Bulletin of the British Museum (Natural Histci) *y Geology series Vol 35 1981 British Museum (Natural History) London 1981 Dates of publication of the parts No 1 26 March 1981 No 2 25 June 1981 No 3 29 October 1981 No 4 17 December 1981 ISSN 0007-1471 Printed in Great Britain by Henry Ling Ltd, at the Dorset Press, Dorchester, Dorset Contents Geology Volume 35 Page No 1 Lower Ordovician Brachiopoda from mid and southwest Wales. M. G. Lockley & A. Williams ... 1 No 2 The fossil alga Girvanella Nicholson & Etheridge. H. M. C. Danielli 79 No 3 Centenary Miscellanea ........ 109 Reassessment of the Ordovician brachiopods from the Budleigh Salterton Pebble Bed, Devon. L. R. M. Cocks & M. G. Lockley Ill Felix Oswald's Turkish Algae. G. F. Elliott 125 J. A. Moy-Thomas and his association with the British Museum (Natural History). P. L. Forey & B. G. Gardiner 131 Burials, bodies and beheadings in Romano-British and Anglo-Saxon cemeteries. M. Harman, T. I. Molleson & D. L. Price 145 The Jurassic irregular echinoid Nudeolites clunicularis (Smith). D. N. Lewis & H. G. Owen . . . . ' . . .189 Phanerotinus cristatus (Phillips) and the nature of euomphalacean gastropods. N. J. Morris & R. J. Cleevely 195 Agassiz, Darwin, Huxley, and the fossil record of teleost fishes. C. Patterson 213 The Neanderthal problem and the prospects for direct dating of Neanderthal remains. C. B. Stringer & R. Burleigh 225 Hippoporidra edax (Busk 1859) and a revision of some fossil and living Hippoporidra (Bryozoa). P. D. Taylor & P. L. Cook 243 No 4 The English Upper Jurassic Plesiosauroidea (Reptilia) and a review of the phylogeny and classification of the Plesiosauria. D. S. Brown . 253 Bulletin of the \ > ' British Museum (Natural History Lower Ordovician Brachiopoda from mid and southwest Wales M. G. Lockley & A. Williams Geology series Vol 35 No 1 26 March 1981 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England. World List abbreviation: Bull. Br. Mus. nat. Hist. (Geol.) Trustees of the British Museum (Natural History), 1981 ISSN 0007-1471 Geology Sries Vol 35 No 1 pp 1-78 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 26 March 19gl Lower Ordovician Brachiopoda from mid and southwest Wales M. G. Lockley' & A. Williams K. Department of Geology, University of Glasgow Gl 2 8QQ, Scotland * GENERAL < 2 7 MAR 198 '- LIBRARY J // Contents gy 6 7 8 Class Inarticuuua .*_-ley 8 Superfamily Lingulacea Menke 8 Schmidtites ? micula (M'Coy), emend 8 Lingulella cf. displosa Williams 10 Palaeoglossaattenuata(i.deC.Sowerby) 12 Pseudolingulagranulata(P\\i\\ips) 13 ? Plectoglossa sp 15 Monobolinaplumbea(Sa\tGr) 15 Monobolina crassa sp. nov 17 Paterula cf. bohemica Barrande 18 Paterulafissura [Addison MS] sp. nov 19 Superfamily Acrotretacea Schuchert 20 ? Conotreta sp 20 Torynelasma sp Genus indet 23 Superfamily Discinacea Gray 23 Trematis evansi [Addison MS] sp. nov. 23 Schizocrania cf. salopiensis Williams 24 Schizocrania multistriata (Reed), emend 26 Schizotreta cf. transversa Williams 26 Schizotreta transversa Williams ffairfachensis subsp. nov. ... 26 Class Articulata Huxley 28 Superfamily Orthacea Woodward 28 Hesperorthis dynevorensis Williams, emend 28 Glyptorthis cf. viriosa Williams 30 Glyptorthis viriosa Williams tumida subsp. nov 31 Corineorthis pustula Williams, emend 33 Corineorthis cf. pustula Williams 35 Corineorthis sp 37 'Present address: Department of Geology, University of Colorado at Denver, 1 100- 14th Street, Denver, Colorado 80202, U.S.A. Bull. Br. Mus. nat. Hist. (Geol.)35 (1): 1-78 Issued 26 March 1981 2 M. G. LOCKLEY & A. WILLIAMS Gelidorthis cennenensissp. nov. . . 37 Mcewanella berwynensis MacGregor 39 Skenidioides sp 40 Superfamily Enteletacea Waagen 41 Dalmanella parva Williams 41 Horderleyella convexa Williams, emend 42 Horderleyella sp 45 Tissintia prototypa (Williams) 46 Tissintia immatura (Williams) 48 Tissintia plana (Williams) 48 Tissintia sp 51 Salopia turgida (M 'Coy), emend 51 Superfamily Gonambonitacea Schuchert & Cooper 54 Kullervo sp 54 Superfamily Tripleciacea Schuchert 55 Triplesia edgelliana (Davidson) 55 Oxoplecia cf. nantensis MacGregor 56 Superfamily Plectambonitacea Jones 58 Sowerbyella antiqua Jones 58 Superfamily Strophomenacea King 61 Murinella sp 61 Macrocoelia llandeiloensis (Davidson) 62 Macrocoelia llandeiloensis (Davidson) elongata subsp. nov. ... 65 Christiania elusa sp. nov 66 Superfamily Porambonitacea Davidson 69 Porambonites sp 69 Parastrophinella parva MacGregor 70 Parastrophinella cf. musculosa Williams 70 Superfamily Rhynchonellacea Gray 71 Rostricellula triangularis Williams, emend 71 Acknowledgements 73 References 73 Index 75 Synopsis A study of Welsh Lower Ordovician Brachiopoda, especially from rocks of the Llanvirn and Llandeilo Series in the Llandeilo and Builth Wells areas, reveals the presence of 45 species and subspecies of which eight, belonging to the genera Christiania, Gelidorthis, Glyptorthis, Macrocoelia, Monobolina, Paterula, Schizotreta and Trematis, are new. Representatives of the genera Plectoglossa, Porambonites, Schmidtitesl and Torynelasma, as well as Conotreta, Gelidorthis and Murinella, were hitherto unknown in Wales; the occurrences of Christiania, Corineorthis, Gelidorthis, Kullervo, Mcewanella, Murinella, Oxoplecia, Parastrophinella, Paterula, Plectoglossa, Porambonites, Skenidioides, Tissintia, Torynelasma, Trematis and Triplesia constitute the earliest records of these taxa in the Anglo-Welsh Province. The fauna is reminiscent of the mainly endemic Anglo-Welsh assemblages from the Shelve area, although the indigenous taxa are supplemented by a number of Baltic stocks like Christiania, Kullervo and Porambonites, and a few of Bohemian affinity, notably Gelidorthis and Paterula. Correlation between the Shelve and other Anglo-Welsh successions is practicable and demonstrates the widespread distribution of distinctive fossil assemblages dominated by inarticulates in pre-Caradoc argillaceous facies. More precise local correlations based on conspecific forms can be effected between both the argillaceous and arenaceous facies of the Llandeilo and Builth Wells areas. LOWER ORDOVICIAN BRACHIOPODA 3 Introduction Since the identification of the 'Llandeilo Flags' as the fourth formation of the Silurian System (Murchison 1839: 222), the Ordovician successions of the Llandeilo area have been a source of much controversy. The issues involved are well known and centre on the merits of the Llandeilo Series as an internationally acceptable time-stratigraphical unit. To some extent this debate was generated by the failure of Murchison (1839: 355-357) and others to realise that fossiliferous sandstones associated with the 'Llandeilo Flags' were not Caradoc but Llanvirn or Llandovery in age. Yet even when the stratigraphical succession had been satis- factorily determined (Strahan et al. 1907: 12; Williams 1953: 1 79), the controversy remained alive. In retrospect this was the consequence of not recognizing that much of the Llandeilo Series is coeval with the lower part of the Nemagraptus gracilis Zone (Williams et al 1972: 5). This particular shortcoming reflected the inherent difficulties of correlating what was until recently believed to be an exclusively shelly Llandeilo facies with contemporaneous graptolitic shales found in the other parts of Carmarthenshire (Dyfed) and in Shropshire. It also arose from the impoverished nature of the Llandeilo shelly assemblages, the brachiopod constituents of which especially seemed to be mainly pandemic species with wide strati- graphical ranges. Between 1866-1883, for example, Davidson (see Cocks 1978) listed several species of brachiopod from the Llandeilo Flags of Wales, some of which were founded on type specimens from post-Ordovician rocks. In 1949, one of us (A.W.) described eleven species of Ordovician brachiopods from the Llandeilo-Llangadog area. The study and a later stratigraphical account in 1953 showed that the brachiopod faunas of the Llandeilo Flags and especially those of the varied sediments constituting the underlying Ffairfach Group are more diverse than had been generally acknowledged. The species were poorly illustrated and inadequately defined, but no revision was contemplated until two features of the distribution of Ordovician marine faunas became evident. The first was the endemic distribution of many of the brachiopod species characteristic of the older Ordovician rocks of Wales and Shropshire. The degree to which these Anglo- Welsh species were distinguishable from contemporaneous taxa in the rocks of Scotland, North America and the Baltic had long been known. But as the Ordovician faunas of Europe and north Africa became more familiar, especially through the admirable researches of Havli5ek (1967, 1971, 1977), the endemicity of those from south Britain became more obvious. Indeed factor analyses of such faunas suggested that the Anglo- Welsh assemblages are diagnostic of a marine province which retained its individuality for the greater part of the Period (Williams 1969, 1973). In this context, the unexpected appearances of taxa which are more characteris- tic of the Baltic, Bohemian or north African successions are records of interprovincial migrations, and their occurrences are important clues to the distribution of tectonic plates during Ordovician times. A renewed interest in the taxonomy of the older Ordovician faunas of the Builth- Llandeilo area was also prompted by a study of the brachiopod assemblages of the Shelve district in west Shropshire (Williams 1974, 1976: 38-44). These researches showed that the assemblages were relics of three associations of low to moderate diversity, each characteristic of a distinctive type of substrate as indicated by the entombing sediments. The brachiopods recovered from the Shelve successions were obviously closely related to those found in contemporaneous shelly facies in Wales. This circumstance afforded an opportunity to explore the geographic and stratigraphical distributions of the associations and the structure and evolution of their constituent communities in the manner outlined by Lockley (1978). In particular, the varied sediments and pyroclastics associated with the volcanic complexes exposed in the Towy Anticline between Llandrindod Wells and Llandeilo (Fig. 1) were known to contain a relatively diverse brachiopod fauna which may have been the Llanvirn antecedents of some of the associations found in the richly fossiliferous Caradoc Series. The community relationship, if any, between these high-diversity associations and the intercalated restricted faunas of the Llandeilo Series was equally intriguing. M. G. LOCKLEY & A. WILLIAMS N Abereiddy Ba ^WLIandrindod Wells ''Z&yf&r Llangadog LLANDEILO Fig. 1 Map of south Wales and the Welsh Borders, showing key localities and outcrop of Ordovician rocks (stippled). In an attempt to unravel these relationships, collections from the Llanvirn and Llandeilo rocks of Powys and Dyfed were made by Dr J. M. Hurst and Dr C. J. Wilcox as well as the authors. Systematic sampling was limited to comparatively few, albeit the most repre- sentative and best exposed, stratigraphical sections. About 500 kg of rock were collected from the type section of the Ffairfach Group and these have yielded over 7500 brachiopods and numerous representatives of several other phyla. Another 130 kg of coeval sediments and ashes from the Coed Duon and Longwood sections, respectively about 4 km SE and SW of Llangadog, provided nearly 3000 brachiopod specimens. In contrast over 1 300 kg of Llandeilo rocks, collected by Dr Wilcox especially from Dynevor Park, Pont-bren-Araeth Dingle and the type section along the Cennen, yielded fewer than 6000 brachiopods; these he kindly placed at our disposal for this study. From the viewpoint of clarifying the relationship between the fossil faunas of the Shelve and Llandeilo areas, those occurring in the Ordovician sediments and ashes of the Builth- Llandrindod Inlier are crucial. We were fortunate to have access to brachiopods from collections made by Mr P. R. Sheldon, obtained mainly from the Carneddau Hills north of Builth Wells and sections along the Howey Brook and near Bwlch-y-cefn Bane, both within 5 km east of Llandrindod Wells, which were invaluable supplements to our own samples. In all, the impressions of over 5000 brachiopods retrieved from the Inlier were available for systematic appraisal. The classic areas of Builth Wells and Llandeilo were, of course, well known to those indefatigable fossil collectors who provided eminent palaeontologists of the last century, like Davidson, M'Coy and J. de C. Sowerby, with so much material for study. In the course of our own researches we have examined as many relevant specimens featured in publications as we have been able to unearth. In this task, we have been aided by Dr L. R. M. Cocks' Review of British Lower Palaeozoic Brachiopods (1978). Such specimens are housed in the British LOWER ORDOVICIAN BRACHIOPODA Museum (Natural History), the Geological Survey Museum, the National Museum of Wales and the Sedgwick Museum; their whereabouts are identified by the prefixes B or BB, GSM, NMW and SM respectively. New numbers assigned to specimens during this study fall within the sequences BB 92265-499, BB 94036-77 and BB 94216-48, and SMA 104410-9, SMA 104446-65 and SMA 105827-36. Table 1 Stratigraphical distribution of brachiopod species according to 'subseries' occurrence. LI and L2 represent the Didymograptus bifidus Shales and Ffairfach Group of the Llandeilo area. L3, L4 and L5 represent the Lower, Middle and Upper Llandeilo of the Llandeilo area, respectively. Bl and B2 represent the D. bifidus and D. murchisoni Beds, and B3 and B4 the Glyptograptus teretiusculus and Nemagraptus gracilis Beds of the Builth area. Brackets denote uncertainty about the horizon at which a taxon occurs. Six of the 45 described taxa, which are not listed here, originate from isolated localities referred to in the text. Species LI L2 L3 L4 L5 Bl B2 B3 B4 Christiania elusa sp. nov. Corineorthis pustula Williams Corineorthis cf. pustula Williams Corineorthis sp. DalmanellaparvaWtiViams * * * * Gelidorthis cennenensis sp. nov. Glyptorthis cf. viriosa Williams Glyptorthis viriosa Williams tumida subsp. nov. Hesperorthis dynevorensis Williams Horderleyella convexa Williams Horderleyella sp. Kullervo sp. Lingulella cf. displosa Williams Macrocoelia llandeiloensis (Davidson) Macrocoelia llandeiloensis (Davidson) elongata subsp. nov. Mcewanella berwynensis MacGregor Monobolina crassa sp. nov. Murinella sp. Oxoplecia cf. nantensis MacGregor Palaeoglossa attenuata (Sowerby) Parastrophinella cf. musculosa Williams (*) Parastrophinella parva MacGregor Paterula cf. bohemica Barrande Plectoglossa sp. (*) Porambonites sp. Pseudolingula granulata (Phillips) Rostricellula triangularis Williams Salopia turgida (M'Coy) Schizocrania cf. salopiensis Williams Schizotreta cf. transversa Williams Schizotreta transversa Williams ffairfachensis subsp. nov. Schmidtites ? micula (M'Coy) ( ) Skenidioides sp. Sowerbyella antiqua Jones Tissintia immalura (Williams) Tissintia plana (Williams) ^ # Tissintia prototypa (Williams) # Tissintia sp. Triplesia edgelliana (Davidson) 6 M. G. LOCKLEY & A. WILLIAMS Faunal distribution In the wake of current palaeoecological studies, the traditional faunal list showing the stratigraphical range and frequency of occurrence of fossil species may be misleading. A lithostratigraphical unit as comprehensive as a Group or a Formation may contain a number of benthic associations which are more or less exclusive of one another. In the varied sediments and ashes composing the Ffairfach Group, for example, there are as many as eight distinctive faunal assemblages; and, although some of them may have been ecotones, the majority must have been fully independent associations. It may, therefore, seem pointless to present an introductory list of species recorded from, say, the Ffairfach Group. Yet there are two benefits from doing so which prompt us to continue the tradition. The first is that if there are two or more benthic associations in a stratigraphical unit, they are likely to reflect an orderly sequence of events affecting the palaeoenvironment and will, through such facies relationships, impart a diagnostic pattern on the faunal list for the entire unit. Such sequences of associations are well seen in the Ffairfach Group and will be discussed elsewhere. Meanwhile it is noteworthy that part of the sequence identified at Ffairfach is found in the Upper Llanvirn sediments and ashes of the Builth area which, of course, accounts for the similarities in the two faunal lists. Secondly the changing nature of fossil associations also contributes to their general usefulness when they are used to compile faunal lists. Throughout geological time, associa- tions have evolved by replacement of their constituent taxa, by speciation and by assimila- tion of immigrant stocks. Since such changes commonly occurred, a particular climax association is normally diagnostic of a short segment of geological time. The exceptions appear to be some low diversity associations like those characteristic of the Middle and Upper Llandeilo successions. Even these, however, are informative if only for palaeoecological purposes. As the list in Table 1 shows, 45 brachiopod taxa have been identified in the Llanvirn-Llandeilo rocks of mid-Wales, with 16 species belonging to the Inarticulata, 18 to the Orthida, 2 to the Triplesiidina, 5 to the Strophomenida, 3 to the Pentamerida and 1 to the Rhynchonellida. Their stratigraphical distribution is also given. In both the Llandeilo and Builth Wells areas, the pattern of distribution is essentially the same. The Lower Llanvirn (Didymograptus bifidus) successions are dominated by the opportunistic articulate species Tissintia prototypa. Thereafter diversity dwindles from a maximum in the variable sediments and ashes of Upper Llanvirn age where articulate species may outnumber inarticulates by 6 or 7 to 1 . Reduction in diversity resulted from an elimination of articulate species which although dominant in the Lower and Middle Llandeilo successions of the type area were no more numerous than the inarticulates in the Upper Llandeilo, while in the Builth area they had almost entirely disappeared by the end of Lower Llandeilo times (Nemagraptus gracilis Zone). As for the endemicity of the faunas, the study has confirmed the dominance of the Anglo- Welsh stocks, but sporadic immigrant genera from the Baltic province are well in evidence in late Llanvirn times when volcanic activity briefly provided shallow belts of clean-washed pyroclastics and derived sediments for colonization by Christiania, Hesperorthis, Kullervo and Triplesia. Stratigraphical terminology The stratigraphical nomenclature used in the systematic section is based almost entirely on the stratigraphy proposed by Williams (1953) for the Llandeilo region and by Elles (1939) and Jones & Pugh (1941, 1948, 1949) for the Builth district. In both areas the existing nomenclature may be lithostratigraphical, biostratigraphical and chronostratigraphical and in the Builth district is complex and in some respects repetitive. At Llandeilo the D. bifidus shales are overlain by the Upper Llanvirn Ffairfach Group which was divided by Williams (1953: 180) into five lithostratigraphical units, referred to LOWER ORDOVICIAN BRACHIOPODA 7 here as formations. In the overlying Llandeilo Series, which consists of a biostratigraphically based chronostratigraphical sequence of stages (Lower, Middle and Upper), the succession can be further subdivided into distinctive lithostratigraphical and biostratigraphical units (Williams 1948) which have been used in the descriptions of the range of some species. In the Builth-Llandrindod area, the 'stratigraphical and palaeontological succession' of zones proposed by Elles (1939: 389), largely on the basis of graptolite biostratigraphy, included only one lithologically-based unit, the Main Volcanic Series. This was sub- sequently subdivided by Jones & Pugh (1941 , 1949), who recognized four volcanic series and numerous laterally variable or impersistent lithostratigraphical units in the succession as a whole. The terminology proposed by these authors remained essentially unchanged except for minor modifications introduced by Hughes (1969) until, with the publication of the Geological Society Special Report no. 3 on the Ordovician (Williams et al. 1972) and the 1976 I.G.S. map of the 'Llandrindod Wells Ordovician Inlier', the stratigraphical terminology of this area was modified to conform to modern lithostratigraphical practice (e.g. Elles' D. murchisoni 'zone' is now the Upper and Lower D. murchisoni Shales). Some confusion arises in the classification adopted for the I.G.S. map, which has modified the lithostratigraphical terminology established by Jones & Pugh in such a way that the lower part of the Builth Volcanic Series becomes the 'Main tuff group' while the Grey Felspar Sands and Pyritous Felspar Sands become amalgamated equally informally into the 'Coarse felspathic sandstones'. Although the nomenclatorial modifications incorporated in the I.G.S. map have been accepted for identifying the successions from which the described taxa have been recovered, we prefer a formalized classification (i.e. Main Tuff Group) and we adhere to the chronostratigraphy proposed by Williams et al. (1972). Systematic methods The procedure adopted in the taxonomic study of the fossil collections at our disposal has been governed by the need to define all taxa as precisely as possible in preparation for palaeoecological researches on the older Ordovician brachiopod faunas throughout Wales and the Welsh borderland. The work has necessarily involved rectifying a number of nomenclatural errors, some of which have been perpetrated by an author of this monograph. The major commitment, however, has been to conduct taxonomic surveys of large samples collected at closely-spaced intervals. Thus in the type section of the Ffairfach Group, which consists of about 100 m of nearly continuous exposures including the mainly unfossiliferous basal Grit Formation (>20m), 77 samples, each yielding an average of almost 100 identifiable brachiopod remains, were taken from these outcrops at a mean stratigraphic interval of 85 cm. The commonest fossil was Dalmanella parva, which constituted more than half of all the brachiopods recorded in 36 out of the 64 samples in which it occurs. It would have been too daunting a task to subject every collection of Dalmanella parva to an exhaustive series of statistical tests. Instead spot checks were carried out to confirm the morphological homogeneity of the stock throughout the entire range of its occurrence, and having found this to be so, one of the better-preserved samples was chosen to represent the Ffairfach populations in statistical comparisons with Dalmanella from other localities. When, as was true of Sowerbyella, a genus appeared to be represented by morphologically distinguishable populations within a section, an appropriate number of samples were comprehensively compared so as to define the variation. The statistical procedures adopted here conform to those outlined by one of us (Williams 1962: 69-79) and permit direct comparisons with data presented subsequently (e.g. Williams 1974) which were also derived from analyses conducted in a similar or identical manner. Economy precludes publication of the 103 statistical tables on which our systematic studies have been based. The tables, laid out according to standard format and arranged in the taxonomic sequence of this paper, have been deposited in the Palaeontology Library of the British Museum (Natural History) and are available for consultation. In addition, the means, 8 M. G. LOCKLEY & A. WILLIAMS variances and numbers of measurements taken for all quantified characters are incorporated in the following systematic descriptions. Where a bivariate estimate of a feature had been calculated, a coefficient of correlation (r) is also given. These parameters should prove sufficient to enable readers to carry out the standard univariate and bivariate statistical tests. In particular, they can be used for the calculations of the rate of growth (a), the index of residual shape (b) and even the natural logarithms of means (variances) when allometric effects of shell growth can be demonstrated (Kermack & Haldane 1950). The abbreviations used in the text for presentation of these statistics identify certain vectors of measurement expressed in millimetres. They are: 1, length;^, mean maximum length; w, width; w, mean maximum width; th, mean maximum depth; Isc, mean maximum length of a muscle scar;Ts, mean maximumjength of the median (or compound) septum; dl, mean maximum length of dental lamellae; Ic, mean maximum length of the dorsal cardinalia (brachiophore bases unless otherwise stated); and Isr, the mean maximum length of the dorsal socket ridges. The variances (var) are always given with these statistics, as are the mean maximum length and variance of the complete valve, i.e. 1 mm (var 1), and the coefficient of correlation (r) when the relative growth of a valve or one of its features is being described. Taxonomic descriptions Under 'material' all measurements given are in mm; p.v. = pedicle valve, b.v. = brachial valve. Lectotypes were mainly selected by Cocks (1978). Class INARTICULATA Huxley, 1 869 Order LINGULIDA Waagen, 1 885 Superfamily LINGULACEA Menke, 1828 Family OBOLIDAE King, 1846 Subfamily OBOLINAE King, 1846 Genus SCHMIDTITES Schuchert & Le Vene, 1929 Schmidtites ? micula (M'Coy), emended (Figs 2-10) 1851 Siphonotreta micula M'Coy : 389 1 852 Siphonotreta micula M'Coy; M'Coy in Sedgwick & M'Coy : 1 88; pi. 1 H, fig. 3. 1866 Siphonotreta micula M'Coy; Davidson : 76; pi. 8, figs 2-6. 1974 Schmidtites ? simplex Williams : 26; pi. 1, figs 1 1-15. 1978 Helmersenia ? micula (M'Coy) Cocks : 29. DIAGNOSIS. Subequally biconvex, circular obolids with valves almost as wide as long and 13% as deep as long ventrally, ornamented by concentric fila, fine overlapping lamellae and microscopic radial striations; pseudointerareas narrow, obscure, dorsal interior with fine median ridge. DESCRIPTION. Suboval, gently biconvex obolids with obtuse posterior beaks subtending an angle of about 120 and rounded anterior margins; shells averaging between 92 and 99% as wide as long in 5 samples (e.g. 33 valves from Pen Cerrig: 1 mm (var 1) 3'31 (0'314), w mm (var w) 3'25 (0 - 350), r 0'967)^nd averaging 13% as deep as long ventrally in 2 samples (e.g. 21 valves from Pen Cerrig: 1 mm (var 1) 3'32 (0'188), th (var th) 0'43 (O'Oll), r 0'359); ornamented by strong closely-spaced fila, finely-developed overlapping lamellae and very fine striations; valves with narrow arcuate divided pseudointerareas; brachial valve interior with a low variably-developed median ridge up to 50% of valve length. LOWER ORDOVICIAN BRACHIOPODA 9 MATERIAL length width Lectotype, exterior of p.v SM A.45444 3'6 3-6 Articulated valves BB 92473 2-5 2'6 BB 92470 1-5 (1-4) SMA.44903 4-7 4-5 Exterior of p.v GSM 16457 4-1 4-1 GSM 16472 3-0 2-9 SMA.104456 5-0 4-4 Internal and external mould of b.v. . . . BB 92301 3-0 2*7 Internal mould of b.v SMA.104419 (5-5) 5-0 HORIZONS AND LOCALITIES. SM A.44879-90 and A.45441-57 from the black Llandeilo shales (probably high Glyptograptus teretiusculus or low Nemagraptus gracilis Zones) of Pen Cerrig, 2-5 km north of Builth Wells, (National Grid ref. SO 042540); SM A.4489 1-927 from the olive shales belonging to N. gracilis Zone exposed in Harper's Quarry, Wellfield 2km north of Builth Wells (SO 037534); BB 92301-6 and BB 92470-3 from Upper Didymogmptus murchisoni Shales exposed in Howey Brook, 4 km east of Howey, NW of Builth Wells (SO 091592); GSM 16456-8 from the black Llandeilo Flags of Wyeford, Builth Wells; GSM 16469-73 from an unknown horizon and locality in the 'Upper Llandeilo of the Llandeilo area'; NMW 68. 376.G. 167-70 from Upper Llanvirn shales exposed in stream 600 m SW of Shaky Bridge, 2 km east of Llandrindod Wells (SO 079609); SM A.44838-9, A.44858, 104419 and 104453-6 from basal N. gracilis Shales exposed in Gwern-y-fed-fach quarry, 1-5 km north of Builth (SO 030526). DISCUSSION. The observation that Schmidtites ? simplex Williams (1974) and Siphonotreta micula M'Coy (1851) are synonyms highlights a problem of taxonomic classification which can only be resolved by establishing whether the morphological characteristics of the samples used in this review of the species are of lingulacean or siphonotretacean affinity. An examination of the lectotype (SM A.45444) and paralectotypes from Pen Cerrig near Builth indicates that the shell ornament consists only of concentric growth-lines devoid of spine bases, and fine microscopic radial striations which are best seen under the scanning electron microscope. Occasionally, specimens show a finely granular ornament in associa- tion with concentric growth lines. Specimens from the olive shales of Wellfield are similarly characterized by strong concentric growth lines with some individuals displaying in addition a coarsely granular to pustulose relief on both the internal and external surface of valves. This extra 'ornamentation' is attributable to a microscopic rucking of the entombing shales. The shells of all specimens are very thin and often crumpled or split peripherally, particu- larly posteromedially as in the lectotype. Williams (1974 : 27) noted a similar 'transverse or even radiating' wrinkling of the shells of Schmidtites described by him from the Shelve area. Indeed a symmetrical posteromedian splitting of the shell is such a common phenomenon, particularly in the Pen Cerrig sample, that it has almost certainly been mistaken for a pedicle groove and must account for the fanciful reconstructions of Sedgwick & M'Coy (1852: pi. lH)and Davidson (1866: pi. 8). The presence of an internal dorsal median ridge, divided interarea and microscopic radial striations indicates that these thin-shelled inarticulates have clear lingulacean affinities and can therefore be assigned with some confidence to the genus Schmidtites. Comparisons between the Stapeley, Rorrington, Meadowtown, Pen Cerrig, Wellfield, Wyeford, Howey Brook and Shaky Brook samples reveals that the Stapeley subspecies, identified here as Schmidtites ? micula subcircularis Williams, is different from all seven other samples in being significantly more elongate (p>0'05, p>0'05, p>0'001, p>0'001, p>0'01, p>0'05 and p> 0*005 respectively). Comparisons of the remaining samples reveal that they are all essentially identical. Only the Rorrington and Howey Brook samples differ significantly at the 5% level (0'05 *,***** *~***%*v,Xv ',' ',"*,'* ***'*! ! 59b BHBHBB59C Figs 59a-c Trematis evansi [Addison MS] sp. nov. Holotype BB 36126, dorsal view x 4, ventral view x4 and detail of dorsal ornament of conjoined valves x 10, from Upper Llandeilo lime- stones, St Clears, Dyfed.. 24 M. G. LOCKLEY & A. WILLIAMS has 12 alternately arranged concentric pit rows between 4 and 5 mm anteromedially; pit size, pit and pit row spacing increase progressively from umbo to anterior margin with quincuncial pattern consistent over entire valve surface except where growth pauses result in minor attenuations. Interiors of both valves incompletely known except for pair of rounded dorsal muscle scars on either side of an obscure low median ridge. HOLOTYPE. Complete specimen, BB 36126; length 1 1-0 mm, width 12*3 mm. HORIZON AND LOCALITY. Upper Llandeilo Bryn-glas Limestone, exposed in old quarry 400 m north of Lower Court Farm near St Clears (SN 307 1 52). DISCUSSION. Addison (1974: 144, not formally published) reviewed the affinities of T. evansi, showing that amongst quincuncially-ornamented Trematis it can only be considered to resemble T. parva Cooper and T. mellijlua Reed in density and style of ornament. Both these poorly-known species, however, have sulcate brachial valves and are distinguishable from T. evansi at least in this respect. Genus SCH1ZOCRAN1A Hall & Whitfield, 1 875 Schizocrania cf. salopiensis Williams (Figs 60-65) 1 866 Discina crassa Hall?; Davidson pars : pi. 6, fig. 6, non fig. 7. cf. 1974 Schizocrania salopiensis Williams : 44; pi. 6, figs 22-26. DIAGNOSIS. Schizocrania with subcircular to suboval outline and a broadly triangular pedicle opening, about one-third as deep as long dorsally with a posteriorly placed umbo and dichotomizing capillae developed with variable density; ventral ornament of concentric growth lines. DESCRIPTION. Pedicle valve subcircular to transversely suboval with broadly triangular pedicle opening with straight sides and a well-developed listrum; brachial valve subcircular to suboval, between 85% and 91% as long as wide in 5 specimens, with an evenly convex transverse profile, longitudinal profile asymmetrically convex with posterior umbo overhanging smoothly rounded posterior margin; ventral exterior ornamented by concentric growth lines surrounding the central point at the anterior apex of the triangular pedicle opening and extending onto listrum; dorsal exterior ornamented by sporadically occurring faint growth lines and prominent capillae disposed radially from umbo and numbering 9 to 1 5 per mm at the anterior margin in a sample of small shells, and between 10 and 16 per mm at 5 mm anteromedially in a group of larger shells; capillae reflexed posterolaterally. Dorsal interior with well-defined slightly divergent posterior adductor scars extending forward for about one-fifth of valve length and separated by a median ridge; radial mantle canals number 6 and 7 per mm, 5 mm anteromedially of the umbo, in 4 and 1 specimens from Penddol with each canal therefore corresponding to about two external capillae; ventral interior unknown. FIGURED MATERIAL Exfoliated exterior part of b.v. Internal mould of b.v. . External mould of b.v. Internal part of exfoliated b.v. External mould of p.v. BB 92277 BB 92300 SM A.46560 SM A.44861 SM A. 1044 18 SMA.104413 length 4-1 4-2 10-5 11 8 (6-5) width 4-5 4-6 11-5 12 10-5 (9-0) HORIZONS AND LOCALITIES. BB 92492-4 from Flags and Grits of the Ffairfach Group exposed at Coed Duon, Llangadog (SN 709256); BB 94248 from Flags and Grits, Ffairfach Group type section (SN 6282 11); BB 92275-7 from Upper Llandeilo Flags exposed in dingle 200 m south of Crug, Llandeilo (SN 627229); BB 92278 from Lower Llandeilo LOWER ORDOVICIAN BRACHIOPODA 25 Sowerbyella Limestones exposed in old trackway SW end of Deer Park, Dynevor Park, Llandeilo (SN 609223); BB 92298-300 from the grey shelly sandstones ('Pebbly Felspar Ash' of the Main Tuff Group) 200m SW of Cam peak, Carneddau Hills, 1*25 km ESE of Newmead Farm near Builth Wells (SO 065539); SMA.44861 from the 'Llandeilo Limestone' (Nemagraptus gracilis Zone) of Harper's Quarry, Wellfield, 2 km north of Builth (SO 037534); SM A.46560 from Llanvirn Didymograptus murchisoni shales exposed in Howey Brook (SO 090592); SM A. 1044 18 from Glyptograptus teretiusculus shales exposed in the stream section east of Pen Cerrig (SO 948537). SM A. 1044 13 from G. teretiusculus to N. gracilis Zone shales exposed at Penddol Rocks, in the Wye river section 1 km NW of Builth (SO 03 1522); GSM 16762 from the 'Llandeilo flags of Builth', probably the same horizon, precise locality unknown. DISCUSSION. The specimens, including GSM 16762, SM A.44850, A.46560 and A.51157 from localities in the Builth region, allow us to add information on the species' ventral morphology and the variability of its dorsal ornament to the existing description (Williams 1974:44). The small Llandeilo and the usually larger Builth specimens exhibit more variable dorsal capillae densities than apparently conspecific brachial valves from Shelve. Such pronounced variability in ornament may be indicative of specific variation in the Welsh stock. The stratigraphical range of Schizocrania cf. salopiensis in the Builth region and in the Llandeilo area (Bassett et al. 1974 : 9; Wilcox 1979 : 42) falls within that of the Shelve stock which is known from Upper Llanvirn, Llandeilo and basal Caradoc rocks. Figs 60-65 Schizocrania cf. salopiensis Williams. Fig. 60, SM A. 1044 13, internal mould of a pedicle valve x 6, from Llandeilo shales, Penddol, Builth. Fig. 61, SM A.46560, latex cast of an external mould of a brachial valve x 8, from Llanvirn shales in Howey Brook, Llandnndod. Fig. 62 SM A. 1044 18, internal part of an exfoliated brachial valve x4, from Llandeilo shales, Pen Cerrig, Builth. Fig. 63, BB 92277, a partially exfoliated brachial valve x 8, from Upper Llandeilo shales, Llandeilo. Fig. 64, BB 92300, exterior of a brachial valve x 8, from Llanvirn sandstones, Newmead, Builth. Fig. 65, SM A.44861, latex cast of external mould of a brachial valve x2, from Llandeilo shales, Wellfield, Builth. Fig. 66 Schizocrania multistriata (Reed). Holotype SM A.34039, exterior of a brachial valve x 2, from Lower Caradoc limestones, Lampeter Velfrey, Dyfed. 26 M. G. LOCKLEY & A. WILLIAMS Schizocrania multistriata (Reed), emended (Fig. 66) 1905 Trematis multistriata Reed : 446; pi. 23, figs 1-la. DESCRIPTION. Subcircular to transversely suboval Schizocrania with brachial valve about nine-tenths as long as wide and one-fifth as deep as long, dorsal transverse profile evenly convex, longitudinal profile unevenly convex with maximum depth posterior to midline; ornament of a few indistinct concentric growth lines and prominent radial capillae numbering 10 per mm at 5 mm anteromedially of the dorsal umbo, capillae reflexed posterolaterally. Interior of brachial valve and pedicle valve unknown. HOLOTYPE. Exterior of brachial valve, SM A.34039; length 16*5 mm, width 18*5 mm. HORIZON AND LOCALITY. The only known specimen is from the Lower Caradoc Bryn Bane Limestone exposed in old quarry 250 m NW of Lampeter Velfrey Church (SN 1 53 146). DISCUSSION. This species, of which the holotype is still the only specimen known, resembles S. salopiensis Williams in its external dorsal morphology and may prove to be a senior synonym of the Shelve species. Meanwhile it is transferred to the genus Schizocrania, to which it undoubtedly belongs. Family DISCINIDAE Gray, 1840 Subfamily ORBICULOIDEINAE Schuchert & Le Vene, 1929 Genus SCHIZOTRETA Kutorga, 1848 Schizotreta cf. transversa Williams (Fig. 70) cf. 1974 Schizotreta transversa Williams : 47; pi. 7, figs 2, 3, 7. A single external mould of a Schizotreta pedicle valve (BB 92274) from the Lower Llandeilo Lloydolithus lloydi Flags exposed in the old quarry in Castle Wood, Dynevor Park, Llandeilo (SN 61521 7), compares most closely with the contemporaneous Shelve species S. transversa Williams. The similarity in the relatively transverse outline and lack of well-developed fila is especially noteworthy in view of the distinctive form from the Ffairfach Group described below. Schizotreta transversa Williams ffairfachensis subsp. nov. (Figs 67-69) DIAGNOSIS. Suboval Schizotreta slightly longer than wide, with a surface ornament consisting of well-developed fila. NAME. From Ffairfach. DESCRIPTION. Subcircular to suboval Schizotreta with truncated posterior margin and pronounced concentric ornamentation consisting of strongly-developed fila numbering 5-6 per mm in the anterior and lateral parts of the valve and with fine, closely spaced radial striations over the whole shell; pedicle valve subconical, almost as wide as long (averaging 95% in three specimens) and 20% as deep as long in 2 valves, with a well-developed pedicle track occupying median portion of the valve posterior to apex which is situated medially 29% of valve length anterior of posterior margin; brachial valve 88% as wide as long (1 mm (var 1)7-64(1-413), w mm {var w) 6-44 (4-798), r 0-968 in 5 valves) and 13% as deep as long (1 mm (var 1 ) 7'64 (1-41 3), th (var th) 0'98 (0- 1 37), r 0*966 in 5 valves) with apex situated at 14% of valve length forward of the posterior margin. LOWER ORDOVICIAN BRACHIOPODA 27 Figs 67-69 Schizotreta transversa ffairfachensis subsp. nov. Figs 67a, b, holotype BB 92290a, b, internal and external parts of an exfoliated brachial valve x4; Fig. 68, paratype BB 92292, internal part of an exfoliated pedicle valve x 4; Figs 69a, b, paratype BB 9229 la, b, internal and external parts of an exfoliated brachial valve x 4; all from Flags and Grits, Ffairfach Group, type section. Fig. 70 Schizotreta cf. transversa transversa Williams. BB 92274, exterior of a pedicle valve x 8, from Lower Llandeilo Limestones, Dynevor Park, Llandeilo. Fig. 71 Schizotreta transversa transversa Williams. B 21766, latex cast of external mould of a pedicle valve x 4, from Llandeilo Flags, Shelve, Shropshire. TYPE MATERIAL Holotype, external and internal parts of exfoliated b.v BB 92290 Paratype, internal part of an exfoliated p.v. . BB 92292 Paratype, internal and external parts of exfoliated b.v. BB 92291 length 10-0 11-0 width 9-0 10-5 5-9 HORIZON AND LOCALITY. All specimens are from the argillaceous lower part of the Flags and Grits of the Ffairfach Group, Ffairfach railway cutting, Llandeilo (SN 6282 1 1). DISCUSSION. The Ffairfach Schizotreta differs from the smaller form S. transversa Williams (1974 : 47) from the Shelve area in its more elongate shape and particularly in the strong development of concentric fila. These features are considered important enough to warrant subspecific recognition of the Llanvirn form. The recognition of Schizotreta in the Arenig (Williams 1974 : 48) and Llanvirn Series reveals an extended stratigraphical range for the genus and indicates that the characteristic elongately suboval shape of Scottish and American stocks like S. corrugata (Cooper) and S. medioradiata (Reed) was also typical of early Anglo- Welsh forms. An examination of specimen B 21766 (Fig. 71) from the 'Llandeilo Flags' of Shelve, referred to by Davidson (1866: pi. 7) as Orbiculoidea forbesi Davidson and since referred to Orbiculoidea sp. by Cocks (1978 : 176), reveals that it can be assigned to S. transversa trans- versa Williams. 28 M. G. LOCKLEY & A. WILLIAMS Class ARTICULATA Huxley, 1869 Order ORTHIDA Schuchert & Cooper, 1 932 Suborder ORTHIDINA Schuchert & Cooper, 1932 Superfamily ORTHACEA Woodward, 1 852 Family DOLERORTHIDAE Opik, 1934 Subfamily HESPERORTHINAE Schuchert & Cooper, 193 1 Genus HESPERORTHIS Schuchert & Cooper, 1931 Hesperorthis dynevorensis Williams, emended (Figs 72-80) 1949 Hesperorthis dynevorensis Williams : 226; pi. 1 1 , figs 1 , 2. DIAGNOSIS. Planoconvex to ventribiconvex Hesperorthis with up to 34 regularly spaced costae, numbering 2 per mm at 5 mm anterior of the umbones, and a ventral muscle scar averaging 38% as long as pedicle valve. DESCRIPTION. Planoconvex to ventribiconvex Hesperorthis with suboval outline and slightly obtuse cardinal angles; pedicle valve averaging 82 to 89% as long as wide and 24 to 28% as deep as long (e.g. 1 mm (var 1) 9-08 (5-778), th (var th) 2-18 (0-325), r 0*738 in 13 valves from Newmead) in three samples; brachial valve averaging 74 to 84% as long as wide (e.g. T mm (var 1) 11-14 (18-842), w mm (var w) 14-28 (27-855), r 0-980 in 30 valves from Llanelwedd) and 14 to 17% as deep as long in same three samples (e.g. 1 mm (var 1) 14-75 (4-175), th (var th) 2-33 (0-351), r 0-826 in 6 valves from Llanelwedd); ventral interarea curved apsacline and up to one-quarter as long as valve with narrow delthyrium defined by parallel to slightly divergent boundaries subtending an angle of up to 30; dorsal interarea flat, anacline and up to 15% as long as valve with well-developed fine transverse growth lines numbering 10-1 5 per mm; notothyrium open with divergent boundaries forming an angle of about 60; exterior ornamented by up to 35 regularly spaced rounded costae depending on size; mean amplitude of costae 0*5 mm at 5 mm anteromedially of umbones of 9 valves; counts of 25, 26, 27, 28, 29, 30, 3 1,32, 33, 34 and 35 costae were recorded in 1,0,0, 1,3,4,6,4, 1,0 and 2 valves of specimens from the type horizon at Llandeilo compared with 2, 3, 6, 7, 6, 2, 1, 0, and valves from the Builth area (see Table 2, p. 30). Ventral interior with deep delthyrial cavity bounded by dental lamellae supporting small teeth about 20% as long as interarea with maximum anterior extension at lateral margins of tooth; ventral muscle scar sHghtly raised, averaging between 37 and 39% as long as valve (e.g. 1 mm (var 1) 9'65 (4-563), Isc (var Isc) 3'54 (0'605), r 0'865 in 1 1 valves from Newmead) and 65 to 83% as wide as long in three samples (e.g. 1 mm (var 1) 3-54 (0-550), w~mm (var w) 2*93 (0-424), r 0-885 in 12 valves from Newmead); paired adductor scars, thin and long, represented anteriorly by fine ridges separated and bounded by fine grooves extending beyond the anterior margin of diductor scars to link with vascula media. Dorsal interior with simple low, thin plate-like cardinal process narrowing posteriorly and dividing a broad, low, smooth notothyrial cavity; brachiophores short divergent about 30% as long as_interarea, with bases averaging 19 to 22% as long as valve (e.g. 1 mm (var 1) 11-11 (1 7-50 1 ), Ic (var Ic) 2'38 (0'79 1 ), r 0'947 in 30 valves from Llanelwedd) and 78 to 86% as long as wide (e.g. 1 mm (var 1) 2-45 (0-738), w"mm (var w) 3'41 (1-901), r 0-951 in 18 valves from Llanelwedd); sockets deep; adductor scar pattern quadripartite with anterior adductors smaller than more deeply impressed posterior pair and extending forward for about 60% of the length of the valve, entire field about two-thirds as wide as long and one-third as wide as valve. Figs 72-80 Hesperorthis dynevorensis Williams. Fig. 72, BB 92308, latex cast of the internal mould of a pedicle valve x2; Fig. 73, BB 92307, internal mould of a brachial valve x2; both from Pebbly Sandstones, Ffairfach Group, type section. Fig. 74, BB 923 13, latex cast of internal mould of a brachial valve x 4; Fig. 75, BB 923 14, internal mould of a pedicle valve x 3; Fig. 76, BB92312, internal mould of a brachial valve x4; all from Upper Llanvirn sandstones, Newmead, Builth. Figs 77-78, NMW 68. 376.G. 153-3, latex casts of external moulds of two brachial valves, both x2; Fig. 79, NMW 68. 376. G.I 5 1-1, latex cast of external mould of a pedicle valve x2; all from Llanvirn Sandstones, Tan y Craig, Builth. Fig. 80, BB 92475, latex cast of external mould of a pedicle valve x2, from Upper Llanvirn Ashes and Lavas, Coed, Duon, Llangadog. FIGURED MATERIAL Internal mould of b.v Latex cast of internal mould of p.v. . Latex cast of external mould of p.v. . Latex cast of external mould of b.v. Latex cast of internal mould of b.v. . Internal mould of p.v length width BB 92307 19 26 BB92312 8 12 BB 92308 24 24 BB 92475 20 26 NMW68.376.G.151-1 17 19 NMW68.376.G.153-3 15 22 NMW68.376.G.153-3 13 16 BB92313 11 13-5 BB92314 11 13-5 HORIZON AND LOCALITIES. BB 92307-1 1 from the upper part of the Pebbly Sands Formation of the Ffairfach Group, Ffairfach railway cutting, Ffairfach, Llandeilo (SN 6282 11); 30 M. G. LOCKLEY & A. WILLIAMS BB 923 12-4 from the lower to middle part of the Main Tuff Group with Lower Didymograptus murchisoni Shales, outcrop SW of cairn on Carneddau Hills 1 km ESE of Newmead Farm (SO 065539); BB 92474-5 from rhyolitic conglomerates in the Ashes and Lavas Formation of the Ffairfach Group exposed at Coed Duon, 3 km south of Llangadog (SN 709256); NMW 68. 376.G. 150-61 from tuffaceous sandstones exposed in quarry east of Tan y Grait 1 km north of Llanelwedd, north of Builth Wells (SO 047528). DISCUSSION. Comparisons between penecontemporaneous Hesperorthis from Ffairfach, Newmead and Llanelwedd are affected by the restricted size range of individuals composing the samples and by the fact that the Ffairfach specimens are, on average, more than twice as big as those from Newmead. Consequently, if allometric changes had occurred during growth, tests of significance might show the samples to be different. However, the Ffairfach and Llanelwedd samples show no significant difference in the eight characters tested and the Newmead sample differs only from the Ffairfach specimens in the shape of its shorter ventral muscle scar (0'02^^138 139 SP^I^i Bps- 31 BUI Figs 131-138 Harder ley ella convexa Williams. Figs 13 la, b, BB 92344a, b, internal and external moulds of a brachial valve, both x 3; Fig. 132, BB 92343, internal mould of a pedicle valve x 3; Figs 133a, b, BB 92342a, b, internal and external moulds of a pedicle valve, both x 3; Fig. 134, BB 92346, internal mould of a brachial valve x 4; Fig. 135, BB 92345, latex cast of brachial valve x 3; all from Flags and Grits, Ffairfach Group, type section. Fig. 136, SM A. 33345, internal mould of a brachial valve x4; Figs 137-138, GSM 75247 and 75246 respectively, latex casts of the external moulds of a pedicle and a brachial valve, both x 4; all from Llanvirn Ashes, Bryntowy, Bethlehem. Figs 139-142 Horderleyella sp. Fig. 1 39, BB 92477, a brachial valve exterior x 4, from Llanvirn beds, Coed Duon, Llanadog (p. 46). Fig. 140, GSM TCC. 362, internal mould of a brachial valve x 2, from Lower Llandeilo beds, Cwrt-y-Gorphwys, Ffairfach (p. 46). Figs 141-142, BB 92348 and 92355 respectively, internal and external moulds of pedicle valves, both x 5, from the Pebbly Sands, Ffairfach Group, type section. LOWER ORDOVICIAN BRACHIOPODA 45 Horderleyella sp. (Figs 141-142) DIAGNOSIS. Small transverse mucronate Horderleyella with pedicle valve about 70% as long as wide and simple fascicostellate ornament developing from juvenile costate condition. DESCRIPTION. Small ventribiconvex mucronate Horderleyella with carinate pedicle valve, sulcate brachial valve and strongly angular fascicostellate bundles. Pedicle valve transverse subtriangular in outline, about 70% as long as wide (1 mm (var 1) 6 - 23 (0'934), w mm (var w) 8'85 (1'305), r 0'770 in 1 1 valves) and about 32% as deep as long in 6 specimens, ventral interarea apsacline; brachial valve wider than long and about 15% as deep as long with anacline interarea; external ornament essentially consisting of up to 8 coarse angular costae but exhibiting the following fasciocostellae in a few pedicle valves: It5, lala, 1, 2a2a, 3a and 3a; ribbing patterns on brachial valves poorly known; concentric growth lines fine, regularly spaced throughout. Ventral interior with dental plates and muscle scar respectively extending anteriorly for about one-quarter and two-fifths of valve length. Dorsal interior with open notothyrial cavity divided by simple cardinal process consisting of slightly differentiated myophore and shaft, notothyrial platform extending forward into broad median septum; brachiophores relatively short, blade-like, supported by bases which are almost as long as wide, sockets supported by fulcral plates with underlying crural pits; adductor scars deeply impressed posteriorly in region of adductor pit but poorly known anteriorly. FIGURED MATERIAL External mould of p.v. Internal and external moulds of p.v. OTHER MATERIAL Internal and external moulds of b.v. Internal and external moulds of p.v. External mould of p.v. BB 92348 BB 92355 BB 92349 BB 92352 BB 92350 BB 92351 BB 92353 BB 92354 length (6-0) 5-6 5-0 7-9 8-0 (6-0) 7-3 5-0 width 8-0 9-2 (6-0) 10-0 10-0 8-0 11-0 8-0 HORIZON AND LOCALITY. BB 92348-55 from the upper part of the Pebbly Sands in the lower part of the Ffairfach Group, Ffairfach railway cutting, Ffairfach, Llandeilo (SN 62821 1). DISCUSSION. Attempts to collect taxonomically useful samples of both H. convexa and H. lata from their respective type localities in Long Wood near Bethlehem (SN 693259 and 695262) have proved unsuccessful. However, a large sample of//, convexa from the middle part of the Ffairfach Group, which compares with the type material in every respect, has allowed us to assess the variability of this species. Although we have examined the lectotype (GSM 75245), the paralectotypes (GSM 75246-7) and topotypes (SM A.33345-6) of//, lata, insufficient material is available to assess the variability of this 'species' thoroughly. Our observations suggest that, apart from differences in size, these two Horderleyella 'species' do not differ substantially in any fundamental morphological respect. Such differences as the differentiation of the ventral muscle scar and the prominence of internal ribbing noted by Williams (1949 : 171-173) are most readily attributable to changes in size. We therefore conclude that H. lata should be regarded as a synonym of//, convexa. The small sample of Horderleyella from the Pebbly Sands is particularly distinctive in that the specimens appear to be persistently coarsely costate; similar forms from the Lower Llandeilo (e.g. BB 94058) are associated with strongly fascicostellate individuals (e.g. BB 94059, see below) of the same size. Although unweathered and partially exfoliated shells, e.g. BB 92476 and 92477 (Fig. 139), appear more strongly costate than any external moulds 46 M. G. LOCKLEY & A. WILLIAMS of comparable size, such differences are not entirely preservational as Horderleyella moulds from the Pebbly Sands (i.e. BB 92348-55) are undoubtedly more coarsely costate than similar-sized moulds of//, convexa such as GSM 75246-7 and BB 94059. Detailed collecting throughout most of the key Upper Llanvirn and Lower to Middle Llandeilo sections of the Llandeilo area revealed the sporadic occurrence of Horderleyella in a variety of facies including: BB 92476-7 from loose calcareous blocks from the Flags and Grits Formation of the Ffairfach Group exposed at Coed Duon, 3 km south of Llangadog (SN 709256); BB 94058-9 from the Lower Llandeilo Sowerbyella Limestones exposed to the east of the Old Castle in Dynevor Park, Llandeilo (SN 6 122 17); and GSM TCC.362 (Fig. 140) from the Late Lower Llandeilo calcareous Flags exposed 320m NE of Cwrt-y- Gorphwys Cottage SW of Ffairfach (SN 520207). In addition H. convexa occurs abundantly in sandstones and tuffaceous sandstones at various horizons in the Ffairfach Group. In this respect therefore there are no indications that specific Horderleyella stocks are exclusively confined to particular stratigraphical horizons, although the forms assigned here to Horderleyella sp. cannot at present be shown to be conspecific with H. convexa. Numerous harknessellid species are known from the Ordovician rocks of the Welsh Borderland (Bancroft 1945, Williams 1974), north Wales (MacGregor 1961, Williams 1963) and west Wales (Addison 1974). Until a thorough revision of the family is undertaken, however, we consider it premature to discuss the affinities of//, convexa. Family HETERORTHIDAE Schuchert & Cooper, 1932 Genus TISSINTIA HavliCek, 1970 A study of the various Tissintia samples recovered during the present investigation has shown that there are three distinct forms of this genus within the Llanvirn and Llandeilo successions of south and central Wales. All three were first identified by Williams in 1949. Two of them, T. prototypa and T. immatura, have already been revised (Williams 1974 : 108-1 14) and only the form originally designated 'Resserella immatura var. plana' (Williams 1949 : 167) has to be formally emended. Tissintia prototypa (Williams) (Figs 143-1 51) 1949 Dalmanella prototvpa Williams : 168; pi. 8, figs 7-10. 1974 Tissintia prototypa (Williams) Williams: 108; pi. 17, figs 15-19; pi. 18, figs 1-9, 11. A large sample of T. prototypa specimens (including BB 92369, BB 94062-4 and BB 94237- 40) from a bivalve-dominated coquina in the Didymograptus bifidus Beds exposed in the upper reaches of Camnant Brook, 8 km NE of Builth (SO 088576), resembles the samples described by Williams (1974 : 107-1 14) from contemporaneous horizons in the Shelve and Llandeilo areas. Comparisons between the Builth and the Llandeilo samples reveal no significant differences although in the former sample the costellae are slightly coarser, numbering 4 and 5 per mm, 5 mm forward of the umbones in 8 and 2 brachial valves. Although the Shelve specimens also have a higher density of costellae than those from Builth, the two samples differ significantly (O'Ol p). These differences in the dorsal morphology of Sowerbyella recovered from the Rhyolitic Conglomerates seem to be of limited taxonomic significance. They have been demonstrated for a pooled sample of four collections of predominantly adult valves. The absence of young and immature valves from the shell residues preserved in the Rhyolitic Conglomerates could well account for the statistically identifiable difference in outline because large shells in all samples tend to be more mucronate than young ones. Moreover, the interiors of brachial valves undergo considerable changes during growth. In particular, secondary shell accretion on the socket ridges and septa supporting the adductor muscle bases changes the morphology of such apophyses. The ridges tend to become more stumpy and the septa become ankylosed to form a cleft platform (e.g. Figs 197-198). The diagnosis and description of S. antiqua have recently been amended (Williams 1974 : 1 30). The revision was based on a study of well-preserved moulds of Sowerbyella from Coed Duon which are here shown to be indistinguishable from basal Llandeilo Sowerbyella from Maes y fallen where the syntypes of the species were collected. In these circumstances, no further amendment is called for and the formal diagnosis for the species and its 'variety' remains that cited above. 60 M. G. LOCKLEY & A. WILLIAMS Figs 196-212 Sowerbyella antiqua Jones. Fig. 196, BB 94060, internal mould of a pedicle valve x2-5; Fig. 197, BB 92499, latex cast of internal mould of a brachial valve, showing marginal follicular embayments, x 3; Fig. 198, BB 92495, latex cast of internal mould of a brachial valve x3; Fig. 199, BB 92496, internal mould of a pedicle valve x 3; all from Lower Llandeilo sandstones, Maes y fallen, Llandeilo. Fig. 200, BB 92401, latex cast of internal mould of a pedicle valve x 6; Fig. 201, BB 92403, external mould of a brachial valve x 6; Fig. 202, BB 92404, internal mould of a pedicle valve x 6; all from the Rhyolitic Conglomerates, Ffairfach Group, type section. Figs 203-204, BB 92408 and 92407 respectively, exteriors of a pedicle and of a brachial valve, both x 3, both from the Flags and Grits, Ffairfach Group, type section. Figs 205-206, BB 94039 and 94037 respectively, internal moulds of pedicle valves, both x6; Figs 207-208, BB 94038 and 94036 respectively, internal moulds of brachial valves, both x 6; all from Lower Llandeilo Limestones, Old Castle, south Dynevor Park, Llandeilo. Figs 209-210, BB 94041 and 94040 respectively, exteriors of a pedicle and of a brachial valve, both x 3, from Lower Llandeilo Limestones, west Dynevor Park, Llandeilo. Figs 211-212, NMW 68.376.153-3, internal moulds of a brachial and of a pedicle valve, both x6, from Llanvirn sandstones at Tan y Graig, Builth. LOWER ORDOVICIAN BRACHIOPODA 61 HORIZONS AND LOCALITIES. GSM 10292 (holotype of S. antiqua by original designation) and SM A.I 1314-5 from unknown horizons and localities in the Llandeilo Limestone of Llandeilo, possibly the Lower Llandeilo Sowerbyella limestones of Dynevor Park. Paratypes, GSM 37533-4 from the Llandeilo Limestone in the old quarry 275 m SW of Ffynonddewi near Nantgaredig (SN4785 2075) and GSM 32152a, 32152b 2 from Lower Llandeilo Basal Sandstones 275m SW of Maes y fallen (SN 649210); SM A. 34101-5, BB 942495-9 and BB 94060 also from the same Maes y fallen locality; GSM 7526S 3 and 75270 (syntypes of the llandeiloensis variety, Williams 1949:234) and BB 94040-3 from Lower Llandeilo Sowerbyella Limestones in old quarry 70 m west of Boat House, western Dynevor Park (SN 609223); SMA.34094-5, GSM 75267 and GSM 75269 (also llandeiloensis syntypes, Williams 1949:235) from Lower Llandeilo Sowerbyella Limestones 5m north of Old Dynevor Castle, southern Dynevor Park (SN 61 15 2176); BB 94036-9 from the same horizon 50m east of the Old Castle (SN 6123 2172); BB 92407-8 from limestones in the upper part of the Flags and Grits Formation of the Ffairfach Group in the type section (SN 6282 11); BB 9240 1-6 from the uppermost Ffairfach Rhyolitic Conglomerates Formation at the same locality; BB 35524-34 from the Flags and Grits Formation of the Ffairfach Group exposed on the western side of Coed Duon near Llangadog (SN 709256); NMW 68. 376. G.I 50-1 61 from quarry east of Tan y Graig 1 km north of Llanelwedd near Builth Wells (SO 048528). Superfamily STROPHOMENACEA King, 1846 Family STROPHOMENIDAE King, 1846 Subfamily FURCITELLINAE Williams, 1965 Genus MURINELLA Cooper, 1956 Murinella sp. (Figs 2 13-2 15) DESCRIPTION. Piano- to slightly biconvex Murinella with slightly obtuse cardinal angles and semi-elliptical outline; pedicle valve about 90% as long as wide and 14% as deep as long with apsacline interarea characterized by a narrow pseudodeltidium with a median fold: brachial valve at least 60% as long as wide and about 10% as deep as wide with anacline interarea characterized by a low, broad, medially indented chilidium extending laterally for 1 5-20% of hinge width; exterior of both valves ornamented by fine parvicostellae numbering 6 per mm at 5 mm anteromedially of the umbo, crossed by fine regularly developed fila (numbering about 25 per mm) and sporadically-occurring growth lines. Ventral interior exhibiting ill-defined muscle field medially divided by indistinct low, narrow ridge marked by very fine longitudinal median striations within muscle field; muscle scar about 77% as long as wide extending anteriorly for about 32% of valve length and bounded posterolaterally by short, widely divergent dental plates extending anteriorly for about 18% of valve length and supporting simple teeth; diductor component of scar characteristically elongate and occupying the lateral sectors of the muscle field. Dorsal interior with a bilobed cardinal process with a small undifferentiated median process flanked by well-developed subsidiary ridges acting as diductor bases, and socket ridges extending laterally, parallel to the hinge line, for about 45% of valve width; median septum broad and low extending anteriorly from the low, broad notothyrial platform for about 40% of valve length with a fine narrow median ridge continuing to commissure and dividing well-impressed, semicircular dorsal adductor scars which are further divided by a few variably developed fine, radiating ridges and grooves. 2 Spjeldnaes (1957 : 84) chose this as the lectotype of antiqua, a choice which was also recognized by MacGregor (1961 :201). However, since the holotype had already been designated by Jones (1928:419; pi. 21), this was superfluous. 3 Cocks (1978 : 97) chose this as the lectotype of llandeiloensis. 62 M. G. LOCKLEY & A. WILLIAMS FIGURED MATERIAL Internal and external moulds of b.v. Internal and external moulds of p.v. External mould of p.v. BB 92461 BB 92462 GSM 10889 length (12) 8-5 16-5 width 20 9-5 18-0 HORIZON AND LOCALITY. BB 92461-3 from the Flags and Grits in the middle of the Ffairfach Group, Ffairfach railway cutting, Llandeilo (SN 6282 11); GSM 10889 probably from the same horizon and locality. DISCUSSION. The brachial valve BB 92461 exhibits pronounced growth distortions on both anterolateral flanks; each side of the valve initially grew radially but then irregularly towards the anteromedian and anterolateral sectors of the shell. A single pedicle valve ofMurinella is known from the Lower Llandeilo Meadowtown Beds of the Shelve area (Williams 1974: 141) and compares with the Ffairfach valve in the arrangement of the muscle field. In the absence of further material, however, it is not possible to do more than conclude that these penecontemporaneous forms may be closely related. Figs 213-215 Murinella sp. Fig. 213, BB 92462, internal mould of a pedicle valve x 3; Fig. 214, GSM 10889, latex cast of the external mould of a pedicle valve, x 2; Figs 2 1 5a, b, BB 9246 la, b, internal mould and latex cast of the external mould of a brachial valve, both x 2-5; all from the Flags and Grits, Ffairfach Group, type section. Subfamily OEPIKINAE Sokolskaya, 1960 Genus MACROCOELIA Cooper, 1956 Macrocoelia llandeiloensis (Davidson) (Figs 2 16-226) 1871 Strophomena compressa (J. de C. Sowerby) var. Llandeiloensis Davidson : 316; pi. 46, figs 11-1 4. 1959 Rafmesquina ? llandeiloensis (Davidson) Spjeldnaes : 16; pi. 1 . figs 1-7. 1961 Macrocoelia llandeiloensis (Davidson) MacGregor : 206; pi. 23, figs 1-10. DIAGNOSIS. Planoconvex Macrocoelia, becoming slightly geniculate in late adult stages of growth with subperipheral rim; finely ornamented with about 8 parvicostellae per mm, 10 mm anterior of umbo, and with a flabellate ventral muscle scar over one-third as long as the adult pedicle valve. DESCRIPTION. Planoconvex semicircular to semi-elliptical Macrocoelia with slightly obtuse rounded cardinal angles and maximum width at, or more commonly just anterior to, the hinge line, variably but evenly convex in transverse profile and unevenly convex in longi- tudinal profile with incipient geniculation developed in adult specimens; pedicle valve averaging 73% and 78% as long as wide in two samples from Ffairfach and Coed Duon respectively, with smaller and larger mean size (N = 35 and_N = 7), and averaging 14% jis deep as long (range 8 to 22%, e.g. 20 valves from Ffairfach: 1 mm (var 1) 10-88 (5'383), th LOWER ORDOVICIAN BRACHIOPODA 63 (var th) 1-56 (0*585), r 0'64 1); brachial valve _averaging 72 to 77% as long as wide in the same two samples (e.g. 20 valves from Ffairfach: 1 mm (var 1) 8*87 (8' 165), w mm (var w) 12*35 (16-792), r O974); ventral interarea long apsacline with transverse growth lines parallel to hinge and a well-developed pseudodeltidium and low chilidium, dorsal interarea short anacline; radial ornamentation unequally parvicostellate with 5, 6, 7, 8 and 9 ribs per mm at 10 mm anteromedially of the umbones of 1 , 4, 5, 8 and 2 brachial valves respectively. Ventral interior with widely divergent dental plates supporting simple teeth and extending anteriorly for an average of 14 to 21% of valve length (e.g. 20 valves from Ffairfach: T mm (var 1) 9-64 (9'343), dl (var dl) 1-40 (0-304), r 0-770) and for an average of 45 to 49% of their maximum lateral extension in the same two samples (e.g. 23 valves from Ffairfach: I mm (var 1) 1-37 (0-302), w mm (var w) 3'05 (2-010), r 0892); muscle field poorly developed, equidimensional in small specimens with rare incipient flabellate patterns at lateral margins but with strongly-developed, flabellate paired diductor scars (in larger specimens) averaging 85% as long as wide and 38% as long as valve and enclosing narrow longitudinal adductor scars medially; pedicle tube short and narrow in small specimens but vestigial or absent in larger individuals with tendency for closure to begin anteriorly. Dorsal interior with stout bifid cardinal process characterized by flattened anterior attach- ment platforms disposed roughly parallel to the interarea and extending anteriorly for up to 8% of vajye length (in larger specimens, e.g. 13 valves from Ffairfach: I mm (var 1) 8*88 (9-048), Ic (var Ic) p-52 (0-031), r 0*886) and for 63 to 73% of their maximum lateral extension in respective samples (e.g. 13 valves from Ffairfach: 1 mm (var 1) 0'52 (0*031), w mm (var w) 0'83 (0*073), r 0*903); processes tapering posteriorly towards umbo and ankylosed anterodorsally to widely divergent socket ridges which extend anteriorly for 10 to 14% of valve length (e.g. 12 valves from Ffairfach : I mm (var 1) 8*88 (8*778), Is? (var Isr) 0*89 (0*055), r 0*872) and for 31 to 39% of their maximum lateral extension in respective samples (e.g. 12 valves from Ffairfach: 1 mm (var 1) 0*89 (0*055), w mm (var w) 2*86 (0*737), r 0-862); genital markings, consisting of rounded raised pustules, are confined to socket region between socket ridges and hinge line and developed sporadically; notothyrial platform low, obscure, passing anteriorly into a low, broad median septum dividing muscle field; musculature only well-defined in adult individuals where adductor scars average 38% as long as valve (e.g. 10 valves from Coed Duon; T mm (var 1) 19*45 (8*914), Isc (var Isc) 7*75 (2*014), r 0*875) and 77% as wide as long (e.g. 1 1 valves from Coed Duon: 1 mm (var 1) 7*52 (2*404), w mm (var w) 9* 71 (4*081), r 0*931); radial ridges at an angle of about 30 to median line and dividing the adductor scars into anteromedian and posterolateral sectors. The greater parts of the posteromedian sectors of both valves are characterized by fine pustules which are strongly elongated along radial lines and tend to become thickened and amalgamated into irregular, slightly anastomosing radial ridges in some larger individuals; these markings are absent towards the periphery which in large forms invariably exhibits a pronounced subperipheral rim which is fluted, particularly dorsally, by radial mantle canals of a lemniscate system. FIGURED MATERIAL Internal mould of b.v. . Internal and external moulds of b.v. Internal mould of p. v. . Internal and external moulds of p.v. . Internal and external parts of exfoliated p.v. exterior length width BB 92421 BB 92445 BB 92432 BB 92431 B 13618 BB 92433 BB 69520 BB 92442 NMW68.376.G.176.3 BB 92443 BB 92444 22 21 11 10*5 16-5 14*5 22-5 12 17 11 25 28 15 15 19 23 27 23 16 22 14 64 M. G. LOCKLEY & A. WILLIAMS Figs 216-226 Macrocoelia llandeiloensis (Davidson). Fig. 216, BB 92421, latex cast of the internal mould of a brachial valve x2; Fig. 217, BB 69520, internal mould of a medially split pedicle valve x 1 ; both from Llanvirn shales, Coed Duon, Llangadog. Fig. 2 1 8, B 1 36 1 8, internal mould of a pedicle valve x2, from the Ffairfach Group, type section. Fig. 219, BB 92445, internal mould of a brachial valve x 1-5, from the Ashes and Lavas, Ffairfach Group type section. Fig. 220, BB 92433, internal mould of a pedicle valve x 2; Fig. 22 1 , BB 92444, internal part of an exfoliated pedicle valve x 3; Figs 222-223, BB 92442 and 92443 respectively, internal moulds of pedicle valves, both x2; Figs 224-225, BB 92432 and 92431 respectively, internal moulds of brachial valves, both x 3; all from the Flags and Grits, Ffairfach Group, type section. Fig. 226, NMW 68. 376. G.I 76. 3, internal mould of a pedicle valve x 2, from Llanvirn shales at Tanlan. Builth. Figs 227-230 Macrocoelia llandeiloensis elongata subsp. nov. Fig. 227, holotype BB 92409, internal mould of a pedicle valve x 2; Fig. 228, BB92410, internal mould of a pedicle valve x 2; Fig. 229, BB 924 1 1 , internal mould of a pedicle valve x 3; Fig. 230, BB 924 1 8, internal mould of a brachial valve x 4; all from Llanvirn sandstones, Howey Brook, Llandrindod. LOWER ORDOVICIAN BRACHIOPODA 65 HORIZONS AND LOCALITIES. BB 92421-8 and BB 69520-5 from argillaceous shell beds in the Flags and Grits Formation on the west side of Coed Duon ridge, 3 km south of Llangadog (SN 709256); BB 92429-39 from argillaceous lower part of the Flags and Grits in the Ffairfach Railway Cutting, Ffairfach (SN 62821 1); BB 92441 from the uppermost part of the underlying Pebbly Sands Formation at the same locality and BB 92442-4 from the lime- stones in the upper part of the Flags and Grits Formation exposed immediately to the south in the railway embankment; BB 92445 from the lower ashy part of the overlying Ashes and Lavas Formation at the same locality; BB 92440 from the sandy beds of the Ffairfach Grit Formation exposed in the field beside the Longwood road 300 m NW of Beili-dyffryn Farm (SN 693257); B 13618 from an unknown horizon and locality at Ffairfach (probably from the same locality as BB 92441); NMW 68.376.G. 176.2-3 from Upper Llanvirn sandstones exposed 40 m from gate on hill road from Tanlan, 4 km NNE of Builth Wells (SO 057547); SM A. 34098-1 00 from Lower Llandeilo sandstones exposed 275 m SW of Maes y fallen, Llandeilo (SN 6492 10); SM A. 34096-7 from Lower Llandeilo beds exposed at Dynevor Park Old Castle, Llandeilo (SN 6 1 1 5 2 1 76). DISCUSSION. Comparison between the two samples described here indicate that the smaller- sized Ffairfach individuals are significantly more transverse in ventral outline than the larger individuals from Coed Duon (0'02 ^^2 < -S -S -o-o ^ S S fe o C O O O oo oo 22 ^ -| -|J <*5 ON K g -^ c I Q Q Q .Jj .y .^J a a a > cxo 83 G C 5 .2 o O C -2 =o s "c 3 22 J > C G , cd c$ ^* t-l ^ CGCCCCGGGGC Q cOcOcOcOcOcOcOcOcOcOcO 'c 'c 'G 00 !T\ r^ Si (U crj c^ rt o ^ ^ - _ llllllU C G C S S co eo co 2 2 (U >>UU '- 1 CCCCCCaCCCC 3 Q.CUCJDDUDUDDDCJ CJ OJ (U J2 vo SO vo sO ^O so SO ^O r_~ SO SO vo ^O SO SO SO SO Os Os Os Os Os Os Os Os ^ m C i^ OS t/5 -o H cO G .. m so fr, "O OS ^ ^ m VO OS ^O -Q ^ ; e ccc CGC^COO oSooooooo2o ' fjOS^Gflc/JliOoOC/JC/lOSt/J m-c^CCcCCGC 'C E m/S c^-^^^xusx: GJ= S OsQcOOOoOOOOcOOg ^^2>-> | - | - ( -> | - | - | ->g- G J2.2 "^ ro O 6 O (^ J5 U. o ^3^ g : J2Ug o ^^ "^8 ^ ^ rt *O ^^ w^ O OO ^^ ^^ O * * -^"^^ ^^ 53 43 Q, *5 o, M ^ tSQCO-^^^^^^^'CO^ C S c gS-S !c = '1 .2 'S US I'll I's-Gll cx^uDCKffiKEKXuffim C/3 3 O (X S' * .i-i wv r ^ ^^ *-*. 5^oS6uu^ ^^^Scx I g tfl o G B x: u oo -7^ ^ c S 2 2 1 ^ 'Q ^"^ "Q 8 5 o ^ o 'ii O Cy (3 S a -s: cv G. ;."* CL 2 S2 S ID vo t ooosO '(Nm^v^o so r- < ^ HfNrSfNCNCNrS 1 /^ rs rs f^> ro 2 55 o -~ c ' -~ CO a I a a a 32 rs S^ S c^S aw^K^ci,^ .^S^a-S^^-ScS^feS ^ ii^ll^liNiS ^ o *j2 s^S^^^cc to -C) ^ 'ir^i^ s: Ci, d, x. t<, x -~ > ^j ^ a i; S 1? S| || 1.1 C ^3 53 N 'B s S c ^2 2 OO^^i^QC.^ - 1 J 1 1 84 f i c c co J5 co co I o S 1 2 o o CO U t- co .2 *-! '- &N 3 00 ? ~^ w -j f ~^ f |^ ^ C o O a JZ'G'G & O c . O g i- c; ._2 S b b fe IT il 1 f M M C J5 S C C C C a> U 5 (U ' (j .2 "^ ^ u cj ^; IX o SE E g a a a-g C-> (L> QJ OJ -rt '2 0000'%'% <3 g 2 i c ^ CO -Q ^ T3 32*DDD3 O O O O M i) J nJ i 1 J 2 (X c^ c^ c^ c^ ^ ^ O 03 03 03 CQ ^ ^ -J 52 S 00 ON Author Paul 1937 J.H.Johnson 1946a J.H.Johnson 19466 J.H.Johnson 19466 Bilan & Golonka 1973 Yabel912 Kulik 1973 Bilan & Golonka 1973 Paul 1938 Flugel & Flugel-Kahler Homan 1972 Pia 1937 J.H.Johnson 1950 Wetheredl890 Wetheredl889 Wetheredl890 Wetheredl890 Bornemann 1886 Dragastan 1975 ) Yabe&Toyama 1928 Romanes 1916 Colin & Vachard 1977 c | 1 2 c s op "co W 'r* co CTJ i: O o ^5 *C 53 . - c 2.a c 00 S SG "rt K ~2 ?J _C O d, . oo -c s: c 'C S- S^S O Ju" ob ^ [7* vo r~~ oo ON vo o " *O (N CO ^" i/~> vo r~~ oo ON o -~ (N ro 1 C ^ 2 ! ^ o "2 "^ s; .s: o s: O} Q .a >w o > (3 ~S "^ .^ ^ fN" CO Q o st *. ^ ^ ?S a i2 li i 2 '% -g 1 "o U a 00 ^ d, O &o'? s; * 1 11-1 'IS, 1 !>"! 1 ^ ^ cx} t - * *^ ^ >5 oj St 2 *^* 2 j^ 13 ^ ^ -S 'S,'i -S -s .^ i c ,2 'S a 5 ex THE FOSSIL ALGA GIRVANELLA 85 microscope. Acetate peels were also tried, but were found unsuitable because of the fine grain-size of the limestone, and because of certain artefacts inherent in their use. Registered numbers given with the halftone illustrations (Figs 2, 3, 5) are those of the British Museum (Natural History), Dept. of Palaeontology. Measurements were made with an eyepiece graticule. This method gives results to about the same degree of accuracy as Wood's (1957 : 24) technique of measuring from photo- graphic prints. The correction of errors introduced by photographic processes is cancelled by the small scale of the graticule. The number of tubes present in growths varied a great deal, as did the degree of preser- vation. Frequently the boundaries of a tube were obscured by neomorphism or by contact with other tubes. It was necessary to measure both diameters for each tube since comparisons of the two diameters were needed. For this reason, only tubes with both inner and outer boundaries clearly apparent were measured, and other tubes were ignored. All the tubes, up to a total of ten, with both boundaries present were measured in each growth. As a result some growths are represented by measurements on only one or two tubes, but no justification could be found for omitting them. In any case measurements were analysed by considering the entire sample, rather than individual growths. Ten was chosen as the upper limit because most growths had only ten or fewer suitable tubes. Where two distinct size-ranges occurred in the same growth, ten tubes were measured from each. A plane section cut through tubes arranged more or less randomly will contain a range of sections of tubes from cross to longitudinal. Perfect cross sections are circular, because of the cylindrical shape of the tubes, but oblique sections are elliptical (Fig. 1). It can be shown that the smallest diameter of such an oblique section is a true diameter of the tube. All measure- ments have therefore been made on this smallest diameter. The tube thickness may have irregularities, so that estimation was sometimes necessary for measurements of the external diameter. However, it was possible in most cases to measure the two diameters along the same line. For scanning electron microscopy, rock samples were prepared by smoothing rock slices on carborundum powders and polishing with a series of diamond laps, finishing with a 1 (im lap. This was followed by etching for 30 sees in 1 vol. % HC1. The slices were then attached to specimen stubs with a conducting paint, Durofix or double-sided sellotape, and were coated with gold-palladium in a vacuum-evaporator. Samples of Recent algae were obtained from a variety of sources, either by personal collection or through the courtesy of other workers. These samples were stored at 3 C, in 0'8% glutaraldehyde buffered to pH 7'4, and were prepared for electron microscopy as follows: 1. Immersion fixation in 2- 5% glutaraldehyde, buffered to pH 7'4 with KH 2 PO 4 and K 2 HPO 4 , for two hours. 2. Washing in several changes of buffer. 3. Immersion in 1% OsO 4 , in buffer, for 30 minutes. 4. Gradual transfer to water-free acetone. 5. Critical point drying. The scanning electron microscopes (S.E.M.) used were Cambridge S2A models, operated at 18 KV at lower magnifications and at 28 KV to 30 KV at magnifications of 5000x or more. The final aperture diameters were 200 urn for lower and 140 urn for higher magnifi- cations. Stubs were held at 30 to the electron beam to reduce the effects of charging, with a working distance of about 10 mm. Historical review Species ofGirvanella have been described, under various generic names, from rocks of Upper Proterozoic (Korde 1973 : 212) to Middle Cretaceous (Colin & Vachard 1977) age. A list of species is given in Table 2. 86 H. M.C. DANIELLI Girvanella was first described by Nicholson & Etheridge (1878 : 23) as an agglutinating foraminiferan related to Rhizammina algaeformis Brady. They gave the specific name problematica to their fossil, for convenience, and provided the following diagnosis: Generic characteristics: 'Microscopic tubuli, with arenaceous or calcareous (?) walls, flexuous or contorted, circular in section, forming loosely compacted masses. The tubes apparently simple cylinders, without perforations in their sides, and destitute of internal partitions or other structures of a similar kind'. Specific characteristics: Tubes from 1 -600th to 1 -700th of an inch in diameter, not observed to taper, twisted together in loosely reticulate or vermiculate aggregations of a rounded or irregular shape, which seem to be mostly from l-20th to l-10th of an inch across'. Nicholson (1888 : 22) gave a further description of the organism, but did not redefine it. Apart from assigning a different range of diameters to the genotype (16 um to 40 um) he added nothing to the original publication, and still believed the organism to have been a foraminiferan. This view was held by some workers for a long time, Rhumbler (1895) proposing a subfamily Girvanellinae of the Rhabdamminidae Rhumbler to contain the genus. Seely (1885) described a genus of calcareous sponge, Strephochetus, from Middle Ordovician limestones in Vermont, and in 1886 Bornemann described Siphonema from Sardinian deposits of the same age. Bornemann compared his material with epilithic cyanophytes, and considered that Siphonema was related to them. Hinde (1887:227) recognized Strephochetus and Siphonema to be synonyms of Girvanella, but disagreed with both suggested affinities. Rothpletz (1891 :301) assigned Girvanella to the Codiaceae ((Trevis) Zanardini 1843) because of the dichotomous branching shown by some species, and its similarity to Sphaerocodium Rothpletz 1891. The latter has been assigned to the Siphonae, a taxon con- taining both the Codiaceae and Dasycladaceae (Endlicher) Cramer 1 888. This assignment to the green algae was accepted by Brown (1 894 : 203). Wethered (1893 : 246) accepted that some forms of Girvanella appeared to have been plants, and might have been calcareous algae. In the discussion of this paper (p. 248), Reid suggested that the calcareous tubes typical of Girvanella were the result of inorganic encrustations on filamentous plants. Seward (1898 : 125) and Pollock (1918 : 255) compared Girvanella to the calcified sheaths of some Recent Cyanophyta, thus confirming Bornemann's (1 886) opinion. Pia (1927 : 37) placed Girvanella and similar genera in an artificial group, the Porostro- mata. This was accepted for some years. In 1935 Fremy & Dangeard proposed that the Jurassic species G. minuta Wethered should be renamed Symploca jurassica, because of its resemblance to the Recent species S. hydnoides Kutz. However, Johnson & H0eg (1961 : 54) expressed doubts about the assignment of Girvanella to the Cyanophyta, and suggested that it was a member of the Chlorophycophyta. This proposal seems to have been based on his own removal of many of the Porostromata to the Codiaceae, and also on the belief that Fremy & Dangeard had described G. minuta as a chlorophyte. Riding (1975 : 174) has restored these genera to the Porostromata. Dricot & Tsien (1977), in a discussion of the validity of the genus Rothpletzella Wood 1948, have pointed out the partial synonomy between Girvanella and Sphaerocodium Rothpletz 1891. Several of the species of the latter were therefore assigned to Girvanella by these authors, in part or completely (Table 2, p. 83). Korde (1973 : 2 12) has mentioned several Upper Proterozoic and Lower Cambrian genera which appear to be girvanellids. The genera involved are Nicholsonia Korde, Fistulella Korde, Botominella Reitlinger (1959), Kenella Korde, and Batinevia Korde (1966). Mamet & Roux (1975) have commented on the resemblance between the last three and Girvanella. Korde (1973) proposes a new class of Rhodophyta, the Protobangiophyceae Korde, to contain these genera. But some of the descriptions of the protobangiophycean genera seem to bear little resemblance to the figures provided, and the reasons for the separation of these forms are not expressed in such a way that they can be evaluated. THE FOSSIL ALGA G1RVANELLA 87 Nicholsonia in particular seems to have been interpreted on the understanding that the original organism was a member of the Rhodophyta. Structures which are absent from the fossil are said to have been uncalcified. Korde's descriptions are not included in the systematic analysis below, as it is difficult to reconcile her accounts with other descriptions. Riding (1977) has related the impregnated sheath of Plectonema gloeophilum Borzi, a modern species with rare branching, to Girvanella, and extends the range of the fossil to Recent times on this basis. However, the branching habit of Plectonema does not agree with the 'simple cylinders' definition of Girvanella (Nicholson & Etheridge 1878 : 23). The ultra- structures of the carbonate are also rather different, that of P. gloeophilum consisting of calcite needles in various orientations. Girvanella tubes comprise equidimensional micrite and radially-arranged needles, the structure having greater regularity and much lower porosity (Fig. 2). The genus Girvanella Further descriptive remarks can now be added to the original definition of the genus (Nicholson & Etheridge 1878). Growths vary in size from single tubes to clusters more than a centimetre across. Growths with many tubes may be almost circular in section, completely irregular, or in some intermediate form, and may have a core body such as a detrital particle or, more commonly, another fossil. The boundaries between growth and rock matrix may be distinct, eroded, micritized, diffuse or neomorphosed. Often the matrix seems to have neomorphosed more easily than the carbonate tubes. Fig. 3 shows a range of growth habits from the Stinchar Limestone. In addition to the problems of defining the genus Girvanella some confusion has arisen in its subdivision. A list of species is given in Table 2, pp. 82-84, in which there are about 40 Lower Palaeozoic taxa. Only a few of these have diagnoses not contradicting the generic diagnosis, as shown in Table 3, and many have indistinguishable characteristics. The holotype of G. problematica Nicholson & Etheridge, the type species of the genus, has been destroyed. Wood (1957 : 23) discovered this and designated a neotype for the species. However, although he did not emend the generic diagnosis of Nicholson & Etheridge (1878), he redescribed the type species in a way which differs from the original generic diagnosis. This anomaly means that Girvanella and its type species problematica are still defined according to Nicholson & Etheridge (1878). Wood's neotype is kept in the British Museum Table 3 Lower Palaeozoic species ascribed to Girvanella with characteristics contradicting the original generic definition; brackets indicate observations made from figures. (atratus) incompta problematica* pusilla (brainierdi) media p. lumbricalis ramosa Branching effusa mexicana p. typica sarmenta o High density convoluta (effusa) media ocellatus pusilla (ramosa) 1 S incompta problematica* sarmenta 8^ Adherence effusa fragila incompta media prolixa pusilla ramosa sarmenta Septation conferta problematica* prolixa 'as emended by Wood (1957). 88 H. M. C. DANIELLI (Natural History) collection, registration number V 34566, and since it does in fact conform to the original generic diagnosis, it is accepted here as the neotype of Girvanella. Considerable information has accumulated since 1878 about both Girvanella itself and the Cyanophyta. Nicholson & Etheridge's (1878) diagnosis needs to be re-examined and related to this new body of information. To facilitate this, the characteristics used to define the genus will be considered: (1) microscopic; (2) tubular; (3) arenaceous or calcareous; (4) sinuous; (5) circular in cross section; (6) low growth density; (7) simple cylinders; (8) imperforate; (9) no septa or other internal structures. Of these, points 1,2, and 8 have never been disputed. Most species have been described as circular (point 5), but J. H. Johnson (1950 : 61) gave G. texana as 'nearly circular'. Wood (1963 : 26) described irregular tubes, attributing them to post mortem collapse. No non- calcareous tubes have been reported (3), although some associations with iron oxide or pyrite have been published (H. M. Johnson 1966 : 51; Lewis 1942 : 52). Playford et al. (1976 : 558) noted that bacteria are more likely to be responsible for such relationships than the algae themselves. Silicified Girvanella has also been described (Lewis 1 942 : 5 1 ), but this example was a secondary replacement. The hollow nature of the tubes (2) has never been questioned, and Lewis (1942 : 52) gives evidence for it. However, both branching and constrictions (7) have been described by several authors (Table 3). The figures provided in these cases commonly show tubes which cross or bend out of the plane of section, leading to misinterpretation (cf. H. M. Johnson 1966 : pi. 12). Septate girvanellids (9) have also been described, but these reports are the result of observation of refraction across grain boundaries within the cell space (Wood 1957 : pi. 5). Fig. 2, opposite, illustrates the causes of apparent branching, constrictions and septa. Many authors have described growths with close-packed, adherent tubes, although Nicholson & Etheridge (1878) describe the genus as being loosely-compacted. In fact a great variety of growth densities (defined here as the ratio of tube volume to matrix volume in a growth) occurs in many populations, and the value of density as a generic or specific characteristic needs reconsideration in the light of population studies. The same applies to tube sinuosity (4). Some examples demonstrating the variability of single populations are given in Fig. 3. Girvanella and Recent cyanophytes Some modern Cyanophyta will calcify under natural conditions, when an excess of calcium is present, although they will not do so readily in culture. Lewin (1962) and Golubic (1973) have given reviews of the subject, and Pentecost (1978) has conducted a detailed study of cyanophyte calcification. It is usually said that cyanophytes calcify only in freshwater environments, although many fossils, supposedly calcified cyanophytes, are primarily marine in origin. Monty (1977) has discussed the matter in terms of stromatolites. Some Recent forms do calcify in regions where storms may cover them with sea water from time to time, and further research may find fully marine forms which calcify. Almost the only thing which seems to be common to the environments in which cyanophytes calcify is the high calcium level. Some calcium is clearly necessary, but Fig. 2 Fine structure of Girvanella and Rothpletzella. (a), Girvanella with apparent branching. White lines mark tube directions. Bar = 7 urn. (V 60469). (b), Girvanella with apparent branching due to crossed tubes (SEM). Note the rounded ends of tubes a and b, against c. Bar=10um. (V 60473). (c), apparent septa, marked s. The tube itself (t) comprises equidimensional micrite, but this is missing from the 'septa'. They are grain boundaries, lying across the tube axis. Bar = Sum. (V 60473). (d), tubes with a prismatic ultrastructure (p). Bar= 10 urn. (V 60473). (e), Rothpletzella with equidimensional micrite forming its tubes (t). Bar = 5 urn. (V 60 1 03a). (f), Rothpletzella with radial prisms (p). Bar = 1 urn. (V 60 1 03a). SB, 90 H. M.C. DANIELLI blue-green algal deposits seem to occur only in waters which are saturated with it. Neither high light intensities nor warmth are essential, and water-flow rates vary considerably. These generalities apply to the group as a whole, however, and individual strains may be more exacting. This may be the cause of many of the difficulties met in laboratory studies of calcification. It seems then, that the ecology of modern cyanophytes is of little help in the interpretation of fossil forms. That Girvanella had a wide ecological range implies only that several biologically distinct taxa were involved. The genus occurs in shallow-water marine lime- stones (from faunal evidence), and may be quite rare or may form 80% of the rock volume (Williams 1962 : 19). It is found with benthic faunas in calcareous muds or silts, and in reef environments. Often it is the only cyanophyte represented, but other porostromates may be present and calcified eucaryotes may also occur. Palaeocological evidence suggests growth in quiet or only moderately turbulent conditions, but such estimates are questionable. It is likely that uncalcified algal mats were present in the same environments, stabilizing sediments and preventing the development of current structures. Modern cyanophytes may either impregnate or encrust their mucilaginous sheaths. Many organisms do both. The nature of the internal diameter of the carbonate tube depends on these differences. Distinguishing between them is therefore important: this matter is con- sidered in more detail below. The carbonate itself is normally calcitic and low in magnesium, iron and strontium, although Monty & Hardie (1976 : 463) have described carbonates with 16mole%Mg, formed in association with Scytonema myochrous (Dillwyn) Agardh. Modern cyanophyte carbonates are therefore stable in most cases, and are not subject to rapid diagenesis. However, since most modern calcifying blue-greens are freshwater organisms, comparisons with the predominantly marine Girvanella should be drawn with caution. The fine structures of modern cyanophyte carbonates vary considerably, from the acicular type described by Gleason (1972 : 155) to the micritic textures figured by Schafer & Stapf (1978 : fig. 4). Flajs (1977) describes the carbonates of three members of the Rivulariaceae as similar to those ofChaetophora (Chlorophyta). In these the filaments become encrusted with very fine calcite grains, which merge to form larger aggregates. Flajs believed all calcifying cyanophytes to follow this pattern. The close similarity between them and certain green algal carbonates, which he demonstrated, is of considerable interest. Differences certainly occur from cyanophyte to cyanophyte, as shown in Fig. 4, but the differences lie in grain size and arrangement, rather than in grain shapes. Acicular carbonates like those described by Gleason (1972) or Krumbein & Potts (1978) are com- paratively rare. There is no clear evidence that these ultrastructures can be used to distinguish cyanophyte genera in the biological sense, but some similarities within families seem to exist, especially if taken with other factors (Danielli in prep.). A description of the ultrastructure of Girvanella has, then, a place in the diagnosis of the genus. Some similarities are shown by the fine structures of Girvanella and Rothpletzella (Fig. 2, p. 89), emphasizing the close relationship of these two genera. Early systematic studies of modern Cyanophyta were based on morphological and ecological grounds (see Geitler 1932). Drouet (1962, 1963) showed that cyanophyte morphology is very dependent on the environment of growth, and considerably condensed the number of taxa in his revised classification of the Oscillatoriaceae (Drouet 1968). The systematic study of Rippka et al. (1979) seems to be even more revolutionary, but will put Fig. 3 Growth habits of Girvanella in the Stinchar Limestone at Aldons Quarry, (a), moderate density, variable parallelism and adherence. Bar = 20 urn. (V 15956). (b), elongate growth with external zone of subparallel adherent tubes, and internal zone of random adherent tubes, high density throughout. Bar =100 urn. (V 60470). (c), subparallel tubes with lower density and adherence. Bar = 50 urn. (V 60471). (d), high density adherent tubes with random arrangement. Two tubes are in complete circles (c). Bar = 20 urn. (V 15965). (e), isolated tube with a cement overgrowth. Bar= 5 urn. (V 60472). THE FOSSIL ALGA GIRVANELLA 91 ^ i ,.. -^ '*m .-V ^^&-.$S& > -'I, 92 H. M. C. DANIELLI identification on a more objective basis. The organisms are cultured on standard media, so that a comparison of genotypes becomes possible. The method has been discussed by Whitton, Holmes & Sinclair (1978 : 64) and by Potts & Whitton (1980). Elliott (1964 : 569) has described possible heterocysts in the Triassic form Zonotrichites lissaviensis Bornemann (1887). However, heterocysts and akinetes are not normally distinguishable in fossil material. H. M. Johnson's (1966 : pi. 6) example of the former in Girvanella may be neomorphic spar. In fossil algae, gross filament morphology is the principal tool available to the systematist. Thus whether or not a particular characteristic is useful should be decided on the basis of its biological nature. If a size or shape variation can be shown to have some direct relationship to the genotype of the organism, it may be of great value. Indeed, it is on this that identification by standard culture is based. If all the organisms tested give the same response under the same conditions, the assumption that they have the same genotype may be justified. Point 1 in the list of Girvanella 's generic characteristics (p. 88), is of descriptive value only. Most cyanophytes are microscopic 1 . Unless the calcification is incomplete, the struc- ture is bound to be tubular (2), and no noncalcareous, mineralized forms have yet been reported (3). Some modern cyanophytes can approximate rectilinear growth (4), but it is not a consistent feature at generic level. It is also difficult to say when a curve is slight enough to be treated as incidental, except perhaps in a population study. Some modern genera are habitually spiral (Spirulina Turpin); others are spiralled at times (Oscillatoria Vaucher) and many are never more than sinuous (Rivularia Agardh.). The characteristic seems to be both environmentally and genetically controlled. Comparison with Oscillatoria suggests that occasional spiralled filaments may be acceptable, but that populations with habitually spiralled filaments such as G. problematica var. spiralis Lewis should be excluded from the genus. Since procaryote cells seem to be either circular or oval in cross section (5), their trichomes are bound to have this shape. However, the carbonate tubes which form around them are not trichomes, and their inner surfaces may have several controls. If the carbonate is an impregnation, or a combined impregnation and encrustation, its inner surface may conform to the surface of the trichome. It will then be circular in section. No such restriction applies to the inner surface of an encrustation, which will more or less follow the surface of the sheath. Many cyanophytes have smooth sheaths with circular cross sections. However, the sheath may be irregular, or it may have a sculpture such as spiral or annular ribbing. This is probably an environmentally-determined character, in part if not entirely. It may also reflect the fine structure of the sheath itself. Scytonema Agardh., for example, has a strongly fibrillar sheath (Singh 1954). Unless the nature of the calcifi- cation is known, therefore, it would seem best to describe the calcareous tube of Girvanella as approximately circular in section, and perhaps to give an acceptable range of variation. It has been suggested that the calcareous tube of Girvanella is an impregnated sheath (Seward 1898) but there is evidence for the presence of both impregnations and encrus- tations in the Stinchar Limestone girvanellids (Danielli 1977). It is possible that these should 'Some bacterial carbonates are visible to the naked eye and resemble porostromates, but their importance in the fossil record is not yet known. Fig. 4 Types of calcification in Recent blue-green algae. Several genera are shown, and grain size varies considerably, but the grain shapes are on the whole similar, (a), encrustation of equidimensional grains on a relatively smooth filament. Note the uncalcified background filaments. Bar = 1 5 urn. (b), impregnation, indicated by the granular appearance of the sheath surface. Bar = 2 urn. (c), embedding in a mass of mucus strands, c calcite grains, m mucus strands. Bar = 2 urn. (d), an empty but impregnated sheath which might be preserved as a porostromate fossil. Bar = 2 urn. (e), encrustation and impregnation of the same sheath, the two carbonates having similar textures. Bar= 1 um. (0, a rivulariacean calcite. The trichome and sheath lay in the hollow, becoming encrusted with equigranular calcite. Organic matter has been removed with 14% NaCIO in this case, exposing the micritic texture of the calcite. In the light microscope these tubes often appear unicrystalline owing to optical continuity of the grains. Bar = 50 um. 94 H. M. C. DANIELLI be separated into two genera, although clear support for such a split cannot at present be obtained from modern organisms. Rivularia, for example, can both encrust and impregnate at the same time, and the two often have inseparable textures. In addition, the criterion of the internal to external diameter ratio, used for example by Riding (1977) to distinguish between them, does not hold in many cases. The impregnated sheath may be very thick in proportion to the protoplast diameter, and encrustations may be thin. These variations seem due mainly to ecological factors. The cylindrical nature of the tubes (7) implies that the diameters are fairly constant along the tube axis, and that the structure is not branched. Some modern cyanophyte trichomes have a constant diameter (e.g. Phormidium Kuetz.), some taper (Rivularia), and some have pronounced constrictions at cross walls (Nostoc Vaucher). Whether or not this would be reflected in the carbonate again depends on the nature of the structure. A constricted tube would probably indicate that the filament it contained was also constricted, and would also be quite good evidence that the filament was simple, but a carbonate tube without constric- tions gives no evidence either way. Tapering would probably be reflected by the internal diameter of the tube, although not necessarily by the external diameter, if the tube were an impregnation. Since part of the cajbonate of Girvanella tubes may well have been an impregnation, tapering might be expected to be apparent. It is then possible to say that the trichomes of Girvanella were probably not tapered. The presence or absence of branching is used at generic level in the identification of modern cyanophytes (West & Fritsch 1927 : 454). It is not possible to separate false from true branching unless the trichomes are present, of course. Some porostromates, such as Ortonella Garwood 1914, have clear branching, but Girvanella is by implication an unbranched form (Dricot & Tsien 1977 : 232). Branching need not be frequent, so careful searches are necessary to establish its absence from any population. Some modern cyanophytes have thin strips of sheath lying between the cells of their trichomes (Nostoc piscinale Kuetz.). Other forms often have fragmented trichomes, with sheath separating the fragments at more or less regular intervals (Scytonema fulginosum Tilden). It is theoretically possible for the strips of sheath to become impregnated with carbonate, producing a septate tube which could be preserved (9). However, no example of a septate Girvanella has been described which will support careful study. Having scrutinized the original diagnosis of Girvanella above, it is clear that an emendation would be of some value. An attempt is offered below, taking these points into consideration. Systematics Kingdom PROCARYOTAE Buchanan et al., 1974 Division CYANOPHYTA Smith, 1938 Class uncertain Family POROSTROMATA Pia, 1927 Genus GIRVANELLA Nicholson & Etheridge, 1878 Fig. 5 The neotype of Girvanella, and some contrasting growths in the same thin section, BM(NH) V 34566. (a), the growth which Wood (1957) designated as neotype of G. problematica. Note the low density central region, with denser growth at the boundaries. Bar = 55 um. (b), a detail of (a), corresponding to Wood's (1957) figure, showing low density, moderate sinuosity, local parallelism and adherence. Bar= lOOum. (c), detail of (b), with a variety of cell space textures. Bar = 50 um. (d), part of the same growth showing variable sinuosity, density and parallelism. Bar = 100 um. (e), another growth in the same section, with parallel, adherent tubes; note the cement overgrowth (c). Bar= lOOum. (f), a third growth in the same slide, with thin sinuous tubes arranged more or less randomly. Bar = 200 um. THE FOSSIL ALGA GIRVANELLA 95 96 H. M. C. DANIELLI [= Argirvanellum Rothpletz, 1916; Batinevia Korde, 1966; Botominella Reitlinger, 1959; Fistulella Korde, 1973; Kenella Korde, 1973; Nicholsonia Korde, 1973; Siphonema Bornemann, 1886; StrephochetusSeely, 1 885; Stromatocerium Miller, 1882]. RANGE. Upper Proterozoic to Middle Cretaceous. DIAGNOSIS. Microscopic tubular encrustations and/or impregnations of sheaths of fila- mentous organisms; filaments arranged at random or prostrate, rarely vertically; filaments single or in growths of variable size, shape and density; orientation parallel to random; filaments unbranched and slightly to highly sinuous; cell space approximately circular in cross section, sometimes compressed or with slight irregularities; cell space usually occupied by cement spar but sometimes micritic, cement developing either by growth from grains in the micritic tube or independently, as druses or equidimensional particles; aseptate micritic calcite tube comprising prisms or equidimensional grains, or both, the prisms with their long axes arranged perpendicularly to the filament axis and in a single layer, sometimes with smaller intercalated prisms. NEOTYPE. The specimen shown on Fig. 5, after Wood (1957). Thin section kept by the Palaeontology Dept. of the British Museum (Natural History), London, reg. no. V 34566. HORIZON AND LOCALITY. Stinchar Limestone (Ordovician, Lower Caradocian Series), Tormitchell Quarry near Girvan, Ayrshire, Scotland. Specific subdivision As Wood (1957) found, specific subdivision of Girvanella is a difficult problem. He drew a graph of internal diameters and obtained a bimodal curve for the sample he took from the percentage frequency 25 - 20 - 15 - 10 - 5 _ 1-2 '3-4 '5-6 '7-8^-lo'n- 'l4- l 16- h.8- '20- ! 22- ' 24 13 15 17 19 21 23 Class intervals, Urn Fig. 6 Internal diameters of Girvanella tubes from the Stinchar Limestone. Plain line, Aldons Quarry (Danielli herein, n=1060). Pecked line, Benan Burn and Tormitchell Quarry (after Wood 1957, with permission; n = 633). THE FOSSIL ALGA GIRVANELLA 97 Stinchar Limestone. The external diameters gave a less useful curve. Internal diameters from the sample collected for the present study gave a unimodal plot with a slightly longer range (Fig. 6). Wood's material came from several localities, and thus could have been derived from environments which differed in biological or biochemical factors but which are indistinguishable by modern geological methods. If this is so, the algae are likely to have been different in each. Even if the same species were present in all, several ecophenes may have been represented. The Aldons Quarry sample came from a single outcrop and is not therefore strictly comparable with Wood's. Mamet & Roux (1975 : 137) surveyed the Carboniferous and Devonian species of Girvanella, and condensed the various taxa into four species. Their method of measuring from the published figures rather than the type material is questionable, but the results are of some interest. They based the four species on a graph of tube thickness plotted against internal diameter. Unfortunately, girvanellids in the Stinchar Limestone at Aldons Quarry do not conform to Mamet & Roux's method of subdivision. Nor does the method allow for the presence of a fifth species. The internal diameter and tube thickness of the Aldons sample are similar to those of Mamet & Roux, but the measurements are distributed over most of the graph and do not fall into clusters. In a case like this reappraisal of the genus as a whole is necessary. One approach might be to reanalyse the previously-designated types, and reorganize them. However, as pointed out by Raup & Stanley (1971 : 177), many types do not adequately represent their species. Skevington (1973 : 43) has come to a similar conclusion in the case of graptolites, and Hughes, Drewry & Laing (1979 : 5 1 5) have regretted the type and synonomy arrangement of taxa under the rules of botanical nomenclature for a similar reason. Since a complete reassessment of the species ofGirvanella would require a work of far greater length than this, and fundamental studies in greater depth, no more is intended here than to suggest a possible approach using the definition of the type species problematica. The species is taken as described by Nicholson & Etheridge (1878). The existing classification ofGirvanella species has been investigated, but the value of species inquirendae in a systematic re-evaluation is open to question. It will be shown that some new basis for subdivision of the genus is needed, if the classification is to reflect the biology of the organisms. The characteristics used by Nicholson & Etheridge (1878) give a good description of the fossil, and may be summarized as follows: (1) sheath diameter; (2) tapering; (3) parallelism; (4) density; (5) sinuosity; (6) growth shape; (7) growth size. No tapering species have been described (2). Indeed, since the generic diagnosis implies that the tube diameter is constant, the question might be considered out of place here. The density of the growths (4) has also been discussed under generic characteristics. Points 3, 5, and 6 are concerned with the spatial relations between tubes. Authors delineating species have often described the sinuosity of tubes, and sometimes their degree of parallelism. Unfortunately no attempt has been made to define these numerically. Much the same can be said of growth shape and size. There is, however, some justification for this if one agrees with Maslov's (1949) view that they are environmentally-determined characteris- tics. This is likely to be the case, in view of the phenotypic plasticity of Recent filamentous cyanophytes (Desikachary 1970, Rippkaef a/. 1979). We are left with point 1, sheath diameter. It has been used by many workers, in both definition and identification of species. The diameters of many of the species listed in Table 2, pp. 82-4, are compared in Fig. 7. The first problem encountered is the meaning of 'diameter'. As shown in Fig. 1 (p. 80) there are two diameters cd, ef to any cross section. In Girvanella they may differ by a factor of two or more. Not all authors ofGirvanella species have explained which diameter they were giving. The diameters shown have considerable overlap, and the ranges differ in length. Measure- ment of a single tube will not allow assignment of the tube to one species. If a number of tube diameters are known for a sample, questions of sample size and statistical significance are raised. Wood (1957) measured about 600 tubes, and found the range 5 urn to 22 H.ITI for the Stinchar Limestone. Over 1000 were used in the present study, and a slightly wider range was 98 H. M.C. DANIELLI obtained (Fig. 6). This might have been extended even further if 1500 tubes had been measured. Green (unpublished, 1959) conducted a statistical study of some Silurian girvanellids, but the conclusions were published without the basic data (H. M. Johnson 1966). The external diameter is a function of the internal diameter and tube thickness. Its signifi- cance has been discussed by Wood (1957 : 26), who considered it to be unreliable because of its dependence on environmental factors. However, Mamet & Roux (1975) have shown that the relationship between tube thickness and internal diameter may be useful in classification. Girvanellids in the sample taken from the Stinchar Limestone at Aldons Quarry agree with the generic definition of Nicholson & Etheridge (1878) in most respects. However, growth densities vary from 10% to almost 100% (Fig. 3). Strictly, the high density growths are not Girvanella, but in every other way the growths fit the definition. Since growth density is probably an environmentally-determined characteristic, there seems to be no reason for separating these forms as a different genus or species from the lower density forms. The internal diameter from the Aldons Quarry sample ranges between 4 urn and 22 um (the area indicated in Fig. 7). Some 26 species are contained in this interval, and it is overlapped by 10 others. It is possible that, if the sample size were larger, some of these species would lie completely within the Aldons range. The external diameter of each tube was measured, and tube thickness was determined as half the difference between the two diameters. A range of 0'5 um to 22 um was found, which includes all those species whose definitions include a value for tube thickness. To illustrate the variability of the growths as well as that of individual tubes, Fig. 8 gives the internal diameter and tube thickness ranges found in some particular growths. While examples of the extreme values are rare, objections of a statistical nature can be made to the exclusion of specimens on grounds of rarity. A study of Girvanella populations throughout its fossil record is evidently the next stage in this reassessment. The example of Rippka et al. (1979), in leaving this to the future, is followed here, but some aspects of the organism which might be considered are the detailed microscopic morphology in relation to the ecology of the organism, the nature of the carbonate tube and its relation with the enclosed trichomes, and the ultrastructure of the carbonate tube in relation to other Porostromates and to modern calcified cyanophytes. The work of Drouet is applicable to the first of these suggestions. Though now his classifi- cations are generally rejected, he did demonstrate the dependence of cyanophyte morphology on evironmental factors, describing ecological varients or ecophenes of, for example, Schizothrix calcicola (Ag.) Gom. (Drouet 1963). These varieties have essentially the same genotype. Girvanella comes from a wide geographical and temporal range, as well as a wide range of ecologies. There must be many genotypes represented, so that the ecophene approach will only be tenable in some circumstances. It might be used with advan- tage within a single outcrop, and preferably at a single horizon, for determining local varia- bility within a population. Huxley (1938) defines a cline as any variational trend in space. If the end members of a cline are separated from each other, they may develop into distinct species. In the case of fossil material especially, the full cline may not be preserved and the end-members may .be taken for different taxa. Thus genotypic variation occurs along the cline. This picture, suggested by Dr M. A. Edington (personal communication), fits Girvanella a little better than does that of ecophenes. Fig. 7 The diameters of Girvanella species. Ranges are given in chronological order, and keyed to Table 2, p. 82-4. Most authors give internal diameters, but some do not specify which they provide. Since the sparry cell space is often easy to see, it is likely that most of these undefined diameters are in fact internal. The range of values from the Aldons Quarry sample is indicated by two vertical broken lines. The variability of the lengths, overlapping positions, and lack of frequency distributions of the individual species makes the diameter of low value as a taxonomic criterion at present. 52 \-\ 51 50 49 48 46 53 M.Crel L. Crel 47 , Jurassi 42 40 HH 3ft HI , 41 Permia 1 1 39 37 ' 35 34 33 31 J 29 > 28 . 27 Carbonil 1 1 26 _ H o 5 Devonia 24 v H- < 21 H- ' 1 16 '"IT 14 Siluria 11 8 H L H 6 h-H 5 * 1 3 2 1 1 ' 4 Ordovicia Cambrian// 1 1 ! 1 1 i i "T" "TT" 22 Diameter, 100 H. M. C. DANIELLI Frequency i i i ^^ i r^ T^^ r 3-4 5-6 7-8 9-10 11- 13- 15- 17- 19- 21 12 14 16 18 20 Class intervals, pm ^ a) Internal diameter; a. N=1O, b. N=1O. Frequency 0.9 I 1 I 2 l 3 ' 4 ' 5 ' 6 ' 7 ' 8 ' 9 ' 10 ' 11 ' 12 ' 13 Class intervals, b) Tube thickness; a. N=7, b. N=1O. Fig. 8 Internal diameters (a) and tube thicknesses (b) for Girvanella growths in the Stinchar Limestone at Aldons Quarry. Plain lines, maximum range for single growth. Pecked lines, minimum range for single growth. The study of clines requires sampling over substantial areas, since clinal variation is a geographical phenomenon. Therefore an investigation of Girvanella clines would require collection along the same horizon, over some kilometres if possible. Since no study of this kind has yet been published, it is not possible at present to decide whether Girvanella does or does not form clines. The second suggestion for future studies concerns the carbonate tube. Its relationship to the original sheath of the organisms greatly affects the acceptable range of variation of characters such as the internal diameter. Unless such questions can be answered, these ranges must remain arbitrary. Turning to the third suggestion, there is in fact some reason to doubt the value of ultra- structure as a taxonomic tool in the case of cyanophyte carbonates. Recent organisms show no clear relationship between carbonate fine structure and species or even genus. In addition THE FOSSIL ALGA GIR VANELLA 1 1 the fossil porostromate Rothpletzella gotlandicum Wood 1948 appears to have a similar ultrastructure to that of Girvanella (Fig. 2, p. 89). Kobluk & Risk (1977: 1077) have reported girvanellids with similar textures to those of the Stinchar Limestone forms. This apparent stability of the carbonate texture from place to place is interesting, but not very promising for specific subdivision. The importance of population studies in work of this kind cannot be over-emphasized, especially with organisms that show such degrees of variation. In these cases a range of variation should be given for every characteristic. Conclusions The morphology of Girvanella has been subject to a good deal of misinterpretation. The genus has been discussed here in terms of modern cyanophytes, and the limitations of the fossil material have been outlined. Direct comparison can only be made at the level of fila- ments, since there is at present no unequivocal evidence concerning trichome shape, or number per filament. The generic diagnosis of Nicholson & Etheridge (1878) has been emended to take account of this discussion, and the specific subdivision of the genus has been considered briefly. Studies involving sampling over wide areas, at a single horizon, are considered necessary so that the variation shown by the fossil can be investigated. The position of the genus in relation to the groups of modern filamentous organisms is at present indeterminable, the only guide being the lack of branching shown by Girvanella. Russian scientists such as Kulik (1973 : 39) have defined species as members of the Hormagoneae and the assignment is very possibly correct, but care should be used in such determinations. The remarks made by Edhorn (1979) concerning the mobility of the fila- ments appear to present an over-extension of the evidence available from the fossil material and would be difficult to test. However, comparisons like that of Edhorn, between the growth habits of Girvanella and those of Recent organisms, may shed some light on the ecology of the fossil if applied with caution. It is at present possible to say very little about the organisms concerned in the formation of Girvanella tubes. Population studies coupled with ecological work on the communities associated, and with sedimentology, may be useful in this connection. Since cyanophytes are important members of many modern communities, and are major sediment-producers at the present time, fossil forms might be expected to be of considerable palaeoecological value. Acknowledgements Dr Isles Strachan (University of Birmingham) and Dr M. A. Edington (University College, Cardiff) gave invaluable advice and encouragement to this study. I am grateful to Dr G. F. Elliott (British Museum (Natural History)), Dr B. A. Whitton (University of Durham), Professor C. L. V. Monty (Universite de Liege), Dr R. E. Riding, Dr K. Benson-Evans, Mr G. Hillman and Professor W. T. Dean (all of University College, Cardiff) for their help with various aspects of the work. Thanks are also offered to Professor D. H. Griffiths (University of Birmingham) and Professor M. Brooks (University College, Cardiff) for facilities in their departments. Thin sections were made by Mr R. Flynn (University of Birmingham). Mr B. S. Harris (University of Birmingham) gave instruction in scanning electron microscopy, and Mr R. W. Jones and Mrs C. Winters (University College, Cardiff) gave additional help. The S.E.M. at Birmingham was owned by N.E.R.C. (GR3/2555) and that in Cardiff by the Department of Metallurgy and Materials Science; I am grateful for the opportunity to use these. The manuscript was typed by Mrs P. Thomas and Mrs C. Carrel (University College, Cardiff). References Banks, M. R. & Johnson, J. H. 1957. Maclurites and Girvanella in the Gordon River Limestone (Ordovician) of Tasmania. J. Paleont., Menasha, 31 : 632-640. 102 H. M.C. DANIELLI Bilan, W. & Golonka, J. 1973. Upper Devonian/Lower Carboniferous calcareous algae from the Fore- Carpathian Depression. Acta geol. poi, Warsaw, 23(1) : 149-168,8 pis. Bornemann, J. G. 1886. Die versteinerungen des Cambrischen Schichtensystems der Insel Sardinien nebst vergleichenden Untersuchungen iiber analoge Vorkommnisse aus andern Landern. Nova Acta Acad. Caesar. Leop. Carol., Halle, 51 : 1-148, pis 1-33. (See Hinde 1887). 1 887. Geologische Algenstudien. Jb. preuss. geol. Landesanst. BergAkad., Berlin, 1886 : 1 1 6-1 34. Brown, A. 1894. On the structure and affinities of the genus Solenopora, together with descriptions of new species. Geol. Mag., London, (dec. 4) 1 : 149-1 5 1 . Chapman, F. 1907. On the relationship of the genus Girvanella, and its occurrence in the Silurian Limestones of Victoria. Rep. Australas. Ass. Advmt Sci., Adelaide, 11 : 377-386, 2 pis. Colin, J. P. & Vachard, D. 1977. Une 'Girvanelle' dulcaquicole du Cenomanien du sud-ouest de la France. Rev. Palaeobot. Palynol, Amsterdam, 23 (4) : 293-302. Conil, R. & Lys, M. 1964. Materiaux pour 1'etude micropaleontologique du Dinantien de la Belgique et la France (Avesnois). Mem. Inst. geol. Univ. Louvain 23. 290 pp., 40pls. Danielli, H. M. C. (1977). The fine structure and taxonomy of Lower Palaeozoic forms of the calcareous alga Girvanella. M.Sc. thesis, Univ. of Birmingham (unpubl.). (in prep). Calcified Cyanophyta from freshwater environments. Dawson, W. 1896. Note on Cryptozoon and other ancient fossils. Can. Rec. Sci., Montreal, 7:203-219. Desikachary, T. V. 1970. Taxonomy of blue-green algae problems and prospects. Schweiz. Z. Hydroi, Basel, 32 : 490-494. Dragastan, O. 1975. Upper Jurassic and Lower Cretaceous microfacies from the Bicaz Valley Basin (East Carpathians). Mem. Inst. Geol. Geophys. Bucarest 21. 87 pp., 103 pis. Dricot, E. & Tsien, H. H. 1977. Le nom du genre Rothpletzella Wood 1948 (algue calcaire paleozoique) est legitime et correct. Mem. Inst. geol. Univ. Louvain 29 : 23 1-240. Drouet, F. 1962. Gomont's ecophenes of the blue-green alga, Microcoleus vaginatus (Oscillatoriaceae). Proc. Acad. nat. Sci. Philad. 114 (6) : 191-205. 1963. Ecophenes of Schizothrix calcicola (Oscillatoriaceae). Proc. Acad. nat. Sci. Philad. 115:261-281. 1968. A revision of the classification of the Oscillatoriaceae. Monogr. Acad. nat. Sci. Philad. 15.370pp. Edhorn, A.-S. 1979. Girvanella in the 'Button Algae' horizon of the Forteau Formation (Lower Cambrian), western Newfoundland. Bull. Cent. Rech. Explor.-Prod. Elf-Aquitaine, Pau, 3 (2): 557-567. Elliott, G. F. 1964. Zonotrichites (Calcareous Algae) from the Arabian Triassic. Eclog. geol. Helv., Basel, 57 (2) : 568-570, pis I-II. Flajs, G. 1977. Die Ultrastrukturen des Kalkalgenskeletts. Palaeontographica, Stuttgart, (B) 160:69-128. Fliigel, E. & Fliigel-Kahler, E. 1980. Algen aus den Kalken der Trogkofel-Schichten der Karnischen Alpen. In Fliigel, E. (ed.), Die Trogkofel-Stufe im Unterperm der Karnischen Alpen. Carinthia II, Klagenfurt, 36 : 1 13-182, 1 1 pis in text. Fremy, P. & Dangeard, L. 1935. Sur la position systematique des Girvanelles. Bull. Soc. linn. Normandie, Caen, (ser. 8) 8 : 101-1 12, 2 pis. Garwood, E. J. 1914. Some new rock-building organisms from the Lower Carboniferous beds of Westmorland. Geol. Mag., London, (dec. 4) 1 : 265-271. 193 1 . The Tuedian Beds of northern Cumberland and Roxburghshire, east of the Liddel Water. Q. J I geol. Soc. Lond. 87 : 97-1 57. Geitler, L. 1930-32. Cyanophyceae (Blaualgen). In Rabenhorst, L. (ed.), Kryptogamen- Flora 14. 1 196 pp. Leipzig. Gleason, P. J. (1972). The origin, sedimentation and stratigraphy of a calcite mud located in the southern fresh-water Everglades. Ph.D. dissertation, Pennsylvania State Univ. (unpubl.). Gollesstaneh, A. 1974. The biostratigraphy of the 'Khami-group' and the Jurassic-Cretaceous boundary in Pars province (southern Iran). Bull. Bur. Rech. geol. min. Paris (2e ser., sect. IV) 3-1974 : 165-197, 3 charts [In French and English]. Golubic, S. 1973. The Relationship between Blue-Green Algae and Carbonate Deposits. In Carr, N. G. & Whitton, B. A. (eds), The Biology of Blue-Green Algae : 434^72. Oxford (Botanical Monographs 9). Green, H. M. (1959). Calcareous Algae of the Silurian of the Welsh Border. Ph.D. Thesis, Univ. of Wales (unpubl.). THE FOSSIL ALGA GIR VAN ELLA \ 03 Giivenc, T. (1965). Etude Stratigraphique el Micropaleontologique du Carbonijere et du Permien des Taurus Occidentaux dans I'Arriere Pays d'Alanya (Turquie). Ph.D. thesis, Acad. of Science, Paris (unpubl.). Hinde, G. J. 1887. [Review of] The Fossils of the Cambrian Strata of the Island of Sardinia, with a Comparative Study of similar forms from other countries, by Dr J. G. Bornemann. 1st part. Geol Mag., London, (dec. 3)4 : 226-229. Hpeg, O. A. 1932. Ordovician Algae from the Trondheim area. Skr. norske Vidensk-Akad. mat.-nat. Kl., Oslo, 1932 (4) : 63-96, pis 1-11. I Ionian, W. 1972. Unter- und tief-mittelpermische Kalkalgen aus den Rattendorfer Schichten, dem Trogkofel-Kalk und dem Tressdorfer Kalk der Karnischen Alpen (Osterreich). Senckenberg. leth., Frankfurt a.M, 53: 135-313, 10 pis. Hughes, N. F., Drewry, G. E. & Laing, J. F. 1979. Barremian earliest angiosperm pollen. Palaeontology, London, 22 (3) : 5 1 3-535. Huxley, J. S. 1938. Clines: an auxiliary taxonomic principle. Nature, Land. 142 : 2 1 9-220. Johnson, H. M. 1966. Silurian Girvanella from the Welsh Borderland. Palaeontology, London, 9(1): 48-63, pis 6-12. Johnson, J. H. 1946a. Mississippian algal limestones from the vicinity of St Louis, Missouri. J. Paleont., Menasha, 20 : 1 66-1 7 1 . 19466. Lime-secreting algae from the Pennsylvanian and Permian of Kansas. Bull. geol. Soc. Am., New York, 57: 1087-1119. 1950. A Permian Algal-Foraminiferal Consortium from west Texas. J. Paleont., Menasha, 24:61-62. 1952. Studies of Organic Limestones and Limestone-building Organisms. Colo. Sch. Mines Q., Golden, 47 (2) : 29-56. 1961. Limestone-building Algae and Algal Limestones. 297 pp., 139 pis in text. Golden, Col. (Colorado School of Mines). & H0eg, O. A. 1961. Studies of Ordovician Algae. Colo. Sch. Mines Q., Golden, 56 (2). vii + 120 pp. ,39 pis in text. Kobluk, D. R. & Risk, M. J. 1977. Micritization and carbonate-grain binding by endolithic algae. Bull. Am. Ass. Petrol. Geol., Stillwater, Okla., 61 (7) : 1069-1082. Korde, K. B. 1966. Novye materialy k sistematike i evolutsiya krasnykh vodoroslei rannego paleozoya. Dokl. Akad. Nauk SSSR, Leningrad, 166 (6) : 1440-1442, pi. 1 . 1973. [Cambrian Algae]. TrudJ paleont. Inst., Moscow, 139. 349 pp. [In Russian]. Krumbein, W. E. & Potts, M. 1978. Girvanella-like structures formed by Plectonema gloeophilum Borzi (Cyanophyta) from the Borrego Desert in southern California. Geomicrobiol. J., New York, 1 (2): 175-181. Kulik, E. L. 1973. Opisanie organicheskikh ostatkov. In Einov, O. L. (ed.), [Stratigraphy and fauna of the coal deposits of the Shartim River (southern Urals)} : 39^8, pis 1-4. Lvov (Ural. Geol. Upravlenie). [In Russian]. Lewin, J. C. 1962. Calcification. In Lewin, R. A. (ed.), Physiology and Biochemistry of Algae : 457-465. London. Lewis, H. P. 1942. On Girvanella in the 'Shumardia Limestone' of Levis, Quebec. Ann. Mag. nat. Hist., London, (ser. 11)9: 49-55, pi. 1 . Luchinina, V. A. 1974. See Nitikin et al. Mamet, B. & Roux, A. 1975. Algues Devoniennes et Carboniferes de la Tethys Occidental. Revue Micropaleont., Paris, 18 (3) : 134-187. Maslov, V. P. 1937. [On the distribution of calcareous algae in east Siberia]. Problemy Paleont., Moscow, 2-3 : 327-348. [In Russian, Engl. summary]. 1949. Vodorosl Girvanella, ee ekologiya i stratigraficheskoe znachenie. Byull. Mosk. Obshch. Ispyt. Prir. (Otdel Geol.) 24 (2) : 89-100. [In Russian]. 1956. [Fossil calcareous algae of the USSR]. Trudy Inst. geol. Nauk Mosk. 160. 301 pp., 86 pis. [In Russian]. Miller, S. A. 1882. Description of two new genera and eight new species of fossils from the Hudson River Group, with remarks upon others. J. Cincinn. Soc. nat. Hist. 5(1): 34-44, pis 1-2. Monty, C. L. V. 1977. Evolving concepts on the nature and the ecological significance of Stromatolites: a review. In Fliigel, E. (ed.), Fossil Algae : 1 5-36. Berlin. & Hardie, L. A. 1976. The geological significance of the freshwater blue-green algal marsh. In Walter, M. R. (ed.), Stromatolites. Devs Sedimentol., Amsterdam, 20 : 447^78. 104 H. M. C. DANIELLI Nicholson, H. A. 1 888. On certain anomalous organisms which are concerned in the formation of some of the Palaeozoic Limestones. Geol. Mag., London, (dec. 3)5 : 15-24. & Etheridge, R., jr. 1878-80. A monograph of the Silurian fossils of the Girvan District . . ., 1. 341 pp. ,24 pis. London. Nitikin, I. F., Gnilovskaya, M. B., Zhuralevna, I. T., Luchinina, V. A. & Mayagkova, E. I. 1974. [Anderken bioherm range and the history of its origin]. Trudy Inst. Geol. Geofiz. sib. Old., Novosibirsk, 84 : 1 22-1 59, pis 23-29. [In Russian]. Parks, W. A. 1933. New species of stromatoporoids, sponges, and corals from the Silurian strata of Baie des Chaleurs. Univ. Toronto Stud. geol. Ser. 33. 40 pp., 8 pis. Paul, H. 1937. Unterkarbonische Kalkalgen und Calcisphaeren Deutschlands. Jb. preuss. geol. Landesanst. BergAkad., Berlin, 58 : 276-28 1 , pi. 2 1 . 1938. Girvanella grabaui sp. nov. aus dem unterpermischen Chuanshan-Kalk Siidchinas. Bull. geol. Soc. China, Peking, 18 : 21 1-214. Pentecost, A. 1978. Blue-green algae and freshwater carbonate deposits. Proc. R. Soc., London, B 200:43-61. Pia, J. 1927. Thallophyta. In Hirmer, M. (ed.), Handbuch der Paldobotanik 1 : 1-136. Munich & Berlin. 1932. Algenkalkknollen aus dem Russischen Devon. Izv. Akad. Nauk. SSSR, Leningrad, (Otdel Mat. estestv. Nauk) 1932 (9) : 1345-1360, 1 pi. 1937. Die wichtigsten Kalkalgen des Jungpalaozoikums und ihre geologische Bedeutung. C.r. Congr. Av. El. Strat. Geol. Carbon., 2e Heerlen, 2 : 765-856, pis 85-97. Playford, P. E., Cockbain, A. E., Druce, E. C. & Wray, J. L. 1976. Devonian stromatolites from the Canning Basin, Western Australia. In Walter, M. R. (ed.), Stromatolites. Devs Sedimentol., Amsterdam, 20 : 543-563. Pollock, J. B. 1918. Blue-green algae as agents in the deposition of marl in Michigan lakes. Rep. Mich. Acad. Sci., Lansing, 20 : 247-260. Potts, M. & Whitton, B. A. 1980. Vegetation of the intertidal zone of the lagoon of Aldabra, with particular reference to the photosynthetic prokaryotic communities. Proc. R. Soc., London, B 208: 13-55, pis 1-2. Raup, D. M. & Stanley, S. M. 1971. Principles of Paleontology. 388 pp. San Francisco. Reid, C. 1 893. See Wethered, E. Reitlinger, E. A. 1959. Atlas microscopicheskikh organicheskikh ostatkov i problematiki drevnikh tolshch Sibiri. Trudy geol. Inst. Leningrad 25. 59 pp., 22 pis. [In Russian]. Rhumbler, L. 1895. Entwurf eines natiirlichen Systems der Thalamophoren. Nachr. Ges. Wiss. Gottingen (Math.-phys. Kl.) 1895 (1) : 51-98. Riding, R. E. 1975. Girvanella and other algae as depth indicators. Lethaia, Oslo, 8: 173-179. 1977. Calcified Plectonema (blue-green algae), a Recent example of Girvanella from Aldabra Atoll. Palaeontology, London, 20 (1) : 33^6. Rippka, R., Derulles, J., Waterbury, J. B., Herdman, M. & Stanier, R. Y. 1979. The Cyanobacteria. J. gen. Microbiol., Cambridge, 111 : 1-61. Romanes, M. F. 1916. Note on an algal limestone from Angola. Trans. R. Soc. Edinb. 51 : 581-584, Ipt. Rothpletz, A. 1891. Fossile Kalkalgen aus den Familien der Codiaceen und der Corallineen. Z. dt. geol. Ges., Berlin, 43: 296-322. 1916. Uber die systematische Deutung und die stratigraphische Stellung der altesten Versteiner- ungen Europas und Nordamerikas mit besonderer Beriicksichtigung der Cryptozoen und Oolithe. II. Teil: Uber Cryptozoon, Eozoon, und Atikokania. Abh. bayer. Akad. Wiss., Munich, 28 (4) : 1-91, 8 pis. Schafer, A. & Stapf, K. R. 1978. Permian Saar-Nahe Basin and Recent Lake Constance (Germany): two environments of lacustrine algal carbonates. Spec. Publs int. Ass. Sedimentol., Oxford, 2:83-107. Seely, H. M. 1885. A new genus of Chazy sponges, Streptochetus. Am. J. Sci., New Haven, (ser. 3) 30 : 355-357. 1902. Some sponges of the Chazy Formation. Rep. St. Geol. Miner. Ind. Geol. Vt, Burlington, 3: 15 1-1 61, pis 56-59. Seward, A. C. 1 898. Fossil Plants 1. 452 pp., figs. Cambridge. Shuiskii, V. P. 1973. [Calcareous reef-forming algae of the Lower Devonian in the Urals]. 155 pp., 34 pis. Moscow (Nauka, Ural. Sci. Cent. Inst. Geol. Geochem. A.N. Zavaritsky). [In Russian]. THE FOSSIL ALGA GIRVANELLA 105 Singh, R. N. 1954. Electron micrographs of the mucilage of the blue-green algae. In Vandermeersche, G. (ed.), Rapp. Eur. Congr. T.E.M., Ghent: 63-67. Skevington, D. 1973. Graptolite fauna of the Great Paxton borehole, Huntingdonshire. Bull. geol. Surv. Gt Br., London, 43 : 47-57, pis 8-9. Smith, G. M. 1938. Cryptogamic Botany 1. 545 pp. New York. (2nd edn, 1955). Tilden, J. E. 1910. Minnesota Algae 1. 319 pp., 20 pis. Minneapolis (Report of the Survey. Botanical Series 8). West, G. S. & Fritsch, F. E. 1927. A Treatise on the British Freshwater Algae, 2nd edn. 534pp. Cambridge. (Reprint 1967 (Cramer, J., ed.) Biblthca Phycol. 3). Wethered, E. 1889. On the microscopic structure of the Jurassic pisolite. Geol. Mag., London, (dec. 3) 6: 196-201, pi. 6. 1 890. On the Occurrence of the Genus Girvanella in Oolitic Rocks, and remarks on Oolitic Structure. Q. Jl geol. Soc. Lond. 46 : 270-281. 1893. On the Microscopic Structure of the Wenlock Limestone, with Remarks on the Formation generally. Q. J I geol. Soc. Lond. 49 : 236-248. Williams, A. 1962. The Barr and Lower Ardmillan Series (Caradoc) of the Girvan District, south-west Ayrshire, with descriptions of the Brachiopoda. Mem. geol. Soc. Lond. 3. 267 pp., 25 pis. Whitton, B. A., Holmes, T. N. H. & Sinclair, C. 1978. A coded list of 1000 freshwater algae of the British Isles. 335 pp. Reading (Dept. Environment Water Data Unit, Water Archive Manual 2). Wood, A. 1941. The Lower Carboniferous calcareous algae Mitcheldeania Wethered and Garwoodia, gen. nov. Proc. Geol. Ass., London 52 (3) : 216-226, pis 13-15. 1948. 'Sphaerocodiuni*, a misinterpreted fossil from the Wenlock Limestone. Proc. Geol. Ass., London, 59 : 9-22. 1957. The type-species of the Genus Girvanella (Calcareous algae). Palaeontology, London, 1 : 22-28, pis 5-6. 1963. The British Carboniferous species of Girvanella (Calcareous algae). Palaeontology, London, 6 -.264-273, pis 38^0. Yabe, H. 1912. Ueber einige Gesteinbildende Kalkalgen von Japan und China. Sci. Rep. Tohoku Univ., Sendai,(ser.2)l(l): 1-8. & Osaki, K.-E. 1930. Girvanella in the Lower Cambrian of South Manchuria. Sci. rep. Tohoku Univ., Sendai, (ser. 2) 14 : 79-83, pi. 25. & Toyama, S. 1928. On some rock-forming algae from the younger Mesozoic of Japan. Sci. rep. Tohoku Univ., Sendai, (ser. 2) 12 (1) : 141-152, pis 18-23. Zanardini, G. 1843. Saggio di classificazione naturale delle Ficee. 65 pp., 1 pi. col. Venice. (Review, 1 844, in Bot. Ztg., Berlin, 2 (22) : 404-408). Index [Principal references are in bold type; an asterisk (*) denotes a figure.] akinetes 92 Albian 84 Aldons Quarry 80, 8 1 *, 9 1 , 96*, 97-8, 100 algae, blue-green 90, 93* green 86, 90 Argirvanellum 96 bacteria 88, 92 f/n Ballantrae Ophiolites 80-1 Bathonian 84 Batinevia86,96 bayankolica 82 ramosa 82 Benan Burn 96* Benan Conglomerate 8 1 Botominella 86, 96 lineata 82 calcification 88, 90, 92, 94; Cyanophyta, Recent. Cambrian 82, 84, 86,99 Caradocian79,83,96 Carboniferous 83-4, 97, 99 Chaetophora 90 characteristics, defining 88, 97 Chlorophycophyta 86 Chlorophyta79,90 clines98, 100 Codiacea 86 Cretaceous 84-5, 96, 99 see 106 Cyanophyta 79, 86, 88, 94 Recent 8 8-94, 9 7-8, 100-1 Dasycladacea 86 Devonian 83, 97, 99 ecophenes 98 electron microscopes, scanning 85 Famenian 83 fine structure 88, 89*, 90 Fistulella 86, 96 decipiens 82 sanashtykgolica 82 foraminifera 86 Girvan, Ayrshire 80, 96; see Aldons Quarry, Tormitchell Quarry Girvanella passim; 87-8 description 87 diagnosis 86, 96 diagrams 80 diameters 96*, 97-8, 99*, 100* fine structure 88, 89*, 90 growth habits 91* neotype, see G. problematica population studies 101 samples 8 1 species list 82-4 specific subdivision 96-101 systematics 94, 96 thickness of tube 100* ultrastructure90,98, 100-1 ambigua 82 amplefurcata 83 antiqua 82 atratus82,81 bayankolica 82 bornemanni 83 brainierdi&2, 87 catenoides 84 composita 82 confer ta%l,%l convoluta 82, 87 decipiens 82 densa 83 distans 83 due ii 83 chuvashovi 83-4 kasakiensis 83 effusa 83, 87 embergeri 84 fragila 83, 87 glomerata 82 grabaui 84 graiw/is82,87 incompta 83, 87 incrustans 82-4 /MCI/ 84 H. M.C. DANIELLI intermedia 84 johnsoni 84 jurassica 84 kordei 84 liebusi 83 lineata 82 magna 84 goksuensis 84 yatani 84 manchurica 82, 87 maplewoodensis 84 was/o v/ 8 3 mafia 83, 87 mexicana&2, 87 minima 84 minuta 84, 86 moniliformis 87; see problematica moorei 84 nicholsoni 83 ornata 82 ottonosia 83 palustris 84 permica 84 pisolitica 84 problematica 82, 86-7, 97 holotype 87 neotype 87-8, 95*, 96 lumbricalis&2, 87 moniliformis 82, 87 spiralis82,92 typical, SI pr0/uaz83,87 /a 83, 87 ramosa 82-3, 87 richmondense 82 sanashtykgolica 82 sarmenta 83, 87 shirazica 84 5i'6i>icfl 82, 87 silesiaca 84 silurianaSl, 87 sinensis 84 stamenia 83 subparallela 84 tasmaniensis82 texana 84, 88 tosaensis 84 wether edii 83 zimmermanni&3 Girvanellinae 86 graptolites 97 heterocysts 92 historical review 85-7 Hormagoneae 101 $l THE FOSSIL ALGA GIRVANELLA 107 Iran 84 Jurassic 82, 84, 99 Kenella 86, 96 ornata 82 Liassic 84 Llandeilo 82 Llandovery 83 Llanvirn 82 Ludlovian 83 Mesozoic 84 methods of study 81,85 morphological terms 80 Namurian 84 Nicholsonia 86-7, 96 composita 82 glomerata 82 grandis 82 involutans 82 Nostoc 94 piscinale 94 ophiolites, see Ballantrae Ordovician82,86,96,99 Ortonella 94 Oscillatoria 92 Oscillatoriaceae 90 Oxfordian 84 Palaeozoic 8 7 Permian 83^, 99 Phormidium 94 plants 86; see algae Plectonema gloeophilum 87 Porostromata 79, 86, 92 f/n, 93*, 94, 101 Procaryotae, procaryotes 92, 94 Proterozoic82,85-6,96 Protobangiophyceae 86 Rhabdamminidae 86 Rhizammina algaeformis 86 Rhodophyta 79, 86-7 Rivularia92,94 Rivulariaceae 90, 93* Rothpletzella 79, 86, 89*, 90 gotlandicum 101 Sardinia 86 Schizothrix calcicola 98 Scytonema 92 fuliginosum 94 myochrous 90 sections 8 5 silicification 88 Silurian 82-3, 99 Siphonae 86 Siphonema 86, 96 incrustans 82-4 Sphaerocodium 86 bornemanni 83 incrustans 83 zimmermanni 83 Spirulina 92 sponge 86 Stinchar Limestone 79-80, 81*, 87, 91 96-8, 100-1 Strephochetus 86, 96 atratus 82 brainierdi 82 ocellatus 82 prunus 82 Stromatocerium 96 richmondense 82 stromatolites 88 Symploca hydnoides 86 jurassica 84, 86 Tormitchell Quarry 80, 96* Tournaisian 83 Triassic 92 trichomes92,93*,94,98, 101 Vermont 86 Visean 83-4 Wenlockian 83 Westphalian 84 Zonotrichites lissaviensis 92 Accepted for publication 2 1 July 1980 The Earth Generated and Anatomized by William Hobbs An early eighteenth century theory of the earth Edited with an introduction by Roy Porter March 1981. 158pp. 6 plates, 8 text figures Bulletin of the British Museum (Natural History) Historical Series Vol. 8 Paper covers, 16.00 Wherein is shewn What the Chaos was: How and when the Oyster-shells, Cockle-shells, and all the other Marine productions, were brought upon, and incorporated in the Rocks and Mountains of the Earth. Proving that it was not at, or by the Deluge, as is Vulgarly Supposed. Also Why and When the Said Hills and Mountains were raised. As also shewing not only the certain Cause of the Ebbing and Flowing of the Tides, But even the Two places where they are once in XV Dayes originally Moved; and where they ultimately meet each other. By which, as also by diverse Arguments, the Vulgar Notion of the Moon's Governing them, is fully confuted. Together with Many other Philosophical Doctrines and Discoveries; Suitable to such a Subject; not before advanced. (From Hobbs' title page) The Bulletin of the British Museum (Natural History) is published in five Series, Zoology, Entomology, Botany, Geology, and Historical. Details and complete lists are free on request. Publication Sales British Museum (Natural History) Cromwell Road London SW7 5BD Titles to be published in Volume 35 Lower Ordovician Brachiopoda from mid and southwest Wales. By M. G. Lockley & A. Williams The fossil alga Girvanella Nicholson & Etheridge. By H. M. C. Danielli Centenary Miscellanea The English Upper Jurassic Plesiosauroidea (Reptilia) and a review of the phylogeny and classification of the Plesiosauria. By D. S. Brown Printed by Henry Ling Ltd, Dorchester Bulletin of the British Museum (Natural Centenary Miscellanea Geology series Vol 35 No 3 29 October 1981 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries should be sent to : Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England. World List abbreviation: Bull Br. Mus. not. Hist. (Geol.) Trustees of the British Museum (Natural History), 1981 ISSN 0007-1471 Geology series Vol 35 No 3 pp 109-252 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 29 October 1981 Centenary Miscellanea f< & GENERAL *' To mark the centenary of the British Museum (Natural History) in 1981, this number of the Bulletin comprises nine short papers, each of which re-examines historical material in the collections of the Department of Palaeontology. Contents Page Reassessment of the Ordovician brachiopods from the Budleigh Salterton Pebble Bed, Devon. By L. R. M. Cocks & M. G. Lockley Ill Felix Oswald's Turkish Algae. By G. F. Elliott 125 J. A. Moy-Thomas and his association with the British Museum (Natural History). By P. L. Forey & B. G. Gardiner 131 Burials, bodies and beheadings in Romano-British and Anglo-Saxon cemeteries. ByM. Harman, T.I. Molleson & J. L. Price 145 The Jurassic irregular echinoid Nucleolites clunicularis (Smith). By D. N. Lewis & H.G.Owen 189 Phanerotinus cristatus (Phillips) and the nature of euomphalacean gastropods. By N. J. Morris & R. J. Cleevely .... . 195 Agassiz, Darwin, Huxley, and the fossil record of teleost fishes. By C. Patterson . .213 The Neanderthal problem and the prospects for direct dating of Neanderthal remains. By C. B. Stringer & R. Burleigh 225 Hippoporidra edax (Busk 1859) and a revision of some fossil and living Hippoporidra (Bryozoa). By P. D. Taylor & P. L. Cook 243 Reassessment of the Ordovician brachiopods from the Budleigh Salterton Pebble Bed, Devon L. R. M. Cocks Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 5BD M. G. Lockley Department of Geology, University of Colorado at Denver, 1100 14th Street, Denver, Colorado 80202, U.S.A. Synopsis The Ordovician part of the brachiopod fauna from the quartzite pebbles found in the Triassic Budleigh Salterton Pebble Bed, Devon, is reassessed, chiefly using the nineteenth-century collections made by Vicary, Valpy and others and originally figured by Salter and Davidson. Davidson's overlooked conclusion of two separate Ordovician ages is confirmed. The older Arenig has a bizarre fauna of large inarticulate brachiopods, and there is a younger and more abundant fauna of late Llandielo age. The European affinities of the fauna are confirmed and the taxonomy of the late Llandielo fauna (including some comparative material from France) is revised. The species Corineorthis erratica (Davidson), Tafilaltia valpyana (Davidson) and Salopia? pulvinata (Salter) are redescribed in detail, and other species placed in synonymy with them. Introduction The value of old collections is well demonstrated by the brachiopods and other shelly fossils found during the nineteenth century in pebbles from the conglomerates of Triassic age near Budleigh Salterton, Devon (Fig. 1). Although it was a Mr Carter who found the first fossils in the pebbles in about 1835, it was an enthusiastic amateur geologist, W. Vicary, who first made an extensive collection of the shells from about 1860, and brought them to the attention of the well-known palaeontologist J. W. Salter. Salter encouraged Vicary to give a short account of the rocks to the Geological Society of London (Vicary 1864). Salter (1864) appended a 'note on the fossils', containing descriptions and illustrations of 37 taxa of many phyla. Vicary 's collection was bequeathed to the British Museum (Natural History) in 1903, and it was joined there in 1910 by the collection of another enthusiastic amateur, R. H. Valpy, after the latter's death. Salter was astute in perceiving the central European affinities of the Budleigh Salterton fauna, which was quite unlike any of those faunas found elsewhere in Britain (apart from Cornwall) in rocks of the same age. Salter assessed this age as 'Lower Silurian' (i.e. Ordovician in modern terms), although he was puzzled by some shells which he interpreted as the oldest spiriferide brachiopods. The shells were subsequently reassessed by T. Davidson (briefly reported in 1866:67 and elaborated in 1870), the greatest of the nineteenth-century brachiopod workers. He then recognized that the pebbles, although all similar in their quartzite lithology, could be assigned to two separate ages, Ordovician and Devonian; it was the latter which included the questionable spiriferides. Davidson figured the brachiopod fauna in a series of publications (1866-71, 1870, 188 la). Davidson's collection from Budleigh Salterton, largely given to him by other collectors but containing many type specimens, was bequeathed to the Museum in 1885. Good specimens are extremely difficult to collect from the Budleigh Salterton Pebble Beds and thus the Museum Bull.Br.Mus.nat.Hist.(Geo\.)35(3): 111-124 Issued 29 October 1981 111 112 L. R. M. COCKS & M. G. LOCKLEY SARTHE Le Mans km Fig. 1 Locality map showing the relative present-day positions of Budleigh Salterton, Devon, Gorran Haven, Cornwall and the Armorican Peninsula, France. possesses the largest, and indubitably the most important, collection of brachiopods from that area, which has not been critically reassessed for a century. The present paper deals only with the Ordovician part of the fauna; the Devonian brachiopods will form the topic of a subsequent publication. Davidson, in a later review (1880 : 339), also realized that there were two separate Ordovician ages represented by the Budleigh Salterton brachiopods. The older of these he identified as 'Ores Armoricain - Lowest portion of Llandeilo', and the younger as 'Ores de May - Caradoc' (Davidson's stratigraphical terminology compared with modern usage can be found in Cocks 1978 : fig. 1). This important conclusion appears, however, to have been overlooked by all subsequent workers. The older horizon is repre- sented only by the large inarticulate brachiopods and the younger by the more common pebbles dominated by 'Orthis budleighensis\ The Ordovician brachiopods found in the Budleigh Salterton pebbles are very similar to those found in situ in the Armorican Peninsula of Brittany and Normandy, France, and also to some extent in the Gorran Haven area of Cornwall (Fig. 1). Sadler (1974) has reviewed the Gorran Quartzites and their trilobite fauna, assigned to them a Llandeilo, possibly a late Llandeilo, age and (1974 : 74) recorded comparable trilobites from Budleigh Salterton pebbles. During the last 20 years much has been published on the Ordovician of the Armorican massif, the stratigraphy of which is now well known (see summaries by Babin et BUDLEIGH SALTERTON BRACHIOPODS 1 1 3 al. 1976 a, b). While a few French brachiopods have been revised recently (Melou 1973, 1975), most of the Armorican brachiopod fauna remains undescribed and we take the opportunity here of illustrating and measuring some brachiopods from Normandy, based on collections made by Sir R. I. Murchison, the founder of the Silurian system, which came to the Museum in 1911 with all the other foreign specimens of the Geological Society of London. The Arenig Fauna The brachiopod fauna of the Ores Armoricain in the Armorican Peninsula, France, was originally described (without illustrations) by Rouault (1850), and later revised and extended by Davidson (1880, 1881 b, c). Davidson's original French material, presented to him by Lebesconte, Guillier, Moriere and others, is also now in the British Museum (Natural History), together with his notes and correspondence and the original drawings for the published plates. No primary work has been carried out since then on the brachiopod fauna, although new generic names were subsequently given to two of the species. Henry (1980) has revised the sparse Ores Armoricain trilobite fauna, which consists ofOgyginus armoricanus (Tromelin & Lebesconte), Platycoryphe heberti (Lebesconte) and P. dangeardi Henry, and endorsed the Arenig age deduced by previous workers. Although there is no internal evidence to be sure of the detailed age of the Ores Armoricain faunas within the Arenig, Dr R. A. Fortey has suggested to us that a middle Arenig age appears the most probable from an assessment of contemporary faunas elsewhere in transgressive situations comparable to that seen in the Ores Armoricain. Davidson's work (1880 on Brittany, 188 \b on Normandy and 1881c on Sarthe) revealed the following brachiopods to be present in the French Ores Armoricain (with updated generic names): Brittany Normandy Sarthe Ectenoglossa leseueuri (Rouault, 1850) x x x Lingulepis crassipyxis Havlicek, 1 980 x Lingulobolus brimonti (Rouault, 1850) x x Lingulobolus hawkei (Rouault, 1850) x x Pseudobolus ? slateri (Davidson, 1866) x x Tomasinacriei (Davidson, 1881) x 2 Figs 2-5 Inarticulate brachiopods of Arenig age from the Budleigh Salterton Pebble Bed. Fig. 2, B 21518 Lingulobolus hawkei (Rouault, 1850), xl'O; the specimen is that figured by Salter (1864 : pi. 17, fig. 4) as Lingula rouaulti sp. nov., and was subsequently selected (Cocks 1978 : 12) as the lectotype of the latter species; W. Vicary Coll. Fig. 3, B 21504 Pseudobolusl salteri (Davidson), x 1-2; the lectotype (sel. Cocks 1978 : 13), originally figured by Davidson (1866: pi. 1, fig. 28); W. Vicary Coll. Fig. 4, B 21516 Ectenoglossa lesueuri (Rouault, 1850), x 1-0; the specimen previously figured by Salter (1864: pi. 17, fig. 1) and Davidson (1866: pi. 1, fig. 3); W. Vicary Coll. Fig. 5, B 21514 Lingulobolus brimonti (Rouault, 1850), x 1-3; the specimen figured by Salter (1864 : pi. 17, fig. 2) and Davidson (1866 : pi. 1, fig. 2 1 ) as Lingula hawkei Rouault; W. Vicary Coll. 114 L. R. M. COCKS & M. G. LOCKLEY No Arenig trilobites have yet been noted in the Budleigh Salterton faunas, but the brachiopods Ectenoglossa leseueuri (Fig. 4), Lingulobolus brimonti (Fig. 5), L. hawkei (Fig. 2) and Pseudobolusl salteri (Fig. 3) are all known (Salter 1864, Davidson 1870, 188 la), and thus match exactly the Ores Armoricain inarticulate brachiopod fauna of Brittany and Normandy. The lack of illustrations in Rouault's original work led the first wave of subsequent revisers astray in their identifications of brimonti and hawkei. The attributions published by Salter (1 864) and Davidson (1 866, 1 870) are thus in error, but this was rectified subsequently by Davidson (1880, 188 la) and led to the correct synonymizing of Salter's rouaulti (whose lectotype is refigured here, Fig. 2) with hawkei. Davidson's illustrations are so good that further redescription of the fauna is not given here, although the fauna as a whole could usefully be revised in the light of a thorough reappraisal of inarticulates and possibly bivalves of Arenig age from the whole central European area, including Bohemia and the Iberian Peninsula. Havlicek (1980) has described a comparable fauna from the Montagne Noire, southern France. The Llandeilo Fauna In contrast to the bizarre inarticulate fauna found in the pebbles of Arenig age, the Llandeilo age pebbles from Budleigh Salterton show an overwhelming dominance of articulate brachiopods, in particular the enteletacean previously commonly identified as Orthis budleighensis and which is now known to be Tafilaltia valpyana. The brachiopod fauna is not a diverse one, and consists of the three orthide species described in detail below, rare lingulides identified by Davidson as Lingula morieri Tromelin, 1876, and a single specimen, BM(NH) Palaeont. Dept. B 21 525, identifiable as Porambonites sp. This last was figured by Salter ( 1 864 : pi. 17, fig. 9) and identified correctly by him, even though it was later unaccountably placed in the synonymy of pulvinata by Davidson (188 la : 358). There are no other fossils on the same block, and thus it is just possible that the specimen could be of Arenig age; however this large (length 28'5 mm, width 39'2 mm) brachial valve appears to be more similar to Llandeilo and Caradoc forms of the genus. The specimen (B 2 1531) figured as Terebratula 7 sp. by Davidson (1870: pi. 4, fig. 11; 188 la: pi. 41, fig. 23) is a monoplacophoran resembling Vallatotheca, kindly reidentified by Dr N. J. Morris. There are also bivalves, gastropods, crinoids and trilobites in the late Llandeilo fauna, all awaiting revision apart from Neseuretus tristani, but all relatively uncommon when compared with the dominant brachiopods. The fauna has its nearest geographical comparison with that from the Gorran Quartzites of the Gorran Haven area, Cornwall (Sadler 1974), sharing with it the trilobite Neseuretus tristani (Brongniart). However, the Cornish beds also yield a common large orthid (figured by Davidson, 188 la, as Orthis calligramma and to be revised by M. G. Bassett), which is not present in the Budleigh Salterton pebbles, and the other Cornish brachiopods include congeneric Corineorthis and heterorthids, but not species in common. The Budleigh Salterton Llandeilo pebbles probably have their closest faunal affinity with the fauna from the Gres de May Formation in Normandy. Detailed comparison is difficult, since the French brachiopod fauna has not yet been properly described, but the opportunity is taken here (see systematic section below) to demonstrate the specific identity of the commonest form at both Budleigh Salterton and in the Gres de May, Tafilaltia valpyana (Davidson). It is relevant to note that the same species (previously described as Tafilaltia dalmanelloides ) is also present in the Llandeilo age Skalka Quartzite of Bohemia (Havlicek 1970). The age of this part of the Budleigh Salterton pebble fauna is almost certainly late Llandeilo, although the upper part of the Gres de May represents deposition which probably continued on into early Caradoc times (Babin et al. 1976a:381). BUDLEIGH SALTERTON BRACHIOPODS 1 1 5 Systematic Palaeontology All the figured and quoted specimens are in the British Museum (Natural History) (B and BB). In general the specimens are well preserved, but the recemented quartzite lithology precludes the preservation of the very fine details of ornament. Class ARTICULATA Huxley, 1869 Order ORTHIDA Schuchert & Cooper, 1932 Suborder ORTHIDINA Schuchert & Cooper, 1932 Superfamily ORTHACEA Woodward, 1 852 Family PLECTORTHIDAE Schuchert & Le Vene, 1929 Subfamily PLECTORTHINAE Schuchert & Le Vene, 1929 Genus C0/?//VE0/?7mSStubblefield, 1939 Corineorthis erratica (Davidson, 1 869) (Figs 6-11) 1 869 Orthis Berthoisi ? var. erratica Davidson : 233; pi. 32, figs 2 1-28. 1870 Orthis Berthoisi ? var. erratica Davidson; Davidson : 83; pi. 5, figs 13-16. 1 88 1 a Orthis Berthoisi var. erratica Davidson; Davidson : 356; pi. 41. figs 1-9. 1978 Svobodaina ? erratica (Davidson) Cocks : 74. DESCRIPTION. Large dorsibiconvex to convexoplane orthide with obtuse cardinal angles and incipient development of weak plication; pedicle valve with subcircular to subpentagonal outline and averaging 83% as wide as long in 5 specimens; brachial valve averaging 80% as wide as long in 7 specimens, and between one-quarter and one-third as deep as long; ventral interarea slightly curved and apsacline, with prominent growth striations; open, wide delthyrium with conspicuous pedicle callist; short anacline dorsal interarea, with open notothyrium largely filled by posterior part of myophore; radial multicostellate ornament with 3 to 4 ribs per mm at 5 mm length; a few prominent concentric growth lines. Teeth short, stout, and supported by dental plates extending anteriorly for up to 21% of valve length and diverging at about 90; well-impressed ventral diductor muscle scars elongately bilobed, about 85% as wide as long, extending anteriorly for an average of just over 50% of valve length and enclosing elongate adductor scars which extend antero- medially to link with the vascular impressions. Cardinal process simple, with well-developed triangular myophore and simple blade-like shaft extending anteriorly across the notothyrial platform onto median septum; elongate brachiophore bases average about one-quarter valve length, initially diverging for about two-thirds of their length, then converging across the flanks of the platform to merge anteriorly with the short median septum. Faint elongate muscle scars, about 85% as wide as long, expanding anteriorly, with maximum width in the mid-part of the valve, although the posterior pair of scars are most deeply impressed to form pits. MATERIAL. Lectotype of erratica (selected Cocks 1978 : 74) B 20936 (Figs 7, 8, 10), external and internal moulds of a pedicle valve, originally figured by Davidson (1869 : pi. 32, figs 21-23); R. H. Valpy Collection, from late Llandeilo pebble in Trias, Budleigh Salterton Pebble Bed, Devon. Additional material in the British Museum (Natural History) : 1 1 brachial valves and 4 pedicle valves also from the Budleigh Salterton Pebble Bed. Specimens of Corineorthis also probably attributable to erratica, e.g. B 13289, are known from the Ores de May of St Germain, France. DISCUSSION. Examination of the material described as Orthis berthoisi var. erratica by 116 L. R. M. COCKS & M. G. LOCKLEY Davidson (1869) shows clearly that it is congeneric with material from Cornwall described as Corineorthis decipiens by Stubblefield (1939), itself recently recognized by M. G. Bassett as a junior synonym of Orthis berthoisi var. cornubiensis Davidson, 1881. Davidson (1881a : 355-7) also discussed the problem of Orthis berthoisi itself, which comes from the slates of La Couyere, Normandy (Rouault 1849 : 68; pi. 2, figs 4-4c). Topotype material of berthoisi (e.g. B 13233 in the Davidson Collection ex Tromelin), although severely crushed, leads us to agree with Davidson that erratica and berthoisi are not the same: until Rouault's types are revised, perhaps berthoisi is best regarded as a nomen dubium. The La Couyere slates are now known to form part of the Riadan Formation, which is late Ordovician in age (Babin et al. 19760 : 374). 11 Figs 6-11 Corineorthis erratica (Davidson, 1869) from pebbles of late Llandeilo age, Budleigh Salterton Pebble Bed, Devon. Figs 6, 9, B 20936 (additional brachial valve on lectotype block and latex cast of it), x 2*0. Figs 7, 8, 10, B 20936, latex cast of internal mould, latex cast of external mould and internal mould of pedicle valve, lectotype (sel. Cocks 1978 : 74), originally figured by Davidson (1869: pi. 32, figs 2 1-23), x 1-5; R. H. Valpy Coll. Fig. 11,B 13284, latex cast of pedicle valve internal mould, x 1-5; T. Davidson Coll. ex W. Vicary. Available statistical data indicate that C. erratica and C. pustula Williams, from the type Llandeilo area, although showing no significant differences in the outline of either valve or the shape of the cardinalia, do differ in their ventral diductor scars, which are more splayed in erratica (85% of length, compared with 62% in pustula). The coarse texture of the Budleigh Salterton sediments makes it hard to determine whether or not the internal pustules and characteristic exopunctae of pustula are present in erratica. C. erratica also has a better- developed myophore than pustula, presumably indicating a better-developed diductor attachment site to complement the larger ventral scars, and we continue to recognize them as two separate species. C. erratica has muscle bounding ridges which are laterally curved, in contrast to the straight ridges present in C. cornubiensis. BUDLEIGH SALTERTON BRACHIOPODS 117 Superfamily ENTELETACEA Waagen, 1884 Family HETERORTHIDAE Schuchert & Cooper, 193 1 Genus TAFILALTIA Havlicek, 1970 Tafilaltia valpyana (Davidson, 1869) (Figs 12-23, 30, 31) 1 864 Orthis redux Barrande; Salter : 294; pi. 1 7, fig. 7 (non Barrande, 1 848). 1 869 Orthis redux Barrande; Davidson : 224; pi. 28, figs 6-9 (non Barrande, 1 848). 1869 Orthis testudinaria Dalman; Davidson : 226 pars; pi. 28, fig. 22, non figs 13-21, 23, 24 (non Dalman, 1828). 1869 Orthis Valpyana Davidson : 235; pi. 32, figs 29-33. 1 870 Orthis redux var. budleighensis Davidson : 82; pi. 5, figs 9-12. 1 870 Orthis Valpyana Davidson; Davidson : 83; pi. 5, figs 23-25. 1 88 1 a Orthis budleighensis Davidson; Davidson : 358; pi. 4 1 , figs 1 2-20; pi. 42, figs 1 6-25. 18810 Orthis Valpyana Davidson; Davidson : 361; pi. 41, figs 21, 22. 1 970 Tafilaltia dalmanelloides Havlicek : 20; pi. 4, figs 1-8. 1977 Tafilaltia dalmanelloides Havlicek; Havlicek : 1 1 5; pi. 1 1 , figs 3-5, 7, 8. 1 978 Howellites ? budleighensis (Davidson) Cocks : 65 . 1978 Heterorthina valpyana (Davidson) Cocks : 74. DESCRIPTION. Fairly small planoconvex to ventribiconvex transverse heterorthinid with brachial valve averaging 80% as long as wide in 46 specimens; pedicle valve averaging 83% as long as wide in 33 specimens and 25% as deep as long in 10 specimens; ventral interarea short, apsacline, with wide, open delthyrium, dorsal interarea short anacline, notothyrium completely filled by cardinal process; radial ornament of fine costellae, numbering 5 to 6 per mm 5 mm anteriorly of the umbo, which curve posterolaterally to merge with the hinge line. Ventral interior with small teeth supported by low, widely-divergent dental lamellae extending laterally for about half the valve width and anteriorly for about one-quarter the valve length; pedicle callist well developed; muscle field large, usually poorly impressed anteriorly but averaging 56% of valve length and 99% as long as wide in 4 specimens in which the anterior edge was more strongly impressed; muscle field slightly flabellate with lateral margins of diductor scars less well developed than the median margins, which enclose poorly differentiated adductor scars; vascular canals, separated by ridges, can be quite deeply impressed. Dorsal interior with well-developed, ponderous cardinal process with a raised posterior myophore, supported anteriorly by a low, broad shaft barely distinguishable from the supporting median septum; triangular posterior platform of myophore with prominent median ridge raised above lateral margins; brachiophores short and blade-like, with bases which average 17% as long as valve length and 64% as long as wide in 33 specimens; sockets narrow and well defined, and without fulcral plates; muscle scars elongate and variably impressed, quadripartite and averaging 46% of valve length and 85% as wide as long in 15 specimens. MATERIAL. Lectotype of valpyana (selected Cocks 1978 : 74) B 21533, external mould (Fig. 12) and counterpart internal mould of a pedicle valve, the original of Davidson (1869 : pi. 32, figs 29-31); W. Vicary Collection, from late Llandeilo pebble in Trias, Budleigh Salterton, Devon. Additional material: over 300 specimens from other blocks in the Budleigh Salterton Pebble Bed in the Vicary and Valpy Collections, including B 21616, the lectotype of budleighensis Figs 15, 16). Also known from Perhaver Quartzite, Gorran Haven, Cornwall; the Gres de petit May, Normandy, France; and the Skalka Quartzite of Bohemia, Czechoslovakia, including the holotype of dalmanelloides, from Chrbina Hill, near Nenacovice (Havlicek 1970 : 20-21). DISCUSSION. A few relatively large heterorthid specimens, distinguished as the new species valpyana by Davidson (1869), closely resemble the much more numerous and smaller form 118 L. R. M. COCKS & M. G. LOCKLEY Figs 12-22 Tafilaltia valpyana (Davidson, 1869) from pebbles of late Llandeilo age, Budleigh Salterton Pebble Bed, Devon. Fig. 12, B 21533, latex cast of external mould of a pedicle valve, lectotype (sel. Cocks 1978 : 74), originally figured by Davidson (1869 : pi. 32, fig. 29), x2'0; W. Vicary Coll. Fig. 13, B 21 524, internal mould of a brachial valve, Davidson (1869 : pi. 32, figs32,32a), x 1-8; W. Vicary Coll. Fig. 14, B 21509, latex cast of the internal mould of a brachial valve, x 2 - 0. Figs 15, 16, B 21616, latex cast and internal mould of a brachial valve, lectotype of budleighensis (sel. Cocks 1978:65), originally figured by Davidson (1870 : pi. 5, fig. 12 lower right), x 3'0. Fig. 17, latex cast of the interior of another brachial valve on the same slab as Figs 15 and 16, x 2-0. Fig. 18, BB 95940, latex cast of the external mould of a brachial valve, x3'0; W. Vicary Coll. Fig. 19, B 2 1621, latex cast of two internal moulds of pedicle valves, x3'0; W. Vicary Coll. Figs 20, 21, BB 95941, internal mould of a brachial valve and latex cast of it, x 2-4; W. Vicary Coll. Fig. 22, BB 95942, internal mould of a pedicle valve x 3-0; W. Vicary Coll. BUDLEIGH SALTERTON BRACHIOPODS 119 15 10 10 width 15 20 Fig. 23 Length-width measurements of brachial valves of Tafilaltia valpyana (Davidson, 1869), from pebbles of late Llandeilo age in Trias, Budleigh Salterton, Devon. The solid squares represent the three specimens figured by Davidson as valpyana; the open triangles represent 45 specimens, including the lectotype, of budleighensis from the same locality. The average (of the budleighensis sample) and regression line of the length on the width are also shown. with which they occur. This smaller form had been recognized by Salter (1864), in his first descriptions of Budleigh Salterton fossils, as being similar to species from the Bohemian province, and he had identified it as Orthis redux Barrande, a species since made the type of Drabovia Havlicek, 1 95 1 . It was not until 1 870 that Davidson felt he could formally separate the two forms, when he gave the Devon brachiopod the varietal name of budleighensis. We can now demonstrate (Fig. 23) that valpyana and budleighensis exhibit similar growth ratios, and so consider the better- known latter name to be a junior synonym. The species can be readily identified as an heterorthid, and our initial reaction was to compare it with an early undescribed species of Heterorthis from the late Llandeilo of Dyfed, Wales (Addison 1974). Similarly, M. G. Bassett has assigned closely related material from Cornwall to the genus Heterorthina. However, comparisons between valpyana and praeculta, the type species of Heterorthina, whilst showing no significant difference in valve outline, relative length of muscle scars and dorsal cardinalia, do show a significant difference in the relatively greater depth of the pedicle valve (5%>p>2%), and we have also noted that H. praeculta has more clearly differentiated and more deeply impressed ventral muscle scars, particularly towards their anterior. The opportunity is taken here to illustrate some topotype specimens of Heterorthina praeculta (Figs 24-29), including the lectotype (selected Cocks 1978 : 74) which has not been figured since its original illustration by Bancroft (1928: 120 L. R. M. COCKS & M. G. LOCKLEY 4f4& 27 Figs 24-29 Topotype specimens of Heterorthina praeculta Bancroft, 1928, from the Cheney Longville Flags (Caradoc: Marshbrookian), south side of road from Cwm Head to Marshbrook, Shropshire, Grid Reference SO 437896. Fig. 24, BB9152, internal mould of a pedicle valve, lectotype (sel. Cocks 1978 : 74), originally figured by Bancroft (1928 : pi. 2, fig. 18), x 1-5; B. B. Bancroft Coll. Figs 25, 29, BB 7288, internal mould of a brachial valve and latex cast of the external mould, x 2-0; J. M. Hurst Coll. Fig 26, 28, BB 68827, internal mould and latex cast of a brachial valve, x 2'0; B. B. Bancroft Coll. Fig. 27, BB 72289, latex cast of the external mould of a brachial valve, x 3*0; J. M. Hurst Coll. pi. 2, fig. 18). Although the cardinalia ofvalpyana resembles some variants of Heterorthina kerfornei, such as the specimen figured by Melou (1975 : pi. 23, fig. 5), other specimens of kerfornei, and indeed most Heterorthina generally, differ from valpyana in possessing a cleft shaft supporting the myophore (cf. Hurst 1979). In our opinion, the species is best assigned to Tafilaltia Havlicek, a genus considered ancestral to Heterorthis by Havlicek (1977) and in its earlier species showing poorly differen- tiated ventral musculature, as in valpyana. The ponderous tripartite myophore and absence of chilidium is diagnostic of Tafilaltia, and indeed Havlicek (1970: 17-21) included material from the Ores de May within his Tafilaltia dalmanelloides, which is accepted here, and is why we have suppressed the latter species as a junior synonym of valpyana. Comparisons between valpyana from Budleigh Salterton and apparently conspecific French material (Fig. 30) from the Ores de petit May, May, Normandy (B 85258), showed no significant differences in any of the nine characters tested, i.e. the outline of both valves (data for 23 brachial valves included in Fig. 31), the relative depth of the pedicle valve, the shape and relative length of the dorsal cardinalia, and the shape and relative length of both dorsal and ventral muscle scars. Similarly no significant differences in dorsal outline or shape and relative length of ventral muscle scars can be detected through comparisons made with Addison's unpublished statistics (1974: tables 14 and 1 5) of the small heterorthid mentioned above (Heterorthis sp. nov.) from late Llandeilo to early Caradoc sandstones of Lampeter Velfrey, Dyfed. However, this undescribed form differs from valpyana in its sharply- differentiated ventral muscle scars, and perhaps represents a direct link between Tafilaltia and better-known Caradoc Heterorthis such as Heterorthis alternata (J. de C. Sowerby) (see Williams 1963). BUDLEIGH SALTERTON BRACHIOPODS 121 Fig. 30 Tafilaltia valpyana (Davidson, 1869) from the Ores de petit May, May, near Caen, Normandy, France. B 85258, part of a slab containing numerous external and internal moulds of both valves, x 3*0; Sir R. I. Murchison Coll. per Geological Society of London Coll., pres. 1911. 10 i I i I I I /idth 10 15 Fig. 31 Length-width measurements of 23 brachial valves of Tafilaltia valpyana (Davidson, 1869) from the Ores de petit May, May, near Caen, Normandy, France, all on slab B 85258 (Fig. 30). The average and regression line of the length on the width are also shown. 122 L. R. M. COCKS & M. G. LOCKLEY Family LINOPORELLIDAE Schuchert & Cooper, 193 1 Genus SALOP1A Williams, 1955 Salopia ? pulvinata (Sa\ter, 1864) (Figs 32-35) 1 864 Orthis pulvinata Salter : 294; pi. 1 7, fig. 8. 1870 Orthis pulvinata Salter; Davidson : 83 paw; pi. 5, figs 17, 19, ? fig. 18. 1881a Orthis pulvinata Salter; Davidson : 357; pi. 41, figs 10, 1 1. 1978 Pionodema ? pulvinata (Salter) Cocks : 80. DESCRIPTION. Large, transverse linoporellid with brachial valve averaging 71% as long as wide (range 71-72%), and between 19 and 26% as deep as long in 3 valves. Dorsal interarea short, flat, anacline, with wide open notothyrium. Ornament poorly known, consisting of faint fine radial costellae. Cardinal process simple, blade-like, consisting of anteriorly widening shaft merging with notothyrial platform. Brachiophores short, with convergent bases extending forward for about one-fifth to one-quarter valve length; sockets narrow, with small curved fulcral plates. Dorsal adductor scars elongate and quadripartite, averaging 49% as long as valve (range 47-52%) in three valves, with anterior pair of scars larger than posterior pair. The muscle field is divided by a broad median septum. Pedicle valve unknown. 35 Figs 32-35 Salopia ? pulvinata (Salter, 1864) from late Llandeilo pebbles in Trias, Budleigh Salterton, Devon. Fig. 32, B 21523, internal mould of a brachial valve, lectotype (sel. Cocks 1978:80), originally figured by Salter (1864: pi. 17, fig. 8),xl-5. Figs 33, 34, BB 70910, internal mould of a brachial valve and latex cast of it, x2'5; W. Vicary Coll. Fig. 35, B 13051, internal mould of a brachial valve, x 2'5; T. Davidson Coll. ex Winwood. MATERIAL. Lectotype B 21523 (sel. Cocks 1978 : 80), the part and counterpart of a brachial valve (Fig. 32), figured Salter (1864: pi. 17, fig. 8); from Ordovician pebble in Trias, Budleigh Salterton, Devon. Other material: two other brachial valves, B 13051 and BB 70940, both internal moulds only, from the same locality. DISCUSSION. Only three brachial valves and no pedicle valve are known of pulvinata; although Davidson (188 la : 358) states that 'M. de Tromelin informs me that O. pulvinata occurs in company with O. Berthoisi, var. erratica, at Saint Germain-sur-Ille, La Bouexiere, Champeaux (Ille-de-Vilaine), and in other places', no French material is available for comparison. In the absence of any pedicle valves the generic determination can only be provisional, hence the query, but the three brachial valves are certainly similar to Salopia turgida (M'Coy), as recently revised by Lockley & Williams 1981, particularly in their simple blade-like cardinal processes and quadripartite adductor muscle scars. However, pulvinata is more transverse (71% as long as wide, as compared with 88% for turgida), the brachiophores of turgida are relatively long in comparison with the quite short brachio- phores of pulvinata, and the adductor muscle scars are more deeply impressed and less elongate in pulvinata than in turgida. A full revision of these species, and their relationship BUDLEIGH SALTERTON BRACHIOPODS 123 to the other species of Salopia, such as S. globosa (Williams, 1949), S. triangularis (J. de C. Sowerby, 1839) and the type species, S. salteri (Davidson, 1869) and its subspecies gracilis Williams, 1955, must await the discovery of pedicle valves at Budleigh Salterton or the collection of more material in France. Acknowledgements We are grateful to Dr A. W. A. Rushton, Dr R. A. Fortey, Dr M. G. Bassett and Dr A. Williams for discussion, to Dr C. Babin for information on French localities, and to Dr C. H. C. Brunton for reading a draft of this paper. References Addison, R. (1974). The bio stratigraphy of the Llandeilo fades of South Wales. Ph.D. thesis, Univ. of Belfast (unpubl.). Babin, C., Arnaud, A., Blaise, J., Cavet, P., Chauvel, J. J., Deunff, J., Henry, J.-L., Lardeux, H., Melou, M., Nion, J., Paris, F., Plaine, J., Quete, Y. & Robardet, M. 1976a. The Ordovician of the Armorican Massif (France). In Bassett, M. G. (ed.), The Ordovician System: 359-385. Cardiff. , Chauvel, J.-J., Lardeux, H., Paris, F. & Robardet, M. 19766. Lexique des Formations de 1'Ordovicien armoricain. Bull. Soc. geol. miner. Bretagne, Rennes, Num. spec. : 1-31. Bancroft, B. B. 1928. On the notational representation of the rib-system in Orthacea. Mem. Prnr Manchr lit. phil. Soc. 72 : 53-90, pis 1-3. Cocks, L. R. M. 1978. A review of British Lower Palaeozoic brachiopods, including a synoptic revision of Davidson's monograph. Palaeontogr. Soc. (Monogr.), London. 256 pp. Davidson, T. 1866-71. A Monograph of the British Fossil Brachiopoda. Part VII. The Silurian Brachiopoda. Palaeontogr. Soc. (Monogr.), London. 397 pp., 50 pis. 1 870. Notes on the Brachiopoda hitherto obtained from the "Pebble-bed" of Budleigh Salterton, near Exmouth, in Devonshire. Q. Jl geol. Soc. Lond. 26 : 70-90, pis 4-6. 1880. On the species of Brachiopoda that characterize the "Ores Armoricain" of Brittany, together with a few observations on the Budleigh Salterton "Pebbles". Geol. Mag., London, (2) 7: 337-343, pi. 10. 18810. Monograph of the British Fossil Brachiopoda. Vol. IV, part IV. Devonian and Silurian Brachiopoda that occur in the Triassic Pebble Bed of Budleigh Salterton in Devonshire. Palaeontogr. Soc. (Monogr.), London, : 317-368, pis 38-42. 18816. Note sur les Lingules du Ores armoricain de la Sarthe. Bull. Soc. geol. Fr., Paris, (3) 9 -.372-377, pi. 7. 1881c. Note sur les Brachiopodes trouves par M. Moriere, dans le Gres armoricain de Bagnoles (Orne). Bull. Soc. linn. Normandie, Caen, (3) 5 : 89-93, pi. 2. Havlicek, V. 1970. Heterorthidae (Brachiopoda) in the Mediterranean Province. Sb. geol. Ved. Praha (P) 12: 7-40, pis 1-11. 1977. Brachiopods of the Order Orthida in Czechoslovakia. Rozpr. ustred. Ust. geol., Prague, 44: 1-327, pis 1-56. 1980. Inarticulate brachiopods in the Lower Ordovician of the Montagne Noire (South France). Mem. Soc. Etud. sci. Aude, Carcassonne, 1 : 1-1 1 , pis 1 , 2. Henry, J.-L. 1980. Trilobites ordoviciens du Massif Armoricain. Mem. Soc. geol. miner. Bretagne, Rennes, 22: 1-250, pis 1-48. Hurst, J. M. 1979. The stratigraphy and brachiopods of the upper part of the type Caradoc of south Salop. Bull. Br. Mus. nat. Hist., London, (Geol.) 32 (4) : 1 83-304, 557 figs. Lockley, M. G. & Williams, A. 198 1 . Lower Ordovician Brachiopoda from mid and southwest Wales. Bull. Br. Mus. nat. Hist., London, (Geol.) 35 (1) : 1-78, 263 figs. Melou, M. 1973. Le genre Aegiromena (Brachiopode - Strophomenida) dans 1'Ordovicien du Massif armoricain (France). Annls Soc. geol. N., Lille, 93 : 253-264, pis 33-36. 1975. Le genre Heterorthina (Brachiopoda, Orthida) dans la formation des Schistes de Postolonnec (Ordovicien), Finistere, France. Geobios, Lyon, 8 : 191-208, pis 20-24. Rouault, M. 1849. Memoire 1 sur la composition du test des Trilobites; 2 sur les changements de formes dus a des causes accidentelles, ce qui a pu permettre de confondre des especes differentes. Bull. Soc. geol. Fr., Paris, (2) 6 : 67-89, pis 1 , 2. 124 L. R. M. COCKS & M. G. LOCKLEY 1 850. Note preliminaire sur une nouvelle formation decouverte dans le terrain silurien inferieurde la Bretagne. Bull. Soc. geol. Fr., Paris, (2) 7 : 724-744. Sadler, P. M. 1974. Trilobites from the Gorran Quartzites, Ordovician of south Cornwall. Palaeontology, London, 17 : 71-93, pis 9, 10. Salter, J. W. 1864. Note on the fossils from the Budleigh Salterton Pebble-bed. Q. Jl geol. Soc. Lond. 20: 286-302, pis 15-1 7. Stubblefield, C. J. 1939. Some Devonian and supposed Ordovician fossils from South West Cornwall. Bull. geol. Surv. Gt Br., London, 2 (5) : 63-7 1 , pi 4. Vicary, W. 1864. On the Pebble-bed of Budleigh Salterton. Q. J I geol Soc. Lond. 20 : 283-286. Williams, A. 1963. The Caradocian brachiopod faunas of the Bala District, Merionethshire. Bull. Br. Mus. nat. Hist., London, (Geol.) 8 : 327-471, pis 1-16. Felix Oswald's Turkish Algae G. F. Elliott Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 5BD Synopsis Algae from the Cretaceous and Caenozoic of Turkey, collected by Felix Oswald in 1898, are re- examined in the light of over eighty years of subsequent algal studies. Introduction The collections of the British Museum (Natural History) are incredibly rich, not merely in the major treasures of the national collection, but in original historical specimens, unique in their day, which were the foundations of so much that followed. Algae are not usually attractive fossils when collected. Although studied by numerous nineteenth- and early twentieth-century microscopists, the organized development of algal studies generally as an important branch of micropalaeontology did not take place until after the second world war. It was largely initiated as part of the oil industry's researches at that time, to supplement their extensive pioneer use of Foraminifera. At the BM (NH), a separate subsection of fossil algae was not individualized from the fossil plants until 1969, when I was entrusted with this task. Because of the extensive connections of the oil industry with the Middle East, the national collection of fossil algae contains much material from this area, mostly from sampling carried out between 1930 and 1960. It is therefore of interest to note the presence in the collections of a few Middle East samples collected in 1898, and recognized as algal at the very beginning of the present century. Felix Oswald (1866-1958; obituaries by Swinnerton 1958, 1959) accompanied W. N. B. Lynch on his second tour of eastern Turkey (then known as Turkish Armenia) in 1898. Oswald's detailed geological observations in this then little-visited area were submitted as an academic thesis in 1905 and fully published with illustrations in 1906, in a book which he type-set and produced himself (Oswald 1906; preface). He acknowledges the help of R. Bullen Newton, then on the staff of the BM(NH), in connection with his palaeontology. There are four relevant samples in the Museum's collections. They are of Lower Cretaceous, Palaeocene-Lower Eocene and Miocene (two) geological age; all but the second were figured by Oswald. Their re-study, and discussion, is set out below. Discussion 1. Lower Cretaceous Oswald (1906) : 'Munieria'; plate facing p. 234 and pp. 236, 340. Buff-grey limestone from Akhveran, 42 km ESE of Bayburt (40 15' N, 40 16' E; eastern Turkey). Turkish Geological Map 1 : 800,000, Sheet 4, Erzerum (1943). The large dasyclad in Oswald's original thin section was identified (probably by Bullen Newton) as a Munieria (Deecke 1883); the comparison given is with Hovelacque (1900 : pi. 46, fig. 2). This latter, however, does not show a Munieria but Salpingoporella sp. Bull. Br. Mus. nat. Hist. (Geol.)35 (3) : 1 25-1 29 Issued 29 October 1 98 1 125 126 G. F. ELLIOTT Fig. 1 Euspondyloporella sp. Oswald's original section, x40. BM(NH) Palaeont. Dept., reg. no. V11063a. (Conrad 1970 : 70, and personal communication). Oswald's actual section (refigured here, Fig. 1) shows a Triploporella or related genus; the section is oblique and of an individual showing pressure-displacement of the structure, so it is not precisely diagnostic. However, in one of further thin sections now prepared from the small original sample, another dasyclad section (uncrushed) was revealed, showing branches of a different form (Fig. 2). It seems unlikely that these two sections are of different taxa and they may well be of the same individual. It is stated of Euspondyloporella duplicata Sokac & Nikler (1973 : 23) that 'the primary [branches] are represented by two forms. In the club-shaped [top] portion of the alga, they are thin and tubular, slightly thickened in the distal part. Below the top, the primary branches consist of a handle occupying \ of the length, and of an elongated egg-shaped thickening occupying \ of the total length of the branch'. In the Oswald material, the original section shows the second pattern, and the new section the first. Further evidence for this determination lies in the large number of primary and small number of secondary branches, the spore-packed branches seen (when not replaced by infill calcite) and OSWALD'S TURKISH ALGAE 127 Fig. 2 Euspondyloporella sp.; a second section showing club-shaped branches from the apex of the thallus, x 30. Reg. no. VI 1063b. gross dimensions, all shown both by Sokac & Nikler's Jugoslav type material and Oswald's Turkish material. It would seem, therefore, that Oswald's dasyclad can certainly be identified as an Euspondyloporella (Triploporelleae), and probably as E. duplicata, though suitable additional sections would be necessary to confirm this. The other accompanying organisms are a typical Tethyan Lower Cretaceous assemblage for this facies, which is widespread through the circum-Mediterranean and Middle East. They comprise the microproblematicum Carpathoporella fontis (Patrulius) (see Jaffrezo 1974 for the involved synonymy), the algae Cayeuxia sp. and Solenopora sp. and the problematic Lithocodium aggregatum Elliott. Bivalve and echinoid fragments occur. Oswald (1906:339,340) stated that the algal limestone was succeeded by radiolarian limestone. South of the Turkish frontier, in the Lower Cretaceous of Iraqi Kurdistan, the organic Qamchuqa Formation (with algae) intertongues with the basinal radiolarian 128 G. F. ELLIOTT Balambo Formation (Dunnington, Wetzel & Morton 1959 : 50, 230), and Oswald's Turkish account seems compatible. The exact level of the type-material of Euspondyloporella in the Lower Cretaceous was given as probably Barremian-Aptian (Sokac & Nikler 1973 : 8). The Oswald sample, to which he assigned Hauterivian age, does not show orbitolines etc., and if therefore from a pre-orbitoline horizon, a Hauterivian-Barremian age seems likely. 2. Palaeocene-Lower Eocene Oswald (1906) '.'Lithothamnion'; pp. 249, 418. Dark-grey limestone from Chorak Khan, 45 km NW of Bayburt. Turkish Geological Map 1 : 800,000, Sheet 3, Sivas (1946). 'Lithothamnion' was for a long time used as a general term for a very wide variety of coralline algae, Recent and fossil. The examples in Oswald's rock are cylindrical units of the segmented coralline Amphiroa; probably a new species, but the rock and its fossil content are markedly affected by mineralization and diagenesis - Oswald refers to the rock as a marble (Oswald 1906 : 248, 418) - and most examples of the fossil are obscured by this. Associated are fragments of Archaeolithothamnium sp., ? Pycnoporidium, lElianella (Parachaetetes auctt.), and what from outline and traces of structure remaining is probably the feather-alga, Distichoplax biserialis (Dietrich) Pia. Molluscan and echinoid debris is also present. This is probably a Palaeocene-Lower Eocene assemblage; a better-preserved sample could confirm this. All of these genera and species occur in rocks of that age in northeastern Iraq, south of the Turkish frontier, and so Oswald's assigned age of Middle-Upper Eocene can be modified. His ''Lithothamnion' is an Amphiroa sp. (Fig. 3) showing wide peripheral perithallus bordering the distinctive zones of the medullary hypothallus, and it is not the same as the Iraqi Palaeocene Amphiroa elliotti (Johnson 1964). The Turkish species is not formally described as new by reason of the preservation. tivl ' -;-*i*CTH JSmn'S -.v/s Fig. 3 Amphiroa sp.; original section of Oswald, x 50. Reg. no. VI 1064a. OSWALD'S TURKISH ALGAE 129 3. Miocene Oswald (1906) : ' ' Lithothamnion ramossissimum Reuss'; p. 52 and facing pi., fig. 9; p. 452. Light-brown limestone from Madrak, 19 km SSE of Erzerum. Also p. 81 and facing pi., fig. 1; p. 453. Creamy-pink limestone from Kanjean, 48 km north of Malazgirt. Malazgirt (3909'N, 4230'E) is 138km SE of Erzerum. Turkish Geological Map 1 : 800,000, Sheet 4, Erzerum (1943). Specimens from these two samples are preserved in the collections of fossil bryozoa at the BM(NH), reg. nos D 7958-7967 incl. The thin sections shows a rich algal-bryozoan assemblage similar to that of the European Vienna-Basin Miocene (Leithakalk), the algal microflora of which was revised in detail by Conti (1946). Oswald's figured Madrak specimen appears to be Lithophyllum piai Conti and his Kanjean specimen Palaeo- tharnnium archaeotypum Conti. Conclusions It is noticeable how, in spite of the relatively rudimentary knowledge of fossil algae available at the beginning of this century, Oswald assigned his material to approximately the right geological ages. He did not, of course, depend solely upon the algae; stratigraphy and other fossils were available. What is remarkable is his thoroughness in doing all that could be done to determine the algae, then regarded as of very little value. Modern sampling of these localities would yield more and perhaps better-preserved materials, but his pioneer effort is noteworthy. References Conrad, M. A. 1970. Barremian and Lower Aptian Dasycladaceae in the area surrounding Geneva (Switzerland). Geologica romana, 9 : 63-100. Conti, S. 1946. 1. Revisione criteria di Lithothamnium ramossissimum Reuss. 2. Le Corallinacee del calcare miocenico (Leithakalk) del bacino di Vienna. Pubbl. 1st. Geol. Univ. Genova, ser. A (Paleont.) 1-2 : 1-68. Deeke, W. 1883. Ueber einige neue Siphoneen. NeuesJb. Miner. Geol. Paldont. Jahrg., Stuttgart, 1883 I: 1-14. Dunnington, H. V., Wetzel, R. & Morton, D. M. 1959. Mesozoic and Palaeozoic. In Iraq. Lexique Strat. Int., Paris, Asie lOa. 333 pp. Hovelacque, M. 1900. Album de microphotographies de roches sedimentaires. 14 pp., 69 pis and expl. Paris. Jaffrezo, M. 1974. Les algues calcaires du Jurassique superieur et du Cretace des Corbieres (2me Partie). Rev. Micropaleont., Paris, 17 : 23-32. Johnson, J. H. 1964. Paleocene calcareous red algae from northern Iraq. Micr op ale ontology, New York, 10: 207-2 16. Oswald, F. 1906. A Treatise on the Geology of Armenia. 516 pp., 35 pis. Beeston, Notts., England. Privately published. (Printed text of thesis, not illustrated, same title, issued 1905, 425 pp. London). Sokac, B. & Nikler, L. 1973. Calcareous algae from the Lower Cretaceous of the environs of Niksic, Crna Gora (Montenegro). Palaeont. jugosl., Zagreb, 13. 57 pp., 16 pis. Swinnerton, H. H. 1958. (Obituary of F. Oswald). Nature, Land. 182 (4649) : 1 549. 1959. (Obituary of F. Oswald). Proc. geol. Soc., London, 1572 : 1 5 1-1 52. J. A. Moy-Thomas and his association with the British Museum (Natural History) P. L.Forev Department or Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 5BD B. G. Gardiner Department of Biology, Sir John Atkins Laboratories, Queen Elizabeth College, Campden Hill Road, London W8 7AH Introduction The British Museum (Natural History), as a national institution for the repository and care of natural history specimens and furtherance of their understanding, owes its efficiency not only to those employed there but also to that very large body of dedicated amateurs and professional scientists who collect, donate and work on the collections. Over the last hundred years their contributions have been very significant and there has been a history of close collaboration. For instance, Richard Lydekker (1849-1915), who worked for the Indian Geological Survey, compiled the catalogues of the Museum's collections of fossil mammals, reptiles and birds. A. W. Wrigley (1885-1953), a draughtsman, worked closely with Museum staff to produce many papers on Tertiary molluscs and Eocene foraminifera. It seems appropriate in this, the Museum's centenary year, to acknowledge the contri- butions made by our non-Museum colleagues. One such was J. A. Moy-Thomas, who was closely associated with the fish section immediately prior to the second world war. His association with the Museum was unfortunately brief, but during those few years he collected and examined a variety of interesting fossil fishes in our collections and, above all, typified that essential collaboration between Museum staff and others which has so often proved rewarding. J. A. Moy-Thomas James Moy-Thomas, eldest son of Mr and Mrs Alan Moy-Thomas, was born in 1908. Billy, as he was affectionately known, was sent to Harrow from whence he obtained an open scholarship to Christ Church, Oxford. After three years of study he was awarded a first class honours degree in zoology in 1930. The following session he remained up at Oxford and attended classes in geology. For most of this period his tutor was Dr G. de Beer, with whom he was later to form a close friendship. In the summer of 1932 at the instigation of de Beer he visited the Zoology Department of the University of Glasgow. There he enjoyed the hospitality of Professor Graham Kerr and was encouraged to re-examine developmental stages of Polypterus, making use of the specimens brought back by the late John Samuel Budgett, including the material used by Kerr himself in 1907. Moy-Thomas's work on Polypterus was published in 1934. In the meantime he worked as an assistant to Professor Walter Garstang at Leeds where he met and married Miss Joy Mitchell. This period also saw the publication of his first paper, in collaboration with T. H. Harrison of Pembroke College. It consisted of a short note to Nature on the St Kilda house mouse. Two more papers followed in 1933 and in the ensuing years he published a further 34 papers. He returned to Oxford in the summer of 1933 to the post of University demonstrator in the Department of Zoology and Comparative Anatomy. This was in essence Bull. Br. Mus. not. Hist. (Geol.)35 (3) 1 3 1-144 Issued 29 October 198 1 131 132 P. L. FOREY & B. G. GARDINER a research fellowship and in his first year back at Oxford J. A. Moy-Thomas attended various lecture courses, including those of Mr J. Z. Young who was later to become a close friend. During the next seven years he also became a good friend and colleague of Dr E. I. White of the British Museum (Natural History), and they both frequently attended meetings of the newly-formed Tetrapods Club (a dining club in London, founded in 1930 for those interested in vertebrate zoology). E. I. White introduced him to Sir Arthur Smith-Woodward (erstwhile keeper of the Department of Geology) who in 1935 gave Moy-Thomas permission to use his many unpublished notes on fossil fishes. In 1936 Moy-Thomas shared the Rolleston Memorial Prize with B. G. Maegraith (a fellow of Exeter College) and in the following year he was re-elected to the post of University Demonstrator and lecturer. Shortly afterwards he became the first holder of the E. T. Browne Fellowship at Queen's College. In each of the years 1935, 1937, 1938 and 1940 he received small sums of money from the Godman Exploration Fund (20 in 1935, 30 in the subsequent years) to enable him to collect fossils for the British Museum (Natural History). He was a diligent collector with infinite patience who furnished the Museum with some 127 specimens of fossil fishes (49 in counterpart), mainly from Glencartholm, only 9 of which were purchased. On many of his collecting trips he was accompanied by his wife but at Glencartholm in 1933 he was helped by Mr W. S. Bullough (later Professor of Zoology at Birkbeck College, London). Ironically it was Bullough who found the specimens of Tarrasius on which Moy-Thomas's 1934 Zoological Society paper was based. In 1939 Moy-Thomas joined the English-Norwegian-Swedish (E.N.S.) expedition to Spitzbergen (not Greenland as is erroneously reported in his obituary notices) which was the result of an intimate collaboration between the British Museum (Natural History), the Paleozoological Department of the Riksmuseum in Stockholm and the Paleontological Museum in Oslo (Fig. 1). The initiator was Professor E. A. Stensio and the expedition was financed by all three countries. The English members were Dr E. I. White, J. A. Moy-Thomas, J. Brough and W. N. Croft. The Norwegians were Professor A. Heintz, Sven F0yn and the student Aarhus. The Swedish members were Professor E. A. Stensio, leader of the expedition, E. Jarvik and G. Wangsjo, and they were accompanied by Dr N. Delia of Riga. At the beginning of the war Moy-Thomas started in the intelligence service but soon volunteered for flying work. Having successfully completed his tour of operations as a night fighter observer/navigator he was posted to R.A.F. Defford which was a non-operational unit concerned with the development and pre-service testing of air interception apparatus (airborne radar). This apparatus was being developed by the Telecommunication Research Establishment (T.R.E. Malvern). Here, together with Professor Derek Jackson F.R.S. (Spectroscopy, Oxford), he flew as a radar observer in Mosquitoes and Beaufighters. Moy-Thomas was killed in a motor accident while on duty on February 29, 1944. He was a sociable person with a great gaiety, sense of humour and zest for life. Everyone with whom he came into contact seems to have liked him. He was a very good golfer (handicap 3), enjoyed a game of darts and was an avid stamp collector. At Oxford he is remembered as a most successful teacher and tutor and at the British Museum (Natural History) as an endearing character with a sense of humour and a first-class memory. He greatly admired both Professors W. Garstang and E. S. Goodrich and for the latter he erected the genus Goodrichia (a large shark from Glencartholm) in 1936. The name was unfortunately pre-occupied (by a mollusc) and it was subsequently changed posthumously to Goodrichichthys (Moy-Thomas, 1951). During his short career J. A. Moy-Thomas worked mainly on fossil fishes (32 papers), particularly shark-like forms and palaeoniscids. However, he still found time for experi- mental work and as late as 1940 was examining the dermal bones of the skull of the trout in an effort to determine whether or not their development was influenced by the neuromast organs. Although one of his longest papers was on coelacanths perhaps his most notable was on Palaeospondylus. J. A. MOY-THOMAS 133 Fig. 1 J. A. Moy-Thomas (left) shown here with W. N. Croft (1915-1953, a palaeobotanist in the British Museum (Natural History)) during the English-Norwegian-Swedish expedition to Spitzbergen (1939). Croft, like Moy-Thomas, died at a young age. He collected a large number of Old Red Sandstone fishes and Devonian plants for the Museum. Interpretation of Palaeospondylus Genus PALAEOSPONDYLUS Traquair, 1890 Fig. 2 A- c TYPE SPECIES. P. gunni Traquair, 1 890. Towards the end of his short career Moy-Thomas turned his attention to Palaeospondylus gunni Traquair, an enigmatic fossil from the Middle Old Red Sandstone of Caithness. This little fossil, barely reaching 60 mm in length, has been a palaeontological conundrum since its first description by Traquair (1890); Dean (1904 : 425) remarks 'Palaeospondylus, like Gloster, seems to have been born to bite the world'. Two Caithness men, Alexander and Marcus Gunn, 'delivered' Palaeospondylus into the scientific world by bringing it to the attention of Traquair. Their collecting efforts at Achanarras were recognized by Traquair who named it after them. Palaeospondylus has certainly had a 'bite' at many palaeon- tologists and zoologists, who have reacted by referring it to one or other of the many fish groups or, in some cases (Gill 1896; Dean 1898, 1900), to specially-erected classes or subclasses. When first described Palaeospondylus was regarded as an agnathan and this opinion received some initial consensus (Traquair 1890, 18930, 18936, 1894, 1897; Howes 1892; Woodward 1892, 1898; Dean 1895; Stensio 1927; Bulman 1931; Ayers 1933; White 1935). This consensus was challenged on numerous occasions. Sollas & Sollas (1903) suggested it to be an elasmobranch. Kerr (1900) and Miller (1930) compared it to a larval 134 P. L. FOREY & B. G. GARDINER dipnoan, Dawson (1893) to a larval amphibian, Kyle (1926) to a larval herring, Dean (1904) to a holocephalan, and Jarvik (1980) regarded it as a larval Osteolepis. Huxley, it is said (Dean 1900), thought it to be a larval Coccosteus while Dean (1896, 1898, 1900) and Abel (1912) regarded it as a larval arthrodire. Finally, Moy-Thomas (1940) crystallized its placoderm relationships by suggesting it to represent a stegoselachian (a naked placoderm). Most modern text books (e.g. Parker & Haswell 1963, Romer 1966) deal with Palaeospondylus as an appendix to the placoderms. Customarily, uncertainty about relationships of fossils arises from material which is poorly preserved or scanty, or both. But Palaeospondylus is known by, literally, thousands of specimens, almost all of which have been found at a single slate quarry on Achanarras Hill where it is distributed through several beds (Rayner 1963). At least one specimen was found in contemporaneous strata at Niandt (Traquair 1909) and a further example in the Sandwick Fish Bed at Cruaday Hill, Orkney (Trewin 1976). The Museum collections contain some 450 specimens. It is true that many are poorly preserved, leading Traquair (1890 : 485) to describe the head as a 'flat crushed mass of bony bars'. Other specimens, however, are preserved almost as well as other representatives of the Achanarras fish fauna (Coccosteus, Homostius, Cheirolepis, Cheir -acanthus, Mesacanthus, Rhadinacanthus, Diplacanthus, Pterichthyodes, Dipterus, Osteolepis, Glyptolepis), the relationships of which have rarely been in doubt. Palaeospondylus has also been the subject of various preparation techniques, examples of which are in the Museum's collections. Sollas & Sollas (1903) used Palaeospondylus as one of their first experimental materials to produce wax-plate reconstructions (P.9856, P.9859-61). Bulman (1931) produced whole mount preparations (P. 1 6 1 20-5) and one specimen (P.22393) has been prepared by the acid transfer technique of Toombs & Rixon (1950). So the plea that the material is poor or insufficient fails in this case. We suggest, instead, that the problem with Palaeospondylus is one generally inherent in fossils: that is, they are data 'in search of interpretation' (Nelson 1978 : 329). There are three main aspects to this problem. The first is incompleteness, in the sense that only the hard parts or a limited amount of the soft parts are preserved, thus reducing the amount of com- parative information available for interpretation. Secondly, 'a fossil is meaningless until it can be interpreted in the light of a Recent model' (Patterson 1977 : 621). Finally a fossil is, by its very nature, in danger of being furnished with an ancestral status by over-zealous palaeontologists. The problems posed by Palaeospondylus fall into all three categories. The last-mentioned area was particularly characteristic of the early studies of Palaeospondylus, as the following remarks made by Dean (1904) show: 'For if the remains of Palaeospondylus are so poorly preserved that they cannot be definitely described, why do we continue to add papers to the troublesome literature? The only possible excuse 'is that the creature is seductive, full of suggestions as to the origin of the gnathostomes, and the mode of evolution of the jawless vertebrates.' We shall begin by examining the earliest theories of the relationships of Palaeospondylus with the agnathans. Traquair (1890) noted the similarity in the shape of the head between Palaeospondylus and Myxine and, considering the period in which Traquair was working, it is not surprising that a relationship between the two should have been suggested. Toward the end of the nineteenth century palaeontology, through its concern with time, was beginning to emerge as the 'authority' on questions of phylogeny. Darwin predicted that, were the fossil record more carefully examined, progenitors of modern groups and links between groups would be found. So, one searched the rocks for suitably primitive and stratigraphically suitable candidates. Palaeospondylus was judged to be such an approximation to an ancestor of modern agnathans or, more specifically, to myxinoids (Traquair 1890; Bulman 1931). Palaeospondylus is naked like modern agnathans and was originally thought to have no jaws or fins. Its nakedness made it a better candidate than other jawless vertebrates the heavily armoured heterostracans and osteostracans. Several characters have been used to suggest relationship between Palaeospondylus and Recent agnathans: 1 -no jaws (Traquair 1890, 18930, b, 1894; Stensio 1927); 2 -no limbs (Traquair 1890, 1893a, b, 1894, 1897; Woodward 1892; Stensio 1927; Bulman 1931); J. A. MOY-THOMAS 135 3 - cirri surrounding a circular opening at the anterior end of the head, interpreted as a mouth (Woodward 1 892; Traquair 1 893a) or a nasal opening (Traquair 1 893/7, 1 894; Stensio 1927; Bulman 193 1); 4 - single median recess at the anterior end of the cranium, interpreted to house a single nasal organ (Woodward 1892; Traquair 18936; Stensio 1927; Bulman 1931); 5 -V-shaped branchial pouch supports behind the head (Bulman 1931); 6 - dichotomized radials (often incorrectly referred to as fin rays) in caudal region (Traquair 1894); 7 - protocercal tail (Traquair 18930, b, 1894, 1897); 8 -no discrete ossifications in the braincase (Traquair 1894). Characters 1 and 2 are primitive, present in any non-gnathostome, and are not therefore agnathan characters any more than they are echinoderm or nematode characters. If the 'mass of bony bars' lying on the (presumed) ventral surface are not evidence of visceral arches then what are they? Three interpretations have been offered; that they are labial cartilages (Woodward 1892), that they are, in fact, ridges on the underside of the neurocranium and represent interbranchial ridges similar to those in cephalaspids (Stensio 1927), or that they are remains of various cartilages supporting the tongue as in myxinoids (Bulman 1931). Most authors agree, as do we, that the 'bony bars' are visceral structures and that many are free or articulate with the braincase. This would seem to rule out any comparison with either lampreys or hagfishes, in which the visceral skeleton is unjointed and continuous with the neurocranium. Only gnathostomes show a jointed visceral skeleton which articulates with the neurocranium. Character 3 has been commonly used for suggesting agnathan relationships, particularly with myxinoids. The more popular interpretation is that the 'cirri' (rostralia) surround a circular opening which represents a single nasal opening. Only hagfish, amongst agnathans, have a series of rostral cartilages (of different lengths) reaching in front of the nasal region. Myxine also has a row of cartilaginous rings around the nasal tube (Cole 1 909 : fig. 1 ). Moy-Thomas (1940) examined several well-preserved specimens (P. 22394, P.22401, P.22410) and found that considerable variation exists between individuals in the size and, to a lesser extent, the number of 'cirri'; he suggested that these together represented a fene- strated capsule(s). We would concur with this interpretation. As to character 4, a median recess at the anterior end of the cranium is found in a variety of craniate embryos and is not an agnathan character. Character 5 is an interpretation of structures which have been otherwise interpreted as pectoral girdles and/or fins (Moy- Thomas 1940; Dean 1896). Character 6 is also found in elasmobranchs and lungfishes (Furbringer 1 904) and character 7 was shown to be a mistaken observation (Traquair had restored the tail upside down). The tail is now regarded as being asymmetrical with a slightly larger lower lobe. In any event, a protocercal tail is found in Recent lungfishes (Miller 1930). Character 8 is difficult to evaluate since the nature of the preserved material is not clear. Microscope sections show no structure and the chemical composition, like that of other fossils from Achanarras, 'now consists of coal' (Sollas & Sollas 1903 : 273). The skeletal material has been interpreted as bone (Traquair 1890), or as calcified cartilage (Traquair 1 839a), but there is no evidence to favour either of these suggestions. So, in our opinion, not one piece of evidence has been produced suggesting Palaeo- spondylus to be an agnathan and interpretation on an agnathan model (hagfish or lamprey) is not justified. The presence of ring centra would also militate against agnathan relationships. Hagfish have no chondrification or ossification around the notochord, lampreys have cartilaginous dorsal arcualia and the only evidence of a vertebral column in ostracoderms is the impressions of (presumably) neural arches in some cephalaspids (Janvier 1980). The presence of ring centra in Palaeospondylus as an agnathan ancestor was explained by assuming that absence of skeletal ossification is a derived condition of Recent agnathans. We do not wholly share this view. Hagfishes, which we regard as the sister group of lampreys and gnathostomes (L^vtrup 1977; Hardisty 1979; Janvier & Blieck 1979), never possessed bone or paired fins in their history. This might also be true of lampreys, in which case we would regard lampreys and anaspids as the sister group of osteostracans and gnathostomes (Janvier, personal communication). 136 P. L. FOREY & B. G. GARDINER We accept that Palaeospondylus has visceral arches, albeit they are difficult to interpret, and agree with Moy-Thomas (1940) that there is evidence for both pectoral and pelvic fins. We therefore believe that Palaeospondylus is either the sister-group of gnathostomes or a member of some gnathostome subgroup. Several suggestions have been offered (p. 134). Sollas & Sollas (1903) considered it to be an elasmobranch but one that 'proceeded in its subsequent development along an independent course, losing its limbs, if it ever possessed them, and acquiring a highly organised vertebral column, homoplastic in character with that of cyclo-spondylous Selachians' (1903:290-291). The main reason given by these authors for elasmobranch affinities is the general similarity in the shape of the head: the eye is situated immediately in front of the otic capsule, and there is a saddle-shaped ledge on the ventral surface of the neurocranium, marking the position of the pituitary body. We find nothing particularly elasmobranch about these features: we cannot confirm the presence of a 'saddle-ledge' but note that the ventral profile ofSqualus is very similar to that of larval Acipenser and Neoceratodus (de Beer 1937). The particular shape of the ventral profile may thus be no more than a general early ontogenetic feature of gnathostomes. Dean, in perhaps light-hearted mood (1904:425), suggested that Palaeospondylus is a holo- cephalan. He based this suggestion on four characters: continuous dorsal fin (no evidence), protocercal tail (incorrect observation), ring vertebrae (known elsewhere - chondrichthyes, larval teleosts, sarcopterygians) and a huge head. Moy-Thomas regarded Palaeospondylus as an adult and to be a stegoselachian, a group of placoderms in which there is little development of armour - Stensioellidae and Rhenanida. He chose to make comparisons with these but also with acanthodians, presumed close relatives of placoderms (Watson 1937), and arthrodires. Moy-Thomas's model was therefore a placoderm + acanthodian morphotype (an aphetohyoidean, established by Watson as a grade group). Moy-Thomas mentioned eight characters to support his argument, citing precedents within selected placoderms and acanthodians: 1 - heterocercal tail; 2 - anterior position of the pelvic fins (like Pseudopetalichthys and Rhamphodopsis - a ptyctodont); 3 - ventral mouth and small size of lower jaw; 4 - palatoquadrate ossified in more than one piece in which the 'tauidion' (Sollas & Sollas 1903) represents the medially united anterior ossifications; 5 - ring-like centra (Gemuendina and Pseudopetalichthys); 6 - short occipital region (like Jagorina, but most placoderms have long occipital regions); 7 - well-developed rostral region (like Nessariostoma); and 8 - hyomandibular not supporting jaws (but the hyomandibular is involved in jaw suspension of most placoderms (Miles 1971) and the jaw suspension of acanthodians is known to have a suspensory hyomandibular (Miles 1973)). Once again, all these characters can be matched outside placoderms and acanthodians and we find no placoderm characters in this list. Jarvik (1980) has recently suggested that Palaeospondylus is a larval Osteolepis. His argument is in two stages. He first notes the similarity between Palaeospondylus and anuran tadpoles (presence of ossified ring-like centra and external shape of the tail). He then argues (1980 : 218) that ' . . . because osteolepiforms have been shown to be close to the ancestry of the Anura it is tempting to suggest that Palaeospondylus may be a larva of Osteolepis macrolepidota, an osteolepiform which is also common in the flagstones at Achanarras'. To this we would make three comments: the similarities between anuran tadpoles and Palaeospondylus are not unique; immediate relationship between anurans and osteolepi- forms is not beyond doubt (Rosen et al. 1981); Osteolepis is very rare at Achanarras, there being only approximately 16 specimens known (Trewin, personal communication). One of the most distinctive structures of Palaeospondylus are the so-called 'post-occipital lamellae' which are readily visible and which, we believe, offer a clue about relationships. The 'post-occipital lamellae' are represented by a pair of rods which lie on either side of the anterior centra. Amongst living fishes there are very few comparable structures. Moy- Thomas regarded them as part of the branchial arch series although 'why they are so much enlarged is still a mystery' (1940 : 401). In this he finds agreement with Dean (1896) and Sollas & Sollas ( 1 903). Other interpretations include: parachordals (Jaekel 1 927), pronephric lamellae of cephalaspids (Stensio 1927), the posterior lingual cartilage of a myxinoid J. A. MOY-THOMAS 137 (Bulman 1931), a rudimentary dorsal shield (Woodward 1892), elements of a shoulder girdle (Kyle 1 926) or a cranial rib of a dipnoan (Kerr 1 900). The last interpretation seems particularly promising to us since the comparable structures in Recent dipnoans are so similar in shape and position, and are distinctively large. The structures in question have rounded, presumably articulatory, heads (Fig. 2s) and always lie against the posterior edge and slightly on the ventral surface of the neurocranium. This was therefore their position in life. They are associated (? articulated) with smaller, angulated structures (branchial arches - Sollas & Sollas 1903, Moy-Thomas 1940), which may be interpreted as occipital neural arches. Our comparisons of cranial ribs seem most favourable with Protopterus and Lepidosiren (Agar 1906 : figs 9, 16). We agree with Kerr (1900, 1919) and Miller (1930) that these enlarged structures are cranial ribs and this allows us to interpret Palaeospondylus as a lungfish. Our interpretation of the visceral structures is given in Fig. 2e and is based on comparisons with the illustra- tions of lungfish larvae provided by Agar (1906), Kerr (1919) and Fox (1965). The many specimens of Palaeospondylus show considerable variation but we have found it impossible to relate differences in structure to absolute size. This is chiefly because the several visceral elements are very difficult to interpret; it is not easy to decide where there are points of articulation or, in some cases, whether an element is separate or not. Diagenesis must have affected such an obviously delicate animal. For this reason we do not intend to describe the skull. Instead we will simply point out several important areas of agreement and disagree- ment between the restoration given here and those offered by Moy-Thomas (1940 : figs 2, 4). We agree that the 'rostralia' are expanded distally and that they may be fused distally, thereby forming together a fenestrated capsule. The resemblance between this and (partially macerated) nasal capsules of Lepidosiren was noted by Kerr (1900) and Miller (1930 : fig. 5). The 'tauidion' (Sollas & Sollas 1903) is a prominent structure which we compare to a dipnoan vomer. We agree that the paired elements which lie beneath the otic capsules and converge anteriorly are ceratohyals (always large in dipnoans and urodeles), and that the small unpaired element immediately behind the anterior ends of the ceratohyals is a basibranchial. We disagree with Moy-Thomas over the interpretation of those visceral elements anterior and lateral to the otic capsule. These have been the most problematical elements, interpreted differently by Sollas & Sollas (1903), Bulman (1931) and Moy-Thomas (1940). We believe that the 'hyomandibular' is nothing more than the thickened edge along the anterolateral margin of the otic capsule and may be compared to the otic process of the palatoquadrate; that the 'gammation' and posterior trapezial bar of Sollas & Sollas (1903) compose a single element (specimen P.22392) representing the quadrate region of the palate; and that the 'anterior trapezial bar', 'pregammation' and perhaps the 'hemidome septum' are palatal elements. Since the last-mentioned elements are the most variable they might represent developing tooth ridges (cf. Kerr 1919 : fig. 164c). Like Moy-Thomas, we consider that the palatoquadrate is fused to the neurocranium and find it interesting that Moy-Thomas chose to make his comparison of the skull with the urodele Hynobius. Identification of the lower jaw is difficult. It is possibly represented by the outer curved element, representing a Meckelian ossification/chondrification which forms the outer edge of the 'hemidome'. This is expanded posteriorly and articulated with that element identified here as the quadrate. Anteriorly it curves to meet its antimere immediately in front of the vomer, as it would in a dorsally flattened head. Moy-Thomas rejected dipnoan affinities of Palaeospondylus, chiefly on the ground that he considered it to be an adult. He noted the existence of centra and the advanced degree of ossification. The nature of the skeleton has never been established beyond the fact that it is chemically like coal. Centra are known in larval teleosts indeed, this was one of the reasons that led Kyle (1926) to consider it as a larval herring. If Palaeospondylus were an adult we find it unlikely that fin rays would be absent, that any trace of dermal covering would be absent, that the girdles would be such insignificant structures and that the neurocranium would be so open and trough-like dorsally. These features are those of larvae. The logical conclusion is that Palaeospondylus is a larval Dipterus, the only lungfish 138 P. L. FOREY & B. G. GARDINER pectoral girdle pelvic girdle fused anterior palatine elements palatoquadrate (in part) posterior portion of palatoquadrate hyomandibular gill arch elements nasal capsule vomer lower jaw tooth plate of palatoquadrate quadrate otic process ceratohyal (+ hypohyal ?) basibranchial occipital neural arch cranial rib 2 mm occipital neural arch cranial rib otic process nasal capsule pectoral girdle labial cartilages ceratohyal Meckelian cartilage Fig. 2 A, reconstruction of Palaeospondylus. Skull in dorsal view, posterior portion of trunk and tail twisted to appear in right lateral view. Based on Moy-Thomas (1940: fig. 7), skull redrawn. B, restoration of the head of Palaeospondylus in ventral view. Labelling on right-hand side represents our interpretation, that on the left Moy-Thomas's. C, Lepidosiren, stage 38. Reconstruction of the head of the embryo in lateral view, tooth plates omitted. After Agar (1906: pi. 3, fig. 16). present in the fauna. Centra are not normally recorded for Dipterus but Jarvik (1952) records them and specimen P. 106 13 shows centra in at least the tail region. The tail of Dipterus, unlike that of many Dipnoans, is also asymmetrical. We can, however, suggest no morpho- logical character which would refer it specifically to Dipterus. We recognize that our interpretation of Palaeospondylus fossils, even at such an elementary level as deciding whether a structure is fused or articulated, is governed by our J. A. MOY-THOMAS 139 initial choice to use a particular model. Such is the nature of palaeontology. The difference between our attempt and those of most other workers who have considered Palaeospondylus (with the exception of Kerr) is that we have attempted to identify a synapomorphy with a Recent group, rather than rely on obviously primitive features (as did Traquair and Bulman), features which are not characters of groups (Moy-Thomas) or use other fossil groups as models (Stensio). We leave Palaeospondylus as larval dipnoan to 'seduce' or to 'bite' other palaeontologists. Jamoytius and Moythomasia Two genera of fossil fishes have subsequently been named after Moy-Thomas. The first was Jamoytius by E. I. White in 1946 and the second Moythomasia by W. Gross in 1950. Genus JAMOYTIUS White, 1946 Fig. 3 DIAGNOSIS. See White 1946. TYPE SPECIES. J. kerwoodi White 1946. REMARKS. When first described by White (1946 : 93) the naked Jamoytius was considered not only as a likely ancestor for amphioxus but also for the Craniata. White consequently erected the new order Euphanerida to incorporate it. These suggestions, however, did not meet with universal favour and Jamoytius was subsequently regarded as a larval thelodont by Wangsjo (1952 : 566) and as an anaspid by Robertson (1953 : 734). Smith (1957 : 394) and Stensio (1958:239) concurred with Robertson and suggested that 'the carbonized remains of the body muscles (myocommata)' were scales. Tarlo (1960 : fig. 5), after a re- examination of the holotype, not only confirmed both Smith's (1957) and Stensio's (1958) contentions but also claimed the presence of ridge scales. More recently Newth (in Young 1962 : 128) has suggested it to be the ammocoete larva of an ostracoderm and Wickstead (1969 : 422) that it corresponded to a metamorphosing amphioxus. Ritchie (1960, 1968) has published accounts based on new material, in which he claimed to have confirmed the presence of scales. Jamoytius kerwoodi White 1 946 1946 J. kerwoodi White : 89. 1960 J. kerwoodi White; Tarlo : 1 13, fig. 5. 1960 J. kerwoodi White; Ritchie : 647, fig. 1. 1968 J. kerwoodi White; Ritchie : 26; pis 3-6. DIAGNOSIS (emended). A naked cyclostome with diphycercal tail and branchial basket. REMARKS. We cannot find any trace of scales on the specimens examined (BM(NH) and Royal Scottish Museum). Furthermore, the carbonized remains bear no resemblance to anaspid scales. The so-called scales seen by Ritchie (1960, 1968) were only observed by him after the specimens had been treated. No such structure is visible on any specimen we have examined and we suggest that both Tarlo (= Halstead) and Ritchie were mistaken in their observations. Futhermore, no good evidence has been presented against regarding the segmental structures as the remains of body muscles. In no specimen is there anything more than carbonized remains and occasionally a rather amorphous, tarry, surface structure. There is likewise no evidence for a lateral fin fold either on the holotype (cf. White 1946 : fig. l)pron RSM 1966. 3.1 (cf. Ritchie 1968 : pi. 4). In our estimation the fossil looks very similar to the present-day lamprey and like it has a branchial basket (Fig. 3) with horizontal struts and a diphycercal tail. The branchial basket has no more than seven openings and the appearance of paired structures (eyes) and an annular cartilage argues against it being a larval amphioxus. 140 P. L. FOREY & B. G. GARDINER 10 mm branchial basket Fig. 3 Jamoytius kerwoodi White. Sketch of anterior portion of head as preserved in BM(NH) P.47787, showing the lamprey-like branchial basket. Genus MO YTHOMASIA Gross, 1950 [ = Aldingeria Gross 1 942 : 43 1 , non Moy-Thomas 1 942] DIAGNOSIS. See Gross, 1942 : 431. TYPE SPECIES. M. perforata (Gross). REMARKS. The name Aldingeria was first used by Moy-Thomas in October 1942 for a Carboniferous palaeoniscid from East Greenland. Two months later in December of that year the same name was used by Gross (1942 : 43 1) for a very different palaeoniscid from the Upper Devonian of the Baltic. When Gross realized that Aldingeria Moy-Thomas took priority over Aldingeria Gross he replaced his genus with the name Moythomasia (Gross 1950 : 145) in honour of the British palaeontologist. When first described by Gross, Moythomasia perforata appeared to be just another palaeoniscid. Fortunately, however, this genus also occurs in the calcareous Devonian rocks of Australia (M. durgaringia Gardiner & Bartram 1977) and it has turned out to be far more interesting than ever Gross could have imagined. Moythomasia durgaringia Gardiner & Bartram 1977 DIAGNOSIS. See Gardiner & Bartram 1977 : 238. REMARKS. This species closely resembles Mimia toombsi Gardiner & Bartram, from which it only differs significantly in having a short ascending process. Moythomasia is regarded as the sister-group of the Actinopteri (Rosen et al. 1981), sharing with them a differentiated propterygium in the pectoral fin, a pelvic fin exclusively supported by preaxial radials, fringing fulcra, ganoid scales, acrodin caps on marginal teeth, an endopterygoid, a dentary with a sensory canal and the absence of the jugal/infraorbital canal junction. Moythomasia and Mimia do not possess a stem to the parasphenoid, or a process on the post-temporal. The ventral fissure and otico-sphenoid fissures were cartilage- filled and not completely bridged by dermal bone (i.e. the parasphenoid), and the lateral occipital fissure is perichondrally lined in both. All of these features suggest that these two genera are the sister-group of all other actinopterygians, with the exception of Cheirolepis which lacks acrodin caps on its teeth and fringing fulcra, and Polypterus which lacks fulcra and ganoid scales. J. A. MOY-THOMAS 141 Conclusion During Moy-Thomas's life-time significant advances were made in our understanding of fossil fishes. New methods of examining fossils were introduced, including acid preparation, the use of fine dental hammers, serial sectioning and microscope examination under liquids of varying refractive indices. And there was also a concomitant influx of literature. Perhaps Moy-Thomas's most important contribution was to summarize this information in the light of his own research, and to embody it in his book Palaeozoic Fishes (1939). Beyond this, however, he was typical of those numerous individuals who rightfully use this National Museum. We hope that the next hundred years will produce as many fruitful collaborations as the last. Below, we print a complete bibliography of Moy-Thomas's contributions. Bibliography of Moy-Thomas Beer, G. R. de & Moy-Thomas, J. A. 1935. On the skull of Holocephali. Phil. Trans R. Soc., London, (8)224:287-312. Harrison, T. M. & Moy-Thomas, J. A. 1932. St Kilda House Mouse. Nature, Land. 129 : 1 3 1 . Moy-Thomas, J. A. 1933a. The anatomy and affinities of Tarrasius problematicus Traq. Nature, Lond. 132: 171-172. 1933-35. Notes on the Types of fossil fishes in the Leeds City Museum. I - Elasmobranchi. Proc. Leeds phil. lit. Soc. 2:451-454 (19336). II- Acanthodii, Dipnoi and Crossopterygii. loc. cit. 3: lll-116(1935a). 1934a. Notes on the development of the chondrocranium ofPolypterus senegalus. Q. Jl microsc. Sci., London, 76 : 209-230. 19346. The structure and affinities of Tarrasius problematicus, Traquair. Proc. zool. Soc. Lond. 1934 : 367-376. 1934c. On the teeth of the larval Belone vulgaris, and the attachment of teeth in fishes. Q. Jl microsc. Sci., London, 76 : 481^*98, 1 pi. 19356. A synopsis of the coelacanth fishes of the Yorkshire Coal Measures. Ann. Mag. nat. Hist., London, (10) 15 : 37-46, 3 pis. 1935c. The coelacanth fishes of Madagascar. Geol. Mag., London, 72 : 2 1 3-227. \935d. The structure and affinities of Chondrenchelvs problematica, Tr. Proc. zool. Soc. Lond. 1935:391-403. 1935e. On the Carboniferous shark Peirodus patellijbrmis, M'Coy. Proc. Leeds phil. lit. Soc. 3:68. 1935/ Recent trends in the study of fossils. The Listener, London, Feb. 27 1935 : 355-356, 1 pi. 1936fl. The evolution of the pectoral fins of fishes and the tetrapod fore-limb. Sch. Sci. Rev., London, 68: 592-599. 19366. On the structure and affinities of the Carboniferous cochliodont Helodus simplex. Geol. Mag., London, 73 : 488-503, 2 pis. 1936c. The structure and affinities of the fossil elasmobranch fishes from the Lower Carboniferous rocks ofGlencartholm, Eskdale. Proc. zool. Soc. Lond. 1936 : 761-788. \936d. The evolution of the elasmobranchs. Rep. br. Ass. Advmt Sci., Blackpool, 1936 : 367. 1937a. On the Carboniferous fish Eucentrurus paradoxus, Traquair. Geol. Mag., London, 74: 183-184. 19376. The palaeoniscids from the Cement Stones of Tarras Waterfoot, Eskdale, Dumfriesshire. Ann. Mag. nat. Hist., London, (10) 20 : 345-356. 1937c. The Carboniferous coelacanth fishes of Great Britain and Ireland. Proc. zool. Soc. Lond. 1937:383^15. 1938a. Carboniferous palaeoniscids from Northumberland and Berwickshire. Geol. Mag., London, 75: 308-3 18,1 pi. 19386. A revision of the fishes referred to the genus Canobius from localities other than Glencartholm. Ann. Mag. nat. Hist., London, (1 1)9 : 291-299. 1938c. On the teeth of a new elasmobranch from the Calciferous Sandstones ofGlencartholm, Eskdale, Dumfriesshire. A nn. Mag. nat. Hist., London, (1 1)9 : 612-613, 1 pi. 1938^. The problem of the evolution of the dermal bones in fishes. In Beer, G. R. de (ed.), Evolution. Essays on aspects of evolutionary biology presented to Professor E. S. Goodrich on his seventieth birthdav : 305-3 19. Oxford. 142 P. L. FOREY & B. G. GARDINER 1939a. The early evolution and relationships of the elasmobranchs. Biol. Rev., Cambridge, 14: 1-26. 19396. Palaeozoic Fishes, xviii + 149 pp. New York and London. 1940. The Devonian fish Palaeospondylus gunni Traquair. Phil. Trans. R. Soc., London, (8)230:391-413, pis 22-25. 1941. Development of the frontal bones of the rainbow trout. Nature, Lond. 147 : 681-682. 1942. Carboniferous palaeoniscids from East Greenland. Ann. Mag. nat. Hist., London, (11) 9: 737-759, pi. 12. 1951. On the name of the fossil shark Goodnc/na. Ann. Mag. nat. Hist., London, (12)4 : 304. & Dyne, M. B. 1938. The actinopterygian fishes from the Lower Carboniferous of Glencartholm, Eskdale, Dumfriesshire. Trans. R. Soc. Edinb. 59 : 437^80, 2 pis. & Westoll, T. S. 1935. On the Permian coelacanth, Coelacanthus granulatus. Geol. Mag., London, 72: 446^57. & White, E. I. 1934. The pectoral fin of Coelacanthus tingleyensis. Nature, Lond. 133 : 149. 1939a. Notes on some Carboniferous palaeoniscids. Ann. Mag. nat. Hist., London, (11)3:622-625. 19396. On the Palatoquadrate and Hyomandibula of Pleuracanthus sessilis Jordan. Geol. Mag., London, 76 : 459^63. White, E. I. & Moy-Thomas, J. A. 1940-41. Notes on the nomenclature of fossil fishes. I. Homonyms A-C. Ann. Mag. nat. Hist., London, (11)5: 502-507 (1940). II. Homonyms D-L. he. cit. 6 : 98-103 (1940). III. Homonyms M-Z. he. cit. (11)7: 395^00 (1941). References Abel, O. 1912. Grundzuge der Paldobiologie der Wirbeltiere. 708 pp. Stuttgart. Agar, W. E. 1906. The development of the skull and visceral arches in Lepidosiren and Protopterus. Trans. R. Soc. Edinb. 45 : 49-64, pis 1-3. Ayers, H. 1933. Palaeospondylus. Science, N. Y. 78 : 458-459. Beer, G. R. de 1937. The development of the vertebrate skull. 552 pp. Oxford. Bulman, O. M. B. 1931. Note on Palaeospondylus gunni Traquair. Ann. Mag. nat. Hist., London, (10)8: 179-190, pi. 4. Cole, F. J. 1909. A monograph on the general morphology of the myxinoid fishes, based on a study of Myxine. Part III - Further observations on the skeleton. Trans. R. Soc. Edinb. 46 : 669-680, 1 pi. Dawson, J. W. 1 893. Some salient points in the science of the earth. 499 pp. London. Dean, B. 1 895. Fishes, living and fossil. 300 pp. New York. 1896. Is Palaeospondylus a cyclostome? Trans. N. Y. Acad. Sci. 15 : 101-104. 1898. Remarks on the affinities of Palaeospondylus gunni in reply to Dr R. H. Traquair. Proc. zool. Soc. Lond. 1898 : 343-347. 1900. The so-called Devonian lamprey, Palaeospondylus. Mem. N. Y. Acad. Sci. 2 : 1-32. 1904. Still another memoir on Palaeospondylus. Science, N. Y. 19 : 425-426. Fox, H. 1965. Early development of the head and pharynx of Neoceratodus with a consideration of its phylogeny. J. Zool., Lond. 146 : 470-554, pis 1-4. Fiirbringer, K. 1904. Beitrage zur Morphologie des Skeletes der Dipnoer, nebst Bemerkungen iiber Pleuracanthiden, Holocephalen und Squaliden. In Semon, R. (ed.), Zoologische Forschungsreisen in Australien und der Malayischen Archipel. I. Band: Ceratodus. IV. Liefrung. Denkschr. med.-naturw. Ges. Jena 4 : 423-510, pis 37^1. Gardiner, B. G. & Bartram, A. 1977. The homologies of ventral cranial fissures in osteichthyans. In Andrews, S. M., Miles, R. S. & Walker, A. D. (eds), Problems in Vertebrate evolution, 1 : 227-245. London & New York. Gill,T. 1896. Fishes, living and fossil. Science, N.Y.3: 909-917. Gross, W. 1942. Die Fischfaunen des baltischem Devons und ihre biostratigraphische Bedeutung. KorrespBl. NaturJVer. Riga 64 : 373-436. 1950. Umbenennung von Aldingeria Gross (Palaeoniscidae; Oberdevon) in Moythomasia n. nom. NeuesJb. Geol. Paldont. Mh., Stuttgart, 1950 (5) : 145. Hardisty, M. W. 1979. Biology of the Cychstomes. xiv + 428 pp. London. Howes, G. B. 1892. On the affinities and inter-relationships and systematic position of the Marsipobranchii. Proc. Trans. Lpool biol. Soc. 6 : 122-145, pis 8-10. Jaekel, O. 1927. Der Kopf der Wirbeltiere. Ergebn. Anat. EntwGesch., Munich & Berlin, 27 : 815-974. J. A. MOY-THOMAS 143 Janvier, P. (1980). Les Osteostraci de la formation de Wood Bay (Devonien inferieur, Spitsberg) el le probleme des relations phylogeniques entre Agnathes et Gnathostomes. 1 vols. Ph.D. Thesis (unpubl.), Univ. Pierre et Marie Curie, Paris. & Blieck, A. 1979. New data on the internal anatomy of the Heterostraci (Agnatha), with general remarks on the phylogeny of the Craniota. Zoologica Scr., Stockholm, 8 (4) : 287-296. Jarvik, E. 1952. On the fish-like tail in the ichthyostegid stegocephalians. Meddr GrQnland, Copenhagen, 114 : 1-19,21 pis. 1980. Basic structure and evolution of the vertebrates, 1. 576 pp. London. Kerr, J. G. 1900. The zoological position of Palaeospondylus, Traquair. Proc. Camb phil Soc 10 : 298-299. 1919. Text-book of embryology, 2. 591 pp. London. Kyle, H. M. 1 926. The biology of fishes. 390 pp. Hong Kong (T. F. H. edn). Lavtrup, S. 1977. Phylogeny of the Vertebrata. 330 pp. London. Miles, R. S. 1971. The Hoionematidae (placoderm fishes), a review based on new specimens of Holonema from the Upper Devonian of Western Australia. Phil. Trans. R. Soc London (B)263: 101-234. 1973. Relationships of acanthodians. In Greenwood, P. H., Miles, R. S. & Patterson, C. (eds), Interrelationships of fishes: 63-104. London. Miller, A. E. 1930. Note on the tail skeleton of Lepidosiren paradoxa, with remarks on the affinities of Palaeospondylus. Proc. zool. Soc. Lond. 1930 : 783-789, pi. 1. Moy-Thomas, J. A. 1940. The Devonian fish Palaeospondylus gunni Traquair. Phil. Trans. R. Soc., London, (B) 230 : 391-413, pis 22-25. 1942. Carboniferous palaeoniscids from East Greenland. Ann. Mag. nat. Hist., London (11) 2: 737-759, pi. 12. Nelson, G. 1978. Ontogeny, phylogeny, paleontology, and the biogenetic law. Syst. Zool., Washington, 27 : 324-345. Parker, T. J. & Has well, W. A. 1963. A text-book of Zoology, 6th edn, 2. xxiii + 758 pp. London. Patterson, C. 1977. The contribution of palaeontology to teleostean phylogeny. In Hecht, M. K. & Goody, P. C. (eds), Major Patterns of Vertebrate evolution : 579-643. New York. Rayner, D. H. 1963. The Achanarras Limestone of the Middle Old Red Sandstone, Caithness, Scotland. Proc. Yorks. geol. Soc., Leeds, 34 : 1 1 7-1 38, pis 1 7-18. Ritchie, A. 1960. A new interpretation of Jamoytius kerwoodi White. Nature, Lond. 188 : 647-649. 1968. New evidence on Jamoytius kerwoodi White, an important ostracoderm from the Silurian of Lanarkshire, Scotland. Palaeontology, London, 2 : 2 1-36, pis 3-6. Robertson, G. M. 1953. Some attempts at phylogeny of early vertebrates. Proc. Iowa Acad. Sci., Des Moines, 60: 725-737. Romer, A. S. 1966. Vertebrate Paleontology, 3rd edn. 468 pp. Chicago. Rosen, D. E., Forey, P. L., Gardiner, B. G. & Patterson, C. 1981. Lungfishes, tetrapods, paleontology and plesiomorphy. Bull. Am. Mus. nat. Hist., New York, 167 : 159-276. Smith, I. C. 1957. New restorations of the heads of Pharyngolepis oblongus Kiaer and Pharyngolepis kiaeri sp. nov., with a note on their lateral-line systems. Norsk geol. Tidsskr., Oslo, 37 : 373^402, 6 pis. Sollas, W. J. & Sollas, I. B. J. 1903. An account of the Devonian fish Palaeospondylus gunni Traquair. Phil. Trans. R. Soc., London, (B) 196 : 267-294, pis 16-17. Stensio, E. A. 1927. The Downtonian and Devonian vertebrates of Spitzbergen, Part 1. Family Cephalaspidae. Skr. Svalbard Nordishavet, Oslo, 12 : 1-391, 1 12 pis. 1958. Les Cyclostomes fossiles ou ostracoderms. In Grasse, P. (ed.), Traite de Zoologie 13. Agnathes et Poissons, anatomic, ethologie, systematique (\Q fasc.) : 173-425. Paris. Tarlo, L. B. H. 1960. The invertebrate origins of the vertebrates. Int. geol. Congr. 21 (Norden) : 113-123. Toombs, H. A. & Rixon, A. E. 1950. The use of plastics in the 'transfer method' of preparing fossils. Museums J., London, 50 : 105-107, pi. 8. Traquair, R. H. 1890. On the fossil fishes at Achanarras Quarry, Caithness. Ann. Mag. nat. Hist., London, (6) 6: 479^86. 1893a. A further description of Palaeospondylus gunni Traquair. Proc. R. phvs. Soc. Edinb. 12: 87-94, pi. 2. 1 893/7. A still further contribution to our knowledge of Palaeospondylus gunni Traquair. Proc. R. phys. Soc. Edinb. 12 : 3 12-32 1 , pi. 9. 1894. Palaeospondvlus gunni Traq. from the Caithness Flagstones. Ann. Scot. nat. Hist., Edinburgh, 1894 : 94-99, pi. 3. 144 P. L. FOREY & B. G. GARDINER 1897. Note on the affinity of Palaeospondylus gunni Traquair, in reply to Dr Bashford Dean of New York. Proc. zool. Soc. Lond. 1897 : 3 14-3 1 7. 1909. Note on the Fossil Fishes from the Niandt limestone, one-eighth of a mile S. W. of the mouth of Forse Burn. Mem. geol. Surv. Summ. Progr., London, 1908 : 92-93. Trewin, N. H. 1976. Correlation of the Achanarras and Sandwick Fish Beds, Middle Old Red Sandstone, Scotland. Scott. J. Geol, Edinburgh, 12 : 205-208. Watson, D. M. S. 1937. The acanthodian fishes. Phil. Trans. R. Soc., London, (8)228:49-146, pis 5-14. Wangsjo, G. 1952. The Downtonian and Devonian vertebrates of Spitzbergen. IX, Monographic and systematic studies of the Spitzbergen cephalaspids. Skr. norsk Polarinst., Oslo, 97. 611 pp. & addendum and supplementary note, 1 18 pis. White, E. I. 1935. The ostracoderm genus Pteraspis Kner and the relationships of the agnathous vertebrates. Phil. Trans. R. Soc., London, (B) 225 : 38 1-457, pis 25-27. 1946. Jamoytius kerwoodi, a new chordate from the Silurian of Lanarkshire. Geol. Mag., London, 83 : 89-97. Wickstead, J. M. 1969. Some further comments on Jamoytius kerwoodi White. Zool. J. Linn. Soc., London, 48: 42 1-422,1 pi. Woodward, A. S. 1892. The forerunners of back-boned animals. Nat. Sci., Lond. 1 : 596-602, 1 pi. 1898. Outlines of Vertebrate Palaeontology. 470 pp. Cambridge. Young, J. Z. 1962. The life Vertebrates, 2nd edn. 767 pp. Oxford. Burials, bodies and beheadings in Romano-British and Anglo-Saxon cemeteries M. Harman 3 Hampden Close, Tollgates, Battle, East Sussex T. I. Molleson Subdepartment of Anthropology, Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 5BD J. L. Price Department of Radiology, Royal Surrey County Hospital, Guildford, Surrey Synopsis Five late Romano-British sites, Cassington, Queensford Mill, Radley, Stanton Harcourt and Curbridge, all within a restricted area of the upper Thames valley, have produced over 200 extant skeletons. These have provided information for a study of the population, its health and its burial practices. Details of each individual are tabulated. The age at death, average adult height and frequency of some non-metric variables are given. Health is considered in detail: dental health, injuries, frequency of fractures, the incidence of osteo-arthritis and the occurrence of other disease, including the earliest recorded case of tuberculosis in Britain and an early example of possible osteosarcoma. The association of decapitated and prone burials in three of the cemeteries is noted and considered in a survey of the recorded instances of both practices in Romano-British and Anglo-Saxon cemeteries. Introduction Students of early British populations consulting the collection of the Subdepartment of Anthropology of the British Museum (Natural History) may be surprised to find some of the boxes labelled 'Oxford Collection'. The origins of this collection go back to the mid- nineteenth century, when George Rolleston, Professor of Human Anatomy at Oxford, assembled a large collection of skeletons from all over the world. The British archaeological material was derived from all parts of the country, but particularly from local sites. Rolleston's contemporaries rescued material from quarries and railway cuttings, and carried out their own excavations, especially on visible monuments such as barrows. Rolleston himself was no mean antiquary and retrieved much material. The tradition of personal involvement was continued by his successors in the Department of Human Anatomy, Dudley Buxton and Miss Beatrice Blackwood. E. T. Leeds, Keeper of the Ashmolean Museum, and his colleagues contributed remains from several cemeteries, and the Department received bones from other archaeologists. Thus the collection continued to grow. In the late 1940's, the major part of this collection, an outstanding group of skeletal material, well catalogued and organized, was given to the British Museum (Natural History) where it became known as the 'Oxford Collection'. The bones from two excavations, Cassington and Radley, have never been published, and these, together with some from more recent excavations at Curbridge, Stanton Harcourt and Queensford Mill, provide a sample of over 200 individuals from a restricted area of the upper Thames valley attributed to the late Romano-British period (third to early fifth centuries AD). These have provided information for the study of the population, its health, and its burial practices, particularly decapitation and prone burial. Both of these practices occur not Bull. Br. Mus. not. Hist. (Geol.) 35 (3) : 145-1 88 Issued 29 October 198 1 145 146 M. HARMAN, T. I. MOLLESON & J. L. PRICE Fig. 1 Decapitated and prone burial from Cassington as depicted in Captain Musgrave's field notebook. Original notebook in Ashmolean Museum, Oxford. Reproduced with permission. BURIALS, BODIES AND BEHEADINGS 147 uncommonly in southern England in late Romano-British times, and are recorded also in pagan Anglo-Saxon cemeteries. The significance of these practices is considered in the light of other examples noted from the literature. While decapitation has been discussed recently (Clarke 1979:192-193, 372-375; MacDonald 1979:414-421) prone burial and its association with decapitation has received little comment. The study of a group from a limited area increases knowledge of both practices without, perhaps, adding to the range of possible explanations for them. The sites At Cassington several burials were examined in the early 1930s by Leeds and others, but most of the bodies were retrieved by Captain Musgrave and Miss Blackwood in 1935 during the construction of the Oxford bypass. Over 100 skeletons were found, in a cemetery of unknown size, and the remains of 72 were added to the Oxford Collection, the rest being reburied. 60 are still available, although three which were sent, in 1932, to the University of Albuquerque, New Mexico, in exchange for Pueblo Indian remains, have not been examined. Information is available from the Oxford Catalogue and from Musgrave's notes and drawings of the skeletons in the graves (Fig. 1 ). Only brief notes have been published, probably by Leeds (JRS 1937:237)'. In the following descriptions the skeletons are generally referred to by the numbers given in the field by Musgrave. In a few cases, where there is no field number, the Oxford Collection Catalogue number has been used. The concordance between the two systems is given in Appendix I, p. 1 70. At Radley a complete small cemetery of 35 graves was hastily excavated in 1945 by R. J. C. Atkinson, in advance of gravel digging. A report of the excavation, though not of the human remains, was published (Atkinson 1952 : 32-34) and most of the bones were deposited in the British Museum (Natural History). Unfortunately the bones themselves were not numbered, labelling was inadequate, and in several instances two or three skeletons are confused. While generally the postcranial bones appear to be consistent with the skulls bearing the same number, these numbers cannot be correlated with the numbered graves on the site plan, making studies of spatial distribution impossible. The remains of at least 3 1 people are available for study. Recently rescue excavations have been carried out by the staff of the Oxfordshire Archaeological Unit. At Curbridge, graves containing 21 individuals, on the line of the Witney bypass, were investigated in 1975 by R. A. Chambers (1976 : 38-55; 1978 : 252). In 1978, at Stanton Harcourt, part of a cemetery threatened by gravel quarrying was excavated by N. McGavin. The excavation report with a brief note on the remains of the 36 individuals is published (McGavin 1981). At Queensford Mill a rapid rescue excavation of part of a very large cemetery took place in advance of gravel extraction in 1972 (Durham & Rowley 1972 : 32-37). A summary of the information about the human remains has been published (Harman, Lambrick, Miles & Rowley 1978 : 4-6). Decapitated bodies were discovered in the first four cemeteries, and in three of those prone burial also occurred. All five cemeteries, Cassington, Radley, Stanton Harcourt, Queensford Mill and Curbridge are considered in detail, while other smaller groups from the same area but which do not contribute to the general survey of the population are considered in the comment on burial practices. The preservation of the bones from Cassington, Radley, Stanton Harcourt and Queensford Mill, all gravel sites, was generally good, though some from Cassington and Radley had suffered when the topsoil was stripped from the sites. The graves at Radley were shallow and many of the skulls were damaged. The skeletons at Curbridge were buried shallowly in clay and their condition varied; some were well preserved but others were badly decayed. Most of 'A key to references in this form, more familiar to workers in archaeological subjects, will be found at the beginning of Appendix VI, p. 184. 148 M. HARMAN, T. I. MOLLESON & J. L. PRICE the skeletons from Stanton Harcourt and Queensford Mill were reasonably complete; some from Curbridge were incomplete owing to fragility and decay; some from Cassington were complete though many were represented by the skull, pectoral and pelvic girdles and all limb bones, or only a selection from these; at Radley most of the skeletons are represented by the skull, pelvic girdle and all limb bones. The populations: normal variation Appendices I-V (pp. 170-183) show the primary information deduced about each skeleton from the five cemeteries considered. The sex of adult individuals was decided where possible from the relevant features of the skull and pelvic girdle and from the size and ruggedness of the skeleton as a whole. The age of individuals has been assessed from the state of epiphyseal fusion and of tooth eruption, and the degree of tooth wear, based on the criteria given by Brothwell (1963 : 59, 60, 69) and from the length of the diaphyses in the case of juveniles, using the chart prepared by Miss R. Powers. The height of adults has been calculated where possible from the lengths of the long-bones, using the regression formulae of Trotter and Gleser (Brothwell 1963 : 102). The state of dental health is indicated by showing the incidence of caries, abscess and ante- mortem tooth loss. Where possible the presence of normal variations in the skeleton, such as metopism, wormian bones and vertebral anomalies, has been noted, and in the tables both these and any evidence of disease or injury are listed. Table 1 shows the number of skeletons from each site arranged according to age and sex, and from all sites combined. At Cassington there are almost twice as many males as females, while at Queensford Mill the males are outnumbered, though less dramatically. The discrepancy at Cassington seems too great to be accounted for by the relatively small sample size; both here and at Queensford Mill it could be explained as a result of segregated burial in a partially excavated cemetery, though the spatial distribution of 53 adult burials of known sex from the latter site does not show segregation. It is clear that children are under-represented in all the cemeteries. At Queensford Mill only one infant of less than a year was found, and this appeared to be a premature baby accompanying the mother, not an independent burial. This is true also of one infant from Cassington, two (possibly twins) from Curbridge, and one from Stanton Harcourt, thus reducing the number of independent infant burials by half. Infant burials are not uncommon on occupation sites, and have occurred in relatively large numbers on nearby sites such as Mount Farm, Berinsfield (Lambrick, in preparation) and Barton Court Farm, Abingdon (Miles, in press). While this may account for the deficiency of young children, the extreme paucity of burials of children under fifteen years in all cemeteries save Queensford Mill is remarkable. It is possible that infants and children buried in shallow graves may have been removed with topsoil, unknown to the excavators, but this was unlikely at Stanton Harcourt. At Queensford Mill, excavated in similar circumstances to the other cemeteries, a significantly larger proportion of child burials was found, and this seems to be a real difference between that cemetery and the others. No segregation according to age was noted at Queensford Mill. While this may have occurred at Cassington, Curbridge or Stanton Harcourt, Radley was completely excavated, and any children's corpses from that community must have been buried elsewhere. Since neither the true proportion of child to adolescent deaths nor the age of those regarded as 'over 40' or 'over 45' is known, it is not possible to calculate average age at death, but it can be observed that of those surviving beyond the age of 20 who could be aged with more accuracy than the mere term 'adult', 69 were males and 70 females: 50% of the males survived beyond the age of 40 years, but only 40% of the females. The difference may be due to a higher female mortality during child-bearing years. Extreme tooth wear and loss and some patho- logical conditions suggest that some individuals survived considerably beyond the age of 40 years. BURIALS, BODIES AND BEHEADINGS Table 1 Distribution of individuals according to age and sex. 149 Ac ',e in yi A J Site Sex 5 10 15 20 25 30 35 40 45 + Adult Total Cassington <5 _ _ _ _ 3 _ 4 1 1 8 25 42 9 - - - 1 1 2 - 2 2 5 9 22 7 31-1 1 1 7 Total - 31-2 4 3 5 3 3 13 34 71 Curbridge d 1 _ _ _ _ 1 _ _ 2 2 1 6 9 _ - - - 1 2 2 5 ? 2 - - 1 1 3 7 Total - 2 1 2 - - 3 7 3 18 Queensford d 1 _ _ _ _ 2 3 5 2 10 1 23 Mill 9 - - - 3 7 4 6 1 1 8 4 34 ? 1 7651 2 2 24 Total 1 7 6 5 2 4 9 7 11 3 1 20 7 81 Radley ef - - - 1 1 _ 2 _ 8 5 17 9 - - - 1 3 - - 1 7 2 14 ? . 1 1 Total - - - - 2 4 - 3 1 - 15 7 32 Stanton d 1 _ _ _ _ 1 6 1 1 5 1 15 Harcourt 9 - - - 1 5 2 2 - 2 3 15 9 _ 3 - - 1 2 6 Total - 3 - - 2 6 8 3 1 7 6 36 All d - - - 1 8 9 12 6 1 33 34 103 Cemeteries 9 - - 6 16 8 8 5 3 24 20 90 ? 1 15 7 5 4 1 1 2 5 4 45 Total 1 15 7 5 11 25 18 22 11 4 62 58 238 2 In more detail child mortality at Queensford Mill was as follows: Foetus of 7-8 months in utero, 1.1-2 years, 2. 2-3 years, 1 . 3 years, 2. 3^t years, 1 . 4-5 years, 1 . 5-6 years, 1 . 6 years, 1 . 7-8 years, 3.9-10 years, 1.10 years, 1.12 years, 2. 13-14 years, 2. The average height of 62 males was 5'6|" (1'70 m) and of 60 females 5'2^" (1*59 m), a little shorter than the modern British averages; the female figure is biased by the large proportion of very small individuals found at Queensford Mill, where of 21 females whose height could be calculated, eight were between 4' 1 0" ( 1 '47 m) and 5'0" ( 1 -52 m). Several skeletal variables have been noted. The incidence of lambdoid wormian bones is high in Romano-British populations -71% (Brothwell 1963:96); in the five cemeteries considered here it is 5 1 %. 150 M. HARMAN, T. I. MOLLESON & J. L. PRICE At Cassington, 14 of 29 possible cases have lambdoid wormians. Two of these also have an open metopic suture, but a third with this feature has no wormian bone. Nine individuals have the right arm longer than the left, probably not noticeable during life. Three of these also have minor sacral anomalies; incomplete fusion of the first sacral vertebra, or spina bifida occulta of the sacrum. Another individual has a sacralized fifth lumbar vertebra, and another, congenital fusion of two cervical vertebrae. At Radley, 10 of 16 possible individuals have lambdoid wormian bones, and of these three also have a wormian on the right side, in the suture between the temporal and occipital, just below asterion. This rare anomaly occurs in four of sixteen individuals in whom this part of the skull was observed. There is also one sagittal wormian bone and one case of an open metopic suture. Two vertebral anomalies occur; a sixth lumbar vertebra, and a sacralized lumbar vertebra. Atkinson notes that at Radley the graves fall into two groups, an eastern and a western group, with very slightly different orientations: he suggests that this may reflect custom, or a difference in date, or possibly family grouping, and it is unfortunate that it is no longer possible to correlate the skeletons with the graves, as anomalies of the type described might support the hypothesis of relationship. At Curbridge, three of ten possible individuals have lambdoid wormian bones, and a fourth has an inca bone. There are four cases of open metopic sutures, two of which also have lambdoid wormians. All the persons decapitated have open metopic sutures. At Stan ton Harcourt, 14 people have lambdoid wormian bones of a possible 21, and these include all three cases of an open metopic suture. There seems to be a correlation between the presence of lambdoid wormian bones and coffin nails; while this may suggest family relationship between those people buried in coffins, the incidence of lambdoid wormians is so high that such a conclusion may be unwarranted. There are five cases of congenital fusion of two vertebrae, usually cervical. This is a remarkable frequency of a fairly rare anomaly that tends to be familiar. Two individuals have an extra lumbar vertebra. At Queensford Mill, in a possible 44 individuals, lambdoid wormian bones occur in 17, of whom two also have coronal wormian bones, two have sagittal wormian bones, two have inca bones and one has an open metopic suture. Two other people have inca bones, and another has a sagittal wormian bone, and in addition an open metopic suture; this occurs in another two individuals who have no other cranial anomalies. Vertebral anomalies occur in eight individuals of 40 with reasonably complete and well-preserved vertebral columns. Three have separate neural arches on the fifth lumbar vertebra; one of these also shows sacral spina bifida occulta, and a second individual showing this has a cleft neural arch on the fifth lumbar vertebra. Another person has a separate neural arch on the fourth lumbar vertebra, and cleft neural arches occur in a further two individuals, one in the first cervical vertebra, and one in the eleventh thoracic vertebra, a very unusual site. One woman had congenital fusion of two cervical vertebrae. While it is clear from the above that some skeletal anomalies overlap others, and that their incidence varies, it is only in one instance that valuable information is provided; at Curbridge, where all three persons who had been decapitated had an open metopic suture, which strongly suggests that they were quite closely related. Total congenital absence of the third molars occurred in four of a total of 58 complete pairs of maxillae and mandibles; four have one absent, two have both lower teeth absent, one has both uppers absent, and one has three absent. Other incomplete jaws show one or two third molars absent. Unerupted and malpositioned upper canines occurred in one individual at Curbridge, three females at Stanton Harcourt, one of these with both upper canines affected, and one male at Queensford Mill, who also had a reduced adjacent incisor and the lower canine and last molar on the same side not developed. Generally people with unerupted canines seem to have retained the deciduous canine, and the adult teeth, though partially visible in the maxillae, did not protrude in unexpected directions and cause trouble. One female at Stanton Harcourt had retained the second lower left deciduous molar, the premolar showing no signs of having developed. BURIALS, BODIES AND BEHEADINGS 151 Very few postcranial bones manifested congenital variants. The femora of a young man from Cassington (14) had the angle of the femoral neck with the axis of the femur widened. On the radiograph, the bone texture was normal, so the widened angle would appear to be a congenital variant. The left arm of an elderly person from Queensford Mill (33) showed complete fusion of the lower end of the humerus with the radius and ulna. Radiographs showed absence of a joint space and revealed no evidence of infective disease. Bone trabeculae and the cortex passed without interruption across the joint and the appearances are therefore those of a congenital bony syntosis. There was a possible case of club foot in a young adult male from Queensford Mill (7). The articular surfaces of the right calcaneus and talus showed signs of degeneration. The talus appeared to have moved forward on the calcaneus and there was some hypoplasia of the anterior portion of the calcaneus. The osteo-arthritic changes in the talo-calcaneal joint are the dominant lesion and this could be the late result of congenital talipes equino varus (club foot). Broth well (1967) discusses an earlier and more severe case from a neolithic longbarrow at Nether Swell, Gloucestershire. Two cases of growth abnormalities were noted -both from Cassington. Flattening of the femoral head in an adult male (31) may be due to an old slipped epiphysis, while in El 1.8.597, a woman of 40-45 years, it may be the result of either a congenital dislocation or possibly Perthes disease. Perthes disease, where the femoral head becomes fragmented, describes a condition in which a localized death of bone and cartilage cells occurs in the primary or secondary centres of ossification. It is thought to result from a defective blood supply to the affected area and may in many cases be the result of minor injury. It is often painful and if untreated osteo-arthritis is an inevitable sequel (Davies 1969). Observations on pathology Dental health is considered to have been poor in the Romano-British population, and this is comfirmed by Table 2, which shows the incidence of caries, abscess and ante-mortem tooth loss in the total number of teeth and tooth sockets seen, for different age groups, in the different cemeteries and in all cemeteries combined. While there are some striking differences in the figures for different cemeteries, the small sample size from which most of these figures are derived reduces their importance. It is clear that in most cases there is an increased deterioration in dental health with increasing age. This deterioration was often accompanied by serious periodontal disease. Some individuals had totally edentulous jaws. In some cases teeth were worn to the roots, or very unevenly worn where opposing teeth were missing. Caries had also reduced some teeth to roots. In the rest of the skeleton the commonest findings were healed injuries and degenerative changes. Cuts other than those associated with decapitation occurred in one individual, Queensford Mill 59, a female of between 30 and 35 years, who had three cuts on the skull which did not appear to be recent. One small cut about 1 5 mm long is on the right parietal; another, about 80 mm in length, extends from the left side of the frontal to the left parietal region, lying obliquely to the sagittal line; and the third, roughly parallel to the sagittal and about 100 mm long, extends from the left parietal to the occipital. The skull has an inca bone and a lambdoid wormian, but the skeleton is otherwise unremarkable. The appearance and site of the injuries are not dissimilar to those occurring in a decapitated skull from Cassington, a man (4) whose injuries are described below. Table 3 shows the total number of reasonably complete limb bones from adults in all cemeteries, divided into pairs and odd lefts or rights. Fractures are not common. These are all the fractures seen on visual inspection, but old well-healed fractures may not have been recognized. Only two, possibly three, women with healed fractures were noted. Queensford Mill 22 has an old greenstick fracture of the left radius which had healed with angulation of the distal 152 M. HARMAN, T. I. MOLLESON & J. L. PRICE Table 2 Incidence of caries, abscess and ante-mortem tooth loss in teeth and tooth sockets seen, arranged according to age groups. Age in Years Caries Abscess Loss Cassington 20-30 30^*0 40+ 17/194 09/164 24/136 9% 6% 18% 08/206 15/203 39/215 4% 7% 18% 07/220 11/212 126/313 3% 5% 40% Curbridge 20-30 30+ 40+ 00/43 09/33 08/40 0% 27% 20% 00/47 08/46 02/36 0% 17% 6% 00/48 08/61 25/84 0% 13% 30% Queensford Mill 20-30 30-40 40+ 07/252 35/246 33/179 3% 14% 18% 00/286 18/298 59/314 0% 6% 16% 06/302 27/326 168/470 2% 8% 36% Radley 20-30 30^0 40+ 08/82 23/98 16/58 10% 23% 28% 07/109 12/139 20/121 6% 9% 17% 03/112 28/168 142/266 3% 17% 53% Stanton Harcourt 20-30 30^0 40+ 26/301 24/91 20/75 9% 26% 27% 14/352 19/118 27/121 3% 16% 22% 16/392 06/122 81/204 4% 5% 40% All Cemeteries 20-30 30^0 40+ 58/872 100/632 101/488 7% 16% 21% 29/1000 72/804 147/807 3% 9% 18% 32/1074 80/889 542/1337 3% 9% 41% Table 3 Total numbers of intact adult limb bones and numbers of fractures recognized. P = pairs; S = singles. Stanton Queensford Cassington Radley Curbridge Harcourt Mill Total PSPSPSPSPSPS Fractures Clavicle 29 7 4 1 3 2 25 2 50 2 111 14 4 cases Scapula 25 4 1 2 3 2 25 50 1 104 9 1? Humerus 31 8 23 4 5 2 26 2 48 4 133 20 Radius 29 7 19 6 4 2 28 50 1 130 16 3 Ulna 24 9 22 4 5 2 27 51 129 15 2 Femur 38 6 25 3 4 3 26 1 51 1 144 14 1 Tibia 31 6 23 6 3 3 27 2 50 1 134 18 2 Fibula 21 1 19 2 3 2 25 2 50 118 7 3 Ribs 4 cases end of the radius so that the wrist would have been slightly deformed. Cassington 39 has a well-healed spiral fracture of the left mid-femoral shaft. The appearance of Radley 21 suggests a healed fracture of the left scapula but this was not confirmed by the radiograph. Seven cases of healed fractures of clavicle and ribs occur among the male skeletons, suggesting recovery from severe falls or impact blows to the chest and back. Fractures to the lower arm, particularly the left, suggest attempts to parry a blow as in fighting (although BURIALS, BODIES AND BEHEADINGS 153 there is little evidence from cut wounds of battle fighting). These lower arm fractures usually heal well (but see Cassington El 1.8. 587) with little deformity, as unless both bones are fractured, the uninjured bone acts as a splint for the injured. Healed spiral fractures of the tibia and fibula are of the kind associated with a rotatory force a boot or shoe gets caught in the stirrup, plough or rut in circumstances when the unshod foot would pull away or suffer lesions itself. One man, Queensford Mill 35, also had fractured his left clavicle and, more recently, his ribs. Some infective changes were present on the clavicle. The discrepancy in the incidence of fractures between the sexes might indicate that male occupations or diversions were more likely to result in this type of injury. It is curious that most of the fractures occur on the left side. Splinting was known in ancient times but traction of fractures was not used until the 13th Century AD so that all the fractures of the lower limb, although well-healed, show some degree of angulation and overlap of the fragments causing shortening of the bone. Bony spurs have developed at the site of a muscle attachment on the left femur of Cassington 5, a young adult male, and Radley 7, a much older male. These could be related to injury at the muscle insertion and are sometimes associated with horse-riding. An older male from Cassington, 27, shows a quite marked cortical thickening of the anterior shaft of the right tibia. This is the type of thickening that is associated with sub-periosteal new bone formation and could be related to repetitive minor trauma. Osteo-arthritis is a slowly progressive condition which eventually leads to destruction of a joint. Insult to the joint from infection or injury, or increased stress from congenital anomalies, may predispose to or initiate this condition. As a result of further wear and tear on the joint surfaces further deterioration occurs, although the more flagrant manifestations of the disease may not be apparent for many years. Severe osteo-arthritis of the hip joint with osteophytic lipping of the acetabulum and femoral head were noted in two males from Cassington, 19 and 43 (Fig. 4, p. 1 58; two adults from Radley, 29 and 1 1; and two males from Stanton Harcourt, 29 and 67. In the last case, the osteo-arthritis of the right hip is almost certainly associated with increased stress on the joint following a malunited fracture of the left tibia and fibula. Usually the right hip is more severely affected than the left. Injuries to the knee, in each case the right, were probably the predisposing factors leading to severe arthritis observed in two adult females, Queensford Mill 19 and Radley 22, and another old adult, Radley 32. In the case from Queensford Mill there is a small hollow in the articular surface of the medial condyle of the right femur. It could be due to osteochondritis dessecans. This is a disorder in which a small portion of bone and cartilage from the joint surface becomes detached and forms a loose body within the joint space. The condition is thought to result from defective blood supply to the affected area and may be brought about by a minor injury. Osteochondritis dessecans is seen most frequently in adolescents and young adults and usually occurs in the larger joints, especially the knee joint. It gives rise to pain and an excess of fluid in the joint space. Recurrent dislocation of the patella may have led to the osteo-arthritis of the right knee of Radley 22. The anterior aspect of the lateral condyle is mainly affected with anterior fringe osteophytes suggesting chondromalacia patellae as a predisposing cause. This is a disorder in which degenerative changes occur in the articular cartilage covering the posterior surface of the patella (Davies 1969). The sacrum of this woman has spina bifida occulta of the first sacral vertebra. Spondylotic changes on the vertebrae were very common throughout the sample and there were few adult bodies in any cemetery which did not show some evidence of this condition (Table 4). Spondylosis is a result of degenerative changes in the intervertebral discs and causes abnormal movement of the vertebral bodies and increased strain on the small posterior synovial joints. Small bony outgrowths (osteophytes) and spurs appear on the margins of the vertebral bodies and extend into the adjacent ligaments and may eventually bridge the intervertebral spaces and provide some stability. At the same time osteo-arthritic 154 M. HARMAN, T. I. MOLLESON & J. L. PRICE Table 4 Total numbers of adult vertebral columns, the number of those affected by osteo-arthritis being shown in brackets. (a) 'Complete' columns (missing less than 5 vertebrae) Age Cassington Radley Curbridge Stanton H. Queensford Mill Total 20-30 5 (0) _ _ 8 (2) 4 (0) 17 (2) 30^0 2 (2) 1 (1) 5 (4) 6 (2) 14 (9) 40+ 3 (3) 1 (1) 5 (5) 11 (10) 20 (19) Adult 2 (1) 3 (1) 5 (2) (b) Partially complete columns (about half the vertebrae or more present) Age Cassington Radley Curbridge Stanton H. Queensford Mill Total 20-30 2 (0) _ 2 (0) 1 (1) 4 (0) 9 (1) 30^40 - - - ' 2 (0) 2 (0) 40+ 1 (1) - 1 (1) 2 (2) 4 (4) Adult 5 (5) - 1 (1) 1 (1) 7 (7) Also 1 6 sacra from Radley. changes occur in the posterior facetal joints. In modern populations minor spondylotic changes are very common over the age of 40 but the grosser manifestations do not usually occur until the late 50s. It is thought that minor congenital anomalies may predispose to spondylosis. In the cervical spine heredity certainly plays a part in its genesis, as a familial incidence has been demonstrated. The adult male from Cassington 41 is an example of its association with a congenital anomaly. The bodies of cervical vertebrae Cl and C2 are partially fused and the laminae are completely fused. It is clear from a consideration of these individuals, and from smaller groups of vertebrae surviving from others, that vertebral osteo-arthritis was unusual in persons of less than 30 years of age, as it occurred in this age group only at Stanton Harcourt where evidence of slight growth on some of the lower thoracic and lumbar vertebrae was noted in five individuals. In the 30 to 40 year-old age group generally the same areas were affected, but more vertebrae were involved. At Stanton Harcourt in addition four people in this group had some degeneration and growth on the cervical vertebrae. In those persons regarded as over 40 years of age, only one, a female from Queensford Mill (67), appears to have remained unaffected by osteo-arthritis in the vertebrae. The degree of degeneration in this age group is variable, probably reflecting the range of ages over several decades; many have slight degeneration in all parts of the column, though in general the lower thoracic and lumbar vertebrae show more moderate to severe signs and in a few individuals fusion occurs. Many cases effusion were shown on radiography to be congenital. However, in a man from Queensford Mill (5) spondolytic growths had united the final cervical and first thoracic vertebrae and also two other upper thoracics. The body of a lower thoracic vertebra had collapsed. Two thoracic vertebrae, the third and fourth, were fused in a man from Cassington (65). At Radley, in sixteen persons of varying ages from whom only the sacral part of the spine had been retained, one ageing female (22) seems to have had the final lumbar vertebra joined to the sacrum by bony growth, and in one man (11) the final lumbar vertebra had also been kept, as it was joined to the sacrum by osteophytic spurs and bridges. One individual from Curbridge (13) had eburnation on the articular facets of some vertebral fragments, and an elderly man from Stanton Harcourt (16) showed moderate evidence of spondylosis on all the vertebrae, the lower part of the column being more severely affected, accompanied by some collapse of the lumbar vertebral bodies. BURIALS, BODIES AND BEHEADINGS 155 Fig. 2 The cranium of Radley 1 7 showing the probable osteosarcoma of the left parietal bone. The wormian bone in the unusual position just below asterion can also be seen. 156 M. HARMAN, T. I. MOLLESON & J. L. PRICE Fig. 3. Pott's disease or spinal tuberculosis in a woman from Queensford Mill (no. 157). Thus the general picture is one of increasing deterioration with advancing age, the lower part of the back normally being first affected. People from Stanton Harcourt seem to have suffered more and at an earlier age than those from Queensford Mill but this is only an impression based on a small number of people, although it may be related to the high frequency of congenital fusion of vertebrae. Other cases of degenerative disease were noted in an elderly woman from Cassington (22) BURIALS, BODIES AND BEHEADINGS 157 and an elderly man from Queensford Mill (30). The woman showed collapse of the body of the fifth lumbar vertebra which was probably due to osteoporosis. This is a condition in which the supporting connective tissue of bone is defective and a generalized osteoporosis is common in elderly subjects, especially women. Paget's disease of bone is rare under the age of 40 but the incidence rises to 5% in the 6th decade and it is slightly more common in men than in women. The disease causes softening and thickening of the bone with disorganization of the trabecular pattern which takes on a spongy appearance. Clinical symptoms are not common but in its advanced form pain and deformity of the bone may be a feature of the disease. Evidence for Paget's disease is to be found in the case of the elderly man from Queensford Mill (30). Radiographically the cortex of the right radius is seen to be thickened and the trabecular pattern is coarse and disorganized. (He also has a healed fracture of the medial end of the left clavicle). Malignant bone tumours are not commonly found in early populations so the spectacular parietal bone tumour from Radley (17) is of particular interest (Fig. 2). It ranks with the equally important, though somewhat later, osteosarcoma of the knee from the Anglo-Saxon cemetery at Standlake (Brothwell 1967). Radiographs of the Radley skull show fine spicules of bone arising from and forming a boss on the right parietal bone. There appears to be a destructive process in the skull vault but this is patchy and partly obscured by sclerosis. The inner surface and the vascular channels are normal. The changes are most probably due to sclerosing proliferative osteogenic sarcoma. This tumour is usually highly malignant, grows rapidly and early secondary blood-borne deposits in the lungs and other organs are almost inevitable in untreated cases. Differentiation from meningioma can be difficult, if not impossible. Brothwell (1961) originally described the pathology as a sarcoma but on reconsidering the evidence (Brothwell 1967) preferred to interpret the changes to the skull vault as being caused by a meningioma or angioma. A laminar periosteal reaction at the lower end of the forearm bones of Cassington 2 arouses the suspicion of a hypertrophic pulmonary osteo-arthropathy. The changes in the lower limbs are less convincing. The bones are otherwise remarkably healthy and are presumably of a young adult. The condition is usually associated with chronic inflam- matory lung diseases, congenital heart disease or intrathoracic and pleural tumours. The bones would be less robust in congenital heart disease so that a chronic infective condition is more likely. Very little evidence for infective bone disease was noted as is usually the case with earlier British populations. One possible case of Pott's disease or spinal tuberculosis was diagnosed in an adult woman from Queensford Mill (157) (Fig. 3). The vertebrae T9-L3 are affected and there is acute angulation in the dorso-lumbar portion of the spine, with collapse and destruction of the eleventh and twelfth thoracic vertebral bodies. Anterior buttressing is present with well consolidated new bone and some sclerosis. The condition is long-standing and is almost certainly due to healed tuberculosis. Rheumatoid arthritis may have been present among the people of Cassington and also Queensford Mill. It is an inflammatory disorder of connective tissue which usually commences in early adult life and shows a much higher incidence in females than in males. Common sites for its onset are the small joints of the hands and feet, which become painful, stiff and swollen. The left hip of an adult male from Cassington (43) shows flattening of the femoral head with some cystic changes and lateral drift of the head within the acetabulum. There is gross osteophytic lipping of the acetabulum and osteophytosis of the inferior margin of the femoral head with some buttressing and new bone formation in the inferior margin of the femoral neck. The appearances are those of osteo-arthritis but this could be imposed on old rheumatoid disease (Fig. 4). An elderly adult from Queensford Mill (33) shows osteo-arthritic changes to the right shoulder, congenital fusion of the left elbow and erosion of the head of the third metacarpal. This could be rheumatoid arthritis or gout, as the articular surface is intact. An older adult male from Cassington (42) had survived fractures to the ribs and to his left ulna but radiographs of the ulna show that healing did not proceed normally and a chronic 158 M. HARMAN, T. I. MOLLESON & J. L. PRICE Fig. 4. Severe osteo-arthritis of the left hip in a man from Cassington (no. 43). Radiograph below. BURIALS, BODIES AND BEHEADINGS 1 59 infective process may have been involved in the cortical thickening and periosteal irregularity of the bone. The evidence for nutritional diseases is extremely tenuous. A single femur from Radley (3) showed changes that could be due to healed rickets. There is marked bowing of the femur. Radiographs show the bone texture to be undisturbed although there is post-mortem pitting of the outer cortex. Some buttressing and remodelling had occurred on the inner border of the curve. This is unlikely to be due to physiological bowing, as the bone is usually remodelled by adult life, so a tentative diagnosis of healed rickets is suggested. As noted elsewhere the skeletons from Radley are otherwise remarkable for the evidence they show for survival to an old age. In assessing a population it is useful to consider the diseases that are not found. In these upper Thames valley Romano-British cemeteries no evidence of syphilis or leprosy was found although a case of leprosy has been documented from the probably contemporaneous, though Christian, cemetery at Poundbury, Dorset (Reader 1974). There was little bone infection present, neither osteomyelitis nor periosteitis being recorded. There was, already noted, one probable case of tuberculosis of the spine from Queensford Mill, and the significance of this case may prove to be considerable in the light of future work in the area. In the meantime it may well stand as the earliest case of tuberculosis so far recorded for Britain. Table 5 Frequency of orbital osteoporosis in late Romano-British populations ^ ' Extent of orbital osteoporosis None Slight Moderate Radley 17 12 4(4, 11,14/16C, 18A) 1(120) Cassington 32 28 4 (2, 6, 37, El 1.8/587) Stanton Harcourt 24 22 2(79,104) Queensford Mill 49 41 8(10, 16, 18,31,45,48,70, 178) Poundbury (adults) 166 125 9 32 Other diseases of poverty and overcrowding such as rickets and scurvy are also not recorded with certainty from these cemeteries. The one possible case of healed rickets from Radley (above) rests on such slender evidence that a strong diagnosis cannot be hazarded. However, mild cases of rickets have been recognized from Poundbury (Molleson, in prep.) and we may be seeing here and at Radley the first signs of deterioration in living conditions in Britain, in consequence of starvation, famine or overcrowding. It is worth recalling in this context that several of the women at Queensford Mill and one from Cassington (15) were of short stature. No bone evidence of thallasaemia or sickle cell anaemia was recorded, although malaria was probably endemic in Britain at the time (Howe 1972). The severity of orbital and parietal osteoporosis was not great: this in marked contrast to the evidence from studies on the Romano-British population at Poundbury (Table 5). If the development of cranial osteoporosis can be taken as a response to hyperdevelopment of bone marrow consequent on chronic iron deficiency anaemia (Angel 1966; CafTey 1937) then the mildness of these conditions in the upper Thames valley populations can be taken as an indication of the generally good dietary health of the people. Decapitated and prone burials Distribution Of the five cemeteries considered, four are remarkable for both the absence of child burials 160 M. HARMAN, T. I. MOLLESON & J. L. PRICE Table 6 Decapitated and prone individuals from Romano-British cemetries in the upper Thames region. Site, No. Sex Age Decapitated Prone Position of head, cuts, other comments Abingdon 3 E c o t/3 S S E o oo C fe "O c O _H -. , CU C ess <" & E u ^ Q ^ Jr< CD u 1 * * S 1 3 c ^ " 6 O o r3 j= o -= 5 o 1 z z c ' & ."2 I'S - 'S S^ ^ CU g c C/5 o T3 C C ? o r- 13 ^ U O t/5 6 S O tn fe S X Z Z O cu j= S U X 1 03 | S U. S iSf! fflzo 3 X UJ ~ : 8' 1 1 "3 ^Jsquinu t"-- oo ON co ^~ NO r OO ON O O rs fN^ CO ^ m NO O 1 1 i_l T OO oo OO ON ON ON ON ON ON o o O O O o o o O 'II3 P JO J X O in in m in uo >n in m in ^o ^o NO \o 'O NO NO NO UOIJBJU3UO C/5 C/5 Z 1 UJ Z z UJ Z Z Z S P^id^a X X X X X X 3UOJJ X o- r- NO NO CO n Tj. NO p ON xgpui IBIUBJ;} NO ^^ ^vj" CO 00 -4- CO ^T- X r- oo r- r ~- * oo f* *" *"* 3 UIJB iqSu JsSuoq 1 SUBIUIJOM piopquiBq O 1 1 co I O O 1 1 1 1 fM 1 rj- CO snuqj.re-09jso |Bjq9JJ9A XX X oo oo NO in (N fN -n No^ OO fN ON ^ NO C/3 M fN fN CO CO -- fN CO r-) fN " O O O O NO in O ON CO m CO NO J O O O O O O O O O fN Is oo t NO in NO r- in >n Tt OO 00 ON rs en fN N ' fN fN O fN fN rN O ~ JD O __ O O n-^- in-^-in in in .1, UO wo wo iy^i in in _ 1 1 m <- >-> o <- m m m m o o <- in m U E o 00 7 33^ 3 ? f ^^^ ^ 00 Is ^^ ^^ fN ^1 CO fN "3 7 o- o o -b o "0 O 03 < 09 Z *~ ' fN CO * in NO r~~ BURIALS, BODIES AND BEHEADINGS 171 "rt S ON 00 I' 52 o o I ro rt U S "c ^ 03 to CN.i uu- fN oo oo J= tt iM _0 .c on ^ '35 a .^ .SP _* C/) c (U t/5 3 X) .s O 1/5 09 ^ 1 (A 00 c ^ o "C t*- C 'C ^ f f" o t/3 t: C3 CU ^ ^ "S o 'C '35 - f~\ OO S -c c t CM O 00 hi f"N 73 t s 3 ~ 73 6 "3. J C o ed ill tfl c II '5.2 0.5 U-S fN ro Tt I/ i NO r-~ oo SS S fN r -1 fN fN oo ua % UJ Z X X X X oo oo ro [^. t~~- -^r '/""> r- r- r- r~ X X X X 11*8 TD X <" _c g '^ o S S O ro ro NO ^O 7 00 Z w ro OO O O X I I I I fN oo O O oo O O O ro O o o o o r-~ o o fN ro fN ro O O ro fN ro O ro fN 00 fN o fN - fN ro O O O ^ o O O NO O o o o o o o o o ro O fN O S O O O oo ro fN O O ro 5 NO fN o O fN fN ON fN O o O O o \ \ fN O o o 1 o o r- O fN O O ro O fN O ^S 5 E j E 5 E ^~* i/"> ^^ i/~i r^- t ^^ i **o T^-^-.lO V^^- IT) ~- 1 - "oo ^" fN ro Tf in ^O t~~- OO ON O O fNfN fN fN fNfN fN fN fN ro ro 172 M. HARMAN, T. I. MOLLESON & J. L. PRICE J5 "0 >> u: <~ => rt o 3 C O -o S2 Q. EaojS^o -o !*! ^* b^ "Z- ^ ' ^ ^f ;J if. y O o O t o'_5 r ^^"'"c" cf^c u *^^ ^^ u c ' ^^ T t O !- -".j on~ in.." * ^ Sj * I** o '-S^"" 1 S_.C - )~^ vO UOUBJU3UO ^ UJ UJ ^ UJ ^ Z Z wuj Z Z ZZZZwZ ^Z pajejtdEoaQ X XXX X 3UOJJ X XXXX XXXX \3pUl JBIUBJ2) oo r~- r- r~ IUJB jqSu J3uo'] XX X sumiujOA\ piopquiBT 0~ II 100 |4 Illlll IIIIn o m 00 00 *-i c ton cc _M '5 I v e v E v e v s v E V e v e v ; 5 E c '$ o U o o SB 'Y T 33 3 . 3 ^ 3"o 3 3 3333.3 3 /i. . .L "^ ^Q *^ .T "Q L "^3 *^ *^ *O "^ "O "O *O i/-\ *O "O ."2 X 0i o *o oo *o "o *o "o '^oc^-c^-'ooo c^-c^-^o^o^o" ^o c^- 3 flQ 6 Z ^O r^- o^ ^^ ^^ r^i ro ^^ ^o ^o r^- oo o^ ^ rs r*^ ^- fN ^ i/~> ^o r^- r*^ ro r*" 1 ) ^f ^f ^f ^ ^ ^^ ^" ^^ ^" ^ */^ */^ */^ ^O ^O "O ^O ^O ^O BURIALS, BODIES AND BEHEADINGS 173 UJ w CO I I I I 33333 T3 TJ T3 -O -O "o "o "o "o "o OO ON O ' fN vo ^o r^^ r"-~ r^- UJ o U "O O O "o S I 1 C-- CQ O >< W J= ,000 c a 3 > _00 Si M U 53 2 i u '^ oo O flj o o Q, co o "J^ c C -'C c OO f-| 2 o O r~- 3 g 5 00 -O oo oo oo 1 '" t u ^ | 51 O (L) "S 1> :- 3 *2 >- r- oo 3 C O oa'o^ Osteo-arthriti UO11BJU3UO ^ Z Z w U ZZ WWLUZZZZ ZZ Z pwidKBci X X X uistdopj^ X X X X sumuuoM piopquiei (N + , 3 e o, CO en in os (N 2 ^ o O O __ \O __ oo m o t~- >n fS| 1 O O O o O O O O O O o 00 aa U Tfr ON CO __ ^c ON (N T -_ O oo c^- O- {^. e^- i 1/3 1 o "O o+ to e- o- "o o- o- *b 00+ "0 B 3 ae 6 ?: CO-* ft r- oo fNr^ Tj-moNfNroinr- OO . CO f- 2 - - E o CO E 1 t 1 >Ji w a 5 o 3 ^5 CO CO J3 (-1 1-" ^^ (U TD -G ^ _Q O O t pC C P- ^ (U J= ^ o '*" " E t*3 t^ *"*** c ' - "Trt *^ u aj C 00 C 3 2 c ~* > a o 'c O O .2 o *j ~ c. ^ (U C o o C/3 5; C/3 OQ _J 55 CO "O (/3 '5 S T3.2 _ S . ^ E fc ^x o 2 * o o o o O O O O O c o a CM OO sO so ON ^Q ^j. ^. -rf C*~, O sO (N ^t oo (N r- rt o. C/3 co ro r ] ^ ^~ (N| _. O O in in in in ^t-in^ o O in O in in O TT m *j O in V) a fi 00 12 O O r ^ + T -^ r- in o J Tt 1^1 7 o (N T3 j^. ~^i m a> 4. a. 3 i ^^^5 r-A < 't ^t < sj Tj- in so r- ON m TI- in so r oo ON O fN T}- m CC Z 176 M. HARMAN, T. I. MOLLESON & J. L. PRICE CO C CO 8 o H cO P u I _" xl a jo _ *- ^ X) X) *- "^ *^ O ^ 2 S" -^ X) ^_ *- i Xi ^ ."ti t) JZ 6 OX) OO ^ ^ JS (U 0) t/3 C/3 > > 00 3 op || O. aj "O C/5 S CO CO U op 'ob O CO t" O 'OJD'OO c3 cd C/5 C/3 "O on i| ll O 'O - o O O - rs rs rs so SO "i- ON rs SO OO in CN^ S? en en en en -^ O rs en O rs rs fN) -H o m i O m en o o O ^ en en o o o o 1 O O O 00 o O o o o o o c/3 aa n ~ in "~" t/i m - m- m- ">-"'> n m ** O m o 3 n mo O O Tj" (N en *O rs en rs en en 00 O in m tJ m en in O in rs rs rs co 1 oors O r~~ f^N 4- | | 1 1 -KN 1 1 ON O r~ rn *~> ( ^~> X u C/3 o "o o e- "o - CM- "O o- cc~- o o- o- "o -o o C~ CH- O ?-?* O+ "O Q SO ON O rs en rj- in SO I OO ON O en w > o o '5 S 'u CO o UN o ~ JS 3 u oo E _^ flj Cv c oo trt o O c C oo OO o O .^ 4 U c>- CO " CO C CO oo "rt "eO 6 U O X> u 3 C i_ 3 U C o o _W w op c r i n rf\ some cervical and some P- al vertebrae, but C3 and C4 teo-arthritis on cervical thritis on lower thoracic ic. Infective arthritis of n sternum, il vertebrae, ankles. ulosis of spine. C J3 o u. u Q % 2 *o ? ! b g o g a 1 S <$ c^ O =; 111 C/3 3^ C 2 s o u ?l o t: u <*- 15 JO CO CO 00 c/5 "Sb-O C r; J= cJ -c on _3 o - j= S c>- to 1 > O i- ^ CO ^> 00 K.2 S| 11 00 05 oa gj 2* PT>! C n O O O fN O O ro O O oo ON O O L ro ^f so O fN >n oo O ro fN t~~- O O so r^ O CO fN O fN fN fN fN t o fN m o o o o o o o o o r^ O TJ- O fN O O O O o O m O O O fN O ro O O O O fN m in TJ- O O so fN O fN fN O ON fN SO ON O SO Tf fN fN O fN m 1 o o ^r o o o fN O O O O O Tf O O fN O O O O O O O fN fN O O O in O O CO IE 5 S E, E , 6 5 l w , : g j E ; s ; v O ON r" T O >n ro 2 p ON"^ ^o OO O v fN O OO -wOO 'Sf r ^ v ro oo r- J m in rS'rJ m so ^ so O S .V .V m f "" In- In in i >n i n O >n o fN -^- ro in O O co t^* in co ro * * fN r*"> * 1 2 I 3=rA ^- <, fN *!. ro ^j ^t fN T fN fN 00 + Tt in T o >n O >n + 7 + in o "n TJ- m fN o+ o- CH o- o c>- O CH *b "O o o o+ 'b 0*0- Of Of 0- c-- "o "0 Of "0 "O Of Of co in >n r~- oo o SO so so so r-~ oo o fN m Tf oo OO O r- so O O in fN >n ro >n r~- in in >n 178 M. HARMAN, T. I. MOLLESON & J. L. PRICE CO o co 00 c! u co g _3 C *2 J= c H 1 E 2 o 3 o u C o U o ^ 11 ll ^^ ^^ *o C/3 M ~ Tt - .2 3 d r^j ^^ */^ ^o i * oo ^^ fiQ Z t i * t r r~- r~^ r BURIALS, BODIES AND BEHEADINGS 179 E W 3 C IS " r" U. "^ CO 3 s " o o j- t 7 c O O >, JZ co 1> **" . O , O co CL C/-3>OU-lO x X. "c ^ & uj O O S <5 p3JEJlde03Q X X OO OO ON oo m CN X3DUI IBtUBJ") fN fN /"} ^j' ^O ^^ ' ' r~ r- r~- r- r~ r-- UIJE jqSu J3uoq X X e^- x o- x UEfUIJOM OIUOU3JSV O O X XX > sumiujOM piopqiueq ^^ ^D f^l fN O O O O o \& c^j */^ ^o >i ^^ I uo- ^-Sn-^o-^- uo-in- -> .0 in m ^ _ ^ 9! u 00 5 "5 CO z (^(Nr^TtUsONO 00 ON ON ON ~ _ 180 M. HARMAN, T. I. MOLLESON & J. L. PRICE 00 iii ,! c "^ t/T . o. s s i ^ a 3 -Si 3 * ^^ ^ -"* O ... ^V t- C/3 ^ "^* gc rt ? ,GJ *-. w CH -^ OD 3 U^'S 3 5 O J= C/3 -J O -J 2 C/3 3 o t/5 Sc -S S "^cd^^ * * ^g^Ofti on on-o xapui [BiuBJ3 >/"> ON I f~- O OS O 00 ^! r^ r- UIJB jqSu J38uoi X X X XX UBIUIJOM D|UOU3JSV OOO X O O O o o o SUBIUIJOM piopquiBq (N O O 5 J o o o o o o (/I u i Z5 O rs) fN *- *- *O iu 00 + ^ + "^"^~^+T+"^"^+ T + - "o "o "o "o Ot O O "O "O is < CQ U (N s 6 60 Z ~ ~ ~ - Z rsl U^ 00 ON BURIALS, BODIES AND BEHEADINGS 181 >> >> ^ oj 'c'c'c 000 = JJ _u _0 .o .o .o u 'O T3 T3 rt C C C -N< r^ ro f^i r^ r* 1 ^ (N fN "V2 W Q3 c5 >". E o g - . 2 o o obo a CO = 73 O"" tf ' C -OT3 c >^ c c ^^ c^ 8,8 o& 2 >- 'ol a> o *"* w C8 O II <- O C C O IU U CQ 111 P-r^ o | = *J C -I- s \ ^r^ 12 e-s e^ s s 182 M. HARMAN, T. I. MOLLESON & J. L. PRICE CU CO a U *s la S C aJ o i_ CO ^ ^ 2 a^ CO CO 1 | ^ !c H > 2 n ^ o 2 C r} cd C/3 "ZJ 3 ^5 tu "n c If S 2 c.2 CU "O ~" ^ "t^ ^ O - > . ^^ f] .5 c ^ ^ ^ li t; c c oo > co QJQ C >" u C/3 U .22 o 'H " || ^ CO o g o g =i j= t 8 t E 5 t: '5 '-S o ' co ca ^ 6 ^ U C C/5 " ^" o o u U .C 6 -C en ?S 6 ^ cu %5 c/3 rN ou jg.g t; - cj 7 c '5. S| f 7| O cU . *- * O aa 2 ililll ilif! S<^ g I'S-S 4 > ^_ * U> Qf c/3 ro c/5 O ^ "" O U O U O O rN so n ^^ ^^ ** o o O O O O O O O O O C/3 cu ON SO fN CO Tt so uo 00 CO O r- o fN rM O O fN fN fN CO CO u CO i O 00 O O O fN >n m O O O O o 41 E , , E, v c 00 i oo i O I OO -KN SO i OO -! -w ^VTrO wTf '5 so SO co SO \o ^o ^^ in so so r^- r*~ co so so r~- TJ- so u 3C m wn ^- wn iin >n n m -n O co _o g u 00 CO fN 1 1 1 O T r- rN fN fN fN o o o o o + + + O n o T 7 O oo < - CO fN ^t fN fN fN ^ ^ ^ CO C o CO i u c CO 1 CO X) _ CO i U > > ^ > ' > > E > 4J > 3 3 Osteo-arthritis on C2 & 3 fused cong Hole in sternum. Metopic. Buried with 62. Osteo-arthritis on L tibia & fibula. Osteo-arthritis on 'C JS c ? 6 u iyi O _E: DO 55 Osteo-arthritis on Slight osteo-arthri Upper third molai Osteo-arthritis on same grave. Osteo-arthritis on Underlies 93 insa Osteo-arthritis on C2 & 3 fused cong Osteo-arthritis on Disturbed and reb Disturbed and reb Z oo Z Z ZZ Z Z * ZZ Z Z ZZ C/5 Z t/3 C/3 Z X X X X X X X X X X X X X fN fN SO C* o - o - O -r SO fN ON fN OO fN fN TT ON fN OO fN O SO CO fN CO fN CO fN CO CO fN fN fN fN CO CO in o O " CO O SO O O fN -j- fN ON in in O O O O O O O O O O fN O fN r- o fN OO ON so oo OO O r- oo r- co O fN fN o co O O fN in O so O fN O O O O /"> CO r~ o fN O O O O O o o O o O co O OO fN OO _ SO ON fN fN uo CO fN m CO O en CO fN fN fN O fN fN fN O fN O O O o "-I" CO >/~> fN fN TT O SO fN m fN O Q O O O O O O O O O O O O , . ^ ; E E E E ; E wE ,^-rf Z-^ fN ^-wr- o oo v CO v co 5 oo -wOO ^ O OQ r~~ t~~- t-~ m so " ^j" oo r-- oo r~- fN in SO sO r *> in-in - in - ^j- - In in - In in - ^t in 15 O O m w o o O *-> o m ,_! _ in 6 *- co O m Tf fN ? 7 m O fN fN C co co 1 1 C fN fN T fN + 7 O in ^- fN *(. fN CO "3 II T + O E l + Cj ^ CO in "'" fN sj e- o- 0- 0- "0 "0 O+ O+ 0- t> t> CH o+ "o O "O O+ ^0 o- -o "o -o o+o- "o e- oo Tf fN TT r- O CO SO ON Tf r- o CO SO OO TT oo in r-~ r-- rj- m m SO r~- r- r- OO OO ON ON ON ON o o o fN SO ON fN O O m oo fN m oo <*- ON fN so co sO SO 00 ^t '^f in *o sO sO sO r- r* r- oo oo oo ON ON ON ON O O m m 184 M. HARMAN, T. I. MOLLESON & J. L. PRICE Appendix VI Gazetteer of cemeteries in Great Britain containing decapitated and prone burials The following gazetteer has been prepared from a survey of the literature. The sites, which are arranged alphabetically, are divided chronologically into two broad groups, Romano-British and Anglo-Saxon; there is one site of doubtful period. Abbreviations used in the gazetteer (a) Journals, &c. A = Archaeologia AC= Archaeologia Cantiana AJ = Archaeological Journal B = Britannia BAR = British Archaeological Reports CB = Crania Britannica CBARR = Council for British Archaeology Research Report C/ = Glevensis JBAA = Journal British Archaeological Association JRS = Journal of Roman Studies MA = Mediaeval Archaeology N&QS&D = Notes and Queries for Somerset and Dorset O = Oxoniensia PC A S= Proceedings of the Cambridge Antiquarians Society PDNHAS= Proceedings of the Dorset Natural History and Archaeological Society PHFC = Papers and Proceedings of the Hampshire Field Club PSA = Proceedings of the Society of Antiquaries RCAHM= Royal Commission on the Ancient and Historical Monuments VCH = Victoria County History VEHSRP=Va\e of Evesham Historical Society Research Papers WAM = Wiltshire Archaeological Magazine (b) Other abbreviations A-S = Anglo-Saxon F = female B = burial M = male C = century P = prone burial Cr = cremation R-B = Romano-British D = decapitated burial Ref. = reference Ex = excavated The entries list the names of the site; county; National Grid reference; number of burials- B; number of cremations - Cr; number of decapitations - D; number of prone burials - P; dating - R-B or A-S; year(s) of excavation - Ex; and references - Ref. References not listed in full, other than to journals, &c., abbreviated as above, will be found on p. 169. Each site is preceded by a number by which it can be located on Fig. 5, p. 163). Gazetteer of Romano-British and Anglo-Saxon cemeteries containing decapitated and prone burials A. Romano-British cemeteries 1. Abingdon, Oxon. SU 483974. Bl 1. Dl; head between knees, adult M. PO. Probably late R-B. Ex. 1974. Ref. 5(1975)6 : 279; Parrington, M. (1978), CBARR 28 : 23-25, 36-37, 92. 2. Alcester, Warwicks. SP 089572. Bl 1: IF, 10 infants. Dl: head between legs, young F. PO. Late C4 or later. Ex. 1 975. Ref. B (1976) 7:331. 3. Baldock, Herts. TL 247339. B8: 2 adults, 6 juveniles. DO. PI: adult. R-B. Ex. 1968. Ref. 7/?S(1969)59:222. 4. Beckford, Worcs. SO 982361. BIO. Crl. D5: heads between knees or feet, some also had hobnails, one possibly in a coffin. PO. R-B. Ex. 1972. Ref. B (1973) 4 : 287; G (1973) 7 : 6-7; VEHSRP(\915) 5: 1-12. 5. Bloxham, Oxon. SP 4236. B30. Dl : adult M. P3 or 4. R-B. Ex. 1930s. Ref. Knight, W. E. J. (1938). BURIALS, BODIES AND BEHEADINGS 185 6a. Cassington, Oxon. SP 449103. B100+, Cr 2. D16: 7 buried prone. P14 or 16: 7 also decapitated. R-B. Ex. 1930s. Ref.JRS(