Etude comparee des Bryozoaires helvetiens de 1'Egypte avec les Bryozoaires vivants de la Mediterrannee et de la Mer Rouge. Mem. Inst. £gypte, Paris, 6 : 185-236, pis. 10-13. - 1913. Contributions a 1'etude des Bryozoaires fossiles. Bull. Soc. geol. Fr., Paris (4) 13 : 124-131, i fig. — igi^a. Les Bryozoaires fossiles des terrains du sud-ouest de la France, VII. Lutetien. Bull. Soc. geol. Fr., Paris (4) 13 : 298-303, pis. 4, 5. - 1914^. Les Bryozoaires fossiles des terrains du sud-ouest de la France, VIII. Rupelien de Gaas. Bull. Soc. geol. Fr., Paris (4) 14 : 465-474, pis. 14, 15. - 1915. Bibliographie paleontologique relative aux Bryozoaires du Bassin de Paris. Bull. Soc. geol. Fr., Paris (4) 15 : 293-305. - 1917. Les Bryozoaires fossiles des terrains du sud-ouest de la France, X. Burdigalien. Bull. Soc. geol. Fr., Paris (4) 16 : 127-152, pis. 2, 3. Les Bryozoaires fossiles des terrains du sud-ouest de la France, XI. Rupelien. Bull. Soc. geol. Fr., Paris (4) 17 : 350-361, pis. 12, 13. - 1918^. Les Bryozoaires fossiles de la region des Corbieres. Bull. Soc. geol. Fr., Paris (4) 18 : 294-314, pis. 7-12. - 1920. Bryozoaires cretaces des Pyrenees. Bull. Soc. geol. Fr., Paris (4) 19: 186-211, pis. 4-6. • -- 1925. Les Bryozoaires du Lutetien de Saint-Germain-en-Laye. Bull. Soc. Sci. nat. med. Seine-et-Oise, Versailles (2) 6 : 46-48. - 1926. Bryozoaires bartoniens du Bassin Franco-Beige. Bull. Soc. geol. Fr., Paris (4) 25 : 741-761, pis. 26-30. - 1931. Curieuse biologic des Orbitulipores, Bryozoaires de 1'Oligocene. C. r. Congr. Socs sav. Paris. Sect. Sci., 1931 : 144-147, i fig. CANU, F. & BASSLER, R. S. 1917. A synopsis of American Early Tertiary Cheilostome Bryozoa. Bull. U.S. natn. Mus., Washington, 96 : 1-87, pis. 1-6. — 1919. In VAUGHAN, T. W. Ed. Geology and paleontology of the West Indies. Fossil Bryozoa from the West Indies. Carnegie Inst. Washington Publ, 291 : 73-102, pis. 1-7. - 1920. North American Early Tertiary Bryozoa. Bull. U.S. natn. Mus., Washington, 106 : i-x, 1-879, pis. 1-162. - 1923. North American Later Tertiary and Quaternary Bryozoa. Bull. U.S. natn. Mus., Washington, 125 : i-vii, 1-302, pis. 1-47. no CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX CANU, F. & BASSLER, R. S. 1927. Classification of the Cheilostomatous Bryozoa. Proc. U.S. natn. Mus., Washington, 69 (14) : 1-42, pi. i. - 1928. Fossil and Recent Bryozoa of the Gulf of Mexico region. Proc. U.S. natn. Mus., Washington, 72 (14) : 1-199, pis. i~34- I929«. Bryozoaires eocenes de la Belgique. Mem. Mus. r. Hist. nat. Belg., Bruxelles, 39 : 1-69, pis. 1-5. - 19296. Contributions to the biology of the Philippine Archipelago and adjacent regions. Bryozoa of the Philippine region. Bull. U.S. natn. Mus., Washington, 100 (9) : i-xi, 1-685, pis. 1-94. - 1930. Bryozoaires marins de Tunisie. Ann. Stn oceanogr. Salammbo, Tunis, 1 : 1-91, pis. 1-13. - 1931. Bryozoaires oligocenes de la Belgique. Mem. Mus. r. Hist. nat. Belg., Bruxelles, 50 : 1-24, pis. 1-4. CANU, F. & CAILLOT, F. 1932. Les Adeones d'fizanville (Seine-et-Oise) . Bull. Soc. Sci. nat. med. Seine-et-Oise, Versailles (2) 13 : 3-15, pis. i, 2. CANU, F. and LECOINTRE, G. 1927. Les Bryozoaires cheilostomes des faluns de Touraine et d'Anjou. Mem. Soc. geol. Fr., Paris (n.s.) 4 : 1-18, pis. 6-10. CHEETHAM, A. H. 1957. Eocene-Oligocene boundary, eastern Gulf Coast region. Trans. Gulf-Cst Ass. geol. Soc., New Orleans, 7 : 89-97, 5 n§s- — 1962. Eocene Bryozoa from the McBean Formation in Georgia. Micropaleont., New York, 8 : 323-336, pis. 1,2. - i963«. The Polyzoan genus Ditaxiporina Stach. Ann. Mag. nat. Hist., London (13) 5 : 485-490. i fig. — 19636. Late Eocene zoogeography of the eastern Gulf Coast region. Mem. geol. Soc. Amer., New York, 91. xii +113 pp. 3 pis. CHEETHAM, A. H. & SANDBERG, P. A. 1964. Quaternary Bryozoa from Louisiana mudlumps. /. Paleont., Menasha, 38 : 1013-1046, 59 figs. COOK, P. L. 1964. Polyzoa from West Africa, I. Notes on the Steganoporellidae, Thala- moporellidae and Onychocellidae (Anasca, Coilostega). Res. sci. Campagnes " Calypso " , Paris, 6 : 43-78, pi. i. CURRY, D. 19580. In CURRY, D. & WISDEN, D. E. Geology of the Southampton area. Bracklesham Bay and the Selsey Peninsula, Sussex. Geol. Ass. Guide, London, 14 : 13-16, i fig- — 19586. In WHITTARD, F. & SIMPSON, S. Ed. England, Wales, and Scotland. Palaeo- gene. Lexique Stratigraphique International, Paris, 1 (3a) : 1-82. — 1962. A Lower Tertiary outlier in the central English Channel with notes on the beds surrounding it. Q. J. Geol. Soc. Lond., 118 : 177-205, 2 figs. — 1962 MS. Le Paleogene de 1'Angleterre. Colloque Paleogene, Bordeaux, 1962 (circular). DARTEVELLE, E. 1933. Contribution a 1'etude des Bryozoaires fossiles de 1'Eocene de la Bel- gique. Ann. Soc. r. zool. Belg., Bruxelles, 63 : 55-116, pis. 2-4. - 1935. Bryozoaires du Lutetien de Gomerfontaine. Ann. Soc. r. zool. Belg., Bruxelles, 65 : 111-126, 2 figs. - 1936. Bryozoaires du Ledien du Bassin de Paris. Ann. Soc. r. zool. Belg., Bruxelles, 66 : 21-32. - 1937. Bryozoaires et Brachiopode du Bruxellien de Nil-Saint-Vincent. Ann. Soc. r. zool. Belg., Bruxelles, 67 : 108-112. — 1942. Notes pur servir a 1'etude des Bryozoaires fossiles, II. Un Bryozoaire nouveau du Ledien du Bassin beige, Heterocella lediensis nov. sp. Bull. Soc. beige Geol. Paleont. Hydrol., Bruxelles, 50 : 148-151, pi. 3, figs, i, 2. - 1952. Bryozoaires fossiles de 1'Oligocene de I'Allemagne. Palaont. Z., Berlin, 26 : 181- 204, 2 figs. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX in DAVID, L. & MONGEREAU, N. 1961. Un example d'etude statistique en paleontologie : Cellaria fistulosa Auct. (Bryozoa, Cheilostomata) du Vindobonien de la region lyonnaise. Bull. Bur. Recherches geol. et Min., Paris, 3 : 29-41. DAVIS, A. G. 1928. The geology of the City and South London Railway Clapham-Morden extension. Proc. Geol. Ass. Lond., 39 : 339-352. -192913. In VENABLES, E. M. On the London Clay of the Bognor district. Proc. Geol. Ass. Lond., 40 : 50. - 19296. tiber die Fauna eines Eozangescheibes von Cothen (Anhalt) : Z. Gescheibeforsch., Berlin, 5 : 111-113. — 1934. English Lutetian Polyzoa (Eocene). Proc. Geol. Ass. Lond., 45 : 205-244, pis. The London Clay at Bracknell, Berks. Proc. Geol. Ass. Lond., 47 : 140-144. — 19366. The London Clay of Sheppey and the location of its fossils. Proc. Geol. Ass. Lond., 47 : 328-345, i fig. — 1940. In WRIGLEY, A. The fauna! succession in the London Clay, illustrated in some new exposures near London. Proc. Geol. Ass. Lond., 51 : 236. — 1962. In CURRY, D. A Lower Tertiary outlier in the central English Channel with notes on the bed surrounding it. Q. J. Geol. Soc. Lond., 118 : 194. DAVIS, A. G. & ELLIOTT, G. F. 1957. The palaeogeography of the London Clay sea. Proc. Geol. Ass. Lond., 68 : 255-277, 2 figs. DEFRANCE, J. P. M. 1828. In Diet. Sci. nat., Paris, 57 : 1-546. — 1829. In Diet. Sci. nat., Paris, 58 : 214. DENIZOT, G. 1957- In DENIZOT, G. Ed. France, Belgique, Pays-Bas, Luxembourg. Tertiaire. Lexique Stratigraphique International, Paris, 1 (4a) : 1-217. DIXON, F. 1850. The geology and fossils of the Tertiary and Cretaceous formations of Sussex. 422 pp., 40 pis. London. DUVERGIER, J. 1921. Note sur les Bryozoaires du Neogene de 1'Aquitaine. Act. Soc. linn. Bordeaux, 72 : 145-181, pis. 1-4. — 1924. Deuxieme note sur les Bryozoaires du Neogene de 1'Aquitaine. Act. Soc. linn. Bordeaux, 75 : 145-190, pis. 1-6. FABRICIUS, O. 1824. Nye zoologiske bidrage. K. danske Vidensk. Selsk. Phys. Skr., Copen- hagen, 1 : 25-80, pis. i, 2. FAURA Y SANS, M. & CANU, F. 1916. Sur les Bryozoaires des terrains tertiaries de la Catalogne. Inst. Cat. Hist. Nat., Barcelona, 1916 : 59-193, pis. 1-9. FISHER, O. 1862. On the Bracklesham Beds of the Isle of Wight Basin. Q. J. Geol. Soc. Lond., 18 : 65-93, 4 figs. FRANKE, F. 1939. Die Bryozoen der Unteroligocans von Magdeburg. Abh. Mus. Nat. Vor- gesch. Magdeburg, 7 : 59-67, pis. 2, 3. GHIURCA, V. 1962. Contributions a la connaissance de la faune de Bryozoaires de Transyl- vanie, III. Revision taxonomique des Bryozoaires de 1'Eocene superieur du N.O. de la Transylvanie publics jusqu'a present. Studio, Univ. Babes-Bolyai (Geol.-Geogr.) Cluj, 1962 : 71-74 (in Rumanian with French and Russian summaries). GORODISKI, A. & BALAVOINE, P. 1962. Bryozoaires cretaces et eocenes du Sendgal. Bull. Bur, Recherches geol. et Min., Paris, 4 : i-io. GREGORY, J. W. 1893. On the British Palaeogene Bryozoa. Trans, zool. Soc. Lond., 13 : 219-279, pis. 29-32. HAGENOW, K. F. VON. 1851. Die Bryozoen der Maastrichter Kreidebildung. xvi+m pp., 22 pis. Cassel. HARMER, S. F. 1926. The Polyzoa of the Siboga Expedition, 2. Cheilostomata Anasca. Siboga Exped. Rep., Leiden, 28B : i-viii, 181-501, pis. 13-34. - 1934- The Polyzoa of the Siboga Expedition, 3. Cheilostomata Ascophora, I. Family Reteporidae. Siboga Exped. Rep., Leiden, 28G : i-vii, 503-640, pis. 35-41. 112 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX HARMER, S. F. 1957- The Polyzoa of the Siboga Expedition, 4. Cheilostomata Ascophora, II. Siboga Exped. Rep., Leiden, 28D : i-xv, 641-1145, pis. 42-74. HASSALL, A. H. 1842. Remarks on the genus Lepralia of Dr. Johnston with descriptions of six undescribed species, and notices of two other zoophytes. Ann. Mag. nat. Hist., London (i) 9: 407-414- HASTINGS, A. B. 1932. The Polyzoa with a note on an associated hydroid. Rep. Gt Barr. Reef Exped., 4, 12 : 399-458, pi. i. HASWELL, W. A. 1881. On some Polyzoa from the Queensland coast. Proc. Linn. Soc. N.S. W., Sydney, 5 : 33-44- pis. 1-3. HINCKS, T. 1877. On the British Polyzoa. Ann. Mag. nat. Hist., London, (4) 20 : 212-218, 520-532- — 1887. On the Polyzoa and Hydroida of the Mergui Archipelago collected for the Trustees of the Indian Museum by Dr. J. Anderson. J. Linn. Soc. Lond. (Zool.) 21 : 121-135, pi- I2- ILLIES, G. 1953- Variationsstatistische Untersuchungen an Rhiniopora cacus (Brydone) (Bryozoa Cheilostomata) aus der Oberkreide von Hemmoor Niederelbe. Mitt, miner .-geol . St Inst. Hamb., 22 : 76-101, pis. 15, 16. JOHNSTON, G. 1838. A history of British zoophytes. xii + 34i pp., 44 pis. Edinburgh. JULLIEN, J. 1882. Dragages du " Travailleur ", Bryozoaires, especes draguees dans 1'Ocean atlantique en 1881. Bull. Soc. Zool. Fr., Paris, 7 : 497-529, pis. 13-17. JULLIEN, J. & CALVET, L. 1903. Bryozoaires provenant des campagnes de " 1'Hirondelle". Result. Camp, scient. Prince Albert I., 23 : 1-188, pis. 1-18. KLUGE, G. A. 1962. Bryozoa of the northern seas of the U.S.S.R. 584 pp., 404 figs. Acad. Sci. U.S.S.R., Pub. Zool. Inst., 76 (in Russian). KOSCHINSKY, C. 1885. Ein Beitrage zur Kentniss der Bryozoenfauna der alteren Tertiar- schichten des siidlichen Bayerns, I. Cheilostomata. Palaeontographica, Stuttgart, 32 : 1-73, pis. 1-7. KYRI, M. M. 1951. La revision des Bryozoaires cheilostomates eocenes contenus dans le materiel collectionn6 par Heljjas. Studii Cere, stiint., Cluj, 2 : 68-83. (in Rumanian with French summary). LAGAAIJ, R. 1952. The Pliocene Bryozoa of the Low Countries and their bearing on the marine stratigraphy of the North Sea region. Meded. geol. Sticht., 's-Gravenhage (n.s.) 5 : 1-233, pis. 1-26. - 1959. Some species of Bryozoa new to the Bowden Beds, Jamaica, B.W.I. Micropaleont., New York, 5 : 482-486, 5 figs. 1963. New additions to the Bryozoan fauna of the Gulf of Mexico. Publs. Inst. mar. Sci. Univ. Tex., 9 : 162-236, pis. 1-8. LAMARCK, J. B. P. A. DE. 1816. Les Polypes. Histoire naturelle des Animaux sans Vertebres, II. 586 pp. Paris. LARWOOD, G. P. 1962. The morphology and systematics of some Cretaceous Cribrimorph Polyzoa (Pelmatoporinae). Bull. Brit.Mus. (Nat. Hist.}, Geol., London, 6 : 1-285, pis- !-23. LEVINSEN, G. M. R. 1902. Studies on Bryozoa. Vidensk. Meddr. dans naturh. Foren. Kobenhavn, 1902 : 1-31. — 1909. Morphologic and systematic studies on the Cheilostomatous Bryozoa. vii + 43i pp., 27 pis. Copenhagen. LONSDALE, W. 1850. In DIXON, F. The geology and fossils of the Tertiary and Cretaceous formations of Sussex : 159-161, 268-322, pis. i, i8A, B. London. — 1878. In DIXON, F. The geology and fossils of the Tertiary and Cretaceous formations of Sussex. 2nd edit. : 176-207, 289-360, pis. i, 18, i8A, B. London. MAcGiLLiVRAY, P. H. i88o. In McCov, F. Prodromus of the zoology of Victoria, 5 : 27-52, pis. 45-49- — 1886. Descriptions of new or little known Polyzoa. Trans. Proc. R. Soc. Victoria, 22 : 128-139, pis. 1-3. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 113 MAcGiLLiVRAY, P. H. 1895. Monograph of the Tertiary Polyzoa of Victoria. Trans. Proc. R. Soc. Victoria (n.s.) 4 : 1-166, pis. 1-22. MALECKI, J. 1963. Bryozoa from the Eocene of the central Carpathians between Grybow and Dukla. Prace Geol., Pol. Akad. Nauk, Oddzial w Krakowie. Kom. Nauk Geol., Warsaw, 16 : 1-158 (in Polish with English and Russian summaries). MARCUS, E. 1920. Bryozoa from the Swedish scientific expeditions to Australia during 1910- 1913. K. svenska Vetensk Akad. Handl., Stockholm, 61 : (5), 1-34, pis. i, 2. — 1921. Indo-Pacifische Bryozoen aus dem Riksmuseum in Stockholm. Ark. Zoo/., Stockholm, 14, 7 : 1-23. - 1922. Bryozoen von den Aru-Inseln. Abh. Senckenb. Naturf. Ges., Frankfurt a. M., 35: 421-446, pis. 24, 25. MAYR, E., LINSLEY, E. G. & USINGER, R. L. 1953. Methods and principles of systematic zoology. 328 pp. New York. McGuiRT, J. H. 1941. Louisiana Tertiary Bryozoa. Louisiana Dept. Conserv. Geol. Bull., Baton Rouge, 21 : 1-177, pis. 1-31. MEUNIER, A. & PERGENS, E. 1886. Bryozoaires du Systeme montien. 14 pp., 3 pis. Brussels. MICHELIN, J. L. H. 1840-1847. Iconographie zoophytologique. Description par localites et terrains des Polypiers fossiles de France et pays environnants . 348 pp., 79 pis. Paris. MILNE EDWARDS, H. 1836. In LAMARCK, J. B. P. A. DE. Histoire naturelle des Animaux sans Vertebres. II. Histoire des Polypes. 2nd edit. 683 pp. Paris. MORREN, C. F. A. 1828. Descriptio coralliorum fossilium in Belgio repertorum. Ann. Acad. Groningen, 1828 : 1-76, pis. 1-22. MORRIS, J. 1854. A catalogue of British fossils, x + 222 pp. London. MOURLON, M. F. 1881. Geologic de la Belgique, II. xvi + 392 pp., 55 figs. Bruxelles. NEVIANI, A. 1898. Briozoi Neozoici di alcune localita d'ltalia. Boll. Soc. Romana Studi ZooL, 7 : 34-49, 97-I09, 5 ngs. NORMAN, A. M. 1903. Notes on the natural history of East Finmark — Polyzoa. Ann. Mag. nat. Hist., London (7) 11 : 567-598, pi. 13. ORBIGNY, A. D.'. 1849. Prodrome de paleontologie stratigraphique universelle, I. 394 pp. Paris. 1851-1854. Paleontologie francaise. Terrains cretaces. V. Bryozaires. 1192 pp., pis. 600-800. Paris. OSBURN, R. C. 1914. The Bryozoa of the Tortugas Islands, Florida. Pap. Tortugas Lab. Carnegie Inst. Washington, 182 : 183-222, 23 figs. — 1927. The Bryozoa of Curacao. Bijdr. Dierk., Amsterdam, 25 : 123-132, 7 figs. 1940. Bryozoa of Porto Rico with a resum6 of the West Indian Bryozoan fauna. Sci. Surv. P. Rico &• Virgin Is., New York, 16 : 321-486, pis. 1-8. 1947- Bryozoa of the Allan Hancock Expedition, 1939. Rep. Allan Hancock Atlantic Exped., Los Angeles, 5 : 1-66, pis. 1-6. 1950. Bryozoa of the Pacific coast of America, I. Cheilostomata-Anasca. Publ. Allan Hancock Pacific Exped., Los Angeles, 14 : 1-269, pis. 1-29. 1952. Bryozoa of the Pacific coast of America, II. Cheilostomata-Ascophora. Publ. Allan Hancock Pacific Exped., Los Angeles, 14 : 271-611, pis. 30-64. PERGENS, E. 1887. Note pr61iminaire sur les Bryozoaires fossiles des environs de Kolosvar. Bull. Seanc. Soc. r. malac. Belg., Bruxelles, 22 : 3-7. 1889. Zur fossilen Bryozoenfauna von Wola Lu'zanska. Bull. Soc. belg. Geol. Paleont. Hydrol., Bruxelles, 3 : 59-72. POKORNY, V. 1944. Nellia ocullata (sic) Busk, eine neue Bryozoenart fur das inneralpine Wiener Becken. Vest. csl. Spol. zool. Nauk, Prague, 11 : 1-14, 2 figs. REUSS, A. E. 1848. Die fossilen Polyparien des Wiener Tertiarbeckens. Haidingers Natur- wiss. AbhandL, Vienna, 2 : 1-109, pis. i-u. 1865. Zur fauna des deutschen Oberoligocans, II. Anthozoen und III. Bryozoen. Sitzb. K. Akad. Wiss. Wien, 50 : 614-691, pis. 6-15. ii4 CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX REUSS, A. E. 1866. Die Foraminiferen, Anthozoen und Bryozoen des deutschen Sept- arienthones. Denkschr. K. Akad. Wiss. Wien, 25 : 117-214, pis. i-n. — 1867. Uber einige Bryozoen aus dem deutschen Unteroligocans. Sitzb. K. Akad. Wiss. Wien, 55 : 216-234, pis. 1-3. — 18690. Palaontologische Studien iiber die alteren Tertiarschichten der Alpen, II. Die fossilen Anthozoen und Bryozoen der Schichtengruppe von Crosara. Denkschr. K. Akad. Wiss. Wien, 29 : 215-298, pis. 17-36. — 18696. Zur fossilen Fauna der Oligocanschichten von Gaas. Sitzb. K. Akad. Wiss. Wien, 59 : 446-488, pis. 1-6. - 1874. Die fossilen Bryozoen des osterreichisch-ungarischen Miocans. Denkschr. K. Akad. Wiss. Wien., 33 : 141-190, pis. 1-12. REYMENT, R. A. 1963. Notes on the description of post-Paleozoic fossil Ostracods. J. Paleont., Menasha, 37 : 682-687. ROBERTSON, A. 1921. Report on a collection of Bryozoa from the Bay of Bengal and other eastern seas. Rec. Indian Mus., Calcutta, 22 : 33-65, n figs. SILEN, L. 1941. Cheilostomata Anasca (Bryozoa) collected by Prof. Dr. Sixten Bock's expedition to Japan and the Bonin Islands 1914. Ark. Zool., Stockholm, 33A, 12 : 1-130, pis. 1-9. - 1944. The main features of the development of the ovum, embryo, and ooecium in the oeciferous Bryozoa Gymnolaemata. Ark. Zool., Stockholm, 35A, 17 : 1-34, 23 figs. — 1947. Conescharellinidae (Bryozoa Gymnolaemata) collected by Prof. Dr. Sixten Bock's expedition to Japan and the Bonin Islands 1914. Ark. Zool., Stockholm, 39A, 9 : 1-61, pis. 1-5. SIMPSON, G. G., ROE, A. & LEWONTIN, R. C. 1960. Quantitative zoology. Rev. edit. vii + 440 pp., 64 figs. New York. SMITT, A. F. 1873. Floridan Bryozoa collected by Count L. F. de Pourtales. K. svenska Vet- ensk. Akad. Handl., Stockholm, 11 : 3-83, pis. 1-13. STACK, L. W. 1936. Correlation of zoarial form with habitat. /. Geol., Chicago, 44 : 60-65, i fig. - 1937. The application of the Bryozoa in Cainozoic stratigraphy. Rep. Aust. N.Z. Ass. Advmt. Sci. (Auckland], 23 : 80-83. STOLICZKA, F. 1862. Oligocane Bryozoen von Latdorf in Bernburg. S.B. preuss. Akad. Wiss., Berlin, 45 : 71-94, pis. 1-3. SYLVESTER-BRADLEY, P. C. 1958. The description of fossil populations. /. Palaeont., Menasha, 32 : 214-235, 16 figs. THOMAS, H. D. & DAVIS, A. G. 1949. The Pterobranch Rhabdopleura in the English Eocene. Bull. Brit. Mus. (Nat. Hist.) Geol., London, 1 : 1-19, pis. 1-3. THORNELY, L. R. 1905. In HERDMAN, W. A. Report on the pearl oyster fisheries of the Gulf of Manaar. Report on the Polyzoa, supplementary report, 26 : 107-130., pi. i. London. - 1907. Report on the marine Polyzoa in the Indian Museum. Rec. Indian Mus., Calcutta, 1 : 179-196, 8 figs. VIGNEAUX, M. 1949. Revision des Bryozoaires n6ogenes du Bassin d'Aquitaine et essai de classification. Mem. Soc. geol. Fr., Paris (n.s.) 28 : 1-155, pis- i-u- VINE, G. R. 1889. Notes on British Eocene Polyzoa. Proc. Yorks. geol. polytech. Soc., 11 : 154-169, pi. 5. VOIGT, E. 1951. Das Maastricht- Vorkommen von Ilten bein Hannover und seine Fauna mit besonderer Beriicksichtigung der Gross-Foraminiferen und Bryozoen. Mitt, miner .-geol . St Inst. Hamb., 20 : 15-109, pis. i-io. WATERS, A. W. 1883. Fossil Chilostomatous Bryozoa from Muddy Creek, Victoria. Q. J. Geol. Soc. Lond., 39 : 423-443, pi. 12. 1887. Bryozoa from New South Wales, North Australia, etc. Ann. Mag. nat. Hist., London (5) 20 : 81-95, I8i-2O3, 253-265, pis. 4-7. CHEILOSTOMATOUS POLYZOA FROM THE EOCENE OF SUSSEX 115 WATERS, A. W. 1891. North Italian Bryozoa, I. Chilostomata. Q. J. Geol. Soc. Lond., 47 : i-34. Pls- x-4- 1905. Notes on some Recent Bryozoa in d'Orbigny's collection. Ann. Mag. nat. Hist., London (7) 15 : 1-16, pi. i. 1909. Reports on the marine biology of the Sudanese Red Sea, XII. The Bryozoa : /. Linn. Soc. Lond. (Zool.) 31 : 123-181, pis. 10-18. 1913. The marine fauna of British East Africa and Zanzibar from collections made by Cyril Crossland in the years 1901-1902, Bryozoa-Cheilostomata. Proc. zool. Soc. Lond., 1913 1458-537, pis. 64-73. 1919- Batopora (Bryozoa) and its allies. Ann. Mag. nat. Hist., London (9) 3 : 79-94, pi. 6. WOOD, S. V. 1844. Descriptive catalogue of the zoophytes from the Crag. Ann. Mag. nat. Hist., London (i) 13 : 10-21. WRIGLEY, A. 1934. A Lutetian fauna at Southampton docks. Proc. Geol. Ass., Lond., 45 : 1-16, i fig. WRIGLEY, A. & DAVIS, A. G. 1937. The occurrence of Nummulites planulatus in England, with a revised correlation of the strata containing it. Proc. Geol. Ass. Lond., 48 : 203-228, pis. 17, 18. PRINTED IN GREAT BRITAIN BY ADLARD & SON LIMITED BARTHOLOMEW PRESS, DORKING \' FOSSIL MAMMALS OF AFRICA No. 21 MIOCENE RHINOCEROSES OF EAST AFRICA D. A. HOOIJER BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 13 No. 2 LONDON: 1966 FOSSIL MAMMALS OF AFRICA No. 21 ^11 AUG1% MIOCENE RHINOCEROSES OF EAST AFRICA BY DIRK ALBERT HOOIfER D.Sc. Rijksmuseum van Natuurlijke Historic, Leiden Pp. 117-190 ; 15 Plates BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 13 No. 2 LONDON: 1966 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 ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 13, No. 2 of the Geological (Palaeon- tological) series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Period- icals. Trustees of the British Museum (Natural History) 1966 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 3 August, 1966 Price £3 35. FOSSIL MAMMALS OF AFRICA No. 21 MIOCENE RHINOCEROSES OF EAST AFRICA By D. A. HOOIJER A en juger d'apres 1'assurance avec laquelle certains auteurs ont attribue des noms specifiques aux restes les plus insignifiants de Rhinoceros fossiles, on pourrait croire que la determination des animaux de ce groupe est chose aisee. Ceux qui se sont serieusement occupes de ce sujet savent que le contraire est vrai. H. G. STEHLIN CONTENTS Page I. INTRODUCTION AND ACKNOWLEDGEMENTS . . . . . .119 II. SYSTEMATIC DESCRIPTIONS ........ 122 Dicerorhinus leakeyi sp. nov. ........ 122 Aceratherium acutirostratum (Deraniyagala) . . . 136 Brachypotherium heinzelini Hooijer . . . . . . .142 Chilotherium sp. . . . . . . . . . .150 POSTCRANIAL SKELETON OF Dicerorhinus AND Aceratherium . .152 III. DISTRIBUTION OF RHINOCEROSES OVER EAST AFRICAN MIOCENE SITES . 183 IV. TIME PLACEMENT OF THE MlOCENE EAST AFRICAN FAUNAS . . . 185 V. REFERENCES .......... 187 SYNOPSIS Four species of Rhinocerotidae are described from the Tertiary (Miocene) of East Africa, including a new species Dicerorhinus leakeyi. The bearing of these Rhinoceroses on the time placement of the Miocene East African faunas is discussed, resulting in a tentative correlation with the Burdigalian of Europe, although some of the Rusinga sites appear to be younger, later Miocene or even Pliocene. I. INTRODUCTION AND ACKNOWLEDGMENTS RHINOCEROSES have a reputation for being difficult animals to deal with as fossils. In spite of an enormous amount of scientific literature, the present state of our knowledge and comprehension of this group is comparable only to that of O. C. Marsh's grasp of the equids.1 H. F. Osborn's oft-cited " Phylogeny of the rhinoceroses of Europe " (1900) has never been followed up by a monographic treatise, and the first part of his " The extinct rhinoceroses " (1898), the only part ever published, is devoted to generalities and the acerathere rhinoceroses of the White River Beds of Nebraska and the Dakotas. A comprehensive paper on the Tertiary Rhinocerotidae of Eurasia is sadly lacking and treatment of the fossil material in the scattered literature is very uneven and incomplete. For a sound diagnosis of a new fossil species we need the whole skull, with the 1 This sagacious comment was made by Dr. Stanley Westoll in the discussion following my paper at the Symposium of Vertebrate Palaeontology and Comparative Anatomy in Bristol on 23rd September, 1964. GEOL. 13, 2. 8 120 MIOCENE RHINOCEROSES OF EAST AFRICA incisors and canines, if any (premaxillaries have often been lost) . We need also the skeleton, in particular the metapodials, which are rarely found associated with a skull. These requirements, therefore, are seldom fulfilled (cf. my motto, taken from Stehlin 1925 : 106). Too much reliance has often been placed upon individually variable molar crown structures or labile cingula. New fossil species (and genera) have occasionally been based on milk teeth mistaken for permanent teeth, or on female specimens of previously described forms. Fragmentary fossil remains of rhinoceroses have more than once been described as hippopotami, and vice versa. Foot bones have been confounded with those of anthracotheres or chalicotheres. Wrong identifications, once published, have a habit of perpetuating themselves in the literature, to the detriment of a better understanding. In order to soften these pontifical remarks I hasten to add that I have not been able always to avoid these pitfalls in my own rhinoceros work either. In the last few decades representatives of three genera of Rhinocerotidae from the Tertiary of Eurasia have been found in Africa. These are Dicerorhinus , Aceratherium and Brachypotherium. The so-called phyletic line of Dicerorhinus Gloger comprises a number of evidently collateral forms with slender limbs and feet ranging from the Aquitanian (Upper Oligocene) through the Pleistocene of Europe, and up into the Holocene of Asia. It seems unlikely that the fossils should all be referred to the same genus, and that this is the genus of the extant Sumatran species. D. sumatrensis (Fischer) is the most primitive among the five surviving species of rhinoceroses, and may truly be said to represent a Miocene stage of evolution of teeth and skeleton, but its immediate ancestry is unknown apart from what can be derived from subfossil remains found in Sumatran caves (Hooijer ig^6a, b). This suggests a decrease in tooth and limb size since the formation of the cave deposits (presumably Early Holocene), which is a common phenomenon. The further use of the generic name Dicerorhinus for the Tertiary and Pleistocene forms, however, is to be recommended ; we have far too many generic names in the Rhinocerotidae anyway (many monotypical) , and it is a relief to see a case in which the generic limits are drawn as broadly as in Dicerorhinus. The first African representative to become known is the well-documented Dicero- rhinus primaevus Arambourg (1959) from the Pontian (Lower Pliocene) of Wad el Hammam in Algeria ; earlier records are ambiguous. The genus Aceratherium Kaup, with its persistently tetradactyl fore feet, and limbs as slender as in Dicerorhinus, ranges from the Stampian (Middle Oligocene) up into the Pontian in Europe and Asia, showing some phylogenetic advance. It was first recorded from East Africa (Moruaret Hill near Losodok or Lothidok, Kenya) by Deraniyagala (1951) as Turkanatherium Deraniyagala ; Arambourg's earlier record of a lower molar, an epistropheus, an astragalus, two metatarsals and some phalanges from Losodok as Aceratherium ? spec, might also belong to Dicerorhinus. The dentition of Aceratherium acutirostratum (Deraniyagala) has recently been described from the Miocene of the Karugamania region, Lake Albert, Western Rift Valley in Congo (Hooijer 1963). In the short-limbed and -footed genus Brachypotherium Roger of Europe (Burdi- MIOCENE RHINOCEROSES OF EAST AFRICA 121 galian (Lower Miocene) through Pontian), often placed in the North American genus Teleoceras Hatcher that may have been derived from it (Osborn 1910 : 292), there is some evolutionary progress (in the shortening of the limbs and metapodials par- ticularly). In 1920 Fourtau recorded this genus (as Teleoceras) from the Burdigalian of Moghara in Egypt, and I have added a more progressive species from the Miocene of the Sinda-Mohari region, Lower Semliki, Congo (Hooijer 1963). Thus, there are various previous records of Tertiary Rhinocerotidae from Africa (in this connexion mention should be made of the well-described but specifically unidentified last upper molar from Karungu, Kenya, recorded by Andrews 1914). Sonia Cole (1950) listed Aceratherium from the Miocene of Rusinga Island, Karungu, and Maboko Island, as well as Teleoceras from Rusingu and Karungu (Cole 1950 : 29), and also published a photograph of a rhinoceros skeleton in the process of being excavated from the Lower Hiwegi Beds in Rusinga in 1947 (1950, pi. i). In a provisional list of the Miocene faunas of East Africa, Le Gros Clark & Leakey (1951 : 5) recorded Rhinocerotidae from the following nine sites : Karungu, Rusinga Island, Chianda Uyoma, Ombo, Maboko Island, Songhor, Losodok, Loperot and Tambach. The collections described in the present paper, upon which the above cited locality records are based, are for the most part in the National Museum Centre for Prehistory and Palaeontology, Nairobi, Kenya, and were generously offered to me for study and report by Dr. L. S. B. Leakey in April, 1963. Much Miocene East African rhinoceros material is in the Department of Palaeontology of the British Museum (Natural History), London, and this I have been lent. Dr. W. W. Bishop of the Kampala Museum, Uganda, has sent me Miocene material from the Napak volcanics, Kara- moja, Uganda, which is likewise described in the present paper. I am very much indebted to Dr. L. S. B. Leakey for entrusting this interesting material to me, as well as to Dr. W. W. Bishop for the Napak material and to Dr. A. J. Sutcliffe for arranging to have the British Museum material made available. I am very grateful to Mrs. S. C. Coryndon, Mrs. Sonia Cole and Dr. T. Whitworth for valuable information and kind advice. A systematic account of the genera and species of Rhinocerotidae from the East African Miocene is given in the following chapters. The specimens from the British Museum (Natural History) have numbers preceded by an M. The conventional dental nomenclature has been used and the measurements of the cheek teeth have been taken at the base of the crown, the length (ant. post.) in the upper P and M externally except in M3, where it is taken internally. Most of the specimens, in addition to the catalogue number, bear letters indicating the sites, such as R. for Rusinga Island, Rs. for a surface find in Rusinga, followed by a sub-site number, e.g., R.I, R.2, etc. (Le Gros Clark & Leakey 1951 : 10) ; K stands for Karungu, KB or MB for Maboko (= Kiboko) Island, and S or Sgr for Songhor (Whitworth 1958 : 2). Maps showing the location of the various sites in Kenya and Uganda will be found in Whitworth (1958 : 2) and Bishop (1958), a map showing the sub-sites in Rusinga was given by Le Gros Clark & Leakey (1951 : 9). On the advice of Dr. Leakey the relationship of the sites to the fossiliferous strata in Rusinga Island has been omitted at this stage. 122 MIOCENE RHINOCEROSES OF EAST AFRICA II. SYSTEMATIC DESCRIPTIONS Genus DICERORHINUS Gloger 1841 Dicerorhinus leakeyi sp. nov. (Pis. 1-3 ; PI. 4, figs, i, 4, 7, 8 ; PI. 5, figs. 1-3 ; PL 6, figs. 7, 8, 12 ; PI. 7, figs. 5, 6 ; PI. 10, figs. 4, 5 ; PL ii ; PL 13, figs. 1-5 ; PL 14, fig. i ; PL 15) DIAGNOSIS. Dicerorhinus species with frontal and nasal horns ; upper incisors smallish, small incisors between lower canines ; inferior squamosal processes united below subaural channel. Occiput as highly elevated as in D. sansaniensis (Lartet). Lower border of mandible nearly straight as in D. sansaniensis, not curved upward in symphysial region as in D. schleiermacheri (Kaup), D. orientalis (Schlosser) and D. ringstroemi Arambourg. Size of skull as in D. schleiermacheri and D. orientalis, larger than in D. sansaniensis and smaller than in D. ringstroemi. Teeth inter- mediate in size between those of D. schleiermacheri and those of D. sansaniensis. Upper premolars with protoloph and metaloph united internally up to at least 15 mm. from crown base, cingulum weak, protocone not markedly constricted off. Upper molars with low and wide lingual entrance to medisinus, internal cingulum very weak or absent, protocone not or hardly constricted off, antecrochet not prominent, not blocking medisinus, ectoloph depressed between the roots, crochet and crista weak or absent, M3 bulging out at junction of ectoloph and metaloph. The specific name is given in honour of Dr. L. S. B. Leakey, who collected the type in 1-935 • HOLOTYPE. The skull and associated mandible from Rusinga (PL i ; PL 2, figs. 1,2). HORIZON AND LOCALITY. Lower Miocene ; Rusinga Island, Kenya. DESCRIPTION. The skull and associated mandible were collected at Rusinga by Dr. L. S. B. Leakey in 1935. Both are somewhat crushed and in part restored or fortified with plaster. The crushing is mostly laterally : the two upper tooth-rows are only 30 mm. apart in the premolar region and 50 mm. between the last molars. The tooth-rows are somewhat displaced longitudinally : the right tooth-row is shifted 2 cm. backward relative to the left. The palate is broken. The whole of the post- dental basal portion of the skull is lost except for the right zygomatic arch and glenoid cavity with the postglenoid and post-tympanic processes. The zygomatic arch is pushed inward and slightly backward, reducing the width of the temporal fossa to a mere 3 cm. Of the glenoid cavity the outer portion is displaced backward ; the outer angle of this cavity is on the same transverse level as the huge postglenoid process. This distortion evidently took place while the condyle of the mandible was lodged in the cavity, for the condyle is deformed in the same way, with its outer part pushed backward, and it fits exactly into the cavity as it is. Only a small portion of the left parietal is preserved, and it is thrust upward. When viewed from the right side, however, the fronto-parietal surface is relatively well preserved, rising backward and upward from the orbit in a gentle curve which MIOCENE RHINOCEROSES OF EAST AFRICA 123 seems unaffected by distortion. The top of the occiput has broken off. The right fronto-parietal crest is preserved, and evidently did not meet its fellow on the other side of the skull, the least distance between the crests being approximately 3 cm. The temporal crest is for the most part preserved, but undamaged only in its lower portion. Below the external auditory meatus the two inferior squamosal processes are seen to be firmly united. On the ventral surface of the skull nothing remains behind the last molars but the base of the pterygoid process on the left side. The hinder margin of the palate is on a level with the front of M2. The frontal region of the skull is only superficially damaged, and it is clear that there is a median boss just above the anterior border of the orbit, indicating the presence of a frontal horn. The postorbital processes of the frontals are damaged, and the width of the skull at this point cannot be determined. The anterior border of the orbit is above the anterior border of M2. The nasal bones are well preserved on both sides, although laterally compressed, and are convex and rugose above, pointing to the presence of a nasal horn. The tips of the nasals are slightly bent downward. The naso- maxillary notch is well shown on either side, extending backward to above the an- terior border of the first tooth, the well-worn DM1. The depth of the notch is 16 cm. from the nasal tips on the left, and 18 cm. from the tips on the right side. Only 25 mm. behind it there is the infraorbital foramen, placed above P2. The premaxillaries are fortunately preserved in the skull, forming two strong, 4 cm. high, converging bones that are slightly inclined downward. They each carry a relatively small incisor, but no other teeth behind these. The premaxillary-maxillary suture is obliterated, but together with the maxillary processes to which they are attached the premaxillaries have a length of 14 cm., projecting only slightly less forward than the nasals. The height from the lower surface of the tips of the pre- maxillaries to the tips of the nasals is n cm. ; the height of the skull from the alveolar margin of M1 to the upper surface of the frontals is approximately 16 cm. The mandible belonging to the skull is well preserved on the right side ; the left body is broken off behind M3. The high ascending portion with the coronoid and condyloid processes fits well into the laterally compressed temporal fossa, and the distorted condyle articulates in the glenoid cavity when the tooth-rows are in occlusion. The lower canines are close to their antagonists, the upper incisors. Between the canines the mandible shows two small incisors. The profile of the mandible is nearly straight ventrally, with the symphysial portion only slightly curved upward, and the angular process is broad and well rounded behind. The characters of the present skull and mandible, notably the indications of the presence of a nasal and of a frontal horn, the presence of smallish upper incisors, the small incisors between the lower canines and the union of the two inferior squamosal processes, leave no doubt as to their belonging to the genus Dicerorhinus. To this genus a number of Tertiary and Pleistocene species have been referred, evidently forming several collateral lines, in Europe as well as in Asia, where it survives as the Sumatran rhinoceros, Dicerorhinus sumatrensis (Fischer). The extant form is not the most advanced : although it lacks the central lower incisors present in the fossil I24 forms it has an opening between the postglenoid and the post-tympanic process, a supposedly less specialized character than the union of these seen in fossil Dicero- rhinus (see Flower 1876 : 456). A second individual of Dicerorhinus leakeyi from Rusinga is represented by an upper dentition and right zygomatic arch, and a mandible of which only the right half is entirely preserved. These specimens are marked no. 2, R.I, 1947. The upper dentition agrees perfectly with that of the holotype skull of D. leakeyi in characters, and the lower jaw differs only in being slightly convex anteroposteriorly below, longer, less high below M3, and in the absence of Pl (PI. 2, figs. 3-4). Measurements of the holotype skull and mandible of D. leakeyi as well as of the lower jaw no. 2 from Rusinga are given in Table I together with those of Dicerorhinus sansaniensis (Lartet) from the Vindobonian of Sansan (Kaup 1854 : 3 ; Filhol 1891 : 200), Dicerorhinus schleiermacheri (Kaup) from the Pontian of Eppelsheim (Kaup 1834 : 40-41), Dicerorhinus orientalis (Schlosser) from the Pontian of Pikermi (Gaudry 1862-67 : 184, as Rh. pachygnathus , and p. 206, as Rh. schleiermacheri ; see Ringstrom 1924 : 12), and Dicerorhinus ringstroemi Arambourg (1959 : 73) from the Pontian of North China (Ringstrom 1924 : 12, as D. orientalis}. Table I shows that the Rusinga form is rather similar in size to D. schleiermacheri and D. orientalis, D. sansaniensis being smaller, and the Chinese species larger. TABLE i Measurements of skull and mandible of Dicerorhinus (mm.) D. leakeyi , D. sansan- D. schleier- D. orien- D. ring- Length from occipital crest to tip of nasal bones Width over postorbital processes Height of occiput from basion Depth of naso-maxillary notch from tip of premaxillaries From naso-maxillary notch to anterior border of orbit From anterior border of orbit to external auditory meatus From tip of premaxillaries to anterior border of P2 Length of mandible from front to back of angular process Length from front to P2 Length of symphysis Height below M3 Height of coronoid process Height of condyloid process Type c. 630 no. 2 ^ens^s 468 C. 145 — c. 135 — 260 — 150 — 500 530 no 90 130 120 105 85 c. 275 270 230 22O 95 2l8 470 70 192 macheri 640 200 200 142 1 60 246 154 142 278 248 talis 645 2IO stroemi 745 236 220 150 c. 500 90 635 . 95 132 The Dicerorhinus skull from Rusinga differs from that of D. schleiermacheri as figured by Kaup (1834, pl.io, fig. i ; 1854, pi- I0> ng- *) in tne occiput being more MIOCENE RHINOCEROSES OF EAST AFRICA 125 elevated ; the fronto-parietal profile is more concave. There is a marked sagittal crest in D. schleiermacheri , and the premaxillari.es are not inclined downward as in the Rusinga skull but project straight forward. The symphysial portion of the mandible is much more curved upward in D. schleiermacheri than in D. leakeyi so that the upper I and the lower C nearly touch each other just the same. The infraorbital foramen is placed further back in D. schleiermacheri than in D. leakeyi, viz., c. 40 mm. behind the naso-maxillary notch and above P3. The lower border of the mandible is not straight, but slightly concave behind the symphysis in D. schleiermacheri ; the angular process, however, is equally rounded. The almost equally large skull of D. orientalis from Pikermi (Gaudry 1862-67 : pi. 32, fig. i, as Rh. schleiermacheri) likewise has the occiput less elevated than the Rusinga skull, but it does not have a sagittal crest. The premaxillaries are in- complete and carried either reduced incisors or none at all (Ringstrom 1924 : 18-20). The infraorbital foramen is nearer to the naso-maxillary notch (12-18 mm.) and is placed above P3. The mandible of D. orientalis (Gaudry 1862-67 : pi- 2&> fig- I> as Rh. pachygnathus but referred to orientalis by Ringstrom 1924 : 21) has a slightly convex lower margin ; the angular portion is incomplete, but the lower canines are quite reduced. The skull of D. ringstroemi from China (Ringstrom 1924 : 6, text-figs, i, 2) again does have the flat profile of D. schleiermacheri and D. orientalis ; the fronto-parietal crests do not meet and form no sagittal crest (least distance 45 mm.). The maxillary processes and the premaxillaries have broken off. The mandible (Ringstrom 1924 : 10, text-figs. 3, 4) has a slightly convex lower profile and quite reduced canines, as in D. orientalis ; the symphysis is curved upward. It is only in the skull and mandible of D. sansaniensis (Duvernoy 1853 : pi. i, fig. la ; Kaup 1854 : pi- IO> fig- 2 '> Filhol 1891, pis. 13, 14) that the salient characters of D. leakeyi are present. The occiput is raised to the same extent as in the Rusinga skull, and the mandible is not much curved upward in its symphysial portion and nearly straight below in profile. Unfortunately the premaxillaries are missing in the Sansan skull, but the lower C are present. They are relatively well developed and between them there are two small incisors. Unfortunately the available skull of Dicerorhinus primaevus Arambourg (1959) from the Pontian of Wad el Hammam in Algeria is that of a young individual with milk teeth, and M1 erupting, lacking the posterior portion and the premaxillaries. The skull of Dicerorhinus caucasicus Borissiak (1938) from the Vindobonian Chokrak beds in North Caucasus is deformed and incompletely known. Hence, only the dentition and skeleton of these forms are available for comparison. Regarding the dentition of Dicerorhinus leakeyi, it is most convenient to deal with all the dental material in the East African Miocene collection that may be safely referred to Dicerorhinus. Both upper incisors are preserved in the holotype skull of D. leakeyi and of the upper dentition no. 2, R . i (PL 4, fig. 7). The crowns are elongated anteroposteriorly, with the lateral surface convex and the medial undulating, convex in front and behind and depressed in between. The lateral surface is more worn down than the medial. 126 MIOCENE RHINOCEROSES OF EAST AFRICA The root is transversely compressed, diminishing slightly in diameters toward its blunt apex. There are no less than eight isolated upper incisors agreeing well in shape and size with those of D. leakeyi, three of which are in the British Museum (Natural History) : a left specimen marked Rs.3, Rusinga, a right specimen marked R. i, Rusinga, and a left specimen marked R.n, Rusinga. The National Museum specimens consist of four from the right side : no. 109, 1949, West side of Hiwegi, Rusinga (PL 4, fig. 8), no. 275, 1949, Kathwanga, Rusinga, no. 81, 1950, R.i-ia, Rusinga, and F.3056, Kathwanga, Rusinga, and a left specimen numbered F . 3060 , Rusinga. Measurements are given in Table 2. TABLE 2 Measurements of upper incisors of Dicerorhinus leakeyi (mm.) Type dext. sin. dext. sin. Rs.3 R.I R.n no. 109 no. 275 Ant. post. 38 37 35 34 3& 33 34 Transv. 18 18 16 15 14 15 14 15 15 no. 8 1 F.3O56 F.3o6o Ant. post. 32 32 Transv. 14 15 The root is well preserved in most specimens, varying from 3-5 to 4 cm. in length. The amount of individual variation within this series is not very great. A very much larger upper incisor from Rusinga will be dealt with below under Brachypotherium. The upper incisors of D. schleiermacheri (Kaup 1834 : 34. pi- IT> fig8- 3> 4) are similar in crown size to those of D. leakeyi : diameters 35 by 16 mm. and 33 by 13 mm., but the root appears to be more massive in the Eppelsheim species. Moreover, there is a small second incisor behind the larger one in each premaxillary, 8 mm. in diameter (Kaup 1834 : 34) ; of these there is no trace in the Rusinga skull. In D. orientalis the upper incisors have not been found, and from the tapering (incomplete) premaxillaries it may be concluded that they were either reduced or absent. The same applies to D. ringstroemi. The foremost teeth present in the holotype skull D. leakeyi are the first upper milk molars, DM1. They are, however, too worn and damaged to reveal much of their structure ; the milk dentition of Dicerorhinus will be dealt with later. The upper premolars P2~4 of the type skull of D. leakeyi as well as those of the dentition no. 2, R.I (PI. 2, fig. 3, PI. 5, figs. 2, 3), (P2 sin. is not preserved) are much worn down. The external surface of the left P2 and P4 as well as that of the right P3 in the type skull are incomplete, and the right P2 and P3 are incomplete internally. P2 is worn to such a degree that a small portion of the medisinus only remains on the occlusal surface. P3 and P4 both have protoloph and metaloph connected internally forming a high lingual wall up to at least 12 mm. (P3) or 15 mm. (P4) above the enamel MIOCENE RHINOCEROSES OF EAST AFRICA 127 base ; the teeth are worn down to this level. There is a weak cingulum on the in- ternal surface, 8 mm. high on the protoloph, and slowly rising behind. In the least worn premolar (P4) the lingual third of the medisinus, well cut off from the internal wall, and the postsinus are shown on the worn surface. The internal surface of the crowns of P3 and P4 has a weak vertical depression at the junction of protoloph and metaloph ; the protocone is only weakly constricted off. There are various isolated upper premolars, all much worn, that should be referred to Dicerorhinus, viz., No. 752, 1951, Rusinga, P2 sin., incomplete postero-internally, No. 1385, 1951, Rusinga, P3 dext. (PI. 6, fig. 12), No. 2549, 1952, Rusinga, P3 sin., damaged behind, and No. 80, 1950, R. i-ia, Rusinga, P4 dext. These specimens all have the high internal wall of the Dicerorhinus type ; the internal cingulum is hardly visible in some specimens. Measurements have been entered in Table 3. TABLE 3 Measurements of upper premolars of Dicerorhinus leakeyi (mm.) Type no. 2 no. 752 P2, ant. post. 26 31 29 ant. transv. 34 35 35 post, transv. 36 37 no. 1385 no. 2549 P3, ant. post. 29 33 31 ant. transv. 43 46 46 47 post, transv. 44 43 no. 80 P4, ant. post. 31 38 34 ant. transv. 51 50 52 post, transv. 49 47 49 The upper molars of the holotype skull of Dicerorhinus leakeyi are characterized by their low and wide lingual medisinus entrances. All except M2 dext. lack portions of the external surface, while M1 dext. is incomplete antero-internally as well. The molars of the upper dentition no. 2, R. i, Rusinga (only those from the right side are preserved), are very similar to those of the holotype and worn to a slightly less degree ; the M2 is incomplete antero-externally and M3 is incomplete behind. There is no manifestation of a lingual cingulum in the molars except for a small tubercle at the medisinus entrance of the M1 in no. 2. The medisinus is not blocked by the ante- crochet, which is hardly visible ; the protocone is not or hardly constricted. The crochet is weak and there is no crista. The ectoloph is much depressed between the roots. 128 MIOCENE RHINOCEROSES OF EAST AFRICA The M3 has a character that lends it a primitive look, viz., the bulging out of the outer surface at the junction of ectoloph and metaloph. This bulge, representing the metacone and supported by a heavy postero-external root, gives the crown a trape- zoid, somewhat M1-2-like outline quite different from the nearly triangular outline found in the M3 of Aceratherium acutirostratum (Deraniyagala) (Hooijer 1963, pi. 7, fig. i), in which ectoloph and metaloph are confluent without a bulge to mark their junction. In this character the M3 of Dicerorhinus is definitely more primitive than that of Aceratherium. The projection of the metastyle and posterior half of the metacone in M3 is lost early in the history of the Rhinocerotidae, as fully discussed by Wood (1927). In the Eocene Hyrachyus (Wood 1934, pi. 22) there is a marked posterior projection of the metacone and metastyle in M3. This projection is no longer discrete in Dicerorhinus, having merged into the outer surface, but the basal bulge is still there.1 It is interesting to note that in Recent Dicerorhinus sumatrensis M3 shows the same development : ectoloph and metaloph, although confluent, form a wide angle with a basal bulge (Hooijer 19460, pi. 2, fig. 8, pi. 3, fig. i). There are numerous isolated upper molars in the East African Miocene collection that present the Dicerorhinus type, as follows : No. 1163, 1950, R.I, Rusinga, M1 sin., ectoloph incomplete, No. 1161, 1950, R.I, Rusinga, M2 sin. (PI. 6, fig. 7), No. 82, 1950, R.i-ia, Rusinga, M1 sin. without parastyle and metastyle, No. 37, 1947, Songhor, anterior outer fragment of left upper molar, No. 485, 1948, Kathwanga, Rusinga, M3 dext., outer and anterior parts lost, No. 14, 1949, R.I, Rusinga, M3 dext., No. 711, 1949, Gumba, Rusinga, M3 dext., much worn, outer surface broken off, two unnumbered portions of M3 sin., Rusinga, one much worn down and with the enamel incomplete anteriorly, the other the posterior portion only, No. 1162, 1950, R.I, Rusinga, M3 sin., corroded enamel and incomplete behind. The following specimens are in the British Museum (Natural History) : M1 dext., R.I, Rusinga, lacking metastyle, M2 dext., R.2, Rusinga, metastyle and part of protoloph wanting, M2 sin., Rusinga, lacking outer portion, M3 sin., R.I, Rusinga (PI. 7, figs. 5, 6) and M3 dext., Rs.ioS, Rusinga, lacking posterior portion, weak crista. Measurements of these specimens are presented in Table 4. In a collection from Napak, Karamoja, Uganda, entrusted to me by Dr. W. W. Bishop, there is the anterior surface of the protoloph of a left upper molar (no. 509, Napak VI, 1961 (i)) with a very weak protocone fold that I have no doubt should be referred to Dicerorhinus. It can be exactly duplicated in the Rusinga material listed above. 1 Recently Wood (1963) described a primitive true rhinoceros from the Late Eocene of Mongolia as Pappaceras confluens, the specific name referring to the essentially confluent ectoloph and metaloph of M3, a rather advanced structure in such an early Tertiary form. MIOCENE RHINOCEROSES OF EAST AFRICA 129 TABLE 4 Measurements of upper molars of Dicerorhinus leakeyi (mm.) Type no. 2 no. 1163 no. 82 R.I M1, ant. post. 36 4i — — 40 ant. transv. 50 52 52 53 50 post, transv. 46 49 49 49 47 p. tr. : a. tr. 0-92 0-94 0-94 o-93 0-94 no. 1161 R.2 M2, ant. post. 4i 43 42 43 ant. transv. 55 55 54 55 post, transv. 46 47 48 49 p. tr. : a. tr. 0-84 0-85 0-89 0-89 no. 14 no. 711 no. 1162 R.I M3, ant. post. 42 46 44 45 42 ant. transv. 5i 53 5i — c. 50 5° 1. outer surface 52 — 55 — 53 Rs.ioS Rs. 43 50 54 57 The ratio of the postero-transverse to the antero-transverse diameter shows that M2 is relatively narrower behind than is M1. In M3 the metacone bulge can be seen in most specimens ; the M3 sin. from R . I is entire behind and shows the bulge very well. In Recent D. sumatrensis the bulge is as marked : in an M3 sin. f rom a Sumatran cave (Dubois Collection no. gioa) as in nearly all the Sumatran cave teeth, the roots have been eaten away by porcupines and only the enamel of the crown, nibbled at the edge, remains, but in basal view the trapezoid outline of the crown is well seen, correspond- ing with that made by the heavy rounded root of the metacone in the Rusinga tooth. The posterior width (metacone-hypocone) happens to be the same in the two speci- mens (37 mm.), and the antero-transverse diameter is also the same in the two (50 mm.). The posterior cingulum is somewhat better developed in the Rusinga speci- men than in that from Sumatra, but this is a highly variable feature : in another Sumatran M3 (Dubois Collection no. 663a) the posterior cingulum is as well developed as that in the Rusinga M3. The upper premolars and molars of Dicerorhinus schleiermacheri as figured by Kaup (1834, pi. n, fig. 5) are larger than those in the Rusinga Dicerorhinus ; the length P2-M3 is 250 mm. as figured by Kaup against 205 mm. in the holotype skull and 220 mm. in dentition no. 2 of D. leakeyi. The maximal transverse diameters given in the text (Kaup 1834 : 37) are •' P4 60 mm. ; M1 62 mm. ; M2 65 mm., whereas in D. leakeyi the antero-transverse diameters are at most 52 mm. in P4, 53 mm. in M1 and 55 mm. in M2. In structure, however, the Eppelsheim premolars and molars resemble those of Rusinga very closely : there is a weak internal cingulum in the premolars above which protoloph and metaloph are confluent, the protocone is weakly constricted, while in the molars the internal cingulum is absent, the medi- sinus entrance low and wide, the protocone constriction weak, and M3 has a bulge at the metacone base. The crista and the crochet are weak or absent in D. leakeyi but these projections into the medisinus recede towards the base and hence show better i3o MIOCENE RHINOCEROSES OF EAST AFRICA in less worn teeth such as those of D. schleiermacheri (in various molars of D. leakeyi there are traces of crista and crochet). The upper dentition of Dicer orhinus san- saniensis (Filhol 1891 : 201, pi. 14, fig. 6 ; Pavlow 1892, pi. 5, fig. 15) agrees in charac- ters with that of schleiermacheri and leakeyi (it shows no crista and a weak crochet), but in size it is less than that of the Rusinga Dicerorhinus (length P2-M3 190 mm., maximal width of P4 45 mm., of M1 46 mm., and of M2 47 mm.). Hence, the upper teeth of D. leakeyi are intermediate in size between those of D. schleiermacheri and D. sansaniensis. The upper molars available of D. primaevus show this to be a larger species than D. leakeyi : the width of M1 is 57-5 mm. and that of M2 62-5 mm. (Arambourg 1959 : 59). In D. caucasicus the upper premolars as described by Borissiak (1938 : 9-13) have the crests internally confluent upon wear, as in D. schleiermacheri, D. sansaniensis and D. leakeyi, but in size D. caucasicus is nearly as small as D. sansaniensis (width of P4 37 mm., of M1 51 mm., of M2 48 mm.). The canines in the mandible of the holotype of D. leakeyi are subtriangular in cross- section, with rounded upper and lower outer edges and a sharp edge internally. At the base of the crown the diameters are 30 mm. horizontally and 21 mm. vertically. The distance between them amounts only to n mm., but between them, or rather crowded out below the inner edges of the canines there are two small incisors, round in section and with pea-shaped crowns 8 mm. in diameter. In mandible no. 2 there are only the alveoli of the incisors and canines, but they resemble those in the holotype very closely. The depth of the alveolus of the left canine, exposed laterally, is just over 10 cm. An isolated pair of lower canines thought to belong to no. 2 (PI. 4, figs. 4, 5) have crowns agreeing exactly in shape and size with those of the type ; the crown length is 4-5 cm. and the roots, which are straight and gradually taper toward the apex, have a length of 9 cm. The tip of a left lower canine, no. 980, 1950, Kiangata, Rusinga, is 30 mm. trans- versely at crown base and very nearly round in section at the root (30 mm. horizon- tally and 28 mm. vertically) . The length of the worn crown is only 27 mm. , indicating perhaps that it was less procumbent than the others. Whether this canine belongs to Dicerorhinus or to some other genus is uncertain. We have further the lower canines of the Aceratherium specimen described under that head. The lower incisors and canines of D. schleiermacheri are similar in size to those of D. leakeyi ; the diameter of the incisor alveoli is 8 mm., and those of the canine 30 mm. horizontally and 22 mm. vertically (Kaup 1834, pi. n, fig. 8 ; there is a cast in the British Museum (Natural History), no. M.2782). In the mandible of D. sansaniensis there are likewise small incisors between the canines (see Filhol 1891 : 201, pi. 14, fig. i ; Roger 1900 : 51, pi. i, fig. 2), diameters 6 mm., and 20 mm., respectively, smaller than in the Rusinga mandibles. In D. orientalis as well as in D. ringstroemi the lower incisors or canines are much reduced (Ringstrom 1924, text-figs. 14, 15 and 3, 4). In D. primaevus only traces of the alveoli of lower incisors have been found (Arambourg 1959 : 60), and in D. caucasicus, as in D. schleiermacheri to which it may be ancestral, there are subtriangular lower canines (erupting) and two small cylindrical incisors in between (Borissiak 1938 : 16, pi. i, fig. 4). The lower canine described from the Miocene of the Sinda-Mohari region, Congo MIOCENE RHINOCEROSES OF EAST AFRICA 131 (Semliki no. 527, Hooijer 1963 : 48), as possibly belonging to Dicerorhinus is similar to that of D. leakeyi although smaller (basal crown diameters 22 by n mm.). All the lower premolars and molars of the holotype of D. leakeyi are present, in- cluding a small and slightly worn Pr Those of dentition no. 2, R . I, 1947, are slightly larger ; only the right side of the mandible is intact and it shows no trace of a Px, while nothing is preserved of the left ramus except the isolated P4, M2 (PI. 6, fig. 8) and M3. These teeth are characterized by having a deep vertical groove externally between the two lophids, the metalophid in front and the hypolophid behind, and by the absence of an external cingulum. There are a few more mandibles presenting the same characters, as follows : M. 1892 1 is a mandible from Rusinga that has alveoli for the I and C as in the Dicerorhinus specimens. The teeth preserved are the right P4 and M2_3, and the left P3-M3. The ascending portions of the rami are restored with plaster behind. The teeth present no differences from those in the holotype and no. 2. No. 231, 1949, a right mandibular ramus from R.2-4, Rusinga, has P4 and M2_3 in situ. The symphysial portion is missing, and of the postdental portion only the angular process is preserved. The teeth agree with those of Dicerorhinus in size and characters. Measurements are given in Table 5. The length of the mandibular tooth-row P2-M3 is 250 mm. in D. schleiermacheri and 196 mm. in D. sansaniensis (Kaup 1854 : 3), longer, and shorter, respectively, TABLE 5 Measurements of lower premolars and molars of Dicerorhinus leakeyi (mm.). Type no. 2 M. 18921 no. 231 P1( ant. post. 18 trans v. n P2, ant. post. 26 28 ant. trans v. 16 16 post, transv. 17 18 P3, ant. post. 31 34 ant. transv. 21 20 21 post, transv. 23 23 24 P4) ant. post. 35 37 35 36 ant. transv. 26 26 26 post, transv. 28 28 27 ML ant. post. 41 ant. transv. 29 28 post, transv. 30 31 M2, ant. post. 41 45 45 43 ant. transv. 32 30 31 post, transv. 31 31 31 27 M3, ant. post. 44 51 49 44 ant. transv. 30 30 31 28 post, transv. 28 29 27 28 Length P2-M3 215 235 225 i32 MIOCENE RHINOCEROSES OF EAST AFRICA than in the D. leakeyi specimens (Table 5). In the figured mandible of D. schleier- macheri the anterior premolar Px is absent, but it occurs in one out of twenty-four mandibles (Kaup 1834 : 37, see pi. n, fig. 8a), and it is about n by 7 mm. in crown diameters. In D. sansaniensis a Px is present in the figured mandible, diameters 15 by 10 mm. (Filhol 1891 : 201). In the second mandible of D, leakeyi P1 has not developed. It is evident that the presence or absence of this tooth is not a character of great significance. Apart from a few isolated lower molars that have a flattened external surface and therefore belong to Brachypotherium there remain a number of lower jaw fragments with teeth and isolated lower teeth that have the groove externally between meta- and hypolophid. Although they are very similar in size to those of Dicerorhinus they do not necessarily belong to that genus. They may in part represent Aceratherium (the lower molars of Aceratherium cannot be distinguished from those of Dicerorhinus), and some of them may even represent Brachypotherium if the flattened outer groove is not constant in the Rusinga and Karungu Brachypotherium. However, since Dicerorhinus is the most common of the East African Miocene rhinoceroses, most of the lower teeth probably belong to that genus, and are enumerated below: No. 786, 1948, West Hiwegi, Rusinga, left mandibular ramus with P3-M3, teeth all fractured and incomplete, P4 slightly worn, M3 erupting, No. 788, 1948, West Hiwegi, Rusinga, a right ramus, teeth broken off, No. 1063, Kiahera Hill, Rusinga, P2 sin., No. 990, 1950, Gumba Rusinga, P3 dext., No. 1397, 1950, Rusinga, Mx dext., No. 223, 1949, Wakondu, Rusinga, right ramus fragment with M2_3, No. 429, Ngira, Karungu, unerupted crowns of P2_4 dext., slightly worn Mj dext., and unworn incomplete crown of M2 dext., No. 596, 1947, junction R.r and R.ia, Rusinga, left ramus with M2_3, Five isolated teeth, all marked RS, Rusinga, representing the left P4 and the right and left M2_3, all much worn down, No. 342, 1950, Ngira, Karungu, incomplete P3 dext., A ramus fragment marked " Aloir, '39, Owen " bears the damaged crowns of the left P4 and Mlf A ramus fragment marked " Karungu, 1937, Owen" has a much worn left M2 that is incomplete externally behind, Nos. 84-86, 1950, R.i-ia, Rusinga, incomplete P4 and M2_3 dext., possibly belonging together, F.3O62 and F.3057 are two portions of a left lower molar from Rusinga, No. 83, 1950, R.i-ia, Rusinga, a much worn and incomplete Mx or M2 sin. Mx is wider behind than in front. In M2 the posterior lophid is either wider or narrower than the anterior, and in M3 either the width is greatest in front, or (no. 231 ; Table 5) there is no difference between anterior and posterior width. M3 is further characterized by the slight development of the posterior cingulum and (but this only upon a certain amount of wear) the absence of a posterior pressure scar. Measurements of the Rusinga and Karungu specimens are given in Table 6. MIOCENE RHINOCEROSES OF EAST AFRICA TABLE 6 Measurements of lower teeth from Rusinga and Karungu (mm.) 133 P2, ant. post. ant. transv. post, transv. P3, ant. post. ant. transv. post, transv. P4, ant. post, ant. transv. post, transv. Mj, ant. post, ant. transv. post, transv. M2, ant. post. ant. transv. post, transv. M3, ant. post. ant. transv. post, transv. no. 786 23 39 26 26 29 27 no. 1063 27 16 19 no. 990 35 20 22 no. 1397 42 26 28 43 28 29 45 29 26 no. 429 28 15 18 38 25 27 44 27 no. 223 no. 596 26 27 44 27 26 RS 24 27 no. 342 nos. 84-86 21 23 Aloir '39 25 Karungu '37 47 45 28 30 For the sake of completeness I mention the lower molar fragments that comprise only the hypolophid, and do not show whether the groove between it and the meta- lophid was deep or flattened. These are R.io6, Rusinga; no. 346, 1950, Ngira, Karungu; no. 446, Kachuku, Karungu, and F.3O5I, Rusinga. The last two are posterior portions of M3, right and left, and 26-28 mm. wide. Nothing can of course be said as to the generic position of these fragments. In the British Museum (Natural History) there is a set of upper milk molars, DM1"4 sin., in situ in a maxillary, very well preserved, and marked Rs.26, Rusinga (PL 5, fig. i). DM1 is a small triangular tooth the ectoloph of which bears a vertical ridge in the middle, the mesostyle. The protocone, which is placed internally at the same level as the mesostyle, is an isolated cusp. The front part of the crown narrows anteriorly and is bounded lingually by a low ridge and buccally by the ectoloph crest, which unite at the antero-external angle so as to enclose a shallow fossa. The hypocone behind the protocone is connected with the metacone by a low ridge separa- ting the medisinus from the postsinus. DM2 has a mesostyle (not present in the more GEOL. 13, 2. 9 i34 MIOCENE RHINOCEROSES OF EAST AFRICA posteriorly placed milk molars or in the permanent dentition). The protoloph and the metaloph are damaged internally but appear to be equally long transversely. Both the paracone style and the metacone style are weak. The entrance to the medisinus is low and V-shaped. There is a crista united with the crochet, closing off a medifossette. DM3 has no crista, and the crochet extends only half way across the medisinus. The ectoloph is depressed between the roots. The paracone style is only in part preserved and the parastyle has broken off. The medisinus entrance is wider than that in DM2 and likewise unobstructed. The grooves delimiting the protocone are faintly shown. DM4, of which the paracone style is incomplete, has the valley entrance again wider, the crochet longer, but the protocone constriction as little developed as in DM3. The anterior cingulum is more prominent, but, like in DM3, it does not quite extend to the antero-internal crown angle. There is no trace of a cingulum internally. The posterior moiety of the ectoloph with the faint meta- cone style is much more inclined inward than the anterior half. The DM4 of the Rusinga milk dentition agrees so well with the M1 of the Dicero- rhinus dentitions from Rusinga, differing mainly in lesser size (Table 7) and thinner enamel, as may be expected, that the juvenile Rusinga specimen may also be referred to D. leakeyi. TABLE 7 Measurements of upper milk teeth of Dicerorhinus leakeyi (mm.) DM1, ant. post. 24 DM3, ant. post. 32 trans v. 20 ant. transv. 37 DM2, ant. post. 26 post, transv. 35 ant. trans v. c. 30 DM4, ant. post. 36 post, transv. c. 30 ant. transv. 40 post, transv. 39 p. tr. : a. tr. o There are two specimens of DM4 from Rusinga that should be referred to Acera- therium, and they will be dealt with later. There are also a much worn right DM1 (no. 546, 1950, R.3, Rusinga) with a length of c. 25 mm. and another DM1 dext. from Rusinga that is 22 mm. wide. The upper milk dentition of D. schleiermacheri as figured by Kaup (1834, pi- TI» fig. 7) is larger than that of D. leakeyi ; the overall length of DM2~4 is 120 mm. as opposed to 103 mm. in the Rusinga specimen, and the width of DM3 is 40 mm., that of DM4 42 mm. (Kaup (1834 : 37) gives 52 mm. for the width of DM4, but a cast of the figured specimen in the British Museum (Natural History), no. nob (O.C.), shows it to be 42 mm.). The crochet is better developed in the Eppelsheim milk teeth than in those from Rusinga. The milk dentition of D. sansaniensis (Pavlow 1892, pi. 5, fig. 8) has a length DM2-4 of 104 mm., and widths of DM3 and DM4 of 35 and 37 mm. respectively, as measured from the figures, slightly less than in D. leakeyi. The milk dentition of D. primaevus as given by Arambourg (1959 : 59) is larger than that of D. leakeyi (length DM2-4 123-5-131-5 mm., width DM3 42-5-45 mm., of DM4 44-50-4 MIOCENE RHINOCEROSES OF EAST AFRICA 135 mm.). It is within these limits that falls the DM4 of Dicer or hinus (?) spec, from the Pontian of Gravitelli, Sicily, the width of which is 47 mm. (Hooijer, 194.60 : 322). The lower milk dentition of Dicerorhinus leakeyi is best represented in a specimen from Napak V (August, 1962), Uganda, sent to me for identification by Dr. W. W. Bishop. It comprises DM2_4 dext., with the crowns unworn ; the metalophids of DM 3 and DM4 are damaged apically behind. The posterior valley of DM2 is closed inside, the anterior valley of DM3 nearly so. In DM3 the metalophid is bilobed in front, the parastylid well developed (PI. 4, fig. i). This dentition tallies well in size with the upper milk dentition of Dicerorhinus from Rusinga and may be referred to the same species. There are more remains of the lower milk dentition from Karungu and Rusinga some of which present larger dimensions than the Napak specimens, but the difference is small. These are: No. 429, Ngira, Karungu, right DM3 and part of DM4 in a ramus fragment (this number also includes the unerupted premolars evidently of the same in- dividual), No. 405, 1951, Rusinga, right ramus with DM3_4 and erupting Mx, No. 1580, 1950, Rusinga, part of left DM3 in ramus fragment, F.3058, Rusinga, posterior portion of DM3 and anterior portion of DM4 sin., F.3059, Rusinga, anterior portion of DM4 dext., and No. 485, 1948, Kathwanga, Rusinga, posterior portion of DM4 sin. Measurements of the lower milk molars are given in Table 8. TABLE 8 Measurements of lower milk teeth from Napak, Karungu, and Rusinga (mm.) nos. 3058/9 Napak V no. 429 no. 405 no. 1580 and 485 DM2, ant. post. 27 ant. transv. 13 post, transv. 16 DM3, ant. post. 37 36 40 ant. transv. 17 19 19 18 post, transv. 20 21 21 DM4) ant. post. 34 37 ant. transv. 19 22 23 19 post, transv. 20 20 DM2 in Z>. schleiermacheri (Kaup 1834 : 38, pi. n, fig. 10) measures 28 mm. antero- posteriorly and 15-5 mm. in width ; it has the posterior valley closed inside. The various milk dentitions of D. primaevus (Arambourg 1959 : 63) are larger than the East African DM2_4 except in the width of DM2 which is given as 13-5-15 mm. ; the width of DM3 is 22-26 mm., and that of DM4 23-5-26-5 mm. The posterior valley of DM 2 is open inside, making it a narrow tooth in D. primaevus, but as observed by 130 MIOCENE RHINOCEROSES OF EAST AFRICA Arambourg (1959 : 62) the valley may be either open or closed in DM2 of D. schleier- macheri, and this varies in the Recent species also : the posterior valley in DM2 becomes either shallower or deeper as it passes outwards. In the latter case it may become isolated as an enamel island upon wear and show a lingually closed valley (Hooijer 19460 : 32). Genus ACERATHERIUM Kaup 1832 Aceratherium acutiro stratum (Deraniyagala) (PL 3 ; PI. 4, figs. 2, 3 ; PL 5, figs. 4, 5 ; PL 6, figs. 1-4 ; PL 9, figs. 2-4 ; PL 14, figs. 4, 5) 1 95 1 a Turkanatherium acutirostratus Deraniyagala : 24. 1951 Turkanatherium acutirostratus Deraniyagala : 134, pi. i. 1953 Turkanatherium acutirostratus Deraniyagala ; Deraniyagala : 13, pi. i ; pi. 2, figs, b, c ; pi. 3, figs, b, d. The skull from Moruaret Hill near Losodok described by Deraniyagala (1951, 1953) as Turkanatherium acutirostratus has elongate, weak nasals and a weak double sagittal crest. There is neither a nasal nor a frontal horn. The premaxillaries are well developed and, although incomplete, extend forward beyond the nasals. They evidently carried incisors, but these are lacking. The occiput is elevated, giving a concave fronto-parietal profile. The dentition of the Turkana skull is characterized by the antecrochet blocking the medisinus in the premolars, which have a prominent internal cingulum, and a markedly constricted protocone and strong antecrochet in the molars. These are characters occurring in Aceratherium, and I have recently described a number of teeth from the Miocene of Congo as Aceratherium acutirostra- tum (Deraniyagala) (Hooijer 1963 : 43, pi. 6, figs. 1-3 ; pi. 7, figs, i, 3-5, 8 ; pi. 8, fig. 2). Earlier, Arambourg (1959 : 74) had already stated that Turkanatherium appeared to be a synonym of Aceratherium. There are various Aceratherium species in the Tertiary of Europe known by more or less complete skulls, and the Turkana Aceratherium may be compared with these. The high occiput is found only in the Pontian Aceratherium incisivum (Kaup 1834, pi. 10, fig. 2) ; in the earlier species the occiput is less elevated (see Wang 1928, text-fig, i). Aceratherium incisivum, as first discovered by Osborn (1899), has a frontal horn, of which there is no evidence in Aceratherium acutirostratum or in the pre-Pontian European species. Elongate weak nasals are common in Aceratherium ; those in the Turkana skull appear to agree best with those of Aceratherium lemanense (Pomel) (Mermier 1896, pi. i). In this Stampian and Aquitanian species the naso- maxillary notch extends backward only to above the middle of P3, as in the Turkana skull ; in A ceratherium platyodon Mermier of the Burdigalian and in A ceratherium tetradactylum (Lartet) of the Vindobonian the notch extends backward to above the middle of P4, whereas in A. incisivum it extends to above the front of M1 (Mermier 1896, pi. i ; Breuning 1924 : 13). In the Turkana skull the anterior border of the orbit is above the anterior border of M2, as in A. lemanense, A. platyodon, and A. tetradactylum ; in A. incisivum it is above the middle of M2 (Mermier 1896, pi. i). MIOCENE RHINOCEROSES OF EAST AFRICA 137 Aceratherium lemanense has a true sagittal crest (skull from Gannat figured by Roman 1912, pi. 8, figs, i, 10) as well as one of the Aceratherium incisivum skulls (Kaup 1834 : pi. 10, fig. 20) ; in the other skull of A. incisivum (Kaup 1834, pi. 10, fig. 26) the fronto-parietal crests do not meet. In the collection from R.I, Rusinga, there are remains of an Aceratherium skull, numbered 850, 1947. Unfortunately the skull is in fragments, and it has been possible only to assemble the nasals (PI. 9, fig. 3) and the fronto-parietal portion of the skull, not fitting together. In restoring these from the fragments both parts appeared to be distorted by pressure mostly from the right side. The nasals evidently did not carry a horn : there is no eminence or rugosity but a groove in the median line in- stead. The length from the tip to the beginning of the downward curve for the naso- maxillary notch is 14 cm. ; the width of the nasals from 5 cm. behind the blunt tip backward to the narial notch is only 8 cm. throughout. Thus, the nasals are long and slender, as in A. lemanense. The fronto-parietal portion shows at least the absence of a true sagittal crest ; the least distance between the two fronto-parietal crests is 4^ cm., as in one of the A. incisivum skulls referred to above. There are no indications of the presence of a frontal horn in the Rusinga skull. Neither the greatest width at the postorbital processes nor the dorsal profile can be made out from the Rusinga specimen. Of the upper dentition of skull no. 850 only a few fragments remain, and they do not show any characteristic features except for one that shows a pit inside the cingulum at the medisinus entrance ; all the crown fragments are much worn down. The mandible of the skull is preserved, but much fragmented and distorted. The symphysial portion is poorly preserved, and the canine alveoli cannot well be traced. The rami are incomplete below, and nothing is preserved of the postdental portions except for the tip of the left coronoid process. The dentition, P3-M3 of both sides, is well worn down, and will be dealt with later. The upper dentition of Aceratherium is less well represented at Rusinga than is that of Dicerorhinus ; there are, however, a few characteristic premolars. Two specimens, nos. 231 and 232, 1950, R.2-4, Rusinga, comprise a P4 dext. with part of the M1 attached to it, and a P4 sin. of the same individual. The antero-external corners of both premolars are broken off (PI. 6, figs. 1-4). The P4 is worn down slightly less than in the Dicerorhinus dentitions, viz., to 16 mm. from the internal enamel base (the M1 is worn down to 10 mm. ; it erupts earlier than the P4). There are marked differences between the P4 of nos. 231-232 and those of Dicerorhinus. The inner entrance to the medisinus is much lower in nos. 231-232, obstructed only by a ridge 8 mm. high, forming part of the inner cingulum, slightly rising behind and bordering a shallow pit. The bottom of the medisinus, which is only 7 mm. above the internal base of the enamel of the crown just inside the cingulum, rises inside the tooth (toward the external side) over a horizontal distance of 12 mm. to a level of 16 mm. above the enamel base. At this point the antecrochet extends completely across the medisinus. Buccally of the antecrochet the medisinus deepens again, and changes its course from transverse to obliquely forward and outward. The constriction of the protocone by anterior and posterior grooves is very marked. 138 MIOCENE RHINOCEROSES OF EAST AFRICA The differences may be tabulated as follows : Dicerorhinus P4 Aceratherium P4 proto- and metaloph united internally medisinus open lingually to 7 mm. from up to at least 15 mm. from crown base ; crown base ; internal cingulum internal cingulum weakly developed ; prominent, forming ledge between protocone constriction slight. proto- and metaloph ; marked proto- cone constriction ; antecrochet prominent. The type of P4 represented in the sample nos. 231-232 is that of Aceratherium acutirostratum as described and figured from the Miocene of the Western Rift, Congo (Hooijer 1963 : 43, pi. 6, figs. 1-3 ; pi. 8, fig. 2). The Congo P4 is less worn than the Rusinga specimens, and somewhat larger. It has a more developed internal cingulum, a ridge about 12 mm. high, at which level the medisinus opens internally, and ex- tending along the protoloph, too. Some 14 mm. inside the tooth the medisinus is blocked completely by a narrow antecrochet. The constriction of the protocone is marked. The following premolars should also be referred to Aceratherium : No. 991, 1950, Gumba, Rusinga, P2 sin., external portion missing; heavy cingulum internally, Two specimens in the British Museum (Natural History), one marked Rs.gi, Rusinga, P3 dext., slightly worn, inner cingulum only at medisinus entrance, narrow antecrochet across the medisinus, and the other a much worn P3 sin. from R.I, Rusinga, with internal cingulum almost absent but antecrochet across medisinus, F.3054 and F.3063 (one specimen), Rusinga, P3 dext., worn and incomplete but showing the pit inside the inner cingulum at medisinus entrance, No. 187, 1947, South of Kiahera Hill, Rusinga, P3 dext., worn and corroded, show- ing protocone constriction and antecrochet across medisinus, K.343, 1950, Ngira, Karungu, a P3 or P4 dext., sides broken off except internally, showing pit inside cingulum. Measurements are given in Table 9. TABLE 9 Measurements of upper premolars of Aceratherium acutirostratum (mm.) Lake Albert, Congo nos. 231-232 RS.QI R.I P3, ant. post. 32 ant. transv. 46 post, transv. 38 42 P4, ant. post. 46 ant. transv. 60 post, transv. 53 46 The upper molars of Aceratherium acutirostratum from Lake Albert, Congo (Hooijer MIOCENE RHINOCEROSES OF EAST AFRICA 139 1963 : 43, pi. 7, figs, i, 3-5, 8) are larger than those of Dicerorhinus leakeyi and differ in the protocone being markedly constricted off. Immediately external to the posterior protocone fold the antecrochet bulges out and basally extends all across the medisinus as seen in the figured M1. This is not the case in Dicerorhinus. The anterior cingulum is strongly developed and encroaches upon the internal surface of the protoloph. The M3 differs from that of Dicerorhinus in that the metacone bulge is not developed, making the crown triangular rather than trapezoid, and in the protocone being markedly constricted. The Congo M3 is unworn at the paracone, and the height of the crown at this point (49 mm.) is much less than the length of the outer surface (65 mm.), making this a brachyodont tooth (cf. Cooper 1934 : 575-581). No unworn M3 of Dicerorhinus leakeyi is available, but from the amount of crownward convergence of the parastyle and metastyle the relative height of the external surface does seem to be much the same as that in Aceratherium acutirostratum. Upper molars of Aceratherium from Rusinga comprise three specimens only: No. 231, 1950, R.2-4, Rusinga, anterior portion of M1 dext. adhering to P4 (PL 6, figs. 3, 4), No. 1630, 1950, Rusinga, upper M. dext., damaged on all sides but medisinus complete, and No. 515, 1951, Rusinga, upper M sin., ectoloph and most of metaloph missing. Among the rhino material from Napak, Karamoja, Uganda, sent to me by Dr. W. W. Bishop there is a specimen (no. 502, Napak I, 1958 (13)) showing the inner portion of the protoloph of a right upper molar with a deep protocone fold and an antecrochet all across the medisinus. This specimen represents the Aceratherium from Congo and Kenya. Of the upper molars from Rusinga no measurements can be given, but some of the premolars are smaller than those from Congo. All are, however, characterized by the antecrochet extending all across the medisinus. The inner cingulum is stronger in some than in others. This varies among the European species of Aceratherium also. In the British Museum (Natural History) there is a cast of a left upper dentition of Aceratherium lemanense from Auvergne (M. 29624). I have compared this with upper dentitions of Aceratherium incisivum in the same museum, viz., a cast of the skull figured by Kaup (1834, pi- IO> fig- 2) with the teeth well preserved (¥.2781), a left P2-M3 (Enniskillen collection, M.233), a left P2"4 (Hastings collection, M. 27464), and an isolated M3 dext. (Enniskillen collection, M.2370). All are from Eppelsheim. As shown in Table 10 the dentition of A. lemanense is similar in size to those of A. incisivum, but the postero-transverse diameter of M1 and M2 is equal to the antero- transverse diameter, and M3 is relatively larger. Other primitive characters shown in the A. lemanense dentition are the stronger inner cingulum and the absence of crista and crochet. In both, however, we find the strong antecrochet and protocone constriction typical of Aceratherium molars. The crista is variously developed in A. incisivum : there are traces in P3 and P4 in M.278i and M.233 > it is more deve- loped in M . 27464, even forming a medifossette with the crochet in P2. In the figured dentition (Kaup 1834 : pi. 14, fig. 5) there is a medifossette in P4 as well. The crochet is well developed in the molars, and in M . 2370 even forms a medifossette with the I4o MIOCENE RHINOCEROSES OF EAST AFRICA crista. The internal cingulum of the premolars either is confined to the medisinus entrance or extends forward and backward from this point. TABLE 10 Measurements of upper teeth of Aceratherium (mm.) A . lemanense A . incisivum M. 29624 M.278i M.233 M. 27464 P2, ant. post. 31 31 32 26 ant. transv. 40 42 39 35 post, transv. 42 45 43 P3, ant. post. 34 34 37 31 ant. transv. 52 54 53 47 post, transv. 50 53 51 44 P4, ant. post. 39 35 38 37 ant. transv. 56 59 52 51 post, transv. 53 56 49 46 M1, ant. post. 39 41 42 ant. transv. 55 60 54 post, transv. 55 57 49 p. tr. : a. tr. i-oo 0-95 0-91 M2, ant. post. 43 44 45 ant. transv. 58 63 58 post, transv. 58 60 p. tr. : a. tr. i-oo 0-94 M.237O M3, ant. post. 47 44 44 44 ant. transv. 58 59 51 55 1. outer surface 65 61 57 56 The Aceratherium dentition from Congo certainly resembles that of A. incisivum more than that of A. lemanense ; there is a well-developed crochet, and at least a trace of a crista in P4, the M1 and M2 are narrower behind than in front (ratios 0-90 and 0-89, respectively), and M3 is less wide than M2 (Hooijer 1963 : 43). However, not all of the A. lemanense dentitions are as primitive-looking as that mentioned above, and even the antecrochet is not constant. The latter is well developed across the medisinus in the premolars of the dentition from Cindre (Allier) figured by Roman (1912 : 59, pi. 8, fig. 20), but may be much reduced or absent as in the dentitions figured by Osborn (1900 : 242, text-fig. SB) and by Viret (1929 : 258, pi. 27, fig. 2). Again, the prominent antecrochet blocking the medisinus in the premolars is seen in " Teleoceras aquitanicum " Repelin (1917 : 37, pi. 5, figs. 7, 8), which is probably the same as A. lemanense (Wang 1928 : 207), as well as in the Burdigalian Aceratherium platyodon (Mermier 1896, pi. 2, fig. 4), but it is absent in the premolars of " Teleoceras aginense " Repelin (1917 : 12, pis. I, 3, 4), which, as pointed out by Lavocat (1951 : 114) is identical with Aceratherium lemanense. Also, the first and second molars of A . lemanense are not always as wide behind as in front as seen from the figures cited. In A . tetradactylum we find much variation in tooth size, development of antecrochet in the premolars, and the inner cingulum (Wang 1928 : 189). The teeth of this MIOCENE RHINOCEROSES OF EAST AFRICA 141 Vindobonian species are close to those of the Pontian A . incisivum with which it has often been united. The lower canines numbered 850, 1947 (R.i) (PI. 4, figs. 2, 3) belong to the same individual and differ from those of Dicerorhinus in being curved, with the convex sides facing each other. The diameters at crown base are 32 mm. horizontally and 24 mm. vertically. The cross-section is a transverse oval, and the diameters of the root are 28 by 23 mm. This pair agrees with Aceratherium in curvature and cross-section ; the root length is over n cm., more than in a specimen of A. tetradactylum from Georgensgmiind recorded by Wang (1928 : 189, text-fig. 2A) that has nearly the same diameters. In A. incisivum (Kaup 1834 : 52> pi- I4> fig- 9) the lower canines are much larger (length 27-30 cm., width 38-44 mm.), and very large lower C are also known in the Aquitanian A. lemanense (Repelin 1917 : 24, pi. 5, fig. 3). The flattening of the canine in A. acutirostratum is less marked than in A. platyodon (Mermier 1895, text-fig. 5) but more than in A. incisivum (Mermier 1895, text-figs. 6-8). The only premolars and molars of the lower jaw that may be referred to A ceratherium acutirostratum are those in the poorly-preserved mandible numbered 850, 1947, belong- ing to the skull remains mentioned above (PI. 9, fig. 4) . The teeth are much worn down, devoid of external cingula and with a sharp groove between meta- and hypolophid. They cannot be distinguished from those of Dicerorhinus leakeyi (Table 5) but P3, M2 and M3 are slightly longer. In Table n the Rusinga teeth may be compared for size with those of A. tetradactylum (no. i after Filhol 1891 : 204 (Sansan), nos. 2-4 after Wang 1928 : 190 (Georgensgmiind), and no. 5 after Rinnert 1956 : 33 (Viehhausen)). The lower teeth of A. incisivum recorded by Kaup (1834 : 53. pi- T4> fig- 9) are very large either. TABLE n Measurements of lower teeth of Aceratherium (mm.) A. acutirostratum Aceratherium tetradactylum no. 850 f •*- — ^ A . incisivum 12345 P,t ant. post. 36 31 33 30-5 30 32 30 ant. transv. 20 23 20 14 16 18 post, transv. 22 15 21 21 20-5 26 P4, ant. post. 36 36 33 35 34 33-5 ant. transv. 25 27 23 25 22 22 post, transv. 26 18 26 23 24 26 Mx, ant. post. 37 37 34 35 37.5 34 ant. transv. 26 27 20 23 22 24-8 post, transv. 29 27 26 25-6 25 M2, ant. post. 48 40 38? 42 37 40-5 34 ant. transv. 30 29 22 25 25 27 26 post, transv. 31 27 25 27-2 — M,, ant. post. 53 42 38 43 47 42 39 ant. transv. 30 27 19 25 22 28 25 post, transv. 29 20 26-5 24 28 Length P2-M3 c. 240 215 216 220 205 i42 MIOCENE RHINOCEROSES OF EAST AFRICA Of the milk dentition of Aceratherium acutirostratum there are two specimens, both DM4 : no. 142, 1949, Kamasengere, Rusinga, DM4 dext. (PI. 5, fig. 4), and no. 218, 1948, R.loy, Rusinga, DM4 sin (PI. 5, fig. 5). They differ from their homologue in the Dicerorhinus dentition from Rs.26 in having the protocone well defined. The anterior cingulum extends to the internal angle of the crown, the antecrochet is conspicuously developed, and the crochet is smaller. The parastyle is broken off in no. 218. These milk molars agree so closely with M1 of Aceratherium acutirostratum in their antecrochet and cingular development and in the degree of individualization of the protocone that they may safely be regarded as Aceratherium. They are also larger than the Dicerorhinus DM4 as will be seen by comparing Table 12 with Table 7. TABLE 12 Measurements of upper milk molars of Aceratherium acutirostratum (mm.) no. 142 no. 218 DM4, ant. post. 40 38 ant. transv. 47 48 post, transv. 45 44 p. tr. : a. tr. c-95 0-92 Genus BRACHYPOTHERIUMBorissiak 1927 Brachypotherium heinzelini Hooijer (PI. 4, fig. 6 ; PI. 6, figs. 5, 6, 9 ; PL 7, figs. 3, 4 ; PL 8 ; PL 9, fig. i ; PL 10, figs. 1-3, 6-8 ; PL 14, fig. 3) 1963 Brachypotherium heinzelini Hooijer : 45, pi. 6, figs. 4-6 ; pi. 8, fig. i. A right upper incisor, no. 79, 1950, R.i-ia (PL 4, fig. 6) lacks only a portion of the root. The crown diameters are 78 mm. anteroposteriorly and 24 mm. transversely. The root is hardly higher than the crown length. This specimen, much larger than the upper I referred to Dicerorhinus leakeyi (Table 2), agrees well with an upper incisor of Brachypotherium goldfussi1 (Kaup 1854 : 2, pi. i, fig. 13 : 81 by 26 mm.), which also has a short massive root. Large incisors referred to Aceratherium incisivum (Kaup 1834 : 51, pi. 14, figs. 1-4 ; 1854 : 9, pi. i, figs. 6-9) have the root higher than the crown length. If this differential character may be relied upon, the Rusinga incisor should be referred to Brachypotherium. Such large upper incisors have also been found at the type site of Brachypotherium heinzelini (Sinda no. 15) and another site, Lake Albert no. 446, from which a characteristic tooth of this species has been derived (Hooijer 1963 : 47, pi. 7, fig. 2). There are two sets of P3~4 from Rusinga that should be referred to the present 1 This Pontian Eppelsheim Brachypotherium is considered merely a large variety of B. brachypus (Lartet) of the Vindobonian by Dep6ret (1887 : 226). The upper incisor figured with the dentition of B. brachypus by Depe'ret (1887, pi. 23, fig. 3) has a slender and long root. It has been considered probably referable to Aceratherium by Rinnert (1956 : 38), and is placed with "Dicerorhinus" (= Aceratherium) simorrensis (Lartet) by Viret (1961 : 67). MIOCENE RHINOCEROSES OF EAST AFRICA 143 species, viz., no. 270, 1949, R-73, Rusinga, P3~4 dext. (PI. 8, figs. 5, 6), and no. 409, 1947, West side of Hiwegi, Rusinga, P3~4 sin. (PI. 8, figs. 3, 4). The Rusinga premolars of Brachypotherium agree with the type of B. heinzelini Hooijer (1963 : 45, pi. 6, figs. 4-6 ; pi. 8, fig. i) in the flattening of the ectoloph behind the paracone style (the metastyle of P4 in no. 270 is missing), in the develop- ment of the inner cingulum (weaker in no. 409 than in no. 270 in which latter it forms a ridge and not a mere tubercle), and in the slight development of the antecrochet, which does not block the medisinus. The external cingulum, however, so markedly developed in the holotype, is only weak in no. 409, and virtually absent in no. 270. Variations in the development of the external cingulum have been noted in B. brachypus also (Viret 1961 : 69). An isolated P4 sin. originating from Napak IIC, Karamoja, Uganda, collected and sent to me by Dr. W. W. Bishop in 1964, unmistakably belongs to Brachypotherium heinzelini. The specimen (PI. 8, fig. i) has the crown surface broken and distorted externally and behind the medisinus, but the protoloph is well preserved, showing the weakly developed antecrochet and the tubercle at the medisinus entrace in which it agrees perfectly with the type P4. The damage to the ectoloph notwithstanding it is clear that there was no external cingulum (well developed in the holotype, weak or absent in the Rusinga P4) . The basal diameters that can be exactly taken (Table 13) are intermediate between those of the two Rusinga P4. Neither in P3 nor in P4 of B. heinzelini the antecrochet is as strongly developed as in B. aurelianense (Nouel) of the Burdigalian (see Osborn 1900 : 250, text-fig, n ; Mayet 1908 : 100, text-fig. 29, pi. i, figs, i, 31). In B. brachypus of the later Vindo- bonian the antecrochet in P3~4 is reduced or wanting (Osborn 1900 : 25), although the upper dentition of B. brachypus from La Grive Saint- Alban figured by Deperet (1887, pi. 23) is only slightly worn and therefore does not show much of the ante- crochet, which becomes more marked toward the base as seen in the right upper molar figured (Deperet 1887, pi. 23, fig. 2). The external cingulum, absent in B. aurelianense, is either present or reduced in B. brachypus (Viret 1961 : 69). The criterion of the internal cingulum, present in B. brachypus and weak or absent in B. aurelianense, is not absolute (Viret 1961). As we shall see further on, the external cingulum is either present or absent in the upper molars of B. heinzelini. The first and second upper molars of Brachypotherium are rare in the East African Miocene. There is a much worn M1 dext. from Karungu, 1937 (PL 6, figs. 5, 6) that has the characteristic, if not constant, feature of the species, the external cingulum. The M2 sin. from the Sinda area, Lower Semliki, Congo, described and figured by me as Aceratherium cf. tetradactylum (Lartet) (Hooijer 1963 : 44, pi. 8, figs. 4-6) belongs here, too. It does lack the external cingulum, and has been identified thus because of its marked resemblance to certain large molars from Beaugency and Pontlevoy figured by Mayet (1908 : 96, 271, pi. 3, fig. 7 ; pi. 10, fig. 4) as Aceratherium aff. tetra- dactylum, and Aceratherium tetradactylum mut. pontileviensis , respectively. However at the time I overlooked the fact that Stehlin (1925 : 108) had referred these to 1 And also Mayet 1908 : 100, text-fig. 30 : Diceratkerium douvillei Osborn that would be based on female specimens of Brachypotherium aurelianense (Stehlin 1925 : 113). i44 MIOCENE RHINOCEROSES OF EAST AFRICA Brachypotherium brachypus. Stehlin notes that in their morphological characters these supposed Aceratherium molars do not differ from those of Brachypotherium brachypus, and further that their large size is no valid reason for referring them to Aceratherium tetradactylum, which is smaller than Brachypotherium brachypus. More- over, Stehlin notes that in the Pont-Levoy-Thenay deposits there is only one type of lower molar of corresponding size, belonging to Brachypotherium. Therefore, the M2 from Sinda no. 2 should have been referred to B. heinzelini. The Congo molar differs from those of Brachypotherium brachypus compared in the internal cingulum being weaker (manifested by a large tubercle at the medisinus entrance only), and the crochet being more developed. There is one specimen of the last upper molar in the Kenya collection that is referable to Brachypotherium heinzelini. This is the M3 dext. from Karungu described and figured but not specifically identified by Andrews (1914: 176-177, pi. 28, fig. 3). The specimen (M. 10632 ; PL 7, fig. 3) is somewhat larger than that of Aceratherium acutirostratum from Lake Albert, Congo (Hooijer 1963 : 43, pi. 7, figs. I, 3) in anterior width and the length of the outer surface (Table 13) but has the protocone only weakly constricted off, not by sharp grooves as in Aceratherium. It lacks the marked metacone bulge of Dicerorhinus from Rusinga in which M3 is smaller still. The cingulum forms a mere ridge at the medisinus entrance but is otherwise absent internally. The antecrochet is weak, and so is the crochet, while there is a trace of a crista. The upward convergence of parastyle and metastyle as seen from the TABLE 13 Measurements of upper teeth of Brachypotherium (mm.) P3, ant. post. ant. transv. post, transv. P4, ant. post. ant. transv. post, transv. M1, ant. post. ant. transv. post, transv. p. tr. : a. tr. M2, ant. post. ant. transv. post, transv. p. tr. : a. tr. M3, ant. post. ant. transv. 1. outer surface B. heinzelini r B. brachypus Sinda Rusinga Napak M. 33527 no. 270 no. 409 33 — 36 — — 51 — 54 — — 52 — 54 56 — 37 49 — 48 — 57 62 61 66 64 55 58 56 62 59 51 70 60 0-86 51 68 59 0-87 56 68 73 Karungu 63 77 63 0-82 62 52 O' 55 65 68 MIOCENE RHINOCEROSES OF EAST AFRICA 145 external side (PL 7, fig. 4) shows the brachyodonty or mesodonty of the crown. The M3 of Br achy pother ium brachypus (see Mayet 1909 : 25, text-figs, n, 12, both from the right side) are indistinguishable from the Karungu specimen but are smaller (antero-transverse diameters 53-59 mm.). The M3 dext. from Beaugency referred by Mayet (1908 : 98, pi. 3, fig. 6) to Aceratherium cf. tetradactylum, but which Stehlin (1925 : 108, footnote i) states should be referred to B. brachypus, is larger (antero- transverse diameter 68 mm.). The left upper tooth series of B. brachypus, the measurements of which are given in Table 13, is from Villefranche d'Astarac (Gers), M. 33527, and is made up of teeth of different individuals (the molar placed as M2 is more worn than that in the place of M1) but shows the characteristic features of the species. There are external as well as internal cingula in all premolars and molars ; the antecrochet is weak and the proto- cone constriction slight. The crochet is likewise weakly developed, and the crista is absent except for a trace in P4. In the European species, however, the external cingulum may be reduced (Viret 1961 : 69), and so it is in some of the specimens of B. heinzelini. In size the African teeth agree rather well with those of B. brachypus : the premolars are larger than those of B. brachypus presented in the Table, but Viret (1961 : 69) cites a P3 from La Grive 61 mm. wide, and P4 68-70 mm. wide, exceeding the large Rusinga specimen no. 409, whereas the Sinda M2 is larger, the Karungu M1 and M3 are however smaller than those in B. brachypus. The Sinda M2 is relatively narrower behind than that in the B. brachypus dentition, but the M1 of that dentition hardly differs from the Karungu M1 in this respect. The upper teeth of Brachypotherium are larger than those of Dicerorhinus leakeyi and Aceratherium acutirostratum, and differ further in being relatively narrower behind. Even when the external cingulum is absent the superior size and relatively smaller posterior width are distinctive of Brachypotherium heinzelini. The absence of a marked protocone constriction and of a large antecrochet differentiates Brachy- potherium heinzelini molars from those of Aceratherium acutirostratum, while the flattened ectolophs, and the absence of the metacone bulge in M3 serve to distinguish Brachypotherium heinzelini molars from those of Dicerorhinus leakeyi. Although both Osborn (1900 : 250) and Mayet (1908 : 107) state that the external surface of the lower molars of the Burdigalian B. aurelianense is flattened, in actual fact this characteristic applies only to the more advanced brachypotheres. Most of the lower molars of B. aurelianense have a marked groove on the external surface between metalophid and hypolophid (Stehlin 1925 : no, in). In the Upper Burdi- galian and Lower Vindobonian B. stehlini Viret (1961 : 71), which is generally larger than B. aurelianense, the lower molars have the external groove between meta- and hypolophid almost completely flattened out, and moreover almost invariably present an external cingulum (cf. Roman & Viret 1934 : 33, pi. 10, figs. 7, 8). The same characters are found in typical B. brachypus of the Upper Vindobonian (Viret 1961 : 72) in which the tendency toward hypsodonty is more marked, as well as in the terminal B. goldfussi (Kaup 1834 : 63, pi. 12, figs. 13, 14). Few lower molars are present in the Rusinga, Karungu, and Napak collections in which the external groove is very shallow or nearly flattened out, and in none of them is 146 MIOCENE RHINOCEROSES OF EAST AFRICA there an external cingulum. These specimens doubtless belong to Brachypotherium heinzdini, but whether they are the only ones is not known ; some of the grooved specimens may also belong here if this character is not constant in the African species, as happens in B. aurdianense. One of the specimens definitely belonging to Brachy- potherium heinzelini, that from Napak II A, 1964, is rather large, exceeding the Dicerorhinus leakeyi lowers in size ; the others are not particularly large. The specimens are as follows : No. 546, 1949, R.I, Rusinga, posterior portion of Mx or M2 sin. (PI. 6, fig. 9), No. 345, 1950, Ngira, Karungu, posterior portion of M.^ or M2 dext., Napak II A, 1964 (kindly forwarded by Dr. W. W. Bishop), M2 sin., nearly entire (pi. 8, fig. 2), and M. 25186, R-7, Rusinga, M3 sin., incomplete in front. TABLE 14 Measurements of lower molars of Brachypotherium heinzelini (mm.) Rusinga Karungu no. 546 no. 345 Mx or M2, ant. post, ant. trans v. post, trans v. 29 31 Napak II A M2, ant. post. 56 ant. trans v. 37 post, transv. 35 Rusinga M. 25186 M3, ant. post. ant. transv. 31 post, transv. 29 No upper milk molars of Brachypotherium appear to be present in the East African Miocene collections. The DM4 of Brachypotherium cf. brachypus from Chevilly figured as an M2 of Diceratherium douveillei by Mayet (1908, pi. 3, fig. 5 ; see Stehlin 1925 : 114, footnote) has a stronger cingulum and the protocone less well marked off than the Aceratherium DM4 of Rusinga. The post-cranial skeleton of Brachypotherium is easily distinguished from that of Dicerorhinus and Aceratherium by the shortness and great relative width of the individual bones. All of the metacarpals are represented in the Rusigna collection : M. 18813 and M. 18812, Rs.6a, Rusinga, associated Me. Ill and Me. IV dext. (PI. 10, figs, i, 2). Part of the shaft of the third metacarpal below the magnum facet medially is missing, F.3269, R. VII. 1941, Rusinga, Me. II dext. (PI. 10, fig. 3), and M. 18822, Rs.7, no. 451, 1947, Kathwanga, Rusinga, Me. IV sin. (PI. 10, fig. 8). The median metacarpal is as much shortened as that in the Vindobonian Brachy- potherium brachypus, whereas the second and fourth metacarpals from Rusinga are MIOCENE RHINOCEROSES OF EAST AFRICA 147 even relatively wider, and more shortened than their homologues in B. brachypus, as will be seen by comparing Table 15 with Table 16 (after Roger 1900 : 22). TABLE 15 Measurements of metacarpals of Brachypotherium heinzelini (mm.) Me. II Me. Ill Me. IV Me. IV F.326Q M. 18813 M. 18812 M. 18822 Median length 125 137 no 113 Proximal width 52 65 45 42 Proximal ant. post, diameter 47 51 53 46 Middle width 45 53 37 38 Middle ant. post, diameter 25 25 24 23 Greatest distal width 50 73 52 52 Width of distal trochlea 40 59 44 47 Distal ant. post, diameter 38 47 39 43 TABLE 16 Measurements of metacarpals of Brachypotherium brachypus (mm.) Me. II Me. Ill Me. IV Median length 100-132 130-147 116-130 Middle width 38-40 5°-55 37~39 In Brachypotherium stehlini of the Upper Burdigalian of La Romieu the metapodials are longer than those in B. brachypus : Me . II has a median length of 160 mm. by a middle width of 40 mm., and Me. Ill has a median length of 170 mm., a proximal width of 70 mm. and the least width of shaft 53 mm. (Roman & Viret 1934 : 33, text- fig. 14, as B. cf. brachypus ; cf. Viret 1961 : 71). Thus, as already observed by Stehlin (1925 : 138), the Brachypotherium of the Burdigalian has the metapodials less short- ened than that of the Vindobonian of La Grive, etc. Of the Brachypotherium from the Miocene of Moghara in Egypt, described as Teleoceras snowi Fourtau, no metacarpal, but a left metatarsal III is available. This bone (Fourtau 1920 : 46, text-fig. 30) is longer than that in B. brachypus, and in fact is as long as that in Brachypotherium stehlini (Roman & Viret 1934 : 33, text-fig. 13 A). The ranges of size in B. brachypus given in Table 17 are after Roger (1900 : 26) and Viret (1961 : 70). TABLE 17 Measurements of metatarsal III of Brachypotherium (mm.) B. snowi B. stehlini B. brachypus Median length 151 150 110-137 105-112 Middle width c. 50 55 45-5° 47~5X Greatest distal width 67 72 up to 70 67 148 MIOCENE RHINOCEROSES OF EAST AFRICA It follows from this that the Rusinga B. heinzelini is more progressive than that of Moghara in Egypt in having more shortened metapodials. Among the isolated postcranial material from Rusinga there is one entire right radius, M . 18908, Rs.6a, shorter and wider than the other radii recorded elsewhere in the present paper (PI. 9, fig. i). This bone very nearly agrees in dimensions with one of the Steinheim radii recorded by Roger (1900 : 17, table, no. 5) as Brachypotherium. The radius of Brachypotherium stehlini from La Romieu (Roman & Viret 1934 : 35, pi. n, fig. i) has the same proximal and distal widths but is much longer, as seen in Table 18. TABLE 18 Measurements of radius in Brachypotherium (mm.) B. heinzelini B. br achy pus B. stehlini Median length Proximal width Middle width Greatest distal width Rusinga 293 95 52 95 Steinheim 300 98 La Romieu 357 99 101 There are some twenty astragali in the East African Miocene collections, and of these there is one, from the left side, no. 538, Gumba, Rusinga, 1949 (PI. 14, fig. 3) in which the total width greatly exceeds the medial height, nearly to the same extent as in the astragalus from the type site of Brachypotherium heinzelini (Hooijer 1963 : 47, pi. 5, fig. 10 ; pi. 8, fig. 7), and in B. brachypus (Table 19). In this Table, the La Grive specimen is after Deperet (1887 : 225, pi. 24, fig. 4), the Stein- heim specimen after Roger (1900 : 24), M. 33529 is from Villefranche d'Astarac, and M . 7760 is a cast from Thenay (Loir-et-Cher) . The Steinheim astragalus figured by Fraas (1870, pi. 7, fig. 6) is clearly Brachypotherium as already surmised by Deperet (1887 : 225), and so is the calcaneum of fig. 9 of Fraas's plate ; both are identified as Rh. incisivus by Fraas (1870 : 302). TABLE 19 Measurements of astragalus of Brachypotherium (mm.) B. heinzelini B. brachypus Lateral height Medial height Total width Ratio medial height/total width Trochlea width Width distal facets Y ^ Sinda Rusinga La Grive Steinheim M. 33529 M-776o — 57 — 73 — — 68 60 60 68 64 64 IO2 82 93 96 IOO 98 0-67 °'73 0-64 0-71 0-64 0-65 68 85 85 86 85 80 85 The first phalanx of a median digit, F.2I26, Rusinga, 1941 (PI. 10, fig. 6), is shorter than any of its homologues in the East African Miocene collections, and more- MIOCENE RHINOCEROSES OF EAST AFRICA 149 over has a proximal width that represents the maximum in our series. The distal articular surface is evenly concave transversely, and flat anteroposteriorly instead of weakly convex as in the longer first phalanges. Its measurements are presented in Table 20 ; the data on Brachypotherium brachypus and on Aceratherium are after Roger (1900 : 27). TABLE 20 Measurements of phalanx I, median digit (mm.) B. heinzelini B. brachypus Aceratherium Length 28 28-30 35-45 Proximal width 55 53~55 45~57 The second phalanx of a median digit of Brachypotherium heinzelini is represented by no less than three specimens: F.2I25, Rusinga, 1941; M. 18854, RS.IOI, Rusinga, and M. 18862, Rs.6a, Rusinga (PI. 10, fig. 7). These bones are shorter and wider than their homologues from Rusinga, and their proximal articular surface is evenly convex transversely and flat anteroposteriorly, not raised in the middle in front and behind as in the longer second phalanges. Measurements of B. brachypus and Aceratherium in Table 21 again after Roger (1900 : 27). TABLE 21 Measurements of phalanx II, median digit (mm.) B. heinzelini F.2I25 M. 18854 M. 18862 B. brachypus Aceratherium Length 22 20 20 20 20-35 Proximal width 58 61 59 60 40-65 The phalanges from Viehhausen provisionally referred to Brachypotherium cf. brachypus by Rinnert (1956 : 37) appear to me to belong to either Aceratherium or Dicer orhinus. The first phalanx of a lateral digit, M . 18859, Rusinga, is relatively shorter and wider than the others, and agrees with F. 2 126 in the shape of the distal articular surface. It may therefore be referred to Brachypotherium heinzelini. The data on B. brachypus and Aceratherium given by Roger (1900 : 27) are few and apparently questionable ; I have added in Table 22 the measurements of some of the first phalanges of lateral digits from Rusinga, including those of the skeleton no. 2 from R.I, Rusinga, 1947. TABLE 22 Measurements of phalanx I, lateral digit (mm.) B. heinzelini B. brachypus Aceratherium Rusinga Length 28 (20-25)? 25-35 37 33 29 28 Proximal width 43 (5°)? 45 40 38 35 40 GEOL. 13, 2. 10 i5o MIOCENE RHINOCEROSES OF EAST AFRICA There remains one second phalanx of a lateral digit, no. 1055, S. E. of Kiahera Hill, Rusinga, that is relatively wider than the others, and its proximal surface is evenly convex transversely as in Brachypotherium. The data on B. brachypus and Aceratherium in Table 23 are as given by Roger (1900 : 27); the data on some of the Rusinga second phalanges of lateral digits are added, including those of no. 2, R. i, 1947- TABLE 23 Measurements of phalanx II, lateral digit (mm.) Rusinga B. heinzelini B. brachypus Aceratherium ( -*• ^ Length 22 (10-15)? 20-30 25 27 25 22 21 Proximal width 42 (47-50)? 45 39 37 35 37 27 This completes the account of the postcranial material from Rusinga referable to Brachypotherium heinzelini. The present species, first described from the Miocene of the Sinda-Mohari region, Lower Semliki, Congo (Hooijer 1963 : 45) on a P4, the external cingulum of which has now been shown to be a variable feature, is more advanced than the European Brachypotherium brachypus which it resembles dentally, in the lateral metacarpals being more shortened and relatively wider. Genus CHILOTHEfl/l7MRingstr6m 1924 Chilotherium sp. (PI. 6, figs. 10, ii ; PL 7, figs, i, 2) An M3 dext. originating from Loperot 1948 (PI. 7, figs, i, 2) represents a genus not before recorded from the African Tertiary. It is well preserved, lacking only a small portion of the internal cingulum at the metaloph, and it is remarkable for two main features, viz., the great height of the crown, and the very weak development of the parastyle fold and the paracone style, characters not seen in Dicerorhinus , Acera- therium, or Brachypotherium. The external surface of the M3 (the joined ectoloph and metaloph) is peculiarly flattened and has the sides (parastyle and metastyle) much less converging crown- ward than in the other genera mentioned : the metastyle is slightly concave basally but becomes very nearly vertical at a level only 15 mm. from the base of the crown, while the parastyle is very steep also. The full basal length of the external surface is 6l mm., which length is reduced to 53 mm. at a height of 15 mm. from the enamel base, and still amounts to 46 mm. at 45 mm. from the base, at which level the crown is worn. Over this vertical distance the thickness of the ectometaloph has hardly reduced, and it is clear that the crown has already undergone a considerable amount of wear. The full height of the unworn crown may well have been some 25 to 35 mm. more than that to which it is worn down, that is, about 70-80 mm. The parastyle fold is hardly visible, and the paracone style is weakly developed in the upper portion MIOCENE RHINOCEROSES OF EAST AFRICA 151 of the crown, flattening out in the basal two-thirds. Near the base, there is a meta- cone bulge that is, however, placed nearer to the inner end of the external surface than in a Dicerorhinus molar. The protocone of the Loperot M3 is very markedly denned and flattened internally ; the anterior and posterior grooves delimiting it become sharper and deeper toward the base, where the posterior groove curves inward to end at the medisinus entrance. The antecrochet becomes very prominent basally, reducing the entrance to the medisinus to a narrow V, while the crochet, which nearly extends all across the medisinus apically, recedes towards the base. There is a weak crista, a projection from the ectoloph opposite the paracone, not yet touched by wear, not reaching the crochet, and like it receding basally. The cingulum is well developed along the anterior surface, absent along the inner surface of the protoloph, forming a cusp entering the medisinus, and continuing along the metaloph, rising to a point 15 mm. high at the metacone bulge posteriorly, beyond which it falls off steeply. There is no cingulum along the external surface. The great hypsodonty of the Loperot M3 and the flattening of the external surface as well as the very marked protocone constriction, prominent antecrochet, and metacone bulge near the internal angle, are all characters pointing to its belonging to the genus Chilotherium Ringstrom (1924 : 26). This remarkable genus of extinct rhinoceroses with its orbit near the upper surface of the skull, the exaggerated symphysis width and huge, widely separated canines had often been regarded as Aceratherium until Ringstrom (1924) recognized its true character and gave it separate generic status. The specimen from Loperot described above is exceedingly similar in shape and size to the M3 of Chilotherium anderssoni Ringstrom (1924 : 34, 35. pl- 3> ngs- 3. 4) from the Pontian of Shansi, North China. The length of the external surface of the M3 of C. anderssoni is 58-60 mm., the anterior width 53-59 mm., and the full height of the crown, presumably the same as in M2, about 85 mm. In C. anderssoni the metacone bulge is as in the Loperot M3 ; in C. habereri (Schlosser) there is no bend at the junction of ecto- and metaloph, the external surface being evenly rounded throughout, and the posterior cingulum is more developed, free at its apex (Ringstrom 1924 : 41, pl. 4, fig. 3). C. habereri as well as the other species of Chilotherium described from Shansi and Shensi by Ringstrom are smaller than C. anderssoni. In the M3 of C. wimani Ringstrom (1924, pl. 7, fig. 3) there is a slender crista uniting with the crochet ; the internal cingulum is rather variable in the molars but invariably developed at the medisinus entrance. Table 24 gives the measurements of the Loperot M3 in conjunction with those of Dicerorhinus leakeyi, Aceratherium acutirostratum, and Brachypotherium heinzelini. While in size the Chilotherium M3 exceeds its homologue in Dicerorhinus, it is near to the M3 of Aceratherium and Brachypotherium in length, but less wide anteriorly and with the external surface shorter. In the Rusinga collection there are two incomplete right upper molars that appear to belong to Chilotherium as well, viz., No. 695, 1949, Gumba, Rusinga, lacking outer portion (Pl. 6, fig. 10), and No. 506, 1950, Wakondu, Rusinga, protoloph only (Pl. 6, fig. n). Both of these show the very markedly constricted and internally flattened proto- i52 MIOCENE RHINOCEROSES OF EAST AFRICA TABLE 24 Measurements of M3 in various genera (mm.) Dicerorhinus A ceratherium Brachypotherium Chilotherium sp. leakeyi acutivo stratum heinzelini Loperot Rusinga Lake Albert Karungu ant. post. 56 42-46 57 55 ant. transv. 60 50-54 62 65 1. outer surface 61 52~57 65 68 cone, delimited by grooves the posterior of which curves inward basally, the prominent antecrochet, the strongly developed anterior cingulum, and the cingular cusp enter- ing the medisinus : the characters of the Loperot specimen. In the more complete specimen (no. 695) the medisinus is not very well preserved, the molar having been assembled from fragments, but from what is left of the crochet it is evident that it is very prominent, extending forward and outward beyond the antecrochet, as in Chilotherium. Moreover, the protoloph (no. 506) is less worn than that in the Loperot M3, the internal height of the protocone is 34 mm. Seen from the internal side its crownward taper is less marked than in either A ceratherium or Dicerorhinus molars, pointing to a high crown. There does not appear to be any postcranial skeletal material of Chilotherium in the East African Miocene collections available to me at present ; Ringstrom (1924) has established that Chilotherium is a brachypothere rhinoceros with the limb and foot bones even more shortened than in Brachypotherium (although the metapodials are not so broad at the middle : Ringstrom 1924 : 58, cf. Table 16), and the excellent Rusinga bones here referred to Brachypotherium heinzelini. To the genus Chilotherium Ringstrom refers forms from the Pontian of Samos and of Maragha, Iran, formerly placed in Aceratherium (Ringstrom 1924 : 83-89), and the genus occurs also in the Vindobonian of Portugal and Spain (Villalta & Crusafont 1955). The earliest occurrence of the genus in Eurasia is in the Burdigalian Bugti Beds of Baluchistan (whence it was originally described as a hippopotamus : Cooper 1934 : 595-596). If the Rusinga and Loperot faunas in which we now have the first evidence of the presence of Chilotherium in Africa would be Burdigalian, the appear- ance of the genus is as early as that in Baluchistan, and earlier than that in Europe. POSTCRANIAL SKELETON OF Dicerorhinus AND Aceratherium There is an abundance of postcranial skeletal material in the East African Miocene collections referable to long-limbed rhinoceroses, including much associated or supposedly associated material. Unfortunately, the extensive literature notwith- standing, our knowledge of the skeleton of the Tertiary rhinoceroses of Europe is far from satisfactory, and in particular the distinction between the limb and foot bones of Dicerorhinus and Aceratherium is well-nigh impossible. In European collections in which these two genera are represented by cranial or dental material, the identifica- MIOCENE RHINOCEROSES OF EAST AFRICA 153 tion of unassociated postcranial material is made mainly on the ground of size, a very uncertain guide. At Steinheim, where both Dicerorhinus germanicus Wang and Aceratherium tetradactylum (Lartet) (as well as the easily recognizable Brachypo- therium br achy pus (Lartet)) occur, the larger bones have usually been ascribed to Aceratherium, the smaller to Dicerorhinus, in accord with the size of the teeth in the two slender-limbed forms (Fraas 1870 : 203 ; Roger 1900 : 41 ; Wang 1928 : 203) . There is a difference, however slight, between Dicerorhinus and Aceratherium, for A ceratherium is tetradactyl, having a fair-sized metacarpal V, whereas in Dicerorhinus the fifth metacarpal is reduced. In the living form of Dicerorhinus, D. sumatrensis (Fischer), metacarpal V is represented by a rudiment, a sesamoid-like bone that, however, has distinct facets for both metacarpal IV and the unciform. Such rudi- mentary fifth metacarpals have also been found in the Vindobonian Dicerorhinus caucasicus Borissiak (1938 : 38) and in Pleistocene D. hundsheimensis (Toula 1902 : 47, pi. 8, figs, i, 6) ; in both cases the rudiment articulates with unciform as well as with metacarpal IV (Toula describes the fifth metacarpal as " das aussere Sesam- knochelchen "). These mammiform bones are hardly longer than their proximal width. On the other hand, in Aceratherium tetradactylum metacarpal V is about 80 mm. long, and has a relatively well-developed distal articulation carrying at least two phalanges, although it is narrow proximally and does not bear proximal facets more extensive than those in Dicerorhinus. The fifth metacarpal appears to be developed in all aceratheres, such as Plesiaceratherium gracile Young (1937) from Shantung, China, Aceratherium lemanense (see Duvernoy 1853, pi. 7, fig. 2, as A. gannatense ; Repelin 1917, pi. 12, as " Teleoceras aginense " (Lavocat 1951 : 114), and Roman 1924 : 51-52, text-figs. 23, 24), Aceratherium tetradactylum (Duvernoy 1853, pi. 7, fig. i ; Stehlin 1925 : 132, 139), and the terminal Aceratherium incisivum (Kaup 1834 : 58, pi. 15, fig. 4 ; 1854, pi. 9 ; 1859). Stehlin (1917) notes that the Acera- therium metacarpal V from Sansan (A. tetradactylum} is weaker than that in the earlier A. lemanense. Professor H. Tobien (private communication) notes that the two skeletons of Aceratherium incisivum of the Pontian of Howenegg (see Tobien 1956) have a metacarpal V in much the same degree of development as in the Aquitanian Aceratherium lemanense as figured by Roman (1924 : 51, text-fig. 23). Detailed studies of the Howenegg skeletons will yield important data that are, however, not yet available at the time of writing. Thus, in the absence of an associated metacarpal V, the distinction between the hands of Dicerorhinus and Aceratherium cannot be made. This does not mean that no attempts have been made to establish distinguishing characters in the postcranial skeleton of the two genera, however, but these do not appear to me to have been successful. They will be referred to as we deal with the skeleton in the pages that follow. The median metacarpals and metatarsals, considered of the greatest value in rhinoceros taxonomy, are remarkably alike in two (cranially and dentally) well-known species such as Dicerorhinus sansaniensis and A ceratherium tetradactylum of the Vindo- bonian of Europe, as is evident from the measurements proffered by Pavlow (1892 : 212) as follows (Table 25) : i54 MIOCENE RHINOCEROSES OF EAST AFRICA TABLE 25 Measurements of median metapodials of Dicerorhinus and Aceratherium (mm.) Me. Ill Mt.III length width length width Dicerorhinus sansaniensis 170 43 150 40 Aceratherium tetradactylum 165 42 153 40 There are no significant proportional differences between these bones. It is futile to insist on the value of such minor differences. The great caution exercised by Stehlin (1925 : 125-139) and Arambourg (1933 : ri) in identifying isolated bones of non-brachypothere rhinoceroses is exemplary, and has, I hope, saved me from draw- ing conclusions not warranted by the evidence. Among the associated postcranial material of rhinoceroses from the Miocene of East Africa there is first of all that of two skeletons found in 1947 at R.I, Rusinga. Of these, Maclnnes (1951 : 2) writes as follows: " The only two complete skeletons yet recovered from the Rusinga deposits have been those of Rhinocerotids. One of these had been exposed on the surface for some considerable time before discovery, and most of the bones were almost weathered away. In the second example, however, the skull, vertebrae, ribs and limbs of the lower side were almost wholly intact and articulated, whilst in the upper half the limbs had been torn off and discarded, though still almost completely articulated, within a few feet of the body. The ribs of the upper side had been forced forwards and backwards from a central point, indicating that the scavengers had penetrated the softer parts of the belly, but had apparently been unable to do any appreciable damage to bones of such bulk. " In her book An outline of the geology of Kenya Mrs. Sonia Cole published a photo- graph of the two skeletons in situ (Cole 1950, pi. i), showing one skeleton lying on its side with about twelve ribs arranged neatly in their natural position. This is the most conspicuous feature shown in the picture ; the limb and foot bones lie scattered around although some are in articulation, such as a tibio-fibula and a foot near the centre foreground. The skull (or what remained of it) is not shown in the illustration, and apparently had already been removed before the picture was taken. A scapula with the entire spine and tuber spinae is isolated in the right foreground. The specimen in the picture is from the left side, but in the collection it is from the right side, so that the photograph may have been printed back to front. Whether the thorax was lying with its right side up or with the left side up has not been recorded by Maclnnes, but in the published picture the twelve or so ribs, which are those of the lower side (we are looking into the emptied thoracic cavity), are from the left side (the vertebral extremities of the ribs are to the right, and the short and wide first rib is foremost in the picture) ; thus, the right side would have been up before removal from the deposit. In the collection the best-preserved ribs of this skeleton are from the right side, those from the left side preserved being fewer in number and rather fragmentary except for the first and last. This is evidence for the distorted side of the thorax (due to penetration by scavengers) having been the left side, which coro- MIOCENE RHINOCEROSES OF EAST AFRICA 155 borates the conclusion drawn from the scapula that the published photograph has been reversed. The " de-reversed " picture is given on PL 3 of the present paper. The more complete skeleton of the two from R.I, Rusinga, 1947, belongs to the upper dentition and mandible of Dicerorhinus leakeyi bearing no. 2, R.r, 1947, although there are a great many bones that are not numbered (the catalogue of the collection sent from the National Museum indicates that all specimens not otherwise marked are thought to belong to skeleton no. 2, R.I, 1947). In addition, we have the skull remains and mandible marked no. 850 (R.i), 1947, that represent Acera- therium acutirostmtum. The skeleton belonging to this skull is the one of which Maclnnes writes that it had been exposed on the surface for some considerable time, with most of the bones almost weathered away. Some of the bones of this Acera- therium skeleton, viz., the scapula sin., the atlas no. 717, and the left radius and ulna no. 850, are in exactly the same state of preservation as those of the no. 2 Dicerorhinus skeleton. These bones are fragmented and distorted, the cracks filled with matrix (or plaster occasionally) , whitish in colour and evidently preserved in the broken state in which they had been found. There are a number of bones in a different state of preservation, all much deformed and treated with shellac, which gives a brown staining. These bones, thirteen in all, are invariably from the right side (humerus, radius, ulna, scaphoid, lunar, cuneiform, pisiform, femur, patella, tibia and fibula, astragalus, and calcaneum). It is hardly possible to take measurements of these deformed bones, but the radio-ulna is of the same size as the left numbered 850, slightly smaller than that in D. leakeyi, and the shellaced right tibia is also slightly shorter than the homologous bone in the no. 2 skeleton of Dicerorhinus. These brownish bones I have no doubt should be regarded as belonging to the no. 850 skeleton of Aceratherium acutirostratum ; there is no duplication anywhere and the radio-ulnae from both sides agree in size (length). Thus, of the same skeleton we have a few untreated bones from the left side (scapula, radio-ulna) and an atlas, and a great many right bones that have been shellaced. The picture emerging from these considerations is that of two adult rhinoceros bodies, one (Dicerorhinus leakeyi) lying on its right side, the left side of the thorax ripped up by scavengers that tore off some limbs. Of the skull and mandible only the right halves are preserved, the left halves gone. In general, the bones from the right side, more deeply embedded in the sediment, are more complete than the left, although all the larger bones are cracked, distorted and deformed in the course of the fossilization process. The bones of the Dicerorhinus skeleton, excluding the smaller elements, are laterally flattened as a result of vertical compression in the fossil deposit. The other body skeleton, slightly smaller and belonging to Aceratherium acutirostratum, must have been lying on its left side. The thirteen bones of the right fore and hind limb, still in articulation, had almost weathered away and had to be treated with shellac. The better protected left side of this skeleton, a few bones of which have been saved, could be left untreated. Here, then, we have two skeletons of different genera, Dicerorhinus and Acera- therium which would have provided an excellent and rare opportunity to study the 156 intergeneric postcranial skeletal differences, if any, but the state of preservation of the Aceratherium bones and many of the Dicerorhinus bones is so poor that they are of no use for detailed morphological comparison. We can only state that the Aceratherium individual was slightly smaller than the Dicerorhinus individual. To avoid needless repetition of tables and to facilitate comparison I decided not to describe the Dicerorhinus leakeyi skeleton of no. 2 separately under its head, and the Aceratherium acutirostratum bones of no. 850 apart (nothing of value can be derived from the latter anyway) . Instead, I shall deal with all of the postcranial material of non-brachypothere rhinoceroses, including many specimens of uncertain generic position, in the present work bone for bone. But let us first consider the skeleton as a whole. The skeleton of Dicerorhinus leakeyi from Rusinga, relatively complete as it is, provides a welcome opportunity to compare the lengths of the limb segments in one and the same individual with those in the Recent Sumatran species and in Tertiary species from Europe of which an associated skeleton is known. In the literature we find metrical data on the skeleton of the oldest Dicerorhinus species known, viz., the small D. tagicus Roman (1924) from the Aquitanian of Budenheim, as well as that of D. caucasicus Borissiak (1938) from the Vindobonian Chokrak beds in North Caucasus, the former older, the latter younger than D. leakeyi. The data are not so complete as may be desired. The skeleton of D. leakeyi lacks the metacarpals, and for the length of this segment I have substituted that of an Me. IV (M. 18814), likewise from R.I, to be dealt with later. Although this bone may be Aceratherium, it fits well with the bones of D. leakeyi. As this is the only complete metacarpal of a slender-limbed rhinoceros in the collection available from these beds I have used Me . IV of the other skeletons as well. The radius of the D. leakeyi skeleton is incomplete ; for the length of this limb segment I have used the length of the ulna from the processus anconaeus. The skeleton of D. tagicus does not provide more than approximate lengths of ulna and Me. IV, but those of the remaining limb segments are accurately recorded. Of the skeleton of D. caucasicus the lengths of ulna and Mt.III are not known as these bones are incomplete, and the required lengths of humerus and tibia TABLE 26 Limb segment lengths and ratios in Dicerorhinus (mm.) D. tagicus D. leakeyi D. caucasicus D. sumatrensis Length of humerus (greatest) 235 450 c. 400 370 Length of ulna (from beak) c. 200 400 345 Length of metacarpal IV c. 100 150 137 130 Length of femur (greatest) 305 545 450 423 Length of tibia (greatest) 250 420 c. 375 313 Length of metatarsal III 122 180 144 Humero-femoral ratio °'77 0-83 c. 0-89 0-87 Ulno-humeral ratio c. 0-85 0-89 o-93 Metacarpo-humeral ratio c. 0-43 °'33 c- °'34 °'35 Tibio-femoral ratio 0-82 o • 77 c. o • 83 o • 74 Metatarso-femoral ratio 0-40 °'33 °'34 MIOCENE RHINOCEROSES OF EAST AFRICA 157 are only approximate. However, the available data lead to some interesting con- clusions as to the adaptations to speed and weight in the fossil species. The measure- ments of D. sumatrensis are those of an adult male skeleton from Sumatra in the Leiden Museum (cat. ost. g). The relative lengths of the limb segments shown by the length ratios in Table 26 differ to a greater or less degree in the four skeletons compared. In the skeleton of D. leakeyi the hind limb is less elongated relative to the fore limb than in D. suma- trensis ; the ulna is shorter relative to the humerus, but the tibia longer relative to the femur in the Miocene than in the Recent form. The metapodials of both limbs are very nearly equal in length relative to the proximal limb segments in the two skeletons. In D. tagicus, the oldest (Aquitanian) species, the hind limb is still less elongated relative to the fore limb than in D. leakeyi, the ulna again shorter relative to the humerus, the tibia still longer relative to the femur ; in all these points the Oligo- cene skeleton differs more from the Recent than does the Miocene skeleton from Rusinga. But the metapodials, especially of the fore foot, are markedly more elongated relative to the proximal limb segments in D. tagicus than in either D. leakeyi or D. sumatrensis, which differ little in this respect. The less completely preserved skeleton of D. caucasicus shows that the hind limb was probably longer relative to the fore limb than in D. sumatrensis ; the metacarpo- humeral ratio is approximately the same as in D. leakeyi and D. sumatrensis, while the tibio-femoral ratio is about as in D. tagicus. When years ago I compared the Pleistocene skeleton of Rhinoceros sondaicus Desmarest from Java with the Recent skeleton of the Javan rhinoceros I found that the Pleistocene skeleton had radius, tibia, and metapodials longer relative to humerus and femur than the Recent, which I interpreted as evidence of the Pleistocene Rh. sondaicus having been a more swiftly-moving type, able to make greater speed, than the Recent. The Pleistocene Rh. sondaicus is almost identical in limb segment ratios to Recent Dicerorhinus sumatrensis, whereas the Recent Rh. sondaicus has the skeletal proportions of the Recent Indian rhinoceros, Rh. unicornis L. (Hooijer 19466). The intraspecific differences in limb segment ratios found in Rh. sondaicus are of the same order of magnitude as those now found between Dicerorhinus leakeyi and D. sumatrensis. The Aquitanian species D. tagicus, with its long manus and pes, represents a definitely more swiftly-moving type than the Miocene and the Recent species. It is interesting to note, however, that D. tagicus appears to be fully tridactyl ; no facet for Mc.V was found on Me. IV by Roman (1924 : 30). D. caucasicus, and probably D. leakeyi as well, have a facet for a fifth metacarpal on their Me. IV, as has the living D. sumatrensis. D. tagicus is the smallest Dicerorhinus known, and has been placed at the beginning of the Dicerorhinus " line "(it was unknown to Osborn (1900), who began the line with D. sansaniensis of the Vindobonian) . The progres- sion into a more slow-moving type along the " line " is shown by the shortened meta- podials in the later species. In the relative length of the tibia D. caucasicus, although geologically later than D. leakeyi, still resembles D. tagicus ; the long humerus of i58 MIOCENE RHINOCEROSES OF EAST AFRICA D. Caucasians appears to exceed even that in the living species, however. It is of course probably fortuitous that D. leakeyi should have various limb segment ratios intermediate between those of D. tagicus and D. sumatrensis, and this gradation should not be regarded as evidence for direct phyletic relationship. However limited the value of the above observations may be, I decided not to leave them out as even the slightest information we can derive from a palaeontological find such as the Rusinga skeleton may be needed later when more comparable Tertiary skeletons in Africa or Europe are discovered. An atlas (no. 717, Rusinga, 1947, thought to belong to no. 850, R.I, Rusinga, the Aceratherium acutostratum skeleton) has incomplete wings, and is crushed from above downwards. The dorsal arch shows the two intervertebral foramina the distance between which is 82 mm. (83-5 mm. in the atlas of Dicerorhinus primaevus (Aram- bourg (1959 : 64) ; 51 mm. in D. sumatrensis, Leiden Museum, cat. ost. g). The anterior articular cavities for the occipital condyles are relatively well-preserved, and the width across them is 148 mm. (145 mm. in D. primaevus, 139 mm. in D. schleier- macheri (Kaup 1834 : 41), and III mm. in D. sumatrensis). The posterior articular surfaces are damaged, but the median ventral tubercle is entire, 15 mm. long and wide, and 20 mm. high. This tubercle, present in the atlas of D. schleiermacheri (Kaup 1834, pl- I3» fig- T) as weU as m that of D. sumatrensis, Arambourg (1959 : 64) notes to be almost completely effaced in the specimen of D. primaevus. The scapula is represented by a right and left specimen of the same individual, that of the skeleton of Dicerorhinus leakeyi no. 2, R.I, Rusinga, 1947. The spine is broken off in the left, but present and nearly entire in the right scapula ; the tuber spinae is a massive process measuring 75 mm. anteroposteriorly (PI. n, fig. i). The portion of the scapula anterior to the spine (supra-spinous fossa) is not completely preserved in either specimen. The thin anterior border is present in the left ; above the neck portion it is nearly straight. The posterior border of the scapula is concave throughout, and thickened in the middle portion. In the right scapula the posterior angle is missing. The glenoid cavity is incomplete costally in the left specimen. Measurements are given in Table 27 below : TABLE 27 Measurements of scapula of Dicerorhinus and Aceratherium (mm.) Dicerorhinus A ceratherium leakeyi Dicerorhinus acutirostratum ( A ^ sumatrensis sin. dext. sin. Height from anterior border of glenoid 495 505 355 cavity to upper end of spine Ant. post, diameter above tuber spinae c. 270 220 Ant. post, diameter of neck 115 115 76 no Ant. post, diameter over tuber scapulae 145 145 105 140 and glenoid cavity Ant. post, diameter of glenoid cavity 97 97 76 90 Transverse diameter of idem 78 60 67 Transverse diameter of tuber scapulae 61 62 32 c. 45 MIOCENE RHINOCEROSES OF EAST AFRICA 159 The left scapula of the A ceratherium acutirostratum skeleton unfortunately is much less complete, lacking the posterior portion above the neck and most of the spine. The glenoid cavity, however, is distinctly smaller than that in the pair of scapulae of the Dicerorhinus leakeyi skeleton from the same site. Further, it may be noted that the tuber scapulae is less extended transversely in the A ceratherium than in the Dicerorhinus specimens. The measurements have been entered in Table 27. The proximal portion of a left scapula from Rusinga (M . 18917, marked Rs . 23a) is still less complete ; the anteroposterior diameter of the glenoid cavity is 95 mm. Nothing can be said as to its generic position. In all these scapulae there is a small coracoid process on the costal surface of the tuber scapulae ; this process is well developed in a specimen of Dicerorhinus primaevus Arambourg (1959 : 64) of which no measurements have been given. Of D. schleier- macheri there is a figure of a scapula (Kaup 1834 : 42> pi- r3> fig- 3) with most of the spine missing and incomplete distally. The diameters of the glenoid cavity are given as 79 by 67 mm., smaller than in the specimens of D. leakeyi. In D. sumatrensis the glenoid cavity is not very much smaller. Of the humerus we have both the right (PI. II, fig. 2) and the left specimen in the skeleton of Dicerorhinus leakeyi no. 2, R. i, 1947. Both are fragmented and crushed to a considerable extent. The right is flattened so that the posterior part of the lateral tuberosity is placed very nearly between the caput and the anterior portion of that tuberosity. Half of the caput and all of the medial tuberosity are gone. The posterior surface is not much damaged proximally, but the deltoid tuberosity is missing and the musculo-spiral groove is flattened. Of the distal half of the right humerus only the lateral condyloid crest and the lateral epicondyle are preserved ; the trochlea is missing. The left humerus has most of the trochlea, but the proximal half of the bone is crushed and flattened anteroposteriorly, as are the head and the tuberosities except for the anterior part of the lateral tuberosity, which miraculously escaped damage. Nevertheless, a few measurements can be given, all approximate (Table 28). A poorly preserved right humerus, shellaced and belonging to the skeleton no. 850 of Aceratherium acutirostratum, is too much deformed for measurement. There are also two humerus portions, the proximal portion of a left specimen (M . 18915) of which the width over caput and lateral tuberosity is only 125 mm., and the distal half of a left specimen (M . 18916) with damaged trochlea and a least shaft width of only 50 mm., greatest distal width 130 mm., of the size of Recent D. sumatrensis. The humerus of D. schleiermacheri (Kaup 1834 : 42, pi. 13, fig. 4) appears to agree rather well with those of D. leakeyi (the scapula referred to by Kaup (above) is much smaller). The humerus mentioned by Gaudry (1862-67 : 2°6) to D. orientalis is more expanded proximally and distally, and so is that of D. primaevus (Arambourg 1959 : 66). The proximal width of a specimen of D. orientalis measured by Aram- bourg (1959) is only 154 mm., however. The trochlea width of the humerus of D. ringstroemi is no mm. (Bohlin 1946 : 219). Of two humeri from Steinheim, Roger (1900 : 17) gives measurements entered in Table 28 ; no. i he regards as probably belonging to Brachypotherium because of its 160 MIOCENE RHINOCEROSES OF EAST AFRICA TABLE 28 Measurements of humerus of Dicerorhinus (mm.) D. leakeyi dext. sin. D. schleiermacheri D. orientalis D. primaevus Greatest length (laterally) 450 457 442 Length from caput to medial c. 390 400 370 condyle Width over caput and posterior c. 150 180 190 191 part of lateral tuberosity Width at deltoid tuberosity M5+ 150 156 Least width of shaft c. 75 69 68 Greatest distal width 140+ 142 160 167 Width of trochlea c. no no Steinheim D. sumatrensis no. i no. 2 La Romieu Greatest length (laterally) 370 420 350 366 Length from caput to medial condyle 337 Width over caput and posterior part of 127 140 140 lateral tuberosity Width at deltoid tuberosity 108 Least width of shaft 48 Greatest distal width 115 155 105 103 Width of trochlea 81 no 80 77 greatest distal width (which exceeds the proximal width) ; no. 2, which would belong to either Dicerorhinus or Aceratherium, agrees well in comparative slenderness with the humerus from La Romieu referred to Aceratherium cf. platyodon Mermier by Roman & Viret (1934 : 32, pi. n, fig. 10). A left radio-ulna belonging to the skeleton of Dicerorhinus leakeyi no. 2, R.I (PI. n, fig. 3) has the ulna nearly entire but the radius incomplete medially and with the distal end shattered. There is no right radio-ulna of the same skeleton in the Rusinga collection sent to me, but there are a right radius and ulna, shellaced and belonging to the skeleton of Aceratherium acutirostratum no. 850, R.I, 1947, as well as the untreated left radius (PI. 9, fig. 2) and ulna marked no. 850, which are slightly smaller than the corresponding bones in D. leakeyi (Table 29). Apart from these associated radio-ulnae there are specifically unidentifiable radii and ulnae. There is a right radio-ulna, laterally flattened and restored with plaster, which is marked R.2 (Rusinga), too badly preserved for measurements to be taken. Then, the proximal part of a right radius (M . 18911, marked Rs . 21, Rusinga) 97 mm. wide, a proximal end of a left (no number discernible) 83 mm. wide, the distal end of a right radius (M. 18910, marked Rs.3i, Rusinga) 78 mm. wide, and three distal ends from the left side, M. 18914, M. 18909, and M. 18912, measuring about 75 mm., 80 mm., and 103 mm. in width, respectively. MIOCENE RHINOCEROSES OF EAST AFRICA 161 TABLE 29 Measurements of radius and ulna (mm.) Steinheim i A. acuti- D. orien- D. ring- D. prima- D. suma-( -- — A — — — ^ D. leakeyi rostratum talis stroemi evus trensis no. i no. 2 no. 3 Radius : Median length Proximal width 355 83 Greatest distal w. 90 350 330 70 385 c. no 1 08 379 105 IO2 293 83 85 370 105 IOO 340 87 90 315 75 70 A. acutirostratum D. leakeyi dext. Ulna: Greatest length 440 Length from processus anconaeus 400 c. 395 (" beak ") Length of olecranon (from 160 155 " beak ") Width at semilunar notch Middle width Greatest distal diameter 67 sin. D. primaevus D. sumatrensis 44° 395 73 45 145 45 67 396 345 125 69 39 56 The radius of D. orientalis (Gaudry 1862-67 : 2°6, pi- 32, fig. 4) is somewhat more slender than that of D. leakeyi distally ; that of D. ringstroemi (Ringstrom 1924 : 15 ; Bohlin 1946 : 221) on the other hand is wider distally and agrees rather well with the average of five radii of D. primaevus given by Arambourg (1959 : 67). The ulna of D. primaevus does not differ much in size from those of Dicerorhinus and Aceratherium from Rusinga, but in one of the primaevus ulnae (no. 197) the olecranon is rather short (88mm.?). Eight radii from Steinheim (Roger 1900 : 17) vary in length from 295 to 370 mm. and in distal width from 65 to 100 mm. Those with the greatest distal width (100 mm.) Roger regards as probably referable to Brachypotherium, but the specimens that have this distal width range in length from 300 to 370 mm. That of which the length is 300 mm. (no. 5 in Roger's table) agrees well with a Rusinga radius that I refer in the present paper to Brachypotherium heinzelini (Table 18). The measure- ments of the largest three Steinheim radii of Roger's are given in Table 29 ; Roger would refer only no. 3 to either Dicerorhinus or Aceratherium, but nos. i and 2 do not differ much in dimensions from those in various fossil Dicerorhinus species and may belong to this genus or to Aceratherium. Of the skeleton of Aceratherium acutirostratum no. 850, R.I, 1947, we have the three proximal carpals, scaphoid, lunar and cuneiform (PI. 14, figs. 4, 5) as well as the pisiform from the right side ; these bones were evidently in articulation in the deposit and are much deformed as a whole, fitting well on to each other but inadequate for accurate measurement. 162 MIOCENE RHINOCEROSES OF EAST AFRICA There are three isolated left scaphoids, one from Karungu (M . 18897) (PI. 14, fig. 6), and two marked KB. A, Maboko (= Kiboko) Island (M. 18896 and M. 18898), and also the anterior portion of a left specimen (M . 18899) • These bones are similar in proportions and differ only in size ; the smallest is even exceeded in size by the scaphoid in D. sumatrensis (Table 30). TABLE 30 Measurements of scaphoid (mm.) M. 18897 M. 18898 M. 18896 M. 18899 D. sumatrensis Posterior height 71 65 50 56 Anterior height 55 49 41 55 55 Proximal width 49 44 38 47 Distal width 47 41 37 47 Ant. post, at middle 66 58 51 65 The Karungu and Maboko Island bones resemble that of D. schleiermacheri (Kaup 1834 : 43, pi. 13, fig. 9) in shape except that their posterior height is relatively greater. The proximal projection behind the saddle-shaped radius facet and the distal pro- jection behind the trapezium facet are more developed than in the Eppelsheim specimen, and the height taken over these projections greatly exceeds the height taken in front, over the convexity of the radius facet and the ridge between the facets for trapezium and magnum. Kaup gives neither of these measurements in the text, but from his figures it is clear that the posterior height in D. schleiermacheri is only slightly the greater, and so it is in D. sumatrensis. Kaup does give the length of the three distal facets (for trapezium, trapezoid, and magnum) as 61 mm., and that of the proximal facets as 49 mm. ; these figures agree closely with those in the largest of our scaphoids. The total length of the Eppelsheim bone is given as 90 mm. ; the greatest diameter of the Karungu scaphoid, measured over the posterior upper end and the distal outward projection is slightly less (86 mm.). This projection does not extend outward beyond the radius facet, and thus the distal width does not exceed the proximal width, as it does in Brachypotherium in contradistinction to Aceratherium (and Dicerorhinus} (Roger 1900 : 19). The scaphoid of D. ringstroemi (Ringstrom 1924 : 15, fig. 6 ; Bohlin 1946 : 222, text-fig. 78-2) is much larger than any of our specimens : the greatest diameter is 104 mm., the anterior height 69 mm., although the width of the proximal facet is only 47 mm. Arambourg (1959 : 67) gives measure- ments of the scaphoid of D. primaevus, viz., maximal height 55 mm., and " longueur maxima " (evidently anterposteriorly) 61 mm., which makes the bone intermediate in size between M. 18897 and M. 18898. The deformed scaphoid of the proximal carpal series of Aceratherium acutirostratum (no. 850) is approximately 55 mm. high anteriorly. One lunar, marked Rs., Rusinga (M. 18906), is from the right side and incomplete behind ; another, marked R . 2, Rusinga (M . 18907), from the left side and incomplete medially. Both have a proximal lateral facet for the ulna, as in Aceratherium and Dicerorhinus, although in M. 18906 it is very small. The last-mentioned specimen MIOCENE RHINOCEROSES OF EAST AFRICA 163 (Table 31) closely agrees with the lunar of D. sumatrensis in shape as well as size. The anterior height of M . 18907 is greater, as in a lunar from Steinheim referred to Aceratherium by Roger (1900 : 19). The bone in question of D. schleiermacheri is figured only by Kaup (1854, pi- 7) > that of D. ringstroemi is figured by Bohlin (1946 : 223, text-fig. 79-2), with measurements, while that of D. primaevus (Aram- bourg 1959 : 67) is unfigured. The deformed lunar in the associated proximal carpal series of no. 850 (Aceratherium acutirostratum} is about 55 mm. high and wide. TABLE 31 Measurements of lunar (mm.) D. suma- M. 18906 M. 18907 D. ringstroemi D. primaevus trensis Steinheim Anterior height 40 47 61 40 48 Proximal width 45 62 50 44 Greatest ant. post. 68 77 67 65 64 diameter A right cuneiform from Karungu (M. 18903), and a left from Rusinga (M. 18904), the latter damaged anteriorly below, are rather small, smaller than that in D. sumatrensis, but closely similar to it. A left cuneiform from Rusinga (M. 25184) is incomplete externally below, and slightly larger. This bone in D. primaevus (Aram- bourg 1959 : 67, " Pyramidal ") is rather extended horizontally. The cuneiform of Aceratherium acutirostratum (no. 850) has an anterior height of about 48 mm. TABLE 32 Measurements of cuneiform (mm.) M. 18903 M. 18904 M. 25184 D. primaevus D. sumatrensis Anterior height 38 50 46 46 Distal width 32 33 38 Proximal ant. post. 25 26 32 29 diameter Greatest horizontal 36 50 48 diameter The pisiform is available only in the deformed right carpal series. It is over 60 mm. in length, and at least 40 mm. in distal height (50 mm., and 33 mm. in D. sumatrensis}. The proximal facets are injured, but one fits well on to the cuneiform. A right magnum, marked RS.IOI, Rusinga (M. 18902) is incomplete anteriorly and medially and lacks most of the posterior downward process. A right and a left magnum marked R .3, Rusinga (M . 18900 and M . 18901), however, are well preserved. The posterior process is much more developed transversely in the right than in the left specimen, D. sumatrensis being intermediate in this respect (Table 33) ; D. primaevus is unfigured. Two well-preserved unciforms, one right, marked Rs.3i, Rusinga (M. 18884), and one left, Kathwanga, Rusinga (M .25191) (PI. 14, fig. 7) resemble that in D. sumatrensis 1 64 MIOCENE RHINOCEROSES OF EAST AFRICA TABLE 33 Measurements of magnum (mm.) M. 18902 M. 18900 M. 18901 D.primaevus D. sumatrensis Greatest anterior c. 33 25 27 33 30 height Greatest anterior 38 36 39 42 width Proximal ant. post. 65 53 57 67 diameter Greatest diameter 82 75 85-5 79 very well indeed. The unciform of D. schleiermacheri (Kaup 1834 : 43, pi. 13, fig. 8) is rather small, apparently of a young individual ; those of D. ringstroemi (after Bohlin 1946 : 225) present dimensions greater than the Rusinga specimens (Table 34). TABLE 34 Measurements of unciform (mm.) D. schleier- D. ring- D. suma- M. 18884 M. 25191 macheri stroemi trensis Greatest anterior height 45 51 49 Greatest width 60 68 54 c. 74 78 61 Greatest ant. post, diameter 80 94 105 108 77 In the Rusinga collection there is an entire Me. IV sin., M. 18814, marked R.I, Rusinga, that is associated with the proximal portion of an Me. Ill sin., M. 18841, with the same inscription. That these bones belong to the same individual cannot be doubted ; their state of preservation is exactly the same and there could not be a better fit (PI. 12, figs. 2, 3). Then there is the proximal portion of an Me. II sin. (M. 18843) that would seem to belong to the same individual but that is marked Rs., which means that it is a specimen picked up from the surface and not found in situ (cf. Le Gros Clark & Leakey 1951 : 3). Its preservation is exactly that of the Me. III-IV, and size and proximal articulation with Me. Ill could not be more fitting. There is another set of metacarpals from Rusinga representing one or two indi- viduals, and again from the left side. Here are the proximal portions only, that of Me. II (M. 18842) marked Rs. (surface find), but those of Me. Ill (M. 18837) and Me. IV (M. 18840) marked Rs.3i (PL 12, fig. 4). These bones are all incomplete behind, Me . IV laterally too, and smaller than the set first mentioned. In both Me. IV (M. 18814 and M. 18840) there is a small lateral proximal facet, placed anteriorly, making an obtuse angle with the large proximal unciform facet and only 6-8 mm. wide by an anteroposterior diameter of some 20 mm. There is no posterior interproximal facet on the lateral side of Me . IV. Now this facet articulated with Me . IV, but whether this was a mere rudiment or a sizable though small meta- carpal it is impossible to tell. As related above (p. 153) Dicerorhinus so far as known has a mammiform Mc.V as extended proximally as the short Mc.V in Aceratherium, MIOCENE RHINOCEROSES OF EAST AFRICA 165 giving the same small facets for articulation with its neighbour and with the unciform (cf. Kaup 1859). Only tne nnd of an Me . V, associated, would tell the genus to which the other metacarpals belong. Hence, it is uncertain whether the two sets of meta- carpals belong to Dicerorhinus or to Aceratherium, and we have to leave it at that. Without mentioning these facets, Roger (1900 : 41) writes that the proximal (unciform) facet of Me. IV in Aceratherium is distinctly wider in front, and more rapidly reduces in width posteriorly than that in Dicerorhinus, which has a relatively narrower and more anteroposteriorly elongated unciform facet. However, the unciform facet in Recent D. sumatrensis Me. IV is shaped exactly as that in the Rusinga Me. IV, and of Aceratherium I have no reliable material for comparison. The character mentioned as distinctive of the fourth metatarsal of Aceratherium by Roger (1900) is found in that of Dicerorhinus leakeyi and will be mentioned later on. It seems best to leave the generic position of the metacarpal sets in doubt, although Me . IV of the larger set fits well with the other bones of D. leakeyi as can be seen in Table 26. There is further an isolated Me. IV, from the right side (M. 18811, Rusinga), that lacks the posterior proximal portion but permits of the median length to be taken (PI. 12, fig. i) ; it is more slender than M . 18814 but less so than M . 18840, and its measurements have likewise been included in Table 35. The metacarpals of D. schleiermacheri are an Me. Ill sin. and an Me. II sin., the latter erroneously described as a right outer (fourth) metacarpal (Kaup 1834 : 43> pi. 13, figs. 13 and 12 ; 1854, pi. 7). The greatest length of Me. Ill is given as 204 mm., that of Me . II as 178 mm. (200 mm., and 177 mm., in Kaup 1854). The median length of the median metacarpal of D. schleiermacheri would be some 190-194 mm., 10 mm. less than the greatest length, at any rate exceeding that of D. primaevus as given by Arambourg (1959 : 68). The Me. Ill of D. orientalis (Gaudry 1862-67 : 205/6, pi. 32, fig. 6) has a median length of 164 mm. and a greatest distal width of 61 mm. Of D. ringstroemi the left metacarpals III and IV have been made known by Ringstrom (1924 : 15, fig. 10). The Me. Ill has a median length of 187 mm., and a greatest distal width of 68 mm. (even 73 mm. in an incomplete Me . Ill) . The Me . IV of D. ringstroemi has a median length of 153 mm. and a greatest distal width of 50 mm., measurements that agree very well with those of M . 18814. Of D. primaevus measure- ments have been entered in Table 35 ; Me . IV is shorter and less expanded distally than that of D. ringstroemi and the Rusinga specimen, while Me. Ill of D. primaevus is likewise shorter and less wide distally than that of D. ringstroemi. The longest metacarpals are those of D. schleiermacheri, and its fourth metacarpal, when known, would probably exceed that of Dicerorhinus leakeyi in size. In D. primaevus Me. IV is slightly more shortened relative to Me. Ill than in D. sumatrensis. It will be observed that the metacarpals of the living species are relatively more expanded distally than those in the Rusinga Dicerorhinus or A ceratherium. The relative width of the fourth metacarpal in D. sumatrensis, however, is exactly the same as that in the Rusinga form. An Me. Ill from Steinheim referred to Aceratherium by Roger (1900 : 40) is as long as that of D. schleiermacheri : its length is 192 mm. and the middle width 55 mm. GEOL. 13, 2. II M. 18841 M. 18837 166 MIOCENE RHINOCEROSES OF EAST AFRICA TABLE 35 Measurements of metacarpals (mm.) Me. II M.I 8843 M.I 8842 D.primaevus Median length 156 Proximal width 43 39 Proximal ant. post, diameter c. 47 Middle width c. 38 c. 33 Middle ant. post, diameter Greatest distal width 41 Width of distal trochlea Distal ant. post, diameter Me. Ill Median length 179 Proximal width 64 49 Proximal ant. post, diameter 55 c. 40 Middle width c. 50 42 Middle ant. post, diameter 18 Greatest distal width 61 Width of distal trochlea Distal ant. post, diameter Me. IV M. 18814 M. 18811 Median length 150 145 140 Proximal width 51 41 Proximal ant. post, diameter 47 Middle width 34 34 Middle ant. post, diameter 21 21 Greatest distal width 50 44 38 Width of distal trochlea 44 40 Distal ant. post, diameter 41 37 Ratio middle width /length 0-23 0-23 D. sumatrensis 136 35 37 34 18 47 36 158 57 47 45 18 59 48 42 130 42 42 30 17 45 38 4i 0-23 Three Steinheim Me. IV referred to Aceratherium by the same author are 150-160 mm. in length, and 30-31 mm. in middle width, very slightly more slender in the shaft than the Rusinga form. TABLE 36 Measurements of metacarpals (mm.) Proximal width Proximal ant. post, diameter K.4) M. 18845 M. 18848 1950 M. 18838 M. 18839 M. 18851 M. 25183 41 39 41 55 58 53 48 42 5«> 46 45 In addition to the above-mentioned metacarpals there are various proximal metacarpal portions that belong either to Dicerorhinus or to Aceratherium, viz., M. 18845, Rusinga, Me. II sin., damaged behind, M. 18848, Kachuku, Lower Series, Karungu, Me. II dext., incomplete behind, K-4, 1950, Ngira, Karungu, Me. II sin., articular surface incomplete, MIOCENE RHINOCEROSES OF EAST AFRICA 167 M. 18838, Rusinga, Me. Ill dext., M. 18839, Rs.i03, Rusinga, Me. Ill dext., M. 18851, Kachuku, Lower Series, Karungu, Me. Ill sin., and M. 25183, Rusinga, Me. Ill sin. Four phalanges are unnumbered and associated ; they are thought to belong to the skeleton of Dicerorhinus leakeyi no. 2, R . i, 1947. They comprise the first and second phalanges of the median digit and of a lateral digit, and if they belong to the no. 2 skeleton they form part of the fore foot as the median phalanges of both hind feet of the skeleton are present. Measurements are given in Table 37. TABLE 37 Measurements of anterior phalanges (mm.) Rusinga D. primaevus f A s t A v D. sumatrensis -ICfVUCfcU digit AdUV^l EU digit JUU.WUCUJL digit ICbbW* CIJ. digit II III IV 36 37 39 42 34 33 3i 50 c. 41 55. 39'5 39 48 40 26 25 27 28 23 26 23 53 39 55 40 34 52 35 Phalanx I, length Proximal width Phalanx II, length Proximal width There is a considerable number of fragments of ribs, some annotated, some not, all belonging to the skeleton of Dicerorhinus leakeyi no. 2, R.I, 1947. Encouraged by the photograph published by Cole (1950, pi. i), which shows that at least the ribs of one side had been mostly there in a complete state, I have tried to reassemble the fragments, and arrange them, after determining the side to which the ribs belong (possible in most cases), in such a way as to obtain a graded series. The head and tubercle, the relative size and position of which gradually change as one passes along the series, are only rarely available, but using Recent rhinoceros skeletons of various genera (Diceros, Rhinoceros, and Dicerorhinus} in the Leiden Museum as a guide, it has been possible to determine all but a few of the twenty right ribs that were on the lower side of the skeleton before excavation, and about half that number of left ribs. The numbering of the fossil ribs is, of course, to some extent provisional. To begin with, the ribs, and especially the wide anterior ribs up to about the ninth, are distorted, as are most of the larger bones of this skeleton. In some of the ribs, when laid with their posterior border on a table, the body does not stand up but is nearly flat on the table. The sternal ends are better preserved than the vertebral. The first rib, easily recognizable because of its large tubercle, shortness, and rapid increase in greatest diameter toward the sternal end, is present from both sides. The left rib is best preserved, the head only being lost. The lateral flattening and distortion (there is a peculiar S-twist near the middle of the body that is unnatural) are severe, however. Its greatest length is 280 mm., the maximal diameter at the sternal end 50 mm., while at the narrowest point below the (missing) head it measures 28 mm. anteroposteriorly and 17 mm. transversely in cross section. In the right first rib both extremities have been lost ; the cross section as taken on the left rib is 22 by 168 MIOCENE RHINOCEROSES OF EAST AFRICA 20 mm., showing that the right rib is less laterally compressed than the left. In an adult male of D. sumatrensis used for comparison with the fossil skeleton throughout the first rib is 230 mm. long, 33 mm. in greatest sternal diameter, and 20 by 13 mm. in section below the head. The second rib, from the right side, is a slightly curved specimen without head or tubercle but with the sternal end. Length as preserved 380 mm. (the length of the complete second rib in D. sumatrensis}, and greatest diameter at middle of body 40 mm. (24 mm. in D. sumatrensis}. Of the left second rib the proximal portion only is present in the collection. The head and tubercle are preserved in the anterior of two right ribs that are still held together by matrix proximally. They agree best with the third rib in Recent skeletons in the configuration of the vertebral end. The third rib, then, is incomplete sternally ; length 490 -f- mm. (440 mm. in D. sumatrensis}, and greatest diameter at middle 36 mm. (26 mm.). The fourth right rib, the sternal end of which is preserved, had a length of about 600 mm. (520 mm.) when complete, by a greatest diameter at middle of about 50 mm. (29 mm.). The (?) fifth right rib lacks a portion of the body that I have been unable to find among the fragments ; diameter at middle of about 50 mm. (32 mm.). The (?) sixth right rib which has the sternal end but no head, is 530 mm. long as preserved, and 52 mm. (30 mm.) in greatest middle diameter. The (?) seventh right rib is without the vertebral end, a body fragment only 275 mm. long and 49 mm. in greater diameter. What is probably the sixth or seventh left rib is an entire specimen, 750 mm. long along the curve, and 50 mm. in diameter along most of its length. In D. sumatrensis the sixth and seventh ribs are 630-670 mm. long and 27-30 mm. in diameter. The (?) eighth right rib is entire, with a length of 900 mm. (700 mm.) and a greatest diameter of 42 mm. (22 mm.). A number of right ribs, all without the vertebral end, have the same curvature as the (?) eighth, and either the same or a smaller thickness. These I have arranged so that the sternal end decreases in size posteriorly. The best preserved is the (?) eleventh rib, of which the costal tubercle at least (already much reduced) is present. The total length is probably 900 mm. (730 mm.) and the greatest diameter 34 mm. (23mm.). From the eleventh rib onwards the ribs in Recent species become very thin medio- laterally (least so in Rhinoceros}, whereas the fossil ribs assigned to this region are more robust. The curvature remains the same until the seventeenth rib has been reached. We have the greater part of what appear to be the twelfth to seventeenth ribs from the right side, and fragments of the ninth to thirteenth ribs from the left side, the side into which the scavengers penetrated. The greatest diameters of the incomplete specimens vary from 30 to 37 mm. (21-26 mm. in D. sumatrensis}. The almost entirely preserved (?) eleventh rib must have been one of the longest, as in the Recent species the length starts to decrease backward from about the eleventh rib. The penultimate and last ribs are very thin at their free end, which are preserved in what appear to be the eighteenth, nineteenth, and twentieth ribs from the right side. But the last rib (twentieth?) from the left side even has the vertebral end too, MIOCENE RHINOCEROSES OF EAST AFRICA 169 in which tubercle and head have become united. Its length along the curve is 410 mm., the greatest vertebral diameter 39 mm., at the free end 10 mm. In D. sumatrensis the twentieth rib is absent although the nineteenth is still 420 mm. long. In Diceros bicornis (L.) one skeleton (Leiden Museum reg. no. 5738) has the twentieth and last rib with the same diameters at the ends as in the fossil, but its length is only 280 mm. Of the pelvis of the skeleton of Dicerorhinus leakeyi no. 2, R. i, 1947, there is only a portion from the right side, which shows the acetabulum, part of the shaft of the ilium, and the pubis and ischium around the obturator foramen, but the specimen is fragmented and distorted to such a degree that no measurements can be recorded. The left femur is one of the best preserved bones of the skeleton of D. leakeyi no. 2, R.I, 1947. Its distal part is slightly displaced relative to the proximal portion, but the usual flattening is not extensive (PI. 13, fig. i). The right femur of the same individual has the distal end crushed and incomplete, and the proximal end missing. It has, however, the (flattened) third trochanter, broken off in the left specimen. Measurements are recorded in Table 38. TABLE 38 Measurements of femur of Dicerorhinus and Aceratherium (mm.) Greatest length Proximal width Least width of shaft Greatest distal width Distal ant. post. diameter, medial side Transverse diameter caput D. leakeyi sin. 545 215 75 c. 145 1 80 95 D. orientalis (Pikermi) D. schleier- macheri 555 153 1 60 (Gaudry) (Arambourg) . D. ring- stroemi D. prima' evus •\ 54° 491 499 — 538 — 192 198 — 223 90 75 75 — 82 155 J45 146 170 150 — — — c- 235 — Steinheim (Roger 1900) Greatest length Proximal width Least width of shaft Greatest distal width Distal ant. post, diameter, medial side Transverse diameter caput D. sumatrensis r ~\ no. i no. 2 no. 3 no. 5 423 162 54° 533 450 395 56 125 150 138 no 105 150 — — — — A ceratherium incisivum 123 149 73 The right, shellaced femur of the skeleton of Aceratherium acutirostratum no. 850, R.I, 1947, consists of the proximal and distal portions that do not fit, and the de- formation does not permit of measurements to be taken. There is also the isolated caput of a femur, no. 991, Rusinga, 1947, that measures 79 mm. transversely, almost as small as in D. sumatrensis. i?o MIOCENE RHINOCEROSES OF EAST AFRICA The femur of D. leakeyi agrees very well in size with the largest femur of D. primaevus (Arambourg 1959 : 70) given in the Table ; that of D. schleiermacheri (Kaup 1834 : 44» pl- I3» n§- 5) *s l°nger than either of them. The femur of the Pikermi D. orientalis as given by Gaudry (1862-67, pi. 32, fig. 7) is rather similar, but the femora recorded by Arambourg (1959 : 70) as belonging to the same species are noticeably shorter. The bone in D. ringstroemi (Bohlin 1946 : 227) is wider distally than any of the others presented in Table 38. Of the Steinheim femora recorded by Roger (1900 : 17) the largest (no. i) is like that of D. leakeyi in the few measurements given. Roger would refer most of the Steinheim femora to Brachypotherium except the slender (and short) no. 5, which he would place with Aceratherium. The femur referred to Aceratherium incisivum by Kaup (1834 : 59, pi. 15, fig. i) has a greater distal width than Roger's nos. 3 and 5, but its length is unknown. The Aceratherium (or Dicerorhinus!} femur agrees with that of D. sumatrensis in size. The patella is represented by five specimens. No. 718, which belongs to the skeleton of Dicerorhinus leakeyi no. 2, R. i, 1947, is corroded, rather small, and drawn out in a point distally. There are a large right specimen belonging to the skeleton of Aceratherium acutirostratum no. 850, distorted and incomplete, and three unnumbered specimens, one left and two right, the last of which bears the mark Rs.3i, Rusinga. TABLE 39 Measurements of patella (mm.) D. leakeyi A. acuti- dext. dext. rostratum sin. dext. Rs.3i D . sumatrensis Length 76+ 84 98 105 91 Width 60+ 85 83 87 82 77 It may seem peculiar that the patella of the Dicerorhinus from Rusinga (it fits well on to the articular surface of the right femur, and certainly belongs to the no. 2 skeleton) is so much smaller than that of the A ceratherium, whereas in the other bones of the two skeletons D. leakeyi is (slightly) the larger as far as can be seen. The patella, however, may be just abnormally stunted in the D. leakeyi skeleton ; it is unfortunate that we do not have this bone from the left limb also. The patella of D. schleiermacheri (Kaup 1834 : 44, pi. 13, fig. 7) differs from that referred to Aceratherium incisivum (Kaup 1834 : 60, pi. 15, fig. 12) in being more squarish ; that of A. incisivum (it should be noted that it has been figured upside down) is more drawn out medially. The length is 90 mm. in both ; the width about 80 mm. in D. schleiermacheri against 94 mm. in A. incisivum. A left tibia and fibula (PI. 13, fig. 2) but only the right fibula, are labelled as belonging to the skeleton no. 2, R. i, 1947 ; Dicerorhinus leakeyi. The tibia is some- what laterally flattened in its proximal part, but the distal end is well-preserved. Fortunately the right tibia of the shellaced limb of Aceratherium acutirostratum, though laterally flattened, permits a few measurements to be taken, which show it to be slightly shorter than that of D. leakeyi (Table 40). A right tibia marked KB.S, MIOCENE RHINOCEROSES OF EAST AFRICA 171 Maboko (= Kiboko) Island, lacks the lateral proximal portion. The proximal portion of a right tibia (M . 18920), marked Rs . 105, Rusinga, the lower two-thirds of a right tibia marked R.I, Rusinga, the distal end of a right tibia (M. 18919) marked Rs.3i, Rusinga, and that of a left tibia (M. 18918) marked Rs., Rusinga, are the remaining specimens in the collection ; measurements so far as possible are recorded in Table 40. TABLE 40 Measurements of tibia of Dicerorhinus and Aceratherium (mm.) D. leakeyi A. acuti- sin. rostratum KB.S M. 18920 R.I M. 18919 M. 18918 Greatest length 420 c. 410 380 Medial length 375 c. 370 350 Proximal width 125 + Distal width 100 82 95 92 95 Distal ant. post. c. 80 79 c.8o 73 72 diameter Steinheim D. schleier- /— — •*- v macheri D. orientalis D. primaevus D. sumatrensis no. i no. 2 Greatest length 388 313 340 300 Medial length 350 340-372 282 Proximal width 130 126 119-130 116 no 120 Distal width 96 98-109 82 75 100 Distal ant. post. 68 59 diameter The tibia of D. schleiermacheri has a greatest length less than that in D. leakeyi (taken from Kaup 1854 ; the other measurements after Kaup 1834 : 44). That of D. orientalis (Gaudry 1862-67 : 2O7> pi- 32> ng- 8) is within the limits of eleven (!) tibiae of D. primaevus (Arambourg 1959 : 71) or nearly so. Of the Steinheim tibiae recorded by Roger (1900 : 18) no. I is regarded as representing Aceratherium, the shorter one (no. 2) is as short as in Brachypotherium but not quite so broad ; in Brachypotherium br achy pus the proximal width is one-half the length. The fibula belonging to the skeleton of Dicerorhinus leakeyi no. 2, R.I, 1947, is longer than the three fibulae recorded from Steinheim by Roger (1900 : 18). Of these, that with a length of 280 mm. (see Table 41) is stated to belong to the supposed Aceratherium tibia that is 340 mm. in greatest length (Steinheim no. I in Table 40) ; the others must have belonged to even longer tibiae. The preservation of the right fibula of the skeleton of D. leakeyi is perfect ; it is an enigma to me why the right tibia of the skeleton has not been preserved. There is an abundance of astragali in the present East African collection : some twenty specimens in all one of which (no. 538, Gumba, Rusinga, 1949) is that of Brachypotherium and is dealt with under the head B. heinzelini in the present paper. All the others belong to either Dicerorhinus or Aceratherium. I72 MIOCENE RHINOCEROSES OF EAST AFRICA TABLE 41 Measurements of fibula of Dicerorhinus and Aceratherium (mm.) D. leakeyi D. sumatrensis Steinheim dext. ( A pi- r3> ng- n) is a rather high specimen, the lateral height being given as 85 mm., the trochlea width as 70 mm., and the greatest width as 81 mm. From Kaup's illustrations it seems that the external height has been taken over the ridge between the cuboid and the navicular facet, and that the trochlea width has not been taken over the lateral and medial surfaces but perhaps at the top of the ridges. In the British Museum (Natural History) there is a cast of a right astragalus from Hessen-Darmstadt, M.2786, catalogued as being of the specimen figured by Kaup (1834, pi. 13, fig. n) as D. schleiermacheri. Its measurements are given in Table 43 ; they correspond well with those of the East African astragali although the trochlea width exceeds the medial height. The astragalus of D. ringstroemi (Ringstrom 1924 : 16, text-fig. 8) is larger than the largest Rusinga specimen ; that of D. primaevus (Arambourg 1959: I74 MIOCENE RHINOCEROSES OF EAST AFRICA 72) in the few measurements given seems to agree well with our no. 3 ^.3264, from Rusinga). Under Aceratherium incisivum Kaup (1834, pi. 15, fig. 2) figures an astragalus from Oppenheim1 that is relatively much higher than an astragalus from Budenheim (pi. 15, fig. 10). According to the measurements given by Kaup (p. 60) the trochlea width (stated to have been taken over the external and internal surfaces) much exceeds the medial height in the Budenheim specimen, although it evidently does not have the great total width characteristic of Brachypotherium. An astragalus from Steinheim recorded by Roger (1900 : 24) as Aceratherium is rather high laterally (perhaps measured over the distal ridge between cuboid and navicular facets), but the ratio of medial height to total width is as in various Karungu and Rusinga specimens. The astragalus associated with teeth of Aceratherium tetradactylum (Rinnert 1956 : 34, pi. 3, figs, i, 3) is intermediate between the two Aceratherium specimens figured by Kaup in medial height as well as in trochlea width. The one and only definite Aceratherium astragalus in our collection (no. 10) cannot be measured because of its poor state of preservation. That from Losodok cautiously identified by Arambourg (1933 : 10) as Aceratherium^ sp. corresponds with various of our specimens in height and total width (74, and 86 mm., respectively) and, as Arambourg realized, rep- resents either Aceratherium or Dicerorhinus. TABLE 43 Measurements of astragalus of Dicerorhinus and Aceratherium (mm.) D. schleier- D. ring- D. prima- D. suma- Oppen- Buden- Vieh- Stein- macheri stroemi evus trensis heim heim hauseri heim Lateral height 66 63 71 — 76 Medial height 75 79 68 61 67 64 66 Total width 93 100 90 82 — 78 Ratio medial height/ 0-81 0-88 0-83 0-85 total width Trochlea width 80 89 70 63 77 67 Width of distal facets 76 66 68 66 tuber In sharp contrast to the abundance of astragali, there are only three specimens of the calcaneum in the East African Miocene collection, the left calcaneum of the skeleton of Dicerorhinus leakeyi no. 2, R.I, 1947, well preserved (PL 14, fig. i), a 1 The Oppenheim astragalus has the medial and the distal calcaneum facets confluent, which is regarded as typical for Aceratherium by Ringstrom (1924 : 74, text-fig. 43). In the Budenheim astragalus these two facets are separate. The difference is apparently trivial ; there may, or may not, be a shallow non- articular groove between these two facets. In most of the East African astragali the junction area of the medial and distal calcaneum facets is damaged, but in M. 18881 and M. 18882 the two facets are confluent, in no. 679 separate, whereas in the astragalus of D. leakeyi there appears to be a very narrow groove bet- ween the two, an intermediate condition we see also in D. sumatrensis. In two astragali from Steinheim figured by Fraas (1870 : 302, pi. 7, figs. 7, 8) as Rh. sansaniensis and Rh. brachypus (that is, Dicerorhinus and Brachypotherium) respectively, the medial and distal calcaneum facets are separate. The larger specimen (pi. 7, fig. 8) agrees in shape and size with Aceratherium tetradactylum, while the smaller (pi. 7, fig. 7) may belong to Dicerorhinus (Rinnert 1956 : 36). Needless to say, I do not think that the presence or absence of a bridge between the medial and the distal calcaneum facet of an astragalus is more than an individual variation, useless for intergeneric comparison. MIOCENE RHINOCEROSES OF EAST AFRICA 175 much deformed shellaced specimen belonging to the right limb of Aceratherium acutirostratum, and no. 679, R.io6, Rusinga, 1947. The last is from the left side but evidently is of the same individual as the right astragalus bearing the same number (no. 5 in Table 42). The tuber and the cuboid facet are not complete, and approxim- ate measurements only can be given. TABLE 44 Measurements of calcaneum of Dicerorhinus and Aceratherium (mm.) D. leakeyi D. schleier- D. prima- D. suma- Buden- Vieh- Stein- Lateral height Greatest width Ant. post, cuboid facet Transv. cuboid facet Greatest diameter tuber Transv. diameter tuber sin. c. 25 73 54 no. 679 macheri 135 c. 60 c. 40 48 evus 132 76 24 c. 60 trensis 1 06 71 39 23 63 46 heim 122 hausen heim no 129 71 82 48 21 81 45 57 — 46 The calcaneum of D. schleiermacheri (Kaup 1834 : 45 » pi- I3> ng- IO) appears to correspond well with that of D. leakeyi ; the Rusinga bone is only somewhat longer. The greatest width of the Eppelsheim specimen is about 80 mm. That of D. primae- vus, the length and width only of which are known, is smaller though it is near to the calcaneum from Steinheim recorded by Roger (1900 : 23) (which is perhaps the same as the specimen figured by Fraas 1870, pi. 7, fig. 10). The bone in question, figured as Rh. sansaniensis (hence, Dicerorhinus), is considered indistinguishable from that of Aceratherium tetradactylum by Rinnert (1956 : 36). The Budenheim calcaneum ascribed to Aceratherium incisivum (Kaup 1834 : 60, pi. 15, fig. n) differs from that of A. tetradactylum from Viehhausen figured by Rinnert (1956, pi. 3, figs, i, 2) only in size and the configuration of the tuber. Of the skeleton of Dicerorhinus leakeyi no. 2, R.I, 1947, we have the entire left navicular, and also an anterior fragment of the right (separately catalogued under no. 360). There are also two right and two left naviculars, viz., M. 18887, R-i» Rusinga, M. 25187, Kathwanga, Rusinga, 1947, No. 64, 1950, R.I, Rusinga, and M. 25188, Kathwanga, Rusinga, 1947. These bones are very similar in shape and differ mainly in dimensions ; the navi- cular of D. sumatrensis is only relatively wider (the width of these bones has not been recorded for D. ringstroemi (Ringstrom 1924 : 16, text-fig. 8) or for D. primaevus}. The navicular of D. schleiermacheri (Kaup 1834 : 45) is larger than that in Aceratherium incisivum (Kaup 1834: 60, pi. 15, fig. ga-c), which is similar in size to that of the recent species ; its total width is about 50 mm. 176 MIOCENE RHINOCEROSES OF EAST AFRICA TABLE 45 Measurements of navicular of Dicerorhinus and Aceratherium (mm.) D. leakeyi M. 18887 M. 25187 no. 64 M. 25188 D.primaevus Greatest anterior height 31 31 26 25 20 30 Total width 49 52 47 47 44 Ant. post, diameter 69 71 60 64 54 60 Aceratherium D. ringstroemi D. sumatrensis incisivum Greatest anterior height 27 20 Total width 50 Ant. post, diameter 75 54 56 There is the left cuboid of the skeleton of Dicerorhinus leakeyi no. 2, R.I, 1947 (PI. 13, figs. 4, 5) and in addition no less than eight isolated cuboids, six right and two left: M. 18892, Karungu, M. 18890, R.I, Rusinga (PI. 13, fig. 6); M. 18891, Kachuku, Lower Series, Karungu, No. 440, Ngira, Karungu, M. 18894, Rs.io5, Rusinga, M. 18893, R.8, Rusinga, M. 18895, Kachuku, Karungu, and M. 18889, Rs-> Rusinga. These cuboids appear to fall into two groups, viz., one in which the anterior height is nearly equal to the anterior width, and one (M . 18890 and M . 18893) in which the anterior surface is distinctly higher than wide (Table 46). In Brachypotherium the TABLE 46 Measurements of cuboid of Dicerorhinus and Aceratherium (mm.) D. leakeyi M. 18892 M. 18890 M. 18891 no. 440 M. 18894 M. 18893 Anterior height 48 48 46 42 41 41 36 Anterior width 47 47 36 43 42 40 30 Greatest ant. post. 64 73 63 69 c. 65 66 diameter Aceratherium M. 18895 M. 18889 D.primaevus D. sumatrensis (Roger) (Rinnert) Anterior height 40 34 48 40 46 40 Anterior width 37 35 39 40 34 30 Greatest ant. post. 56 diameter cuboid is distinctly wider than high anteriorly (Roger 1900 : 24) ; the present speci- mens represent either Dicerorhinus or Aceratherium. In D. schleiermacheri the anterior height of the cuboid is equal to the anterior width (Kaup 1834 : 45) > and so it MIOCENE RHINOCEROSES OF EAST AFRICA 177 is in the cuboid of D. leakeyi and in that of Recent D. sumatrensis. However, the cuboid of D. primaevus as appears both from the measurements and from the illustra- tion (Arambourg 1959 : 72, text-fig. 326) is decidedly higher than wide. This is also the case in the Steinheim Aceratherium recorded by Roger (1900 : 24) and in Acera- therium tetradactylum from Viehhausen (Rinnert 1956:34, pi. 3, fig. i). In A. incisivum (Kaup 1834 : 45, pi. 15, fig. 9) the anterior width of the cuboid appears to be somewhat greater than the height. Therefore, it would seem that the relative height of the cuboid cannot be used in intergeneric differentiation. Both ectocuneiforms of the skeleton of Dicerorhinus leakeyi no. 2, R.I, 1947, are available. There are four other ectocuneiforms, all from the right side : M. 18905, R.I, Rusinga, M. 18886, Rs.io5, Rusinga, M. 18885, Rs.30, Rusinga, and M. 18888, Rs., Rusinga. Of these, M. 18905 fits on to the navicular M. 18887, and belongs to the same individual. It is higher, but not wider than the ectocuneiform of D. leakeyi (Table 47). The ectocuneiform of D. sumatrensis is less elongated antero posteriorly than the fossil bones. That of D. primaevus (Arambourg 1959 : 72) appears to be rather low ; the width, given as 23 mm., is omitted in the Table as it is probably a misprint (53?). TABLE 47 Measurements of ectocuneiform of Dicerorhinus and Aceratherium (mm.) D. prima- D. suma- D. leakeyi M. 18905 M. 18886 M. 18885 M. 18888 evus trensis Anterior height 27 31 26 26 25 21-5 24 Anterior width 55 51 50 44 43 47 Ant. post. 53 51 52 50 48 60 43 diameter Only the right mesocuneiform of the skeleton of Dicerorhinus leakeyi no. 2, R.I, 1947, has been saved. It is more elongated anteroposteriorly than that in D. suma- trensis (Table 48). TABLE 48 Measurements of mesocuneiform of Dicerorhinus (mm.) Height Width Ant. post, diameter The left entocuneiform only of the skeleton of Dicerorhinus leakeyi no. 2, R.I, 1947, is available ; the proximal portion articulating with navicular and meso- cuneiform has broken off, and only the facet for Mt.II remains. The posterior D. leakeyi D. primaevus D. sumatrensis 19 21 16 22 26 19 41 29 178 MIOCENE RHINOCEROSES OF EAST AFRICA tuberosity is much larger than that in the entocuneiform of D. sumatrensis (in parentheses) : height 47 mm. (27 mm.), and width 30 mm. (22 mm.). All the metatarsals from the left side, as well as the Mt.II and IV from the right side of the skeleton of Dicerorhinus leakeyi no. 2, R.I, 1947, are preserved (PI. 13, fig. 3 ; PI. 15). No other entire metatarsals are in the East African Miocene collec- tion. Roger (1900 : 41) has tried to separate the metatarsal IV of Dicerorhinus from that of Aceratherium, and states that Mt.IV in Aceratherium has a postero-lateral incurvation of the large proximal cuboid facet, which makes this facet trilobate or trefoil-shaped, with the shaft forming a prominence behind the incurvation, whereas in Mt . IV of Dicerorhinus the cuboid facet is distinctly narrower and more antero- posteriorly elongated. My observations do not bear this out ; on the contrary the cuboid facet on Mt . IV in Dicerorhinus sumatrensis and that in D. leakeyi answer to the description of the facet of Roger's Aceratherium, and the cuboid facet of Mt . IV of Aceratherium tetradactylum associated with the dentition at Viehhausen is des- cribed by Rinnert (1956 : 34) as broadly elliptical. The right second metatarsal of the D. leakeyi skeleton shows an interesting patho- logical condition (PI. 15, figs. 1-3). The distal half is thickened, and the distal articulation wholly deformed. The swollen surface shows irregular growths all over and appears spongy. Without radiological or histological examination the attribu- tion of such an affection of the bone to a specific disease is impossible, but the out- ward appearance of the fossil is suggestive of something like Paget's disease. No other bones of the skeleton (nor the skull so far as preserved) appear to be afflicted with this disease (which may occur quite localized in the human skeleton) ; it may have developed of course in some of the missing elements like the metacarpals or the right tibia. The phalanges of this digit were certainly affected, but these are not present in the collection. The metatarsals of D. leakeyi are remarkable for their length. Mt.IV is longer than the Me. IV of Dicerorhinus or Aceratherium from the same Rusinga deposits (M. 18814). In D. primaevus (Arambourg 1959 : 72 and 68) Mt.IV is longer than Me. IV although Mt.III is shorter than Me. III. In D. sumatrensis there is only a slight difference in length between Mt . IV and Me . IV. In general, metatarsals are more shortened than the metacarpals within the same species. The metatarsals of D. schleiermacheri are unknown ; its metacarpals, however, are longer than those in orientalis, ringstroemi, and primaevus, and probably would have exceeded those of D. leakeyi in length. An Mt . Ill of D. orientalis from Pikermi recorded by Gaudry (1862-67 : 207> P^ 32> %• 9) nas a length only of 160 mm. by a greatest distal width of 52 mm., less than in D. leakeyi. The metatarsals of D. primaevus are also shorter than those of D. leakeyi. Mt . II and Mt . IV of D. primaevus are relatively less expanded distally than in D. sumatrensis, as is also the case with Mt . II and Mt . IV of D. leakeyi. In all three forms the median metatarsal is approx- imately 10% longer than the metatarsals on either side of it, as it is in Aceratherium incisivum recorded by Ringstrom (1924 : 192) and listed in the last column of Table 49. In a set of metatarsals from Budenheim recorded as A. incisivum by Kaup (1834 : 61, pi. 15, fig. 9) the lengths are less than those studied by Ringstrom. MIOCENE RHINOCEROSES OF EAST AFRICA 179 TABLE 49 Measurements of metatarsals of Dicerorhinus and Aceratherium (mm.) S £ i g B fl I ! 1* S i 11 'S 1 ^ *S >9 6 § So -^ 3 fe S -55 -I -8 .9 W o -IS P ^v Mt. II, median length 162 149 126 140 135 146 Proximal width 29 29 21 Prox. ant. post, diameter 48 34 36 Middle width 32 30 23 30 Middle ant. post, diameter 30 21 20 Greatest distal width 42 35 40 Width of distal trochlea 36 35 26-5 Distal ant. post, diameter 41 39 33 Ratio middle width /length 0-20 0-24 0-16 0-21 Mt. Ill, median length 180 165 144 168 150 158 Proximal width 57 53 Prox. ant. post, diameter 49 37 Middle width 50 41 42 35 Middle ant. post, diameter 25 19 Greatest distal width c. 60 52 53 46 Width of distal trochlea 51 44 Distal ant. post, diameter 41 38 Ratio middle width /length 0-28 0-28 0-25 0-22 Mt . IV, median length 160 147 126 160 140 126 140 145 Proximal width 44 41 33 4° Prox. ant. post, diameter 46 41 36 36 + Middle width 29 25 33 21 28 Middle ant. post, diameter 21 25 23 Greatest distal width 38 38 37 39 Width of distal trochlea 37 32 25 30 Distal ant. post, diameter 40 38 32 33 Ratio middle width /length 0-18 — 0-20 0-21 0-15 0-22 — The Mt.III and Mt.IV from Losodok described by Arambourg (1933 : ir, pi. i, figs. 4, 5) were noted to be nearly identical in dimensions with those of Aceratherium tetradactylum, but differing in their wider extremities and the shape of their articular surfaces. The fourth metatarsal Arambourg noted to be longer and more slender than that in the Sansan Aceratherium (tetradactylum}. Arambourg (1933) preferred to leave the generic identity of the Losodok bones uncertain, naming them " Acera- therium} sp.". The Mt.IV of D. leakeyi, it will be observed, corresponds with its homologue from Losodok in length and distal width, but has a more slender shaft. i8o MIOCENE RHINOCEROSES OF EAST AFRICA The Rusinga Mt . IV is intermediate in relative shaft width between a slender Mt . IV of Aceratherium tetradactylum from Viehhausen (Rinnert 1956 : 34) and one from Freimann recorded by Stromer (1928 : 29) as probably referable to A. tetradactylum. In the Aquitanian of Laugnac there is a very slender Mt.IV figured by Repelin (1917 : 40, pi. 6, figs. 3, 4) as " un type special de Rhinocerotide " that has a great resemblance to the same bone in A . tetradactylum (which is Vindobonian) ; its length is 130 mm., and its middle width only 20 mm., giving a ratio of 0-15. The Laugnac bone is found in the same deposits as Aceratherium lemanense ( — Teleoceras aginense Repelin: Lavocat 1951:114) that has less slender metapodials (Mt.IV length 99-103 mm., middle width 28-30 mm.). The Mt.III from Losodok nearly falls within the range of length of this bone in A . tetradactylum as given by Osborn (1900 : 246 : Mt . Ill 135-165 mm.) ; the Rusinga Mt . Ill is longer but relatively less slender. The Rusinga Mt.II is again less slender than that from Viehhausen recorded by Rinnert (1956 : 34, pi. 3, fig. 4), but is about equal in relative shaft width to that of A. incisivum as given by Ringstrom. A right Mt.II from the Upper Burdigalian of La Romieu figured by Roman & Viret (1934 : 36, pi. 9, fig. 12) is about 128 mm. in median length and very slender (no measurement given), just as is that from Vieh- hausen. The La Romieu bone has been identified only as " Ceratorhinus sp.? ". The conclusions from all this may only be that we are not able as yet to distinguish between the metapodials of Dicerorhinus and Aceratherium. Two proximal portions of right second metacarpals have to be recorded, viz., M. 18844, R.I, Rusinga (proximal width 32 mm., ant. post. 46 mm.), and M. 18847, Rs.io5, Rusinga (proximal width 29 mm., ant. post. 49 mm.). These bones are very much like their homologue in the skeleton of D. leakeyi. All of the phalanges of the left hind foot of the skeleton of Dicerorhinus leakeyi no. 2, R.I, 1947, are present (PI. 13, fig. 3), and only the third phalanx of the median digit is incomplete. Of the right hind foot there are the first and second phalanges of the median digit (PL 10, figs. 4, 5), and none of the other digits. Measurements will be found in Table 50. TABLE 50 Measurements of posterior phalanges of Dicerorhinus (mm.) D. leakeyi D. primaevus D. sumatrensis I III IV digit II III IV Phalanx I, length 37 40 33 42 31 35 30 Proximal width 40 55 38 48 38 47 35 Phalanx II, length 27 31 25 37 23 27 22 Proximal width 37 58 35 33 33 48 32 Phalanx III, length 33 — 30 Greatest diameter 60 58 In D. leakeyi as well as in D. sumatrensis the phalanges of digit IV are smaller than those of digit II. The width of the second phalanx of the lateral digit of D. primaevus MIOCENE RHINOCEROSES OF EAST AFRICA 181 (Arambourg 1959 : 69) is probably 43 instead of 33 mm. Comparison with Table 37, in which the measurements of the anterior phalanges are given, shows that the lateral digit phalanges are more reduced in size relative to those of the median digit in the hind foot than in the fore foot. To end the account of the foot skeleton of Dicerorhinus leakeyi mention should be made of the sesamoids. Some of the proximal sesamoid bones of the left hind foot are preserved in situ. The two situated behind the distal end of metatarsal III are 41 mm. long and 22 mm. wide ; those attached to metatarsal II are 32 mm. long by a width of 17 mm. In D. sumatrensis the proximal sesamoids have the same width (17 mm.) ; those of the median digit are 38 mm. long, and those of the second digit 30 mm. There remains a number of distal ends of metapodials, phalanges, and sesamoid bones enumerated here for the sake of completeness. The specimens definitely belonging to Brachypotherium have been sorted, and are recorded under B. heinzelini. Those listed in the following pages are either Dicerorhinus or Aceratherium. Distal ends of median metapodials (measurements in mm.) Greatest Trochlea Ant. post. width width diameter No. 430, Karungu, 1947 57 45 M.I 88 1 8, Rusinga 45 38 M. 18823, RS.IOI, Rusinga 51 M.I 8836, Ombo c. 53 c. 44 41 M. 18834, marked Nt 58 48 42 Distal ends of lateral metapodiais M. 18829, R.I, Rusinga 38 34 M. 18825, Rusinga 38 35 32 M. 18826, Rs. 31, Rusinga 39 37 34 M. 18821, Rusinga 34 31 M.I 8824, Rusinga 39 34 34 M. 18820, Rusinga 39 M. 18832, R.I, Rusinga 41 39 M. 18819, Rs. 8 1, Rusinga 35 M. 18833, R.I, Rusinga 43 40 39 M. 18816, Rs. 31, Rusinga 38 35 36 M. 18817, Rs. 31, Rusinga c. 39 M.I 883 1, Rusinga 32 33 M. 18827, Rs. 105, Rusinga 42 38 40 M. 18815, R.2, Rusinga 39 43 M. 1 8828, Rs. 3, Rusinga 40 35 35 M. 18835, Kachuku, Lower Series, 33 Karungu M. 18830, Rusinga 34 32 GEOL. 13, 2. 12 182 MIOCENE RHINOCEROSES OF EAST AFRICA Phalanx I, median digit (measurements in mm.) M.I 8858, R.I, Rusinga No. 938, 1947, Gumba red-beds, Rusinga No. 197, 1947, S. of Kiahera Hill, Rusinga No. 237, 1950, R.2-4, Rusinga M.I 8860, Rs.3i, Rusinga Phalanx II, median digit M.I 8863, Rs.3o, Rusinga M.I 8861, Rusinga M.I 8864, Rs.30, Rusinga M.I 8867, Ngira, Karungu No. 1060, S.E. of Kiahera Hill, Rusinga Phalanx I, lateral digit M.I 8856, Rusinga M.I 8857, Rusinga No. 1 152, R.2-4, Rusinga, 1950 M. 18853, Rs.3i, Rusinga M. 18868, Ngira, Karungu Phalanx II, lateral digit M. 18855, Rs. 104, Rusinga M.I 8866, Rs.30, Rusinga M.I 8865, Rusinga No. no, 1949, W. Hiwegi, Rusinga 27 29 24 25 22 Length 36 35 29 28 24 22 21 22 Prox. width 55 55 52 48 48 43 53 42 Prox. width 38 38 35 40 39 36 37 27 29 There remain one third phalanx of a median digit, no. 498, Rusinga, 1950 (length 24 mm., greatest width 63 mm.), a third phalanx of a lateral digit, M . 18852, Rusinga (length 40 mm., greatest diameter 64 mm.), and an incomplete third phalanx of a lateral digit, no. 845, Kathwanga, Rusinga (length 29 mm.). Proximal sesamoids, median digit Length Width M.I 887 1, Rusinga 48 25 No. 921, 1947, N. of Kiahera Hill, Rusinga 49 26 M.I 8869, Rs.6a, Rusinga 47 25 M. 18874, Rs. 2 1, Rusinga 45 24 No. 238, 1950, R.2-4, Rusinga 43 26 M. 18870, R.I, Rusinga 40 21 No. 536, Chianda Uyoma 40 24 M. 18873, Rs.2i, Rusinga 23 No. ? (possibly belonging to skeleton no. 2, R.I, 40 21 1947) MIOCENE RHINOCEROSES OF EAST AFRICA 183 Proximal sesamoids, lateral digit No. 239, 1950, R.2-4, Rusinga 34 16 M. 18872, Rs. 38, Rusinga 32 17 No. 820, 1947 (possibly belonging to skeleton no. 2, 31 20 R.I, 1947) No. 820, 1947 (idem, second specimen) 30 18 No. ? (idem) 30 17 No. ? (idem) 31 16 What is probably the distal sesamoid (situated behind the junction of the second and third phalanges) of the median digit (unnumbered, possibly belonging to skeleton no. 2, R. r, 1947) is 31 mm. wide transversely and 7 mm. high at the articular surface. The tail vertebrae labelled as belonging to the skeleton no. 2, R. I, 1947, of Dicero- rhinus leakeyi range from what is probably the fourth from the sacrum to nearly the last. The largest vertebra has only the left transverse process, 42 mm. in antero- posterior diameter (23 mm. in D. sumatrensis) , and an arch that appears to have been higher than the body but crushed dorso-ventrally. The spinous process has a thickened summit. The greatest width of the vertebra is about 95 mm. (73 mm.). The second largest caudal vertebra has both transverse processes, greatest width 82 mm. (69 mm.), but these processes are much reduced anteroposteriorly to 21 mm. (14 mm.). The small and distorted arch is bifid behind, and probably not higher than the body. An isolated double summit of an arch intermediate in size between the last and the vertebrae to be mentioned next indicates that the body of at least one caudal vertebra has been lost. The next has a body still as long as that of the second largest of the lot, viz., 39 mm. (29 mm.), only traces of a transverse process, and a very small arch, which was probably open dorsally. This vertebra is crushed laterally. Of the remaining thirteen caudal vertebrae only the largest two have two ridges dorsally, the others being without a trace of an arch. These vertebrae are not distorted and seem to form an unbroken series. The length and anterior height of the body decrease from 35 and 27 mm. in the first, over 26 and 12 mm. in the middle (seventh) of the series, to 16 and 8 mm. in the last. The caudal vertebra in D. sumatrensis that shows the same reduction of the arch as the anterior of our series of thirteen is the ninth caudal ; it is 25 mm. long and 23 mm. high anteriorly. The fifteenth caudal vertebra in D. sumatrensis is 26 mm. long and 14 mm. high, while the twenty-first is 19 mm. long and 7 mm. high. It is followed by three more vertebrae, the last one of which has a pointed end. From this comparison it follows that the tail vertebrae of D. leakeyi reduce in length more rapidly as one passes along the series than in the corresponding section of the tail of D. sumatrensis, that the relative anterior height is greater half-way along the tail in D. sumatrensis, but that it diminishes more rapidly toward the end than in D. leakeyi. III. DISTRIBUTION OF RHINOCEROSES OVER EAST AFRICAN MIOCENE SITES In Table 51 are given the locality records of those specimens of which the generic position has been determined. GEOL. 13, 2. I2§ 184 MIOCENE RHINOCEROSES OF EAST AFRICA TABLE 51 Distribution of Rhinocerotidae in the East African Miocene Dicerorhinus Aceratherium Brachypotherium Chilotherium Rusinga X X X (no sub-site given) R.I x x . R.i-ia R.2 x R.2-4 X Rs.s X Rs.6a . . x V R.y, Rs.y R.n Rs.26 x R-73 RS.QI RS.IOI . R.loy Rs.ioS X . Gumba XX X West side of Hiwegi X . X Kamasengere x Kathwanga x X S. of Kiahera Hill . X . S.E. of Kiahera Hill x Wakondu . .X Karungu (Andrews . X . 1914 and 1937) Ngira, Karungu .XX Songhor X . Moruaret Hill . X . (Deraniyagala) Loperot . .X Napak I x II A and C X V X ... VI X ... Generically uncertain material of Rhinocerotidae has also been obtained from sub- sites of Rusinga and other sites in Kenya whence no generically identifiable rhinoceros specimens have come, as follows : R.4 (astragalus), R.8 (cuboid), Rs.2i (radius, astragalus, two proximal sesa- moids), Rs.23a (scapula), Rs.30 (ectocuneiform, two phalanges II of median digits, phalanx II of lateral digit), Rs.3i (radius, unciform, Mc.III-IV, patella, tibia, astragalus, three lateral metapodials, phalanx I of median digit, phalanx II of lateral digit), Rs.38 (astragalus, proximal sesamoid), Rs.Si (lateral metapodial), Rs.iO3 (Me. Ill), Rs.io4 (phalanx II of lateral digit), Rs.io5 (Me. II, tibia, cuboid, ecto- cuneiform, lateral metapodial), R.io6 (astragalus, calcaneum), Kiahera Hill (P2), N. of Kiahera Hill (proximal sesamoid), Kiangata (lower C), Kiune (astragalus), MIOCENE RHINOCEROSES OF EAST AFRICA 185 and the following sites other than Rusinga : Aloir, 1939 (mandibular ramus and astragalus), Chianda Uyoma (proximal sesamoid), Ombo (median metapodial), Maboko (= Kiboko) Island (two scaphoids, tibia, astragalus), Kachuku, Lower Series, Karungu (Mc.II-III, cuboid, lateral metapodial) , Kachuku, Karungu (cuboid), Losodok (Arambourg, 1933) (lower M, axis, astragalus, Mt.III-IV, three phalanges I of median and lateral digits) , and Arongo Chianda, 25 . x . 1939 (astragalus) . I have not seen any material from Tambach, and am unable to confirm the record of rhinoceros from that locality (cf. Le Gros Clark & Leakey 1951 15). In addition to the nine sites in Kenya from which Rhinocerotidae have been recorded in 1951 there are Aloir, 1939, and Arongo Chianda, 25. x. 1939, both with a generically unidenti- fiable non-brachypothere astragalus. All groups of mammals found in Rusinga are known to be represented at Mfwanganu Island, Kenya, except for the rhinoceroses and insectivores (Whitworth 1961), and indeed the only rhinoceros-like specimen that I have seen from that island, a proximal metapodial fragment, is Brachyodus aequa- torialis Maclnnes, the large Rusinga anthracothere, which will be reported later. IV. TIME PLACEMENT OF THE MIOCENE EAST AFRICAN FAUNAS The Miocene faunas of East Africa are generally regarded as Early Miocene, corresponding to the Burdigalian stage of Europe. Dr. Leakey kindly informs me that the geology of Rusinga is much more complicated than had been previously thought, and not all of it may be of the same age. Loperot, at present being in- vestigated by the Harvard Expedition, is a considerable area with many different sites that may not be contemporaneous. Most of the Loperot sites are probably much younger than most of Rusinga. Further studies on elements of the Proconsul fauna are being undertaken. Potassium- Argon dates have been published during the last few years, and are still being worked on, and these have not invariably had the mammalian palaeontologists' approval. From a number of K/A dates for Rusinga, including two of over 100 million years (!), Evernden, Savage, Curtis & James (1964 : 176 : KA 336) consider 15-3 ± 1*5 million years the best estimate and only meaningful age ; this would approximately correspond with Late Miocene (Vindobonian) . However, the age that has recently been determined for Napak, Uganda, viz., 19 i 2 million years (Bishop 1964) points to Early Miocene (Burdigalian). A Middle Miocene age for Rusinga has been proposed on faunal grounds by Thenius (1959 : 268), and the geological setting of the Western Rift deposits of Congo, whence a typical Rusinga fauna has been described (Hooijer 1963), even leaves room for a Late Miocene age of part of the fauna. The slightly different faunules of Malembe and Bololo in the Atlantic coastal region of Congo are Burdigalian as the associated fish fauna indicates (Hooijer 1963 : 5, 64). Radiometric dates are not as yet available for the various Western and Eastern Congo sites. What now is the bearing of the Rhinocerotidae of Rusinga and Napak on the problem of the age of these deposits? Let us summarize the salient characters and similarities to European Tertiary rhinoceroses. 186 MIOCENE RHINOCEROSES OF EAST AFRICA Of the four genera and species of Rhinocerotidae from Rusinga and other Miocene sites in East Africa two are in keeping with either Burdigalian or Vindobonian, and two rather with Vindobonian, in the European sense. Aceratherium acutirostratum is unique in the combination of a shallow naso-maxillary notch (Aquitanian in Europe) and an elevated occiput (Pontian in Europe) . Its teeth are more advanced in struc- ture than those in the European Oligocene forms ; either Burdigalian or Vindobonian would seem fitting for this species. Chilothenum, now found for the first time in Africa (the two Rusinga M need not be specifically the same as the Loperot M3)1, ranging from Burdigalian into the Pontian in Asia and from Vindobonian into Pontian in Europe, could be either Burdigalian or Vindobonian as well ; the earliest chilotheres are as fully-fledged as the Pontian (Cooper 1934 : 596). Dicerorhinus Leakey i has the skull shape of the Vindobonian D. sansaniensis although it is larger, and its teeth agree in characters with those of this as well as some larger Pontian forms. It has no close relationship with tapir-sized, slender-footed Aquitanian and Burdigalian D. tagicus. Brachypotherium heinzelini almost duplicates the Late Vindobonian B. brachypus ; only its lateral metacarpals are relatively shorter and wider. It is definitely more advanced in progressive metapodial abbreviation than the Moghara B. snowi, which has only reached the stage of the Late Burdigalian and Early Vindobonian B. stehlini. The four forms occur together in the Gumba beds of Rusinga. The same assemblage of rhinoceroses, except for the rarest Chilotherium sp., occurs at Napak, K/A dated as Early Miocene, Burdigalian. The fauna of Napak is exceedingly similar to that of the Kenya sites ; in the latest survey of the fauna (Bishop 1962) this was brought out by various specialists. Rhinocerotidae and Anthracotheriidae were not mentioned in the 1962 paper as no data were available at the time. Among the dental material from Napak kindly sent to me from time to time by Dr. W. W. Bishop there is a very characteristic upper molar of Brachyodus aequatorialis Maclnnes (1951), indistinguishable from the Rusinga type. The speci- men originates from Napak II C, and other from Napak V and VIII, and from Moroto I and II ; this will be described later. Thus, the faunal likeness between Napak and Rusinga is further enhanced by the Anthracotheriidae as well as by the Rhinocerotidae. The fauna of Fort Ternan, a site already famous for Kenyapithecus wickeri Leakey (1962), more advanced than Proconsul, has a totally different aspect. It comprises a small Trilophodon and a suid more evolved than the Rusinga forms, ruminants with incipient horns (unknown in the Miocene), and a highly intriguing form transitional between Brachyodus and hippo (Leakey, in Howell & Bourliere (editors), 1963 : 554). Anthracotheres are considered ancestral to hippopotami ; for these animals no other ancestry can be made plausible. Brachyodus occurs in the Burdigalian of Europe, and the first Hippopotamus appears in the Pontian of Europe (Hooijer 19460 ; Aguirre 1963). The Fort Ternan anthracothere or ancestral hippopotamus, therefore, would best be accorded a Vindobonian or very early Pontian age. Now this is just what the radiometric datings indicate : ± 12 million yearsj (Leakey, in Howell & Bourliere 1 The results of the extensive Harvard Expedition to Loperot, which include parts of four skeletons, will be reserved for a later paper. MIOCENE RHINOCEROSES OF EAST AFRICA 187 1963 : 554), 14 million years (Leakey 1963 : 138 ; Evernden, Savage, Curtis & James 1964 : 174). We might therefore say that Fort Ternan has been K/A dated to the satisfaction of the mammalian palaeontologist. The rhinoceroses of Fort Ternan have not yet been described, but with this fauna are bound to be different from the Rusinga and Napak species ; their study is eagerly awaited. The " best estimate " of a date for Rusinga, 15-3 i i'5 million years as pro- pounded by Evernden et al., would seem to differ too little from that of Fort Ternan (12-14 million years) for such a faunal change to have taken place. A date like that of Napak (19 i 2 million years) seems much more fitting for Rusinga, and is in keeping with palaeontological data. A fauna cannot well remain virtually unchanged for a period of several million years (if the difference in K/A dates for Napak and Rusinga amounts to that much ; both have appreciable standard errors, and the difference may be more apparent than real). One might therefore well wonder whether Rusinga has not been considered too young. The Rhinocerotidae of Rusinga and Napak, as we have seen, would broadly corres- pond with the Burdigalian and Vindobonian stages in Europe. Intercontinental correlation on forms of this kind is, however, only approximate. None of the East African species is identical with any in Europe ; they probably were products of independent evolution in Africa although contemporaneous in origin with those of Eurasia. The pre-Miocene history of the rhinoceroses in Africa is sadly unknown ; none are, for example, found in the Fayum Series, at which times there had been faunal interchange between Africa and Eurasia. In Africa, rhinoceroses appear first at the Rusinga stage (unless the so-called Burdigalian fauna of Moghara, Egypt, which shows little affinity to that of Rusinga, is older). It is feasible that Dicerorhinus leakeyi and Br achy pother ium heinzelini represent more progressive evolutionary stages than the forms living at the same time in Europe (and North Africa, witness the Brachypotherium of Moghara), and actually are as old as the Napak K/A date indicates. Exact correlations cannot be made on the fauna so far as known. We may say that the East African Miocene fauna is approximately equivalent to the Burdigalian of Europe, but application of this Deperetian term to the East African faunal stage may easily impart a false sense of precision. At this stage, all that can be said is that most of the Rusinga sites are tentatively accepted as correlative with the Burdigalian, the Lower Miocene of Europe, but that some sites on the island and elsewhere in East Africa appear to be younger, later Miocene or even Pliocene. V. REFERENCES AGUIRRE, E. DE. 1963. Hippopotamus crusafonti n. sp. del Pliocene inferior de Arenas de Rey (Granada). Notas Inst. geol. Esp., Madrid, 69 : 215-230, pis. 1-3. ANDREWS, C. W. 1914. On the Lower Miocene Vertebrates from British East Africa collected by Dr. Felix Oswald. Quart. J. geol. Soc. Lond., 70 : 163-186, pis. 1-3. ARAMBOURG, C. 1933. Mammiferes miocenes du Turkana, Afrique Orientale. Ann, Paleont., Paris, 22 : 121-147, pis. i, 2. i88 MIOCENE RHINOCEROSES OF EAST AFRICA ARAMBOURG, C. 1959. Vertebres continentaux du Miocene superieur de 1'Afrique du Nord. Publs. Sew. Carte geol. Alger. (n.s.) Paleont., 4 : 161 pp., 18 pis. BISHOP, W. W. 1958. Miocene Mammalia from the Napak volcanics, Karamoja, Uganda. Nature, Lond., 182 : 1480-1482, 3 figs. — 1962. The mammalian fauna and geomorphological relations of the Napak volcanics, Karamoja. Rec. geol. Surv. Uganda, Entebbe, 1957-1958 : 1-18, pis. 1-5. — 1964. More fossil Primates and other Miocene mammals from north-east Uganda. Nature, Lond., 203 : 1327-1331, fig. BOHLIN, B. 1946. The fossil mammals from the Tertiary deposit of Taben-buluk, Western Kansu, Part II. Palaeont. Sinica, Peking (n.s.), C 8B : 259 pp., 9 pis. BORISSIAK, A. A. 1938. A new Dicerorhinus from the Middle Miocene of North Caucasus. Trav. Inst. Paleozool. Acad. Sci. URSS, Leningrad, 8, 2 : 1-68, pis. i, 2. BREUNING, S. 1924. Beitrage zur Stammesgeschichte der Rhinocerotidae. Verh. zool.-bot. Ges. Wien, 73 : 5-46, 36 figs. COLE, S. M. 1950. An outline of the geology of Kenya, xiii + 58 pp., 5 pis. Nairobi. COOPER, C. FORSTER 1934. The extinct rhinoceroses of Baluchistan. Philos. Trans., London (B) 223 : 569-616, pis. 64-67. DEPERET, C. 1887. Recherches sur la succession des faunes de vertebres miocenes de la Vallee du Rhone. Arch. Mus. Hist. not. Lyon, 4 : 45-313, pis. 12-25. DERANIYAGALA, P. E. P. 1951. A Hornless Rhinoceros from the Mio-Pliocene Deposits of East Africa. Spolia Zeylan., Colombo, 26, 2 : 133-135, pi. i. — 1 95 1 a. A new Genus and Species of Hornless Mio-Pliocene Rhinoceros. Proc. Ceylon Ass. Sci., jth Ann. Congr., 2 : 24. — 1953. The Extinct and Living Subfamilies of African Rhinocerotidae. Spolia Zeylan., Colombo, 27 : 13-14, 3 pis. DUVERNOY, M. 1853. Nouvelles etudes sur les Rhinoceros fossiles. Arch. Mus. Hist. nat. Paris, 7 : 1-144, 8 pis. EVERNDEN, J. F., SAVAGE, D. E., CURTIS, G. H. & JAMES, G. T. 1964. Potassium-Argon dates and the Cenozoic mammalian chronology of North America. Amer. J. Sci., New Haven, 262 : 145-198, fig. FILHOL, H. 1891. Etudes sur les mammiferes fossiles de Sansan. Ann. Sci. geol. Paris, 21 : 1-319, 46 pis. FLOWER, W. H. 1876. On some cranial and dental characters of the existing species of rhinoceroses. Proc. zool. Soc. Lond., 1876 : 443-457, 4 figs. FOURTAU, R. 1920. Contribution a I'etude des vertebres miocenes de I'Egypte. xii + 122 pp., 3 pis. Cairo (Govt Press). FRAAS, O. 1870. Die Fauna von Steinheim. Jh. Ver. vaterl. Naturk. Wurtt., Stuttgart, 26 : 145-306, pis. 4-13. GAUDRY, A. 1862-1867. Animaux fossiles et Geologic del' Attique. 476 pp., atlas 75 pis. Paris. HOOIJER, D. A. 1946. Prehistoric and fossil rhinoceroses from the Malay Archipelago and India. Zool. Meded., Leiden, 26 : 1-138, 10 pis. 1946. The evolution of the skeleton of Rhinoceros sondaicus Desmarest. Proc. Acad. Sci. Amst., 49 : 671-676. — 1946. Notes on some Pontian mammals from Sicily, figured by Seguenza. Archs. Neerl. Zool., Leiden, 7 : 301-333, 2 figs. 1963. Miocene Mammalia of Congo (with a chapter by A. Gautier & J. Lepersonne). Ann. Mus. r. Afr. centr., Tervuren (8°) Sci. Geol., 46. ix -f 77 pp., 10 pis. HOWELL, F. C. & BOURLIERE, F. (editors). 1963. African Ecology and Human Evolution. viii + 666 pp., figs. Chicago. KAUP, J. J. 1834. Description d'ossements fossiles de mammiferes inconnus jusqu'a present, qui se trouvent au Museum grand-ducal de Darmstadt, III : 33-64, atlas pis. 10-15. Darmstadt. 1854. Beitrage zur naheren Kenntniss der urweltlichen Saeugethiere, I. viii + 31 PP-, 10 pis. Darmstadt. MIOCENE RHINOCEROSES OF EAST AFRICA 189 KAUP, J. J. 1859. Uber den vierten Finger des Aceratherium incisivum. N. Jb. Min. Geogn. Geol. Petrefact., Stuttgart, 1859 : 163-167, pi. 2. LAVOCAT, R. 1951. Revision de la faune des mammiferes oligocenes d'Auvergne et du Velay (ed. Sciences et Avenir) : 1-153, 26 pis. Paris. LEAKEY, L. S. B. 1962. A new Lower Pliocene fossil Primate from Kenya. Ann. Mag. Nat. Hist., London (13) 4 : 689-696, pi. 18. — 1963. Adventures in the search for Man. Nat. Geogr. Mag., Washington, 123 : 132-152, 7 figs. LE GROS CLARK, W. E. & LEAKEY, L. S. B. 1951. The Miocene Hominoidea of East Africa. Fossil Mammals of Africa, 1. v + 117 PP-, 9 pis. British Museum (Nat. Hist.), London. MACINNES, D. G. 1951. Miocene Anthracotheriidae from East Africa. Fossil Mammals of Africa, 4. 24 pp., 4 pis. British Museum (Nat. Hist.), London. MAYET, L. 1908. Etude des mammiferes miocenes des Sables de 1'Orleanais et des faluns de la Touraine. Ann. Univ. Lyon (n.s.) 1, 24 : 336 pp., 12 pis. — 1909. Etude sommaire des mammiferes fossiles des faluns de la Touraine proprement dite. Ann. Univ. Lyon (n.s.) 1, 26 : 72 pp., 30 figs. MERMIER, E. 1895. Sur la decouverte d'une nouvelle espece d' 'Aceratherium dans la molasse burdigalienne du Royans. Ann. Soc. linn. Lyon, 42 : 163-189, i pi. — 1896. Etude complementaire sur \' Aceratherium platyodon de la molasse burdigalienne superieure des environs de Saint-Nazaire en Royans (Drome). Ann. Soc. linn. Lyon, 43 : 225-2/10, 2 pis. OSBORN, H. F. 1898. The extinct rhinoceroses. Mem. Amer. Mus. nat. Hist., New York, 1 : 75-164, pis. I2A-2O. — 1899. Frontal horn on Aceratherium incisivum. Relation of this type to Elasmotherium. Science, New York (n.s.) 9 : 161-162, i pi. 1900. Phylogeny of the rhinoceroses of Europe. Bull. Amer. Mus. nat. Hist., New York, 13 : 229-267, 1 6 figs. — 1910. The Age of Mammals in Europe, Asia and North America, xvii + 635 pp., 220 figs. New York. PAVLOW, M. 1892. Etudes sur 1'histoire paleontologique des Ongules, VI. Les Rhinoceridae de la Russie et le developpement des Rhinoceridae en general. Bull. Soc. Nat. Moscou (2) 6 : 137-221, pis. 3-5. REPELIN, J. 1917. Les Rhinoc6rotides de 1'Aquitanian superieur de 1'Agenais (Laugnac). Ann. Mus. Hist. nat. Marseille, 16 : 1-45, 14 pis. RINGSTROM, T. 1924. Nashorner der Hipparion-Fauna Nord-Chinas. Palaeont. Sinica, Peking (C) 1, 4 : 1-156, 12 pis. RINNERT, P. 1956. Die Huftiere aus dern Braunkohlenmiozan der Oberfalz. Palaeonto- graphica, Stuttgart, 107, A, i : 1-65, 6 pis. ROGER, O. 1900. Ueber Rhinoceros Goldfussi Kaup, und die anderen gleichzeitigen Rhino- cerosarten. Ber. naturw. Ver. Schwaben, Augsburg, 34 : 1-52, 2 pis. (explanation p. 69). ROMAN, F. 1912. Les Rhinocerides de 1'Oligocene d'Europe. Arch. Mus. Hist. nat. Lyon, 11, 2 : 1-92, 10 pis. 1924. Contribution a l'£tude de la faune de mammiferes des Littorinenkalk (Oligocene superieur) du Bassin de Mayence. Les Rhinoceros. Trav. Lab. Geol. Univ. Lyon (7) 6 : 1-54, 5 Pis. ROMAN, F. & VIRET, J. 1934. La Faune de mammiferes du Burdigalien de la Romieu (Gers). Mem. Soc. Geol. Fr., Paris (n.s.) 9, 21 : 1-67, 12 pis. STEHLIN, H. G. 1925. Catalogue des ossements de mammiferes tertiaires de la collection Bourgeois. Butt. Soc. Hist, nat., Loir-et-Cher , 18 : 77-277, 31 figs. (Reprint : 1-205). STROMER, E. 1928. Wirbeltiere im obermiocanen Flinz Miinchens. Abh. bayer. Akad. Wiss., Miinchen, 32 : 1-71, 3 pis. THENIUS, E. 1959. Tertiar. II. Wirbelttierfaunen. Handbuch der stratigraphischen Geologic, 3, 2 : xi + 328 pp., 10 pis. Stuttgart. igo MIOCENE RHINOCEROSES OF EAST AFRICA TOBIEN, H. 1956. Zur Okologie der jungtertiaren Saugetiere vom Howenegg (Hegau) und zur Biostratigraphie der europaischen Hipparion-Fauna. Schr. Ver. Gesch. Naturg. Baar, Tubingen, 24 : 208-223, 3 fig3- TOULA, F. 1902. Das Nashorn von Hundsheim. Rhinoceros (Ceratorhinus Osborn) hunds- heimensis nov. form. Abh. geol. Reichsanst., Wien, 19 : 1-92, 12 pis. VILLALTA, J. F. DE & CRUSAFONT, M. 1955- Chilotheriuwi quintanelensis Zbysz., sinonimo de Hispanotherium matritensis (Prado). Notas Inst. geol. Esp., Madrid, 37 : 1-9, 4 figs. VIRET, J. 1929, Les faunes de mammiferes de 1'Oligocene superieur de la Limagne bourbon- nais. Ann. Univ. Lyon (n.s.) 47 : 1-328, 32 pis. — 1961. Catalogue critique de la faune des mammiferes miocenes de la Grive Saint- Alban (Isere) (suite du fascicule III. 1951). N. Arch. Mus. Hist. nat. Lyon, 6 : 53-81, pi. 5. WANG, K. M. 1928. Die obermiozanen Rhinocerotiden von Bayern. Palaont. Z., Berlin, 10 : 184-212, pis. 7-10. WHITWORTH, T. 1958. Miocene ruminants of East Africa. Fossil Mammals of Africa, 15. 50 pp., 18 figs. British Museum (Nat. Hist.), London. 1961. The geology of Mfwanganu Island, Western Kenya. Overseas Geol. Min. Resour., London, 8, 2 : 150-190, i pi. WOOD, H. E. 1927. Some Early Tertiary rhinoceroses and hydracodonts. Bull. Amer. Paleont., Ithaca, 13, 50: 1-89, 7 pis. 1934- Revision of the Hyrachyidae. Bull. Amer. Mus. nat. Hist., New York, 67 : 181- 295, pis. 20-24. 1963. A primitive rhinoceros from the Late Eocene of Mongolia. Amer. Mus. Novit., New York, 2146 : n pp., 2 figs. YOUNG, C. C. 1937. On a Miocene Mammalian Fauna from Shantung. Bull. Geol. Soc. China, Peking, 17 : 209-244, 3 pis. a Bull. BM. (N.H.) Geol. 13, 2 PLATE 1 PLATE 2 Dicerorhinus leakeyi sp. nov. FIG. i. Holotype skull, Rusinga, 1935, lower view. x£. FIG. 2. Holotype mandible, Rusinga, 1935, upper view, FIG. 3. P2-M3 dext., no. 2, R.I, 1947, crown view. xf. FIG. 4. Mandible, no. 2, R.I, 1947, right view. xj. Bull. B.M. (N.H.) Geol. 13, 2 PLATE 2 '-** Bull. B.M. (N.H.) Geol. 13, 2 PLATE 3 w U) TJ ' j.J *rj rij hH I— t I-H t— * O O O O • • w • OO^-J ' M $X^ i- fl> 1 CJ p M JU § 1 1 § 13 ts n> H <-h w j— i p H -^ & M crq' o' >-i 0 0 0) X r-t- M 0 B h2 »• a 0 1950, external \ Brachypoth oo Oi p M M Aceratherium a 01 0 Hh W s 03. ? external view. c-h rt- O h- 1 CO ^ X 3 • S- ^ V** ^— ^ H Jj HH g £^ o ffi 8 ^d $ 5 1 CD *% i .3 i- Cfd i Bull. BM. (N.H.) Geol. 13, 2 PLATE 4 -• '"':^' . t; . > ^ PLATE 5 Dicerorhinus leakeyi sp. nov. FIG. i. DM1'4 sin., Rs.26, crown view. xi. FIG. 2. ps-4 sjn-) no> 2( R.I; 1947, crown view. xf. FIG. 3. Same, internal view. xf. Aceratherium acutirostratum (Deraniyagala) FIG. 4. DM4 dext., no. 142, Kamasengere, Rusinga, 1949, crown view. FIG. 5. DM4 sin., no. 218, R.ioy, 1948, crown view. xf. Bull. B.M. (N.H.) Geol. 13, 2 PLATE 5 PLATE 6 Aceratherium acutirostratum (Deraniyagala) FIG. i. P4 sin., no. 232, R.2-4, 1950, crown view. x£. FIG. 2. Same, internal view. xf. FIG. 3. . P4 dext. with portion of M1, no. 231, R.2-4, 1950, crown view, x f. FIG. 4. Same, internal view, x £ . Br achy pot her ium heinzelini Hooijer FIG. 5. M1 dext., Karungu, 1937, crown view. xf. FIG. 6. Same, external view, x £. FIG. 9. Posterior portion of Mx or M2 sin., no. 546, R. i, 1949, crown view, x f. Dicerorhinus leakeyi sp. nov. FIG. 7. M2 sin., no. 1161, R.I, 1950, crown view. xf. FIG. 8. M.Z sin., no. 2, R.I, 1947, crown view. xf. FIG. 12. P3 dext., no. 1385, Rusinga, 1951. Crown view. x$. Chilotheriutn sp. FIG. 10. M1 or M2 dext., no. 695, Gumba, Rusinga, 1949, crown view. xf. FIG. ii. Protoloph of right upper M, no. 506, Wakondu, Rusinga, 1950, crown view. x f. Bull. B.M. (N.H.) Geol. 13, 2 PLATE 6 II *rcj ^rj ^rj hcj O O O O O\ Oi 4* uo 3 H 3 en " CL " CL P g' » p .* j? <-t- W rh " r+- " C6 . CD hH (B |— I 1-1 H 3 C* ""* n O X ^ M X * A w •». ? 51 ? 3 3 8 X c' 3 5* > <% tfl a- 1 1 1 " T3 Bull. B.M. (N.H.) Geol. 13, 2 PLATE 7 p p p p p p CTi Oi 4». U> M W P f§ a I-L. - j» f P » B i a i M p * p i co_ 22. » * H r rt- f1" r+- XH — i (t - fD a P P g 0 g o -a >o B p § * p ^ < ^ < ,§ h-j M So 0 V> O v ^ ^ x co 4>. 5 ^^ q -fc ° S 2 VO Hf, g ^ o ffi ^ ^ 3 ^ «• ^- § 1 1? to < - o ^' c x x s* CO cnMk_i^ J< O 5 3- H S w 3- »«. K Bull. EM. (N.H.) Geol. 13, 2 PLATE 8 Hj ^ h-( hH P P 4* OJ P P a w tC £D [1. ^ . & S O 00 oo p Oi p ^j M M M 4^ 3 p to P 0 p M P 5' CL CD rt- H 00 O 0 f c ^o "^J Tl n> ^ < < o en o g " s F s q 5 a. - ^52- |.p?&? s' g .» p- s ^. K> r^ " K) 5 • • 3 " .^ ^-H P » MM' H *£> - • X M^ ^M «*» o " • vo ' 4^- QJ H -|i' 5 p - ^ rt- M „ IK* a «D fi 2. 3.-* » ?^ ? p > p g- 1-1 a. » 2 < a> n- ^ M. 5 2 S. ^ 03 i ^ b X T) 5* t-1 X 3* ' ^ iH^ g H *" W J5* H J^ OO »• a $ C/J y 0 < Bull. B.M. (N.H.) Geol. 13, 2 PLATE 13 ^&£ft*> PLATE 14 Dicerorhinus leakeyi sp. nov. FIG. i. Astragalus and calcaneum sin., no. 2, R.I, 1947, anterior view. xf. Dicerorhinus or Aceratherium sp. FIG. 2. Astragalus sin., Aloir, 1939, anterior view. xf. FIG. 6. Scaphoid sin., M. 18897, Karungu, anterior view. xf. FIG. 7. Unciform sin., M. 25191, Kathwanga, Rusinga, anterior view. xf. Brachypotherium heinzelini Hooijer FIG. 3. Astragalus sin., no. 538, Gumba, Rusinga, 1949, anterior view. xf. Aceratherium acutirostratum (Deraniyagala) FIG. 4. Scaphoid, lunar and cuneiform dext., no. 850, R.I, 1947, associated, anterior view. Xf. FIG. 5. Same, proximal view. xf. Bull. B.M. (N.H.) Geol. 13, 2 PLATE 14 p p p p -^ OJ 10 M g C/) C/5 O P5 P g|g &e« t?" X O i^i SS' «> P • 1 8- §' 8 g x .^ r -M ' X ? ITI «*» M p X E 1 2! t-1 3 > § H , W Bull. B.M. (N.H.) Geol. 13, 2 PLATE 15 PRINTED IN GREAT BRITAIN BY ADLARD & SON LIMITED BARTHOLOMEW PRESS, DORKING THYSANOPELTIDAE (TRILOBITA) FROM THE BRITISH DEVONIAN E. B. SELWOOD BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 13 No. 3 LONDON: 1966 THYSANOPELTIDAE (TRILOBITA) FROM THE* BRITISH DEVONIAN BY EDWIN BRIAN SELWOOD, Ph.D. University of Exeter \, I Pp. 191-220 ; 3 Plates ; 3 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 13 No. 3 LONDON: 1966 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 ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 13, No. 3 of the Geological (Palaeontological) series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. Trustees of the British Museum (Natural History) 1966 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 21 September, 1966 Pnce £i zs. THYSANOPELTIDAE (TRILOBITA) FROM THE BRITISH DEVONIAN By E. B. SELWOOD CONTENTS Page I. INTRODUCTION ......... 193 II. ACKNOWLEDGMENTS ........ 194 III. OCCURRENCE AND RANGE ....... 194 IV. VARIABILITY . . . . . . . . . .195 V. ECOLOGY .......... 202 VI. DETERIORATION OF THE BASAL PART OF THE GLABELLA . . 203 VII. SYSTEMATIC DESCRIPTIONS ....... 204 VIII. REFERENCES . . . . . . . . .219 SYNOPSIS A revision of the British Devonian Thysanopeltidae is made and the following species de- scribed : Scutellum (Scutellum) costatum costatum Pusch, S. (Scutellum) costatum whidbornei subsp. n., S. (Scutellum) costatum lummatonensis subsp. n., S. (Scutellum) delicatum delicatum (Whidborne), S. (Scutellum) delicatum tigrinum (Whidborne), 5. (Scutellum) pardalios (Whid- borne), S. (Scutellum) flabelliferum (Goldfuss). With the exception of S. (Scutellum) costatum costatum, which is known only from Upper Devonian localities, and 5. (Scutellum) flabelliferum, which is of Upper Couvinian age, the fauna comes principally from two localities of Givetian age, namely Wolborough Quarry (Zone of Maenioceras molarium), near Newton Abbot, and Lummaton Quarry (Zone of Maenioceras terebratum), Torquay. The Givetian material is abundant and shows considerable variability ; this is described and the evidence suggests that the species represented are closely related. Two lineages are recognized within the variants of both S. (Scutellum) costatum and S. (Scutellum) delicatum from the Middle Devonian ; these are distinguished by separate subspecific names. I. INTRODUCTION THYSANOPELTID trilobites from the Middle Devonian were first noted by Phillips (1841), who referred specimens from " Newton " to Bronteus flabettifer Goldfuss. Later, in a major study of the Devonian faunas, Whidborne (18890) included details of all species and localities then known. From material assembled principally from Chercombe Bridge and Wolborough, near Newton Abbot, and Lummaton, near Torquay, he recognized six species of which three (Bronteus pardalios, B. tigrinus, and B. delicatus) were new. Passing reference to Whidborne's work has been made in a number of publications on Continental faunas, but, so far as is known, no specimens have been held to be conspecific with his species. Additional records of Bronteus are included in the Newton Abbot Memoir of the Geological Survey (Ussher 1913). With the exception of Scutellum alutaceum (Goldfuss), which is not well established, the present investigation has confirmed the diversity of species noted by Whidborne. GEOL. 13, 3. 13 194 BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) At the same time, it is evident that Whidborne only selected for description certain specimens which could be readily distinguished from species established earlier, and by no means fully described the variability of each species. Sufficient material is now available to permit a quantitative investigation of the variation, and the species from Devonshire are here redescribed in the light of this variation. The terminology employed is essentially that used in the Treatise of Invertebrate Paleontology, (0) Arthropoda I, but is also supplemented by that of Snajdr (1960). The latter work (an account of the thysanopeltid fauna of Bohemia) includes an interpretation and terminology of the lateral glabellar markings, which differs from that given by R. & E. Richter and used in the Treatise, Snajdr's interpretation is based primarily on the nature and evolution of the muscle impressions seen on the inner surface of the exoskeleton beneath the glabellar furrows. These details cannot be recognized on the material described, but the interpretation appears well founded and is followed here. Studies of the British material suggest that the broad (sag.) occipital furrow, which is common within the Thysanopeltidae, is double and contains a preoccipital glabellar lobe. This, with its implication of a hitherto unrecognized glabellar furrow, further complicates the terminological difficulties. The nomenclature for the glabellar furrows is set out below. Richter Snajdr This paper 3P 3g 4P Supplementary groove 2g 3p 2p ig 2p IP ip (preoccipital glabellar furrow) II. ACKNOWLEDGMENTS I wish to thank particularly Dr. W. T. Dean, Dr. R. Goldring, and Professor S. Simpson, who read and criticized the original manuscript, and Mr. J. Saunders, technician in the Department of Geology at Exeter, who is responsible for the photographs. Dr. V. Loudon and Dr. M. Stone offered helpful advice on statistical problems. Dr. R. C. Blackie (Royal Albert Memorial Museum, Exeter), Mr. A. G. Brighton (Sedgwick Museum, Cambridge), Dr. M. L. K. Curtis (City Museum, Bristol), Dr. W. T. Dean (British Museum (Natural History)), Dr. R. J. G. Savage (Geology Museum, Bristol University) and Dr. F. S. Wallis (Torquay Natural History Museum) kindly permitted the examination and arranged the loan of specimens in their care. III. OCCURRENCE AND RANGE The provenance of much of the museum material is incomplete and frequently ambiguous. However, almost certainly only four main localities are involved : Wolborough Quarry and Chercombe Bridge Quarry near Newton Abbot, and Lummaton Quarry and Barton Quarry now within Torquay. In each town con- BRITISH DEVONIAN THYSANOPELTID AE (TRILOBITA) 195 fusion arises because of the close proximity of the quarries and by the application of a variety of names, some of which have fallen into disuse, and others which have no precise geographical location. The principal alternatives and grid references are noted below. 1. Wolborough Quarry (8X852705) : Woolborough, Wolvery, Newton Abbot, Newton, Newton Bushell, Woolbury. 2. Chercombe Bridge Quarry (8X832711) : Chircombe, Cherecombe, Bradley Woods, Newton Bushell. 3. Lummaton Quarry (8X914665) : Happaway, St. Mary Church, Babbacombe, Lummaton-Barton. 4. Barton Quarry (8X913671) : No alternatives appear to be used. Only two species of thysanopeltid trilobite recorded from South West England, Scutellum flabelliferum (Goldfuss) and Scutellum costatum Pusch, occur outside Britain. The former, which is characteristic of the Couvinian on the Continent, has only been identified positively by rare occurrences at Chercombe Bridge, probably from an horizon near to the Givetian /Couvinian boundary (Selwood 1965). This species represents the earliest thysanopeltid trilobite so far recognized from the British Devonian. Scutellum costatum, which has been recovered in considerable --nttrnbers from Wolborough, Lummaton and Barton, is relatively long ranging and continues both in Continental and British successions from the Givetian to the lower part of the Upper Devonian (Manticoceras Stufe). This particular species shows considerable variability and it has been found that forms closest to the original figures and descriptions are rare and recorded only from localities of Upper Devonian age (notably Ransleigh Quarry, near East Ogwell), whereas variants of the species from the Givetian at Lummaton and Wolborough are distinctive and referred to separate subspecies. The British species of Scutellum are all of Givetian age, and occur within the Shell- Bed at Lummaton. The age of the Shell-Bed has been discussed by House (1963), who refers the horizon to the zone of Maenioceras terebratum. However, it seems that goniatites have not been identified positively from the Shell-Bed : recent collecting indicates that goniatites characteristically occur in a black bituminous bedded limestone, which appears as a faulted wedge in the quarry. Since these limestones, as well as the Shell-Bed, have yielded Scutellum costatum lummatonensis and Scutellum costatum whidbornei, there is some presumptive evidence for the contemporaneity of the two lithologies. With the exception of Scutellum delicatum tigrinum, which has not been recognized, museum collections from Wolborough contain an identical thysanopeltid fauna to that from Lummaton. This particular locality has yielded goniatites (House 1963) indicating a lower horizon in the Givetian and referred to the zone of Maenioceras molar ium. IV. VARIABILITY Scutellum occurs more commonly than any other trilobite in the British Middle Devonian, and museum collections contain large numbers of specimens. Within BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) s '1 1 •I c ir a < J in " 0 < OS w 1 ^H 0 1 •S • t-< O rH s x ; ^ 2 r* fl B 0 ^ SB 0 4) -^ .&" £::;;:: x o bo ^ P a ^ 'C -S 1 1 1 S g ^ S en 1 1 1 J 1 .1 1 _g g ^3 ;g g C/D 5^ _S Qi ^0 2? as § ^S g ^ '-S .§ 1 i i •§ •§ •§ j§* Q 5S Q Vi vi *^3 ^i 3 1 1 1 4 •! -§. -PS I—, •- ^ "^ ^- *~ ~- if 9 f IP if f f S G S | \T) CO \O ^2 *^» ,H u_i p c^ 10 O ON w 10 H rsi ^^ *^ W O <3 ONOO iOrOONl>.N (N-S-S •4-> "SobMHOMOoS^> ° s s JH ^ tx, \O 00 t*>« VO vO CO C4 "^^ **** t^~i M-( " tx M O* tx HH tx M M ""^ ^i Wor2HOpOHwOO'§'§ -4.; "oooooooo^^ 03 r^ ^ I + + + + + + + ^j rO C^l vO iS1* vO O w *sO • cO O^ ^" ^d" *O cO *O ^" pd O O rocoioioo O •^ -^ u to BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) Number lation of icient specimens 94 59 5. costatum whidborne 89 65 and 84 64 S. costatum lummatonem N 0 oo to 0 1 H 00 ON CO ON O M ON ON j CO IO IO >O VO ON Tt~ vO s 1 to ON 10 »o tr^ M VO 00 ON ON ON H bB ON ON ON O\ ON ON ON ON ON ON ON ON ON ON ON ON B cS 8 ° O O O 0 0 O O O O 0 O O O O O a .2 ^ •ef* M CO H H CO M O >o oo r^ H Tj- ON O O 5° (3 *i CO IO M O >sO ^O t^ CO M M O M vO O t^» & 2 § *° M CO CO CO CO Tf IO M O O M CN vO ft O en H -J ON O H T}- ON 00 CO M 0 O H O M N M O HH •S Ctf M M 1— 1 HH ft W |H ^j o ft « H W*3 <0 0 X ^ S o CO >O CO N r^ O 00 CO CO !"">• ^* ON *O H HI N M IO O -f 00 H N CO C^ CO lO O 00 ON £**• *O t^» "fi O M HI O CO CO M O O O •<*• o M HMO C/) M o E H M O ON CO CO M Tt- O- M M o CO ON O O CO O CO O VO O" HI IO HI O 4J M o O 0 O 0 0 O 0 o 0 O O O HI 0 0 in CO CO Tf- 10 u-> 10 ON 10 r^ tx, vO ^^ CO M £*>, £-%, CO t^- M H IO M • t^- O • r^ o >o ^ M CO • M H • N ^> H H H ^5 ON CO M H {"•<••. *O ^^ O"1 0 M O ^o i—i 0 N r-- • oo CO O •o 00 t^ 0 Tf ON C-> •* co t^ VO IO O CO iO 0 O N 10 o O M Tf o O N -o-Q5 0-032 >o-o5 0-599 >o-05 Ba/Bp 3'86i o-o5 La/Lp 3-117 o-Q5 5-070 o-O5 1-851 >o-O5 Scutellum flabelliferum is stratigraphically older than the other species but three of the four growth indices are not significantly different from those of Scutellum costatum. In some specimens the median rib is reduced in size and this suggests that the species could provide a range of variation from which Scutellum costatum could have been derived. These comparisons thus broadly confirm the close relationship which was suggested to exist between Scutellum costatum and Scutellum delicatum and Scutellum flabelli- ferum but they do suggest that Scutellum pardalios is distinct. It should be remem- bered however that that the reduced major axes of species other than Scutellum costatum are based on relatively few specimens so that the calculated probabilities could well be underestimated. V. ECOLOGY The significance of the paliferum and companiferum types of body plan has been discussed at some length by R. & E. Richter (Richter, R. 1926 ; Richter, R. & E. 1934), who interpret certain structural elements as strengthening mechanisms for BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) 203 the exoskeleton. Whilst these two types are not recognizable precisely in the terms denned by Richter within the British material, both highly vaulted and flattened forms do occur (p. 197) and these appear to have ecological significance. In all thysanopletid trilobites the pygidium is the most prominent part of the exoskeleton, commonly equalling half the length of the body, and undoubtedly the distinctive shield-like form had considerable adaptive significance. This is con- firmed by the exceptional amount of diversification within the group. The thin delicate dorsal exoskeleton, the high proportion of surface area to body volume, and the distribution of individuals through a wide range of lithologies, suggest that the flattened forms were active and free-swimming. In complete specimens it is not uncommon to find the thorax flexed so that the pygidium and cephalon are inclined in opposite directions, either upwards or downwards. Whilst this could be a post mortem phenomenon it is possible that swimming was aided by paddle-like movements of the pygidium. The effect of such movements would be enhanced by the considerable expansion of the pleural region. The small size of the axis and the considerable breadth of the doublure indicate that the pygidial appendages were most probably reduced ; this might be expected if the swimming function had been, in part, taken over by the action of the whole pygidium. Such a reduction of the appendages would seemingly affect the efficiency of the respiratory system, but this could well be offset by direct diffusion of gases in solution through the thin integument. The strongly vaulted forms are more robust and not so clearly adapted for active swimming. Without exception, they are all restricted to the Shell-Bed lithology at Lummaton and Wolborough, which has been interpreted by Elliot (1961) and Dineley (1961) as shell debris infillings of original depressions on a stromatoporoid reef. They may well have lived actually browsing on the reef. VI. DETERIORATION OF THE BASAL PART OF THE GLABELLA Within the Thysanopeltidae the " occipital furrow " is normally broad (sag.}, deeply convex and frequently marked at the lateral extremities by tumour-like swellings. In the native British species the furrow is more deeply excavated at its anterior and posterior limits, and in Scutellum pardalios it shows a moderately elevated band within the furrow. This band together with the tumour-like swellings are coarsely ornamented. The presence of ornamentation within the cranidial furrows is unknown and suggests that a preoccipital glabellar lobe is present, which is divided into median and lateral lobes (the latter being represented by the tumour- like swellings). Thus the anterior excavation of the "occipital furrow" would appear to represent a transverse preoccipital glabellar furrow, the ornamented section the preoccipital glabellar lobe, and the posterior groove the true occipital furrow. Although Whidborne fully described the character of the " occipital furrow " in his original description of Scutellum pardalios, no significance was attached to it by him or by later authors. A somewhat similar arrangement is also indicated in Scutellum alutaceum geesensis (see Richter, R. & E. 1956, pi. 7, fig. 44A) but no comment 204 BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) has been offered. The present interpretation, affecting as it does the nature of the segmentation of the cephalon must be of significance and be represented in all members of the family. It is interesting to note that Eobronteus (Text-fig. 3A), which is usually considered to be close to the origin of the family, shows a narrow occipital furrow and, in addition, the base of the glabella is weakly constricted at its lateral margin. The latter could well indicate the presence of the basal glabellar lobe which was reduced in the course of evolution. It is suggested therefore that the broad " occipital furrow " which is usual in the family, has a composite structure formed through the disappearance of the pre- occipital glabellar lobes and the coalescence of the preoccipital glabellar furrow and occipital furrow. Where the tumour-like swellings occur, it may be interpreted that the median glabellar lobe only has been lost. FIG. 3. Deterioration in the basal part of the glabella in the Thysanopeltidae. A, Eobronteus laticauda (Wahlenberg) . Upper Ordovician. Occipital furrow narrow, ip indicated at the axial furrow ; B, Scutellum (Scutellum) pardalios (Whidborne). Middle Devonian, ip transglabellar, preoccipital ring divided into three small lobes ; C, Scutellum (Scutellum) delicatum (Whidborne). Middle Devonian. ip transglabellar, median lobe of preoccipital ring absent ; D, Scutellum (Thysanopeltis] speciosum (Hawle & Corda). Middle Devonian. Preoccipital lobes absent, a coalesced groove area only is left. (Figs. A and D after R. & E. Richter 1956). Thus it appears that there is, within the Thysanopeltidae a deterioration of the basal part of the glabella which is quite comparable to that seen in the Phacopidae and the Proetidae (Richter, R. & E. 1939). Text-fig. 3 illustrates the principal modifications of the basal part of the glabella seen in the family ; the specimens are arranged in a morphological series but this is not intended to imply stratigraphical or evolutionary significance. VII. SYSTEMATIC DESCRIPTIONS Family THYSANOPELTIDAE Hawle & Corda 1847 Genus SCUTELLUM Pusch 1833 TYPE SPECIES. Scutellum costatum Pusch 1833. Scutellum (Scutellum) pardalios (Whidborne) (PI. i, figs, i-n) 1889 Bronteus pardalios Whidborne : 29. 18890 Bronteus pardalios Whidborne ; Whidborne : 35, pi. 3, figs. 1-7. TYPE LOCALITY. Wolborough Quarry (8X852705) near Newton Abbot, Devon. 2O5 HORIZON. Middle Devonian. Givetian. Zone of Maenioceras molarium (Wol- borough Quarry) and of M. terebratum (Lummaton Quarry). LECTOTYPE. Here selected BM I.ioo. A pygidium figured by Whidborne (18890, pi. 3, fig. 5). This figure represents a restored specimen; the original is refigured (pi. I, figs. 1-3). SYNTYPES. SMC H. 4,108-113. BM 111.34924. FIGURED SPECIMENS. BM IT. 495 (PI. i, figs. 4, 5) ; BM IT. 496 (PL i, fig. 6) ; SMC H. 4,112 (PI. i, fig. 7) ; RAMME 7149 (PL i, figs. 8-n). MEASUREMENTS OF LECTOTYPE (in mm.) Length of pygidium 73-0 Length of axis 15-0 Breadth of pygidium 84-0 Breadth of axis 29-5 DIAGNOSIS. A species of Scutellum (Scutellum) showing a highly vaulted, but thin, dorsal exoskeleton. Furrows, of both cranidium and pygidium less prominent and less deeply impressed than usual for subgenus. Ornament consisting of coarse, closely spaced tuberculation. On pygidium and more sporadically on cephalon, ornament developed as prominent hollow truncated cones. Median and lateral preoccipital glabellar lobes raised and ornamented. Occipital ring ornamented by a massive mesial spine. DESCRIPTION In plan the glabella is broader than it is long, and expands forward rapidly to reach a maximum breadth (2 '5 times greater than at the posterior) just short of the anterior margin of the cranidium. The anterior margin of the glabella forms a broad arc which at the sagittal line extends to the anterior border. The glabellar furrows are not prominent. 4p is short, transverse and only recognizable at the lateral margin of the glabella. It deepens towards the axial furrow and is placed one-third of the distance from the anterior margin of the glabella to the preoccipital glabellar lobe. 3p is frequently scarcely recognizable, but is sometimes seen as a small transversely elliptical depression lying midway between the axial furrow and the sagittal line at a position half way between 4p and 2p. 2p is broader and more prominent than the glabellar furrows to the anterior. It is horseshoe-shaped and is marginally placed on the glabella so that the anterior and posterior branches run into the axial furrow. That part of the glabella so isolated forms a prominent node. The posterior branch is more deeply impressed than the anterior. The preoccipital furrow (ip) is transglabellar, broad (sag.) and deeply impressed and curves gently towards the anterior distally. It shows a symmetrical cross-section. The occipital furrow is less deeply impressed and shows an asymmetrical cross-section, rising more steeply to the anterior than to the posterior. The preoccipital glabellar lobe is low ; the median lobe is isolated and forms a narrow flat band, but the lateral lobes are more strongly elevated and recognizable as swellings arranged oblique to the preoccipital furrow. The glabella is defined by deeply impressed axial furrows ; the section that lies posterior to 2p is short and almost parallel to the sagittal line ; 206 BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA it then curves sharply outwards at the occipital furrow, thus considerably increasing the breadth (trans.) of the occipital ring. The section of the axial furrow anterior to 2p diverges from the sagittal line at approximately 45° ; it is less deeply impressed than the posterior section and deepens towards a pit, obliquely placed across the furrow, which is positioned adjacent to glabellar furrow 4p. The occipital ring is prominent and broad, reaching half the maximum breadth of the glabella ; it is characterized by a prominent backwardly-directed mesial tubercle. The frontal area of the cranidium is small and only recognizable as slightly flattened areas situated at the anterior lateral extremities of the glabella. The anterior border furrow is weak and clearly defined only away from the sagittal line. The anterior border is narrow and shows a more or less level surface ornamented by 2-3 continuous terrace lines. From the anterior margin it is recurved sharply to the posterior ; this section is ornamented by 3-4 continuous terrace lines. At the antero-lateral margin of the cranidium this border is strongly upstanding but it is scarcely raised at the sagittal line. The fixed cheeks are large and inflated, and attain their maxi- mum breadth posteriorly, but narrow rapidly anteriorly between the facial suture and the axial furrows. From the highest part of the fixed cheek, the cheek area falls away steeply to the posterior but more gently to the anterior. The posterior border furrow of the cranidium is positioned close to the margin. From its origin which is just removed from the axial furrow, it runs obliquely towards the anterior, making an angle of approximately 20° with the posterior border. In this section the furrow has a sharp V-shaped cross-section. At a point below the palpebral lobe the furrow turns posteriorly to run more or less parallel to the facial suture ; in this section the furrow has a broad evenly rounded cross-section. Posteriorly the furrow narrows and fails to reach the posterior margin. This furrow is not observed in the librigenae. The posterior border forms a flattened plane sloping towards the posterior border furrow ; from its origin near the glabella it widens rapidly distally. The palpebral lobe is small and semicircular in outline ; its anterior lies opposite 2p and its posterior is positioned opposite the posterior end of the glabella. The palpebral lobe is placed farther from the sagittal line than the anterior lateral border of the cranidium. The palpebral furrow forms a shallow open groove ; it has not been traced beyond the anterior limit of the lobe but continues posteriorly, pro- gressively shallowing, following a course broadly parallel to the posterior branch of the facial suture for about one-third of the length of the latter. The anterior branch of the facial suture is long and gradually approaches the axial furrow towards the anterior. From the palpebral lobe this branch diverges outwards at approximately 10° from the exsagittal line ; opposite 3p it swings inwards but then almost immedi- ately resumes its course to the anterior and runs more or less parallel to the sagittal line, finally cutting the anterior margin of the cephalon almost at right angles. The posterior branch of the facial suture is short, from the posterior end of the palpebral lobe it swings outwards through a semicircle to run almost parallel to the posterior border. It then curves sharply to the posterior to cut the posterior border obliquely at a distance from the axial furrow equal to four-fifths of the breadth (trans.) of the occipital ring. BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) 207 The frontal profile is made up of 5 prominent and distinct arches comprising the central glabella, bounded by the fixed cheeks and the palpebral lobes. The glabella is strongly inflated and subtriangular in outline and its sides decline from the crest at approximately 60°. The crestal point, at the sagittal line, is rounded and lies well above the adjoining parts of the fixed cheeks. Deeply incised axial furrows separate the glabella from the fixed cheeks ; these have an open V-shaped cross- section. Each fixed cheek is bounded adaxially by the axial furrow and abaxially by the palpebral furrow. The highest part of the cheek area is rounded and lies three-quarters of the distance from the axial furrow to the palpebral furrow. From this point the cheeks slope down to the axial furrow at approximately 60° and rather more gently to the palpebral furrow. The palpebral lobe is flat or gently inclined towards the weakly impressed palpebral furrow. In side view the glabella is strongly inflated and stands well above the adjoining parts of the cranidium. It rises steeply from the preoccipital furrow and then slopes more gently to a crestal point one-third the distance from the preoccipital furrow to the anterior border. The anterior section of the curve slopes gently to the anterior but steepens up at the anterior border furrow. No preglabellar field is seen. The border furrow is evenly concave and rises anteriorly into a low insigni- ficant border. The preoccipital furrow is relatively long (sag.) and deep and passes into a low evenly inflated median preoccipital lobe. The occipital furrow is narrower and less deeply impressed than the preoccipital furrow. The occipital ring is asymmetrical and rises almost vertically from the posterior to the crestal point, which approximates to the level of the base of the glabella ; it then slopes more gently to the occipital furrow. The preoccipital ring is about one-third of the length (sag.) of the occipital ring. The free cheek is of moderately large size and is subtriangular in outline. The eye is prominent and strongly curved longitudinally and is seen to occupy the highest part of the free cheek but lies below the level of the palpebral lobe. The eye surface is convex, directed upwards and outwards, and marked by numerous holochroal facets which give a machine turned appearance to the surface. A prominent evenly rounded eye socle separates the eye from the rest of the free cheek. The cheek area is longer than broad and slopes steeply away from the eye to a broad, concave furrow which bounds the lateral part of the cheek. Towards the anterior, the cheek area slopes more gently than laterally. The border is small and appears as a narrow rim to the cheek ornamented by discontinuous raised lines directed obliquely backwards. In plan the pygidium is semielliptical in outline and broader than long, with a length : breadth ratio of 5 : 7 ; the maximum breadth is attained just posterior to the axis. The articulating half-ring is small, delicate and broadly curved. The articulating ledge which forms the inner part of the anterior margin is straight and measures three-fifths of the maximum pygidial breadth. This ledge terminates in a small articulating wing ; from this point the outer part of the anterior margin curves in an even arc to the lateral margin of the pygidium. The axis is small and subtriangular. At the anterior margin its measurements exceed one-third of the GEOL. 13, 3. I4 208 BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) maximum breadth recorded for the pygidium and at the sagittal line it approaches one-sixth of the total pygidial length. A broad, deep and evenly rounded furrow clearly separates the axis from the articulating half-ring. This furrow is divided into three equal sections by two weak constrictions marking longitudinal furrows which are continued into the axis and effect a weak trilobation. From the anterior margin of the axis these furrows are directed posteriorly parallel to the sagittal line. At a point half the distance (exsag.) from the anterior margin to the axial furrow, they curve evenly inwards so that the straight sections are united by an almost semicircular furrow which, at the median line, lies only one-sixth of the total distance from the posterior margin of the axis to the anterior. The median lobe of the axis is evenly inflated and the lateral lobes slope steeply to the pleural furrows. The axial furrows are broad, smooth, and particularly deeply set at the anterior margin. At the midpoint the axial furrows are constricted ; the posterior section of each furrow makes an angle of 40° with the sagittal line whilst the anterior section, if projected, would make an angle of 50°. No segmentation of the axis has been observed. The flanks of the pygidium are strongly vaulted. The inner part of the pleural region forms a pronounced elevated platform which, at the sagittal line, extends two- thirds of the distance from the axial furrow to the posterior margin. The outer boundary of the platform crosses the ribs at their midpoints to reach the anterior margin at the articulating wing. Broadly, the platform is at two levels for ribs 1-2 lie at a distinctly lower level than ribs 4-7 ; rib 3 occupies an intermediate position. The median rib lies in a weakly depressed position on the higher platform. From the margin of the platform, the pleurae slope steeply to the periphery of the pygidium. The pleural field is divided by narrow moderately impressed intercostal furrows into 7 paired ribs and a single medianly placed rib, which radiate out from the axis. The intercostal furrows weaken towards and fail to reach the axis and also stop short of the periphery of the pygidium. The first pair of intercostal furrows is more deeply impressed than the following, so that ribs I and 2, and parti- cularly rib i, are more convex than the succeeding ribs. The two anterior pairs of intercostal furrows are almost straight but towards the posterior the furrows become weakly S-shaped, each furrow curving weakly to the anterior at the margin of the pygidium. Each rib attains its maximum breadth at the periphery and decreases in breadth progressively towards the axis. The second pair of ribs tapers very strongly and barely reaches the axis. Ribs 3-6 are moderately broad at the axis (about one- third of the maximum peripheral breadth), but rib 7 and the median rib are narrower (about one-fifth of the maximum peripheral breadth). All of the ribs are broad and flattened at the periphery but increase in convexity progressively towards the platform. On the peripheral parts of the platform the ribs are moder- ately inflated but, with the exception of rib i, all flatten somewhat as the axis is approached. In side view the axis extends posteriorly for a quarter of the total sagittal length of the pygidium ; it is strongly delimited anteriorly from the articulating half-ring by a deep evenly curved furrow and posteriorly by a less prominent groove. The axis is strongly inflated, rising steeply from the articulating furrow but descending BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) 209 rather more gently to the posterior. The platform is very lightly and evenly bowed and extends posteriorly from the axis for a distance approaching two-thirds of the total post axial length (sag.). Posteriorly the platform passes gradually into a broadly concave slope which extends to the posterior margin of the pygidium. Viewed from the posterior the axis occupies a prominent elevated position and appears weakly trilobed in outline. The evenly rounded median lobe of the axis passes gently into the more flattened lateral lobes ; the latter curve sharply down to deeply impressed pleural furrows. The outer section of the lateral lobe, the axial furrow, and the inner section of the anterior margin are obscured by the elevated posterior section of the pygidium, which rises in this view to a height approaching half that of the axis. The anterior pleural region is relatively low and weakly S-shaped and slopes gently to the outer margin. The doublure is thin and is closely applied to, and follows the vaulting of, the dorsal exoskeleton. In the pygidium its inner margin reaches the platform. In the ventral view, the position of the intercostal furrows is marked by moderately convex elevations. The anterior border of the cranidium is ornamented by 3-4 unbranched hair-like lines which run parallel to the margin of the cranidium. The posterior border of the fixed cheeks is smooth. The glabella is characterized by massive, pointed tubercles directed obliquely backwards. These tubercles, which are frequently broken, are smaller and more closely spaced towards the anterior, particularly on the median part of the glabella. Towards the posterior of the glabella the tubercles are prominent and frequently elongated transversely. In the posterior glabellar region a tubercle, larger than any of the others, lies on the sagittal line just anterior to the preoccipital furrow. Usually the occipital ring is broken medianly, but in perfect specimens it is characterized by a massive mesial tubercle which swells to occupy the whole length (sag.) of the ring. This tubercle is drawn out into a sturdy spine, directed obliquely backwards. The lateral parts of the ring are ornamented by small, irregular tubercles of which 2-3 would be encountered on an exsagittal traverse ; these tubercles are comparable in size to those at the anterior end of the glabella. The median lobe of the preoccipital ring is marked by two rows of small tubercles, and the lateral lobes show 2-3 tubercles of varying size. The fixed cheeks are tuberculate but with smaller tubercles than on the glabella ; these are of fairly uniform size but do increase in size slightly towards the palpebral lobe. The latter is also tuberculate and shows one particularly large tubercle on its posterior border. All of the furrows of the cranidium are smooth. The free cheek is ornamented by closely spaced tubercles comparable in size to those occurring on the outer parts of the fixed cheeks. These tubercles, which are almost invariably broken at their summits are directed obliquely backwards and become more closely spaced and regular in size towards the genal angle. Finer tubercles appear towards the borders. The pleural region of the pygidium is bounded by a narrow margin which is ornamented by delicate hair-like lines ; at the anterior these lines run broadly parallel to the border but posteriorly lie oblique to the margin. The ribs and axis 210 BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) are covered with coarse, closely spaced tubercles which appear as truncated cones directed upwards and backwards. The larger tubercles, which approach I mm. in diameter at their summits in large specimens, tend to be more abundant towards the periphery of the pygidium and are irregularly associated with smaller tubercles. Across the ribs (trans.) some 6 tubercles are recorded at the periphery and two near the axis. Inevitably many of the tubercles are damaged in the course of collecting the specimens and this gives the impression that they represent the broken bases of delicate hollow spines. However, there is a remarkable uniformity in the heights of the tubercles, which would not necessarily be expected if spines were represented as a continuation of the dorsal exoskeleton. Further, it is evident from perfectly preserved tubercles that the exoskeleton is not broken at the summit of the cone but is recurved slightly to form a narrow flat ring surrounding a depressed central area which is marked by broken secondary calcite. Thus it is unlikely the dorsal exoskeleton was ever continued into prominent spines. All of the furrows of the pygidium are smooth. Sectioned material containing pygidia embedded in matrix has revealed the presence of exceedingly delicate setae, one of which fits into the central pit of each tubercle. These setae are hollow and now filled with fibrous calcite which is apparently continuous with the secondary calcite filling the body space between the dorsal exoskeleton and the doublure. It may be presumed that the setae were movable and responsive to tactile stimuli ; the extreme delicacy of the walls suggests that the setae may also have been the receptors of chemical stimuli. The ventral surface of the doublure is ornamented by clear asymmetrical furrows which are continuous or occasionally branched and run more or less parallel to the border of the cephalon and pygidium. The furrows are crowded at the periphery but become gradually more widely spaced until at the inner margin the interval between them is six or seven times as great as at the periphery. On the dorsal surface of the doublure these furrows are represented as weakly raised areas. STRUCTURE OF THE EXOSKELETON. All specimens possess, for their size, an exceedingly thin and delicate exoskeleton ; in a pygidium 40 mm. long the total thickness is only 0-09 mm. Sectioned material reveals that the exoskeleton is made up of three layers ; a thin outer (dorsal) layer, a middle layer 4 times as thick as the outer layer, and a thin inner layer twice as thick as the outer layer. The inner and outer layers are distinguished by being rather more heavily pigmented than the middle layer. This threefold division of the dorsal exoskeleton has been recorded in a number of other trilobites (Harrington 1959 : 85) ; in all cases the middle layer is the thickest but the relative thickness of the other layers varies. Since Kielan (1954) records that the outer layer is thicker than the inner layer in Scutellum flabelliferum it is evident that these relative thicknesses are not uniform for the genus. On the cephalon the exoskeleton thins over the tubercles ; it has not been possible to determine whether all of the layers are represented here, but the thickness observed suggests that more than the outer layer is represented. It is unlikely that the mineralized exoskeleton was ever continuous over the tubercles on the pygidium (p. 210). BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) 211 Considering the large size of this species and the thickness of the dorsal exoskeleton it is perhaps surprising that the specimens, particularly the pygidia, have survived intact. However, R. & E. Richter (1934) have pointed out the very considerable reinforcement given by the broad doublure closely applied to the dorsal exoskeleton ; this is supplemented in Scutellum pardalios by thickening of the exoskeleton in the narrow furrows. Thus, the highly vaulted pygidia are supported by thickened arches (intercostal furrows) which radiate out from the axis to the periphery. The size and the general absence of malformed specimens suggest that the exoskeleton was quite adequate for its functions. Where break up of the specimens has been observed it is quite clearly posthumous and results from the effects of either compac- tion or tectonic activity, and reflects not the fragile exoskeleton but the large area covered by each specimen. The minor displacements associated with the fracturing of the exoskeleton frequently give the impression that the exoskeleton is thick, for it reveals beneath the true exoskeleton layers of fibrous calcite which grew around the specimen during diagenesis and appear superficially as a continuation of the dorsal exoskeleton. COMPARISONS. The character of the ornamentation serves to distinguish the species from all others in the subgenus Scutellum (Scutellum) . Within this subgenus the pygidium is most closely related to the highly vaulted forms of Scutellum (Scutellum) costatum whidbornei, but is distinguished from it by its ornament, size, and by the narrower and less clearly impressed interpleural furrows. The glabella is broader and much more strongly vaulted than in 5. (Scutellum) costatum. Many of the features of this species are characteristic of the subgenus S. (Parale- jurus) but it is distinguished from all members of this subgenus by the thin exo- skeleton, the prominent ornamentation and by the more deeply impressed intercostal furrows of the pygidium. Scutellum (Scutellum) costatum costatum Pusch (PI. 2, fig. 10) 1833 Scutellum costatum Pusch : 119. 1926 Scutellum costatum Pusch; Richter, R. & E. : 117, pi. 7, figs. 18-23 (includes earlier synonymy of the species). 1954 Scutellum (Scutellum) costatum Pusch ; Kielan : 31. 1955 Bronteus costatus (Pusch) Maksimova : 32, pi. i, figs. 5-9. 1956 Scutellum (Scutellum) costatum Pusch ; Richter, R. & E. : 96, 101, pi. 2, figs. 7, 8. REMARKS. Scutellum (Scutellum) costatum s. str. is rare and has only been posi- tively identified from horizons of Upper Devonian age in Devon (p. 195), and it appears to be the end member of a continuously varying series extending from the Lower Middle Devonian Scutellum flabelliferum through Scutellum costatum whid- bornei subsp. n. in the Upper Middle Devonian. The possibility of such a relationship existing between Scutellum flabelliferum and Scutellum costatum was indicated by R. & E. Richter (1926 : 121). FIGURED SPECIMEN. TM 6.490 (PI. 2, fig. 10). 212 BRITISH DEVONIAN THYS ANOPELTID AE (TRILOBITA) Scutellum (Scutellum) costatum whidbornei subsp. n. (PI. 2, figs. 9, 11-18) 1861 Bronteus fldbellifer Goldfuss ; Pengelly : 343, pi. 7, fig. 2. i88ga Bronteus granulatus Goldfuss ; Whidborne : 40, pi. 3, fig. 10. DERIVATION OF NAME. After the Rev. G. F. Whidborne. TYPE LOCALITY. Wolborough Quarry (8X852705) near Newton Abbot, Devon. HORIZON. Middle Devonian. Givetian. Zone of Maenioceras molarium (Wol- borough Quarry) and of M. terebratum (Lummaton Quarry). HOLOTYPE. BM 1.5,078 (PI. 2, figs. 17, 13). PARATYPES. BM IT. 497 (PI. 2, fig. 9) ; BM IT. 498 (PI. 2, figs, n, 12) ; SMC H. 3,778 (PI. 2, figs. 18, 14) ; SMC H. 3,782 (PI. 2, figs. 15, 16). MEASUREMENTS OF HOLOTYPE (in mm.) Length of pygidium 27-0 Length of axis 5-0 Breadth of pygidium 31-0 Breadth of axis 9-5 DIAGNOSIS. A subspecies distinguished from 5. (Scutellum) costatum costatum Pusch by the prominent median rib of pygidium clearly united with axis. Platform less clearly defined from rest of pleural region. REMARKS. The variability of this subspecies renders the selection of a holotype difficult. A specimen near to the norm has been chosen ; this specimen may be distinguished from Scutellum (Scutellum) costatum costatum by the character of the median rib as indicated in the diagnosis. Specimens which lie away from the norm inevitably show additional distinguishing features which affect the vaulting of the dorsal exoskeleton, the character of the median rib, and intercostal furrows. The changes affecting these characters are indicated on page 199. Scutellum (Scutellum) costatum lummatonensis subsp. n. (PI. 2, figS. 1-8) DERIVATION OF NAME. From Lummaton, Torquay, Devon. TYPE LOCALITY. Lummaton Quarry (8X914665) Torquay, Devon. HORIZON. Middle Devonian. Givetian. Zone of Maenioceras molarium (Wol- borough Quarry) and of M. terebratum (Lummaton Quarry). HOLOTYPE. Pygidium SMC H. 2,254 (PI- 2> ng- 7)- PARATYPES. RAMME 66 /oi (PI. 2, figs, i, 2) ; SMC H. 3,801 (PI. 2, fig. 3) ; RAMME 66/02 (PL 2, figs. 4-6) ; SMC H. 2,253 (PI. 2, fig. 8). MEASUREMENTS OF HOLOTYPE (in mm.) Length of pygidium 16-0 Length of axis 3-0 Breadth of pygidium 20-0 Breadth of axis 5-0 BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) 213 DIAGNOSIS. A subspecies of S. (Scutellum) costatum Pusch showing similar range of variation to S. (Scutellum) costatum whidbornei but distinguished by absence of all strong ornamentation. Scutellum (Scutellum) delicatum delicatum (Whidborne 1889) (PL 3, figs. 4, 7-9) 1889 Bronteus delicatus Whidborne : 29. i88ga Bronteus delicatus Whidborne ; Whidborne : 33, pi. 3, figs. 14, 140!, 15. TYPE LOCALITY. Wolborough Quarry (8X852705) near Newton Abbot, Devon. HORIZON. Middle Devonian. Givetian. Zone of Maenioceras molarium (Wol- borough Quarry) and of M. terebratum (Lummaton Quarry). LECTOTYPE. Cranidium SMC H. 4,107 (PL 3, figs. 8, 9). Here designated. SYNTYPES. Pygidium BM 1.1,094; SMC H. 4,106. FIGURED SPECIMENS. BM 1.1,084 (PI- 3> fig- 4) » SMC H. 3,804 (PL 3, fig. 7). MEASUREMENTS OF LECTOTYPE (in mm.) Length (sag.) of cranidium 28-25 Length (sag.) of occipital ring 3-0 Max. breadth of glabella 26-25 Length (sag.) of glabella 21-5 (excluding preoccipital ring) DIAGNOSIS. A species of Scutellum (Scutellum) typically lacking coarse ornamenta- tion. Anterior section of glabella shows pronounced asymmetrical wrinkling of dorsal exoskeleton, weakening and becoming discontinuous posteriorly. Weak elongated nodes appear in same direction along crests of wrinkles ; at first these nodes only represented along flanks of glabella but later also present medianly. Pygidium normally smooth but showing fine, widely spaced pustulation in oblique light. Pygidium characterized by prominent tumid axis weakly divided by longi- tudinal furrows into three lobes ; median lobe continued posteriorly into prominent undivided median rib larger than all other ribs. Intercostal furrows defining median rib frequently appearing continuous with axial furrows. DESCRIPTION In plan the glabella is broader than it is long, and expands rapidly forward to reach a maximum breadth (3 times that at the posterior) near to the anterior margin of the cranidium. The anterior margin of the glabella forms a broad arc reaching to the anterior border at the sagittal line. The glabellar furrows are weakly impressed, the anterior furrows only being recognizable in low, oblique light. 4p is short, transverse and forms a broad but shallow furrow extending from the axial furrow rather more than half the distance towards the sagittal line. It is most deeply impressed at its midpoint, but shallows more rapidly towards the axial furrow than towards the sagittal line. At the axial furrow it lies half the distance (exsag.) between the anterior of the glabella and the preoccipital glabellar furrow. 3p forms a shallow traversely elliptical depression lying rather nearer to 2p than 4p 214 BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) and midway between the axial furrow and the sagittal line. 2p is broader and more prominent than the glabellar furrows to the anterior. It is horseshoe-shaped and marginally placed on the glabella so that the anterior and posterior branches run into the axial furrow ; that part of the glabella, so isolated, forms a low node. The anterior branch is broader and shallower than the posterior ; the latter is most deeply set at the axial furrow. The preoccipital furrow (ip) is transglabellar, deep and broad (sag.) and curves gently towards the anterior distally. The median preoccipital glabellar lobe is low and flattened and is inclined from anterior to posterior but scarcely rises above the preoccipital and occipital furrows. The lateral preoccipital glabellar lobes are weakly elevated and recognizable as small nodes lying oblique to the preoccipital furrow. The axial furrows are most deeply impressed posterior to 2p. Between 2p and the posterior of the glabella, the axial furrows are straight and deeply impressed and follow a course almost parallel to the sagittal line. Further to the posterior they shallow, broaden and are deflected weakly outwards at the preoccipital furrow and yet more strongly outwards at the occipital furrow, so that the occipital ring is about one and a half times the breadth (trans.) of the glabella posterior. Anterior to 2p the axial furrows shallow and broaden, and diverge from the exsagittal line at 45° and finally unite anteriorly with the anterior border furrow. The anterior border furrow is absent at the median line and only weakly developed at the anterior lateral termination of the glabella. The occipital ring is lens-shaped and prominent, being one and a quarter times longer (sag.) than the preoccipital ring. The fixed cheeks are gently inflated ; at the posterior they are broad but narrow appreciably towards the anterior. The posterior border furrow is shallow and placed close to the posterior margin of the fixed cheek. The palpebral lobe is small and semicircular in outline ; its anterior margin lies opposite the posterior end of the glabella and its posterior margin lies close to the posterior border furrow. At its extremity the palpebral lobe is farther from the sagittal line than any other part of the cranidium. The anterior branch of the facial suture is long and straight. From the anterior end of the palpebral lobe this branch diverges at approximately 10-15° from the exsagittal line and cuts the anterior border close to the axial furrow. The posterior branch of the facial suture is short and swings outwards through a semicircle from the posterior end of the palpebral lobe, to run almost at right angles to the sagittal line. On the line (exsag.) of the outer part of the palpebral lobe the suture curves sharply to the posterior to cut the posterior margin three-fifths of the distance from the axial furrow to the genal angle. Viewed from the front the whole cranidium forms a broad, moderately inflated curve which is broken into 3 sections by the axial furrows into a central glabellar region bounded by the fixed cheeks. The glabella forms a broad, low curve sloping gently away from the crestal point at the sagittal line towards the axial furrows which, in this view, do not appear to be strongly impressed. The palpebral lobe is small and is differentiated from the rest of the fixed cheek only by being slightly more elevated ; it rises distally to the height of the crest of the glabella The palpebral furrows are not recognizable. The palpebral lobe and fixed cheek slope BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) 215 gently towards the axial furrow, but dip down rather more steeply at the furrow itself. The combined breadth (trans.) of the palpebral lobe and fixed cheek is greater than that of the glabella. In side view the glabella is low and scarcely rises above the level of the highest parts of the fixed cheeks. The curve of the glabella forms a low flattened arc which rises very slightly from the shallow preoccipital furrow, then slopes gradually to the anterior. Two-thirds of the distance from the posterior of the glabella to the anterior the gradient increases so that the anterior of the glabella forms a more steeply inclined arc. No preglabellar field is apparent. The anterior border is low and insignificant. The preoccipital furrow is shallow and the median lobe of the preoccipital ring is scarcely recognizable ; in this view it is seen only as a gently inclined surface sloping from the occipital furrow to the preoccipital furrow. The occipital ring is prominent rising above the level of the posterior end of the glabella at its crestal point. The occipital furrow is only weakly impressed. The free cheek is large and subtriangular in shape and is most probably continued postero-laterally into a short, sharp genal spine. The eye is prominent and is strongly curved longitudinally and is elevated to form the highest part of the free cheek, rising to a level equal to half the height of the glabella. The eye surface is convex and marked by many holochroal facets ; it rises steeply from a narrow flattened platform but is separated from it by a weak eye socle. The cheek area is longer than broad and slopes steeply, in an even curve, away from the platform to a broad but shallow lateral border furrow. The lateral border is insignificant and appears as a thin raised area. The posterior border furrow is absent. In plan the pygidium is semielliptical in outline, being broader than it is long with a length : breadth ratio of 4 : 3. The maximum breadth is attained just anterior to the posterior termination of the axis. The articulating ledge which forms the inner part of the anterior margin is straight and measures approximately one-fifth of the maximum pygidial breadth. This ledge terminates in a small articulating wing, from this point the outer part of the anterior margin curves in an even arc to reach the lateral margin at the point of maximum breadth. The articulating half- ring is broad (trans.) and short. The axis is small and subtriangular in outline. At the anterior margin its measurements exceed one-third of the maximum breadth of the pygidium and at the sagittal line it approaches one-fifth of the total pygidial length. The axis is separated from the articulating half-ring by a deep, moderately broad, furrow which is arched forwards medianly and, to a lesser extent abaxially. The axial furrows are narrow and not deeply impressed. The axis is strongly inflated and faintly trilobed. At the anterior border the axis is divided into 3 approximately equal sections by 2 weakly impressed furrows ; these are continued posteriorly broadly parallel to the sagittal line. At the posterior extremity of the axis, the median lobe is of the same breadth as the median rib. The median lobe of the axis is uniformly rounded and the lateral lobes triangular in outline and they become somewhat flattened at the lateral border. The posterior termination of the axis is rounded. The axial furrows are broad and smooth but not deeply impressed so that they are 216 BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) scarcely marked at the posterior termination of the tail. The central section of the furrows is straight and would, if projected, make an angle of 45° with the sagittal line. Anterior from this central section, the furrows curve outwards rather more gently and posteriorly the axial furrows become continuous with the furrows denning the median rib. No segmentation of the axis has been observed. The pleural region forms a low flattened surface sloping gently away from the low platform bounding the axis. The junction between the platform and the rest of the pleural region is perfectly gradational. The pleural region is divided by narrow deeply impressed intercostal furrows into 7 paired ribs and a single median rib. Paired ribs i and 2 are straight and radiate out from the axis. The remaining ribs are gently curved ; from the axis they are directed to the posterior, broadly parallel to the sagittal line, and then sweep posteriorly outwards towards the periphery. All of the intercostal furrows fail to reach the periphery of the tail and, with the exception of those denning the median rib, all furrows weaken towards and fail to reach the axial furrows. Towards the axis the intercostal furrows are deep and have evenly rounded bottoms but, as the furrows broaden towards the periphery, they flatten to reach a maximum of one-quarter of the breadth (trans.) of the adjoining ribs. With the exception of the median rib all of the ribs decrease evenly in size from the periphery towards the axis and at the same time the cross-section of the ribs changes from broadly arcuate to convex. At the axis the ribs are a quarter to one-fifth of their width at the periphery. Rib 2 lies between strongly converging intercostal furrows and does not reach so far towards the axis as the adjoining ribs. The median rib is larger than any of the paired ribs ; it is bottle shaped and continuous with the axis. In side view the axis extends posteriorly for rather less than one-quarter of the total sagittal length of the pygidium and is strongly delimited anteriorly from the articulating half-ring by a deep, evenly rounded furrow. Posteriorly it passes, without a break, into the pleural area. The articulating half-ring is small, at the sagittal line extending anteriorly for a distance equal to the maximum breadth of the articulating furrow. The axis is inflated but somewhat flattened ; it rises steeply from the articulating furrow and then flattens ; the posterior two-fifths of the axis forms a broad arc descending from the flattened area. The platform is flattened to very gently convex, and extends posteriorly for a distance of seven- twelfths of the total post-axial length (sag.) and then passes through a gentle curve into a broadly concave slope extending to the posterior margin of the pygidium. Viewed from the posterior the axis occupies a prominent position and shows a weakly triangular cross-section but is weakly constricted at the midpoints of the slopes away from the sagittal line. The pleural region forms a broad low arc. The platform slopes gently from the axis to its periphery, where it curves in a broad concave slope to the margin of the pygidium. The proximal sections of ribs I and 2 are obscured by the slightly more elevated posterior section of the pygidium. The doublure is broad and very thin and extends more than two-thirds of the distance from the margin of the pygidium to the axis. It is closely applied to the dorsal section of the exoskeleton and ornamented by delicate finely branched terrace BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) 217 lines which run more or less parallel to the border of the pygidium. The spacing of the terrace lines increases towards the axis. The anterior section of the glabella is ornamented by prominent asymmetrical wrinkles, which represent infoldings of the whole exoskeleton. These are continuous and run more or less parallel to the anterior border of the glabella. Towards the posterior the wrinkling weakens and becomes discontinuous, and at the same time low elongated elevations, which first appear on the lateral flanks of the glabella, become evident along the wrinkle lines. This ornamentation is represented over the whole of the posterior half of the glabella. The posterior parts of the fixed cheeks show a low irregularly spaced granulation but the anterior parts are smooth. The ornamentation of the occipital ring is imperfectly known, but the lateral parts appear to be ornamented by low elongated elevations which are broadly convex to the anterior and lie oblique to the transverse line. The anterior border is orna- mented by up to 3 terrace lines. The pygidium is smooth but in oblique light small irregularly spaced hollow pustules are discernible. It appears that the exoskeleton represented over the pustules is much thinner than that over the rest of the pygidium, so that it becomes transparent and reveals the secondary calcite filling the pustule. All of the furrows on the cranidium and pygidium are smooth. Scutellum (Scutellum) delicatum tigrinum (Whidborne) (PL 3, figs. 5, 6, 10-12) 1889 Bronteus tigrinus Whidborne : 34, pi. 3, fig. 12. 18890. Bronteus delicatus Whidborne ; Whidborne, pi. 3, fig. 13. 18890 Bronteus granulatus Goldfuss ; Whidborne, pi. 3, fig. 8. TYPE LOCALITY. Probably Lummaton Quarry (8X914665) Torquay, Devon. HORIZON. Middle Devonian. Givetian. Zone of Maenioceras terebratum. LECTOTYPE. BM In. 34923 (PI. 3, fig. n). Here designated. SYNTYPE. SMC H. 2241. This cranidium is imperfectly preserved and cannot be referred to any species with certainty. FIGURED SPECIMENS. SMC H. 3,785 (PI. 3, fig. 5) ; BM ^.58,998 (PI. 3, fig. 6) ; BM IT. 499 (PI. 3, fig. 10) ; BM ^.34,923 (PI. 3, fig. n) ; SMC H. 4,106 (PL 3, fig. 12). MEASUREMENTS OF LECTOTYPE (in mm.) Length (sag.) of cranidium 18-0 Length (sag.) of occipital ring 2-0 Max. breadth of glabella 19-0 Length (sag.) of glabella 1275 (excluding preoccipital ring) DIAGNOSIS. A subspecies of Scutellum (Scutellum) delicatum (Whidborne) dis- tinguished from type species by coarser ornamentation of dorsal exoskeleton. Ornamentation consists of moderately fine granulation of ribs and axis of pygi- 218 BRITISH DEVONIAN THYS ANOPELTID AE (TRILOBITA) dium, and moderately pronounced granulation of posterior parts of glabella. Glabella broad, showing maximum length : maximum breadth ratio of 2:3. REMARKS. Whidborne (18890) established the species Bronteus tigrinus on the basis of two cranidia only and distinguished it from Bronteus delicatus by the broader glabella and by the presence of prominent pits within the axial furrows. Recent collecting has revealed additional material including an almost complete specimen from which it is apparent that the pygidium, like the cranidium, is closely comparable to that of 5. (Scutellum) delicatum. However, it is distinguished quite clearly by a moderate granulation of the dorsal exoskeleton. This granulation is more closely spaced and prominent on the axis than on the ribs. In addition, the axis is less tumid and its threefold longitudinal division is more marked than in S. (Scutellum) delicatum. The ornamentation of the cephalon is also distinctive ; the anterior part of the glabella is characterized by a weak wrinkling giving a broadly continuous series of lines at the anterior margin. These become less prominent and more discontinuous towards the posterior. Small transversely elongated tubercles arise from the lines at the antero-lateral parts of the glabella ; these increase in prominence towards the posterior and come to occupy the whole of the posterior section of the glabella. At the same time the tubercles become less clearly elongated and appear as rounded bosses at the glabella posterior. The occipital ring is more finely ornamented than the posterior parts of the glabella, but does show a large prominent mesial tubercle. With the exception of the anterior parts, the fixed cheeks are moderately strongly granulated. The free cheeks appear to be ornamented only by light irregular wrinkling of the dorsal exoskeleton. Whidborne's two species are thus distinguished primarily by differences in orna- mentation ; Bronteus delicatus being smooth and Bronteus tigrinus granulated. This situation is paralleled exactly by the British Middle Devonian forms of Scutellum costatum. It appears therefore unwise to attribute more than subspecific significance to the differences in ornamentation and it is proposed to distinguish the granulated form as 5. (Scutellum) delicatum tigrinum and the smooth form as 5. (Scutellum) delicatum delicatum. The possibility of the two forms being varieties of the same species was envisaged by Whidborne (18890:34). The separation of the two subspecies on the basis of the ornamentation necessitates the inclusion of one of the specimens (SMC H. 4,106) figured by Whidborne (pi. 3, fig. 13) as Bronteus delicatus within the subspecies 5. (Scutellum) delicatum tigrinum. The re-allocation of this specimen then allows a further distinction to be drawn ; namely that the glabella of the subspecies tigrinum shows a maximum breadth : maximum length ratio of 3 : 2, which is appreciably greater than in the subspecies delicatum. A fairly complete specimen of this species (BM IT. 499, PI. 3, fig. 10) shows that 10 uniform segments are present in the thorax. The axis of each segment is broad, but narrower than each pleura, and forms a broad subtriangular arch rising sharply from the flattened pleural areas. Each ring is convex in section and slopes more gently towards the anterior than to the posterior, and is continued anteriorly into a broad (trans.} but short (sag.) articulating half-ring. The axial furrows are not BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) 219 prominent. The proximal parts of the pleurae show evenly convex median sections which bear narrow flattened anterior and posterior articulating flanges. The distal sections of the pleura are not known. COMPARISONS. 5. (Scutellum) delicatum is most closely related to 5. (Scutellum) costatum but is distinguished from it by the absence of coarse ornamentation and by the prominent tumid axis and median rib on the pygidium. Scutellum (Scutellum) flabelliferum (Goldfuss) (PL 3, ngs. 1-3) 1839 Bronteus flabellifer Goldfuss: 361, pi. 33, fig. 3« (excluding pygidium), y. 1954 Scutellum (Scutellum) flabelliferum (Goldfuss) ; Kielan : 28, pi. 4, figs. 6-9, text-figs. 20-23 (includes earlier synonymy of the species). 1955 Bronteus (Bronteus) cf. flabellifer Goldfuss ; Maksimova : 43, pi. i, figs. 2, 3. 1956 Scutellum (Scutellum) flabelliferum (Goldfuss) ; Richter, R. & E. : 96, pi. 7, figs. 40-42. REMARKS. Whidborne (18890) and R- & E. Richter (1925) correctly pointed out that because of confusion over specimens figured by Goldfuss (1839, I^43) many specimens from Devon ascribed to this species are properly placed in S. (Scutellum) costatum Pusch (= Bronteus granulosus Goldfuss) . Scutellum flabelliferum is of rare occurrence and has only been identified from the Chercombe Bridge Quarry where it occurs in a black limestone in association with Dechenella (Dechenella) setosa, Whidborne. The preservation of this material is, in some cases, particularly good ; one large pygidium (PI. 3, fig. 2) shows delicate tooth-like extensions arising from the anterior-lateral border. The median rib of the specimen figured by Whid- borne (here re-figured : PI. 3, fig. 3) is unusual for the species for it has the form more characteristic of Scutellum costatum. The presence of this variant within the assemblage tends to confirm that the latter species was derived from Scutellum flabelliferum. VIII. REFERENCES DINELEY, D. L. 1961. The Devonian system in South Devonshire. Fid. Stud., London, 1 : 121-140. ELLIOTT, G. F. 1961. A new British Devonian alga, Palaeoporella lummatonensis, and the brachiopod evidence of the age of the Lummaton Shell-Bed. Proc. Geol. Ass. Lond., 72 : 251-260, pis. 9-10. GOLDFUSS, A. 1839. Beitrage zur Petrefaktenkunde. B. Beitrage zur Familie der fossilen Crustaceen. Nova Acta Acad. Caesar. Leop. Carol., Halle a.S., 19 : 353-364, pi. 32. — 1843. Systematische Ubersicht der Trilobiten und Beschreibung einiger neuen Arten derselben. Neues Jb. Min. Geol. Palaont., Stuttgart, 1843 : 537-567, pis. 4-6. HARRINGTON, H. J. 1959. General description of Trilobita. In R. C. Moore (editor), Treatise on Invertebrate Paleontology-O , Arthropoda i. (Trilobitomorpha) . xix + 560 pp., 415 figs. Lawrence & Meriden. HAWLE, I. & CORDA, A. J. C. 1847. Podrom einer Monographie der bohmishen Trilobiten. 176 pp., 7 pis. Prague. HOUSE, M. R. 1963. Devonian ammonoid successions and facies in Devon and Cornwall. Quart. J. Geol. Soc. Lond., 119 : 1-27, pis. 1-4. 220 BRITISH DEVONIAN TH YS ANOPELTID AE (TRILOBITA) KIELAN, Z. 1954. Les Trilobites m6sodevoniens des Monts de Sainte-Croix. Palaeont. polon., Warsaw, 6 : 1-50, pis. 1-7. MAKSIMOVA, Z. A. 1955. Trilobity srednego i verkhnego devona Urala i severnykh Mugodzhar. Trud. vsesoyuz. nauch.-issled. geol. Inst., Moscow (n.s.) 3 : 1-244, P^s- 1-18. MILLER, R. L. & KAHN, J. S. 1962. Statistical Analysis in the Geological Sciences, xiii + 483 pp. New York & London. PENGELLY, W. 1861. On the Devonian age of the World. The Geologist, London, 4, 332-347, pi. 7. PHILLIPS, J. 1841. Figures and Descriptions of the Palaeozoic Fossils of Cornwall, Devon, and West Somerset, xii + 231 pp., 60 pis. London. PUSCH, G. G. 1833. Geognostische Beschreibung von Polen, so wie der ubrigen Nordkarpathen- lander, 1 : 338 pp. Stuttgart & Tubingen. RICHTER, R. 1926. Von Bau und Leben der Trilobiten IV. Die Versteifungen der Schale und deraus hervorgehende Konvergenzen. Pal. hungarica, Budapest, 1 : 90-108. R. & E. 1925. Unterlagen zum Fossilium Catalogus, Trilobitae III. Senckenbergiana, Frankfurt a. M., 7 : 239-244. — 1926. Die Trilobiten des Oberdevons. Abh. preuss. geol. Landesanst., Berlin (N.F.) 99 : 1-314, 12 pis. — 1934. Missbildungen bei Scutellidae und konstruktive Konvergenzen Senckenbergiana, Frankfurt a. M., 16 : 155-160. 1939. Ueber Namuropyge n.g. und die Basisolution der Trilobiten Glatze. Bull. Mus. R. Hist. nat. Belg., Brussels, 15, 3 : 1-29. 1956. Grundlagen fiir die Beurteilung und Einteilung der Scutelluidae (Tril.). Senck. leth., Frankfurt a. M., 34 : 79-124, pis. 1-7. SELWOOD, E. B. 1965. Dechenellid trilobites from the British Middle Devonian. Bull. Brit. Mus. (Nat. Hist.) Geol., London, 10 : 319-333, i pi. SNAJDR, M. 1960. Studie o c"eledi Scutelluidae (Trilobitae). Rozpr. geol. Ust. csl., Prague, 26 : 1-264, pls- I-36- USSHER, W. A. E. 1913. The Geology of the Country around Newton Abbot. Mem. Geol. Surv. U.K., London, vi + 149 pp., 3 pis. WHIDBORNE, G. F. 1889. On some Devonian Crustacea. Geol. Mag., London (3) 6 : 28-20. — i88ga. A monograph of the Devonian fauna of the South of England. Vol. i. The fauna of the Limestones of Lummaton, Wolborou^h, Chircombe Bridge, and Chudleigh, Pt. i : 1-46, pis. 1-4. Palaeontogr. Soc., [Monogr.~\, London. EXPLANATION OF PLATES All specimens whitened with ammonium chloride before photographing. Speci- mens prefixed BM., RAMME, SMC and TM. are housed respectively in the British Museum (Natural History), London, the Royal Albert Memorial Museum, Exeter, the Sedgwick Museum, Cambridge, and the Torquay Natural History Museum. PLATE i Scutellum (Scutellutri) pardalios (Whidborne) FIGS. 1-3. Pygidium (BM I.ioo) Lectotype. [Wolborough Quarry], i. Plan view, xi. 2. Posterior view, xi. 3. Side view, xi. Figured by Whidborne i88ga, pi. 3, fig. 5. FIGS. 4-5. Cranidium (BM IT. 495) Lummaton Quarry. 4. Plan view, X2. 5. Side view, X2. FIG. 6. Cranidium. (BM IT. 496) Lummaton Quarry. Plan view, x6. Occipital ring with prominent mesial tubercle. FIG. 7. Cephalon (SMC H. 4,112) Lummaton Quarry. Plan view, X2-5. Figured by Whidborne 18890, pi. 3, figs. 4, 4». FIGS. 8-1 1. Pygidium (RAMME 7149) Wolborough Quarry. 8. Side view, xi. 9. Plan view, xi. 10. Posterior view, xi. n. Detail of ornamentation, X3- Bull. B.M. (N.H.) Geol. 13, 3 PLATE 1 ->•'' /• W--V *<*'/.' PLATE 2 FIGS, i, 2. Scutellutn (Scutellum) costatum lumrnatonensis subsp. n. Pygidium (RAMME 66 /oi) [Wolborough Quarry], i. Plan view, xi. 2. Side view, xi. FIG. 3. Scutellum (Scutellum) costatum lumrnatonensis subsp. n. Pygidium (SMC H. 3,801) Wolborough Quarry. Plan view, xi-5- FIGS. 4-6. Scutellum (Scutellum) costatum lumrnatonensis subsp. n. Pygidium (RAMME 66/02) [Lummaton Quarry]. Cast. 4. Plan view, xi. 5. Posterior view, xi. 6. Side view, x i . FIG. 7. Scutellum (Scutellum} costatum lummatonensis subsp. n. Pygidium (SMC H-2,254) Holotype. Lummaton Quarry. Plan view, xi-5- FIG. 8. Scutellum (Scutellum) costatum lummatonensis subsp. n. Pygidium (SMC H.2,253) Lummaton Quarry. Plan view, X2. FIG. 9. Scutellum (Scutellum) costatum whidbornei subsp. n. Pygidium (BM IT. 497) Lummaton Quarry. Plan view, x i . FIG. 10. Scutellum (Scutellum) costatum costatum Pusch. Pygidium (TM 6.490) Ransleigh Quarry. Plan view, xi-5. FIGS, ii, 12. Scutellum (Scutellum) costatum whidbornei subsp. n. Pygidium (BM IT. 498) Lummaton Quarry, n. Plan view, xi. 12. Side view, xi. FIGS. 18, 14. Scutellum (Scutellum) costatum whidbornei subsp. n. Pygidium (SMC 11.3,778) [Wolborough Quarry]. 18. Side view, xi. 14. Plan view, xi. FIGS. 15, 16. Scutellum (Scutellum) costatum whidbornei subsp. n. Pygidium (SMC 11.3,782) [Wolborough Quarry]. 15. Side view, xi. 16. Plan view, xi. FIGS. 17, 13. Scutellum (Scutellum) costatum whidbornei subsp. n. Pygidium (BM 1.5078). Holotype. Wolborough Quarry. 17. Plan view, xi-25. 13. Side view, xi-25. Bull. B.M. (N.H.) Geol. 13, 3 PLATE 2 10 PLATE 3 FIGS, i, 2. Scutellum (Scutellum) flabelliferum (Goldfuss). Pygidium (RAMME 77/04) Chercombe Bridge Quarry, i. Plan view, xi. Cast. 2. Detail showing toothed margin at the anterior lateral margin, x 10. Cast. FIG. 3. Scutellum (Scutellum) flabelliferum (Goldfuss). Pygidium (BM 1.5058). Chercombe Bridge Quarry. Plan view, X3. Figured by Whidborne 18890, pi. 3, fig. 16. FIG. 4. Scutellum (Scutellum) delicatum delicatum (Whidborne). Pygidium (BM 1.1084) Lummaton Quarry. Plan view, xi. Figured by Whidborne 18890, pi- 3, fig. 15. FIG. 5. Scutellum (Scutellum) delicatum tigrinum (Whidborne). Pygidium (SMC 11.3,785) Lummaton Quarry, Plan view, xi'5- Cast. FIG. 6. Scutellum (Scutellum) delicatum tigrinum (Whidborne). Pygidium (BM In. 58998) Lummaton Quarry. Plan view, xi. Figured by Whidborne 18890, pi. 3, fig. 8. FIG. 7. Scutellum (Scutellum) delicatum delicatum (Whidborne). Pygidium (SMC 11.3,804) [Lummaton Quarry]. Plan view, xi. FIGS. 8, 9. Scutellum (Scutellum) delicatum delicatum (Whidborne). Cranidium (SMC 11.4,107) Lectotype. Wolborough Quarry. 8. Plan view, xi-5- 9. Side view, x 1-5. Figured by Whidborne 18890, pi. 3, figs. 14, 140. FIG. 10. Scutellum (Scutellum) delicatum tigrinum (Whidborne). Cephalon (BM IT. 499) Lummaton Quarry. Plan view, X2-5. FIG. ii. Scutellum (Scutellum) delicatum tigrinum (Whidborne). Cranidium (BM In. 34923) Lectotype. [Lummaton Quarry]. Plan view, X2-5. Figured by Whidborne 1889, a pi. 3, fig. 12. FIG. 12. Scutellum (Scutellum) delicatum tigrinum (Whidborne). Cranidium (SMC H. 4,106) Lummaton Quarry. Plan view, X4. Figured by Whidborne 18890, pi- 3. fig- I3- Bull. B.M. (N.H.) Geol. 13, 3 PLATE 3 PRINTED IN GREAT BRITAIN BY ADLARD & SON LIMITED BARTHOLOMEW PRESS, DORKING EXINE STRUCTURE IN SOME FOSSIL AND RECENT SPORES AND POLLEN AS REVEALED BY LIGHT AND ELECTRON MICROSCOPY J. M. PETTITT BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 13 No. 4 LONDON: 1966 EXINE STRUCTURES IN SOME FOSSIL AND RECENT SPORES AND POLLEN AS REVEALED BY LIGHT AND ELECTRON MICROSCOPY BY JOHN MICHAEL PETTITT Pp. 221-257; 2I Plates; I Text-figure BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 13 No. 4 LONDON: 1966 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 ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 13, No. 4 of the Geological (Palaeontological) series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. Trustees of the British Museum (Natural History) 1966 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 4 October, 1966 Price £2 2s. EXINE STRUCTURE IN SOME FOSSIL AND RECENT SPORES AND POLLEN AS REVEALED BY LIGHT AND ELECTRON MICROSCOPY By JOHN MICHAEL PETTITT CONTENTS Page I. INTRODUCTION .......... 223 II. METHODS OF INVESTIGATION ....... 224 III. DESCRIPTIONS .......... 225 SPORES OF LIVING AND FOSSIL HETEROSPOROUS PTERIDOPHYTES . 225 SPORES OF LIVING AND FOSSIL HOMOSPOROUS PTERIDOPHYTES . 24! POLLEN OF FOSSIL AND LIVING GYMNOSPERMS .... 243 THE MEGASPORE MEMBRANE OF PALAEOZOIC SEEDS AND SEED-LIKE STRUCTURES ......... 247 IV. GENERAL DISCUSSION ......... 249 V. REFERENCES .......... 254 SYNOPSIS This paper gives descriptions of the wall structure of the spores and pollen grains from a variety of living and fossil plants. The spores of Palaeozoic heterosporous plants are compared with those of extant homo- sporous and heterosporous pteridophytes, and Carboniferous pteridosperm and cordaite pollen is compared with that of living cycads and conifers. It has been found that the structure of the megaspore membrane in some Palaeozoic gymno- spermous ovules more closely resembles the spore exine of some free-sporing heterosporous plants than the megaspore membrane of living gymnosperms. Similarly, the exine structure of pteridosperm and cordaite pollen is closer to that of some pteridophyte spores than to the pollen of living gymnosperms. Special attention has been paid to the origin and distribution of the various wall layers, particularly the mesospore and the perispore. I. INTRODUCTION THE discovery by Ehrlich & Hall (1959) that fine structural detail can be preserved in the exines of Eocene angiosperm pollen has prompted Larson & Lewis (1961) to suggest that the evolutionary development of the pollen wall could be subjected to direct study. More recently it has been shown that Classopollis, a Mesozoic conifer pollen pos- sesses a distinctive wall structure, the complexity of which is unmatched even among living angiosperms (Pettitt & Chaloner 1964). Afzelius (1956) has investigated the structure of the exine in the spores of living pteridophytes and the pollen of living gymnosperms and angiosperms, and has GEOL. 13, 4. 15 224 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN suggested that certain submicroscopic structures in the exine may be more primitive than others. She suggests that a pollen grain with a thin lamellated layer in the exine may be considered phylogenetically more advanced than one with a thick lamellated layer. Larson & Lewis (1961) discuss this hypothesis and advance the suggestion that fine structural variations in the exine could be related to functional differences. Such a hypothesis would appear to be open to experimental verification, and if it is eventually substantiated any phylogenetic interpretations based on exine structure alone would have to be qualified accordingly. It has been found that exine fine structure can be preserved in the fossil spores of free-sporing Palaeozoic pteridophytes and in the fossil pollen of Carboniferous seed plants, and that the stratification revealed in ultra- thin sections can be directly compared with the exine structure in the spores of extant pteridophytes and the pollen of living gymnosperms. The following is an attempt to synthesize the information so far obtained. II. METHODS OF INVESTIGATION The fossil material was released from the enclosing sediment with cold commercial (40%) hydrofluoric acid. Where necessary dilute nitric acid was used to clean and soften the spores and pollen prior to dehydration and embedding. Details of the fixation, embedding and histological procedures used in this investigation are given in Pettitt (1966). Ultra-thin sections were cut either on a Huxley-pattern or a Reichert ultramicro- tome using glass knives and mounted on copper grids with or without carbon reinforcement for observation with a Siemens' Elmiskop la or an A.E.I. EM6B electron microscope operating at 60 kV. Some use has been made of section stains and where they have been employed the fact is recorded in the figure legends. Thicker sections (about 0-2 /*) of methacrylate and epoxy resin embedded speci- mens examined under phase and anoptral contrast illumination have been used fairly extensively in the investigation of the thicker exines. The terminology adopted for the exine subdivisions revealed in the spore and pollen walls follows that proposed by Faegri (1956). Although this system of nomenclature was applied to explain the subdivisions detectable in the exine of angiosperm pollen following alcoholic basic fuchsin staining, Larson (1964) by using electron stains has shown that it is probably also applicable to the spores and pollen of non-angiospermous plants. A cknowledgments I am indebted to Mr. Clive Jermy of the Department of Botany, British Museum (Natural History) for supplying a great deal of the modern pteridophyte material, to Professor R. W. Baxter of the University of Kansas for allowing me to work on the Lawrence Shale fossils, to Dr. T. N. Taylor of the University of Illinois for sending me seed petifications and to Dr. W. G. Chaloner for providing the samples of Scottish coals. I should like to express my very sincere thanks to Professor EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 225 Dennis Lacy of the Department of Zoology, St. Bartholomew's Medical College, London who has allowed me free use of all the facilities in his department. I am also grateful to Professor Lacy and to Dr. W. G. Chaloner for their helpful advice and discussion. III. DESCRIPTIONS SPORES OF LIVING AND FOSSIL HETEROSPOROUS PTERIDOPH YTES Spores of Archaeopteris cf. jacksoni Dawson The occurrence of heterospory in this Devonian species has already been described (Pettitt 1965) and a full description of the morphology of the spores is given in that publication. Megaspores. The subdivision of the megaspore exine into an outer granular layer and an inner apparently homogeneous layer has been detected with the optical microscope (Pettitt 1965). At that time it was recorded that some, but not all, of the megaspores possessed a conspicuous inner body or mesosporium. It was not possible, however, to see how the inner body was formed nor to determine precisely how it was related structurally to the remainder of the spore wall. Ultra-thin sections of the megaspores from a sporangium of A . cf . jacksoni examined in the electron microscope support the interpretation based on the observations with the optical microscope. The thin, inner layer (endexine) which appears homogeneous in sections under the light microscope is seen to have a lamellate d structure, whilst the outer optically " granular " layer (ektexine) is composed of three-dimensional, anastomosing sporopollenin units giving the layer a spongy appearance (PI. 2, figs, i, 2 ; PI. 3, fig. 3). The outer spongy layer accounts for the greater part of the thickness of the wall. The subdivision of the inner lamellated layer or endexine into a series of concentric lamellae is evident in all the micrographs. In some sections there is partial or complete separation of this layer into an inner and an outer zone (PL 3, fig. 3). The separation would appear to be along one of the periods between the lamellae and the fact that it does not occur in all the megaspores, together with the observation that the actual site of the separation within the layer varies in the different spores would suggest that it is not a regular feature of ontogeny. From the evidence it seems reasonably certain that the separation of the endexine in these megaspores is responsible for the presence of the inner body or so-called mesosporium that can be observed in entire spores examined by transmitted light (Pettitt 1965). In the earlier publication I provisionally assigned the megaspores of A. cf. jacksoni to the megaspore genus Biharisporites Potonie (1956) using the presence of a meso- spore in some of the specimens as one of the features indicating a close correspondence with the spores of that genus. However, the discovery that the mesospore of these fossils is probably formed by the artificial separation of the inner layer of the exine into two immediately calls to question the validity of the feature as a taxonomic character. 226 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN The structure and formation of mesospore-like bodies in the spores of other fossil plants and in the spores of living pteridophytes is discussed later. Electron micrographs of sections through Archaeopteris megaspores that are situated at the outside of the sporangial mass show an unusual feature. On the surface of these spores dense aggregations of droplets occur in close association with the underlying spongy layer of the exine. The droplets vary in size, and consist of an outer stratum of homogeneous sporopollenin which becomes granular towards the centre of the body. The droplets have a hollow central lumen which is traversed by irregular wefts or threads of material with a coarsely granular fine structure (PI. 2, fig. i). There is no evidence of structural connexion between any of these bodies and the outer part of the exine, and they cannot be regarded as part of the wall ornamentation in the strict sense. Recent observations on the formation of sporopollenin droplets in the degenerating tapetal cells of gymnosperm pollen sacs (Pettitt 1966) leaves little doubt that the droplets in the Archaeopteris sporangia have a similar origin. Rowley (1963) has called somewhat similar droplets Ubisch bodies in the pollen sacs of angiosperms, but as Ubisch bodies seem to have a func- tional role in pollen wall formation in these plants it would be unwise to apply the name to the droplets in the Archaeopteris megasporangia. They are perhaps more comparable to the objects which Ueno (1959, 1960) has described as the perine in the pollen of some gymnosperms. This point is more fully discussed in a later section. Finally, it is evident that such deposits on the outer surface of the spore exine could easily be mistaken for sculptural elements in an uncritical examination. Microspores. The microspore exine of Archaeopteris cf. jacksoni is divided into two distinct layers which, in their ultrastructure, closely resemble the layers of the megaspore exine. The inner layer (endexine), which lines the lumen of the spore, is lamellated, whilst the outer layer (ektexine) is composed of ramifying units of sporopollenin giving the layer a spongy appearance (PI. 3, fig. i). In addition to the fine structure and stratification of the exine, thin sections through the spore wall show some other features that warrant description. The electron micrograph shown on PI. 3, fig. 2 is of a section which passes through one of the triradiate commissures. The section clearly shows that the lamellated endexine passes upwards through the commissure and forms a lining to the aperture. On the surface of the spore the endexine overlaps on to the outer layer of the exine and gradually thins out. It seems probable that the structures just described are lips, although lips surrounding the triradiate laesurae of these microspores were not detected with transmitted light (Pettitt 1965). The exine ornamentation of the microspores of A . cf . jacksoni has been described as being composed of minute conical elements about 1-1-5 /* m height, and I fi or less wide at the base (Pettitt 1965), and on the basis of this they were provisionally assigned to the genus Cyclogranisporites Potonie & Kremp 1954. The electron micrographs show how the conical elements are constructed from the sporopollenin units forming the outer exine layer (PI. 3, fig. i). EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 227 The spores of Selaginella and Isoetes Megaspores of Selaginella. The electron micrographs of the acetolyzed megaspore of Selaginella selaginoides published by Afzelius, Erdtman & Sjostrand (1954) show that the wall is composed of a " three-dimensional network of rounded bars ". An almost identical fine structure has recently been observed in Selaginella myosurus and Selaginella kraussiana by Stainier (1965), and from the observations made during the present investigation it would seem that the outer layer of the megaspore exine in a number of Selaginella species is of this composition. The following descrip- tion is, however, based mainly on the megaspores of Selaginella pulcherrima Liebm. Sections of the megaspore wall examined by phase contrast illumination show that there are structural variations within the outer exine layer. These are evident as concentric zones in the ektexine that appear to result from variations in the amount of interstitial space between the elements composing this layer of the wall. Electron micrographs show that at the outer edge of this outer layer the sporopollenin units become discontinuous and the wall texture is more open or porous in appear- ance. Towards the inner limit of the layer the sporopollenin elements gradually become reduced in thickness and less widely spaced and the boundary is seen as a zone of more compact structure (PL 4, figs, i, 2). Martens (1960) and Stainier (1965) have detected obvious zoning in the megaspore exine of S. myosurus and they find that the sporopollenin units in certain parts of the exospore have a definite orientation which would seem to be unique to this species. The acetolyzed megaspore wall of Selaginella consists mainly of a very thick, spongy outer layer or ektexine. On the inside of this, however, is a much thinner layer, the endexine, which accounts for not more than one-twentieth of the total thickness of the exine. Although the endexine has not been adequately photo- graphed in the electron microscope it can be clearly seen in thin sections that were examined and photographed by anoptral contrast illumination (PI. 5, fig. i). In these sections the endexine appears to be faintly lamellated, and the junction between it and the ektexine is very distinct. Fitting (1900) published a very detailed account of megaspore development and structure in Isoetes and Selaginella. As his observations are of significance to the present study the more relevant of them should perhaps be briefly reviewed. Fitting recognizes three or sometimes four layers in the megaspore wall of Sela- ginella. The outer layer in some species is a thick silicified " perispore ", but this is not always present. Underlying the perispore is the exospore which is often differentiated into two layers. Within the exospore is the mesospore, a very thin, yellowish membrane that is easily separated from the overlying layer. The inner- most layer recognized by Fitting is a cellulosic endospore. Fitting followed the development of the sporoderm in Selaginella megaspores and sets out a very precise series of events. He found that the exospore is the first formed layer and that this increases in thickness to between 0-5-1-2 /* (depending on the species) before the mesospore is formed. According to Fitting the meso- spore is formed as a new layer and does not arise as a result of splitting in the exo- spore. The fact that from its inception there is sometimes a gap between the 228 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN mesospore and the exospore is advanced as evidence for formation de novo. Follow- ing its formation the mesospore rapidly increases in thickness. In this connexion it is worth noting that Fitting observed two rather relevant differences in the megaspores of Selaginella galeotii on the one hand and Selaginella spinulosa and Selaginella helvetica on the other. In the first species a distinct margin between the exospore and the mesospore is not evident, whilst in 5. spinulosa and 5. helvetica there is a division in the middle of the exospore extending from the proximal to the distal surface which divides the layer into two membranes of equal thickness. Fitting noticed that as the megaspores increase in size the exospore and the mesospore become separated from each other and from the spore protoplast. The layers remain firmly attached at the spore apex and the separation can be detected first in the distal region of the spores and later in the equatorial region. There is apparently some variation in the stage at which this separation occurs and in the extent to which it finally develops. An extreme example would seem to be the megaspores of S. spinulosa where the two layers remain firmly attached at all times. In the megaspores of some species bars or trabeculae, which have the same composi- tion as the spore wall, were seen to link the exospore and the mesospore together. These structures are rather short-lived, and with the increase in size of the spores they become thinner and finally rupture. Fitting regards the presence of the trabeculae in Selaginella as indicative of close connexion between the exospore and the mesospore in young megaspores, and he believes that they originate as a result of polymerization or coalescence of the two membranes at their points of contact. The trabeculae are formed at the sites of contact as the membranes draw apart during spore growth and the space between them increases. The space between the two wall layers and between the mesospore and the spore protoplast is filled with an interstitial substance (more evident in older spores) which Fitting found to give a pectin reaction and also to stain with Congo red and Analine blue. The interstitial substance gradually disappears as the megaspores mature and the mesospore expands until it makes contact with the exospore. Both layers then grow in circumference with the result that in the mature megaspores the thick mesospore present in the early stages of development is present as a thin membrane closely associated with the exospore. Somewhat later the cellulosic endospore is formed and the protoplast increases in mass and fills the spore lumen. About this time the perispore is formed from the lamellae of the special mother cell wall. Of the two papers published by Lyon (1901, 1905) describing exine formation in Selaginella the latter is the most pertinent to the present account. In this, Lyon questions some of the conclusions reached by Fitting (1900) and revalues the earlier work of her own. Lyon detected only two megaspore wall layers in the material she investigated ; the exospore, which is formed first, and the endospore, which is lamellated. Some confusion arises here because the layer that Lyon (1905) calls the endospore is the layer which Fitting (1900) calls the mesospore. Lyon does not find a mesospore sensu Fitting in her material. EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 229 Lyon's investigations lead her to conclude that the exospore and the endospore (mesospore of Fitting) are sometimes formed simultaneously, and that the space which occurs between these two layers is an artifact caused by shrinkage. Lyon's evidence for this suggestion is based on the observation that in some megaspores the membranes do not separate. Lyon also finds that in young megaspores there is a tendency for the exospore to split into an outer and an inner layer and that the inner part of the exospore, to which the endospore is attached, could be interpreted as a separate wall layer. Campbell (1902) suggests that the cavity between the wall layers of the mega- spores originates as a result of the exospore enlarging more rapidly than the meso- spore. According to this author the mesospore closely resembles the exospore in appearance and could arise by the separation of an inner layer from the exospore. More recently, Stainier (1965) has recognized four wall layers in the megaspore of 5. myosurus, which correspond to the four detected by Fitting (1900), and three in 5. kraussiana. This author finds that the mesospore is separated from the exospore in both these species and concludes that the separation in 5. myosurus is a natural phenomenon which occurs during spore growth but in S. kraussiana it is probably a fixation artifact. The descriptions given by Fitting (1900), Lyon (1905), Campbell (1902) and Stainier (1965) show that whilst there is some measure of agreement in the naming of the various exine subdivisions that these authors detected in the megaspores of Selaginella it is far from being absolute. It is difficult, for example, to accept that the layer which Fitting describes and illustrates as the mesospore in young Selaginella megaspores is identical with the layer to which Campbell gives the same name. It is apparent that the term mesospore is being applied to very different structures. The importance of an inner body or mesospore as a microscopical character in the recognition of a number of fossil microspore and megaspore genera has already been mentioned. There is some difference of opinion as to whether this structure is an artifact (see Kremp 1965) and the conflicting evidence from living material hardly helps to resolve these differences. It is of some interest, therefore, to try to establish the true nature of the mesospore. Sections of 5. pulcherrima megaspores containing mesospores of the type illustrated by Fitting (1900, pi. 6, fig. 23) in his immature megaspores were subjected to a variety of histochemical tests. It should be pointed out that possibly only one or two of the four megaspores in any one sporangium contained mesospores of this type. Judging solely from their external appearance all the spores at first seemed to be perfectly normal. However, sections showed that normal protoplasmic contents, whilst present in some of the spores, are completely lacking in the spores that contain a mesospore. This observation completely contradicts those of Fitting (1900) and Lyon (1905). Structurally the central part of the mesospore appears fibrous under the light microscope at high magnifications and the fibrosity gives way to a more granular structure towards the periphery of the body. In the centre is a hollow cavity (PI. 5, fig. 4). Trabeculae linking the mesospore with the ektexine (exospore of 230 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN Fitting) are sometimes present. The endexine could not be detected in mega- spores that contained a mesospore. Sections stained with Sudan black B show that the entire mesospore and the ektexine are strongly sudanophilic (PL I, fig. 5). Following treatment by the periodic acid/Schiff (PA/S) procedure the central, fibrous zone of the mesospore gives a vivid red colour, but there is no reaction from the outer granular zone, nor from the ektexine (PL i, fig. 3). This staining pattern is repeated in sections stained with ruthenium red. It is, however, reversed in sections stained with Nile blue sulphate. The inner fibrous zone stains faintly, if at all, whilst the outer granular zone of the mesospore and the ektexine stain a strong blue colour (PL I, fig. 4). From these tests it can be reasonably concluded that the inner fibrous zone of the mesospore gives positive indications of the presence of lipids and carbohydrates. Further characterization of the nature of the body was obtained after treating sections with mercuric bromophenol blue, developed as a protein indicator by Mazia, Brewer & Alfert (see Pearse 1961). It was found that the ektexine and the outer granular layer of the mesospore remains unaffected by this test, but that the inner fibrous zone is coloured faintly blue-green (PL 5, fig. 3).1 As a measure of control, sections of megaspores that did not contain a mesospore were subjected to the same histochemical procedures. Following acetylated Sudan black B staining the spore protoplast colours strongly and three sudanophilic wall layers are recognizable ; the ektexine, the endexine which is frequently detached from the ektexine except at the spore apex and a thin membrane closely associated with the spore protoplast. The identity of this last layer is only clearly recognizable where the protoplast has become separated from the outer wall layer (PL i, fig. 6). At the apex of the megaspore, where cellular differentiation has occurred, the mem- brane is seen to have anticlinal walls which closely match the position of the cell walls of the underlying gametophyte tissue. So far, this layer has been detected only in the megaspores of S. pulcherrima and in this respect these spores would seem to be unique. Between the membrane just described and the endexine is an optically structure- less layer that does not colour with the Sudan stain but which stains intensely with ruthenium red. This layer, which is not always detectable can, by analogy with pollen wall terminology, be called the intine and is presumably that which Fitting (1900) and Stainier (1965) call the endospore. The ektexine is coloured bright blue after Nile blue sulphate staining, whilst a subtending layer, which from its position would appear to be the intine, stains a strong purple colour. The blue coloration of the ektexine following the Nile blue sulphate method indicates that this layer contains a high proportion of acidic lipids, whilst the purple colour of the inner layer following the same test would suggest that this contains predominantly neutral lipids. The discovery that the inner layer is composed of pectic substances that are stained with ruthenium red and neutral lipids is somewhat surprising, and perhaps the analogy with the pollen 1 It has proved impossible to obtain a colour-faithful photomicrograph of this test and, consequently, it has not been included in PI. i. EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 231 inline extends only to the location of the layer and not to its composition. Stainier (1965) has shown that the endospore of 5. myosurus is cellulosic. Following section treatment by the PA/S test the intine and the thin sudanophilic membrane that surrounds the protoplast give a strong red colour characteristic of a positive Schiff reaction. The protoplasmic contents of the megaspore stain faintly (PL i, figs. I, 2). Preparations treated with mercuric bromophenol blue reveal the presence of proteins in the megaspore protoplast which is coloured blue-green after this pro- cedure (PL 5, fig. 2, see footnote on p. 230). No reaction is seen from the megaspore wall layers. The evidence from these histochemical methods strongly suggests that the meso- spore, sensu Fitting, in these spores is not composed of sporopollenin. Although the presence of lipid in the ektexine, the endexine and the mesospore is clearly demon- strated by Sudan black B staining, the mesospore, or more correctly the central zone of it, also gives positive reactions for carbohydrates, pectin and proteins. Carbohydrates, pectin and proteins were not detected in the ektexine or endexine, but carbohydrates and proteins do occur in the normal megaspore protoplast. The only conclusion that can be drawn from these results is that the mesospore is chemi- cally more closely allied to the spore protoplast than to the exine. The occurrence of the mesospores in megaspores without a recognizable protoplast together with the histochemical evidence suggests that the inner fibrous zone of the mesospore represents an abortive protoplast that has undergone serious deformation at some stage during its development. The outer granular zone of the body behaves chemically like the exine material and the presence of this layer may be correlated with the apparent absence of an endexine in these spores. Presumably the resistance of the mesospore to acetolysis could be explained by the presence of the lipid and protein. An entirely different interpretation can, however, be placed on the structure which Fitting (1900) and Stainier (1965) designate the mesospore in mature spores of Selaginella. In mature megaspores of S. pulcherrima and Selaginella padangensis Hieron. with normally developed protoplasts the endexine is frequently found detached from the ektexine. This is clearly seen in sections of acetolyzed mega- spores where the endexine is present as a distinct inner layer, separated from the ektexine by a narrow cavity, and enclosing a central lumen. These preparations also show that the ektexine is frequently split, particularly in the apical region, a phenomenon to which Lyon (1905) draws attention. Sections through the apical region of the megaspores show that the endexine is attached to the ektexine in the region of the triradiate sutures in precisely the way that Fitting (1900) describes for his " exospore " and " mesospore ". Not infrequently it can be seen attached over a much greater part of its circumference. The frequent apical attachment of the endexine to the ektexine is explained by serial sections cut tangentially through the proximal region of the megaspores. These show that the endexine passes up into the triradiate commissures of the spore in exactly the same way as in the micro- spores of Archaeopteris (PL 5, fig. 5). 232 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN It would appear that the mesospore recognizable in the normal, mature and presumably viable megaspores of Selaginella is in no way identical to the structure given that name by Fitting (1900) in the spores which he believed to be immature but which are probably abortive. In normal, mature Selaginella megaspores the mesospore is formed by the endexine separating from the ektexine around most of its circumference as suggested by Stainier (1965) for 5. kraussiana, but remaining attached to the outer layer at the triradiate laesurae. Fossil megaspores containing precisely this type of structure are described in the previous section. In contrast, the central body or so-called " mesospore " of abortive Selaginella megaspores is probably a protoplasmic remnant which has undergone considerable structural deformation. Whether or not this structure which will survive acetolysis and oxidative maceration would also survive fossilization is questionable. Microspores of Selaginella. Ultra-thin sections of microspores of 5. pulcherrima show that the entire thickness of the exine is composed of sporopollenin that appears homogeneous or faintly fibrous in the electron microscope (PL 5, fig. 6). There is no evidence of an inner lamellated endexine. The fine structure of the microspore wall of this species is, therefore, very different from that of the megaspore. Surrounding the entire microspore and interrupted only at the triradiate com- missures is a thin, electron-opaque layer which has a somewhat granular ultra- structure. In the megaspores of the same species a well-demarcated, phase-white, granular layer can be seen on the outside of the exine when sections are examined by anoptral contrast illumination (PL 5, fig. i). Fitting (1900) describes a granular silicified layer on the outer surface of the Selaginella megaspores to which he gives the name " perispore " and it is reasonable to assume that the granular electron- dense and phase- white layers in S. pulcherrima could bear the same interpretation. The optical appearance of the layer suggests that it is an inorganic substance (prob- ably silica) that has been deposited on the outer surface of the exine. The question of whether this should rightly be called a " perispore " in the sense that the term has now been defined is discussed later. Following histochemical tests with Sudan black B, Nile blue sulphate, PA/S and mercuric bromophenol blue the microspores of Selaginella behave in much the same way as those megaspores lacking a mesospore. The exine is seen to be sudanophilic and colours blue after treatment with Nile blue sulphate, but gives no reaction after the PA/S test. The siliceous " perispore ", however, of both the microspore and megaspore stains vividly with celestin blue (PL I, fig. 2). The spore protoplast is stained faintly red after PA/S treatment, and a strong blue-green colour after section treatment with mercuric bromophenol blue. The microspore exine is not coloured by the protein test. Megaspores of Isoetes. In his account of the structure of the megaspore of Isoetes Fitting (1900) divides the wall into four layers which are, from the outside inwards, a strongly silicified perispore, an exospore (which is often split into three sub-layers), a thin mesospore and a cellulosic endospore. The extent to which the perispore is developed varies somewhat in different species and it forms the ornament EXINE STRUCTURE IX FOSSIL AND RECENT SPORES AND POLLEN 233 in those megaspores which are ornamented. The ornamentation is repeated in the underlying exospore. There is, therefore, a close association between the perispore and the exospore in Isoetes and some interbedding is not uncommon. Fitting's study shows that the perispore is laid down only when the megaspores are mature. In most species of Isoetes Fitting finds that the exospore is clearly divided into three subdivisions or lamellae. This feature is particularly clear in those species which have a thick exospore, and apparently transitions exist within the genus from spores in which exospore stratification is clearly recognizable to those in which none can be seen. Fitting describes the exospore of Isoetes as being composed of small, dark brown " crystalline " rods which are arranged parallel to the surface of the spore. The rods are more closely packed in the middle zone of the layer than in the outer or the inner zones. The mesospore is described as a thin layer that in both young and mature spores is always somewhat separated from the exospore and which has the same chemical composition as the latter. The cellulosic endospore is intimately connected with the mesospore which, according to Fitting, exhibits some stratification following treatment with cupram- monium. Between the endospore and the spore protoplast is a thin " pellicle " that gives a positive reaction for pectin. In Isoetes, as in Selaginella, the exospore is the first wall layer to be formed and the mesospore, which Fitting believes to arise de novo as in Selaginella, is formed somewhat later. The formative events closely parallel those of Selaginella and as the megaspore increases in size the mesospore separates from the exospore, remaining attached at the proximal pole. In some species, mesospore separation does not take place at all or only to a very limited extent. Endospore deposition occurs when the megaspores have almost attained their full size. One further point is worth recording. Fitting (following Tschistiakoff) recom- mends that the term " epispore " should be applied to the membrane or membranes formed from the " Epiplasma " (the tapetum ?) and the term perispore to that of those formed from the special mother cell wall. Following these definitions, the outer silicified layer of Isoetes megaspore walls, which Fitting describes as originating from the inner layers of the special wall, would be a perispore. Ultra-thin sections for electron microscopy which show exine structure in Isoetes megaspores have not so far been obtained. Thin sections for examination by optical microscopy are much easier to cut and form the basis of the following descrip- tion. PI. 5, figs. 7, 8 are anoptral contrast photographs of thin sections of megaspores of Isoetes humilior F. Mull, ex A. Braun that were soaked in hydrofluoric acid to remove the siliceous material from the exine and acetolyzed prior to embedding in epoxy resin. The resolution attainable by this method of examination is sufficient to confirm Fitting's (1900) observation that the exospore (ektexine) is composed of small " crystalline " rods arranged parallel to the surface of the spore. Histochemi- 234 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN cal tests on megaspores not subjected to hydrofluoric acid treatment show, however, that the exine elements or rods are composed of sporopollenin. Although the arrangement of the sporopollenin elements forming the ektexine in the megaspores of this genus somewhat resembles that seen in Selaginella, there are some marked differences. Comparison of the micrographs of Selaginella (PL 4, figs, i, 2 ; PL 5, fig. i) with those of Isoetes show that in the latter the ramifying rodlets of the ektexine are thinner and much less densely packed than in Selaginella, the general effect being that the Isoetes exine appears to have a more " open " texture. A second difference is that an inner lamellated endexine was not detected in the megaspores of Isoetes. In a number of megaspores of Isoetes that were examined as both sections and entire objects the inner part of the ektexine is seen to be separated from the re- mainder (PL 5, fig. 8 ; PL 6, fig. i) as described by Fitting. The separation would seem to be quite unrelated to the age of the spores, and evidently occurs along a plane of weakness in the exine. It is never complete around the full circumference of the megaspore and attachment is always maintained at some point, usually the proximal pole. Microspores of Isoetes. Electron micrographs of Isoetes echinospora Dur. micro- spores show that the exine consists of two layers, the outer is lamellated and the inner homogeneous (PL 6, figs. 2, 3). The lamellae of the outer layer appear to anastomose in some sections and the layer then appears structurally not unlike the ektexine of the megaspores. The arrangement of these subdivisions is particularly interesting because the microspores of this genus are the only spores so far examined where the outer layer of the exine is lamellated. Nilsson & Praglowski (1963) however, call this layer the perine and do not, therefore, regard it as part of the exine sensu stricto. This interpretation, together with the question of just what does and what does not constitute a perine, is discussed later. The electron micrographs and phase contrast pictures show that in Isoetes, although there are some structural differences between the exine of the microspores and that of the megaspores, they are not so marked as they are in Selaginella. The phase contrast pictures of sections of acetolyzed microspores from a single sporangium of Isoetes echinospora (PL 7, figs. 1-8) show that the inner layer of the exine is invariably detached from the outer, but the occasional spore exhibiting complete unity between these two layers, presumably the normal condition (PL 7, fig. i), serves to suggest that the separation is either an unusual feature of develop- ment or an artifact induced by fixation. Variations in the extent to which the layers have separated are shown. In some of the microspores the inner layer is attached only at the proximal pole (PL 7, figs. 4, 5) and in others only at the distal pole (PL 7, figs. 2, 3). One example shows a condition where the inner membrane is attached laterally and is completely free at both the proximal and distal poles (PL 7, fig. 8) and others where the inner membrane is completely free from the outer around its entire circumference (PL 7, figs. 6, 7). Thus, within a single micro- sporangium of Isoetes, spores can be found that could be classified when dispersed EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 235 as either cavate or not so, if the current definition for this structural condition was applied to them (see Dettman 1963, Kremp 1965). Somewhat similar variations in the arrangement of the wall layers occur in the microspores of Marsilea that are described below. The wall structure of Laevigatisporites cf. glabratus Megaspores identified as Laevigatisporites glabratus (Zerndt) Potonie & Kremp 1954 were recovered from the Lawrence Shale (Virgillian) of Lone Star Lake, Kansas by digesting the sediment in cold hydrofluoric acid. Megaspores of the Laevigati- sporites type have been recorded from a number of Carboniferous Sigillarian fructi- fications (see Chaloner 1953). Electron micrographs (PI. 7, fig. 10) and phase contrast photographs (PI. 6, fig. 5) show that the entire thickness of the exine is composed of the familiar ramifying units of sporopollenin. The presence of an inner wall layer (endexine) bordering the spore lumen was not detected, but Schopf (1941) has found two wall layers in the microspores of Mazocarpon oedipternum (a Sigillarian cone with L. glabratus megaspores) . The spores of Marsilea and Regnellidium Megaspores of Marsilea. Feller (1953) and Boterberg (1956) have both described the structure and development of the megaspores and microspores in Marsilea — • Feller in Marsilea hirsuta and Boterberg in Marsilea diffusa. Both authors compare their observations with those published by Meunier (1888) on the structure of the sporocarp and spores of Pilularia globulifera and Boterberg compares his with those of Bonnet (1955). Although Feller's and Boterberg' s respective accounts of spore ontogeny in Marsilea differ in some relatively minor details they both conclude that the mega- spore in the species they have investigated is composed of an inner endospore and an outer exospore. Three outer, rather unique gelatinous layers surround the exospore, but these are not considered in this account. According to these two authors the endospore is the first megaspore wall layer to develop and when fully differentiated in the mature megaspore it consists of two layers. From the evidence presented by these authors the inner endospore probably corresponds to the intine of an angiosperm pollen wall ; the outer layer is sudano- philic. In young megaspores the endospore is .often found detached from the remainder of the wall. The structure of the exospore of mature Marsilea megaspores is very complex. Immediately overlying the endospore is the inner division of the exospore which Feller (1953) and Boterberg (1956) call the " reticulated layer ". This layer is continuous around most of the megaspore but gradually thins and disappears in the region of the germinal sutures. The " prismatic layer " forms the outer division of the exospore and also the outer layer of the megaspore exine sensu stricto. Both authors give a precise and detailed account of the development of the " prismatic 236 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN layer " in Marsilea and the reader is directed to their publications for this informa- tion. In mature megaspores the " prismatic layer " consists of a basement mem- brane upon which is arranged a number of radially-directed elements that, under high magnifications, can be resolved as small prismatic chambers. According to Boterberg (1956) the chambers are open to the outside and during development are penetrated by digitations of the tapetal periplasmodium. Towards the spore apex the " prismatic layer " decreases in thickness and is not developed in the region of the germinal sutures. The value of Ehrlich's haematoxylin and Orange G, Congo red, Fast green and Analine blue used by Feller (1953) and Boterberg (1956) in detecting the first appear- ance and subsequent deposition of sporopollenin in the wall layers is very limited. For example, Boterberg (1956) states that in mature or almost mature megaspores the " reticulated layer " is stained by Fast green, Analine blue, Congo red and Orange G. None of these stains is specific or even selective for lipoidal substances. The electron micrographs published here (PL 8, fig. 2) show that the " reticulated layer " in acetolyzed megaspores of Marsilea quadrifolia L. is composed of anastomos- ing rodlets of acetolysis-resistant material, and that the spaces between could have been occupied by some material that did not survive acetolysis. It is not impossible that the stains employed by Feller and by Boterberg are in fact colouring some interstitial substance between the sporopollenin elements rather than the exine material proper, even assuming that this had in fact been formed. It is a well known fact that in the megasporangium of Marsilea only one fertile spore is produced. Although all the sporocytes in the megasporangium undergo meiosis, only a single spore from one of the tetrads so formed develops to maturity ; the remainder abort. Both Feller (1953) and Boterberg (1956) give an account of this very interesting phenomenon. They find that the three abortive spores that remain attached to the developing megaspore are surrounded by a membrane, the endospore, which is formed at the same time as the endospore of the fertile megaspore. Wall layers external to the endospore are not formed around the abortive members of the tetrad. Acetolyzed megasporangia from mature sporocarps of M. quadrifolia and Marsilea drummondii A. Br. yield megaspores which often bear small acetolysis-resistant bodies attached at the apex (PI. 9, figs. I, 2). This situation is very reminiscent of that seen in the Devonian and Carboniferous megaspore tetrads of Cystosporites. Thin sections through the apical region of acetolyzed M. drummondii megaspores examined by phase contrast illumination (PI. 9, fig. 2) show that the abortive spores are represented by small, irregularly-shaped masses of sporopollenin with small hollow vesicles, and appear very much as shown in the illustration of Boterberg (1956). There is none of the complex structural organization seen in the exine of the fertile megaspores. Electron micrographs of ultra-thin sections through the exine of a fertile mega- spore of M. quadrifolia illustrate that the exine stratification detected by Feller and by Boterberg in the megaspores of M. hirsuta and M. diffusa is the same in this species. EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 237 The inner layer of the exine (endexine) which is equivalent to the outer endospore of Feller and Boterberg is composed of sporopollenin which exhibits a granular structure in sections stained with potassium permanganate. On the inner border of this layer is a zone which is slightly more electron opaque than the remainder (PI. 8, fig. 2). The " reticulated layer " of the French authors is seen in electron mircographs to be composed, as mentioned above, of anastomosing rodlets of sporopollenin (PL 8, figs. 2, 3). The rodlets are more or less circular in cross-section and are surrounded by clear spaces that in the living spore were presumably filled with some interstitial substance. At the junction of the " reticulated layer" and the subjacent endexine the rodlets of the former are seen to arise directly from the somewhat irregular surface of the latter. There is thus structural continuity between the endexine and the innermost layer of the ektexine. The most clear indication of the structural organization of the " prismatic layer " is obtained from thin sections examined by phase contrast illumination (PL 8, fig. 4). The outer terminations of the radially-directed chambers are seen to be covered by a thin, continuous, acetolysis-proof membrane. A membranous cover closing the prismatic chambers was not detected in M. hirsuta (Feller 1953) nor in M. diffusa (Boterberg 1956). The radial walls of the chambers are perforated with small, irregularly-shaped apertures which in sections of non-acetolyzed material are some- what obscured by aggregations of small phase-dark granules adhering to the walls. There is some visual evidence to suggest that at their innermost limit the walls of the prismatic chambers divide, although the junction between the " prismatic layer " and the " reticulated layer " appears quite distinct under the optical microscope (PL 8, fig. 4). Electron micrographs of this region show, however, that the elements of the " prismatic layer " are in places continuous with those of the " reticulated layer" (PL 8, fig. 3). The fine structural continuity between the three layers comprising the megaspore exine of M. quadrifolia suggests perhaps that sporo- pollenin deposition in the megasporangium is a continuous process. The implication here is that the very marked stratification in the megaspore wall of this genus does not necessarily mean that waves or cycles of sporopollenin deposition are separated by intervals when the mechanism, or mechanisms, responsible for wall formation is inactive. Microspores of Marsilea. Electron micrographs of radial sections through aceto- lyzed spores from the microsporangia of M. quadrifolia and M . dmmmondii clearly show the so-called " striated " or " prismatic " layer of the exine detected by Feller (1953) and Boterberg (1956). Tangential sections through this layer reveal, how- ever, that the striations are in fact a system of more or less radially-aligned tubes (PL 9, fig. 5) that are closed at their apices, but interconnected at their bases (PL 9, fig. 4) . The inner exine layer (endospore of Feller and Boterberg) appears to have a finely granular ultrastructure. Both these authors in their respective accounts of microsporogenesis in Marsilea show that the microsporangia contain; in addition to the normal and presumably fertile microspores, a variable number of abortive spores. Perhaps the most inter- GEOL. 13, 4. l6 238 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN esting of these from a morphological point of view are the pseudospores. Pseudo- spores are described as armcleate, aplasmic spore-like inclusions that develop within small vacuoles in the tapetal periplasmodium. Their formation begins with droplets, which Bonnet (1955) identifies as tannoids but which Feller's evidence would suggest are sporopollenin, becoming surrounded by a thin, sudanophilic membrane. As development proceeds the surrounding membrane assumes an appearance very reminiscent of the outer membrane of the microspores. The final stage of pseudo- spore development is marked by the formation of one or more gelatinous layers on the outside of the " striated " layer. Feller (1953) recognizes four types of pseudo- spores that are separable on the form of the central inclusion, and there are structural intermediates between the four types. None of these contains a nucleus or cyto- plasm and they do not appear to form in tetrads. " Microspores volumineuses " are a second type of abnormal spore described by Feller (1953) and Boterberg (1956) but these, unlike the pseudospores, are formed from the archesporium. Their formation is due to a blockage in meiosis at or before metaphase II (Boterberg 1956), with the result that individual haploid microspores are not formed. " Microspores volumineuses " are, therefore, diploid, but the French authors find that exine development proceeds more or less normally and eventually the diploid spores are surrounded by a sudanophilic wall consisting of a thick endospore and a thinner " striated " overlying membrane. In the space between these two layers are small droplets of various shapes and sizes. Acetolyzed microsporangia of M. quadrifolia and M. drummondii have been found to contain a large number of spores which from their structure when examined as entire objects could be equated with one or other of Feller's (1953) four types of pseudospores (PL 10, fig. i). Examination of thin sections of these structures with anoptral contrast illumination suggests that the situation in my material cannot be fully explained on the basis of the interpretation of pseudospore formation given by Feller. The sections show that each pseudospore is surrounded by an outer exine layer composed of radially-aligned tubes which seem to be exactly similar to the outer membrane of the normal, fertile microspores (PI. 10, figs. 3-8). There is, however, considerable variation in the extent to which the inner wall layer is developed. In some of the smaller pseudospores it is not present as a distinct membrane but is represented merely by small, discrete, phase-white droplets of acetolysis-resistant material adhering to the inside of the external layer. The partial or nearly complete coalescence of these droplets in some specimens appears to result in a structure not unlike a membrane (PI. 10, fig. 4). In somewhat larger pseudospores the inner layer can be seen as either a small central body having no structural connexion with the outer layer (PI. 10, fig. 5) or as a much folded membrane that has contracted away from the outer layer (PL 10, figs. 7, 8). In both these forms (some specimens of which may be " microspores volumineuses ") small acetolysis-resistant, phase- white droplets are attached to the inner surface of the outer membrane. PL 10, fig. 6 shows a section of a spore the wall of which seems to differ from the normal condition (PL 10, fig. 3) only in the slight thickening of the inner layer and the EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 239 partial contraction of this layer away from the external layer of the exine. It should perhaps be emphasized here that the abnormal spores or pseudospores constitute the greater part of the contents of the microsporangium of M. quadrifolia and M . drummondii. The sections of these structures show that in all of them, no matter how contorted the inner membrane may be, a central lumen is always present. Cytoplasmic contents could not, however, be detected in the sections of non- acetolyzed material treated with appropriate nuclear and cytoplasmic stains. Nevertheless, the presence of a central lumen would seem to suggest, although not prove, that at some time they did possess cytoplasmic contents. The evidence available from the sections indicates that the abnormal spores have to some extent followed the same developmental pattern as the fertile spores, but that at some stage during their development the process of wall formation has deviated from the normal. There are several relevant factors to be considered in this connexion. Firstly, both the normal and abnormal spores occur together in the same sporangium. Secondly, in all cases it seems to be only the inner exine layer of the abnormal spores that is malformed ; the outer membrane is precisely the same as in the normal spores. Electron micrographs confirm the optical observa- tions on this point. Finally, the outer wall layer often assumes a shape that is quite independent of that assumed by the inner membrane. Feller (1953) and Boterberg (1956) have found that in the development of the normal microspores the endospore is the first layer formed around the protoplast. If it can be assumed that this is also true for the abnormal spores then the endospore in these must have been formed around either a normal protoplast which sub- sequently degenerated or a protoplast that was degenerate from the start. The fact that the outer membrane in these spores is completely normal and does not necessarily follow the contours of the endospore would seem to lend support to the first alternative. However, there are points which still remain to be explained. Whilst the degeneration and subsequent contraction of the protoplast could be responsible for the collapse of the endospore it is difficult to see how it could account for the increased thickness of this layer in the abortive spores. The presence of acetolysis-proof droplets which may represent the inner layer in some spores and those that occur between the two wall layers in others also needs to be explained. It is difficult to reconcile Feller's (1953) interpretation of the origin and mode of development of aplasmic, anucleate pseudospores with the level of structural organi- zation revealed in the abortive spores described above. Heslop-Harrison (1963^, b) has shown that the protoplast of pollen grains is involved in the establishment of the exine pattern and it is of more than local interest to establish precisely how the spore wall is formed in the abortive microspores of Marsilea, and at what stage during the process the spore protoplast begins to degenerate. One further point should be mentioned here. Feller (1953) and Boterberg (1956) refer to small granules of material, which chemically resemble the exine substance, occurring in blebs that arise from the tapetal periplasmodium adjacent to the developing microspores. Feller, who follows their development, believes that these granules contribute towards the formation of the microspore exine. The granules 24o EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN may well be identical with the Ubisch bodies described by Rowley (1963) or the tapetal plaques detected by Heslop-Harrison (1962, 19636) in the pollen sacs of angiosperms. Shattuck (1910) described how the degree of spore abortion in the sporangia of Marsilea can be controlled by varying the external conditions under which the plants are growing. He shows that there is some correlation between the time during spore development at which the stimulus (a change in temperature or light intensity) is applied and the resultant fertility of the spores. The critical period in this respect would seem to be from the time the mother cells enter synapsis until the tetrads are well formed. Shattuck advances the suggestion that a mechanism of this sort could have been responsible for the origin of heterospory in Marsilea and possibly for its origin in all heterosporous plants. By varying the external conditions and reducing the number of developing sporocarps on a plant Shattuck was able to induce the formation of large spores or " megaspores " in microsporangia and small spores or " microspores " in megasporangia. Shattuck is not unaware of the very significant differences between heterospory and heterothally and he attempted, without success, to germinate the experimentally produced " megaspores " and " microspores ". Thus, the only reliable criterion that would test the validity of his claims, the production of megagametophytes and microgametophytes, could not be applied. Nevertheless, the possible significance of his observations cannot be ignored. The megaspores of Marsilea have, unquestionably, the most complex exine struc- ture of any megaspores from either living or fossil plants so far investigated. It is considerably more complex than that seen in the megaspores of the lycopods Selaginella and Isoetes. Although the exine structure of the microspores of Marsilea appears superficially similar to that of the megaspores, ultra-thin sections show that there is a large measure of difference between the two. Megaspore of Regnellidium. Chrysler & Johnson (1939) describe the wall of the mature megaspore of Regnellidium diphyllum Lindm. as consisting of two principal layers, the endospore and the epispore ; the latter being subdivided into three sublayers. According to their interpretation the endospore lies next to the spore protoplast and is continuous under the apical papilla. The homogeneous inner layer of the epispore is sudanophilic, lies closely applied to the endospore and is not continuous under the papilla. The thickest layer of the megaspore wall is formed by the middle layer of the epispore, the prismatic layer. This is composed of hollow radiating chambers or prisms the outer ends of which are closed by a membranous cap. The prismatic layer is also sudanophilic. The outer layer of the epispore is termed the papillate layer but since microchemical tests show it to be composed of cellulose it will not be considered further. Finally, a thick layer of mucilagenous material was detected between the papillate layer and the sporangium wall, but the authors do not regard this as part of the megaspore wall. Although Chrysler & Johnson (1939) do not record the presence of a layer internal to the endospore in the megaspores of Regnellidium they do illustrate sections of spores which clearly show a membrane in this position and which they regard as a EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 241 layer of the cytoplasm. This layer is continuous over the floor of the papilla and appropriate histochemical tests would determine whether it is analogous to the intine of angiosperm pollen. The megaspore wall of Regnellidium has not so far been examined in the electron microscope, but sections of acetolyzed material examined by phase contrast micro- scopy show that the exine stratification, although somewhat similar to that of Marsilea megaspores, differs sufficiently to warrant description. Under phase contrast the inner layer of the megaspore wall (endexine or endospore of Chrysler & Johnson) appears to be homogeneous (PI. 10, figs. 9, 10). The layer immediately overlying the endexine (inner layer of the epispore of Chrysler & Johnson) is not homogeneous, but structurally resembles the " reticulated layer " of Marsilea megaspores. In Regnellidium, however, the " reticulated layer " is thicker than in Marsilea and subdivided into two very distinct zones, the outer one of which has a coarser structure than the inner (PI. 10, fig. 10). In the megaspores of Marsilea the radially-arranged hollow chambers or prisms are covered by a continuous membrane which passes without interruption across their outer surfaces. In Regnellidium the situation is somewhat different. The mem- brane terminating each chamber is not continuous and the result, as Chrysler & Johnson (1939) point out, is that the outer surface of the layer appears papillate (PI. 10, fig. 9). Enclosed at the outer end of each of the chambers is an accumulation of acetolysis-resistant material in the form of small phase-dark granules. The microspore of R. diphyllum is, according to Chrysler & Johnson, a small edition of the megaspore. Two wall layers can be detected, an endospore and an outer prismatic layer. The authors observed several cases where the prismatic layer surrounded an entire tetrad of microspores or, in some cases, possibly even two tetrads. SPORES OF LIVING AND FOSSIL HOMOSPOROUS PTERIDOPH YTES The spores of three genera of living homosporous pteridophytes (Lycopodium, Psilotum and Asplenium) have been examined to determine how the exine of these spores compare with the microspores of living and fossil heterosporous pteridophytes. In addition, one fossil dispersed spore from the Devonian [Archaeotriletes] , the parent plant of which is unknown but is assumed to have been homosporous, has been sectioned. Spores of Lycopodium The electron micrographs of acetolyzed and chlorinated Lycopodium clavatum spore walls published by Afzelius, Erdtman & Sjostrand (1954) show that the exine is divided into an outer lamellated and an inner granular layer. Afzelius (1956) shows how the fine structure of the spore exine can be destroyed by over-chlorination or treatment with chromic acid. Acetolyzed spores of Lycopodium selago L. (PL 10, fig. n ; PL n, fig. i) do not display the striking lamellation detected in the spores of L. clavatum, and the exine 242 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN appears to be more or less granular to faintly fibrous. On the proximal side of the spores a discontinuous granular zone underlies the outer wall layer and this is ultrastructurally similar to the granular layer illustrated by Afzelius, Erdtman is Sjostrand (1954). The rather irregular occurrence of the layer in L. selago, however, suggests that it is produced by the breakdown of the outer wall layer and should not be regarded as a separate stratum of the exine in these spores. The fine structure of the spore wall in L. selago differs sufficiently from that detected in L. clavatum to justify an investigation to determine the full range of variation within the genus. Spores of Psilotum An acetolyzed exine of a Psilotum nudum Griseb. spore section stained with lead hydroxide is shown on PI. n, fig. 3. It can be seen that the wall is weakly stratified. A finely granular layer surrounds the spore lumen and to the outside is a somewhat thicker fibrous stratum. The undulating outer and inner margins of the wall have borders of rather higher electron density than the rest of the exine. Spores of Asplenium The spores of the polypodiaceous fern genus Asplenium have, according to Bower (1923) and Sorsa (1964) a distinct perispore or perine. The perispore is regarded as being the last layer formed by the tapetum. In the majority of species it is said to be represented by a thin, loosely-attached, often highly ornamented membrane (see, for example, Sorsa 1964). Erdtman (1952) does not consider the perispore as part of the exine in the strict sense, but rather as an extra-exinous layer of the sporoderm. Electron micrographs of ultra- thin sections of acetolyzed spores of Asplenium adiantum-nigrum L. show that the wall is stratified into three distinct layers. The outer wall layer, or perispore, is very electron opaque. Stratification is clearly detectable within the perispore itself and is seen as a series of electron-dense mem- branes (two outer and two inner) enclosing two granular interstitial layers (PL 12, figs, i, 2). The electron micrographs show that only the two outer perispore mem- branes and the outer granular layer that they enclose contribute to the formation of the elements which form the ornamentation of the spore. The two inner perispore membranes and the inner granular layer continue uninterrupted beneath the projections, and as a result they have a hollow central cavity. Between the perispore and the inner layer of the spore wall is a space, the presence of which presumably accounts for the fact that the perispore is easily detached. The inner layer of the wall would constitute the whole of the exine sensu stricto according to Erdtman's (1952) generally accepted terminology for perinate spores. This layer in A . adiantum-nigrum is structurally homogeneous in most sections, but appears faintly granular in a few (PI. 12, figs, i, 2). The inner and outer edges of the layer are bordered by two very thin slightly more electron-opaque zones. The exine fine structure of the spores of A. adiantum-nigrum is quite unlike the EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 243 microspores of Selaginella, Isoetes and Marsilea or the spores of Lycopodium and Psilotum. By definition the perine of the Asplenium spores is a wall layer which has no homologue in these species, but this interpretation can be questioned and is discussed later. The fact that the inner wall layer or exine of the Asplenium spores is homogeneous and completely undifferentiated is not altogether a surprising discovery. Any discussion as to its possible significance would, however, be almost valueless without substantially more information on the range of ultrastructural variation in fern spores, particularly in the phylogenetically more primitive genera. The wall structure of Archaeotriletes The fossil spores provisionally referred to this rather broad genus were isolated with 40% hydrofluoric acid from a sandstone of Upper Devonian age from Scaumenac Bay, Quebec, Canada. The botanical affinity of the spores is not known, but they possess, in common with a number of other genera of the same age, an ornamentation which consists of appendages that terminate in grapnel-shaped hooks (see Naumova 1953). The exine of Archaeotriletes has not been studied as completely as some of the other spores described in this account, and the observations are included as a first assess- ment which may have to be modified later. Ultra-thin sections of Archaeotriletes show that the outer part of the exine which forms the wall ornamentation is composed of anastomosing rodlets of sporopollenin similar to those seen in the spores of Archaeopteris , the megaspores of Selaginella and Isoetes and the Carboniferous lycopod megaspore Laevigatisporites. The impression gained from some sections is that the orientation of the elements changes slightly towards the inside so that the wall appears to be weakly stratified (PL 12, fig. 4). So far a layer internal to the spongy layer has not been detected. Lastly, it can be seen that the spines which constitute the ornamentation of the spores are formed by the spongy layer and that they have a hollow central cavity (PL 12, ng- 3)- POLLEN OF FOSSIL AND LIVING GYMNOSPERMS The exine structure of Schopfipollenites Pollen belonging to the genus Schopfipollenites Potonie & Kremp 1954 was obtained from the Lawrence Shale, Lone Star Lake, Kansas. The Lawrence Shale is of Pennsylvanian (Virgillian) age. Schopf (1949) has described pollen of the Schopfipollenites type in the medullosan pteridosperm fructifications Dolerotheca and Whittleseya, and there is no doubt that Schopfipollenites is a genus based on pollen. However, because the grains have both a proximal monolete aperture (a feature characteristic of some types of spores) and two distal grooves or sulci (a single sulcus is a feature characteristic of some pollen) Schopf has adopted Renault's term " prepollen " for the fossils. In view of Afzelius's (1956) suggestion that some submicroscopic features are more 244 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN primitive than others it was considered of particular interest to examine the wall structure of Schopfipollenites by electron microscopy and to compare it with the structure detected in the microspores of the free-sporing vascular cryptogams and the pollen of living gymnospermous seed plants. The exine of Schopfipollenites is divided into two very distinct layers which can be clearly seen with the optical microscope (Schopf 1949). In ultra-thin sections examined with the electron microscope the fine structural details of the two layers is resolved. The inner exine layer (endexine) is distinctly lamellated (PI. 13, figs. 2, 3 ; PI. 14, fig. i) but in places the lamellation is scarcely detectable doubtless as a result of compression during the fossilization process. In one section (PI. 13, fig. 3) it appears as though the endexine is itself subdivided into two layers. The outer layer of the exine (ektexine) is constructed of the familiar anastomosing units of sporopollenin giving the wall a spongy appearance. Structural distortion due to compression is also recognizable in this layer. Continuity exists between the endexine and the ektexine and in some of the micrographs (PI. 13, fig. 2 ; PI. 14, fig. i) it can be seen that the innermost units of the ektexine arise directly from the outer surface of the endexine. Towards the outer edge of the pollen grain the sporopollenin units of the ektexine are progressively more coalescent and at the periphery are completely fused to form a solid margin. Under the optical microscope entire specimens of Schopfi- pollenites from the Lawrence Shale appear to be minutely punctate (PI. 13, fig. i). The electron micrographs show, however, that the solid margin is not interrupted by punctae (PI. 13, fig. 2). One possible explanation to account for this discrepancy is that the interstitial spaces between the outermost sporopollenin units of the ektexine can be mistaken for punctate surface ornamentation when the pollen is examined by transmitted light. In the possession of a lamellated endexine and a spongy ektexine the exine of the pteridosperm pollen Schopfipollenites very closely resembles that of the spores of Archaeopteris and the megaspores of Selaginella. This fossil pollen is, in fact, ultrastructurally considerably more " spore-like " than " pollen-like ". This topic will be returned to later. Pollen of Encephalartos Sections of acetolyzed pollen from the living cycad Encephalartos villosus Lem. clearly show that the exine is stratified into three distinct layers. The proximal surface of the grain is covered by an outer ektexinal layer of homo- geneous to faintly fibrous sporopollenin that extends laterally round the sides of the grain and enters the distal sulcus where it rapidly thins and finally disappears (PI. 15, figs, i, 2). The inner and outer margins of this layer are interrupted by small cavities or pits (PI. 14, fig. 2). Immediately underlying the outer ektexine is the middle wall layer or inner part of the ektexine which in thin sections appears to be prismatic. This layer is not well developed on the proximal face of the grain, but becomes thicker towards the sides and is thickest in the distal sulcus where, EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 245 with the disappearance of the outer ektexine, it is the outer layer (PI. 15, figs. I, 2). In sections that pass obliquely through the inner ektexine it can be seen that the elements forming the layer are a system of radially-aligned closed tube-like processes. The outer extremities of the elements are connected to the inner surface of the outer ektexine, and the inner extremities to the innermost exine layer. In the distal sulcus, where the outer ektexine is not formed, the elements of the middle layer are terminated by a thin, continuous membrane which passes across their outer ends. This membrane could be considered to represent the outer ektexine in this region. The endexine in E. villosus pollen is a very thin lamellated layer (PI. 14, fig. 2). Not infrequently, the endexine and the inner ektexine become detached from the outer ektexine at the sides of the pollen grain so that they curve into the pollen lumen ; the effect is to make the grain appear saccate (PI. 15, fig. r). Afzelius (1956) and Ueno (1960) have published electron micrographs showing the exine structure in the pollen of Cycas. Afzelius finds that the pollen exine of Cycas revoluta is composed of two layers, the inner layer being lamellated and the outer one being composed of branched bacula-like elements. Ueno, on the other hand, detected three layers in the pollen exine of Cycas taiwaniana which appear to correspond to the three layers in the pollen of E. villosus, although there are some small differences. The stratification in the pollen wall of Ceratozamia mexicana described by Larson (1964) closely resembles that found in Cycas. In C. mexicana also the innermost layer of the exine is lamellated. Compared with the Carboniferous pteridosperm prepollen Schopfipollenites the exine stratification revealed in the pollen of living cycads is quite complex. One notable feature shown by Ueno's (1960) illustrations of C. taiwaniana and those of E. villosus illustrated here (PI. 14, fig. 2 ; PI. 15) is that the exine of cycad pollen grains is thinner in the distal sulcus or germinal furrow than in the rest of the wall. But this is not so in Schopfipollenites. Perhaps the distal sulci in the pteridosperm prepollen are not true germinal furrows but merely a harmomegathic region and germination may have occurred through the proximal suture, as in the spores of free-sporing pteridophytes. Pollen of Taxus In glutaraldehyde-fixed pollen of Taxus baccata L., post-fixed in osmium tetroxide, almost the entire thickness of the exine is conspicuously lamellated (PI. 16, figs. 2, 3). On the outside of the lamellated layer is a zone rather than a uniform layer consisting of small, electron-dense droplets of exine material. Erdtman, Berglund & Prag- lowski (1961) call these structures orbicules, and they undoubtedly account for the pollen wall ornamentation. The orbicules vary somewhat in size and many bear small conical projections which radiate from their surfaces. The arrangement of these bodies on the surface of the lamellated layer seems to be quite haphazard. Some lie directly on the lamellated layer and in close contact with adjacent orbicules so that there is a suggestion of partial coalescence between them. Others lie com- pletely free and some distance above the lamellated layer. At high magnifications a substance with a granular fine structure can be detected between the orbicules 246 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN and the lamellated layer. This substance appears to bind the innermost orbicules to the subtending layer and in some cases to cement the orbicules together (PI. 16, fig. 2). Between the lamellated layer and the pollen cytoplasm is a fairly thick, electron- transparent intine which is structureless in material fixed in glutaraldehyde and post-fixed in osmium tetroxide. Ueno (1959) considers the small spinous droplets covering the Taxus exine as perine which is united with the outer exine layer or sexine. The sexine is the name applied by Ueno to the granular substance situated between the droplets and the lamellated layer. Certainly, the very scattered distribution and somewhat irregular appearance of the orbicules are very unusual features which perhaps have their nearest homologue in the perispore of the pteridophytes rather than in a true exine layer. Of some significance in this connexion are the small electron-dense droplets which occur in the innermost pollen sac tapetal cells of Taxus. These droplets closely resemble those attached to the pollen exine (PL 16, fig. 4). Similar droplets also occur between adjacent pollen grains and seemingly lie free in the pollen sac locule. They can be compared with the Ubisch bodies described by Rowley (1963) and the tapetal plaques of Heslop-Harrison (1962, 19636). Sporopollenin droplets of this kind are not confined to Taxus and Ueno (1959) has shown that they occur in a number of other coniferous genera. Afzelius (1956) also draws attention to their presence in Cephalotaxus. There is also some apparent structural homology between these structures in the conifers and those that form the outer layer or zone of the spore wall in the moss D-icranum scoparium Hedw. described below. Pollen of Pinus The ultrastructure of the pollen wall of Pinus has been described by Miihlethaler (1955), Ueno (1958) and by Ting & Tseng (1965). The electron micrographs of exine structure in Pinus balfouriana published by the latter authors show that the inner wall layer (endexine) in this species is granular in composition. Overlying the endexine is a thicker ektexine which they subdivide into a " foot layer " or " limen inferum ", a series of columellae with " depressions " and " funnels " and a tectum. Their pictures show that the ektexine expands to form the sacci of the pollen grain whilst the granular endexine alone remains to form the floor of the sacci. In contrast to this, Ueno (1958) has shown that in Pinus thunbergii only the tectum and the columellae contribute to the sacci, and the saccus floor is formed by the " foot layer ". This also seems to be the case in Pinus nigra (Miihlethaler 1955). Sections of Pinus sylvestris L. pollen that has been fixed in glutaraldehyde and post-fixed in osmium tetroxide show features which are structurally in accord with those published by Ueno (1958) in P. thunbergii. The inner layer of the exine (endexine) in P. sylvestris is composed of lamellated sporopollenin (PI. 17, fig. 4). An ektexine overlies the endexine and consists of a basal layer (" foot layer "), a series of columellae which are sometimes branched (termed the " tigna " by Ting & Tseng 1965) and a tectum. The tectum is folded in places and possibly inter- rupted by " depressions " and " funnels " similar to those detected in P. balfouriana EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 247 by Ting & Tseng (PL 17, fig. 3). The basal membrane (" foot layer ") of the ektexine appears to be composed of a series of distinct droplets of the exine material which lie in close contact and which, in places, are partly coalescent. However, it seems likely that the droplets are arranged along a thin and continuous membrane which they are obscuring (PI. 17, fig. 4). One notable feature is the difference in the construction of the sacci in P. sylvestris and in P. balfouriana. In the former, unlike the latter, the saccus is composed only of the columellae and the tectum of the ektexine (PI. 17, fig. 2 ; PI. 18, fig. i), the inner layer or " foot layer " has not expanded but remained to form the floor of the saccus precisely as in P. nigra (Miihlethaler 1955) and P. thunbergii (Ueno 1958). Another marked difference between the two species is in the fine structure of the endexine. In P. balfouriana it is granular, in P. sylvestris it is distinctly lamellated. Although Ueno (1958) does not comment directly about the structure of this layer in P. thunbergii, judging solely from his drawings it appears to be lamellated. The structural variation of the endexine in the pollen of Pinus could be of taxono- mic and possibly of phylogenetic importance, and a full comparative ultrastructural investigation into the wall structure of both living and fossil species would be worth- while. Lastly, and merely to record the observation, Ubisch body-like objects can also be detected adjacent to the tapetal cell walls in Pinus sylvestris pollen sacs (PI. 18, fig. 2). THE MEGASPORE MEMBRANE OF PALAEOZOIC SEEDS AND SEED-LIKE STRUCTURES It has already been pointed out that the structure of the megaspore membrane in some Carboniferous cordaite and pteridosperm ovules is very different in construc- tion from that of the ovules of living gymnosperms (Pettitt 1966). The observations described below present the evidence in support of this statement. The megaspore membrane of Carboniferous cordaitalean and pteridosperm ovules The cordaitalean ovules from which the megaspore membranes used in this study were obtained occur as compression fossils in the Lawrence Shale (Virgillian) of Lone Star Lake, Kansas. The cuticle morphology of the ovules (to be described elsewhere) is very characteristic and leaves no doubt that they are cordaitalean. The pteridosperm megaspore membranes were isolated with hydrofluoric acid from " nucule " casts of Trigonocarpus from the British Coal Measures. Ultra-thin sections of a methacrylate-embedded cordaite megaspore membrane from the Lawrence Shale are shown on PI. 19, figs. 1-3. It can be seen that the entire thickness of the membrane is composed of a three-dimensional network of sporopollenin elements giving a spongy appearance. At higher magnifications (PL 19, fig. 2) the homogeneous structure of the elements is apparent. At the 248 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN inner edge of the membrane, in contrast to the outer, the sporopollenin units are fused to form a distinct margin around the lumen (PL 19, fig. 3). The very real structural resemblance of the megaspore membrane of these fossil ovules to the wall structure detected in the megaspores of Selaginella, the spores of Archaeopteris and the Carboniferous lycopod megaspore Laevigatisporites is of considerable interest. A repetition of this structure in the megaspore membrane of Trigonocarpus (PL 18, fig. 3) the ovule of a Carboniferous pteridosperm and in Taxospermum, a petrified cordaitalean ovule (PL 18, fig. 4) adds significantly to the statement that structurally the megaspore membranes of the early seed plants more closely resemble the spores of the free-sporing heterosporous pteridophytes than the megaspore membranes of the modern gymnosperms (Pettitt 1966). The phylogenetic significance of this is discussed below. Megaspores of Lepidocarpon Megaspore tetrads belonging to the genus Cystosporites are known to occur in the Carboniferous seed-like structures Lepidocarpon, and there is a great deal of evidence to suggest that all lepidocarp seeds contained tetrads of the Cystosporites type. Tetrads of Cystosporites are characterized by having one large, sac-like, presumably fertile spore and three, smaller, presumably abortive spores attached at the apex of the large one. It is quite probable that most species included in the genus do represent lepidocarp seed megaspores. Only the wall structure of the large fertile spore is described in this account. Megaspore tetrads referable to Cystosporites giganteus (Zerndt) Schopf 1938 were isolated from a Lower Carboniferous coal from Fife in Scotland by digesting the matrix in dilute nitric acid. Chaloner (1952) has shown that megaspore tetrads of C. giganteus occur inside the lepidocarp seed Lepidocarpon waltoni. The wall of the large spore of C. giganteus exhibits a very characteristic mesh-like or fibrous texture when examined in surface view with the light microscope (see Chaloner & Pettitt 1964). In contrast to this the walls of the three abortive spores appear to be homogeneous or faintly granular when examined in this way. Examination of sections of the fertile spore by electron microscopy clearly reveals the arrangement and fine structure of the fibrils that compose the outer wall layer. It can be seen (PL 20, figs. I, 2) that the elements form a lose-textured stratum. The electron micrographs show that there are two distinct sizes of fibrils. The larger ones are themselves fibrillar in structure, whilst the smaller ones are homogeneous and more or less circular in cross section. There is evidence from electron microscopy of sections and from phase contrast studies of the surface which suggests that the two types of fibrils are interconnected or that the large fibrous type branch and give rise to the smaller variety. The inner layer of the Cystosporites exine is very thin and is composed of homo- geneous sporopollenin (PL 20, figs i, 2). In his description of Cystosporites Schopf (1938) draws particular attention to the fibrous character of the fertile spore exine. He supposes that this type of structure is correlated with the passage of food reserves from the parent plant into the mega- EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 249 spore whilst it is enclosed in the sporangium and after the exine was formed. This interpretation would obviously not be applicable to those species of the genus, for example, Cystosporites verrucosus and C. devonicus (see Chaloner & Pettitt 1964) in which the exine is not fibrous but more compact and homogeneous. The spongy texture of the megaspore wall of Laevigatisporites , a sigillarian megaspore described earlier, is not without significance in this connexion. Sigillaria is a free-sporing heterosporous plant which had, presumably, an endosporic gametophyte that developed in much the same way as the present day Selaginella or Isoetes. Whilst the possible physiological significance of a porous megaspore exine surrounding the female gametophyte of Cystosporites can be appreciated, that of an equally porous membrane surrounding the gametophytes of Selaginella, Isoetes and Laevigatisporites is not so readily apparent. The exine structure of Didymosporites The inclusion of an account of the exine of the fossil megaspore tetrads Didymo- sporites under the general heading of seeds and seed-like structures is not meant to imply that these tetrads are known to occur in seeds or even seed-like organs. However, it has been suggested by some authors (see Andrews 1961) that the Lower Carboniferous fossil Bensonites, the parent sporangium of Didymosporites (Chaloner 1958), may represent a morphological stage intermediate between heterospory and the seed habit. For this reason, as well as that of convenience, the account has been included at this point. Tetrads of Didymosporites are composed of two large, thin- walled, presumably fertile spores and two minute, presumably abortive ones that are situated at the junction of the two large spores (PI. 21, fig. 4). Chaloner (1958) has isolated Didymo- sporites tetrads from the megasporangium (Bensonites fusiformis] of the Lower Carboniferous coenopterid Stauropteris burntislandica. He also obtained tetrads of Didymosporites (D. scotti} from a Lower Carboniferous (Dinantian) coal out- cropping east of Pittenweem Harbour in Fife, Scotland and it was from this material that the megaspore tetrads described below were obtained. The exine of the large, presumably fertile spores of D. scotti, when examined as thin sections with the electron microscope, are completely homogeneous and show no detectable stratification (PL 21, fig. i). Anoptral contrast pictures of sections through the entire tetrad (PI. 21, figs. 2, 3) show that the structural homogeneity extends throughout the fertile spore. The exines of the abortive spores are also homogeneous. A thin, structureless membrane surrounds each tetrad, and this presumably represents, as Chaloner (1958) suggests, a remnant of the tapetum. IV. GENERAL DISCUSSION The foregoing descriptions of the structure of the spore and pollen grain walls as revealed by the electron microscope show that there is a considerable amount of ultrastructural variation in the exines of the various genera and the question 250 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN arises as to whether the observed structural variations have any phylogenetic significance, and to what, if anything, they can be attributed. In their work on the structure of the spores of the Musci, McClymont & Larson (1964) find that the wall stratification in all the spores that they examined was basically similar. They also show that the exine of moss spores is rather more simple in construction than that of the vascular cryptogams described above. In the mosses they detected two exine layers, the inner of which is very thin and featureless. The outer exine layer is electron-opaque, acetolysis-resistant and appears to be homogeneous. This layer, according to the authors, corresponds to that which Afzelius (1957) has called the exine in Funaria spores. In the spores of some genera (e.g. Physcosmitrium and Encalyptra) this layer does not participate in the formation of the wall ornamentation, whilst in others (e.g. Polytrichum and Astonum) it forms either the entire sculptural element or only the base. In the latter case it seems that the major part of the sculptural element is composed of a third layer which McClymont & Larson call the perine. It is also suggested that the perine forms the entire sculptural element in those spores where the exine takes no part in the formation of the wall ornamentation (e.g. Physcosmitrium and Encalyptra) . Electron micrographs of acetolyzed spores of the moss Dicranum scoparium are shown on PL 20, figs. 3, 4. In these the exine is composed of a homogeneous to faintly granular layer which bears on its outer surface a number of irregularly arranged and variously shaped droplets that represent the wall ornamentation. These elements closely resemble those which McClymont & Larson (1964) refer to as perine. Low magnification pictures of sections through an acetolyzed capsule of Dicranum (PI. 20, fig. 4) show that the droplets are not only arranged on the spore exines but also randomly distributed in the spaces between the spores and are concentrated in a narrow zone adjacent to the remnants of the capsule wall. The occurrence of the droplets adjacent to the capsule wall strongly suggests that they were in some way associated with the tapetal tissue which has, of course, been destroyed by acetolysis. If these droplets do in fact originate from the tapetal cells, then McClymont & Larson's suggestion that they are perinous (sensu Erdtman 1952) would be strictly correct. There is an unescapable resemblance in the arrange- ment and occurrence of the droplets forming the spore ornamentation in the mosses and the droplets associated with the pollen wall of Taxus baccata described earlier (p. 245) and which Ueno (1959) regards as being a perine. Before continuing there is perhaps some justification for discussing the precise interpretation of the perine. The term perine (or one of its synonyms) has been applied to a variety of exinous or extra-exinous layers of the spore or pollen wall with, seemingly, very little concern as to whether the layer in question is truly perinous. Arising from this is a more important issue ; whether all the layers to which the term has been applied are homologous. Erdtman (1952) regards the perine as an outer, extra-exinous layer of the spore wall of certain mosses and ferns which is formed by the activity of a tapetal plas- modium with " perinogeneous properties ". He points out that it is virtually EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 251 impossible to determine whether a particular layer is exinous or perinous without undertaking a study of the development of the spores. Harris (1955) also discusses the origin of the perispore or perine and suggests that if the ornamentation of spores arises as a result of the periplasmodial substance gelling on the outer surface of the spore the layer thus formed is the perispore. This definition is essentially the same as that proposed by Bower (1923) and it would seem to restrict the use of the term perine to the spores from plants with a plasmodial tapetum. Rowley's (1963) observation that Ubisch bodies (tapetal plaques, tapetal droplets or orbicules) are apparently only formed in plants with a secretion tapetal system is very relevant in this connexion. Bower (1923) considers that the most striking development of the perispore is in the Salviniaceae, and regards the massulae in the microsporangium and the glochidia and " swimming organ " of the megasporangia of Azolla as perispores with a specialized function. Eames (1936) on the other hand, restricts the term to the meshed structure formed around the megaspore by the fourth massula. Arnold (1955) has shown that this structure survives in fossils. In the mature microsporangia of Salvinia the spores are surrounded by an amor- phous mass of vacuolate, sudanophilic, acetolysis-resistant material (PI. 21, fig. 5) which is formed from the tapetal periplasmodium. Eames (1936) calls this the perispore and, in the sense that this term is defined by Erdtman (1952) and Harris (1955) his usage is correct. A very similar vacuolate mass also occurs in the mega- sporangium of Salvinia. Fitting (1900), as mentioned before, recommends that the term perispore should be applied only to wall layers that are formed from the special mother cell wall and the term epispore to those formed from the tapetum. In the spores of Asplenium adiantum-nigrum, however, the outer layer of the wall or so-called perispore (see Nilsson & Praglowski 1963) appears, ultrastructurally, to resemble a true exine layer. It is difficult to believe that the perispore in Asplenium or in Isoetes microspores for that matter, originates directly from a periplasmodium by the process of gelling suggested by Harris (1955). There is clearly a need for a full comparative developmental investigation into the formation of these wall layers. The correct interpretation can, as Erdtman has said, only be obtained from such a study. The results might well provide an ex- planation for the presence of the tapetal droplets in the pollen sacs of some conifers and in the capsules of mosses. It would surely clarify the situation with regard to the homology, if it exists, between the structurally distinct perispore in the Salvini- aceae, the Polypodiaceae and the conifers and mosses. It might even result in a much-needed, clear and authenticated terminology! Before leaving the question of tapetal derivatives, one further point is of particular relevance to the present study. It has already been shown (Pettitt 1966) that in certain gymnospermous ovules the structure which is called the megaspore membrane is, in fact, formed by the degeneration of the tapetum to form a thick, acetolysis- resistant membrane. The homologue of this tapetal membrane in the sporangia of free-sporing plants with a plasmodial tapetum would possibly be the perispore. 252 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN The tapetal membrane in the ovule would be, for example, homologous to the vacuolate sudanophilic mass of the Salvinia microsporangia and to the outer mem- brane of the spores of A . adiantum-nigrum. While there may be certain, admittedly somewhat tenuous, structural similarities between the tapetal membrane of the ovules and the perispore of Salvinia, there are less between the former and the perispore of A. adiantum-nigrum. McClymont & Larson's (1964) study of moss spores shows that the exine in these plants is less well differentiated than it is in the vascular cryptogams. At the present state of our knowledge it is probably premature to discuss whether this means that they are also more primitive in an evolutionary sense. Such a discussion would have relatively little value until we are more sure of the criteria for judging the " simple " level of organization (as opposed to " reduced " and " derived ") in spore and pollen walls. If complex structural differentiation in the exine is a measure of phylogenetic advancement then the spores of the mosses, of Psilotum, Lycopodium and the microspores of Selaginella have a primitive exine structure ; certainly more primitive than the pteridosperm prepollen Schopfipollenites, for example. Accepting Afzelius's hypothesis (1956) that pollen grains with a thin lamellated layer in the exine are phylogenetically more advanced than those with a thick lamellated layer, it is not surprising to find that the prepollen Schopfipollenites is phylogenetically less advanced than the pollen of Encephalartos. This hypothesis would, however, be difficult to apply in cases where an inner lamellated layer is not formed. In these spores the absence of such a layer could be taken to mean either that it has been completely reduced, and therefore the spores were advanced, or that it had not developed at all. It is not clear precisely how the microspores of Isoetes, with a well-developed lamellated layer as the outer wall stratum, would fit into this scheme, unless one accepts Nilsson & Praglowski's (1963) suggestion that this layer is the perine (and therefore not strictly part of the exine) . From a palaeobotanical standpoint one of the most interesting facts to emerge from the present study is that of the structural similarity between the megaspores of some of the free-sporing heterosporous plants and the megaspore membranes of the fossil gymnosperms. In the heterosporous Upper Devonian pteridophyte Archaeop- teris the exine structure of the microspore and megaspore is almost identical. The situation is made the more striking by the structural dissimilarity of the microspores and megaspores in S. pulcherrima. It would be convenient, to say the least, to consider the spores of Archaeopteris as indicating that heterospory in this plant was not far developed from a homosporous stage and that the structure of the micro- spores and megaspores had not differentiated much beyond that of some ancestral isospores. This situation can be contrasted with the structure of the microspores and megaspores in Selaginella, a plant in which heterospory has long prevailed. It would seem that there is some correlation between a spongy or fibrous exine structure and the retention of the female gametophyte within the megaspore. This type of wall texture is seen in the megaspores of Selaginella, Isoetes, in the Carboni- ferous spores Laevigatisporites and Cystosporites, and also in some Palaeozoic gymno- EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 253 sperm ovules. Didymosporites would appear to be an exception. At the risk of overstressing the significance of this it could be surmised that the continued evolution of a megaspore of the Archaeopteris type could result in further development of the spongy layer and the suppression of the inner lamellated layer. This process would result from, or accompany, the retention of the female gametophyte within the megaspore. The final structure would resemble the megaspore of Selaginella. Resulting from the retention of the megaspore within the megasporangium to form an ovule all that would be necessary to form a megaspore membrane of the type contained in the Carboniferous pteridosperms and cordaites is the complete sup- pression of the thin, inner, lamellated layer and a slight reduction in thickness of the spongy layer. It is worth repeating here that this structure is totally different from that found in modern gymnospermous ovules (Pettitt 1966). The hypothesis is, however, substantially weakened by the fact that a spongy exine is not confined to the spores of free-sporing heterosporous plants and the megaspores of fossil ovules. Larson (1964) has found that the ektexine of Linum pollen is composed of ramifying sporopollenin rodlets and this type of structure also occurs in the pollen of Juglans and Gary a (Stone et al. 1964). The structural similarity of the Archaeopteris exine and that of the Carboniferous prepollen Schopfipollenites suggests that very little, if any, fine structural alteration need accompany the change from microspore to pollen grain, if indeed such a change occurred. The comparison of the prepollen exine with that of the pollen from the living cycads and conifers shows quite clearly that the exine of the fossil is ultra- structurally considerably closer to a spore than to a pollen grain. A spore-like exine structure can also be seen in the Carboniferous pollen Florinites (PI. 19, fig. 4). This material was recovered as pollen masses from the same horizon in Kansas as the cordaitalean ovules. The genus Florinites represents the pollen of the cordaites and has been found in unquestionably cordaitalean pollen sacs by Florin (1936), Delevoryas (1953) and Fry (1956). Sections through the central body of the grain show that it is composed of an outer spongy layer and an inner, much thinner, granular layer (PI. 19, fig. 5). The inner layer shows no trace of lamellation. The structural and /or physiological significance of an inner lamellated layer in the exine is some way from being understood, but Larson's (1964) discovery that the inner wall layer of Bruchia brevifolia spores is lamellated only in the harmo- megathic region is of significance in this connexion. It would be of considerable interest to know the implications of fine structural changes in the exine, and to establish precisely what is responsible for the observed differences. Judging from the very apparent dissimilarity of the microspores and megaspores of 5. pulcherrima, for example, it might be supposed that the mechanism involved is very fundamental. It is noteworthy that Manton (1950) has shown that in forms of " Cystopteris fragilis " a difference in spore size can be correlated with a difference in chromosome number ; large spiny spores are associated with a high chromosome number (n = 126) and somewhat smaller spiny spores with a low chromosome number (n = 84). Examination with an electron microscope GEOL. 13, 4. I? 254 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN would show whether the difference here is purely one of spore size or if it extends to the wall structure. A final comment of taxonomic interest arising from this study of spore and pollen wall structure must be included. The separation of the inner exine layer or layers that was observed in the spores of Archaeopteris, Selaginella and Isoetes is clearly relevant to the correct interpreta- tion of spore and pollen grain morphology, particularly in dispersed fossil specimens. Spores in which the inner layer of the exine is separated from the outer by a cavity are described as cavate (Dettman 1963, Faegri & Iverson 1964) and at least in some cases the inner layer or membrane is termed the mesospore (see Kremp 1965 : 89). Fitting's (1900) application of this term has already been fully discussed. The amount of variation that can occur in the formation of a mesospore, even in the spores of a single sporangium, is illustrated by the microspores of Isoetes (PI. 7, figs. 1-8) and Botrychium jenmanii Underw. (PI. 6, fig. 6), and there is certainly a need for caution in assessing the taxonomic value of such structural features. V. REFERENCES ANDREWS, H. N. 1961. Studies in paleobotany. 487 pp. New York & London. ARNOLD, C. A. 1955. A Tertiary Azolla from British Columbia. Contr. Mus. Paleont. Univ. Mich., Ann Arbor, 12 : 37-45, pis. i, 2. AFZELIUS, B. M. 1956. Electron microscope investigations into exine stratification. Grana palynol., Stockholm, 1 : 22-37, 2 pls- I957- On new methods in physical cell research and their application in studies of pollen grains and spores. In ERDTMAN, G. Pollen and spore morphology, Plant taxonomy. Gymnospermae, Pteridophyta, Bryophyta. 151 pp., 265 figs. Stockholm. , ERDTMAN, G. & SJOSTRAND, F. S. 1954. On tne nne structure of the outer part of the spore wall of Lycopodium clavatum as revealed by the electron microscope. Svensk. hot. Tidskr., Stockholm, 48 : 151-161, pis. i, 2. BONNET, A. 1955. Contributions a 1'Etude des Hydropteridees I. Recherches sur Pilularia globulifera L. et P. minuta Dur. Cellule, Louvain, 57 : 131-239, pis. 1-8. BOTERBERG, A. 1956. Etude sur les Hydropteridales, IV. Genese et differentiation des porois sporales chez Marsilea diffusa Lepr. Cellule, Louvain, 58 : 81-106, pis. 1-6. BOWER, F. O. 1923. The ferns, 1, 359 pp., 309 figs. Cambridge. CAMPBELL, D. H. 1902. Studies on the gametophyte of Selaginella. Ann. Bot., London, 16 : 419-428, pi. 19. CHALONER, W. G. 1952. On Lepidocarpon waltoni, sp. n., from the Lower Carboniferous of Scotland. Ann. Mag. nat. Hist., London, 5 : 572-582, pi. 21. 1953- On the megaspores of Sigillaria. Ann. Mag. nat. Hist., London, 6 : 881-897, pi. 22. — 1958. Isolated megaspore tetrads of Stauropteris burntislandica. Ann. Bot., London, 22 : 197-204, i pi. & PETTITT, J. M. 1964. A seed megaspore from the Devonian of Canada. Palaeontology, London, 7 : 29-36, pis. 3, 4. CHRYSLER, M. A. & JOHNSON, D. S. 1939. Spore production in Regnellidium. Bull. Torrey bot. Club, New York, 66 : 263-279, pi. 5. DELEVORYAS, T. 1953. A new male cordaitean fructification from the Kansas Carboniferous. Am. J. Bot., Lancaster, Pa., 40 : 144-150, 3 pis. DETTMAN, M. E. 1963. Upper Mesozoic microfloras from South-Eastern Australia. Proc. R. Soc. Viet., Melbourne, 77 : 1-148, pis. 1-27. EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 255 A> UlMAJl G B fy::/: H D K : L FIG. i. Comparative diagrammatic drawings of exine stratification in : A, Archaeopteris microspores (left) and megaspores (right) ; B, Selaginella megaspores ; c, Selaginella microspores ; D, Isoetes microspores ; E, Laevigatisporites ; H, Lycopodium ; i, Asplenium; j, Schopfipollenites ; K, Taxus and L, a cordaite seed megaspore. The drawings are not to scale. 256 EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN EAMES, A. J. 1936. Morphology of vascular plants. Lower groups. 433 pp., 215 figs. New York & London. EHRLICH, H. G. & HALL, J. W. 1959. The ultrastructure of Eocene pollen. Grana palynol., Stockholm, 2 : 32-35, pis. i, 2. ERDTMAN, G. 1952. Pollen morphology and plant taxonomy. Angiosperms, 539 pp., 261 figs. Stockholm. — , BERGLUND, B. & PRAGLOWSKI, J. 1961. An introduction to a Scandinavian pollen flora. Grana palynol., Stockholm, 2 : 3-92, pis. 1-74. FAEGRI, K. 1956. Recent trends in palynology. Bot. Rev., 22 : 639-664. - & IVERSEN, J. 1964. Textbook of pollen analysis. (2nd ed.). 228 pp., pis. 1-9. Oxford. FELLER, M. J. 1953. fitude sur les Hydropteridales II. Sporocarpe et sporogenese chez Marsilea hirsuta R. Br. Cellule, Louvain, 55 : 307-377, pis. i-io. FITTING, H. 1900. Bau und Entwickelungsgeschichte der Makrosporen von Isoetes und Selaginella und ihre Bedentung fur die Kenntniss des Wachsthums pflanzlicher Zellmem- branen. Bot. Ztg., Berlin & Leipzig, 58 : 107-156, pis. 5, 6. FLORIN, R. 1936. On the structure of the pollen-grains in the Cordaitales. Svensk. hot. Tidskr., Stockholm, 30 : 624-651, pis. 5-7. FRY, W. L. 1956. New cordaitean cones from the Pennsylvanian of Iowa. /. Paleont., Tulsa, Okla., 30 : 35-45, pis. 6-8. HARRIS, W. F. 1955. A manual of spores of New Zealand. Pteridophyta. Bull. N.Z. Dep. Scient. ind. Res., Wellington, 116 : 1-186, pis. HESLOP-HARRISON, J. 1962. Origin of exine. Nature, Land., 195 : 1069-1071. An ultrastructural study of pollen wall ontogeny in Silene pendula. Grana palynol., Stockholm, 4 : 7-24, 5 pis. - 19636. Ultrastructural aspects of differentiation in sporogeneous tissue, ijth Symp. Soc. exp. Biol., Cambridge, 315-340, pis. 1-13. KREMP, G. O. W. 1965. Morphographic encyclopedia of palynology . 185 pp., 38 pis. Tucson. LARSON, D. A. 1964. Further electron microscope studies of exine structure and stratification. Grana palynol., Stockholm, 5 : 265-276, pis. i— 8. - & LEWIS, C. W. 1961. Fine structure of Parkinsonia aculeata pollen. I. The pollen wall. Am. J . Bot., Lancaster, Pa., 48 : 934-943, 4 pis. LYON, F. M. 1901. A study of the sporangia and gametophytes of Selaginella apus and Selaginella rupestris. Bot. Gaz., Chicago, 32 : 124-141 & 170-194, pis. 5-9. — 1905. The spore coats of Selaginella. Bot. Gaz., Chicago. 40 : 285-295, pis. 10, n. MANTON, I. 1950. Problems of cytology and evolution in the Pteridophyta. 316 pp., 279 figs. Cambridge. MARTENS, P. 1960. Sur une structure microscopique orientee dans la paroi megasporale d'une Selaginelle. Nouvelles observations sur la structure des parois megasporales de " Selaginella myosurus " (Sow.) Alston. C.r. hebd. Seanc. Acad. Sci. Paris, 250 : 1599-1602, 1774-1775, pis. i, 2. McCLYMONT, J. W. & LARSON, D. A. 1964. An electron microscope study of spore wall structure in the Musci. Am. J. Bot., Lancaster, Pa., 51 : 195-200, pis. 1-3. MEUNIER, A. 1888. LaPilulaire. Etude anatomico-genetique du sporocarpe chez la Pilularia globulifera. Cellule, Louvain, 4 : 319-400, pis. 1-6. MUHLETHALER, K. 1955- Die Strucktur einiger Pollenmembranen. Planta, Berlin, 46 : 1-13, 5 Pis. NAUMOVA, S. N. 1953. Complexes sporo-polliniques du Devonian Superieur de la Platforme russe et leur signification stratigraphique. Trudy. Inst. geol. Nauk. Mosk. 143 (Geol. 60) : 1—204. (In Russian). NILSSON, S. & PRAGLOWSKI, J. 1963. Spore and pollen key. In ERDTMAN, G., PRAGLOWSKI, J. & NILSSON, S. An introduction to a Scandinavian pollen flora II. 89 pp., 58 pis. Stock- holm, EXINE STRUCTURE IN FOSSIL AND RECENT SPORES AND POLLEN 257 PEARSE, A. G. E. 1961. Histochemistry . Theoretical and Applied, 2nd ed. 998 pp., 245 figs. London. PETTITT, J. M. 1965. Two heterosporous plants from the Upper Devonian of North America. Bull. Brit. Mus. Nat. Hist. Geol., London, 10 : 83-92, pis. i, 2. — 1966. A new interpretation of the structure of the megaspore membrane in some gymno- spermous ovules. /. Linn. Soc. (Bot.), London, 59 : 253-263, pis. 1-4. - & CHALONER, W. G. 1964. The ultrastructure of the Mesozoic pollen Classopollis. Pollen Spores, Paris, 6 : 611-620, pi. i. POTONIE, R. 1956. Synopsis der Gattungen des Sporae dispersae, I. Sporites. Beih. Geol. Jb., Hannover, 23 : 1-103, P^3- i-n. — & KREMP, G. 1954. Die Gattungen der palaozoischen Sporae dispersae und ihre Strati graphic. Geol. Jb., Hannover, 69 : 111-194, pls- 4-20. ROWLEY, J. R. 1963. Ubisch body development in Poa annua. Grana palynol., Stockholm, 4 : 25-36, i pi. SCHOPF, J. M. 1938. Spores from the Herrin (No. 6) Coal Bed of Illinois. Rep. Invest. III. St. geol. Surv., Urbana, 50 : 5-55, pis. 1-8. — 1941. Contributions to Pennsylvanian Paleobotany. Mazocarpon oedipternum, sp. nov. and sigillarian relationships. Rep. Invest. III. St. geol. Surv., Urbana, 75 : 3-40, pis. 1-6. — 1949. Pteridosperm male fructifications : American species of Dolerotheca, with notes regarding certain allied forms. Rep. Invest. III. St. geol. Surv., Urbana, 142 : 681-724, pis. 104-115. SHATTUCK, C. H. 1910. The origin of heterospory in Marsilea. Bot. Gaz., Chicago, 49 : 19-40, pis. 3-6. SORSA, P. 1964. Studies of the spore morphology of Fennoscandian fern species. Ann. Bot. Fenn., Helsinki, 1 : 179-201, pis. 1-4. STAINIER, F. 1965. Structure et infrastructure des parois sporales chez deux Selaginelles (Selaginella myosurus et S. kraussiana). Cellule, Louvain, 65 : 221-244, pis. 1—5. STONE, D. E., REICH, J. & WHITFIELD, S. 1964. Fine structure of the walls of Juglans and Carya pollen. Pollen Spores, Paris, 6 : 379-392, pis. 1-5. TING, W. S. & TSENG, C. C. 1965. Electron microscope studies on the pollen wall of Pinus balfouriana Grev. et Balf. Pollen Spores, Paris, 7 : 9—10, pis. i, 2. UENO, J. 1958. Some palynological observations of Pinaceae. /. Inst. Polytech. Osaka Cy. Univ., Osaka, 9 : 163-186, pis. 1-3. •~ 1959 • Some palynological observations of Taxaceae, Cupressaceae and Araucariaceae. /. Inst. Polytech. Osaka Cy. Univ., Osaka, 10 : 75-87, i pi. — 1960. On the fine structure of the cell walls of some gymnosperm pollen. Biol. J . Nara Women's Univ., 10 : 19-25, pis. i, 2. PLATE i Selaginella pulcherritna (Megaspores) p. 227 FIG. i. Section treated by periodic acid/Schiff procedure. Intine and thin membrane surrounding the spore protoplast coloured red. Spore protoplast is faintly pink. The exine is not coloured. X4OO. FIG. 2. Section treated by the periodic acid/Schiff procedure. The thin membrane (red) surrounding the spore protoplast is shown more clearly than in Fig. i. X425. FIG. 3. Section treated by periodic acid/Schiff procedure. The inner, fibrous zone of the mesospore is coloured a vivid red. The outer, granular zone, is not coloured. X46o. FIG. 4. Section treated by Nile blue sulphate method. Ektexine and outer, granular zone of mesospore stained blue. Inner, fibrous zone of mesospore very faintly coloured, x 157. FIG. 5. Section stained with Sudan black B. The mesospore and the ektexine are strongly sudanophilic. x 390. FIG. 6. Section stained with acetylated Sudan black B. Three sudanophilic membranes can be seen ; the ektexine (black arc at top of figure), the endexine and the one closely associated with the spore protoplast. The intine is not stained. Protoplast and inner wall layers have pulled away from the ektexine. X685. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 1 GEOL. 13, 4 18 PLATE 2 Archaeopteris cf. jacksoni Dawson p. 225 FIG. i. Outer part of megaspore wall showing associated spherical bodies, x 10,000. FIG. 2. Inner part of megaspore wall, x 30,000. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 2 GEOL. 13, 4. 185 PLATE 3 Archaeopteris cf. jacksoni Dawson p. 225 FIG. i. Section of microspore wall, x 30,000. FIG. 2. Section of microspore wall passing through a commissure. The endexine passes upwards through the opening and overlaps onto the outer surface of the spore. x 7,500. FIG. 3. Section of megaspore wall showing separation of the endexine into two layers, x 15,000. Bull.B.M. (N.H.) Geol. 13,4 PLATE 3 PLATE 4 Selaginella pulcherrima Liebm. p. 227 FIG. i. Outer part of megaspore wall. x 25,000. FIG. 2. Inner part of megaspore wall. The spore lumen is to the lower right of the micro- graph, x 11,250. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 4 PLATE 5 FIG. i. is present FIGS. 2, protoplast XI28. FIG. 4. FIG. 5. triradiate FIG. 6. Selaginella pulcherrima Liebm. p. 227 Section of an acetolyzed megaspore wall showing structure of ektexine. The endexine as a thin, phase-white layer. Anoptral contrast. x 1,000. 3. Sections of megaspores treated with mercuric bromophenol blue. The spore in Fig. 2 and the mesospore in Fig. 3 are coloured blue-green. Fig. 2, x 100. Fig. 3, Acetolyzed megaspore containing a mesospore. x 100. Tangential section through proximal region of megaspore. commissures. X 125. Stained with Bismark brown. Section of microspore wall. x 7,500. Isoetes humilior F. Mull. The endexine lines the p. 232 FIG. 7. Section of acetolyzed megaspore wall. Anoptral contrast, x 1,000. FIG. 8. Section of acetolyzed megaspore wall showing separation of inner part to form a distinct membrane. See also PL 6, fig. i. Anoptral contrast, x 1,000. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 5 PLATE 6 Isoetes humilior F. Mull. p. 232 FIG. i. Section of acetolyzed megaspore showing partial separation of inner part of wall. Anoptral contrast, x 106. Isoetes echinospora Dur. p. 234 FIG. 2. Section of acetolyzed microspore wall. x 10,000. Section stained with potassium permanganate. FIG. 3. Section of acetolyzed microspore wall showing separation of inner homogeneous layer from outer lamellated layer. x 7,000. Laevigatisporites cf. glabratus (Zerndt) p. 235 FIG. 4. Entire megaspore from Lawrence Shale, Kansas. x 20. FIG. 5. Section of wall. Phase contrast, x 640. Botrychium jenmanii Underw. p. 254 FIG. 6. Acetolyzed spores showing separation of wall layers. See also PL 7, fig. 9. x 370. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 6 PLATE 7 Isoetes echinospora Dur. p. 234 FIGS. 1-8. Acetolyzed microspores showing (except in Fig. i) separation of wall layers. Figs, i, 2, 5, 6, 8 sections under phase contrast. Figs. 3, 4, 7 entire spores. For full explanation see text. Fig. i. x 2,200; Figs. 2, 4. x 1,700; Fig. 3, x 1,500; Fig. 5. x 2,100; Fig. 6, x 1,900; Fig. 7. x 1,000; Fig. 8. x 1,850. BotrychiumjenmaniiUndervf. p. 254 FIG. 9. Section of acetolyzed spore showing separation of wall layers. See also PI. 6, fig. 6. Phase contrast, X925. Laevigatisporites cf. glabratus (Zerndt) p. 235 FIG. 10. Section of wall, x 30,000. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 7 PLATE 8 Marsilea drummondii A. Br. p. 235 FIG. i. Entire, acetolyzed megaspore. x^2. FIG. 4. Section of acetolyzed megaspore wall. Phase contrast, x 460. Marsilea quadrifolia L. p. 235 FIGS. 2, 3. Sections of acetolyzed megaspore wall. Fig. 2 is the inner part and Fig. 3 the outer. Both micrographs x 10,000. Sections stained with potassium permanganate. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 8 PLATE 9 Marsilea quadrifolia L. p. 235, 237 FIG. i. Apex of acetolyzed megaspore showing abortive spores. x 310. FIGS. 4, 5. Sections of acetolyzed microspore walls. Fig. 4, x 15,000. Fig. 5, x 12,500. Sections stained with potassium permanganate. Marsilea drummondii A. Br. p. 235, 237 FIG. 2. Section through apex of acetolyzed megaspore showing structure of abortive spores. Anoptral contrast, x 470. FIG. 3. Section of acetolyzed microspore wall, x 7,500. Bull. BM. (N.H.) Geol. 13, 4 PLATE 9 PLATE 10 Marsilea drummondii A. Br. p. 237 FIG. i. Mass of normal and abortive spores from acetolyzed microsporangium. x 103. FIG. 2. Acetolyzed microspore. X333- Marsilea quadrifolia L. p. 237 FIGS. 3-8. Sections of acetolyzed microspores showing variation in development of inner wall layer. For full explanation see text. All are anoptral contrast pictures. Fig. 3. X 520; Fig. 4. x8oo; Fig. 5. x 880; Fig. 6. X 590; Fig. 7. x 500; Fig. 8. x 600. Regnellidium diphyllum Lindm. p. 240 FIGS. 9, 10. Sections of acetolyzed megaspore wall. Phase contrast. Fig. 9. x 660; Fig. 10. x 1,300. Lycopodium selago L. p. 241 FIG. ii. Section of acetolyzed spore wall. x 5,000. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 10 PLATE ii Lycopodium selago L. p. 241 FIG. i. Section of acetoly zed spore wall, x 30,000. Psilotum nudum Griseb. p. 242 FIG. 2. Entire, acetolyzed spore. X44O. FIG. 3. Section of acetolyzed spore wall. x 10,000. (Section stained with lead hydroxide.) Bull. B.M. (N.H.) Geol. 13, 4 PLATE 11 GEOL. 13, 4. PLATE 12 Asplenium adiantum — nigrum L. p. 242 FIGS, i, 2. Sections of acetoly zed spore walls. Fig. i, x 3,000. Fig. 2, x 12,500. FIG. 5. Entire, acetolyzed spore. x 770. Archaeotriletes sp . p. 243 FIGS. 3, 4. Sections of wall. Fig. 3, x 10,000. Fig. 4, x6,25o. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 12 PLATE 13 Schopfipollenites sp. p. 243 FIG. i. Entire grain from Lawrence Shale, Kansas. x 200. FIGS. 2, 3. Sections of wall. The outer edge of the grain is at the top left of fig. 2, the lumen is at the bottom right. Fig. 2, x 22, 500. Fig. 3, x 4,000. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 13 PLATE 14 Schopfipollenites sp. p. 243 FIG. i. Sections of the wall. x 10,000. Encephalartos villosus Lem. p. 244 FIG. 2. Section of acetolyzed pollen wall. x 10,000. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 14 PLATE 15 Encephalartos villosus Lem. p. 244 FIGS. 1,2. Sections of acetolyzed pollen walls. Fig. i is more or less transverse and Fig. 2 more or less longitudinal. Both micrographs x c. 2,500. FIG. 3. Entire, acetolyzed pollen grain. x 1,600. Bull. B.M. (N.H.) Geol. 13, 4 • • PLATE 15 GEOL. 13, 4. PLATE 16 Taxus baccata L. p. 245 FIGS. 1-3. Sections of pollen grains fixed in glutaraldehyde and post-fixed in osmium tetroxide. Fig. i, xc. 6,000. Fig. 2, x 20,000. Fig. 3, x 15,000. FIG. 4. Micrograph showing small, electron-dense droplets in tapetal cell cytoplasm. (Fixed in glutaraldehyde, post-fixed in osmium tetroxide.) X 15,000. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 16 PLATE 17 Pinus sylvestris L. p. 246 FIG. i. Entire, acetolyzed pollen grain, x 280. FIGS. 2-4. Sections of pollen grains fixed in glutaraldehyde and post-fixed in osmium tetroxide. Fig. 4 shows the " foot layer " forming the floor of the saccus and the underlying endexine. Fig. 2, x c. 2,000. Fig. 3, x 7,500. Fig. 4, x 30,000. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 17 PLATE 18 Pinus sylvestris L. p. 246 FIG. i. Section of pollen grain showing origin of a saccus. (Fixed in glutaraldehyde and post-fixed in osmium tetroxide.) x 10,000. FIG. 2. Micrograph showing the Ubisch body-like objects adjacent to tapetal cell wall. Part of saccus of pollen grain forms an arc in lower part of figure. (Fixed in glutaraldehyde and post-fixed in osmium tetroxide.) x 7,500. Trigonocarpus sp. p. 247 FIG. 3. Section of megaspore membrane. Anoptral contrast, X57O. Taxospermum undulatum Neely p. 247 FIG. 4. Surface view of megaspore membrane. Phase contrast, x c. 600. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 18 PLATE 19 FIGS. 1-3. Sections of megaspore membrane from cordaite ovule. Fig. 3 shows the inner edge of the membrane. Fig. i, x c. 2,000. Fig. 2, x 30,000. Fig. 3, x 6,250. Florinites sp. p. 253 FIG. 4. A mass of pollen grains from Lawrence Shale, Kansas, x 500. FIG. 5. Section of pollen wall. The inner, granular layer is towards the bottom of the micrograph, x 10,000. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 19 PLATE 20 Cystosporites giganteus (Zerndt) p. 248 FIGS, i, 2. Sections of wall of large, presumably fertile megaspore. Fig. i, x c. 1,500. Fig. 2m x c.i, 800. Dicranum scoparium Hedw. p. 250 FIG. 3. Section of acetolyzed spore, x 10,000. FIG. 4. Section of part of acetolyzed capsule showing electron-dense droplets concentrated in zone adjacent to remnants of capsule wall and randomly distributed between spores, x c. 2,000. Section stained with potassium permanganate. Bull. B.M. (N.H.) Geol. 13, 4 PLATE 20 PLATE 21 Didymosporites scotti Chaloner p. 249 FIGS. 1-3. Sections of wall of large, presumably fertile megaspores. Fig. i is an electron micrograph, Figs. 2, 3 are anoptral contrast pictures. Fig. i, x 11,250. Figs. 2, 3, x 1,600. FIG. 4. Complete tetrad from the Lower Carboniferous of Fife, Scotland. Abortive spores can be seen on left of figure between two large, presumably fertile spores, x 120. Salvinia auriculata Aubl. p. 251 FIG. 5. Mass of resistant material containing microspores released from an acetolyzed microsporangium. x 254. Bull. EM. (N.H.) Geol. 13, 4 PLATE 21 ""W^plWP^ PRINTED IN GREAT BRITAIN BY ADLARD & SON LIMITED BARTHOLOMEW PRESS, DORKING NEW GYMNOSPERMS FROM THE TICO FLORA, SANTA CRUZ PROVINCE, ARGENTINA S. ARCHANGELSKY BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 13 No. 5 LONDON: 1966 NEW GYMNOSPERMS FROM THE TICO FLORA, SANTA CRUZ PROVINCE, ARGENTINA BY SERGIO ARCHANGELSKY -X Museo de Cicncias Naturales, La Plata Research worker, National Research Council, Argentina Pp. 259-295 ; 8 Plates ; 39 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 13 No. 5 LONDON: 1966 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 ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 13, No. 5 of the Geological (Palaeontological) series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. Trustees of the British Museum (Natural History) 1966 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 17 November, 1966 Price £i i2s. NEW GYMNO SPERMS FROM THE TICO FLORA, SANTA CRUZ PROVINCE, ARGENTINA BY By SERGIO ARCHANGELSKY SYNOPSIS The Lower Cretaceous plants described here were collected between 1958 and 1963 from dif- ferent localities on the outcrop of the Baquero Formation. A new locality, Bajo Grande, yielded good mummified plant fragments. Two new genera, Trisacocladus and Apterocladus are des- cribed and referred to the family Podocarpaceae. Trisacocladus tigrensis gen. et sp. nov. from Bajo Tigre is based on leafy shoots with male cones attached, as well as on female structures probably also connected organically to leafy branches. The male cones yielded 3-saccate pollen grains of the type usually referred to Trisaccites. Apterocladus lanceolatus gen. et sp. nov., from Bajo Grande, is also based on leafy shoots with attached male cones possessing pollen grains with three rudimentary air bladders. The following new species of fern-like cutinized fronds are described : Pachypteris elegans, Almargemia incrassata and Ticoa lamellata ; together with the conifers Podocarpus dubius, Tomaxellia biforme, and two new species of Araucarites (A. minimus and A. baqueroensis) both of which are based on ovuliferous scales. The original diagnosis of Tomaxellia degiustoi is revised in the light of new and better preserved material from the type localty (Tico Amphitheatre). The description of these taxa adds con- siderably to the Lower Cretaceous flora of Patagonia which presently contains about 70 species, many with cuticular structure preserved. I. INTRODUCTION THE fossil plants described in the present paper were collected in the Tico Amphi- theatre, Bajo Tigre and Bajo Grande of Santa Cruz Province (Southern Argentina), during the years 1958-63. References to the Tico and Bajo Tigre localities have already been made (Archangelsky 1963, 1965). The new locality, Bajo Grande, has been mentioned but without descriptions of its fossil plants. The Baquero Formation in Bajo Grande constitutes most of the " cliffs " (barranca in Spanish) which surround a large geographic depression (bajo in Spanish). In Bajo Grande, as in most localities, this formation is composed of nearly horizontal strata which cover sediments referred to the Matilde Formation of Middle to Upper Jurassic age. Bajo Grande is the third locality which has yielded mummified plant fragments ; the sediments being more indurated yielded a large variety of species. In Tico and Bajo Tigre, the plants occur in lenticular beds at different levels in the sections. Each bed usually contains a few species only, but these are represented by many specimens, some of which are dominant in the association. The plants in Bajo Grande are usually more fragmentary and may have drifted for a longer distance before deposition. The only bed so far discovered in Bajo Grande has been named the Araucarites bed, because ovuliferous scales of this genus are very common, although other taxa are well represented. GEOL. 13, 5. 2O 262 NEW GYMNOSPERMS FROM TICO, ARGENTINA The Baquero Formation in Bajo Grande and elsewhere is divided into an upper and a lower member. The lower member, which includes the mummified plant material, has sediments of a somewhat darker colour and coarser grain (conglomerates, sand- stones, etc. of grey, violet and similar colours) and is absent in some areas. The upper member is always white to yellow in colour and consists usually of volcanic ash (tuff). It is geographically widespread and may contain plant impressions or be totally devoid of fossils. The total thickness of the Baquero Formation in all the areas so far surveyed, seldom exceeds 100 metres. 70 68 , LOS MANANTIALES BAOO GRANDE • TAD"E E H|DA * • C° ALT . EL TRANQUIIJO * ROCA BLANCA TICO . . BA:O . PUNTA DEL BARCO Map showing the main localities mentioned in the paper. In addition, the following three localities yielding rich floras are included : La Golondrina (Permian), El Tranquilo (Triassic) and Roca Blanca (Liassic). NEW GYMNOSPERMS FROM TIC6, ARGENTINA 263 On the evidence of the rich flora a Lower Cretaceous age is postulated for the Baquero Formation, and all references to an Upper Cretaceous age should be abandoned. Consequently all correlations with the Upper Cretaceous Chubut Series (especially its Yellow Tuffs Formation or " Tobas Amarillas ") are erroneous. All previously recorded localities and those referred to in the present paper are represented in the map on p. 262. The fossil plants described are all Gymnosperms. In 1965 references were made to a conifer associated with Ginkgoites tigrensis, having male and female cones in organic connection. This material, from Bajo Tigre, is here fully described and referred to the family Podocarpaceae. Good specimens of Tomaxellia degiustoi from the type locality in Tic 6 are also described and add to our knowledge of its gross morphology. The rest of the material described comes from the new Bajo Grande locality. Three species of cutinized fern-like fronds are referred to the genera Pachypteris, Almar- gemia and Ticoa. Two species of ovuliferous scales are included in the genus Arau- carites (some material from Bajo Tigre and Tic 6 is also included in one of the species). Leafy branches with male cones in organic connection are referred to a new genus of the Podocarpaceae, Apterocladus, while sterile leafy branches are referred to a new species of Podocarpus. Finally a new species of Tomaxellia is described. The re- maining material from Bajo Grande, which will be described in the near future, includes conifers of Brachyphyllum type, isolated male and female cones, Bennettitales and a few fern fragments. The usual maceration technique (nitric acid followed by alkali) has been followed. Some delicate cuticles were treated only with very dilute alkali. A few transfers were also made. The material studied belongs to the following collections : La Plata Natural History Museum (LP), Argentina ; Lillo Institute of Tucuman (LIL), Argentina and the British Museum (Natural History) (BMNH, V.) Great Britain. II. SYSTEMATIC DESCRIPTIONS GYMNOSPERMAE Abundant fern-like fronds possessing a thick cuticle are present in the Tic 6 flora. New generic names have been established for these based on cuticular characters (Archangelsky 1963) ; they were included in the Gymnosperms. The new material described here confirms the importance of these plants which occur in most of the localities where the Tico flora has been found. In the absence of fructifications the natural relationship of these leaves is uncertain, but their cuticular structure strongly suggests their inclusion in the Pteridospermae and Cycadales. Three new species are described in the present work, Ticoa lamellata which resembles some living Cycads in epidermal characters, Almargemia incrassata also placed in or near the Cycads, and Pachypteris elegans which has characters in common with some Recent members of the Cycadales, although it also has similarities with the Pteridosperms. These leaves being abundant fossils must belong to species important in the Lower Cretaceous vegetation of Patagonia, 264 NEW GYMNOSPERMS FROM TIC6, ARGENTINA Genus PACHYPTERIS Brongniart 1828 Pachypteris elegans sp. n. (PI. i, fig. i ; PL 2, figs. 10-12 ; Text-figs. 1-5) DIAGNOSIS. Leaf bipinnate, oblong, 5 cm. long x 2 cm. wide (base missing). Leaf rachis strong, 2 mm. wide. Pinnae opposite to subopposite, separated by 3-4 mm., lanceolate, up to 13 mm. long x 5 mm. wide, forming an angle of 45-60° with main rachis. Pinna rachis up to 1-5 mm. wide. Pinnules subopposite to alternate, arising at about 45° to pinna rachis, oblong to lanceolate with rounded apex and margins entire, up to 1-5 mm. long x 0-75 mm. wide ; lower margin decurrent and connecting with upper margin of pinnule below. Lower catadromic pinnule often once lobed and seated on the axil of pinna and main rachises. Mid- vein present up to near apex of pinnule, decurrent at base. Few simple lateral veins given off at acute angles. Substance of lamina thick. Upper and lower cuticles up to 5 fi thick. Upper cuticle without stomata. Cells isodiametric, 25-50 JLL, becoming rectangular near margins and on rachis. Anticlinal walls thick, unpitted, up to 5 JLI, deeply cutinized. Cell surface finely granular. Lower cuticle with stomata ; shape and size of epidermal cells as for upper cuticle, except on veins which are marked by bands of rectangular cells in more or less longi- tudinal files. Trichome bases present on veins and rachises, sometimes between veins. Unicellular hairs have been observed. Stomata haplocheilic, densely disposed between veins on lower cuticle, typically monocyclic, sometimes imperfectly dicyclic, round or oval, having no definite orienta- tion ; subsidiary cells often shared by two stomata. Guard cells little or not at all sunken. Subsidiary cells not much specialized, 6-7, not differentiated into polar and lateral, with cuticle considerably thickened on distal side (both periclinal and anticlinal walls). Guard cells cutinized, typically 40-50 fi long with dorsal and ventral walls sometimes thickened. A marked ring-shaped thickening surrounding aperture. HOLOTYPE. LP 6233. MATERIAL. In addition to the holotype, LP 6234-6249, BMNH, ¥.52266-75, LIL 2746-2752. Slides LP 167-173 ; BMNH, ¥.52276-78. HORIZON AND LOCALITY. Lower Cretaceous, Baquero Formation, lower member, Araucarites bed ; Estancia Bajo Grande, Santa Cruz Province, Argentina. DESCRIPTION. Pachypteris elegans is a common element in the Araucarites bed of Bajo Grande. Its cuticle when macerated is yellow to pale brown in colour. The lower epidermis is readily distinguished by the presence of stomata which are completely absent on the upper cuticle. The guard cells are feebly cutinized and they present three pairs of rim-like thickened zones which surround the aperture. The inner, near the aperture (ventral) is the less marked. The outer thickening (dorsal), near the contact with subsidiary cell is well marked. Finally a median thickening between the inner and outer is sometimes developed and may be fused with the inner one. When present, it may be finely granular. NEW GYMNOSPERMS FROM TICO, ARGENTINA 265 6 8 FIGS. 1-5. Pachypteris elegans sp. n. Figs. 1-3, Stomata showing strong cutin thickenings (black) and less marked ridges (dotted) on guard cells and on subsidiary cells. Slide LP 169, X37O. Fig. 4, Hair base. Slide LP 169, X37O. Fig. 5, Section of stomata, FIGS. 6-8. Almargemia incrassata sp. n. Fig. 6. Outline of holotype. LP 6255, xi. Fig. 7. Portion of upper epidermis showing two types of cells. Slide LP 184, xi8o. Fig. 8. Portion of lower epidermis showing stomata and two types of epidermal cells, Slide LP 184, xi 80, 266 NEW GYMNOSPERMS FROM TIC6, ARGENTINA The stomatal aperture is slightly sunken, the poles of the guard cells being somewhat raised. Stomata with only two large subsidiary cells have been seen. DISCUSSION. Harris (1964) emended the original diagnosis of Pachypteris (and included the genus Pachydermophyllum Thomas & Bose in it). This new diagnosis is based on general morphology of the frond as well as on its cuticular structure. The present species is included in Pachypteris because there is agreement in the shape of the fronds and their pinnules as well as in venation. There are, however, some differences in the cuticle : (a) the stomata are restricted to the lower epidermis in P. elegans, while those of the European species are amphistomatic (although very few stomata occur on the upper epidermis) ; (b) the guard cells of stomata are not much sunken in P. elegans and have thickened dorsal and ventral walls, plus the ring- shaped structure round the aperture. These differences characterize the new species, the third one with cuticle preserved (the other two being P. lanceolata and P. papillosa] . Among similar fern-like fronds from the Tic 6 flora, Mesosingeria although having comparable pinnules, differs in the lack of a definite mid-vein and in its cuticular structure, especially the stomata which are deeply sunken with a protective tube of cutine projecting outwards. Ticoa differs mainly in its stomatal apparatus (usually with a large epistomatal chamber) and in the shape of its epidermal cells which are normally elongated and not isodiametric. Moreover, the basal catadromic pinnules are never decurrent on the pinna rachis. The stomatal apparatus of Mesodescolea has in common the thickening of the guard cells on the dorsal walls as well as the median ring-like thickening, which is more pronounced. Ventral walls of subsidiary cells of this genus may also be thickened. In its gross morphology, the leaf of P. elegans is three times pinnate and has a fern-like appearance, while Mesodescolea is twice pinnate and has a wide pinnular lamina. There are some differences in cuticular structure. Ruflorinia also has guard cells with a median thickening round the aperture, but the shape of the guard cells is oval, corresponding to the shape of the stomatal apparatus. The lips of the guard cells are also different (strongly thickened) while no dorsal thickenings occur on these cells. The distribution of stomata and the number of subsidiary cells is markedly different. Some species of the form genus Scleropteris (e.g. 5. pomelii Sap.) may be compared with P. elegans, but as there is no information about the cuticle further comparisons are not possible. Scleropteris is a highly unnatural assemblage of species, probably uniting taxa of different natural orders (Archangelsky 19630;). The peculiar thickenings on the guard cells of P. elegans are comparable with those found in the living Cycadales Stangeria paradoxa and Bowenia spectabilis. The stomata of both living genera have, like P. elegans, little or not at all sunken guard cells. The ventral walls of the guard cells are cutinized in all three taxa. Stangeria and P. elegans have dorsal walls of subsidiary cells with cutinized ridges (sometimes in the fossil species there are two parallel ridges of cuticle, one on the guard cell and the other on the limit of guard and subsidiary cell) . The median ring-like thickenings on the guard cells are absent in Stangeria but may be seen in Bowenia. In all three NEW GYMNOSPERMS FROM TIC6, ARGENTINA 267 taxa the stomata are monocyclic to imperfectly dicyclic. The external morphology of the leaves in Stangeria and Bowenia differs considerably from the fronds of Pachypteris elegans. This Patagonian fossil may well be one of the latest representatives in Southern floras of that imperfectly known group of " Mesozoic Pteridosperms ", which on the one hand has some similarities with the true Cycadales while on the other has charac- ters of its own. The fructification Pteroma (Harris 1964) which is supposed to be the microsporophyll of Pachypteris papillosa, may throw some light on the true relationships of this group. Genus ALMARGEMIA Florin 1933 Almargemia incrassata sp. n. (PI. i, figs. 3, 4 ; PI. 3, figs. 13, 14 ; Text-figs. 6-10, 13) DIAGNOSIS. Leaf (fragmentary) simply pinnate. Rachis delicate, 1-5 mm. wide. Pinnae subopposite, inserted at angle of about 45° (more acute towards apex), with decurrent catadromic side and constricted anadromic side, oblong, with broadly dentate apex, up to I cm. long x 0-4 cm. wide at the middle ; lateral margins dentate towards apical sector. Uniform veins, 3-4 per pinna, parallel to margins, converging at base. Fine lines between veins clearly differentiated. Both cuticles 2 jn thick, with markedly elongated and similar epidermal cells, about 20 fi wide. They are of two kinds : (i) numerous, with thin anticlinal walls (i /*) and (2) less numerous, with thick anticlinal walls (7 /*), and less elongated. The thick-walled cells form longitudinal rows one, or occasionally more than one, cell broad. Cells on rachises markedly elongated with the same two types of thin and thick-walled cells and occasional papillae on the thick-walled cells. Stomata present on lower cuticle only, longitudinally orientated, sometimes obliquely, placed in rows of thick-walled cells, typically monocyclic to imperfectly dicyclic. Subsidiary cells typically 4-6 in number. Polar subsidiary cells occasion- ally differentiated ; longer than lateral subsidiary cells which sometimes slightly overhang guard cells. All subsidiary cells usually thick-walled (there are a few exceptions). Stomata sometimes with subsidiary cells in contact. Guard cells with poles on epidermal surface and aperture slightly sunken, 50-55 fi long. Poles, ventral and dorsal walls strongly cutinized. Longitudinal ridge of cutin occasionally present near dorsal wall of guard cells. HOLOTYPE. LP 6255. MATERIAL. In addition to the holotype, LP 6254 ; BMNH, V . 52264. Slides. LPi84; BMNH, ¥.52265. HORIZON AND LOCALITY. Lower Cretaceous, Baquero Formation, lower member, Araucarites bed ; Estancia Bajo Tigre, Santa Cruz Province, Argentina. DESCRIPTION. Almargemia incrassata sp. n. is represented by only two fragmen- tary specimens. LP 6254 i§ a lea-f fragment about 3 cm. long. It shows a few pinnae on both sides of a slender rachis. The pinnae are poorly preserved but show the 268 NEW GYMNOSPERMS FROM TICO, ARGENTINA FIGS. 9, 10, 13. Almargemiaincrassatasp.n. Figs. 9, 10. Stomata. Slide LP 184, X5oo. Fig. 13. Section of stoma, xsoo. FIGS, n, 12, 14, 15. Ticoa lamellata sp. n. Fig. n. Stoma showing sunken guard cells, large pit and remnants of lamellae. Slide BMNH. no. ¥.52263, xsoo. Fig. 12. Guard cells of a stoma showing remnants of lamellae. Slide LP 174, X5oo. Fig. 14. A trichome. Slide BMNH. no. ¥.52263, xsoo. Fig. 15. Section of stoma, X5oo. NEW GYMNOSPERMS FROM TICO, ARGENTINA 269 main morphological characters. The cuticle is well preserved and is resistant to maceration, after which it remains as an almost colourless film. The other leaf fragment (LP 6255) bears three complete pinnae, clearly showing the shape with its characteristic dentate margins. The fine lines between veins mentioned in the diagnosis clearly correspond to rows of thick- walled epidermal cells. This character of thick and thin walled cells on the same epidermis is known to occur among Recent Cycadales and has been fully discussed by Pant & Nautiyal (1963). The stomata are typically exposed with their poles of guard cells curved towards the surface, while the aperture is slightly sunken. This character, which is also common to both living and fossil Cycadales, is also found in the genus Ticoa although not so sharply marked as in Almargemia. The ventral and dorsal thickenings of the guard cells are conspicuous ; the contact area between guard and subsidiary cells (dorsal wall of guard cells) is also thickened, usually a little less than the other thickenings mentioned. In a few cases, guard cells with strong dorsal ridges parallel to the dorsal walls of the guard cells were observed. Also the subsidiary cells may slightly over- hang the guard cells. All these characters related to the thickenings of the guard cells are found in most of the living Cycadales. Thus Almargemia incrassata is another element of the Tico flora which may be placed in or near the Cycadales, in its strict sense, and agrees with Florin's (1933) observations on the type species of the genus, A. dentata (Heer) Florin. DISCUSSION. Among fossil leaves with a similar pinnate plan, A. incrassata is most similar to the type species of the genus, A. dentata (Florin 1933 : 100). Leaves of the European species are pinnate with dentate margins, and are borne on slender rachises with a similar plan of venation. Differences in morphology are of specific significance : the size of the pinnules and the number of veins is greater in A . dentata, while in A. incrassata the margins of pinnae are more dentate. The cuticles agree in having epidermal cells of two sorts, but in A. incrassata thick walled cells form definite longitudinal rows while in A. dentata they are usually isolated or occur in small groups. Stomata of both species are typically surrounded by thick-walled cells with 4-6 subsidiary cells. The guard cells are alike in both species. Teixeira (1948) figures A. dentata from Almargem (pi. 18, figs. 9-11) and from Belas (pi. 21, figs. 4-7), but adds no diagnostic characters to the original description. In these specimens, the shape of the pinnules is clearly different from A. incrassata. A. dentata is recorded from the Lower Cretaceous (Aptian) of Portugal. Thus the geographic distribution of the genus and probably its geological range are extended. Genus TICOA Archangelsky 1963 Ticoa lamellata sp. n. (PI. i, figs. 2, 8, 9 ; PI. 3, figs. 15, 16 ; Text-figs, n, 12, 14, 15) DIAGNOSIS. Leaf at least bipinnate. Segments incomplete, at least 3 cm. wide with rachis 2 mm. wide. Pinnae lanceolate up to 1-4 cm. long x 0-5 cm. wide, ending with acute pinnule, inserted at an angle of about 75°, alternate, slightly 270 NEW GYMNOSPERMS FROM TIC6, ARGENTINA overlapping adjacent pinnae. Pinnules alternate, inserted at an angle of about 70°, up to 4 mm. long X 1-5 mm. wide, slightly decurrent, lanceolate, constricted at base. Those towards proximal part of pinnae (near rachis) more oblong, with rounded apex and margins entire (not lobed). Veins not seen. Both cuticles of similar thickness, 3 /*. Upper cuticle having no stomata but numerous trichomes. Cells markedly elongated, parallel to margins, with straight walls. Cells of lower cuticle elongated near margins and on veins, somewhat iso- diametric or wedge-shaped near stomata and trichomes, slightly elongated on rest of lamina. Cell surface flat ; anticlinal walls straight, sometimes pitted. Cells on rachises markedly elongated ; trichomes and elongated stomata present. Anticlinal walls thicker than those of lamina. Stomata usually dicyclic, avoiding veins and margins, longitudinally orientated, parallel to veins and margins, 25-40 per sq. mm. Guard cells strongly sunken in round or oval pit, formed by subsidiary and encircling cells. Encircling cells typically 6-8, forming sides of pit, about 40-50 ju, wide. Mouth of pit typically 40-50 ju, wide, situated at the same level as epidermis or slightly raised. Subsidiary cells small, at bottom of pit ; guard cells 50-60 /* long, well cutinized, with cutin thickenings on dorsal and ventral walls. Poles of guard cells raised and strongly cutinized. Transverse striations occur on dorsal walls of guard cells, extending on both sides. They may belong to the remnants of lamellae. Trichomes single, sometimes in pairs, composed of an isodiametric cell, 25-35 /* in diameter, with thick anticlinal walls, about 7-9 /£. Free part composed of single cutinized hair with thick lateral walls. HOLOTYPE. LP 6250. MATERIAL. In addition to the holotype, LP 62456; BMNH, ¥.52262. Slides LP 174; BMNH, ¥.52263. HORIZON AND LOCALITY. Lower Cretaceous, Baquero Formation, lower member, Araucarites bed ; Estancia Bajo Grande, Santa Cruz Province, Argentina. DESCRIPTION. The material consists of fragmentary specimens preserved as black compressions on a pale grey matrix. When macerated the cuticle appears as a pale yellow film and is quite resistant to chemical treatment. The impressions of the pinnules show no traces of veins. However, on the lower cuticle a median longi- tudinal strip devoid of stomata may well represent the mid- vein. Lateral veinlets were not observed. DISCUSSION. This species is closely comparable with Ticoa harrisii in the size of pinnules. However T. lamellata has pinnae and pinnules inserted at a wider angle (70-75° against 45° in T. harrisii) . Pinnules of the new species are usually constricted at the base ; no lobation of basal pinnules was seen. The cuticle is of the same thickness on lower and upper epidermis while in T. harrisii the upper cuticle is thicker than the lower. The stomatal number is slightly higher in T. lamellata while the number of encircling cells is lower (6-8 against 8-10 in T. harrisii). Hair bases in T. lamellata do not show inwardly cutinized extensions as seen in T. harrisii and T. magnipinnulata. Finally, the guard cells in T, lamellata have their poles more NEW GYMNOSPERMS FROM TICO, ARGENTINA 271 strongly cutinized ; also, the probable remnants of lamellae while constant in the new species, were not observed in the others. T. magnipinnulata on the other hand, differs in the size and shape of pinnules as well as in stomatal structure (it has 3 cycles of cells forming the stomatal pit). Ticoa is another leaf genus whose leaves may be compared with those of other genera so far as their shape is concerned. All three species are alike in cuticular and especially in stomatal structure. The shape and size of leaves, however, are more varied, T. harrisii and T. lamellata being closer to each other than to T. magni- pinnulata. The two former species may be compared with Pachypteris only in the shape of pinnules, although the lamina of basal pinnules does not extend to the seg- ment rachis. The cuticles of the two genera differ in the shape of their epidermal cells and in stomatal structure. Trichome bases with cutinized anticlinal walls are constantly present in Ticoa but absent in other similar leaf genera. CONIFERALES The six species of conifers so far known in the Tic 6 flora have been referred to the genera Brachyphyllum, Athrotaxis and Tomaxellia. Most of them are numerically important in the different beds of the Tico Amphitheatre. Other conifers were subsequently found in Bajo Tigre and Bajo Grande, and some of them will be des- cribed here. Isolated ovuliferous scales are referred to the Araucariaceae, confirming the presence of this family in the Tico flora. The Podocarpaceae are represented by several species, some with male and female structures. The suspected presence of this family (Archangelsky 1963) is therefore also confirmed. A new species of Tomaxellia is also described. The Podocarpaceae and the Araucariaceae appear to be the best represented families of conifers in the Lower Cretaceous of Patagonia. They are represented by numerous species each with a large number of individuals. Many other conifers have been collected from different localities of the Tico flora (some of them probably belonging to these families) . The pollen content of some beds shows a high number of bisaccate pollen grains (as well as trisaccate) which may well have been derived from species of the Podocarpaceae. Family ARAUCARIACEAE Genus ARAUCARITES Presl 1838 The generic name Araucarites is used for cones, isolated ovuliferous scales and sterile twigs having " araucarian affinity " (Seward 1919 : 256). While it is possible to relate female structures with some security to the living Araucariaceae, the assignation of sterile twigs is less secure even when the cuticular structure is preserved. The two species described here are isolated ovuliferous scales showing a single median embedded ovule and a ligule supported by wide lignified bracts. 272 NEW GYMNOSPERMS FROM TICO, ARGENTINA Araucarites baqueroensis sp. n. (PI. i, fig. 5 ; Text-fig. 17) 1951 Araucaria (section Colymbea), Feruglio : 65. DIAGNOSIS. Cuneiform ovuliferous scale with lateral wings and acuminate apex about 3 mm. long. Width of scale 2 cm., length 2-3 cm. A single oval seed im- mersed in the scale is typically 1-1-5 cm. long X 4-5 mm. wide, broadest near distal end. Delicate wings expanded laterally up to 0-8 cm. from edge of seed. Ligule short, not exceeding external margin of bract, maximum width about i cm. HOLOTYPE. LP 57666. (Cerro Testigo bed.) MATERIAL. In addition to the holotype, LP 57676, 57686, 5770 (Cerro Testigo bed) ; LP 6345-46; BMNH, ¥.52254 (Araucarites bed) ; LP 52776 (Cladophlebis tripinnata bed) ; LP 53696 (Ticoa harrisii bed). HORIZON AND LOCALITY. Lower Cretaceous, Baquero Formation, lower member. Estancia Bajo Tigre (Cerro Testigo bed) ; Estancia Bajo Grande (Araucarites bed) and Tico Amphitheatre (Cladophlebis tripinnata and Ticoa harrisii beds), Santa Cruz Province, Argentine. DESCRIPTION. Araucarites baqueroensis is based on numerous ovuliferous scales from different localities, those from the Cerro Testigo bed in Bajo Tigre yielding the most abundant and best preserved. They are usually fragmentary and are associated with abundant broad twigs referred to Brachyphyllum. Only impressions were found in this bed which consists of a fine-grained, pale brown sediment, similar to the Tico or other Bajo Tigre horizons bearing mummified plant fragments. However, in this bed oxidation has destroyed all cuticles. The scales are large, much larger than those of the other species, A . minimus from Bajo Grande. In A. baqueroensis the ligule is well defined, being separated laterally from the wing of the bract by a visible ridge (which is a mould of the cavity left in the original compression, between the tissue of the bract and the ligule). Near the apex, the ligule is broader but does not exceed the length of the bract. The field between the external distal margin of the bract and the seed is large, more than 5 mm. long. Most specimens are broken ; seeds or seed ligules are commonly found isolated from the bract, but a few specimens show clearly the seed ligule and their bracts which have wing-like lateral expansions. The acuminate apex is usually broken but can be seen in some specimens projecting for a short distance. Two fragmentary remains of what I believe to be the same species were found in Bajo Grande. One of them (LP 6346) shows a distinct acuminate apex of the bract projecting for 3 mm., distally broken. The specimen from the Cladophlebis tripinnata bed in Tico is also frag- mentary but agrees with the typical material. The one from the Ticoa harrisii bed is slightly larger (about 3 cm. long, incomplete), and shows the characters of the species. DISCUSSION. Araucarites baqueroensis differs in size and shape from the other species (A . minimus] described in the present paper. Araucaria sp. as described by Berry (1924) from the Jurassic of Santa Cruz NEW GYMNOSPERMS FROM TICO ARGENTINA 273 Province (Argentina) is similar in size and general shape, but the ligule appears much narrower than in the present species, especially in its distal sector. Feruglio's Araucaria sp. from Cerro Cuadrado and Punta del Barco may well belong to Araucarites baqueroensis, but only a brief description and no figures are available. It was found in the same formation but in the upper member where the plant association is somewhat different. Some specimens from Graham Land described as Araucarites cutchensis Feistm. by Halle (19130;, pi. 8, figs. 5, 7, 8) are similar in size and shape but they have a pointed acuminate apex and a narrower ligule. The type specimens described by Feistmantel (1876) are also similar to A. baqueroensis but they show a long acuminate apex (Feistmantel 1876, pi. 9, fig. i), although this is lacking in some of the specimens. In subsequent publications, the same author included markedly different forms in this species (cf. Feistmantel 1877, pi. 14, figs. 6, 8). 16 17 18 20 19 FIGS. 16, 18-20. Araucarites minimus sp. n. Fig. 16. Ovuliferous scale (reconstructed), Xi. o, ovule ; I, ligule ; e, bract. Fig. 18. Cells of outer membrane. Slide LP 233, X3yo. Fig. 19. Cells with straight and unpitted walls of membrane (3) of diagnosis. Slide LP 235, X37O. Fig. 20. Cells of granulose membrane (2) of diagnosis. Slide LP234, X37°- FIG. 17. Araucarites baqueroensis sp. n. Ovuliferous scale (reconstructed), xi. Lettering as for Fig. 16. 274 NEW GYMNOSPERMS FROM TICO, ARGENTINA Araucarites minimus sp. n. (PI. i, figs. 6, 7 ; Text-figs. 16, 18-20) DIAGNOSIS. Broadly cuneiform ovuliferous scales with well developed lateral wings and acuminate apex up to 3 mm. long. Width of scale typically 8 mm. length at least 7 mm. broken at base of attachment. Single oval seed immersed in scale, typically 3-5-5 mm. long x 2-2-5 mm- wide, broadest near distal end. Delicate wings expanded laterally up to 3-5 mm. from edge of seed, with irregularly lobed margins, showing sinuous bands of probably fibrous cells. Short ligule, not exceeding external margin of bract, about 3-4 mm. wide, sometimes seen. Four cutinized membranes present : (i) Outside of scale, including outside of integument of seed, rather strong, finely granular with more or less isodiametric cells, about 20-30 ft in diameter, usually with rounded contours and thick anticlinal walls (up to 4 /*) . Membrane on wings becoming very delicate and losing all traces of cell contours. Dark contents usually adhering to this membrane, forming definite bands disposed in close files. (2) Thin granular membrane showing markedly elongated cells, 15-20 fji wide, with anticlinal pitted and minutely crenulate walls. (3) Thin membrane with markedly elongated cells, 5-15 /* wide, but showing no granules on surface, with straight, unpitted walls. (4) Structureless strong membrane, probably corresponding to megaspore wall. HOLOTYPE. LP 6329. MATERIAL. In addition to the holotype, LP 6328, 6330-44 ; LIL 2739-2745 ; BMNH, ¥.52255-60. Slides LP 231-235 ; BMNH, ¥.52261. HORIZON AND LOCALITY. Lower Cretaceous, Baquero Formation, lower member, Araucarites bed ; Estancia Bajo Grande, Santa Cruz Province, Argentina. DESCRIPTION. Araucarites minimus is abundant in Bajo Grande. The delicate wings may be lacking, but this is probably an accident of preservation. The size of the seed is fairly uniform. The seed is embedded in the scale and covered by its surface. It occupies a hollow cavity which may be filled with mineral substances (powdery calcite or limonite) or preserved as a carbonized body. These coal frag- ments when treated with dilute alkali show three types of membranes (membranes 2-4 of the diagnosis). The two membranes with narrow and elongated cells might both belong to the nucellus cuticle but I prefer merely to designate them as mem- branes 2 and 3 until more material shows their true nature. The rest of the scale is also covered by a carbonaceous film which is broken into small fragments. These fragments after treatment with dilute alkali, show the same type of membrane, assumed to be the outside of the scale. It is similar on the acuminate apical sector, on the base and top of the seed. The ligule is usually broken, but in a few specimens it is seen to be similar to that of Araucarites phillipsii described by Kendall (1949 : 155, text-fig. IB). DISCUSSION. I have referred these scales to the Araucariaceae because they have one seed and the remnants of a ligule embedded in the bract. The size of the scales makes it possible to estimate the approximate size of the female cone, which might NEW GYMNOSPERMS FROM TICO, ARGENTINA 275 have been about 2 cm. in diameter (relatively small for the Araucariaceae) . Most of the living species have large cones and so has the fossil Araucaria mirabilis (Spegazzini) described and figured by Calder (1953) from the Jurassic of Santa Cruz Province. Furthermore, it is possible to include A. minimus in the Eutacta section of the genus Araucaria, because the scales of the other two sections, viz. Bunya and Colymbea are of a different type. As already pointed out by Berry (1924) and others, it is curious that the two living representatives of the family in South America, Araucaria araucana (Mol.) K. Koch and A. angustifolia (Bertolini) O. Kuntze, belong to the Colymbea section, while all fossil representatives so far found in the same area may be classed in the Eutacta section, including the two species from Tic 6. Fossil material referred to Araucarites (cone scales) has been described from Jurassic and Lower Cretaceous strata in South America. Halle (19130;) describes several female cone scales from the Middle Jurassic of Graham Land. All this material is included in Feistmantel's Araucarites cutchensis described from the Jurassic of India which represents a rather unnatural assemblage. All the figured specimens (Halle 1913^, pi. 8, figs. 3-8) are much larger than ours and have a longer acuminate apex. The original description of A. cutchensis is poor, but judging from Feistmantel's (1876) figures A. minimus differs in shape and is much smaller. The specimens of A . cutchensis figured by Feistmantel (1877) are also different in size and shape. There are also differences in the shape of the scales figured by Feistmantel (1879), although they are closer in size to A. minimus. Araucarites macropterus Feistmantel is much larger than A. minimus. Berry (1924) described cone scales of Araucaria sp. from the Laguna del Carbon, in the Gran Bajo de San Julian, Santa Cruz Province.1 These scales differ in shape and size from A . minimus by having a less developed wing and a shortly acuminate apex. Feruglio (1951) described scales of Araucaria from the Laguna del Molino of the Gran Bajo de San Julian, Santa Cruz, which he considered as Upper Jurassic or Lower Cretaceous. No figures were given, but the description suggests that they are larger. In the same paper, Feruglio briefly described other scales from the Punta del Barco and Cerro Cuadrado localities, comparing his material with that described by Berry (1924), and therefore different from ours. Araucarites nipaniensis Singh (1957) from the Jurassic of India, although only slightly larger than A . minimus differs in the shape of the seed and in having a large distal field between seed and margin of bract. No acuminate apex has been reported in this species. Araucarites phillipsi Carruthers from the Jurassic of Yorkshire, as described by Kendall (1949) is larger (up to 1-7 cm. long). It agrees in having a short ligule but the acuminate apex of the bract is not as prominent as in our species. The wings 1 1 have doubts as to the age of these sediments which Berry described as Rhaetic. Feruglio (1951) considers that the Gran Bajo de San Julian strata which contain these plants are either Jurassic or Lower Cretaceous. Stipanicic (1957) in a complete geological survey of this area, concludes that the Laguna del Carbon section (as well as all other similar exposures in the Gran Bajo de San Julian) belong to his Matilde Formation, Middle to Upper Jurassic in age. The facts given to ascertain this age are not conclusive ; there are few plants, most of them useless for precise age determination, such as Cladophlebis, Sphenopteris, Ptilophyllum and Podocarpus. The mention of Athrotaxis cf. ungeri points to a relation with the Baquero Formation or similar strata from the Lago San Martin area, also in Santa Cruz Province. GEOL. 13, 5. 21 276 NEW GYMNOSPERMS FROM TICO, ARGENTINA in A . minimus are also more developed. The epidermis of the cone scale differs in its more elongated cells but the cells of the nucellus cuticle are similar and so is the megaspore membrane. Family PODOCARPACEAE Genus TRISACOCLADUS nov. DERIVATION OF NAME. Trisaco, three air bladders of the pollen grains, cladus, leaf. DIAGNOSIS. As for the only species, Trisacocladus tigrensis sp. n. Trisacocladus tigrensis sp. n. (PI. 4, fig. 21 ; PI. 5, figs. 22-39 '< PI- 8, figs. 56-67 ; Text-figs. 21-25) DIAGNOSIS. Twigs branched, widest seen 0-6 cm. and longest 5 cm. (base and apex broken). Leaves disposed in a close spiral, when detached from branch leaving a marked rhomboidal cushion with a slight depression, probably representing the vascular supply. Leaves radially disposed or spreading in one plane, with rounded apex, decurrent, with margins entire, typically 5-6 mm. long x 0-8 mm. wide. A median longitudinal depression suggesting the mid- vein sometimes seen. Substance of lamina thick. Male cones axillary, attached and shortly pedunculate, oval, gradually tapering towards apex, smallest seen 9 mm. x 3 mm., largest 15 mm. x 4-5 mm. Central axis straight, up to 2 mm. wide. Microsporophylls spirally arranged, inserted at right angles to axis (sometimes more than 90°) composed of a main branch expanding at distal end into a head also placed at right angles (parallel to cone axis). Head extending upwards as short laminar projection, up to 3 mm. long, covering base of next sporophyll. Head also extending downwards as short keel. Two (or more?) oval pollen sacs present, up to I mm. long. Cuticle of sporophyll delicate, showing more or less elongated cells, 10-15 /* wide, with straight walls. Membrane of pollen sac appearing structureless and irregularly granulose. Pollen usually of Trisaccites type, varying in shape and size from 12-32 /JL in equatorial diameter. Equatorial outline somewhat triangular, sometimes round or oval. Three (sometimes two) small air bladders usually present on distal half of body. Exine of body appearing psilate or faintly foveolate, while on wings it may be reticulate. Margins of air bladders finely crenulate. Female cone-like structures probably laterally inserted on branches, subtended by a few sterile leaves. Longest cone seen 2 cm. x 7 mm. wide. Central fleshy axis, 2-6 mm. wide, bearing irregularly (or bilaterally?) erect ovules (and small linear bracts up to 5 mm. long?). Ovules orthotropous, close together, typically oval, 2-3 mm. long x 1-5-2 mm. wide composed of three cutinized layers : (i) innermost delicate membrane, finely granular, showing no structure, assumed to be the mega- spore wall ; (2) thin membrane showing elongated narrow cells, 5-15 /£ wide, pro- jecting into short acuminate apex ; cells becoming much shortened towards base and apex, assumed to be the nucellus ; (3) outer membrane showing no structure but NEW GYMNOSPERMS FROM TICO, ARGENTINA 277 many circular granules irregularly disposed or tending to form ill-defined files assumed to be the integument. Impressions of ovules show longitudinally elongated bulging meshes crowded with small circular granules, tending to be disposed in files. Seeds developing a stone, showing a marked micropilar projection and meshes on outer surface, crowded with small granules. HOLOTYPE. LP 5826 (Trisacodadus bed). MATERIAL. In addition to the holotype LP 5827-42, 5844-50, 6157-69, 6175-81, 6211-32; BMNH, V. 52235-38, ¥.52240-50, ¥.52252-53 (Trisacodadus bed); LP 6182-91 (Ginkgoites tigrensis bed) ; LP 58186, 6170-74, 6430-39 ; BMNH, ¥.52239, ¥.52251 (Upper bed). Slides LP 54-65, 155-166, 244, 300-302 ; BMNH, V. 52234. HORIZON AND LOCALITY. Lower Cretaceous, Baquero Formation, lower member, Trisacodadus, Ginkgoites tigrensis and Upper beds. Estancia Bajo Tigre, Santa Cruz Province, Argentina. DESCRIPTION. This plant occurs in a newly discovered bed of the Bajo Tigre here referred to as the Trisacodadus bed. Deep horizontal tunnels made by coal miners enabled me to collect abundant material. The bed extends horizontally and is located on the roof of these tunnels. It yielded abundant Trisacodadus tigrensis branches and cones and also freshwater phyllopods but no other fossils. The rock is pale brown in colour and the fossils are dark brown. The matrix is fine grained, but a few centimetres below this layer the grain becomes coarser and the rock includes many coal fragments, some of considerable size. Numerous branches have radially disposed leaves but one specimen shows leaves spreading in one plane. Both types are here considered as belonging to the same plant. No trace of cuticle was found. All attempts, including transfers, failed. The leaves are usually covered by a very fragile film which may be the mineralized remnant of the cuticle, but it shows no structure. The phyllotaxis of the leaves is probably 3/8. Fragments of this conifer were also found in other beds of Bajo Tigre (Ginkgoites tigrensis and Upper beds, about 5 metres above). The holotype shows clearly the attachment of a male cone to a short fragmentary branch. On the axis of this attachment a few linear leaves up to 6 mm. long are present, and they are comparable to those found on the sterile branches. All the other male cones were found isolated. As their shape, size and pollen content are very similar to the holotype, they are all placed in Trisacodadus tigrensis. It is not clear whether the pollen sacs are placed on the adaxial or abaxial surface of the sporophyll, but as they are believed to be of podocarpaceous affinity, I presume their position is abaxial. Many cones have their pollen sacs still preserved as coaly oval bodies which may be easily separated and treated with very dilute alkali. In such sacs the pollen is also well preserved. Similar cones yielding pollen grains of the same type are also found in the Ginkgoites tigrensis bed associated with comparable sterile twigs. One of these cones (LP 6184) has at its base a few leaves similar to those described for the holotype. In the Upper bed, Ginkgoites is missing while Trisacodadus becomes abundant. Many twigs and male cones are found in associa- 278 NEW GYMNOSPERMS FROM TICO, ARGENTINA FIGS. 21-24. Trisacocladus tigrensis gen. et sp. n. Fig. 21. Two membranes seen on a translucent specimen. Inner (dotted) belongs to the megaspore membrane ; the outer is the nucellus membrane. Slide LP 159, x 25. Fig. 22. Outlines of the elongated bulging meshes (two with crowded small granules) seen in impressions of seeds. LP 5849, X25. Fig. 23. Ideal section of the ovule showing megaspore membrane, nucellus and in- tegument (with inner content). X25. Fig 24. Cells of cuticle of the microsporophyll. Slide LP 50, X370. tion. The pollen grains and the gross morphology of twigs and leaves are in all respects similar. When preparing the pollen sacs, many grains were found still attached to the irregularly granulose membrane, although most of them spread free when the sac burst. The three small air bladders give the grains a typical appearance, similar to the sporomorphs described as Trisaccites from Australia, Tasmania and New Zealand (Cookson & Pike 1954, Couper 1960). In the present material the bladders are flattened in the distal half of the body (unexpanded) . However, in the same NEW GYMNOSPERMS FROM TIC6, ARGENTINA 279 pollen sacs a few grains with turgid bladders have been found. Such grains closely resemble the sporomorph Microcachrydites, found also in Australia, New Zealand, Tasmania and Antarctica. A detailed study of the pollen grains from Bajo Tigre has been published by Gamerro (1965). Only one specimen (LP 5846) shows what may be an organic attachment of a female cone to a shoot bearing Trisacocladus tigrensis leaves. A few leaves are placed in the axil of the shoot and the presumed base of the cone. The typical fleshy axis of the cone is conspicuous and shows many adhesions, mainly quartz grains and Triletes type megaspores ; it was probably sticky for irregularly placed leaf fragments are also found glued to the surface. There is little doubt that the axis was originally cylindrical. It usually bears ovules on two sides of the compression, but in some specimens they occur in other positions. Some axes show definite scars which may belong to the bases of fallen ovules. If this is true, then the cone axis was covered by radially disposed ovules. No bract or any sterile appendage was seen in relation to the ovules and the axis. Some sterile linear bracts (or leaves?) may be seen crossing the exposed surface of the cone axis, but only in a few specimens, and I am not sure that they bear any relation to the ovules (PL 5, fig. 39). The ovules are erect and show no covers or outgrowths. The integument is thick and possibly had some cell contents which formed the granules mentioned in the diagnosis. The nucellus is strongly cutinized down to near the base (chalaza), its cells near the base and apex being much shorter than at the middle. The anticlinal walls are also more strongly cutinized towards the base and apex. The ovules still attached to the cone axis have some of their membranes preserved, and they were treated in the usual way with dilute alkali. They show the three membranes of the diagnosis, the inner and structureless one being the megaspore wall. The chalaza is clearly seen in some specimens, being somewhat elliptical and transversely elongated (probably the ovules were somewhat flattened). The scars seen on the axis are similar in size and shape and may correspond to the chalaza. Detached seeds referred to T. tigrensis are preserved in two ways : 1. The stone is well developed within a cavity and shows the characteristic meshes on its outer surface. Usually no inner cuticle is preserved. 2. The seed is flattened as a disc from which good cutinized membranes can be obtained. The meshes on the outside are usually seen clearly. The stone is not preserved as a separate recognizable layer. Seeds still attached to the cone axis are always preserved in this manner. Similar venation in the preservation of conifer seeds has been noted occasionally in other floras. Where the external meshes are seen clearly they distinguish the seed of T. tigrensis from that of the associated Karkenia incurva Archangelsky (1965). These cones are identified with the foliage already described as Trisacocladus tigrensis for the following reasons : (i) there is one specimen which shows what I believe is an organic connection ; (2) in the bed where these cones occur, only this type of foliage and male cones are known and no other plant remains have so far been found, among hundreds of vegetable fragments ; (3) in a lower bed, where Ginkgoites 28o NEW GYMNOSPERMS FROM TIC6, ARGENTINA tigrensis is abundant, Trisacocladus tigrensis foliage and male and female cones also occur. (4) In another (Upper) bed which is probably a lateral equivalent to the Trisacocladus bed, distant about 700 metres from it, seeds of the cones and shoots plus male cones also occur together. Thus, these elements are associated in three different beds and there is one probable organic connection. DISCUSSION. In the absence of cuticle it is not possible to relate Trisacocladus with Tomaxellia, which occurs in the same formation but at different localities. There is agreement in the shape of leaves, although individual leaves of Trisacocladus are smaller. In Tomaxellia leaves spreading in one plane were not observed. All Elatocladus species described by Halle (1913^) from Hope Bay, Graham Land, are different. E. heterophylla is somewhat similar but differs in having leaves with an acute apex (on both short and long leaves) and in the variation of the size of the leaves found on different shoots (not so marked in Trisacocladus tigrensis as in E, heterophylla). There is also a difference in age, the Antarctic species being probably Middle Jurassic. The male cones of T. tigrensis are of the type usually referred to the organ genus Masculostrobus. Several species of this genus are known from Jurassic and Cre- taceous strata. Only one, M. sahnii Vishnu Mittre (1956) from the Jurassic of India, has pollen grains bearing three air bladders. The Indian cones are smaller (7 mm. x 3 mm.). Pollen grains usually bear 3 bladders, but 2 or 4 were also seen. The equatorial diameter of the grains is somewhat greater (32-39 ju, against 12-32 ju, in the present species). A few leaves on the basal portion of the cone of M. sahnii are known. They agree in shape with those of Trisacocladus tigrensis, but are smaller. Vishnu Mittre (1957) considers that the female structures of M. sahnii probably belong to Nipanioruha granthia Rao discussed later. Pollen grains similar to those found in the cones, were observed in many preparations of dispersed pollen grains in sediments of other beds. They are commonly preserved with unexpanded air bladders and subtriangular equatorial outline. It is interesting to note that Tri- saccites grains (like ours) are known from Upper Jurassic, Cretaceous and Early Tertiary sediments of Australia, New Zealand and Tasmania. The same geological range applies to Microcachrydites, another palynomorph, which has been found in the same localities and also in Antarctica (Cookson 1947). Among Recent conifers Trisacocladus can be matched only within the Podocar- paceae. The leaves are similar to those found in Podocarpus and Dacrydium. Its male cones resemble in shape some cones of Podocarpus, but there is a clearly denned sporophyll axis in Trisacocladus instead of a more or less laminar micro- sporophyll found in Recent Podocarpaceae. On the other hand, pollen with three air bladders is found in some Podocarpus (e.g. P. dacrydioides] , in Microcachrys and in Microstrobos (—Phaerosphera). The leaves of the two last genera are, however, different, being small and scale-like. It is noteworthy that in one species of Phyllo- cladus (P. glaucus), three air bladders are sometimes found in the pollen (Cranwell 1940). In the Jurassic of Nipania (Rajmahal Series of India), many petrified coniferous remains have been described, some belonging to the Podocarpaceae. Nipanioruha NEW GYMNOSPERMS FROM TICO, ARGENTINA 281 (as described by Rao 1946, and lately emended by Vishnu Mittre 1957) bears leaves comparable in shape with those of Trisacocladus and there is much more information available on the anatomy of these leaves which is lacking in our material. Male cones of the Indian genus are surrounded by leaves and therefore differ from Tri- sacocladus male cones which are shortly pedunculate and not covered by leaves. Both types of cones produce pollen with three air bladders. 25 26 FIG. 25. Tvisacocladus tigrensis gen. et sp. n. Pollen grain showing three air sacs with their basal depressions. Slide LP 244, x noo. FIG. 26. Apterocladus lanceolatus gen. et sp. n. Pollen grain showing three rudimentary equatorial air bladders. Slide LP 242, x noo. Among fossil conifers, our female cones may be compared with Mehtaia Vishnu Mittre (1957), a Jurassic Podocarp from India. Mehtaia has erect ovules, no epi- matium but small bracts. It differs, however, in its much thinner axis and in its markedly curved micropyle (straight in our material). Furthermore, the leaves of Mehtaia are different from those of Trisacocladus. All three species of Mehtaia have much smaller cones (the longest is 10 mm. against 20 mm. in Trisacocladus}. Other female cones of the same formation from India differ in having inverted ovules (including those of Nipanioruha granthia Rao mentioned above), well developed bracts and sometimes also epimatia. As far as I am aware, there are no other similar female structures comparable with ours. The affinity of the Trisacocladus cones is close to some Recent Podocarpaceae, but there are a few important differences (some probably due to lack of information about accessory structures in our material, such as the bracts). Recent Podocar- paceae have a well developed epimatium which covers the ovules to a varying degree. They also possess bracts in relation to the ovules. Our material may have bracts but no structure which could possibly suggest an epimatium. The ovules are thus naked. I believe that this structure (epimatium), if originally present, would surely 282 NEW GYMNOSPERMS FROM TIC6, ARGENTINA have been preserved together with the ovules in at least some of the numerous speci- mens examined. Although this is negative evidence, it is nevertheless suggestive. Among living Podocarpaceae those having cones with erect ovules come closer to Trisacodadus. Phyllodadus and Microstrobos (=Phaerosphera] have erect ovules but also well developed epimatia (in Phyllodadus there is also an aril which, however, may be considered as a secondary feature) . The cones of the living genera are usually smaller and have fewer ovules. Microstrobos is known only from Tasmania and South-east Australia. Recent Phyllodadus is known from New Zealand, Tasmania, Borneo and the New Guinea area. As a fossil it has been recorded from the Tertiary of Australia and New Zealand (Cookson & Pike 1954). All characters of Trisacodadus (leaves, male and female cones) have been matched separately among living representatives of the Podocarpaceae. All other conifers may be excluded from consideration as they differ in their reproductive structures. The Recent genera Podocarpus, Dacrydium, Microstrobos, Phyllodadus and Micro- cachrys share some characters with the fossil form. Female cones with erect ovules are not known in Podocarpus, Dacrydium or Microcachrys. Pollen grains with three air bladders, on the other hand, are unknown in Dacrydium and unusual in Phyllo- dadus (and most species of Podocarpus). Long linear leaves are known only in Podocarpus and Dacrydium. It thus appears that no Recent genus of the Podo- carpaceae shares all morphological characters with Trisacodadus. The Pata- gonian plant may well belong to a new genus of the family which became extinct in the Upper Mesozoic or Lower Tertiary. The geographical distribution of these plants was at least mainly gondwanic, being present in India, Australia, New Zealand, Antarctica and South America. This assemblage constitutes an interesting stock of morphological characters from which many modern genera of Podocarpaceae may have arisen. These considerations support the views expressed by Florin (1963) as to the origin and distribution of the Podocarpaceae in the past. The wide gap in our knowledge of this family in the Mesozoic is only partly filled by the impressions of leafy shoots, placed in noncommital genera such as Pagiophyllum or Elatodadus. But evidence is slowly accumulating in respect of this important family which must have played an outstanding role in southern lands. Our material shows how hazardous it may be to accept the presence of a living genus in old rocks on the basis of only one part of the plant. This is certainly the case with the leaves which are related to the living genus Podocarpus, and which may subsequently prove to have closer relation- ship with other members of the family, when fertile structures become available. The same may be said regarding the dispersed pollen in sediments. While it may definitely indicate a family link, it may also lead to erroneous conclusions when generic relations are established. Genus APTEROCLADUS nov. DERIVATION OF NAME. Aptero, no wings, cladus, leaf. DIAGNOSIS. As for the only species, Apterocladus lanceolatus sp. n. NEW GYMNOSPERMS FROM TIC6, ARGENTINA 283 Apterocladus lanceolatus sp. n. (PI. 6, figs. 40-48 ; PL 7, figs. 52, 55 ; PL 8, figs. 68-70 ; Text-figs. 27, 28, 34) DIAGNOSIS. Woody plants. Lateral axis of penultimate order up to 2 mm. in diameter, bearing leafy branches at a wide angle, 8 mm. apart (seen in one specimen). Branchlets straight or slightly curved, longest seen (incomplete) 3 cm. x 1-5 mm wide (excluding leaves) or 1-7 cm. wide (including leaves). Leaves homomorphic bifacial, spirally disposed but expanded in approximately the same plane by twisting of the decurrent leaf bases. Leaves firm, coriaceous, entire and flat, straight, with one median vein, spreading 50-80°, lanceolate with acute apex, 6-8 mm. long x 1-5-2 mm. broad, maximum breadth near middle of blade, gradually tapering towards apex, abruptly contracted on base, broadly decurrent, each sheathing the axis for a distance of about 1-5-2 mm. downwards. Cuticle 2-3 /i thick. Stomata present on one cuticle, forming two bands near margins. Stomata in close bands, forming ill-defined rows, longitudinally orientated ; subsidiary cells of neighbouring stomata often in contact (lateral and polar); sometimes two stomata sharing a subsidiary cell. Epidermal cells in stomatal bands usually tending to be isodiametric, sometimes rectangular. Epidermal cells on margins and middle of lamina usually rectangular or square, forming definite files. Epidermal cells on nonstomatiferous cuticle square or rectangular, forming definite files, all about 15-25 /i wide. Anticlinal cell walls unpitted, slightly sinuous, 2 fi thick. Surface wall finely granular. No papillae or hairs seen. Stomata typically haplocheilic, monocyclic (sometimes imperfectly dicyclic). Stomatal apparatus oval. Guard cells slightly sunken, very feebly cutinized, sur- rounded by typically 5-6 subsidiary cells. Subsidiary cells may or may not be differentiated into polar and lateral. When differentiated, polar subsidiary cells usually square. Subsidiary cells sometimes with thickened surface wall. Mouth of pit oval or rectangular, with marked rim of cutin formed by fusion of subsidiary cells. Subsidiary cells sometimes with strong external cutin ridge parallel to mouth of pit, which may fuse with neighbouring thickening of adjacent subsidiary cells, forming a sort of ring. Hypodermis not cutinized. Male cones shortly pedunculate, laterally attached to branches, oval, 0-6 x 0-35- 0-5 cm. wide, with microsporophylls spirally inserted in central axis, close together, composed of main branch and distal (outer) leafy rhomboidal and acuminate head. More than one pollen sac present on each sporophyll. Pollen grains sometimes with three rudimentary air bladders, almost spherical, sometimes a little flattened in polar axis. Equatorial diameter about 45 /i and polar diameter (flattened specimens) about 37 fi. Exine delicate, finely granular in texture and thinner on one half. HOLOTYPE. LP 6296 (Bajo Grande, Araucarites bed). MATERIAL. In addition to the holotype, LP 5665, 5666, 5698-5700, 5730, 5749 (Bajo Tigre, Ptilophyllum bed) ; LP 6302-6306, 6317-6327 ; BMNH, ¥.52282-93 (Bajo Grande, Araucarites bed). Slides LP 228-230, 242, 243, 245, 246 ; BMNH, ¥.52294-95 (Bajo Grande, Araucarites bed). 284 NEW GYMNOSPERMS FROM TIC6, ARGENTINA HORIZON AND LOCALITY. Lower Cretaceous, Baquero Formation, lower member, Bajo Grande, Araucarites bed and Bajo Tigre, Ptilophyllum bed, Santa Cruz Province, Argentina. DESCRIPTION. Apterocladus lanceolatus is common in Bajo Grande, and occurs in the same bed as the other conifers Araucarites minimus, Podocarpus dubius and Tomaxellia biforme. Its preservation is similar, although the cuticle is not easy to prepare. The leaf is coaly and broken into small fragments which cannot be pulled out together. The largest fragments were treated with dilute alkali and then bits of cuticle were separated. However, the cuticle usually adheres very strongly to the rock matrix and better results were obtained by first dissolving the matrix with HF to free the cuticles. In both cases the cuticle appears very fragmentary but shows some diagnostic characters. One specimen (LP 6296) shows clearly a male cone in organic attachment to a leafy shoot. Pollen was obtained by scraping the organic remains on the microsporophylls. All grains are of one kind. Similar isolated male cones are abundant in the same bed. They all possess the same type of pollen grains and have therefore been included in the species. One of the isolated male cones (LP 6323) has a few compressed scale leaves at the base which after treatment with dilute alkali, showed very small fragments of cuticle with elongated epidermal cells. These cuticular fragments are similar to the leaf cuticle of Apterocladus lanceolatus. The cones are single or may be present in pairs (LP 6321, BMNH, ¥.52293). Apterocladus lanceolatus also occurs abundantly in Bajo Tigre, in the Ptilophyllum bed (named here for the first time). These specimens lack a cuticle, but are other- wise identical. Only the impressions of square and rectangular epidermal cells can be seen, arranged in definite longitudinal files. Fragmentary remains, possibly belonging to this species were found in the Tic 6 Amphitheatre. (Otozamites grandis bed, LP 5149 and Ticoa harrisii bed LP 5370). One specimen yielded cuticle frag- ments with square or rectangular epidermal cells, arranged in files. One other speci- 27 28 FIGS. 27, 28. Apterocladus lanceolatus gen. et sp. n. Fig. 27. Male cone showing im- pressions of the outside of some microsporophylls and one leaf near base. BMNH. no. V. 52291, x8. Fig. 28. Male cone showing central axis and insertion of a few micro- sporophylls. LP 6326, x8. NEW GYMNOSPERMS FROM TIC6, ARGENTINA 285 men from the Taeniopteris bed (LP 5279) shows a few leaves more than I cm. long, with tapering base and apex, one vein, but no cuticle is preserved. These fragments suggest that A. lanceolatus may be of widespread occurrence (it is over 50 km. from Tic 6 to Bajo Grande). The pollen grains are interesting in that they have a clear equatorial thickening which is sometimes trilobed. These lobes may well represent incipient air sacs, in which case there is further evidence for including Apterocladus in the Podocar- paceae. Many similar pollen grains found detached in the sediments show this character. A detailed study of the pollen grains has been published by Gamerro (1965). DISCUSSION. Apterocladus lanceolatus differs from all other conifers so far studied in the Tic 6 flora. The leaves are only comparable to Podocarpus dubius although there are many differences in shape and in cuticular structure. Stomata of both species are placed in bands, always longitudinally orientated, with oval or rectangular mouth of pit. These characters definitely indicate a Podocarpaceous affinity. Some Taxodiaceae have leaves of similar habit, but usually they are amphistomatic. Moreover, the orientation of stomata in Taxodiaceae is irregular (they may be longi- tudinally, obliquely or transversely orientated). Only Sequoia, Taxodium and Glyptostrobus have leaves of similar habit to Apterocladus. The pollen grains of A . lanceolatus differ from those of Taxodiaceae in the lack of a papilla-like germinating pore. In this respect they are more like those of Recent Araucariaceae, although presenting some differences in the structure of the exine. Most Recent Podocarpaceae have markedly winged pollen grains, and therefore differ from our species. Only Saxegothaea resembles our species in having wingless grains, but the exine structure is rather different and the size of the grains is smaller in the Recent genus. The vegetative part of Saxegothaea also bears some resemblance to our fossil. The leaves are, however, much longer (although contracted at the base) while the male cones are shorter than those of Apterocladus. The cuticular structure is similar in that Saxegothaea bears stomata on one epidermis, placed in two bands and longitudinally orientated. However, in the Recent genus the stomatal rows are better defined than in Apterocladus. Finally, epidermal cells in the Recent genus although rectangular and placed in files, have markedly sinuous walls, a character lacking in Apterocladus. Nevertheless, Apterocladus appears to be more closely related to Saxegothaea than to any other living conifer. This genus is now known in the Patagonian Andes (Chile and Argentina), not far from where the Baquero Formation is exposed (less than 500 km.). It may be worth mentioning that Saxegothaea has been recorded from the Oligocene of Fueguia (about 700 km. south- wards), although there is some doubt about Dusen's determination (1899). The pollen grains of Apterocladus may well be compared (or confused) with some Araucaricites found in dispersed condition. Its grains have the equatorial thickened zone which might represent incipient sacci, and the thinner exine of one of the halves might represent a colpus. Thus a winged grain may be derived from Apterocladus, such as the grains of Phyllocladus, some of which have very small incipient sacs. On the other hand, the incipient sacci could represent a step in a reductionary series, 286 NEW GYMNOSPERMS FROM TIC6, ARGENTINA in which case the next step would be represented by non-saccate grains (as in Saxegothaea) . The contracted leaf bases of Apterocladus distinguish it from all the species of Elatodadus described by Halle (19130) from the Jurassic of Graham Land. There may be some resemblance to Elatodadus sp. figured by Halle (1913, pi. 5, fig. 8) from the Lower Cretaceous of Rio de los Fosiles, Santa Cruz ; but here the contraction at the base of the leaves is gradual, while it is more marked in Apterocladus. Coronelia molinae Florin (1940), from the Tertiary of Chile shares with Apterocladus the character of contracted leaf -bases, although in the Chilean species this constriction is much more pronounced. There are also differences in cuticular structure, mainly in the presence of papillae and peculiar hairs in C. molinae. There is some resemblance to the two species of Palissya described by Frenguelli (1949) from the Upper Jurassic of Chubut Province, Argentina. The leaves are similar and have contracted bases, but nothing is known about their cuticular structure. Genus PODOCARPUS L'Heritier Podocarpus dubius sp. n. (PI. 6, figs. 49, 50 ; PI. 7, figs. 53, 54 ; Text-figs. 29-33) DIAGNOSIS. Woody plants. Lateral axis of penultimate order up to 3-5 mm. in diameter, bearing leafy branchlets at an angle of about 45°, 1-1-5 cm- apart (in a row). Branchlets usually straight, sometimes slightly curved, longest seen (incomplete) 4 cm. x 2 mm. (not including leaves) or 2 cm. wide (including leaves). Leaves dimorphic. Large leaves usually of bilateral type, spirally disposed, linear, straight or falcate, spreading (40-80°), tapering to acute or acuminate apex, with uncon- tracted, markedly decurrent base, sheathing axis for about 4-5 mm., uninerved, largest 1-3 cm. X 1-5 mm. broad in free portion. Short leaves usually situated at base of branchlets or near apex, or on branches of penultimate order ; they are spirally disposed, adpressed to axis or spreading in all directions, markedly decurrent (decurrent sector sometimes larger than free part of leaf). Short leaves may occasionally be placed among long leaves. Both cuticles of same thickness, about 3-4 fji. Stomata present on both cuticles, few on decurrent sector and forming ill-defined rows. In middle sector of leaf, stomata placed in two well defined bands per epidermis, tending to be near margins leaving a conspicuous median sector devoid of stomata (this sector may correspond to real lateral margins of the leaf). Stomata in bands, forming more or less defined longitudinal rows, longitudinally orientated, sometimes obliquely, never transversely. Stomata of a row sometimes with subsidiary cells in contact, but never shared. Epidermal cells in stomatal bands isodiametric or elongated, with rounded con- tours. Square or rectangular epidermal cells on same rows as stomata. Those on margins and middle of lamina typically rectangular, about 20 /£ wide, placed in files. Median longitudinal sector sometimes occupied by markedly elongated rectangular cells. Cell walls straight, unpitted, 3-4 /JL thick, with finely granular surface. Round delicate papillae, occupying most of cell surface, sometimes seen. NEW GYMNOSPERMS FROM TICO, ARGENTINA 287 Stomata haplocheilic, monocyclic to imperfectly dicyclic. Stomatal apparatus oval. Guard cells slightly sunken, very feebly cutinized, surrounded by typically 4-5 subsidiary cells. Subsidiary cells occasionally with cutin ridges parallel to mouth of pit. Polar subsidiary cells sometimes differentiated, when differentiated usually square. Mouth of pit oval, sometimes rectangular, constricted by a strong rim of cutin formed by fusion of subsidiary cells. Rim of cutin of stomata less conspicuous in one of the two cuticles. Hypodermis not cutinized. HOLOTYPE. LP 6309. MATERIAL. In addition to the holotype, LP 6307-13, 6414-29 ; LIL 2753-62 ; BMNH, ¥.52296-99. Slides ¥.52300-01 ; LP 221-227. HORIZON AND LOCALITY. Lower Cretaceous, Baquero Formation, lower member, Araucarites bed. Estancia Bajo Grande, Santa Cruz Province. DESCRIPTION. This is one of the commonest species in the Araucarites bed of the Bajo Grande locality. The substance of the leaves is coaly and broken into small fragments which can be treated with dilute alkali in order to obtain small bits of cuticle. The leaves of most specimens examined are usually of the long type, and in one rank of a branchlet their length may vary considerably. Short leaves were seen in a few specimens ; they usually occur at the base of branches. Unfortunately it was impossible to obtain cuticular fragments from these small leaves. Long leaves are considered to be of the bilateral type because there is no evidence of torsion at the base and yet they tend to be disposed in one plane. The base of these leaves is wide and they sheath the axis for a considerable distance. Following Laubenfels's classification of coniferous leaves (1953) they fall into his type I, tetragonal in cross- section and falcate in profile. Stomatal bands are clearly defined. Although only small portions of cuticle were obtained I believe that there are more stomata in the bands of the lower cuticle, especially on the unexposed surface. Text-fig. 33 shows a sector of a long leaf. The two black lines may correspond to a keel and thus the stomatal bands between these lines correspond one to half the lower cuticle and the other to half the upper cuticle. DISCUSSION. This species differs clearly from all conifers so far studied in the Tico flora. Tomaxellia degiustoi is the one which comes closest, but it differs in leaf arrangement (radially disposed and not spreading in one plane). Also the cuticular structure is different ; in Tomaxellia stomatal bands have no definite stomatal rows. There is some resemblance to Elatocladus conferta (Oldham) from the Jurassic of Hope Bay (Halle 1913^), although the leaves of Podocarpus dubius sp. n. are usually longer. Florin (1940 : 31) considers E. conferta to be similar to Recent Podocarpus of the Eupodocarpus section. Elatocladus jabalpurensis (Feistm.) and Elatocladus sp. (Halle 19150, pi. 9, figs. 7-9) may also be compared with our species. Florin (1940) considers these two species as Podocarpus, subgenus Stachy carpus. Frenguelli (1949) described some fertile specimens of probable Jurassic age from the Canadon Asfalto, near Paso de los Indios in the Chubut Province. These plants 288 NEW GYMNOSPERMS FROM TICO, ARGENTINA 34 33 FIGS. 29-33. Podocarpus dubius sp. n. Figs. 29-31. Stomata. (Fig. 29, slide LP 225. Fig. 30, slide LP 226, Fig. 31, slide LP 221). All X37O. Fig. 32. A stomatiferous row on lower cuticle. Slide LP 225, x85. Fig. 33. Disposition and orientation of stomata on a bilateral leaf (upper half, probably upper cuticle). Slide LP 226, X3o. FIG. 34. Apterocladus lanceolatus gen. et sp. n. Slide LP 228, X370. Two stomata. NEW GYMNOSPERMS FROM TICO, ARGENTINA 289 were referred to Palissya conferta (Oldham) and Palissya jabalpurensis Feistm. Both species differ from ours in their contracted leaf bases. The female cones, judging from Frenguelli's illustrations (1949, pis. I, 2) are rather obscure structures. His reconstruction of the cone scales (text-fig, i) is tentative and must be critically considered. All previous references of Palissya from India have been discarded. Most of the species have been included in the artificial genus Elatocladus in order to differentiate them from typical species of Palissya better known from shoots and female cones in Europe (Florin 1963). This status should be accepted until better material of the Gondwana floras becomes available. Frenguelli's specimens would be better placed in the genus Elatocladus. Podocarpus? palissyafolia (Berry) Florin from the Baquero Formation in Cerro Cuadrado and Punta del Barco was originally described as Elatocladus palissyafolia, on the basis of a fragmentary specimen (Berry 1924). It differs from Podocarpus dubius in the contracted leaf bases but is similar in other respects. The leaves are not bilateral as in our species. Unfortunately, this material is too fragmentary for exact identification and no new material is available from the type localities. Podocarpus inopinatus Florin, from the Tertiary of Chile, has similar but smaller leaves (4-7 mm. long against up to 1-3 cm. long in our species). There is general agreement in cuticular structure : the leaves are amphistomatic, with marginal bands of stomata and epidermal cells arranged in rows. However, in P. inopinatus neighbouring stomata may share a polar subsidiary cell, but this has not been observed in our material. Moreover the Chilean species has dicyclic stomata while P. dubius has monocyclic to imperfectly dicyclic stomata. Pitting of the anticlinal cell walls reported for the Chilean species is unknown in our material. Podocarpus inopinatus belongs to the section Dacry carpus of the genus (according to Florin). The other two species of Podocarpus described by Florin in the same paper differ in the shape and insertion of their leaves, and in cuticular structure. Among the Indian species of Elatocladus, there are two with bilateral leaves similar to those of P. dubius. In E. plana (Feistmantel 1879, Seward 1919, text-fig. 802, Sahni 1928) the leaves are much longer and disposed closer to each other. E. tenerrima (Feistmantel 1877, Sahni 1928) differs in the typically wide angle of inser- tions of leaves (almost perpendicular to the branches) ; in P. dubius, the leaves normally form an acute angle with the branch. Both Indian species may, however, be closely related to P. dubius in having bilateral and amphistomatic leaves. So far as comparison with living conifers is concerned, Podocarpus dubius, having bilateral leaves may be closely compared with the Podocarpaceae. Among the family, this condition is known to occur in the genera Acmopyle, Dacry dium and Podocarpus sect. Dacry carpus. I believe that our species in leaf morphology and cuticular structure is better placed in Podocarpus. Lacking other characters, this combination is to be preferred (following Florin's (1940) paper on Tertiary material from Chile). The difference in age is, however, considerable and I am not sure that our species really belongs to Podocarpus as we know it from Recent material. The main difference between Podocarpus dubius and other genera with similarly shaped leaves is the bilateral condition of the leaves in P. dubius, not found in any ago NEW GYMNOSPERMS FROM TICO, ARGENTINA other families except the Podocarpaceae. The arrangement of stomata in bands is a common character in both Podocarpaceae and Taxodiaceae, but in the former, more or less well defined files of stomata occur in bands as they do in P. dubius. On the other hand all living Taxodiaceae are exclusive to the Northern Hemisphere, with the exception of Athrotaxis, which is clearly different from our species and is known living and fossil from the Southern Hemisphere. Among Araucariaceae there are no similar species, all possess leaves of one kind (with no foliar dimorphism) . Podocarpus dubius may be considered as another fossil member of the Podocar- paceae which inhabited Patagonia during Lower Cretaceous times. Genus TOMAXELLIA Archangelsky 1963 : 86 EMENDED DIAGNOSIS. Woody plants with branches up to the fourth order. Leaves decurrent, spirally inserted, homomorphic (long) or dimorphic (long and short) . Long leaves narrow, ending with acute apex, widest at their decurrent part, rhom- boidal in transverse section, with sharp lateral angles. Cuticle thick. Stomata present on both cuticles. On lower cuticle, at base of leaves stomata irregularly distributed and indistinctly orientated ; stomata forming irregular bands in middle of lamina. On upper cuticle stomata usually forming two bands from base to near apex. Orientation of stomata on free part of both cuticles oblique to longitudinal, sometimes transverse. Rows of stomata in bands ill- defined. Stomata when in a row usually separated, sometimes with outer encircling cells in contact, never sharing subsidiary cells. Epidermal cells between stomatal bands, rows and margins elongated, rectangular. Stomata similar on both cuticles, monocyclic to dicyclic. Stomatal apparatus oval or circular (never rectangular). Guard cells sunken in a pit, surrounded by typically 4-5 subsidiary cells. Hypodermal cells cutinized, markedly elongated. DISCUSSION. Only one species of Tomaxellia (T. degiustoi] has so far been des- cribed (Archangelsky 1963). New material from Bajo Grande includes a new species, T. biforme, and this together with additional specimens from the type locality in Tic 6 Amphitheatre are described below Comparisons have already been made with Recent genera of Araucariaceae, Taxodiaceae and Podocarpaceae, and with the fossil genera Elatides and Elatodadus. Vishnu-Mittre (1957) described a new genus, Indophyllum for shoots having leaves with free margins of lamina extending beyond the leaf cushion (as in Pagiophyllum] but they are homomorphic and much shorter than in Tomaxellia. One of his three species, Indophyllum raoi, has leaves which are rhomboidal in cross-section. In all three species, the stomata are longitudinally orientated. Only in /. raoi, on the lower cuticle, are the stomata irregularly orientated but they are longitudinally orientated on the upper epidermis and therefore differ from Tomaxellia. No detailed comparison of the stomatal structure is possible, because many characters are not preserved in the Indian material. Indophyllum is of Upper Jurassic age. Other leaves described from the same formation differ in having definite Podocarpaceous characters such as orientation of stomata in bands and stomatal structure. NEW GYMNOSPERMS FROM TICO, ARGENTINA 291 Tomaxellia degiustoi Archangelsky (PI. 4, %. 17) 1963 Tomaxellia degiustoi Archangelsky : 87, pi. 7, figs. 39, 40; pi. 8, figs. 58-60; pi. u ; pi. 12, fig. 79 ; text-figs. 72-74. The new material collected in 1962 from the Ticoa harrisii bed in the Tic 6 Amphi- theatre, yielded some good fragments of this species. When originally described, T. degiustoi was based on two very small fragments of branchlets and isolated leaves found in the rock matrix. With these new specimens knowledge of the gross mor- phology of branches and branchlets is now available. The leaves are spirally disposed on the branches and do not spread in one plane, as they do in many of the Podocar- paceae. The angle formed by the leaves with the axis is always very acute (never more than 45°), usually about 20-30°. Sometimes the leaves are almost adpressed to the axis. The broadest branches seen show clearly the oval scars of the leaf cushions. The branchlets are irregularly disposed on the branches, and may be widely separated or almost opposite. The cuticles are identical with those of typical material. The original diagnosis has been enlarged in respect of the external morphology, the rest has been left unchanged. EMENDED DIAGNOSIS. Woody plants. Branches straight or slightly curved, 3 mm. in diameter, longest seen n cm. (base and apex truncated) bearing irregularly inserted branchlets at distances varying from 1-3-5 cm- Branchlets forming acute angle with branch (about 10-50), straight or slightly curved, 4-5 cm. long (incomplete) bearing homomorphic leaves spirally inserted and spreading at acute angle or tending to be adpressed to axis. Leaves falcate, long and narrow, slightly decurrent, almost acicular, up to 1-3 cm. long x 1-5 mm. wide, ending with acute apex. Leaves broadest at base, gradually tapering towards apex. Leaf cushion on branches large, persistent, up to 8 mm. long X 2-5 mm. broad. Leaf cushions on branchlets small, oval or rhomboidal, 2 mm. long x I mm. broad. Free part of leaf rhomboidal in transverse section, but upper and lower angles somewhat rounded ; in middle region of leaf lower angle bearing wide short papillae. Lateral angles sharp, entire. Cuticle as for original diagnosis. MATERIAL. In addition to the material quoted in 1963 ; LP 5052, 5059, 5117, 5342. HORIZON AND LOCALITY. Lower Cretaceous, Baquero Formation, lower member, Ticoa harrisii bed. Estancia La Magdalena, Tico Amphitheatre, Santa Cruz Province, Argentina. Tomaxellia biforme sp. n. (PL 4, figs. 18-20 ; PL 7, fig. 51 ; Text-figs. 35-39) DIAGNOSIS. Main branches 3 mm. wide, giving off in different planes numerous lateral branchlets of a second order, up to 6 cm. long ; these branches give off at GEOL. 13, 5 22 292 NEW GYMNOSPERMS FROM TIC6, ARGENTINA acute angle lateral shoots of a third order, which may branch once again to give branchlets of ultimate order. Leaf cushions small, oval or rhomboidal, about 1-5 mm. long x i mm. broad. Leaves spirally disposed, mainly of two kinds : (i) short, with rounded apex and adpressed to axis, and (2) long, linear, with acute apex, with free part spreading outwards at almost right angles. Longest leaf, 4-7 mm. x 2 mm. wide ; shortest, 1-5 mm. x 0-8 mm. wide. Length-breadth ratio from 4-3 : i to 1-6 : i. Leaves widest at their decurrent part (on leaf base cushion). Apex of leaves may project towards the stem. Leaves in section probably rhomboidal, with shar- pened lateral angles and rounded lower angle ; upper angle not prominent. Veins not seen. Cuticle 4-5 IJL thick. Stomata present on both cuticles. On lower cuticle, at base (decurrent sector) stomata irregularly distributed and variably orientated ; in middle of lamina forming two irregular bands near lower angle which becomes thinner towards apex. Stomata on upper cuticle forming two bands from base towards apex, leaving clearly two marginal and one central sector devoid of stomata. Orien- tation of stomata on both cuticles tending to be oblique or longitudinal. Two mar- ginal sectors, 10-15 cells wide, devoid of stomata and clearly marked on lower cuticle. On both cuticles, rows of stomata in stomatal bands may be present ; usually not well defined. Stomata present near apex on both cuticles. Stomata when in a row, usually separated, sometimes with outer encircling cells in contact, but never sharing subsidiary cells. Epidermal cells on base (decurrent sector) about 30-35 /* in diameter, isodiametric (square or with rounded contours). Epidermal cells in stomatal bands isodiametric or rectangular (longitudinally or transversely orientated) . Epidermal cells towards margins becoming more rectangular, longitudinally orien- tated, up to more than 100 /* long. On margins, where both cuticles join, epidermal cells project outwards, forming a serrate edge with free apices of cells, this character being more prominent in smaller leaves, and almost non-existent in larger leaves. Cell surface flat, with no trichomes or papillae, sometimes faintly granular. Anti- clinal walls straight, unpitted, strongly cutinized down to hypodermis, about 5 fi thick. Stomata similar on both cuticles, typically imperfectly monocyclic to dicyclic. Stomatal apparatus usually oval or circular (never rectangular). Guard cells sunken in a pit, very feebly cutinized, surrounded by typically 4-5 inner subsidiary cells more or less isodiametric, finely striated, forming mouth of pit, slightly sunken below general epidermal surface. Outer encircling cells strongly cutinized with inner lateral walls slightly overhanging depression of the epistomatal chamber. Hypodermal cells cutinized, markedly elongated. HOLOTYPE. LP 6278. MATERIAL. In addition to the holotype, LP 6279-80, 6283-87, 6289-93, 6297 ; BMNH, ¥.52279-82. Slides LP 199-218 ; BMNH, ¥.52283-86. HORIZON AND LOCALITY. Lower Cretaceous, Baquero Formation, lower member, Araucarites bed ; Estancia Bajo Grande, Santa Cruz Province, Argentina. DESCRIPTION. This is a very common species in the Araucarites bed of Bajo NEW GYMNOSPERMS FROM TIC6, ARGENTINA 293 36 39 38 FIGS. 35-39. Tomaxellia biforme sp. n. Fig. 35. Stoma. Slide LP 202, X37O. Fig. 36. Section of a stoma, X3yo. Fig. 37. A short leaf. Slide LP 200, X3O. Fig. 38. A long (normal) leaf. Slide LP 199, X3O. Fig. 39. Distribution and orientation of stomata on upper epidermis. Slide LP 202, X3O. 294 NEW GYMNOSPERMS FROM TIC6, ARGENTINA Grande. The leaves are usually of the long type, but short leaves are not uncommon ; this short type may occur singly or in groups or series, followed by series of longer leaves. This dimorphism is known to occur among both fossil and Recent conifers. There is great variation among the leaves of this species, but they are alike in cuti- cular structure, and in stomata in particular. The distribution of stomata on the upper epidermis is characteristic ; at the base of the leaf they tend to be transversely orientated. On the lower cuticle, stomata may be present directly on the lower angle or else there may be a zone devoid of stomata. Epidermal cells are usually devoid of papillae, but in one specimen (LP 6286) several leaves showed papilla-like cutin thickenings on the apical sector of the lower cuticle, occurring on the transverse anticlinal walls of rectangular cells. Also in these specimens, the lateral serrate margins are more conspicuous than in other forms of the species, but this is a varying character. Stomatal apparatus is alike in all specimens examined. Sometimes 6 subsidiary cells are present, but usually there are 4-5. The striation of inner sub- sidiary cells is constant, and the striae tend to be placed transversely to the mouth of the pit. ACKNOWLEDGMENTS I would like once more to express my sincere thanks to Professor T. M. Harris (Reading University) for many important suggestions, and to Mr. F. M. Wonnacott (British Museum, Natural History) for critically revising the manuscript. I am also indebted for aid in field excursions to the National Oil Company of Argentina (YPF). Mr. L. Ferreyra of the La Plata Museum took the photographs of the material, and Mr. J. C. Gamerro supplied the photographs of the pollen grains, in- cluded in the present paper. REFERENCES ARCHANGELSKY, S. 1963. A New Mesozoic Flora from Tic6, Santa Cruz Province, Argentina. Bull. Brit. Mus. (Nat. Hist.) Geol., London, 8 : 45—92, pis. 1-12. 1963^ Notas sobre la Flora Fosil de la zona de Tic6, Provincia de Santa Cruz. Ameghiniana, Buenos Aires, 3 : 113-122. 1965. Fossil Ginkgoales from the Tico Flora, Santa Cruz Province, Argentina. Bull. Brit. Mus. (Nat. Hist.) Geol., London, 10 : 119-137, pis. 1-5. BERRY, E. W. 1924. Mesozoic Plants from Patagonia. Amer. J. Soi., Connecticut, 7 : 473-482. COOKSON, I. C. 1947. Plant microfossils from the lignites of the Kerguelen Archipelago. Brit. Austr. N.Z. Antarctic Res. (1929-31), Rep., A2 (8) : 129-142. PIKE, K. M. 1954. The fossil occurrence of Phyllocladus and two other Podocarpaceous types in Australia. Austr. Bot. 2 : 60-68, 2 pis. COUPER, R. A. 1960. New Zealand Mesozoic and Cainozoic Plant Microfossils. N.Z. Geol. Surv. Pal. Bull., Wellington, 32 : 1-87, pis. 1-12. CRANWELL, L. M. 1940. Pollen Grains of the New Zealand Conifers. N.Z. Sci 6- Techn., Wellington 22 (i B) : i 6-17 B. DUSEN, P. 1899. Uber die Tertiare Flora der Magellanslanderen. I. In Nordenkjold, O., Wiss. Ergebn. schwed. Exped. Magellansland 1895-1897. 1 : 87-108, 5 pis. NEW GYMNOSPERMS FROM TIC6, ARGENTINA 295 FEISTMANTEL, O. 1876. Jurassic (Oolitic) Flora of Kach. Pal. Indica, Calcutta (u) 1 : 1-80, pis. 1-12. 1877. Flora of the Jabalpur Group (Upper Gondwanas) in the Son-Narbada Region. Pal. Indica, Calcutta (n) 2 : 1-25, pis. 1-14. 1879. Outliers of the Madras Coast. Pal. Indica, Calcutta (n) 1, 4. FERUGLIO, E. 1951. Piante del Mesozoico della Patagonia. Publ. 1st. Geog. Univ. Torino, 1 : 35-8o, 3 pis. FLORIN, R. 1933. Studien iiber die Cycadales des Mesozoikums. K. svenska VetenskAkad. Handl., Stockholm (3), 12, 5 : 1-134, Pls- 1-16. 1940. The Tertiary Fossil Conifers of South Chile and their Phytogeographical significance. K. svenska VetenskAkad. Handl., Stockholm, 19, 2 : 1-107, Pls- I~6- 1963. The distribution of Conifer and Taxad genera in Time and Space. A eta Horti Bergiani, Uppsala, 20 : 121-312. FRENGUELLI, J. 1949. Addenda a la Flora del Gondwana Superior en la Argentina. I. Physis, Beunos Aires, 20 : 139-146, 2 pis. GAMERRO, J. C. 1965. Morfologia del polen de la conifera Trisacocladus tigrensis Archang. de la Formation Baquero, Provincia de Santa Cruz. Ameghiniana, Buenos Aires, 4 : 31-38. 196513. Morfologia del polen de Apterocladus lanceolatus Archang. (Coniferae) de la Formacir-n Baquer6, Provincia de Santa Cruz. Ameghiniana, Buenos Aires, 4 : 133-138. HALLE, T. G. 1913. Some Mesozoic Plant-bearing deposits in Patagonia and Tierra del Fuego and their Floras. K. svenska VetenskAkad. Handl., Stockholm, 51, 3 : 1-58, pis. 1-5. 1913^. The Mesozoic Flora of Graham Land. Wiss. Ergebn. schwed. Sudpolarexped. (1901-1903), Stockholm, 3, 14 : 1-123, pis. 1-9. HARRIS, T. M. 1964. The Yorkshire Jurassic Flora. II. Caytoniales, Cycadales and Pterido- sperms. viii + 191 pp., 7 pis. British Museum (Natural History), London. KENDALL, M. 1949. A Jurassic member of the Araucariaceae, Ann. Bot., London (n.s.) 13 : 151-161. LAUBENFELS, D. J. DE. 1953. The external morphology of Coniferous leaves. Phytomor- phology, Manchester, 3 : 1-20. PANT, D. D. & NAUTIYAL, D. D. 1963. Cuticle and epidermis of recent Cycadales. Leaves, sporangia and seeds. Senckenberg. biol., Frankfurt am Main, 44 : 257-348. SAHNI, B. 1928. Revision of Indian Fossil Plants. Part I. Coniferales (a. Impressions and Incrustations). Pal. Indica, Calcutta (n.s.) 11 : 1-49, 6 pis. SINGH, G. 1957. Araucarites nipaniensis sp. nov., a female Araucarian cone-scale from the Rajmahal Series. Palaeobotanist, Lucknow, 5 : 64-65, i pi. STIPANICIC, P. N. 1957. El " Complejo Porfirico de la Patagonia Extraandina " y su fauna de anuros. I. Parte geologica. Acta Geol. Lilloana, Tucuman, 1 : 185-230. TEIXEIRA, C. 1948. Flora Mesozoica Portuguesa, 1. Lisboa. ViSHNU-MiTTRE. 1956. Masculostrobus sahnii sp. nov., a petrified Conifer male cone producing three-winged and one and four-winged abnormal pollen grains, from the Jurassic of the Rajmahal Hills, Bihar. Grana Palynologica, Stockholm, 1 : 99-107. I957- Studies on the Fossil Flora of Nipania (Rajmahal Series), Bihar. Coniferales. Palaeobotanist, Lucknow, 6 : 82-112, pis. 1-12. PLATE i Pachypteris elegans sp. n. FIG. i. General aspect of holotype. LP 6233, xi. Ticoa lamellata sp. n. FIG. 2. General aspect of holotype. LP 6250, x 1-3. FIG. 8. Lower cuticle (middle) and upper cuticle (right). Slide LP 174, X3O. FIG. 9. Lower cuticle showing distribution of stomata and small hair bases. Slide LP 174, Xi oo. Altnargetnia incrassata sp. n. FIG. 3. Fragmentary pinnae. Left, BMNH, no. ¥.52664; right, LP 6254, Xi-3- FIG. 4. General aspect of holotype. LP 6255, x 1-3. Araucarites baqueroensis sp. n. FIG. 5. General aspect of holotype. LP 57666, xi. Araucarites minimus sp. n. FIG. 6. General aspect of holotype. LP 6329, X3- FIG. 7. A small scale showing acuminate apex. LP 6330, X3- Bull.B.M. (N.H.) Geol. 13,5 PLATE 1 GEOL. 13, 5. PLATE 2 Pachypteris elegans sp. n. FIG. 10. Lower cuticle of undifferentiated pinna, showing distribution of stomata and venation. Slide LP 169, X3O. FIG. ii. Lower and upper cuticles of a pinnule. Slide LP 170, X3O. FIG. 12. Stomata on lower cuticle. Slide LP 169, X5oo. Bull. B.M. (N.H.) Geol. 13, 5 PLATE 2 PLATE 3 Almargemia incrassata sp. n. FIG. 13. Lower cuticle (left) and upper cuticle (right) showing two types of epidermal cells. Slide LP 184, xioo. FIG. 14. Lower cuticle showing stomata. Slide LP 184, Ticoa lamellata sp. n. FIGS. 15, 16. Stoma focused at two different levels. 15, mouth of pit ; 16, guard cells. Slide LP 174, X500. Bull B.M. (N.H.) GeoL 13, 5 PLATE 3 : Mtl y RrT • *'- V v • • ^ 'J i> ?<»&£>>££ *> W3a$&^ : PLATE 4 Totnaxellia degiustoi Archangelsky FIG. 17. Branched specimen. LP 5117, Xi'3- Tomaxellia biforme sp. n. FIG. 18. General aspect of holotype. LP 6278, XO-75. FIGS. 19, 20. Lower (19) and upper (20) cuticles, showing distribution of stomata. Slide LP 202, x 30. Trisacocladus tigrensis gen. et sp. n. FIG. 21. Two fragmentary branches. LP 5845, XI. Bull. B.M. (N.H.) Geol. 13, 5 PLATE 4 19 PLATE 5 Trisacocladus tigrensis gen. et sp. n. FIG. 22. Same specimen as Fig. 21, X3- FIG. 23. Branched twig with leaves spreading in one plane. LP 5827, Xi'5- FIG. 24. Holotype. Pollen sacs are seen as paler structures among microsporophylls. LP5826, xi -5. FIG. 25. The same, X4- FIGS. 26-28. Male cone, three different magnifications, showing microsporophylls. LP 5829, Xi, X3, x8. FIGS. 29, 30. Wide branch showing leaf scars. BMNH. no. V. 52238, xi and X3- FIGS. 31, 32. Branch with female cone, showing probable attachment. LP 5846. Fig. 31, Xi and Fig. 32, X3. ca, cone axis ; o, ovules still attached to cone axis. FIG. 33. A detached seed. LP 5847, x 10. FIG. 34. A cone axis showing irregular scars. LP 5842, x i. FIGS. 35-37. Cone showing axis and a few ovules laterally attached. LP 5838, x i and X3 ; counterpart LP 6231, X2. The white dots are remnants of megaspores. FIG. 38. Cone axis and attached ovules. LP 6218, X2. FIG. 39. Cone axis with probable bracts (?) transversely disposed. LP 6215, X2. Bull. B.M. (N.H.) Geol. 13, 5 PLATE 5 38 39 PLATE 6 Apterocladus lanceolatus gen. et sp. n. FIG. 40. Holotype, showing attached male cone. LP 6296, xi. FIG. 41. Same, X3- FIGS. 42-44. Branched specimens. Fig. 42, LP 6302, xi; Fig. 43, LP 6304, X2 ; Fig. 44, LP 6303, xi. FIGS. 45, 46. Male cone showing external surface of microsporophylls. BMNH. no. V. 52291, X3 and x8. FIGS. 47, 48. Twin male cones. LP 6321, xi and X3. Podocarpus dubius sp. n. FIG. 49. A branch. LP 6302, xi. FIG. 50. The holotype. LP 6309, xi. Bull. B.M. (N.H.) Geol. 13, 5 » PLATE 7 Tomaxellia biforme sp. n. FIG. 51. Stomata of upper cuticle. Slide LP 202, Apterocladus lanceolatus gen. et sp. n. FIG. 52. Stomata in a band. Slide LP 228, xsoo. FIG. 55. General aspect of upper non-stomatiferous (left) and lower stomatifierous (right) cuticles. Slide LP 228, x 100. Podocarpus dubius sp. n. FIG. 53. General aspect of both stomatiferous cuticles. Slide LP 225, x 100. FIG. 54. A stoma. Slide LP 226, X5oo. Bull. B.M. (N.H.) Geol 13, 5 PLATE 7 i V • * *••> -.>*>• v? %m t \ IA *f ; p^S;' "1|| %\ ;-^^!v r^ i IP o o c. o o c> c: _V Upper Ludlow Shale 3- Aymestry Limestone Upper Limestone Nodular Beds Lower Limestone Lower Ludlow Shale Wenlock Limestone Wenlock Shale FIG. 2. Stratigraphic succession of Silurian cystoid bearing rocks at Dudley. 304 THE BRITISH SILURIAN CYSTOIDS coming from Hurst Hill, also in the Sedgley Inlier. All the Wenlock rocks in the Sedgley Inlier are confined to the immediate vicinity of Hurst Hill while the vast majority of the Sedgley Inlier, including that part on which Sedgley itself is built, is of Ludlow rocks and both Mr. Coope and Dr. Strachan, of the Geology Department, Birmingham University, inform me that it is likely that these specimens came from the Sedgley Limestone at Sedgley. In this case they are considerably younger than all the other species described here. IV. RELATIONSHIPS TO OTHER FAUNAS Schizocystis and Glansicystis are endemic to Britain and the last representatives of the Echinoecrinitidae. All other echinoencrinitids are confined to the Ordovician of the Baltic States and Scandinavia. This element of the fauna had a Baltic origin therefore and represents an evolutionary dead end. Lovenicystis angelini (Jaekel) from Gotland, Sweden, is the only member of the Callocystitidae, apart from the British forms, to occur outside North America. The family ranges from the Upper Ordovician to the Middle Devonian and the British fauna probably represents an evolutionary sideline which migrated across to Europe. It is possible that the Staurocystinae evolved in the British Isles and later migrated back to North America. This subfamily apparently attains its maximum diversity in Britain and appears earlier here than in the United States. V. FUNCTIONAL MORPHOLOGY AND MODE OF LIFE In dealing with the morphology of the cystoids it is convenient to consider the stem, the theca, and the subvective system separately. a. The Stem All the species described had a well developed stem and all except Schizocystis probably had a root structure, but it has not been possible to confirm this in Tetra- cystis oblongus or Prunocystitesfletcheri. The stem is characteristic of the superfamily Glyptocystitida, to which all the present species belong, and is divisible into proximal and distal portions. The proximal portion of the stem is composed of thin, ring-like columnals with a wide lumen, and usually tapers rapidly away from the base of the theca. There are two sorts of proximals : large outer proximals and smaller inner proximals. The outer proximals fit outside the inner proximals and are provided with an internal flange against which the two adjacent inner columnals fit. Exter- nally they may have a flange or a ring of spines or tubercles. Alternating with the outer proximals throughout the proximal portion of the stem are the inner proximals. These are thin rings which are frequently hidden from view by the outer proximals. The internal flange of the outer proximals is provided with two swellings set opposite each other which are interpreted as fulcra. Similarly the upper and lower surfaces of the inner proximals are provided with slight swellings which were probably facets. In Cheirocrinus it is known that the fulcra on the upper and lower surfaces of the outer proximals are set at slightly different positions so that the axis of rotation was different for successive pairs of proximals. It has been impossible to confirm that THE BRITISH SILURIAN CYSTOIDS 305 the same arrangement obtained in the British Silurian cystoids but it is certain that the proximal stem was very flexible. Several have been preserved quite tightly curved but still articulated. The distal portion of the stem was also composed of two sorts of columnals but arranged one above the other. This alternation is not always obvious but was almost certainly present. The distals of both types were relatively tall and narrow in diameter with a narrow lumen. The distal stem generally tapers much less rapidly and in forms with a root structure may not taper at all. The larger distals are the morphological homologues of the outer proximals and are flanged in the species with flanged outer proximals. The smaller distals are, correspondingly, the homologues of the inner proximals. At the termination of the distal stem there is frequently a swelling or some irregular branches which form a root structure. All species with a root structure are believed to have been attached permanently and therefore relied on food-bearing currents coming to them. A flexible stem would be of advantage in that it would allow them to sweep a considerable area around the point of attachment. In forms like Glansicystis it would allow the apex of the theca to be directed upwards at all times. It would also allow flexure in currents and thus reduce the chances of the stem snapping. Finally it would raise the theca up off the substrate thus allowing the food collecting apparatus to collect falling food particles before they reached bottom dwelling forms and keep the rhombs and food collecting apparatus well clear of the turbulent zone just above the bottom, reducing the chances of fouling by mud particles. All species without a root structure are believed to have been free and vagrant. Among the present species this includes only Schizocystis armata which is believed to have used the stem as an organ of locomotion. Schizocystis is a special case and will be dealt with in a separate section. b. The Theca In Pseudocrinites the theca is more or less distinctly lenticular but in all the other species it is ovoid. The mouth was apical and the anus lateral and set fairly high up on the theca. The latter opened in a periproct covered by an anal pyramid surrounded by a single circlet of auxiliary plates. This circlet may be complete or incomplete, apparently even in the same species, and it is difficult to see what function it could have served. It may simply be the remnant of the extensive plated periproctal membrane of the ancestral group, the Cheirocrinidae. The anal pyramid forms a one-way valve system closed either by ligamental or muscular action. There is no indication of any musculature to open the valve and this must be assumed to have been achieved by peristaltic action. As a result the faeces would have been pellet-like and very unlikely to have fouled either respiratory or food bearing currents. Despite its high position on the theca the anus is placed well away from the pec- tinirhombs which are at the same level as the anus and the single rhomb below it is on the opposite side of the theca. In the Callocystitidae, where the arms are folded down the theca, the food grooves of the brachioles are directed towards the centre of the arms and therefore away from the anus in the ambulacra adjacent to it. There are two orifices higher up on the theca usually accepted as the hydropore and 306 THE BRITISH SILURIAN CYSTOIDS gonopore. One is circular and frequently covered by a pyramid of plates resembling a small anal pyramid. This orifice almost certainly acted as an exit and is therefore accepted as the gonopore. Adoral to it, and curved round it in many cases, is a slit-like orifice provided with a rim and usually swelling out at either end. This is accepted as the hydropore. These two orifices are always above rhomb Li : R5 and are taken as the fundamental point for orientating the theca. The ambulacrum which passes between the anus and the hydropore and gonopore in the Glypto- cystitida is always ambulacrum 5. The theca of the glyptocystitid cystoids is composed of a definite number of thecal plates arranged in definite circlets. The formula 4 basals (BB), 5 infra- laterals (ILL), 5 laterals (LL), 5 radials (RR), and 5 orals (OO) is usually given. The ancestral genus, Cheirocrinus, had 6 radials and occasionally examples of other species occur with extra plates but on the whole the formula applies for the first four circlets. In the case of the orals however, it is possible that the usual pentamery of echinoderms has caused the acceptance of too low a number. There were apparently 7 orals in the Glyptocystitida. Kesling (19610, 19616, 19620, 19626) and Kesling & Mintz (1961) have figured the oral areas of Glyptocystites ehlersi Kesling, Jaekelo- cystis hartleyi Schuchert, Rhombifera bohemica Barrande, Cheirocrinus anatiformis (Hall) and Lepadocystis moorei (Meek). In all cases there were 7 plates in the oral area and Oi was interpreted as being compound. To these species can be added Glansicystis baccata (Forbes), Schizocystis armata (Forbes), Apiocystites pentre- matoides Forbes, Staurocystis quadrifasciatus (Pearce) and Pseudocrinites bifasciatus Pearce, all with 7 orals. Pleurocystites probably had 7 orals too. If all these inter- pretations are correct 7 orals are known in six of the seven families within the Glyptocystitida and this situation would seem to be of general occurrence. Text-fig. 3 shows the interpretation adopted here. OO2-5 are the same as in all previous interpretations. Oi is the plate bearing the left portions of the gonopore and hydropore as viewed facing these pores. 67 bears the other portions of these orifices and the small plate, adoral to 67, is O6. This interpretation commends itself as in FIG. 3. Schematic representation of the arrangement of oral plates in the Glyptocystitida. OO, orals. THE BRITISH SILURIAN CYSTOIDS 307 some cases the common suture of 67 and O6 runs from Oi to R5 and in this case the 7 orals form a ring with the plates numbered consecutively. When fully developed all the rhombs are disjunct pectinirhombs. This type of disjunct pectinirhomb is formed of pore-canals, called dichopores, which are basically folds in the surface of the theca. The walls of adjacent dichopores are confluent and all the dichopores open collectively in vestibules. As with all rhombs they are shared between two thecal plates and the common plate suture divides the rhomb into two half-rhombs. The width of a rhomb is measured along the plate suture. At right angles to the suture lies the rhomb axis which divides the rhomb into two demi-rhombs and along which the length of the rhomb is measured. Rhombs in which the length exceeds the width are called compressed and the converse depressed. When the two half-rhombs are unequal the rhomb is unequal and when the two demi-rhombs are unequal it is asymmetrical. Rhombs are developed in three positions in all the species dealt with. These are across the sutures between plates Ba : IL2, Li : R5 and L4 : R3. As far as is known B2 : IL2 is not developed in Prunocystites fletcheri and Schizocystis armata sometimes lacks Li : R5. Otherwise all three rhombs are developed in all species. When fully developed there is a constant difference between the two half-rhombs of each rhomb. The half-rhombs B2, R5, and R3 all have a raised rim on the absutural side of the series of slits and the slits are generally longer than the corresponding slits of the other half-rhombs. In half-rhombs IL2, Li, and L4 there is a continuous raised rim running around the slits on both the ad- and absutural sides. In several cases the slits are very slightly wider in these half-rhombs than in the preceding half- rhombs. Two trends can be seen in the shape of the rhombs and these can, to a certain extent, be related to the maximum size of the thecae in which they occur. In Pseudocrinites, and other genera or species with relatively large thecae, the rhombs tend to be large, depressed and asymmetrical. They are angular in outline and provided with a large number of dichopores. In forms like Glansicystis, with relatively small thecae, the rhombs tend to be distinctly compressed, symmetrical and the half- rhombs tend to be semi-circular in outline. The number of dichopores rarely exceeds 10. Schizocystis, which has this type of rhomb and a relatively large theca, has an unusually large number of dichopores in its rhombs (up to about 20) but still less than in the rhombs of Pseudocrinites. The constant difference between the two half-rhombs is interpreted as being asso- ciated with the functioning of the rhombs. Sinclair (1948) postulated that there were ciliary currents flowing through the dichopores in life. In the rhombs of Glyptocystites, which have no absutural rims but an adsutural rim on one half-rhomb, he suggested the rim assisted in preventing re-circulation of the currents. The half-rhomb with the rim was therefore the exit and the other the entrance. All further investigations made by the present writer tend to confirm this interpretation. The closed rim of the rhombs in the present species is interpreted as a mechanical barrier to re-circulation. In Staurocystis it is known that the slits of one specimen (BU. Hoi, rhomb Li : R5) are narrower in the incurrent half-rhomb than in the 308 THE BRITISH SILURIAN CYSTOIDS excurrent. This may be a device to prevent harmful particles or organisms entering and clogging the dichopores. The extra length of the incurrent slits may be the result of an attempt to achieve the same cross-sectional area while decreasing the width. Alternatively in some cases the slits of the two half-rhombs appear to be of the same width and in these cases the excurrent would have been stronger than the incurrent. This is known in other cystoids with different types of rhombs and also in blastoids. Thus Macurda (1966) has been able to show that the excurrent from the hydrospires of Globoblastus norwoodi was six times as fast as the incurrent. In this case, and in others the present writer has investigated, there can be no doubt as to which way the currents were flowing. Such an arrangement may be of advantage in directing a strong current away from the theca and thus reducing the chances of re circulation by surface ciliary currents. The trends in the outline of the rhombs are probably a result of the maximum size of the thecae in which they occur. Assuming the rhombs were respiratory, larger thecae would require more oxygen and therefore larger rhombs. This could be achieved by either lengthening the dichopores or increasing their number. The latter is likely to be more efficient since there must be a length in which all the available oxygen will be removed. Increasing the length beyond this will not increase the efficiency of the rhombs. It is significant that in one of the largest rhombs examined by the writer (Li : R5 in Pseudocrinites gordoni, BMNH. E. 23122) the longest dichopore was 6-45 mm. and there were approximately 85 dichopores in 14-85 mm. However, in rhomb L4 : R3 of Echinoencrinites cf. reticulatus (RM. EC. 5515) the longest dichopore is 67 mm. but there are only 9 dichopores in a width of 3-7 mm. Apparently the longest dichopores of both types of rhomb are about the same length. It must be pointed out, however, that the size and shape of a rhomb is controlled by the plates in which it is developed. Thus a rhomb developed across a short suture has a much smaller potential width than one developed across a long one. This in turn is to a certain extent, controlled by the size and shape of the thecae as a whole. Considering the theca as a whole it is possible to deduce a current pattern for the three rhombs. If the above interpretations are correct then water entered the rhombs at the ab- and adoral extremities of the theca and passed out near the ambitus. In all forms of cystoids in which the current patterns have been determined they are the same, i.e. in at the extremities and out near the ambitus. In Glyptocystites and Cheirocrinus there were some additional currents running around the theca and Cystoblastus and Rhombifera represent a departure to a condition where all the cur- rents ran around the theca. Unfortunately sufficiently well preserved specimens of these genera are not available and the current patterns remain unknown in detail. c. The Subvective System By analogy with living crinoids this system functioned as the food collecting apparatus of the cystoids. The fixed species may be conveniently divided into two groups on the basis of their ambulacral structure. These are : species with the ambulacra folded down the sides of the theca, and those with the ambulacra res- THE BRITISH SILURIAN CYSTOIDS 309 tricted to the apical region of the test. In the former there are usually four ambulacra but Pseudocrinites represents a special case and only has two ambulacra. By analogy with species of callocystitids with the full five ambulacra, the four present in most British species are accepted as i, 2, 4 and 5, the last being the ambulacrum passing between the anus and the gonopore. For descriptive purposes it is convenient to number the brachioles in each ambulacrum in the order in which they developed. FIG. 4. Staurocystis quadrifasciatus (Pearce). Camera lucida drawing of apical region to show arrangement of the ambulacra and brachiole numbering system. G, gonopore ; H, hydropore ; M, approximate position of mouth. The arrow points to the periproct. The ambulacrum passing between the anus and gonopore is ambulacrum 5 . BMNH. 481 97. Thus the first formed brachiole of ambulacrum 2 is 21 (see Text-fig. 4). In all the British species the first formed brachiole of each ambulacrum is branched off to the left, looking down the ambulacrum from the mouth to the tip. The second formed brachiole is on the right in all species with four ambulacra and the brachioles alternate regularly throughout the length of the ambulacrum. In both species of Pseudo- crinites described here there are two facets on the left adapically in each ambulacrum. It is suggested that the two adoral facets are all that remains of the two ambulacra missing in this genus (see Text-figs. 16, 21). Kesling (19616) has shown that the two fully developed ambulacra in Pseudocrinites are 2 and 5 and so it is suggested here that the three adoral brachioles in ambulacrum 2 are i1, 21 and 22, while those in ambulacrum 5 are 41, 51 and 52. The ambulacra in the first group are composed of two series of alternating large and small flooring plates (see Text-fig. 10). The larger alternate about the median line and the smaller are inserted between them at the outside edge. One each of both types contributed to the formation of a brachiole facet with the smaller forming 3io THE BRITISH SILURIAN CYSTOIDS the adoral portion in all cases. Along the zigzag suture between the two rows of primary flooring plates ran the main food groove. At each change of direction it gave off lateral branches running to the brachioles. These were biserial structures with a food groove running down the side facing the main food groove. Both the " arms " and the brachioles were provided with cover plates over their food grooves. It is possible that the cover plates of the " arms " were immobile and therefore that food collection was confined to the brachioles. There were apparently three cover plates to every brachiolar in the brachioles. The structure of the brachioles indi- cates that they were incapable of twisting about their axes and therefore it would be advantageous to have the food grooves facing any food bearing currents. From the disposition of the ambulacra food collecting would seem to have been confined to four, or in Pseudocrinites two, zones of collection. The brachioles were capable of curling inwards over the ambulacrum. It is not certain whether this was a pro- tective action or simply the position adopted when at rest. When feeding the brachioles would have been extended and this was probably achieved by hydrostatic action. If so the orientation of the facets gives an indication of the width and cross- section of the zones of collection. Pseudocrinites would seem to have been at a disadvantage compared with the other species. Not only did it have half the potential number of brachioles of its nearest relative, Staurocystis quadrifasciatus, but the brachioles were set further apart. There were also only two dozen brachiolars in each brachiole in P. bifasciatus as opposed to three dozen in Staurocystis. From the lenticular shape of the theca of P. bifasciatus one may infer that if it turned edge-on to any currents it would have brought one zone of collection to face the current and reduced resistance to the current. If the current periodically reversed both zones could function with very little twisting of the stem. The most obvious reversing current is a tidal one. Pseudocrinites pyriformis presumably had a similar mode of life although it had longer brachioles and ambulacra which did not extend as far down the sides of the theca. In Glansicystis, and possibly also Prunocystites, the brachioles are grouped about the mouth and the food grooves on the theca are very short. The brachioles have their food grooves on the adoral side and thus form an apical cone of collection. This arrangement is much better adapted to collecting falling food particles. From the position of the food collecting apparatus it may be inferred that the British Callocystitidae were better adapted to take advantage of currents in a relatively agitated sea while Glansicystis was better able to take advantage of falling food particles in a relatively still sea. These appear to be two extreme conditions which are found randomly distributed among stemmed Pelmatozoa. d. Schizocystis armata (Forbes) As mentioned previously this species differs in most respects from all the others dealt with. It is found over a much wider, though still restricted, area. It is apparently confined to the Wenlock Shale and thus lived in a muddy environment. In its morphology it has one very important difference in lacking a root structure. THE BRITISH SILURIAN CYSTOIDS 311 Kirk (1911) has shown that many stemmed pelmatozoa were in fact vagrant. It is unfortunate that he chose as some of his examples British Silurian species which are now known to have been attached permanently but his main contention is not invalidated by this. It is not unreasonable to assume that all stemmed pelmatozoans which lack a definite root structure were voluntarily vagrant, at least temporarily, during their life-time. If food supplies were erratic temporary fixation would only be of advantage if the animal could voluntarily free itself. One must infer that forms like Ancyrocrinus literally weighed anchor and sought new feeding grounds in times of short supply. Such activity is known in Recent free living crinoids (which in general lack a stem) and may be inferred for extinct ones. On the above reasoning Schizocystis was probably vagrant, and used its stem as a tail. The theca, like the stem, is on the same basic plan as the other species but has differences in detail. It is very variable in form and ornamentation but is generally compressed adorally and this may be an attempt to develop a bilateral symmetry. Taking the anus as posterior the compression is lateral and the two flattened areas correspond to the left and right sides. In many examples the right side is gently convex throughout but the left side is concave or flat adorally and convex aborally. The rhomb Li : R5, which is often undeveloped, lies on the left side of the theca. There are two ambulacra, running posteriorly and anteriorly along the crest of the theca. The posterior ambulacrum is the shorter and sometimes reaches the peri- proct. A curious feature of the ambulacra is that brachioles are developed pre- dominantly on one side only. Brachioles are mainly developed on the left sides of both ambulacra and as a result the food grooves face the left side of the theca in the posterior ambulacrum and the right side in the anterior ambulacrum. It is reasonable to assume that this species was benthonic, and since it lacks a root structure, that the theca lay on the substrate. Placing a theca on a flat surface the most stable position for it is lying on the left side. In the two-rhombed forms this leaves both rhombs clear of the substrate and the mud is unlikely to have fouled the rhombs. This is also true of the gonopore and hydropore which are developed on the left side in a raised portion of the test which is vertical in this orientation. The brachioles of the posterior ambulacrum could browse off the mud while those of the anterior ambulacrum would have their food grooves facing upwards and could possibly have fed off falling food particles. When the third rhomb is present it would have been useless in this orientation as it would have been fouled by the mud as would the gonopore and hydropore. The weight of the stem may have been used as a counterbalance to the theca and so the adoral part may have been kept off the substrate. There are difficulties in this suggestion. If the substrate was muddy the theca would have sunk in to some extent and the zone immediately above the surface would have been muddy if there was even slight turbulence. Possibly three-rhombed forms could tolerate having one inefficient rhomb since the two-rhombed forms toler- ated the complete loss of one. It is difficult to explain the orientation of the food collecting apparatus and either this species was relatively poorly adapted to its new mode of life or it was trying to get the best of both worlds, feeding in the usual manner for the Echinoencrinitidae and browsing off the mud as well. 3I2 THE BRITISH SILURIAN CYSTOIDS VI. GROWTH a. The Stem Plotting the graph (Text-fig. 5) of thecal height against the number of proximal columnals in Staurocystis quadrifasciatus indicates that there is considerable variation. A slightly better grouping is obtained by plotting the diameter of the top-most columnal against the total number of proximals (Text-fig. 6). Growth of the proxi- mal stem was almost entirely confined to the addition of new columnals and this most likely took place at the base of the theca, as in living crinoids. In the distal stem growth was partly achieved by the addition of new distals and partly by the lengthening of existing distals. Since the large and small distals alternate regularly, in many cases, addition of new distals could not have been by the intercalation of " internodals " as in some crinoids. It is very unlikely that proximals migrated relatively down the stem and turned into distals after reaching a given point. This would require considerable changes in shape and the inner proximals would have to emerge from within the outer proximals. Almost certainly new distals were added at the junction of the distal and proximal portions of the stem and in some cases the topmost distals are thinner than succeeding ones and appear to be relatively newly formed. THECAL HEIGHT mm. 30 • 25- 20 15 10 5 FIG. 5. 10 20 30 40 50 NUMBER OF PROXIMALS Staurocystis quadrifasciatus (Pearce). Graphical representation of relationship between thecal height and number of proximal columnals. b. The Theca All thecae known have the same number of thecal plates and within the size range examined growth was achieved by increasing the size of these plates and not by the addition of new plates. The smallest thecae known are still quite large, about 5 mm. in thecal diameter, so the earliest growth stages remain unknown. Fine growth lines THE BRITISH SILURIAN CYSTOIDS 313 are not very often preserved in these cystoids but growth can be shown to have been holoperipheral in all thecal plates. Some specimens of all species show colour banding on the plates which is associated with growth. This is most clearly shown in SM. Am6o, a worn specimen of Tetracystis oblonga showing, on all plates visible, concentric light and dark bands. These bands are thickest where growth has been greatest and thinnest where growth has been least. 63 shows a central light coloured area followed outwards by four alternating dark and light bands and the suggestion of another dark band at the suture. These bands can be matched in other plates but in general the central light area is smaller relatively, particularly on the only radial plate clearly visible (R5). Measuring outwards from the centre of 63 towards the common suture with IL3, the thicknesses of the bands are as follows : 0-6 mm. 0-23 mm. 0-53 mm. 0-33 mm. and 0-27 mm., the vague suggestions of the outermost dark band are too indefinite to measure. The darker bands are generally thinner as is the case above (see PI. 10, figs. 6, 13). Apparently the growth was periodic and the periods when darker calcite was deposited were either shorter or less favourable to calcite secretion. This was almost certainly a seasonal effect in which case the animal commenced growing in season favourable to calcite secretion and survived two lean periods, possibly up to the onset of a third. If these periods were annual, which is the most obvious suggestion, then the dark bands probably represent winter and this cystoid reached its present size in three years. No example seen exceeds five years. There is apparently no information on longevity and growth rates in living crinoids but in echinoids corollary growth rates can be found. The information available in four species gives coronal diameters of 24-90 mm. in three years (Hyman 1955 : 588), whereas the cystoid is 14 by c. 8 by c. n mm. The mode of growth of DIAMETER OF TOPCOLUMNAL mm. 10 10 20 30 40 50 NUMBER OF PROXIMALS FIG. 6. Staurocystis quadrifasciatus (Pearce). Graphical representation of relationship between diameter of topmost proximal columnal and number of proximal columnals. GEOL. 13, 6. 25 3*4 THE BRITISH SILURIAN CYSTOIDS echinoids is quite different at least initially, and so not much importance can be attached to this information. The primary orifices of the thecae of the species described here all open across plate sutures and this is interpreted as being for mechanical convenience in growth. Enlargement of these orifices was achieved by a lack of secretion of calcite in the plates surrounding them and not by resorbtion. The slits of the pectinirhombs also grew initially by primary non-secretion of calcite. Once the slits become disjunct they could only increase in size by resorbtion. In Lovenicystis angelini the evidence suggests that no resorbtion occurred but apparently it did occur in Schizocystis armata. All the rhombs commenced as a single slit and grew by the addition of slits laterally. As the plates bearing the two half-rhombs became larger the slits became longer and beyond a certain length they became disjunct. In a fully developed rhomb all the slits are disjunct and the rims are well developed. In the present type of rhombs the spacing of slits is very constant as is their width. Internally both the dichopores and inter-dichopore spaces are approximately o-io mm. wide and in the majority of cases there is one dichopore and one inter-dichopore space in a width of 0-19-0-21 mm. Almost certainly this constancy of width is due to the combined thicknesses of cilated epithelium, and lengths of cilia, lining the dichopores. The fact that the inter-dichopore spaces in this type of rhomb are also of the same width indicates that these too were lined with ciliated epithelium. c. The Subvective System The ambulacra grew by the addition of flooring plates distally. In the Callo- cystitidae the ambulacra are wider, and the individual flooring plates thicker and wider, adorally. This suggests the last formed flooring plates were distal. There is no evidence for the addition of flooring plates adorally and it would be difficult to THECAL25r HEIGHTmm. 20 15 10 50 100 150 200 250 NUMBER OF BRACHIOLES FIG. 7. Staurocystis quadrifasciatus (Pearce). Graphical representation of relationship between thecal height and total number of brachioles. THE BRITISH SILURIAN CYSTOIDS 315 maintain the arrangement of brachioles in which the most adoral brachiole in each ambulacrum was branched off to the left. Each pair of flooring plates added gave rise to another brachiole and these apparently grew rapidly as in no case have distal brachioles been found significantly shorter than the adoral ones. The brachioles grew by the addition of brachiolars distally. Kesling (19616) found that the spacing of brachioles in Pseudocrinites gordoni Schuchert remained almost constant throughout growth and was a useful specific character. This can be confirmed for Pseudocrinites bifasciatus and Stauro- cystis quadrifasciatus (see Text-fig. 7) but the spacing of brachioles in Tetracystis oblongus and Apiocystites pentrematoides varies considerably. This variation is not necessarily related to growth. In species with an almost constant spacing of brachioles it is found that the spacing is slightly wider adorally, presumably due to the increased thickness of the ambulacral flooring plates. In species with a constant spacing of brachioles, and probably in the others as well, there was considerable readjustment of the flooring plates with respect to the thecal plates throughout growth. As growth proceeded any one flooring plate migrated adorally relative to the thecal plates. Thus, for example, in Pseudocrinities bi- fasciatus, BU. Hoi2 has a thecal height of 7 mm. and an estimated 12 brachioles in ambulacrum 2. Brachiole 212 lay on IL2 just above the common suture with 63. In SM. Ai2593 (thecal height 16 mm.) brachiole 212 lay on R2 just above the common suture with L/}. Thus if these brachioles are homologous, as is suggested here, brachiole 212 migrated across an infralateral and a lateral onto a radial plate between the two growth stages represented by these two specimens. VII. SYSTEMATIC DESCRIPTIONS Superfamily GLYPTOGYSTITIDA Bather 1899 Family CALLOGYSTITIDAE Bernard 1895 Subfamily STAUROGYSTINAE Jaekel 1899 DIAGNOSIS. A subfamily of Callocystitidae with 2-4 strongly protruding ambulacra developed on smooth bands on theca and not in shallow grooves. In the classification which will be adopted in the forthcoming Treatise on inverte- brate palaeontology (Kesling 1963) only two genera are referred to this subfamily, Staurocystis and Pseudocrinites. Tetracystis, which has hitherto been assigned to the Apiocystitinae, has very similar protruding ambulacra developed on smooth bands on the theca, and is here included in this subfamily on these grounds. Pruno- cystites fletcheri, the type species of Prunocystites, also has protruding ambulacra and so this species is included here as well. Genus STAUROCYSTIS Haeckel 1896 Lepadocrinus of authors. TYPE SPECIES. Pseudocrinites quadrifasciatus Pearce 1843. 3i6 THE BRITISH SILURIAN CYSTOIDS DIAGNOSIS. A genus of Staurocystinae with ovate theca and four ambulacra all reaching the stem, brachioles closely spaced, relatively long lateral food grooves and ambulacral cover plates frequently preserved. At present there is only one species. Haeckel (18960 : 166) introduced the name S. cruciata but this is a nomen nudum. S. oblongus (Forbes) Haeckel = Tetracystis oblongus (Forbes), see p. 330. Staurocystis quadrifasciatus (Pearce) (PI. i, figs. 1-8; PI. 2, figs. 1-3; PI. 10, figs. 1-5, 12; Text-figs. 4-13) Pseudocrinites quadrifasciatus Pearce Pseudocrinites quadrifasciatus Pearce Pseudocrinites quadrifasciatus Pearce Pseudocrinites quadricopuladigiti Garner : 160, pi. a, fig. 14. Pseudocrinites quadrifasciatus Pearce Pseudocrinites quadrifasciatus Pearce Pseudocrinus quadrifasciatus (Pearce) Pseudocrinites quadrifasciatus Pearce 'Pseudocrinites quadrifasciatus Pearce Pseudocrinites quadrifasciatus Pearce Pseudocrinites quadrifasciatus Pearce 160. 472. 265. Forbes : 498 ; pi. 13, figs. 1-13. Murchison : 216, text-fig. 37, fig. 2. Pictet : 298, pi. 99, fig. 13. Murchison : 222, 513, text-fig. 55, fig. 2. Bigsby : 26. Salter : 127. Bailey : 54, pi. 18, figs. 6a-c. 1876 Pseudocrinites quadrifasciatus Pearce ; Quenstedt : 680, pi. 113, figs. 77, j8a-d. 1891 Lepadocrinus quadrifasciatus (Pearce) Woods : 43. 1891 Pseudocrinus quadrifasciatus (Pearce) ; Carpenter 17, 10. i896a Staurocystis quadrifasciatus (Pearce) Haeckel : 134, pi. 3, figs. 1—3. 1897 Lepadocrinus quadrifasciatus (Pearce) ; Bather : 101, fig. 156. 1899 Staurocystis quadrifasciatus (Pearce) ; Jaekel : 286, text-fig. 61, pi. 14, figs, i, 2. 1900 Lepadocrinus quadrifasciatus (Pearce) ; Bather : 62, text-fig. 30. 1904 Staurocystis quadrifasciatus (Pearce) ; Schuchert : 227. 1938 Lepadocrinus Spencer, pi. 2, fig. b. 1943 Staurocystis quadrifasciatus (Pearce) ; Bassler & Moodey : 53, 191. 1964 Lepocrinites quadrifasciatus (Pearce) Castell : 100, pi. 24, fig. i. DIAGNOSIS. As for the genus. HOLOTYPE. Pearce's original specimen (Bristol Museum Ca.786o) was destroyed in an air-raid on 24th November, 1940. As the species is readily recognized and there is no point of taxonomy to be settled there is no need to designate a neotype. MATERIAL. About 140 specimens in BMNH., BU., DL., GSM., OUM., RSM., SM. HORIZON AND LOCALITY. Wenlock Limestone, Middle Silurian ; Dudley (Tivi- dale, Wren's Nest, Castle Hill) and Walsall. DESCRIPTION. The theca is subovate to subquadrate and provided with four ambulacra running down the sides from the apex to the insertion of the stem. The mouth, hydropore and gonopore are apical and the anus lateral. The theca was attached to the substrate by a stem provided with a root structure. The thecal plates are arranged according to the usual formula. The arrangement of the plates of the first four circlets is as in Text-fig. 8 and of the orals as in Text-fig. 9. The position of the ambulacra may vary slightly and four plates may surround the THE BRITISH SILURIAN CYSTOIDS 317 FIG. 8. Staurocystis quadrifasciatus (Pearce). Sketch of arrangement of thecal plates of the first four circlets. BB, basals ; ILL, infra-laterals ; LL, laterals ; RR, radials. Based on specimen SM. A. 28230. periproct in some cases. Otherwise there is little variation in the plating arrange- ment. The four basal plates unite abapically to form a circular cicatrix 4-4 mm. in diameter in SM. A28230. All the plates are ornamented with radiating and con- centric ridges which are often discontinuous and give the surface a malleated appear- ance. This ornament is only developed on the portions of the plates which do not under-lie the ambulacra. These latter portions are smooth bands on which traces of the outlines of the ambulacral flooring plates show in well preserved examples. Three disjunct pectinirhombs are developed in all examples and are placed in the three inter-ambulacral areas which do not contain the periproct. B2 : IL2 lies in Al FIG. 9 Staurocystis quadrifasciatus (Pearce). Camera lucida drawing of apical region of slightly weathered specimen, showing arrangement of orals and positions of ambulacra. A A, ambulacra ; G, gonopore ; H, hydropore ; I A, inter-ambulacrum ; OO, orals. The arrow points to the periproct. BU. Hoi. 3i8 THE BRITISH SILURIAN CYSTOIDS the anterior inter-ambulacrum, Li : R5 in the left and L4 : R3 in the right. In young specimens all or some of the pectinirhombs may be completely or partially conjunct. B2 : IL2. (BU. Ho6g) this rhomb is depressed in outline, angular and measures 5-0 mm. by 3-7 mm. externally. 62 has an open rim while IL2 has a closed rim. There are 25 dichopores in a width of 4-35 mm. and the longest reaches 3-2 mm. The slits are 0-07 wide and reach a length of 0-9 mm. on B2 but only half that length on IL2. BMNH. £.16353-54 are two sections through this rhomb from a single specimen. Both sections are transverse and approximately parallel to the suture. In £.16353 (PI- Io» figs- 2> 3) all the dichopores open at the surface apparently. There are 19 dichopores in a width of 3-9 mm. The outline of the section is a symmetrical triangle with the deepest dichopore at the centre. This dichopore is at least i -2 mm. deep. The dichopores and inter-dichopore spaces are both very evenly spaced and average about o-i mm. in width. The dichopore walls are very thin and average about 0-015 mm. By comparison with a complete theca it is possible to infer that the section was through the half-rhomb 62. This is the only section which could pass along an ambulacrum as this one does. £.16354 is m the same rhomb but cut nearer the suture apparently, as the deepest dichopore is deeper than in £.16353. There are only 18 dichopores in a width of 3-75 mm. They decrease in depth towards the left of the section in much the same way as in £.16353 DUt at a slower rate towards the right. The deepest dichopore is 1-75 mm. but the right marginal dichopore is still 1-6 mm. deep. The width of the dichopores and inter-dichopore spaces is the same however. There is an odd appear- ance of a membrane surrounding the entire set of dichopores internally which, under crossed nicols, coincides with a slight change in optical orientation of two calcite crystals. This could conceivably be due to the section being slightly oblique to the plate suture between 62 and IL2 and therefore passing through calcite in optical continuity with both plates. Similar "membranes" described as surrounding the hydrospires of some blastoids also coincide with optical discontinuities between calcite crystals filling the thecae (Joysey, personal communication, July, 1966). Li : R5. (BU. Hoi). This rhomb is depressed, angular and measures 3-0 mm. by 5-33 mm. externally. There are 24 or 25 dichoposes in a width of 4-8 mm. and the longest dichopore is 2-53 mm. long. The slits on R5 are regularly 0-07 mm. wide and the inter-dichopore spaces 0-125 mm. The longest slit in R5 reaches 1-07 mm. and that in Li o -67 mm. Although the adsutural portion of the rhomb rim is not com- pletely developed the thecal wall is more strongly raised adjacent to the slits on Li than on R5. The slits in Li are 0-09 mm. wide and separated by 0-107 mm- L4 : R3. (BU. Ho42). It is slightly immature depressed, asymmetrical and with an angular outline. It measures 5-13 mm. by 2-93 mm. externally and has 27 dicho- pores in a width of 5-0 mm., the longest reaching 2-37 mm. There are 9 dichopores in the demi-rhomb adjacent to ambulacrum 4, and 18 in the demi-rhomb adjacent to ambulacrum 2. The slits in half-rhomb R3 are 0-07 mm. wide and reach a maxi- mum length of 0-9 mm. adaxially. Those in £4 vary from 0-09-0-107 mm. in width and reach approximately 0-6 mm. in length. The slits in R3 are separated by 0-125 mm. and those in £4 by 0-09-0-107 mm. In this rhomb one or two marginal dicho- THE BRITISH SILURIAN CYSTOIDS 319 10 MFG FIGS. 10, ii. Staurocystis quadrifasciatus (Pearce). Fig. 10. Sketch of impression of ambulacral flooring plates on the thecal plates, near top of an ambulacrum. P, primary flooring plate ; S, secondary flooring plate. Based on specimen SM. A. 28230. Fig. n. Camera lucida drawing of part of an ambulacrum. Br, brachiole ; F, facet ; LFG, lateral food groove ; MFG, main food groove ; P and S as in Fig. 10. The arrows point to the mouth. Based on specimen SM. A.IH56. pores are conjunct and the adsutural rim is virtually undeveloped. However BMNH. 47908 shows a fully developed rhomb which is even more asymmetrical and has a complete closed rim on 1,4 and only an absutural rim on R3. BMNH. 57385 is particularly interesting as the demi-rhomb adjacent to ambulacrum 4 is only partly formed and apparently its growth was inhibited by the presence of the ambulacrum (see PI. 2, fig. 2). The mouth has not been directly observed but presumably lay apically at the confluence of the four ambulacra (see Text-fig. 4). It was surrounded by 7 orals arranged as in Text-fig. 9. The ambulacra were developed on broad bands on the theca. These bands are 4-0 mm. wide at the apex in SM. A2823O and taper very gradually towards the tips. The primary ambulacral flooring plates were very thin and narrow and the secondary flooring plates were very small in comparison and did not reach the ambulacral mid-line (see Text-fig. 10) . The brachiole facets are small kidney-shaped areas and are in contact with each other. Their spacing appears to have been almost constant throughout growth. The average of ten measurements, taken from five specimens (one near the tip and one near the mouth for each) was 5 brachioles in 3-59 mm. The variation was from 5 in 3-2 mm. to 5 in 3-8 mm. The constancy of this feature during growth is also indicated to a certain extent by Text-fig. 7. The main and lateral food grooves of the ambulacrum are provided with cover plates which apparently did not articulate with the flooring plates in life and formed a permanent roof over the food groove. They are very narrow plates, about 320 THE BRITISH SILURIAN CYSTOIDS PeH FIG. 12. Staurocystis quadrifasciatus (Pearce). Camera lucida drawing of area around periproct, showing four plates surrounding the periproct. AA, ambulacra ; An, anal pyramid ; Au, auxiliary circlet ; ILL, infralaterals ; LL, laterals ; Pe, periproct (com- prising anal pyramid and auxiliary circlet) ; Pe.H, periproct hood. Cf. PI. i, fig. 7 ; PL 2, fig. 3. BMNH. 48197. half the width shown in Text-fig, n, and are apparently unornamented. Between the lateral food grooves the primary flooring plates are ornamented with a large central tubercle and occasionally another smaller tubercle close to the main food groove (see PI. 2, fig. 3). The brachioles are composed of about 3 dozen regularly alternating brachiolars and are about 6-8 mm. long. They have a central straight food groove, about 0-2 mm. wide and provided with cover plates. There are three cover plates per brachiolar. In BMNH. £.7512 the brachioles are 07 mm. wide at the base and taper gradually throughout their length to 0-4 mm. near the tips. The periproct is surrounded by three thecal plates in most cases. These are, in increasing order of contribution to the border : IL5, IL4, L5. Occasionally L4 also reaches the periproct and contributes to a very small portion of the border. L5 frequently surrounds half the periproct and is provided with a special structure immediately adoral to the periproct here termed the periproct hood (Text-fig. 12). The periproct may be circular or bell-shaped and contains an anal pyramid reaching 3 mm. in diameter and composed of 5-7 anals. The anal pyramid is surrounded by a circlet of auxiliaries which may be open or closed and consists of about 9 plates. Both the anals and auxiliaries may have slightly raised margins. The gonopore and hydropore are apical and shared by Oi and 67, as is usual. The gonopore is a small circular opening 0-25 mm. in diameter in BU. Hoi. Adoral to it and slightly curved around it is the hydropore which is provided with a rim. BMNH. 57381, which has the oral area ground down, shows an embayment in the internal margin of the thecal wall just below the point of emergence of these orifices (see Text-fig. 13). The stem is very variable in length and relative proportions of the proximal and THE BRITISH SILURIAN CYSTO1DS 321 LI FIG. 13. Staurocystis quadrifasciatus (Pearce). Sketch of apical region ground down. Note embay ment above rhomb Li : R5 at position of gonopore and hydropore. A, am- bulacrum ; IA, inter-ambulacrum ; TC, thecal cavity. The arrow points to the periproct. BMNH. 57381. distal portions (see PI. 10, figs, i, 4, 5). The proximal portion tapers rapidly and the outer proximals have a smooth external flange. The distal portion is almost without any taper and the alternation of the distals is indistinct after the first few pairs. BMNH. £.7553, 38696 and BU. Hoig show distinct branched root structures. REMARKS. This is the commonest species from the Wenlock Limestone and was gregarious. A slab of Wenlock Limestone in the Geological Survey & Museum, London, has the remains of 8 specimens preserved on it and one in the Dudley Library has remains of 10 specimens. Another slab, in the Sedgwick Museum, has 2 specimens of this species and one of Tetracystis oblongus (Forbes) on it. It is interesting that despite the complexity of the food collecting apparatus in this genus, it is confined to the Wenlock Limestone, whereas Tetracystis and Pseudocrinites are more widely distributed and survived until the Devonian. Genus PSEUDOCRINITES Pearce 1843 TYPE SPECIES. Pseudocrinites bifasciatus Pearce 1843, by designation of Haeckel (18960 : 135). DIAGNOSIS. A genus of Staurocystinae with lens-like, biconvex theca ; with two ambulacra. Modified from Kesling (1963 : in). Ten species of Pseudocrinites have so far been described in addition to the one Schuchert described as Trimerocystis peculiaris, an abnormal individual which had three ambulacra instead of the usual two. Kesling (1961^ : 257) considered P. abnormalis Schuchert and T. peculiaris Schuchert to be synonymous with Pseudo- crinites gordoni Schuchert, and I believe that Pseudocrinites magnificus Forbes is probably a very large individual of P. bifasciatus Pearce. This reduces the number of species to 8 and it might be possible to reduce it even further if a comprehensive study of the American species were undertaken. One new species is described here 322 THE BRITISH SILURIAN CYSTOIDS based on two specimens from the Holcrof t collection (BU) . Both differ considerably from P. bifasciatus but quite closely resemble the American species P. perdewi Schuchert. REGIONAL DISTRIBUTION. England : Dudley, Walsall, Sedgley. United States : W. Virginia, Pennsylvania and Maryland. STRATIGRAPHIC RANGE. Wenlock Limestone (Middle Silurian) to Manlius Formation (Devonian, Gedinnian). Pseudocrinites bifasciatus Pearce (PL 2, figs. 4-8 ; PI. 3, figs. 2-6 ; PI. 10, figs. 7, 8 ; Text-figs. 14-19) i843a Pseudocrinites bifasciatus Pearce : 160. 18436 Pseudocrinites bifasciatus Pearce : 472. 18436 Pseudocrinites bifasciatus Pearce : 265. 1844 Pseudocrinites bicopuladigiti Garner : 160, pi. A, figs. 8-13. 1848 Pseudocrinites bifasciatus Pearce ; Forbes : 496, pi. n, figs. 1-7. 1848 Pseudocrinites magnificus Forbes : 497, pi. 12, figs. 1-4. 1854 Pseudocrinites bifasciatus Pearce ; Murchison : 216. 1854 Pseudocrinites magnificus Forbes ; Murchison : 216, text-fig. 37, fig. i. 1857 Pseudocrinus bifasciatus (Pearce) Pictet ; 298, pi. 99, fig. 12. ^57 Pseudocrinus magnificus (Forbes) Pictet : 298. 1867 Pseudocrinites bifasciatus Pearce ; Murchison : 222, 513. 1867 Pseudocrinites magnificus Forbes ; Murchison : 222, 513, text-fig. 55, fig. i. 1868 Pseudocrinites bicopuladigiti Garner ; Bigsby : 26. 1868 Pseudocrinites bifasciatus Pearce ; Bigsby : 26. 1868 Pseudocrinites magnificus Forbes ; Bigsby : 26. 1873 Pseudocrinites bifasciatus Pearce ; Salter : 127, 2 figs. 1876 Pseudocrinites bifasciatus Pearce ; Quenstedt : 678, pi. 113, figs. 74 (3 figs.), 75. 1891 Lepadocrinus bifasciatus (Pearce) Woods : 43. 1891 Pseudocrinus bifasciatus (Pearce) ; Carpenter : 7, 10. J895 Pseudocrinus bifasciatus (Pearce) ; Bernard : 206, text-fig. 73^. ?i8g6 Pseudocrinus, Haeckel : 401. i8g6a Pseudocrinites bifasciatus Pearce ; Haeckel : 135, pi. 3, figs. 25, 26. i8g6a Pseudocrinites magnificus Forbes ; Haeckel : 136. 1899 Pseudocrinites bifasciatus Pearce ; Jaekel : 285, pi. 14, fig. 3, text-fig. 60. 1899 Pseudocrinites magnificus Forbes ; Jaekel : 285, pi. 14, fig. 4. 1900 Pseudocrinus Bather : 62, text-fig. 29. 1904 Pseudocrinites bifasciatus Pearce ; Schuchert : 227, 229. 1904 Pseudocrinites magnificus Forbes ; Schuchert : 229. 1943 Pseudocrinites bifasciatus Pearce ; Bassler & Moodey : 53, 185. 1943 Pseudocrinites magnificus Forbes ; Bassler & Moodey : 53, 185. DIAGNOSIS. A Pseudocrinites with circular outline to theca ; very broad ambulacra reaching stem ; periproct composed of anal pyramid and complete circlet of auxiliaries ; thecal plates sculptured with coarse irregular ridges giving theca a malleated appearance. HOLOTYPE. Pearce's original specimen (Bristol Museum Ca.7862) was destroyed in an air-raid on 24th November, 1940. Again as this species is readily recognized and there is no point of taxonomy to be settled no neotype is designated. THE BRITISH SILURIAN CYSTOIDS 323 MATERIAL. About 35 specimens in BMNH., BU., DL., GSM., OUM., RSM., SM. HORIZON AND LOCALITY. Wenlock Limestone, Middle Silurian ; Dudley (Wren's Nest and Castle Hill). Before describing the species it is necessary to settle the synonymy between P. bifasciatus Pearce and P. magnificus Forbes. There is a constant number of brachioles per unit length in the ambulacra of Staurocystis quadrifasciatus and, according to Kesling (i96ia), in Pseudocrinites gordoni Schuchert. If this holds true for P. bifasciatus there will be a linear relation- ship between thecal diameter and the total number of brachioles. Text-fig. 14 clearly shows that this is so and that P. magnificus falls very close to the predicted value for a large P. bifasciatus. This is a particularly important point for in his original description Forbes (1848 : 497) stated that " the arms extend completely down the sides, and are remarkable for the great number of fingers [brachioles] on each, being no fewer than 34. [Forbes only counted the brachioles on one side of each ambulacrum.] They are much more closely packed, longer, slenderer, and apparently more flexible than in the last species. [P. bifasciatus} ". The number of brachioles was the main point distinguishing the two species and since there is in fact very little real difference they must be considered synonymous. DESCRIPTION. The theca is circular in outline, biconvex and lens-like. There are two ambulacra running around the periphery of the theca and usually reaching the THECAL 35 DIAMETER mm. 30 - 25 20 15 10 50 100 150 NUMBER OF BRACHIOLES FIG. 14. Graphical representation of relationship between diameter and total number of brachioles in Pseudocrinites bifasciatus Pearce and P. magnificus Forbes. The square dot represents the holotype of P. magnificus. 324 THE BRITISH SILURIAN CYSTOIDS stem. The mouth, gonopore and hydropore are apical and the anus lateral. The anus and rhomb I>4 : R3 were in one lateral face, the periproctal face, while rhombs B2 : IL2 and Li : R5 were in the other face. The gonopore and hydropore are therefore in this face which is referred to as the gonal face. The thecal plates of the first four circlets are arranged in the usual order (see Text-fig. 15) . The exact positions of the ambulacra may vary slightly but otherwise there is very little variation. As far as it is possible to determine the arrangement of the visible portions of the orals is as in Text-fig. 16 but they extended over a larger area beneath the ambulacra. The four basals combine abapically to form an oval cicatrix 5-5 mm. by 5-0 mm. in SM. Ai2593. All the plates are ornamented with FIG. 15. Pseudocrinites bifasciatus Pearce. Sketch of arrangement of thecal plates of first four circlets. BB, basals ; ILL, infra-laterals ; LL, laterals ; RR, radials. Based on specimen SM. A. 12593. irregular and discontinuous radiating and concentric ridges very much like those of Staurocystis. The portions of the thecal plates underlying the ambulacra bear the impressions of the flooring plates. These are much less closely packed than in Staurocystis (see Text-fig. 18). Three disjunct pectinirhombs are present, of these 62 : IL2 and Li : R5 are about the same size and L-4 : R3 is distinctly larger. 62 : IL2. (SM. Ai2593). This rhomb is depressed, angular and almost perfectly symmetrical. It measures 4-7 mm. by 3-8 mm. externally and has 22 dichopores in a width of 3 -9 mm., the longest reaching 3 -o mm. The slits in B2 are 0-07 mm. wide and separated by 0-125 mm. They reach a maximum length adaxially of 0-87 mm. Those in IL2 are 0-107 mm. wide and separated by 0-09 mm. and reach a maximum length of 0-62 mm. The closed rim on IL2 protrudes about 0-4 mm. from the general surface of the plate. Li : R5. (SM. Aioi92). This rhomb is depressed, more or less equal and sym- metrical. It measures 4-35 mm. by 3-65 mm. externally and has 20 dichopores in a width of 3-35 mm., the longest reaching approximately 3-0 mm. The slits in R5 are THE BRITISH SILURIAN CYSTOIDS 325 4' 5' FIG. 1 6. Pseudocrinites bifasciatus Pearce. Sketch of apical region, showing arrangement of the ambulacra, interpretation of the brachioles and the exposed portions of the oral plates. OO, orals. SM. A. 10192. 0-07 mm. wide and separated by 0-125 mm- The longest reaches i-i mm. Unfor- tunately the slits in Li are obscured by matrix but could not have exceeded 0-65 mm. in length. Li has a closed rim extending about 0-4 mm. from the level of the plate surface. L4 : RS. (SM. Ai2593). This is depressed, very asymmetrical and angular. It measures 6-0 mm. by 3-0 mm. approximately (it is slightly damaged). There are 32 dichopores in a width of 6-0 mm. and the longest reaches approximately 2-35 mm. The slits in R3 are 0-07 mm. wide and separated by 0-125 mm., the longest reaching 0-87 mm. Again unfortunately those in L4 are obscured by matrix but could not have exceeded 0-53 mm. in length. The mouth lay apically between facets 41 and i1 and shows as a narrow curved slit in BMNH. 47833. This specimen shows the internal view of the gonal face and apex. There is a slight groove internally in R5 below the gonopore and the hydro- pore. There are two broad ambulacra which coalesce at the apex and usually reach the stem. They are 5 mm. wide at the apex in SM. Ai2593 and taper gradually to about 3 mm. at their rounded tips. The primary flooring plates are not as thin as those of Staurocystis resulting in a more distant spacing of the brachioles. The ambulacra are more prominent than in Staurocystis and reach a height of 2 mm. in BMNH. 40189. The brachiole facets are kidney-shaped areas which are quite prominent and set at an angle of about 45 degrees to the line of the ambulacrum. The spacing of the facets appears to be constant throughout growth and a specific character. The average of ten measurements was 4 brachioles in 5-3 mm. with maximum and minimum of 4 in 6 -o mm. and 4 in 4 -9 mm. These measurements were taken from thecae varying in diameter from n mm. to 24 mm. The main and lateral food grooves of the ambulacra are provided with cover plates which are much broader than those of Staurocystis or Pseudocrinites pyriformis sp. nov. and are frequently 326 THE BRITISH SILURIAN CYSTOIDS ornamented with tubercles (see Text-fig. 19). Apparently these cover plates formed a permanent roof over the food grooves. Between lateral food grooves the primary flooring plates are ornamented with a slightly sinuous ridge. The brachioles taper very rapidly near the base so that although the facets are almost in contact adjacent brachioles are widely separated. They are composed of about 24 to 28 brachiolars and reach a length of about 6-6-5 mm. Basally they are i mm. wide and taper rapidly to start with, then gradually to about 0-5 mm. at the tips. The brachioles are provided with cover plates which are 3 or 4 times as numerous as the brachiolars. Most of these features are well preserved on BMNH. £.29068. MFG FIGS. 17-19 Pseudocrinites bifasciatus Pearce. Fig. 17. Sketch of side of an ambulacrum, showing primary and secondary flooring plates and a brachiole tapering rapidly at the base. Fig. 18. Sketch of impression of ambulacral flooring plates on the thecal plates. Fig. 19. Sketch of part of an ambulacrum. F, facet ; LFG, lateral food groove ; MFG, main food groove ; P, primary flooring plate ; S, secondary flooring plate. The arrows point to the mouth. Based on specimen SM. A. 10192. The periproct is surrounded by four plates. In increasing order of contribution to the border these are : IL5, IL4, L4, L5, the latter contributing about one-third the total. There is no periproct hood in L5. The periproct is covered by an anal pyramid, composed of 5 or 6 anals and up to 2-65 mm. in diameter, and an auxiliary circlet which is always complete and composed of 10 plates in BMNH. £.29068 and 40198. Both anals and auxiliaries are apparently unornamented. The gonopore and hydropore are apical and shared by Oi and Oy. The gonopore is a small circular opening 0-25 mm. in diameter in BU. Hoi2 which is a very im- mature specimen. The hydropore is oblong in this specimen and about twice the size of the gonopore. No gonals have been observed. As previously mentioned, there is a groove in R5, which corresponds in position to the embayment seen in Staurocystis, below the position of these two orifices. THE BRITISH SILURIAN CYSTOIDS 327 The stem is composed of the usual proximal and distal portions and, in BMNH. 40189, has a root structure. The outer proximals are provided with a smooth external flange and the distals are not clearly recognizable as large and small through- out most of the length. In BMNH. 40189 there are 14 outer proximals and about 21 distals of which the first six are in distinct pairs of large and small. The stem is about 27 mm. long. REMARKS. Schuchert (1904 : 228-229) grouped the American species of this genus into three groups : P. gordoni group. Thecal outline circular, ambulacra reaching the stem, auxiliary circlet complete. P. clarki group. Thecal outline elongate, ambulacra reaching the stem, auxiliary circlet incomplete. P. perdewi group. Thecal outline pyriform, ambulacra not reaching the stem, auxiliary circlet complete. Schuchert assigned P. bifasciatus and P. magnificus to the gordoni group and this seems to be correct. However, some doubt may be cast as to the distinctness of the first two groups as some examples of P. bifasciatus seem to have elongate thecae, and the nature of the auxiliary circlet is known to vary within a single species in other genera of Staurocystinae. The third group, that of P. perdewi, is apparently quite distinct from the other two and the next species described in this paper belongs in it. There may be some grounds for separating these two species from the others sub- generically but without a thorough re-examination of the American species this would seem a hasty action. Pseudocrinites pyriformis sp. nov. (PI. 3, figs, i, 7 ; PI. 4, figs. 1-5 ; PI. 5, fig. i ; Text-figs. 20, 21) DIAGNOSIS. A Pseudocrinites with pyriform outline to the theca ; broad am- bulacra extending only half way down the sides ; thecal plates showing coarse irregular sculpture adorally and fine irregular sculpture associated with fine radiating ridges aborally. HOLOTYPE. BU. Hoio. MATERIAL. In addition to the holotype, paratype BU. Hon. HORIZON and LOCALITY. Sedgley (Aymestry) Limestone, Upper Silurian ; Sedgley, Staffs. DESCRIPTION. The theca is pyriform in outline and bilaterally compressed with ambulacra in a broad band running across the top of the theca and extending about half way down the sides. The mouth, gonopore and hydropore are apical and the anus lateral. Rhombs 62 : IL2 and Li : R5 are in the gonal face, and the periproct and rhomb 1,4 : R$ in the periproctal face. The theca was provided with a stem. The arrangement of the thecal plates of the first four circlets is as in Text-fig. 20. The arrangement of the orals is unknown. Neither of the specimens available shows all the thecal plates, and those plates shown in dashed outline are not clearly visible. 328 THE BRITISH SILURIAN CYSTOIDS In BU. Hoi I the ambulacra extend a little further down the theca and encroach upon ILL2 and 5, but this seems to be the only difference. Where the plates are not covered by the ambulacra they have a very distinctive ornament. In the adoral half of the theca this is of very irregular ridges and pits accompanied by fine tubercles throughout. Aborally however the ornament is of much finer irregular ridges, which do not protrude as far from the general surface, fine tubercles and some fine radiating ridges. In the holotype the basal plates appear to be almost smooth but this is partly due to weathering. The basals unite abapically to form a cicatrix 7-3 mm. by 6-0 mm., but as this part of the theca is slightly crushed the cicatrix may have been circular originally. FIG. 20. Pseudocrinites pyriformis sp. nov. Sketch of arrangement of thecal plates of first four circlets. AA, ambulacra ; BB, basals ; ILL, infra-laterals ; LL, laterals ; RR, radials. The plates indicated by a dashed outline are not clearly visible. Based on specimen BU. Hoio. The usual three disjunct pectinirhombs are developed. Both specimens show Li : R5 clearly, the holotype shows half-rhomb R3, and neither show B2 : IL2 clearly although part of it is visible. B2 : IL2. Part of IL2 shows in the holotype. It has a closed rim and shows 18 slits in a width of 4-2 mm., the longest reaching i-o mm. The slits are 0-161 mm. wide and separated by 0-107 mm. which is unusually wide. Li : R5. (Holotype). This is very depressed, more or less angular and symmetri- cal. It measures 12-3 mm. by 5-7 mm. externally and has 43 dichopores in a width of 117 mm., the longest reaching 4-0 mm. The slits in R5 are 0-107 mm. wide and separated by o -125 mm. They reach a length of 2 -o mm. Those in Li are o -143 mm. wide and separated by 0-09 mm. They reach a maximum length adaxially of 0-95 mm. Due to the development of the closed rim around this half-rhomb these slits open on a surface sloping away from the suture at an angle approaching 45 degrees (see PI. 4, fig. 2). (BU. HOII). This rhomb is much the same and has about 40 dichopores in a width of 12 mm. approximately. The slits in R5 reach a length of 2-0 mm. while THE BRITISH SILURIAN CYSTOIDS 329 those in Li reach i-o mm. The slits are again narrower, and consequently more widely separated, in R5 than in Li. 14 : R3. (Holotype). This rhomb is only half visible and shows 45 dichopores in R3 in a width of 12-0 mm. but there were probably a few more which are now hidden by matrix. It is asymmetrical, there being 18 dichopores in the demi-rhomb adjacent to the periproct and therefore at least 27 in the other demi-rhomb. The slits in R3 are 0-107 mm. wide and separated by 0-125 mm- The longest is 2-33 mm. long. The mouth has not been directly observed but presumably lay between facets I1 and 41 as in P. bifasciatus. The ambulacra are developed on broad bands which taper relatively rapidly. They are 7 mm. wide apically and taper to about half that just before their rounded tips. Apically the ambulacral flooring plates are 3-5 mm. high in the holotype and 4-0 mm. in Hon. There are 14 facets in ambulacrum 5 and 16 in ambulacrum 2 of Hoio, the corresponding figures for Hon being 20 and 24. The arrangement of the flooring plates is much the same as in P. bifasciatus but the brachioles are much more widely spaced. The average of four measurements was FIG. 21. Pseudocrinites pyriformis sp. nov. Camera lucida drawing of apical region. H, hydropore ; M, approximate position of mouth. Cf. Text-figs. 4, 16. Based on specimen BU. Hoio. 4 facets in 10-4 mm. with maximum and minimum distances of 10-0 and n-o mm. These are almost twice the distances for P. bifasciatus. The facets are large kidney- shaped areas which are almost planar. The main and lateral food grooves are provided with highly ornamented cover plates which are relatively long and narrow. They have the same general proportions as those of Staurocystis but differ in being much larger and in being covered with strongly raised ridges. Between the lateral food grooves the primary flooring plates are highly ornamented with raised tubercles, and both types of flooring plates are ornamented with ridges laterally. The main food grooves are provided with a narrow ledge on which the cover plates rested in life. The lateral food grooves were relatively short. The brachioles are composed of alternating brachiolars ornamented with tubercles and ridges. The penultimate GEOL. 13, 6. 26 330 THE BRITISH SILURIAN CYSTOIDS brachiole in ambulacrum 5 of Hon is 9-35 mm. long and has 22 brachiolars in this distance. It is incomplete however. The terminal brachiole apparently exceeds 12 mm. in length and this too is incomplete. The penultimate one tapers from 1-45 mm. wide at the base to 0-65 mm. where it is broken off. The cover plates are apparently ridged and about 2-3 per brachiolar. The periproct is probably surrounded by four plates but cannot be seen completely in either specimen. Parts of the anal pyramid and auxiliary circlet can be made out in the holotype but the total number of plates in each is unknown. In Hon the auxiliaries were ornamented with ridges. There is no periproct hood in L5. The gonopore and hydropore lay apically presumably across the common suture of Oi and Oy. The gonopore is small, circular and provided with a pyramid of gonals, of which 2 can be seen in the holotype. It is 0-35 mm. wide in the holotype and 0-65 mm. wide in Hon. The hydropore is curved around the gonopore adapically and provided with a conspicuous rim bearing ridges externally. It is 1-85 mm. wide in the holotype and 3-45 mm. wide in Hon. 21-5 mm. of the proximal stem is preserved in the holotype. It tapers from 7-0 mm. to 4-5 mm. in this distance and there are 26 outer proximals each provided with an external flange that has tubercles developed on the less weathered portions. REMARKS. This species is quite distinct from the preceding but closely resembles P. perdewi Schuchert. It differs from the latter species in the brachiole and dicho- pore spacing, in having more prominent ambulacra and less prominent rhombs. The holotype has 45 dichopores in L4 : R3 whereas a specimen of P. perdewi of about the same dimensions would have about 70 dichopores in the same rhomb. The ornamentation of the plates of the two species is also quite different. As a general rule the British species of Staurocystinae seem to have more prominent ambulacra and less well developed rhombs than their American counterparts. Genus TETRACYSTIS Schuchert 1904 TYPE SPECIES. Teiracystis chrysalis Schuchert 1904. DIAGNOSIS. A genus of Staurocystinae with four protruding ambulacra all reaching the stem ; relatively widely spaced brachioles ; wide main food groove and short lateral food grooves ; some secondary ambulacral flooring plates reaching ambulacral mid-suture ; ambulacral cover plates rarely preserved. Hitherto the genus Tetracystis has been assigned to the subfamily Apiocystitinae. Kesling (1963 : 108) lists among the characters of the Staurocystinae that the ambulacra are strongly protuberant. In both genera attributed by Kesling to the Staurocystinae this is due to the ambulacra being developed on smooth bands on the thecal plates and not in shallow grooves as in the Apiocystitinae. In both these respects Tetracystis resembles Staurocystis and Pseudocrinites and is therefore included in the Staurocystinae. The specimens described by Forbes as Pseudocrinites oblongus apparently belong to this genus. They have four ambulacra three pectinirhombs and four plates bordering the periproct. In Regnell's scheme for the Apiocystitinae (Regnell THE BRITISH SILURIAN CYSTOIDS 331 1945 : 91) there are only three genera which fall into this category. These are Lovenicystis Regnell 1945, Jaekelocystis Schuchert 1903 and Tetracystis Schuchert 1904. The present specimens differ from Lovenicystis in having all the ambulacra reaching the column and from Jaekelocystis in having angular pectinirhombs in which one half-rhomb is not reduced to a spout-like pore. They agree with Tetracystis in these respects and also in the nature of the ambulacra which has caused their removal to the Staurocystinae. There are only three other species known at present, T. chrysalis Schuchert, T. bifarius Amsden and T. fenestratus (Troost). The British species appears to differ from all three in having more prominent ambulacra. REGIONAL DISTRIBUTION. England : Dudley ; United States : Tennessee and W. Virginia. STRATIGRAPHICAL RANGE. Middle Silurian to Basal Devonian. Tetracystis oblongus (Forbes) 1848 (PI. 5, figs. 2-7, 11-13 ', PI- I0> ngs- 6, 13 ; Text-figs. 22-25) 1848 Pseudocrinites oblongus Forbes : 499, pi. 14, figs. 1-14. 1854 Pseudocrinites oblongus Forbes ; Murchison : 217. 1857 Pseudocrinus oblongus (Forbes) Pictet : 298. 1867 Pseudocrinites oblongus Forbes ; Murchison : 222, 513. 1868 Pseudocrinites oblongus Forbes ; Bigsby : 26. 1873 Pseudocrinites oblongus Forbes ; Salter : 127. 1876 Pseudocrinites oblongus Forbes ; Quenstedt, pi. 113, fig. 79, (3 figs.). 1891 Pseudocrinus oblongus (Forbes) ; Carpenter : 10. 1896^ Staurocystis oblongus (Forbes) Haeckel : 134. 1899 Apiocystites oblongus (Forbes) Jaekel : 282. 1900 Lepadocrinus oblongus (Forbes) Bather : 62. 1943 Lepocrinites oblongus (Forbes) Bassler & Moodey : 52, 173. DIAGNOSIS. A Tetracystis with broad, prominent ambulacra ; up to 30 brachioles per ambulacrum and up to 30 dichopores per rhomb ; periproct relatively depressed. LECTOTYPE. Forbes did not designate a type but the specimens he examined were in the Fletcher and Gray collections. Among the specimens in the Fletcher collection is the original of Forbes' pi. 14, fig. I (SM. Aioi4o), here selected as lectotype. MATERIAL. In addition to the lectotype, 9 specimens in BMNH. GSM., SM. HORIZON AND LOCALITY. Wenlock Limestone, Middle Silurian ; Dudley, Worcs. DESCRIPTION. The theca is subquadrate with four ambulacra reaching the stem. The mouth, gonopore and hydropore are apical and the anus lateral. No complete stem has been seen. The arrangement of the thecal plates of the first four circlets is shown in Text-fig. 22. Very few specimens are known and there is little variation among them. The four basal plates unite abapically to form a circular cicatrix 4 mm. in diameter in the lectotype. The thecal plates are ornamented with irregular radiating and concentric ridges which are often discontinuous and strongly resemble the ornamentation in Staurocystis and Pseudocrinites bifasciatus. Beneath the ambulacra the traces of the 332 THE BRITISH SILURIAN CYSTOIDS FIG. 22. Tetracystis oblongus (Forbes). Sketch of arrangement of thecal plates of first four circlets. BB, basals ; ILL, infra-laterals ; LL, laterals ; RR, radials. Based on specimen SM. A. 10140. ambulacral flooring plates can be made out. As far as can be determined from BMNH. 57411 there were 7 orals arranged in much the same way as in Staurocystis. The usual three disjunct pectinirhombs are present. They are all angular and depressed in outline. 62 : IL2 (Lectotype). This is 3-15 mm. by 2-15 mm. and has 19 dichopores in a width of 2-9 mm. The longest dichopore is 1-8 mm. The slits in IL2 are obscured by matrix, those in B2 are up to 0-054 mm- wide and separated by 0-125 mm. IL2 is provided with a closed rim. FIGS. 23, 24. Tetracystis oblongus (Forbes). Fig. 23. Sketch of part of an abulacrum. Fig. 24. Sketch of impression of ambulacral flooring plates on the thecal plates. P, primary flooring plate ; S, secondary flooring plate. The arrows point to the mouth. Based on specimen SM. A. 10140. THE BRITISH SILURIAN CYSTOIDS 333 Li : R5 (Lectotype). This is 375 mm. by 275 mm. externally and has 23 dicho- pores in a width of 3 -6 mm. and the longest dichopore is 2 -27 mm. The slits in both half-rhombs are 0-07 mm. wide, and separated by 0-125 mm. Those of Li reach 0-54 mm. in length while those in R5 reach i-o mm. L4 : R3 (Lectotype). The rhomb measures 4-8 mm. by 3-0 mm. and is asym- metrical, possibly due to the proximity of ambulacrum 4. There are 28 dichopores, all of which are disjunct, in a width of 4-45 mm. and the longest is 2-4 mm. The slits in L4 are 0-07 mm. wide and separated by 0-107 mm. while those the longest reaching 0-67 mm. Those in R3 are only 0-055-0-06 mm. wide and separated by o-r mm. The longest is 0-95 mm. long. The mouth was apical and slit-like (BMNH. 57411) and apparently surrounded by 7 orals. The four ambulacra meet in the oral area and the first formed facets are on the left as in other species of Staurocystinae (see Text-fig. 25). The four am- bulacra are developed on relatively narrow bands which, contrary to Forbes' state- ment, hardly taper at all. They are 2-5 mm. wide apically and still 2-0 mm. wide near their tips. They are much more conspicuous in this species than in the American species. The arrangement of the ambulacral flooring plates is shown in Text-fig. 24. The secondary flooring plates are much larger relatively than in Staurocystis or Pseudocrinites and quite frequently reach the ambulacral mid line. The brachiole facets are rather inconspicuous but slightly raised and the lateral food grooves are relatively short. Both the lateral and main food grooves were covered with am- bulacral cover plates which were subquadrate and unornamented. Apparently the primary flooring plates were also unornamented between the lateral food grooves. The spacing of the brachiole facets varies within much wider limits than in the pre- vious two genera. The average of 8 measurements was 4 brachioles in 5-73 mm. but varied from 4 in 4-67 mm. to 4 in 7-0 mm. This may vary within a single speci- men and is not directly associated with growth as the largest theca had an inter- mediate value. The brachioles are narrow, being 0-7 mm. wide basally in SM. Ai26o3, and they taper gradually. No complete brachiole has been seen but 20 brachiolars are preserved in some brachioles on SM. Ai26o3. FIG. 25. Tetracystis oblongus (Forbes). Sketch of apical region of theca. M, approximate position of mouth ; OO, position of exposed portions of some of the oral plates. Based on specimen SM. A. 10140. 334 THE BRITISH SILURIAN CYSTOIDS The periproct is surrounded by four plates, which are in increasing order of contri- bution to the border; 14, IL4, IL5, £,5. L4 only just reaches the periproct as in some examples of Staurocystis. There is no periproct hood in L5- The periproct consists of an anal pyramid of 5 or 6 anals surrounded by a closed circlet of auxiliaries. These are unequally developed so the border of the anal pyramid is narrower adapic- ally as shown in Text fig. 22. There were 9 or 10 auxiliaries in the lectotype and probably more in BMNH. £.29075. Both types of plate may have been slightly ornamented with irregular ridges. The gonopore and hydropore are apical, presumably developed across the suture between OOi and 7. Both are inconspicuous and it is impossible to confirm the multiple perforations of the gonopore described by Schuchert (1904). No complete stem has been seen so it is impossible to confirm that a root structure was developed but this seems likely. SM. Ai26o3 has 10 outer proximals preserved. All have a smooth external flange. Genus PRUNOCYSTITES Forbes 1848 TYPE SPECIES. Prunocystites fletcheri Forbes 1848. DIAGNOSIS. A genus of Staurocystinae with (probably 4) short ambulacra confined to apex of theca ; with few relatively long and slender brachioles. This genus was erected by Forbes to contain the single species P. fletcheri, of which at the time only the holotype (SM. Aioi87) was available. The genus was therefore based on this single specimen which is partly buried in matrix and hence, as Forbes stated (1848 : 503), the original description was imperfect. The holotype, and all subsequently found specimens, are small and it is difficult to assess whether they are immature or adult and, if they are immature, which characters are the result of immaturity. Only one of the specimens available is isolated from the matrix and this has several plates obscured. If all the available specimens are conspecific it is possible to build a composite picture of the characters of the genus, but it must be emphasized that this picture is composite. Forbes (1848 : 503-504) did not discuss the affinities of P. fletcheri but later (1848 : 526) he placed it in a position intermediate between Echinoencrinites s.l. and Pseudocrinites s.l. which roughly correspond to the Echinoencrinitidae and Stauro- cystinae. Jaekel (1899) considered P. fletcheri and " Echinoencrinites " baccatus Forbes to be congeneric and therefore assigned the latter to Prunocystites, and both species to his family Scoliocystidae (= Echinoencrinitidae). This course has been accepted subsequently. " E." baccatus is commoner than P. fletcheri and its characters are better known. The concept of Prunocystites has therefore been partly based on this species which undoubtedly belongs in the Echinoencrinitidae s.l., where it was placed by Forbes. This accounts for Prunocystites having been considered as an echinoencrinitid. P. fletcheri differs from all other British Callocystitids in having ambulacra confined to the apex of the theca but this could be the result of the small size (? and immaturity) of the thecae known. If the specimen figured in PL 9, fig. 10 is really P. fletcheri it THE BRITISH SILURIAN CYSTOIDS 335 would indicate that this is a feature of immaturity. The staurocystinid type of ambulacrum can be confirmed in the holotype of P. fletcheri if it is wetted with alcohol and examined under a fairly high magnification. On this evidence the genus Prunocystites is transferred to the Staurocystinae. There emerges the possi- bility that all or some of the thecae grouped under P. fletcheri are young specimens of other staurocystinid species. They are not young of Pseudocrinites bifasciatus which are well known and easily recognized. The long, slender and relatively widely spaced brachioles would seem to preclude Staurocystis quadrifasciatus and Tetracystis oblongus but young specimens of these species are apparently unknown. In view of the general uncertainty about Prunocystites it is suggested that the name be retained for the specimens described below until such time as it may become possible to settle the characters and affinities of the genus. This leaves " E." baccatus Forbes without a genus and, as there is no other name available, the genus Glansicystis nov. (glans = acorn) is proposed for it (see p. 344). Prunocystites fletcheri Forbes (PI. 3, figs. 8, 9 ; PI. 8, fig. 4 ; PI. 9, fig. n, ?io ; Text-figs. 26-29) 1848 Prunocystites fletcheri Forbes : 504, pi. 16, figs. 1-4. 1854 Prunocystites fletcheri Forbes ; Murchison : 217, text-fig. 37, fig. 3. 1857 Prunocystites fletcheri Forbes ; Pictet : 299, pi. 99, fig. 16 (2 figs.). 1867 Prunocystites fletcheri Forbes ; Murchison : 222, text-fig. 55, fig. 3. 1868 Prunocystites fletcheri Forbes ; Bigsby : 26. 1873 Prunocystites fletcheri Forbes ; Salter : 127. 1876 Prunocystites fletscheri (sic) Forbes ; Quenstedt : 685, pi. 113, fig. 90. 1891 Prunocystites fletcheri Forbes ; Woods : 43. 1899 Prunocystites fletscheri (sic) Forbes ; Jaekel : 282. 1900 Prunocystites fletcheri Forbes ; Bather : 60, text-fig. 24. 1943 Prunocystites fletcheri Forbes ; Bassler & Moodey : 53, 185. DIAGNOSIS. As for the genus. HOLOTYPE. SM. Aioi87. MATERIAL. BMNH. 40207, 57395-96 ; DL. 307 ; SM. A.ioi87-88, A.I2579 ; ?BU. IVa.25. HORIZON AND LOCALITY. Wenlock Limestone, Middle Silurian ; Dudley, Worcs. DESCRIPTION. The theca is ovate and very small, being only 5 mm. high in the holotype. The brachioles are developed apically on short ambulacra that protrude from the theca. The mouth was presumably apical, the anus lateral and high up on the theca. The gonopore and hydropore are unknown. In all but one theca there are only two rhombs, B2 : IL2 being undeveloped. The proximal stem is nearly as wide as the theca in some cases, tapers rapidly and is very short. The distal stem in SM. Aioi88 is very long and slender. The thecal plates are arranged in a definite order and although it has not been possible to confirm the complete arrangement in any one specimen a composite pic- ture may be built up from the specimens available. This composite picture suggests 336 THE BRITISH SILURIAN CYSTOIDS 26 U^l 27 FIGS. 26, 27. Prunocystites fletcheri Forbes. Fig. 26. Camera lucida drawing of thecal plates visible in holotype. SM. A. 10187. Fig. 27. Camera lucida drawing of thecal plates visible in specimen SM. A. 10188. A, ambulacrum ; BB, basals ; ILL, infra- laterals ; LL, laterals ; RR, radials. Note there is no rhomb B2 : ILa. that the arrangement was the same as in other species of Staurocystinae but a certain amount of variation occurs. The arrangements of the plates visible in three speci- mens are shown in Text figs. 26-29. The basals unite abapically to form a cicatrix, the ILL and LL circlets are closed but the exact arrangement of the RR and OO is incompletely known. SM. Ai2579 ano^ DL. 307 show thecal plates ornamented with fine tubercles and larger tubercles at the umbones of the ILL plates. L5 interrupts the RR. circlet in DL. 307. Two pectinirhombs are present in all but BU. IVa 25 which is quite probably not a specimen of this species. The shape and degree of separation of the half-rhombs vary considerably and are difficult to interpret. BMNH. 40207 shows half-rhomb L4 well and although it is very small it is completely surrounded by a strong raised rim This suggests that the rhomb is fully developed and by inference that the theca was fully grown, although only 5 mm. high. SM. Ai257Q shows the same rhomb. It is nearly twice as wide (1-7 mm. as opposed to i-o mm.) completely conjunct and some- what angular. This suggests the rhomb was incompletely developed and therefore that the theca was immature. Li : R5 is distinctly disjunct in BMNH. 40207 but apparently conjunct in the holotype. No rhomb has been seen with more than 7 dichopores in it. In SM. Ai257Q the slits were 0-07 mm. wide and separated by 0-125 mm- m half-rhomb R3. The peristomial area is not clearly visible in any specimen but there were definite ambulacra which protruded from the theca. These are best seen in SM. Ai2579 and DL. 307 where they are confined to the apical region and each ambulacrum has 4 to 6 brachioles. Three ambulacra can be seen and the flooring plates can be distinguished. The brachioles are long and slender, reaching 12 mm. in DL. 307 and tapering from 0-53 mm. wide at the base to 047 mm. at the tips. There are at least 44 brachiolars in the most complete brachioles of DL. 307 and the holotype and THE BRITISH SILURIAN CYSTOIDS 337 \L5 Bl FIGS. 28, 29. Prunocystites fletcheri Forbes. Fig. 28. Camera lucida drawing of posterior lateral view. Fig. 29. Camera lucida drawing of right lateral view. Note the pro- truding ambulacra confined to the apex of the theca. A, ambulacrum ; BB, basals ; ILL, infra-laterals ; LL, laterals ; Pe, periproct ; RR, radials. SM. A. 12579. they are individually rather tall and slender. The cover plates of the brachioles are small and pentagonal. There are approximately three to every brachiolar. The periproct is close to the ambulacra and high up on the theca. This may be a feature of immaturity. In SM. Ai2579 and DL. 307 the periproct is surrounded by four plates : IL/j., IL5, 14 and I>5 but in BMNH. 57396 IL4 does not reach the periproct. In the latter specimen there is an anal pyramid of 6 anals and a complete outer circlet of auxiliaries, probably 10 in number. In SM. Ai2579 the details are not clear but there appears to be only an anal pyramid. The gonopore and hydropore have not been detected. The stem is composed of proximal and distal portions. In SM. Aioi88, which has the most complete stem known, the proximal portion is composed 12 columnals, 6 outer and 6 inner. It is 1-6 mm. long and tapers from 1-25 mm. to 0-6 mm. in this distance. The distal portion is composed of an unknown number of tall thin colum- nals, tapers from 0-6 mm. to 0-5 mm. and the preserved portion is 28 mm. long. The holotype has 16 proximals, the outer proximals being provided with a smooth external flange aborally. BMNH. 57395, the largest theca that can definitely be assigned to this species, has 20 proximals. Subfamily APIOGYSTITINAE Jaekel 1899 DIAGNOSIS. A subfamily of Callocystitidae with 4 or 5 ambulacra developed in shallow grooves in thecal plates, not strongly protruding. Genus APIOCYSTITES Forbes 1848 TYPE SPECIES. Apiocystites pentrematoides Forbes 1848. DIAGNOSIS. A genus of Apiocystitinae with open circlet of radials ; 3 or 4 plates surrounding periproct ; pectinirhombs. four ambulacra all reaching the column ; three small 338 THE BRITISH SILURIAN CYSTOIDS Both Regnell (1945 : 92) and Kesling (1963 : in) state the following : 1. LL and RR circlets closed. 2. Periproct surrounded by three plates (not L4 according to Kesling). 3. Regnell adds that ILL-4 and 5 are not elongated. Forbes (1848) described this genus and attributed to it the single species A. pentre- matoides which becomes type species by monotypy. In this species the RR circlet is open, L-5 being inserted between R4 and R5, the periproct is surrounded by four plates, namely ILL4 and 5, and LL4 and 5, and finally ILL4 and 5 are elongated in an ad-abapical direction. The genus Apiocystites has been considered to be synonymous with Lepocrinites Conrad, 1840 (=Lepadocrinus of authors). This latter genus is said to have LL and RR circlets closed and the periproct surrounded by three plates (not L4). It also differs from Apiocystites in having ambulacra which do not reach the stem, and in having larger, angular pectinirhombs. In Regnell's scheme for the Apiocystitinae (1945 : 91) there are only two genera which have an open RR circlet and four plates around the periproct. These are Tetracystis Schuchert 1904 and Jaekelocystis Schuchert 1903. It is also necessary to compare the present genus with Lovenicystis Regnell 1945, the type species of which has been described as a species of Apiocystites. Tetracystis differs in having protruding ambulacra which are developed on smooth bands on the theca and are characteristic of the Staurocystinae. It also has large angular pectinirhombs with a relatively large number of dichopores. Jaekelocystis differs in having specialized pectinirhombs in which one half is reduced to a small pore 0-5 mm. in diameter. Lovenicystis differs in having irregular ambulacra which do not reach the stem in all cases, (which do and which do not varies with the indivi- dual), in having larger angular rhombs and an incomplete circlet of auxiliaries in the periproct. This latter feature is found rarely in A. pentrematoides. Apparently despite the changes in the diagnosis of the genus Apiocystites that have become necessary after re-examination of the type species, it has no junior synonyms and remains a distinct genus. Bassler & Moodey (1943) accepted four species of Apiocystites including the type species of Lovenicystis. Of the remaining three species it has not been possible to trace a description of A. anna Safford and consequently it is impossible to comment on it. A. elegans Hall 1851, judging from the figures, is closely related but apparently differs in having three plates around the periproct. This is not a very reliable character (it varies in Staurocystis quadri- fasciatus] and so the diagnosis has been extended to include this species. Apiocystites pentrematoides Forbes (PL 5, figs. 8-10 ; PL 6, figs, i-io ; PL 7, fig. i ; PL 10, figs. 9, 14 ; Text-figs, 30-35) 1848 Apiocystites pentrematoides Forbes : 503, pi. 15, figs. 1-9. 1854 Apiocystites pentrematoides Forbes ; Murchison : 217, text-fig. 37, fig. 4. 1857 Apiocystites pentrematoides Forbes ; Pictet : 298, pi. 99, fig. i^a-c. 1867 Apiocystites pentrematoides Forbes ; Murchison : 222, text-fig. 55, fig. 4. 1868 Apiocystites pentrematoides Forbes ; Bigsby : 24. 1873 Apiocystites pentrematoides Forbes ; Salter : 127 (i fig.)- THE BRITISH SILURIAN CYSTOIDS 339 1876 Apiocystites pentrematoides Forbes ; Quenstedt : 680, pi. 113, fig. 80. 1891 Apiocystis pentremitoides (sic) (Forbes) Carpenter : 9, 10 (5 figs.), 81. 1899 Apiocystites pentrematoides Forbes ; Jaekel : 282, pi. 15, fig. 2. 1904 Apiocystites pentrematoides Forbes ; Schuchert : 211. 1943 Apiocystites pentrematoides Forbes ; Bassler & Moodey : 52, 130. DIAGNOSIS. An Apiocystites with four plates surrounding periproct. HOLOTYPE. BMNH. 40187. MATERIAL. More than 45 specimens in BMNH., BU., GSM., OUM., ?RSM., SM. HORIZON AND LOCALITY. Wenlock Limestone, Middle Silurian ; Dudley, Worcs. DESCRIPTION. The theca is ovate elongate and somewhat rounded in cross- section. The mouth, hydropore and gonopore are apical and the anus lateral. The stem was provided with a root structure. The ambulacra are inconspicuous as they are developed in shallow grooves and are flush with the surface of the thecal plates. The thecal plates are arranged in the usual manner and the arrangement of the first four circlets is as in Text-fig. 30. The arrangement of the orals and adjacent ambulacral flooring plates is shown in Text-fig. 32. There is some variation in the exact position of the ambulacra but little variation in the thecal plating arrangement. FIG. 30. Apiocystites pentrematoides Forbes. Sketch of the arrangement of thecal plates of first four circlets. BB, basals ; ILL, infra-laterals ; LL, laterals ; RR, radials. Based on specimen SM. A. 12601. The basals combine abapically to form a circular cicatrix about 4 mm. in diameter in SM. Ai26oi. Unlike the majority of species of Callocystitidae described the thecal plates are ornamented with fine tubercles and not irregular ridges. This ornamenta- tion is not developed in the shallow curved grooves in which the ambulacra lie. These grooves are smooth and show faint traces of the outlines of the overlying ambulacral flooring plates. The usual three disjunct pectinirhombs are present. When immature these rhombs tend to be subangular in outline but more often the half-rhombs are distinctly rounded. B2 : IL2 SM. Ai259Q. Th*3 *s a compressed rhomb, with distinctly separated halves and measures 3-0 mm. by 2-35 mm. externally. It has 9 dichopores in a 34° THE BRITISH SILURIAN CYSTOIDS FIG. 31. Apiocystites pentrematoides Forbes. Camera lucida drawing of apical region, showing arrangement of food grooves, facets and visible portions of oral plates. Br, brachiole ; F, facet ; FG, food groove ; G, gonopore ; H, hydropore ; OO, orals ; P, primary flooring plate ; S, secondary flooring plate. SM. A. 12601. width of 1-93 mm., the longest reaching 2-47 mm. The slits in B2 are 0-095 mm. wide, separated by o-io mm. and reach a length of 075 mm. Those in IL2 are obscured by matrix but could not have exceeded 0-5 mm. in length. Li : R5 SM. Ai26oi. This is a slightly immature, angular, rhomb and it is depressed. It measures 2-0 mm. by 2-2 mm. externally and has 9 dichopores in a width of 1-9 mm., the longest reaching 1-47 mm. The slits in Li are 0-14 mm. wide and separated by 0-07 mm. while those in R5 are 0-09 mm. wide and separated by 0-125 mm- The slits in Li reach a length of 0-47 mm. while those in RS reach 0-6 mm. L4 : R3 SM. Ai26oi. This is an equal symmetrical rhomb measuring 2-2 mm. by 2-2 mm. It has 9 dichopores in a width of 1-8 mm., the longest reaching 2-63 mm. The slits in 14 are slightly wider than those in R3 and also shorter. The mouth was elongate, slightly curved and placed apically at the union of the ambulacra. It was surrounded by the usual 7 orals plates arranged as in Text-fig. 32. The ambulacra are relatively narrow and, in SM. Ai26oi, taper gradually from 1-6 mm. wide apically to about i-o mm. at their tips. They broaden considerably apically forming an area 4-2 mm. by 3-3 mm. around the mouth in this specimen. They are developed in shallow grooves and constructed in the same way as other callocystitid ambulacra. They differ from the other species described in having the secondary flooring plates about half the size of the primaries and regularly reaching the ambulacral mid-suture (see Text-fig. 34). The main food groove is relatively narrow and gives off lateral food grooves at a very oblique angle. This adds to the different appearance of this type of ambulacrum (see Text-fig. 33). The flooring plates are apparently unornamented between the lateral food grooves. Both types THE BRITISH SILURIAN CYSTOIDS 341 FIG. 32. Apiocystites pentrematoides Forbes. Camera lucida drawing of apical region of a weathered specimen, showing arrangement of oral plates and adjacent ambulacral flooring plates. A, ambulacrum ; F, position of facets ; G, gonopore ; H, hydropore ; IA, inter-ambulacrum ; M, mouth ; OO, orals. The arrow points to the periproct. BU., unregistered. of food groove are provided with small cover plates which are also unornamented. The brachiole facets are small circular pits in the flooring plates. The brachioles are slender and relatively widely spaced. This last feature varies considerably. The average of ten measurements was 4 brachioles in 472 mm. but varied from 4 in 373 mm. to 4 in 5-6 mm. The brachioles may reach 27 mm. in length and are only 0-3 mm. wide at the base. They are composed of approximately 30 brachiolars. The periproct is surrounded by four plates which are : ILL4 and 5 and LL,4 and 5. LS contributes approximately a third of the border, and the other three contributing about equally to the remaining two- thirds. There is no periproct hood in L5. The periproct is almost level with the general surface of the theca and I -6 mm. in diameter in SM. Ai26oi. Within the narrow but complete circlet of auxiliaries in this speci- men there is a slightly elevated anal pyramid I -3 mm. in diameter and composed of 6 anals. In some examples the auxiliary circlet may be incomplete. This is the case in BMNH. 57392 where there are only 5 auxiliaries abapically. In the complete circlet of BMNH. 40188 there are only 6 auxiliaries and 7 in BMNH. 25533 and £.29074. Both the anals and auxiliaries are unornamented. The gonopore and hydropore are apical and shared by Oi and 67. The hydropore is semicircular and somewhat slit-like and is adapical to the gonopore. In BMNH. 40153 the gonopore is covered by a pyramid of 4 gonals (see Text-fig. 35). The hydropore and gonopore may be quite separate as in SM. Ai26oi (Text-fig. 31) or the hydropore may be wrapped around the gonopore as in BMNH. 40153 (Text-fig. 35). 342 THE BRITISH SILURIAN CYSTOIDS 33 34 I FIGS. 33, 34. Apiocystites pentrematoides Forbes. Fig. 33. Sketch of part of an ambula- crum. Based on specimen SM. A.I2599. Fig. 34. Sketch of impression of ambulacral flooring plates on the thecal plates. P, primary flooring plate ; S, secondary flooring plate. The arrows point to the mouth. Based on specimen SM. A. 12597. The stem is typical of the Callocystitidae but from what is known the distal portion is reduced, and may merge gradually into the proximal portion. SM. Ai2599 nas a complete stem terminating in a root structure (see PL 10, fig. 14). It is 20 mm. long compared with a thecal height of 15 mm. The proximal portion is relatively large, FIG. 35. Apiocystites pentrematoides Forbes. Sketch of the area around the gonopore and hydropore, showing gonopore closed by pyramid of four gonals and closely approxi- mated to hydropore. F, facet ; G, gonopore ; H, hydropore ; OO, orals. Cf. Text-fig. 31. Based on specimen BMNH. 40153. THE BRITISH SILURIAN CYSTOIDS 343 at least 13 mm. long, and tapers from 4-5 mm. at the base of the theca to about 2-5 mm. The distal portion is very short and narrow. In BMNH. 48192 the stem is shorter than the thecal height and the distal portion consists of only 5 columnals. There are 8 outer proximals in this specimen and 17 or 18 in SM. Ai26oi. The outer columnals are provided with an external flange which is slightly beaded with fine tubercles in well preserved examples. Family ECHINOENCRINITIDAE Bather 1899 DIAGNOSIS. A family of Glyptocystitida with typically produced periproct composed of anal pyramid and single circlet of auxiliaries ; with short ambulacra restricted to apex of theca. Modified from Kesling (1963). Within this family the rhombs are all disjunct pectinirhombs, but there are two different types. In Echinoencrinites, and the other Ordovician genera except Scolio- cystis, all the dichopores of each rhomb are discrete entities and they open externally in slits which are level with the surface of the thecal plates. The slits are isolated from each other and provided individually with very narrow raised rims. In the Silurian genera, and Scoliocystis, the walls of individual dichopores in a rhomb are continuous with those of the adjacent dichopores at the external surface and the dichopores are effectively folds in the thecal plates. All the slits of one half -rhomb open in a depression or vestibule which is provided, as a whole, with a raised rim. This rim is open in one half-rhomb and closed in the other. The Echinoencrinites type of disjunct pectinirhomb was common in the Ordovician but the other type predominates in the Silurian and Devonian. The latter type was more efficient and was independently acquired by several families of Glyptocystitida. If Jacket's figures of Scoliocystis (Jaekel 1899, pi. n, figs. 10, loa, n) are accurate then it must be the ancestor of the British Silurian echinoencrinitids, and these three genera may be distinguished from the others by their advanced type of pectinirhomb. Jaekel (1899) proposed the family Scoliocystidae which was unfortunately pre-occupied by Bather's Echinoencrinitidae published earlier in the same year. The name is now resurrected for a subfamily of the Echinoencrinitidae containing the genus Scolio- cystis. The Echinoencrinitidae may therefore be divided into the following two subfamilies : Subfamily EGHINOENCRINITINAE nov. DIAGNOSIS. A subfamily of Echinoencrinitidae with disjunct pectinirhombs composed of dichopores opening as discrete slits, individually surrounded by slight rims and level with external surface of thecal plates. Subfamily SCOLIOCYSTINAE Jaekel 1899 emend DIAGNOSIS. A subfamily of Echinoencrinitidae with disjunct pectinirhombs composed of dichopores opening as slits within vestibules, collectively surrounded by a rim. 344 THE BRITISH SILURIAN CYSTOIDS Genus GLANSICYSTIS nov. As stated earlier the removal of Prunocystites fletcheri from the Echinoencrinitidae to the Staurocystinae necessitates the description of a new genus for " Echino- encrinites " baccatus Forbes. DIAGNOSIS. A genus of Scoliocystinae with ovate theca, four plates surrounding periproct, L5 interrupting the RR circlet, three pectinirhombs. TYPE SPECIES. Echinoencrinites baccatus Forbes 1848. 1848 1854 1857 1867 1868 1873 1876 18960 1899 1900 1943 Glansicystis baccata (Forbes) (PI. 7, figs. 2-18 ; PL 8, figs. 1-3 ; PL 10, figs. 10, n ; Text-figs. 36-39) Echinoencrinites baccatus Forbes : 506, pi. 17, figs. i-n. Echinoencrinites baccatus Forbes Echinoencrinites baccatus Forbes Echinoencrinites baccatus Forbes Echinoencrinites baccatus Forbes Murchison : 217, text-fig. 37, fig. 5. Pictet : 300. Murchison : 222, text-fig. 55, fig. 5. Bigsby : 25. Echinoencrinites baccatus Forbes ; Salter : 128. Echinoencrinus baccatus (Forbes) Quenstedt : 676, pi. 113, fig. 70. (3 figs.). Echinocystis baccata (Forbes) Haeckel : 146. Prunocystites baccatus (Forbes) Jaekel : 263. Prunocystites baccatus (Forbes) ; Bather : 60. Prunocystites baccatus (Forbes) ; Bassler & Moodey : 53, 184. DIAGNOSIS. As for the genus. LECTOTYPE. GSM. 7380. The original specimen of Forbes (1848, pi. 17, figs, i, 3, 4). MATERIAL. About 30 specimens of which three (BMNH. £.29057 ; GSM. 7380, 103101) are of the typical form. These are in BMNH., BU., GSM., OUM., SM. HORIZON AND LOCALITY. Wenlock Limestone, Middle Silurian ; Dudley, Worcs. and Walsall, Staffs. DESCRIPTION. The theca is elongate ovate with brachioles developed apically around the mouth. The anus is lateral and just over half way up the theca while the FIG. 36. Glansicystis baccata (Forbes). Sketch of arrangement of thecal plates of first four circlets. BB, basals ; ILL, infra-laterals ; LL, laterals ; RR, radials. Based on specimen SM. A.I2578. THE BRITISH SILURIAN CYSTOIDS 345 gonopore and hydropore are apical. Three disjunct pectinirhombs and the usual type of stem are present. The thecal plates are arranged in definite circlets and the usual formula may be given. The typical plating arrangement of the first four circlets is shown in Text-fig. 36. There is considerable variation however. Text-fig. 37 shows an arrangement in which ILLi-4 form a closed circlet and Text-fig. 38 another unusual arrangement in which there are 6 RR and L5 does not interrupt the RR circlet. Both these specimens appear to be anomolous. The four basals combine abapically to form a circular cicatrix 4 mm. in diameter in SM. Ai2577. There were 7 orals, 6 of which reach the peristome and contribute to its margin, as in the other Glyptocystitida. The odd oral is 67 (Text-fig. 39) . The size of the area occupied by the orals varies and is associated with two distinct trends within this species. In the lectotype and two other specimens (GSM. 103101, BMNH. £.29057) the area occupied by the orals is small (PL 7, fig. 8) while in the other examples this area is larger (PI. 7, fig. 2). The plates of the type and the two examples mentioned above are ornamented with FIG. 37. Glansicystis baccata (Forbes). Sketch of unusual arrangement of thecal plates in which IL5 has been elevated in the theca allowing ILLi-4 to form a closed circlet. BB, basals; ILL, infra-laterals ; LL, laterals ; RR, radials. SM. A.I2577- ramifying irregular raised pustules while those of all other specimens are ornamented with fine granules. There is also a tendency for the plates of the typical form to be much more tumid and the sutures deeper than in the other form. There are three disjunct pectinirhombs with half-rhombs with semi-circular outlines. All have few dichopores and prominent rims which are closed on IL2, Li and L4. These rims are more strongly developed on the typical form than on the others. B2 : IL2 BMNH. £.29057. This is a compressed rhomb measuring 2-8 mm. by 2-3 mm. with 8 dichopores in a width of 173 mm., the longest reaching 2-23 mm. The slits in both B2 and IL2 are 0-09 mm. wide and separated by 0-125 mm. In both half -rhombs they reach a maximum length of o -65 mm. Li : R5 (Lectotype). This is a compressed rhomb measuring 3 -53 mm. by 2-8 mm. externally with 8 dichopores in a width of 173 mm. The extra millimetre in the GEOL. 13, 6. 27 346 THE BRITISH SILURIAN CYSTOIDS width is entirely taken up by the prominent rhomb rim in Li. The longest dichopore is approximately 175 mm. long. The slits in Li are 0-107 mm- wide and separated by 0-125 mm. while those in R5 are only 0-07 mm. wide and separated by 0-14 mm. The slits in Li reach a maximum length of 0-6 mm. and those of R5 marginally longer. The same rhomb in SM. Ai2577 is highly compressed, being twice as long as wide, has 7 dichopores and a much less conspicuous rim on Li. L4 : R3 (BMNH. £.29057). This is a compressed rhomb measuring 4-33 mm. by 3-05 mm. with 10 dichopores in a width of 2-47 mm., the longest reaching 3-0 mm. The slits on 14 are obscured by matrix but those on R3 are 0-07 mm. wide and separated by 0-14 mm. They reach a length of approximately 0-75 mm. The same rhomb in SM. Ai2577 has only 8 dichopores, a much less conspicuous rim on L4, and is again more distinctly compressed. The ambulacra in the typical form are restricted to the oral plates and give rise to .two large, and possibly two small, facets in the lectotype. In the other form the ambulacra may extend on to radial plates (PL 7, fig. 2) and have up to 3 facets in one ambulacrum. The maximum number of facets seen was 7. There may well be FIG. 38. Glansicystis baccata (Forbes). Sketch of unusual arrangement of thecal plates in which there are six radial plates forming a closed circlet. BB, basals ; ILL, infra- laterals ; LL, laterals ; RR, radials. BMNH. £.29057. four ambulacra of which I and 4 terminate in a single large facet (the first formed) while ambulacra 2 and 5 may add up to three facets. This interpretation is not certain. The orals are ornamented between the ambulacra in the same way as the other thecal plates. The food grooves are broad and deep and covered by alternating cover plates which are unornamented. Apparently the cover plates were unable to articulate with the orals or flooring plates and were lodged in grooves in these plates. These grooves enlarged as the ambulacra grew. The mouth is a more or less central, oval opening. The facets are set in quite deep depressions in the oral or flooring plates. The brachioles are preserved in BU. Ho63 where they are at least 12 mm. long (none appears to be complete) and are composed of at least 50 brachiolars. As far as can be determined they do not taper in the preserved portion and are 0-75 mm. wide. The food groove was provided with cover plates and these are 2 to 3 times as numerous as the brachiolars. The periproct is surrounded by four plates which are : ILL4 and 5 and LL4 and 5. Apart from the two anomolous specimens figured in Text-figs. 37 and 38 the arrange- ment is constant. L-5 contributes just less than a third of the border and the other THE BRITISH SILURIAN CYSTOIDS 347 three plates share more or less equally in the remaining two-thirds. In the typical form there appear to be 12 auxiliaries forming a complete circlet but only half this number are found in SM. Ai2577 and possibly 7 in SM. Ai2578, still forming a complete circlet. The anal pyramid is composed of 6 anals in SM. Ai2577, the oru figs. 2-18; PL 8, figs. 1-3; PL 10, figs. 10, n Glauconome canalifera 34 prismatica 34 tetragona 54 Globoblastus norwoodi 308 Glyptocystites 307, 308 ehlersi 306 Glyptocystitida 299, 306*, 315, 343, 352 360 INDEX Glyptostrobus 285 Gymnolaemata 22, 23-108 Heterocella 53, 54 lerichei 54 monstruosa 54 pentagona 54 vicksburgica 54 sp. 5. 54 Heteropora 108 glandiformis 106 Hiantopora magna 62 Hincksinidae 22, 23, 24 Hippopleurifera 8, 64, 65 canui 7, 9, n, 15, 18, 63*, 64, 65 Hippopodinidae 80, 8 1 Hippoporina 5, 80, 81 beyrichi 75 globulosa 8-9, 12, 18, 80*, 80, 81 Hippopotamus 186 Hippothoa mucronata 64 HOOIJER, D. A. 117-190 Hyrachyus 128 Indophyllum 290 raot 290 Isoetes 227, 232-234, 240, 243, 249, 251, 252, 254, 255* echinospora 235, PI. 6, figs. 2, 3; PI. 7, figs. 1-8 humilior 233 ; PI. 5, figs. 7, 8; PI. 6, fig. i Jaekelocystis 331 hartleyi 306 Juglans 253 Karkenia incurva 279 Kenyapithecus wickeri 187 Kionidella 5, 101, 102, 103 dactylus 103 excelsa 101, 103 hastingsae 8-10, 13, 18, 100*, 102, 103 obliqueseriata 10, 103 Labioporella 9, 42-44 dartevellei 8-10, 15, 18, 41, 43*, 44 Labioporellidae 21 Laevigatisporites 243, 248—249, 252, 255* glabratus 235; PL 6, figs. 4, 5; PL 7, fig. 10 Lagarozoum 37, 46, 47 Lepadocrinus 315 bifasciatus 322 oblongus 331 quadrifasciatus 316 Lepadocystis moorei 306 Lepidocarpon 248 waltoni 248 Lepocrinites oblongus 331 quadrifasciatus 316 Lepralia auriculata 77 gattyae 60 hoernesi 73 nitida 60 schizogaster 77 Linum 253 Louisiana State University 5, 6, 12 Lorenicystis 331 angelini 304, 314 Lunulites n, 28, 29, 30 fcowef 30 distans 30 fenestra 30 jacksonensis 30 ligulata 30 tintinabula 30 transiens 5, 8-n, 18, 28, 29*, 30 truncata 30 urceolata 10, 28, 30 Lunulitidae 14, 28, 29, 30 Lycopodium 241, 243, 252, 255* clavatum 241, 242 1, 242; PL 10, fig. u; PL n, fig. i Maenioceras molarium 193, 195, 196, 205, 212, 213 terebratum 193, 195, 196, 205, 212, 213, 217 Mamillopora 101 Mamilloporidae 101, 102, 103 Manticoceras 195 cor datum 196 Margarettidae 21 Marsilea 235-237, 239-241, 243 di#ksa 235-237 drummondii 236-239; PL 8, figs. I, 4; PL 9, figs. 2, 3; PL 10, figs, i, 2 hirsuti 235—237 quadrifolia 236-239; PL 8, figs. 2, 3; PL 9, figs, i, 4, 5: PI- I0> ngs- 3-8 Masculostrobus 280 sahnii 280 Mazocarpon oedipternum 235 Mehtaia 281 Membranipora angulosa 31 oculata 48 vulnerata 37 Membraniporella 60—62 aragoi 61 bioculala 61 compressa 61 monilifera 61 planula 61 radiata 8-9, 18, 59*, 61, 62 subagassizi 61 ulrichi 61 INDEX Meniscopora 88 bigibbera 88 clavata 88 Mesodescolea 266 Mesosingeria 266 Metrarabdotos 92 Metroperiella 77 Microcachrydites 279, 280 Microcachrys 282 Micropora 35 articulata 40 Microporidae 14, 34, 35-39 Microporina 9, 34, 35, 36 magnipora 8-9, n, 18, 35, 36* Microstrobos 280, 282 Mucronella alifera 66 angustoecium 71, 73 pyriformis 88 Nellia 14, 48, 49-53 appendiculata 53 bifaciata 49, 52 bituberculata 49 midway anica 53 oculata 48, 49, 52 tenella 8-10, 15, 18, 48, 49, 50*, 51, 52 tenuis 53 ventricosa 8-10, 18, 50*, 52, 53 Nipanioruha 280 granthia 280-281 Nummulites 28 laevigatus 6 planulatus 6 variolarius 6, 7 Ogivalina 22, 23, 24 dimorpha 8-10, 18, 22, 23* eximpora 22, 24 OnychocellaSl, 32, 33 angulosa 31, 33 dimorpha 22 parisiensis 31, 33 subpyriformis 8-n, 18, 31, 32*, 33 Onychocellidae 31 , 32-34 Orbitulipora 27, 103, 104-106 haidingeri 103 petiolus 8-10, 13, 18, 25, 27, 103, 104*, 105, 106, 108 Orbituliporidae 103, 104-108 Otozamites grandis 284 Pachydermophyllum 266 Pachypteris 263, 264, 265-267 elegans 261, 263, 264, 265*, 266, 267; PI. i, fig. i; PI. 2, figs. 10-12 lanceolata 266 papillosa 266-267 Pagiophyllum 282, 290 paliferum group 197, 202 Palissya 286, 289 conferta 289 jabalpurensis 289 Palmicellaria lerichei 69, 71 PAUL, C. R. C. 297-355 Pelmatozoa 310 Perigastrella 73 Peristomella 65 alifera 66 Petralia mucronata 64 PETTITT, J. M. 221-257 Phaerospora 280, 283 Pharciceras lunulicosta 196 Pholadomya 10 Phyllocladus 282, 285 glaucas 280 Physcosmitrium 250 Pinularia globulifera 235 Pinus 246—247 balfouriana 246, 247 »«'gya 246 sylvestris 246, 247; PI. 17, figs. 1—4; PI. 18, fig. i, 2 thunbergii 246, 247 Plesiaceratherium gracile 153 Pleurocystites 306 Podocarpaceae 276, 277-294 Podocarpus 263, 275, 280, 282, 286, 287-290 dacrydioides 280 dubious 261, 284-285, 286, 288*, 289, 290, PI. 6, fig. 49, 50; PL 7, figs. 53, 54 inopinatus 289 palissyafolia 289 Polytrichum 250 Porella 69 variabilis 69 Poricellaria 9, n, 39, 40, 41 a/ata 5, 8-9, 1 8, 38*, 39, 40, 41 complicata 40 crassomuralis 41 diplodidymioides 41 limanowskii 40-41 negrisi 41 ratoniensis 15, 40, 41 vernoni 40 Poricellariidae 14, 39, 40, 41 Poristoma 87, 88 clavata 88 incisa 88 parisiensis 88 Poropeltarion 36 lebanonese 35 newelli 35 Proconsul 185, 186 Prunocystities 310, 334, 335 baccatus 300, 344 362 INDEX fletcheri 300, 304, 307, 315, 334, 335, 344; PI. 3, figs. 8, 9; PI. 8, fig. 4; PI. 9, figs. 10, ii Pseudocrinites 299, 305, 307, 309, 310, 315 321, 330, 334 abnormalis 321 bifasciatus 300, 301, 306, 310, 315, 321, 322, 323*-326*, 327, 329, 331, 335; PI. 2, figs, 4-8; PI. 3, figs. 2-6, PI. 10, figs. 7, 8 clarki 327 gordoni 308, 315, 321, 323, 327 magnificus 321, 322, 323*, 327 oblongus 330, 331 perdewi, 322, 327, 330 pyriformis 300, 301, 303, 310, 325, 327, 328*, 329*, PI. 3, figs, i, 7; PI. 4, figs. 1-5; PL 5, fig. 10 quadrifasciatus 315, 316 Pseudocrinus 322 bifasciatus 322 oblongus 331 quadrifasciatus 316 Psilotum 241-243, 252 nudum 242; PI. n, figs. 2, 3 Pteroma 267 Ptilophyllum 275, 283, 284 Peullina 60 Rectonychocella dimorpha 22 Regnellidium 235, 240, 241 diphyllum 240, 241; PI. 10, figs. 9, 10 Reptadeonella 98 Reptescharellina globulosa 80 Retepora marginata 73 Rhagasostoma dutempleana 47 spiniferum 46 Rhinoceros 167, 168 brachypus 174 incisivus 148 pachygnathus 124, 125 sansaniensis 174, 175 schleiermacheri 124, 125 sondaicus 157 unicornis 157 Rhombifera 308 bohemica 306 Romancheina gouetensis 66, 68 Ruflorinia 266 Salicornaria borealis 34 dichotoma 48 Salvinia 251, 252 auriculata PI. 21, fig. 5 Saxegothaca 285, 286 Schismoporella 77 Schizocystis 304, 307, '311, 348, 353 armata 300, 301, 305-307, 310, 314, 348, 349*-353* Schizomavella 77, 78 trigonostoma 8, 9, 18, 76*, 77, 78 Schizoporella hoernesi 78 var. procumbens 78 magnoaperta 91 Schizoporellidae 75, 76-81 Schizostoma 91 aviculiferum 91 crassum 91 denticulatum gi gibbosum 91 heveticum 92 Schizostomella 14, 91, 92-95 curryi 8-9, n, 12, 15, 18, 89*, 91, 92, 93 gibbosa 93 liancourti 8, 9, 18, 91, 93, 94*, 95 Schopfipollenites 243-245, 252, 253, 255* sp. PI. 13, figs. 1-3; PL 14, fig. i Scleropteris 266 pomelii 266 Scoliocystidae 334, 343 Scoliocystinae 343, 348 Scoliocystis 343 Scutellum 195, 204, 205-219 (Paralejurus) 211 (Scutellum) 205, 211, 213 costatum 193, 211, 213, 219 costatum 193, 196, 211, 212; PL 2, fig. 10 lummatonensis 193, 196, 197, 200, 212, 213; PL 2, fig. 10 whidbornei 193-197, 200, 211, 212, 213; PL 22, figs. 9, 11-18 delicatum 193, 204*, 217-219 delicatum 193-196, 213, 214-218; PL 3, figs. 4, 7-9 tigrinum 193, 196, 217, 218, 219; PL 3, fig. 5, 6, 10-12 flabelliferum 193, 196, 219; PL 3, figs. 1-3 pardalios 193, 196, 204*, 204, 205-211; PL i, figs, i-n (Thysanopeltis) speciosum 204* alutaceum 193 geesensis 203 costatum 195, 199, 202, 204, 211, 218, 219 lummatonensis 195, 198*, 199, 201, 202 whidbornei 195, 198*, 199, 201, 202, 211 delicatum 198*, 199, 201, 202 tigrinum 195 flabellifer 195, 198*, 199, 201, 202, 210, 211, 219 pardalios 198*, 199, 201-203, 211 Sedgley Limestone 304 Selaginella 227, 229, 232-234, 240, 243, 244, 248, 249, 252-255* galeotii 228 helvetica 228 kraussiana 227, 229, 232 wyoswws 227, 229, 231 padangensis 231 pulcherrima 227, 229, 231, 232, 252, 253; PL i, figs. 1-6 INDEX 363 selaginioides 227 spinulosa 228 Selenariidae 14, 21, 24, 25-27 SELWOOD, E. B. 191-220 Semieschara dutempleana 47 parisiensis 31 Sequoia 285 Sertella 8, 73-75 marginata 8-10, 18, 72*, 73 Sertellidae 37, 73, 74, 75 Setosella 37, 38, 39 fragilis 8, 9, 18, 37-39, 3§* vulnerata 37 Setosellina 9, 24 capriensis 27 goesi 27 gregoryi 8-10, 12, 18, 25, 26*, 27 roulei 24 Sigillaria 249 Siphonella cylindrica Siphonoporella 44 Smittia coccinea var. alifera 66 variabilis 69 (Mucronella) hoernesi 71 (Porella) variabilis 69 Smittina 69 angulata 69 orbavicularia 69 reticuloides 69 telum 69 variabilis 69 Smittinidae 68, 69-71 Smittipora 33, 34 fragilis 41, 42 midwayanica 34 sp. 8, 9, n, 18, 32*, 33, 34 Smittoidea 9, 68-71 prolifica 68 variabilis 8-10, 12, 18, 67*, 69, 70, 71 Sphenopteris 275 Stach-Brown classification 12 Stachycarpus 287 Stamenocella dimorpha 22 Stangeria 266, 267 paradox a 266 Staurocystinae 300, 315, 330 Staurocystis 307, 315, 324, 325, 329-331, 334 cruciata 316 oblongus 316, 331 quadrifasciatus 300, 301, 306, 309*, 310, 3i2*-3i4*, 315, 317*, 3i9*-32i*, 323, 335; PI. i, figs. 1-8; PL 2, figs. 1-3; PI. 10, figs. 1-5, 12 Stauropteris burntislandica 249 Steganoporella 15, 19, 44, 93 elegans 35 fragilis 41 parvicella 17 variabilis 44 Steganoporellidae 21, 41, 42-46 STONE, M. 194 SUTCLIFFE, A. J. 121 Sycostoma 7 Taeniopteris 285 Taxodium 285 Taxospermum 248 undulatum PL 18, fig. 4 Ta#ws 245, 246, 255* baccata 245, 250; PL 16, figs. 1-4 TAYLOR, T. N. 224 Teichopora 87, 88-91 clavata 5, 8, 9, 15, 18, 87, 88, 89* syringopora 91 Teleoceras 121 aginese 140, 153, 180 aquitanicum 140 snowi 147 Tellinae 7 Tetracystis 300, 315, 321, 330, 331 bifarius 331 chrysalis 330, 331 Jenestratus 331 oblongus 300, 304, 313, 315, 316, 321, 331, 332*, 333> 335-' PI- 5, figs. 2-7, 11-13; PI- 10, figs. 6, 13 Thysanopeltidae 204, 205-219 Ticoa 263, 266, 269, 270, 271 harrisii 270-272, 285, 291 lamellata 261, 263, 268*, 269-271; PL i, figs. 2, 8, 9; PL 3, figs. 15, 16 magnipinnulata 270, 271 Tobas Amarillas 263 Tomaxellia 263, 271, 280, 287, 290, 291-294 biforme 261, 284, 290, 291, 292, 293*, 294; PL 4, fig. 18-20 degiustoi 261, 263, 287, 290, 291; PL 4, fig. 17 Trigonocarpus 247, 248 sp. PL 18, fig. 3 Trigonopora monilifera 69, 71 Triletes 279 Triophodon 186 Trimerocystis peculiaris 321 Trisaccites 276, 278, 280 Trisacocladus 261, 276, 277-282 tigrensis 261, 276, 277, 278*, 279, 280, 281*, 282; PL 4, figs. 21 ; PL 5, figs. 22-39; PL 8, figs. 56, 67 Tubucella 5, 84-87 mamillaris $,, 54~58 davisi 8, 9, 15, 18, 54, 55*, 57, 58 fragilis 53, 54 haidingeri 54 INDEX 54 lediensis 54 monstruosa 8, 9, 15, 18, 54, 55*, 56, 57 polymorpha 54 subsymmetrica 54, 58 75 Wenlock Shale 299, 301, 303 Whittleseya 243 PRINTED IN GREAT BRITAIN BY ADLARD & SON LIMITED BARTHOLOMEW PRESS, DORKING