t= = est SHISSUN EH: paste hs i ite SEES Ei staineg sti piste sianititie strsrececssrssesres Stitiststetsissras: apsesseatastetecs spay rotirsssress Esta instestesreseriesessegaeteseoriasy iste Seaetstsspstesctestestsets eas tH 333 a = oe Cy 2) SEP 1966 BULLETIN OF ae AL WSS THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY VOL. X 1964—1965 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) LONDON: 1966 DATES OF PUBLICATION OF THE PARTS No. No. No. No. No. No. No. No. No. No. ro. No. 11. 2 10 July 2 December . 26 February 27 April 3 June 7 July. 12 July 15 December 24 November 3 December . 3, December . PRINTED IN GREAT BY ADLARD & SON BARTHOLOMEW PRESS, BRITAIN LIMITED DORKING 1904 1964 1905 1905 1965 1965 1905 1965 1905 1905 1965 Not I INO? 2 INO! 3 No. 4 No. 5 No. 6 No. 7 No. 8 No. 9 No. 10 No. Ir CONTENTS GEOLOGY VOLUME X PAGE Middle Jurassic Ostracoda from the Millepore Series, Yorkshire. Ree BADE: I Revision of British marine Cretaceous Ostracoda with notes on additional forms. P. KAYE 35 Two heterosporous plants from the Upper Devonian of North America. J. M. PETritr 81 Silurian Polyzoa from Benthall Edge, Shropshire. D. E. OWEN 93 Fossil Ginkoales from the Ticé Flora, Santa Cruz Province, Argentina. S. ARCHANGELSKY 11g The generic position of Osmundites dowkeri Carruthers. M. E. J. CHANDLER 139 Fossil Mammals of Africa No. 18: East African Miocene and Pleistocene Chalicotheres. P.M. BUTLER 163 Fossil Mammals of Africa No. 19: The Miocene Carnivora of East Africa. R. J. G. SAVAGE 239 Dechenellid Trilobites from the British Middle Devonian. E. B. SELWOOD 317 Cretaceous Ammonites and Nautiloids from Angola. M. K. HowaARrRTH 335 The Fauna of the Portrane Limestone, III. D. Karjo & E. KLAAMANN 413 Index to Volume X 435 “MIDDLE JURASSIC OSTRACODA ROM THE MILLEPORE SERIES, é, YORKSHIRE R. H. BATE ea BULLETIN: OF- TISH MUSEUM (NATURAL HISTORY) | <3 ie De Vol. to No. 1 Si LONDON: 10964. MIDDLE JURASSIC OSTRACODA FROM THE MEE PORE, SERIES, YORKSHIRE Wy Za \ JUL 1964 ee: BY RAYMOND HOLMES BATE, Ph.D. Pp. 1-33 ; 14 Plates; 1 Text-figure BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY | Vol. 10 No. 1 \ LONDON: 1964 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), <¢mnstituted im 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper 1s Vol. 10, No. i of the Geological (Palaeontological) series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Perrodicals. © Trustees of the British Museum (Natural History) 1964 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued July, 1964 Price Thirty-five Shillings MILLEPORE SERIES, YORKSHIRE By R. H. BATE CONTENTS I. INTRODUCTION AND ACKNOWLEDGEMENTS II. SySTEMATIC DESCRIPTIONS Order Podocopida Miiller Suborder Platycopina Sars Family Cytherellidae Sars Genus Cytherelloidea Mexander : Cytherelloidea catenulata (Jones & Shadoova Suborder Podocopina Sars : : : : : Superfamily Bairdiacea Sars Family Bairdiidae Sars. Genus Bairdia M‘Coy Bardia hilda Jones Superfamily Cypridacea Baird Family Paracyprididae Sars Genus Paracypris Sars Parvacypris bajociana Bate Superfamily Cytheracea Baird Family Bythocytheridae Sars Genus Monoceratina Roth 5 Monoceratina vulsa (Jones & Ghecbou) Family Progonocytheridae Sylvester-Bradley ; Subfamily Progonocytherinae Sylvester-Bradley Genus Progonocythere Sylvester-Bradley . Progonocythere cristata Bate Genus Acanthocythere Sylvester-Bradley Subgenus Protoacanthocythere Bate Acanthocythere (Protoacanthooythere) Waucolaia Bate c Genus Aulacocythere Bate : Aulacocythere punctata Bate Aulacocytheve reticulata Bate Genus Fuhrbergiella Brand & Malz . Subgenus Praefuhrbergiella Brand & Malz Fuhrbergiella (Praefuhrbergiella) avens Bate Fuhrbergiella (Praefuhrbergiella) minima sp. nov. Genus Micropneumatocythere Bate Micropneumatocythere convexa Bate Micropneumatocythere globosa sp. nov. Genus Pneumatocythere Bate . Pneumatocythere bajociana Bate Pneumatocythere cavinata sp. nov. Subfamily Pleurocytherinae Mandelstam . Genus Pleurocythere Triebel Pleurocythere kivtonensis Bate Pleurocythere nodosa Bate Family Cytherideidae Sars Subfamily Cytherideinae Sars UV Ss) og iS OCOODOKCODOO WMWADADMDBDDDHDWH OHO WO WO OF [ooo oo ooo) HAA A He Ae A A A A AW Ae AW AW a OW Ot Oa ts ss da SPPHHAARWWWHDNHHH OO 0 0 0000 MIDDLE JURASSIC OSTRACODA FROM THE 4 OSTRACODA FROM THE MILLEPORE SERIES, YORKSHIRE © f Page Genus Dolocytherve Mertens _ _. a ae Eee : : 14 Dolocythere maculosa Bate. : : : é 14 Family Schulerideidae Mandelstam_. : : : : 9 15 Subfamily Schulerideinae Mandelstam : . : : 15 Genus Asciocythere Swain : : : : 6 : 15 Asciocytheve acuminata sp.nov. . ‘ : : 15 Asciocythere lacunosa Bate . : : : dnc lO Genus Eocytheridea Bate 4 5 ; F : : 16 Eocytheridea ? acuta sp. nov.. ‘ 3 : ‘ 16 Eocytheridea ? astricta sp. nov. : : : 3 17 Eocytheridea cavinata sp. nov. : 5, : : 18 Eocytheridea elongata Bate . : 5 : : 19 Eocytheridea ? evugata sp. nov. : F j ; 19 Eocytheridea faveolata sp. nov. 5 : ‘ : 20 Eocytheridea lacunosa Bate . : : 6 : 21 Eocytheridea reticulata sp. nov. ‘ , PS : 21 Genus Praeschuleridea Bate. ; : 22 Pyaeschuleridea subtrigona (Jones & Shewbinea), : 22 Pyaeschuleridea subtrigona subtrigona (Jones & Sherborn) : ; 8) 1890 Cytheropteron cuspidatum tricuspidata Jones & Hinde : 38, pl. 3, figs. 6, 7. 1936 Monoceratina tricuspidata (Jones & Hinde) Veen : 9, 42, 43, pl. 2, figs. 4-11. 1941 Monoceratina tricuspidata (Jones & Hinde) ; Bonnema : 40, pl. 6, figs. 77-80. 1941 Monoceratina tricuspidata (Jones & Hinde) ; Triebel : 353. Driacnosis. Small Monoceralina with three prominent ventro-lateral spines. Other subsidiary tubercles occur over the lateral surface but no reticulation. Eye tubercle well developed. LecToTyPeE. B.M.N.H., Io. 1583 (Jones & Hinde 1890), Upper Chalk, Keady Hill. OTHER MATERIAL. B.M.N.H., lo. 1202, Io. 1582, two specimens from the Upper Chalk, coranguinum Zone, Sonning, Berks. MEASUREMENTS Length Height Right valve (B.M.N.H., Io. 1583, lectotype) . 0:65 mm. 0:25 mm. REMARKS. The species is similar in shape to MW. montuosa but is smaller and has three very prominent lateral spines. It occurs throughout the Upper Chalk in Britain, but is never very abundant. Monoceratina umbonata (Williamson) (Pl. 4, figs. 3, 4, 6-8) 1847 Cytherina umbonata Williamson : 82, pl. 4, fig. 78. 1849 Cythere umbonata (Williamson) Jones : 12, pl. 2, figs. 3a-g. 1870 Cytheropleron umbonatum (Williamson) Jones : 74, 76. 1872 Cytheropteva umbonata (Williamson) Williamson : 136. ?1880 Cythere umbonata (Williamson) ; Marsson : 45, pl. 3, figs. 15a—c. 1890 Cytheropteron umbonatum (Williamson) ; Jones & Hinde : 40, pl. 1, figs. 21-26. 1890 Cytheropteron umbonatum longispinata Jones & Hinde : 41, 42, pl. 3, figs. 11, 12; pl. 4, figs. 30, 31. 1893 Cytheropteron umbonatum (Williamson) ; Chapman & Sherborn : 347. 21934 Monoceratina umbonata (Williamson) Alexander : 62, pl. 8, fig. 9. 1941 Monoceratina umbonata (Williamson) ; Bonnema : 29, pl. 6, figs. 54-62. MATERIAL. (i) B.M.N.H., In. 51595-51601, figured Jones (1849, pl. 2, figs. 3a-g) from the Chalk Detritus at Charing. (ii) B.M.N.H., Io. 314 (Morris) Charing ; B.M.N.H., In. 19382-85 (Hinde) Upper Greensand, Warminster; B.M.N.H., I. 2676-77 (Chapman) Gault Clay, Folkestone. (iii) B.M.N.H., Io. 1203, Io. 1590, Cambridge Greensand, Barrington (Cambs.). (iv) Hull University 17.C.8.1, Red Chalk, Speeton, E. Yorks. MEASUREMENTS Length Height Adult left valve (B.M.N.H., Io. 1203) . 2) 0-78 mm: 0-34 mm. REMARKS. This species is extremely variable in ornament and degree of inflation, the ornament being reticulate but also often strongly spinose. The original figure by Williamson is drawn from an oblique angle and those of Jones (1849) from the same locality are generally taken as typical. Most of Jones’ figured specimens are pre-adults, the larger adult forms being rare. The spinose variants were grouped by Jones & Hinde (1890) into a distinct variety : var. longispinata. All BRITISH MARINE CRETACEOUS OSTRACODA 57 intermediates between the spinose and simple reticulate forms are however found and therefore no separation can be made on this basis. The forms figured by Jones & Hinde as var. longispinata were adult specimens which would be expected to have a more strongly developed ornament than the young moults. The length and changes in shape of the lateral spine described as a varietal difference are due to breakage and forms described as having a short, broad, flat topped spine have the spine broken. The degree of inflation is particularly variable. The lateral surface anterior to the median sulcus is generally flattened in true M. wmbonata. Forms with inflated anterior lateral areas described by Jones & Hinde (1890) as var. acanthoptera Marsson are renamed Monoceratina umbonatoides (see below). These forms, often with a pronounced node on the antero-lateral area, have almost certainly evolved from M. umbonata and transitional forms do occur. M. umbonata is largely confined to Albian and Cenomanian sediments, however, whilst the Upper Cretaceous members of this plexus are found to be M. wmbonatordes. M. umbonata differs from M. pedata pedata principally in shape. It has a well marked median sulcus and the reticulation is often arranged concentrically around the lateral spine. Monoceratina umbonatoides nom. nov. (Pl. 4, figs. 5, 6) 1890 Cytheve umbonatum acanthoptera (Marsson) ; Jones & Hinde: 41, pl. 1, figs. 11-13 ; pl. 4, figs. 22-29. Lectotype. B.M.N.H., Io. 1592 (Jones), Magee, Antrim, here designated. OTHER MATERIAL. (i) B.M.N.H., Io. 374 (Jones & Hinde) Upper Chalk, Keady Hill, Derry. (ii) B.M.N.H., Io. 1205, Io. 1591, three specimens from the Upper Chalk, coranguinum Zone, Sonning, Berks. MEASUREMENTS Length Height Left valve (B.M.N.H., Io. 1205) : . 0.65 mm. 0-39 mm. Remarks. This species differs from M. wmbonata in the strong inflation of the antero-lateral area. A large node is usually present antero-dorsal to the median sulcus. The species seems to be restricted to the Upper Chalk and is probably a direct descendent from M. wmbonata. It differs from Marsson’s form (1880, pl. 3, figs. 14a—c) in the reticulate ornament and in having the spine posterior to rather than below the median sulcus. Family PROTOCYTHERIDAE Genus PROTOCYTHERE Triebel 1938 Protocythere consobrina Triebel (Pl. 5, figs. 17-19) 1938 Protocytheve consobrina Triebel : 184, pi. 1, figs. 6, 7. MatTeriAL. B.M.N.H., Io. 1190, Io. 1603-05 from the Lower Gault, Culham, Oxfordshire. 58 BRITISH MARINE CRETACEOUS OSTRACODA MEASUREMENTS Length Height Male left valve (Io. 1605) . ; : . 1:05 mm. 0°57 mm. Female left valve (Io. 1190) F : . 0-92 mm. 0:57 mm. REMARKS. This species is very similar to Protocythere triplicata (Roemer) from the Hauterivian and Barremian. The principal differences are that the ribs are less convex and the anterior hinge element is set slightly out of line, the median groove in the right valve passing above the anterior tooth. Specimens of P. triplicata from the Middle Barremian at Speeton, E. Yorkshire, are figured on PI. 5, figs. 12, Topelese P. consobrina lacks the ventral riblets of P. lineata and has smooth intercostal areas. The instars have subdued ribbing whilst those of P. lineata are still strongly emphasised, obscuring the dorsal margin. Protocythere lineata (Chapman & Sherborn) (Pl. 5, figs. 1-8) 1849 Cytheve (Cytherveis) tviplicata (Roemer) ; Jones : 18, pl. 3, figs. ga—/h. 1870 Cythereis tyviplicata (Roemer) Jones : 75-76. 1890 Cythereis triplicata (Roemer) ; Jones & Hinde : 19, pl. 1, figs. 56-61. 1893 Cythereis triplicata lineata Chapman & Sherborn : 348, pl. 14, fig. 5. 1898 Cythereis triplicata lineata Chapman & Sherborn ; Chapman : 338. 1938 Protocythere jonest Triebel : 186, pl. 1, figs. 8—ro. 1956 Protocythere jonesi Triebel ; Deroo : 1514. Diacnosis. Protocythere with three prominent longitudinal ribs. The ventral one bears small longitudinal riblets upon its surface. Intercostal areas with irregular network of small cross ribs. LectotyPe. B.M.N.H., I. 2704, figured Chapman & Sherborn (1893, pl. 14, fig. 5) from the Gault Clay, Middle Albian at Folkestone, Kent, here designated. OTHER MATERIAL. (1) B.M.N.H., In. 51665-66, In. 51668-72 figured Jones (1849, pl. 3, figs. ga—-g) from the Detritus at Charing and from Folkestone Gault Clay. (ii) B.M.N.H., Io. 309 (Morris) from the Detritus at Charing. (iii) B.M.N.H., I. 2464 (Jones & Hinde 1890) from the Chalk marl at Didcot. (iv) B.M.N.H., Io. 1187, Io. 1596-1600 from the Upper Gault at Burwell, Cambs. MEASUREMENTS Length Height Right valve (B.M.N.H., I. 2704, lectotype) . 0-60 mm 0-30 mm. REMARKS. This species identified by Jones (1849) as Cythereis triplicata (Roemer). was renamed by Triebel (1938). Unfortunately a specimen described as a subspecies by Chapman & Sherborn (1893) has been found to be a young stage of C. tviplicata sensu Jones. This subspecies C. tviplicata lineata has priority over Triebel’s subse- quent correction and therefore provides the valid specific name. The species has been well described by Triebel. Its particular characteristics include the longitudinal riblets upon the ventral rib and the numerous small ribs which cross the inter-costal areas. The prominence of these ribs is variable, a fact which is particularly apparent amongst specimens of different states of preservation. P. lineata is closely related to P. consobrina which occurs in equivalent strata in BRITISH MARINE CRETACEOUS OSTRACODA 59 Southern England. The latter is, however, more strongly inflated and lacks the ventral riblets. The intercostal areas are smooth and the dorsal and ventral ribs are less arched. Another related form is P. tricostata which has the intercostal areas strongly reticulate. The longitudinal ribs are longer and keel-like and do not obscure the dorsal and ventral margins to the same extent. Protocythere tricostata Triebel (Bie 5 tigsi4. nO) 1938 Protocythere tricostata Triebel : 190, pl. 2, figs. 17-22. MATERIAL. B.M.N.H., Io. 1188, four specimens from the Middle Albian at Speeton, E. Yorkshire. MEASUREMENTS Length Height Male left valve (B.M.N.H., Io. 1188) . =) 0:90 mm: 0-47 mm. Female left valve (B.M.N.H., Io. 1188) . 0-80 mm. 0-47 mm. Protocythere rudispinata (Chapman & Sherborn) (Pl. 5, figs. 9-11) 1893 Cythereis rudispinata Chapman & Sherborn : 348, pl. 14, figs. 6, 7. DiaGnosis. Small Protocythere with three longitudinal rows of large flat topped spines running across the lateral surface. Anterior margin bearing a row of spines. Intercostal areas smooth. LectotyPe. B.M.N.H., I. 2705 figured Chapman & Sherborn (1893, pl. 14, fig. 6), a left valve from the Lower Gault at Folkestone, here designated. PARALECTOTYPE. B.M.N.H., I. 2705 figured Chapman & Sherborn (1893, pl. 14, fig. 7), a right valve from the same locality. OTHER MATERIAL. B.M.N.H., Io. 1189 from the Lower Gault at Henfield, Sussex. MEASUREMENTS Length Height Left valve (B.M.N.H., I. 2705, lectotype) . 0:59 mm. 0:33 mm. DEscRIPTION. Valves relatively small, compressed laterally. Dorsal and ventral margins straight and subparallel. The lateral surface bears three longitudinal rows of stout flat-topped spines. A further row of similar spines runs along the ventral surface, whilst smaller spines are often found between the two major ventral rows. The weak anterior marginal rib bears a double row of laterally directed spines. Anterior and posterior margins tuberculate. Intercostal areas smooth. Normal pore canals rare, usually connected with a spine on lateral surface. Duplicature broad and crossed by numerous radial pore canals which curve upwards antero-dorsally. Hinge crenulate merodont with strongly divided stepped terminal elements. REMARKS. The shape, marginal features and hinge place Chapman & Sherborn’s specimens in the genus Pyvotocythere rather than Cythereis. The form described by Triebel (1940, pl. 4, figs. 47-50) and by Deroo (1956) differs fundamentally, being larger, compressed dorsally, more strongly convergent posteriorly and lacks the 60 BRITISH MARINE CRETACEOUS OSTRACODA prominent hinge ear in the left valve. It also differs in the details of the marginal area, hinge, and normal pore canals and falls within the latter genus. Further differences are in the shorter nature of the spines and the absence of spines along the ventral surface. Genus VEENIA Butler & Jones 1957 Veenia barringtonensis sp. nov. (Pl. 6, figs. 1-3) DraGnosis. Veenia, with three longitudinal ribs which almost join posteriorly Ventral rib connected to anterior margin at one-third height, also bears row of laterally divided tubercles at posterior end. HoLotyrPeE. B.M.N.H., Io. 1172, a female left valve from 1 ft. below the Cambridge Greensand, Barrington (Cambs.). PARATYPES. Four specimens, B.M.N.H., lo. 1173-76, from the same horizon. MEASUREMENTS Length Height Male left valve (B.M.N.H.., lo. 1173, paratype) 0-87 mm. 0-47 mm. Female left valve (B.M.N.H., Io. 1172 holotype) 0-79 mm. 0°47 mm. DESCRIPTION. Valves elongate, compressed, anterior broadly rounded, posterior pointed at mid-height in the right valve but forming a blunt point at the postero- dorsal angle in the left valve. Dorsal margin straight, ventral margin straight or weakly convex. Three longitudinal ribs cross the lateral surface. The dorsal rib is rather sinuous and is equal in length to the median hinge element, obscuring the margin in its central part but curving downwards anteriorly and posteriorly. The median rib is shorter and straight, being connected with a low muscle node anteriorly. The ventral rib is strongly convex. Anteriorly it is connected to the margin at one-third height by a short horizontal cross rib, posteriorly it bears a row of 5 or 6 small laterally directed tubercles on its crest. The anterior and posterior margins are tuberculate, each tubercle corresponding to the extremity of a radial pore canal. Duplicature broad, crossed by a few, thick, radial pore canals. These number 10 anteriorly and 6 posteriorly, being concentrated antero- and postero-ventrally, the upper ones curving dorsally. Inner margin and line of concrescence coincide. Hinge strongly amphidont having in the right valve two high, divided, terminal teeth separated by a long, locellate, median groove deepened anteriorly into a smooth socket. The socket and groove are open ventrally but are bounded dorsally by a high, smooth bar. The median groove extends somewhat above the terminal elements. In the left valve there are two strong, divided sockets, separated by a high strongly denticulate bar. The bar bears a prominent smooth tooth at its anterior end and is separated from the dorsal margin by a narrow shelf. In front of the anterior socket the margin is elongated into a keel-like process which fits into a depression above the anterior tooth in the right valve. RemARKS. JV. barringtonensis differs from the closely related V. harris:ana in the shape of the longitudinal ribs and the greater emphasis of the ornament. BRITISH MARINE CRETACEOUS OSTRACODA 61 Veenia harrisiana (Jones) (Pl. 4, fig. r ; Pl. 6, figs. 4-11) 1849 Cytherve (Cythereis) interrupta (Bosquet) ; Jones : 16, pl. 2, figs. 6a—g. 1849 Cythere (Cythereis) quadrilaterata (Roemer) ; Jones : pl. 4, figs. 10h, 2. 1870 Cythere harvisiana Jones : 75, 76 (new name). 1890 Cytheve havvisiana Jones ; Jones & Hinde : 16, pl. 1, figs. 47-52. 1890 Cythere harvisiana veticosa Jones & Hinde : 18, pl. 1, fig. 46. 1890 Cythere harvisiana setosa Jones & Hinde : 17, pl. 1, figs. 43-45. 1890 Cythereis auriculata (Cornuel) ; Jones & Hinde : 19, pl. I, figs. 53-55. 1893 Cythere harrisiana Jones ; Chapman & Sherborn : 346. 1893 Cythere havrisiana rveticosa Jones & Hinde ; Chapman & Sherborn : 346. 1893 Cythere harvrisiana setosa Jones & Hinde ; Chapman & Sherborn : 346. 1893 Cythere auriculata (Cornuel) ; Chapman & Sherborn : 346. 1893 Cythere lineatopunctata Chapman & Sherborn : 348, pl. 14, fig. 4. 1893 Cythere koninckiana (Bosquet) ; Chapman & Sherborn : 348, pl. 14, fig. 2. 1898 Cythere harvvisiana Jones ; Chapman : 335. 1898 Cythere havrisiana reticosa Jones & Hinde ; Chapman : 336. 1898 Cythere harvisiana setosa Jones & Hinde ; Chapman : 335, 336. 1898 Cythere koninckiana (Bosquet) ; Chapman : 337, 338, figs. 4a, b. 1898 Cytheveis auriculata (Cornuel) ; Chapman : 338. 1938 Protocythere auviculata (Cornuel) Triebel : 195, pl. 2, figs. 27-31. 1956 Protocythere triebeli Deroo : 1515 (new name). 1963c Veenia triebeli (Deroo) Kaye : 233, pl. 18, figs. 10, 11. ?1963c Homocythere reticulata Kaye : 234, pl. 18, figs. 8, 9. DiaGnosis. Veenia with three low, inflated, straight parallel longitudinal ribs. Hinge ears prominent in left valves. Lateral surface smooth-pitted. Marked changes occur in ornament throughout ontogeny. LectotyrPe. B.M.N.H., In. 51663 figured Jones (1849, pl. 2, fig. 6d), a pre-adult right valve from the Gault Clay, Folkestone, here designated. PARALECTOTYPES. B.M.N.H., In. 51657-62 figured Jones (1849, pl. 2, figs. 6a—c, e-g), figs. 6b, c, g from the Gault at Folkestone, figs. 6a, e, f from the Detritus at Charing. OTHER MATERIAL. (i) B.M.N.H., I. 2689 (Chapman & Sherborn) ; B.M.N.H., I. 2690 (Chapman & Sherborn var. reticosa) ; B.M.N.H., I. 2691, Io. 1610 (Chapman & Sherborn var. setosa) ; B.M.N.H., I. 2680 (Chapman & Sherborn C. auriculata) ; B.M.N.H., I. 2703 (Chapman & Sherborn C. lineatopunctata pl. 14, fig. 4) all from the Gault Clay at Folkestone. (ii) S.M.B. 40574-77, B. 40580-81, B. 40585-88, B. 40621 (Chapman 1898) all from the Cambridge Greensand at Swaffham. (iii) B.M.N.H., Io. 1606-09 from the Middle Albian at Speeton, E. Yorkshire. MEASUREMENTS Length Male left valves. ‘ y ‘ : . 0:96-1:00 mm. Female left valve . : : : ‘ . 0:86-0:90 mm. Penultimate instars : : : : . 0:66-0-75 mm. Instars group A . : : : : . 0°52-0°56 mm. Instars groupB . : : : : . 0°43-0-45 mm. Instars groupC . : ; : : . 0:36-0:38 mm. Instars group D_. 3 ; : : . 0:30-0:32 mm. 62 BRITISH MARINE CRETACEOUS OSTRACODA REMARKS. This species must be amongst the most confused of all Cretaceous ostracoda. Most of the early references refer to pre-adult valves as the adult form was not described until 1890. The difference in shape and ornament between the adults and pre-adults led to their being considered as separate species for a considerable time. The adults, first included within Protocythere auriculata (Cornuel) were renamed P. triebeli by Deroo in 1956. The pre-adults, which offer the first valid specific name, were further subdivided on a basis of ornament. This ornament varies from completely smooth to strongly reticulate and all intermediate stages are usually seen within the same sample (PI. 4, fig. 1). Ona basis of the amphidont hinge the writer (1963c) placed the species within the genus Veenza. A further confusing feature is that weak sexual dimorphism is shown by the penultimate moults ; the valve proportions being interpreted as varietal differences by early authors. The adult specimens differ from the pre-adults in the following ways :—The long margins are parallel, and a prominent hinge ear is developed in the left valve. The ornament of longitudinal ribs is increased in length and prominence and an anterior marginal rib is introduced. The muscle node is also subdivided. The duplicature doubles in width and the radial canals become longer and curve upwards antero- dorsally. The hinge is strengthened and changes from merodont to amphidont. The synonomy only includes references with figures or those of which the author has seen the actual specimens. Homocythere reticulata Kaye is here tentatively included in the synonomy. Its intimate occurence with pre-adults of V. harrisiana where adults are absent may indicate that it is a form of the adult found under unusual ecological conditions. Family TRACHYLEBERIDIDAE Genus CYTHEREIS Jones 1849 Cythereis corrigenda nom. nov. (Pl. 7, figs. 6, 9) 1940 Cythereis rudispinata Chapman & Sherborn ; Triebel, 200, pl. 4, figs. 47-50. 1956 Cythereis vudispinata Chapman & Sherborn ; Deroo : 15106. MaTerRIAL. B.M.N.H., Io. 1198, Io. 1616-17, from the Lower Gault, dentatus Zone, Culham, Oxfordshire. MEASUREMENTS Length Height Male left valve (B.M.N.H., Io. 1617) . . 0-87 mm. 0-42 mm. Female left valve (B.M.N.H., Io. 1198) 0-7-7, mim. 0-42 mm. Remarks. Triebel’s specimens differ significantly in shape, size, nature and distribution of ornament, marginal features and hingement from Chapman & Sherborn’s specimens (1893 : 248, pl. 14, figs. 6, 7). The latter specimens falling within the genus Protocythere. Triebel’s form has therefore been given a new name. Cythereis matronae Damotte & Grosdidier (1963) is very similar to C. corrigenda, but it is stated to differ in its greater size and lack of reticulation. Triebel’s speci- mens, however, are of similar size and have smooth intercostal areas. BRITISH MARINE CRETACEOUS OSTRACODA 63 Cythereis folkstonensis nom. nov. (Pl. 7, figs. 1-5) 1849 Cythere (Cythereis) quadrilatevata (Roemer) Jones : 18, pl. 3, figs. 10a—c, e-f (non pl. 3, fig. rod ; pl. 4, figs. g—h). 1870 Cythere quadrilaterata (Roemer) ; Jones : 75, 76. 1890 Cytheveis quadrilatevata (Roemer) ; Jones & Hinde : 20, pl. 1, figs. 69-71, 74-75. Diacnosis. Large Cythereis with three longitudinal rows of tubercles. Lateral surface devoid of reticulation. Eye tubercle and muscle node prominent. Lectotype. B.M.N.H., In. 51678 figured Jones (1849, pl. 3, fig. 10a) from the Gault Clay at Folkestone, here designated. PARALECTOTYPES. B.M.N.H., In. 51679-80, In. 51682~-83 figured Jones (1849 pl. 3, figs. 10), c, e, f) from the Gault Clay at Folkestone. OTHER MATERIAL. B.M.N.H., Io. 1192, Io. 1614-15, from the Upper Gault at Burwell, Cambs. MEASUREMENTS Length Height Left valve (B.M.N.H., In. 51678, lectotype) . 1:12 mm. 0-60 mm. ReMARKS. This species having been wrongly attributed by Jones is here renamed. C. folkstonensis appears to be restricted to the Albian, most of the specimens from the Chalk are referable to such species as C. lurmannae, C. cornueli, Veemia harrisiana, ete. The most diagnostic features of C. folkstonensis are the rows of tubercles along the longitudinal ribs ; the separation of the median longitudinal rib and the muscle node ; the smooth intercostal areas. Cythereis folkstonensis is most closely related to Cythereis glabrella Triebel but is less inflated and has spines on the longitudinal tibs. It is likely that it is related to its reticulate counterpart Cytherets reticulata (Jones & Hinde). Cythereis lonsdaleiana Jones (Ble less 7010) 1849 Cythere (Cythereis) lonsdaleiana Jones : 20, pl. 5, figs. 12a, b (non fig. 12c). 1870 Cytheve lonsdaleiana Jones : 75, 76. 1880 Cythere filicosta Marsson : 43, pl. 3, figs. 12a, b. 1890 Cythereis lonsdaleiana Jones ; Jones & Hinde : 27, pl. 1, figs. 64, 65. 1941 Cythereis filicosta (Marsson) ; Bonnema : 132, pl. 4, figs. 48-53 (non pl. 7, figs. 55-58). Diacnosis. Cythereis with keel-like longitudinal ribs. Dorsal rib formed of series of short oblique cross ribs. Median rib short and joined to large smooth muscle node. LectotypPe. B.M.N.H., In. 39012 figured Jones (1849, pl. 5, fig. 12b) from the Upper Chalk at Norwich, here designated. PARALECTOTYPE. B.M.N.H., In. 39011 figured Jones (1849, pl. 5, fig. 12a) from the same locality. OTHER MATERIAL. B.M.N.H., Io. 1196 and Io. 1618 from the Upper Chalk at Norwich. MEASUREMENTS Length Height Right valve (B.M.N.H., In. 39012, lectotype) 0-63 mm. 0:37 mm. 64 BRITISH MARINE CRETACEOUS OSTRACODA REMARKS. Jones’ pl. 5, fig. 12c (B.M.N.H., In. 39013) is of a much larger, differently ornamented form. Marsson’s species Cythere filicosta as redescribed after examination of the original types by Bonnema (1941) fits Jones’ original description exactly. The multiple nature of the dorsal rib was the most diagnostic feature according to Bonnema who used this criterion for separating the form from Cythereis semiplicata (Reuss). The interior of the valves are shallow, with a wide duplicature. The hinge is strongly amphidont with high, weakly lobed terminal teeth in the right valves. Specimens referred to this species by Chapman and other authors from the Gault Clay of S.E. England (B.M.N.H., I. 2683) are instars of Cythereis reticulata and allied forms. Cythereis macrophthalma (Bosquet) (Pl. 6, figs. 12-15, 17) 1847 Cypridina macrophthalma Bosquet : 16, pl. 3, figs. 3a—d. 1936 Cythereis macrophthalma (Bosquet) Veen : 7, pl. 2, figs. 43-48. 1958 Cythereis macrophthalma (Bosquet) ; Howe & Laurencich : 212. MATERIAL. (i) B.M.N.H., Io. 354 from the Chalk Rock at Dunstable ; B.M.N.H., Io. 351 from the Chalk at Norwich ; B.M.N.H., Io. 345 from the Chalk at Colchester all mounted by Jones & Hinde (1890) as C. quadrilaterata. (ii) B.M.N.H., lo. 1193 from the Upper Chalk at Norwich. MEASUREMENTS Length Height Male carapace (B.M.N.H., Io. 345) . 0-85 mm. 0:47 mm. Female left valve (B.M.N.H., lo. 345). Org mnie 0-47 mm. REMARKS. Cythereis quadrilaterata sensu Jones appears to be confined to Albian sediments and the specimens from the Chalk so labelled by Jones are all referable to C. macrophthalma. Bosquet’s original figures are so bad that identification is here largely based on Veen’s illustrations. In shape the species is somewhat akin to Protocythere or Veenia but possesses typical hinge and marginal features of the genus Cythereis. The intercostal areas are smooth and the muscle node is separated from the median rib. The species lacks the spination and inflation of true Cytherers quadrilaterata. Cythereis ornatissima s.]. (Reuss 1846) (EUS S higSack, Zen) 1846 Cytherina ornatissima Reuss : 104, pl. 24, figs. 12, 18. 21846 Cytherina ciliata Reuss : 104, pl. 24, fig. 17. ?1874 Cytheve orvnatissima Reuss : 146, pl. 2, figs. 5, 6. ?1887 Cythere ornatissima Reuss ; Kafka : 15, fig. 30. 1963 Cythereis ovnatissima (Reuss) and subspecies ; Pokorny : 8-26, pl. 1, figs. 1-3 ; pl. 2, fig. 1; pl. 3, fig. 3; pl. 4, figs. 1-9 ; pl. 6, figs. 1, 2, 5,6; pl. 7, fig. 3. MATERIAL. (i) B.M.N.H., In. 53097, 53164, 53266, 53272, Ilo. 1622-25 (Rowe) from Upper Chalk, Norwich. REMARKS. This species has been greatly confused in the past and large number of forms have been referred to it. BRITISH MARINE CRETACEOUS OSTRACODA 05 Triebel (1940) figured topotypic material from the Turonian of Bohemia and showed that forms attributed to this species by the majority of the early authors were almost without exception quite different. Jones (1849, plate 4, figs. 11a—h) figured a form which he referred to as Cytherevs ciliata (Reuss). In a later paper (1870) he stated that Reuss had decided that the two species C. ciliata and C. ornatissima were con-specific and that the latter had preference. Jones & Hinde (1890) therefore referred to the 1849 material under C. ovnatissima (Reuss). Jones’ (1849) figured material came from the Detritus at Charing and the Gault at Folkestone but he recorded the species from other Cretaceous horizons (Chalk marl). The specimens figured by Williamson (1847) as Cythere echinulata were also included by Jones & Hinde (1890) in the synonomy. Jones states that C. orvnatissima is most abundant in the Gault Clay. Between 1849 and 1890 a wide variety of forms were figured as C. ornatissima (Reuss) and in 1890 Jones & Hinde separated off five subspecies leaving Jones, 1849 material as C. oynatissima s.s._ A form previously described by Jones (1849, pl. 5, figs. 13a-d) as Cythereis cornuta (Roemer) was included as a subspecies under the name Cytherets ornatissima nuda. Later work, particularly by Triebel (1940) has separated off many of those later forms, refiguring some and erecting new species for others. Jones and Jones & Hinde’s specimens belong to a wide range of forms but lack of material makes the bulk of the varieties questionable. Dr. Triebel of the Senckenberg Museum, Frankfurt am Main, has kindly sent me a topotypic specimen of his 1940 published material. On examination it was found that though none of the figured specimens attributed to this form by Jones belongs there, some of the specimens from the British Museum collection are comparable. These are included in slides from the Dunstable Chalk (I. 2466, Io. 388, Io. 1626) and from the Upper Chalk of Keady Hill, N. Ireland (Io. 375). Triebel’s specimens have now been referred to Cythereis longaeva longaeva by Pokorny (1963). A number of specimens from the Rowe Norwich collections are larger and more spinose than Triebel’s and probably belong to Cythereis ornatissima s.l. They are in slides B.M.N.H., In. 53097, 53164, 53266, 53272 and Io. 1622-25. Further study of the distribution in these and related forms throughout the Chalk are required before a final decision can be made. All the species of Cythereis from the British Upper Cretaceous show a wide variability of ornament ; particularly is the emphasis of the reticulation and spination. A thorough investigation of large numbers of specimens to determine the variability of the ornament is needed in most cases. Recently Pokorny (1963), working with topotypic material and some of Reuss’ original material now deposited in the Natural History Museum, Vienna, has completely revised Cytherets ornatissima (Reuss) erecting two new species and three subspecies. The limited nature of this latter material does not entirely clarify the situation and in his opinion the two forms C. ornatissima and C. ciliata are by no means definitely conspecific. The wide variety of forms grouped by Reuss into C. ornatissima show that a large number of specimens are necessary for accurate study and the limited material of Jones is hard to place into Pokorny’s excellent systematic divisions. Triebel’s specimen (PI. 8. 66 BRITISH MARINE CRETACEOUS OSTRACODA fig. 5) is certainly C. longaeva longaeva and the Jones material from Dunstable and Keady Hill (Pl. 7, fig. 12, Pl. 8, fig. 3) seems closest to that form. Rowe’s specimens are closest to C. ornatissima altinodosa in lacking surface reticulation and matching well in the tuberculation and ribbing. Cythereis ornatissima paupera Jones & Hinde 1890 Cytheveis ovnatissima paupera Jones & Hinde : 23, pl. 2, figs. 10, 11. REMARKS. There is no trace of the figured material for this variety. The two slides of material from Dunstable in the Jones (1890) collection do not appear to resemble the figured specimens. These slides (B.M.N.H., I. 2466 and Io. 388) contain specimens of Cythereis ornatissima (Reuss) together with Cythereis glabrella Triebel. From the figure C. ovnatissima paupera appears to be similar to Cytherets nuda or Cythereis lurmannae and possibly it is conspecific with one of these forms. The spinose outline makes it unlikely that this species could be C. glabrella and the absence of surface reticulation distinguishes it from C. ovnatissima s.s. Cythereis ornatissima radiata Jones & Hinde 1890 Cythereis ornatissima vadiata Jones & Hinde : 25, pl. 4, fig. 13. ReEMARKS. There is no trace of Jones’ original specimen of this form nor is there any comparable material from the Cambridge Greensand. From an inspection of residues from the Cambridge Greensand (Barrington) I have found no form re- sembling the figure. The Mockler collection (1909 ; B.M.N.H.) of ostracoda from the Cambridge Greensand contains a number of slides labelled C. ornatissima (Reuss). These slides (B.M.N.H., In. 53344-56) contain a wide variety of forms : Cythereis veticulata (Jones & Hinde), C. lwymannae Triebel, C. thorenensis Triebel, C. folkstonen- sis, C. bonnemai Triebel, [socythereis fortinodis Triebel and Veenta harrisiana (Jones). From Jones & Hinde’s figure it seems likely that C. ornatissima radiata is a worn C. thorenensis. Cythereis lurmannae Triebel (Pl. 8, figs. 11-15) 1890 Cythereis ornatissima vax. stvicta Jones & Hinde : 25, pl. 1, fig. 63. 1940 Cythereis lurmannae Triebel : 201, pl. 6, figs. 63-66. 1956 Cythereis luvmannae Triebel ; Deroo : 1510. MATERIAL. (i) B.M.N.H., Io. 346 (Jones & Hinde 1890) from the Chalk marl at Didcot. (ii) B.M.N.H., In. 39007-08 figured Jones (1849; C. ciliata Pl. 4, figs. 11g, g’) from the Gault at Folkestone. (i) B.M.N.H., In. 51686-88 figured Jones (1849 ; C. cornuta pl. 5, figs. 13c, d) from the Detritus at Charing. (iv) B.M.N.H., Io. 1194, lo. 1629-33, from the Upper Gault at Barrington, Cambs. RemMARKS. The single figured specimen of Jones & Hinde (1890) is not identified as such but a slide in the Jones collection from Didcot (B.M.N.H., Io. 346) appears to contain this or a strictly comparable form together with three C. thorenensis Triebel and one C. reticulata Jones & Hinde. The specimen is now found to belong to BRITISH MARINE CRETACEOUS OSTRACODA 07 / Cythereis lurmannae Triebel. Jones, however, has figured C. stricta in so many different contexts that Triebel’s name is much better upheld, particularly as Jones described the specimen as a variety and not a subspecies. The specimen labelled by Chapman (1893) as C. stricta (B.M.N.H., I. 2687) belongs to Cythereis bonnemai Tniebel. Cythereis nuda Jones & Hinde (Wedh, 7, 11S, bers, ane, 3d5)) 1849 Cythere (Cythereis) lonsdaleiana Jones : 20, pl. 5, fig. 12¢ (non fig. 12a, b). 1849 Cythereis covnuta (Roemer) ; Jones : 21, pl. 5, fig. 13b (non figs. 13a, c, d). 1849 Cythereis ciliata (Reuss) ; Jones: pl. 2, fig. 11h’. 1890 Cythereis ornatissima nuda Jones & Hinde : 23, pl. 2, fig. 9 (nom figs. 8, 12-14). 1893 Cythereis wrighttt Jones & Hinde ; Chapman : 370. 1898 Cythereis ovnatissima nuda Jones & Hinde ; Chapman : 339. 21956 Cythereis nuda Jones & Hinde ; Deroo : 1519, pl. 4, figs. 62-64. LectotyPe. B.M.N.H., In. 51685 figured Jones (1849, pl. 5, fig. 13) from the Detritus at Charing, here designated. Remarks. As the bulk of the specimens previously referred to C. ornatissima nuda can be attributed to well known species, the remaining specimen (B.M.N.H., In. 51685) is taken as lectotype of Cythereis nuda. The additional material figured by Jones & Hinde (1890, pl. 1, fig. 76; pl. 4, fig. 14) appears to be lost and cannot, therefore, be determined. A specimen attributed to Jones and labelled var. nuda from Keady Hill (B.M.N.H., Io. 376), differs considerably from the earlier forms, belonging either to a new species of Cythereis or to C. wrighttt Jones & Hinde. A specimen in the Chapman collection from the Gault at Folkestone (B.M.N.H., I. 2685) is a young form of C. reticulata but the form described by Chapman (1898) from the Cambridge Greensand is a true C. nuda (Sedgwick Museum B.40597). Further specimens of C. nuda are those mentioned by Chapman (1893) as Cythereis wrightit Jones & Hinde from the phosphatic Chalk at Taplow, B.M.N.H., I. 2607. True C. wrightit was described from Keady Hill (Wright collection) and appears to approximate to the form described as C. ornatissima nuda, slide B.M.N.H., Io. 376 (see above), the figure, however, is that of a right valve whilst the British Museum specimen is a left valve. The absence of the figured specimen precludes further study of that species. The form described by Chapman & Sherborn (1893 (pl. 14, fig. 9) as C. wrightii var. aculeata (B.M.N.H., I. 2707) is a pre-adult of Cythereis reticulata Jones & Hinde. Jones’ specimen of Cythereis lonsdaletana, B.M.N.H., In. 39013 (pl. 5, fig. 12c) from the Upper Chalk of Norwich is also Cythereis nuda. Cythereis reticulata Jones & Hinde (Pl. 8, figs. 16-19) 1890 Cythereis ornatissima reticulata Jones & Hinde : 24, pl. 1, fig. 68, ; pl. 4, figs. 9-12. 1940 Cythereis reticulata Jones & Hinde ; Triebel : 192, pl. 5, figs. 51-56. 1950 Cythereis reticulata Jones & Hinde ; Deroo : 1518, pl. 5, figs. 68-82. MATERIAL. (i) B.M.N.H., Io. 1195, Io. 1634-37, from the Lower Gault Clay, Culham, Oxfordshire. 68 BRITISH MARINE CRETACEOUS OSTRACODA REMARKS. There is no trace of the figured material of this form. The only labelled specimen in the Jones collection is one from Keady Hill (not among his original localities) which is now seen to be of C. ornatissima s.l. Subsequent authors (Triebel, Deroo) have established the true nature of the species by reference to Albian forms which have the median longitudinal rib well developed. Of Jones’ figures, pl. 1, fig. 68 and pl. 4, figs. 9-12 fit the species best in its now accepted sense. Such specimens are found commonly in the Gault Clay from which Jones inspected material and the species is now restricted to forms of Albian and Cenomanian age. References to the species from higher horizons are most likely to be of C. ornatissima s.s. Two slides from the Chapman (1893) collection from the Gault Clay Folkestone (B.M.N.H., I. 2686, I. 2684) contain Cythereis thorenensis Triebel. The records by Chapman (1893, 1898) and by Weber (1934) do not belong to C. reticulata. Cythereis hirsuta described by Damotte & Grosdidier (1963) is very similar to C. reticulata but differs in the prominence of the median rib and muscle node in the latter. The spination of the ribs is most pronounced in the former. Cythereis thorenensis Triebel (PIS tesa Aes) 7) 1849 Cythereis ciliata (Reuss) ; Jones : 19, pl. 2, figs. 11a—f (non figs. 11g, h). 1870 Cythereis ornatissima (Reuss) ; Jones : 75. 1890 Cythereis ornatissima (Reuss) ; Jones & Hinde : 21, pl. 2, figs. 1-5. 1940 Cythereis thovenensis Triebel : 195, pl. 5, figs. 57-59. MATERIAL. (i) B.M.N.H., In. 39001—06 figured Jones (1849, pl. 2, figs. I11a—d), figs. I1a—e from the Detritus at Charing, fig. 11f from the Gault clay at Folkestone. (i) B.M.N.H., Io. 1197, Io. 1619-21 from the Upper Gault at Maidstone, Kent. REMARKS. Most of Jones’ original specimens attributed to C. ornatissima from the Gault Clay and Detritus (Albian—Cenomanian) belong to Triebel’s species. Genus PLATYCYTHEREIS Triebel 1940 Platycythereis gaultina (Jones) (Pl. 8, fig. 9) 1849 Cythere (Cythereis) gaultina Jones : 17, pl. 2, figs. 7a-c. 1870 Cytheve gaultina Jones ; Jones : 75, 76. 1890 Cytheve gaultina Jones ; Jones & Hinde : 18, pl. 1, figs. 35, 36. 1893 Cythere gaultina Jones ; Chapman & Sherborn : 346. 1893 Cythereis excavata Chapman & Sherborn : 348, pl. 14, fig. 8. 1898 Cytherve gaultina Jones ; Chapman : 336. 1940 Platycythereis gaultina (Jones) Triebel : 219, pl. 7, figs. 81-85 ; pl. 8, figs. 86, 87. 1956 Platycythereis gaultina (Jones) ; Mertens : 209, pl. 11, figs. 59, 60. D1aGNosis. Small Platycythereis with strongly reticulate lateral surface. Complex anterior marginal rib but no longitudinal ribs. Hook-like process present over region of muscle scars. LectotyreE. B.M.N.H., In. 52631 figured Jones (1849, pl. 2, fig. 7a) from the Gault Clay at Folkestone, here designated. BRITISH MARINE CRETACEOUS OSTRACODA 69 OTHER MATERIAL. (i) B.M.N.H., In. 52632 and In. 51664 figured Jones (1849, pl. 2, figs. 7b, c) from the Gault Clay at Folkestone. (ii) B.M.N.H., I. 2688, I. 2706 (Chapman & Sherborn 1893 ; latter figured pl. 14, fig. 8). Gault Clay, Folkestone. REMARKS. This well known species occurs throughout the Albian. No description further to that of Triebel (1940) is required. The form figured by Chapman & Sherborn (1893) as Cythereis excavata is conspecific with P. gaultina and is therefore included in the synonymy. The specimen figured by Chapman (1808, text-figs. 2a, b) which has been taken as a typical P. excavata by later authors (Triebel 1940) differs fundamentally from the initial figures and is renamed below. Platycythereis chapmani nom. nov. (PI. 6; figs: 16, 18, 20) 1898 Cythere gaultina excavata (Chapman & Sherborn) ; Chapman : 336, text-figs. 2a, b. 1898 Cythere subtuberculata Chapman : 337, text-figs. 3a, b. 1940 Platycythereis excavata (Chapman & Sherborn) ; Triebel : 315, pl. 7, figs. 78-80 ; pl. Io, fig. 110. DiacGnosis. Large Platycythereis with prominent keel-like anterior marginal rib joined dorsally to the eye tubercle and ventrally to the ventral longitudinal rib. Lateral surface compressed and strongly reticulate. HoLotyrPe. A right valve, Sedgwick Museum B4o6109, figured Chapman (1808, Text-figs. 2a, b) from the Cambridge Greensand of Swaffham, Cambs. OTHER MATERIAL. (i) S.M.B. 40620 figured Chapman (1808, text-figs. 3a, b) from Swaffham. (ii) B.M.N.H., Io. 1201, Io. 1612-13 from the Cambridge Greensand at Barrington, Cambs. Remarks. As the initial specimens attributed to this species now prove to be wrongly identified a new name is required for the later forms. Chapman’s (1898) later figured specimen is taken as holotype. Due to the rarity of the species at Swaffham, Chapman evidently did not recognise the form he figured as C. subtuberculata to be merely an instar of P. chapman. P. chapmani is very similar to P. laminata Triebel, figures of which are included here for comparison. The major differences are that in P. Jaminata there is no rib along the ventral surface and the anterior marginal rib is not as distinct and is not continued antero-dorsally to join the eye tubercle. P. Jaminata is more triangular in shape, particularly the left valve ; the dorsal marginal rib is also strongly developed. Platycythereis laminata Triebel (Pl. 6, fig. x9) 1940 Platycythereis laminata Triebel : 217, pl. 8, figs. 88-90. 1956 Platycythereis laminata Triebel ; Deroo : 1520. MaTERIAL. B.M.N.H., Io. 1198 and Io. 1611 from the Lower Gault at Henfield, Sussex. 70 BRITISH MARINE CRETACEOUS OST RACODA Genus TRACHYLEBERIDEA Bowen 1953 Trachyleberidea acutiloba (Marsson) (BIS Sriligsw75eCunkO) 1880 Cythere acutiloba Marsson : 42, pl. 3, fig. 11. 1890 Cythereis spinicaudata Jones & Hinde : 28, pl. 2, figs. 17, 18. 1940 Cythereis acutiloba (Marsson) Bonnema : 132, pl. 4, figs. 59—66. MATERIAL. (i) B.M.N.H., I. 2487 figured Jones & Hinde (1890, pl. 2, fig. 17) from the Upper Chalk, Keady Hill. (1) B.M.N.H., Io. 359, Ilo. 1627-28 (Jones & Hinde 1890) from the Chalk Rock, Dunstable. (iii) B.M.N.H., lo. 1209, from the Upper Chalk, coranguinum Zone, Sonning. MEASUREMENTS. Length Height Left valve (B.M.N.H., I: 2487). : . 0°65 mm. 0-35 mm. Remarks. As suggested by Bonnema (1940) Jones & Hinde’s specimens appear to be conspecific with Marsson’s. The shape and internal features of the species, however, fall within Haskin’s (1963) redefinition of the genus Tvachyleberidea. The strongly convergent dorsal and ventral margins and lateral compression are the most distinct features of the species. The strong reticulation, weak dorsal, ventral and anterior marginal ribs together with the low muscle node are also well seen. The hinge is strongly amphidont, having the anterior tooth smooth and the posterior tooth divided in the right valve. Suborder PLATYCOPINA Family CYTHERELLIDAE Genus CYTHERELLOIDEA Alexander 1929 Cytherelloidea chapmani (Jones & Hinde) (Pl. 9, figs. 15-19, 22) 1890 Cytherella chapmani Jones & Hinde : 49, pl. 3, fig. 70. 1893 Cytherella chapman Jones & Hinde ; Chapman & Sherborn : 340. ?1898 Cythervella chapmani Jones & Hinde ; Chapman : 345. 21956 Cytherelloidea chapmani (Jones & Hinde) Deroo : 1909. Diacnosis. Cytherelloidea with the dorsal longitudinal rib connected to the anterior end of the median rib. LecToTyPeE. B.M.N.H., Io. 1641 (Chapman 1893) from the Lower Gault, Folkestone. OTHER MATERIAL. (i) B.M.N.H., I. 2669, I. 2671 (Chapman) from the Lower Gault, Folkestone. (ii) B.M.N.H., Io. 1293, Io. 1642-46, from the Lower Gault, Culham, Oxon. MEASUREMENTS. Length Height Left valve (B.M.N.H., lo. 1641, lectotype) . 0-52 mm. 0-29 mm. Right valve (B.M.N.H., Io. 1642) ; . 0-60 mm. 0:37 mm. DESCRIPTION. Carapace elongate, subrectangular in lateral view. Lateral surface covered with a series of inflated ribs. A high, anterior marginal rib, dis- continuous dorsally, is continued along the ventral margin as a low flat shelf. The valves are swollen posteriorly to form large connected postero-dorsal and postero- BRITISH MARINE CRETACEOUS OSTRACODA 71 ventral nodes. A long, high, slightly arcuate ventral longitudinal rib runs from the postero-ventral node, whilst a short, horizontal rib runs from the postero-dorsal node to terminate at two-thirds the valve length from the anterior margin. This rib is connected to a low oblique dorsal rib which traverses the central part of the dorsal margin. The dorsal rib is joined anteriorly to the anterior end of a prominent ventrally convex median rib. This latter rib is not joined posteriorly to either of the posterior nodes. The intercostal areas are smooth. Remarks. The original Jones & Hinde specimen of this species is now lost so the Chapman specimen mentioned by Jones is here erected lectotype. The lecto- type is a juvenile and adult specimens from Culham are figured here. C. chapmani most closely resembles C. pavawilliamsoni Kaye but differs in having the median and dorsal ribs joined anteriorly and posteriorly. The prominent median rib differentiates it from C. knaptonensis Kaye and C. stricta (Jones & Hinde). Cytherelloidea globosa sp. nov. (Pl. 9, figs. 7, 9, Io) DiacGnosis. An inflated species of Cytherelloidea with prominent vertical median sulcus limited laterally by longitudinal swelling. Hototyre. B.M.N.H., Io. 1283, a right valve from the Cambridge Greensand at Barrington, Cambs. PaRATYPES. B.M.N.H., Io. 1284-87 from the same locality. MEASUREMENTS. Length Height Right valve (B.M.N.H., Io. 1283, holotype) 0:55 mm. 0:34 mm. Left valve (B.M.N.H., Ilo. 1284, paratype) . 0-53 mm. 0-30 mm. DEscRIPTION. Valves small, elongate, subrectangular. Dorsal and ventral margins straight and subparallel ; anterior and posterior margins semicircular. Lateral surface inflated but divided into two halves by a deep, prominent median sulcus. Below the sulcus lies a large smooth elongated node. The anterior lateral area is strongly and evenly inflated. The posterior lateral area bears two large nodes, the dorsal one being larger and more elongate than the ventral one. The postero-ventral node is connected by a swollen area to the postero-dorsal node but is separated from the ventral node by a prominent depression, which runs obliquely to join the median sulcus. In certain specimens the postero-ventral lobe is not developed and is possibly a dimorphic feature. REMARKS. The strong sulcus, anterior and posterior inflation and lack of well- defined ribs distinguish this species from other described forms. Cytherelloidea granulosa (Jones) (PI. 9, figs. 24-26) 1849 Cytherella williamsoniana var. granulosa Jones : 31, pl. 7, fig. 267. 1880 Cytherella williamsoniana bosqueti Marsson : 33, pl. 2, figs. 8d, e. 1890 Cytherella williamsoniana var. granulosa Jones ; Jones & Hinde : 49, pl. 3, figs. 68, 69, 72. 1940 Cytherelloidea williamsoniana (Jones) ; Bonnema : 95, pl. I, figs. 44-47. Diacnosis. Large Cytherelloidea with lateral surface covered with a series of 72 BRITISH MARINE CRETACEOUS OSTRACODA prominent pustules. Short separate dorsal and ventral longitudinal ribs are present but no median rib. Anterior marginal rib prominent. Lectotype. B.M.N.H., In. 51609 figured Jones (1849, pl. 7, fig. 267) from the Upper Chalk, Norwich. OTHER MaTERIAL. (i) B.M.N.H., I. 2484 (Jones & Hinde 1890), Upper Chalk, Magheramorne, Antrim. (ii) B.M.N.H., In. 53110, In. 53232, Io. 1647-48 (Rowe), from the Upper Chalk, Norwich. MEASUREMENTS. Length Height Right valve (B.M.N.H., In. 51609, lectotype) 0-80 mm. 0-45 mm. DeEscRIPTION. Valves quadrangular in shape with straight parallel dorsal and ventral margins and semicircular anterior and posterior margins. A high, semi- circular anterior marginal rib occurs which is often connected to a postero-ventral node by a flattened marginal shelf. A further large node occurs postero-dorsally which is somewhat elongated along the dorsal margin. The two posterior nodes are entirely separated. A low ventrally arcuate rib lies in line with the lower of these nodes but is not connected to it. A shallow muscle pit occurs centrally with a small culmination immediately above it on the dorsal margin. Except for the ribs and posterior nodes the whole of the lateral surface is covered with a series of prominent pustules. Juveniles are fairly common and have the ribbing subdued but maintain the strong pustulation. REMARKS. This species, originally described as a variety of C. williamsomiana by Jones, is characteristic of the Upper Chalk, and has not been found by the author below the cor-anguinum Zone. The strong pustulation makes the species distinct from others of the genus found in the Chalk. Specimens of Marsson’s C. williamsoniana var. bosqueti from Rugen, kindly sent to the author by Dr. E. Herrig, show that the latter is conspecific with C. granulosa. Cythereolloidea hindei sp. nov. (BINopstigss 45.8510) Diacnosis. Cytherelloidea, with anterior marginal rib, ventral longitudinal rib and lower sinuous dorsal longitudinal rib all connected. Hototyre. B.M.N.H., Io. 1288 a right valve from the Upper Chalk at Norwich. ParaTyPEs. B.M.N.H., Io. 1289-92, from the same locality. OTHER MaTeERIAL. (i) B.M.N.H., Io. 344. Chalk, Colchester ; Io. 339, Chalk, Luton. (ii) B.M.N.H., In. 53140, In. 53234. (Rowe), Upper Chalk, Norwich. MEASUREMENTS. Length Height Right valve (B.M.N.H., Io. 1288, holotype) 0°75 mm. 0-42 mm. Left valve (B.M.N.H., Io. 1289, paratype) . 0-68 mm. 0°35 mm. DEscRIPTION. Valves elongate, subrectangular in shape. Dorsal and ventral margins straight and parallel ; anterior and posterior margins evenly rounded. Surface ornamented by a series of inflated ribs. Two large nodes joined by a short, high connecting rib occur posterior-dorsally and postero-ventrally. A high, straight rib runs along the ventral margin from the postero-ventral node and is BRITISH MARINE CRETACEOUS OSTRACODA 73 continued without a break into a prominent anterior marginal rib. A low, sinuous rib runs along the dorsal margin being connected to the anterior marginal rib anteriorly and weakly joined by a cross rib to the postero-dorsal node posteriorly. A ventrally convex median rib runs below the muscle scar pit being entirely separated both anteriorly and posteriorly. The intercostal areas are smooth. In the larger right valves a low flattened area lies between the dorsal rib and the margin but in both valves the ventral rib is not separated from the margin. Remarks. This species is easily distinguished by the connection of the ventral, anterior and dorsal ribs and the concurrence of the ventral rib and the margin. Cytherelloidea knaptonensis Kaye (RISO aigss 20.210) 1963 Cytherelloidea knaptonensis Kaye : 114, pl. 19, figs. 10-12. MATERIAL. B.M.N.H., lo. 1297 from the Upper Gault at Leighton Buzzard. REMARKS. This species occurs in the Gault Clay at various levels at Speeton, Leighton Buzzard and Burwell. It is closely related to C. parawilliamsoniana Kaye and C. chapmani (Jones & Hinde). It differs from them in the poor development of the median rib. Cytherelloidea oblinquirugata (Jones & Hinde) (Pl. 9, figs. 12-14) 1890 Cytherella oblinquirugata Jones & Hinde : 50, pl. 3, fig. 73. MATERIAL. B.M.N.H., Io. 1299, Io. 1638—40 from the Upper Chalk at Norwich. MEASUREMENTS. Length Height Adult left valve (B.M.N.H., Io. 1640) . O75 mm. 0-40 mm. Pre adult right valve (B.M.N.H., Io. 1299) . 0-50 mm. 0-34 mm. Pre adult left valve (B.M.N.H., Io. 1299) . 0-50 mm. 0-30 mm. REMARKS. This species was originally erected on a juvenile specimen and consequently requires revision. Throughout the ontogeny of this species the median longitudinal rib becomes increasingly less prominent and at maturity is barely discernible. Sexual dimorphism is expressed by the existence of two large nodes postero-dorsally and postero-ventrally in the females. The characteristic flattened shelf along the ventral margin and the sinuous nature of the dorsal rib are the most characteristic features of the species. The ventral rib is also more strongly joined than in related forms such as C. williamsoniana, particularly in the males and juveniles. Cytherelloidea parawilliamsoniana Kaye (Pl. 9, fig. 23) 1963 Cytherelloidea pavawilliamsoniana Kaye : 115, pl. 20, figs. 22, 23. REMARKS. This species is strongly allied to the other members of the genus found in the Gault. It has not been found by the writer at any locality outside Yorkshire. It differs from the other forms in having a strong median rib which is separated anteriorly. 74 BRITISH MARINE CRETACEOUS OSTRACODA Cytherelloidea stricta (Jones & Hinde) (Pl. 9, figs. 1-3, 5, 6) 21847 Cytherina servata Williamson : 79, pl. 4, fig. 79. 1849 Cytherella wiliamsoniana Jones : 31, pl. 7, figs. 26a—d, g, h (non figs. 26e, f). 1890 Cytherella williamsoniana Jones ; Jones & Hinde : 48, pl. 3, figs. 57-62. 1890 Cytherella williamsoniana stricta Jones & Hinde : 48, pl. 3, fig. 71. 1893 Cytherella williamsoniana stricta Jones & Hinde ; Chapman & Sherborn : 346. 1956 Cytherelloidea stricta (Jones & Hinde) ; Deroo : 1509, pl. 1, figs. 7, 8. 1958 Cytherelloidea stvicta (Jones & Hinde) ; Howe & Laurencich : 270. 1963 Cytherelloidea stricta (Jones & Hinde) ; Kaye : 117, pl. 19, figs. 14, 15. DraGnosis. Cytherelloidea with prominent straight dorsal and ventral longitu- dinal ribs but no median rib. The ventral rib is separate both anteriorly and posteriorly and set off from the margin by a shelf formed as a continuation of the anterior marginal rib. Dorsal rib joined to postero-dorsal process. Lectotype. B.M.N.H., In. 51604 figured Jones (1849, pl. 7, fig. 26c) Gault Clay, Folkestone. OTHER MATERIAL. (i) B.M.N.H., In. 51602-03, In. 51605, In. 51608, figured Jones (1849, pl. 7, figs. 26a, 6, d, h) Gault Clay, Folkestone. (ii) B.M.N.H., I. 2762 (Chapman & Sherborn 1893), Gault Clay, Folkestone. (iii) B.M.N.H., Io. 1294, Upper Gault, Leighton Buzzard. MEASUREMENTS. Length Height Carapace (B.M.N.H., In. 51604, lectotype) . 0-71 mm. 0-37 mm. Carapace (B.M.N.H., In. 51605) ; 5 0°75 mama: 0:37 mm. Left valve (B.M.N. ee In. 51607) ‘ 0-73 mm. 0-37 mm. REMARKS. C. eA cpsowicead has a pareticalanlyy confused past and almost all Cretaceous species of what is now the genus Cytherelloidea were included in it by early authors. Jones included a variety of forms one of which pl. 7, fig. 26f, being the only clear external illustration was proposed as lectotype by Howe & Laurencich (1958). Unfortunately the specimen relating to this figure was lost long before 1958 and the concept of the species becomes nomina dubium being restricted to the single figure without specimens. The specimen from fig. 26e has been separated off into a new species by Kaye (1963) and the rest of the material, which is found to be conspecific recognised as C. stricta (Jones & Hinde) 1890. Thus the specimens from Jones (1849, pl. 7, figs. 26a—d, g, h) though originally defined as C. williamsoniana must now be withdrawn from that species in its restricted sense and included in C. stricta. Due to the absence of material from the Jones & Hinde 1890 collection one of these specimens is here erected lectotype. The species is rather variable in the strength but not in the distribution of the ribbing. Sexual dimorphism is shown by greater inflation of the posterior ead of the valves in females and such differences may have been thought to be valid varietal differences by Jones & Hinde in their erection of var. stricta. The most marked features of the species are the absence of a median longitudinal rib, the long, straight dorsal rib and the short, arcuate ventral rib which is entirely separated from the posterior nodes. The strong anterior marginal rib is continued as a shelf along the ventral margin. The posterior margin bears a series of small BRITISH MARINE CRETACEOUS OSTRACODA 75 tubercles. As C, williamsoniana, which was made the type species of the genus (Alexander 1929), is a nomina dubium it is necessary to seek I.C.Z.N. ratification of an alternative type species for Cytherelloidea. IV. SUMMARY The various species described by Jones and Chapman etc. are listed below in tabular form together with their new classification ; of the 98 different specific references some 55 specific names are considered valid. JONES 1849 : ORIGINAL IDENTIFICATION (1) Cythere hilseana (Roemer) (2) Cythere punctatula (Roemer) Cythevre punctatula var. virginea Cythere umbonata (Williamson) Cythere baivdiana sp. nov. Cythereis tviplicata (Roemer) Cythereis quadrilaterata (Roemer) Cythereis ciliata (Reuss) ) ) ) ) ) ) ) ) Cytherets corvnuta (Roemer) ) Cythereis alata (Bosquet) ) Bairdia siliqua sp. nov. . ) Bairdia siliqua var. « ) Bardia harrisiana sp. nov. ) Bairdia angusta (Munster) ) Cythereis interrupta (Bosquet) ) Cythereis gaultina sp. nov. ) ) (20) Cythere williamsoniana var. granulosa JONES 1870: ORIGINAL IDENTIFICATION (1) Cytheridea perforata (Roemer) (2) Cytheropteron concentricum (Reuss) (3) Cythere harvisiana sp. nov. (4) Cythereis ovnatissima (Reuss) JONES & HINDE 1890 : ORIGINAL IDENTIFICATION (1) Pontocypris trigonalis sp. nov. (2) Pontocypris bosquetiana sp. nov. (3) Pontocypris triquetra (Jones) (4) Macrocypris wrightii sp. nov. . (5) Macrocypris concinna sp. nov. (6) Bythocypris veussiana sp. nov. (7) Cythere harrisiana var. setosa . (8) Cythere havvisiana var. rveticosa (9) Cythereis auriculata (Cornuel) . Cythereis macrophthalma (Bosquet) . PRESENT IDENTIFICATION . Schulevidea jonesiana (Bosquet) Neocythere (N.) vanveeni Mertens + Neocythere (Centrocythere) denticulata Mertens . Neocythere (Physocythere) virginea (Jones) . Monoceratina umbonata (Williamson) ? Macrodentina sp. . Protocytheve lineata (Chapman & Sherborn) . Cythereis folkstonensis nom. nov. Cythereis thovenensis Triebel Cythereis lonsdaleiana Jones Cytheveis nuda Jones & Hinde . Alatacythere robusta (Jones & Hinde) . Macrocypris siliqua (Jones) . Macrocypris muensteriana Jones & Hinde . Pontocyprella harvisiana Jones . Dolocytheridea bosquetiana (Jones & Hinde) Veenia harrisiana (Jones) . Platycythereis gaultina (Jones) . Amphicytherura chelodon (Marsson) Cytherelloidea williamsoniana (Jones) + Cytherelloidea stricta (Jones & Hinde) Cytherelloidea granulosa (Jones) PRESENT IDENTIFICATION . Schulevidea jonesiana (Bosquet) . Neocythere (N.) vanveeni Mertens + Neocythere (C.) denticulata Mertens Veenia harvisiana (Jones) Cythereis thovenensis Triebel PRESENT IDENTIFICATION . Eucythere trigonalis (Jones & Hinde) . Dolocytheridea bosquetiana (Jones & Hinde) . Dolocytheridea bosquetiana (Jones & Hinde) . Macrocypris wrighti (Jones & Hinde) ? Macrocypris simplex Chapman . Dolocytheridea bosquetiana (Jones & Hinde) Veenia harvisiana (Jones) Veenia harvisiana (Jones) Veenia harvisiana (Jones) ) Cytherets ) Cythereis ) Cythereis ) Cythereis ) Cythereis ) Cythereis ) Cythereis ) Cythereis ) Cythereis ) Cythereis Cythereis Cythereis BRITISH MARINE CRETACEOUS OSTRACODA ornatissima paupera ovnatissima nuda ovnatissima reticulata ornatissima vadiata ornatissima stricta wrightti sp. nov. tubevosa sp. nov. tuberosa var. symmetrica icenica sp. NOv. icenica var. quadrata vallata sp. nov. spinicaudata sp. nov. No material . Cythereis nuda Jones & Hinde (in part) . Cythereis reticulata Jones & Hinde ? Cythereis thovenensis Triebel . Cythereis lurmannae Triebel Cythereis wrightit Jones & Hinde No material . No material . Amphicytherura chelodon (Marsson) No material . No material . Tvachyleberidea acutiloba (Marsson) . Brachycythere ci. sphenoides (Reuss) Cytheropteron alatum var. robusta Cythevopteron alatum var. fortis Cytheropteron alatum var. cornuta . Alatacythere robusta (Jones & Hinde) . Alatacythere vobusta (Jones & Hinde) ? Alatacythere vobusta (Jones & Hinde) ) . 2? Alatacythere robusta (Jones & Hinde) ) . Alatacythere phylloptera (Bosquet) ) . No material ) . Monoceratina montuosa (Jones & Hinde) ) Cytheropteron cuspidatum var. tricuspidata Monoceratina tricuspidata (Jones & Hinde) ) Cytheropteron pedatum (Marsson) . Monoceratina pedata pedata (Marsson) ) ) ) ) ) ) ) ) 2) Cytheropteron sphenoides (Reuss) ) ) ) Cytheropteron hibernicum sp. nov. Cytheropteron ? phyllopteron (Bosquet) Cytheropteron cuspidatum sp. nov. Cytheropteron cuspidatum var. montuosa 2) Cythevopteron pedatum salebrosa . Monoceratina pedata salebrosa (Jones & Hinde) Cytheropteron umbonatum acanthoptera . Monocervatina umbonatoides nom. nov. Cytheropteron umbonatum longispina . Monoceratina umbonata (Williamson) Cytheropteron sherbovm sp. nov. . Monoceratina sherborm (Jones & Hinde) Cytherella williamsoniana chapmani sp. nov. Cytherelloidea chapmani (Jones & Hinde) Cytherelloidea oblinquirugata (Jones & Hinde) CHAPMAN & SHERBORN 1893: ORIGINAL IDENTIFICATION (1) Cythere ? spinifera sp. nov. (2) Cythereis triplicata lineata (3) Cythereis vudispinata sp. nov. (4) Cythereis wrightii aculeata (5) Cytheridea votundata sp. nov. . (6) Cythereis excavata sp. nov. (7) Cytheridea perforata var. insignis (8) Cythere koninckiana (Bosquet) (9) Pseudocythere simplex (Jones & Hinde) CHAPMAN 1808: ORIGINAL IDENTIFICATION (1) Macrocypris simplex sp. nov. (2) Cythere gaultina var. excavata (3) Cythere subtuberculata sp. nov. . OTHER COMPARATIVE SPECIES: ORIGINAL IDENTIFICATION (1) = (2) Cythere slavantensis Veen (3) Cythere acanthoptera Marsson . (4) a PRESENT IDENTIFICATION . Schulevidea jonesiana (Bosquet) . Protocythere lineata (Chapman & Sherborn) . Protocythere rudispinata (Chapman & Sherborn) Cythereis veticulata (Jones & Hinde) . Schuleridea jonesiana (Bosquet) . Platycythereis gaultina (Jones) . Schuleridea jonesiana (Bosquet) Veenia harvisiana (Jones) Dolocytheridea bosquetiana (Jones & Hinde) PRESENT IDENTIFICATION Macrocypris simplex Chapman . Platycythereis chapmani nom. nov. . Platycythereis chapmani nom. nov. PRESENT IDENTIFICATION Macrocypris exquisita sp. nov. . Neocythere (Physocythere) virginea (Jones) . Monoceratina acanthoptera (Marsson) Monoceratina bonnemai sp. nov. BRITISH MARINE CRETACEOUS OSTRACODA Cythere longispina Bosquet Monoceratina laevoides Bonnema Protocythere consobrina Triebel Protocythere jonesi Triebel Protocythere tricostata Triebel . ( ( ( Protocythere triebeli Deroo Cythereis vudispinata (Chapman & Sherborn) Triebel (5) (6) (7) (8) (9) Io) II) 12) (13) Cythere filicosta Marsson. : < (14) Cypridina macrophthalma Bosquet . (15) Cytherina ornatissima Reuss (16) Platycytheveis laminata Triebel (17) Cythere acutiloba Marsson (18) Cythere laticvistata Bosquet (19) Cytherelloidea knaptonensis Kaye (20) = (21) Cytherelloidea pavawiliamsomiana Kaye (22) os (23) Cytherella williamsoniana bosqueti Marsson “I N . Monoceratina longispina (Bosquet) . Monoceratina pedata laevoides Bonnema . Protocythere consobrina Triebel . Protocythere lineata (Chapman & Sherborn) . Pyrotocythere tricostata Triebel Veenia barringtonensis sp. nov. Veenia harrisiana (Jones) Cythereis corvrigenda nom. noy. . Cythereis lonsdaleiana Jones . Cythereis macrophthalma (Bosquet) . Cythereis ornatissima (Reuss) . Platycythereis laminata Triebel . LTvachyleberidea acutiloba (Marsson) . Brachycythere laticristata (Bosquet) . Cytherelloidea knaptonensis Kaye Cytherelloidea hinder sp. nov. . Cytherelloidea pavawilliamsoniana Kaye Cytherelloidea globosa sp. nov. Cytherelloidea gvanulosa (Jones) V. REFERENCES ALEXANDER, C.I. 1929. econ. Geol., Austin, 2907 : 10 pls. 137 PP., Ostracoda of the Cretaceous of North Texas. Bull. Univ. Tex. Bur. 1933. Shell structure of the Ostracode genus Cythevoptevon and fossil species from the Cretaceous of Texas. J. Paleont., Tulsa, 7 : 181-214, pls. 25-27. 1934. Ostracoda of the genera Monocevatina and Ovthonotacythere from the Cretaceous of Texas. J. Paleont., Tulsa, 8 : 57-67, pl. 8. ATH, A. 1850. palaeontologische Beschreibung der nachste Umgebung von Lember. Haidingers naturw. Abh., Vienna, 3 : 171-284. BonneMA, J. H. 1940-41. Ostracoden aus der Kreide des Untergrundes des norddéstlichen Neiderlande. Natuurh. Maandbl., Maastricht, 27 : 91-95, 104-108, 115-118, 129-132, pls. 1-4 ; 28: 8-10, 21-24, 26-29, 40-43, 56-60, 70-72, pls. 5-7. BosgueEt, J. 1847. Description des Entomostracés fossiles de la Craie de Maestricht. M/ém. Soc. Sci. Liége, 4: 353-378, pls. 1-4. 1852. Description des Entomostracés fossiles des terrain Tertiares de la France et de la Belgique. Mém. Acad. R. Belg., Brussels, 24 : 142 pp., 6 pls. 1854. Les Crustacées fossiles du terrain Crétacé du Limbourg. Verh. comm. geol. beschr. Kaurt. v. Nederl., Haarlem, 2 : Butier, E. A. & Jones, D. E. 1957. Domes, Bienville Parish, Louisiana. CHAPMAN, F. 1893. Geol. Ass. Lond., 13: 1894. Lond., 50: 369. 677-692, pls. 33, 34. 1898. On Ostracoda from the “‘ (7) 2: 331-346. CHAPMAN, F. & SHERBORN, C. D. Mag., Lond. (3) 10: CORNUEL, J. 1846. Département de la Haute Marne. 1848. Département de la Haute-Marne. 1893. The Bargate beds of Surrey and their microscopic contents. Cambridge Greensand ”’ On the Ostracoda of the Gault at Folkestone. 345-349, pl. 1, fig. 14. Description des Entomostracés fossiles de terrain Crétacé Inférieur du Bull. Soc. géol. Fr., Description des nouveaux fossiles microscopiques du terrain Crétacé Inférieur du Bull. Soc. géol. Fr., 13-137, pls. I—Io. Cretaceous Ostracoda of Prothro and Reyburns Salt Bull. geol. Surv. La., Baton Rouge, 32 : Note on some microscopic fossils from the chalk of Swanscombe. I—49, pls. 1-6. Proc Quart. J. Geol. Soc. Ann. Mag. Nat. Hist., London Geol. Paris (2) 1 : 193-205, pl. 7 Paris (2) 3: 241-246, pl. I. 78 BRITISH MARINE CRETACEOUS OSTRACODA / Damorte, R. & GRospIDIER, E. 1963. Quelques Ostracodes du Crétacé de la Champagne Humide. 1. Albien-Cénomanien. Rev. Micropaléont., Paris, 6 : 51-66, pls. I-31. DeRoo, G. 1956. Etudes Critiques au sujet des Ostracodes marins du Crétacé Inférieur et. Moyen de la Champagne Humide et du Boulonnais. ev. Inst. franc. Pétrole, Paris, 11 : 1499-1545, pls. 1-5. Dupper, A. 1952. Uber das Cenoman in Neidersachisten Bergland. Paldont. Z., Stuttgart, 26 : 177-188, pls. 25-27. Haskins, C. W. 1963. Revision of the ostracod genus Tvachyleberidea Bowen., Micro- paleontology, New York, 9: 71-74, pl. 1. q Howe, H. V. & LaAurencicH, L. 1958. Introduction to the study of Cretaceous Ostracoda. 536 pp. Baton Rouge. Jones, T. R. 1849. x,, figs. 16,6); Pl a2 eiigd) Arnold (1936) describes and illustrates fertile pinnae of Archaeopteris from Scaumenac Bay, Quebec, as probably referable to Avchaeopteris jacksont. The UPPER DEVONIAN HETEROSPOROUS PLANTS 87 fertile pinnae from the Escuminac formation of the same locality which I have examined resemble Arnold’s material so closely that I consider that they probably belong to the same species. Two of the specimens (V.51312, V.51316) from which spores were obtained are fragments of fertile pinnae (PI. 1, figs. 2, 3), one consisting of the distal ends of two pinnae about 2 cm. long, the other of four pinnae about 2-5 to 3 cm. long. The pinnae on each specimen are so arranged that they have obviously been part of a parallel series on the same leaf. The third specimen (V.44711) is a large fertile primary pinna, 24 cm. long and bearing 18 or Ig pinnae (PI. 1, fig. 1). The smaller specimens are preserved in a very soft sandstone from which the entire spore-masses could be dissected out with a needle. The third, more complete specimen is preserved in a much more indurated, finer grained sandstone and the remains of the sporangia were removed from this with cellulose nitrate film pulls. When dissected out the spore-masses were treated with hydrofluoric acid to remove any adherent mineral matter, individually macerated in Schulze’s solution followed by dilute ammonia and mounted in glycerine jelly, Canada Balsam or “ Clearcol’’. The cellulose nitrate film pulls when removed were treated with dilute hydrofluoric acid, washed and dried, and mounted in Canada Balsam. In the two smaller specimens, from which the most complete sporangial remains were obtained, the sporangia are represented only by spore-masses ; no remnant of the sporangium wall cuticle has been preserved. The macerated spore-masses are of two kinds, both usually 0-3 to 0-5 mm. wide, but some 1-7 to 2:8 mm. long consist- ing of several hundred microspores 45-70 in diameter and others 1:2 to 2-6 mm. long of 9-48 (usually about 15-25) megaspores I10—370y in diameter (PI. 1, figs. 5, 6). By teasing the spore-masses with a pair of needles the spores were separated. Megaspores. The megaspores of Archaeopteris cf. jacksoni when flattened in the equatorial plane are more or less circular in outline. The triradiate mark extends between one to two-thirds of the spore radius and is either in the form of simple commissures or laesurae with labra (lips) about 5y wide. A conspicuous inner membrane (mesosporium?) can be seen in some spores. Paraffin sections of the megaspore-masses (for embedding and sectioning technique see Chaloner & Pettitt 1964) cut at intervals of 6 show that the spore exine is composed of two distinct layers ; an inner homogenous layer about 2 thick is surrounded by a granular layer 6-7y in thickness. The exine sculpture of the spores is somewhat variable. In some specimens the entire spore coat is evenly covered with minute rounded to conical projections (coni) I-2u high and Ip broad at the base (PI. 2, fig. 1), whilst in others it is unevenly covered either with elements that are more or less circular in radial projection and about ry or less in height (grana) or with elements in which the height (1~2y) is greater than the basal diameter and in which the upper end is not much broader than the base (baculae). In some spores the sculptural elements on the contact areas are rather smaller than those covering the rest of the exine, and in others the distal limits of the contact areas are marked by weak curvaturae formed by coalescent sculptural elements. Megaspores with a mesosporium and a uniform decoration of coni can be included 88 UPPER DEVONIAN HETEROSPOROUS PLANTS in the genus Brharisporites Potonié (1956), and clearly some of the Archaeopteris megaspores could also be included in this genus. However, the variation in exine sculpture of some of the spores makes it difficult to assign them to a single genus based purely on morphographic characters. Two species of Biharisporites have been described from the Upper Devonian of Canada by Chaloner (1959) and one by McGregor (1960). One of Chaloner’s species, B. ellesmerensis is within the size range of the Archaeopteris megaspores but differs primarily in having considerably larger sculptural elements ; McGregor’s species B. submamullaris is larger (280-610y). Microspores. The equatorially flattened microspores of Archaeopteris cf. jacksont are circular to subtriangular in outline. The triradiate mark extends over about two-thirds of the spore radius, in some specimens nearly to the equator, and is formed by a simple suture. The exine is about 2—4p thick and is evenly covered with an ornament of small conical elements 1-1-5y high and Ip or less wide at the base. In some spores an inner membrane (mesosporium?) is present (Pl. 1, fig. 9), but in others it is not seen. Circular miospores with an ornament of minute conical projections can be referred to the form genus Cyclogranisporites Potonié & Kremp (1954). This genus is ubiquitous throughout the Carboniferous, and a Lower Carboniferous form very similar to the microspores of A. cf. jacksoni has recently been described by Playford (1962) as Cyclogranisporites lasius. Chaloner (1963) has recorded the genus in sediments of Lower or Middle Devonian age from Southern England. It has proved impossible to determine the precise arrangement of the two types of sporangia on the fertile pinnae. However, three adjacent spore-masses belonging to the same pinnule (ringed on PI. 1, fig. 3) were dissected out and macerated. It was found that two of the spore-masses were composed of microspores and one of megaspores, and it seems therefore, that both microsporangia and megasporangia are borne on the same pinnule in A. cf. jacksont. Each of the spore-masses is enclosed in a coat of acid-resistant cutinised material in the form of globules or as a continuous non-cellular layer adhering to the spores. In some of the spore-masses this residue extends beyond the end of the mass and forms a short protrusion about 60 in length which might represent the remains of the sporangium stalk (Pl. 1, fig. 4). The coat of cutinised material is presumably a residue of the same nature as that reported in Psilophyton sporangia by Lang (1931) which he terms a “‘ tapetum’’, in the sporangia of Archaeopteris latifolia by Arnold (1939), in the sporangia of Svalbardia polymorpha by Heeg (1942), and is probably what Beck (1960) calls non-cellular reticulate thickenings in the sporangia of Archaeopteris ci. macilenta. Feller (1953) and Boterberg (1956) have described inclusions associated with the formation of pseudospores during microsporogenesis in Marsilea which somewhat resemble the globules of tapetal substance in the sporangia of Archaeopteris cf. jacksomt. Boterberg believes that in Marsilea the pseudospores are formed from the residual mass of plasmodial material which results from a lessening of meiotic activity. UPPER DEVONIAN HETEROSPOROUS PLANTS 89 COMPARISON WITH SPORANGIA OF OTHER SPECIES OF ARCHAEOPTERIS Although heterospory has been inferred in several species of Archaeopteris (Krausel & Weyland 1941) it has only hitherto been positively demonstrated in one, Archaeopteris latifolia, from the Upper Devonian of Pennsylvania (Arnold 1939). The spore-masses of A. latifolia are about as large as those of A. cf. jacksont, and although the diameter of the megaspores in the two is very similar, the number per spore-mass is greater in the Scaumenac species. The microspores in A. latifolia are somewhat smaller, being only 35y in diameter. Beck (1960) has found spores of only one size in the sporangia of A. cf. macilenta, but as he later pointed out (Beck 1962) this could mean that the species was dioecious or bore the mega- sporangia and microsporangia on different leaves or branches. I have had the opportunity to examine some fertile material of A. latifolia from the Port Allegany locality presented to the British Museum (Natural History) by Dr. W. G. Chaloner. Due possibly to a slight difference in preservation this material has given a certain amount of information additional to Arnold’s original account. Arnold (1939) describes only spore-masses from his material of A. latifolia and does not give any information about the sporangium wall. In the British Museum material of this species, bulk maceration of the shale results in the release of isolated, incomplete sporangium cuticles the largest measuring 2-5 mm. in length by 0-3 mm. in width, bearing the clear impression of a cellular reticulum (PI. 2, figs. 4, 5). The cells of the reticulum are isodiametric, measuring about 60-80u across, and on certain of the cuticles a somewhat thinner zone of cells runs longi- tudinally along the length of the sporangium. The cells of this thinner band are more or less elongated, measuring 80 by 6o0u and are uniseriate (Pl. 2, fig. 5). Although no definite dehiscence mechanism has been demonstrated in the sporangia of Archaeopteris it has been suggested that spore release was preceded by a simple longitudinal splitting of the sporangium wall (Beck 1960). The longitudinal band of cells in the cuticles of A. latifolia would probably facilitate dehiscence of this type by presenting an area of weakness along which splitting could occur. Tapetal residues in the form of small acid-resistant cutinised globules are also present in the sporangia of A. latifolia, and in many forms a thick covering on the inside of the cuticle (PI. 2, fig. 8). Adhering to the inside of most of the sporangium cuticles are more or less circular spores 35-50u in diameter (PI. 2, fig. 7). A clear triradiate mark extends between one-half to three-quarters of the spore radius and is formed by a simple suture. The exine is about Ip thick and is evenly covered with small conical elements Ip high and ty or less broad at the base. The morphology of these spores is essentially the same as that of the microspores of A. jacksoni and consequently they could also be referred to the genus Cyclogranisporites. The occurrence of spores inside the sporangium cuticle is too frequent to be the result of chance association, and several spores can be found on some of the larger fragments of cuticle. The lower size limit of these spores corresponds to that given by Arnold for the microspores of A. latifolia, but Arnold does not record any sculptural elements on the exines of the microspores he isolated. Beck (1960) has go UPPER DEVONIAN HETEROSPOROUS PLANTS reported the occurrence of spores, 44—68y in diameter, with finely spinose exines in the sporangia of A. cf. macilenta. To judge solely from his illustrations of these spores (pl. 27, figs. 8, 9) they appear to be of the Cyclogranisporites type. If, as suggested by Beck, his inability to demonstrate heterospory in A. cf. macilenta was due to the species being dioecious, or bearing the megasporangia and micro- sporangia on different branches or leaves, the spores described by him could be the microspores of another heterosporous species of Avchaeopteris, and this is, as Beck points out, a much more acceptable alternative than to consider the genus as includ- ing both homosporous and heterosporous species. In the material of A. latifolia from Port Allegany, none of the sporangium cuticles which I have examined contained megaspores. However, a large number of mega- spores was recovered from the maceration residues of the matrix. These megaspores are more or less circular in polar view and 300—400y in diameter. The triradiate mark extends from one-half to three-quarters of the spore radius and has lips 7u wide. The exine is about 7-8y thick with frequent secondary folds. The orna- mentation ranges from conical elements in which the length is more than twice the basal diameter (spinae), to raised ridges forming an irregular reticulate sculpture about 5u high (muri or cristae). On the contact faces of all these forms the sculptural elements are smaller than those covering the rest of the exine. An inner membrane (mesosporium?) can be seen in some specimens (PI. 2, fig. 6). The extremes of variation in exine ornamentation in the megaspores makes it impossible to assign them to any one form genus. Those forms with an ornamentation of conical appendages could be included in the megaspore genus Biharisporites, and megaspores of essentially this type have been found in the megasporangia of A. cf. jacksont. The megaspores of A. latifolia described by Arnold (1939) are within the size range of the megaspores described here, but no highly developed sculpturing is present on his specimens, the exine being only “ slightly roughened’. _ It is possible that the megaspores described above are those of A. latifolia and that differences in preservation or in maceration procedure can account for the more pronounced ornamentation in my material. However, because proof of organic connection is lacking this suggestion is at the most very tentative, and is based merely on the association of the spores and sporangia. Discussion. One of the most interesting facts to emerge from the present study of the sporangia of Avchaeopteris is the similarity of the spores in the various species. The microspores of A. cf. jacksoni and of A. latifolia are almost identical and both are referable to the genus Cyclogranisporites, and those of A. cf. macilenta described by Beck (1960) are clearly similar. In addition, the megaspores of A. cf. jacksoni and possibly those of A. latifolia are morphologically alike. That the various species of a plant genus should have spores that are morpho- logically similar is in no way unusual (see for example the microspores of Selaginella eggersit and Selaginella radiata figured by Erdtman (1957, text-figs. 177, 180)). However, spores very similar to the microspores of Archaeopteris have also been found in some other Devonian plant genera, e.g., Sporogonites exuberans Halle from the Lower Devonian of Réragen in Norway (Halle 1916) and Svalbardia UPPER DEVONIAN HETEROSPOROUS PLANTS 91 polymorpha from the upper Middle Devonian or lowermost Upper Devonian of Spitsbergen (Hgeg 1942) and would therefore be of limited taxonomic value. IV CONCLUSIONS The present study of the fructification of Avchaeopteris cf. jackson in which the microsporangia and megasporangia are in organic connection further demonstrates heterospory in this genus and supports Beck’s and Krausel & Weyland’s supposition that, in all probability, all Avchaeopteris is heterosporous. The occurrence of both microspores and megaspores in the sporangia of Barinophy- ton vichardsoni demonstrates that the genus is definitely heterosporous and is the more noteworthy for being so, as it differs conspicuously from other Devonian heterosporous plants. It would seem that heterospory in the Upper Devonian appeared independently in more than one line of plants, and it has already been shown (Chaloner & Pettitt 1963, 1964) that at least one group had by that time reached a level of heterospory that is the hallmark of the seed. REFERENCES Ananiev, A. R. 1954. Flore du Dévonien inférieur de la partie Sud-Est de la Sibérie occidentale. In Pyvoblémes de la géologie de l’Asie, 1: 287-324, pls. 1-5. Akad. Nauk. S.S.S.R., Moscow. [In Russian]. 1957. Nouveaux végétaux fossiles du Dévonien inférieur de Torgachino dans la partie Sud-Est de la Sibérie occidentale. Bot. Zh., Moscow, 42 : 691-702, pls. 1-3. [In Russian]. ARNOLD, C. A. 1936. Observations on fossil plants from the Devonian of Eastern North America, I. Plant remains from Scaumenac Bay, Quebec. Contr. Mus. Paleont. Univ. Mich., Ann Arbor, 5 : 37-47, pls. 1-4. 1939. Observations on fossil plants from the Devonian of Eastern North America, IV. Plant remains from the Catskill Delta deposits of Northern Pennsylvania and Southern New York. Contr. Mus. Paleont. Univ. Mich., Ann Arbor, 5 : 271-313, pls. 1-10. 1947. An Introduction to Paleobotony. ix+433 pp., 187 figs. New York and London. 1958. Petrified cones of the genus Calamostachys from the Carboniferous of Illinois. Contr. Mus. Paleont. Univ. Mich., Ann Arbor, 14: 149-164, pls. I-12. Beck, R. 1960. The identity of Avchaeopteris and Callixylon. Brittonia, N.Y.,12 : 351-368, pls. 24-29. 1962. Reconstructions of Archaeopteris and further consideration of its phylogenetic position. Amer. J. Bot., Lancaster, 49 : 373-382, 2 figs. BoTERBERG, A. 1956. Etude sur les Hydroptéridales, IV. Genése et différenciation des parois sporales chez Marsilea diffusa Lepr. Cellule, Louvain, 58 : 81~106, pls. 1-6. CHALONER, W.G. 1959. Devonian Megaspores from Arctic Canada. Palaeontology, London, 1: 321-332, pl. 55. 1963. Early Devonian spores from a borehole in Southern England. Gyvana Palynologica, Stockholm, 4 : 100-110, pl. I. CHALONER, W. G. & Pettitt, J. M. 1963. A Devonian Seed Megaspore. Nature, Lond., 198 : 808-809, 3 figs. 1964. A Seed Megaspore from the Devonian of Canada. Palaeontology, London, 7 : (In Press). Dun, W.S. 1897. On the Occurrence of Devonian Plant-bearing Beds on the Genoa River, County of Auckland. Rec. Geol. Surv. N.S.W., Sydney, 5 : 117-121, pls. 10, 11. ErpDTMAN, G. 1957. Pollen and Spore Morphology, Plant Taxonomy. Gymnospermae, Ptevidophyta, Bryophyta. 151 pp., 265 figs. Stockholm. g2 UPPER DEVONIAN HETEROSPOROUS PLANTS FELLER, M. J. 1953. Etude sur les Hydroptéridales, II. Sporocarpe et sporogenése chez Marsilea hirsuta R. Br. Cellule, Louvain, 55 : 307-377, pls. I-10. Hare, T.G. 1916. A fossil sporogonium from the Lower Devonian of Réragen in Norway. Bot. Notisey, Lund., 1916 : 79-81, 1 fig. HoeEc,O.A. 1942. The Downtonian and Devonian Flora of Spitsbergen. Novges Svalbard-og- Ishavs-Undersokelser. Skrift., Oslo, 83 : 7-228, pls. 1-57. KRAUSEL, R. & WEYLAND, H. 1941. Pflanzenreste aus dem Devon von Nord-Amerika, II. Die Oberdevonischen Floren von Elkins, West-Virginien, und Perry, Maine, mit beriicksichtigung einiger stticke von der Chaleur-Bai, Canada. Palgontographica, Stuttgart, 86, B : 1-78, pls. 1-15. Lane, W. H. 1931. On the Spines, Sporangia and Spores of Psilophyton princeps Dawson, shown in specimens from Gaspé. Phil. Tvans., London, 219, B : 421-442, pls. 27, 28. 1932. Contributions to the Study of the Old Red Sandstone Flora of Scotland, VIII. On Arthrostigma, Psilophyton and some associated Plant-remains from the Strathmore Beds of the Caledonian Lower Old Red Sandstone. Tvans. Roy. Soc. Edinb., 57: 491-521, pls. 1-4. MAHABALE, T.S. 1956. Trends of specialisation in the sporocarp and spores in the living and fossil Marsileaceae. Palaeobotanist, Lucknow, 5: 66-72, pls. I, 2. McGrecor, D.C. 1960. Devonian spores from Melville Island, Canadian Arctic Archipelago. Palaeontology London, 3: 26-44, pls. 11-13. Naumova, S. N. 1953. Complexes sporo-polliniques du Dévonian Supérieur de la Platforme russe et leur signification stratgraphique. Tvav. Inst. Sci. geol. Akad. Nauk. S.S.S.R., Moscow, 143 (Geol., 60) : 1-204. [In Russian]. Pant, D. D. & SHRivastava,G. K. 1961. Structural studies on Lower Gondwana megaspores, Part 1. Specimens from the Talchir Coalfield, India. Palaeontographica, Stuttgart, 109, B: 45-61, pls. 30, 31. PLAYFORD, G. 1962. Lower Carboniferous microfloras of Spitsbergen, I. Palaeontology, London, 5: 550-618, pls. 78-87. Potoni£, R. 1956. Synopsis der Gattungen der Sporae dispersae, I. Sporites. Beth. Geol. Jb., Hannover, 23 : 1-103, pls. 1-11. PoToni£, R. & Kremp, G. 1954. Die Gattungen der palaozoischen Sporae dispersae und ihre Stratigraphie. Geol. Jb., Hannover, 69 : 111-194, pls. 4-20. RIcHARDSON, J. Middle Old Red Sandstone spore assemblages from the Orcadian Basin, North East Scotland. Palaeontology, London. (In Press). ScHorr, J. Witson, L. R. & BenTALt, R. 1944. An annotated synopsis of Paleozoic fossil spores and the definition of generic groups. Rep. Invest. Ill. Geol. Surv., 91 : 7-66, pls. 1-3. Waite, D. 1905. Jn SmitH, G. O. & WuHiTE, D. The Geology of the Perry Basin in South Eastern Maine. Pyvof. Pap. U.S. Geol. Surv., Washington, 35 : 9-92, pls. 2-6. : ; _ PLATE 1 Archaeopteris cf. jacksoni (Dawson) Upper Devonian ; Scaumenac Bay, Quebec Fic. 1. Part of a fertile primary pinna bearing 18 or 19 pinnae, x}. V.4471I. Fics. 2, 3. Fertile pinnae from which spore-masses were dissected out. The ring (Fig. 3) surrounds two microsporangia and one megasporangium on the same pinnule. Fig. 2 X1. Fig. 3, X2. V.51316, V.51312. Fic. 4. Small megaspore-mass with a cutinised basal projection, x50. V.51326. Fic. 5. Microspore-mass, x50. V.51327. Fic. 6. Megaspore-mass, X50. V.51327. Fic. 9. Microspore separated from microspore-mass, X 500. V.51316. Barinophyton richardson (Dawson) Upper Devonian ; Perry, Maine Fics. 7, 8. Specimens from which spores were isolated (part and counterpart), XI. V.51350, V.51351. Fic. 10. Microspore, x 500. V.51357. Figs. 1-3, 7, 8 were photographed under xylol. I PLATE Bull B.M. (N.H.) Geol. 10, 3 PLATE 2 Fic. 1. Avchaeopteris cf. jacksont (Dawson). Megaspore, X 200. V.51325. Fic. 2. Bavrinophyton richavdsoni (Dawson). Megaspore, X200. V.51357. Archaeopteris latifolia Arnold Upper Devonian ; Pennsylvania. Fics. 3, 6. Megaspores recovered from maceration residues of matrix, xX 200. V.5131I, V.51310. Fics. 4, 5. Incomplete sporangium cuticles showing a clear cellular reticulum. A thinner longitudinal zone is seen towards the right in Fig. 5. The circular objects are microspores. The background has been painted out. 50. V.51302, V.51303. Fic. 7. Microspore inside sporangium cuticle, x 500. V.51303. Fic. 8. Sporangium cuticle with adherent tapetal globules, x 450. V.51303. PLATE 2 Bull. B.M. (N.H.) Geol. 10, 3 _ >... Sah ids aaa / SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE ee A 7 San. q i Ps . ak Jip ~ 2 ol com _— wy jn _D. E. OWEN “ Ber BULLETIN OF BRITISH MUSEUM (NATURAL HISTORY) LOGY stg : Vol. 10 No. 4 et LONDON: 1965 mewrRIAIN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE BY DEP OW EN, PhD: (The Manchester Museum) Pp. 93-117 ; Plates 1-6 BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 10 No. 4 LONDON : 1965 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), tmnstituted im 1949, 1s ussued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 10, 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) 1965 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued April 1965 Price Twenty-seven Shillings SeORIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE By DAVID ELYSTAN OWEN CONTENTS Page I. INTRODUCTION ‘ ; : j . ‘ ; : . 96 II. AGE OF THE MATERIAL : . . 5 : : : ¢ 97 III. MerTHOD OF SECTIONING . : ‘ ‘ ‘ ; : : 98 IV. SySTEMATIC DESCRIPTIONS : ; P F : , F 98 Order Cyclostomata Busk : : : : : ‘ : 98 Family Diastoporidae Gregory . : ‘ : : : 98 Genus Mitoclema Ulrich . : ; 5 F 5 A 98 M. vegularis (Vine) : : : : ; : 99 Family Ceramoporidae Ulrich. p : . ; 5 99 Genus Ceramopora Hall ; ; d ‘ : 5 99 Ceramopora sp. . 4 : ; : 100 Genus Favositella Etheridge & Foard : : : é 100 F. inteypuncta (Quenstedt) ; : : : : 100 Family Fistuliporidae Ulrich : : ; ¢ : : 101 Genus Fistulipora McCoy : : : : : ; 101 F. crassa (Lonsdale) ; : : : : ; 101 Order Trepostomata Ulrich . : : : : : é 102 Family Batostomellidae Miller. 5 ‘ : : : 102 Genus Eridotrypa Ulrich . : : : 3 : : 102 E. cylindrica sp.nov. . é : : 3 ‘ 103 E. cavasp.nov. . 0 c 0 : 5 : 104 Evidotrypasp .. : é : : 104 Family Stenoporidae Waagen & Wentzel F : : : 105 Genus Leioclema Ulrich . : : ; ‘ : ‘ 105 L. denstpovum sp. nov. .« : 5 : : F 105 L. asperum (Hall) . L : ‘ 3 A : 106 L. vamosum sp. nov. 2 : : . ° : 107 Family Constellariidae Ulrich : é 3 ; : : 107 Genus Nicholsonella Ulrich : : : : ; ; 107 N. parva sp. nov. : : : é 4 : 108 Family Halloporidae Bassler : c : : é : 109 Genus Hallopora Bassler . : : : : ; : 109 H. elegantula (Hall) : ; : : ; : 109 H. striata (Hall). : c : 0 5 6 110 Family Amplexoporidae Miller. : ‘ : : 5 II Genus Monotrypella Ulrich : 3 é : : c IIl M. benthallensis sp. nov. : : : ‘ é 111 Genus Monotrypa Nicholson’ . é : 6 : 9 112 M. flabellata Owen : . ¢ 5 5 : 112 Order Cryptostomata Vine. : é : j : : 113 Family Rhabdomesidae Vine : : : ; : : 113 Genus Rhombopora Meek ‘ ‘ : 5 : é 113 R. mawi sp. nov. . ; 3 : : 5 . 113 Family Rhinidictyidae Ulrich. : : 5 5 : 114 Genus Pachydictya Ulrich ‘ : : : : 5 114 P crassa (Fall). : : : 5 : : T15 V. ACKNOWLEDGEMENTS : c 6 . . : ‘ : 116 VI. REFERENCES . : : ; : ; : , ; : 116 96 SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE SYNOPSIS A small collection of Polyzoa from Wenlockian strata of Benthall Edge, Shropshire, contains seventeen species, four of which were described from American Silurian rocks and seven of which are new. I. INTRODUCTION WuEN Davidson was writing his great series of monographs on British Fossil Brachiopoda for the Palaeontographical Society he had some assistance from George Maw of Benthall Hall, south-west of Ironbridge. Maw was a keen amature geologist and was wonderfully placed for collecting from the Wenlock series. The methods he employed to acquire a really representative fauna were a cross between modern wholesale collecting and factory labour of the Industrial Revolution and were charmingly described (Maw im Davidson & Maw 1881 : 100-101) with colourful details. Some twenty tons of shale from Wenlock and Ludlow beds from approxi- mately forty localities were collected, crushed, washed and sieved, and women were employed at one shilling and sixpence a day to pick out the fossils. Further, the old quarries on Benthall Edge were picked over by hand. The brachiopods were all handed over to Davidson, who described the additions to his earlier species in a special supplement (1881). The Polyzoa came into the hands of G. R. Vine of Sheffield. Vine had been interested in Polyzoa at least as early as 1877 when he started a series of papers on Carboniferous forms from Yorkshire, many of which were published in the Pro- ceedings of the Yorkshire Geological and Polytechnic Society. Vine wrote to Maw and asked permission to examine the Polyzoa from the washings, and was sent approximately two and a half hundredweight of small fragments which he proceeded to work through with a hand lens. He described how he picked out upwards of two hundred thousand specimens of small corals, Polyzoa, Entomostraca, etc., and embarked on the description of the Polyzoa, publishing his first paper in the Quarterly Journal of the Geological Society in 1882. He described a number of species, a few of them new, and was very strong on the Ctenostomata and on certain of the simple Cyclostomata. It is clear, however, that he was very uncertain of the great mass of Trepostomata, of Ceramoporoid Cyclostomata and of Cryptostomata other than the reticulate forms. Such names as Drymatopora problematica Vine, Polypora problematica Vine and Thamniscus problematica Vine hint of his difficulties, and his last paragraph (1882 : 68) reads “‘ I have endeavoured, in the above paper, to give as few microscopical details as possible, because these seemed to me to be rather out of place. I cannot, however, let the paper pass beyond my control without saying that every species recorded has been examined macroscopically and microscopically. The sections prepared have revealed many unexpected features that will help to throw some light at least upon the development of the Polyzoa generally, and upon the biological history of the Silurian Polyzoa in particular’’. In his list of species, Monticulipora sp., which he thought, with Nicholson (1879 : 253), was a coral, was noted (p. 47) as “very rich, but the whole wants working’’. Later (1886a : 228) he wrote “‘ had I met someone who would have cooperated with me, the Actinozoa division, chiefly the Monticuliporidae of my list, would have been much fuller than ELS SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE 97 The following year (1883) Vine read a second paper to the Geological Society of London, but this appears only in abstract. Thirty-nine genera of corals and Polyzoa were listed and two new species, Leioclema granatus Vine and L. pulchellus Vine, were described. Unfortunately the manuscript is not in the library of the Geological Society. After that, Vine returned to his native Yorkshire and his final two Silurian Polyzoa papers (1886 and 1886a) were in the Proceedings of the Yorkshire Geological and Polytechnic Society. In these again he described and listed species, but his emphasis was largely on Ctenostomata which he knew and understood well. He also wrote at length on the classification of the group both here and for the British Association as secretary of a Committee making a study of them (1881). Thus, at a time when Ulrich was publishing papers on Ordovician, Silurian, Devonian and Carboniferous Polyzoa, superbly illustrated with lithographs showing the species in section, and Nicholson sectioning Monticuliporids, the opportunity of describing completely the Silurian polyzoan fauna from such a wealth of material was lost. Thirty years after his death, Vine’s son handed over his geological col- lections to Sheffield Museum. All that is left there of the two hundred thousand specimens picked out are about two hundred mounts each containing one or more specimens, nearly half of which are corals. Amongst them are a very few sections too poor and thick to be of much use. In addition, there is a tray containing small boxes of unsorted material from the numbered localities and a further tray containing other unsorted material. Besides the Sheffield collection, there are a number of mounts with similar specimens and thick microscope sections in the collections of the British Museum (Nat. Hist.), the Geological Survey Museum and the Welsh National Museum at Cardiff. Throughout, many of the mounts often contain more than one species. In addition, there are three trays in the British Museum (Nat. Hist.) which, through the courtesy of Dr. H. Dighton Thomas, I have been able to examine. Two are not particularly rich, and their labels suggest that they contain mixed portions from several localities. The third, labelled Benthall Edge, contains a quantity of first-class material. It is from this tray that all the specimens described in this paper have been separated. It was not until I had spent a considerable time isolating, sectioning and mounting these specimens that I found the Sheffield material (at the suggestion of Professor L. R. Moore). Having examined it carefully, I do not think any useful purpose would be served by including it in this review. Throughout this paper I have followed the custom of the British Museum (Nat. Hist.) and referred to this group as Polyzoa. In common with many workers over- seas I have previously used the name Bryozoa. I have, however, recently re-read Thompson (1830) and am quite satisfied that the name Polyzoa properly refers to these creatures and should be accepted on grounds of priority. Il. AGE OF THE MATERIAL The material consists of some thousands of fragments mostly around a few milli- metres in length and 0°5 to 3 mm. in diameter. It appears to be a portion from one sifting and the label ‘“‘ Benthall Edge” suggests that it is the washings from that 98 SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE locality, north-east of Much Wenlock. There the massive Wenlock Limestone rests on the very fossiliferous Tickwood Beds, and it seems most likely that the washings were from those shales. The old quarries and exposures are heavily overgrown to-day, but similar material occurs at this level. The Tickwood Beds are described (Whittard 1952 : 169) as being the uppermost members of the thick Wenlock Shale and to lie in the zone of Cyrtograptus lundgreni, and there is little doubt that this zone also includes at least the lowermost beds of the Wenlock Limestone. The probability, then, is that the Polyzoa described here are from beds of this age. The species belong to the three orders Cyclostomata, Trepostomata and Crypto- stomata, and the two typical Silurian genera Lezoclema and Evidotrypa are both well represented. The deposit from which the material appears to have been collected is very similar in lithology to a number of levels in the Ludlovian, yet the polyzoan fauna is almost entirely different. The only three species common to both are Fistulipora crassa (Lonsdale), Favositella interpuncta (Quenstedt) and Monotrypa flabellata Owen. This suggests that the smaller Polyzoa may be of some use for zonal purposes over a limited area. On the other hand, only four species—Hallopora elegantula (Hall), Hallopora striata (Hall), Letoclema asperum (Hall) and Pachydictya crassa (Hall)—appear to be common to this deposit and the highly fossiliferous Rochester Shales of New York State whose Polyzoa were described by Bassler (1906). I have looked in vain for more of the characteristic species from that deposit. In the same way I have looked for these British species in the Russian literature, particularly of Astrova (1959) and Nekhoroshev (1961), but have not found them. Ill. METHOD OF SECTIONING A very large number of the specimens for study were small, measuring only a few millimetres in length and perhaps one or two millimetres in diameter. In order to be sure of getting the proper sections correctly orientated, I cut such specimens into three parts using a small diamond cutting wheel mounted on a dental drilling apparatus. One portion was then mounted for reference, and the other two em- bedded in a proprietary brand of hard plaster of paris. It was then possible to grind away sufficient of each specimen to show the tangential and the vertical sections, to mount them on glass, and to complete the microscope slide. Besides allowing the right planes to be ground on the material, the plaster was of great value in holding together friable specimens which otherwise tended to break up before the section was sufficiently thin to show such features as the laminae in the walls. IV. SYSTEMATIC DESCRIPTIONS Order CYCLOSTOMATA Busk 1852 Family DIASTOPORIDAE Gregory 1899 Genus MITOCLEMA Ulrich 1882 Ulrich described this genus to include simple Polyzoa with slender ramose zoaria with more or less prominent apertures arranged in transverse series around the SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE 99 branches or in an irregular spiral. He compared it with both Spzvopora Lamouroux and Entalophora Lamouroux, but pointed out that those forms were Jurassic to Recent and that there were no known links with the Palaeozoic species, then des- cribed only from the Ordovician. He therefore felt that these last should be placed in a separate genus. Silurian forms are now known to be fairly common and links may yet be found in the younger Palaeozoic and the older Mesozoic beds. The family is an interesting one in the simplicity of its form and structure and the fact that it continues with little change from Ordovician to Recent times. Mitoclema regularis (Vine) comb. nov. Rin iwatios. 1.2: 1882 Spivopora vegularis Vine : 55, text-figs. 4-6. MATERIAL. PD 4233-309. DESCRIPTION. The zoaria consist of slender cylindrical tubes with raised circular zooecial apertures on all sides forming an irregular spiral. Simple zooecia arise centrally and run parallel to the axis, finally curving out slightly to reach the surface at an angle of about 45 degrees. No diaphragms, mesopores, or acanthopores occur. MEASUREMENTS. Diameter of zoaria : : ; : : ‘ I mm. Apertures raised up é 5 ; ‘ ; 5 Orang Diameter of apertures. : =) 072mm: Number of apertures in 2 mm. Woneraddinally ; : 15 Number of apertures in 2 mm. laterally : : 6 Remarks. This beautiful little species is moderately common in the collection and is very characteristic. In worn specimens, as was noted by Vine, the outer zooecial wall below the aperture is often perforated, showing a single long cavity from the aperture or even an aperture and a hole beneath it. Vine placed the species in the Mesozoic genus Sfivopora, which it closely resembles. Bassler (1952 : 381) introduced Mitoclemella to take the species of Mitoclema with zooecial apertures spiralling round the zoarium. Ulrich’s original description of Mztoclema allowed for such species and I therefore retain this species in his genus. Family CERAMOPORIDAE Ulrich 1882 Genus CERAMOPORA Hall 1851 Hall (1851 : 400) described this as incrusting or flattened hemispherical with cells arranged in alternating or imbricating series, the apertures arching or triangular with the apex above. Though he placed it among the corals, he noted that it was probably a polyzoan. Ulrich (1890: 463, pl. 39, figs. 1-1) chose as type species C. imbricata from the Rochester Shale at Lockport, New York, and he re-described it and figured its internal structure as shown by sections. He stated that the lower or basal portion of the zoarium was composed of a cellular or spongy tissue from which the zooecia grew out more or less obliquely. He further pointed out that the 10o SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE intercellular space between the non-tabular zooecia was occupied by irregular flexuous rows of mesopores, the zooecia and tubes interconnecting by perforations in the walls. Lunaria were said to be small but well-marked. Ceramopora sp. Plbaientie Seagyna: MATERIAL. PD 4240-42. DESCRIPTION. Zoaria small and discoid. Zooecia occur as simple inverted cones and cylinders with thick granular walls. Mesopores are common, sometimes closed, with numerous curved diaphragms forming in places vesicular tissue which is much thicker at the base of the specimen. Apertures are large, nearly circular, with small lunaria, in contact or separated by polygonal mesopores. Acanthopores are wanting. MEASUREMENTS. Diameter of zoarium : F : : : : 5 mm. Thickness of zoarium. ; é : 2mm. Size of apertures . : 7 5 O- ae Oo: 5 X 0:4-0°55 mm. Width of lunaria . ‘ ‘ ; : . 0*16—0-20 mm. Size of mesopore apertures. : closed to 0-2 & 0:2 mm. Number of apertures in 2 mm. : ; : 4 Thickness of walls : | (OF oyeo; 15 mm. REMARKS. This species is meneniad on a sind specimen whose slightly hollowed upper surface is typical of the genus. The thick mass of vesicular tissue resolves itself into mesopores, most of which die out before reaching the surface. Though similar in zoarial shape to the species of the Niagara Limestone and Roch- ester Shales of New York State and to the Silurian forms from Gotland, it differs in having greater regularity of zooecial form and less looseness. As this is a single specimen whose preservation is not perfect, it is not here described as a new species. Genus FAVOSITELLA Etheridge & Foord 1884 Favositella interpuncta (Quenstedt) 1878 Favosites inteypunctus Quenstedt : ro, pl. 143, fig. 9. 1884 Favositella interpuncta (Quenstedt) Etheridge & Foord : 473, pl. 16, figs. 1-1f. 1911 Favositella interpuncta (Quenstedt) ; Bassler : 100, fig. 35. 1962 Favositella interpuncta (Quenstedt) ; Owen : 197, pl. 28, figs. 1, 2. MATERIAL. PD 4243-45. RemARKS. A single, typical, small, encrusting specimen of this species occurs in the collection. It shows the uneven perforated walls with dark granular centres, the tabulated mesopores and even the enclosed “ brown bodies” or “ pearls ”’ described by Oakley (1934). It is a relatively poor specimen. With regard to the description of both genus and species, I have nothing to add to my comments (1962 : 197). SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE IOI Family FISTULIPORIDAE Ulrich 1882 Genus FISTULIPORA McCoy 1849, emend. Nicholson & Foord 1885 The genus was first described by McCoy (1849 : 130) in the following words :— “ Corallum incrusting, composed of long, simple, cylindrical, thick-walled tubes, the mouths of which open as simple equal circular cells on the surface, and having transverse, funnel-shaped diaphragms at variable distances ; interval between the tubes occupied by a cellular network of small vesicular plates’. Later, Nicholson & Foord (1885 : 500) re-described it more fully, noting that it was variously shaped, that the cylindrical zooecia had lunaria, that diaphragms were horizontal, that tabulate mesopores often coalesce to give rise to vesicular tissue, and that occasional acanthopores occur in the zooecial walls forming blunt spines at the surface. This was the first time that the terms “‘ mesopore ”’ and “ acanthopore ”’ were used and the lunules (lunaria) were also named. It should be remembered that Nicholson still considered the genus to belong to the corals. Acanthopores are not now con- sidered to occur in the genus. This amended description seems to sum up the genus fairly accurately though the number of species now described is so great that it may need further study. Fistulipora crassa (Lonsdale) eee ties Sete PIS 2 shes.) ieee) 1839 Heteropora cvassa Lonsdale : 680, pl. 15, figs. 14—14a. 1884 Fuistulipora crassa (Lonsdale) Nicholson : 118, pl. 7, figs. 1—2a. 1885 Fistulipora crassa (Lonsdale) ; Nicholson & Foord : 506, pl. 15, fig. 1. 1962 JF istulipora crassa (Lonsdale) ; Owen: 197. MATERIAL. PD 4246-67. DESCRIPTION. Zoarium ramose or encrusting, in the latter case often covering other species of Polyzoa, and consequently difficult to distinguish macroscopically from the ramose form. Ramose examples are occasionally slightly flattened. Neither maculae nor monticules are seen and lunaria occur as complete rings, slightly raised around the apertures. Zooecia are simple, thin-walled tubes, in ramose forms running parallel and then curving gently outwards to reach the surface at right angles, and in encrusting forms arising from a basal epitheca at a low angle and curving gently to the surface, with occasional diaphragms. Mesopores occur as a vesicular mass, but in ramose forms are to be found in the exozone only. Apertures are circular to oval, completely ringed by lunarial tissue ; mesopore apertures are polygonal, often closed at the surface by calcareous tissue. MEASUREMENTS. Diameter of ramose zoaria_.. : ; i . 2-4 mm. Thickness of encrusting zoaria : : : 0°5-1:25 mm. Size of apertures . : . Or15-0O'2 mm. X O-I-O0-15 mm. Thickness of lunarial tissue. : , . 0*02-0:04 mm. Size of mesopore apertures . 0:08-0:12 mm. X O-I-0-16 mm. Number of apertures in 2 mm. ‘ : i ; 6-9 102 SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE REMARKS. The ramose forms are common in the collection and the encrusting forms rather less so. The similarity in the exozone of the ramose forms and the entire encrusting forms is such that I have no hesitation in placing them in the same species. They are indistinguishable in tangential section and a complete incrustation of another ramose polyzoan makes it impossible to distinguish them without a vertical or transverse section. Furthermore, both forms are notable for the lunaria completely encircling the apertures, a fact noted by Nicholson (1884 : 118) when he re-described the species. Both have a thin cortex of calcareous tissue which often hides the mesopore apertures. Both Lonsdale and Nicholson had specimens from Benthall Edge, Lonsdale’s figured specimen (1839, pl. 15, fig. 14) being Io mm. in diameter and Nicholson’s specimens being described as 2-12 mm. across. Nicholson & Foord (1885, pl. 15, fig. 1) figure a very large specimen. The ramose specimens described here vary from 2—4 mm. and, like those in Nicholson’s description (1884 : 118), are slightly flattened. Nicholson (1884 : 119) described Fistulipora ludensis as a small encrusting form otherwise very similar to F. crassa. It was said to differ in having spiniform tubules (acanthopores) and rather more mesopores, as well as in its zoarial form. I have not seen any specimens of this species and neither have I seen acanthopores in the genus. It seems possible that Nicholson mistook certain thickenings in the lunaria for acanthopores and that this species is the encrusting form of Ff. cvassa (Lonsdale). Fistulipora dobunica (Nicholson & Foord 1885: 511, pl. 17, figs. 3-36) was also described as an encrusting form from the Wenlock Limestone, but its well-marked maculae and its very tiny apertures (12 in 2 mm.) distinguish it. Fistulipora lockportensis Bassler is the only one of four species of F'zstulipora described by Bassler (1906 : 23, pl. 7, figs. 1-3) from the Rochester Shale which is comparable. While the other three species have horseshoe-shaped lunaria, in that species they appear as a complete ring. F. lockportensis is, however, a large and massive form, the zoaria measuring as much as 10 cm. across and the apertures 4 to 2 mm. Order TREPOSTOMATA Ulrich 1882 Family BATOSTOMELLIDAE Miller 1889 Genus ERIDOTRYPA Ulrich 1895 Ulrich introduced this genus to include certain species formerly described as Batostomella, which he proposed to restrict to the Carboniferous, and for a number of new species for which he could find no existing genus. He described it (1895 : 264) as ramose with slender branches. ‘‘ Zooecia more or less oblique, with thick walls, the tubes intersected by diaphragms only. The latter may be wanting in the axial region, are in most cases absent for a short distance within the apertural edge, but are always present and closest together in the turn from the axial into the narrow peripheral region. Mesopores with close-set diaphragms, varying in number, some- times abundant, at other times very few. Acanthopores small, never numerous, sometimes wanting ’’. The most noticeable features of species in this genus to me are the short, thickened walls of the exozone and the mesopores, often closed in this SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE 103 region. In many species the base is expanded or encrusting and the ramose portion rises from this. Where visible, the laminae show Leioclemid wall structure and possibly link Evidotrypa to Leioclema, which differs mainly in the great development of large acanthopores. Eridotrypa cylindrica sp. nov. BZ ties. 3.04. Diacnosis. Evidotrypa with thin-walled zooecia without diaphragms in the endo- zone and with thick mass of laminated tissue forming exozone and showing Leio- clemid wall structure. MATERIAL. Holotype PD 4268-70 (specimen and sections). Paratypes PD 4271-78. DESCRIPTION. Zoarium cylindrical, ramose. Zooecia occur as long, thin-walled tubes in the endozone, bending sharply into the exozone where the walls thicken and straighten out to reach the surface at about 70°. Wall laminae curve distally in a marked figure U. Mesopores often closed, occur in exozone only. Diaphragms in mesopores and occasionally in zooecia in exozone. The laminae of the diaphragms show Leioclemid wall structure perfectly (cf. Boardman 1960 : 30, 31). Apertures rounded to oval or polygonal with rounded corners, ringed around with dark tissue. Mesopore apertures polygonal. Occasional small acanthopores, often difficult to tell in tangential section from a nearly closed mesopore. MEASUREMENTS. Diameter of zoaria : : 5 ; : . 2-3mm. Thickness of exozone. : : ; : o-8-1I-0 mm. Size of apertures . . O'I14-0:2mm. X O-I-0-17 mm. Thickness of spending ring : . 0°02—0:05 mm. Size of mesopore apertures very saline to 0:06 x 0-04 mm. Diameter of acanthopores ‘ : : ; . 0°02 mm. Number of apertures in 2 mm. : ; 3 ; 8 Remarks. This is fairly easily recognisable macroscopically in an unworn speci- men for the apertures appear polygonal and the mesopores and small acanthopores are not visible. With the thick mass of laminated tissue in the exozone and Leio- clemid wall structure, it differs from all other species in the collection, in which it is relatively common. The thick tissue in the exozone, the sturdy rings round the apertures, the closed mesopores and the presence of only very small acanthopores lead me to place this species in Evidotrypa though its wall structure would seem to ally it to Leioclema. In appearance it compares fairly closely with E. similis Bassler (1g06 : 31, pl. 12, figs. 10-14; pl. 26, figs. 1, 2) but differs in being smaller with smaller apertures, no thin diaphragms and smaller acanthopores. 104 SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE Eridotrypa cava sp. nov. Phe 2yfigss 5-6: DraGnosis. Evidotrypa with massive walls and apertures sunk in hollows. In the endozone the zooecia are closely tabulate, in the exozone they are ringed with dark tissue which shows up in tangential section. MATERIAL. Holotype PD 4285. Paratypes PD 4279-84 and PD 4286-88. DESCRIPTION. Zoarium ramose. Zooecia arise centrally, run parallel and then turn sharply into exozone to reach the surface at or near a right angle. Walls in endozone thick and diaphragms numerous at about half tube-width intervals. In exozone, walls greatly thickened, integrate, with laminae which arise parallel to the inner edge of the wall and curve distally in a broad U-shape, though the centre is marked by an uneven black line. Only a few diaphragms enter the exozone. Small mesopores arise in the exozone and acanthopores are also visible in this region. Apertures oval and often partly closed by the thickness of the walls. The dark rings around them, in which lie the acanthopores, give them the appearance of lying in hollows, and in the solid specimen they are seen to do this. Mesopore apertures are often partially or even completely closed. MEASUREMENTS. Diameter of zoaria : : ; : 2 I°5-2°5 mm. Thickness of exozone : : ; . 0:4-0:8 mm. Size of apertures. : 5 O-I-0°3 mm. X 0:05-0:2 mm. Size of mesopore apertures . from closed to 0-08 x 0-06 mm. Diameter of acanthopores : ; : . 0:02-0:04 mm. Number of apertures in 2 mm. : : : ; 6 Thickness of wall between two apertures . j 0-06—0-2 mm. REMARKS. This is a common species and is easy to recognise in the unworn specimen through the massive walls and the apertures resting in hollows. Micro- scopically it is equally characteristic both in tangential and vertical sections, and differs in tabulation in the endozone and in massiveness in the exozone both from other species of Evidotrypa and from other forms in the deposit. Eridotrypa sp. Rigs ties: 2) MATERIAL. PD 4289-95. DESCRIPTION. Zoarium encrusting, thin, but occasionally thickening. Zooecia arise from a basal epitheca. In thin portions the endozone is very short and the fairly thick walls thicken rapidly into the exozone, which consists of a mass of thick calcareous tissue with threadlike hollow acanthopores running through. In the expanded portion the thin zooecial walls arise at a relatively low angle from the basal epitheca but quickly turn up towards the surface. They remain thin until they expand into the calcareous exozone. Diaphragms few. No mesopores. The polygonal apertures are surrounded by thick walls containing numerous small acanthopores. SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE 105 MEASUREMENTS. Thickness of zoaria in thin encrustations . ‘ 0-2—0°3 mm. Thickness of zoaria in expansion. 4 : : 2 mm. Breadth of expanded portion . ; : : 7 2-5 gm. Thickness of exozone . : ‘ ; : O-I-0:4 mm. Size of aperture . ‘ . O14-0'18 mm. X 0O-I—-0-16 mm. Number of apertures in 2 mm. ; : : . 10-12 REMARKS. This species is represented in the deposit by three specimens, two encrusting fragments of crinoid stem and the third encrusting a small coral. In PD 4293 there is a considerable expansion on one side, and the species is similar to Enidotrypa umbonensis Owen, differing from it in its lack of mesopores and less even shape of the apertures. The material is not sufficiently complete and well preserved to allow a new species to be based on it. Family STENOPORIDAE Waagen & Wentzel 1886 Genus LEIOCLEMA Ulrich 1882 Ulrich described this genus to include forms with encrusting, lamellar, subglobose or ramose zoaria, with zooecia whose apertures often become petaloid by the en- croachment of large acanthopores in the walls, and with abundant mesopores which may even take on the vesicular appearance typical of Fistuliporids. Boardman (1960 : 30) described a typical Leioclemid wall structure of diaphragm-wall units which can often be traced across two or three adjacent mesopores. In Silurian forms, I have not always found this wall structure though it shows clearly in some. This may be due in part to imperfect preservation. The other features described by Ulrich seem to me to distinguish it from associated genera. Leioclema densiporum sp. nov. Pls; fies..32A:. Di1aGnosis. Ramose Lezoclema with polygonal mesopores and many acanthopores surrounding and indenting each aperture. MATERIAL. Holotype PD 4302-04 (specimen and sections). Paratypes PD 4299-301 and PD 4305-08. DEscRIPTION. Zoarium ramose, usually slender. Zooecia thick-walled through- out, arise centrally and curve gently to reach the surface at or near a right angle. Numerous mesopores and acanthopores develop in the exozone. Diaphragms few. Wall structure fairly clear, showing few laminae, with the typical Leioclemid wall not visible. Apertures oval, indented by the many moderate sized acanthopores, and separated by polygonal mesopores. 106 SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE MEASUREMENTS. Diameter of zoaria : : ; : : I°5—2:5 mm. Thickness of exozone. ‘ é 0:3-0-6 mm. Size of apertures . yO 14-0: 18 mm. X 0:10-0:12 mm. Size of mesopore aperhines . 0:08-0:16 mm. X 0:02-0:06 mm. Diameter of acanthopores : ; : . 0:0I-0-:03 mm. Number of zooecia in 2 mm. : : : 8 Thickness of zooecial wall at sumPace : ‘ _ (0:02) mir Number of acanthopores around aperture j 8-12 REMARKS. The five specimens in the collection all show the same features quite clearly. In the unworn figured holotype the indented apertures, spines marking acanthopores, and polygonal mesopores are all clearly visible. I hesitate to leave this species in Lezoclema as it does not show the typical Leioclemid wall and diaphragm structure, though this is not unusual in other Silurian species. It is, however, very typical of the genus as described originally by Ulrich (1882). It is not unlike the description of L. multiporum Bassler though that species is encrusting. Bassler (1906: 34) notes, however, that all of a small lot collected in Rochester, N. Y., differed from the usual method of growth and formed dwarfed branches. It differs, however, mainly in having larger zooecia and fewer mesopores. Leioclema asperum (Hall) Pl.3; figs. 530: 1852 Callopora aspera Hall: 147, pl. 4o. 1890 Leioclema asperum (Hall) Ulrich : 425. 1906 Lioclema (sic) asperum (Hall) ; Bassler : 32, pl. 11, figs. 1-3 ; pl. 24, figs. 14-16. MATERIAL. PD 4300-11. DESCRIPTION. Zoarium small, encrusting or massive. Zooecia simple tubes arising from an epitheca and running directly to the surface with relatively few diaphragms. Numerous tabulate mesopores with diaphragms approximately tube- width apart. Large hollow acanthopores extend the whole depth. Once more it is difficult to comment on the wall structure which is more granular than laminar. Apertures circular or oval, deeply indented by the few very large acanthopores. Mesopores oval or polygonal with well rounded corners. MEASUREMENTS. Diameter of figured specimen ; : : 25 me Depth . ‘ : : : : o)) 25 om: Size of apertures . : ; © ae —0'2 mm. X 0-12—0'15 mm. Size of mesopore apertures .0:07-0‘I mm, X 0:04 X 0:07 mm. Diameter of acanthopores : i ; : 0-05-0'I mm. Number of apertures in 2 mm. : ; ; 8-9 Number of acanthopores to each ante ‘ : 2-4 Thickness of zooecial wall at surface F . . 0:02 mm. SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE 107 REMARKS. This single specimen fits in very closely with Bassler’s description (1906) and with specimens identified by him in the British Museum (Nat. Hist.) collections. It is distinguished from other forms of Lezoclema in the deposit by its encrusting habit and its particularly large acanthopores which show up clearly in vertical as well as tangential sections. Leioclema ramosum sp. nov. Bi 3; figs’7, & DraGnosis. Ramose Leioclema with thick walls in endozone and Leioclemid wall structure clearly marked in exozone. Acanthopores one to three per zooecial aperture. MATERIAL. Holotype PD 4315-17 (specimen and sections). Paratype PD 4312-14. DESCRIPTION. Zoarium ramose. Zooecia arise centrally and curve gently out to reach the surface at or even beyond a right angle. Walls fairly thick throughout, but thickening markedly in exozone in which both mesopores and acanthopores occur. The wall is typically Leioclemid (Boardman 1960 : 30-31), wall-diaphragm units extending across many mesopores, the diaphragms occuring approximately a tube-width apart. Diaphragms also occur in the exozone in the zooecia. Meso- pores fairly numerous, sometimes closed. Apertures oval or polygonal with rounded corners, mesopore apertures similar but smaller. Acanthopores relatively few and very large, often indenting the side of the zooecial aperture, one, two, or at most three to an aperture. MEASUREMENTS. Diameter of zoaria ; : 3 3 : 2-2°5 mm. Thickness of exozone . : : 0-75-I mm. Size of apertures . ‘ 0-12-0° 16 mm. X 0:08-0:12 mm. Size of mesopore apertures. . Closed to 0-I x 0-05 mm. Diameter of acanthopores ; F ; . 0°05-0:09 mm. Number of apertures in 2 mm. : , , 10 Thickness of walls : ; : Up to O-I mm. Remarks. This is one of the few Sider species of Lezoclema which shows the wall structure clearly. It is not very different from Lzoclema [sic] rvamulosum Bassler (1906 : 35, pl. 11, figs. 11-13 ; pl. 25, figs. 9, 10) but differs in having relatively thick walls in the endozone and in having diaphragms. It differs from other species of Lezoclema in the collection in its beautifully formed Leioclemid wall structure, but otherwise compares closely particularly in tangential sections. Family CONSTELLARIIDAE Ulrich 1890 Genus NICHOLSONELLA Ulrich 1890 This genus was introduced by Ulrich (1890 : 374, 421) to include four or five Ordovician species, one of which had already been described by him as a species of 108 SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE Heterotrypa Nicholson. The type species, N. ponderosa Ulrich, is very large, andall were described as having irregularly intertwining flattened branches or fronds. Zooecia were said to be tubular with circular apertures and a faint granular peristome. The walls were thin and traversed longitudinally by minute tubul. The numerous angular mesopores were said to isolate the zooecia, but acanthopores were not mentioned. With age the interzooecial spaces were said to become filled with a perforated calcareous deposit. The figures showed the shadowy nature of the walls. Bassler (1906 : 38, 39, pl. 14, figs. 10-14 ; pl. 24, figs. 1, 2) added two Silurian species from the Rochester Shale of New York State. Both were explanate expansions and were very similar to Ulrich’s Ordovician forms with the same shadowy walls and the calcareous deposit filling interzooecial spaces, but both had many large granular acanthopores. I have examined sections of Nicholsonella florida (Hall) in the British Museum (Nat. Hist.) collections named by Bassler, and feel that the genus would be more properly placed in the Cyclostomata near Fistulipora. In describing F. umbrosa Owen (1960 : 69, 70, pl. 16, figs. 1, 2) I was struck by the similarity of the shadowy walls to those of species of Nicholsonella, but felt that it was otherwise a very typical Fistuliporid. Such walls occur frequently in Ceramoporoids and less often if at allin the more typical Trepostomata. Nicholsonella parva sp. nov. PinAy igss 2: DiaGnosis. Small Nicholsonella with numerous large acanthopores which only slightly indent the apertures. MATERIAL. Holotype. PD 4321. Paratypes PD 4318-20. DESCRIPTION. Zoaria hemispherical or encrusting, small, with spiny projections marking the position of stout acanthopores. Zooecia short, straight or curved with moderately thick granular walls and no diaphragms. Mesopores common, tabulated, diaphragms approximately a tube-width apart. Acanthopores large with hollow centres and granular walls about three to a zooecium. Apertures circular or oval, touching or separated by polygonal mesopores, frequently slightly indented by acanthopores. MEASUREMENTS. Breadth of hemispherical zoarium_ . : . . 4-5 mm. Height of hemispherical zoarium : : 5 3 mm. Size of apertures . : F ; 0-3-0:4 mm. diameter. Size of mesopore apertures . 0:05-0:08 mm. X 0-06—-0-I mm. Diameter of acanthopores : 5 ; . 0:05-0:08 mm. Number of zooecia in 2mm. . . 4-5 REMARKS. The shadowy nature of the walls and the simple tubular zooecia place this species clearly in Nicholsonella. Only three specimens occur in this collection, two hemispheres and one thin incrustation. The zooecia are slightly smaller than those of N. florida (Hall 1852 : 146, pl. 40, figs. 2a—f), the mesopores less SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE 109 vesicular, the acanthopores fewer and relatively larger, and the apertures less flori- form, but otherwise this species is very like Bassler’s description (1906 : 38, pl. 14, figs. 10-11 ; pl. 24, figs. 1, 2) of Hall’s species, which has, however, large explanate zoaria. It also differs from N. vinguebergi Bassler (1906 : 39, pl. 14, figs. 12, 13) in having larger zooecia whose apertures are not so deeply indented as to be petaloid as is the case in that species. Family HALLOPORIDAE Bassler 1911 Genus HALLOPORA Bassler 1911 This was introduced as a new name for Callopora Hall (non Gray 1848 : 109, 146). Hall described Callopora (1851 : 400) as “‘ ramose or incrusting with a columnar structure ; cells tubular with the apertures circular or petaloid, not contiguous, and having the intermediate spaces occupied by angular cell-like openings which are transversely septate ; tubular cells rarely septate’. The type species, C. elegantula Hall, was well figured in 1852 (pl. 40, fig. 1-17), and it shows clearly the characters of the genus. Bassler added, in his renaming (1911 : 325), that zoaria of Hallopora were almost always solid, ramose and bushy, and in the perfect state the apertures were closed by perforated ornamental covers, which, as growth proceeded, formed the diaphragms of succeeding layers. Hallopora elegantula (Hall) Bl eArniess 3,04: 1852 Callopova elegantula Hall: 144, pl. 40, figs. 1-17. 1882 Callopova elegantula Hall ; Ulrich : 250, pl. 11, figs. 6—6b. 1884 Callopova nana Nicholson : 120, pl. 7, figs. 4—4b. 1906 Callopora elegantula Hall ; Bassler : 41, pl. 17, figs. 11-15 ; pl. 26, fig. 12. 1911 Hallopova elegantula (Hall) ; Bassler : 334, text-fig. 210. MATERIAL. PD 4322-33. DESCRIPTION. Zoaria ramose often relatively stout. Zooecia arise from centre and curve gently out to reach the surface at right angles. Diaphragms numerous and closely spaced in endozone, becoming fewer in exozone where numerous closely tabulated mesopores occur. Zooecial wall integrate in the inner part of the exozone, with laminae running a short distance nearly parallel to the wall and forming a V distally, which shows up as a black line in section, but the V and the black line become much less marked near the surface. Mesopore diaphragms show Leioclemid wall structure, the laminae of one diaphragm running into the wall distally and curving back from a blunt V to run into the diaphragm of the next mesopore. No acanthopores. Apertures circular or oval, separated by polygonal mesopores. MEASUREMENTS. Diameter of zoaria : : : : 5 . 2-5 mm. Size of apertures . : : 0:3-0'4 mm. X 0:25—-0°3 mm. Size of mesopore apertures. O-I-0'2 mm. X 0:05-0-2 mm. Number of apertures in 2 mm. f : : : : 5 110 SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE REMARKS. This is the commonest polyzoan in the collection and one of the commonest in the Wenlockian. It is easy to distinguish macroscopically by its relatively stout form and the circular apertures clearly separated by mesopores. Microscopically in all sections it is quite distinctive. It tallies exactly with Hall’s description and figures and with specimens from the U.S.A. named by Bassler. It also compares with Vine’s sections of Callopora nana Nicholson. In his original description of C. nana, Nicholson mentioned C. elegantula Hall but did not compare it with that species, even though he compared it with two American Ordovician species. The description, figures and the specimen so named in the British Museum (Nat. Hist.) collections are so like C. elegantula Hall that I have no doubt they are conspecific and that C. nana Nicholson is a junior synonym of C. elegantula Hall. Nicholson made the point in his species that both zooecia and mesopore apertures are elongated along the long axis of the zoarium, and I note this to occur occasionally but not generally. He described (1884 : 122, pl. 7, figs. 5-5) C. fletchert (Edwards & Haime) (1885 : 267, pl. 62, figs 3, 3a) as having circular apertures, and figured several mesopores adjoining one another, but he also noted the great thickening of the wall in the exozone which distinguishes it completely from this species. [The possible synonymy of C. fletcheri with Hallopora ramulosa (Phillips) has been dis- cussed by Stubblefield (1938 : 30).] Hallopora striata (Hall) comb. nov. Pl. 4, figs. 5, 6. 1852 Tvematopora striata Hall: 153, pl. 40, figs. 7a—d. 1906 Evidotyvypa striata (Hall) Bassler : 32, pl. 12, figs. 4-6 ; pl. 24, figs. 3-6 ; pl. 25, fig. 14. MATERIAL. PD 4296-08. DESCRIPTION. Zoarium cylindrical, branching. Zooecia arise centrally and curve gently to reach the surface at or near a right angle. Walls integrate, fairly thick throughout, but thickening markedly in the exozone where the laminae are relatively straight, V-ing distally to form a central dark line. Mesopores common, sometimes closed. Diaphragms occur regularly throughout the length of the zooecia at intervals of a half to one tube-width, though they are less frequent at the surface, and in mesopores at approximately the same intervals. The laminae in the thick diaphragms show that the diaphragms form a unit with the wall after the manner of the Atachtotoechids (Boardman 1960 : 32), though the unit does not appear to continue into the neighbouring zooecium in the same way. Apertures oval. Acanthopores wanting. MEASUREMENTS. Diameter of zoaria : ; : : , . 2-2°3mm. Size of apertures . : : 0:3-0°4 mm. X 0:2—-0:25 mm. Size of mesopore apertures. ‘ . 006mm. X 0:04 mm. Number of apertures in 2 mm. ‘ F : i - 4-5 SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE III ReMARKS. This small species came to light in sectioning numerous specimens of other species indistinguishable macroscopically. The integrate walls, tabulate zooecia and lack of acanthopores are typical of Hallopora Bassler. I compared it with sections in the British Museum (Nat. Hist) from Lockport, New York, labelled by Bassler Evidotrypa striata (Hall), and found it to be identical. Bassler’s tan- gential section shows dark dots which he clearly took to be acanthopores though I am by no means certain that this is their real identity. The three specimens in the collection do not permit of sufficient examination for re-description of the species, but I consider that they, and Bassler’s specimens mentioned above, belong to Hallopora Bassler and not Evidotrypa Ulrich. Family AMPLEXOPORIDAE Miller 1889 Genus MONOTRYPELLA Ulrich 1882 Ulrich introduced this genus to include species very like those belonging to the ramose genus Amplexopora Ulrich but differing in the absence of acanthopores and in the presence in some species of what he described as closely tabulated interspaces that simulated mesopores. In his figures these appear to be typical mesopores. It is the presence of numerous acanthopores that is the greatest difference between Amplexopora Ulrich and Monotrypa Nicholson, for both genera are always without mesopores. There seems to be little difference between species of Monotrypella without mesopores (or tabulated interspaces), and Monotrypa, though the latter was founded on a large massive “ coral’ and the former on a small ramose “ polyzoan ”’. Monotrypella benthallensis sp. nov. PIS tes. iy. Diacnosis. Monotrypella with polygonal zooecial apertures of two sizes and Atactotoechid wall structure. MATERIAL. Holotype PD 4334-36 (specimen and sections). Paratype PD 4337. DESCRIPTION. Zoaria ramose with groups of slightly larger zooecia but no true monticules. Zooecia run parallel for some distance and then curve gently out, finally making an angle or elbow to reach the surface at right angles. Zooecial walls thin in endozone, with diaphragms few or wanting, but thickening in exozone with a number of diaphragms mostly thin but a few rather thicker. Wall structure integrate, Atactotoechid (Boardman 1960 : 32), with laminae making a small angle with the walls and forming a V distally which shows as a dark line. Diaphragms continue forward into the wall but are easily lost in its structure. Mesopores and acanthopores wanting. Apertures polygonal, and integrate structure observable in thin black line which separates them. 112 SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE MEASUREMENTS. Diameter of zoaria : ; : : : . 2-3 mm. Thickness of exozone. ‘ : 0:2—0'°6 mm. Size of larger apertures . : 0: oye —0° P mm. X 0-2—0:25 mm. Size of smaller apertures : - 0:2-0:25 mm. X 0-15 mm. Number of larger apertures in 2 mm. : : : : vi Number of smaller apertures in 2mm. . ; ‘ 9 Maximum thickness of zooecial wall in exozone 0-06-0-Il mm. REMARKS. The exozone may be very short and the wall structure is then less easy to determine. The species is represented by four specimens and is easy to recognize macroscopically by the polygonal shapes of the zooecia with their larger groups, and by the lack of mesopores, and microscopically by the sharp angle which the zooecia make in the exozone. The vertical sections are very similar to those of Enidotrypa echinata Hall sp. (1879 : 112, pl. 11, figs. 1-5) named by Bassler in the British Museum (Nat. Hist.) collections, but the tangential sections of that species appear to show numerous very small acanthopores which are not present here. Furthermore, I do not believe this species belongs to the genus Evidotrypa. Although the wall structure appears to be Atactotoechid there are many differences between this and species of that genus. Absence of cystiphragms and of intermittent thick- ening of walls show that it is no near relative of Atactotoechus. The integrate wall structure and absence of mesopores suggest the Amplexoporidae, and the lack of acanthopores, the genus Monotrypella. 1 place it in this genus and not in Monotrypa Nicholson as it is a small ramose polyzoan and differs in form from typical Silurian species of Monotrypa. Genus MONOTRYPA Nicholson 1879 This genus was first separated from Monticulipora by Nicholson (1879) and the subject was further elaborated in 1881. The essential features were the absence of mesopores, though the presence of larger and smaller zooecia, the former often collected in monticules, was noted. Acanthopores were absent except in M. discoidea (James) which was later removed from the genus. The walls were said to be thin, seemingly structureless, and apparently amalgamated to one another in some species, but in others were considerably thickened. In either case they were said to preserve the original lines of demarcation separating each zooecium. Diaphragms were entire, uniformly distributed, sometimes few or wanting. This has since proved to be a very easily recognizable genus and many species have been described. Crenulate walls have proved a feature in a number. Where the walls are thick, their integrate nature and the black line formed by the V-ing of the wall laminae are clearly observable in both tangential and vertical sections. Monotrypa flabellata Owen Pi Opies: a92) 1960 Monotrypa flabellata Owen: 72, pl. 16, figs. 10-11; text-fig. 6. 1962 Monotrypa flabellata Owen ; Owen: 109, pl. 32, figs. 1, 2. SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE rea MATERIAL. PD 4338-47. DESCRIPTION. Zoarium small, encrusting or hemispherical, with groups of larger zooecia showing on surface. Zooecia arise from the epitheca and are simple tubes with crenulate walls and numerous simple diaphragms. The walls are markedly integrate, with laminae arising from the inner sides at a low angle and running distally to form a V which shows up as a black line through the length of the wall. Apertures polygonal with the dark line of laminae clearly visible. No mesopores, though occasional, smaller zooecia are seen in sections. No acanthopores. MEASUREMENTS. Breadth of zoaria : : : : : up to 7 mm. Thickness of zoaria ; : : : up to 2 mm. Size of aperture of larger zooecia_.. . O5 mm. X 0:33 mm. Size of aperture of normal zooecia > Yo-3smim:, X< 0-2 mm. Number of larger zooecia in 2 mm. . : : : . 5-6 Number of normal zooecia in 2 mm. : : ; ; 7 Thickness of zooecial walls. : : 0:2-0:3 mm. Remarks. I place the specimens of Monotrypa in tine collections in this Ludlovian species though there are certain slight differences. The occasional groups of larger zooecia have not been observed in Ludlovian forms and the walls are, if anything, even stouter. There are rather more diaphragms, though diaphragms are occasion- ally numerous in Ludlovian specimens. The zoaria, too, are all smaller, though zoaria of other species in the deposit are also small. The thick wrinkled walls with the dark central line marked by the V-ing of wall laminae are so similar in these and the typical Ludlow forms as to make me consider them conspecific. Order CRYPTOSTOMATA Vine 1883 Family RHABDOMESIDAE Vine 1883 Genus RHOMBOPORA Meek 1872 The genus was introduced for the Carboniferous species, R. lepidodendroides Meek, to include forms with slender, ramose, solid zoaria, and zooecia with vestibules within a very thick outer wall, numerous acanthopores and no mesopores. Ulrich inseveral papers (1890 ef alia) described many species in this easily recognizable genus. Some had hemisepta serving to demarcate the vestibules. Moore (1929 : 134) discussed the genus, and drew attention to the fact that the type species had no hemisepta, and also that many forms described from older strata had hemisepta. Bassler (1953 : G. 134) noted “no hemisepta’”’, but it seems that this is not a diagnostic generic feature. Rhombopora mawi sp. nov. Pl. 5, figs. 3-5. Diacnosis. Khombopora with hemisepta and mesopores, sometimes closed, with thick diaphragms. 114 SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE MATERIAL. Holotype PD 4350-52 (specimen and sections). Paratypes PD 4348-49 and PD 4353-67. DESCRIPTION. Zoaria ramose, some extremely slender and fragile and others rather stouter and stronger. Zooecia thin-walled tubes in endozone curving slightly from the axis and then turning more or less sharply to reach the surface at or neara right angle. The exozone is very thick and solid in the stouter specimens but thinner in the slim ones, and there are intermediate examples and even some with the exozone thicker on one side than the other. Well-marked vestibules particularly in stouter forms are completed with superior hemisepta and occasionally inferior hemisepta. Wall laminae make a low angle with the wall and curve distally into a U-shape. Occasional thin diaphragms occur. Mesopores are numerous in exozone, sometimes closed, containing many thick diaphragms whose laminae run up into the wall laminae after the manner of Leioclemids. Apertures are circular or oval, surrounded by a well-marked ring. Large hollow acanthopores are very numerous, particularly in the stouter forms, occurring in the thick calcareous tissue between the apertures, sometimes breaking the ring but never cutting into the aperture. Macroscopically the positions of the closed mesopores are marked by small depressions. MEASUREMENTS. Diameter of zoaria : 3 : 2 : 1-8—2°5 mm. Thickness of exozone. : ; : : 0-2-0-7 mm. Diameter of endozone . : : o-8-1:2 mm. Size OlvapeLeurenmr : , O 08- O-I2 mm. X 0:05-0-Il mm. Thickness of surrounding ring ; ; : = 0-02) Number of apertures in 2 mm. : as g-II Number of acanthopores surrounding amperes ; 5-10 Diameter of acanthopores ; : : . 0-OI—0:03 mm. REMARKS. This is the second most common species in the collection. It is easy to distinguish macroscopically in unworn specimens, and is very distinctive micro- scopically where its thick cortex and deep vestibules distinguish it from other species. It is typically Rhabdomesid and I place it in the genus Rhombopora although it has well-developed hemisepta. The tabulated mesopores are another feature not normally associated with the genus. Perhaps a new genus should be introduced to take such species of Rhombopora with hemisepta and occasionally with mesopores. Such forms are common in both Silurian and Carboniferous rocks of England. Family RHINIDICTYIDAE Ulrich 1895 Genus PACHYDICTYA Ulrich 1882, emend Ross 1961 In the emended genus Ross (1961 : 338) emphasized the salient features, including the microstructure of the walls. There is nothing to add to her description. SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE II5 Pachydictya crassa (Hall) Pl. 5. figs. 6-8. 1852 Stictopora cvassa Hall: 45, pl. 18, figs. 4a—c. 1893 Pachydictya crassa (Hall) Ulrich : 147. 1906 Pachydictya crassa (Hall) ; Bassler : 57, pl. 18, figs. 11-12 ; pl. 21, figs. 14-16. MATERIAL. PD 4368-76. DEscRIPTION. Zoaria bifoliate, branching, with apertures in longitudinal rows on both sides. Very occasionally triangular in cross-section. Narrow strip along edges of ribbons without apertures. Occasional large zooecia stand out raised up on surface having the appearance of small volcanoes with central craters surrounded by side craters. Zooecia arise on both sides of the mesotheca, which is pierced by numerous tubuli, and make an angle of about 60° with it, quickly becoming normal to the surface. The walls consist of a central laminate portion between the zooecia which contains the tabulate interspaces typical of the genus, and numerous acanthopores, and a clearer portion adjoining the zooecia. The larger zooecia have diaphragms. Aper- tures are oval and surrounded by dark rings and the spaces between them contain numerous small acanthopores. Mesopores wanting. MEASUREMENTS. Breadth of zoaria F : , ‘ ‘ 4-5 mm. Thickness of zoaria 4 : . I-I'°5 mm. Thickness of zoaria at an oaleneedl a zooecium 22°25 mm. Breadth of outer strip. : : : : 0-5-0'7 mm. Number of rows of zooecia_.. F : ; commonly 7 Size of aperture . : 0:25-0'4 mm. X O-I—O-14 mm. Longitudinal amaenepace- : : : : 0-2-0°3 mm. Lateral interspace : : 0°3-0:35 mm. Number of apertures in 2 mm. longitudinally : : : 4 Number of apertures in 2 mm. laterally . , 5 : 5 REMARKS. This is a common species and is easy to pick out macroscopically. It is notable for the occasional, enlarged zooecia which are, however, similar in section to those of normal size. It appears to be similar both to Hall’s figures of Stictopora crassa (1852) and to Bassler’s figures and descriptions of Pachydictya crassa (1906), though there is no sign of a linear ridge separating the zooecial rows. It differs from the Llandoverian-Wenlockian P. holmi Hennig (1905: 25, text-figs. 22-32, pl. 1, fig. 4) in having smaller apertures and less thickness, and in the zooecia having few diaphragms. Like that species the zoarium is very occasionally triangular in section. It differs from the Llandoverian P. dichotoma Nekhoroshev (1961 : 156, pl. 34, figs. 2, 3) in having fewer rows of apertures to a branch. Nekhoroshev noted this as the main difference between his species and P. crassa (Hall), and stated that later forms show a reduction in the number of rows. 116 SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE V. ACKNOWLEDGEMENTS The author would like to record his sincere thanks to Dr. H. Dighton Thomas, who has always been ready with help and advice, and to Mr. H. Spencer, who has taken great pains to perfect the photography. VI. REFERENCES Astrova, G. G. 1959. Silurian bryozoa from Central and Eastern Touva. Tyvav. Inst. Paleont. Acad. Sct. U.R.S.S., Moscow, 79 : 1-72, pls. 1-12. [In Russian. ] BAssLER, R. S. 1906. The bryozoan fauna of the Rochester Shale. Bull. U.S. Geol. Suvv., Washington, 292 : 1-137, pls. 1-31. 1911. The Early Paleozoic Bryozoa of the Baltic Provinces. Bull. U.S. Nat. Mus., Washington, 77 : I—xxi, 1-382, pls. 1-13. —— 1952. Taxonomic notes on genera of fossil and recent Bryozoa. J. Wash. Acad. Sci., 42 : 381-385, figs. 1-27. 1953. Tveatise on Invertebrate Paleontology, Part G. Bryozoa : G1-253. Geol. Soc. Amer. & Univ. Kansas Press. BoarpMan, R.S. 1960. Trepostomatous bryozoa of the Hamilton Group of New York State. Prof. Pap. U.S. Geol. Suvv., Washington, 340 : 1—87, pls. 1-22. 1960a. A revision of the Ordovician bryozoan genera Batostoma, Anaphragma and Amplexopora. Smithson. Misc. Coll., Washington, 140, 5 : 1-28, pls. 1-7. Davipson, T. 1881. A Monograph of the British Fossil Brachiopoda, 5, 1 : 65-72. Supple- ment to the British Silurian Brachiopoda. Palaeontogr. Soc. [Monogr.], London. Davipson, T. & Maw, G. 1881. Notes on the physical character and thickness of the Upper Silurian Rocks of Shropshire, with the Brachiopoda they contain grouped in Geological Horizons. Geol. Mag. Lond. (2) 8 : 100-109. Epwarps, H. M. & Harme, J. 1851. Monographie des Polypiers fossiles des Terrains palaeo- zoiques. Arch. Mus. hist. nat. Paris, 5 : 1-502, pls. 1-20. —— 1854 [1855]. Corals from the Silurian Formation. A Monograph of the British Fossil Corals, 5 : 245-299, pls. 57-72. Palaeontogr. Soc. [Monogr.], London. GRAY, J. E. 1848. List of specimens of British animals in the collection of the British Museum, I. Centroniae ov Radiated Animals. 173 pp. London. Hatt, J. 1851. New genera of fossil corals from the report by James Hall, on the Palaeont- ology of New York. Amer. J. Sci., New Haven (2) 11: 398-401. 1852. Natural History of New York. Palaeontology, 2. viii + 362 pp., 88 pls. Albany. 1879. The Fauna of the Niagara Group of Central Indiana. Rep. N.Y. St. Mus., Albany, 28 : 99-203, pls. 1-32. Hennic, A. 1905. Gotlands Silur-Bryozoer, 1. Avk. Zool., Uppsala, 2, 10 : 1-37, pls. I, 2. —— 1908. Gotlands Silur-Bryozoer, 3. Avk. Zool., Uppsala, 4, 21 : 1-64, pls. 1-7. LONSDALE, W. 1839. Jn Murcuison, R. I. The Silurian System: 577-712, pls. 1-27. London. McCoy, F. 1849. On some new genera and species of Palaeozoic Corals and Foraminifera. Ann. Mag. Nat. Hist., London (2) 3 : 119-136, 4 figs. 1851-1855. In SrEpGwicK, A. & McCoy, F. A synopsis of the classification of the British Palaeozoic vocks, with a systematic description of the British Palaeozoic Fossils in the Geological Museum of the University of Cambridge. xcviil+661 pp., 25 pls. London & Cambridge. Moore, R.C. 1929. A bryozoan faunule from the Upper Graham Formation, Pennsylvanian, of North Central Texas. J. Paleont., Tulsa, 3 : 1-27, 121-156, pls. 1-3, 15-18. NEKHOROSHEYV, V. P. 1961. Ordovician and Silurian Bryozoa from the Siberian Platform. Order Cryptostomata. Tvudy vses. nauchn.-isslzed geol. Inst. (VSEGEI), Leningrad, 41: 1-246, pls. 1-37. [In Russian.] SILURIAN POLYZOA FROM BENTHALL EDGE, SHROPSHIRE Ly, Nicuorson, H. A. 1879. On the Structure and Affinities of the “‘ Tabulate Corals’’ of the Palaeozoic Period. xii + 342 pp., 15 pls. Edinburgh & London. —— 1881. On the Structure and Affinities of the Genus Monticulipora and its Subgeneva. xvi + 240 pp., 6 pls. Edinburgh & London. —— 1884. Contributions to Micro-Palaeontology. Notes on some Species of Monticuliporoid Corals from the Upper Silurian Rocks of Britain. Ann. Mag. Nat. Hist., London (5) 13: 117-127, pl. 7. Nicuorson, H. A. & Foorp, A. H. 1885. On the genus Fistulipora McCoy with descriptions of several species. Ann. Mag. Nat. Hist., London (5) 16 : 496-517, pls. 15-18. Oakley, K. P. 1934. Phosphatic calculi in Silurian Polyzoa. Proc. Roy. Soc. Lond. (B) 116 : 296-314, pls. 12-14. Orpieny, A.D. 1850. Prodvome de Paléontologie stratigvaphique universelle,1. 1x + 394 pp. Paris. OweEN, D. E. 1960. Upper Silurian Bryozoa from Central Wales. Palaeontology, London, 3: 69-74, pl. 16. 1961. On the species Orvbignyella fibyosa (Lonsdale). Geol. Mag. Lond., 98 : 230-234, pl. 14. —— 1962. Ludlovian Bryozoa from the Ludlow district. Palaeontology, London, 5: 195- 212, pls. 28-32. Pocock, R. W. & WuITTARD, W. F. 1948. The Welsh Borderland. Brit. Reg. Geol., 2nd ed. Geol. Surv. Gt. Britain. QuENSTEDT, F. A. 1878. Petvefactenkunde Deutschlands, 6. ‘Korallen. 1093 pp., 42 pls. Leipzig. Ross, J. R. P. 1960. MRestudy of Types of Seven Ordovician Bifoliate Bryozoa. Palaeont- ology, London, 3 : 1-25, pls. I-10. —— 1961. Larger Cryptostome Bryozoa of the Ordovician and Silurian, Anticosti Island, Canada, 2. J. Paleont., Tulsa, 35 : 331-344, pls. 41-45. STUBBLEFIELD,C.J. 1938. The Typesand Figured specimens in Phillips and Salter’s Palaeont- ological Appendix to John Phillips’ Memoir on ‘ The Malvern Hills Compared with the Palaeozoic Districts of Abberley, etc.’ Swmm. Progr. Geol. Surv. Gt. Brit., 1936, 2: 27-51. Tuompson, J. V. 1830. Zoological Researches, and Illustrations ; or Natural History of Nondescript or Imperfectly Known Animals, a series of Memoirs. On Polyzoa, a new animal discovered as an inhabitant of some Zoophites—with the description of the newly instituted Genera of Pedicellaria and Vesicularia, and their Species. Memoir 5 : 89-102, pls. 1-3. Cork. Urricu, E.O. 1882. American Palaeozoic Bryozoa. J. Cincinn. Soc. Nat. Hist., 5 : 121-175, 232-257, pls. 6-8, Io, 11. —— 1890. Palaeozoic Bryozoa. Geol. Surv. Illinois, 8 : 283-728, pls. 29-78. —— 1895. On Lower Silurian Bryozoa of Minnesota. The Geology of Minnesota. Final Report, 3, 1 : 96-332, pls. 1-28. Minneapolis. Vine, G. R. 1881. Second Report of the Committee consisting of Professor P. M. Duncan, F.R.S., and Mr. G. R. Vine, appointed for the purpose of reporting on fossil Polyzoa. Rep. Brit. Ass., London, 1881 : 161-175. —— 1882. Notes on the Polyzoa of the Wenlock Shales, Wenlock Limestone, and Shales over Wenlock Limestone. Quart. J. Geol. Soc. Lond., 38 : 44-68. —— 1883. Notes on the Corals and Bryozoans of the Wenlock Shales (Mr. Maw’s Washings). Quart. J. Geol. Soc. Lond., 39 : 69-70 (Proc.). —— 1886. Notes on the Polyzoa of the Wenlock Shales, etc. Part. 1. Pyvoc. Yorks. Geol. (Polyt.) Soc., Leeds, 9 : 179-201. —— 1886a. Notes on the Palaeontology of the Wenlock Shales of Shropshire (Mr. Maw’s Washings 1880). Proc. Yorks. Geol. (Polyt.) Soc., Leeds, 9 : 224-248. WHITTARD, W.F. 1952. A Geology of South Shropshire. Proc. Geol. Ass. Lond., 63 : 143-197. PD4239. PD42306. PD4241. PD4242. PD4267. PLATE tf Mitoclema vegularis (Vine) Vertical section showing simple zooecia curving to the surface. X50. Solid specimen with apertures occurring in an irregular spiral. x13. Ceramopora sp. Vertical section of PD4240 showing vesicular tissue. X50. Tangential section of same specimen showing apertures with lunaria. Fistulipora crassa (Lonsdale) Vertical section of ramose form. X50. PLATE 1 Bull. B.M. (N.H.) Geol. 10, 4 PLATE 2 Fistulipora crassa (Lonsdale) Fic. 1. PD4251. Vertical section of PD4250, encrusting form. X50. Fic. 2. PD4261. Tangential section of another specimen, PD4260, showing lunaria encirc- ling apertures. X50. Evidotrypa cylindrica sp. nov. Fic. 3. PD4269. Vertical section of holotype, PD4268, showing Leioclemid wall structure. x 50. Fic. 4. PD4270. Tangential section of holotype showing ringed apertures. 50. Evidotrypa cava sp. nov. Fic. 5. PD4285. Vertical section (holotype) showing numerous diaphragms in endozone. x 50. = Fic. 6. PD4286. Tangential section of another specimen showing apertures thickly ringed round, small mesopores and acanthopores. xX 50. Bull. B.M. (N.H.) Geol. 10, 4 PLATE 2 Fie. X 50. FIG. FIG. x 50. Fic. Fie. x 50. Fic. Fic. x 50. Fic. PD4293. PD4294. PD4303. PD4304. PD4311. PD4310. PD43106. PD4317. PLATE 3 Evidotrypa sp. Vertical section of PD4292 showing both thin and expanded portion. Tangential section of same specimen. X50. Leioclema densiporum sp. nov. Tangential section of holotype, PD4302, showing numerous acanthopores. Vertical section of holotype showing thick walls. x50 Leioclema asperum (Hall) Vertical section of PD4309 showing mesopores and stout acanthopores. Tangential section of same specimen showing large acanthopores. X 50. Leioclema vamosum sp. nov. Vertical section of holotype, PD4315, showing Leioclemid wall structure. Tangential section of holotype showing few large acanthopores. 50 PLATE 3 Bull. B.M. (N.H.) Geol. 10, 4 + te: kt > = Fig. 1. PD4321. walls. X50. Fic. 2. PD4321. x 50. Fic. 3. PD43209. Fic. 4. PD4330. mesopores. X50. Fie. 5. PD4298. Fic. 6. PD 4297. PLATE 4 Nicholsonella parva sp. nov. Vertical section of holotype showing tabulate mesopores and shadowy Tangential section of holotype showing mesopores and acanthopores. Hallopora elegantula (Hall) Vertical section of PD4328 showing Leioclemid wall structure. x50. Tangential section of same specimen showing apertures separated by Hallopora striata (Hall) Vertical section showing Atactotoechid wall structure. x50. Tangential section of another specimen showing integrate walls. 50. Bull. B.M. (N.H.) Geol. 10, 4 PLATE 5 Monotrypella benthallensis sp. nov. Fic. 1. PD4337. Vertical section showing Atactotoechid wall structure. X50. Fic. 2. PD4335. Tangential section of holotype, PD4334, showing different sizes of apertures and integrate wall structure. 50. Rhomboporva mawi sp. nov. Fic. 3. PD4351. Vertical section of holotype, PD4350, a specimen with a thick exozone. Note the tabulate mesopores. X50. Fic. 4. PD4365. Vertical section through another specimen with a relatively thin exozone. Note the hemisepta. x 50. Fic. 5. PD4366. Tangential section of another specimen showing acanthopores. X50. Pachydictya crassa (Hall) Fic. 6. PD4369. Tangential section. Note rings round apertures and numerous acantho- pores. X50. Fic. 7. PD4368. Transverse section of another specimen showing tabulate zooecium and tubules in mesotheca. X50. Fic. 8. PD4370. Vertical section (of another specimen) of a piece which has broken away from the mesotheca. This shows outer wall (a) and inner wall (6) embedded in a translucent wall (c) which adjoined the mesotheca. Specimen is encrusted by a Fistuliporid (d). X50. Bull. B.M. (N.H.) Geol. 10, 4 PLATE 6 Monotrypa flabellata Owen Fic. 1. PD4344. Vertical section showing zooecia with integrate walls arising from a basal epitheca. X50. Fic. 2. PD4343. Tangential section showing a group of zooecia with larger apertures (lower right) adjoining those of the more normal size. X50. PLATE 6 Bull. B.M. (N.H.) Geol. 10, 4 S. ARCHANGELSKY _ BULLETIN OF _ MUSEUM (NATURAL HISTORY) Ren Vol. 10 No. 5 Rare. LONDON: 1965 FOSSIL GINKGOALES FROM THE TICO FLORA, SANTA CRUZ PROVINCE, ARGENTINA IB; SERGIO ARCHANGELSKY Museo de Ciencias Naturales, La Plata Research worker, National Research Council, Argentina Pp. 119-137 ; 5 Plates ; 19 Text-figures BUELETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. to No. 5 LONDON: 1965 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), ‘nstituted im 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical serves. Parts will appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 10, 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) 1965 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued June, 1965 Price Twenty-eight Shillings FOSSIL GINKGOALES FROM THE TICO FLORA, SANTA CRUZ PROVINCE, ARGENTINA By SERGIO ARCHANGELSKY SYNOPSIS The present paper deals with Lower Cretaceous Ginkgoales and associated seeds found in the Ticé Flora, Santa Cruz Province, Argentina. The plants were collected from two different localities, Tico Amphitheatre and Bajo Tigre Estancia. Two new species of leaves are referred to the genus Ginkgoites (G. tigvensis and G. ticoensis). A female structure found in close asso- ciation with G. tigvensis is referred to a new genus, Karkenia. Fragments of short shoots probably belonging to the same plant and roots found in the same bed are described. A new seed, associated with G. ticoensis, is described as Allicospermum patagonicum sp. n. INTRODUCTION THE material described in the present paper was collected by the writer during the years 1958-59 (Tico) and 1962-63 (Bajo Tigre). The first two excursions covered only the Tico Amphitheatre and included only a short visit to a new exposure with similar sediments bearing mummified plants in the Estancia Bajo Tigre. A longer excursion to this new locality was undertaken in 1962 when it became evident that the same formation known from Punta del Barco and the Ticd Amphitheatre extended N. and E. to cover a large area. In the Bajo Tigre the fossil plants are preserved in the same way as in Tico and occur in similar lenticular beds of brownish colour which makes them easy to distinguish in the field from the typical, mostly sterile, white tuffs. Several plant beds were discovered, the three main ones con- taining (1) Ginkgoites and other remains described here, (2) Piilophyllum and associated conifers of Brachyphyllum type, and (3) one long-leaved conifer with male and female cones in organic connexion, together with abundant freshwater mollusca. In 1963, further collecting in the same locality was undertaken, and several short excursions made to new areas which confirmed the extension of the same formation more than 100 km. northwards. Among the new localities discovered, one yielding excellent plant mummifications is known as Bajo Grande situated a few kilometres SW of the Petrified Forest of Santa Cruz (containing the petrified female cones of Araucanria mirabilis Speggazini). In Bajo Grande a rich plant association was found composed of many conifers (some with cones), Bennettitales and cutinized fern-like fronds, but no Ginkgoales. As a result of these excursions, a large amount of material has been collected, the description of which will take some years of future work. The age of the plant bearing beds (and the whole formation) has previously been considered as Upper Jurassic or Lower Cretaceous. However, the pollen content of the strata seems to confirm a Lower Cretaceous age, possibly Barremian to Hauterivian and the flora may well be the last plant association before the advent of the Angiosperms ; it may mark the uppermost limit of survival of some important fossil taxa. Little is known about Upper Mesozoic Ginkgoales in Argentina. The most comprehensive papers deal mainly with the abundant Triassic impressions. There are no records of Ginkgoites in the Lower and Middle Jurassic strata, but they are 122 FOSSIL GINKGOALES FROM TICO, ARGENTINA present in Upper Jurassic and Lower Cretaceous formations from Lago San Martin, Santa Cruz Province and from Graham Land, Antarctica (Halle 1913). Other records are from Tertiary strata in Chubut Province, Patagonia. The present material fills a gap in the knowledge of the cuticle of Ginkgoites leaves in Argentina and throws further light on the development of the female structures in this group. SYSTEMATIC DESCRIPTIONS GINKGOALES Genus GINKGOITES Seward 1919 Ginkgoites ticoensis sp. n. (PI. a, figs..5,'05; Pl 3; figs. 19-21 5 Bloat ttew27 ) Lext-tiges1 0) Dracnosis. Leaves with petiole 1 cm. long x 1 mm. wide ; lamina with radius of 2-3 cm., divided into 4 segments ; basal angle about 90°. Segments linear to oblong, 4-6 mm. wide, apices blunt, rounded, sometimes slightly lobed. Veins conspicuous, dichotomously forked, up to 12 per lobe at a concentration of 2 per mm. ; margins entire, substance dense, mesophyll thick, transversely cracked. Resin bodies absent. Both cuticles of same thickness (1-2, measured in folds). Upper cuticle having no stomata ; cells polygonal, not forming rows or bands, each cell bearing a strong hollow papilla ; cells between veins 25-35 in diameter, on veins somewhat rect- angular or elongated, also bearing papillae. Anticlinal walls straight, delicate, closely pitted. On veins, lateral anticlinal walls sometimes strongly thickened. Periclinal walls finely granular ; granules also present on papillae. Lower cuticle showing bands of rectangular cells along veins and margins, alter- nating with bands of polygonal cells between veins. Vein bands 6-10 cells wide, marginal bands more than Io cells wide. Cells along veins 15-20y wide ; lateral anticlinal walls may be strongly cutinized. Cells between veins polygonal, iso- diametric, about 20 ; anticlinal walls straight or slightly sinuous, closely pitted ; periclinal walls finely granular, but not as densely as on upper cuticle. A strong hollow papilla usually present on each cell, but sometimes absent. Stomata not placed in files, well spaced, variably orientated, rarely sharing subsidiary cells. Stomatal apparatus round, more often oval. Subsidiary cell group round or oval, composed of 5, sometimes 6 similar cells. Encircling cells usually present but not forming a complete ring, not differentiated from neigh- bouring epidermal cells. Subsidiary cells sometimes differentiated into polar and lateral, usually thickened on the edge of the pit, almost closing the pit and forming a cutinized ring ; occasionally (especially on polar subsidiary cells) there is a strong hollow papilla. Rarely there is no thickening on subsidiary cells. Guard cells feebly cutinized or not cutinized, slightly sunken in an oval pit. Mouth of pit 20-25 long. Trichomes absent. 123 FOSSIL GINKGOALES FROM TICO, ARGENTINA Neaw ec) oe Cio MED () eu g. I. Lower cuticle showing stomatiferous area. Fics. 1-6. Ginkgoites ticoensis sp.n. Fi Slide LP 24, Xitoo. Fig. 2. Upper cuticle. Slide LP 24, x1oo. Fig. 3. Two stomata with subsidiary cells in contact. Slide LP 24, x500. Figs. 4-6. Stomata. Slide LP 24, x 500. 124 FOSSIL GINKGOALES FROM TICO, ARGENTINA Ho.otyre. LP 5800a. MATERIAL. In addition to the holotype, LP 5801a, 5802, 5803a-5805a ; LIL PB 2559 (4); Brit. Mus. (Nat. Hist.), nos. V.51566, V.51579, V.51926. Slides LP 21-25. HORIZON AND LOCALITY. Lower Cretaceous, Baqueré Formation, lower member, Brachyphyllum nirandai Bed ; Ticd Amphitheatre, Santa Cruz Province, Argentina. Discussion. Ginkgoalean leaves are common in the Mesozoic. There is a remarkable uniformity in morphological and cuticular characters of the leaves referred to Ginkgo—Ginkgoites. It is difficult to differentiate species based only on size and shape of leaves ; the Recent Ginkgo biloba shows a remarkable variation of such characters. The cuticle probably constitutes a better argument for the differ- entiation of species. The two species of Ginkgottes described here are clearly referable to the Ginkgo complex, but the name Ginkgoites is preferred for them because of the age and the closely associated female structures which are very different from those known in the Recent genus. As most of the specimens possess good cuticle, comparisons have been made with those Ginkgoites taxa with similar preservation. Harris (1935) described several species of Ginkgoites from the Rhaeto-Liassic of Greenland. Only two of them are comparable to G. ticoensis. G. acosima Harris is typically larger (leaf-radius up to 8 cm.) and the distal parts of the segments are notched. Resin bodies have been observed. The basa! angle of the leaf varies more than in G. tzcoensis, and concentration of veins per centimetre is 15 against 20 in the Ticé species. The upper cuticle of G. acosima bears a few stomata while none is present in the Patagonian species. Ginkgoites taeniata (Braun) closely resembles the Tico species in size and shape. It differs in having resin bodies and fewer veins per centimetre (or per lobe). The upper cuticle has few stomata and papillae. Ginkgo huttont (Sternberg) from the Jurassic of Yorkshire (Harris 1948 : 192) is a widely variable leaf, and G. ticoensis can be matched in the range of variability. However, the lobes of G. huttoni are usually wider, having more veins (20-40) than in G. ticoensis. There are resin bodies and the cuticle is thicker (5 the upper and 2-3 the lower). Stomata and trichomes are also present on the upper cuticle. Ginkgoites longifolius (Phillips), also from Yorkshire, resembles G. ticoensis in size and shape. However, there are fewer veins in each lobe and resin bodies have been observed. Although the thickness of the cuticles is alike, there are no papillae on the epidermal cells of the upper cuticle in G. longifolius. Ginkgo ex gr. huttoni (Sternberg) described from the Wealden of East Siberia (Vachrameev & Doludenko 1961) approaches the Ticé species in having no resin bodies or stomata on its upper cuticle. But the epidermal cells are larger and trichomes have been observed on the lower cuticle. Also there are more veins per lobe (14-18). FOSSIL GINKGOALES FROM TICO, ARGENTINA 125 Ginkgoites obrutschewt Seward (1911), from the Jurassic of Chinese Dzungaria, has larger lobes and fewer veins per centimetre. Abundant resin bodies have been observed and there are no papillae on the epidermal cells of the upper cuticle. Ginkgoites cf. stbivica (Heer) as described by Yabe & Oishi (1933) from the Middle Jurassic of Manchuria, has fewer veins per segment and there are rudimentary stomata on the upper cuticle and no papillae. Ginkgoites marginatus (Nathorst) as described by Lundblad (1959) is usually smaller and has no resin bodies in the mesophyll ; there are more veins per lobe (12). The upper cuticle is without stomata and the epidermal cells have a strong median papilla. Lundblad (1959) considers Ginkgoites hermelini (Hartz) from the Liassic of Greenland and G. cf. s¢bivica as described by Yabe & Oishi (1933), to be synonymous with G. marginatus. Baiera cf. australis M‘Coy, as described by Halle (1913) from Lago San Martin, Santa Cruz Province, is similar to G. ficoensis, although Halle did not describe its cuticle. During a reinvestigation of the original material from Lago San Martin in the Stockholm Museum of Natural History I found some poor epidermal fragments which add to the knowledge of this species. The lobes of Halle’s specimens are more deeply dissected down the lamina than in G. ticoensis and there are fewer veins in each segment (5-10). The size and shape of the epidermal cells are similar, and so is the stomatal apparatus. However, no papillae are seen on the cells of the upper cuticle. The inclusion of the Lago San Martin specimens in Bazera is question- able (as indeed Halle states). They probably belong to Ginkgoites, because the leaves are clearly petiolate and the lamina is well developed and not wedged as are most Bazera species. The Lago San Martin formation which bears these fossils, is comparable in age to the Baquerd Formation (probably Lower Cretaceous). Some similar species from both floras have already been mentioned (Archangelsky 1963). Although the specimens described as Bavera cf. australis by Halle would be better placed in Ginkgoites, they are specifically different from G. ticoensis, but they may well be closely related forms. Ginkgoites tigrensis sp. n. (Pl. 1, figs. 1-4 ; Pl. 3, fig. 22 ; Pl. 4, figs. 23-26 ; Text-figs. 7-11) Dracnosis. Leaves with petiole up to 5 cm. long x 2:5 mm. wide. Lamina with a radius of 1-5 cm., usually divided into 4-8 segments ; basal angle go-180°. Segments lanceolate with rounded or obtuse apex, 3-8 mm. wide, margins entire ; veins conspicuous, dichotomously forked, crossing the lamina at a concentration of about 18-24 per centimetre, up to 15 present in a full sized lobe ; two veins seem to be present in the petioles. Oval, round or fusiform bodies between veins are rather few and scattered. 126 FOSSIL GINKGOALES FROM TICO, ARGENTINA Upper cuticle up to 3-4 thick (measured in folds). Epidermal cells rectangular on base of lamina and on veins of lobes, 20-25 wide ; between veins becoming more isodiametric, about 20-25 1n diameter. Anticlinal walls straight, thick, up to 5u, pitted ; periclinal walls with strong ridges, sometimes forming parallel striae, markedly granular ; papillae occasionally observed. Stomata absent on the petiole and base of lamina, but present on lobes between veins, not forming rows, variably orientated, scattered, sometimes sharing subsidiary cells. Fic. 7. Ginkgottes tigvensis sp.n. Outlines of different leaves to show variation in shape and size. All x1. Lower cuticle very thin ; shape and sculpturing on epidermal cells as for upper cuticle. Anticlinal cell walls thin (1-2), pitted. Stomata present. Stomatal apparatus on both cuticles circular or oval, with 4—7 similar haplocheilic subsidiary cells (usually 5-6). Encircling cells sometimes present (apparatus im- perfectly dicyclic). Sculpturing on subsidiary cells as for common epidermal cells, except for marked thickening (which occasionally is a papilla) on edge of pit ; thickenings sometimes fused to form a continuous rim of cutin. Guard cells feebly cutinized, slightly sunken. Mouth of pit 25—40u long. Trichomes absent. HoiotypPe. LP 5806. Counterpart, B.M. (N.H.) no. V.51571. FOSSIL GINKGOALES FROM TICO, ARGENTINA 127 MATERIAL. In addition to the holotype, LP 5541-54, 5557-71, 5573-74, 55934, 55944, 5631-330), 5636-39, 56434, 5644a, 5647-49), 5650, 5672, 5807-14, 5824-25 ; British Museum (Nat. Hist.) Nos. V.51490-V.51501, V.51572-78, V.51924-25. Slides LP 30-40, 145. Fics. 8-11. Ginkgoites tigvensis sp. n. Fig. 8. Upper cuticle showing distribution of stomata. Slide LP 30, x1o0o. Figs. 9, 10. Stomata. Slide LP 30, x500. Fig. 11. Resin bodies. Slide LP 32, x 40. 128 FOSSIL GINKGOALES FROM TICO, ARGENTINA HORIZON AND LOCALITY. Lower Cretaceous, Baqueré Formation, lower member; Bajo Tigre, Santa Cruz Province, Argentina. DESCRIPTION. This species occurs in the Estancia Bajo Tigre, about 10 miles E. of Ticd. Several fossiliferous beds have been discovered in sediments of similar colour and texture to those found in the Ticé Amphitheatre. The plants are also mummified in the same way. In the bed containing Ginkgoites tigrensis it is the dominant element, the associated plants being abundant female structures and their dispersed seeds, a few ferns and some twigs of conifers. I have included two slightly different types of leaf in G. tigvensis, they both occur together. Type A has four segments, type B has up to eight. Their cuticles are very similar but in type B the anticlinal walls are sometimes thinner and the sub- sidiary cells more often project as papillae instead of forming a continuous rim round the mouth of a stoma. Sculpturing of the cell surface may be more marked in type A. Intergradation in these features of the cuticle does, however, occur. The basal angle of the leaves is usually about 130°, but in small specimens it is up to 180°. It is very difficult to separate the delicate lower cuticle from the upper, but a few fragments were obtained by pulling with nail varnish and then treating them with dilute KOH. In the same locality but from a different bed (where Ptilophyllum and Brachy- phyllum are abundant) I collected two small leaves which may be compared with the small specimens found in the G. tigrensis Bed (LP 5824-25). Although cuticular fragments are small and show no important characters, the morphology of the leaves coincides. The largest petiole seen (Pl. 1, fig. 4) shows clearly two longitudinal furrows which I believe are veins. Discussion. Ginkgoites tigrensis differs from G. ticoensis in shape, size and cuticular structure. A character in which G. tigrensis differs from all other Ginkgoites, is the marked tendency of its resin bodies to be concentrated mainly along the margins of the segments. In all other species they are placed between the veins but scattered generally over the lamina as in Ginkgo biloba. Ginkgo huttont (Sternberg) usually has larger leaves and more veins per centimetre ; it also has trichomes on the epidermis and a thicker cuticle. Ginkgoites longifolius (Phillips) has a thinner cuticle and no stomata on the upper side of the leaf. There are also fewer veins per lobe (4-9). The lobes of G. tigrensis are wider than those of G. marginatus (Nathorst). The concentration of veins is 4-18 per centimetre while in the Patagonian species it is 18-24 ; also there are more veins per segment in G. tigrensis. Baiera cf. australis M‘Coy from Lago San Martin, Santa Cruz Province (Halle 1913) is smaller and the lobes are deeply dissected. Ginkgo biloba is clearly different in shape and size. The distribution of the resin bodies is also different as well as the stomatal apparatus. FOSSIL GINKGOALES FROM TICO, ARGENTINA 129 SEEDS AND FEMALE STRUCTURES CLOSELY ASSOCIATED WITH GINKGOITES LEAVES ALLICOSPERMUM Harris 1935 Allicospermum patagonicum sp. n. (Piva, figs’ 7—o Pli5) figs 28 3 ext-tig™ 12) DiaGnosis. Seeds oval, originally somewhat flattened, with slightly acuminate apex, typically 4-5 mm. long by 3 mm. wide. Seed consisting of an outer flesh and an inner stone enclosing various cuticles. Outer flesh about 1 mm. thick (usually represented by an empty space) ; stone 3-0-3-5 mm. long by 2°5 mm. wide, with micropilar prolongation 0-5-1-0 mm. long. Surface of stone marked with longitudinal bulges. On maceration, seed yielding the following cuticles. (1) The inner (megaspore membrane), densely and finely granular, showing no cell walls, thick (2~3u in folds) and resistant to maceration. (2) Thin cuticle (I-1-5p in folds) described as nucellus, partly covering megaspore membrane (probably not more than one half of it). Cells markedly elongated (120 or more) and 8-I0u wide. Cell surface flat, not ornamentated. Cell walls straight, becoming thicker and pitted towards apex ; end walls straight. (3) Poorly preserved thin cuticle (less than ry in folds), finely granular ; cells isodiametric (15) or slightly elongated (24u X 5p) with straight walls. Small hollow papillae, one per cell, are sometimes present. This membrane is regarded as the inner lining of the integument. (4) Thick cuticle (outer cuticle of integument) enclosing stone and flesh. Cuticle faintly marked with somewhat isodiametric cells, 10-15 in diameter, with straight thick walls. Surface may be granulose with many adherences. Stomata absent. The apex of the nucellus where a pollen chamber might be situated, and a cutinized lining of the micropylar canal were not seen. A round scar sometimes seen at the base of the stone probably represents the hilum. IFOLOTYPE. LP 5821a. MATERIAL. In addition to the holotype, LP 5804), 5822a, 5823a, 5863c ; LIL PB 2559(3) ; British Museum (Nat. Hist.), V.51580—81(2). Slides LP 49-53, 125-129. Horizon AND Locarity. Lower Cretaceous, Baqueré Formation, lower mem- ber, Brachyphyllum mivandai Bed ; Ticd Amphitheatre, Santa Cruz Province, Argentina. DeEscRIPTION. The seeds are preserved in one of two ways. A. Without compression. The seed then forms a cavity enclosing the stone which is itself filled with fine sediment. On the surface of this stone (or possibly the internal cast of the stone) there are some coaly fragments which yield a few membranes when macerated. These membranes are situated on the outer surface of the coal. Two cuticles are usually present. The outer and thicker is similar to cuticle 4 of the diagnosis, while the inner, poorly preserved, corresponds to cuticle 3 (inner lining of the integument). Cuticles 2 and I are sometimes also present, adhering to the inner 130 FOSSIL GINKGOALES FROM TICO, ARGENTINA surface of the coaly fragments. The gap seen outside the coal may be due to the shrinkage of the stone. B. With compression, but no infilling with sediment. The seed then forms a disc and it is possible to prepare its cuticles. The flesh forms a compressed border round the thicker substance of the stone. In such specimens the megaspore membrane and the nucellus are usually better preserved. Clearly, the nucellus cuticle is single and was not seen fused to the inner lining of the integu- ment as in the seed described by Harris (1944 : 427, text-fig. 3D). Fics. 12-16. Allicospermum patagonicum sp. n. and Karkenia incurva gen. et sp. n. Fig. 12. Allicospermum patagonicum sp. n. Diagrammatic section of seed to show the probable extent and position of the different membranes (the stone is dotted). Io. Figs. 13-16. Karkenia incurva gen. et sp. n. Fig. 13. Diagrammatic section of seed to show the probable extent and position of the different membranes (small dots represent the stone ; thick dots are resin cavities). 10. Fig. 14. Cells of nucellus. Slide LP 42, X 425. Fig. 15. Outermost layer of cells (outer integument) with a few resin cavities. Slide LP 44, x500. Fig. 16. Cells of the inner integument, faintly marked. Slide LP 42, x 425. Discussion. Allicoshermum retimivum Harris from the Jurassic of Yorkshire is similar in size and shape, though slightly wider. Fine differences are : A. patagoni- cum has a granulose outer cuticle instead of a smooth one, and the stone, on macera- FOSSIL GINKGOALES FROM TICO, ARGENTINA 131 tion, yields no reticulum as does A. vetimirum. The Yorkshire species has not been identified with any leaf. Seeds looking rather like A. patagonicum are associated with Ginkgoites leaves in various floras and have sometimes been more or less definitely linked with them. There is, for example, A. xistwm with Ginkgoites taemiata Harris from Scoresby Sound, Greenland, the cuticles of that seed being known. A. patagonicum is associated with Ginkgoites ticoensis in the Brachyphyllum nmuvandait Bed. Neither of these taxa has been found in other horizons or localities so far studied. Fics. 17-18. Karkenia incurva gen. et sp. n. Fig. 17. Reconstruction of the entire female structure (based on LP 5817). 4:5. Fig. 18. A few inverted ovules inserted on the main axis. x5. (C. Freile del.) 132 FOSSIL GINKGOALES FROM TICO, ARGENTINA Genus KARKENIA nov. , which in the language of the Tehuelche ” The name Karkenia is from “ karken ’ Indians (Southern Patagonia) means “‘ female or woman DiaGnosis. Oval or elongated seed-bearing structures composed of a central axis with irregularly disposed pedunculate ovules. Ovules round or oval, curved (atropous) facing axis with micropylar end, densely packed, composed of four cutinized membranes belonging to the megaspore, nucellus, inner and outer integu- ments. Seeds developing a conspicuous stone. For discussion of genus see below. TYPE SPECIES. arkenia incurva sp. n. Karkenia incurva sp. n. (Pl. 1; fig: 10 3 -Pl. 2; figs: 1a, 145 16, 18%; Pil 5) shes. 20-32) -sihexta hie saeres 1G) DiAGNosis. Seed-bearing structure up to 4:5 cm. long by 1-3 cm. wide, tapering gradually towards base and apex. Up to 100 ovules present, densely packed and irregularly disposed, attached by delicate peduncle to main central axis I-2 mm. wide. Ovules curved (atropous) with micropylar end close to main axis, round or oval, 3 mm. long by 2-2:5 mm. wide. Ovules composed of several cutinized membranes, commencing from the inside : (1) Megaspore membrane ; structureless, finely granulose, less than Ip thick (in folds). (2) Nucellus membrane ; usually Fic. 19. Tentative reconstruction based on the dwarf-shoots, leaves (Ginkgoites tigrensis) and female structures (Karkenia incurva). x2. (C. Freile del.) FOSSIL GINKGOALES FROM TICO, ARGENTINA 133 closely fused to membranes 1 and 3, extending down to near base, I-1°5y thick (in folds), showing markedly elongated cells more than 1ooy long by 15u wide, surface flat, not granulose. At the top, nucellus projecting as a short acute micropylar beak. (3) Inner layer of integument, probably extending down to near base of ovule, showing no definite structure but small granules. (4) Outer layer of integument, faintly cutinized membrane, showing cells about 50u long by 15-25 wide, finely granulose with occasional small hollow papillae. Between membranes 3 and 4, small round resin bodies (?) from 2—30y in diameter occur. HototypPe. LP 5816. MATERIAL. In addition to the holotype, LP 5580-84, 5598a, 5599, 5631a—33a, 5635, 5640-41, 5647a—-49a, 5814), 5815a, 5817b-5819; British Museum (Nat. Hist.), V.51499-503, V.51582-84. Slides LP 41-47, 131-144. HORIZON AND LOCALITY. Lower Cretaceous, Baqueréd Formation, lower mem- ber, Ginkgoutes tigrensis Bed ; Bajo Tigre, Santa Cruz Province, Argentina. DESCRIPTION. Seed-bearing structures, all of one kind, are very abundant in the bed where Ginkgoites tigrensis occurs. Very few other plant remains are present in association, all of which are rare : two ferns, one referred to Cladophlebis and the other to Sphenopteris, and a conifer with long linear leaves. This conifer becomes more abundant in an upper bed and is known to occur with male and female cones organically attached to the branches. G. tigrensis and Karkenia incurva are not only abundant fossils in this bed but are always found in close association. Many of the fructifications are found entire or slightly broken, but with the ovules still attached to the main axis. Detached or shed seeds are also very abundant and they clearly show an egg-shaped stone, finely striated in surface view, with a marked acuminate apex. The stone is surrounded by the remnants of the outer fleshy layer about 1 mm. thick. The size of the ovules and the seeds is similar. There is no trace of the peduncle when the seeds are found isolated, except an occasional slight thickening at the hilum area. Several transfers of these fructifications have been prepared. All show the irregular insertion of the ovules which have no definite phyllotaxis. The ovules face the axis of the fructifications with their micropylar end, or may be slightly turned from that position, but never erect. The peduncles are short, not much longer than the total length of the ovules. Ovules are crowded and compact, suggesting a cone-like structure. The peduncles are attached only to the hilum sector of the ovules, where a slight expansion may sometimes be seen ; for the rest they are free and easy to separate from the ovules. Therefore, the inverted position of the ovules corresponds to an atropous incurved type. No bracts or laminar appendages were observed in relation to the ovules or the peduncles. Karkenia may be defined as having a central axis bearing pedunculate, naked and inverted 134 FOSSIL GINKGOALES FROM TICO, ARGENTINA ovules. I suggest that the peduncle-ovule structure is morphologically a lateral branch of the main central axis, being analogous but not homologous to the mega- sporangiophores of other groups. There is no specialization of the peduncles, which are merely lateral appendages of the central axis. As for the cutinized membranes of the ovules, there is no doubt about the shape and size of the megaspore. The nucellus, closely attached to the megaspore and sometimes to the inner layer of the integument, is also cutinized down to near the base of the seed. The micropylar projection is seen as a very short apical extension of the nucellus and is also cutinized. The micropylar canal was not clearly seen and no pollen was found in connection. There is doubt about the structure and extent of the inner lining of the integument. It is a structureless membrane, sometimes showing very faint marks which may be cell outlines, but this is not sure. Granules are clearly seen. Also, there is doubt about the structure of the external surface of the integument, although some cells have been observed. It is similar to the inner lining of the integument but thicker. Granules and some papillae were also seen, but there are no stomata. Between these two membranes round bodies are found isolated or in large groups of 20-30 or more. They may well be resin bodies, which are more likely to be preserved than the mucilage cavities present in the Recent Ginkgo biloba. I believe these round bodies are natural features of the seeds, because they are constant in all the specimens observed and are of the same type ; often, when detached, they leave a round impression on the integument membrane. Associated dwarf-shoots and roots. In close association with Karkenia and Ginkgoites I have found small, short branches which are probably dwarf-shoots of the same plant. One of these specimens shows a shoot, 3:3 cm. long by 5 mm. wide, bearing three dwarf-shoots at intervals of about 0-8 cm. The largest dwarf-shoot is 1-5 cm. long by 4 mm. wide. The widest seen was I cm. Each of these shoots is crowded with spirally disposed rhomboidal scars, their longest axis being horizon- tal. The width of these scars is 1-2 mm. and corresponds to the size of the main axis of the female structures and the petioles of the leaves. In the middle of these cushions one or two (?) small circular scars are seen. They may correspond to the vascular bundles. In its distal part one of these dwarf-shoots shows the remnants of an axis and a few ovules of the type described for Karkenia. The organic attach- ment between shoot, axis and ovules may be inferred from the continuous brown colour which is clearly different from the adjacent light colour of the matrix (PI. 2, fig. 12). In close association with the previously described material, many fragments of roots occur (Pl. 2, figs. 15, 17). Some of them cross the sedimentary layers obliquely, while others are lying in the sedimentary planes, which, however, are not clearly defined. The Ginkgoites leaves, complete Karkenia structures, dwarf-shoots and the roots, are situated in the boundary of two different sediments ; the lower sector, bearing most of the organic remains, is a pale brown, fine-grained rock succeeded FOSSIL GINKGOALES FROM TICO, ARGENTINA 135 by a white coarse-grained sediment. The plants were found in abundance only a few millimetres above and below this boundary plane. On top of this sector, only detached Ginkgoalian leaves and fragmentary conifers are present. The presence of roots may well indicate that the most productive part of the plant bed was deposited 7m situ, and therefore the plants included have not suffered a long transport. The roots are composed of a main root about 0-4 cm. wide, giving off secondary roots, irregularly disposed, at acute or right angles. These secondary roots give rise to delicate rootlets which are typically crowded with round bodies, 1-2 mm. in diameter, irregularly situated. No organic remains were found except for a few carbonized fragments which dissolved completely under maceration. These round bodies may well belong to some type of mycorrhiza. MATERIAL. Dwarf-shoots : LP 5587-88, 5642, 56430, 56446, 5645-46 ; British Museum (Nat. Hist.), V.51575. Roots: LP 5575-79, 55930, 55946, 5598) ; British Museum (Nat. Hist.), V.51504—-05. Discussion. Guinkgoites tigrensis, Karkenia incurva and the dwarf-shoots described may belong to the same plant. The close association and the absence of other forms which could possibly bear female structures are the only arguments to suggest this identity. Comparisons of Karkemia incurva can only be made with the Recent Ginkgo biloba and with Trichopitys heteromorpha Saporta, a Permian Ginkgoalean plant whose female structures are inadequately known. Trichopitys heteromorpha Saporta as described by Florin (1949) has in common with Karkema incurva the irregular distribution of the ovules on a main axis, and their inverted position. The number of the ovules is, however, smaller, but the size is similar. The pedicels which bear the ovules and the main axis are wider in Trichopitys, and the whole fructification may be longer. Also, the ovules of the Permian genus are separated and do not form a compact structure as in Karkenia. The leaves of Tvichopitys are very different, not having a developed lamina. No dwarf-shoots are known to occur. Neither in Trvichopitys nor in Karkenia is there a collar at the base of the ovule. Ginkgo biloba has a female structure composed of one long stalk bearing two terminal ovules, one of which usually aborts. At the base of the ovules there is a cup-like structure known as a collar. The stalks are spirally disposed on short branches (dwarf-shoots) in the axils of young leaves. Abnormal cases do occur ; one of them shows several ovules irregularly disposed on a main axis. These ovules have long pedicels but are not inverted. This case suggests, as stated by Florin (1949), that the ancestors of the group must have been multiovulate structures, like Trichopitys and Karkenia. The main differences are the absence of a collar and the inverted position of the ovules, which Karkenia has probably retained from primitive forms. The collar is considered by Florin (1949) as a secondary feature, related to the insertion of the erect ovules, and Karkenia (as well as Trichopitys) is in accordance because there is no collar but an inverted position of the ovules. 136 FOSSIL GINKGOALES FROM TICO, ARGENTINA The absence of any laminar structure in direct relation to the ovules, suggests that the abnormal cases of leaves bearing ovules (found in the Recent Ginkgo) are secondary phenomena. Karkenia may well be an intermediate type of female structure (“ flower ’’) between Tvrichopitys and Ginkgo, having undergone some fusion and reduction processes since Permian times, but still retaining some primitive features. It is difficult to establish the degree of relationship between Tvichopitys and Karkenia. Possibly each of the “ sporangial trusses ”’ of Tvichopitys can be homo- logous with the single ovule and its peduncle of Karkenia. In such a case, the ‘ sporangial trusses ’’ must have fused to form a compact structure, while the main axis was strongly reduced. The leaves (sterile telomes) of such branches became reduced and further disappeared, while they persisted on the entirely sterile branches During all these changes, probably the “short shoot” habit was attained, with sterile and fertile telomes clearly differentiated. There is no information about all these possible intermediate types (Permian—Lower Cretaceous). It is perhaps easier to understand the processes which followed in order to reach the Ginkgo type of flower. Every compact structure of Karkenia may be homologous with the Ginkgo peduncle and ovules. This state was attained by reduction of Karkenia peduncles, and fusion of its ovules, followed by their erection (forming a collar as a secondary feature). It may be suspected that the erect position of the ovules and the formation of the collar, is probably a rather recent phenomenon, possibly post-Neocomian, when the Angiosperms began to dominate. Text-fig. Ig is a reconstruction of Karkenia borne on dwarf-shoots, together with Ginkgoutes tigrensis leaves. ACKNOWLEDGEMENTS I would like to express my gratitude to Professor T. M. Harris (Reading University) for many important suggestions ; to Mr. F. M. Wonnacott for critically revising the manuscript. Thanks are due to Professor O. Selling for permission to examine specimens from the Halle Collection in the Stockholm Museum of Natural History. I am indebted to the National Oil Company (Y PF) and the Alumine Mining Company, both from Argentina, for help during the field excursions. I am obliged to Mr. L. Ferreyra of La Plata Museum of Natural History for the photographs which are included in the present paper. FOSSIL GINKGOALES FROM TICO, ARGENTINA 137 REFERENCES ARCHANGELSKY, S. 1963. A New Mesozoic Flora from Ticd, Santa Cruz Province, Argentina. Bull. Brit. Mus. (Nat. Hist.) Geol., London, 8 : 45-92, pls. 1-12. Fiorin, R. 1949. The Morphology of Trvichopitys heteromorpha Saporta, a Seedling Plant of Palaeozoic Age, and the Evolution of the Female Flowers in the Ginkgoinae. Acta Horti Bergiani, Stockholm, 15 : 79-109, pls. 1-4. Harte, 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, pls. 1-5. 1913a. The Mesozoic Flora of Graham Land. Wiss. Evgebn. schwed. Sudpolarexped. (1901-1903), Stockholm, 3, 14 : 1-123, pls. 1-9. Harris, T. M. 1935. The Fossil Flora of Scoresby Sound, East Greenland, IV. Ginkgoales, Coniferales, Lycopodiales and isolated fructifications. Medd. Gvonland, Kjobenhavn, 112 : 1-176, pls. 1-29. 1944. Notes on the Jurassic Flora of Yorkshire, 11. Allicosbermum vetimivum sp. nov. Ann. Mag. Nat. Hist., London (11) 11 : 424-428, text-fig. 3. 1946. Notes on the Jurassic Flora of Yorkshire, 30. Ginkgoites longifolius (Phillips) n.comb. Ann. Mag. Nat. Hist., London (11) 13 : 20-24, text-figs. 6, 7. —-— 1948. Notes on the Jurassic Flora of Yorkshire, 38. Ginkgo huttoni (Sternberg) Heer. Ann. Mag. Nat. Hist., London (12) 1 : 192-207, text-figs. 4-7. — 1948). Notes on the Jurassic Flora of Yorkshire, 39. Ginkgo digitata (Brongniart) Heer. Ann. Mag. Nat. Hist., London (12) 1 : 207-213, text-figs. 7, 8. Lunpsiap, A. B. 1959. Studies in the Rhaeto-Liassic Floras of Sweden. II:1. Ginkgo- phyta from the mining district of N.W. Scania. KK. svenska VetenskAkad. Handl., Stock- holm, 6, 2 : 1-38, pls. 1-6. OIsHI, S. 1933. A study on the cuticles of some Mesozoic Gymnospermous plants from China and Manchuria. Sci. Rep. Téhoku Imp. Univ., Sendai (2, Geol.) 12 : 239-252, pls. 1-4. SEWARD, A.C. 1911. Jurassic Plants from Chinese Dzungaria. MJém. Com. Géol. St. Pétersb. (n.s.) 75 : 1-61, pls. 1-7. VACHRAMEEV, V. A. & DoLuDENKO, M. P. i961. Upper Jurassic and Lower Cretaceous Floras of the Burenska Basin. Tvud. geol. Inst. Akad. Nauk S.S.S.R., Moscow, 54 : 1-136, pls. 1-60. [In Russian.] YasBeE, H. & Olsu1,S. 1933. Mesozoic Plants from Manchuria. Sci. Rep. Téhoku Imp. Univ., Sendai (2, Geol.) 12 : 195-238, pls. 1-6. PLATE 1 Ginkgottes tigvensis sp. 1. Fics. 1-3. Different leaves showing variation in size, shape and lobation of the segments. Fig. 1 (LP 5807) X12; Fig. 2 (LP 5824) x2; Fig. 3 (B.M.N.H. V.51571) counterpart of holotype, X I-I. Fic. 4. Leaf with long petiole. LP 5552, XI. Ginkgoites ticoensis sp. n. Fic. 5. Holotype (LP 5800) x1-5. Fragments of Brachyphyllum mivandai Arch. and Ruflovinia sierra Arch. are also seen. Fic. 6. Fragmentary leaf to show venation. LP 5801, X1°5. Allicospermum patagonicum sp. n. Fics. 7, 8. Isolated seeds showing carbonized remains of the outer fleshy integument adhering to the main body. Fig. 7, LP 5804, x8; Fig. 8, LP 5822, x8. Fic. 9. Several seeds in different positions. LP 5821, X1°5. Karkenia incurva gen. et sp. n. Fic. 10. Analmost complete fertile structure (left) together with a leaf of Ginkgoites tigrensis. B.M.N.H. V.51582, XII. Bull. B.M. (N.H.) Geol. to, 5 PLATE 1 PILI, Karkenia incurva gen. et sp. n. Fic. 11. Enlarged fragment of a female structure, showing main axis (bottom) and several ovules (some are inverted). LP 5817, 4:5. Fic. 14. Several fragments of female structures and part of a Ginkgoites tigvensis leaf. ILIP Fw, S< it Fic. 16. Fragments of female structures. Towards the left, a small fragmentary pinna of Cladophlebis sp. is also seen. LP 5815, X1. Fic. 18. Two isolated seeds. LP 5818, X1°5. DwarF SHOOTS Fic. 12. Enlarged fragment showing at the top two seeds of Karkenia incurva. LP 5645, x 8. Fic. 13. An almost complete branch showing rhomboidal scars. LP 5643, x 4. Roots Fic. 15. Enlarged rootlets showing round bodies attached. LP 5598, x8. Fic. 17. A root (white arrow) traversing the sediment. LP 5593, XI. Bull. B.M. (N.H.) Geol. 10, 5 IPL AMINE, 2 ee PLATE 3 Ginkgoites ticoensis sp. n. Fic. 19. General aspect of both cuticles (the lower towards the right). Slide LP 24, x 180. Fic. 20. Lower cuticle. Sector between veins showing distribution of stomata. Slide P25) <180: Fic. 21. Two stomata showing strong thickenings overhanging mouth of pit. Slide LP 25, x 800. Ginkgoites tigvensis sp. 0. Fic. 22. Fragment of leaf segment showing three resin bodies near left margin. Slide ILIP Sy), SX 1). PLATE 3 Bull. B.M. (N.H.) Geol. 10, 5 PLATE 4 Ginkgoites tigvensis sp. n. Fic. 23. Fragment of upper cuticle showing few scattered stomata. Slide LP 30, x 180. Fic. 24. Fragment of upper cuticle showing elongated cells on veins and a sector between veins, almost devoid of stomata. Slide LP 40, x 180. Fic. 25. Stoma showing an almost continuous rim of cutin overhanging mouth of pit. Slide LP 4o, x 850. Fig. 26. Stoma. Slide LP 30, x 800. Ginkgottes ticoensis sp. n. Fic. 27. Stoma showing guard cells slightly cutinized. Slide LP 22, x 800. Bull. B.M. (N.H.) Geol. 10, 5 PLATE 4 PLATE 5 Allicospermum patagonicum sp. Nn. Fic. 28. Nucellus membrane (left) and megaspore membrane (right). Slide LP 51, X175. Karkenia incurva gen. et sp. n. Fic. 29. Part of a seed showing groups of resin (?) bodies. Slide LP 139, x60. Fics. 30. 31. Two apical portions of nucelli. Fig. 30, Slide LP 47, x175 ; Fig. 31, Slide R425 x60; : Fic. 32. Megaspore membrane. Slide LP 45, 175. (N.H.) Geol. 10, 5 Bull. B.M. PRINTED IN GREAT BRITAIN BY THOMAS DE LA RUE & COMPANY LIMITED LONDON. 7 ia . . . ‘ i. Ve . ‘ Wik» « } THE GENERIC POSITION OF be OSM UNDITES DOWKERI CARRUTHERS ME. J. CHANDLER Vol. 10 No. 6 LONDON: 1965 imtroGeENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS BY MARJORIE E. J. CHANDLER Pp. 139-161 ; 12 Pls. ; 2 Text-figures BULLELIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 10 No. 6 LONDON: 1965 Issued July, 1965 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, 1s issued im 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. 10, No. 6 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) 1965 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Price Two Pounds Two Shillings Her GENERIC POSHMION OF OSMUNDITES DOWKERI CARRUTHERS By M. E. J. CHANDLER SYNOPSIS A silicified rhizome from the Thanetian of Herne Bay, Kent, formerly described as Osmundites dowkevi Carruthers, is now referred to the sub-genus Plenasium of the living Osmunda. This has been possible owing to the discovery of another better preserved rhizome in Thanet, coupled with fuller knowledge of the family Osmundaceae. Recent studies by W. Hewitson show clearly characters which distinguish the three living genera and various sub-genera of the Osmundaceae. The fossil material is described in detail. Especial attention is given to the distribution of sclerenchyma in the leaf base and wing stipules. The presence of two protoxylem groups in the leaf traces within the cortex is demonstrated. The form and number of the xylem bundles in the cylinder of the rhizome is displayed. A close affinity with rhizomes described by C. A. Arnold from the Eocene of Clarno, Oregon, is emphasized. A possible connexion between the Thanetian rhizomes and the common Eocene foliage described as Osmunda lignitum (Giebel) by Heer and Gardner is suggested in view of the fact that this foliage appears also to belong to Plenasium. INTRODUCTION THE fossil species Osmundites dowkeri Carruthers, based on a single rhizome from the Thanetian of Herne Bay, Kent, has been known since 1870. It has been described or mentioned in several publications but without any full and accurate descriptions of the anatomy. This omission was due in the first place to the fact that the im- portance of some of these details was not realized but in later works is, in part at least, to be attributed to the poor condition of the solitary specimen hitherto known. The published evidence shows clearly the Osmundaceous characters but no attempt was made previously to establish the true generic position, hence the name Osmun- dites. The particular choice of name may have been dictated originally by the inherent mistrust which some palaeobotanists invariably showed about the use of Recent generic names for incomplete fossil material. Its retention in later work was due to the immense amount of research needed on living genera before relationship with Todea and Leptopteris could be excluded. Such research of course involved the preparation of many slides from rather intractable material to discover what characters in rhizomes of the different genera were of diagnostic value so that the knowledge could be applied to fossils. This, coupled with the difficulty of obtaining for dissection a sufficiently large range of living forms, has been a stumbling block to further research. But without it no sound opinion on the particular generic affinity of the fossil could be given. 142 THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS RECENT INCENTIVES TO RESEARCH ON OSMUNDITES DOWKERI The discovery of a better preserved fern rhizome at Herne Bay by D. J. Jenkins (Chandler 1961 : 51, pl. I, figs. 1, 2) stimulated the desire to determine the generic position of Osmundites dowkert more accurately. Fortunately in 1962 Hewitson published a comprehensive study of the family Osmundaceae demonstrating that even if the rhizomes only are known, Osmunda can be distinguished from Todea and Leptopteris. He further showed the range of characters within the Recent genus Osmunda and how these could be applied in separating its three sub-genera, Osmunda, Osmundastrum and Plenasium. Hewitson’s research made it clear beyond doubt that the two Thanetian rhizomes not only belong to the same genus and species but to the sub-genus Plenasium of Osmunda itself. They should therefore henceforward be known as Osmunda (sub-genus Plenasium) dowkeri (Carruthers). SUMMARY OF PREVIOUS WORK AND GENERAL DESCRIPTION OF MATERIAL Both known specimens are silicified, the holotype being a large piece of a mature rhizome, whereas the newly found specimen is smaller and younger, apparently representing the subapical region of a young plant or a young branch of a plant. The holotype (V. 29629) has twice been described by Carruthers (1870 : 349, pl. 24, figs. 1-3 ; pl. 25, figs. 1, 3, 4 ; and more briefly 1872 : 52, pl. 2, fig. 8). It was also mentioned by Gardner & Ettinghausen (1880 : 53) and by Seward & Ford (1903 : 254) but these authors made no attempt to redescribe the rhizome or to discuss its affinities in any detail. Kidston & Gywnne-Vaughan (1907 : 768) gave a further account but apart from an excellent description of the diarch roots they added little to what was already known although they did stress the strong curve taken up by the xylem of the leaf trace almost immediately after it has left the stele of the stem. They also gave a diagrammatic transverse section of the wing stipules and leaf base (1907, pl. 6, fig. 5) which they believed provided the only distinctive specific character in Osmundaceous stocks. The fungus infested condition of the tissues and resultant deterioration prevented them from giving any further description of the transverse section which they show in pl. 4, fig. 21. The magnification of this figure is too small to show the really significant features which are also much obscured by partial disorganization both of the specimen and of the slide. Arnold (1952 : 72), in describing two Osmundaceous rhizomes from the Eocene Clarno Beds of Oregon, referred briefly to Osmundites dowkeri because it closely re- sembled his new species, Osmundites Chandleri, in the strong curve of the emergent leaf trace. Nevertheless he stated (p. 75) that the two were so remote geographically and geologically as not to justify even a consideration that they might be the same. In both Thanetian specimens, as in all Osmundaceae rhizomes, the small true stem is surrounded by a thick mantle of spirally arranged leaf bases which accounts for most of the thickness seen. In neither is evidence of fronds or fructifications preserved. Hewitson (1962: 88) confirmed Kidston & Gywnne-Vaughan’s views as to the importance of the distribution of sclerenchyma in the leaf bases and demonstrated that this character could be used to discriminate between species, THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS 143 sub-genera and genera. On the evidence of petiole base structure it can be stated definitely therefore that the two Herne Bay specimens are specifically and generically identical despite differences of size and age. The specimen V. 29630 is much better preserved than the original holotype. Added to this the late W. N. Croft prepared from it an exceptionally fine thin section for he was a master craftsman in this as in all such matters. The better preserved material is therefore described here, before considering the detailed characters still visible in the holotype, for it has much to contribute towards the fuller understanding of the larger older rhizome. DETAILED EXAMINATION OF V.29630 General Considerations. The rhizome found on the shore at Hampton, Swalecliff, Herne Bay, and like the holotype presumed to come from the Thanetian was figured by Chandler (1961 : 51, pl. 1, figs. 1, 2) to show the gross characters. The maximum length preserved was 53:2 mm. and the diameter 35 x 45 mm. The rhizome broadens slightly upwards the maximum diameter at the lowest point being only about 27 mm. The cross section is elliptical. The outer surface has been abraded to such a degree that the stipes above the stipule wings and the extreme upper ends of these wings have always been removed. The upper surface of the specimen is a deep basin-like depression with rim of unequal height owing to differential breakage and abrasion. The form of this basin is dictated by the angles the stipe bases form with the true stem, about 23°. Such a narrow angle is most nearly approached in the living sub- genus Osmundastrum (15°-25°) but this has proved to be very unlike the fossil in other respects. In Plenasium the corresponding angle is 30°-45° ; in Osmunda (sub-genus) 25°—40° ; in Todea and Leptopteris about 30° (Hewitson 1962: 73). The walls of the basin are formed by an amorphous silica casing which must have filtered in solution into the interstices between the silicified petiole bases. Asa result the true ventral surfaces of petioles and wing stipules are not exposed except in minute patches where the casing has chipped away. The length and form of the wing stipules cannot therefore be recorded but the 30 mm. depth of the basin with wing stipules still in transverse section on its upper rim indicates that they must have been more than 30 mm. long. The bottom of the basin is formed by pith, xylem cylinder and amorphous silica occupying the space once filled by cortex, this tissue having disappeared. The basin like form of the upper end of the specimen indicates that growing tip and younger leaves had been torn away prior to fossiliza- tion leaving the tougher tissues of the somewhat older leaves just below. These leaf remains have become silicified cell by cell around the true stem. A similar basin was seen in Todea barbara when tip and youngest fronds were removed. After silicification superficial chemical action appears to have operated differentially in the apical region. The silicified xylem cylinder with protruding roots has been left in relief by etching out most of the inner cortex. The cavities resulting from this etching are now filled by redeposited coarse irregular grains of silica. The silicified leaf traces beyond the outer cortex are structurally intact except that the stout 144 THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS cylinder of sclerenchyma which delimited each has been dissolved leaving cylindrical or variously distorted hollows. These reproduce the irregular forms of the leaf stipes in section. The siliceous casing which penetrated between the stipes now remains as a complicated upstanding network (Pl. 5, fig. 11). Within the casing are embedded hairs, roots and wing stipules all full of structural detail. The re- moval of what in the living plant was the strongest and most resistant part of the stipe niay be due to incomplete penetration by silica of the dense thick walled sclerenchyma cells. These changes are secondary and purely superficial phenomena for the sclerenchyma cells are perfectly preserved inside the rhizome as shown in a section about half way down the specimen. But even inside the rhizome the parenchyma of the inner cortex has gone, its place being taken in the slide by amor- phous silica except in one small patch. The slide (V.29630a) is the outcome of the late W. N. Croft’s technical skill and patience. He describes its preparation in his working notes in the following words : “ ground one side of slide flat. Treated this with polystyrene with much solvent (benzene) in which alizarin had been ground. Scraped off dried crust with razor blade. Examined surface with binocular after wetting with cedar oil. Staining was fairly satisfactory stain having been taken up by some of the xylem strands, although patchily. Hardening of slice was in any case necessary as it was somewhat porous ”’ V.29630 was clearly a relatively young rhizome for its stipes were soft and the wing stipules flexible. Hence the regularity of arrangement seen in the firm older holo- type is not present. (cf. description of stipes and wing stipules on p. 146). Anatomical Structure. The Pith, about 1:5 to 2 mm. in diameter, is formed of typical parenchyma as seen in transverse section (part only being represented by amorphous silica). The cells are commonly 0-05 to 0-1 mm. in diameter, rarely 0:14 mm. At the circumference of this tissue there are a few rows of cells, varying in number, only about half this size. Some of the larger cells towards the circumference show dark staining. In view of the limited material, no longitudinal section is available. The Xylem Cylinder and Leaf Traces. The counting of the xylem bundles which form the cylinder has been carried out consistently in all sections examined whether of this or the holotype and in accordance with a plan suggested by Hewitson to secure uniformity of treatment. Without such a plan the number would vary considerably with the personal factor. Hewitson, throughout his research, treated bundles connected by even a single tracheid as one. Otherwise, he explains, “cases are encountered where it is difficult to make a decision’’. On this basis there are twenty strands in V.29630a, a large projecting horse-shoe trace being regarded as one although one of its limbs is almost but not quite severed from the arc (Pl. 2, fig. 3; Pl. 3, fig. 6). A noticeable feature in the xylem ring is a somewhat oblique horse-shoe on a radius at right angles to the larger horse-shoe above described. It has one limb completely separated (counted separately therefore) the remaining one forming a query-shaped bundle with the hook towards the circumference of the stem (Pl. 3, fig. 6). On the opposite side of the xylem ring (below right) is a query-shaped THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS 145 pair of bundles with adjacent hooks (PI. 2, fig. 3 ; Pl. 3, fig. 5). There is also a query-shaped bundle bending to unite with an oval bundle (PI. 2, fig. 3 at 2 o’clock). Various stages of the development of one of the large horse-shoes from the fusion of two query-shaped ones can be seen in the slide. In addition there are a number of pointed or pointed-oval bundles one or two of which show a tendency to be hooked on account of a slight excavation of the outline on one radial margin. The development of leaf traces can be admirably seen by comparing slide V.29630a (Pl. 2, fig. 3) with the smooth lower surface of V.29630 from which it was cut, this surface representing a slightly higher level in the rhizome (PI. 4, fig. 7). The largest horse-shoe strand of the slide has separated into two distinct bundles each with a small median notch on its inner side (PI. 4, fig. 7). A completely separated deeply C-shaped leaf trace formed by the separated apex of the horse-shoe has already passed into the outer cortex. All departing leaf traces have a pronounced C-shape. Within the inner cortex of the slide one leaf trace shows an initial stage of separation. Others are completely detached. The outer limit of the five-sided light coloured inner cortex is easily traced (Pl. 2, fig. 3). Its sides, slightly concave, alternate with sharp angles. They stand out clearly from the darker coloured outer cortex surrounding it and forming the outermost part of the true stem. The outer cortex is in its turn readily distinguished from the leaf sheath by its darker colour and by the delimiting sclerenchyma of the leaf bases seen wherever an included leaf trace projects in any degree at all beyond the cortical tissue (Pl. 1, fig. 2). The structure of the outer cortex appears to be homogeneous with well preserved parenchymatous cells except as stated above where the bounding sclerenchyma of projecting leaf traces is developed externally. There are two protoxylem groups in all traces within both regions of the cortex. They are visible at the inner angle of each arm of the C-shaped xylem strands (PI. 2, fig. 3; Pl. 4, figs. 8,9; PI. 5, fig. 10). In thin sections they are not very easy to detect at first because of the blurring resulting from the oblique sections of leaf trace in which the cells themselves are also frequently tilted slightly by the grinding processes. This is more or less inevitable for leaves arising at an angle from the central xylem cylinder are bound to be sectioned somewhat obliquely in a transverse cut across the rhizome axis. In slide V.29630a a trace in the outer cortex on the opposite side of the xylem cylinder to the large horse-shoe strand clearly shows one of its protoxylems as a group of about six small cells at the inner angle of one limb (Pl. 2, fig. 4) while the second is seen in a blurred section in a corresponding position on the other limb. By tilting the slide slightly under the microscope a true cross section of a trace in the inner cortex which has just separated from the stele can be observed. It lies adjacent to roots (Pl. 3, fig. 5). By using strong reflected light on the polished solid surface of V.29630 from which the section was cut two proto- xylems are more readily apparent under the microscope on traces within the true stem for here the details are not blurred by tilting of the cells (Pl. 4, figs. 7-9). In both xylem strands and leaf traces the position of the protoxylem may further be indicated by a slight elongation and convergence of the adjoining metaxylem strands in transverse section (PI. 2, fig. 4 ; Pl. 3, fig. 5). 146 THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS The number of leaf traces in the cortical region is regarded by Hewitson as of some importance. He counts only those traces which are actually free from the stele but whose outer limits, as shown by absence of sclerenchyma ring, do not project beyond the outer cortex (see p. 145). The slide shows four leaf traces in the above restricted sense of which two lie wholly in the outer cortex. For living Osmundaceae Hewitson (1962 : 73) gives the following figures : Sub-genus Plenasium 3 to 8 traces (o to I inner cortex ; 3 to 8 outer) Sub-genus Osmunda 8 to 14 traces (2 to 4 inner ; 5 to II outer) O. lancea is exceptional with 12 to 22 (1 to 5 inner; 11 to15 outer) Sub-genus Osmundastrum 11 to 27 traces (4 to 12 inner ; 7 to 15 outer) Genus Todea 6 to 12 traces (2 to 5 innner ; 3 to 8 outer) Genus Leptopteris 4 to 15 traces in small rhizomes in two of the three living species (0 to 3 inner ; 3 to 13 outer) 9g to 27 traces (ina large rhizome of L. superba) (0 to 8 inner ; 4 to 20 outer). As can be seen the numbers vary in the different genera and sub-genera. The closest to the fossil is Osmunda (Plenasium) banksiaefolia with four traces in the entire cortex but differing in that there are none in the inner cortex and four in the outer. The Leaf Mantle. Outside the true stem in the leaf mantle, the newly departed leaf bases still show two protoxylem groups with C-shaped or reniform xylem bands (Pl. 1, fig. 2 ; Pl. 2, fig. 4, bottom right). Passing towards the circumference of the rhizome, i.e. in a position equivalent to a higher level on the emerging stipes, the traces develop a broader larger opening on the adaxial side, gradually becoming broader and flatter themselves (Pl. 1, fig. 1). As a result of this development the cutermost, oldest, petioles preserved in the mantle have a broad xylem band with incurved ends and wide opening. In these older outer stipes the protoxylem has divided into a number of separate strands which are seen in section lying along the inner concave outline of the xylem band. (PI. 5, fig. 12; Pl. 6, fig. 14, where the strands are just visible in the photographs as deeply stained patches. They are very clearly seen in the slides themselves). As in all Osmundaceae, petioles which have emerged from the stem are surrounded entirely by a stout thick ring or ellipse of sclerenchyma often about 0-14 to 0:34 mm. thick (Pl. 5, fig. 12). In the emerging traces still partly embedded in the outer cortex this sclerenchyma belt is obvious only on the protruding outer surface of the petiole as described (PI. 1, fig. 2). No wing stipules have been preserved on the first two whorls of petioles outside the stem (PI. 6, fig. 14 below, left) but in subsequent whorls of the loosely arranged and somewhat flexible young stipes the’ stipules are much twisted and curved. Some stipes are tangentially compressed and radially elongate with much distorted xylem as seen in section (PI. 6, fig. 14 right), others are narrow and elongate tangentially (PI. 6, fig. 14 left, above and centre). On first emerging typical stipe dimensions are as follows, the tangential measure- ment being given first in every case : 2°55 by I'I4 mm. ; 3:34 by I'I4 mm. ; 3:07 THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS 147 by o1r mm. Radially elongate stipes are 2-39 by 2:28 mm. ; 3:07 by 2:28 mm. At the extreme circumference typical measurements are: 8 by 7 mm. ; 3°5 by 42mm. ; 7 by 28mm. ; 2:55 by 7mm. One of the largest outermost stipes lies parallel with the greatest diameter of the elliptical section and has a total breadth of about 26 mm. of which the wings measure 8 and 9 mm. respectively. The maximum radial diameter of this leaf base is 4-5 mm. The wing stipules are formed of coarse celled light brown parenchyma with scattered, distinctly separated, patches of sclerenchyma as seen in transverse section (representing the cut ends of long scleren- chyma strands). The patches lie mainly at one level in the wing but are sometimes seen at different levels where the wing is thickest (Pl. 5, fig. 12 ; Pl. 6, fig. 14). There may be from six to nine patches in each wing but apparently the number is variable. Towards the thin lateral extremities of the wings the sclerenchyma patches diminish progressively in size. Sclerenchyma has also developed within the stipe itself starting in the young leaf base at the apex of the bay on the adaxial side of the stele. Initially only a few sclerenchyma cells are seen in this position. How- ever passing upwards, as the leaf develops, this small patch increases in size and later divides into three as can be observed in stipes a little further out in the mantle. Later still when the trace broadens the sclerenchyma forms several partly united patches lying near the inner curve of the xylem. Towards the circumference of the section it has again broken up to form distinct separated patches as is clearly shown in PI. 5, fig. 12, lowest stipe, and Pl. 6, fig. 14. More sclerenchyma occurs within the sclerotic ring of the petiole base on both the adaxial and abaxial sides of the xylem as well as laterally. Many small scattered quite separate patches are visible (Pl. 5, fig. 12 ; Pl. 6, fig. 14). SUMMARY OF CHARACTERS WHICH INDICATE RELATIONSHIP WITH OSMUNDA, SUB-GENUS PLENASIUM IN V.29630 The features described above distinguish the specimen unmistakeably from Todea and Leptopteris in the light of Hewitson’s researches. Moreover certain characters ally it with Osmunda and within that genus with the sub-section Plenasium. These characters are : The distribution of the sclerenchyma in wing stipules and leaf base. As described above. The interrupted character of the adaxial sclerenchyma adjacent to the leaf trace is encountered in the Japanese and Chinese species Osmunda (Plenasium) banksiaefolia (Presl) Kuhn. The sub-genus Osmunda has, in contrast, a long continuous sclerenchyma band in each wing in two of the Recent species and in all three (Osmunda regalis, O. japonica and O. lancea) little or no scattered sclerenchyma within the continuous sclerenchyma ring, nor is the adaxial scleren- chyma associated with the leaf trace arranged as in the fossil (cf. Hewitson 1962, fig. 7 A,B,C). In the sub-genus Osmundastrum the continuous sclerenchyma ring is very distinct from that of the fossil in that it is formed of two kinds of sclerenchyma. Todea, although it has scattered sclerenchyma within the wings and continuous 148 THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS ring of the stipe, lacks any sclerenchyma on the adaxial side of the leaf trace in Hewitson’s material while Kidston & Gywnne-Vaughan (1907, pl. 6, fig. 7) show, diagrammatically, only a very weak and limited development of such sclerenchyma in Todea barbara. Leptopteris has either a few well developed sclerotic patches in each wing or numerous very poorly developed strands, but within the continuous sclerenchyma ring of the stipe scattered strands are lacking in all species. The xylem characters. Especially the two protoxylem groups in the leaf trace prior to its separation from the stem and the query-shaped bundles with hook directed towards the outside in the xylem cylinder. Two protoxylem groups in such positions are normal in all Plenasium species in which also, as in the fossil, the xylem trace has a marked C-shape immediately after it departs from the stele. Hewitson records having seen one specimen of Osmunda regalis from India with two protoxylem groups but this is a rare condition in the sub-genus Osmunda. In view of the other Plenasium characters which the fossil shows it seems reasonable to regard its two protoxylems as indications of this sub-genus rather than as an aberrant type of the sub-genus Osmunda which in the other respects it does not resemble. The number of xylem bundles (twenty) in the cylinder of the stem is rather high for Plenasitum which has about three to twelve, the sub-genus Osmunda varies from four to eleven, Osmundas- tvum seven to twenty-two, while in the genera Todea (two to seven) and Leptopterts (three to twelve) low bundle numbers are found with marked confluence of the bundles. From species to species in each sub-genus and genus there is some variation. Thus Osmunda (Plenasium) javanicum has three to eleven, Osmunda (P.) vachellit nine, Osmunda (P.) bromeliaefolia three to twelve and Osmunda (P.) banksiaefolia four tonine. In this respect therefore the fossil is specifically distinct with about twenty to twenty-one entirely separate bundles in the only two specimens seen. Character of the cortex (Pl. 1, fig. 2). There is clear separation between the inner and outer cortex and between the outer cortex and leaf mantle due in the latter case to the sharp definition of the sclerenchyma bands on the external surface only of the emerging leaf traces as already described. In Todea and Leptopteris the outer cortex is of two cell types, the ring around the stipes being thick walled, the remaining tissues of thinner walled cells with larger lumen. In these two genera, therefore, the sclerenchyma ring is clear all round the trace while still within the cortex whereas throughout the whole genus Osmunda as shown above it is only apparent where the stipes protrude. In the fossil there are four traces in the whole cortex (those, that is, whose outer limits as shown by sclerenchyma do not project from the cortex) two of which lie in the inner cortex. The number is determined by the narrowness of this tissue (external diameter of outer cortex 14 by 9:5 mm. and of the inner cortex 5 by 4-5 mm.) combined with the sharp angle of departure of the traces (23°). At its greatest width the outer cortex may be about 3-75 mm. and the inner 0-75 mm. with a least THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS 149 width of about 0-25 mm. The sub-genus Osmunda, and the genera Todea and (usually) Leptopteris are similar to one another in having up to fifteen traces in the whole cortex while Osmundastrum differs even more from V.29630 in having eleven to twenty-seven traces, from four to twelve of these being in the inner cortex which is relatively wide. Plenasium, on the other hand, resembles the fossil in the small number of traces (three to eight) with normally nought to one in the inner cortex. However Hewitson stresses that in this respect the point on the rhizome at which the section is taken is important, the number of traces increasing with “ an increas- ing fraction of phyllotaxy and an increasing stem size ’’, but in Plenastum the low number of traces is real, seven traces being the largest number he had seen in a very large rhizome of Osmunda javanicum of which one only was in the inner cortex. RE-EXAMINATION OF THE HOLOTYPE V.29629, V.29629a and b and slides V.29629c—k and Kidston Collection K.1248 General Considerations. Having now described and discussed the better preserved specimen it remains to add a few new facts about the holotype and to indicate the reasons for regarding both rhizomes as belonging to a single genus and species. Carruthers in his original description gave natural size drawings but no dimensions in figures. The specimen when found by Dowker must obviously have been longer than the 110 mm. which now survive, for so many sections have been cut from the central region. These inevitably must have meant the grinding away of an appre- ciable length. The rhizome is now represented by an upper (V.29629a & b) anda lower (V.29629) portion. The upper part is 47 mm. long, the lower 63 mm. Carruthers’ figures (1870, pl. 1, figs. I, 2) show the two portions to have been 67 and 73 mm. respectively at that time. This means a loss of some 30 mm. of length from the two pieces. It is not clear whether two complete sections made by Carruthers (V.29629c and V.29629h) were cut before his drawings of the hand specimen were made but probably they were, for the thirty missing millimetres would scarcely cover the preparation of these two thick slides, of slides V.29629d-g, and of the Kidston slide also (K.1248, figured Kidston & Gywnne-Vaughan 1907, pl. 4, fig. 21). K.1248 was prepared commercially by F. Krantz in Bonn in or shortly before 1907. The production of three serial peel sections by Walton in 1930 must also have entailed a further slight diminution of the length of the upper fragment, V.29629a & b. V.29629d was formerly V.2432 and was then entered in the Register (in 1889) as presented by “‘the late Dr. Millar, March, 1888” (one specimen). V.29629e (formerly V.40193) is a recent purchase, in 1958, which formed part of the Dufty Collection. The two slides V.29620f and g are all that remain of four registered in 1902 as “ V.7103 (one specimen) purchased executors of late George Dowker, 1899 ”’ and “‘ V.7104 (three specimens) Sections of Osmundites Dowkeri’”’. The register indicated that the four slides were all part of V.6126 the former registration number of the holotype. There is nothing to indicate whether Dowker’s slide, V.7103, was one of the two survivors or whether both of these survivors belonged formerly to V.7104. The two slides were apparently already missing in 1952 when the whole of the Osmundites dowkeri material then extant in the Museum was re-registered 150 THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS TRAcE TRACE / N eet Na —~— — ETCHED SURFACE hhh hhh hhh dh dh dbdbdbdbbddbddidhdb dbs dddbidddbbbbhbdldidds 3 Peer. SECTIONS ea l¥S5O VUSTESIESLLIL IIOP IOLILITIMMOOSOLEISOLTLITLELEETIOOISSOLSS, Ph hhh Lh hehehehe Shh heh hahadaahan cadadeceabcahedhathabentidcthatAachakdkabed V. Pa. 96 29 rI-K. — — CARRUTHERS Cal87O V:29629H. UTZ LALLA LL Ad hdd chcieceedebchd bd N ORMAN V-2 9. 62 9 ——@ D777 7 a hdhehed bedded dhhdhhhdchubahchchdhdidddddbddbehhdddbdddth R ELATIVE PosiTio N NOT DETE RNIN ED PesiITioN OF ORIGINAL CUT RELATIVE TO SLIDES. [PIT TTTTOOOOC OCI OC DLL LIL LOL LLL LLL LL LLL WITTDI OLLIE CIOL LILLIA LL LLL LILO TOIT ITITIIIO OT OL LOL LILA O LL IDA LALLA AD ALLL LL LLL LLL IIIT ITIITT TOP COCCI LIL IAAL OIA LLL NORMAN V.29629 D-E. RELATIVE POSITION NOT DETERMINED —— CARRUTHERS Cal870 V.29629¢ — Kiwstron cal7o7 K:1243 —— —PourSHED SURFACE Fic. 1. Diagram showing relative position in rhizome of slides. For details see text p. 150. Vertical distances not to scale. Sections cross hatched. V.29629b is the counterpart half of the sectioned upper fragment of the rhizome so cannot be shown in the figure. THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS 151 as V.29629 (seven specimens). One specimen was the hand specimen (then in two fragments) and six were slides which can be accounted for as follows : Two Carruthers slides V.29629c¢ and h (formerly part of V.6126) ; one peel section now V.29629; ; V.296209d (see above) and V.29629f and g (which alone represent the former four specimens V.7103-04). The slides, V.29629d-g appear to predate Carruthers’ work. They have one feature in common in that all were made by a dealer whose printed label incorporat- ing the words “ Norman. Preparator’”’ is on each. Probably the slide from Dowker’s executor was one of Norman’s. Some of these have been reassembled from other sources (see above). Perhaps the dealer disposed of them in the course of business ; perhaps also of others not traced. This, although irritating, is relatively unimportant in that we now have all the information we need. Of Norman’s available slides V.29620d is far the best as it includes a good tangential arc of the xylem ring. V.29629e is much disorganized. V.29629f and g are incomplete portions of the transverse section, f showing a fragment of xylem ring and mantle on the abraded part of the rhizome and g a piece of the leaf mantle from the broad, less abraded side of the stem towards the exterior of the rhizome. Study of the slides themselves, taken together with a consideration of the fore- going statements, while it cannot give precise distances between the slides shows their former relative positions in the rhizome. These are drawn diagrammatically in Text-fig. 1. Although no record of the sectioning appears to have been kept the sequence of events seems to have been that Dowker, or probably Norman, cut his rhizome in half and slides were prepared professionally by Norman from the two cut ends. The two halves of the hand specimen were then transferred to Carruthers who made two complete transverse sections from the cut ends about 1870. Kidston’s section and the peels were the last to be prepared, two of the latter having been acquired in 1963. As a result of the way they were made slides from the upper fragment have the coverslips on the upper side of the slide. Those from the lower fragment must be reversed with the coverslip lying on the underside of the slide in order to place them in correct sequence. The transverse diameter of the hand specimen V.29629 is 45 by 63 mm. Its leaf mantle is closely compacted. The surface shows the abraded ends of the petiole bases arranged in a steep spiral of about 35° with the axis. As in V.29630 these petioles are always worn away below the upper end of the wing stipules. Consequently neither their transverse sections nor such surfaces as are preserved can show whether the stipules were fused at the apex across the face of the petiole producing a curved commissure. Alternatively they could have persisted as two separate wings one each side of the stipe throughout their length (Text-fig. 2). Kidston & Gywnne- Vaughan (1907 : 766) believed the presence or absence of this commissure to be the only constant superficial distinction between Todea stocks on the one hand (including species now assigned to Leptopteris) and Osmunda on the other, this commissure being found only in Todea and Leptopteris. Fortunately however it has been shown by Hewitson that there are other cogent anatomical grounds on which the two groups can be distinguished. 152 THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS Of the hand specimen the lower and upper portion still remain. The upper with its rough unpolished surface has also been sectioned longitudinally into two fragments (V.29629a and b) but the section runs slightly obliquely passing only in the lower half through the true stem. The grinding processes apparently removed about 6 mm. of the breadth and the pith is only exposed at the lower end of the longitudinal section. The leaf traces are seen to arise from the stem at about 23°, a figure which agrees with that deduced for V.29630. A few details of pith, sclerenchyma, xylem and phloem can be discerned in spite of the difficulty of examining microscopically such dark material by reflected light. Because the material was so limited, no thin longitudinal section was cut. The lower piece (V.29629) of the rhizome is deeply excavated below, only the outer leaf bases remaining at its circumference while the inner leaf bases have been worn in such a manner as to produce the deep conical basal cavity. The upper transversely cut and polished surface of V.29629 lies some 55 mm. above the lowest part of the mantle which forms its circumference. The transverse section shows the true stem to have a diameter of about 13 mm. It lay excentrically in the leaf mantle as the result of abrasion prior to fossilization. Consequently on one side of the specimen all but about two or three layers of the mantle are missing, whereas eight or nine layers are still present along the opposite radius of the rhizome (cf. Kidston & Gywnne-Vaughan, 1907 pl. 4, fig. 21). The least distance between the true stem and the present circumference of the specimen of the more abraded side is only about 5 mm., but along the opposite radius about 34 mm. The original diameter when the rhizome was perfect would have been about 80 to 85 mm. along these radii if abrasion on one side had not been so great. The diameter may well indeed have exceeded the figures suggested for it is probable that some leaf bases may have disappeared from the least abraded side of the rhizome also. Of the transverse sections which still exist, only six show the whole rhizome. Three of these are valuable peel sections (V.296297, 7, k) which were made very close together at a high level in the rhizome, where the xylem cylinder was less damaged than elsewhere and its tissues were less obscured by fungal hyphae. A comparison of V.296297 and k shows admirably the changes which have occurred as xylem ring and leaf traces passed upwards (cf. Pl. 8, fig. 16 showing a lower section, V.296291, and Pl. 9, fig. 17). It is regrettable that the section (K.1248) figured by Kidston & Gywnne-Vaughan (1907 : 768, pl. 4, fig. 21) was made at a level where fungal infestation was great and considerable distortion of xylem had occurred, especially on one side, the result of decay combined with radial compres- sion along the shortest diameter. This probably explains why the two authors did not describe the anatomy in greater detail (apart from that of the roots). It also explains the large number of xylem strands which they reported since the partial union of some adjacent strands, now counted as one, is obscured by the radial crushing in this particular section. THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS 153 Anatomical Structure. The Pith is about 3-5 mm. in diameter, formed of typical parenchymatous cells frequently about 0-057 mm. in cross section. Around the outer margin there are about six or seven layers, sometimes only three, of somewhat larger cells which appear denser and darker in colour but are shown by the longitudinal section to be normal thin walled equiaxial parenchymatous tissue. In the limited area of pith visible in the longitudinal section no isolated tracheids have been detected but it must be remembered that the section does not pass through the central region of the pith. The pith is continuous with the “‘rays”’ of tissue between the xylem strands and no indication of an inner endodermis has been seen. The Xylem Cylinder and Leaf Traces. The xylem cylinder can be examined in section on the cut surfaces of the rhizome and in seven slides including the peel sections. It is about 4:5 to 5 mm. in diameter and is formed of about twenty or twenty-one entirely separated strands, using again Hewitson’s method of counting (cf. p.144 ). Owing, however, to the radially crushed state of part of the cylinder as explained above the number cannot be seen in most of the available sections for on the side where the crushing occurred it is usually impossible to say whether two adjacent strands are or are not connected as Hewitson specifies “‘ even by one tracheid’’. The clearest sections for counting are the peel sections (V.296297 and k). The xylem strands are separated by some five or six layers of radially elongate parenchyma cells which pass outwards into a parenchyma layer seen in places surrounding the xylem ring. The strands vary much in shape. Two large horse- shoes opening inwards project beyond the outer circumference of the cylinder (Pl. 9, fig. 17). A similar large horse-shoe is seen on the polished lower surface of the upper fragment (V.29629a) of the rhizome (PI. ro, fig. 18). It also shows a leaf trace, just separated from the two arms of a horse-shoe lower down in the rhizome (left in Pl. 9, fig. 17), whose rounded distal end forms the C-shaped trace on this surface. Occasionally traces are elongate at one extremity owing to the initial development of a root (Pl. 9, fig. 17 at 2 o’clock). The origin of a pair of roots is beautifully displayed in peels V.296297 (PI. 8, fig. 16) and7. Some adjacent strands are united at their inner ends giving rise to a U or V opening outwards. Such a U is seen to the right of a large horse-shoe (Pl. 9, fig. 17). Two pairs of strands form two adjacent question marks, one reversed, which are well displayed in the peel section V.29629 (Pl. 9, fig. 17 top centre, top right). It also shows simple ovals and three slightly united strands forming an S at 6 o’clock. Between the arms of united strands there is parenchyma. Most of the obvious tracheids are large metaxylem elements, smaller ones sometimes occurring at their outer ends. On the walls of the tracheids several lines of narrow pitted or scalariform thickening can be seen. Some difficulty is encountered in the study of the thin sections because the angle at which the traces spring from the stem again causes transverse sections to cut the leaf traces slightly obliquely. Further the cells themselves may lie slightly obliquely on those slides which are more than one cell thick and in the grinding process some 154 THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS disorganization of tissues has occurred. For these reasons many of the cell walls have a blurred outline. However, as in the case of V.29630a, a slight tilting of the slide in an appropriate direction sometimes clarifies the cell walls. Once again in the holotype a study by reflected light of the opaque polished surfaces of the rhizome itself assists in the understanding of the sections. More especially it is a help in locating the protoxylems for on the solid surfaces distortion and disorganization are at a minimum. Any attempt to reduce further the thickness of the sections might readily lead to worse disorganization. Indeed the thinnest of all, Kidston Collection slide K.1248 is much disrupted. There is good evidence visible on the polished lower surface of the upper fragment (V.29629a@) of the rhizome close to the remaining small arc of xylem at the longitudinally cut edge (Pl. 10, figs. 18-20). Besides displaying a large horse-shoe trace, it shows V traces opening outwards (PI. 10, fig. 18) and the initial stages of separation of the rounded end of a horse-shoe to form a trace. In addition there are well preserved completely separated traces in the inner cortex while four are still wholly immersed (in Hewitson’s sense) in the outer cortex. Several of these traces offer unmistakeable evidence of small proto- xylem strands at the inner angles of the arms of the C-shaped xylem (PI. 10, figs. Ig, 20) although owing to reflections from the polished surface they are difficult to show by photography. The evidence is best seen if the examination of these solid surfaces is made not by artificial light but in bright daylight without direct sunlight. A low power objective in the microscope shows it clearly. Of the thin slides the evidence most easily seen is in the peels V.296297 and k. The successive stages in the development of the leaf trace can be better observed in the holotype than in V.29630 owing to its wider diameter. As the trace passes outwards and upwards the xylem sheath becomes more deeply C-shaped or reniform with only a narrow gap on the adaxial side occupied by about five or six radiaily elongate cells. This form persists into the outer cortex of the stem. The xylem, by this time horse-shoe shaped with thickened arms, has a metaxylem which may be four or five cells thick within the arms but only about two cells or even a single cell thick at the apex of the horse-shoe. Throughout both regions of the cortex the leaf trace is surrounded by a clear dark line corresponding to the phloem and its associated cells. By the time the trace passes into the outer cortex it may be appreciably larger than it was in the inner. Beyond and surrounding the phloem in the outer cortex there is an oval or ovate belt of thin-walled parenchyma clearly delimited from the normal denser tissue of this region (PI. 7, fig. 15). The maximum diameter of the inner cortex is about 9 mm., its greatest width from the xylem cylinder to its circumference being about 2 mm. Its outline has seven or eight points separated by slightly concave sides (Pl. 7, fig. 15; Pl. 9, fig. 17). The maximum diameter of the outer cortex is about 15 mm. the greatest width between its inner and outer limits being some 3 to 3:5 mm. _ It also has a seven or eight-rayed outline. Unlike V.29630 the cell structure happens to be well preserved both in the inner and outer cortex (Pl. 9, fig. 17). In this mature rhizome the relatively wide cortex of necessity means that a larger number of leaf traces are sectioned within the true stem than in V.29630. In the entire cortex about ten to twelve completely THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS 155 immersed traces are visible. In the inner cortex there are six in V.29629¢, five in V.29629h-7, seven in V.29629k and five on the polished surface of the lower end of the rhizome (PI. 7, fig. 15). In the outer cortex there are five in V.29629¢, 7, k and on the polished surface of V.29629 ; six in V.29629h, 7. That there is a somewhat greater number than in living species of Plenasium is no doubt correlated with the greater number of xylem strands in the stem cylinder. Leaf Mantle. Beyond the true stem for the first four or five whorls the emerged leaves have a more deeply reniform transverse section and then begin to develop a broader larger opening on the adaxial side (Pl. 11, fig. 21). The development of sclerenchyma in the bay of the xylem follows the same course as that described on p- 147 for V.29630. The arrangement of scattered sclerenchyma strands within the continuous ring of the petiole is also similar (Pl. 12, fig. 22 ; Kidston & Gywnne- Vaughan, 1907, pl. 6, fig. 5). The sclerenchyma shows less clearly in the photographs of the thick older slides of the holotype than in the actual slides themselves. It is much clearer in the thinner Kidston slide (cf. wing stipules Pl. 5, fig. 13 ; Pl. 11, fig. 21). The identical character of the leaf bases and wing stipules affords clear evidence that the two rhizome fragments belong to a single species. In contrast to the young stock in V.29630 the leaf bases are tightly and geometrically packed. Kidston & Gywnne-Vaughan (1907 : 769) believed that in close proximity to the true stem the stipules were all concrescent. Since the stipule outlines are perfectly distinct it seems more probable from experience with other fossils that the appearance of concrescence was due to infiltration of silica which later formed a cement. In some newly emerged stipes one wing stipule only may have developed but normally a pair is seen, the stipules here being very short in transverse section (Pl. 12, fig. 23). At this stage the diameter from tip to tip may be only 3:5 to 4:5 mm. and the stipe itself may measure 2:5 mm., the dorsiventral thickness being 2 to 2:75 mm. The stipules broaden laterally upwards as shown in successively older stipes toward the circumference of the mantle but, as stated, in no case is a sufficient length of petiole preserved to show the distal termination of the stipule. This end always appears to have been removed by abrasion. The breadth of the wing stipules is greatest and their dorsiventral thickness least where they lie parallel with the greatest diameter of the rhizome near the circumference (PI. 11, fig. 21). The largest examples measure 18 mm. in breadth, the actual stipe itself and each wing being about 6 mm., the dorsiventral thickness about 2:75 mm. Midway between the extremes of measurement at the base on the one hand and towards the more distal end of the stipules as preserved on the other there are measurements of about 7 mm. from tip to tip with stipe breadth of 3 mm. and dorsiventral thickness of 2:25 mm. In the highest part of the stipules seen the tips are extremely narrow and may be somewhat curved ; they are presumably less rigid here on account of their thinness. This suggests that little of the distal end is missing so that their total length may not greatly have exceeded 55 mm. A stipule of Todea shown in Text-fig. 2 was 45 mm. long which suggests that Osmunda dowkeri must have had large stout fronds. Other Tissues. A sheath of parenchyma surrounds the xylem strands and is thickest in the outer part of the leaf gap, thinnest opposite the strands themselves. 156 THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS Fic. 2. Two Recent petiole bases. a. Osmunda zeylanica. 8. Todea barbara. In both the stipe (s¢) flanked by the wing stipules (w) but in Todea these stipules are fused across the adaxial side of the stipe producing a curved commissure line c. The thick diver- gent lines and in Todea vertical ones below the commissure, indicate sclerenchyma within. Clearly a transverse section at the level 7 in the two cases would appear quite different for in Todea the stipules would lie in front of and quite separate from the stipe, whereas in Osmunda they would flank it on both sides. This would be apparent in the section of a fossil stipe at the appropriate level. Between the xylem strands it constitutes the “ medullary rays ”’ of some writers on Osmundaceae. It is encircled externally by phloem which is followed by tangen- tially elongate cells. The large horse-shoes of the incipient leaf traces cause a bulge in the surrounding parenchyma and phloem. When the leaf traces first appear they are surrounded by a branch of phloem and tangentially elongate cells (much obscured by fungi). V.29620d (PI. 12, fig. 24) shows very clearly the emergence of a root from one angle of an incipient leaf trace not yet separated from the xylem ring. Tracheids of the trace can be seen passing directly into the root. Further evidence as to root develop- ment is shown in Pl. 8, fig. 16 ; Pl. 9, fig. 17. THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS 157 SPECIFIC IDENTITY OF THE TWO SPECIMENS DESCRIBED It should be noted that the holotype not only agrees with V29630 in the character of the leaf bases of the mantle but also in the two protoxylem strands of traces within the cortex, in the form and number of xylem strands in the stem cylinder and in the deeply curved form of the xylem in the newly separated leaf traces. Hence all lines of evidence point to the specific identity of the two specimens and to their affinity with Osmunda rather than with Todea or Leptopteris. Within the genus Osmunda the relationship is with the East Asian sub-genus Plenasiwm. SIMILARITY TO OSMUNDITES CHANDLERI ARNOLD Only one fossil species resembles the Thanetian Osmunda dowkeri at all closely. It is Osmundites chandlert Arnold from the Eocene Clarno Beds of Oregon, U.S.A. (Arnold 1952: 68, pls. 7, 8). In this case the resemblance is so close that in spite of the geographical distance between Oregon and Southern England it cannot be disregarded. Features they possess in common are the presence of two protoxylem strands in the young leaf trace of the inner and outer cortex (well seen by reflected light on the smooth surface of a rhizome kindly supplied by Professor Arnold and perhaps in Arnold’s specimen 1952, pl. 7, fig. 12), the C-shaped xylem of the newly emerged leaf, above all the character of the leaf bases both as regards the form of the xylem band and the distribution of sclerenchyma. In Osmundttes chandler scleren- chyma within the continuous ring which surrounds the stipe is scattered laterally, adaxially and abaxially. It also occurs on the adaxial side close to the xylem arc and that in the outermost leaves preserved appears to break up into distinct strands as in Osmunda dowkeri although a short distance within it is only partially separated into about twelve masses. In the outside whorl of the specimen complete separation had occurred in one or two places and it is possible that larger rhizomes with suffi- ciently mature stipes would show the same degree of separation that Osmunda dowkeri displays (cf. also Arnold 1952, pl. 8, figs. 17, 19). Within the stipule wings of Osmundites chandler: the distribution of distinct sclerenchyma patches is identical in the two ; most of the patches are arranged in one line, but they sometimes occur at more than one level. Arnold reports thirty-four oval or horse-shoe shaped xylem strands in the stem cylinder (cf. Kidston & Gywnne-Vaughan thirty). If Hewitson’s method of counting were adopted a reduction in this number could be expected but it is not possible from the published figure to make an accurate count as the focus of the print does not show whether any of the strands are partially united. Arnold’s pl. 8, fig. 5 shows a half cylinder in which there appear to be about eleven distinct strands. The rhizome received does not clarify this point as the cylinder has broken down on one side. No clear and unmistakeable query-shaped strands could be seen in this specimen although a tendency for the development of this form may be indicated where two bundles turn to one another. Some approximation to this form is seen in Arnold’s pl. 7, fig. 13 below the left-hand limb of the just separated trace. See also his pl. 8, fig. 15. Large projecting horse-shoes are a conspicuous feature of the xylem ring. Arnold himself did not press the possible relationship of Osmunda dowkert and Osmundites chandleri although he commented on the resemblance 158 THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS between them. He considered that the geological and geographical separation of the sites from which they came was too great to allow of relationship. It must be borne in mind that distance in these senses does not always exclude specific identity for as Scott (1954) has already shown and is to show still further (unpublished work), identical extinct genera and even identical species do occur among the fruits and seeds of the Eocene Clarno Beds of Oregon and the London Clay of England. As regards difference of age it is now clear that the Lower Tertiary flora persisted at least from the beginning of the Tertiary period into the Oligocene. Should the re- examination of material of Osmundites chandlert confirm the suggested relationship to Osmunda dowkert then the former should be referred to Carruthers’ species. In any case it seems reasonably certain that Osmundites chandleri should be transferred to the living Osmunda and to the sub-genus Plenasium within it. A POSSIBLE CONNEXION BETWEEN OSMUNDA (PLENASIUM) DOWKERI AND OSMUNDA (PLENASIUM) LIGNITUM It is natural at this point to enquire what evidence there is as to the relationship of the foliage described as Osmunda lignitum (Giebel) with living sub-genera of Osmunda. ‘The species is represented by beautiful impressions in the Bournemouth Marine Beds (Gardner & Ettinghausen 1880 : 49, pl. 4, figs. 1-3 ; 1882 : 66) and by much broken coriaceous remains in the Bovey Tracey Lignite of Devon (Heer 1862: 1068, pl. 55, figs. 4-6; pl. 56, figs. 1-11 ; pl. 57, figs. 1-7) as well as in numerous Oligocene horizons on the Continent. In a letter dated 3.8.60 Dr. R. E. Holttum drew attention to the strong resemblance between Osmunda ligmitum and the Japanese and Chinese species Osmunda banksiae- folia. This species, he added, had been included in Osmunda javanicum in Synopsis Filicium (Hk. & Bak.) although probably distinct. Various species included at one time in O. javanicum are closely related forms belonging to the sub-genus Plenasium. Gardner & Ettingshausen (1880 : 53) had already noticed the close resemblance between this fossil foliage and “‘ Osmunda javanicum’”’ which ranged, they noted, from Kamschatka to Java and Ceylon. At a later date in a Revision of Eocene Ferns for which Gardner alone was apparently responsible (Gardner & Ettingshausen 1882 : 66), the variations of the “ species’? Osmunda javanicum in the different latitudinal areas of its range are described. The statement there occurs that, “‘ It is in the more average-sized pinnae from Formosa, latitude 24° that we meet with the most absolute identity, as far as the fragments admit of comparison, with our fossil forms ”’. Additional support for the view that O. lignitwm belongs to the sub-genus Plena- sium is provided in Hewitson’s (1962: 61, text-figs. I-4) account and figures of foliage in the Osmundaceae. While there is general agreement that the species of Plenasium are in need of reappraisal, it is certain that Osmunda lignitum has its closest affinities within this sub-genus. Thus the species Osmunda javamicum, O. vachellit, O. bromeliaefolia and O. banksiaefolia, discussed by Hewitson, are all characterized, as is O. lignitum, by once-pinnate fronds. The sub-genus Osmunda THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS 159 is excluded from close relationship on account of its bipinnate foliage. (Within it Hewitson includes Osmunda lancea because it, too, is bipinnate.) His description of the nervation of this sub-genus shows that it is unlike that of the fossil Osmunda ligntum. The American sub-genus Osmundastrum (Osmundastrum cinnamomea and O. claytoniana) although it has once-pinnate fronds is distinguished from Osmunda lignitum by the deeply dissected pinnae. In Plenasium, whatever the ultimate renaming of its species, there is some variation of the margin in the pinnae. It is entire in Osmunda vachellit ; entire or toothed in O. javanicum ; toothed with narrow pinnae in O. bromeliaefolia ; coarsely toothed with wider pinnae as in the fossil in O. banksiaefolia where the resemblance is very close indeed. In Osmunda lignitum the lateral nerves sometimes give off a greater number of forked tertiary nerves, five or six being shown by Heer on the lower side of the leteral (secondary) nerve (1862, pl. 57, figs. 1, 4), while in the text he mentions as many as seven or eight. In his other figures, however, (cf. Heer 1862, pl. 57, fig. 5 for example) there is complete agreement with Hewitson’s text-fig. 41 of Osmunda banksiaefolia. In the upper part of the pinnule nearer the tip, O. lignitum shows fewer nerves which close to the tip may be undivided. Again in O. lignitum, the lowest tertiary nerves are markedly curved and enter the sinus between adjacent teeth where sometimes they unite (Heer 1862, pl. 57, fig. 2). Although most of the tertiary nerves actually spring from a secondary, occasionally a forking nerve arises from a primary one where it passes directly to the sinus (Heer 1862, pl. 6, figs. 1-5 ; cf. Hewitson 1962 : 65, text-fig. 41). The coriaceous character of the pinnules in itself and quite apart from a different nervation, serves to distinguish O. lignitum from any species of the filmy ferns Leptopteris. Todea, too, is quite unlike O. lignitum in that its fronds are bipinnate while the lateral nerves of the pinnules have a simple fork or may be unbranched. The existence of Plenasiwm in the Lower Tertiary of Western Europe on this entirely independent evidence provided by the foliage, demonstrates at least that there is no phytogeographical reason why the rhizomes should not be referred to that sub-genus of Osmunda. It further raises the question whether the rhizomes and the foliage belong to a single Lower Tertiary species, having regard to the wide distribution in space and time of many Tertiary plants. There is no direct evidence in support of such a connexion and probably such will never be forthcoming but the possibility must be borne in mind. Should the relationship ever be established, then the specific name dowkeri would have to give place to the earlier designation lignitum. SUMMARY OF CONCLUSIONS AS TO THE RELATIONSHIP OF OSMUNDA DOWKERI TO LIVING OSMUNDACEAE The Thanetian species, Osmunda dowkeri (Carruthers), now represented by two rhizomes, belongs to Osmunda, not to Todea or Leptopteris. This is shown : (1) By the homogeneity of the sclerenchyma in the outer cortex around the traces and their accompanying parenchyma which causes the sclerenchyma ring of the leaf trace to be apparent only on the abaxial side where the trace bulges beyond the limits of the cortex and true stem (see p. 147). 160 THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS (2) By the form and distribution of sclerenchyma in the wing stipules of the leaf base combined with the distribution of sclerenchyma in the continuous ring of the emerged stipe and the arrangement of sclerenchyma in the bay of its C-shaped xylem (cf. Hewitson 1962, Text-fig. 7A—M). Within the genus Osmunda in the broad sense, relationship of the fossil is with the section or sub-genus Plenasium. This is also demonstrated (1) By the sclerenchyma distribution (again cf. Hewitson 1962, text-fig. 7A-I1). (2) By the deeply curved C-shaped form of the leaf trace as soon as it separates from the xylem cylinder. (3) By the presence of a pair of protoxylem strands at the inner angles of the C-shaped traces in the inner and outer cortex. (4) By the presence of query-shaped strands in the xylem cylinder (p. 144). (5) By the low number of leaf traces within the cortex (p. 148). Osmunda (Plenasium) dowkert is distinguished from any living species of Plenasium by the greater number of xylem strands in the stem cylinder (twenty or twenty-one approximately) and by the narrow angle at which its stipes emerge (23°). The possible specific identity of Osmundites chandler Arnold which should also be referred to Osmunda, sub-genus Plenasium, cannot be lightly dismissed. The relationship of the Bournemouth and Bovey Tracey foliage of Osmunda lignitum (Giebel) to Plenasiwm is clear (p. 158). The possibility that it may be the foliage of the species Osmunda (Plenasium) dowkeri, known only from its rhizomes, must be borne in mind having regard to the long range in time and space of many members of the older Tertiary flora. The presence of Plenasitwm in the Lower Tertiary of Western Europe, based on independent foliar evidence supports the determination of the rhizomes as Plenasium. The finding of this East Asiatic fern genus accords with the phytogeographical indications provided by many Angiosperm families in older Tertiary deposits. THE GENERIC POSITION OF OSMUNDITES DOWKERI CARRUTHERS 161 ACKNOWLEDGEMENTS The late W. N. Edwards stimulated this research by his reluctance to use the name Osmunda for Osmundites dowkert without a preliminary thorough investigation as to the possibility of distinguishing the Recent genera of Osmundaceae from their rhizomes alone. The work of Dr. W. Hewitson of Harvard has provided this important stage in the research and he himself has supplied valuable help and com- ment. The extraordinarily beautiful thin section prepared by the late W. N. Croft from a newly discovered rhizome from Thanet has provided fresh information and cleared up points left in doubt by study of the original material. Great gratitude is due to Dr. R. E. Holttum for the interest he has taken in this work, for obtaining Recent material from Kew and for calling attention to Dr. Hewitson’s research as well as for various helpful suggestions. Dr. K. I. M. Chesters has as usual typed this manuscript and she and Mr. F. M. Wonnacott have kindly criticized while reading and editing it. The Photographic Department of the British Museum (Natural History) deserve a special word of thanks for the trouble they have taken in producing the excellent photographs which were not possible with my own apparatus. Finally the Regius Professor of Botany, University of Glasgow, has kindly lent slide K.1248 from the Kidston Collection. REFERENCES ARNOLD, C. A. 1952. Fossil Osmundaceae from the Eocene of Oregon. Palaeontographica, Stuttgart, 92, B: 63-78, pls. 6-8. CARRUTHERS, W. 1870. On the Structure of a Fern Stem from the Lower Eocene of Herne Bay, and on its Allies, Recent and Fossil. Quart. J. Geol. Soc. Lond., 26 : 349-354, pls. 24, 25. 1872. Notes on some Fossil Plants. Geol. Mag., Lond.,9 : 49-59, pl. 2. CHANDLER, M.E. J. 1957. The Oligocene Flora of the Bovey Tracey Lake Basin, Devonshire. Bull. Brit. Mus. (Nat. Hist.) Geol., London, 3 : 71-123, pls. 11-17. 1961. The Lower Tertiary Flovas of Southern England, 1. Palaeocene Flovas. London Clay Flova (Supplement). xi+354 pp., 34 pls. Brit. Mus. (Nat. Hist.), London. GARDNER, J. S. & ETTINGSHAUSEN, C. VON 1879-82. A Monograph of the British Eocene Flova,1: Filices. 86 pp.,13 pls. (Mon. Palaeont. Soc., London.) HEER, O. 1862. On the Fossil Flora of Bovey Tracey. Philos. Tvans., London, 152 : 1039- 1086, pls. 55-71. HeEwitson, W. 1962. Comparative Morphology of the Osmundaceae. Ann. Mo. bot. Gdn., St. Louis, 49 : 57-93, pl. I. Kipston, R. & GyWNNE-VAuUGHAN, D. T. 1907. On the Fossil Osmundaceae, I. Tvans. Roy. Soc. Edinb., 45 : 759-780, pls. 1-6. Scott, R. A. 1954. Fossil Fruits and Seeds from the Eocene Clarno Formation of Oregon. Palaeontographica, Stuttgart, 96, B : 66-97, pls. 15, 16. SEWARD, A. C. & Forp, S. O. 1903. The Anatomy of Todea with Notes on the Geological History and Affinities of the Osmundaceae. Tvans. Linn. Soc. Lond., 6: 237-260, pls. 27-30. PLATE 1 Fic. 1. Complete transverse section across rootstock showing central xylem cylinder and limits of outer cortex of stem (dark with angular outline). It also shows the variable form of distorted young leaf bases in surrounding mantle (contrast Kidston & Gywnne-Vaughan 1907, pl. 4, fig. 21). 3. (slide V.29630a.) Fic. 2 Central area of above showing dark outer cortex enclosing seven leaf traces, two only completely immersed. A thick sclerenchyma band is seen on outer margins of five traces which abut on edge of cortex. Inner cortex a narrow lighter region around xylem (represented by amorphous silica) enclosing two distinct traces and a third (on left) in process of separating from xylem. , from cast, BMNH. Mi2112. All x 3. Some of the isolated molars, and all the incomplete teeth, cannot be definitely sorted into second and third molars. Recourse was therefore made to probability paper, by which means and standard deviations could be estimated, assuming equal numbers of M, and Mg in the collection. Altogether, 5 independent specimens containing lower molars are referred to the large form, as against 22 referred to the typical form. The three molars are so much alike that a single description will suffice for all. There are two roots, flattened from front to back and inclined somewhat posteriorly. The posterior root is the stouter of the two, especially in Mg, where its posterior surface is more rounded. The length of the roots on M, and M, is 20-24 mm. A low inter-radicular crest is present on the base of the tooth. Of the two V-shaped buccal cusps, the protoconid is a little lower than the hypo- conid and occupies a smaller area on the crown. The anterior crest of the protoconid falls rapidly as it passes in a curve along the anterior border of the crown, and at its lingual end it becomes a cingulum ledge anterior to the base of the metaconid. There is no paraconid, though a notch may be worn in the crest to give the appearance of that cusp. A narrow cingulum on the anterior edge of the crown does not extend to the buccal side of the protoconid. The posterior arm of the protoconid remains nearly horizontal, crossing the crown to join the metaconid. The latter cusp is higher than the protoconid ; its base extends forward so as partly to close the trigonid basin. Closely applied to the posterior surface of the metaconid is the metastylid, the tip of which is lower than that of the metaconid. In three specimens the metastylid is represented only by a sharp posterior crest of the metaconid. The arms of the hypoconid diverge at a greater angle than those of the protoconid. The anterior arm ends between the metaconid and the metastylid, so that the latter cusp EAST AFRICAN CHALICOTHERES 187 projects backwards partly closing the talonid basin. The posterior arm of the hypoconid runs to the tip of the entoconid, a conical cusp, somewhat lower than the metaconid, and placed at the posterolingual corner of the crown. The trigonid basin opens lingually anteriorly to the base of the metaconid, and the trigonid basin opens between the metastylid and the entoconid. The posterior edge of the crown is occupied by a cingulum, somewhat variable in development. When best developed it rises towards the lingual side to form a vertical rib on the posterior face of the entoconid ; this rib represents the hypoconulid. Fic. 5. A-£, Right M, of Chalicotherium rusingense from Rusinga, X I, A, crown view ; B, buccal view ; c, lingual view ; D, anterior view ; E, posterior view. F, lower pre- molars of C. rvusingense, R 1782.50 x I. G, M,—P3 of C. grande, BMNH. M4o821 and (P3) a Paris specimen, x 3. H, M,—P3 of C. goldfussi, from cast, BMNH. M2719, x 3. The effect of wear is to expose a strip of dentine on the crests of the protoconid and the hypoconid. These cusps are thus reduced in height. The metaconid and entoconid however are worn mainly on their buccal sides, and their height is reduced more slowly. Comparisons. The lower molars of C. rusingense are much smaller than those of C. grande or C. wetzlert, but a little larger than those of C. pilgrimi. Typical speci- 188 EAST AFRICAN CHALICOTHERES mens are similar in size to those of Schizotherium turgaicum and somewhat smaller than those of S. priscum. In Schizotherium the hypoconulid of M, is enlarged to form a posterior heel, probably the relic of a larger structure present in Eomoropinae. In C. rusingense, as in C. grande and other species of Chalicother1um, M, resembles Mg. The metastylid is distinctly developed in Schizotherium, as well as in C. pilgrimi, C. wetzleri and (with a few exceptions) in C. rusingense. In C. grande and other late species it is more or less merged into the metaconid. The width/length index of the lower molars is somewhat greater in C. grande (56-58 in specimens of M, from Sansan) than in C. rusingense (M2 : 53-55). Wehrli’s (1939) measurements show that in C. goldfussi still broader molars can occur (M, : 55-62). Comparatively narrow molars (index of M, below 55) occur in C. weézlert, C. pilgrimi and species of Schizotherium. Molar Occlusal Relations. (Text-fig. 6.) The functional inter-relations of the upper and lower molars were studied by examining the wear facets and by fitting teeth together. In the centric position, in which the teeth are pressed together as closely as possible, the lower molar covers the lingual part of the upper molar, the inner borders of both teeth being in line. The inter-relations of the cusps in the centric position may be tabulated as follows :— Hypoconid tip Centre of central valley, between tips of paracone and metacone. Protoconid tip Cingulum anterior to the protoconule. Entoconid Groove between protocone and hypocone. Metaconid-metastylld Space between protocone and the more anterior hypocone. Paracone tip Buccal side of tooth, in valley separating protoconid and hypoconid. Metacone tip Embrasure between two lower molars, partly filled by posterior cingulum. Protoconule Buccal side of metastylid. Protocone Between metastylid and entoconid. Hypocone tip Against anterior arm of protoconid, in which it wears a notch. During chewing, the crests of the protoconid and hypoconid slide up the lingual surface of the ectoloph. This can take place only when the lower jaw is displaced to the lateral side of the centric position. The movement was almost certainly ectal, starting when the lower teeth are placed so that the protoconid and hypoconid touch the parastyle and mesostyle respectively ; from this position the lower teeth move medially, upwards and slightly forwards to the centric position. The facets of wear produced in chewing can without difficulty be homologized with those distinguished in other perissodactyls (Butler 1952). EAST AFRICAN CHALICOTHERES 189 7 (€ 3a Fic. 6. Molar occlusion of Chalicotherium rusingense. imposed in centric relation. 3b 10 ? A, upper and lower molars super- B, C, wear facets on the molars. Io. Lower facet Near the edge of the anterobuccal surface of the anterior crest of the protoconid. Posterior surface of the protoconid-meta= conid crest. (a) Buccal surface of metaconid and meta- stylid, near the tip. (b) Posterior face of metastylid, near the tip. Upper facet Lingual face of the posterior crest of the metacone. Anterior face of paracone and protoconule. Lingual face of protoconule. Anterior crest of protocone. (These two facets are variable in development, and may be absent, especially 3 (b). may unite.) (Metaconid-hypocone contact is absent.) Edge of anterior crest of protoconid. Anterior face of the buccal part of the hypoconid-metastylid crest, near the edge. Posterior face of the hypoconid-entoconid crest. A posterobuccal facet on the entoconid, involving the vertical rib. Anterior surface of entoconid. When the protocone-protoconule crest is well developed the two facets Edge of posterior crest of hypocone. Posterolingual face of the paracone- meso- style crest. Anterior face of metacone and metaloph. Anterolingual face of hypocone, near its tip. Posterior crest of protocone. (Contact slight and sometimes absent.) Lingual surface of hypoconid. Of these, 1, 2, 6 and 7 from the main chewing surfaces. lingual cusps (3, 8 and 9g) are probably significant only as stops. Buccal surface of protocone. Contacts between the When the lower jaw swings to the lingual side of the centric position, contacts 5 and 10 would ensure occlusal balance while the teeth on the opposite side of the mouth are in use. 190 EAST AFRICAN CHALICOTHERES Lower Premolars. (Text-fig. 5.) Lower premolars are poorly represented in the collection. The mandible F 3608 shows their roots in the left ramus ; three isolated teeth (R 1782.50) probably represent P,—P, of a single individual, but they are in poor condition. Two more worn and isolated teeth (R 738.50 and R 739.50) may be Ps; and P, of another individual ; they differ in some respects from R 1782.50 ,and may not be correctly identified. Finally, there is a broken specimen of P, (R 12.48). F 3608 shows that P, was absent. The total length of P,—P, is 37 mm., or 53% of the total length of the molars. All the premolars are two-rooted ; their length diminishes from P, to P,. In R 1782.50, Py is shorter and proportionately broader than M, (15.5 x I0 mm.). Its roots are flattened anteroposteriorly, but are less distinctly separated at the base than those of the molars. The crown-pattern differs from that of the molars in the following respects : the anterior arm of the protoconid runs in a rather more anterior direction to the middle of the anterior edge of the tooth, where it turns lingually, fal- ling rapidly to form a cingulum that merges into the anterior base of the metaconid ; there is no metastylid ; the hypoconid is much lower than the protoconid, and the crest which connects the hypoconid to the buccal surface of the metaconid is corres- pondingly low ; the entoconid (broken off) appears to have occupied a smaller proportion of the crown than in the molar, and the posterior cingulum is probably absent. TABLE IV. Measurements of lower molars of two specimens of C. rusingense. F 3608 1782.50 Trigonid Talonid Trigonid Talonid Length Width Width Length Width Width M, 26°8 14:0 13°6 26:8 I4°4 14°3 M, 23°8 13°0 12°7 26:4 14:0 13°9 M, 183 105 II'5 19°8 — — P3 of R 1782.50 is smaller in all dimensions than P, (13 x 9 mm.), and is much less molariform. The anterior arm of the protoconid forms at the anterior end of the crown an angulation that is probably to be interpreted as a small paraconid. The metaconid does not rise to a distinct point, but is united with the protoconid to form a ridge that runs transversely to the crown and slightly backwards. There is no meta- stylid. The hypoconid is only about half as high as the protoconid, and, though the postero-lingual part of the crown is broken, the entoconid must have been very low and cingulum-like. On P, (10 x 6.5 mm.) the protoconid is the only well-developed cusp. An anterior ridge inclined lingually represents the paraconid, and a lingual EAST AFRICAN CHALICOTHERES 191 ridge the metaconid. The talonid is rudimentary. From the small hypoconid a ridge passes forward to merge with the posterobuccal surface of the metaconid ridge. There is a low, cingulum-like entoconid. R 738.50 differs from P,4 of R 1782.50 in its greater posterior breadth (16 x 11.8 mm.), due to the presence of an accessory cusp on the buccal side of the hypoconid (probably an abnormality). The tooth is very worn, and the presence of a metastylid cannot be affirmed, though it is probable. The entoconid, lower than the hypoconid, stands rather independently from the other cusps. There is no posterior cingulum. R 739.50 (13.3 X 9 mm.) is a specimen of P3 which agrees with R 738.50 in the presence of a small cingulum cusp buccal to the hypoconid, and it probably comes from the same individual. The entoconid is a small elevation on a cingulum at the posterolingual edge of the crown. In R 12.48, which is a specimen of P,, most of the trigonid has flaked away. The hypoconid is even more reduced than in R 1782.50, and the entoconid is represented only by a fragment of cingulum. Comparisons. The premolar/molar index of C. rusingense (53) is close to those of C. wetzleri (54 in the Bollingen specimen) and Schizotherium priscum (54, from Osborn’s figure, 1913) ; it is exceeded by S. cf. avitum (59, Teilhard de Chardin 1926). In some specimens of C. grande the index is reduced (47 in two specimens from Sansan, 44 in a Spanish specimen figured by Villalta & Crusafont 1943). In Schizotherium priscum (Osborn 1912) and S. turgaicum (Borissiak 1928) P, possesses a weak metastylid, possibly present in one of the specimens of C. rusingense. P, of S. priscum and S. cf. avitum is more equal to P, than in C. vusingense, and more similar in pattern ; in C. grande P, is smaller and simpler than P,, its metaconid being incompletely differentiated, as in C. rusingense. P, of C. grande is more reduced than in C. rusingense ; the metaconid ridge is weak and the roots may fail to separate. Thus reduction of the premolars has proceeded farther in C. rusingense than in Schizotherium, but not quite as far as in C. grande. TABLE V. Dimensions of lower molars of C. rusingense, inclusive of large form. N m Ss V M, + M, Length 17 24°9 mm. 1°75 7:0 Trigonid width 24 13°4 mm. 0°88 6:6 Talonid width 20 13°3 mm. 0°54 4°1 (range) M, Length 3 18-4 mm. 17:2-19'8 Trigonid width 2 10°6 mm. 10*5—10°7 Talonid width 3 II°5 mm. 10:7-12°4 Canines and Incisors. (Text-figs 7, 9.) In the holotype a diastema of 26 mm. separates P? from the upper canine. This is a small, curved tooth, measuring 11-5 < 10:3 mm. at the base of the crown, and 192 EAST AFRICAN CHALICOTHERES 2a ‘Gis 9 Zags aa i a ‘ t 1 ' Fic. 7. Partial reconstruction of the face of Chalicotherium rusingense, based on the holotype. Side view and palatal view, x 4. 19 mm. from the tip to the posterior border of the alveolus. It is placed procumb- ently in the jaw, its root making an angle of 60° to the line of the cheek teeth. The tipis blunt. There is a posterior crest near the tip, and a more rounded ridge on the buccal side (corresponding to the rib on the buccal slope of the paracone in the premolars and molars). The lingual surface is poorly preserved, but there was probably a short lingual crest near the tip. The anterior surface is evenly rounded. Anterior to the canine the alveolar border is preserved for a length of 12 mm., but it contains no teeth. The premaxilla is very slender, and though its tip has been broken off it is unlikely that any upper incisors were present. In F 3608 and R 283.48 a lower canine root can be seen on either side. The lower canine must have been smaller than the upper canine, measuring in section about 10 X 7mm. It was procumbent, and was separated from P, by a diastema, 26 mm. long in F 3608 and 28 mm. in R 283.48. Three incisor alveoli can be seen on each side of F 3608, immediately anterior to the canine. Their labial walls are broken away, but it is clear that the alveolus for I, was the largest. The incisor from Malembe, Congo Republic, described by Hooijer (1963) as an upper incisor of Macrotherium (?) spec., is not referable to C. rusingense. It is much too large to fit even into the largest lower incisor alveolus. The molar fragment EAST AFRICAN CHALICOTHERES 193 from Malembe is also much larger than in C. vusingense : its talonid width is 25 mm., compared with a mean talonid width of 13-3 mm. for M, and Mj of C. rusingense. Comparisons. The anterior teeth of Schizothertum priscum are unknown. An isolated tooth tentatively identified by Teilhard de Chardin (1926) as the lower canine of S. cf. avitwm measures 18 < 9:5 mm. ; this is larger in proportion to M, than in C. rusingense. The juvenile mandible from St. Gérand-le-Puy described by Filhol (1879) appears to belong to Phyllotillon. A juvenile mandible of C. grande from Sansan, figured by Filhol (1891) and now in Paris, shows alveoli for a canine and three incisors. These probably belong to the milk dentition, for gubernacular foramina are present in association with the first two incisors. Id, is the largest of the incisors, and the canine is larger than any of the incisors. One of Lartet’s specimens of “‘ Amisodon magnum” contains the broken alveolus of a lower canine, separated from P, by a diastema 29:5 mm. long. An isolated tooth lying shortly anterior to the lower jaw of the skeleton of C. grande described by Filhol (1891) may be an incisor (see Viret 1958, pl. 1). Nothing is known of the anterior upper dentition of C. grande. A lower milk canine appears to have been present in C. goldfussi, if C. baltavarense is correctly identified with that species (see p. 168). In C. brevivostyis upper canine and incisors are absent, as in Nestoritherium sivalense. Face and Palate. (Text-fig. 7 ; Table VI.) Of the skull, only part of the face and palate are known, mainly from the holotype, but supplemented by fragments of maxilla. Anterior to M! the length of the face is 1:23 times the length of the upper molar series. The external narial opening extends back to above P? ; it is flanked by a slender process of the small premaxilla. The zygomatic process arises above the posterior end of M?, and the anterior end of the orbit was probably situated above the anterior part of M*. The infraorbital foramen is preserved only in R 483.51, where it opens above M1. In other specimens the floor of the infraorbital canal can be traced back to the orbit. The canal and the foramen are situated fairly high on the face, about 25 mm. above the alveolar border of the cheek teeth. The maxillary sinus is represented by a small cavity between the infraorbital canal and the roots of the molars. TABLE VI. Measurements of face and palate, C. rusingense, holotype. Antemolar length . 5 ; : : : . 123 mm. Length diastema 2 . . é : : : 22 mm. M1_M§ inclusive : : : : Z ; Fi 75 mm. P2_P# inclusive ; F : : : : 3 38 mm. C—P* inclusive . ¢ : 3 ; 80 mm. Infraorbital Feeameneeelveotan orden 3 3 ‘ 32 mm. Lower edge of orbit—alveolar border . j ‘ : 41 mm. Height of zygoma_ . 3 ; : 30 mm. Width of palate between first molars (est. Nc . . 58 mm. 194 EAST AFRICAN CHALICOTHERES The palate is transversely arched. Its width between the first molars must have been about 45 mm., and it probably increased somewhat in width posteriorly. It is pierced by a foramen medially to M!. The posterior border of the palate has not been preserved. Comparisons. The infraorbital foramen is above the anterior part of M?! in C. rusingense, C. grande, C. brevirostris and Schizotherium priscum, but in C. pilgrimi, probably due to the youth of the specimen, it is a little farther forward, above the posterior part of P*. The height of the foramen on the face is least in Schizotherium, most in C. grande : the index, distance of foramen from teeth/molar length, is 25 in S. priscum, 29 in C. pilgrimi, 32 in C. rusingense and 43 in C. grande. In S. priscum and C. grande the zygoma arises above the posterior end of M?, as in C. rusingense, but in C. brevirostris it arises a little farther forward (mid M#?) ; in juvenile specimens, such as the holotype of C. pilgrimi, it reaches the level of the anterior end of M?. The length of the face is unknown in Schizotherium and in C. grande. It has almost certainly been shortened in C. brevirostris, for in this species it is much shorter than in C. rusingense. The anterior margin of the orbit in C. brevirostris is above the i Fic. 8. Mandibles of Chalicotherium. A, C. vusingense, from Rusinga, F 3608. B, C. wetzleri (holotype of Palaeotheriwm schinzii von Meyer) froma cast. Cc, C. gvande, Paris specimen from Sansan. All x 4. EAST AFRICAN CHALICOTHERES 195 posterior part of M*. Unfortunately the known skulls of C. grande are so badly crushed that the exact position of the border of the orbit is uncertain. The distance between the orbit and the infraorbital foramen appears, however, to be greater in C. rusingense than in either C. grande or C. brevirosins. Mandible. (Text-figs. 8,9 ; Table VII.) F 3608 is the only specimen in which more than a fragment of the mandible is preserved. It consists of both horizontal rami, but the posterior part of the mandible has been broken off on both sides. R 283.48 shows the region of the symphysis. The ramus is moderately deep, its depth increasing posteriorly. Below the molars the rounded lower border is very slightly concave in lateral view, falling and be- coming sharper behind M, to indicate the angular process (most of which is missing). The lateral surface below the teeth is nearly flat, but shows a weak longitudinal groove about 14 mm. above the lower border. This groove fades out below the premolars, where there is a slight convexity. The masseteric fossa is very flat and hardly noticeable. The coronoid process appears to be inclined at an angle of 45° to the alveolar border. Internally, the ramus is moderately convex below the molars; about 15 mm. above the lower border there is a slight groove, presumably for the mylohyoid muscle. The mandibular foramen is low, its lower border being 22 mm. below the level of the alveolar border. The pterygoid fossa is very shallow. Fic. 9. Chalicotherium rusingense, F 3608, symphyseal part of mandible, in dorsal and ventral views, x }. The two rami converge at an angle of 30°. The symphysis extends back to the posterior end of P,. The symphyseal region is nearly in line with the horizontal ramus, but shows a slight upward tilt. Its vertical depth is less than that of the ramus, and there is a step in the alveolar border anterior to P,. The lower surface of the symphysis is evenly convex in a transverse direction, and its upper surface is concave to form a gutter. In the region of the diastema the alveolar border forms a sharp ridge, which turns laterally immediately behind the canine, where it overhangs the lateral surface of the jaw. The mental foramen is placed below the diastema, about midway between P, and the canine. Above it is a small foramen, close to the alveolar border. 196 EAST AFRICAN CHALICOTHERES Comparisons. C. rusingense resembles C. grande and C. wetzleri in the general shape of the horizontal ramus of the mandible. The anterior part of the mandible of C. grande is known mainly from a juvenile specimen figured by Filhol (1891, pl. 44). This differs from C. vusingense in the shorter symphysis and diastema. In an adult specimen (Filhol 1891, pl. 45) the diastema is only 58% as long as the premolar series, compared with 80°% in C. rusingense, and it is probable that the anterior part of the mandible, and therefore also the tip of the snout, was more abbreviated in C. grande than in C. rusingense. A further difference is that the mandibular foramen of C. grande is in line with the teeth, whereas in C. rusingense it is at a lower level. The mandible of Schizotherium is very poorly known. From the figure of Osborn (1912) it may be seen that the alveolar border is stepped down anteriorly to P, as in Chalicotherium. In Schizotherium sp. (Bohlin 1946) the symphysis ends a little anteriorly to P,, and the diastema was evidently longer than in C. vusingense. In this specimen the coronoid process appears to rise rather steeply, resembling Eomoropus (Osborn 1913). In S. cf. avitum (Teilhard de Chardin 1926) the diastema is longer than the premolar series, and the symphysis ends far anterior to Py, as in Eomoropinae. Shortening of the anterior part of the mandible appears to be a trend of chalicotherian evolution, and it is probable that C. rusingense is more primitive than C. grande in this respect. TABLE VII. Measurements of lower jaw, F3608. Antemolar length (est.) . : ; é ; : 95 mm. Between M, and the canine alveolus . ‘ ‘ ; 63:5 mm. Length of diastema . 6 : : 4 ; : 27°5 mm. Length of symphysis (est.) : ; : : ; 63 mm. Length M,—M, inclusive . 6 6 5 : : 70 mm. Length P,—P, inclusive. : : . : ¢ 36 mm. Depth at posterior end of M, (perpendicular to lower border) . ‘ 2 : : : : : 49°5 mm. Depth at anterior end of M, : : > 9 ; 34 mm. Depth at diastema . : : ; : : ; 22 mm. Width between canines . 5 : : : c 14 mm. Minimum width behind canines . ; : 5 . 26 mm. Width across M, talonids . ‘ : : : : 92 mm. Scaphoid. (Text-fig. 10 A-F.) The only specimen of this bone is F 2077, which belongs to the left side. It isa proximodistally flattened bone with a distal (or ‘‘centrale’’) process, situated towards the ulnar and dorsal (= anterior) sides. Its dimensions are : height 31 mm., dorsovolar length 31 mm., radio-ulnar width 44-5 mm. In proximal view, the bone is transversely widened ; most of the proximal surface is occupied by the facet for the radius. This facet is concave in a dorsovolar direction ; its ulnar edge meets the proximal margin of the lunate facet in an acute angle, and its margin is slightly EAST AFRICAN CHALICOTHERES 197 raised near the middle of the dorsal side and again at the ulnar end of the volar side. The radial third of the proximal surface of the scaphoid forms the radial process, and the middle of the volar side extends as a small volar process. LIGA 1 AG, Fic. 10. A-—F, left scaphoid of Chalicotherium rusingense, F 2077. A, proximal view ; B, distal view ; c, radial view ; D, dorsal view ; E, volar view ; F, ulnar view. G-1, corresponding views of scaphoid of C. gvande, Paris specimen from Sansan. All x }4. Labelling of facets: J/.d., distal lunate ; /.p., proximal lunate; m, magnum; 17, radius ; ¢, trapezoid. 198 EAST AFRICAN CHALICOTHERES In dorsal view the scaphoid is wider than long. The dorsal surface is divided by a ridge which runs from the elevation in the dorsal margin of the radius facet to the distal process. This ridge divides a smooth, triangular proximo-ulnar area from the remainder, which is roughened. The roughening is very marked in a line from the radial process along the edge of the trapezoid facet, presumably for ligamentary attachment. The trapezoid facet occupies a large area on the distal side of the bone, extending from the radial process to the radial side of the distal process. It is slightly saddle- shaped, being weakly divided by a ridge which arises from the tip of the distal process, near the dorsal side, and crosses the facet to end below the volar process. Radially, the trapezoid facet does not reach the tip of the radial process, but it leaves a small area of smooth bone that might have articulated with a trapezium. On the ulnar face of the scaphoid there is a proximal facet for the lunate, but no distal lunate facet, the ulnar surface of the distal process being rough. The magnum probably articulated with a small convex area at the tip of the distal process. Comparisons. The scaphoid of Schizotheriwm is unknown. In comparison with C. grande, of which four examples were studied in Paris, the scaphoid of C. rusingense appears to be compressed in a proximodistal direction. Its distal process is shorter and less slender. In C. grande the proximal lunate facet is hardly indicated, but there is a large distal lunate facet on the ulnar side of the distal process ; in C. rusingense, on the other hand, the proximal lunate facet is well developed, but the distal facet is absent. In Phyllotillon betpakdalensis (Borissiak 1946) and Movropus elatus (Holland & Peterson 1913) both facets are equally developed, and this is probably the primitive condition. The two species, Chalicotherium grande and C. rusingense, agree in the cylindrical concavity of the radius facet, which is saddle-shaped in Phyllotillon and flat in Moropus, and in the small and indistinct contact with the magnum, the distal process terminating in a rounded point, instead of being truncate as in the other genera. Third Metacarpal. (Text-fig. 11 A-E.) F 2070 is the proximal end of a right metacarpal III. The transverse diameter of its head slightly exceeds the dorsovolar diameter (37°5 < 35°5mm.). The proximal surface is crossed by two dorsovolar keels, which demarcate the rectangular facet for the magnum. The ulnar keel is much the higher of the two. On its ulnar side is the unciform facet, and the radial part of the proximal surface is occupied by a facet for metacarpal II. The unciform facet extends onto the proximal surface of the prominent dorso-ulnar process, which overhangs a cavity on the ulnar side of the head of the metacarpal for articulation with metacarpal IV. The facet for metacarpal II is triangular, tapering in a volar direction. It is convex in a dorsovolar direction ; its volar part is approximately horizontal and its dorsal part nearly vertical. The dorsoulnar part of the facet faces radially, extending onto the radial side of the keel that divides it from the magnum facet. The meta- carpal facet overhangs the radial side of the bone. The magnum facet is also convex in a dorsovolar direction, but its dorsal part does not reach so far distally as EAST AFRICAN CHALICOTHERES S oe, SS SIN <3 SSS =e. \4l »y | of RS? Yi My Sl Nae A-E, right fourth metacarpal of Chalicotherium vusingense, from site R 106. FIG. 12. E, proximal view. A, dorsal view ; B, volar view; c, ulnar view ; D, radial view ; F-J, corresponding views of C. gvande, Paris specimen from Sansan. All x 4. EAST AFRICAN CHALICOTHERES 203 In dorsal (anterior) view the astragalus is much broader than high: its transverse diameter is 67 mm. and its height on the tibial side is 35-5 mm. The trochlea is correspondingly reduced in height, with the keels only moderately developed. Each keel subtends an angle of about 120°, but the fibular keel faces less proximally than the tibial keel ; in dorsal view it reaches farther proximally and distally. The neck of the astragalus, distally to the trochlea, is very short (about 5 mm. high). In volar (posterior) view may be seen a large, slightly concave, ectal facet on the fibular side, separated by a groove from the small, rounded, slightly convex sustentacular facet on the tibial side. The ectal facet extends more than half-way across the astragalus. Proximally it meets the trochlea in an acute edge, and distally on the fibular side it meets the small, flat distal calcaneal facet in a blunt angle. The sustentacular facet is separated proximally from the tibial part of the trochlea by a groove that leads to the astragalar foramina. In distal view the fibular half of the astragalus is seen to be much narrower than the tibial half : the dorsovolar diameter at the fibular keel is 24 mm., and at the tibial keel it is 43 mm. Most of the distal surface is occupied by the navicular facet, which extends from the tibial border more than half-way across the bone. It is divided by a rounded convexity into two slightly concave areas, the larger one being dorsal, and the smaller one tibial and volar. Owing to breakage the presence of a cuboid facet cannot be determined ; it could not have been more than of limited extent. Fic. 13. Left astragali of Chalicotherium. Left, C. rusingense, MFW 1208.55, anterior (dorsal), posterior and distal views. Centre, corresponding views of C. grande, Paris specimen from Sansan. Right, corresponding views of C. goldfussi, from Pikermi, BMNH. M11351. All x 4. Key: c, cuboid facet ; e, ectal facet ; ”, navicular facet ; s, sustentacular facet. 204 EAST AFRICAN CHALICOTHERES The tibial surface of the astragalus is much roughened, and near the trochlea it is marked by a deep pit for a ligament. Comparisons. The astragalus of Schizotherium turgaicum (Borissiak 1946, Belyaeva 1954) is proportionately much higher than in C. rusingense, but the neck is equally short, the greater height being shown in the trochlea. The trochlear keels are more acute, and, on the volar side, the sustentacular and ectal facets are more elongated proximo-distally. In distal view the astragali of the two species are more similar, except that in C. rusingense the fibular part of the bone is narrower relatively to the tibial part. There is no cuboid facet in S. turgaicum. In C. grande the astragalus is lower than in C. rusingense, and has blunter trochlear keels. The ectal facet is deeper, and does not extend so far towards the tibial side (see Wegner 1913). In distal view the fibular part of the astragalus is narrowed to the same extent as in C. rusingense. A cuboid contact is present in C. grande in the form of a strip along the posterior edge of the navicular facet (figured by Holland & Peterson 1913). The navicular facet is tilted towards the tibial side to a more marked degree in C. grande than in C. rusingense, and, probably associated with this, the tibial height of the astragalus is only 75% of the fibular height in C. grande, compared with 87°% in C. rusingense. Second Metatarsal. (Text-fig. 14 ; Table VIII.) There are three examples of this bone. The best preserved is MFW1213.55. F2071, which is somewhat weathered, is similar but a little longer. Sgr82.48 is a small, juvenile specimen. The proximal end is triangular in outline, with dorsal and volar processes, situated towards the fibular side, and a tibial process. Most of the proximal surface is occupied by the facet for the mesocuneiform, slightly concave in the tibio-fibular direction. On the fibular side of this is the facet for the ectocuneiform, which in MFW1213.55 is differentiated into dorsal and volar portions. The dorsal portion faces in a fibular direction, and is separated from the mesocuneiform facet by a sharp crest ; the volar portion faces proximally, and is separated from the meso- cuneiform facet only by alow crest. In F2071 the dorsal portion is missing, perhaps due to weathering. There is no entocuneiform facet. On the fibular side, below the ectocuneiform facet, is an overhung facet for metatarsal III ; this apparently does not extend to the extreme volar side of the bone. Below it again is a roughening of the surface for ligamentary attachment. The dorsal process projects beyond the fibular side of the bone and extends for a short distance down the shaft. The middle of the shaft is oval in section, somewhat flattened in a dorsovolar direction. The bone as a whole is short in comparison to its width: the total length is 2-o—2-1 times the maximum width of the distal end. The distal articulating surface is placed obliquely to the long axis, its tibial side being more proximal than its fibular side. In F2071 its dorsal border is separated from the shaft by a shallow groove, but this is not present in MFW1213.55. Dorsally, the articulating surface projects beyond the level of the shaft, and forms a smoothly convex area for the phalanx. The volar part of the articulation, for EAST AFRICAN CHALICOTHERES 205 STW 4 15 nae po C.grande Fic. 14. Right second metatarsals of Chalicotherrum. C. vusingense, MFW 1213.55, in dorsal, ulnar, volar, radial, proximal and distal views. C. pilgvimi, BMNH. M12168, corresponding views. C. gvande, Paris specimen from Sansan, corresponding views. All x 4. 206 EAST AFRICAN CHALICOTHERES the sesamoids, is divided by a median keel which starts at the most distal end of the bone. The groove on the tibial side of this keel is more deeply cut than the groove on the fibular side. Proximally to the distal articulation, apophyses for ligaments are developed on the lateral sides of the bone ; below each, but most marked on the fibular side, is a pit. Comparisons. In Schizotherium priscum (Filhol 1893) and S. turgaicum (Borissiak 1921) the second metatarsal is much more elongated (length/distal width 4-0 and 4:3 respectively, cf. 2:0-2-1 in C. rusingense). In C. pilgrima (BM.12168) it is shortened to the same degree as in C. vusingense (2:1) but in C. grande it is somewhat shorter (I-°7—1:9). In proximal view the head is more compressed in a tibio-fibular direction in Schizotherium than in Chalicotherrum. C. pilgrimi is very much like C. rusingense in the shape of the proximal end, but in C. grande the head is trapezoidal rather than triangular, owing to broadening on the volar side. In S. turgaicum the ectocuneiform facet is divided into two. In the species of Chalicotherium it is single, but C. rusingense and C. pilgrim: show more clearly than C. grande a functional division between a more vertical dorsal part and a more proximal volar part. In C. grande the crest separating the ectocuneiform and mesocuneiform facets is blunt throughout its length, whereas in C. rusingense and C. pilgrimi its dorsal half is acute. TABLE VIII. Measurements (mm.) of metatarsals of C. rusingense. Second Third MFW Ser. 1213°55 F2071 84:48 | F2076 R30 R648-47 Rot Width of head : : ; 2, 26 21°5 25 23°5 175 23 Dorso-volar, head . 5 : 28 24°5 20 28 30°5 24 25 Minimum width, shaft ; : 20°5 23°5 15'5 23 22°5 15°5 19 Distal width, across articulation. 28 29 21 20'5 27 21°5 24 eng thi: ; : ; ; 58 58°5 50 69 62°5 58 57°5 The overhung facet for metatarsal III is double in S. turgaicum, single in the three species of Chalicotherium. At the middle of its length the shaft is more nearly circular in cross-section in Schizotherium than in Chalicotherium where it is somewhat broadened. In C. grande the shaft is broader in comparison with the distal width than in C. vusingense or C. pilgrim. Third metatarsal. (Text-fig. 15 ; Table VIII.) Four specimens are identified as this bone, although they differ somewhat in size and proportions. The longest specimen is F2076 (left). Another specimen EAST AFRICAN CHALICOTHERES 207 (left) from R3o is of the same width but slightly shorter. 648.47 (left) and a specimen from Ror (right) are smaller and more slender, and may be juvenile. The ratio of length/distal width ranges from 2-2 to 2:6, and it is probable that metatarsal III was slightly longer than metatarsal II of the same individual, though of similar width. There is some variation in the shape of the section of the middle of the shaft : thus in F2076 the transverse diameter slightly exceeds the dorsovolar diameter, but in R648.47 the dorsovolar diameter is the greater, while the specimen from R30 is nearly circular in section. The proximal surface is roughly triangular or trapezoidal. It is inclined distally towards the dorso-fibular apex, and is occupied by a very slightly convex facet for the ectocuneiform. There is no distinct facet for the cuboid. The fibular edge, which is nearly perpendicular to the straight dorsal edge, is slightly concave in F2076, straight in the other specimens. The third edge, from the tibio-dorsal apex to the fibulovolar apex, is convex. In all specimens the dorsovolar diameter of the head is greater than the transverse diameter, but less markedly so in the specimen from Ror than in the others. Fic. 15. Right third metatarsals of Chalicotherium. Above, C. rvusingense, F 2076, in dorsal, ulnar, proximal and distal views. Below, C. gvande, Paris specimen from Sansan, corresponding views. All x 4. 208 EAST AFRICAN CHALICOTHERES On the most proximal part of the fibular surface of the metatarsal are two facets for articulation with metatarsal IV, slightly overhung by the fibular border of the proximal surface. Each facet occupies the proximal part of a pit in the fibular surface. The more dorsal facet is the larger ; it is somewhat concave, and faces slightly in a volar direction. The volar facet is smaller and flatter. The distal end is tilted towards the tibial side to a greater extent than in metatarsal II. The fibular ligamentary apophysis is much better developed than its tibial counterpart, whereas on metatarsal II they are more equal. The dorsovolar diameter of the distal end is relatively greater than in metatarsal II, exceeding the transverse diameter. The sesamoid articulation is accordingly more extensive than the phalangeal articulation. The fibular sesamoid groove is more deeply incised than on metatarsal II, though it remains shallower than the tibial groove. Comparisons. This metatarsal is much shorter in comparison with its width than in Schizotherium priscum (Filhol 1893) and S. turgaicum (Borissiak 1921). It is much more like C. grande, though only the specimen from R30 agrees with that species in proportions, the other specimens being relatively longer. In Schizotherium metatarsal II is much more slender than metatarsal III, but in C. grande the two metatarsals are of equal width, and this appears to be so also in C. rusingense. C. rusingense resembles Schizotherium in the proportions of the proximal end of the bone ; in C. grande it is rather broader. In C. grande there is a distinct cuboid facet, absent in C. rusingense and in Schizotherium. C. rusingense also agrees with Schizotherium in the possession of two distinct facets for metatarsal IV ; in C. grande the volar facet is apparently absent. The distal ends of metatarsals II and III of C. grande differ in the same ways as those of C. rusingense, but the differences are less marked. Basal phalanges. (Table IX.) Of 43 basal phalanges, 23 are referred to the manus, because of their larger size and their resemblance to phalanges of the manus of C. grande. The 23 specimens fall into three groups, regarded as representing digits II, III and IV. Digit II of manus. (Text-fig. 16A). This type of phalanx is so identified from its resemblance to the basal phalanx of B.M. M8638, a complete digit of C. grande, including the metacarpal. It is broader proximally than distally. The proximal border forms two lobes of which the radial is the more prominent. The facet for the metacarpal occupies the proximal half of the dorsal surface of the bone. It is concave, and faces very slightly proximally and radially. Its ulnar border is marked by a prominent crest which reaches its greatest height at the disto-ulnar extremity of the facet. Its radial border is lower, broadening out distally to form a crescent- shaped flattened area that possibly marks the insertion of an extensor tendon. Distally to the metacarpal facet the phalanx narrows to a waist, especially conspicuous in lateral view. On the volar side there is a broad longitudinal concavity between two marginal ridges. Each of these ridges is developed into a proximal prominence, and about half-way along the bone there is a second, smaller prominence EAST AFRICAN CHALICOTHERES 209 TABLE IX. Measurements (mm.) of some phalanges, C. rusingense. Manus II Manus III Manus IV Pes Basal phalanges MFW R450.47 R136a.49| R189.47 Ro45.47| R136.49 F2074 | 1214.55 R452.47 Ulnar length 56°5 49'3 60-2 57 63 56-2 42°4 39 Radial length 59 50 58-3 54°5 59°7 51:2 40°2 30°5 Length of metapo- dial articulation 25°7 20°8 25:2 21°5 27°5 25°7 Da 20:2 Length distal to metapodial arti- culation 22 20°5 23°1 23°3 24 19'7 15 13°5 Proximal width 34 31 36°5 33 39°5 33°5 30°5 27°4 Distal width 23 22°5 25°1 22B 27 24:2 23 19°8 Thickness 25°5 23°6 26:7 22°90 20°5 23°2 19°5 19°4 Manus Pes Middle phalanges MFW R136b.49 R757.47 R748.50 | 1217.55 R233.51 R844.48 Maximum length . : ; F 45 39 38 30 28 25°5 Proximal width . : : : 27°5 21 24 21°5 18:5 18°5 Width across trochlea . A ; 22 15'5 19 18°5 15'5 14 Height of trochlea. : ; : 37 28 30 24 22 21°5 Manus Pes Ungual phalanges R136c.49 R134.49 R160.49 Rg22.47 Maximum length . 3 c . 96 — — 64e Length of base. : : : 66 60 48 41 Height of articulation . ; : 30 26e 23 21 Height at cleft : ; : ‘ 44 43 40 32 Width of base ; : : 6 25 20°5 25 20 Width of articulation . j : 22°5 18 19 16 (intermediate volar tubercle). Distally, the marginal ridges merge into the keels that form the edges of the articulation for the middle phalanx. This articulation, which occupies the distovolar part of the bone, is in principle a trochlea, with median groove and lateral keels. However, it usually shows an incomplete subdi- vision into a distal part and a volar part. The bottom of the groove in passing from one part to the other forms a slight but distinct angulation ; the lateral keels diverge in the distal part, but are parallel in the volar part ; the joint surfaces on the sides of the groove, instead of being evenly curved, are incompletely differentiated into two areas corresponding to the two parts of the groove and keels. The articula- tion is not symmetrically placed in relation to the phalanx as a whole, but is arranged 210 EAST AFRICAN CHALICOTHERES Fic. 16. Basal phalanges. a-c, Chalicotherium rusingense. A, R 450.47 (left manus), dorsal, radial and distal views. 8B, unnumbered specimen from site Rs 91 (right manus, reversed), dorsal, radial and distal views. c, unnumbered specimen from site Rs 30 (left manus), dorsal and radial views, and F 2075 (right manus, reversed), distal view. D, Schizotherium priscum, three phalanges from Paris Museum, dorsal and lateral views. rE, C. gvande, three types of basal phalanx of the manus, Paris Museum, in dorsal view. All x 4. EAST AFRICAN CHALICOTHERES ya so that its most dorsal end is displaced towards the ulnar side and its volar end towards the radial side. The lateral surfaces of the phalanx are marked by proximal and distal roughenings for ligaments. The ulnar surface is deeper and flatter than the radial surface, which is more convex and less distinctly marked off from the dorsal surface. There are 8 examples of this bone, ranging in ulnar length from 44-3 to 56-5 mm. (mean 50-4), in radial length from 44-3 to 59 mm. (mean 52:4). Digit III of manus. (Text-fig. 168.) This phalanx averages rather longer and narrower than that of digit II. The metacarpal facet occupies less than half the dorsal surface ; the shaft distal to this facet is relatively longer and narrower, and the distal trochlea is also relatively narrower. The flattened area on the radial edge of the metacarpal facet is less conspicuous. The lobes of the proximal end are more equal, the ulnar lobe exceeding the radial lobe slightly. The intermediate volar tubercles are distal to the metacarpal facet. Seven complete specimens range in ulnar length from 50-2 to 60:2 mm. (mean 57°3). There is also a larger specimen of which the ulnar length must have been about 68 mm. This phalanx is identified as belonging to the third digit because it is the longest and most symmetrical of the three types. Digit IV of manus. (Text-fig. 16c.) -In this type the ulnar lobe of the proximal border is much more prominent than the radial lobe. It is shorter than the phalanx of digit III, and resembles that of digit Il in length. The metacarpal facet occupies rather more than half the dorsal length, and the intermediate volar tubercles are placed opposite the distal part of the facet, as in digit II. This phalanx differs from that of digit II in being more stoutly constructed, in lacking a waist, and in the lack of a conspicuous area of flattening on the radial margin of the metacarpal facet, as well as in the different shape of the proximal border. There are 7 examples of this bone, ranging in ulnar length from 46-5 to 65 mm. (mean 555). Pes. (Text-fig. 17a—c.) The basal phalanges of the pes are smaller than those of the manus (ulnar length of 10 complete specimens, 37—44:3 mm., mean 40-8). The metatarsal facet faces more dorsally, and always occupies more than half of the dorsal surface. Intermediate volar tubercles are usually absent. The phalanges of the three digits of the pes appear to differ in much the same way as those of the manus, but to a smaller degree. Two specimens, one from Rusinga (unnumbered) and one from Songhor (Sgr 32°47), though small (ulnar length 35-7 and 35:3 mm. respectively) are more slender than specimens referred to the pes. They are probably juvenile specimens from the manus. Comparisons. The basal phalanges of the manus of C. grande are from one-third to one-half larger than those of C. rusingense. There are some differences in proportions : the proximal width is relatively greater, and the metacarpal facet occupies a greater proportion of the dorsal surface than in most specimens of EAST AFRICAN CHALICOTHERES N H nN Fic. 17. Basal phalanges of the pes. a-—c, Chalicotherium rusingense. A, R 452.47. B, R 281.51. c, Rusinga, no data. vb, two phalanges of C. grande, Paris Museum. All x 4. C. rusingense. The intermediate volar tubercles are very weakly developed. As in C. rusingense, phalanges of the pes are much smaller than those of the manus. The phalanges of the manus of C. grande fall into three groups, presumably corresponding to the three digits, as in C. rusingense. (Text-fig. 16£.) The differences between them are not so great, however ; for example, phalanges referred to digit III are less obviously elongated, in comparison with those of digits II and IV, than in C. rusingense. Unfortunately, the articulated left manus described and figured by Gervais (1877) and by Holland & Peterson (1913) cannot be used to identify the phalanges of the different digits, for the digits appear to have been reconstituted artificially : the basal phalanx of digit IV belongs more probably to digit III of the right side ; that of digit II seems to belong to digit IV, and that of digit III to digit IT. Two basal phalanges may tentatively be referred to C. pilgrimi : B.M. 12170 and 12172 (Forster-Cooper 1920). From its symmetry, 12172 may be interpreted as from digit III of the manus. It is a little shorter than most specimens of C. rusingense of this type, agreeing in length more with those of digits II and IV. It is broader than any of the phalanges of C. rusingense. Its metacarpal facet occupies rather more than half of the dorsal surface, as in C. grande. Intermediate volar tubercles are well-developed. 12170 appears to belong to digit II, as the metacarpal facet extends more proximally on the radial side than on the ulnar side. It is somewhat larger than in C. rusingense, and its metacarpal facet is more extensive. A phalanx of the pes, referred to C. wetzleri by Viret (1929) is very similar in size and shape to specimens of C. vusingense. It has an intermediate volar tubercle, like some of these. EAST AFRICAN CHALICOTHERES 213 Several basal phalanges of Schizotherium turgaicum are figured by Belyaeva (1953, pl. 3), and five specimens of S. friscum were examined in Paris (Text-fig. 16D). In each species the phalanges show a variety of size and form, but they cannot be allocated to the different digits on the same basis as in Chalicotherium. Filhol’s (1893) drawings of the metapodials of S. priscum indicate that the phalanges of the pes are probably larger than those of the manus, at least on digits III and IV, as in Grangeria (Colbert 1934) : the largest phalanx in Paris is about as wide as the distal end of the third metatarsal. Most specimens resemble in general proportions those of digit III of the manus of C. rusingense, but they may be distinguished by the much more proximal orientation of the metacarpal (metatarsal) facet, which occupies less than one-third of the dorsal surface. In side view the distal trochlear keels are less convex, permitting less rotation of the middle phalanx. Intermediate volar tubercles are frequently present, as in C. rusingense. In a few specimens (one of S. priscum and two of S. turgaicum) the metacarpal facet faces more dorsally and occupies nearly half the dorsal surface, as in C. rusingense. In Moropus and Phyllotillon a similar dorsal orientation of the facet distinguishes the basal phalanx of digit II of the manus, and this may well have been true also of Schizotherium. Middle Phalanges. (Text-fig. 18 ; Table IX.) There are 28 middle phalanges in the collection. They vary in size: the larger ones probably belong to the manus and the smaller ones to the pes. The small Tic. 18. Middle phalanges. Left, Chalicotherium rusingense, R 136.49, in lateral, dorsal and proximal views. Centre, C. vusingense, from site Rs 38. Right, Schizotherium priscum, Paris Museum. All x 4. 214 EAST AFRICAN CHALICOTHERES specimens, making up about half of the series, are more alike in size and shape than the large specimens, and it appears that the middle phalanges of the different digits of the manus are less uniform than those of the pes. The most compressed specimens, with narrow proximal and distal joint surfaces, probably belong to digit III of the manus, and the small phalanges, supposedly from the pes, are compara- tively broad in relation to their length and height. The phalanges are short, laterally compressed bones. The proximal surface is occupied by the articulation for the basal phalanx. This consists of a pair of lateral concave joint surfaces separated by a median keel. The keel runs from a pointed dorsal process, which in many specimens projects beyond the general dorsal surface of the phalanx, to a broader, truncated or slightly emarginate volar process. The volar part of the keel broadens out, the volar parts of the articulatory facets diverg- ing from each other. The dorsal and volar parts of the keel often meet in a more or less distinct angle, corresponding to the angle in the groove of the distal articulation of the basal phalanx. In some specimens the lateral joint surfaces also show signs of a division into a dorsal part, facing proximally, and a volar part, facing more dorsally. The articulation for the basal phalanx is not symmetrically placed in relation to the median plane of the bone, but faces slightly towards the ulnar (fibular) side. The distal end of the phalanx is occupied by the articulation for the ungual phalanx. It has the form of a trochlea, consisting of a median groove and prominent lateral keels, semicircular in lateral view. Dorsally the keels are parallel, but towards the volar side they diverge slightly and the groove between them becomes shallower. The lateral surfaces of the phalanx are flattened, except at the distal margin of the proximal articulation, which projects laterally ; near the middle of each lateral surface is a pit for hgamentary attachment. Comparisons. Except for their larger size, the middle phalanges of C. grande are indistinguishable from those of C. vusingense. In C. grande the phalanges of the pes are smaller and proportionately broader than those of the manus, as was postu- lated for C. rusingense. A middle phalanx of Schizotheriwm priscum in Paris fits the larger basal phalanges. It compares in size with some phalanges of the manus of C. vusingense, but is relatively broader. Some middle phalanges of S. turgaicum described by Belyaeva (1954) are smaller than any of C. rusingense, and relatively broader. In the Paris specimen, the proximal articulation as a whole faces more dorsally than in C. rusingense, the volar part being more extensive and the dorsal part reduced. The distal trochlea faces more ventrally, its keels are more widely separated at their dorsal ends, and the groove between them is much shallower. S. turgaicum appears to possess the same characters. The middle phalanges of Schizotherium thus approach those of Phyllotillon and Moropus. Ungual phalanges. (Text-fig. 19 ; Table IX.) Seventeen ungual phalanges in the collection all have a similar structure, though they differ in size and proportions. The phalanx is approximately triangular in side view, with a curved dorsal border which extends proximally to form a prominent dorsal process (preserved intact only EAST AFRICAN CHALICOTHERES 215 in R1r36.42, and almost complete in Rg22.47). The articulation for the middle phalanx occupies most of the proximal surface and extends along the lower side of the dorsal process. It is curved in an arc of about 100°, and consists of a pair of elongated joint facets separated by a median keel. Below the articulation is a median pit, probably for the flexor ligament, and on either side of this a foramen through which blood vessels passed to the bed of the claw. The terminal part of the phalanx is split by a deep cleft, extending back more than halfway along the volar and dorsal surfaces. The volar surface proximally to the cleft 1s swollen into a rounded boss. The lateral surfaces of the phalanx are flattened and rather rugose. The bone is not quite symmetrical : in relation to the plane of the cleft the dorsal process is directed slightly towards the ulnar (fibular) side. The articulation is also asymmetrically arranged: its dorsal end (on the dorsal process) is ulnar (fibular) to its volar end. The volar boss is displaced a little towards the ulnar (fibular) side. iP SD °< Fic. 19. Ungual phalanges. a, Chalicothevium rusingense, R 136.49, lateral, volar and proximal views. B, C. vusingense, R 922.47, lateral and volar views. c, Three phalanges of Schizotherium priscum, Paris Museum. All x 4. 216 EAST AFRICAN CHALICOTHERES The phalanges fall into two groups, which are probably to be referred to the manus and pes. Those of the pes are broader in proportion to length and height, and are more deeply cut by the claw cleft, the dorsal limit of this being perpendicularly above the volar boss. The distinction is particularly clear in the region of the volar boss : in the manus the horizontal length of the boss, measured from the end of the claw cleft, is about the same as the width of the phalanx (index 80-105) : in the pes it averages a little more than half the width (index, 47-70). The phalanges of the manus are on the average higher than those of the pes (height at volar boss : manus 38-52 mm., pes 32-42 mm.), and also longer in volar length (manus 50-66 mm., pes 41-52 mm.), but the ranges of width are similar. The radius of curvature of the proximal articulation is greater in the manus than in the pes. It has not been possible to allocate phalanges to individual digits with any certainty. It is likely that the largest phalanges belonged to digit II of the manus (e.g. R736.50 and R136.42) and pes (e.g. R497.42), but the preponderance of the claw on this digit of the manus was certainly much less than in Moropus. Comparisons. There are four specimens of ungual phalanges of Schizother1um priscum in Paris, including a large specimen which compares in measurements with Ancylotherium gaudryi Filhol (1880). They are all rather broad, comparing in length/width relations with the phalanges of the pes of C. vusingense. Their height/width indices are, however, much less than in any specimen of C. vusingense, and there are various differences in detail : the dorsal process is less developed ; the proximal articulation has a less prominent median keel, and its curvature is less ; the volar surface is flattened and is separated from the lateral surfaces by sharp edges. An ungual phalanx of S. turgaicum figured by Belyaeva (1954) shows similar characters, but is even lower and broader. The largest specimen of S. priscum approximates in size to the smallest specimens of C. rusingense. The ungual phalanges of C. grande are so similar to those of C. vusingense that it is possible to distinguish them only by size. The Digit as a Whole. (Text-fig. 20.) The second metatarsal MFW1214.55 fits the basal phalanx MFW1213.55 so well that they probably belong to the same individual. A satisfactory, but not perfect fit was found between the third metatarsal from Roi and the basal phalanx F2082. When the basal phalanx was placed so that the dorsal border of its articulating facet coincided with the dorsal border of the facet on the metatarsal, the angle between the dorsal surfaces of the two bones was about 72°. As in other chali- cotheres therefore, the basal phalanx is capable of considerable hyperextension. In this position, no appreciable rotation of the phalanx is possible round its own axis, but lateral sliding would result in some abduction or adduction of the digits. There is no evidence of the notches noticed by Matthew (1929) on the margin of the metacarpal facet of the basal phalanx of digit II in Movopus, and interpreted by him to imply two alternative positions of the digit. It is possible however that his “lateral notch’ corresponds to the flattened area on the radial margin of the facet in C. rusingense, interpreted here as the insertion of an extensor ligament. EAST AFRICAN CHALICOTHERES 217 In maximum flexion the dorsal angle between the phalanx and the metatarsal opens out to no more than about 110°. Rotation beyond this point is prevented by the median keel on the metatarsal, which meets the notch in the proximal border of the phalanx. Presumably the more posterior part of the metatarsal joint surface was occupied by the sesamoid bones, which would remain in contact with the proxi- mal end of the phalanx throughout the movement, being tied to it by a ligament, equivalent to the middle sesamoidal ligament of the horse. Although the distal ends of the metacarpals of C. rusingense are not known, the similarity of the basal phalanges of the manus to those of the pes is such that con- siderations based on the pes may be taken to apply also to the manus. This is certainly true of C. grande. Fic. 20. Reconstructed toes. a-c, Chalicotherium rusingense. A, from the manus, walking position ; 8B, from the pes, walking position ; c, from the manus, clinging position. D, Schizotherium priscum, walking position. E, Plagiolophus annectens, walking position. The joint between the basal and middle phalanges shows a partial differentiation into dorsal and volar portions, representing the areas of greatest pressure during extension and flexion respectively. By fitting phalanges together it may be seen that even in maximum extension the middle phalanx turns down on the basal phalanx ; the median axis of the middle phalanx makes an angle of 10-15° with 218 EAST AFRICAN CHALICOTHERES the long axis of the basal phalanx. The rotation from maximum extension to maximum flexion is 30—40°, so that at maximum flexion the angle between the two phalanges is about 45°. Movement between the middle phalanx and the ungual phalanx is greater : the angle of rotation appears to be about 60°. In maximum extension, the volar surface of the ungual phalanx lies approximately parallel to the median horizontal plane of the middle phalanx, but at a lower level, leaving space beneath the middle phalanx for the flexor tendon. The total amount of bending in the digit itself would be rather more than a right angle, to which must be added 40° or so at the metapodial-phalangeal joint. When walking, the weight must have been received on the proximal ends of the basal phalanges, and presumably also on the sesamoids ; it is likely that a plantar pad existed in this region. The toes themselves did not play any part in supporting the animal. Reconstruction of complete digits shows that the claw probably rested on the ground, the toe being extended but forming an arch. By extreme extension the claw could be raised a little, which would be necessary when the animal was walking over rough ground. This is a much less specialized condition than that found in Ancylotherium pentelicum by Schaub (1943), in which the toes were held back against the dorsal sides of the metacarpals. Maximum hyper- extension of the digit in C. vusingense would seem to be accounted for by pressure of the claw against the ground, aided by the main extensor ligament attached to the dorsal process of the ungual phalanx, but in Ancylother1um Schaub found it necessary to postulate the existence of elastic ligaments. In Schizotherium the position of the proximal articulating surface of the basal phalanx shows that the degree of hyperextension must have been much less than in Chalicotherium ; it is doubtful whether the distal end of the phalanx could be raised above the proximal end. The known middle and ungual phalanges appear to indicate a digit that was straight, inclined downwards at a small angle, and placed so that the flattened volar surface of the ungual phalanx rested on the ground. The elevation of the proximal part of the basal phalanx above the ground seems to have been small, and much of the weight of the animal might well have been sup- ported by a pad at the base of the digit ; nevertheless, the ungual phalanx must have taken some of the weight, and Schizotherium was digitigrade rather than “metacarpograde”’. Gyvangeria again seems to have had a digitigrade foot, though its ungual phalanges are unknown. Some degree of hyperextension at the metapodial-phalangeal joint is found in all perissodactyls, and the chalicotheres seem to have exaggerated this, first becoming digitigrade by modification of the distal end of the metapodial and probably receiving part of the weight on a pad in this region ; then the whole weight was received by the pad, thus permitting a higher degree of specialisation of the claws. In C. rusingense and C. grande, if the basal phalanx is placed so that its most volar surface is horizontal in the transverse direction, the keels for articulation with the middle phalanx are approximately vertical, but the metatarsal is inclined so that its proximal end is medial to its distal end. This is probably the natural EAST AFRICAN CHALICOTHERES 219 position, for the greater length of the third metatarsal is such as to bring its basal phalanx to the same level as that of the second digit. In C. grande the metatarsals increase in length from the second to the fourth. If they were held vertically, only the fourth digit would reach the ground, as Matthew (1929) pointed out. There are objections to believing that Chalicotherium walked on the outside of its foot. The fourth metatarsal and metacarpal are no stouter than the others ; in fact in the manus the second metacarpal is the stoutest. In Movopus and Ancylotherium, again, the second metatarsal is shorter but stouter than the third, though unlike Chalicotherium the fourth is shorter than the third. Schaub con- cluded that in Ancylotheriwm the weight was taken by the radial digits (II and III). It seems highly probable therefore that in Chalicotherium, as well as in Ancylotherium, the second digit touched the ground. This must imply an inclination of the whole manus or pes towards the radial or tibial side. If the feet were orientated in a normal manner, with their dorsal sides facing forwards, the limbs would have to be spread out far laterally. It would be mechanically more efficient to bring the feet beneath the body by rotating them so that the toes pointed somewhat inwards, the shortest metapodial being then the most posterior, and it is suggested that this was the position in which Chalicotherium walked. In maximum hyperextension the basal phalanx does not lie in line with the metapodial, but is turned so that its distal end is more lateral. The plane of rotation at the metapodial-phalangeal joint is inclined, so that the two bones come to lie in the same vertical plane only in maximum flexion. The two interphalangeal joints are not in the same plane ; if the joint between the basal and middle phalanges is considered as vertical, the joint between the middle and ungual phalanges is inclined so that its dorsal side is more lateral. The effect is that when the digit is extended it is curved towards the lateral side, but as it is flexed the claw is moved medially, till it lies parallel to and almost in line with the basal phalanx. Thus in the walking position the toes would point forwards, in spite of the medial rotation of the feet, while when the toe is flexed in order to use the claw, the metapodial, phalanges and claw are nearly in the same plane. The dorsal position of the metacarpal facet on the basal phalanx of digit II of the manus in Phyllotillon, Moropus and probably in Schizotherium, may be related to the enlargement of the claw of that digit. If the claw rested on the ground in walking, the degree of hyperextension that would be required at the metapodial-phalangeal joint would be greater the larger the claw. Schaub suggests that the elevation of the claws in Ancylotherium was an adapta- tion to avoid blunting when walking on hard ground. The lack of this adaptation in Chalicotherium would imply that the animal frequented ground too soft to cause serious damage to the claws. This accords with the view of Abel (1920) that Chalicotherium (“ Macrotherium’’) was an inhabitant of forest, where the ground would be covered with litter, while Ancylotherium (‘ Chalicotherium’’) preferred more open country (savannah). The shortness and stoutness of metacarpal II might be interpreted as an adaptation to clinging to tree-trunks in the manner postulated by Borissiak (1945) : the more €.1—-Z.1 oS—€h Z-I-€0-.1 G.o€—z.bz | ° : * 4 pI L.1-G.1 ¢.'g-9S 9:1 zb-L€ : : * yQsusy ‘sog tr.1-1l.0 9.6z—-S.gI Z-I-89:0 62-91 L.1-v.1 99-S.zS F.I-€.1 9.G€-z-o€ | - > Up Q-I-Z-1 4~8b—9-1€ | g-I-z-I ,6b—-6z G.Z-1-Z L6-S.gL | 9-2-S.z S.Lo-L.LG | - + YQsusl ‘TTT snueyy Q-1-¥.-1 og-1¢ G.1-b-r ¢.6€—-g.z€ | - * Uypln F.Z—-0.2 €6—-bZ ¥.z—-€.2 €9q-L.zS | - * yAsuel ‘T] snueyy ‘sosurreyd jeseq ¥.G QzI Z£.$-1.G ozz—Ler : s + “AT c.G of! 1-S-0.¢ +61-Ser : ; J. ST 6.4 Cir €.b-9.€ to1-obt : : : II uys8ueT ao $3.0 CE CI-1 Lz g-1 69-S9 G.1 LE : ssouyory} ea) €L.0 61 £g-0 Oz €.1 ob—Lb ZI of : : YQPIM ‘peoy AT 2 96-0 Gz L.1-t-1 ¢.bg-Sh V1 ¢.GE : ssouyoryy 5 1g-0 Iz G.1-Z-1 LGS—Sh GS. G.LE : : UPI ‘pesy TIT Ss G.1 of bz z6 €.z gs * AT-ITI WPI pourqurog 2 G.z gs G.€ of * AI-II WPI peurquiog a ‘spedieorjoyy Z Z-I-0O-1 G.bb—-GS.CE Q-1 ob ; : SsoUyoryy = €.I-1.1 6b—-©.1h Q-1 ob : é * 4ysTey x C.1-b.1 ¢.96—¢.64 €.z LG : : * yp ‘proydeos a oryel “Wu oryel “UU orjer “WU orze1 “UU 2 = gz C.€z (¢-Z€ ues) $.g€—S.9€ | (6.bz ueoul) F.gz—F-Ez | ° : * (uu) “yy yQ8ueT wmnasirg *S WNIWSANY “S apuwas “9d asuasUIsnA “Dd ‘(J JO YASUI] 0} VATVLJOI) s}USWOINSeaUT 9} eUOTIOdOIg ‘X AAV 220 N N AFRICAN CHALICOTHERES EAST V.z z6.0-S.0 Z-I-09:0 ‘poysinSunstp useq you oaey sed puv snueu oy} Jo sosuvreyd ‘wnz1ayj02149S Jo sotoads 94} UT x gil gil gS cL 0.S—€.F LS F.1 Z-1-QO-1 0-7-9: vz—-€1 Lv.o-1+.0 y,1€-L.G1 | $S.o-1S.0 LEI gII-zol €€ 6z-Sz Lb—-G.€vF II-oO1 y£I-Z1 8° Z6.0 1-2 Z-I-I-1 Co.1-£0-1 Z-7—-6.-1 €.€ g-I-S.1 6.1 6.z-L.z zg-0-14.0 ¢.1-b.1 £6.0-Sg-0 6-I-g-1 tor C.vE 6L Chih C.ob—G.LE ¢.C9-69 CzI 19-S.S¢ ¢.zl—g9 III—Oo1 ¢.1€—-9z ¢.6S—G.06 o£-1€ €Z—-Lo o-1-S6.0 L.z-S.z Z-I-I-I I-I-Zo-f C.Z-Z-Z F.1 8-1 L.z 1-1—-$g-0 g-1-+.1 g6-0-1g-0 6.I-9-1 Gz—G.€z 69-S.z9 6z—gz Lz—-9z ¢.gS—9S GCE ¢.€r Lo o£—o0z €b-zE 9Qz-61 6F-LE yysuey ‘AT peoy Y}prm * -yysue] ‘TIT TeysIp YFP peoy YPM yysuey “TT YFPIM peurquio+) ‘sTesIe} eI yystey [eIpeu ssouyoryy * YPM ‘snyese1ysy qyPIM “ yysrey ‘Sed . * YypIM > qystoy ‘snueyy ‘sosuvreyd tensuy), 222 EAST AFRICAN CHALICOTHERES lateral digits would extend round the sides of the trunk and their greater length would be advantageous, but most of the pressure would be exerted by the second digit, which would be applied most perpendicularly to the surface of the trunk. In the pes, the short second digit might be associated with a straddling of the legs to give a firmer support, the claws presumably being dug into the ground. Proportionate Sizes of Teeth and Feet. In the absence of associated remains it is possible to make only a very rough estimate of the size of the feet in comparison with the teeth. The method used was to divide measurements of the bones of the feet by the mean length of M,. Where more than one specimen of a bone is known, the largest specimen is compared with the largest specimen of M, and the smallest with the smallest specimen of M,, obviously juvenile specimens being neglected. This was done also for specimens of C. grande from Sansan, for Schizotherium turgaicum and for S. priscum. (Table X.) C. rusingense differs only slightly from C. grande. The greatest difference is its proportionately larger scaphoid, which suggests that the single specimen of this bone comes from an unusually large individual. Otherwise the differences are hardly significant : the head of the fourth metacarpal is smaller, the basal phalanx of the third digit of the manus is longer and narrower, the metatarsals are longer, and the ungual phalanges of the pes are broader. Both species of Chalicotherium differ considerably from Schizotherium. The metacarpals are relatively broader, as is the manus as a whole. In C. grande metacarpals II and III are somewhat shorter than in S. priscum, but metacarpal IV is approximately of the same relative length. The pes of C. grvande is much broader than that of S. priscum. In both species of Chalicotherium the astragalus is proportionately broader and thicker than in S. turgaicum, but of similar relative height, and the metatarsals are much shorter than in S. priscum. The phalanges of the manus are proportionately much larger in all dimensions in Chalicotherium than in Schizotherium. RELATIONSHIPS. In the foregoing description C. vusingense has been compared mainly with C. grande, the best-known member of the Chalicotheriinae, and with species of Schizotherium (S. priscum and S. turgaicum), the most primitive genus of Schizothe- riinae. It shows resemblances to both these forms, but the resemblances to C. grande are of greater taxonomic value, whereas those to Schizotherium are primitive characters inherited from the common ancestor of Chalicotherium and Schizotherium, perhaps in the Lower Oligocene. C. rusingense is Close to C. grande in (1) the upper molar pattern, (2) the loss of the hypoconulid of Ms, (3) the reduction of the scaphoid-magnum contact, (4) characters of the head of metacarpals III and IV, (5) the broad, short metatarsals, those of digits II and III being equal in width, (6) the astragalus, which is reduced in height, and also reduced in thickness in the fibular half, (7) the greater size of the phalanges of the manus as compared with those of the pes, (8) the form of the articulating facets EAST AFRICAN CHALICOTHERES 223 on the phalanges, associated with the development of a metacarpograde stance, and (9) the ungual phalanges, which are narrower and less hoof-like than in Schizotherium. At the same time C. vusingense is more primitive than C. grande in several respects in which it approaches Schizotheriwm : (1) the smaller size, (2) some features of the upper molar pattern—the protocone is frequently connected to the protoconule by a ridge, the ridge on the buccal slope of the paracone is sharper, and the accessory rib in the postfossette is frequently present, (3) the presence in most specimens of a metastylid on the lower molars, (4) the smaller degree of reduction of the premolars, (5) the greater elongation of the anterior part of the jaws, (6) the astragalus, in which the trochlea is higher and the navicular facet is less tilted towards the tibial side. (7) the less thickened metatarsals, (8) the absence of a distinct cuboid facet on metatarsal III, and the presence of two separate facets for articulation with metatar- sal IV, (9) the longer basal phalanges, especially on digit III of the manus, and (10) the better development of intermediate volar tubercles on the basal phalanges. C. pilgrimi agrees with C. rusingense in size, in the presence of a metastylid on the lower molars, and in the characters of metatarsal II, but the upper molars of C. pilgyimt are more primitive in that the paracone and metacone have not receded so far from the buccal edge of the tooth. C. wetzleri agrees with C. rusingense in the metastylid, and perhaps in the elongation of the snout and characters of the phalanges. C. salinum, though much nearer to C. grande, shows some primitive features in the sharpness of the buccal paracone crest and the retention of the protocone-protoconule crest. C. rusingense must therefore be regarded as a persistently primitive form not closely related to other species of the genus. Its interest lies mainly in the light it throws on the evolutionary changes involved in the derivation of the Chalicotheriinae from a form close to Schizotherium. Subfamily SCHIZOTHERIINAE ANCYLOTHERIUM AND RELATED GENERA Ancylotherium pentelicum (Gaudry & Lartet 1856) is a characteristic member of the Pontian fauna of Pikermi, Samos and other localities in S.E. Europe, extending to Maragha in Iran (de Mecquenem 1924). Thenius (1953) pointed out the re- semblances between this species and Metaschizotherium fraast von Koenigswald (1932), from the Upper Miocene of Germany and France, and proposed to include the genus Metaschizotherium in Ancylotherium. Viret (1949) had previously con- sidered M. fraasi to be almost identical with Phyllotillon naricus (Pilgrim 1908, 1910) from the Lower Miocene of Baluchistan. It is also necessary to consider Phyllotillon betpakdalensis (Flerov 1938), from the Upper Oligocene of Kazakhstan, which has been described in great detail by Borissiak (1946). The upper molars of A. pentelicum (Thenius 1953, Wagner 1857), M. fraasi (Fraas 1870, Depéret 1892, von Koenigswald 1932), P. naricus (Pilgrim 1912, Forster Cooper 1920) and P. betpakdalensis (Borissiak 1946) are so much alike that the relationship of the species cannot be doubted. The molars of P. naricus 224 EAST AFRICAN CHALICOTHERES show a considerable range of size (length of M? 40-48 mm.; length of M? 40-49 mm.): the specimens identified as milk-molars by Pilgrim (1912, pl. 12, fig. 3) are small examples of permanent molars. M. fraasi falls within the lower part of the size range of P. naricus. It is doubtful whether M. bavaricum von Koenigswald (1932) is specifically distinct from M. fraast. P. betpakdalensis is very variable and reaches a larger size (length of M? 42-60 mm.). 296 Family FELIDAE Gray . : 5 : E 2 4 302 Subfamily NrmRAVINAE Trouessart . : : ; 6 302 Metailurus africanus (Andrews) . 5 ° : : 304 III. CONCLUSIONS AND THE AGE OF THE FAUNA ; : : ¢ 309 IV. REFERENCES i F ; : ‘ : é , : 312 242 MIOCENE CARNIVORA OF EAST AFRICA SYNOPSIS Eighteen species of carnivores are described from the Tertiary lacustrine tuffs of East Africa, mostly from the Kavirondo Gulf in Kenya and the Napak region in Uganda. Six of the twelve genera (Kelba, Tevatodon, Anasinopa, Leakitherium, Hecubides and Kichechia) and fifteen species are new. Kelba is referred to the Arctocyonidae and Teratodon is placed in a new family of oxyaenoid creodonts. Anasinopa, Metasinopa, Dissopsalis, Metapterodon, Pterodon, Leakitherium and Hyaenodon are all included in the Hyaenodontidae. Hecubides, an amphicyonine canid, Kichechia, a viverrid, and a species of the felid Metailurus are also described. Conclusions on the age of the deposits based on the carnivores, other mammalian elements and radiometric dating converge on Lower Miocene (Burdigalian), with some sites of probable Middle or Upper Miocene (Vindobonian or Pontian) age. I. INTRODUCTION AND ACKOWLEDGMENTS In the first publication of Fossil Mammals of Africa Clark & Leakey (1951) gave an account of the discovery of Miocene fossils in Kenya, listed the sites in the Kavirondo and recorded the fauna as then known. They discussed the probable age of the deposits, based on the known range of mammalian genera recognised in the sequence. Two of the genera used were carnivores, referred to then as Amp/ucyon and Pseudaelurus. They concluded that the most probable age for the fauna was Lower Miocene. Whitworth (1954) in a later publication gave a brief account of the stratigraphy on Rusinga Island, together with a location map of sites. Accounts of the Tertiary geology of the Kavirondo are to be found in Kent (1944), Shackleton (1951), Whitworth (1953, 1961), Bishop & Whyte (1962) and Bishop (1963). The mammal faunas from the Kavirondo pyroclastic deposits are extremely rich. The insectivores, bats, primates, lagomorphs, tubulidentates, hyracoids, anthraco- theres and ruminants have already been monographed. With the publication of detailed studies on the flora by Chesters (1957) and on the mollusca by Verdcourt (1963), the basis exists for profitable studies on the palaeoecology. Fifteen out of the eighteen species of carnivore described below are new, six out of the twelve genera are new and there is one new family. This measure emphasises the novelty of the fauna, whose age equivalent is little known elsewhere on the African continent. Specimens quoted in the text are mostly in the collections of the British Museum (Natural History) and the National Museum of Kenya, Nairobi; their registration numbers are prefaced respectively by the initial letters M. and CMF. Occasional specimens from other institutions are referred to and these named in full in the text. Throughout the systematic description no reference is made to horizon, since the conclusions on stratigraphic age depend on identification of the faunal elements. The subject is dealt with in the final discussion. My thanks are extended to Dr. L. S. B. Leakey who generously offered me the carnivores for detailed study and kindly arranged for me to spend a field season in Kenya visiting the Kavirondo sites and collecting on Mfwanganu Island. Dr. W. Bishop has kindly allowed me to study the Uganda carnivores. MIOCENE CARNIVORA OF EAST AFRICA 243 My thanks are also due to Dr. E. I. White and his predecessor as Keeper, the late Mr. W. N. Edwards, for facilities at the British Museum (Natural History). For advice, criticism and discussion I am indebted to Dr. W. W. Bishop, Professor P. M. Butler, Dr. A. T. Hopwood and Dr. T. Whitworth. Mrs. Shirley Coryndon’s careful and painstaking numbering and cataloguing of the Kavirondo fossils has been quite invaluable. Mr. D. Erasmus is responsible for the drawings, excepting Nos. 2, 6-8, 41-43, 47-49 which are by the author. To Mr. E. W. Seavill and Mr. R. Godwin of Bristol University I am indebted for the photography. Il. SYSTEMATIC DESCRIPTIONS Order CARNIVORA Bowdich Suborder CREODONTA Cope Superfamily ARCTOCYONOIDEA Trouessart Family ARCTOCYONIDAE Murray Diacnosis. Creodonta with teeth primitive, tritubercular, varying towards sectorial or bunodont. No carnassials or specialised shearing teeth. Premolars mostly simple, acute, an inner cusp on P4, sometimes on P? and P,,; canines large, acute, incisors small. Skull moderately long, brain-case small, sagittal and occipital crests strong, occiput narrow and high, tympanic bulla not ossified. [after Matthew 1937]. Remarks. Matthew’s last revision (1937) is still the basis for arctocyonid studies. Simpson (1945) differed from Matthew in only one detail; Matthew distinguished four subfamilies, namely Oxyclaeninae, Chriacinae, Arctocyoninae and Triisodon- tinae, while Simpson does not recognise the Chriacinae as sufficiently distinct to merit subfamilial rank; Simpson’s interpretation is followed here and the Chriacinae are included with the Oxyclaeninae. The remaining two subfamilies are small, com- prising only seven genera. Simpson lists a further four genera which he places in the Arctocyonoidae incertae sedis, and to these may now be added Opsiclaenodon (Butler 1947) and from them we may subtract Paroxyclaenus (Russell & McKenna 1961). Subfamily OXYCLAENINAE Matthew DracGnosis. Molars tritubercular, sectorial or bunodont, hypocone rudimentary or distinct; paraconid distinct; premolars simple save sometimes fourth. REMARKS. The inexactness of the diagnosis emphasises the arbitrary status of the subfamily. It comprises several little known groups and as it stands represents the best solution to the problem of affinity. The Arctocyoninae are distinguishable by their quadrate bunodont molars which have well developed hypocone; the Triisodontinae have round conical cusps on tritubercular molars, the hypocone is weak and the paracone and metacone are progressively connate. Among the incertae sedis genera none displays any features which would suggest affinity with the new genus described below. 244 MIOCENE CARNIVORA OF EAST AFRICA The Oxyclaeninae as defined comprise 15 genera, 14 listed in Simpson (1945) plus Colpoclaenus Patterson & McGrew (1962). All are restricted to the Palaeocene and Lower Eocene of North America save Arctocyonides which occurs in the Upper Palaeocene of Europe. In three genera, Carcinodon, Prothryptacodon, and Spanoxyo- don, no upper dentition is known. Genus KELBA nov. Diacnosis. Upper molars tritubercular; paracone and metacone equal sized, conical but not connate, parastyle prominent; protocone V-shaped with paraconule and metaconule; cingulum continuous all round, carrying distinct hypocone. TYPE SPECIES. Kelba quadeemae sp. nov. Kelba quadeemae gen. et sp. nov. (Pl tie, is Lext-tigs. 2) Diacnosis. This is the only known species and the diagnosis is the same as that for the genus. The name is derived from the Arabic kelb meaning dog and quadeem meaning ancient. HototyPe. M.19087. Isolated right upper molar, probably M?. Locatity. The holotype is from Rusinga Island, Kavirondo Gulf, Kenya. ADDITIONAL MATERIAL. In addition to the holotype there is an isolated left upper molar from Mfwanganu Island, Kavirondo Gulf, Kenya, and another molar from Napak in Uganda. DEscRIPTION. The holotype is an isolated right upper molar, probably M2. The tooth is little worn and all details are clearly seen; the outline is quadrate and slightly asymmetrical due to the development of parastyle and hypocone on opposite corners. The paracone and metacone are equally low cones, separated by a vestigial meso- style; the large protocone is V-shaped, its apex equal in height to the paracone and metacone; the arms of the V carry distinct metaconule and paraconule; the cingulum is continuous round all sides, though not equally developed throughout; it carries a low parastyle and low hypocone and is expanded slightly on the anterior and posterior margins, while being retracted lingually. The tooth is three-rooted, the Fics. 1, 2. Kelba quadeemae gen. et sp. nov. (1) Right M?. Holotype (M.19087), Rusinga Is. (2) Left M3. (M.19095), Napak. x3. MIOCENE CARNIVORA OF EAST AFRICA 245 root below the protocone being much larger than those below the paracone and metacone. The second specimen (CMF .4028) from Mfwanganu is a left upper molar, also probably M2 ,and not so well preserved as the holotype. The teeth are similar in size but show slight differences in detail. In the Mfwanganu molar the cingulum is slightly more expanded anteriorly and posteriorly, and the mesostyle slightly more prominent. An isolated upper molar, (M.19095) from Napak IV in Uganda is also referred to the species. This tooth is probably a left M%, transversely elongated, with width almost twice the antero-posterior length: the protocone is a large V-shaped cusp, the paracone rises to the same height as the protocone and the metacone is very small; a cingulum is developed on either side of the protocone and anterobucally to the paracone. The three roots are situated below the paracone, protocone and metacone; the protocone root is the largest and the metacone root very small. REMARKS. Isolated molars with a basic tribosphenic pattern such as the teeth described above are very difficult to identify with certainty. At this level of organi- sation there is extremely little difference between the teeth of insectivores, primates, tillodonts, creodonts, condylarths, pantodonts and dichobunodonts. Beyond the superficial similarity of these groups, the closest structural resemblances to Kelba are to be found among species usually referred to creodonts or condylarths. The condylarths were examined carefully and the case for the inclusion of Kelba there was found rather less convincing than with the creodonts. Among the Condylarthra, the Phenacodontidae lack a well developed V-shaped protocone, the Periptychidae possess more or less symmetrical hypocone and ectocone, and the Meniscothertidae tend towards lophodont molars. Among the Hyopsodontidae, the Hyopsodontinae are all small insectivorous mammals with sexi-tubercular molars and the Mio- claeninae have tritubercular molars which either lack or have a rudimentary hypo- cone.. Although Kelba is thus placed here among the Arctocyonidae, it will be necessary when more material is available to review this assessment. The differences between the two specimens of M? are so slight, and in view of the proximity of Mfwanganu and Rusinga, there seems little doubt they belong to the same species. The new record extends the range of the arctocyonids to Africa. Most of our knowledge of the group is based on North American finds, and surprisingly the new species shows no close affinity to either the known European or Asian arctocyonids. Affinity is greatest with the Oxyclaeninae, and in particular close similarities can be seen with Metachriacus, Deltatherium, Tricentes and Loxolophus; these genera all occur in the Palaeocene of North America, the first three in the Middle and the last in the Lower Palaeocene. In Deltatherium the molars are more sectorial and less bunodont, the mesostyle is undeveloped and the cingulum extends lingually beyond the protocone. In Metachriacus the upper dentition is not fully known (M? in the only specimen with molars is broken lingually), but enough is available to make a close comparison. Metachriacus molars lack a parastyle, the paracone is slightly larger than the metacone, and in M! and possibly M? the protocone is asymmetrical ; 246 MIOCENE CARNIVORA OF EAST AFRICA otherwise the dentition is similar to the new genus. The closest affinities to Kelba are probably to be found in T7icentes; this genus with Loxolophus has few features which vary from Kelba. Both Palaeocene genera have bunodont molars, the para- cone is rather larger than the metacone, and a small parastyle is present. In Loxolophus the cingulum extends more lingually than in Tricentes. The four American genera mentioned above differ from each other as greatly as they differ from Kelba. No clear ancestry for Kelba is obvious among the Palaeocene forms; on the whole Tvicentes is perhaps closer than any of the others. Kelba would appear to represent a late survivor of a primitive stock, as probably are Didymoconus and Ardynictis, two genera of arctocyonids described by Matthew & Granger (1924, 1925) from the Lower Oligocene of Mongolia. Measurements (in mm.) for Kelba quadeemae M2 M? M3 M.19087 CMF.4028 M.19095 Holotype Ant-post. 10°2 9°6 71 trs. 12°3 T7222 Teer Superfamily OXYAENOIDEA Osborn 1910 Family TERATODONTIDAE nov. Diacnosis. M3 present; M? transverse; M2 main carnassial, M3 less functional as carnassial. Premolars large, bunodont, tubercular with thick enamel: P4 larger than M!. Lower molars with small talonid and metaconid present; M, larger than M,. Jaw relatively short. REMARKS. The superfamily Oxyaenoidea Osborn I910 is synonymous with Pseudocreodi Matthew 1909, Osborn’s name having validity. The superfamily com- prises only two families, Oxyaenidae and Hyaenodontidae, to which is here added a third. In the Oxyaenidae M3 are absent in all known genera, M?is transverse and M3 are the functional carnassials: the premolars may enlarge and P® has progressively developed protocone: the jaw is short and the symphysis robust. The Hyaenodonti- dae is a much less compact family: in all genera P? lacks a distinct protocone. Within the Hyaenodontidae are four subfamilies: the Limnocyoninae and Machaeroidinae lack M3 and M®? is always transverse, (Prolimnocyon is the one exception—M 7? is transverse and a vestigial M? is present): Hyaenodontinae lacks metaconid on lower molars: Proviverrivinae retains M? and metaconid on lower molars, the skull is narrow and jaws long. Gazin (1946) proposed that the Limnocyoninae and Mach- aeroidinae be given family recognition as the Limnocyonidae. As our concern here is with the Hyaenodontidae sensu stricto we shall not pursue the wrangle. Clearly the status of the Creodonta as we know them today is measured; arctocyonids and hyaenodontids are likely to be among the first victims. MIOCENE CARNIVORA OF EAST AFRICA 247 Most of the known characters of the Teratodontidae can be found within the subfamilies of the Oxyaenidae and Hyaenodontidae, but the combination is unique to the Teratodontidae. In particular the extraordinary premolars mark out the teratodontids from all oxyaenids and hyaenodontids. The new family is established to accommodate two species of Tevatodon gen. nov. described below and in addition I would transfer to the family Quercytherium tenebrosum Filhol from the Upper Eocene—Middle Oligocene of France. Genus TERATODON nov. Diacnosis. Teratodontid of medium size, jaws relatively short. M1! and M? metacone slightly larger than and connate with paracone; elongate metastyle; M? slightly larger than M!. P4 bitubercular and larger than M1; protocone almost as large as paracone. Lower molars with well developed metaconid, trigonid cusps high, talonid small, paraconid-protoconid shear very oblique. Pg, large with low single cusp. Generic name derived from the Greek fevas, a monster or strange creature and odous a tooth. TYPE SpEcIES. Teratodon spekez sp. nov. SPECIES AND DISTRIBUTION. The type species is known from Koru and Songhor. There is in addition a second species from Songhor described below. Teratodon spekei gen. et sp. nov. (Pl. 1, figs. 2, 3; Text-figs. 3-11) DiaGnosis. Species about size of Vulpes vulpes; metastyle on M? elongate transversely but not extending beyond the level of the parastyle. The trivial name is in memory of Captain John Hanning Speke of Dowlish Wake, Somerset, who in 1859 discovered Lake Victoria. HorotypPe. M.14307. Left maxilla with P4, M!+? and alveolus of M3; from Koru near Kavirondo Gulf, Lake Victoria, Kenya. ADDITIONAL MATERIAL. In addition to the holotype, the type locality has yielded the following additional specimens: M.14215. Right maxilla with P*and Mand broken root of M!. (Paratype) M.14310. Anterior maxillae with canine and P? on both sides and alveoli of P1. (Paratype). [These two specimens probably belong to the same individual as the holotype.] M.14216. Left mandibular fragment with M, and Mg. M.14308. Right mandibular fragment, symphysial region with two very worn premolars and part of a third. The following two specimens from the type locality are referred to the species: M.14204. Mandible fragment with very worn premolar and root of another premolar. M.14225. Right lower canine. 248 MIOCENE CARNIVORA OF EAST AFRICA Fics. 3-5. Tevratodon spekei gen. et sp. nov. Left maxilla with P4, M! 2. (3) Occlusal aspect. (4) Lateral aspect. (5) Medial aspect. Holotype (M.14307), Koru. x2. Songhor, a site about 10 miles NW of Koru, has yielded the following mandibular remains: CMF.4039. Right mandible with DP.,, P,, M,,9, and unerupted Msg. CMF.4040. Left mandible with M,,, and unerupted Mg. [These two specimens probably belong to the same individual. | CMF.4041. Left mandibular fragment with P,, 9. DEscRIPTION. The holotype comprises a posterior maxillary fragment with the anterior part of the jugal arch, P*, M1 in place and the alveolus of M3: attached to the specimen is a piece of frontal bone showing the anterior line of the temporal muscle origin. The strength of the jugal arch and the depth of the temporal insertion (together with the robustness of the mandible) suggest a strongly built skull with massive temporal muscles. The right maxilla, M.14215, probably belongs to the same individual as the holotype; they come from the same site, are the same size and have the same degree of wear on the teeth. The anterior fragment of maxilla M.14310 may also belong to this individual. Mtand Mare closely similar, M? being slightly larger than M!. On the paratype little more than the roots of M+ remain, while on the holotype the crown of M? is rather worn; but from what can be seen, it does not differ structurally from M?. M?has low connate paracone and metacone, small parastyle and metastyle which is elongated transversely; the metacone is larger than the paracone; the protocone is V-shaped and attenuated transversely. The alveolus for M? extends transversely across the palate as far as that of M2, and this evidence taken with the fully developed M., indicates a sizeable transverse tooth. P? is startlingly different from the molars; it is present in both holotype and paratype. The tooth is massive and robust, with two thick roots and the crown sur- face area is slightly greater than that of M!. The enamel is thick and although in both specimens the cusps are worn flat, the bases of two cusps are recognizable; these must have been low tubercular cusps. P? is unknown though presumably present. MIOCENE CARNIVORA OF EAST AFRICA 249 P? is an extremely stout tooth with two strong roots; the crown forms a low symmetrical ovate cone, with slight ridge near the base on the antero-lateral side and suggestion of a cingulum on the posterior margin. P! was very much reduced and double rooted; it would appear that the size relationships between P! and P? were similar to those between P, and P, described below. The upper canine is well but not strongly developed; the antero-posterior diameter at the base of the crown is much less than that on P?. The canine is flattened laterally, more so on the inner side with anterior and posterior keels. The tip is not preserved and no trace of a saw-edge is evident on the proximal parts of the tooth. The premaxillae are missing though fragments of the nasals survive. Fics. 6-8. Tevatodon spekei. (6) Maxillae with canines and P2?, occlusal and lateral aspects. (M.14310), Koru. (7) Right mandible with P, ,, occlusal and lateralaspects. (M.14308), Koru. (8) Left mandible with P, ,, occlusal and lateral aspects. (CMF.4041), Songhor. AES 5: Material of the lower dentition is more abundant and between the specimens almost a complete dentition is known. Specimens CMF.4039, 4040 and 4041 probably all belong to the same individual, while specimens M.14308 and M.14216 probably belong to another individual which could well be the same as the holotype, M.14215 and M.14310. The Songhor individual is immature with the permanent dentition still erupting while the Koru animal is a fully mature adult with well worn dentition. Of the three molars, M, is the smallest and M, the largest, and all three are 250 MIOCENE CARNIVORA OF EAST AFRICA basically similar. M, has a high pointed trigonid, the paraconid-protoconid cusps are about equally high and form an oblique shearing blade; the metaconid is small and a stud is present on the cingulum below the shearing edge of the paraconid: the talonid is slightly shorter than the trigonid and it faces steeply downward and inward from a high buccal wall to the low lingual edge. In the adult specimen (M.14216) M, trigonid cusps are worn flat from grinding. M, on the two juvenile jaws (CMF. 4039 and 4040) is similar to Mg, but smaller; the trigonid is worn flat making in- effective the shearing paraconid-protoconid edge; the paraconid-protoconid edge is less obliquely inclined. Mg, fully erupted on M.14216, has a trigonid which is rather bigger than that of M, although the talonid is no larger; the high paraconid- protocnid shearing blade shows it to be the main carnassial, occluding with M2. In the juvenile jaws the unerupted M, can be clearly seen in X-ray photographs and in both specimens the bone is just opened so that the tip of the protoconid is visible. On specimen M.14216 only M,,. are preserved, together with the posterior root of M,: the jaw is remarkably robust for the size of the teeth, being thick in proportion to its depth. Mandible CMF.4040 is not preserved anterior to M,. Specimen CMF.4039 is preserved up to and including the symphysis: the coronoid process and condyloid articulation are missing, as is the tip of the angular process. The body of the mandible is slender, relatively short and the symphysis large: the symphysial area is rough and D-shaped. A mental foramen is present on the lateral border behind P,. The two teeth preserved anterior to the molars are taken to be fully erupted and worn DP, and P, erupting: P, would have formed in jaw cavity between DP. and M,. DP gis as large as M,: it has two roots, but the crown is worn flat almost to the base and no details can be discerned. P, is just erupting above the symphysis and is much the largest tooth in the series: it is oval with two roots, with thick enamel on crown which forms a massive but very low pointed cusp. Between the anterior root of P, and the symphysis is a small cavity, probably for the canine root tip. In the large cavity within the mandible between M, and DP, was an undeveloped tooth; only a single cusp tip is calcified and nothing further can be traced; it is presumed that this tooth is the developing P,. Specimen CMF.4041 preserves P,,. around the symphysial region; P, is identical to that in CMF.4039 and is also just erupting through the bone. P, is preserved on the left side, a small laterally compressed two rooted tooth with low anteriorly placed cusp. Specimen M.14308 from the type locality has two very worn premolars preserved, taken to be P, and Ps, on basis of size and position in mandible relative to symphysis and mental foramina; the anterior half of P, also survives and is again worn almost to the base of the crown. This robust mandible fragment could belong to the same individual as M.14216 and the holotype. M.14204 has very worn premolar preserved, probably P., and behind it the roots of P,. M.14225 is an isolated canine from Koru, ovate and more compressed on the inner side, curved more strongly than the upper described; it is of the size and shape expected of a lower canine of T. spekev. REMARKS. The premolars are the outstanding feature of Tevatodon. No other MIOCENE CARNIVORA OF EAST AFRICA 251 Fics. 9-11. Tevatodon spekei. Right mandible with DP, P,, M, 4. (9) Occlusal aspect. (10) Lateral aspect. (11) Medial aspect. (CMF .4039), Songhor. x1°5. carnivore or carnivorous mammal known to me has quite such an extraordinary combination of premolars and molars. Their highly exceptional character, seen in the maxillae and mandibles from Koru and the mandibles from Songhor makes the linkage between these two sites (about Io miles apart) plausible. Associated with the bunodont premolars is the robust build of the jaws and their heavy musculature. The functioning of this dentition presents occusal and mechanical problems. M# are clearly the main carnassial teeth, both on basis of structure and wear. The blades are high, but their obliqueness may be regarded as a primitive feature. Anterior to these come the grinding premolars: the wear surfaces on the upper premolars are almost flat and directed slightly dorsally and anteriorly; there is little definite trace of direction of movement on the surface, but from faint striations it would appear more likely to have been transverse than longitudinal. Unfortunately 252 MIOCENE CARNIVORA OF EAST AFRICA the condyloid process is unknown and hence we are ignorant of any articular modifications. It is difficult, from a purely mechanical viewpoint, to see how a typical carnivore can perform efficiently the dual function of shearing at the back and grinding anterior to this: the molars seem likely to impede grinding movements. Both processes require powerful movements, best achieved with the occusal surfaces near the fulcrum and a long moment arm from the fulcrum to the line of action of the muscles (temporal and masseter mainly). The flattened surface of M! and the trace of wear on the cusps of M? suggest the molars are inevitably occluded during pre- molar grinding. The premolars could only grind if their height carried them above the level of the molars; height alone would be useless and must be accompanied by thickened enamel or more complex tooth structure if it is not to be rapidly lost. Teratodon premolars are large and have thick enamel. The only fossil material with which comparison can usefully be made is Quercy- therium from the Upper Eocene of Phosphorites du Quercy and Gard in France. Here too can be seen the enlarged premolars, the second being the largest in the series. Piveteau (1961) has suggested that the specialized dentition of Quercytherium can be compared with that of hyaenids. The anatomical comparison does not stand up to detailed examination but functionally there is parallelism in that both combine crushing and shearing teeth. In my osteological collection I have a dog skull from Ounianga Kebir, an inhabited oasis in the midst of the Sahara desert. This skull illustrates what can happen to a typical carnivore when forced to feed very largely on vegetable material. The people of the oasis live largely on a diet of rice and dates, with occasional goat and chicken on festive occasions. The dogs subsist largely on dates, and these mostly rejects too sandy for human consumption. In the desert without the utmost care, sand covers all food near ground level: the combination of sand and date stones soon wears down teeth. On this particular dog skull, all the premolars and molars are worn to flat surfaces, the M? no less than P?: only the outer rim of P4 remains rather above the levelled dentition. Further, these teeth show well marked transverse striations indicating lateral grinding movement. The feeding habits of Tevatodon remain something of a mystery. The premolar structure shows a departure from normal and the feeding habits appear to make use of these modifications. The dentition as a whole seems unbalanced, and far from being a satisfactory compromise, it seems to get the worst of both: the carnassials cannot function efficiently because of the large premolars and the premolars cannot grind efficiently because of their position and the presence of shearing molars behind. The jaw movements were probably not dissimilar to those of the Ounianga dog. Instead of desert sand and date stones, we can imagine volcanic dust and stones of the savannah fruits (well fossilized on Rusinga, see Chesters 1957). A small stud at the base of the paraconid on the lower molars has been described. It cannot occlude with anything as it is too low, yet it is very well developed on My and Mg, though less so on M,. I suggest the stud acted as a guide to erupting teeth, keeping them in true alignment: if the carnassials erupt with lateral displacement, the blades will not shear: so long as the posterior edge of the proceeding molar is MIOCENE CARNIVORA OF EAST AFRICA 253 medial to the stud, then the shear should function. Tevatodon was probably quite vulnerable to displaced eruptions owing to the shortened jaw. ArFinities. The classification of Tevatodon presents difficulties: on the basis of the molars alone, it is clearly to be numbered among the Proviverrinae; but the premolar specialities and associated shortening of the jaws rule this out. Enlarged premolars are not uncommon in the Oxyaenidae and are to be found in some of the carnivorous marsupials, Borhyaenidae: the differences in molar structure, however, rule out possibility of affinity with these families and the expanded premolars reflect homeomorphy. The molar teeth of Tevatodon are comparable with those of Anasinopa described below and with Sinopa. Anasinopa is much larger and the shear on the carnassials less oblique; Sznopa is intermediate between the other two. M2 M, Angle between shear and Angle between shear and paracone-metacone line paraconid-metaconid line Teratodon 80° 50° Sinopa 60° 45° Anasinopa 50° 40° The Teratodontids could be regarded as an early offshoot of the Oxyaenoidea, close to the Proviverrinae, retaining M? and molars with very oblique shear, while specializing in the development of crushing premolars. TABLE I Measurements (in mm.) for Tevatodon spekei Cc ip pe Mt M2? P, 1B P, M, M, Mg, M.14310 ap gO 12°9 trs 5°7 83 M.14307 a—p 6:7 6:8 6:8 Holotype trs 9°5 9°5 10°7 M.14215 a—p 6°5 6°7 trs 9°8 II‘O M.14216 a-—p 73} 9°4 trs 5°0 6°3 M.14308 a-—p 12°76 108 trs 7°4 8:8 CMF .4039 a-—p II*5* gop — 7°4 trs 6°5 4°9 3°8 5°1 CMF .4040 a—p 6:0 7°3 trs 3°8 4°9 CMF .4041 a-—p (G40) 11 trs Py) 7:0 *Tooth erupting; measurement approximate. D, Deciduous premolar, probably DP3 Teratodon enigmae sp. nov. (Pl. x figs. 4, 5; Text-figs. 12-18) Diacnosis. Differs from the type species in having shorter and more robust jaws, 254 MIOCENE CARNIVORA OF EAST AFRICA and very heavy premolars. On M? the metastyle extends laterally beyond the level of the parastyle. Both upper and lower canines are large and the snout is blunt. HorotypPeE. M.19088. Facial region with dentition fairly complete behind the incisors. From Songhor, near Kavirondo Gulf, Kenya. ADDITIONAL MATERIAL. A left mandible, M.19089, from the same site and possibly belonging to the same individual as the holotype, is referred to the species. DESCRIPTION. The holotype comprises a reasonably complete facial region with most of the dentition posterior to the incisors. Maxillae, nasal and palatine bones ‘ \ it it Fics. 12-14. Tevatodon enigmae sp. nov. Maxillary region. (12) Occlusal aspect. (13) Right lateral aspect. (14) Left lateral aspect. Holotype. (M.19088a), Songhor. ae MIOCENE CARNIVORA OF EAST AFRICA 255 are preserved and the sutures visible. The complete nasals are entirely horizontal, elongate and with parallel sides: the anterior edge is transverse with a very short lateral arm adjoining the premaxilla. The premaxillae are broken anteriorly, but the root of one incisor remains on the right side; posteriorly they extend back and overlap the nasals for some 15 mm. Between the large root of the upper canine and the premaxillary border of the maxilla is a deep groove, the lower part of which may have accommodated the lower canine. There are in addition fragments of right frontal and parietals, the latter showing high sagittal crest. The upper molars are similar to Tevatodon spekei but beyond this close similarity ends. The jaw is short and very much constricted in the premolar region. The mandible and beginning of the jugal arch are heavily built. Other than fractures due to fossilization, the bone is in good condition and shows no sign of fracture during life or any other abnormality. In contrast to this the dentition is bizarre. The beast is presumed to have had three molars, of which M? and Mare well preserved on both sides. M?is distinguishable from that of Tevatodon spekei only by the more elongate metastyle, which is extended well beyond the level of the parastyle. Mis a trans- verse molar of the size and proportions expected of Tevatodon spekei: it has V-shaped protocone, connate paracone and metacone, the paracone slightly larger than the metacone, elongated parastyle extending to meet the metastyle of M2. In the short gap between the canine and M? is crowded a grotesque array of ‘premolars’, which almost defy description. These ‘teeth’, as will be seen from the illustrations, cannot be numbered P!4; they possess massive roots, they are not symmetrical on left and right, the crowns are worn into a longitudinal concave arc and are without trace of a cingulum, the largest tooth is midway between the canine and M2 and is so broad that a palatal gap of only 4 mm. is left. Fic. 15. Teratodon enigmae. Maxillary region, anterior aspect. Holotype (M.19088a), Songhor. X1I°‘5. The robust mandible has a large symphysis and two mental foramina, the larger and more posterior under ?P.. The five alveoli at the back of the mandible presum- ably are for the three molars, though it is far from clear to see how three teeth, each of which could be expected to have two roots, can fit into five alveoli. The root of the canine is visible beside the symphysis and on its outer edge the root tip of ? Pj. 256 MIOCENE CARNIVORA OF EAST AFRICA Between this root tip and the molar alveoli are four tooth stumps, so worn that no characters remain: a little enamel is left on the posterior edge of the last of these four teeth: the wear surface forms a longitudinal concave arc as on the upper dentition. Fics. 16-18. Tevatodon enigmae. Left mandible. (16) Occlusal aspect. (17) Medial aspect. (18) Lateral aspect. (M.19089), Songhor. X1I‘5. MIOCENE CARNIVORA OF EAST AFRICA 257 RemARKS. Although the skull bones are in places broken, the joins are clear and there is no doubt the pieces have been assembled correctly. Hence the possibility that pieces of several individuals, or even several species, being assembled together is ruled out. It seems inconceivable that this individual represents the norm of the species. The predominance of transverse striations on the upper and lower ‘premolars’, together with the shape of the facets, suggests that these are genuine occlusion facets and not weathering surfaces. It must follow that the specimen is that of an abnormal individual. If the abnormality was caused by damage to the jaw and or tooth germs, then more asymmetry would be expected, and some sign of bone repair inevitable. No pathological cause is known which would produce such bizarre patterns. This leaves only a congenital cause for the abnormality and the individual must be a mutant. The norm of the species probably represents something quite different from Teratodon speket. The parallel sided nasomaxillary region, expanding rapidly from the springing of the jugals, the flat topped and blunt nosed snout with large maxillo- turbinal cavity, recall proportions seen in Enhydra, the sea-otter. The cheek denti- tion in Enhydra is relatively large and the enamel thick—adaptations to shell-crush- ing. It is tempting to think of T. enigmae as a shell-crushing aquatic form, breaking molluscs loose with its strong canines and crushing them with the heavy premolars. Without insisting that the above reasoning is water-tight and that no other solutions are possible, I submit that on the available evidence it seems the most plausible explanation. It would greatly help to have more material: the population may represent one of those interesting short periods of genetical instability so rarely preserved, when many new prototypes are appearing and disappearing in the process of establishing a few new strains. Measurements (in mm.) on Tevatodon enigmae (M.19088a) : C M2 M3 Right side a—p Te 2 7°0 53 trs TeBe I1°6 10°4 Left side a—p 13°6 71 5°2 trs 9°2 II*2 I0°4 *Measurement taken on root. Family HYAENODONTIDAE Leidy DiacGnosis. Creodonta with upper molars either three or reduced to two; two front upper molars specialised as carnassial teeth either tuberculo-sectorial or completely sectorial; last upper molar, when present, transversely extended; all the lower molars specialised as carnassial teeth ; P+ two-rooted, except in some specialised genera; primitive forms with long and slender skulls; tail long and heavy; later forms with more robust skull, claws blunt; cursorial adaptations to a varying extent. [after Pilgrim 1932]. REMARKS. The diagnosis omits details of post-cranial characters, to be found in the diagnoses of Matthew (1909) and Denison (1938). The Hyaenodontidae together with the Oxyaenidae make up the superfamily Oxyaenoidea. (=Pseudocredodi of Matthew 1909 and Denison 1938). Of the four subfamilies of Hyaenodontidae, 258 MIOCENE CARNIVORA OF_EAST AFRICA Denison (1938) grouped together the Limnocyoninae and Machaeroidinae as short broad skulled types, and the Proviverrinae and Hyaenodontinae as long faced narrow skulled types, while Gazin (1946) separated the Limnocyoninae and Machaeroidinae in a new family, the Limnocyonidae. Subfamily PROVIVERRINAE Matthew DiaGnosis. Hyaenodontidae with narrow skull and long face; M3; molars tritubercular above, tuberculo-sectorial below; metaconids present on lower molars; carnassial specialization less advanced [After Matthew 1909]. RemMARKS. The skull and facial characters distinguish the subfamily from the Limnocyoninae and Machaeroidinae: the tooth characters distinguish it from the Hyaenodontinae. Of the 12 genera comprising the sub-family, 7 of these are listed by Simpson (1945) and four have been added since; Stovall (1948) added/schnognathus and three are due to Matthes (1952), Prodissopsalis, Leonhardtina and Getselotherium. The Eocene of North America has yielded Sinopa and Tritemnodon, and from the European Eocene come Prorhyzaena, Proviverra, Paracynohyaenodon, Prodissopsalis, Leonhardtina and Geiselotherium. (Simpson also includes in his list Cynohyaenodon and Galethylax, both of which I consider synonymous with Proviverra). The Oligocene has yielded Jschnognathus in North America and Metasinopa in Egypt. The only Miocene form is Dissopsalis from India. To these is now added a further genus from East Africa. CoOMMENT.—Matthes (1952) described a new creodont fauna from the Middle Eocene lignite beds of Geiseltal; the six new proviverrine species described are placed by Matthes in four new genera. Unfortunately the photographic plates have reproduced very poorly and there are no diagrams of the dentitions, hence interpre- tation is seriously impeded. None of the new species or genera is very close to the new African genus described below, but the taxonomy in the paper calls for some comment. I consider two of the species, Imperatoria gallwitzt and I. hageni to be identical. Both are known only from mandibles and lower dentitions: they have identical morphological characters and the size differences are so slight that they are well within the range of individual variation, as seen in the following figures (from Matthes 1952): I. gallwitz (mm.) I. hageni M)-3 38 40 P,-M; 80 85 Length M, Il 12 > M, 13 14 M3 14 14 Imperatoria is known only from mandibles and lower dentitions and Prodissopsalis is known only from skulls and maxillary dentitions. Both occur in the same beds at the same sites; both are the same size. The dentition of Imperatoria corresponds exactly in composition, pattern and size with that which could be envisaged for the MIOCENE CARNIVORA OF EAST AFRICA 259 lower dentition of Prvodissopsalis, as shown in the following figures where comparison is made with the upper and lower dentition of the closely related genus Sznopa. Sinopa grangervi a M1-8 22-7 mm. UGE’ 8. c P1_M? 56:8 EOC (Matthew 1906) bi, .2o7 mm) ib) 5 dP,-M,601 d Ave Prodissopsalis eocaenicus a M!-3 32 mm c P1_M? 81 mm. (Matthes 1952) 10a IOC : — = 815 —— = 9°76 Imperatoria b d gallwitz b M,-, 38 mm. d P,-M, 83 mm. (Matthes 1952) I therefore consider Imperatoria a nomen nudum and all material previously referred to it to be synonymous with Prodissopsalis eocaenicus. Genus ANASINOPA nov. Dracnosis. Proviverrine with dental formula 2:43; Skull elongate and jaws slender: P, two-rooted; lower premolars compressed, crowded posteriorly, length slightly greater than height; P, with a distinct talonid; P*4 tubercular, parastyle smaller than metacone: M1!*? tritubercular, triangular, metacone and paracone close together but not connate, metastyle shearing, metaconule and paraconule present; protocone V-chaped; M3 transverse; M,_, tuberculo-sectorial, metaconid present, Mg, largest and M, smallest, protoconid and paraconid subequal, their height approximately equal to trigonid length, metaconid much smaller, talonid basined; M,, talonid length slightly less than trigonid, M., talonid much reduced. TyPE SPECIES. Anasinopa leakeyi sp. nov. The only species. Anasinopa leakeyi gen. et sp. nov. (Eat, tes Ong bi 2= Wext-igs, 19-22) DiaGnosis. The generic characters form the basis of the diagnosis. Species about the size of the European wolf (Canis lupus). C,-—M,=92 mm; P, — M, = 84 mm; M, — M, = 41 mm; M! — Mest. = 34 mm. HoLotyPe. Five pieces comprising maxillae and mandibles of one individual. M.19g081 a _ Left maxilla with P4, M! and alveoli of P?+®. - b Right ,, » M?*2 and alveolus of M3. f c Right mandible with C, P,,, M,-3. »” a Left yy ” Moi: d a”) ” ” C and P,. 260 MIOCENE CARNIVORA OF EAST AFRICA Locaity. Rusinga Island, Lake Victoria, Kenya. ADDITIONAL MATERIAL. From Rusinga Island, Site 106 :— CMF.4044 Right M2 CMF.4045 Right M, From Rusinga Island, unsited:— CMF.4018 Right mandible fragment with M,_, CMF.4019 Right M2 CMF.4020 Right M1 CMF.4047 Left P4 CMF.4048 Right M? CMF.4049 Left M, (trigonid only) CMF.4050 Left P, CMF.4051 Left M, (trigonid only) CMF.4052 Left P, CMF .4054 Right M, CMF.4055 Right mandible fragment with P, , CMF.4056 Right M, (trigonid only) CMF.4058 Left M2 (broken) From Karungu:— CMF.4046 Left P4 From Maboko Island :— CMF.4043 Left M, From Mfwanganu Island :— CMF.4053 Right mandible fragment with C, P, roots, P, and part of P, CMF .4057 Right M, Site unknown :— CMF.4059 Right P, DEscriIPTION. None of the skull is preserved beyond that surrounding the teeth. The alveoli of the two-rooted P? and P® are preserved and the infra-orbital foramen is present above the posterior alveolus of P?. P*is tubercular; the transverse width is approximately equal to the antero-posterior length; paracone is conical with small parastyle anteriorly and metacone posteriorly ; protocone well developed and slightly anterior to paracone, its posterior border continuous with metacone base, and anterior border constricted and separate from parastyle; order of cusp size commencing with the largest is paracone—protocone—metacone—parastyle; deep valley between paracone and protocone; metacone more or less connate with paracone. M! tuberculo- sectorial; transverse width slightly greater than antero-posterior length; paracone and metacone tubercular, metacone slightly larger than paracone, both cones close together but not fully connate; small parastyle; metastyle trenchant, connate with metacone and with weak oblique shear; external cingulum; large lunate protocone with small paraconule and metaconule on the arms. M2? structurally similar to M+ but slightly larger and metastyle more sectorial. M* unknown; small transverse two-rooted tooth. MIOCENE CARNIVORA OF EAST AFRICA 261 None of the lower incisors is preserved, but the narrow symphysial region indicates three closely packed teeth. Only the base of the lower canine is preserved and this indicates a slender and moderate sized tooth. P, follows immediately 20 <2) \ y if HN ZY) \\ A \ wl ( PAY >I Ab 4 eel Oe 1G: p \ . \ WD gx S ( A Fics. 19-22. Anasinopa leakeyi gen. et sp. noy. (19) Reconstruction of right maxilla with P4, M? 2; based on M.19081a,b; occlusal aspect. (20) Right mandible with C, P,-4, M;-,; occlusal aspect. (M.19081c). (21) same as Fig. 20; lateral aspect. Fig. 20, medial aspect. (22) same as Holotype (M.19081r), Rusinga Is. x1. 262 MIOCENE CARNIVORA OF EAST AFRICA behind the canine without any diastema; it is a small two-rooted tooth with a posterior cingulum. P, is much larger and P gis slightly larger again, with a posterior cusplet. P,is the largest premolar; to the main cusp is joined posteriorly a short talonid with high external cusp and low internal ridge. M, trigonid with metaconid considerably smaller than subequal paraconid and protoconid, the latter two cusps apparently sectorial with weak oblique shear, but trigonid cusps worn to stumps; talonid slightly shorter than trigonid, shallow basin with high buccal and low lingual bordering ridges. M, structurally similar to M,, but rather larger. M, with trigonid much larger than M,; metaconid small, paraconid and protoconid with strongly developed oblique shear; talonid very small, less than half length of trigonid, with buccal bordering ridge and also an oblique ridge crossing inwardly over the sloping basin. The mandible is long and slender. The symphysis is three times as long as it is high, extending as far back as P,. The condyle is rounded and elongated transversely ; the slender curved angular process for the masseter reaches back to the level of the condyle. The coronoid with marked anterior ridge arises immediately behind M, and sweeps high above the condyle in a typically carnivore fashion. The anterior mental foramen is below a point between P, and P,; the posterior foramen is below Pe REMARKS. It is regrettable that so many of the genera in the sub-family are poorly known, often only from fragments. The presence of metaconids on the lower molars however distinguishes them from the hyaenodontines. Proviverra is a very small form with connate paracone and metacone on the upper molars; premolars are short and high, and lower molars have high metaconid almost equalling protoconid. In Paracynohyaenodon the paraconid is very low, smaller than the metaconid. IJschnognathus is known only from a fragmentary mandibular symphysis and its relationship to the proviverrines must remain in doubt. Metasinopa is very small, lacks P, and has very reduced metaconids on lower molars. In Dis- sopsalis the parastyle is lacking on P4 and vestigial on the upper molars; metaconule and paraconule are not present on M!*?. The diagnosis of Geiselotherium given by Matthes (1952) comprises nine negative statements about the genus, making it almost impossible to recognize: size alone excludes it from consideration with Anasinopa leakeyi, Other characters could be listed which differentiate these genera from Anasinopa, but those given are sufficient to establish the distinction. On the basis of molar tooth structures, the closest similarities to Anasinopa are to be found in Sinopa and Tritemnodon. Tritemnodon lacks a parastyle on P4; the upper molars have connate paracone and metacone, and lack paraconule and metaconule. Sinopa species have large parastyle on P4, larger than in Anasinopa; M!*? have widely separate paracone and metacone: in the lower dentition P, is single-rooted; P, shows little or no development of talonid; Mg, tends to be smaller than M, and the talonid little reduced; the talonids of M,_, are more fully basined than in Anasinopa. Anasinopa appears to represent a stage of evolution between Sinopa and Tritemno- don (Middle Eocene of N. America). It is less advanced than Metasinopa (Lower MIOCENE CARNIVORA OF EAST AFRICA 263 Oligocene of Egypt) which has lost P, and has very reduced metaconids on lower molars. TABLE 2 Measurements (in mm.) on holotype of Anasinopa leakeyi (M.19081 a-e) ae M! M2 M3 M.19081 a a—p 113308) 12°4 trs 13°3 13°6 M.19081 b a—p 12°2 14°2 trs 13°0 15°3 (e Py Py Ps in M;, Ms, Ms, M.19081 c a—p 83 6:0 9°7 I1°6 13'0 12°3 14° 16°6 trs 5°8 3°5 51 5°8 6:9 Fit 8-2 8°5 M.1g081 d,e a-p 8-1 13°2 14°5 15°5 trs 5°6 73 79 8°7 Genus METASINOPA Osborn 1909 Diacnosis. ‘P;, Mg. As in Pterodon and Apterodon a basal talonid is preserved, which distinguishes this animal from Hyaenodon. A persistent metaconid on M, and M, distinguishes this animal from Pterodon and Afterodon and relates it to Sinopa and Tritemnodon. The lower premolars are small and P, absent. Heels of the lower molars small, trenchant.’ [after Osborn 1909]. Type Species. Metasinopa fraasi Osborn. Nearly complete left mandible (Amer. Mus. No. 14453) from the Lower Oligocene of Fayim, Egypt. REMARKS. Osborn (1909) established the genus on the mandible and tentatively referred to the same genus a maxilla from the same beds (Amer. Mus. No. 14452). Osborn further suggested that Sinopa ethiopica Andrews (1906) was probably a species of Metasinopa. The holotype of S. ethiopica is a left mandible with P,, M,_., and Andrews provisionally placed it in Simopa: Osborn’s suggested reference to Metasinopa seems reasonable. Metasinopa napaki sp. nov. (Text-figs. 23, 24) Diacnosis. Mz, with talonid half as long as trigonid; metaconid present, proto- conid and paraconid trenchant, talonid sloping downward and lingually from buccal ridge. HorotyrPe. M.19097. Left mandible fragment with broken Mg. Locatity. Napak I, Karamoja, N.E. Uganda. DEscRIPTION. Only a broken left M, is known. The protoconid and paraconid are well developed and have trenchant outer face: only the base of the metaconid is present and it appears to be a small cusp: the talonid is half as long as the trigonid and much narrower; the incipient basin has high outer and low inner margin. 204 MIOCENE CARNIVORA OF EAST AFRICA REMARKS. The fragmentary evidence does not warrant any firm deductions. The specimen is provisionally included in Metasinopa largely for convenience and because there is no evidence for separation. To the same species is also provisionally referred a maxillary fragment with P*+4 (M.19096). The specimen possesses the root of P? and complete single cusped P3: P* has strong protocone but no parastyle; the metacone is smaller than protocone; a cingulum is present anteriorly and buccally. The absence of a parastyle on P4 prevents its inclusion with Sinopa, Anasinopa, Dissopsalis or Prodissopsalis. The animal was the same size as the holotype of M. napaki. LS 23 24 Fics. 23, 24. Metasinopa napaki sp. nov. (23) Left mandible with Mg. Holotype (M. 19097), Napak. (24) Maxilla with P**4. (M. 19096), Napak. x2. Metasinopa napaki is smaller than M. fraasi and larger than M. ethiopica, and differs from both in having a proportionately longer talonid on My. Until more material is available it would be best to retain Metasinopa for the inclusion of the following specimens :— Metasinopa fraasi (type species) Amer. Mus. 14453 Left mandible Lower Oligocene of Fayim, Egypt. Metasinopa (?) sp. Amer. Mus. 14452 Left maxilla Lower Oligocene of Fayim, Egypt. Metasinopa ethiopica Geol. Mus. Cairo C.10193 Left mandible Lower Oligocene of Birket-el-OQurun, Egypt. Metasinopa napaki M.19097 Left mandible Napak I, Karamoja, Uganda. ” M.19096 Right maxilla Napak I, Karamoja, Uganda. MIOCENE CARNIVORA OF EAST AFRICA 265 Measurements in mm. on Metasinopa napaki M, IPs) p4 ant-post 12°0 M.19097 lat 6°6 trigonid length 79 M.19096 ant-post 8-6 9°8 lat 5:0 9°8 Genus DISSOPSALIS Pilgrim 1910 Dracnosis. Dental formula I ?, C+, P 4, M 3: carnassials MZ and M32: protocone prominent, especially on P4, and placed anterior to and remote from para- cone: parastyle reduced: metastyle prolonged into shear: premolars robust with well developed cingulum: P* almost as large as M1; P, larger than M,. Molars trenchant ; M!*? with large protocone, connate paracone and metacone, shearing metastyle; M® very small: M, three cusped trigonid and basined talonid; M, paraconid-proto- conid shear strong, without metaconid, talonid reduced to small peg. [After Colbert 1933]. Type Species. Dissopsalis carnifex Pilgrim. The generic name reflects the double carnassial shear on two sets of molars, M5 and M3. LocaLity AND Horizon. Type species from Chinji, Salt Range, Siwalik Hills, India; Chinji stage, ?>Middle Miocene. In addition to the type species Pilgrim (1910, 1914) described a second and smaller species, D. ruber, from the same horizon and locality. REMARKS. Pilgrim’s descriptions, based on fragmentary material, are a remark- able example of his insight. Colbert (1933) had available a skull of D. carnifex collected on the American Museum Expedition, and has given a full account of it, together with a referred mandibular fragment. Dissopsalis pyroclasticus sp. nov. (Pl. 3; Text-figs. 25-27) DIAGNOsIS. Species much larger than D. ruber and approximately same size as D. carnifex but jaw shorter and teeth crowded. No diastema between premolars nor between P, and canine. Metaconid progressively reduced; small on M,,, vestigial on M, and only pin-point on Mg. Shear progressively improved from M, to Mg. Talonid basined on M,, 4; minute peg on Msg. HoLotyPe. M.19082. Right mandibular ramus containing P,, M,_, and alveoli of canine, P, 5. No other specimens can be assigned to the genus with certainty. Locatity. Kaboor, Northern Frontier District, Kenya. 2606 MIOCENE CARNIVORA OF EAST AFRICA eM ff Si 25 Fics. 25-27. Dissopsalis pyroclasticus sp. nov. Right mandible with P,, M,-.. (25)Occlusal aspect. (26) Lateral aspect. (27) Medialaspect. Holotype(M.19082), Kaboor. x1. DeEscRIPTION. The mandible is robust and heavy. The symphysial junction extends back as far as the middle of P,. The canine alveolus is not fully preserved but suggests a normal sized canine. P, is single rooted and crowded behind the canine. P,and P, both had large double roots and were tightly packed against each other close behind P,. P, is a large heavy tooth; the cusp is keeled anteriorly and posteriorly, the posterior keel continuing into a small accessory cusp, trenchant buccally and sloping down to a cingulum on the lingual side. M, is smaller than P,; the trigonid is greatly worn, and of the three cusps the protoconid was larger than the paraconid, while the metaconid was very much smaller than either of the others; the talonid is broken, but must have been about the same length as the trigonid, basined with high buccal rim and low lingual rim. M gis larger than M, but structur- ally similar; the metaconid is reduced to a minute peg and the paraconid-protoconid MIOCENE CARNIVORA OF EAST AFRICA 267 has a strong oblique shear. M, has a larger trigonid than M, and hence bigger shear area on paraconid-protoconid; the metaconid is detectable only as a pin-head projec- tion of enamel on the postero-internal slope of the protoconid; the talonid is reduced to a small peg. REMARKS. The similarity of the new species to D. carnifex is striking. The differences are so small that it is difficult to determine which is the more or less advanced. The shorter jaw and more crowded dentition, together with the more reduced talonid on M, suggest that D. pyroclasticus is slightly more advanced than D. carnifex. The robust jaw and heavy premolars are reminiscent of Quercytherium from the Phosphorites du Quercy, though in this genus the molars are less specialised. Measurements (in mm.) on holotype of Dissopsalis pyroclasticus (M.19082) P, M, M, M, a—p 16°0 ca. 13°0 17°5 16°5 trs 9:0 7°8 8°7 9°0 Subfamily HYAENODONTINAE Trouessart DiaGnosis. Hyaenodontidae with narrow skull and long face; M3 or M3; molars sectorial, length greater than width; M3 small and transverse or absent; M1!*? with paracone and metacone completely or nearly connate, protocone reduced or absent; lower molars without metaconid, talonid vestigial or absent. REMARKS. Simpson (1945) listed seven genera in the subfamily; of these, four are clearly good genera, Pterodon, Apterodon, Metapterodon, and Hyaenodon: Propterodon is less well known. Hemipsalodon is a synonym of Pterodon and Dasyurodon a synonym of Apterodon. To these is here added a new genus, Leakitherium. TABLE 3 Distribution of Hyaenodontine Genera EUROPE AFRICA ASIA N. AMERICA LOWER MIOCENE Hyaenodon UPPER OLIGOCENE Hyaenodon MIDDLE OLIGOCENE = AHyaenodon Hyaenodon Hyaenodon A pterodon LOWER OLIGOCENE Hyaenodon Hyaenodon Hyaenodon Hyaenodon A ptevodon A pterodon Pterodon Ptevodon Pterodon Metapterodon Leakitherium UPPER EOCENE Hyaenodon Hyaenodon Hyaenodon Ptevodon Pterodon Ptevodon Propterodon MIDDLE EOCENE Propterodon 268 MIOCENE CARNIVORA OF EAST AFRICA Genus METAPTERODON Stromer 1926 D1aGnosis. Hyaenodontine with M3; P*-M3 slowly increase in size; M? small and transverse; M!*? sectorial with buccal cingulum, parastyle minute or absent, protocone present; P® simple two rooted. TYPE spEcIES. MW. kaiseri Stromer from Elizabethfeldern, S.W. Africa; horizon stated by Stromer (1926) to be Lower Miocene. A second species from Rusinga is described below. Pterodon biincisivus Filhol (1876) from the Lower Oligocene of Phosphorites du Quercy, France is here trans- ferred to the genus Metapterodon. REMARKS. Stromer’s original diagnosis placed much emphasis on skull features, in particular the position of the infra-orbital foramen. I consider these plastic architectural modifications and thus variable from species to species, depending largely on size and adaptation requirements. From Schlosser’s remarks, it appears that he compared Metapterodon kaisert with only two species of Pterodon, the type species P. dasyuroides and P. africanus from the Fayam. He makes no mention of the other five species of Ptervodon described prior to 1926. A critical phrase in Stromer’s diagnosis is ““Zahngrésse von P? bis M? stark zunehe- mend’’. With this I disagree; the tooth size, as seen in Table 4 does increase from P? to M2, but not greatly. M? is only about one-third as long again as P*. The determinative feature is that the increase is much less than that found in Pterodon species. In Metapterodon the reduced parastyle and well developed protocone (as noted by Stromer) form clear generic distinctions from Pterodon. On the basis of the above diagnosis Pterodon biincisivus Filhol falls within the genus Metapterodon: it also has the infra-orbital foramen above the border of P? and P“as in M. katseri. Metapterodon kaiseri Stromer (Pl. 4, fig. 1; Text-fig. 28) 1926 Metapterodon kaisert Stromer: 110-112, pl. 40, figs. 13, 14 Diacnosis. Metapterodon species of about size of Alofex: skull elongate and slender, infra-orbital foramen above border of P3-P4, P? simple two-rooted: upper molars with outer cingulum, parastyle absent from P4 and M}, rudimentary on M?; protocone well developed on M1**. [After Stromer 1926]. HototyrPe. Left skull fragment with P?-M?, from Elizabethfeldern, S.W. Africa. Stromer (1926) considered the deposit to be Lower Miocene in age on the basis of similarity of fauna with East African fauna. 1926 < I Munich. ADDITIONAL MATERIAL: CMF.4038. Right maxilla with P?—M? from Karungu, Kavirondo Gulf, Kenya. CMF.4066a Left maxillary fragment with P4-M?, teeth broken. CMF.4066b Right mandibular symphysis with broken canine root and two broken premolars. CMF.4066c Left mandibular fragment with roots of molars. All from Rusinga Island, Kavirondo Gulf, Kenya. DESCRIPTION: On specimen CMF.4038 little more than the bone around the teeth is preserved. Posteriorly the root of the jugal arch is preserved: this rises nearly MIOCENE CARNIVORA OF EAST AFRICA 269 Fics. 28, 29. Metapterodon kaiseri Stromer and Metapterodon zadoki sp. nov. (28) M. kaiseyi. Right maxilla with P3-M$; occlusal, medial and lateral aspects. (CMF.4038), Karungu. (29) M.zadoki. Right maxilla with M! 2; occlusal, medial and lateral aspects. Holotype (M.19094), Rusinga Is. Both 1.5. 270 MIOCENE CARNIVORA OF EAST AFRICA vertically and shows no tendency to spread horizontally. The infra-orbital canal issues to the bone surface in the space between the root tips of P? and P4. P®isa simple two-rooted tooth with prominent internal cingulum and small posterior accessory cusp. P* is three rooted, larger than P?, with robust central cusp, well developed posterior accessory cusp and protocone of about same size; the protocone is centrally placed opposite the main cusp and the external cingulum is well marked. M lis slightly larger than P4; the protocone is missing, but from the root it appears to have been about as large as that on P4; the paracone and shearing metacone are about equal in length though in the holotype both are much worn. The external cingulum is prominent and continues anteriorly round the paracone. M? is larger than M!, with prominent protocone placed well anteriorly and remote from the paracone, which is a stout conical cusp from whose anterior border arises a minute parastyle; the metacone is about the same length as the paracone and forms a strong shearing blade, separated by a cleft from the paracone; the angle of shear is very low, being almost parallel to the longitudinal axis; as in M! the external cingu- lum fold continues anteriorly around the paracone. Mis a small peg-like transverse tooth, with a single transversely flattened root: the crown is worn but enough remains to discern the presence of a protocone and larger paracone, beyond which probably lay a small parastyle. The crowns of P*+4 and the paracone of M1! are worn flat and the shearing meta- cones of M!+? show evidence of much wear: neither the protocone on P* nor M? shows any wear, suggesting a deep or very reduced talonids on the lower molars. REMARKS. The strongly sectorial M1*2, together with the prominent protocone suggest a degree of evolution comparable with Pterodon. The genus has not the advanced specialization of Hyaenodon, nor the more tubercular features of A pterodon. A picture emerges of Metapterodon species as medium sized hunters comparable with foxes, as opposed to the heavier built and larger Pterodon species, more comparable to the hyaenas. There are no features on which the S.W. African and Rusinga specimens can be seen to differ. The table of measurements for both specimens shows the close similarity in size: (the figures for the S.W. African specimen are taken from Stromer (1926): in this the Mis missing and the P* is rather narrower). On specimen CMF.4066a all three teeth are broken (P4-M?2), but enough of M? is preserved to make identification certain. The only difference from the specimen described above is a slightly greater size; this however is small and not considered to be of taxonomic importance in view of the proximity of Karungu and Rusinga, and the general resemblance of their mammal faunas. Metapterodon zadoki sp. nov. (Pl. 4, fig. 2; Text-fig. 29) Diacnosis. Slightly larger than the type species ; upper molars robust with strong shear, parastyle absent from M1*?, protocone very reduced on M1*?. The specific name pays tribute to Zadok, the keen-eyed Luo collector on Rusinga. MIOCENE CARNIVORA OF EAST AFRICA 271 HorotyPe. M.19094. Right maxillary fragment with M12. Rusinga Island, Lake Victoria, Kenya. DescripTIon. The holotype is the only known specimen of the species. None of the maxilla save that which supports the teeth is preserved. A posterior fragment of P*is present. M!has a high conical paracone and trenchant metacone, both of about equal width; the shear of the metacone is continued onto the paracone, thus providing a large shearing surface: there is no parastyle, but a prominent external cingulum wraps round the anterior margin of the tooth: the vestigial protocone is little more than a low enamel-capped root placed far anteriorly, clear of the extended shearing metacone-paracone. M7? is an enlarged edition of M}, with which it corresponds in all details. RemMARKS. The most notable differences between M. zadoki and M. kaiseri relate to the carnassial specialization. M. zadoki is the more advanced, having a shear extending onto the paracone, reduced protocone and being without parastyle: this is a stage of development which could easily be derived from M. kaisert. TABLE 4 Measurements of Metapterodon species (in mm.) Metapterodon | Metapterodon | Metapterodon | Metapterodon | Metapterodon Raiservi kaisert kaisevt zadokt biincisivus Stromer 1926 CMF. 4038 CMF.4066a M.19094 Filhol 1876 S.W. Africa Kenya Kenya Kenya Phosphorites Holotype Holotype du Quercy Holotype py a—p 8 7:8 — 12:0 trs 3 49 5°4 ipy a—p 75 8-6 9:0. trs 6°5 7°4 8-2 IIo Mt a—p 9 9°4 9°4. 15°0 trs 7 8-5. JF TIO) M2 a—p I0'5 10°6 I2°3 16:2 trs 10°5 8°5 9°33 14°6 M3 a—p 2 2:8 — 4:8 trs eh 1g 6-4 12‘0 *approximate 272 MIOCENE CARNIVORA OF EAST AFRICA Genus PTERODON de Blainville 1839 Diacnosis. Hyaenodontine with M3; P?—M? rapid increase in size; premolars short and high; M+*? with small protocone, parastyle large; lower molars with strong paraconids, talonids reduced; molars strongly trenchant: infra-orbital foramen above BS TYPE SPECIES. Ptevodon dasyuroides de Blainville from the Lower Oligocene of the Phosphorites du Quercy, France. Also recorded from the same horizon in several parts of France, Germany and the Isle of Wight. The following species have also been described :— P. grandis (Cope 1885). Lower Oligocene, White River Beds, Saskatchewan, Canada. P. magnus Rutimeyer (1891). Middle Eocene, Switzerland. P. africanus Andrews (1903). Lower Oligocene, Fayim, Egypt. P. leptognathus Osborn (1909). ,, at P. phiomensis Osborn (1909). _,, “ ee He P. hyaenoides Matthew & Granger (19250). Upper Eocene, Shara Murun, Mongolia. P. californicus Stock (1933). Upper Eocene, California. ” ” REMARKS. ALl species are strikingly similar and vary mainly in size. None is plentiful or fully known. Three other species described from Europe, P. parisiensis de Blainville (1841), P. cuvierr Pomel (1847b) and P. coguandi Pomel (1853), are considered synonyms of P. dasyuroides. P. biincisivus Filhol (1876) is discussed above and considered to be a species of Metapterodon. Pterodon africanus Andrews (Pl. 4, fig. 3; Text-figs. 30-32) 1903 Ptevodon africanus Andrews: 342, text-fig. 3. Di1AGNosis. Species distinguished from others in the genus on size; all except P. grandis and P. nyanzae are smaller. P. grandis is two-thirds as large again; P. nyanzae lacks an anterior keel on P* and Mt}. HotoryPe. M.8503. Right ramus of mandible with P, ,, M,_, from the fluvio- marine beds, Lower Oligocene, north of Birket-el-Qurun, Fayim, Egypt (Andrews 1906: 220, pl. 19, fig. 3). OTHER MATERIAL. In addition to the holotype there exist skulls and man- dibles described by Andrews (1906) and Schlosser (1911). Further specimens from Uganda and Kenya are referred to below. M.1gogo. Left maxilla with P4, M!?, Napak I, Karamoja, Uganda. CMF.4024. Right P4, Rusinga Island, Kavirondo Gulf, Kenya. DESCRIPTION. On the new maxilla from Napak the crowns of the teeth are broken but otherwise in good condition. On P* the strong paracone has a posterior keel and beyond a keeled accessory cusplet; the protocone has a very heavy root, 273 MIOCENE CARNIVORA OF EAST AFRICA 32 a (9) nm 5 ee (Ss) folie x ©, RO (se) e a = TA aS x" a & ah =a BUS ACS ae a6 * cp gi (a0) ak H 3 _+r n BH oe ow g 2 < 2 45 aS) So ~~ I “4 No) mY Se So Se SRG 2) oO oO Ay n @ Fics. 30-32. 274 MIOCENE CARNIVORA OF EAST AFRICA but is not elevated into a cusp and forms instead an internal shelf; the parastyle is broken externally and is slightly smaller than the posterior cusplet. M1! is much larger than P4; the small protocone is placed very anteriorly and its border projects beyond that of the smaller parastyle: the large paracone is characterised by the presence of a strong external groove and weak internal groove, posterior to which the cusp is trenchant internally; (the groove is indicative of the fusion of the paracone and metacone): a notch separates the paracone from the trenchant blade of the elongate metastyle. Mis a massive tooth with a high paracone, keeled posteriorly to meet the long trenchant metastyle; the parastyle is slightly larger than the proto- cone, which arises anteriorly, well removed from the paracone; the paracone-meta- style blade lies at about 20° to the longitudinal axis of the jaw. The small alveolus denotes a vestigial transverse M°. REMARKS. The similarity of the East African material to the Fayim specimens is so close and the differences so trivial, I can find no case for separating them into different species. The size of the Faytm and Napak teeth are very similar, though few measurements in the table are precise due to damage affecting nearly all teeth. The Faytm skull (C.10192) is rather lighter in build than the Napak maxilla, while the holotype jaw has a massiveness more akin to the new find. These differences are trivial and amount to no more than individual variations; possibly the holotype and the Napak specimen are male and the Faytim skull female. The external groove on the paracone-metacone of M! is barely noticeable on the Faytm skull, but distinctive on the Napak maxilla: otherwise there are no distinguish- ing features. The Rusinga premolar tooth, CMF.4024, referred to the species is a very worn and isolated P?. Considerable interest attaches to the specific linkage of East African sites with Fayum, since the Faytim stratigraphy is well dated and the hyaenodontids appear to be good stratigraphic indicators. This aspect is discussed at the end of the paper. Pterodon nyanzae sp. nov. (Text-figs. 33-35) DiaGnosis. Species larger than P. africanus and smaller than P. grandis. Dis- tinguished from P. africanus by presence of anterior keel on P4 and M1. Name derived from Nyanza, the province of Kenya in which the species occurs. HototyrPe. M.rgogr. Isolated right P4 from Ombo, Kavirondo Gulf, Kenya. PARATYPES. In addition to the holotype, the type locality has yielded two para- type specimens M.19092 isolated left P4, broken and M.19093 isolated right M1}, broken. Another specimen CMF.4007, a very broken left P4 from Rusinga Island, site 3, is also referred to the species. A right M? (UMP 64.33) is recorded from Napak II A. MIOCENE CARNIVORA OF EAST AFRICA 275 Description. The holotype is a nearly complete right P*. The tooth has three roots, anterior, posterior and internal; all are heavy straight-sided roots,the internal the largest and the anterior the smallest. The tooth has a large central conical cusp, accessory cusplets and an internal shelf. The apex of the central cusp is worn flat and the thick enamel is elevated into a ridge anteriorly and posteriorly, the posterior ridge being the higher. The anterior accessory cusplet is much worn by occlusion with P, and the posterior cusplet is strongly keeled: the internal sloping shelf is broad and without any cusp development. 34 35 Fics. 33-35. Ptevodon nyanzae sp.nov. Right P4. (33) Occlusal aspect. (34) Medial aspect. (35) Lateral aspect. Holotype (M.19091), Ombo. xT. The paratype M.1g092 is a left P4 and although part of the internal shelf and all the roots are missing, it is an exact mirror image of the holotype, displaying the same degree of wear. The paratype M.19093 is a right M! with only the anterior half pre- served: the tooth is three rooted and probably the anterior root is the largest. The paracone and metacone are completely fused though there remains a distinct groove externally on the conical cusp, truncated by wear. CMF.4007 consists only of the posterior half of the central cusp and the posterior keeled cusplet. In size and character there is nothing to distinguish it from the two P* teeth described above. REMARKS. The three teeth from the same site may be from the same individual, judging from the degree of wear on the crowns. The table of measurements shows they are considerably larger than P. africanus teeth, much more so than would be expected by individual variation. The well marked anterior keel on both P4 andM! clearly distinguishes these teeth from the P. africanus specimens of Faytm and of Napak. 276 MIOCENE CARNIVORA OF EAST AFRICA TABLE 5 Measurements (in mm.) for Pterodon africanus and Pterodon nyanzae ps M1 M2 P. africanus a—p 25 29 36 C.10192 trs 20 20 23 Fayim, Egypt P. africanus a—p 26 32 38 M.19090 trs 22 Pit 29 Napak, Uganda P. africanus a—p 25 CMF .4024 trs 22 Rusinga, Kenya P. nyanzae (Holotype.) a—p 30 M.19091 trs 24 Ombo, Kenya P. nyanzae a—p 29 M.19092 trs 24 Ombo, Kenya P. nyanzae a—p — M.19093 trs 28 Ombo, Kenya Genus LEAKITHERIUM nov. Dracnosis. Hyaenodontine without M3; M1? highly sectorial, protocone greatly reduced on M?; molars with connate paracone and metacone and shearing metastyle; P4 with protocone and prominent parastyle, central paracone, metacone and trenchant metastyle. TYPE SPECIES. Leakitherium hiwegi sp. nov. Leakitherium hiwegi sp. nov. (Pl. 4, figs. 4, 5; Text-figs 36, 37) Diacnosis. As for genus. Species about size of leopard. HototyrPe. M.19083. Left maxillary fragment with M!+? from Rusinga Island, Lake Victoria, Kenya. The only species. PARATYPE. CMF.4025. Left maxilla with M!and P4 from Rusinga Island, site 3, Lake Victoria, Kenya. This is the only other specimen of the species. DESCRIPTION. None of the skull other than the bone surrounding the teeth is known. The holotype has the bone preserved to the level of the orbit and the begin- MIOCENE CARNIVORA OF EAST AFRICA 277 ning of the jugal is discernible. P* has high central cusp and prominent protocone opposite: the parastyle is smaller than the protocone: the tooth is not well preserved posteriorly but the metacone probably formed a low cusp followed posteriorly by a trenchant metastyle. M1is larger than P*: the metacone is connate with and slightly larger than the paracone; the cusps are divided buccally by a groove and are sectorial on the inner side. The protocone is placed anteriorly, level with the paracone: the parastyle is displaced buccally and much smaller than on P?*: the trenchant meta- style continues posteriorly the shear of the metacone. M? is about the same size as M! and structurally similar, save that the protocone is very reduced and the parastyle absent. ) ‘ [\ Re Z a i Z — i Sy SOY Fics. 36, 37. Leakitherium hiwegi gen. et sp. nov. (36) Left maxilla with M! ?, occlusal, medial and lateral aspects. Holotype (M.19083), Rusinga Is. (37) Left maxilla with P4, M!, occlusal, medial and lateral aspects. (CMF .4025), Rusinga ls. x1°5. Remarks. The presence of two carnassial upper molars and the absence of M? places the species immediately in the Hyaenodontinae. Within this subfamily only Hyaenodon is known to lack M? and on Hyaenodon the protocones and talonids are 278 MIOCENE CARNIVORA OF EAST AFRICA also lacking. Leakitherium displays the clear tendency to greater carnassial efficiency as seen in the morphological series A pterodon—Pterodon—L eakitherium—H yaenodon. The strongly sectorial molars, with cutting plane directed anteroposteriorly and not transversely, are characters found elsewhere among the Hyaenodontidae only in Pterodon and Hyaenodon. The loss of M? in Leakitherium without loss of protocones save reduction on M2, implies a less advanced specialization than in Hyaenodon. Measurements (in mm.) for Leakitherium hiwegi CMF.4025 M.1908 3 1D M1 M1 M2 ant-post ? 16°4 16:0 16°5 16°0 lat 113 14°3 6:0 ? 13°0 Genus HYAENODON Laizer & Parieu 1838 Driacnosis. Dental formula 34-42; M? with shallow groove on completely connate paracone and metacone; molars without protocone; metastyle strongly elongated, especially on M?. Main carnassial pair M2, with M3 as accessory carnas- sials. M, without talonid, M,,. with or without vestigial talonid. TYPE SPECIES. Hyaenodon leptorhynchus Laizer & Parieu. STRATIGRAPHIC RANGE. Upper Eocene to Upper Oligocene, Europe; Upper Eocene to Middle Oligocene, Asia and N. America; Lower Oligocene to Lower Miocene, Africa. The following species have been attributed to the genus:— Europe America N. Hi. leptorhynchus Laizer & Parieu 1838 H. horrvidus Leidy 1853 H. brachyrhynchus de Blainville 1842 H. cruentus Leidy 1853 H. vulpinus Filhol 1876 H. crucians Leidy 1853 H.. compressus* Filhol 1876 H. mustelinus Scott 1894 H. minor Gervais 1848-52 H. paucidens Osborn & Wortman 1894 H. bavaricus Dehm 1935 H. montanus Douglass 1901 H. aimi Cooper 1926 H. leptocephalus Scott 1887 FH. pavisiensis Gervais 1848-52 HZ, vetus Stock 1933 H. martini Depéret 1917 H. minutus Douglass 1901 H. Cayluxi Filhol 1876 HZ. dubius* Filhol 1872 Asia FH. vequieni Gervais 1846 H. pervagus Matthew & Granger 1924 H. aymardi* Filhol 1881 H. eminus Matthew & Granger 19254 H, herberti Filhol 1876 H. yuanchensis Young 1937 H. milloquensis Martin 1906 Hi. filholi Schlosser 1887 Africa FH. ambiguus* Martin 1906 H. brachycephalus Osborn 1909 H. gervaisi Martin 1906 H. andrewsi sp. nov. FH. laurillardi Pomel 1853 H. matthewi sp. nov. Hi. exiguus Gervais 1876 H. pilgvimi sp. nov. *species also occur in Asia MIOCENE CARNIVORA OF EAST AFRICA 279 Remarks. No genus among the hyaenodonts is in so much need of revision as Hyaenodon. No less than 33 species are recorded in the literature and the genus has a much greater stratigraphic range than any other in the subfamily. A survey of the species makes it clear that either there is a very wide range of variation within the genus or several genera are involved, or possibly both factors operate together to produce the present chaotic assemblage of forms. The diagnoses of species within the genus rest almost entirely on size differences; morphological characters used have been found to be inconstant. Much of the known material comprises mandibular remains and in an attempt to unravel the species I plotted the distribution of the following six characters :— a. Presence or absence of Pj. b. P, with r or 2 roots. c. Mental foramina below P,, between P, and Pg, or below P.. d. Presence or absence of buttress on antero-external margin of M.. e. Presence or absence of trace of talonid on Mg. f. Size relation of protoconid to paraconid. P, is almost always present and usually has two roots. There is invariably a mental foramen below P,, frequently below P, and sometimes one or more either below P, or between P,and P,. On M. the buttress is highly variable, being present on some individuals and absent from others of the same species: it is more usually present than absent. Relatively few specimens show trace of a talonid on Mg. The protoconid tends to be larger than the paraconid, sometimes by a considerable margin, and occasionally the two are almost equally long. None of these characters, either singly or in combination, can be used for specific determination. Size is a not very satisfactory criterion on which to base a diagnosis; there is probably a wide range within each species, if only this could be checked, and much overlap. Nine North American species are described, ranging from the very large H. horridus to the small H. mustelinus. Three species are recorded from Mongolia and China and one from the Faytm of Egypt. The remaining 20 species are European and of these 4 also occur in Asia. Many are poorly known and synonyms are inevit- ably rampant. The stratigraphic distribution in Europe illustrates an Upper Eocene group of species and a Middle-Upper Oligocene group. The Upper Eocene group has a possible forerunner in the Middle Eocene, Propterodon: this form from Egerkingen is poorly known but may be a link in the line from proviverrines to hyaenodontines. In the absence of a full scale revision of the genus I find it best to make a com- promise. The African species appear to have one important character in common which is rare if not truly absent from all others, and on this basis they are grouped into a new subgenus. When more material is available, it will probably be possible to elevate this taxon to generic rank. 280 MIOCENE CARNIVORA OF EAST AFRICA Subgenus ISOHYAENODON nov. D1aGnosis. Hyaenodon species in which the protoconid and paraconid of Mg are approximately equal in length. Type SPECIES. Hvyaenodon (Isohyaenodon) andrewst sp. nov. In addition to the type species, the following are included in the subgenus Isohyaenodon: Hyaenodon brachycephalus Osborn, H. matthewt sp. nov., and H. pilgrimt sp. nov. The three new species, H. andrewsi, H. matthew and H. pilgrim are all smaller than H. brachycephalus, and H. pilgrimz is by far the smallest known Hyaenodon species. TABLE 6 Distribution of Hyaenodon species EUROPE AFRICA ASIA N. AMERICA LOWER andrewst MIOCENE matthew pilgrim UPPER milloquensis OLIGOCENE bavaricus leptorhynchus gevvaist MIDDLE leptovhynchus pervagus mustelinus OLIGOCENE gevuaisi aymardi paucidens brachyrhynchus ambiguus leptocephalus vulpinus compressus horvvidus laurillardi dubius cruentus exiguus CYUCIANS compressus martini cayluxt dubius ambiguus fitholi LOWER aymardt brachycephalus yuanchensis horvidus OLIGOCENE andrewst cruentus crucians montanus minutus UPPER aim eminus vetus EOCENE minor parisiensis vequient hevbert1 MIOCENE CARNIVORA OF EAST AFRICA 281 Hyaenodon (Isohyaenodon) andrewsi sp. nov. (Pl. 4, fig. 6; Text-figs. 38-40) Dracnosis. Jsohyaenodon of about the same size as H. minor. M, with vestigial talonid. The trivial name is a tribute to C. W. Andrews (1866-1924), a profound student of fossil mammals who made a singular contribution to our knowledge of the Faytim faunae. Horotyre. M.15048. Right mandibular fragment with M, 5. From Ombo, Kavirondo Gulf, Kenya. ADDITIONAL MATERIAL. In addition to the holotype the following are referred to the species: CMF.402t Right mandible with P,,,, Mo.5, alveoli of C, Py.5, Mj. Moruorot South, Northern Frontier District, Kenya. CMF.4022 Right mandibular fragment with M, and alveoli of My,9. Songhor, Kavirondo Gulf, Kenya. CMF.4023 Broken left M,. Rusinga Island, site 3, Kavirondo Gulf, Kenya. C.8812-13. Fragment of right mandible with M,: from Fluviomarine beds, Birket-el-Qurun, Fayim, Egypt. DeEscripTIOn. The holotype is much broken and little of the mandible remains other than that part enclosing the tooth roots, but the surviving parts give the impression of robustness. M, is small and much worn, in contrast to all other teeth which are unworn, M. being not even fully erupted. On M, the protoconid is rather larger than the para- conid: the summit of the protoconid is truncated by wear and the paraconid appears to have been damaged during life since on the broken anterior surface the fracture edges are not fresh; these two cusps are separated by a shallow cleft, and the tren- chant labial face is longitudinally aligned: the small low talonid slopes antero- medially. M, is much larger than M,; the paraconid and protoconid form good shearing facets; the talonid is proportionately smaller than on M, and a small buttress is present near the base of the antero-labial edge of the paraconid. The large M. is almost fully erupted and completely unworn; the paraconid is as broad as but not as high as the protoconid and the cleft which separates them extends almost to the base of the tooth; the carnassial shear is only slightly oblique and a small buttress is again present on the outer edge of the paraconid: a vestigial talonid is present on the postero-internal border. An anterior mandibular fragment with three premolars has been glued to the holotype described, although its true association with this is dubious. There is very little actual bone contact and this does not form a neat join. I suspect that at best this is an anterior fragment with a middle section between the two parts missing, or else it has nothing to do with H. (I.) andrews. This second fragment, if it truly 282 MIOCENE CARNIVORA OF EAST AFRICA Fics. 38-40. Hyaenodon (Isohyaenodon) andvewsi sp. nov. Right mandible with M,-,. (38) Occlusal aspect. (39) Lateral aspect. (40) Medial aspect. Holotype (M.15048), Ombo. X1°‘5. belongs to the same individual, may represent three milk premolars and part of the alveolus of the canine; the premolars are small in comparison with the molars of the holotype. The first is a small two-rooted tooth with posterior accessory cusp lying very close behind the canine. The second is similar but slightly larger. The third premolar is more robust with high pointed central cusp and low posterior cusp; a cingulum is present anteriorly and postero-internally. The first two premolars appear to belong to the same dentition; compared with the permanent molars, their small size suggests milk teeth; their position with respect to the canine makes it likely that they are DP, 9, though it is possible they could be DP, 5, allowing for a very small first milk premolar, or even its absence. MIOCENE CARNIVORA OF EAST AFRICA 283 The second mandible, from Moruorot, is less broken but fewer teeth are present. The jaw is long, shallow and strongly built; anteriorly there is a long symphysis and posteriorly the mandibular foramen issues well behind M,. Mental foramina are present under P, and Py. The angular process of the jaw is stout and dorso-ventrally flattened; the masseteric fossa deep. The posterior part of the canine alveolus is large and ovoid in section. P, and P, are absent: P, was small and apparently single rooted, and P, much larger and two-rooted, the posterior root being preserved in the alveolus. The crown of P,is much worn and broken: it appears to have had a conical central cusp and small posterior accessory cusp. P, is complete, the tip of the cusp is somewhat worn and its anterior and posterior borders become slightly concave towards the base, with a prominent posterior cingulum around the accessory cusp. M, is missing and M, broken with only the base remaining; this indicates a strong paraconid-protoconid shear, slightly oblique, behind which lay a small antero- medially sloping talonid. Mg, is complete and indistinguishable from that on the holotype; the tooth is fully erupted in a crowded jaw, and lies obliquely resulting in a transverse shear. Specimen CMF .4022 contains a Mg, with broken protoconid, but otherwise little worn. The fully erupted tooth in situ has an oblique shear, which runs at 40° to the longitudinal axis of the tooth row. CMF.4023 comprises a very broken tooth stump which is referred to the species on size, absence of talonid and metaconid, and obliqueness of shear. Andrews (1906) referred a specimen (C.8812-13) from the Fluvio-marine beds north of Birket-el-Qurun in the Faytim, to the genus Hyaenodon. On page 219 he described this right mandibular fragment and associated Mg. The tooth is structurally indistinguishable from the holotype described above and nearly the same size. Nothing debars the specimen from inclusion in the species H. andrewst. Hyaenodon (Isohyaenodon) matthewi sp. nov. (Text-figs. 41-43) Diacnosis. Isohyaenodon of rather smaller size than H. (I.) andrewsi and Ms with less oblique shear; details in accompanying table. Trivial name is a tribute to W. D. Matthew, a British geologist whose signal contributions to mammalian palaeontology are unsurpassed. HoLotyPe. M.19098. Left mandibular fragment with M, and broken M, from Songhor, Kavirondo Gulf, Kenya. OTHER MATERIAL. The following additional specimens are referred to the species: CMF.4060 Left M, from Rusinga Island, Kavirondo Gulf, Kenya. CMF.4061 Right M, from Rusinga Island, Kavirondo Gulf, Kenya. DEscCRIPTION. A fragment of mandible surrounding M, is preserved. The build of the jaw is lighter than in H. (I.) andrews; the masseteric fossa is pronounced and a groove is present on the lower anterior face of the coronoid crest. M4, is similar to that on H. (I.) andrews: but proportionately smaller. The protoconid is higher than 284 MIOCENE CARNIVORA OF EAST AFRICA 42 Al 43 Fries. 41-43. Hyaenodon (Isohyaenodon) matthewi sp. nov. Left mandible with M,-5. (41) Occlusal aspect. (42) Medial aspect. (43) Lateral aspect. Holotype (M.19098), Songhor. XI. the paraconid and the shear plane convex outward; the obliqueness of shear is 20°, that is much less than in H. (I.) andrewsi. A vestige of the talonid is present and on this unworn tooth a very minute prong is visible on the internal cingulum at about the place where a metaconid would arise if one was present. The small talonid of M, abuts against the paraconid of Mg, the alignment being maintained by a small buttress; no more of M, is preserved on the holotype. On a left M, (CMF.4060) from Rusinga the protoconid is decidedly higher than the paraconid and a small talonid is present; the shear plane is convexly curved and not as oblique (at 15°) as that of Ms. CMF.4061 is an isolated right Mg, indistinguishable from the holotype, but slightly broken on the cusp tips. Hyaenodon (Isohyaenodon) pilgrimi sp. nov. (Text-figs. 44-49) D1aGnosis. Small species of Isohyaenodon,; about half the size of H. filholi and H. mustelinus. The trivial name is a tribute to G. E. Pilgrim (1874-1943), an outstanding scholar of European and Asiatic Tertiary mammals. HoLotyPe. M.1g100a-c. Pair of complete mandibles with posterior fragment of skull and 7 cervical vertebrae; from Rusinga Island, Kavirondo Gulf, Kenya. OTHER MATERIAL. CMF.4062 Right mandible with P,, M,,5. Rusinga Island, Kavirondo Gulf, Kenya. CMF.4063 Left mandible fragment with P,,,, M,. Rusinga Island, Kavi- rondo Gulf, Kenya. CMF.4064 Upper left M!. Songhor, Kavirondo Gulf, Kenya. CMF.4065 Left lower canine. Rusinga Island, Kavirondo Gulf, Kenya. DESCRIPTION. This species is about the same size as the polecat, Mustela putorius. The two mandibles are complete though separate and only the incisors and P, are missing. The mandible is about the same size as that of a pine marten; it is lightly built with a long symphysis. The coronoid is high and the crest rounded. The condyle is transverse, well rounded and on a level with the tooth row; below is the MIOCENE CARNIVORA OF EAST AFRICA 285 short stout angular process. The masseteric fossa is not deep and the mandibular foramen issues just posterior to Mg. On the labial side of the jaw there is a row of mental foramina, below Py, P35, M, and Mg. C HR@aseeo - 44 \ Fies. 44-46. Hyaenodon (Isohyaenodon) pilgvimi sp.nov. Partial reconstruction from right and left mandible with C, P,-M,. (44) Occlusal aspect. (45) Medial aspect. (46) Lateral aspect. Holotype (M.19100), Rusinga Is. x2. No incisors are known and from the proximity of the canine to the symphysis they must have been very small. The canine is slightly flattened on the medial side and tapers upward with a gentle curve backwards. P, is absent on both sides, and was apparently a small single-rooted tooth. P, is two-rooted, the cusp is centred over the anterior root and it slopes backward to a small accessory cusp over the posterior root. On P, the cusp is centrally placed, with a longitudinal keel running anteriorly to a slight cingulum and posteriorly to a small accessory cusp. P, is similar structurally to P, only slightly larger. M, is broken on both sides, but was clearly a small obliquely shearing molar, with protoconid rather larger than the paraconid and a vestigial talonid. M4, is similar to M, but considerably larger and Mgis slightly larger than My. On M, the well developed shear is oblique, the proto- 286 MIOCENE CARNIVORA OF EAST AFRICA conid is higher than the paraconid and the only trace of the talonid is a slight bump of enamel on the posterior edge of the protoconid. A minute buttress is present on the antero-labial edge of M, and Mg. A left M1 from Songhor is referred to the species. L 2 | ) 48 47 49 Fics. 47-49. Hyaenodon (Isohyaenodon) pilgvrimi sp. nov. Left M1. (47) Occlusal aspect. (48) Lateral aspect. (49) Medial aspect. (CMF.4064), Songhor. x4. Together with the two mandibles, the posterior part of the brain-case and seven cervical vertebrae are preserved; all are broken and partly crushed, particularly the posterior part of the skull, so that of this no details can be discerned. In size and build the cervical vertebrae approach those of Mustela putorius, the polecat. The atlas vertebra possesses a fairly wide dorsal and narrow ventral arch, the latter with well developed longus colli tubercle. No rectus capitus posticus minor muscle scar is visible on the anterior face of the dorsal arch. The wings, though broken, can be seen to be light and did not project far laterally, no more than one third the width of the neural arch. The oblique foramen issues dorsally behind the cotylar process. The vertebrarterial canal is very short, the ventral and posterior openings being close together at the base of the wing. The axis is closely comparable with that of the polecat, and differs from it only in having light non-tuberculate posterior zygapophyses and the posterior extension of the spine beyond the neural arch, though broken, was probably thinner and shorter. The remaining cervical vertebrae are partially crushed, especially on lateral and ventral faces. Their dorsal surfaces reveal that the neural spines were vestigial on third to fifth inclusive; the sixth cannot be seen and on the seventh a small spine was present. The close comparison in size and proportions of the jaws and vertebrae with those of the polecat is instructive. In appearance Isohyaenodon pilgrinu must have looked very like a polecat, but perhaps without the latter’s strength and agility. There is a consistent weakness in the development of dorsal musculature, a feature found in modern aquatic carnivores. REMARKS ON ISOHYAENODON SPECIES The most striking thing about the three new species is their similarity to each other and difference from other Hyaenodon species. Secondly the Isohyaenodon material falls readily into three groups on basis of size differences. The only other described species from Africa, H. brachycephalus can be included in the subgenus Isohyaenodon on the basis of near equality of paraconid and protoconid on Mg. Its size is greater than any of the three species described above, and it possesses a relatively short mandible compared with others in the subgenus. The distribution of this character of mandibular length is another variable in the genus, which on available material, does not form any meaningful pattern. 287 MIOCENE CARNIVORA OF EAST AFRICA ‘ozewurxoidde,. 40-61 L.€€ €.z€ bozEr ‘ON ‘SHI “Ioury snyoydaotyovaq *(T) “H o.€ S.F toot INO ss DE TS Crya OOS CLAS €90h' AWO - Ze oian ic OG 6.1 1b zooh AWO Le QoS (Seon Osh La, G-EMQs Osc val y.c | <<< Satan — yYSII ONE CHS wd TAG Gow AS Gee GS epi iS ims EHS Suize g-b1 3J°1 OO16I IW wmiss)ig *(T) “HH L.9 Q.zI 190F INO ie g-S QI o90h TINO i avoy Dicank g6061 WW amayyoue “(T) “HH ol o.€1 €-z1g9"9 ce €.L 40.41 Ezorq INO $.g €.91 zzob WO “ 9-8 4.61 — 6.721 GG 1 — Zrr x0-LE 40.9€ 1zov TINO . xG-L V.GI 0.9 0.71 I-F 6.8 9-9£ gboSi Ww asmaapun *(T) “HT sy} d-e sy d-e sy d-e sm d-e sy d-e sy d-e sx} d-e "wow "dtd °W "Wl iN val fq Sl TN ‘sotoads (uopouavkyosy) uopouavAy uo (Wu ut) syusWIoInseoyy “2 ATAV IL 288 MIOCENE CARNIVORA OF EAST AFRICA Suborder FISSIPEDA Blumenbach Superfamily CANOIDEA Simpson Family CANIDAE Gray Diacnosis. Arctoidea, with a moderately high skull, brain-case not expanded; auditory bulla originally small, remote from the paroccipital process, as in Amphi- cyoninae, but in most lineages ultimately becoming enlarged and inflated and brought into contact with the paroccipital process; alisphenoid canal present. Dental formula 3-+-+-2-~: P* elongate, protocone prominent in early genera, later much reduced; M1! 3- or 4-tubercular, often with intermediate cusps, always broader than long, becoming progressively larger along many lines; M? similar, only very exceptionally lost; M® present in some early genera, and in most of the Amphi- cyoninae; lost later; M, with metaconid strong in primitive genera and in the Amphicyoninae, progressively weaker along other lines; talonid with trenchant hypoconid; entoconid present either as a ridge, shelf, or tubercles in the early genera, and retained on most lines; M, long; Mg only exceptionally lost in some highly specialized genera; digitigrade; primitive members with five digits, later forms with first digit both in manus and pes reduced. [after Pilgrim 1931]. REMARKS. About 60 genera of fossil canids are recognised and in addition there are 12 living genera. The classification of these numerous genera within the family Canidae presents many difficulties and numerous attempts have been made. None is completely satisfactory. The relative abundance of fossil forms increases rather than lessens the difficulties. Simpson (1945) has stated of the Canidae that their “status as a single family ...can be upheld without serious doubt” and then adds “the whole group is extremely polyphylectic’—two statements which appear irreconcilable. It is beyond the scope of this work to attempt yet a further revision of the classification (a task begun by Hough (1948) for American fossil genera and by Hiirzeler (1944) and Ginsburg (1955) for some of the European genera): It will suffice here to adopt the status of the subfamily Amphicyoninae as defined below, without reference to its relationship to other subfamilies. Subfamily AMPHICYONINAE Trouessart D1acnosis. Canidae, with auditory bulla small, little inflated, remote from the paroccipital process; mastoid process prominent and broad; molars progressively enlarged, premolars and carnassials progressively reduced; upper molars trituber- cular, with broad postero-internal shelf; M? and M, present in most forms, but lost in advanced members; P* with progressively reduced protocone; M, generally with metaconid, talonid with hypoconid and ridged entoconid. M, with bicuspid triginid and crested talonid; limb bones massive; humerus with entepicondylar foramen; manus and pes 5-dactyl. {after Pilgrim 1931.| REMARKS. Only one genus (Wammocyon Loomis 1936) has been added to the subfamily since Pilgrim (1931) discussed its status. Arambourg (1961) described a very worn mandibular fragment from Gebel Zeltan in Libya as A frocyon, a new genus MIOCENE CARNIVORA OF EAST AFRICA 289 of amphicyonid. Until more material is available nothing useful can be added to this statement. In the confusion that exists regarding the affinities of the canid genera, it would be futile to attempt a concise formulation of any one subfamily. The new genus described below is included with the Amphicyoninae on the basis of its close parallels to Amp/icyon, rather than on a rigorously definitive basis. When canid systematics are usefully revised, the two genera will probably be placed close together. Genus HECUBIDES nov. Diacnosis. Amphicyonine with long face; dental formula 2:+4:2; anterior premolars well spaced; P* reduced in comparison with molars; protocone not so anteriorly placed as in Amphicyon and parastyle absent; carnassial blade not oblique; M! only slightly larger than M2; both sub-triangular and transverse width greater than length, lunate protocone and extensive internal cingulum more asym- metrical and smaller on M!; paracone and metacone equal on M2 and metacone only slightly smaller than paracone on Mt. M? small: M, with well developed trenchant hypoconid and ridged entoconid. TYPE SPECIES. Hecubides euryodon sp. nov. In addition to the type species a second new species, H. macrodon, is described below and two others are referred to the genus, H. americanus (Wortman) and H. lemanensis (Pomel). Remarks. The probable affinities of the new genus are described below in the remarks on the type species. It is quite distinct from both the European Amphicyon and the American Daphoenis and appears to represent a separate stream of develop- ment. Hecubides euryodon gen. et sp. nov. (Pl. 5, fig. 1; Text-figs. 50-54) Diacnosis. Medium sized Hecubides species: M! about 18 mm. transversely and 15 mm. longitudinally: M1 with strong internal cingulum, no crenulation on either protocone or internal cingulum of either M1! or M?. DERIVATION OF NAME. Generic name from Hecuba, princess in Greek mythology who was changed into a stone dog. The trivial name is from the Greek ewrys, wide or broad. HorotyPe. M.19084. Maxillary fragment with P?+4, M!+? and alveoli of P? and M? on both sides. Locatity. Locality I, Napak, Karamoja, Uganda. PARATYPE. In addition to the holotype, the type locality has yielded an isolated right M, (M.19085). 290 MIOCENE CARNIVORA OF EAST AFRICA ADDITIONAL MATERIAL. The following specimens are referred to the species :— M.14313 Right M}, broken Koru, Kavirondo Gulf, Kenya. M.19099 Left M, Locality I, Napak, Karamoja, Uganda. CMF.4026 Right M, Songhor, Kenya. CMF.4027 Right M, Rusinga Island, Kavirondo Gulf, Kenya. CMF.4067 Left M,, talonid only Rusinga Island, Kavirondo Gulf, Kenya. CMF.4068 Left M?, protocone only Mfwanganu Island, Kavirondo Gulf, Kenya. CMF.4069 Left M,, trigonid only Rusinga Island, Kavirondo Gulf, Kenya. H.M.V.5830 Left M, Locality IV, Napak, Karamoja, Uganda. UMP64.32 Right M, Locality I, Napak, Karamoja, Uganda. DEscrIPTION. The holotype consists of an incomplete maxillary dentition together with the adjacent parts of the maxillae bones. P4, M1*? are preserved on both sides: P? is complete on the right side, broken on the left: parts of the alveoli of P* and M® can be distinguished. The dentition gives an overall impression of robustness; the teeth have thick enamel, low cusps and are relatively wide. The premolar teeth are well spaced and indicate a relatively long but strong jaw. Nothing anterior to the alveolus of P? is known; of this tooth only the posterior root cavity remains and it was presumably a smaller version of P? which is a single cusped birooted tooth, narrow and elongated. In P‘ the fully sectorial paracone-metastyle blade is directed antero-posteriorly; the paracone is higher and larger than the metastyle and the two are separated by a narrow cleft: no parastyle is present, though a slight bump can be detected on the ridge of enamel which falls anteriorly from the summit; the protocone is small, low, close to the paracone and lies midway between the paracone summit and its anterior border. M!? is a large triangular tubercular tooth, with its transverse width greater than its length; the paracone is slightly larger than the metacone and both have prominent antero-posterior keels. The crescentic protocone, separated by a wide basin from the paracone and meta- cone, is slightly asymmetrical, being shorter but heavier anteriorly; lingual tothe protocone is a thick cingulum, and a narrow cingulum runs buccal to the paracone and metacone. M? is only slightly smaller than M! and structurally very similar; the protocone is symmetrical and less high than in M1 and the lingual cingulum is larger and more expanded. The only evidence of Mis a trace of the alveolus indicat- ing a small, transverse two rooted tooth. Mandibular teeth referred to the species are first and second molars. M, is a robust tooth, the trigonid is about twice as long as the talonid; paraconid and protoconid are sectorial, protoconid is much the largest cusp; the metaconid is small, adhering MIOCENE CARNIVORA OF EAST AFRICA 291 to the protoconid with crest on level of the paraconid; the talonid is basined, length and breadth about equal, hypoconid well developed and entoconid forms a low ridge. M, is a stout rectangular tooth; the prominent protoconid is paired with a smaller metaconid and there is no paraconid; posteriorly the hypoconid is keeled and continues in line with the protoconid; the entoconid presents a curved ridge linking the paraconid and hypoconid. The other specimens do not call for any special comment. Most are M, and few are unbroken. REMARKS. Hecubides is known from three sites in the Kavirondo region of Kenya and from Karamoja in Uganda. These, with Afrocyon from Libya, are the earliest records of canids in Africa, the next being Canis and Vulpes in the Lower Pleistocene. In comparing Hecubides with other canids, we may limit study to those genera grouped by Simpson (1945) in the subfamilies Caninae, Amphicyonodontinae and Amphicyoninae. The American subfamily Borophaginae is quite different and need not be considered: the octocyoninae, with one living African genus and two possible Pleistocene precursors, has very atypical molar characters: all simocyonines are characterised by the absence of M? and very reduced M2. Of the numerous American canid genera, none is as close to Hecubides as some European Amp/icyon species. The following characters clearly differentiate the American canids: many are either without M? (as Nothocyon, Cynodesmus, Mammo- cyon and Pliocyon) or the tooth is very reduced as in Proamphicyon: in most genera M? is considerably smaller than M! (e.g. Hesperocyon, Daphoenus and Campylocyno- don): the development of the protocone on P‘is large with poor or oblique shearing blade in Daphoenus and Daphoenodon, the protocone is small and the blade strongly sectorial in Mesocyon and Mammocyon: the metaconid is a free and fully developed cusp in the M! of Daphoenus and Parictis, and the talonid has high entoconid and hypoconid cusps in Tomarctus and Leptocyon: a paraconid is present on M, in Leptocyon and Tephrocyon. Omitting the oasis of synonyms, and other genera either so different or so poorly known that they do not warrant discussion, only the American species referred to Amp/icyon remain and these are discussed below together with the old world species. Among the European Tertiary canids, the closest affinities with Hecubides are to be found among the Amphicyon group. Among the Hemicyon group of genera (Hemicyon, Harpalaeocyon, Dinocyon, Plithocyon, Phoberocyon) there are fairly close similarities in the structure of the upper and lower molars, but all are more specialised for crushing; the teeth are more tuberculose, the upper molars wide and almost square in some cases, the protocone of P4 large and medianly placed. The essential differ- ences between Hecubides and the remaining genera can be briefly listed :Cynodictis (with Plesiocyon and Pachycynodon) possesses viverrid-like characters—very reduced M?and high tricusped trigonid on M ,: Cephalogale, Alopecodon and Pseudamphicyon all lack M? and have reduced M?: Amphicynodon (synonyms Cynodon and Paracyno- don) has tricusped trigonid on M, and a paraconid is present on My. 292 MIOCENE CARNIVORA OF EAST AFRICA SS = pS Vir ATI \\\NS \ Fies. 50-54. Hecubides eurydon gen. et sp. nov. (50) Maxilla with P3-M!, based on the right and left sides of dentition; occlusal aspect. Holotype (M.19084), Napak. (51) same, lateral aspect. (52) M; and Mg, based on M.19085, M.19099 and CMF.4027; occlusal aspect. (53, 54) Mj, based on M.19085 and M.19099; medial and lateral aspects. All x2. MIOCENE CARNIVORA OF EAST AFRICA 293 Only two additional genera require examination to survey the Asiatic Tertiary canids. Vishnucyon bears no resemblance whatever to Hecubides; its P4 is without a protocone, the M! is deeply waisted and M? very reduced. Arctamphicyon, known from M1!*+2, suggests ursid affinities in its narrow but transversely extended molars. The type species of Amphicyon is A. major Blainville. This species is clearly generically distinct from Hecubides, the most striking differences being in M2, which on Amphicyon major is asymmetrical transversely, has a crenulated internal cingulum and the paracone is much larger than the metacone: the molar teeth are subsquare rather than sub-triangular and the internal cingulum is less extensive in both teeth than in those of Hecubides. All four premolars are present though P tis vestigial and there is a diastema between it and the canine, which reaches the proportions of a sabre-tooth: the diastema behind the upper canine suggests further an elongate lower canine. The asymmetry of the buccal border of M? and the striking difference in size of paracone and metacone, are features which clearly mark off the Amphicyon group from Hecubides. About 70 species have been referred to the genus Amphicyon. Apart from a profusion of synonyms it is clear that several genera are involved and some of the species bear little resemblance to the type species. Only those which are clearly nearer to the genus Hecubides than to the type species A. major will be discussed below. None of the eight Asiatic species of Amphicyon come within this category. Of the 18 American species attributed to Amphicyon, A. americanus more closely resembles Hecubides than A. major. I have been able from a cast to confirm Matthew’s remarks (1924: 106) that A. sinapius is closer to A. major than to A. lemanensis and has no proximity to Hecubides. None of the other American species comes within the scope of the discussion and only A. americanus is transferred to the new genus Hecubides. Among the European Amphicyon species, A. lemanensis stands out as quite distinct from all others, and close to Hecubides euryodon. Both species are about the same size; P4, M 1:2 are almost identical in each, M! of A. lemanensis has a crenulated protocone and the postero-internal cingulum is much larger and more asymmetrical, while the M? is proportionately broader transversely and the external border is directed postero-internally. Both species are undoubtedly closely allied and the European species is thus placed in the new genus. A. dehmi Crusafont, from the Burdigalian of Vallés-Penedés, N.E. Spain appears to fall between Hecubides and Amplicyon. The rather squarish molars, M! slightly asymmetrical, M? with pos- periorly crenulated internal cingulum and a paracone which is slightly larger than the metacone tend to suggest a closer proximity to Amphicyon sensu stricto. Crusafont (1955) has justifiably placed the species in a new subgenus Jctzocyon of the genus Amphicyon. In conclusion therefore, the new genus Hecubides has four known species, the type species H. euryodon and another new one from East Africa, H. macrodon. To these are added H. americanus (Wortman) from Nebraska (age unknown), and H. lemanen- sts (Pomel) from the Aquitanian of France and Germany. Hecubides appears to be 294 MIOCENE CARNIVORA OF EAST AFRICA an earlier offshoot of the dogs than Amphicyon: the latter could be said to be more specialised in having molars more nearly square than triangular, a more elaborate internal cingulum on M? and a greater development of the paracone at the expense of the metacone. Hecubides macrodon sp. nov. (Pl. 5, fig. 2; Text-fig. 55) Diacnosis. Large sized Hecubides species; M! about 25 mm. transversely and 20 mm. longitudinally; internal cingulum proportionately smaller and external cingulum thinner than in type species. HorotyPe. M.19086. Left M1. Locality. Site 31, Rusinga Island, Kavirondo Gulf, Kenya. DeEscripTIon. Mis structurally similar to that of H. euryodon, but larger, with rounded and less pronounced features. The paracone is slightly wider than the metacone: the protocone forms a broad and shallow crescent and the internal cingulum, best developed posteriorly, is proportionately smaller than in H.ewryodon: the external cingulum is very thin and forms only a skin on the lower edges of the paracone and metacone. I\S \ Fic. 55. Hecubides macrodon sp. nov. Left M!; occlusal aspect. Holotype (M.19086), Rusinga Is. x1°5. REMARKS. Few deductions can be made from an isolated tooth, but its difference from H. euryodon in size and minor details of structure, seem sufficient to merit specific distinction. The remarks on the type species regarding the relationships to Amplhicyon apply also to this species. A right P4 (CMF.4070) from Rusinga is also referred to H. macrodon,; the tooth is very broken and only the outer edge of the paracone and metastyle survive. MIOCENE CARNIVORA OF EAST AFRICA 295 TABLE 8 Measurements (in mm.) on the dentitions of Hecubzdes. is pe Mt M2 M3 M, M2 Hecubides euryodon P3_M? M.19084 (Holotype) a—p 103, GOH EAH) AUG) —- ARO BOF) M, M.19085 (Paratype) M, CMF.4027 trs 5°3 10°3 17°5 16°4 — 9°6 10°3 Hecubides euryodon a—p 23°5 Hunt. Mus. V.5830 trs org) Hecubides macrodon a—p 20°5 M.19086 (Holotype) trs 24°2 Hecubides americanus a—p 15 27 20 17 8 (Wortman) (Holotype, approx.) trs 8 17 27 22 12 Hecubides lemanensis a—p 17°4 14°5 (Pomel) (B.M.N.H., no. 30879) trs 10°3 18°4 Hecubides lemanensis a—p 15'1 12°0 (Pomel) M.7643 trs 19°5 WAP Hecubides lemanensis a—p 20° (Pomel) (B.M.N.H., no. 26733) trs g'I Amphicyon (Ictiocyon) a—p 16°6 15°6 II‘5 7°4 18:0 dehmi Crusafont (Holotype) trs TACT 17°5 164 UH) 9°7 Amphicyon major de Blainville a-p 32°3 PagfAk 22°4 (B.M.N.H., no. 29615) (Cast of holotype) trs 19°2 35°0 Bey Superfamily FELOIDEA Simpson Family VIVERRIDAE Gray Diacnosis. Skull elongate, low with long snout; auditory bulla composite with ecto- and ento-tympanic parts, wholly or only partially ossified. Dental formula $14.2: M1+2 large, tritubercular: P* with well developed protocone; parastyle and metastyle usually present: M, long with tritubercular trigonid and basined talonid; trigonid cusps usually high; oe esee P4/M ,, truly sectorial except in few specialized genera. [After Pilgrim 1931]. 296 MIOCENE CARNIVORA OF EAST AFRICA REMARKS. On teeth alone it is impossible to separate with certainty the miacids from the viverrids and the auditory region is essential for this purpose. This region is missing from the specimens described below and hence their place in the Viverridae must be regarded as provisional. The later miacids and early viverrids are so similar that it is impossible to make a sharp division. Gregory & Hellman (1939) included miacids within their family Viverridae though this practice has not been generally accepted. The miacid subfamily Viverravinae is closest to the Viverridae, all its members lacking M, as in viverrids. Simpson (1945) recognised seven subfamilies in the Viverridae, of which only three, Stenoplesictinae, Viverrinae and Herpestinae are known in the fossil record: the latter two subfamilies contain half the 42 recog- nised genera in the family. Subfamily HERPESTINAE Gill DiaGnosis. External auditory meatus long. Carnassial teeth not strongly trenchant: molars rather more sectorial than tubercular. REMARKS. The Herpestinae are essentially less specialized in the carnassial direction than the Viverrinae, though the trenchant character of the teeth is not so reduced as in Paradoxurinae and Hemigalinae. It is on this basis that the fossils described below are included in the Herpestinae, in leu of any knowledge of the auditory region. No extinct genera are ascribed to the subfamily and only Herpestes among the ten genera listed by Simpson (1945) has a fossil record, which in Europe extends into Upper Oligocene. The lack of differentiation of distinct fossil genera reflects difficulties of establishing diagnostic characters. Genus KICHECHIA nov. DiAGnosis. Herpestine with upper dental formula 3.1.4.2. Teeth not com- pressed ; canine long and slender; parastyle present only on P4; upper molars without conules and without hypocone; protocone crescentic and without anterior and posterior wings. TYPE SPECIES. Kichechia zamanae sp. nov. REMARKS. Only the holotype and isolated teeth or partial dentitions are known and they possess no characters which would preclude them from the Viverravinae. The sum of the dental characters is diagnostic, though individually several of them are to be found in other genera. Kichechia zamanae gen. et sp. nov. (Pl. 5, fig. 3; Text-figs. 56-60) Diacnosis. The only known species, diagnosis as for genus. The name is derived from the Swahili word kichechi, a mongoose, and zamani meaning ancient. HoiotyPe. M.19077a, b. Facial region of skull and anterior part of braincase with complete upper dentition on right side except P?. LOCALITY. PARATYPES. M.19078 Right mandible with Canine root; P, 4; root of M,. Rusinga Island. M.19079 Right M,. Rusinga Island, Kavirondo Gulf, Kenya. M.19080 Right mandible with P,; M, 4. Songhor, Kenya. MIOCENE CARNIVORA OF EAST AFRICA Site R 1, Rusinga Island, Kavirondo Gulf, Kenya. ADDITIONAL MATERIAL. From Rusinga Island, Kavirondo Gulf, Kenya. CMF.4003 CMF.4004 CMF.4006 CMF.4008 CMF .4009 CMF.4010 CMF. 4011 CMF.4012 CMF .4014 CMF.4015 CMF.4016 CMF.4017 CMF.4029 CMF.4030 CMF.4031 CMF .4032 CMF.4033 CMF .4034 CMF.4035 CMF.4036 CMF.4037 CMF.4071 CMF.4072 CMF.4074 CMF.4075 CMF.4076 CMF.4077 CMF.4078 CMF.4005 From Moruorot, Northern Frontier District, Kenya. CMF.4013 Right mandibular fragment with P, root, Pg 4, My. Right M1. Left M1land alveolus of P4. Site 12. 297 Site 2, Left mandible with C root, P, alveolus, Py 4, My. Site 1. Left mandibular fragment with P,, M,. Left mandibular fragment with C, P, roots, Pg 4. Left mandibular fragment with M,. Site 1. Right mandibular fragment with P,, M,; roots of P, ,and Mg. Site I. Right mandibular fragment with P, ,, M,; roots of P, and C. Right M,. Right mandibular fragment with P, ,, M,; roots of Py and Mg. Right mandibular fragment with P,, M,; roots of P, and Mg. Site Ia. Right mandibular fragment with P,; roots of M, 5. Right mandibular fragment with P,, M,; roots of Mg. Site 1. Left mandibular fragment with M,, broken P,. Left mandibular fragment with P, and broken M,. Left mandibular fragment with M,; root of Mg. Left mandibular fragment with P,_,. Beit ite Site 6: Right M,. Right P,. Left P4. Site r. Right maxillary fragment with P4, M! 2, Left P4. Left upper canine. Left mandibular fragment with P,, M,. Left mandibular fragment with Pg. Left M, in mandibular fragment. Left P,_, in mandibular fragment. From Mfwanganu Island, Kavirondo Gulf, Kenya. Right M1. 298 MIOCENE CARNIVORA OF EAST AFRICA From Songhor, Kenya. CMF.4073 Anterior facial region of skull with nasals, maxillae, frontals, palatines; no teeth. From Napak, Karamoja, Uganda. UMP64.35 Left mandibular fragment with C and roots of P, 4. Napak V. UMP64.34 Right M,. Napak IV. DescriPTION. In the holotype the facial region of the skull is complete but is broken off about the fronto-parietal junction and the whole of the posterior is missing. The skull is fractured and partly displaced, but may have had size and proportions similar to a living mongoose. The face is long, low and narrow. The premaxilla has a long, tapering ascending ramus which reaches back to the line of P!. The maxilla is almost wholly vertical on its outer face and posteriorly carries the stout base of the zygomatic arch: the infra-orbital foramen is smaller than the canine alveolus and lies almost immediately above P3. The frontal bones are flattened dorsally. The extremities of the post- orbital processes are broken on each side, but from their roots it can be judged that they were well developed: since the zygoma is broken off near its anterior root, it is not possible to estimate how fully the orbit was enclosed posteriorly. From the postorbital process a ridge sweeps posteriorly toward the mid-line; these two ridges meet and continue medianly backward, but do not form a true sagittal ridge; the ridge indicates the upper limits of the origin of the temporal muscles and their meeting medianly suggests powerful musculature to the mandible. The skull is very constricted immediately behind the postorbital processes, narrowing to 9 mm. after which it expands rapidly to 20 mm. width; posteriorly to this it is missing. The anterior palatine foramen is about the size of the alveolus of I3, lying near the median plane between I? and C in the narrow pre-maxillary region. The posterior palatine foramina are smaller and lie opposite the posterior end of P,. The palatine bone is almost the same length as the palatine portion of the maxilla, extending backward 18 mm. from the anterior edge of P4: the pterygoid process is broken. The horizontal ramus of the mandible is preserved in one of the paratypes (M.19078): the bone is slender in transverse section and relatively shallow dorso- ventrally compared with its length. The teeth are closely packed; a large mental foramen occurs below P, and there are several smaller ones posteriorly. DENTITION. The dental formula is 2++4:5. The holotype has all three incisors and the canine on each side, together with P2, 3+4, M1*? on the right side. The three incisors lie transversely on a slight curve, all close together, I? is slightly larger than I1, and I is much larger than I?: I! and I? are spatulate. I?is conical with a groove ~ cutting postero-buccally across it and worn by friction with a ridge on the antero- lingual border of the lower canine. A diastema 3.1 mm. long separates I? from C. The canine is long, slender, gently tapering and slightly curved: in transverse section it is ovoid, more flattened lingually than buccally: anteriorly the tooth is rounded and posteriorly keeled: there is a slight ridge on the antero-lingual margin. P! follows immediately behind C without any gap: the tooth is absent on both MIOCENE CARNIVORA OF EAST AFRICA 299 Fias. 56,57. Kichechia zamanae gen. etspnov. Facial region, left and right sides united in reconstruction. (56) Occlusal aspect. (57) Lateral aspect. Holotype (M.19077), Rusinga Is. x2. sides and only the small single alveolus remains. P?is a two-rooted tooth with single cusp, whose height is equal to its antero-posterior width at the level of the continuous basal cingulum. P? is slightly larger than P?; cusp height is again equal to antero- posterior width at the base of the crown; the cingulum is continuous and most prominent posteriorly. P* is relatively broad and stout with the carnassial shear oblique: the large prominent paracone continues anteriorly into a small parastyle: the protocone is a low cusp, well developed; the metacone short and trenchant, intermediate in height between paracone and protocone: a cingulum is present buccally and posteriorly. M! is transversely broad; paracone and metacone are 300 MIOCENE CARNIVORAOF EAST AFRICA equally developed; parastyle absent, but buccal cingulum present; protocone is symmetrical, crescentic, and bounded internally by a cingulum. M?is very similar to M! but smaller; cingulum less well developed. The continuity of bone behind M2 testifies to the complete absence of M3. In the mandibular dentition no incisors are known. Only the root of the canine is preserved and this extends posteriorly under P,,,. In section the canine is roughly ovoid at the base and smaller than the upper canine. P, is unknown, but the single small alveolus indicates an almost vestigial single cusped tooth. On P, the cusp is asymmetrical, more steeply inclined and smaller on the anterior half; a cingulum is present posteriorly and there is the trace of one anteriorly. P.,is an enlarged version of P,; its cusp height is approximately equal to its antero-posterior length; the posterior cingulum is well marked but the anterior one is small; a minute accessory cusp 1s sometimes present on the posterior keel of the main cusp. P, is larger than Iics. 58-60. Kichechia zamanae gen. et sp. nov. Mandibular dentition, P,-M,; composite reconstruction based on M.19078, M.19079 and M.1g080. (58) Occlusal aspect. (59) Lateral aspect. (60) Medial aspect. x3. MIOCENE CARNIVORA OF EAST AFRICA 301 P.; the main cusp rises to about the same level as that of P, and is steeper buccally than lingually; a small cingulum is present on the antero-lingual extremity. The posterior slope of the main cusp carries a prominent accessory cusp on the buccal side while postero-lingually is developed a low cingulum. On M, the protoconid is the highest of the three cusps on the trigonid, with the paraconid more robust than and very slightly higher than the metaconid: the buccal border of the protoconid and paraconid is trenchant and the two cusps are separated by a deep notch: the talonid is about the same length as the trigonid; the prominent hypoconid is separated from the trigonid by a deep cleft and the lingual border of the talonid is fringed by a slightly crenulated cingulum. M, is present only on one specimen (M.1g080) and on this is damaged: it is a small two-rooted tooth with apparent low protoconid and metaconid; the talonid is slightly larger than the rest of the tooth, carrying a hypoconid buccally which continues lingually as a low cingulum. REMARKS. Among the living herpestines, the dentition of Bdeogale is closest to Kichechia, both animals being about the same size. In Bdeogale, 1? is larger than in Kichechia, while the upper canine of the fossil genus is slightly larger and curved posteriorly unlike the straight canine of Bdeogale. P? in the living genus is large, with an internal cusp not found in Kichechia. P4is very similar in both genera,the metastyle being slightly larger and the protocone more anteriorly placed in Kichechia. The first upper molars are essentially similar, the fossil form having a slightly less prominent external cingulum and more prominent internal one. The proportion of M1/M? is alike in the two genera. In the mandibular dentition the premolars are similar, but the molars display differences. M, trigonid in Bdeogale is unusual, the metaconid being connate with the paraconid, and the protoconid and paraconid being separated by a trough, at the base of which arises an incipient cusp (paraconulid); the talonid has a prominent hypoconid: M, is proportionately much larger than in Kzchechia with well developed trigonid and talonid. These differences in the lower dentition are striking, but two points of importance are first, the lower teeth referred to Kichechia are not asso- ciated with maxillary parts in the same specimen and hence their relationship is only an inferred one; secondly the mandibular molars of Bdeogale are highly exceptional, differing from other genera of herpestines. The essential character of the dentition of Kichechia—the formula, tooth pro- portions, cusp development and degree of sectorial development—all suggest close affinity with Herpestinae. The Viverrinae are more specialized in the sectorial direction than the Herpestinae, but on dental characters alone Kichechia could represent the common stock from which both lines evolved. Comparison with the Miacidae, especially the Viverravinae, is valid in terms of dental formula and tooth structure, but differs in detail. Kzchechia lacks the strong parastyle on M! and the high M, trigonid with weak talonid so characteristic of viverravines. Kichechia has no close affinity with the Stenoplesictinae, the only other Tertiary viverrids. Kichechia is the earliest known example of a viverrid in Africa, the next record being in the Pleistocene. 302 MIOCENE CARNIVORA OF EAST AFRICA TABLE 9 Dental measurements (in mm.) of Kichechia zamanae C ips Ip pe Pe M1 M2 M.19077b ant-post 4°1 — 4°0 4°4 6:6 4°4 3°2 lat 2°8 _— 2°2 3°7 6:0 6°9 5°4 12a IP Vea M, M, CMF.4006 ant-post 3°8 4°5 5°60 65 lat 2°1 2°7 3°3 3°8 M.19078 ant-post 3°60 4°4 5°8 lat 2 2°9 3°4 M.19079 ant-post 6°7 lat 3°7 M.19080 ant-post 4°0 lat 2°8 Family FELIDAE Gray Diacnosis. ‘‘Aeluroidea, primitively with long skull, becoming progressively shorter, especially the face; rather inflated braincase; alisphenoid canal only present in primitive forms; entotympanic portion of auditory bulla very large, separated by a high septum (occasionally doubtfully so) from the laterally placed, smaller ecto- tympanic; external auditory meatus short; paroccipital process separated from the mastoid process, stretched out against the hinder part of the bulla; dental formula g--3-4-7: canines strongly developed; M, with two converging blades developed from paraconid and protoconid; primitive forms with strong metaconid, progressively becoming fused with protoconid, talonid only present in primitive forms, trenchant, progressively disappearing. M+ and M, always small; premolar series progressively reduced; humerus usually with entepicondylar foramen; extremities relatively long and slender, digitigrade; manus 5-dactyl; pes generally 4-dactyl; claws retractile, except in Acinonyx and allied genera; os penis rudimentary.” (Pilgrim 1931). REMARKS. The division of the family into four subfamilies as given in Simpson (1945) is adopted here. Subfamily NIMRAVINAE Trouessart Diacnosis. Felids with large incisors; upper canine enlarged and lower canine normal or slightly reduced; carnassial teeth deeply notched; P* large, P* with well developed protocone, strong paracone and parastyle present; anterior premolars absent or vestigial. REMARKS. The large upper canines and incisors distinguish the subfamily less from the Felinae than the deep notches on the carnassial teeth distinguish it from the Machairodontinae. Scott & Jepsen (1936) erected the subfamily to accommodate Archaelurus and Nimravus and specifically excluding Pseudaelurus and Metailurus. Teilhard de Chardin (1945) suggested the erection of the subfamily Pseudaelurinae to accommodate Pseudaelurus and Metailurus. Simpson (1945) placed all the above MIOCENE CARNIVORA OF EAST AFRICA 303 genera in the Nimravinae. The latter grouping is followed here giving a total of ten genera; Adlurictis and Dinailurictis from the European Eocene and Oligocene: Dinictis, Nimravus, Dinaelurus, Archaelurus and Pogonodon from the Oligocene and Lower Miocene of North America; Pseudaelurus from the Miocene of Europe and North America; Metailurus from the Upper Miocene of Europe and Asia. Kitts (1958) erected the genus Nimravides to accommodate the North American Pliocene species Pseudaelurus thinobates. The subfamily shows features in advance of the Proailurinae and probably includes ancestral stocks of both felines and machairo- dontines. Genus METAILURUS Zdansky 1924 Diacnosis. Nimravine with P1+? absent, P? large and P* with strong paracone and well developed parastyle: P, if present very reduced. TYPE SPECIES. Metailurus major Zdansky. In addition to the type species, Zdansky (1924) described another species, M. minor from the same Pontian beds of China. Colbert (1939) described M. mongoliensis from the Vindobonian of Mongolia, and Thenius (1951) transferred Felis letodon Weithofer to the genus as M. parvulus. Andrews (1914) described a mandible from Karungu, Kenya as Pseudoaelurus africanus and below this is transferred to the genus Metailurus. Remarks. Matthew (1929: 496) wrote “Metailurus does not seem to me to be separable generically from Pseudaelurus, although it represents an intermediate stage between that genus (typically) and Felis. Nor do I find any reason for removing the American species from Pseudaelurus, with the typical species of which they agree more nearly than they do with the types of Metailurus.”’ Stock (1934) summarized the characters of Metailurus as follows:— ““Metailurus Lower Pliocene. Dentition $4:5:+ P*4 with well developed para- style. M, with heel considerably reduced. Diametral index of superior canine, 63.8 (M. major), 66.3 (M. minor). Anterior end of mandibular ramus without flange or angulation. Condylar and carotid foramina closely connected with foramen lacerum posterius. No alisphenoid canal. Tympanic bulla completely ossified.”’ Taken together with his list of characters for Pseudaelurus, the generic distinction is perfectly clear. Metailurus is at present better known than Pseudaelurus; though represented by fewer species, they are much more complete than anything known of Pseudaelurus. Pseudaelurus may be distinguished from Metailurus by having 3-4 premolars and on P* the parastyle is weak. The progressive trends from Pseudaelurus through Metailurus to Felis are the reduction of the anterior premolars and the reduction of the protocone with corresponding increase in the size of the parastyle on P4. The size changes in the upper canine do not appear to follow a definable trend. Metailurus represents an intermediate stage morphologically between Pseudaelurus 304 MIOCENE CARNIVORA.OF EAST AFRICA and Felis and it seems preferable to retain this generic distinction; Pseudaelurus for the European and American species, Metailurus for the Asiatic species. This holds true for all but two species, “F’elzs’’ lecodon and Pseudaelurus africanus, the latter being discussed below. A fragment of a right mandibular ramus from Pikermi was described by Weithofer (1888) as ‘Felis’ letodon: this has been shown by Thenius (1951) to be conspecific with another mandible from Pikermi described by Hensel (1862) as Machairodus parvulus. On the strength of a newly described maxillary dentition from Pikermi, Thenius regarded all Pikermi specimens as generically comparable with Metailurus minor from the Chinese Pontian and in consequence has renamed “‘Felis’’ letodon as Metailurus parvulus (Hensel). Teilhard de Chardin (1945: 18-23) referred to “Metailurus tunggurensis Colbert 1939, p. 78, fig. 18’’: this is a mistake for Metailurus mongoliensis. Metailurus africanus (Andrews) (Pl. 5, fig. 4; Text-figs. 61, 62) 1914 Pseudaelurus africanus Andrews: 178-179, pl. 29, figs. 1a, b. Diacnosis. Dental formula #27; a Metailurus intermediate in size between M. major and M. minor. The anterior process of the nasal bone elongate: upper canine somewhat more ovate than in M. major: P* paracone and metacone equal length; large maxillary and mandibular canine-premolar diastema; vestigial P, present. HoLotyPe. M.10634. Left mandibular ramus with Ig, C, P3;4, from Bed 31 at West Kachuku, Karungu, Victoria Nyanza, Kenya. ADDITIONAL MATERIAL. M.19076. Facial region of skull, the maxillae with all dental alveoli and P#+4 present, described and figured below. From Site 18, Rusinga Island, Kavirondo Gulf, Kenya. CMF.4001 Isolated P, from Songhor. DescriPTIon. Most of the anterior of the skull is preserved though much crushed. The premaxilla has a long ascending ramus which probably almost reached the frontal; centrally the anterior palatine foramen is about the same size as the alveolus of 1*. The convex surface of the maxilla is evidence of the deep roots for the canine tusks. The infra-orbital foramen is drop-shaped, its height being about half the length of the canine alveolus: the lower border of the foramen is situated about 1.5 cms. above the base of the main cusp of P®. The nasal bone is long and broad, with a prominent anterior descending ramus overlapping the premaxilla. The frontal bones, though incomplete, suggest a narrow interorbital region: the post-orbital process was short and from its posterior border arises the ridge demarcating the MIOCENE CARNIVORA OF EAST AFRICA 305 anterior limit of the temporal muscle; this ridge ascends rapidly and meets the sagit- tal line about 1.5 cms. behind the postorbital process. The jugal is very robust; posteriorly it is broken and cannot be traced beyond the orbit. P3+4 are preserved on both sides but only the alveoli of the other teeth remain. The alveolus of I? is slightly larger than that of I! and much compressed laterally. The alveolus of I*is very much larger than that of I?, more or less circular, and on the right side contains the tooth root. The three incisor alveoli lie close together in an arc and are separated by a short diastema from the large oval canine alveolus. The canine alveolus measures 16.2 mm. antero-posteriorly and the maximum transverse width is 9.0 mm. The canine js separated from the premolar series by a diastema almost as long as the canine alveolus. There is no trace on either side of any pre- molar anterior to the bi-rooted P?. The prominent central cusp of P* has an anterior keel terminating in a small anterior cusp: the posterior half of the central cusp is broken on both teeth: the posterior cusp is larger than the anterior one and a cingulum terminates the tooth posteriorly. In P* the paracone is slightly higher than the metacone and equal to it in length: internally the two cusps are sectorial and externally are separated by a deep trough: the metacone terminates in a ridge, the paracone in a point: the parastyle is in line with the metacone and paracone and is larger than the protocone. The alveolus of M! indicates a small bi-rooted tooth lying transversely close behind P4. RemArkKS. From the table of measurements on the dentition, WM. africanus can be seen to be comparable in size with M. mongoliensis and M. parvulus, and inter- mediate between M. major and M. minor. From the alveolus, the upper canine of M. africanus appears to have been intermediate in size between M. major and M. minor, though more ovoid than either of these two species and in this character similar to M. parvulus. P?+4 are proportionately more similar to those of M. major than to other species. The equality of paracone and metacone length on P# allies the species to the Chinese forms and differentiates it from M. parvulus. The incisor- canine diastema is of similar size in M. africanus and M. major and much larger in the smaller species M. minor. The canine-premolar diastema in M. africanus is much larger than that of any other species. The anterior process of the nasal is longer in African species than in either of the two Chinese species. The holotype mandible described by Andrews fits the above skull perfectly; the size and spacing of the teeth correspond precisely. P, in the holotype must have been minute judging from the pinhole alveolus; thus the effective diastema extended from canine to P, and the outer concavity of the mandible in this region gave room for the long upper canine when the jaw was closed, a feature better developed in M. major than in M. minor. The additional tooth referred to the species (CMF.4001) is a P,. It measures 5-6 mm. laterally and 11.5 mm. antero-posteriorly. The tooth is indistinguishable in character from P, on the holotype, but is slightly smaller; their ratios of length to breadth are identical. MIOCENE CARNIVORA OF EAST AFRICA 306 ee a ’ aM 1 ) Y 7 G3 y f ' 62 Facial region, distortion corrected. (M.19076), Rusinga Is. x1. Metailurus africanus (Andrews) (62) Lateral aspect. Fics. 61, 62. (61) Occlusal aspect. MIOCENE CARNIVORA OF EAST AFRICA 307 The stratigraphic distribution of the species of Pseudaelurus and Metailurus is as follows: Pseudaelurus quadridentatus Gervais Metailurus (type species) lorteti Gaillard transitorius Depéret tournauensis (Hoernes) marint Villalta & Crusafont aluroides MacDonald pedionomus MacDonald intrepidus Leidy marsht Thorpe martint (Hibbard) kansensis (Hibbard) major Zdansky (type species) minor Zdansky parvulus (Hensel) mongoliensis Colbert africanus (Andrews) Vindobonian, Europe Vindobonian, Europe Vindobonian, Europe Vindobonian, Europe Vindobonian, Europe Barstovian, N. America Clarendonian, N. America Barstovian-Clarendonian, N. America Clarendonian, N. America Hemphillian, N. America Hemphillian, N. America Pontian, Asia Pontian, Asia Pontian, Europe Sarmatian, Asia ““Miocene’’, Africa The European Pseudaelurus species are all Vindobonian in age and the North American species higher, ranging from the Barstovian to the Hemphillian. The Mongolian species of Metailurus is Sarmatian (Tung Gur formation), the Chinese and Pikermi species are all Pontian. Pseudaelurus is more primitive, occurs earlier in the stratigraphic record and appears to persist longer than Metailurus. The species and distribution of both genera are, however, not sufficiently abundant to enable any firm stratigraphic conclusions to be drawn for the African occurrence, save to suggest that Middle to Late Miocene is likely. 308 MIOCENE CARNIVORA OF EAST AFRICA TABLE I0 Measurements (in mm.) on dentitions of Metailurus *—alveolus measured Metailurus Metailurus Metailurus Metailurus Metailurus africanus major minor mongoliensis parvulus (Andrews) Zdansky No. 3+4 AM.26599 ex. Thenius M.19076 Pontian: Zdansky Colbert Pontian: and M.10634 China Pontian: Miocene: Pikermi Miocene: China Mongolia Kenya i lat — 3°0 2°8 — Ev aa AW 3°6 a i12 lat 3°2* 4:0 3°5 — a—p 5:0* 5°4 4°3 = 18 lat 4°8* 6°7 4°8 — a—p Gp Be 8-2 5:2 — Gc lat 9:0* II°5 7°8 6°9* a—p 16°2* 18°7 12°3 12°9* P3 lat 5°7 8-9 6°6 6°7 a—p 13°0 20°2 13°7 13°5 ID lat 10°3 14°0 10°6 9°5 a—p 21°0 31°2 24°0 21°4 Mt lat I0'0 II°9 10'0 9°3 a—p yo) SS) 47 Art I, lat — BOG] 2 23 — a-p ait 33 2°7 2°0 = I, lat — 3°8 2°9 3°0 — a—p aa 4°2 3°0 2°7 = I, lat 2°8 5°6 4° Be = a-p ays) DIS 3°9 3°3 a € lat 55 9:0 65 13 6°6 a—p Io°l 12°7 8:8 II'5 g'I P3 lat 4°6 8-4 53 5°7 = a—p Iorl 15°5 9°9 12°0 — Py, lat 6:8 953 6°5 6:8 6°5 a—p 13°9 21°O 14°5 I5'0 15'0 M, lat — Iovl 72 7°2 7°0 a—p -— 23°2 18° W763 17°8 Ratios (lateral/anteroposterior) C o>. “61 "65 = oh Bs "44 "44 “48 = "50 Ee "49 "45 "44 = "44 M1 “35m "46 "47 aa "44 C 54 ‘71 "74 65 72 Ps 45 "54 "53 47 a Py 49 "44 45 45 43 Mi = "43 "40 "42 "39 Diastemae Le-€ 4°4 4°3 62 o = C-Ps 13°0 5°8 3°4 = 37 C_-P, 20°8 18°5 84 5°5 Gpo72 MIOCENE CARNIVORA OF EAST AFRICA 309 Til. CONCLUSIONS AND THE AGE OF THE FAUNA The carnivores do not reveal much about the environment, being predators mainly dependent on the herbivores in the fauna. They range from very small species about the size of a stoat to the large hyaena-like Pterodon. The picture of forests on the volcanic slopes, swamp with gallery type vegetation and savannah with flash floods is well described by Chesters (1957) and by Bishop (1963). The carnivores described in this paper are recorded from nine localities. Two of these, Rusinga and Napak, are subdivided into a number of sites and these are quoted where known; unfortunately many of the best finds were made on Rusinga before site designation was initiated. Rusinga Island, Mfwanganu Island and Karungu in western Kavirondo are associated with the Rangwa volcanic centre. Ombo is a high level site in eastern Kavirondo. Songhor and Koru are in Nyanza, east of the Kavirondo Gulf and associated with the Tinderet volcanic centre. Moruorot is in the Northern Frontier District of Kenya and Napak in Karamoja, Uganda. Kaboor is in Turkana, Northern Kenya. Age analysis can be based on three lines of evidence; the relationships of the carnivores to other carnivore faunas, the deductions obtained for other faunal elements and radiometric dating of the fossiliferous tuffs. TABLE II Distribution of the carnivores by sites Elizabethfeldern Fayuim + Napak Moruorot + Mfwanganu + Rusinga Karungu Ombo Maboko Songhor Kaboor Kelba quadeemae Teratodon spekei Teratodon enigmae Anasinopa leakeyi Metasinopa napaki + Dissopsalis pyroclasticus ar Metapterodon kaiseri Metapterodon zadoki Pierodon africanus af ae Ptevodon nyanzae Leakitherium hiwegi Hyaenodon andrewsi + ar Hyaenodon matthewti Hyaenodon pilgrimi Hecubides euryodon + Hecubides macyodon Kichechia zamanae + al. Metailurus africanus a a a a Sinbaln + Koru ++4++++4+4++4+4+44 ++ +444 310 MIOCENE CARNIVORA OF EAST AFRICA The carnivore evidence for stratigraphical dating can be summarized for individual genera. Kelba, if an arctocyonid, has its closest relationships among the Palaeocene and Eocene arctocyonids of North America, which have two late survivors in the Lower Oligocene of Mongolia. Tevatodon finds closest comparison with Quercytherium from the Phosphorites du Quercy (Upper Eocene to Middle Oligocene) of France. Anasinopa is comparable with Sinopa and Tritemnodon from the Middle Eocene of North America and Europe, and is more primitive than Metasinopa from the Fluviomarine Series (Sannoisian, Lower Oligocene) of the Fayim, Egypt, the latter genus being also known from Napak I. The Dissopsalis species from Kaboor is closely comparable with D. carnifex from the Chinji Stage (probably Middle Miocene) of India. Metapterodon from Karungu and Rusinga is also known from Southwest Africa and Stromer (1926) on the basis of this and other faunal elements suggested a similar age for both deposits. Pterodon africanus is known from the Kavirondo sites and from the Sannoisian of the Fayim. The genus Hyaenodon ranges in Europe from Upper Eocene to Upper Oligocene, with more restricted ranges in Asia and North America; one species, H. andrewsi is common to the Sannoisian of the Fayim and to East Africa. Hecubides may be regarded as a primitive ‘Amphicyow’ ; this form genus is recorded from the Middle Oligocene to Late Miocene (Stampian-Pontian). The non-African species of Hecubides are H. lemanensis from the Aquitanian of France and H. americanus from Nebraska (horizon unknown). Kichechia is a herpestine, the subfamily being recorded from Upper Oligocene times in Europe. Metailurus africanus is closely comparable with two Pontian species from China. The evidence is thus equivocal, the creodonts suggesting Oligocene and the fissipeds Miocene dating; since they are of holarctic origin the creodonts might be expected to survive later in Africa and this is borne out by the novel character of the fauna. If a single age is required by other evidence, then Lower Miocene is most likely. Proboscideans and anthracotheres are useful mammalian taxa in comparative age analyses. Andrews (1914) designated a Lower Miocene (Burdigalian) age to the Karungu deposits on the basis of the close affinity of Deinotheriwm hobleyi with D. cuviert from France. The association of a small species of Deinotherium with Gomphotherium angustidens in the Kavirondo is also well known outside Africa from the Burdigalian deposits of Sables de l’Orleanais, France; El Papiol, Spain; Koty- haza, Hungary and Bugti Hills, Baluchistan. G. angustidens occurs at all these sites and the Deinotherium species are virtually indistinguishable. Most of the sites also contain anthracotheres comparable with East African species, but carnivores are poorly represented, usually by Amp/icyon fragments. Burdigalian faunas have been identified in other parts of Africa. In Southwest Africa Stromer (1926) found no proboscideans or anthracotheres, but the carnivore, hyracoid and lagomorph elements support his argument for comparison with Kavirondo sites. The Moghara site, west of Cairo, yielded G. angustidens but no Deinotheriwm and only one carnivore (Hyaenaelurus) (Fourtau 1920). At Gebel Zeltan in Central Libya Deinotherium hobleyt occurs in association with Gomphotherium angustidens, anthracotheres, hyaenodont, felid and canid carnivores (Savage 1965). Recently the two probos- MIOCENE CARNIVORA OF EAST AFRICA 311 cideans have been found at new sites in Algeria and Tunisia. All this evidence strengthens the case for a late Burdigalian age for the East African faunas, at least in part. PALAEOCENE EOCENE OLIGOCENE MIOCENE PLIOCENE Arctocyonidae (for Ivelba) Quercytherium (for Tevatodon) Sinopa (for Anasinopa) Metasinopa Dissopsalis Ptevodon Hyaenodon Amphicyon (for Hecubides) Herpestinae (for Kichechia) Metailurus Pseudaelurus TABLE. 12. Stratigraphic range of genera or nearest taxon where genera nova. Broken line where record doubtful. @ Stratum with closely comparable species. 4 Stratum with identical species. Both Chesters (1957) and Verdcourt (1963) emphasised the uniformity and modernity of the flora and mollusca throughout the succession, supporting the concept of a single biotic assemblage. This evidence, while not directly useful in dating, does not necessarily conflict with the Burdigalian estimate. 312 MIOCENE CARNIVORA OF EAST AFRICA The fauna of Maboko (=Kiboko) Island in the Kavirondo Gulf has been stated to be of two ages; Hopwood (in Shackleton 1951) argued for Burdigalian and Helvetian ages on the basis of the proboscideans; Leakey (72 Whitworth 1958) suggested Vindobonian or Pontian for the younger elements. Recent discoveries by Leakey (1961) at Fort Ternan, a site associated with the Tinderet volcanic centre as are Songhor and Koru, suggest the fauna is of Pontian age. Radiometric dating of rock samples using K-A,4, is not yet complete. Preliminary results for Napak I give 19 million years (Bishop 1964) ; for the basal (Kiahera) series on Rusinga Island 15.3 and for Fort Ternan 14 million years (Evernden et al 1964). The Napak figure would be consistent with a late Burdigalian age. The Rusinga figure seems too young, but the sample gave ages ranging from 15.3 to 167 million years and may not have been from the Kishara Series. My current assessment is that more than one fauna is represented in the Kenya- Uganda Tertiary sites; that one of these is Burdigalian in age with numerous sites in east and west Kavirondo and in Karamoja. A younger fauna is present at Fort Ternan; there is a hint of a post-Burdigalian fauna at Maboko and Kaboor, and possibly at Rusinga, Songhor and Karungu, mainly on basis of Metailurus and Dissopsalis. For the younger fauna a Vindobonian (Middle Miocene) and or Pontian (Upper Miocene) age is probable. IV. REFERENCES ANDREws, C. W. 1903. Notes on an expedition to the Faytim, Egypt, with descriptions of some new mammals. Geol. Mag., London (4) 10 : 337-343, 2 figs. —— 1906. A descriptive catalogue of the Tertiary Vertebrata of The Fayum, Egypt. xxxvii+324 pp., 26 pls. British Museum (Natural History), London. 1914. On the Lower Miocene Vertebrates from British East Africa, collected by Dr. Felix Oswald. Quart. J. Geol. Soc. Lond., 70 : 163-186, pls. 27-29. ARAMBOURG, C. 1961. Note préliminaire sur quelques Vertébrés nouveaux du Burdigalien de Libye. C. R. Soc. géol. Fy., Paris, 1961 : 107-109, 1 fig. BisHop, W. W. 1963. The later Tertiary and Pleistocene in Eastern Equatorial Africa. In Howe tt, F. C. & BourRLIERE, F. (Editors) African Ecology and Human Evolution. Chicago. 1964. Mammalia from the Miocene volcanic rocks of Karamoja, East Africa. Proc. Geol. Soc. Lond., 1617: 91-94. BisHop, W. W. & WuvytTe, F. 1962. Tertiary mammalian faunas and sediments in Karamoja and Kavirondo, East Africa. Nature, Lond., 196 : 1283-1287, 2 figs. BLAINVILLE, H. M. D. de. 1839. Sur les insectivores fossiles d’Auvergne. Amn. franc. étrang. Anat. Phys., 3 : 60. 1841. Ostéogvaphie des mammiferes vécents et fossiles, 2. Carnassiers. Paris. Butter, P. M. 1946. An arctocyonid from the English Ludian. Ann. Mag. Nat. Hist., London (11) 13 : 691-701, 2 figs. CuHESTERS, K. I. M. 1957. The Miocene Flora of Rusinga Island, Lake Victoria, Kenya. Palaeontographia, Stuttgart, 101 B : 30-71, pls. 19-21. CHARDIN, P. Teilhard de. 1945. Les Félidés de Chine: les formes fossiles. Jn CHARDIN, P. T. de & Leroy, P. Les Félidés de Chine. Publ. Inst. Géobiol., Pekin, 11 : 1-58, 21 figs. Crark, W. E. Le G. & Leakey, L. S. B. 1951. The Miocene Homionidea of East Africa. Fossil Mammals of Africa, 1 : 117 pp., 9 pls. British Museum (Natural History), London. CoLBertT, E. H. 1933. The skull of Dissopsalis carnifex Pilgrim, a Miocene Creodont from India. Amer. Mus. Novit., New York, 603 : 1-8, 4 figs. MIOCENE CARNIVORA OF EAST AFRICA 313 CoLBERT, E. H. 1939. Carnivora of the Tung Gur Formation of Mongolia. Bull. Amer. Mus. Nat. Hist., New York, 76 : 47-81, 19 figs. Corr, E. D. 1885. The White River beds of Swift Current River, Northwest Territory. Amer. Nat., Boston, 19 : 163. CrusaFont, M., Vitratta, J. F. & TRuyots, J. 1955. El Burdigaliense continental de la cuenca del Vallés-Penedés. Mem. Comun. Inst. Geol. Barcelona, 12 : 1-272, pls. 1-11. Deum, R. 1935. Ueber tertiare Spaltenfiillungen im Frankischen und Schwabischen Jura. Abh. bayer. Akad. Wiss., Munchen, 29 : 1-86, pls. 1-5. Denison, R. H. 1938. The broad-skulled Pseudocreodi. Ann. N. Y. Acad. Sct., 37 : 163-256, 32 figs. DepéreT, A. 1892. La faune de mammiféres miocenes de la Grive St. Alban (Iseére) et de quelques autres localités du bassin du Rhéne. Arch. Mus. Hist. nat. Lyon, 5, 2 : 1-95, pls. 1-4. 1917. Monographie de la faune de mammiferes fossiles du Ludien inférieure d’Euzet-les- Bains (Gard). Ann. Univ. Lyon (1) 40 : 1-288, pls. 1-25. Dove tass, E. tgo1. Fossil mammalia of the White River beds of Montana. Tvans. Amer. Phil. Soc., Philadelphia (2) 20 : 237-279, pl. 1. EVERNDEN, J. F., SavaceE, 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: 145-108. Firuor, H. 1872. Recherches sur les mammiferes fossiles des dépdts de phosphate de chaux dans les départements du Lot, du Tarn et de Tarn-et-Garonne. Amn. Sci. géol., Paris, 3, 7 : I-31, pls. 13-19. 1876. Recherches sur les phosphorites du Quercy. Bibl. Ec. haut. Etud., Paris, 15 : 1-220, pls. 1-27. 1881. Etude sur les mammiferes fossiles de Ronzon (Haute-Loire). Bzbl. Ec. haut. Etud., Paris, 24 : 1-270, pls. 1-26. FORSTER Cooper, C. 1926. Hyaenodon aimi sp. n. from the Headon beds at Hordle. Ann. Mag. Nat. Hist., London (9) 18 : 370-373, 1 fig. Fourtau, R. 1920. Contribution a l’étude des vertébvés miocénes de l’Egypte. 121 pp., 3 pls. Egypt. Survey Dept., Cairo. GAILLARD, C. 1899. Mammiféres miocénes nouveaux ou peu connus de la Grive St. Alban (Isere). Avch Mus. Sci. nat. Lyon, 7, 2 : 1-79, pls. 1-3. Gazin, C. L. 1946. Machaeroides eothen Matthew, the saber-tooth creodont of the Bridger Eocene. Pyvoc. U.S. Nat. Mus., Washington, 96 : 335-347, pls. 45, 46. Gervais, P. 1846. Mémoire sur quelques mammiferes fossiles du département de Vaucluse. C. R. Acad. Sci. Paris, 22 : 845-846. 1848-52. Zoologie et paléontologie frangaises. vill+271 pp. Atlas 150 pp., 80 pls. Paris. 1876. Zoologie et paléontologie generales. 2° ser. 72 pp., 12 pls. Paris. GrInsBuRG, L. 1955. De la subdivision du genre Hemicyon Lartet. (Carnassier du Miocene). Bull. Soc. géol. Fr., Paris (6) 5 : 85-99, 6 figs. Grecory, W. K. & Herrman, M. 1939. On the evolution and major classification of the civets (Viverridae) and allied fossil and recent Carnivora; a phylogenetic study of the skull and dentition. Pyvoc. Amer. Phil. Soc., Philadelphia, 81 : 309-392, pls. 1-6. HENSEL, R. F. 1862. Uber die Reste einiger Sdugethierarten von Pikermi in der Miinchener Sammlung. Mber. k. preuss. Akad. Wiss. Berlin, 14 : 560-569, pl. 1. HIBBARD, C. W. 1934. Two new genera of Felidae from the Middle Pliocene of Kansas. Tvans. Kans. Acad. Sci., Topeka, 37: 239-255, pls. 1-3. HoErNES, R. 1881. Vorlage von Sdugethierresten aus den Braunkohlen-Ablagerungen der Steiermark. Verh. geol. Reichsanst. Wien, 1881 : 338-339. Houeu, J. R. 1948. The auditory region in some members of the Procyonidae, Canidae and Ursidae. Bull. Amer. Mus. Nat. Hist., New York, 92 : 67-118, pls. 9-15. 314 MIOCENE CARNIVORA-OF EAST AFRICA HURZELER, J. 1944. Zur Revision der europaischen Hemicyoniden. Verh. naturf. Ges. Basel, 55 : 131-157, 17 figs. Kent, P. E. 1944. The Miocene Beds of Kavirondo, Kenya. Quart. J. Geol. Soc. Lond., 100 : 85-116, pls. 6, 7. Kitts, D. B. 1958. Nimyvravides, a new genus of Felidae from the Pliocene of California, Texas and Oklahoma. J. Mammail., Baltimore, 39 : 368-375, pl. 1. LaizER, L. de & de PartEv. 1838. Description et détermination d’une machoire fossile, Hyaeno- don leptorhynchus. C. R. Acad, Sci. Paris, 7 : 442. LEeakeEy, L.S. B. 1961. A new Lower Pliocene fossil primate from Kenya. Ann. Mag. Nat. Hist., London (13) 4; 689-696, pl. 18. Lripy, J. 1853. Remarks on a collection of fossil mammalia from Nebraska. Pyvoc. Acad. Nat. Sci. Philad., 6 : 392-394. —— 1858. Notice of remains of extinct vertebrata, from the valley of the Niobara River, Nebraska. Proc. Acad. Nat. Sci. Philad., 1858 : 20-29. Loomis, F. B. 1936. Three new Miocene dogs and their phylogeny. J. Paleont., Chicago, 10 : 44-52, 6 figs. LYDEKKER, R. 1884. Siwalik and Narbada Carnivora. Palaeont. indica., Calcutta (10) 2 :178— 351, pls. 1-20. Macponatp, J. R. 1948. A new species of Pseudaelurus from the Lower Pliocene of Nebraska. Univ. Calif. Publ., Geol. Sci., 28 : 45-52, 4 figs. 1954. A new Pseudaelurus from the Lower Snake Creek fauna of Nebraska. J. Paleont., Chicago, 28 : 67-69, 1 fig. Martin, R. 1906. Revision der obereocaenen und unteroligocaenen Creodonten Europas. Rev. suisse Zool., Geneve, 14 : 405-600, pls. 1-4. Mattues, H. W. 1952. Die Creodontier aus der mitteleozdnen Braunkohle des Geiseltales. Halle. Jb. mdtsch. Evdgesch., 1 : 201-240, pls. 15-40. MattHew, W. D. 1909. The Carnivora and Insectivora of the Bridger Basin, Middle Eocene. Mem. Amer. Mus. Nat. Hist., New York, 9 : 291-567, pls. 42-51. 1924. Third contribution to the Snake Creek fauna. Bull. Amer. Mus. Nat. Hist., New York, 50 : 59-210, 63 figs. 1929. Critical Observations upon Siwalik Mammals. Bull. Amer. Mus. Nat. Hist., New York, 56 : 437-560, 55 figs. 1937. Paleocene Faunas of the San Juan Basin, New Mexico. Tvans. Amey. Phil. Soc., Philadelphia, 30 : 1-510, pls. 1-65. MatrHew, W. D. & GRANGER, W. 1924. New Carnivora from the Tertiary of Mongolia. Amer. Mus. Novit., New York, 104 : 1-09, 7 figs. 1925a. New creodonts and rodents from the Ardyn Obo formation of Mongolia. Amer. Mus. Novit., New York, 193 : 1-7, 9 figs. 1925. New mammals from the Shara Murun Eocene of Mongolia. Amer. Mus. Novit., New York, 196 : 1-11, to figs. OsBorN, H. F. 1909. New carnivorous mammals from the Fayim Oligocene, Egypt. Bull. Amer. Mus. Nat. Hist., New York, 26 : 415-424, 9 figs. 1910. The Age of Mammals in Europe, Asia, and North America. xvii+635 pp., 220 figs. New York. Osporn, H. F. & Wortman, J. L. 1894. Fossil Mammals of the Lower White River Beds. Bull. Amer. Mus. Nat. Hist., New York, 7 : 199-228. PatrersoN, B. & Maccrew, P. O. 1962. A new Arctocyonid from the Paleocene of Wyoming. Breviova, Mus. Comp. Zool. Harvard, 174 : 1-10. Pircrim, G. E. 1910. Notices of new mammalian genera and species from the Tertiaries of India. Rec. Geol. Suvv. India, Calcutta, 40 : 63-71. —— 1914. Description of teeth referable to the Lower Siwalik Creodont genus Dissopsalis. Rec. Geol. Surv. India, Calcutta, 44 ; 265-279, pl. 1. MIOCENE CARNIVORA OF EAST AFRICA 315 Piterim, G. E. 1931. Catalogue of the Pontian Carnivora of Europe in the Department of Geology. vi+174 pp., 2 pls. British Museum (Natural History), London. —— 1932. The Fossil Carnivora of India. Paleaont. indica, Calcutta, 18 : 1-232, pls. I-10. PIVETEAU, J. 1961. Tvaité de Paléontologie, 6, 1. vili+-1138 pp., 1 pl. 970 figs. Paris. Pome, A. 1847a. Note sur des animaux fossiles découverts dans le département de 1’Allier. Bull. Soc. géol. Fr., Paris (2) 4 : 378-385, pl. 4. 1847b. Note sur le Pterodon. Bull. Soc. geol. Fr., Paris (2) 4: 385-393. 1853. Catalogue méthodique et descriptif des vertébrés fossiles de la Loire et de 1’ Allier. An. sci. litt. industy. Auvergne, 26 : 81-229. RussELL, D. E. & McKenna, M.C. 1961. Etude de Pavoxyclaenus, mammifere des phosphorites du Quercy. Buill. Soc. géol. Fr., Paris (7) 3 : 274-282, 2 figs. RUTIMEYER, L. 1891. Die eocane Saugethier-Welt von Egerkingen. Abh. schweiz. paldont. Ges., Zurich, 18 : 1-153, pls. 1-8. SAVAGE, R. J. G. 1965. Two mammal faunas from the early Tertiary of central Libya. Pyoc. Geol. Soc. Lond., 1623: 89-91. ScHLosserR, M. 1887. Die affen, Lemuren, Chiropteren, Insectivoren, Marsupialier, Creodonten und Carnivoren des europaischen Tertiars, I. Beitr. Paldont. Geol. Ost.-Ung., Wien, 6 : 1-227, pls. 1-5. —— 1911. Beitrage zur Kenntnis der oligozinen Landsangetiere aus dem Faytim (aegypten), Ibid, 14: 51-167, 6 pls. Scott, W. B. 1887. Preliminary account of the fossil mammals from the White River formation. Bull. Mus. Comp. Zool. Havv., 13 : 151-171, pls. I, 2. 1894. The Osteology of Hyaenodon. J. Acad. Nat. Sci. Philad., 9 : 499-535, 10 figs. Scott, W. B. & JEepsEeN, G. L. 1936. The Mammalian Fauna of the White River Oligocene. Pt. I. Insectivora and Carnivora. Tvans. Amer. Phil. Soc., Philadelphia, 28 : 1-153, pls. 1-22. SHACKLETON, R. M. 1951. A contribution to the geology of the Kavirondo Rift Valley. Quart. J. Geol. Soc, Lond., 106 : 345-392, pls. 1-6. Stmpson, G. G. 1945. The Principles of Classification and a Classification of Mammals. Bull. Amer. Mus. Nat. Hist., New York, 85 : 1-350. Stock, C. 1933. Hyaenodontidae of the Upper Eocene of California. Pvoc. Nat. Acad. Sci., Washington, 19 : 434-440, pl. I. 1934. Skull and Dentition of the American Miocene cat, Pseudaelurus. Bull. Geol. Soc. Ameyr., Rochester, N.Y., 45 : 1051-1058, pls. 1, 2. STOVALL, J. W. 1948. Chadron vertebrate fossils from below the Rim Rock of Presidio County, Texas. Amer. J. Sci., New Haven, 246 : 78-95, pls. 1, 2 STROMER, E. 1926. Reste land- und siisswasser-bewohnender Wirbeltiere aus den Diamanten- feldern Deutsch-Stidwestafrikas. Jn Kaiser, E. Die Diamantenwiiste Stidwestafrikas, 2. vli+ 535 pp., 48 pls., 32 stereo photos. Berlin. TuHENIuS, E. 1951. Zur odontologischen Charakteristik von “‘Felis’”’ Jeiodon aus dem Pont von Pikermi (Griechenland). N. Jb. Min. Geol. Paldont. Monatsh., Stuttgart, 3 : 88-96, 1 fig. TuHorPE, M. R. 1922. Some Tertiary Carnivora in the Marsh collection, with descriptions of new forms. Amer. J. Sci., New Haven, 3 : 446-447, I fig. VERDCOURT, B. 1963. The Miocene non-marine Molusca of Rusinga Island, Lake Victoria and other localities in Kenya. Palaeontographica, Stuttgart, 121, A: 1-37, 64 figs. VILLALTA, J. F. de & Crusaront, M. 1943. Los vertebrados del Mioceno continental de la cuenca Vallés-Panadés I, II. Insectivoros y Carnivoros. Bol. Inst. Geol. Min. Espana, Madrid, 56 : 145-336, pls. 1-17. WEITHOFER, A. K. 1888. Beitrage zur Kenntniss der Fauna von Pikermi bei Athen. Beitr. Geol. Paldéont. Ost.-Ung., Wien, 6 : 225-292, pls. 10-19. WHItTWoRTH, T. 1953. A contribution to the geology of Rusinga Island, Kenya. Quart. J. Geol. Soc. Lond., 109 : 75-06, pls. 2, 3. 316 MIOCENE CARNIVORA OF EAST AFRICA WuitwortH, T. 1954. The Miocene Hyracoids of East Africa. Fossil Mammais of Africa, 7: 58 pp., 7 pls. British Museum (Natural History), London. 1958. Miocene Ruminants of East Africa. Fossil Mammals of Africa, 15 : 50 pp., 18 figs. British Museum (Natural History), London. 1961. The Geology of Mfwanganu Island, Western Kenya. Overseas Geol. Min. Resources, 8 : 150-190, pl. I. WortMan, J. L. t901. A New American species of Amphicyon. Amer. J. Sci., New Haven, 11 : 200-204, 2 figs. Youne, C. C. 1937. An Early Tertiary vertebrate fauna from Yuanchu. Bull. Geol. Soc. China, Peking, 17 : 413-438, 16 figs. ZDANSKy, O. 1924. Jungtertiare Carnivoren Chinas. Palaeont. sinica, Peking (C) 2, 1 : 1-155, pls. 1-23. IBID RAND, a X12 approx Fic. 1. Kelba quadeemae gen. et sp. nov. Right M2; occlusal aspect. Holotype (M.19087), Rusinga Island. Fic. 2. Tevatodon spekei gen. et sp. nov. Left maxilla with P4-M?; occlusal aspect. Holotype (M.14307), Koru. Fic. 3. Tevatodon spekei gen. et sp.nov. Maxillae with C, P?; occlusal aspect. (M.14310), Koru. Fic. 4. Tevatodon enigmae sp. nov. Facial region; occlusal aspect. Holotype (M.19088a), Songhor. Fic. 5. Tevatodon enigmae sp.nov. Left mandible; occlusal aspect. (M.19089), Songhor. Fic. 6. Anasinopa leakeyi gen. et sp. nov. Left maxilla with P4 and M}; occlusal aspect. Holotype (M.19081a), Rusinga Island. Fic. 7. Anasinopa leakeyi gen. et sp. nov. Right maxilla with M! 2; occlusal aspect. Holotype (M.190816), Rusinga Island. Bull. B.M. (N.H.) Geol. 10, 8 PLATE 1 PIL AILS, 2 X1I°2 approx. Anasinopa leakeyi gen. et sp. nov. Holotype (M.19081c), Rusinga Island. Right mandible with C, P,-M3. Fic. 1. Occlusal aspect. Fic. 2. Lateral aspect. Bull. B.M. (N.H.) Geol. 10, 8 PEATE 2 PLATE 3 XI*2 approx. Dissopsalis pyroclasticus sp. nov. Holotype (M.19082), Kaboor. Right mandible with P,-M,. Fic. 1. Occlusal aspect. Fic. 2. Lateral aspect. Bull. B.M. (N.H.) Geol. 10, 8 PLATE 3 Fic. Fic. Fie. Fic. Fic. Fic. PLATE 4 X1I‘2 approx. Metapterodon kaisert Stromer. Right maxilla with P’-M8; occlusal aspect. (CMF.4038), Karungu. Metapterodon zadoki sp. nov. Right maxilla with M? 2; occlusal aspect. Holotype (M.19094), Rusinga Island. Pievodon africanus Andrews. Left maxilla with P4-M?; occlusal aspect. (M.19090), Napak. Leakitherium hiwegi gen. et sp. nov. Left maxilla with M! 2; occlusal aspect. Holotype (M.19083), Rusinga Island. Leahitherium hiwegi gen. et sp. nov. Left maxilla with P*M1}; occlusal aspect. (CMF.4025), Rusinga Island. Hyaenodon (Isohyaenodon) andrewsi sp. noy. Right mandible with M,-.; occlusal aspect. Holotype (M.15048), Ombo. Bull. B.M. (N.H.) Geol. 10, 8 PLATE 4 PLATE 5 XI°2 approx. Fic. 1. Hecubides euryodon gen. et sp. noy. Maxilla with P?—-M?; occlusal aspect. Holotype (M.19084), Napak. Fic. 2. Hecubides macyodon sp.nov. Left M ;occlusalaspect. Holotype (M.19086), Rusinga Island. Fic. 3. Kichechia zamanae gen. et sp. nov. Facial region of skull, left and right sides; occlusal aspect. Holotype, Rusinga Island. Fic. 3a (M.19077a), Fic. 3b (M.190775). Fic. 4. Metailurus africanus (Andrews). Facial region; occlusal aspect. (M.19076), Rusinga Island. Bull. B.M. (N.H.) Geol. 10, 8 PLATE 5 js VLION 2 poeci96s } | ra | 7) & = ae NG py ; “Se By \ ‘ f , } 4 ae ba d ' JEAGe . 8 PRINTED IN GRE: BY THOMAS DE Ly PEON OAR 88) OMS 9 OMA NES MIT ee a bi Hat :’ ‘ { { i 4 As Ho ben tN ’ ‘) , 1 / r E. B. SELWOOD BULLETIN OF BRITISH MUSEUM (NATURAL HISTORY) GY ‘ Vol. 10 No. 9 - LONDON: 1965. i ae Lt, Crea) Wiles ei % 12 at rs <& ‘s be 2 - DEC 1965 Z df, ony ® MUg DECHENELLID TRILOBITES > FROM@P EE BRITISH MIDDLE DEVONIAN ay ay 2) L we BY EDWIN BRIAN SELWOOD, Ph.D. (University of Exeter) f Pp. 317-333 » 1 Plate ; 6 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 10 No. 9 LONDON : 1965 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), «instituted im 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at wrregular intervals as they become veady. 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. 10, No. 9 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) 1965 ID IRIS, WY 1B IBS}; ©) 18) ANISIID, JERI AN ILS ol NEO S IDOE (ONT ORONIE Tet) aN (O) ike )} Issued November, 1965 Price Eleven Shillings DECHENELLID TRILOBITES FROM THE BRITISH MIDDLE DEVONIAN By E. B. SELWOOD CONTENTS Page I. INTRODUCTION : ‘ : ‘ 0 . : : : 319 Il, ACKNOWLEDGMENTS ‘ : : : : : : : 320 III. VARIATION . . : : : : . 0 : 0 321 IV. AGE OF THE Fauna. : . c : F ; . - 324 V. EcOLOGY AND DISTRIBUTION . : . 5 . 5 : 324 VI. SysTEMATIC DESCRIPTION : . : : : : : 326 VII. REFERENCES . b : : : c : 4 ; a 333 SYNOPSIS Dechenella has a restricted occurrence in South West England. A single species only is known, Dechenella (Dechenella) setosa Whidborne, 1889. This is redescribed and variation seen in the pygidia and cranidia is examined. The species is bimorphic, each bimorphic group of pygidia showing a range of variation that can be correlated with size and hence presumably with age. The species is probably of late Couvinian age. I. INTRODUCTION ALTHOUGH Dechenella is a characteristic Middle Devonian genus with a wide distri- bution, relatively few species are known, and in Britain only one species, Dechenella setosa, has been recognised. This species, unknown outside Devonshire, was first described by Whidborne (1889, 1889a) from the Middle Devonian Limestones at Chercombe (sometimes spelt Chircombe or Cherecombe) Bridge near Newton Abbot, and subsequently redescribed by Richter (1912). Much of the museum material is incompletely localized but from the lithologies it seems that, with rare exceptions, all comes from an extensive disused quarry on the north bank of the River Lemon, 200 yards east of Chercombe Bridge (National Grid Reference: SX832711), where 70 feet to 80 feet of well bedded limestones are still exposed. The individual lime- stone beds, which vary in thickness from a few inches to several feet, are dominantly pale grey in colour and yield a fauna of brachiopods, corals, stromatoporoids and polyzoa. In the higher structural horizons of the quarry there are, interbedded with these pale limestones, horizons of dark bituminous limestone (which are no doubt the “black marbles’ yielding trilobites mentioned by Whidborne 188ga: 28). 320 BRITISH DEVONIAN DECHENELLID TRILOBITES Recent collecting in this quarry has failed to produce any trilobites from the pale coral limestone, but much new material has been obtained from one horizon, 12 inches thick, of black fine-grained limestone exposed in an overgrown section at the entrance to the quarry, a few yards east of the old lime kiln. This is a lithology identical to that of the museum material, and there seems no reason to doubt that the earlier collectors found a similar restricted distribution of trilobites. Two speci- mens in the British Museum (Natural History) are labelled ‘““West Hill, East Ogwell’’ and were probably collected from the largely overgrown quarry south-east of Cher- combe Bridge. Searchingin this quarry has yielded a further pygidium froma lithology identical to that in the Chercombe Bridge Quarry. The West Hill quarry lies only 100 yards along the strike from the latter quarry, and it is most likely that the same horizons are represented in both quarries. A significant difference exists between the old collections and those made recently ; Whidborne based his description of the species upon 37 pygidia and 2 cranidia, and in all the museum material examined a preponderance of pygidia exists, but recent collecting has demonstrated that pygidia and cranidia occur in approximately equal numbers. This strongly suggests bias in the original collecting. As the material is sparsely distributed through the rock, it is likely that much of it was collected by workmen in the quarry as chance finds came to light. Such collecting would undoubt- edly be biased in favour of “‘attractive’” specimens; in this case, the perfectly pre- served pygidia certainly catch the eye much more readily than the dissociated cranidia and free cheeks. Sufficient material is now available to show that there is, particularly in the pygidia, a considerable variation of morphological features. Although comparisons of extreme variants might suggest that more than one species is involved, the variation appears to be continuous and there is no reason to suspect the presence of more than one species. The sample is interpreted as an assortment of individuals of different ages, since it is most unlikely that any of the material was collected through a considerable thickness of sediment. There is a notable absence of larval and small specimens in the museum material, and a similar gap has been found in recent collecting. This suggests that their absence is to be accounted for by some ecological factor, and that only adults migrated into the area. Since the published descriptions do not take account of the variation, and particu- larly since the pygidium described by Richter (1912) is not characteristic, the variation of Dechenella setosa is first described and the species then redefined in more general terms. Il. ACKNOWLEDGMENTS I wish to thank particularly Dr. W. T. Dean and Professor S. Simpson, who read and criticized the original manuscript, and Mr. J. Saunders, technician in the Department of Geology at Exeter, who rediscovered the trilobite band and gave valuable assistance in subsequent collecting. Mr. Saunders is also responsible for the photographs. BRITISH DEVONIAN DECHENELLID TRILOBITES 321 Dr. R. C. Blackie (Exeter City Museum), Mr. A. G. Brighton (Sedgwick Museum, Cambridge), Dr. W. T. Dean (British Museum (Natural History)), Dr. J. D. D. Smith (Geological Survey and Museum), and Dr. F. S. Wallis (Torquay Natural History Museum) kindly permitted the examination and arranged the loan of specimens in their care. Dr. W. Struve allowed me to see type material of Continental species held in the Senckenberg Museum, Frankfurt-am-Main. This part of the work was completed whilst in receipt of a travel grant from the British Council; this is gratefully acknowledged. III. VARIATION Measurements have been made on the pygidia of 44 specimens and, where possible, four standard measurements taken; the breadth of pygidium (Bp), the breadth of axis (Ba), the length of pygidium (Lp), and the length of axis (La) (Text-fig. 1, Table I). TABLE I Standard Mean Deviation Breadth of pygidium 16:79 + 1'479 4°89 Length of pygidium 12:04 + 1:026 3°39 Breadth of axis 4°31 + 0-368 TIO. Length of axis 10°43 + 0°884 2:92 All measurements in millimetres lg La ie Bp a inde ast ! Fic. 1. Standard measurements of cranidia and pygidia of Dechenella setosa. Bc, breadth of cranidium; Bg, breadth of glabella; Lc, length of cranidium; Lg, length of glabella. Bp, breadth of pygidium; Ba, breadth of axis; Lp, length of pygidium; La, length of axis. 322 BRITISH DEVONIAN DECHENELLID TRILOBITES Frequency diagrams of dimensions and ratios of different dimensions are con- sistent with the hypothesis that all specimens can be referred to a single species, since each shows a single well defined peak. 20 20 20 15 154 hs 15 E E E = € E € E ¢ c s 4 a 10 ° 10 a 10 5 5 54 5 y= 2-42x-0-04 y= 1-17x - 0-13 y=2:78x+0-05 2 4 6 8 ae 10 15 2 4 6 8 Ba (in mm) La (in mm) Ba (in mm) = 4 x = b = 5} Q = 4! Fic. 2. Scatter diagrams showing variation in pygidia of Dechenella setosa. The reduced major axis is included for each scatter. 3074 304 30 254 254 204 se E E € iS E is E154 s oy S Q a a a 104 104 sy | y=1:44x- 0-86 " y=1-69x - 0-88 y= 4:02x -0:53 5 10 15 20 5 10 15 2 4°68 Lp (in mm) La (in mm) Ba (in mm) = 1 x 2s Aes Fic. 3. Scatter diagrams showing variation in pygidia of Dechenella setosa. The reduced major axis is included for each scatter. BRITISH DEVONIAN DECHENELLID TRILOBITES 323 Scatter diagrams (Text-figs. 2, 3) relating the length of the standard measurements show, on arithmetic co-ordinates, a markedly rectilinear distribution. The reduced major axis (y=b+kx) has been drawn for all scatters (see Miller & Kahn 1962 : 204), and these clearly indicate that the sample may be interpreted as a single species showing isometric growth. In the absence of young and larval specimens, it is not possible to state the course of the reduced major axis when extrapolated downwards into the smaller size ranges, but from the known ontogenies of trilobites (Palmer 1958) it is probable that the growth here is allometric. TABLE II Standard Standard Dispersion Reduced Correlation evvor of evvoy of avound No. of major axis coefficient slope intercept R.M.A.* specimens Lp/Bp I°44x —o-86 0:94 0:076 0:095 2:062 42 Ba/Lp 2:78x +0:05 0:94 O-141 0-634 I'205 42 Ba/Bp 4°02X —0°53 0-89 0-187 1:287 2°399 42 Ba/La 2°42K —0-04 0:90 0-160 0-174 1°387 44 La/Lp I'I7xX —O'13 0-99 0:024 0:256 0°589 42 La/Bp I-69x —o0:88 0:92 0:090 1'033 2:091 42 * R.M.A. — Reduced major axis When the detailed characteristics of the pygidia are examined, much variation can be observed which is closely related to the size, and hence presumably to the age, of the individuals. At the same time, the sample falls into two groups of approxi- mately equal number, each showing the variation suggested to be associated with age. The two groups may be distinguished by the character of the axial furrows: in Group A they are weakly constricted between rings 7-8, whilst in Group B they are straight (Text-fig. 4). The measured characters of the pygidium are quite independent of the bimorphic characters, specimens referred to Groups A and B show a random distribution within all of the scatters prepared in Text-figs. 2 and 3. Bimorphic variation has also been recorded amongst the cranidia. The glabellas of specimens referred to Group C are more pointed than those of Group D, and 8 on the anterior branch of the facial suture is placed more anteriorly in Group C than in Group D. In side view, the occipital ring is seen to lie below the level of the glabella in Group C, whilst it reaches the height of the posterior part of the glabella in Group D. Ornamentation in the form of a fine granulation has only been observed on the glabella of specimens belonging to Group D (Text-fig. 5). The bimorphism described above is probably sexual but the lack of complete specimens makes it impossible to determine if a correlation exists between the bimorphism seen in the pygidia and cranidia. When comparisons are made with other species (page 331) it is found that, although the range of variation of Dechenella setosa would appear to include a number of species, there is some correlation between those features which distinguish the bimorphic individuals of Dechenella setosa, and those which distinguish the species Dechenella vernemili and Dechenella nittbergensis. 324 BRITISH DEVONIAN DECHENELLID TRILOBITES It is thus tempting to ascribe the cranidium D to pygidium A (both showing verneuila characters) and cranidium C to pygidium B (both showing r7ttbergensis characters). Verification can only await the discovery of complete specimens. om : (coal) a (my) Fic. 4. Pygidia of Dechenella setosa. Group A. (a) Plan view, note constricted axis; (b) Side view; (c) Posterior view. Group B. (d) Plan view, note straight axis; (e) Side view; (f) Posterior view. Both pathological and cicatrized pygidia have been observed in the sample, and it is evident that the species possessed considerable powers of regeneration (BIS, fie 8): IV. AGE OF THE FAUNA Both Calceola sandalina (Linné) and Strvingocephalus burtint Defrance have been recorded from the Chercombe Bridge Quarry (Ussher 1913 : 20). Although neither has been positively identified during the present investigation, Middleton (1959) has confirmed that the limestones are at least in part of middle Middle Devonian age. Richter (1912) suggested that Dechenella setosa was of Givetian age, and this is supported by the acknowledged Givetian age of all the closely related species. However, Scutellum (Scutellum) flabelliferum (Goldfuss) also occurs in the same horizon; this trilobite is a characteristic Couvinian form. A somewhat stronger case can thus be made out for a Couvinian age, though the presence of Sivingocephalus burtimt in the same quarry must indicate an horizon near to the Couvinian/Givetian boundary. The limestones of the quarry have so far failed to yield conodonts. V. ECOLOGY AND DISTRIBUTION Since the trilobites are restricted to the black limestone, it seems that sea floor conditions controlled their distribution and that the species was benthonic. The fine-grained bituminous character of the limestone, the presence of pyrite and the great reduction of coral and stromataporoid growth suggest bottom sediments BRITISH DEVONIAN DECHENELLID TRILOBITES 325 Fic. 5. Cranidia of Dechenella setosa. Group C. (a) Plan view, anterior of glabella pointed; (b) Side view, occipital ring low ; (c) Anterior view. Group D. (d) Plan view, glabella rounded anteriorly, posterior surface of glabella granulated; (e) Side view, occipital ring high; (f) Anterior view. containing a considerable amount of decaying organic matter; such an environment would be well-suited to mud feeding organisms. The absence of young and larval stages from the sample might be explained by these stages being planktonic and settling on to the sea floor only on reaching maturity. Small thin shelled bivalves and the trilobite Scutellum (Scutellum) flabelliferwm occur together with the dechenellid trilobites. Scutellum occurs in a wide variety of lithologies in the Torquay district and was probably planktonic. The distribution of the fauna would not seem to be entirely dependent upon ecological conditions, for comparable lithologies to that yielding the dechenellid trilobites exist both in the Chercombe Bridge Quarry and in many of the other Middle Devonian Limestone outcrops in the South Devon area. To some extent the localization of the fauna may be more apparent than real in that much collecting remains to be done, but at the same time the main limestone outcrops have been extensively quarried in the past and it is surprising that more specimens have not come to light. The variation seen in Dechenella setosa does not suggest that the British species evolved in isolation, for it shows a range of variation allowing close comparisons to be made with species described from the Rhenish Mountains, Morocco, and Bohemia (page 331). It is suggested that the centre of dispersal of European Middle Devonian dechenellids lay some distance outside the British area and that the record at Chercombe Bridge represents an isolated and unusual migration into the British 326 BRITISH DEVONIAN DECHENELLID TRILOBITES area. Since the adult specimens of Dechenella setosa were probably benthonic, the isolated occurrence might be explained by the chance distribution of planktonic larvae in currents. Vi; SYSTEMATIC DESCRIPTION Family PROETIDAE Salter 1864 Subfamily DECHENELLINAE Pribyl 1946 Genus DECHENELLA Kayser 1880 Subgenus DECHENELLA Kayser 1880 Dechenella (Dechenella) setosa Whidborne (Pl. 1, Text-figs. 1-6) 1889 Dechenella setosa Whidborne: 29. 1889 Dechenella setosa Whidborne: 27, pl. 2, figs. 15-17. 1912 Dechenella (Eudechenella) setosa Whidborne; Richter : 310, pl. 20, figs. 8, 9. 1950 Dechenella (Dechenella) setosa Whidborne; Richter, R. & E.: 178. Dracnosis. A bimorphic species of Dechenella with weakly impressed glabellar furrows. The frontal area is large and more than one quarter of the total length of the glabella at the sagittal line. The anterior border is broad. Specimens of Group C are distinguished from those of Group D by their smooth, more pointed glabellas. The axis of the pygidium is narrow and constricted between rings 7 and 8 in Group A, but straight in Group B. The length of the pygidium is more than three fifths of its maximum breadth. 18 rings and 12 ribs may be distinguished clearly. The pleural furrows are narrow and deep anteriorly but they shallow and widen posteriorly. Lectotype. Cephalon (BM., I. 5039). Pl. 1, fig. 14. Designated by Richter, R. & E. (1950). LocaLity AND Horizon. Chercombe Bridge Quarry (Nat. Grid Ref. SX832711), near Newton Abbot, Devon; Middle Devonian, probably late Couvinian. ADDITIONAL MATERIAL. The following specimens exemplify the characteristic features of each of the four groups of Dechenella setosa described in this paper: Group A, GSM.6987 (PI. 1, figs. 1-3); Group B, BM., I. 5056 (Pl. 1, figs. 5-7); Group C, BM., I. 5039 (PI. 1, fig. 14); Group D, BM., IT.1o1 (PI. 1, fig. ro). DESCRIPTION Cramdium. Side view. The glabella is broadly curved. From the crestal point, lying one third of the distance from the glabella posterior to the glabella anterior, the glabella descends quite steeply to the pre-glabellar field, but rather more steeply in Group D than in Group C. The preglabellar field declines gently forwards and BRITISH DEVONIAN DECHENELLID TRILOBITES 327 passes into a broadly concave anterior border furrow. The anterior border is gently inflated and rises at a low angle from the border furrow, but then falls sharply to the margin. The occipital furrow is deep and rounded and passes posteriorly into a symmetrical and flattened occipital ring. This ring continues the curve of the glabella in Group D but falls slightly below this in Group C. In sectioned material, the posterior border of the occipital ring is seen to be reflexed onto the ventral surface and is carried forward for a distance equivalent to half the length of the occipital ring. Frontal View. In profile, the glabella is weakly triangular and rises gently and regularly from poorly defined axial furrows to the sagittal line. The weak keel so developed is more evident in Group C than in Group D. The palpebral lobes are broad and flat. Plan View. The glabella, which is slightly longer than its maximum breadth, is weakly trefoiled, being constricted at 2p in Group C and at 3p in Group D. The anterior part of the glabella is broadly rounded and well defined in Group D but tapers rather more sharply in Group C and reaches slightly farther to the anterior margin. The maximum glabellar breadth is measured from 5-6, thereafter the glabella narrows slightly to the occipital furrow. Four lightly impressed glabellar furrows are recognisable on the dorsal exoskeleton. Ip is most strongly developed; it has a broadly arcuate course from a position somewhat anterior of the mid-point of the palpebral lobe, towards a point on the occipital furrow rather more than two thirds of the distance from the axial furrow to the sagittal line. Approximately half the distance along its course, the furrow curves more sharply towards the posterior; at this point a weak fork can be distinguished in some specimens, in others a faint pit can be seen on the line of, but separate from, the anterior section of 1p. The glabellar furrow 1p shallows noticeably towards, and fails to reach, both the occipital and axial furrows. 2p is less clearly impressed and runs parallel to the anterior section of 1p. 3p runs parallel to 2p; it is short and frequently just a faint mark on the glabella. 2p and 3p fail to reach the axial furrows. The distance between Ip and 2p is one and a half times greater than that between 2p and 3p. 4p appears as a shallow depression and can only be recognised on large specimens. A weak keel extends along the sagittal line from the posterior border of the glabella for a distance approaching one third of the total glabellar length. The occipital furrow is deep and narrow. It is arched forwards medianly and terminated laterally in deeply impressed pits, which are placed on the line of the axial furrow and orientated oblique to the occipital furrow. The occipital ring is a flat band marked by a small mesial tubercle. This ring averages one tenth of the total sagittal length of the cranidium. The frontal area is more than one quarter of the total sagittal length of the cranidium. The preglabellar area is smooth and slopes gently to a well defined anterior border furrow and is continued into the anterior area of the fixigena as a broad smooth surface sloping to the anterior border. The palpebral area is broad and flat and the posterior area of the fixigena is small. The anterior margin of the border is broadly curved; the anterior facing part of the border is marked by three to four discontinuous terrace lines, whilst the posterior part is 328 BRITISH DEVONIAN DECHENELLID TRILOBITES smooth and gently declined to border furrow. In Group D the border is larger than in Group C. The anterior branch of the facial suture diverges from the glabella at angles varying between 30°—-40°. y is rounded and placed opposite to glabellar furrow 3p, and is clearly separated from the axial furrows. 8 is evenly rounded in Group D, where it is positioned level with the anterior border of the glabella on the line of the anterior border furrow. The palpebral lobe is long and flattened. 3 is rounded and placed at approximately the same distance from the sagittal line as 8. ¢ lies farther from the axial furrow than y and is situated at a distance from the sagittal line approximating to half the maximum glabellar width. The posterior branch of the suture is short and turns sharply outwards at ¢ until a distance from the median line similar to that of 3 is reached; it then runs broadly parallel to the posterior border of the cranidium and eventually cuts the border at a distance from the axial furrow approximately equal to half the width of the occipital ring. The posterior border is broadly rounded to flattened, and its length (sag.) approximates to half that of the occipital ring. The internal mould is imperfectly known, but the glabellar furrows are broader and more clearly impressed than on the dorsal exoskeleton. Librigena. The cheek area is moderately inflated and slopes more steeply to the posterior border furrow than to the lateral border furrow. The eye platform widens laterally and to the posterior, but it is generally poorly defined. A broad but shallow lateral border furrow defines the lateral border. This border is triangular in cross section; the outward facing surface is steeper than that facing inwards and is orna- mented by four to five discontinuous terrace lines. The posterior border furrow, which has an open V-shaped cross section, unites with the lateral border furrow and continues for a short distance into the genal spine. The posterior border is only weakly inflated, and about two thirds of the width of the lateral border. The genal spine is short (about one half the maximum breadth of the librigena) and sturdy, being ornamented by two to three fine lines. The eye, which is large and crescentic, rises more steeply from the cheek area posteriorly than anteriorly. Its visual surface is smooth, and evenly convex and is separated from the cheek area by a weak groove. The doublure is flat and its breadth is comparable to that of the lateral border. Its surface is irregularly pitted and ornamented by six fine lines parallel to the margin of the cheek. The free border of the doublure is slightly recurved dorsally. Pygidium. Plan View. In outline, the pygidium is longitudinally elliptical but specimens of Group A are more rounded than those of Group B. The length of the pygidium is more than three fifths of the breadth, and the axis is narrow, ranging between one fifth and one third of the pygidium breadth. In Group B, the axis narrows evenly between straight axial furrows, but it constricts slightly between rings 7 and 8 in Group A. The posterior termination of the axis is rounded and reaches to the border furrow in small specimens of both groups, but it becomes pro- portionately shorter with the increase in size and then stands clear of the border. 18 (+2) rings may be recognised, of these 14-15 are clear for the posterior segmenta- tion is indistinct. Narrow ring furrows reach to the axial furrows in segments 1-8, but thereafter they weaken at the axial furrows and become less distinct. The dorso- BRITISH DEVONIAN DECHENELLID TRILOBITES 329 lateral parts of the rings are marked by weak longitudinal notches which define a weakly inflated area. These notches are deepest at the anterior border of the rings, and shallow rapidly to the posterior and do not affect the posterior border. Longi- tudinal grooves produced by the notches are developed with varying intensity; they are best shown in small specimens, particularly those of Group B. Rings of the larger specimens are more characteristically narrow flattened bands. I2 (+2) weakly S-shaped ribs may be recognised; of these 8-9 are clear. The pleural grooves of ribs 1-4 are well defined, narrow and deep, and reach almost from the axial furrow to the border. Thereafter the grooves become progressively less well defined, shallower and broader and fail to reach to the border furrow. All of the anterior ribs are notched at the border. Smaller specimens show ribs with flattened to broadly rounded cross section, but the ribs of larger specimens are triangular in cross section and show crestal lines either medianly or slightly posteriorly placed, and with a steeper fall to the posterior pleural furrow than to the anterior. This crestal line may be traced from the axial furrow across the flattened section of the pleural lobes. Interpleural furrows are faint, and not always recognised on all ribs; their presence has been recorded up to rib 8. The furrows which are medianly or slightly posteriorly placed on the rib are clear at the axial furrow but become less well defined towards the periphery. The border is flattened to weakly convex, and declines towards the margin. Initially narrow, it widens posteriorly and attains its maximum width at rib 5. In Group A, this width is maintained, but in Group B the maximum width of the border is attained at the posterior lateral part of the pygidium. The border furrow is only weakly developed. Posterior View. In Group B, the axis shows a semicircular cross section in specimens of small and intermediate size, but the section becomes more gently convex in large specimens. In Group A, the axis is more nearly semi-circular at all sizes. The flanks are strongly rounded in all small specimens but they become distinctly flattened with increased size in Group B. Group A is more strongly rounded at all sizes. The border slopes gently to the periphery and is differentiated from the pleural lobes only by a weak concavity. Side View. The axis curves gently down from the anterior to the posterior; occasionally the anterior part of the axis is rather flattened. All of the rings decrease in size posteriorly. The articulating half ring is distinct but narrow and the articulat- ing furrow is sharp. The first 5-6 rings are clear in this view and are separated by deeply impressed ring furrows; the individual rings are planar and slope anteriorly to the preceding ring furrow. Thereafter the rings are flat and the ring furrows scarcely impressed. The border is clearly marked off from the axis by a well defined re-entrant angle. In young specimens the border slopes to the margin at an angle of 45°. This slope decreases and flattens in adult specimens of both Groups A and B. The doublure is narrow and increases in breadth from the anterior margin of the pygidium towards the posterior, but it is weakly constricted postaxially. Its structure is continuous with the dorsal exoskeleton both at the periphery of the pygidium and at its anterior margin. The inner margin is free but closely applied to the ventral surface. Anteriorly the doublure is strongly inflated and evenly rounded but it 330 BRITISH DEVONIAN DECHENELLID TRILOBITES becomes somewhat flattened towards the posterior. The surface of the doublure is ornamented by 7-8 terrace lines but it is not pitted or granulated. Internal Mould. The rings are narrow and sharp and separated by wide deep furrows which decrease in intensity towards the posterior; all of the furrows are much clearer than the comparable furrows on the dorsal exoskeleton. The ribs are narrow and angular and appear as furrowed ridges between broad, deep pleural furrows. All of the ribs are much clearer and the posterior ribs extend further towards the border than on the dorsal exoskeleton. The border is clearly defined, smooth and flat. Ornamentation. With the exception of the furrows, a fine pitting has been recorded on all parts of the cranidium; this being most strongly expressed on the median parts of the glabella. The pitting of the fixed cheeks and the border is fine and irregular and not clearly defined below magnifications of «20. No granulations have been recorded in specimens referred to Group C, but a collection of pustules grouped at the posterior end of the glabella characterises Group D. At lower magnifications the free cheek is smooth, but at x 20 the surface is seen to be finely and irregularly pitted. With the exception of the furrows, all dorsal surfaces of the pygidium are pitted. The pits on the rings are considerably finer than those on the pleural lobes, where they are usually, but not invariably, arranged in two rows either side of the inter- pleural furrow. The pits of the border are evenly distributed. The rings and border are normally finely granulated. The intensity of granulation is variable; on the ribs it is normally confined to that part of the rib posterior to the interpleural furrow, and most strongly developed on the outer two thirds of the ribs. The granulation of the border is also of variable intensity ; normally the greatest concentration of granules is on the posterior-lateral part of the border. A few specimens in both Groups A and B appear to be completely smooth. Delicate, frequently sigmoidally shaped, raised lines ornament the steeper peripheral part of the border. These are oblique to the margin and usually more or less parallel to the sagittal line. Rarely the raised lines branch. Posteriorly the raised lines come to lie progressively more nearly parallel to the margin of the tail, and where preservation is complete run parallel to the margin of the tail at its extremity. The ventral surface of the dorsal exoskeleton is smooth at low magnifications but a fine granulation of the surface can be seen at 30. MEASUREMENTS (in mm.) GSM.6987.— BM. I. 5056 Length of pygidium 14:0 12:0 Breadth of pygidium 175 145 Length of axis Tees 10:25 Breadth of axis 5:0 4:0 BM.,1I.5039 BM.,IT.101 BM.,IT.102 Length of cranidium 14:0 (est.) 6:5 7°5 Breadth of cranidium 10-0 (est.) 5:0 6:0 Length of glabella 9°75 3°75 4°75 Breadth of glabella 8-0 3:0 4:0 Breadth of cephalon 22:0 BRITISH DEVONIAN DECHENELLID TRILOBITES 331 (a) (b) Lp (in mm) Bp (in mm) 0 5 10 15 La (in mm) Ba (in mm) 4 D. verneuili 6 D. rittbergensis ° D. gigouti e D. struvei Fic. 6. Scatter diagrams showing variation in some Continental species of Dechenella. The appropriate reduced major axis for Dechenella setosa has been added to each scatter. COMPARISONS WITH OTHER SPECIES. Arguments have been advanced in the fore- going pages in support of the contention that the specimens from Devonshire con- stitute a single bimorphic species with a wide range of variation. The variants of Dechenella setosa show many similarities to Continental species of Givetian age, and a close relationship is indicated. However, the fine and precise differences which separate these species make comparisons exceedingly difficult. Characters said to be diagnostic occur in varying combinations in the British material. This could have suggested, had less material been available, that several new species are represented. The pygidia of Dechenella setosa may be compared most closely to Dechenella verneuili (Barrande), Dechenella rittbergensis Zimmerman and Dechenella gigoutt R. & E. Richter, but no single pygidium can be found which agrees exactly with the diagnoses given for these species. Scatter diagrams (Text-fig. 6) prepared 332 BRITISH DEVONIAN DECHENELLID TRILOBITES from the published measurements, and measurements taken from the figures of the above species show a marked rectilinear distribution. This suggests that the species may be conspecific, and that future collecting may reveal a wider range of variation than has previously been suspected. A close relationship also exists between these scatters and those prepared for Dechenella setosa, for they all fall within the observed range of the British species. This is apparent on the scatter diagrams (Text-fig. 6) where the appropriate reduced major axis for Dechenella setosa has been added. Text-figure 6a is included as an example of a close correlation between the reduced major axis of Dechenella setosa and the scatter of the Continental species. Text-fig. 60 is included as an example with less perfect fit. Thus, it may prove difficult to separate these species of Dechenella in the future. Meanwhile, some of the more significant differences between Dechenella setosa and the published descriptions of other species are noted. The number of ribs and rings present in the pygidia of Dechenella setosa is com- parable to that in Dechenella rittbergensis, the number being distinctly higher than that of the other species. The shape of the pygidium, rather blunted and broadly- rounded posteriorly, is said to be closely comparable to Dechenella gigouti (Richter, R. & E. 1950). Although there is a similarity between Dechenella gigouti and the pygidium of Dechenella setosa figured by Richter (PI. 1, figs. 5-7), this particular shape is distinctly unusual in the sample; by far the more common shape is that ascribed to Dechenella vernewilt. Similarly, the very broad border of Dechenella gigoutt is not characteristic of Dechenella setosa, though variants with moderately broad border, as figured by Richter, do exist. The outline of the pygidium both in cross and long section, which is specifically important in other described species, is found to vary with size in Dechenella setosa. Unlike Dechenella rittbergensis the axis is normally separated from the post axial region by a clear re-entrant angle. The cross section of the ribs also varies with size; the rounded ribs (characteristic of Dechenella vernewilt) are most common in small specimens, whilst those with a more flattened section (characteristic of Dechenella viltbergensis) and triangular section (characteristic of Dechenella struver R. & E. Richter) typify the larger specimens. The granulated surface of the test of Dechenella setosa would appear to distinguish the species from Dechenella vernewili and Dechenella rittbergensis which are essentially smooth forms; however, occasional smooth forms occur in both Groups A and B. Those specimens of Group A can be said to show a “vernewili trend’’, for this group has the constricted axis characteristic of Dechenella verneuili: the smooth forms of Group B, which have straight axial furrows can likewise be said to show a “‘vzttber- gensis trend’. There is no regular variation of test pitting, such as has been used to distinguish between Dechenella rittbergensis and Dechenella vernemli. The cranidia may be compared most closely to Dechenella gigouti, Dechenella rittbergensis and Dechenella vernewili. The broad frontal area is characteristic and serves to distinguish the species from Dechenella nittbergensis and Dechenella verneutlt, but is less broad than that observed in Dechenella gigouti, where the length of the frontal area is equal to half the glabellar length. The presence of terrace lines on the BRITISH DEVONIAN DECHENELLID TRILOBITES 333 anterior border and the pitting of the surface serve to distinguish Dechenella setosa from Dechenella gigouti, which lacks both these features. Apart from the broad border, specimens referred to Group C appear superficially like Dechenella rittbergensis but the glabella is less pointed and the glabellar furrows are less clearly defined. Group D, on the other hand, is more comparable to Dechenella vernewilt but the glabella is less broad and the glabellar furrows run broadly parallel to one another. With the knowledge of the variation in Dechenella setosa, it is tempting to suggest that Dechenella rittbergensis and Dechenella vernewli represent bimorphic forms of the same species, but the geographical separation of localities yielding these species renders this improbable. VIL. REFERENCES Kayser, E. H. E. 1880. Dechenella, eine devonische Gruppe der Gattung Phillipsia. Z. disch. geol. Ges., Berlin, 32 : 703-707, pl. 27. Mrippieton, G. V. 1959. Devonian tetracorals from South Devonshire, England. J. Paleont. Tulsa, 33 : 138-160, pl. 27. Mirter, R. L. & Kaun, J.S. 1962. Statistical Analysis in the Geological Sciences. xiii + 483 pp. New York & London. Patmer, A. R. 1958. Morphology and Ontogeny of a Lower Cambrian ptychoparioid trilobite from Nevada. J. Paleont., Tulsa, 32 : 154-170, pls. 25, 26. PRIBYL, A. 1946. Notes on the recognition of the Bohemian Proetidae (Trilobitae). Bull. int. Acad. tchéque Sci., Prague, 46 : 91-131, pls. 1-4. RicHTER, R. 1912. Beitrage zur Kenntnis devonischer Trilobiten. 1. Die Gattung Dechenella und einige verwandte Formen. Abh. senckenberg. naturf. Ges., Frankfurt a.M., 31 : 239-340, pls. 18-21. RIcHTerR, R. & E. 1950. Arten der Dechenellinae (Tril.). Senchenbergiana, Frankfurt a.M., 31 : 151-184, pls. 1-4. SALTER, J. W. 1864. A Monograph of the British Trilobites from the Cambrian, Silurian, and Devonian Formations, 1 : 1-80, pls. 1—6. Palaeontogr. Soc., [(Monogy.], London. UssHer, W. A. E. 1913. The Geology of the Country around Newton Abbot. Mem. Geol, Surv. U.K., London. vi + 149 pp., 3 pls. WHIDBORNE, G. F. 1889. On some Devonian Crustacea. Geol. Mag., London (3) 6 : 28-29. 1889a. A Monograph of the Devonian Fauna of the South of England, Vol. 1. The fauna of the Limestones of Lummaton, Wolborough, Chircombe Bridge, and Chudleigh, Pt. 1 : 1-46, pls. 1-4. Palaeontogr. Soc., (Monogr.], London. PLATE 1 Fics. 1-4. Dechenella (Dechenella) setosa, Whidborne. Group A. Pygidium (GSM. 6987). 1, Plan view, axial furrows weakly constricted, x2.3 2, Side view, X2.3 3, Posterior view, xX2.3 Pygidium (TM., B.490). 4, Plan view, x1.8 Fics. 5-9. Dechenella (Dechenella) setosa, Whidborne. GroupB. Pygidium(BM., 1.5056). 5, Plan view, axial furrows straight, x2.3 6, Side view, x2.3 7, Posterior view, x2.3 Pygidium (BM., I. 5050). 8, Plan view, cicatrized specimen. Wound has caused axis to grow asymmetrically, x 2.3 Pygidium (BM., I. 1110a). 9, Plan view, large specimen showing prominent crestal lines on ribs, x1.8 Fics. 10,11. Dechenella (Dechenella) setosa, Whidborne. Group D. Cranidium (BM., IT. ror). tro, Plan view, 5.5 Cranidium (BM., IT. 104). 11, Plan view, x4.1 Fics. 12,13. Dechenella (Dechenella) setosa, Whidborne. Group C. Cranidium (BM., IT. t02). 12, Plan view, x2.3 Cranidium (BM., IT. 103). 13, Plan view, x4.1 Fic. 14. Dechenella (Dechenella) setosa, Whidborne. Group C. Cephalon (BM., I. 5039). Plan view, x1.4 Lectotype. Fic. 15. Dechenella (Dechenella) setosa, Whidborne. Free Cheek (BM., IT. 105), x5.5 All specimens whitened with ammonium chloride before photographing. Specimens with numbers prefixed BM., GSM. and TM. are housed respectively in the British Museum (Natural History), London, the Geological Survey & Museum, London, and the Torquay Natural History Museum. Bull. B. M. (N. H.) Geol. 10, 9 BERATED 1 DECHENELLA (DECHENELLA) SETOSA Ny cd ¥y: a ais niece igie bag 6) 4 we oe fi i 2) ‘ Mal “PRINTED IN BY THOMAS DE 1 : dias He: yi CRETACEOUS AMMONITES AN _ NAUTILOIDS FROM ANGOLA _ M. K. HOWARTH 4 y i? My ni i j : & Wa : tay re ye ba Oy : a : ae Cae 7a is Mok: to UNO. CE a ee Be onc. 1965 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA BY MICHAEL KINGSLEY HOWARTH, Ph.D. Ph. 335-412 ; 13 Plates ; 23 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 10 No. 10 LONDON: 1965 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 10, No. 10 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) 1965 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued December, 1965 Price £2 15s. CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA By MICHAEL KINGSLEY HOWARTH CONTENTS Page I INTRODUCTION ; p : : . é : < ‘ 339 II AMMONITE FAUNAS OF THE MARINE CRETACEOUS OF ANGOLA. 6 340 (a) Mogamedes basin . : c 0 : : : é 341 (b) mae and Cuanza basins . : : : : : 341 . Albian . : : 6 : : : : 341 De Canemanicin Seintionten : j : : ; ‘ 342 3. Campanian and Maastrichtian . : 3 é : 342 Ill canes DESCRIPTIONS : : : : 343 Family DOUVILLEICERATIDAE Dao & Wonencl F : : 343 Genus Douvilleiceras Grossouvre. 343 Douvilleiceras mamnullatum (Schlotheim) ? var. LUCE nodum (Quenstedt) . ; 6 5 : 343 Douvilleiceras orbignyt Hyatt : : : : 345 Family BRANCOCERATIDAE Spath . c 5 : : : 346 Subfamily Brancoceratinae Spath ; ; 3 a : 346 Genus Neokentrocevas Spath . . > : 2 : 346 Neokentroceras curvicornu Spath . : : : 348 Neokentroceras singulave Haas 5 : : : 350 Neokentroceras subtuberculatum Spath . 4 : 351 Neokentroceras trituberculatum sp.nov. . : ; 352 Neokentroceras pseudovaricosum Spath . 5 i 353 Neokentroceras cvassicostatum sp.nov. . j 3 355 Family PHYLLOCERATIDAE Zittel . ‘ : 0 : 0 356 Genus Neophylloceras Shimizu : 0 é : : 356 Neophylloceras ultimum Spath 5 : . : 356 Family TETRAGONITIDAE Hyatt . : 6 . c : 357 Subfamily Gaudryceratinae Spath : 2 : . c 357 Genus Anagaudryceras Shimizu : : ‘ 5 357 Anagaudryceras mikobokense Collignon : : : 358 Genus Gaudryceras Grossouvre : , : : : 360 Gaudryceras vavagurense Kossmat . ; , : 361 Family BACULITIDAE Meek . : 5 . 9 é : 362 Genus Baculites Lamarck , : 6 5 0 é 362 Baculites anceps Lamarck . 2 : : 0 363 Baculites subanceps Haughton 0 : : : 368 Family NOSTOCERATIDAE Hyatt . : ; : ; : 371 Genus Didymoceras Hyatt ; ; é j : 374 Didymoceras subtuberculatum sp.nov. . : 6 374 Didymoceras cf. californicum Anderson . c ; 376 Didymoceras cf. hoynbyense (Whiteaves) . : : 377 Didymoceras cf. angolaense (Haughton) . 6 : 378 338 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA Genus Nostoceras Hyatt. Nostocevas hyatti Siseasncon. Nostoceras cf. keynense (Anderson) . Nostocevas votundum sp. nov. Nostoceras helicinum (Shumard) Nostoceras (2?) obtusum sp. nov. Family DIpPLOMOCERATIDAE Spath. Genus Polyptychoceras Yabe Polyptychoceras pseudogaultianum (Woloyania) Family DESMOCERATIDAE Zittel Subfamily Puzosinae Spath Genus Kitchinites Spath. Kitchinites angolaensis sp. nov. Subfamily Desmoceratinae Zittel Genus Desmophyllites Spath . : : Desmophyllites diphylloides (Forbes) Subfamily Hauericeratinae Matsumoto Genus Oiophyllites Spath Oiophyllites angolaensis Spat Family PACHYDISCIDAE Spath 3 Genus Eupachydiscus Spath Eupachydiscus pseudogrossouvrer Collisnen Family PLACENTICERATIDAE Hyatt Genus Hoplitoplacenticeras Paulcke . Hoplitoplacenticeras cf. marrott (Coquand) Hoplitoplacenticeras cf. costulosum (Schliiter) Hoplitoplacenticeras spp. indet. Family SPHENODISCIDAE Hyatt Genus Manambolites Hourcq . Manambolites dandensis sp. nov. Genus Sphenodiscus Meek Sphenodiscus sp. indet. . Family NAuTILIDAE d’Orbigny Genus Eutrephoceras Hyatt Eutrephoceras simile Spath IV AGES OF THE FAUNAS DESCRIBED : (a) Douvilleiceras fauna of Dombe Grande (b) Neokentroceras fauna of Praia do Jombo (c) The Egito fauna (d) The Barra do Dande Payne (e) The Carimba fauna (f) The Benguela and San Nicolau Pareiies VY REFERENCES . CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 339 SYNOPSIS Five separate taunas are described from a collection of over 250 Cretaceous ammonites and nautiloids from Angola. They are as follows: (1) 9 Douvilleiceras from Dombe Grande, which fix the age as the Mammillatum Zone, Lower Albian. About 50 Neokentroceras from Praia do Jombo, north-east of Lobito, which are the best specimens yet found of the genus, and are of low Upper Albian age. (3) 85 ammonites of the genera Anagaudryceras, Gaudryceras, Didymoceras (including D. subtuberculatum sp. nov.), Polyptychoceras, Kitchinites (K. angolaensis sp. nov.), Desmophyllites, Oiophyllites, Eupachydiscus, Hoplitoplacenticeras and Tetragonites from Egito, which fix the age as the Vari Zone, Upper Campanian. 26 ammonites of the genera Neophyllocevas, Baculites, Nostoceras (including N. votundum and N. (?) obtusum spp. nov.), Polyptychoceras, Manambolites (M. dandensis sp. nov.) and Sphenodiscus from Barra do Dande, of which the Sphenodiscus is probably Upper Maastrichtian, while all the remainder are either Polyplocum Zone, Upper Campanian, or basal Maastrichtian. (5) © Baculites and Didymoceras from Carimba, of Upper Campanian age. The description of Baculites subanceps from Carimba has made necessary a full revision of the European species B. anceps Lamarck; the type population from Manche, France, is described, a selection of specimens are figured and a neotype is designated. (2 — = I INTRODUCTION TuE bulk of the ammonites described in this paper were collected between 1928 and 1931 by Henrique O’Donnell and Alexandre Borges, both of Servico da Carta Geologica of Angola (now superseded by the Servicos de Geologia e Minas at Luanda). O’Donnell sent 221 cephalopods (210 ammonites and 11 nautiloids) to Dr. L. F. Spath for determination and description in 1930. They are now in the collections of the British Museum (Natural History) and consist of the following: 81 ammonites from the Upper Albian of Praia do Jombo. 7 ammonites from the Cenomanian of Salinas. I ammonite and 2 nautiloids from the Senonian of San Nicolau. 7 ammonites (1 now lost) from the Campanian of Carimba. Io ammonites from the Campanian of Benguela. 85 ammonites and g nautiloids from the Campanian of Egito. 1g ammonites from the Campanian and Maastrichtian of Barra do Dande. Efforts were made by Spath from September 1930 to October 1935 to reach agreement with O’Donnell and later with Fernando Mouta on the cost and place of publication of a full description of this collection, but satisfactory terms could not be agreed upon. The negotiations with Mouta were reopened in 1950 with the result that Spath was then able to publish his “Preliminary notice” of the collection in 1951. Spath was never able to start on the full description he had wanted to publish for so long, and only now, 35 years after the collection first arrived here, is it possible to present the full description that the ammonites clearly deserve. Previous references to the collection as a whole were made by Mouta & O’ Donnell (1933: 64) and Mouta (1938: 33). The 7 Cenomanian ammonites from Salinas were described by Spath (1931: 316), and they all belong to species described by Douvillé (1931). The Egito 340 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA ammonites were also referred to by Spath (1940a@: 52). Finally both the Egito and Barra do Dande faunas were listed and discussed by Spath (1953: 49-50) in his paper on the Antarctic Campanian cephalopod fauna, and five of the specimens were figured: (Spath 1953; pl. 2, fig. 6; pli 3, fig: 6} pl. (6; fie. 6; pls 7, fe 75 plaagaties Ze Descriptions of the lamellibranchs, gastropods and echinoids collected at the same time by O’Donnell were also delayed; the lamellibranchs and gastropods were described by Rennie (1945) and the echinoids by Dartevelle (1952: 27; 1953). Alexandre Borges was less successful in sending his ammonites to Spath for description. He had concentrated on collecting examples of Dowvilleiceras from localities between Benguela and Dombe Grande in 1930 and 1931 and had finally obtained over 50 specimens. These he attempted to send to Spath in 1931 and again in 1932, but for some unknown reason the collection never left Angola. Through the kindness of the Director of the Servicos de Geologia e Minas at Luanda I have been able to see and describe the 9 specimens which are all that now remain of the original 50 Douvilleiceras. Other collections of Angolan ammonites in the British Museum (Natural History) that are described here are the 7 heteromorph ammonites collected by Mr. Beeby Thompson at Barra do Dande in about 1915 that were referred to by Spath (1921: 56), 6 examples of Neokentroceras from Catumbella purchased from Dr. W. J. Ansorge in 1905, and the Neokentroceras in Professor Gregory’s collection that were described by Spath (1922). All these collections, and two smaller ones, also contain many examples of the well-known Upper Albian ammonites of Angola described by Spath, Haas and others, but none of them belongs to undescribed species or warrants further description. Acknowledgements. Loans of type or figured specimens or of plaster casts of type specimens were kindly made available by Carlos A. Neves Ferrao, Director of the Servicos de Geologia e Minas, Luanda, Dr. N. P. Newell, of the American Museum of Natural History, New York, Dr. L. Cahen, Director of the Musée royal de l Afrique centrale, Tervuren, Dr. A. W. Crompton, Director of the South African Museum, Cape Town, and Dr. J. Sornay of the Muséum national d’histoire naturelle, Paris. Mr. C. W. Wright made some helpful suggestions on the species of Baculites discussed. The majority of the photographs were taken by the author, but a few of the larger specimens were taken by Mr. N. Tanti. Measurements. Whorl dimensions are quoted in millimetres in the following order: Diameter: whorl height, whorl breadth, width of umbilicus. Figures in brackets following each of the last three figures express that figure as a proportion of the diameter. II AMMONITE FAUNAS OF THE MARINE CRETACEOUS OF ANGOLA A complete bibliography of Angolan geology can be found in Andrade & Andrade (1958) and a bibliography of the Jurassic and Cretaceous rocks in Haughton (1959). General accounts of the Cretaceous of Angola are available in Mouta (1954: 53-58) and Haughton (1963: 277-283). Cretaceous sediments are confined to the coastal strip of Angola and stretch from Mogamedes in the south to Cabinda in the north. CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 341 Except in the Cuanza basin where the width reaches 140 km., the strip is nowhere more than a few kilometres wide. Along the whole coastal strip lagoonal and continental deposits of Aptian or pre-Aptian age rest directly on the basement complex. Marine conditions first appear towards the top of the Aptian and a thick series of sediments (attaining 2000 metres in places) was laid down, containing representatives of all the stages up to the top of the Cretaceous. The marine beds of this coastal strip can conveniently be divided into five basins (Neto 1961: 63). Ammonites have not been found in the two northerly basins of Cabinda and Congo, but occur in considerable numbers in the Cuanza, Benguela and Mocgamedes basins. (a) Mocdmedes basin. The general succession in this small southerly basin is as follows (Carvalho 1960: 37-48; 1961: 27-93, 210-212; Haughton 1963: 278-279): Maastrichtian. Fauna of fish teeth. Santonian or Campanian. Basalt. Cenomanian. Limestone with concretions. ?Albian. White Limestone. Unfossiliferous sandstone. Conglomeratic facies of torrential origin. 1. Lagoonal facies. Lamellibranchs and gastropods are common in divisions 4, 5 and 7 and have been described by Rennie (1929; 1945), but the ages of the lower two were wrongly stated to be Senonian. From a single bed only 0-4 metres thick in the upper part of division 5 at Salinas came the fine ammonite fauna described by Douvillé (1931). Spath (1931: 316; 1932a: 124) reviewed the determinations of those ammonites and established that, contrary to Douvillé’s assertions that Barremian to Turonian forms were represented, all were Cenomanian in age. A single specimen from Salinas was figured by Haas (1952: 2-4, figs. 3, 4) as Desmoceras latidorsatum (Michelin), var. inflata Breistroffer, and referred to the Albian. Its position in the succession at Salinas is not known. It agrees closely with species of known Albian age in the Benguela basin, but the species and variety occur in both the Upper Albian and Lower Cenomanian in other areas, so it may have been part of the Cenomanian fauna at Salinas described by Douvillé. The only other ammonite known from the succession of this basin is the specimen from division 7 at San Nicolau recorded as Baculites aff. asper (Morton) by Spath (1951: 9), which is not specifically determinable and may be Santonian or Campanian. (b) Benguela and Cuanza basins. The succession in the Benguela basin has been described by Neto (1960: 89-99; 1961: 63-93) and Haughton (1963: 279-281), and stratigraphical descriptions and maps for the Cuanza basin can be found in Brognon & Verrier (1958: 61-74), Hoppener (1958: 75-82) and Freneix (1959: 111-113). The succession and ammonite faunas of both basins are similar and may be considered together. 1. Albian. The celebrated Albian ammonite fauna of Angola is known mainly from localities close to Benguela. Important early works on the stratigraphy and ammonites of the Albian of that region by Choffat (1888; 1905) and Gregory (1910; Cee Ga SN eo 342 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 1922) were summarized and expanded by Spath (1922) when he described an extensive collection made by Professor Gregory. Further Albian collections were described by Haughton (1925), Airachi (1931), Thiele (1933) and Haas (1941). The whole of the previous work was again summarized by Haas (1942) when he described a large collection from the Albian of Hanha. Further Albian ammonites have been described by Haas (1945; 1952—but some are Campanian and are listed below), Sornay (1951; 1953) and Soares (1959). Almost the whole of this Albian fauna is of Upper Albian age. It is rich in specimens of Hysteroceras, Neokentroceras, Mortoniceras, Elobiceras, Puzosia and Hamitidae, of which the lowest in the succession are Hysteroceras and Neokentroceras (see p. 400 below) and date the base of the series as low in the Upper Albian. Specimens of Stoliczkaia figured by Choffat (1888: 69, pl. 2, figs. 5-9) and recorded by Haughton (1925: 270) and Mouta & O’Donnell (1933: 63) (none were seen by Spath or Haas) occur in beds above the main Mortoniceras bearing beds, and led to the proposal (Mouta & O’Donnell 1933: 58-63; Mouta 1954: 128) of a formation characterized by Stoliczkaia. This is still Upper Albian in age, probably the upper half. (A Stoliczkaza figured by Douvillé (1931: 29, pl. 2, fig. 2) from the Salinas fauna is almost certainly Cenomanian, like the remainder of Douvillé’s ammonites). Beds below the Upper Albian succession contain specimens of Douvilleiceras and Puzosia in some abundance (Neto 1960: 96; 1961: 69), indicating a Lower Albian age, but only one of the examples of Dowvilleiceras has ever been figured (Choffat 1888: 71, pl. 3, fig. 1). In the collections with which the present paper is concerned there are examples of most of the Upper Albian genera, but nothing new, except in the case of Neokentroceras, which is represented by a splendid series of specimens that are described in detail below and greatly extend our knowledge of this genus. A small collection of Dowvilleiceras is also described, which establishes the age of the beds from which they come as Lower Albian. 2. Cenomaman—Santoman. All the stages of the Upper Cretaceous are probably present in the Cuanza basin (Brognon & Verrier 1958) and ammonites from most of them have been mentioned by Hoppener (1958: 79-81). A Cenomanian Acanthoceras and a Turonian Mammutes were recorded by Thiele (1933), and Haas (1942a) described two poorly preserved ammonites, ?Mantelliceras and Sharpeiceras goliath, that are probably Cenomanian, a poor Turonian ?Romaniceras, and three well preserved Texanites of Santonian age. An Upper Santonian Placenticeras, P. reineckei, was figured by Haughton (1925: 271, pl. 13, figs. 4, 5). More recently Basse (1963: 871-875, pls. 22-24) has described a new collection of Upper Turonian—Lower Coniacian ammonites from Cape Ledo, Cuanza basin, which consists of examples of Prionocyclus and Subprionocyclus, and the new genus Ledoceras. Nothing comparable with the excellent Cenomanian ammonites of the Mocamedes basin has been found in the Benguela or Cuanza basins. 3. Campanian and Maastrichtian. The presence of ammonites of these two stages has often been mentioned, but only a few have been described hitherto. Their first mention was by Spath (1921: 56), referring to the Barra do Dande Nostoceras collected by Beeby Thompson that are described herein. The first to be described and figured were Haughton’s (1925) Campanian and Maastrichtian ammonites from CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 343 Carimba, consisting of species of Nostoceras, ?Didymoceras, ?Solenoceras, Baculites, Menuites and Libycoceras. Further ammonites from the Carimba district and from Capolo were described by Haas (1943), all consisting in this case of the heteromorphs Nostoceras, Axonoceras, Solenoceras, and Baculites. Preliminary identifications of the Campanian and Maastrichtian ammonites described herein were given by Spath (1951; 1953: 49, 50). Two fine specimens of Nostoceras from Barra do Dande were figured by Sornay (1951), and a fragment of a large Didymoceras from the same locality was figured by Silva (1961). All the above ammonites are from the Cuanza basin, but a Campanian succession also occurs in the Benguela basin, especially at Egito. Part of the Egito Campanian ammonite fauna was described unwittingly by Haas (1952) and wrongly referred to the Albian. Haas (1952: 16) said “‘the Albian age of the Ammonoidea here described is beyond any doubt’, but at Egito, Campanian beds he unconformably on Upper Albian, and ammonites from the Egito Campanian were mixed in his descriptions with Albian ammonites from other localities. “Puzosia lytoceroides” Haas (1952: 8-11, figs. 14-17) is the Upper Campanian form Gaudryceras varagurense (Kossmat), “Gaudryceras aenigma’ Haas (1952: II-12, figs. 18-20) is Anagaudryceras mikobokense Collignon, and the smaller of the two specimens figured as Tetragonites jurinianus angolana Haas (1952: 12-15, figs. 21, 23-25 only) is probably Campanian and is best identified as Tetragonites cf. epigonus (Kossmat). The other ammonites described by Haas are all from Albian localities in the Benguela region, the only doubtful ones being the four Egito specimens listed as Desmoceras latidorsatum (Michelin) var. inflata Breistroffer (Haas 1952: 3, 4), none of which was figured. Measurements of these four given by Haas agree with those of the Albian Catumbela specimens, and it is doubtful whether anything in the Campanian has such thick and depressed whorls, so their reference to the Albian species is probably correct. The rich Campanian and Maastrichtian collections from Egito and Barra do Dande that were summarized by Spath (1951) are described below, and there are smaller faunas from the Senonian of Carimba, Benguela and San Nicolau. The collection is richer than any Upper Cretaceous ammonites previously described from Angola. IT SYSLEMALLO DESCRIPTIONS Family DOUVILLEICERATIDAE Parona & Bonarelli 1879 Genus DOUVILLEICERAS Grossouvre 1894 Douvilleiceras mammillatum (Schlotheim) ? var. aequinodum (Quenstedt) Plate I, figs. 1-4 1846 Ammonites monile aequinodus Quenstedt: 138, pl. Io, fig. 2. 1888 Acanthoceras mamuillarve (Schlotheim); Choffat: 71, pl. 3, fig. 1. 1925 Douvilleiceras monile (J. Sowerby); Spath: 73, pl. 5, fig. 5. 1962 Douwvilleiceras mammillatum (Schlotheim) var. aequinodum (Quenstedt); Casey: 271, pl. 40, fig. 5, pl. 41, figs. 5-7, pl. 42, fig. ro. 344 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA MATERIAL. 8&8 specimens, D.G. 294, 300, 306-309, 312 and 322, from Dombe Grande, Benguela basin, in the collection of the Servicos de Geologia e Minas, Luanda. Description. The collection consists of four complete specimens, the largest being 63 mm. diameter, and four fragments of less than half a whorl. The whorl section is depressed in all specimens, with a height/breadth ratio of 0-75 to 0-80. There are 28 to 30 ribs per whorl at 60 mm. diameter. Up to about 40 mm. diameter all the ribs appear to commence at the umbilical edge, but at larger diameters about one-third are intercalated and commence only very weakly at the umbilical edge or in the middle of the side of the whorl. The ribs on all whorls are inclined slightly backwards. Each rib bears seven tubercles on each side of the whorl, which are equal in size and approximately evenly spaced after 45 mm. diameter. The largest specimen has an eighth tubercle on ribs near its aperture. At sizes smaller than 45 mm. diameter a much larger single lateral tubercle occurs on alternate ribs. These ribs have a small umbilical tubercle, the large lateral tubercle and three small ventro-lateral tubercles, and the ribs with which they alternate bear 5 or 6 small tubercles. At sizes smaller than 25 mm. diameter the large lateral tubercle appears to be developed on every rib. The mid-ventral sulcus is only a slight depression in the ribs which are continuous across the venter; it is roughly equal to the distance between the first and third ventro-lateral tubercles. Only poor traces of septa and suture-lines are to be seen in the specimens, but the three largest complete examples appear to have about two-thirds of a whorl of body chamber each, and are presumably immature. Measurements of the four best examples are as follows: D.G. 294. —————: 24:0, 29°9, ——. D.G. 306. At58mm.: 24:1 (0°42), 29:2 (0°50), 20-6 (0°36). D.G. 308. At 58-6 mm.: 24:0 (0:41), 29°7 (0:51), 19:0 (0°32). D.G. 309. At 55:2 mm.: 22-6 (0-41), 28-2 (0-51), 18-3 (0°33). Remarks. The full synonymy of Douvilleiceras mammuillatum (Schlotheim) and its varieties has been given by Casey (1962: 265-274). The present Angolan specimens are referred to var. aequinodum of that species rather than to the type variety because of the fairly high rib density of 28 to 30 ribs at 60 mm. diameter. The development of a large lateral tubercle up to 45 mm. diameter and the relatively wide ventral sulcus are more like the type variety of the species, while the umbilical width could be that of either variety when compared with Casey’s (1962: 267, 271) measurements. The strength of the ribs is variable in the Angolan specimens due to the varying preservation, but in some places the ribs are as strong and wide as in typical English and French examples of the species. The only previously figured Angolan Dowvilleiceras (Choffat 1888: 71, pl. 3, fig. 1) shows exactly the same characters as the present collection and is here referred to the same species and variety. It has the large lateral tubercles on the inner whorls and is drawn with eight tubercles on each rib near its aperture at about 70 mm. diameter. CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 345 In Britain and France this species is confined to and characterizes the Mammillatum Zone, the upper half of the Lower Albian. The species also occurs in Madagascar (e.g. Besairie 1936: 158, fig. 10d; Collignon 1949: 76; 1950: 46; 1963: figs. ?1238, 21239, 1241, ?1242, ?1248) where its stratigraphical position, characterizing a zone at the top of the Lower Albian, is again accurately known (Besairie & Collignon 1956: 32-36; 1960: 68-74). The only other record from Africa is of specimens from Somalia recorded and figured by Tavani (1942: 33, pl. 3, fig. 10; 1949: 37). Douvilleiceras orbignyi Hyatt Blatesmé fies .5 1841 Ammonites mammuillaris Schlotheim; d’Orbigny: 249, pl. 73, figs. 1-3. 1903 Douvilleiceras orbignyi Hyatt: 110. 1923 Douvilleicevas mammuillatum (Schlotheim) var. baylei Spath: 70, pl. 5, fig. 4. 1962 Douvilleiceras orbignyi Hyatt; Casey: 279, pl. 40, figs. 6-8, pl. 42, figs. 12, 13. MATERIAL. One specimen, D.G. 305, from Dombe Grande, Benguela basin, in the collection of the Servicos de Geologia e Minas, Luanda. DEscripTIon. The specimen consists of a half whorl fragment of about 50 mim. maximum diameter. The whorl section is greatly depressed, the whorl height and breadth being 17:5 and 25-6 mm. respectively near the aperture. Very large ventro- lateral tubercles, each divided by three spiral ridges, define a deep U-shaped mid-ventral sulcus. Large spinose mid-lateral tubercles occur on alternate ribs and below these there are tiny umbilical tubercles. Just before the broken aperture there is a single low rib with small tubercles; otherwise the ribbing is very weak throughout. REMARKS. This single specimen agrees with typical English and French examples of D. orbignyi at similar sizes. It is readily distinguished from any of the varieties of D. mammuillatum by the large ventro-lateral tubercles. D. orbignyi has these tubercles without, or with very few, intermediate small ribs. D. alternans Casey differs in having one intermediate rib at 50 to 100 mm. diameter, and D. magnodosum Casey and D. imaequinodum (Quenstedt) both have two intermediate ribs at similar sizes. These characters appear to be constant in these species, although there is considerable variation in other details of the ornament. In D. orbignyi the size of the lateral and ventro-lateral tubercles and the size of the ribs show much variation amongst the French neotype and topotypes and English specimens figured by Casey (1962: 279, figs. 99, 100, pl. 40, figs. 6-8, pl. 42, figs. 12, 13). The Angolan specimen has weak ribs more like the neotype than some of the other figured examples. The Madagascan species D. benonae Besairie (1936: 164, pl. 15, figs. 15, 16; Collignon 1963, fig. 1244) has even larger ventro-lateral tubercles and many more intermediate ribs. In Britain and France D. orbignyi is an associate of D. mammullatum in the Mammillatum Zone. Several specimens are known from Madagascar from the same zone at the top of the Lower Albian (Besairie 1936: 160; Collignon 1949: 76; 1950: 46; 1963: figs. ?1240, 1248). 346 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA Family BRANCOCERATIDAE Spath 1933 Subfamily BRANCOCERATINAE Spath 1933 Genus NEOKENTROCERAS Spath 1921 TYPE SPECIES. Neokentroceras curvicornu Spath 1921. EMENDED DIAGNOSIS. Small size; largest known adult is 40 mm. diameter when complete. Evolute, whorl section quadrilateral or rounded-quadrilateral. Ribs weak or absent in most species, but moderately strong in some. Umbilical tubercle strong; lateral tubercle weak or absent; ventro-lateral tubercle strong, clavate or spiny, but is sometimes absent when ribbing is strong. Keel present in all species. On final part of the adult body chamber the ventro-lateral tubercles become replaced by high ribs which curve strongly forwards to form continuous folds across the venter, and the keel is almost completely lost. AGE AND DISTRIBUTION. Lower half of the Upper Albian. Angola, ? Nigeria, ? Brazil. Remarks. The first proposal of Neokentroceras (Spath 1921a: 306) consisted merely of four words of description and the designation of a type species which was a nomen nudum. In a paper the following year Spath (1922: 105-107, 139-143, text-fig. D) gave full descriptions and discussion of the genus, its type species, and three further new species. Only 21 specimens were available to Spath; of these, 6 were not described (these are N. trituberculatum sp. nov. described below), and only 4 out of the remaining 15 were reasonably complete specimens. Spath’s poor illustrations made interpretation of the genus and its species very difficult for later workers except by reference to his original specimens. A second collection of Neokentroceras from Angola was described at length by Haas (1942: 46-66, pls. 7-10). It consisted of 117 specimens, of which all except about 9 were fragments of less than half a whorl. Haas adopted the 4 species and I variety of Spath and proposed a further 4 species and 4 varieties. The only other discussion of this genus is Reyment’s (1955: 39-41) description of three Nigerian specimens, which were referred to the type species of the genus, but two of them were made the type specimens of a newly created subspecies. The present collection yields a considerable amount of new information on Neokentroceras, for, although there are only 48 additional specimens, 33 of them are fairly complete and 20 have adult body chambers or adult suture-lines preserved. All are from a single locality at Praia do Jombo, Benguela basin, Angola. There are several specimens in each of the species described below which are complete up to the adult mouth border, and they all show a similar type of modification of the ornament on the final part of the body chamber. The ribs on the side of the whorl strengthen in all cases; where ventro-lateral tubercles are present they diminish in size and lose much of their tuberculate nature to become merely raised portions of the ribs; the ribs form chevrons or folds that are continuous across the venter, and the keel diminishes markedly, almost disappearing in many cases. In all instances CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 347 where the final suture-lines and the modified ornament on the body chamber are seen in the same specimen, the final suture-lines are approximated. The most highly developed species of Neokentroceras is the heavily tuberculate type species, NV. curvicornu, which has characters that are clearly different from any other genus of comparable age. The two ribbed species N. pseudovaricosum Spath and N. crassicostatum sp. nov., are, on the other hand, not far removed from some species of Hysteroceras. Amongst the Angolan forms Hysteroceras varicosum (J. Sowerby) var. angolana Haas (1942: 21-24, pl. I, fig. 21, pl. 2, figs. I-17) bears considerable resemblance to the ribbed species of Neokentroceras. Small inner whorls of N. pseudovaricosum and H. varicosum var. angolana are so alike as to be virtually indistinguishable, and it is probable that the two species are closely related. Hysteroceras orbignyi (Spath), a well-known species of the English Gault which is also found in Angola, is related to ribbed species of Neokentroceras. In all species of Hysteroceras the keel rapidly disappears and the ribs are continuous across the venter on most or all of the body chamber, whereas in Neokentroceras this stage is only reached on a short portion of the adult body chamber just before the mouth border. In view of the characters of the adult Neokentroceras that are now known, the two Nigerian specimens figured as the new subspecies N. curvicornu crassicornutum by Reyment (1955: 41, pl. 4, figs. 7, 8) cannot be placed in this genus. The smaller specimen (pl. 4, fig. 8) has large swollen ventro-lateral tubercles, small lateral tubercles and umbilical tubercles, while the larger specimen (pl. 4, fig. 7) shows a larger body chamber (? adult) with a broad flat venter and a strong keel, and quadrituberculate main ribs and some trituberculate intercalated ribs that do not reach the umbilical edge. These are clearly characters of the subgenus Mortoniceras (Durnovarites) to which both Reyment’s specimens should be referred, as MM. (D.) crassicornutum, a species with particularly large ventro-lateral tubercles on the inner whorls. Reyment’s third specimen (1955: 41, pl. 4, figs. 9, 9a), from a slightly lower horizon and preserved in a different matrix, is a fragment of a whorl which may well be a genuine N. curvicornu. From the body chamber characters of the largest specimen referred to above, Reyment (1955: 39) deduced that Neokentroceras was related to the quadrituberculate forms of Mortoniceras, a relationship first suggested by Spath (1922: 106). This is not correct for the proper adult characters now known for Neokentroceras point strongly to this genus being a late end-form development from Hysteroceras, a view adopted later by Spath (1934: 472-473) and also by Haas (1942: 47-48). Its age is therefore low in the Upper Albian. Neokentro- ceras should be referred to the subfamily Brancoceratinae rather than to the Mortoniceratinae, for it is unlikely to have had any direct connections with members of the latter family. Records of Neokentroceras indicating a world-wide distribution were listed by Spath (1922: 105-107) and by Haas (1942: 46-47). Spath’s (1934: 472) later view, that the genus was restricted to Angola, seems to be more correct, for records of the genus from Texas, Tunisia, India and Borneo are all based on figured specimens which are examples of Mortoniceras (s.1.), Spathiceras or Dipoloceras. The single 348 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA specimen from Nigeria figured by Reyment (1955: pl. 4, figs. 9, 9a) appears to be a genuine Neokentroceras, and the record of N. tectorius (White 1887: 225, pl. 20, figs. 6, 7) from Brazil, based on a diagrammatic drawing of a single specimen, is doubtful. Thirteen specimens from the Upper Albian of Madagascar figured by Besairie (1931: 633, pl. 65, figs. 4~7) and Collignon (1932: 16-17, pl. 3, figs. 1-9) are close to Neokentroceras. However they are all inner whorls of less than 15 mm. diameter, ribs are present in all cases together with small umbilical and ventro-lateral tubercles, and on the venters of even the most tuberculate specimens the keel is always dominant over the ventro-lateral tubercles. Their reference to Hysteroceras is more logical, for several of them are close to the more tuberculate examples of H. orbignyt Spath and H. binum (J. Sowerby) from the English Gault. Thus apart from possible occurrences in Nigeria and Brazil, there do not appear to be any records of genuine Neokentroceras from outside Angola. The present division of Neokentroceras into species is unsatisfactory, owing to the poor illustrations in Spath’s (1922) original paper, and to the fragmentary nature of nearly all the specimens described by Haas (1942). The collection described here contains the first complete specimens to be found. Haas’s division into 8 species and 5 varieties 1s excessive for many of his forms can be seen to represent individual variation within a species amongst the present collection. The amount of variation within one species is well illustrated in the 7 specimens of N. pseudovaricosum described and figured below (PI. 3, figs. 5-11). They show variation in adult size giving a factor of nearly 2:1 between the largest and smallest, and marked variation in rib density and whorl thickness. There are four or five different combinations of adult size, rib density and whorl thickness, yet all are united by the possession of highly characteristic ribs, which are broad and flattened at the ventro-lateral edge. It is clearly more correct to unite these under one specific name than to divide them into 4 or 5 species. Similar amounts of variation are found amongst the 29 specimens of N. curvicornu, and the smaller numbers of other species. In the collection as a whole divisions can be made at natural breaks in the variation, except in one case where two species seem to be very close, and the result is that only four of the previously described species are recognized and two new species are described. Neokentroceras curvicornu Spath Plate 2, figs. 1-9 1921a Neokentroceras curvicornu Spath: 306 (nomen nudum). 1922 Neokentroceras curvicornu Spath: 139-140, figs. D 1, Ia, 2. 1942 Neokentrocevas speciosum Haas: 61-63, figs. 6n, o, pl. 8, figs. 14-17, pl. 9, fig. ro. 1942 Neokentroceras speciosum var. vudis Haas: 63, fig. 6p, pl. 8, figs. 18, 19. ?1955 Neokentroceras curvicornu curvicornu Spath; Reyment: 41, pl. 4, figs. 9, 9a. Horotyes, “C) 20116 (Plpz, nee a): MATERIAL. In addition to the holotype, 28 specimens, including four paratypes (C. 20117-18, C. 20123, C. 20289) from the shore at landing place near Hanha, CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 349 and C. 52551-54, C. 52556-73, C. 52575 and C 52584 from Praia do Jombo. Dimensions of holotype: at 24 mm.: 7:7 (0:32), , 10:0 (0:41). Dimensions of C. 52552: at 26-5 mm.: 9-0 (0-34), 8-3 (0°31), 11-6 (0°44). Diacnosis. Evolute, whorl section quadrilateral, thickness and height of whorl approximately equal. Umbilical tubercle large and radially elongated. Ventro- lateral tubercle very large and forms outwardly pointing spine, sometimes curved slightly backwards. Keel well formed but lower than tubercles in all cases. Ribs weak or absent, except on final part of body chamber where tubercles diminish markedly in size, the keel disappears and the ribs are projected on the venter to form large folds. Length of adult body chamber about five-eighths of a whorl. REMARKS. The holotype was badly figured by Spath, but the specimen is poorly preserved and from the figure given here (PI. 2, fig. 1) it can be seen that the specific characters are barely discernible. However, the collection contains 24 specimens in addition to those seen by Spath, and as many of these are well preserved and show all stages of growth, the species can now be adequately described. Ten specimens have recognizable adult features, and these have mouth borders at sizes ranging from 21 to 35 mm. diameter, and approximated last suture-lines at diameters between 14°5 and 20-5 mm. Modification of the ornament on the adult body chamber consists of a tendency for the last three pairs of ventro-lateral tubercles to become elongated and to join across the venter as high forwardly curving ribs, and at this stage the keel almost disappears. On the middle one-third of the length of the adult body chamber the ventro-lateral tubercles become elongated into very large and widely spaced spines, but there is considerable variation in the degree of coarseness attained and in the direction of the spines, some of which are straight while others curve backwards even in the same specimen. The tubercles on the first one-third of the body chamber and the preceding septate whorls are fairly constant in size and density. At diameters between 20 and 28 mm. the numbers of ventro-lateral and umbilical tubercles average 15 and 10 per whorl respectively, and only on occasional specimens do these figures rise as high as 18 and 12. On the holotype there are 13 ventro-lateral and 10 umbilical tubercles per whorl at 27-3 mm. diameter, and the low figure of 13 reflects the wide spacing of the ventro-lateral tubercles on the last quarter whorl. It is probable that this part of the holotype is the middle one-third of the length of the body chamber, but this cannot be confirmed by the suture-lines which are not preserved. Lateral tubercles are not found at any stage of growth. Rib development is weak. The umbilical and ventro-lateral tubercles are some- times joined by a rib at the beginning and end of the body chamber, but these always show loss of relief when crossing the side of the whorl, and on other parts of the whorls ribs are absent or only very weak. The keel is small and is always lower than the ventro-lateral tubercles. This is the most evolute and most strongly tuberculate species of Neokentroceras, in which the development of keel and ribs is always very weak in comparison with the tubercles. N. singulare Haas is closely related and is discussed in detail below. 350 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA N. trituberculatum sp. nov. is also similar but develops a lateral tubercle from an early growth stage. N. speciosum and its variety rudis described and figured by Haas (1942) are synonyms of JN. curvicornu. His specimens all fall well within the range of variation of the latter species. The holotype of N. speciosum (Haas 1942, pl. 8, fig. 14) is close to the adult growth stage, and compares well with the specimen figured in Pl. 2, fig. 7. The type specimen of var. rudis (Haas 1942: pl. 8, fig. 18) can also be matched amongst the present collection (it is fairly close to Pl. 2, fig. 3), and the varietal name is not worth retaining. The three specimens figured by Haas (1942, pl. 8, figs 20-22) as N. curvicornu? are small fragments which are not really specifically determinable. Three Nigerian specimens were figured by Reyment under this specific name. One of them (Reyment, 1955: pl. 4, figs. 9, 9a) compares well with the specimen figured here in Pl. 2, fig. 9, and although fragmentary and poorly preserved it is probably a genuine N. curvicornu. The other two specimens (Reyment, 1955: pl. 4, fig. 7, 8) were made the types of the new subspecies N. curvicornu crassicornu- tum: this has been discussed above (p. 347) and shown to be probably a valid species of Mortoniceras (Durnovarites). Neokentroceras singulare Haas Plate 2, figs. 10-15 1942 Neokentroceras singulave Haas: 64-66, fig. 67, s, pl. 9, fig. 11, pl. Io, fig. I. MATERIAL. 14 specimens, C. 52555, C. 52574, C.52576-83, C.52585-87 and C. 52597 all from Praia do Jombo. DiaGnosis. Close to N. curvicornu, but distinguished by being slightly more involute, with flat and smooth whorl sides, and more compressed whorl shape. Umbilical tubercles of moderate size; ventro-lateral tubercles large and elongated into spines curving backwards. Ribs weak throughout. Adult body chamber similar to N. curvicornu. Remarks. If a much larger collection were available, a complete gradation might be found between this species and N. curvicornu, and singulare would then be considered a variety of the latter species. There are sufficient distinguishing features in the present collection of only 14 specimens, however, to justify their separation from N. curvicornu as a distinct, but very closely related species. At all stages the whorls are more compressed and a little more involute than in N. curvicornu. The sides of the whorl are nearly smooth in most individuals, and the tubercles are somewhat smaller than in the latter species. In the seven specimens that show adult characters, the diameter at the mouth border ranges from 23 to 35 mm., and the diameter at the final approximated suture-lines ranges from 15:5 to 22 mm. The body chamber occupies five-eighths of a whorl and has modifications of the ornament similar to those in NV. curvicornu, i.e. the ventro-lateral tubercles are large and widely spaced on the middle part of CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 351 the body chamber, and just before the mouth border the last 3 or 4 tubercles are modified to form raised ribs which curve forwards and cross the venter as folds (PI. 2, mess. IL; 15). On the septate whorls and the beginning of the body chamber the sides of the whorls are flat and almost smooth and neither the umbilical nor the ventro-lateral tubercles project markedly outwards from the side of the whorl as in N. curvicornu. Ribbing is very weak on these whorls and only occasionally do low ribs join the umbilical and ventro-lateral tubercles. On whorls between 20 and 25 mm. diameter there are 10-12 umbilical and 15-18 ventro-lateral tubercles per whorl, but the number of the latter may fall to 13 if the middle part of the adult body chamber is included. Lateral tubercles are never developed. The keel is about as high as the ventro-lateral tubercles on the septate whorls, but diminishes on the body chamber and disappears just before the mouth border. The single specimen and holotype of the species described by Haas (1942, pl. Io, fig. 1) appears to be an almost complete adult, for it has ribs just before the aperture which cross the venter as folds. The mouth border is just missing and the maximum size when complete would have been about 25 mm. diameter. It is very closely matched by the complete adult figured here in Pl. 2, fig. 15. Neokentroceras subtuberculatum Spath Plate 3, fig. 1 1888 Schloenbachia lenzi Szajnocha; Choffat: 64, pl. 1, fig. 3 (non figs. 4-6). 1922 Neokentroceras subtuberculatum Spath: 141-142, figs. D 8, 8a. 1922 Neokentroceras choffati Spath: 106. ?1942 Neokentroceras choffati Spath; Haas: 49-51, fig. 6a, pl. 7, figs. 15-18, pl. 9, fig. 4. 1942 Neokentroceras choffati Spath var. crassinoda Haas: 50, pl. 7, fig. 19. 1942 Neokentroceras costatum Haas: 52-53, figs. 6b-d, pl. 7, figs. 20-25, pl. 9, fig. 5. 1942 Neokentrocervas magnum Haas: 53-56, fig. 6g, pl. 8, figs. 2-6, pl. 9, fig. 6. 1942 Neokentroceras cf. subtuberculatum Spath; Haas: 56-58, figs. 6h, 2, pl. 8, fig. 7, pl. 9, fig. 7. HoLotyre. C. 20042 (Pl. 3, fig. 1), the only specimen, from near Benguela. REMARKS. The species is poorly known, but the two diagnostic features appear to be the development of ribs throughout growth and a small lateral tubercle on the adult body chamber, in addition to small umbilical tubercles and moderate sized ventro-lateral tubercles. The holotype was so badly figured by Spath that the species could not be properly interpreted from his descriptions. The specimen is, in any case, poorly preserved, and a full description of the species will only be possible when complete, well- preserved specimens are found. A second specimen, C. 20061, which was referred to this species by Spath is very badly preserved and is specifically indeterminate. The only part of the holotype that is at all well preserved is the final half whorl. This is probably part of the adult body chamber, though it is not possible to prove this as no suture-lines are preserved and the mouth border is missing. On this part of the specimen the whorl shape is quadrangular, slightly higher than broad, and has 352 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA parallel, almost flat sides. The umbilical tubercles are small and radially elongated, while the ventro-lateral tubercles are of medium size and point outwards, and have a tendency to be clavate. Small lateral tubercles are connected to both the inner and outer tubercles by low ribs. In some cases the radial elongation of the umbilical tubercle is angled slightly forwards on the ventral side of the tubercle, and the rib commences from the middle of the tubercle and runs behind this elongation. On other ribs the elongation is part of the rib itself. There are 18 ventro-lateral tubercles on the final whorl at about 35 mm. diameter; the number of umbilical tubercles is less than this, but an accurate count cannot be made owing to the poor preservation. The holotype of N. choffati Spath was figured by Choffat (1888: pl. 1, fig. 3). This shows all the characters typical of N. subtuberculatum and there can be little doubt that it isa synonym. The specimens figured as NV. choffati by Haas are all too small to be referred with certainty to the present species, though this will probably be possible when the septate whorls of N. subtwberculatum are properly known. The variety N. choffati var. crassinoda and the two species N. costatum and N. magnum of Haas are also included in the synonymy of N. subtuberculatum. A collection of more complete specimens will be necessary to confirm this synonymy, but the characters of the fragments figured and described under these names agree closely with those of N. subtuberculatum. Neokentroceras trituberculatum sp. nov. Plate 3, figs. 2-4 Hororyver. (©€Fy20285 (Ply 3) fig. 2). MATERIAL. In addition to the holotype, five paratypes, C. 14818-21, C. 20284 all from near Catumbella, Benguela, Angola. Diacnosis. Whorls robust with quadrangular section. A lateral tubercle of moderate size occurs in addition to a moderate-sized umbilical tubercle and a large clavate ventro-lateral tubercle. Ribs weak throughout and often absent on septate whorls. Keel well formed but lower than ventro-lateral tubercles. REMARKS. The six specimens referred to this species form part of a collection purchased from Dr. W. J. Ansorge, and they have been referred to by Spath (1922: 140) and Reyment (1955: 39). The specimen C. 36204 referred to by Reyment (1955: 39) as belonging to a related but different trituberculate species, is only 16 mm. diameter, and consists of small inner whorls of Mortoniceras (Durnovarites), as was recognized by Spath (1942: 713). This species is characterized by a well developed lateral tubercle which is placed slightly ventral of the middle of the side of the whorl and is developed from an early stage of growth. The umbilical tubercles are sharp and pointed on the inner whorls, becoming radially elongated on later whorls, while the ventro-lateral tubercles are large and tend to become clavate. Weak ribs are developed between the tubercles on the adult body chamber, but on the septate whorls they are still weaker or absent. CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 353 The holotype is septate up to 25 mm. diameter, but owing to the preservation it cannot be determined whether the final suture-lines are approximated; these are followed by five-eighths of a whorl of body chamber ending at a maximum diameter of 37:5 mm. This specimen is not quite complete, but has a small portion of the umbilical wall part of the mouth border preserved, indicating a diameter of 40 mm. when complete. The ornament on the side of the whorl shows no significant modifi- cation towards the end of the body chamber and the venter at this point is poorly preserved. There are 16 ventro-lateral and 15 umbilical tubercles on the last whorl at 38 mm. diameter. C. 20284 is a second large specimen having a maximum diameter of 38 mm., but the preservation is such that no suture-lines can be seen. It has 16 ventro-lateral and 16 umbilical tubercles on its final whorl. Specimens C. 14818-21 are parts of the inner whorls of four individuals; they all have maximum diameters between 15 and 19 mm. and suture-lines are only preserved in one specimen. The lateral tubercle first appears at about 14 mm. diameter in these specimens, at which size the umbilical and ventro-lateral tubercles are well developed. All other species of Neokentroceras, except N. subtuberculatum, differ in having no lateral tubercles. N. subtuberculatum differs in having smaller lateral tubercles that only appear at later growth stages and in having stronger ribs. Neokentroceras pseudovaricosum Spath Plate 3, figs. 5-11 1922 Neokentroceras pseudovaricosum Spath: 142, fig. D 4, 5, 5a. 1922 Neokentroceras pseudovaricosum var. compressa Spath: 142, fig. D 6. 1942 Neokentroceras costatum var. tenuis Haas: 53, figs. 6e, f, pl. 7, figs. 26, 27, pl. 8, fig. 1. HoLotyPe. C. 20125 (Pl. 3, fig. 5), from the shore landing place near Hanha. MATERIAL. In addition to the holotype, 6 specimens; C. 20120, C. 20122 (paratypes), C. 20124 from the shore at landing place near Hanha, and C. 52590-92 from Praia do Jombo. Dimensions are as follows: C2201255 At Z21.0;mmn: 7-0 (033)10:6" (0-31), — At 18-9 mm.: 6:5 (0°34), 6-0 (0:31), —. C. 20120. At 15-7 mm.: 6-2 (0°39), 5:6 (0°35), —. C. 20122. At 21-0 mm.: 7°8 (0°37), 7-0 (0-33), —. C. 20124. At 18-9 mm.: 6:5 (0:34), 5:8 (0°30), c.7-0 (0°37). Adult size c.21 mm., 10 ribs per half whorl at 19-7 mm. diameter. C. 52590. At 27-5 mm.: 9-0 (0°32), 8-0 (0-29), II-9 (0°43). At 21-0 mm.: 7:4 (0°35), 6°7 (0°32), 8-3 (0°39). Adult size c.37 mm., 22 ribs per whorl at 18-4 mm. diameter, 23 at 23°7 mm., 23 at 29 mm. 354 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA C. 52591. At 18-2 mm.: 6-2 (0°34), 6:0 (0°33), 7:6 (0°42). Adult size c.20 mm., 23 ribs per whorl at 18-8 mm. diameter. C. 52592. At 30:4 mm.: 10°7 (0°35), 9:3 (0°30), 12:7 (0:42). At 18-9 mm.: 7:3 (0°38), 6-6 (0°35), 7:0 (0:37). Adult size c.37 mm., 26 ribs per whorl at 21 mm. diameter. Diacnosis. Coiling less evolute than in other species of Neokentroceras. Whorls compressed with rounded quadrangular section. Ornament consists of small pointed umbilical tubercles and falcoid ribs which are wide and flattened at the ventro- lateral angle and swing forwards on the venter. The ribs are roughly associated in pairs with the umbilical tubercles, but in most cases connections between ribs and tubercles are very vague. There are no ventro-lateral tubercles. Keel of moderate size, fading on the last part of the adult body chamber, where the ribs are continuous across the venter as chevrons. REMARKS. Spath’s illustrations of the holotype and one of the paratypes of this species were so poor that Haas (1942: 58-61) was quite unable to interpret the species correctly. The specimens he figured as typical forms of the species (Haas 1942, figs. 67, k, pl. 8, figs. 8, 9, pl. 9, fig. 8) are small fragments which are difficult to place, but they have very large tubercles and are probably rather coarsely tuber- culate specimens of N. curvicornu, while the specimens he figured (Haas 1942, figs. 61, m, pl. 8, figs. Io-13, pl. 9, fig. 9) as N. pseudovaricosum var. gracilis are examples of either N. curvicornu or N. singulare. However, his figured specimens of N. costatum var. tenuis Haas (1942: 53, figs. 6¢, f, pl. 7, fig. 26, 27, pl. 8, fig. 1) appear to be genuine specimens of N. pseudovaricosum. The distinguishing character of N. pseudovaricosum is the type of ornament at the ventro-lateral edge. Commencing indistinctly at small sharp umbilical tubercles, the ribs rapidly strengthen and are falcoid on the side of the whorl, then they become broad and flattened at the ventro-lateral edge and curve strongly forwards on the venter. True tubercles are not formed at the ventro-lateral edge. The 7 specimens referred to this species are characterized by these distinctive ribs and are clearly marked off from all other species of Neokentroceras. In other characters, however, there is wide variation, as can be seen from the dimensions listed above. C. 20124 and C. 52591 (Pl. 3, figs. 8, 10) are nearly complete adult specimens with the mouth border only just missing in each case; the diameters when complete would have been 21 and 20 mm. respectively. C. 52592 (PI. 3, fig. 11) isa much larger adult with part of the mouth border preserved at 37 mm. diameter, while C. 52590 (PI. 3, fig. 9) has final approximated suture-lines at 25-5 mm. diameter followed by a quarter of a whorl of body chamber, indicating a size when complete of about 37 mm. diameter. Rib density varies between 20 and 26 ribs per whorl at 18-22 mm. diameter. C. 52590 has 11 umbilical tubercles and 23 ribs at 29 mm. diameter. Whorl breadth ranges from 30 to 35% of the diameter at 16-21 mm. diameter. The low whorl breadth of 30% in C. 20124 led Spath to the proposal of var. compressa for this specimen. It can be seen, however, that the difference in whorl breadth between this specimen and the holotype of the species at the same CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 355 diameter is only 0-2 mm. or 1%, which is a negligible difference and is much less than the range of variation in the species. Of the three whorl sections figured by Spath (1922: 141, figs. D4, 5a, 6), his fig. D4 is accurate, fig. D 5a is drawn much too wide, for in this specimen the whorl breadth is always less than the height, while fig. D6 is drawn much too compressed, for the whorl breadth should be 5:8 mm. The earlier part of the whorl in the latter specimen is slightly crushed by compression, and it was probably this that led Spath to the proposal of the name var. compressa. It is not advisable to use any varietal names for this species until the full variation is better known, for it is unlikely to be completely expressed in a collection of only 7 specimens. The degree of variation of adult size in this species is comparable with that found in much larger collections of other small species. The modifications of the ribs just before the mouth border in the smallest and largest adults (C. 52591 and 52592) are very similar. In both cases the last 3 or 4 ribs lose much of the broadening and flattening at the ventro-lateral edge and curve strongly forwards on the venter to join from opposite sides, while the keel almost disappears. Neokentroceras crassicostatum sp. nov. Plate 2, fig. 16, Pl. 3, figs. 12-15 1922 Neokentroceras sp., Spath: 143, figs. D 7, 7a. HoLotyre. C. 52593 (PI. 3, fig. 12) from Praia do Jombo. MATERIAL. In addition to the holotype, 7 specimens (paratypes); C. 20126 from the shore at landing place near Hanha, and C. 52594—96, C. 52598—600 from Praia do Jombo. Dimensions: C. 52593. At 32 mm.: 9°8 (0-30), 9:4 (0:29), 16-0 (0°50). 22 ribs and 12 umbilical tubercles at 34 mm. diameter. C. 52600. At 23-6 mm.: 8-3 (0°35), 7°7 (0°32), —. Diacnosis. Allied to N. pseudovaricosum, but more evolute, has larger and more widely spaced ribs clearly connected to umbilical tubercles, and small sharp ventro-lateral tubercles surmounting ribs on inner whorls. Intercalated ribs not connected with tubercles also occur. REMARKS. Of the eight specimens referred to this species, five have adult body chambers. The holotype (PI. 3, fig. 12) has final approximated suture-lines at 26 mm. diameter followed by three-eighths of a whorl of body chamber and would have been about 40 mm. diameter at the adult mouth border. C. 52596 (PI. 3, fig. 14) has nearly half a whorl of body chamber but the final septa are missing; the last 3 or 4 ribs before the aperture curve forwards on the venter and meet from opposite sides, indicating near proximity to the adult mouth border which would have occurred at 356 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 37 mm. diameter. C. 52599 (Pl. 3, fig. 13) is a much smaller adult, having approxi- mated suture-lines at 17:5 mm. diameter, half a whorl of body chamber, similar modification of the last 4 or 5 ribs, and the adult mouth border would have occurred at about 25 mm. diameter. C. 52598 (PI. 3, fig. 15) consists of one-third of a whorl, half septate, half body chamber; the final two suture-lines are only slightly approxi- mated, but the body chamber has the bold widely spaced ribs characteristic of this part of the adult. C. 52600 (PI. 2, fig. 16) has approximated suture-lines at Ig mm. diameter and nearly half a whorl of body chamber. The other three specimens are fragments, and one of them (C. 20126) was described briefly by Spath (1922: 143). The small ventro-lateral tubercles on the inner whorls are seen well on the holotype where they occur on the septate whorls up to about 24 mm. diameter. On other specimens they disappear at a smaller size, probably corresponding to about half a whorl before the beginning of the adult body chamber. On the holotype there are 22 ribs and 12 umbilical tubercles per whorl at 34 mm. diameter, and C. 52600 has 21 ribs per whorl at 25 mm. diameter. The ribs in this species are similar in form to those in NV. pseudovaricosum, but they differ in being stronger and more widely spaced and show distinct connections with the umbilical tubercles. Occasional ribs are intercalated and commence at about the middle of the side of the whorl and are not connected with the umbilical tubercles. Distinct ventro-lateral tubercles on the inner whorls also serve to distinguish this species from N. pseudovaricosum. The species shows considerable resemblance in side view to Hysteroceras varicosum var. angolana (Haas 1942, pl. I, fig. 21, pl. 2, fig. 1). In the latter variety, however, the ribs are bold and continuous across the venter of the whole of the body chamber, while the keel and small ventro-lateral tubercles of the septate whorls are lost at about the beginning of the body chamber. In N. crassicostatum the keel is present on the whole of the body chamber and only the last 3 or 4 ribs are continuous across the venter in the form of V-shaped chevrons. Family PHYLLOCERATIDAE Zittel 1884 Genus NEOPHYLLOCERAS Shimizu 1934 Neophylloceras ultimum Spath 1953 Neophylloceras ultimum Spath: 4, 49, pl. 7, fig. 7. HoLotyPe. C. 41477, the only specimen, from Barra do Dande. REMARKS. Several comparable species of Neophylloceras have been described since Spath (1953: 4, pl. 7, figs. 7a, b) named and figured this single Angolan specimen. Its greatly subdivided and complex suture-line shows through the very thin transparent shell, but the continuity is not sufficient to allow it to be figured. The extremely fine and closely spaced striae can be seen clearly on Spath’s figures and this character serves to distinguish N. wltimum from all other species. In fact 32 striae cross a length of 10 mm. of venter immediately preceding the aperture of this specimen, and this density is nearly twice that of the nearest comparable species at a CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 357 similar size. Such comparable forms are N. vamosum (Meek) which ranges from the Turonian to the Upper Campanian and probably the Lower Maastrichtian in western north America and Japan (Matsumoto 1959): 1-5, pl. 1, fig. 1), N. hetonazense (Matsumoto 19590: 5) in the Campanian and Maastrichtian of western North America, Japan and Graham Land, N. lambertense Usher (1952: 50, pl. 1, figs. I-3) in the Upper Campanian and Lower Maastrichtian of British Columbia, and N. neva (Forbes) from the Campanian or Maastrichtain in southern India (Kossmat 1895: 166, pl. 16, fig. 2). N. vamosum and N. hetonatense have recently been figured from Upper Campanian or Lower Maastrichtian beds in Alaska (Jones 1963: 22, pl. 6, pl. 7, figs. 1-5) and one of the specimens (pl. 6, figs. 2, 4-6) of the former species has extremely fine striae almost comparable with those of N. ultimum. The position of the genus Neophylloceras and its relationship with its Lower Cretaceous ancestor Hypophylloceras have been discussed by Matsumoto (1959a@: 55-58). Other phylloceratid species in the Campanian and Maastrichtain belong mainly to the genus Epiphylloceras Collignon 1956. The type species, E. surya (Forbes), occurs in Angola (Haughton 1925: 268, pl. 12, figs. 3-5) and southern India (Kossmat 1895: 158, pl. 16, fig. 1) and is characterized by bundled ribbing, one rib of each bundle being usually enlarged. Several other species occur in the Maastrichtian of Madagascar (Collignon 1956: 24-25). Family TETRAGONITIDAE Hyatt 1900 Subfamily GAUDRYCERATINAE Spath 1927 Genus ANAGAUDRYCERAS Shimizu 1934 TYPE SPECIES. Ammonites sacya Forbes 1846. The interpretation and characteristics of this genus have been discussed at length by Wright & Matsumoto (1954: I1I-113) and Matsumoto (1959: 73; 1959a: 138). Wiedmann’s (1962: 156-158) relegation of Anagaudryceras to the synonymy of Gaudryceras is not accepted. The ornament of all but the adult stage of Anagaudry- ceras is so fine that the shell appears to be smooth, while Gaudryceras is characteris- tically covered with fine ribs. This difference is sufficient for generic distinction in keeping with the scale of differences usually adopted for Cretaceous genera (e.g. by Wright 1957; Matsumoto 1959, 1959a, 19590). The only additional point concern- ing the morphology of Anagaudryceras which can usefully be made here concerns the suture-line. Spath determined the species described below as an evolute species of Pseudophyllites. When one of the suture-lines was etched out, however, it proved to have quadrifid lateral saddles (basically bifid, with each arm divided again) and a single large saddle in the internal suture-line. In Pseudop/yllites the lateral saddles are basically trifid and there are two or more saddles in the internal suture- line. These differences in the suture-lines enable involute species of Anagaudryceras to be distinguished from evolute species of Pseudophyllites where there are few or no other differences. 358 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA Species of Anagaudryceras range from the Upper Albian to the Maastrichtian and have a world-wide distribution: examples have been described from Europe, North Africa, Angola, Madagascar, India, Japan, Alaska, California, New Zealand and Antarctica. The wide distribution has led to a multiplicity of specific names, but none of the faunas contains more than a very few specimens and it is not yet possible to make an assessment of the variation within a species. The 17 available specific names are probably far too many (listed in Collignon 1956: 66, 68-70). One of the main species is A. sacya (Forbes; see Matsumoto 1959: 72) which has broad band- like ribs on the body chamber and well marked constrictions on earlier whorls. It ranges from the Upper Albian to the Turonian and possibly higher; A. buddha (Forbes), A. vevelatum (Stoliczka) and A. limatum (Yabe) are synonyms; Yabe’s variety obscura is probably a true variety, but larger and more complete specimens of the New Zealand Campanian species A. subsacya (Marshall) and A. crenatum (Marshall) are required before it can be determined whether they also are synonyms. The second main species has only weak constrictions and no band ribs on the body chamber. In the Maastrichtian this is A. mzkobokense Collignon and A. aureum (Anderson) is clearly a synonym; in India Cenomanian ammonites of very similar morphology have the three specific names A. involutum (Stoliczka), A. madraspatanum (Stoliczka) and A. wtaturense Shimizu, and in Japan A. yamashitar (Yabe) has only a marginally smaller umbilicus, but is Santonian. This leaves A. politissimum (Kossmat) from the Turonian to Santonian of India which has a smaller whorl height and whorl breadth at the same diameter as A. mikobokense and may represent a different species, and A. swubtililineatwm (Kossmat) from the Campanian or Maastrichtian of India which is too fragmentary for identification. Both A. multiplexus (Stoliczka) from the Cenomanian of India and A. coalingense (Anderson) from the Maastrichtian of California represent an evolute many-whorled species, with constrictions but no known ribs; much larger collections are needed to determine whether these are conspecific. Such a specific classification could be expected to emerge from a comparison of abundant material of Anagaudryceras if it were available. The collection described below consists, however, of only 13 specimens, yet this is one of the largest collections of a single species of the genus found so far. Anagaudryceras mikobokense Collignon Plate 4, figs. 1-3; Text-fig. 1 1938 Gaudryceras politissimum Kossmat; Collignon: 92, pl. 7, fig. 2. 1952 Gaudrycevas aenigma Haas: 11, figs. 18-20. 1956 Anagaudryceras mikobokense Collignon: 59, pl. 8, fig. I. 1958 Lytocevas (Gaudryceras) auveum Anderson: 184, pl. 71, fig. 1. 1959a Anagaudryceras mikobokense Collignon; Matsumoto: 139, pl. 38, fig. 1. MATERIAL. 13 specimens, C. 52636—48, from 1 km. north of Egito, Angola. Description. The innermost whorls are exposed and evolute, while the degree of involution increases slightly with increasing size. The whorl height and breadth CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 359 are equal at about 34 mm. diameter; at smaller sizes the breadth exceeds the height ; at larger sizes the height progressively exceeds the breadth, and at 80 mm. diameter the breadth/height ratio has a range of 0-80 to 0-90. The largest specimen is wholly septate at its maximum diameter of 85 mm., and none shows any adult characters. The shell is smooth and unornamented up to about 30 mm. diameter; at larger sizes there are straight radial growth striae on well preserved parts of the shell, and irregularly developed low, widely spaced radial ribs which tend to increase in strength with increase in size. These ribs are inclined forwards at the umbilical margin, curve slightly backwards on the side of the whorl and become radial in crossing the venter. There are no constrictions, although where the ribs cross the venter they often have the appearance of a low flare, of the sort that are sometimes associated with constrictions. In the suture-line there are four equal-sized folioles terminating each of the first and second lateral saddles and the first auxiliary saddle, then there are two smaller auxiliary saddles before the umbilical edge and a single large lateral saddle in the internal suture-line. There are small upright saddles in the middle of the first lateral and first auxiliary lobes. At large sizes the suture-lines become somewhat further subdivided. Remarks. These are the specimens originally identified as “Pseudophyllites sp. nov. (a more evolute form than P. indra Forbes sp.)”’ by Spath (1940a@: 52; 1951: 8; 1953: 49). They are now referred to Anagaudryceras rather than to Pseudophyllites because of their suture-line characters, which, as described above, enable the two genera to be separated. Two specimens of average characters are figured in Pl. 4, figs. I, 3, a slightly more involute specimen is figured in Pl. 4, fig. 2, and a complete suture-line is shown in Text-fig. I. Fic. 1. Complete suture-line of Anagaudrycevas mikobokense Collignon from venter to dorsum at whorl height of 28 mm. C. 52646, from Upper Campanian, 1 km. north of Egito, Angola. x2:6. Variation among the 13 specimens is not large and is mainly in the whorl propor- tions. Graphs of whorl height and umbilical width plotted against the diameter show a range in whorl height of 28-3 mm. to 32 mm. at 70 mm. diameter, and in umbilical width of 22 mm. to 24 mm. at the same size. Whorl! breadth is more constant, the greatest variation being less than 2 mm. at any diameter between 40 mm. (15:5 mm. whorl breadth) and 80 mm. (29-5 mm. whorl breadth). All the specimens have the 360 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA very reduced ornament on the outer surface of the shell, and this is even more reduced on the inner surface. An example of the present species from Egito was described and figured by Haas (1952: 11, figs. 18-20) as Gaudryceras aenigma Haas. This specimen is a perfect match for the one figured here in PI. 4, fig. 3. The true G. aenigma (Haas, 1942: 167, pl. 42, fig. 3, pl. 44, fig. 2) is an Albian species and has the fine sharp ribs characteristic of Gaudryceras. The Angolan specimens compare very closely with the holotype from the Lower Maastrichtian of Madagascar described by Collignon. That holotype differs only in the possession of very faint, rare constrictions. The Californian specimens described by Anderson and by Matsumoto are also very similar to the Angolan specimens and undoubtedly conspecific. Both Collignon (1956: 59) and Matsumoto (1959a@: 139) have included the two specimens from the Campanian of Antarctica figured by Kilian & Reboul (1909: 14, pl. 1, figs. 7, 8) in the synonymy of A. mikobokense. Although these Antarctic specimens appear to be smooth, one (fig. 7) shows what appears to be a strongly curved constriction, and both show traces of strongly prorsiradiate fine ribs. They are probably specimens of Gaudryceras with the ribs worn away, as is undoubtedly the case in the two further Antarctic specimens figured by Kilan & Reboul (1909, pl. 1, fig. 6) and Spath (1953: 12, pl. I, fig. 10). Other specimens of similar morphology to A. mikobokense but of different ages have already been briefly mentioned above, but in view of the small number of specimens involved and the difficulties of comparison, further discussion would not be of value. Genus GAUDRYCERAS Grossouvre 1894 TYPE SPECIES. Ammonites mitis Hauer 1866. The characteristics and synonymy of this genus have been discussed by Wright & Matsumoto (1954: 111-113) and Matsumoto (1959a: 141), who concluded that subdivision of the genus is not necessary. The specific nomenclature of Gaudryceras is in an even greater state of confusion than that of Anagaudryceras discussed above. Species of Gaudryceras have a world- wide distribution similar to that of Anagaudryceras, and also include examples known from South-East Africa, British Columbia and South America. About 27 specific names have been proposed (for lists see Collignon 1956: 67-60), plus G. alamedense (Smith 1898), G. devallense Anderson 1958, G. filicinctwum (Whiteaves 1876), G. navarrense Wiedmann 1962, G. sachalinense (Schmidt 1873) and G. vascogoticum (Wiedmann 1962). The number of known specimens of Gaudryceras is much greater than in the case of Anagaudryceras, and from the rich faunas in Madagascar and Japan it should be possible to work out the variation within each species and a good specific classification. It has been pointed out by Yabe (1903: 14) and Collignon (1956: 48-49) that at about 100 mm. diameter the whorls of many species become rapidly more massive, with whorl height and breadth increasing markedly and umbilical size decreasing markedly. Outer and inner whorls of the same species CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 301 often look very different, therefore, and many have been given different specific names. Interpretation of the type species, G. mute, is difficult, because the holotype figured by Hauer (1866: 305, pl. 2, figs. 3, 4) is distorted to an elliptical shape, and the low whorl breadth may also be due to crushing. Grossouvre (1893: 227, pl. 26, fig. 4, pl. 27) figured another small specimen and also a much larger specimen which is a good example of the much more massive appearance of the whorls at large sizes. This species retains fine ribs up to the largest known sizes, but further study of the holotype and of a topotype collection is necessary before it can be properly defined. Gaudryceras varagurense Kossmat Plate, fe: 5 Pls eties., 1, 2 1895 Lytoceras (Gaudryceras) vavagurense Kossmat: 122, pl. 17, fig. 9, pl. 18, figs. 2a—c. 21909 6Lytoceras (Gaudryceras) vavagurense Kossmat; Kilian & Reboul: 12, pl. 1, fig. 6. ?1929 ©6Lytoceras (Gaudryceras) vavagurense Kossmat; Barrabé: 180, pl. 9, fig. 16. ?1930 Gaudryceras vavagurense Kossmat; Besairie: 569, pl. 21, fig. 4. 1931 Lytoceras (Gaudryceras) vavagurense Kossmat; Basse: 14, pl. 1, figs. 25, 26. 1931 Lytoceras (Gaudryceras) vavagurense Kossmat; Collignon: 11, pl. 1, figs. 5, 6, pl. 8, fig. 2. 1952 Puzosia lytoceroides Haas: 8, figs. 14-17. 1953 Gaudryceras (Neogaudryceras) pictum (Yabe); Spath: 12, pl. r, fig. ro. 1956 Gaudryceras vavagurense Kossmat; Collignon: 56, pl. 5, fig. 6. 1962 Gaudrycevas navarvense Wiedmann: 158, pl. 9, fig. 3. MATERIAL. II specimens, C. 52649-59, from 1 km. north of Egito, Angola. DeEscripTIon. The whorls are evolute and the umbilicus shallow. The whorl section is rounded with greatest breadth at or near the umbilical edge. Whorl height and breadth are equal at about 40 mm. diameter; at smaller sizes breadth exceeds height, at larger sizes height exceeds breadth. The ornament consists of fine ribs, some of which bifurcate on the side of the whorl near the umbilical edge; there are also single ribs which do not divide, and a few intercalated ribs that do not reach the umbilical edge. The ribs are inclined strongly forwards on the umbilical walls, bend slightly backwards on the side of the whorl, then bend forwards again on the venter. There are 6 to 8 constrictions per whorl which follow the line of the ribs exactly, and appear as constrictions on the internal cast with a ridge or collar behind, whilst on the shell they appear as thickened ribs only. REMARKS. Spath (1951: 8) determined these specimens as “Gaudryceras sp. (vavagurense, auct, non Kossmat)”’. Haas (1952: 8, figs. 14-17) had two examples of this species from Egito amongst his collection. He figured one of them and made it the holotype of a new species, Puzosza lytoceroides, but it is clearly a fine example of Gaudryceras varvagurense and compares very closely with the specimen figured here mo Pl. 5, fig. 1. The largest specimen in the present collection is 85 mm. diameter, and does not show the massive whorls which the species develops at about 100 mm. diameter. Measurements of the whorl proportions obtained from 6 specimens were inadequate 362 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA for an assessment of the variation of the species, but they were plotted graphically and could be compared with the proportions of other specimens. The fine ribs con- tinue up to the aperture of the largest individual, and from a comparison with the type specimens of species which develop coarser ribbing, it is deduced that the present collection belongs to a species which retains fine ribbing throughout growth. Such species with coarse ribs are known especially from Japan (e.g. G. densiplicatum (Jimbo)) and all of them show the coarse ribbing well before the growth stage reached by the Angolan specimens. In another more closely related species, or group of species, the inner whorls are indistinguishable from those of the Angolan specimens, but at 50 to 70 mm. diameter the ribs, while remaining small, become more widely spaced. This is known in western north America as G. demanense (Whiteaves) (Usher 1952: 59, pl. 4, fig. 1), in Japan as G. tenwilivatum (Yabe 1903: 19) and in Madagascar as G. lauteli Collignon (1956: 57, pl. 7, fig. 1), all of which are either conspecific or closely related. For the fine ribbed species to which the Angolan specimens belong the oldest name is G. varagurense (Kossmat 1895), for the difficulties of interpretation outlined above make it inadvisable to use G. mute (Hauer 1866) until it can be properly defined. The holotype from the Santonian of southern India is a broken and incom- plete specimen, but its ornament seen on several whorls up to 100 mm. diameter, and its dimensions obtained from a reconstruction of the spiral, compare closely with those of the Angolan specimens. Other species which are very close to or conspecific with G. varagurense are: G. analabense Collignon (1956: 54, pl. 6, fig. 3) and G. beantalyense Collignon (1956: 53, pl. 5, fig. 3) both from the Coniacian of Madagascar, and G. variocostatum van Hoepen (1921: 7, pl. 2, figs. 10-12) from the Santonian of Pondoland, which is based on a specimen of only 40 mm. diameter. G. cinctum Spath (1921: 41; 1922a: 118, pl. 9, fig. 3) from the Santonian of Natal appears to be more involute, but it is too small for proper comparisons. The Japanese Santonian to Maastrichtian species G. stviatum (Jimbo 1894: 35, pl. 6, fig. 6) and its variety G. striatum var. pictum Yabe (1903: 33, pl. 4, fig. 6) are also fine ribbed, but again the material is too small and poorly known. G. navarrense Wiedmann (1962: 158, pl. 9, fig. 3) from the Campanian of northern Spain appears to be a typical G. varagurense showing all the normal characters. G. propemite Marshall (1926: 142, pl. 28, figs. 3, 4) from New Zealand and G. delvallense (Anderson 1958: 183, pl. 41, fig. 4) from California have strongly curved and wiry ribs and probably represent a different species. Any attempt to sort out the synonymies outlined here must await the description of the rich Japanese faunas and a re-assessment of the Madagascan specimens in the light of the results. Family BACULITIDAE Meek 1876 Genus BACULITES Lamarck 1799 TYPE SPECIES: B. vertebralis Lamarck (1801: 103) by subsequent designation by Meek (1876: 391). CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 363 Baculites vertebralis was introduced by Lamarck without any description, but with references to figures of Faujas (1799: 141, pl. 21, figs. 2, 3) and Bourguet (1742, pl. 49, figs. 313-316). Although the figures of these two authors showing short smooth septate fragments are very poor, it can at least be seen that the whorl sections of all of them are circular or elliptical. In a later work Lamarck (1822: 647) discarded the specific name vertebralis in favour of B. faujasii which he proposed for the same species, with a short description now added, a reference to the figure of Faujas only, and a statement that the type specimen (in his collection) came from the mountain of Saint-Pierre, near Maastricht (in south Limbourg, Holland). This type specimen is lost. J. de C. Sowerby (1828; 186, pl. 592, fig. 1) was able to interpret B. “‘faujasi”’ correctly from this description, stating that the venter and dorsum were equally rounded, the whorl section elliptical and the shell smooth. Topotypes from St. Pierre, Limbourg, were well figured by Binckhorst (1861: 40, pl. 5d, figs. Ia-h). This interpretation of B. vertebralis as a smooth species with an elliptical whorl section and of Maastrichtian age is now well established (e.g. Nowak 1908: 346, fig. 8a, pl. 14, fig. 8). The second important early species of Baculites is B. anceps Lamarck, 1822. Its interpretation will have to be discussed at length because the Angolan specimens are very close to a form from the Pacific region which has been referred to a subspecies of B. anceps. The type area for B. anceps is the outcrops of the Calcaire a Baculites, in Manche, France. B. anceps shows considerable variation, and as B. vertebralis occurs in the same beds, it is important to establish the identification of the latter species, so that its clear separation from B. anceps can be demonstrated. The necessity for designating a type specimen for B. anceps and describing the characters of the type population has been stressed by Matsumoto (1959a@: 130-136) and Matsumoto & Obata (1963: 59-63), for until this is done no further progress can be made in describing similar species from other parts of the world. Application will be made to the ICZN to have the specimen designated below as neotype officially recognized. Baculites anceps Lamarck Plate 4, fig. 4; Pl. 5, figs. 4, 5; Pl. 6, figs. 1-5; Text-figs. 2, 3, 5-12 1816 Baculites vertebvalis Lamarck; Defrance: supplement p. 60, pl. 22, figs. 1-3 (date of plate uncertain). 1822 Baculites anceps Lamarck: 648. 1825 Baculites vertebvalis Lamarck; Blainville: 380, pl. 12, figs. 1-3. 1831 Baculites anceps Lamarck; Deshayes: 224, pl. 6, fig. 2. 1837 Baculites anceps Lamarck; Bronn: 732, pl. 33, fig. 6. 1842 Baculites anceps Lamarck; d’Orbigny: 565, pl. 139, figs. 1-7. 1876 Baculites anceps Lamarck; Schliiter: 145, pl. 40, figs. 2, 6. 1888 Baculites anceps Lamarck; Prestwich: pl. 12, fig. 16. 1889 Baculites anceps Lamarck; Griepenkerl: 106, pl. 11, fig. 2. ?1891 Baculites valognensis Bohm: 50, pl. 1, fig. 13. 1908 Baculites anceps Lamarck var. valognensis Bohm; Nowak: 335, figs. 1-4 (p.331), figs. 6, 7, 9, 12 (p.337); pl. 14, figs. 6, 7. 364 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA NreotyrPeE. BM(NH) 32573, from the “Calcaire a Baculites’’ of Manche, France, is here designated as neotype. It was originally part of Mantell’s collection. Diacnosis. A species of Baculites in which the venter is always sharpened, and sometimes a keel is differentiated by slight grooves on either side. The dominant form is smooth at all growth stages and unconstricted, but others occur in which ribs varying between weak and strong are formed on either body chamber or septate portion. These ribs are large and arcuate on the dorsal half of the shell, then they swing well forwards and are reduced almost to striae that are straight up to the keel where they form slight crenulations in some cases. Fine striae occur between the main ribs on the external surface of the shell. A minority of specimens have constrictions, which occur indiscriminantly on smooth or ribbed forms and vary in strength between weak and well marked. DESCRIPTION. The type population occurs in the Calcaire a Baculites in Manche, Normandy. The locality from which most specimens have been obtained is Valognes. The largest collection is that in the British Museum (Natural History) and consists of 84 specimens, 47 of them obtained by Sowerby from de Gerville. The following description is based on this collection, plus two of d’Orbigny’s originals and four specimens from de Vibrayes’ collection in the Muséum national d’histoire naturelle, Paris, sent by Dr. J. Sornay, a total of 90 specimens. The largest specimen is a body chamber fragment with the mouth border missing, and has a cross section height of 32 mm. and a width of 22 mm. at the broken aperture. The height of the shell at the final suture-line before the body chamber varies between 14 and 26 mm., but some of the smaller specimens are probably immature. Only one specimen (PI. 5, fig. 4) has characters which in a spirally coiled ammonite would be taken as indicative of an adult: its mouth border is flared and the final two suture-lines are much closer together than any of the preceding ones; it is a small specimen compared with many of the others, the shell height at the mouth border and final suture-line being 16 and 14:5 mm. respectively. Such flared mouth borders are seen in a number of specimens, and they all have a long rostrum on the venter and a smaller one on the dorsum, as shown in one of the specimens figured by d’Orbigny (1842, pl. 139, figs. 3-5). In all specimens the cross section is sharpened or keeled on the venter, broad and slightly flattened on the dorsum, and has well rounded sides, so that even though the venter and dorsum are markedly different the thickest part of the shell is close to the mid-point of the side. Shallow grooves defining a distinct keel are present in a few specimens. The ornament shows considerable variation. The two variables are the presence or absence of ribs and constrictions, and the following table shows the number of specimens belonging to each of the nine possible combinations in the collection of gO specimens. Although there are no clear divisions between the groups, such a grouping expresses the variation, and shows that 59 of the 90 specimens are smooth, 16 have weak ribs and 15 strong ribs; in each of these three groups between one-third and one-quarter of the specimens have constrictions. The largest group, 42 smooth and CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 365 unconstricted specimens, accounts for nearly half the collection. The ribs are highly arcuate and strongly developed just dorsal of the middle of the side; they are inclined strongly forwards, straight and reduced to striae on the ventral half of the side, and reach the venter to form slight crenulations on the keel in some cases; they are also straight on the dorsum over which they pass without interruption, but are inclined less strongly forwards. The constrictions are similar to the ribs on the dorsal half of the side, but on the ventral half they at first follow the ribs, then bend slightly backwards before swinging well forwards again to reach the venter. CONSTRICTIONS Absent Weak Strong Absent 42 4 13 59 n is Weak iit 3 2 16 og | Strong II fo) 4 15 64 7 19 | 90 The specimens figured here to show the range of morphology are: the neotype (Pl. 5, fig. 5) which has a smooth body chamber, very weak ribs on the septate part and no constrictions; a smooth, unconstricted specimen, showing the final two suture-lines close together and part of the flared mouth border (PI. 5, fig. 4); a specimen with ribs of moderate strength (Pl. 6, fig. 1); a fragment with strong ribs (Pl. 4, fig. 4); two smooth specimens with constrictions (Pl. 6, figs. 2, 5); and two ribbed specimens with constrictions (Pl. 6, figs. 3, 4). REMARKS. Matsumoto (1959a: 130, Matsumoto & Obata 1963: 59) has already stated that the interpretation of Baculites anceps must be stabilized by the designation of a type specimen, and has suggested (quoting Wright 7m litt.) that such an interpretation should be based on d’Orbigny’s (1842) figures of the species. An examination of the original description and the type population of the species leads to somewhat different conclusions. Lamarck (1822: 648) described the species as follows: “Baculite gladiée. Baculites anceps. B. testa recta, compressiuscula, ancipiti, laevi; uno latere subacuto, altero crassiore, obtuso; siphone marginale ad latus acutum. Habite . . . Fossile d’Angleterre. Mon cabinet. Elle atteint jusqu’a 13 pouces de longueur.” Lamarck’s original specimens are lost, but from this description it is clear that this is a smooth species of Baculites with a subacute venter and a flat dorsum. It is most unlikely that Lamarck’s specimens came from England where the species is still not known to occur, and from interpretations of the species shortly after Lamarck it is much more likely that his originals came from the Calcaire a Baculites 366 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA in Manche, Normandy. Specimens probably from Manche were described by Defrance (1816: 160, pl. 22, figs. 1-3) and Blainville (1825: 380, pl. 12, figs. 1-3) under the name B. vertebralis; in both descriptions it is clear that the species referred to is B. anceps, and the figured specimen appears to be entirely smooth. Further specimens from Manche, collected by de Gerville at Valognes, were described by Deshayes (1831), this time under the correct name B. anceps; the figured specimen is a short smooth fragment, said to be keeled. The best interpretation of B. anceps prior to d’Orbigny is that of Bronn (1837), who figured a fine specimen from Manche, that is 265 mm. long, keeled, with a smooth septate portion and fine ribs on the body chamber; this was certainly sufficient to fix the identity of the species. The Swedish specimen figured as B. anceps by Hisinger (1837: 31, pl. 6, fig. 2) has a cross section that is close to a perfect ellipse and does not belong to this species. In view of the establishment of B. anceps as a mainly smooth species, it is surprising that d’Orbigny (1842) chose as his figured specimens two Manche examples that had large ribs on their body chambers. D’Orbigny was well aware of the variation of the species between such ribbed forms and entirely smooth forms, and his was the first good description of this variation. He also stated that the species was known only from Manche, even though he included (wrongly) Hisinger’s Swedish specimen in his synonymy. Authors following d’Orbigny added little to his interpretation of B. anceps. Binckhorst (1861: 42, pl. 5d, fig. 3) referred a specimen to this species which is, in fact, a Eubaculites with a tabulate venter from the Maastrichtian of Limbourg. Schliiter (1876) and Griepenkerl (1889) recorded the species from Germany. Griepenkerl followed d’Orbigny in considering the variety with large ribs as the normal form, and he proposed the name var. swblaevis for the smooth form, a name that is not necessary because the smooth form is the dominant form in the type population. Although the specimen figured by Prestwich (1888: 332, pl. 12, fig. 16) was labelled ‘‘Upper Chalk” and in the absence of a stated locality would be taken for an English specimen, it is from Manche, and the best one figured hitherto; it is 205 mm. long, one half septate with very fine striate ribs, and the other half smooth body chamber. The German specimen figured by Bohm (1891) as Baculites valognensis may be an example of B. anceps, but it is a small fragment and not really recognizable. Nowak (1908) interpreted Bohm’s species as merely a variety of B. anceps, and figured two Manche specimens, one with fine ribs on the septate part, the other with somewhat larger ribs on the body chamber. As these are also part of the normal range of variation in the type population (they are not like the really boldly ribbed forms) the varietal name is again unnecessary. Nowak (1908: 328, figs. 1-5 (p.329), pl. 14, figs. 1-5, 10) also proposed a new variety leopoliensis for a form from the Cretaceous of Poland that has bold ribs on the body chamber and fine secondaries retained on the venter to a large size. But the whorl section of this form does not show a consistent keel on the venter, which is the most characteristic feature of B. anceps, and the variety should be excluded from B. anceps. No further descriptions or figures of B. anceps have been given. From the discussion above it is clear that B. anceps was interpreted as a smooth or finely ribbed species prior to d’Orbigny, and it is now known that smooth forms CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 3607 are dominant in the type population. The specimen chosen as neotype is therefore an almost smooth example and is not like the coarsely ribbed examples figured by d’Orbigny ; it is the specimen, B.M. (N.H.) 32573, originally figured by Prestwich (1888, pl. 12, fig. 16), and was from Mantell’s collection, presented to the British Museum (Natural History) with the label ““Baculites anceps, Normandy’’. It is typical of the Fics. 2-4. Suture-lines of Baculites. Fig. 2. Baculites anceps Lamarck. Fourth suture-line from body chamber. Neotype, B.M.(N.H.) 32573, from Lower Maastrichtian, Calcaire a Baculites, Normandy. x4:°5. Fig. 3. B. anceps Lamarck. Suture-line at cross section height of 16 mm. B.M.(N.H.) 6408, same horizon and locality. 4:8. Fig. 4. B.subanceps Haughton. Last suture-line at cross section height of 11-5 mm. C. 52730, from Upper Campanian, Carimba, Angola. x5. dominant form of the species, having no constrictions, and only very fine ribs on the septate part. Plaster-casts of the two best specimens in d’Orbigny’s own collection (no. 7204) were kindly sent to me by Dr. J. Sornay, but neither is good enough to be made the type specimen, nor were they the originals of any of d’Orbigny’s figures. Four specimens from de Vibrayes’ collection in Paris, also sent on loan by Dr. Sornay, are smaller and less well preserved than the neotype. Amongst the material in the British Museum (Natural History) from which the above description of the type population was drawn, the one designated as neotype is the best specimen for which a definite locality is known (even though this is only ““Normandy’’). It is slightly 3608 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA better preserved and more complete than the best of the 47 Valognes specimens obtained by Sowerby from de Gerville and forming the main part of this collection. The characteristic feature of B. anceps is the keeled or sharpened venter. Variation in other characters is considerable, ranging from completely smooth to boldly ribbed types and including unconstricted and constricted specimens. There is no reason to believe that any of these should be separated specifically, for all intermediates exist, even specimens with very weak constrictions, and all are united by the keeled venter. B. anceps is very common at only one horizon in the Calcaire a Baculites of Manche (Grossouvre, 1901: 286, the lowest bed), where it has every appearance of forming a normally variable single population. b. vertebralis Lamarck, which occurs less commonly in the same bed, has a completely different cross section. The age of this bed is Lower Maastrichtian. The specimen described and figured by Desmarest (1817: 49, pl. 2, figs. 4-6) as B. dissimilis has a whorl section that is close to elliptical with no marked difference between venter and dorsum. The specific name is not a senior synonym of B. anceps. The Californian and Japanese specimens described by Matsumoto (1959a: 130-136, pl. 34, fig. 3; pl. 35, fig. 1) and Matsumoto & Obata (1963: 59-63, pl. 20, fig. 3) as B. anceps pacificus also lack the keel of B. anceps, and should be excluded from that species. They belong to B. subanceps as described below. Baculites subanceps Haughton Plate 5, fig. 3; Pl. 6, figs. 6, 7; Pl. 7, fig. 1; Text-figs. 4, 13=15 1925 Baculites subanceps Haughton: 278, pl. 14, figs. 6-8. 1959@ Baculites aff. B. anceps Lamarck; Matsumoto: 130, pl. 34, fig. 3; pl. 35, fig. 1. 1963 Baculites anceps pacificus Matsumoto & Obata: 59, pl. 20, fig. 3. LecToTyPe. South African Museum No. 6829 (PI. 6, fig. 6) from Carimba. MATERIAL. In addition to the lectotype, ten paralectotypes all numbered 6829 in the collection of the South African Museum, Capetown, and four specimens in the British Museum (Natural History) (C. 52729-32). All from Carimba. DEscRIPTION. The largest specimen (PI. 7, fig. 1) is a fragment of a body chamber with a short part of the flared mouth border preserved. The height and width of the cross section close to the mouth border are 37 and 30 mm. respectively. This fragment is 120 mm. long and no part of the septate shell is preserved. Its cross section is close to elliptical, but has two wide and shallow grooves on either side of the evenly rounded venter (Pl. 7, fig. 1). The shell is nearly smooth and there is a long ventral rostrum and a short dorsal rostrum at the mouth border. All the other specimens are smaller; some (e.g. Pl. 5, fig. 3) have presumably immature, unflared mouth borders, and the largest final suture-line before a body chamber occurs at a cross section height of 23-5 mm.; the cross section (Text-figs. 13-15) in all of them shows a flattened dorsum, a narrower but well rounded venter, and rounded sides; none has any signs of grooves on either side of the venter. There are no final crowded suture-lines. The ornament is similar to that of the ribbed forms of B. anceps; prominent arcuate ribs on the dorsal half of the side are projected strongly forwards CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 369 and reduced to striae on the ventral half, then (unlike the ribs of B. anceps) they increase in strength again and bend slightly back to pass over the rounded venter as prominent crenulations; there are between 2 and 3 times as many such ribs crossing the venter as there are arcuate ribs on the side. The ribs are also reduced to striae across the dorsum and are projected only slightly forwards. All the specimens bear such ribs except the single largest one which is nearly smooth. There are no constrictions. Remarks. Haughton (1928: 278) had “‘numerous examples’’ of this species and figured three of them. A holotype was not designated, so the whole collection consists of syntypes and a lectotype designation can be made. The original of Haughton’s pl. 14, fig. 6 is lost or not available for study, and as this figure does not show any of the ornament it is not suitable to be a lectotype. The specimen from which Haughton obtained the cross section of his pl. 14, fig. 8 (refigured here Pl. 7, fig. 1) is a very large smooth body chamber fragment, with shallow ventral grooves peculiar to its large size, and is also unsuitable for a lectotype. The lectotype designated is therefore the medium-sized specimen figured in Pl. 6, fig. 6, which is half septate, half body chamber, and shows the ornament well. It is one of the best preserved syntypes. Two further specimens are figured (Pl. 5, fig. 3; Pl. 6, fig. 7; Text-figs. 13, 15) which show the ornament and cross section at different sizes. Little variation in the ornament can be detected in the 15 specimens, for all the medium-sized examples are ribbed and the two largest body chambers become nearly smooth. The venter is smoothly rounded in all cases with no trace of sharpening or of a keel. an Fics. 5-15. Cross sections of specimens of Baculites figured in the plates. For details of individual specimens see plate explanations indicated below. Figs. 5-12. Baculites anceps wamanrcka hice 5)— elvan ton 4 bigs 61— Pla tic: 4; bies7 — Pll 5) fies 5. Hig. 8 — PG, ini, ae Jeikee, @) == 1eal, (6), 1alee Be IOree, eo) —— PIL, (6), satay, Sip Vente ery = IPA, Gp sales, 719 Velikex, 3s —— IPI Gy, fig. 5. Figs. 13-15. Baculites subanceps Haughton. Fig. 13 = PI. 5, fig. 3; Fig.14 = Pl. 6, Ide (5 lekfeg, a0) —— eal Gy salen G7, All figures natural size. 370 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA The rounded shape of the venter in B. subanceps is sufficient to separate it specifically from B. anceps, for the keel or sharpened venter of the latter is the one constant character of an otherwise variable species. No intermediate examples occur in the type poluations of either species. The Angolan examples described belong to the subspecies B. subanceps subanceps. In the Pacific region a second subspecies occurs, B. subanceps pacificus Matsumoto & Obata, which was originally described as a subspecies of B. anceps. However it has the rounded venter of B. subanceps with no trace of sharpening as in b. anceps. The Pacific subspecies differs from the type subspecies in Angola in having many more closely spaced arcuate ribs on the side of the shell. Comparison of the holotype of pacificus (Matsumoto 1959a: pl. 34, fig. 3) with the lectotype of subanceps (PI. 6, fig. 6) shows that paczificus has between two and three times as many arcuate ribs as subanceps. The density of the ribs on the venter is approximately the same in the two subspecies, as is also the whorl shape, the shape of the mouth border and the curve of the ribbing. There are no other differences between the two forms, which appear to be genuine contemporaneous subspecies that are geographically separate. B. subanceps pacificus can be dated as Upper Campanian in both Japan and California, while B. subanceps subanceps is definitely of Upper Campanian age in Angola, as deduced from the associated Libycoceras and the many heteromorph ammonites. The only other Baculites known from Angola are the fragments with keeled or sharpened venters described by Haughton (1925: 279, pl. 14, fig. 9), and the apparently similar forms described by Haas (1943: 13-15, figs. 15-19) as B. anceps, all from localities near Capolo. They are poorly preserved and from the wide variety of forms of the venter, some distortion has probably occurred making specific deter- mination doubtful. Two Middle or Upper Campanian specimens from Madagascar figured by Collignon (1938: 88, pl. 6, figs. 4, 5) are also poorly preserved and of doubtful affinities. The Pondoland and Zululand Baculites described by Woods (1908) and Spath (1921a) and further Madagascan species described by Collignon (1931) all appear to be of Santonian age, and the only other high Cretaceous specimen known from south of Sahara is the example from the Maastrichtian of Nigeria figured by Reyment (1955: 15, pl. 1, fig. 5) as B. cf. asper Morton. The most closely related species to B. subanceps is the Polish Upper Campanian species B. leopoliensis Nowak (1908: 328, pl. 14, figs. I-5, 10), which differs in that its arcuate ribs are retained to a large size (at least 40 mm. cross section height), its ribs do not form marked chevrons on the venter, and the thickest part of the arcuate ribs are approximately at the centre of the side, not dorsal as in B. subanceps. B. palestinensis Picard (1929: 438, pl. 10, figs. I-7) is another closely related species from the Upper Campanian of Israel; its ribs are less strongly curved and less projected on both venter and dorsum than in B. subanceps, and the thickest part of the arcuate rib is in the middle of the side. The Alpine species B. fuchst Redtenbacher (1873: 134, pl. 30, fig. 15) and the Californian species B. fairbankst Anderson (1902: 92, pl. 7, figs. 152, 153; 1958: 190, pl. 49, fig. 4) are both based on holotypes that are too poorly preserved to be interpreted satisfactorily, as has been pointed out by Matsumoto (1959a@: 134). CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 371 Family NOSTOCERATIDAE Hyatt 1894 The difficulties in arriving at a satisfactory generic classification of this family have been pointed out by Anderson (1958: 195) and Matsumoto (1959@: 157-158). The character usually considered to be of greatest generic significance is the mode of coiling of all the whorls, and especially of the adult body chamber; the presence or absence of tubercles, constrictions and flared ribs are characters used to a lesser extent. The classification arrived at by Wright (1957: L 222-224) is based mainly on mode of coiling and is a sound division of the family. Although few species of Nostoceratidae are known from large collections, in those where ten or more specimens are known from a single locality (including the Angolan collections of Didymoceras subtuberculatum sp. nov. and Nostoceras hyatti Stephenson described below), the mode of coiling shows little variation. At species level it is no more variable than any other ammonite character, and groups of species with similar coiling make satisfactory generic divisions. But considerable difficulties arise with the Campanian—Maastrichtian genera Civroceras Conrad (1868), Didymoceras Hyatt (1894), Nostoceras Hyatt (1894) and Bostrychoceras Hyatt (1900), to which all the Angolan examples belong, and these will have to be discussed more fully here. Wiedmann’s (1962) solution was to refer all these forms to the oldest of them, Cirroceras (Jouaniceras Basse 1939, was also included as a synonym), but it is open to the objections that Civroceras is not generically identifiable, Nostoceras is a compact and useful group of species, and the type species of Bostrychoceras, B. polyplocum (Roemer), was somewhat mis-interpreted by Wiedmann. Cirroceras Conrad (1868). TyPE SPECIES: Ammonceratites convadi Morton (1841). The holotype is from the Upper Campanian or Lower Maastrichtian of New Jersey, and is a fragment of less than one whorl from the middle growth stage of the ammonite. It is poorly preserved, crushed and seems to have been non-septate. It was refigured by Whitfield (1892: 2609, pl. 45, figs. g-1I), and is now apparently lost, but figures of a plaster cast of it have recently been given by Reeside (1962: 120, pl. 70, figs. 1-3). Other specimens referred to Morton’s species by Whitfield (1892: pl. 45, figs. 12, 13) and Weller (1907: 833, pl. 108, figs. 5-8) belong to the well-known species Nostoceras hyatti Stephenson. Identification of Civroceras must rest on the holotype alone, but it is clearly too fragmentary to distinguish between Didymoceras and Emperoceras. It closely resembles one of the superb specimens of Emperoceras simplicicostatum figured by Whitfield (1902: 68, pl. 25, fig. 2), but it could equally well be one of several species of Didymoceras (e.g. D. (2) newtont Whitfield 1880: 449, pl. 15, figs. 1-4, or D. hornbyense (Whiteaves), Usher 1952: 103, pl. 27, figs. I, 2, pl. 28, fig. 2), or a Nostocevas such as N. dvaconis Stephenson (1941: 413, pl. 82, figs. 5-9). (Emperoceras is quite distinct from the other genera (Didymoceras, Bostrychoceras and Nostoceras) described here because of its two long parallel arms in early growth stages (Whitfield 1902)). Discovery of more complete specimens from the same horizon and locality would hardly settle its identity satisfactorily, because the doubt would always remain as to whether they really were the same as the holotype. The absence of early and of adult whorls precludes generic identifica- 372 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA tion, and Civvoceras must therefore be considered a nomen dubium, an unusable generic name. Apart from Nostoceras which forms a compact, closely defined group (discussed below), there remain a large number of species to which the names Didymoceras and Bostrychoceras have been applied with differing interpretations and limits. The question to be decided is whether species such as B. elongatum (Whiteaves) (Usher 1952: pl. 28, figs. 3, 4), usually considered typical of Bostrychoceras, are to be separated generically from those like D. hornbyense (Whiteaves) (Usher 1952: pl. 27; pl. 28, fig. 2), usually considered typical of Didymoceras. The complete range of variation is perhaps better illustrated by two specimens figured by Stephenson (1941: pl. 83, figs. 6, 7 and fig. 13) that are very different, and would be referred to different genera by most workers. There are many species representing different combinations of tight or loose coiling and presence or absence of tubercles or constrictions between these extremes, and when the type species of Didymoceras and Bostrychoceras are examined, both are found to be relatively close to the centre of the variation. Didymoceras Hyatt, 1894. TyprE SPECIES: Ancyloceras nebrascense Meek & Hayden 1856. The holotype (Meek 1876: 480, pl. 22, fig. 1) is a half whorl fragment in which the whorls were probably in contact. A better specimen was figured by Whitfield (1880: 451, pl. 14, fig. 9, pl. 15, fig. 6) and consisted of two helically coiled whorls in contact. In the specimen figured by Hyatt (1894: 574, pl. 14, figs. 13, 14) only the last helical whorl before the body chamber is preserved; this is not in contact and is followed by a rounded body chamber loop. Regular bituberculation is present on the body chamber and the last one or two whorls of the spire. Another typical species is D. hornbyense (Whiteaves 1895) (Usher 1952: 103, pl. 27, pl. 28, fig. 2), in which only the last whorls of the spire are loosely coiled. Bostrychoceras Hyatt, 1900. TypE SPECIES: Turrilites polyplocus Roemer 1841. Wiedmann’s (1962: 198-200) interpretation of this species is open to criticism. Roemer (1841: 92, pl. 14, figs. 1, 2) figured two syntypes of his species, and Schliiter (1872: 112) specifically excluded Roemer’s fig. 2 from his synonymy of this species. But this can hardly be considered as a selection of Roemer 1841, pl. 14, fig. I as lectotype of the species, because in the next part of his work Schliiter (1876: 135) included the same fig. 1 of Roemer (and again specifically excluded fig. 2) in a new species T. saxonicus. This is not corrected in the corrigenda to Schliiter’s work, and from his wide interpretation of T. polyplocus it is not clear which of Roemer’s two figures he wished to include in the species. Wiedmann (1962: 198) must be credited with having selected Roemer’s fig. 1 as lectotype. This lectotype has very irregular tubercles only on its last whorl just before it becomes uncoiled, and from the way they are drawn it could even be doubted whether they are tubercles at all. At least the whole of the closely coiled spire is non-tuberculate in a typical B. polyplocum. Of the specimens figured by Schliiter (1872) the two non-tuberculate specimens of pl. 33, figs. 3 and 4 are the most typical of B. polyplocum, while pl. 33, fig. 5 has the last three whorls loosely coiled and tubercles on the body chamber. The remainder might all be different species—pl. 33, fig. 6 and pl. 35, fig. 8 are bituberculate and CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 373 loosely coiled, pl. 34, fig. 1 is similar but tightly coiled at large sizes, pl. 33, fig. 8 is constricted and probably tightly coiled, pl. 34, figs. 2-5 has depressed whorls and has been re-named Cirroceras depressum Wiedmann (1962: 199), and pl. 35, figs. 1-7 have been referred by Wiedmann (1962: 204) to Didymoceras schloenbachi (Favre). The species B. secoense Young (1963) from Texas differs from B. polyplocum only in being consistently bituberculate on all whorls. Further work on much larger collections from Germany is necessary to determine the limits of variation in B. polyplocum, but it seems most likely that the tightly coiled, non-tuberculate forms and those with the last one or two whorls loose and tubercles on the body chamber, are conspecific. A fine specimen of the former type from Madagascar was figured by Boule, Lemoine & Thevenin (1907: 61, pl. 14, fig. 1). B. andicwm (Stoliczka) (Kossmat 1895: 143), to which Wiedmann (1962: 200) referred these tightly coiled, non-tuberculate forms, differs in having constrictions on all whorls and is of Coniacian to Santonian age. Matsumoto’s (1959@: 159) suggestion that the constricted group being Santonian and older might be separated subgenerically from the non-constricted group which are Campanian and Maastrichtian, seems to be defeated by the presence in the Maastrichtian of such strongly constricted forms as B. sauwndersorum (Stephenson 1941: 416, pl. 83, figs. 6-8). Another typical species of Bostrychoceras is B. elongatum (Whiteaves) (Usher 1952: 105, pl. 28, figs. 3, 4). Thus the type species of Didymoceras and Bostrychoceras are very similar, differing only in the slightly looser coils and more persistent tubercles of the former. The difference does not warrant generic separation. A form which is perhaps midway in morphology between the two type species is D. californicum Anderson (1958: 197, pl. 72, fig. 6). A large number of other forms carry the range of variation far beyond the characters of the two type species. B. condamyi (Collignon 1932: 39, pl. 9, figs. 1, 2) and B. otsukai (Yabe 1904: 14, pl. 3, fig. 9, pl. 4, figs. 1-3) have ribbing like a typical Bostrychoceras but loosely coiled whorls. A specimen of the latter species figured by Matsumoto (1959a: 160, pl. 40, fig. 2) differs from a typical Didymoceras (e.g. Usher 1952: pl. 27) only by its finer ribs and lack of tubercles. B. boulei (Collignon, 1932: 40, pl. 9, fig. 4) has fine ribs and no tubercles, but very loosely coiled whorls. D. navarroense (Shumard) (Stevenson 1941: 417, pl. 83, figs. g-13) is equally loosely coiled but has heavy bituberculation; while D. subtuber- culatum sp. nov. described below has smaller tubercles and periodic flared ribs. Amongst the closely coiled species, D. stevensoni (Whitfield 1880) is typical of Didymoceras except that some examples (Whitfield rgo1: 219, pls. 29, 30) are closely coiled up to the end of the final whorl, B. colubriformis (Stevenson I941: 412, pl. 81, figs. 1-3) is dwarf and has constrictions and fine bituberculate ribs, and B. saundersorum (Stevenson 1941: 416, pl. 83, figs. 6-8) is particularly tightly coiled, non-tuberculate and has marked constrictions. D. schloenbachi (Favre 1869; see Basse 1931: 19, pl. 2, figs. 11-15; Wiedmann 1962: 204) is equally tightly coiled, but is bituberculate and constricted and as its body chamber is not known it might be a Nostoceras. In addition there are many fragmentary specimens described under different specific names by Gabb (1864), Meek (1876) and Anderson (1958) that are not even generically identifiable. 374 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA The three main variables—coiling (loose or tight), tubercles and constrictions (present or absent in each case)—give eight different combinations, to each of which one or more species could be referred. To use coiling alone as the basis for generic division would leave a large number of intermediate species that have partly loose whorls. As the choice is between one genus or about eight genera, all these forms are best referred to Didymoceras Hyatt, of which Bostrychoceras Hyatt is considered a subjective synonym. Nostoceras Hyatt 1894. TyPE species: N. stantont Hyatt. The two syntypes were figured by Stephenson (1941: 407, pl. 80, figs. 2-5) together with several other species of the genus. This is a relatively closely defined genus, characterized by a closely coiled spire followed by a U-shaped retroversal body chamber that breaks away suddenly from the spire. In Nostoceras s.s. the body chamber hangs vertically below the spire, in the subgenus Anaklinoceras Stephenson (1941: 414) it turns upwards and surrounds the spire. All have well developed bituberculate ribbing on all whorls. Most of the known species are the North American forms described by Anderson & Hanna (1935: 22), Stephenson (1941) and Anderson (1958). In addition there are the Angolan forms described below, some undescribed badly crushed examples from Syria and Iraq, and possibly N. schloenbaci (Favre 1869—see above), N. pauper (Whitfield) (Reeside 1962: 118, pl. 68, figs. 10-13), N. natalense (Spath Ig2ta: 248, pl. 22, fig. 2) and N. subangulatum (Spath 1921a: 250, pl. 22, fig. 3) of which the body chambers are not known. In the descriptions of the Angolan fauna the following terms are used for the helically coiled forms. With the spire in an upright position and the apex pointing upwards, the upper part of the whorl is the upper surface between the venter (the siphuncle is usually just above the outermost point of the whorl) and the dorsum, the lower part of the whorl is the lower surface between venter and dorsum. When the whorls become detached and form a U-shaped loop as in Nostoceras, the ornament of the spire becomes twisted so that what was the upper part of the whorl forms the back of the loop, while what was the lower part of the whorl forms the front of the loop. The venter runs around the periphery of the loop (or occasionally just to the back of the periphery). Genus DIDYMOCERAS Hyatt 1894 Didymoceras subtuberculatum sp. nov. Plate 7, figs. 2-6, Pl. 11, fig. 4 HOLoTyPE. C. 52701 (Pl. 7, fig. 2) from 1 km. north of Egito, Angola. MATERIAL. In addition to the holotype, 16 specimens, including 15 paratypes (C. 52693, C. 52695-709) and C. 52694 which shows some variation. Same locality. Diacnosis. The spire consists of a loose helicoid spiral, dextrally or sinistrally coiled, distance between adjacent whorls in uncrushed material roughly equal to cross section diameter of shell at that point. Earliest whorls and shape of adult body chamber not seen in material preserved. Whorl section approximately circular at CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 375 all growth stages seen. Ornament consists of ribs, tubercles, constrictions and flared ribs, all developed at all growth stages. Ribs fine and dense and not interrupted on either venter or dorsum; they cross the dorsum radially, are inclined backwards on both upper and lower sides in crossing from dorsum to venter, but the inclination is considerably more on the upper side, so that they are inclined forwards in crossing the venter from upper to the lower side. Occasional ribs bifurcate on upper and lower sides and a few intercalated ribs cross the venter. Three or four flared ribs per whorl, present on all the whorls preserved, flares often immediately preceded by a slight constriction. Two rows of tubercles occur on all whorls; upper row just below mid-ventral line (i.e. line of siphuncle) and lower row just ventral of middle of lower surface. These paired tubercles joined by two ribs with 2 to 6 non-tuberculate ribs between each pair; much more widely spaced flared ribs usually coincide with tubercle spacing and are therefore tuberculate. Remarks. Of the 17 fragmentary specimens of this species, Io are dextrally coiled and 7 sinistrally coiled. Only a few are not distorted or crushed. Four of the smallest specimens that are relatively uncrushed (Pl. 7, figs. 4, 5) show a loose helical spiral that must be close to the original shape of the conch. Two of the medium-sized specimens are crushed by pressure along the axis of the spire so that the whorls are nearly in contact, but one of these is selected as holotype for its has 14 complete whorls and shows the characters of the species better than any other specimen (Pl. 7, fig. 2). Two other specimens are crushed by pressure at right angles to the spire axis and are less well preserved (PI. 7, fig. 3). The largest specimens are only short fragments but are not badly crushed and show the whorl shape and ornament well (Pl. 7, fig. 6). The largest and smallest whorls preserved have cross section diameters of 43 mm. and 7 mm. respectively. Suture-lines are poorly preserved and difficult to follow in all specimens, but septal surfaces up to 40 mm. diameter occur in several of the large specimens, indicating that adults reached sizes at least a half to one whorl larger than the largest fragment preserved. None of them shows evidence of modified adult body chamber coiling. Significant variation from the remainder of the collection can only be seen in one specimen: C. 52694 is a short septate fragment with a whorl section diameter of 34 mm., and has particularly strongly curved and oblique ribs which do not appear to have any tubercles. Spath (1951: 8; 1953: 49) made two specific determinations for these specimens: “Bostrychoceras polyplocum (Romer) Schliter, pars’ and “‘Bostrychoceras sp. nov. (cf. punicum?, Pervinquiere)”’. The second determination can be discarded for it was presumably given to some of the smaller specimens, which after extraction from the matrix have proved to be the same as the remainder of the collection. The reference to part of Schliiter’s interpretation of Didymoceras polyplocum is presumably to the bituberculate and loosely coiled specimens (Schliiter 1872: pl. 33, fig. 6, ? pl. 34, fig. 1, pl. 35, fig. 8) that have been shown above (p. 372) to be different from the restricted D. polyplocum. There are no flared ribs on these specimens and only one of them (pl. 35, fig. 8) has a constriction on what appears to be the body chamber, so they are probably specifically distinct from the flared and constricted Angolan forms. In fact the 17 specimens of D. subtuberculatum show little variation, and the 376 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA specific diagnosis drawn from them is certainly not wide enough to include the German forms. There are few other Campanian or Maastrichtian forms that are closely comparable with D. subtuberculatum. The Pondoland and Zululand specimens figured by Woods (1906: 339, pl. 42, figs. 4, 5) and Spath (1921a: 252, pl. 24, fig. 2) appear to be loosely coiled and two of them have flared ribs, but none has any tubercles and they are only small fragments. Several loosely coiled examples from Madagascar figured by Collignon (1932: 40, pl. 9, fig. 4; 1938: 87-88, pl. 5, fig. 4, pl. 6, fig. 2) are also without tubercles or flares. At first sight D. subtuberculatum seems to resemble several Turonian to Santonian species of Hyphantoceras, such as the Japanese species H. venustum (Yabe 1904: II, pl. 5, figs. 1, 2 (holotype), pl. 3, fig. 4—Euhyphantoceras maestrichtiense Shimizu 1935, 1s an objective synonym and is Santonian, not Maastrichtian) and the north American species H. buttense, H. ceratopse and H. laqueum described and figured by Anderson (1958: 207-210), all of which might be synonyms of H. venustum (see Matsumoto 1959@: 158). There is even some resemblance to Schliiter’s figures (1872, pl. 32, figs. 13-20) of the type species Hyphantoceras reussianum (d’Orbigny). In Hyphantoceras, however, the arrangement of the ornament is different; the flared ribs are more frequent and each bears 2 to 4 tubercles, while all the ribs between the flares are non-tuberculate; in D. subtuberculatum the flares are fewer and more widely spaced, and tubercles occur at smaller intervals on non-flared as well as flared ribs. Didymoceras cf. californicum Anderson Plate 8, fig. 1 1958 Didymoceras californicum Anderson: 197, pl. 72, fig. 6. MATERIAL. One specimen, C. 52727, from Carimba, Angola. DEscRIPTION. The specimen consists of one and a quarter coiled whorls and part of another smaller whorl. The whorls are in contact, and siphuncle and septa occur up to the largest stage preserved. The whorl section is approximately circular and the diameter of the largest whorl is 23 mm. The ornament consists of simple ribs, approximately 42 per whorl, which cross the venter inclined at an angle to the whorl but roughly parallel to the axis of the spire. The ribs bear two rows of small insignificant tubercles; the upper row is exactly along the line of the siphuncle, while the lower row is some distance below this. Only one rib bifurcates at a tubercle in this specimen, all the remaining ribs being single. REMARKS. This specimen and Anderson’s species are readily compared with Didymoceras polyplocum (Romer) and D. elongatum (Whiteaves). The differences between the three species are mainly the density of the ribs, and the occurrence of small tubercles on the septate whorls of D. californicum. At approximately the size of the specimen described here, the rib density is 25 per whorl in the lectotype of D. elongatum (Usher 1952: 105, pl. 28, fig. 3), about 42 per whorl in the present specimen, and 55-60 per whorl in D. polyplocum (Rémer 1841: pl. 14, fig. 1; Schliiter CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 377 1872: pl. 33, figs. 3-5). The Angolan specimen differs from the other two species by its regular small tubercles (in D. polyplocum tubercles are irregular and rare on all whorls before the body chamber). Anderson’s holotype (the only specimen) came from the Upper Campanian or Lower Maastrichtian of California, and shows the uncoiled body chamber commencing at about the maximum size attained by the Angolan specimen. Agreement in whorl section, coiling and ornament is close, except that rib bifurcation at tubercles is probably more common in the Californian specimen. The only comparable African specimens are those figured by Basse (1931: 18, pl. 1, figs. 16, 17) from Madagascar and by Reyment (1955: 15, pl. 1, fig. 4) from Nigeria; both are Maastrichtian, but are fine-ribbed and closer to D. polyplocum than to the Angolan specimen. The Texan species D. secoense (Young 1963: 42, pl. 3, figs. 1-5, pl. 4, figs. 4, 8; Adkins 1928: pl. 37, figs. 1, 3) differs from D. folyplocum only in its possession of regular bituberculation on every second or third rib. It resembles D. californicum but its rib density is greater. It seems unlikely that D. californicum is a synonym of D. hornbyense (Whiteaves) as Claimed by Matsumoto (1960: 54), for it has a considerably smaller apical angle, its tubercles are much smaller, and it shows no evidence of slow loosening of the last one or two septate whorls before the body chamber as in D. hornbyense. Didymoceras cf. hornbyense (Whiteaves) Plate 8, fig. 4 1895 Hetevoceras hornbyense Whiteaves: 316. 1903 Hetevocervas hornbyense Whiteaves: 332, pl. 42, figs. 1-4. 1921a Didymoceras hornbyense (Whiteaves) Spath: 251. 1925 Didymoceras hornbyense (Whiteaves); Haughton: 276, pl. 15, fig. 2. 1952 Nostoceras hornbyense (Whiteaves) ; Usher: 103, pl. 27, figs. 1, 2, pl. 28, fig. 2, pl. 31, fig. 23. MATERIAL. One body chamber fragment, C. 52737, from Barra do Dande, Angola. Remarks. The single specimen is well preserved with neither distortion nor crushing, and consists of a quarter of a whorl of body chamber 95 mm. long with the last septum preserved. The whorl section is circular, 27 mm. diameter at the smaller end, 33 mm. at the larger end. The coiling is dextral. Towards the smaller end the venter is considerably eroded and the ribs and tubercles almost removed. On the dorsum the ribs are reduced to small striae. Pairs of large tubercles occur irregularly on every third or fourth rib, and there are some looped ribs between the tubercles. Another fragmentary example from Angola, figured by Haughton, is smaller than the present specimen, but has closely similar ornament. Both specimens show the typical characters of D. hornbyense as seen in Whiteaves’ original specimens and Usher’s (1952: 103) revision. The example figured here compares well with the largest figured by Usher (1952: pl. 27). Reference to Didymoceras rather than to Nostoceras is favoured on account of the large size of the body chamber, which does not form the hanging U-shaped body chamber characteristic of Nostoceras. Its possible relationship with N. helicinwm is discussed below. 378 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA Didymoceras cf. angolaense (Haughton) Plate 8, fig. 2 1925 Nostocervas angolaense Haughton: 275, pl. 15, fig. 1. ?1943 + Nostocevas cf. angolaense Haughton; Haas: 5-6, figs. 2, 8. MATERIAL. One specimen, C. 52739, from Barra do Dande, Angola. ReMARKs. The specimen consists of four whorls closely coiled into a dextral helical spire of small apical angle. The whorl section is rounded between the ribs, but is angular over the tubercles as described by Haas (1943: 5, fig. 2). The ribs, tubercles and constrictions are similar to those in the type specimen described by Haughton, although the 20-22 tubercles in each row of the present specimen appear to be slightly more than in the holotype. There are 33 ribs on the last whorl, but no suture-lines can be seen in this specimen, which is preserved as a limonite-stained shell filled with crystalline calcite. The specimens described by Haas are fragmentary and have no distinct ribs, and cannot be referred with certainty to this species. The species is referred to Didymoceras rather than to Nostoceras because the last whorl of the holotype is loose, and does not change suddenly to the downwards curving body chamber typical of Nostoceras. The most closely related species are D. splendidum (Shumard) (Stephenson 1941: 415, pl. 82, figs. 1-4) from the Lower Maastrichtian of Texas and D. excelsus (Anderson 1958: 194, pl. 72, fig. 4) from the top of the Campanian or the Lower Maastrichtian of California. Both these species differ in having even more acutely angled spires, a pair of tubercles on each rib and no non-tuberculate ribs as in D. angolaense. Genus NOSTOCERAS Hyatt 1894 Nostoceras hyatti Stephenson Plateio; Ply 10) tigen Vext-fies 16 1892 Heteroceras convadi (Morton); Whitfield: 269-271, pl. 45, figs. 12, 13, non figs. 9-11, 14. 1907 Heteroceras conradi (Morton); Weller: 833, pl. 108, figs. 5-8. 21935 ‘“‘Hamites’’ vancouverensis Gabb; Anderson & Hanha: 23, pl. 7, figs. 2-4, pl. 8, fig. 5. 1941 Nostocevas hyatti Stephenson: 410, pl. 81, figs. 9-12. 1951 Didymoceras sp. nov. ind., Sornay: 274, pl. 4, fig. 4. 1962 Nostocevas sp., Reeside: 119, pl. 69, figs. 7-12. MATERIAL. I0 specimens, C. 52740-43, C.52747-52, from Barra do Dande, Angola. Description. The collection consists of seven looped body chambers, one having the last spiral whorl preserved, the other six having a short septate portion or the last septum preserved and the entire spire missing, and three fragments of body chambers. The spire is coiled dextrally in four specimens and sinistrally in three (the coiling is not determinable in the three fragments). The last whorls of the spire are in contact right up to the point at which the rapid change to the U-shaped body chamber occurs. The two limbs of the body chamber are nearly straight and close together, leaving only a narrow gap between them which diminishes in width at the CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 379 level of the mouth border. The length of the body chamber varies slightly, for in some the last septum is at the beginning of the first straight arm opposite the mouth border, while in others it occurs earlier at about the position at which the whorl breaks away from the closely coiled spire. The axis of the spire is inclined at a small angle (c. 20°) to the plane of the body chamber. The whorl section is approximately circular throughout although the diameter at right angles to the plane of the body chamber increases on the middle part of the body chamber in most examples. The ribs are sharp and mainly single throughout but occasional irregular bifurcation occurs at the tubercles or at the edge of the dorsum, and irregular looping or zigzagging occurs between some of the tubercles. The strength of the ribbing is considerably reduced on the dorsum of all whorls. On the body chamber the ribs are markedly stronger and more widely spaced than on the spiral whorls but they tend to increase in density again on approaching the mouth border. On the spiral whorls the ribs are inclined at a small angle to a whorl section plane; on the body chamber ribs are generally more radial (i.e. annular) but they are somewhat irregular and on the back part (see descriptive terms for helically coiled ammonites p. 374 above) of the first straight limb the ribs are arched upwards and particularly strong. Narrow constrictions occur at wide intervals on the spiral whorls, but not on the main part of the body chamber. The mouth border is immediately preceded by a constriction, then a collar-like rib, followed by a narrow flat portion up to the slightly sinuous mouth border. Two rows of medium sized tubercles occur fairly constantly on alternate ribs on the spiral whorls; the upper row overlies the siphuncle, the lower row occurs just below the line of contact between adjacent whorls. On well preserved parts complete with the shell the tubercles are elongated into short pointed spines. On the body chamber the tubercles occur on every rib, gradually L 4 ph Pd Fics. 16,17. Suture-lines of Nostoceras. Fig. 16. Nostocevas hyatti Stephenson. Last suture- line of an adult at 21 mm. ventro-dorsal diameter. C. 52743, from Upper Campanian, Barra do Dande, Angola. x3. Fig. 17. N. obtusum sp. nov. Suture-line at 10 mm. ventro-dorsal diameter. Holotype, C. 52744, from Upper Campanian, Barra do Dande, Angola. x5. 380 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA increasing in size to become large round the U-shaped bend, then rapidly diminishing towards the mouth border. The line of the two rows moves outwards on the first part of the body chamber until they are situated on the periphery of the main bend and the final limb of the body chamber. One dextrally coiled specimen shows half a complete suture-line lying to the right of the siphuncle and the left of the dorsum (i.e. on the “top” of the whorl) (Text-fig. 16). The first and second lateral saddles are similar in shape and size, both being divided into two by a minor lobe of moderate depth then each half divided into two again. The dorso-lateral saddle is smaller, but is also bifid and is bounded by the short narrow dorsal lobe in the middle of the dorsum. The first and second lateral lobes are large and deep and each is divided by a large minor saddle. The largest and smallest specimens have mouth border diameters of 27 mm. and 22 mm., while the transverse and ventro-dorsal diameters in the middle of the loop of the largest specimen are 29 mm. and 23 mm. respectively. The distance between the mouth border and the lowest point of the periphery of the loop varies between 45 and 65 mm. The largest diameter of the single helically coiled whorl preserved is 37 mm. at the point of break away of the body chamber, and at this point the cross section diameter is 17 mm. REMARKS. None of the determinations given by Spath (1951: 10; 1953: 50) appears to fit these specimens. Stephenson’s originals consist of both spire and body chamber from the top Campanian or Lower Maastrichtian of Texas, and the Angolan examples agree with them in all respects. The other specimens listed in the synonymy are all body chambers only. Whitfield, Weller and Reeside figured five fine body chambers from New Jersey that show the normal characters of N. hyatti; the interpretation of Heteroceras conrad: (Morton) has been discussed above (p. 371)—it cannot be shown to be a Nostoceras and it is certainly not conspecific with N. hyatt. The two body chamber fragments from California figured by Anderson & Hanha also agree closely with N. hyatt; the interpretation of Hamites vancouverensis Gabb (1864: 70, pl. 13, fig. 18) is difficult because the holotype is a small fragment, and again it is most unlikely to be the same as N. /yatti: because of its larger size and different tuberculation. Matsumoto (1960: 54) was probably correct in uniting H. vancouverense, Didymoceras fresnoense and Exiteloceras bennisoni of Anderson (1958: 197, pl. 68, fig. 2; 201, pl. 72, fig. 7); all differ from Nostoceras in their open or rounded-U-shaped body chambers, and belong to either Dzdymoceras or Emperoceras. Another North American top Campanian or Lower Maastrichtian species, Ammonites cooperi Gabb (1864: 69, pl. 14, fig. 23), included in H. vancowverensis by some authors, is also difficult to interpret because of a fragmentary holotype. The interpretations of Whiteaves (1903: 336, pl. 43, fig. 1) and Usher (1952: 107, pl. 29, fig. 1) are probably correct and put A. cooperi into the genus Emperoceras. Finally the single body chamber figured by Sornay (1951: 274, pl. 4, fig. 4) is from Barra do Dande, Angola, and is a good example of N. /yatti. It closely resembles one of the examples figured here (PI. 10, fig. 1). Nostoceras lvyatti is characterized by its large size and bold ribs and tubercles. N. sternbergi Anderson & Hanha (1935: 22, pl. 7, fig. 1) differs in having finer ribs CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 381 and smaller tubercles, N. dvaconis (Stephenson 1941: 413, pl. 82, figs. 5-9) has finer ribs and a depressed spire of high apical angle, N. kernense (Anderson 1958: 196, pl. 65, fig. 1) has regular plicate ribbing, and N. mexicanwm (Anderson 1958: 196, pl. 58, fig. 3) has finer ribs and small irregular tubercles. The species is known only from Texas, New Jersey, California and Angola. Nostoceras cf. kernense (Anderson) Plate 8, fig. 6 1958 Didymoceras kernense Anderson: 196, pl. 65, figs. 1, 2. MATERIAL. One specimen, C. 52746, from Barra do Dande, Angola. Remarks. This fragment of a U-shaped body chamber is compared with Anderson’s species because of its markedly branching ribs. It can be seen from the figure that the position of branching, density and angle of the ribs show a close resemblance to Anderson’s holotype. In some places the branching is virgatotome with up to three secondary ribs leaving the primary rib in succession. The Angolan specimen differs in being considerably smaller and by having shorter straight arms on the body chamber (judging from the rapid narrowing of the gap between them). In a few places where the shell is preserved the ribs can be seen to cross the venter as in Anderson’s holotype, but most of the specimen is an internal mould on which the band between the rows of tubercles is nearly smooth. The only comparable species is N. stervnbergi Anderson & Hanha (1935: 22, pl. 7, fig. 1) from California. This shows similar multiple rib branching, but it has constrictions and differs in details of ribbing, including a sudden change to bold ribs on the final straight limb. Nostoceras rotundum sp. nov. Plate 10, fig. 3 1951 Didymoceras angolaense Sornay: 274, pl. 4, figs. 1-3. HOLOTYPE. C. 52745, the only specimen, from Barra do Dande, Angola. Diacnosis. Medium sized species of Nostoceras, with close-coiled helical whorls, followed by a hanging body chamber consisting of a semicircular loop. Ornament on helical coils consists of 14 to 15 tubercles per whorl and weak ribs; tubercles and moderately strong ribs on body chamber. DESCRIPTION. The single specimen in the present collection consists of three whorls closely coiled into a dextral spire, followed by a body chamber loop which breaks suddenly away from the spire and is angled obliquely downwards forming a nearly perfect semicircular loop; the mouth border faces obliquely upwards towards the last whorl of the spire. Suture-lines are too poorly preserved to reveal details, but the last suture-line is clearly seen to occur on the penultimate whorl directly 382 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA above the point at which the loop leaves the spire, so that the body chamber consists of the whole of the last whorl of the spire plus the loop. The angle between the plane of the loop and the axis of the spire is 45°. Whorl sections in the spire are roughly circular with a small flat portion at the position of contact between whorls, and in the loop the transverse diameter is slightly greater than the ventro-dorsal diameter. Ornament consists of obliquely aligned pairs of tubercles forming two rows. On the spire the upper row forms the outermost point of the whorl and contains 14 tubercles per whorl, while the lower row occurs above the position of contact between whorls and contains 15 tubercles per whorl. The tubercles increase slightly in strength on the loop and the rows twist so that the lower row occurs around the outermost periphery of the loop, while the upper row goes onto the back of the loop. Ribs are poorly developed on all whorls. Between the rows of tubercles on the spire and the loop only vague undulations occur, but above and below the rows of tubercles weak ribs occur on the spire, and these strengthen on the loop to form simple curved ribs above the upper row and both simple and plicate ribs below the lower row. The dorsum of the loop is smooth. The siphuncle occurs just above the upper row of tubercles. No constrictions occur on any part, but just before the end of the body chamber the whorl contracts laterally, then flares out and ends in a gently sinuous mouth border of exactly circular section. Diameter of the mouth border 23 mm. ; diameter of the semicircle of the loop 52:5 mm.; total height of specimen as preserved 88 mm.; total height extrapolated to apex 110-115 mm.; diameter of the final spiral whorl 40 mm. REMARKS. From the description and figures of Sornay’s single specimen from Barra do Dande it is difficult to identify his specimen with the one figured here. However Sornay’s original was kindly made available on loan by Dr. L. Cahen, Director of the Musée royal de 1’ Afrique centrale, Tervuren, and the great similarities between the two were then revealed. Sornay’s specimen consists of three-quarters of a whorl coiled into a sinistral helical spire, followed by a downwards twisting portion, then two-thirds of a semicircular loop. On the spiral whorl there is a marked depression where the whorls were in contact right up to the point where the body chamber breaks away suddenly from the spire. The final suture-line is at the beginning of the spiral whorl, so that the body chamber occupied three-quarters of the last whorl of the spire plus the loop. From a comparison with the holotype it is probable that the mouth border is only just missing; the pair of tubercles before the broken aperture are probably the last ones. The tubercles and ribs agree exactly with those of the holotype. The diameter of the final spiral whorl is 54 mm., and the cross section diameter close to the mouth border is 29 mm. The two specimens described above agree in having a semicircular loop with no straight arms, a body chamber that occupies the last whorl of the spire as well as the loop, bold ribs on the last part of the loop, and large tubercles throughout. These characters serve to distinguish the species from N. /yatti which has distinct straight arms in the loop, a body chamber occupying only the loop, and ribs dominant over tubercles on all but the bend of the loop. Reference of N. votwndwm to Nostoceras or Didymoceras is arbitrary, and it is included here in Nostoceras because of its tightly CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 383 coiled spire and large tubercles. The new specific name, N. votundum, is necessary because NV. angolaense (Sornay 1951) is pre-occupied by N. angolaense Haughton, 1925. Specimen C, 52745 is chosen as holotype rather than Sornay’s larger specimen because it is more complete and free from matrix. N. mariateresianum Haas (1943: 6, 7, figs. 1b, 9), a closely related species from Angola, is known only from a single fragment that has similar ornament with dominant tuberculation, but has 22 to 24 tubercles per whorl in each row and 2 to 3 constrictions per whorl. Another Angolan species Didymoceras angolaense (Haughton), has a smaller apical angle, sharp ribs and small tubercles and a loose Didymoceras body chamber. The Zululand species Nostoceras (?)subangulatum (Spath 1g921a: 250, pl. 22, fig. 3) has stronger ribs on the spiral whorls than in N. votundum, and it is close to Didymoceras (or ? Nostoceras) stevensont (Whitfield 1880: 447, pl. 14, figs. 5-8; Igor: 219, pls. 29, 30). Whitfield’s (1901) figured specimen, though much larger, shows a similar loop to that of N. rotundum. Nostoceras helicinum (Shumard) Plate 8, figs. 3, 5 1861 Turrilites helicinus Shumard: tot. 1894 Nostocevas helicinum (Shumard) Hyatt, 573. 1941 Nostocevas helicinum (Shumard); Stephenson: 410, pl. 80, figs. 11, 12. 1943 Nostocevas helicinum (Shumard); Haas: 2-5, figs. Ia, 6, 7. MATERIAL. Two specimens, C. 52738 and C. 52753, from Barra do Dande, Angola. REMARKS. Both specimens are sinistrally coiled with the whorls in contact, the larger specimen, C. 52738, consisting of one whorl with a maximum diameter of 30 mm., the smaller specimen, C. 52753, consisting of nearly two whorls, the maximum diameter of the larger being 24 mm. Although suture-lines are not well preserved the larger specimen appears to have three-quarters of a whorl of presumably immature body chamber, while the smaller specimen is septate up to shortly before its aperture. Deep narrow constrictions occur at roughly 180° intervals on both specimens, and pairs of small tubercles are present on all whorls. Two Angolan specimens were described and figured by Haas (1943: 2-5), and the two further specimens now figured agree with these in all respects. Haas’s description was much more detailed and complete than that of Stephenson (1941: 410, pl. 80, figs. II, 12) who designated the neotype of the species. The four Angolan specimens belong to the normal variety of N. helicinum, which has fine ribs and a spire angle of 80-go0°. Two varieties that have been separated are N. helicinum var. humile Stephenson (1941: 412, pl. 81, figs. 4-6) which has a more depressed spire (larger spire angle) and rather coarser ribs and tubercles, and var. crassum (Stephenson 1941: 412, pl. 81, figs. 7, 8) which has much coarser ribs and tubercles. N. stantoni and its varieties (Stephenson 1941: 407-410, pl. 80, figs. I-10) have a smaller spire angle and few tubercles, N. hyatti has a smaller spire angle and coarser ribs and tubercles, and the somewhat similar species Didymoceras hornbyense (Whiteaves) (Usher 1952: 103, pl. 27, pl. 28, fig. 2) attains much larger sizes and has larger ribs and tubercles. 384 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA Nostoceras (2?) obtusum sp. nov. Plate ro, fig. 2; Text-fig. 17 HoLotyPe. C. 52744, the only specimen, from Barra do Dande, Angola. Diacnosis. Coiled whorls consisting of depressed, obtuse-angled closely coiled spire; adult body chamber unknown. Two rows of tubercles are the dominant ornament, with small ribs crossing the whorl above and below them and low ribs connecting tubercles between rows. DeEscrRIPTION. The single specimen consists of two whorls closely coiled into a sinistral helical spire that has a large apical angle of 125-130°. The first one and a quarter whorls are septate and the final three-quarters of a whorl is body chamber, probably that of an immature specimen, for the final septa are not approximated and there are no signs of adult body chamber modification of the mode of coiling. The whorl section has angles at the position of the tubercles, a flat or slightly impressed portion at the position of contact with earlier whorls just above the dorsum, and rounded upper and lower sides. Tubercles are the dominant feature of the ornament; the upper row occurs just below the mid-ventral line and forms the outermost point of the whorl, while the lower row occurs near the middle of the lower side and is the lowest point of the whorl. On the upper side of the whorl small straight radial ribs are connected to the tubercles in pairs, and on the dorsal side of the lower row of tubercles similar small ribs are connected singly or in pairs to the tubercles. These ribs are greatly reduced in crossing the dorsum. Between the two rows of tubercles low undulations join or occasionally zigzag between opposite tubercles; by comparison with the ribbed part of the whorl this band is nearly smooth. There are no constrictions. Maximum diameter of the final whorl 47 mm. ; width of ‘‘umbilicus” on underside 20 mm.; whorl height (dorsum to top of outer tubercle) 15 mm.; the outer whorl has about 55 ribs on the upper side of the whorl, about 31 tubercles in the upper row, 25 tubercles in the lower row and 41 ribs on the lower side of the whorl. As much of the suture-line as is visible is shown in Text-fig. 17. REMARKS. The combination ofa large apical angle, giving a very depressed spire, slender whorls, angled whorl section, and large tubercles with the lower row in the middle of the lower surface, serves to distinguish this species from any other Nostoceras. Other flat whorled species and varieties, such as N. helicinum var. humile Stephenson (1941: 412, pl. 81, figs. 4-6) and N. dvaconis (Stephenson 1941: 413, pl. 82, figs. 5-7, 8, 9), have much smaller tubercles and round whorl sections. The Zululand species NV. (?) natalense Spath (1921a: 248, pl. 22, fig. 2) has much larger and more massive whorls, with large tubercles, bold ribs and a small “umbilicus” on the underside of the spire. N. obtusum is referred to Nostoceras rather than to Didymoceras because of the tight coiling and the ornament, which compare with other more completely known species of Nostoceras. CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 385 Family DIPLOMOCERATIDAE Spath 1926 Genus POLYPT YCHOCERAS Yabe 1902 Polyptychoceras pseudogaultianum (Yokoyama) late nin shion 2 1890 Ptychoceras pseudogaultianum Yokoyama: 181, pl. 20, figs. 1, 2, ?3. MATERIAL. 10 specimens, C. 52718—26 from 1 km. north of Egito, and C. 52754 from Barra do Dande, Angola. DESCRIPTION. The single specimen from Barra do Dande consists of a straight arm 57 mm. long, followed by a complete U-bend and a short portion of the next larger straight arm 11 mm. long. The whorl section is nearly circular throughout, being 6:5 mm. diameter at the smaller end and 9 mm. X I0 mm. at the larger end. The two arms are close together, the maximum width of the gap between them near the hook being only I mm. The ribs are relatively widely spaced and are broad and flattened on the internal mould. On the side of the whorl they are inclined slightly forwards towards the venter, which they cross unchanged, but the dorsum is smooth or is crossed by striae only. Immediately before the hook there is a constriction preceded by a collar on the venter. On the hook the ribs are smaller and more striate, and just beyond the hook there is a second constriction on the short portion of the larger arm. No suture-lines are visible on this specimen. The other specimens are all fragments of straight arms up to 50 mm. long and have dimensions similar to the single specimen described above. All have slightly oblique broad ribs, and two of them have shallow constrictions between two adjacent ribs. Septal surfaces and fragments of suture-lines are present in several specimens. REMARKS. Spath (1953: 49, 50) determined the Barra do Dande specimen as Phylloptychoceras sp. nov. and the Egito specimens as Polyptychoceras cf. pseudo- gaultianum (Yokoyama). However, all are clearly conspecific and are referred here to Yokoyama’s species, from which they show no significant differences. Yokoyama (1890: pl. 20, figs. 1-3) figured three syntypes, the two largest (figs. 1, 2) being comparable in size with the Angolan specimen and showing the same type of ribs, while the smallest (fig. 3) is more densely ribbed and has occasional constrictions. Wiedmann (1962: 185) referred this fine-ribbed syntype to the north German Upper Santonian and Campanian species P. (?) obliquecostatwm (Schliiter). Whether this is correct or whether the specimen falls within the variation of P. pseudogaultianum must await the analysis of a larger Japanese topotype collection, and also a proper generic assessment of Schliiter’s species which is known only from short straight fragments. P. pseudogaultianum occurs in both the Santonian and Campanian of Japan. Four other Japanese species, P. havadanum (Yokoyama), P. subquadratum (Yokoyama), P. subundulatum (Yokoyama) and P. obstrictwm (Jimbo) differ in size and details of ribbing. P. vancouverense (Whiteaves) (Usher 1952: 101, pl. 26, figs. 5, 6) is a closely related species from the Upper Campanian of British Columbia. It is slightly larger and has 386 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA more widely spaced, flattened band-like ribs than P. pseudogaultianum. The Graham Land specimens described by Spath (1953: 18, pl. 7, fig. 5) as Polyptychoceras sp. Juv. indet. are indeterminable and could equally well be Glyptoxoceras or Diplomoceras. Subptychoceras has ribs arranged in groups on low bulges and Phylloptychoceras has undulating folds on the sides of the whorl and some striae; both are best con- sidered subgenera of Polyptychoceras. The lectotype of Phylloptychoceras sipho (Forbes), the type species of the subgenus, is figured here (Pl. 11, fig. 1) because previous determinations of this species have had to rely on the inadequate drawings of Forbes (1846: 118, pl. 11, figs. 5a-g) and Stoliczka (1865: 194, pl. go, figs. 5-9), and the figure of the suture-line given by Spath (1953: pl. 11, fig. 7). Family DESMOCERATIDAE Zittel 1895 Subfamily PUZOSIINAE Spath 1922 Genus KITCHINITES Spath 1922 Kitchinites angolaensis sp. nov. Plate 11, figs. 4-6 HOLotyPe. C. 52675 (Pl. 11, fig. 5), from 1 km. north of Egito, Angola. MATERIAL. In addition to the holotype, 8 paratypes (C. 52676—-83) all from 1 km. north of Egito, Angola. DIMENSIONS: C. 52675. At 64 mm.: 28-7, —, 16:5. C. 52680. At 48-5 mm.: 22:0, —, II‘9. DiaGnosis. Whorls moderately involute, inner whorls about one half concealed. Whorl section compressed, with only slightly convex whorl sides, vertical umbilical walls and an angled umbilical edge. On whorls up to 40 mm. diameter the ornament consists of fine, slightly sigmoidal ribs which curve gently forwards on approaching the venter; primary ribs cross whole side of whorl and intercalated secondaries occur on ventral half only. Between 40 and 50 mm. diameter ribs gradually fade on inner half of whorl leaving ribs near the venter only. At larger sizes ventral ribs show marked increase in strength. 4 to 6 constrictions per whorl are present, but poorly developed; on inner whorls they are of similar shape to the ribs but inclined more strongly forwards and cut across the ribs; at larger sizes they are nearly straight on side of whorl and curve slightly forwards on the venter. On outer surface CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 387 of shell, constrictions usually represented by or preceded by a collar on the venter. Suture-lines not visible in detail. REMARKS. These nine specimens were identified by Spath (1951: 8) as “Gen. nov. (Kitchinites ?) sp. nov.”, but their generic characters are those of Kitchimites and there are no adequate reasons for separating them. K. angolaensis is characterized by fine sigmoidal ribs at small sizes, followed at larger sizes by smooth whorl sides and considerably stronger ventral ribs. All the specimens are partly crushed making estimation of the whorl thickness difficult. The most closely related species is K. darwint (Steinmann 1895: 73, pl. 5, fig. 3) from the Quiriquina Beds of Chile, which has the same smooth whorl sides, but differs in its larger umbilicus, thicker whorls with more convex sides and its much stronger constrictions. The New Zealand species K. brevicostata (Marshall 1926: 183, pl. 24, fig. 3, pl. 43, fig. 2) is also close to K. angolaensis, but it has a larger umbilicus and nearly straight ribs that do not fade on the side of the whorl. The type species Kvtchinites pondicherryanus (Kossmat 1897: 40, pl. 6, fig. 6) has much stronger straight ribs throughout and there is no evidence of the ribs fading on the sides of the whorl. The genus Neopuzosia Matsumoto 1954, was proposed for the two Japanese species, NV. japonica (Spath), the type species, and N. ishikawaz (Jimbo) (see Matsu- moto 1954: 89-95), which have sigmoidal ribs, at least on the inner whorls, that are strongly projected on the venter. The whorl is generally thicker and not so flattened as in Kitchinites. Neopuzosia is now generally admitted as a subgenus of Kitchinites, but K. brevicostata mentioned above is intermediate between the two in most of its characters, and now K. angolaensis shows mixed rather than intermediate characters. Its sigmoidal and fine ribs are like those of Neofuzosia, its compressed and flattened whorls are like those of Kitchinites, while its smooth whorl sides at larger sizes are shared only with K. darwint, usually placed in Kitchinites s.s. Neopuzosia is Santonian and Lower Campanian in age, Kitchinites s.s. ranges from Campanian to Lower Maastrichtian, and K. angolaensis is Upper Campanian. The horizon of K. darwimi is not accurately known. Proposal of further generic names will confuse what is probably a closely related group of species; K. angolaensis is referred here to Kitchinites s.1., rather than to either subgenera which are of doubtful value. Five species of Kitchinites s.1. from the Lower Campanian of Madagascar were described by Collignon (1961: 55-58). All of them have considerably thicker whorls than any of those listed above, but from their convex whorl sides and ribs projected on the venter they would probably be referred to Neopuzosia rather than to Kitchimites s.s. Of Collignon’s five specific names, K. busnardot, K. quadratus and K. fascigerus (Collignon 1961: pl. 6, figs. 3, 4, pl. 23, fig. 3) represent a species with coarse ribs, while K. flabelliformis and K. enayi (Collignon 1961: pl. 6, fig. 5, pl. 23, fig. 2) represent a species with much finer ribbing. The ornament of the latter species is hardly distinguishable from that of K. angolaensis, but the whorl thickness of the Madagascan species must be about twice that of the Angolan species, even allowing for the crushing in the latter. The Upper Turonian “Neopuzosia”’ matsumotot figured by Collignon (1961: 54, pl. 23, fig. 1) appears to be a Mesopuzosza close to the holotype of M. pacifica Matsumoto (1954: 82, pl. 15, fig. I). 388 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA Subfamily DESMOCERATINAE Zittel 1895 Genus DESMOPHYLLITES Spath 1929 Desmophyllites diphylloides (Forbes) Plate 11, fig. 3 1846 Ammonites diphylloides Forbes: 105, pl. 8, fig. 8. 1953 Desmophyllites diphylloides (Forbes); Spath: 21, 49, pl. 2, figs. 5, 6. 1955 Desmophyllites diphylloides (Forbes); Matsumoto & Obata: 121, pl. 24, figs. 1-5, pl. 30, fig. I. 1959b Desmophyllites diphylloides (Forbes); Matsumoto: 9, pl. 3, fig. 3. 1961 Desmophyllites diphylloides (Forbes) ; Collignon: 61-65, pl. 24, figs. 4, 5, pl. 25, figs. 1-8. MATERIAL. 8 specimens, C. 41475 and C. 52661-67, from 1 km. north of Egito, Angola. DESCRIPTION. The eight specimens vary in size from 20 to 39 mm. diameter and all are wholly septate, but only three are well preserved and free from lateral crushing. Dimensions of these three are as follows: C241473, “At 30mm): 1675, 12°5).2-7- C. 52663. At 28-5 mm.: 15:0, 11°8, 2:5. C. 52664. At 20 mm.: 10-7, 8:3, —. All the specimens are unornamented except for constrictions on the internal mould which are biconcave forwards on the sides of the whorl and are projected forwards on the venter. There are 6 or 7 constrictions per whorl. REMARKS. Full synonymy and description of this species have been given by Matsumoto & Obata (1955), Matsumoto (1959b) and Collignon (1961). The best of the Angolan specimens and the largest of Forbes’s three paratypes were figured by Spath (1953, pl. 2, figs. 5, 6), and the lectotype is now figured (Pl. 11, fig. 3) for the first time since Forbes’s original drawing. D. diphylloides shows a considerable amount of variation in whorl dimensions, and strength and shape of the constrictions. Collignon (1961: 61—65) has expressed this variation by dividing the Madagascan specimens into four varieties; var. besairiei differs from the normal variety in having the constrictions prolonged into a long narrow rostrum on the venter, var. imevmis has greatly reduced constrictions which are sometimes almost absent, and var. lata has a more compressed whorl shape, of which the dimensions listed by Collignon (1961: 64) fall just below the lower limit of the species indicated on the whorl height/breadth graph of Matsumoto & Obata (1955: 124). Undoubtedly these varieties express the normal amount of variation that is now held to occur in many species of ammonites. The Angolan specimens show about average characters of the species and would be referred to the normal variety. The species has a world-wide distribution in India, Japan, western north America, Angola, Madagascar and south-east Africa. It has a relatively long range for an ammonite species, which Matsumoto (1959): 11; 1959c: 70) gave as the whole of the Campanian in Japan and extending up into the Lower Maastrichtian in some other CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 389 areas. Collignon (1961: 61-65), however, recorded many of his specimens from both Lower and Upper Santonian, and it seems that a range from Lower Santonian to Lower Maastrichtian must be admitted. Subfamily HAUERICERATINAE Matsumoto 1938 Genus OIOPHYLLITES Spath 1953 Oiophyllites angolaensis Spath 1953 Otophyllites angolaensis Spath: 21, pl. 6, fig. 6. HoLotyPe. C. 41476, the only specimen, from 1 km. north of Egito, Angola. Remarks. The holotype of this species has been adequately figured by Spath (1953: pl. 6, fig. 6), and consists of wholly septate and somewhat eroded inner whorls of 22 mm. maximum diameter. No part of the shell is preserved and the surface of the internal mould shows no trace of ornament. The whorl section is elliptical with the greatest width near the umbilical edge, and the venter is smoothly rounded with no keel, although this does not exclude the possibility of a keel occurring at this size on the external surface of the shell. Dimensions: at 22 mm.: 10-0, 6:2, 5-7. This specimen was referred to Ozop/yllites because of its resemblance to the five Graham Land specimens of O. decipiens Spath (1953: 21, pl. 4, figs. 7, 8), the only other species of the genus. The Graham Land specimens are even more poorly preserved, but the shell is present in some places and shows that there is no keel on the shell at 22 mm. diameter, and that sigmoidal striae cover the surface of the shell. There are no constrictions. Matsumoto & Obata (1955: 136-137, text-fig. 6) refigured two of the Graham Land specimens and suggested that the genus was an offshoot of Hauericeras, and Collignon (1961: 21) has relegated Ozophyllites to a subgenus of Hauericeras. The generic status and position of Ozophyllites cannot be decided until much larger and better preserved material is available. The single specimen of O. angolaensis may be merely the inner whorls of Hawuericeras as suggested by Matsumoto & Obata (1955: 137), for larger specimens might reveal the presence of a keel at a later growth stage. The Graham Land Ozophyllites were associated with a fauna containing Maorites, a genus that can be accurately dated as Lower Campanian in Madagascar. Family PACHYDISCIDAE Spath 1922 Genus EUPACHYDISCUS Spath 1922 Eupachydiscus pseudogrossouvrei Collignon Plate 12, figs. I, 4 1931 Pachydiscus grossouvrei Kossmat; Basse: 26, pl. 3, figs. 8, 9 (non pl. 2, figs. 16, 17). 1932 Parapachydiscus besairviei Basse; Collignon: 28, pl. 8, fig. 2. 1955 Eupachydiscus pseudogrossouvrei Collignon; 42, pl. 8, figs. I, 2. MATERIAL. 7 specimens, C. 52668—74, from 1 km. north of Egito, Angola. DESCRIPTION. All the specimens are crushed and distorted to some extent, but in places the shell and ornament are well preserved. The largest is 105 mm. diameter 390 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA and has the best preserved whorl shape with the following approximate dimensions: at 85 mm.: 40 (0-47), 40 (0:47), 20 (0:24). The whorl section is circular with smoothly rounded umbilical walls. The ribs are radial and nearly straight on the sides of the whorl, and curve forwards to form a slight projection on the venter. In most cases long and short ribs alternate, the long ones starting at the umbilical seam and the short ones starting some way up the side of the whorl. There are no tubercles on the ribs at the umbilical edge. The largest specimen has 33 or 34 ribs on its outer whorl at 105 mm. maximum diameter. Septa and suture-lines are poorly preserved and are only seen in a few places, and the largest specimen is septate at its maximum size. REMARKS. Spath (1951: 8; 1953: 49) determined these specimens as “Eupachy- discus sp. (cf. havadat, Jimbo sp.)’’. E. havadai (Jimbo) has been described at length by Matsumoto (1954a: 281-287, pl. 8, fig. 2, pl. 9, pl. ro, figs. 1-3; 19590: 33-38) from the Japanese type material and examples from California and Canada, and some of the Canadian examples were separated as a slightly more compressed sub- species. Madagascan specimens of E. havadai have been figured by Collignon (1938: 78, pl. 4, fig. 4; 1955: 44, pl. 9, fig. 1). All these examples differ from the Angolan specimens in having thicker whorls, stronger ribs surmounted by tubercles at the umbilical edge, and a constriction-like depression associated with some of the major ribs. The Angolan specimens have straighter and less strong ribs and no tubercles or constrictions, and they agree exactly with the species EF. pseudogrossouvrei from Madagascar. The age of this species is known to be the upper part of the Middle Campanian in Madagascar (Collignon 1955: 88-89; Besairie & Collignon 1960: 77-78). Another Madagascan specimen was separated by Collignon (1955: 43, pl. 8, fig. 2) as var. undulatocostata on account of its slightly closer and gently curved ribs. The single specimen on which this variety was founded occurs somewhat lower in the Middle Campanian than the normal form, and if it is genuinely separable, then the Angolan specimens agree with the normal variety with straight ribs. Most other species of Eupachydiscus (for lists see Collignon 1955: 79) have tubercles or bullae at the umbilical edge and stronger and more widely spaced ribs, and the only one which is close to E. pseudogrossouvrei is E. launayi (Grossouvre 1894: 184, pl. 19). In France E. launayi is known only from the single type specimen from the Lower Campanian, but 15-20 specimens have been described by Collignon (1938: 60, pl. 1, fig. 2; 1955: 36-38, pl. 5, fig. 1) from the Lower Campanian of Madagascar, well below E. pseudogrossouvrei in the Middle Campanian (Collignon 1955: 89; Besairie & Collignon 1960: 78). E. launayi differs from EF. pseudogrossouvrei only marginally in having slightly higher and thicker whorls and feeble tubercles on the ribs at the umbilical edge. Specimens very close to E. launayi were described by Collignon (1955: 39, pl. 6, fig. 2) from a higher level in Madagascar and overlapping with the horizon of E. pseudogrossouvret. The ribs in this form are more dense, the tubercles are larger and the periodic larger ribs appear at an earlier stage than in E. launayt, so its resemblance to E. pseudogrossouvrei is less close. CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 391 Family PLACENTICERATIDAE Hyatt 1900 Genus HOPLITOPLACENTICERAS Paulcke 1906 TypE species. Hoplites plasticus Paulcke 1906 (ICZN Opinion 554); the lectotype of the species is Paulcke 1906, pl. 13, figs. 1, ra-d. The specific classification of Hoplitoplacenticeras is in a state of confusion owing to the apparently large amount of variation in a single species, and the nine rather poorly preserved Angolan specimens do not clarify any of the difficulties. Hoplitoplacenticeras is, however, one of the most important genera for dating the beds at Egito, and it will be useful to outline the basis on which this genus has been accurately dated. Evidence relating to the position of Hoplitoplacenticeras in the classical sections of France and Germany was summarized by Grossouvre (1901: 801-803, table 35), who found that with the exception of H. lafresnayanum (d’Orbigny) (known from only one, or perhaps a very few, specimens from the Calcaire a Baculites of Manche that contains other ammonites characteristic of the Neubergicus Zone, Lower Maastrichtian), all the other species of the genus characterize a zone at the top of the Campanian, which was named after the most typical species, H. vari (Schliiter). More recent work by Jeletzky (1951: 18, 74) has shown that the Upper Campanian is divisible into two zones, of which the lower one is the zone of H. vavi and contains all the species of Hoplitoplacenticeras. Besairie & Collignon (1960: 74-80) have summarized the accurate stratigraphical work of Hourcq (1950: 64-85) and earlier workers in Madagascar and have shown that the few specimens of Hoplitoplacenticeras in that island are confined to the H. vari Zone, taken there as comprising the whole of the Upper Campanian. Direct evidence as to the age of the H. plasticum fauna in Patagonia is poor, for the only associated forms at the same locality, Cerro Cazador f (Paulcke 1906: 235-240), are several long-ranging species of Tetragonitidae, and Psewdokossmaticeras paulcki Collignon (1955a: 44) which might be of Upper Campanian age, although most species of the genus are Maastrichtian. Hoplitoplacenticeras vancouverense occurs in the Cedar District Formation in British Columbia, which can be dated fairly accurately as Upper Campanian (Usher 1952: 38-39). A specimen of Hoplitoplacenti- ceras found in Wyoming (Cobban 1963: C60) has allowed one point in the established zonal sequence of baculitids in the western interior of the United States to be correlated against the standard Campanian sequence of Europe. Hoplitoplacenticeras cf. marroti (Coquand) late i2 howe bl ah fie. 3 1859 Ammonites marroti Coquand; 995. Cf. 1867 Ammonites coesfieldensis Schliiter: 14, pl. I, figs. 2, 3, 5, non figs. I, 4. Cf. 1867 Ammonites costulosus Schliiter: 17, pl. 2, fig. 1, non figs. 2-4. Cf. 1872 Ammonites striatocostatus Schliiter: 65, pl. 20, figs. 1-4. Cf. 1872a Ammonites vari Schliiter: 92. 392 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA Cf. 1876 Ammonites vari Schliiter; Schliiter: 160. 1894 Ammonites vari Schliiter var. marroti Coquand; Grossouvre: 118, pl. 8, fig. 3, pl. 9, figs. 2, 3. 1898 Hoplites vari var. marroti (Coquand) ; Choffat: 80, pl. 20, figs. 1-5. Cf. 1906 Hoplitoplacenticeras plasticum costatum Paulcke: 34, pl. 11, fig. 2, pl. 12, figs. 1-3, Goll, 305}, 1st, B. 1925 Hoplitoplacenticeras vari (Schliiter); Diener: 178. 1929 Hoplites cf. vari (Schliiter); Barrabé: 181, pl. 9, figs. 11-13. ?1931 Hoplites vari (Schliiter) ; Basse: 35, pl. 5, figs. 1-3, pl. 12, fig. 2, pl. 13, fig. 1. 1947 Hoplitoplacenticeras vari (Schliter); Chavan: 129, pl. 2, fig. 1. 1963 Hoplitoplacenticeras marroti (Coquand) Young: 63, pl. 2, figs. 5, 15, 17, pl. 17, figs. 3, 4, PlZontesN2 3) Plaza hoSw ANDi i mt cel MATERIAL. Two specimens, C.52684-85, from 1 km. north of Egito, Angola. DESCRIPTION. Both specimens consist of about half a single whorl, roughly 60 mm. and 45 mm. diameter, which are rather poorly preserved and slightly distorted. The whorl breadth is about two-thirds of the height and the whorl section is angled at the tubercles. Long, slightly sigmoidal primary ribs alternate with less prominent secondary ribs which commence at the middle of the side of the whorl. The ribs cross the venter but are much reduced between the ventral tubercles. There are small umbilical tubercles, small mid-lateral tubercles, moderate, clavate lower ventro-lateral tubercles, and small upper ventro-lateral tubercles. REMARKS. The holotype of H. marroti was figured by Grossouvre (1894, pl. 8, fig. 3), and this specific name has priority over H. vari Schliter (1872a) which was substituted by Schliiter for his Ammonites striatocostatus Schliter (1872:65) already preoccupied by Meneghini (1856). The type specimens of H. vari are therefore those described and figured by Schliiter (1872: 65, pl. 20, figs. 1-4.) Other specimens referred to H. vari by Schliiter are those listed as Ammonites coesfieldensis and A. costulosus in the synonymy above, but the proper interpretation of H. vavi must await a full revision of Schliiter’s types and further topotype material. Judging from the best of Schliiter’s figured specimens (1872, pl. 20, figs. 1, 2), H. vari is probably conspecific with H. marroti, but it may be found that a varietal distinction, H. marroti var. vari, is necessary. H. praematura (Imkeller 1901: 58, fig. 1) from the northern Alps may be another variety of H. marroti, but from its apparently early loss of tubercles full specific distinction may be advisable. The two Angolan specimens agree well with the holotype of H. marroti, although they differ in the development of a small mid-lateral tubercle, and the lower ventro- lateral tubercle is as large as, or larger than, the upper ventro-lateral tubercle. Their fragmentary nature and preservation does not allow further comparisons to be made. The Portuguese examples figured by Choffat, the Madagascan examples figured by Barrabé, the Palestine example figured by Chavan and the Texas examples figured by Young, all listed in the synonymy above, conform more-or-less closely with H. marroti. Further Madagascan examples figured by Basse (1931: 35, pl. 5, figs. I-3) are more compressed and have less prominent ribs. The other European species of Hoplitoplacenticeras differ markedly: H. dolbergense (Schliiter 1876: 159, pl. 44, figs. 1-4) is the closest, but its ventro-lateral tubercles are larger and the CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 393 ribs are looped to them; H. coesfieldensis (Schliiter 1867: 14, pl. I, figs. I, 4 only), and H. lemfordense (Schliiter 1876: 160, pl. 44, figs. 8, 9) have dense, well marked ribs; and H. lafresnayanum (d’Orbigny 1842: 326, pl. 97, figs. 3-5; Grossouvre 1894, pl. 23, fig. 4) is a Lower Maastrichtian species that has bold and strongly inclined ribs. Amongst the Patagonian forms of H. plasticum, H. plasticum costatum Paulcke (see synonymy) agrees with the Angolan specimens in strength of ribs and tubercles, but its ribs are almost straight, not sigmoidal. Hoplitoplacenticeras cf. costulosum (Schliiter) Platesrs: fige2 1867 Ammonites costulosus Schliiter: 17, pl. 2, figs. 2-4, non fig. I. 1872 Ammonites costulosus Schliiter; Schliiter: 66, pl. 20, figs. 5, 6. 1906 Hoplitoplacenticeras plasticum laeve Paulcke: 45, pl. 14, figs. 3, 4, pl. 15, figs. 2, 3. 1931 Hoplites (Hoplitoplacenticeras) plasticum Paulcke; Basse: 36, pl. 4, figs. 5, 6, pl. 12, fig. 3. MATERIAL. Three specimens, C.52686-88, from 1 km. north of Egito, Angola. DeEscriPTION. The three specimens are 32 mm., 27 mm. and 22 mm. diameter respectively, and the outer whorl of the largest is fairly well preserved. The whorl shape is compressed, with almost flat whorl sides tapering towards a narrow flat venter. The sigmoidal ribs are of low relief, but are broad and flat, and the inter- spaces are narrow. The umbilical tubercles are only small raised portions of the ribs. The lower ventro-lateral tubercles are clavate ends to the ribs, and the upper ventro- lateral tubercles are smaller and are situated on the venter. There are no mid-lateral tubercles. Remarks. The largest Angolan specimen compares well with the most strongly ribbed of those figured by Schliiter (1867, pl. 2, fig. 2) and with the smoothest specimen figured by Paulcke (1906 pl. 15, fig. 2), except that both Schliiter’s and Paulcke’s figures show only one ventro-lateral tubercle, while the Angolan specimens have both upper and lower ventro-lateral tubercles close together. Schliiter (1867, pl. 2, figs. 3, 4) also figured specimens in which the ribs are striate, and Paulcke (1906, pl. 15, fig. 3) figured one which develops prominent umbilical tubercles. The relationship of Schliiter’s and Paulcke’s species cannot be deduced until their respec- tive ranges of variation are worked out, and further specimens figured photographi- cally to show the type of ventro-lateral tubercles developed. Two Madagascan specimens figured by Basse (1931, pl. 4, figs. 5, 6) are similar to the Angolan speci- mens. H. vancouverense (Meek 1976a: 370, pl. 6, fig. 1; Usher 1952: 93, pl. 25, figs. I, 2) also has reduced ribs at all growth stages, but it differs in its thicker whorls and much larger vertro-lateral tubercles. H. lafresnayanum (d’Orbigny) (Gros- souvre: 1894: 121, pl. 23, fig. 4) has a similar pattern of tubercles, but it has con- siderably stronger ribs. Hoplitoplacenticeras spp. indet. MATERIAL. Four specimens, C.52689-92, from 1 km. north of Egito, Angola. 394 CRETACEOUS AMMONITES AND -NAUTILOIDS FROM ANGOLA DescriPTION. Three of the specimens (C.52689-91) are the inner whorls of an indeterminate species of this genus. The fourth specimen (C.52692) differs markedly from any hitherto described species. It is 26 mm. diameter and the preservation is sufficiently good to see that the whorl is very broad, with a height to breadth ratio of about 0-7, there are large tubercles or spines on the side of the whorl, and the wide, flat venter has four rows of small tubercles, the inner pair of rows bounding a well marked mid-ventral groove. The pattern of tubercles is similar to that of one of Paulcke’s specimens (1906, pl. 13, fig. 2), but the wide, flat, centrally grooved venter is more exaggerated and the maximum size of the specimen is only 26 mm. diameter. Family SPHENODISCIDAE Hyatt 1900 The type of subdivision of the first lateral saddle of the suture-line has been regarded by most workers as the most important generic character in this family. A primary bifurcation of the first lateral saddle has always been taken as distinctly different from a primary trifurcation of the saddle. Within the two groups thus formed genera have been separated according to the degree of indentation of the saddles and to major differences in ornament and whorl shape. The nomenclature is complicated by a considerable number of ammonites having a primary bifurcation of the first lateral saddle followed by another bifurcation of the ventral half of the saddle, the resulting pattern of “‘secondary trifurcation” remaining clear throughout growth. Such forms have been variously referred to the nearest existing genera or made the basis of new generic names. The classifications adopted by Picard (1929: 452-453), Olsson (1944: 108-112), Hourcq (1949: 113-115) and Basse (1954: 866-869) were based on these lines, where primary consideration was given to the subdivision of the first lateral saddle. Wright (1957: L437) was the first to point out that details of suture-lines can be misleading in this family, and the classification which he adopted showed a more balanced appraisal of all the characters. With the discovery of the Angolan speci- mens described below which appeared to be typical Manambolites, except that the first lateral saddle showed primary trifurcation, not bifurcation, the possibility that this character was not of generic value, and perhaps not even of specific value, required investigation. The characters of the species referred to all the sphenodiscid genera can be summarized as follows (see Wright 1957: L437 for details of nomen- clature) : 1. Luibycoceras Hyatt 1900. All saddles entire. First lateral saddle shows either bifurcation or secondary trifurcation. All species are ornamented, except L. acuto- dorsatum and the unfigured and undescribed species L. chargense Blanckenhorn (1900: 45) which are smooth. The species showing secondary trifurcation is L. acutodorsatum (Noetling 1897: 76, pl. 21, fig. 3) which has always been referred before to Sphenodiscus, but all its saddles are entire, wholly unlike even the simplest suture-line of Sphenodiscus. Paciceras Olsson (1944: II0-I12) is a synonym, and it shows the beginnings of a secondary bifurcation of the outer half of the bifid first lateral saddle. CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 395 2. Indoceras Noetling 1897. Like Libycoceras, with first lateral saddle bifid, but smooth and with rounded venter at least on the two outer whorls. 3. Manambolites Hourcq 1949. All parts of the first lateral saddle are indented and sometimes the second lateral saddle also. Remaining saddles entire. First lateral saddles either bifid or trifid, and some bifid examples have a tendency to a second bifurcation of the outer half. Smooth or feebly ornamented. The trifid species is the Angolan form described below. A considerable amount of variation in suture-line details was shown to exist by Hourcq (1949: 112, figs. 21, 22). Mzezzemceras Basse (1954: 868, pl. 17, fig. 2) is a synonym. 4. Coahwilites Bose 1927. Suture-lines generally like those of the simpler ones of Sphenodiscus and show the same narrow-necked, kidney shaped saddles. Of the three species described by Bése (1927: 279-293), the type species has a bifid first lateral saddle, while in the other two species this saddle is bifid then the outer half is bifid again. It differs from Sphenodiscus by its well marked ribs and tubercles, and rounded or flat venter at some stage. Daradiceras Sornay & Tessier 1949, is an extreme development of Coahuilites showing large ribs and tubercles, and might be considered a subgenus of Coahuilites. 5. Sphenodiscus Meek 1871. All saddles of the suture-line usually indented, but some or all of the auxiliary saddles may be entire. Saddles narrow-necked and kidney shaped in complicated suture-lines. First lateral saddle usually trifid, but examples are known where this saddle is primarily bifid, with the outer half bifid again. One series of such examples were made the basis of the genus Austrospheno- discus Olsson (1944: 108-110), and the Texan species S. plewrisepta (Conrad) has a suture-line in which the range of variation includes both trifid and bifid examples— the suture-line of an example which is clearly bifid, with the outer half again bifid, is shown in Text-fig. 22. Smooth or only weakly ornamented. The alternative to admitting this amount of variation in the suture-lines of sphenodiscid genera is the further multiplication of generic names by creating new genera for Libycoceras acutodorsatum and the Angolan species described below, according generic status to Austrosphenodiscus, and possibly creating a new genus or subgenus for the specimen of S. flewrisepta referred to above. Such a purely morphological classification would obscure relationships, and tend to separate into different genera even conspecific specimens. The age of spenodiscid genera by dating against associated ammonites of zonal value is not as well established as is often assumed, for the mere presence of any sphenodiscid has too often been taken as an indication of a Maastrichtian age. The type species of Manambolites occurs in the Middle Campanian in Madagascar and at one locality it occurs as low as the base of the Middle Campanian (Hourcq 1949: 113; Besairie & Collignon 1960: 77-79). The other species of Manambolites are probably Upper Campanian only. Evidence that the type species of Libycoceras and the associated L. chargense Blanckenhorn are Upper Campanian in age in north Africa and the Middle East was presented by Reiss (1962); they occur in the Zone of Bostrychoceras polyplocum, taken as the top of the Upper Campanian. No other 396 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA species of Libycoceras (including Paciceras) are accurately dated. Indoceras is not accurately dated against associated ammonites. The earliest species of Coahuilites are probably Upper Campanian, but later ones are undoubtedly Maastrichtian. Sphenodiscus is well dated at many localities as Maastrichtian (see pp. 403-404), and it is not known in the Upper Campanian. In the present state of knowledge derivation of the whole family from Ewulophoceras at the end of the Lower Campanian is the most likely phylogeny. Genus MANAMBOLITES Hourcq 1949 Manambolites dandensis sp. nov. Plate 12, fig: 2, Pl. 13, fig. 1; Text-figs, 18-21 1953 Gen. nov. (“Sphenodiscus’’) sp. nov. aff. Manambolites spatht, Picard sp.; Spath: 49, pl. 3, fig. 6. Horotyee. C. 41474 (Pl. 13, fig. 1), from Barra do Dande, Angola. MATERIAL. In addition to the holotype, C. 52734, C. 52736 (paratypes), and C. 52735, from Barra do Dande, Angola. D1acGnosis. Smooth or very feebly ornamented species, of which the first lateral saddle of the suture-line is divided into three by two adventitious lobes. The three parts of the first lateral saddle are slightly indented, all other saddles entire. Description. The holotype is an adult specimen measuring 134 mm. diameter at the nearly complete mouth border. The adult body chamber occupies slightly less than half a whorl and has extensively modified features. The whorl height is markedly lowered away from the true spiral; from the beginning of the body chamber the venter alters rapidly from sharp to evenly rounded, and near the mouth border becomes almost tabulate; and near the mouth border the thickness of the dorsal half of the whorl is greatly contracted. The mouth border curves gently forwards on approaching the venter, but no part of it is preserved on the venter itself. The half whorl before the body chamber is preserved complete with the shell; the umbili- cus is a pin hole, the whorl shape is oxycone with a sharp venter, and the evenly convex sides of the whorl are interrupted just before the venter by a slight rounded ridge. Sinuous growth striae cover the shell surface, and there are very low radial undulations on the outer half of the whorl side which reach as far as the low spiral ridge; at the middle of the side of the whorl there are very small radially elongated raised portions on each undulation. The final suture-line is completely exposed together with parts of the two previous ones, and these appear to be crowded though they cannot be compared with earlier ones. The first lateral saddle is divided into three by two adventitious lobes, the outer lobe is smaller than the inner one but the saddle as a whole is clearly trifid rather than bifid, and the three saddles thus formed are moderately indented. The second lateral saddle and the seven auxiliary saddles are entire. The last three auxiliary saddles just before the umbilicus are markedly retracted. CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 397 The two paratypes are wholly septate specimens. The larger one (PI. 12, fig. 2) has a maximum diameter of 75 mm., and at this size the last three suture-lines are approximated. This is considerably smaller than the 102 mm. diameter at which the final suture-lines occur in the holotype. This specimen has very low and rudi- mentary tubercles at the middle of the whorl side from which equally low undula- tions run to a very slight ridge at the side of the venter. The venter itself is sharpened to a knife edge. The smaller paratype consists of only one-third of a whorl of about 50 mm. diameter. In both paratypes the suture-lines (Text-figs. 18, 19) have the same basic pattern as in the holotype, and in the earliest suture-line visible in the smaller paratype at 38 mm. diameter, the first lateral saddle is clearly trifid. In almost every case the three parts of the first lateral saddle are indented, and all the other saddles are entire. Remarks. Three of the specimens described here formed the basis of Spath’s (1951: 9, 1953: 49) three determinations—“‘Manambolites sp. nov. aff. M. spathi (Picard)” (also “Gen. nov. (“Spenodiscus’’) sp. nov.” in 1953: 49, pl. 3, fig. 6), “Libycoceras angolaense Haughton” and “‘Libycoceras sp. nov.’’. There is a fourth specimen in the collection, C. 52735, which is a fragment of part of a whorl of roughly 120 mm. diameter, and has the ventral parts of six suture-lines (Text-fig. 21). The last three of these are noticeably closer together and they may be the adult suture- lines. Those parts of the suture-line that can be seen agree with the present species, but the specimen is broken before the first lateral saddle is reached in each case, so the specimen can only be identified as Manambolites cf. dandensis. The most closely comparable species is Manambolites piveteaui Hourcq (1949: III, pl. 3, fig. 1) from Madagascar, which has the same adult body chamber modifications, but differs in its clearly bifid first lateral saddle. The suture-line of this species shows considerable variation in details as can be seen from Hourcq’s figures (1949: 112, figs. 21, 22), but the basic pattern of a first lateral saddle is constant. It is from the Middle Campanian of Madagascar (Besairie & Collignon 1960: 77~79). The only other described species of Manambolites are M. spathi (Picard 1929: 449, fig. Io) from the Upper Campanian of Palestine, which has the characteristic suture-line with a bifid first lateral saddle, but is otherwise poorly preserved, and M. pervinguiert (Basse 1954: 866, pl. 17, fig. 2) from Tunisia (probably from the Upper Campanian), which is very close to M. piveteaw, has the same bifid first lateral saddle, but may have slightly stronger ornament. M. pervinquiert was made the type species of Mzezzemceras Basse (1954: 868) used as a subgenus of Coahwilites, but its relation- ships to Manambolites are so close (it may even be conspecific with M. piveteam, the type species of the genus) that it must be considered a synonym of that genus. “Manambolites” ricensis Young (1963: 127, figs. 8/, 9m, p, 11h, pl. 2, figs. 14, 16, 19, pl. 72, fig. 4, pl. 74, fig. 2) is, as was pointed out by its author, an enigma. It is undoubtedly from the top of the Campanian in Texas, but it has a suture-line like Paralenticeras or Eulophoceras. The trifid first lateral saddle and bifid second lateral saddle, of which all parts are well frilled (Young 1963: figs. 8/, gm), are distinctly like those of Paralenticeras of the Upper Coniacian and Lower Santonian, and resemble to some extent those of Eulophoceras which ranges as high as the Lower Campanian, 398 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA PEER, OOO ORR iS) i) Sey, Fics. 18-23. Suture-lines of Sphenodiscidae. 18-21. Manambolites dandensis sp. nov. Upper Campanian, Barra do Dande, Angola. Fig. 18, Paratype, C. 52736, at 26 mm. whorl height, x2. Fig. 19, Paratype, C. 52734, at 42 mm. whorl height, x 1-6. Fig. 20, Holotype, C. 41474, last suture-lines at 63 mm. whorl height, x 1-1. Fig. 21, First lateral saddle of C. 52735, at approx. 75 mm. whorl height, x1-5. Fig. 22. Sphenodiscus pleuri- septa (Conrad). Maastrichtian, Upper Escondido Formation, Honda Creek—Rock Crossing, Medina County, Texas. C. 53965, at 68 mm. whorl height, x1. Fig. 23. Sphenodiscus sp. indet. Maastrichtian, Barra do Dande, Angola. C. 52733, at 108 mm. whorl height. x tI-I. The most complicated suture-line of Manambolites (Hourcq 1949: 112, fig. 22-7) is considerably different. The only Sphenodiscid ammonites previously recorded from Angola are the three specimens of Libycoceras angolaense Haughton (1925: 269-270, pl. 14, figs. I-5). These have bifid first lateral saddles in every case, all their saddles are entire, and their ornament is characteristic of the genus Libycoceras. Genus SPHENODISCUS Meek 1871 Sphenodiscus sp. indet. Text-fig. 23 MATERIAL. One specimen, C. 52733, from Barra do Dande, Angola. ReEMARKS. The specimen is an internal mould and consists of one quarter of a wholly septate whorl preserved on one side only. Its maximum whorl height is CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 399 125 mm., which indicates a diameter of about 200 mm. The whorl section is oxycone and highly compressed, and there is no ornament on the internal mould. The suture- line (Text-fig. 23) has a trifid first lateral saddle, and greatly indented parts to the first and second lateral saddles. There are six auxiliary saddles of which the three ventral ones are slightly indented. Few species of Sphenodiscus have suture-lines as complicated as that of the present fragment. The large specimen of S. Jobatus (Tuomey) figured by Hyatt (1903: 66, pl. 7) is closely comparable in most characters, but its whorl section is much less compressed and its saddle endings differ in detail. Several other north American specimens figured by Hyatt (1903, pl. 6, figs. 3, 4, pl. 8, figs. 3-7, pl. 9, figs. 7-10) differ only in details of the saddle indentations and their number of auxiliary saddles. The suture-line of S. szva (Forbes 1846: 110, pl. 7, fig. 6; Stoliczka 1865: 59, pl. 33, fig. 3; Kossmat 1895: 177, pl. 22, fig. 2) is very complicated at the largest known diameter of 65 mm., and larger specimens of this species might be closely comparable with the Angolan specimen. Family NAUTILIDAE dOrbigny 1840 Genus EUTREPHOCERAS Hyatt 1894 Eutrephoceras simile Spath 1909 Nautilus blanfordianus Kilian & Reboul: 8, pl. 1, figs. 1, 2. 1953 LEutrephoceras simile Spath: 40, pl. 12, fig. 4, pl. 13, figs. 1-5, 7. 1950 Eutvephoceras egitoense Miller & Carpenter: 34, pl. I, figs. 1-4. MATERIAL. Nine specimens, C. 41480 and C. 52710-17, from 1 km. north of Egito, Angola. REMARKS. Sixteen specimens from the same horizon and locality as the present collection were described as EF. egitoense by Miller & Carpenter (1956: 34, pl. I, figs. 1-4). One specimen from the present collection (C. 41480) was figured as E. aff. simile by Spath (1953, pl. 13, fig. 7). From Miller & Carpenter’s description and figures, and from a comparison of the new Angolan material with the Graham Land material, itisclear that the Angolan specimens are EF. simile, and that E. egitoense isasynonym. A specimen from Snow Hill Island, Graham Land, obtained after Spath’s original description shows the sharp reticulate pattern of transverse and longitudinal striae in the young growth stages (20-30 mm.) much better than in any of the originals available to Spath. The transverse striae have a slight sinus in the middle of the venter. The Angolan specimens do not differ in any way from the Graham Land examples—they have a whorl height/breadth ratio of about 0-70 at 50 mm. diameter, reticulate striae on the inner whorls, an approximately central siphuncle on the inner whorls that becomes more ventral at larger sizes, and a straight, radial external suture-line that is retracted at the edge of the umbilicus, all characters typical of E. simile. As with the collection described by Miller & Carpenter (1956) the specimens are rather poorly preserved and distorted, preserved in a light brown calcareous sandstone, and they do not add anything to the description of those authors. 400 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA IV AGES OF THE FAUNAS DESCRIBED (a) Douvilleiceras fauna of Dombe Grande. The nine specimens of Dowvilleiceras mammuillatum (Schlotheim)? var. aequinodum (Quenstedt) and D. orbignyi: Hyatt from Dombe Grande, are all that now remain of more than 50 examples of Dowvilleiceras collected in 1930 and 1931 by Alexandre Borges from various localities south-west of Benguela. Borges stated (im litt. to Spath) that they all came from beds above the Pholadomya beds. In fact Borges had searched the Pholadomya beds for ammonites for several years and had found none. Recent work on the stratigraphy of the area south-west of Benguela by Neto (1960: 89-99; 1961: 65-77) has confirmed the placing of these Dowvilleiceras in beds above the Pholadomya pleuromyaefornus beds, and in fact a formation character- ized by Nerinea (and some indeterminate ammonites) and reaching 400 metres in thickness comes between the Pholadomya beds below and the ‘“‘Acanthoceras”’ (i.e. Douvilleiceras) beds above. Therefore Choffat’s (1888: 20, 71) placing of one of his two specimens of Dowvilleiceras in the Pholadomya beds is probably incorrect (Mouta & O’Donnell 1933: 58-61). The age of all the Douvilleiceras seen so far is Mammillatum Zone, Lower Albian. This, not Upper Albian, is the age of the ““Acanthoceras’’ Formation of Neto (1960: 95; 1961: 69, 74). (b) Neokentroceras fauna of Praia do Jombo. The Jombo beach lies in the Benguela basin just south-west of the mouth of the Hanha (=Cubal) River and 16 km. north-east of Lobito. This is the same locality as that (‘“‘shore at landing place near Hanha’’) from which most of Spath’s original Neokentroceras came, and the specimens described by Haas (1942) came from a cliff on the bank of the same river near Hanha. Henrique O’Donnell, who collected the present specimens, said (zt litt. to Spath) that they came from beds in the lower part of the Upper Albian formation. This is the only direct evidence for their stratigraphical position; but combined with the morphological evidence that they are probably an end-form development of Hysteroceras, it is fairly certain that their age is low in the Upper Albian. This is the age of the lowest dateable part of the “Pervinquerta”’ Formation of Neto (1960: 95; 1961: 69, 75), and higher parts of this formation which contain the abundant fauna of Mortoniceras, Elobiceras, Puzosia and Hamitidae are higher in the Upper Albian. The Middle Albian is either repre- sented by a disconformity between the ““Acanthoceras” and “‘Pervinquieria” Forma- tions, or, more likely, is represented by beds in these formations that do not contain ammonites. The full fauna of Neokentroceras at Praia do Jombo consists of the following species; N. curvicornu Spath, N. singulare Haas, N. subtuberculatum Spath, N. trituberculatum sp. nov., N. pseudovaricosum Spath and N. crassicostatum sp. nov. (c) The Egito fauna, The Senonian outlier from which the Egito fauna came, is in a small bay 1 km. north of Egito. It is exposed for 1200 m. along the shore, but the variable width of the beds never exceeds 300 m. The formation consists of horizontal marly limestones CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 401 and coarse sandstones with unfossiliferous clays below, and is 40 m. thick. It rests unconformably on Upper Albian beds containing Stoliczkaia dispar in the upper part. According to O’Donnell (7 litt. to Spath) this outlier is the only one that contains ammonites amongst several similar ones (but with more sandstone and conglomerate) that occur between Egito and Lobito. The full cephalopod fauna from Egito is: Anagaudryceras nikobokense Collignon Gaudryceras varagurense Kossmat Didymoceras subtuberculatum sp. nov. Polyptychoceras pseudogaultianum (Yokoyama) Kitchimites angolaense sp. nov. Desmophyllites diphylloides (Forbes) Otophyllites angolaense Spath Eupachydiscus pseudogrossouvrei Collignon Hoplitoplacenticeras cf. marroti (Coquand) H. cf. costuloswm (Schliiter) H. spp. indet. Eutrephoceras simile Spath Tetragonites sp. indet. All the above species except the last one have been described in the systematic part of this paper. Tetrvagonites sp. indet. is represented by only one specimen that is poorly preserved and not specifically determinable. The presence of Hoplitoplacenti- ceras is sufficient to place the fauna in the zone characterized by this genus (Hoplito- placenticeras vari Zone) in the Upper Campanian. If the Zone of Bostrychoceras polyplocum is also put into the Upper Campanian (see discussion of Barra do Dande fauna below) rather than in the Maastrichtian, then the Egito fauna is referable to the lower half of the Upper Campanian. This does not clash with Spath’s (1951: 8; 1953: 49) dating of the fauna as the “‘very top of the Campanian’’, because Spath was following Haug’s classification where the Polyplocum Zone was placed as the basal zone of the Maastrichtian. The second view expressed by Spath in the same papers, that the Egito fauna “‘could equally well be considered to be basal Maastrichtian’’, reflected his suspicions that Hoplitoplacenticeras might not be confined to the zone that it is said to characterize (Spath 1953: 52) and also his desire to lower the base of the Maastrichtian still further so that it included the Vari Zone as well. But Spath’s fears that Hoplitoplacenticeras might occur outside the Vari Zone in Mada- gascar are not justified, for apart from the anomalous species H. lafresnayanum, the genus is a good zonal indicator (see p. 391 above). Wherever the Campanian— Maastrichtian boundary is placed, the Egito fauna can be definitely referred to the Vari Zone. No indication was given by O’Donnell that any part of the ammonite fauna was collected from any particular bed at Egito. The fauna can only be considered as a single unit, and none of the ammonites other than Hoplitoplacenticeras conflicts with this placing of the assemblage in a single zone—the Vari Zone of the Upper Cam- panian. Eupachydiscus pseudogrossouvrer occurs in the upper half of the Middle 402 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA Campanian in Madagascar, in the Zone of Delawarella subdelawarensis (Besairie & Collignon 1960: 77). This does not conflict with its presence in the Vari Zone in Angola, especially as Hoplitoplacenticeras is not common in Madagascar and its full range there might not yet be known. Kztchinites angolaense is not accurately dateable against any other species of the same genus, which are generally of Cam- panian or Lower Maastrichtian age. Desmophyllites diphylloides has a long range from the Lower Santonian to the Lower Maastrichtian, while Oiophyllites angolaensis can only be compared with O. decipiens which occurs in the Lower and ? Middle Campanian in Antarctica. Of the three lytoceratid species, Anagaudryceras mikobokense occurs in the Lower Maastrichtian in Madagascar and the Lower Maastrichtian or top of the Campanian in California, Gaudryceras varagurense occurs in the Santonian in India and has been recorded from the Santonian and Campanian of many other localities, and the Tetvagonites sp. indet. cannot be accurately dated. Such lytoceratids tend to be relatively long ranging and the presence of all three in the Upper Campanian does not clash with any previous records. The two hetero- morph ammonites in the Egito fauna cannot be used for accurate dating : Didymoceras subtuberculatum is a new species not clearly related to any other species of the genus that occur in the Campanian or Maastrichtian; Polyptychoceras pseudogaultianum occurs in the Santonian and Campanian in Japan, and a closely related species occurs in the Upper Campanian of British Columbia. (d) The Barra do Dande fauna The Barra do Dande ammonites collected by Henrique O’Donnell and Beeby Thompson consist of the following species: Neophylloceras ultimum Spath Baculites sp. indet. Didymoceras cf. hornbyense (Whiteaves) D. cf. angolaense (Haughton) Nostoceras hyatti Stephenson N. cf. kernense (Anderson) N. rotundum sp. nov. N. helicinum (Shumard) N. (2?) obtusum sp. nov. Solenoceras sp. indet. Polyptychoceras pseudogaultianum (Yokoyama) Manambolites dandensis sp. indet. Sphenodiscus sp. indet. All the above species, except Solenoceras and Baculites, have been described in the systematic part of this work. Solenoceras is represented by one very small fragment that has fine, slightly oblique ribs and two rows of small tubercles on the venter. It is not worth describing and may belong to either S. binodosa (Haughton 1925: 278) or S. bembense Haas (1943: II, figs. 4, 14) from Angola. There is one short indeterminate fragment of Baculites. Spath’s first list of the Barra do Dande ammonites (Spath 1951: 9, 10) corresponds CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 403 exactly with the list given above, but his later list (1953: 49, 50) included in addition all the Angolan Maastrichtian ammonites described by Haughton (1925) and Haas (1943) which came from other localities. Beeby Thompson, whose collection consisted only of six specimens of Nostoceras hyatti and one N. helicinum, gave no details of the beds at Barra do Dande. Henrique O’Donnell gave the following details (im litt.) to Spath: the Senonian at the mouth of the River Dande is about 60 m. thick, and consists of thick beds of more-or-less laminated marls, alternating with thin beds of limestone that are often fossiliferous and typically lenticular with hard crystalline centres; the limestones sometimes contain thin bituminous layers, and the whole series is characterized by the gigantic Imoceramus langi. O’ Donnell did not indicate that any of his ammonites came from any particular part of the series. To Beeby Thompson’s and O’Donnell’s collections must be added the two specimens of Nostoceras described by Sornay (1951) and the large Didymoceras described by Silva (1961). These three ammonites are definitely recorded as having come from bed 5 of the Barra do Dande section as described by Freneix (1959: 111-113). This is the best and most detailed description of the section, but further details were given by Darteville & Casier (1943: 85-86, fig. 46; 1959: 267-268). Bed 5 is a soft sandy limestone containing a band of black silica and is only 0-30 m. thick; it has been said to be of Campanian age because of the Nostoceras it contains. The overlying bed 6, also only 0-30 m. thick, is a gritty limestone containing plant debris, silicified gastropods, coprolites and fish teeth; the latter have been used to date it as Maastrichtian. Bed 7 consists of Recent deposits. The beds below bed 5 account for the remainder of the 60 metres of beds at Barra do Dande, and there is no indication that any of them contained the ammonites collected by O’Donnell. If all the specimens of Didymoceras and Nostoceras in O’Donnell’s collection are considered to come from bed 5, then the matrix of the specimens of Baculites, Solenoceras, Polyptychoceras and Manambolites agrees with them exactly—all are preserved in a hard white limestone, with varying amounts of iron-staining, and the septate whorls usually consist of recrystallized calcite. The very large specimen of Sphenodiscus is somewhat different, for there is no recrystallized calcite and no iron-staining, and it might have come from a different (? higher) bed. This Sphenodiscus undoubtedly indicates a Maastrichtian date, and probably Upper Maastrichtian. In Madagascar Sphenodiscus is known in one area, and occurs in the Upper Maastrichtian above beds with good Lower Maastrichtian ammonites (Besairie & Collignon 1960: 74, 79). In Europe the main Sphenodiscus fauna occurs in the Upper Maastrichtian, but one species, S. wbaghsit Grossouvre, also occurs in the upper half of the Lower Maastrichtian according to the zonal distribution table of Jeletzky (1951: 18-19). The occurrences of Sphenodiscus in the Middle East and India are not accurately dateable against other ammonites within the Maastrichtian. In North America the best stratigraphical sequence of species of Sphenodiscus is known in northern Mexico (Bose & Cavins 1927). The five zones in the Maastrichtian are based on Coahuilites and Sphenodiscus, and if the base of this succession corresponds to the base of the Maastrichtian, then at least the zone of Sphenodiscus 404 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA lenticularis (the second zone from the bottom) would come in the Lower Maastrich- tian. This is a large lenticular species with a highly complicated suture-line. The zones in Mexico and in the Gulf Coast of the United States have been discussed by Young (1960), but the question as to where the lower boundary of the Maastrichtian should be placed is not yet resolved. The other rich faunas of Sphenodiscus in the United States described and figured by Hyatt (1903) and by more recent workers (e.g. Reeside 1962: 136) are all Maastrichtian and some are Upper Maastrichtian, but the majority cannot be dated more accurately. The Angolan Sphenodiscus, therefore, probably indicates Upper Maastrichtian, but it could be as low as the upper half of the Lower Maastrichtian. The remainder of the Barra do Dande ammonite fauna belongs to either the Polyplocum Zone at the top of the Upper Campanian or to the lower half of the Neubergicus Zone at the base of the Lower Maastrichtian, and a decision as to which zone it belongs to does not seem possible in the present state of knowledge of ammonites from these zones. The most obvious correlation is with the ammonite fauna of the Nacatoch Sand of Texas described by Stephenson (1941) which contains two of the same species of Nostoceras, several Didymoceras and Solenoceras, and occurs below the horizons with Sphenodiscus in the Kemp Clay. Young (1960: 252, 256) is undecided as to whether the Nacatoch Sand is top Campanian or basal Maastrichtian, but on the whole favours the latter. The presence of Manambolites in the Angolan fauna appears to favour a top Campanian age, for no species of this genus can be proved to be Lower Maastrichtian (see above p. 395). It is tempting to make a comparison with “Manambolites” ricensis Young (1963: 127) which can be proved to come from the top of the Campanian in Texas, but Young’s species is so atypical of the genus as regards its suture-line, that it ought to be ignored for correlation purposes. The other Barra do Dande ammonites are useless for correla- tion; species of Neophylloceras are relatively long ranging, and Polyptychoceras pseudogaultianum, the only species common to both the Egito and Barra do Dande faunas, is said to range throughout the Santonian and Campanian in Japan. The evidence tends to favour the placing of all the Barra do Dande ammonites, except Sphenodiscus, in the Polyplocum Zone, Upper Campanian, rather than the Lower Maastrichtian, but the exact range of the various species of Didymoceras and Nostoceras has yet to be worked out. The position of the Campian-Maastrichtian boundary adopted here is between the Bostrychoceras polyplocum and Pachydiscus neubergicus Zones. This position, rather than at the base of the Polyplocum Zone, is more likely to be adopted by a majority of ammonite and micro-palaeontologists. The succession of ammonites and zones in the European Campanian and Maastrichtian has been discussed at length by Jeletzky (1951; 1958) who included the Polyplocum Zone in the Cam- panian on historical grounds. The lower position of the boundary adopted by Haug (1910), Spath (1953) and other workers is less satisfactory. Reiss (1962) favours the higher position of the boundary as used by Jeletzky, and has used it in establishing the Polyplocum Zone age of the phosphate deposits in Israel. Finally Young (1960; 1963: 19-20, 64) accepts this higher position for the boundary and has applied it to CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 405 his descriptions of the succession in the Gulf Coast of the United States. (e) The Carimba fauna The six ammonites from Carimba consist of the four specimens of Baculites subanceps, the single specimen of Didymoceras cf. californicum described above, and one fragment (C. 52728) of a very large indeterminable nostoceratid. The third fragment of a helically coiled ammonite listed by Spath (1951: 11) as possibly “Didymoceras hornbyense (Whiteaves) Haughton”’ is missing from the collection. These ammonites came from the Teba Formation, presumably from the upper part which is said to be rich in macrofossils (Mouta 1956: 43), and from which Haughton (1925: 264) obtained his fine fauna of Nostoceras, Didymoceras, Solenoceras, Baculites, Menuites and Libycoceras, and Haas (1943) his further examples of Nostoceras, Solenoceras and Axonoceras. The presence of Libycoceras is sufficient to establish the Upper Campanian age of at least part of the Teba Formation, for all accurately dated occurrences of this genus are in the Polyplocum Zone, Upper Campanian (Reiss 1962: 7-12). Another sub- species of the Angolan form Baculites subanceps subanceps occurs in the Upper Campanian of California (Matsumoto 1959a@: 130) and Japan (Matsumoto & Obata 1963: 59) (see p. 370 above), and the two are thought to be contemporaneous. Of the other Teba Formation ammonites, Menuites is relatively long-ranging (? San- tonian—Lower Maastrichtian) and the heteromorphs could be either Upper Campanian or basal Lower Maastrichtian in age. (f) The Benguela and San Nicolau faunas. Nothing can be added to the discussions and details of these faunas given by Spath (1951: 6, 9). The Benguela fauna consists of 16 crushed specimens that are not accurately determinable and not worth describing. The San Nicolau fauna consists of two specimens of the nautiloid listed by Spath and one indeterminate Baculites. V REFERENCES ADKINS, W.S. 1928. Handbook of Texas Cretaceous fossils. Bull. Univ. Tex. Bur. econ. Geol. Tech., Austin, 2838: 1-303, pls. 1-37. ArRACHI, C. 1931. Di alcuni Ammoniti della Creta superiore di Lobito (Angola). R.C. Ist. lombardo, Milan (2) 64: 845-852, pl. 2. ANDERSON, F. M. 1902. Cretaceous deposits of the Pacific coast. Proc. Calif. Acad. Sci., San Francisco (3) 2: 1-154, pls. 1-12. 1958. Upper Cretaceous of the Pacific coast. Mem. geol. Soc. Amer., New York, 71: 1-378, pls. I-75. — & Hanna, G. D. 1935. Cretaceous geology of Lower California. Pyoc. Calif. Acad. Sci., San Francisco (4) 23: 1-34, pls. I-11. ANDRADE, M. M. DE & ANDRADE, J. M. DE. 1958. Achega para una bibliografia geoldgica completa de Angola. Publ. Mus. Lab. min. geol. Porto (3) 74: 1-41. BaRRABE, L. 1929. Contribution a l'étude stratigraphique et pétrographique de la partie médiane du pays Sakalava. Mém. Soc. géol. Fr., Paris (n.s.) 12: 1-270, pls. I-10. Basse, E. 1931. Monographie paléontologique du Crétacé de la province de Maintirano. Mém. géol. Serv. Min. Madagascar, Tananarive, 1931: 1-86, pls. I-13. 406 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA Basse, E. 1939. Sur quelques Mollusques crétacés des Corbieres méridionales. Bull. Soc. géol. Fy., Paris (5) 9: 35-58, pl. 3. 1954. Sur la présence du genre Coahuilites Bose 1927 dans le sud Tunisien et du genre Eulophoceras Hyatt 1903 dans la Proche-Orient. Bull. Soc. géol. Fv., Paris (6) 3: 865-871, jo, 27 1963. Quelques ammonites nouvelles du Crétacé supérieur d’Angola. Bull. Soc. géol. Fr., Paris (7) 4: 871-876, pls. 12-14. BESAIRIE, H. 1930. Recherches géologiques a Madagascar. Contribution a l’étude des ressources minérales. Bull. Soc. Hist. nat. Toulouse, 60: 345-616, pls. 1-27. 1931. Les rapports du Crétacé Malgache avec le Crétacé de 1’ Afrique australe. Bull. Soc. géol. Fr., Paris (4) 30: 613-643, pls. 64—67. 1930. Recherches géologiques a Madagascar. La géologie du Nord-Ouest. Mém. Acad. Malgache, Tananarive, 21: 9-259, pls. 1-24. Besairiz, H. & CoLLricNon, M. 1956. Lewvique stvatigyvaphique international 4, 2 (Madagascar). 93 pp. Paris. 1960. Lexique stratigvaphique international 4, 2 (Madagascar, supplément). 190 pp. Paris. Biyxuorst, J. T. 1861. Monographie des gastropodes et des céphalopodes de la craie supérieure du Limbourg. vi + 83 pp. (gastropods) + 44 pp. (cephalopods), 14 pls. Brussels. Brainvit_eE, H. M. D. DE. 1825-1827. Manuel de malacologie et de conchyliologie. 664 pp. (1825), 87 pls. (1827). Paris. BLANCKENHORN, M. 1900. Neues zur Geologie und Palaontologie Aegyptens. Z. dtsch. geol. Ges., Berlin, 50: 21-47. Boum, J. 1891. Die Kreidebildungen des Fiirbergs und Salzbergs bei Siesdorf in Ober-Bayern. Palaeontographica, Stuttgart, 38: 1-106, pls. 1-5. BoéseE, E. 1927. Cretaceous ammonites from Texas and northern Mexico. Bull. Univ. Tex. Bur. econ. Geol. Tech., Austin, 2748: 143-357, pls. 1-19. Bose, E. & Cavins, O. A. 1927. The Cretaceous and Tertiary of southern Texas and northern Mexico. Bull. Univ. Tex. Bur. econ. Geol. Tech., Austin, 2748: 7-142. Bouts, M., Lemoine, P. & THEVENIN, A. 1906-1907. Paléontologie de Madagascar. III— Céphalopodes crétacés des environs de Diego-Suarez. Ann. Paléont., Paris, 1: 173-192, pls. 14—20 (1906); 2: 1-56, pls. 1-8 (1907). BourGueEt, L. 1742. Traité des petrifications. xvi + 163 + 91 pp., 60 pls. Paris. BroGnon, G. & VERRIER, G. 1958. Note sur la stratigraphie du bassin du Cuanza en Angola. Bol. Soc. geol. Portug., Porto, 12, 3: 61-74. Bronn, H. G. 1834-1838. Lethea Geognostica. 1346 pp., 47 pls. Stuttgart. CARVALHO, G. S. DE. 1960. Sobre os depésitos cretacicos do litoral de Angola. Bol. Serv. geol. Min. Angola, Luanda, 1: 37-48. 1961. Geologia do deserto de Mocamedes. Mem. Jta. Invest. Ultvamar, Lisbon, 26: 1-227. CasEy, R. 1962. A monograph of the Ammonoidea of the Lower Greensand, Part 4: 217-288, pls. 36-42. Palaeontogr. Soc. (Monogy.), London. CHAVAN, A. 1947. La faune campanienne de Mont des Oliviers d’apres les matériaux Vignal- Massé. J. Conchyliol., Paris, 87: 125-197, pls. 2-4. CuHorrat, P. 1888. Matériaux pour l’étude stratigraphique et paléontologique de la province d’Angola. Mém. Soc. Phys. Genéve, 30: 1-116, pls. 1-8. 1898. Recueil d’etudes paléontologiques sur la faune crétacique du Portugal. 1, Especes nouvelles ou peu connues. Part 2, Les ammonées du Bellasien, des Couches a Neolobites vibyayeanus, du Turonien et du Sénonien. Tvav. géol. Portugal, Lisbon, 1898: 41-86, pls. 3—22. 1905. Nouvelles données sur la zone littorale d’Angola. Comm. Sevv. géol. Portug., Lisbon, 1905: 31-78, pls. 1-4. CoBBAN, W. A. 1963. Occurrence of the late Cretaceous ammonite Hoplitoplacenticervas in Wyoming. Prof. Pap. U.S. geol. Suvv., Washington, 475-C: 60-62. CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 407 COLLIGNON, M. 1931. Faunes sénoniennes du nord et de l’ouest de Madagascar. Ann. géol. Serv. Min. Madagascar, Tananarive, 1: 1-66, pls. 1-9. 1932. Les ammonites pyriteuses de |’Albien supérieur du Mont Raynaud a Madagascar. Ann. géol. Serv. Min. Madagascar, Tananarive, 2: 1-36, pls. 1-4. 1938. Ammonites Campaniennes et Maastrichtiennes de l’ouest et du sud de Madagascar. Ann. géol. Serv. Min. Madagascar, Tananarive, 9: 55-118, pls. 1-9. 1949. Recherches sur les faunes albiennes de Madagascar. 1. L’Albien d’Ambarimaninga. Ann. géol. Serv. Min. Madagascar, Tananarive, 16: 1-128, pls. 1-22. 1950. Recherches sur les faunes albiennes de Madagascar. 3. L’Albien de Komihevitra. Ann. géol. Serv. Min. Madagascar, Tananarive, 17: 21-54, pls. 3-9. 1955. Ammonites néocrétacées du Menabe (Madagascar). II. Les Pachydiscidae. Ann. géol. Serv. Min. Madagascar, Tananarive, 21: 1-08, pls. 1-28. 1955a. Ammonites néocrétacées du Menabe (Madagascar). III. Les Kossmaticeratidae. Ann. géol. Serv. Min. Madagascar, Tananarive, 22: 1-54, pls. 1-12. 1956. Ammonites néocrétacées du Menabe (Madagascar). IV. Les Phylloceratidae; V. Les Gaudryceratidae; VI. Les Tetragonitidae. Ann. géol. Serv. Min. Madagascar, Tananarive, 23: 1-106, pls. I-11. 1961. Ammonites néocrétacées du Menabe (Madagascar). VII. Les Desmoceratidae. Ann. géol. Serv. Min. Madagascar, Tananarive, 31: 1-115, pls. 1-32. 1963. Atlas des fossiles cavacteristiques de Madagascay (Ammonites). X Albien. xv-+- 184 pp., pls. 241-317. Service géologique, Tananarive. ConrAD, J. A. 1868. Synopsis of the invertebrate fossils of the Cretaceous Formation of New Jersey. In Cook, J. H. 1868. Geology of New Jersey. Geol. Surv. New Jersey, Newark, Appendix A: 721-732. Coguanp, M. H. 1859. Synopsis des animaux et des végétaux fossiles observés dans le formation crétacée du sud-ouest de la France. Bull. Soc. géol. Fv., Paris (2) 16: 945-1023. DARTEVELLE, E. 1952. Echinides fossiles du Congo et de l’Angola. 1. Introduction historique et stratigraphique. Ann. Mus. Congo belge 8v0, Tervuren, 12: 1-71. 1953. Echinides fossiles du Congo et de l’Angola. 2. Description systematique des Echinides fossiles du Congo et de l’Angola. Ann. Mus. Congo belge 8vo, Tervuren, 13: I-240, pls. I-19. DARTEVELLE, E. & CasiER, E. 1943. Les Poissons fossiles du Bas-Congo et des régions voisines, 1. Ann. Mus. Congo belge 4to, Tervuren (A, 3) 2, 1: 1-200, pls. 1-16. 1959. Les Poissons fossiles du Bas-Congo et des régions voisines, 3. Ann. Mus. Congo belge 4to, Tervuren (A, 3) 2, 3: 257-568, pls. 23-39. DEFRANCE, M. J. L. 1816. Im Dictionnaire des Sciences naturelles, dans lequel on traite méthodiquement des differens Etres de la Nature . . . 1816-1830. 60 vols. text, 12 vols. plates, Paris. Vol. 3 (1816), 492 pp., +174 pp. in supplement. Plates—Zoologie, Conchyliologie et Malacologie, by H. M. D. de Blainville, 1816-1830: 36 pp., 118 pls. DesHayeEs, M. G. P. 1831. Description de coquilles cavactévistiques des terrains. vii + 264 pp., 14 pls. Paris. Desmarest, A. G. 1817. Mémoire sur deux genres de coquilles fossiles cloisonnées et a siphon. J. Phys. Chim. Hist. nat., Paris, 85: 42-51, pls. 1, 2. DIENER, C. 1925. Ammonoidea neocretacea. Fossilium Catalogus, 1: Animalia. Pars. 29. 244 pp. Berlin. DovuvitLE, H. 1931. Contribution a la géologie de l’Angola. Les ammonites de Salinas. Bol. Mus. min. géol. Univ. Lisboa, 1: 17—46, pls. 1-4. FaujJAsS DE SAIntT-Fonp, B. (1799). WHistoive naturelle de la montagne de Saint-Pierre de Maestricht. 263 pp., 53 pls. Paris. (Date of publication uncertain.) ForBeEs, E. 1846. Report on the Cretaceous fossil invertebrates from southern India, collected by Mr. Kaye and Mr. Cunliffe. Tvans. Geol. Soc. London (2) 7: 97-174, pls. 7-19. Favre, E. 1869. Description des Mollusques fossiles de la Craie des environs de Lemberg en Galicie. xii + 187 pp., 13 pls. Geneve. 408 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA FRENEIX, S. 1959. Mollusques fossiles du Crétacé de la Cote occidentale d’Afrique, du Cameroun alAngola. Ann. Mus. Congo belge 8vo, Tervuren, 24: i-xv, 1-126. GaBB, W. M. 1864. Geological Survey of California. Palaeontology, 1: 57-217, pls. 9-32. Philadelphia. GREGORY, J. W. 1916. Contribution to the geology of Benguella. Tvans. Roy. Soc. Edinb., 51: 495-536, pls. I, 2. 1922. Supplementary note on the geology of Benguella in relation to its Cephalopods and the history of the south Atlantic. Tvans. Roy. Soc. Edinb., 53: 161-163. GRIEPENKERL, O. 1889. Die Versteinerungen der senonen Kreide von K6nigslutter im Herzogthum Braunschweig. Paldont. Abh., Jena, 4: 305-4109, pls. 34-45. GRossoUVRE, A. DE. 1894. Recherches sur la Craie supérieure. Part 2, Paléontologie, les ammonites de la Craie supérieure. Mém. Carte géol. dét. Fy., Paris. 264 pp., 39 pls. 1901. Recherches sur la Craie supérieure. Part 1, Stratigraphie générale, avec une mono- graphie du genre Micrastey par J. Lambert. Mém. Carte géol. dét. Fr., Paris. 1013 pp. Haas, O. 1941. A case of inversion of suture-lines in Hystevoceras varicosum (Sow.). Amer. J. Sci., New Haven, 239: 661-664, pl. 1. 1942. The Vernay collection of Cretaceous (Albian) ammonites from Angola. Bull. Amer. Mus. Nat. Hist., New York, 81: 1-224, pls. 1-47. 1942a. Some Upper Cretaceous ammonites from Angola. Amer. Mus. Novit., New York, 1182: 1-24. 1943. Some abnormally coiled ammonites from the Upper Cretaceous of Angola. Amer. Mus. Novit., New York, 1222: 1-18. 1945. A recently acquired Albian ammonite from Angola. Amer. Mus. Novit., New York, 1286: 1-4. 1952. Some Albian desmoceratid and lytoceratid ammonites from Angola. Amer. Mus. Novit., New York, 1561: 1-17. HAuveER, F. von. 1866. Neue Cephalopoden aus den Gosaugebilden der Alpen. S.B. Akad. Wiss. Wien, 53: 300-308, pls. I, 2. Haua, E. 1910. Traité de Géologie, 2. Les périodes géologiques, 2 (Systeme jurassique et Période crétacée) : 929-1396, pls. t00-119. Paris. HaucGuton, S. H. 1925. Notes on some Cretaceous fossils from Angola (Cephalopoda and Echinoidea). Ann. S. Afy. Mus., Cape Town, 22: 263-288, pls. 12-15. 1959. Geological bibliography of Africa south of the Sahara. Bibliogvaphy of the Jurassic and Cretaceous Systems. Commission de co-operation technique en Afrique au sud du Sahara. iui + 83 pp. London. 1963. The stratigraphic history of Africa south of the Sahara. xii + 365 pp. Edinburgh & London. HISINGER, W. 1837. Lethaea Svecica, seu Petrificata Sveciae, iconibus et chavacteribus illustrata. 124 pp., 36 pls. Holmiae. HoeEpeEn, E. C. N. vAN. 1921. Cretaceous Cephalopoda from Pondoland. Ann. Tvansv. Mus., Pretoria, 8: 1-48, pls. 1-11. HoppeneEr, H. 1958. Brief report on the palaeontology of the Cuanza basin—Angola. Bol. Soc. geol. Portug., Porto, 12, 3: 75-82. Hourco, V. 1949. Paléontologie de Madagascar. XXVIII. Sur quelques ammonites du Sénonien. Ann. Paléont., Paris, 35: 89-117, pls. 11-13. 1950. Les terrains sédimentaires de la région de Morondava. Amn. géol. Serv. Min. Madagascar, Paris, 20: 1-109. Hyatt, A. 1894. Phylogeny of an acquired characteristic. Pyvoc. Amery. Phil. Soc., Philadelphia, 32: 349-647, pls. I-14. —— 1900. Cephalopoda. In Zittel, K. A. von. 1896-1900. Textbook of Palaeontology, trans- lated by C. R. Eastman. Pp. 502-604. London. —— 1903. Pseudoceratites of the Cretaceous. Monogr. U.S. Geol. Surv., Washington, 44: I—351, pls. 1-47. CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 409 IMKELLER, H. t1gor. Die Kreidebildungen und ihre Fauna am Stallauer Eck und Enzenauer Kopf bei T6lz. Palaeontogrvaphica, Stuttgart, 48: 1-64, pls. 1-3. JeLetzxy, J. A. 1951. Die Stratigraphie und Belemnitenfauna des Obercampan und Maastricht Westfalens, Nordwestdeutschlands und Danmarks sowie einige allgemeine Gliederungs- Probleme der jiingeren borealen Oberkreide Eurasiens. Beith. geol. Jb., Hannover, 1: I-142, pls. 1-7. 1958. Die jiingere Oberkreide (Oberconiac bis Maastricht) Stidwestrusslands und ihr Vergleich mit der Nordwest- und Westeuropas. Beth. geol. Jb., Hannover, 33: 1-157. Jimpo, K. 1894. Beitrage zur Kenntniss der Fauna der Kreideformation yon Hokkaido. Paldont. Abh., Jena, 6: 149-194, pls. 1-9. Jones, D. L. 1963. Upper Cretaceous (Campanian and Maastrichtian) ammonites from southern Alaska. Prof. Pap. U.S. Geol. Surv., Washington, 432: 1-53, pls. 1-41. Kiran, W. & REBOUL, P. 1909. Des céphalopodes néocrétacés des iles Seymour et Snow Hill. Wiss. Evgebn. schwed. Sudpolavexped., Stockholm, 3, 6: 1-75, pls. I-20. Kossmat, F. 1895-1898. Untersuchungen tiber die Siidindische Kreideformation. Beitr. Paldont. Geol. Ost.-Ung., Wien & Leipzig, 9 (1895): 97-203, pls. 15-25; 11 (1897): 1-46, pls. 1-8; 11 (1898): 89-152, pls. 14-109. Lamarck, J. B. P. A. DE M. DE. 1801. Systéme des Animaux sans vertébres. viii + 432 pp. Paris. 1822. Histoive naturelle des Animaux sans vertébres. 7. 711 pp. Paris. MARSHALL, P. 1926. The Upper Cretaceous ammonites of New Zealand. Tvans. N.Z. Inst., Wellington, 56: 129-210, pls. 19-47. Matsumoto, T. 1954. Family Puzosiidae from Hokkaido and Saghalien. Mem. Fac. Sci. Kyushu Univ. (D) 5: 69-118, pls. 9-23. 1954a. Selected Cretaceous leading ammonites in Hokkaido and Saghalien. In Matsumoto, T. (editor). 1954. The Cretaceous System in the Japanese Islands. Appendix: 243-313, pls. 17-36. Tokyo. 1959. Cretaceous ammonites from the Upper Chitina Valley, Alaska. Mem. Fac. Sct. Kyushu Univ. (D) 8: 49-90, pls. 12-29. 1959a. Upper Cretaceous ammonites from California. Part 1. Mem. Fac. Sci. Kyushu Univ. (D) 8: 91-171, pls. 30-45. 1959). Upper Cretaceous ammonites of California. Part 2. Mem. Fac. Sci. Kyushu Univ. (D) Special vol. 1: 1-172, pls. 1-41. 1959c. Zonation of the Upper Cretaceous in Japan. Mem. Fac. Sci. Kyushu Univ. (D) 9: 55-93, pls. 6-11. 1960. Upper Cretaceous ammonites of California. Part 3. Mem. Fac. Sci. Kyushu Univ. (D) Special vol. 2: 1-204, pls. 1, 2. Matsumoto, T. & OBata, I. 1955. Some Upper Cretaceous desmoceratids from Hokkaido and Saghalien. Mem. Fac. Sci. Kyushu Univ. (D) 5: 119-151, pls. 24-30. 1963. A monograph of the Baculitidae from Japan. Mem. Fac. Sci. Kyushu Univ. (D) 13: I-116, pls. I-27. MEEK, F. B. 1871. Preliminary palaeontological report, consisting of lists of fossils, with descriptions of some new types. Ann. Rep. U.S. Geol. Surv. Terr., Washington, 4: 287-318. 1876. A report on the Invertebrate Cretaceous and Tertiary fossils of the Upper Missouri country. Rep. U.S. Geol. Surv. Tery., Washington, 9: i-lxiv, 1-629, pls. 1-45. 1876a. Descriptions and illustrations of fossils from Vancouver and Sucia Islands and other northwestern localities. Bull. U.S. Geol. Geogr. Surv. Tery., Washington, 2: 351-374, pls. 2-6. MEEK, F. B. & HaypeEn, M. D. 1856. Descriptions of new species of Gasteropoda and Cephalo- poda from the Cretaceous formation of Nebraska Territory. Proc. Acad. Nat. Sci. Philad., 8: 70-72. Mitier, A. K. & CARPENTER, L. B. 1956. Cretaceous and Tertiary nautiloids from Angola. Estud. Ens. Docum. Junta Invest. Ultvamar, Lisbon, 21: 1-48, pls. 1-6. 410 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA Morton, S. G. 1841. Description of some new species of organic remains of the Cretaceous group of the United States: with a tabular view of the fossils hitherto discovered in this formation. J. Acad. Nat. Sci. Philad., 8: 207-227, pls. Io, 11. Mouta, F. 1938. Notice géologique sur l’Angola. Comun. Serv. geol. Portug., Lisbon, 20: 19-37. 1954. Noticia explicativa do esboco geologico de Angola. 148 pp., 13 pls. Lisbon. 1956. Lewxique stratigvaphique internationale. 4, Afrique, part 7b, Angola. 54 pp. Paris. Moura, F. & O’DoNNELL, H. 1933. Carte géologique de l’Angola. 87 pp., 12 pls. Lisbon. Neto, M. G. M. 1960. Géologie de la région Benguela-Cuio. Bol. Serv. geol. min. Angola, Luanda, 1: 89-99. 1961. As bacias sedimentares de Benguela e Mocamedes. Bol. Serv. geol. min. Angola, Luanda, 3: 63-93. NoeEtLinG, F. 1897. Fauna of the Upper Cretaceous (Maestrichtian) beds of the Mari Hills. Palaeont. indica, Calcutta (16) 1, 3: 1-79, pls. I-23. Nowak, J. 1908. Untersuchungen tiber Cephalopoden der oberen Kreide in Polen. 1 Teil. Genus Baculites Lamarck. Bull. int. Acad. Cracovie (Acad. pol. Sci.), 1908: 326-353, pl. 14. Orsson, A. A. 1944. Contributions to the palaeontology of northern Peru. The Cretaceous of the Paita region. Bull. Amer. Paleont., Ithaca, 28: 1-164, pls. 1-17. OrpiGNny, A. bd’. 1840-1842. Paléontologie francaise. Terrains crétacés. 1, Céphalopodes: I-120 (1840), 121-430 (1841), 431-662 (1842), 148 pls. Paris. PAULCKE, W. 1906. Die Cephalopoden der oberen Kreide Siidpatagoniens. Ber. naturf. Ges. Freiburg i. B., 15: 167-248, pls. 10-19. PRESWICH, J. 1888. Geology. 2, Stratigraphical and physical. xxviii + 606 pp., 16 pls. Oxford. Picarp, L. 1929. On Upper Cretaceous (chiefly Maestrichtian) Ammonoidea from Palestine. Ann. Mag. Nat. Hist., London (10) 3: 433-456, pls. 9, Io. Reiss, Z. 1962. Stratigraphy of the phosphate deposits of Israel. Bull. geol. Sevv. Israel, Jerusalem, 34: 1-23. REDTENBACHER, A. 1873. Die Cephalopoden fauna der Gosauschichten in den nordéstlichen Alpen. Abh. geol. Reichsanst. (Bundesanst.), Wien, 5: 91-140, pls. 22-30. REESIDE, J. B. 1962. Cretaceous ammonites of New Jersey. Im Richards, H. J. e¢ al., 1958— 1962. The Cretaceous fossils of New Jersey. 2: 113-137, pls. 68-75. Geol. Surv., New Jersey, Trenton. RENNIE, J. V. L. 1929. Cretaceous fossils from Angola (Lamellibranchia and Gasteropoda). Ann. S. Afy. Mus., Cape Town, 28: 1-54, pls. 1-5. —— 1945. Lamelibranquios e Gastropodos do Cretacico Superior de Angola. Mem. Jta. Miss. geogr. colon., Lisbon (Geol.) 1: 1-141, pls. 1-4. ReEyMENT, R. A. 1955. The Cretaceous Ammonoidea of southern Nigeria and the southern Cameroons. Bull. Geol. Surv. Nigeria, Kaduna, 25: 1-112, pls. 1-25. R6m_Er, F. A. 1840-1841. Die Versteinerungen des Norddeutschen Kvreidegebirges: 1-48, pls. 1-7 (1840), 49-145, pls. 8-16 (1841). Hannover. QuENstTEDT, F. A. 1845-1849. Petvefactenkunde Deutschlands. Die Cephalopoden: 1-104, pls. 1-6 (1845), 105-184, pls. 7-14 (1846), 185-264, pls. 15-19 (1847), 265-472, pls. 20-29 (1848), 472-580, pls. 30-36 (1849). Tiibingen. SCHLUTER, C. 1867. Beitrag zuy Kenntnis der jiingsten Ammoneen Norddeutschlands. 36 pp., 6 pls. Bonn. —— 1871-1876. Die Cephalopoden der oberen deutschen Kreide. Palaeontographica, Stuttgart, 21: 1-24, pls. 1-8 (1871) ; 21: 25-120, pls. 9-35 (1872) ; 24: 1-144 (121-264), pls. 36-55 (1876). —— 1872a. Probeabdriicke seines Cephalopoden-werkes tiber die obere deutsche Kreide. Verh. naturh. Ver. preuss. Rheinl., Bonn, 29, Correspondenzblatt 1: 91-92. Scumipt, M. F. 1873. Uber die Petrefakten der Kreideformation von der Insel Sachalin. Mém. Acad. Sci. St. Pétersb. (7) 19, 3: 1-37, pls. 1-8. Suimizu, S. 1934. Ammonites. Im Shimizu, S. & Obata, T. 1934. Cephalopoda. Iwanami’s lecture series of Geology and Palaeontology. 137 pp. Tokyo. (In Japanese.) CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA 411 SHIMIZU, S. 1935. The Upper Cretaceous cephalopods of Japan. Part 1. J. Shanghai Sci. Inst. (2) 2: 159-226. SHUMARD, B. F. 1861. Descriptions of new Cretaceous fossils from Texas. Pyvoc. Boston Soc. Nat. Hist., 8: 188-205. Sttva, G. H. DA. 1961. Ammonite nouvelle du Campanien de la Barra do Dande (Angola). Mem. Mus. min. geol. Univ. Coimbra, 51: 19-24, pls. 1-3. SmitH, J. P. 1898. Development of Lytoceras and Phylloceras. Proc. Calif. Acad. Sci., San Francisco (3) 1, 4: 129-160, pls. 16-20. Soares, A. F. 1959. Contribuicao para o estudo da fauna fossil da regiao de Entre Lobito e Catumbela (Angola). Garcia de Orta, Lisbon, 7: 135-154, pls. I-5. SORNAY, J. 1951. Ammonites albiennes et sénoniennes de |’Angola et de |’Afrique égatoriale frangaise. Rev. Zool. Bot. afr., Brussels, 44: 271-277, pls. 2-4. 1953. Ammonites nouvelles de l’Albien de l’Angola. Rev. Zool. Bot. afy., Brussels, 47: 52-59, pls. 1-3. SoRNAY, J. & TESSIER, F. 1949. Ammonites nouvelles du Maestrichtian du Sénégal. Bull. Soc. géol. Fr., Paris (5) 19: 245-249, pl. 7. SOWERBY, J. DEC. 1823-1846. The Mineral Conchology of Great Britain. 4 (pars)—7, pls. 384— 648. London. SpaTH, L. F. 1921. On Upper Cretaceous Ammonoidea from Pondoland. Ann. Durban Mus., 3: 39-57, pls. 6, 7. 1g21a. On Cretaceous Cephalopoda from Zululand. Ann. S. Afr. Mus., Cape Town, 12: 217-321, pls. 19-20. 1922. On Cretaceous Ammonoidea from Angola, collected by Prof. J. W. Gregory, D.Sc., F.R.S. Trans. Roy. Soc. Edinb., 53: 91-160, pls. 1-4. 1922a. On the Senonian ammonite fauna of Pondoland. Trans. Roy. Soc. S. Afr., Cape Town, 10: 113-147, pls. 5-9. 1923-1943. A monograph of the Ammonoidea of the Gault. Palaeontogr. Soc., Monogyr., London: 787 pp., 72 pls. 1932a. Review of “H. Douvillé, 1931, Les Ammonites de Salinas’. Geol. Zbl., Leipzig, rae 1940a. On Upper Cretaceous (Maestrichtian) Ammonoidea from Western Australia. J. Roy. Soc. W. Aust., Perth, 26: 41-57, pls. I, 2. 1951. Preliminary notice on some Upper Cretaceous ammonite faunas from Angola. Comun. Serv. geol. Portug., Lisbon, 32: 123-130. 1953. The Upper Cretaceous Cephalopod fauna of Grahamland. Scientific Reports, Falkland Islands Dependencies Survey, 3: 1-60, pls. I-13. STEINMANN, G. 1895. Die Cephalopoden der Quiriquina-Schichten. Newes Jb. Min., Stuttgart, 10: 64-94, pls. 4-6. STEPHENSON, L. W. 1941. The larger invertebrate fossils of the Navarro Group of Texas. Bull. Univ. Tex. Bur. econ. geol. Tech., Austin, 4101: 1-641, pls. 1-95. STOLICczKA, F. 1863-1866. The fossil Cephalopoda of the Cretaceous rocks of southern India. Palaeont. indica, Calcutta (3) 1: 41-56, pls. 26-31 (1863); 2-5: 57-106, pls. 32-54 (1864); 6-9: 107-154, pls. 55-80 (1865); 10-13: 155-216, pls. 81-94 (1866). TAvaANI, G. 1942. Paleontologia della Somalia. III, Fossili del Cretaceo. 6, Molluschi del Cretaceo dell Somalia. Palaeontogr. ital., Pisa, 32, 4, 2: 7-47, pls. 1-5. 1949. Fauna Malacologia Cretacea della Somalia e Dell’Ogaden. Part 2. Palaeontogr. italica, Pisa, 45: 1-76, pls. 1-11. THIELE, S. 1933. Neue Fossilfunde aus der Kreide von Angola mit einem Betrag zur Stam- mesgeschichte der Gattung Pervinquieria Bohm. Zbl. Min. Geol. Paldont., Stuttgart, 1933, B: 110-123. UsHER, J. L. 1952. Ammonite faunas of the Upper Cretaceous of Vancouver Island, British Columbia. Bull. Geol. Surv. Can., Ottawa, 21: 1-182, pls. I-30. 412 CRETACEOUS AMMONITES AND NAUTILOIDS FROM ANGOLA WELLER, S. 1907. A report on the Cretaceous palaeontology of New Jersey. Geol. Surv. New Jersey, Trenton (Paleont.) 4: ix + 871, 111 pls. WHITEAVES, J. F. 1876. On some invertebrates from the coal-bearing rocks of the Queen Charlotte Islands. Geol. Surv. Can., Montreal, Mesozoic Fossils, 1, 1: 1-92, pls. I-10. 1895. Notes on some fossils from the Cretaceous rocks of British Columbia. Canad. Rec. Sci., Montreal, 6: 313-318. 1903. On some additional fossils from the Vancouver Cretaceous, with a revised list of species therefrom. Geol. Surv. Can., Montreal, Mesozoic Fossils 1, 5: 309-409, pls. 40-51. Waite, C. A. 1887. Contribuicoes a Palaeontologia do Brazil. Avch. Mus. nac., Rio de J., 7: 1-273, pls. 1-28. WHITFIELD, R. P. 1880. Paleontology of the Black Hills. Jn Newton, H. & Jenney, E. M. 1880. Report on the Geology and resources of the Black Hills of Dakota. U.S. Geol. Surv., Washington: 325-468, pls. 1-16. 1892. Gasteropoda and Cephalopoda of the Raritan Clays and Greensand Marls of New Jersey. Mon. U.S. geol. Surv., Washington, 18: 1-402, pls. 1-50. 1g01. Note ona very fine example of Helicoceras stevenson preserving the outer chamber. Bull. Amer. Mus. Nat. Hist., New York, 14: 2109, pls. 29, 30. 1902. Observations on and emended description of Heterocervas simplicostatum Whitfield. Bull. Amer. Mus. Nat. Hist., New York, 16: 67-72, pls. 23-27. WIEDMANN, J. 1962. Ammoniten aus der Vascogotischen Kreide (Nordspanien). 1. Phyllo- ceratina, Lytoceratina. Palaeontographica, Stuttgart, 118, A: 119-237, pls. 8-14. Woops, H. 1906. The Cretaceous fauna of Pondoland. Aun. S. Afy. Mus., Cape Town, 4: 275-350, Pls. 33-44. Wriacut, C. W. 1957. In Treatise on Invertebrate Paleontology. Part L, Mollusca 4. Cephalopoda, Ammonoidea (Ed. by R. C. Moore). xxii + 490 pp. Kansas & New York. Wricut, C. W. & Matsumoto, T. 1954. Some doubtful Cretaceous ammonite genera from Japan and Saghalien. Mem. Fac. Sci. Kyushu Univ. (D) 4: 107-134, pls. 7, 8. YaBE, H. 1903. Cretaceous Cephalopoda from Hokkaido. Part 1. J. Coll. Sci. Tokyo, 18, 2: I-55, pls. 1-7. 1904. Cretaceous Cephalopoda from Hokkaido. Part 2. J. Coll. Sci. Tokyo, 20, 2: 1-45, pls. 1-6. Yoxoyama, M. 1890. Versteinerung aus der japanische Kreide. Palaeontographica, Stuttgart, 36: 159-202, pls. 18-25. Younc, K. 1960. Later Cretaceous ammonite successions of the Gulf Coast of the United States. xxi Int. Geol. Congr. Copenhagen, 21: 251-260. 1963. Upper Cretaceous ammonites of the Gulf Coast of the United States. Bull. Univ. Tex. Bur. econ. Geol. Tech., Austin, 6304: 1-373, pls. 1-82. \ PLATE 1 Fics. 1-4. Douvilleiceras mammillatum (Schlotheim) ? var. aequinodum (Quenstedt). “Acanthocevas’’ Formation, Dombe Grande, Angola. Servi¢cos de Geologia e Minas, Luanda, nos. D.G. 306, 308, 309 and 294 respectively. Fic. 5. Douvilleicevas orbignyt Hyatt. Same formation and locality. D.G. 305. All figures natural size. PLATE 1 Bull. B. M. (N. H.) Geol. 10, 10 PLATE 2 Fics. 1-9. Neokentroceras curvicorvnu Spath. Fig. 1, holotype, shore at landing place near Hanha, Angola; B.M. C. 20116. Figs. 2-7, Praia do Jombo; C. 52556, C. 52553, C. 52573, C. 52560, C. 52558, C. 52552 respectively. Fig. 8, paratype, shore at landing place near Hanha; C. 20123. Fig. 9, Praia do Jombo; C. 52554. Fics. 10-15. Neokentrocevas singulave Haas. Praia do Jombo, Angola. C. 52583, C. 52586, C. 52585, C. 52574, C. 52579, C. 52597 respectively. Fic. 16. Neokentrocervas crassicostatum sp. nov. Paratype, Praia do Jombo, Angola. C. 52600. Figs. I, 6, 7, 9, 11, 12-16 natural size. Figs. 2-5, 8, 10 — X1°5 Bull. B. M. (N.H.) Geol. 10, 10 PAE 2 10a 10b 4a 14b PLATE 3 Fic. 1. Neokentroceras subtuberculatum Spath. Holotype, near Benguela, Angola. C. 20042. Fies. 2-4. Neokentroceras trituberculatum sp. nov. Near Catumbella, Benguela, Angola. Fig. 2, holotype, C. 20285. Figs. 3, 4, paratypes, C. 14819 and C. 20284. Fies. 5-11. Neokentrocevas pseudovaricosum Spath. Fig. 5, holotype, shore at landing place near Hanha, C. 20125. Figs. 6-8, paratypes, same locality, C. 20120, C. 20122, C. 20124. Figs. 9-11, Praia do Jombo, C. 52590, C. 52591, C. 52592. Fics. 12-15. Neokentroceras crassicostatum sp. noy. Praia do Jombo. Fig. 12, holotype, B.M. C. 52593. Figs. 13-15, paratypes, B.M. C. 52599, C. 52596, C. 52508. Figs. I, 2, 4, 9, 11, 12, 14 — natural size. Figs. 3, 5-8, 10, 13, I5 — X1I°5. Bull. B. M. (N. H.) Geol. 10, 10 PLATE 3 PLATE 4 Fics. 1-3. Anagaudrycevas mikobokense Collignon. «1 km. north of Egito, Angola. C. 52637, C. 52643 and C. 52641 respectively. Fic. 4. Baculites anceps Lamarck. Side and ventral views. Calcaire a Baculites, Manche, France, C. 382. Fic. 5. Gaudrycevas vavagurense Kossmat. 1 km. north of Egito, Angola. C. 52657. All figures natural size. PLATE 4 Bull. B. M. (N. H.) Geol. 10, 10 PLATE 5 Fics. 1, 2. Gaudryceras vavagurense Kossmat. 1 km. north of Egito, Angola. C. 52658 and C. 52656. Fic. 3. Baculites subanceps Haughton. Carimba, Angola. Side, dorsal and ventral views. C. 52730. Fics. 4, 5. Baculites anceps Lamarck. Calcaire a Baculites, Manche, France. Fig. 4, Valognes, Manche, C. 70597. Fig. 5, neotype, side and ventral views, ‘““Normandy’’, Mantell Collection, B.M. (N.H.). 32573 x7/8. All figures natural size, except fig. 5, x0-9 approx. PLATE 5 Bull. B. M. (N. H.) Geol. 10, 10 PLATE 6 Fics. 1-5. Baculites anceps Lamarck. Calcaire a Baculites, Manche, France. Fig. 1, 6409. Figs. 2-5, Valognes, Manche, C. 70595, C. 70596, C. 70600 and C. 70630 respectively. Fics. 6, 7. Baculites subanceps Haughton. Carimba, Angola. Fig. 6, lectotype, South African Museum no. 6829. Fig. 7, C. 52729. All figures natural size. 6 a PLATE 4 Bull. B. M.(N. H.) Geol. 10, 10 PLATE 7 Fic. 1. Baculites subanceps Haughton. Cross section of large specimen. Carimba, Angola. Paralectotype, South African Museum, no. 6829. Fics. 2-6. Didymoceras subtuberculatum sp. nov. 1 km. north of Egito, Angola. Fig. 2, holotype, C. 52701. Figs. 3-6, paratypes, C. 52703, C. 52708, C. 52705, C. 52696 respectively. Fig. 5c is a view of the lower surface of the spire showing the paired tubercles. All figures natural size. Bull. B. M. (N. H.) Geol. 10, 10 AL AMIN 7 PLATE 8 Fic. 1. Didymoceras cf. californicum Anderson. Carimba, Angola. C. 52727. Fic. 2. Didymoceras cf. angolaense (Haughton). Barra do Dande, Angola. Fig. 2b, view of top of spire. C. 52739. Fias. 3,5. Nostocevas helicinum (Shumard). Barra do Dande, Angola. Figs. 5a, b, c, views of top, side and base of spire. C. 52753 and C. 52738 respectively. Fic. 4. Didymoceras cf. hornbyense (Whiteaves). Barra do Dande, Angola. 4a, b, c, views of top, side and base of spire. C. 52737. Fic. 6. Nostoceras cf. kevnense (Anderson). Barra do Dande, Angola. Views of side and base of body chamber hook. C. 52746. All figures natural size. PIL ANIIS, £5) Bull. B. M. (N. H.) Geol. 10, 10 PLATE 9 Fics. 1, 2. Nostoceras hyatti Stephenson. Barra do Dande, Angola. Figs. _ views of the outer periphery of the body chamber hook. C. 52743 and C. 527 All figures natural size. Bull. B. M. (N.H.) Geol. 10, 10 PLATE 9 PLATE to Fic. 1. Nostocevas hyatti Stephenson. Barra do Dande, Angola. C. 52742. Fic. 2. Nostoceras (?) obtusum sp. nov. Holotype, Barra do Dande, Angola. C. 52744. Fic. 3. Nostoceras votundum sp. nov. Holotype, Barra do Dande, Angola. Fig. 3d is a view of the base of the body chamber. C. 52745. All figures natural size. Bull. B. M. (N. H.) Geol. 10, 10 PLATE 10 IPILJNIP IS, iar Fic. 1. Phylloptychoceras sipho (Forbes). Lectotype, Valudayur Beds (Campanian— Lower Maastrichtian), Pondicherry, India. Fig. 1c is a view of the top of the loop. C. 51153. Fic.2. Polyptychoceras pseudogaultianum (Yokoyama). Barra do Dande, Angola. C. 52754. Fic. 3. Desmophyllites diphylloides (Forbes). Lectotype, Valudayur Beds (Campanian— Lower Maastrichtian), Pondicherry, India. C. 22682, x1I°5. Fics. 4-6. Kitchinites angolaensis sp. nov. 1 km. north of Egito, Angola. Fig. 4, paratype, with fragment of Didymoceras subtuberculatum in matrix, C. 52682. Fig. 5, holotype, C. 52675. Fig. 6, paratype, C. 52680. All figures natural size, except fig. 3, x1°5. Bull. B. M. (N. H.) Geol. 10, 10 MILANI, JU Fics. 1, 4. Eupachydiscus pseudogrossouvyei Collignon. 1 km. north of Egito, Angola. PLATE 12 C. 52670 and C. 52674 respectively. Fic. 2. Fie. 3. Manambolites dandensis sp.nov. Paratype, Barra do Dande, Angola. C. 52734. Hoplitoplacenticeras cf. maryoti (Coquand). 1 km. north of Egito, Angola. C. 52685. All figures natural size. Bull. B. M. (N. H.) Geol. 10, 10 ALAN, V2 PLATE 13 Fic. 1. Manambolites dandensis sp. nov. Holotype, Barra do Dande, Angola. C. 41474. Fic. 2. Hoplitoplacenticeras cf. costulosum (Schliiter). 1 km. north of Egito, Angola. C. 52686. Fic. 3. Hoplitoplacenticeras cf. marvoti (Coquand). 1 km. north of Egito, Angola. C. 52684. All figures natural size. Bull. B. M. (N. H.) Geol. 10, 10 PLATE 13 bb . 4 5 wee ‘ { vt * : ; * e b oh ’ i i + ist i | My wv) sat. _— i \ M > i , ¥, Be At i nN Wet a We eae feta Nae oe TY wi) " me : ia 8 7 4 LIMESTONE, Ul THE CORALS id ; - -D. KALJO anp E. KLAAMANN A SSA hey} BULLETIN: OF | BRITISH MUSEUM (NATURAL HISTORY) BRE Mca | Vol. 10 No. 11 ~ LONDON : 1965 | re" ioe FAUNA OF THE PORTRANE LIMESTONE, tH 7 DEC1965 a LHe CORALS se S Ss Seay We BN DIMITRI KALJO AND EINAR KLAAMANN (Institute of Geology, Academy of Sciences of Estonian S.S.R., Tallinn) Php. 413 — 434; 4 Plates ; 1 Text-figure BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. to No. 11 LONDON : 1965 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted im 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become veady. 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. 10, No. 11 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) 1965 ADIRGIONS AN 1DIB.S) (O)iEY THE BRITISH MUSEUM (NATURAL HISTORY) Issued December, 1965 Price Eighteen shillings THE EAUN AOR’ THEY PORTRANE LIMESTONE, Ill TRE (CORALS By D. KALJO anp E. KLAAMANN CONTENTS Page I INtrROopDUCTION : é : : : : : : 416 Il SySTEMATIC DESCRIPTIONS : c : : : . : 410 RuGosa By D. Karjo . : F é : : 416 Family StTREPTELASMATIDAE Nicholson : : : : : 417 Kenophyllum sp. : : : : : : 417 Kenophylium cf. inflatum (Dy bore ski) . 3 : Z ; 417 Streptelasma fragile Wilson 0 : ; é : : 418 Streptelasma distinctum Wilson . ; 0 : : , 418 Streptelasma cf. vusticum (Billings) : : : ; : 419 Grewinghia europaea (Roemer) . : é ° ; : 420 Grewingkia hibeynica sp.nov. . : : : : : 420 Brachyelasma ct. duncani (Dybowski) . ; . : : 421 Dalmanophyllum subduplicatum (MW ane : : : j 422 TABULATA BY E. KLAAMANN : c 3 : : 22 Family Syringophyllidae Pocta . F : F : c 422 Sarcinula sp. . ‘ : F : : : : 22 Family Lyoporidae ies , : 3 ; ; é ‘ 423 Reuschia sp. . 4 : : ; é : : : 423 Reuschia ? sp. , : . : : 424 Family lee idee aginenes & Evan : 3 : F é 424 Catenipora tapaensis (Sokolov) . ‘ : : ‘ ¢ 425 Catenipora wrighti sp. nov. : : : 2 , : 427 HELIOLITOIDEA By E. KLAAMANN . é . ; : : 428 Family Coccoserididae Kiaer : : : : : 428 Coccoserididae gen. et sp. indet. P : ‘ ; ' 428 Esthonia astevisca (Roemer) . : : ; . : 429 Pragnellia cf. avborescens Leith : é : 5 5 430 Family Heliolitidae Lindstr6ém . : ; : 3 5 430 Wormsipora hirsuta (Lindstrém) i : ; : : 430 Wormstpora portranensis sp.nov. . ; ; ¢ : 431 Family Proheliolitidae Kiaer . 6 . . 2 . 432 Proheliolites dubius (Schmidt) . : : i : : 432 Ill REFERENCES . 5 7 . A 2 j 3 : - 433 SYNOPSIS The present work forms part of a series of papers dealing with the fauna of the Portrane Limestone and contains a description of 20 species of corals—Rugosa, Tabulata and Heliolitoidea —of which three are new. The corals from the Portrane Limestone have much in common with those of the Norwegian 5a and Estonian Pirgu Stages. For this reason it may be assumed that the Portrane Limestone is of early Ashgill age. All the figured specimens are now in the British Museum (Natural History). 4106 THE CORALS OF THE PORTRANE LIMESTONE I INTRODUCTION THE present work, representing a part of the study of the Portrane Limestone fauna, is based on a small collection of corals kindly put at the disposal of the authors by Dr. A. D. Wright of Queen’s University, Belfast, to whom they wish to express their sincerest thanks. The corals were etched out from the rock by chemical methods. As they are considerably silicified and rather fragile, it was not possible to apply the usual method of study by means of thin sections, and the examination was restricted to external observations, and, in the case of the Rugosa, to an examination of the visible part of the calice only. In a number of cases, in the course of preparation of the corals by the chemical methods, some skeletal elements were more dissolved than others, or even disappeared altogether, and, as a result, the natural structure had undergone alteration. Therefore the entire collection was not determinable to an equal degree of exactitude, and a large part of it had to be left aside. However, it has been possible to define an interesting complex of corals, a description of which follows. For each species, the distribution outside Ireland is given. Here it is of interest to note that the majority of the fauna studied is known to occur in Norway and Estonia and some species in the Richmondian of North America. Among them are such typical Upper Ordovician species as Grewingkia europaea (Roemer), Catentpora tapaensis (Sokolov), Proheliolites dubius (Schmidt) and others. A detailed examination of the fauna of the Portrane Limestone proves it to have a great resemblance to that of the Norwegian 5a and Estonian Pirgu Stages. In the complex described, the typical species of the Norwegian 5b and Estonian Porkuni Stages are practically absent. Thus the Portrane Limestone clearly belongs to the early Ashgill. Recently Kaljo, Klaamann & Nestor (1963) showed that the Ashgillian coral faunas of Estonia and Norway have many features in common. On this basis it may be briefly stated that, commencing with the Ashgillian, a uniform coral fauna was developed throughout the North European zoogeographical province and that it was distributed over a wide area. II SYSTEMATIC DESCRIPTIONS RUGOSA By D. KALJO The Rugosa of the Portrane Limestone are rather varied and numerous. Unfortunately the poor state of preservation does not allow all the details of their structure to be observed, the best preserved and most clearly seen parts being the calices. Whilst studying this material, the author came to the conclusion that the recent investigators of corals have paid too little attention to the calice. The details of its structure are features worthy of specific importance at least. However, the present state of knowledge of the Rugosa is such that, before the structure of the calice can be used in taxonomy, it will be necessary to study good material in order THE CORALS OF THE PORTRANE LIMESTONE 417 to correlate the characteristic features of the calice with those internal features which are usually given taxonomic importance. It is of interest to note in the Portrane specimens the presence of a varying scar of fixation which was observed in nearly all of the species examined. These scars may be narrow, slit-shaped (Pl. 1, fig. 6), of varying size, sole-shaped (PI. 1, fig. 12), or burrow-shaped. Their frequent occurrence and rather considerable size (the scar in Pl. 1, fig. 12, being 32 mm. long and 16 mm. wide in a corallum only 55 mm. long) shows that these rugose corals lived in relatively mobile water where it was necessary to have a firm attachment. The varying shape of the scars points to the fact that fixation occurred either directly on to the rough detrital bottom or on to larger bodies of an elongate or flat shape. Family STREPTELASMATIDAE Nicholson Genus KENOPHYLLUM Dybowski 1873 Kenophyllum sp. (Bl tigss 12) DEscriPTION. The corallum is simple, medium-sized, horn-shaped, with the proximal part trochoid and the distal ceratoid; the incomplete length is 36 mm. and the diameter 22 mm. The septa are thick and fill the whole lumen in the proximal part, but distally they become slightly thinner, with narrow interseptal loculi between them. The exact number of major septa is not certain, but it approaches 36-38. They reach to the axis, but their axial parts are to some extent separated from the rest of the septa and anastomose with each other, forming a compact axial structure, the elements of which are mostly of an irregular, oblong shape. The peripheral stereozone is narrow. The tabulae are not seen. REMARKS. The state of preservation of the specimen studied does not allow an exact identification, but it is obviously an undescribed species. The closest form seems to be Kenophyllum canaliferum Kaljo (Nabala Stage, Upper Ordovician, Estonia), which, however, differs from the Portrane specimen in external appearance, the number of septa and, probably, the shape of the calice. With respect to the number of septa and, to some extent, the external appearance, the form described also resembles K. subcylindricum (Dybowski) (Upper Ordovician, Nabala, Vormsi and Pirgu Stages, Estonia), although the latter has well-developed minor septa. Kenophyllum cf. inflatum (Dybowski) (PI. 1, figs. 3-5) DESCRIPTION. The corallum is simple, small, mostly trochoid, the proximal part slightly curved. A deep, saucer-shaped calice makes up about two-thirds of the length of the corallum. The largest specimen measured is 24 mm. in length and 20 mm. in diameter, but more commonly the length varies from 16-19 mm. with a 418 THE CORALS OF THE PORTRANE LIMESTONE diameter of 12-16 mm. The septa in the calice are short, but in the proximal part the major septa reach the axis. The major septa vary in number with the growth of the calical diameter, as follows (diameter in mm.: the number of major septa) :— 12: 24; 13: 28; 15:30; 20:33. It may be noted that the increase ofthe septa distally slows down in comparison with the growth of the diameter—in other words, in the distal direction the intervals between the septa become larger. The minor septa are very short. The cardinal septum may be placed on either the convex or (though less frequently) the concave side of the corallum. The cardinal fossula is small and narrow. No tabulae were observed. REMARKS. Dybowski (1873: 347) gave a very brief, unillustrated description of Petraia inflata of the Estonian Upper Ordovician, which is very closely related to the Portrane specimens, but which is larger, with a greater number of septa. Kaljo (1958: 24), pointing out that P. inflata differs from Kenophyllum siluricum (Dybowski) only in external appearance, treated both forms as one species. At present, however, the author attributes a greater taxonomic significance to the details of the calice and considers it necessary to revise this group of Rugosa; therefore for the time being he considers the Portrane forms as K. cf. inflatum (Dybowski). In Estonia the corals of the znflatum type occur in the Upper Ordovician Vormsi Stage and are rather rare in the lower part of the Pirgu Stage. Genus STREPTELASMA Hall 1847 Streptelasma fragile Wilson (Biz figs: 6577) 1926 Streptelasma fragile Wilson: 11, pl. 1, figs. 1, 2 DESCRIPTION. The corallum is small, slender, ceratoid, the proximal part slightly curved. The largest specimen measured is 24 mm. in length and 9 mm. in diameter. The calice is funnel-like, deep. Septa thin, the cardinal septum placed on the convex side, generally off the plane of the greatest curvature. In the proximal part, the major septa reach the axis, their number being 16-20 in the case of a diameter of 5-6 mm. The minor septa are short, but distinctly visible. Walls thin; no tabulae have been observed. REMARKS. This splendid, small Rugose coral was described by Wilson (1926) from the Richmondian of the Rocky Mountains. The Portrane specimens agree well with its description, the Canadian specimens differing only in possessing a somewhat deeper calice. Streptelasma distinctum Wilson (Pl. 1, figs. 8, 9) 1926 Streptelasma distinctum Wilson: 12, pl. 1, figs. 6, 7. 1958 Streptelasma bystvowi Reiman: 33, pl. 1, figs. 4—6. THE CORALS OF THE PORTRANE LIMESTONE 419 DESCRIPTION. A small or medium-sized, horn-shaped, ceratoid or trochoid simple coral. Calice funnel-shaped with a wide bottom, one side of the calice being more sloping than is usual. The walls thin, the edge sharp. Major septa long, reaching the axis, where they may communicate with each other to some extent, or even intertwine. In the axial region some sparse pali-like structures may be observed. The number of major septa in the calice amounts to 43 (the diameter of the corallum at the bottom of the calice being 18-20 mm.). Minor septa are very short, the peripheral stereozone narrow. Tabulae are strongly convex at the periphery, and in the middle they are wavy. REMARKS. This species belongs to the group of Streptelasma corniculum Hall, one of whose peculiar features is a simple axial complex. In the number of septa, it bears a similarity to S. orventale Kaljo (1958: 21, pl. 2, figs. 1-4—-from the upper part of the Estonian Middle Ordovician), which, however, differs from it in the arrange- ment of the septa and the nature of the tabulae. S. poulseni Cox (1937: 9, pl. 2, figs. 8a—c, ga—-b—Cape Calhoun formation, Greenland) is even more similar but differs mainly from it in size and in the flatter and more numerous tabulae. It is obvious that this group of species requires a thorough revision, based on reliable material. S. distinctum has been previously described from the Richmond formation of the Rocky Mountains, British Columbia, aid from the Pirgu Stage of Estonia. Streptelasma cf. rusticum (Billings) (Pl. 1, fig. 10) DEscRIPTION. Simple, medium-sized, ceratoid coral. The specimen is broken and its incomplete length is only 23 mm. The cup-shaped calice has thick vertical walls and a convex floor. The external diameter of the calice is 23-27 mm. and the internal diameter 15-17 mm. The depth of the calice is 12 mm. The major septa almost reach the axis but leave the axial area free, revealing an axial structure which consists of separate elements. The number of major septa at the bottom of the calice of a corallum with a diameter of 19-21 mm. is 41. Minor septa in the calice are rather short, but lengthen to some extent in the proximal direction. The calical edge is rather thin. Tabulae are convex in the peripheral part. REMARKS. Streptelasma rusticum (Billings) has been described by Billings (1858)! and Lambe (1901) from the Hudson River formation of Canada, by Cox (1937) from the Richmondian of Canada and Ohio, by Wang (1948) from the Girvan Stinchar Limestone Group, and by Kaljo (1961) from the Estonian Pirgu Stage (Upper Ordovician). In these descriptions there is a considerable difference as to the axial structure: according to Lambe the major septa are “passing to the centre, where they are often considerably twisted’’, and further, ““Dissepiments . . . forming with the twisted inner ends of the primary septa the confused central structure charac- 1The author has not been able to obtain this work. The reference is taken from Lambe (1901 :110). 420 THE CORALS OF THE PORTRANE LIMESTONE teristic of Streptelasma’’ (Lambe 1901: 111). The later authors, however, deal with the forms of S. vusticwm whose major septa nearly reach the axis and whose axial structure is formed of the isolated inner ends of septa. Cox alone notes that the larger forms may have an anastomosing axial structure. The present author is of the opinion that in this instance two different species were dealt with, the larger forms coming from the Arctic being the genuine vusticum and representing the real Grewingkia, whereas the Ohio, Girvan and Estonian forms are nearer to Stveptelasma and obviously form a new species. The final solution of this problem, however, requires better and more complete material than that currently at the author’s disposal. The Portrane form, by the number of its septa, shape of the calice and wide stereozone is very closely related to the Girvan form, differing to some extent from the Estonian specimens. It also resembles S. craigense (M’Coy), though the charac- teristic feature of the latter is the occurrence of long minor septa. Genus GREWINGKIA Dybowski 1873 Grewingkia europaea (Roemer) (Elaers festa) 1861 Stveptelasma euvopaeum Roemer: 16, pl. 4, figs. 1a-f. 1933 Kiaerophyllum anguineum Scheffen: 23, pl. 3, figs. 3, 4. 1948 Streptelasma (Kiaerophyllum) europaeum (Roemer) Wang: 103, pl. 7, figs. 1a—b, text-fig. 4. 1961 Streptelasma (Grewingkia) euvopaeum euvopaeum (Roemer) Kaljo: 57, pl. 2, figs. 2-9, text-fig. 3. DESCRIPTION. Simple, conical, medium-sized corallum. The cup-shaped calice is of medium depth, with thin, almost vertical walls, and its floor convex. Septa somewhat thickened in ephebic stages. The major septa number 40 in the case of a diameter of 16 mm. and reach the axial structure, which is wide and consists of connected septal ends of different shapes. The peripheral stereozone is narrow, I-1.5 mm. The minor septa project to some extent from the stereozone. The tabulae are abruptly convex at the periphery. REMARKS. The Portrane specimens, though somewhat smaller, are very closely related to the Estonian and other forms in their general structure. G. europaea differs from the species of the G. buceros group in possessing a very narrow stereozone and an axial structure which is not compact. The species is found in the Pirgu Stage of Estonia, in the Norwegian Stage 5a, and in the Stinchar Limestone Group of Girvan. Grewingkia hibernica sp. nov. (Pl. 1, figs. 12-14) Diacnosis. Grewingkia of moderate size, straight or moderately curved, the THE CORALS OF THE PORTRANE LIMESTONE 421 calice shallow. Septa short, but in the calice relatively long. The axial structure wide, about one-third of the diameter of the corallum. HoLotTyPe. R.453109. FIGURED PARATYPES. R.45318, R.45512-13. DEscrRIPTION. Mostly medium-sized, straight or moderately curved Rugose corals. The length of the corallum attains 70 mm., the greatest diameter 40 mm. The epitheca is covered with poorly preserved rugae and sometimes reveals horizontal growth-rings. The calice is shallow (9 mm. in the holotype), and, as a result of the axial structure penetrating to some extent into the calice, has a slightly convex floor. The septa in the calice are long at the base and only gradually shorten towards the upper rim of the calice. The septa are moderately thick. The major septa reach the axial structure where they may form compact groups connected by septal ends. The holotype has a diameter of 25-28 mm. and 50 major septa; in another specimen with the diameter of 33-38 mm. their number is 54. The minor septa are relatively long, about one-fifth to one-quarter of the length of the major septa. The stereozone is narrow (I-2 mm.) and segmented. The axial structure is wide, occupying approximately one-quarter of the diameter of the corallum, and consists of a complex of granular and oblong axial elements, intertwined with each other. The tabulae, which could be examined only in some fragments, showed a curvature at the periphery. ReMARKS. The Portrane form differs from all the other Grewingkia species in the peculiar structure of its calice. Genus BRACHYELASMA Lang, Smith & Thomas 1940 Brachyelasma cf. duncani (Dybowski) (Pl. 1, fig. 15) DESCRIPTION. Corallum simple, ceratoid to cylindrical, somewhat compressed, medium-sized. The corallum is about 40 mm. in length and the maximum diameter 18-20 mm. The calice is cup-shaped, with thin vertical walls and flat floor. The stereozone is narrow. Septa thin and short, the axial area open, containing only some single “‘pali’”. Minor septa short. Tabulae rather curved at the periphery, flat or slightly wavy in the middle. RemMArRKS. The form described from the Portrane Limestone is closest to B. duncami (Dybowski) (5a of Norway, Pirgu Stage of Estonia) in its general size and shape as well as in the width of the stereozone and number of septa. Owing to ignorance of some details of the structures (e.g. that of the stereozone and of the axial zone in ontogeny), it is impossible to give a precise identification of the Portrane specimens. B. undulatum (Scheffen), from Stage 5a of Norway, is also closely related, but the latter has a greater number of septa. The other species of Brachyelasma have either longer septa or a different axial structure and a wider stereozone. 422 THE CORALS OF THE PORTRANE LIMESTONE Genus DALMANOPHYLLUM Lang & Smith 1939 Dalmanophyllum subduplicatum (M Coy) (Pl. 1, figs. 16-18) 1850 Petraia subduplicata M’Coy: 279. 1878 Lindstrémia subduplicata (M’Coy) Nicholson & Etheridge: 86, pl. 6, figs. 2-2f, text-fig. 4. DESCRIPTION. Corallum simple, small, ceratoid. Maximum dimensions: length 22 mm., diameter 15 mm. Epitheca covered with fine rugae. The funnel-shaped calice has a sharp rim. A strong axial column, consisting of a large central element surrounded by intertwined, separated, small septal ends, penetrates into the calice, below the floor of which it is rather wide, occupying about one-half of the diameter of the lumen. The number of major septa is 27-31 when the diameter of the calice is I2-15 mm. The minor septa are rather short. Between the major septa, particu- larly in the distal part of the coral, the interseptal loculi are wide. The peripheral stereozone is narrow. The tabulae are convex. ReEMARKS. The Portrane Limestone specimens are very similar to those from Girvan described by Nicholson & Etheridge (1878), except that the septa in the proximal part of the former are less thickened by stereome than those of the latter. I am not altogether convinced that some of the interseptal loculi in the proximal part of the corallum have not become wider in the course of the chemical preparation. D. subduplicatum has been described as occurring in the Craighead Limestone (Caradoc) and Upper Llandovery of Girvan, the Upper Ordovician of Wales, and in Stage 5a of Norway (Kiaer 1897). TABULATA By E. KLAAMANN Order SARCINULIDA Family SYRINGOPHYLLIDAE Poéta 1902 Genus SARCINULA Lamarck 1816 Sarcinula sp. (Pl. 2, figs. 1-4) DEscRIPTION. Corallum nodular, irregularly shaped, the maximum diameter being 40 mm. and the maximum height 35 mm. On account of the poor state of preservation at the surface of the colonies, only the rounded openings of the inner cavities can be observed there: their diameters are about 2:5—3-0 mm. and they are placed at intervals of 2-5-3-0 mm. from each other. In some cases it was possible to measure the true diameter of the corallites, namely, 3:5 to 3-7 mm. and in rare cases even 4:°0mm. The walls of the corallites are thick, 0-7-1-0 mm. The numerous pores penetrating the walls of the corallite in horizontal rows can be clearly seen. These rows of pores open between the plates connecting neighbouring corallites. The THE CORALS OF THE PORTRANE LIMESTONE 423 connecting plates are very closely arranged, without plate-free intervals, and their average number is 6 in 5 mm. Thus the intercorallite tissue shows considerable resemblance to that of Calapoecia. The tabulae are badly preserved; they are slightly concave or curved and spaced at intervals of 0-5 to 3:5 (?) mm. The septal apparatus is represented by short laminar septa, whose number in a corallite exceeds 20 (probably amounts to 24). In a number of places thin ray-like “‘costae’’ are seen to diverge beyond the limits of the corallite forming a peculiar septal halo around it. REMARKS. The relatively large distance between the corallites and the nature of the walls and of the septal apparatus show that, in spite of the intercorallite tissue resembling that of Calapoecia, we have here a typical representative of Sarcinula. Unfortunately, the unsatisfactory state of preservation of all the structural detail does not permit of an exact identification with any known species. The Irish specimens bear the strongest resemblance to S. luhai Sokolov (1951: 92-94, pl. 16, figs. 6-7; pl. 17, figs. 1-2) of the Pirgu Stage of Estonia and Stage 5a of Norway. This species possesses corallites of almost the same diameter (most frequently 3:5-3-7 mm.), closely proximate rows of pores, and connecting plates without plate-free intervals between them. In the size of its corallites, the Irish species also resembles S. latwm Sokolov (1951: 91-92, pl. 16, figs. 3-5), but the corallum of the latter always has a flat discoid shape, and between the connecting plates there are sharply outlined free intervals of 0-5—3-00 mm. in width. S. latwm is also known from the Pirgu Stage of Estonia. Until recently it was assumed that the only representative of Sarcinula was S. organum (Linné). However, the latest researches in the Baltic area reveal that, e.g., in the Upper Ordovician of Estonia, this species is represented very rarely and by only a very limited number of individuals, and that the main réle is played by other species, and in particular by S. luhai and S. latum. S. organum differs from those species in the much smaller diameter of its corallites (2°5-3-0 mm.), and thus is but a rather distant relation to them as well as to the Sarcinula from the Portrane Limestone. It is possible that the forms from England described and illustrated by Milne-Edwards & Haime (1854: 295, pl. 71, figs. 3-30) as Syringophyllum organum are very closely related to the Portrane Sarcinula. Order LICHENARIIDA Family LYOPORIDAE Kiaer 1930 Subfamily EOFLETCHERIINAE Sokolov 1955 Genus REUSCHIA Kiaer 1930 Reuschia sp. (Pl. 2, fig. 5) DESCRIPTION. The collection contains a small fragment (15 x 15 mm.) of a colony consisting of 17 thick-walled, tubular corallites, which now and then are in contact with each other along their whole length, or stand at a distance of I-1-5 mm. from each other. The diameter of the corallites varies from 1-8 to 2:2 mm. compared 424 THE CORALS OF THE PORTRANE LIMESTONE to which the thickness of the walls seems striking—o-6—0:8 mm. and in some cases even 1 mm. Hence the interior vacuity is very narrow (PI. 2, fig. 5) and seldom exceeds 0:5-0'8 mm. in diameter. At the openings of the corallites the thick walls show a trabecular structure, but owing to the poor state of preservation, it was impossible to determine the number of trabeculae. Tabulae were not observed. REMARKS. Up to the present time the only representatives of Reuschia which have been described are those of the Upper Ordovician of Norway (Kiaer 1930; Hill 1953) and China (Yi 1960). All these investigators define only one species, R. aperta Kiaer. The Portrane specimen differs from typical Norwegian repre- sentatives by having corallites of smaller diameter which are in closer contact. The Chinese F. aperta, however, has even larger corallites than the Norwegian form, and in addition has thicker walls; thus it seems that in this instance a new, separate species should be instituted. Reuschia? sp. (Pl. 2, fig. 6) DESCRIPTION. The small bushy colony does not exceed 40 mm. in height. It is composed of cylindrical corallites budding like Auwlopora and forming bunches of corallites, the central ones of which are orientated in a more or less vertical direction, whilst those at the outside bend slightly towards the periphery. The diameter of the corallites is constant throughout their whole length, mostly amounting to 2:5-3-0 mm. The calices are deep, with a circular cross-section and smooth rims. The better preserved corallites show a slight contraction at the opening, the calices developing a barrel-like shape. The thickness of the walls varies between 0-5 and 0-7 mm. They are composed of narrow trabeculae whose ends, in some corallites, extend inwards beyond the stereozone and, owing to their dense arrangement, form low vertical rugae. The latter are the only structures in the interior cavity of the corallites, since tabulae are missing altogether. REMARKS. The mode of budding and the form of the corallites stress the great similarity of this species to Eofletcheria. However, the total absence of tabulae and the comparatively great thickness of the corallite walls indicate that we may consider the specimen to belong to Rewschia. Of the representatives of this genus, R. aperta, described by Yui (1960:97-08, pl. 9, figs. 4-5; pl. Io, figs. 5-8) from the Upper Ordovician of China, bears the greatest resemblance to the Portrane specimen, from a consideration of the size of the corallites, at least. Order HALYSITIDA Family HALYSITIDAE Edwards & Haime 1850 Subfamily CATENIPORINAE Hamada 1957 Genus CATENIPORA Lamarck 1816 The Halysitida are represented in the collection by more than 60 small irregular THE CORALS OF THE PORTRANE LIMESTONE 425 colonies or parts of colonies. The specimens are, however, poorly preserved, so that in many instances it was impossible to examine the tabulae and the septal spinules and to establish the original thickness of the walls of the corallites. Accordingly, of the principal criteria that normally serve as a basis for taxonomic determination of the order, only the dimensions of the corallites and the form of the lacunae could be used. These two characters do, however, indicate two groups of forms, one of which is certainly identical with Catenipora tapaensis (Sokolov), a species of wide distribution in the Upper Ordovician of Baltoscandia. Catenipora tapaensis (Sokolov) (Ple2, fie: 12) 1854. Halysites catenularia (part.); Edwards & Haime: 270-272. 1858. Catentpora labyrinthica Fischer (part.) ; Schmidt: 229. 1860. Halysites eschavoides Lam. (part.); Eichwald: 507. 1915. Halysites eschavoides Fischer-Benzon; Yabe: 34 (10), pl. 6 (2), figs. 3, 4. 1951. Palaeohalysites tapaensis Sokolov: 81-82, pl. 14, figs. I, 2. 1951. Palaeohalysites prirsaluensis Sokolov: 84-85, pl. 14, figs. 5-7. 1951. Palaeohalysites kuruensis Sokolov: 85-86, pl. 15, figs. 3, 4. 1955- Palaeohalysites privsaluensis Sokolov; Sokolov, pl. 65, fig. 2. DESCRIPTION. The corals have bushy, hemispheric or weakly cushion-shaped colonies whose diameter does not exceed 10 cm. The corallites form irregular nets on the sides of whose meshes are 1 to 6 corallites, in most cases 2 to 4. The lacunae are irregularly polygonal, mostly curved and oblong; their maximum diameter varies between 3 and 15 mm. Transversely the corallites are elliptical, the longer axis I-3 to 1-7 mm., the shorter 0-9 to 1:2 mm. (Text-fig. 1, I)—there are almost no variations of these dimensions. The corallite walls are rather thick, 0-2-0-3 mm., and those between adjoining corallites are about twice as thick. The interval between the horizontal or slightly concave tabulae amounts to 0:5 to o-°8 mm. On account of the silicification, the septal spinules have been poorly preserved, occurring rarely; the corallites probably possessed twelve rows of spinules originally. Remarks. The Portrane forms reveal almost absolute similarity to Baltoscandian representatives of C. tapaensis, differing from them only in the somewhat lesser convex form of the corallite walls. This is expressed by the fact that the average long transverse diameter of the corallites of the former is approximately 0-I mm. greater and the shorter diameter about 0-1 mm. less than the corresponding average measurements of the Estonian and Norwegian representatives of this species. As seen in the synonymy, the conception of the species C. tapaensis in the present work is much wider than that of the author of the species who distinguished three separate species. This subdivision was based on small differences in the form of the corallites, in the thickness of their walls and in the development of septal spinules. The investigation of a great number of specimens of Catenipora from the Upper Ordovician of Estonia has shown that C. tapaensis (sensu Sokolov), C. piirsaluensis and C. kuruensis, established by Sokolov (1951) on the basis of limited material, are 426 THE CORALS OF THE PORTRANE LIMESTONE min Dp /4 2,0 if 1,0 Long diameter of coraliite Q5 1,0 15 mm Short diameter of corallite Fic. 1. Average dimensions of Cateniporva tapaensis (Sokolov) (1) and C. wrighti sp. n. (II) computed from 22 and 9g colonies, respectively. really connected with each other by a large number of transitional forms. Inasmuch as all the quoted forms have not only a morphological similarity but also equal stratigraphical distribution in Estonia, it can be concluded that they belong to one and the same species, which should be called C. tapaensis (Sokolov) according to page priority. In all likelihood, a part of the Halysitids from Portrane described by Edwards & Haime (1854: 272) under the name Halysites catenularia belong to the species discussed. DIsTRIBUTION. Ireland, Portrane Limestone; Norway, Ringerike (Stavnae- stangen), Upper Ordovician, Stage 5a; Estonia, Upper Ordovician, Vormsi and Pirgu Stages. THE CORALS OF THE PORTRANE LIMESTONE 427 Catenipora wrighti sp. nov. (Pl. 2, figs. 7-11) Diacnosis. Corallum bushy, diameter not exceeding 50mm. Corallites connected into small irregular meshes, the sides of which are composed of 4 corallites or less: corallite diameter 0-7-1-0 X I-I-I-5 mm. Tabulae and septal spinules present. HOLoTyPE. R.45329, a small corallum, 20 mm. in diameter. FIGURED PARATYPES. R.45330-33. LocaLiry AND Horizon. Ireland, Portrane; Upper Ordovician, Portrane Limestone. DESCRIPTION. Corallum small, bushy, of irregular form, with a diameter ranging from 20 to 40 mm. The small elliptical corallites are joined into meshes, at the sides of which there are usually from I to 4 corallites, though the maximum is 8. As a result, the shape of the lacunae varies from the more common irregular polygon, with a diameter of 3-20 mm., to an elongated and meandering form. In single cases there is a locally dense disposition of the chains, so that the lacunal area is reduced toa minimum. The average diameter of the corallites is 0-9 x I-3 mm., but it varies from 0-7-I'0 X I:I-I°-5 mm. The relation of the short diameter of the corallite to the long one is approximate 1: 1:5 (Text-fig. 1, II). The thickness of the exterior walls varies from 0-15 to 0:2 mm., the interior walls (those between the corallites) being somewhat thicker. Tabulae horizontal, their distance apart about 0:5-0-7 mm. Owing to the changes in the material as well as to the chemical processing, only the basal part of the septal spinules is preserved, which in places are represented by short blunt tubercles on the interior walls of the corallites. RemARKS. Among the undisputed Ordovician Halysitid species known at present (whose number is less than 20), the dominant representatives belong to Catentpora and Quepora. These genera have an equal vertical range, and the only criterion for distinguishing them is the presence or absence of septal spinules in the visceral chamber of the corallites, a character depending to a great extent on the state of preservation. As observed by the author, the septal spinules of the Halysitids may be destroyed by minor diagenetic processes without any striking changes in the other elements of the skeleton. This circumstance renders a practical application of these characters extremely complicated. In my opinion, in the current systematics of the subfamilies Cateniporinae and Halysitinae, too great a significance has been attached to septal spinules, or, rather, to their absence. For these reasons a comparison of the new species with the Ordovician species of Quepora as well as Catentpora is given. Catenipora wrighti shows the greatest similarity (particularly in the size of corallites) to C. tapaensis (Sokolov) described above, as well as to Quepora aequabilis (Teichert) from the Trenton of Arctic Canada, Q. quebecensis (Lambe) from the Middle Ordovician of Quebec, and Q. (?) parallela (Schmidt) from the Ashgillian of Estonia (Pirgu and Porkuni Stages). C. tapaensis has larger colonies and thicker walls, in particular those between corallites, and a greater diameter of corallites 428 THE CORALS OF THE PORTRANE LIMESTONE (cf. I and II, Text-fig. 1). The majority of the corallites of the new species have a diameter of 0-9 x 1-3 mm., whereas in C. tapaensis the dimensions 1-0-1:2 1‘6 mm. are of most common occurrence. Q. aequabilis is distinguished by very small lacunae and much narrower corallites —only 0-4-0-7 mm. Q. quebecensis has larger and thicker-walled corallites from which no septal spinules are known. The Baltic-species, Catenipora parallela Schmidt (1858: 229), attributed to Quepora by Hamada (1957), differs only slightly in the dimensions of the corallites (0-75-0°9 X I:2-I'4 mm.) and thickness of the walls (0-15-0-2 mm.). However, it is clearly distinguishable by the almost straight parallel walls of the corallites and the long, curved, commonly unconnected chains. It would appear from this last mentioned character that Q. parallela ought to be referred to Eocatentpora. Such Ordovician species as Q. (?) agglomeratiformis (Whitfield), Q. delicatula (Wilson) (both from the Richmond of Canada), Catentpora obliqua (Fischer-Benzon) (Nabala Stage of Estonia) and C. tractabilis (Sokolov) (Vormsi Stage of Estonia) are very different, having considerably larger dimensions of the corallites than C. wright. HELIOLITOIDEA By E. KLAAMANN Order PROTARAEIDA Family COCCOSERIDIDAE Kiaer 1899 Gen. et sp. indet. (Pl. 3, figs. 1-6) DESCRIPTION. Colonies irregular, flat, with a thickness of 7-15 mm. and a diameter seemingly exceeding 50-60 mm. A number of specimens have fully or partly retained a somewhat wavy basal epitheca on the lower surface of the colony. The upper surface is poorly preserved, showing in single cases indefinite, low tubercles. The interior of the colony has been etched out; only some parts which border the lower and upper surfaces of the corallum are preserved. In the cross- sections of these parts a great number of quadrangular and hexagonal “‘tubules’’ are observed, the diameter of which keeps within the limits of 0-4-0-7 mm. The vertical sections, however, clearly reveal the pinnate microstructure of these “tubules”, a proof of their being, in reality, coenenchymal trabeculae which, as a result of secondary changes, have acquired a form of prismatic tubules of the type observed in Heliolites. No horizontal structures were discovered. RemARKs. The laminar corallum, the presence of rough, vertical trabeculae and the absence of tabulae all indicate that the poorly preserved forms described above belong to the Coccoserididae. But as the structure of the corallites is unknown to THE CORALS OF THE PORTRANE LIMESTONE 429 us, a more detailed definition of the material cannot be made. It is most probable that here we are confronted with representatives of Pvotavaea, a genus widely distributed in the Upper Ordovician of Baltoscandia, Great Britain and North America. Subfamily ACIDOLITINAE Sokolov 1950 Genus ESTHONIA Sokolov 1955 Esthonia asterisca (Roemer ) (Plea; figs. oN 10) 1858. Heliolites inordinata (part.); Schmidt: 228. 1861. Heliolites interstincta Linné; Roemer: 24, 25, pl. 4, figs. 4a, b. 1880. Heliolites intricatus Lindstrom var. lamellosus (part.) Lindstrém: 32, 33, pl. I, fig. 5. 1883. Heliolites asteriscus (part.) Roemer: 505. 1897. Heliolites asteriscus (part.) Roemer; Roemer: 505. 1899. Heliolites intricatus var. lamellosa Lindstré6m; Kiaer: 42-44, pl. 5, fig. 13; pl. 7, figs. 3-5; text-fig. Io. 1899. Acantholithus asteriscus (Roemer) (part.) Lindstré6m: 113, 114, pl. 11, figs. 31-35. 1903. Acantholithus asteviscus (Roemer); Kiaer: 10-12. 1955. Esthonia asterviscus (Roemer) Sokolov, pl. 71, figs. 1-6. 1955. Esthonia lamellosa (Kiaer) Sokolov, pl. 71, fig. 7. 1962a. Esthonia asteriscus (Roemer); Sokolov, pl. 2, figs. 1a—d. DEscrIPTION. The material consists of fragments of irregular, wavy and laminar coralla with a thickness of only 1-5—4-5 mm. On the upper surface of the laminae are Clearly discernible small, shallow calices placed at a distance of 0:5-1-5 mm. from each other. The diameter of the calices is 0-9-1'2 mm. They are clearly distinguished from the internal skeleton consisting of angular, thick-walled, intermediate tubules, about 0-15—-0-2 mm. in cross-section. The twelve septa penetrate the interior chamber of the corallites to a depth of 0-2—-0:25 mm. In the central part of the calice are what appear to be fine tubercles, but which are in reality formed by ends of axial trabeculae. No tabulae were detected either in the corallites or in the inter- mediate tubules. Remarks. The Portrane specimens discussed doubtless belong to typical representatives of Esthonia asterisca (Roemer), a form frequently described from the Upper Ordovician of Baltoscandia and most commonly referred to as Heliolites intricatus var. lamellosa Lindstr6m and Acantholithus asteriscus (Roemer) (see synonymy). As indicated by Lindstrém himself (1899), the first-mentioned name is not correct, since at the time of the establishment of the variety the author erroneously connected under this name two different species of Roemer—Heliolites parvistellus and H. asteriscus. For these reasons it cannot be considered correct to restore the name lamellosa, as was done by Sokolov (1955, pl. 71, fig. 7), for designating the forms of Esthonia asterisca which have a very thin encrusting colony. In the course of time the species Esthonia asterisca (Roemer) was defined with greater precision. In order to avoid possible misunderstandings concerning this 430 THE CORALS OF THE PORTRANE LIMESTONE species it would be advisable first of all to establish its type; if Roemer’s original collection is lost, we recommend that the specimen from the Vormsi Stage of Estonia depicted by Sokolov (1955, pl. 71, figs. 1, 2) be used as neotype. Further, it is imperative to determine with greater precision the date of the establishing of the species discussed. The year usually quoted—1861—is incorrect, since the name H. asteriscus was first proposed by Roemer (1883) in his comment on the species Heliolites inordinata. DistrisuTion. Ireland, Portrane Limestone; Norway, 5a in Asker and Ringerik e ; Estonia, Vormsi and Pirgu (?) Stages. Genus PRAGNELLIA Leith 1952 Pragnellia cf. arborescens Leith (PL. 3, figs. 7, 8) DeEscRIPTION. The coralla have a curved, branching form, but at the base of the colony they swell out and encrust. The diameter of the branches varies from 3 to 8 mm., the length is about 30 mm. Corallites small, rounded, with a diameter mostly about I-o mm., the maximum being 1:2 mm. Calices very low, with short septa which have a considerably thickened base. The central part of the calice often slightly bulges and, when in a good state of preservation, reveals a fine punctuation caused by the ends of septal trabeculae. The coenenchyme on the surface of the colony is represented by small tubercles (o-I mm. in diameter). The cross-sections of colonies also reveal a trabecular structure, the trabeculae being in radial arrange- ment. The rest of the details of the structure of the corallum have been destroyed. REMARKS. Pvagnellia is a rare representative of the Heliolitoidea of which only two species are known at present. The Portrane form resembles most of all P. arborescens Leith (1952: 795, pl. 11b, figs. 1-8) from the Upper Ordovician of North America, differing from it by greater intervals between calices on the surface of the colony and by a lesser forking of branches. Sokolov (1962, 1962a) gave the following data on the distribution of Pragnellia: Upper Middle Ordovician—Upper Ordovician of the Urals, Altai, North America; Pirgu Stage of Estonia. Family HELIOLITIDAE Lindstrém 1873 Genus WORMSIPORA Sokolov 1955 Wormsipora hirsuta (Lindstr6m) (Pl. 4, figs. 1-3) 1899. Nicholsonia megastoma M’Coy; Kiaer: 37-39, pl. 6, figs. 8, 9; pl. 7, figs. I, 2. 1899. Heliolites hivsutus Lindstrom: 64, pl. 11, figs. 18-22. 1903. Propora hivsuta (Lindstr6ém) Kiaer: 9, 12, 39-42. 1955. Wormsipora hirsuta (Lindstrém) Sokolov, pl. 74, figs. 1-3; pl. 81, figs. 3, 4. 1962. Wovmsipora hirsuta (Limdstr6m) ; Sokolov, pl. 4, fig. 2. THE CORALS OF THE PORTRANE LIMESTONE 431 DESCRIPTION. Corallum irregular, slightly elongated in the vertical direction, varying from 15-30 mm. in width, 20-40 mm. in height; composed of uniform, star-shaped corallites which often touch each other and whose diameter is I-7—2:0 mm. The corallites have clear contours, since their walls are considerably thicker than those of the coenenchymal tubules, which, in addition, have broken contours in cross- section. The septal apparatus serves as the most important character of the present species. It consists of numerous coarse spinules bent upwards and penetrating the corallites to a depth of 0-6 mm. The ends of the spinules are sometimes cleft. The spinules are arranged in distinct, vertical rows and placed close to each other, creating the impression not of spinules, but of 12 septal ribs indentated at the rim. Fine septal growths are to be noticed in places on other septal elements as well. The cavities of the corallites and coenenchymal tubules are dissected by convex tabulae and diaphragms. The average distance between tabulae in corallites is 0-4-0°7 mm. REMARKS. Judging by the good figures of the lectotype and exhaustive descriptions presented by Lindstrém (1899) and Sokolov (1955, 1962), the Portrane specimens discussed are identical in minutest details with Wormsipora hirsuta (Lindstrom) of Estonia. DISTRIBUTION. Portrane Limestone, Ireland; Vormsi Stage, Estonia. Wormsipora portranensis sp. nov. (Pl. 4, figs. 4-9) Diacnosis. Corallum small, hemispheric or irregular in shape. Diameter of corallites 2:5-3-°0 mm. Coenenchyme of thick-walled tubules with interrupted contours in tangential section. Septal spinules, joined at their bases, form 12 coarse ribs. Tabulae of corallites and tubulae horizontal or gently curved. HorotyPe. No. R.45344, an irregular colony, 25 X 35 mm. FIGURED PARATYPES. R.45345-47. Locatitry AND Horizon. Ireland, Portrane; Upper Ordovician, Portrane Limestone. DeEscriPTION. Hemispheric or oblong coralla with a diameter of 15-30 mm., on the surfaces of which open deep star-shaped calices of corallites, which may be at a distance of about 2 mm. from each other or may touch each other with their rims. The rims of the calices are somewhat raised in respect to the coarse-grained surface of the coenenchymal tissue filling the narrow intervals between the corallites. The diameter of the corallites keeps within the limits of 2-5-3-0 mm. Longitudinal sections reveal that the coenenchymal tubules are rather thick-walled with regularly distributed horizontal diaphragms. The latter are placed more densely than the tabulae in the corallites. The septal apparatus is represented by 12 coarse ribs, formed as a result of the union of the thick bases of spinules. Spinules long (0-5- 0-7 mm.), diverging upwards at a sharp angle from the ribs and frequently penetrating the overlying tabulae. Interval between the tabulae varying from 0-3 to 1:2 mm. 432 THE CORALS OF THE PORTRANE LIMESTONE REMARKS. In its structure, the new species is strikingly similar to Wormsipora hirsuta (Lindstrém), from which it differs in the greater diameter of the corallites and in the joined bases of the spinules, leading to the formation of coarse, spinose septal ribs. In its external appearance, the colony is extremely like the form described by Lindstrém (1880, pl. I, fig. 6) as Plasmopora conferta Edwards & Haime. In all likelihood this form is also Wormsipora, having, however, even larger corallites of 4 mm. diameter. Order PROPORIDA Family PROHELIOLITIDAE Kiaer 1899 Genus PROHELIOLITES Kiaer 1897 Proheliolites dubius (Schmidt) (Pl. 4, figs. 10-12) 1858. Heliolites dubia Schmidt: 226. 1861. Heliolites dubia Schmidt; Roemer: 26-27, pl. 4, figs. 5a—5b. 1880. Heliolites dubius Schmidt (part.); Lindstr6m: 32, pl. 1, figs. 3, 4 (only). 1883. Heliolites dubius Schmidt; Roemer: 505-500. 1897. Heliolites dubius Schmidt; Roemer: 505-5006. 1897. Proheliolites dubius (Schmidt) WKiaer: ro. 1899. Proheliolites dubius (Schmidt) (part.); Kiaer: 21-26, pl. 3, figs. 5, 6; pl. 6, fig. 5. 1899. Proheliolites dubius (Schmidt) (part.); Lindstr6m: 70-71, pl. 11, figs. 13-17. 1903. Proheliolites dubius (Schmidt); Kiaer: 6, 12. 1955. Proheliolites dubius (Schmidt); Sokolov: pl. 75, figs. 6, 7. 1956. Proheliolites dubius (Schmidt); Hill & Stumm: F461, text-fig. 348, 6a—6b. 1962a. Proheliolites dubius (Schmidt) ; Sokolov, pl. 6, figs. 4a—b. DEscrIPTION. Corallum irregular, hemispheric or elongated, the maximum diameter not exceeding 40 mm. Corallites with compact walls, rounded, but owing to their very dense arrangement, often assuming a polygonal form. They are surrounded by fine and sparse, mostly triangular or quadrangular coenenchymal tubulae, whose maximum number around one corallite is four. Corallites with a diameter of 0-9 mm. predominate, others varying within the limits of 0-8-1-0 mm. The diameter of the tubulae is about 0-3-0:5 mm. The septal apparatus is very peculiar, being arranged in 12 rows of short, unconnected spinules which, unlike those of the other Heliolitoidea, bend downwards. In cross-sections the spinules are seen as 12 points connected to the inner wall of the corallites. Tabulae horizontal, in the tubulae rather dense (0:2-0-3 mm. apart) and in the corallites sparser (o-5-I mm. apart). REMARKS. The only difference between the forms described and those from the Baltic consists in the smaller colonies. However, this character is typical of all the THE CORALS OF THE PORTRANE LIMESTONE 433 Tabulata and Heliolitoidea of Portrane discussed and may have been caused by ecological factors. DISTRIBUTION. Portrane Limestone, Ireland; 5a and 5b, Ringerike (Stavnaestan- gen) and Asker, Norway; Boda Limestone of Dalarne, Sweden; Pirgu-Stage, Estonia. III REFERENCES Bitincs, EF. 1858. Report for the Year 1857 of E. Billings, Esq., Palaeontologist. Geol. Surv. Canada Rep. Progr., 1857: 147-192. Cox, I. 1937. Arctic and Some Other Species of Streptelasma. Geol. Mag., London, 74: 1-19, Dlsten 2. Davis, W. J. 1887. Ientucky Fossil Corals—a Monograph of the Fossil Corals of the Silurian and Devonian Rocks of Kentucky, Pt. II. MKentucky Geol. Surv. [4] + i-xiii, 139 pls. Frankfort. Dysowsk!, W. 1873. Monographie der Zoantharia sclerodermata rugosa aus der Silurformation Estlands, Nord-Livlands und der Insel Gotland. Arch. Naturk. Liv-, Ehst- u. Kwil., Dorpat (1) 5: 257-414, pls. I, 2. Epwarps, H. M. & Haine, J. 1854 [1855]. A Monograph of the British Fossil Corals. Pt. 5. Corals from the Silurian formation. Palaeontogr. Soc. [Monogr.], London: 245-299, pls. 57-72. EICHWALD, C. E. Von. 1860. Lethaea Rossica ou Paléontologie de la Russie, 1. 1657 pp., atlas 38 pls. Stuttgart. Hamapa, T. 1957. On the classification of the Halysitidae I, Il. J. Fac. Sci. Tokyo Univ., 10, 3: 393-4309. Hiri, D. 1953. The Middle Ordovician of the Oslo Region, Norway. 2. Some Rugose and Tabulate Corals. Norsk geol. Tidssky., Bergen, 31: 143-168, 5 pls. Hirt, D. & Stumm, E. C. 1956. Tabulata in Treatise on Invertebrate Paleontology. Part F. Coelenterata: '444—-F 477, text-figs. 340-357. (Editor R. C. Moore). Geol. Soc. Amer. and Univ. Kansas Press, Lawrence. Karjo, D. L. 1958. On the taxonomy of the genus Streptelasma Hall and a description of some new Rugose corals. Geoloogia Inst. Uurim., Tallinn, 2: 19-26, pls. 1, 2 (In Russian with English summary). 1961. Some additional data on the study of Ordovician streptelasmids in Estonia. Geoloogia Inst. Uurim., Tallinn, 6: 51-67, pls. 1-4 (In Russian with English summary). , Ktaamann, E. R. & Nestor, H. E. 1963. Features in common in the Ashgillian coral and stromatoporoid fauna of Estonia and Norway. Geoloogia Inst. Uurim., Tallinn, 13: 75-81 (In Russian with English summary). KIAER, J. 1897. Faunistische Uebersicht der Etage 5 des norwegischen Silursystems. Shr. VidenshkSelsk. Christiama (1) 3: iv + 76. 1899. Die Korallenfaunen der Etage 5 des norwegischen Silursystems. Palaeontographica, Stuttgart, 46: 1-60, pls. 1-7. 1903. Revision der mittelsilurischen Heliolitiden und neue Beitrage zur Stammesgeschichte derselben. Shr. VidenskSelsk. Christiania, 10: 1-58, 14 figs. 1930. Den fossilforende ordovicisk-siluriske lagrekke pa Stord og bemerkninger om de ovrige fossilfunn i Bergensfeltet. Bergens Mus. Aarb., 1929, 11: 1-75, pls. 1-5. 1932. The Coral Fauna of the Kalstad Limestone in Meldalen. Skr. norske VidenskAkad., Oslo (1) 1932, 4: 103-113, pls. 12-17. LamBe, L. M. rgor. A revision of the genera and species of Canadian Palaeozoic Corals. The Madreporaria Aporosa and the Madreporaria Rugosa. Cont. Can. Palaeont., Ottawa, 4, 2: 97-197, pls. 6-18. Lerirn, E. I. 1952. Schizocoralla from the Ordovician of Manitoba. J. Paleont., Tulsa, 26: 789-796, pls. 114-116. Linpstr6m, G. 1880. Fragmenta Silurica. 59 pp., 30 pls. Holmiae. 434 THE CORALS OF THE PORTRANE LIMESTONE Linpstr6m, G. 1899. Remarks on the Heliolitidae. WK. svenska VetenskAkad. Handi., Stockholm, 32: 1-140, pls. 1-12. M’Coy, F. 1850. On some new genera and species of Silurian Radiata in the Collection in the University of Cambridge. Ann. Mag. Nat. Hist., London (2) 6: 270-290. Nicuorson, H. A. & ETHERIDGE, R. 1878. A Monograph of the Silurian Fossils of the Girvan District in Ayrshire, 1. ix + 135 pp., 9 pls. Edinburgh & London. Reman, V. M. 1958. Neue Tetrakorallen des baltischen Oberordoviziums und des Llandovery. Geoloogia Inst. Uurim, Tallinn, 2: 33-47, pls. 1, 2. (In Russian with German summary). Roemer, F. 1861. Die fossile Fauna der silurischen Diluvial-Geschiebe von Sadevitz bei Oels in Nieder-Schlesien. xvi +- 82 pp., 8 pls. Breslau. 1883. Lethaea geognostica, 1. Lethaea palaeozoica, 1. 688 pp., 2 pls. Stuttgart. 1897. Lethaea geognostica, 1. Lethaea palaeozoica, 1, 3. vi + 688 pp., 2 pls. Stuttgart. SCHEFFEN, W. 1933. Die Zoantharia Rugosa des Silurs auf Ringerike im Oslogebiet. Shr. norske VidenskAkad., Oslo (1) 1932, 5: 1-64, pls. 1-11. SCHMIDT, FR. 1858. Untersuchungen tiber die Silurische Formation von Ehstland, Nord-Livland und Oesel. Arch. Naturk. Liw-, Ehst- u. Kurl., Dorpat (1) 2: 1-248. SoKxoLoy, B.S. 1951. Tabuljaty paleozoja Evropeiskoi chasti SSR. I. Ordovik Zapadnovo Urala 1 Pribaltiki. Tyvud. vsesoyuz. neft. geol. -vazy. Inst., Moscow-Leningrad (n.s.) 48: I-132, pls. 1-18. 1955. Labuljaty paleozoja Evropeiskoi chasti SSR. Vvedenije. Obchije voprosy siste- matiki i istorii razvitija tabuljat. Tvud. vsesoyuz. neft. geol. -vazy. Inst., Moscow-Leningrad (n.s.) 85: 1-328, pls. 1-90. 1962. Subclass Tabulata in Osnovy paleontologit; Gubki, arheociaty, kishechnopolostnye, chevvi: 192-265, pls. 1-18. Moscow. [Ed. B. S. Sokolov.] 1962a. Subclass Heliolitoidea in Osnovy paleontologii; Gubki, arheociaty, kishechnopolostnye, chervt.: 266-285, pls. 1-6. Moscow. [Ed. B. S. Sokolov.] 1962b. Biostratigraficheski 1 biogeograficheski obzor tabuljatomorfnyh korallov paleozoja SSSR. Geol. geofis., Novosibirsk, 10: 53-67. Wana, H.C. 1948. Notes on Some Rugose Corals in the Gray Collection from Girvan, Scotland. Geol. Mag., London, 85: 97-106, pl. 7. Witson, A. E. 1926. An Upper Ordovician fauna from the Rocky Mountains, British Columbia. Mus. Bull. Can. geol. Surv., Ottawa, 44: 1-34, pls. 1-8. YABE, H. 1915. Einige Bemerkungen iiber die Halysites-Arten. Sci. Rep. Tohoku Imp. Univ., Tokyo (2) 4: 25-38, pls. 5-9. Yu, CH. 1960, Pozdneordovikskije korally Kitaja. Acta palaeont. sin., Peking, 8, 2: 65-132, pls. I-15. FIGS. I, Fic. 6. INWe5 7. IG. 3: Fic. 9. FIG. ro. FTG re Ihe, 12. PLATE 1 Kenophyllum sp. : : é : ; Pp. 47 2. Two views of the corallum showing the axial structure. IR.45309. x 1.25. Kenophyllum cf. inflatum (Dybowski) . 3 . 0 [Da Ain? Corallum. R.45310. xX 2. ,5. Calice and corallum. KR.4531I. x 2. Streptelasma fragile \Vilson 9 3 . 5 Ds Alanis) Corallum showing a slit-shaped scar of fixation. KR.45312. x 2. Two small corals fixed on the bottom with curved sides of the apex. R.45313. xX 2. Streptelasma distinctum Wilson . : c > 196 Aas) Bottom of calice. R.45314. xX 1.3. Longitudinal section. R.45315. x 2. Streptelasma ci. rusticum (Billings) . ‘ P - p. 419 CalliGeseks7 52 lOmin or Grewingkia europaea (Roemer) . é F . pp. 420 Calicemss4 56 me = Grewingkia hibernica sp. noy. : : : . p. 420 Corallum with very wide proximal part and large scar of fixation. R.45318. x 1.2 Fics. 13, 14. Calice and cross-section of holotype. R.45319. x 1.2. Brachyelasma ci. duncani (Dybowski) 5 0 D420 Fic. 16. FIG. 17. Fic. 18. Corallum, R.45320. x 1.2 On Dalmanophyllum subduplicatum (MCoy) . 6 Paa22 Corallum with wide scar of fixation near the apex. .4532I. x Specimen showing the axial structure. K.45322. x 2. Calice with axial structure. KR.45323. x 2. to Bull. B. M. (N.H.) Geol. 10, 11 PRATER! PAV ee Sarcinula sp. . ; ; F : ae A22 Tics. 1, 2. Transverse and longitudinal sections of an etched corallum. R.45324. x 2. Fic. 3. Interior of an etched corallum. R.45325. x 2. Fic. 4. Fragment of a corallum showing horizontal rows of pores and laminated structure of intercorallite tissue. R.45326. x 2. Reuschia sp. . : . . - 0 Po 4125) Fic. 5. A small corallum formed by thick-walled cylindrical corallites. R.45327. x 2. Reuschia ? sp. . : : : : eps 424 Fic. 6. An irregular bushy corallum. R.45328. x 2. Catenipora wrighti sp. nov. . : ‘ 4277 ENG 7) ELOlOby pes N« 45920 nm a: Fics. 8-11. Surface views of small colonies. R.45330-33. x 2. Catenipora tapaensis (Sokolov) . 5 : DEA 25 Fic. 12. Surface view of an etched corallum showing the varying form and diameter of lacunae. R.45334. x 2. PLATE 2 Bull. B. M. (N.H.) Geol. 10, 11 PLATE 3 Coccoserididae gen. et sp. indet. - : : - p. 428 Fic. 1. Basal epitheca of a laminar corallum. R.45335. x 2. Fic.2. Thesamespecimen. Part of the upper surface showing Protavaea-like structure. x 8. Fic. 3. An etched laminar corallum. R.45336. x 2. Fic. 4. Part of the same specimen showing vertical sections of coenenchymal trabeculae. Figs. 5,6. Typical surface views of a laminar colony. R.45337. x 2 and x 8. Pragnellia cf. arborescens Leith . : - SP 430 Fies. 7, 8. Irregular cylindrical colonies. R.45338 and R.45339. x 2. Esthonia asterisca (Roemer) . ° . - p. 429 Fic. 9. Surface view of a thin lamellar corallum. R.45340. x 2. Fic. 10. Fragment of a lamellar corallum. R.45341. x 8. Bull. B. M. (N.H.) Geol. 10, 11 PAL ANI 3) PLATE 4 Wormsipora hirsuta (Lindstrém) . . - ED ASO Fics. 1, 2. Upper surface and interior view respectively of a small hemispherical corallum. R.45342. X 2. Fic. 3. Longitudinal section of corallites showing rugae-like rows of septal spinules. R.45343 x 2. Wormsipora portranensis sp. nov. . : . - p. 431 Fics. 4,5. Vertical section and upper surface of holotype. R.45344. x 2. Fics. 6, 7. Surface views of two cylindrical coralla showing the interrupted structure of the coenenchymal tissue and 12 rows of septal spinules. R.45345, R.45346. x 2. Fics. 8, 9. Two longitudinal sections of corallites showing massive septal spines directed upwards. R.45347. x 7 and x 4, respectively. Proheliolites dubius (Schmidt) : : > © p2432 Fics. to, 11. Upper surface and interior view respectively of an etched corallum. Vertical section shows horizontal tabulae in the corallites and coenenchymal tubes. R.45348. x 2. Fic. 12. Thesame specimen. Between the corallites rare polygonal coenenchymal tubes can be observed. x 8. Bull. B. M. (N.H.) Geol. 10, 11 PLATE 4 LON (: e {7 oec1965 } z s) Lig, WSS INDEX*TO: VOLUME X New taxonomic names and the page numbers of the principal references are printed in Bold type. An asterisk (*) indicates a figure. Acanthoceras 342 Acanthocythere 10 faveolata to Acantholithus asteviscus 429 Acidolitinae 429 Acinonyx 302 Actinozoa 96 Afrocyon 291 Ailurictis 303 Alatacythere 51 phylloptera 51, 52, 76; Pl. 2, figs. 17, 19 vobusta 51, 52, 75, 76 ; Pl. 2, figs. 7-16, 18 Allicospermum 129 patagonicum 121, 129, 130*, 131; Pl. 1, figs. 7-9 ; Pl. 5, fig. 28 vetimirum 130, 131 xistum 131 Alopecodon 291 Ammonceratites convadi 371% Ammonites coesfieldensis 391, 392 cooperi 380 costulosus 391, 393, diphyllides 388 mamillaries 345 marroti 391 mitis 360 SACYA 357 styiatocostatus 321, 392 vart, 391 var. mayroti 391 Amphicynodon 291 Amphicyon 242, 289, 291, 223-295, 310, 311 americanus 293 dehmi 293, 295 lemanensis 293 MAJOY 293, 295 sinapius 293 (Ictiocyon) dehmi 295 Amphicynoninae 241 288 289, 291 Amphicyonodontinae 291 Amphicytherura 49 cholodon 49, 75, 76; Pl. 1, figs. 6, 7 icenica 49 Amplexopora 111 Amplexoporidae 111 Anagaudryceras 337, 338, 357-360 aureum 358 buddha 358 coalingense 358 crenatum 358 tnvolutum 358 limatum 358 madraspatanum 358 mikobokense 337, 343, 358, 359*, 360, 401, 402 ; Pl. 4, figs. 1-3 multiplexus 358 povitissimum 358 vevelatum 358 subsacya 358 subtililineatum 358 utaturense 358 yamashitai 358 Anaklinoceras 374 Anasinopa 242, 253, 259, 262, 264, 310, 311 leakeyi 241, 259, 260, 261*, 262, 263, 309 ; IPA, i, tiles, O; WB IB 2 Ancylopoda 167 Ancylotherium 166, 169, 218, 219, 223, 226, 227*, PNR, PIO} Ey PEN gaudryi 216 hennigi 165, 166, 226, 227*, 228*, 229*, 230, Pe, DEVI DEA Py.) pentelicum 165, 166, 167, 218,223-226, 227*, DANS PIX} 5 PIBNO), GB, PVA, PEI A GLY Angola 337-412 Anisodon larteti 168 magnum 168, 193 schinzit 171% Antsodon (Choelichotherium) 168 Anoplotherium 168 grande 168 magnum 168 Apterodon 263, 267, 278 Avaucaria mirabilis 121 Archaelurus 302, 303 Archaeopteris 86, 88, 89 jacksoni 83, 86, 87, 88, 90, 91 ; Pl. 1, figs. 1-6, @) @ Jel, 13 latifolia 88-90 ; Pl. 2, figs. 3-8 macilenta 88, 89, 90 ARCHANGELSKY, S. 119-137 Arctamphicyon 293 arctocyonid 310 Arctocyonidae 241, 242, 243, 311 Arctocyonides 244 Arctocyoninae 243 Arctocyonoidae 243 Arctocyonoidea 241, 243 Ardynictis 246 ARNOLD, C. A. 141, 142 A sciocythere 15 acuminata 15, 16 ; Pl. 2, figs. 10-12 ; Pl. 3, figs. I-10 lacunosa 16, 23, 24 Ashgillian 416 Atactotoechus 112 436 INDEX Aulacocythere 10 punctata 10 reticulata 10 Aulopora 424 Austrosphenodiscus 395 Axonoceras 343, 405 Baculites 337, 339, 343, 362-305, 367*, 360%, 379, 402, 405 anceps 337, 339, 363, 364-366, 367*, 368, 369%, 3700 el atic. ae) lies ties 4S ieleOy figs. 1-5 var. pacificus 368 sublaevis 366 valognensis 363 aspey 370 dissimilis 368 fairbanski 370 fawasii 363 fuchsi 370 leopoliensis 370 palestinensis 370 Sp. 402 subanceps 337, 339, 367*, 368, 369*, 370, 405 ; TAL, By, titey, 3) 5 IRIE Oy ike, ©, 77 3 elle 7, ithe subanceps pacificus 370, 405 subanceps subanceps 370 valognensis 363 vertebralis 362, 363, 366, 368 Baculitidae 337, 362 Baiera 125 australis 125 Bairdia 8, 46 angusta 46, 75 fullonica 8 harvisiana 75 hilda 8 stliqua 43, 75 Bairdiacea 8, 42 Bairdiiadae 8 Baluchistan 173 Barinophyton 83, 84, 85*, 86 citrulliforme 83, 86 dawsoni 86 obrutschevii 85 obscurum 86 vichavdsoni 83, 84, 85, 86, 91 ; Pl. 1, figs. 7, 8, rie 3 IPA, Ap Sikes, Bate, R. H. 1-33 Batostomella 102 Batostomellidae 102 Bdeogale 301 Benguela basin 341-342 Bennettitales 121 Benthall Edge 96 Bhorophaginae 291 Biharisporites 88, 90 ellesmevensis 88 submamillaris 88 Borhyaenidae 253 Borissiakia 165, 225 betpakdalensis 226 Bostrychocevas 371-374 boulei 373 colubriformis 373 condamyt 373 elongatum 373 indicum 373 otsukat 373 polyplocum 371-373, 375, 395, 401, 404 saundersovum 373 SECOENSE 373 Brachycytheve 50 laticvistata 50, 77 ; Pl. 2, figs. 1-6 sphenordes 50, 76 ; Pl. 2, fig. 5 Brachycytheridae 50 Brachyelasma 421 duncani 415, 421 undulatum 421 Brachyphyllum 121 Brancoceratidae 337, 346 Brancoceratinae 337, 346, 347 Bryozoa 97 ButTLeER, P. M. 163-237 Bythocypris 46 vyeussiana 46, 75 stlicula 46 Bythocythere aliena 9 Bythocytheridae 9, 52 Calamospora 84. Calamostachys 84 Calapoecia 423 Callopora 109 aspera 106 elegantula 109, 110 fletchevi 110 nana 109, 110 Campylocynodon 29% Canidae 241, 288 Caninae 291 Canis 291 Canoidea 241, 288 Carcinodon 244 Carnivora 241, 243 Catentpova 424, 425, 427 Ruruensis 425 labyrinthica 425 obliqua 428 parallela 428 piirsaluensis 425 tapaensis 415, 416, 425, 426*, 427, 428; Pl. Zrii az, tvactabilis 428 wrighti 415, 426*, 427, 428 ; Pl. 2. figs. 7-11 Cateniporinae 424 Centrocythere 47 denticulata 47, 75 ; Pl. 1, figs. 8, 12, 13 Cephalogale 291 Ceramopora 99 imbricata 99 sp. 100); Ply i figs. 354 Ceramoporidae 99 Chalicothere 226 INDEX Chalicotheres 165, 166 Chalicotheridae 226 Chalicotheriinae 166 Chalicotherium 165-167, 171, 173, 180, 184, 186*, 188, 196, 203*, 206, 207*, 213, 219, 222, 224, 225 anisodon 168, 169 antiquum 167, 168 baltavarense 193 baltavarensis 167, 168 brevivostris 165, 166, 172, 184, 193, 194, 195 goldfussi 165, 166, 167, 168, 169, 172, 175*, 180, 181, 183, 184, 187*, 188, 193, 203* gvande 165-167, 168, 169-173, 175*, 180, 181, 183-185, 186*, 187*, 188, 191, 194*, 195, 196, 197*, 198, 199*, 200, 201, 202*, 203%, 204,205*, 206, 207*, 208, 210*, 211, 212*, 214, 216-220, 222, 223 grande rhodanicum 165-167, 170, 183, 185 magnum 169 var. secundarium 170 modicum 170, 224 pilgrimi 165, 167, 179, 175*, 180, 181, 186*, 187, 188, 194, 205*, 206, 212, 223 rusingense 165, 166, 173, 174, 175*, 176-178, 179*, 180-182, 183*, 184, 185, 186*, 187*, 188, 189*, 190, I9I, 192*, 193, 194*, 195*, 196. TOW, UK), UCLh, Alo, Aoi, AOA, HOt", Zonk, ZOD Pe ZOOM207~ 208,200) 2tOn, 2im 22, Bip 2TA 20S S200, lin a2 On 220), 222), 223 salinum 165, 166, 171, 172, 180, 181, 186*, 223 sp. 165, 170 styviacum 165, 170 wetzlert 165, 166, 170, 171, 173, 185, 187, 188, IQI, 196, 212, 223 Chalicotherium (Anisodon) 168 Chalicothevium (= Macrotherium) 165, 166 CHANDLER, M. E. J. 139-161 Choelichotherium 168 Chriacinae 243 Cirroceras 371, 372 depressum 373 Cladophlebis 133 ; Pl. 2, fig. 16 Coahuilites 395, 397, 403 Coccoserididae 415, 428 gen. et sp. indet. 428, 429 ; Pl. 3, figs. 1-6 Colpoclaenus 244 Constellariidae 107 Creodonta 241, 243, 246, 257 creodonts 242 Cryptostomata 96, 98, 113 Ctenostomata 96, 97 Cuanza basin 341, 342 Cyclogranisporites 88, 89, 90 lasius 88 Cyclostomata 96, 98 Cynodesmus 291 Cynodictis 291% Cynodon 291 Cynohyaenodon 258 Cypridacea 9, 40 437 Cypridina 64 macrophthalma 64, 77 Cytheracea 9, 45 Cythere 43, 45-65, 68-70, 75-77 acanthoptera 52, 76 acutiloba 70,77 auriculata 6% bairdiana 37,49, 75 ; Pl. 1, fig. 9 chelodon 49 concentrica 48 echinulata 65 filicostata 43, 64, 77 gaultina 68 gaultina excavata 69, 76 harvisiana O01, 75 harrisiana veticosa 61, 75 harrisiana setosa 61, 75 hilseana 45, 75 intervupta 61 Roninckiana 61, 76 laticristata 50, 77 lineatopunctata 61 ; Pl. 6, fig. 7 longispina 53, 76 lonsdaleiana 63, 67 macrophthalma 49 ornatissima 64 pedata 54 pedata laevis 55 phylloptera 5% punctatula 47, 48, 75 quadrilaterata 61, 63 siliqua 43 slavantensis 48, 76 sphenoides 50 spinifera 45, 46, 76 subtuberculata 69, 76 triplicata 58 umbonata 52, 53, 75 umbonata acanthoptera 57 williamsoniana 75 var. gyanulosa 75 Cythereis 49, 51, 58, 62-68, 70, 75-77 acutiloba 70 alata 51, 75 auriculata 61, 75 bonnemai 66, 67 ciliata 65-68, 75 cornueli 63 cornuta 65-67, 75 corrigenda 62, 77 ; Pl. 7, figs. 6, 9 excavata 68, 69, 76 filicostata 63 folkstonensis 63, 66, 75 ; Pl. 7, figs. 1-5 gaultina 68, 75 glabrella 63, 66 hirsuta 68 icenica 49, 76 interrvupta 61, 75 longaeva longaeva 65, 66; Pl. 7, fig. 12; Pl. 8, figs, 3, 5 lonsdaleiana 63, 67, 75,77 ; Pl. 7, figs. 7, 10 lurmannae 63, 66, 67, 76 ; Pl. 8, figs. 11-15 438 INDEX Cythereis—cont. perforata 46 macrophthalma 49, 64, 75, 77; Pl. 6, figs. sevvata 74. 12-15, 17 umbonata 56 nuda 67, 75, 76; Pl. 7, figs. 11, 13, 16 Cytheroptera 56 ovnatissima 64, 65, 66, 68, 75 ; PI. 8, figs. 1, 2, umbonata 56 4, 6 Cytheropterina 24 altinodosa 66 plana 24, 25 ; Pl. 8, figs. 7-10 ; Pl. 9, figs. 1-4 nuda 65-67, 76 triebeli 25 paupera 66, 76 Cytheropteron 25, 47, 50, 51, 53-55 vadiata 66, 76 alatum cornuta 51, 76 reticulata 67, 76 fortis 51, 76 stricta 66 vobustum 51, 76 Phylloptera 5% concentyicum 47, 75 quadrilatevata 61, 63, 64, 75 cuspidatum 76 veticulata 63, 64, 66, 67, 68, 76; Pl. 8, figs montuosa 53, 76 16-19 tvicuspidata 56, 76 vudispinata 59, 62, 76, 77 hibernicum 51, 76 semiplicata 49, 64 pedatum 54, 76 spinicaudata 70, 76 salebrosa 54, 76 stvicta 67 phyllopterum 51, 76 thorenensis 66, 68, 75 ; Pl. 7, figs. 14, 15, 17 punctatula var. virginea 47 triplicata 58, 75 purum 25 lineata 58, 76 sherborni 55, 76 tuberosa 76 sphenoides 50, 76 var. symmetrica 76 umbonatum 56 vallata 76 umbonatum acanthoptera 76 wrightit 67, 75, 76; Pl. 7, fig. 8 umbonatum longispinata 56, 76 aculeata 76 Cytherura 54 Cytherella 70, 71 spoonert 54 chapmani 70, 76 Cytheruridea 24, 49 oblinquivugata 76 williamsoniana 74 Dalmanophyllum 422 bosqueti 71, 72, 77 subduplicatum 415,422; Pl. 1, figs. 16-18 granulosa 71 Daphoenodon 291 stricta 74 Daphoenus 289, 291 Cytherellidae 8, 70, 75 Daradiceras 395 Cytherelloidea 8, 70, 75 Dasuryodon 267 catenulata 8 Dechenella 319, 326, 332 chapmani 70, 71, 76 ; Pl. 9, figs. 15-19, 22 gigoutt 331-333 globosa 71 ; Pl. 9, figs. 7, 9, 10 vittbergensis 323, 324, 331-333 granulosa 71, 72, 75 ; Pl. 9, figs. 24-26 characters 324 hindei 72, 73 ; Pl. 9, figs. 4, 8, 11 trend 332 knaptonensis 71, 73, 77 ; Pl. 9, figs. 20, 21 setosa 319, 320, 321*, 322, 323, 324*, 325*, 326, oblinquirugata 73, 76 ; Pl. 9, figs. 12, 13, 14 331-333 parawilliamsoni 7% styuvei 331 parawilliamsoniana 73, 77; Pl. 9, fig. 23 verneuilt 323, 324, 331-333 stricta 71,74, 75 ; Pl. 9, figs. 1-3, 5, 6 characters 324 wilwamsoniana 71-73, 75 trend 332 Cytheridea 9, 45, 46 Dechenella (Dechenella) setosa 319, 326—333 ; Pl. bosquetiana 46 I, figs. 1-15 jonesiana 45, 46 Dechenella (Eudechenella) setosa 326 perforata 45, 75 Dechenellina 326 insignis 45, 46, 76 Deinotherium 310 votundata 45, 46, 76 . cuviert 310 subtrigona 22 hobleyi 310 vulsa 9 Delawarella subdelawarensis 402 Cytherideidae 14, 45 Deltatherium 245 Cytherideinae 14 Desmoceras Cytherina 46, 47, 56 latidorvsatum 341, 343 ciliata 64 var. inflata 341, 343 concentrica 47 Desmoceratidae 338, 386 ornatissima 64, 77 Desmoceratinae 338, 388 INDEX 439 Desmophyllites 338, 339, 388 diphylloides 338, 388, 401, 402 ; Pl. 11, fig. 3 var. besairier 388 inermis 388 lata 388 Diastoporidae 98 Didymoceras 337, 338, 343, 371-374, 377, 380, 382, 384, 402, 404, 405 angolaense 337, 378, 381, 402 ; Pl. 8, fig. 2 californicum 337, 373, 376, 377, 405; Pl. 8, fig. I elongatum 376 fresnoense 380 hornbyense 337, 371, 372, 377, 383, 402, 405; Pl. 8, fig. 4 kernense 381 NAVAYVOENSE 373 newtont 371 polyplocum 375-377 schloenbacht 373 SeCOENSE 377 sp. 378 stevensont, 373 subtuberculatum 337, 338, 371, 373, 374, 375, 376, 401, 402 ; Pl. 7, figs. 2-6 Didymoconus 246 Dinaelurus 303 Dinailurictis 303 Dinictis 303 Dinocyon 291 Diplomoceras 386 Diplomoceratidae 338, 385 Dipoloceras 347 Dissopsalis 242, 258, 262, 264, 265, 310-312 carnifex 265, 267, 310 pyroclasticus 241, 265, 266, 267, 309; Pl. 3 vubey 265 Dolocythere 14 maculosa 14 Dolocytheridea 20, 46 bosquetiana 20, 46, 47, 75, 76; Pl. 1, figs. 18, 19, 20 hilseana 46 intermedia 20 Douvilleicevas 337, 339, 342-344, 400 alternans 345 benonae 345 inaequinodum 345 magnodosum 345 mammillatum 344, 345 var. aequinodum 337, 343, 344, 345, 400; Pl. 1, figs. 1-4 orbignyi 337, 345, 400 ; Pl. 1, fig. 5 Douvilleiceratidae 337, 343 Drymatophora problematica 96 Ektyphocythere 26 triangula 26, 27 Elobiceras 342, 400 Emperocevas 371, 380 simplicicostatum 371 Enhydra 257 Entalophora 99 Eocatenipora 428 Eocytheridea 16-22 carinata 18, 19; Pl. 4, figs. 6-11; Pl. 5, figs. 1-8 elongata 19 faveolata 20, 21 ; Pl. 6, figs. 4-9 lacunosa 21 reticulata 2152275 Rl) 6; hess to; ors vel, figs. I-5 Eocytheridea? 16-20 acuta 16, 17, 20; Pl. 3, figs. 11-14 astricta 17, 18, 20; Pl. 4, figs. 1-5 erugata 19, 20; Pl. 5, figs. 9-12; Pl. 6, figs. I-3 Eofletcheria 424 Eofletcheriinae 423 Eomoropus 165, 196 Epiphylloceras 357 SUTYA 357 Eridotrypa 98, 102, 103, 104, 111, 112 cava 104 ; Pl. 2, figs. 5, 6 cylindrica 103 ; Pl. 2, figs. 3, 4 echinata 112 similis 103 Spy lod os.) bls 3) figs, 1, 2 styiata 110 umbonensis 105 Esthonia 429 astevisca 415, 429, 430 ; Pl. 3, figs. 9, Io asteviscus 429 lamellosa 429 Eubaculites 366 Eucythere tyvigonalis 75 Euhyphantoceras maestrichtiense 376 Eulophoceras 396, 397 Eupachydiscus 338, 339, 389, 390 havadai 390 launayt 390 pseudogrossouvret 338, 389, 390, 401 ; Pl. 12, figs. I, 4 var. undulatocostata 390 sp. 390 Eutrephoceras 338, 399 egitoense 399 simile 338, 399, 401 Exiteloceras bennisoni 380 Favositella 100 interpuncta 100 Favosites interpunctus 100 Felidae 241, 302 Felinae 302 Felis 304 leiodon 303, 304 Feloidea 241, 295 Fissipeda 241, 288 Fistulipora 101, 102 crassa 98, 101, 102 ; Pl. 1, fig. 5; Pl. 2, figs. 1 dobunica 102 lockportensis 102 440 INDEX Fistulipova—contd. ludensis 102 Fuhrbergiella 10, 11 avens 10 favosa to minima 11 ; PI. 1, figs. 1-8 Galethylax 258 Gaudryceras 337, 338, 357, 360 aenigma 343, 358, 360 alamedense 360 analabense 362 beantalyense 362 cinctum 362 delvallense 362 demanense 362 densiplicatum 362 devallense 360 filicinctum 360 lauteli 362 mite 361, 362 navarrense 360-362 politissimum 358 propemite 362 sachalinense 360 sp. 361 striatum 362 var. pictum 362 tenuilivatum 362 vavagurense 337, 343, 301, 362, 401, 402; Pl. File, nike, Gy 1edl, Gy aateey, a6, variocostatum 362 vascogoticum 360 Gaudryceras (Neogaudryceras) pictum 361 Gaudryceratinae 337, 357 Geiselotherium 258, 262 Ginkgo 136 biloba 124, 134, 135 complex 124 huttoni 124 Ginkgoales 121, 122 Ginkgoites 121, 122, 124, 125, 131, 134 acosima 124 hermelini 125 longifolius 124 marginatus 125 obrutschewi 125 sibivica 125 taeniata 124, 131 ticoensis 121, 122, 123*, 124, 125, 131; Pl. 1, figs. 5,6; Pl. 3, figs. 19-21 ; Pl. 4 fig. 27 tigrensis 121, 125, 126*, 127*, 128, 132*, 133, resis Meh, we ae, art Sell, 3, tare veal PIL, 3}. fig. 22 ; Pl. 4, figs. 23-26 Glyptoxoceras 386 Gomphotherium angustidens 310 Grangervia 213, 218 Grewinghia 420, 421 buceros 420 europaea 415, 416,420 ; Pl. 1, fig. 11 hibernica 415,420 ; Pl. 1, figs. 12-14 Hallopora 109, 111 Hallopora—cont. elegantula 98, 109, 110 ; Pl. 4, figs. 3, 4 vamulosa 110 striata 98,110, 111 ; Pl. 4, figs. 5, 6 Halloporidae 109 Halysites catenularia 425, 426 escharoides 425 Halysitida 424, 425 Halysitidae 415, 424 Hamites vancouverensis 378, 380 Hamitidae 342, 400 Haplocytheridea jonesiana 45 Harpaleocyon 291 Hauericeras 389 Hauericeratinae 338, 389 Hecubides 242, 289, 291, 293, 295, 310, 311 americanus 289, 293, 295, 310 euryodon 241, 289, 290, 291, 292*, 293-295, 309 ; Pl. 5, fig. x lemanensis 289, 293, 295, 310 macrodon 241, 289, 293, 294*, 295, 309 ; IDE 5, wiles 2 Heliolites 429, 430 432 astevicus 429, 430 dubia 432 dubius 432 hirsutus 430 inordinata 429, 430 interstincta 429 intricatus var. lamellosa 429 parvistellus 429 Heliolitidae 415, 430 Heliolitoidea 415, 428, 430, 432, 433 Hemicyon 291 Hemigalinae 296 Hemipsalodon 267 Herpestes 296 Herpestinae 241, 296, 301, 311 Hesperocyon 291 Hetevoceras convadi 378, 380 hornbyense 377 Hetevopora crassa 101 Hewitson, W. 141, 142 Homocythere reticulata 61, 62 Homocytheridea 30 Hoplites 391-393 plasticus 391 (Hoplitoplacenticeras) plasticum 393 Vayt 392 var. marroti 392 Hoplitoplacenticervas 338, 339, 391, 401, 402 coesfieldensis 393 costulosum 338, 393, 401 ; Pl. 13, fig. 2 dolbergense 392 lafresnayanum 391, 393, 401 lemfordense 393 marroti 338, 391, 392, 401 ; Pl. 2, fig. 3; Pl. 13, fig. 3 var. Vari 392 plasticum 391, 393 plasticum costatum 392, 393 laeve 393 INDEX 441 Hoplitoplacenticeras—cont. Hysteroceras—cont. praematura 392 varicosum 347, 350 spp. 338, 393, 394, 401 var. angolana 347, 350 vancouverense 391, 393 vari 391, 392, 401 Ictytocyon 293, 295 Howartu, M. K. 335-412 Imperatoria 258, 259 Hyaenaelurus 310 gallwitzi 258, 259 Hyaenodon 242, 263, 267, 270, 277, 278, 279, hagent 258 280, 286, 310, 311 Indoceras 395, 396 aimi 278, 280 Ischnognathus 258, 262 ambiguus 278, 280 Isocytheveis 66 andrewst 278, 280, 281, 283, 309, 310 fortinodis 66 aymardi 278, 280 Isohyaenodon 280, 281, 283, 284, 286 bavaricus 278, 280 brachycephalus 278, 280, 286, 287 Kayo, D. 413-434 brachyrhynchus 278, 280 Karkenia 121, 132-136 Cayluxi 278, 280 incurva 130*, 131*, 132*, 133-135; Pl. 1, compressus 278, 280 fig. 10 ; Pl. 2, figs. 11-18 ; Pl. 5, figs. 29-32 crucians 278, 280 Kaye, P. 35-79 cruentus 278, 280 Kelba 242, 244, 246, 310, 311 dubius 278, 280 quadeemae 241, 244*, 245, 246, 309; PI. 1, eminus 278, 280 fig. I exiguus 278, 280 Kenophyllum 417 filholi 278, 280, 284 canaliferum 417 gevvaist 278, 280 inflatum 415,417, 418 ; Pl. 1, figs. 3-5 herberti 278, 280 silivicum 418 horridus 278-280 sp. 415, 417 ; Pl. 1, figs. 1, 2 lauvillardi 278, 280 subcylindricum 417 leptocephalus 278, 280 Kenya 165, 173, 242, 309 leptorhynchus 278, 280 National Museum 174 martini 278, 280 Kiaerophyllum anguineum 420 matthewi 278, 280, 309 Kichechia 242, 296, 301, 310, 311 milloquensis 278, 280 zamanae 241, 296, 297, 298, 299*, 300*, 301, minor 278, 280, 281 302, 309; Pl. 5, fig. 3 minutus 278, 280 Kinkelinella bajociana 28 montanus 278, 280 tenuicostati 28 mustelinus 278-280, 284 Kirtonella 25, 26 parisiensis 278, 280 plicata 25, 26 paucidens 278, 280 reticulata 25, 26 ; Pl. 9, figs. 5-16; Pl. ro, pervagus 278, 280 figs. I-2 pilgrimi 278, 280, 309 Kirtonellinae 25 vequient 278, 280 Kitchinites 338, 339, 386, 387 vetus 278, 280 angolaense 401, 402 vulpinus 278, 280 angolaensis 338, 339, 386, 387; Pl. 11, yuanchensis 278, 280 figs. 4-6 Hyaenodon (Isohyaenodon) 287 brevicostata 387 andrewsi 241, 280, 281, 282*, 283, 284, 287 ; busnardot 387 Pl. 4, fig. 6 darwint 387 matthewi 241, 280, 283, 284*, 287 enayt 387 pilgrimi 241, 284, 285*, 286*, 287 fascigerus 387 Hyaenodontidae 241, 242, 246, 247, 257, 267 flabelliformis 387 Hyaenodontinae 241, 246, 258, 267, 277 pondicherryanus 387 Hyphantoceras 376 quadratus 387 buttense 376 sp. 387 cevatopse 376 KLAAMANN, E. 413-434 laqueum 376 vyeussianum 376 Leakitherium 242, 267, 276, 278 venustum 376 hiwe$i 241, 276, 277*, 278, 309 ; Pl. 4, figs. 4, Hypophylloceras 357 5 Hysteroceras 342, 347, 348, 400 Ledoceras 342 binum 348 Leioclema 97, 98, 103, 105, 106, 107 orbignyt 347, 348 asperum 98, 106, 107 ; Pl. 3, figs. 5, 6 442 Leioclema—cont. densiporum 105, 106 ; Pl. 3, figs. 3, 4 gvanatus 97 pulchellus 97 ramosum 107 ; Pl. 3, figs. 7, 8 Leiotriletes atavus 84. nigvatus 84 Leonhardtina 258 Leptocyon 291 Leptocythere? 14 Leptopteris 141-143, 146-149, 151, 157, 159 superba 146 Libycoceras 343, 370, 394-396, 398, 405 acuto-dorsatum 394, 395 angolaense 397, 398 chargense 394, 395 Lichenariida 423 Limnocyonidae 246, 258 Limnocyoninae 246, 258 Lindstrémia subduplicata 422 Lioclema vamulosum 106, 107 Lophiodon goldfussi 167 Lophodentina 14 Loxolophus 245, 246 Lyoporidae 415, 423 Lytocevas (Gaudryceras) aureum 358 vavagurense 301 Machaeroidinae 246, 258 Machairodontinae 302 Machairodus parvulus 304 Macrocyprididae 42 Macrocypris 42 avcuala 43 concinna 75 equisita 42, 43, 76; Pl. 4, figs. 12, 16 muensteriana 43, 75 ; Pl. 4, figs. 9, 10 siliqua 43, 44, 75 ; Pl. 4, figs. 11, 14, 15, 18 simplex 44,75; Pl. 4, fig. 13 wrightit 44, 45,75; Pl. 4, fig. 17 Macrodentina sp. 49, 75 Macrotherium 166, 168, 172, 219 brevivostvis 172 giganteum 167-169 gvande 166, 169 var. vyhodanicum 170 magnum 169 oggenhausense 169 salinum 171 sansaniense 168, 169 sp. 172 wetzlevi 17% Macrotherium (2) 192 Mammites 342 Mammocyon 291 Manambolites 338, 339, 395, 396-398, 403, 404 dandensis 338, 339, 396, 397, 398*, 402; Vell ai, if, 3 1eNL, ang) ie pervinquieri 397 piveleaut 397 YICENSIS 397, 404 spathi 397 INDEX Manis gigantea 167 sindiense 172 sindiensis 171% Mantelliceras 342 Maorites 389 Marsilea 88 Marsileaceae 85 Maw, G. 96 Menuites 343, 405 Mesocyon 291 Mesopuzosia 387 pacifica 387 Metachriacus 245 Metailurus 242, 302, 303, 304, 307, 311, 312 africanus 241, 304, 305, 306*, 307-310; Pl. 5, fig. 4 MAJOY 303-305, 307, 308 MINOY 303-305, 307, 308 mongoliensis 303-305, 307, 308 parvulus 303-305, 307, 308 tunggurensis 304 Metapterodon 242, 267, 268, 270, 272, 310 biinciswwus 271 Raiservi 241, 268, 269*, 271, 309 ; Pl. 4, fig. x zadoki 241, 269*, 270, 271, 309 ; Pl. 4, fig. 2 Metaschizotherium 105, 169, 184, 223, 225 bavaricum 184, 224 fraasi 223-225 hennigi 226 tvansvaalensis 226, 234 Metasinopa 242, 258, 262-264, 310, 311 ethiopica 264 fraasi 263, 264 napaki 241, 263, 264*, 265, 309 sp. 264 Miacidae 301 Micropneumatocythere 11 convexa 11 globosa 12 ; Pl. 1, figs. 9-20 Mitoclema 98 regularis 99 ; Pl. 1, figs. 1, 2 Mitoclemella 99 Mocamedes basin 341, 342 Mongolia 172 Monoceratina 9, 52, 53-55 acanthoptera 52, 76 ; Pl. 3, fig. 2 bonnemai 52, 53, 76 ; Pl. 3, figs. 5, 6 laevoides 55, 77 longispina 52, 53, 77 ; Pl. 3, fig. 1 montuosa 53, 54, 56, 76; Pl. 3, fig. 3 pedata laevoides 55, 77 ; Pl. 3, fig. 17 pedata 54, 55-57, 70; Pl. 3, figs. 9-14 salebrosa 54, 55, 76; Pl. 3, figs. 15, 16 sherborni 55, 76 ; Pl. 3, fig. 4 tricuspidata 54, 56, 76; Pl. 3, figs. 7, 8 umbonata 54, 56, 57, 75, 76; Pl. 4, figs. 3, 4, 6-8 umbonatoides 55,57, 76 ; Pl. 4, figs. 2, 5 vulsa 9 Monotrypa 111, 112, 113 discoidea 112 flabellata 98, 112, 113 ; Pl. 6, figs. 1, 2 INDEX 443 Monotrypella 111, 112 benthallensis 111, 112 ; Pl. 5, figs. 1, 2 Monticulipora 112 sp. 96 Monticuliporidae 96 Moropus 168, 169, 198, 200, 213-215, 219, 224, 225 betpakdalensis 225 elatus 198, 201 matthewt 167 merviamt 167 Mortoniceras 342, 347, 400 Mortoniceras (Durnovarites) 347, 350, 352 crassicornutum 347 Mortoniceratinae 347 Mustela putorius 285, 280 Mzezzemceras 395, 397 Nautilidae 338, 399 Nautilus blanfordianus 399 Neocythere 47, 48 denticulata 47, 75 ; Pl. 1, figs. 8, 12, 13 vanveent 48, 75 ; Pl. 1, fig. 10 virginea 48, 75, 76; Pl. 1, figs. 11, 14-17 Neokentrocervas 337, 338, 342, 346, 347-349, 353, 354, 400 choffati 351, 352 var. cvassinoda 351, 352 costatum 351, 352 var. tenuis 353, 354 crassicostatum 337, 347, 355, 356, 400 ; Pl. 2, fig. 16; Pl. 3, figs. 12-15 curvicornu 337, 340, 347, 348, 349, 359, 354, 400 ; Pl. 2, figs. 1-9 crassicorvnutum 347, 350 curvicornu 348 magnum 351, 352 pseudovaricosum 337, 347, 348, 353, 354-356, 400 ; Pl. 3, figs. 5-11 var. compressa 353 gracilis 354 singulave 337, 349, 350, 351, 354, 400; Pl. 2, figs. 10-15 SP. 355 speciosum 348, 350 var. vudis 348 subtuberculatum 337, 351, 352, 353, 400 ; Pl. 3, fig. I tectovius 348 trituberculatum 337, 346, 350, 352, 400 ; Pl. 3, figs. 2-4 Neophyllocevas 337, 339, 35, 357, 404 hetonaiense 357 lambertense 357 vamosum 357 ultimum 337, 356, 357, 402 Neopuzosia 387 ishikawai 387 japonica 387 matsumotoi 387 Nestoritherium 165 simense 181 Nestoritherium—cont. sivalense 193 Nicholsonella 107-109 florida 108 parva 108, 109 ; Pl. 4, figs. 1, 2 vinguebergi 109 Nicholsonia megastoma 430 Nimraivides 303 Nimravinae 241, 302, 303 Nimravus 302, 303 Nostoceras 338, 339, 342, 343, 371-374, 377, 378; 379*, 380-382, 384, 402, 404, 405 angolaense 378, 383 avaconts 371, 381, 384 helicinum 338, 377, 383, 402, 403 ; Pl. 8, figs. 3,5 var. cvassum 383 humile 383, 384 hornbyense 377 hyatti 338, 371, 378, 379*, 380, 382, 383, 402, Mos) = Il, G) A IZAl; We), wiles, a kernense 338, 381, 402 ; Pl. 8, fig. 6 mariateresianum 383 mexicanum 381 natalense 384 obtusum 338, 339, 379*, 384, 402 ; Pl. ro, fig. 2 pauper 374 rotundum 338, 339, 381, 382, 383, 402; Pl. 10, fig. 3 schloenbacht 374 sp. 378 stantoni 374, 383 sternbergi 380, 381 subangulatum 374 Nostoceratidae 337, 371 Nothocyon 291 Octocyoninae 291 Oiophyllites 338, 339, 389 angolaense 401 angolensis 338, 389, 402 decipiens 389, 402 Olduvai 226, 232, 233 Opsiclaenodon 243 Orthonotacythere voigteiensis 30 Osmunda 141-143, 146-149, 151, 156-161 banksiaefolia 158, 159 bromeliaefolia 158, 159 japonica 147 javanicum 158, 159 lancea 140, 147, 159 lignitum 141, 158-160 vegalis 147, 148 vachellit 159 zeylanica 156* Osmunda (Plenasium) 160 banksiaefolia 146-148 bromeliaefolia 148 dowkeri 142, 158, 160 javanicum 148 lignitum 158 vachellii 148 444 INDEX Osmundaceae 141, 142 Osmundastrum 142, 143, 146-149, 159 cinnamomea 159 claytoniana 159 Osmundites 141-161 chandlevi 142, 157, 158, 160 dowkeri 141-161 ; Pls. 1-12, 24 figs. Owen, D. E. 93-117 Oxyaenidae 246, 247, 253, 257 Oxyaenoid creodonts 242 Oxyaenoidea 241, 246, 257 Oxyclaeninae 241, 243, 244 Pachycynodon 291% Pachydictya 114 crassa 98, 115; Pl. 5, figs. 6-8 dichotoma 115 holmi 115 Pachydiscidae 338, 389 Pachydiscus grossouvret 389 neubergicus 404 Paciceras 394, 396 Palaeohalysites kuvuensis 425 pursaluensis 425 tapaensis 425 Palaeothevium 171 cvassum 171 magnum 171 schinzit 170, 171, 194* Pangolin gigantesque 167 Pavacynodon 291 Paracynohyaenodon 258, 202 Paracyprididae 9 Paracypris 9 bajociana 9 Paradoxurinae 296 Paralenticeras 397 Parapachydiscus besairiei 389 Parictis 291 Paroxyclaenus 243 Perissodactyls 188 Petraia inflata 418 subduplicata 422 Pettitt, J. M. 81-92 Phoberocyon 291 Pholadomya beds 400 Phylloceratidae 337, 356 Phylloptychoceras 386 sipho 386 ; Pl. 11, fig. 1 sp. 385 Phyllotillon 165, 171, 184, 193, 198, 214, 219, 224, 225, 226 betpakdalensis 165, 174, 198, 223, 224, 225 NAVICUS 173, 223-225 Physocythere 75 virginea 48, 75, 76; Pl. i, figs. 11, 14-17 Placenticeras 342 veinecket 342 Placenticeratidae 338, 391 Plagiolophus annectens 217* Plasmopova conferta 432 Platycopina 8, 70 Platycythereis 68 chapmani 69, 76 ; Pl. 6, figs. 16, 18, 20 excavata 69 gaultina 68, 75, 70; Pl. 8, fig. 9 laminata 69, 77 ; Pl. 6, fig. 19 Plenastum 141, 142, 143, 146-149, 157-160 Plesiocyon 291 Pleurocythere 14 kivtonensis 14 nodosa 14 Pleurocytherinae 14 Pliocyon 291 Plithocyon 29% Pneumatocythere 13, 14 bajociana 13, 14 carinata 13, 14 ; Pl. 2, figs. 1-9 Podocopida 8, 42 Podocopina 42 Pogonodon 303 Polypova problematica 96 Polyptychoceras 338, 339, 385, 386, 403 havadanum 385 obliquecostatum 385 obstrictum 385 pseudogaultianum 338, 385, 386, 401, 402, 404 ; TEAL, eit, aay, subquadrvatum 385 subundulatum 385 vancouverense 385 Polyzoa 97 Pontocyprella 47 harvisiana 47, 75 triquetva 46, 75 Pontocypris 46 bosquetiana 46, 75 trigonalis 75 Portrane limestone 415-434 Praefuhrbergiella 10, 11 avens 10 favosa 10 minima 11 ; Pl. 1, figs. 1-8 Praeschuleridea 22 subtrigona 22 magna 23, 24; Pl. 7, figs. 6-11; Pl. 8, figs. 1-6 subtrigona 22 ventviosa 24 angulata 24 ventviosa 24 Pragnellia 430 avborescens 415, 430 ; Pl. 3, figs. 7, 8 Prionocyclus 342 Proamphicyon 291 Procytheridea triangula 26 ventyiosa 24 Procytheropteron 48 virgineum 48 Prodissopsalis 258, 259, 204 eocaentcus 259 Proetidae 326 Progonocythere 9 cristata 9, 1o Progonocytheridae 9, 47 Progonocytherinae 9 Progymnospermopsida 86 Proheliolites 432 dubius 415, 416,432 ; Pl. 4, figs. 10-12 Proheliolitidae 415, 432 Prolimnocyon 246 Propora hirsuta 430 Proporida 432 Propterodon 267, 279 Prorhyzaena 258 Protaraeida 428 Prothryptacodon 244 Protoacanthocythere 10 faveolata to Protobarinophyton 86 obrutschevii 86 Protocythere 57 auriculata 61, 62 consobrina 57, 58, 77 ; Pl. 5, figs. 17-19 jonest 58, 76 lineata 58, 59, 75-77; Pl. 5, figs. 1-8 vudispinata 59, 60, 76; Pl. 5, figs. g-11 tricostata 59, 76, 77 ; Pl. 5, figs. 14, 16 triebeli 61, 62, 76 triplicata 58 ; Pl. 5, figs. 12, 13, 15 Protocytheridae 25, 57 Proviverra 258, 262 Proviverrinae 241, 246, 253, 258 Pseudaelurinae 302 Pseudaelurus 302, 303, 307, 311 africanus 303, 304 ailuroides 307 intvepidus 307 kansensis 307 lortett 307 mayint 307 marshi 307 martint 307 pedionomus 307 quadridentatus 307 thinobates 303 tournauensis 307 transitovius 307 Pseudamphicyon 291 Pseudocreodi 246, 257 Pseudocythere 47 simplex 47, 76 Pseudokossmaticeras paulcki 391 Pseudophyllites 357, 359 indva 359 SP. 359 Psilophyton 84, 88 princeps 84 Pteridophyta 83 Pterodon 242, 263, 267, 270, 272, 278, 309, 311 africanus 241, 267, 272, 273*, 274-276, 309, BLor, PIE fies 3 biincisivius 268 californicus 272 coquandi 272 cuviert 272 INDEX Ptervodon—cont. dasyuroides 268, 272 grandis 272, 274 hyaenoides 272 leptognathus 272 magnus 272 nyanzae 241, 272, 274, 275*, 276, 309 parisiensis 272 phiomensis 272 Pterygocythereis 51 phylloptera 5% Ptilophyllum 121 Ptychoceras pseudogaultianum 385 Puzosia 342, 400 lytoceroides 343, 361 Puzosiinae 338, 386 Quepora 427, 428 aequabilis 427, 428 agglomeratiformis 428 delicatula 428 pavallela 427, 428 quebecensis 427, 428 Quercytherium 252, 310, 311 tenebrosum 247 Reuschia 423, 424 aperta 424 Sp. 415,423,424 ; Pl. 2, figs. 5, 6 Rhabdomesidae 113 Rhinidictyidae 114 Rhombopora 113 mawi 113, 144 ; Pl. 5, figs. 3-5 Romaniceras 342 Rugosa 415, 416, 417, 418 Sarcinula 415, 422, 423 latum 423 luhai 423 organum 423 sp. 422 ; Pl. 2, figs. 1-4 Sarcinulida 422 SAVAGE, R. J. G. 239-316 Schizotheriinae 222 Schizotherium 167, 171, 173, 184, 188, 191, 196, 198, 206, 208, 213, 214, 218, 219, 221— 225 avitum 191, 193, 196 pilgrimi 173 priscum 167, 170, 171, 173, 175*, 180, 183-185, 188, 191, 193, 194, 200, 201, ZOOS ZTON 2m ed ely 2 OM aes 222, 224 sp. 173, 196, 200 turgaicum 167, 172, 180, 188, 201, 204, 208, 213, 214, 216, 220, 222, 224 wetzlevi 171 Schloenbachia lenzi 351 Schuleridea 45 jonesiana 45, 46, 75, 76; Pl. 1, figs. 1-5 Schulerideidae 15 Schulerideinae 15 Scutellum 325 445 194, 223, 181, 206, 220, 206, 446 INDEX Scutellum (Scutellum) flabelliferum 324, 325 Selaginella 90 eggersit 9o vadiata 90 SELWOOD, E. B. 317-333 Sharpeiceras goliath 342 Sheffield Museum 97 Sinopa 253, 258, 259, 262-264, 310, 311 ethiopica 263 grangervi 259 Solenoceras 343, 402-405 bembense 402 binodosa 402 Sp. 402 Southcavea 27 bajociana 27, 28, 29 grandis 27, 29; Pl. 11, figs. 5-13; Pl. 12, fig. I reticulata 27, 28, 29 ; Pl. io, figs. 3-14 ; Pl. II, figs. 1-4 Spanoxyodon 244 Spathiceras 347 Sphenodiscidae 398* Sphenodiscus 338, 339, 394, 395, 396, 398, 399, 403, 404 lobatus 399 pleurisepta 395, 398* siva 399 Sp. 338, 396, 397, 398*, 402 ubaghsi 403 Sphenopteris 133 Spivopora 99 Sporogonites 90 exubevans 90 Stenoplesictinae 296, 301 Stictopova cvassa 115 Stoliczkaia 342 dispar 401 Streptelasma 418, 420 bystrowt 418 corniculum 419 craigense 420 distinctum 415, 418, 419 ; Pl. 1, figs. 8,9 euvopaeum 420 fragile 415, 418 ; Pl. 1, figs. 6, 7 orientale 419 poulsent 419 vusticum 415, 419, 420 ; Pl. 1, fig. 10 Streptelasma (Grewingkia) europaeum euvopaeum 420 (Kiaerophyllum) europaeum 420 Streptelasmatidae 415, 417 Stringocephalus burtini 324 Subprionocyclus 342 Subptychoceras 386 Svalbardia polymorpha 88, 90 Syringophyllidae 415, 422 Syrvingophyllum organum 423 Systenocythere 29, 30 extlofasciata 29, 30 sp. 30 ; Pl. 12, figs. 2-5 Tabulata 415, 422, 433 Tanganyika 165, 226 Tephrocyon 291 Teratodon 242, 247, 250, 252, 253, 255, 310, 311 enigmae 241, 253, 254*, 255*, 256*, 257, 309 ; Pl. 1, figs. 4, 5 spekei 241, 247, 248*, 249*, 250, 251*, 253, 255, 257, 309; Pl. 1, figs. 2, 3 Teratodontidae 241, 246, 247 Tetragonites 339, 401, 402 epigonus 343 jurinianus angolana 343 Tetragonitidae 337, 357 Texanites 342 Thamniscus problematica 96 Tickwood Beds 98 Ticéd Amphitheatre 121, 124 Todea 141-143, 146-149, 151, 156, 157, 159 barbara 148, 156* barnea 1.43 Tomarctus 291 Trachyleberidea 70 acutiloba 70, 76, 77 ; Pl. 8, figs. 7, 8, 10 Trachyleberididae 62 Trematopora striata 110 Trepostomata 96, 98, 102 Tricentes 245, 246 Trichopitys 135, 136 heteromorpha 135 Triisodontinae 243 Trileites langi 84 Tritemnodon 258, 262, 263, 310 Turrilites helicinus 383 polyplocus 372 SAXONICUS 372 Uganda 232, 242, 309 Veenia 60, 62 barringtonensis 60, 77 ; Pl. 6, figs. 1, 2, 3 harrisiana 60, 61, 62, 63, 66, 75-77; Pl. 4 fig. 1; Pl. 6, figs. q—11 triebeli 61 VINE, G. R. 96-97 Vishnucyon 293 Viverravinae 296, 301 Viverridae 241, 295, 296 Viverrinae 296, 301 Vulpes 291 Welsh National Museum 97 Wormispora 430, 432 hirsuta 415,430, 431, 432; Pl. 4, figs. 1-3 portranensis 415,431, 432 ; Pl. 4, figs. 4-9 ae lta. PRINTED IN G rad ‘THOMAS DE LA a a Sees th. ieataceres 5: ite i Po22 202353 Stirs steel af: eset: Be tere (Sze ptet senaeaerees