PALAEONTOLOGIA AFRICANA ANNALS OF THE BERNARD PRICE INSTITUTE FOR PALAEONTOLOGICAL RESEARCH UNIVERSITY OF THE WITWATERSRAND JOHANNESBURG VOLUME XI 1968 Printed in the Republic of South Africa by THE NATAL WITNESS (PTY) LTD., Pietermaritzburg, Natal. BERNARD PRICE INSTITUTE FOR PALAEONTOLOGICAL RESEARCH BOARD OF CONTROL S. P. JACKSON (Chairman), 1. GL'fN THOMAS, B. I. BALINSKY, R. A. DART, T. W. GEVERS, S. H. HAUGHTON, F. G. HILL, B. D. MALAN, B. W. PRICE, P. V. TOBIAS, H. C. KOCH STAFF Honorary Director and Editor : S. H. HAUGHTON, B.A. (Cantab), D.Sc., Hon. LL.D. (C.T.), Hon. D.Sc. (Rand), Hon. D.Sc. (Nata l), F.R.S., F.G.S. , F.R.S.S. Afr. Assistant to Director and Editor: A. R. I. CRUICKSHANK, B.Sc. (Edin), Ph.D. (Cantab), M. Inst .Biol. Supervising Field Officer: J. W. KITCHING Honorary Research Fellow Palaeo-anthropology: R. A. DART, M.Sc. (Q'1 'd), M.B., Ch.M. (Syd.), Hon. D.Sc. (Rand), F.R.S.S.Afr. Field Officer Pleistocene: B. MAGUIRE, B.Sc. (Rand) Honorary Research Fellow Karroo Vertebrates: A. S. BRINK, M.Sc. (Stell ), Ph.D. (Lond.) Honorary Research Fellow Palaeobotany: EDNA P. PWMSTEAD, D.Sc. (Rand), F.R.S.S.Afr. C.S.I.R. Research Fellow Invertebrate Palaeontology, Micropalaeontology and Microstratigraphy: R. J. DAVEY, B.Sc. (Southampton), Ph.D. (Nottingham) Research Assistants: P. J. RYAN, M.Sc. (Rand), Ecca Stratigraphy. To 29/2/68 T. STRATTEN, M.Sc. (Potch), Dwyka Stratigraphy. To 31 /8/68 J. ANDERSON, B.Sc. (Rand), Palynology M. L. DE GASPARTS, B.Sc. (Naple~), Micropalaeontology. To 31 / 1/68 P. B. BEAUMONT, B.A. (C.T.), Archaeology-Swaziland Project J. S. McKINNEY, M.Sc. (Oklahoma), Coal Stratigraphy. To 31 /5/68 Technical Assistant: D. WOLFAARDT Secretary: Miss J. J. KALLENBACH. To 14-/6/68 Miss E. G. BARRETT Non-European Laboratory Assistants: REGENT HUMA ALLISON MHLONGA 9 Preparators CONTENTS REPORT OF THE HONORARY DIRECTOR FOR THE REMAINDER OF THE YEAR 1967 .. GONYAULACYSTA PARORTHOCERAS, A NEW SPECIES OF DINO­ FLAGELLATE CYST. By R. J. Davey TOOTH STRUCTURE IN RHIZODUS HIBBERTI Ag."A RHIPIDISTIAN FISH. By A. R. I. Cruickshank . . . . . . . . . . . . . . 3 ON THE STRUCTURE OF THE SKIN IN URA NOCENTRODO N ( RHINESUCHUS) SENEKALENSIS VAN HOEPEN. By G. H. Findlay 15 A COMPARISON OF THE PALATES OF PERMIAN AND TRIASSIC DICYNODONTS. By A. R. I. Cruickshank . . 23 ON THE CLASSIFICATION OF THE GORGONOPSIA. By D. Sigogneau. . 33 ON THE SCALAPOSAURID SKULL OF OLIVIERIA PARRINGTONI BRINK. WITH A NOTE ON THE ORIGIN OF HAIR. By G. H. Findlay .. 47 ON THE LYSTROSAURUS ZONE AND ITS FAUNA WITH SPECIAL REFERENCE TO SOME IMMATURE LYSTROSAURIDAE. By J. W. Kitching .. 61 THE CONFUSED STATE OF CLASSIFICATION WITHIN THE FAMILY PROCYNOSUCHIDAE. By J. M. Anderson.. 77 THE CULTURAL IMPLICATIONS OF THE RHINOCEROS TEETH FROM LIMEWORKS, MAKAPANSGAT. By J. M. Anderson 85 THE LITHIC INDUSTRY IN THE MAKAPANSGAT LIMEWORKS BRECCIAS AND OVERLYING SURFACE SOIL. By B. Maguire .. 99 SOME MAIN CONCLUSIONS DRAWN FROM A BASINAL ANALYSIS OF THE DWYKA SERIES IN THE KARROO BASIN . By T. Stratten .. 127 SOME CONCLUSIONS DRAWN FROM A BASINAL ANALYSIS OF THE ECCA SERIES IN THE KARROO BASIN. By P. J. Ryan .. 133 REPORT OF THE HONORARY DIRECTOR FOR THE REMAINDER OF THE YEAR 1967 This report covers the period from April 1st 1967 to the end of that calendar year, in order that its terminal date shall coincide with that of the financial year. It should be stressed at the outset that the term' 'palaeontological research" , which forms part of the Institute's title, has in modern times a wider significance than it had for earlier geologists. While it still entails the collection, preparation, and study of fossils with the object of deciding the genera and species to which they can be attributed, it is becoming more and more de Ti811ellT to accept that they are the remains of once-living animals and plants, and to endeavour to deduce from all the available evidence their modes of life. That is why we in the Institute welcome as co-workers the scientists who are endeavouring to elucidate, by field and laboratory studies, the conditions and the environments in which the fossil-bearing sediments were deposited. STAFF During the period under review the staff was enriched by the coming of Dr A. R. I. Cruickshank, who arrived from Edinburgh on October 19th and assumed duty on that day as Assistant to the Honorary Director and Scientific Officer. Dr Cruickshank's special studies will be concerned with the vertebrates of Karroo age, a continuation in a field with which he has been familiar for some years. As Assistant to the Honorary Director, a good proportion of the adminis­ trative work of the Institute falls to be done by him; and he has speedily familiar­ ized himself with this aspect of his duties. Other welcome additions to the permanent staff are Miss]. Kallenbach , who became typist in April 1967, and Mr D. Wolfaardt who assumed duty as a technical assistant at the end of the year. It is pleasing to report that the longest-serving member of the staff, Mr J. W. Kitching, has been accepted by the University as a candidate for the degree of M.Sc., for which he will submit a thesis entitled "On the distribution of the Karroo vertebrate fauna with special reference to certain genera and the bearing of this distribution on the zoning of the Beaufort beds". The preparation of this thesis is now in hand. Mr Kitching has had to shoulder most of the day-to-day supervisory work during the period under review, a task he has performed with his customary efficiency. Dr E. P. Plumstead and Dr R. J. Davey, both C.S.I.R. Fellows, have continued their valued work at the Institute. The former, as President of the International Subcommission on Gondwana stratigraphy, attended a symposium held in the Argentine and Brazil. She also attended a symposium on Continental Drift at Montevideo, organized under the joint auspices of J.U.G.S. and U.N.E.S.C .O . With the consent of the Government of South Africa she issued, at the Argentine meeting, an invitation to hold a Gondwana symposium in the Republic in 1970, an invitation which was cordially accepted. An organizing committee for this symposium has been established by C.S.I.R. Dr Davey has acted, inter alia, as an adviser to S.O.E.K.O.R.; this Corporation has agreed to the appointment, as from the beginning of 1968, of Mr J. Anderson as a palynologist, working on their behalf, on the staff of the Institute and to subsidize the Institute for this work. Professor Dart's assistants, Messrs. B. Maguire and P. Beaumont, have continued to work at the Institute, as have the two subsidized sedimentary geologists Messrs P. Ryan and T. Stratten. Mr A. Keyser, of the Geological Survey staff, is preparing a thesis on certain anomodont reptiles, and has made considerable use of the facilities which we have placed at his disposal. PUBLICATIONS During the period under review, Volume X of Palaeontologia Africana was issued and distributed. The major part of this volume consists of Dr Plumstead's description of the fossil plants of the Cape system (Palaeozoic), in the preparation of which she examined almost every specimen available in collections in South Africa. The paper is illustrated by 25 photographic plates . Financial contributions to the cost of this volume made by C.S.I.R. and by the Geological Survey of South Africa are gratefully acknowledged. The preparation of Volume Xl is in hand. RESEARCH WORK A. Palaeoanthropolo8Y Until his departure for the United States of America in August to occupy a chair in Philadelphia, the work of Messrs Maguire and Beaumont was supervised by Emeritus Professor R. A. Dart. Mr Maguire had completed, by July, his analysis of more than 16,000 suspected chert artefacts found in the surface soil overlying the australopithecine pink breccias at the Makapansgat Limeworks site and presented his findings on these very primitive worked cherts at the S2 A ~ meeting in Pretoria. During two visits to Makapansgat he collected some 200 artefacts from the Phase 2 breccia. He is continuing with an examination of quartzite and vein-quartz artefacts from soils above the breccias and of the stone artefacts from the solution cavities which penetrate the surface of these breccias. The collection under study consists of approximately 40,000 pieces. Mr Beaumont has continued with the investigations into the pre-European mining industry in Swaziland which he has been undertaking with the support of the Anglo American Corporation. Plans have been drawn by Professor P. V. Tobias and a band of co-workers for a further detailed attack on the problems posed by the famous site at Sterk­ fontein which is under the control of the Univers ity of the Witwatersrand, ii whose Council has charg~ the Board of Control of this Institute with its super­ vision. These plans include new excavations within a defined area-now sur­ rounded by a security fence. Preliminary clearing of rubble has begun, and from this rubble some 380 artefacts, some with adherent breccia, had been recovered by the end of August. From the rubble there have also been extracted bone­ bearing breccia and rodent bones in addition to the artefacts of various cultures (early Acheulean, MSA, and LSA); all have been labelled and stored. The direct responsibility for the field activities rests on Mr A. R. Hughes, Supervisor of Laboratories of the University's Department of Anatomy. The main objects of the renewed studies are to ascertain the full extent of the cave deposit, to throw further light on the stratigraphy of the breccias, to obtain (if possible) absolute radioactive dating of the deposits, and to increase the knowledge of the fauna. B. Karroo Fossil Reptiles Since his arrival at the Institute Dr Cruickshank has begun a study of the diapsid genus Chasmatosaurus and of its relationship to the later form Erythrosuchus. He has prepared certain specimens by removal of matrix; among them is an almost complete foot thought to belong to the second of these genera. The study was, of course, merely in its preparatory stages at the end of 1967. Mr A. Keyser, of the Geological Survey staff, has made considerable use of the Institute's facilities and collections, including specimens from the Rubidge collection, in preparation for a doctoral thesis on the skull characters of certain Anomodont genera. Much of the material that he has handled has been further developed in order to display features in which he is particularly interested for comparative purposes. As an initial step in the preparation of his thesis Mr J. W. Kitching has compiled a record on cards of the exact locality of each of the known and identified reptiles from the Karroo beds in South Africa and has pictorially represented this information on an extensive wall-map formed of sheets of the Topocadastral maps on the scale of 1 :250,000. He has begun the compilation of a detailed chart which will show the zonal distribution of the known genera. There can be no doubt that this study will lead to a revised zonification of the Beaufort series. We have received a typescript copy (in two volumes, one consisting of photographs) of the thesis submitted in Paris by Dr D. Sigogneau. The thesis deals with the classification and anatomy of the Gorgonopsia and is based almost entirely on studies which she undertook in the Institute over a period of nearly two years. The contents of the thesis, which is in French, are very detailed; but the conclusions are of such importance that we hope to be able to publish an extended summary in Palaeonto]oBia Ajricana. C. Micropa]aeontoloBY Dr R. J. Davey has undertaken work on Cretaceous microplankton oc­ curring in borehole core material from Zululand. He has paid particular attention iii to three genera of hystricospheres and their genetic relationships and has shown that by the systematic change of certain variables an evolutionary series has been produced, possibly as a result of the adaptation of the organisms to environment. He considers that the evolutionary stages can be correlated with positions in the Upper Cretaceous sedimentary seguence. In this work the Institute's Zeiss photomicroscope has proved of great value. Dr Davey has spent some time in revising his Ph. D. thesis on non-calcareous microplankton from the Cenomanian of England, France and orth America, which will be published by the British Museum (Natural History) in 1968. He has also supervised the preparation of palynological material from borehole cores submitted by S.O.E.K.O.R. D. Palaeobotany Dr Plumstead has continued her studies of fossil plants from various parts of Africa and has examined material from a new area of Antarctica sent to her by the British Antarctic Survey. Borehole cores from Somkele, Zululand, sent by the Geological Survey, yielded excellent samples of two plants, one of which is new to South Africa, while specimens from the deepened Sambokkraal borehole near Merweville, C.P., sent by S.O.E.K.O.R. , yielded a new species. Two visits to the Orange Free State yielded specimens from a new shaft on St Helena G.M. and from a number of borehole cores. Collections of plant fossils were recorded from the Ulco Limeworks and from the area of the Verwoerd dam. For publication, Dr Plumstead prepared" A Review of Contributions to the Knowledge of Gondwana mega-plant fossils in Africa published since 1950" which was presented at the Gondwana Symposium in Argentina in October 1967, of which she was President, and' 'Fossil Floras and Latitudes" which was presented at the Continental Drift Symposium in Uruguay in the same month. Mr J. Anderson, an M.Sc. student, worked on Triassic fossil plants which he had collected from the Oliviershoek Pass area in Natal. E. SedimentoloBY Throughout the period under review Mr T. Stratten and Mr P. J . Ryan continued work at the Institute. The former is making a detailed study of the Dwyka series of the Karroo basin; the latter almost completed his description of the sedimentology of the Ecca series of the same basin, which he has presented as a Ph.D. thesis to the University of the Witwatersrand. The detailed work of both investigators has an important bearing on the understanding of the ecological conditions under which plant and animal life flourished in the region in Lower Karroo times. TEACHING The institution by the University of an Honours B.Sc. course in Palaeontology necessitates the regularization of formal instruction and the extension, in a limited degree, of the assistance that, for some years past, has been given by members of the Institute. iv As a part-time lecturer attached to the Department of Geology, Dr Plum­ stead has given courses in Palaeobotany and in the Geology and Petrology of Coal to 2nd, 3rd and 4th year students, while Mr Kitching has been called upon to give talks on Karroo vertebrates and Karroo stratigraphy to undergraduates. For the year 1968, Regulations bearing upon the Honours B.Sc. in Palaeon­ tology have been drawn up and submitted to the University for approval. If adopted, they will necessitate close liaison between the Institute's staff and the Professor of Anatomy and his staff. v GONYAULACYSTA PARORTHOCERAS A New Species if DINOFLAGELLA TE CYST by Roger J. Davey INTRODUCTION Da vey (1968 in press) emends the species Gonyaulacysta orthoceras Eisenack and also transfers it to the genus Cribroperidinium Neale and Sarjeant. By emending Eisenack's species the specimens described by Sarjeant (1966) as G. orthoceras, from the Lower Cretaceous of England, must now be excluded from this species. Hence, a new species is here erected to accommodate them. SYSTEMATIC DESCRIPTION Gonyaulacysta parorthoceras sp. nov. 1966 Gonyaulacysta orthoceras (Eisenack); Sarjeant: 121, pI. 14, figs. 5, 6; text-fig. 29 DIAGNOSIS: HOLOTYPE: Shell ovoidal with strong, tapering apical horn accounting for about one-fifth to one-quarter of overall length. Shell wall of moderate thickness, granular to tuberculate. Tabulation 4', 1 a, 6", ?6c, 7"', Ip, 1""; crests low bearing small spinelets. Cingulum strongly spiral, of moderate breadth: sulcus broad and extending to antapex. B.M. (N.H.) slide V.51730, specimen 3. Speeton Clay, Shell West Heslerton Borehole No.1, West Heslerton, Yorks, at 19.25 metres depth. Lower Cretaceous (Upper Barremian). REFERENCES DA VEY, R. J., 1968. Non-Calcareous Microplankton from the Cenomanian of England, Northern France and North America. Bull. Br. Mus. nat. Hist., London. Vol. 17, no. 3 (Part 1) (in press). EISENACK, A., 1958. Microplankton aus dem norddeutschen Apt nebst einigen Bemerkungen tiber fossile Dinoflagellaten. Neues Jb. Geo]. Palaont, Abh., 106, 3: 383-422, pIs. 21-27. NEALE, J. W. & SAR]EANT, W. A. S., 1962. Microplankton from the Speeton Clay of Yorkshire. Geol. Mag ., 99, 5: 439-458, pIs. 19, 20 . SAR]EANT, W. A. S., 1966. Dinoflagellate Cysts with a Gonyaulax-type tabulation, in Studies of Mesozoic and Cainozoic Dinoflagellate Cysts. Bull. Br. Mus. nat. Hist., London. Suppl. 3: 107-156. TOOTH STRUCTURE IN Rhizodus hibberti AG., A RHIPIDISTIAN FISH by A. R. J. Cruickshank Department of Zoology, Edinburgh University, West Mains Road, Edinburgh, 9, Scotland Present address: Bernard Price Institute for Palaeontological Research, University of the Witwatersrand, Jan Smuts A venue, Johannesburg, South Africa ABSTRACT The structure of the large laniary teeth of the Lower Carboniferous rhipidistian fish Rhizodu hibberti Ag. is interpreted in terms of a new theory of tooth development. The structure of these teeth is found to correspond almost exactly to that of the synchronomorial scale as defined by Orvig (195 1) and Stensio (1961 ; 1962). These labyrinthodont teeth are thus shown to be composed of many tooth primordia, and are not a single unit of dentine. Some isolated Rhizodus teeth are described in which the entire labyrinthodont structure is missing, leaving an empty space inside the tooth. From this, a non-mechanical tooth remova l mechanism is postulated. I TRODUCTION New interpretations of vertebrate hard tissues (bone, dentine and enamel) have been made by Orvig (1951) and Stensio ( 1961; 1962). The former reviewed the characteristics of the various types of dentine and introduced into his discussion a new theory on which to base descriptions of teeth and scales. This is termed the" Lepidol11orial Theory". The lepidol11orial theory proposes that each tooth or scale in any vertebrate is composed in general of severallepidol11oria or units of dentine. Lepidol11oria can be combined in two basic ways, to give what are known as e ither adesmic or monodesmic teeth or scales. Stensio (1961 ; 1962) elaborated the lepidomorial theory and introduced the terms' cyclomorial' and 'synchronomorial' to describe the state of affairs if more than one generation of lepidol11oria were involved in tooth or scale formation. In the adesmic or cyclo­ morial state the lepidomoria lay down hard ti ssue before the papillae have time to fuse and in the monodesmic or synchronomorial scale the opposite holds good (figs. 1 & 2). 3 Synchronomorial scales thus contain at an early stage one large compound pulp cavity. It is surrounded by a thin layer of mantle dentine which bulges outward into vertical ridges, marking the position of the individuallepidomorial crowns. Between these ridges, the mantle dentine projects inwards into the pulp cavity as comparatively low intrapulpar crests. These crests represent the vestiges of the mantle dentine in the lepidomorial crowns where the latter fused together. During further development they grow inward and join, thus forming several partition walls which divide the embryonic monodesmic pulp cavity into secondary lepidomorial units. Circumvascular dentine is gradually developed on the partition walls so that the interior of the crown is almost completely filled with hard tissue (fig. 2c). Running through the simple pulp cavities of the lepidomoria of Palaeozoic elasmobranchs was a vascular loop. It entered the scale by means of a neck canal and left through a basal canal (fig. 1). These canals persist in considerable numbers in the composite scales of these fish and must be correlated with the vascular canals of Williamson as seen in ganoid scales (fig. 4). Dentinal tubules in each lepidomorial unit or crown issue from the upper part of each ascending vascular canal of Williamson. Orvig (1951) concludes that all hard tissues in the vertebrates are related and laid down by cells which, although differing in anyone animal, have a common ancestry. Thus bone, dentine and enamel are closely related tissues. Enamel is further differentiated into that belonging to the lower vertebrates and that belonging to the mammals. In the former it is of mesodermal origin and in the latter it is of ectodermal origin (Kvam 1946; Poole 1956a & b; Scott & Symons 1964). As opposed to many earlier authors (e.g. Widdowson 1928) Orvig recognises only three types of dentine in the lower vertebrates. These he calls osteodentine, tubular dentine and orthodentine. Osteodentine is composed of dentinal osteons (modified primary osteons as opposed to Haversian systems proper) and an interstitial bony substance which mayor may not contain bone cells. The bony substance arises first as trabeculae in the pulp tissue and the dentinal osteons are then deposited on the trabeculae without any resorption of the hard tissue. Before the formation of the dentinal osteons the intervascular trabeculae are lined with cells, frequently osteocyte-like, which during the process of deposition of the dentinal osteons modify into true odontoblasts. Tubular dentine, as defined by Moy-Thomas (1939b), is composed of dentinal osteons with an interstitial enamel-like substance. However, Poole (1956a & b) regards this sort of enamel as being similar to dentine and thus tubular dentine may be only a variety of orthodentine. Orthodentine consists of an outer layer of pallial (mantle) dentine and an inner layer of circumvascular (circumpulpar) dentine. The latter consists of large dentiml osteons, either lining vascular cavities in the teeth or the whole pulp cavity. The mantic dentine is pierced by odontoblast processes and may be very thin. Vasodentine (Tomes) is regarded as a variety of orthodentine showing secondary loss of odontoblasts and the incorporation of blood capillaries in the dentine. It is not known which type of dentine is the oldest. Orvig (1951) considers that teeth originally consisted of an outer layer of mantle dentine with an inner mass of osteodentine. In the majority of cases the osteodentine has been lost and the teeth are thus formed from orthodentine. Tarlo (1963; 1964) and Orvig (1965) discuss the nature of "aspidin", an acellular bone-like tissue found in Ordovician ostracoderms. In support of his earlier work, Orvig considers it to be secondarily acellular, whereas Tarlo concludes that the condition is primitive, seeing that it is the oldest vertebrate hard tissue known. Bystrow (1938; 1939; 1942) has described the structure of teeth and dermal bones of osteolepid and dipnoan fish and labyrinthodont amphibia. This present publication is the first to describe the well-known labyrinthine tooth structure of Rhipidistians in terms of the lepidomorial theory. The similarity of the labyrinthodont tooth to the synchronomorial scale is notable (figs. 2 & 3). In addition, tooth replacement phenomena are described here which agree with those illustrated by Bystrow, but which do not agree with the usually accepted patterns. MATERIAL AND METHOD The specimens used in the descriptions are from the unregistered collection of the Royal Scottish Museum, except for the tooth sectioned longitudinally (figs. 5 & 6). This latter specimen is from the Hugh Miller collection and has the number E. (Nat. Hist.) 352. The techniques used in preparing the teeth for study were small variations of standard rock-section cutting. An attempt to make serial transverse sections of one tooth was only partially successful. The apparatus used for this attempt was a screw slot-cutter with a reinforcing washer. However it tended to shatter both the tooth and itself if the specimen was not held absolutely rigid. DESCRIPTION OF MATERIAL The fish from which the teeth came is classified by Romer (1966, p. 361) as follows: order Crossopterygii, sub-order Rhipidistia and family Rhizodontidae. Rhizodus is common in the Lower Carboniferous of both Europe and orth America. In general the fish of this family have slender bodies, paired fins with short obtuse lobes and cycloid scales. In some forms there is a heterocercal tail and the backbone has ring-like centra (Rhizodopsis), while in others the tail is diphycercal and the notochord is unconstricted (Moy-Thomas 1939a). Hugh Miller, quoted by Barkas (1875), estimated that some individuals of Rhizodus must have measured at least 40 feet overall although the only complete specimen recorded was 9 feet long (Stock 1880). 5 Two sorts of tooth occur on the mandibl es of Rhizo dus : a fringe of denticles 1.5-2.0 cms . high on the dentary and a shorter series of very much larger teeth on shelves of bone (coronoids) inside the dentary (fig. 5). Each shelf of bone supporting Jan iary teeth has room for two, only one of which is normally fully erupted (Pander 1860, Tab. 10 & II) . The external appearance of both types of teeth is identical, except for their size. The height of the large laniary teeth on a mandible 25 cm. long diminishes from 6 cm. at the fro nt to just over 2 cm . at the back. There were four functional laniary teeth on this mandibl e, one laniary tooth in the process of being shed, and seven denticl es . In other specimens the number of denticles is very much greater, and it is thought that this low number is aberrant. The lan iary teeth are slender cones which curve backwards and inwards. They are oval in cross section with the long axis roughly parall e l to the jaw and the oval may be drawn out into one or two fine cutting edges. The basal third of the crown is cut by deep grooves and the whole c rown is covered in fine striae. The lan iary teeth of Rhizodus have the same macrostructure as the synchro­ nomorial scale described by Orvig ( 195] ). Distally there are no intrapulpar crests dividing the pulp cavi ty and on the ex terior there is a moderately thi ck layer of enamel. Towards the root intrapulpar crests develop and finally a rather compl icated infolding of the dentine is seen (the' Plicidentine' of Tomes) (f ig. 3c-d). The ename l layer thins off towards the neck of the tooth. Also, in the grooves, it is th inner than normal. Smart (personal communication) suggests that the pack ing together of the lepidomorial units typical of 'Plicidentine' would l1'lake an ext reme ly strong, yet light structure designed to withstand heavy shocks. It is known that the majority of fish and aquat ic arthropods of thc Ca rbonifero us were heavily armoured and a strong piercing tooth would be essenti?1 to thcse predacious Rhipidistians. In longitudi nal sect ion the ascending vascu lar ca nals of Will iamso n are c learly seen. These gradual ly fuse to form the single monodesmic pulp cavity (fig. 6). No neck canals of Williamson are present. The ascendi ng vascu lar canals 0 1 Williamson connect directly w ith the vascular spaces in the jaw and the ce ll spaces become visible as the dentine becomes more bone-like. Eventuall y the cell spaces take the fo rm of true osteocytes, in a simi lar manner to those of bone of attachment. Typically, odontob last spaces are not found in orthodentine. However, in the teeth of this spec imen of R. hibberti they occurred frequently in the dentine lining the pu lp cav ity . They measure about 7fL by 2fL. The dentinal tubes are about lfL in width, and the "growth lines" of the circumpulpar dent ine coincide with the point where they branch. (In fact it is not a branching, but a fusion of the inwardly migrat ing odontob lasts whi ch causes these rings to appear (fi g. 7).) No other ce ll spaces are present in the dentine, but near the exter ior of the tooth a granular layer containing many cell-li ke spaces is seen. This may be sim ilar to the 'G ranu lar layer of Tomes' as found in mammals, but in position is 6 more akin to the regions of globular dentine described by Bystrow in Osteolepis and other labyrin thodonts. It is however not near! y so ext ensi ve as in these other ani mals. The teeth of Rhizodus are thecodont and are held in the jaw with a well developed bone of attachment. Basally and radially this bone of attachment passes into true bone with cell spaces, and is distinct from the dentine at all levels except the lowest. Bone of attachment is seen in the crevices on the exterior of the tooth. As the grooves widen towards the base of the crown, this bone of attachment tends to cover more and more of the outer surface until at the level of the jaw it surrounds the entire crown. Where enamel is present, bone of attachment always overlies it. In the lower basal sections there are some spaces between the trabeculae of bone of attachment rather bigger than normal. These spaces may mark regions where bone has not been formed, or, having been formed, is in the process of being resorbed prior to the tooth being shed (fig. 3d) (see also under Discussion). Cell spaces do not occur in the bone of attachment except near the base of the tooth and where it merges into the bone of the jaw. At first these cell spaces are compact with few processes and measure approximately 30fL by 8fL. Their long axes are parallel to the trabeculae. Towards the jaw their shape gradually changes and more processes are seen. The cells here measure 40fL by 5fL (Roux 1942). Bone of attachment presents rather higher birefringence in polarised light than does dentine and a more even colour than ordinary bone. This is possibly due to the presence of smaller crystallites more constantly orientated and coincides with the idea that bone of attachment was being constantly removed and relaid while the tooth erupted. Resorption of tooth tissue is known to occur in most vertebrates from the outside initiated through mechanical stimulus . Whereas this probably occurred in Rhizodus, there is also evidence that dentine was resorbed through the pulp cavity. Transverse sections of an isolated tooth found in rock, and especially those sections from the base of this tooth, show a marked difference from the cor­ responding sections of teeth still attached to jaws (fig. 9a & b). In the latter all structures associated with the labyrinthodont pattern are seen, but in the former most of the dentine has been removed and all that remains is the outem10st layer. Bone of attachment shows signs of erosion, but enamel seems to be untouched. The complete jaw referred to on page 6 has a pair of laniary teeth with their roots contiguous. If, as is most likely, one of these teeth is replacing the other, then signs of dentine removal are to be expected on the side of the older tooth, nearest the replacing one. No evidence of this can be seen. In figure 3d it will be noticed that alongside the fully functional tooth there lies remnants of another. The matter is discussed below. DISCUSSION .. The interpretation of the labyrinthodont tooth given in this paper, using Orvig's descriptions of the synchronomorial scale as a basis, gives an entirely new 7 orientation to one of the best known features of the Palaeozoic fish and amphibia. A consideration of the generalised synchronomorial scale shows that the equi valent of many generations of tooth material fuse to form a single tooth. This is the state of affairs seen in the labyrinthodont tooth of Rhizodus hibberti of the Lower Carboniferous. Under Tomes' classification the dentine in labyrinthodont teeth was known as 'Plicidentine'. It can now be interpreted more accurately as a variety of orthodentine, with an outer layer of mantle dentine and an inner layer of circum­ vascular dentine (Orvig 1951, pp . 347-359) . Mantle dentine is characterised by the presence of odontoblast processes; it is of even structure and does not have the concentric lines similar to growth rings that circumvascular dentine has (figs. 7 & 8). Therefore it would appear to represent an initial rapid deposition of hard tissue and would be expected to predominate in the tip of the tooth. As far as can be ascertained this is so. A granular layer similar to that in the dentine is seen in the enamel, becoming more marked towards the tip. At the same time the layer in the dentine becomes reduced (fig. 8). The presence of this layer in the enamel can be explained if this enamel is formed by the dentine organ and is subjected to similar influences as the dentine during development. Therefore the enamel is probably mesodermal in origin. In postulating that the dentine of these teeth was removed under control of a non-mechanical system, rather than in the more normal way, the following facts must be examined. Firstly, each shelf of bone supporting the laniary teeth has room for only two, and, of which, it seems normal that only one was functional at anyone time. On the lower jaw of Rhizodus referred to on page 6, a pair of teeth lie side by side. There is no evidence of erosion where they are in contact, which would be expected if mechanical stimulus was the normal method of removing the older tooth. Secondly, transverse sections of adjacent tooth sockets show remnants of dentine with labyrinthodont structure lying round the edge of the empty socket and other sections of isolated (presumably shed) teeth identified as belonging to Rhizodus, reveal that all internal dentine is absent. These observa­ tions suggest that the dentine is removed from the interior before the tooth is .shed. Any mechanism responsible for dentine resorption could also trigger off the eruption of the other tooth of the pair. It could be postulated that the isolated teeth with no labyrinthodont struc­ tures present were in fact newly erupted teeth broken away from the pair as a natural hazard of life. However in figure 9b it will be seen that the remnants of dentine and enamel are in general only slightly thinner than the walls of the tooth in figure 9a. The dissociated nature of the fragments can be explained only if the supporting structures in the centre of the tooth had been eroded from the inside. ewly erupted teeth figured by Bystrow (1938) for Benthosuchus always have the labyrinthodont structures thinly developed, as are the tooth walls themselves. They are much thinner than the tooth wall in figure 9b. Bystrow (1942) discusses the pattern of dermal bone growth in Dipterus 8 with reference to Westoll-lines. In sections of dermal bones he shows resorption spaces occurring within the hard tissues and, in other publications (1938; 1939), he shows similar resorption spaces in the bases of teeth of Benthosuchus, Glyptolepis, Ho10ptychius and Eusthenopteron. Regarding the Westoll-lines, Jarvik (194-8) agrees that they must represent growth stages, an0 Bystrow's analysis of their nature must lead to the conclusion that some cyclical mechanism of dentine resorption from within the tooth must have operated in these vertebrates. In some mammals milk teeth are shed even if the permanent tooth-germ is excised (Oberstzyn 1963). Therefore it seems possible that replacement teeth do not need to erupt from directly below the functional tooth and that the basal dentine can be removed by some process other than mechanici'.l stimulus. ACKNOWLEDGMENTS The descriptions of the laniary teeth of Rhizodus hibberti in this paper were carried out as part of the requirem.ents for the degree of B.Sc. (Hons.) in Zoology in Edinburgh University. I am very grateful to Dr C. D. Waterson of the Royal Scottish Museum for the loan of the material, to Mr G. F. Friend for supervising the work, and the technical staff of both the Geology and Zoology Departments for the time and trouble taken helping to cut the sections and take the photographs. The Cambridgeshire Education Authority supplied a very generous grant during the course of the year. Dr R. Sprinz and Dr Alun Maddy kindly read this paper in manuscript and made many valuable suggestions for its improvement. Mr G. M. Burnett undertook the onerous task of translating the German texts, for which I am very grateful. REFERENCES BARKAS, W. J., 1875. On the microscopical structure of fossil teeth from the Northumberland Coal Measures. Mon. rev. dent. sura., 4, 393- 394-. BYSTROW, A. P., 1938. Zahnstruktur der Labyrinthodonten. Acta zool. Stockh., 19, 357-4-25. BYSTROW, A. P., 1939. Zahnstruktur der Crossopterygier. Acta zool. Stockh., 20, 283-338. BYSTROW, A. P., 194-2. Deckknochen und Zahne der Osteolepis und Dipterlls. Acta zool. Stockh., 23, 263-289. JARVIK, E., 194-8. On the morphology and taxonomy of the Middle Devonian osteolepid fishes of Scotland. K. Svenska Vetensk-Akad. Handl., 25, 1- 301. KVAM, T., 194-6. Comparative study of the ontogenetic and phylogenetic development of dental enamel. Norske Tandelaeaiforen. Tid. (supplement). 56, 1-198. Moy-THOMAS, J. A., 1939a. Palaeozoic fishes. London, Methuen. Moy-THOMAS, J. A., 1939b. Evolution of the elasmobranchs. Biol. Rev., 14, 1-26. 9 OBERSZTYN, A., 1963. Experimental investigations of factors causing resorption of deciduous teeth. J. dent. Res., 42, 660-674. ORVIG, T., 1951. Histologic studies of placoderms and fossil elasmobranchs. I. The endoskeleton. Ark. Zo01., 2, 321-456. ORVIG, T., 1965. Palaeohistological notes. 2. Ark. Zool., 16,551-556. PANDER, C. H., 1860. Uber die Saurodipterinen, DendrodOriten, Glypto1epiden und Cheirolepiden des devonischen Systems. St Petersburg. POOLE, D. F. G., 1956a. Fine structure of scales and teeth of Raia clavata. Q.)I. microsc. Sci., 97, 99-107. POOLE, D. F. G., 1956b. Structure of teeth in mammal-like reptiles. Q. )1. microsc. Sci., 97, 303-312. ROMER, A. S., 1966. Vertebrate palaeontology. 3rd edition. UniverSity of Chicago Press. Roux, G. H., 1942. Minute structures in the scales of Latimeria. S. Afr. J. med. Sci., 7, 1-18. SCOTT, J. H. & SYMONS, N. B., 1964. Introduction to dental anatomy. Edinburgh and London, E. & S. Livingston. STENSIO, E. A., 1961. Permian vertebrates in Geology ?Jthe Arctic. Ed. G. O. Raasch. Toronto University Press. pp. 231-247. STENSIO, E. A., 1962. Origin et natur des ecailles placoids et des dents. Colloq. into Cent. naw. Rech. scient., 104, 75-85. STOCK, T., 1880. Note on the discovery of an entire specimen of Rhizodus sp. in the Wardie Shales. Trans. Edinb. geol. Soc., 4, 38. TARLO, L. B. H., 1963. Aspidin: the precursor of bone. Nature, London, 199, 46-48. TARLO, L. B. H., 1964. The origin ?Jbone in Proceedings ?J the first bone and tooth symposium, Oiford. 1963. Oxford, Pergamon Press. WlDDOWSON, T. W., 1928. Dental anatomy and physiology and dental histology. London, Bale, Sons & Danielson Ltd. 10 Fig. 1. Fig. 2. Fig. 3. Fig. 4. Fig. 5 Fig. 6. Fig. 7. Fig. 8. Fig. 9. EXPLANATION OF TEXT FIGURES The mode of formation of Cyclomorial scales. From brvig, 1951. (a) Vertical section of the primordial lepidomorium, consisting of a dentine crown and a bony base. (b) A more advanced stage of development. The hard tissues of the second lepidomorium have begun to form. (c) Two lepidomoria in the final stages of development. lpca-lepidomorial pulp cavity; cr-crown of primordial lepidomorium; ve-vessels entering the pulp cavity through both the neck and basal canals of Williamson; bp-basal plate of the primordiallepidomorium; can. W. n. - neck canal of Williamson; p. Idu' ­ papilla of soft tissue of the first areal zone of growth; ldup-ldu'-crowns of the primordial and second lepidomorium respectively; bpa-basal plate of the Cyclomorial scale; can. W. b-basal canal of Williamson; bp'- basal plate of the second lepidomorium. The mode of formation of a Synchronomorial scale. From brvig, 1951. (a) Horizontal section of the crown at early stage of development. A thin layer of mantle dentine has formed round the embryonic synchronomorial pulp cavity. (b) More mantle dentine has formed, and the intrapulpar crests have grown out to form partition walls. (c) Formation of dentinal osteons within the secondary lepidomorial pulp cavities formed in stage two. pd-pallial (mantle) dentine; ipcr- intrapulpar crests; v. sp.-secondary lepidomorial pulp cavities; v. can-vascular canals; osd-dentinal osteons; empca-embryonic monodesmic pulp cavity. Transverse sections of laniary tooth of Rhizodus hibberti Ag. Unregistered collection Royal Scottish Museum. See also Fig. 5. (a) 2 cm. from apex. (b) 2.26 cm. from apex. (c) 2.46 cm. from apex. (d) 4.11 cm. from apex . Pc-main pulp cavity; ipcr- intrapulpar crests; ce-cutting edge of tooth; Sp-large spaces between tooth and jaw-bone; Od-dentine of functional tooth; Od' -dentine of alternate tooth; Ba-bone of attachment; Bj-bone of jaw. lines represent 1 cm. Vertical section of ganoid scale to show ascending canals of Williamson (can. W), but no neck canals. From brvig, 1951. Tooth of Rhizodus hibberri Ag. prior to being sectioned transversely . See also Fig. 3. Unregistered collection, Royal Scottish Museum. Bj- bone of jaw; Sh-shale in which tooth and jaw remnant is embedded. Line represents 1 cm . Longitudinal section of incomplete laniary tooth of R. hibberri. Registered number E (nat. hist.) 352. Hugh Miller collection, Royal Scottish Museum. Pc-main pulp cavity; Ba-bone of attachment; Bb- basal bony region of tooth; Bj- bone of jaw; can. W-ascending canal of Williamson; V. can . J-vascular canals of the jaw. Line is 1 cm. long. Orthodentine in laniary tooth of R. hibberti, showing mantle dentine (pd) and circum­ vascular dentine (Cpd). The latter with "growth rings" caused by synchronous fusion of inwardly migrating odontoblasts (Gr). Ba- bone of attachment. Line is 1 mm. long. High power view of figure 3b. to show granular layer in both dentine and enamel. e-enamel; gr. e-granular layer in enamel; gr. d-granular layer in dentine; pd-mantle dentine. Line is 50fL long. Transverse sections of Rhizodus teeth at equivalent levels. (a) Section of tooth immediately above level of jaw-bone. c.f. fig. 3c. (b) Basal section of isolated tooth with dentine missing from interior of tooth. Distortion of tooth caused by pressure of overlying sediment. Od-dentine. 11 Fig. 1 i-.'·,X:·_ ... ::/ b'p FiJ.3 ~,.,,:-':~~~t:.,,:.~:~::;.~-.~.:,; . );' \. con·""n 1. 12 1b 3c Fig.6 13 ON THE STRUCTURE OF THE SKIN IN URANOCENTRODON (RHINESUCHUS) SENEKALENSIS, VAN HOEPEN By G. H. Findlay In the famous collection of fossil remains of the labyrinthodont Urano­ centrodon, housed since 1911 in the Transvaal Museum, the bony skin armour from the ventral surface of the body of at least six individuals has been preserved. In spite of this lavish quantity of material only a few notes on the osseous skin structure were included in van Hoepen' s (1915) description, and all later papers have passed it by almost completely. The two most recent contributions with full bibliographies on the skeleton of Uranocentrodon are those of Romer (194-7) and Watson (1962). Most of the sandstone slabs from the L)'Strosaurus (basal Triassic) zone of Senekal which contain this skin material have been grouped together for many years in a large closed display case. Three slabs apparently unconnected with each other had not been cemented into the display and were separately available at the Museum for detailed study. They proved to belong together, and fitted to yield a strip of armour some 25 inches long down the ventral midline and in places up to 4- inches broad on one or other side of the midi ine. On these portions the present study is mainly based. They had been presented to the museum by the Municipality of Senekal after the main collection had been purchased. The complete skeleton of Uranocentrodon in the ational Museum, Bloem­ fontein, has a length between the limb girdles of some 4-0 inches, clearly covered by ventral armour which extends laterally to a maximum of 5-6 inches from the midline. From these measurements, it may be concluded that the study material available which comprised this 25 inch strip gave a fairly extensive view of the skin structure, pending opportunities to compare it in detail with the rest of the collection. ORIENTATION OF SPEC/ME S The dermal ossicles in Uranocentrodon form an inverted-V or chevron pattern, with the point of the chevron lying in the ventral midline directed like an arrowpoint towards the head end of the animal. In the specimen studied the bilaterally diverging lines of bones meet at an angle of 60° at the anterior end of the specimen falling to 50° over the slenderer parts farther back. The angles of divergence from the midline itself decline therefore from 30° to 25 ° respectively. Mostly it was the internal aspect of the skin armour which showed uppermost on the stone slabs from Senekal. In some cases the skin had obviously slipped out of line with the vertebral column and was clearly indented from above by dis­ placed parts of the axial skeleton. 15 Because the ossicles overlap one another, it is their medial ends which are seen on the internal surface of the armour, and their lateral ends on the external surface. Therefore if the free ends of the ossicles point towards the midline, it is the internal surface which is being examined. By establishing the midline, the direction of divergence and the pointing directions of the free ends of the ossicles, one can orientate the specimen in all planes. Confirmation may be obtained from the position and orientation of the vertebral column where present. When the midline is not visible on the specimen, as is the case in some of the slabs collected at Senekal, the morphology of the individual bony ossicles must be used to determine the facing of the specimen. ANATOMY OF THE TYPICAL DERMAL OSSICLE The dermal ossicles may be considered as belonging to two regional types: those which lie away from the ventral midline, comprising up to 8 clear rows in our material, and those which abut on the midline itself, limited naturally to a single row contributed from either side. The laterally placed ossicle (Fig. 1), seen from the internal or dorsal aspect, has a smoothly pointed flame-shaped medial end (G). Its lateral end (C) after narrowing somewhat, exhibits a thin saucer-like expansion with a slight anterior deflection. These ends are frequently broken off in the specimens. Such an ossicle measures about 35 mm. from end to end. Although both ends taper, the ossicles measure about 5-6 mm. across at the centre where they are broadest. The most conspicuous feature of the ossicles, and indeed the only feature to work from if the ends are missing, is a conical buttress (F) running along the posterior edge of the ossicle. This cone (Fig. 1 F) has its apex near the posterior edge of the spoon or saucer-like lateral end of the ossicle where it is about 1 mm. broad, and it swells evenly in thickness to a breadth of 2-3 mm. just medial to the midpoint of the ossicle. Here the buttress flattens out over the posterior part of the flame-shaped medial end of the ossicle. This buttress converts the dorsal aspect of the ossicle into a grooved or gutter-like channel (E). In the gutter-like part of the ossicle, two structures are clearly housed. One is the next most lateral ossicle of the same row, whose flame-shaped medial end lies about 8 111m. away from the inner tip of the ossicle in which it lies (Fig. 1). The posterior edge of the overlying ossicle impinges on the anterior faCing of the conical buttress, while its anterior edge abuts on or overrides the buttresses in the next head ward row of ossicles (Fig. 1). The second group of structures which lie in the guttered surface of the internal aspect of the scale are the grooves and foramina presumably for nerves and vessels. These lie tucked in the angle formed where the ossicle rises up to form the buttress. Here they would be sheltered from and protected by the overlying ossicle (Fig. 5). The medial or flame-shaped end of the ossicle shows on its inner aspect a number of grooves suggesting muscular, ligamentous or tendinous attachments passing horizontally forward, interspersed with vascular grooves and foramina. 16 'the external aspect of the dermal armour could not be directly studied in plan view, but sections cut through one of the slabs at right angles to the line of th~ laterally placed ossicles showed the following features. The outer surface was s