sin- V-\ , vv . BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY VOL. Ill 1956-1959 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) LONDON: 1956-1959 DATES OF PUBLICATION OF THE PARTS No. i. 1 8 April ....... 1956 No. 2. 12 October ...... 1957 No. 3. 13 December ...... 1957 No. 4. 31 January ...... 1958 No. 5. 14 February ...... 1958 No. 6. 20 February ...... 1958 No. 7. 20 February ...... 1958 No. 8. 26 August ...... 1958 No. 9. 12 September ..... 1958 No. 10. 6 February ...... 1959 PRINTED IN GREAT BRITAIN AT THE BARTHOLOMEW PRESS DORKING BY ADLARD AND SON, LTD. CONTENTS GEOLOGY VOLUME III PAGE No. i. The Structure, Evolution and Nomenclature of the Ostracod Hinge. P. C. SYLVESTER-BRADLEY i No. 2. Eocene Mollusca from Nigeria : A Revision. F. E. EAMES 23 No. 3. The Oligocene Flora of the Bovey Tracey Lake Basin, Devonshire. M. E. J. CHANDLER 71 No. 4. The Structure of some Leaves and Fructifications of the Glossopteris Flora of Tanganyika. D. D. PANT 125 No. 5. Lidgettonia, a new type of fertile Glossopteris. H. HAMSHAW THOMAS 177 No. 6. The Faunal Succession in the Caradoc Series of South Shropshire. W. T. DEAN 191 No. 7. A new Labyrinthodont (Pamcyclotosaurus] from the Upper Trias of New South Wales. D. M. S. WATSON 233 No. 8. An Early Pleistocene Mammalian Fauna from Bethlehem. D. A. HOOIJER 265 No. 9. The Upper Permian Flora of England. H. M. M. STONELEY 293 No. 10. Blue-Green Algae from the Middle Devonian of Rhynie, Aberdeenshire. W. N. CROFT AND E. A. GEORGE 339 Index to Volume 3 355 THE STRUCTURE, EVOLUTION AND NOMENCLATURE OF THE OSTRACOD HINGE P. C. SYLVESTER-BRADLEY BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 3 No. i LONDON: 1956 BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY The following papers appeared in Volume I (i949~52) : Pnce No. i (1949). The Pterobranch Rhabdopleum in the English Eocene. H. D. Thomas & A. G. Davis . • 75- ^d- No. 2 (1949). A Reconsideration of the Galley Hill Skeleton. K. P. Oakley & M. F. Ashley Montagu . 5*- No. 3 (1950). The Vertebrate Faunas of the Lower Old Red Sandstone of the Welsh Borders. E. I. White. Pteraspis leathensis White, a Dittonian Zone-Fossil. E. I. White . 7*. «• No. 4 (1950). A New Tithonian Ammonoid Fauna from Kurdistan, Northern Iraq. L. F. Spath . . 1O5- No. 5 (1951). Cretaceous and Eocene Peduncles of the Cirripede Euscal- pellum. T. H. Withers . . 5«- No. 6 (1951). Some Jurassic and Cretaceous Crabs (Prosoponidae) . T. H. Withers ... 55- No. 7 (1952). A New Trochiliscus (Charophyta) from the Downtonian of Podolia. W. N. Croft . . IOS- No. 8 (1952). Cretaceous and Tertiary Foraminifera from the Middle East. T. F. Grimsdale. . No. 9 (1952). Australian Arthrodires. E. I. White *5S- No. 10 (1952). Cyclopygid Trilobites from Girvan. W. F. Whittard The following papers appeared in Volume II (i953-I956) : No. i (1953). The Deer of the Webourn Crag and Forest Bed of Norfolk. A. Azzaroli ... • ' ' r ' No. 2 (1953). A Carboniferous Petrified Forest in Patagonia. Mary Gordon Calder .... No. 3 (1953). The Solution of the Piltdown Problem. By J. S. Weiner, K. P. Oakley & W. E. Le Gros Clark • 3*. &• No. 4 (1954). Some Upper Cretaceous & Eocene Fruits from Egypt. Marjorie E. J. Chandler . No. 5 (1954). The Carboniferous Flora of Peru. W. J. Jongmans . I5«- No 6 (1955). Further contributions to the Solution of the Piltdown Problem. J. S. WTeiner, W. E. Le Gros Clark, K. P. Oakley & others ... • • • i1 No. 7 (1955). The Schizaeaceae of the South of England in early Tertiary Times. Marjorie E. J. Chandler . I5$- No. 8 (1956). The Brachyopid Labyrinthodonts. D. M. S. Watson. THE STRUCTURE, EVOLUTION AND NOMENCLATURE OF THE OSTRACOD HINGE BY PETER C. SYLVESTER-BRADLEY University of Sheffield Pp. 1-21; Pis. 1-4 ; 2 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 3 No. i LONDON : 1956 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical Series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 3, No. i of the Geological series. PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued April, 1956 Price Twelve Shillings and Sixpence THE STRUCTURE, EVOLUTION AND NOMENCLATURE OF THE OSTRACOD HINGE By P. C. SYLVESTER-BRADLEY [University of Sheffield] SYNOPSIS A revision of Middle Jurassic ostracods from the Fullers' Earth Clay has revealed a series of hinge-types which clearly demonstrate transitional stages between those commonly known as merodont and those known as amphidont. Their study has justified the classification of hinge-types according to the stage reached in a morphogenetic series. The primitive hinge is tripartite in character, and more specialized hinge-types originate by the subdivision of the median element of the basic three. The morphology of the ostracod hinge is discussed, a rational classification is proposed, and in the light of this classification, the terminology of both the hinge as a whole and that of its constituent parts is reviewed. The evolution of the post- Palaeozoic Cytheracea is traced from Palaeozoic ancestors with a similar hinge but different muscle-scar pattern, and a new superfamily, the Quasillitacea, is proposed for these Palaeozoic forms. Two new genera are proposed, Acanthocythere (Middle Jurassic) and Dictyocythere (Upper Jurassic). The latter is divided into two subgenera, Dictyocythere sensu stricto and Rhysocy there nov. The distribution of the species of Dictyocythere in the Upper Jurassic of England and northwest Germany suggests that the " Purbeck " Beds of the Aylesbury and Swindon districts are earlier than Middle Purbeckian, and that the " Wealden 2 " of the German sequence is probably equivalent to the upper part of the Middle Purbeck Beds of Dorset. The development of the hinge in Dictyocythere is shown to be palingenetic. i. TERMINOLOGY OF THE OSTRACOD HINGE THE hinge structures of fossil ostracods play an important part in their taxonomic determination, but unfortunately the terms used for their description are ambiguous. Thus the terms " taxodont " and " heterodont " have been used to describe the compound hinge of ostracoda, not always with quite the same meaning. As used by Bold (1946), Kingma (1948), Grekoff (1952) and Malkin (I953),1 " taxodont " refers to the presence of denticulate elements, and " heterodont " to the development of high, pointed teeth and a hinge-bar. The terms (which have been borrowed from those used to describe the lamellibranch hinge) are not altogether appropriate in their new context, and seem likely to lead to ambiguity. The term " taxodont " as used in lamellibranchs refers to the occurrence of an alternate tooth and socket arrange- ment, essentially similar in both valves. An exactly similar structure is not known to occur in the Ostracoda, although an alternate tooth and socket arrangement (here termed " interdentate ") is not infrequently found in some or all of the elements 1 Malkin (1953) also defines the terms adont, dysodont, desmodont, crasidont and archidont. These terms are not referred to further in this paper, as they do not describe hinge structures here dealt with. GEOL. Ill, I. I 4 STRUCTURE, ETC. OF THE OSTRACOD HINGE of the compound hinge. This arrangement might well be called " taxodont " and the term has in fact been so used by Kesling (1951) ; this use of the term is not, however, equivalent to that of the earlier definitions, and it would seem advisable to abandon it to prevent further confusion. Other authors have used different terms descriptive of the same structures ; thus " merodont " = " taxodont ", and " amphidont " =" heterodont " (e.g. Triebel 1950 ; Pokorny, 1952). These terms are not open to the same objections, and are adopted here and fully defined below. Additional terms are also introduced ; some of these are new, some are after Triebel. Technical terms, in standard English, or anglicized from Latin roots, are used to designate structures of the individual elements ; terms derived from the Greek and terminating in " -dont " are used to designate the hinge structure as a whole. 2. DEFINITIONS OF HINGE ELEMENTS The hinge structure of the ostracod carapace may be simple or compound. The simple hinge (" adont " of Grekoff, Bold and Kingma) may be thought of as made up of a single element, which consists of a groove along the margin into which fits the edge of the other valve (Text-fig, i, A). The hinge is not further differentiated. (/anterior median posterior —''anterior "-xanterbmedkfri posteromedian posterior A B TEXT-FIG, i. — Diagrams of left valves of three ostracods, dorsal view, to illustrate division of hinge into "elements". (Stippled areas represent sockets). The compound hinge is divided into three or four elements in each valve. Primi- tively there are three elements, anterior, median and posterior (Text-fig, i, B). Usually these elements alternate so that if (as is usual in the right valve) the anterior and posterior are ridges, the median element separating them will be a groove ; or if (as is usual in the left valve) the terminal elements are grooves or sockets, the median will be a ridge. Though genera are known which are exceptions to this (in Haplo- cytheridea, for example, all three elements of the left valve are grooves, those of the right being ridges), the distinction between the constituent elements remains clear, the terminal elements being more coarsely dentate than the median. It will be shown below that the more primitive compound hinge with three elements STRUCTURE, ETC. OF THE OSTRACOD HINGE 5 evolves, by subdivision of the median element, into a more advanced type in which there are four elements — anterior, anteromedian, posteromedian and posterior. The division into two parts of the median element is at first not pronounced — the anterior part being merely more coarsely dentate than the posterior (Text-fig. 2, c), but later developed forms have a more clear-cut distinction, the anteromedian becoming rather similar in proportions to the anterior element, the posteromedian usually being much the longest element of the four (Text-fig, i, c) . Rarely the anterior element also becomes differentiated, with denticles of two sizes (e.g. Amphicy there and Dictyocythere ; see p. 14). These hinge elements themselves display a great deal of variation, and the following definitions are of terms used in their description : A. Ridge and groove. i. A hinge element may consist of a smooth ridge fitting into a corresponding groove in the opposite valve, e.g. the median element of Camptocythere (Text-fig. 2, A). A. Lophodont B. Merodont C. Entomodont D. Lobodont E. Schizodont F Amphidont TEXT-FIG. 2. Diagrams illustrating the six main types of dentition developed in the ostracod hinge. Each pair of valves in dorsal view. (Stippled areas represent sockets). 6 STRUCTURE, ETC. OF THE OSTRACOD HINGE 2. A dentate element consists of a ridge divided into projections (denticles) which fit into a socketed groove termed loculate. 3. The median element may be furnished with finer denticles, when it is termed denticulate, the groove locellate (Text-fig. 2, B). 4. A ridge may be divided into separate teeth between which lie alternating sockets for the reception of projections from the complementary groove. Such a ridge is termed interdentate, the groove interloculate. 5. The median element of the left valve (which is usually a ridge or a bar) often lies under an elongated, sub-triangular, excavated area termed the accommodation groove. This, it would seem, serves to receive the project- ing dorsal edge of the right valve when the carapace is open. B. Bar. 6. The median element may consist of a ridge which is raised up from the shell margin behind it ; this is differentiated as a hinge-bar. C. Boss and pit. 7. An element may be raised up as a hemispherical boss, which may show division into two or more lobes, when it is referred to as bi-, tri-, or multi- lobate ; a lobate boss fits into a loculate pit in the complementary valve (Text-fig. 2, D). D. Tooth and socket. 8. An element may consist of a single tooth, projecting from a platform which itself projects more or less beyond the hinge margin, when it is referred to as stirpate (e.g. the anterior element in Trachyleberis] (Text-fig. 2, F). 9. The tooth may be a simple, more or less conical projection ; 10. or it may have more or less parallel sides, when it is called pessular ; 11. or it may be split into two lobes, when it may be called bifid (Text-fig. 2, E) ; 12. or it may be crescentic, half-surrounding a socket ; 13. or it may be a slightly elongated and slightly curved projection, when it is termed reniform (as in the posterior element of many amphidont hinges (Text-fig. 2, F). A reniform tooth may be smooth (Hemicythere or Trachy- leberis} or lobate (Bradley a] or dentate (some species of Cythereis}. 3. CLASSIFICATION OF HINGE TYPES Using the above definitions of the elements of the compound hinge as a basis, it is possible to classify hinges developed in the Cytheracea in the six main groups outlined below. This classification does not take into consideration hinge-types developed in other superfamilies of the Podocopa. For example, whereas the Quasillitacea (here defined) have hinges very like the Cytheracea (though apparently confined to the more primitive lophodont and merodont types), the Bairdiacea have a more complicated compound hinge not always homologous with that in the Cytheracea — compare, for example, Bairdoppilata (Coryell, Sample & Jennings, 1935) with Macrocypris (Sylvester-Bradley, 1948^) and Triebelina (lophodont ; Triebel, 1948). The Cypridacea, STRUCTURE, ETC. OF THE OSTRACOD HINGE 7 on the other hand, have mostly simple hinges with a single element, though a few are lophodont (e.g. Cypridea ; Sylvester-Bradley, 1949). i. Lophodont: e.g. Bythocythere, Camptocythere, in which the hinge elements are divided into three in each valve, all consisting of ridges and grooves (Text-fig. 2, A). The most usual arrangement is as follows : Left valve Right valve Anterior element . . Groove . Ridge. Median element . . Ridge . Groove. Posterior element . . Groove . Ridge. 2. Merodont: ( = " taxodont " of some authors) e.g. Cythere, in which anterior and posterior elements are dentate or loculate, the median element either smooth or denticulate, but not subdivided (Text-fig. 2, B). The most usual arrangement is as follows : Left valve Right valve Anterior element . . Loculate groove . Dentate ridge. Median element . . Denticulate ridge . Locellate groove. Posterior element . . Loculate groove . Dentate ridge. The elements in some merodont genera (e.g. the terminal elements of Pleurocythere) are interdentate and interloculate rather than dentate and loculate. Many species of Middle Jurassic age have a merodont hinge with smooth median element (e.g. Schuleridea) , and this condition may well be the more primitive. 3. Entomodont: e.g. Lophocy 'there, Progonocythere and Xenocythere (Triebel, 1949, 1951), in which the median element becomes subdivided, the anterior part remaining dentate, the posterior smooth or denticulate (Text-fig. 2, c). A common arrangement is as follows : Left valve Right valve Anterior element . . Loculate groove . Dentate ridge. Anteromedian element . Short dentate ridge • . Short, wide loculate groove. Poster omedian element . Long denticulate ridge . Long, narrow locellate groove. Posterior element . . Loculate groove . Dentate ridge. The genus Macrodentina has a hinge transitional between entomodont and lobodont (seep. 14). 4. Lobodont: e.g. Acanthocythere gen. nov., in which the anterior element and the anteromedian elements are lobed bosses (see PL i, figs. 3, 4, 8, 9 ; Text-fig. 2, D). Left valve Right valve Anterior element . . Loculate pit . Lobate boss. Anteromedian element . Lobate boss . Loculate pit. Posteromedian element . Smooth or denticulate bar . Smooth or locellate groove. Posterior element . . Loculate groove . Dentate ridge. 8 STRUCTURE, ETC. OF THE OSTRACOD HINGE The genus Amphicythere has a dentition transitional between lobodont and amphi- dont. The anterior element is lobate, but the antero-median element is almost entire (see p. 14). 5. Schizodont: e.g. Schizocy there, Palmenella, Paijenborchella (Triebel, 1950), in which teeth of the anterior and anteromedian elements are bifid (Text-fig. 2, E). Left valve Right valve Anterior element . . Biloculate socket . Bifid, stirpate tooth. Anteromedian element . Bifid tooth . Biloculate socket. Posteromedian element . Denticulate bar . Locellate groove. Posterior element . . Loculate socket . Lobate, reniform tooth. 6. Amphidont: ( = " heterodont " of some authors ; Triebel uses the term amphidont in a wider sense than that proposed here — to include dentitions of both schizodont and lobodont types) e.g. Dictyocythere, Trachyleberis , in which the median element is further differentiated into a single anterior tooth and a posterior bar or ridge. A common arrangement (Text-fig. 2, F) is as follows : Left valve Right valve Anterior element . . Socket . Stirpate tooth. Anteromedian element . Conical or pessular tooth . Socket. Posteromedian element . Smooth or denticulate bar . Smooth or locellate groove. Posterior element . . Loculate socket . Lobate, reniform tooth. Young moults of amphidont species usually have merodont or entomodont hinges (see p. 19). The ontogeny is therefore recapitulatory . This classification is not comprehensive. The hinge structures listed form a useful morphogenetic series, but complications are developed in some genera which are not covered — for example, the differentiations of the median element found in some genera of the Cytherideinae (e.g. Cytheridea, Cyprideis), and the development of crescentic teeth in the terminal elements in Loxoconcha and other genera. 4. THE EVOLUTION OF THE HINGE AND ITS BEARING ON THE CLASSIFICATION OF THE PODOCOPA Well-defined compound hinge-structures are particularly characteristic of the superfamily Cytheracea, which in post-Palaeozoic times provided a far greater number of species of marine ostracods than all the remaining superfamilies taken together. The series of hinge-types classified above is a morphogenetic series and not phylo- genetic. Genera have passed through the various stages at different times during their racial history. The most advanced stage reached by Middle Jurassic times is the lobodont, but the amphidont hinge appears before the close of the Upper Jurassic (e.g. Dictyocythere, see p. 14) . It is not unlikely that earlier genera with an amphidont hinge remain to be discovered. In Recent and Tertiary ostracods, merodont and amphidont hinge-structures predominate but entomodont hinges also occur (e.g. Leptocythere) ; in Middle Jurassic times the Cytheracea have hinges which are pre- STRUCTURE, ETC. OF THE OSTRACOD HINGE 9 dominantly either merodont (usually with a smooth median element) or entomodont ; lobodont hinges are rare. Elsewhere (Sylvester-Bradley, 19486) attention has been drawn to the series leading from the Middle Jurassic Oligocythereis (entomodont) through the Cretaceous Cythereis to the Tertiary and Recent Trachyleberis (amphi- dont). This series is thought to be broadly phylogenetic, and the three genera in question are placed in the same subfamily, the Trachyleberidinae. Several other undescribed genera from the Middle Jurassic have reached various stages in similar series. It seems unlikely that any of these forms are directly related to the Trachyle- beridinae ; they probably represent other lineages undergoing parallel evolution. The acquisition of the amphidont hinge is considered to be a well-marked trend. If this view is correct, the diagnosis of the subfamily Progonocytherinae (Sylvester- Bradley, 19480 : 189) for Cytheracea with entomodont hinge, needs drastic revision, and it is now suggested that Lophocythere, Progonocy there and Oligocythereis should not be grouped together in one subfamily, although they all have entomodont hinges. Before attempting to assess which of the hinge-types described can truly be taken as primitive, it is necessary to refer to older faunas and seek possible Palaeozoic ancestors of the Cytheracea. The genus Monoceratina is the longest lived of the Cytheracea according to present ideas. The type species is Carboniferous in age, and Recent species have been recorded (Stephenson, 1946 ; Bold, 1946). However, the hinge structure is rather obscure ; the shape differs much from the majority of the Cytheracea ; the muscle- scar patterns differ sufficiently in different species to warrant query as to whether there has not been a mistake in the identification of some of the post-Palaeozoic material, and it is not in any case easy to postulate that Monoceratina exhibits an ancestral type of hinge-structure. Perhaps the Palaeozoic family that most invites comparison with the Cytheracea is the Middle Devonian Ropolenellidae first described by Coryell & Malkin (1936), especially the two genera Ropolenellus and Euglyphella. By the courtesy of Mr. Raymond R. Hibbard, who presented me with several samples rich in ostracods, which he collected from the Middle Devonian of New York State, I am able to describe the more intimate details of the shell of some species of Euglyphella. Some of the earlier authors who dealt with the Ropolenellidae nave oriented the ostracods so that the higher end is posterior. The investigation of the hinge reveals that they are closely analogous, if not homologous, with those found in the Cytheracea ; the higher end is therefore here regarded as anterior in conformity with the known orientation of the Cytheracea. Consequently the terms " left " and " right " used here have the reverse meaning to those used in description by some previous authors. The shape of these two genera of the Ropolenellidae, in marked contrast to many other described species of Palaeozoic ostracods, is similar to many genera of the Cytheracea (PI. 2, figs. 5-7). A broadly rounded anterior tapers to a narrow, rounded or triangular posterior. Either or both ends may be spinose. The hinge-line is straight, delimited by well-marked anterior and posterior cardinal angles and is shorter than the length of the shell. The ventral border is straight or slightly concave. The genus Euglyphella was proposed by Warthin (1934) with Strepula sigmoidalis Jones as type species. Several other species have since been assigned to the genus. GEOL. Ill, I. 2 io STRUCTURE, ETC. OF THE OSTRACOD HINGE It possesses a striking ornament of carinae not unlike those developed in many post- Palaeozoic Cytheracea (e.g. Lophocythere , from the Bathonian ; see Sylvester- Bradley, 19480). The hinge in Euglyphella is divided into three elements as follows : Left valve Right valve Anterior .... Rather long, deeply over- . Projecting ridge, forming a hung, loculate groove continuation of the selvage. Median .... Bar, lying below a long, . Long groove. narrow accommodation groove Posterior. . . . Short, curved loculate groove . Short, curved projecting ridge. The hinge margin of this genus is in fact almost exactly similar to that of many post-Palaeozoic Cytheracea. The loculate nature of the terminal elements in the left valve cannot be made out in more than a few specimens, and I have not detected any specimen showing a corresponding dentation on the terminal elements of the right valve, but this may well be due to indifferent preservation. The free margin is also similar to that seen in many Cytheracea. There is a wide duplicature, no vestibule, but a pronounced selvage. The muscle-scar pattern, however, is quite unlike anything known in the post-Palaeozoic Cytheracea ; a central, circular muscle-scar pit bears a cluster of small oval scars in the form of slight tubercles, which are in close juxtaposition. Another genus referred by Coryell & Malkin (1936) to the Ropolenellidae is Bufma (PI. 2, figs. 3, 4). Species of this genus differ somewhat in shape from Euglyphella and Ropolenellus in that the dorsal margin of the left valve is markedly curved. The characteristic ornament of the genus is unlike that of any post-Palaeozoic ostracod known to the author, though Henningsmoen (1953) has pointed out that it is shared by other Palaeozoic genera such as Ponderodictya (see below). The details of hinge, duplicature and muscle-scar, however, are as described for Euglyphella, and though Coryell & Malkin were clearly correct in concluding that the genera are related, Henningsmoen is probably right in removing Bufina to the Quasillitinae. The Ropolenellidae are apparently only known from Middle Devonian strata, in which they are often fairly common. They seem to represent a successful though short-lived line of evolution which in many respects anticipated the much later Cytheracea. Another family well represented in the Middle Devonian of North America is the Quasillitidae. An examination of Quasillites (PI. 2, figs, i, 2) shows that it has a hinge-structure very similar to that of the Ropolenellidae and the two families should almost certainly be grouped together. Henningsmoen (1953) regarded them as subfamilies of one family ; it is here suggested that they are related families in the same superfamily, the Quasillitacea (see p. n). Near to Quasillites is the Carboni- ferous genus Graphiadactylus. It may also be mentioned that the genus Ponderodictya (described by Coryell & Malkin from the Middle Devonian and placed by them in the Cytherellidae ; see figs, in Triebel, 1954) bears a hinge comparable to that of the Ropolenellidae, differing mainly in that the anterior element is longer in proportion (as long, in fact, as the STRUCTURE, ETC. OF THE OSTRACOD HINGE n median element). The ornament is not unlike that developed in Bufina. The genus has been placed in the Healdiidae as a member of the Platycopa (Henningsmoen, 1953 ; Triebel, 1954), but the hinge makes this assignment a little doubtful. Pondero- diclya may perhaps be a member of the Quasillitacea, and as such one of the Podocopa. It is clear that in Devonian times the Healdiidae, Quasillitidae and Bairdiidae possessed characters in common which suggest that all three families had been derived from a common ancestor of not much greater antiquity than the Devonian. Subsequently the three families diverged in morphology so that they have been placed far apart in classification. From the foregoing brief survey it will be seen that several families of Palaeozoic ostracods possessed features in common with the later Cytheracea. It seems probable that the first members of the Cytheracea were derived from one or more of the Quasil- litacea towards the close of Palaeozoic time. Somewhat similar views have been advanced by Kellett (1943). Thus the superfamilies Quasillitacea and Cytheracea form part of the suborder Podocopa, and are distinguished by the muscle-scar pattern, which in the Quasillitacea is aggregate (i.e. consisting of a group of individual scars crowded in close juxtaposition ; PI. 2, figs. 3, 4, 6), and in the Cytheracea discrete (i.e. consisting of a series of individual scars separated from each other). The primitive hinge was therefore lophodont. Whether the elements were primi- tively dentate or smooth cannot be certainly stated, for the apparently smooth elements in much Palaeozoic material may well be due to a loss of finer structure on recrystallization. The fact that loculate grooves have been detected in the terminal elements of both Euglyphella and Quasillites shows that at least the terminal elements were dentate in these genera by Devonian times. Some Palaeozoic specimens show a well-developed accommodation groove, so that structure cannot be taken as an advanced feature. It. seems to be the inevitable consequence of any considerable difference in size between the two valves. 5. SYSTEMATIC DESCRIPTIONS Suborder PODOCOPA Superfamily QUASILLITACEA (Coryell & Malkin). (First introduced here as a superfamily) TYPE GENUS. Quasillites Coryell & Malkin. (Middle Devonian.) DIAGNOSIS. Podocopa with a primitively tripartite hinge and an aggregate muscle-scar pattern. FAMILIES INCLUDED. Quasillitidae (including Graphiadactyllidae) and Ropolenel- lidae. Other families no doubt remain to be described. REMARKS. Distinguished from Cytheracea by muscle-scar pattern ; from Cypridacea and Bairdiacea by hinge and muscle-scar pattern ; from Healdiidae (Platycopa) by hinge. DISTRIBUTION. So far known only from the Palaeozoic. Mesozoic Platycopa with an aggregate muscle-scar pattern are, however, known (Ogmoconcha, Lower Lias) GEOL. Ill, I. 2§ 12 STRUCTURE, ETC. OF THE OSTRACOD HINGE and it may be expected that the Podocopa underwent a parallel development ; Mesozoic Quasillitacea may yet be discovered. Superfamily CYTHERACEA Family uncertain Genus ACANTHOCYTHERE nov. TYPE SPECIES. Cythere sphaerulata Jones & Sherborn, 1888 : 253, pi. i, fig. 6. DERIVATION OF NAME. anavBa, a thorn or prickle + genus Cythere. Gender: feminine. DIAGNOSIS. Cytheracea with lobodont hinge, surface more or less spiny, carapace plump, eye tubercles shiny, rather prominent. OCCURRENCE. The two species of the genus here described are so far known only from the Upper Fullers' Earth Clay (Bathonian) of the Bath district. Acanthocy there sphaerulata (Jones & Sherborn) (Plate i, figs. 1-4) 1888. Cythere sphaerulata Jones & Sherborn, p. 253, pi. i, fig. 6. DIAGNOSIS. Carapace oblong, tumid, ornamented with closely-set blunt spines arranged in a faint reticulate pattern. HOLOTYPE. Geol. Dept. Brit. Mus. (N.H.) no. 1.1835. A complete carapace. TYPE LOCALITY. " Yellow Fullers-earth Clay, Midford " (Jones & Sherborn, 1888 :• 254). OTHER FIGURED SPECIMENS. Geol. Dept. Brit. Mus. (N.H.) nos. 1^.42433-34. DESCRIPTION. Both valves with sub-parallel dorsal and ventral margins, no taper, anterior and posterior ends evenly rounded. Sexual dimorphism rather pro- nounced, the presumed males longer than females. Dimensions (mm.) Proportions L H W L H W Carapace FIGS, i, 4. Left valve male (external and internal lateral views). In. 42434 (0-67 mm. long). FIGS. 2, 3. Right valve male (external and internal lateral views). 1^42433 (0-69 mm. long). Acanthocythere spiniscutulata n. sp. FIGS. 5, 8. Left valve (external and internal lateral views), holotype. 1^42435 (0-57 mm. long). FIGS. 6, 9. Right valve (external and internal lateral views), paratype. ^.42436 (0-56 mm. long). FIG. 7. Carapace (dorsal view), paratype. In. 42437 (0-57 mm. long). All figures approx. x 90 (exact lengths given above), from pencil drawings by Miss D. Robinson based on camera lucida sketches by the author. All specimens from the economic Fullers' Earth (Upper Fullers' Earth Clay, Bathonian, Middle Jurassic) from the Fosse Way mine, nr. Bath, collected by the author, now in the British Museum (Nat. Hist.). Ru)l. B. M.\N.H.) Geol. Ill, i. 1LATE I ACANTHOCYTHERE GEN. NOV. (MIDDLE JURASSIC) PLATE 2 Quasillites sp. FIG. i. Right valve, external lateral view. The irregularly shaped blank space below the centre is obscured by matrix. In. 48624 (1-16 mm. long). FIG. 2. Lett valve, internal lateral view. In. 48623 (1-09 mm. long). Bufina sp. FIG. 3. Left valve, external lateral view. In. 48622 (0-89 mm. long). FIG. 4. Right valve, external lateral view. In. 48621 (0-93 mm. long). Euglyphella sp. FIGS. 5, 6. Right valve, external lateral and internal lateral views. In. 48619 (0-99 mm. long). FIG. 7. Left valve, internal lateral view. In. 48620 (1-13 mm. long). Figs, i, 2, approx. x 70 ; 3-7, x 65. Exact lengths given above. All are pencil drawings by Miss D. Robinson based on camera lucida sketches by the author. All specimens from the Pleurodictyum Bed (Ludlowville, Hamilton Group, Middle Devonian) of Bay View, Erie County, New York, U.S.A., collected by Raymond R. Hib- bard, now in the British Museum (Nat. Hist.). Bull. B. M. (N.H.) Gcol. Ill, t. PLATE 2 MIDDLE DEVONIAN OUASILLITACEA PLATE 3 Dictyocythere (Dictyocythere) decorata (Anderson) FIG. i. Left valve, external lateral view. In. 48618 (0-94 mm. long). Cythere Marl, Swindon Series, Swindon (? Purbeckian) . Dictyocythere (Dictyocythere) mediostricta n. sp. FIG. 2. Carapace $, left lateral view. In. 48606 (1-02 mm. long). FIGS. 3-6. Carapace $, right lateral, left lateral, dorsal and ventral views, holotype. In. 48607 (1-1-2 mm. long). Middle Purbeck Beds, Worbarrow Tout, Dorset. Dictyocythere (Dictyocythere) retirugata (Jones) FIG. 7. Carapace <$, ventral view. Specimen accidentally destroyed. FIG. 8. Right valve $, external lateral view. In. 48611 (0-84 mm. long). FIG. 9. Left valve $, external lateral view. In. 48610 (0-80 mm. long). FIG. 10. Right valve $, external lateral view. In. 48608 (0-97 mm. long). All from Cythere Marl, Swindon Series, Swindon (? Purbeckian). Dictyocythere (Rhy so cy there) transiens (Jones) FIGS. 11-13. Carapace (right), lateral, dorsal and ventral views. In. 48617 (0-60 mm. long). Upper Portland Beds, Bugle Pit, Hartwell. All figures from photographs by the author, x 60 approx. The specimens whitened with magnesium oxide before being photographed. Bull. 13. M. (N.H.} Geol. Ill, i. PLATE 3 DICTYOCYTHERE GEN. NOV. (UPPER JURASSIC) PLATE 4 Diet} ocvthere (Rhy so cy there] rugulata (Jones) FIGS, i, 2. Lett and right valves $, dorsal views. In. 48604 (0-84 mm. long), In. 48605 (0-86 mm. long). Dictyocythere (Dictyocy there) retirugata (Jones) FIGS. 3, 4. Left and right valves $, dorsal views. In. 48610 (0-80 mm. long), In. 48611 (0-84 mm. long). Dictyocythere (Rhy socy there) rugulata (Jones) FIG. 5. Internal view of centre of right valve to show muscle-scar pattern. In. 4861 2 (x 200). FIG. 6. Left valve $ photographed under water by reflected light to show normal pore canals and minute punctation behind and below central region. In. 48604 (0-84 mm. long). FIGS. 7, 8. Dorsal views of left and right valves of juvenile moults showing entomodont hinge. ^.48614 (0-68 mm. long), 111.48613 (0-67 mm. long). FIGS. 9, 10. Anterior hinge elements of juvenile and adult right valves in dorsal view. In. 48613, In. 48605 (x 200). Dictyocvthere (Dictyocythere) retirugata (Jones) FIG. ii. Anterior element of right valve in dorsal view. In. 48611 (X 200). Dictyocythere (Rhy socy there) rugulata (Jones) FIG. 12. External lateral view of left valve of <$. In. 48615 (0-97 mm. long). FIGS. 13-15. (J carapace in right lateral, ventral and dorsal views. 111.48616 (0-98 mm. long). Dictyocythere (Dictyocythere) retirugata (Jones). FIGS. 16, 17. Dorsal views of left and right valves $. In. 48609 (0-92 mm. long), 111.48608 (0-97 mm. long). Figs, i, 2, 6-10, 12-15 from so-called Purbeck Beds of Bugle Pit, Hartwell. Figs. 3, 4, 5, n, 16, 17 from Cythere Marl, Swindon Series, Swindon (? Purbeckian). All figures from photographs by the author. Figs. 1-4, 6-8, 12-17 X 55 approx. Specimens in Figs. 12-14 whitened with magnesium oxide before being photographed. null. B. M. (N.H.) Gcol. HI, I. PLATE 4 PICTYOCYTHERE GEN, NOV. (UPPER JURASSIC) EOCENE MOLLUSCA FROM NIGERIA : A REVISION F. E. EAMES BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 3 No. 2 LONDON : 1957 BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY The following, papers appeared in Volume I (1949-52) : Price No. I (1949). The Pterobranch Rhabdopleura in the English Eocene. H. D. Thomas & A. G. Davis . . . . . . . . 75. 6d. No. 2 (1949). A Reconsideration of the Galley Hill Skeleton. K. P. Oakley & M. F. Ashley Montagu ....... 55. No. 3 (1950). The Vertebrate Faunas of the Lower Old Red Sandstone of the Welsh Borders. E. I. White. Pteraspis leathensis White, a Dittonian Zone-Fossil. E. I. White 75. 6d. No. 4 (1950). A New Tithonian Ammonoid Fauna from Kurdistan, North- ern Iraq. L. F. Spath ........ IDS. No. 5 (1951). Cretaceous and Eocene Peduncles of the Cirripede Euscal- pellum. T. H. Withers ........ 55. No. 6 (1951). Some Jurassic and Cretaceous Crabs (Prosoponidae). T. H. Withers .......... 55. No. 7 (1952). A New Trochiliscus (Charophyta) from the Downtonian of Podolia. W. N. Croft los. No. 8 (1952). Cretaceous and Tertiary Foraminifera from the Middle East. T. F. Grimsdale ......... 105. No. 9 (1952). Australian Arthrodires. E. I. White .... 155. No. 10 (1952). Cyclopygid Trilobites from Girvan. W. F. Whittard . 6s. The following papers appeared in Volume II (1953-56) : No. I (1953). The Deer of the Weybourn Crag and Forest Bed of Norfolk. A. Azzaroli . . . . . . . . . . £i 55. No. 2 (1953). A Coniferous Petrified Forest in Patagonia. Mary G. Calder I2S. No. 3 (1953). The Solution of the Piltdown Problem. J. S. Weiner, K. P. Oakley & W. E. Le Gros Clark . . . . 35 6d. No. 4 (1954). Some Upper Cretaceous and Eocene Fruits from Egypt. Marjorie E. J. Chandler ........ 165. No. 5 (1954). The Carboniferous Flora of Peru. W. J. Jongmans . . 155. No. 6 (1955). Further Contributions to the Solution of the Piltdown Problem. J. S. Weiner, W. E. Le Gros Clark, K. P. Oakley & others ........... £i No. 7 (1955). The Schizaeaceae of the South of England in early Tertiary Times. Marjorie E. J. Chandler . . . . . . 155. No. 8 (1956). The Brachyopid Labyrinthodonts. D. M. S. Watson . . £i EOCENE MOLLUSCA FROM NIGERIA A REVISION BY FRANK E. EAMES Pp. 23-70 ; Pis. 5-10 BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 3 No. 2 LONDON: 1957 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series, corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 3, No. 2 of the Geological series. PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued October, 1957 Price Twenty-five Shillings EOCENE MOLLUSCA FROM NIGERIA: A REVISION By F. E. EAMES CONTENTS Page I. INTRODUCTION AND GENERAL REVIEW ...... 29 II. SYSTEMATIC DESCRIPTIONS ........ 30 A. GASTROPODA .... . -3° Family POTAMIDIDAE ........ 30 Genus Exechostoma Cossmann . . . . .30 Exechostoma cossmanni Newton ..... 30 Genus Terebralia Swainson ....... 30 Terebralia nigeriensis sp. nov. ..... 30 Terebralia amekiensis sp. nov. . . . . .31 Family TURRITELLIDAE . . . . . . . .31 Genus Turritella Lamarck . . . . . . .31 Turritella amekiensis sp. nov. . . . . .31 Subgenus Coeloconica nov. ....... 32 Turritella (Coeloconica) mauryana Newton . . .32 Family ARCHITECTONICIDAE ....... 32 Genus Architectonica (Bolten MS.) Roding .... 32 Subgenus Nipteraxis Cossmann ...... 32 Architectonica (Nipteraxis) bendeica sp. nov. ... 32 Subgenus Stellaxis Dall . . . . . . -33 Architectonica (Stellaxis) bicingulata (Newton) . . 33 Subgenus Solariaxis Dall ....... 33 Architectonica (Solariaxis) amekiensis sp. nov. . . 33 Family SCALIDAE ......... 34 Genus Acrilla H. Adams ....... 34 Acrilla nigeriensis sp. nov. ...... 34 Family CALYPTRAEIDAE ........ 35 Genus Calyptraea Lamarck ....... 35 Calyptraea newtoni sp. nov. ...... 35 Genus Turbocalyptraea nov. ....... 35 Turbocalyptraea scabrosa sp. nov. ..... 35 Genus Crepidula Lamarck ....... 36 Subgenus Concavimargo nov. ...... 36 Crepidula (Concavimargo) falconeri Newton ... 36 Family XENOPHORIDAE ........ 36 Genus Xenophora Fischer von Waldheim . . . 36 Xenophora nigeriensis (Newton) ..... 36 Family STROMBIDAE ........ 36 Genus Tibia (Bolten MS.) Roding 36 Tibia bidigitata (Newton) ...... 36 Genus Cyrtulotibia nov. . . . . . . .36 Cyrtulotibia unidigitata (Newton) . . . . -37 GEOL. Ill, 2. 26 EOCENE MOLLUSCA FROM NIGERIA Genus Semiterebellum Cossmann . . . . . -37 Subgenus Africoterebellum nov. . . . . . -37 Semiterebellum (Africoterebellum) elongatum Newton . 38 Genus Amekichilus nov. ....... 38 Amekichilus suturocostatum (Newton) .... 39 Family AMPHIPERATIDAE Genus Eovolva Schilder ........ 39 Eovolva nigeriensis (Newton) ..... 39 Genus Sphaerocypraea Schilder ...... 39 Sphaerocypraea sudanensis (Schilder) . . . -39 Family NATICIDAE ......... 39 Genus Neverita Risso ........ 39 Neverita amekiensis sp. nov. ..... 39 Genus Sinum (Bolten MS.) Roding ..... 40 Sinum africanum Newton ...... 40 Sinum nigeriense sp. nov. ...... 40 Family CYMATIIDAE ........ 40 Genus Varicohilda nov . . . ... . .40 Varicohilda turriculata (Newton) . . . . .41 Family MURICIDAE . . . . . . . . .41 Genus Hexaplex Perry . . . . . . . .41 Subgenus Paziella Jousseaume . . . . . .41 Hexaplex (Paziella) bendeica sp. nov. . . . .41 Genus Pterynotus Swainson . . . . . . .42 Pterynotus newtoni sp. nov. ...... 42 Family BUCCINIDAE ........ 43 Genus Bendeia nov. ........ 43 Bendeia africana (Newton) ...... 43 Genus Laccinum nov. ........ 43 Laccinum lugardi (Newton) ...... 44 Genus Janiopsis Rovereto ....... 44 Janiopsis nigeriensis Newton ..... 44 Family VOLEMIDAE ......... 44 Genus Pseudomazzalina nov. . . . ' . . . .44 Pseudomazzalina nigeriensis (Newton) . . . .45 Family FUSINIDAE ......... 45 Genus Clavilithes Swainson . . . . . . .45 Subgenus Africolithes nov. ...... 45 Clavilithes (Africolithes) africanus (Newton) ... 45 Genus Leucozonia Gray ........ 46 Leucozonia pseudominax sp. nov. . . . .46 Family VOLUTIDAE ......... 46 Genus Volutocorbis Dall ....... 46 Volutocorbis multispinosa (Newton) .... 46 Genus Bendeluta nov. ........ 46 Bendeluta conicoturrita (Newton) ..... 47 Family OLIVIDAE .......... 47 Genus Pseudoliva Swainson . . . . . . .47 Subgenus Buccinorbis Conrad . . . . . -47 Pseudoliva (Buccinorbis) kitsoni (Newton) . . '! . 47 Family STREPSIDURIDAE ........ 4^ Genus Strepsidura Swainson ....... 48 Subgenus Strepdiduropsis nov. ...... 48 Strepsidura (Strepsiduropsis) spirata Newton ... 48 EOCENE MOLLUSCA FROM NIGERIA 27 Page Family CANCELLARIIDAE ........ 48 Genus Sveltia Jousseaume ....... 48 Subgenus Africosveltia nov. ...... 48 Sveltia (Africosveltia} multiplicis (Newton) ... 49 Genus Bonellitia Jousseaume ...... 49 Subgenus Admetula Cossmann ...... 49 Bonellitia (Admetula) amekiensis sp. nov. ... 49 Subgenus Africostoma nov. ...... 49 Bonellitia (Africostoma) decorata (Newton) ... 50 Family CONIDAE ......... 50 Genus Conus Linn6 . . . . . . . .50 Subgenus Leptoconus Swainson ...... 50 Conus (Leptoconus) amekiensis sp. nov. ... 50 Family TURRIDAE . . . . . . . .51 Genus Eopleurotoma Cossmann . . . . . .51 Eopleurotoma nigeriensis Newton . . . . .51 Genus Turricula Schumacher . . . . . .51 Subgenus Surcula H. & A. Adams . . . . .51 Turricula (Surcula) africana (Newton) . . . .51 Genus Surculites Conrad . . . . . . .51 Subgenus Clinura Bellardi . . . . . . .51 Surculites (Clinura) ingens (Mayer-Eymar) . . .51 Genus Mitrelloturris nov. . . . . . . .51 Mitrelloturris casteri (Chavan) . . . . .52 Genus Amekicythara nov. . . . . . . .52 Amekicythara douvillei (Newton) . . . . -53 B. LAMELLIBRANCHIA 53 Family NUCULIDAE ........ 53 Genus Nucula Lamarck . . . . . . .53 Nucula costaeimbricatis Newton . . . . .53 Family GLYCYMERIDAE . . . . . . . .53 Genus Africarca nov. ........ 53 Africarca nigeriensis (Newton) . . . . .53 Family NOETIIDAE ......... 54 Genus Protonoetia MacNeil . * . . . . .54 Protonoetia nigeriensis (Newton) ..... 54 Genus AY cop sis von Koenen . . . . . . .54 Arcopsis africana (Newton) ...... 54 Genus Rectangularca nov. ....... 54 Rectangular ca africana (Newton) . . . . -55 Family PLICATULIDAE . . . . . . . .55 Genus Plicatula Lamarck . . . . . . -55 Plicatula polymorpha Bellardi . . . . -55 Family OSTREIDAE ......... 55 Genus Ostrea Linn6 ........ 55 Ostrea amekiensis sp. nov. . . . . . -55 Ostrea pseudomarginidentata sp. nov. .... 56 Subgenus Crassostrea Sacco ...... 56 Ostrea (Crassostrea) lugardi (Newton) . . . .56 Family CARDITIDAE ........ 57 Genus Glans M. von Muhlfeldt . . . . . -57 Glans nigeriensis sp. nov. . . . . . -57 28 EOCENE MOLLUSCA FROM NIGERIA Page Subgenus Bendeglans nov. . . . . . . 57 Glans (Bendeglans} costaeirregularis (Newton) . . .58 Subgenus Amekiglans nov. ...... 58 Glans (Amekiglans} costaenodulosis (Newton) . . -59 Subgenus Divergidens nov. ...... 60 Glans (Divergidens} triparticostata (Newton) . . .60 Family LUCINIDAE ......... 60 Genus Phacoides Gray ........ 60 Phacoides eaglesomei Newton ..... 60 Genus Gibbolucina Cossmann . . . . . .61 Subgenus Eomiltha Cossmann . . . . . .61 Gibbolucina (Eomiltha} (?} subrhomboidalis (Newton) . 61 Genus Pompholigina Dall . . . . . . .61 Subgenus Eodivaricella Chavan . . . . . .61 Pompholigina (Eodivaricella) oppenheimi (Newton) . . 61 Family CARDIIDAE ......... 61 Genus Fragum (Bolten MS.) Rodin g . . . . .61 Subgenus Africofragum nov. ...... 61 Fragum (Africofragum} newtoni sp. nov. . . .61 Family VENERIDAE ......... 62 Genus Tivelina Cossmann . . . . . . .62 Tivelina newtoni sp. nov. ...... 62 Genus Pitar Roemer ........ 63 Pitar amekiensis sp. nov. ...... 63 Genus Chionella Cossmann . . . . . . -63 Subgenus Costacallista Palmer ...... 63 Chionella (Costacallista} elongatotrigona (Newton) . . 63 Subgenus Microcallista Stewart ...... 64 Chionella (Microcallista) kitsoni (Newton) . .64 Genus Sinodia Jukes-Browne ...... 64 Sinodia heward-belli Newton ..... 64 Genus Sinodiopsis nov. ....... 64 Sinodiopsis coxi sp. nov. ...... 65 Family MACTRIDAE ......... 65 Genus Spisula Gray ........ 65 Subgenus Crepispisula nov. ...... 65 Spisula (Crepispisula} amekiensis sp. nov. . . -65 Family TELLINIDAE ......... 66 Genus Macoma Leach ........ 66 Subgenus Bendemacoma nov. ...... 66 Macoma (Bendemacoma} nigeriensis (Newton) . . 66 Family MYIDAE ......... 67 Genus Raetomya Newton ....... 67 Raetomya schweinfurthi (Mayer-Eymar) . . -67 Family CORBULIDAE ........ 67 Genus Varicorbula Grant & Gale ...... 67 Varicorbula amekiensis sp. nov. . . . . .67 Family KITSONIIDAE nov. ....... 68 Genus Kitsonia nov. ........ 68 Kitsonia eocenica (Newton) ...... 68 III. ACKNOWLEDGMENTS ........ .68 IV. REFERENCES 68 EOCENE MOLLUSCA FROM NIGERIA 29 SYNOPSIS The molluscan fauna described by R. B. Newton (1922) from the Eocene of Bende Ameki (Nigeria) is revised. Fifteen new gastropod species and eight new lamellibranch species are described ; ten new genera and seven new subgenera of gastropoda and four new genera and six new subgenera of lamellibranchia are proposed, one of the lamellibranch genera belonging to a new family. The Eocene age of the fauna is confirmed, but it is believed to be of Upper Eocene (Bartonian), rather than Middle Eocene (Upper Lutetian), age. I. INTRODUCTION AND GENERAL REVIEW DURING the course of exploration in Nigeria by the joint Shell and British Petroleum organization it has frequently been necessary to consult R. B. Newton's work on Eocene Mollusca from Nigeria (1922). Certain problems concerning the age of the beds arose, and these led the writer to investigate the fauna more fully. A detailed study of the material described by Newton, which is in the collections of the British Museum (Natural History), was undertaken, with the result that, while the con- clusions agreed fairly closely with Newton's opinion as to the age, it became apparent that many of the generic determinations required revision, and that not only was only one European species present, but also that all forms compared by Newton with European ones were specifically, and in some cases even generically, distinct. A full description of many of the forms is given in Part II of this paper. On the basis of the fauna he described, Newton concluded that the beds were of Upper Lutetian age, although he recognized that the Bartonian facies was very apparent. The result of the present revision has been to show that only two lamelli- branchia— Plicatula polymorpha Bellardi and Raetomya schweinfurthi (Mayer- Eymar) — occur outside Nigeria, in the Upper Mokattam Beds of Egypt ; the Plicatula has been found also in the Priabonian of Italy, and the Raetomya in the Eocene of Senegal and the Cameroons. Of the gastropod genera, Strepsidura (sensu lato) ranges from Eocene to Oligocene, and Exechostoma and the subgenus Buccin- orbis range from Maestrichtian to Eocene ; not one of the gastropod species has yet been found elsewhere. The cephalopod genus Belosepia is restricted to the Eocene. Although the bryozoan genus Cupuladria, hitherto believed to range only from Miocene to Recent, has been found in the beds by palaeontologists of the Shell Oil Company, the Eocene age indicated by the mollusca is also confirmed by the fish remains recorded, which include the genus Cylindr acanthus. The coral Turbinolia, which is also of fairly common occurrence in the beds, is known only from Eocene and Oligocene deposits. Although two of the lamellibranch species have been recorded from the Upper Mokattam Beds of Egypt, these beds have also yielded Nummulites beaumonti and Orbitolites complanatus ; these latter fossils suggest a rather earlier age than Bartonian (to which stage the Upper Mokattam Beds have been assigned), and appear to suggest that the Upper Mokattam is not entirely of Bartonian age. In Angola, Dartevelle & Roger have recently clearly shown that the beds (Quim- briz Beds) in which Platyodon klinghardti (mistakenly referred to Raetomya schwein- furthi by Caster) and Macrocallista palmerae (to which the Nigerian form here named Sinodiopsis coxi was erroneously referred) are really of Miocene age. Any similarity 30 EOCENE MOLLUSCA FROM NIGERIA of the Angola fauna to that of the Nigerian Bende Ameki Beds is evidently purely superficial, no species being in common. Dartevelle & Roger refer to Raetomya as a mactrid, but Newton has clearly shown that it is a myid ; indeed, Dartevelle & Roger had some doubts about referring the species klinghardti to the American genus Platyodon, and it may well be that it also is a Raetomya. As a result of the revision of the molluscan fauna from Bende Ameki, the writer believes that the evidence and the general relationships of the fauna, while confirming Newton's opinion as to the Eocene age, suggest that an Upper Eocene (Bartonian) rather than a Middle Eocene (Upper Lutetian) age, is indicated. II. SYSTEMATIC DESCRIPTIONS GASTROPODA Family POTAMIDIDAE Genus EXECHOSTOMA Cossmann, 1889 Exechostoma cossmanni Newton 1922. Exechostoma cossmanni Newton, p. 45, pi. 5, figs. 8-9. REMARKS. Specimen G.42I77 (Newton's fig. 9) is now selected as lectotype. Genus TEREBRALIA Swainson, 1840 Terebralia nigeriensis sp. nov. 1922. Terebralia sp.A. Newton, p. 46, pi. 4, fig. 10. MATERIAL. The holotype (G. 42 174). DESCRIPTION. Apical whorls missing and aperture broken. Of medium size, turriculate-conic, spire angle 19°. Whorls (of which four or five are preserved) almost flat-sided, their height very slightly more than half their width ; sutures deeply impressed, distinctly undulatory. Ornament consisting of moderately coarse and fairly widely-spaced ribs crossed by strong spiral threads ; ribs somewhat irregularly spaced, sometimes only slightly narrower than, but sometimes (especially on the later whorls) only half the width of, their intervals, gently opisthocline adapically, their forward-directed face slightly concave, more strongly so abapically, where they tend to be slightly swollen. On the whorls preserved there are n strong spiral threads with intervals of about the same width. There is no noticeable development of varices. Abapical portion of last whorl poorly preserved, without ribs, but with about nine incised spiral grooves separated by intervals of approxi- mately half their width. Columella moderately concave and callous ; other apertural characters not determinable. DIMENSIONS. Height (incomplete) 49-7 mm. Diameter of last whorl 21-0 mm. REMARKS. The above characters are sufficient for specific determination. EOCENE MOLLUSCA FROM NIGERIA 31 Terebralia amekiensis sp. nov. 1922. Terebralia sp.B. Newton, p. 48, pi. 4, fig. n. MATERIAL. The holotype (6.42175). DESCRIPTION. Apical whorls missing and aperture broken, about five whorls preserved. Of small-medium size, turriculate-conic, spire angle i7°-i8°. Whorls flat-sided, their height about two-fifths of their width ; sutures linear and moder- ately deep, gently undulatory. Ornament consisting of relatively fine and closely- spaced ribs crossed by strong spiral threads ; ribs of about the same width as their intervals, their forward-directed face sometimes gently concave, vertical or slightly opisthocline abapically. On the whorls preserved there are eight to ten fairly strong spiral threads with intervals of about the same width ; no noticeable development of varices. Base broken, without ribs, but carrying strong spiral threads with intervals of about the same width. " DIMENSIONS. Height (incomplete) 38-4 mm. Diameter of last whorl 15-3 mm. REMARKS. The above characters are sufficient to define the species. Family TURRITELLIDAE Genus TURRITELLA Lamarck, 1799 Turritella amekiensis sp. nov. (PI. 5, figs, i, 2) 1922. Turritella cf. sulcifera Deshayes : Newton, p. 50, pi. 5, fig. 7. MATERIAL. Numerous specimens, including the holotype (G. 42215, Newton's fig. 7)- DESCRIPTION. Of medium size, turriculate-conic, spire angle 15°, last whorl about one-eighth of the total height. Protoconcti not preserved, evidently small. Whorls distinctly convex, their greatest convexity slightly below the middle, upper slope less convex than the lower. Sutures linear. Height of whorls about seven- twelfths of their width. The earliest whorls seen carry four spiral threads, the number increasing to about 18 on the last spire whorl, on which a few very fine additional threads are present in some of the intervals. Base of last whorl with threads of a similar type. Columella gently concave ; columellar lip narrow, especially abapically. Aperture evidently rounded-subquadrate. No siphonal fascicle. Growth lines concave forwards, distinctly prosocline adapically, very slightly prosocline abapically, the maximum concavity at about three-fifths the height of the whorl. DIMENSIONS. Height 27-4 mm., width 7-5 mm. REMARKS. The Nigerian form is consistently smaller than T. sulcifera Deshayes of the Paris Basin Lutetian ; its whorls do not tend to be concave above, and the details of the spiral ornament are different. 32 EOCENE MOLLUSCA FROM NIGERIA Subgenus COELOCONICA nov. TYPE SPECIES. Turritella mauryana Newton. SUBGENERIC CHARACTERS. Of moderate size to fairly large, distinctly coeloconoid. Protoconch not preserved, evidently small. Early whorls subcylindrical to slightly convex, with three sharp spiral threads the uppermost of which is fairly close to the suture ; traces of a fourth thread at the lower suture ; all four threads with sharp, closely-spaced, vertically diposed crenulations where crossed by growth lines ; rest of surface with microscopic spirals. With growth the third thread from the adapical suture gradually develops into a very prominent, sharp flange. Base of last whorl with a second prominent, but feebler, keel and two obscure threads below, as well as microscopic spirals. Aperture rounded-subquadrate, as in Turritella. Growth-lines with a deep U-shaped sinus the apex of which is slightly above the second keel ; markedly prosocline at the adapical suture, orthocline at the abapical suture, their lower end immediately below the upper. REMARKS. The high, gently coeloconoid spire, the details of ornament (including development of the strong flange), and the form of the growth-lines differentiate this from other described groups of Turritella. Turritella (Coeloconica) mauryana Newton 1922. Turritella mauryana Newton, p. 48, pi. 5, figs. 4-6. MATERIAL. Several specimens, including the lectotype (G. 42209, Newton's fig. 4) here selected. Family ARCHITECTONICIDAE Genus ARCHITECTONICA (Bolten MS.) Roding, 1798 Subgenus NIPTERAXIS Cossmann, 1915 Architectonica (Nipteraxis) bendeica sp. nov. (PI. 5, figs. 3«-c) 1922. Solariaxis cf. canaliculata (Lamarck): Newton, p. 54. MATERIAL. The holotype (G.42373) and a few other specimens. DESCRIPTION. Moderately small, low-conic to slightly cyrtoconoid, apical angle I4O°-I45°. Protoconch smooth, anastrophic ; 3-4 flattened, very slightly convex spire whorls with deep, sunken sutures. Ornament consisting of five crenulated spiral threads, the uppermost one the most prominent and with the strongest crenulations, the others subequal. Last whorl sharply rounded at the periphery where there is a sixth and broader cordon, and with a fine subsidiary thread in each interval ; on large specimens a second order of intercalaries may appear. Base gently convex, with seven crenulated or beaded spiral threads, the outer three finer and moderately widely spaced, the inner three coarser and with coarser beading, the innermost one well within the umbilicus ; blunt, accentuated growth lines join EOCENE MOLLUSCA FROM NIGERIA 33 the crenulations. Umbilicus deep, wide and completely visible, occupying about two-fifths of the diameter of the base. Growth lines gently prosocline, slightly sinuous, becoming orthocline at the upper suture. Aperture rounded kite-shaped, peristome discontinuous ; columella thin, slightly excavated, with two small furrows corresponding to the two innermost spiral threads on the base, the upper furrow more distinct. DIMENSIONS. Holotype : height 5-7 mm., width 12-0 mm. A topotype (0.42374) has a width of 13-8 mm. REMARKS. Compared with A. canaliculata (Lamarck), this form lacks the sharp peripheral keel, and the details of the ornament are quite distinct. Subgenus STELLAXIS Ball, 1892 Architectonica (Stellaxis) bicingulata (Newton) (PL 5, ng. 4) 1922. Stellaxis bicingulata Newton, p. 52, pi. 5, figs. 10-11. MATERIAL. Numerous specimens, including the lectotype (G. 42355, Newton's fig. 10) here selected. SUPPLEMENTARY DESCRIPTION. Of medium size, solarioid, spire slightly cyrto- conoid ; spire angle decreasing during growth from about 125° to about a right angle or slightly less ; last whorl forming slightly more than one-third of the height. Protoconch smooth and loosely coiled, anastrophic. About five spire whorls, which are gently convex, the main surface separated from a narrow but distinct, smooth, abapical sutural cord by a spiral groove ; main surface smooth in the earlier stages, later developing extremely vague, irregular spiral threads. Sutures linear. Growth lines almost straight, prosocline. Marginal keel on last whorl with a few very faint threads, occasionally appearing to be slightly bifid. Base gently convex medially, flatter marginally where there is one smooth spiral cord close to the peripheral keel. Umbilicus very deep, wide and completely visible, occupying about a third of the diameter of the base, margined by blunt teeth which are sometimes made to appear vaguely bifid by the accentuated growth-lines, its inner wall vertical and with a spiral thread at about two-thirds of its height ; a fine spiral groove limits the teeth externally in young forms, but is often obsolete or absent in adults. Aperture kite-shaped, with a small indentation level with the row of umbilical teeth, and a still fainter one by the umbilical thread above. Subgenus SOLARIAXIS Dall, 1892 Architectonica (Solariaxis) amekiensis sp. nov. (P. 5, figs. $a-c) 1922. Solariaxis cf. spectabilis (]. de C. Sowerby): Newton, p. 53, pi. 5, figs. 12-13. MATERIAL. A few specimens, including the holotype (0.42361, Newton's fig. 12). DESCRIPTION. Of medium size, moderately low, gently cyrtoconoid, spire angle 34 EOCENE MOLLUSCA FROM NIGERIA decreasing during growth from about 130° to about 105° (excluding the peripheral downturning of the last whorl). Protoconch not well preserved, anastrophic. About five flattened spire whorls with deep sutures, slightly imbricate. Ornament consisting of seven rather finely crenulated, not very coarse spiral threads, the fourth and sixth (from the adapical suture) finer than the others, the second, third and fifth coarser, the first even a little coarser than these. Last whorl bluntly angular at the periphery, its upper four primary threads rather wide-spaced, with a faint double intercalary, a fine single intercalary, and a stronger double intercalary in the upper, median and lower intervals, respectively ; the five lower primary threads more closely-spaced, the upper interval with a fine intercalary, the peripheral thread double. Base flattened, only very slightly convex, with seven crenulated or beaded spiral threads increasing in strength towards the umbilicus, the wall of the latter with two finer, widely-spaced threads ; growth lines serrate but not thickened. Umbilicus deep, wide and completely visible, occupying about a third of the diameter of the base. Growth lines distinctly prosocline, fairly straight, but becoming orthocline at the adapical suture. Aperture oval, kite-shaped ; peristome discontinuous. Columella thin, gently concave, with four slight furrows corresponding to the two umbilical threads and the two innermost spiral threads on the base. DIMENSIONS. Holotype : height 13-4 mm., width 22-8 mm. REMARKS. Compared with that of A. (Solariaxis) spectabilis (J. de C. Sowerby) the spiral ornament is coarser and differently disposed. Family SCALIDAE Genus ACHILLA H. Adams, 1860 Acrilla nigeriensis sp. nov. (PI. 5, ng. 6) 1922. Acrilla cf. affinis (Deshayes) : Newton, p. 50, pi. 3, figs. 10-12. MATERIAL. The holotype (0.42285, Newton's fig. 10), and a few other specimens. DESCRIPTION. Specimens incomplete, apical whorls missing and aperture broken. The largest specimen (the holotype) has about eight whorls preserved, spire angle 11° ; they are rather loosely coiled, with deep, linear sutures, and are very strongly convex, with a tendency to angularity at about three-quarters or more of their height. Ornament consisting of fine, sharp, prominent axial ribs with subdued spiral threads in the intervals. The ribs are considerably narrower than their interspaces, occasion- ally slightly varicose, straight or gently concave forward, gently prosocline, and more strongly bent forward at the adapical suture. There are some 30-35 vague, irregular spirals on the last-preserved whorl of the holotype. A topotype (G. 42286) shows a basal disk limited by a fairly sharp carina ; the ribs continue over the disk, but are much flattened, and there are some 20 spirals similar to those on the spire whorls. Aperture oval, a little higher than wide ; columellar lip slightly concave, more callous abapically, where a small auricle is developed. EOCENE MOLLUSCA FROM NIGERIA 35 REMARKS. This species was compared by Newton with A. affinis (Deshayes) of the Paris Basin Eocene, but is distinctly more aciculate, the spire angle being smaller. The whorls are distinctly more convex, and the sutures more deeply sunk, the shell being more loosely coiled. Family CALYPTRAEIDAE Genus CALYPTRAEA Lamarck, 1799 Calyptraea newtoni sp. nov. (PI. 5, figs. 7a, b) 1922. Calyptraea crepidularis Lamarck : Newton (pars), p. 59. MATERIAL. The holotype (0.42438) and one topotype (0.42439). DESCRIPTION. Of moderate size, calyptraeiform, irregularly conic. Protoconch small, smooth, dextral, Nerita-like, obliquely set. Surface ornamented, at least in the later stages, with oblique threads bearing papillae, the threads becoming more nearly parallel to the apertural margin as they approach it. Aperture subcircular, with margin lying in one plane, septum with very concave edge. DIMENSIONS. Height 4-7 mm., width 12-8-13-5 mm. REMARKS. C. crepidularis Lamarck, of the Paris Basin Eocene, with which Newton identified this form, is much flatter, is subrectangular in outline, has a markedly eccentric apical region, and does not possess oblique threads bearing papillae. Genus TURBOCALYPTRAEA nov. TYPE SPECIES. T. scabrosa sp. nov. GENERIC CHARACTERS. Of medium size to moderately small, calyptraeiform, tending to turbinate, apex markedly eccentric. Whorls strongly convex, with several spirally disposed rows of short, hollow, forward-directed spines together with obliquely disposed threads. Aperture as in Calyptraea, subcircular to suboval. A distinct, sunken, gently concave septum, the margin of which is noticeably concave in its upper part and gently convex in its lower part, occupies nearly half the aper- ture. A small, deep umbilicus is developed in the upper part of the septum. REMARKS. The form, ornament, character of the septum, and presence of an umbilicus together characterize this new genus. Turbocalyptraea scabrosa sp. nov. (PI. 5, figs. 8, ga, b) 1922. Calyptraea crepidularis Lamarck : Newton (pars), p. 59, pi. 4, figs. 18-19. MATERIAL. A few specimens, including the holotype (0.42437, Newton's fig. 19). DESCRIPTION. As above. REMARKS. C. crepidularis Lamarck, from the Eocene of the Paris Basin, with which Newton identified this species as well as the last, differs markedly in its almost flat 36 EOCENE MOLLUSCA FROM NIGERIA form and subrectangular outline, its very small spire, its lack of coarse ornament, and in the characters of the septum. Genus CREPIDULA Lamarck, 1799 Subgenus CONCAVIMARGO nov. TYPE SPECIES. Crepidula falconeri Newton. SUBGENERIC CHARACTERS. Like Crepidula, but teleoconch whorls not in contact ; surface smooth ; body cavity deep ; aperture oval ; septum rather deeply sunk and with a cavity extending underneath the inner lip ; edge of septum distinctly concave. Crepidula (ConcavimargoJ falconeri Newton « 1922. Crepidula falconeri Newton, p. 58, pi. 2, figs. 13, 130. MATERIAL. Several specimens, including the lectotype (0.42411, Newton's fig. 13) here designated. Family XENOPHORIDAE Genus XENOPHORA Fischer von Waldheim, 1807 Xenophora nigeriensis (Newton) (PI. 6, figs. la-c) 1922. Tugurium nigeriense Newton, p. 51, pi. 4, figs. 20-21. MATERIAL. Numerous specimens, including the lectotype (0.42291, Newton's fig. 21) here designated. REMARKS. The presence of numerous agglutinated objects covering most of the whorl surface indicates that the species is better regarded as a Xenophora ; an umbilicus as small as the one it possesses may occur in this genus. Family STROMBIDAE Genus TIBIA (Bolten MS.) Roding, 1798 Tibia bidigitata (Newton) 1922. Rostellaria bidigitata Newton, p. 12, pi. 4, figs. 8-9. MATERIAL. Several specimens, including the lectotype (0.41643, Newton's fig. 8) here designated. Genus CYRTULOTIBIA nov. TYPE SPECIES. Rostellaria unidigitata Newton. GENERIC CHARACTERS. Form somewhat like that of a Tibia with a very short, inclined siphonal canal ; often developing a strong shoulder on the last whorl and thus recalling Cyrtulus. Protoconch conic, of two or three smooth, moderately EOCENE MOLLUSCA FROM NIGERIA 37 convex whorls. Spire like that of Tibia, conic, of five or six gently convex whorls ; early stages with fine axial riblets, which are orthocline above and opisthocline below (i.e., concave forwards), crossed by increasingly broad spiral threads ; coarse, swollen varices developed occasionally. On later spire whorls the axial riblets be- come obsolete and are represented by accentuated growth lines only. Last whorl slightly to very strongly shouldered, the more strongly shouldered specimens with a narrow callous band (an extension of the callus of the posterior sinus) extending back some one and a half to two whorls along the sutural region. Last whorl smooth except for spiral threads on the base, rather conic, base very slightly excavated, neck oblique. Rostrum short, curved to the right. Aperture oval, with a narrow, slit-like posterior sinus curving back on to the suture ; columellar lip callous, developing a prominent, raised knob of callus limiting the inner side of the posterior sinus ; a broad notch to the right of the rostrum is delimited on the right by a short spine. Outer lip rather thick, not varicose, internally smooth, with no additional spines, gently parasigmoidal, distinctly opisthocline as a whole. Remarks. The general form, short inclined rostrum, single abapical labial spine, the contour of outer lip, broad abapical notch, and extremely strong adapical parietal callus (extending back along the suture for two or three whorls) readily distinguish this genus from Tibia. Cyrtulotibia unidigitata (Newton) 1922. Rostellaria unidigitata Newton, p. 14, pi. 4, figs. 3-7. MATERIAL. Many specimens, including the lectotype (0.41688, Newton's fig. 3) here selected. Genus SEMITEREBELLUM Cossmann, 1889 Subgenus AFRICOTEREBELLUM nov. TYPE SPECIES. Semiterebellum elongatum Newton. SUBGENERIC CHARACTERS. Form much like that oT Terebellum, but more narrowly fusiform. Protoconch consisting of some three to four smooth, moderately convex whorls, less acute than the shell as a whole. About six spire whorls, at first gently convex, becoming flatter with growth. Ornament in the earlier stages consisting of a sharp, fine adapical thread finely crenulated by growth lines, with a narrow, shallowly excavated band below it, the abapical half of the remaining portion of the whorls with four to six fine, incised spiral lines ; ornament obsolete on later whorls. Last whorl with numerous irregular spiral threads (finer and with broader intervals abapically) on the base, which is only vaguely concave ; neck short, gently swollen. Rostrum extremely short, its end barely projecting more than the outer lip abapically. Aperture oval-subtriangular, narrower adapically, with a broad, shallow notch to the right of the rostrum. Columellar callus thin, especially medially, with a low, oblique ridge adapically, forming the upper edge of the short posterior sinus which extends only slightly above and back along the suture. Outer lip opisthocline as a 38 EOCENE MOLLUSCA FROM NIGERIA whole, forming a strongly projecting, rounded lobe to the right of the abapical channel, thin, not varicose, internally smooth. REMARKS. This subgenus differs from Semiterebellum (sensu stricto) in being more lanceolate, in the less projecting rostrum, in the posterior sinus which does not ascend partly up the spire, and in the more opisthocline outer lip which is more lobate below. Semiterebellum (Africoterebellum) elongatum Newton (PL 5, figs. ioa, b) 1922. Semiterebellum elongatum Newton, p. 17, pi. 2, figs. 14-15. MATERIAL. Numerous specimens, including the lectotype (0.41762, Newton's fig. 15) here selected. Genus AMEKICHILUS nov. TYPE SPECIES. — Semiterebellum suturocostatum Newton. GENERIC CHARACTERS. Of small-medium to medium size, having the general form of Ectinochilus. Protoconch helicoid-trochoid, consisting of about four smooth, moderately convex whorls. Spire gently cyrtoconoid ; last whorl slightly more than half the height of the shell. Whorls only slightly convex, sutures distinct ; five to six spire whorls. An incised line separates a narrow juxtasutural band which is always crenulated or beaded in the early stages ; later, the band is either smooth or beaded, and may be vaguely bifid. Last whorl oval, base declivous ; neck short, not swollen. Rostrum short, scarcely projecting, inclined gently to the right, with a broad, very shallow notch to its right, the notch limited externally by a short, sharp spine. Aperture oval, rather small, with a long, narrow, callous posterior sinus which curves over and slightly down on to the suture of the last whorl, extending only a short distance back. On those specimens which have a heavily beaded juxtasutural thread, the upper part of the posterior sinus extends back right up the spire as a gently convex callous band occupying the lower half of the whorls. Columella gently concave, smooth, with thick callus developing an even thicker ridge along the upper side of the posterior sinus. Outer lip thin, varicose, internally smooth, but with a slight internal thickening, externally strongly varicose, especially at the posterior sinus, nearly straight and orthocline, slightly convex adapically. Ornament consisting of a few incised spiral lines with minute pits, widely spaced posteriorly, often obsolete on the middle of the last whorl, changing to more closely spaced spiral threads on the base. REMARKS. This form is shorter and more oval than Semiterebellum, and has a less projecting rostrum, a distinct juxtasutural band, and an abapical labial spine ; the posterior sinus does not ascend above the suture of the last whorl. Its closest relatives seem to be the Ectinochilus and Dientomochilus group of shells, but there are no obvious varices and the posterior sinus curves over at the suture of the last whorl. The posterior sinus in Africoterebellum is similar, but in that genus the form of the shell, ornament and labial spine are quite different. EOCENE MOLLUSCA FROM NIGERIA 39 Amekichilus suturocostatum (Newton) 1922. Semiterebellum suturocostatum Newton, p. 15, pi. 4, figs. 14-17. MATERIAL. Many specimens, including the lectotype (0.41757, Newton's fig. 14) here selected. Family AMPHIPERATIDAE Genus EOVOLVA Schilder, 1932 Eovolva nigeriensis (Newton) (PI. 6, figs. 2, 3) 1922. Amphiperas nigeriensis Newton, p. 18, pi. 3, figs. 14-15. 1932. Eovolva nigeriensis (Newton) : Schilder, p. 212. MATERIAL. Several specimens, including the lectotype (0.41786, Newton's fig. 14) here selected. REMARKS. This is the monotype of Eovolva. Genus SPHAEROCYPRAEA Schilder, 1927 Sphaerocypraea sudanensis (Schilder) (PI. 6, figs. 40, b) 1922. Cypraea cf. bowerbanki J. de C. Sowerby : Newton, p. 18, pi. 3, fig. 13. 1929. Sphaerocypraea bowerbankii (J. de C. Sowerby) : Schilder, p. 305. 1932. Eocypraea (Sphaerocypraea) bowerbanki (J. de C. Sowerby) var. sudanensis Schilder, p. 218. MATERIAL. A few specimens, including the lectotype (0.41780, Newton's fig. 13) here selected. Remarks. In spite of its general similarity to 5. bowerbanki, Schilder (1929) sus- pected that the Nigerian form was probably specifically distinct. The details of the fossula cannot be compared as they are not seen in available British specimens, but the outer lip of the Nigerian form is somewhat broader and is distinctly more callous and margined, and the shell was evidently a little more globose. It seems advisable to regard the Bende Ameki specimens as constituting a distinct species for which Schilder's name must be adopted, in spite of its unsuitability. Family NATICIDAE Genus NEVERITA Risso, 1826 Neverita amekiensis sp. nov. (PI. 6, figs. 50, b) 1922. Neverita cf. calvimontana (Deshayes): Newton, p. 55, pi. 5, figs. 16-17. MATERIAL. The holotype (G. 42383). 40 EOCENE MOLLUSCA FROM NIGERIA DESCRIPTION. Of small-medium size, having the form of Polinices rather than Neverita, a little less flattened than is usual in the latter genus. Protoconch conic, very low, of about two smooth, gently convex whorls. Last whorl forming about four-fifths of the height of the shell. Spire consisting of about two and a quarter almost flat, smooth whorls with fine linear sutures. Last whorl very large, rather flattened adapically, its flank moderately sharply rounded ; base declivous. Aper- ture semilunar, not much produced to the right. Columella straight, its callus very heavy, a solid, gently convex plug almost, but not quite, filling the umbilicus. REMARKS. In N. calvimontana (Deshayes), from the Eocene of the Paris Basin, the umbilical callus is less extensive and solid, the whorls are more distinctly convex, and the sutures are not quite linear. Genus SINUM (Bolten MS.) Roding, 1798 Sinum africanum Newton (PI. 6, figs. 6a-c) 1922. Sinum africanum Newton, p. 57, pi. 4, figs. 12-13. MATERIAL. Several specimens, including the lectotype (0.42406, Newton's fig. 12) here selected. Sinum nigeriense sp. nov. (PI. 6, figs. 7«, b) 1922. Sinum cf. clathratum (Gmelin): Newton, p. 56, pi. 2, figs. 16-17. MATERIAL. Several specimens, including the holotype (0.42390, Newton's fig. 16). DESCRIPTION. Of small-medium size, not very thin-shelled, not very auriform, outline more like that of Polinices. Protoconch conic, very low, of two smooth, slightly convex whorls with a small nucleus. One and a quarter almost flat spire whorls with 13-15 wavy, fine spiral threads the course of which is slightly deflected at each growth line ; up to three microscopic spirals developed in the intervals on the penultimate whorl. Last whorl very large, occupying most of the height of the shell, adapical portion rather extensive and flattened, flanks rounded, base declivous ; ornament as on the spire whorls, but primary threads more numerous (approximately 40 in number). Aperture rounded-subquadrate, somewhat produced abapically and to the right. Umbilical callus narrow, but distinct, leaving a very small umbilical opening. Outer lip markedly prosocline and gently convex. REMARKS. S. clathratum is more auriform and compressed, has a lower spire and a more ample aperture, and does not have the adapically flattened whorls which in the new species produce a rather conic appearance. Family CYMATIIDAE Genus VARICOHILDA nov. TYPE SPECIES. Hilda turriculata Newton. GENERIC CHARACTERS. Of small-medium size, having the general form of Hilda. EOCENE MOLLUSCA FROM NIGERIA 41 Protoconch not well preserved, evidently smooth and naticoid. Last whorl consti- tuting about half the height of the shell. Spire consisting of four to five whorls which are about half as high as wide and carry a quite sharp median angulation ; ornament consisting of narrow, rather widely-spaced, straight, orthocline or slightly opistho- cline axial ribs crossed by fairly prominent spiral threads which are a little broader than their intervals ; no varices on spire whorls. Last whorl ornamented like the spire whorls, with the addition of fine intercalary spiral threads near the shoulder and a very strong varix 240° back from the outer lip ; base excavated abapically, the ribs becoming fainter but the spiral ornament persisting ; neck short, swollen, inclined to the left. Aperture narrow and parallel-sided, rather like that of Anachis, a fairly broad, oblique adapical part limited by a spiral parietal fold ; a fairly short, narrow siphonal canal, slightly inclined to the left, has a deep lateral notch at the end. Siphonal fasciole swollen. Outer lip almost orthocline, with a very strong varix slightly behind its sharp edge, internally thickened and with long lirae. Columellar lip straight and vertical, with three strong, spiral columellar folds, forming angular junctions with the parietal lip and with the siphonal canal ; callus well-developed, spreading a little over the base, distinctly limited, becoming slightly detached by the siphonal fasciole, with numerous knobs and transverse wrinkles between its outer margin and the columellar and parietal folds. REMARKS. Although generally resembling Hilda, this genus differs in having a slightly shorter spire, angular whorls, a strong varix in addition to the labral varix on the last whorl, a narrower and more parallel-sided aperture, the canal inclined to the left, and three strong columellar folds in addition to wrinkles and knobs on the inner lip. Varicohilda turriculata (Newton) (PI. 7, figs. la-c) 1922. Hilda turriculata Newton, p. 29, pi. 4, figs. 24-25. MATERIAL. Many specimens, including the lectotype (0.41971, Newton's fig. 24) here selected. v Family MURICIDAE Genus HEXAPLEX Perry, 1811 Subgenus PAZIELLA Jousseaume, 1880 Hexaplex (Paziella) bendeica sp. nov. 1922. Poirieria cf. calcitrapa (Lamarck) : Newton, p. 31, pi. 3, figs. 22-23. MATERIAL. The holotype (0.41983, Newton's fig. 22). DESCRIPTION. Of small-medium size, fusiform, ribs aligned in seven irregular axial series, each rib being, on the last three whorls, slightly behind that on the whorl above. Protoconch not preserved. Last whorl forming about half the height of the shell. Spire conic, consisting of five to six whorls which are distinctly angulated medially, irregularly flattened and shelving above, subcylindrical or even slightly GEOL. Ill, 2. c 42 EOCENE MOLLUSCA FROM NIGERIA inturned below, with linear, wavy sutures. Adapical shelf with two or three very vague spirals near the middle ; one spiral thread on the angulation, and one close to the abapical suture, the latter on later whorls with a faint spiral just above it. Ribs nodular and subspinose at the shoulder on early whorls, a little narrower than their intervals, becoming considerably narrower than the intervals with growth and also developing short upturned spines at the shoulder, the spines being channelled on the forward-facing side. Last whorl inflated, base well excavated, neck moderately long ; siphonal fascicle bulging, carrying widely-spaced scales. Three primary threads and vague intercalaries on the flank of the last whorl. Aperture oval, with no distinct adapical channel, with a moderately long, narrow, gently curved siphonal canal which is gently inclined to the left. Columella gently excavated, smooth, twisted, and with a vague fold at the beginning of the canal. Columellar lip callous, not widely spread, becoming detached abapically, leaving a small false umbilicus be- tween it and the siphonal fascicle. Outer lip thin, with a varix close behind it, gently parasigmoidal and definitely prosocline adapically, with a deep lateral notch at the spine on the shoulder, internally thickened, dentate below the notch. Growth lines serrate on forward-facing side of varices. REMARKS. H. (P.) calcitrapa has a much more inflated last whorl and a lower spire. Poirieria has five, not seven, axial rows of varices. Genus PTERYNOTUS Swainson, 1833 Pterynotus newtoni sp. nov. 1922. Pteropurpura cf. tricarinata (Lamarck) : Newton, p. 30, pi. 4, figs. 26-27. MATERIAL. The holotype (G.4I982, Newton's figs. 26-27). DESCRIPTION. Of medium size, fusiform, markedly triangular when viewed from above on account of three rows of lamellar, non-spinose varices, each varix being slightly behind the corresponding one on the preceding whorl. Protoconch (nucleus missing) apparently rather tectiform, of about three smooth, gently convex whorls. Last whorl forming slightly more than half the height of the shell. Spire conic, consisting of nearly five convex whorls with linear, undulatory sutures. Early whorls carry two ribs, later whorls only one, between the varices ; ribs nodular and crossed by muricate spiral threads, three orders of which are present on the last whorl ; serrate, rather widely-spaced growth threads are also present. Last whorl inflated ; base well excavated, neck long and straight except for the protuberant, tubular siphonal fasciole. Aperture oval, with an abapical channel and a long, narrow siphonal canal which is slightly longer than the height of the aperture, inclined to the left but vaguely curved towards the right, and then curved well back at the tip. Columella gently concave, twisted at the start of the canal. Inner lip callous, a little wider adapically, rather detached medially, well detached abapically, smooth except for a spiral ridge limiting the adapical channel. Outer lip with a broad, leaf-like, non-spinose varix, orthocline, its edge thin and dentate, internally thickened and dentate. Growth lines serrate on forward-facing side of varices. EOCENE MOLLUSCA FROM NIGERIA 43 REMARKS. Pterynotus (Pteropurpura) tricarinatus (Lamarck) has the varices up- turned and spinose adapically. Family BUCCINIDAE Genus BENDEIA nov. TYPE SPECIES. Liomesus africanus Newton. GENERIC CHARACTERS. Of small-medium size, buccinoid, intermediate in general form between Cominella and Strepsidura, oval-conic. Protoconch naticoid, of two to two and a half smooth, moderately convex whorls. Last whorl forming slightly less than three-quarters of the height of the shell. Spire consisting of about two and a half gently convex whorls with conspicuous, deep sutures, slightly stepped ; two incised spiral lines close to the adapical suture define two raised cords. Last whorl inflated-oval, ornamented like the spire whorls, base well excavated and with 11-13 spiral threads which are narrower and stronger abapically ; neck moderately long and swollen. Aperture rather narrowly oval, with an adapical channel, and with a siphonal canal of moderate length which is inclined to the left and notched. Colu- mellar callus not thick or extensive ; columella with a fairly strong fold limiting the canal, and a series of wrinkles where the spiral threads of the base pass under the columellar callus. Siphonal fasciole only moderately swollen, carrying five to eight longitudinal threads, limited above by a strong, sharp, raised thread. Outer lip thin, internally smooth, almost orthocline, slightly prosocline adapically. REMARKS. Although superficially resembling Liomesus, this genus is more strep- siduriform (i.e. it has a shorter spire, and is more produced abapically), has a longer and more inflected canal, a less callous inner lip, and a distinct siphonal fasciole limited above by a fine, raised carina ; moreover, spiral ornament is present on the upper part of the whorls, the base of the last whorl is more excavated, the aperture is narrower, and there is a distinct columellar fold in an abapical position. The relationships evidently lie more with Cominella than with Liomesus or Strepsidura. Bendeia africana (Newton) (PI. 7, figs. 2a, b) 1922. Liomesus africanus Newton, p. 38, pi. 3, figs. 20-21. MATERIAL. Several specimens, including the lectotype ^.42096, Newton's fig. 21) here selected. Genus LACCINUM nov. TYPE SPECIES. Athleta lugardi Newton. GENERIC CHARACTERS. Attaining a large size, thick-shelled, conic to buccinoid, with a low spire. Protoconch (worn in available specimens) evidently not large and bulbous as in many Volutidae. Shell completely smooth except for growth lines which are somewhat accentuated on the siphonal fasciole, although extremely vague ; fine spirals can sometimes be distinguished. Last whorl forming about ten-thirteenths 44 EOCENE MOLLUSCA FROM NIGERIA (younger specimens) to seven-eighths (larger specimens) of the height of the shell. Spire distinctly coeloconoid, composed of about five flattened, very gently convex whorls with distinct, linear sutures, relatively narrower in later stages of growth. Last whorl very large, its shoulder even more sharply rounded in large specimens than in juveniles, flank subcylindrical, base slightly excavated, neck moderately long and swollen by the siphonal fasciole. Aperture narrowly oval, with a broad, flat channel situated adapically at the shoulder, and with a fairly short, poorly- defined siphonal canal which is slightly inclined to the left and deeply notched. Outer lip thin, smooth internally, orthocline as a whole, gently bisinuous. Columella gently concave, vaguely bent at its junction with the canal ; inner lip with a thick, moderately wide layer of callus which becomes slightly detached abapically at the extensive siphonal fasciole, and which develops a large, protruding knob of callus beside the adapical channel. No columellar folds. REMARKS. The above characters indicate that this genus is not related to Athleta, nor even a volutid. It is evidently a buccinid, and seems best placed near Lacinia, from which it differs in its more cylindrical last whorl, the lack of an umbilicus, and its less ample aperture and better defined siphonal canal. Laccinum lugardi (Newton) 1922. Athleta lugardi Newton, p. 25, pi. 5, figs. 1-3. MATERIAL. Several specimens, including the lectotype (0.41860, Newton's fig. i) here selected. Genus JANIOPSIS Rovereto, 1899 Janiopsis nigeriensis Newton 1922. Janiopsis nigeriensis Newton, p. 41, pi. 4, figs. 22-23. MATERIAL. Several specimens, including the lectotype (0.42131, Newton's fig. 22) here selected. Family VOLEMIDAE Genus PSEUDOMAZZALINA nov. TYPE SPECIES. Bulbifusus nigeriensis Newton. GENERIC CHARACTERS. Of large-medium size, rather thin-shelled, inflated-fusiform, entirely smooth. Protoconch unknown. About six moderately convex spire whorls which are broadest a little below the middle, height about two-fifths of the width. Sutures linear. Last whorl forming one-half to three-fifths the height of the shell, inflated, base evenly excavated, neck moderately long, straight, vertical. Aperture oval, with a narrow adapical channel, vaguely constricted abapically where it is extended into a moderately oblique and wide, notched siphonal canal the length of which is about one-half the height of the aperture proper. No siphonal fasciole. Columella gently excavated, rather vaguely bent at the beginning of the canal, with a narrow, thin layer of callus, without columellar folds. Outer lip thin, strongly EOCENE MOLLUSCA FROM NIGERIA 45 convex in its median and anterior parts, receding strongly adapically at an angle of about 45° and becoming nearly orthocline close to the suture. Some 10-12 irregular, elongate lirae are developed well inside the aperture in its upper half. REMARKS. Compared with Mazzalina (of which Bulbifusus is a synonym) this genus is less inflated and more fusiform, and has a higher spire, a narrower aperture and canal, no columellar folds, no spiral ornament on the base, and an outer lip which is deeply excavated adapically. Levifusus is distinctly ornamented and has a longer canal. Sycostoma is less fusiform and has heavier columellar callus ; its outer lip is less excavated adapically. Pseudomazzalina nigeriensis (Newton) 1922. Bulbifusus nigeriensis Newton, p. 35, pi. 4, figs. 1-2. MATERIAL. Several specimens, including the lectotype (0.42057, Newton's fig. 2) here selected. Family FUSINIDAE Genus CLAVILITHES Swainson, 1840 Subgenus AFRICOLITHES nov. TYPE SPECIES. Rhopalithes africanus Newton. SUBGENERIC CHARACTERS. Of large-medium size, fusiform. Protoconch not pre- served. Last whorl forming about half the height of the shell. Spire conic (apex rather cyrtoconoid), of about eight gently convex whorls, with a slight spiral depression just below the suture. First five or six whorls with moderately fine spiral threads crossing strong nodular ribs which are not in alignment from whorl to whorl. In the later stages the ribs become obsolete adapically, and on the last two whorls are absent completely, the spiral threads also becoming feebler. Base of last whorl excavate ; neck long, straight, vertical. No siphonal fascicle. Imperforate. Aper- ture oval, with a small adapical channel and a long, straight siphonal canal only slightly inclined to the left. Outer lip thin, broadly concave, internally smooth. Columellar callus not widely spread, detached externally. Columella straight, joining the canal without any twist. No columellar folds. REMARKS. Clavilithes is considerably less fusiform. Rhopalites has rather shouldered whorls, a more oblique canal, and a heavier adapical apertural callus, and it tends to have a siphonal fascicle and small pseudumbilicus. Chiralithes, from the Upper Eocene of Peru, is somewhat similar, but has a considerably broader aperture. Perulithes, also from the Upper Eocene of Peru, has a higher spire, and its later whorls are quite smooth. Clavilithes (Africolithes) africanus (Newton) 1922. Rhopalithes africanus Newton, p. 32, pi. 2, figs. 9-10. MATERIAL. Several specimens, including the lectotype (0.42004, Newton's fig. 10) here selected. 46 EOCENE MOLLUSCA FROM NIGERIA Genus LEUCOZONIA Gray, 1847 Leucozonia pseudominax sp. nov. (PI. 6, fig. 8) 1922. Cornulina minax (Solander) : Newton, p. 34, pi. 3, figs. 6-7. MATERIAL. The holotype (G. 42052, Newton's figs. 6-7). DESCRIPTION. Of medium size, similar in form to Cornulina minax. Protoconch not preserved. Last whorl forming a little more than half the height of the shell. Probably about four spire whorls forming a conic spire ; whorls angulated at about one-third of their height, cylindrical or with sides inclined slightly inwards abapically, sloping and gently concave adapically. Ornament of spiral threads crossing vague, broad ribs which form blunt, slightly upturned spines on the keel. Last whorl large, base excavated, lower portion missing ; spines well developed and protruding both on the main keel and on a subsidiary one developed abapically, with a narrow slit on their forward-facing side and thus evidently hollow. Aperture rounded-oval, with a small adapical channel limited below by a small spiral ridge ; evidently with a fairly short siphonal canal distinctly inclined to the left. Columella concave, with at least two strong columellar folds abapically. Outer lip thin, gently parasigmoidal, orthocline as a whole, coarsely but vaguely fluted internally. The broken lower end of the columella is solid. REMARKS. Apart from the fact that the ornament on the lower part of the last whorl differs from that of Cornulina minax in consisting of widely-spaced, sharp spiral threads instead of numerous closely-spaced threads of several orders, the presence of strong columellar folds indicates that the Nigerian form is not a Cornulina. Since the genus Fascioplex has an umbilicus, a lower spire, and a last whorl which is less inflated adapically, the Nigerian form seems best placed in the genus Leucozonia. Family VOLUTIDAE Genus VOLUTOCORBIS Ball, 1890 Volutocorbis multispinosa (Newton) 1922. Volutospina multispinosa Newton, p. 28, pi. 3, figs. 3-5. MATERIAL. Numerous specimens, including the lectotype (0.41949, Newton's fig. 3) here selected. REMARKS. This species is now removed from Volutospina, as the more scabrous ornament, less spinose ribs, and more oval form indicate that it is a Volutocorbis. Genus BENDELUTA nov. TYPE SPECIES. Volutospina conicoturrita Newton. GENERIC CHARACTERS. Of medium size, not very thick-shelled, in general form somewhat similar to Volutospina and related genera. Protoconch conical, not large, of EOCENE MOLLUSCA FROM NIGERIA 47 about three smooth, moderately convex whorls. Spire conic. Four flat-sided spire whorls, a little wider than high, ornamented with rather weak and narrow, straight axial ribs crossed by low, flat spiral ribbons ; whorls shouldered above, shoulder bearing increasingly prominent but short, upturned, hollow spines, both shoulder and horizontal sutural ledge free of spiral ornament ; sutures linear, undulatory. Last whorl large, inflated, with a second row of short, laterally directed spines at the level of the adapical end of the aperture, this level forming the widest part of the whorl ; ribs dying out below and spiral ornament obsolete on flanks ; base moderately sharply excavated at its junction with the rather broad neck at the middle of which a vague spiral depression demarcates the slightly swollen siphonal fascicle ; base, neck and siphonal fascicle with irregular spiral threads. Aperture elongate, rather irregular in shape on account of the constriction of the shell above the neck and on the flank, rather deeply notched below, and with a moderately short, oblique, poorly-differentiated siphonal canal. Columella oblique, well set off above, slightly convex, its upper half with three spiral folds of which the upper one is a little weaker and is a little closer to the median fold than is the lower one. Parietal callus spread adapically over half the ventral surface of the last whorl as a very thin glaze, almost absent medially and abapically. Lip straight and orthocline as a whole, slightly receding and sinuous abapically, strongly prosocline on the narrow shoulder, thin- edged, internally smooth. REMARKS. The more globose last whorl (the widest part of which bears the lower row of spines), the thicker neck, and the shorter and more irregularly shaped aperture, together with the other characters mentioned above, readily distinguish this from Volutospina. Bendeluta conicoturrita (Newton) 1922. Volutospina conicoturrita Newton, p. 27, pi. 3, figs. 1-2. MATERIAL. Several specimens, including the lectotype (0.41901, Newton's fig. i) here selected. Family OLIVIDAE' Genus PSEUDOLIVA Swainson, 1840 Subgenus BUCCINORBIS Conrad, 1865 Pseudoliva (Buccinorbis) kitsoni (Newton) 1922. Buccinorbis kitsoni Newton, p. 37, pi. 3, figs. 16-19. MATERIAL. Many specimens, including the lectotype (0.42069, Newton's fig. 16) here selected. REMARKS. The umbilicus is mostly covered by the umbilical callus, but the presence of such a distinct depression as is present in this species is more a feature of Buccinorbis than of Pseudoliva. 48 EOCENE MOLLUSCA FROM NIGERIA Family STREPSIDURIDAE Genus STREPSIDURA Swainson, 1840 Subgenus STREPSIDUROPSIS nov. SUBGENERIC CHARACTERS. Like Streps^d^lra, but with one strong spiral fold on the columella at the beginning of the canal, and above it a convex pad on which are grouped six more folds ; anterior part of shell more constricted, siphonal canal a little longer and more twisted ; outer lip internally thickened and weakly crenulated. REMARKS. Mazzalina has no carinate siphonal fascicle, and its canal is shorter and less oblique. Strepsidura (Strepsiduropsis) spirata Newton (PL 7, ng. 3) 1922. Strepsidura spirata Newton, p. 33, pi. 3, figs. 24-25. MATERIAL. Many specimens, including the lectotype (0.42008, Newton's fig. 25) here selected. Family CANCELLARIIDAE Genus SVELTIA Jousseaume, 1887 Subgenus AFRICOSVELTIA nov. TYPE SPECIES. Cancellaria multiplicis Newton. SUBGENERIC CHARACTERS. Of small-medium size, having the form of Sveltia or Sveltella, but with a rather shorter spire. Protoconch unknown. Spire conic ; three to four spire whorls which are convex and tend to be slightly subangular just above the middle ; sutures linear, undulatory. Ornament consisting of strong, widely- spaced spiral threads crossing solid axial ribs which are narrower than their intervals ; ribs fairly strong, becoming very solid on the last whorl, gently prosocline. Last whorl oval, narrower abapically, base declivous. Aperture oval, narrower and gently emarginate (not channelled) abapically. Columella straight, with three columellar folds of which the uppermost is the strongest. Columellar callus moder- ately widely spread adapically, narrower and detached abapically, leaving a narrow but distinct pseudumbilicus which is limited externally by a broad, vague swelling rather than by a siphonal fasciole. Lip straight to gently concave, slightly prosocline, thin-edged, with a strong varix close behind it, lirate and slightly thickened internally. REMARKS. The aperture has no siphonal notch as in Cancellaria. Compared with Sveltia s.str., in the new subgenus the spire is lower, there is a distinct pseudum- bilicus, and there are three (not two) columellar folds. Sveltella has no noticeable umbilicus, a higher spire, and only two columellar folds. EOCENE MOLLUSCA FROM NIGERIA 49 Sveltia (Africosveltia) multiplicis (Newton) (PL 7, figs. 4a-c) 1922. Cancellaria multiplicis Newton : p. 42, pi. 3, fig. 26. MATERIAL. The holotype (0.42137). Genus BONELLITIA Jousseaume, 1887 Subgenus ADMETULA Cossmann, 1889 Bonellitia (Admetula) amekiensis sp. nov. (PL 7, figs. 5«-c) 1922. Bonellitia cf. evulsa (Solander) : Newton, p. 44, pi. 5, figs. 14-15. MATERIAL. The holotype (0.42171, Newton's fig. 15) and several other specimens. DESCRIPTION. Small, like Admete and Admetula in form. Protoconch unknown. Last whorl slightly less than half the height of the shell. Spire conic, consisting of about four strongly convex whorls with deep, linear, undulatory sutures. Ornament consisting of solid axial ribs, practically straight and gently prosocline, equal to or slightly narrower than their intervals, occasionally varicose, crossed by spiral threads. Last whorl swollen-ovate, narrower abapically, base declivous and with the ribs feebler. Aperture rounded-oval, not notched abapically, but with a broad, slightly emarginate channel inclined to the left. Columella vertical, with three well-developed columellar folds of which the lower two are the more closely spaced. Parietal callus thin ; columellar callus narrow, its outer edge separated from the region of the neck by a well-marked, linear, vertical depression, but not by an umbilicus. No siphonal fascicle. Outer lip fairly straight, gently prosocline, blunt, limited behind by a varix, internally thickened and lirate, the lirae continuing over the floor of the aperture. REMARKS. Although Admetula is placed in the synonymy of Bonellitia by Wenz, the name seems worth retaining for those forms wj,th less muricate ornament. The Nigerian form differs from B. (A.} evulsa in its more oval-conic outline and less inflated form, its relatively finer and more numerous spiral threads, and its more callous and more detached columellar lip. Subgenus AFRICOSTOMA nov. TYPE SPECIES. Trigonostoma decorata Newton. SUBGENERIC CHARACTERS. Small, having the general form of a Bonellitia except for the marked, canaliculate ramp behind the shoulder. Protoconch naticoid, of two smooth, convex whorls. Last whorl forming slightly less than half the height of the shell. Spire conic, of two and a half subcylindrical, gently convex whorls which are slightly inturned below and have a sharp, upturned, undulatory carina on the shoulder angle ; ramp horizontal as a whole, gently concave, smooth except 5o EOCENE MOLLUSCA FROM NIGERIA for irregular continuations of the ribs. Ornament consisting of delicate, gently prosocline axial ribs crossed by spiral threads, both being narrower than their intervals ; the occasional very strong varices develop even stronger lobes on the upturned carina than do the ribs. Last whorl large, slightly convex abapically, ornamented like the spire whorls. No neck or siphonal fascicle. Aperture oval- pyriform, wider adapically, slightly emarginate abapically (not notched), with a slight projection at the shoulder carina. Columella vertical, with a double fold abapically (limiting the very short, oblique siphonal canal), and with a well-separated median fold. Parietal callus thin and not widely spread above, thicker and tending to be slightly detached below. Lip straight, gently prosocline, thin-edged but with a broad, strong varix just behind it, internally thickened and lirate. REMARKS. The absence of an umbilicus together with other characters described above indicate that this species is not a Trigonostoma ; it seems best regarded as belonging to a new subgenus of Bonellitia. Bonellitia ( Ajricostoma) decorata (Newton) (PI. 7, figs. 6a-c) 1922, Trigonostoma decorata Newton, p. 43, pi. 5, figs. 18-19. MATERIAL. Several specimens, including the lectotype (0.42138, Newton's fig. 18) here selected. Family CONIDAE Genus CONUS Linne, 1758 Subgenus LEPTOCONUS Swainson, 1840 Conus (Leptoconus) amekiensis sp. nov. (PI. 7, figs. 7«-c) 1922. Conospirus cf. parisiensis (Deshayes) : Newton, p. 24, pi. 3, figs. 8-9. MATERIAL. Several specimens, including the holotype (0.41837, Newton's fig. 9). DESCRIPTION. Of small-medium size, biconic. Protoconch rather tectiform, of three smooth, slightly convex whorls. Last whorl forming at least five-eighths of the height of the shell. Spire gently coeloconoid, of five to five and a half whorls with small nodes on an angulation close to the abapical suture ; the narrow portion below the angulation vertical or sloping slightly inwards and causing a slightly imbricate appearance ; upper part flattened or vaguely concave, shelving, with four or five spiral threads on its abapical two-thirds, the threads being slightly crenulated where crossed by the deeply concave growth-lines. Last whorl inverted-conic, shoulder with small, low nodes, base slightly excavated and neck slightly swollen ; coarse, irregular spiral ribbons, which have narrow intervals and are fainter adapically, are vaguely beaded where crossed by low, flat ribs extending from the small nodes on the shoulder ; the ribs are of about the same width as their intervals. Aperture EOCENE MOLLUSCA FROM NIGERIA 51 narrow and parallel-sided, vaguely constricted near the abapical end, emarginate but not notched terminally. REMARKS. Conus parisiensis, with which Newton compared this species, is more produced abapically and has compound crenulations on the uppermost thread of the spire whorls (not small nodes on the lower angulation) ; its spiral ornament is not beaded and is restricted to the abapical half of the last whorl. Family TURRIDAE Genus EOPLEUROTOMA Cossmann, 1889 Eopleurotoma nigeriensis Newton 1922. Eopleurotoma nigeriensis Newton, p. 23, pi. 2, figs. 11-12. MATERIAL. A few specimens, including the lectotype (0.41829, Newton's fig. 12) here selected. Genus TURRICULA Schumacher, 1817 Subgenus SURCULA H. & A. Adams, 1853 Turricula (Surcula) africana (Newton) 1922. Surcula africana Newton, p. 22, pi. 2, figs. 6-8. MATERIAL. Several specimens, including the lectotype (0.41809, Newton's fig. 7) here selected. Genus SURCULITES Conrad, 1865 Subgenus CLINURA Bellardi, 1875 Surculites (Clinura) ingens (Mayer-Eymar) 1922. Surcula ingens (Mayer-Eymar) : Newton, p. 20, pi. 2, figs. 3-5 (cum syn.). MATERIAL. Two specimens. Genus MITRELLOTURRIS nov. TYPE SPECIES. Asthenotoma (Endiatoma) casteri Chavan. GENERIC CHARACTERS. Having the general form of Mitrella, elongate oval-conic to rather turriculate-conic. Protoconch unknown. About six spire whorls (only four preserved), which are flat-sided, their height a little more than half their width. Sutures linear, slightly stepped. A narrow, flat, slightly raised juxtasutural band is limited abapically by a vague, narrow, spiral depression ; whorls otherwise smooth. Last whorl about two-fifths of the height of the shell, narrowly oval, base narrowly and gently excavated, the neck of moderate length and swollen. Base and neck with spiral threads becoming increasingly stronger and more closely spaced abapically. 52 EOCENE MOLLUSCA FROM NIGERIA Aperture narrow and parallel-sided, with a short, distinct siphonal canal notched at the end. Columellar callus not extensive, smooth. Small adapical channel present. Outer lip (broken) evidently thin, not varicose, internally smooth, gently convex abapically, orthocline or slightly prosocline at the suture, with a broad, shallowly V-shaped sinus at a position corresponding to two-thirds of the height of the spire whorls. The two last whorls with a very vague angulation of the surface at the apex of this sinus. REMARKS. Although obviously related to Asthenotoma and Endiatoma, this genus differs in the complete lack of axial ornament, the reduced spiral ornament, the smooth columella, and the absence of folds inside the outer lip. Mitrelloturris casteri (Chavan) (PI. 7, figs. 8a, b ; PL 8, fig. i) 1952. Asthenotoma (Endiatoma) casteri Chavan, p. 80, text-fig. MATERIAL. The holotype (0.69600). Genus AMEKICYTHARA nov. TYPE SPECIES. Cominella douvillei Newton. GENERIC CHARACTERS. Of small-medium size, form resembling that of the Cythara group of the Turridae. Protoconch naticoid, consisting of two and a half smooth, convex whorls. Last whorl forming four-sevenths of the height of the shell. Spire conic, consisting of three to four gently convex whorls which are moderately angulated at about two-thirds of their height ; sutures linear. Ornament delicately cancellate, Ficus-like, with tiny crenulations at the intersections of the axial and spiral elements. Last whorl oval, gently angulated adapically like the spire whorls and similarly ornamented ; base declivous and gently excavated at the origin of the moderately short neck ; axial ornament obsolete and spiral ornament irregular abapically. Aperture narrowly oval, with a short siphonal canal which is gently inclined to the left and flares a little at its deeply emarginate end. Columella gently concave, with one prominent fold (well inside the aperture) at half the height of the aperture proper. Columellar lip with thin callus which is moderately widely spread adapically but narrow abapically. No siphonal fasciole. Imperforate. Outer lip thin at edge, moderately convex medially, slightly concave at a level correspond- ing with that of the columellar fold, receding adapically, and with a very shallow, rounded sinus adjacent to the suture ; thickened internally and with prominent, short lirae. REMARKS. The above characters indicate that the species is a member of the Turridae. Cominella, in which genus Newton placed the species, is a buccinid genus with fundamentally different characters. 53 Amekicythara douvillei (Newton) (PI. 8, figs. 2a-c) 1922. Cominella douvillei Newton, p. 39, pi. 2, figs. 18-19. MATERIAL. Several specimens, including the lectotype (0.42100, Newton's fig. 19) here selected. LAMELLIBRANCHIA Family NUCULIDAE Genus NUCULA Lamarck, 1799 Nucula costaeimbricatis Newton 1922. Nucula costaeimbricatis Newton, p. 73, pi. 6, figs. 10-11. MATERIAL. The holotype (£,.48096, Newton's figs. 10-11). Family GLYCYMERIDAE Genus AFRICARCA nov. TYPE SPECIES. Glycimeris nigeriensis Newton. GENERIC CHARACTERS. Of small-medium size, moderately thick-shelled, moder- ately strongly inflated, oval-subtrapezoidal in outline. Beaks submedian, almost orthogyrous. Hinge-line almost straight, forming only about two-thirds of the length of the shell. Greatest height anterior to the median line, at about two-thirds of the length, the shell being slightly produced antero-ventrally. Cardinal area inclined mainly at an angle of about 45° to the surface of the hinge-plate. Hinge-plate fairly solid, straight medially, gently arched downwards on the flanks. Teeth taxodont, numerous, anterior ones slightly more numerous than the posterior, the eight most anterior and six most posterior ones larger and oblique ; a minute gap between anterior and posterior series. Muscle scars fairly large, not buttressed. No byssal sinus. Numerous low, square-cut ribs. Surface without carina. Ventral margin internally fluted. REMARKS. Although the hinge is somewhat like that of Glycymeris, the sub- trapezoid (rather Arcopsis-like) outline and other characters described above readily distinguish this genus. Africarca nigeriensis (Newton) (PI. 8, figs. 3a, 6) 1922. Glycimeris nigeriensis Newton, p. 72, pi. 8, figs. 8-10. MATERIAL. Several specimens, including the lectotype (1,48441, Newton's fig. 8) here selected. 54 EOCENE MOLLUSCA FROM NIGERIA SUPPLEMENTARY CHARACTERS. Beaks very slightly anterior to the middle line, the extreme tips tending to be slightly prosogyrous. Hinge-line sloping vaguely downwards and outwards away from the beak. Dorso-lateral angles obtusely rounded, the anterior a little the more angular. Anterior end gently convex, joining the ventral margin in a well-rounded curve. Ventral margin gently convex. Posterior end obliquely truncated, slightly convex, joining the ventral margin in a rounded, blunt angle. Cardinal area moderately narrowly triangular, placed symmetrically beneath the beak, slightly concave where abutting against the slight projecting ridges constituting the dorsal margin of the shell, carrying fairly numerous ridges, slightly wider than their intervals medially and narrower than their intervals on the flanks, at right angles to the hinge-line. Adult with 24 anterior and 20 posterior teeth, distally converging ventrally ; median 30 teeth all small, especially medially. Surface with nearly 70 rather low, square-cut, finely beaded ribs of about the same width as their intervals, rather finer on the flanks. Surface evenly convex. Family NOETIIDAE Genus PROTONOETIA MacNeil, 1938 Protdnoetia nigeriensis (Newton) 1922. Anadara nigeriensis Newton, p. 70, pi. 8, figs. 4-7. 1938. Protonoetia nigeriensis (Newton) : MacNeil, p. 25, pi. 4, figs. 1-3. MATERIAL. Several specimens, including the lectotype (L. 48544, Newton's fig. 4) here selected. REMARKS. This is the type species of Protonoetia. Genus ARCOPSIS von Koenen, 1885 Arcopsis africana (Newton) (PI. 8, fig. 4) 1922. Fossularca africana Newton, p. 68, pi. 8, figs. 14-17. MATERIAL. Several specimens, including the lectotype (L. 481 13, Newton's fig. 14) here selected. REMARKS. Fossularca is a synonym of Arcopsis. Genus RECTANGULARCA nov. TYPE SPECIES. Striarca africana Newton. GENERIC CHARACTERS. Of small-medium size, not very thick-shelled, rather strongly inflated, subrectangular in outline. Beaks median, practically orthogyrous, the extreme tips slightly opisthogyrous. Cardinal area narrowly triangular, the portion anterior to the beak slightly the shorter ; with one fine chevron-shaped groove very close to the two shorter sides, the remainder of the surface with EOCENE MOLLUSCA FROM NIGERIA 55 numerous, fine, closely-spaced ridges perpendicular to the hinge-line ; when the valves are in contact the two areas are in one plane, occasionally pouting slightly along the line of junction. Hinge-line straight, forming about seven-ninths of the length of the shell. Dorso-lateral angles obtuse, the anterior a little the more so. Anterior end gently convex near the dorso-lateral angle, rather sharply rounded below. Ventral margin almost straight, gently upturned near the ends. Posterior end fairly straight, joining the ventral margin in a sharply rounded curve. No byssal depression. Hinge-plate very narrow, straight, its base horizontal, slightly expanded at the extreme ends ; a small, narrow gap, slightly depressed, beneath the beak, with 22- 23 taxodont teeth on each side, the teeth being short and vertical except for the five or six flanking ones which are slightly longer and oblique (converging ventrally) . Surface not carinate, with fine, very numerous radial riblets of three orders minutely beaded by fine concentric threads. Muscle scars large, subequal, not buttressed. Valve margins smooth. REMARKS. The more elongate and subrectangular form, characters of the hinge, lack of a buttress to the posterior muscle scar, and smooth valve margins distinguish this genus from Striarca. Breviarca is much shorter and has a much more convex ventral margin. Rectangularca africana (Newton) (PL 8, figs. 5«, b, 6) 1922. Striarca africana Newton, p. 69, pi. 8, figs. 11—13. MATERIAL. Several specimens, including the lectotype (1^.48172, Newton's fig. n) here selected. DIMENSIONS. Height 9-6 mm., length 18-3 mm., thickness (two valves) 7-8 mm. Family PLICATULIDAE Genus PLICATULA Lamarck, 1801 Plicatula polymorpha Bellardi 1922. Plicatula polymorpha Bellardi : Newton, p 65, pi. 6, figs. 6-9 (cum syn.). MATERIAL. Fourteen specimens. Family OSTREIDAE Genus OSTREA Linne, 1758 Ostrea amekiensis sp. nov. (PL 8, fig. 7) 1922. Ostrea cf. ludensis Deshayes : Newton, p. 61, pi. 8, figs. 2-3. MATERIAL. The holotype ^.48195, Newton's figs. 2-3). DESCRIPTION. The single specimen is a right valve, rounded-subtriangular, a little 56 EOCENE MOLLUSCA FROM NIGERIA narrower in the upper third near the ligament area which (although worn) is much enrolled forward but otherwise typical. The valve is almost flat, very slightly concave medially, vaguely out-turned laterally and at the rounded end. Outer surface, apart from growth lines, with numerous, fine vermicular riblets like those of Placuna (about six to the mm. near the rounded end). Muscle scar slightly posterior to the middle. Valve margin smooth internally. DIMENSIONS. Height 28-7 mm., length 26-3 mm. REMARKS. Although known only by a right valve, this species is well character- ized by its Placuna-like. ornament. As shown by Deshayes's illustration, the right valve (called " left valve ") of 0. ludensis is larger and less transverse, and has a larger ligament area and ornament of a different type (not Placuna-like). Ostrea pseudomarginidentata sp. nov. (PI. 8, fig. 8) 1922. Ostrea cf. marginidentata S. V. Wood : Newton, p. 60, pi. 6, figs. 2-5. MATERIAL. Several specimens, including the holotype (1,48203). DESCRIPTION. Of medium size, moderately thick-shelled, linguiform. Left valve (incomplete) with a large, flattened attachment area occupying most of the surface, leaving a narrow area, inturned ventrally almost at right angles, with numerous rather small radial ribs of about the same width as their intervals (four occupying 5-2 mm. ventrally, where they are widest). Valve margin correspondingly frilled. Right valve irregularly flat. Outer surface smooth except for growth lines and a few irregularities. A narrow, inturned margin, flanking the ligament area, has rather prominent, short ribs which are irregularly spaced but normally narrower than their intervals ; further from the umbo these merge into relatively coarse, much wider- spaced crenulations on the inner margin of the valve. Muscle scar fairly large, posterior to the middle. DIMENSIONS. Holotype (a left valve) : height (incomplete) 24-3 mm., length 19-6 mm. Topotype (Newton's illustrated right valve, 1.48202) : height, 36-5 mm., length 22-8 mm. REMARKS. While showing some superficial similarity to 0. marginidentata S. V. Wood, this species differs not only in being smaller, but in being usually higher, and in having a larger attachment area, and plications which are finer on both valves. Subgenus CRASSOSTREA Sacco, 1897 Ostrea (Crassostrea) lugardi (Newton) 1922. Crassostrea lugardi Newton, p. 62, pi. 6, fig. i ; pi. 7, fig. i ; pi. 8, fig. i. MATERIAL. Several specimens, including the lectotype (£,.48217, Newton's pi. 6, fig. i and pi. 7, fig. i) here selected. EOCENE MOLLUSCA FROM NIGERIA 57 Family CARDITIDAE Genus GLANS M. von Miihlfeldt, 1811 Glans nigeriensis sp. nov. (PL 8, figs, ga, b) 1922. Cardita cf. planicosta J. Sowerby : Newton, p. 80. MATERIAL. The holotype (1^.48318). DESCRIPTION. The solitary right valve is small, subtrapezoidal in outline, and rather strongly inflated. Umbo moderately prominent, distinctly prosogyrous, situated at about one-third of the length from the anterior end. Lunule small, indis- tinct, limited in its early stages by a fine, raised thread. Escutcheon well defined. Antero-dorsal margin gently convex, more steeply descendent than the longer, straight postero-dorsal margin. Anterior end rather sharply rounded, situated rather low down. Ventral margin almost straight. Posterior end truncated slightly obliquely, joining the ventral margin in a rounded angle which is only a little more than a right angle, and the postero-dorsal margin in a rounded angle of about 155°. Surface ornamented with 23 rather low, moderately broad, rather vaguely beaded ribs which are slightly wider than their intervals over the main body of the shell but more closely spaced at the extreme ends ; they have rather square-cut, shallow intervals, and are flattened on top, although slightly depressed along the middle line. Valve margins coarsely fluted. Right valve hinge : AI small, pointed, with a small, shallow socket above it ; 3a broken, evidently thin and moderately oblique ; 3b solid, rather narrowly triangular, its dorsal pointed portion projecting into the shell cavity ; a small PHI may have been present, but cannot be observed as the margin of the shell is rather worn posteriorly. DIMENSIONS. Height 11-9 mm., length 14-7 mm. REMARKS. From its general form and ornamentation, as well as from the charac- ters of the hinge, it is evident that this specimen is not a juvenile of the large Venericardia (Venericor) planicosta (Lamarck). Subgenus BENDEGLANS nov. TYPE SPECIES. Cardita costaeirregularis Newton. SUBGENERIC CHARACTERS. Of small-medium size, only moderately thick-shelled, subtriangular-cuneiform, moderately strongly inflated ; anterior and posterior parts of surface with flat ribs separated by linear intervals, median portion with only three very wide, rather high and flat-topped ribs separated by broad intervals, the intervals showing as raised broad ridges on the inner surface of the shell ; right valve with small AI, low 3a, solid 3b, lamellar 5b, and small PHI ; left valve with All, fairly solid 2 and 4b, PII, and small PIV. GEOL. Ill, 2. 6 58 EOCENE MOLLUSCA FROM NIGERIA Glans (Bendeglans) costaeirregularis (Newton) (PI. 9, figs, i, 2) 1922. Cardita costaeirregularis Newton, p. 81, pi. 9, figs. 30-31. 1944. Cyclocardia costaeirregularis (Newton) : Chavan, p. 35. MATERIAL. Many specimens, including the lectotype (1.48324, Newton's fig. 30) here selected. SUPPLEMENTARY SPECIFIC DESCRIPTION. Beaks pointed, rather high, prosogyrous, situated anterior to the middle line at about two-thirds of the length. Lunule small, cordiform, concave. Escutcheon rather short, narrow. Anterior part of surface with nine depressed, flatly rounded, unbeaded ribs separated by linear intervals ; posterior part with n similar but rather narrower ribs ; median part with three very broad, elevated, flat-topped, unbeaded ribs separated by broad, subrectangular intervals ; the anterior ribs increase in size posteriorly, the ninth being almost as large as the adjacent one on the median part and separated from it by a linear interval ; the most anterior of the posterior group is a little larger than the remainder ; valve margins fluted according to the ribbing, the three main depressions of the outer surface showing as ridges on the inner surface. Posterior adductor impression of normal shape, anterior one rather elongate. Right valve : AI small, low, close to the end of 33 ; 3a obscure, thin, low, more or less fused to the valve margin, only slightly inclined forwards ; 3b solid, elevated, not very broadly triangular, inclined slightly backwards ; 5b long, thin and lamellar, almost straight, fused to the lower edge of the nymph from which it is separated by a narrow groove ; PHI lamellar, moderately long, fused to the continuation of the ridge limiting the escutcheon. Left valve : All short, fused to the ridge limiting the inner margin of the lunule ; 2 solid, elevated, narrowly triangular, slightly inclined forwards ; 4b solid, elevated, rather more narrowly triangular, oblique ; PI I small ; apparently a still smaller tooth (PIV) above and behind PI I. Nymph narrow, relatively short. Escutcheon of right valve with a groove outside it, radiating from beneath the beak and enlarging with growth. REMARKS. Chavan placed this species in Cyclocardia, but in C. borealis (Conrad), as represented by specimens in the British Museum (Natural History), the shell is suborbicular and has beaded ribs in youth, and vague, flatly V-shaped ribs in the adult, with no median enlarged ribs ; there are no posterior lateral teeth, and the cardinal teeth of the right valve differ in orientation. The hinge of Cardita (Cyclo- cardia) granulata Say, as figured by Chavan, differs in a similar manner. The greatly enlarged median ribs and the hinge characters warrant the placing of the Nigerian species in a new subgenus. Subgenus AMEKIGLANS nov. TYPE SPECIES. Cardita costaenodulosis Newton. SUBGENERIC CHARACTERS. Of small-medium to medium size, very thick-shelled, oval-subtriangular to cuneiform, inflation moderate ; extreme posterior ribs flat EOCENE MOLLUSCA FROM NIGERIA 59 and rather low, slightly wider than their intervals, the remainder very narrow and high, with rounded tops, finely beaded, with deep, broader, smooth, U-shaped intervals ; right valve with AI, low 3a, solid 3b, lamellar 5b, and distinct PHI ; left valve with short, pointed All, solid triangular 2, narrower and longer 4b, and distinct PI I. Glans (AmekiglansJ costaenodulosis (Newton) (PI. 9, figs. 3, 4) 1922. Cardita costaenodulosis Newton, p. 82, pi. 9, figs. 22-25. 1938. Cossmannella costaenodulosis (Newton) : Chavan, pp. 3-10, fig. i. 1944. Cossmannella costaenodulosis (Newton) : Chavan, p. 35. MATERIAL. Many specimens, including the lectotype (1.48360, Newton's figs. 24-25) here selected. SUPPLEMENTARY SPECIFIC DESCRIPTION. Beaks pointed, rather high, prosogyrous, situated anterior to the middle line at about two-thirds of the length. Lunule small, narrowly cordiform, depressed, bulging medially. Escutcheon narrow, rather short. Extreme posterior end with five rather flattened, unbeaded ribs with slightly nar- rower intervals, the remainder of the surface with 14 very narrow and high, finely beaded ribs with rounded tops, separated by smooth, broadly U-shaped intervals of nearly three times their width. Valve margins coarsely fluted. Right valve : AI rather small, fairly close to and in line with the end of 3a ; 3a low, not very long, sloping forwards at about 45°, partially fused to the inner margin of the lunule, from which it is separated by a slight groove ; 3b solid, elevated, triangular, its anterior edge vertical ; 5b rather long, very thin and lamellar, fused to the lower edge of the nymph ; PHI distinct, moderately long, rather distant, situated fairly near the valve margin. Left valve : AH small, pointed, situated at the forward end of the inner edge of the lunule ; 2 solid, elevated, triangular, its posterior edge vertical ; 4b solid, more narrowly triangular, oblique at about 45° ; PII distinct. Posterior adductor impression tending to be obliquely subrectangular ; anterior adductor impression moderately elongate. REMARKS. Chavan placed this species in Cossmannella. None of the specimens of Cardita fayumensis Oppenheim (C. aegyptiaca Fraas non Monterosato) , the type species of Cossmannella, in the British Museum (Natural History), shows the hinge. It is not recorded as having any lateral teeth, but Chavan has intimated to me that the hinges of the two species are identical as judged from specimens in his own collections. The ribs of C. fayumensis are recorded as being " few, strong, sharp, slightly tripartite ", but while specimens in the British Museum (Natural History) do occasionally show tripartite ribs, this is due to wear. The ribs are considerably more numerous and less high than in costaenodulosis, the intervals being much less conspicuous, and the shell is not normally so pointed posteriorly, often being trun- cated. In view of the doubts concerning the exact characters of the hinge of Cossmannella, and since its form and ribbing apparently differ appreciably from those of the Nigerian species, a new subgeneric name seems warranted for the latter. Go EOCENE MOLLUSCA FROM NIGERIA Subgenus DIVERGIDENS nov. TYPE SPECIES. Cardita triparticostata Newton. SUBGENERIC CHARACTERS. Of small to small-medium size, fairly thick-shelled, rather strongly inflated, outline subtriangular, moderately high. A few simple, narrow ribs posteriorly, the remainder of the surface with broad, low, square ribs with deep, narrow, almost linear, square-cut intervals, the middle third of each rib bearing a strong, rounded, beaded cord. Right valve with AI, obsolete and very oblique 3a, solid and rather broadly triangular 3b, and distinct PHI ; left valve with All, strongly divergent 2 and 4b (the latter the more oblique), PII and weak PIV. Glans (Divergidens) triparticostata (Newton) (PI. 9, figs. 50, b, 6, 7) 1922. Cardita triparticostata Newton, p. 83, pi. 9, figs. 26-29. 1938. " Venericardia " triparticostata (Newton) : Chavan, p. 7. MATERIAL. Several specimens, including the lectotype (1^.48319, Newton's fig. 26) here selected. SUPPLEMENTARY SPECIFIC DESCRIPTION. Beaks pointed, prominent, moderately prosogyrous, situated anterior to the middle line at about three-fifths of the length. Lunule rather small, moderately broadly cordiform, smooth, not much sunk. Escutcheon short. Extreme posterior end with seven low, gently rounded, unbeaded ribs with rather narrower intervals, the remainder of the surface with 18 low, flat, broad ribs with deep, narrow, almost linear, square-cut intervals, the middle third of each rib carrying a strong, rounded, beaded cord. Valve margins coarsely fluted. Right valve : AI small but distinct, slightly elongate ; 3a obscure, long, low, thin, lamellar, very oblique and only gently descendent, fused to the inner margin of the lunule ; 3b solid, elevated, rather broadly triangular, its anterior edge sloping gently forwards, its posterior edge sloping more obliquely backwards ; PHI distinct. Left valve : All distinct, slightly elongated ; 2 and 4b strongly divergent, solid, elevated, narrowly triangular, their inner margins forming an angle of about 80°, 4b slightly more oblique than 2 ; PII distinctly elongated ; PIV similar but weaker. Nymph rather narrow, moderately long. Posterior adductor impression normal in shape ; anterior adductor impression moderately elongate. REMARKS. Chavan tentatively referred this species to Glyptoaxis, but the right cardinal is not curved and elongate ; he has recently suggested to me that it may be a Cardiocardita (group of Cardita beaumonti), but Cardiocardita has no anterior laterals. In its form, ornament, and hinge characters the Nigerian species appears to belong to a new subgenus. Family LUCINIDAE Genus PHACOIDES Gray, 1847 Phacoides eaglesomei Newton 1922. Phacoides eaglesomei Newton, p. 75, pi. 8, figs. 18-20. MATERIAL. Numerous specimens, including the lectotype (£,.48237, Newton's fig. 18) here selected. EOCENE MOLLUSCA FROM NIGERIA 61 Genus GIBBOLUCINA Cossmann, 1904 Subgenus EOMILTHA Cossmann, 1910 Gibbolucina (Eorniltha) (?) subrhomboidalis (Newton) 1922. Phacoides subrhomboidalis Newton, p. 77, pi. 7, figs. 4-5. MATERIAL. The holotype (1,48244). REMARKS. The characters of the hinge are unknown since the valves of the single specimen available cannot be separated. Chavan (in correspondence) has suggested that the species may be an Eomiltha. Genus POMPHOLIGINA Dall, 1901 Subgenus EODIVARICELLA Chavan, 1951 Pompholigina (Eodivaricella) oppenheimi (Newton) 1922. Divaricella oppenheimi Newton, p. 78, pi. 7, figs. 2-3. 1951. Pompholigina (Eodivaricella) oppenheimi (Newton) : Chavan, p. 23, fig. 27. MATERIAL. A few specimens, including the lectotype (L.48O97, Newton's fig. 2) here selected. REMARKS. This is the type species of Eodivaricella. Family CARDIIDAE Genus FRAGUM (Bolten MS.) Roding, 1798 Subgenus AFRICOFRAGUM nov. TYPE SPECIES. Cardium cf. obliquum Lamarck : Newton = Fragum (Africofragum) newtoni sp. nov. SUBGENERIC CHARACTERS. Small, outline Fragum-like, length and height nearly equal, posteriorly subcarinate. Ornament of flattened ribs which have rather narrower intervals, and are smooth except for the most anterior four or five which bear small prickles ; ribs on posterior area more irregular and producing short spines at the margin. Postero-dorsal margin, of left valve only, with a row of six short, upturned and outward-bent spines. Hinge less arched than in Fragum, anterior lateral teeth a little closer to cardinals than posterior lateral teeth. Nymph short. Fragum (Africofragum} newtoni sp. nov. (PI. 9, figs. Sa-c) 1922. Cardium cf. obliquum Lamarck : Newton, p. 74, pi. 7, figs. 6-9. MATERIAL. Many specimens, including the holotype (L.484i6, Newton's fig. 6). DESCRIPTION. Beaks moderately prominent, not large, prosogyrous, situated slightly anterior to the median line. Antero-dorsal margin short, straight or slightly 62 EOCENE MOLLUSCA FROM NIGERIA convex, joining the anterior end in an obtusely rounded angle. Anterior end well rounded, receding a little ventrally. Ventral margin convex, its posterior part the straighter, joining the posterior end in a blunt angle of a little more than 90°. Posterior end obliquely truncated, gently convex. Postero-dorsal margin short, nearly straight. About 34 ribs. Left valve : All obscure, below the forward end of a well-developed AIV ; 2a massive, upturned, pointed ; 2b small ; PII and PIV small. Right valve : AI larger than AIII ; 3a small ; 3b massive, upturned, pointed ; PI well developed ; no PHI below the straight shell margin, above which are the spines. Valve margins fluted. REMARKS. Cardium obliquum Lamarck appears to be a Loxocardium, and is quite different from the Nigerian form in that it is transversely oval in outline, not sub- carinate, and has distinct transverse scales or beads on the ribs, but no spines. Family VENERIDAE Genus TIVELINA Cossmann, 1886 Tivelina newtoni sp. nov. (PL 9, figs, ga, b, loa, b) 1922. Tivelina cf. sphenarium (Bayan) : Newton, p. 90, pi. 9, figs. 7-10. MATERIAL. Many specimens, including the holotype ^.48504, Newton's fig. 7). DESCRIPTION. Small to small-medium. Moderately thick-shelled, oval-subtri- angular. Beaks small, prosogyrous, high, situated anterior to the middle line at about two-thirds of the length. Inflation moderate ; greatest height anterior to the middle line, at the position of the beaks. Lunule large, narrowly cordiform, limited by a fine raised thread. Escutcheon narrow, rather short. Antero-dorsal margin gently convex, steeply descendent. Anterior end well rounded, a little produced antero-ventrally. Ventral margin convex, straighter (even vaguely emarginate) posteriorly. Posterior end moderately produced, situated rather low down, sharply rounded. Postero-dorsal margin long, slightly convex, rather steeply descendent. Surface with numerous strong, rather irregular, fairly sharp, concentric threads. Right valve : AI and AIII small, short, lamellar ; 3a rather short, simple, thin, inclined forwards ; i vertical, narrowly triangular, simple ; 3b oblique, deeply grooved. Left valve : All prominent ; 2a thin, simple, inclined forwards ; 2b narrowly triangular, simple, inclined backwards ; 4b oblique, thin, lamellar, simple, moderately long. Nymph moderately long, thin. Pallial sinus subtriangular, rather short, its apex not reaching the middle line, its upper arm subhorizontal, its lower arm steeply descendent. Valve margins smooth. REMARKS. Comparison with specimens in the British Museum (Natural History) shows that this is not Bayan's species ; the ornament is more serrate and less Costacallista-like, the pallial sinus is larger, and the hinge-plate has no rectangularly ending projection beneath the anterior lateral teeth, as in T. sphenarium. EOCENE MOLLUSCA FROM NIGERIA 63 Genus PITAR Roemer, 1857 Pitar amekiensis sp. nov. (PL 9, figs, na-c, 12 ; PL 10, fig. i) 1922. Cordiopsis incmssata (J. Sowerby) : Newton, p. 85 (pars). MATERIAL. The holotype (1.48253) and two topotypes (L. 48245-6). DESCRIPTION. Of medium size, moderately thick-shelled, rather strongly inflated, equivalve, subtriangular in outline. Beaks rather prominent, prosogyrous, situated anterior to the middle line at about one-fifth of the length. Escutcheon moderately long, narrow. Lunule large, rather narrowly cordiform, limited by a vague incised line. Antero-dorsal margin straight, steeply descendent. Anterior end rather sharply rounded. Ventral margin convex, with a gentle median bulge, straighter anteriorly and posteriorly. Posterior end vaguely truncated, joining the ventral margin in an obtusely rounded angle. Postero-dorsal margin moderately long, gently convex. Surface with irregular concentric threads, often with noticeably narrower intervals. Right valve : AI and a smaller, shorter AIII ; 3a vertical, rather thin, its forward face less perpendicular to the surface of the hinge-plate than its posterior face ; I sloping slightly backwards, its posterior face the less nearly perpendicular to the hinge-plate ; 3b oblique, moderately long, deeply grooved. Left valve : All pro- minent, pointed ; 2a thin, lamellar, vertical ; 2b oblique at about 45°, narrowly triangular, higher posteriorly ; 4b moderately long, simple, gently curved down- wards near its end. Nymph of medium length (about twice as along as the posterior cardinal tooth). Muscle impressions not very large. Pallial sinus rather acutely triangular, its upper arm gently ascendent, its apex narrowly rounded and situated slightly less than half-way across to the anterior adductor impression, its lower arm very steeply descendent. Valve margins smooth. DIMENSIONS. Holotype: height 27-0 mm., length 30-7 mm. Topotype (1,48245) : height 30-9 mm., length 35-0 mm. REMARKS. These three specimens, which were labelled " Cordiopsis incmssata " by Newton, together with those here recorded as Sinodiopsis coxi sp. nov., are quite different from that species ; they are more inflated anti more triangular, the beaks are much higher and less anterior in position, the hinge-plate is placed less forward and less inclined, the pallial sinus is a little shorter and more ascendent, the nymph is shorter, the teeth differ in the details of their orientation, the ventral margin is more bulging, and the concentric ornament is less serrate. Genus CHIONELLA Cossmann, 1886 Subgenus COSTACALLISTA Palmer, 1927 Chionella (Costacallista) elongatotrigona (Newton) 1922. Callista elongatotrigona Newton, p. 88, pi. 9, figs. 1-5. MATERIAL. Many specimens, including the lectotype (1,48065, Newton's fig. 3) here selected. 64 EOCENE MOLLUSCA FROM NIGERIA Subgenus MICROCALLISTA Stewart, 1930 Chionella (Microcallista) kitsoni (Newton) (PI. 10, figs. 2a, b, 3) 1922. Callista kitsoni Newton, p. 89, pi. 9, figs. 6, 11-14. MATERIAL. Several specimens, including the lectotype (L.48o69, Newton's fig. n) here selected. REMARKS. The left anterior cardinal tooth (2a) is simple, not grooved as in " Callista " auct. ( = Costacallista Palmer, 1927). Genus SINODIA Jukes-Browne, 1908 Sinodia heward-belli Newton 1922. Sinodia heward-belli Newton, p. 86, pi. 9, figs. 18-21. MATERIAL. Several specimens, including the lectotype (L.48iO2, Newton's fig. 21) here selected. REMARKS. The left anterior cardinal tooth (2a) is not faintly grooved as in Sinodia, but in all other characters the species agrees with that genus and seems best retained in it. Genus SINODIOPSIS nov. TYPE SPECIES. Cordiopsis incrassata (J. Sowerby) : Newton = Sinodiopsis coxi sp. nov. GENERIC CHARACTERS. Of medium size, moderately well inflated, transversely oval- subquadrate in outline, equivalve. Beaks rather small, prosogyrous, not prominent, well recurved, situated anterior to the middle line at about one-quarter to one-fifth of the length. Escutcheon fairly long, narrow. Lunule vague, large and moderately narrowly cordiform, limited by a fine incised line. Surface ornamented with rather irregular, strong, closely-spaced concentric threads. Hinge-plate moderately well developed ; not as thick, as inclined, or as high as in Cordiopsis and Sinodia. Right valve : AI rather solid and triangular ; AIII small ; 3a thin, lamellar, gently inclined forwards ; i rather triangular, its forward edge vertical, high, lamellar, its surface shelving downwards posteriorly ; 3b oblique and deeply bifid. Left valve : All very prominent and pointed ; 2a thin, lamellar, vertical ; 2b not so thin as 2a but lamellar at the apex, oblique at about 45° ; 4b simple, thin, long, gently arched. Nymph long, narrow, gently arched. Muscle impressions rather large, situated rela- tively more dorsally than in Cordiopsis and Sinodia. Pallial sinus bluntly and moderately acutely triangular, its upper arm horizontal or slightly ascendent, its apex sharply rounded, its lower arm descendent at about 45°, the apex situated at about mid-length of the shell. Valve margins smooth. EOCENE MOLLUSCA FROM NIGERIA 65 Sinodiopsis coxi sp. nov. (PI. 10, figs. 4, 5) 1922. Cordiopsis incrassata (J. Sowerby) : Newton, p. 85 (pars), pi. 9, figs. 15-17 (non J. Sowerby). 1938. Macrocallista palmerae Caster, p. 66 (pars) (non pi. i, figs. 9-10 ; pi. 8, fig. 7). MATERIAL. The holotype (1^.48250) and several topotypes. REMARKS. This form is quite distinct from Sinodia (Cordiopsis} orbicularis (Goldfuss) (= Venus incrassata J. Sowerby non Brocchi) in its outline, hinge, and pallial sinus ; the species was erroneously referred by Caster to the Angola Miocene species Macrocallista palmerae, but it differs still more from Macrocallista. While the hinge and pallial sinus are of the same type as in Sinodia, the form is quite different, the hinge-plate is less massive and less inclined, and tooth 2a is not grooved. Cordiopsis, which also has a much more massive and more inclined hinge and a different outline, has a shorter pallial sinus. The species is named after Dr. L. R. Cox. Family MACTRIDAE Genus SPISULA Gray, 1837 Subgenus CREPISPISULA nov. TYPE SPECIES. Mactra semisulcata Lamarck: Newton = Spisula (Crepispisula) amekiensis sp. nov. SUBGENERIC CHARACTERS. Of medium size, moderately thin-shelled, subtriangular, rather Hecuba-like in outline, rather strongly inflated, subcarinate anteriorly and posteriorly. No lunule or escutcheon. Ornament of coarse incrementals. Left valve : All rather short, prominent ; 2a and 2b meeting in a right angle dorsally, projecting ; resilium pit moderately narrowly triangular, its anterior side limited by a high, thin lamella, its posterior side by a very slight ridge ; PII prominent, a little longer than All ; posterior and anterior lateral teeth at about equal distances from the cardinal teeth. Right valve : AI a little larger and stronger than All I ; 3a and 3b simple, divergent, 3a a little the more oblique ; resilium pit as in the left valve ; PI and PHI better developed than the opposing anterior lateral teeth. A narrow ligament slit extends from the dorsal side of the hinge-plate to the tip of the beak. Pallial sinus narrow, forming a gently ascending tongue with its apex at mid- length of the shell. Valve margins smooth. Spisula (Crepispisula} amekiensis sp. nov. (PI. 10, figs. 6a, b, 7) 1922. Mactra semisulcata Lamarck : Newton, p. 93, pi. 7, figs. 10-13. MATERIAL. Several specimens, including the holotype ^.48224, Newton's figs. 10-11). 66 EOCENE MOLLUSCA FROM NIGERIA SUPPLEMENTARY SPECIFIC DESCRIPTION. Beaks rather high, narrow, prosogyrous. Antero-dorsal margin long, straight. Anterior end narrowly rounded, even bluntly pointed. Ventral margin straightest posteriorly, bulging somewhat downwards along its anterior half. Posterior end sharply and narrowly rounded to bluntly pointed. Postero-dorsal margin arched, obtusely angulated in two places. Anterior carination formed by a vague depression anterior to which the concentric ornament becomes less closely spaced than on the middle of the shell. Posterior angulation formed of two very obtuse carinae, the ornament behind the posterior one being likewise relatively less closely spaced. REMARKS. Although this form is somewhat similar in outline to the Recent genus Scissodesma Gray, it has definite concentric ornament, its posterior carina is less marked, the ligament slit from the dorsal side of the hinge-plate to the tip of the beak is distinctly smaller, and the lateral teeth are not crenulated. Mactra semi- sulcata Lamarck, a Paris Basin Eocene species with which Newton identified the Nigerian form, possesses a similar ligament slit, but its outline differs in being less triangular and less bulging antero-ventrally, it has a less definite anterior ridge, and its ornament is weaker. Family TELLINIDAE Genus MACOMA Leach, 1819 Subgenus BENDEMACOMA nov. TYPE SPECIES. Peronaea nigeriensis Newton. SUBGENERIC CHARACTERS. Of large-medium size, rather thick-shelled, transversely oval-subtriangular, length considerably exceeding height, inflation moderate. Beaks small, moderately prominent, prosogyrous. Surface ornamented with accentuated growth-lines, posteriorly with two very vague carinae. Escutcheon long, narrow. Lunule narrow, shorter than escutcheon, limited by a fine incised line. Left valve : 2a vertical, distinctly grooved dorsally ; 2b moderately oblique, very thin and lamellar, simple ; no lateral teeth. Right valve : 3a rather solid, directed moderately forwards, grooved dorsally ; 3b a little longer than 3a, directed backwards moder- ately obliquely, well grooved ; no lateral teeth. Nymph long and rather narrow. Pallial sinus rather narrowly tongue-shaped, its upper part rising slightly for a short distance from the posterior adductor impression, then gently descendent for most of its length, narrowly rounded at its apex, its lower part coalescent with the pallial line posteriorly for nearly half its length. Valve margins smooth. Macoma (Bendemacoma) nigeriensis (Newton) 1922. Peronaea nigeriensis Newton, p. 91, pi. n, figs. 1-3. MATERIAL. Several specimens, including the lectotype ^.48513 , Newton's fig. i) here selected. REMARKS. The characters of the hinge and the form of the pallial sinus indicate that this species is not a Peronaea ; it appears to belong to a new subgenus of EOCENE MOLLUSCA FROM NIGERIA 67 Macoma characterized by the shape of the shell, the grooved 3a, and the form of the pallial sinus. Family MYIDAE Genus RAETOMYA Newton, 1919 Raetomya schweinfurthi (Mayer-Eymar) 1922. Raetomya schweinfurthi (Mayer-Eymar) : Newton, p. 96, pi. 10 (cum syn.). 1942. Raetomya schweinfurthi (Mayer-Eymar) : Rossi, p. 182, pi. n, fig. 3. 1952. Labiosa (Raeta) schweinfurthi (Mayer-Eymar) : Tessier, pp. 350-351, pi. 29, figs. 7-8. 1954. Labiosa (Raeta) schweinfurthi (Mayer-Eymar) : Dartevelle & Roger, pi. 5 fig. 4. 1955. Raeta schweinfurthi (Mayer-Eymar) : Dartevelle & Roger, pp. 164-167. REMARKS. This is the type species of Raetomya. Tessier placed it as a subgenus of the mactrid genus Labiosa Schumacher, 1817, which is evidently a synonym of Anatina Lamarck, 1816. Raeta is also a mactrid genus, but schweinfurthi has been well described by Newton and is undoubtedly a myid. Since it was recorded from Egypt and Nigeria, it has also been recorded from the Eocene of Tripolitania and the Cameroons and from the (reputed) Lower Lutetian or Ypresian of Senegal ; the horizon stated in the latter record is rather low, and it might be advisable to re- investigate the evidence on which the age was assigned. The Angola Miocene speci- mens referred to this species by Caster actually belong to Platyodon klinghardti ( J. Bohm) ; Dartevelle & Roger express doubts concerning the generic assignation of the latter species, and it may well be that it is a distinct Miocene species of Raetomya. Family CORBULIDAE Genus VARICORBULA Grant & Gale, 1931 Varicorbula amekiensis sp. nov. (PI. 8, figs. 10, ii ; PI. 10, figs. 8a-c) 1922. Corbula rugosa Lamarck : Newton, p. 98, pi. 7, figs. 14-18. MATERIAL. Many specimens, including the holotype ^.48264, Newton's fig. 14). DESCRIPTION. Small, thick-shelled, inequivalve, right valve larger and more strongly inflated than the left, beaks submedian (on the right valve a little anterior to the middle line on account of the greater rostration), prosogyrous, right valve umbo strongly enrolled. Outline subtriangular. Right valve very strongly inflated, umbo prominent but not broad, posterior end with a moderately short, slightly upturned rostrum, with two blunt carinae pos- teriorly, limiting the upper and lower portions of the rostrum. Anterior end sharply rounded. Ventral margin gently convex, straighter posteriorly. Ornament con- sisting of strong, rounded concentric folds of about the same width as or slightly wider than their intervals, regularly increasing in size with growth. Valve margin internally smooth, with a fine incised line some distance from the edge for the recep- 68 EOCENE MOLLUSCA FROM NIGERIA tion of the margin of the smaller left valve. Tooth i solid, triangular, strongly upturned and pointed, with a deep, triangular chondrophore behind it ; PI short, lamellar. Pallial sinus widely rounded, reaching about one-third of the distance towards the anterior adductor impression. Inner half of rostrum with two short grooves extending downwards and outwards, one near the top, the other half-way down ; these are not associated with a left valve siphonal plate (such as is present in Caestocorbuld), no trace of which has been found on any of the specimens. Left valve fitting within the margin of the right, inner portion of right valve rostrum remaining visible. Valve oval-subtriangular, inflated (a little flattened posteriorly near the blunt carina), beak narrow, umbo with fine concentric folds, main portion of surface with growth-lines only, apart from four very fine, widely- spaced radial threads on its middle third. Tooth 2b oblique, simple, with a deep, triangular chondrophore in front of it ; there is a trace of a very small, pointed All at the anterior corner of the chondrophore. Pallial sinus as in the right valve. Valve margin smooth. REMARKS. The illustrations of Cossmann & Pissarro, and specimens from France in the British Museum (Natural History), suggest that the name Corbula rugosa Lamarck has been applied to two forms, one relatively finely ornamented, the other (probably a Varicorbula) more strongly ornamented, higher, and more inflated. The Nigerian form is not conspecific with either of these ; the rostrum and umbonal region are narrower, the shell is more triangular, and there is no distinct initial stage separated off by a constriction. Family KITSONIIDAE nov. TYPE GENUS. Kitsonia gen. nov. FAMILY CHARACTERS. Lithophaga-like in form. No gape. Not nacreous externally. Sinupalliate. Internally, that part of the surface behind the pallial sinus is highly polished and nacreous. Ligament external. No hinge-plate, the two right valve teeth and the one left valve tooth, all cardinals, project markedly beyond the plane of commissure. Apparently a boring form. REMARKS. This form belongs to a new family which is provisionally regarded as being related to the Clavagellidae and is placed in the Clavagellacea. Genus KITSONIA nov. TYPE SPECIES. Coralliophaga eocenica Newton. GENERIC CHARACTERS. Small, thin-shelled, equivalve, very elongate, ventral margin slightly excavated, posterior end slightly curved downwards. Beaks small, low, prosogyrous, situated slightly behind the anterior end. Surface smooth except for growth lines. Anterior adductor impression moderately broad and large ; posterior adductor impression situated medially just below the dorsal margin. Pallial sinus apparently extending for two-fifths of the length of the shell. That part of the surface anterior to the pallial sinus dull, white, with obscure, raised, rather vermicular, radial markings (reminiscent of the Lucinidae), that part posterior EOCENE MOLLUSCA FROM NIGERIA 69 to the sinus highly polished and nacreous. No lunule or escutcheon. Ligament narrow, rather long, external. Right valve : two lamellar subumbonal teeth, sub- horizontal and slightly ascending posteriorly, the anterior one relatively short, highest distally ; posterior tooth very obliquely bifid, its posterior portion very close to the anterior portion and largely behind it. Left valve : one long, lamellar, simple, subumbonal tooth, highest distally, subhorizontal, and slightly ascending posteriorly. No other teeth. Valve margins smooth internally. REMARKS. As noted by Newton, this form is quite different from the edentulous and gaping Gastrochaena and from the mytilid genus Lithophaga, which is integri- palliate and likewise edentulous. However, the lack of a hinge-plate, the entirely different dentition, and the lack of radial ornament and of wide-spaced concentric frills distinguish it readily from Coralliophaga. Kitsonia eocenica (Newton) (Plate 9, figs. 13, 14) 1922. Coralliophaga eocenica Newton, p. 99, pi. n, figs. 4-5. MATERIAL. Two specimens (a right valve and a left valve) ; lectotype ^.48192, Newton's fig. 5) here selected. III. ACKNOWLEDGMENTS The writer is indebted to the authorities of the British Museum (Natural History) for facilities to undertake the work ; to the Chairmen and Directors of the joint Shell and British Petroleum organization for permission to publish the results ; to Dr. L. R. Cox for advice at all stages of the investigation ; to palaeontologists of the Shell organization for stimulating discussions concerning the fauna and its age ; and to Mr. A. Chavan, who has collaborated in the study of the Lucinidae. IV. REFERENCES CASTER, K. E. 1938. Macroscopic fauna of the Quimbriz (Eocene) formation on the Luculo River, Angola. Com. Serv. geol. Portugal, Lisboa, 20 : 53-85, pis. i-io. CHAVAN, A. 1938. Un groupe africain des Carditides ; Cossm'annella Mayer-Eymar. Bull. Soc. geol. Fr., Paris (5) 8 : 3-10. 1944- Sur la portee d'un remarquable 61argissement des cotes chez les V6n6ricardes. C. R. Soc. geol. Fr., Paris, 1944, 4 : 34-36. 1951. Essai critique de classification des Divaricella. Bull. Inst. Sci. nat. Belg., Bruxelles, 27, 18 : 1-27. • 1952. A new fossil gastropod from the Middle Eocene of Ameki, Nigeria. Proc. Malac. Soc. Lond., 29 (2-3) : 80-82, fig. i. DARTEVELLE, E. 1952. Echinides fossiles du Congo beige et de 1'Angola, I. Introduction, his- torique et stratigraphique. Ann. Mus. Congo beige (8°) Sci. geol., Tervuren, 12 : 1-14, 23 figs. DARTEVELLE, E. & ROGER, J. 1954. Contribution a la connaissance de la faune du Miocene de 1'Angola. Com. Serv. geol. Portugal, Lisboa, 35 : 227-312, pis. 1-5. • !955- Platyodon klinghardti (J. Bohm, 1919). LameUibranche du Miocene d'Afrique occidentale. Bull. Mus. Nat. Hist. nat. Paris (2) 27, 2 : 164-169. 70 EOCENE MOLLUSCA FROM NIGERIA MACNEIL, F. S. 1938. Species and Genera of Tertiary Noetinae. U.S. Geol. Surv. Prof. Paper, 189A : 1-49, pis. 1-6. NEWTON, R. B. 1922. Eocene Mollusca from Nigeria. Geol. Surv. Nigeria Bull., 3 : 1-114, P^s- i-n. Rossi, C. 1942. Molluschi paleogenici della Sertica. Pubbl. 1st. Geol. Univ. Milano (P) 33 : 109-193, pis. i-n. SCHILDER, F. A. 1929. The Eocene Amphiperatidae and Cypraeidae of England. Proc. Malac. Soc. Land., 18 : 298-311. 1932. Cypraeacea. Fossilium Catalogus, i Animalia. Pars 55. 276 pp. Berlin. TESSIER, F. 1952. Contribution a la stratigraphie et a la paleontologie de la partie ouest du Senegal (Crdtace et Tertiare). These, Univ. Marseille, IIIe Partie : Paleontologie. WENZ, W. 1938-44. Gastropoda. In Schindewolf, O. H. Handbuch der Paldozoologie, 6, i. Berlin. WHITE, E. I. 1955. Notes on African tertiary sharks. Colonial Geology & Mineral Resources, London, 5 : 319-325, pis. i, 2. PLATE 5 FIG. i. Turritella amekiensis sp. nov. Holotype (0.42215). x i. FIG. 2. Turritella amekiensis sp. nov. Paratype (0.42222). x 3. FIGS. $a-c. Architectonica (Nipteraxis) bendeica sp. nov. Holotype (0.42373). X 3. FIG. 4. Architectonica (Stellaxis) bicingulata (Newton). Syntype (0.42356). X i. FIGS. 5 locule cordate at the base, carpel wall basally thickened). Family BETULACEAE Genus CORYLUS (Tourn.) Corylus sp. (PI. 13, figs. 49-5i) Part of a male inflorescence with anthers (PI. 13, figs. 49, 50) enclosing pollen examined and determined by Dr. J. B. Simpson. The pollen, which was flattened, contained no intine ; its condition made it clear that it was a true fossil, not Recent material accidentally introduced. The grains, like those of Corylus, show the typical thickening of the exine in the neighbourhood of the pores, and the zone of granules around each pore, the granules being longer and not so close-set as in other regions of the exine. Various modern species of Corylus differ slightly from one another in the size and prominence of the zone of granules, but Dr. Simpson has not determined the nearest living species. Genus CARPINUS Linnaeus Carpinus boveyanus (Heer) pars (PI. 13, figs. 52-67) 1862. Carpolithes boveyanus Heer, p. 1077, pi. 70, figs. 7-14 (in part). DIAGNOSIS. Fruits sub-ovoid, 2-5-4 mm- l°ng. J '75-2*5 mm- broad. Sometimes twinned. NEOTYPE. A fruit showing fibres and scar of attachment. Brit. Mus. (H. N.), No. V.3387O. DESCRIPTION. Fruit : Attached to a bract rarely preserved and then only as a fragment at the base. Very variable in shape and size, enclosed by the abraded accrescent perianth rarely with remains of short superior perianth segments (PI. 13, fig. 57). On the whole sub-ovoid, bisymmetric and slightly compressed. Scar of attachment to bract basal, large, sub-circular (PI. 13, figs. 52-54). Surface in the THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 93 better-preserved specimens showing slender longitudinal vascular bundles which arise from the margin of the basal scar ; the strands themselves are often abraded but their position may be indicated by furrows (PI. 13, figs. 52, 54, 56). Surface smooth, formed of small, oblong, longitudinally aligned cells about 0-008 mm. in diameter. A few specimens show two fruits grown together. If, as appears likely, they really belong to Carpinus they may be due to the rare development of two fruits in a bract, for two female flowers are present in the early stages in living Carpinus. I have not been able to find any living twinned fruits to bear out this suggestion. Twinned fossils are shown in PI. 13, figs. 59-65). Endocarp : One-loculed, agreeing with the fruit in shape, style terminal forming a mucro when well preserved, showing remains of two style bases. Dehiscence along a marginal suture in the plane of symmetry. Cavity not well seen. Seed : When preserved much shrivelled, its placentation obscure. Length of best preserved fruits, 2-5-4 mm.; breadth, 1-75-2-5 mm. Scar breadth, 0-9-1 mm. REMARKS AND AFFINITIES. A few fruits among a mass of small, black, crushed fruits and seeds are attributable beyond doubt to Carpinus. Possibly some of these were included in Heer's Carpolithes boveyanus which certainly included specimens of Myrica. The presence of an undoubted accrescent superior perianth, and the basal scar of attachment to a bract make the relationship to Carpinus clear. It is unfortunate that there is no evidence so far of the pendulous anatropous seeds which together with perianth and scar serve to distinguish this species from Myrica. The same or a closely allied species of Carpinus occurs in the Bournemouth Marine Beds, at Cliff End near Mudeford, and in the Lower Headon of Hordle. Family FAGACEAE Genus FAGUS Linnaeus Fagus minima n. sp. (PL 13, figs. 68-70) DIAGNOSIS. Endocarp unusually small, 3-75-4-5 mm. long ; lateral faces, 2-5 mm., 2-2 mm. and i-i mm. broad respectively. HOLOTYPE. An endocarp. Brit. Mus. (N. H.), No. ¥.33886. DESCRIPTION. Endocarp : Trigonous with three unequal flat or concave sides. Broadest outline semi-oval, style apiculate, base with scar of attachment, i mm. in diameter, reaching the margin of the narrow face but not continued on to it as in the case of the other two faces, thus showing it to be one of a pair of nuts which abutted along the narrow face ; angles sharp but not flanged. Surface smooth, but cell walls raised so as to form minute, irregular, sinuous, longitudinal crumples 0-008 mm. apart, also having irregularly rounded dimples 0-012-0-019 mm. in diameter which are so crowded towards the apex as to produce a rough surface ; they may represent hair-bases. Length of endocarp, 3-75 mm.; breadth of broadest side, 2-5 mm.; 94 THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN maximum breadth of narrowest side, i-i mm. (? originally about 1-75 mm. now folded in) ; maximum breadth of medium side, 2-2 mm. Length of endocarp found by C. & E. M. Reid (see below), 4-5 mm. REMARKS AND AFFINITIES. One endocarp. Two others were figured by C. & E. M. Reid from Bovey (1910, pi. 16, figs. 67, 68). The angled form, character of attach- ment scar and surface, position and character of style, all indicate relationship with Fagus. The small size distinguishes it from any living species seen. No other fossil species so small has been recognized. The compressed form consequent on the narrowness of one side, and the correspondingly narrow triangular scar which terminates at the margin of the narrow side suggest that the endocarp was developed in a laterally compressed cupule. Fagus leaves are of common occurrence in Cre- taceous and Tertiary deposits. Leaves of Fagus and of Nothofagus (Bandulska, 1924) occur in the Bournemouth Freshwater Beds. Family ULMACEAE Section CELTIDOIDEAE Genus ZELKOVA Spach. Zelkova boveyana n. sp. (PL 13, figs. 71-73) DIAGNOSIS. Endocarp much inflated. Maximum diameter, 1-5 mm.; maximum diameter in plane of symmetry, 0-9 mm.; maximum diameter at right angles to plane of symmetry, 1-25 mm. HOLOTYPE. An endocarp. Brit. Mus. (N. H.), No. ¥.33887. DESCRIPTION. Endocarp : Approximately bisymmetric about a plane through the attachment and style, marked by a conspicuous marginal ridge (PL 13, figs. 72 73) ; asymmetric in this plane of symmetry about a line between the attachment and style so that the outline is gibbous on one margin near the attachment, and on the other near the apex (PL 13, fig. 71). The endocarp is much inflated producing a rounded-quadrilateral outline at right angles to the plane of symmetry (PL 13, fig. 72). Attachment indicated by a small depression from which obscure ridges diverge, a few also diverge from the marginal ridge. Surface rough with indefinite depressions that give rise to an obscure network, wall formed of small equiaxial cells 0-012 mm. in diameter. Maximum diameter, 1-5 mm.; maximum diameter in plane of symmetry, 0-9 mm.; maximum diameter at right angles to plane of symmetry, 1-25 mm. REMARKS AND AFFINITIES. One endocarp. The form and structure so far as it has been seen, relate the fruit to Zelkova, a genus represented by about six living species in North Temperate regions. All species seen are much larger than the fossil, but they vary much in size, so that the relationship in size between the fossil and Z. keaki (for example) is comparable with that between Z. keaki (3 mm. maximum diameter) and Z. sinica (7 mm.). Size alone could not therefore exclude the fossil from this genus. Z. sinica resembles THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 95 the fossil closely in form, but most living species are more markedly curved, and some are actually hooked in the stylar region. The apparent absence of superficial fibres in the fossil may be due to its worn condition. There appear to be no other grounds for excluding it from the genus Zelkova. Family MORACEAE Section MOROIDEAE Genus MO RO IDE A nov. DIAGNOSIS. Unidentified genera of the section Moroideae, family Moraceae. Moroidea boveyana n. sp. (PI. 13, fig. 74) DIAGNOSIS. Fruit markedly asymmetric in plane of symmetry. Stylar projection narrow. Length incomplete ; breadth in plane of symmetry, 1-5 mm.; thickness, 07 mm. HOLOTYPE. A fruit, broken at the rounded base. Brit. Mus. (N. H.), No. V.33888. DESCRIPTION. Fruit : Sub-circular in outline, somewhat flattened laterally but sub-cuneate in transverse section, the narrower edge crested along the whole length preserved, the opposite edge and base of the fruit rounded. Style prominent, terminal at the apex of the crested margin, closely adjacent to a sub-terminal curved projection which marks the point of entry of the funicle to the sub-apical placenta inside the rounded margin. Carpel wall 0-05-1 mm. thick, formed of a few layers of small cells aligned radially so as to give a columnar appearance in section, producing superficially a closely punctate surface with pits 6-oi2-o-oi6 mm. in diameter. Length of fruit, incomplete ; breadth, 1-5 mm.; thickness, 0-7 mm. REMARKS. One fruit, broken at the rounded base. The interior and structure of the seed have not been seen. A similar fruit was found at Hordle giving evidence of both internal structure and of seed structure. In the Hordle specimen the relation to Moraceae was clear, and the evidence indicated a connexion either with the section Moroideae or with Artocarpoideae, probably with the former. The chief distinctions between the Bovey and Hordle fossils lie in the style which is a narrow projection in the Bovey fruit, and a broad flat one with a broad flat stylar canal in the fruit from Hordle, and in the greater symmetry of the Hordle fruit. Such differences are probably of specific value, but this conclusion can only be established when a greater range of living and fossil material is available for comparison. In the meantime the Bovey and Hordle fruits are treated as specifically distinct. 96 THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN Family NYMPHAEACEAE Genus BRASENIA Schreber Brasenia ovula (Brongniart) (PI. 13, fig. 75) 1862. Nymphaea doris Heer, p. 1072, pi. 70, figs. 32-37. 1925. Brasenia sp. (B. ovulum Brongn. ?) Chandler, p. 23, pi. 3, figs. ja-d. 1926. Brasenia ovula (Brongn.) : Reid & Chandler, p. 99, pi. 6, figs. 15-18. DESCRIPTION. Seed : Obovoid, now much crumpled and distorted, having an aperture at one end, measuring 0-45 mm. in diameter, from which the embryotega has come away. Surface black, glistening, with longitudinal corrugations about 0-075 mm. in breadth at the middle of the seed. Surface cells, 0-05-0-075 mm. in diameter, with interlocking digitations, the length of individual digitations often more than one-third of the total diameter of a cell ; surface of cells finely punctate. Length of seed, 2-25 mm.; breadth, 1-5 mm. REMARKS. One seed from Heathfield. Heer recorded numerous seeds from Bovey under the name Nymphaea doris. His seeds were 2-5-3-5 mm. long and 2-3 mm. broad, the diameter being slightly increased no doubt by the flattening of the the seeds. The species ranges in the British Tertiary from Bournemouth Marine to Hamstead Beds. Family MAGNOLIACEAE Genus MAGNOLIA Linnaeus Magnolia boveyana n. sp. (PI. 13, figs. 76-80) 1910. Magnolia attenuata Weber : C. & E. M. Reid, p. 165, pi. 15, figs, i, 2. DIAGNOSIS. Seeds longer than broad. Length 6-7 mm., breadth 2-75-4-5 mm. HOLOTYPE. Brit. Mus. (N. H.), No. ¥.33890. DESCRIPTION. Seed : Anatropous, ovate in outline, narrow, longer than broad, much compressed (compression doubtless emphasized by f ossilization) , gently convex, angled longitudinally on one face with a shallow longitudinal depression on the other (the raphe side). Chalaza terminal at the broad end, marked by a small plug or scar which is pierced at the centre. Surface marked by a fine " finger- print " pattern due to polygonal cells, o-oi mm. in diameter, cells aligned in rows, the rows being grouped in clusters. Testa (represented only by the inner hard coat) 0-4 mm. thick at the middle of the seed, formed of equiaxial cells arranged radially in a columnar manner, the columns about 0-016 mm. broad. Tegmen thin, translucent, longitudinally striate, structure obscure. Length of a seed, 6 mm.; breadth, 2-75 mm. Length of a second seed, 6 mm.; breadth, 4 mm. Length of a seed found by C. & E. M. Reid (1910 : 165), 7 mm.; breadth 4-5 mm. THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 97 REMARKS AND AFFINITIES. Six seeds and several fragments. They resemble Magnolia seeds of the American longer-than-broad type. The species is larger, more ovate in outline, and less triangular in transverse section than M. angusta from the London Clay (Reid & Chandler, 1933 : 177, pi. 5, figs. 6-8). Comparable living species are M. grandiflora Linn., and M. glauca Linn., but the former is a large, and the latter is a smaller species. C. & E. M. Reid (1910 : 165) named these seeds M. attenuata Weber, identifying them with Magnolia seeds found abundantly in the Rhine lignite ; but they were careful to indicate that as the type of M. attenuata was a leaf, and as another Magnolia species also occurred in the same deposit, the connexion of the Rhineland seeds with the leaves was not conclusively established. On this account Kirchheimer (1936^ : 85-86) recently instituted a new specific name M . sinuata for a species from Salzhausen in which he included the relatively narrow seeds of M. attenuata Weber and a broader-seeded species M. hoffmani Ludwig on the grounds of histological identity. M. sinuata, although some of its forms resemble the Heathfield specimens, appears to exhibit much greater variation both of form and size. On the whole it is appreciably larger, the length of the seed, 6-10 mm.; breadth, 5-9 mm., whereas the maximum length of the Heathfield seeds is 7 mm.; maximum breadth, 4-5 mm. Hence a distinct specific name, Magnolia boveyana, has been given to the seeds from the Bovey basin. Family LAURACEAE Genus CINNAMOMUM Blume Cinnamomum is represented at Bovey, according to Heer, by three species, two based on leaves (C. rossmassleri and C. lanceolatum) , the third on leaves and flowers (C. scheuchzeri). Probably some of the cupules and berries hereafter recorded may belong to Cinnamomum. At present, however, they are referred to the family Lauraceae only, the evidence being insufficient for definite determination. C. rossmassleri Heer was represented by two leaf-fragments both with the apex missing. Hence Heer himself regarded the determination as doubtful. The species is therefore omitted in the list on p. 77 as the figures (Heer, 1862, pi. 67, figs. 17, 18) are inconclusive and not distinguishable with certainty from those of C. scheuchzeri. C. scheuchzeri and C. lanceolatum are better represented, and there can be no reason to doubt the occurrence of the genus Cinnamomum in the Bovey Lake deposits. No new leaves of either species have been found and the genus is not therefore included in the plant list on p. 77. VARIOUS GENERA AND SPECIES (PI. 14, figs. 81-91) In addition to the genus Cinnamomum, the Lauraceae are represented by leaves referred to Laurus primigenia Unger and to Daphnogene ungeri (Heer, 1862 : 1064, g8 THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN pi. 65, figs, i, 2, 6). Neither of these determinations are wholly satisfactory. There are a number of cupules and berries also. These were never seen by Heer or their affinities were not recognized. C. & E. M. Reid (1910, pi. 16, figs. 64-66) figured three small wrinkled cupules clearly belonging to Lauraceae. Similar specimens (all small) are common both at Bovey and Heathfield. They vary from about 1*75-3 mm. in diameter. Sometimes they are cup-like with simple margins, sometimes they are notched or divided above into sepals. In a few the small unripe berry still lies within the calyx. In all the skin is much wrinkled, shining, formed of very small cells which may produce a finely striate effect. The flattened skins or epicarp of larger, detached, formerly ovoid berries are also common. The epicarp is leathery, shining, often yellowish-brown and semi- translucent, mainly formed of equiaxial cells which may vary considerably both in form and size ; they are often about 0-025 rnm. in diameter. In some specimens, overlying these cells at the apex is a thin layer of polygonal cells about 0-05 mm. long and 0-025 rnm. broad. These diverge from the apical scar, but quickly die out and become obscure so that they cannot be traced a short distance below the apex. Remains of the mesocarp commonly adhere to the epicarp and in some specimens enclose numerous ovoid or globular yellow oily (?) bodies about 0-05 mm. in diameter. The largest berry seen is 6 mm. long (incomplete) by 5-25 mm. broad (breadth increased by flattening). It is possible, but unlikely in view of the immature condition, that detailed com- parative study of cuticle and mesocarp structure in living and fossil material might serve to distinguish some at least of the genera represented, but it would require very long research and an abundance of living material for comparison. The berries and cupules can therefore only be referred to the family Lauraceae without suggestions as to the generic relationship. Family CAPPARIDACEAE Genus CAPPARIDISPERMUM nov. DIAGNOSIS. A form-genus to embrace seeds of Capparidaceae of which the nearer relationship is unknown. Capparidispermum boveyanum n. sp. (PI. 14, figs. 92-96) DIAGNOSIS. Seeds transversely oboval about 2-3 mm. in minimum and 2-75-3-75 mm. in maximum diameter. Contiguous walls of the curved limbs form a condyle, they appear to be fused for most of its length. Testa tubercled, the tubercles oriented parallel with the margin of the seed. Surface cells equiaxial. HOLOTYPE. Brit. Mus. (N. H.), No. ¥.33904. DESCRIPTION. Seed : Woody, transversely oboval in outline, almost flat, approximately bisymmetric, splitting for germination in the plane of symmetry. THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 99 Locule with curved unequal limbs separated by a narrow curved condyle, the micropylar limb longer and narrower than the other (PI. 14, fig. 96). The sutures in the plane of dehiscence are smooth finished surfaces both along the margins of the seed and on the condyle. Marginal suture variable in breadth, 0-18 mm. broad at the distal end of the seed ; 0-4 mm. broad near the proximal end on the concave outer curve of the short limb. Hilar scar large, oval, marginal between the limbs, sometimes sunk in an emargination (PI. 14, fig. 94). Micropyle small, terminal on the longer limb. The walls which form the condyle appear to be completely fused for the greater part of their length except at the hilar end where they enclose a shallow V-shaped cavity (PL 14, fig. 96), 0-8 mm. long, 0-3 mm. wide at the margin. The condyle is sometimes indicated externally by an elongate prominence which is smoother than the rest of the surface (PL 14, fig. 92) . Testa tubercled externally, the tubercles often elongate parallel with the margin of the seed ; breadth of bases of tubercles, 0-05-1-5 mm.; height, 0-025-0-05 mm., surface cells equiaxial about 0-025 mm. in diameter giving the margins of the tubercles a clawed or digitate appear- ance. Testa in section formed of equiaxial cells, 0-017 nun. in diameter. Lining cells of seed-cavity, 0-017 mm. in diameter, equiaxial, in rows transverse to the length of the cavity. Dimensions of four seeds respectively : 2-75 by 2 mm., 3-75 by 3 mm., 3-5 by 2-25 mm., 3-5 by 2-25 mm. (the last represented by one valve only). REMARKS. Four seeds and two fragments. The curvature of the locule, marginal hilum between the limbs and the long narrow condyle indicate relationship with Capparidaceae. No living genus seen combines the characters of form, size and ornamentation, hence the reference to a form-genus Capparidispermum. It is possible that the specimen shown in PL 14, fig. 95, represents a second species, it is larger, flatter, and the short limb has a more marked spiral incurve. In the absence of more evidence it is regarded as a slightly abnormal seed of the same species. Family HAMAMELIDACEAE Genus ? sp. (PL 14, figs. 97, 98) DESCRIPTION. Seed : Sub-oboval in outline, rounded at the apex, slightly excavated at the base, laterally compressed (compression exagerrated by fossiliza- tion), with a large, bilobed, slightly concave hilar scar lying across the base having one lobe on each flat face ; the lobes are unequal in length and breadth but occupy about half the length of the seed. The arrangement of the other organs has not been seen. Surface much abraded, rough as preserved, the polygonal cells being about 0-03 mm. in diameter and somewhat sinuous. In a few places the testa has cracked transversely on drying owing to the transverse alignment of one or more of its layers. As seen in section near the apex it is 0-25 mm. thick but its cells cannot ioo THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN here be distinguished although at one point there is a false appearance of curved columns, 0-016 mm. broad, the results of the fracture of tissue which has become vitrified in fossilization. Lining of seed-cavity formed of equiaxial cells o-o 12-0-016 mm. in diameter. Length of seed, 5-25 mm.; breadth, 3-25 mm.; thickness, 2 mm. REMARKS AND AFFINITIES. One seed. The form and the large lobed hilar scar extending over the base and continued on each side, indicate a seed of Hamamelidaceae. Although many seeds in this family show similar hilar scars, no living genus seen has so large a scar as the fossil, while in many it is considerably smaller. In certain living genera there are two distinct scars one on each side near the base, but not united over it. Sinowilsonia (length of seed, 6-5 mm.; breadth, 4 mm.) has a similar scar occupying almost half the length of the seed. Fortunearia, with a scar extending about one-third of the length, has a much larger seed as have many species of Hamamelis. While the relationship of the fossil to the family Hamamelidaceae is certain, the evidence is insufficient to determine the generic position, but perhaps the closest living genus is Sinowilsonia. It is interesting to note that Dr. J. B. Simpson (1936 : 99) records the occurrence of pollen belonging to Bucklandia, Corylopsis, Fortunearia, Loropetalum, Dicoryphe and Distylium in the Scottish Tertiary coals of Ardnamurchan and Mull. Genus ? sp. (PL 14, fig. 99) DESCRIPTION. Fruit : Represented only by a fragment of septum and fibrous axis adhering to the seed near its apex. Seed : Originally ovoid (much compressed in fossilization) . Hilar scar sunk, long and narrow, lateral but continued across the base where it terminates without extending on to the opposite face. Testa black, shining, showing the cell-structure fairly clearly ; around the scar the cells are elongate, approximately parallel with its margin, but over most of the surface there are fine parallel striations diverging obliquely from the middle which appear to be a secondary consequence of compression. Length of seed, 4-75 mm. ; maximum breadth, 2-5 mm. Length of scar on lateral face, 2 mm.; breadth, 0-3 mm. REMARKS AND AFFINITIES. One seed from Heathfield. The form, hilar scar, and surface all indicate relationship with Hamamelidaceae. It has not been possible to relate it to a genus. In Hamamelis the hilar scar is shorter and the seed larger. In Corylopsis, which has a long, sunk scar, there is also a marked facetting on the opposite side. F other gilla has a small, sunk scar and the walls of the surface cells are much thicker. The second species here described is quite distinct from the first in which the hilar scar is large, broad and bilobed. THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 101 Family ROSACEAE Genus RUBUS Linnaeus Rubus microspermus C. & E. M. Reid (PL 14, figs. 100-109) 1910. Rubus microspermus C. & E. M. Reid, p. 169, pi. 15, figs. 13-17. DESCRIPTION. Endocarp : Laterally compressed, semi-circular, semi-oval or sub-ovate in outline, ventral margin straight or almost straight, dorsal margin semi-circular or markedly convex, base rounded, margin rimmed all round. Surface reticulate with conspicuous pits, angular in outline, separated by thin, sharp, clearly defined ridges, occasionally the ridges are prolonged on to the margin. Carpel wall formed superficially of small equiaxial cells o-oi mm. in diameter. Length of endocarp, 1-2-5 mm- (commonly 2-25 mm.) ; breadth, 1-1-8 mm. (commonly 1-25 mm.). REMARKS AND AFFINITIES. Endocarps of this species are common at Bovey and Heathfield. Their relationship to Rubus was discussed by C. & E. M. Reid (1910 : 169) but a larger range of material is here shown (PI. 14, figs. 100-109) to demonstrate variation in size and form. Prickles of Rubus which were reasonably assumed to belong to the same plant were also described and figured by C. & E. M. Reid (1910 : 169, pi. 15, figs. 16, 17). The endocarps have been compared with Rubus acutiformis Chandler which occurs at Hordle, Cliff End, Sandbanks, Branksome Dene and Stud- land, but the two species are clearly distinguishable although both are of unusually small size. Rubus acutiformis is commonly more pointed and narrower at the apex, and is therefore more ovate in outline as a rule, while the ventral margin is not infrequently very slightly concave, the apex being curved somewhat towards the ventral side. The well-marked marginal flange is most conspicuously developed on the ventral margin. > Family LEGUMINOSAE Genus ? (PL 15, figs. 110-112) DESCRIPTION. One perfect compressed seed, and fragments of three others must be referred to Leguminosae. The original shape was probably lensiform (but the seeds are now flat owing to compression), the marginal hilar scar (obscured by margi- nal cracking) must have been small. The surface is formed of inconspicuous concave equiaxial cells 0-02 mm. in diameter, the walls are 0-25 mm. thick as seen in one of the incomplete specimens, 0-45 mm. thick around the hilar aperture, the cells being arranged in a columnar manner, the radial columns about 0-012 mm. broad ; a shallow pocket, presumably connected with the radicle, lies immediately beneath the hilum, it is delimited on the surface of the cavity by elongate cells. Lining of main seed-cavity of convex equiaxial cells 0-012 mm. in diameter. Diameter of the perfect seed, 3-5-3-75 mm.; other specimens larger but incomplete. 102 THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN Family RUTACEAE Section ZANTHOXYLEAE Genus RUTASPERMUM nov. DIAGNOSIS. A form-genus to include seeds of Rutaceae of which the nearer relationship is unknown. Rutaspermum exaratum (Heer) 1862. Carpolithes exaratus Heer, p. 1079, pi. 70, figs. 24-27. The seed was described by Heer as 3-5 mm. long, 3-3 mm. broad ; the figures showed it as sub-circular in outline, much inflated, with a long, narrowly-triangular, sunk hilar scar (the " umbilical fissure " of Heer). The surface was ornamented with nodular ridges aligned parallel with the rounded dorsal margin, much as in a species to be described from Sandbanks, and ornamented also with fine polygonal cells or pits (the " innumerable dots " of Heer) ; but the species is larger than the Sandbanks seeds. The characters are clearly those of Zanthoxyleae, of the type referred to the form-genus Rutaspermum. In the absence of actual specimens and of more accurate figures than the diagrammatic illustrations given by Heer, the species cannot be clearly defined. Available evidence, however, suggests that it is of a very distinctive type. Family SABIACEAE Genus MELIOSMA Blume Meliosma reticulata (C. & E. M. Reid) (PI. 15, figs. 113-118) 1910. Calvarinus reticulatus C. & E. M. Reid, p. 169, pi. 15, figs. 18-20. DIAGNOSIS. Endocarp about 5-5-25 mm. long, 3-75-4 mm. in maximum trans- verse diameter. External surface with about eighteen to twenty raised ribs over the proximal half which branch and anastomose to form a sharp network over the distal half. NEOTYPE. A perfect but laterally compressed endocarp. Brit. Mus. (N. H.), No. ^33924. DESCRIPTION. Endocarp : Woody, obovoid, slightly compressed laterally (com- pression increased by fossilization), the transverse diameters being in the proportion of 7 : 10 ; bisymmetric about a plane which passes through the attachment, funicular canal and a marked marginal angle ; splitting in the plane of symmetry into two valves. Funicular canal oblique, about I mm. long, placenta basi-lateral or sub- basal. External surface ornamented with about eighteen to twenty raised ribs diverging from the attachment over the lower half of the fruit and uniting, branching, and anastomosing, to form a network over the upper half of the fruit. Surface formed of polygonal cells o-oi mm. in diameter. Walls i-25-i'5 mm. thick. Surface THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 103 of locule rough, cell-structure very obscure, but the centipede-type of cells with interlocking walls characteristic of Meliosma can be traced oriented parallel with the lateral ribs on one small fragment ; the length of individual cells is obscure, but near the base their width is about 0-037 nim. Length of endocarp, 5-25 mm.; breadth, 4-25 mm. Length of a second endocarp, 5 mm.; breadth, 3-75 mm. REMARKS AND AFFINITIES. Six endocarps or valves and a number of fragments. The form, surface ornamentation, structure, and short oblique funicular canal relate these fossils to Meliosma. So far as it has been possible to study the different living species, the Bovey fossils are unique in the number of their longitudinal ribs and the fineness and sharpness of the apical network, also in the narrow oboval form. In living species the three diameters (two transverse and one longitudinal) are usually approximately equal but occasionally the transverse diameters may be slightly longer or shorter than the longitudinal. The considerable differences in the diameters of the fossil which give rise to the elongate form and lateral com- pression have not, however, been seen in the living forms. C. & E. M. Reid (1910) described the species under the generic name Calvarinus, and referred it to the family Boraginaceae. There were two specimens from Heath- field and one from Bovey. Reid and Chandler also obtained endocarps from both localities. Family VITACEAE There are several types of vine seeds at Bovey and Heathfield. Heer (1862 : 1070, pi. 69, figs. 25-29) distinguished two species which he named Vitis britannica and V. hookeri. The figures are poor and the descriptions inadequate but the outstanding features are clear. C. & E. M. Reid (1910 : 165, pi. 15, figs. 3-6) distinguished three species, V. hookeri Heer, V. teutonica A. Br. and V. ludwigi A. Br. The greater number of specimens are now referred to Parthenocissus britannica (Heer). One imperfect seed appears to belong to V. hookeri Heer. The specific relationship is discussed in the following pages where the species are described. Two new species have been added, Parthenocissus boveyana from Bovey, and Vitis stipitata from Heathfield. Genus PARTHENOCISSUS Planchon Parthenocissus britannica (Heer) (PL 15, figs. 119-122) 1862. Vitis britannica Heer, p. 1071, pi. 69, figs. 25, 26. 1910. Vitis ludwigi A. Br. : C. & E. M. Reid, p. 166, pi. 15, fig. 6 (not fig. 4 as in text). 1910. Vitis teutonica A. Br. : C. & E. M. Reid, p. 166, pi. 15, figs. 4, 5 (not fig. 6 as in text). DIAGNOSIS. Seed pointed-obovate in outline, smooth dorsally, slightly emarginate at the apex, chalaza elongate-ovate, surface grooved between chalaza and base ; sharply angled ventrally with infolds occupying more than half the length, deep, GEOL. Ill, 3. 9 104 THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN narrow, straight but diverging upwards. Length, 4-75 mm.; breadth, 2-25-2-8 mm.; thickness, 1-5-2 mm. NEOTYPE. A typical seed. Brit. Mus. (N. H.), No. V.^g2y. DESCRIPTION. Seed : Pointed-obovate in outline, slightly emarginate at the apex, pointed at the base, with smooth contours, ventral and dorsal faces meeting at an acute angle, ventral face sharply facetted so as to form a conspicuous raphe ridge extending almost the whole length of the seed, the facets being flat or concave and meeting approximately at a right angle ; at the apex the ridge gives place to the shallow channel which produces the slight apical emargination. Ventral infolds deep, narrow, straight, occupying more than half the length of the seed arising near the base, diverging upwards so that the raphe ridge is broader above than below. Dorsal face flat, slightly emarginate at the apex having a shallow groove along which the raphe passes into the narrow elongate-ovate chalaza which lies above the middle of the seed. From the chalaza a well-marked groove passes to the base. Testa formed of two coats, the outer thin, its irregular elongate polygonal cells about 0-02 mm. in shortest diameter producing transverse striations which diverge from the chalaza and ventral infolds to the margin. The inner coat is hard, averaging o-i mm. in thickness, formed of cells 0-016 mm. in diameter arranged in radial columns. These cells give a finely pitted surface to the coat. Length of seed, 4-75 mm.; breadth, 2-25-2-8 mm.; thickness, 1-5-2 mm. (somewhat distorted). REMARKS AND AFFINITIES. Ten seeds and several fragments from Bovey. The smooth seed with long divergent infolds suggests relationship with the living Partheno- cissus. The same species was apparently described by C. & E. M. Reid (1910) under the name Vitis teutonica A. Br. They state that " Heer's type specimens of V. brilannica appear to be nothing but badly compressed seeds of this vine [V. teutonica}, though his figures are scarcely recognisable". Their description is " ovate-acuminate, gradually narrowed into the beak, granulate all over, inner face with long shallow pits, outer convex longitudinally sulcate with a narrow pyriform chalaza, length 4 mm." The identity of these specimens with Vitis teutonica A. Br. from the German lignite cannot now be maintained. V. teutonica was originally based on leaves, but as seeds were associated with them, Unger applied the name to the seeds also (Braun, 1845 : 172 ; Unger, 1860 : 23, pi. 9, figs. 1-8). Unfortunately neither Unger's figures or diagnosis really define the species. Later German workers have also referred a variety of seeds from the Brown Coal to V. teutonica. Thus Krausel (1920, pi. 25, figs, i, 2) illustrates seeds with short, wide, divergent lateral infolds, a marked apical groove on the ventral side, a short elongate-obovate chalaza and shallow furrow between the chalaza and base on the dorsal side (pi. 24, figs. 20-23). His seeds appear more inflated than the Bovey specimens, and are sometimes fluted. Kirchheimer (1934 : 35, pi. 9, figs. 3-6) describes and figures a somewhat fluted seed under the name V. teutonica. It has a relatively small oval chalaza and appears quite distinct from the Bovey seeds. Later Kirchheimer (1938, pi. 4, figs. 12-15 i I939> pl- 2, fig. 3) figures other seeds under this name. The first shows a small median oval chalaza and very wide sub-parallel ventral THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 105 infolds. The second shows a much larger chalaza which is elongate-oval in shape occupying the upper half of the seed, while wide and divergent ventral infolds and a markedly stipitate base arise out of the rounded, smooth, lower half of the seed. The Bovey fossil appears to be sufficiently distinctive in its shape and chalazal character and in the narrow upwardly divergent infolds to merit a distinct specific name and Parthenocissus britannica (Heer) is here retained for this purpose. Another Bovey seed described and figured by C. & E. M. Reid (1910 : 166, pi. 16, fig. 6) as V. ludwigi ? is also probably a differently distorted specimen of P. britannica, although in the description it is said to have a more slender form with an oval, not pyriform, chalaza. Such individual differences may, however, occur among living seeds within a species. Parthenocissus boveyana n. sp. (PI. 15, figs. 123-125) DIAGNOSIS. Seed obovate in outline, not emarginate at the apex, contours smooth, chalaza median oval, angle of raphe ridge about 135°, ventral infolds markedly divergent upwards. Length, 3-5 mm.; breadth, 2-75 mm.; thickness, i mm. HOLOTYPE. Brit. Mus. (N. H.), No. ¥.33929. DESCRIPTION. Seed : Obovate in outline, not emarginate at the apex, pointed at the base, contours smooth, ventral face facetted, the facets being slightly concave and meeting at an angle of about 135° so that the raphe-ridge is not very sharp or conspicuous ; the raphe itself — a stout cord — is preserved above the middle on the ventral face and is continued on to the dorsal face passing into the external chalaza. Ventral infolds deep and narrow, extending from near the base to a quarter of the length from the apex, they diverge above and are slightly convex towards the raphe-ridge which is therefore triangular. Dorsal face almost flat, very slightly convex, not emarginate at the apex, or if it is grooved, the groove is obscured by the preservation of the raphe ; there is a shallow median groove between the chalaza and the base. Chalaza oval situated rather above the middle of the seed, gradually narrowing above into the raphe. Surface rather rough especially at the base and apex and on the chalaza, cells diverging from the lateral infolds and from the chalaza, many of them elongate in the direction of divergence, others polygonal 0-012 mm. in diameter. Wall as seen in section columnar, the columns about 0-012 mm. in diameter. Thickness of wall, o-i mm. Length of seed, 3-5 mm.; breadth, 2-75 mm.; maximum thickness, i mm. REMARKS AND AFFINITIES. One seed, and an imperfect specimen possibly referable to this species. The seed is relatively broader than seeds of V. britannica, and has a larger, broader chalaza, and more divergent lateral infolds. It is less rounded and stipitate than Kirchheimer's figures of V. teutonica (1939, pi. 2, figs. ja-e) and lacks the emarginate apex. Its chalaza is much larger than that in Kirchheimer's figure (1938, pi. 4, fig. 14). io6 THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN Genus VITIS Linnaeus Vitis hookeri Heer (PI. 15, figs. 126, 127) 1862. Vitis hookeri Heer, p. 1070, pi. 69, figs. 27-29. 1910. Vitis hookeri Heer : C. & E. M. Reid, p. 165, pi. 15, fig. 3. DESCRIPTION. Seed : Broadly obovate in outline but scarcely emarginate at the apex, pointed at the base, contours smooth and rounded. Ventral face facetted so as to form a conspicuous raphe-ridge, ventral infolds broad widening upwards, about half as long as the seed. Dorsal face rounded with slight flutings diverging from the ovate chalaza which is situated above the middle of the seed. Between chalaza and base is a deep median groove. Testa, 0-075 mm. thick, formed of cells 0-0125 mm. in diameter which have a columnar radial arrangement ; they give rise superficially to a finely and evenly but deeply punctate surface. Length of seed, 3-75 mm.; estimated breadth (actually incomplete), 3-5 mm. REMARKS AND AFFINITIES. One incomplete seed from Heathfield ; also a second from Bovey which may possibly be referred to this species. This second specimen is much distorted being compressed from base to apex, but its features can be seen and measured, and agree with those described above. It is slightly stipitate. Except that the better preserved specimen is not stipitate, it shows characters which agree with those of Vitis hookeri Heer of which Heer found one seed at Bovey. There is general agreement in size with Heer's specimen (length, 3-5 mm; breadth, 3 mm.), the contours are in general similar to those in Heer's inadequate figure; the chalaza is of comparable size and, although not round, occupies a very similar position on the dorsal surface when due allowance is made for its displacement by distortion which has brought it nearer to the apex than it originally lay. In the second much distorted specimen, the chalaza is round and but slightly above the middle. Comparison of the ventral face cannot be satisfactorily made both because of the distortion of the new material and because Heer's figure cannot really represent the ventral aspect of any vine. While, therefore, identity with Heer's species V. hookeri is not indisputably established, it seems probable that the new specimens should be referred to that species. The presence or absence of stipitation is not in itself of great importance, for Recent grape-seeds show that this character varies considerably in a single species. C. &. E. M. Reid (1910 : 165, pi. 15, fig. 3) described as V. hookeri Heer a seed which is probably correctly so named, although its chalaza is somewhat smaller than that shown in Heer's type. But variation in size of the chalaza also occurs in living species. The ventral side of this seed was not shown. Family TILIACEAE Genus TILIA Linnaeus Tilia sp. (PI. 15, figs. 130-132) Two groups of anthers were obtained from Heathfield and were sent to Dr. J. B. Simpson for examination. In a letter (27.^.38) he reported that they yielded THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 107 typical pollen of Tilia, adding " I am not well acquainted with the pollen of many of the tropical genera of Tiliaceae, and so cannot positively exclude the other genera except Grewia and Luhea, but certainly they [the pollen grains] agree so perfectly with Tilia, that I feel sure it must be this genus they represent ". A further report after a greater variety of living material had been obtained and examined was expected but has not come to hand. But in a letter dated i.iv.38 Dr. Simpson adds in writing of the fossil pollen " The furrows are short and deep and have the appearance of pits and occur in the middle of the sides not at the rounded corners of the triangular grain. The exine ... is irregularly reticulate with a fine mesh ". Ross, working quite independently of Simpson recorded " cf. Tilia " based on pollen. Family LYTHRACEAE Genus MICRODIPTERA nov. DIAGNOSIS. Fruit many-seeded. Seeds anatropous, compressed at right angles to the germination valve and in the plane of symmetry. Valve an oval operculum as in Diclidocarya menzeli E. M. Reid in the lower part of the dorsal surface. Seed- body flanked by two thin lateral wings with spongy internal tissue. Wings more or less equally developed. TYPE SPECIES. Microdiptera major n. sp. from the Eocene of Sandbanks (awaiting full description). Brit. Mus. (N. H.), No. ¥.34249. Microdiptera parva n. sp. (PI. 15, figs. 133-149 '. Text-fig. 2) DIAGNOSIS. Lateral wings very thin, markedly concave on the ventral side, raphe straight and narrow, triangular form of seed rare. Maximum length of seed so far recorded, 1-25 mm.; maximum breadth, 1-75 mm. (2 mm. in Cliff End specimen) . HOLOTYPE. A seed. Brit. Mus. (N. H.), No. ¥.33934. DESCRIPTION. Seed : Anatropous, much compressed dorsi-ventrally ; oboval, sub-circular, transversely oval, irregular in outline, or occasionally broadly triangular, differentiated into a median elongate-oval body and thin lateral wings, convex on the dorsal surface, concave on the ventral. Raphe prominent, linear, median longitudinal on the ventral face flanked by the marked concavities of the wings, hilum marginal. Germination by an oval operculum on the lower half of the dorsal face, associated with the basal micropyle. Internal chalaza circular at the apex of the seed (Text-fig. 2). Testa formed superficially of a shining, black coat giving a netted surface with large cells or pits equiaxial at the apex, more irregular and elongate below. Over the operculum the cells are very regular, equiaxial, about 0-05 mm. in diameter, arranged in about seven regular longitudinal rows. Some specimens show in addition evidence of small cells, 0-01-0-012 mm. in diameter, also aligned in longitudinal rows. Within the superficial coat and clearly seen on io8 THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN abraded seeds is a layer of fine cells 0-012 mm. in diameter ; they are arranged so as to give rise to striations which curve around the operculum, across on to the wings, and lie parallel with the margin of the wings near the circumference of the seed. Near the margin, the wings are only 0-025 mm. thick, but they thicken towards the seed-body close to which they measure 0-06 mm. in thickness. The external coats are close-textured and no cells can be distinguished in section ; near the body the dorsal coat is 0-025 mm- thick, the ventral 0-012 mm., the middle layer is 0-025 mm- thick and is much less compact than the outer layers, but its cell-structure cannot be clearly seen. The coat surrounding the seed-cavity is hard and compact, it is 0-05 mm. thick on the dorsal side, 0-037 mm. thick on the ventral side. The seed- cavity is finely striate longitudinally. The tegmen is semi-translucent and is fused with the testa at the large black chalazal scar. FIG. 2. Microdiptera parva n. gen. et sp. Diagrammatic longitudinal section through seed-cavity, x 45 approx. The dimensions of a series of seeds are as follows : Length Breadth (mm.) (mm.) (1) • 1-25 - i-75 (2) . i-o . 1-25 (3) • i-o . 1-5 (4) i-o i-75 (5) • i-25 • i-o Typical dimensions of germination opercula are respectively 0-5 by 0-3 mm., 0-5 by 0-4 mm., 0-6 by 0-4 mm., 0-5 by 0-4 mm., 0-55 by 0-35 mm., 0-4 by 0-3 mm. REMARKS AND AFFINITIES. The characters are those of Lythraceae, more especially the position of the organs, the form of the anatropous seed and the structure and arrangement of the wings. The systematic position will be discussed in greater detail when the type species from Sandbanks is described in a forthcoming catalogue. At Bournemouth not only isolated seeds, but some still in the position of growth in a fragment of fruit were found. THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 109 Family NYSSACEAE Genus NYSSA Gronov. The genus Nyssa from Bovey is in urgent need of revision. Four supposed species, truncate at one end, must now be referred to Symplocos. They are Nyssa europaea Unger (Heer, 1862 : 1066, pi. 69, figs. 11-17 '> C. & E. M. Reid, 1910 : 67, pi. 15, fig. 9), Nyssa laevigata Heer (1862 : 1066, pi. 69, fig. 18), Nyssa microsperma Heer (1862 : 1067, pi. 69, fig. 24 ; C. & E. M. Reid, 1910 : 167) and Nyssa obovata Weber (C. & E. M. Reid, 1910 : 168, pi. 15, fig. 10). One specimen figured as Nyssa ornithobroma Unger by C. & E. M. Reid (1910 : 168, pi. 15, fig. n) does not agree with Unger's species and may be a Mastixia. Others are of uncertain relationship : N. striolata (Heer, 1862 : 1067, pi. 69, figs. 20-23) • The figures show an ovoid ribbed fruit with finer, closer, more regular ribbing than that of N. boveyana (now to be described). Heer's fig. 23 shows a mass of fruits lacking both the finer ribbing seen in his figs. 20-22, and the coarser ribbing of N. boveyana. The types in the Geological Survey Museum were so decayed in 1910 that the true nature of this endocarp could not be discovered, all finer features being obliterated. There remains in these earlier records Nyssa vertumni Unger (C. & E. M. Reid, 1910 : 168, pi. 15, fig. 12) which is undoubtedly Nyssa. There is indeed abundant evidence of the genus which occurs both at Bovey and Heathfield but the evidence is of one species only, now described as N. boveyana n. sp. Nyssa boveyana n. sp. (PI. 16, figs. 150-157) DIAGNOSIS. Endocarp sub-ovoid or less commonly sub-obovoid, tending to be arched, the dorsal side being less convex than the ventral, ventricose in the upper part of the ventral face. Germination valve triangular, relatively narrow in propor- tion to its length, 2-5-4-5 mm.; breadth, 2-5-3-25 mm. Length of endocarp, 6-8-11-5 mm.; breadth, 3-1-5-5 mm. HOLOTYPE. Brit. Mus. (N. H.), No. ¥.33944. DESCRIPTION. Endocarp : Sub-ovoid or less commonly sub-obovoid with a tendency to be less convex on the dorsal than on the ventral side, frequently ventricose in the upper part of the ventral surface, inflated and rather narrowed towards the base (but now usually flattened and often distorted in f ossilization) . The surface shows conspicuous, broad, rounded, longitudinal ribs and thin strands of fibres in the intervening narrow furrows ; the ribs, six to twelve in number, extend from base to apex on the ventral and lateral faces, are in general smoothly rounded but have a tendency to become nodular especially over the ventricosity, they are symmetrically no THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN placed, four or five being ventral and two lateral ; along the crests of the lateral pair at the top of the endocarp are the sutures of the dorsal germination valve. Valve sub-triangular, gaping at the apex and along the sides which are slightly incurved towards the base, breaking irregularly along its base which is also that of the triangle. On the dorsal surface there is a thin thread-like median rib, sometimes flanked on each side near the base by a broad scarcely raised rib, these die out before reaching the valve. Length of endocarp, 6-8-11-5 mm.; breadth, 3-1-5-5 mm. The dimensions of a series of endocarps, in millimetres, are as follows : 9-5 x 4-6, 10 x 4-6, 10 x 4, 9 x 5-1, 9 x 6, ii x 3-1, 9 x 4-2, n x 4-2,7-2 x 4-2, 9-8 x 4-5, 8 x 3-8, 9-8 x 5'5, 9x4, 10-5 x 4-8, 9 x 4-6, 7-4 x 4-9, 8x4, 7-2 x 4-1, 8 x 37. 9-5 x 5, n-5 x 5, 6-8 x 8-9, 8-4 x 4. REMARKS AND AFFINITIES. One of the most abundant fossils in the Bovey lignite basin occurring in profusion in the upper beds at Heathfield and in the lower beds at Bovey (C. & E. M. Reid, 1910 : 167). The structure of the endocarp and the characteristic short dorsal germination valve place them beyond doubt in Nyssa. Although now almost invariably flattened they were obviously much inflated in life. The many directions of distortion and compression show that originally they were but slightly compressed dorsi-ventrally below the valve. The valve itself was flat or slightly concave while there was considerable ventricosity in the corresponding part of the ventral face. Some specimens have been folded upon themselves from top to bottom, the folding always being towards the dorsal side. A few have been flattened laterally ; these show the difference in curvature of the two surfaces described. The position of crushing has affected the appearance of the ridges which may be almost obliterated or folded longitudinally so as to appear as sharp ridges. The specific determination of fossil Nyssa presents great difficulties as was recognized by Krausel (1920) and later, in a series of papers by Kirchheimer. Initially Kirchheimer (1934) regarded the abundant Brown Coal species described by Krausel (1920), Gothan & Sapper (1933) and himself as agreeing both in morphology and histology with the Recent Nyssa sylvatica Marsh and indistinguish- able from it. In later reviews of the subject (1938, 1939) he included all the Brown Coal " species " (including Middle Oligocene to Pliocene forms) in a single " form- species ", Nyssa disseminata (Ludwig), on the grounds that the characters did not permit of true specific determination. The length of N. disseminata was given as 0-8-1-9 cm.; breadth as 0-5-1-2 cm. (1938 : 339) ; length, 1-2 cm.; breadth, 0-5-1-2 cm. (19390 : 270). It has not been possible to make a study from actual Brown Coal material and it is abundantly clear that at present species cannot be distinguished on histological grounds. It may, however, be possible to separate some species at least by size and shape of the endocarp plus shape of the germination valve. Thus the Pliocene Reuverian endocarps recorded as N. sylvatica by C. & E. M. Reid (1915) are dis- tinguishable on these grounds from the Bovey species although apparently agreeing closely with the living N. sylvatica. THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN N. sylvatica Number of speci- mens used N. sylvatica (Recent) 40 N. sylvatica var. biflora (Recent) (Pliocene ; (Reuver) 49 Length of endo- 6-9-8 mm. . 7-5-10-5 mm. . 5-5-11 mm. carp Breadth of endo- 4-5-6 mm. . 5-2-7 mm. . 3-5-6 mm. carp Length of valve . 3-3-5 mm. . 3-5-3-6 mm. . 2-25-3-25 mm. (5 specimens (2 specimens (8 specimens) only) only) Breadth of valve . 4-5 mm. . 4-5-5 mm. . 3-4 mm. N. disseminata (Brown Coal) N. boveyana Kirchheimer's . 42 measurements (1938 ; I939«) 8-20 mm. 5-12 mm. 6-8-11-5 mm. 3-1-5-5 mm. 2-5-4-5 nun. (13 specimens) 2-5-3-25 mm. The valve measured in Kirchheimer's figure of a Brown Coal Nyssa (Kirchheimer, 1938, pi. 4, fig. 24) appears to be : length about 4-5 mm.; breadth, 3-5 or 4 mm. It will be noted that the Bovey endocarps are relatively longer, and narrower than those of N. sylvatica while the valves have a relatively narrow triangular form. These characters can be seen in the published figures. It is possible, basing the suggestion on experience of the wide range shown by many other Eocene and Oligocene genera and species, that the Brown Coal species of Nyssa, at least in the older beds, may indeed be a true single species. Whether it should in that case be referred to N. disseminata Ludwig (1857) or to N. rugosa Weber (1852) is not within the scope of this paper. It does, however, raise the question of the relationship of the Bovey Nyssa to the Brown Coal forms, for it might be reasonably supposed that the same wide-ranging species occurred in Germany and Britain. The figures already quoted, together with an examination of the . published illustrations of N. disseminata (or N. rugosa) suggest that the Bovey Nyssa should provisionally be regarded as distinct. The germination valve in the two species is of similar narrow triangular character differing markedly from tljat of N. sylvatica from any source. On the other hand the Bovey Nyssa is appreciably smaller on the whole than the Brown Coal endocarps judging by Kirchheimer's measurements quoted above. Moreover if his figures (1938, pi. 4, figs. 21-24) are typical they show that N. disseminata tends to produce endocarps which are normally broadest above the middle : those from Bovey show a majority which are broadest at the middle. In view of these features a distinct name, Nyssa boveyana, has been given to the British material. But it obviously more closely resembles N. disseminata than N. sylvatica whether Recent or Pliocene. Family MYRTACEAE Section MYRTINAE Genus MYRTOSPERMUM nov. DIAGNOSIS. Seeds referable to the family Myrtaceae and probably to the section Myrtinae, with curved or U-shaped cavities, marginal hilum at the end of a condyle H2 THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN between the limbs of the curved cavity. Chalaza close to the hilum, terminal or sub-terminal on the inner side of one limb, micropyle adjacent to the hilum, terminal on the other limb. TYPE SPECIES. Myrtospermum variabile n. sp. Bournemouth Freshwater Beds (awaiting description). Brit. Mus. (N. H.), No. ¥.34248. Myrtospermum boveyanum n. sp. (PL 16, figs. 160-168) 1910. " compressed winged seed " C. & E. M. Reid, p. 173, pi. 16, figs. 61, 62. DIAGNOSIS. Seed of variable shape, much compressed, glossy, with thin testa. External surface ornamented with very regular concentric rows of cells 0-05-0-1 mm. in diameter (pits or convexities). Diameters of seeds, 1-25-2 mm. HOLOTYPE. Brit. Mus. (N. H.), No. ¥.33950. DESCRIPTION. Seed : Bisymmetric, flat (the degree of compression emphasized by fossilization, sometimes distorted and folded on itself) , sub-oval, transversely-oval, sub-circular, or rounded-triangular, sometimes slightly emarginate at the hilum ; limbs of the U-shaped cavity somewhat unequal in length and breadth, the micro- pylar limb being the longer and narrower. Hilar-scar elongate-oval, marginal between the limbs, micropyle small terminal on one limb, chalaza small sub-terminal on the other. Surface shining, ornamented externally with polygonal or hexagonal pits 0-05-0-1 mm. in diameter, but a few specimens give clear evidence that the pits are actually highly convex areas which have now collapsed ; the pits are aligned parallel with the margin of the seed except near the middle where they diverge from the narrow median area between the limbs. In this median area individual pits are usually very obscure but when visible they are longer and narrower in this part of the testa than over the rest of the surface. In certain specimens the outlines of the surface pits are confused possibly owing to the presence of fine parenchymatous cells which form their walls as in other species of Myrtospermum. Testa only about 0-075 mm. thick in section. The outer part is formed by the coarsely pitted coat, the inner part shows evidence of equiaxial or rectangular cells 0-012-0-016 mm. in diameter ; its structure is often obscure owing to intense compression. Diameter of seeds, 1-25-2 mm. REMARKS AND AFFINITIES. Numerous seeds, all much crushed so that they are reluctant to split in the plane of symmetry. Fortunately one imperfect seed had split naturally and shows the U-shaped cavity and marginal germination. The form of the cavity was confirmed in other specimens by treatment with nitric acid, potassium chlorate and ammonia which rendered them semi-translucent. The curved form and unequal limbs are also clearly indicated by the alignment of the surface sculpture. The species occurs both at Bovey and Heathfield. The relationship of the fossil to Myrtaceae is fully discussed in a forthcoming catalogue on the Bournemouth flora. The species here described as M. boveyanum differs from others from Bovey, Heathfield, the Bournemouth Beds and elsewhere in its extreme degree of compression THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 113 which must be, in part at least, original. It differs also in its thin walls, glossy surface and in the extreme regularity of its surface sculpture. It is also larger than the common Tertiary species awaiting description in a forthcoming catalogue as Myrtospermum variabile. Two specimens were figured and described by C. & E. M. Reid (1910 : 173, pi. 16, figs. 61, 62) but not named. Myrtospermum dubium n. sp. (PI. 16, figs. 169-172) DIAGNOSIS. Seed sub-circular or irregular in outline. Surface pits 0-025-0-05 mm. in diameter near the circumference, smaller near the hilum and between the limbs. Diameter of seed, 1-25-2 mm. HOLOTYPE. Brit. Mus. (N. H.), No. V.33957. DESCRIPTION. Seed : Sub-circular or irregular in outline, somewhat inflated, approximately bisymmetric but with a tendency to be more convex on one side than on the other. Two specimens have a slight elevation over the condylar area between the limbs of the U-shaped seed-cavity ; the exact form of the cavity is not exposed. Marginal hilum large, oval, micropyle small. Surface pitted, pits polygonal or hexagonal, about 0-025-0-05 mm. in diameter near the circumference, smaller and more obscure towards the hilum and between the limbs, tending to be arranged in rows parallel with the margin and to diverge from the area between the limbs. Tegmen thin and translucent, cells not seen. Diameter of four seeds respectively, 1-6 mm., 1-6 by 1-25 mm., 2 mm., 1-75 by 1-25 mm. REMARKS. Five seeds all much compressed and carbonized. One was fractured transversely whereupon it showed the two limbs of the cavity in transverse section; owing to the mode of preservation, the structure of the wall is obscure. The diameter is twice as great as that of typical seeds of M. variabile. Myrtospermum sp. (PI. 16, figs. 173, 174) DESCRIPTION. Seed : Bisymmetric, broadly sub-oval in outline, slightly truncate at the hilar end, inflated, but having a somewhat depressed median area ; U-shaped cavity with a maximum diameter of 0-35 mm.; micropylar limb longer and narrower than the chalazal limb, micropyle terminal at the end of the longer limb, chalaza sub-terminal on the other. Surface deeply pitted, pits equiaxial and hexagonal near the margin where they are 0-07 mm. in diameter, becoming irregular in form and size away from the margin, narrow and elongate over the condylar area between the limbs where they may measure o-i by 0-03 mm. Their walls appear not to be built of small cells. They form thick ridges with a median groove along which there is a marked tendency for splitting to occur. Testa thick, maximum thickness at the end opposite to the hilum (0-12 mm.). The outer part of the testa as seen in section is a single coat of large simple prismatic cells with their longest axes (0-09 mm.) n4 THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN at right angles to the surface. The collapsed outer ends of these cells form the surface pits. Inner part of testa formed of uniform parenchyma, the cells being 0-025 mm. in diameter and radially arranged. Diameter of seed, i-i by 0-9 mm. REMARKS. One seed (¥.33960) , now broken at the hilar end and split marginally (irregularly) so as to expose the internal structure. It differs in its small size from M. boveyanum and M. dubium. From M. boveyanum it also differs in its inflated form and less regular pitting. From M, dubium it also differs in its surface sculpture. The single layer of prismatic cells forming the outer coat of the testa is characteristic. There is no indication, as in M. variabile from Cliff End, Sandbanks, Woolwich and Reading Beds, that it was formed of fine parenchyma. The middle lamella between the prismatic cells is clearly indicated by the narrow grooves along the middle of the ridges between the pits. Family CORNACEAE Section MASTIXIODEAE Genus MASTIXIA Blume Mastixia boveyana n. sp. (PI. 17, figs. 175-178) 1910. Mastixia n. sp. C. & E. M.Reid, p. 166, pi. 16, figs. 73, 74. DIAGNOSIS. Endocarp with smoothly rounded, interrupted, longitudinal external ribs ; median infold broad, rounded ; wall with an external coat of parenchyma. Length about n mm. (estimated). HOLOTYPE. A broken endocarp showing the infold. Brit Mus. (N. H.), No. ¥.33961. DESCRIPTION. Endocarp : One-loculed, ribbed longitudinally, the ribs (about nine at each end, the number in the middle doubtful) smoothly rounded, interrupted, and with additional intercalated short ribs like elongate nodulations towards the middle of the endocarp. A large longitudinal germination valve occupies nearly half the breadth of the nut bearing on its inner surface a broad, rounded, median, longitudinal infold (0-4 mm. across, 0-8 mm. deep in the one specimen in which it could be measured) ; the length of the valve cannot be determined from the imperfect material available. Endocarp wall hard and woody, variable in thickness from about 0-4 mm. as measured through the grooves, to 0-5 mm. through the ribs, the locule being smooth, not ribbed in agreement with the external surface of the endocarp. In section the walls show the following structure : An inner coat of horizontally aligned sclerenchyma many layers thick giving a transversely striate locule-surface ; the cells forming the outer layers of this coat gradually pass from a horizontal to an oblique, and from an oblique to a radial alignment, the radially aligned portion forming the main thickness of the wall. Along the margins of the valve, however, THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 115 the oblique alignment is never lost but passes across the walls to the periphery constituting planes of weakness along which the valve separates ; outside the radial cells are a few peripheral layers of parenchyma, superficially the cells which form the layers are irregularly polygonal with an average diameter of about 0-025 mm., and no definite alignment. These cells give a smooth surface to the endocarp. Length of endocarp unknown, estimated to be about n mm.; diameter of one fragment (probably basal end), 2-7 mm. (possibly increased by distortion). Seed : Not seen ; cells of the testa, 0-03-0-05 mm. in diameter, preserved as impressions on the locule wall superposed on the transverse striations of the locule-lining. REMARKS AND AFFINITIES. Two fragments of Mastixia endocarp were figured and briefly described by C. &. E. M. Reid (1910 : 166, pi. 16, figs. 73, 74) but no attempt was made at specific diagnosis. Six more fragments have now been recog- nized in their collection, two are the ends of nuts showing the complete circumference, both being distorted obliquely. Another fragment is from the middle of a nut with both ends and the valve missing ; a fourth, incomplete at the ends, appears to have broken along the edge of the valve. It suggests a length of about n mm. for the perfect specimen. All fragments are recognizable by their ribbing and micro- scopic structure, they therefore appear to offer a sufficient basis for specific determination. Detailed histological studies of Mastixia and allied fossil genera from the Brown Coal of Germany have been published by Kirchheimer in papers from 1934 to 1939. Many details of the sclerenchyma and parenchyma of the endocarps are shown. The genus Mastixia was first recognized as a fossil by C. & E. M. Reid (1910). It has also been recorded more recently from the London Clay (Reid & Chandler, 1933:448, pi. 25, figs. 1-17). Family PRIMULACEAE Section LYSIMACHINAE * Genus LYSIMACHIA (Tourn.) Lysimachia boveyana n. sp. (PI. 17, figs. 179, 180) DIAGNOSIS. Seeds as in Lysimachia, 0-8-0-85 mm. long, 0-62-0-7 mm. broad. Surface rugosities forming small areoles with nodular boundaries. There are also semi-translucent tubercles especially well seen around the margin. HOLOTYPE. Brit. Mus. (N. H.), No. ¥.33964. DESCRIPTION. Seed : Rounded triangular in outline, originally gently convex on the dorsal face, facetted on the ventral face the two facets meeting to form a longitudinal median angle extending from margin to margin. Hilum elongate about the middle of this angle. Surface rugose, rugosities forming small areoles with nodular boundaries about 0-032 mm. in diameter on the ventral side, and 0-032 mm. or larger in diameter on the dorsal side. On the ventral side they are aligned n6 THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN in rows directed from the median angle to the margin, on the dorsal side no such definite arrangement is apparent. Around the margin semi-translucent tubercles, 0-025 mm- long and 0-022 mm. broad, are visible. Similar tubercles, or their remains can also be seen over parts of the surface but they are less prominent than around the margin. Length of seed, 0-8-0-85 mm.; breadth, 0-62-0-7 mm- REMARKS AND AFFINITIES. Two seeds. The form, character of testa, and median hilum as described, are only to be found combined in the family Primulaceae. Seeds of this shape with a simple hilar ridge extending from edge to edge have been seen only in the Lysimachinae and closely comparable structure in the genus Lysimachia. The majority of species have larger seeds, but in L. japonica some seeds are comparable in size (0-85 by 0-65 mm. and 0-9 by 0-65 mm. for example) although average sized seeds are larger (1-05 by 0-9 mm and 0-9 by 0-75 mm.). The genus Lysimachia is distributed through the temperate and sub-tropical regions of the whole world. Family SYMPLOCACEAE Genus SYMPLOCOS Jacquin Symplocos anglica n. sp. (PL 17, figs. 181-186) 1862. Nyssa europaea Unger : Heer, p. 1066, pi. 69, figs. 11-17. 1910. Nyssa europaea Unger : C. & E. M. Reid, p. 167, pi. 15, fig. 9. DIAGNOSIS. Endocarp ovoid to obovoid or oblong, three-loculed, apical depression with gently sloping edges markedly truncating the apex of the endocarp. Base rounded. Outer surface finely wrinkled longitudinally. Length, 3-5-6-7 mm.; breadth, 2-2-4-4 mm. HOLOTYPE. A somewhat compressed endocarp. Brit. Mus. (N. H.), No. V-33965- DESCRIPTION. Endocarp : Syncarpous, three-loculed, locules arranged around a central canal, one or two occasionally being abortive, opening by wide apertures into the large apical depression the edge of which slopes gently inward being neither thickened nor rounded. At the rounded base is a small attachment scar from which the funicle passes into the central canal. The endocarp may be either obovoid, oblong with rounded base, or ovoid, markedly truncated at the top by the apical depression, but the truncation may be obscured either by the persistent base of the style or by oblique distortion when it appears ovoid. Wall possibly formed of fused mesocarp or endocarp (the former preserved only in part if present), or the mesocarp may be entirely absent ; thickness of wall averaging about 0-25 mm.; its outer surface is finely wrinkled longitudinally, it usually shows about twelve to THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 117 sixteen irregular, sometimes interrupted and nodular rather obscure longitudinal ribs ; surface cells mostly very small with obscure outlines but occasionally (when the mesocarp is abraded ?) the wall is finely and evenly pitted. Length of endocarp, 3-5-6-7 mm.; breadth, 2-2-4-4 mm- Average length, 4-8 mm.; average breadth, 3-3 mm.; average breadth if uncompressed, about 2-1 mm.; breadth of apical depression as compressed, 1-5-2 mm. REMARKS AND AFFINITIES. About thirty-four specimens, all much crushed, some symmetrically others obliquely. The species was described and figured as Nyssa both by Heer and by C. & E. M. Reid. However it is clear that such truncated specimens could not be Nyssa whereas they agree in character and size with the species of Symplocos here described. Other fossil species based on fruits are commonly much larger and differ in other ways. S. gregaria Unger (1866 : 31, pi. n, figs, ig-h) is one of the more comparable species but is more variable in size ranging from 4'i-i2-5 mm. in length, and from 2-1-6-2 mm. in breadth (Kirchheimer, 19360 : 95, pi. 9, figs. 50, b) gives the dimensions as 4-10 mm. length, 2-5-6 mm. breadth. The fruits of 5. gregaria are represented as larger, more markedly truncate, with finer, more regular and numerous grooves and with thicker walls. S. kirstei Kirchheimer (1939 : 285, pi. 3, fig. 3) is also similar but larger (length, 5-5-11 mm.; breadth, 3-5-6 mm.) and relatively longer and narrower. A new name, Symplocos anglica, has therefore been given to the Bovey fruits. Symplocos headonensis Chandler (PL 17, figs. 187, 188) 1910. ? Nyssa obovata Weber : C. & E. M. Reid, p. 168, pi. 15, fig. 10. 1926. Symplocos headonensis Chandler, p. 40, pi. 7, fig. 3 ; text-fig. 24. DESCRIPTION. Endocarp : Syncarpous, four-loculed, the locules arranged around a central canal and opening above by wide apertures into a large apical depression ; broadly ovoid, urceolate, or sub-globular, conspicuously truncated* by the apical depression ; margin of the apical depression thickened, somewhat rounded ; attach- ment indicated by a small sunk scar at the base, thickness of wall, 0-2 mm.; thickness of septum, 0-05 mm. Surface without ribs, but uneven, surface cells angular, unequal in size, with a tendency to be aligned in longitudinal rows near the base thus producing obscure fine striations. Length of an obscurely four-lobed and slightly urceolate specimen (possibly immature), 6 mm.; breadth, 5-5 mm. Length of a second well-developed specimen, 7 mm.; breadth, 5-5 mm. Estimated diameter when uncompressed, 3-5 or 3-6 mm. REMARKS AND AFFINITIES. Three specimens one of which is in the Geological Survey Museum if the specimen figured by C. & E. M. Reid (1910, pi. 15, fig. 10) really is of this character. The illustration indicates a larger, broader form than S. anglica. This species is near in size, form, and general appearance to S. headonensis Chandler from Hordle which see for the probable relationship of this species to living forms. The species occurs both at Bovey and Heathfield. u8 THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN Family SOLANACEAE Genus SOLANISPERMUM nov. DIAGNOSIS. Seeds of unknown generic relationship agreeing with Solanaceae in form and structure. Solanispermum reniformis n. sp. (PI. 17, figs. 189-191) 1910. Carpolithus sp. 5 C. & E. M. Reid, p. 174, pi. 16, fig. 72. A number of seeds of Solanaceae with distinctive rugose surface which characteristic- ally weathers into fibres occur in the Lower Bagshot, Bournemouth Freshwater and Marine Beds, and the Cliff End Beds near Mudeford as well as at Bovey. The most perfect specimen was found at Branksome Dene and an extremely good seed showing the hilar aperture was found at Sandbanks. In order that the description should be as complete as possible it has been largely based on material from the Bournemouth area. DIAGNOSIS. Seed transversely oval or reniform in outline, occasionally hooked, surface normally with coarse, interrupted, sinuous rugosities or tubercles which produce a pitted effect in places. An outer coat, rarely preserved, shows " pits " with sinuous outlines. The rugose coat shows fine striae at right angles to the tubercles. Splitting along the striae on drying produces a fibrous effect. Inner coat spongy formed of equiaxial cells. Maximum diameter of seeds about 3'5-4'8 mm. HOLOTYPE. A perfect seed figured by C. & E. M. Reid (1910, pi. 16, fig. 72). Geol. Surv. Colin. No. 1805. DESCRIPTION. Seed : Bisymmetric, flattened or slightly inflated, transversely oval or reniform in outline, occasionally hooked. Hilum usually marginal occupying part of the concave margin in reniform seeds and one of the longer margins in oval seeds, large and gaping, elongate-oval leading into a small cavity separated from the main seed-cavity by a thin curved partition seen in longitudinal sections of the seed. Probably the funicle lay close to this partition within the hilar cavity. A few seeds are distorted (in growth) so that a gaping hilum is twisted on to one of the broad surfaces. Dorsi- ventral flattening of such distorted seeds may have occurred. Micropyle usually marginal, adjacent to the hilum. Its position is most apparent in hooked seeds where it occupies the extremity of the hook. Surface occasionally showing traces of an outer coat with coarse digitate cells, but in most specimens this is worn away. As normally preserved ornamented with interrupted sinuous rugosities which diverge from the hilum. They produce a pitted effect in places, the pits being about o-05-o-i mm. in diameter. The rugose seed-coat is about o-i mm. thick ; it appears striate, the striae crossing the rugosities more or less at right angles and lying parallel with the margin near the circumference of the seed. Splitting tends to occur along the striae in weathered or dried specimens producing THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 119 a fibrous effect. Close examination of the " fibres " shows them to be formed of fine equiaxial cells 0-012 mm. in diameter. Several layers of such " fibres " occur in this integument. Within it is a spongy coat, 0-4 mm. thick, formed of compact soft parenchyma. The lining of the seed-cavity is striate, the striae diverging from the neighbourhood of the hilum (actually from the closely associated chalaza) . Maximum diameter of seeds, 4-8 mm.; commonly 3-5 mm.; diameter at right angles to it, 2-25-3-6 mm. REMARKS AND AFFINITIES. One seed figured by C. & E. M. Reid (1910) from Heathfield and another (now broken) from the same pit collected by E. M. Reid and Chandler in 1932. The curved outline, flattened seed, large gaping marginal hilar cavity and even the mode of distortion which brings the gaping hilar cavity on to one of the broad surfaces all suggest Solanaceae. Seeds of Capsicum and allied genera show a general resemblance but the succession of coats cannot be matched exactly in any genus examined. Striate fibrous coats, or striate fibrous outgrowths of the testa are found in Lycopersicum and Cyphomandra. Reference to a living genus cannot be made and even the reference to the family is somewhat tentative until living seeds with a closely comparable succession of coats have been found. Meanwhile these readily recognizable specimens are referred to a new genus Solanispermum as it is in Solanaceae that the closest resemblance has so far been traced. INCERTAE SEDIS Carpolithus sp. (PI. 17, figs. 192-194) Fruit : Inferior with remains of three small triangular patent perianth segments at the apex. Elongate having three broad surfaces two of which are ventri-lateral and one dorsal. Surfaces separated by longitudinal angles, the lateral angles sharp and almost flanged, the ventral one fibrous. The broad perianth segments each lie opposite one of the broad surfaces. In profile the fruit is long and narrow. The ventral angle is straight. The dorsi-lateral angles are convex hence the two ventri-lateral surfaces are more or less semi-oval (but unequal). The dorsal face is narrowly oval. Ventral surfaces smooth formed of fine, close, elongate, obliquely aligned fibres, 0-008 mm. broad. Dorsal surface with similar fibres transversely aligned. Fibrous surfaces often concealed by a rough coat of equiaxial cells, 0-025 mm. in diameter. Length, 2-25-3-5 mm.; breadth, 1-1-3 mm- Breadth of ventral faces in one specimen, i and 0-3 mm. respectively, and in a second specimen, 0-5 and 0-52 mm. respectively. REMARKS. Thirteen endocarps from Heathfield. The relationship of these small fruits has not been established. The difference in form of the dorsal and two ventral faces or facets suggests that the fruits grew in close association with one another, the ventral faces possibly in actual contact. GEOL. Ill, 3. IQ 120 THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN Carpolithus sp. (PL 17, figs. 195-197) A three-, four- or possibly five-partite capsule, the segments united still at the base. Segments lanceolate, three only preserved, two being attached, the third represented by a detached distal end. They diverge from one another and have strongly incurved tips. The length of one segment, 4-2 mm.; breadth, 0-9 mm. (but the full length is not shown owing to the strong incurving of the tip) . The second (attached) segment 3-6 mm. with deeply incurved tip by 1-5 mm. The number of segments in the perfect fruit can only be guessed by the angle between existing segments. Segments highly rugose and even nodular on both surfaces ; margins greatly thickened and almost revolute. Hence the segments are concave externally, the broader one having a slight median ridge which makes it biconcave. All show traces of spines near the base, one each side of the median line in the shorter of the two attached segments. Cells of outer surface equiaxial, about 0-016 mm. in diameter. Carpolithus sp. (PI. 17, figs. 198, 199) Endocarp (?) : Syncarpous, sub-ovoid, three-lobed and three-carpelled, or two- lobed and two-carpelled by suppression, the base being sunk between the lobes. Dehiscing loculicidally from the apex (which is always broken irregularly), almost to the base. No central axis seen. Wall close-textured, hard, 0-033 mm. thick (cells indistinct owing to intensely carbonized condition in section) . Surface smooth, of equiaxial cells 0-016 mm. in diameter usually evenly distributed but occasionally aligned in obscure longitudinal rows. Septa very thin, columnar in section. Locule lined by equiaxial cells, 0-012 mm. in diameter. Length of longest specimen, 5-5 mm. (incomplete at apex) ; breadth, 3-25 mm. Length of second specimen, 4 mm.; breadth, 2-1 mm. Seed (lying within the locule near the broken apex) : Linear with a median ridge on the concave side at one end, somewhat tufted at the other end, too decayed to show cell-structure. Length, 2-7 mm.; breadth, 0-6 mm. at the broadest estimate, but actually bent so as to appear only 0-5 mm. broad. REMARKS. Two specimens from Bovey, one from Heathfield. Also three very imperfect specimens from Bovey which may belong to this species. The relationship is undiscovered. Carpolithus sp. (PL 17, figs. 200-202) Several much collapsed and immature fruits are narrow-obovoid or urceolate, three-lobed, three-loculed, the lobes being rather slender and much smaller than those figured in PL 17, figs. 198, 199. Base sunk between the lobes. Apex in best THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 121 preserved specimens with three minute patent persistent perianth segments one corresponding to each lobe. Dehiscence by loculicidal splitting from the apex downwards showing apical canals (or canal) (style ?). Surface dull, finely rugose, cells aligned in obscure longitudinal rows. Locule transversely striate, striae 0-0125-0-017 mm. apart due to small cells aligned in transverse rows. Thickness of walls about 0-025 mm.; thickness of septum (where seen near the apex), 0-016 mm. Length and breadth of four specimens respectively : 4-6 mm. (broken at one end) X 2-4 mm., 4-2 mm. x 2 mm., 4-2 X 2 mm., 4-6 mm. x 2 mm. Seed : Occupying the whole length of the locule, flattened. Longitudinally striate. REMARKS. The form of this tiny fruit and the three minute patent perianth segments recall the rather larger, flattened, immature fruits of Eomastixia bilocularis found in the Bournemouth Beds and the considerably larger immature specimens of another Eomastixia from Lake (awaiting description). There is nothing in the limited evidence available to exclude such a relationship although it should be noted that living Mastixia perianth segments are four- to five-partite. A species of Mastixia occurs in the Bovey lignites but any suggestion as to the relationship must be regarded as tentative only. Carpolithus sp. (PL 17, fig. 203) An elongate ovoid seed with a few longitudinal angles or crumples. No definite organs are visible but the alignment of cells at the pointed end and crumples at the broad end suggest the presence of organs in these positions. Surface formed of inflated longitudinally aligned cells about 0-025 mm. long and o-ooi mm. broad, but over much of the surface the cells have been abraded so that no cell-structure can be seen. Length of seed, 2-6 mm.; breadth, 0-9 mm. Bulbil ? * (PL 17, fig. 204 ; Text- fig. 3) Two sub-spherical bodies, slightly flattened on one side with a large deep depression at the middle of the flat surface, and a second similar depression on one side, may FIG. 3. Bulbil ? Diagrammatic section showing overlapping layers of thick bracts, x 4 approx. GEOL. in, 3. to§ 122 THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN be bulbils or buds. They show superficially an obscure network of shallow furrows. When fractured they reveal overlapping layers of thick bracts. Diameter about 6-6-7 mm. Relationship not known. REFERENCES BANDULSKA, H. 1923. A Preliminary Paper on the Cuticular Structure of certain Dicotyle- donous and Coniferous Leaves from the Middle Eocene Flora of Bournemouth. /. Linn. Soc. Bot., London, 46 : 241-268, pis. 20, 21. 1924. On the Cuticles of some Recent and Fossil Fagaceae. /. Linn. Soc. Bot., London, 46 : 427-441, pis. 39, 40. BRAUN, AL. 1845. Die Tertiar Flora von Oeningen. N. Jb. Min. Geol. Palaont., Stuttgart, 1845 : 164-173. CHANDLER, M. E. J. 1921. Note on the Occurrence of Sequoia in the Headon Beds of Hordwell, Hants. Ann. Bot., London, 139 : 457. 1922. Sequoia couttsiae, Heer, at Hordle, Hants : A study of the characters which serve to distinguish Sequoia from Athrotaxis. Ann. Bot., London, 143 : 385-390, 5 figs. 1923. The Geological History of the Genus Stratiotes : An Account of the Evolutionary Changes which have occurred within the Genus during Tertiary and Quaternary Times. Quart. J. Geol. Soc. Lond., 79 : 117-138, pis. 5, 6. 1925. The Upper Eocene Flora of Hordle, Hants, 1. 32 pp., 4 pis. Mon. Palaeont. Soc., London. 1926. The Upper Eocene Flora of Hordle, Hants, 2. vii + 20 pp., 4 pis. Mon. Palaeont. Soc., London. FLORIN, R. 1919. Eine Ubersicht der fossilen Salvinia — Arten mit besonderer Beriick- sichtigung eines Fundes von Salvinia formosa Heer im Tertiar Japans. Bull. Geol. Inst. Uppsala, 16 : 243-260, pi. n. 1938-45. Die Koniferen des Oberkarbons und des Unteren Perms, 1-8. Palaeonto- graphica, Stuttgart, 85 (B) : 729 pp., 186 pis. GARDNER, J. S. 1879-82. A Monograph of the British Eocene Flora, 1 : Filices. 86 pp., 13 pis. Mon. Palaeont. Soc., London. 1883-86. A Monograph of the British Eocene Flora, 2 : Gymnospermae. 159 pp., 27 pis. Mon. Palaeont. Soc., London. GOTHAN, W. & SAPPER, J. 1933. Neues zur Tertiarflora der Niederlausitz. Inst. Palaobot. Petrogr. Brennsteine, Berlin, 3 : 1-44, pis. 1-7. HEER, O. 1862. On the Fossil Flora of Bovey Tracey. Philos. Trans., London, 152 : 1039- 1086, pis. 55-71. i862a. On certain Fossil Plants from the Hempstead Beds of the Isle of Wight. Quart. J. Geol. Soc. Land., 18 : 369-377, pi. 18. KIRCHHEIMER, F. 1929. Die Gattung Salvinia in den Tertiarfloren der Wetterau und des Vogelsberges. Ber. oberhess. Ges. Nat.-u. Heilk., Giessen (n.f.) 12 : 140-160. 192913. Die fossilen Vertreter der Gattung Salvinia Mich., I. Ein Beitrag zur kenntnis der Mikrosporangien der Salvinia formosa Heer. Planta, Berlin, 9 : 388-406, 8 figs. 1930. Die fossilen Vertreter der Gattung Salvinia Mich., I. Die bisherigen Funde von Sporangienresten und Sporen tertiaren Salvinien. Zbl. Min. Geol. Palaont., Stuttgart, 1930 : 339-349- 1930^. Die fossilen Vertreter der Gattung Salvinia Mich., II. Uber Sporangienrest einer miozanen Salvinie. Planta, Berlin, 11 : 169-206, 19 figs. 1931. Die fossilen Vertreter der Gattung Salvinia Mich., III. Uber einen neuen Func von Resten der Mikrosporangien einer miozanen Salvinie. Planta, Berlin, 13 : 102-1 13, 5 fi£ 1932. Zur morphologie der Salvinia macrophylla Kirch, aus dem miozanen Ton vor Lauterbach (Oberhessen) . Palaont. Z., Berlin, 14 : 309-314, 2 figs. • 1934. Das Hauptbraunkohlenlager der Wetterau. 51 pp., 10 pis. Hanau. THE OLIGOCENE FLORA OF THE BOVEY TRACEY LAKE BASIN 123 • 1935. Bau und botanische Zugehorigkeit von Pflanzenresten aus deutschen Braunkohlen. Bot. Jb., Leipzig, 67 : 37-122, pis. 1-13. 1936. Zur Kenntnis der Friichte rezenter und fossiler Mastixioideen. Bot. Zbl., Dresden, 55 : 275-300, pis. 5-8. 19360. Beitrage zur Kenntnis der tertiarflora Friichte und Samen aus dem deutschen Tertiar. Palaeontographica, Stuttgart, 82 (B) : 73-141, pis. 7-13. 1937- Palaobotanische Beitrage zur Kenntnis des Alters deutscher Braunkohlenschichten, II. Die braunkohlefiihrenden Tone von Siegburg (Rheinland) und Karaenz (Oberlausitz). Braunkohle, Halle, 50 : 893-931, 26 figs. • 19370- Grundzeige einer Pflanzenkunde der deutschen Braunkohlen. 1. 53 pp., 117 figs. Saale. • 1938. Beitrage zur naheren Kenntnis der Mastixioideen-Flora des deutschen Mittel- bis Oberoligozans. Bot. Zbl., Dresden, 58 (B) : 303-375, pis. 3-8. — 1939. Tertiare Dikotyledonenreste und ihr systematischer Wert. Flora, Jena, 133 : 239-296, pis. 1-3. i939a. Uber die botanische Zugehorigkeit weiterer Frucht- und Samenreste, besonders aus den Braunkohlenschichten Sachsens. Planta, Berlin, 29 : 262-278, 3 figs. KRAUSEL, R. 1920. Nachtrage zur Tertiarflora Schlesiens. Jb. preuss. geol. Landesanst., Berlin, 39 : 329-417, pis. 16-27. LUDWIG, R. 1857. Fossile Pflanzen aus der jiingsten Wetterauer Braunkohle. Palaeonto- graphica, Stuttgart, 5 : 81-110, pis. 16-23. 1860. Fossile Pflanzen aus der altesten Abtheilung der Rheinisch- Wetterauer Tertiar- formation, Palaeontographica, Stuttgart, 8 : 39-154, pis. 6-60. PENGELLEY, W. 1863. The Lignites and Clays of Bovey Tracey. Philos. Trans., London, 153 : 1019-1038. REID, C. & E. M. 1910. The Lignite of Bovey Tracey. Philos. Trans., London, 201 (B) : 161-178, pis. 15, 16. 1915. The Pliocene Floras of the Dutch-Prussian Border. Meded. Rijksopsp. Delfst., Amsterdam, 6 : 178 pp., 20 pis. REID, E. M. 1920. Recherches sur quelques graines Pliocenes du Pont-de-Gail (Cantal). Bull. Soc. geol. Fr., Paris (4) 20 : 48-87, pis. 3, 4. REID, E. M. & CHANDLER, M. E. J. 1926. The Bembridge Flora. Catalogue of Cainozoic Plants in the Department of Geology, i. viii -f- 206 pp., 12 pis. Brit. Mus. (Nat. Hist.), London. 1933. The Flora of the London Clay, viii + 561 pp., 33 pis. Brit. Mus. (Nat. Hist.), London. SIMPSON, J. B. 1936. Fossil Pollen in Scottish Tertiary Coals. Proc. Roy. Soc. Edinb., 56 : 90-108, pis. 1-3. UNGER, F. 1860. Sylloge Plantarum Fossilium, I. Denkschr. Akad. Wiss. Wien, 19 : 1-48, pis. 1-2 1. 1866. Sylloge Plantarum Fossilium, III. Denkschr. Akad. Wiss. Wien, 25 : 1-76, pis. 1-24. PLATE ii Osmunda lignitum (Giebel) FIG. i. Two sporangia which have burst and become interlocked. x 28. (¥.33833.) FIG. 2. Diagram to explain Fig. i. Annulus at (a). FIG. 3. The same pair of sporangia, opposite surface, x 28. FIG. 4. Diagram to explain Fig. 3. Annulus at (a). FIG. 5. A single burst sporangium showing the annulus at (a), x 60 approx. (V. 33833.) FIG. 6. A group of spores embedded in remains of sporangium from the specimen in Figs, i, 3. The spores show the fine granulations of the surface. x 400. (V. 33833^) All the above are from Bovey. Salvinia boveyana n. sp. FIG. 7. A sporocarp showing the globular thin- walled sporangia projecting through the wall, x 28. (V.33834.) FIG. 8. Another less mature specimen showing a closely compacted mass of sporangia. Also figured C. & E. M. Reid, 1910, pi. 16, fig. 57. x 12. Geol. Surv. Mus. Colin. No. 76682. FIG. 9. A sporangium showing spores embedded in a froth-like mass, x 150. (¥.33835.) FIG. 10. Another sporangium with stalk, x 150. ^.33835.) FIG. ii. Part of another, x 400. (¥.33835.) All the above are from Bovey. Polamogeton tenuicarpus C. & E. M. Reid FIG. 12. An endocarp, side, with keel on the right beginning to gape. The seed (s), protrudes at the apex. x 15. (V. 33836) Bovey. FIG. 13. Another endocarp showing the gap between the ends of the curved carpel. x 15. (V.33837) Bovey. FIG. 14. Pollen-grain probably belonging to this species. x 1000. J. B. Simpson Colin. Heathfield. Stratiotes websteri (Brongniart) FIG. 15. A seed, ventral side, (c] collar. x 6-5. (¥.33838.) FIG. 1 6. The same, dorsal, looking on to the keel. The specimen is somewhat crushed dorsiventrally. x 6-5. FIG. 17. Part of a seed which has begun to split and burst. It shows the smooth rounded collar and the keel (k) arising out of it. x 6-5. (V. 33839.) FIG. 1 8. Valve of a seed, broken at the micropylar end, inner surface showing the short transverse raphe (r). x 6-5. (¥.33840.) FIG. 19. Another broken valve as in Fig. 1 8. X 6-5. (¥.33841.) All the above are from Bovey. Caricoidea nitens (Heer) FIG. 20. A laterally compressed fruit, truncate at the base. Also figured Heer (1862, pi. 70- fig. 1 8). x 6 approx. FIG. 21. Another fruit. The impression of the calyx is clearly seen in profile at the base. Also figured Heer (1862, pi. 70, fig. 16). x 6 approx. FIG. 22. Base of a dorsiventrally compressed fruit showing the basal calyx scar in the centre of which is the plug closing the passage to the locule. x 6 approx. FIG. 23. Holotype. Figured Heer (1862, pi. 70, fig. 20). A laterally compressed fruit from which a tangential slice had been cut to display the small locule. x 6 approx. (¥.33842.) All the above are from Bovey. Figs. 8, 20-23: Photo C. Reid; Figs. 9-11 : Photo W. N. Croft; Fig. 14: Photo J. B. Simpson. Bull. B.M. (N.H.) Geol. 3, 3 PLATE 11 14 20 21 22 OSMUNDA, SALVINIA, POTAMOGETON, STRATIOTES, CARICOIDEA PLATE 12 Calamus daemonorops (Unger) FIGS. 24, 25. Two young fruits or female flowers showing three styles, x 15. (V. 33846 — 47-) FIG. 26. Somewhat older fruit ; style bases are preserved but the three-fid style has disappeared, x 15. (¥.33848.) FIG. 27. A fruit showing clearly the inner and outer perianth segments. x 15. (V. 33849.) FIG. 28. A better developed fruit also showing the two whorls of persistent perianth seg- ments, x 15. (V. 33850.) FIG. 29. Base of immature female fruit, the striate bracts in two whorls each of three bracts. X 15- (V.3385I-) FIG. 30. Small immature fruit showing overlapping reflexed scales, x 15. (V. 33852.) FIG. 31. A larger, better developed but much compressed fruit showing the overlapping reflexed scales, x 15. (¥.33853.) FIG. 32. The three-partite bract-like perianth of a male flower, x 15. (¥.33854.) FIG. 33, 34. Two more male flowers, x 15. (¥.33855-56.) FIG. 35. Fragment of a fruiting axis, x 6-5. (¥.33857.) FIG. 36. A smaller fragment of an axis. X 6-5. (V. 33858.) FIG. 37. Immature seed extracted from a fruit, x 15-5 (¥.33859.) FIGS. 38-41. Spines and spine bases, x 2-8. (¥.33860-63.) FIG. 42. A pollen-grain (doubled on itself), x 1000. J. B. Simpson Colin. FIG. 43. A crumpled immature seed possibly belonging to this species, x 15-5. (¥.33864.) FIG. 44. The same, opposite side, x 15-5. All the above (except Figs. 37 and 42 from Heathfield) are from Bovey. Myrica boveyana (Heer) FIG. 45. An endocarp showing somewhat rugose surface, x 6-5. (V. 33865.) FIG. 46. A laterally flattened endocarp which may belong to this species. X 6-5. (¥.33866.) FIG. 47. One valve of an endocarp, internal surface, showing coat of equiaxial cells : (/) funicle, (st) stylar canal, x 15. (¥.33867.) FIG. 48. Neotype. Another valve from a broader fruit. Interior. Lettering as above. X 15- (V.33868.) All the above are from Bovey. Fig. 42 : Photo J. B. Simpson. Bull. B.M. (N.H.) Geol. 3, 3 PLATE 12 CALAMUS, MYRICA PLATE 13 Corylus sp. FIG. 49. Fragment of an inflorescence of male flowers with anthers. X 15. (V. 33869.) Heathfield. FIG. 50. The same, opposite side. X 15. FIG. 51. Pollen grain from the above, x 1000. J. B. Simpson Colin. Carpinus boveyanus (Heer) FIG. 52. Neotype. Fruit with accrescent calyx preserved showing longitudinal furrows associated with fibres which arise from the margin of the basal scar (s) of attachment to the wing-like bract, x 6-5. (¥.33870.) FIG. 53. Another longer, narrower fruit ; (s) as above, x 6-5. (¥.33871.) FIG. 54. Another fruit, x 6-5. (V. 33872.) FIG. 55. Another, x 12. (¥.33873.) Heathfield. FIG. 56. Another, x 15. (¥.33874.) FIG. 57. A small fruit with remains of superior perianth at the apex of the accrescent calyx ; one of the two styles is preserved, x 6-5. (¥.33875.) FIG. 58. A specimen with styles preserved. Perianth much worn, x 6-5. (¥.33876.) FIG. 59. A fruit, the lateral scar near the base indicates that it was one of a pair of fruits, the other being but little developed, x 12. (¥.33877.) Heathfield. FIG. 60. A twinned fruit, x 12. (¥.33878.) Heathfield. FIG. 61. A small twinned fruit, x 12. (¥.33879.) Heathfield. FIGS. 62-65. Four twinned fruits. In Fig. 62 part of the basal end of one of the pair is broken away exposing the locule. The upper part is splitting in the plane of symmetry. Fruits in Figs. 64, 65 very unequally developed, x 6-5. (¥. 33880-83.) FIG. 66. Base of dorsiventrally compressed fruit. Carpinus ? x 6-5. (¥.33884.) FIG. 67. One valve (interior) of another dorsiventrally compressed fruit probably Carpinus. X 6-5. (V.33885.) Fagus minima n. sp. FIG. 68. Holotype. An endocarp : (s) basal scar of attachment, x 6-5. (¥.33886.) FIGS. 69, 70. Two endocarps. x 6. The figures are reproduced from C. & E. M. Reid (1910, pi. 16, figs. 67, 68). Geol. Surv. Mus. Colin. No. 76683. Zelkova boveyana n. sp. FIG. 71. Holotype. Endocarp, lateral aspect : (st) style, (a) attachment, x 15. (V.33887.) FIG. 72. The same, marginal view, x 15. FIG. 73. The same, as in Fig. 72 but more tilted to show the attachment (a). X 15. Moroidea boveyana n. sp. FIG. 74. Holotype. Fruit, side. The rounded base is broken, (st) style; (/) projection marking the point of entry of the funicle to the sub-apical placenta, x 15. (¥.33888.) Brasenia ovula (Brongniart) FIG. 75. A somewhat crumpled and distorted seed showing longitudinal corrugations of the surface due to effects of contraction and alignment of the cells : (e) position of embryotega- X 15- (¥.33889.) Heathfield. Magnolia boveyana n. sp. FIG. 76. Holotype. A seed, concave surface with raphe : (ch) chalaza. x 6-5. (¥.33890.) Heathfield. FIG. 77. The same, opposite convex surface. x 6-5. FIG. 78. Another seed, raphe side. X 6-5. (¥.33891.) Heathfield. FIG. 79. The same, opposite side, x 6-5. FIG. 80. A seed longitudinally sectioned showing the canal (ca) through the chalaza-plug. X 6-5. (¥.33892.) Heathfield. Unless otherwise stated all the above are from Bovey. Fig. 51 : Photo J . B. Simpson ; Figs. 55, 59-61, 70 : Photo C. Reid. Bull. B.M. (N.H.) Geol. 3, 3 PLATE 13 76 ~ 77 ^78 ^ 79 80 CORYLUS, CARPINUS, FAGUS, ZELKOVA, MOROIDEA, BRASENIA, MAGNOLIA PLATE 14 LAURACEAE FIG. 81. An empty cupule with simple rim tilted to show the inner surface and scar of attachment of the berry, x 6-5. (¥.33893.) FIG. 82. Another, x 6-5. (¥.33894.) FIG. 83. A wrinkled leathery type of cupule with entire rim. x 6-5. (¥.33895.) FIG. 84. Leathery cupule with distinct sepals at the rim. x 6-5. (¥.33896.) FIG. 85. Another of similar type but much more slender, x 6-5. (¥.33897.) FIG. 86. Small cupule with distinct sepals enclosing berry, x 6-5. (¥.33898.) FIG. 87. Cupule with entire rim, berry enclosed, x 6-5. (¥.33899). FIG. 88. Small cupule with berry. x 6-5. (¥.33900.) Bovey. FIG. 89. A detached berry, imperfect below. X 6-5. (¥.33901.) FIG. 90. Incomplete berry having glandular secretions beneath the skin, x 6-5. (V. 33902.) Bovey. FIG. 91. Another, x 6-5. (¥.33903.) Bovey. Unless otherwise stated all the above are from Heathfield. Capparidispermum boveyanum n. sp. FIG. 92. Holotype. A seed showing curved form and ornamentation : (h) hilum. x 15. (V.33904.) FIG. 93. The same, opposite side, x 15. FIG. 94. Another seed showing the hilar aperture (h) very clearly, x 15. (¥.33905.) FIG. 95. A larger but more compressed seed in which the cotyledonary limb is somewhat more incurled. x 15. (¥.33906.) FIG. 96. One valve of a seed, inner surface, showing the curved cavity. The form of the partition between the limbs is highly characteristic of Capparidaceae. x 15. (V. 33907.) All the above are from Bovey. HAMAMELIDACEAE Genus ? FIG. 97. A seed, side, showing a lobe of the large hilar scar at (/). x 6-5. (¥.33908.) Bovey. FIG. 98. The same seed at right angles to Fig. 97. The hilar scar (h) crosses the proximal end of the seed and has a lobe on each of its broad faces, x 6-5. FIG. 99. A seed of another species showing the long narrow hilar scar (h). At the opposite end of the seed a small portion of the septum between two locules adheres and shows a strand of fibres from the axis of the carpel, x 6-5. (¥.33909.) Heathfield. Rubus microspermus C. & E. M. Reid FIGS. 100-105. Six endocarps showing variations in form and size, x 15. (¥.33910-15.) Bovey. FIGS. 106-109. Four endocarps. x 15. (¥.33916-19.) Heathfield. Bull. B.M. (N.H.) Geol. 3, 3 PLATE 14 106 ' 105 107 108 " 109 LAURACEAE, CAPPARIDISPERMUM, HAMAMELIDACEAE, RUBUS PLATE 15 LEGUMINOSAE Genus ? FIG. no. A crushed seed. X 6-5. (¥.33920.) FIG. in. Another, x 6-5. (¥.33921.) FIG. 112. Part of a third seed showing hilum and the testa in section. x 6-5. (¥.33922.) The above are from Heathfield. Meliosma reticulata (C. & E. M. Reid) FIG. 113. Valve of an endocarp, external surface showing reticulations. X 6-5. (V. 33923.) FIG. 114. The same, interior : (/) funicular canal. X 6-5. FIG. 115. Neotype. A perfect but laterally compressed endocarp. X 6-5. (V. 33924.) FIG. 116. The same, opposite surface, x 6-5. FIG. 117. A dorsiventrally compressed endocarp, looking on to the attachment (centre of figure), x 6-5. (¥.33925.) FIG. 118. The lower part of one valve of an endocarp, interior showing the attachment and short funicular canal (/). X 6-5. (¥.33926.) All the above are from Bovey. Parthenocissus britannica (Heer) FIG. 119. Neotype. A typical seed, dorsal, showing long narrow chalaza. x 6-5. (¥.33927.) Bovey. FIG. 1 20. The same, ventral, showing lateral infolds diverging upwards, x 6-5. FIG. 121. Another seed, dorsal. X 6-5. (¥.33928.) Bovey. FIG. 122. The same, ventral, x 6-5. Parthenocissus boveyana n. sp. FIG. 123. Holotype. Seed, dorsal, x 6-5. (¥.33929.) Bovey. FIG. 124. The same, ventral, x 6-5. FIG. 125. Another seed with outer coat removed, dorsal (base broken) . X 6-5. (¥.33930.) Bovey. Vitis hookeri Heer FIG. 126. Seed (imperfect above on the left), dorsal x 6-5. (¥.33931.) Heathfield. FIG. 127. The same, ventral, x 6-5. Vitis stipitata n. sp. FIG. 128. Seed (imperfect on the left), dorsal, x 6-5. (¥.33932.) Heathfield. FIG. 129. The same, ventral, x 6-5. Till a sp. FIG. 130. A group of anthers, x 15. ^.33933.) FIG. 131. Another group of anthers, x 15. FIG. 132. A pollen grain from anthers in Fig. 131. x 1000. J. B. Simpson Colin. The above are from Heathfield. Microdiptera parva n. gen. et sp. FIG. 133. Seed, dorsal, with pitted oval germination valve and striate wing surface, x 15-5. (V-33934-) FIG. 134. The same, ventral, showing median longitudinal raphe ridge also a furrow on each side of it flanked by a wing, x 15 -5. FIG. 135. Seed with pointed wings, dorsal : (h) hilum, x 15-5 (¥.33935.) Bull. B.M. (N.H.) GeoL 3, 3 PLATE 15 138 143 145 LEGUMINOSAE, MELIOSMA, PARTHENOCISSUS, V1TIS, TILIA. MCRODIPTERA FIG. 136. The same, ventral, x 15-5. FIG. 137. Seed, dorsal. The germination valve has come away and the cavity is full of sand, x 15-5. (¥.33936.) FIG. 138. The same, ventral, showing wings and raphe ridge very clearly, x 15-5. FIG. 139. Seed, dorsal, pitted valve very clear, x 15-5. (V. 33937.) FIG. 140. Same, ventral, x i5'5- FIG. 141. Seed, dorsal. Xi5'5. (¥.33938.) FIG. 142. Same, ventral. Cavities flanking raphe ridge show clearly as they are filled with sand, x I5'5- FIG. 143. Seed with asymmetrically developed wings and clearly denned valve beginning to gape. x 15-5- (V-33939-) FIG. 144. Seed, dorsal. Symmetric wings show striations very clearly, x I5'5- ^.33940.) FIG. 145. Same, ventral, x 15*5. FIG. 146. Small narrow seed, dorsal, x 15-5. (¥.33941.) FIG. 147. Same, ventral, x 15-5. FIG. 148. Small seed, dorsal, x 15-5. (V. 33942.) FIG. 149. Same, ventral, x i5'5- All the above are from Bovey. Fig. 132 : Photo by J. B. Simpson. PLATE 16 Nyssa boveyana n. sp. FIG. 150. Holotype. Endocarp, dorsal, showing the valve (v). x 6-5. (V. 33944.) FIG. 151. Same, ventral, x 6-5. FIG. 152. Another endocarp, dorsal, valve beginning to open, x 6-5. ^.33945.) FIG. 153. The same, ventral, x 6-5. FIG. 154. An endocarp, dorsal, valve detached exposing part of locule. x 6-5. ^.33946.) FIG. 155. The valve removed from the above. x 6-5. FIG. 156. Apical end of an endocarp showing valve in position exceptionally clearly, x 6-5. (V-33947-) FIG. 157. A large laterally compressed endocarp. x 6-5. (V. 33948.) FIG. 158. Detached valve from endocarp of Nyssa sylvatica var. europaea for comparison. X 6-5. (V.33949.) Pliocene ; Reuver. FIG. 159. Detached valve from N. sylvatica var. biflora for comparison, x 6-5. Recent. Myrtospermum boveyanum n. sp. FIG. 160. Holotype. Seed showing curved form and characteristic pitting, x 15. (V.33950-) FIG. 161. Typical seed, usual preservation, showing indications of curved form and condyle between the limbs, x 15. (V. 3 3951.) FIG. 162. Same, opposite side, x 15. FIG. 163. Another seed with clear indications of curved cavity, x 15. (V. 33952.) FIG. 164. Seed showing clear indications that the pits were originally inflated, x 15. (V.33953-) FIG. 165. Somewhat distorted seed, x 15. ^.33954.) FIG. 1 66. Incomplete seed, internal surface of one valve showing condyle : (h) hilum leading into raphe cavity in condyle. Structures are rarely seen in these highly compressed seeds of which the valves are reluctant to separate, x 15. ^.33955.) FIG. 167. Two closely adpressed seeds in position of growth, one ? abortive, x 15. (V.33956-) FIG. 168. The same pair of seeds, opposite side, x 15. Myrtospermum dubium n. sp. FIG. 169. Sub-circular seed, x 15. ^.33957.) FIG. 170. Opposite side of same seed, x 15. FIG. 171. Seed of irregular form, x 15. (V. 33958.) FIG. 172. Another somewhat larger seed, x 15. ^.33959.) Myrtospermum sp. FIG. 173. Seed, imperfect at hilar end. X 15. (V. 33960.) Heathfield. FIG. 174. One valve of the above, internal surface, showing curved cavity, broad triangular canal in condyle and columnar testa cells as seen in section, x 15. Unless otherwise stated all the above are from Bovey. Bull. B.M. (N.H.) Geol. 3, 3 PLATE 16 169 170 ^HP~ 173 NYSSA, MYRTOSPERMUM PLATE 17 Mastixia boveyana n. sp. FIG. 175. Holotype. One end of an endocarp. x 6-5. (V. 33961.) FIG. 176. Opposite aspect of same fragment showing locule in transverse section with its valve and infold, x 6-5. FIG. 177. One end of another endocarp, side. x 6-5. (¥.33962.) FIG. 178. Fragment of endocarp, external surface, at about the middle showing ribbing. X 6-5. (¥.33963.) The above are from Heathfield. Lysimachia boveyana n. sp. FIG. 179. Holotype. Seed, ventral, showing two facets and ridge on which the attachment lies, x 28. (¥.33964.) FIG. 1 80. The same, dorsal. The longitudinal shadow marks a crack in the testa. x 28. Symplocos anglica n. sp. FIG. 181. Holotype. An endocarp somewhat compressed and showing the apical depres- sion, x 6-5. (¥.33965.) FIG. 182. Endocarp, side, slightly distorted so as to show part of the apical depression. X 6-5. (V.33966.) FIG. 183. The same, more tilted, so as to show the three apical apertures. x 6-5. FIG. 184. Another, side, showing truncated apex and wrinkled surface, x 6-5. (¥.33967.) FIG. 185. Small endocarp, distorted so that two of the three apical apertures are shown. X 6-5. (¥.33968.) FIG. 186. A broader endocarp, probably belonging to this species. x 6-5. (¥. 33969.) Symplocos headonensis Chandler FIG. 187. Endocarp somewhat distorted so as to show the apex with four apertures. X 6-5. (V.33970.) FIG. 1 88. Another, side. The apical depression is seen in profile. X 6-5. (V. 3 3971.) All the above are from Bovey. Solanispermum reniformis n. gen et sp. FIG. 189. Perfect seed, x 6. Geol. Surv. Mus. Colin. Na. 76684. FIGS. 190-191. Two fragments of one seed which broke on removal from the matrix. X 15-5- (V-33972-) Carpolithus spp. FIG. 192. A carpel, side, showing one of the ventral facets. ¥entral angle on the right. Note three-fid perianth at apex, x 15. (V. 33973.) Heathfield. FIG. 193. Another, dorsal side, dorsi ventral compression shows part of one of the ventral facets on the right, x 15. (¥.33974.) Heathfield. FIG. 194. The same, ventral side, showing two facets and ventral angle. Persistent superior perianth can be seen, x 15. FIG. 195. Two lobes of a dehisced capsule, inner surface, x 8. (¥.33975.) FIG. 196. The same, outer surface, x 8. FIG. 197. Outer surface of a third detached lobe from the same specimen, x 8. FIG. 198. Endocarp, incomplete at one end, broken so as to show one of the septa, x 6-5. (¥.33976.) Heathfield. FIG. 199. Another, broken at one end. x 6-5. ^.33977.) FIG. 200. A three-lobed slender fruit, x 6-5. (¥.33978.) FIG. 201. Another somewhat broader specimen, x 6-5. (¥.33979.) FIG. 202. Another fruit with three minute patent perianth segments, x 6-5. (¥.33980.) FIG. 203. A seed, x 15-5. (¥.33981.) Bulbil ? FIG. 204. Bud or Bulbil. The large hollow may be the burrow of an insect, x 2-8. (^.33982.) Unless otherwise stated all the above are from Bovey. Fig. 189 : Photo by C. Reid. Bull B.M. (N.H.) Geol. 3, 3 PLATE 17 198 199 200 201 202 MASTIXIA, LYSIMACHIA, SYMPLOCOS, SOLANISPERMUM, CARPOLITHUS, BULBIL? THE STRUCTURE OF SOME LEAVES AND FRUCTIFICATIONS OF THE GLOSSOPTERIS FLORA OF TANGANYIKA D. D. PANT BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 3 No. 4 LONDON: 1958 BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY The following papers appeared in Volume I (1949-52) : Price No. i (1949). The Pterobranch Rhabdopleura in the English Eocene. H. D. Thomas & A. G. Davis . . 75. 6d. No. 2 (1949). A Reconsideration of the Galley Hill Skeleton. K. P. Oakley & M. F. Ashley Montagu . . 55. No. 3 (1950) . The Vertebrate Faunas of the Lower Old Red Sandstone of the Welsh Borders. E. I. White Pteraspis leathensis White, a Dittonian Zone-Fossil. E. I. White ..... . js.6d. No. 4 (1950). A New Tithonian Ammonoid Fauna from Kurdistan, Northern Iraq. L. F. Spath ios. No. 5 (1951). Cretaceous and Eocene Peduncles of the Cirripede Euscalpellum. T. H. Withers . . 55. No. 6 (1951). Some Jurassic and Cretaceous Crabs (Prosoponidae) . T. H. Withers . . 5*- No. 7 (1952). A New Trochiliscus (Charophyta) from the Downtonian of Podolia. W. N. Croft ..... ios. No. 8 (1952). Cretaceous and Tertiary Foraminifera from the Middle East. T. F. Grimsdale .... ios. No. 9 (1952). Australian Arthrodires. E. I. White 155. No. 10 (1952). Cyclopygid Trilobites from Girvan. W. F. WThittard 6s. The following papers appeared in Volume II (i953~56) : No. i (1953). The Deer of the Weybourn Crag and Forest Bed of Norfolk A. Azzaroli . . . • • • ^ 5s- No. 2 (1953). A Coniferous Petrified Forest in Patagonia. Mary G. Calder ....... 125. No. 3 (1953). The Solution of the Piltdown Problem. J. S. Weiner, K. P. Oakley & W. E. Le Gros Clark . . 3-s. 6d. No. 4 (1954). Some Upper Cretaceous and Eocene Fruits from Egypt. Marjorie E. J. Chandler No. 5 (1954). The Carboniferous Flora of Peru. W. J. Jongmans . No. 6 (1955). Further Contributions to the Solution of the Piltdown Probem. J. S. WTeiner, W. E. Le Gros Clark, K. P. Oakley & others ..... & No. 7 (1955). The Schizaeaceae of the South of England in Early Tertiary Times. Marjorie E. J. Chandler i.5s~- No. 8 (1956). The Brachyopid Labyrinthodonts. D. M. S. Watson . £i THE STRUCTURE OF SOME LEAVES AND FRUCTIFICATIONS OF THE GLOSSOPTERIS FLORA OF TANGANYIKA BY DIVYA DARSHAN PANT (Botany Department, The University, Allahabad, India) Pp. 125-175 ; Plates 18-21 ; 21 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 3 No. 4 LONDON: 1958 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 3, No. 4 of the Geological series. PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued January 1958 Price Twenty-one Shillings THE STRUCTURE OF SOME LEAVES AND FRUCTIFICATIONS OF THE GLOSSOPTERIS FLORA OF TANGANYIKA By DIVYA DARSHAN PANT CONTENTS Page I. INTRODUCTION .......... 127 II. DESCRIPTION OF THE MATERIAL ....... 128 Genus Glossopteris Brongniart . . . . . . .128 G. fibrosa n. sp. ........ 130 G. hispida n. sp. ........ 134 G. colpodes n. sp. . . . . . . . 141 Glossopteris sp. A . . . . . . . .149 Genus Rhabdotaenia nov. ........ 149 R. danaeoides (Royle) n. comb. . . . . .151 R. harkini n. sp. ........ 154 Scale leaves .......... 159 Sporangia .......... 160 Specimens from Mhukuru . . . . . . .161 Specimens from Australia . . . . . . .163 Genus Spermatites Miner . . . . . . . .165 S. crystallinus n. sp. ....... 165 S. tetrapterus n. sp. . . . . . . . .169 III. SUMMARY ........... 172 IV. ACKNOWLEDGMENTS . . . . . . . . .173 V. REFERENCES . . . . . . . . . .173 SYNOPSIS The epidermal and cuticular structure of three new species of Glossopteris, a species of Rhab- dotaenia gen. nov. (Taeniopteris in part), some scale leaves, microsporangia and two new species of seeds (Spermatites Miner) from the Mhukuru Coalfield, Tanganyika are described. Taeniopteris danaeoides Royle from India is referred to Rhabdotaenia on the basis of its epidermal structure. Two discs bearing microsporangia, from Newcastle, New South Wales, are described and their sporangia compared with those from Africa. I. INTRODUCTION THE study of the cuticular structure of fossil plant compressions has rapidly advanced but comparatively little work has yet been published on the structure of plants of the Glossopteris flora. The cuticular structure of Glossopteris itself, the most widespread genus of the flora, is but briefly described for only two leaves, G. indica Schimper (Zeiller, 1896 : 369, fig. 13) and G. angustifolia Bgt. (Sahni, 1923 : 277, GEOL. Ill, 4. II 128 THE GLOSSOPTERIS FLORA OF TANGANYIKA pi. 17, figs. 2, 3). These cuticles are, however, so dissimilar that they have always pointed to the need for more extensive work on the subject and the present paper is an attempt in this direction. The material described here was obtained from six borehole cores from the Mhukuru Coalfield in Tanganyika. The cores have been described by Harkin (1953) and in the same paper Professor J. Walton (Appendix IV : 28) gave a pre- liminary report on their flora, a typical Glossopteris assemblage. Professor Walton later handed the cores to me for detailed study. The work was begun in his laboratory in Glasgow and completed at Professor Harris's laboratory in Reading. The cores are only about 5 cm. in diameter, hence the specimens are small, but the matrix is fine and is entirely unoxidized and the plants are beautifully preserved compressions. An unexplained feature of the preservation is that while some leaves are complete, as is ordinarily the case, others show only one epidermis with more or less of the mesophyll adherent to it ; the veins may be present or they may be missing and merely represented by the epidermal cells. It looks as though the leaf had split through the middle before preservation and the two halves were preserved separately. As a result the single epidermis is thin enough to be trans- parent and it shows its cell outlines very clearly, often far more clearly than can be seen in the cuticle. It is easily removed from the rock by celloidin pulls. It also shows the lignine thickenings of the guard cells which are destroyed when the cuticle is prepared. The smaller fossils (scale leaves, microsporangia and seeds) described in this paper owe their interest to the excellent preservation of their cuticles. One of the small Vertebraria axes which show their tracheids and ray tissue has already been described (Pant, 1956). The other fossils occurring in the cores consist of a single fragment of a leaf sheath of Schizoneura, some equisetaceous stems, a few megaspores, abundant two-winged pollen grains, a few three-winged, one-winged and unwinged examples and numerous slender roots showing scalariform metaxylem and spiral and annular protoxylem with or without a surrounding sheath of fibres. These are not described here. Walton (1953 : 30) had, in addition, recorded the occurrence of two specimens of Noeggerathiopsis in these cores (a third is a counterpart) but after examining the specimens and preparing their cuticles I think that they are poorly preserved midrib regions of Glossopteris leaves. According to Harkin (1953 : 9) the plant-bearing shales at Mhukuru belong to the " Upper Coal Measures " and to the " upper part of K3 bed " regarded as corres- ponding to the upper part of the Ecca Series in South Africa (see Harkin, 1953 : 7, Table 2). Walton (1953 : 28) also regards the age as Ecca. His determinations were, however, based on a preliminary identification of the fossils which are here regarded as new species and therefore do not indicate a precise age. Genus GLOSSOPTERIS Brongniart The epidermis of the various species described here is of a single general type and it may well be that all these species belong to one true genus. Glossopteris indica THE GLOSSOPTERIS FLORA OF TANGANYIKA B 129 FIG. i. A-E, Glossopteris fibrosa n. sp. ; F, G. hispida n. sp. A, details of venation of a typical leaf showing fibres in certain meshes. ¥.34444. x 10. B, leaf apex with short meshes. ¥.34444. x 2. c, leaf apex with long meshes ; a few anastomoses may have been missed. ¥.34446. x 2. D, leaf base showing narrow meshes near the margin. The midrib shows lumps but no bundles. ¥.34447. x 2. E, two leaf frag- ments, on the left, middle region, on the right, near base ; some anastomoses may have been missed. Holotype, ¥.344400;. x 2. F, middle region of leaf showing veins, midrib and part of the margin. Holotype, ¥.34450^. x 2. 130 THE GLOSSOPTERIS FLORA OF TANGANYIKA of Zeiller (1896) however, appears from the figure to be different in cuticle and may belong to another genus ; (I was unable to find Zeiller's original slide in Paris). The fructifications described by Zeiller (1902) as Ottokaria, and by Plumstead (1952, 1956) as Scutum and Lanceolatus also suggest generic differences but at present I consider it premature to make new genera out of Glossopteris. Glossopteris fibrosa n. sp. (PL 18, figs. 1-5 ; PL 19, fig. i ; Text-figs, i, A-E ; 2, 3) DIAGNOSIS. Leaf long lanceolate, average length estimated at between 13 and 30 cm.; width in middle region 4-6 cm. Basal part tapering very gradually, margins becoming almost parallel to midrib at base, two sides of lamina often of unequal width. Apex more or less obtuse. Petiole not known. Midrib up to 0-5 cm. wide below, persisting to apex but becoming narrow, probably depressed above and convex below. Midrib sometimes showing small lumps. Margins normally entire or slightly undulate, rarely lobed, slightly curved downwards and thickened. Veins arising from midrib at a very acute angle in all parts of the leaf but soon bending outwards for about i cm. and then continuing at an angle of 6o°-8o° to the midrib except in basal and apical parts where veins less arched making angles of 40°-5o° only with midrib. Veins crossing lamina at a concentration of about 20-30 per cm. near midrib and 32-44 per cm. near margin (measured at right angles to majority of veins). Veins anastomosing in all parts of lamina but more frequently near midrib. Average width of meshes 0-5 mm. (ranging from 0-4 to 0-9 mm.) near midrib and 0-3 mm. near margin. Meshes in middle of lamina of varied length, average length 7 mm.; extreme base of leaf with lamina only one or two short meshes wide. Veins bending forwards at margin. Veins normally 90-160 11 thick (some up to 250//,), prominent on underside. Substance of lamina rather thin, meshes showing elongated fibres running parallel with main veins, about 5-14 /£ wide, fibres occasionally crossing connecting veins or moving to join main veins ; also showing palisade-like cells 15-30 JJL wide and spongy mesophyll cells elongated transversely to the veins averaging 60 /* long and 20 ju, wide. Upper epidermis of lamina usually without stomata. Cells between veins averaging 86 ju, long and 44 fi wide tending to form rows parallel with veins. Lateral (anticlinal) walls about 3 /* thick, arched or nearly straight, never sinuous. Surface wall either without any papilla but appearing finely mottled or occasionally with several small papillae with or without an obscure larger median papilla. Sometimes numerous small papillae tend to be in longitudinal rows or are replaced by longitudinal striations. Cells over veins and often above fibres somewhat narrower and longer. In basal part of leaf and near margin upper epidermal cells become isodiametric but elongated along the margin itself. Cells over midrib with thicker walls (about 6 /i thick), elongated or short, rectangular or polygonal, tending to occur in longitudinal rows, stomata present but rare. Midrib cells often with a moderately conspicuous papilla. Trichomes absent. Upper cuticle of lamina rather thick (up to 3 /* thick). Cell outlines thin, straight, THE GLOSSOPTERIS FLORA OF TANGANYIKA V-\ 131 :•': .:.'k-