Petre Le oes ae) CCE ey Pe Oe, carte = + iT eens rer fetitaele® Seen 2 seers “+8! Tats 5 ay ena ise atti 2 Herat aye ast SHE et ees hee Petit Telit tates, tp abet eT titthne Apart ave ye mee 65 * 7 ¢ © \7 a) te ete G soa te RS Pae iarstetresetts islet Nate Prcely + MN pe itaets Ct ie Fae kD Site Veco Ok Se * S ¢ a Sete ‘ats See Rute Gilde ak Ae bare es SS \\ | Neo, fs SEMIO de ae JOM BUEN [s GENER a’ BULLETIN OF ( 2 5 smunes) THE BRITISH MUSEUM pee (NATURAL HISTORY) GEOLOGY VOL. 23 1973 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) LONDON: 1981 DATES OF PUBLICATION OFTHE PARTS No. 1 . 2 March 1973 No. 2 . 8 March 1973 No. 3 29 June 1973 No. 4 31 May 1973 No. 5 12 July 1973 No. 6 13 July 1973 No. 7 20 July 1973 ISSN 0007-1471 PRINTED IN GREAT BRITAIN BY HENRY LING LIMITED, AT THE DORSET PRESS, DORCHESTER DTI IHD CONTENTS GEOLOGY VOLUME 23 Marsupialia, Insectivora, Primates, Creodonta and Carnivora from the Headon Beds (Upper Eocene) of southern England. P. E. CRAY Some British Cretaceous gastropods belonging to the families Procerithi- idae, Cerithiidae and Cerithiopsidae (Cerithiacea). H. L. ABBASS A review of some English Palaeogene Nassariidae, formerly referred to Cominella. C. P. NUTTALL & J. COOPER Palaeozoic coral faunas from Venezuela, II. Devonian and Carboniferous corals from the Sierra de Perija. C. T. SCRUTTON Prodeinotherium from Gebel Zelten, Libya. J. M. HARRIS Cirripedes from the Upper Cretaceous of Alabama and Mississippi, eastern Gulf Region, U.S.A. I. Palaeontology. J. S. H. CoLLINs II. Geology. F. F. MELLEN Fenestrate bryozoa from the Viséan of County Fermanagh, Ireland. R. TAVENER-SMITH Index to Volume 23 PAGE 103 ET 221 283 349 381 389 495 i pedis. hier? ae Son paidaiey' | a a +¢@a>? 2 ) a ' 5 ty oe > ew Al ‘ : ~~ 2° i sv ev Le wTiot 4. -.. eee ee | ftpmect any? « y neg? 7 '@ ) Sani ok INDEX TO VOLUME 23 New taxonomic names and the page numbers of the principal references are printed in bold type. An asterisk (*) denotes a text-figure. AsBass, H. L. 103-75 Acanthocladiidae 483-90 Actinostoma 409, 411 Acinophyllum stramineum 231 vermetum 230-1 Adapidae 65-6 Adapis 65, 66-73 magnus 16, 66, 67-71, 72-3; pl. 1; pl. 2. fig. 1 minimus 66 parisiensis 27, 66, 71-3; pl. 2, figs 2-3 priscus 66, 71-2 ruetimeyeri 66, 71-2 sciureus 66, 71-2 Adapisorex 38-9 Adelomys depereti 12, 16 siderolithicus 12, 27 Adunator 39 Africa 291 Afrocyon burolleti 288 Ageria 106, 110, 156-60, 169 ? angustata 57-8; pl. 8, fig. 13 costata 108, 158, 159; pl. 8, figs 7-12 gaultina 108, 148, 156, 157-9; pl. 8, figs 1-6 Agerina 74 Albian 106-8, 114, 119, 123-4, 127-8, 130, 133-4, 137, 155, 157, 159-60, 167 Alum Bay 18-19, 24, 199 Amphichiromys europaeus 63 Amphilemur 66 Amplexizaphrentis 239-42 cassa 242 sutherlandi 224, 230, 232, 239, 240-2, 241*, 281; pl. 1, figs 9-13 Amplexus intermittens 234 Anchilophus sp. 16 Anchomomys 65-6 Anchura carinella 124 Ancillaria bretzi 208 Andagularia 142 Androconus 48 Annona 351, 358-9, 379, 381-2 Anomia 409 Anoplotherium 71 anthracotheroid incertae sedis 16 Anthracotherium 15 Apatemyidae 61—5 Apatemys 61-2 bellus 63 Aphanolemur 66 Apheliscinae 47 Aphronorus 48 appendicular skeleton 330 Aporrhaidae 107 Aptian 106-9, 111, 113, 118, 120, 122-3, 126-7, 132-3, 138, 162, 165 Aquitanian 288 Archaeofenestella 461 Arcoscalpellum 362-74, 381 bakeri 351-2, 367-8, 381; pl. 3, figs 10-13 campus 351-3, 369-71, 373, 381; pl. 3, figs 14-16; pl. 4, figs 1-4 comptium 371 conradi 352 fossula 353, 366-7, 371, 373 gracile 373 hubrichti 351-3, 357, 363-7, 365*, 370, 373, 381-4; pl. 2, figs 13-20; pl. 3, figs 1-9 maximum 353, 367 michellotianum 363 simplex 362 solidulum 368 sulcatum 367 velutinum 363 withersi 351-3, 366, 370, 371-4, 381; pl. 3, figs 17-18; pl. 4, figs 5-15 Arctica morrisi bed 203, 214 Artiodactyla 16 Arvaldus 44-5, 46-7 stintoni 27, 34, 45-7, 46* Asia 84, 89-90, 134, 291 Aston Clinton 136 astragalus 295, 340-1 496 INDEX atlas vertebra 290, 319-21 Atherfield 107, 113, 131-2, 142, 144, 146, 162, 165 Austinian 381 Australia 180 Austria 152, 184 Auversian 183-4, 193, 208 axial skeleton 319 axis vertebra 290, 322-4 Babylonia (Peridipsaccus) archambaulti 205-6; pl. 5, fig. 8a, b Barton Beds 182, 186-8, 190-1, 199-201 Bartonia 181 canaliculata 181, 182*, 183, 192 Ballyshannon Limestone 393 Barytherium 344 Barytichisma 239 Bathonian 150 Bathraspira 110, 123, 124-30, 169 beaudouini 126 brightoni 108, 123, 129-33; pl. 3, figs 9, 11, 14, 15 cleevelyi 108, 124, 125*, 127-8; pl. 3, fig. 10 ervynym 129 excavata 127 fouadi 108, 125*, 126, 128-9, 130; pl. 3, figs 7-8, 12-13 multistriata 126 neocomiensis 123-7, 129 ornata 126 shanklinensis 109, 124, 126-7; pl. 3, figs 3-6 subornata 127 tecta 108, 123-6, 125*, 127, 129; pl. 3, figs 1-2 Batillaria concava 22 Belgium 62, 138, 161, 184, 188, 193, 205-6, 212 Benbulben Shales 393 Bessoecetor 48-50 levei 48, 50 Billingsastraea 256, 263 Billingsastraeinae 256-64 bivalves, see lamellibranchs Blackdown 107, 130, 148, 157, 159-60 Bognor Rock 188, 203 Bordeaux 188, 212 Bowenelasma 231, 242-3, 244-8 breviseptata 224, 230, 245-7, 280; pl. 3 typa 224, 230, 243-5, 253, 280-1; pl. 2 brachiopods 228 Brachylepadidae 380 Brachylepadomorpha 353, 380 Brachylepas 380 angulosa 351-2, 380, 381; pl. 5, fig. 16 niassanti 380 Brachysphingus 180-1, 183, 208-9 gibbosus (208); pl. 8, figs 3-4 Bracklesham Beds 186-8, 193 Bramshaw 187, 193-4 Briantelasma 231, 243, 247-51, 280-1 americanum 248 oliveri 224, 230, 248-51, 249*, 253, 280-1; pl. 4 Brightonella 123 sandlingensis 123 turris 123 British Museum (Natural History) 183, 186, 188, 196 Brockenhurst Bed 187-8, 191, 193, 197, 205, 210-11 Brontops 341 Brook 193-4, 211 Brotia 138 bryozoa 398, 409 Buccinacea 179, 189-215 Buccinanops 188; see Bullia Buccinidae 179-83 Buccinum acies 201 armata 196 auversiense 208; pl. 7, figs 6, 7 bicorona 212-3, 215 bolli 212 bullatum 187, 193 aspera 189 canaliculatum 180-1, 192 cassidaria 212 cinctum 180 desertum 180, 183, 189-90, 192, 198 desori 201-2 excavatum 190 flexuosum, see Cominella fusiforme 183, 192-3, 195; pl. 3, figs 7, 8 fusiopsis 193, 195-6 gossardi 180, 212 latum 201 lavatum 192 montense 180, 183-4, 215; pl. 9, fig. 8a, b ovatum 201 porcatum 180 quoesitum 213-5 semicostata 196 solandri 198; see Cominella solanderi suturosum 204; pl. 6, fig. la, b testudineum 180 uniseriale 212 ventricosum, see Cominella sp. nov. 198 Buccitriton 180, 213 Bullia 181, 188 (Buccinanops) monilifera 181 (Molopophorus) striata 208 Bundoran Shale 393 Bunolistriodon massai 288 Burdigalian 184, 212, 288 Burham 136 Burnupena 180 cincta 180 Buxolestes 48, 50 Caenopithecus 65-6 Calantica saskatchewanensis 352 (Titanolepas) martini 352 Calcaire Grossier 188 Calenscio Serir 286 California 180, 208-9 Callista kickxi horizon 205 Calp Sandstone 393 Campanian 111, 117, 351, 379, 381-2 Campanile 109, 138, 150, 169 breve 152 carezi 152 cenomanica 108, 151-2; pl. 6, figs 1, 2 giganteum 152 morgani 152 Campanilinae 151 Campo Chico Formation 226 Cano Colorado 240, 281 Cano del Noreeste Formation 229 Cano del Norte 227, 252, 281 Cano del Oeste 226, 281 Cano del Sur 280 Cano Grande Formation 226, 280 Cano Indio Formation 229 Cano Los Guineos Formation 228, 280 Cano Pescado 281 carina 354-9, 363-4, 367-75, 378, 380; pls 1-4 Carnivora 16, 89-93 Carrick Lough 392, 394, 397, 399-400, 404 Catodontherium sp. 16, 23 Cebochoerus 12, 16 INDEX Cenomanian 108, 111, 115, 119, 121, 136, 138, 140-1, 151, 155 Centetodon 44-5 Cerithiella 160, 165, 167-8 atherfieldensis 109, 161*, 162; pl. 6, fig. 10 cloacina 161*, 162 devonica 108, 160-2, 161*; pl. 6, fig. 9 metula 161* nodoliratum 160, 161* prealonga 162; pl. 6, fig. 8 semirugatum 161* Cerithiidae 150-9 Cerithiopsidae 160-7 Cerithiopsis 146 meeki 164 quadrilirata 164 Cerithium 118, 122 aculeatum 109, 112-3 aptiense 133 ascheri 109, 138-40 attenuatum 109, 113-4 belgicum 138, 140 bicostatum 161 binodosum 110, 157-8 chargense 111 cingillatum 152-3 clementinum 109, 141-3, 145 cribriforme 166-7 cuckhamsliense 109, 116 davonstianum 110 delpeyae 109, 137 derignyanum 135 (Fibula?) detectum 122 disparile 133 distinctum 149 dupinianum 145 ervynum 124 excavatum 115 fertile 137 forbesianum 109, 141, 143-5 frickei 109, 114 gallicum 109 gottfriedi 109, 113 gurgitus 125 hector 110, 157 hispidulum 109 inauguratum 109 kaunhoweni 149 leufroyi 165 metula 160 millepunctatum 141 mosense 109, 135-6 navasi 110, 154 neocomiensis 130 ornatissimum 135-7 peregrinorsum 146 Phillipsi 110, 143-5 pseudoclathratum 154 pustulosum 110, 153-4 rhodani 135 ricordeanum 109, 139 rochati 115 sabandianum 137 sanctaecrucis 109, 139 saundersi 110, 153 schrammeri 126 subattenuatum 113-4 497 498 INDEX subspinosum 118-9 bicoronata 214 terebroides 109, 139 bonnecarrei 201 tectiforme 166 canaliculatum 181 tectum 123 deserta 180-1, 190, 193, 195-6 trimanile 133 solanderi 192, 198 tuberculatum 133 ‘ desori 202 turriculatum 109, 131, 133 flexuosa 208, 210 Cervella cervoidea 435 gossardi 180 cervical vertebra 290, 325-6 montense 185 Chadronia 48 porcatum 180 Chalons-sur-Vesle 188 praecedens 212 Chattian 184 solanderi, solandri 189, 198, 200 Cheilostomata 402, 459 supracostata 201 China 65 suturosum 204 Chinji 305 testudinea 180, 182*; pl. 9, fig. 1la, b Choeropotamus depereti 16 ventricosa 204, 206; pl. 6, fig. la, b Cimolithium 109-10, 138-41, 169 (Buccinum) flexuosa 210 ascheri 109, 138-40; pl. 2, figs 6, 7, 9, 10 solanderi 198 belgicum 140-1 ventricosa 204 aff. eleanorae 108, 138, 140-1; pl. 2, fig. 5 Cominellidae 179, 181 gallicum 138 Compressiphyllum 251 hispidulum 138 Condranema 471 insuguratum 138 Coon Creek 352, 383-4 komarense 140 Cooper, J. 177-219 miyakoense 139-41 corals, Middle Devonian 224 Cirripedia 354 Lower Pennsylvanian 224 Cirsocerithium 110, 118-9, 120-2 Corbula bed 203, 214 aptiense 121 Cornulina minax 196-7 gauthieri 119 Cray, P: E. 1-102 harborti 120-1 Creodonta 16, 26, 83-9 kirkaldyi 109, 118, 120-1; pl. 2, fig. 8 Cretiscalpellum 353, 354, 381, 384 nooryi 108, 119, 121-2; pl. 2, figs 3, 4 glabrum 355 peroni 119, 121 harnedi 351-3, 355, 357-8, 381; pl. 1, figs 10-12 reticulatum 118 macrum 351-3, 354-5, 357-8, 381; pl. 1, figs 1-6 subspinosum 108, 118, 119-20, 121-2; pl. 2, striatum 357 figs 1, 2 subcarinatum 357 Clava 122, 153 unguis 354-62 Coffee Formation 351, 358-9, 379, 382 vallum 351-2, 357, 358-62, 361*, 381-4; pl. 1, CoLuins, J. S. H. 349-80 figs 14-22; pl. 2, figs 1-12 Colombia 231 venustum 351-3, 355, 356-7, 381; pl. 1, figs Columella 179, 181, 182* 7-9, 13 Colwel] Bay 197-8, 210 Cryptaulax 156 Colwellia 180-1, 208-9, 212 angustatum 158 antiquata 184, 208; pl. 7, fig. 11 Cryptopithecus 48-50 auversiensis 184, 208-9, 211; pl. 7, figs 6, 7 sideroolithicus 49 bretzi 184, 208; pl. 7, figs 8, 9 Cryptostomata 409 cretacea 184, 208 Cuisian 184, 201, 204 flexuosa 184, 208-9; 210-2; pl. 7, figs 1-5 cuneiform 290, 295, 330, 336-7 tejonensis 184, 208 Cyathaxoniicae 233 Cominella 179-81, 182*, 183, 190, 209, 215; see Cyathophyllidae 255-64 Northia Cyathophyllum 257 annandalei 212 goldfussi 269 aturensis 212 limbatum 269 bicorona 214 (Peripaedium) turbinatum 257 venezuelense 252-4, 257 Cylindrophyllum 224, 256, 262-4, 262* elongatum 230, 232, 262*, 263-4, 281; pl. 7, figs 3-6 Cyllena 188, 209 Cystiphyllidae 269-72 Cystiphyllina 269-72 americanum 269-70 conifollis 272 Czechoslovakia 157-8 Dacrytherium ovinum 16 Dalmanophyllum dalmani 243 Danian 134 Dartry Limestone 393 Dasyurodon 84 Deinotheriidae 285, 291, 344 Deinotherioidea 291 Deinotherium 285, 292-3, 345 bavaricum 289, 292, 294-5 bozasi 291-3, 296, 312, 345-6 cuvieri 296 giganteum 286, 289, 291-3, 296, 305-6, 308, 312, 337-9, 341-5 gigantissimum 293, 329, 342 hobleyi 292, 296 indicum 291-3, 345 levius 312 pentapotamiae 292 Delaware Basin 399 dentition 290, 295, 313; pls 4—5 Derrygonelly 392 Desorinassa 180-1, 201-2, 208-9, 213 acies 184, 201, 203-4; pl. 5, fig. 5 bonnecarrei 184, 201 desori 184, 201, 202-3, 204; pl. 4, figs 4-7 lata 184, 201; pl. 4, figs 8, 9 ovata 184, 201, 203-4; pl. 5, fig. 4 supracostata 184, 201 williamsi 180, 184, 188, 201, 203-4; pl. 5, figs 1-3 Devonian 229 Lower 231, 251 Middle 224, 231, 233, 248, 251-2 Diatinostoma 150 Dichobune 13 Dichodon 15, 19-20 cervinum 12, 16 cuspidatus 15-16, 21, 27 Didelphidae 27, 28-36 Didelphidectes 28 Didelphinae 28 Didelphis 29 Didymictis 90 INDEX 499 Didymoconus 49 Dinantian 393 Diplocynodon 25 Diploschiza melleni 381 Disphyllidae 255 Ditoecholasma dalmani 243 Ditretus 150 Dor el Talha 289 Dorsanum 185, 188 Dover 135, 155 Durhamina 224, 230, 232, 265-8, 281 cordillerensis 268 sp. nov. 266-8; pl. 8 ? sp. 265-6; pl. 7, figs 7-8 Durhaminidae 264-5, 266-8 Dyspterna 48-50, 53, 57-61 helbingi 57 hopwoodi 26-7, 57, 58-61, 59*, 60* woodi 57-9 Egypt 134 Eifelian 231 Elephas 285, 295, 305, 329, 336, 338, 342, 344-5 maximus 286, 289, 295, 330, 331*, 332, 334-43 Emmonsia 272 carmani 274 radiciformis 274 Emsian 231 Engelhardtia 25 Enniskillen 392 Enniskillenia 239 Eocene 180 Eochiromys 62 Eodevonaria 231 Epapheliscus italicus 57 Eppeslheim 305 skull 306-8, 310 Europe 28, 36, 48, 62, 89-90, 138, 179-80, 189, 291 Exechocirsus 110, 152-6, 169 pustulosus 154; pl. 7, figs 6, 7 saundersi 107-8, 153-5, 169; pl. 7, figs 3—5 aff. saundersi pl. 7, figs 1, 2, 8, 9, 12-14 aff. subpustulosus 108, 155-6; pl. 7, figs 10, 11 Exelissa 169 Favosites 224, 230, 272, 273-5 arbuscula 230-1, 274-5, 280; pl. 10, figs 3-6 digitata 275 gothlandicus 272 venezuelensis 230, 273-4, 280; pl. 10, figs 1-2 Favositidae, Favositina 272 : femur 290, 295, 330, 339-40 500 INDEX Fenestella 401-2, 409-10, 424, 440, 459, 462, 465, 472-5, 482, 484 albida 477 cf. albida 406, 451-3; pl. 14, figs 1-6 alternata 477 antiqua 410 cf. arthritica 406, 431-3; pl. 7, figs 8-9; pl. 8, figs 1-4 benskiensis 470 binodata 468, 470 cerva 448 colymaensis 470 compressa 430-1, 433, 443 cornuta 435 crassa 409-10 crockfordae 450, 462 delicatula 406 cf. delicatula 407, 455-7, 469; pl. 15, figs 5-8; pl. 16, figs 1-3 demaneti 430-1 eichwaldi 450 ejuncida 428 fanata carrickensis 406-7, 436-8; pl. 9 fenestratum 413, 415 cf. filistriata 406, 444-6; pl. 11, figs 5-10; pl. 12, fig. 1 frutex 406, 411-3, 416; pl. 1 funicula 443 cf. funicula 406, 441-4; pl. 10, figs 7-10; pl. 11, figs 1-4 ghzelensis 443, 448 gracilis 462 hemispherica 406-7, 421-3, 425; pl. 4, figs 7-9; pl. 5, figs 1-4 iowensis 427-8 irregularis 406, 460-2; pl. 17, figs 2-7 ivanovi 406-7, 413-5, 466-7; pl. 2 limbata, limbatus 411, 415 lyelli 411 matheri 413 milleri 410 modesta 406, 418-21; pl. 4, figs 1-6 multiporata 457, 465 multispinosa 406, 416-8; pl. 3 var. karakubensis 418 ninae 466 nododorsalis 433 oblongata 406, 453-5; pl. 14, figs 7-8; pl. 15, figs 1-4 oculata 470, 472 parallela 406, 423-5; pl. 5, figs 5—9 placida 444 plebeia 406, 410, 428-31, 441; pl. 7, figs 1-7 var. longifenestrata 430 plebeja 428 plummerae 466-7 polyporata 406, 457-60, 462, 465; pl. 16, figs 4-8; pl. 17, fig. 1 praemagna 406, 433-6, 448; pl. 8, figs S—11 praerhomboidea 466-7 prisca 410 pseudovirgosa 406, 448-51; pl. 13 quadradecimalis 462 regalis 450, 477 reticulata 410 rudiformis 427-8 rudis 426-7, 430-1 major 427 multinodosa 406, 425-8; pl. 6 sevillensis 418 cf. spinacristata 406, 439-41; pl. 10, figs 1-6 subantiqua 409-10 subspeciosa 406, 446-8; pl. 12, figs 2-8 tenax 407 tenuifila 421 tenuivirgata 457 triserialis 470 varicosa 473, 475 varifenestrata 436 virgosa 448 Fenestellidae 409, 483 Fenestrella 409 Fenestrellina 409-10 multispinosa 416 Ferruginous Sands 107, 120, 122, 126 Fibula, see Cerithium detectum fibula 290, 330, 340 Fistulipora 420, 440, 462 Floresta 231 Folkestone 107, 114, 119, 124, 128, 130, 133-5, BIG IST, WG Fort Benton 352 France 5, 62, 65-6, 71, 74-5, 80-2, 84, 90-1, 119, 124, 130, 133, 136, 146, 157, 183-4, 192-4, 202, 212, 214-5 Fusus desertus 190, 193 Galba spp. 24-5 Galethylax 48 Gebel Zelten 285-6 Geolabis 44 Germany 48, 63, 74, 80, 91, 119, 124, 126, 133, 136, 138, 184, 187, 205 Gesneropithex 65—6 Glass Mountains 403 Glencar Limestone 393 Goniacodon 49 Gorgonia antiqua 410 INDEX 501 Grewinkia buceros 243 europheum hosholmensis 243 Guildford 132 Gymnocerithium 122 Gymnophyllum wardi 237, 239 Hadrophyllidae 237 Hadrophyllum 230, 232, 237-9, 238*, 281 orbignyi 237 sp. 237-9, 238*; pl. 1, fig. 8 Hamilton Group 230-1 Hampshire Basin 180 Haplaletes 49 Haplobunodon lydekkeri 16, 27 sp. 16, 20 Hapsiphyllidae 239 Harris, J. M. 283-348 Headon Beds 186-8, 191, 193, 205-6, 210, 212 Heliophyllum 256-63 halli 227, 230-1, 253, 256-60, 262*, 280-1; pl. 5, figs 6-8; pl. 6, figs 1-6 wellsi 224, 230, 261-2, 262*, 280; pl. 6, figs 7-8; pl. 7, figs 1-2 Hemicerithium 118 Hemipsalodon 84 Hemitrypa 435, 440, 475-8 burulica 478 hibernica 406, 435, 475-8; pl. 22 oculata 475 reticulata 475 Herne Bay 214 Herpetotherium 28 Heterochiromys 63 fortis 63 gracilis 63 Heterohyus 12, 62-5 armatus 62-5 europaeus 63-5 heufelderi 63-5 nanus 62-5 quercyi 62-5 Heterophrentis 224, 243, 251-5 inflata 253 prolifica 253 simplex 230-1, 253-5, 280; pl. 5, figs 3-5 spissa 253 venezuelensis 227, 230, 252-3; pl. 5, figs 1-2 Highcliffe 199 Hitchin 116, 154 Holland 184, 205 Hordle 187 Howenegg 286, 289, 343 Humbertia 90 angustidens 9 hastingsiae 93 helvetica 90 humerus 290, 295, 330, 332, 334 Huntingbridge 183, 193-4 Hyaenodon 13, 84-9 aimi 15, 85-7 ‘hantonensis’ 85 leptorhynchus 84 cf. minor 16, 27, 85-9; pl. 5 Hyaenodontidae, Hyaenodontinae 83-9 Hyainailouros 84 Hyopsodontidae 47 Hypacodon 45 Ictidopappus 90 India 137 Insectivora 16, 36-65 Todictyum 476 Ischnognathus 84 Isle of Wight 107, 120, 122, 127, 131, 144, 146, 165 Japan 118, 127, 133, 138, 209 Jepsenella 61 jugal 301 Kansas 362, 400 Keepingia 180-1, 188, 201, 209, 211-2 annandalei 184, 212 aturensis 184, 212 bolli 184, 212; pl. 8, fig. 8 cassidaria 184, 188, 211-2; pl. 8, fig. 9 cancellata 211 gossardi 184, 188, 211-2; pl. 8, figs 5-7 praecedens 184, 212 tarbellica 184, 212 uniseriale 184, 212 Keriophyllum dahlemense 257 Kiaerophyllum semilunatum 243 Kindle Formation 242 Kotyhaza 294, 343 Labidolemur 61-2 lacrimal 301 lamellibranchs 398, 409 Lantianus 65 Latdorf 193, 205-6 Lattorfian 184, 187, 205, 212 latus, carinal 359-60, 363-4, 366, 369-71, 373, 375-6; pls 2-5 inframedian 358, 362; pl. 2 rostral 358, 362, 371, 373, 375-6; pls 2-3, 5 subcarinal 358-60; pl. 2 upper 358-60, 363-4, 366, 373, 375; pls 1, 3-4 502 Leakitherium 84 Leighton Buzzard 108, 124, 127 Leitrim 393 Lepadomorpha 354 Lepidosteus 14 Leptacodon 36, 38, 44 Jepseni 37 ladae 37-8 minutus 37 tener 37-8, 44 Leptocoelia 231 Levifenestella 463-5, 482 maeve 463 undecimalis 406, 463-5; pl. 17, figs 8-9; pl. 18, figs 1-5 Lewes 136 Libya 286 Limnaea Marl 8, 10 Lindstroemiidae 233 Litomylus 49 London Clay 188 Lonsdaleia cordillerensis 265 Loricultina texanum 352 Lough Erne 393 Lough Melvin 393 Loxodonta 341 africana 286, 289, 319, 321*, 323*, 324-6, 326* lunar 337 Lutetian 183-4 Lyndhurst 187, 197 Maastrichtian 111, 122, 149, 161, 166, 351, 353, 357, 367, 371-2, 374-5, 378-82 Maboko Island 342 Madagascar 119, 124, 134, 137 Mainz Basin 188 magnum 295, 337-8 mandible 290, 311; pls 4-5 Manitoba 352 manus 295 Manzati 342 Marada Formation 287 Marisastrum sedgwicki 256 Marmosa 34 Marsupialia 16, 26-36 Mastodon 316, 327, 336, 338, 344 angustidens 289, 295, 319, 333*, 334-5, 337-41, 343-4 Pygmaeus 288 maxilla 290, 300; pl. 5 Megalopterodon 84 Melania scalariformis 141 Melanopis attenuatum 114 MELLEN, F. E. 381-8 INDEX Melongena (Cornulina) minax 196-7 Merida Andes 229 Merstham 136 Mesalia (Bathraspira) tecta 123 Mesestoma 157 Messelina 38 , metacarpals 295 Metacerithium 109-10, 131-6, 169 abictiforme 134 amudasiensis 134 dentatum 137-8 ornatissimum 108-9, 134-5, 135-7; pl. 4, figs 4-5, 11-12 renngarteni 137 rikuchuense 132-3 stoliczkai 134 trimonile 108, 131-2, 133-5, 137; pl. 4, figs 1-3, 6-8, 10 turriculatum 107, 109, 131-3, 137; pl. 5, figs 8-11 sp. nov. 108, 137-8; pl. 4, fig. 9 Metachriacus 49 Metapterodon 84 Mexico 140 Miacidae 89-93 Microchoerinae 73-83 Microchoerus 13, 21-6, 29, 74-80, 81 edwardsi 75, 80 erinaceus 12, 16, 23, 74, 75-80; pls 3-4 ornatus 75, 80 Mimotricentes 49 Minilya 465-72 binodata 406, 468-70; pl. 19, figs 5-7; pl. 20, fig. 1 duplaris 465 mimica 468 ninae 467 nodulosa 406-7, 470, 490 oculata 406, 470-2; pl. 20, figs 2-8 plummerae 406, 466-8; pl. 18, figs 6-8; pl. 19, figs 1-4 praerhomboidea 467 rhomboidea 468 Miocene 287, 291-2, 296 Mississippian 242 Molopophorus 180-1, 183, 209 antiquatus 209; pj. 7, fig. 11 cretacea 209 gabbi 180 striatus 180-1, 208-9; pl. 8, figs 1-2 tejonensis 208 Montian 215 Mooreville, Middle 381-2 Munich skull 308-10 Myolestes dasypelix 37, 44 Myrica 25 Nannopithex 74, 81 Naples, Bay of 182 Nari Stage 184, 212 nasal 301 ‘Nassa’ 183; see Nassarius antiquata 208-9 cretacea 208-9 tarbellica 212 Nassarlidae 177-219 Nassarius 188, 190, 213 arcularius (181); pl. 9, fig. 9 Nassidae 180; see Nassariidae Nasua 61 Navarroan 381 Necrolemur 74-5, 81 antiquus 75 parvulus 80 raabi 80 Neocomian 109, 143, 159 Neogastropoda 189-215 Neogene 296 Nerinaea 153 Nerineopsis 109, 110-8, 169 aculeatum 109, 112-3, 115; pl. 1, figs 5-8 adeli 108, 111, 116, 117-8; pl. 1, fig. 4 attenuatum 113 claxbiensis 109, 111; pl. 1, fig. 3 coxi 108, 114-5; pl. 1, figs 10-11 cuckhamsliensis 107-8, 111, 116-7; pl. 1, figs 3-4 davonstiana 110-1, 114 gottfriedi 115 longi 115 melburiensis 108, 111, 115—6, 117; pl. 1, figs 1-2 subattenuatum 109, 111, 113-4, 115; pl. 1, figs 9, 12 Nettleton 111, 147, 159 Neurogymnurus major 49-SO, 57 neuropophysis 290 New Forest 186 New Jersey 352 New York 230 New Zealand 179-80, 182-3 Ngorora 296 North America 28, 36, 48, 61-2, 84, 89-90, 184 Northern Ireland 107, 117 Northia (Cominella) angusta 185 bullata f. armata 196 suturosa 205 Nudivagus 110, 122, 123 ? cooperensis 122 INDEX 503 morrisi 109, 122-3; pl. 2, figs 11-12 simplicus 122 Nummiulites elegans 199 NUTTALL, C. P. 177-219 Nyctitheriidae, Nyctitheriinae 36-47 Nyctitherium 36-9, 44 daspyelix 37 serotinum 37 velox 44 Oligocene 180 Oliva 209 Onesquethaw 224, 230-1, 236, 252 Onondaga Limestone 230-1, 251 Ontario 231 Opsiclaenodon 48-57, 58 major 12, 16, 49, 50-7, 52*, 53*, 54*, 55*, 57-61 Oregon 180, 208 Orthochetus 165-8; pl. 6, fig. 15 charlesworthi 166*, 167 elongatus 166*, 167 hantoniensis 109, 165-7, 166*; pl. 6, fig. 12 helmyi 108, 167-8; pl. 6, figs 13-14 leufroyi 166*, 167 mapeulensis 166* tectiformis 166* Oxyclaenus 49 Pagonomus 48, 50 Pakistan (West) 184, 212 Palaeocene 180, 352 Palaeocoryne 435, 443, 459 Palaeogene 177-219 Palaeosinopa 50 Palaeotherium 13, 15-16, 19, 21, 26-7 medium 15 muehlbergi muehlbergi 27 praecursum 16, 23, 27 palatine 302 Palencia 307 skull 309-10 Palmarito Formation 228-9 Paloplotherium 13 annectens 15 Paludina lenta \7 palynomorphs 229 Pantolestes 48, 50 Pantolestidae, Pantolestinae 47-61 Pantomimus 48, 50 Paracerithium 110, 118-9, 121, 131, 142 parietal 302 Paris Basin 187-8, 194, 204 pelvis 339 504 Penniretepora 440, 489 Pennsylvania 402 Pennsylvanian, see corals Pentagonia 231 Peratherium 12, 16, 23, 27, 28-36, 30*, 31*, 35* crassum 28 cuvieri 28 Taurillardi 29 Periconodon 66, 74 Peridipsaccus, see Babylonia Peripaedium, see Cyathophyllum turbinatum Perissodactyla 16 Permian 402 Phillipsastraeidae 255 Phos 180, 212-3 Pikermi 343 Pinus 25 Pirinella 147 Plagiolophus annectens 16, 27 minor 16 Planorbina 24-5 Plasmophyllinae 269 Plasmophyllum 224, 230, 269-72, 280 secundum americanum 230-1, 269-71, 280; pl. 9, figs 1-4 conistructum 270-1 sp. 271-2; pl. 9, figs 5-7 Pleistocene 291 Pleurodictyum venezuelense 273 Pliocene 291-2, 296 Pollia 183, 215 lavata 192 montense 185, 215 Polypora 398, 410, 438, 464, 479-82 dendroides 406, 478-80; pl. 23 radialis 413 septata 413 stenostoma 406, 490 verrucosa 406, 480-2; pl. 24 Pontian 292 Potamides 147, 153 (Tympanotonus) radula 113 Potamomya plana 17 Prairie Bluff 351, 375, 378, 381-2 premaxilla 300 Primates 16, 65-83 Priabonian 180, 184, 187, 190, 197, 199, 204-6, 208, 210 Proboscidea 291 Procerithtidae 110-50 Procerithium 110, 141, 146 Procynodictis euzetensis 91 Prodeinotherium 285-6, 293-330 bavaricum 291, 294, 312, 343, 345 INDEX hobleyi 291, 294, 296, 297*, 298*, 299*, 304*, SK Sl, Bie BiG, SIG), SAO, Gv Syst 325, 326; 33)l=5332,9395 554,.555—46-mpls 1-5 pentapotamiae 291, 294, 305, 312, 345 Prolibytherium magnieri 288 _ Pronycticebus 65-6 Propalaeosinopa 48-9 Prophet Formation 242 Propterodon 84 Protoadapis 65-6 Protictis 90 Protofusus 110, 142 Pseudamphimeryx hantonensis 16 Pseudocominella 180-1, 184, 188, 189-90, 201-2, 204, 208, 213 armata 184, 188-93, 193-6, 198; pl. 3, figs 1-8 bullata 184, 189, 193, 196, 198; pl. 2, fig. 9a, b aspera 184, 189 deserta 182*, 184, 188, 190-3, 196-7, 200-1; pl. 1 semicostata 180, 184, 189-90, 196-8; pl. 3, figs 9-10; pl. 4, figs 1-3 solanderi 184, 188-9, 193, 196, 198-201, 213; pl. 2, figs 1-8 Pseudoloris 74, 80-3 abderhaldeni 80 parvulus 16, 80, 81-3, 82* reguanti 80 Pseudophaulactis 243 Pterodon 84 Ptilofenestella 401 carrickensis 406, 467, 490 Ptiloporella 460, 471, 472-5, 489 varicosa 406, 473-5; pl. 21 Ptylopora 398, 471, 473, 475, 483-7 pluma parva 406, 483-7; pl. 25 Quaternary 296 Quercygale 90-3 angustidens 90-1 hastingsiae 16, 91-3; pl. 6 helvetica 91 radius 290, 295, 330, 333*, 334-5 Reculver 202 Retepora polyporata 457 reticulata 410 undulata 452-3 Rhabdocolpus 109-10, 140-6, 169 clementinum 109, 142-3; pl. 5, figs 6, 12-13 forbesianum 109, 143-5, 146; pl. 5, figs 1-5 manselli 142 melvillei 109, 145-6; pl. 5, fig. 7 INDEX 505 quehenensis 142 Siphonalia 190 undulatus 142 bicarinata 190 Rhabdomeson 471 Siphonophrentis 243 Rhinoceros 341 Sirte, Basin 287 rib 290, 327 Gulf of 286 Rio Cachiri Group 226, 230-2 skull 290, 295-6; pls 1-3 Rio Palmar Formation 232, 240, 281 Sligo 393 Ripley Formation 351-2, 357, 367, 369, 371-2, Smilodectes 66 374-5, 378-82 South Africa 180 Rodentia 16 South America 181 rostrum 380; pl. 5 Spain 74-5, 80, 82 Royden 187, 197, 205-6 Spalacodon 13, 37, 41 Rugosa 233-72 Spanoxyodon 49 rugose horn-corals 398 Sparnacian 184 Rupelian 184, 204-5, 212 Sphaeronassa mutabilis 182* sphenoid 303 Sabaneto Group 229 Spinulicosta 231 Sagenella 180, 213 Squama spissa 352 Santonian 381-2 squamosal 304 Sarmatian Flinzsande 308 Stehlinella 62 Saturninia 36-8, 44—5 Stereolasma 224, 230, 233-4, 280 Scalpellidae 354 sp. 233; pl. 1, figs 6-7 Scalpellum 374 Stewartophyllum 224, 230, 232, 234-5, 281 gabbi 374 intermittens 235 inaequiplicatum 352 sp. 235; pl. 1, figs 4-5 scapula 290, 295, 330-2, 331* Stramentum canadensis 352 Scenopagus 38 haworthi 352 Schloenbachia varians 136 Strepsidura 190 Scraeva 23, 37-43, 44-5 armata 183, 189, 193, 196, 198; pl. 3, figs 9-10; hatherwoodensis 23, 37, 39-41, 40*, 43 pl. 4, figs 1-3 woodi 12, 16, 37, 40, 41-3, 42*, 45-6 deserta 190, 193 Scrupocellaria 424 armata 196 SCRUTTON, C. T. 221-81 semicostata 193, 198 scutum 354-5, 357-9, 363-4, 369-75, 379; pls 1-5 turgida 211; pl. 9, fig. 10 Seila 160, 162-4, 167-8 (Buccinum) armata 196 dextroversa 164*, 168 semicostata 196 iglali 108, 163-4, 164*, 169; pl. 6, fig. 11 Streptelasma rectum 233 meeki 164* Streptelasmatidae 242 mundula 164 Streptelasmatina 233 quadrilirata 164* Strombodes helianthoides 256 Selma Chalk 351, 381-2, 384 simplex 254 Senonian 108, 134, 157-8, 352, 354, 356, 358-9, Strombus 113, 213 363-4, 368, 379 subcarina 358, 360; pls 1-2 Septopora 392, 487-90 subrostrum 359, 362; pl. 2 cestriensis 487, 489 supraoccipital 302 hibernica 406, 487-90; pl. 26 Switzerland 63, 65-6, 74-5, 91, 124, 133, 136-7 subquandrans 490 Syringaxon 224, 230, 236, 280 Serratocerithium 153 arnoldi 236 Shaftesbury 115 rudis 236 Sierra de Perija 229, 251 sp. 236; pl. 1, figs 1-3 Silurian 231, 251 Simpsonictis 90 Tabulata 272-5 Sinclairella 62 Tabulipora 462 Sinodia suborbicularis 22 Tapocyon 90-1 506 INDEX tarsal bones 340 Unichinia 62 Tarsiidae 73-83 Unio 10, 17 Tarsius 73 United States 138, 180, 213 TAVENER-SMITH, R. 389-493 Upper Calp 393 Tayloran 381 WESSSaReeIey teeth 293, 317 Teilhardia 62 Valladolid 289, 343 Teliostoma 153 Venus Bed 188, 197, 206, 210 Tennessee 352 Vincetown Formation 352 Terebellum 209 Virgiscalpellum 374-80, 381-3 Terebrella 139 beisseli 374 tergum 354-60, 363-79; pls 1-5 darwinianum 353, 378 Tertiary 352 gabbi 352-3, 374, 377-80, 383-4 Texas 122, 352 apertus 351-2, 378, 381; pl. 5, fig. 15 Thamnopora 272, 274 gabbi 352-3, 374-8, 377*, 381; pl. 4, figs venezuelensis 273 16-18; pl. 5, figs 1-11, 14 Thanetian 184, 201, 203, 213-4 hagenowianum 353, 377-8 Thanet Sands 188, 202, 214 radiatum 377 Thanetinassa 180-1, 201, 212-3 spp. 379-80; pl. 5, figs 12-13 bicorona 184, 188, 212, 213-5; pl. 9, figs 1-7 Viverra angustidens 90-1 Thericium 153 Viverravinae 89-93 theridomyid Spp. 12, 16-17, 20, 723}, Dil Viverravus 90 Thoracica 354 angustidens 90-1, 93 Titanolepas, see Calantica euzetensis 91 Togo 184 hastingsiae 93 Tongrian 212 vomer 302 Tortonian 184 Voysa 141 Trionyx 14-18 es horis aie ee Hs 162 Waagenophyllidae 265 a eagie Washington State 180, 208-9 ae idolep ius) 23) Westeregeln 187, 191 Turonian 108, 111, 116, 131, 146-7, 150, 153, 380 Whitecliff Bay 194, 197, 199, 205-6 Turritella angustata 110, 157 Whitecliffia 180-1, 204 CE a ah suturosa 184, 204-6, 208; pl. 5, figs 6-11; pl. 6, cingulatocostata 157 figs 1-4 Bee recenne nia 1ST tumida 180, 184, 188, 204, 206-8; pl. 6, figs tusk 290 510 tympanic bulla 305 : : Tympanotonus 110, 152-3; see Potamides Whitley ese AOE a (Exechocirsus) subpustulosus 155 Wieht (sec egcu Maem . Wrotham 127 Uchauxia 108, 110, 141, 145, 146-50, 169 . badri 108, 147, 148-9, 150; pl. 6, figs 3-4 Yorkshire 112, 138, 155 distinctum 147 Ypresian 180, 184, 203 forbesiana 109 peregrinorsa 110, 146-8 Zaphrenticae 242-52 Phillipsi 149 Zaphrentis curvulena 239 richei 148 prolifica 251-2 solida 147 simplex 254 wiset 109, 147-8, 150; pl. 6, figs 5-6 spatiosa 251-2 sp. nov. 149-50; pl. 6, fig. 7 zoarial features 406 Uintatherium 341 Zonophyllum 271-2 ulna 290, 330, 335-6 zooecial chamber 405, 485 unciforms 338 features 405 os 7. | a> a) { 2 8 MAR }37 7 MARSUPIALIA, INSECTIVORA, @,,&.~ PRIMATES, CREODONTA AND _ CARNIVORA FROM THE HEADON BEDS (UPPER EOCENE) OF SOUTHERN ENGLAND Neg ft ; MARSUPIALIA, INSECTIVORA, PRIMATES, CREODONTA AND CARNIVORA FROM THE\ HEADON BEDS (UPPER EOCENE) OF SOUTHERN ENGLAND lee PETER EDGAR CRAY Bath Academy of Art, Contam Pp. 1-102; 6 Plates, 23 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vols 23) Nos 1 LONDON : 1973 ‘ f 2 8) "Aix iy THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted im 1949, 1s issued in five series corresponding to the Departments of the Museum, and an H istorical series. Parts will appear at irregular intervals as they become veady. Volumes will contain about three to four hundred pages, and will not necessarily be completed within one calendar year. In 1065 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 23, No. 1 of the Geological (Palaeontological) series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. World List abbreviation Bull. Br. Mus. nat. Hist. (Geol.). © Trustees of the British Museum (Natural History), 1973 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 2 March, 1973 Price £4.45 MARSUPIALIA, INSECTIVORA, PRIMATES, CREODONTA AND CARNIVORA FROM THE HEADON BEDS (UPPER EOCENE) OF SOUTHERN ENGLAND By P. E. CRAY CONTENTS I. INTRODUCTION AND ACKNOWLEDGEMENTS . 5 II. FossiL HORIZONS AND FAUNAL LISTS . : : j : 6 (a) Introduction : : : F : 6 (b) Horizons at Hordle Cliff, Heepshire 8 (i) Introduction 8 (ii) Lower Headon Beds . : : : » 3 10 Rodent Bed . . : . : : 5 10 Crocodile Bed : ¢ 0 : : ; 13 Mammal Bed. : : 14 Mammalian fauna of the MererenallGraeadile Bed 16 Minor horizons . : : : 2 17 (c) Horizons at Headon Hill, Isle of Wight c 0 : 18 (i) Introduction . : : : : ; ; 18 (ii) Lower Headon Beds . 5 2 : : : Bik Minor horizons : 5 : : é : 21 (iii) Upper Headon Beds . 3 0 , : 0 21 Microchoerus Bed . F : ‘ é é 21 Lignite Bed . ; ‘ : 2 : < 24 Il]. SysTEMATIC DESCRIPTIONS c . : ; c : : 27 Order MARSUPIALIA 2 : F ‘ j : j 27 Family DIDELPHIDAE Gray . : . : 5 : 27 Subfamily DIDELPHINAE Simpson : 2 ‘ : 28 Genus Pevatherium Aymard : . : : : 28 ‘Peratherium species A’. : ; : : 29 ‘Pevatherium species B’ é : : : 34 Order INSECTIVORA : : : : : 36 Family NYCTITHERITDAE Spann ; : : ; 36 Subfamily NYCTITHERIINAE Se : 6 ¢ 36 Genus Scyaeva gen. nov. . : : : : 37 Scvaeva hatherwoodensis sp. nov. - : ; 39 Scraeva woodi sp. nov. : : A é 2 41 Genus Avvaldus gen. nov. . : 2 ° : 6 44 Arvaldus stintoni sp. nov. . c : : 6 45 4 UPPER EOCENE MAMMALIA Page Family PANTOLESTIDAE Cope : : : . c 47 Subfamily PANTOLESTINAE Simpson . F : : 47 Genus Opsiclaenodon Butler 5 : : C 48 Opsiclaenodon major (iiss) : : ¢ 50 Genus Dyspterna Hopwood : ; : 2 5 57 Dyspterna hopwoodi sp. nov. : c c 7 58 Family APATEMYIDAE Matthew . : : : 61 Genus Heterohyus Gervais . : : : : ¢ 62 Heterohyus sp. . 2 : : 5 c 63 Order PRIMATES . : : : : 6 65 Family ADAPIDAE piroueesare! : : : : 0 65 Genus Adapis Cuvier . 2 ¢ : : 4 66 Adapis magnus Filhol : ; c A 67 Adapis parisiensis (Blainville) . 4 0 : 71 Family TARSIIDAE Gray : : : : 73 Subfamily MICROCHOERINAE Tedei lees : : : 73 Genus Microchoerus Wood . : c ; ° 5 74 Microchoerus evinaceus Wood. c 2 : 75 Genus Pseudoloris Stehlin . z ¢ : : < 80 Pseudoloris pavvulus Saye : ‘ 5 : 81 Order CREODONTA . : ; : : c 83 Family HYAENODONTIDAE Let : . ¢ 5 83 Subfamily HYAENODONTINAE Trouessart . : : 83 Genus Hyaenodon Laizer & Parieu c : : 84 Hyaenodon c.f. minoy Gervais . - ‘ : 85 Order CARNIVORA . : : : : : é c 89 Family MIACIDAE Cope . : : : o : 89 Subfamily VIVERRAVINAE Matthew : : . 5 89 Genus Quercygale Kretzoi_ . ; : : b : 90 Quercygale hastingsiae (Davies) . : : - QI IV. CORRELATION AND AGE OF THE HEADON BEDS . : - 5 93 V. REFERENCES . : : 5 ; 5 P é = : 96 SYNOPSIS The present work incorporates a revision of part of the mammalian fauna of the Headon Beds (Upper Eocene) of Hampshire and the Isle of Wight. Stratigraphical and historical aspects of the main mammal-yielding horizons in the Headon Beds are presented, together with provisional lists of the complete mammalian fauna. Marsupialia, Insectivora, Primates, Creodonta and Carnivora are considered in the systematic descriptions. Fourteen species of mammals are described, of which four are new. Two new insectivore genera, Scraeva and Avvaldus are referred to the Nyctitheriidae. Avvaldus is known from a single species, A. stintoni; two species of Scrvaeva are recognised, S. hathevwoodensis and S.woodi. A partial upper dentition has been referred to the pantolestid insectivore Opsiclaenodon major, previously known only from the lower dentition. Well preserved material of a new species of Dyspterna, D. hopwoodi, supports reference of the genus, formerly almost a nomen dubium, to the Pantolestidae, and confirms the close relationship of Dyspteyna and Opsiclaenodon. Hetero- hyus sp., Adapts parisiensis and Pseudoloris parvulus all constitute new records from the Headon Beds. The evidence from those mammals given detailed study, indicates that at present the Upper Headon Beds are best grouped provisionally with the Lower Headon Beds in the Upper Eocene (Lower Ludian). The Eocene/Oligocene boundary should be sited at some position above the Upper Headon Beds. HEADON BEDS OF ENGLAND 5 I. INTRODUCTION AND ACKNOWLEDGEMENTS THE first mammals recorded from the Headon Beds were listed by Wood (1844); these specimens, listed again by Charlesworth (1845), came from the Lower Headon Beds of Hordle Cliff. In a further paper, Wood (1846-47) listed and excellently illustrated this and further material from Hordle Cliff. The first record of mammals from the Upper Headon Beds of Headon Hill was in a stratigraphical account by Prestwich (1846). The earliest detailed description of mammals from the Headon Beds was by Owen (1848). In a preliminary paper in 1848, Barbara, Marchioness of Hastings recorded mammals from Hordle Cliff. She published, in 1852 and 1853, the results of six years’ collecting from this section. In addition to mammals, the Marchioness recorded fish, reptile and bird material, together with valuable information about occurrence and preservation. Also in 1852, Wright recorded an isolated mandible from the Lower Headon Beds of Headon Hill; this was described by Owen (18572). Some of the Marchioness of Hastings’ specimens were described by Owen (1857b) and Davies (1884). The only detailed systematic work was by Lydekker in his ‘Catalogue of the Fossil Mammalia in the British Museum’, published in five parts between 1885-87. The Headon Beds material was incompletely and sometimes inaccurately listed, but Lydekker made a valuable contribution, consolidating Owen’s work and describing more of the Marchioness of Hastings’ collection. Some of Lydekker’s results were published as a few short papers (1884, 1885b, 1885d). Stehlin’s monograph ‘Die Saugetiere des schweizerischen Eocaens’ (1903-16) contained critical discussions of some of the Headon Beds material. In a faunal paper, Stehlin (910) listed sixteen species from Hordle Cliff. Depéret (1917), in a revision of the mammalian fauna of Euzet-les-Bains, listed four species from Headon Hill and sixteen from Hordle Cliff. Contributions on aspects of the fauna were made by Cooper (1910, 1925, 1926, 1928). The most recent work, on isolated genera, has been by Butler (1946), Simons (1961) and Franzen (1968). The present work comprises a revision of part of the mammalian fauna of the Headon Beds. In the faunal lists, identification of Rodentia, Perissodactyla and Artiodactyla is tentative; these elements of the Headon fauna have not yet been studied in detail. Throughout the systematics, the age of the Headon Beds is considered provisionally as Upper Eocene (Lower Ludian). This problem is discussed later in the light of the evidence afforded by study of the mammals. Traditional stratigraphical nomenclature has been retained for the subdivision of the succession; as the purpose of the present work is primarily palaeontological, no attempt has been made to bring this nomenclature into concordance with modern international stratigraphical practice. The registration numbers of museum specimens are prefixed as follows: B.M. British Museum (Natural History) S.M. Sedgwick Museum, Cambridge G.S.M. Geological Survey Museum 6 UPPER EOCENE MAMMALIA D.M.S.W. D.M.S. Watson collection; now in University Museum of Zoology, Cambridge In addition, some specimens are from the private collection of Mr F. C. Stinton; these are prefixed F.C.S. Where differentiation of several specimens grouped under a single museum registration number has been necessary, numbers in parentheses have been added after the museum registration number. In the tables of measurements, all dimensions are in millimetres. Antero-posterior length (a-p) and transverse width (trs) are maximum values. Measurements in parentheses represent approximations, based on damaged teeth or taken from roots. I am very grateful to Dr R. J. G. Savage, not only for supervision of the initial research, but also for his continued encouragement, advice and stimulating discussion since that time. I also wish to thank him for so readily making available his personal library and for critically reading the draft manuscript. I am indebted also to Dr D. E. Russell for much encouragement and advice, for his generosity in permitting examination of type specimens and manuscript from unpublished work and for critically reading the draft manuscript. My thanks are due to the following for facilities to study collections or for the loan of specimens: Dr A. J. Sutcliffe and Dr R. Hamilton of the British Museum (Natural History); Dr C. L. Forbes and Dr R. B. Rickards and their predecessor Mr A. G. Brighton of the Sedgwick Museum, Cambridge; Professor J.-P. Lehman of the Muséum National d’Histoire Naturelle, Paris; Mr J. M. Edmonds of the University Museum, Oxford, and the late Mr A. T. Grapes of the Museum of Isle of Wight Geology, Sandown, Isle of Wight. Thanks are due also to Mr F. C. Stinton for the loan of specimens from his private collection. I am indebted also to Dr J. W. Murray for helpful discussion of certain aspects of the stratigraphy; to Dr J. M. Pallot for the processing and identification of pollen samples and to Dr N. Edwards for much helpful discussion and for the use of information from unpublished manuscript. My thanks are due also to Mr J. Martin for the drawing of figs 9, 10 and 12, and to Mr R. J. Godwin for the photography. The initial research, of which this paper forms part, was carried out during the tenure of a studentship from the (then) Department of Scientific and Industrial Research. This grant and the facilities offered by the late Professor W. F. Whittard in the Geology Department of the University of Bristol are gratefully acknowledged. Il. FOSSIL HORIZONS AND FAUNAL LISTS (a) Introduction A general account of the geological structure and Tertiary stratigraphy of the Hampshire Basin can be found in Chatwin (1960). ~— ~— 20 13 — S.M. C30707 ih On O77 meant ae Oo/75 el: OE T lei 2 he LT nO S.M. C53476 —- —- — — ~— — 16 135 18 13° — S.M. C30708 — — 185 105 — — B.M. 36812b —- —- —> — — 18 o9 B.M. 30348 —- — ~—~— ~—~— ~~ ~ ~ ~ 19 12 #%1:8 0 M2 M3 a-p trs a-p trs S.M. C54171 — — 18 2:0 S.M. C54170 195 20 — — 34 UPPER EOCENE MAMMALIA which may indicate a greater individual age. Hough observed (1961), that in recent didelphids, spaces tend to appear between the premolars in the mature adult. This feature of the premolars in S.M. C30707, may be similarly a variation due to individual age. ‘Peratherium species B’ (Text-fig. 12) MATERIAL. All specimens, except where otherwise stated, are from the Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff. B.M. 13250 (2). Fragment of left mandibular ramus with Mzy—My and the basal part of the coronoid. Teeth badly worn. B.M. M13251. Fragment of left mandibular ramus with M2—M, and alveoli of C—P3 and M,. Teeth badly worn. B.M. 13250 (3). Fragment of right mandibular ramus with M3—My. Teeth badly worn. S.M. C54155. Fragment of left mandibular ramus with M,—Mg and alveoli of M3. Teeth badly worn. Lower Headon Beds, Hordle Cliff. The Sedgwick Museum label gives ‘Rodent Bed’ as the locality but the preservation and adhering white sand suggests probable derivation from the Mammal/Crocodile Bed. S.M. C53474. Anterior fragment of left mandibular ramus with Mg and alveoli of the canine (base), P}-M; and Ms. Teeth moderately worn. B.M. 44142 (1). Fragment of right mandibular ramus lacking teeth, with alveoli of Py—M,. (2). Small fragment of right mandibular ramus with alveoli of uncer- tain position in the tooth row. S.M. C53480. Fragment of right mandibular ramus with posterior root of Mg and the roots of M3—Mg. S.M. C54169. Isolated trigonid fragment of left lower molar. Badly worn. Lower Headon Beds (Rodent Bed), Hordle Cliff. DESCRIPTION. The posterior mental foramen lies below M; (B.M. M13251) or Mg (S.M. C53474); the anterior mental foramen is situated below P; (B.M. M13251) or Pz (S.M. C53474). The subhorizontal symphysial surface extends to below Py. S.M. C53474 possesses a nutrient foramen (as in S.M. C30707, referred to ‘Peratherium species A’) opening through the symphysial surface. No description is given of the dentition which is structurally almost identical to previously described specimens referred to ‘“Peratherium species A’. REMARKS. Because of the much larger dimensions of the teeth (especially the length of the tooth row) compared with ‘Peratheriwm species A’, these specimens are considered to represent a distinct species. Hough stated (1961), that this criterion is less subject to variation than depth of mandible and spacing of premolars, and has been used in the determination of Marmosa species. There is some variation in mandible depth and premolar spacing; as in ‘Pevatheriwm species A’ this probably reflects individual age differences. HEADON BEDS OF ENGLAND 35 Two mandibular fragments without teeth (B.M. 44142 (2) and S.M. C53480) are tentatively referred to this species. B.M. 44142 (1) is considerably longer and more slender than the other referred specimens with widely spaced anterior premolars; it is considered tentatively as an old adult of ‘Peratherium species B’. el Ml B ily Al 7 i, 4 i tae a Si , Dw) WW Nc iN - \ IAW Pe MZ il RW Fic. 12. ‘Pevatherium species B’. Fragmentary left mandibular ramus with M2—M4. B.M. M13251 x6. Lower Headon Beds, Hordle Cliff. (a) Occlusal view. (8) Internal view. (c) External view. 36 UPPER EOCENE MAMMALIA A pathological condition is seen in S.M. C53480. The mandible evidently had been damaged during life internally below Mz and Ms, and shows some formation of callus in this area. TABLE 6 Measurements (in mm) of ‘Pevatherium species B’ My M2 Ms M4 a-p trs a-p trs a-p trs a-p trs B.M. 13250 (2) — — 2°5 75 eee 1:6 2°5 I°4 B.M. M13251 a —- 2°5 1:6 2°55 1:6 2°25 I'4 B.M. 13250 (3) 2:5 ito 2°5 1°35 S.M. C53474 — _ 2:6 I*4 S.M. C54155 2, I°4 2°7 a — Order INSECTIVORA Family NYCTITHERIIDAE Simpson, 1928 REMARKS. Simpson’s family Nyctitheriidae was abandoned by McKenna (r1960a) as an unnatural unit. Van Valen (1967) placed Nyctitheriwm together with Clinopternodus Clark, 1937, as a subfamily (Nyctitheriinae) within the Adapisoricidae. On the basis of Robinson’s (then unpublished) work on the group, McKenna (1968) recognised ‘a restricted family-group taxon based on WNyctitherium and several referred genera’, into which he placed Leptacodon. In his review of the family, Robinson (1968) upheld the status of Simpson’s family Nyctitheriidae, subdividing it into three subfamilies: Nyctitheriinae Simpson, 1928; Geolabidinae McKenna, 1960; and Micropternodontinae (Stirton & Rensberger, 1964). The Geolabidinae, placed originally by McKenna (1960b) in the Erinaceidae, had been grouped by Van Valen (1967) within the Adapisoricidae. Robinson’s view is adopted here. Subfamily NYCTITHERIINAE Simpson, 1928 DIAGNOSIS. Pi semi-molariform, Ps, P3 with two roots each, P; with one root, jaw long and slender, with no apparent shortening. Probable dental formula i, ie i, 5. (Robinson, 1968 : 129). ReMARKS. Three genera were included by Robinson (1968) in the subfamily: Nyctitheriwm Marsh, 1872. Middle—Late Eocene; North America. Leptacodon Matthew & Granger, 1921. Middle (?), Late Palaeocene—Early Eocene; North America. Saturnimia Stehlin, 1940. Late Eocene; Europe. To these are added two new genera described below. Leptacodon has undergone a varied history. McKenna (1960a : 52and 1960b : 156) removed Leptacodon (sensu stricto) from the Leptictidae; this genus, composed of L. HEADON BEDS OF ENGLAND 37 teney Matthew & Granger, 1921 (the type species) ; L. Jadae Simpson, 1935 (apparently, but not mentioned specifically by McKenna); L. minutus Jepsen, 1930, and L. jepsent McKenna, 1960a, was considered to be near the base of erinaceid phylogeny and possibly closely related to the ancestry of the Geolabidinae. This tentative reference of Leptacodon (sensu stricto) to the Geolabidinae was accepted by Russell (1964 : 46). Van Valen (1967) placed Leptacodon (sensu stricto) in the Adapisoricidae (subfamily Adapisoricinae) and more recently, McKenna (1968) has stated that the genus is perhaps best considered as a nyctithere. Robinson (1968) in his review of the Nyctitheriidae, clarified the systematics of the genus Nyctitherium ; he indicated that the type specimen of Myolestes dasypelix Matthew, 1909, is a nyctitherine, although he has agreed with McKenna’s allocation of a referred specimen ‘c.f. Myolestes dasypelix’ to the Geolabidinae (McKenna, 1960b : 147). In addition to Nyctitherium dasypelix, Robinson (1968) recognised two further species: the type species, N. velox Marsh, 1872, and N. serotinum (Marsh, 1872). Inthe same paper, he rejected Stehlin’s reference of Saturninia to the Soricidae, maintaining for the genus, a close relationship with Nyctitherium. Genus SCRAEVA gen. nov. EtyMoLtocy. Screawa—(Anglo-Saxon), a shrew. Dracnosis. In Py—-Ms, talonid narrower transversely than trigonid; talonid and trigonid subequal in length; trigonid with markedly angular outline, with straight or slightly convex internal margin and moderately to sharply pointed anterior apex; protoconid larger than metaconid; talonid angular with internal and external margins subparallel, except on Ms. M,—M3 with a mesoconid; hypoconid and entoconid subequal in size; hypoconulid median or slightly internal of midline, in Mg posteriorly projecting as a prominent rounded lobe. Pq hypoconulid minute, slightly external to midline. Internal face of coronoid process with antero-posterior ridge at molar alveoli level. TYPE SPECIES. Scraeva hatherwoodensis sp. nov. Upper Headon Beds (Mzcro- choerus Bed), Headon Hill. REMARKS. One other species has been distinguished: Scraeva woodi sp. nov. Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff. In 1844, Charlesworth (in Wood, 1844 : 350, footnote) recorded a fragmentary mandibular ramus, obtained by Flower from Hordle Cliff, which he named Spalacodon ; this was again recorded by Charlesworth (1845 : 50). A year later, the specimen was figured by Wood (1846, pl. 2, figs 5, 5a-c) and listed as ‘an imperfect lower jaw’. Charlesworth and Wood did not designate a species and the specimen was referred to as ‘Spalacodow or ‘Spalacodon sp.’ Because of the inadequate original descriptions and the loss of the specimen, Spalacodon is considered here to be a nomen nudum. Comparison of the type material of Scraeva woodi with Wood’s illustrations indicate the probable specific identity of the two forms. This contention is strengthened by the probability that Spalacodon and the material of Scraeva woodi were obtained from the same level, the Crocodile Bed of Hordle Cliff. The specimen described by 38 UPPER EOCENE MAMMALIA Charlesworth and Wood (from Wood’s figures, a right mandibular ramus with Pz and P4s—Ms) has not been located during the present study and the material is considered to be lost. Scraeva is referred to the Nyctitheriidae mainly on the basis of strong resemblances to Leptacodon and Nyctitherium. The soricid-like antero-posterior internal ridge on the coronoid, characteristic of Nyctitherium (Robinson, 1968), is well developed in Scraeva. A talpid characteristic, occasionally developed in Nyctitherium and present also in Saturninia, is a continuation of the crista obliqua up the posterior wall of the metaconid; this is prominently and invariably developed in Scvaeva. The Headon genus is distinguished from other nyctitheres as follows: In Nyctitherium (see Robinson, 1968), Pg has a smaller talonid basin and lacks an entoconid; in the molars, the hypoconulid is more internal in position and smaller and the paraconid is more internally situated. In this study the type specimen of Saturninia has not been examined, but from Stehlin’s illustrations (1940, fig. 1c, fig. 2c) the Pg is relatively smaller with a less basined talonid, and the molar paraconids are more internal in position. In addition, in both Nyctitherium and Saturninia, the outline of P4—Mz is more rounded and the P,—Ms trigonid and talonid are of similar transverse width. The Py, of Scraeva is very similar to that of the type species of Leptacodon, L. tener, except for a rounded posterior talonid margin and a sharper, more separated para- conid in Leptacodon tener. The molars are also similar although in Leptacodon tener, the metaconid is as high as the protoconid; the paraconid is somewhat sharper and more separated; the Mg hypoconulid is less projecting posteriorly and the trigonid cusps are not so elevated above the talonid. Less resemblance is apparent in the P, of Leptacodon ladae which has a more isolated median paraconid, a more anteriorly placed protoconid and a much shorter talonid. Scraeva also shows marked resemblances to Scenopagus McKenna & Simpson, 1959, Messelina Tobien, 1962, and Adapisorex Lemoine, 1883; these genera have been generally regarded as ‘primitive probable erinaceids’, although extensive regrouping and familial separation from the Erinaceidae has been proposed by Van Valen (1967 : 261). Inseveralrespects, particularly the elevated trigonids and the structure of Mz, Scraeva is similar to Scenopagus but there are significant differences: in Scenopagus, the talonid cusps of the molars are slightly higher; the trigonid cusps are compressed antero-posteriorly and the paraconid crest is larger and more oblique, resulting in a blunter anterior tooth apex; the hypoconid crest runs to the internal base of the protoconid, not the metaconid as in Scraeva, and the hypoconulid is only slightly developed in Mj-Mz. Py has not been adequately described or figured. In Messelina, the lower molars are much narrower transversely ; the talonid and trigonid are subequal in width and the hypoconulid is internal; the trigonid of the molars is not elevated, and in M,—Mg the trigonid and talonid cusps are subequal in height. Also, the Py of Messelina has a rounded outline and lacks a hypoconulid. Greater resemblance to Scraeva is seen in Adapisorex: in the molars of both genera, the protoconid is larger than the metaconid (but not so extremely developed in Scraeva) and the basined talonid has a median hypoconulid. The P4 is similar in Scraeva and Adapisorex, with a small metaconid, a large protoconid anterior to the metaconid HEADON BEDS OF ENGLAND 39 which dominates the trigonid (less extremely so in Scrvaeva) and a small median hypoconulid. However, Adapisorex is distinguished from Scraeva by having much lower-placed cusps, the trigonid being little elevated above the talonid; the entoconid and hypoconid are much larger and the trigonid and talonid are subequal in width. In basic structure, particularly of Ps-Mz, Adunator Russell, 1964 (originally considered by Russell as a leptictid but placed by Van Valen, 1967, in the Adapisori- cinae) shows similarities to Scraeva, although the Py paraconid is much more separated anteriorly as a distinct cusp. Also, in Adwnator the Mz hypoconulid is more internal and not so projecting as an isolated lobe, and in the molars the metaconid and entoconid are higher than the protoconid and hypoconid respectively. Scraeva hatherwoodensis sp. nov. (Text-fig. 13) EtymMoLocy. hatherwoodensis—from Hatherwood Point, off Headon Hill, Isle of Wight, the sea cliff locality. Dracnosis. The smaller species. P4-Me trigonid with a slight posterior tilt. P4-M2 with straight internal trigonid margin and protoconid and metaconid more internally placed than in S. woodi. P4 metaconid and protoconid slightly lower than corresponding cusps of M;; paraconid median, at same height as molar paraconids. P4 metaconid more posteriorly placed in relation to protoconid than in S. woodz, in M, metaconid only slightly posterior to protoconid. Hototyre. B.M. M26051. Fragment of left mandibular ramus with P,s-M2 and alveoli of Iz-P3 and M3. Lower external surface embedded in matrix. Teeth slightly worn. Upper Headon Beds (Microchoerus Bed), Headon Hill. The only specimen of the species. Description. The long slender mandible has a large oval posterior mental foramen just below the alveolar level of Ps. There is a small angular process (tip damaged on the specimen) at molar alveoli level. The coronoid and condyle are not preserved. The mandibular foramen is visible just below alveolar level. On the internal face of the coronoid, a prominent antero-posterior ridge runs from the level of the M3 alveoli towards the mandibular foramen. A similar ridge was observed by Simpson (7m McGrew, 1959) on a toothless mandibular fragment from the Middle Eocene of Wyoming, tentatively referred to the Soricidae. This structure also occurs in Nyctitherium (Robinson, 1968 : 130). M3 is double-rooted and known only by alveoli. In Mj—Mg the talonid and trigonid are subequal in length, but the talonid is narrower transversely. The internal and external talonid margins are straight and subparallel. The trigonid is angular with a sharp anterior apex and a straight internal margin; sharply elevated, it falls almost vertically to a low basined talonid. There is a prominent antero-external cingulum. The protoconid is larger than the metaconid and slightly more anteriorly placed (on both My and Mg the metaconid is damaged). The paraconid is moderately well 40 UPPER EOCENE MAMMALIA separated and cuspate, forming the sharp anterior extremity of the trigonid; this cusp is median in M; but slightly internal to this position in Mz. The hypoconid and entoconid are subequal; the hypoconulid is a little smaller but comparison is difficult as the entoconid is almost unworn, whereas the hypoconid and hypoconulid are badly worn. The median hypoconulid (in Mg slightly internal to the midline) is posterior to the hypoconid and entoconid, projecting very slightly posteriorly and joined to the hypoconid and entoconid by smalllow crests. From the entoconid a low feeble crest runs anteriorly to the metaconid base. A stronger crista obliqua runs from the hypoconid up the posterior wall of the metaconid; this crest swells into a prominent mesoconid (stronger on M; than Mg) midway along its length. P, is very similar to M, and Mg but differs in several respects: the tooth is more laterally compressed and smaller in all dimensions; the hypoconulid is very small, situated external to the midline near the hypoconid; the metaconid is more posterior in relation to the protoconid; the hypoconid is larger than the entoconid and the protoconid/paraconid crest is smaller, placed slightly external to the midline and less V-shaped than in the molars. P3 and Pz are represented only by alveoli; these teeth were apparently double-rooted and closely spaced. Py is single-rooted, represented only by the alveolus. Known only from the alveolus, the canine was evidently slightly anteriorly inclined, with a root diameter as that of P}. Is is represented by the possible base of the alveolus. REMARKS. This species is clearly related to Scraeva woodi in that both forms exhibit the characteristic squared angular outlines of the teeth. There are a number of small but distinct differences between S. hatherwoodensis and S. woodi (see diagnoses) which indicate at least a specific separation. Also, S. woodi is markedly larger. Fic. 13. Scvaeva hatherwoodensis gen. et sp. nov. Fragmentary left mandibular ramus with Py-Mz. Holotype (B.M. M26051) x10. Upper Headon Beds, Headon Hill. (a) Occlusal view. (8) Internal view. HEADON BEDS OF ENGLAND 41 TABLE 7 Measurements (in mm) of Scyvaeva hatherwoodensis Py My Me a-p trs a-p trs a-p trs Holotype 1°35 o-9 I°5 Ir I'5 I'2 (B.M. M2605r) Scraeva woodi sp. nov. (Text-fig. 14) ETYMOLOGY. woodi—after Searles Wood who first listed and figured insectivore material from the Headon Beds as Spalacodon, a name suggested by Charlesworth, but here considered as a nomen nudem. Dracnosis. Larger than S. hatherwoodensis. Py, with slightly concave internal trigonid margin, in molars slightly convex. P4metaconid and protoconid same height as corresponding cusps of M;; paraconid markedly internal, lower than molar para- conids. P, metaconid slightly posteriorly placed in relation to protoconid, in My, metaconid opposite and in Mg—M3 metaconid slightly anterior to protoconid. HototypPe. S.M. C53475. Fragment of left mandibular ramus with Py-M3 and alveoli of Ps. Teeth unworn. Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff. PARATYPE. B.M. M25112. Left mandibular ramus with P2, Py—-M3 and alveoli of I2—-P;, P3. Badly worn dentition. Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff. MaTERIAL. S.M. C30709. Fragment of left mandibular ramus with Mj—Mp, posterior alveolus of P4, and alveoli of Ms. Teeth unworn but metaconids damaged. Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff. B.M. 36801c. Fragment of left mandibular ramus with Mg and alveoli of Mg Internal margin of tooth damaged and tips of cusps broken off. Patches of coarse white sand matrix adhering. Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff. S.M. C54168. Isolated right Mz. Slightly worn. Lower Headon Beds (Rodent Bed), Hordle Cliff. DESCRIPTION. The mandible is a little stouter than in S. hatherwoodensis, with a large oval posterior mental foramen below the posterior toot of P3. The antero- posterior ridge on the internal face of the coronoid is present as in S. hatherwoodensis. Paratype B.M. Mz25112 shows the wedge-shaped symphysial surface with ridged margins; this occupies nearly the whole depth of the jaw under Is, sloping baso- posteriorly to below the anterior root of Pe. The incisors, canine and P; are known only from alveoli in paratype B.M. M25112. In this specimen, two distinct, well-separated alveoli both slightly smaller than the P; alveolus, indicate at least two incisors, here considered as Ip-I3. Another more 42 UPPER EOCENE MAMMALIA anterior tooth may have been present but evidence is lacking. The canine is re- presented by an alveolus indicating a larger anteriorly-inclined tooth. Pj, single- rooted and well-separated, is represented by the alveolus which is smaller than that of the canine. Pg: is double-rooted and ovalin outline. A single large anterior median cusp slopes antero-basally to a small low antero-internal cuspule; a small heel bears a minute internal cuspule. Pg is known only from the alveoli; the tooth is double- rooted and apparently about the same size as Pp». Fic. 14. Scyvaeva woodi gen. et sp.nov. Fragmentary left mandibular ramus with P4—M3. Holotype (S.M. C53475) x10. Lower Headon Beds, Hordle Cliff. (a) Occlusal view. (B) Internal view. (c) External view. HEADON BEDS OF ENGLAND 43 The trigonid and talonid in P, are subequal in length. In M,—Ms the talonid is slightly more elongated antero-posteriorly and narrower transversely; the internal and external margins are fairly straight and subparallel, except in M3 where the talonid narrows posteriorly (not as extreme as in S. hatherwoodensis). The Mg in paratype B.M. Mz5112 has a somewhat indented external margin. P, and Mg are narrower transversely than M; and Mz. In P4—Ms the trigonid is angular with only a slightly curved internal margin; sharply elevated, it drops steeply to a low basined talonid. There is a prominent antero-external cingulum as in S. hatherwoodensis. The metaconid is smaller than the protoconid, situated posterior to the protoconid in Py, opposite in Mj and slightly anterior in Mz and M3. The paraconid, moderately well separated and cuspate, forms the fairly sharp antero-internal extremity of the tooth—this cusp is moderately internal to the midline in all teeth (P4-M3). The hypoconulid in My; and Mgis a little smaller than the entoconid and median in position ; in Py, it is very small and slightly external to the midline (but in Ms, hypoconulid, entoconid and hypoconid are subequal). The P, hypoconulid in paratype B.M. M25112is indistinct, indicated only bya slight externally situated projection posterior to the hypoconid and entoconid. Ms has a posteriorly projecting hypoconulid lobe, slightly internal to the midline; this lobe is very prominent and apically projecting in paratype B.M. M2511z. In Pg—Mg the hypoconulid is joined to the hypoconid and entoconid by small low crests; a low feeble crest runs antero-internally from the entoconid to the metaconid base. A much stronger crista obliqua runs from the hypoconid up the posterior wall of the metaconid ; this crest bears a distinct mesoconid in M\—Msz, although the cusp is only feebly developed in Mg. REMARKS. Comparison with S. hatherwoodensis has been effected in the diagnosis and description. Only five specimens are known and little idea can be gained of variation within the species. S.M.C30709 and S.M. C54168 differ from the holotype only by having a less distinct mesoconid in the molars. Paratype B.M. M25112 represents a much older individual than the holotype and comparison of the two specimens is difficult because of a high degree of wear on the former specimen. Reference of B.M. Mz25112 to the species is fairly certain, however; when structural differences which can be assigned to excessive wear ate discounted, the two specimens have a high degree of similarity. Also, both specimens are apparently from the same locality and approximate horizon. B.M. 36801c appears closely similar to the holotype, although the former specimen is damaged. TABLE 8 Measurements (in mm) of Scvaeva woodi Pe, P3 12h My, Me Ms a-D tis =ph tts! a—p (Ets) d-p) tts) a-p) irs! fa-p)) irs B.M. M25112 Ti O;0 M1 nT 7 OF NT 75 AS eat Om seICA5 a T8— 1-2 B.M. 36801c SS eS SS SS i a ee YSZ ee S.M. C54168 — — —«-—-S—s T8255 —— — S.M. C53475 Ti), TRO) FUO|G} aS} ESI} ga, acts) ago S.M. C30709 a aS TNS) 1-7) 125) = 44 UPPER EOCENE MAMMALIA Genus ARVALDUS gen. nov. EtymoLocy. Arvaldus—Jutian king of the Isle of Wight, A.D. 686. DraGnosis. P, protoconid markedly anterior to the metaconid; hypoconulid (damaged) appears more prominent than in Scvaeva. P4-Ms posterior margin some- what oblique to length of tooth row; paraconid internal; protoconid larger than metaconid; P, hypoconid opposite, M;—Mg slightly posterior, and in M3 markedly posterior to entoconid. M3 hypoconulid prominent, internal to midline, slightly projecting; talonid angular in outline with straight posterior and postero-internal margins. TyPE SPECIES. Avrvaldus stintoni sp. nov. Upper Headon Beds (Lignite Bed), Headon Hill. The only known species. REMARKS. Avrvaldus shows considerable similarity with Leptacodon, Nyctitherium and Saturninia. The Py of Leptacodon tener (the type species) seems quite close to that of Avvaldus with a well differentiated talonid and the protoconid strongly anterior to the metaconid, although in Py, of L. tener, a hypoconulid is absent and the posterior talonid margin has a rounded outline. Similarity with the Py of Leptacodon ladae is shown by the anterior position of the protoconid and the presence of a small hypoconulid, although in several other respects—the median paraconid, rounded outline and small talonid, the tooth is rather different. The type species of Nyctitherium, N. velox, is distinguished from Arvaldus by the Py talonid being much shorter antero-posteriorly with the protoconid opposite the metaconid ; in the molars, the paraconid is more internal and the hypoconulid is weaker and placed more inter- nally, although in this latter respect, Avvaldus is closer to N. serotinum where the molar hypoconulid is slightly more median in position. In Saturninia the Py, protoconid is opposite the metaconid and the paraconid of the molars is more internally situated. One of the most obvious features which distinguishes Lepiacodon, Nyctitheriwm, Saturninia and Scraeva from Arvaldus is the structure of M3. In Leptacodon there is a strong entoconid opposite the hypoconid and a well separated posteriorly projecting hypoconulid. This tooth in Nyctitheriwm velox is different in outline; the talonid is shorter with a median hypoconulid and an entoconid which is opposite the hypoconid. The Mg talonid in Saturninia is shorter and different in outline with a median posteriorly projecting hypoconulid. In M3 of Scraeva the entoconid and hypoconid are opposite and there is a median posteriorly projecting hypoconulid. In this M3 talonid structure, Avvaldus shows similarities with the American geolabidines, Geolabis, ‘c.f. Myolestes dasypelix’ and Centetodon, particularly the latter. In Centetodon, the Mg talonid bears only a hypoconid and an internal crestiiform hypoconulid (fused entoconid and hypoconulid?) ; the internal wall of the postero-internally projecting hypoconulid curves abruptly baso-posteriorly and there is no sign of an entoconid. A somewhat similar distinctive Ms; talonid structure is seen in Arvaldus, although here it is not so pronounced; the hypoconulid is less posteriorly situated and a small anteriorly placed entoconid is still present. However, in Centetodon and other geolabidines, the structure of Py is different: in the American forms, the Py protoconid and metaconid are opposite and the talonid is much reduced, HEADON BEDS OF ENGLAND 45 with a single large internal cusp. Greater similarity to Avvaldus is seen in the P4 of ‘c.f. Hypacodon sp.’ (= Centetodon sp.) described by McKenna. (1960b : 150). Here the elongation of the tooth is similar to that in Avvaldus and the protoconid is anterior to the metaconid; the talonid is more developed and there is a central hypoconid and a small though distinct entoconid. Robinson (1968 : 134) has considered this specimen as an adapisoricid. Arvaldus differs from Scraeva in a number of important aspects (see diagnoses), the more obvious of these being the non-angular tooth outlines, the structure of P4 and the distinctive Mg talonid; the two genera do not appear to be closely related. To some extent, the structure of Avvaldus is intermediate between nyctitherines and geolabidines and on the basis of the Mg structure, the genus could be tegarded as a primitive relatively unspecialised geolabidine. However, the molarised P4 and the continuation of the crista obliqua up the posterior wall of the metaconid support reference to the Nyctitheriinae. The anterior lower dentition and the upper teeth of Avvaldus are unknown. At present the European record of the family comprises Saturninia, Scraeva and Arvaldus, all from the Upper Eocene. It is likely that the phyletic position of Avvaldus will be clarified with the discovery of nyctitheriids from the Lower Eocene of Europe. Arvaldus stintoni gen. et sp. nov. (Text-fig. 15) EtyMoLocy. stintoni—after Mr F. C. Stinton, the finder of the specimen. Dracnosis. No other species are known and the diagnosis is the same as that for the genus. Hototyre. F.C.S.1. Fragment of right mandibular ramus with Pe, P4sa-Ms3 and the alveoli of Ps. Upper Headon Beds (Lignite Bed), Headon Hill. The only known specimen of the species. DEscrRIPpTION. The mandible has a large posterior mental foramen below P3. The posterior region is not preserved. Py, is very similar to that of the paratype of Scraeva woodt (B.M. M25112). The tooth is double-rooted and oval in outline with a large median anterior cusp, which slopes baso-anteriorly to a small low antero- external cuspule. A small low heel bears a single minute internal cuspule. Repre- sented by only alveoli, P3 is double-rooted and apparently a little smaller than Py. In P4—Ms, the talonid is a little longer antero-posteriorly than the trigonid especially in Mg and about the same width, with an indented internal and external margin. The trigonid has a curved internal margin. The posterior wall drops less steeply than in Scvaeva. There is a prominent antero-external cingulum. The protoconid is larger than the metaconid and situated opposite to it in M;—M3; in P, the protoconid is markedly anterior to the metaconid. The cuspate paraconid is moderately well separated and situated internally in all four teeth as in Scraeva woodi. The Py paraconid is sharper and more anteriorly projecting than in this latter species. In P4—M3 the hypoconid is stouter than the entoconid; the hypoconulid is a little smaller than the entoconid or subequal, except on Mg where it is a little larger; this cusp is 40 UPPER EOCENE MAMMALIA slightly external to the midline in Py, median in M, and slightly internal in Mz and Msg; it is posterior to the entoconid and hypoconid and projecting slightly in all four teeth. The hypoconulid is well separated from the entoconid but is joined to the hypoconid by a small low external crest. In P4, the hypoconid is opposite to the entoconid, in My and Mg slightly posterior and in Mg markedly posterior. From the entoconid, a feeble crest runs antero-internally to the base of the metaconid. A stronger crista oblique runs antero-internally from the hypoconid up the posterior wall of the metaconid; this swells up midway along its length into a prominent mesoconid. The enclosed talonid is deeply basined. The posterior margin of the talonid is slightly obliqua to the length of the tooth in P4—Mg (a tendency also seen in paratype B.M. M25112 of Scraeva woodt). In Msg the straight posterior margin runs from the sharp postero-external extremity of the talonid to the slightly more posteriorly situated hypoconulid. From this latter cusp, the straight talonid margin runs obliquely forwards to the relatively anteriorly-placed entoconid. Fic. 15. Avvaldus stintoni gen. etsp.noy. Fragmentary right mandibular ramus with P2, Pa-M3. Holotype (F.C.S. 1) x10. Upper Headon Beds, Headon Hill. (a) Occlusal view. (B) External view. (c) Internal view. HEADON BEDS OF ENGLAND 47 TABLE 9 Measurements (in mm) of Avvaldus stintoni Pe Ps P4 M, Me M3 a DECLSE a PRRELS sae biS aD cLSee cd —-Dm cise d—Pees LES Holotype 1200756 ee O-Guunl-O) Li T-O5met-2 0) To5) | 10:0 (F.C.S. 1) Family PANTOLESTIDAE Cope, 1884 Diacnosis. Facial part of skull short, basicranial region broad, middle portion of skull elongate. Angle of mandible a short stout process expanded at the tip. Mental foramen usually beneath My, sometimes below Py. Dental formula 3, f, 3, + Molars and premolars with low massive cusps. Canines large, massive, lower canines often semiprocumbent. Upper incisors spaced and _ peg-like. P ,—-P3 elongate and fairly small. P4 relatively enlarged (after Matthew, 1909). REMARKS. On the basis mainly of differences in the structure of Pi, two sub- families have been traditionally distinguished : Pantolestinae Simpson, 1937 Pentacodontinae Simpson, 1937 Gazin (1959) erected a new subfamily, Apheliscinae; this was referred by McKenna (1960a) to the condylarths (Hyopsodontidae). Van Valen (1966) supported Gazin’s reference of the Aphelescinae to the Pantolestidae, but more recently (1967) has agreed with McKenna’s removal of the group to the Hyopsodontidae. Van Valen (1967) in his revision of insectivore classification raised the pentacodonts to family level. As the justification for such a procedure (Van Valen, 1967 : 321) seems no greater than that for retaining them in the Pantolestidae, the old-established two fold subdivision of the Pantolestidae is retained here. Subfamily PANTOLESTINAE Simpson, 1937 Dracnosis. Pj relatively little enlarged. P, compressed, no metaconid, semi- shearing, heel essentially unicuspid and little or not basined. P4 with compressed amphicone, no metacone, low, sharp styles, small protocone without cingula (Simpson, 1937). Remarks. The Pentacodontinae are distinguishable from this subfamily by Pi being much enlarged; Py has a well developed metaconid and basined talonid; P4 has a distinct metacone, large protocone, with expanded cingula and styles are small or absent. 48 UPPER EOCENE MAMMALIA The following genera have been referred, to date, to the Pantolestinae: Propalaeosinopa Simpson, 1927. Middle-Late Palaeocene; North America. Bessoecetor Simpson, 1936. Middle-Late Palaeocene; North America. ?Late Palaeocene; Europe. Pagonomus Russell, 1964. Middlé-Late Palaeocene, ?Early Eocene; Europe. Pantomimus Van Valen, 1967. Middle Palaeocene; North America. Palaeosinopa Matthew, 1901. Late Palaeocene-Early Eocene; North America. Pantolestes Cope, 1872. Middle Eocene; North America. Cryptopithecus Schlosser, 1890. Late Eocene; Germany. Opsiclaenodon Butler, 1946. Late Eocene; Britain. Dyspterna Hopwood, 1927. Early Oligocene; Britain. Androconus Quinet, 1965. Early Oligocene; Europe. Chadronia Cook, 1954. Early Oligocene; North America. Galethylax Gervais, 1848-52. Late Eocene; Europe. Buxolestes Jaeger, 1970. Middle Eocene; Europe. The above generic list follows that of Van Valen (1967) but with the tentative retention of Opsiclaenodon and Bessoecetory as distinct genera, and the addition of Buxolestes. Simpson (1945) listed Propalaeosinopa, Bessoecetor, Palaeosinopa, and Pantolestes. Van Valen (1967) added a new genus Pantomimus, Pagonomus, Cryptopithecus, Dyspterna, Androconus, Chadronia and Galethylax. ‘This latter problematical genus (discussion in Van Valen, 1966 : 73; 1967 : 227) was only provisionally referred to the family. Van Valen originally listed a new genus Pantinomia from the Middle Palaeocene of North America, as a pantolestid (1967 : 259) but in a later view (1967 : 225, footnote) referred it to the arctocyonids. In the same paper, Van Valen (1967) considered Bessoecetor as a synonym (tentative) of Propalaeosinopa and Opsiclaenodon as a synonym of Cryptopithecus, although Russell (in Russell, Louis & Poirier, 1966) retained Bessoecetor as a valid genus, into which he placed tentatively a new species, B. lever. More recently, Dr D. E. Russell (personal communication, 1971) has reaffirmed his belief in the validity of Bessoecetor. Russell (1964) in his type description of Pagonomus did not assign the genus to a subfamily; he noted resemblances of most of the teeth to the pentacodontine, Aphronorus. Additional confirmation of this view was given by a molarised P4 tentatively referred to the genus by Russell (im Russell, Louis & Poirier, 1966). Van Valen (1967) considered the type of Pagonomus (an isolated M2) as a pantolestine. Genus OPSICLAENODON Butler, 1946 Diacnosis. Lower molars with moderately elevated cusps, molars uncompressed antero-posteriorly. My, and Mg very similar in size and structure, Mg much smaller but quite similar structurally. Entoconid moderate-sized and distinct on all molars, hypoconid slightly larger. Hypoconulid of molars small but distinct, not extended posteriorly as a lobe on Mg. Upper molars moderately elongated transversely with HEADON BEDS OF ENGLAND 49 rounded outlines; M! slightly smaller than M?, M? markedly smaller; prominent hypocone in all molars; conules developed in all molars, situated close to metacone and paracone, the metaconule being larger than the paraconule. In M2, conules well developed, but more rudimentary in M! and M’. M1! and M? parastylar region extended antero-externally, and metastylar region extended postero-externally, as prominent rounded lobes. No continuous cingulum round base of protocone. Type SpEcIES. Opsiclaenodon major (Lydekker, 1887). Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff. The only known species. RemarKs. Butler redescribed (1946) Newrogymnurus major Lydekker known only from fragmentary lower dentitions, as a new genus Opsiclaenodon. In his original description, Butler referred (1946 : 691) Opsiclaenodon to the Arctocyonidae, suggesting a relationship with Dyspterna and Didymoconus. McKenna (im Russell and McKenna, 1961 : 281, footnote) suggested that Opsiclaenodon was an oxyclaenine arctocyonid, with affinities to Metachriacus, Mimotricentes and Spanoxyodon. Russell (1964 : 193) observed that reference of Opsiclaenodon to the Oxyclaeninae was unlikely. Van Valen (1966) referred Cryptopithecus and Dyspterna to the Pantolestinae. He later added (1967) Opsiclaenodon as a synonym of Cryptopithecus. Restudy of Butler’s material and examination of additional specimens, has confirmed Van Valen’s view that Opsiclaenodon is a pantolestine. The type of Cryptopithecus sideroolithicus is a fragmentary left mandibular ramus with Mz2—M3 (Bayerische Staatssammlung fiir Palaontologie, Munich, No. A.S. (128) XI 1). Comparison with a cast of this specimen (kindly made available by Dr D. E. Russell) reveals reasonable structural differences (see below) and in the present work, the validity ot Butler’s genus is accepted provisionally, pending discovery of better more complete material of Cryptopithecus. Isolated M1 and M2 (described below) referred to Opsiclaenodon show resemblances to the oxyclaenine arctocyonid, Oxyclaenus. However, in this genus, the internal margin of the upper molars is more angular and the hypocone is rudimentary; also the lower molars show important differences. The same specimens also show similarity within the arctocyonids to the triisodontine, Goniacodon, except for the poor hypocone development in the latter; M; and Mg in these two genera also show similarities. However, these resemblances do not seem significant as Opsiclaenodon shows none of the peculiar adaptive features characteristic of the Triisodontinae. Structurally, the referred upper molars of Opsiclaenodon are strikingly similar to some hyopsodontine condylarths such as Haplaletes and Litomylus. In the latter genus, however, the conules are stronger; tooth outlines are more angular; M3 is triangular-shaped and lacks a hypocone. Also, P4 has a metaconid and a small low paraconid. These characteristics (and those of the lower dentition such as P4 witha metaconid and a small low paraconid) serve to exclude Opsiclaenodon ftom any close relationship with the Hyopsodontidae. Amongst the described members of the Pantolestinae, Propalaeosinopa and Bessoecetor are easily distinguished from Opsiclaenodon by having slender elevated trigonid cusps; Mg is much larger and bulbous and Py is very elongated with a strong anterior basal cusp and incipient basining of the talonid ; M!—M? are very slender and D 50 UPPER EOCENE MAMMALIA transversely elongated with a smaller hypocone and the outline of Mg is quite different. In the European species provisionally referred to Bessoecetor, B. levei, the M2 and M8 are more slender and transversely elongated, with angular parastylar and metastylat areas; the lower molars are more angular with more elevated trigonids and the Mg is longer antero-posteriorly than the other molars, with a posteriorly projecting hypoconulid. The upper molars of Pantomimus are more elongated transversely with prominent externally projecting metastylar and parastylar areas; a hypocone is absent. The M? (type) of Pagomomus is more elongated transversely with more elevated slender cusps and more angular metastylar and parastylar areas; the hypoconal area is more angular and posteriorly extended. In Pantolestes the upper molars are similar in shape to Opsiclaenodon, but the hypocone is very small and the metastylar and parastylar areas are rounded and less prominent; the lower canine is massive but lacks striations and a groove; the molar paraconid is vestigal and the hypoconulid is strong and posteriorly projecting, in M3 forming an enlarged posterior lobe. In Cryptopithecus, the anterior margin of the molars is more rounded and blunt, with the M,; paraconid smaller, less elevated and less separated from the protoconid; the trigonid is compressed antero-posteriorly with crowded elevated cusps; the M; metaconid is more massive and the talonid cusps are better developed than in Opsiclaenodon with a very strong hypoconid. The Ms in Palaeosinopa is as large as M2; the molar trigonids are slender and more elevated, especially Mg and the antero-external cingulum is smaller. The upper molars are more elongated trans- versely with a small hypocone and the posterior mental foramen lies under Mb. Buxolestes has a relatively simple canine section; P4 has a more elongated talonid ; Mg is more rectangular and antero-posteriorly compressed and M3 is enlarged. In the upper molars of Buxolestes, M? is more elongated transversely with a less rounded outline and M3 is triangular-shaped and only slightly reduced with a less developed parastylar area. Apart from the closely related Dyspterna (see later discussion), Opsiclaenodon appears nearest structurally to the contemporaneous Cryptopithecus. The earlier European history of the family is poorly known and none of the described European species from the Palaeocene and Early-Middle Eocene appear to be closely related to Opsiclaenodon and Cryptopithecus. The Late Palaeocene and Early Eocene Palaeosinopa, however, seems to exhibit the basic dental structure from which Opsiclaenodon and Cryptopithecus could be derived. Opsiclaenodon major (Lydekker, 1887) (Text-figs 16-19) 1887 Neurogymnurus major Lydekker: 302 1967 Cryptopithecus major (Lydekker); Van Valen: 87 Dracnosis. As for the genus. Lectotype. B.M. 29718a. Fragment of right mandibular ramus with Py-M3 and alveoli of the canine and P2-Ps. Talonids badly worn. Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff. Listed Lydekker, 1887 : 302. Figured Butler, 1946, fig. 1; fig. 2, (2), (9). HEADON BEDS OF ENGLAND 51 MateriaL. All material is from the Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff, except where otherwise stated. B.M. 29718. Posterior fragment of left mandibular ramus with M,-M3. The coronoid, condyloid and angular processes preserved, though damaged. External surface embedded in matrix. Wear greater than lectotype. Listed Lydekker 1887 : 302. Figured Butler, 1946, fig. I. B.M. 36801. Fragment ofright mandibular ramus with Mz—-M3. Listed Lydekker, 1887 : 303. Listed Butler, 1946 : 601. B.M. 36802a. Fragment of right mandibular ramus with alveoli of M,—Ms3. Listed Butler, 1946 : 691. B.M. 36803. Left mandibular ramus with Ps, Ms the base of the canine and the alveoli of Py; and P3-Mz. Teeth unworn. Listed Lydekker, 1887 : 303. Figured Butler, 1946, fig. 1; fig. 2, (3), (4), (6), (7), (8), (Zo). B.M. 36807. A crushed fragment of left mandibular ramus with Mz—-M3 and alveoli of Pg-My. External surface embedded in brown clayey sand matrix. Teeth unworn. Listed Lydekker, 1887 : 303. Figured Butler, 1946, fig. 2, (1), (5). S.M. C53478. Isolated right My. Unworn. Damaged metaconid. B.M. M12565b. Isolated right Mj. Unworn. S.M. C30611. Fragment of left mandibular ramus with P3—P, and alveoli of Mj, bases of alveoli of Pp and the canine. Unworn. S.M. C54162. Isolated teeth. All from Lower Headon Beds (Rodent Bed), Hordle Cliff. a. Right Pz. Unworn. b. Right Mg. Part worn. c. Right Mg. Part worn. d. Right Mg. Part worn. e. Left M;. Part worn. B.M. 29863. Anterior fragment of left mandibular ramus with the canine, Py-P,, and alveolus of P;. Tip of canine broken off. Symphysis preserved. Matrix of brown clayey sand. B.M. M25094. Anterior fragment of right mandibular ramus with P3 and alveoli of the canine, Pj, Pe, Paand M;. Posterior part of symphysis preserved. B.M. 25084. Isolated right lower canine. Tip damaged. B.M. M25oq1. Isolated right Mg. Moderately worn. S.M. C30713. Isolated right upper molar (M2?)._ Unworn. S.M. C30722. Isolated left upper molar (M??). Badly worn. S.M. C9677. Isolated left upper molar (M3?). Unworn. S.M. C54163. Isolated right upper molar (M1?). Moderately worn. Lower Headon Beds (Rodent Bed), Hordle Cliff. Description. The mandible is moderately elongated and stout with a slight anterior deepening. The symphysial surface (B.M. Mz5094, B.M. 36803 and B.M. 29863) is elongated and oval in outline, ending posteriorly below Pa. The posterior region of the mandible is preserved only in B.M. 29718: the condyle, situated above the level of the tooth row is damaged, but appears to have been 52 UPPER EOCENE MAMMALIA transversely elongated. The angular and coronoid processes are separated from the condyle by fairly prominent notches. The masseteric fossa extends only slightly below the level of the tooth row; the mandibular foramen is situated similarly. The anterior mental foramen occurs below P,/canine (B.M. 36803 and B.M. 29718a) or below P; (B.M. M25094 and B.M. 29863). The posterior mental foramen is also variably sited below Py (B.M. 36803 and B.M. 29718a) or M, (B.M. 36802a, B.M. 25094 and S.M. C30611). An additional small foramen occurs below P3 (B.M. 29863 and B.M. 36803) or below Pe (B.M. Mz25094). Fic. 16. Opsiclaenodon major (Lydekker, 1887). Fragmentary right mandibular ramus with P4-M3. Lectotype (B.M. 29718a) approx. X35. Lower Headon Beds, Hordle Cliff. (a) Occlusal view. (B) External view. (c) Internal view. HEADON BEDS OF ENGLAND 53 The lower incisors ate not known. The lower canine is known from B.M. 36803, B.M. 29863 and B.M. M25084. The upper half of the tooth is enamel covered; this enamel is wrinkled externally, forming regular striations. The upper half of the tooth is triangular in section with sharp postero-external, internal and anteto- internal crests; the latter two crests bound a prominent groove. In B.M. 36803 and B.M. 29863 the groove, crests and enamel covering occur below alveolar level, but in these specimens, the tooth is incompletely erupted. These features are considered to occupy the upper half of the tooth, by analogy with material of Dyspterna (described later) where an almost identical tooth is fully developed and completely preserved. Butler, in his type description (1946 : 694) observed that the canine of B.M. 36803 was grooved and incompletely erupted, but did not record the curious external striations of the enamel and the enamel distribution. In B.M. 29718a the canine alveolus extends to the posterior end of the symphysis; in S.M. C30611 and B.M. M25094, it extends posterior to this. Fic. 17. Opsiclaenodon major (Lydekker, 1887). Fragmentary left mandibular ramus with C, Pe-Ps. B.M. 29863 3:5. Lower Headon Beds, Hordle Cliff. (a) Occlusal view. (B) Internal view. (c) External view. 54 UPPER EOCENE MAMMALIA P, from its alveolus (B.M. 36803, B.M. 29863 and B.M. Mz25094) was evidently small, single-rooted and placed immediately posterior to the canine. Pz is preserved in B.M. 36803, S.M. C54162a and B.M. 29863; the tooth is double-rooted and has a slight forward inclination. From a single large anterior cusp, steep posterior and postero-internal crests fall to the antero-external and antero-internal extremities, respectively of a small slightly basined heel. An anterior crest from the main cusp curves internally at the anterior tooth margin forming a faint cingulum. On the antero-internal wall of the tooth, just posterior to the anterior crest, there is a vertical groove, similar to though less pronounced than that of the canine. Horizontal striations similar to those of the canine, are well developed externally on the main cusp in B.M. 29863 and S.M. C54162a, though less prominent on B.M. 36803. P3 is known from S.M. C3061r and B.M. 29863; these specimens were not listed by Butler (1946) who described only alveoli of this tooth. The tooth is generally similar to Pg but larger with a large crested main cusp. The anterior crest curves internally at the base of the antero-internal groove forming a very faint cingulum- like cuspule (S.M. C30611) or a strong low cusp (B.M. 29863). The posterior crest continues antero-posteriorly across the middle of the well developed heel. Externally, there is some development of irregular, subhorizontal enamel striations. Py, (B.M. 29718a, S.M. C30611 and B.M. 29863) is very similar in structure to P3 but much larger. The antero-internal groove is marked clearly on $.M. C30611 and B.M. 29863, but less prominent on B.M. 29718a. According to Butler (1946 : 695), the groove is absent in B.M. 29718a, the only specimen containing Py, listed by him. The antero-posterior talonid crest in S.M. C30611 swells posteriorly to form a cusp; in B.M. 29863, the median region of the crest is raised as a distinct cusp. Two small internal cuspules occur on the talonid of B.M. 29718a; these are absent in the other specimens. M,-Ms are structurally very similar with rounded trigonid and talonid margins; M, and Mp are subequal; M3 is markedly smaller. Identification of isolated teeth Fic. 18. Opsiclaenodon major (Lydekker, 1887). Composite right M;—Mg all approx. x5. M, (B.M. M12565b) and Mz (S.M.C 54162d). Lower Headon Beds, Hordle Cliff. (A) Occlusal view. (B) External view. (c) Internal view. HEADON BEDS OF ENGLAND 55 as M, or Mgis uncertain ; M, has been distinguished here by a slight anterior narrowing of the trigonid. The metaconid and protoconid are subequal (when unworn), opposite and transversely crested. The small though distinct low paraconid is median or slightly internal to the midline, forming the fairly sharp anterior apex of the tooth; a crest from this cusp meets the anterior protoconid crest. A fairly strong shelf-like antero-external cingulum is present. The hypoconid and entoconid are prominent and opposite, the hypoconid being slightly larger; each cusp is united to the posterior trigonid wall by sharp crests. The hypoconulid is small, though distinct; it is strongly posterior to the hypoconid and entoconid, and posteriorly projecting; the cusp is usually slightly external to the midline and is joined to the hypoconid and entoconid by low crests. This condition is seen in the M3 of B.M. 36807, but this cusp in the Ms of B.M. 29718 and B.M. 36803 is distinct and median. The talonid is strongly basined. Small accessory crests and cuspules are variably developed in the molars: in B.M. 36803, B.M. 36807 and B.M. M25og1, a discon- tinuous crest swings obliquely down the posterior trigonid wall from the metaconid towards the crista obliqua; this cannot be distinguished in other specimens owing to excessive wear. On Mz of B.M. 36807 and S.M. C54162d, there is a small cuspule in the middle of the talonid at the base of the posterior trigonid wall. This is absent in S.M. C54162b and S.M. C54162c. In other specimens wear is too great for this feature to be distinguished. The upper teeth are known only in the isolated state. S.M. C30713 and S.M. C30722, almost identical structurally, are the largest teeth; these are considered as M2. S.M. C54163 is somewhat smaller antero-posteriorly, showing several minor structural differences; this is tentatively considered as M1. S.M. C9677, markedly smaller in length and width than the other teeth, probably represents M3. M2? is elongated transversely and has a rounded outline. The antero-internal crescentic protocone, the largest cusp, dominates the internal half of the tooth. The conical paracone is subequal in height with the protocone; the metacone is smaller. There is a strong shelf-like external and postero-external cingulum round the metacone base, ending externally as a small mesostyle, and a similar anterior cingulum round the base of the paracone (not present externally). The parastylar region is extended antero- externally, the metastylar region extended postero-externally, as rounded lobes. The sharp postero-externally trending protocone crest is interrupted just internal to Fic. 19. Opsiclaenodon major (Lydekker, 1887). Occlusal view of composite M!-M3 x5. Right M1 (S.M. C54163), right M2 (S.M. C30713) and left M® (S.M. C9677), drawing reversed. Lower Headon Beds, Hordle Cliff. 56 UPPER EOCENE MAMMALIA the metacone by a prominent metaconule. A smaller paraconule is similarly placed on the protocone crest just internal to the paracone. There is a small low anterior cingulum and a very strong shelf-like postero-internal cingulum which supports externally a very strong hypocone which is subequal in height with the metacone. S.M. C30722, a well worn M2, has‘a more developed anterior cingulum. M1! is basically similar to M? and although the tooth is more worn than M2 (S.M C307713) and damaged externally, significant differences can still be recorded: the tooth is shorter antero-posteriorly than M2. The hypocone and hypocone-cingulum are less developed; the metaconule is very small and the paraconule, just distinguishable on the slightly worn crest must have been very rudimentary on the unworn tooth. The postero-external cingulum round the base of the metacone is absent and there is only a trace of the antero-internal cingulum. M3 is basically similar to the other molars but much smaller and compressed antero-posteriorly. The metaconule is very small, merging with the antero-external cingulum and an internal paracone crest. The antero-external trending protocone crest joins the paracone; a small swelling on this crest just internal to the paracone indicates a rudimentary paraconule. A small postero-external cingulum is present, but not round the metacone base as in M! and M2; this latter cusp, slightly smaller than the hypocone, is more antero-internal than in M! and M2. RemArRKS. There is considerable variation in the depth of the mandible: B.M. 29718a is greatest, B.M. 29718 intermediate, and B.M. 36803 smallest in size. B.M. 36803 has unworn teeth and an incompletely erupted canine, indicating derivation from a very young individual. This suggests that the variation in mandible depth is, at least in part, related to individual age differences. There are slight variations in the positions of the mental foramina; in the degree of prominence of the antero- TABLE I0 Measurements (in mm) of Opsiclaenodon major Pe P3 Pa Mi Me M3 a-p tis a-p ts a-p tts) a-p ts a-p ts a-p iis B.M. 29718 ee B.M. 29718a —- — — — 45 2:3 40 34 41 3:3 34 26 B.M. 36807 SE OE BM 36802 —_ — -— — — 4°3 29 Bz 233 B.M. M12565b — es B.M. 36803 sys rechs} — = =| SF a6 B.M. 29863 FOR oan s 2 45 24 => =| = => — == B.M. M25094 — — 36 2:0 = SO B.M. Mz25091 Se ia ea iS iS SPEECH S.M. C53478 SS ee), Se OE ee Se S.M. C54162a 30 17 — — — — — — ~~ ~—- —- — » on b eee | ee — 44 32 == = ” » C i SS SS SS iS ST OS oe Sa a ee Se SS a oF SS SS ee ee Oe a ee Oe ee oe ae S.M. C30611 — — 37 19 42 24 — — — — — — HEADON BEDS OF ENGLAND 57 internal groove and antero-external enamel striations in the premolars, and in the development of accessory cuspules. These have been considered in the preceding description. Genus DYSPTERNA Hopwood, 1927 Dracnosis. Lower molars with low blunt cusps, Mg trigonid rather compressed antero-posteriorly. M, and M3 probably subequal in size, Mz larger. Entoconid of molars vestigial or absent with postero-internal talonid margin raised as a crest, hypoconid large. Hypoconulid of molars median and small to vestigial. TYPE SPECIES. Dyspterna woodi Hopwood, 1927. Early Oligocene (‘Sannoisian’), Yarmouth, Isle of Wight. Remarks. The genus was originally based by Hopwood on a single specimen, from the Lower Hamstead Beds, a very fragmentary worn and damaged mandibular ramus with M2—M3 (B.M. M13125), provisionally placed in the Oxyclaeninae. Three years later, Dal Piaz described (1930) a maxillary fragment with P4-M3 which he tentatively referred to Hopwood’s species. This specimen was removed by Van Valen (1966 : 88) from Opsiclaenodon as the type of a new apheliscine, Epapheliscus ttalicus Van Valen, 1966. Dal Piaz also described a new species (1930 : 13) which he tentatively referred to the genus as ?Dyspterna helbingi; the holotype (and only known specimen) is a fragmentary left maxilla with P? and the roots of P4-M?. Recently, Van Valen has observed (1966 : 88) that from the published figure and description, the affinities of this specimen cannot be determined. In his redescription of Neurogymnurus major as Opsiclaenodon major, Butler suggested (1946 : 698) a possible relationship with Dyspterna woodi. The genus was not listed by Simpson (1945) being based on inadequate material, and McKenna stated (7m Russell & McKenna, 1961 : 281, footnote 3) that Dyspterna constitutes practically a nomen dubium. Van Valen (1966 and 1967) referred Dyspterna to the Pantolestidae. In the present work, material collected fairly recently from the Upper Headon Beds has been referred to Dyspterna (on the basis of molar structure) as a new species. The anterior lower dentition of this species of Dyspterna is strikingly similar to that of Opsiclaenodon and the two genera appear to be closely related, thus supporting Butler’s original view. Opsiclaenodon is distinguished generically from Dyspterna by the uncompressed Mg trigonid; the molar entoconid is strongly developed; the hypoconulid is well defined and the cusps are more trenchant and separated. The prominent hypoconid and vestigial or absent entoconid sharply distinguishes Dyspterna from all other pantolestines. These and other specialised features could have arisen by modification of a dental structure similar to that of Opsiclaenodon. Van Valen has stated (1967 : 227) that Opsiclaenodon could be directly ancestral to Dyspterna. This possibility is precluded by the contemporaneous occurrence of Dyspterna hopwoodi sp. nov. and Opsiclaenodon major in the Headon Beds; both genera were probably derived from an earlier common ancestry, perhaps in the Middle Eocene. 58 UPPER EOCENE MAMMALIA Dyspterna hopwoodi sp. nov. (Text-figs 20, 21) ErymoLocy. After A. T. Hopwood. Dracnosis. My, and M3 hypoconid large, entoconid and hypoconulid vestigial, appearing as small swellings on postero-internal marginal crest. Crista obliqua in M, and Ms directed anteriorly to base of prococonid. Mg trigonid more compressed antero-posteriorly than in D. woodt, with the small paraconid strongly antero-internal in position, close to metaconid; M3 hypoconulid vestigial. Sharp condyloid and coronoid crests bordering deep masseteric fossa. HototyrPe. B.M. M26052. Right mandibular ramus with the canine, P2—P4 and Ms, and alveoli of I;-Is, P1 and Mj-Mz. Symphysis and posterior region of mandible are preserved. Slightly worn dentition. Upper Headon Beds (Lignite Bed), Headon Hill. PARATYPES. B.M. M26053. Isolated right My. Slightly worn. Upper Headon Beds (Lignite Bed), Headon Hill. B.M. M26054. Isolated right P}. Unworn. Upper Headon Beds (Lignite Bed), Headon Hill. DESCRIPTION. The mandible is moderately stout with a slight anterior deepening as in Opsiclaenodon. The symphysis is elongated and rectangular in outline, ending posteriorly under Pp: and inclined at about 45° to the level of the tooth row. The posterior region of the mandible is complete in the holotype, except for the condyle which lay just above the level of the tooth row. The coronoid process is well developed and massive, the anterior edge rising steeply from the tooth row level. The notch between the condyle and the coronoid is hardly perceptible; this is in contrast to Opsiclaenodon, where it appears to have been more developed. The angular is fairly prominent and well separated from the condyle; the lower edge is deflected internally as a small flange. The mandibular foramen is situated just below alveolar level; the ridge above this running from the condyle to a position below Mg is more prominent than in Opsiclaenodon. The anterior mental foramen is sited below the canine/P;, the posterior mental foramen below Mj. The bases of the alveoli of I;-Is indicate that these teeth were closely crowded and situated immediately anterior to the canine. The canine is almost identical to that of Opsiclaenodon major with the upper half of the tooth enamel covered; externally the enamel is wrinkled, forming regular striations. The upper half of the tooth is triangular-shaped with sharp antero-internal, internal and postero-external crests, and a prominent deep antero-internal groove. The groove and crests are confined mainly to the enamel-covered upper half of the tooth. Below this level, the crests become rounded but the groove, in contrast to Opsiclaenodon major is continued to alveolar level as a distinct but rounded constriction. From the alveolus in the holotype, it appears that P, was a little separated from the canine; the tooth (B.M. M26054) is single-rooted with a main central cusp and a small postero- internal heel. The cusp is crested anteriorly, postero-internally and posteriorly. There are small but distinct antero-internal and postero-external cingula and a cingulum trace antero-externally. A rudimentary antero-internal groove is present. HEADON BEDS OF ENGLAND 59 Py is slightly larger than P; and double-rooted. There are no antero-external and postero-external cingula; otherwise the tooth is structurally nearly identical to Pj. There is the suggestion of a slight antero-internal groove. Pg is identical to Pe, except for a larger size and a slightly stronger cusp-like antero-internal cingulum. There is a trace of an antero-internal groove. Py, is structurally quite similar to Ps but much larger. The sharp anteriorly crested main cusp falls to a prominent low anterior cusp; this latter cusp is larger than that of Opsiclaenodon and the tooth is generally more pointed. There are small antero-external and antero-internal cingula. The antero-posterior talonid crest ends posteriorly inasmall cusp. There is a slight antero-internal groove. It is difficult to estimate reliably the relative sizes of My and Mg from the alveoli in the holotype of D. hopwoodi because of mandible damage. By comparison with the holotype of D. wood: (about the same size as D. hopwoodt) where M3 is smaller than Fie. 20. Dyspterna hopwoodi sp. noy. Right mandibular ramus with C, Ps—P4 and Ms. Holotype (B.M. M26052). Upper Headon Beds, Headon Hill. (a) Occlusal view x2°5. (B) Internal view x2. 60 UPPER EOCENE MAMMALIA Mg, it has been concluded that an isolated molar (paratype) of D. hopwoodi is best considered as My. This latter tooth is elongated with a rounded margin. The protoconid and metaconid are subequal, opposite and transversely crested. The small but distinct low median paraconid forms the sharp anterior apex of the tooth. There is a moderately strong antero-external cingulum and a smaller antero-internal cingulum. The entoconid and hypoconid are vestigial, appearing as small swellings on a low curved crest which forms the postero-internal talonid margin. The median hypoconulid projects slightly posteriorly; the entoconid and hypoconid are opposite each other. The talonid is slightly basined. Msg is almost the same length as Mj, with a subequal and opposite protoconid and metaconid. The smaller paraconid is strongly internal and anterior to the metaconid; the trigonid appears antero-posteriorly compressed. The almost straight antero-external trigonid margin is formed by a paraconid and protoconid crest. There is a small antero-external cingulum but the antero-internal cingulum is absent. The vestigial hypoconulid is represented by a minute rounded median lobe which is posteriorly projecting and separated from the larger hypoconid. The entoconid is vestigial; a low rounded crest runs anteriorly from the hypoconulid-lobe, forming the postero-internal tooth margin. The talonid is only very slightly basined. Remarks. In the canine of the holotype (B.M. M26052) distinct wear facets are visible: part of the postero-external face has been gauged away by the upper canine in occlusion, to form above alveolar level, a prominent oblique baso-exteriorly directed shelf. Wear has taken place along the whole vertical length of the tooth from the tip to this shelf, but becomes progressively greater towards this latter position; presumably slight wear took place near the canine tip when the tooth was partially erupted (compare Opsiclaenodon major, B.M. 29863). Anteriorly and basally to the prominent groove, there is a shallow vertical oval facet, produced probably by friction against one of the upper incisors. Fic.21. Dyspternahopwoodisp.nov. Right P; (B.M.M26054) and right Mz (B.M. Mz6053). Both x4. Upper Headon Beds, Headon Hill. (a) Occlusal view. (8) Internal view. (c) External view. HEADON BEDS OF ENGLAND 61 The antero-internal groove does not show any signs of wear; the bordering crests are very sharp and quite unworn. The same situation exists in the closely similar canine of Opsiclaenodon major. As Butler noted (1946 : 695), a similar unworn groove occurs on the lower canine of Nasua, a recent procyonid. No information has been found about the significance of this groove in Nasua. On the available evidence, the structure in all three genera would appear to be non-functional, at least in occlusion. In molar structure, D. hopwoodi is closely related to the type species, D. woodi and may be ancestral to the latter. D. hopwoodi sp. nov. appears specifically distinct on the basis of the following differences: in D. woodi, the hypoconid crest is directed antero-internally to meet the base of the metaconid, in Mg running up the posterior wall of the metaconid; the cingula are relatively larger and more shelf-like; the M3 trigonid is less compressed antero-posteriorly with the paraconid slightly more median ; the hypoconulid appears to have been relatively larger, and the coronoid and condyloid processes are lower and more rounded, enclosing a shallower masseteric fossa. More detailed comparison is not possible because of the poor preservation of the holotype of D. woodi. TABLE II Measurements (in mm) of Dyspterna hopwoodi Py Pe P3 Pa My Mz Ms a-p ts ap ts a-p tr a-p trs a-p trs a-p trs a-p trs B.M. M26052 — — 3:6 2:3 4:7 298 66 3:3 — — — — 5:0 3:6 B.M. M26053 — 3 — ~~ —~ —~—~ ~~ ~~ ~~ 54 37 —- ~—- —~— — B.M. M26054 3°44 29 — — — - -—- —_—- — Family APATEMYIDAE Matthew, 1909 DiacGnosis. Genera with a reduced dental formula—probably Z 4 5, ;. Jaw deep and stout in proportion to molars, with posterior mental foramen below M, or Mz. Marked soricid-like development of lower incisors, the roots of which terminate below Mg. Pg blade-like and single-rooted. Molars in early forms trenchant; more typically, bunodont and primate-like in appearance. P%—P4 double-rooted and unicuspid. Upper molars relatively elongated antero-posteriorly and broadly expanded postero-internally with a well developed hypocone, but lacking a mesostyle. Lower molars, except for Mz, have a simple talonid but the trigonid, particularly in M2-Msz shows a distinctive parallelogram arrangement of cusps and crests. (Mainly after Gazin, 1958). Remarks. Of the seven genera assigned to the family, five are known from North America and two from Europe: Jepsenella Simpson, 1940. Middle Palaeocene; North America. Labidolemur Matthew & Granger, 1921. Late Palaeocene; North America. Apatemys Marsh, 1872. Early-Late Eocene; North America. 62 UPPER EOCENE MAMMALIA Stehlinella Matthew, 1929. Late Eocene; North America. Sinclairella Jepsen, 1934. Early-Middle Oligocene; North America. Eochiromys Teilhard de Chardin, 1927. Early Eocene; Belgium. Heterohyus Gervais, 1848-52. Middle-Early Eocene; Europe. Van Valen (1967) considered Teilhardella Jepsen, 1930 was invalid and included it within Apatemys. In addition to the above generic list, McKenna (1963 : 30) suggested that Unuchinia Simpson, 1937, previously placed incertae sedis in the Insectivora, might be an apatemyid. More recently, Van Valen (1967) formerly erected a new subfamily Unuchiniinae Van Valen & McKenna, for the reception of Unuchinia. Stehlinella is known from the anterior portion of the skull with part of the upper dentition and an almost complete mandibular ramus lacking only Py. Sznclairella is known from a complete but crushed skull with part of the upper dentition and a mandibular ramus lacking P3-P, and Mj. The remaining six genera are based on fragmentary mandibles, although Heterohyus is known also from a palatal fragment. Owing to incomplete material and the comparative rarity of specimens, classifica- tion is very tentative. Some genera are probably composite but better material is needed before significant advances can be made. The family has oscillated between Primates and Insectivora for many years. The most recent detailed work on the affinities of this problematical group is McKenna (1963). From a consideration of molar structure in the earliest genera, McKenna referred the family to the Insectivora; this view, followed by Van Valen (1967) is adopted here. Genus HETEROHYUS Gervais, 1848-52 Diacnosis. Usually single mental foramen below Mg; fossa below Py, generally reduced. Enamel covering lower incisors extends into alveolus but does not completely coat the crown; P,4 generally reduced ; Mg talonid usually strongly elongated and Ms posterior trigonid wall usually oblique to longitudinal axis of tooth. TYPE SPECIES. Heterohyus armatus Gervais, 1848-52. Middle Eocene (Early Lutetian) ; Bouxwiller, France. ReMARKS. Distinguished from the other genera as follows: Jepsenella shows only incipient development of apatemyid characteristics; in Eochiromys and Labidolemur the incisor crown is completely covered by enamel. These genera and all other apatemyids possess either a little-elongated Mg talonid or lower molars with a transverse posterior trigonid wall. Five species have been distinguished: H. armatus Getvais, 1848-52, pl. 35, fig. 14. Middle Eocene (Early Lutetian) ; France. H. quercyi (Filhol, 1880), p. 174, figs 1-3. Late Eocene (Early Ludian, Quercy Phosphorites) ; France. H. nanus Teilhard de Chardin, 1922, p. 93, fig. 41; pl. 4, fig. 21. Late Eocene (Early Ludian, Quercy Phosphorites) ; France. HEADON BEDS OF ENGLAND 63 H. europaeus (Riitimeyer, 1890), p. 346, figs 1-2. Middle Eocene (Early-Middle Lutetian) ; Switzerland. H. heufeldert Heller, 1930, p. 32, pl. 5, fig. 4. Middle Eocene (Lutetian) ; Germany. The most complete species, H. guercy1, is based on a crushed skull and almost com- plete upper dentition and a fragmentary mandibular ramus; H. manus and H. aymatus are based on mandibular fragments, H. heufelderi on a complete lower dentition. The type of H. aymatus (the type species) was not located by Teilhard de Chardin (1922 : 89) and here, is considered to be lost. Stehlin (1916 : 1462) erected the genus Heterochiromys for two species from Egerkingen: H. gracilis known from mandibular rami lacking teeth and isolated incisors, and H. fortis based on isolated incisors. It is highly probable, as suggested by Teilhard de Chardin (1922 : 92), that Heterochiromys gracilis Stehlin is synonymous with Heterohyus quercyi (Filhol) and Heterochivomys fortis Stehlin, is synonymous with Heterohyus armatus Gervais. Stehlin (1916 : 1434) also placed several mandi- bular rami lacking teeth, and isolated incisors from Egerkingen in a new genus and species, Amphichiromys europaeus. This was based on minor differences in the incisors. Isolated lower molars tentatively assigned by Stehlin to this genus are of characteristic heterohyid type and apparently do not warrant generic separation. The genus is considered synonymous with Heterohyus. The generic diagnosis is unsatisfactory and probably, Heterohyus is a composite genus. H. nanus does not fit well into this diagnosis and appears to be indistinguish- able from Apatemys bellus. American material has not been examined, so this view is only tentative. The material described below, constitutes the first record of an apatemyid from the Tertiary of Britain. Heterohyus sp. (Text-fig. 22) Material. S.M.C54172. Isolated left M3. Moderately worn. Lower Headon Beds (Rodent Bed), Hordle Cliff. DEscrIPTION. This strongly elongated tooth is composed mainly of a gently basined talonid, the latter being 2} x the length of the trigonid. The trigonid is wider than the talonid and externally projecting. The low blunt cusps and crests of the trigonid exhibit a distinctive parallelogram arrangement. The metaconid and paraconid are almost equal in size, the antero-internal paraconid being directly anterior to the metaconid ; these cusps are united by a massive antero- posterior crest. The smaller protoconid is moderately separated from the metaconid and paraconid, and more anteriorly placed in relation to the metaconid; the posterior trigonid wall is thus markedly oblique to the longitudinal axis of the tooth. From the paraconid a small low crest runs parallel to the posterior trigonid wall, forming the oblique anterior tooth margin. The antero-external extremity of the tooth, directly anterior to the protoconid, is sharp and acute-angled ; here, the crest is very faint and low. 64 UPPER EOCENE MAMMALIA The talonid basin is bounded externally and internally by rounded fairly strong crests which are almost parallel. The higher crista obliqua meets the posterior trigonid wall midway between the metaconid and protoconid; the internal crest is continued faintly up the posterior wall of the metaconid. The internal talonid margin is damaged; it may have curved out internally, with a more rounded outline than the external margin. Around the posterior edge of this damaged area is the possible trace of a small entoconid. From this position, the straight talonid margin runs to a relatively high rounded externally shifted hypoconulid-lobe. From the entoconid, a minute crest dilates as it curves round to meet a prominent hypoconulid, in the anterior part of the hypoconulid-lobe. No hypoconid is distinguishable, although the tooth is damaged anterior to the hypoconulid; a small cusp may have been present. The total antero-posterior length of Mg is 5-8 mm; the trigonid length is 1-6 mm, that of the talonid 3-7 mm. The trigonid transverse width is 2-9 mm, the talonid width is 2-2 mm. REMARKS. This specimen is distinguished from the Mg of known species as follows: H. nanus is much smaller with a short talonid and a slightly oblique trigonid; H. quercyt is much smaller with a slightly oblique trigonid; H. europaeus is larger with a slightly oblique trigonid; H. heufelderi is smaller with a relatively shorter talonid and in H. armatus the talonid is less elongated and wider transversely than the tri- gonid. Also, the tooth outline in all the above is more prominently rounded. The Headon Beds specimen appears closest to H. armatus. However, little is known about variation within the genus; the known species are restricted to the Fic. 22. Hetervohyus sp. Left M3. S.M. C54172 approx. x 8. Lower Headon Beds, Hordle Cliff. (A) Occlusal view. (8B) Internal view. (c) External view. HEADON BEDS OF ENGLAND 65 Lutetian and Early Ludian and the genus already may be oversplit. S.M. C54172 could represent a distinct new species but better material is needed. At present the specimen seems best recorded as Heterohyus sp. TABLE 12 Measurements (in mm) of M3 of Hetevohyus species Angle between Ratio of posterior trigonid wall of length to trigonid and Total a-p talonid longitudinal length length axis H. heufelderi (ex Heller, 1930) 4:0 — = H. quercyi (ex Teilhard de Chardin, 1922) 3°9 1:24 100° H. nanus (ex Teilhard de Chardin, 1922) 2°5 r:14 105° H. armatus (ex Gervais, 1848-52) 7:0 1653 7 130° H. europaeus (ex Stehlin, 1916) 9:0 ig BOs T00° H. sp. (S.M. C54172) 5:8 T 224 120° Order PRIMATES Family ADAPIDAE Trouessart, 1879 Diacnosis. Dental series usually continuous—no marked diastema. Incisors typically wide-crowned and trenchant; upper canines straight, lower canines short, subpremolariform. Upper molars tritubercular or quadritubercular with a large V-shaped protocone; small but true hypocone derived from posterior cingulum; mesostyle usually absent. Paraconid absent or much reduced on lower molars; entoconid retarded; hypoconulid absent or rudimentary except on Mg. Mandible generally short and stout with a greatly expanded angular and a high recurved coronoid process. Condyle elongated transversely and relatively flattened. Skull usually with high sagittal and lambdoidal crests. Small ovoid braincase. Prominent heavy facial region. (after Hill, 1953). REMARKS. Simons (1962) presented a revised classification of European Eocene primates. This revision, together with Gazin’s revival (Gazin, 1958 : 31) of the family Notharctidae Trouessart, 1879, has significantly altered Simpson’s classifica- tion of 1945. The generic list follows Simons (1962) with the addition of Lantianus Chow, 1964. Adapis Cuvier, 1822. Middle-Late Eocene; France, Britain, Switzerland. Pronycticebus Grandidier, 1904. Late Eocene; France. Anchomomys Stehlin, 1916. Middle-Late Eocene; France, Switzerland. Caenopithecus Ritimeyer, 1862. Middle Eocene; Switzerland. Protoadapis Lemoine, 1878. Early-Late Eocene; France, Germany. Gesneropithex Hiirzeler, 1946. Late Eocene; Switzerland. Lantianus Chow, 1964. ?Late Eocene; China. 66 UPPER EOCENE MAMMALIA In addition to Caenopithecus, Simpson, (1945) considered two further genera as Adapidae incertae sedis: Aphanolemur Granger & Gregory, 1917 and Amphilemur Heller, 1935. Aphanolemur is now believed to be a synonym of Smilodectes Wortman, 1903, anotharctid (Gazin, 1958 : 30). . Amphilemur has been considered by McKenna (1960 : 58) as an erinaceoid insectivore. More recently, Simon’s generic list has been accepted by McKenna (1967) and Russell, Louis & Savage (1967) although the latter authors regarded reference of Gesneropithex as provisional and tentatively added Periconodon Stehlin, placed by Simons (1962) and McKenna (1967) in the Omomyidae. Genus ADAPIS Cuvier, 1822 Synonomy. See Hill, 1953 : 464. Diacnosis. Mandible has a greatly expanded angular process; coronoid process separated from condyle by a fairly shallow notch. P ; present, small in size; P4 submolariform but cusps variably developed ; Mj and Mg similar to Py; Mg considerably elongated by hypoconulid which is not isolated distinctly as a third lobe; metastylid present in later species. P! present, small in size; P4 molariform with a bicusped ectoloph. Upper molars lack mesostyle; hypocone not strongly developed, absent in M%; metaconule rudimentary or absent; strong external cingulum; internal cingulum continuous. TYPE SPECIES. Adapis parisiensis (Blainville, 1841). Late Eocene (Late Ludian); France. REMARKS. Anchomomys, Pronycticebus, and Caenopithecus differ from Adapis by having the Mg hypoconulid more distinctly separated as a lobe and by lacking the bicusped ectoloph of P, and the continuous internal upper molar cingulum. Caenopithecus is further distinguished by lacking P! and having a mesostyle. In Protoadapis Pz is much smaller than Pg and the molar paraconids extend more internally. Gesneropithex has a median paraconid; the metaconid—protoconid crest is perpendicular to the longitudinal axis of the tooth and the Mg trigonid is compressed antero-posteriorly. The following species are referred to Adapis: A. parisiensis (Blainville, 1841), p. 112, pl. 9. Late Eocene (Bartonian-Late Ludian) ; France. A. magnus Filhol, 1874, p. 11, pl. 8, figs 9-12. Late Eocene (Bartonian-Early Ludian) ; France, Britain. A. ruetumeyert Stehlin, 1912, p. 1261, pl. 21, figs 5, 8, 12, 14. pl. 22, figs 17-19. Middle Eocene (Late Lutetian) ; Switzerland. A. sciureus Stehlin, 1916, p. 1514, text-figs 368-370, pl. 22, fig. 5. Middle Eocene (Early Lutetian) ; Switzerland. A. priscus Stehlin, 1916, p. 1511, text-figs 366-367. Middle Eocene (Early Lutetian) ; Switzerland. Simons (1962) observed that the holotype of Adapis minimus Heller, 1930, is indeterminate owing to removal of enamel by post-mortem chemical absorption. HEADON BEDS OF ENGLAND 67 Adapis magnus Filhol, 1874 (Ble i e lee tie 1) Diacnosis. Large size. Molars moderately elongated; metastylid developed. Upper canines large with grooved crown; P? transversely elongated with prominent internal cingulum and talonid. Hototypre. Fairly complete skull lacking the incisors and canines of both sides, left P!-P2 and M3. ?Late Eocene/Early Oligocene (Quercy Phosphorites) ; Raynal, France. Described and figured, Filhol, 1874, p. 11, pl. 8, figs 9-12. Specimen not located. MaTeERIAL. All specimens from the Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff. B.M. 30323. Fragment of right mandibular ramus with M\—-M3. Very slightly worn. Brown shelly sand matrix. Listed, Lydekker, 1885a : 263. B.M. 30856. Fragment of right mandibular ramus with Mz-M3. Moderately worn. Brown sand matrix. Listed, Lydekker, 1885a : 263. B.M. 29864. Fragment of right mandibular ramus with P,—P), and posterior root of P3. Moderately worn. Listed, Lydekker, 1885a : 263. B.M. 29741. Fragment of left mandibular ramus with Py, M3, roots of Pz and part of the canine. Teeth and alveoli of P;-Py are badly damaged. External surface embedded in brown clayey sand matrix. Listed, Lydekker, 1885a : 263. B.M. Mio260. Fragment of left mandibular ramus with P3—Mg and alveoli of M3. Fairly well worn. B.M. Mi1o269. Fragment of right mandibular ramus with Pa-Mz. Teeth unworn. B.M. M10267 (3). Fragment of left mandibular ramus with Mj-Mz. Moderately worn. D.M.S.W. M556 (41). Fragment of left mandibular ramus with Py—Mg. B.M. M26057 (1). Right Mp. (2). Right Mj. Both teeth slightly worn. Wear and preservation indicates probable derivation from same individual. B.M. M20560. Right Mg. Badly worn. B.M. M25093. Left M3. Badly damaged. S.M. C30680. Fragment of right mandibular ramus with M,;—-Mz. Slightly worn. S.M. C53483. Left M3. Unworn S.M. C53477. Right Ms. Moderately worn. S.M. C54221. Right Me. Unworn. S.M. C30659. Right My. Moderately worn. S.M. C30660. Right My. Well worn. S.M. C30656. Left My. Well worn. S.M. C30658. Right Mz. Well worn. S.M. C30657. Right My. Well worn. S.M. C31345. Left Me (?). Well worn and badly damaged. B.M. M3572. Fragment of left maxilla with left half of palate and P2-P?, M1-M3 andalveoliofthecanineand P1, P2isdamaged. Teethunworn. Listed, Lydekker, 1887 : 299. 68 UPPER EOCENE MAMMALIA B.M. Mroz261 (1). Left upper molar (M2?). Moderately worn. 2). Right upper molar (M2?). Moderately worn. ). Upper molar. Moderately worn but badly damaged. ). Left upper molar (M?2?). Slightly worn. ). Left P4. Moderately worn. S.M. C53484. Left upper molar (M??). Unworn. S.M. C30661. Right upper molar (M??). Slightly worn. S.M. C30662. Right upper molar (M!?). Slightly worn. Probably from same individual as S.M. C30661. B.M. M25098. Right upper molar (M??). Unworn. B.M. 25092. Right upper molar (M!?). Moderately worn. S.M. C53479. Left upper molar (M??). Moderately worn. B.M. 12564 (1). Left upper molar (M??). Slightly worn. (2). Right P4. Moderately worn. M. C30681. Right upper molar (M3?). Well worn. M. C30679. Left upper molar (M!?). Badly worn. M. C30673. Left upper molar (M2?). Unworn. M. C30674. Left upper molar (M!?). Slightly worn. Probably from same vidual as S$.M. C30673. M. C30678. Left P4. Moderately worn. M. C30675. Left upper molar (M2?). Badly worn. M. C30676. Right upper molar (M?2?). Slightly worn. .M. C30677. Fragment of upper molar. B.M. 30346a. Listed, Lydekker, 1885a : 263. (1) Right upper molar (M?2?). Slightly worn. (2) Left upper molar (M??). Slightly worn. Probably from same individual as B.M. 30346a (I). (3) Left P4. Unworn. (4) Left upper molar (M??). Unworn. (5) Right upper molar (M1?). Moderately worn but badly ( ; B.M. M20560 (1 (2 K NNNNRNANMNW damaged. (6) Left upper molar (M1?). Unworn but slightly damaged. (7) Left upper molar (M??). Unworn but slightly damaged. Probably from same individual as B.M. 30346a (6). (8) Left upper molar (M3?). Unworn. (9) Left P8. Shghtly worn. B.M. M25096. Right P2. Slightly worn. B.M. M10267. (1) Left P?. Unworn. (2) Right upper canine. DEscRIPTION. The base of the strong coronoid crest and the shallow masseteric fossa are preserved in B.M. M1o260. A small mandibular foramen occurs just below alveolar level. The posterior mental foramen lies between P3 and Pa. The lower incisors and P;—Ps are unknown from the Headon Beds. The lower canine is known (B.M. 29741) only by the root. Ps, preserved only in B.M. M1o260, HEADON BEDS OF ENGLAND 69 is oval and double-rooted, with an almost continuous cingulum; it has a main crested cusp which falls to a small anterior cusp and a single-cusped postero-internal talonid. Py (unworn in B.M. M10269) is moderately molarised with the protoconid larger than the metaconid. The metaconid is median and posterior to the protoconid. A slight swelling on the postero-internal wall of the metaconid (distinct in B.M. M10269 but indistinct in B.M. Mro260) indicates a rudimentary metastylid. These cusps, which are crested transversely, form the oblique posterior trigonid wall. There are small external and antero-external cingula. A straight crest from the protoconid falls antero-basally to a small median paraconid which merges into a prominent antero-internal cingulum. From the prominent hypoconid, a crest runs antero-internally (crista obliqua) to meet the posterior protoconid wall, and internally to join a small but distinct entoconid (posterior to the hypoconid). A low crest from the entoconid runs anteriorly to the base of the metaconid. In B.M. Mro260, the P, entoconid is vestigial. In M,, protoconid and metaconid are subequal when unworn. A metastylid is well differentiated, although it is indistinct in B.M. M1o0260; this cusp, with the crested paraconid and metaconid forms the strongly oblique posterior trigonid wall. From the protoconid a crest curves antero-internally to meet the paraconid external to the midline. This latter cusp, as in Py, merges with a strong antero-internal cingulum. There are faint antero-external, external and postero-external cingula. The hypoconid is slightly larger than the more posteriorly situated entoconid; these cusps are joined by crests to a small but distinct median hypoconulid which is slightly posterior to the entoconid. There is a prominent crista obliqua. Mg js closely similar to Mj, except for its larger size. The trigonid of Mg is almost identical with that of M; and Mg, except for a generally poorly developed metastylid (in B.M. 30323 and B.M. 30856, for example). The hypoconid is strongly developed (as in M,—Mg), with a prominent crista obliqua. The entoconid is vestigial (B.M. 30856) or represented by a low swelling (B.M. 30323) posterior to the hypoconid. The median hypoconulid (external to the midline in S.M. C53483) is as large as the hypoconid and surmounts a large rounded posteriorly projecting hypoconulid lobe which greatly extends the length of the talonid. From the hypoconulid, a low crest curves antero-internally, forming the internal talonid margin. The hypo- conulid lobe is well separated from the hypoconid. The upper canine alveolus (B.M. M3572) indicates a fairly large curved tooth. An isolated canine (B.M. Mro267 (2)) has an almost continuous cingulum, more prominent internally. The tooth is grooved vertically on the anterior and posterior faces. P! is known from the Headon Beds only by an alveolus (B.M. M3572) indicating a very small single-rooted externally placed tooth. P?is oval and obliquely set in the tooth row (B.M. 25096) with an almost continuous cingulum which is strong generally except antero-externally. There is a large external anteriorly-crested cusp, and a small internal rounded lobe which is variably basined. P3 is larger than P2, with a large antero-external cusp and a smaller antero-internal cusp. The tooth is set obliquely in the jaw as in P2. The antero-external cusp is crested anteriorly and posteriorly. There is a distinct parastyle which is connected to the protocone by a sharp crest. A cingulum is absent round the protocone but 70 UPPER EOCENE MAMMALIA elsewhere it is prominent (although very reduced in B.M. 30346a (9)). P4is smaller than M1 but larger and more quadrate than P%, and set more squarely in the tooth row. The paracone is slightly larger than the metacone; these cusps are sharply crested anteriorly and posteriorly. The protocone and metacone are subequal. As in P3, a sharp crest runs from the protocone to a distinct parastyle. The cingulum is generally well developed but absent anteriorly. M?! is roughly quadrate in outline and larger than P4. Externally and anteriorly, the tooth is almost identical with P4, except for a small paraconule midway along the protocone-parastyle crest and a stronger cingulum round the protocone base. There is a hypocone sited slightly internal to the protocone and opposite the metacone; this cusp, the smallest of the major cusps, is joined to the posterior cingulum by a low curved crest. M2 is larger than M1, but otherwise structurally nearly identical. M® is similar but slightly narrower antero-posteriorly than M2. The M? internal TABLE 13 Selected measurements (in mm) of Adapis magnus P3 Pa M, Me M3 a-p trs a-p trs a-p trs a-p trs a-p trs B.M. 30323 Se EG BE OH OFS B.M. 30856 TO— 5:4 Cay B.M. 29864 — = 6:4 4°8 Felt 5:1 — — ae iad B.M. M10260 6:3 4:0 7p 4°4 7:0 4°8 7:6 533 — — B.M. M10269 —— — 6:8 4:8 6:9 5°7 7:2 6:1 — = B.M. M10267 (3) = = = — 67 5:0 73 54 oe = D.M.S.W. M556 (41) — — aes 5°6 72 5-0) 8:0) 5-6 -- — B.M. M26057 (1) — — 76 533 ae —_ B.M. M26057 (2) 7I 8646 == = sess = B.M. M25093 — = os —_ — — = = 8-7 4:2 S.M. C54221 = a = = = = Q-7 5-7 = = S.M. C53483 —- -—- |S Fe Oe aD 452 S.M. C53477 ee aS ips p4 M1 M2 M3 a-p trs a-p trs a-p trs a-p trs a-p trs B.M. 3572 4°7 5°8 —— — 6:7 FE, Ff) tI G20) 18°8 B.M. M1o261 (1) — — = = == ae 69 ©6883 Se el B.M. Mroz261 (2) — = = — = =) 0m 53 nas 2s B.M. M20560 (1) — = = = oe, a= 75 83 eX = B.M. M20560 (2) —— — 62 7°5 == = — —_ ee = B.M. M12564 (1) — == —_ == = aa 63 SI = oly B.M. M12564 (2) — — 6-1 7:2 = —_ = = =—_ = S.M. C30678 — Gs ae is ae — $.M. C30674 — — — = 75 9:7 — = —_ = S.M. C30679 = = — = 6:9 8-3 el _ = il S.M. C30681 — = = —_ == —_ 770-88 S.M. C53479 = = = = = = = — =O) ny HEADON BEDS OF ENGLAND 71 cingulum in B.M. 30346a (2), is very rudimentary. In some specimens, the antero- internal molar cingulum is continuous with a crest from the hypocone; in other specimens thisisnotso. This feature is also variable in M1—M of the same individual, as in B.M. 3572. Remarks. The species has long been recognised from the Lower Headon Beds of Hordle Cliff. Adapis magnus (sensu lato) is subject to considerable variation and several distinct varieties have been recognised (Stehlin, 1912). The species is distinguished from other species as follows: A. sciwreus has shorter rounded molars; the protoconid-metaconid crest is only slightly oblique and there is no metastylid; P3 is more simple and Py less molarised with a smaller talonid; the upper molars have a strong metaconule and protoconule and a stronger hypocone. A. priscus possesses rounded less elongated molars; the protoconid-metaconid crest is only slightly oblique and there is no metastylid. In A. ruetimeyert the rounded molars lack a metastylid; the protoconid-metaconid crest is only slightly oblique and there is a strong hypocone. A. farisiensis is much smaller with the lower molars relatively more elongated antero-posteriorly; the upper molars are less extended transversely; P? is extended antero-posteriorly with a smaller internal cingulum and talonid, and the upper canine is smaller and lacks grooves. Adapis parisiensis (Blainville, 1841) (Pl. 2, figs 2, 3) Dracnosis. One third smaller than A. magnus, with molars relatively more elongated ; metastylid variably developed. Upper canine small, generally lacking a grooved crown; P? elongated antero-posteriorly with small talonid and internal cingulum. Hototyre. A skull, according to Hill (1953), now in the Paris Museum. Late Eocene (Late Ludian); France. Described and figured, Blainville, 1841, p. 112, pl. 9 (Genus Anoplotherium). MATERIAL. F.C.S. 7. (1) Fragment of left mandibular ramus with M;-M3. Msg is damaged. (2) Fragment of right mandibular ramus with P2—Mj, Ms, and alveoli of the canine, P; and Mg. Part of the symphysis is preserved. (3) Right M2-talonid only. (4) Left P2 (?). (5) Left P4 with roots embedded in a maxillary fragment. (6) Left M1 (?)—internal fragment with protocone only. All specimens from the same individual. Teeth moderately worn. Upper Headon Beds (Lignite Bed), Headon Hill. Description. The mandible is fairly long. The posterior mental foramen lies below the posterior root of P4; the larger anterior foramen is sited between the canine and P,. A small intermediate foramen lies below the anterior root of Pg. The symphysis is long and oblique. 72 UPPER EOCENE MAMMALIA The canine alveolus indicates a semiprocumbent tooth. P, (from the alveolus) was small and single-rooted. Ps: is an oval double-rooted tooth with a large anteriorly crested cusp and a small postero-internally projecting talonid. The talonid bears a strong antero-posterior crest which posteriorly is raised into a small cusp. There isa fairly strong internal cingulum and a weak postero-external cingulum. Pg3 is double-rooted and similar to P:, except for a longer talonid and a larger talonid cusp. In P4, the protoconid and metaconid appear to be subequal but this is obscured by wear on the one specimen. The metaconid is markedly posterior to the protoconid, with a distinct postero-internal metastylid. A straight anterior pro- toconid crest joins a distinct median paraconid. There is a strong antero-internal shelf-like cingulum and a smaller continuous antero-external, and postero-external cingulum. The hypoconid is joined to the smaller more posteriorly sited entoconid by a curved posterior crest, and joined to the posterior trigonid wall by a strong crista obliqua. M, is almost identical with Py, except that the entoconid and hypoconid are subequal, and there is a small median hypoconulid slightly posterior to the entoconid. M2 differs from Mj only by its larger size. Mg is very similar but broader and more elongated antero-posteriorly than the other molars; the metastylid is very well developed, being higher when unworn, than the protoconid. The talonid is almost the same as in A. magnus: the moderately developed hypoconid is joined to the posterior trigonid wall by an oblique crest ; the smaller entoconid is slightly posterior to the hypoconid; the hypoconulid, the largest talonid cusp is extended posteriorly as a prominent rounded median separated lobe. P! is ovoid with a large external cusp which is crested anteriorly and posteriorly ; internally the tooth is basined. There is a strong cingulum except externally where it is vestigial. P4 is slightly elongated transversely. Of the major cusps, the pro- tocone appears to be the largest and the metacone smallest, with the paracone intermediate in size. The paracone and metacone are crested anteriorly and posteriorly. Astrong broad crest runs from the protocone to a distinct parastyle. A strong external and posterior cingulum is present. Antero-internally there is a short cingulum; this does not extend round the protocone base. Postero-internally there is little sign of a hypocone differentiated from the cingulum. The M? internal fragment shows a strong cingulum which is continuous round the protocone base and a small distinct hypocone. Remarks. Distinguished from other species as follows: A. sciwreus has rounded shorter lower molars; the protoconid-metaconid crest is only slightly oblique and there is no metastylid; P3 is more simple; P, is less molarised with a smaller talonid, and the upper molars have a strong metaconule and protoconule, with a stronger hypocone. A. friscus possesses rounded less elongated lower molars ; the protoconid- metaconid crest is slightly oblique and there is no metastylid. In A. ruetimeyert, the rounded lower molars lack a metastylid; the protoconid-metaconid crest is slightly oblique and the upper molar hypocone is strong. The presence of Pj, the strongly oblique protoconid-metaconid crest and a prominent metastylid, support reference of this specimen to A. magnus or A. parisiensis, However, A. magnus is much larger HEADON BEDS OF ENGLAND 73 and has relatively less elongated lower molars and a slightly more transverse protoconid-metaconid crest. There is considerable variation in the skull, posterior region of the mandible and molars of A. parisiensis. Filhol (1883) recognised six varieties which were reduced to three by Stehlin (1912 : 1235). F.C.S. 7 does not agree closely with any of Stehlin’s figured specimens, differing in the greater development of the metastylid and in mandible shape. The material is perhaps best considered as a variety of A. parisiensis ; it is undoubtedly specifically distinct from A. magnus from the Lower Headon Beds. TABLE 14 Measurements (in mm) of Adapis parisiensis Pe P3 P4 Mi Me M3 Gijo) gs} Elf) mS) GIO) aS} IO) tas) Giejo) ams Elejo) tags BCS. 7 (2) Geir gic OO 7eR oy ACS) iBeC:S: 7) (2) AiO Mn; OMNES 2 een yc ON 4: ON -ONNA- 2 ey —— ——— OcON 4S Family TARSIIDAE Gray, 1825 DraGnosis. Zygomatic arches slender. Brain case much expanded transversely. Bullae much inflated; ectotympanic tubular, extends out of bulla. Foramen magnum directed largely downwards. Very short narrow muzzle; palate narrow anteriorly. Tooth rows converge (V-shaped). Upper canine smaller than anterior incisor. Lower canine larger than upper. Hypocone often small. (after Simons, 1967). REMARKS. Following Simons (1961), the group is divided into two subfamilies: Tarsiinae Gray, 1825. Recent; East Indies—Philippines. Microchoerinae Lydekker, 1887. Middle-Late Eocene; Europe. On the basis of relationship with Tarsius, Simons (1961 : 53) placed the Microchoerinae (previously grouped with the Anaptomorphidae) in the Tarsiidae. Following McKenna (1967) and Russell, Louis & Savage (1967), Microchoerinae Lydekker, 1887 is adopted in preference to the previously more widely used Necrolemurinae Simpson, 1940. Subfamily MICROCHOERINAE Lydekker, 1887 Diacnosis. Skull with comparatively large anteriorly-directed orbits. Snout reduced in size with anterior convergence of the tooth rows. Typically the dental formula is ¢, i i, 3, with total loss of lower incisors except in one species. Lower canine enlarged and P; small and vestigial. Upper canine small and premolariform, the upper enlarged tooth being the most anterior incisor, here considered as I?._ P4 not elevated, having a short heel and tending towards a molariform condition. 74 UPPER EOCENE MAMMALIA P.—P, low and oval in shape. In M2-Ms, paraconid fusing with metaconid, the talonid basined and subovate. Upper molars quadrate with large hypocone usually developed on border ofa postero-internal basin. Metaconule and protoconule developed and in more advanced forms, numerous accessory cuspules. (compiled from Simpson, 1940, and Simons, 1961). REMARKS. Distinguished from the other subfamily as follows: The Tarsiinae has upper molars with a primitive tritubercular crown, a rudimentary hypocone and a reduced metaconule. Five genera are currently referred to the Microchoerinae: Microchoerus Wood, 1884. Late Eocene; Britain, France, Switzerland, Spain. Necrolemur Filhol, 1873. Middle-Upper Eocene; France, Germany. Nannopithex Stehlin, 1916. Middle Eocene; Switzerland. Pseudoloris Stehlin, 1916. Late Eocene; France, Spain. Agerina Crusafont, 1967. Middle Eocene; Spain. Microchoerus and Necrolemur have long been placed in this subfamily. Recent work by Hiirzeler (1948) and Simons (1961 : 61) has added Nannopithex, previously only tentatively assigned to the group. Simons has also shown (1961 : 54) that Pseudoloris, formerly placed by Simpson (1940 : 198) in the Anaptomorphidae as the single member of a separate subfamily, is closely related to the three former genera and should be placed, therefore, in the Microchoerinae. The genus Periconodon from the Middle Eocene of Switzerland, was tentatively assigned to the group by Hill (1955 : 290), but Simons (1962 : 30) referred it to the Omomyidae; more recently, this genus has been placed tentatively in the Adapidae by Russell, Louis & Savage (1967). Crusafont (1967) departed from the generally accepted classification by referring Pseuwdoloris to the Omomyidae. Genus MICROCHOERUS Wood, 1844 Diacnosis. Upper molars with hypocone, mesostyle and numerous accessory cuspules ; Mg witha prominent wide rounded posterior lobe, separated from hypoconid and entoconid by small constrictions; small though distinct paraconid on Mj, not distinct on Mp and M3. Extensive enamel crenulations on lower molars. TYPE SPECIES. Mucrochoerus erinaceus Wood, 1844. Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff. REMARKS. Distinguished from other genera as follows: Pseudoloris is much smaller with sharp cusps; the molars lack enamel crenulation and accessory cuspules; upper molars lack a mesostyle; the lower molars have a distinct paraconid and the M3 hypoconulid lobe is narrow, sharp and well separated. Nannopithex is smaller with only slight enamel crenulation; no mesostyle is developed and the M3 hypo- conulid lobe is narrower and more distinctly separated. Necrolemur has only slight enamel crenulation in the molars and lacks a mesostyle, and the M3 hypoconulid lobe is less developed. HEADON BEDS OF ENGLAND 75 Three species have been described: M. erinaceus Wood, 1844, p. 350. Late Eocene (Early Ludian); Britain, France, Spain. M. edwardsi (Filhol, 1880), p. 124. Late Eocene (Late Ludian); Britain, France. M. ornatus Stehlin, 1916, p. 1377, pl. 22, fig. 12. Late Eocene (Late Ludian); Switzerland, Spain. M. erinaceus and M. edwardsi are known from parts of the maxillae and fragmentary mandibles; M. ovnatus is based on a partial upper dentition. Crusafont (1967) considered that Microchoerus edwardsi was synonymous with M. evinaceus. In the same work, Crusafont (1967) referred Microchoerus erinaceus to Necrolemur and considered Necrolemur antiquus Filhol as a synonym of ‘Necrolemuyr’ evinaceus ; Microchoerus (sensu stricto) was retained for the species M. ornatus. Microchoerus erinaceus Wood, 1844 (Pls3eePl A) Dracnosis. Dental formula é i ;, ;. The largest species. Apparently lacks lower incisors. Upper molars only moderately complex, lacking duplicated paraconule. P3 with rounded outline and a single large external cusp. HototyPe. B.M. 25229 (1). Fragment of left mandibular ramus with Py—Ms3. (2) Palatal fragments with an almost complete dentition, lacking only the left canine and right I? (latter tooth reconstructed in plaster). Fairly badly worn. Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff. Listed and described, Wood, 1844 : 350; Figured, Wood, 1846 : 12, figs 1 and 3; Described and figured, Lydekker, 1885d : 529, fig. (unnumbered); Figured (upper), Lydekker, 1887 : 304, fig. 45; Described and figured, Cooper, 1910 : 40, pl. 3, fig. 1; Described and figured, Simons, 1961 : 59, fig. 2. MaTERIAL. All specimens from the Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff, except where stated otherwise. B.M. 10768. Fragment of right mandibular ramus with Mj—-M3. Fairly badly worn; crowns damaged. B.M. M12563 (x). Fragment of left mandibular ramus with Ps-Mg. (2). Fragment of right mandibular ramus with M\—Ms3. (3). Isolated Mg, probably from (1) above. B.M. 30346. Fragment of left mandibular ramus with Mj—Mg. B.M. 30346b (x). Fragment of left mandibular ramus with Mz and broken Mj. (2). Fragment of right mandibular ramus with M3. (3). Fragment of left mandibular ramus with Py—-M3. (4). Fragment of right maxilla with M1-M3, (5). Right M2. (6). Right P4. Mentioned by Lydekker, 1887, p. 304. (4) above, listed Cooper, Igto : 40. 70 UPPER EOCENE MAMMALIA B.M. M2605b. Isolated My, Mz and M3, from same individual. S.M. C9672. Left mandibular ramus with M;—M3. Posterior region almost complete. Listed and figured, Cooper, Igro : 39; pl. 3, fig. 6. S.M. Cg681. Left mandibular ramus with P3—M3 and the alveoli of the canine, P,; and Pg. The symphysis is preserved. S.M. Cg670. Left mandibular ramus with P2-M3. Listed and figured, Cooper, IgI0 : 40; pl. 3, fig. 4. S.M. C9671. Right mandibular ramus with P3—M3 and roots of the canine, Py and Py. Symphysis preserved. Listed, Cooper, Ig10 : 39. According to Cooper, associated with S.M. C9670. S.M. C9673. Fragmentary right mandibular ramus with Mj—M2, damaged M3 and roots of Py. Listed, Cooper, 1910 : 40. S.M. C9675. Fragmentary right mandibular ramus with Py—Mg. S.M. C9678. Right M3. S.M. C9679. Right Mz. Probably from same individual as S.M. C9678. S.M. C9680. Fragment of right mandibular ramus with Py and M;. Probably from same individual as S.M. C9678. S.M. C54174 (1). Right P3. Unworn, slightly damaged. (2). Left M,;. Talonid only. Almost unworn. Lower Headon Beds (Rodent Bed), Hordle Cliff. S.M. C53482. Fragment of left mandibular ramus with roots of Pa, Mj, and alveoli of the canine and P;—P3. B.M. M26055. Fragment of left mandibular ramus with Ps-Ms. P3 damaged. Teeth badly worn. Upper Headon Beds (Microchoerus Bed), Headon Hill. F.C.S. 3. Left Ms. Upper Headon Beds (Lignite Bed), Headon Hill. B.M. M12563 (1). Right M2. (2). Left M1. S.M. Cg669. Palatal fragment with right and left M?-M%. Listed, Cooper, I9IO, p. 39. S.M. C30710-27. Teeth mounted on a board: S.M. C30712. Left M2. S.M. C30714. Left M?. S.M. C30715. Left M!. S.M. C30716. Right Mj. S.M. C30717. Right Mp. S.M C30718. Left M}. S.M. C30719. Right P4. SAM. €30720., cette. S.M. C30723. Left Mj. S.M. C30724. Right Mg. S.M. C30727. Right upper canine. B.M M12565a =(x). Right Mj. (2). Right My; (or Mg). (3). Right M1. HEADON BEDS OF ENGLAND 77 (4). Right I? (?). (5). Right upper canine (?). B.M. M12565c (1). Left upper canine (?). (2). Right P§. Partially embedded in matrix. B.M. M25090 (i) ettilien(@): (2). Left I3 (?). Both moderately worn. S.M. C54175 (x). Left P§. Moderately worn. (2). Left upper canine (?). Lower Headon Beds (Rodent Bed), Hordle Cliff. DEscrIPTION. The mandible is fairly short and stout; the oval-shaped symphysis (S.M. Cg681, S.M. Cg671, and S.M. Cg670) is steeply inclined, following the line of the canine root. Below P3, the lower edge of the mandible projects internally as a small shelf. The smaller posterior mental foramen lies below Py, (S.M. C9671) and S.M. Cg68r), or slightly anterior to this (S.M. C9670 and S.M. C53482). The position of the larger anterior mental foramen varies from below P3 (S.M. Cg671 and S.M. C9681) to below Pz (S.M. C9670). The posterior region is preserved only in S.M. C9672 (wax restoration of the type excepted). The moderately developed rounded angular process falls considerably below the lower anterior margin of the mandible. The mandibular foramen lies just below alveolar level. The coronoid is small, the posterior margin rather damaged on the specimen. The slender condyle has an articulating surface which is slightly elongated transversely; it lies considerably above the tooth row level. The coronoid crest is more prominent than the condyloid crest, although both are not well developed. The tooth row is compact and continuous. In Ms the protoconid and metaconid are subequal and larger than the hypoconid and entoconid. A small crest which unites the protoconid and metaconid follows the rounded anterior tooth margin; there is no distinct cuspate paraconid. A crest from the hypoconid runs antero- internally towards the metaconid. The posterior region of the talonid is separated into a prominent rounded hypoconulid lobe which is slightly displaced internally; this lobe greatly increases the length of the talonid. In B.M. 10768, the hypoconulid lobe is considerably enlarged relative to the other specimens. There is no distinct cuspate hypoconulid but the lobe bears an internal and an external accessory cusp; these accessory cusps are rather indistinct in some specimens especially after wear. There is a fairly strong external cingulum (only weakly developed in S.M. C9672). Several specimens show slight traces of an internal cingulum. M2 is more quadrate than Mg but otherwise very similar without the hypoconulid lobe. No paraconid is distinguishable as in M3; the protoconid, metaconid and hypoconid are subequal in size, the entoconid smaller. There is a small bilobed median hypoconulid on the hypoconid-entoconid crest (indistinct on S.M. Cg68r). In most specimens, there is an accessory cuspule (mesoconid) on the prominent crista obliqua. A strong external cingulum is present (weak in S.M. Cg672) but there is no trace of an internal cingulum. M;, is similar structurally to Me, with a small duplicated hypoconulid on most specimens (not distinguishable on S.M. C9671) anda mesoconid. There is no internal cingulum. The tooth differs from Mz and Mg by having a more squared anterior 78 UPPER EOCENE MAMMALIA margin and a small distinct paraconid situated directly anterior to the metaconid; the paraconid is separated from the metaconid but united to the protoconid by a low crest. P,is smaller and subtriangular with a smalllow heel. The tooth is dominated by a large median protoconid. A curved antero-basally directed crest from the protoconid forms the anterior tooth margin; the posterior face of the protoconid bears two crests. The smaller separated metaconid in S.M. Cg680, S.M. C9670, S.M. Cg671 and S.M. C9675 is more posteriorly placed than the protoconid and falls to a strong antero- internal cingulum. In the holotype (B.M. 25229) and S.M. C9681, the protoconid and metaconid are opposite and the internal cingulum is only weakly developed. S.M. C9681 and S.M. Cg670 have a well developed external cingulum; this is poorly developed in S.M. C9680, S.M. Cg674 and S.M. Co68r. P3 (well preserved in S.M. Cg681, S.M. C9671 and S.M. C9670) bears a large median anteriorly crested cusp and a small cingulum-like heel; there is a strong internal cingulum. In S.M. Cq67zr and S.M. Co9670 the anterior crest runs parallel to the length of the tooth; the main cusp slopes moderately basally towards the heel, bounded by two sharp crests; an external cingulum is moderately developed. In S.M. Cg681, the anterior crest curves round towards the internal cingulum; the main cusp is more posteriorly situated, falling more steeply to the heel and the two posterior crests are less prominent. Pz (S.M. C9670 and B.M. M26055) is made up of an anteriorly-crested cusp and a small cingulum-like heel. The cusp is concave internally and convex externally. There is a prominent internal cingulum and a smaller external cingulum. Pj, represented only by an alveolus (S.M. Cg681) and a broken root (S.M. C9671) was evidently a small anteriorly inclined tooth, situated on the external edge of the tooth row. The broken root (S.M. Cg671) and a large oval deep alveolus (S.M. Cg681) of the lower canine indicate a large anteriorly-inclined tooth, lying at a much lower level in the jaw than the other teeth. I2-P4 are associated only in the holotype. The most anterior tooth I?, is enlarged with almost vertical sides. I? is considerably smaller with an antero-internally sloping crown and a small external cingulum. The semi-procumbent canine is much enlarged (but not as high as I?) and has an external cingulum. P? is single-rooted and semi-procumbent with a small heel; the tooth is smaller than the canine. P3 and P4 are quadrate and larger than P2; both have a single large external cusp and a strong external cingulum. P4 has a small antero-internal and a small postero- internal cusp which are sharp and unworn in S.M. C30719 and S.M. C30720; these are worn down to an antero-posterior ridge in the holotype and B.M. 30346b (6). M! and M2 are quadrate and subequal in size; the metacone lies on a straight antero-posterior line with the paracone. In the smaller M°, with a more rounded posterior margin, the metacone is more internal to the paracone. The hypocone, which in M1! is almost equal to the protocone, becomes progressively smaller in M2 and M3. A paraconule and metaconule are well developed in all molars, the former sited anterior of a line between protocone and paracone, the latter anterior of a line between metacone and hypocone. The metaconule, equal to the paraconule in M}, becomes progressively smaller in M2 and M8. In all molars, a small accessory cusp HEADON BEDS OF ENGLAND 79 is situated midway between the metaconule and the protocone; this can be regarded as a duplication of the metaconule. In M! and M®?, this accessory cusp is smaller than the main metaconule, but in M it is larger. There is a prominent mesostyle in M! and M? external to a line between the paracone and metacone. This cusp is represented in M3 by a slight cingular swelling. A small but definite hypostyle occurs on M! and M? of B.M. 30346b (4). A parastyle and metastyle are encountered more frequently, appearing as upgrowths from the prominent external cingulum: a parastyle occurs on M? of S.M. C30712 and S.M. C30714 and M! of B.M. 30346b (4), B.M. M12563 (2), S.M. C30715 and S.M. C30718; the metastyle is present on M2? of the holotype (B.M. 25229), B.M. 30346b (4), S.M. C30712 and S.M. C30714 and the M! of B.M. 25229 and B.M. 30346b (4). In all other specimens, these three accessory cusps are indistinct. Remarks. The number of specimens is really inadequate for any satisfactory statistical treatment of variation within the species to be carried out. However, some tentative conclusions may be drawn from the available material. The following variates were considered: length and width of Ms, Me, M; and Py; length of tooth rows M;—Ms and P,—M3 and the internal depth of the mandible below the posterior TABLE 15 Selected measurements (in mm) of Micvochoerus erinaceus Mi- Pe P3 Py, Mi Me M3 M3 aap ciSua-p LES! a-p) tts) ap) trs) a-p) tks! a=p trs ‘a=p B.M. 25229 (1) 3 Onee2 Cra 3 + ON——0(3'-4) 4-3 Ol LTO) B.M. 10768 —- — — — — — (2:5) (2:9) 3°83 3:2 4:5 27 — B.M. M12563 (1) —- — — — 3:0 26 41 3:9 42 39 — — — B.M. M12563 (2) 1-23 AC SSO) AO 3-2 B.M. 30346 ee tee oo ety ALO 3240 AO, Brg B.M. M26055 312 rn 27S 2 IASON 13-0) -4cAy 2-8) —— S.M. C9670 Ones 2g ee 397 340 8-7 34) (45 93°C TTS S.M. C9671 ae ao SS 9 85 87 3540 4-0 3 :P Io S.M. C9681 a — eS 2A 3-4 2-0) 42 3°50 3°Q)) 3-4) 4-2 2:5) 12-0 S.M. C9672 At -OmmrA-ONNnS "5 4- On 3 - Onl 225 S.M. C9673 —- — — ~— —~— — 4:0 3:2 40 3:2 — — — S.M. C9675 BoP eS Oma SMES OM AGNES) a Se p2 P3 p4 M1 M2 M3 B.M. 25229 (2) aan 220 Ooh Soh 4rON 74.0) 93:00 5:0) 3°01) 3°90 B.M. 25229 (2) ZOE ZAR 27S 35) SiO) AZ) 40) VAT 3:9» <4°9. 3°2. -Ar0 B.M. 30346b (4) ai Gaia aa 5On 40 4:9) 2:8) 470 B.M. 30346b (5) 1 a. B.M. 30346b (6) Sg I re ee B.M. M12563 (1) —- — — 39 51 — — B.M. Mr2563 (2) 4:2) 4-0 = — = S.M. C9669 SS SS 39 (48) 31 42 S.M. C9669 —- —- —- —- —- — — 3-5 SOS 4:2 S.M. C30715 = ee SS a 7 S.M. C30719 —- —- — — — 38 — —- —- ~—- —- — S.M. C30720 ee 80 UPPER EOCENE MAMMALIA margins of Ms, M; and P3. The mandible depth in M. erinaceus appears to be highly variable and cannot be correlated with individual age as reflected by dental wear. Frequency distributions utilising dimensions of the posterior lower teeth indicate a separation of the larger specimen B.M. 303,46; this is removed by secondary grouping of the data and is, in all probability, due only to incomplete sampling. P3_P4 of the type specimen (B.M. 25229 (1)) and S.M. C9681 are different structurally from the rest of the material. These differences might be sexual or could possibly indicate an anterior milk dentition; they are unlikely to be of specific importance. Simons (1961) observed an alveolus anterior to the lower canine in some French specimens of M. edwards, indicating at least one lower incisor. There is no evidence for similar teeth in the specimens of M. evinaceus from the Headon Beds. But preservation of this region of the mandible is rare and their absence in M. erinaceus should be treated with caution. Apart from the probable structural difference in dentition, M. evinaceus differs from M. edwardsi by its greater size. Several features of the upper dentition distinguish the species from M. ornatus (known only from upper dentition): P? in M. evinaceus has a rounded outline and a single major external cusp whilst this tooth in M. oyvnatus has a squared outline and twoexternalcusps. Inaddition to the duplicated metaconule present in both species, the M2? of M. ornatus is further complicated by a duplicated paraconule and a prominent hypostyle. MM. erinaceus is significantly larger than M. ornatus: in the holotype of MW. evinaceus, the length of the tooth row, M2-M$ is 11-0 mm, whereas in the holotype of M. ornatus, this measurement is 9:0 mm. Genus PSEUDOLORIS Stehlin, 1916 Diacnosis. Small size. No accessory cuspules and enamel crenulations. Upper molars relatively elongated transversely, lacking mesostyles; cingulum- hypocone distinct on M! and M2, indistinct on M3; lower molar cusps relatively sharp and separated. Mg entoconid and hypoconulid high and separated with distinct narrow hypoconulid-lobe. Distinct small paraconid on My and Mg, reduced on M3. Type species. Pseudoloris parvulus (Filhol, 1890). ? Late Eocene (Quercy Phosphorites) ; France. REMARKS. Thegenus was proposed by Stehlin (1916 : 1397) for a species originally described by Filhol, (1890a) as Necrolemur parvulus. The validity of this genus has been subsequently accepted generally. Simpson (1940: 198) agreeing with Teilhard de Chardin (1916—1921a) that Pseudoloris was significantly Tarsius-like, erected a new subfamily, Pseudolorisinae, for this monospecific genus. Recently Simons (1961 : 54) has shown that the similarities between Psewdoloris and other microchoerines are so great, that separation at subfamily level is not possible. Weigelt (1933) described P. abderhaldeni from the Middle Eocene brown coal deposits of Geiseltal, Germany. Simons (1961 : 61) stated that the type of this species is conspecific with ‘“Necrolemur’ vaabi Heller, the latter species being referable to Nannopithex. Crusafont (1967) described a new species, Pseudoloris reguanti from the Upper Ludian of Spain. HEADON BEDS OF ENGLAND 81 Pseudoloris is distinguished from other genera as follows: Microchoerus has a mesostyle and is markedly larger with lower rounded cusps; the M3 hypoconulid- lobe is wide and rounded and the molars possess numerous accessory cuspules and enamel crenulations. In Necrolemur the M3 hypoconulid-lobe is wide and rounded; the cusps are lower, rounded and less distinct and the genus is much larger. Nanno- pithex has slight enamel crenulations in all teeth; P4 is much larger than the lower molars; the molar cusps are more rounded and lower, and Mg entoconid and hypo- conulid are not as high and separated. Pseudoloris parvulus (Filhol, 1890) (Text-fig. 23) 1890 Necrolemur pavvulus Filhol. Diacnosis. As for genus. HoLotyrPe. Fragment of left mandibular ramus with P3—Ms3 and alveoli of the canine, P;-Py. ?Late Eocene (?Ludian, Quercy Phosphorites); Caylux, France. Described and figured, Filhol (1890a : 39, text-fig., unnumbered). Specimen not located by Teilhard de Chardin (1916—-192I1a: 5). Abundant referred material described and figured by Stehlin (1916 : 1397) and Teilhard de Chardin (1916- 1g21a : 4). MATERIAL. B.M. 36812f. Fragment of right mandibular ramus with M2—Ms3. Teeth slightly worn. Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff. DESCRIPTION. Mg and Mg are elongated antero-posteriorly. The trigonid and basined talonid are well differentiated, but the cusps are low, the trigonid cusps being only slightly higher than those of the talonid. In Mg, the trigonid and talonid are subequal in width but the trigonid 1s compressed antero-posteriorly and is shorter. The protoconid is slightly higher than the metaconid and markedly more anteriorly placed; both cusps are crested transversely. The metaconid is internal in position; the protoconid tip lies just external to the longitudinal midline. There is a well developed antero-external cingulum. A small median cuspate paraconid forms the sharp anterior trigonid apex; this cusp is joined by a small low crest from the metaconid and from the protoconid. The talonid is strongly basined. The entoconid is slightly smaller than the metaconid; a sharp crest runs anteriorly from the entoconid to meet, at the base of the posterior trigonid wall, a sharp metaconid crest. Although the tooth is damaged externally, it appears that the hypoconid was opposite, or just anterior to the entoconid, a strong crista obliqua running from this cusp to the base of the posterior protoconid wall. A rudimentary hypoconulid is represented as a very small projection on the posterior talonid margin, just external to the midline; this is joined to the hypoconid and entoconid by small crests. Owing to damage, the extent posteriorly of the cingulum cannot be observed; it may have been continuous round the base of the hypoconid as in M3. F 82 UPPER EOCENE MAMMALIA The anterior part of Mgis very similar to that of Mg, differing only in minor features: the protoconid is slightly more externally placed, and is subequal with the metaconid; the posterior trigonid margin is less oblique than in Mg and the paraconid is slightly less developed and more crestiform in appearance, forming a more rounded anterior trigonid apex. The anterior part of the talonid is also very similar. Mg differs markedly from Mz by having a hypoconulid-lobe. This long prominent separated lobe is narrow and mainly median although the posterior extremity is slightly internally directed. The lobe, consisting of a single large crestiform cusp, is con- nected with the hypoconid and entoconid by crests. A prominent continuous antero- external, external and postero-external cingulum is present. REMARKS. Recently Pseudoloris parvulus has been recorded by Crusafont from the Early Ludian of Sosis, Spain (1965, 1967) and Euzet-les-Bains, France (1967). The single specimen from the Headon Beds appears to agree precisely with additional material of Pseudoloris parvulus figured by Stehlin (1916), Teilhard de Chardin (1916—1g921a) and Piveteau (1957); it constitutes the first record of the genus from the Early Ludian of Britain. Fic. 23. Pseudoloris parvulus (Filhol, 1890). Fragmentary right mandibular ramus with Me-M3. B.M. 36812f x12. Lower Headon Beds, Hordle Cliff. (a) Occlusal view. (sp) External view. (c) Internal view. HEADON BEDS OF ENGLAND 83 TABLE 16 Measurements (in mm) of Pseudolovis parvulus M2 M3 a-p trs a-p trs B.M. 36812f 16 1:25 19 Te. Order CREODONTA Family HYAENODONTIDAE Leidy, 1869 Diacnosis. Py double-rooted except in some specialized genera; P3 lacking distinct protocone; Mz or Mz functioning as main carnassials; M3 present or absent; cheek teeth row moderate to long; Mg when present transversely extended. Jaw long and shallow to moderately short and deep, with symphysis weak to fairly robust, and angle long, slender, projecting and upcurved at tip. Face long to moderately short; basicranial region narrow to moderately wide; no preglenoid crest (mainly from Denison, 1938 : 117). Remarks. The order Creodonta is retained here in a restricted sense, following Romer, (1966 : 229). The family is known from the Early Eocene-Middle Oligocene of North America; Early Eocene-Late Oligocene of Europe; Late Eocene-Oligocene, Early Pliocene of Asia and the Early Oligocene, Early Miocene of Africa. Denison (1938) distinguished four subfamilies: Proviverrinae, Hyaenodontinae, Limnocyoninae and Machaeroidinae; this was followed by Simpson, (1945). Gazin (1946) proposed a new family, Limnocyonidae, to accommodate the latter two subfamilies. Van Valen (1966) placed all four subfamilies in the Hyaenodontidae but reduced them to the rank of tribes; Gazin’s two fold subdivision was retained as two subfamilies, the Hyaenodontinae and the Limnocyoninae. More recently, Van Valen (1967) has abandoned the use of tribes within the Hyaenodontinae, grouping together the Hyaenodontini and the Proviverrini. In the present work, the earlier scheme of Denison is utilized as it provides a simple working arrangement (see also Savage, 1965). Subfamily HYAENODONTINAE Trouessart, 1885 Diacnosis. Long faced, narrow skulled genera; main carnassials M3 with carnassial specialization very advanced; M3 present, M8 variably present or absent; molars sectorial, length greater than width; M3, where present, small and transverse ; M! and M? with paracone and metacone completely or nearly connate and protocone reduced or absent ; lower molars lack metaconid, talonid vestigial or absent. Body compact, robust; limbs and feet cursorial (after Matthew, 1909 and Savage, 1965). 84 UPPER EOCENE MAMMALIA Remarks. Distinguished from the other subfamilies as follows: in the Proviver- rinae, the metaconid is retained; the carnassial specialization is less advanced and M3 are invariably present. The Limnocyoninae and Machaeroidinae both have M3 functioning as the main carnassials.- Savage (1965) listed five good genera in the Hyaenodontinae: Hyaenodon Laizer & Parieu, 1838. Late Eocene-Late Oligocene; Europe. Late Eocene-Middle Oligocene; North America, Asia. Early Oligocene, Early Miocene; Africa. Pterodon Blainville, 1839. Late Eocene-Early Oligocene; Europe. Late Eocene- Early Oligocene; North America. Early Oligocene, Early Miocene; Africa. Late Eocene; Asia. Apterodon Fischer, 1881. Early-Middle Oligocene; Europe. Early Oligocene; Africa. Metapterodon Stromer, 1926. Early Miocene; Africa. Early Oligocene; Europe. Leakitherium Savage, 1965. Early Miocene; Africa. Simpson (1945) listed Hemipsalodon Cope, 1885; Dasyurodon Andreae, 1887 and Propterodon Martin, 1906. Savage (1965) has considered that Hemipsalodon is a synonym of Pterodon and Dasyurodon a synonym of Apterodon. More recently, Van Valen (1966 and 1967) has asserted the generic validity of Hemipsalodon and has stated (1967) that Metapterodon is a synonym of Pterodon ; he also adds Ischnognathus Stovall, 1948 (Van Valen, 1966 and 1967) and (tentatively) Hyainatlouros Bieder- mann, 1863 (Van Valen, 1967). Pvopterodon is not well known (see Van Valen, 1966 : 75) and Savage has stated (personal communication) that this latter genus will probably prove to be invalid. According to Van Valen (1967), Megalopterodon Dashzeveg, 1964, is synonymous (provisionally) with Pterodon. Apterodon was referred by Van Valen (1966) to the Mesonychidae; the genus was returned to the Hyaenodontidae by Van Valen (1967) following the work of Szalay (1967). Genus HYAENODON Laizer & Parieu, 1838 Diacnosis. Dental formula 3, s i ; ; M? with shallow groove on completely connate paracone and metacone; molars without protocone; metastyle strongly elongated, especially on M2. Main carnassial pair Mj, with M3 as accessory carnas- sials. Ms without talonid, M; +2 with or without vestigial talonid (Savage, 1965). TYPE SPECIES. Hyaenodonleptorhynchus Laizer & Parieu, 1838. Middle Oligocene (Chattian) ; France. REMARKS. Distinguished from other genera as follows: in Pterodon, M? is present and a small protocone and large parastyle occur on M1! and M?. In Apterodon, a protocone is present on M1 and M2, and Mj, possesses a prominent three-cusped talonid. M2? is present in Metapterodon and M1 and M? possess a strong protocone. Leakitherium has a strong protocone on P4—M2?. HEADON BEDS OF ENGLAND 85 A complete list of species attributed to Hyaenodon has been given by Savage (1965). Some 36 species have been described: 9 are from North America, 3 from Asia, 20 from Europe and 4 from Africa, the latter being grouped by Savage as a distinct subgenus. Specific distinctions are based considerably on size differences. Numerous structural features have also been utilized (listed in Savage, 1965), but have nearly all proved highly variable. Many of the early European types are lost and the existing figures and descriptions are frequently quite inadequate. Of the described species, many are undoubted synonyms; several true genera probably are involved. In the absence of a greatly needed thorough revision of the genus, some attempt is made below to clarify the systematics of the Headon Beds specimens. Hyaenodon c.f. minor Gervais, 1848-52 (Pl. 5) 1926 Hyaenodon aimi Cooper: 370, fig. B. Diacnosis. M3 with talonid trace usually present but variably developed. Mj—-Mz2 usually with distinct cuspule at base of antero-internal paraconid face. P; with a single root. HototyrPe. Right mandibular ramus with P2—-M3 and alveoli of the canine and P;. Molars badly worn. Late Eocene (Early Ludian); France. Described and figured, Gervais, 1848-52 : 129; pl. 25, fig. 9. Specimen not located. MateriaL. All specimens from Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff, except where stated otherwise. B.M. 29752. Complete mandible attached at symphysis. With right canine, P.—M3, broken root of P; and alveolus of I3; and left Ps—M3, broken root of canine and alveolus of Pj. Teeth very badly worn—often to root level. The posterior region of both ramiis complete. Labelled in British Museum (N.H.) as “Hyaenodon hantonensis’. Listed by Lydekker, 1884 : 444, 1885a : 25. S.M. C31347 a-c, d, k. Probably all from same individual. Fine greenish-white sand matrix. Teeth well worn. a. Right mandibular ramus with canine, Py—-Ms3, and alveoli of Pj-P3. M1 is badly damaged. Posterior region preserved. Figured, Cooper, 1926 : 372, fig. A. b. Left mandibular ramus with canine, Pz, Py-M3. Ps3 represented by broken roots. c. Fragment of right maxilla with P4, damaged fragment of M!, alveoli of P% and M2. Anterior part of the zygomatic arch preserved. d. Fragment of left maxilla with P4-M2 and roots of P3. k. Isolated left upper canine. Tip damaged. S.M. C31344. Fragment of left mandibular ramus with Py-Mz. Moderately worn. B.M. 29814. Left M3. Badly worn. Listed by Lydekker, 1885a : 26. (erroneously as right Ms). 86 UPPER EOCENE MAMMALIA S.M. C31346. Left Ps. Moderately worn. S.M. C30775. Fragment of left maxilla with I8, P3—-P4, alveoli of the canine, M1—M2 and roots of P! and P?. Teeth damaged. Internal surface embedded in fine greenish sand matrix. 7 B.M. M25073. Fragment of left maxilla with P1-P3, posterior alveoli of P4 and part of the canine alveolus. B.M. M25074. Fragment of right mandibular ramus with Ps—M3. Moderately worn. B.M. M13075. Fragmentary right and left mandibular rami, not attached at symphysis. Same individual. Adhering matrix of shelly orange sand. Holotype of ‘Hyaenodon aims’. Described and figured Cooper, 1926 : 370, fig. B. Right ramus with Iz and the canine, Po-M3 and alveoli of Iz and P,. Tip of canine missing. Left ramus with the canine, P2—M3 and alveolus of P}. Tip of canine missing. B.M. M25075. Fragment of left maxilla with P2-M?2. Unworn. Upper Headon Beds (Lignite Bed), Headon Hill. DESCRIPTION. The mandible is moderately elongated. The posterior mental foramen lies under the posterior root of P3, the anterior mental foramen at a higher level, under the root of Py. The symphysial surface is inclined at 45° to the tooth tow and ends posteriorly under Pz. The posterior mandibular region is well pre- served in B.M. 29752. The prominent transversely elongated condyle, situated below the level of the tooth row, is separated from a small narrow angular process by astrong notch. The angular projects downwards just below the level of the anterior mandibular margin. The coronoid is moderately well-developed with a slightly notched posterior margin. The mandibular foramen is sited markedly posteriorly, just below condyle level. The coronoid and condyloid crests are prominent. Iz is represented in the holotype of ‘Hyaenodon aimi’ (B.M. M13075) by a small alveolus. Is is a small semiprocumbent tooth immediately external to the I, alveolus and pressed up close against the canine. 1; may have been present but no evidence remains. Cooper (1926 : 371) stated that the incisors were not preserved. The lower canine is large and sharp with an ovoid section. In the right ramus of S.M. C31347, this tooth is probably somewhat displaced from its alveolus. P, is represented by the alveolus; this single-rooted tooth was slightly separated from the canine. In S.M. C31347b, the left ramus anomalously lacks a P, alveolus and there is a larger gap between P2 and thecanine. Contrary to the statement by Cooper (1926 : 371, footnote) P; was originally present in the right ramus of B.M. 29752, asa root-filled alveolus is visible. Pz, is double-rooted, posteriorly-directed and immediately posterior to P;; it bears a large central cusp which is postero-externally crested. There is usually a continuous internal and postero-internal cingulum, but in B.M. M13075, it is small and antero-internal in position. Pg is similar to Pp: except that the tooth is longer antero-posteriorly and has a small posterior longi- tudinally crested cusp; there is no distinct posterior cingulum in B.M. M13075. Pa is similar to Ps but larger transversely and strongly inclined posteriorly. The cingulum is usually prominent antero-internally and postero-internally, although HEADON BEDS OF ENGLAND 87 indistinct in B.M. M13075. Where the anterior crest of the main cusp merges with the antero-internal cingulum there is an irregular cuspule (S.M. C31347 and B.M. M25074) or a distinct cusp (S.M. C31344, B.M. M13075 and B.M. 29752). Mj is very small. The protoconid is slightly larger than the paraconid when un- worn; these crestiform cusps are separated by a cleft. A small low talonid bears an elongated cusp. The antero-internal paraconid face (S.M. C31344) bears a small projecting cuspule which is closely pressed against the postero-internal margin of P,; this cuspule is indistinct in B.M. M13075. On the antero-external paraconid face there is a vertical buttress pressed up against the postero-external margin of P4. The cingulum is vestigial. In B.M. 29752, only worn rounded stumps of this tooth remain. Mg is similar to Mj, but of larger size. The crested paraconid and protoconid, well separated by a cleft, form good shearing facets. The antero-internal cuspule and antero-external buttress are usually well developed; the cuspule is higher than in My and is not in contact with the adjacent anterior tooth. The cuspule is indistinct in B.M. M13075. The posterior talonid cusp is prominent. In B.M. 25074, this tooth is structurally similar but very small, being about the same length as Mj. M3 is markedly elongated in length; the crested paraconid is separated from the larger protoconid by a deep cleft; these cusps together form an effective slightly oblique shearing blade. Postero-internally, there is a small talonid rudiment in B.M. M13075 and S.M. C31347 (more strongly developed in B.M. 29752). Antero- externally, a buttress is developed as in the other molars. A trace of the antero- internal cuspule situated high on the antero-internal paraconid face is present in the least worn specimens. Anterior to the orbits, the maxillaries have a great height, as is seen in S.M. C30775. The infra-orbital foramen is narrow and slit-like, opening above P? (B.M. M25073 and S.M. C30775). S.M. C31347c and d indicate a fairly narrow palatal region. I3 is a very small pointed tooth, slightly separated from the canine alveolus (S.M. C30775). The upper canine (S.M. C31347k, S.M. C30775 and B.M. M25073) is a large curved tooth with an ovoid section and a long vertically-striated root. Pl (B.M. M25073) is double-rooted, elongated antero-posteriorly and situated immediate- ly posterior to the canine; the tooth has a main cusp over the anterior root, and a strong continuous internal cingulum. P2? is larger than P!, but otherwise similar; it is separated from P! by a gap. P3 is separated from P2 by a gap. This tooth is slightly larger than P? (B.M. Mz25073) and has two roots. The main cusp is crested antero-internally and posteriorly. There is a distinct crested posterior cusp, separated from the main cusp by a cleft, and a prominent internal cingulum. P4 (S.M. C31347c and d, and S.M. C30775) has three roots and lies immediately posterior to P3. There is a strong protocone lobe but no distinct cusp; the transverse width is greater than the antero-posterior length. The large main cusp is separated from a small though distinct posterior cusp by a cleft; there is a small anterior cusp (more developed in B.M. M13075 and S.M. C30775 than in S.M. C31347c and d). M1 is much smaller than P4 and situated immediately posterior to it; the tooth is 88 UPPER EOCENE MAMMALIA triangular-shaped, the antero-posterior length being about the same as the transverse width. The crested anterior cusp (fused metacone and paracone) and a crested metastyle form an effective shearing blade; the cusps are separated internally by a small cleft and externally by a shallow valley. There is a small indistinct parastyle in S.M. C31347d; this is better developed in B.M. M13075. M? is very similar structurally but much larger and more trenchant. The crested metastyle is tapered and elongated posteriorly ; with the crested metacone/paracone, this forms internally an elongated almost vertical shear plane. REMARKS. With the exception of B.M. M13075, most of the material listed above, for many years has been referred to the Late Eocene species, Hyaenodon minor. The length of the tooth row and the single-rooted P; appear to exclude reference to any of the other European species of Late Eocene age (listed in Savage, 1965 : 280); with the present chaotic state in the systematics of the genus, comparison of the Headon specimens with Hyaenodon minor seems the most reasonable course. Cooper (1926) described a new species, Hyaenodon aimi, based on a well-preserved mandible from the Lower Headon Beds of Hordle Cliff. In the present work, the holotype (and only specimen) of ‘Hyaenodon aimi’ is not considered specifically distinct from the other Headon Beds material. The length of the lower tooth row in the holotype of ‘Hyaenodon aimi’ and in the specimens of Hyaenodon c.f. minor is almost identical (see table of measurements) ; dimensions of individual teeth in the two species overlap considerably. The only difference in dentition worthy of mention is the presence of a distinct cuspule antero-internal to the paraconid on My; and Mg in Hyaenodon c.f. minor, this cuspule is indistinct in the holotype of “Hyaenodon aim’, but its presence or absence is unlikely to be of any significance. Cooper (1926) noted the closely similar length of the lower tooth row in ‘Hyaenodon aima’ and the other Headon Beds specimens, but cited as differences the smaller more slender mandible, the closed tooth row and the curved lower margin of the mandible in ‘Hyaenodon aimi.’ In particular, Cooper emphasised the differences in mandible depth below Mg and considered the possibility of this variation being caused by differences in the sex of individuals; he concluded, by comparison with living carnivores, that the discrepancy in dimensions was too great to be accounted for in this way. It is highly probable that these differences in mandible shape and size are due to variations in individual age. The holotype of ‘Hyaenodon aimi’ is of a young individual with only a slightly worn dentition. Nearly all of the remaining Headon Beds specimens of lower dentitions exhibit greater degrees of wear, indicating more mature individuals; in these animals, increasing age has been accompanied by many changes such as a deepening of the mandible, flattening of the lower margin of the mandible and spacing of the premolars. B.M. M25074 is rather smaller than most specimens, with a curiously small Mp. The specimen may be specifically distinct, but as little is known about intraspecific size variation in Hyaenodon species, it is best grouped at present with the other material. In general aspect, the teeth in B.M. M13075 and B.M. M25075 are slightly more slender and delicate; these specimens may represent females of the species. HEADON BEDS OF ENGLAND 89 TABLE 17 Measurements (in mm) of Hyaenodon c.f. minor P2 P3 Pa My Me M3 Ps3-Ms3 Gisj) wes} Glejo) wes) A) Ww) ely) eas lee) ws} EI) B.M. 29814 — — I6I 70 — B.M. 29752 94 46 12:2 5:0 (Ci) ESS (TET IS B.M. 29752 Ae eee ee OA let 4) a (248)! 1(672)115575 S.M. C31347a —- — — — 12:2 63 (7:99) — I04 5:0 15:0 66 — S.M. C31347b 96 47 — — 128 61 7:6 4:2 10:5 4:8 14:5 6:3 (57-0) S.M. C31344 I22 61 7:6 4:0 10:0 5:2 — — — S.M. C31346 — — 163 5:0 — —~— —~ ~ — — —_—- — — B.M. M25074 Sa ae ce neg 2. ante 7-3 F4rS) LE QM" 597. 45%5 B.M. M13075 OP FO uOG Fee wu OF why Foe) aso GA avies) Fite) Foro) B.M. M13075 WOO BO wine Bet wits xo kyl fle! aetsy ES} auitey (Os GSH) P2 p3 p4 M1 M2 P4-M2 Gi=jo) GS) lef) aes) I=) ass BI=jo) was} jo) was} IR99) B.M. M25073 76 36 ITO 5:2 124 63 — — — —~— — S.M. C31347¢ —- — — ~~ ~~ — 113 107 — ~~ — S.M. C31347d —- — — — Ito t0:2 9:8 7:3 (14:0) 9°3 (32°5) S.M. C30775 pie aetna arn aS) 0.94105, (0:7) = = B.M. M25075 1 OMe ACO Gu hs >2 eI li-O) LO 7-O)) L745 O-O) Aly Order CARNIVORA Family MIACIDAE Cope, 1880 Dracnosis. P4, M; carnassials as in modern carnivores. Post-carnassial teeth more or less tubercular. Metacone smaller than protocone (Matthew, 19009). REMARKS. The family is recorded from the Middle Palaeocene-Late Eocene of North America; Late Eocene of Europe and Late Eocene—?Early Oligocene of Asia. The group has been traditionally divided into two subfamilies (Simpson, 1945 and Piveteau, 1961): Viverravinae Matthew, 1909 and Miacinae Trouessert, 1885. Subfamily VIVERRAVINAE Matthew, 1909 Dracnosis. Molars 5. Antero-external cusp of P4 and posterior accessory cusp of Pyprominent. Elongated oval outline of M2. Mabsent. Skull much elongated. (Partly from Matthew, 1900). Remarks. Distinguished from the other subfamily as follows: in the Miacinae the skull is short or moderately elongated; molars 5 and the antero-external cusp of P4 and the posterior accessory cusp of P4 rudimentary or absent. 90 UPPER EOCENE MAMMALIA The Viverravinae contains five genera: Didymictis Cope, 1875. Middle Palaeocene-Early Eocene; North America. Ictidopappus Simpson, 1935. Early-Middle Palaeocene; North America. Protictis Matthew, 1937. (incl. Simpsonictis MacIntyre, 1962.). Middle (and Late?) Palaeocene; North America. Viverravus Marsh, 1872. Early-Middle Eocene; North America. ?Early Oligo- cene; Asia. Quercygale Kretzoi, 1945. Late Eocene; Europe. In 1962, MacIntyre described a new genus Simpsonictis. More recently, MacIntyre (1965 : 3544) has separated Protictis Matthew, 1937, from Didymictis as a distinct genus: Protictis was considered to include Simpsonictis MacIntyre, 1962. Genus QUERCYGALE Kretzoi, 1945 1965 Humbertia Beaumont: 142. DraGnosis. Viverravinae of medium to large size. P4 with or without parastyle, deuterocone strong and projected. M1 with strong parastyle, paracone robust and flattened externally, metacone smaller and more conical, no true metaconule but a paraconule present. Protocone near the anterior border bearing two crests. No hypocone. M? much smaller than M! and very elongated transversely. Lower premolars have a posterior cuspule low and trenchant. My, high with a metaconid higher than the paraconid. Talonid short with hypoconid trenchant and dominant. Mz small with trigonid complete and talonid reduced and trenchant. Symphysis high and short, coronoid process more or less reduced (modified from Beaumont, 1965). TYPE SPECIES. Quercygale angustidens (Filhol, 1872). ?Early Ludian (Quercy Phosphorites) ; France. REMARKS. Teilhard de Chardin (1914-15) first recognised the viverravine affinities of Viverra angustidens Filhol, placing it in the American genus Viverravus. This European material was separated by Kretzoi (1945) as a new genus Quercygale. Beaumont (1965) placed ‘Viverravus’ angustidens into a new genus Humbertia and described a new species Humbertia helvetica (Ritimeyer, 1862). Humbertia Beau- mont, 1965, is synonymous with Quercygale Kretzoi, 1945 (Beaumont, 1966 : 282). Quercygale differs from the other genera as follows: in Viverravus, the Mg trigonid is high and the talonid is long with a distinct hypoconulid; M1 has a hypocone (usually), a paraconule is absent and the P‘4 parastyle is invariably present. Ictidopappus has upper teeth with a pronounced transverse elongation; the talonid of Py is short with a very small cusp and the lower molar talonids are basined. In Protictis, the My; and Mg trigonid is higher and the talonid is larger and wider trans- versely; Pz and P, have a well separated anterior cusp. As observed by Beaumont (1965), Quercygale appears closest to Didymictis but in this latter genus, the My and Mz talonid is basined and the P, parastyle is invariably present. Beaumont (1965) observed that apart from the absence of Mg, Quercygale is very similar to Tapocyon Stock, 1934, a miacine from the American Late Eocene. Van Valen (1967) has considered that this presence or absence of Ms is inadequate for generic distinction, HEADON BEDS OF ENGLAND gl and that Quercygale is best considered as a subgenus of Tapocyon. Following this, Van Valen has concluded (1967 : 255) that . . . ‘the Miacinae can no longer be unambiguously distinguished from the Viverravinae on the basis of the presence of three molars rather than two’. Three species are recognised here: Q. angustidens (Filhol, 1872), pl. 17, figs 33-35. Late Eocene (Bartonian-Early Ludian) ; France. Q. helvetica (Riitimeyer, 1862), p. 86, pl. 5, fig. 86. Middle Eocene (Lutetian) ; Switzerland. Q. hastingsiae (Davies, 1884), p. 433, pl. 15, figs 1-4. Late Eocene (Early Ludian) ; Britain. Beaumont (1966) noted that Miacis? macintyri Van Valen, 1965, from the Lutetian of Germany, is very similar to Quercygale helvetica. Reference of Miacis? macintyri to Quercygale was supported by Van Valen (1967). The species is not well known (based on an isolated P4 and M!) and may well be synonymous with Quercygale helvetica. Depéret (1917) described a new species Procynodictis euzetensis (later in the same work described as Viverravus euzetensis) from the Early Ludian of France. This species was considered by Gauthier & Favre (1948) to fall within the range of variation of ‘Viverravus angustidens’. This latter view was followed by Beaumont (1965 : 143) who recorded Procynodictis euzetensis as a synonym of Humbertia angustidens. Although the holotype of Procynodictis euzetensis has not been examined and Depéret’s illustrations (1917, text fig. 2; pl. 24, figs 3-6) make precise comparison difficult, Procynodictis euzetensis is synonymised with Humbertia angustidens in the present work, on the authority of Gauthier & Favre, and Beaumont. Quercygale hastingsiae (Davies, 1884) (Pl. 6) 1884 Viverva hastingsiae Davies: 433, pl. 15, figs. 1-4. 1885 Viverva hastingsiae Davies; Lydekker: too, figs 11-12. 1915 Vivervavus angustidens (Filhol); Teilhard de Chardin (partim): 117. 1917 Vivervavus hastingsiae (Davies); Depéret: 248. 1965 Humbertia angustidens (Filhol); Beaumont (partim): 143. Diacnosis. About the same size as Q. angustidens. My trigonid somewhat wider and larger than talonid, with fairly blunt anterior apex; talonid larger than that of _ Q. angustidens with squared outline. Outline of upper teeth rounded; internal lobe of P8 large; posterior margin of M! with strong embayment. Hototyre. B.M. 30203. Skull and mandible in occlusion. Badly crushed and damaged. The outline of the skull is preserved in fine greenish sand matrix, but only the lower part is distinct and relatively undamaged. The right and left halves of the skull have been displaced by crushing along the palatal sutures; the occluded mandibular rami are also similarly displaced relative to each other. To 92 UPPER EOCENE MAMMALIA facilitate examination, the posterior part of the right ramus has been removed (in the past) from its original occluded position. The removed right ramus shows P3—Mg, the exposed right maxilla and premaxilla, I-C-M2. The upper and lower teeth anterior to P3 are damaged, displaced and partly obscured by matrix. Described and figured, Davies, 1884 : 433, pl. 15, figs 1-4. Lydekker, 1885a : 100, figs 11-12. From Lower Headon Beds (Mammal/Crocodile Bed), Hordle Cliff. The only known specimen from the Headon Beds. DESCRIPTION. Pg is rectangular in outline and elongated antero-posteriorly, with a slight anterior narrowing. The anterior margin is squared, the posterior margin more rounded. There is a large crested central cusp and a prominent talonid which bears a small cusp. A slight postero-external and antero-internal cingulum is present. Py, is generally similar to P3 but much larger, and the anterior narrowing is more prominent. The talonid bears an antero-posteriorly aligned crestiform cusp and is basined internally. The largest tooth is Mj; this has a stout bulbous trigonid which is much wider and longer than the talonid. The protoconid, the largest cusp, is opposite the metaconid ; both cusps are anteriorly and transversely crested. The paraconid, the smallest trigonid cusp, is crestiform with its internal edge strongly internal in position. There is a short but wide antero-external cingulum. The small talonid is structurally similar to P4; the internal region is slightly basined and there is an antero-posteriorly aligned crestiform cusp. Mg is very short antero-posteriorly. The trigonid is almost as wide transversely as the talonid of M;, with very low cusps; the protoconid is largest and slightly anterior to the metaconid. A low curved crest running from the internal paraconid to the protoconid, forms the squared anterior tooth margin. The antero-external part of this crest is badly worn; it may represent the site of a small cusp. The talonid, small and narrowing posteriorly, is well worn and shows no evidence of any cusps. The upper teeth anterior to and including P2 have been much displaced from their original positions. One upper incisor is preserved; it is small and simple. The upper canine is large and curved witha strong root. P1is embedded in matrix; it is smaller than the incisor with a single central cusp. The anterior region of P? is missing. The tooth has a central major cusp but is larger and more elongated antero-posteriorly than P!. P% is strongly elongated antero-posteriorly and very much larger than P2. The tooth narrows anteriorly and is dominated by a large central crested cusp. Posteriorly there is a low rounded separated cusp and a postero-external shelf-like cingulum. P# is triangular-shaped in outline and much larger than P38. The tooth is dominated by a large central crested paracone; antero-externally this falls as a crest towards a low rounded parastyle. There is a low rounded protocone which is larger than the parastyle. The anterior margin between parastyle and protocone is strongly embayed. The crestiform metacone is separated from the paracone by a deep cleft which continues externally as a narrow valley; these cusps together form a sharp oblique postero-internal shearing edge. There is a small distinct antero-external cingulum and a faint postero-external cingulum. HEADON BEDS OF ENGLAND 93 Ml is the same width transversely as P4. The metacone is directly posterior to the larger paracone; a posterior metacone crest curves antero-externally along the external tooth margin to the parastyle. The crested protocone is crescentic in shape, an anterior crest running antero-externally towards the paracone; a low paraconule is developed on this crest just antero-internal to the paracone. The parastylar area extends externally as a wing which is separated from the postero- external area by a smallembayment. The parastyle is crestiform; it is joined to an external paracone crest. There is a distinct internal and antero-internal cingulum. M2 is much smaller than M! and a little smaller than P?. The ovoid tooth lies obliquely in the tooth row. The antero-external margin is in contact with M! just posterior to the M1 metacone. The crown of the tooth is badly damaged and no details can be discerned. RemMARKS. Teilhard de Chardin (1915 : 117, footnote 1) considered that ‘Viver- yavus’ hastingsiae was synonymous with ‘Viverravus’ angustidens. Depéret (1917 : 248) recorded several differences, on the basis of which he believed the two species to be distinct. Gauthier & Favre (1948 : 116) emphasised the larger Mj talonid of ‘Viverravus’ hastingsiae but were uncertain whether this merited specific separation. More recently Beaumont (1965 : 143) on the evidence of the more elongated Mj talonid, considered ‘Humbertia’ hastingsiae as a distinct subspecies of ‘Humbertia’ angustidens. Depéret’s view is followed here. In addition to the considerably more elongated talonid of M;, the P? of Quercygale hastingsiae (from comparison with illustrations of Q. angustidens) has a smaller internal heel as Depéret observed. Although this single specimen from the Headon Beds may be aberrant, it seems reasonable at present to regard it as specifically distinct. TABLE 18 Measurements (in mm) of Quercygale hastingsiae P3 P4 Mi M2 a-p trs a-p trs a-p trs a-p trs B.M. 30203 7:0 2:8 Q:1 317) 9°7 6:2 51 3:6 Pps Pp M1 M2 a-p trs a-p trs a-p trs a-p trs B.M. 30203 7:8 — I13 — 6:7 — — —- IV. CORRELATION AND AGE OF THE HEADON BEDS Although there is general agreement concerning the broad correlation of Upper Eocene and Lower Oligocene deposits in Europe, no unanimity has been reached yet regarding the detailed correlation of marine and non-marine sediments in the different depositional basins. One particular problem is the exact siting of the Eocene/ Oligocene boundary. The situation is hindered by a confused stage and substage 94 UPPER EOCENE MAMMALIA terminology, which despite many recent attempts at clarification by authors, has still to be resolved. It is not intended here to present a critical review of this problem, detailed aspects of which have been considered recently by Rey (1964), Krutzsch & Lotsch (1964), Blondeau, Cavelier, Feugueur & Pomerol (1965), Thaler (1966), Denizot (1968), Cavelier (1968), Curry, Gulinck & Pomerol (1969) and others. The complete mammalian fauna of the Headon Beds has not yet been studied in detail, and authoritative use of the mammals in correlation and dating is clearly premature. Nevertheless, in view of the importance of the Headon Beds in current stratigraphical research, some interim conclusions are presented here. The exact age of the Headon Beds and indeed of most of Forbes’ ‘Fluvio-marine series’ has long been controversial. The uppermost beds of the British succession, the Hamstead Beds, are generally accepted to be Oligocene in age. No unanimity has been reached regarding the age of the underlying Bembridge, Osborne and Headon Beds. Three schemes have been proposed in the past for the age of the Headon Beds; two schemes are supported currently by European stratigraphers and palaeontologists, but the third, that proposed by Reid & Strahan (1889) is now obsolete. Reid & Strahan (1889) and Arkell (1947) considered that the Lower, Middle and Upper Headon Beds were Oligocene (Lattorfian of Arkell) and the underlying Barton Beds were uppermost Eocene. The Eocene/Oligocene boundary lay at the base of the Lower Headon Beds. This view, in which is emphasised the general change from marine to brackish and freshwater conditions at the beginning of Lower Headon times, has not found further support. In a currently held view, supported traditionally by most British workers, the Lower Headon Beds are considered as uppermost Eocene (Bartonian) and the Middle and Upper Headon Beds are Oligocene (Lattorfian, Tongrian). The Eocene/ Oligocene boundary is placed, therefore, at the base of the Middle Headon Beds. This view, based on correlation by mollusca and foraminiferida of the Middle Headon Brockenhurst Bed with the type Lattorfian (Lower Oligocene) of Germany, has been adopted by Wrigley & Davies (1937), Bhatia (1955), Denizot (1957, 1968), Curry (1958a, 1966), Chandler (1g61a, 1963), Rey (1964), Franzen (1968) and others. In the second currently supported arrangement, the Lower, Middle and Upper Headon Beds are placed together with the Barton Beds in the Upper Eocene (Lower Ludian). This was adopted by Stehlin (1910) and followed more recently by Krutzsch & Lotsch (1964), Blondeau, Cavelier, Feugueur & Pomerol (1965), Thaler (1966) and Cavelier (1968). In this scheme, the Eocene/Oligocene junction is placed at some position above the Upper Headon Beds. Stehlin (1910) on the basis of what he believed to be a fundamental break in the mammal faunas (‘grande coupure’), considered that the Eocene/Oligocene junction in Britain should be placed at the base of the Hamstead Beds; this has been followed more recently by Krutzsch & Lotsch (1964), Blondeau et al (1965) and Cavelier (1968). Thaler (1966) preferred to place the boundary at the base of the Bembridge Beds. Stehlin (1910) compared the mammalian fauna of St.-Hippolyte-de-Caton (=Euzet- les-Bains) with the Lower Headon Beds, considering them Lower Ludian in age. HEADON BEDS OF ENGLAND 95 The same faunas were compared by Depéret (1917) who also assigned them a Lower Ludian age. Stehlin (1910) also compared the mammalian fauna of Montmartre with a similar fauna from the Bembridge Limestone; he considered these faunas to be Upper Ludianinage. According to Stehlin (1910) the fauna from the Bembridge Marls was similar to that from the Bembridge Limestone. He observed that the Upper Ludian (Bembridge Beds) fauna differed from that of the Lower Ludian (Lower Headon Beds) fauna by the extinction of several species and the arrival of several forms unknown in the Lower Ludian. Stehlin’s correlations of the Lower Ludian faunas of Euzet-les-Bains with the Lower Headon Beds, and the Upper Ludian faunas of Montmartre with the Bembridge Limestone, have now become well established. Recently, Thaler (1965, 1966) proposed to replace the old classification by a series of seven zones defined by theridomyids. Although some of Thaler’s zones transgress the traditional stratigraphical subdivisions, the Euzet zone, Montmartre zone and Ronzon zone correspond exactly with the Lower Ludian, Upper Ludian and ‘Sannoisian’ respectively. Stehlin (1910) noted that as the faunas of the beds above the Lower Headon Beds and below the Bembridge Limestone were poorly known, it was not possible in Britain to determine the precise junction of the Lower and Upper Ludian faunas. More recently, Krutzsch & Lotsch (1964) grouped the Upper Headon Beds and Osborne Beds in the Montmartre zone with the Bembridge Beds; for them, the junction of Lower and Upper Ludian faunas lay at the base of the Middle Headon Beds. The mammalian fauna obtained recently from the Upper Headon Beds appears to be similar to that from the Lower Headon Beds and apparently distinct from the Bembridge fauna. While the sample from the Upper Headon Beds is clearly still incomplete, the Upper Headon Beds, on present evidence, should be grouped, at least provisionally with the Lower Headon Beds in the Upper Eocene (Lower Ludian). The Upper Headon fauna should be placed in the Euzet zone, rather than the Montmartre zone as suggested by Krutzsch & Lotsch (1964). It follows that, on the basis of the mammalian faunas, the traditional British practice of siting the Eocene/Oligocene boundary at the base of the Middle Headon Beds, cannot be sustained. This present view is in agreement with Martini (1970) whose study of calcareous nannoplanckton from the Brockenhurst Bed, has indicated that the Brockenhurst Bed is uppermost Eocene in age, and not equivalent to the type Lattorfian (Lower Oligocene). The Eocene/Oligocene boundary therefore, must be situated at some position above the Upper Headon Beds. Owing to the poorly known mammalian fauna, the exact stratigraphical position of the overlying Osborne Beds is uncertain at present. However, study of this fauna, together with the fauna of the Bembridge Beds, currently in progress by Mr A. Insole at the University of Bristol, may clarify this situation. 96 UPPER EOCENE MAMMALIA V. REFERENCES ARKELL, W. J. 1947. The geology of the country around Weymouth, Swanage, Corfe and Lulworth. Mem. geol. Suvv. U.K., London, 386 pp. BraumontT, G. DE. 1965. Les Viverravinae (Carnivora, Miacidae) de l’Eocéne de la Suisse. Bull. Soc. vaud. Sci. nat., Lausanne, 69 : 133-146, pls. 1-2. 1966. Les Miacinae (Carnivora, Miacidae) de l’Eocéne de la Suisse. Bull. Soc. vaud. Sci. nat., Lausanne, 69 : 273-285, pl. r. BuatTiA, S. B. 1955. The foraminiferal fauna of the late Palaeogene sediments of the Isle of Wight, England. J. Paleont., Chicago, 29 (4) : 665-693, pls. 66-67. BraInviL_te, H. M. DE. 1839-64. Ostéogyaphie ou description iconographique comparée du squeletie et du systéme dentaive des mammiferes vécents et fossiles pour servir de base a la Zoologie et a la Géologie. Paris. BLonDEau, A., CAVELIER, C., FEUGUEUR, L. & PoMEROL,C. 1965. Stratigraphie du Paléogéne du Bassin de Paris en relation avec les bassins avoisinants. Bull. Soc. géol. Fy., Paris, (7) 7 : 200-221. Butrer, P. M. 1946. An arctocyonid from the English Ludian. Ann. Mag. nat. Hist., London, (11) 13 : 691-701, 2 figs. Cavetier, C. 1968. L’Eocéne supérieur et la base de l’Oligocéne en Europe occidentale. im Colloque sur l’Eocéne, Paris, 1968. Mém. Bur. Rech. géol. minier., Paris, 58 : 507-527. CHANDLER, M. E. J. 1922. A recent exposure of the ‘Marine Bed’, Hordle, Hants. Geol. Mag., London, 59 : 224-229, 1 fig. 1925. The Upper Eocene flora of Hordle, Hants, Part 1. Palaeontogy. Soc. (Monogy.), London, 1-32, pls. 1-4. 1926. The Upper Eocene flora of Hordle, Hants, Part 2. Palaeontogy. Soc. (Monogy.), London, 33-52, pls. 5-8. —— 1961a. The Lower Tertiary floras of Southern England. 1. Palaeocene floras. London Clay flora (supplement). 354 pp., 34 pls. Brit. Mus. (Nat. Hist.), London. 1961b. Flora of the Lower Headon Beds of Hampshire and the Isle of Wight. Bull Br. Mus. nat. Hist., London, 5 (5) : 91-158, pls. 24-30. 1963. Revision of the Oligocene floras of the Isle of Wight. Bull. Br. Mus. nat. Hist., London, 6 (3) : 321-384, pls. 27-35. CHARLESWORTH, E. 1845. On the discovery, by Mr Searles Wood, of an alligator in the freshwater cliff at Hordwell, associated with extinct mammalia. Rep. Br. Ass. Adumt Sci., London, 14 : 50. Cuatwin, C. P. 1960. The Hampshire Basin and adjoining areas. By. veg. Geol., 3rd ed, 99 pp., 42 figs. London. CuHow, M. 1964. Rhabdocolpus at one horizon in the Lower Greensand [Crackers], requires further consideration and palaeoecological interpretation. Haas, after investigating South American Triassic material, concluded that the relationship of the genera Protofusus, Paracerithiwm and Rhabdocolpus is much closer than that implied by Cossmann’s classification, although as he pointed out (pp. 212 & 293), their exact position can only be discovered from a thorough investigation of all Triassic Procerithiidae. He showed that the morphological changes occurring in these genera, particularly in their ornament, could be interpreted as a series, with Rhabdocolpus the most highly specialized of the three. However, the evidence and determination of the earliest representatives of these genera is confusing and such a lineage remains doubtful. A further complication is the close resemblance, especially of ornament, of Rhabdocolpus to the Triassic genus Andagularia Haas, which is considered to belong to the Purpurinidae. ?Rhabdocolpus ?clementinum (d’Orbigny) (Pl. 5, figs 6, 12 & 13) ?1843 Cerithium clementinum d’Orbigny: 357, pl. 228, figs 1-3. 1845 Cerithium clementinum d’Orbigny; Forbes: 352, pl. 4, fig. 9. MATERIAL, LOCALITY AND HORIZON. BM(NH) GG 5692-5694, GG 20841-20843, GG 20844 (5), Crackers, Lower Greensand; GG 5708 (5 specimens), Lower Lobster Bed, Lower Greensand; Atherfield, Isle of Wight. DescripTion. A very small, turriculate shell with a narrow spire, flat-sided whorls and a grooved linear suture. The whorls are ornamented by 16-20 axial ribs, PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 143 that are orthocline on the early whorls but soon become strongly opisthocline. Generally, five primary spiral cords are visible on the whorl side, with occasionally a secondary cord between the second and third. Small rounded tubercules occur at the intersections of the axial ribs towards the base of the whorl; these tubercules become somewhat rectangular on the abapical cords. The base is convex and has a varying number of spiral cords ornamenting its surface, although there are always a pair of primary cords forming small carinae some distance from the periphery and another pair of strong cords close to the columella. The entire semi-rounded aperture is flat at its base and has a straight inner lip. In complete specimens, the growth lines are visible on the body whorl and these are initially strongly prosocline, but at the second spiral cord swing back to become strongly opisthocline, producing a rather asymmetric opisthocyrt shape. Discussion. Unfortunately, it has not been possible to discover Forbes’ original specimen in the Geological Survey Collection. The various specimens from the same locality listed above, bear a strong resemblance to his figure of a turriculate, gastropod with cancellate ornament. Judging from the specimens in the British Museum (Natural History) collection, it would appear that in his figure, the ribs are too numerous, and that the dimensions of the whorls are incorrectly drawn. The only other figures of C. clementinum are those given by d’Orbigny (1843 : 357, pl. 228, figs 1-3) and these do not show the typically cancellate and tuberculate ornament of the British specimens. D’Orbigny’s figures show smooth cords and ribs that are orthocline, or only opisthocyrt rather than the strongly opisthocline ribs of the British Aptian fossils. In fact, d’Orbigny’s figure could equally as well be interpreted as an attempt to represent the ornament of C. forbesianum d’Orbigny [= C. phillipst Leymerie of Forbes]. Wollemann (1900 : 169) suspected that Forbes’ Aptian fossil was not identical to those described by d’Orbigny from the Neocomian of Marolles (Dept. d’Aube), although he did not explain his doubts. In his description of C. clementinum, he mentioned that the axial ribs of subsequent whorls are irregularly alternate and that the whorls are convex, while he described the aperture as being quadrangular and depressed. Such a description could not be applied to the specimens referred here to Forbes’ 1845 figure. It is not advisable to introduce a new specific name for the British material, until d’Orbigny’s material has been re-examined and his types recognized. Peron (I900 : 192) commented that the ornament of C. clementinum is extremely variable and that the specimen figured in the ‘Paléontologie francaise’ is a very rare form. ? Rhabdocolpus forbesianum d’Orbigny, 1850 (Pl. 5, figs 1-5) 1845 Cevithium phillipsi Leymerie; Forbes: 352, pl. 4, fig. 12. 1850 Cerithium forbesianum d’Orbigny: 116 1854 Cervithium forbesanum d’Orbigny; Pictet & Renevier: 52, pl. 5, fig. 6. 1864 Cerithium forbesianum d’Orb.; Pictet and & Campiche; 286. 144 BRITISH CRETACEOUS GASTROPODS 1868 Cevithium forbesianum d’Orb.; Verneuil & Loriére: 15, pl. 2, figs 7a-e. 1906 Uchauxia forbesiana (d’Orb.); Cossmann: 57 & 252. 1961 Uchauxia forbesiana (d’Orb.); Casey: 607. MATERIAL AND OCCURRENCE. Four specimens in the collections of the Geological Survey Museum have been accepted as the original material in the Geological Society Collection that was described by Forbes (1845) as C. phillipst. Two of the specimens bear the Geological Society number 2267 and the other two are associated with a label numbered. 2275 and all are from the Lower Greensand of Atherfield, Isle of Wight. Blake (1902) lists R 2267 in his catalogue of type material at the Geological Society and, it is true, that a specimen with this number in a small rock fragment, bears ornament similar to that shown in the original figure on Forbes’ plate 4. However, not one of these specimens is exactly like the figure, which is probably a composite representation of the features shown by these incomplete and poorly preserved specimens. This material must, therefore, be accepted as a type series and one of them, GSM Geol. Soc. 2275 (a)—an incomplete specimen but clearly showing the ornament of later whorls,—is selected here as a lectotype. I have used better preserved, topotype material from the Crackers, Lower Green- sand, in the British Museum (Natural History) collections to describe this species; GG 20845-20849, GG 20850 (6), GG 5609, GG 5695-5697, G 20898 and G 20908. Another specimen G 20983 comes from East Shalford in Surrey, an Atherfield Clay locality mentioned by Casey (1960 : 547). DeEscRIpTION. A very small shell (on average 2-4 mm in height), composed of convex whorls that are separated by a grooved linear suture; in the early whorls the suture may appear to undulate owing to the effect of the axial ribs. The ornament of the early whorls is quite different from that of the later whorls. Initially, ornament consists of several spiral cords and 10-14 opisthocline axial ribs, with slight tubercules occurring where the cords cross the ribs. After 3 or 4 whorls the axial ribs disappear but continue to influence the cords, first as the tubercules briefly persist and then by producing slight crenulations and, or, thickenings in the smooth ribbon-like spiral cords. On later whorls, six raised-primary cords are generally visible but some variation in this number occurs. Frequently, the third and fourth cords from the adapical suture are the strongest and these occur at the centre of the whorl. In other examples, the strongest cords occur towards the abapical suture with the weakest on the adapical shelf. The interspaces between the spirals are considerably wider than the width of the cords. Secondary cords develop in these interspaces, but not according to any definite pattern. The rounded base is ornamented by a varying number of spiral cords with a primary cord close to the periphery and two other primary cords close to the columella. Growth lines are only visible on the whorl sides as striae between the cords, but they are easily seen on the base and swing back towards the columella. The semi-rounded aperture is entire, has a flat basal lip, a thin outer lip, straight inner lip and a smooth columella. Discussion. Several authors have followed d’Orbigny and used the name forbesianum to describe this British Aptian fossil, but few have endeavoured to PROCERITHIIDAE, CERITHIIDAE & CERIDHIOPSIDAE 145 distinguish it from Cerithium phillips: Leymerie 1842. D’Orbigny first used the name forbesianum in his Prodrome (1850 : 116) when recording the British material, which he considered Forbes had misidentified, but did not provide any detailed explanation. Pictet and Campiche (1864 : 287) suggested that the slender elongate form of both species may have led to the confusion, but commented that the rare varices that occurred in C. phillips: provided an easy method of distinguishing the two species. These varices are easily seen in d’Orbigny’s figure of the French species (1843 : pl. 227, fig. 10) but are not shown in Leymerie’s own figure (1842 : pl. 17, figs Ira & b) of this more elongate shell. Verneuil and Loriére (1868 : 15) recognized that C. forbestanum was quite different from C. phillipsi and provided the only adequate figure (pl. 2, figs 7a-e) of the species to show its variable ornament, when describing its occurrence at several localities in the Aptian of the Spanish province of Teruel. Cossmann (1906 : 57) obviously accepted, or also observed, the varices in C. phillipsi Leymerie, for he included the species in his new genus Uchauxza in which the possession of varices is a diagnostic feature. Yet, this evidence still needs to be verified by examining type material, and, in order to ensure that phillips: does belong to Uchauxzia, it is also necessary to ascertain that an anterior canal is present for one has not been figured. At the same time, Cossmann (1906) referred C. forbestanum to Uchauxia commenting that it wasclose to phillipsiand quoted Pictet as his source. The only figure produced by Pictet is that in Pictet and Renevier (1854 : pl. 5, fig. 6) which depicts a shell with ornament between that of forbesianum and pluillips: [as understood from the figures quoted above]. However, the presence of an entire aperture in several of the British specimens of forbesianum listed above, shows that this species cannot belong to Uchauxia (as interpreted by Cossmann, 1906). The turreted and flat-sided whorls of C. clementinum d’Orbigny; Forbes, together with their persistent cancellate ornament, easily distinguish this species from the differently ornamented forbesianum with its strong spiral cords and convex whorls. Yet both have several features in common, such as a grooved linear suture, and a semi-rounded entire aperture, while their early whorls can be confused if the ornament is not adequately preserved. The greater convexity of the whorls and their persistent orthocline axial ribs distinguish C. dupinianum d’Orbigny (1843 : 357, pl. 227, figs 4 & 5) a similar species to forbesianum, described from the Neocomian of Marolles (Dept. d’Aube). ? Rhabdocolpus melvillei sp. nov. (Ele Spatie7) Diacnosis. A small procerithiid with reticulate ornament on its early whorls but which loses axial ornament on later whorls, and has a wider spiral angle than usual in the genus. Hototyre. In the Sedgwick Museum, B. 27340. OTHER MATERIAL. Four paratypes in the Sedgwick Museum, Cambridge, B. 27341-27344, all belonging to the Wiltshire Collection. 146 BRITISH CRETACEOUS GASTROPODS LOCALITY AND HORIZON. Crackers, Atherfield Clay Series, Lower Aptian, Forbesi Zone, near Atherfield Point, Isle of Wight. DescripTIon. The shell is very small and has a wide spiral angle. The whorl outline is strongly convex and the suture impressed. The earlier whorls show about twelve opisthocyrt axial ribs. On laterwhorls, threespirals appearand small rounded tubercles are developed where these intersect the axials. Later, a gap develops between the first and second spiral cords and two further cords are produced on either side of the first. The new spirals eventually become as strong as the earlier formed ones. On the body whorl, the spirals are separated by interspaces greater than their width and the axials are separated by interspaces greater than those between the spirals. The rounded base is ornamented by several less-tuberculated spiral cords. Discussion. This species can be distinguished from ? R. forbestanum by its more strongly-rounded whorl outline, by its greater number of axial ribs, slightly stronger development of tubercles and wider spiral angle. The few specimens of ? R. melvilles that are available do not show any indication of secondary spirals between the primaries, nor the marked differentiation in ornament that occurs in forbesianum. However, the full range of variation within forbestanum is not known and it could be possible that ? R. melvillei is, merely, a variety of that species. Family ? PROCERITHIIDAE Genus ? UCHAUXIA Cossmann, 1906 TYPE SPECIES. By original designation Cossmann (1906: 56), Cersthiwm peregrinorsum d’Orbigny 1843, from the Turonian of Uchaux, Vaucluse, France. Discussion. The material described below, undoubtedly belongs to a common genus and because of considerable similarity in form and ornament to species of Uchauxia, they are regarded as belonging to that genus. However, owing to their incomplete preservation this determination must remain provisional. The genus Uchauxia is also subject to uncertainty, for a crucial factor in its recognition is the apertural shape of its type species, U. peregrinorsa. The various figures of that species provided by d’Orbigny (1843: pl. 231, fig. 3), Cossmann (1906 : pl. 6, figs 37-40) and Roman and Mazeran (1920 : pl. 5, figs 42-44) do not provide sufficient reliable evidence of this feature, although each author mentions a rounded aperture with a short anterior canal in his description. In order to verify these features it 1s necessary to refer to the type material. Unfortunately, this has not been possible during this investigation, nor have other specimens of the species been available for comparison. For the moment, following Cossmann (1906 : 57), it is assumed that these small Cretaceous shells are fully grown and that they resemble Procerithium and Rhabdocolpus, but have a distinct, short anterior canal. Uchauxia, as here understood, is generally larger than Cerithiopsis Forbes & Hanley but does have a similar whorl shape and tends to have the same range of ornament. Cerithiopsis, however, usually has a more advanced type of siphonal canal, which PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 147 does not project above the basal plane and resembles a keyhole in basal view. Cossmann (1906 : 57) pointed out that Uchauxia only occurs in strata that are believed to be entirely marine and for this reason, differs from both Potamides Brongniart 1810 and Pirinella Gray, 1847. Uchausia wisei sp. nov. (Pl. 6, figs 5 & 6) Diacnosis. Uchauxia with four prominently beaded spirals, the abapical penultimate spiral being weaker than the others, and weakly developed axial ornament. Hototyre. BM(NH) GG 6161, C. W. Wright collection. OTHER MATERIAL. Eight Paratypes BM(NH) GG 5611, GG 6158-6160 & GG 6162— 6165. C. W. Wright collection. All incompletely preserved. LOCALITY AND HORIZON. Claxby Ironstone, Valanginian-Hauterivian, Nettleton, Lincs. Description. The shell is small, narrow and turriculate, slightly cyrtoconoid, with a flat-sided whorl outline and linear suture. On the earlier whorls, three primary spiral cords are developed with interspaces of equal width. The anterior and posterior spirals become more prominent on later whorls, with the central spiral moving posteriorly as a secondary spiral cord makes its appearance abapically to it. A secondary spiral is also developed behind the posterior primary spiral. On still later whorls, a secondary spiral cord may appear between the two posterior primary spirals. The spirals cross 20 to 24 axial ribs, producing rounded tubercles at each intersection. The axials are separated by interspaces almost equal to their own width and are orthocline on the early whorls but become opisthocyrt. The base is convex and bordered by two spiral cords; the axials do not extend on to it and the surface is almost smooth. The small, oval aperture has a curved anterior canal. The growth lines follow the axial ridges. Dimensions. Holotype BM(NH) GG 6161. Height (approx. 5 whorls). : . . ‘ 10-0 mm Spiral angle ‘ : : Q : : ‘ a Penultimate whorl height. ' ; : : 1-8 mm Penultimate whorl diameter . ; ; ; : 3°8 mm Discussion. The most characteristic feature of this species, when compared with other species of Uchauxia that have four spiral cords, is the weak influence of any axial component. In the Turonian type species, U. peregrinorsa, the axials are quite strong, and similarly, in the Gosau species U. solida (Zekeli, 1852 : pl. 20, fig. 3) and U. distinctum (Zekeli, 1852 : pl. 19, fig. 6), the exact age of which needs to be investigated (see Kuhn, 1947 : 188). U. wiser is distinguished from the two species described below, in that Uchauxia sp. nov. has only three spiral cords of tubercles ornamenting the whorl and that its tubercules have some axial alignment; while the prosocline axial ornament of U. badri is quite diagnostic. 148 BRITISH CRETACEOUS GASTROPODS The specimen described by Roman & Mazeran (1920: pl. 5, fig. 45) as Uchauxia vichet, although having predominantly spiral ormament, has many more cords, which have a lower degree of tuberculation, than U. wisei. Originally, their speci- men was figured by Cossmann in his examples of U. peregrinorsa and it may still be found to come within the range of variation of that species, when this aspect is fully investigated. Although the figures of the type species in both Cossmann (1906) and Roman & Mazeran (1920) are inadequate for exact comparison, it is possible to decide that peregrinorsa (d’Orb). has several more tuberculate spirals than this specimen of U. wisei. The Turonian species described by Roman & Mazeran, U. richei, has three granular cords but possesses many more primary cords and has a slightly different whorl outline. The British Museum (Natural History) collection included a number of Ageria gaultina sp. nov. from the Gault of Folkestone, which at first sight, have some resemblance to this fossil but the existence of a greater number of axial ribs and a differing growth line pattern provide adequate distinguishing characters. Uchauxsia badri sp. nov. (Pl. 6, figs 3 & 4) Diacnosis. A small, slender, cyrtoconoid, turriculate shell that is ornamented by four tuberculate spiral cords, the tubercules of which are aligned along prosocline axial ribs, giving a distinctive twisted appearance to the shell. HototypPe. Sedgwick Museum, Cambridge, B 44632. HORIZON AND LocALiTy. Upper Greensand, Albian, Blackdown, Devon. Description. A small cyrtoconoid shell with a slender turriculate spire. The whorl outline is slightly convex and the shell has a grooved linear suture. The ornament of the earlier whorls is dominated by strong axial ribs which cross three unequally-spaced spiral cords. On subsequent whorls, the number of cords increases to four and eventually five, producing rounded tubercles at their intersections with the axial ribs. The ribs consequently become less dominant and contribute to the cancellate effect of the ornament. The interspaces between the spiral cords are then equal and also equal to the width of the cords but considerably narrower than the spaces between the axial ribs. The prosocline axial ribs are partly offset in relation to one another on subsequent whorls and this gives a twisted look to the shell. On later whorls, these ribs become reduced in strength, although this may, in part, be due to wear and poor preservation. On the penultimate and body whorl, it is also possible to see the fifth spiral cord, which occurs close to the abapical suture. The axial ribs influence this, but although they may cross it, they then weaken towards the suture. The convex base is covered by faint collabral threads and ornamented by five spiral cords, two of which produce small carinae. The small rounded aperture has a partly twisted anterior canal. PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 149 Dimensions. Holotype, Sedgwick Museum, B 44632. Height 10-6 mm Mean spiral angle ; : : : ; : 20° Height of last whorl . : : : : , 3°0mm Maximum diameter. : : ; 4 2-9 mm Discussion. The distinctive prosocline axial ribs, that give the ornament of this species a slightly twisted look, have been mentioned in the discussion of the pre- ceding species. It is this feature that immediately distinguishes U. badvi from the type species U. peregrinorsa despite the figures of the latter being unsuitable for any further detailed comparison. However, it is also possible to observe that the axial tibs of pevegrinorsa are opisthocyrt and that its spiral cords are much thinner (i.e. weaker) than those of badri. One of the Lower Cretaceous species, forbesianum (d’Orbigny), that Cossmann included in this genus, is now thought to belong to Rhabdocolpus Cossmann, 1906. There is still some confusion concerning the characters of U. phillipsi (Leymerie) and as suggested elsewhere in this paper, type material needs to be re-examined in order to verify the validity of Cossmann’s determination. The ornament shown in the few figures of this slender species is quite different from U. bad, but Leymerie’s figure (1842, pl. 17, figs 11a & b) although not showing varices, does show orthocline axial ribbing. Cerithium kaunhowent Cossmann (1902 : 61), from the Maastrichtian of Maastricht, differs in that, although having cancellate ornament, this is much finer and composed of at least 6 untuberculated spiral cords and numerous opisthocyrt axial threads, while the shell also has a definite anterior canal. Cerithium distinctum Zekeli (1852: pl. 19, fig. 6) from Gosau, although having four predominate tuberculate cords between the sutures, also has a thin cord in the centre of each interspace and has strongly orthocline axial ribs. Uchausia sp. nov. (EL 6. fig 7) MATERIAL AND LOCALITY. BM(NH) G 73788 J. S. Gardner Collection, Gault, Folkestone, Kent. * DEscriPTION. Unfortunately, this unique specimen is rather crushed and measurements are not reliable, but it has been possible to discover details of the ornament by developing out its reverse side. This gastropod can be described as a small, more or less subulate, slender shell with convex whorls. Its moderately deep, linear suture lies at the bottom of a concave furrow formed by adjacent whorls. The initial whorls show bicarinate ornament, but a third spiral cord soon appears just behind the posterior of the two initial cords, and others develop later. These spiral cords are separated by wider interspaces and each interspace is occupied by at least one almost smooth spiral thread. The primary cords are tuberculate and the rounded tubercles are separated by intertubercular spaces that are nearly equal to their width. Usually these are of the same number and strength on each cord, but sometimes those on the posterior cord are slightly weaker. On the early whorls, the 150 BRITISH CRETACEOUS GASTROPODS axial ridges are as dominant as the spiral cords, yet, later, with the increase in size of the whorl they appear to become slightly disrupted and if there is any connection at all, only a thin curved ridge joins the tubercules. These ridges tend to follow the opisthocyrt growth lines. Towards the base the whorl flattens repeatedly; first to a serrated cord making a slight carina and then again to another smooth cord. Only the former is visible on previous whorls and would appear to mark the periphery of the whorl. Traces of a short columella and an aperture with a short, shallow anterior canal are present. Discussion. Cerithiaceans with this general form are common amongst both European and American gastropods and include many marine and non-marine species. However, the Cretaceous age of the specimen and its occurrence in an undoubtedly marine deposit, prompt its inclusion in the genus Uchauxia. The ornament of the specimen serves to distinguish it from both U. wisei and U. badri. It differs from wisez in the number of spiral cords and their interspace width, while the possession of fewer tubercles on each cord and the stronger axial ridges emphasize the differences between the two species. The more reticulate pattern produced by cords, tubercles and ribs, assists in separating it from badri, although it is the markedly prosocline ribs of the latter which separate the two species and which give U. badri a slightly twisted look. Family CERITHIIDAE Fleming, 1828 REMARKS. Cossmann (1906) discussed the relationship of the Procerithiidae and the Cerithiidae (p. 20-21) and the derivation and evolution (p. 60-61) of the very prolific Cerithiidae in detail. He concluded that the anterior canal was the most important diagnostic character in separating the two families, even though the form of the canal was extremely variable within the Cerithiidae and that the demarcation between the subcanaliculate beak of the Metacerithiinae and the truncated canal of the Potamidiinae would appear to be very slight but quite distinct. In his view, the two families can easily be separated by the character of the canal, for even in the most truncated cerithiid the anterior canal always forms a slight ridge or projection beyond the surface of the base, while conversely, the beak of the Procerithiidae never protrudes below the base of the shell. In the Cerithiidae the canal is also generally twisted. Although Cossmann was satisfied that the Cerithiidae were derived directly from the Procerithiidae, it is possible that they may have arisen from the Jurassic family Eustomidae (see Cossmann, 1906 & 1913). Members of this family, consisting of the two genera Diatinostoma and Duitretus, possess a well-marked, but straight, cerithiid-like canal and generally occur in the coral-rudist assemblage at horizons from the Bathonian to the Turonian, (see Cossmann 1906 : 12-14). If this alterna- tive should prove to be correct, the relationship of the Cerithiidae with holostomatous forms might be further apart than Cossmann suspected. On the other hand, the apparent relationship may be explained by homeomorphy. It is particularly noticeable that from Upper Cretaceous to Recent times, the members of the Cerithiidae have frequently been associated with corals. In the PROGR RaGh ED AB, (Ch Rit ri DIA & CHRTTHTORSTDAE 151 Cretaceous, the richly coralline areas occurred further to the South and consequently, the British record of the Cerithiidae at this time is very sparse. The available evidence, which undoubtedly suggests that the Cretaceous was a period of experi- mentation within the family, represents differing degrees of specialisation of the entrance to the mantle cavity and of the anterior siphon. It is interesting to note that typical holostomatous procerithiid forms also occurred at these times but only persisted until the end of the Cretaceous. Sub-family > CAMPANILINAE Wenz, 1940 Genus ? CAMPANILE Bayle (in Fischer), 1884 ? Campanile cenomanica sp. nov. (Pl. 6, figs 1 & 2) Diacnosis. High, conical shell with whorls ornamented throughout by an adapi- cal carina and three prominent spiral cords; the twisted columella lacks any folds. MATERIAL AND OCCURRENCE. BM(NH) GG 18686; its history not certain, for although the original label states “Chalk Marl, Craie Chloritique, Dorset’ and gives the registered number 73597, the information in the register given with that number could not appertain to this particular specimen. Its preservation does suggest that it may come from the Cenomanian of Dorset, yet efforts to find similar material both in collections and in the field have not been successful. DEscriPTION. Unique turreted, conical, slender shell of medium size that is not completely preserved. Flat whorls, having straight whorl outline with a slight adapical shoulder. Adpressed linear suture. Whorl ornament characterized by a high adapical astragal with a rounded edge and bearing beads that are separated by intervals equal to their own width. Three prominent spiral cords are equally distributed below this and in each interspace there is a faint spiral thread. The central cord is somewhat weaker. Beaded ornament gradually develops on the astragal and on the cords. A fourth, but smooth spiral cord can be seen at the basal edge of the penultimate whorl. This cord marks the limit of the shallow convex base, ornamented by numerous equally spaced spiral cords. The spiral ornament is crossed by frequent opisthocyrt growth lines, which straighten a little towards the abapical suture, but after crossing on to the base, curve forward and then finally swing back towards the columella. At intervals corresponding to the occurrence of the beaded tubercles, and following the growth lines, there is a slight alteration in the growth direction of the shell. The poorly preserved aperture is not large in proportion to the remainder of the whorl, as in some species of Campanile. It shows a slightly twisted columella with a distinct anterior canal and has a callused parietal lip. There are indications of a small posterior ‘gutter’ or neck, and the form of the outer lip cannot be interpreted with any certainty. Discussion. The apertural features shown by this specimen, undoubtedly, indicate a strong affinity with the Cerithiidae. Its ornament and general shell morphology are comparable to those of the genus Campanile. However, the D 152 BRITISH CRETACEOUS GASTROPODS apparent atypical growth lines, the shorter anterior canal and lack of columella folds could be interpreted as an indication that it would be better placed in the Cerithiinae. Yet, on referring to the literature, one recognizes that the features generally accepted as those of Campanile, because of its Eocene type species C. gigantewm (Lamarck) 1804, are only one end of the range occurring within that genus. In fact, another Eocene species C. cornucopiae (J. Sow.) 1818, is basically very similar to the British Cretaceous specimen, having a sub-rhomboidal aperture, a small twisted anterior canal below a callused parietal lip, initially a flat whorl outline, a linear suture and the same slight straightening of the opisthocyrt growth lines. The presence of two columellar folds is the only difference between the two species. Two Maastrichtian species from Iran described by H. Douville (1904 : 312, pl. 43), C. morgami and C. breve show the rather simpler ornament of C. cenomanica. Another Maastrichtian species C. cavezi Vidal, (1917; 7, pl. 3, fig. 6) from Sensui, Spain, has the typical coronet ornament of later species of Campanile. All of these, however, possess the distinctive columellar folds of the genus. Whether the British fossil is a juvenile, the early whorls of an incomplete specimen, or an early form of the genus Campanile, remains to be discovered. It certainly does not show the characteristic two fold ornament that develops in Tertiary species of the genus. More material is also necessary before one can attempt to investigate its relationships with other Cretaceous species, or to make any conclusions as to its possible position in the evolution of the campanilids. Cossmann recognized that the genus first appeared towards the end of the Cretaceous, but C. cenomanica, if correctly identified, would probably place their origin farther back than he suspected. Genus EXECHOCIRSUS Cossmann, 1906 TYPE SPECIES. By original designation, Cossmann (1906: 121), Cerithium cingillatum Zekeli, 1852 (pl. 18, fig. 6), from the Upper Cretaceous of Austria. DESCRIPTION. Moderate-sized shells with weakly convex to flat-sided whorls. Distinct adpressed suture, often undulates and is occasionally grooved. Ornament extremely variable in size, strength and pattern, both within the genus and within species; consists basically of tuberculate spiral cords and axial ribs. Adapical ornament frequently prominent and separated from the remainder. Opisthocyrt to opisthocline growth lines. Genus is characterized by the presence of a prominent varix diametrically opposite the aperture on the body whorl. In some species, varices occur on earlier whorls but these are never aligned as in other Cerithiidae. A varix thins towards the adapical and abapical sutures. Small, low, rounded to quadrangular aperture with a short anterior canal and a weakly developed adapical ‘gutter’, or channel. Smooth, curved columella with callused inner lip; basal lip nearly flat, outer lip not expanded outwards. DIscussION. Cossmann (1906: 121) considered Exechociyvsus a subgenus of Tympanotonus Schumacher, 1817 on account of some similarity in ornament, but more particularly, because of their apertural characters. However, despite these PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 153 resemblances it would seem reasonable to reserve the genus Tympanotonus for non- marine gastropods and to place the marine Exechocirsus in a separate genus. Exechocirsus also has certain similarities with other Upper Cretaceous and Lower Tertiary cerithiid genera, e.g. Thericium, Clava (Clavocerithium), Clava (Pseudo- vertagus), Serratocerithium and Teliostoma. There is a common element in shell shape, size and in ornament patterns (in many cases varices occur at intervals), while apertural features are also alike. Further studies may show that Exechocirsus is a precursor of these genera and probably of other groups of the Cerithiacea such as Potamides and even Cerithiwm itself. Certainly, from amongst the few Cretaceous cerithiids, Exechocirsus appears to be the most likely ancestor for many of the Tertiary forms and this warrants its consideration as a separate genus. The possession of essentially cerithiid characters suggests inclusion in the sub-family Cerithiinae (Cerithiidae) rather than to accompany Tympanotonus in the Potamididae. Péhelincev (1953 : 192) has attempted to explain the strong and varied ornament occurring within the genus, by suggesting that it probably increased the mechanical stability of the shells and he cites their prolific occurrence at Gosau, in company with thick shelled Opisthobranchs and Nerinaea, as evidence. With regard to the type species of the genus, Cerithium cingillatum Zekeli, it is possible that this may be the early whorls of Cevithiwm pustulosum J. de C. Sowerby, 1835, a species which is also recorded from the Upper Cretaceous at the same locality (Gosau, Austria). Exechocirsus saundersi (Woods) (Pl. 7, figs 3-5) 1896 Cevithium saunderst Woods: 92, pl. 4. figs 12 & I2a. 1921 Cervithiwm saundersi Woods, Ravn: 47. HoLotyPe. Sedgwick Museum, B 4459, Cuckhamsley, North Farnborough, Berks., Turonian, Holaster planus Zone. OTHER MATERIAL. Chalk Rock, G 66724, Latimer, Bucks.; G 69915, Henley Quarry, ? Medmenham, Bucks; GG 20821, GG 20825-20826, Hitch Wood Pit (Hill End Farm), nr. Hitchin, Herts.; G 67704 H. planus Zone Dover; G 11523 Missen- den, Herts.; GG 20829, Kensworth Quarry, Beds. Forbes (1960: p. 237) records a large specimen from H. planus Zone, Western Colville, Cambs., SM. B 81559. Description. Medium-sized, slightly elongate shell with almost flat-sided whorls. Distinct impressed linear suture, which owing to the shell ornament appears to undulate a little. There are four equal spiral cords, with interspaces from one half to equal width between them. These cords are ornamented by rounded tubercles, which are longitudinally aligned, occur at the junctions of the cords and ribs and conform to the orthocline (or very slightly opisthocline) axial ribs. In general, the axial ribs number about sixteen per whorl, but this varies according to growth, and their interspaces are equal to their width. There are a varying number of fine spiral threads between the tuberculated cords. In some specimens, a smooth 154 BRITISH CRETACEOUS GASTROPRODS partly undulating cord occurs below the main spiral and just adapical to the suture. Woods had described and figured this, but in the holotype this cord appears to be composed of much smaller and more numerous tubercles (see pl. 7, fig. 5). The base is flat to very slightly convex, has a sixth granular, sub-carinate cord near the border and numerous spiral threads crossing its surface. The aperture is only partially exposed but appears to be obovate to quadrangular. It shows a short, well-formed anterior canal that is slightly curved, a callused inner lip and a concave columella. A strong varix occurs opposite the aperture on the last whorl and in a few specimens, there are indications of others on earlier whorls. Discussion. The occurrence of a strong varix on the last whorl diametrically opposite the aperture is characteristic of the genus Exechocirsus. In his original description, Woods distinguished the new species from C. pustulosum, J. de C. Sowerby (1835: pl. 39, fig. 19) by its smaller spiral angle and fewer tubercles. Another feature is that the axial ribs are more definite ; they are also opisthocyrt and only incorporate the tubercles of the lower three cords, those on the adapical cord remaining unconnected. The tubercles on the abapical cord become somewhat elliptical and on successive whorls they gradually diminish in size (see Zekeli, 1852, pl. 19, fig. 5). Finally, there are undoubtedly more longitudinal rows of tubercles, i.e. ribs, per whorl, in E. pustulosum (J. de C. Sowerby). Cerithium pustulosum d’Orbigny (non Sowerby), 1843 resembles E. sawndersi in ornament, but its longitudinal rows of tubercles are strongly opisthocline and are also said to be farther apart (Zekeli, 1852 : 100). In addition, the suture is grooved and the aperture more ovate. Woods also separated E. sawndersi from C. pseudoclathratum d’Orbigny, 1850 (as figured in Geinitz, 1874, pl. 31, fig. 5), yet, although the latter has similar ornament, its true generic position is uncertain. Sohl (1960 : 83) included C. pseudoclathratum with two American, Upper Cretaceous species that he has since (1964 : 364), mentioned as being closely related to Ceritiella Verrill, 1882. The Upper Cretaceous, Cerithium navasi Vidal, 1917 from Sensui, Spain, is easily distinguished from E. saunders, in only having three tuberculate spiral cords and a less prominent, finely-beaded, cord at the abapical suture. It is difficult to decide the relationship of several British specimens which show slight differences from typical sawndersi ornament, owing to the lack of knowledge of the range of ornament variation occurring in that species. In a number of these examples, the tubercles appear to be more numerous and this obscures any regular axial alignment that they may have. Two undescribed, unlocalized specimens from the Sowerby Collection G 60542— 60543, and others in material collected by C. W. Wright from Hitch Wood Pit, Hitchin, Herts. (GG 20822-20824) show distinctive features (see pl. 7, figs 8 & 9g; I,2&12). Basically there is a stronger development of both primary and secondary spiral cords, while the axial ribs are fewer (12-14 per whorl) and consequently wider apart. Among other features, the adapical cord is separated from the others by a much wider interspace, the suture occurs in a groove formed between adjoining whorls and varices frequently occur on earlier whorls. PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 155 In BM(NH) G 67704, (pl. 7, fig. 4) from the Chalk Rock of Dover, some of the very earliest whorls are preserved. Initially, the whorls are ornamented by two tubercu- late spiral cords, this eventually increases to three and within a further two whorls, to four. The spire in this specimen is considerably more slender than those of other species included in Exechocirsus. In another specimen, BM(NH) GG 20821, pl. 7, fig. x) the tubercles are weakly developed but the axial ribs produce a rather longi- tudinal step-like appearance in the shell, which also has a more distinctive suture. Other material BM(NH) GG 5932-5934, GG 5982-5983, GG 5987 and GG 5771 (see pl. 7, figs 13 & 14) collected by C. W. Wright, from the Holaster planus Zone at Kiplingcotes in East Yorkshire, is obviously related to E. saundersit. The slender shape of the shells, their markedly opisthocyrt growth lines and opisthocyrt (rather than opisthocline) ribs suggest they are a different form. Additional support to such a view, is provided by their intermediate spiral threads, which are stronger than in typical saundersi and by the fact, that in general, they are much smaller specimens. It is possible that these features are the result of a phenotypic response to differences in environment and for the moment, the specimens are regarded as Exechocirsus aff. saunderst (Woods). ? Exechocirsus aft. subpustulosus Pcéhelincev, 1953 (Pl. 7, figs 10 & 11) 1953 Tympanotonus (Exechocivsus) subpustulosus Péhelincev: 193-194, pl. 33, figs 13-16. 1960 Tympanotonos (Exechocirsus) subpustulosus Pchelincev, in Orlov, pl. 19, fig. 13. MATERIAL. BM(NH) G 737908 and G 16134, Upper Greensand, Upper Albian, Haldon, Devon. DEscrIPTION. Poorly preserved medium-sized turriculate shells with feebly convex whorls that have a very small adapical sutural ramp, a slightly grooved undulating suture and opisthocyrt growth lines. The whorls are ornamented by about 18 narrow axial ribs, that are orthocline on the early whorls but rapidly become opisthocyrt. These ribs are separated by inter- spaces that are at least equal to their width. On the earlier whorls a narrow spiral groove cuts the ribs to produce an adapical row of tubercles. Other similar spiral furrows occur abapically to this, on later whorls, and these dissect the ribs even more, with the result that the original predominantly axial ornament becomes changed to a spiral pattern. A single spiral thread occurs in each groove. Numerous fine spiral threads can be seen in the interspaces between the ribs. Further spiral grooves occur towards the base on the last whorl. Neither the base of the shell or the aperture can be seen. Discussion. Lacking knowledge of apertural details and without any indication that these specimens possess the characteristic varix of Exechocirsus, it is not possible to refer them to that genus with certainty. They are provisionally referred to the Transcaucasian Cenomanian Exechocirsus subpustulosus Pthelincev (1953) on the basis of ornament resemblance. The axial ribs of swbpustulosus are also dissected - by spirals to produce four rows of tubercles and both species have a rounded base. 156 BRITISH CRETACEOUS GASTROPODS Although the adapical rows of tubercles is the most prominent and is also separated by a wide groove from the lower rows, it has fewer tubercles than that of the British specimens. Pchelincev remarked that there is a tendency for the ornament of subpustulosus to become weaker on later whorls and this effect is also apparent in the Haldon material. It is only the lack of varices on the earlier whorls, together with the narrow apical angle, that enabled Pcthelincev to distinguish his species from E. pustulosus (J. de C. Sowerby). Family ? CERITHIIDAE Genus AGERIA nov. Named in honour of Dr D. V. Ager, Professor of Geology at the University of Swansea, in recognition of his help and advice. TYPE SPECIES. Ageria gaultina sp. nov. DiaGnosis. Turriculate shells, ornamented by spiral cords and axial ribs; the cords are frequently beaded and the abapical cord is characteristically strongly developed and ornamented. GENERIC CHARACTERS. Small-medium sized, slender, turriculate shells. Convex whorls with a slight concavity before an abapical astragal. Flush, slightly grooved, undulating suture. Regular ornament formed by axial ribs and spiral cords. In some species the axial ribs weaken on later whorls, but in all species they have a characteristic opisthocyrt shape. In most species, the primary spiral cords are beaded, but in some the cords are smooth. A strong cord forms a distinctive abapical feature which delimits the base. Opisthocline growth lines influence the spiral ornament and show a median sinus, they swing back towards the columella on the uniform basal slope. Axially elongate, quadrangular aperture and indication of a short anterior canal with a smooth, slightly curved columella. Discussion. The general shell shape and ornament, together with the slight indication of a short anterior canal, suggest that this distinctive group should be included in the Cerithiidae. Yet, the lack of adequately preserved material showing the aperture prevents a definite conclusion being formed; particularly as the indica- tions of an anterior canal in the type species A. gaultina sp. nov., could be the result of crushing. Consequently, it is not possible to assign Agevia to a particular sub- family of the Cerithiidae. Furthermore, its similarity in ornament to the procer- ithiid genus Cryptaulax Tate, 1869 is noticeable and the possibility that Ageria could be derived from this cannot be dismissed. In many respects, the generic characters resemble those found in members of the Turritellidae, i.e. the turreted shell, the whorl outline and the possession of a thin parietal lip with a smooth columella. However, the holostomous, sub-circular aperture of the Turritellidae makes any further comparison rather pointless, while the characteristic lack of axial ornament in that family together with the marked sinuosity of its growth lines and outer lip and the deeply excavate columella provide further distinguishing characters. Two species that can now be referred to Ageria are Turnitella costata J. de C. PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 157 Sowerby, 1827, Upper Greensand, Albian, Blackdown, Devon; and Cerithiwm binodosum Roemer in Frié (1894, text-fig. 87), Lower Senonian, Biezno Beds (= Priesener Schichten) Brezno, Czechoslovakia. Questionably referred species are Turritella angustata d’Orbigny, 1850 from the Neocomian of France (see p. 158) and Cerithiwm hector d’Orb. (1850, p. 156) in Guéranger, 1867 (pl. 14, fig. 2), Cenomanian from Le Mans, France. The original figure of Turritella cingulatolineata J. Miller (1851, p. 33, figured as cingulatocostata, pl. 6, fig. 15) from Aachen appears to be that of an Ageria, having a characteristic prominent abapical cord. Yet, the ornament shown in the figure provided by Holzapfel (1888, pl. 14, fig. 1) when he interpreted Miuller’s species as belonging to the genus Mesostoma, is quite unlike that of Ageria. This figure shows a median row of prominent tubercles, while its quadrangular aperture hints at the presence of a strong abapical cord. Ageria gaultina sp. nov. (Pl. 8, figs 1-5) DiaGnosis. Narrow, tutriculate shell, ornamented by orthocline axial ribs which do not persist on later whorls; its strong abapical cord is carinate. HoLotyPeE. BM(NH) GG 6473 Albian, Gault, Folkestone, Kent. J. S. Gardner Collection. OTHER MATERIAL. BM(NH) G 4362 (26 specimens), G 11318, G 11544, G 20741- 20742, GG 20857-20862, 38085 (2), 48136 (3), Albian, Gault, Folkestone. G 71026, Gault, East Wear Bay, nr. Folkestone, Kent and G 49804, Albian, Osmington, Dorset. Dr H. G. Owen of the Dept. of Palaeontology, BM(NH) considers that the matrix of GG 20742 is consistent with that of Bed V of the Lower Gault at Folkestone, in the Lautus-Nitidus Subzone. DEscrIPTION. Small to medium-sized, turriculate shell, with an acute spire composed of moderately high whorls (height equal to ? of width). Flush, linear suture. Whorl outline slightly convex but briefly becoming concave before an abapical peripheral carina. Whorls ornamented by nine orthocline axial ribs, that are crossed by numerous beaded spiral cords (22 from adapical suture to the peripheral carina in the holotype), with the interspaces separating the ribs equal to twice the width of a rib. These axial ribs are successively displaced from whorl to whorl and this produces a counter spiral pattern in relation to the direction of coiling (see pl. 8). Such ornament occurs in the earliest whorls, but initially the orthocline ribs are much stronger than the simple spiral cords and therefore, in cross section, the whorls are angular. The basal beaded spiral is the first to become differentiated with other cords soon developing and varying in strength and composition. A pattern of five strongly beaded primary cords witha varying number of more finely beaded secondary cords occurring in their interspaces is formed. In some cords the beads are rounded, in others they become elongated in the direction of the growth lines, while the ‘beads’ 158 BRITISH CRETACEOUS GASTROPODS of some secondary cords are the result of the growth lines simply dissecting the cord. As the ornament of the spirals strengthens, the ribs tend to diminish and are almost obsolete on later whorls, where they are only indicated by the presence of strong tubercles on the primary cords. The opisthocline growth lines influence the spiral cords, but do not affect the direction of the axial ribs. The base is delimited by the prominent basal spiral cord and this is followed inwards by a second, less prominent, primary cord, the base then slopes uniformly to the columella. The remainder of the surface is ornamented by alternating granulose spiral cords and smooth spiral threads. After crossing the basal carina, the growth lines swing sharply back towards the columella. Unfortunately, a completely preserved aperture is not available, but it is probably somewhat quadrangular and axially elongate. There are some indications of a short anterior canal but this may have resulted from crushing of the shell. The outer lip is virtually straight in cross section, but the growth lines show a slight median sinus. The inner lip is smooth, partly callused and rather narrow. The columella is very slightly curved, but this could also be a result of crushing. Discussion. There is a little variation within the ornament of this species. A specimen of 5 whorls, found at Osmington, Dorset, BM(NH) G 49804 (see pl. 8, fig. 5) has only thirteen spiral cords between the adapical suture and the basal carina and not one of these cords is beaded, yet it does show a typical sequence of primary cords. Among ornament variations shown by material collected from the Gault of Folke- stone, Kent are more equidistant primary cords; a less prominent peripheral beaded carina; broader (i.e. thicker) spiral cords ornamenting the base and a considerable range of differences in the strength and dominance of both spiral and axial elements on the early whorls (see pl. 8, figs 2, 3, 4 & 6, BM(NH) GG 20741-20742, GG 20857, GG 20862). Agena gaultina is easily distinguished from A. costata (J. de C. Sowerby), described below, by its orthocline axial ribs and the fact that these ribs do not persist on to later whorls. Its more carinate abapical cord also serves to identify the species. The specimen figured as Cerithiwm binodosum Roemer, by Fri€ (1894, text-fig. 87), from the Lower Senonian, Biezno Beds (Priesener Schichten) of Selten, nr. Louny, N. W. Bohemia, Czechoslovakia, is not like any other specimen figured as C. binodosum Roemer, 1841 and undoubtedly belongs to the genus Agevia. It possesses the distinctive, strongly ornamented abapical spiral cord of that genus and has very similar whorl and apertural shapes. The cancellate pattern formed by its axial ribs and spiral cords separates it from other species of Agerza. Cryptaulax angustatum (d’Orbigny), Peron (1900 : 93, pl. 1, fig. 14) from the Neocomian of France, is probably another species of Agevia. The ornament, whorl shape, axially elongated quadrangular aperture, and smooth, slightly curved columella, are features in common with other members of that genus. Yet, its aligned (initially orthocline) prosocline axial ribs are also strongly reminiscent of the genus Cryptaulax Tate, 1869. The apparent lack of a strong abapical cord adds to the uncertainty, while its smooth spiral cords would readily distinguish it from the recognized species of Ageria. All of these features are discernible in the figures PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 159 provided by both d’Orbigny (1842, pl. 151, figs 4 & 6) and Peron, and in a British Neocomian specimen BM(NH) GG 20740, from the Claxby Beds, Nettleton, Lincs. (see pl. 8, fig. 13). Ageria costata (J. de C. Sowerby) (Pl. 8, figs 7-12) 1827 Turritella costata J. de C. Sowerby: 126, pl. 565, fig. 4. 1900 Turritella costaia Sow., Jukes-Browne: 464. Diacnosis. Acicular shell with persistent prosocline axial ribs and a shelf-like abapical cord. Hototyre. BM(NH) 43674, Albian, Upper Greensand, Blackdown, Devon, Sowerby Collection. OTHER MATERIAL. BM(NH) 34844 (2 specimens), G 20830-20832, G 71382, G 16142 (2 specimens) from Upper Greensand, Blackdown, Devon; G 71090-71095, from Bed 12 Peak Hill, nr. Sidmouth, Devon, H. F. Metcalfe Collection. DESCRIPTION. A small, slender, somewhat subulate shell with convex whorls that appear to be twice their height in width. Flush, undulating suture occurs in a very slight groove beneath the abapical astragal. The whorl outline becomes concave just before this astragal and causes a small constriction in its shape. The growth lines are opisthocline and show a definite median sinus. The shell is characteristically ornamented by a consistent pattern of axial ribs and spiral cords. Initially, the eight, or ten ribs are orthocline and dominate the ornament. However, the ribs soon become prosocline and terminate above the basal astragal as described by Sowerby. The spiral ornament consists of four to five beaded to tuberculate primary cords with both weaker beaded cords and smooth threads between these. The peripheral carina, or astragal, is shelf-like and dissected by numerous growth lines. The flattish base is ornamented by rather broader spiral cords than those of the whorl side. A strong cord, just in from the carina, forms the main feature. The aperture (see pl. 8, fig. 9) is quadrangular and elongated axially, with a notch in the outer lip just above the astragal. The parietal lip is quite thin and the columella is not callused. The base of the aperture is flat and an anterior canal is not apparent. Discussion. Although this species is very similar to Agevia gaultina sp. nov. in _ ornament, there are a number of differences between them. The orientation of its axial ribs is markedly prosocline as opposed to the orthocline ribs of gaultina. A further contrast is that in costata the ribs persist and occur on later whorls. The shell also appears to be more needle-like in shape, even though its whorl diameter: whorl height ratio seems lower than that of A. gaultina. The whorls of costata are more convex, while the abapical astragal that delimits the base of the two species, is shelf-like in costata and carinate in gaultina. 160 BRITISH CRETACEOUS GASTROPODS Family CERITHIOPSIDAE Genus CERITHIELLA Verrill, 1882 [Lovenella Sars, 1878, non Hincks, 1869; Newtonia Cossmann, 1892, non Schlegel, 1866; Newtoniella Cossmann, 1893; Cevithiolinum Locard, 1903]. Type by original designation, Cerithium metula, Loven, 1846, Recent. REMARKS. Wrigley (1940 : 13) considered that Cerithiella has a moderately deep sunken suture and that its subordinate axial riblets cross the spiral carinae frequently producing nodes at the intersections. In his view, the related genus Sela has an inconspicuous suture, is composed of essentially flat-sided whorls and its fine axial threads are only visible in the interspaces between the spirals. However, in instances where species have ornamental features similar to those of Seda, the truncate concave columella and short twisted anterior canal of Cerithiella is easily distinguishable from the virtually straight columella occurring in Seda. The aperture of Cervithiella has been described as sub-quadrangular as opposed to the sub-ovate shape found in Sez/a. The Cretaceous species mentioned in the discussion below, were originally assigned to the genus by Sohl (1964 : 364). Cerithiella (Cerithiella) devonica sp. nov. (Pl. 6, fig. 9) MATERIAL. Unique specimen in the Sedgwick Museum, Cambridge, B 44633 from the Upper Albian, Blackdown, Devon. Diacnosis. Acuminate cerithiopsid with cancellate ornament consisting of three equally-spaced tuberculate spiral ribbons and numerous fine axial threads and possessing a short twisted anterior canal. Dimensions. Height = 11 mm; Maximum diameter 3 mm. DESCRIPTION. Small slender shell with turriculate spire. Whorls almost flat- sided and with a very slight adapical ramp. Linear suture in moderately-deep asymmetric channel formed by adjacent whorls. Ornament consists of three equally-spaced spiral ribbons, the adapical ribbon being very slightly weaker. Very faint spiral threads occur on the short ramp below the adapical suture. The concave interspaces are equal to the width of the spirals. The fine axial threads are slightly opisthocline and produce angular tubercles where they cross the spiral ribbons. Ornament of the thirteen preserved whorls is consistent. The basal ornament, growth lines and protoconch are not adequately preserved for description. The aperture is incomplete but there are indications of a short twisted anterior canal. Discussion. The characters of ornament, whorl shape, suture and aperture appear to be sufficient to indicate that this specimen belongs to the genus Cerithiella. Its cancellate ornament of three spirals and numerous axial threads is very close to that of the North American Upper Cretaceous species C. nodolivatum (Wade), 1926. On comparison with the figures given by Sohl (1960; pl. 9, figs 15-16, 24) only a PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 161 slightly longer anterior canal in nodolivatwm and the presence of the fine spirals on the short adapical ramp of devonica could be used as distinguishing features. Yet, these small differences when considered together with the widely separate geo- graphical and stratigraphical occurrence of the two forms would seem to support the belief that they are distinct species. Unfortunately, the basal ornament of the British specimen is not preserved and consequently, it is not possible to verify the distinctive growth line pattern on the base of the shell, described by Sohl (1964 : 365) for the American cerithiopsids. C. semirugatum (Wade) 1926 and Cerithiella sp. nov. Sohl, 1964 are easily dis- tinguished from C. devonica by their ornament of four spiral ribbons. Similarly, the ornament of four variable cords in Cerithium bicostatum Kaunhowen (1808 : pl. 6, fig. 17) from the Maastrichtian of Belgium, separates that species from devonica. Kaunhowen also mentions a short, but straight canal, in his description. Fic. 2. 1, Cerithiella devonica sp. nov., Cretaceous, Upper Greensand, Albian, Blackdown, Devon. x4. Cerithiella metula (Lovén). Type species. Recent, North Sea. 2:6. 3, Cerithiella nodolivatum Wade, Cretaceous, Maastrichtian, Ripley Formation. Tennessee, U.S.A. x5. 4, Cerithiella semivugatum Wade, Cretaceous, Maastr., Ripley Formation, Tennessee. U.S.A. 4:6. 5, Cerithiella cloacina Wrigley, Eocene, London Clay, Whitton, Middx., England. x5. 6, Cerithiella atherfieldensis sp. nov., Cretaceous, Lower Greensand, Aptian, Atherfield, Isle of Wight. x2. 162 BRITISH CRETACEOUS GASTROPODS The more turreted character of the Lower Aptian, Cerithiella atherfieldensis sp. nov. described below, together with the variation in strength of its spiral cords, its sub- carinate whorls and limited axial ornament provide distinguishing characters. In some Tertiary examples of the genus, C. cloacina Wrigley and C. praelonga (Deshayes) the spiral cords are smooth. [Text-fig. 2] Cerithiella (Cerithiella) atherfieldensis sp. nov. (Pl. 6, fig. ro) HoLotyPe. Sedgwick Museum, Cambridge B 27334 from the Lower Aptian, Forbesi Zone, Crackers, Lower Greensand, Atherfield, Isle of Wight. DiaGnosis. Narrow, turreted shell, ornamented by five spiral cords of varying strength (the fourth producing a slight carination before the grooved suture) and limited axial threads. DESCRIPTION. Small, acutely spiral turreted shell of slightly convex, sub- carinate whorls. Distinct suture in narrow groove. Protoconch not preserved. Aperture sub-ovate with a strongly twisted anterior canal; short truncated columella. Shell ornamented by five, more or less equidistant primary spiral cords and fine axial threads. A cord occurs on either side of the suture with three stronger ones in the centre of the whorl. The more abapical of these is the strongest and almost forms a carina. The interspaces are much wider than the width of the cords and a fine secondary spiral cord occurs in each of the two central interspaces. On the base, a prominent interspace separates numerous fine spiral cords from the most abapical primary cord. The fine axial threads are slightly opisthocyrt on the whorl side and are not visible on the base. These axial threads are more prominent on the earlier whorls but always weaker than the spiral cords. Discussion. In this species, the small spiral angle, the limited axial ornament, the somewhat turreted appearance, the variation in strength of its spiral cords, and the abapical carination of the whorls, all serve to distinguish C. atherfieldensis from the other Cretaceous species mentioned in the discussion of C. devonica. It is also the earliest recorded species of this genus. Genus SEILA Adams, 1861 [Cinctella Monterosato, 1884] Type by subsequent designation, Dall (1889 : 250), Tviphoris dextroversus Adams & Reeve, 1848, Recent. Diacnosis. (After Sohl, 1960 : 84 and 1964 : 364). Moderately small to very small, slender, turriculate, thin shell. Smooth, swollen, paucispiral protoconch. Whorls flat-sided and ornamented by prominent, smooth spiral, ribbon-like cords with numerous fine growth-line threads in their interspaces. Indistinct suture. Last whorl moderately high and with a carina at basal periphery. Aperture subovate, short anterior canal with a swelling on base of almost straight columellar lip. Outer lip thin. PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 163 Seila iglali sp. nov. (Bl 6) fig 12) MATERIAL. Unique specimen, BM(NH) G 16084, from the Upper Greensand, Albian, Combe Raleigh Hill, nr. Honiton, Devon, W. Vicary Collection. DiaGnosis. Conical shell, large for genus, ornamented by four raised, flat-topped spiral ribbons on whorl side and numerous smooth spiral cords on base, with trans- verse threads occurring in the interspaces. DESCRIPTION. Small, conical shell (Height: 13:5mm; Maximum diameter: 4:8 mm) with a flat-sided whorl outline. Whorls ornamented by raised, flat-topped spiral cords (i.e. ‘ribbons’) with fine transverse threads in the concave interspaces between these. Distinct linear suture occurs in a symmetrical groove between adjacent whorls. Early whorls appear to be smooth and rather convex (although this may be the result of poor preservation). Traces of axial ornament can be seen adapically on the fourth whorl, while the fifth shows faintly the typical ornament of subsequent whorls. This consists of four prominent flat spiral ribbons that are approximately equal in width but vary in their distance apart (see Text-fig. 5). The fourth ribbon is slightly narrower and forms a basal carina. The concave interspaces are crossed by regularly-spaced, easily visible, axial threads. Interspaces roughly three times the width of a thread. Axial threads essentially orthocline, but towards the base, i.e. between the third and fourth spiral ribbons, and particularly on later whorls, they tend to become opisthocline. On close examination, it is possible to see that the axial threads cross over the spiral ribbons. Within the basal carina, and after a further concave interspace, the flat surface of the base is ornamented by numerous (15-16) identical smooth spiral cords that continue on to the columella and are separated by intervals equal to their width. Growth lines are not visible on the base. Aperture small and sub-ovate ; columella broad and equal to one third of the whorl diameter, and is twisted anteriorly. Anterior canal short and narrow. Discussion. The general flat whorl outline, together with its ornament of raised, ribbon-like spirals, leads one to consider this species as a member of the genus Sezla. However, its size, which is greater than that of most species assigned to this genus, and its distinct linear suture, cause some uncertainty. In addition to these doubtful features, its conical shell shape is also rather unusual and quite unlike the slender, sub-cylindrical form of the type species Sezla dextroversa (Adams & Reeve). Ina number of Recent species of the genus, the base of the anterior canal forms a slit, which extends for one third of the whorl’s circumference. This feature is certainly lacking in S. iglali and while this could be due to poor preservation, it also appears to be lacking in North American Upper Cretaceous species which suggests that it is not present in these earlier forms. Obviously, in order to confirm that this Albian fossil is correctly assigned to the genus Seila, it is necessary to obtain a well-preserved protoconch. Sohl (1960 : 84) refers the North American, Upper Cretaceous species Cerithiopsis 164 BRITISH CRETACEOUS GASTROPODS meeki (Wade, 1926) and Cerithiopsis quadrilirata (Wade, 1926) found in the Ripley Formation (Maestrichtian), to the genus Sez/a, on the basis of their ornament and protoconch features. Since the columella fold found in both of these species is much stronger than in typical Sezla, and, in fact, more reminiscent of other Cerithiopsidae, he qualifies this decision. Yet, as none of these genera possess ornament similar to that of meekt and quadrilivata, he refers the two species to Sezla. Seila iglali is very close to S. meeki (Wade) and apart from their considerable difference in size, the only feature that will separate the two species is the presence of a fourth spiral ribbon in zglali. In both Sela quadrilirata (Wade) and Sela sp. Sohl, (1960 : 85, pl. 9, fig. 23), some of the spiral ribbons are very close together. Among other distinguishing features, the axial threads of S. quadvilivata are finer and more numerous and also described as being prosocline; while in Sez/a sp. no. axial threads are present. Sohl (1964 : 364) has also recorded S. meeki from a slightly older horizon in the Upper Cretaceous [Coffee Sands of Mississippi (Campanian) | and noted that its spiral ribbons were much thinner and consequently farther apart. The limited amount of material available, prevented him from considering the possible significance in this. Among Eocene species, Seila mundula (Deshayes, 1865 : 222) is the closest to iglali, but can be separated by its convex base and very much finer axial threads. In contrast, to the Cretaceous specimens referred to ?Cerithiella, there is no indica- tion of secondary spirals developing in the interspaces of the primary ribbons in S. iglali. If the determination of this specimen is correct, it will be the earliest record of the genus. Fic. 3. 1, Seta dextroversa (Adams & Reeve), type species of genus. Recent. China Seas. xg. 2, Seila iglali sp. nov., Cretaceous, Albian, Devon, England. x2-6. 3, Seila mundula (Deshayes), Eocene, London Clay. Highgate. 3-3. 4, Seila meeki (Wade), Cretaceous, Maastrichtian. U.S.A. x5 approx. 5, Sela quadrilivata (Wade) Cretaceous, Maastrichtian, U.S.A. 3-3 approx. PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 165 Genus ORTHOCHETUS Cossmann, 1889 TYPE SPECIES. By original designation, Cerithium leufroyi Michelin in Deshayes, (1833 : 380), Lutetian, France. DiaGnosis. (After Wenz, 1940 : 779). Medium-sized, slender, turriculate regu- larly conical shell, composed of numerous low whorls that have an abapical keel. Whorls ornamented by spiral cords and axial riblets, frequently producing a reticu- late network. Last whorl approximately one third of shell height. Aperture sub- quadrangular, with a strong, moderately long, truncated canal; thin outer wall, slightly sinuous adapically. Columella straight, with base a little twisted; may be smooth, or with several oblique folds. Growth lines opisthocline to opisthocyrt, but swing back towards the columella on the base. Orthochetus hantoniensis sp. nov. (EIS Osdigs 2) MATERIAL. BM(NH) G 70130, Aptian, Atherfield, Isle of Wight. Diacnosis. Slender, turreted cerithiopsid with an abapical carina and strongly developed axial ribs (fewer in number than normal), which influence five of the six spiral cords to produce a frilled appearance. Columella straight, but uncharacter- istically short and apparently without plications. DESCRIPTION. Small, slender, turreted shell with fairly low whorls. The suture is subcarinate and linear, but axial ribs may give it an undulate appearance. The protoconch is not completely preserved, but early ornament seems to be identical to that of later whorls. This ornament consists of strong, straight axial ribs (15-16 per whorl) that are separated by interspaces equal to their width; and six spiral cords of varying strength. A slight adapical ramp occurs immediately beneath the suture and this is bounded by a strong spiral cord. At equal distances below this follows a very weak cord, a moderately strong cord, then at the abapical carina another strong cord, and, finally, another weak cord. All of these five cords are influenced by the axial ribs, the apparent thickening at their intersections producing a frilled, or tuberculate appearance to the shell. A sixth spiral cord close to the abapical suture, limits the base and is unaffected by the axial ribs. The axial ribs diminish soon after the abapical carina. Numerous faint opisthocline growth lines occur on the whorl side and on the base these swing back towards the columella. Columella moderately short and straight with faint oblique ornament along its length. Small aperture subquadrangular, showing a well-developed, twisted and truncated anterior canal. Thin sinuous outer lip. Discussion. The distinctive ornament and slender turreted spire of this species, together with its general apertural features, are sufficient to refer it to the genus Orthochetus. However, the columella is somewhat shorter than is usual, while its twisted anterior canal is less inclined; in fact, such features are reminiscent of the genus Cervithiella. The strongly developed axial ornament and its consequent 166 BRITISH CRETACEOUS GASTROPODS influence on the shell’s appearance are adequate diagnostic features to separate this species from others that have been described. The lack of plications on the columella might be construed as uncharacteristic of Orthochetus, yet, Wrigley (1940 : 1r & 12) mentioned a number of instances in Eocene species where this feature is obscured by a reflected columella border. Its diagnostic features distinguish O. hantoniensis from the two Maastrichtian species mentioned by Cossmann (1906 : 97) Cerithium tectiforme Binkhorst (1861: 24 pl. 1 figs 3a-c) from Limbourg and O. mapeulensis Douvillé (1904 : 303 pl. 41 fig. 12) from Kouh Mapeul, Persia. The ornament of both these Maastrichtian species is more reticulate, due to the presence of smaller and more numerous axial ribs. Although both O. hantoniensis and O. tectiformis have three primary spiral cords, the presence of secondary spirals in the former and of tubercles in tne latter assist in separating the two species. Douvillé (1904 : 303) referred to Ovthochetus as an ancient group and quoted the occurrence in the Chalk of Cerithium cribriforme Zekeli (1852 : pl. 20, figs 2 & 2’) as evidence. In many respects, this turreted species is very like O. hantoniensis, Fic. 4. 1, Ovthochetus tectifoymis ( Binkhorst), Cretaceous, (Maastrichtian), Limbourg Netherlands. x1-3. 2, Ovthochetus hantoniensis sp. nov., Cretaceous (Aptian). Atherfield, Isle of Wight. 3:3. 3, Orthochetus leufroyi; (Michelin in Deshayes 1833). Eocene. Chaumont, France. Type species. x0-6. 4, Orthochetus charlesworthi (Prestwich), Eocene, London Clay. Sheppey, Kent. 0-6. 5, Orthochetus mapeulensis Douville. Cretaceous, (Maastrichtian). Louristan, S. Persia. 1-3. 6, Ovthochetus elongatus Wrigley. Eocene, London Clay, Sheppey. xo-6. PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 167 having strong axial ribs and three smooth spiral cords undulating over them. Yet its aperture is not shown, and there are no indications of secondary spirals in the figure which does show that C. cribriforme is rather awl-shaped, i.e. subulate. The existence of O. hantoniensis would appear to substantiate Douville’s claim that the genus Orthochetus arose in the Cretaceous, even if C. cribriforme Zekeli should prove to be mis-identified. Several of the larger Eocene species show that the anterior canal and the columella are not necessarily identical to, or as long as that of the type species O. lewfroyi. Comparison of O. charleswortli (Prestwich) and O. elongatus Wrigley illustrates that there is considerable variation in spiral angle and whorl diameter. Wrigley (1940 : 12) also showed that within a species there is a wide variation in the strength and position of its ornament components, e.g. see his figures of O. elongatus (p. 17, figs 13-15). Incertae sedis ? ‘Orthochetus’ helmyi sp. nov. (Pl. 6, figs 13 & 14) MatTerrIAL. BM(NH) G 10500 and G 71988 from the Gault, Albian of Folkestone, Kent. Diacnosis. Turreted, conical cerithiopsid, ornamented by four prominent flat- topped ribbon-like spiral cords and numerous fine opisthocyrt axial threads. Description. Medium-sized, slightly turreted, conical shell. Distinct linear suture lies in a furrow between the abapical and adapical primary spiral cords of adjacent whorls. Protoconch not preserved. Whorl side ornamented by four strong, prominent, flat, ribbon-like spiral cords, with a fifth weaker primary cord occurring at the abapical suture and delimiting the base. The two lower ribbons, i.e. the third and fourth abapically, are stronger and wider. Secondary spiral threads occur in the unequal interspaces between the primary ribbons, but can only be seen easily on the lower halves of later whorls. Fine axial growth threads occur in the interspaces and also continue over the spiral ribbons. These opisthocyrt threads show a broad shallow sinus, which has its apex at mid-whorl; each thread is separated from its neighbour by a space equal to double the width of a thread. The base is flat to feebly convex and within the delimiting fifth primary cord, it is ornamented by numerous thin secondary spirals. The opisthocyrt growth lines, after crossing on to the base, tend to straighten and then swing back towards the columella. The aperture is sub-quadrangular and has a flat basal lip. The short columella is smooth. Discussion. It is difficult to assign this species to one particular genus of the Cerithiopsidae without having a specimen that shows the important details of its aperture and protoconch, for its ornament and whorl shape have characteristics similar to those of Cerithiella, Seila and Orthochetus. Reference to the literature and to collections of Tertiary representatives of these genera does not provide any E 168 BRITISH CRETACEOUS GASTROPODS useful guide, since it is apparent that there is some confusion in distinguishing the genera and in assigning species to them. The criteria that Wrigley (1940 : 13) used, have been referred to under Cerithiella (page 160), but these are often inadequate, e.g. the inconspicuous suture that is considered a diagnostic feature of the genus Sezla. Undoubtedly, a reappraisal of the criteria used to distinguish the genera of the Cerithiopsidae is necessary and the review of species assignations suggested by Sohl could be undertaken at the same time. Some of the features present in ‘helmyz’, i.e. the distinct linear suture and the crossing of the spirals by the axial threads, indicate possible affinities with Cerithiella. The moderate size of the specimen and the fact that its whorls could hardly be described as flat-sided, indicate that it should not be assigned to Sezla, but could easily be recognized as a possible member of Ovthochetus. The presence of promi- nent, ribbon-like spiral cords and the fine axial opisthocyrt threads and the basal features of the whorl, together with the indications of a short straight columella are reminiscent of several species of Seda, including the type species S. dextroversa. However, these features could equally be those of Orvthochetus and therefore after recognizing that the linear suture is very like that of Eocene species of this genus, helmyi is provisionally placed in Orthochetus. In the majority of described species of this genus, and particularly in the earliest member O. hantoniensis, the axial orna- ment is quite prominent and contributes to the typical reticulate ornament pattern. The size, ribbon-like spirals, super-imposed axial ribs and turreted whorl shape of helmyi are considered to be the most important indicators of its probable relationship with Orthochetus. Similar difficulties prevent the assignation of a unique specimen, preserved as an external mould in an ironstone concretion from the Sandgate Beds, found at Parham Park, Sussex (BM(NH) 9145, G. A. Mantell Collection, (see pl. 6, fig. 15). The ornament of this medium-sized, turreted gastropod consists of three primary and three secondary spiral cords. It has a distinct linear, impressed suture and opisthocyrt growth lines between the sutures. The whorls are not flat-sided, (in fact they are concavo-convex) and the spiral cords are carinate, or rounded, not ‘ribbon-like’. Axial ornament, apart from the collabral threads, is lacking. A slight distortion in the smooth opisthocyrt curve wherever the growth line crosses a spiral cord produces a wavy effect. Unfortunately, neither the abapical portion of the body whorl, nor the aperture are preserved. VI. CONCLUSIONS The need for exhaustive re-examination and re-evaluation of existing collections and type material has been repeatedly mentioned in the discussion above and can only be emphasized again here. Before the taxonomy of the Cerithiacea can be revised, it is necessary to improve upon the incomplete information currently available by collecting additional material. Admittedly the task of revision is complicated by the limitations of the characters that are used to distinguish fossil gastropods, while their poor preservation adds to this difficulty. Yet, the careful collecting advocated by present-day palaeontologists should enable these problems PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 169 to be surmounted. The important morphological and ornamental variation of Cretaceous gastropods can be satisfactorily assessed, once accurate information on their stratigraphical occurrence is produced. These conclusions would then con- tribute to and facilitate our understanding of their phylogenetic and palaeoecological relationships. The family Cerithiidae became conspicuous during the Cretaceous and during that period experimented with the form of the aperture and anterior canal. Fossil evidence is sparse and consequently, our knowledge of the forms which are thought to be transitional between either the Procerithiidae, or the more cerithiid-like Eustomidae, and true cerithiids is very limited. A similar situation occurs with their descendants. In order to achieve a re-assessment of these families, a full understanding of the derivation of the anterior canal is vital. Critical genera such as Ageria, Cimolithium and Uchauxia require further investigation, and determina- tion of the phylogenetic position of Metacerithium, Exechocirsus and Campanile is equally important. Cox (1965 : 158) briefly mentioned the possibility that difficult groups might prove to be polyphyletic (he referred to this while discussing the origin of Exelissa, a genus of the Procerithiidae), but did not proceed to discuss the taxonomic implications of such a suggestion. It would provide an easy explanation of the morphological diversity apparent in ‘Rhabdocolpus’ but careful consideration is necessary before such a theory can be followed. The variation in ornament shown by several of the Cretaceous gastropods dis- cussed here, is not thoroughly understood. In some genera, e.g. Metacerithium, ornament is said to be basically stable, while in others, e.g. Bathraspira it would appear to alter very quickly. Further, the significance of sculptural variation within a species, such as that shown by Exechocirsus saundersi (Woods), is also not fully appreciated. It has been suggested that the smaller Yorkshire specimens of this species might be explained as an example of ecotypic variation. The results of the detailed study recommended above should assist in providing an answer to these questions. Among more detailed points for consideration is the verification and explanation of the occurrence of several species of the same genus at a particular horizon, e.g. three species of ?Rhabdocolpus in the Crackers Bed of the Lower Aptian in the Isle of Wight. There are also certain elements of doubt concerning the generic identi- fication of several species that are considerably larger than the normal size of the genus to which they are assigned. In classifying Cretaceous gastropods ‘size’ is frequently used as a guide, in a subconscious and perhaps, quite illogical manner. If some disparity occurs, the correctness of the determination begins to be doubted. This is certainly true for the Upper Cretaceous specimens of Nerineopsis described in the paper and also for the Upper Albian species Sezla iglalz. A better understanding of the relationship of Tertiary and Recent gastropods is to a large extent dependent upon earlier forms. The origin and early evolution of many Recent prosobranch families occurred during the Cretaceous period and the past neglect and inadequate descriptions of gastropods of this age needs to be remedied. It is hoped that this paper will provide some of the preliminary data necessary for 170 BRITISH CRETACEOUS GASTROPODS a more detailed and sophisticated appraisal of Cretaceous gastropods, and also that it will serve to pose a few of the questions needing investigation before such a study can be undertaken. VI. REFERENCES Aspass, H. L. 1961. The evolutionary characters of certain turreted gastropod families and their significance in ontogeny and phylogeny. A review. J. Geol. u. Avab Repub, Cairo, 5, 1, 1961 : 39-54. 1962. The English Cretaceous Turritellidae and Mathildidae (Gastropoda). Bull. Brit. Mus. (Nat. Hist.), Geol., London, 7, 6, 1962 : 173-196, 5 pls. 1963. A monograph on the Egyptian Cretaceous gastropods. Monogr. geol. Surv. un. Arab Repub., Cairo, Pal .Ser., 2, 1963 : xii, 146, 12 pls. Apams, A. 1848. The zoology of the voyage of H.M.S. Samavang, under the command of Captain Siv E. Belcher . . . 1843-1846. London, 1848-1849. 1861. On some new species of Eulima, Leiostvaca and Cevithiopsis from Japan. Amn. Mag. nat. Hist., London, 3, 7, 1861 : 125-131. ApaAMs, A. & REEVE, L. A. 1848. ‘Mollusca’ in Adams, A. 1848-1849. Aiison, E. C. 1955. Middle Cretaceous gastropoda from Punta China, Baja California, Mexico. J. Paleont., Tulsa, 29, 3, 1955 : 400-432, 5 pls. ARKELL, W. J. 1941. The gastropods of the Purbeck Beds. Q. Jl. geol. Soc., London, 97, 1, 1941 : 79-128. ASCHER, E. 1906. Die Gastropoden, Bivalven und Brachiopoden der Grodischter Schichten. Beitr. Paléont. Geol. Ost-Ungans, Wien. 19, 2-3, 1906 : 135-172, 3 pls. BAYLE, E. 1884. in FISCHER, P. 1880-1887. Bennison, G. N. & Wricut, A. E. 1968. The geological history of the British Isles. Heffer, Cambridge, 1968 : x, 406. Binkuorst, J. T. van den B. 1861. Monographie des gastéropodes et des cephalopodes de la craie supérieure du Limbourg. 4° Miller Bros. Maastricht, 1861 : vi, 44, 18 pls. Brake, J. F. 1902. List of the types and figured specimens vecognized by C. D. Sherborn, F.G.S., in the collection of the Geological Society of London. Geol. Soc. Lond., 1902 : 100, XxXxii. BRONGNIART, A. 1810. Sur les terrains qui paroissent avoir formés sous l’eau douce. Ann. Mus. Hist. nat., Paris, 15, 1810 : 357-405. see under 1822, CUVIER, G. Buvienigr, A. 1852. Statistique géologique, minévalogique, minévallurgique et paléontologique du département de la Meuse. Atlas : 52 pp., 32 pls., Paris. CasEy, R. 1961. The stratigraphical paleontology of the Lower Greensand. Paleontology, London, 3, 4, 1961 : 487-621, 8 pls. Corticnon, M. 1949. Recherches sur les faunes Albiennes de Madagascar. 1. L’Albien d’Ambarimanga. Ayn. géol. Serv. Mines, Madagascar, Paris, 16, 1949 : 128, 22 pls. CossmMann, M. 1889. Catalogue illustré des coquilles fossiles de l’Eocene des environs de Paris. V & Suppl. Annis. Soc. v. malacol. Belg., Bruxelles, 24, 1889 : 3-385, 12 pls. 1892. Revue de paléontologie pour l’année 1891. Amn. géol. univ. Rev. géol. paléont., Paris, 8, 1892 : 718-721. 1893. Revision sommaire de la faune du terrain Oligocene marin aux environs d’Etampes (suite). J. conchyliol., Paris, 41, (Ser. 3, 33), 1893 : 203-363, pl. Io. 1894. Revue de paléontologie pour l’année 1892. Amn. géol. univ. Rev. géol. paléont., Paris, 9, 1894 : 763-768. 1895-1925 Essais de Paléoconchologie comparée. Paris. 13 vols. [Volume 7, 1906. Includes Cerithiacea]. 1902. Rectifications de nomenclature. Revue crit. Paléozool., Paris, 6, 3, 1902 : 160-162. 1906. See Essais de paléoconchologie comparée. Vol. 7 : 261, 15 pls. —— 1913. Contribution a la paléontologie frangaise des terrains Jurassique. III. Cerithiacea et Loxenematacea. Mem. Soc. géol. Fr., Paléont., Paris, 46, (19), 3-4 : 88, 4 pls; 46, (20), 3-4 : 89-264, 7 pls. PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 171 CotTEAu, G. 1854. Paléontologie de l)Yonne. Prodrome des mollusques. Gasteropodes. Bull. Soc. Sci. hist. nat. Yonne, Auxerre, 8, 1854 : 201-231. Cox, L. R. 1930. The fossil fauna of the Samana Range and some neighbouring areas. Part IV. Lower Albian Gastropoda and Lamellibranchia. Mem. geol. Surv. India, Calcutta, NS 15, 1930 : 39-49, 1 pl. —— 1960. The British Cretaceous Pleurotomariidae. Bull. Brit. Mus. nat. Hist., (Geol.), London, 4, 8, 1960 : 385-423, 17 pls. 1965. Jurassic bivalvia and gastropoda from Tanganyika and Kenya. Bull. Brit. (Nat. Hist.), London, Suppl. 1, 1965 : 208. Cuvier, G. L.C. F. D. & Bronentart, A. 1822. Description géologique des environs de Paris. 4° Paris, 1822 : 428, 16 pls. Datt W. H. 1889. Report on the Mollusca II. Gastropoda (Report on dredgings.... by U.S. Coast Survey Steamer Blake....). Bull. Mus. comp. Zool. Harvard, Cambridge, Mass., 18 : 1-492. Davis, A. G. 1923. Contributions to the geology and palaeontology of the Croydon regional survey area. No. 1. The Lower Chalk, zone of Ammonites (Schloenbachia) varians. Proc. Tvans. Croydon Nat. Hist. Soc., 9, 1923 : 125-136. DELPEY, G. 1948. Gastéropodes mésozoiques de l’ouest de Madagascar. Ann. géol. Surv. Mines, Madagascar, Paris, 15, 1948 : 25, 7 pls. DESHAYES, G. P. 1824-1837. Description des coquilles fossiles des environs de Paris, Paris, Vol. 2 : 1-178 (1824), 179-306 (1833), 307-434 (1834), 435-562 (1835), 563-690 (1836), 691-814 (1837). 1830-1832. Encyclopédie méthodique ou par ordre de matiéves histoive naturelles des Vers et Mollusques. 3 vols., Paris, 1830-1832 : 1152. 1842. In Leymerie, A. 1864-1865. Description des animaux sans vertebves du Bassin de Paris. Paris, Vol. 3 : I~ 200 (1864), 201-656 (1865). Dovuvitte, H. 1904. Mission scientifique en Perse. J.de Morgan. Vol. 3, Etudes géologiques, Partie IV, Paléontologie, Mollusques fossiles. 4° Paris, 1904 : 191-380. 25 pls. EuprEs-DrEsLonccHamps, E. 1842a, Mémoire sur les cérites fossiles des terrains secondaire du Calvados. Mém. Soc. linn. Normandie, Caen, 7, 1842 : 189-214. 1842b, Mémoire sur les melanies fsssiles des terrains secondaires du Calvados. Mém. Soc. linn. Normandie, Caen, 7, 1842 : 215-232, pls 11-12. FIscHER, P. 1880-1887. Manuel de conchyliologie et de paleontologie conchyliologique. 8° F.Savy, Paris, 1880-1887 : xxv, 1369, 23 pls. Fitton, W. H. 1836. Observations on some of the strata between the Chalk and the Oxford Oolite in the southeast of England. Tvans. Geol Soc., London, 4, 2, 1836 : 103-389. FLEMING, J. 1828. A history of British animals exhibiting the descriptive characters and systematical avvangement of the geneva and species of . . . of the United Kingdom. 8° Edinburgh and London, 1828 : xxiii, 565. Forbes, C. 1960. Field meeting in the Cambridge district. Pvoc. geol. Ass., London, 71, 2 : 234-241. ForBes, E. 1845. Catalogue of Lower Greensand fossils in the Museum of the Geological Society with notices of species new to Britain. Q. J1 geol. Soc., London, 1, 1845 : 237-250, 345-355, 4 pls. Fric, A. 1894. Studien im Gebiete der Béhmischen Kreideformation Palaeontologische Untersuchungen der einzelnen Schichten v. Priesener Schichten. Aychiv. Naturw. Land. Durchforsch., Bohm., Prag, 9, 1, 1894 : 1-134. GeInitz, H. B. 1874. Das Elbthalgebirge in Sachsen. Pt. 2. Der mittlere und obere Quader, V. Gasteropoden und Cephalopoden. Palaeontographica, Stuttgart, 20, abt. 2, 1874 : 161-198, pls 29-36. GMELIN, J. F. 1788-1793. C.alLinné.. . ‘Systema Naturae .. .’,3 vols. 8° Lipsiae, 13th edition, 1788-1793 (in 7 volumes). 172 BRITISH CRETACEOUS GASTROPODS GoipFuss, G. A. 1844. Petvefacta Germaniae, Vol. 3, Folio, Dusseldorf, 1844 : 128, pls 166— 200. Grey, J. E. 1847. A list of the genera of recent mollusca, their synonyms and types. Pyoc. zool. Soc., London, 15, 1847 : 129-219. GUERANGER, M. E. 1867. Album paléontologique du Département de la Sarthe veprésentant au moyen de la photographie les fossiles veceuilles dans cette civconscription. Folio, Le Mans, 1867 : 20, 25 pls. Haas, O. 1953. Mesozoic invertebrate faunas of Peru. Bull. Am. Mus. nat. Hist., New York, 101, 1953 : xi, 328, 18 pls. HOoLzapFEL, E. 1887-1888. Die Mollusken der Aachener Kreide. Palaeontographica, Stuttgart, 34, 1887-1888 : 29-180, pls 4-20. HUCKRIEDE, R. 1967. Molluskenfaunen mit limnischen und brackishen Elementen aus Jura, Serpulit und Wealden N.W.-Deutschlands und ihre palaogeographische Bedeutung. Beth. geol. Jb., Hannover, 67, 1967 : 263, 25 pls. INTERNATIONAL COMMISSION ON ZOOLOGICAL NOMENCLATURE. Inteyvnational code of zoological nomenclature. Int. Trust. Zool. Nomen., London, 1961 : 176. Ivanova, A. N. 1959. [Bivalves, gastropods and belemnites in the Jurassic and Cretaceous deposits of Saratov on the Volga]. Tvudy VNIGRI, Leningrad, 137, 1959 : 267-460, 27 pls. [In Russian]. Jupp, J. W. 1868. On the Speeton Clay. Q. Jl geol. Soc., London, 24, 1868 : 218-250. JuKEs-Browne, A. J. 1900. The Cretaceous vocks of Britain. Vol. 1. The Gault and Upper Greensand of Britain. Mem. Geol. Surv. Gt. Britain, H.M.S.O., London : xiv, 499. KAUNHOWEN, F. 1898. Die Gastropoden der Maestrichter Kreide. Paldont. Abh., Berlin, 8, 1898 : 132, 13 pls. KENNEDY, W. J. 1969. The correlation of the Lower Chalk of South-East England. Proc. Geol. Ass., London, 80, 4, 1969 : 459-560. KIRKALDY, J. F. 1939. History of the Lower Cretaceous period in England. Pyvoc. Geol. Ass., London, 50, 3, 1939 : 379-417. 1963. The Wealden and marine Lower Cretaceous beds of England. Pyoc. Geol. Ass., London, 74, 2, 1963 : 127-146. Ktun, O. 1947. Zur Stratigraphie und Tektonik der Gosauschichten. Sitzber. Akad. Wiss., Wien, 156, Abt. I, 1947 : 181-200. Lamarck, J.B.P.A. de M. de. 1802-1806. Mémoive sur les fossiles des envivons de Paris, comprenant la détevmination des espéces qui appartiennent aux animaux marins sans vertébres, . . . Ann. Mus. Hist. Nat., Paris, 1802-1806 : 284. LampLuGu, G.W. 1889. On the subdivisions of the Speeton Clay. Q. Jl geol. Soc., London, 45, 4, 1889 : 575-618. LEYMERIE, A. 1842. Mémoire sur le terrain Crétacé du Département de |’Aube contenant des considerations générales sur le terrain Néocomien. Mem. Soc. geol. Fr., Paris, 5, 1, 1842: 127, 18 pls. LinNaEus, C. 1788-1793. Systema Naturae, sive regna tria naturae systematice proposita per classes, ordines, genera and species. Editio decima tertia, aucta, reformata, cura J. F. Gmelin, 3 Vols. [in 7] illust. 8° Lipsiae. 1788-1793. Locarp, A. 1903. Les Cerithium et les Cerithiidae des mers d’Europe. Anmnls. Soc. Agric. Lyon, 10, 1903 : 95-128. Loriot, P. de. 1874-1875. Description des fossiles. In Loriot, P de and Perxtrat, E. Monographie paléontologique et géologique des étages supérieurs de la formation jurassique des environs de Boulogne-sur-Mer. Mem. Soc. Phys., Genéve, 23, 1874 : 261-407, I0 pls, 24, 1875 : 1-326, 16 pls. Loven, S. L. 1846. Index Molluscorum litora Scandinaviae occidentalis habitantium. Faunas prodromum. Ofvers. K. VetenskAkad. Forh., Stockholm, 3 1846 : 135-160, 183-204. MIcHELIN, H. 1823. In DESHAYES, G. P. 1824-1837. 1838. Sur une argile dépendant du Gault, observée au Gaty Commune de Gérodot, Departement de l’Aube. Mem. Soc. geol. Fr., Paris, 3, 5, 1838 : 97-103, 1 pl. PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 173 Moore, R. C. and others. 1968. Developments, trends, and outlooks in Palaeontology. J. Paleont., Tulsa, 42, 6, 1968 : 1327-1377. Murer, J. 1851. Monographie der Petrefacten der Aachener Kreideformation. Zweite Abteil. 8° Bonn, 1851 : 88, 4 pls. Munster, G. 1844. In Gotpruss,G. A. 1844. Vol. 3. Nacao, T. 1934. Cretaceous mollusca from the Miyako District, Honshu, Japan. (Lamellibranchiata and Gastropoda). J. Fac. Sci. Hokkaido Univ., Ser. IV. Geol. & Min., Sapporo, 2, 3, 1934 : 177-277, 10 pls. Oprtnton 417, I.C.Z.N. 1956. Rejection for nomenclatorial purposes of volume 3 (Zoologie) of the work by Lorenz Oken entitled “Okens Lehrbuch dey Naturgeschichte’ published in 1815-1816. Opin. & Decl. 1.C.Z.N., London, 14, 1, 1956 : 1-42. Oxen, L. V. 1815-1816. Okens Lehrbuch der Naturgeschichte, Vol. 3 (Zoologie). Jena, Abt. 1 : xxviii, 842, 40 pls; Abt. 2 : xvi, 1270. Orpicny, A. D. 1842-1843. Paléontologie Francaise, terrains crétacés 2. Gasteropodes. 8° Paris : 456, pls 149-2306. 1850. Pyrodvome de paléontologie stratigvaphique universelle des animaux mollusques et vayonnés. Vol. 2. 12° Paris : 428. PéHELINCEV, V. 1927. [The Gault gastropoda from Sokolova Gora near Saratov]. Izv. geol. Inst., Petvogrvad Bull. Com. Geol. Leningrad 45, 9, (49), 1926 : 991-999, 1 pl. [In Russian] 1927a. [The Jurassic and Lower Cretaceous fauna of the Crimea and the Caucasus. Gastropoda of the Cretaceous sandstones in the neighbourhood of Balaclava]. Mém. Com. Géol., Leningrad, 172, 1927 : 148-177, 3 pls. [English summary]. 1953. [Gastropod fauna of the Upper Cretaceous strata in Transcaucasia and Central Asia}. Izd. Akad. Nauk. SSSR, Moscow, Geol. Mus. A.P. Karpinskogo, Ser. Monograph, 1, 1953 : 391, 51 pls. [In Russian]. PCéHELINCEV, V. F. & Korospxov,I. A. 1960. In Or tov, Y. A. [Edit.], ‘Osnovoi Paleontologii, Part 4—Gastropoda’. Izd Akad. Nauk Moscow, 1960 : 360, 28 pls. Peron, M. 1899-1900. Etudes paléontologiques sur les terrains du Département de L’ Yonne. Cephalopodes et gastropodes de l’étage Néocomien. Bull. Soc. Sci. hist. nat. Yonne Auxerre, 53, 3, 1899-1900 : 67-219, 4 pls. Pictet, F. J. & CampicHE, G. 1861-1864. Description des fossiles du terrain Crétacé des environs de Sainte-Croix. Mater. Paléont. Suisse, Geneve, 3, 1861-1864 : 752, 55 pls. Pictet, F. J. & RENEvIER, E. 1854-1858. Descriptions des fossiles du terrain Aptien de la Perte-du-Rhéne et des environs de Sainte-Croix. Mater. Paléont. Suisse, Genéve, Ser. 1, No. 1, 1854-1858 : 184, 23 pls. Pictet, F. J. & Roux, W. 1847-1853. Description des mollusques fossiles qui se trouvent dans les Grés Verts des environs de Genéve. Mém. Soc. phys. Hist. nat. Genéve, 11, 2 : 257— 412; 12 : 157-287; 13 : 73-173, 489-558, 51 pls. Price, F.G.H. 1874. On the Lower Greensand and Gault of Folkestone. Pyoc. Geol. Ass., London, 4, 1874 : 135-150. 1875. On the Gault of Folkestone. Q. Jl. geol. Soc., London, 30, 1875 : 342-366. 1879. The Gault. 8°. London : viii, 81. Quaas, A. 1902. Beitrage zur Kenntniss der Fauna der obersten Kreidebildung in der libyschen Wiiste. II. Die Fauna der Overwegischichten und der Blatterthone in der libyschen Wiiste. Palaeontogvaphica, Stuttgart, 30, Part 2, 1902 : 153-334, 33 pls. Ravn, J. P. J. 1921. Kridtaflejringerne paa Bornholms Sydvestkyst og deres Fauna. III. Senonet. Mus. Minér. Geol. Univ. Copenhague, Commun. Paléont., No. 17, 1921 : 52, 3 pls. Rayner, D. 1967. The stvatigvaghy of the British Isles. Camb. Univ. Press, Cambridge, 1967 : X, 453. Reuss, A. E. 1845. Die Versteinerungen dev Bohmischen Kreideformation. 4°, Stuttgart, 1845 : iv, 148, 51 pls. Roemer, F. A. 1840-1841. Die Versteinerungen des Norddeutschen Kreidegeberges. 4°, Hannover, 1840-1841 : 145, 16 pls. 174 BRITISH CRETACEOUS GASTROPODS Roman, F. & MazErRANn, P. 1920. Monographie paléontologique de la faune du Turonien du Bassin d’Uchaux et de ses dépendances. Avychiv. Mus. Hist. nat. Lyon, 12, 2, 1920 : 137, 11 pls. Sars,G.O. 1878. Mollusca vegionis arcticae Norvegiae Oversight over dei Norges arktiske vegion fovekommende Bloddyy. Christiana, 1878 : 466, 52 pls. SCHUMACHER, H.C. F. 1817. Essai d’un nouveau systéme des habitations de vers testacés. 4°, Copenhague. 1817 : iv, 287, 22 pls. SepGwick, A. & Murcuison, R. I. 1833. A sketch of the structure of the Eastern Alps. Trans. geol. Soc., London, 2nd Ser., 3, 2, 1833 : 301-420, pls 35-40. SEELEY, H.G. 1864. Cuttings from a note-book on Chalk gastropods. The Geologist, London, 7, 1864 : 89-93, 1 pl. SHARMAN, G. & NrEwtTon, E. T. 1896. Note on some Cretaceous fossils from the drift of Moreseat, Aberdeen. Geol. Mag., London, Dec. IV, 3, 1896 : 247-254. Sout, N. F. 1960. Archaeogastropoda, Mesogastropoda and stratigraphy of the Ripley, Owl Creek, and Prairie Bluff Formations. Late Cretaceous gastropods in Tennessee and Mississippi. Prof. Pap. U.S. Geol. Suvv., Washington 331-A, 1960 : 151, 18 pls. 1964. Gastropods from the Coffee Sand (Upper Cretaceous) of Mississippi. Pyvof. Pap. U.S. Geol. Surv., Washington, 331-C, 1964 : 341-394, Pls 53-57. 1968. Post-Paleozoic Gastropoda. In R. C. Moore and Oruers 1968. J. Paleont., Tulsa, 42, 6, 1968 : 1363-1364. 1969. Cretaceous biotic provinces delineated by gastropods. [Abstract]. J. Paleont., Tulsa, 43, 3, 1969 : 899. SoweERBy, J. & SowERBY, J.deC. 1812-1846. The Minerval Conchology of Great Britain. 8°, London, 7 Vols, pls 1-383 (1812-1822) by J. Sowerby; pls 384-648 (1823-1846) by J. de C. Sowerby. SoweERBy, J.deC. 1833. Plates and Plate Descriptions in SepGwick, A. & Murcuison, R. I. 1833. SowErRBy, J. de C. 1836. Descriptive notes respecting the shells figured in plates 11-23. In Fitton, W. H. 1836. [see under Fitton, W. H. 1836]. STEPHENSON, L. W. 1941. The larger invertebrate fossils of the Navarro Group of Texas. Bull. Univ. Texas. Publ. Bur. Econ. Geol. Austin, 4101, 1941 : 641, 95 pls. 1952. Larger invertebrate fossils of the Woodbine Formation (Cenomanian) of Texas. Prof. Pap. U.S. Geol. Surv., Washington, 242, 1952 : 226, pls 8-59. Stoticzka, F. A. 1867-1868. The Cretaceous fauna of Southern India. 2. The Gastropoda of the Cretaceous rocks of Southern India. Mem. geol. Surv. India, Calcutta, Palaeont. Indica, Ser. V. 2, 13, 1867-1868 : xiii, 500, 28 pls. Tate, R. 1869. Contributions to Jurassic palaeontology. 1. Cvryptaulax, a new genus of Cerithiadae. Ann. Mag. nat. Hist., London, Ser. 4, 4, 1869 : 417-419. Tryon, G. W. 1887. Manual of conchology, structural and systematic, Vol. 9, Philadelphia, 1887 : 488, 71 pls. VERNEUIL, P. E. P. de & LoriErRE, G. de. 1868. Description des fossiles du Néocomien supérieur de Utrillas et ses envivons (Province de Tervel). 4°, Le Mans. 1868 : vii, 30, 3 pls. VERRILL, A. E. 1882. Catalogue of marine molluscs added to the fauna of New England during the past ten years. Tvans. Conn. Acad Arts Sci., New Haven, 5, 1882 : 447-587, 5 pls. Vipat, L. M. 1917. Nota paleontolégica sobre el Cretaceo de Catulufia. Assoc. espan. Progr. Cienc., Congr. Sevilla, Barcelona, 5, 1917 : 19, 3 pls. Wane, B. 1917. New and little-known Gastropoda from the Upper Cretaceous of Tennessee. Proc. Acad. nat. Sci., Philadelphia, 69, 1917 : 280-304, 3 pls. 1926. The fauna of the Ripley formation of Coon Creek, Tennessee. Prof. Pap. U.S. geol. Surv., Washington, 137, 1926 : 272, 72 pls. WANNER, J. 1902. Die Fauna der obersten weissen Kreide der libyschen Wiiste. Palaeonto- graphica, Stuttgart, 30, Pt. 2, 1902 : 91-152, 7 pls. PROCERITHIIDAE, CERITHIIDAE & CERITHIOPSIDAE 175 Wenz, W. 1938-1944. Handbuch dev Paldozoologie. Band 6 Gastropoda, in 7 vols., 8°, Berlin, 1, 1938; 2, 1938; 3, 1939; 4, 1940; 5, 1941; 6, 1943; 7, 1944 : 1639 pp, 4211 figs. WoLLEemMaANN, A. tI900. Die Bivalven und Gastropoden des deutschen und hollandischen Neocoms. Abh. K. Preuss. geol. Landes, Berlin, 31, 1900 : 180, 8 pls. 1903. Die Fauna des mittleren Gaults von Algermissen. Jahrb. K. Preuss. geol. Landes, Berlin, 24, 1903 : 22—42, 2 pls. 1906. Die Bivalven und Gastropoden des nordeutschen Gaults (Aptiens und Albiens). Jahrb. K. Preuss. geol. Landes, Berlin, 27, 1906 : 259-300, 5 pls. 1908. Nachtrag zu meinen Abhandlungen tiber die Bivalven und Gastropoden der Unteren Kreide Norddeutschenlands. Jahrb. K. Preuss. geol. Landes, Berlin, 29, Pt 2, 1908 : 151-193, 5 pls. Woop, C. J. 1969. Notes on the sedimentology of the Chalk. In Worssam, B. C. & Taytor, J. H. ‘Geology of the country around Cambridge’. Mem. Geol. Surv. G. B., Sheet 188, New Series, H.M.S.O., London, 1969 : 43-45. Woops, H. 1896. The mollusca of the Chalk Rock, Pt 1. Q. Jl. geol. Soc., London, 52, 1, 1896 : 68-08, 3 pls. WricLey, A. 1940. Some Eocene mollusca, with descriptions of new species. Pyoc. Malac. Soc., London, 24, 1. 1940 : 6-17. ZEKELI, F. 1852. Die Gastropoden der Gosaugebilde. Abh. K.K. geol. Reichsanst., Wien, 1, 2, 1852 : 124, 24 pls. H. L. ApsBass Dept. of Geology, FACULTY OF SCIENCE, Arn SHAMS UNIVERSITY, CAIRO. F PLATE x Fics. 1 & 2. Nerineopsis melburiensis sp. nov. Lower Cenomanian, Chloritic Marl, Melbury Park, nr. Shaftesbury, Dorset. Holotype GSM 93752 Fig. 1. Apertural view. X2°5. Fig. 2. To show ornament of spiral cords. X2°5. Fics 3 & 4. Nerineopsis cuckhamsliensis (Woods) Fig. 3. Holotype, Sedgwick Museum B 4443, Chalk Rock, Turonian, Cuckhamsley, Wilts. Xe Fig. 4. BM(NH) G 48978, Chalk Rock, Holaster planus Zone, Hill End Pit, nr. Hitchin, Herts. 3:5. Silicone rubber mould from natural external shell impression. Fics 5-8. Nerineopsis aculeatum (Sharman & Newton), 1896 Fig. 5. BM(NH) G 55220, To show later ornament. x6. Fig. 6. BM(NH) G 55191, To show general ornament. x6. Fig. 7. BM(NH) G 55222, Lectotype. x6. All from Speeton Clay, Speeton, Yorks., G.W. Lamplugh Coll’n. Fig. 8. GSM 93749, General ornament and columella. 4. Speeton Clay, Speeton, Yorks. Fics9 & 12. Nerineopsis subattenuatum (d’Orbigny) Fig.9. GSM 93687. x3. Fig. 12. BM(NH) GG 5610, C.W. Wright Coll’n. Both from Lower Aptian, Forbesi Zone, Crackers Bed, Lower Greensand, Chale Bay, Isle of Wight. Fics 10 & 11. Nerineopsis coxi sp. nov. Gault, Folkestone, Kent, J.S. Gardner Coll’n. Fig. 10. BM(NH) GG 6474. x4. Fig. 11. BM(NH) GG 6475, Holotype. x4. Fic. 13. Nerineopsis claxbiensis sp. nov. Claxby Beds, Lower Hauterivian, Nettleton, Lincs. BM(NH) GG 5612, Holotype. x4. Fic. 14. Nerineopsis adeli sp. nov. Upper Chalk, Springbank, nr. Coagh, County Derry, Northern Ireland. GSM 93753, Holotype. x3. Bull. Br. Mus. nat. Hist. (Geol.) 23, 2 IPIL ANID, i PATE ee Fics 1 & 2. Cirsocerithium subspinosum (Deshayes) Gault, Albian, Folkestone, Kent. Fig. 1. Sedgwick Museum B 32530. x5. Fig. 2. BM(NH) GG 20746, (specimen is now slightly damaged). » 6. Fics 3 & 4. Cirsocerithium nooryi sp. nov. Lower Chalk, Cenomanian, Ditchling, Sussex. BM(NH) 98208, Holotype. x3. Fic. 5. Cimolithium aff. eleanorae Allison. Cenomanian, Charlton Bay, Devon. BM(NH) G 74550. x2. Fic. 6. Cimolithium ascheri Wollemann. Neocomian, Speeton Clay, Speeton, Yorks. BM(NH) G 67942. x2. Fic. 7. Cimolithium cf. ascheri Wollemann. Barremian, Cementstone Group, Cayton Bay, Yorks. BM(NH) GG 20820. x7. Fic. 8. ? Cirsocerithium kirkaldyi sp. nov. Ferruginous Sands, Lower Greensand, Aptian, Shanklin, Isle of Wight. BM(NH) GG 5603 Holotype. » 4. Fics 9 & 10. Cimolithium ascheri Wollemann. Holotype 414-415, x 2, in Geologisch-Palaeontologisches Institut, Georg-August Universitat, Gottingen; Neocomian, Bohnenkamp nr. Querum, Sarstedt, North Germany. Fics 11 & 12. Nudivagus morrisi sp. nov. Ferruginous Sands, Upper Aptian, ? Group xiv, Lower Greensand, Shanklin, Isle of Wight. Holotype, Geological Survey Museum 2252. 2. Fig. 11. Silicone rubber mould from natural impression of external ornament. Note growth lines. Fig. 12. Latex rubber mould from same specimen, which shows spiral threads. Bull. By. Mus. nat. Hist. (Geol.) 23, 2 PLATE 3 Fic. 1. Bathraspira tecta (d’Orbigny) BM(NH) GG 6227. Silicone rubber cast of external mould preserved in nodule from the Regularis Subzone, Tarde- furcata Zone, Albian, Arnold’s Pit, Billington Crossing, Leighton Buzzard, Beds. C.W. Wright Collection. x4. Fic. 2. Bathraspira ? tecta (d’Orb) BM(NH) GG 20790, Gault, Albian, Folkestone, Kent. J.S. Gardner Collection. 6. Fic. 3. Bathraspira shanklinensis sp. nov. BM(NH) G 20956 Lower Greensand, Ferruginous Sands, ? Perna Bed, Aptian, Redcliff, Sandown, Isle of Wight. Caleb Evans Collection. x2. Fics 4-6. Bathraspira shanklinensis sp. nov. BM(NH) GG 5602, Ferruginous Sands, Lower Greensand, Upper Aptian, Shanklin, Isle of Wight. Holotype. C. W. Wright Collection. Euger4n | <4e Biges55 <3: lene, ©, <3}. Fic. 7. Bathraspira fouadi sp. nov. Paratype, Sedgwick Museum B 32524, Gault Albian, Folkestone, Kent, Wiltshire Collection. x3. Fic. 8. Bathraspira fouadi sp. nov. BM(NH) G 71562 Gault, Albian, Folkestone, Kent. x 3. Fic. 9. Bathraspira brightoni sp. nov. BM(NH) 34843 Upper Greensand, Blackdown, Devon. 3 approx. Fic. 10. Bathraspira cleevelyi sp. nov. BM(NH) GG 5607, Holotype ? Regularis Subzone, Tardefurcata Zone, Lower Albian, Leighton Buzzard, Beds., or Bed 5, Ford Place, nr. Wrotham, Kent. C.W. Wright collection. Silicone rubber cast of external mould in nodule. 2°5. Fics 11 & 15. Bathraspira cf. brightoni sp. nov. BM(NH) GG 20791 Gault, Albian, St. Florentin, Yonne, France. Fig. 11 shows strength of spirals and indications of growth lines. x 4. Fig. 15 shows spiral ornament. » 4. Fic. 12. Bathraspira fouadi sp. nov. BM(NH) G 73795 Hoiotype. Gault, Albian, Folkestone, Kent. J. S. Gardner Collection. 5. Fic. 13. Bathraspira fouadi sp. nov. ? var. BM(NH) G 73797 Gault, Albian, Folkestone, Kent. J.S. Gardner Collection. x2. Fic. 14. Bathraspira brightoni sp. nov. BM(NH) G 73796 Holotype. Gault, Albian, Folkestone, Kent. J. 5S. Gardner Collection. <3 approx. Bull. Br. Mus. nat. Hist. (Geol.) 23, 2 PLATE 3 PILATE 4 Fics 1 & 6. Metacerithium aff. trimonile (Michelin) Gault, Albian, Folkestone, Kent. Fig. 1. BM(NH) G 73793. 3:25. Fig. 6. Geological Survey Museum. 93725. x3. Fics 2, 3, 7,8 & 10. Metacerithium trimonile (Michelin) Gault, Albian, Folkestone, Kent. J. S. Gardner Collection, BM(NH) Fig. 2. G 73794. X3- Eilon 3G G647 2x 3 Higa 7p. GIO“ Os: leis fo, (GG Ogi, <3}. Fig. 10. GG 6476. x4. Fic. 9. Metacerithium ? sp. nov Sedgwick Museum, B 32530, Albian, Gault, Folkestone, Kent. 3. Fic. 4. Metacerithium aff. ornatissimum (Deshayes) Sedgwick Museum B 32518, Gault, Albian, Folkestone, Kent. x3. Fic. 5. Metacerithium ornatissimum (Deshayes) BM(NH) GG 5601. C. W. Wright Collection, Cenomanian, Schloenbachia varians Zone. Aston Clinton, Bucks. x3. Fic. 11. Metacerithium aff. ornatissimum (Desh.) Geological Survey Museum 1698; Gault, Albian, Folkestone, Kent. 3. Fic. 12. Metacerithium ornatissimum (Desh.) BM(NH) G 28908. x3. Cenomanian varians Zone. Bed.1. Merstham, Surrey. Collected. A. G. Davis & Ref’d. Davis (1923). Bull. By. Mus. nat. Hist. (Geol.) ye PLATE 4 PLATE 5 Fics 1-5 ? Rhabdocolpus forbesianum (dOrbigny) Lower Greensand, Crackers Bed, Lower Aptian, Forbesi Zone, Atherfield, Isle of Wight. Figs 1-4. Stereoscan. Fig. 1. BM(NH) GG 20850. x20. Fig. BM(NH) GG 20848, showing the ornament of the early whorls 23. Fig Geological Survey Museum 2275, Lectotype from type series. 23. 2. Be Fig. 4. BM(NH) GG 20845. 56. Fig. 5. BM(NH) GG 5609. 5. Fic. 7. ? Rhabdocolpus melvillei sp. nov. Lower Greensand, Crackers Bed, nr. Atherfield Point, Isle of Wight. Holotype, Sedgwick Museum B 27340, Wiltshire Collection. 5. Fics 6, 12 & 13. 2? Rhabdocolpus ? clementinum (d’Orb) Lower Greensand, Crackers Bed, Aptian, Atherfield, Isle of Wight. Stereoscan. x 23. Fig. 6. BM(NH) GG 20841. Fig. 12. BM(NH) GG 20842. Fig. 13. BM(NH) GG 5694. Fics 8-11. Metacerithium turriculatum (Forbes) Lower Greensand, Crackers Bed, Atherfield, Isle of Wight. Fig. 8. BM(NH) G 74602. x5. Figs 9 & 11. BM(NH) GG 20814, J. E. Lee Collection. x5, x7. Fig. 10. Geological Survey Museum 2253, Specimen on left is the Holotype. 3. nu a) a a (e) o S ~ az) ~ S = = a x 3 ~~ | BS ica) Ss = isa) Pe ATI 6 Fics 1 & 2. ? Campanile cenomanica sp. nov. BM(NH) GG 18686. ? Cenomanian, Dorset. x 2. Fics 3 & 4. Uchausxia badri sp. nov. Sedgwick Museum B 44632. Upper Greensand, Albian, Blackdown, Devon. x6. Fics 5 & 6. Uchausxia wisei sp. nov. C. W. Wright Collection. Claxby Ironstone, Valanginian-Hauterivian, Nettleton, Lincs. Fig.5. BM(NH) GG 5611. x3°5. Fig. 6. BM(NH) GG 6161 Holotype. 3. Fic. 7. Uchausia sp. BM(NH) G 73788. Gault, Albian, Folkestone, Kent. J.S. Gardner Collection. x6. Fic. 8. Cerithiella praelonga (Deshayes) BM(NH) 72024. Middle Eocene, Bracklesham Beds, Bracklesham Bay, Hants. 5. Fic. 9. Cerithiella devonica sp. nov. Sedgwick Museum B 44633. Upper Greensand, Albian, Blackdown, Devon. x5. Fic. 10. Cerithiella atherfieldensis sp. nov. Sedgwick Museum B 27334. Lower Greensand, Crackers Bed, Lower Aptian, Forbesi Zone, Atherfield, Isle of Wight. 3. Fic. 11. Seila iglali sp. nov. BM(NH) G 16084. Upper Greensand, Albian, Combe Raleigh Hill, nr. Honiton, Devon. W. Vicary Collection. x 4. Fic. 12. Orthochetus hantoniensis sp. nov. BM(NH) G 70130. Aptian, Atherfield, Isle of Wight. P. Cambridge Coilection. 5. Fics 13 & 14. ‘Orthochetus’ helmyi sp. nov. Gault, Albian, Folkestone, Kent. x 3. Fig. 13. BM(NH) G 10500, Holotype. Fig. 14. BM(NH) G 71988. Fic. 15. ‘Orthochetus’ sp. nov. BM(NH) 9145, Lower Greensand, Sandgate Beds, Upper Aptian, Parham Park, Sussex. G. A. Mantell Collection. x 3. Bull. By. Mus. nat. Hist. (Geol.) 23, 2 PLATE 6 PLATE 7 Fics 1 & 2. Exechocirsus aff. saundersi (Woods) Cretaceous, Turonian, Chalk Rock, Holaster planus Zone; Hitch Wood Pit, near Hitchin, Herts. Fig. 1. BM(NH) GG 20821; shows variation in ornament i.e. weakly developed tubercles and the presence of intermediate (or secondary) spirals. Both C. W. Wright Collection, and silicone rubber moulds from external impressions. 2. Fics 3-5. Exechocirsus saundersi (Woods) Fig. 3. BM(NH) GG 66724; Turonian, Chalk Rock, Blackwell Hall Farm, Latimer, Bucks. Silicone rubber mould. x 2. Fig. 4. BM(NH) G 67704; Turonian, Chalk Rock, Dover, Kent. A. W. Rowe Collection. Silicone rubber mould. x 4. Fig. 5. Sedgwick Museum B 4459, Holotype; Turonian, Holaster planus Zone, Cuckhamsley, North Farnborough, Berks. Silicone rubber mould. 3. Fics 6-7. Exechocirsus pustulosus (J. de C. Sowerby) 1833. BM(NH) G 17916; Holotype; Upper Cretaceous, Gosau, Austria. TI. Fics 8 & 9. Exechocirsus aff. saundersi (Woods) BM(NH) G 60542-60543. Silicone rubber moulds from impressions of each half of the same shell. No other information available. x3. Fics 10 & 11. Exechocirsus aff. subpustulosus Pchelinceyv Fig. 10. BM(NH) G 16134. Fig. 11. BM(NH) G 73798. Both from the Albian, Haldon, Devon; W. Vicary Collection. x 2. Fics 12-14. Exechocirsus aff. saundersi (Woods) Fig. 12. BM(NH) GG 20823, Hitch Wood Pit, nr. Hitchin, Herts. C. W. Wright Collection. Silicone rubber mould. x 2. Fig. 13. BM(NH) GG 5934. Turonian, Holaster planus Zone, Kiplingcotes, E. Yorks. C. W. Wright Collection. Silicone rubber mould. x 2. Fig. 14. BM(NH) GG 5983. Same locality as above; natural mould. » 4. 7 JPL AD 18 Bull. By. Mus. nat. Hist. (Geol.) 23, 2 PAC aS Fics 1-4 & 6. Ageria gaultina sp. nov. Albian, Gault, Folkestone, Kent. J. S. Gardner Collection. BM(NH) specimens Fig. 1. GG 6473. Holotype, x4. ime, 2, (CG a7 33}. Fig. 3. GG 20742. 3. ? Bedv, Lower Gault (fide Dr. H. G. Owen), shows slight variation in ornament. Fig. 4. GG 20862. Shows the ornament of the earlier whorls. » 4. Fig. 6. GG 20857. x3. Fic. 5. Ageria gaultina sp. nov. Albian, Osmington, Dorset, BM (NH) G49804. 3°5. Fics 7-10. Ageria costata (J. de C. Sowerby) Upper Greensand, Albian; Blackdown, Devon. BM(NH) specimens Fig. 7. 71382. Shows aperture. x4. Fig. 8. 43674. Holotype, Sowerby Collection. 5. Fig. 9. 34844. Also shows aperture. x6. JES, Ti@; (EKG; BOS, >< 3} Fics 12. Ageria costata (J. de C. Sowerby) BM(NH) G 71095, Albian Upper Greensand; Bed 12, Peak Hill, nr. Sidmouth, Devon. H. F. Metcalfe Collection. x 3. Fic. 13. ? Ageria angustata (d Orbigny) BM(NH) GG 20740. Neocomian, Claxby Beds, Nettleton, nr. Caiston, Lincs C. W. Wright Collection. 5. Bull. By. Mus. nat. Hist. BOE: = é = —@ eS A LIST OF SUPPLEMENTS TO THE GEOLOGICAL SERIES OF THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) . Cox, L. R. Jurassic Bivalvia and Gastropoda from Tanganyika and Kenya. Pp. 213; 30 Plates; 2 Text-figures. 1965. 6. . Et-Naccar, Z. R. Stratigraphy and Planktonic Foraminifera of the Upper Cretaceous—Lower Tertiary Succession in the Esna-Idfu Region, Nile Valley, Egypt, U.A.R. Pp. 291; 23 Plates; 18 Text-figures. 1966. {10. . Davey, R. J., Downie, C., SaRGEANT, W. A. S. & Witiiams, G. L. Studies on Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 248; 28 Plates; 64 Text- figures. 1966. {7. . APPENDIX. Davey, R. J., Downie, C., SARGEANT, W. A. S. & WiLLias, G. L. Appendix to Studies on Mesozoic and Cainozoic an meckne Cysts. Pp.ige 1969. 8op. . Ertiott, G. F. Permian to Palaeocene Calcareous Aisi (Dasycladaceae) of the Middle East. Pp. 111; 24 Plates; 17 Text-figures. 1968. £5.12. . RHOovDES, F. H. T., Austin, R. L. & Druce, E. C. British Avonian (Carboni-_ ferous) Conodont faunas, and their value in local and continental correlation. Pp. 315; 31 Plates; 92 Text-figures. 1969. 11. . Cuitps, A. Upper Jurassic Rhynchonellid Brachiopods from Northwestern Europe. Pp. 119; 12 Plates; 40 Text-figures. 1969. £4.75. . Goopy, P. C. The relationships of certain Upper Cretaceous Teleosts with special reference to the Myctophorids. Pp. 255; 102 Text-figures. 1969. £6.50. . OwEN, H. G. Middle Albian Stratigraphy in the Paris Basin. Pp. 164; 3 Plates; 52 Text-figures. 1971. 6. . Sippigu1, Q. A. Early Tertiary Ostracoda of the family Trachyleberididae from West Pakistan. Pp. 98; 42 Plates; 7 Text-figures. 1971. {8. id Ser Printed in England by Staples Printers Limited at their Kettering, Northants, establishment £ q * A REVIEW OF SOME ENGLISH -PALAEOGENE NASSARIIDAE, _ FORMERLY REFERRED TO COMINELLA 4 on MUs BR 4 JULI973 4, & Y, RAL WSS ©. P; NUTTALL ‘ AND _ J. COOPER im rie “fd iM ¥ uly nye . % eA \ ° a ; vel , ¥ : rs rr ts, i N, + ys Zak “ ; ye, Cea Ry Sa e - BULLETIN OF € BRITISH MUSEUM amas HISTORY) Vol. 23 No. 3 Ub a 3 LONDON: 1973 A REVIEW OF SOME ENGLISH PALAEOGENE ASA OS NASSARITDAE, FORMERLY REFERRED TO a i COMINELLA ( 4 JULI973 | &%, e/ NAL WS BY CLIVE PATRICK NUTTALL. ...( AND JOHN COOPER Pp. 177-219; 9 Plates; 1 Text-figure BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 23 No. 3 LONDON : 1973 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) instituted im 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper 1s Vol. 23, No. 3 of the Geological (Palaeontological) series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. World List abbreviation Bull. Br. Mus. nat. Hist. (Geol.) © Trustees of the British Museum (Natural History), 1973 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 29 June, 1973 Price {2°50 A REVIEW OF SOME ENGLISH PALAEOGENE NASSARIIDAE, FORMERLY REFERRED TO COMINELLA By CLIVE PATRICK NUTTALL and JOHN COOPER CONTENTS Page SYNOPSIS ‘ : : : : : : : : ; ‘ 179 INTRODUCTION : : ‘ : : : : ‘ : . 180 MATERIAL STUDIED . : a ; ‘ : 183 TABLE I. Distribution of some early Nassariidae : 4 : 184 Explanation of abbreviations referring to sources of saosin ; : 186 LOCALITIES AND HORIZONS 5 ¢ . é : ¢ : : 186 MODE OF LIFE . : < 0 . . : : c 5 : 188 ACKNOWLEDGEMENTS : : 2 ; : : : : : 188 SYSTEMATIC DESCRIPTIONS : : : 3 : , : ‘ 189 Pseudocominella gen. nov. : 2 : : 5 3 : ; 189 P. deserta (Solander) . : : 5 : , G c 190 P. avmata (J. de C. Gowerby) : 6 : . : : : 193 P. semicostata sp. nov. : é c : : : : : 196 P. solandevt (Cossmann) 5 : c : c : cl . 198 Desovinassa gen. nov. . : : : 5 : ; : ‘ 201 D. desori (Deshayes) . F ‘ ; F ‘ ; : : 202 D. williamsi sp. nov. : . - : : i 4 : 203 Whiteclifia gen. nov. . : 3 : : c : c : 204 W. suturosa (Nyst) . : 2 Z : : 3 : : 204 W.tumida sp.nov. . : 3 : : ¢ . : : 206 Colwellia gen. nov. : 5 5 , 0 ; : 208 Colwellia flexuosa (Edwards) ° : c ; c . : 210 Keepingia gen. nov. ‘ ‘ : j é : : : 6 212 Thanetinassa gen. nov. . : : : : ° : : : 212 T. bicorona (Melleville) : : - : : : 2 213 REFERENCES . : : . 5 5 : : : . 5 215 SYNOPSIS English Palaeogene prosobranch gastropod species, previously assigned to the Tertiary and living New Zealand genus Cominella (Buccinidae or Cominellidae of Buccinacea), are reviewed. New genera are described to accommodate these English species and their European relatives. They are assigned to the Nassariidae (Buccinacea) on the characters of their fascioles and columellae, in particular the presence of a terminal columellar plait. Like living Nassariidae, 180 SOME ENGLISH they probably lived in warm seas, in shallow water, possibly intertidally. Their known distri- bution, both in Europe and beyond, is summarized. Colwellia nov. gen. occurs both in the European and the United States west coast Eocene (California to Washington State) where it is represented by species usually assigned to Molopophorus. Most other species assigned to the latter genus also appear to belong to the Nassariidae. Its type species, M. stviatus (Gabb) seems not to belong to the family but to be a synonym of Brachysphingus which is shown to lack the characteristic terminal columellar plait. Palaeocene Thanetinassa nov. gen. shows certain similarities to the later Palaeogene Phos-like (Buccinidae) genera Tritiavia, Buccitriton and to Sagenella (Nassariidae), all from the U.S. south-eastern province. This publication is a preliminary to a more intensive review of the Nassariidae and their possible relationship with the Phos-like group. The following new genera of Nassariidae are described:—Pseudocominella, Desorinassa, Whitechffia, Colwellia, Keepingia and Thanetinassa. The following new species are described from the Hampshire Basin. Eocene (Ypresian):—Desorinassa williamsi; Late Priabonian :— Pseudocominella semicostata, and Whitecliffia tumida. Several Palaeogene species, including Buccinum canaliculatum J. de C. Sowerby and B. montense Briart & Cornet are shown to be unrelated to the Nassariidae. INTRODUCTION THE present study started as a result of a request for identification of a specimen from the English Lower Eocene which proved to belong to an undescribed species. It is one of a small group of species in need of revision which are normally assigned to the living New Zealand genus Cominella. ~ The genus Cominella was erected by Gray (1850: 72) who assigned to it nine living species from New Zealand, Australia and South Africa. Iredale (1918 : 34) designated as type species the first mentioned, Buccinum testudineum Lamarck from New Zealand. He also selected the living South African species Buccimum cinctum Réding (= porcatum Gmelin, preoccupied) as type species of his new genus Burnupena. During the latter part of the nineteenth century the name Cominella came to be almost universally applied to a group of European species, most of which are now assigned to the new genera described here. The designation by Cossman (Ig0I : 149) of Buccinum porcatum Gmelin as type species of Comznella is invalid as it was not among those listed by Gray. Cossmann, unfortunately, also based his definition of the genus partly on two European Palaeogene species which he regarded as ‘plesio- types’, Buccinum gossardit Nyst and Buccinum desertum Solander. He regarded Molopophorus from the Eocene of California as a probable synonym of Cominella, and (Igor : 201) stated that Molopophorus should be eliminated from the Nassidae (now known as Nassarlidae), though expressing doubts as to the accuracy of Gabb’s figure of the lectotype of M. striatus Gabb, the type species of Molopophorus. Later Cossmann indicated that he had no remaining doubts as to this synonymy after examining the type figures of another species, Molopophorus gabbi Dall (1909 : 45, pl. 3, fig. 8), from the Oligocene of Oregon (see Weaver, 1942 : 466). Subsequent American workers (Stewart, 1926 : 389; Vokes, 1939 : 140-141) have suggested a connection between Molopophorus and European Palaeogene species assigned to Cominella. In the present study four United States West Coast species, previously PALAEOGENE NASSARIIDAE r81 assigned to Molpophorus, are recognized as belonging to Colwellia. A re-examination of the lectotype of Molopophorus striatus suggests that it is a juvenile Brachysphingus, and that neither belong to the Nassariidae (see p. 209). All other species normally regarded as belonging to Molopophorus appear to have little in common either with its type species or with Brachysphingus and seem to belong to more than one undescribed genus. Gabb (1869 : 156, 157, pl. 26, fig. 36) in describing Molo- pophorus as a subgenus of Bullia, compared it with Bullia (Buccinanops) monilifera (Kiener) now living on the Atlantic Coast of South America. Not only the type species of Molopophorus, but also species assigned to that genus later, and Brachy- sphingus have been generally accepted by American authorities as members of the Nassarilidae. Some of the gross shell characters of these American Tertiary forms and the European Tertiary genera described in this paper closely resemble those found in modern Nassariidae and in Cominella which have fasciole regions of a broadly similar type (see Text-fig. 1). The growth lines on the neck region are bent aborally to form a ridge confluent with the adapical margin of the track of the anterior notch. The deepest point of the asymmetrically U-shaped notch is usually fairly close to the ridge. Below the ridge there is some variation. The growth lines range from regular to imbricate and are reverse S-shaped, but the upper and lower halves of the S may be of varying proportions and curvature. In true Cominella the notch corresponds with a raised ridge, as opposed to a groove observed in many definite Nassariidae and the fossil genera under consideration. A more important feature that the latter share with definite Nassariidae is that the columella is bent strongly to the left and bears a terminal plait. Comunella lacks this plait and its columella is not so strongly bent. On these grounds these fossil genera may be assigned with confidence to the Nassariidae rather than to the Buccinidae (as constituted by Wenz, 1941 : II51I-— 1200; 1943 : 1201-1210), which includes the Cominellidae of Powell (1929 : 57). Homoeomorphic, superficial, similarities between these genera and Comuinella include the frequent development of a subsutural platform and concave ramp and the rather flexuous but basically orthocline growth lines. Buccinum canaliculatum (J. de C. Sowerby, 1823 : 14, pl. 415, fig. 2) from the Barton Beds of the English Upper Eocene was placed in Cominella by Newton (1891 : 168). Cossmann (1901 : 121-123, pl. 8, figs 13, 14) selected it as the type species of his monotypic genus Bartonia. He compared it with ‘Cominella’ deserta from the same formation, pointing out the differences in the features of their aper- tures and canals. Bartonia canaliculata lacks a columellar plait, its columella curves to the left and is not truncated, no ridge is formed at the adapical margin of its fasciole and the growth lines in the fasciole region are less strongly curved than in Pseudocominella (Text-fig. 1). It is hoped that Bartonia, which remains assigned to the Buccinidae (sensu Wenz, 1941-1943) will be treated in a future paper. In this paper, five genera, Pseudocominella, Desorinassa, Colwellia, Whitecliffia and Thanetinassa are described to accommodate British and European Tertiary species previously assigned to Cominella. The opportunity is also taken to describe a sixth genus of the group, Keepingia, though it does not occur in England. The 182 SOME ENGLISH Ab Fic. 1. Fasciole and columellar details of 1, Cominella testudinea (Lamarck), type species of Cominella (fam. Buccinidae) Recent, New Zealand, BMZD 1844.7.29.36, showing ridge above fasciole and no columellar plait (also see pl. 9, figs 11a, b), x5. 2, Sphaero- nassa mutabilis (Linné), type species of Sphaeronassa (fam. Nassariidae), Recent, Bay of Naples, BMZD 19722, showing strong ridge above fasciole, strong terminal columellar plait with weaker plaits above, x5. 3, Pseudocominella deserta (Solander) type species of Pseudocominella, Upper Eocene, Lower Barton Beds Bed B, Barton, GG 19675 (Ru), showing strong ridge above fasciole, imbricate growth lines of fasciole, strong terminal columellar plait with weaker plaits above (see also pl. 1 figs 7a, b), x5. 4a, b, Bartonia canaliculata (J. de C. Sowerby), type species of Bavtonia (fam. Buccinidae), Upper Eocene Lower Barton Beds, Barton. 4a, GG 6897/1 (Wr), apertural view, x 2-5; 4b, GG 6897/2 detail, showing no sharp break between neck and fasciole and no columellar plait, «7-5. PALAEOGENE NASSARIIDAE 183 gross shell characters of these new genera are fairly similar, the main differences being in the form of the apical whorls, strength of sculpture and the strength and character of the ridging in the fasciole region. The known distribution of these genera, in Europe and beyond, is summarized on Table 1. We have examined numerous references to Cominella, ‘Nassa’, Molo- pophorus and Brachysphingus but have found few plausible extra-European occurrences of the genera newly described here. The recognition of these additional Palaeogene members of the Nassariidae affects our understanding of the evolution of the family. It is intended to pursue this topic in a future paper. Although it is hoped that the specific and generic diagnoses will be sufficient to distinguish different taxa, the descriptions given here are detailed. In addition, the variation in a species from locality to locality has been illustrated and commented on as fully as possible. This policy is justified because apparently good illustrations and descriptions have proved inadequate and misleading, so that important points have only been resolved by the study of actual specimens. To quote only a few examples, it seems incredible that the European Tertiary species discussed herein should have been accepted for about one hundred years as members of the Recent New Zealand genus Cominella which belongs to another family, that Buccinum montense Briart & Cornet was not recognized as being congeneric with many Eocene species assigned to Pollia of the Buccinidae, that Bartonia canaliculata (J. de C. Sowerby) should have been considered congeneric with Buccinum desertum Solander, and that the French Lutetian and Auversian Buccinum fusiforme Deshayes should have been considered synonymous with the Bartonian B. desertwm Solander rather than with the Auversian Stvepsidura armata J. de C. Sowerby. MATERIAL STUDIED This study of English specimens is based primarily on the collections in the British Museum (Natural History). Material in the Sedgwick Museum, Cambridge, the Geological Survey Museum, London, the Sandown Museum, Isle of Wight, Institut Royal des Sciences Naturelles de Belgique, Bruxelles, Institut de Paléontologie, Muséum National d’Histoire Naturelle, Paris, and Ecole des Mines, Paris has also been examined. In addition, much valuable information has been obtained from the stratigraphically detailed labels of several private collectors. To indicate the authen- ticity of various records, the collector is named when known. The most important single collection is that of F. E. Edwards in the British Museum (Natural History). Much of his collection was obtained from workers such as Henry Keeping, who also collected the fossils from Huntingbridge listed by Fisher (1862 : 79), and it was enriched by exchange with foreign palaeontologists. It forms the basis of many published faunal lists (von Koenen, 1864; Judd, 1880; Keeping & Tawney, 1881). The British material in the Edwards collection was catalogued by Newton (1891), who included many of Edwards’ manuscript names. Both the Edwards collection and Newton’s catalogue may be presumed to be the SOME ENGLISH 184 (‘poururexe useq JOU DALY »% PAYIeUI soroeds yo suourtoads :930u) (viuLsny) vyvssoo0rdns *‘Gixx —_ -- — (OD0]) tassv9aunng ‘Tixx — — DAIDIAAI “Dd pyonbyun “> susuauola, “Dd — 12404q *9D DIAISAD * I — DSOANANS “fy -VSOANANS * _ — DAagsy DI0]1]NG ‘q wpawssos “yy DIONE “d DIV1IN4 ‘d (eseqys Len) DSOANINS * {4 VSOANNS * (NVISINVG “M) 2a/DpUDUUD “Sixx — DUADPISSYI “SJ PADSSOS “3 BIVUAISWUMN ° ST 4 re DIADPISSDI * ST 1199 “M 4199 “ST SHONAAANIDOQ (LSvOD Isa) GNVTIOPL UTHLO VoIuany ANVWUaD walotag HLUON TT 1A0sap *‘T DYUOAOIIG * T. AA DIDUAD “I DIDULAD * asuarssannn * OR UpADSSOs “ST SUSUGANJD “ST suaparavad “Mex 91112940} “MY FONVUT sepliiesseN ApIee oulos Jo WOIZNqIIysSIG I alavy 1Aaosap *‘ DUO0AOIIG * I. smpyyim °C DIDUAD * J uvijouey uereureds uersoid x welsing eon Ty uvISIOANY iapunjos ‘q (‘uemoqerig vyaasap “I vpyuny * M DSOANINS * vsonxayf "9D pIDsoIMas * qT wyaasap *q ANVIOND Ayres) ueuoyieg (uvtuoqetig oyeT) ueyI0}}eT[ ueredny uenyeyD uerueqyinby uelesiping UeIN0}IOT, HOVLS HNAOOAV TVd HNAHOO4 AHNAOVOSITO AHNHOOIWN PALAEOGENE NASSARIIDAE 185 EXPLANATION TO TABLE 1 The table is compiled from: (a) England (this paper). (b) France: Cossmann, 1901; Deshayes, 1865; Glibert, 1960, 1963; Peyrot, 1927. (c) Belgium & Holland: Albrecht & van der Valk, 1943; Glibert, 1954, 1957. ) Germany: Beyrich, 1854; Glibert, 1960, 1963; Sandberger, 1863; von Koenen, 1889. ) North America: Vokes, 1939; Weaver, 1942. f) Austria: Traub, 1938. g) Togo: Furon & Kouriatchy, 1948. h) West Pakistan: Vredenburg, 1925. Notes. An examination of the holotype and paratype (Bayerisches Geol. Landesamt no. 3080, 3081) of Northia (Cominella) angusta H6lzl (1958 : 244, pl. 20, figs 16-16a) suggests that the species should be referred to Dovsanum. Buccinum montense Briart & Cornet (1871 : 30-31), referred by subsequent authors to Cominella is a Pollia (Buccinidae). Many North American species ranging from Paleocene to Miocene in age are omitted from this table. They belong to undescribed genera (see p. 209) close to those described herein. An examination of the holotype of Cominella devtonensis Bellardi (1882: 3, pl. 1 figs ra, rb) from the Middle Miocene (Tortonian) of St Agata-fossili, Italy (Instituto di Geologia, Turin, no. 11436) shows that it is ‘not a member of the Nassariidae and is in no way related to Pseodocominella. Abbreviations of generic names C., Colwellia; D., Desovinassa; K., Keepingia; P., Pseudocominella; T., Thanetinassa; W., Whitecliffia. 186 SOME ENGLISH main sources of the records of species in Britain quoted by Continental authorities such as von Koenen and Cossmann. Keeping’s own collection in the Sedgwick Museum, Cambridge, is of particular value because of the reliability of his labels. The publication (1933) of the late E. St. J. Burton’s letter classification of horizons in the Barton Beds at the type locality Barton-on-Sea has enabled later collectors to indicate horizons more precisely than hitherto. The faunal list published by Burton was by no means complete. His own collection, now in the British Museum (Natural History), is not as valuable as it might have been since he often arranged specimens of a particular species from several horizons in the same box. In the systematic descriptions, the localities and horizons of material which we have examined are given after the synonymy, under the heading ‘Material Studied’. It has not been considered necessary to list here material in other Museums which merely complements the British Museum collections unless the record is of particular interest. Under the subsequent heading ‘Further Occurrences’, we give records in the literature of those other localities and horizons from which we have not studied material. To avoid undue repetition, we have omitted those records that were probably based on the specimens which we have studied. Explanation of abbreviations referring to sources of specimens ANSP, Academy of Natural Sciences of Philadelphia; CAS, California Academy of Science; Cu, D. Curry Colln.; Ed, F. E. Edwards Colln.; EMP, Ecole des Mines, Paris—invertebrate collections now being transferred to Laboratoire de Paléontologie, Institut de Géologie, Université de Paris, Orsay-91, France, under direction of M. J. Manivit; Ho, M. J. Hoare Colln.; G or GG (followed by number), Department of Palaeontology, British Museum (Natural History); JRScNB, Institut Royal des Sciences Naturelles de Belgique, Bruxelles; Je, P. Jennings Colln.; K, H. Keeping Colln.; KGéT, H. Keeping & E. B. Tawney Colln.; Le, D. N. Lewis Colln.; MNP, Institut de Paléontologie, Muséum National d’Histoire Naturelle, Paris; Nw, C. P. Nuttall Colln.; Ru, A. J. Rundle Colln.; St. J. B, E. St. J. Burton Colln. ; Sedg. Mus., Sedgwick Museum, Cambridge; Sand Mus., Sandown Museum & Public Library, Sandown, I.0.W.; S, F. C. Stinton Colln.; U.Cal., Museum of Paleontology, University of California, Berkeley; Wr, A. Wrigley Colln.; ZD, Recent Mollusca Section, Department of Zoology, British Museum (Natural History). LOCALITIES AND HORIZONS Many of the British localities are still in existence, are well-known and need no further introduction. Some others, notably the New Forest (Hampshire) localities of the Bracklesham and Middle Headon Beds are either poorly known or are no longer accessible. Curry (1958) gives details and map references to most. The following points should be noted about locality names attached to the Edwards Collection: PALAEOGENE NASSARIIDAE 187 1. Bramshaw, Upper Bracklesham Beds, is described (Curry, 1958 : 70) under Shepherd’s Gutter Bed. 2. Roydon, Middle Headon Beds, refers to Royden Manor brick pit (now over- grown) at Nat. Grid SU 320005. Curry (1958 : 29) discusses this locality under the heading ‘Brockenhurst Beds’. The section was described by Keeping & Tawney (r88r : 113 and 114 footnote), the latter refers to the fact that Keeping collected the material in the Edwards Collection. 3. Brockenhurst, Middle Headon Beds. Specimens in the Edwards Collection are labelled either Brockenhurst or Whitley Ridge. Keeping & Tawney (1881 : 10g) write ‘The greater part of the fossils from Brockenhurst were collected by the hands of one of the authors (i.e. Keeping) and . . . dispersed into the various public and private collections. They were obtained from the (railway) cutting at Whitley ... about (1858, when an extra track was being laid). Ina footnote (pp. 109-110), they point out that they both visited the cutting in the summer of 1880, and that the rich (fossiliferous) zone would never be seen here again but that fossils could still be obtained from the spoil heaps (left from the cutting for the original single-track railway in about 1838). As it is clear from a further footnote (p. 114) that Keeping collected the Royden and Lyndhurst fossils in the Edwards Collection, it is more than likely that the same applied for Whitley Ridge. A possible explanation for some of Edwards’ material being labelled ‘Brockenhurst’ is that it was obtained from the spoil-heaps. It seems extremely unlikely that the material came from the Victoria Tile Works at SU 316034. This locality was not mentioned by Keeping & Tawney (1881). It was in existence by 1890 and was normally worked to only a shallow depth as the weathered material near the surface produced the best-coloured products, so that the Brockenhurst shell bed was seldom exposed (Davis, 1952 : 215-216). The foregoing, however, suggests that Davis might have been mistaken in stating that Edwards had worked the local exposures. Whitley Ridge Cutting is the type locality of the Brockenhurst Beds (Curry, 1958 : 29). 4. Lyndhurst, Middle Headon Beds. Edwards’ specimens were probably collected by Keeping in 1858 from pits dug on Lyndhurst Hill, to the west of Lynd- hurst, the Brockenhurst Bed outcropping at about the 200 ft contour (Tawney, 1883). The Brander Collection locality ‘Hordwell’ (= Hordle) quoted by Solander (1766) was the nearest village to the Barton coastal section before Barton itself was built and named. The Middle Headon Beds have been regarded by most British workers as being of Lower Oligocene (Lattorfian) age. Recent nannoplankton zonation work by Martini & Ritzkowski (1968 : 244-247), Martini (1969 : 117-159), and Martini & Moorkens (1969 : 125-127) show that the Headon Beds are late Priabonian and that some German localities, previously regarded as Oligocene are distinctly pre- Lattorfian. Westeregeln, in part, is now regarded as being approximately equivalent to the Upper Barton Beds. The ages accorded here to French Paris Basin localities are based on the lists given by Glibert (1960 : 3-5), and by reference to Fritel (1910). 188 SOME ENGLISH MODE OF LIFE Pseudocominella and its relatives appear to have lived in shallow water. Thanetinassa bicorona, according to Briart & Cornet (1871 : 31), is very common at some French localities, notably Chalons-sur-Vesle in the Sables Inférieurs—a shallow-water sandy deposit. The species also occurs in the coarsely-grained Thanet Sands in England. No species of the group are known from the London Clay. The only British species occurring in rocks of equivalent age is Desorinassa williamsi from the shallow-water Bognor Rock. Pseudocominella armata is rare and occurs in the rich fauna of the Upper and more clayey part of the Bracklesham Beds. No species is known from the more sandy Lower Bracklesham Beds. In the Barton Beds, species of Pseudocominella are known from Horizons Az to H. P. deserta is never common. From the appearance of the specimens it would seem that most were obtained from the extremely fossiliferous Horizon E. P. solanderi, however, is extremely common in the lenticular, sandy, shell drifts of Horizon A3. Members of the group also occur in the Brockenhurst and Venus Beds of the Headon Series. These two latter beds represent periods of shallow marine deposition in a largely fresh and brackish water succession. Turning again to the European Continent, many of the occurrences are consistent with a shallow-water mode of life. Examples include P. avmata in the Calcaire Grossier and Keepingia gossardi in the Sables Supérieurs of the Paris Basin, K. gossardi and Whitecliffia tumida in the Belgian Oligocene and K. cassidavia in the late Oligocene deposits of the Mainz Basin. Finally, we note the occurrence of species of Keepingia in the Lower Miocene of the Bordeaux region in conjunction with Nassarius, Dorsanum and Cyllene, all living genera of the Nassariidae. Nearly all the present day Nassariidae live in very shallow water and many inter- tidal species of the three most important genera, Nassarius s.l., Bullia and Buccinanops are known. Dorsanum and Cyllene are rare but are known only from shallow-water. Considering only the British species of the Pseudocominella group, it would seem unlikely that either P. aymata or P. deserta was intertidal. The evidence in the case of the other species is not clear cut; any of them could well have been intertidal but there is no proof. Such a mode of life would help to explain the rather sporadic fossil record of the group; inshore deposits are relatively often eroded after deposition. The seas in which these forms lived would have been considerably warmer than those around Northern Europe now. At the present day the family is important in warmer waters, mostly within 40° North or South of the Equator. The occurrence of a few species of Nassarius s.l. in Northern European seas may be due to the warming influence of the Gulf Stream. Modern members of the family are carrion-eaters (Fretter & Graham, 1962 : 522). ACKNOWLEDGEMENTS Many collectors have allowed us to study their well authenticated material and have willingly presented to the British Museum (Natural History) those specimens that have been asked for. For such help we therefore wish to thank Messrs D. PALAEOGENE NASSARIIDAE 189 Curry, M. J. Hoare, D. N. Lewis, A. J. Rundle and F. C. Stinton as well as Mr and Mrs R. Jennings and Mr and Mrs B. A. Williams. We thank Mr D. Curry also for useful advice on some nomenclatorial problems. We extend our thanks to Dr Horace G. Richards, Academy of Natural Sciences of Philadelphia; Dr Leo G. Hertlein, Academy of Science, San Francisco; Dr Joseph H. Peck, Museum of Paleontology, University of California at Berkeley; Dr Annie Dhondt and Dr M. Glibert, Institut Royal des Sciences Naturelles de Belgique, Bruxelles; Drs J. C. Fischer and P. Brebion, Muséum National d’Histoire Naturelle, Paris; Dr Denise Petitbois, Ecole des Mines, Paris and M. J. Manivit, Université de Paris, Orsay; Dr C. L. Forbes and Dr B. R. Rickards, Sedgwick Museum, Cambridge; and Mr A. E. Baker of Sandown Museum, I.O.W., all of whom either have given access to the collections in their charge or have loaned specimens. SYSTEMATIC DESCRIPTIONS Order NEOGASTROPODA Superfamily BUCCINACEA Family NASSARIIDAE PSEUDOCOMINELLA gen. nov. (Pls 1-3; Pl. 4, figs 1-3) TYPE SPECIES. Buccinum desertum Solander, 1766. DiacGnosis. Bucciniform; moderate sized, normally attaining an adult height of between 20 and 35 mm; protoconch of three convex, naticoid whorls; form of transition into teleoconch unknown; teleoconch of up to seven whorls; aperture broad, only slightly constricted anteriorly, about two-thirds height; posterior siphonal notch present; fasciole narrow, rugose and bounded adapically by a strong ridge and abapically by a strong fold and corresponding with a groove on the external shell surface; spiral ribbing on fasciole sometimes seen on juvenile shells; columella bent to the left and terminating in a raised diagonal plait abapically; columellar callus weak, often not quite obscuring ridge and fold of fasciole which may therefore appear as weak columellar plaits; outer lip usually spirally ribbed within, internal ribs unrelated to external ribs; growth lines wavy, essentially orthocline but bent adorally in sutural region and aborally to meet ridge of fasciole; weak sinus (strom- boid notch) present in abapical part of outer lip; flat subsutural platform present; ramp concave; spiral ribbing fairly strong; collabral ribbing usually developed, with a tendency to become spinose at periphery. OTHER SPECIES ASSIGNED. Strepsidura armata J. de C. Sowerby, 1850; Buccinum bullatum Philippi, 1847; B. bullatwm aspera von Koenen, 1889; Pseudocominella semicostata nov. sp.; Cominella solandert Cossmann, 1889. GEOLOGICAL RANGE. Middle Eocene (Lutetian) to Lower Oligocene (Lattorfian). 190 SOME ENGLISH REMARKS. Vokes (1939 : 141) suggested that ‘Cominella’ deserta might be congeneric with a group of species from the United States west coast Eocene, usually referred to Siphonalia (fam. Buccinidae). S. bicavinata Dickerson, in fact, strongly resembles both P. avmata and P. semicostata. Its familial position is obscure. Like Recent Szphonalia, it possesses a very weak terminal columellar plait and reverse S-shaped growth lines on the fasciole. It is definitely not a Pseudocominella, which has a strongly grooved fasciole and an immediately obvious columellar plait. Some species of Nassarius (s.l.) have fairly similar apices and sculpture to Pseudo- cominella. They differ in often having a constrict:on of the neck above the fasciole whilst the fasciole itself may correspond with a convex fold which is often strongly spirally ribbed. The fasciole of Strepsidura (pl. 9, fig. 10) is similar to that of Pseudocominella but the former genus may be distinguished by its more twisted columella and by its aperture which is approximately four-fifths shell-height. In addition, its protoconch is larger and rather flattened apically. Pseudocominella differs from Comuinella in the form of the fasciole and the lack of a columellar plait; the two genera are not co-familial. Pseudocominella deserta (Solander, 1766) (Pl. 1, figs 1-12) 1766 Buccinum desertum Solander: 13 (pars), pl. 1, fig. 15 (non figs 18, 19) 1823 Buccinum desevtum Brander; J. de C. Sowerby: 14, pl. 415, fig. 1. 1850 Fusus desevtus Morris; d’Orbigny: 363. 1854 Buccinum excavatum Beyrich: 444, pl. to (7), figs ta, 1b, tc. 1864 Stvepsidura deserta (Solander) von Koenen: 100. 1866 Buccinum (Cominella) desertum Lowry et al., pl. 3. 1880 Cominella (Buccinum) deserta (Solander) Judd: 154. 1881 Cominella (Buccinum) deserta (Solander); Keeping & Tawney: 116. 1888 Buccinum desertum Solander; Gardner, Keeping & Monckton: 626. 1889 Buccinum desevtum Brander; Bristow et al.: 289. 1889 Buccinum (Cominella) desertum Sowerby; von Koenen: 241 (pars). 1889 Cominella desevta (Solander); Cossmann: 136 (pars). 1891 Coméinella desevta (Solander); Newton: 168. 1901 Comunella deserta (Solander); Cossmann: 149, 150 (pars). 1901 Coménella deservta (Solander); Cossmann, pl. 6, fig. 4. 1933 Comunella desevta (Solander); St. J. Burton: 156. 1963 Comunella deserta (Solander); Glibert: 67 (pars). 1968 Cominella deserta (Solander); Martini & Ritzkowski: 244-247. LECTOTYPE (designated herein). Upper Eocene, early Priabonian, Barton-on- Sea, Hants, figured Solander, 1766, pl. 1, fig. 15, GG 19667 (Brander Colln.). PARALECTOTYPE (designated herein). GG 19668, same details, but not figured by Solander. MATERIAL STUDIED. ENGLAND, Upper Eocene, early Priabonian, Barton Beds, coastal section between Highcliffe and Barton-on-Sea, Hants. Horizon not indicated: GG 19669/1 (figd. J. de C. Sowerby 1823, pl. 415, fig. 1, the central figure); GG 19670/1-25 (Ed). Lower Barton Beds, horizon Az, GG 19671 (Ru). PALAEOGENE NASSARIIDAE 191 Horizon A3, GG 19672/1-3 (Le); GG 19673 (S); GG 19674/1-13 (Wr). Horizon B, GG 19675 (Ru). Middle Barton Beds, GG 19676/1-g (St/B); GG 19676/1-6 (Wr). Horizon C, (Je); (S). Horizon E, (Cu); (S). Upper Barton Beds, Horizon H, GG 19678/1-3 (Wr); GGu1151 (H. A. Toombs); GG 19725 (Cu). Lower Barton Beds, Alum Bay, I.0.W., Sand. Mus. 3554 (J. F. Jackson Colln.); same (near top), GG 19732 (Wr). Upper Eocene, late Priabonian, Middle Headon Beds, Brocken- hurst, GG 19679 (Ed). Brockenhurst Bed, Whitecliff Bay, I.0.W., Sedg. Mus. C 29122 (K. & T.). GERMANY, Upper Eocene, early Priabonian, Westeregeln, GG 19680/1-3 (Ed). FURTHER OCCURRENCES. GERMANY, Westeregeln, (Beyrich, 1854 : 444; von Koenen, 1869 : 241; Glibert, 1963 : 67; Martini & Ritzkowski, 1968 : 244-247, with nannoflora indicating a similar age to that of the Upper Barton Beds (Bed H). Diaenosis. Pseudocominella up to 36 mm high; up to seven whorls of teleoconch; mean spire angle 55°—60° ; prominent subsutural platform and concave ramp develop- ed; spiral sculpture of narrow ridges often alternately of primary and secondary strength and separated by wide interspaces ;collabral ribbing variable in strength and often rather irregular; outer lip ribbed internally; columellar callus developed. Description. Each whorl is enveloped by the succeeding one just below the periphery. A prominent subsutural platform is developed on the third whorl (first of teleoconch) and reaches a width of 1-5 mm on specimens over 30 mm high. It is normally coronate, being crossed by the axial ribs; in axially multirugose specimens, however, it is often flat on later whorls. It is bordered by a collar bearing on later whorls up to three wavy spiral ribs (two on both the lectotype and paralectotype). The platform is undercut, the sutural ramp being noticeably concave. The whorl shoulder is distinctly adapical to the periphery. On the first and second whorl of the teleoconch one spiral rib marks the margin of the subsutural platform, another the shoulder, and three or four more lie on the rather flat abapical portion of the whorl. These ribs are broad, rather flattened in section, and are separated by narrow grooves. On the third whorl the interspaces broaden and spiral ribs of secondary strength may appear, first on the sutural ramp, then between the primary rib marking the shoulder and the one immediately abapical to it. On later whorls five or six primary spiral ribs of intermediate strength may be distinguished on the increasingly convexly-sided portion of the whorl abapical to the shoulder. These ribs are narrow and are separated by interspaces about double their width. Secondary and sometimes tertiary ribs are often developed with no discernible pattern. On the body whorl between fifteen and twenty-two primary ribs are developed between the shoulder and the adapical margin of the fasciole; up to twenty-five ribs of secondary and tertiary strength may also be present in the same area. On the lectotype there are fifteen primary and fifteen secondary ribs in this region. In the same region there are thirty-nine ribs on the paralectotype but it is difficult to separate them into different orders of strength. The first whorl of the teleoconch bears about thirteen collabral ribs which are well-rounded in section. On the next two whorls the number increases to between fifteen and twenty-five and thereafter remains fairly constant on all succeeding whorls. The lectotype is a 192 SOME ENGLISH particularly multicostate specimen (see table). The spacing and strength of the collabral ribs often varies on a particular shell as well as between different ones. Their strength is normally inversely proportionate to their number. They are strongest at the shoulder which is often slightly nodose. The ribs are usually weaker on the subsutural ramp but stronger on the subsutural platform. They decrease in strength below the periphery. A variable number of internal spiral ribs occur on the inner surface of the outer lip. They are rather irregularly arranged, stretch back into the aperture for varying distances, and some are discontinuous. DIMENSIONS I 2 3 4 5 7 8 9 10 II GG 19667 LT 26°8 14:1 15°6 6+ 25 30 9 23 I-90 1-72 GG 19668 PLT 22 13°4 14'1 6 Ir+ 39 13 19 1-64 1°56 GG 19671 8:8 533 5:0 44 12 23 6 15 1-66 1-76 GG 19672/1 12:6 7-4 TPP 5 20 29 II 16 I-70 1°75 GG 19673/1 17-0 1073 9°4 54 ° 36 II 14 1°66 1-82 GG 19677/11 22°9 12-7 13:2 6 15 36 9 21 1°80 1-74 GG 19677/1 30°9 15 16°5 64 14 26 II 16 2:06 1-87 GG 19670/5 341 19:2 19:64 97, I5+ 32 Eh) <= ~ ae 7s Key (also used in all following descriptions) HT, Holotype; LT, Lectotype; NT, Neotype; PLT, paralectotype; PT, paratype. 1, total height (mm) 2, height of aperture 3, breadth 4, number of teleoconch whorls 5, number of collabral ribs on last whorl 6, number of collabral ribs on penultimate whorl (not shown for this species) 7, number of spiral ribs on last whorl 8, number of spiral ribs on penultimate whorl g, number of spiral ribs inside outer lip 10, height: height of aperture 11, height: breadth REMARKS. Buccinum fusiforme Deshayes, from the Lutetian and Auversian of the Paris Basin, was placed in the synonymy of this species by Cossmann (1889 : 136). A re-examination of Deshayes’ specimens in the Ecole des Mines (see p. 195) shows that they all clearly belong to P. armata. All subsequent continental authors, however, have accepted Cossmann’s views. In the synonymy given here it is assumed that their records of French specimens are of P. aymata and not of P. deserta, and that P. deserta does not occur in France. From examining the lectotype and paralectotype it would appear that other authors have correctly interpreted English specimens of the species. The other specimens figured by Solander (pl. 1, figs 18, 19) as Buccinum desertum are both lost. J. de C. Sowerby (1823, pl. 415, fig. 2) thought that Solander’s fig. 18, and possibly fig. 19 belonged to his new species Buccinum canaliculatum (now Bartonia see p. 181); these determinations have since been disputed. Cossmann (1889 : 137) identified Solander’s figure 18 as Cominella deserta var. solanderi (Edwards MS). D. Curry (personal communication) points out that Solander’s fig. 19 is much more likely to be of Buccinum lavatum J. de C Sowerby (now Pollia). PALAEOGENE NASSARIIDAE 193 In the adult stage P. deserta most clearly resembles P. bullata (Philippi) (pl. 2, figs ga, b) from the Lower Oligocene of Latdorf, Germany. The whorl profile of the latter is more evenly convex, the subsuturalramp being barely grooved. Itssculpture, particularly the collabral elements, is weaker, and the first four whorls of the teleo- conch are much smoother. In some specimens the collabral ribs on the last body whorl are relatively strong in the peripheral region where they form definite nodes. The spiral ribs on the inner surface of the outer lip are more numerous. Comparisons with P. armata and P. solanderi are discussed under the latter named. The records by Morris (1843 : 146) from Bracklesham (as Fusus desertus Brander), by Forbes (1856 : 86) from the Headon Beds of Colwell Bay and by Nyst (1836 : 36) from the Tongrian of Klein Spauwen (Belgium) cannot be confirmed and are almost certainly incorrect. Bristow ef al. (1889 : 289) erroneously state that Keeping & Tawney (1881) record this species from the Bracklesham Beds of the Isle of Wight. This record, in fact, is from the Brockenhurst Bed of the Headon Series (see ‘Material Studied’). Giebel (1864 : 17, pl 1, fig 1) described and figured Buccinum bullatum Philippi from Latdorf. Von Koenen (1865 : 466) redetermined this as Strepsidura deserta (Solander) From the figure it is clear that Giebel’s determination was right and von Koenen’s was wrong. In his more authoritative later publication (1889 : 241) von Koenen does not record P. deserta from Latdorf. Pseudocominella armata (J. de C. Sowerby, 1850) (Pl. 3, figs 1-8) 1835 Buccinum fusiforme Deshayes: 653. 1837 Buccinum fusiforme Deshayes; Deshayes, pl. 87, figs 15, 16, 17. 1850 Buccinum fusiforme Deshayes: d’Orbigny: 420. 1850 Stvepsiduva avmata J. de C. Sowerby in Dixon: 104, 186, pl. 7, fig. 11. 1865 Buccinum fusiopsis Deshayes: 499. 1889 Comunella deservta (Solander) Cossmann: 136 (pars). 1891 Strepsidura avmata J. de C. Sowerby; Newton: 165 (pars). 1901 Comunella deserta (Solander); Cossmann: 149, 150, (pars). 1911 Cominella deseyta (Solander); Cossmann & Pissarro, pl. 37, fig. 178-1. 1948 Cominella deserta (Solander); Morellet, L. & J.: 30, 133, 179, 331, 343- 1963 Cominella desevta (Solander); Glibert: 67 (pars). NEOTYPE (proposed herein). Middle Eocene, Auversian, exact horizon unknown, (probably Bed 19 (Brook Bed) of Fisher 1862 : 74) Upper Bracklesham Beds, Bracklesham Bay, Sussex, GG 19681/1 (Ed). MATERIAL STUDIED. ENGLAND, All from Upper Bracklesham Beds, Auversian. Bracklesham Bay, GG 19681/2~-4 (Ed) (probably same horizon as Neotype) ; G 8869/1— 6 (Ogle Colln.) ; Bed 19 (Brook Bed) of Fisher (1862 : 74), Sedg. Mus. C 64860 (Fisher Colin.) ; G 75897, G 75998 (E. M. Venables Colin.). Brook, Hampshire GG 19683/ 1-4 (Ed); Sedg. Mus. C 67793. Huntingbridge, Hampshire, GG 19684 (Ed). Bramshaw, Hampshire, GG 19685/1-4 (Ed). ‘Fossil bed’ at Bramshaw of Fisher (1862 : 81, lines 2-4), 1 juvenile (Cu). FRANCE, Two separate samples in Deshayes B 194 SOME ENGLISH Colln., Ecole des Mines, Paris (see p. 195). One labelled Le Fayel, Mary and Caumont, the other labelled Vendrest, Auvers and Acy, with no indication as to which shell comes from which locality. All these localities were quoted by Deshayes (1865 : 499), none of them were mentioned by him in his original description (1835 : 653-654). FURTHER OCCURRENCES. FRANCE, Lutetian:—Grignon (Deshayes, 1835 : 654), (Cossmann 1889 : 137). Auversian: Valmondois (Deshayes 1835 : 654) (1865 : 499); Le Guépelle (Deshayes, 1865 : 499); Bezu-le-Guery, Le Fayel (Glibert, 1963 : 67); Auvers (p. 133), Nanteuil-le Haudouin (p. 179), Acy (p. 331), Ru d’Alland valley (p. 343) Morellet (1948) ; Acy (Cossmann & Pissarro, ro1I, pl. 37, fig. 178-1). Exact horizon unknown. Senlis (Deshayes, 1835 : 654). Diacnosis. Pseudocominella with teleoconch of up to six whorls; mean spire angle between 60° and 65°; up to twenty-five collabral ribs per whorl on early part of teleoconch, reducing to about ten increasingly nodose ribs on each of the last two to three whorls; whorl profile convex on early whorls, shoulder becoming increasingly angular on later whorls; spiral sculpture of varying strength; margins of fasciole strongly ridged; fasciole very rugose; outer lip denticulate within in some individuals. DESCRIPTION. All the specimens are damaged and those from the type locality are partly crushed. The protoconch (pl. 3, fig. 4) is typical for the genus and is fairly well-preserved on specimens from Brook and Bramshaw. It consists of about three apparently smooth naticoid whorls. The subsutural platform and close-set collabral ribbing first appear towards the end of the third whorl. On the first three or four whorls of the teleoconch the subsutural platform is bordered by a raised rim which eventually becomes obsolete, finally, on the body whorl the platform tends to be downward sloping. The early whorls of the teleoconch are broadest at the adapical suture and are evenly convex in profile except for a groove under the subsutural platform. In later whorls this groove develops into a broad concave ramp. The profile is further modified by the development of nodes at the shoulder, below which the whorl-sides are vertical. The two strongest spiral ribs on the early teleoconch mark the edge of the sub- sutural platform and the shoulder. On later whorls there are normally two strong ribs in each of these positions. They are joined by others on the ramp itself and increase in number to between seven and fifteen on the ramp of the last whorl. The spiral sculpture differs considerably on specimens from different localities. The patterns of ribbing on the ramp of shells from Bracklesham Bay, Whitecliff Bay and Brook are similar; Bramshaw and Huntingbridge shells differ in having much stronger ribs which may be differentiated into those of primary and secondary strength. Below the shoulder on the peripheral region, up to ten weak spiral ribs may be seen with difficulty on the spire whorls of specimens from Bracklesham Bay. These usually die out on the body whorl of Bracklesham specimens (including the neotype) but always persist on specimens from other localities. The ribs are barely stronger on the Whitecliff Bay and Brook specimens. The Bramshaw and Huntingbridge shells have over forty considerably stronger and more clear-cut ribs of PALAEOGENE NASSARIIDAE 195 both primary and secondary strength on the body whorl. The neck region of specimens from all localities bears between eleven and fifteen spiral ribs. There are between twenty and twenty-five collabral ribs on each of the first two whorls of the teleoconch. These become more irregular in strength and spacing on the next whorls, reducing eventually in number to about nine evenly spaced nodes on the fifth and sixth whorl of the teleoconch. On the earlier whorls they are well-rounded in section, separated by rather broader, smoothly concave interspaces. They are moderately strong and of even strength below the shoulder; above it they markedly decrease in strength and are almost obsolete at the sub-sutural ramp. On the body whorl the collabral sculpture is confined to the area just above and below the shoulder. It consists of strong nodes or sometimes blunt spines. Six of the twelve specimens from the type locality have, on the last whorl, indications of a growth halt during which an aperture similar to the terminal one was developed. Here, the growth lines are more prominent, the outer lip was apparently slightly flared and parietal callus remains above the adapical suture of the previous whorl. One specimen has two such halts and on two of the larger shells these halts occur about 180° before the terminal aperture. Signs of preterminal apertures are common but much less prominent on specimens from other localities. The track of the fasciole is rather deeply grooved, and its bordering ridges, particularly the anterior, are strong. Growth rugae in this region are noticeably imbricate especially in the larger shells. The aperture is damaged in all specimens; none of those from the type locality and only one out of the four from Brook shows any indications of the outer lip being denticulate within. Internal ribs stretch back for about 45° within the aperture in specimens from other localities. In one specimen from Bramshaw these break down into numerous intermittent dashes as the aperture is approached’ DIMENSIONS. I 2 3 4 5 6 9 Io II GG 19681/1 NT 25°5 15'7 16+ 6 9 Io — 1:6 1°57 GG 19681/2 22+ — I5"1 6 be) 10 — — 1°47 GG 19681/3 22 13 I2°9 6 9 12 — 1-69 1-70 GG 19681/4 17+ Ti+ 1073 6 8 ae) — — _ GG 19685/4 12 7 65 5+ 13 18 — I-71 1°84 GG 19685/3 16 9:2 9'7 5+ 12 19 12 1°74 1-65 GG 19685/1 26°7 16:2 16:7 6+ 9 9 15 1°65 1:60 GG 19685/2 28+ 16°5 17'7 6+ 10 9 some I-70 1°58 GG 19684 15°5 933 9°33 6 9 12 13 1-67 1-67 (see p. 192 for key) ReMARKS. The appearance and matrix of the neotype and associated specimens GG 19681/1-6 (Ed) matches that of the specimens listed which are known to come from the Bed 19 (The Brook Bed) at Bracklesham Bay. All the French records of ‘Cominella’ deserta are considered herein to be of P. armata (see pp. 192 and 194). Specimens of Buccinum fusiopsis in the Deshayes Collection (EMP) (no specimens from Deshayes original (1835 : 653-654) localities survive) and the specimen from Acy figured as Cominella deserta by Cossmann and Pissarro (1911, pl. 37, fig. 178-1) have strong sculpture very similar in character to 196 SOME ENGLISH that of specimens from Bramshaw and Huntingbridge in the New Forest (Hampshire). The specimen which Cossmann figures (1go1, pl. 6, fig. 4) as Cominella deserta is stated to be from Barton and the specific identification appears to be correct. All the English specimens of P. avmata have stronger nodes at the periphery of the last whorl than any of the French specimens examined; the closest comparison is between them and the specimen figured from Huntingbridge (cf. pl. 3, figs 6 and 8). P. deservta always has a more rounded whorl profile and a narrower, more concave ramp. The only previous figure of an English specimen of the species is of the relatively smooth form originally described by Sowerby from Bracklesham Bay (1850, pl. 7, fig. 11). Without knowledge of the Bramshaw and Huntingbridge speci- mens, it is hardly surprising that French authorities considered their specimens either to be specifically distinct from P. armata or to be synonymous with P. deserta. The exposures at Bramshaw and Huntingbridge are slightly younger than the Brook Bed (Bed 19) of Brook and of Bracklesham Bay (Fisher, 1862 : 92-93; Curry 1958: 70-71). It seems likely that P. armata gave rise to P. deserta. Changes include an increase in strength of spiral sculpture, an increase in the number of collabral ribs on the later whorls, gradual disappearance of peripheral nodes and the narrowing of the ramp. Deshayes proposed the name Buccinum fusiopsis (1865 : 499) as his earlier name B. fusiforme (1835 : 653) was preoccupied (Borson, 1820 : 222). Sowerby’s name (1850 : 104, 186) therefore stands. The surviving parts of both Dixon’s and the Sowerby family’s collections are housed in the British Museum (Natural History), but the whereabouts of this figured specimen is unknown. Comparisons with P. deserta and P. solanderi follow the description of the latter, whilst those with P. bullata (Philippi) and P. semicostata are discussed under the last named. The name armata has often been applied erroneously to P. semicostata (see its synonymy). The record from the Belgian Oligocene by Glibert & de Heinzelin (1954 : 367) of Northia (Cominella) bullata {. armata is of P. bullata. Pseudocominella semicostata sp. nov. (Pl. 3, figs 9-10; Pl. 4, figs 1-3) 1864 Strepsidura deserta var. armata J. de C. Sowerby; von Koenen: roo. 1880 Strepsidura (Buccinum) armata J. de C. Sowerby; Judd: 154. 1880 Strepsidura (Buccinum) semicostata Edwards MS; Judd: 154 (nom. nud.). 1881 Strepsiduva (Buccinum) armata J. de C. Sowerby: Keeping & Tawney: III, 113, 116. 1881 Strepsiduva (Buccinum) semicostata Edwards MS; Keeping & Tawney: 116 (nom. nud.). 1882 Strepsiduva avmata J. de C. Sowerby; Judd: 477. 1891 Strepsiduva avmata J. de C. Sowerby: Newton (pars): 165. 1891 Strepsidura semicostata Edwards MS; Newton: 165 (nom. nud.). 1901 Strepsiduva arvmata J. de C. Sowerby; Cossmann: 133. tgo1 Strepsidura semicostata Edwards MS; Cossmann: 133 (nom. nud.). 1902 Strepsiduva avmata J. de C. Sowerby; Reid: 4o. 1915 Strepsiduva avmata J. de C. Sowerby; White: 44. 1926 Melongena (Cornulina) minax (Solander) Jackson: 360 (pars). 1926 Strvepsiduva armata J. de C. Sowerby; Jackson: 364. 1960 Stvrepsiduva avmata J. de C. Sowerby; Glibert: 9. PALAEOGENE NASSARIIDAE 197 Hototype. Upper Eocene, Late Priabonian, Middle Headon Beds, Whitley Ridge, railway cutting, Brockenhurst, Hampshire GG 19687/1 (Ed ex K). ParaTyPeEs. All Upper Eocene, Late Priabonian, Middle Headon Beds. Whitley Ridge Railway cutting, Brockenhurst, Hampshire, GG 19687/2-3 (Ed ex kK); Sedg. Mus. C 26453-8 (K). Victoria Tilery, Brockenhurst, (S). ‘Brockenhurst’, exact locality unknown, GG 19688/1-14 (Ed); GG 19689/1-2 (St]B); G7760 (Beckles Colln.). Lyndhurst, Hampshire, GG 19690/1-4 (Ed). Royden, Hampshire GG 19686/1-9 (Ed ex K); Sedg. Mus. C 25070-90 (KG&T). Brockenhurst Bed, Whitecliff Bay, 1.0.W., GG 19692/1-5 (Wr); Sedg. Mus. C 29270-3 (KGT); Sand. Mus. 1930 (J. F. Jackson), recorded as Melongena minax. ?Venus Bed, Colwell Bay, 1.0.W., GG r9691 (Ed). FURTHER OCCURRENCES. Unknown. Diacnosis. Stout Pseudocominella with between six and seven whorls of teleo- conch; mean spiral angle increasing from 60° to 80° with growth; ramp broad; shoulder angular except for early teleoconch, low, at or below mid-height of spire whorls; about twenty-one collabral ribs on early part of teleoconch reducing to nine or ten very nodose ribs on last three whorls; aperture broad; fasciole bordered by strong ridges; fasciole very rugose; outer lip denticulate within in some specimens. DeEscriPTION. The protoconch is decorticated or missing in all specimens; its gross size and shape is similar to that of P. deseyta (Solander). The first two to three whorls of the teleoconch are broadest at the adapical suture and are convex in profile except for a subsutural constriction which rapidly grows into a broad concave ramp. The profile becomes strongly carinate on the fifth and sixth whorl of the teleoconch, with the periphery coinciding with the shoulder. The subsutural platform is horizontal, on later whorls it becomes wavy under the influence of the axial nodes. At the earliest point in the development of the teleoconch at which it is possible to determine detail there are three prominent spiral ribs, and the whorl profile is naticoid. During this whorl, two new ribs develop between the two most adapical. The opisthocyrt collabral costae are of similar strength and form a reticulate pattern with slight nodes at the intersections; they do not extend on to the subsutural platform which first develops on this whorl and is bordered by a spiral rib. The following whorl (third of the teleoconch) is similar except for the broadening of the subsutural ramp which lacks spiral ribbing. The shoulder is marked by the second most adapical spiral rib. The reticulate pattern is made of increasingly spirally-elongate rectangles. On the fourth teleoconch whorl the increasing strength of the collabral ribs compared with the spiral ribs becomes noticeable; within the next whorl or so this leads to the loss of the reticulate appear- ance. Theramp gradually becomes concave and bears two new spiral ribs of secondary strength. The edge of the subsutural platform is marked by two strong spiral ribs and another pair lies on the shoulder. Between the fifth and seventh teleoconch whorls seven to ten spiral ribs are developed on the ramp. Apart from those on the shoulder and bordering the subsutural platform these are seldom of greater strength than those on earlier whorls and are often difficult to follow even on 198 SOME ENGLISH well-preservedshells. On immature specimens with a teleoconch of four whorls, there are on the body whorl between eighteen and twenty spiral ribs of varying strength below the shoulder, seven or eight of these being on the neck and rather stronger. The full adult number of spiral ribs, about twenty-nine in all, of which fourteen are on the neck, is reached by the sixth teleoconch whorl. There are between eighteen and twenty-one collabral ribs on the first four teleoconch whorls; these die away below the peripheral region of the body whorl even in juvenile shells. On the next whorl these weaken on the ramp and become increasingly nodose at the shoulder and grow further apart. From here on nine or ten axial nodes or spines occur on each whorl. The fasciole is strongly ridged and is very rugose in large specimens. Specimens showing ribs on the inside of the outer lip are very rare. DIMENSIONS. I 2 3 4 5 6 7 8 9 Io II GG 19687/1 HT estimated 27°0 — — —- —- — — — — 1-86 1°52 actual 26°1 14°5 17°8 6 13 12 — — 16 — — GG 19686/2 PT 2053 12°8 13'7 53 9 II 36 12 10 1°58 1-48 GG 19686/3 PT 13°8 8-5 8-5 5 12 21 33 10 fo) 1-62 1°62 GG 19686/4 PT 8-7 595 5:2 44 16 20 23 7 ria 1°58 1:67 GG 19686/5 PT 74 4°7 4°3 =n REDS ED 5 O° E57 ear? 4 GG 19688/1 PT 32+ 19°5 21°1 64 8 12 42 (see p. 192 for key) REMARKS. Edwards gave the manuscript name Strepsidura semicostata to his single specimen from Colwell Bay (GG 19691). This was so badly worn that he failed to recognize that it was conspecific with numerous other specimens (listed here) which he identified as S. armata. P. semicostata differs from P. armata in being relatively stouter, having a broader, less sloping ramp, more prominent axial nodes, and a lower shoulder. It has also been confused by von Koenen (1864 : 100) and by Judd (1880 : 154) with P. bullata (Philippi) which may be easily distinguished by its much weaker collabral sculpture on the early teleoconch; strong axial ribbing and nodes are normally confined to its body whorl. Pseudocominella solanderi (Cossmann, 1889) (Pl. 2, figs 1-8) 1766 Buccinum desertum Solander: 13 (pars), pl. 1, fig. 18 (non figs 15, 19). 1887 Cominella solandvi Edwards; Keeping: 71 (nom. nud.). 1888 Buccinum solandvi Gardner, Keeping & Monckton: 620 (nom. nud.). 1888 Buccinum sp. nov.; Gardner, Keeping & Monckton: 627. 1889 Cominella solandvi Edwards; Bristow: 120, 290 (nom. nud.). 1889 Cominella deserta (Solander) var. solandevi Edwards; Cossmann: 137. 1901 Cominella (Buccinum) solanderi (Edwards) Cossmann: 150. 1926 Cominella solandvi Edwards MS; Jackson: 356. 1933 Cominella sp.; Burton: 156. PALAEOGENE NASSARIIDAE 199 NEOTYPE (selected herein). Upper Eocene, early Priabonian, Lower Barton Beds, Horizon A3, coastal section between Highcliffe and Barton, Hampshire, GG 19693/1 (Ho). MATERIAL STUDIED. Same locality, Lower Barton Beds, horizon unspecified GG 19696/1-20 (Ed); 19697/1-10 (Wr). Horizon Az, GG 19703 (Ru). Horizon Az (near top) GG 19698/1-20 (Nu). Horizon A3, GG 19699/1-50 (S); GG 19700/1-3 (StJB); GG 19693/2-3 (Ho); GG 19694/1-18 (Le); GG 19695/1-3 (S); (Cu); (Je). Horizon B, 2 specimens (Ru). Alum Bay, I.0.W., Barton Beds, horizon unknown, Sedg. Mus. C 59712-4; (10 specimens) Sand. Mus. 4720. Lower Barton Beds (near top) GG 19731/1-8 (W7); (2 specimens) Sand. Mus. 1084 (J. F. Jackson Colln.). Whitecliff Bay, I.0.W. Barton Beds, lowest 13 inches with Nummulites elegans, Sedg. Mus. C 64008 (K). FURTHER OCCURRENCES. Lower Barton Beds, Barton-on-Sea, as Cominella sp., horizons A3 and B (Burton, 1933 : 156). Dracnosis. Moderate sized, fairly elongate Pseuwdocominella with a marked horizontal and undercut subsutural platform; mean spire angle between 50° and 60°; collabral ribs present only on early whorls of neoconch; spiral sculpture of close- spaced low ribs; outer lip ribbed internally. DeEscRIPTION. The protoconch is about 1-2 mm high and broad. Its first half whorl is minute; it seems to be disjunct with a semi-erect tip, but this appearance is certainly due to weathering. It is followed by two and a half smooth naticoid whorls. It seems very unlikely that the transition to the teleoconch was abrupt; orthocline collabral ribs of increasing strength occur, these soon become nodose and are joined by spiral ribbing within a quarter whorl. Most adult specimens are badly worn; the sculpture of the body whorl is abraded and the early whorls either decorticated or even missing, the main exceptions are immature specimens under 8 mm high. Teleoconchs of shells 20 mm high would have been of six whorls. Each whorl is enveloped by the succeeding one just below the periphery. The first two-and-a-half to three teleoconch whorls bear twelve, increasing to sixteen rather angular, orthocline, collabral ribs. These die out by a shell height of between 6 and 8 mm, except on a very few of the specimens seen, where they persist above the periphery sometimes until full growth is reached. On succeeding whorls the only collabral sculptural elements are rugae which are distinctly stronger than the growth lines seen on well preserved shells. Spiral ribbing and the subsutural platform appear about a quarter whorl after the first collabral ribs. The subsutural platform, which reaches a width of r mm on the last whorl of large shells is normally horizontal but may slope either upwards or downwards away from the suture. Another fairly strong spiral rib marks the shoulder, which is nodose and carinate for the first two or three whorls of the teleconch. Within the next two whorls these are joined by a weaker spiral rib on the concavely-sloped sutural ramp and three or four more between the shoulder and the suture below. By the fifth teleo- conch whorl the latter have increased in number to eight or nine. Two spiral ribs of lower strength commonly occur on the sutural ramp. On the last whorl other weak ribs are sometimes irregularly disposed, mainly above the periphery. The 200 SOME ENGLISH number of spiral ribs on the last whorl increases with size (see table), most adult shells having between thirty-eight and forty ribs of varying strength. The nine or ten on the neck region are the strongest, apart from the primary ribs on the shoulder and on the edge of the subsutural platform. The spiral ribs are about the same width as the interspaces except on the neck region where they are more widely spaced. In section they are usually gently convex and often almost flat-topped. The inner surface of the outer lip bears ribs which do not correspond with the external ones. They stretch back inside the aperture for nearly a quarter whorl in large individuals, less so in small ones and are resorbed as the shell grows. The number of internal ribs is about half that found on the outer surface and is dependent on size. Two spiral colour-bands may be seen on the neotype, one, about a millimetre broad, on the ramp with a narrower one just below the shoulder, other specimens also show this phenomenon (pl. 2, figs 4, 6, 7, 9). DIMENSIONS. I 2 3 4 5 6 7 8 9 10 II GG 19693/1 NT 14°7 76 7°8 53 fo) ° 28 10 18 1°93 1-89 GG 19695/3 52 370 Oi = 953 19 614) «17 5) 13. 178 9 GG 19694/17 6:2 35 35 4 9 16 24 7, Cees GG 19693/2 8-6 4°7 4°5 44 fe) 14 25 Io Il 1°83 I-QI GG 19694/14 10°8 6-1 655 5 6 15 32+ II 14 72 1-66 GG 19694/10 12-6 74 6:8 5 fo) fo) 38+ II 17 1-70 I-61 GG 19694/6 14°6 8-2 8-2 5 4 15 41 14 17 1°78 1-78 GG 19694/5 17°8 9°5 92 58 ) © 35 19 10 =:8q ao GG 19694/1 20:0 12:0 Ten 6 fo) 15 41 16 ze) 1-67 I-80 GG 19700/3 25°5 13°5 13°8 — fo) — -—— — 20 1:89 1°85 (see p. 192 for key) REMARKS. Cossmann (1889 : 137) in discussing P. deserta wrote ‘les figures qu’en donne Brander sont tres exactes, sauf la figure 18, qui représente une variété pour laquelle a été proposé le nom Solanderi, Edw.’. The specimen figured by Solander is lost. None of the specimens GG 19696/1—20 in the Edwards Collection, which are labelled as solandn, have been chosen as neotype as they are all badly worn. In the circumstances it seems justifiable to select a neotype from better preserved material whose horizon is known more precisely. This species is unknown except from the Lower Barton Beds. One sample, GG 19701/1—-4 (Wr) was labelled as being from the Middle Barton Beds. Examina- tion by several workers suggested that the specimens did in fact, come from the Lower Barton Beds. Moreover glue was present on the specimens, unlike others in Wrigley’s Collection, suggesting that he might have acquired the specimens from an old collection and had assumed that any information with the specimens was correct. Another specimen, GG 19702 (Wr) was found in the same box as three specimens of P. deserta, GG 19678/1-3 (Wr) from the Upper Barton Chama Bed (Bed H). Again, examination by several workers showed that the preservation and matrix of the specimen of P. solanderi was not the same as that of its companions and that it was almost certainly from the Lower Barton Beds. This species is among the twenty most common prosobranchs in horizon A3, of the Lower Barton Beds where P. deserta is extremely rare. The latter species, PALAEOGENE NASSARIIDAE 201 however, is more common in the Middle and Upper Barton Beds in which P. solanderi is unknown. Both species usually attain a height of 20 mm in Bed A3 of the Lower Barton Beds. In the Middle Barton Beds, P. deserta often exceeds 30 mm in height. This increase in size corresponds with almost an extra whorl (making a total of seven for the teleoconch). P. deserta is relatively broader and its sculpture is stronger. The spiral sculpture is the only completely consistent difference between the two species. In P. deserta the interspaces are wider and the ribs are usually alternately of primary and secondary strength. A small proportion of specimens of one species may show one or more of the features listed below which normally distinguish it from the other. In P. deserta the collabral ribs tend to increase in strength with size, whilst they be- come obsolete by the third whorl in all but a few P. solanderi. The subsutural platform is broader and more strongly grooved below in P. deseryta and the carina on the shoulder persists to adulthood in most individuals. In all four of the British species of Psewdocominella, the early teleoconch has such similar sculpture, shape and rate of whorl expansion that juveniles are difficult to tell apart. The nodiferous, carinate, shoulder of this stage is most noticeable and of shortest duration in P. solanderi which is the least strongly sculptured when adult. DESORINASSA gen. nov. (Pl. 4, figs 4-9, Pl. 5, figs 1-5) TYPE SPECIES. Buccinum desor Deshayes, 1865. DracGnosis. Agrees with Pseudocominella except in the following respects: subsutural platform not always developed; sculpture very weak; columella curved to the right above columellar plait; growth lines of fasciole not particularly rugose and not very sinuous; ridge above fasciole not strong; outer lip smooth within. OTHER SPECIES ASSIGNED. *Buccinum acies Watelet, 1853 (I[RScNB) ; **Cominella bonnecarret Furon 1948; *Buccinum latum Deshayes, 1865 (EMP); Buccinum ovatum Deshayes 1835 (EMP, MNP). **Cominella supracostata Traub, 1938; *Desorinassa williamst sp. nov. (see footnote for explanation). GEOLOGICAL RANGE. Palaeocene (Thanetian) to Lower Eocene (Cuisian). REMARKS. Desorinassa may be distinguished from Pseudocominella by its curved columella and its weaker sculpture, in particular the almost complete lack of collabral elements. The fasciole is more like that of Keepingia and Thanetinassa. The specimens of Buccinum latum Deshayes, (Deshayes Colln., EMP) figured herein as pl. 4, fig. 8 and pl. 4, fig. 9 are selected respectively as LEcToTYPE, and PARALECTOTYPE of this species. They are labelled as being from Abbecourt and Bracheux with no indication as to which specimen comes from which locality. These are the only localities quoted by Deshayes in his original description (1865 : In these lists * denotes species in which the protoconch has not been studied, ** denotes that no specimens have been examined. 202 SOME ENGLISH 501). It would appear that his illustration of the apertural view of the species (1866, pl. 93, fig. 27) is composed of the outline of the lectotype upon which has been superimposed the better preserved outer lip and stronger columellar and parietal callus of the paralectotype. Desorinassa desori (Deshayes, 1865) (Pl. 4, figs 4-6) 1865 Buccinum desori Deshayes; 501. 1866 Buccinum desort Deshayes; Deshayes, pl. 93, figs 16, 17. 1889 Cominella desovi (Deshayes) Cossmann: 137. 1901 Cominella desovi (Deshayes); Cossmann: 150. 1911 Cominella desovi (Deshayes); Cossmann & Pissaro, pl. 37, fig. 178-3. 1934 Comznella desov1 (Deshayes); Cooper: 8. 1936 Cominella desort (Deshayes); Farchad: 22. 1963 Cominella desovi (Deshayes); Glibert: 67. MATERIAL STUDIED. ENGLAND, Palaeocene, Thanetian, Thanet Sand, Reculver, Kent GG 19704 (Wr) (see remarks p. 203). FRANCE, Thanetian, Jonchery-sur- Vesle, France, GG 19705/1-3 (Wr); G 19438/1, 2 (L. Staadt). FURTHER OCCURRENCES. ENGLAND, Thanetian, Bishopstone, near Reculver, (Cooper, 1934 : 8) (see remarks, p. 203). FRANCE, ‘Sables inferieurs’ (= Thanetian), Abbecourt; Chalons-sur-Vesle; Jonchery; Saint Martin-aux-Bois (Deshayes, 1865 : 501). Chalons; Chenay; Jonchery; Noailles (Glibert 1963 : 67). Bracheux, near Beauvais (Farchad, 1926 : 22). DiaGnosis. Moderately large, stout, almost smooth Desovinassa with broad, sloping subsutural platform; mean spire angle about 60°; sutural slope reversed just prior to adult aperture; ridge on adapical side of fasciole very weak; outer lip not striate within. DESCRIPTION. The protoconch is worn but appears to have been similar to that of Pseudocominella, consisting of about three smooth, naticoid whorls; initially it may have been somewhat flattened. The teleoconch has up to six whorls which are convex-sided with the periphery lying just above the suture. A cord may be seen below the suture on the first whorl of the teleoconch. The material is not well enough preserved to show the development of the subsutural platform. It is first recognizable by the third whorl of the teleoconch where either two (or in some specimens three) spiral ribs, situated high on the ramp near the suture, form an angulation in the whorl profile. Subsequently the platform becomes flatter, though the slope is never less than 45°. It reaches a breadth of nearly 1-5 mm. No groove is developed but the profile is slightly concave just below the platform. Spiral sculpture is poorly developed, there are two or three weak grooves just below the subsutural platform and a mixture of about fifteen weak grooves and step-like ribs on the neck region of the body whorl. Collabral sculpture is absent except for growth rugae, these are strongest away from the periphery, becoming quite notice- able on and just below the subsutural platform and to a lesser extent on the neck PALAEOGENE NASSARIIDAE 203 region. The ridge marking the adapical margin of the fasciole can hardly be distin- guished in smaller specimens. No ribbing has been seen inside the outer lip. The sutural slope becomes reversed immediately prior to the final aperture. DIMENSIONS. I 2 3 4 7 10 II GG 19704 22+ 11+ 14+ — a= — — GG 19705/3 II+3 6:5 63 4+ 12+ 1°74 1°79 GG 19705/1 1373 8-2 7:9 44 17+ 1-62 1:68 G 19438/2 16°5 99 9°7 = = 1-67 1°70 G 19438/1 17°6 10-0 10°6 5+ 18+ 1:76 1°66 (see p. 192 for explanation) Remarks. D. desori differs from D. acies (Watelet) (pl. 5 fig. 5) and D. ovata (Deshayes) (pl. 5, fig. 4) in being larger, having a subsutural platform and lacking regular spiral sculpture. A comparison with D. williamsi is given under the latter. The single English specimen of D. desort seen is badly damaged but agrees in all essential respects with the French specimens. It is thought that Wrigley acquired many of the more interesting specimens of J. E. Cooper’s collection and GG 19704 was found in a box which was probably made by Cooper. It is therefore possible that this specimen is that listed by Cooper (1934 : 8) and that only one specimen has so far been discovered in English rocks. However, A. J. Rundle (personal communication) after examining its matrix reports that it is a silicified specimen from the Corbula bed in the cliff at Reculver and not from the morrisi bed, as listed by Cooper. Desorinassa williamsi sp. nov. (Pl. 5, figs 1-3) HoLotyPe. Palaeocene, Ypresian, Bognor Rock, foreshore at Bognor, Sussex, GG 19706 (B. A. Williams Colln.). PARATYPES. Some horizon and locality, GG 19707, (Rev. A. Fuller Colln.), and two internal moulds GG 19708/1-2 (Aistory unrecorded). FURTHER MATERIAL AND RECORDS. Unknown. DiacGnosis. Large, stout, almost smooth Desorinassa; mean spire angle of about 60°; subsutural platform lacking; whorl profile convex; interior of outer lip not denticulate. DESCRIPTION. The protoconch is unknown. Four whorls of the teleoconch survive; at least one or two earlier whorls are missing in all specimens. The whorls, which are broadest at the abapical suture, are convex except where recurved at the adapical margin to form an adpressed or almost flush suture. Spiral sculptural elements are lacking except for between six and eight weak ridges on the neck of the body whorl. Irregularly spaced collabral growth rugae of varying strength are developed. These, though weak, are distinctly stronger than growth lines. No ribs are present on the inner surface of the outer lip. Both internal moulds have a collabral groove about a quarter whorl before the aperture. 204 SOME ENGLISH DIMENSIONS. I 2 3 7 10 II GG 19706 HT 27°8+ 17°3 17:0 8+ — — estimated 29 — — — 1-68 I-71 GG 19707 PT 27°2+ 17°2 179 5+ -- — estimated 30 —_ — — 1-74 1:67 (see p. 192 for key) REMARKS. This species may be easily distinguished from D. desort which is smaller, and has a broad, sloping, subsutural platform with spiral ribbing on and near it. Two Cuisian species from the Paris Basin, D. ovata (Deshayes) (pl. 5, fig. 5) and D. acies (Watelet) (pl. 5, fig. 6) are much smaller than D. williamsi and possess fine spiral ribbing. The former is of a similar shape to D. williamsi but the latter is distinctly narrower. WHITECLIFFIA gen. nov. (Pl. 5, figs 6-12; Pl. 6) TYPE SPECIES. Buccinum suturosum Nyst, 1836. Diacnosis. Differs from Pseudocominella only in the following respects: height less than 20 mm; protoconch of one and half whorls, initially deviated and bulbous; suture either weakly canaliculate or with narrow, flat, subsutural platform; no shoulder developed; collabral sculpture either absent or of weak grooves or ribs; columellar and parietal callus distinctly raised above general level of shell surface; fasciole similar to that of Pseudocominella but ridges and growth lines less prominent ; outer lip seldom ribbed within. OTHER SPECIES ASSIGNED. Whitecliffia twmida sp. nov. GEOLOGICAL RANGE. Upper Eocene (Late Priabonian) to Middle Oligocene (Rupelian). REMARKS. This genus seems likely to be an offshoot of Pseudocominella. It may readily be distinguished by its protoconch and by the relative smoothness of the teleoconch. Whitecliffia suturosa (Nyst, 1836) (Pl. 5, figs 6-11; Pl. 6, figs 1-4) 1836 Buccinum suturosum Nyst: 36, pl. 3, fig. 96. 1843 Buccinum sutuyosum Nyst; Nyst: pl. 43, fig. 16. 1845 Buccinum suturosum Nyst; Nyst: 579. 1867 Buccinum suturosum Nyst; von Koenen: 83, pl. 6 (1), fig. 3a, b, c. 1880 Cominella (Buccinum) ventricosa Edwards MS; Judd: 154. 1880 Stvepsidura suturosa (Nyst) Bosquet: 467. 1881 Cominella (Buccinum) ventricosa Edwards MS; Keeping & Tawney: 116. 1886 Stvepsiduva suturosa (Bosquet); Vincent: 7. 1889 Buccinum (Cominella) suturosum Nyst; von Koenen: 242, pl. 21, figs 10a, b, c, Ifa, b. 1901 Cominella suturosum (Nyst); Cossmann: 150. PALAEOGENE NASSARIIDAE 205 1943 Novthia (Cominella) suturosa (Nyst) Albrecht & Valk: 62, pl. 5, figs 140, 141. 1954 Northia (Cominella) suturosa (Nyst); Glibert & de Heinzelin: 367, 382, 392, 393, pl. 7, figs 4a, b. 1954 Babylonia (Peridipsaccus) avchambaulti Meunier; Glibert & de Heinzelin: 368, 382, pl. 7, fig. 6. 1963 Cominella suturosa (Nyst); Glibert: 68. LectotyPE. Middle Oligocene, Lower Rupelian, Callista kickxi horizon; IRScNB IG 3862, Klein Spauwen, Belgium (selected Glibert, 1963 : 68, figured Glibert, & de Heinzelin 1954, pl. 7, fig. 4b). PARALECTOTYPE. Same details, 7RScNB IG 3863, (figured Glibert & de Heinzelin, 1954, pl. 7, fig. 4a). MATERIAL STUDIED. ENGLAND, Upper Eocene, late Priabonian, Middle Headon Beds, Royden, GGuig9717 (Ed). Brockenhurst Bed, Whitecliff Bay, I.0.W., GG 19718 (Wr); (Cu). GERMANY, Lower Oligocene, Lattorfian, Latdorf, GG 19719/ 1-2 (Ed); GG 19720/1-2 (history unknown). Brtcium, Lower Oligocene, Lattorfian (= Lower Tongrian), Grimmertingen [RScNB sample IG 5002 (de Looz). Middle Oligocene, Lower Rupelian, Klein Spauwen, GG 19721 (de Koninck). Berg, IRScNB sample IG 4285 (Bosquet). FURTHER OCCURRENCES. HOLLAND, Lower and Middle Oligocene, South Limburg, in mine shafts (Albrecht & Valk, 1943 : 62). GERMANY, Middle Oligocene, Rupelian, Stettin Sands, Magdeburg, (von Koenen, 1889 : 242). Diacnosis. Whiteclifia with canaliculate suture; mean spire angle between 43° and 80°; whorls evenly convex with no shoulder developed; sculpture absent except for occasional weak spiral grooving; outer lip denticulate within. Description. All the Belgian specimens and Wrigley’s specimen from Whitecliff Bay are too badly worn to show sculptural details. The description is based mainly on the shells from Royden (GG 19717), Latdorf (GG 19719—20) and a specimen from Whitecliff Bay (Cz). The mean spire angles of the lectotype, paralectotype and GG 19721 from Klein Spauwen are all under 50°. The spires of several specimens from Berg are also more acute than average. The English, Grimmertingen and Latdorf specimens are normally more tumid with the spire angle between 50° and 55°. The appearance of the suture is so considerably changed by wear, that this feature cannot be reliably used as a taxonomic feature. The better-preserved English and German specimens as well as most of the Belgian specimens have a narrow canaliculate suture and no subsutural ramp. Many of the Grimmertingen specimens, however, seem to have a fairly broad ramp and a less incised suture. In most specimens the sutures are not channelled for the first three or four whorls. Below the suture the whorl profile is rather gently convex, and its periphery coincides with the abapical suture in most specimens. In shells with relatively acute spires, however, whorls are enveloped well below the periphery. About half the specimens have a relatively strong pad of callus in the parietal and columellar region which stands out above the rest of the shell-surface. The fasciole is not 206 SOME ENGLISH very rugose but the ridges on either side are fairly sharp when unworn. Only one specimen, the largest in sample JRScNB 1G 5002 from Grimmertingen, has ribbing on the inner surface of its outer lip; this feature is also shown in one of von Koenen’s figures (1889, pl. 21, fig. rob). All the spire whorls of GG 19717 bear weak spiral grooves; eight may be counted on the penultimate whorl, but they become obsolete on the body whorl, which is smooth except for ten ribs on the neck. On another larger specimen (Cz), spiral grooves which are too poorly preserved to count, are present on the fourth and fifth whorls whilst the last two are smooth. GG 19719/1 and GG 19720/1, from Latdorf, however, have twelve grooves on the penultimate whorl, twenty-four on the body whorl and thirteen ribs on the neck below. Another specimen from Latdorf GG 19719/2 whose surface is also well preserved shows traces of spiral sculpture only on the neck. Grimmertingen specimens are remarkable in having two or three relatively strong grooves just below the subsutural platform; below this weaker grooves may sometimes be distinguished. All other Belgian specimens seen from Berg and Klein Spauwen lack this grooving. No specimens show any trace of collabral sculpture other than growth lines. DIMENSIONS. I 2 3 10 II factual 14°5 8-3 8-6 — — eeony ‘Lestimated 15°5 — = 1:87 1:80 GG 19720/1 16-1 QI gl 1°77 777 GG 19720/2 16-9 10:0 8-9 I-69 I-90 f actual 16:8 9:2 8-6 — = COs ‘estimated 17°5 — — 1:90 2°03 GG 19719/2 18-8 Io-l 9:0 1-88 2:09 (see p. 192 for key) REMARKS. GG 19717 from Royden was labelled as Cominella ventricosa Edwards MS by Edwards; Keeping & Tawney’s (1881 : 116) record is of this specimen. Judd’s list (1880 : 154) was also compiled from the Edwards Collection, so his record of C. ventricosa from Colwell Bay, not Royden, is almost certainly a clerical error. Von Koenen (1867 : 83; 1889 : 242) quotes Royden asa locality. It therefore seems quite likely that he also examined this specimen, recognizing it to belong to W. suturosa. Grimmertingen specimens seem to represent a fairly marked variety, distinguished by the strength of the spiral grooving and by the usually less acute spire. A damaged specimen, /RScNB figd. spec. no. 4070 I.S.T. from Grimmertingen (pl. 5, figs 8a, b) identified as Babylonia (Pertdipsaccus) archambaulti (Meunier) by Glibert & de Heinzelin (1954 : 368, 382, pl. 7, fig. 6) is definitely a Whiteclifia and may be an exceptionally large, broad example of the present species with a spiral angle of 80°. Whitecliffia tumida sp. nov. (Pl. 6, figs 5-10) HoLotyre. Upper Eocene, Late Priabonian, Middle Headon Beds (Venus Bed), Whitecliff Bay, 1.O.W. GG 19722/r (Cu). PALAEOGENE NASSARIIDAE 207 PARATYPES. Same locality and horizon, GG 19722/2-7 (Cu); GG 19724 (Cu); GG 19723/1-3 (Wr); Sedg. Mus. C 29305 (KGT); (Cu); (S); GG 19756/1-3 (S) from Bed O of Stinton (1971 : 405). FURTHER OCCURRENCES. Unknown. Diacnosis. Broad Whiteclifia with teleoconch of four or five whorls; mean spire angle of about 65°; whorl profile strongly convex; last whorl often rather swollen; collabral sculpture normally of up to sixteen strong, rather angular, costae which becomes less frequent and either die out or are replaced by broad flexures on pre-penultimate or penultimate whorl; spirally multistriate; subsutural collar obsolete; suture becoming slightly canaliculate on last whorl or so; outer lip rarely internally ribbed. DEscripTIon. None of the specimens examined is complete, and the form of the aperture is seen clearly only in one paratype. Whorls are enveloped just below the periphery. The whorl profile is more strongly convex in the early whorls before the collabral ribbing dies away. The whorl profileis almost horizontal at the adapical suture, and usually becomes grooved by the third or fourth whorl. The subsutural platform reaches a maximum width of 0-5 mm but seldom exceeds 0-3 mm wide; it is not separated from the rest of the whorl-side by a rib. The teleoconch starts with numerous weak, slightly opisthocline, collabral ribs. In the next quarter whorl these increase in strength and gradually adopt the orthocline attitude of the collabral ribs of the later teleoconch. There are normally about sixteen rather prominent, angular, collabral ribs on the first whorls of the teleoconch. They are weakest just below the suture and are of maximum strength from the shoulder to the adapical suture. On the third teleoconch whorl they weaken considerably; they may be just discerned on the fourth as faint flexures, which, on the body whorl, can be seen to die away below the periphery. GG 419724 is unusual. On the second whorl of its teleoconch about ten slight collabral crenulations occur for nearly half a whorl. This is strongly reminiscent of the start of collabral ribbing in other speci- mens, but dies away instead of developing further. On this specimen no other colla- bral sculptural elements exist except for growth lines. In the other specimens the spiral sculpture is first noticed within half a whorl of the collabral sculpture appearing. It consists of between six and ten irregularly spaced bands. Usually they can be seen to be ribs separated by interspaces wider than themselves. Often, however, and particularly on the last whorl, this separation is reduced and the shell becomes spirally multi-grooved. Bands of secondary strength may sometimes be seen. Spiral sculpture is either weak or absent from the periphery of the body whorl. Between seven and ten fairly prominent cords are present on the neck region. Internal ribbing of the outer lip is seen clearly only on one specimen (Cu) which has fifteen ribs extending back from the aperture for at least 45°. Irregular collabral colour bands up to about 1 mm broad separated by gaps of similar size may be seen on the last whorl of GG 19722/3 and one specimen (Cz); neither of these specimens shows spiral colour banding. GG 19722/6, however, shows spiral but not collabral colour banding. There isa dark band about 0-5 mm wide on the subsutural platform 208 SOME ENGLISH with a slightly narrower band just below the platform, below this there are indications of two more narrower bands well above the periphery. DIMENSIONS. I 2 3 8 10 II GG 19722/1 HT 12°9 8-0 707] 9+ I-61 1-67 GG 19722/2 PT 5:0 3:2 371 6+ 1°56 1°61 GG 19722/4 PT 6-7 4:0 3°9 8+ 1-67 1-72 GG 19722/5 PT 75 4:7 4:0 9+ I-60 1°88 GG 19722/7 PT 10:2 6-4 6-9 9+ 1°59 1-73 (see p. 192 for key) REMARKS. This species closely resembles W. suturosa in shape and size. In the latter, collabral sculpture is almost non-existent, spiral sculpture though similar is weaker, and grooving of the suture occurs earlier and is more noticeable. As indicated, GG 19724 has unusual sculpture. It seems advisable, however, not to describe a separate species on the basis of a single specimen. COLWELLIA gen. nov. (PL. 7) TYPE SPECIES. Cominella flexuosa Edwards in Lowry, Etheridge and Edwards, 1866. Diacnosis. Agrees with Pseudocominella except in the following respects: seldom exceeds 20 mm height; protoconch of about two and a half smooth whorls, small, about I mm high, initially slightly flattened, first whorl very small; teleoconch sculpture developing gradually with collabral sculpture developing first; teleoconch with spiral and collabral sculpture of varying strength; subsutural platform either poorly developed or absent ; growth lines of fasciole broadly sinuous as in Desovinassa ; outer lip often denticulate within. OTHER SPECIES ASSIGNED. WNassa antiquata Gabb, 1864 (pl. 7, fig. 11) Buccinum auversiense Deshayes, 1865 (pl. 7, figs 6, 7); Ancillaria bretzi, Weaver, 1912 (pl. 7, figs 8, 9); *Nassa cretacea, Gabb, 1864; *Molopophorus tejonensis Dickerson, 1915). (See p. 201 for explanation). GEOLOGICAL RANGE. Middle Eocene (Auversian) to Upper Eocene (late Pria- bonian), Europe. Middle Eocene (Domengine Formation, California and Umpqua Formation, Oregon) to Upper Eocene (Cowlitz Formation, Washington and Tejon Formation, California), U.S.A. REMARKS. The growth lines of the species assigned above to Colwellia are of the type found in Pseudocominella and all the other new genera described herein: they are approximately orthocline but are bent in an adoral direction in the sutural (adapical) region. In the lectotype (pl. 8, fig. ra, b) of the type species of Molopo- phorus, Bullia (M.) striata Gabb, 1869 (ANSP 4249), the growth lines are bent away from the aperture in the sutural region, which appears to be canaliculate, as in Brachysphingus (pl. 8, figs 3, 4). The same feature is shown by U.Cal. 7182-30750 PALAEOGENE NASSARIIDAE 209 (pl. 8, fig. 2). However, another specimen figured as Molopophorus striatus by Dickerson (1915 : 67, pl. 8, fig. 6, given as fig. 7 in explanation to the plate) seems to be misidentified and is almost certainly a Colwellia (pl. 7, fig. 10). The type of growth line found in Molopophorus (s.s.) and Brachysphingus is also found in Cyllene, a rare living genus of Nassariidae. This feature is found in other widely-separated prosobranch superfamilies and is often associated with a gutter or canaliculate suture which accommodates a posterior filament as in Oliva (Volutacea) (Tryon, 1883 pl. 3) and Terebellum (Strombacea) (Jung & Abbott 1967, p. 446, pl. 319). Sections through the columella of Brachysphingus show that it lacks the terminal plait found in all definite Nassariidae including Cyllene. Although no specimens of Molopophorus s.s. were available for sectioning (the lectotype is the only specimen in the ANSP collections—Dr H. G. Richards, personal communication), the external similarities of the two genera are so great that it is doubtful if either belongs to the Nassariidae. Their general shell features including the fasciole ridge and reverse S-shaped growth lines suggest that they may best be placed in the very broad Buccinidae (sensu Wenz, 1941-1943). In fact, it seems that Molopophorus (Gabb, 1869 : 156) is a synonym—based on immature material—of Brachysphingus (Gabb, 1869 : 155), which is here selected as generic name of the taxon. A large number of North American west coast Tertiary species ranging in age from the Meganos Formation (lower Eocene) of California (Clark & Woodford, 1927 : 117) to the Miocene of Washington State (Weaver, 1942 : 463-471) have been assigned to Molopophorus and the genus is also recorded from the Oligocene and Miocene of Japan (Oyama, 1960 : 73-74). These species differ considerably in form. Vokes, moreover, (1939 : 140-141) mentions that there were two or three distinct types of nuclear whorls among the Californian species ascribed to Molopo- phorus and suggested that they were unlikely to be congeneric. Most of these species need assigning to new genera—a task beyond the scope of this work. It would appear, however, that they are fairly closely related to the European forms described herein and would also seem to belong to the Nassariidae. Gabb (1864 : 97) described Nassa cretacea and Nassa antiquata both of which have been regularly assigned to Molopophorus by American workers in this century. Nevertheless, he himself still assigned them to Nassa (1869, 219) in the paper in which he described Molopophorus. The difficulties of interpreting Molopophorus correctly have hindered workers who have suspected a connection between it and European species previously assigned to Cominella. We have assigned to Colwellia only those American species which have a very strong resemblance to C. flexuosa and C. auversiensis. In Desorvinassa the ramp is convex and the collabral sculpture weak. In Colwellia the ramp is concave and the collabral sculpture is much stronger on the earlier whorls. D. williamsi (Ypresian) and C. auversiensis (Auversian) appear to form a link between the two genera. The differences in the protoconch, given in the diagnosis, are sufficient to distin- guish Colwellia from Keepingia, which almost certainly arose from it. Cc 210 SOME ENGLISH Colwellia flexuosa (Edwards, 1866) (Pl. 7, fig. 1-5) 1866 Cominella flexuosa Edwards in Lowry, Etheridge & Edwards, pl. 3. 1880 Cominella (Buccinum) flexuosa Edwards MS; Judd: 154. 1881 Cominella (Buccinum) flexuosa Edwards MS; Keeping & Tawney: 105, 116. 1889 Cominella flexuosa Lowry; Bristow: 290. 1891 Cominella flexuosa Edwards MS; Newton: 168. 1921 Cominella flexuosa Edwards; Wrigley: 139. 1926 Cominella flexuosa Edwards; Jackson: 356. 1963 Cominella flexuosa Edwards in Lowry; Glibert: 67. LECTOTYPE (designated herein). Upper Eocene, late Priabonian, Middle Headon Beds (presumably Venus Bed), Colwell Bay, I1.0.W., GG 19709/1 (Ed). PARALECTOTYPES (designated herein). Same details, GG 19709/2-30. MATERIAL STUDIED. Restricted to Middle Headon Beds. Venus Bed, Colwell Bay, GG 19710/1-11 (Wr); GG 19711 (Nu); (Cu); (S); Sand. Mus. 4778, 748; Sedg. Mus. C 28281-4 (K&T). Headon Hill, 1.0.W., GG 19712/1-3 (St]B); Sedg. Mus. C 28645-9 (KGET). Venus Bed, Whitecliff Bay, 1.0.W., GG 19713 (Cu). Brocken- hurst Bed, Brockenhurst, GG 19714 (Ed). FURTHER OCCURRENCES. Unknown. Diacnosis. Colwellia with narrow sloping subsutural collar and rather broad concave subsutural ramp; mean spire angle between 55° and 70°; angular can- cellate ribbing on early whorls, becoming almost smooth, particularly below peri- phery of last whorl; outer lip often weakly denticulate within; obvious sinus in growth lines of peripheral region, and in some specimens a posterior siphonal notch, developed at maturity. DESCRIPTION. The protoconch is worn in all specimens, the junction with the teleoconch cannot be seen. The aperture is slightly more than two-thirds shell height. The teleoconch consists of up to five whorls. The suture normally lies on the periphery, sometimes on later whorls it lies just below as the sutural slope in- creases slightly. Below the rounded subsutural band the very slightly concave ramp slopes away to the shoulder. Three or four weak spiral ribs occur on and below the shoulder of spire whorls; that on the shoulder is the strongest. These ribs become noticeably weaker on the last whorl. Up to two more spiral ribs, which are usually weaker, may occur on the ramp. Between five and ten rather widely spaced spiral cords are present on the neck of the body whorl. The collabral ribs are stronger than the spiral ones and are like crested-waves in cross-section. There are about twenty-one collabral ribs on the spire whorls of the teleoconch. On the first half of the body whorl there are about sixteen increasingly weaker ribs. Collabral ribbing becomes virtually obsolete on the final half whorl. The main indication of its presence is slight axial nodes on the spiral ribs. Similar nodes are formed on the spire whorls where the collabral and spiral ribs intersect giving a rather cancellate appearance. In some specimens the outer lip is smooth internally, in others there are up to fifteen unevenly spaced weak denticles about 1 mm long PALAEOGENE NASSARIIDAE 211 and set about I mm inside the aperture, and a few shells show another row of denticles about 45° inside the aperture. As well as the broad sinus in the growth lines of the peripheral region of mature specimens there is a slight posterior siphonal notch developed in the outer lip where it meets the suture (pl. 7, fig. 2c). Its strength seems related to the size of the shell and it is sometimes absent in smaller specimens. It may be presumed to mark the reaching of full growth as there is no trace, except occasionally on the last quarter whorl, of its previous existence on earlier growth lines. In a few specimens the colour pattern is clearly visible under ultra-violet light. A pair of spiral bands, about 0-5 mm wide, separated by about 0-2 mm lie on the periphery (thus appearing just above the suture on spire whorls). Another pair of bands, about half the breadth, occurs at the top of the neck. DIMENSIONS. I 2 3 4 5 6 7 8 9 10 II GG 19709/1 LT 14:2 05 QO 65 TAU 2 20 eeted| 6 5 1:49 1°58 GG 19713 — 5°4 3°3 i Lp s20) 5H Aneel 1-630 51-65 GG 19709/6 PLT 10:8 6:8 6-05 te 2222 rz 4 © £59 1-64 GG 19709/2 PLT 14°3 8-3 S:0mnn Pp it SNE 7) 6) 1-72 1-66 GG 19709/4 PLT 14°9 9°4 OR § 18 23; 15-+ 7 OME SOM O2 GG 19709/3 PLT 20°'I 13:0 It2 6 Oo. 25, 28 Io) RG) RON, StF) (see p. 192 for key) REMARKS. Following Wrigley’s (1921 : 139) reasoning, we ascribe the author- ship of this species to F. E. Edwards. Its resemblance to the earlier C. auversiensis (Deshayes) (pl. 7, figs 6, 7) is striking. Both species have teleoconchs so similar to those of Keepingia that it is necessary to compare them with both K. gossardi (Nyst) (pl. 8, figs 5-7) and K. cassidaria (Sandberger) (pl. 8, fig. 9). C. auversiensis differs from both C. flexuosa and K. cassidaria in being less tumid, having a less concave ramp and less marked shoulder whilst its collabral sculpture, though weaker on early growth stages is more persistent and its growth lines less flexuous. Sand- berger figures two varieties of K. cassidaria. The first (1863, pl. 20, figs I, 1a, 1d, Ie) is characterized by about four strong spiral ribs on the spire whorls and about ten on the body whorl. It lacks collabral sculpture on the last and sometimes on the penultimate whorl. The second variety (1863, pl. 20, figs 1b, 1c) which he refers to (p. 230) as var. cancellata, is less common in B.M.(N.H.) collections. It has a cancellate appearance; the primary spiral ribs are separated by secondary ones and numerous collabral rugae are developed. Both forms of K. casstdaria may be readily distinguished from C. flexuosa by their sculpture. K. gossardi differs from C. flexuosa by being larger and having a relatively smaller aperture, less sinuous growth lines, and more definite subsutural collar with a tendency for the suture to be canaliculate. The sculpture of K. gossavdi is much stronger and persists to maturity. It has numerous spiral ribs and about fifteen collabral folds per whorl. Two specimens in the F. E. Edwards Collection (GG 19725-6) were labelled by Edwards as being from the Bracklesham Beds of Brook, Hampshire. The first was correctly identified as flexuosa and appears from the matrix to be from the Brockenhurst Bed of the Brockenhurst area. The second is likely to be from Brook but is a young Strepsidura turgida (Solander). The sculpture on the spire whorls of Ct 212 SOME ENGLISH these two species is remarkably similar. Edwards’ label may have caused Keeping & Tawney (1881 : 105, 116) to indicate that the species occurs in strata older than the Headon Beds. There is a remarkably small variation in size of the known specimens. Most of those in B.M.(N.H.) collections lie between 12 and 16 mm in height. The largest and smallest specimens whose dimensions are given here are exceptional. KEEPINGIA gen. nov. (Pi; 8, figs 5-9) TYPE SPECIES. Buccinum gossardi Nyst, 1836. Diacnosis. Agrees with Colwellia, except in the following respects; sometimes exceeding 20 mm in height; protoconch initially slightly heterostrophic, of one and a half smooth whorls, separated from teleoconch by a whorl on which there are about 20 decreasingly opisthocline collabral folds, with spiral sculpture developing on later part. OTHER SPECIES ASSIGNED. **Cominella annandalei Vredenburg 1925; **Cominella aturensis Peyrot, 1927; Buccinum bolli Beyrich, 1854; Buccinum cassidaria Sand- berger, 1863; **Cominella praecedens Peyrot 1927; Nassa tarbellica Grateloup, 1834; **Buccinum uniseriale Sandberger, 1863. (See p. 201 for explanation.) GEOLOGICAL RANGE. Lower Oligocene (Late Lattorfian) to Lower Miocene (Burdigalian). REMARKS. The similarities of the teleoconch suggest that Keepingia is close to Colwellia but its consistently different protoconch in all species examined justifies its separation. It is the only genus described herein that survives into the Miocene. The first occurrence is of the type species in the Upper Tongrian Henis Clay of Belgium, the species persists into the Rupelian of Belgium and France. The last definite occurrence of the genus is that of K. tarbellica in the Burdigalian of the Bordeaux basin. As indicated above. no specimens of four of the species listed were available for study, so the stratigraphical and geographical range of the genus as shown on Table 1 can only be tentative. This is particularly unfortunate in the cases of C. annandalet from the Nari (Oligocene) of the Indian sub-continent. No species are known from England. THANETINASSA gen. nov. (Pl. 9, figs 1-7) TYPE SPECIES. Buccinum bicorona Melleville, 1843. Diacnosis. Phos-like, relatively tall, up to about 20 mm high; teleoconch of up to six whorls; mean spire angle of about 45°; aperture about half shell height; strong, PALAEOGENE NASSARIIDAE 213 slightly sloping, subsutural platform developed; both spiral and collabral ribbing strong, with beads produced at intersections; outer lip notably convex, only slightly produced anteriorly; growth lines markedly recurved in neck region immediately prior to terminal aperture, to form a ‘stromboid-like’ notch, associated with particu- larly strong spiral ribbing; columella bent to the right above a terminal plait; growth lines of fasciole as in Desovinassa, columellar and parietal callus weakly developed. OTHER SPECIES ASSIGNED. None. GEOLOGICAL RANGE. Palaeocene (Thanetian). REMARKS’ The strong ‘stromboid-like’ notch developed at the final aperture immediately distinguishes this genus from the others described in this paper. All these have a slight sinuosity in a similar position, but nothing so marked. In living Strombus the notch accommodates the right eye peduncle. Both Phos and Nassarius s.l1. have similar notches. It is always strong in Phos, but is of variable strength in Nassarius. The eye tentacles of Phos are joined at their base to form a Y-shaped structure (Tryon, 1881 : 215, pl. 83, fig. 484), in Nassarius (Tryon, 1882, pls 7, 9, 12, 14, 15, 18) the arrangement is rather similar. It seems possible that this structure as a whole, rather than a single tentacle, lay underneath the notch. It is hoped that the relationship between the Nassariidae and the Phos-like genera which were assigned to the Buccinidae by Wenz (1941) will be discussed in a future paper in greater detail than would be appropriate here. The curved columella, the form of the columellar plait, with its strong posterior boundary ridge and the broad fasciole of Thanetinassa are very similar to those of contemporary Desovinassa, which differs in having very weak sculpture The later genus Pseudocominella is strongly sculptured and P. solanderi is similar in shape and size toT. bicorona. Any similarity in their sculpture, however, is superficial. Moreover, the columellar and fasciole features of the two genera are very different. Thanetinassa should be compared with Buccitriton and Tritiaria (both assigned by Wenz, 1941 : 1177-1178, to the Buccinidae) and Sagenella (tentatively assigned by Wenz, 1943 : 1224, to the Nassariidae) All three genera occur in the Lower Tertiary of the South-Eastern United States. In all three the fasciole is bounded posteriorly by a much weaker ridge than in Thanetinassa, and the stromboid notch is either weak or absent and is never associated with strong spiral ribs. In Buccitriton the columella is short, terminating well above the most anterior part of the shell, which in this case is the outer lip. In Thanetinassa, on the other hand, the terminal columellar plait forms the most anterior portion of the shell. In Trttiarza the fasciole is much narrower than in Thanetinassa and is spirally striate. Thanetinassa bicorona (Melleville, 1843) (Pl. 9, figs 1-7) 1843 Buccinum bicorona Melleville: 12, 73, pl. 10, figs 4, 5. 1850 Buccinum bicorona d’Orbigny: 303. 1865 Buccinum quoesitum Deshayes: 503. 214 SOME ENGLISH 1866 Buccinum quoesitum Deshayes: pl. 93, figs 9-12. 1889 Cominella bicoronata (Melleville) nom. mut. Cossmann: 138. 1901 Cominella bicovona (Melleville); Cossmann: 150. 1911 Cominella bicoronata (Melleville); Cossmann & Pissaro, pl. 37, fig. 178-7. 1934 Cominella bicovonata (Melleville); Cooper: 7. 1963 Comunella bicorona (Melleville) ; Glibert: 66. MATERIAL STUDIED. ENGLAND, Palaeocene, Thanetian, Upper Thanet Sands’ Corbula Bed. Herne Bay, Kent, G57073-8 (J. E. Cooper Colln.). Bishopstone’ Herne Bay, G 60746 (J. E. Cooper Colln.); GG 19741/1-2 (Wr); GG 19742 (Wr); GG 19743 (A. G. Davis Colin.). Bishopstone Gap, GG 19744 (Ru). Arctica morrist Bed, foreshore exposure approx. 400 yards E. of Bishopstone Gap, GG 19745 (Ru). FRANCE, Palaeocene, Thanetian, Chalons-sur-Vesle, Marne, G 19357/1-4 (L. Staadt Colln.); GG 19740 (Wr); 5 shells EMP (Deshayes Colln.). FURTHER OCCURRENCES. FRANCE, Sable Inférieurs (Thanetian). TyPE LOCALITY Villiers-Franqueux (Melleville, 1843 : 73). Chalons-sur-Vesle, Brimont, Jonchery (Deshayes, 1865 : 503; Chalons-sur-Vesle (Briart & Cornet, 1871 : 31); Chalons-sur- Vesle, Chenay, Jonchery (Glibert, 1963 : 66). DiaGnosis. As for genus. Description. All the specimens are fragile, abraded, and with their earlier whorls either missing or decorticated. It is impossible therefore to separate the protoconch and teleoconch. GG 19741/2, however, is 12-3 mm high and has eight whorls. Neither definite ribbing nor the subsutural platform appear until the fourth whorl, so the first three may well comprise a rather acutely-spired, naticoid, proto- conch. Most of the spiral ribs are distinctly weaker than the collabral ones and the preservation in all specimens is such that spiral ribs cannot be seen until the third teleoconch whorl. Up to nine spirals ribs occur on the spire whorls and about twenty on the body whorl. The two strongest, which are stronger than any of the collabral ribs, are those forming the sloping subsutural platform and the shoulder. The concave ramp is formed in the interspace between these two ribs. Sometimes two or three other fairly strong ribs occur on the neck above the fasciole, marking a backwardly-directed fold in the growth lines. Sometimes a weak rib is developed above the main rib of the subsutural platform. The collabral ribs are more or less orthocline and decrease in strength below the periphery; on some specimens they may be seen to follow the above-mentioned fold on the neck. They are prominent, steep-sided, and separated by interspaces nearly twice their width. Their number remains fairly constant, throughout growth. On the last whorl, their spacing becomes less regular. An irregular, varix-like, pre-terminal aperture is often developed just after the commencement of the last whorl, being situated on the ventral surface of the shell, slightly to the left of the final aperture. Similar varices may well be developed later on the last whorl, GG 19744 having a total of five. Two weak spiral ribs are occasionally developed on the fasciole. Spiral ribbing has not been seen within the outer lip. PALAEOGENE NASSARIIDAE 215 DIMENSIONS. I 2 3 4 5 6 7 8 9 10 Ev 18°5 — — — —_- — 9 — — — — een rae hee 20:0 = 9:0 6 AS SES SS = 1-77 ee ee OL ONE te ac EON Ol ae | i — — prorat: ea 35 —- — 6 22 —- —- — — — 2°31 GG 19741/2 123. — 60 5-6 15 20 420 8 — — 2:01 155 85 83 6 23) 124 LS 8 — — — GG 19357/1 ee 160 — — —- —- — — ~ ~~ 188 = 1:93 GG 19357/2 146 7:0 83 6 TO} LO 2 8 — 2:08 1:76 GG 19357/3 12: Oo: Onn On) 6 22 Zien nO 7 — 1:70 I-61 Remarks. The English specimens are notably more strongly varicate on the last whorl than those from France. There are, however, no other apparent grounds for specific separation. This species strongly resembles Buccinum montense (pl. 9, figs 8a, b) described by Briart & Cornet (1871 : 30-31) from the Montian ‘Calcaire Grossier’ of Mons; they compare the two species in detail, and both were subsequently assigned to Cominella by Cossmann (1901 : 150). Examination of the fasciole of Buccinum montense from Mons (GG 71604—7, Chavan Colln.) show that this species is not a member of the Nassariidae but is congeneric with European Eocene and Oligo- cene species assigned to Pollia of the Buccinidae. Cossmann (1889 : 138) placed B. quoesitwm Deshayes (1865) in the synonomy of B. bicorona Melleville (1843) remarking that Deshayes had overlooked the earlier name. The whereabouts of Melleville’s specimens is unknown but his figures (1843, pl. ro, figs 4, 5) appear to agree with those of Deshayes (1866, pl. 93 figs 9-12) and Deshayes’ surviving specimens all from Chalons-sur-Vesle (EMP). Melleville’s locality was Villers Franquex, at which the only horizon exposed is the middle fossiliferous zone of the Sables de Chalons-sur-Vesle (Fritel, 1910 : 332). REFERENCES Axssott, R. T. 1960. The genus Stvombus in the Indo-Pacific. Indo-Pacific Mollusca, Philadelphia. 1 (2) :33-146, pls 11-117. ALBRECHT, J. C. H. & VaLtx, W. 1943. Oligocine Invertebraten von Sud-Limburg. Meded. geol. Sticht., Maastricht, Ser.C.IV. 1 : (3). 163 pp., 27 pls. BELLARDI, L. 1882. I molluschi dei terveni terziavi del Piemonte e della Liguria, 3. 253 pp., 12 pls. Turin. Beyricu, H. E. 1854. Die Conchylien des norddeutschen Tertiargebirges. Pt. 2. Z. dt. geol. Ges., Berlin, 6 : 408-500, pls. 9-14. Borson, S. 1820. Saggio di orittografia Piemontese. Memorie R. Acad. Sci. Torino, 25 : 180- 220, pl. 5. Bosguet, J. 1880. Fossiles du systéme Tongrien et de l étage inférieur Rupélien in DEWALQUE, G. Pyrodrome d'une description géologique de la Belgique: 2nd. ed.: 463-473. Brussels. Briart, A. & CorNET, F.-L. 1871. Description des fossiles du calcaire grossierde Mons. Pt. 1. Mém. cour. Sav. étr. Acad. v. Sci. Belg., Brussels, 36 : i—viii, 1-76, pls 1-5. 216 SOME ENGLISH Bristow, H. W. 1889. The geology of the Isle of Wight. 2nd. (revised by C. Reid & A. Strahan) ed. Mem. geol. Surv. U.K., London, xiv+349 pp., 5 pls, 84 text-figs. Burton, E. St. J. 1933. Faunal horizons of the Barton Beds in Hampshire. Pyroc. Geol. Ass., London, 44 : 131-167. Crark, B. L. & Wooprorp, A.O. 1927. The geology and paleontology of the type section of the Meganos Formation (Lower Middle Eocene) of California. Univ. Calif. Publs geol. Sct., Berkeley, 17 : 63-142, pls 14-22, I map. Cooper, J. E. 1934. Oldhaven and Thanet Sand Mollusca of Herne Bay. J. Conch. Lond., 20 : 4-8. Cossmann, M. 1889. Catalogue illustré des coquilles fossiles de 1’Eocéne des environs de Paris. Amnnls Soc. v. malacol. Belg., Brussels, 24 : 385 pp., 12 pls. — 1901. Essais de paléoconchologie comparée. 4, 293 pp., 10 pls. Paris. 1915. Essais de paléoconchologie comparée. 10, 292 pp., 12 pls. Paris. — & Pissarro, G. 1911. Iconographie compléte des coquilles de Eocene des environs de Paris. 2, 65 pls. Paris. Curry, D. 1958. Lexique stvatigyvaphique international, Europe; Angleterre, Pays de Galle et Ecosse, Paléogéne. 1 (3a), XII, 82 pp. Paris. Dat, W. H. tI909. Contributions to the Tertiary paleontology of the Pacific coast. I. The Miocene of Astoria and Coos Bay, Oregon. Prof. Pap. U.S. geol. Suvv., Washington, no. 59, 278 pp., including 22 pls in pagination. Davis, A. G. 1952. The Brockenhurst Beds at Victoria Tilery, Brockenhurst, Hampshire. Proc. geol. Ass., London, 63 : 215-219. DESHAYES, G.-P. 1824-1837. Description des coquilles fossiles des envivons de Paris. 2, 814 pp.; atlas 28 + 51 pp., LXV + 106 pls. Paris. pp. 499-780 (1835); pls. 1-105 (1837). —— 1864-1866. Description des animaux sans vertébves découverts dans le bassin de Paris. 3 : I—200 (1864); 201-658 (1865). 2, atlas 107 pp., 107 pls. (1866). Paris. DEWALQUE, G. see BosguEtT, J. 1880. DIcKERSON,R.E. t915. Fauna ofthe type Tejon: its relation to the Cowlitz phase of the Tejon Group of Washington. Proc. Calif. Acad. Sci., San Francisco, Ser. 4, 5 : 33-98, pls 1-11. Drxon, F. see SowERBy, J. deC. 1850. Farcuap, H. 1936. Etude du Thanetien (Landenien marin) du bassin de Paris. Mém. Soc. geol. Fr., Paris, N.S. 13, mém. 30 : 1-103, pls 1-6. FIsHER, O. 1862. On the Bracklesham Beds of the Isle of Wight basin. Q. Jl geol. Soc. Lond., 18 : 65-94. ForBES, E. 1856. On the Tertiary fluvio-marine formations of the Isle of Wight. Mém. geol. Surv. U.K., London, 162 pp., Io pls., 20 text-figs. FRETTER, V. & GRAHAM, A. 1962. British prosobvanch molluscs. xvi + 755 Pp., 317 figs. London. FRriTeEL, P. H. 1910. Guide géologique & paléontologique de la région Parisienne. 356 pp., 12 pls (maps), 162 text-figs. Paris. Furon, R. & Kourtatcuy, N. 1948. La faune Eocéne du Togo. Mém. Mus. natn. Hist. nat. Paris, N.S. 27 : 95-114, pls 8, 9. Gass, W. M. 1864. Description of the Cretaceous fossils in Paleontology of California, Phila- delphia, 1 : 57-236, pls 9-32. 1869. Cretaceous and Tertiary fossils in Paleontology of California, Philadelphia, 2, 299 pp., 36 pls. GaRDNER, J. S., KEEPING, H. & Monkton, H. W. 1888. The Upper Eocene, comprising the Barton and Upper Bagshot formations. Q. Jl geol. Soc. Lond., 44 : 578-635. GIEBEL, C. 1864. Die fauna dev Braunkohlenformation von Latdorf bei Bernberg. 93 pp., 4 pls. Halle. GLIBERT, M. 1957. Pélécypodes et gastropodes du Rupélien supérieur et du Chattien de la Belgique. Mém. Inst. rv. Sci. nat. Belg., Brussels, 137, 98 pp., 6 pls. 1960. Les Volutacea fossiles du Cénozoique étranger des collections de l'Institut Royal des Sciences Naturelles de Belgique. Mém. Inst. rv. Sci. nat. Belg., Brussels, ser. 2, 61, 109 pp. PALAEOGENE NASSARIIDAE 217 GLIBERT, M. 1963. Les Muricacea et Buccinacea fossiles du Cénozoique étranger des collections de l'Institut Royal des Sciences Naturelles de Belgique. Mém. Inst. v.Sci.nat. Belg., Brussels, Ser. 2, 74, 179 pp. —— & de HEINZELIN de BRAvucourT, J. 1954. L’Oligocéne inférieur belge in Volume jubilaire Victoy van Straelen, 1 : 281-438, pls 1-7. Brussels. GRATELOUP, J. P. S. DE. 1834. Tableau (suite du) des coquilles qu’on recontre dans les terrains tertiares grossiers (faluns) du bassin géologique de l’Adour (Landes). Act. Soc. linn. Bordeaux, 6 : 270-320. Gray, J. E.in Gray, M.E. 1850. Figures of molluscous animals selected from various authors. 4,iv + 219 pp. London. Horz1, O. 1958. Die Molluskenfauna des oberbayerischen Burdigals. Geologica bav., Munich, 38, 348 pp., 2 pls., 5 text-figs. IREDALE, T. 1918. Molluscan nomenclatural problems and solutions—no. 1. Proc. malac. Soc. Lond., 18 : 28-40. Jackson, J. F. 1926. A catalogue of Eocene and Oligocene fossils in the Museum of Isle of Wight Geology, the Free Library, Sandown. Pyroc. Isle Wight nat. Hist. archaeol. Soc., Newport, 1, (6 for 1925) : 340-373. Jupp, J. W. 1880. On the Oligocene strata of the Hampshire Basin. Q. Jl geol. Soc. Lond., 36 : 137-177. 1882. On the relations of the Eocene and Oligocene strata in the Hampshire Basin. Q. Jl geol. Soc. Lond., 38 : 461-489. June, P. & Assort, R. T. 1967. The genus Tevebellum (Gastropoda: Strombidae), Indo- Pacific Mollusca, 1 (7) : 445-454, pls 318-327. KEEPING, H. 1887. On the discovery of the Nummulina elegans zone at Whitecliff Bay, Isle of Wight. Geol. Mag., London, N.S., Dec. 3, 4 : 70-72. & TAwNeEy, E. B. 1881. On the beds at Headon Hill and Colwell Bay in the Isle of Wight. Q. Jl geol. Soc. Lond., 37 : 85-127, pl. 5. KOENEN, A. von. 1864. On the correlation of the Oligocene deposits of Belgium, Northern Germany, and the South of England. Q. Jl geol. Soc. Lond., 20 : 97-102. 1865. Die Fauna der unter-oligocdnen Tertiarschichten von Helmstadt bei Braunschweig. Z. dt. geol. Ges., Berlin, 17 : 459-534, pls 15, 16. 1867. Das marine Mittel-Oligocén Nord-Deutschlands und seine Mollusken-fauna. Palaeontogvaphica, Stuttgart, 16 : 53-127, pls 6, 7. 1889. Das Norddeutsche Unter-Oligocaén und seine Mollusken-fauna. Part 1. Abh. geol. SpecKarte preuss. thir. St., Berlin, 10 : 1-280, pls 1-23. Lowry, J. W., ETHERIDGE, R. E. & Epwarps, F. E. 1866. Chart of the characteristic British Tertiary fossils, stvatigvaphically avvanged. 4 pls, 800 figs. London. Martini, E. 1969. Nannoplankton aus dem Latdorf (locus typicus) und weltweite Para- llelisieurungen im oberen Eozan und unteren Oligozin. Senckenberg. leth., 50 : 117-159, pls 1-4. —— & MoorkeEns, T. 1969. The type-locality of the Sands of Grimmertingen and calcareous nannoplankton from the Lower Tongrian. Bull. Soc. belge Géol. Paléont. Hydrol. Brussels. 78 : 111-130. & Ritzkowski, S. 1968. Was ist das ‘Unter-Oligocan’? Eine Analyse der BEY RICH’- schen und V. KOENEN’schen Fassung der Stufe mit Hilfe des fossilen Nannoplanktons Nachr. Akad. Wiss. Gottingen, II, Mathphys-K1., 13 : 231-250. MELLEVILLE, M. 1843. Mémoire sur les sables tertiaives du bassin de Paris. 88 pp., Io pls. Paris. Morettet, L. & J. 1948. Le Bartonien du bassin de Paris. Mém. Serv. Carte géol. dét. Fr., Paris, 437 pp. Morris, J. 1843. A catalogue of British fossils. x + 222 pp., London. Newton, R. B. 1891. Systematic list of the Frederick E. Edwards collection of British Oligocene and Eocene Mollusca in the British Museum (Natural History). xiii + 365 pp. London. 218 SOME ENGLISH Nyst, P.-H. 1836. Recherches sur les coquilles fossiles de Hoesselt et de Kleyn-Spauwen (province du Limbourg). iii + 40 pp., 4 pls. Ghent. 1843. Description des coquilles et des polypiers fossiles des tervains tertiaives de la Belgique. atlas, 48 pls. Brussels. 1845. Description des coquilles et des polypiers fossiles des terrains tertiaires de la Belgique. Mém. cour. Sav. éty. Acad. y. Sci. Belg., Brussels, 17 : 1-697, pls 1-15. Orpicny, A. D’. 1850. Prodrome de paléontologie stvatigvaphique universelle des animaux mollusques et vayonnés. 2, 428 pp. Paris. Oyama, K., Mizuno, A. & Sakamoto, T. 1960. Illustvated handbook of Japanese Paleogene molluscs. 241 pp., 71 pls. Tokyo. Preyrot, M. A. 1928. Conchologie néogénique de l’Aquitaine. Act. Soc. linn. Bordeaux, 79 suppl. : 5-263, pls 5-14. PuiLippl, R. A. 1846-1847. Verzeichniss der in der Gegend von Magdeburg aufgefundenen Tertidrversteinerungen. Palaeontogyaphica, Stuttgart, 1 : 42-44 (1846); 45-90, pls 7-10, 10a (1847). Pow_E LL, A. W. B. 1929. The recent and Tertiary species of the genus Buccinulum in New Zealand, with a review of related genera and families. Tvans. N.Z. Inst., Auckland, 60 : 57-101, pls 1-4. REEVE, L. 1853-1854. Monograph of the genus Nassa. Conchologica Iconica, 8, 29 pls. ReIp, C. 1902. The geology of the country around Southampton. Mem. geol. Surv. U.K., London, 70 pp., 21 figs. —— & STRAHAN, A. see Bristow, H. W. 1880. SANDBERGER, C. L. F. 1863. Die Conchylien des Mainzer Tertidrbeckens. 458 pp., 1 chart, 35 pls. Wiesbaden. SOLANDER, D.C.in BRANDER,G. 1766. Fossilia Hantoniensia collecta, et in Musaeo Britannica deposita. vi + 43 pp., 9 pls. London. SOWERBY, J. DEC. 1823-1825. The Mineral Conchology of Great Britain. 5 : 1-64, pls 408— 443 (1823); 65-138, pls 444-485 (1824); 139-171, pls 486-503 (1825). London. 1850. Notes and descriptions of new species in Dixon, F. 1850, The geology and fossils of the Tertiary and Cretaceous formations of Sussex : 163-194, pls 2-9, London. STEwarT, R. B. 1927. Gabb’s California fossil type gastropods. Pyvoc. Acad. nat. Sct. Philad., 78 (for 1926) : 287-447, pls 20-32. STINTON, F. C. 1971. Easter field meeting in the Isle of Wight, Thursday 26th March to Tuesday 31st March 1970. Proc. geol. Ass., London, 82 : 403-410. Tawney, E. B. 1883. On the outcrop of the Brockenhurst Bed near Lyndhurst. Geol. Mag., London, Dec. 2, 10 : 157-160. TRavuB, F. 1938. Geologische und palaontologische Bearbeitung der Kreide und des Tertiars im Ostlichen Rupertiwinkel, nordlich von Salzburg. Palaeontogvaphica, Stuttgart, A, 88 : 1-114, pls 1-8. Tryon, G. W. 1881-1883. Manual of Conchology, 3, 310 pp., 87 pls (1881); 4, 276 pp., 58 pls (1882); 5, 276 pp., 63 pls (1883). Philadelphia. VINCENT, G. 1886. Liste des coquilles du Tongrien inférieur du Limbourg belge. nat. Hist. (Geol.) Mus. e* =) Ep anaeges = Bull. Br. PAG NIE, 2 Fics 1-8 Pseudocominella solanderi (Cossmann) p. 198 Upper Eocene, early Priabonian, Lower Barton Beds, Barton-on-Sea, Hampshire. Fics 1a—d. Neotype, GG 19693/1 (Ho). Horizon A3. x4. Fig. 1d. Oblique view to show colour-bands separated by the spiral rib which marks the weak shoulder of the last whorl. Fic. 2. GG 19694/1 (Le), same sample as figs 3, 6, 8, Horizon A3, relatively tumid specimen. Be Fic. 3. GG 19694/2, same sample as figs 2, 6, 8 showing unusually persistent axial ribbing and colour-banding. 5. Fic. 4. GG 19698/1 (Nw). Horizon A2, juvenile shell showing early sculpture, and colour- banding on shoulder. x6. Fic. 5. GG 19695/2 (S). Horizon A3, same sample as fig. 7, juvenile. 6. Fic. 6. GG 19694/3, same sample as figs 2, 3, 8, adult shell with persistent strong axial ornament, and colour-banding. x 3. Fic. 7. GG 19695/1, same sample as fig. 5, mouth of gerontic shell. 4. Fic. 8. GG 19694/4, same sample as figs 2, 3, 6, protoconch, the erect appearance of the first whorl is due to weathering. x 24 (stereoscan). Fics ga, b Pseudocominella bullata (Philippi) pp. 189, 193, 198 Lower Oligocene, Lattorfian, Latdorf, Germany, GG 19736/1 (Ed). Fic. ga. Showing lack of axial sculpture on early whorls. x 2. Fic. 9b. Showing protoconch and development of spiral sculpture. 12. nat. Hist. Mus. Bull. Br. PLATE 3 Fics 1-6 Pseudocominella armata (J. de C. Sowerby) p- 193 Middle Eocene, Auversian, Upper Bracklesham Beds. Fics ta, b. Neotype. GG 19681/1 (Ed.), Bracklesham Bay, Sussex. 2. Fics 2-4. GG 19685/1-3 (Ed), Bramshaw, Hampshire. Fic. 2a. GG 19685/3. x2. Fic. 2b. GG 19685/3. Showing fasciole. x5. Fic. 3. GG 19685/2. Oblique view showing development of strong sculpture and pre- terminal apertures on last whorl. x 3. Fic. 4. GG 19685/1. Early whorls. 12 (stereoscan). Fic. 5. GG 19683/1 (Ed), Brook, Hampshire, oblique view showing relatively weak sculpture. x3. Fic. 6. GG 19684 (Ed), Huntingbridge, Hampshire. x 3:5. Fics 7, 8 Pseudocominella armata (J. de C. Sowerby) Paris Basin, referred to by Deshayes 1865 : 499 as Buccinum fusiforme Deshayes, (Deshayes Colin., EMP). Fic. 7. Specimen from either Le Fayel, Mary or Caumont, Auversian, oblique view showing sculpture. 3. Fic. 8. Specimen from either Vendrest, Auvers or Acy, all Auversian (see p. 194 for explana- (HVA), YK Sf. Fics 9-10 Pseudocominella semicostata sp. nov. p- 196 Upper Eocene, Late Priabonian, Middle Headon Beds, Royden, Hampshire (Ed). Both specimens referred to as Stvepsiduva armata (J. de C. Sowerby) by all previous authors. Fic. 9. GG 19686/5. Juvenile shell. x5. Fics toa, b. GG 19686/3. Larger shell showing gradual development of peripheral spines. er ie) ine) a o rh 2S) ws ee) ~ S “a ~~ SS By. Bull. PLATE 4 FIGs 1-3 Pseudocominella semicostata sp. nov. p. 196 Upper Eocene, late Priabonian, Middle Headon Beds. All specimens referred to as Stvepsiduva avmata (J. de C. Sowerby) by all previous authors. Fics 1-2. Whitley Ridge railway cutting, Brockenhurst, Hampshire (Ed ex Kk). Fics ta, b. Holotype, GG 19687/1. x2. Fic. 2. Paratype, GG 19687/2. 2°5. Fic. 3. GG 19688/14, Brockenhurst, Hampshire, broken paratype showing columella. » 3. FIGs 4-7 Desorinassa desori (Deshayes) p. 202 Palaeocene, Thanetian. Fics 4-6. Jonchery-sur-Vesle, France. Fics 4a, b. GG 19438/1 (L. Staadt Colin). x3. Fig. 4b. Slightly oblique side view to show labial profile, posterior flaring of outer lip and subsutural platform. Fic. 5. GG 19438/2 (L. Staadt Colin), oblique view of younger shell. x 4. Fic. 6. GG19705/1 (Wr). Oblique view of even younger shell with weaker spiral sculpture and showing naticoid protoconch comparable with that of Pseudocominella. 4. Fics 7a, b. GG 19704, Thanet Sand, Reculver, Kent. (Wy), possibly the specimen referred to by Cooper (1934 : 8) (see remarks, p. 203). x2. Fics 8, 9 Desorinassa lata (Deshayes) p. 201 Palaeocene, Thanetian, Abbecourt and Bracheux, France. (Deshayes Colin, EMP) with no indication as to which specimen comes from which locality. Fic. 8. Lectotype, selected herein. Fic. 9. Paralectotype selected herein. Both 1-67. Bull. Br. Mus. nat. Hist. (Geol.) PLATE 5 FIGs 1-3 Desorinassa williamsi sp. nov. p- 203 Palaeocene, Ypresian, Bognor Rock, foreshore at Bognor, Sussex. Fics 1a, b. Holotype. GG19706 (Williams Colin). 2. Fic. 2. Paratype. GG 19707 (Rev. A. Fuller Colin). Oblique view showing details of fine sculpture on last whorl. x4. Fic. 3. Paratype. GG 19708/1, probably from Bognor, history unknown, internal mould. X1°6. Fic. 4 Desorinassa ovata (Deshayes) p. 201 Lower Eocene, Cuisian, Cuise, France. (Deshayes Colin, EMP) probably referred to by Deshayes, 1866 : 497. 4. Fic. 5 Desorinassa acies (Watelet) p. 201 Lower Eocene, Cuisian, Sapicourt, France. IScRNB sample IG 105901, fig. spec. no. 5600 I.S.T., referred to by Glibert (1963 : 66). 3. Fics 6-11 Whitecliffia suturosa (Nyst) p. 204 Lower and Middle Oligocene, Belgium. Fics 6-8. (de Looz Colln., IRScNB) sample IG 5002, Lower Oligocene (Lattorfian = Lower Tongrian) Grimmertingen specimens referred to by Glibert & de Heinzelin (1954 : 382). Fic. 6. Fig. spec. no. 5602 I.S.T., tumid specimen. 2:7. Fics 7a, b. Fig. spec. no. 5603 I.S.T., typical specimen. Fig. 7a, 2-7; Fig. 7b, oblique view to show protoconch and spiral sculpture on ramp. X3°5. Fics 8a, b. (IRScNB). Fig. spec. no. 4070 I.S.T., Grimmertingen, specimen referred to as Babylonia (Peridipsaccus) avchambaulti (Meunier) by Glibert & de Heinzelin (1954 : 368, 382, jl, Hy ste Oe KB Fics 9-11. Middle Oligocene, Lower Rupelian. Fic. 9. GGu19721. (de Koninck Colin), Klein Spauwen. x3. Fies 10, 11. (Bosquet Colln., IRScNB) sample IG 4285, Berg. Specimens referred to by Glibert & de Heinzelin (1954 : 392, 393). Fic. 10. Fig. spec. no. 5604 1.8.T. 2:5. Fic. 11. Fig. spec. no. 5605 1.8.T. 2:5. Bull. By. Mus. nat. Hist. (Geol.) 23, 3 PLATE 5 PAA EG: Fics 1-4 Whitecliffia suturosa (Nyst) p. 204 Fics 1a, b. GG 19717 (Ed). Upper Eocene, late Priabonian, Middle Headon Beds, Royden, Hampshire, specimen referred to as Comnella ventvicosa Edwards MS by Judd (1880 : 154) and by Keeping & Tawney (1881 : 116), and possibly as Buccinum suturosum (Nyst), by von Koenen (1867 : 83). 3. Fics 2-4. Lower Oligocene, Lattorfian, Latdorf, Germany (Ed). Fics 2a, b. GG 19720/1. Specimen with relatively strong spiral grooving. 3. Fic. 3. GG 19720/2. Rear view of specimen with very weak spiral sculpture. x 3. Fic. 4. GG 19719/2. Oblique view showing protoconch, form of suture and sculptural details. 4. Fics 5-10 Whitecliffia tumida sp. nov. p. 206 Upper Eocene, late Priabonian, Middle Headon Beds (Venus Bed), Whitecliff Bay, Isle of Wight. Fics 5a, b. Holotype. GG 19722/1 (Cu). x4. Fics 6-10. Paratypes. (All (Cz) except fig. 9). Fics 6a, b. GG 19722/3. Specimen with inflated body whorl, Fig. 6a x3; Fig. 6b, showing collabral colour banding. 2:7. Fic. 7. GG19722/6. Showing broad colour band on ramp, with narrow one belowit. 2-7. Fics 8a, b. GG19724. Small individual lacking strong axial sculpture on early whorls, Fig. 8a, 5:5; Fig. 8b, x5. Fic. 9. (Sedg. Mus., K&T) C.29305. 3. Fic. 10. GG19722/2. Juvenile shell showing paucispiral protoconch. 11 (stereoscan). ine) 73}, (Geol.) Hist nat. Mus. Bull. Br. PALIN, 7 Fics 1-5 Colwellia flexuosa (Edwards) p. 210 Upper Eocene, Late Priabonian, Middle Headon Beds, Isle of Wight. Fics ta, b. Lectotype. GG 19709/1 (Ed). Colwell Bay. x4. Fics 2a-e. Paralectotype. GG 19709/3 (Ed). Colwell Bay, exceptionally large specimen; 2a, <3; 2b, oblique view of fasciole, x 4; 2c, side view of outer lip, showing changes in growth lines with development of siphonal notch and peripheral sinus (compared with fig. 4), x3; 2d, 2e, ultra-violet photographs showing colour banding. 2. Fics 3a-c. Paralectotype. GG 19709/4 (Ed). Colwell Bay, 3a, x5; 3b, 24 (stereoscan) ; 3c, X27 (stereoscan). Fic. 4. GG 19711 (Nu), Venus Bed, Colwell Bay, side view showing abrupt transition from relatively orthocline to very flexuous growth lines (compare with fig. 2c). 3. Fics 5a, 5b. GG 19713 (Cu), Venus Bed, Whitecliff Bay, juvenile specimen. 6. FIGs 6-7 Colwellia auversiensis (Deshayes) p. 208 Middle Eocene, Auversian, Auvers, Paris Basin. (Deshayes Colln, EMP), specimens probably belonging to the type series of Buccinum auversiense Deshayes (1865 : 498; 1866, pl. 94, figs 1-3), both x 2:5. Fics 8, 9 Colwellia bretzi (Weaver) p. 208 Upper Eocene, Lower Cowlitz Formation, Vader, Washington State, U.S.A. Fig. 8. GG 12783/1. (Wr) x5. Fic. 9. GG 12783/2 (Wr) 16 (stereoscan). Fic. 10 ?Colwellia sp. p. 209 (Bruce Martin Colln, CAS., Loc. 244). Upper Eocene, Tejon Group, Tejon Quadrangle, Kern County, California, U.S.A., east bank of Live Oak Creek about three quarters of a mile from its mouth or from the edge of the San Joaquin Valley, and about three miles due east of the mouth of Grapevine Canyon. Specimen referred to and figured as Molopophorus striatus (Gabb) by Dickerson (1915 : 65, pl. 8, fig. 6 given as fig. 7 in explanation thereof). x6. (See also pl. 8.) Fic. 11 Colwellia antiquata (Gabb) p. 208 (U.Cal. 15939, Loc.A-819), specimen described and figured as Molopophorus antiquatus (Gabb) by Vokes (1939 : 142, pl. 19, fig. 2), oblique view showing protoconch. x6. Bull. By. Mus. nat. Hist. (Geol.) 23, 3 IRE AN ID, 78) FIGs I, 2 Molopophorus striatus (Gabb) p- 209 Type species of Molopophorus, Upper Eocene, Tejon Formation, Tejon, California. Fic. ta, b. Lectotype. (ANSP 4249) selected Stewart (1926 : 389, pl. 29, fig. 14), Ia x9; Ib x6 Fic. 2. (U.Cal. 7182) Showing sculpture on spire. 9. FIGs 3, 4 Brachysphingus gibbosus Nelson p. 209 Palaeocene, Martinez Formation, Ventura County, California (ex U.Cal.). Fic. 3. G72721. Oblique view showing growth lines reversed at suture. 1-7. Fic. 4. GG 12834/1. Section showing absence of columellar plait. x2. Fics 5-7 Keepingia gossardi (Nyst) p- 212 Middle Oligocene, Rupelian. Fias 5a, b. GG 19739/1. Etampes, France. 2:5. Fic. 6. (EMP) unlocalised young shell. 16 (stereoscan). Fic. 7. GG19748/1. ?Klein Spauwen (type locality), Belgium. x3. Fic. 8 Keepingia bolli (Beyrich) [Ds Bur GG 19749/1 (Wr), Upper Oligocene, Chattian, Almatil near Cassel, Germany. x 2. FIG. 9 Keepingia cassidaria (Sandberger) p. 212 GG 19733/1 (Wr), Upper Oligocene, Chattian, Cyrenmergel, Sulzheim, Germany. 4. Bull. Br. Mus. nat. Hist. (Geol.) 23, 3 PLATE 9 FIGs 1-7 Thanetinassa bicorona (Melleville) p. 213 Palaeocene, Thanetian. Fics 1-4. Chalons-sur-Vesle, France. (L. Staadt Colin.) iGai GGO357)/Th ase Fic. 2. GG 19357/3. 4. Fic. 3. GG 19357/2. Oblique view showing ‘stromboid-like’ notch in outer lip. 4. Fic. 4. GG 13357/4. Section showing form of columella. x 3. FIGs 5-7 Corbula Bed, Upper Thanet Sands, Bishopstone Gap, Herne Bay, Kent. Fig. 5. GGu19741/1 (Wr). x4. Fic. 6. GG19744 (Ru). x3. Fics 7a, b. G 60746 (J. E. Cooper Colln.). 3. Fics 8a, b Buccinum montense Briart & Cornet Pp. 215 G 71604 (Chavan Colin.). Palaeocene, Montian, Belgium. Fie. 8a. XQ. Fic. 8b. Oblique view of fasciole. 15. Fic. 9 Nassarius arcularius (Linné) p. 181 Type species of Nassarius, Recent, Indo-Pacific. BMZD.19716 (Hugh Cuming Colln) I. of Ticao,,Philippines, on reef, figd. Reeve (1853), pl. 4, fig. 25a). 1-5. Fic. 10 Strepsidura turgida (Solander) p. 190 Type species of Stvepsiduva, GG 19750 (St/B), Upper Eocene, Bartonian, Barton Beds, Barton-on-Sea, Hampshire. x2. Fics t1a, b Cominella testudinea (Lamarck) p. 180 Type species of Cominella, BMZD.1844.7.29.36 New Zealand. iG elas G2 Fic. 11b. Oblique view of fasciole. 5. Bull. Br. Mus. Hist. (Geol.) 23, 3 A LIST OF SUPPLEMENTS TO THE GEOLOGICAL SERIES OF THE BULLETIN: OF THE BRITISH MUSEUM (NATURAL HISTORY) . Cox, L. R. Jurassic Bivalvia and Gastropoda from Tanganyika and Kenya. Pp. 213; 30 Plates; 2 Text-figures. 1965. £6. . Er-Nacear, Z. R. Stratigraphy and Planktonic Foraminifera of the Upper Cretaceous—Lower Tertiary Succession in the Esna-Idfu Region, Nile Valley, Egypt, U.A.R. Pp. 291; 23 Plates; 18 Text-figures. 1966. £10. . Davey, R. J., DowniE, C., SARGEANT, W. A. S. & WitiiaMs, G. L. Studies on Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 248; 28 Plates; 64 Text- figures. 1966. £7. . APPENDIX. Davey, R. J., Downtz, C., SARGEANT, W. A. S. & WILLIAMS, G. L. Appendix to Studies on Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 24. 1969. 8op. . Ettiott, G. F. Permian to Palaeocene Calcareous Algae (Dasycladaceae) of the Middle East. Pp. 111; 24 Plates; 17 Text-figures. 1968. {5.124. . Ruopes, F. H. T., Austin, R. L. & Druce, E. C. British Avonian (Carboni- ferous) Conodont faunas, and their value in local and continental correlation. Pp. 315; 31 Plates; 92 Text-figures. 1969. £11. . Cuitps, A. Upper Jurassic Rhynchonellid Brachiopods from Northwestern Europe. Pp. 119; 12 Plates; 40 Text-figures. 1969. £4.75. . Goopy, P. C. The relationships of certain Upper Cretaceous Teleosts with special reference to the Myctophorids. Pp. 255; 102 Text-figures. 1969. £6.50. . OWEN, H. G. Middle Albian Stratigraphy in the Paris Basin. Pp. 164; 3 Plates; 52 Text-figures. 1971. {6. . Sippigur, Q. A. Early Tertiary Ostracoda of the family Trachyleberididae from West Pakistan. Pp. 98; 42 Plates; 7 Text-figures. 1971. £8. anlle igh Printed in England by Staples Printers Limited at their Kettering, Northants establishment ar K | , | aay Hise DEVONIAN “g (NATURAL HISTORY) | Vol. 23 No. PALAEOZOIC CORAL FAUNAS FROM VENEZUELA, II. DEVONIAN AND CARBONIFEROUS CORALS FROM THE SIERRA DE PERIJA BY COLIN I SCRUTION- Department of Geology The University Newcastle upon Tyne Pp. 221-281; 10 Plates; 12 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 23° No. 4 KONDON =: 1973 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, 1s issued in five series corresponding to the Departments of the Museum, and an Histoncal series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar ‘year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 23, No. 4 of the Geological (Palaeontological) series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. World List abbreviation Bull. Br. Mus. nat. Hist. (Geol.) © Trustees of the British Museum (Natural History), 1973 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 31 May, 1973 Price £4°35 PALDAEOZOIC CORAL FAUNAS FROM DEVONIAN AND VENEZUELA, IL CARBONIFEROUS CORALS FROM THE SIERRA DE PERIJA By C, Lf. SCRUTLON CONTENTS INTRODUCTION ACKNOWLEDGEMENTS STRATIGRAPHY OF THE Momma SiR | DE earn : AGES AND RELATIONSHIPS OF THE CORAL FAUNAS (a) Rio Cachiri Group (i) Cano Los Guineos Boemetion (ii) Cano Grande Formation (iii) Cano del Oeste Formation (b) Rio Palmar Formation . SYSTEMATIC DESCRIPTIONS RUGOSA . Family LINDSTROEMIIDAE Potts. Genus Steveolasma Simpson S. sp. Genus Grapeophyihon NEG 2S. sp. : Genus Syvingaxon Tendotroen : S. sp. Family HADROPHYLLIDAE iicholean Genus Hadvophyllum Edwards & Haime H. sp. Family HAPSIPHYLLIDAE Grabau Genus Amplexizaphrentis Vaughan A. sutherland sp. nov. Family STREPTELASMATIDAE Nicholson Genus Bowenelasma nov. B. typa sp. nov. . 2 B. breviseptata sp. nov. Genus Briantelasma Oliver B. oliveri sp. nov. 5 Genus Heterophrentis Billings . H. (H.) venezuelensis (Weisbord) H. (H.) simplex (Hall) Family CYATHOPHYLLIDAE Dana Genus Heliophyllum Hall : H. haili Edwards & Haime H. wellsi sp. nov. Genus Cylindrophyllum Simpson C. elongatum Simpson . Page 224 226 226 229 230 230 231 232 232 232 233 233 233 233 234 235 236 236 237 237 237 239 239 240 242 242 243 245 247 248 251 252 254 255 256 257 261 262 263 224 PALAEOZOIC CORAL FAUNAS Family DURHAMINIDAE Minato & Kato F : : 264 Genus Durhamina Wilson & Langenheim : : 2 265 ?D. sp. é : : ; ; : é : 265 D. sp. nov. : : . : ° : : 266 Family CYSTIPHYLLIDAE Edwards & Haime 5 : 269 Genus Plasmophyllum Dybowski : : c 269 P. secundum americanum (Edwards & Haime) : 269 IZ, 80. : : : d : : é 271 TABULATA ‘ : ° : : : c : 3 PP Family FAVOSITIDAE . ; : : s : 5 272 Genus Favosites . ‘ : : 5 d a ‘ 272 F.. venezuelensis (Weisbord) . : c : c 273 F. arbuscula Hall : : j : c 5 274 VI. REFERENCES . ; : é : : : : 5 275 VII. AppENDIX. COMPOSITION AND LOCATION OF SAMPLES . : : 280 SYNOPSIS Rugose and tabulate corals of Middle Devonian, and Lower Pennsylvanian age are described from localities in the northern part of the Sierra de Perija, western Venezuela. The Palaeozoic stratigraphy of the area is briefly reviewed and the ages and relationships of the coral faunas are discussed. The new taxa Bowenelasma typa gen. et sp. nov., B. breviseptata sp. nov., Amplexi- zaphrentis sutherland sp. nov., Briantelasma oliveri sp. nov. and Heliophyllum wellsi sp. nov., as well as species of Steveolasma, ?Stewartophyllum, Syringaxon, Heterophrentis (H.), Cvlindro- phyllum, Durhanuna, Plasmophyllum and Favosites are described. I. INTRODUCTION TuIs is the second of two papers describing rugose and tabulate corals from the Palaeozoic rocks of the Venezuelan Andes. The first paper (Scrutton 1971) dealt with Silurian and Permo-Carboniferous corals from the Mérida Andes. This paper describes material of Devonian and Carboniferous ages from localities in the northern Sierra de Perija (see Fig. 1), some 300 km to the north-west and close to the Venezuelan-Colombian border. The bulk of the material was collected during 1959 by Dr J. M. Bowen (then Compania Shell de Venezuela) but the opportunity is taken here to revise in addition the small samples of corals collected by C. W. Yeakel, P. W. McFarland and R. A. Liddle in 1924 (described by Weisbord 1926) and by R. A. Liddle in 1942 (described by Wells 1943) which came from the same or nearby sections. The richest fauna is one of upper Onesquethaw age (considered equivalent to the early Middle Devonian) in terms of the eastern North American stages (see Oliver 1968, fig I), represented by 14 of Bowen’s samples and including most if not all of the older collections. A further three of Bowen’s samples are also Devonian in age. The Devonian corals, which show very strong affinities with coral faunas of the same age in eastern North America, include sixteen species belonging to eleven genera; one genus and four species are new. Three of Bowen’s samples contain corals which, with some supporting evidence, indicate probable Carboniferous ages. Three species, two of them new, belonging to two genera are involved. The material is housed in the Department of Palaeontology, British Museum (Natural History). FROM THE SIERRA DE PERIJA, VENEZUELA 225 Yee rs Ke i? CARBONIFEROUS DEVONIAN + + Granite E Fossiliferous Somple Loc PRE DEVONIAN Bese ia ny yay EZ YEX. Were, “ SO Kits es TG ~~ ey Fic. 1. Geological map of the northern Sierra de Perij4 showing sample locations (for further information see Appendix). The area covered in detail in Fig. 2 is outlined. Based on maps supplied by J. M. Bowen. 226 PALAEOZOIC CORAL FAUNAS Il ACKNOWLEDGEMENTS The author is most grateful to Dr J. M. Bowen (Shell International Petroleum Corporation), who collected most of the material, for discussions on the stratigraphy of the northern Sierra de Perija and for permission to use information not yet pub- lished. Dr Bowen kindly made available the maps upon which Text-figs 1 and 2 are based. Acknowledgement is due to Compafiia Shell de Venezuela for permission to describe Bowen’s collection which they have presented to the British Museum (Natural History), and to Dr K. V. W. Palmer (Paleontological Research Institution) who has facilitated the loan of material in her care. Dr W. J. Oliver Jr. (U.S. Geological Survey) has kindly read the systematic section of the manuscript and has contributed valuable discussion and information concerning eastern North American coral faunas of Devonian age. Dr J. G. Johnson (Oregon State University) has made valuable comments on the Devonian brachiopod faunas associated with the corals. R. F. Wise and P. J. Green (British Museum (Natural History)) have respectively prepared and photographed material for this paper. E. Lawson (Department of Geology, University of Newcastle upon Tyne) drafted the originals of Text-figs 1-3, 5, 6, 8, 9 and II. III. PALAEOZOIC STRATIGRAPHY OF THE NORTHERN SIERRA DE PERIJA The basic framework of the Palaeozoic stratigraphy of the northern Sierra de Perija was first established by R. A. Liddle and his co-workers during expeditions along the course of the Rio Cachiri in 1924 and 1942. The results were published in some detail in Liddle, Harris & Wells (1943). Subsequently a number of papers have been published referring to the area (including Sutton 1946, Hea & Whitman 1960 and Miller 1962) and knowledge of the Devonian sections was summarised by Weisbord (1968). These contributions, however, did not substantially modify the stratigraphical results of Liddle’s expeditions. The most significant advance in the understanding of the Palaeozoic successions is due to J. M. Bowen’s work in the northern Sierra de Perija during 1959 when the material described in this paper was collected. Both the geological map (Figs I, 2) and the stratigraphic successions (Fig. 3) reproduced here are based on the results of his work (Bowen 1972). From the first descriptions of the rich macrofauna yielded by the oldest fossiliferous rocks of the area, published by Weisbord (1926), the strong resemblance to faunas of the upper Lower Devonian to lower Middle Devonian Onesquethaw interval in eastern North America was recognized. Liddle, in Liddle, Harris & Wells (1943) later distinguished three formations of Devonian age in the Rio Cachiri section, the Cafio Grande Formation, the Cafio del Oeste Formation and the Campo Chico Formation in order of decreasing age, grouped together asthe Rio Cachiri Series (Cachiri Group of Sutton 1946 : 1634). The rich invertebrate fauna was confined to the Cano Grande Formation and Liddle described both the other formations as un- fossiliferous (Liddle, Harris & Wells 1943 : 286, 289), although as Weisbord (1968 : FROM THE SIERRA DE PERIJA, VENEZUELA 227 221) pointed out, Liddle records in his list of samples (Liddle, Harris & Wells 1943 : 313, sample 37) Heliophyllum halli and ‘Heterophrentis venezuelense’ apparently in situ on the Cafio del Oeste Formation outcrop in the Cafio del Oeste section. In view of Liddle’s statement in the text and the fact that Bowen’s more extensive collections from the Cano del Oeste Formation do not contain these corals, it seems likely that sample 37 was in fact a displaced raft from the Cafio Grande Formation such as are known to occur on this section. An even more anomalous situation, however, surrounds the original collection of fossils described by Weisbord (1926) and collected in 1924 by C. W. Yeakel in the company of P. W. McFarland and Liddle. According to Liddle (Liddle, Harris & Wells 1943, map) and Weisbord (1968 : 217) the collection was taken in the Cafio del Norte about 4-5 km north of the Rio Cachiri from the upper part of the Cafio Grande Formation. That the fauna is of Cafio Grande aspect can hardly be doubted, but Bowen (1972; Fig. 4) records no outcrop of this formation in, or even close to the Cafio del Norte. In fact the difference between Liddle’s and Bowen’s maps of this particular section is most striking. Weisbord (1926 : 223) does record that ‘the majority of our fossils were collected from float’ and this may be the explanation. Alternatively, it is possible, in this sort of country, that the site of the collection was wrongly located (fide Bowen). Unfortunately no definite solution ed =—- ~~ ~ ~ F t Lt | t l et l t Campo Chico Formation Cano del Oeste Formation Cano Grande Formation Pre- Devonian metamorphic rocks on ; 0 100 200 300 400 ©3149 Fossiliferous Sample Locality 01 ___1___i_ m Fic. 2. Geological map of the upper reaches of the Rio Cachiri, Sierra de Perijé showing sample locations (for further information see Appendix). The area covered here is indicated in Fig. 1. Based on maps supplied by J. M. Bowen. 228 PALAEOZOIC CORAL FAUNAS to this problem seems possible at the moment and in this paper Yeakel’s collection is considered as most likely to have originated from a displaced raft or possibly faulted slice of Caio Grande Formation in the Cafio del Norte. Bowen (1972) has recognized a fourth division of the Devonian Rio Cachiri Group in the Sierra de Perija. This, the Caio Los Guineos Formation of the Rio Lajas area, he considers to be lower in the sequence than the Cafio Grande Formation although the two formations do not crop out in the same sections. They have a very similar macrofauna and both formations are dated as Lower Devonian (Emsian, equivalent to the middle Onesquethaw) mainly on the basis of brachiopods although an early Middle Devonian age (upper Onesquethaw) is considered more likely here (see p. 230-31). Bowen has also collected a fossil suite dominated by brachiopods from the Cafio del Oeste Formation, previously not satisfactorily dated, from which he deduces a lower Middle Devonian age. Finally, the highest, Campo Chico For- mation has yielded palynomorphs indicating a Middle to Upper Devonian (probably Frasnian) age (Bowen 1972). The Devonian corals, discussed in detail in the following sections, come mainly from the Caio Grande Formation, for which they suggest an upper Onesquethaw age. Liddle (Liddle, Harris & Wells 1943 : 290) assigned all the Permo-Carboniferous ‘red-bed’ sequence and overlying crinoidal limestones in the Rio Cachiri section to the Palmarito Formation of Christ (1927). Bowen (1972) has considerably modi- fied this succession, recognizing in it the lithological units described by Arnold (1966) from the Permo-Carboniferous of the Mérida Andes, about 300 km to the south-east. The basal development of red-beds Bowen assigns to the SabanetaGroup. Equivalent lower and upper divisions are recognized by both Arnold and Bowen which the RIO CACHIRI RIO PALMAR RIO PALMAR FM. CARBONIFEROUS woo CANO INDI C. DEL NOROESTE FM. | CANO DEL NOROESTE FM. Nec CAMPO CHICO FM. CAMPO CHICO FM. DEVONIAN MIDDLE eccal U PRE-DEVONIAN PERIJA FM GRANITE OF RIO LAJAS Fic. 3. Devonian and Carboniferous successions in the northern Sierra de Perija (after Bowen 1972). For comments see text. FROM THE SIERRA DE PERIJA, VENEZUELA 229 latter erects as new formations. The Cano del Noroeste Formation below (equiva- lent to Arnold’s lower clastic member) contains a fauna of fusulinids, brachiopods and palynomorphs indicating a lower Middle Carboniferous Namurian to Lower Westphalian age (Lower Pennsylvanian) according to Bowen (1972) whilst the Cafio Indio Formation above is unfossiliferous. The age of the Cano Indio Formation is fixed as low in the Pennsylvanian, however, by the dating of the overlying Rio Palmar Formation. This contrasts with the as yet poorly dated Sabaneta Group (Sabaneta Formation of Arnold 1966) of the Mérida Andes, the type area, where fossiliferous marine horizons are not known. Here, Arnold (1966 : 2371-2372) records a possible Permian age for the higher parts of the Sabaneta Group on the evidence of spores. The Rio Palmar Formation is a thick carbonate sequence conformably succeeding the Sabaneta Group. It is dated late Middle Carboniferous (late Lr. Pennsylvanian) on the basis of fusulinids and palynomorphs (Bowen 1972) and is the only level in the Permo-Carboniferous from which corals are described in this paper. This new formation is not represented in the Mérida Andes (Bowen 1972) where Arnold (1966 : 2373, 2377) records the Sabaneta Group grading upwards through a transitional zone into the Palmarito Formation. In the northern Sierra de Perija, the typical facies of the Palmarito Formation, which is absent from the Rio Cachiri area but crops out further south, comes in unconformably above the Rio Palmar Formation. According to Bowen (1972), the whole of the Upper Pennsylvanian is apparently missing here, with the Palmarito Formation yielding Wolfcampian fusulinids. In the Mérida Andes, however, the rich marine fauna of the Palmarito Formation suggests that the late Pennsylvanian as well as the early Permian may be represented (Arnold 1966 : 2378; Scrutton 1971 : 189). Thus it would appear that the distinctive carbonate facies of the Rio Palmar Formation is the lateral equivalent of rocks of upper Sabaneta or, less likely, lower Palmarito type to the south-east. Between the Palmarito Formation and the Lower Cretaceous Rio Negro Formation, Bowen (1972) has described a succession of pyroclastics, lavas and terrestrial sediments of ?Permian to Jurassic age. As no material from these younger rocks is described in this paper, they are not considered further here. IV. AGES AND RELATIONSHIPS OF THE CORAL FAUNAS The extent of previous work on South American Palaeozoic coral faunas was briefly reviewed by Scrutton (1971 : 190). Two papers refer to Devonian corals from the northern Sierra de Perija, those of Weisbord (1926) and Wells (1943). Weisbord described the material collected by C. W. Yeakel in 1924 and Wells revised the corals on the basis of further collections made by R. A. Liddle in 1942. In both collections, the corals were all from the Cafio Grande Formation of the Rio Cachiri area. J. M. Bowen’s much more extensive collections from a wider area in the northern Sierra de Perija have more than doubled the species list for the Cafio Grande For- mation as well as yielding small coral faunas from three further horizons. The compositions of the faunas are listed in Table 1 and are discussed below in ascending stratigraphical order. 230 PALAEOZOIC CORAL FAUNAS TABLE I Composition of coral faunas from the Rio Cachiri Group (Devonian) and Rio Palmar Formation (Carboniferous) of the Sierra de Perija. Rio Cachiri Group Cafio Los CafioGrande Cajio del Palmar Species Guineos Fm. Fm. Oeste Fm. Stereolasma sp. Syvingaxon sp. Heliophyllum hall H. wellsi Heterophrentis (H.) simplex x H. (H.) venezuelensis Plasmophyllum secundum americanum P. sp. Acinophyllum vermetum Briantelasma oliveri Bowenelasma typa B. breviseptata Favosites venezuelensis F. arbuscula Cylindrophyllum elongatum — ? — ?Stewartophyllum sp. x Hadrophylium sp. x Amplexizaphrentis sutherlandi x ?Durhamina sp. x Durhemina sp. nov. ? x Xx X DS ON ON OM OS PM OS ESD AIDS OS OOM (a) Rio Cachiri Group (i) Cato Los Guineos Formation. A small fauna of four different corals only is available from this formation. Its composition suggests affinities with Middle Devonian coral faunas of eastern North America where Heliophyllum halli Edwards & Haime in particular is characteristic of the Onondaga Limestone and Hamilton Group of New York and their equivalents elsewhere. Steveolasma appears to have a similar, though less well documented range and Heterophrentis (H.) simplex (Hall) has been recorded so far only from horizons of Hamilton age in New York and the Ohio Valley (see Oliver 1968, fig. 1 for a correlation chart of the eastern North America Devonian). Bowen (1972) lists other fossils, principally brachiopods, from the Cano Los Guineos Formation which he takes to indicate an upper Lower Devonian (Emsian) age. There is some doubt, however, as to the reliability of this brachiopod fauna in distinguishing between a late Lower and an early Middle Devonian age (J. G. Johnson, pers. comm.). This matter is discussed more fully in the following section on the Cafio Grande Formation which has yielded similar but richer coral-brachiopod faunas. In view of the ambiguity of the brachiopod evidence, the affinities of the coral fauna suggest that an upper Onesquethaw age (early Middle Devonian) is more likely than a late Lower Devonian age for this formation. FROM THE SIERRA DE PERIJA, VENEZUELA 231 (ii) Cano Grande Formation. The principal Onondaga—Hamilton species present in this much richer coral fauna are Heliophyllum halli, Heterophrentis (H.) simplex, Plasmophyllum secundum americanum (Edwards & Haime) and Favosites arbuscula Hall. In addition, Acinophyllum vermetum (Weisbord)! is considered to agree in all its essential characteristics with A. stramineum (Billings) from the lower part of the Onondaga Limestone of Ontario (Oliver, in preparation) and Cylindro- phyllum elongatum Simpson, which comes either from the Cafio Grande Formation or the overlying Cafio del Oeste Formation, is also found at the same horizon in New York. This evidence would clearly suggest an upper Onesquethaw (lower Middle Devonian) age for the Cafio Grande coral fauna. Also present, however, is a new species of Briantelasma. This genus has previously been recorded from rocks of Silurian (? Wenlock) and Lower Devonian (Helderberg) age in eastern North America (Oliver 1960a, b, 1963) and is the only element in the coral fauna not previously recorded from the Middle Devonian. The new genus Bowenelasma provides no direct evidence but its general characteristics relate it more to Briantelasma and Silurian streptelasmatids rather than to typical Middle Devonian members of the family. It should be stressed, however, that there is no sign of typical eastern North American middle Onesquethaw corals in the Venezuelan fauna. Bowen (1972) lists a brachiopod fauna from the Cano Grande Formation, very similar to that in the Cafio Los Guineos Formation, which he takes to indicate a late Lower Devonian age. He infers that this is supported by palynomorphs. As previously noted (p. 230), however, Dr J. G. Johnson (pers. comm.) doubts that confidence can be placed in the brachiopod fauna at the moment to distinguish between a middle and an upper Onesquethaw age (that is, late Lower Devonian and early Middle Devonian as currently understood in North America). In particular, Johnson notes that Eodevonaria has been collected from rocks of upper Onesquethaw age, and he regards the upper limits of the range of Leptocoelia to be uncertain. Both genera have been taken to indicate a Lower Devonian horizon. He also draws attention to the occurrence of Tropidoleptus (Morales 1965) and common Pentagonia and Spinulicosta, all more suggestive of an upper Onesquethaw than a Lower Devonian age, in the Floresta fauna of Colombia, with which the Venezuelan faunas have much in common. Johnson thinks it likely that these South American occurrences represent an overlap of late Lower and early Middle Devonian forms. Dr L. Nijssen (pers. comm.), who made the palynological determinations, also has reservations about distinguishing between an Emsian and an early Eifelian age on the basis of the flora. Although much revision of the North American coral faunas is necessary, Oliver (1968 : 741) clearly indicates that an important break in the sequence of coral assemblages there occurs between the middle and upper Onesquethaw. In this light, the strong affinities of the coral fauna with faunas of upper Onesquethaw age in eastern North America suggests that the Cafio Grande Formation should be dated upper Onesquethaw (early Middle Devonian) rather than middle Onesquethaw at the present time. 1 As Dr W. J. Oliver Jr. has a review of this species in preparation it is not described in the systematic section of this paper. 232 PALAEOZOIC CORAL FAUNAS The affinities between the Venezuelan and eastern North American coral faunas indicates that the two areas must have formed part of the same province in early Middle Devonian time. The presence of Bowenelasma and some species endemic to Venezuela suggests that north-south migration may have been partially restricted in some way, although these corals may be recorded from eastern North America in the future. (ii) Cano del Oeste Formation. This formation appears to be poor in corals although it has provided a fairly varied fauna of brachiopods and molluscs which are listed by Bowen (1972). One specimen each of ?Stewartophyllum sp. and Hadrophyllum sp. indicate a tentative Middle Devonian age, whilst the brachiopods have been more precisely dated as early Middle Devonian. Although the specimen of Cylindrophyllum elongatum, which comes from a slipped block, is listed by Bowen (1972) with the Cafio del Oeste fauna, Bowen has in- formed the writer that the block compared lithologically with the Cano Grande Formation outcropping further up the Cafio Pescado and may well have come from that formation. If the block did originate from the Cafio del Oeste Formation, however, then it would be from near the base, whereas the rest of the fauna, pro- viding evidence for the age, derives from the very top of the formation. (b) Rio Palmar Formation This formation has also yielded two corals among a rather limited invertebrate fauna. Amplexizaphrentis sutherlandi sp. nov. and ?>Durhamina sp. indicate a general Carboniferous age and compare most closely with North American cordilleran species. The other faunal elements provide more precise information and indicate a late Lower Pennsylvanian age. Durhamina sp. nov. is known only from a loose block whose stratigraphic relation- ships are obscure. Of the known stratigraphy, it is most likely to have come from the Rio Palmar Formation which is where Bowen (1972) lists it as “durhaminido gen. and sp. nov.’. Possible Mississippian brachiopods were reported from the same block, however, and it may represent an horizon as yet not recognized in situ in the stratigraphic sequence. V. SYSTEMATIC DESCRIPTIONS The terminology used in the following descriptions is that proposed by Smith (1945 : 4-9) and Moore, Hill & Wells (1956) unless otherwise indicated. All the material collected by Bowen is now housed in the Department of Palaeon- tology, British Museum (Natural History). The registered numbers for these specimens are prefixed R and in each case the number of the sample in which the specimen was collected is given in parentheses after the registered number. The locations of Bowen’s samples are plotted on Text-figs 1 and 2 and full details are given in the Appendix. FROM THE SIERRA DE PERIJA, VENEZUELA 233 The sources of other material described below are indicated by the following abbreviations: PRI—Paleontological Research Institution, Ithaca, N.Y.; NYSM— New York State Museum, Albany, N.Y. Order RUGOSA Edwards & Haime 1850 Suborder STREPTELASMATINA Wedekind 1927 Superfamily CYATHAXONIICAE Edwards & Haime 1850 Family LINDSTROEMIIDAE Poéta 1902 1971 Lindstroemiidae; Scrutton: 192, cum syn. Discussion. The family concept was reviewed by Scrutton (1971 : 192). Genus STEREOLASMA Simpson 1900 1g00 Steveolasma Simpson: 205. 1941 Steveolasma Busch: 395. 1949 Steveoelasma Stumm: 7. 1962 Steveolasma Stumm: 234. 1965 Steveolasma Stumm: 14. DiacGnosis. Small ceratoid to trochoid corals. Major septa fuse in the axis to form, with or without additional sclerenchyme, a prominent axial pillar. Minor septa short, usually contratingent. Peripheral stereozone narrow. Cardinal fossula narrow, poorly developed. Tabulae complete or incomplete, strongly arched with a flat or slightly depressed axial area. No dissepiments. TYPE SPECIES (by original designation). Stveptelasma rectum Hall 1876, pl. 10, figs I-13 in part ?=Strombodes? rectus Hall 1843 : 210, text-fig. 87, 5 on p. 209, and no. 48, fig. 5 on p. 44 of tables. Middle Devonian, Hamilton Group; western New York State, U.S.A. Discussion. Reference should be made to Stumm (1949: 7) and Stumm & Watkins (1961 : 445) for comments on and a redescription of Steveolasma rectum. Stereolasma sp. (Pl. 1, figs 6, 7) MATERIAL. R45129 (3849). Cano Los Guineos; Caio Los Guineos Fm., early Middle Devonian (upper Onesquethaw). DESCRIPTION. Small ceratoid coral, incomplete. In the late neanic stage the mean diameter is II mm with 25 major septa. The lumen is completely infilled by laterally contiguous wedge-shaped major septa merging in the axial area with a core of sclerenchyme 3 mm in diameter. This core may be formed by irregularly intertwined septal ends but the structure is not clear. There are signs of rudimentary minor septa as wedges about I mm long between the major septa. 234 PALAEOZOIC CORAL FAUNAS In the ephebic stage (Pl. 1, fig. 6) there are-25 major septa at a mean diameter of 14:5 mm. The major septa, except the cardinal septum, are 0-6—0-75 mm thick where they are very slightly waisted one-third of the way along their length from the periphery. Thecardinalseptumis notably thinnerato-4mm. Theseptaare thickest about two-thirds of their length to the axis where they become laterally contiguous. From this point they taper smoothly into the axis in a slight counter clockwise vortex maintaining contact with their neighbours. The septa are longest in the alar area where they all but meet in the axis. The counter and cardinal septa are shorter, however, each reaching only three-quarters of the radius, leaving a narrow axial gap about 4:5 mm long and 0-5 mm wide in the otherwise solid oval core of fused septal ends, 7 mm x 6 mm in size. The peripheral ends of the septa alternate with contratingent minor septa in a peripheral stereozone I-1:2 mm thick. The minor septa generally appear as short thorny projections from the stereozone mainly in the counter quadrants where they are more free standing. The minor septa flanking the counter septum are 2 mm long and they decrease in size round towards the cardinal septum near which they are 1-5 mm long. The minor septa contra- tingent on the major septa immediately flanking the cardinal septum, however, are again 2 mm long. Septal microstructure consists of an irregular structureless core embedded in calcite showing growth lamellae running at a very low angle to the septal faces axially and inward towards the core. In the longitudinal section (Pl. 1, fig. 7), cut between the two sections described above, the lumen is infilled with septal material except for two small peripheral areas. The septal material shows bowl-shaped laminations, notched in the axis, which are interpreted as growth-lines. In the smaller of the two voids, part of a tabula can be seen curving upward and axially from the peripheral stereozone. There are no dissepiments. The specimen contains circular borings, 0-8-1: mm in diameter, of unknown origin. Discussion. This specimen appears to compare most closely in size and structure with the type species of Stereolasma, S. rectum (Hall) from the Middle Devonian Hamilton Group of eastern North America as illustrated by Simpson (1900, figs 16-18) and Stumm & Watkins (1961, pl. 58, figs 1-16). S. rectum, however, has much thinner skeletal elements and less well developed minor septa except for the counter-lateral minor septa which are strongly accelerated. The Venezuelan specimen may belong to a new species but further material is required to confirm this. Genus STEWARTOPHYLLUM Busch 1941 1941 Stewartophyllum Busch: 393. 1949 Stewartophyllum Stumm: 9. ?1965 Stewartophyllum Federowski: 344. TYPE SPECIES (by original designation). Amplexus intermittens Hall 1876; pl. 32, figs 8-15. Middle Devonian, Hamilton Group; Moscow, N.Y., U.S.A. FROM THE SIERRA DE PERIJA, VENEZUELA 235 DiaGnosis. Small subcylindrical to trochoid corals. Major septa dilated and extending to axis where they fuse in neanic stage, withdrawn from axis in ephebic stage. Cardinal fossula present. Minor septa very short or absent. Tabulae complete and incomplete, flat axially, variably orientated peripherally. No dis- sepiments (based on Stumm 1949 : 9 and Stumm & Watkins 1961 : 447). Discussion. The type specimens of S. intermittens (Hall) have been redescribed by Stumm and Watkins (1961 : 447) who showed that three other species described by Busch should also be referred to Hall’s species. The diagnosis for the genus given above takes account of this redescription. ?Stewartophyllum sp. (Pl. 1, figs 4, 5) MATERIAL. R45110 (3243). Cafio Grande; Cafio del Oeste Fm., early Middle Devonian. DESCRIPTION. Small curved trochoid coral, about 25 mm high and 20 mm in maximum diameter. In the late neanic-early ephebic stage, the lumen is almost entirely infilled. The septa occur in a single series up to 1-25 mm thick at the periphery and wedge-shaped, reaching between three and four-fifths of the way to the axis. They are arranged in an irregularly pinnate pattern and are mostly laterally contiguous with scattered very narrow gaps in the peripheral area between some septa. The axial ends of the septa merge into solid sclerenchyme completely infilling the axis, in which no struc- tural details can be seen. The septal structure compares with that described as pseudotrabecular by Kato (1963, text-fig. ge). In higher sections a subtriangular cardinal fossula containing a short, thin cardinal septum, 4 mm long, is present on the convex side of the coral. There are slightly larger gaps between the septa and signs of tabulae can occasionally be seen. The axis of the coral is filled by a subcircular core of sclerenchyme 7 mm in diameter. As the subcalicular level is approached, the subtriangular cardinal fossula increases in size at the expense of the cardinal half of the core. The pinnate septal symmetry is also more regularly developed immediately below the calice. The diameter here is 16 mm with about 40 septa. No longitudinal section is available. Discussion. This single specimen is tentatively assigned to Stewartophyllum. It appears to have considerable structural similarity with S. intermittens (Hall) from the Middle Devonian Hamilton Group of eastern North America (see Stumm & Watkins 1961, pl. 58, figs 17-28, 37-39) although it is clearly not conspecific. Apart from the much higher septal number in the Venezuelan specimen, the large axial core of sclerenchyme makes even a congeneric assignment open to question. More material is required to assess the variability and significance of the axial core. 236 PALAEOZOIC CORAL FAUNAS Genus SYRINGAXON Lindstrém 1882 1970 Syvingaxon Sutherland: 1125. 1971 Syvingaxon Scrutton: 194, cum syn. Discussion. See Scrutton (1971 : 194) for a diagnosis and full discussion of the status of this genus. Syringaxon sp. (Pl. 1, figs 1-3) MATERIAL. R45130 (3849). Cano Los Guineos; Cafio Los Guineos Fm., early Middle Devonian (upper Onesquethaw). DESCRIPTION. Small conico-cylindrical coral with moderately developed septal grooves. Coral circular in cross-section with a variably corrugated epitheca and a peripheral stereozone about 0-6 mm thick. In the mature stages the major septa are 0-13- 0-2 mm thick midway along their length where they are very slightly waisted; the cardinal septum is the thinnest. They reach approximately seven-eighths of the radius towards the axis. The axial ends of the major septa are rhopaloid causing them to be laterally contiguous and to form, with a small amount of additional sclerenchyme, an aulos 2-2-5 mm in external diameter. In a subcalicular section (Pl. 1, fig. 1), the internal dimensions of the aulos are 1:2 x 0-g mm. Ina section approximately 1-5 mm below this (Pl. 1, fig. 2), the aulos is solidly infilled, possibly by sclerenchyme coating a tabula as the longitudinal section shows the aulos to be open for at least another 2-5 mm below the level of this section. The minor septa are approximately half the radius in length except for the counter-lateral minor septa which are equal in length to the counter septum. They are closely contratingent so that the lumen between a major septum and its contratingent minor may be closed at sub-calicular levels. The longitudinal section is two-thirds infilled by septal material due to its being cut slightly off-centre. There is a very narrow open core in the distal part of the aulos containing irregularly spaced flat tabulae averaging 0-25 mm apart. In the interseptal loculi one tabula can be clearly seen, flat near the peripheral stereozone, and curving upwards towards the aulos. There are faint indications of what may be further tabulae but none can be positively identified as such. The subcalicular section is about 6-8 mm mean diameter with 20 major septa and the section I-5 mm below is 6-4 mm in mean diameter with 18 major septa. Discussion. This single specimen, although slightly larger in size at the same septal number, is quite similar to S. ayvnoldi Scrutton (1971 : 196) from rocks of Ludlovian age in the Mérida Andes of Venezuela. Further material is necessary, however, to determine the status of the Devonian specimen. S. rudis (Girty 1897 : 299, pl. 2, figs 7, 8), from the Lower Devonian of North America, is in need of redescription and cannot be compared with the present material. FROM THE SIERRA DE PERIJA, VENEZUELA 237 Family HADROPHYLLIDAE Nicholson 1889 1937 Palaeocyclidae Bassler: 189, pars. 1949 Hadrophyllidae Stumm: 4. 1955 Porpitidae Jeffords: 12, pars. 1956 MHadrophyllidae Hill: 262. 1961 Hadrophyllidae Fontaine: 69. 1969 Hadrophyllidae Sutherland & Haugh: 27. Tyre GENUS. Hadvophyllum Edwards & Haime 1850 : 67. DiaGNnosis. Small, simple, discoid, patellate, button shaped or depressed turbinate corals. Major septa usually arranged in quadrants and cardinal and alar fossulae often strongly developed. Minor septa short and contratingent. Hori- zontal partitions (?tabulae) developed in some genera. Discussion. A brief resumé of the classification of this group of corals is given by Sutherland & Haugh (1969 : 28). For the moment, Hill’s (1956 : 262) inter- pretation of the Hadrophyllidae is accepted and attention is drawn to the presence of what are probably tabulae in the specimen described below as Hadrophyllum sp. and in Gymnophyllum wardi Howell (see Sutherland & Haugh 1069 : 35). Genus HADROPHYLLUM Edwards & Haime 1850 1850 Hadrophyllum Edwards & Haime: 67. 1851 Hadrophyllum Edwards & Haime: 357. 1937 Hadvophylium Bassler: 197. 1952 Hadrophyllum Le Maitre: 41. 1955 Hadrophyllum Jeffords: 8. 1961 Hadvophyllum Fontaine: 69, cum syn. Dracnosis. Discoid to broadly trochoid. Cardinal septum in oval fossula; other cardinal quadrant septa pinnate and sub-parallel, counter quadrant septa more or less radially arranged with a long counter septum. Minor septa short. Horizontal partitions (?tabulae) may be developed. TYPE SPECIES (by original designation). Hadrophyllum orbignyt Edwards & Haime 1851 : 357, pl. 6, figs 4, 4a. Devonian; Charleston Landing, Indiana and Clark County above Louisville, Kentucky, U.S.A. Discussion. The status of this genus has been discussed by Jeffords (1955 : 8). An amended diagnosis is given here to cover the occurrence of tabula-like structures in the specimen of Hadrophyllum described below. Hadrophyllum sp. (Pl. 1, fig. 8, Text-fig. 4) MATERIAL. R49275 (3243). Cafio Grande; Cafio del Oeste Fm., early Middle Devonian. B 238 PALAEOZOIC CORAL FAUNAS DESCRIPTION. Small, broadly trochoid coral with smooth, slightly worn exterior. Calice not seen. A single section at the ephebic stage shows the lumen largely infilled by thick septa arranged with striking pinnate symmetry. The septa are mostly between I and 1:4 mm in thickness peripherally. The counter and alar septa are longest, all three meeting in the axis. In the counter quadrants, the counter-lateral and next four metasepta are roughly equal in length and about three-quarters the length of the counter septum. The fifth and subsequent metasepta in the counter quadrants become progressively shorter and more wedge-shaped against the alar septa. On the cardinal side of the alar septa the first two metasepta are equal in length to the alar septa and the subsequent metasepta become progressively shorter and more wedge-shaped round towards the cardinal septum. The wall of the cardinal fossula appears to be the result of sclerenchyme coating the successive ends of the metasepta in the cardinal quadrants and is not formed by extra long metasepta flanking the short, 3:5 mm long, cardinal septum. Very narrow gaps, about o-I mm wide, separate the major septa in the peripheral area of the corallum. In several places the gaps are crossed by narrow bars that are probably tabulae. There are short triangular minor septa, about I mm long, inserted between nearly all the major septa. The septal formula is CoA8K7A8C at a mean diameter of 13 mm. No longitudinal section is available. Discussion. The general characters of this coral agree with those of the genus Hadrvophyllum although it cannot be readily assigned to any described species. Comparison is made difficult as virtually no other species of Hadrophyllum has been Fic 4. Septal arrangement in Hadvophyllum sp. (R49275). x5. C = cardinal septum; K = counter septum; A = alar septa. FROM THE SIERRA DE PERIJA, VENEZUELA 239 studied with, or illustrated by, thin sections. The appearance of interseptal structures like tabulae in cross-section has never previously been recorded in a species of Hadvophyllum but this may also be due to a lack of thin section study. For example, rare tabulae have recently been recorded in thin sections of another member of the Hadrophyllidae, Gymnophyllum wardi Howell, by Sutherland and Haugh (1969 : 35). Family HAPSIPHYLLIDAE Grabau 1928 1956 Hapsiphyllidae Hill: 267. 1964 Hapsiphyllidae Rowett & Sutherland: 46. Genus AMPLEXIZAPHRENTIS Vaughan 1906 1906 Ampblexizaphrentis Vaughan: 315. 1958 Amplexizaphrentis Sutherland: 44. 1962 Amplexizaphrentis Armstrong: 30. 1962 Enniskillenia Kabakovitch im Soshkina & Kabakovitch: 323. 1964 Amplexizaphrentis Rowett & Sutherland: 47. DiaGnosis. Small, solitary, trochoid to ceratoid corals in which major septa may or may not unite around a conspicuous cardinal fossula. Septa commonly withdraw from axis in late stages of growth. Alar pseudofossulae usually well developed in early stages, but may become inconspicuous in ephebic stage. Tabulae usually complete, arched or flattened axially except where abruptly depressed in the fossula. No dissepiments. (After Rowett & Sutherland 1964 : 47). TYPE SPECIES (see Opinion 854, I.C.Z.N. 1968). Zaphrentis curvulena Thomson 1881 : 223, 236. Lower Limestone Group, Lower Carboniferous; Brockley, near Lesmahagow, Lanarkshire, Scotland. Discussion. The status of the genus Amplexizaphrentis has been discussed by Sutherland (1958 : 44 et seq.) and his conclusions are accepted here. More recently a case to stabilize the generic name and type species was submitted to the I.C.Z.N. by Shrestha (1966), supported by Mitchell (1966) and approved by the Commission in Opinion 854. This effectively answered the comments on the validity of Amplexizaphrentis expressed by de Groot (1963 : 39). According to Rowett & Sutherland (1964 : 51) Barytichisma (Moore & Jeffords 1945 : 131) is distinguished from Amplexizaphrentis ‘primarily on the basis of the unusually wide peripheral stereozone’. They also note in the diagnosis for that genus that the septa are amplexoid and reach the axis immediately above the tabulae only. Weyer (1965 : 450) also laid emphasis on amplexoid septa in Bary- tichisma in maintaining the genus distinct from Amplexizaphrentis. If this distinc- tion is accepted, the species described below seems to fall more naturally into Amplexizaphrentis than Barytichisma despite a tendency to amplexoid septa at all stages of growth. Certainly A. sutherlandi sp. nov. does not develop an unusually wide peripheral stereozone. 240 PALAEOZOIC CORAL FAUNAS Amplexizaphrentis sutherlandi sp. nov. (Pl. 1, figs 9-13, Text-fig. 5) DERIVATION OF NAME. After Dr P. K. Sutherland (University of Oklahoma). DiaGnosis. Curved ceratoid Amplexizaphrentis with cardinal fossula on concave side of corallum. Major septa thick, tending to amplexoid throughout ontogeny; 34 present in ephebic stage of holotypeat 14mm diameter. Minor septa rudimentary. Tabulae wide, flat with downturned edges, inclined strongly from the counter to cardinal side. HorotyPe. R45127 (3784). Cafio Colorado; Rio Palmar Formation, Lr. Pennsylvanian. PARATYPES. R4g291 (3784; ?distal end of holotype), R49292 (3784). Same locality and horizon as holotype. DESCRIPTION. Curved ceratoid corals up to 60-70 mm long and about 25 mm maximum diameter (estimated because of crushing of the calice). The cardinal fossula is on the concave side of the corallum. Septal grooves weak or absent on a slightly rugate epitheca. In the early neanic stage (Pl. 1, fig. 10) the septa are short and variably developed with a wide, open lumen. On the cardinal side a tabula is cut, on the upper surface of which the cardinal septum and several metasepta in the left cardinal quadrant are well developed. They are approximately 1-3 mm long and 0-5 mm wide at their bases, the cardinal septum being fractionally longer than the others. No septa can be distinguished in the zone where the tabula is in the plane of section and immediately below the tabula, in the alar areas, the septa are very short and thorn- like. In the counter area the counter septum is short and thin, about 1:3 mm long and 0-I mm wide expanding to 0-5 mm wide where it merges into the thin peripheral stereozone. The flanking septa are similar in shape but become progressively shorter, from a maximumof 0-8 mm long, towards the alar areas. The mean diameter is 7 mm; Ig septa can be counted but there are probably 22 or 24 present at this level. In a later neanic section (Pl. 1, fig. 11), approximately 2 mm higher than the first, the septa in the cardinal quadrants are much more strongly developed. The alar septa appear to meet across the axis of the corallum in a continuous band o-5—0-8 mm thick and the metasepta in the cardinal quadrants are arranged in strongly pinnate groups. The ends of successive metasepta outline a large suboval cardinal fossula bisected by a long waisted cardinal septum extending to the axis. The metasepta in the cardinal quadrants are o-4—0-6 mm thick peripherally and taper towards the axis. Where tabulae are sectioned, they and the adjacent septa are thickened by about 0-2 mm towards the periphery. In the counter quadrants, the septa are thin short spines as in the previous section. Measurements are difficult as the peripheral stereozone has been removed around much of these quadrants. The mean diameter is 7-5 mm and the septal formula is C6A4K4A6C. In an ephebic stage section (Pl. 1, fig. 9) 14 mm higher (measured centre to centre) than the previous section, the counter quadrants are fully developed. The counter FROM THE SIERRA DE PERIJA, VENEZUELA 241 septum 1 mm thick, is thicker than all other septa, which range between 0-5—0-7 mm thick, and reaches into the axial area to meet other septal material. The flanking septa are long and pinnately arranged with their axial ends fused and coated with sclerenchyme to form strong walls parallel to the alar septa and suggesting very weak alar fossulae. The arrangement of the septa in the cardinal quadrants is similar to that described in the previous section except that the alar septa are now of similar aspect to the cardinal metasepta and meet in the axis with an angle of about 120°. The cardinal fossula is more pear shaped and the cardinal septum has shortened to 3-4 mm long (its axial end is damaged). Most of the major septa taper towards the axis but some are slightly rhopaloid and may have very thin needle-like axial extensions from the club-shaped thickening. The peripheral stereozone is I mm thick where it is undamaged and slight kinks in it on either side of the counter septum suggest the presence of rudimentary minor septa. The mean diameter is 14 mm and the septal formula is C6AgKgAO6C. A section in the ephebic stage of another specimen (R49292) shows all the septa withdrawn from the axis except the metasepta flanking the cardinal septum. These with a considerable thickening of sclerenchyme, outline an oval cardinal fossula bisected by a thin cardinal septum. The diameter is 15 mm with an estimated 38 septa (C18K?18C). In longitudinal section the tabulae are largely complete with few subsidiary plates and a rather irregular spacing. They slope strongly downwards from the counter to the cardinal side with a flat or slightly undulating surface and down- turned peripheral margins. Successive complete tabulae may be as much as 3-4 mm 20 d (mm) Fic. 5. Number of major septa plotted against diameter for specimens of Amplexi- zaphrentis sutherlandi. The dashed line joins asterisks representing different sections of the holotype. 242 PALAEOZOIC CORAL FAUNAS apart during the late neanic-early epebic stages but become much closer-spaced, less than I mm at the subcalicular level. The tabulae are extensively crested with septal material in the peripheral areas. This material shows fibres directed axially and upwards at varying angles and growth lines steeply dipping at the periphery, curving round to the horizontal in the axial area. The tabulae in the axial area are coated with sclerenchyme showing a zigzag fibre pattern. Although in some places this appears to be an alteration phenomenon based on the calcite cleavage pattern, in others it appears to be original and could reflect lateral movements in the basal ectoderm of the polyp during calcification (see also Sutherland 1958 : 49). There are no dissepiments. Discussion. This species seems to compare quite closely with Amplexizaphrentis cassa Sutherland (1958 : 54, pl. 8, figs 1-5, pl. 10, figs 1-4) from the middle Mississippian Kindle Formation of the Tetsa River area, and particularly closely with A. sp. B (Sutherland 1958 : 57, pl. 11, figs I-5) from the Prophet Formation (?early Mississippian) of the Prophet-Muskwa Rivers area, both British Columbia. A. sutherland is distinguished by the thickness of its septa, its large size and a slightly higher septal ratio than A. cassa. In these features, however, it approaches very closely to A. sp. B of Sutherland although the cardinal fossula in the latter may be larger in proportion. The age of the Rio Palmar Formation appears to be lower Pennsylvanian and thus A. sutherlandi is somewhat younger than either of the Canadian species. Superfamily ZAPHRENTICAE Edwards & Haime 1850 Family STREPTELASMATIDAE Nicholson 1889 1971 Stveptelasmatidae Scrutton: 206. Discussion. The writer has recently outlined his views on the composition of this family (Scrutton 1971 : 206). The new genus described here stands very close to existing members of this group and does not necessitate any reconsideration of the family concept. Genus BOWENELASMA nov. DERIVATION OF NAME. After Dr J. M. Bowen (Shell International Petroleum Corporation) who collected the material described here. DiaGnosis. Curved ceratoid corals with the cardinal septum on the convex side of the corallum. Septa dilated to close the lumen in early ontogeny; strongly dilated and more or less coated with sclerenchyme in the cardinal quadrants at the ephebic stage. Cardinal fossula narrow or poorly developed. Intertwined septal elements in the axial area may form a low boss in the calice. Minor septa well developed. Tabulae steeply sloping axially and upwardsat the periphery, undulating or highly domed in the axial area. No dissepiments. TYPE SPECIES. Bowenelasma typa sp. nov. FROM THE SIERRA DE PERIJA, VENEZUELA 243 Discussion. Bowenelasma is recorded so far only from the early Middle Devonian of Venezuela. The genus is distinguished first and foremost by the strong, consistent dilatation and thickening of the cardinal quadrant septa in the later, but subcalicular, ontogenic stages. This feature appears occasionally in other streptelasmatid corals and it is also characteristic of the genus Pseudophaulactis Zaprudskaya in Ivanovskii (1963 : 32, pl. 6, fig. 2) from the Llandovery of the Siberian Platform. Bowenelasma and Pseudophaulactis are very similar in appearance, the latter differing only in the very attenuate nature of the septa in the counter quadrants and the general lack of twisted septal ends in the axial area (see Ivanovskii 1965, pls 5, 6 and 7, fig. x). It should also be noted, however, that the two genera are widely separated strati- graphically and geographically. Species of other streptelasmatid genera occasionally show significantly prolonged dilatation and thickening in the cardinal quadrants during ontogeny as a specific character or a subspecific variant. They include Kiaerophyllum semilunatum Scheffen (1933 : 21, pl. 2, figs 4-6) from the Upper Ordovician of the Oslo region which is considered a variant of Grewingkia buceros according to Neuman (1969 : 36), Grewingkia europaeum hosholmensis Kaljo (1961 : 58, pl. 3) from the Upper Ordo- vician of Estonia and the specimen identified as Dalmanophyllum dalman by Ivanovskii (1963 : 36, pl. 7, fig. 4) (= Ditoecholasma dalmani in Ivanovskii 1965, pl. 1, fig. 6) from the Llandovery of the Siberian Platform. All these are considerably older than the new Devonian genus. Bowenelasma is clearly not congeneric with Grewingkia, which is principally distinguished by its wide spongy axial structure (Neuman 1969 : 35). Bowenelasma also lacks the distinctive blade-like columella so prominent in illustrations of the type species of Dalmanophyllum (Edwards & Haime’s 1851, pl. 1, fig. 6; Minato 1961, pl. 11, figs 16, 3c, 5b). Comparison with its contemporary streptelasmatids shows Bowenlasma to be readily distinguishable from all of them. This genus appears to be a late and little modified descendant of the primitive streptelasmatid structural pattern and as such contrasts quite strongly with the more structurally divergent Heterophrentis/ Siphonophrentis group. Briantelasma, however, may be more closely related (see discussion under Briantelasma). Bowenelasma typa sp. nov. (Pl. 2, figs 1-11, Text-fig. 6) DERIVATION OF NAME. From typus (L.) = type, indicating the type species. Diacnosis. Bowenelasma with 45 major septa at a mean subcalicular diameter of 27 mm (holotype). In ephebic stage, major septa four-fifths radius in length, dilated, coated with sclerenchyme and laterally contiguous in cardinal quadrants and slightly deflected counter-clockwise. Axial area with few septal ends inter- twined to form low calicular boss. Minor septa up to one-quarter radius in length. Cardinal septum short in narrow fossula. Tabulae simple, convex, steeply sloping upwards and axially at periphery, obscured in axis but apparently gently undulating. 244 PALAEOZOIC CORAL FAUNAS HoLotyPeE. R45094 (3149). Cano Grande; Cafio Grande Formation, early Middle Devonian (upper Onesquethaw). PARATYPES. R45075, R45077 (both 3121), R45100 (3157); same locality and horizon as holotype. R45121 (3618); Cafio Colorado; same horizon as holotype. OTHER MATERIAL. R45082, R49247-9, R4g9251 (all 3121), R49263 (3157); same locality and horizon as holotype. R45118 (3618), Cafio Colorado; R45279 (3323), loose boulder on Cafio del Oeste; both same horizon as holotype. DESCRIPTION. Curved ceratoid corals up to about 90 mm long and 27 mm mean diameter at the subcalicular level. The cardinal septum is located on the convex side of the corallum. The epitheca is lightly rugate with weak septal grooves. In the neanic stage (PI. 2, fig. 4), the lumen is almost completely infilled by thick, irregularly pinnate septa. The major septa are between 0-6—0-7 mm thick near the periphery and taper towards the axis, contiguous with their neighbours over most of their length. Major septa in the counter and alar sectors, 3-0-3-5 mm long, reach into the axial area where they fuse and almost completely fill the lumen. The other septa wedge against these. The cardinal septum is about 1-3 mm long. Minor septa are present as very short wedges 0-7-I-0 mm long, nearly always with very small openings in the lumen beyond their axial ends. The counter septum is flanked by extra long (1-7 mm) minor septa. There are also small openings in the lumen periaxially, beyond the axial end of the cardinal septum and on the counter side of one of the alar septa. The septal formula is C345K5A4C at a mean diameter of 7 mm. Sections in the late neanic stage of the holotype (R45094c-d, Pl. 2, fig. 2) show major septa of very variable length twisted in a wide counter-clockwise axial vortex. Most of the major septa are 0:3-0:4 mm wide and half the radius in length. Only a few irregular septal ends cross the axial area where there may also be a few isolated spots of septal tissue. The cardinal septum is short and heavily invested with sclerenchyme. Some of the other septa in the cardinal quadrants also appear to be thickened but there is no marked contrast with the septa in the counter quadrants at this level. Minor septa cannot be easily distinguished in the wider parts of the very irregularly developed peripheral stereozone. Some of them appear to reach 2-5 mm in length and the stereozone may be up to 4 mm wide. There are 38 septa at a mean diameter of 17 mm. In the ephebic stage (Pl. 2, fig. 1) there is a strong contrast between cardinal and counter quadrants. In both the major septa extend four-fifths of the distance to the axis and are twisted in a weak counter-clockwise vortex. Those in the cardinal quadrants however are between I-I-I:4 mm thick and laterally contiguous for most of their length, whilst those in the counter quadrants are only 0:25-0:35 mm across and more or less parallel sided. The cardinal septum, also contiguous with its neighbours, is 5-5 mm long and situated in a weak narrow fossula closed on the axial side by a tabula and converging septa. The axial area of the coral, about a fifth of the diameter across, is traversed by two or three septal strands and traces of tabulae. There may also be a few isolated spots of septal tissue. In R45077b there is evidence that this structure in the axial area may produce a low boss in the floor FROM THE SIERRA DE PERIJA, VENEZUELA 245 of the calice. Minor septa are also well developed, as wedges of variable size up to I-5 mm across at the base and 3:5 mm long in the cardinal quadrants and as thin spines projecting up to 2 mm beyond the peripheral stereozone in the counter quad- rants. The minor septa flanking the counter septum are markedly longer than their neighbours but measurement is difficult as this part of the coral is nearly always crushed. The peripheral stereozone in the counter quadrants is quite thin, only about 1 mm thick. In the holotype, the septal formula is C1oA13KI10A8C at a mean diameter of 27 mm. At subcalicular levels, the septa in the cardinal quadrants thin and separate from their neighbours first on the axial side of the minor septa. The cardinal septum thins most rapidly leaving a narrow, parallel sided fossula which it bisects. It maintains a length of about half the radius whilst the septa in the cardinal quadrants are still somewhat thicker than those in the counter quadrants but at slightly higher levels it rapidly withdraws to about a fifth of the radius in length. Ina paratype, R45121a (Pl. 2, fig. 10), the minor septa flanking the counter septum are about 5°5 mm long at this level and the septal formula is C1ioA1z2K12A9C at a mean diameter of 28 mm. The longitudinal section is dominated by septal traces and the shape of the tabulae is difficult to see. At both walls there are thick solid septal deposits, rather irregularly developed but much more pronounced on the cardinal side. In the axial area individual septal traces form a complex and irregular pattern breaking up and obscuring the tabulae. The tabulae can be clearly seen in the peripheral areas, slightly convex, moderately spaced and sloping axially and upward at 5~-20° to the wall. In the axial area, the tabulae appear to be gently undulating and in one case (Pl. 2, fig. 6) sloping distinctly downward from the cardinal to the counter side. The angle of slope is difficult to assess but appears to be as much as 45° in some cases. Measurements on the holotype and paratypes are plotted in Text-fig. 6. Discussion. The only species so far assigned to Bowenelasma are the two species from the Sierra de Perija described here. The type species, B. typa, which is the more abundant by a factor of three in Bowen’s collection, is compared with B. breviseptata under the discussion of the latter species. Bowenelasma breviseptata sp. nov. (Pl. 3, figs 1-7; Text-fig. 6) DERIVATION OF NAME. From brevis (L) = short + septum, in reference to the characteristically short major septa. Diacnosis. Bowenelasma with 43 major septa at a mean subcalicular diameter of 33 mm (holotype). In ephebic stage, major septa half radius in length, cardinal quadrants heavily coated with sclerenchyme peripherally. Cardinal septum very slightly longer than adjacent metasepta; no fossula developed. Wide axial area partially filled by thickened and twisted septal ends. Minor septa up to one-third radius in length. Tabulae simple, convex, steeply sloping upwards and axially at the periphery, domed in the axis or obscured by septal traces. 246 PALAEOZOIC CORAL FAUNAS HowotyPe. R45123 (3619). Cafio Colorado; Cafio Grande Formation, early Middle Devonian (upper Onesquethaw). PARATYPE. R45125 (3619). Same locality and horizon as holotype. OTHER MATERIAL. R45105 (3199), Cafio del Sur; ? R4g256 (3154), Cafio Grande; both same horizon as holotype. DeEscRIPTION. Large curved ceratoid corals with the cardinal septum located in the convex side of the corallum. Estimated original length of holotype 80-90 mm. The epitheca has weak septal grooves and is gently rugate. The late neanic stages (PI. 3, fig. 3) are largely infilled by major septa up to 1-7 mm thick and laterally contiguous along most of their length. The major septa are two-thirds the radius in length leaving a wide axial zone mostly infilled by loosely and irregularly twisted septal elements. The cardinal septum is only just over one- third the radius in length with a gap in the lumen beyond its axial end: the other protosepta cannot be distinguished. The minor septa are well developed wedges up to 5 mm long and 1 mm or a little more across the base. In the ephebic stage (Pl. 3, fig. 1) septal thickening is confined to the cardinal quadrants. The septa are about half the radius in length with a slight anticlockwise displacement. In the cardinal quadrants the peripheral half to three-quarters of the septa are thickened to about 2 mm across bringing adjacent septa into contact. 50 40 20 0 10 20 30 40 d (mm) ---- Btypa + holotype —-—-— B. breviseptata 0 holotype & R45105 @ other material + other material Fic. 6. Number of major septa plotted against diameter for specimens of Bowenelasma typa and B. breviseptata. FROM THE SIERRA DE PERIJA, VENEZUELA 247 Otherwise the septa are between 0-6 and 0-7 mm across and parallel sided in both quadrants. The cardinal septum is slightly longer than adjacent metasepta with the peripheral dilatation imperfect ; there is no cardinal fossula. In the axial half of the diameter, septal elements are loosely and irregularly twisted together and thickened in part. This structure forms a low boss in the floor of the calice. The minor septa are very long, and regularly developed reaching 5 mm or a third of the radius in length. Unfortunately those flanking the counter septum cannot be seen in the holotype due to crushing in this area. The minor septa are wedge-shaped in the cardinal quadrants and parallel sided to gently tapering, 0-5—1-0 mm across in the counter quadrants. The peripheral stereozone in the counter quadrants appears to be thin but is everywhere worn and cannot be measured. R45105 (Pl. 3, figs 5-7) differs somewhat from the other specimens through its septa being generally less thickened. The section of the neanic stage available (Pl. 3, fig. 5) shows no lateral contiguity between septa. In the ephebic stage the septa are only slightly thicker in the cardinal quadrants than the counter quadrants, although the cardinal septum itself is notably thick at 1:2 mm across. The minor septa are up to 4-4 mm or two-fifths the radius in length and slightly variable in development. The septal formula is C6AgKgA5C at a diameter of 22 mm. The longitudinal section of the holotype (Pl. 3, fig. 4) is dominated by septal traces and the peripheral stereozone. The latter is particularly thick on the cardinal side. Traces of tabulae can be seen in the peripheral areas, sloping steeply axially and upwards but they cannot be identified in the axial area among the irregular vertical sections of septal elements. In R45105d (PI. 3, fig. 7), however, the structure is very clear. The tabulae consist of large curved plates, convex upwards and outwards. They form a high axial dome and are almost vertical peripherally where they form a narrow trough against the peripheral stereozone. They are irregularly spaced between I and 2:5 mm vertically. Measurements are plotted in Text-fig. 6. Discussion. At present it is not certain how much weight should be given to the differences between R45105 and the holotype. R45105 is considered conspecific with Bowenelasma breviseptata here although with more material and a better understanding of the variation it may prove to be subspecifically or even specifically distinct. B. breviseptata is distinguished from B. typa principally through the much shorter major septa in the mature stages and the lower septal number at comparable diameters. JB. breviseptata also has more strongly developed minor septa and lacks any clear sign of a cardinal fossula. Genus BRIANTELASMA Oliver 1960 1960a Briantelasma Oliver: 89. 1960b Briantelasma Oliver: 6. 1963 Briantelasma Oliver: 26. Diacnosis. Trochoid to cylindrical corals with subpinnately arranged major septa extending half way or more to the axis. Lumen partly or completely infilled 248 PALAEOZOIC CORAL FAUNAS by laterally contiguous septa and axial sclerenchyme especially in early growth stages. Cardinal fossula present. Minor septa usually well developed. Tabulae strongly domed with axial depression, complete and closely spaced. (Based on Oliver 1960b : 6.) TYPE SPECIES (by original designation). Briantelasma americanum Oliver 10960 : 89, pl. 14, figs 1-4. Reef facies of Coeymans Limestone, Helderbergian, Lower Devonian; Madison and Oneida Counties, New York. Discussion. The writer follows Oliver (1960b : 6) in regarding Briantelasma as a member of the Streptelasmatidae in which the structural elements are completely invested in sclerenchyme during almost the full ontogeny. The ancestors of Briantelasma could well belong to the same lineage that evolved Bowenelasma as there are similarities in the basic structural patterns of the two genera. They differ principally, and very obviously, in the manner and extent of the development of sclerenchyme and also in the detailed appearance of the tabularium and the peri- pheral stereozone. There is no doubt that they represent distinct and divergent genera over their presently known range. Briantelasma oliveri sp. nov (Pl. 4, figs 1-7; Text-figs 7, 8) DERIVATION OF NAME. After Dr W. A. Oliver, Jr. (United States Geological Survey). DiacGnosis. Ceratoid to trochoid Briantelasma of variable size up to 39 mm diameter with 60 major septa in the holotype. Lumen completely infilled by septa and sclerenchyme to immediately sub-calicular levels. Major septa withdraw from axis to approximately half radius in length in ephebic stage. Axis plugged by structureless to vermiform loops of sclerenchyme which persists as low boss in calice. Minor septa up to one-quarter radius in length. Tabulae present in floor of calice; not distinguished in sclerenchyme. HoLotyPE. R45090 (3129). Cafio Grande; Cafio Grande Formation, early Middle Devonian (upper Onesquethaw). PARATYPES. R45091 (3129); same locality and horizon as holotype. R45119 (3618) ; Cafio Colorado, same horizon as holotype. OTHER MATERIAL. ? R45099 (3157); same locality and horizon as holotype. ? R4g9290 (3618) ; Cafio Colorado, same horizon as holotype. ? PRI 24426 (= 24433); ? loose block in Cafio del Oeste, same horizon as holotype. ? PRI 24429 (= 24430A); same locality and horizon as holotype. ? PRI 24430B; ? locality of holotype or PRI 24426, same horizon as holotype. DESCRIPTION. Ceratoid to trochoid, straight to slightly horn shaped corals up to 40 mm diameter at the base of the calice and an estimated I00 mm long. The holotype, which is the longest specimen, is incomplete and has a crushed calice. The cardinal septum is on the convex side in curved coralla. Epitheca moderately rugate and bearing weak septal grooves. FROM THE SIERRA DE PERIJA, VENEZUELA 249 In the neanic section R45091a (Pl. 4, fig. 5) the lumen is completely or almost completely infilled by laterally contiguous septa and axial sclerenchyme. The major septa vary between half and the full radius in length and their axial ends merge into the sclerenchyme almost completely infilling the axis. They are between 0-6 and 0-7 mm in width and contiguous with their neighbours except for occasional narrow gaps between one-quarter and one-third the radius from the periphery, probably on the axial side of wedge-shaped minor septa. These minor septa are extremely difficult to distinguish but appear to be present only in the counter quadrants. There is a slight pinnate septal symmetry but again difficult to distinguish mainly due to the character of the preservation. The septal formula is C6A8KQA6C at a mean diameter of Io mm. In the ephebic stage (Pl. 4, figs 3, 6), the laterally contiguous major septa are very evenly developed 0-5 to 0-6 of the radius in length. With the exception of a few specific major septa, they all have open axial ends merging with a solid axial core of sclerenchyme which is either more or less structureless or contains a large scale vermiform pattern (R45119a, b). The septal pattern in the holotype is illustrated in Text-fig. 7. The counter septum, which is fractionally wider and longer than the average, I'5 mm X II mm in the holotype, is characteristically flanked by counter- lateral septa slightly shorter than the average and with closed axial ends. Some septa in the alar areas, particularly new metasepta appearing immediately on the counter side of the alar septa are also less than average length and have closed axial ends. The cardinal septum, which is difficult to distinguish, appears to be about average length and is flanked by shorter metasepta with the next adjacent meta- septa curved slightly towards the cardinal septum at their axial ends. Minor septa are well developed wedges up to a maximum of one-quarter the radius in length. Fic. 7. Septal arrangement in Briantelasma oliveri (R45090a). X2. C = cardinal septum; K = counter septum; A = alar septa. 250 PALAEOZOIC CORAL FAUNAS The lumen is totally infilled. The septal formula is C1ioA13K15A10C at a mean diameter of 23 mm in R45091b Close to the base of the calice (Pl. 4, fig. 1) an opening first appears on the axial side of the cardinal septum which shortens rapidly at this level. A narrow elongate or sub-oval fossula is formed. Subsequently gaps appear between the major septa on the axial side of the minor septa and rapidly enlarge. The septa appear to thin almost simultaneously in all quadrants; contrasted thickening in various quadrants at this level is due to sections slightly angled to the growth lines. As the major septa withdraw to the periphery, with gently tapered axial ends, the axial plug of sclerenchyme is isolated briefly as a low mound in the calice base. At this level in the holotype, there is a peripheral stereozone 4 mm wide at a mean diameter of 39 mm. The stereozone thins in the walls of the calice and the minor septa project from it, thin and narrowly attenuate. In longitudinal sections (Pl. 4, fig. 4), there is evidence of some gaps in between the septa and sclerenchyme at the neanic stage and in the cardinal area at later stages. Two such gaps in the cardinal area occur in the longitudinal section of the holotype (R450g90e) and show signs of tabulae trough shaped against the peripheral stereozone and capped by flat subsidiary tabulae. The septa and sclerenchyme show a strongly developed grain shallowy arched across the axis and sharply down turned towards the peripheral trough-shaped tabulae. This may reflect the shape of 70 60 50 30 20 d (mm) - - -® -- B oliveri * holotype B. americanum === — B. knoxboroense serteseeeseee B. mainense Fic. 8. Number of major septa plotted against diameter for specimens of Briantelasma oliveri. The polygons outline the n/d variation fields for other species of Briantelasma. FROM THE SIERRA DE PERIJA, VENEZUELA 251 tabulae suppressed, or obscured in the sclerenchyme. Unfortunately no longitudinal section is available through the base of the calice, where tabulae are known to be developed, to support this. Measurements on the holotype and paratypes are plotted in Text-fig. 8. There is a considerable variation in size, basal calice diameters ranging from 15:5 mm to 39 mm. Discussion. Previous records of Briantelasma come from Silurian (? Wenlock) and Lower Devonian (Helderberg) rocks in eastern North America. Briantelasma oliveri, from the early Middle Devonian of the Sierra de Perija, thus considerably extends the known range of the genus. The new species can be distinguished from all earlier described species of Briantelasma through the major septa withdrawing from the axis in the ephebic stage. Also, over the ontogenetic range available, B. oliveri lies wholly outside the fields of variation in septal number with diameter shown by the other species (see Text-fig. 8). It has higher septal ratios at corres- ponding diameters and reaches a higher maximum diameter than the North American species. Genus HETEROPHRENTIS Billings 1875 Discussion. Oliver (1958 : 817, 825) suggested that Compressiphyllum Stumm (1949 : 13), which differs from Heterophrentis s.s. only through a pronounced com- pression of the corallum at right angles to the counter-cardinal plane, was best considered as a subgenus of Heterophrentis. Subgenus HETEROPHRENTIS (HETEROPHRENTIS) Billings 1875 1875 Heterophrentis Billings: 235. 1938 Heterophrentis Stewart: 20. 1949 Heterophrentis Stumm: 11. 1958 Heterophrentis (Heterophrentis) Oliver: 825. 21964 Heterophrentis Webby: 7. 1965 Heterophrentis Stumm: 18. 1968 Hetevophrentis Altevogt: 762. DiaGnosis. Ceratoid to trochoid corals, circular in cross-section. Septa thin throughout ontogeny; major reach more or less to axis, minor up to one-third radius in ephebic stage; peripheral stereozone variable in thickness. Cardinal fossula prominent. Tabulae complete or incomplete, flat axially, arched peripherally. No dissepiments. TYPE SPECIES (see Miller 1889 : 193). Zaphrentis spatiosa Billings 1858 : 178 = Zaphrentis prolifica Billings 1858 :176 (according to O’Connell 1914 : 183). Onondaga Limestone, early Middle Devonian; Rama’s Farm, near Port Colborne, Ontario, Canada. Discussion. Numerous species of Heterophrentis (H.) have been described and the scope of the snbgenus is reasonably well defined despite the fact that, according 252 PALAEOZOIC CORAL FAUNAS to Sutherland (1958 : 45), the original specimens of neither Z. spatiosa, nor Z. prolifica have been sectioned. The traditional interpretation of the subgenus is followed here and would surely need to be stabilised even if Z. spatiosa proved not to have the characters attributed to Heterophrentis (H.). Heterophrentis (H.) is common in the Lower and Middle Devonian of eastern North America. The subgenus has also been recorded in the Middle Devonian of Spain and north-west Africa and is confirmed here in the early Middle Devonian of Venezuela. Heterophrentis (Heterophrentis) venezuelensis (Weisbord 1926) (PI. 5, DSS 1,.2) 1926 Cyathophyllum venezuelense Weisbord: 4 pars., pl. 1, fig. 5 only. 1943 Heterophrentis venezuelensis (Weisbord) Wells: 97 pars., pl. 10, fig. 7 only. Diacnosis. Heterophrentis (H.) with major septa two-thirds radius in ephebic stage, tapering axially; about 64 at 33 mm mean diameter in lectotype. Wide axial area occupied by tabulae only. Minor septa confined to peripheral stereozone one-quarter radius in width. Cardinal fossula present but weak. Tabulae mainly complete, gently undulating to depressed axially, sharply downturned peripherally, very close spaced beneath calice floor. LeEcTOTYPE (here chosen). PRI 21594 (= PRI 24421, same specimen). Loose block (?) in Cafio del Norte; Cafio Grande Fm., early Middle Devonian (upper Onesquethaw). DEscriPTIon. The lectotype is an incomplete curved conical coral 80 mm high between a diameter of 23 mm and the top of the calice. The epitheca is lightly rugate. The counter-cardinal plane is at right angles to the plane of curvature. In the ephebic stage, 10 to 15 mm below the base of the calice, the major septa are regularly developed, 9-10 mm or two-thirds the radius in length. They are between I and 1-3 mm thick at the edge of the peripheral stereozone and taper gently towards the axis. Their arrangement is more or less radial except near the cardinal septum which is flanked on either side by three or four shorter metasepta and fronted axially by the concave side of a tabula. The cardinal septum itself cannot be easily distinguished but may be slightly longer than the flanking metasepta. The axial area of the corallum contains sections of tabulae but no septal material. The peripheral stereozone is between 3:5 and 4 mm wide which is also the length of the wedge shaped minor septa. There are at least 60 and an estimated 64 major septa at a mean diameter of about 33 mm and thus a septal ratio of about I-94. In longitudinal section, the tabulae are mainly complete and spaced about I mm or slightly less apart at the subcalicular level. Towards the proximal end, the spacing becomes wider and less regular. The tabulae are gently undulating or very slightly depressed across the axial area and are sharply downturned peripherally. FROM THE SIERRA DE PERIJA, VENEZUELA 253 They appear to be inclined downwards from the convex to the concave side of the corallum. Subsidiary tabulae occur mainly on the periphery of the wide axial plateau formed by the complete tabulae. Evidence from the cross-section suggests that the tabulae are depressed in the cardinal fossula. Many tabulae have signs of septa strongly developed on their upper surfaces and some tabulae are thickly coated with septal deposits. The closer spaced tabulae in the 6-7 mm below the calice floor, however, are not affected in this way. Discussion. The syntypes of Cyathophyllum venezuelense Weisbord and the additional specimens assigned to the species by Wells when he redescribed it as Heterophrentis venezuelensis have been re-examined with the aid of a number of new slices from which acetate peels have been prepared. No less than five species and four genera are thought to be represented. Only two of the original syntypes are species of Heterophrentis and these are not conspecific. One, PRI 21594 (= PRI 24421), has been selected as the lectotype and only known example of H. (H.) venezuelensis and is the specimen described above. The other, PRI 21593, is a speci- men of H. (H.) simplex (Hall) and is described below. The only additional material of either species among Bowen’s collection is a single specimen assigned to H. (H.) simplex. Of the other syntypes PRI 21592 (Weisbord 1926, pl. 1, fig. 1) is probably Bowenelasma typa, and PRI 21791 (Weisbord 1926, pl. I, fig. 4) is Heliophyllum hall. Of the additional material studied by Wells, PRI 24426 (two thin sections cut from PRI 24433 and figured by Wells 1943, pl. 10, figs 8, 9), PRI 24429 (cut from PRI 24430A) and PRI 24430B are all questionably assigned to Briantelasma oliveri. Wells (1943 : 98) did observe that Heterophrentis venezuelensis as then understood would probably prove, with the study of additional material, to contain more than one species. Heterophrentis (H.) venezuelensis is difficult to compare with the many species of Heterophrentis (H.) named from the Devonian of North America as very few com- prehensive illustrations and descriptions for these species exist. Furthermore, the Venezuelan species is represented only by the lectotype so that its range of variation is unknown. As far as can be determined at the moment, the major septa are comparatively short in H. (H.) venezuelensis, with a weak cardinal fossula and minor septa confined to a well developed peripheral stereozone. This appears to distin- guish the species, at least from H. (H.) prolifica (Billings) and H. (H.) spissa (Hall), the two species with which Wells (1943 : 98) chiefly compared H. (H.) venezuelensis. Stumm (1965 : 20) considered H. (H.) sfissa conspecific with H. (H.) inflata (Hall) which he suggested in turn might be a junior synonym of H. (H.) prolifica. All these ‘species’ appear to be characteristic of the early Middle Devonian in eastern North America. H. (H.) venezuelensis certainly appears to differ markedly from the holotype of H. (H.) inflata illustrated by Stumm (1965, pl. 4, figs 11, 12). As noted above the writer believes that H. (H.) simplex, a third species compared with H. (H.) venezuelensis by Wells (1943 : 98), is actually represented by a single specimen among the syntypes. This specimen and the holotype of H. (H.) simplex are figured here (Pl. 5, figs 3-5) and described below. They can be seen to contrast strongly with the lectotype of H. (H.) venezuelensis. Cc 254 PALAEOZOIC CORAL FAUNAS Heterophrentis (Heterophrentis) simplex (Hall) 1843 (Pl. 5, figs 3-5; Text-fig. 9) 1843 Strombodes simplex Hall: 210, text-fig. 87, 6 on p. 209, no. 48, fig. 6 on p. 44 of tables. 1876 Zaphrentis simplex (Hall) Hall pars, pl. 21, figs 5, 8-10 (non figs 6, 7, 11). 1926 Cyathophyllum venezuelense Weisbord: 4 pars, pl. 1, figs 2, 3 only. ?1938 Heterophrentis simplex (Hall) Stewart: 23, pl. 2, figs 5-7. 1965 Heterophrentis simplex (Hall); Stumm: 21, ?pars, pl. 11, fig. 19, pl. 14, figs 5, 6. ?1968 Heterophrentis simplex (Hall); Stumm: 38, pl. 1, figs 6, 7. DiaGnosis. Ceratoid Heterophrentis (H.). Major septa three-quarters radius in ephebic stage, thin and tapering axially with a very slight counter-clockwise deflection in axis. Cardinal septum short in suboval fossula. Minor septa one quarter radius in length, confined to or more commonly projecting from variably developed peripheral stereozone. HoLotyPe. NYSM 360. Middle Devonian, Hamilton Group; Moscow, New York. OTHER MATERIAL. PRI 21593. Loose block (?) in Cato del Norte; Cafio Grande Fm., early Middle Devonian (upper Onesquethaw). R49293 (3849). Camo Los Guineos; Cafio Los Guineos Fm., early Middle Devonian (upper Onesquethaw). DistTRIBUTION. Middle Devonian, Hamilton Group of New York, U.S.A. and equivalent horizons elsewhere in eastern North America; early Middle Devonian of Sierra de Perija, Venezuela. DESCRIPTION. The Venezuelan specimens are slightly curved incomplete ceratoid corals lacking the proximal end. 60 50 n 40 30 20 10 20 30 d (mm) Fic. 9. Number of major septa plotted against diameter for specimens of Hetevophrentis (H.) simplex. The holotype is indicated by an asterisk. The dashed lines join points representing different sections of the same specimen. FROM THE SIERRA DE PERIJA, VENEZUELA 255 In the neanic stage of PRI 21593 (PI. 5, fig. 4), the lumen is partially infilled by major septa between 0-8 and 1-0 mm thick, parallel sided and with blunt axial ends. These septa are mostly laterally contiguous along the axial third of their length where they are involved in a weak counter-clockwise vortex. A small area in the axis is free of septal material. The cardinal septum is short, 3-5 mm or half the radius in length, and flanked by short curved metasepta concave towards the cardinal septum. Towards the periphery the major septa are separated by narrow, parallel sided, ovoid or gradually widening wedge-shaped gaps terminated by the peripheral stereozone. The stereozone is rather variably developed up to 1:25 mm thick and contains traces of rudimentary wedge-shaped minor septa. The septal formula is C7A8K8A7C at a diameter of I5 mm. In the ephebic stage of PRI 21593 (PI. 5, fig. 5), the major septa are very slim and taper gently from the peripheral stereozone where they are about 0-4 mm thick to within a quarter of the radius or slightly less of the axis. There is a weak counter clockwise vortex. A few isolated bars of septal material may occur in the axis. In the cardinal quadrants, three or four metasepta flanking the cardinal septum curve to meet each other, or join sections of tabulae, across the axial end of the cardinal septum, delimiting a strong suboval fossula. The fossula is bisected by the cardinal septum, 4:5—-5 mm or two-fifths the radius in length. The peripheral stereozone is variably developed from 3:3 mm thick where the minor septa are completely con- fined to the stereozone, to 1 mm thick from which the minor septa may project up to a further 2 mm towards the axis. There are 42 major septa at a diameter of 24 mm, a septal ratio of 1-75. Measurements are plotted in Text-fig. 9. No longitudinal section is available. Discussion. The Venezuelan specimens are considered conspecific with the holo- type of H. (H.) simplex which is figured here on PI. 5, fig. 3. The polished ephebic section of Hall’s specimen is slightly larger—30 mm diameter with 53 major septa— but in other features the specimens are almost identical. H. (H.) simplex has slimmer, longer major septa and a much more distinct cardinal fossula than H. (H.) venezuelensis. Furthermore, the septal ratio in this the smaller species (I-77 at 30 mm in the holotype) is lower than that in H. (H.) venezuelensis (I-94 at 33 mm in the lectotype). Family CYATHOPHYLLIDAE Dana 1846 1963 Cyathophyllidae Birenheide: 367. 1966 Cyathophyllidae Pedder: 182, pars. 1969 Phillipsastraeidae Jell: 62, pars. Discussion. Following his valuable revision of the Cyathophyllidae, Birenheide (1963 : 368) commented on the structural similarities between this family and the Disphyllidae, and Pedder (1966 : 182) later united the two groups of corals as sub- families within the Cyathophyllidae. Jell (1969; see particularly pp. 64, 69), who on the other hand united the disphyllids with the phillipsastreids in the Phillip- sastraeidae, has also discussed this relationship. Although disputing Pedder’s 250 PALAEOZOIC CORAL FAUNAS classification, he did point out the close similarities of the Billingsastraeinae, a new subfamily of his Phillipsastraeidae, to the cyathophyllids. The writer is of the opinion that Bzllingsastraea (as widely interpreted) and Cylindrophyllum belong together with Heliophyllum in the same suprageneric group. This is essentially the Billingsastraeinae of Jell (1969 : 68) with the addition of Heliophyllum. Birenheide (1963 : 368) included Heliophyllum in his undivided Cyathophyllidae but the Heliophyllum—Billingsastraea group, although considered here to be more closely related to the cyathophyllids than the disphyllids, deserves to be distinguished from the former on the basis of septal characteristics (see Text- fig. 11). Itis therefore proposed that the Billingsastraeinae and the Cyathophyllinae be considered subfamilies of the Cyathophyllidae. The Phillipsastraeidae is interpreted in the sense of Scrutton (1968) (= Phillip- sastraeinae of Jell 1969 : 65) but admitting the Marisastrinae (sensu Jell 1969 : 66) if the presence of rhipidacanths can be confirmed in Marisastrum sedgwicki. The Disphyllidae (= Disphyllinae of Jell 1969 : 68 and ? Paradisphyllinae Jell 1969 : 67) is also accorded full family status. Perhaps the affinities of the three families Cyathophyllidae, Disphyllidae and Phillipsastraeidae would be best recognized at the superfamily level. Subfamily BILLINGSASTRAEINAE Jell 1969 Genus HELIOPHYLLUM Hall 1846 1846 Hehophyllum Hall in Dana: 356. 1938 Helophyllum Stewart: 35. 1938 Heliophyllum Fenton & Fenton: 209. 1945 Heliophyllum Smith: 25. 1947 Helhophyllum Le Maitre: 30. 1949 Helhophyllum Stumm: 21, pars. non 1949 Heliophyllum Soshkina: 88. 1962 Heliophyllum Stumm & Tyler: 267. 1963 Heliophyllum Birenheide: 404. Dracnosis. Solitary or colonial rugose corals with long slender major and minor septa bearing well developed yard-arm carinae in the dissepimentarium. Septa dilated to fill the lumen in early ontogeny; dilatation may persist in tabularium of cardinal quadrants in ephebic stage. Cardinal fossula long, narrow and weak. Dissepimentarium of several rows of small globose dissepiments, horizontal or reflexed peripherally, steeply sloping downwards axially. Tabulae complete or incomplete, usually low domes but may be flat or slightly sagging. TYPE SPECIES (by monotypy). Strombodes helianthoides? Hall (non Phillips, nec Goldfuss) 1843 : 209, fig. 87, 3 on p. 209, fig. 48, 3 on p. 44 of tables = Helio- phyllum halli Edwards & Haime 1850 : 69. York, Livingstone Co., N.Y., U.S.A.; Ludlowville Fm., Middle Devonian (Hamilton Group) (Wells 1937 : 9). FROM THE SIERRA DE PERIJA, VENEZUELA 257 Discussion. Birenheide (1963) in a revision of cyathophyllids from the German Devonian has confirmed the views of Wang (1948 : 11) and others that Keriophyllum Wedekind is not congeneric with Heliophyllum. He (1963 : 392) assigned the type species of Keriophyllum to Cyathophyllum (Peripaedium) turbinatum, type species of Peripaedium Ehrenberg. Birenheide (1963 : 405) did regard Keriophyllum dahle- mense Haller (1936 : 615, pl. 33, figs 1-2, pl. 34, fig. 1), however, as a valid species of Heliophyllum. Heliophyllum is particularly common in the Middle Devonian of eastern North America but it has also been recorded from Europe, Asia, North Africa and South America. Some of these records may be species of Cyathophyllum but the acceptance of K. dahlemense as a species of Heliophyllum and the confirmation of records of H. halli from Spain (Altevogt 1968 : 764), North Africa (Le Maitre 1947 : 31) and Afghanistan (Brice 1971 : 263) supports this distribution in part. It is doubtful, however, if true Heliophyllum occurs in Australia. Heliophyllum halli Edwards & Haime 1850 (Pl. 5, figs 6-8; Pl. 6, figs 1-6; Text-figs 10, 11) 1850 Heliophyllum halli Edwards & Haime: 69. 1926 Cyathophyllum venezuelense Weisbord: 224 (4), pars., pl. 35 (1), fig. 4 only. (PRI 21791). 1937 Heliophyllum halli Edwards & Haime; Wells: 9, pl. 1, figs 1-15. 1943 Heliophyllum halli Edwards & Haime; Wells: 363 (95), pl. 36 (10), figs 1-2. 1947 Heliophyllum halli Edwards & Haime; Le Maitre: 31, pl. 1, figs 1-7, pl. 2, figs 1-6, pl. 3, fig. I. 1965 Helophyllum halli Edwards & Haime; Stumm: 36, pl. 32, figs 5, 6, 20-23. 1968 Helophyllum hall Edwards & Haime; Altevogt: 764, pl. 2, figs 5a, b, c. 21969 ©=Hehophyllum hal Edwards & Haime; Kaplan: 28, pl. 3, fig. 1. 1970 Heliophyllum halli Edwards & Haime; Brice: 263, pl. 18, fig. 6. Diacnosis. Usually curved, ceratoid to turbinate solitary corals but capable of considerable shape variation and rare increase. Major septa long, uniformly attenu- ate in dissepimentarium but may be thickened in tabularium of cardinal quadrants. Cardinal septum slightly shorter than other metasepta in long narrow fossula, other metasepta usually arranged with bilateral symmetry about counter-cardinal plane. Minor septa one-third to one-half radius in length. Septa in dissepimentarium with well developed yard-arm carinae. Dissepiments small, globose, horizontal or inclined towards epitheca peripherally, steeply sloping towards the axis. Tabulae wide low-domed complete plates with subsidiary domed plates in axis and small, inclined subsidiary plates periaxially. MATERIAL. R45076, R45078, R45080-1, R45084-6, R49250, R49252, R49296 (3121); R45087 (3124); R45092-3, R49253 (3145); R49254 (3149); R45096-7, R49255, R49258, R49260 (3154); R49262 (3157); all Cafio Grande; Cafio Grande Formation. R45104, R49268-9 (3199); R45107—-9, R49274 (3221); all Cafio del Sur; Cafio Grande Formation. R45120, R45122, R49286, R49288-9 (3618); R45124 258 PALAEOZOIC CORAL FAUNAS (3619) ; all Cafio Colorado; Cafio Grande Formation. R45111-5, R49277-8, R49281, R49284 (3323); all from loose boulder in Cafio del Oeste; Cafio Grande Formation. R45128, R49294-5 (3849); all Cafio Los Guineos; Cafio Los Guineos Formation. All early Middle Devonian (upper Onesquethaw). DISTRIBUTION. Widespread in Middle Devonian of eastern North America and also present in the Middle Devonian of North Africa, Spain, Afghanistan and Venezuela. Records of this species from other areas doubtful. DESCRIPTION. Curved conical trochoid to turbinate corals attaining a maximum mean diameter of 38 mm and a height of about 50 mm. Corallum shape not greatly variable. Epitheca with very weak septal grooves and gently rugate. The cardinal fossula is invariably located on the convex side of the corallum. In neanic sections, the lumen is almost entirely closed by thickened septa. In R45104c (Pl. 6, fig. 2), the major septa, o-8-1-1 mm thick extend almost to the axis where their ends are coated with sclerenchyme which fills the axial area. Along their length they are laterally contiguous in parts and separated elsewhere by a narrow sub-parallel sided gap between 0-05 mm and 0-2 mm wide crossed at intervals by the traces of tabulae. There is a distinct bilateral symmetry, the cardinal septum being longer than all the other septa and flanked by pinnate groups of metasepta. Minor septa appear mostly as short wedges up to 2:8 mm long and 0-7 mm wide between the peripheral ends of the major septa forming a solid septal stereozone around the cardinal quadrants. In the counter qudrants, however, there is a very narrow unthickened zone around the periphery, about half as wide as the length of the minor septa, in which major and minor septa alike are 0-2—0-3 mm wide. Signs of carinae are seen only rarely as incipient structures within the thickened peripheral parts of some septa. R45104c is 14:5 mm in mean diameter with 35 major septa. In higher sections (R45108a, R45111c), the unthickened peripheral zone in the counter quadrants grows in width and gradually extends circumferentially across the alar areas towards the cardinal septum. Yard-arm carinae appear, one or two per septum, on the peripheral unthickened parts of the septa. In the axis, where there may be a weak counter-clockwise vortex, the core of sclerenchyme disappears in most cases and the ends of the septa, usually just short of the centre, may be gently tapered in an axial zone about 0-2 of the diameter across. The cardinal septum shortens and the arrangement of the flanking metasepta defines a long narrow cardinal fossula, well developed in some sections but hardly distinguishable in others. Septal thickening is developed on the crests of the innermost series of dissepiments and the tabulae. It begins to diminish first in the counter quadrants and often disappears here completely by the late neanic—early ephebic stage. The boundary between the unthickened and thickened portions of the septa remains sharp until all trace of thickening disappears. The cardinal quadrants, however, are usually heavily thickened into and throughout much of the ephebic stage, often with the septa laterally contiguous except in the long narrow cardinal fossula. The outer edge of the thickened zone moves inwards towards and usually across the dis- sepimentarium/tabularium boundary, and the zone itself may be as small as half the FROM THE SIERRA DE PERIJA, VENEZUELA 259 diameter in width in some ephebic sections. On the cardinal septum and one or two flanking metasepta, however, the thickening starts very much closer to the periphery and serves to emphasise the cardinal fossula. In ephebic sections the epitheca is about 0-25 mm thick and the peripheral parts of the septa about 0-I mm across. The major septa, with the exception of the cardinal septum and a few flanking metasepta, all reach the axial area usually with tapered ends bilaterally arranged about the counter-cardinal axis. Nearly always the major septa are slightly thickened in the inner part of the tabularium. Sections such as R49278a (PI. 6, fig. 5) with no notable septal thickening at all are rare. One specimen R45115a-b (PI. 5, fig. 6) shows an irregular network of septal material in an axial area about a fifth the diameter across. Minor septa range from a third to just over half the radius in length and those flanking the counter septum may be slightly longer than their neighbours. Yard-arm carinae are developed on major and minor septa in the dissepimentarium, rarely with a strong contrast between counter (better developed) and cardinal quadrants. Variation from specimen to specimen is notable with individual carinae between 0-4 and 0-65 mm in average width, 0-6 to 1:2 mm in average spacing and between means of 3 to 6 perseptum. The intensity of carina- tion tends to increase with increasing size and septal number and with decreasing septal thickening. The appearance of the cardinal fossula is also variable with the cardinal septum ranging between extremes of a quarter to two-thirds of the radius in length in a narrow, parallel sided fossula closed axially by the incurving ends of the adjacent metasepta. Dissepiments are intercepted on average 0-5 mm apart, usually uniserial between adjacent septa. Up to three rows closest to the epitheca are concave abaxially, the rest concave adaxially. They may in some cases loop between carinae rather than from septum to septum. Traces of tabulae are wider spaced and more diffuse. In longitudinal section, the lower parts and much of the peripheral areas par- ticularly on the cardinal side are infilled by sclerenchyme. The dissepimentarium is composed of small globose dissepiments up to 1:2 mm high but normally fairly evenly developed around 0-75 mm high. The width of the dissepimentarium increases towards the calice and contains up to six rows of dissepiments. They are horizontal to slightly reflexed peripherally and steeply sloping downwards axially. The dissepimentarium is notably modified in the cardinal fossula (Pl. 6, fig. 6) where the peripheral two or three rows of normal globose dissepiments are followed axially after a sharp break by large axially and downward sloping vesicles. The tabulae are complete and incomplete. The complete series are broad low domes I-2 mm apart. The incomplete tabulae are mainly subsidiary broad low domes across the axial four-fifths of the tabularium diameter with a few small flanking vesicles. The sclerenchyme coating the septa is commonly seen cresting the tabulae as well. Traces of carinae are confined to the dissepimentarium. Close to the epitheca they slope inwards and upwards at no more than 10° to the wall, but towards the axis they curve over to an angle of about 45° and occasionally as low as 30° to the horizontal. Septal microstructure is only clear in parts, where uniserial monacanthine trabeculae can be seen. The cross-bar carinae which cut, and definitely 260 PALAEOZOIC CORAL FAUNAS precede the septa in formation have the same general appearance as septal tissue but no details of their trabecular structure can be determined. Measurements of septal number against diameter are plotted in Text-fig. ro. Discussion. Although the Venezuelan specimens have a considerably smaller range of adult diameters than topotypic Heliophyllum halli (see Text-fig. 10) there appears to be no structural distinction of note between the two groups of specimens. Size and size linked factors such as septal number and intensity of carination, and possibly also the more prevalent septal thickening in the Venezuelan sample are probably all of ecological rather than genetic importance. Furthermore, although the legendary variability of H. halli in external form is well documented (Wells 1937), variation in internal structure is not so well known neither for this species nor for the host of other North American ‘species’ of Heliophyllum (see for example Stumm (1965 : 35-38)). For these reasons it seems preferable not to distinguish the Venezuelan material in any way from H. halli. 60 50 40 30 0 10 20 30 40 50 60 70 80 90 d (mm) ———e —— Hholli (venezuela) -—-—-—- H. halli (New York State) including ———4——— H.wellsi (holotype) x H.halli OH-hallireflexum + H arachne Fic. 10. Number of major septa plotted against diameter for specimens of Hehophyllum hall. from Venezuela and New York State, and for specimens of H. wellst. FROM THE SIERRA DE PERIJA, VENEZUELA 261 Heliophyllum wellsi sp. nov. (PIG, figs-7, 8; Pll 7, figs 1, 2; Vext-figs ro; r1) DERIVATION OF NAME. After Professor J. W. Wells (Cornell University, Ithaca, NEY..; U.S:A.}. Diacnosis. Large solitary Heliophyllum with dissepimentarium one-half to three-fifths the radius. Attenuate major and minor septa densely carinate in dissepimentarium. Carinae yard-arm, sometimes with branching ends; they may form reticulate network with septa close to the epitheca. Axial core of reticulate septal lamellae about one-fifth the diameter across. Dissepiments small, globose, numerous; tabulae incomplete, tabularium horizontal or slightly domed. HoLotyrPe. R45088 (3126). Cano Grande; Caio Grande Formation, early Middle Devonian (upper Onesquethaw). PARATYPE. R45089 (3126). Same locality and horizon as holotype. DEscripTION. The specimens are large conico-cylindrical fragments of the mature stages of two coralla. The epitheca is moderately rugate. In cross-section the epitheca is 0-2 mm thick and only slightly penetrated by the expanded and shallowly convex ends of the septa. The septa are extremely thin across the broad dissepimentarium, measuring 0:05-0:10 mm across; the peripheral expansion is variable but not more than 0:30 mm across and less than 1 mm long. There is no sign of a cardinal fossula. In the dissepimentarium, the septa bear abun- dant yard-arm carinae, averaging 22-24 per septum, spaced 0-7 mm apart and reaching 2 mm in length near the periphery. The carinae towards the inner dis- sepimentarium are shorter and closer spaced normal cross-bars but many of those in the outer dissepimentarium have bifurcating or geniculate ends. Insome instances septa and carinae may form an intricate reticulate network just inside the epitheca. The minor septa are about as long as the dissepimentarium is wide, that is half to three-fifths the radius. A slight but well defined septal thickening to 0-3-0-4 mm in width takes place at the dissepimentarium-tabularium boundary, gently tapering towards the axis over some 5~7 mm. _ Incipient carinae can be clearly seen embedded in these thickened parts of the major septa in higher ephebic sections. The axial ends of the major septa terminate against an axial area 10-20 mm across containing a dense and rather irregular reticulate network of septal material which is occasionally weakly carinate. The details of this axial structure show considerable variation in the sections available. In longitudinal section the dissepimentarium is composed of numerous small globose dissepiments on average 0-7 mm high, horizontally arranged near the periphery and curving over to slope steeply downward towards the axis. The tabulae are incomplete longer, flatter, vesicular plates arranged horizontally or slightly domed across the axis. In the axial three-fifths of the tabularium diameter, the tabulae are disrupted by numerous, vertical and sub-parallel, wavy septal lamellae. The dissepimentarium is crossed by the traces of carinae, very regularly spaced 0:5-0-6 mm apart normal to their length. The carinae are inclined 25—30° to the epitheca at the periphery, curving over gently through 40° to lie approximately 262 PALAEOZOIC CORAL FAUNAS 40° to the horizontal at the inner edge of the dissepimentarium. They may just penetrate the tabularium, bending sharply upwards through about 20° as they do so. Diameter and septal number in the holotype are plotted on Text-fig. 10. The paratype, of comparable size, is too crushed to measure accurately. The fine structure of the epitheca, dissepiments and tabulae can be clearly seen to be fibronormal. The septa appear to be composed of unitrabecular monacanths and the fine structure of the carinae is of septal type but is otherwise obscure. Discussion. Heliophyllum wellsi is distinguished in general by its large size, the intensity of the septal carination and the apparent lack of a cardinal fossula. The most striking feature of the new species, is however, the axial core of reticulate septal lamellae which appears to be unique to H. wellsi among species of Heliophyllum. Genus CYLINDROPHYLLUM Simpson 1900 1938 Cylindrophyllum Stewart: 43. 1949 Cylindrophyllum Stumm: 33. 1949 Cylindrophylium Ehlers & Stumm: 23. Fic. 11. Comparative septal carination in (a) Cylindvophyllum elongatum (R45116a) x25, (b) Heliophyllum halli (R45115a) x10, and (c) Heliophyllum wellsi (R45088a) x6. FROM THE SIERRA DE PERIJA, VENEZUELA 263 Diacnosis. Dendroid or phaceloid rugose corals. Corallites with attenuate major and minor septa and radial symmetry. Septa with variably developed yard- arm carinae in dissepimentarium, the major septa irregular in course with occasional lateral projections in tabularium and frequently withdrawn from axis. Dis- sepiments small, globose, usually sloping steeply downwards towards the axis. Tabulae mainly incomplete; tabularium surface usually bowl-shaped, occasionally flat or more rarely arched. TYPE SPECIES (by original designation). Cylindrophyllum elongatum Simpson 1900 : 217, text-fig. 42. Onondaga Limestone (Edgecliff Member), early Middle Devonian; Clarksville, Albany Co., N.Y., U.S.A. Discussion. Cylindrophyllum has been commonly placed in the Disphyllidae but the septal characteristics in particular suggest that the genus is more likely to be related to the Heliophyllum-Billingsastraea group (see discussion of the Cyatho- phyllidae). Cylindrophyllum elongatum Simpson 1900 (Pl. 7, figs 3-6; Text-fig. 11) 1900 Cylindrophyllum elongatum Simpson: 217, fig. 42. DracGnosis. Phaceloid Cylindrophyllum with corallites up to about 20 mm diameter. Major and minor septa almost equal in length, more or less confined to dissepimentarium. Septa straight or zigzagged, carinae usually well developed but may be variable to almost absent in some corallites. Dissepiments small, globose, up to four rows deep, strongly sloping axially and downwards. Tabulae complete or incomplete; tabularium surface flat or bowl-shaped, rarely arched. MATERIAL. R45116 (3539). Slipped block in Cafio Pescado; Cano Grande Formation or lower part of Cafio del Oeste Formation, early Middle Devonian (upper Onesquethaw). DISTRIBUTION. Onondaga Limestone; New York, U.S.A. Cafio Grande Fm. or lower Cafio del Oeste Fm. Sierra de Perija, Venezuela. Both early Middle Devonian. DEscrRIPTION. The specimen is a portion, 120 mm xX I00 mm xX 70 mm, of a phaceloid colony in which most of the corallites are partially crushed. There are no signs of connecting processes between corallites in this colony. In cross-section the corallites were nearly all originally circular with a regularly developed fibronormal epitheca 0-1 mm thick. In a few cases, adjacent corallites grow with short lengths of their epithecae cemented back to back. Major and minor septa are uniformly attenuate, 0-03 mm thick on average, except in the peripheral 1 mm or less in which they gradually thicken towards the epitheca. The septa have very bluntly wedge-shaped peripheral ends reaching 0-25 mm in thickness which may penetrate up to half the thickness of the epitheca. Major and minor septa appear to be formed of unitrabecular monacanthine trabeculae and are dif- ferentiated by no more than 0-3 to 0-4 mm in length. They reach between a half 264 PALAEOZOIC CORAL FAUNAS and two-thirds the radius to the axis, the minor septa ending at the tabularium/ dissepimentarium boundary, the major septa just penetrating the tabularium. The septa may be straight or zigzagged. A few corallites have scattered carinae on one or two septa only but most corallites have moderately developed carinae on all or nearly all septa. The carinae appear to be of the yard-arm type as developed in Heliophyllum but many are geniculate with a central portion displacing the septum obliquely and only the extremities orientated normal to the septal face, thus giving the appearance of alternate carinae on zig-zag septa. Dissepiments usually appear as a single series between septa of straight to slightly curved bars 0-3-0-5 mm apart. There is a wide axial area between a third and a half the radius across free of septa and crossed only by curved sections of tabulae. No signs of the type of increase were observed. Because of the crushing, accurate measurements were possible on only a few corallites. The largest individual was 10-5 mm in diameter with 29 major septa. In longitudinal section the dissepiments are small, globose and regularly developed 0-4-0-6 mm high. The dissepimentarium is usually four rows, or sometimes possibly five rows of dissepiments deep. The surface slopes axially and downwards at 45°— 55° to the horizontal. The carinae crossing the dissepimentarium may be almost parallel to the epitheca at the periphery but incline more and more towards the axis when traced towards the inner edge of the dissepimentarium where they are 45°—50° to the horizontal. The tabulae are complete and incomplete, arranged horizontally or with a slight axial depression. The complete tabulae are either slightly arched or more usually slightly sagging and are supplemented or partially replaced by large vesicular plates. Discussion. This specimen appears to be conspecific with photographs of the holotype and other specimens from the Edgecliff member of the Onondaga Limestone of New York kindly provided by Dr W. A. Oliver, Jr. (U.S. Geological Survey). The only points of difference appear to be a tendency for the New York material to have a higher percentage of complete tabulae and to reach a larger corallite diameter. The diagnosis for the species given here has been constructed with the aid of these photographs. Family DURHAMINIDAE Minato & Kato 1965b 1965b Durhaminidae Minato & Kato: 28. 1967 Durhaminidae Stephens: 426. 1967 Durhaminidae Federowski: 25. DracGnosis. Solitary, fasciculate, cerioid or plocoid rugose corals. Axial structure with median counter-cardinal plate in early ontogeny, poorly developed or lost in ephebic stage and replaced by few irregularly arranged septal ends or septal lamellae. Minor septa variably developed and scattered third order septa may occur. Septa usually thin, sinuous or zigzagged in dissepimentarium, slightly dilated in tabularium. Lonsdaleoid dissepiments may be variably developed, the others concentric or herringbone: all dip steeply axially and downwards. Tabulae mainly incomplete, FROM THE SIERRA DE PERIJA, VENEZUELA 265 peripherally flat or trough-shaped, axially flat-topped or arched domes. Clinotabellae usually present. Axial tabellae may or may not be clearly differentiated. TYPE GENUS. Durhamina Wilson & Langenheim 1962 : 504. Discussion. The specimen described below as Durhamina sp. nov. definitely possesses poorly developed sporadic third order septa and an examination of Easton’s (1960, text-fig. 18) illustration of the holotype of Lonsdaleia cordillerensis (lower right quadrant of largest corallite) shows them to be present to a similar extent in the type species of Durhamina itself. This serves to underline the extremely close relationship between the Durhaminidae and Waagenophyllidae and is of particular interest in view of Minato & Kato’s (1965a : 13) observations on tertiary and higher orders of septa in the Waagenophyllidae. The diagnosis given above is emended accordingly. Members of the Durhaminidae occur in the Middle Carboniferous to Middle Permian rocks of Eurasia, western North America and South America. Genus DURHAMINA Wilson & Langenheim 1962 1962 Durhamina Wilson & Langenheim: 504. 1965b Durhamina Minato & Kato: 34. Diacnosis. Fasciculate Durhaminidae with weakly constructed axial structure and scattered, weakly developed third order septa. Lonsdaleoid dissepiments sporadically developed. TYPE SPECIES (by original designation). Lonsdaleia cordillerensis Easton 1960 : 580, Text-figs 17, 18. Lr. Permian; nr. Ruth, eastern Nevada, and other localities in eastern Nevada and southern California, U.S.A. Discussion. Wilson & Langenheim (1962 : 504) provided a very full comparison of Durhamina with similar genera, supplemented by Minato & Kato (1965b : 34). It is only necessary here to draw attention to the scattered third order septa (see discussion of family) not hitherto recorded in the genus. ?Durhamina sp. (Pl. 7, figs 7, 8) MATERIAL. R49285 (3495). Cafio del Norte; Rio Palmar Formation, Lr. Pennsylvanian. DEscrIPTION. Solitary cylindrical fragment 9 mm maximum diameter. In cross-section the epitheca is thin and smooth. The major septa are between a half and the full radius in length and are very slightly thicker in the tabularium than dissepimentarium. The longer septa meet in the axial area, the shorter septa have free tapered ends. Minor septa are fairly evenly developed, up to a third the radius in length. Dissepiments are concentric between them and their flanking 266 PALAEOZOIC CORAL FAUNAS majors but weakly herringbone beyond their axial ends when they are particularly short. There are no lonsdaleoid dissepiments. There are 21 major septa at a diameter of 7-5 mm. In longitudinal section the dissepimentarium is narrow, a quarter the radius in width, and composed of steeply inclined vesicles averaging I mm long and 0-25— o-3 mm high. The tabulae are complete and incomplete but mostly the latter. The complete tabulae are flat or gently sagging across the axis and downturned peripherally. The incomplete tabulae are undulating plates, inclined gently axially and upwards to meet the vertical trace of a septum in the axis. There are scattered clinotabellae. Discussion. This fragment can be closely matched with corallites from species of Durhamina in which the axial structure is weak, and is therefore assumed to be part of a corallite detached from a fasciculate coral colony. In view of the limi- tations of the single specimen, however, the identification must be treated with reserve. Durhamina sp. nov. (Pl. 8, figs 1-10; Text-fig. 12) MATERIAL. R49276 (3270). From a boulder in the Cano Grande; horizon un- certain, ? Rio Palmar Fm., Lr. Pennsylvanian, but see p. 232. DESCRIPTION. The specimen is a large fragment, some I50 x I50 x 100 mm before cutting, of a phaceloid or in parts cerioid colony set in a tough fine-grained reddish-brown calcareous matrix. No external details of the corallites can be observed. In cross-section, the corallites are circular when free and uncrushed. There are areas in the colony where groups of three to eight corallites become laterally con- tiguous and assume a cerioid habit. Young hysterocorallites have slim major septa, three-fifths to the full radius in length from corallite to corallite but of fairly uniform length within a single corallite. In the smallest corallite (Pl. 8, fig. 5), 4 mm in diameter, the major septa are about 0-07 mm thick near the periphery, tapering towards but not reaching the axis. The counter septum, however, crosses the axial area with a slim blade-like expansion centred on the axis, 0-8 mm long and 0-12 mm thick, which imparts a strong bilateral symmetry to the corallite. In slightly larger corallites (Pl. 8, figs 6, 7) with longer major septa, several of them meet in the axis often deflecting the end of the counter septum or obscuring its identification and the bilateral symmetry is weakened or lost. The peripheral ends of the septa are often, but not in all corallites, interrupted by a single series of lonsdaleoid dissepiments. When the septa are not interrupted they are usually somewhat irregular in course and thickness at the periphery. The dissepiments are very variable in appearance. Among the lonsdaleoid series and the inner two or three series of similar sized normal dissepiments are scattered very much smaller FROM THE SIERRA DE PERIJA, VENEZUELA 267 and more highly globose dissepiments. The dissepiments are irregularly arranged in a single series or weakly herringbone between adjacent major septa. Minor septa are absent or sporadically developed as short spines on the dissepiments. In mature corallites, most major septa extend from the epitheca to the axis with occasional interruptions along their length. The epitheca is a very smoothly and regularly developed fibronormal wall 0-12 mm thick. It shows no signs of septal grooves. The septa which have slightly expanded, flat to shallowly convex or wedge shaped peripheral ends show little or no penetration of the inner face of the epitheca. The major septa are slim and are usually irregularly zigzagged or sinuous in the dissepimentarium due to dissepimental interference. They are not frequently interrupted, however, and there is no well defined series of lonsdaleoid dissepiments. Most septa show a slight and often sharp dilatation at the inner edge of the dissepi- mentarium associated with a slight thickening of the innermost series of dissepiments. Major septa are between 0-1 and 0-2 mm thick at this point from which they taper smoothly towards the axis. In any one corallite about half the major septa will meet in the axis usually resting against an irregular median plate formed by the extended end of what is assumed to be the counter septum. These septa may be slightly thickened in the immediate axial area. The rest of the major septa taper to a point just short of the axis. A few of the septa may be interrupted periaxially thus lending definition to the axial structure but no additional septal lamellae are involved. Also, tabulae are intercepted more frequently in the axial area. Minor septa are extremely irregularly developed. In some corallites they extend to or just across the dissepimentarium/tabularium boundary, or very rarely up to half way across the tabularium to the axis. They are very thin and zigzagged in the dissepimentarium or may be frequently interrupted. At the other extreme the minor septa may be no more than scattered spines on the crests of a few dissepiments. The dissepi- mentarium is between 0-3-0-4 of the radius in width and disordered in appearance. An irregular herringbone pattern is developed either between major septa when the minor septa are weak, or between major and minor septa when the minor septa are strong. In the latter case, scattered third order septa are developed as very short pegs on the epitheca and as short spines cresting dissepiments (PI. 8, fig. 2). Septal structure, although clearly trabecular, cannot be distinguished in detail and the style of increase cannot be determined in cross-section. In longitudinal section the dissepiments are mixed larger, less inflated, rather elongate vesicles and smaller, more globose vesicles. The longer vesicles may reach up to 3-4 mm long and 1 mm high normal to the dissepimental surface, whilst the smaller average about 0-75 mm long and 0-4 mm high. The dissepiments slope steeply and fairly uniformly axially and downwards at about 50° to the horizontal. The tabularium is wide, between 0-6 and 0-7 of the diameter and consists of complete and incomplete plates. Complete tabulae are rare, fairly wide flat-topped domes with the peripheral extremities sigmoidally curved downwards and meeting the dissepimental wall horizontally. The incomplete tabulae are axially dome-shaped and periaxially trough-shaped plates with occasional clinotabellae, combined to give a broad axial plateau to the tabularium surface, which may be gently domed or depressed across the axis. There is no sign of any great modification of the tabulae 268 PALAEOZOIC CORAL FAUNAS in the axial area corresponding to the weak axial structure in cross-section. In longitudinal section, increase appears to be peripheral non-parricidal or lateral. Measurements of corallite diameter and septal number are plotted in Text-fig. 12. Corallites reach a maximum of 18 mm diameter but data are restricted by the partial crushing of many corallites. Discussion. Durhamina sp. nov. compares most closely with D. cordillerensis (Easton) from the Lr. Permian of Nevada and California although it can be distin- guished from that species on a number of points. The most striking is the more regular development and greater simplicity of the axial structure in the Venezuelan species. In D. cordillerensis the axial structure is of quite variable density and may rarely be absent. When it is well developed, however, the septal lamellae involved are clearly distinguished from the major septa, in contrast to the very primitive arrangement in Durhamina sp. nov. The latter also has a more strongly developed dissepimentarium than D. cordillerensis and indeed most other species of Durhamina. Finally the Venezuelan species has larger corallites than the North American species. Mature corallites are about 14 mm mean diameter with 22-25 major septa, reaching a maximum of 18 mm diameter in D. sp. nov., compared to mean diameter about 8-9 mm with 22 to 28 major septa (Easton 1960 : 580) or 20-22 major septa (Wilson & Langenheim 1962 : 506) and maximum diameter about 15 mm in D. cordillerensis. The Venezuelan species has not been formally named here as it is represented by only a single specimen from a loose block of uncertain stratigraphic position. 30 20 di (mm) Fic. 12. Number of major septa plotted against diameter for corallites in one colony of Durhamina sp. nov. FROM THE SIERRA DE PERIJA, VENEZUELA 269 Suborder CYSTIPHYLLINA Nicholson 1889 Family CYSTIPHYLLIDAE Edwards & Haime 1850 Subfamily PLASMOPHYLLINAE Dybowski 1873 1964 Plasmophyllinae; Birenheide ; 15. Genus PLASMOPHYLLUM Dybowski 1873 1873 Plasmophyllum Dybowski : 340 1964 Plasmophylium (Plasmophylium); Birenheide: 16 cum syn. Diacnosis. Solitary or weakly colonial corals in which the lumen is filled by vesicular tissue which may or may not be thickened with sclerenchyme. Septa are absent or represented by short irregularly developed spines. The vesicles in general increase in size towards the axis and are arranged with a shallow bowl-shaped to inverted sub-conical surface. Thickening when present is based on this surface and may extend from epitheca to axis or be confined to a single concentric zone. TYPE SPECIES (by subsequent designation of Schltiter 1885 : 10). Cyathophyllum limbatum Quenstedt 1879 : 465, pl. 158, fig. 37 (non fig. 38) = Cyathophyllum goldfussi Edwards & Haime 1851 (non Castelnau 1843). Gerolstein, Eifel, Germany; Ahrdorf- Schichten, Middle Devonian (Eifelian). Discussion. Birenheide (1964) has provided a most valuable review of the Devonian cystimorph group of corals, placing them all, with a few exceptions, in the subgenera Plasmophyllum (Plasmophyllum) and Plasmophyllum (Mesophyllum). The writer feels, however, that these two groups deserve full generic status based on the distinction in septal development between them and that their close relationship is best indicated at the subfamily level. Plasmophyllum secundum americanum (Edwards & Haime) 1851 (Pl. 9, figs 1-4) 1851 Cystiphyllum americanum Edwards & Haime: 464, pl. 13, figs 4, 4a. 1964 Plasmophyllum (Plasmophylium) secundum americanum (Edwards & Haime) Birenheide: 23, pl. 24, fig. 117 (cum syn.). Diacnosis. Trochoid or turbinate tending to subcylindrical Plasmophyllum with vesicles clearly separable into two concentric zones. Vesicles of the inner zone slightly larger than in the outer zone and increasing in size towards the axis in both zones. Short blunt septal spines developed, particularly in the inner half of the outer zone. Usually sclerenchyme is heaviest in a broad concentric band or bands coating or completely obscuring the vesicles in the outer part of the inner zone. MATERIAL. R45008 (3157). Cafio Grande; Cafio Grande Formation, early Middle Devonian (Upper Onesquethaw). R45106, R49270, R49272 (3221). Cafio del Sur; Cafio Grande Formation, early Middle Devonian (upper Onesquethaw). D 270 PALAEOZOIC CORAL FAUNAS DIsTRIBUTION. Widespread in rocks of Middle Devonian age in North America. Cano Grande Fm., early Middle Devonian; Sierra de Perija, Venezuela. Also present in the Middle Devonian of Spain, U.S.S.R., south-east Asia and Australia according to Birenheide (1964 : 24, synonymy). DESCRIPTION. Turbinate corals up to 40 mm diameter and 70 mm height. Calice a shallow bowl with a strongly everted rim. Epitheca moderately rugate. In cross-section, the epitheca is fibronormal and of somewhat variable thickness, 0-I-0-3 mm thick in the small part preserved. The lumen is entirely filled by vesi- cular tissue, fine in an outer zone one-third the radius across in smaller sections to one-half the radius in the largest subcalicular section, and coarser in the axial zone. In the peripheral area the vesicles are highly globose and often circular or sub- circular in cross-section, but somewhat variable in shape and generally ranging in diameter from 0-5 to 2mm. They may be thin walled or lightly coated with fibro- normal material locally extended into short, blunt septal spines up to I mm long and 0-5 mm across at the base (Pl. 9, fig. 3). The coating has a general concentric disposition in the corallum but does not form obvious zones. Towards the axis the vesicles become less and less globose and the circular sections disappear. In the general region of the boundary between the zones of large and small vesicles the thickening abruptly increases. It may form a single broad somewhat ill defined concentric band extending well into the zone of larger vesicles and ringing a rela- tively narrow axial zone of lighter thickening about one-fifth the diameter across. In other sections this may grade into a series of relatively narrow concentric bands of thickening in the zone of larger vesicles. The thickening partly obscures the septal spines which are coarser here and are developed to within a short distance of the axis. Often isolated spots of sectioned septal spines occur in the immediately periaxial area. The larger vesicles average about 2 mm across periaxially and increase in size towards the axis. In longitudinal section, the relative width of the zone of smaller dissepiments and the degree of eversion of its upper surface is seen to increase with increasing size. The vesicles gradually increase in size towards the axis and the septal spines are best developed just outside the zone of larger vesicles and heavy thickening. The spines may reach I-7 mm in length. Their microstructure is not perfectly clear but they appear to consist of localized broad fans in the fibronormal material. At the inner edge of the zone of smaller vesicles the surface slope is about 50° to the horizontal. The larger vesicles are much thickened and may be completely obscured in a dense mass of sclerenchyme. Towards the edge of the zone the thickening can be seen as separate cones on the surface of the vesicles but the cones tend to merge in the axial area. The vesicles in the axial area do not exceed 3:5 mm across their bases, and are arranged with a shallow bowl-shaped surface across the axis. Discussion. Birenheide (1964 : 23), in his comprehensive revision of the Devonian cystimorphs, made Cystiphyllum americanum a subspecies of the Middle Devonian species Plasmophyllum secundum. In particular he noted that larger examples of P. secundum americanum could scarcely be distinguished from the West European subspecies P. secundum conistructum Quenstedt. To judge from Birenheide’s FROM THE SIERRA DE PERIJA, VENEZUELA 271 figures, however, and from a general comparison of American and German material, there would appear to be several points of difference. Plasmophyllum secundum americanum has much more regularly developed vesicles, both in size and shape, particularly in the peripheral areas of the corallum. Furthermore, the development of short blunt wedge-shaped septal spines, most conspicuous in the inner part of the zone of smaller vesicles, is a characteristic feature of P. secundum americanum but does not appear to be present in P. secundum conistructum. On the other hand, typical P. secundum americanum, for example R45008 figured here (PI. 9, figs 1-4), compares remarkably well with Wedekind & Vollbrecht’s (1931, pl. 17, figs 1-2) illustrations of Paralytophyllum crassum (Wedekind) from the Givetian of Germany. In particular, the development of septal spines and the size are more or less identical, as is the calice shape, although this appears to be more variable in P. secundum americanum as a whole. Birenheide (1964 : 28) listed Parvalytophyllum crassum in his synonymy for P. secundum pseudoseptatum (Schulz). The fact remains that but for the geographical isolation, any sort of distinction between the American material and European P. secundum would be difficult to maintain. Birenheide’s synonymy for P. secundum americanum is accepted uncritically. Plasmophyllum sp. (PI. 9; figs 5= 7) 1943 Zonophyllum sp. Wells: 98, pl. ro, figs 5, 6. MATERIAL. R45095, R49259 (3154), PRI 24425 (PRI 24427 cut from same specimen). Cafio Grande; Cafio Grande Fm., early Middle Devonian (upper Onesquethaw). DESCRIPTION. Small subcylindrical fragments 12-16 mm in diameter. The corallum may be strongly constricted at intervals as a result of rejuvenation. In cross-section the vesicles appear rather flattened sub-semicircular and variable in size from about 0-5 mm to 5 mm across their bases. The epitheca and two or three concentric internal surfaces are heavily thickened with fibronormal scleren- chyme. The external wall thus varies between 0-6 and 1-5 mm in thickness. When it is thinner, short blunt septal spines can be distinguished extending inwards about I mm from the periphery and approximately 0-6 mm apart. Thickening of the wall tends to obscure these spines completely. Similar septal spines may or may not occur associated with inner zones of thickening. The vesicles show a general increase in size towards the axis but the gradation, although irregular, is more or less continuous and inner and outer size zones cannot be distinguished. In longitudinal section, the vesicles slope axially and downwards at the periphery and flatten out across the axis giving a bowl-shaped calice. The upper surface of the vesicles is frequently thickened. At intervals up the length of the corallite the lumen is entirely filled by sclerenchyme, more heavily deposited at the periphery so that vesicular tissue is seen only infrequently to reach the epitheca unobscured. D* 272 PALAEOZOIC CORAL FAUNAS Discussion. Wells (1943 : 98) originally described some of this material as Zonophyllum sp. Following Birenheide’s (1964) revision of the cystimorphs, the Venezuelan material clearly belongs to the group of Zonophyllum he regards as congeneric with Plasmophyllum. Wells (op. cit.) noted that the Venezuelan material was similar in structure to Cystiphyllum contfollis Hall (1876, pl. 30, figs 3-9) as redescribed by Fenton & Fenton (1938 : 232, fig. 20, pl. 23, figs 9, 10, pl. 24, figs 4-7). Unfortunately the Fentons did not section and figure Hall’s type material, all of which looks from Hall’s figures to be too highly cystose to be closely related to the Venezuelan species. On the other hand, one specimen figured by Fenton & Fenton (1938, pl. 24, figs 5-7) from the Hamilton Group (Moscow Fm.) of Leicester, N.Y. as ‘Cystiphyllum’ conifollis looks extremely similar to the Venezuelan specimens and may indeed be conspecific with them. It seems best therefore neither to use the name conzfollis nor to erect a new species until Hall’s original material has been redescribed. Order TABULATA Edwards & Haime 1850 Suborder FAVOSITINA Sokolov 1962 Family FAVOSITIDAE Dana 1846 Genus FAVOSITES Lamarck 1816 1816 Favosites Lamarck: 204. 1851 Emmonsia Edwards & Haime: 152, 245. 1939 Favosites Lecompte: 80, cum syn. 1953 Favosites Ross: 48, cum syn. 1960 Favosites Philip: 196. 1965 Favosites Stumm: 60. 1965 Emmonsia Stumm: 65. DraGnosis. Massive, ramose, foliaceous or encrusting cerioid corals. Corallites generally contiguous, polygonal, of similar diameter throughout their length. Corallite walls unthickened internally, but may be variably dilated in a thin peripheral zone. Mural pores confined to corallite wall faces, usually one or two but may be three to a face, arranged in vertical series. Septa present in some species, may be variably developed usually as spines in vertical rows but also as septal striae. Squamulae also present in some species, with or without septal spines, and also may be variable in development. Tabulae usually complete and horizontal, frequently suspended in squamulate forms. TYPE SPECIES (see Edwards & Haime 1850: 60). F. gothlandicus Lamarck 1816 : 205 = Specimen . . . corallinum Gothlandicum Linnaeus 1745 : 39, fig. 27 and 1749 : 106, pl. 4, fig. 27. Discussion. Considerable discussion of the genus Favosites has been published. It is sufficient here to state that the writer accepts Philip’s (1960 : 190) opinion on the status of squamulae in Favosites and also considers that Thamnopora should be maintained as a distinct genus. FROM THE SIERRA DE PERIJA, VENEZUELA 273 Favosites venezuelensis (Weisbord) 1926 (Pl. ro, figs I, 2) 1926 Pleuvodictyum venezuelense Weisbord: 6, pl. 1, figs 8, 9. 1943 Thamnopora venezuelensis (Weisbord) Wells: 99, pl. Io, fig. ro. Diacnosis. Favosites, apparently nodular in form, with monomorphic corallites, up to 2 mm diameter, having walls unthickened within the colony. Corallite walls heavily thickened in a subcalicular zone 5~7 mm deep so as to completely close the lumen in most cases. Squamulae present, particularly well developed in thickened zone. Mural pores uniserial, sporadically biserial, 0-3 mm diameter. Tabulae complete, horizontal, 7 in 5 mm. Hototyre. PRI 21597 (PRI 24428 same specimen). Loose block (?) in Cafio del Norte; Caio Grande Fm., early Middle Devonian (upper Onesquethaw). MATERIAL. R45079 (3121). Cafio Grande; same horizon as holotype. R45101 (3182). Cano del Sur; same horizon as holotype. DEscrRIPTION. Nodular colonies; R45101 is partly crushed, 85 mm long and about 50 mm maximum diameter. Internally, the corallites are polygonal in cross-section, 3 to 4 sided in immature stages and mainly 6 to 7 sided when mature. They range up to almost 2 mm dia- meter when measured between the axial plates of opposite walls but mature corallites average about I-4 mm diameter. The corallite walls are thin with the axial plate about 0-01—0-02 mm thick coated on either side by fibronormal tissue 0-02 to 0-04 mm thick in mid-wall positions. The fibronormal tissue thickens slightly into the corners of the corallites which are slightly rounded. Mural pores appear sporadically in mid- wall or slightly offset positions. No definite indications of septal spines or squamulae are seen in cross-sections through the centre of the colony. Tangential sections close to the colony surface show variable degrees of thickening of the corallite walls (Pl. ro, fig. 2). Thickening first builds up in the angles of the corallite walls so that the lumen becomes increasingly rounded. At the same level increasing signs of squamulae appear, of variable size with truncate or bluntly tapering ends. There are also very rare indications of what may be septal spines. Towards the colony surface, wall thickening may build up to almost or completely close the corallites. The mural pores however appear to be least affected by this excess thickening and even maintain close to their usual diameter when the corallites they connect are apparently closed. At the colony surface, the thickening rapidly diminishes leaving calices rounded and separated by walls about 0-3 mm thick at their midpoints. In longitudinal section the tabulae are complete, flat or slightly arched or sagging, and fairly wide spaced 7in5 mm. They either cross from wall to wall or link facing squamulae. The squamulae are flat to very faintly trough-shaped plates directed horizontally in most cases. They are up to 0-8 mm long, 0:5 mm wide and become more strongly developed as wall thickening begins to build up where they average 0-6 mm apart vertically. Corallites are normal to the colony surface for about Io mm or rather more of their length and the corallite walls start to thicken quite 274 PALAEOZOIC CORAL FAUNAS rapidly about 5-7 mm from the surface. Associated with the thickening, both squamulae and tabulae become more numerous, the former also becoming increasingly thickened. Mural pores are sectioned more frequently due to the thicker walls. They are large and circular about 0-3 mm diameter and usually 1 mm apart in single vertical rows. Most corallite walls show occasional doubling up of the pores with variable horizontal spacing. Many pores appear to be capped by squamulae but the relationship is apparently not invariable. In longitudinal section, the calices appear to be bow] or funnel shaped and I to 2 mm deep for the most part but there is considerable variability. Most but not all of them are floored by a zone where the infilling of the corallites is more or less complete. Discussion. This description is based mainly on the new material collected by Bowen, particularly R45101, and considerably supplements the previous descriptions. In particular, the well developed squamulae in this species have not been recorded before. The species should clearly be placed in Favosites as that genus is now understood. The character of the wall thickening in F’. venezuelensis although apparently unique in its development is comparable with that seen in many other species of Favosites (for example see Ross 1953, pl. 21, fig. 6; Philip 1960, pl. 24, fig. 4), and is essentially a mature surface feature of the colony. In Thamnopora the corallites are more or less rounded throughout the colony. F.. venezuelensis shows some similarities with Emmonsia radiciformis (Stewart 1938 : 69, pl. 16, figs 8-10, non Rominger 1876 : 34, pl. 12) and to a lesser extent with Emmonsia carmani Stewart (1938 : 67, pl. 15, figs 1-3). Both these species from the Middle Devonian of Ohio, however, appear to be more intensely squamulate and lack the strong thickening of the distal corallite walls. Favosites arbuscula Hall 1876 (Pl. ro, figs 3-6) 1876 Favosites arbuscula Hall, pl. 36, figs 1-9. 1953 Favosites (Emmonsia) arbuscula (Hall); Ross: 54, pl. 14, figs 2-9. 1965 Emmonsia arbuscula (Hall); Stumm: 65, pl. 64, fig. 13. Di1aGNosis. Ramose Favosites with branches averaging 14 mm diameter. Coral- lites oblique or perpendicular to surface, 1:2 to 1°38 mm mean mature diameter. Walls 0-1 mm thick with little or no thickening at aperture. Mural pores 0-2 mm diameter, uniserial. Squamulae abundant, sometimes concentrated in distinct zones. Tabulae thin, often suspended from squamulae, I-1-5 mm apart. MATERIAL. R45102-3, R49264-5 (3182). Cano del Sur; Caio Grande Formation, early Middle Devonian (upper Onesquethaw). DisTRIBUTION. Middle Devonian (Hamilton Group and equivalent units) of eastern North America. Cafio Grande Fm. (early Middle Devonian) of the Sierra de Perija, Venezuela. FROM THE SIERRA DE PERIAJ, VENEZUELA 275 DESCRIPTION. Colonies ramose with branches averaging 14 mm diameter, close packed or diverging. Largest fragment 80 mm long branching twice in planes at right angles to each other. In cross-section the corallites are three or four sided when immature, six or seven sided when mature. There is a tendency for mature and immature corallites to be arranged in a regular pattern. Mature corallite diameters reach 1-8 mm and average about 1-5 mm. Their walls are thin, the axial plate about 0-02 mm thick flanked on either side by fibronormal tissue 0-04 mm thick. There is only slight rounding of corallite anglesinternally. Mural pores are infrequently sectioned, always close to the mid-wall position and signs of septal spines or squamulae are extremely rare. At the colony surface, corallite walls are slightly thickened up to about 0-25 mm from side to side and often irregularly scalloped or provided with very short stubby spines. The corallites are nearly all polygonal with rounded corners and are only rarely subcircular. Mural pores are sectioned more frequently and occasional signs of squamulae are seen. In longitudinal section, the corallites diverge at a low angle to the axis of the branch and curve over to open at the surface somewhat obliquely. Internally corallite walls show slight irregular thickening and thinning with a more marked but only slightly thicker zone within 1 to 1-5 mm of the branch surface. Mural pores are infrequently sectioned up to 0-2 mm diameter and of indeterminate vertical spacing. Squamulae are present as generally unthickened plates of flat to shallow trough-shaped section, horizontally or slightly upwardly directed and about 0-5 mm long. They are concentrated about 0-25 mm vertically apart in distinct zones I to 2 mm deep and with a vertical frequency of 4 to 5 mm. The tabulae are complete, unthickened, flat to gently undulating and mostly wide spaced I to 1-5 mm but sometimes up to 4mm apart. They are often suspended from the squamulae. Discussion. The Venezuelan material appears to agree in most respects with Favosites arbuscula Hall as redescribed by Ross (1953 : 54). Whether or not Favosites digitata Rominger (1876 : 39, pl. 15) is conspecific with and therefore a senior synonym of F. arbuscula cannot be determined in the absence of a modern description and illustration of Rominger’s species. For the present it seems preferable to accept the well described F. arbuscula. The Venezuelan material possesses a combination of characters within the range of F. arbuscula but not matched exactly in any of Ross’ (1953 : 55) subspecies. No new subspecies designation is felt to be justified, however, and the material is referred to F. arbuscula sensu lato. VI. REFERENCES ALTEvocT, G. 1968. Devonian tetracorals from Spain and their relation to North American species. In Oswatp, D. H. (Ed.) International Symposium on the Devonian System, Vol. 2 : 755-769, 2 pls. Calgary. ARMSTRONG, A. K. 1962. Stratigraphy and palaeontology of the Mississippian System in southwestern New Mexico and adjacent southeastern Arizona. Mem. Inst. Min. Technol. New Mex., Socorro, 8 : i-vi, 1-99, pls 1-12. ARNOLD, H. C. 1966. Upper Palaeozoic Sabaneta-Palmarito sequence of Mérida Andes, Venezuela. Bull. Am. Ass. Petrol. Geol., Tulsa, 50 : 2366-2387. 276 PALAEOZOIC CORAL FAUNAS BASSLER, R.S. 1937. The Paleozoic rugose coral family Palaeocyclidae. J. Paleont., Tulsa, 11 : 189-201. Biriincs, E. 1858. Report for the year 1857 of E. Billings, Esq., Palaeontologist, addressed to Sir W. E. Logan, F.R.S., Director of the Geological Survey of Canada. Geol. Suv. Canada, Report of Progress for the year 1857 : 147-192. 1875. On some new or little known fossils from the Silurian and Devonian rocks of Ontario. Can. Nat., Montreal, N.S., 7, 4 : 230-240. BIRENHEIDE, R. 1963. Cyathophyllum- und Dohmophyllum- Arten (Rugosa) aus dem Mittel- devon der Eifel. Senckhenburg. leth., Frankfurt a.M., 44 : 363-458, pls 46-62. 1964. Die ‘Cystimorpha’ (Rugosa) aus dem Eifeler Devon. Abh. Senckenb. naturforsch. Ges., Frankfurt a.m., 507 : 1-120, 28 pls. Bowen, J. M. 10972. Estratigrafia del Precretaceo en la parte norte de la Sierra de Perija. Boln Geol. Dir. Geol. Publ. Espec., Caracas, 5 : 729-761. Brice, D. 1971. Etude paléontologique et stratigraphique du Dévonien de 1’Afghanistan. Contribution 4 la connaissance des brachiopodes et des polypiers rugueux. Notes Mém. Moyen-Orient, Paris, 11 : i-viii + 1-364, pls 1-20. Buscu, D.A. 1941. An ontogenetic study of some rugose corals from the Hamilton of western New York. J. Paleont., Tulsa, 15, 4 : 392-411, 73 text-figs. Curist, P. 1927. La coupe géologique le long du chemin de Mucuchachi a Santa Barbara dans le Andes Vénézueliennes. FEclog. geol. Helv., Lausanne, 20, 4 : 769-792. Dana, J. D. 1846. Genera of fossil corals of the family Cyathophyllidae. Am. J. Sci. Arts., New Haven, (2), 1 : 178-189, text-figs 1-5. DE Groot, G. E. 1963. Rugose corals from the Carboniferous of Northern Palencia (Spain). Doctor. Thesis Wis. Naturk. Rijksuniv. Leiden, 1963 : 1-123, pls 1-26. DysBowsk1, W. N. 1873-4. Monographie dey Zoantharia sclerodermata vugosa aus dey Silur- formation Estlands, Norvd-Livlands und der Insel Gotland . . ., pp. 257-532, pls 1-5, Dorpat. Easton, W. H. 1960. Permian corals from Nevada and California. J. Paleont., Tulsa, 34, 3: 570-583, 18 text-figs. Epwarps, H. M. & Hatme, J. 1850. A Monograph of the British Fossil Corals, Part 1. Introduction, etc.: i-Ixxxv + 1-71, pls 1-11. Palaeontogr. Soc. [Monogy.], London. 1851. Monographie des Polypiers fossiles des Terrains palaeozoiques. Avch. Mus. natn. Hist. nat. Paris, 5 : 1-501, pls. 1-20. Euters, G. M. & Stumm, E. C. 1949. Corals of the Devonian Traverse Group of Michigan. Part II, Cylindvophyllum, Depasophyllum, Disphyllum, Endophyllum and Synaptophyllum. Contr. Mus. Paleont. Univ. Mich., Ann Arbor, 8, 3 : 21-41, 8 pls. FEDOROWSKI, J. 1965. Lindstroemiidae and Amplexocariniidae (Tetracoralla) from the Middle Devonian of Skalty, Holy Cross Mountains, Poland. Acta palaeont. pol., Warszawa, 10 : 335-363, pls 1-6. 1967. The Lower Permian Tetracoralla and Tabulata from Treskelodden, Vestspits- bergen. Norsk Polavinstitutt Skriftey, Oslo, 142 : 1-44, 7 pls. Fenton, C. L. & Fenton, M. A. 1938. Heliophylium and ‘Cystiphyllum’ corals of Hall’s ‘Illustrations of Devonian corals’. Ann. Cavneg. Mus., Pittsburgh, 27 : 207-250, pls 1-24. Fontaine, H. 1961. Les Madréporaires paléozoiques du Viét-Nam, du Laos et du Cambodge. Archs géol. Viét-Nam, Saigon, 5 : 1-276. Atlas, pls 1-35. Girty, G. H. 1897. A revision of the sponges and coelenterates of the Lower Helderberg Group of New York. New York State Mus., Ann. Rept, Albany, 48, 2 : 259-322. Hat, J. 1843. Geology of New York, pt. 4, comprising the survey of the Fourth geological district, pp. i-xxvill + 1-683, tables pp. 1-68, pls 1-19, Albany, N.Y. 1876. Illustrations of Devonian fossils; corals of the upper Helderberg and Hamilton groups, pp. 1-7, pls 1-39. Albany N.Y. Haier, W. 1936. Einige biostratigraphische Untersuchungen in der Rohrer Mulde unter besonderer Beriicksichtigung der Keriophyllen. Jb. preu8. geol. Landesanstalt, Berlin, 56 : 590-632, pls 32-43. FROM THE SIERRA DE PERIJA, VENEZUELA 277 Hea, J. P. & Wuitman, A.B. 1960. Estratigrafia y petrologia de los sedimentos precretacios de la parte norte-central de la Sierra de Perija, Estado Zulia, Venezuela. Boln Geol. Dir. Geol. Publ. Espec., Caracas 3, I : 351-376. Hitt, D. 1956. Rugosa. In Moore, R. C. (Ed.) Treatise on Invertebrate Paleontology, F : 233-324, text-figs 165-219. Lawrence, Kansas. Ivanovskil, A. B. 1963. Rugozy ordovika i Silura Sibirskoi Platformy, pp. 1-160, 33 pls. Nauka, Moskva. (In Russian). 1965. Dyrevneisie vugozy, pp. 1-152, 39 pls. Nauka, Moskva. (In Russian). Jerrorps, R.M. 1955. Septal arrangement and ontogeny of some porpitid corals. Palaeont. Contr. Univ. Kans., Lawrence, Coelent. Art 2 : 1-16. JELL, J. S. 1969. Septal microstructure and classification of the Phillipsastraeidae. In CAMPBELL, K. S. W. (Ed.) Stratigraphy and Palaeontology : Essays in honour of Dorothy Hill: 50-73, pls 7, 8. Canberra. Karjo, D. L. 1961. Dopolneniya k izucheniyu streptelazmid ordovika Estonii. Geoloogia Inst. Uurim., Tallinn, 6 : 51-67, pls 1-4. Kapran, A. A. 1969. Chetyrekhluchevye korally zhivetskikh otlozhenii beyanaula, chingiza ipredchingizya. Izv.vyssh.ucheb. Zaved. geol. vaz. Moscow, 12, 1 : 19-29, 4 pls. (In Russian.) Kato, M. 1963. Fine skeletal structures in Rugosa. J. Fac. Sci. Hokkaido Univ. (4, geol. min.) 11 : 571-630, 3 pls. Lamarck, J. B. P. DE 1816. AHistoive naturelle des animaux sans vertébres, 2 : 1-568. Paris. LrecompteE, M. 1939. Les tabulés du Dévonien moyen et supérieur du bord sud du bassin de Dinant. Mém. Mus. r. Hist. nat. Belg., Bruxelles, 90 : 1-229, pls 1-23. Le Mairre, D. 1947. Le récif coralligene de Ouihalane. Notes Mém. Serv. Mines Carte géol. Maroc, 67 : 1-112, pls 1-24. 1952. La faune du Dévonien Inférieur et Moyen de la Saoura et des abords de l’Erg el Djemel (Sud-Oranais). Matér. Carte géol. Algéy., Alger, 1. Paléont., 12 : 1-171, 22 pls. Lippte, R. A., Harris, G. D., & WELLS, J. W. 1943. The Rio Cachiri section in the Sierra de Perija, Venezuela. Bull. Am. Paleont., Ithaca, 27 : 269-368 (1-100), pls 27-36 (1-10). Linnaeus, C. 1745. Dissertatio, Coralia Baltica adumbrans, Quam... submittit Henricus Fougt, pp. viii + 1-40, 1 pl. Upsala. 1749. Corallia Baltica ... proposita ab Henrico Fougt. Amoen. Acad. Stockholm & Leipzig, 4 : 74-106, pl. 4. MILLER, J. B. 1962. Tectonic trends in Sierra de Perija and adjacent parts of Venezuela and Columbia. Bull. Am. Assoc. Petyol. Geol., Tulsa 46 : 1565-1595, 15 figs. Mititer, S. A. 1889-1897. North American geology and palaeontology .. . 3vd Edition, pp. 1-664 (1889). First Appendix, pp. 665-718 (1892). Second Appendix, pp. 719-793 (1897). Cincinnati. Mryato, M. 1961. Ontogenetic study of some Silurian corals of Gotland. Stockh. Contr. Geol., 8, 4 : 37-100, pls I-22. Minato, M. & Kato, M. 1965a. Waagenophyllidae. J. Fac. Sci. Hokkaido Univ. (4, geol. min.), 12, 3-4 : 1-241, 20 pls. 1965b. Durhaminidae (Tetracoral). J. Fac. Sci. Hokkaido Univ. (4, geol. min.), 13, I : 11-86, 5 pls. MirtcHEeLL, M. 1966. Comment on the proposed designation of a type-species for Amplext- zaphventis Vaughan, 1906 (Anthozoa). Z.N. (S.) 1669. Bull. zool. Nomencl. London, 23, 2-3 : 82-83. Moore, R. C., Hirt, D. & WELLs, J. W. 1956. Glossary of morphological terms applied to corals. In Moore, R.C. (Ed.) Treatise on Invertebrate Paleontology, F : 245-251. Lawrence, Kansas. Moorg, R. C. & JeFrorps, R. M. 1945. Description of Lower Pennsylvanian corals from Texas and adjacent areas. Publs. Bur. econ. Geol. Univ. Tex. Austin, 4401 : 77-208, 214 figs. Morates, P. A. 1965. A contribution to the knowledge of the Devonian faunas of Colombia. Boln Geol. Fac. Petrol. Univ. ind. Santander 19 : 51-111, 9 pls. 278 PALAEOZOIC CORAL FAUNAS NeEumAN, B. 1969. Upper Ordovician streptelasmatid corals from Scandinavia. Bull. geol. Instn Univ. Uppsala, N.S. 1, 1 : 1-73, 59 figs. O’ConNELL, M. 1914. Revision of the genus Zaphrentis. Ann. N.Y. Acad. Sci., 23 : 177-192. OxtiverR, W. A., Jr. 1958. Significance of external form in some Onondagan rugose corals. J. Paleont., Tulsa, 32, 5 : 815-837, pls 104-106. 1960a. Rugose corals from reef limestones in the Lower Devonian of New York. Jf. Paleont., Tulsa, 34, 1 : 59-100, pls 13-19. 1960b. Devonian rugose corals from northern Maine. Bull. U.S. Geol. Surv., Washington, 1111A : 1-23, pls I-5. 1963. Anew Kodonophyllum and associated rugose corals from the Lake Matapedia area, Quebec. Prof. Pap. U.S. geol. Suvv., Washington, 430-C : 21-31, pls 9-14. 1968. Succession of rugose coral faunas in the Lower and Middle Devonian of eastern North America. In Oswatp, D. H. (Ed.) Inteynational Symposium on the Devonian System, Vol. 2 : 733-744. Calgary. OPINION 854. 1968. Amplexizaphrentis Vaughan, 1906 (Anthozoa) : designation ofa type-species under the plenary powers. Bull. zool. Nomencl. London, 25, 2-3 : 82-83. PEDDER, A. E. H. 1966. The Devonian tetracoral Haplothecia and new Australian phacello- phyllids. Proc. Linn. Soc., N.S.W., Sydney, 90 : 181-189, pl. 6. Puitip,G.M. 1960. The Middle Palaeozoic squamulate favositids of Victoria. Palaeontology, London, 3 : 186-207, pls 30-34. QUENSTEDT, F. A. von. 1878-81. Petvefactenkunde Deutschlands, 6. Die Rohren- und Stern- korallen., pp. 1-144 (1878); 145-624 (1879); 625-912 (1880); 913-1094 (1881); and Atlas. Leipzig. ROMINGER, C. 1876. Palaeontology. Fossil Corals. Geol. Surv. Michigan, Lansing, 3, 2 : I-161, 55 pls. Ross, M. H. 1953. The Favositidae of the Hamilton Group (Middle Devonian of New York). Bull. Buffalo Soc. nat. Sct., 21 : 37-89, pls 12-27. Rowett, C. L. & SUTHERLAND, P. K. 1964. Wapanucka Rugose Corals. Bull. Okla geol. Suvv. Norman, 104 : 1-124, pls 1-9. SCHEFFEN, W. 1933. Die Zoantharia Rugosa des Silurs auf Ringerike im Oslogebiet. Shr. norske Vidensk-Akad. mat.-nat. Kl., Oslo, 1932, 5 : 1-64, 11 pls. SCHLUTER, C. 1885. Ueber neue Korallen aus dem mitteldevon der Eifel. Sber. niederrhein. Ges. Nat. u. Hetlk., Bonn, 1885 : 6-13. Scrutton, C. T. 1968. Colonial Phillipsastraeidae from the Devonian of south-east Devon, England. Bull. By. Mus. nat. Hist. (Geol.), 15, 5 : 181-281, 18 pls. 1971. Palaeozoic coral faunas from Venezuela, I. Silurian and Permo-Carboniferous corals from the Mérida Andes. Bull. Br. Mus. nat. Hist., (Geol.), 20, 5 : 183-227, 5 pls. SHRESTHA, C. L. 1966. Amplexizaphrentis Vaughan, 1906 (Anthozoa): proposed designation of a type-species under the plenary powers. Z.N. (S.) 1669. Bull. zool. Nomencl. London, 22, 5-6 : 348-350. Srupson, G. B. 1900. Preliminary descriptions of new genera of Paleozoic rugose corals. Bull. N.Y. St. Mus. Albany, 8, 39 : 199-222. SmitH, S. 1945. Upper Devonian corals of the Mackenzie River region, Canada. Spec. Pap. geol. Soc. Am., New York, 59 : vill + 1-126, pls 1-35. SosHKINA, E. D. 1949. Devonskie korally Rugosa Urala. Tyvudy paleont. Inst., Leningrad, 15 : 1-160, pls 1-58. (In Russian.) SosHxina, E. D. & Kapaxovicu, N. V. 1962. Podotryad Streptelasmatina. In ORLov, U. A. (Ed.) Osnovy paleontologii, 2. Gubki, Arkheotsiaty, Kishechnopolostnye, Chervi: 317-324. Moscow. (In Russian.) STEPHENS, C. H. 1967. Leonardian (Permian) compound corals of Nevada. J. Paleont., Tulsa, 41, 2 : 423-431, pls 52-54. STEWART, G. A. 1938. Middle Devonian corals of Ohio. Sp. Pap. Geol. Soc. Am., New York, 8 : vii + 1-120, 20 pls. FROM THE SIERRA DE PERIJA, VENEZUELA 279 Stumm, E. C. 1949. Revision of the families and genera of the Devonian Tetracorals. Mem. geol. Soc. Am., New York, 40 : viii + 1-92, pls 1-25. 1962. Corals of the Traverse Group of Michigan. Part VIII, Steveolasma and Hetero- phrentis. Conty. Mus. Paleont. Univ. Mich., Ann Arbor, 17, 10 : 233-240, 2 pls. 1965. Silurian and Devonian corals of the Falls of the Ohio. Mem. geol. Soc. Am., New York, 93 : x + 1-184, 80 pls. (Dated 1964 but published January 1965, vide Dr W. A. Oliver, Jr.). 1968. The corals of the Middle Devonian Tenmile Creek Dolomite of northwestern Ohio. Contry. Mus. Paleont. Univ. Mich. Ann Arbor, 22, 3 : 37-44, 3 pls. Stumm, E. C. & Tyrer, J. H. 1962. Corals of the Traverse Group of Michigan. Part IX, Heliophyllum. Contr. Mus. Paleont. Univ. Mich., Ann Arbor, 17, 12 : 265-276, 3 pls. Stumm, FE. C. & Watkins, J. L. 1961. The metriophylloid coral genera Steveolasma, Amplexi- phyllum and Stewartophyllum from the Devonian Hamilton Group of New York. /. Paleont., Tulsa, 35, 3 : 445-447, pl. 58. SUTHERLAND, P. K. 1958. Carboniferous stratigraphy and rugose coral faunas of Northeastern British Columbia. Mem. Geol. Surv. Can., Ottawa, 295, 1-177, pls 1-33. 1970. A redescription of the Silurian rugose coral Syvingavon siluriense (McCoy). J. Paleont., Tulsa 44, 6 : 1125-1128, pl. 152. SUTHERLAND, P. K. & HauGu, B. N. 1969. The discoid rugose coral Gymnophyllum : growth form and morphology. Jn CampBeLL, K. S. W. (Ed.) Stratigyvaphy and Palaeontology : Essays in honour of Dorothy Hill: 27-42, pls 3, 4. Canberra. Sutton, F. A. 1946. Geology of Maricaibo Basin, Venezuela. Bull. Am. Ass. Petrol. Geol., Tulsa, 30 : 1621-1741, pls 1-9. THomeson, J. 1881. Onthe genus Alveolites, Amplexus, and Zaphrentis, from the Carboniferous System of Scotland. Proc. R. phil. Soc. Glasg., 13 : 194-237, 4 pls. VauGHAN, A. 1906. An account of the faunal succession and correlation. In Mat Ley, C. A. & VAUGHAN, A. The Carboniferous rocks at Rush (County Dublin). Q. Jl geol. Soc. Lond., 71: 1-52. Wane, H. C. 10948. The Middle Devonian rugose corals of eastern Yunnan. Contr. Geol. Inst., Nat. Univ. Peking, 33 : 1-45, 5 pls. Wessy, B.D. 1964. Devonian corals and brachiopods from the Brendon Hills, West Somer- set. Palaeontology, London, 7, 1 : 1-22, pl. 1. WEDEKIND, R. & VOLLBRECHT, E. 1931. Die Lytophyllidae des mittleren Mitteldevon der Eifel. Palaeontographica, Stuttgart, 75 : 81-110, pls 15-46. WEISBORD, N. E. 1926. Venezuelan Devonian Fossils. Bull. Am. Paleont., Ithaca, 11 : 221-256 (1-36), pls 35-41 (1-7). 1968. The Devonian System in Western Venezuela. Jn Oswatp, D. H. (Ed.) Inter- national Symposium on the Devonian System, Vol. 2 : 215-226, 2 figs. Calgary. We tts, J. W. 1937. Individual variation in the rugose coral species Heliophyllum halk E. & H. Palaeontogr. Am., Ithaca, 2, 6 : 1-22, pl. 1. 1943. Anthozoa. Jn LippiE, R. A., Harris, G. D. & WELLS, J. W. The Rio Cachiri section in the Sierra de Perija, Venezuela. Bull. Am. Paleont., Ithaca, 27 : 363-368 (95- too), pl. 36 (10). WeEYER, D. 1965. Uber Amplexus zaphrentiformis White 1876 (Pterocorallia; Oberkarbon, Pennsylvanian). Geologie, Berlin, 14 : 449-463, 2 pls. Witson, E. C. & LaNGENHEIM, R. L., Jr. 1962. MRugose and tabulate corals from Permian rocks in the Ely Quadrangle, White Pine County, Nevada. /. Paleont., Tulsa, 36, 3: 495-520, pl. 86-89. 280 PALAEOZOIC CORAL FAUNAS VII. APPENDIX—COMPOSITION AND LOCATION OF SAMPLES The most accurate system for the location of samples available is by means of the Maracaibo Cathedral co-ordinates employed by Compania Shell de Venezuela on their I : 10,000 topographic maps. The co-ordinates are in km. All the localities are in the northern part of the Sierra de Perija, Zulia State, Venezuela and the locations of 7m situ samples are plotted in Text-figs I and 2. Cano Los Guineos Formation Caiio Los Guineos section: 3849 Heliophyllum hall Siik-32 W 97°58 Heterophrentis (H.) simplex Stereolasma sp. Syringaxon sp. Cano Grande Formation Canio Grande section: 3121 Bowenelasma typa N Io-11 W 89:36 Favosites venezuelensis Heliophyllum hal 3124 Heliophyllum hall N 10-04 W 89:34 3126 Heliophyllum wellsi N 10-02 W 89:32 3129 Briantelasma oliveri N 9:99 W 89:28 3145 Heliophyllum hall N 9-81 W 89:11 3149 Bowenelasma typa N 9:70 W 89:10 Heliophyllum halli 3154 ? Bowenelasma breviseptata N Heliophyllum hal Plasmophyllum sp. 3157 Bowenelasma typa N 9:81 W 89:04 ? Briantelasma oliveri Heliophyllum hall Plasmophyllum secundum americanum 9°79 W 86-06 Cano del Sur section: 3182 Favosites arbuscula N 9:56 W 88-78 F. venezuelensis 3199 Bowenelasma breviseptata N 9:29 W 89:25 Heliophyllum halla 3221 Briantelasma sp. N 9:05 W 89-79 Heliophyllum halla Plasmophyllum secundum americanum FROM THE SIERRA DE PERIJA, VENEZUELA Cafio del Oeste section (boulder) : 3323 Bowenelasma typa Brniantelasma sp. Heliophylium hall Cafio Colorado section: 3618 Bowenelasma typa Briantelasma oliveri Heliophyllum halli 3619 Bowenelasma breviseptata Heliophyllum hall Cano Grande Fm. or lower Cano de Oeste Fm. Cafio Pescado section (slipped block): 3539 Cylindrophyllum elongatum Cano del Oeste Formation Cafio Grande section: 3243 Hadrophyllum sp. ? Stewartophyllum sp. Rio Palmar Formation Cano del Norte section: 3495 ? Durhamina sp. Cano Colorado section: 3784 Amplexizaphrentis sutherlandi ? Rio Palmar Formation Cafio Grande section (boulder) : 3270 Durhamina sp. nov. DR CoLiIn T. SCRUTTON, Department of Geology THE UNIVERSITY NEWCASTLE UPON TYNE. NE 7RU. E N 12°83 N 6-01 N 6-01 N 6:32 N 9-69 N 16:97 Ne 3-5r N 10:51 W 87-79 W 91:98 W o1-98 W 89:83 W 88-11 W 85:97 W 90°47 W 86-68 281 PLATE1 Syringaxon sp. Fics 1, 2. Cross-sections (slides). R45130b, a. x6. Fic. 3. Longitudinal section (slide). R45130k. x6. Catio Los Guineos; Cafio Los Guineos Fm., early Middle Devonian. ?Stewartophyllum sp. Fics 4, 5. Cross-sections (slides). R45110b, a. X2°5. Cafio Grande; Cafio del Oeste Fm., early Middle Devonian. Stereolasma sp. Fic. 6. Cross-section (slide). R45129a. x3. Fic. 7. Longitudinal section in counter-cardinal plane; C on left (slide). R45120g. x 3. Cano Los Guineos; Cano Los Guineos Fm., early Middle Devonian. Hadrophyllum sp. Fic. 8. Cross-section (polished surface). R49275. 2:5. Cano Grande; Cafio del Oeste Fm., early Middle Devonian. Amplexizaphrentis sutherlandi sp. nov. Fic. 9. Cross-section (peel), early ephebic stage. R45127b (taken from holotype). x3. Fic. 10. Cross-section (slide), early neanic stage 16 mm below R45127b. R45127a (cut from holotype). x3. Fic. 11. Cross-section (peel), neanic stage 14 mm below R45127b. R45127d (taken from holotype). 3. Fic. 12. Longitudinal section in counter-cardinal plane; C on right (slide). R45127e (cut from holotype). x3. Fic. 13. Longitudinal section in counter-cardinal plane; C on left (slide). R4g291b. x2. Both Cano Colorado; Rio Palmar Fm., Lr. Pennslyvanian. All cross-sections orientated with the cardinal septum at top centre. E* Bull. Br. Mus. nat. Hist. (Geol.) 23, 4 IIE ANAS, Ft PAE 2 Bowenelasma typa gen. et sp. nov. Fic. 1. Cross-section (slide), ephebic stage. R45094b (cut from holotype). 2. Fic. 2. Cross-section (peel), neanic stage. R45094d (taken from holotype). 2. Fic. 3. Longitudinal section in counter-cardinal plane; C on left (slide). R45094e (cut from holotype). x2. Fic. 4. Cross-section (slide), early neanic stage. R45100b. 4. Fic. 5. Cross-section (slide), ephebic stage. R45100a. 2. Fic. 6. Longitudinal section in counter-cardinal plane; C on left (slide). R45100c. x2. Fic. 7. Cross-section (slide), ephebic stage, to illustrate septal structure. R45094b (cut from holotype). x6. Both Cano Grande; Cano Grande Fm., early Middle Devonian. Fics 8, 9. Cross-sections (peels), neanic stage. R45121f,e. x2. Fic. 10. Cross-section (slide), sub-calicular level. R4512ta. x2. Fic. 11. Longitudinal section in counter-cardinal plane; C on left (slide). R4512th. x2. Cano Colorado; Cano Grande Fm., early Middle Devonian. All cross-sections orientated with the cardinal septum at top centre. Bull. Br. Mus. nat. Hist. (Geol.) 23, 4 PLATE 2 E** PLATE 3 Bowenelasma breviseptata sp. nov. Fics 1, 2. Cross-section (slide), ephebic stage. R45123a (cut from holotype). Fig. 1 Kee ie, 2, 3<- Fic. 3. Cross-section (slide), neanic stage. R45123b (cut from holotype). x2. Fic. 4. Longitudinal section in counter-cardinal plane; C on left (slide). R45123d (cut from holotype). x2. Cano Colorado; Cano Grande Fm., early Middle Devonian. Fic. 5. Cross-section (peel), neanic stage. R45105¢c. x2. Fic. 6. Cross-section (slide), ephebic stage. R45105b. 2. Fic. 7. Longitudinal section (slide). R45105d. 2. Cafio del Sur; Caio Grande Fm., early Middle Devonian. All cross-sections orientated with the cardinal septum at top centre. Bull. Br. Mus. nat. Hist. (Geol.) 23, 4 PLATE 3 PLATE 4 Briantelasma oliveri sp. nov. Fics 1, 2. Cross-section (slide), sub-calicular. R45090b (cut from holotype). Fig. 1, «2; figs 25 XxX Ae Fic. 3. Cross-section (slide), ephebic stage. R45090a (cut from holotype). x2. Fic. 4. Longitudinal section (slide) in counter-cardinal plane; C on right. R45o0g0e (cut from holotype). x2. Fic. 5. Cross-section (slide), neanic stage. R45091a. x3. Fic. 6. Cross-section (slide), ephebic stage. R45091b. x2. Both Cano Grande; Cano Grande Fm., early Middle Devonian. Fic. 7. Cross-section (slide), sub-calicular. R49290a. 2. Cano Colorado; Cano Grande Fm., early Middle Devonian. All cross-sections orientated with the cardinal septum at top centre. Bull. Br. Mus. nat. Hist. (Geol.) 23, 4 PLATE 4 PLATE 5 Heterophrentis (H.) venezuelensis (Weisbord) Fic. 1. Cross-section (peel), ephebic stage. PRI 21594/PRI 24421 (taken from lectotype). XK 2. Fic. 2. Longitudinal section (peel). PRI 24421 (taken from lectotype). 1-5. ?Loose boulder in Cano del Norte; Cano Grande Fm., early Middle Devonian. Heterophrentis (H.) simplex (Hall) Fic. 3. Cross-section (polished surface), ephebic stage. NYSM 360 (holotype). 1-5. Moscow, New York; Middle Devonian (Hamilton Group). Fic. 4. Cross-section (polished surface), neanic stage. PRI 21593. 1-5. Fic. 5. Cross-section (polished surface), ephebic stage. PRI 21593. X1I°5. ?Loose boulder in Cano del Norte; Cano Grande Fm., early Middle Devonian. Heliophyllum halli Edwards & Haime Fic. 6. Cross-section (slide), ephebic stage. R45115a. 2. Loose boulder in Cano del Oeste; Cano Grande Fm., early Middle Devonian. Fic. 7. Cross-section (slide), ephebic stage. R49289a. x2. Fic. 8. Longitudinal section in counter-cardinal plane; C on right (slide). R49289b. x2. Cano Colorado; Cano Grande Fm., early Middle Devonian. All cross-sections orientated with the cardinal septum at top centre. PLATE 5 nat. Hist. (Geol.) 23, 4 Mus. Bull. Br. PLATE 6 Heliophyllum halli Edwards & Haime Fic. 1. Cross-section (slide), ephebic stage. R45104a. 2. Fic. 2. Cross-section (slide), neanic stage. R4510q4c. 2. Cano del Sur; Cano Grande Fm., early Middle Devonian. Fic. 3. Cross-section (slide), ephebic stage. R45092b. 2. Fic. 4. Longitudinal section in counter-cardinal plane; C on left (slide). R45092Cc. Cano Grande; Cano Grande Fm., early Middle Devonian. Fic. 5. Cross-section (slide), ephebic stage. R49278a. 2. Fic. 6. Longitudinal section in counter-cardinal plane; C on left (slide). R49278d. Loose boulder in Cano del Oeste; Cano Grande Fm., early Middle Devonian. All sections orientated with the cardinal septum at top centre. Heliophyllum wellsi sp. nov. Fic. 7. Cross-section (slide), ephebic stage. R45088b (cut from holotype). I. Fic. 8. Longitudinal section (slide). R45088e (cut from holotype). 1-5. Cano Grande; Cano Grande Fm., early Middle Devonian. Mus. nat. Hist. (Geol.) 23, 4 Bull. Br. PLATE 7 Heliophyllum wellsi sp. nov. Fics 1, 2. Cross-sections (slides), illustrating character of axial structure in ephebic stage. R45088a, b (both cut from holotype). x2. Cano Grande; Cano Grande Fm., early Middle Devonian. Cylindrophyllum elongatum Simpson Fic. 3. Cross-section (slide). R45116a. 2°5. Fics 4-6. Longitudinal sections (slides). R45116h, j, g. 2:5. Loose block in Cano Pescado; either Cano Grande Fm. or low in Cano del Oeste Fm., early Middle Devonian. ?Durhamina sp. Fic. 7. Cross-section (slide). R49285a. x4. Fic. 8. Longitudinal section (slide). R49285b. x 4. Cano del Norte; Rio Palmar Fm., Lr. Pennsylvanian. Bull. By. Mus. nat. Hist. (Geol.) 23, 4 PLATE PLATE 8 Durhamina sp. noy. Fic. 1. Cross-section (slide) of corallite arrangement in colony. R49276b. xt. Fic. 2. Cross-section (slide) of corallite with third order septa. R49276b. x6. Fics 3-4. Cross-sections (slides) showing variation in appearance of mature corallites. R49276c, a. 2. Fics 5-7. Cross-sections (slide) of immature corallites. R49276a. 4. Fics 8-10. Longitudinal sections (slides), fig. 8 showing increase. R49276h, i,j. 2. Loose block in Cano Grande; Carboniferous, (?Rio Palmar Fm., Lr. Pennsylvanian). 8 Dy PEACE Bull. Br. Mus. nat. Hist. (Geol.) 23, 4 PLATE 9 Plasmophyllum secundum americanum (Edwards & Haime) Fic. 1, 2. Cross-sections (slides). R45098b, c. X1°5. Fic. 3. Cross-section (slide) showing form and arrangement of septal spines. R45098b. Fic. 4. Longitudinal section (slide). R45098d. x 1°5. Cano Grande; Cano Grande Fm., early Middle Devonian. Plasmophyllum sp. Fics 5-6. Cross-sections (slides). R45095a, b. 3. Fic. 7. Longitudinal section (slide). R45095c. 3. Cano Grande; Cano Grande Fm., early Middle Devonian. x6. PLATE 9 nat. Hist. (Geol.) 23, 4 Mus. Bull. Br. PLATE to Favosites venezuelensis (Weisbord) Fic. r. Cross and longitudinal sections (slide). R45101b. 3. Fic. 2. Tangential section (slide). Rq45rora. 3. Cano del Sur; Catto Grande Fm., early Middle Devonian. Favosites arbuscula Hall Fic. 3. Tangential section (slide). R45103a. 3. Fics 4-5. Cross-sections (slides). R49264a, R45103b. x3. Fic. 6. Longitudinal section (slide). R49264b. x3. Cano del Sur; Cano Grande Fm., early Middle Devonian. Bull. By. Mus. nat. Hist. (Geol.) 23, 4 PLATE ro >" ——_ > A LIST OF SUPPLEMENTS TO THE GEOLOGICAL SERIES OF THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) . Cox, L. R. Jurassic Bivalvia and Gastropoda from Tanganyika and Kenya. Pp. 213; 30 Plates; 2 Text-figures. 1965. {£6. . EL-NaGcear, Z. R. Stratigraphy and Planktonic Foraminifera of the Upper Cretaceous—Lower Tertiary Succession in the Esna-Idfu Region, Nile Valley, Egypt, U.A.R. Pp. 291; 23 Plates; 18 Text-figures. 1966. 10. . Davey, R. J., DowntE, C., SARGEANT, W. A. S. & Wititams, G. L. Studies on Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 248; 28 Plates; 64 Text- figures. 1966. £7. . APPENDIX. DAvEy, R. J., DowniE, C., SARGEANT, W. A. S. & WILLIAMS, G. L. Appendix to Studies on Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 24. 1969. 8op. . Ettiott, G. F. Permian to Palaeocene Calcareous Algae (Dasycladaceae) of the Middle East. Pp. 111; 24 Plates; 17 Text-figures. 1968. £5.12}. . Ruopes, F. H. T., Austin, R. L. & Druce, E. C. British Avonian (Carboni- ferous) Conodont faunas, and their value in local and continental correlation. Pp. 315; 31 Plates; 92 Text-figures. 1969. {II. . Cuitps, A. Upper Jurassic Rhynchonellid Brachiopods from Northwestern Europe. Pp. 119; 12 Plates; 40 Text-figures. 1969. £4.75. . Goopy, P. C. The relationships of certain Upper Cretaceous Teleosts with special reference to the Myctophorids. Pp. 255; 102 Text-figures. 1969. £6.50. . OwEN, H. G. Middle Albian Stratigraphy in the Paris Basin. Pp. 164; 3 Plates; 52 Text-figures. 1971. 6. . Sippigul, Q. A. Early Tertiary Ostracoda of the family Trachyleberididae from West Pakistan. Pp. 98; 42 Plates; 7 Text-figures. 1971. {8. ae. BCR Printed in England by Staples Printers Limited at their Kettering, Northants establishment as = ae See é . rf wer eS ae ee * Se ate A are Zo aw Ae a PRODEINOTHERIUM FROM GEBEL ZELTEN LIBYA | AEMWUSE I few ee = sae Of ft) aN 4 i = 4X >! 4 S \ \ BY JOHN MICHAEL HARRIS || Centre for Prehistory and Palaeontology, Nairobi — Pp. 283-348; 5 Plates, 16 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 23 No. 5 LONDON : 1973 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 10949, 1s issued in five series, corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become veady. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 23, No. 5 of the Geological (Palaeontological) series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. World List abbreviation Bull. Br. Mus. nat. Hist. (Geol.). © Trustees of the British Museum (Natural History), 1973 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 12 July, 1973 Price £2.85 PRODEINOTHERIUM FROM GEBEL ZELTEN, LIBYA By J. M. HARRIS CONTENTS Page I. ABBREVIATIONS AND ACKNOWLEDGEMENTS . . 9 . : 286 II. INTRODUCTION. : : : 5 : : e : c 286 III. SYSTEMATIC DESCRIPTIONS. : 3 : : 5 6 291 Order PROBOSCIDEA . : : : : : : 291 Suborder DEINOTHERIOIDEA : : é : C 291 Family DEINOTHERIIDAE . : : : : : 291 Genus Deinotherium . : : : F : ; 292 Genus Prodeinotherium . 6 : : : : 5 293 Prodeinotherium Roo : : a . : : 296 Skull. ; : : : : ; 296 OTHER DEINOTHERIID SKULLS . : 3 ; 5 305 Mandible. : : 311 COMPARISON WITH OTHER DEINOTHERE MANDIBLES. 312 Dentition 2 : : : 3 : 313 TOOTH FUNCTION IN DEINOTHERES . : : é 318 Axial skeleton : : : : : d c 319 Appendicular skeleton . ; 330 SKELETAL EVIDENCE ON THE LOCOMOTION OF DEINO- THERES . c ¢ : 5 : ¢ é 341 IV. CoNCLUSIONS . 3 : : 5 : : : : : 344 V. REFERENCES . : : : . ; : , : . 346 SYNOPSIS A large assemblage of vertebrate fossils has been collected from Miocene continental and near-shore marine lithosomes at the Gebel Zelten, Libya. Deinotheres were not prolific but are of interest in that they include the most complete skull yet discovered, plus the associated skull, mandible and partial skeleton of an immature individual. These are described and compared with other deinothere remains and with the postcranial skeleton of fossil and recent elephantoid proboscideans. Features of the skull, skeleton and dentition indicate that the family Deinotheriidae possessed two genera and Ehik’s taxon Prodeinotherium is resurrected for the earlier and more primitive genus to which the Gebel Zelten specimens belong. The facial region of the Libyan Prodeino- therium skull has a downcurved rostrum that is modified to bear a proboscis. The occipital region, basicranium and anterior cervical vertebrae are adapted to impart a powerful downward thrust to the tusks. The cheek teeth are separated into an anterior crushing battery and a posterior shearing battery. The postcranial skeleton of Prodeinotherium and Deinotherium exhibit graviportal adaptations for support of the body weight. The atlas and axis vertebrae of deinotheres are distinctive and other differences in the postcranial skeleton from that of elephantoids may be detected. The scapula and foot bones of Prodeinotherium may be readily recognized but the long bones resemble those of Elephas. More advanced characters shown by Deinotherium include reduction of the scapular spine, shortening of the femur, elongation of the manus and pes, and reduction of the first digit thus leading to functional tetradactyly. These may all be interpreted as cursorial adaptations. 286 PRODEINOTHERIUM I ABBREVIATIONS AND ACKNOWLEDGEMENTS Individual specimens from Gebel Zelten are referred to by their University of Bristol accession numbers, e.g. 6404: 44, 6418: 20 (M.26665). Specimens of Deinotherium giganteum from Howenegg, Germany, and currently at the Johannes Wurtemberg Universitaat, Mainz, had not been given permanent accession numbers at the time of writing and are referred to in tables of measurements as ‘Howenegg’. Recent specimens of Loxodonta africana used for comparison in the section on the axial skeleton include OM 2209 from the National Museum of Kenya and 1961.8.9.82 from the British Museum (Natural History). Comparison between the appendicular skeleton of Prodeinotherium and Elephas maximus was based on the skeleton of an Indian elephant in the University of Bristol Department of Geology (catalogue number U.B.20189). Deinothere material with KNM catalogue numbers is from the Centre for Prehistory and Palaeontology, Nairobi. The catalogue numbers of mammalian specimens from the British Museum (Natural History) are prefaced by B.M.(N.H.)M. or (in tables of measurements) M. In the tables, measurements are given in cms. unless otherwise stated; * = approximate measurement, -+- = incomplete specimen. This paper formed part of a Ph.D. thesis from the University of Bristol that was supervised by Dr R. J. G. Savage for whose help and encouragement I am very grateful. The work was financed by a grant from N.E.R.C. Iam grateful also to Mrs D. M. Leakey who read and criticised the manuscript. Prodeinotherium skull M.26665 was prepared by Mr A. Rixon in the Department of Palaeontology, British Museum (Natural History). Deinothere material from the 1964 and 1966 University of Bristol expeditions was prepared by Mr. M. White of the Department of Geology, University of Bristol. The photographs were prepared by Mr R. Godwin and line drawings of the postcranial material by Mrs M. Clapp. Field work in Libya was made possible by generous co-operation from Esso Standard Libya and Oasis Oil Company of Libya Inc. I am indebted to the staff of the various institutions that have permitted me to examine material in their collections, and especially the late Professor C. Arambourg (Institut de Paleontologie, Paris), Professor H. Tobein (Johannes Wurtemberg Universitaat, Mainz), Professor R. Dehm (Institut ftir Palaontologie, Munich), Dr E. Aguirre (Museo de Sciencias Naturales, Madrid), Dr L. S. B. Leakey (Centre for Prehistory and Palaeontology, Nairobi), and Dr A. Sutcliffe (Department of Palaeontology, British Museum (Natural History)). Il: INTRODUCTION THE Gebel Zelten is an elongate mesa lying approximately 200 kms inland from the Gulf of Sirte and is bounded by the co-ordinates 19}$°-203°E and 28-29°N. The gebel extends for some 140 kms in a northwest-southwest direction. It attains a height of some 40-60 metres at its western extremity but gradually decreases in elevation towards the east until it merges with the Calenscio Serir. At the eastern extremity of the gebel the northern and southern scarps are separated by a distance of some 50 kms. The scarps converge towards the west and near the centre of the gebel they are only 8 kms apart. West of this point the southern scarp diverges FROM GEBEL ZELTEN, LIBYA 287 sharply from the northern before merging at the extreme western extremity (see text-fig. 1). The Gebel Zelten is deeply dissected by wadis that often reach 3 kms in length. Most of the vertebrate sites are located in branches of the wadi systems. The rocks exposed at the Gebel Zelten are of Miocene age and were attributed by Selley (1966) to the Marada Formation. The rocks represent the mixed fluvio- marine facies of the southern part of the Sirte Basin. The succession at Gebel Zelten is predominantly of sandstone, sandy shales and shales, the units dipping northward at an angle of about six seconds (Doust, 1968). According to Doust, the base of the section at the northern scarp consists of intertidal and fluviatile sands and shales that pass upward into bioturbated calcarenites and calcsilicates representing intertidal and subtidal facies. The top of the succession comprises calcareous crossbedded sandstones from an estuarine channel facies. On the southern scarp a mixed intertidal and fluviatile facies passes upwards into intertidal sediments that are ultimately overlain by the estuarine channel facies. Detailed discussions of the geology are given in Selley (1966, 1967, 1968), Doust (1968), and Magnier (1962, 1968). whe NB Y C\ Wadi el'ldem Transportation Corner Fic. 1. Vertebrate localities of the Gebel Zelten. Sites D, E, F and G are located between sites C and H. 288 PRODEINOTHERIUM The vertebrate fauna of the Gebel Zelten was discovered in 1958 by personnel of the Compagnie des Petroles Total (Libye) during a regional reconnaissance of Libya (Magnier, 1962). In December 1960 an expedition from the Museum National d’Histoire Naturelle, Paris, visited Gebel Zelten to exploit the vertebrate remains (Arambourg & Magnier, 1961). Only an area 20 km long at the centre of the northern scarp of the Gebel was prospected in detail but vertebrate specimens were collected from all exposures examined. The material retrieved is currently housed in the Institut de Paléontologie, Paris. A faunal list was given by Arambourg and Magnier (1961), and later revised by Arambourg (1963b). Of the elements listed only Bunolistriodon massai has been described in detail (Arambourg, 1963a) but preliminary descriptions of Afrocyon burollett, Prolibyther.um magniert, Mastodon pygmaeus and an aeopyornithid tibia have been published (Arambourg Ig61a, 1g61b). From 1964, expeditions from the University of Bristol under the leadership of Dr R. J. G. Savage have systematically collected vertebrate material from Gebel Zelten. Exploitation of the southern scarp was undertaken in 1964 and of the northern scarp in 1966. The 1967 and 1968 expeditions expanded and consolidated the results of earlier investigations. In all, well over a thousand specimens, representing several tons of unprepared material, have been collected by University of Bristol parties. This is in addition to the material collected by the 1960 French expedition and to the specimens donated by oil company personnel to the British Museum (Natural History) and American Museum of Natural History. A list of the faunal elements collected was given in Savage and White (1965) and revised in Savage (1967) and Selley (1968). Most of the vertebrate fossils have been collected from channels and flood plains of the fluviatile facies, though remains are not infrequent in the tidal flat facies. At the time of writing 24 sites had been discovered and these were assigned letter prefixes in alphabetical order of discovery (see fig. I). Permanent site numbers and other pertinent data are now on file in the Department of Geology, University of Bristol. All the Gebel Zelten sites occur near the base of the exposed section and range in areal extent from one to ten acres. A table showing the distribution of the Gebel Zelten mammal fauna by sites was given by Savage (in Selley, 1968). Desio (1935) referred the rocks on the southside of the Gebel Zelten to the Aquitanian stage and those on the north scarp to the Burdigalian. The vertebrate fauna was allocated a Burdigalian age by Arambourg (1963b) although it is not clear whether this age was based on the mammalian elements or on the marine invertebrates. Savage and White (1965) confirmed a Burdigalian age for the mammals, and Savage (in Selley, 1968) later refined the age of the fauna to early Burdigalian or even late Aquitanian. Definite conclusions on the age of the Gebel Zelten assemblage must necessarily be delayed until all the mammalian groups represented have been thoroughly investigated. To date only the carnivores (Savage, 1973) and giraffoids (Hamilton, 1973) have been completed. The Gebel Zelten deinothere skulls are not only the earliest skulls, though not the earliest deinothere specimens (which are from East Africa), but are also unusually complete. The best preserved specimen (6418 : 20) was collected in 1966 and was FROM GEBEL ZELTEN, LIBYA 289 deposited in the British Museum (Natural History). The accession number of the specimen is M.26665. All the permanent teeth are erupted and show wear but the zygomatic, squamosal, auditory and pterygoid regions are poorly preserved. A virtually complete skull of an immature individual was collected by the writer from Site D in 1967. This specimen (6404 : 44) was completely buried and was only discovered during the excavation of deinothere limb material, for which purpose site D had been revisited. The skull was lying on its side (the only deinothere skull from Gebel Zelten that was not lying upside down) and had been slightly deformed by compaction after burial. The bone was in an extremely fragile state and the right zygomatic arch disintegrated during excavation. The occipital region frag- mented and became partially deformed during transportation to Bristol. Permanent P3_M1 are fully erupted but only M! shows wear. M2 is only partly erupted. The squamosal, auditory and orbital regions are better preserved than in M.26665. Several of the facial sutures are still unfused. The mandible (6404 : 45) and much of the skeleton belonging to this individual (M.26667a-1) were excavated from the adjacent part of site D. An incomplete deinothere skull (6404 : 14) and much of a mandible (6404 : 13) were collected in 1964 from within a few yards of skull 6404 : 44. Unfortunately skull 6404 : 14 was broken in transit from the Gebel Zelten and only the premaxilla, jugal and dentition are now in a reasonable state of repair. The teeth of the 1964 specimens are in different stages of eruption and wear and are not thought to belong to one individual. The 1968 Bristol expedition to Gebel Zelten and Dor el Talha discovered an in- complete and partly crushed skull at Gebel Zelten site L. Pieces of this skull were collected but all except a fragment of the inferior half of the occiput are still in Libya. The ventral edges of the lateral wings of the squamosal are better preserved on this specimen than on M.26665. Two other partly eroded skulls were discovered at site R by the 1966 expedition. These were allocated field numbers (6418 : 4, 6418 : 19) but have yet to be collected. In addition to the deinothere postcranial material from site D, a cuneiform and isolated vertebrae have been collected from other Gebel Zelten sites. Gebel Zelten deinothere material is appreciably smaller than, and differs morphologically from, equivalent specimens of Mastodon angustidens from Gebel Zelten. In the ensuing descriptions, comparisons with recent elephant remains are based on vertebrae of Loxodonta africana from the British Museum (Natural History), the manus and pes of L. africana from the Osteology Department of the Centre for Prehistory and Palaeontology, Nairobi, and fore and hind limb material of Elephas maximus in the University of Bristol collections. Comparative deinothere material includes the partial skeleton of ‘Deinotherium bavaricum’ in the Naturhistorisches Museum of Vienna, the partial skeleton of D. giganteum from Valladolid in the Museo de Sciencias Naturales of Madrid, and carpals and tarsals of D. gigantewm from Howenegg in the University of Mainz collections. Conclusions on deinothere taxonomy, morphology and functional anatomy were made after viewing specimens in the following institutions: British Museum (Natural History), London; Centre for Prehistory and Palaeontology, Nairobi; Department 290 PRODEINOTHERIUM of Geology Museum, University of Bologna; Museo de Sciencias Naturales, Madrid; Institut de Paleontologie, Paris; Naturhistorisches Museum, Vienna; Institut fir Palaontologie und Historisches Geologie, Munich; Hessisches Landesmuseum, Darmstadt; Naturhistorisches Museum, Mainz; Institut fiir Palaontologie, Johannes Wurtemberg Universitat, Mainz; Naturhistorisches Museum, Wiesbaden. Exten- sive use has been made also of data given by Bergounioux & Crouzel (1962a, 1962b), Dehm (1963), de Pauw (1908), Graf (1957), McInnes (1942), Pacheco (1930), Palmer (1924), Sahni & Tripathi (1957), Stefanescu (1897, 1907), and Tobein (1962). The results of these investigations suggest that some revision of deinothere systematics is necessary. Diagnoses are given at the beginning of the section on systematic description because the Gebel Zelten deinotheres do not appear to belong to the type genus. TABLE I Deinothere specimens from Gebel Zelten Field Accession Field Accession number number Description number number Description 6401: I Left M8 6404: 47 M.266671 Distal end of left 6401: 4 Right maxilla with humerus P3_M3 6405: 98 Left cuneiform 6404: 13 Mandible with tusks 6409: 41 M? fragment and cheek teeth 6410 2 Deciduous Pa 6404: 14 Skull fragments with 6412: I0 Tuskless mandible dentition with worn cheek 6404 : 30 M.26667a Distal end of fibula teeth 6404: 31 M.26667b Right ulna 6412: 39 Left Me 6404 : 32 M.26667c Proximal end of right as eo ae = cf eae 6413: 18 Left M8 6404: 33 M.26667d Right femur 6416 : 120 Tusk tip 6404 : 34 M.26667e Left ulna 6416 : 130 Race 6404: 36 M.26667{ Right scapula 6418: 20 Adult skull 6404: 37 M.26667g Left innominate 6418: 22 Mandible with tusks 6404 : 38 M.26667h Left rib head and cheek teeth 6404 : 39 M.26667i Right neuropophysis 6419: 15 Atlas vertebra of 3rdcervicalvertebra 6419: 16 Tusk alveoli 6404: 41 M.26667j Proximal and distal 6419: 17 Enamel fragment end of left radius 6421 a 2 Left M2 6404: 44 Immature skull 6421: 15 Molar fragment 6404: 45 Fragmented 6423: 13 Enamel fragment immature mandible 6423: 45 Enamel fragment 6404 : 46 M.26667k Distal end of right 6423: 56 Enamel fragment humerus **6424 : 79 Axis vertebra ** — Site unknown FROM GEBEL ZELTEN, LIBYA 291 III. SYSTEMATIC DESCRIPTIONS Order PROBOSCIDEA Illiger, 1811 Suborder DEINOTHERIOIDEA, Osborn, 1921 Family DEINOTHERIIDAE Bonaparte, 1845 DiaGnosis. Large herbivorous graviportal mammals. Dental formula $03 for the deciduous teeth and 2°:2:* for the permanent dentition; DP} and P3 with well developed external crest; DP¢ and Mi} trilophodont ; the remainder of the cheek teeth are bilophodont. Horizontal tooth replacement not developed so that all permanent teeth may be erupted at the same time (cf. elephantoids). Mandibular symphysis and lower tusks curved downwards so that the tusk tips are vertically or nearly vertically aligned. Skull low with deep rostral trough, retracted external nares, low orbit, inclined occiput, high occipital condyles, elongate paroccipital processes, and diploe. MIOCENE PLIOCENE PLEISTOCENE rd P. hobleyi > mn rad .@) > P. bavaricum Cc Pe) O me) m™m D. P. pentapotamige > (op) > D. indicum Fic. 2. Temporal and geographic distribution of deinotheriid taxa. 292 PRODEINOTHERIUM REMARKS. The family Deinotheriidae originated in Africa. The earliest and most primitive forms are from East African localities of early Miocene age. During the Burdigalian the early representatives of this family migrated into Asia and Europe. By the middle of the Miocene epoch, larger and more advanced repre- sentatives had appeared in Europe and Asia. The primitive and advanced forms coexisted until the end of the Miocene in Asia and the end of the Pontian in Europe. The more advanced forms lingered on until the middle Pliocene in Eurasia. Only the early and primitive form is known from the Neogene of Africa; a larger and more advanced form occurred in the Pleistocene but the two forms are not known to have coexisted. Deinotheriid species have been commonly allocated to a single genus—Deino- therrum—erected by Kaup in 1829. The species were defined upon geographic distribution and characters of the mandible and dentition. While the cheek teeth exhibit a great deal of variation in cusp morphology from specimen to specimen, these differences do not teadily lend themselves to species differentiation, especially when the several species are interpreted as belonging to a single genus. Moreover the concept of a single deinothere genus does not adequately express the fundamental division of the family into an early and primitive group of species (D. hobleyi, D. bavaricum and D. pentapotamiae) and a later, more advanced group (D. giganteum, D. indicum and D. bozasi). The recent discoveries of deinotheriid skull and post- cranial material, coupled with the re-examination of well documented specimens confirms that the two groups of deinotheres may be separated morphologically. It would appear therefore that division of the family into two genera is warranted. This idea was originally put forward by Ehik (1930) but has subsequently been ignored. On evidence currently available, however, Ehik’s premise is substantiated. The two genera are defined below. Genus Deinotherium Kaup, 1829 REVISED DIAGNOsIS. Large deinotheres. Dental formulae as for the family; tendency for the development of subsidiary styles on P3~4 and for simplification of the postmetaloph ornamentation of M2-3 when compared to Prodeinotherium. The skull rostrum not parallel to the mandibular symphysis and nearly horizontally aligned; rostral trough and external nares wide; preorbital swelling sited anteriorly on the rostrum; skull roof short and narrow at the temporal fossae; occiput slopes gently posteriorly; occipital condyles elevated above the level of the external auditory meatus; paroccipital processes very elongate. Postcranial skeleton with cursorial modifications to graviportal structure; scapular spine reduced with no acromion or metacromion; carpals and tarsals narrow with dolichopodous meta- podials exhibiting functional tetradactyly. FROM GEBEL ZELTEN, LIBYA 293 REMARKS. Most of the diagnostic characters of Deznotherium concern the skull and postcranial characters. The teeth, however, are the most commonly preserved remains. Teeth that may be attributed to Deinotherium are almost always larger than the equivalent teeth of Prodeinother1wm. Both genera exhibit a tendency to increase in size through geologic time, and in Europe there is an overlap in absolute size of the cheek teeth of the two genera. This is due to the greater numbers and greater longevity of deinothere specimens from that continent, but at any one horizon the teeth of Deinotheriuwm are always unmistakably larger than those of Prodeino- therium. On the basis of specimens examined and of those reported by Graf (1957), McInnes (1942) and Sahni & Tripathi (1957), the minimum parameters of Deino- therium teeth would appear to be as follows: Tooth Length (in mm) Breadth (in mm) Se 63 55 p4 59 62 M1 71 64 M? 70 74 M? 70 74 Ps 51 43 P, 61 48 Mi 74 54 Me 73 63 Ms 75 65 These parameters agree well with those noted by Bergounioux & Crouzel (1962a) who also define parameters for the deciduous teeth. TYPE SPECIES. Deinotherium giganteum Kaup, 1829 (including D. gigantissimum Stefanescu, 1897). OTHER RECOGNISED SPECIES: D. indicum Falconer, 1845; D. bozast Arambourg, 1934. A detailed synonymy of D. gigantewm was given by Graf (1957) and of D. indicum by Sahni & Tripathi (1957). Genus Prodeinotherium Ehik, 1930 REVISED DIAGNOSIS. Small deinotheres. Dental formulae as for the family; M2-3 with well defined postmetaloph ornamentation. Skull rostrum turned down parallel to the mandibular symphysis; rostral trough and external nares narrow; preorbital swelling close to orbit ; external nares anteriorly sited and nasal bones with anterior median projection; skull roof relatively longer and wider than in Deinotherium; occiput more vertically inclined; occipital condyles sited more ventrally than in Deinotherium and level with the Frankfurt Plane; paroccipital 294 PRODEINOTHERIUM processes short. Postcranial skeleton graviportally adapted; scapula with well defined spine and stout acromion and metacromion; tarsals and carpals narrow but not dolichopodous. ReMARKS. This genus was originally proposed by Ehik (1930) for a small species of deinothere from Kotyhaza, Hungary. The Kotyhaza specimens comprised a jaw, some teeth, anda few incomplete limbelements. These specimens were generally smaller than is normal in D. bavaricum but were subsequently attributed to this taxon without objection (Graf, 1957; Bergounioux & Crouzel, 1962a; etc.). I agree that the taxa bavaricum and hungaricum are probably conspecific but propose to allocate them to Prodeinotherium. Ehik’s generic taxon is therefore resurrected to encompass the small, early and primitive species of deinothere from the Neogene of Africa and Eurasia. The slight change in spelling from Prodinotherium (Ehik, 1930) to Prodeinotherium is, in view of Kaup’s spelling of the type genus of the family Deinotheriidae, considered to be a justifiee emendation according to the International Code of Zoological Nomenclature (Stoll et al, 1961) section 9g, article 50, para graph (c). Specimens that may be attributed to species of Prodeinotherium are observed to increase in size through geologic time. The maximum parameters for the permanent cheek teeth of Prodeinotherium, based on personal observation and data from Graf (1957), McInnes (1942), and Sahni & Tripathi (1957) are as follows: Tooth Length (in mm) Breadth (in mm) Ps 63 60 p4 61 62 Mt 73 69 M? 73 75 M3 69 72 P3 51 43 Py 60 52 My 73 53 M2 75 70 M3 81 68 Similar figures are given by Bergounioux & Crouzel (1962a) who also define size limits for the deciduous teeth. TYPE SPECIES. Pyrodeinotherium bavaricum (von Meyer), 1831. OTHER RECOGNIZED SPECIES. P. pentapotamiae (Falconer), 1868; P. hobleyi (Andrews), IgII. A detailed synonymy of P. bavaricum was given by Graf (1957) and of P. pent- apotanuae by Sahni & Tripathi (1957). Skull. Dentition. Skeleton. FROM GEBEL ZELTEN, LIBYA 295 TABLE 2 Comparison of Prodeinotherium and Deinotherium Prodeinotherium rostrum ventrally flexed rostrum narrow and deep preorbital swelling close to external nares external nares almost as deep as wide external nares surmounted by median projection of nasals orbit above P8 skull roof nearly flat and inclined anteriorly occiput steeply inclined paroccipital processes longer than in elephantoids occipital condyles cut ventrally by Frankfurt Plane P3-4 usually lack mesostyles M2-3 with well developed postmetaloph ornamentation tusks nearly vertical scapula—stout spine, metacromion and acromion; supraspinous fossa _ well developed humerus—lateral epicondyle tapers proximally radius—medial half of head larger than lateral half; distal epiphysis more massive than in Mastodon angustidens lunar—radial facet covers most of proximal surface; magnum facet is concavo-convex cuneiform—posterolateral process of similar length to that of Elephas but does not articulate with unciform; unciform facet is concavo-convex unciform—cuneiform facet roughly tri- angular; largest distal facet is for McV magnum—proximal surface has large posteromedial projection metacarpals—laterally compressed but of similar size to Mastodon angustidens manus more plantigrade femur—similar length to Elephas maxi- mus astragalus—tibial facet equant and convex; prominent posteromedial pro- cess Deinotherium rostrum almost straight rostrum wide and shallow preorbital swelling more anteriorly sited external nares much wider than deep no median projection of nasals orbit above P4 skull roof shorter and narrower at the temporal fossae occiput gently inclined paroccipital processes longer than in Prodeinotherium occipital condyles Frankfurt Plane P3-4 often possess mesostyles M2-3 with reduced postmetaloph orna- mentation tusks longer and may be recurved beneath symphysis scapula—reduced spine and no meta- cromion or acromion; supraspinous fossa greatly reduced humerus—lateral epicondyle does not taper proximally radius—medial half of head corres- pondingly larger; distal epiphysis more massive than in Prodeinotherium lunar—radial facet extends less far posteriorly; magnum facet almost flat elevated above cuneiform—posteriolateral process is relatively longer and more ventrally inclined than in Prodeinotherium; unci- form facet is biconcave unciform—cuneiform facet extends far- ther posterolaterally and distally, and tapers more abruptly posteriorly; largest distal facet is for McIV magnum—posteriomedial projection is less pronounced metacarpals—more compressed lateral- ly and distinctly more elongate than in Prodeinotherium manus more digitigrade femur—proportionately 30% than in E. maximus astragalus—tibial facet equant but nearly flat posteromedial process more reduced than in Prodeinotherium shorter 296 PRODEINOTHERIUM Teeth are the most common deinothere remains, but although there is a great deal of variation in minor topographic features on the cheek teeth the basic tooth morphology is conservative throughout the known history of the group. Indeed, when only a single genus—Deinotheriwm—was accepted, tooth morphology was of far less importance for specific differentiation than absolute size of the teeth or geographic distribution of the specimens. In effect only two distinct species could be recognized—D. gigantewm and D. bavaricum. However, with these two taxa separated by generic rank, tooth morphology assumes greater diagnostic importance. It is entirely possible that detailed revision of known material will furnish evidence for distinguishing different lineages within the two genera incorporating variations in cheek tooth morphology. Such a revision is, however, beyond the scope of this paper. Only two deinothere species are currently recognized from fossil mammal sites in Africa—Prodeinotherium hobleyi of the Neogene and Deinotherium bozasi of the Quaternary. The Gebel Zelten deinothere teeth are more advanced than specimens of P. hobleyi from the earliest Miocene of East Africa but more primitive than those from Ngorora (Pliocene) in Kenya. The observed differences are insufficient at present to warrant the erection of a separate taxon for the Libyan material. Prodeinotherium hobleyi (Andrews), 1911 1911 Deinotherium hobleyi Andrews: 943. 1919 Deinotherium cuviert Kaup; Brives: go. 1957 Deinotherium bavaricum v Meyer; Graf: 152. 1967 Deinotherium cuviert Kaup; Savage: 263. Skull (Text-figs 3-6; Plates 1-3) MATERIAL REFERRED. 6404 : 14, fragmentary adult skull with only the dentition, right premaxilla, left jugal and occipital condyles well preserved; 6404 : 44, complete but slightly distorted immature skull; M.26665, virtually complete skull but with zygomatic arches, squamosal, auditory and pterygoid regions poorly preserved; 6418 : 4 and 6418 : 19, incomplete and uncollected skulls. DescriPTION. The skull of Prodeinotherium hobleyi from Gebel Zelten is elongate and low, the length of the cranium being twice that of the rostrum. The skull is characterized by a relatively flat skull roof which ascends gently to the lambdoidal crest, a small orbit that is open posteriorly, a broad and downturned rostrum that is surmounted by massive preorbital swellings, high occipital condyles, and a back- wards sloping occipital region. The rostrum is appreciably narrower than in Deinotherium giganteum and is excavated to form a deep rostral trough anterior to the external nares. The occipital region is very wide owing to the well developed lateral wings of the squamosal bones. Much of the skull is formed from cancellous bone, and where the surface bone has been eroded diplée similar to those of ele- phantoid skulls may be detected. Key to letters used in text-figures 3, 4, 5 & 6. b = auditory bulla en = externalnares f = frontal fo = foramen ovale fr = foramen rotundum ga = glenoid articulation surface in = internal nares io = infraorbital foramen lc = lamboidal crest ls = lateral wing of squamosal m = maxilla n = nasal ns = nasal sinus o = orbit FROM GEBEL ZELTEN, LIBYA —) oc = occipital condyle p = parietal pf = posterior palatine foramen pgf = postglenoid fossa plf = posterior lacerate foramen pm = premaxilla po = postorbital process of frontal pp = paroccipital process ps = preorbital swelling of maxilla pt = pterygoid process rt = rostral trough Ss = squamosal tf = temporal fossa zm = zygomatic process of maxilla zs = zygomatic process of squamosal Fic. 3. Prodeinotherium hobleyi skull, anterior view. Scale = 5 cm. 297 98 tf Fic. 4. PRODEINOTHERIUM Prodeinotherium hobleyi skull, dorsal view. Scale = 5 cm. FROM GEBEL ZELTEN, LIBYA Fic. 5. Prodeinotherium hobleyi skull, ventral view. Scale = 5 cm. 299 300 PRODEINOTHERIUM Premaxilla. The dorsal surface of the deep rostral trough (text-figs 4, 5) anterior to the external nares is lined by the premaxillae. The anterior portion of the premaxilla forms a small flange projecting anteriorly from the upstanding rugose walls of the rostral trough (Pl. 1a). The premaxillary-maxillary suture extends from the lateral edge of this flange along the medial edge of the wall of the rostral trough until it meets the anterior wing of the nasal (text-figs 3, 4). The suture has completely fused in skull M.26665 but the premaxillary-maxillary contact is denoted by the junction of smooth bone, typical of the medial walls and floor of the trough, with the rugose bone of the lateral walls and dorsal surface. The suture is still visible in skull 6404 : 44. The posterior limit of the premaxilla beneath the external nares is not clear owing to the fusion of sutures in M.26665 and the cracked nature of the bone in 6404 : 44. Perhaps the limit is marked by the anterior edge of the cavernous nasal sinus (text-fig. 4, Pl. rb) that occurs on either side of the external nares. Viewed from the dorsal aspect, each premaxilla is roughly triangular with the posterior edge of the bone forming the shortest side. In transverse section the distal part of the premaxilla is L-shaped, the ascending arm lining the medial wall of the rostral trough while the body forms the floor. Medial to the preorbital swelling of the maxilla (text-figs 3, 4, 5; Pls. 1-2), the ascending process of the premaxilla bulges medially in front of the nasal sinus, and here the surface of the bone is curved sigmoidally. The facial process of the premaxilla ascends to contact both the nasal and the anteriormost prolongation of the frontal bone (text-figs 3, 4), thereby ex- cluding the maxilla from contact with the nasal. Ventrally the premaxilla is visible only at the anterior third of the rostrum where the premaxillary flanges appear above the diverging tips of the maxillae (text-fig. 5; Pl. 3b). The flanges of the premaxillae converge ventrally. Maxilla. This is the largest bone of the facial region, extending from the dental alveoli up to the level of the external nares, and from the anterior tip of the rostrum to behind the dentition. The facial process of the maxilla is contiguous with the frontal, lacrimal and premaxilla. It continues forwards and downwards from the orbit as the external wall of the rostral trough. Just anterior to the frontal-maxillary suture the dorsal and dorsolateral parts of the rostral walls are swollen in skull M.26665 to form a massive preorbital swelling. In 6404 : 44 the swelling is less pronounced but this portion of the maxilla is considerably elevated above the remainder of the dorsal rim of the rostral trough. The infraorbital foramen (Pl. 2a and b) is large and the anterior opening lies above the anterior edge of the third premolar. In skull M.26665 two deep grooves extend anteriorly from the infra- orbital foramen. One runs for a short distance down the lateral surface of the rostrum; the other extends parallel to the dentition to terminate on the dorsal surface of the rostral trough immediately in front of the preorbital swelling. These grooves mark the proboscideal branches of cranial nerves V and VII. In skull 6404 : 44 only the former groove is well developed. The jugal process of the maxilla is stout, expanding posteriorly to define the posteroventral limit of the orbit (Pl. 2a). The posterior opening of the infraorbital canal is sited above the base of the jugal process of the maxilla and beneath the postorbital process of the frontal (Pl. 2b). FROM GEBEL ZELTEN, LIBYA 301 In ventral aspect the maxillae diverge behind the anterior third of the rostrum, revealing the margins of the premaxilla. Behind the anterior third of the rostrum the intermaxillary suture extends posteriorly down the midline of the palate as a cristiform ridge (Pl. 3b). The maxillary-palatine suture is fused in skull M.26665 but is presumed to have been level with the posterior root of the second molar and to have incorporated the paired posterior palatine foramina. Shallow grooves extend anteriorly from these foramina until level with the fourth premolars. The lingual border of the alveolar region is continued forward from the anterior root of the premolar as a broad, rounded ridge, converging on its fellow towards the tip of the rostrum. On the medial side of this ridge a very weak crest extends anteriorly to meet the ventral exposed portion of the premaxilla. The plate is vaulted in the vicinity of the premolars and narrows as it continues forwards beneath the rostrum. In the region of the molars the palate is convex. Nasal. In skull M.26665 the nasal-frontal suture is fused but there is a pro- nounced break of slope some 7 cms behind the external nares that probably indicates the junction of these bones. The surface of the bone immediately above and behind the external nares is much smoother than the adjoining bone. In 6404 : 44 the lateral wing of the nasal extends laterally until level with the front of the orbit and the suture with the premaxilla is well defined. In dorsal plan the nasals are crescentic (text-fig. 4). In M.26665 the nasals continue anteriorly at the intranasal suture to form a well developed median projection (Pl. Ia and b) though this is less well developed than that of the extant elephants. In 6404 : 44 the anterior portion of the intra-nasal suture is marked by an indentation rather than by a projection; this is attributed to the relative immaturity of the specimen. Apparently the nasal bones do not contribute to the cavernous nasal sinuses on either side of the external nares. Lacrimal. Due to fusion of the sutures and to incomplete preservation the lacrimal is not well delineated in skull M.26665 but in 6404 : 44 the lacrimal is seen to form the anteroventral quadrant of the rim of the orbit. The lacrimal bone is triangular with the base of the triangle elevated a short distance above the infra- orbital canal. From the centre of the ventral half of the bone a lacrimal foramen opens into a lacrimal canal that passes outwards to the anteroventral rim of the orbit. Above the lacrimal foramen is a large traction epiphysis. The lacrimal is contiguous only with the frontal and maxilla and does not form part of the infra- orbital swelling of the maxilla (cf. Andrews, 1921; 530). Jugal. This bone is missing from skull M.26665 but is known from 6404 : 44 and from the incomplete skull 6404 : 14. The jugal joins the zygomatic process of the maxilla at the posteroventral edge of the orbit, at which point the zygomatic process of the maxilla attains its greatest height. From the vertical jugal-maxillary suture the jugal increases in height posteriorly while simultaneously becoming flattened from triangular to oval in transverse section. The posterior portion of the jugal is overlapped by the zygomatic process of the squamosal. The posterior end of the jugal forms a smooth convex curve and gives off a small conical process from the posterior ventral surface. A jugal from skull 6404 : 44 is less massive and lacks the conical process on the posterior ventral surface. 202 PRODEINOTHERIUM Palatine. This is fused to the palatine process of the maxilla level with the posterior root of M? and extends for some distance behind M3. The palatine is pierced by two asymmetrically placed posterior palatine foramina. In M.26665 the anteriormost part of the palatonarial border is sited some way behind M? and is gently curved with the lateral, more posterior, portions missing. The internal narial opening is almost circular. Vomer. The vomer extends back behind the level of the anterior part of the palatonarial border and forms a V-shaped crest along the roof of the internal nares. Immediately above the posterior end of the vomer a small chip of bone is missing in skull M.26665 thus revealing part of the cavernous ethmoidal fossa. Frontal. In M.26665 the frontal bones occupy just over half of the skull roof (text-fig. 4). The nasal-frontal suture is fused but the anterior edge of the frontals may be determined as above. Part of the frontal-parietal suture is revealed in the centre of the skull roof of M.26665 as a sinuous interdigitating groove but this cannot be traced laterally. The frontals are slightly domed convexly in the midline and over the large postorbital processes. There are no supratemporal ridges. Both the frontal and parietal portions of the skull roof extend ventrally to contribute to the temporal fossa. The postorbital processes of the frontals (text-fig. 4) are blunt protruberances from each of which a ridge descends ventrally and posteriorly to continue as the free end of the alisphenoid. In front of these processes the frontals demarcate the upper rim of the orbit. Parietal. The parietal bones form the posterior part of the skull roof. They increase in width antetiorly from the lambdoidal crest. The lambdoidal crest forms the most elevated portion of the skull roof and strongly delineates the latter from the occipital region. The dorsal surface of the parietals is slightly concave in transverse section and there is no sagittal crest. Fusion of the sutures on the posterior half of the cranium makes the boundary of the parietals indistinct. Even in the immature skull (6404 : 44) the junctions of the parietals, squamosals, alisphenoids and frontals are not clear. Beneath the sharply delineated rim of the temporal fossa the bone in M.26665 is sculpted by deep and wide grooves denoting the point of attachment of the temporal muscle. Ventral to the sculpted surface the bone is much cracked and broken. In 6404 : 44 the walls of the temporal fossae are better preserved but the bone shows no temporal muscle scars. No trace of an interparietal may be detected. Supraoccipital. This contributes to the central and dorsal regions of the occiput (Pl. 3a) and is concave both sagittally and transversely. The perimeter of this bone is fused but it must have been roughly triangular; whether or not it formed the dorsal border of the foramen magnum cannot be determined. The dorsal edge contributes to the posterior side of the central portion of the lambdoidal crest. The lateral junction with the squamosal may be indicated by a pronounced ridge that extends from the paroccipital process to the lamboidal crest. The supraoccipital is excavated in the midline by two deep triangular pits that form fossae for the nuchal ligament. The dorsal edge of the nuchal fossa lies some 5 cms below the lambdoidal crest; the ventral edge some 8 cms above the foramen magnum. As in elephantoid skulls, the two halves of the nuchal fossa are separated by a narrow ridge of bone. The floor FROM GEBEL ZELTEN, LIBYA 303 of each pit is comprised of sculpted bone, as is that region between the top of the nuchal fossa and the lambdoidal crest. Exoccipital. The paired exoccipital bones form the ventrolateral regions of the occiput and support very large and prominent occipital condyles. From the dorsolateral edge of the nuchal fossa of skull M.26665 a prominent crest curves ventrally and laterally to terminate on the lateral edge of the paroccipital process. This ridge is interpreted as separating the exoccipital from the lateral wing of the squamosal and is continuous with the lateral limit of the supraoccipital. The condyles are triangular and converge as they taper ventrally. The condyles are limited ventrally by a deep groove (the condyloid fossa of Palmer, 1924) that is also present, though less prominent, on the dorsal edge. The condyles are most elevated from the rest of the exoccipital bone at their anteromedial edge at which point they protrude almost at right angles from the plane of the occiput. The paroccipital processes are incomplete on M.26665 but on 6404 : 44 are seen to extend for some distance below the base of the occipital condyles, merging laterally with the ventral edge of the lateral wing of the squamosal. Basioccipital. Fusion of the sutures and incomplete preservation prevents delinea- tion of the periphery of this bone in M.26665. The basioccipital descends steeply from the foramen magnum and becomes progressively more narrow anteriorly (text-fig. 5; Pl. 3b). The posterior part is deeply notched by the foramen magnum and extends laterally to the medial edges of the occipital condyles. Anteriorly and ventrally the body is keeled. The anteroventral junction with the basisphenoid is probably level with the glenoid articulation surface. The basisphenoid passes uninterruptedly from the basioccipital but the keel becomes less pronounced anteriorly and disappears before reaching the internal narial opening. The basicranium of 6404 : 44 has been distorted by crushing and the ventral surface is now parallel with the left lateral side. Slightly anterior to, and on either side of, the notch excavated by the foramen magnum is a large boss. This probably served for the insertion of the rectus anticus muscles. The right boss is level with, but somewhat below, the paroccipital process. The left boss is less well preserved. From this point the basioccipital narrows appreciably as it continues anteriorly. Some 5 cms in front of the anterior edge of the foramen magnum a pronounced keel extends down the midline. This keel extends for at least 9 cms until level with the eustachian opening of the auditory bulla but in front of this point the bone is too badly displaced to allow further observation of the keel. On either side of the posterior end of the keel are excavated large fossae for insertion of the rectus anticus major muscles. An elongate notch, the posterior lacerate foramen, is sited anterior to the paroccipital process and lateral to the basioccipital at the posterior end of the keel. Sphenoid. The junctions of the various component parts of this bone with each other and with contiguous elements are for the most part indistinct. The anterior boundary of the alisphenoid forms part of the prominent cristiform outer wall of a deep groove that extends from the pterygoid process to the postorbital process of the frontal. The optic foramen, that probably perforated the orbito- sphenoid, lies some distance above and anterior to the slit-like anterior lacerate foramen, and the orbitosphenoid-alisphenoid suture undoubtedly lay somewhere 304 PRODEINOTHERIUM between the two. Both foramina lie in a deep groove beneath the alisphenoid- frontal ridge. From the optic foramen a shallow groove extends forwards to the orbit, marking the course of the optic nerve. The foramen rotundum lies posterior to the anterior lacerate foramen and opens from the alisphenoid flap into a large alisphenoid canal. The mandibular branch of the fifth cranial nerve emerged from the posterior opening of the alisphenoid canal while the facial and maxillary branches emerged from the anterior opening. The position of the maxillary branch is marked by a deep groove, partly covered by a flap of bone, that extends towards the ptery- goid process. In skull M.26665 this groove extends almost vertically (Pls. 2a and b) but in 6404 : 44 it passes anteriorly at an angle of 45°. Behind the posterior exit of the alisphenoid canal, and above the anterior portion of the auditory bulla, is a large foramen ovale (text-fig. 6). The pterygoid region is very poorly preserved in skull M.26665. Ventrally and medially it passes uninterruptedly into the presphenoid and posteriorly into the alisphenoid. Nearly all of the ventral part that, together with the palatine, forms the lateral border of the internal nares is missing. The posterior surface of the pterygoid forms a smooth crest that. passes laterally to the anterior tip of the petrosal. This region is also poorly preserved in skull 6404 : 44. Squamosal. This is one of the largest bones of the skull and forms the postero- ventral wall of the temporal fossa. It is only poorly preserved in M.26665 and in many places the surface bone is missing to reveal diplée beneath. The squamosal continues anteriorly and laterally from the supraoccipital to project as a vast lateral wing forming the widest part of the skull (text-fig. 4; Pl. rb). The external edge of this wing passes uninterruptedly from the lambdoidal crest into the paroccipital process of the exoccipital. In lateral aspect the junction of the squamosal with both the alisphenoid and the parietal is indistinct. The zygomatic process of the squamosal is less stout than that of the maxilla, being almost quadrate in transverse section near the point of origin but becoming more slender as it extends forward Fic. 6. Pvrodeinotherium hobleyi right auditory bulla. Scale = 5 cm. FROM GEBEL ZELTEN, LIBYA 305 to overlap the jugal. The auditory meatus opens onto the external surface of the lateral wing immediately posterior to, and slightly above, the zygomatic process. In M.26665 this region is poorly preserved but in 6404 : 44 the passage of the meatus is clearly seen. The passage of the meatus is gently inclined medially. The glenoid articulation surface is convex and elongate transversely. There is no postglenoid process and behind the glenoid surface there is a shallow transverse channel forming the floor of the auditory meatus as in Elephas. Tympanic bulla. The tympanic bullae are incompletely preserved in M.26665 but are seen to be relatively larger than in the skull of D. gigantewm from Eppeslheim, Germany. Much of the surface bone on the Gebel Zelten specimen is missing and sufficient detail is preserved to show only the eustachian opening below and behind the foramen ovale. In skull 6404 : 44 both bullae are present but the left bulla is incomplete and crushed and the right bulla (text-fig. 6) is incomplete. The bullae are bounded medially by the basioccipital and basisphenoid (contributing in part to the posterior lacerate foramen), posteriorly by the paroccipital process, and dorsally by the squamosal and alisphenoid. The bulla projects ventrolaterally as a semi- circle of bone. In dorsal aspect it may be divided into two unequal portions—a posterior concave surface and an anterior (slightly larger) convex area. The regions are separated by a ridge that runs posteriorly and laterally from the foramen ovale to merge with the lateral perimeter level with the posterior lacerate foramen. The concave area thus enclosed presents an irregular surface pitted with numerous small foramina. The anterior portion of the dorsal surface is convex and consists mainly of smooth bone although it is pierced by several small foramina and canals near the anterolateral edge. The eustachian opening occurs at the anteromedial tip. On skull 6404 : 44 the ventral surface of both bullae is incomplete but sufficient bone is preserved to permit a composite description. The ventral surface is divided sagittally into two unequal portions by a crest that extends from the paroccipital process to the anterolateral margin of the bulla. The portion of the bulla lateral to the crest is the smaller. The ridge is most pronounced posteriorly where it separates the posterior lacerate foramen from the large stylomastoid foramen, and becomes progressively less distinct anteriorly. The bone anterior to the stylo- mastoid foramen and lateral to the ridge is concave transversely and heavily sculpted. This part of the bulla is equivalent to the tympanohyal pit. Medial to the crest the surface of the bulla is smooth but cut by transverse shallow grooves. This portion of the left bulla is crushed and fragmented; on the right bulla the surface bone is missing from the anteromedial region to reveal the cavernous tympanum. In comparison with that of the extant elephants, the tympanic bulla of Prodeino- therium is less elongate sagittally and wider transversely with a larger stylomastoid foramen. OTHER DEINOTHERIID SKULLS. Prodeinotherium pentapotamiae. An incomplete skull of P. pentapotanuae was collected in 1912 by Pilgrim from the Lower Chinji Sandstone, four miles from Chinji in the Salt Range of India. Two portions of the skull are preserved—a palate with cheek teeth (I.M. A.460) and a basicranium (I.M. A.461). These are of similar size to those of the Gebel Zelten skulls and were 306 PRODEINOTHERIUM described by Palmer (1924). The skull foramina that Palmer identified are sited similarly to those of the Gebel Zelten skulls. Only the base of the vertical tapering pillar that Palmer identified as the hyoid is still preserved but the basal tympanohyal portion would have fitted well into the elongate and narrow tympanohyal pit anterior to the stylomastoid foramen in Gebel Zelten skull 6404: 44. Unlike Palmer, I do not believe that the right paroccipital process of the Indian skull can be far from its original position. Palmer had correctly identified the tympanic bullae, unlike Andrews (1921), but these are now only poorly preserved on the Indian specimen. Eppelsheim skull. The skull of Deinotherium gigantewm from Eppelsheim, Germany, was originally described by Klipstein and Kaup (1836) and the most recent description of this specimen was given by Andrews (1921). In the light of the information provided by the Gebel Zelten Prodeinotherium material it is now necessary to revise some of Andrews’ interpretations of the Epplesheim specimen. For example, on page 527 Andrews remarks: . de Blainville suggests that possibly the occipital surface had been crushed down towards the floor of the skull. Careful examination of the specimen, however, does not seem to support this idea and it seems probable that . . . it represents the condition of the living animal.’ TABLE 3 Prodeinotherium hobleyi skull measurements Parameter M.26665 6404: 44 Total length of skull . ue e.O4 73 Width anterior tip of rostrum. Be Bees IO-l Length rostrum to external nares eae 34 Width skull at external nares : 56 38 Width rostral trough at external nares. 17:8 118 Length external nares to lambdoidal crest 40 31 Minimum width skull roof . , ae Ae 28 Width lambdoidal crest ; 59°6 35 Height foramen magnum to lambdoidal crest : é . A 2554 = Height nuchal fossa : ; : 2» 12-3 12°6 Width nuchal fossa. : ; E59 EL Width occipital condyles. : D272 — Height foramen magnum . : ; 6-4 Width foramen magnum. Fi] — Height molar alveoli to Jambdoidal crest 41-9 oi Width glenoid condyles : 13°7 — Length palatonarial border to rostrum tip 56 84 Width palate at P? ; : : 73 — Width palate at M!_ .. : : if OTEO — Width palate at M3. : : ; 83 — FROM GEBEL ZELTEN, LIBYA 307 The cracked condition of the bone and asymmetry of the left and right lateral wings of the squamosals reveal that the Eppelsheim skull has undergone at least a modicum of postmortem distortion, either during burial or after excavation. Differences between the cast of the ventral side and the specimen itself indicate that some post-excavation restoration has been necessary. Indeed, much of the squamosal region is now plaster of paris. Certain anomalous features of the Eppelsheim skull are discussed later and indicate that restoration has not always been skilful or accurate. On page 528 Andrews stated ‘In Deinotheriwm no development of cellular bone seems to have taken place.’ This is clearly wrong. In the Eppelsheim skull, as in the Gebel Zelten skulls, portions of the surface bone are missing to reveal cellular bone beneath. This phenomenon was noted in de Blainville’s earlier description of the skull (de Blainville, 1837) and was confirmed by Palmer (1924) in his descrip- tion of the skull of P. pentapotamiae. Skull 6404: 14 from Gebel Zelten was unfortunately shattered in transit to Bristol but the fragments salvaged leave no doubt that much of the skull was composed of cellular bone. On page 529 Andrews noted that ‘in front of the orbit the maxilla forms a great mass of bone terminating in a rugose, somewhat concave surface for the attachment of muscles’. These concavities are preservation anomalies and not original mor- phological features. That on the left side is much smaller than the right concavity. In corollary the preorbital swellings of the Eppelsheim skull are unequal, that on the left being much larger. During restoration a large piece of bone has been cemented on to the anterior part of the left swelling thus increasing its total size. This piece of bone is the same length as the concavity on the right side and, if it were to be transferred there, the preorbital swellings would be (a) symmetrical, (b) of similar size and (c) convex. Further, on page 530, Andrews declared: ‘The point of union of the premaxilla (with the maxilla) is probably marked by the sudden narrowing of the snout about 40 cms in front of the orbit.’ This sudden narrowing marks the anterior termination of the preorbital swelling. The ‘suture’ to which Andrews refers in the next sentence occurs on both the dorsal and ventral surface of the rostrum and bisects both the premaxilla and maxilla. It appears likely that the anterior part of the rostrum was once detached and the ‘suture’ marks the point of breakage. The actual suture between the premaxilla and maxilla is well seen on Gebel Zelten skull 6404 : 44 and there is no reason to believe that the geometry of these bones differed appreciably on the Epplesheim skull. The anterodorsal surface of the external nares appears to be complete and is similar in morphology to the equivalent portion of a skull fragment of Deinotherium giganteum from Palencia, Spain. There is no anterior median process in contrast to the condition in Gebel Zelten skull M.26665. The lacrimal bone cannot be observed in the Eppelsheim skull but was more likely to have occupied a similar position to that in 6404 : 44 than to have formed part of the preorbital swelling (cf. Andrews, Ig2I : 530), especially as the latter feature is found farther forward 308 PRODEINOTHERIUM than in the Gebel Zelten skulls. The ‘irregularly arranged pits’ at the anterior end of the premaxilla (Andrews, Ig2I : 531) are not, as Andrews suggested, rudimentary tusk alveoli but diploe revealed by erosion of the external surface of the bone at this point. Any comparison made between the occipital regions of the Eppelsheim and Gebel Zelten skulls is liable to be suspect for the roof of the Eppelsheim specimen is incomplete posteriorly. The lambdoidal crest appears to be missing and much of the bone between the foramen magnum and the lambdoidal crest is either absent or distorted (cf. Andrews, 1921 : 527). The shape of the occipital surface of the Eppelsheim skull appears to be more oval than in the Gebel Zelten skulls and the nuchal fossa is shaped differently. The nuchal fossa of the Eppelsheim skull may be divided into two portions—an anterior deep and wide oval pit (long axis trans- versely orientated) and a posterior triangular trough that is shallower and tapers ventrally. The latter trough may be all that remains of the true nuchal fossa on the Eppelsheim skull. The ‘large post-tympanic flanges’ (Andrews, Ig2I : 527) are equivalent to the paroccipital processes of other mammals and are, as Andrews described, formed partly of the exoccipitals and partly of the squamosals. The paroccipital processes are smaller in the Gebel Zelten skulls. The ‘tongue-like process of bone’ (Andrews, 1921 : 527) that is attached to the left paroccipital process was incorrectly identified by Andrews as the paroceipital process itself. This piece of bone is seen only on the left side of the skull and is more likely to represent part of the hyoid bone (cf. Palmer, 1924: 4). The large convex bulbs that Andrews (1921 : 530; fig. 3) interpreted as tympanic bullae are seen also on Gebel Zelten skull 6404 : 44 where I have interpreted them as bosses for the insertion of the rectus anticus muscles. The tympanic bullae of the Epplesheim skull are more likely to be represented by a strong crest of bone that terminates posteriorly at the hyoid and extends forward apparently to merge with the pterygoid flange. A petrosal of D. gigantewm was described by Claudius (1865) but the original specimen is now lost and the accuracy of its identi- fication must be regarded as questionable (pers. comm. H. Tobein). Munich skull. Stromer (1938) described the remains of a skull, mandible and partial skeleton of D. gigantewm from the Sarmatian Flinzsande of Munich. Un- fortunately all except the cast of the palate were subsequently destroyed and only Stromer’s description and diagrams are now available for study. The skull was originally incomplete, lacking the lower part of the occiput, the basicranium, the ventral walls of the temporal fossa and part of the skull roof. However the frag- ments that were collected are of importance in that they are less distorted than some portions of the Eppelsheim skull. The rostral flanges of the premaxilla and maxilla are separated in the midline as in the Eppelsheim skull, and are flexed at a similar angle, i.e. less steeply than Prodeinotherium. The preorbital swellings were either not developed or have not been preserved. The surface of the skull roof was relatively flat and, as restored by Stromer, may have been more horizontal posteriorly than in the Gebel Zelten skulls. Only the upper portion of the occiput was preserved and was described as FROM GEBEL ZELTEN, LIBYA 309 strongly inclined posteriorly, though this cannot be detected from the illustrations given by Stromer. The lateral wing of the squamosal was more vertically aligned than in the Eppelsheim skull and strongly resembles that of Prodeinotherium, indicating that the Eppelsheim skull was badly deformed in this region. Palencia skull. The skull roof and palate of D. giganteum from Castrillo de Villa- vega, Palencia, Spain, is housed in the Madrid Natural History Museum. The nasal bones are represented but are incompletely preserved on the right side. They surmount an external narial opening that is greatly compressed dorsoventrally. The lateral sinuses, which are large on the Gebel Zelten Prodeinotherium skulls, are much less prominent in the Palencia specimen. The nasal-frontal suture is fused but the nasals are demarcated as a flat surface projecting anteriorly at an angle from the remainder of the skull roof. The intra-nasal suture is fused and there is no anterior median projection of the nasals over the external nares. The skull roof forms a hollow behind the posterior edge of the nasals and in front of the lambdoidal crest. The lambdoidal crest is incomplete on the right side but on the left extends down the lateral wing of the squamosal to contribute to the widest part of the skull. The postorbital process of the frontal is very large and the anterior portion is located in front of the external nares. Behind the postorbital process the lateral surface of the temporal fossa is strongly delineated from the skull roof, the angle between the two surfaces being in the order of 45°. Only the dorsal edge of the occipital surface is preserved and this had been in- completely prepared. It is not therefore possible to determine at present whether the occiput was inclined as in the Eppelsheim skull or the Gebel Zelten specimens. The nuchal fossa is incomplete. The shape indicated by the upper part of the fossa is approximately triangular and the dorsal edge of the fossa sited only some 5 cms below the lambdoidal crest. Although the orbit is not represented on this specimen the position of the postorbital processes suggests that the orbit was sited in advance of the external nares in contrast to the situation in the Gebel Zelten specimens. Discussion. Comparison of the skulls of Prodeinotherium and Deinotherium, as represented by the Gebel Zelten, Eppelsheim and Palencia specimens, yields evidence of both positive and negative allometric changes. Using the overall skull length as a basis for comparison, parameters that remain the same in both genera are skull width at the external nares, width at the lambdoid crest, and length of the rostrum. The only major increase in width is in the rostral trough of Deinotheriwm and this is exhibited in both the Palencia and Eppelsheim skulls. In both these skulls the external nares are also sited farther posteriorly than in Prodeinotherium and are considerably less deep than wide—converging on the elephantoids in both position and shape. The length and width of the skull roof are proportionately smaller in Deinotherium than in Prodeinotherium. The facial/cranial region ratio remains constant and the decrease in skull roof length in Deinotherium is directly propor- tional to the increase in area of the occipital region due to the anterior inclination of the latter. The shorter skull roof of Deinotheriwm is also correspondingly narrower. This is a compensatory feature to permit a large area of origin of the temporal muscles despite a shorter skull roof. 310 PRODEINOTHERIUM The majority of the differences between the cranial morphology of Deinotherium and Prodeinotherium involve only relatively minor allometric changes and some of these are foreshadowed by the differences between the immature (6404 : 44) and mature (M.26665) Prodeinotherium skulls. Thus the rostrum is more steeply down- turned in M.26665 than in Deinotheriuwm and the rostral flexure of 6404 : 44 is even more pronounced. Decrease in rostrum curvature in Deinotherium, together with increase of rostral width and retraction of the external nares, suggests the develop- ment of a more powerful and effective proboscis in the later genus. The length measurements of 6404 : 44 and M.26665 are very similar but the skull width para- meters are relatively smaller in the immature specimen. In Deinotherium the elevation of the occipital condyles, elongation of the paroc- cipital processes and increase in inclination of the occiput are associated with functional changes enhancing the downward movement of the head (and thereby the tusks). The greater inclination of the occiput has contributed to a decrease in the length of the skull roof, although this is also affected by the more posterior siting of the external nares. The shorter skull roof limits the potential area of attachment of the temporal muscles but this is compensated for by excavation of the temporal fossa and decreased width of the skull roof. In this way major changes in cranial morphology in the Deinotheriidae are associated with proboscis development and with increase in potential movement of the skull on the neck, both factors suggesting changes in feeding methods and habits. TABLE 4 Deinotherium skull measurements Parameter Eppelsheim Munich Palencia Total skull length 109 ad = Width anterior tip rostrum 34°7 — — Length rostrum to external nares 53°5 — — Width skull at external nares 79 — 69 Width rostral trough at external nares 49°8 as 32* Length external nares to lambdoidal crest 29 33 34 Minimum width skull roof 48 40 45 Width at lambdoidal crest 98 — — Height lambdoidal crest to foramen magnum 47°73 — — Height lambdoidal crest to nuchal fossa 9°4 — — Height nuchal fossa 19°5 — — Width nuchal fossa ice) ~- — Width occipital condyles 30°5 — —_ Height foramen magnum 5°6 — —= Width foramen magnum 14-1 -— — Height lambdoidal crest to molar alveoli 50* — — Width glenoid condyles 27 21 — Length palatonarial border to rostrum tip 68 —- — Width palate at P38 15°8 77 a Width palate at M! 18-7 = —— Width palate at M3 05 10 -- FROM GEBEL ZELTEN, LIBYA 311 Mandible (Plates 4-5) MATERIAL REFERRED. 6404 : 13, incomplete horizontal rami with tusk fragments and cheek teeth; 6404 : 45, fragmented immature mandible with tusks and dentition ; 6412 : 10, tuskless mandible with worn cheek teeth; B.M.(N.H.)M.26666, mandible with tusks and cheek teeth; 6419 : 16, symphysis fragment with tusk alveoli. DESCRIPTION. Four nearly complete mandibles and a number of mandibular fragments have been collected from various sites at Gebel Zelten. None of the specimens is complete but sufficient information can be derived from the specimens to provide the following composite description. The horizontal ramus is sigmoid, rising gently in front of Mg and flexing ventrally at M, until at the tusk alveolus it subtends an angle of 115° to the alveolar border. The cheek teeth are erupted in almost parallel rows on the median dorsal surface of the horizontal ramus. The alveolar border is produced anteriorly to form a strong ridge extending from the third premolar to the tusk alveolus. These ridges converge anteriorly and enclose a gutter-like suprasymphysial depression similar to that of the elephantoids. The symphysis itself is broad and extends anteriorly from the vicinity of the posterior root of the third premolar. In one specimen (6412 : 10) a wide and shallow pit is preserved on the.ventral and medial side of the horizontal ramus beneath the posterior root of Ps. The pit may represent the point of insertion of the anterior belly of the digastric muscle. On each ramus there are three mental foramina. In mandible 6404 : 13 the posterior foramen is slightly anterior to the front root of P; in 6412 : Io it is level with the hind root of this tooth. In both specimens the anterior mental foramen is sited about 8 cms in front of the posterior mental foramen, and the intermediate foramen varies in location between the two. Beneath the third premolar the horizontal ramus is oval in transverse section, being deeper than wide and slightly more convex laterally than medially. Beneath Ms the horizontal ramus has flat and parallel lateral and medial sides and develops a small keel on the ventral surface. The dorsal surface of the horizontal ramus increases in width posteriorly and the lateral surface becomes the leading edge of the ascending ramus when level with M3, from which it is separated by a wide, oval and concave platform. The leading edge of the ascending ramus is nearly at right angles to the horizontal ramus. The coronoid process is small and directed anteriorly. The mandibular condyle is elevated only a short distance above the coronoid process and is a little wider than the latter is long. The condyle is almost cylindrical. The inferior and posterior borders of the ascending ramus are appreciably thicker than the adjoining bone. There is a pronounced notch beneath the dental foramen on the inferior border of mandible 6412 : 10 that may mark the insertion of the sternomandibular muscle. The medial surface of the ascending ramus is divided unequally by a ridge extending from the base of the mandibular condyle to the alveolar border beneath Ms. A second ridge descends posteriorly from the anteromedial face of the coronoid process and joins the first beneath the hind edge of the coronoid process. The 312 PRODEINOTHERIUM triangular area thus enclosed serves for the insertion of the temporal muscle (which also occupies a depressed area of similar size on the lateral face of the ascending ramus). The masseter muscle inserts on the lateral surface of the angle of the jaw. The internal pterygoid muscle inserts similarly on the medial surface of the angle and the external pterygoid muscle inserts above the dental foramen and between the mandibular condyle and coronoid process. The digastric muscle inserts on the ventromedial surface of the mandible beneath and in front of the coronoid process. COMPARISON WITH OTHER DEINOTHERE MANDIBLES. The degree of curvature and length of the mandibular symphysis varies appreciably with the various deino- there taxa. The mandibular symphysis of Deinotherium giganteum is longer and has a more vertically aligned symphysial region than that of Prodeinotherium bavaricum. D. bozasi of the African Quaternary has a relatively short mandibular symphysis but this is flexed at right angles and therefore more abruptly than in any other deinotheriid taxon. The symphysis of D. levius (here interpreted as synony- mous with D. giganteum) is intermediate in length and degree of flexure between D. giganteum (sensu stricto) and P. bavaricum (Graf, 1957). Arambourg (1934) believed the apparently more acute flexure of D. Jevius separated this taxon from D. giganteum, and Laskarev (1944) deduced that D. levius had African rather than European affinities. The mandibular symphysis of P. hobleyi from Gebel Zelten is similar to that of P. bavaricum. Sahni & Tripathi (1957) separated the two Asian deinotheriid taxa on the basis of the transverse section of the horizontal ramus level with M3. In D. indicum the jaw bulges on either side of Mg, the lateral bulge being more pronounced and giving rise to a flat platform on the lateral side of this tooth and in front of the ascending ramus. In contrast, the jaw of P. pentapotamiae has a narrow and elliptical cross section. In this respect P. hobley: from Gebel Zelten is similar to D. indicum. Graf (1957) and Bergounioux & Crouzel (1962a) showed that the mental foramina of the mandible varied greatly in position. Those of D. giganteum and D. bozasi tended, however, to be sited more anteriorly than those of P. bavaricum. The mental foramina of the Gebel Zelten Prodeinotherium are also sited farther forward than in the examples of P. bavaricum quoted by Graf (1957 : 168). TABLE 5 Prodeinotherium hobleyi mandible measurements Left horizontal ramus (6412 : IO) Total length . : : ‘ : =< 163 Height at tusk alveolu : : , : 127 Height at Ps. : : : : : 13°53 Height at M,_ . : : F ; E I2°3 Height at Ms... : . : : 3 Lit Width at Ps. ; : : : : 7°3 Widthat M; : 7 d ! ‘ 7h | Width at M3_«tj ; : : ; : 10-7 FROM GEBEL ZELTEN, LIBYA 313 Left ascending ramus (6412 : IO) Length . : ; ; : a i22"6 Height at coronoid process : : 5 26 Height at mandibular condyle : ‘ 206 Mandibular condyle ee = Length ©. : ; : ‘ 4°7 Width ; : : : ; ; ‘ 103 Dentition (Text-fig. 7; Plates 4-5) MATERIAL REFERRED. 6401 : 4, right maxilla with P?-M3; 6404 : 13, mandible with tusks and cheek teeth; 6404 : 44, immature skull with cheek teeth; 6404 : 45, fragmented immature mandible with tusks and cheek teeth; 6412 : 10, tuskless mandible with worn cheek teeth; M.26665, adult skull with cheek teeth; M.26666, mandible with tusks and cheek teeth; 6401 : 1, LM3; 6404 : 14, LP?-M? and RP3-4; 6409 : 41, M2 fragment; 6410 : 2, RP4; 6412 : 39, LM2; 6412 : 40, LMj; 6412 : 54, LP4; 6413 : 18, LM®; 6416 : 120, tusk tip; 6416 : 130, tusk; 6421 : 2, LM2; 6404 : 28, 6405 : 48, 6419 : 17, 6421 : 15, 6423 : 13, 6423 : 45, 6423 : 56, enamel fragments. DEscripTIon. Teeth are the most commonly preserved deinotheriid remains and differ sufficiently in size and morphology from other Cenozoic mammals to permit ready identification even on incomplete specimens. The terminology employed in previous descriptions of deinothere teeth has varied greatly from author to author. The teeth are basically lophodont, the lophs originating through coalescence of adjacent cusps in a bunodont ancestral stock. Terms used here are based on the standard cusp nomenclature (Osborn, 1907) and those used by Osborn for probos- cidean teeth (Osborn, 1942). The terms are here used only to express topographic equivalence with the dental characters of other mammals and have no genetic significance. Superior dentition. Downturning of the rostral region involved the posterior retreat of the external nares, accompanied by withdrawal of the premaxillae to line the floor of the rostral trough. The incisors, canines and anterior cheek teeth were lost during, if not before, the hindward migration of the premaxillae. All lophs and cingula on the upper teeth are rugose until worn. The lingual edges of the transverse lophs are taller than the labial edges and show more wear. A well developed facet is formed on the anterior surface of each transverse loph and extends for the entire width of the loph. Parallel facets are seen also on the anterior and posterior cingula when the tooth is worn sufficiently for these to occlude with the lower dentition. The third premolar is equant and possesses two internal cusps and an ectoloph (text-fig. 7A). The protocone is the larger cusp and is separated from the hypocone by a deep but very narrow valley. The protocone extends labially and anteriorly to the base of the ectoloph, thus forming a low, incipient protoloph. The hypocone bears a small mesostyle on its anterolingual face. The hypocone is elongate labially but does not forma metaloph. The strong ectoloph is formed by the coalition of the paracone, metacone, and an intermediate cusp that is taller than either. The 314 PRODEINOTHERIUM posterior end of the ectoloph is continuous lingually with the posterior cingulum that terminates lingually on the posterolingual face of the hypocone. The median valley is wide and open posteriorly ; it is deepest between the labial bases of the protocone and hypocone. The anterior cingulum connects the base of the protocone to that of the ectoloph. There are no lingual or labia cingulla. The fourth premolar is bilophodont and subequant, being longer than broad. The ectoloph is less strong than in the third premolar and the intermediate cusp is reduced to a swelling on the anterior face of the metacone. The transverse lophs are not straight, the protoloph being convex anteriorly and the metaloph convex posteriorly. Both lophs are connected labially with the ectoloph. From the posterior lingual face of the protocone a ridge extends posterolabially to the centre of the median valley and a similar ridge extends down the anterior face of the A Fic. 7. Pyrodeinotherium hobleyi upper teeth. A = left P?, B = right M!; Not to scale. ac = anterior cingulum, h = hypocone, me = mesostyle, p = protocone, T = tritoloph, E = ectoloph, M = metaloph, P = protoloph, pe = posterior cingulum. FROM GEBEL ZELTEN, LIBYA 315 hypocone. The metaloph of the fourth premolar is the only loph in the upper dentition to bear a posteriorly directed wear facet. The median valley is widest labially next to the ectoloph and narrows, opening lingually between the protoloph and metaloph. It is almost, but not quite, dammed by a small mesostyle on the anterior face of the hypocone. The anterior and posterior cingula are well developed. There is a minute internal cingulum. The anterior molar is trilophodont and subrectangular, being longer than wide and narrowest posteriorly (text-fig. 7B). All three lophs bear ridges directed towards the midline of the tooth from their posterolingual and posterolabial edges. The tritoloph (posterior loph) is much less wide than either the protoloph or metaloph and the ridges from the posterolingual and posterolabial edges of the loph are less pronounced than in the protoloph or metaloph. Between the protoloph and meta- loph the anterior median valley is blocked labially by a small bune between the paracone and metacone, and lingually by a short spur connecting the protocone to the hypocone. This spur bears a small mesostyle. The posterior median valley, between the metaloph and tritoloph, is much less deep. It is blocked lingually by a short spur connecting the lingual edges of the metaloph and tritoloph, and labially by the junction of sharp ridges extending down the posterior face of the metaloph and anterior face of the tritoloph. The anterior cingulum is strong but the posterior cingulum is less well developed than in the premolars. The second molar is bilophodont and slightly longer than wide. The protoloph is similar to, but taller than, the metaloph. The centrally directed ridges on the hind face of the protoloph are well developed and stronger than in M1. The ridge from the lingual edge of the loph extends beyond the midline of the tooth; that from the labial end of the loph is more posteriorly directed and fuses with a similar ridge extending from the crown of the metacone. The median valley is open labially but is bounded lingually by a short spur joining the protoloph to the metaloph. The postmetaloph ornamentation is rather complex. A strong ridge extends from behind the hypocone to join the lingual edge of the posterior cingulum. A more ventrally directed ridge connects that from the hypocone to a posterolabially directed ridge from the metacone, thus forming a subsidiary loph behind the meta- loph. The posterior cingulum bears a large style on its labial extremity. The third molar is subequant, and tapers posteriorly. Like the second molar it is bilophodont. The topography of the protoloph is similar to that of M2. The metaloph is less wide than the protoloph, and the postmetaloph ornamentation is simpler. A ridge descends from the hypocone to the lingual termination of the posterior cingulum. The metacone gives off a strongly developed ridge that descends posteriorly and then centrally to the midline of the tooth to connect with the posterior cingulum. In some teeth the metacone ridge is also connected to the hypocone ridge and thus the postmetaloph ornamentation of M3 is more variable than in M2. There is no style on the posterior cingulum. Inferior dentition. Tusks are present in mandibles 6404 : 13, M.26666, and 6404: 45. The tusks are oval in transverse section with the long axis aligned anteroposteriorly. Enamel was preserved only on the tusks of the immature mandible (6404 : 45) where it covered the entire erupted surface of the tooth. The c 316 PRODEINOTHERIUM tusks in 6404 : 45 are short and nearly straight but in mandible M.26666 they curve downwards until the tip is nearly vertical. The dentine shows no ‘engine turning’ structure as in Mastodon but forms bands parallel to the periphery of the tooth. As in the superior cheek teeth, all lophs, cingula and isolated cusps of the lower molars and premolars are rugose until worn. Except on Py, all wear facets on the transverse lophids occur on the posterior surface. The labial edges of the lophids are less high but show more wear than the lingual edges. The third premolar is triangular and tapers anteriorly. The protoconid is the largest and tallest cusp. It is elongate sagittally and is sited in the midline of the tooth. The protoconid is continued forwards as a high narrow ridge extending to the lingual edge of a very abbreviated anterior cingulum. It is also connected posterolingually to a tall metaconid. From the junction of the protoconid and meta- conid, a narrow and less elevated ridge proceeds posteriorly to merge with an elongate hypoconid. This in turn connects with an elongate entoconid via a weakly developed, low, posterior cingulum. A C-shaped hypoconulid may or may not connect with the posterolabial edge of the entoconid. The protoconid and outer face of the ectolophid became worn more rapidly than the remainder of the tooth. The fourth premolar is bilophodont and represents a wider and more elongate version of Ps. The protoconid and metaconid are set farther apart and are con- nected by a protolophid that is curved with the convex face posteriorly. The anterolingual extension of the protoconid extends more laterally than in P3. The protoconid is also joined to the hypoconid by a low ectolophid. There is no hypo- conulid but the hypoconid is connected to the entoconid by a hypolophid that is straighter and less tall than the protolophid. There are no well defined wear facets on the hypolophid but the lateral extremities become worn faster than the central portion and small wear facets occur on both sides of the lophid. The hypoconid is connected by a short spur to the strong posterior cingulum. The first molar is trilophodont, the protolophid and hypolophid being wider than the tritolophid. A strong anterolingual ridge on the anterior face of the labial edge of the protolophid projects towards, but does not join, a weaker anterolabial ridge from the lingual extremity of the lophid. Similar but less strongly developed ridges may be traced on the anterior faces of the hypolophid and tritolophid. The anterior and posterior median valleys are open at both sides. The anterior cingulum is less well developed than the short, centrally placed posterior cingulum that stands proud at the rear of the tooth. The second and third molars are bilophodont and are differentiated only by the posterior taper on the Mg and the better developed posterior cingulum on that tooth. From the labial edge of the protolophid a strong medial ridge extends postero- ventrally to meet the anterior cingulum. A similar but less strong ridge extends from the lingual extremity of the lophid. These ridges are duplicated in a less pronounced fashion on the anterior face of the hypolophid. The median valley is open at bothends. As in Mj, the wear facets on the lophids face posteriorly. The posterior cingulum of Mg is less wide than that of Mg but is taller, more curved, and extends farther posteriorly. p4 M2 M3 My Mg ap prot trit * FROM GEBEL ZELTEN, LIBYA TABLE 6 Prodeinotherium hobleyi tooth measurements (in mm.) Upper premolars 6401: 6404: 6404: 6412: M.26665(1) + 14(I) 14(r) 55 ap 47°5 564 Ceieiae == 40°5 prot — 55:0 — 511 — met 50°1 55°60 — — 47°5 ap Aare 47°1 = Te 47°9 prot — 62:7 64:0 — 5855 met — 60°3 = — 56-7 Upper molars 6401: 6401: 6404: 6413: M.26665(1) I 4 14(l) 18 ap —_ 68-6 76:6 — 70°5 prot — 58-8 60-7 —— 62°5 met — 56-7 60-6 — 61°5 eet 47°7 49°2 = = ap — 67°8 732 — 70:6 prot — 67°9 69:0-+ —_— — met — 66:5 64:0-+ — — apy 9 075°7 67-2 = 59°3 728+ prot 77:2 67°9 —_ 62-0 77:6 met 66-9 62:6 — — 739 Lower premolars 6404: 6404: 6404: 6404: 6410: 6412 13(1) 13(r) 45(1) 45(r) 2 10(1) ap 43°3 45°5 491 5°°7 a 39°6 tr 32°3 35°0 40°7 39°6 a 31°7 ap = 544 531 = 58-9 == 44°3 prot 43-0 43°9 = 46'9 38°4 38-6 hyp 45°6 43°2 a 43°3 34°9 = Lower molars 6404: 6404: 6404: 6404: 6412: 6412 6412: 13(I) 13(r) 45(1) 45(r) ro(1) 10(r) 39 ap = 70°0 65°3 82-7* = 55°0 55°3 — prot 48-8 47:0 47°4* —_— 38°8 41°8 — hyp 51:0 49°1 48-2* = 39°5 39°8 = trit 45:9 44°3 38-3 = 33°8 34°2 = apo 71-9 om 76°3 57°0 54°8 = prot 65°5 61-6 — 61-1 503 50°4 52°8 yp 59°0 = 62:2 48-0 49°5 = ap — — a= 84-0 66-9 67:8 _— prot — = = 63°4 541 54°9 = hyp == = = 59°3 47°6 46°4 = = greatest sagittal length: tr = transverse width; = width at protoloph(id) ; = width at tritoloph(id) ; = approximate measurement. met = width at metaloph(id) ; hyp = width at hypolophid; M.26665(r) 43°7 46°1 44:9 58°4 57°6 M.26665(r) 715 62:9 317 318 PRODEINOTHERIUM TOOTH FUNCTION IN DEINOTHERES. The function of deinothere tusks has yet to be satisfactorily explained. Conventional interpretations include pulling branches down to mouth level and digging for food. It is unlikely that deinothere tusks were used for digging for the following reasons. The tusks would fail to reach the ground unless the animal was either kneeling or standing at a lower elevation than the surface at which digging occurred. Such wear facets as have been observed are not consistent with a digging function. Food excavated from below ground surface would inevitably be contaminated with soil which would result in rapid attrition of the cheek teeth; no compensatory increase in hypsodonty of the deinothere cheek teeth has been observed. The tusks of immature deinotheres are covered by enamel. Enamel production apparently ceases shortly after the eruption of the tusks. The absence of enamel on the tusks of mature deinotheres suggests that the enamel-bearing portion was subse- quently removed by wear. I have observed wear facets in two places on deinothere tusks—on the posterior surface of the tusk tip (P. hobleyz) and on the anteromedial surface of the tusk tip (P. hobleyi, P. bavaricum, D. bozasi). The anteromedial facets may be attributed to constant abrasion by the proboscis but might conceivably have been produced during the action of stripping bark or branches from vegetation. A detailed description of the cheek tooth function in deinotheres is in the process of preparation. A summary of the preliminary conclusions is given below. Deinothere cheek teeth occlude orthally and may be functionally divided into anterior crushing teeth and posterior shearing teeth. In the premolars some shearing is effected during occlusion of the ectoloph and ectolophid. The transverse crests on the premolars serve to prevent anterior dislocation of the jaw during the occlusal stroke of the mandible and perform a crushing function during the recovery stroke. The transverse crests of the deciduous fourth premolars and of the molars have well developed shearing facets—on the anterior surface of the lophs and rear of the lophids. These facets are maintained until the transverse crests have been almost completely removed by wear. In unworn teeth the facets are aligned almost vertically but become progressively more horizontal with wear. The change in angle of the shear facet is seen not only between individuals of different maturity but also within single tooth rows, the facets on the posterior end of the tooth row being oriented more vertically than on the anterior molars. The anterior crushing battery and posterior shearing battery are reconstituted three times during the life of the individual as follows: anterior battery posterior battery Juvenile animal DP2-3 DP4 Immature adult P3-4 M1 Mature adult P3-4, Mr M2-3 The anterior trilophodont molar performs different functions at different stages of wear. Initially it serves as a shearing tooth in contrast to the crushing premolars. By the time the second and third molars are fully functioning the anterior molar is often too worn to provide an efficient shearing action and becomes incorporated into the anterior crushing battery. FROM GEBEL ZELTEN, LIBYA 319 The width of the anterior molar, its attitude in the alveolus, and its alignment in the tooth row all suggest that it functions primarily as a posterior premolar rather than as an anterior molar. The shearing action of this tooth is of limited duration but sufficient to cope during the transition between deciduous and permanent dentition. The bifunctional nature of the tooth row may well explain why deino- theres failed to develop horizontal tooth replacement as in the elephantoids. Axial skeleton (Text-figs 8-13) MATERIAL REFERRED. M.26667h, left rib head; M.266671, right neuropophysis of third cervical vertebra; 6419 : 15, atlas vertebra; 6424 : 79, axis vertebra. DescripTIon. Adlas vertebra. The atlas vertebra of Prodeinotherium hobleyi (text-fig. 8, 6419 : 15) is smaller than that of Mastodon angustidens (6412 : 185) from Gebel Zelten. Although the transverse processes of the mastodont atlas are incomplete the Prodeinotherium atlas was probably wider. The facets for articula- tion with the occipital condyles are comma-shaped in both specimens but the masto- dont facets are larger, relatively taller, less wide and apparently more concave. Differences on the dorsal edge of the neural arch are especially pronounced. The neural spine of the mastodont atlas forms an elongate transverse ridge with large and widely separated scars for the insertion of the rectus muscles. The Prodeino- thertum atlas spine is a shorter, blunt protruberance that is bounded by rectus muscle scars that are smaller and face more ventrally than in Mastodon. The transverse processes of the mastodont specimens are incomplete but were apparently more horizontal than those of Prodeinotherium and are sited higher on the neuro- pophysis, extending for some 4 cms above the dorsal edge of the axis facet. The odontoid fossa of the mastodont atlas is also incomplete but appears to rise more steeply anteriorly and to be more concave than that of Prodeinotherium. The atlas attributed to Prodeinotheriwm does not articulate well with a Prodeino- therium axis vertebra (6424 : 79) from an unknown locality at Gebel Zelten, mainly due to differences in the articular facets. The general proportions of the occluding facets and odontoid fossa and processes are however sufficiently similar for the poor articulation to be attributed to variation between individual specimens. TABLE 7 Atlas vertebva measurements Parameter Prodeinotherium Mastodon Loxodonta hobleyi angustidens africana 6418315 M.26656 1961.8.9.82 Maximum length 10°4 10.7-+-* 78 Maximum height 17-2 20:0 17'9 Width at transverse process 28-7 24°3+ 30°74 Width at occipital facets 18°4 20°7 Pita) Width neural canal Or 7°3 8:0 Height neural canal 5:0 33) 3°9 Width odontoid fossa 56 6°5 Ps Height odontoid fossa 6-2 56 6:2 Width at atlas facets 16°7 16-3* 17°7 PRODEINOTHERIUM Fic. 8. Prodeinotherium hobleyi atlas vertebra. 5 cm. anterior view, B = posterior view. Scale = A= FROM GEBEL ZELTEN, LIBYA 321 ————uc“K« Sa 22 Fic. 9. Lowxodonta africana atlas vertebra. A = anterior view, B = posterior view. Scale = 5 cm. 322 PRODEINOTHERIUM Axis vertebra. The sole deinotheriid axis from Gebel Zelten (6424 : 79) is incomplete dorsally and ventrally (text-fig. 10). In overall size it is slightly smaller than the axis vertebra of the extant African elephant (text-fig. 11) but the odontoid process, although eroded, is longer and stouter in Prodeinotheriwm. The transverse processes are broken at their bases and the entire ventral half of the centrum is missing. The cephalic articular surfaces are oval, extending posteroventrally from above the odon- toid process. It is not clear whether, as in Loxodonta, the articular facets extend to meet below the odontoid process because they are incomplete. The neural canal is circular in transverse section. The neural archisnotched anteriorly and posteriorly for the passage of the second and third cervical spinal nerves, the posterior notch being the deeper. v/ (' lll \ Fic. 10. Pyrodeinotherium hobleyi axis vertebra. = anterior view, B = posterior view, C = right lateral view. Scale = 5 cm. FROM GEBEL ZELTEN, LIBYA 323 The spine is the most diagnostic feature of the Prodeinotheriwm axis. In the extant elephants ‘the neuropophyses blend together above and develop a thick bifurcate spine before coalescing with the centrum’ (Owen, 1866 : 547). In the Gebel Zelten axis the neuropophyses are less tall but longer anteroposteriorly than in Loxodonta. The bifurcation of the spine is extreme, beginning at the anterior tip of the spine and diverging at an angle of 55° (cf. 30° in Loxodonta). The twin crests Fic. 11. Lowxodonta africana axis vertebra. A = anterior view, B = posterior view, C = left lateral view. Scale = 5 cm. 324 PRODEINOTHERIUM thus formed ascend backwards less steeply than in the elephant but reach farther posteriorly and are separated by a U-shaped valley that widens posteriorly and descends at an angle of nearly 30° from the horizontal. In Loxodonta this valley appears only at the posterior portion of the spine and is horizontal. The post- zygapophyses of the elephant axis are convex; the greater part of the articular surface faces ventrolaterally but the dorsomedial corner faces posteriorly. Much of the articular surfaces of the zygopophyses of the Gebel Zelten Prodeinotherium axis are eroded but they are seen to be larger than in the elephant, to be concave and to face ventrolaterally. A small triangular facet may be discerned along the dorsomedial edge of, and at right angles to, the main articular surface but this may have served for the attachment of muscles or ligament rather than for articulation. Above the postzygapophyses much of the bone is missing but this region of the spine must originally have been greatly swollen. Although the neuropophyses are appreciably wider and longer than those of the elephant, the centrum is only slightly longer. It is not until the relative areas of the neural canals are taken into consideration that the axis of Prodeinotherium is seen to be more massive and appreciably longer than that of the African elephant. The larger surface of the neural spine affords a greater area of attachment for the rectus capitus posterior major muscle, obliquus capitus inferior muscle, senuspinalis capitus and semispinalis cervicus muscles. All four muscles serve to extend the head on the neck or, acting unilaterally, to rotate or flex the head on the neck. The narrow width of each diverging supraspinous crest suggests that the nuchal ligament was inserted only on the posterior portion of the dorsal surface of the supraspinous crest. The great width of the spine reflects the great width of the nuchal ligament that is in turn reflected by the wider fossae for the nuchal ligament on the occiput of Prodeino- therium than in elephantoids. The relatively low dorsoventral height of the spine may reflect the elevated condition of the occipital condyles and/or the less elevated head of deinotheres. TABLE 8 Axis vertebva measurements Parameter Prodeinotherium Loxodonia . hobleyi africana 6424:79 1961.8.9.82 Maximum length 17°1 13°4 Maximum width 17°5-+ 17°5 Maximum height 17-7 24°6 Width at atlas facets 176+ 17-2 Width neural canal 5:2 6-2 Height neural canal 5:0 5°8 Width odontoid process 379+ 3°4 Height odontoid process 4°7 36 Maximum length neuropophysis 15:2 I1°9 Minimum length neuropophysis 5°5 36 Minimum width neuropophysis 15°3 TF, Height neuropophysis above neural canal 76+ 8-1 Length postzygopophysis 52+ 5°5 Width postzygopophysis 5255 Si FROM GEBEL ZELTEN, LIBYA 325 Third cervical vertebva. An incomplete right neuropophysis of a third cervical vertebra M.266671 was associated with immature Prodeinotherium skeletal elements from Gebel Zelten site D. The specimen (text-fig. 12) is not well preserved but incorporates the region between the neural spine and the vertebrarterial canal. The neural spine is very low and could have been elevated only slightly above the valley separating the two wings of the axis spine. The neural canal is compressed dorsoventrally. The neural arch is longer anteroposteriorly but less stout than in Loxodonia. The prezygapophysis is convex and faces anterolaterally. The post- zygapophysis is concave and faces ventrally. Both zygapophyses are aligned more horizontally than in Loxodonta. The neural arch is notched anteriorly and posteriorly for the passage of the spinal nerves, the posterior notch being deeper and extending - YA (6 \ Wy We tll Ww Fic. 12. Prodeinotheyium hobleyi third cervical vertebra. A = anterior view, B = posterior view, C = lateral view. Scale = 5 cm. 326 PRODEINOTHERIUM to the dorsal edge of the vertebrarterial canal. The surface of the bone is not well preserved and muscle attachment areas cannot be detected. It is interesting that the articular surfaces of the zygapophyses have reversed curvature when compared to the elephantoids in which the postzygapophysis and not the prezygapophysis is convex. This is believed to have some bearing on the direction of movement of the neck in elephantoids as opposed to deinotheres. The centra of the cervical vertebrae of the Gebel Zelten deinothere shows some flattening anteroposteriorly as in the elephantoids but the neuropophyses are considerably more massive. Taking into consideration the diameter of the neural canal, the cervical vertebrae of the Gebel Zelten Prodeinotherium were more elongate than those of Loxodonta. TABLE 9 Measurements of third cervical vertebra Parameter Prodeinotherium Loxodonta hobleyi afvicana M.266671 1961.8.9.82 Maximum length neuropophysis WeWse Fal Length neuropophysis at neural arch 371 2:8 Width transverse process above vertebrarterial canal 2:8 Ig Width neural canal 70+ 6:8 Height neural canal 4°4 5°7 Height neural arch above centrum 5:9 9:0 Fic. 13. Loxodonta africana third cervical vertebra. A = anterior view, B = posterior view. Scale = 5 cm. FROM GEBEL ZELTEN, LIBYA 327 Rib One left rib was collected with the immature Prodeinotherium skeleton from site D. Other rib fragments were noted but were too incomplete to warrant collection. The head of the rib M.26667h is large, rounded and divided by a deep transverse groove into two convex oval facets for articulation with two thoracic vertebrae. The transverse groove is widest medially, and the cranial articulation facet is larger than the caudal. The neck is short and a little wider than the head. On the dorsal edge of the neck is an oval, slightly convex tubercule that articulates above the caudal facet of the head. On the caudolateral side of the tubercule is a rounded prominence, one fourth the size of the tubercle, that may have served for attachment of the longissimus dorsi muscle. The shaft of the rib extends ventro- laterally from the neck but flexes downwards at nearly 40° at the angle. The dorsal surface of the angle bears an elongate prominence for the attachment of the iliocostal muscle. On the cranial surface of the rib, between the head and the angle, there are two elongate fossae. On the caudal surface a deep groove appears midway between the head and angle near the lateral border of the shaft. A trans- verse section of the shaft proximal to the angle resembles the shape of an hourglass; distal to the angle it is concavoconvex (concave caudally). Distal to the caudal depression the cranial face turns laterally and the transverse section becomes oval and progressively flattened towards the end of the rib. Extending over the cranial surface of the shaft from the angle are a number of irregular prominences that serve for the attachment of various parts of the serratus posticus muscles. The bicipital head of the rib suggests that it was articulated towards the front of the rib cage. TABLE I0 Measurement of Prodeinotherium vib M.26667h Parameter Dimension Length head . ; ; é . 2 j 53 Width head. j : é : : : 5:0 Width neck. : ; : : : : 5:2 Length neck . F ; : s : 3 Br7 Length neck to angle 4 ; b 5 : 10°7 Length at angle ‘ : é : ‘ . 6:8 Breadth at angle. c é : : : 2° Discussion. In mammals generally, only one muscle is attached to the anterior face of the arch of the atlas vertebra. This muscle, variously recorded as the rectus capitis dorsalis minor (Sisson & Grossman, 1947), rectus posterior minor (Jayne, 1898), and rectus capitis lateralis posterior minor (Shindo & Mori, 1957) inserts on the occiput lateral to the nuchal fossa and serves to extend the head on the neck. The reduced width of the neural spine of the deinothere atlas compared to that of elephantoids is due, at least partly, to the fact that the sites of origin of the rectus muscles are closer together and face upwards and outwards rather than forwards. The line of action of the muscles is thus less horizontal and more vertically aligned than in Mastodon or the elephants. The change in line of action of the musculature 328 PRODEINOTHERIUM reflects the different geometry of the occiput of Prodeinotherium which has a more dorsally placed nuchal fossa than has that of the elephantoids. The more posteriorly inclined transverse processes may be partly due to a more vertical alignment of the muscles extending from the transverse process to the occiput, or may reflect the importance of muscles originating from the ventral surface of the transverse process to insert on the basicranium and paroccipital process. The Gebel Zelten Prodeinotherium axis differs in many features from that of the elephantoids, the more striking differences being associated with the development of the neuropophyses. The muscles attaching thereto serve to move both the neck on the thorax and the head on the neck. The cross sectional area of the anterior opening of the neural canal of the axis vertebra of Loxodonta (ca. 4000 mm?) is half as large again as that of the Gebel Zelten deinothere axis (ca. 2600 mm2) and if this area may be taken to suggest the relative size of these two mammals (cf. Radinsky, 1967) then the Gebel Zelten deinothere, although smaller than the African elephant, has much larger and longer neuropophyses than the living animal. It is obvious therefore that the axis of Prodeinotherium plays a more important role in the move, ment of the head and neck than that of Loxodonta. Five major muscles are attached to the axis spine. The rectus capitis posterior major, obliquus capitis inferior, and semispinalis capitalis originate on the axis and move the head about the atlantooccipital joint; the scalenus and semispinalis cervicus insert on the axis and move the head on the thorax. The lamellar portion of the nuchal ligament also inserts on the spine of the axis. Three muscles possess considerably larger attachment areas on the spine of the Prodeinotherium axis than on that of the elephant. These are the obliquus capitis inferior, the rectus capitis posterior minor and the semispinalis cervicus. The former rotates, extends and fixes the atlantooccipital joint (and thereby the head), while the rectus capitis extends the head and the semispinalis cervicus extends the neck. All three are attached to the upper part of the neuropophysis. That the spine is less elevated than in Loxodonta suggests that its function as a lever for moving the head on the thorax is less important than its capacity to move the head on the neck. The external walls of the spine are orientated more laterally than in Loxodonta thus affording a more nearly perpendicular line of action for the muscles moving the head and thereby enhancing the mechanical efficiency of these muscles. The sagittal length and divergent nature of the neural spine of the axis also affords a greater area of attachment for the digitation of the lamellar portion of the nuchal ligament. The postzygapophyses of the axis and zygapophyses of the succeeding cervical vertebrae of deinotheres are more horizontally aligned than in the elephantoids. The curvature differs also—the postzygapophyses of deinotheres are concave and the prezygapophyses convex in direct contrast to the condition normally found in the elephantoids. Thus while the zygapophyses of the anterior cervical vertebrae of elephantoids permit an upward flexure of the neck, those of Deinotherium and Prodeinotherium flex in the opposite direction. The differences in direction of flexure may be related to the different ways in which the tusks function in deino- theres and elephantoids. FROM GEBEL ZELTEN, LIPYA 329 The centrum of the axis is slightly longer in Prodeinotherium than in Loxodonta and is appreciably longer when the area of the neural canal is used as a basis for comparison. The neural arch of the third cervical vertebra from Gebel Zelten is also longer than its equivalent in Loxodonta and it is probable also that this was true of the remaining vertebrae of the neck. Increase in the length of the neck amounts to a decrease in the lever arm of the musculature supporting the head on the neck and moving the head on the thorax. The decrease in lever arm of these muscles reduces their mechanical efficiency, and a larger and stronger musculature is therefore required to support the head on the neck. This is reflected by the large area of insertion of the semispinalis cervicus muscles in Prodeinotherium. Without access to the cervical vertebrae of Deinotherium one cannot deduce whether the relatively long neck of the Gebel Zelten deinotheres is a primitive feature, the neck becoming shorter and mechanically more efficient in later forms, or whether it was a functional adaptation to permit the head to be nearer the ground and thereby enhancing the potential use of the tusks for feeding. The thoracic vertebrae of Deinotheriwm are incompletely known, the anterior thoracics being best known from the skeleton of D. ‘gigantissimum’ (= D. giganteum) from the Pliocene of Manzati, Rumania. The neural spines of these vertebrae are inclined steeply posteriorly. Slyjper (1946) pointed out that the neural spines act as levers to transmit muscular force to the centra. The force applied determines the height and direction of the spines. Increase in total body size is accompanied by a relative increase in muscle strength, the size and shape of the spines being therefore dependent also on body size, weight of the head, and length of the neck. The most favourable direction of the neural spine is perpendicular to the muscular force acting on the spine, but intermediate directions are achieved when two or more forces act in different directions upon one spine. In the elephant three main forces are applied to the neural spine of the anterior thoracic vertebrae by the nuchal ligament, the sfinalis dorsi, and the multifidus spinae muscles. Tension in the nuchal ligament exerts a force pulling the spines anteriorly whilst contraction of the muscles pulls the spines posteroventrally. The funicular portion of the nuchal ligament exerts most of its force on the anterior thoracic vertebrae. The anterior thoracic neural spines of Loxodonta are more perpendicular to the centra than those of Elephas, although in both animals the spines slope backwards. This indicates that the nuchal ligament of Loxodonta is more horizontally aligned and, conversely, that the head of Elephas is more elevated with regard to the anterior thoracic vertebrae. Because the nuchal ligament functions most effectively when the lever arm (neural spine) is perpendicular to the line of action of the ligament one may deduce that mammals whose heads are sited above the withers will have backward pointing spines whereas those mammals whose skulls are on the same horizontal plane as the withers will have upright spines. The few mammals with skulls held below the level of the withers will also have vertical neural spines but because the line of action of the ligament will not be perpendicular to the spines the lever arm of the ligament must increase to compensate for reduction in efficiency. The neural spines of these mammals will therefore be 330 PRODEINOTHERIUM correspondingly longer than those of mammals in which the skull and shoulders are level. The vertebral column of D. ‘gigantissimum’ was restored and mounted by de Pauw (Stefanescu, 1907; Pl. 1). The neural spines of thoracic vertebrae 2-5 are far more steeply inclined than in those of either of the extant elephants. The original specimens from Manzati, Rumania, were incomplete and their exact length cannot be determined. Backward inclination of these neural spines may be inter- preted in two ways: either the head of the deinothere was carried above the shoulders (the backward inclination being influenced by the line of action of the nuchal liga- ment) or the vertebral column was rigid (the position of the spines being determined by the multifidus muscles). Total length as well as the angle of inclination of the neural spines must be considered before either alternative can be fully investigated. With regard to the rigidity of the vertebral column, none of the extant graviportal mammals can flex their lumbar and thoracic vertebrae. In view of the large size of the deinotheres it seems likely that the vertebral column would have been similarly modified and perhaps the extreme spinal inclination suggests extreme inflexibility. Appendicular skeleton (Text-figs 14-16) MATERIAL REFERRED. M.26667a, distal end of fibula; M.26667b, right ulna; M.26667c, proximal end of right humerus; M.26667d, right femur; M.26667e, left ulna; M.26667f, scapula; M.26667¢, left innominate; M.26667j, proximal and distal end of left radius; M.26667k, distal end of right humerus; M.26667]1, distal end of left humerus; 6405 : 98, left cuneiform. DESCRIPTION. Scapula. An almost complete right scapula (M.26667f, text-fig. 14) was associated with the immature skull from Gebel Zelten site D. The left scapula was also present but was partly overlain by skull 6404 : 44 and was destroyed during excavation of the latter. The fragility of scapula M.26667f resulted in the collapse of the vertebral border and cranial and caudal angles during transportation from the site but these portions were subsequently restored with the aid of photo- graphs taken at the site. The articular fossa is broad and concave. The scapular tuberosity is poorly developed but the coronoid process is well defined. The scapular notch is deeply indented. The spine divides the lateral surface of the scapula into supraspinous and infraspinous fossae of similar proportions to those of Elephas maximus. The spine attains its highest elevation in its ventral third at the point of divergence of the acromion and metacromion. From here it descends gradually to the neck and caudal angle. Near the neck the cranial surface of the spine is concave but becomes convex near the cranial angle. The reverse is true of the caudal surface. Although the scapula is similar in size to that of the Indian elephant, the caudal angle is sited dorsal to the point of divergence of the acromion and metacromion. In the Indian elephant the base of the acromion is level with the caudal angle. The FROM GEBEL ZELTEN, LIBYA 331 B Fic. 14. Right scapula of (A) Pyodeinotherium hobleyi and (B) Elephas maximus. Scale = 5 cm. D 332 PRODEINOTHERIUM head of the Prodeinotherium scapula is wider and less concave than that of the elephant and the spine ascends more gently from the neck than in Elephas. The scapular notch is deeper and the scapular tuberosity less pronounced than in Elephas. TABLE II Scapula measurements Parameter Prodeinotherium Elephas hobley1 maximus M.26667j U.B.20189 Length head to vertebral angle 52+ 62 Transverse width at caudal angle 35 42 Length caudal angle above head 25 15 Length caudal angle behind spine 26 30 Length spine 40 44 Length acromion 9 17 Length metacromion 6 16 Maximum width spine behind metacromion 5 3-7, Width head (straight) 1533 1453 Width head (curved) 17 16 Depth head 10* 8-3 Humerus. The distal extremities of a right and left humerus M.26667k and M.266671) were associated with the immature skull 6404 : 44. They are complete from the condyles to the deltoid tuberosity and proximal end of the lateral supra- condyle ridge. This part of the bone is slightly shorter proximodistally than the equivalent part of the humerus of the Indian elephant but is appreciably wider and deeper suggesting a larger area of origin of the many flexor and extensor muscles that arise from this portion of the humerus. In cranial aspect the distal end of the musculospiral groove is seen to curve laterally towards the proximal end of the lateral supracondylar ridge. This is a strong ridge, straighter and more salient laterally than its equivalent in the elephant. Asin Elephas the lateral epicondyle does not project very far laterally but it is stronger cranially in Prodeinotherium. Medial to the lateral epicondyle there is a small but deep fossa separated by a low ridge from the very wide but shallow coronoid fossa. The median epicondyle of the trochlea is larger than the lateral, the disparity in size being more pronounced than in Elephas. In caudal aspect the lateral and medial epicondyloid ridges are considerably wider than their counterparts in Elephas, as are also the epicondyles. Thus, despite the great overall width of the distal portion of this bone, the olecranon fossa is con- siderably narrower than in Elephas. In Elephas the epiphysial suture between the median epicondyle and the condyle is aligned diagonally whereas in Prodeinotherium it runs medially, thereby increasing the area of the condyle. The anconeal process of the Prvodeinotherium humerus is narrower but no more salient than that of Elephas. FROM GEBEL ZELTEN, LIBYA 333 Fic.15 Fic. 16 Fic. 15. Radius of Prodeinotherium hobleyi and Mastodon angustidens. A = P. hobleyi left radius (M.15429), cranial view B= M. angustidens right radius (6412 : 166), cranial view C = P. hobleyi head of left radius (6404 : 41), dorsal view D = P. hobleyi head of left radius (6404 : 41), caudal view Fic. 16. Radius of Prodeinotherium hobleyi and Mastodon angustidens. A = P. hobleyi left radius (M.15429), caudal view B = P. hobleyi left radius (M.15429), distal view C = M. angustidens right radius (6412 : 166), caudal view D = M. angustidens right radius (6412 : 166), distal view Scale = 5 cm. Dt 334 PRODEINOTHERIUM TABLE 12 Humerus measurements Parameter Prodeinotherium Elephas hobley1 maximus M.26667k M.266671 U.B.20189 Width at dorsal edge of lateral epicondyle 18-5 183 14°6 Length from lateral epicondyle to distal end of condyle 19°7 19°7 14:6 Width at median epicondyle 20°5 20:9 19:0 Length from median epicondyle to distal end of median condyle 14'0 13°4 10°5 Width of articular surfaces on condyles 16:2 16-7 1593 Depth at median epicondyle 14:2 14°8 13°4 Minimum width shaft above lateral epicondyle 974 1O'L 77 Minimum depth shaft above lateral epicondyle 9:2 9:3 7:6 Depth of articular surface of median condyle 12:0 10:2 I2'I Depth of articular surface of lateral condyle 10-2 10:2 II‘O Radius. Only two proboscidean radii have been collected from Gebel Zelten— the head and neck of a left radius of P. hobleyi (text-fig. 15C and D; M.26667j) and a nearly complete right radius (M.26654; text-figs 15 and 16) of Mastodon angustidens. Two left proboscidean radii from East African Miocene localities are present in the collections of the British Museum (Natural History). That from Rusinga (B.M. (N.H.)M.21872) was attributed to P. hobleyi by MacInnes (1942) but belongs to Mastodon angustidens. The other specimen (B.M.(N.H.)M.15429) is from Maboko Island and belongs to Prodeinotherium (text-figs 15 and 16). The proximal articular surface of the radius of Mastodon angustidens is trapezoidal in outline, with the medial side longer than the lateral. The capitular depression for articulation with the humerus is not large and extends through the centre of the articular surface parallel with the caudal and cranial edges. At right angles and distal to the capitular depression, articulation facets for the ulna occur on the medial and caudal sides, the medial facet being the smaller. Beneath the neck and caudal edge of the median articulation facet for the ulna, the radius is swollen to produce a radial tuberosity for the insertion of the biceps brachii muscle. This tuberosity is more pronounced than in Prodeinotherium or Elephas. Immediately below the radial tuberosity the shaft is approximately rectangular in transverse section, the caudal and cranial edges being the longest and the lateral edge being slightly shorter than the medial. Distally the medial and lateral edges diverge and the shaft is compressed craniocaudally but whereas the shaft of M. angustidens retains this shape throughout its length, that of Prodeinotherium becomes more equant in its distal half. Halfway down the shaft of M. angustidens the medial edge of the cranial surface is flattened for the insertion of the pronator teres muscle. The interosseous crest extends down the medial side of the caudal face from beneath the radial tuberosity. Distal to the insertion of the pronator teres, and at the proximal end of the distal third of the shaft, the interosseous crest diverges, a strong ridge extending towards the medial edge of the epiphysis and a weaker ridge towards the lateral. Lateral FROM GEBEL ZELTEN, LIBYA 335 to the interosseous crest is a deep depression extending from the radial tuberosity to the distal fourth of the bone. The radius of Prodeinotherium is smaller than that of M. angustidens and is less straight proximodistally and craniocaudally. The radial tuberosity is comparatively less well developed and the interosseous ridge diverges halfway down the shaft. The distal extremity is less wide and deeper craniocaudally than in M. angustidens. The radii of both M. angustidens and P. hobleyi are of similar length to the radius of Elephas maximus but the shafts are much wider and the entire bone stouter. TABLE 13 Radius measurements Parameter Prodeinotherium Mastodon Elephas hobleyi angustidens maximus M.26667} M.15429 M.26654 M.22872 U.B.20189 Length head to distal epiphysis — 48 63 59 43 Width proximal epiphysis 5°8 50+ 7:0 6:5 4°6 Depth proximal epiphysis 9:0+ 8-5 10°3 85+ 8-5 Width at distal epiphysis a 10-6 13°7 T4°5 Io-l Depth at distal epiphysis = 8-3+ Viet 6-9 6°5 Width at centre of shaft == 54+ 1 Te 43 Depth at centre of shaft — 471 375 3°7 371 Ulna. Right and left ulnae were associated with skull 6404 : 44 but the shaft of the left ulna (M.26667e) is shattered and only the right ulna M.26667b could be prepared completely. The right ulna lacks the distal and olecranon epiphyses and also the proximal portion of the semilunar notch including the anconeal process. The bone is approximately equal in length to that of the extant Indian elephant. The proximal extremity is the stoutest part of the bone and the olecranon, even with- out the epiphysis, is larger than the olecranon plus epiphysis of the Indian elephant. On the posterolateral side of the semilunar notch is a large scar marking the insertion of the anconeous muscle. A coronoid process projects cranially on either side of the radial notch, the medial process being the larger. The ulnar tuberosity, located on the craniomedial side of the ulna and separated from the medial coronoid process by a large depression, is less well developed for the insertion of the brachialis muscle than is the equivalent tuberosity in the elephant. The middle third of the ulna shaft is triangular in transverse section with cranial, medial and lateral faces but reverts to a pentagonal cross-section at the distal epiphysis. The interosseous ridge extends distally along the medial edge beneath the ulnar tuberosity but dies out at the proximal edge of the distal third of the shaft. Flat surfaces on the caudal aspect of the bone mark the areas of attachment 336 PRODEINOTHERIUM of the external pollicis longus, extensor communis and flexor digitorum profundus muscles and the ligament humeroantibrachialis. These are all developed much as in the Indian elephant. The shaft is of similar length to that of the Indian elephant but is considerably deeper and wider. TABLE 14 Ulna measurements Parameter Prodeinotherium Elephas hobleyi maximus M.26667b M.26667e U.B.20189 Length from olecranon to distal epiphysis 61* 53* 63 Width olecranon 13°6 12-2 15:2 Depth olecranon below semilunar notch HOF 12-2 13-2 Width semilunar notch 7-0 = 6-7 Width coronoid processes 19°4 a 15°4 Width distal end of shaft I1°8 10:9 12:2 Depth distal end of shaft 13°7 I4'7 II‘g Width midpoint of shaft 8-4 = 673 Depth midpoint of shaft 7:0 — 6-3 Carpal bones. A small number of carpal bones have been collected from Gebel Zelten sites. MacInnes (1942) reported several carpals from Rusinga Island, Kenya, and these he attributed to P. hobleyi. Examination and comparison of the East African and Gebel Zelten material has revealed that only three of MacInnes’ speci- mens may definitely be attributed to Prodeinotherium—an incomplete left cuneiform (M.21875), a right magnum (M.21877) and a left unciform (M.21882). These are redescribed with the Gebel Zelten material. Cuneiform. The left cuneiform from Gebel Zelten (6405 : 98) is better preserved than that from Rusinga. The proximal surface that articulates with the ulna bears two articulation facets (a median rectangular facet and a lateral triangular facet) separated by a sagittal ridge. In Elephas and Mastodon the proximal articular surface of the cuneiform is saddle-shaped with a transverse depression through the centre. The distal surface of the deinothere cuneiform bears a triangular concavo- convex facet for articulation with the unciform, much as in the elephant or mastodon. On the caudal border there is a triangular facet for articulation with the pisiform, the hypoteneuse of the triangle being bounded by a deep groove extending down the large posterolateral process from the caudal junction of the proximal facets. The distal surface of the posterolateral process of the Gebel Zelten cuneiform is eroded but, as in Elephas, it bears a facet for articulation with the fifth metacarpal. The medial surface bears a cranial and (larger) distal facet for articulation with the lunar. FROM GEBEL ZELTEN, LIBYA 337 TABLE 15 Cuneiform measurements Parameter P. hobleyi D. giganteum = M. angustidens E. maximus 6405 :08 (Howenegg) M.21874 U.B.20189 Maximum length 8:5 15°6 8-6 7-9 Maximum width Tez 19°3 Iro+ II‘ Depth cranial face 48 9°3 5°0 4°6 Depth caudal face 5:0 8-1 5°3 5°1* Length proximal facet 673 I1'9 5:9 5°5* Width proximal facet or 14-7 84+ 77 Length distal facet 71 14'8 8-2 6:5 Width distal facet 75 13°6 8-0 75 Length posterolateral process 36 7:0 4:0 3:9 Width posterolateral process 5°0 6-3 — 3°5 Depth posterolateral process 76 9°5 — 7a Lunar. An incomplete lunar (M.21876), identified by MacInnes (1942) as the left lunar of Prodeinotherium hobleyi is almost certainly mastodont. However the right lunar of both P. hobleyi and Mastodon angustidens are present in the collections of the Centre for Prehistory and Palaeontology, Nairobi. The Prodeinotherium lunar (KNM-RU 12091) is triangular and almost complete although the lateral edges are somewhat irregularly preserved. The facet articulating with the radius occupies nearly the whole of the proximal surface and extends nearly to the caudal edge of the bone. In contrast, that of Mastodon terminates more cranially. The radial facet is convex in the anteroexternal corner but concave over the remainder of the cranial surface. The ulnar facet at the anteroexternal corner is rhomboid and not triangular as in Mastodon. The magnum facet on the distal surface is convex cranially and concave caudally. TABLE 16 Lunar measurements Parameter P. hobleyi D. giganteum M.angustidens E. maximus KNM-RU1i291 (H6wenegg) KNM-RU 1260 U.B. 20189 Maximum length 8-7 16:0 10°3 8-1 Maximum width 8-7 15°4 10°8 8-1 Maximum depth 5°8 10:8 71 53 Length radial facet 73 11-4 75 6-7 Width radial facet 73 13°6 8-5 6:0 Length magnum facet 707 139 9:0 71 Width magnum facet 8-0 1374 8-6 Tez Length ulnar facet 4-2 7:0 5°4 3°2 Width ulnar facet 3°55 4°3 4:0 2:9 Magnum. The right Prodeinotherium magnum from Rusinga (M.21877) is incompletely preserved but is more complete than a left magnum of Mastodon angustidens from Gebel Zelten (M.26652) that provides a basis for comparison. In Prodeinotheriwm the proximal surface of the magnum articulates with the lunar 338 PRODEINOTHERIUM and is L-shaped, a postero-medial projection (the foot of the L) standing proud. In Mastodon or Elephas this projection merges gradually with the body and is relatively less distinct. The proximal facet of the deinothere magnum is divided into two portions—an almost flat cranial surface for articulation with the cranial part of the lunar and a convex caudal region that occludes with the caudal part of the lunar and scaphoid. The scaphoid articulates with the caudomedial projection of the magnum. The distal articular facet is pear-shaped (stem posterior) and concave for articulation with the third metacarpal. Four facets may be discerned on the medial side. The cranioproximal, craniodistal and caudoproximal facets are tri- angular and denote points of contact with the trapezoid. A mediodistal facet extends anteriorly beneath the craniodistal facet to articulate with the second metacarpal. The lateral side of the magnum adjoins the unciform and four facets may be observed. Two large proximal facets are separated by a central groove in contrast to the situation in the elephantoids in which the facets are contiguous. The distal unciform facets on the magnum are also separated, in contrast to the situation in Elephas, but the craniodistal facet is poorly developed as in Mastodon. TABLE 17 Magnum measurements Parameter P. hobleyi D. giganteum = M.angustidens E.maximus M.21877 (H6wenegg) M.26652 U.B.20189 Maximum length 8°5 15°3 II‘2 8-7 Width cranial edge 5°7 10:2 6-9 For Width caudal edge 6:8 10:0 6:8 71 Depth cranial edge 57 10°4 6-7 6-2 Depth caudal edge TRG) 13°4 oo 78 Length proximal facet 78 14°7 9:2 76 Width anterior edge proximal facet 4°7 9-9 6:6 7X Width posterior edge proximal facet 6-6 10-0 6:8 71 Length distal facet 6:8 14°0 — 6-7 Width distal facet 4°5 6:2 53+ 4°5 Unciform. Three unciforms of Mastodon angustidens M.26653, M.26655, M.26657) are known from Gebel Zelten and contrast strongly in several features with the left Prodeinotherium unciform (M.21882); described by MacInnes (1942). In Prodeino- thertum, the dorsal facet for articulation with the cuneiform is roughly triangular inoutline. It reaches its maximum width at the anterior edge and tapers posteriorly. The cranial and caudal portions are convex but are separated by a transverse concave depression. The facet reaches its highest elevation at the caudomedial edge. The medial surface of the Prodeinotherium unciform bears a large proximal and two smaller distal facets for articulation with the magnum. Similar facets are also exhibited by Elepbhas and Mastodon. Three facets are apparent on the distal surface of the unciform. The medial facet is elongate, faces distomedially, is divided into two by a concave transverse groove and supports the lateral edge of the third metacarpal. The central facet is triangular (tapering posteriorly), is concave and FROM GEBEL ZELTEN, LIBYA 339 articulates with the fourth metacarpal. It is separated from a larger rectangular, convex facet for Metacarpal V by a sagittal groove that is deepest posteriorly. Adjacent to the caudolateral edge of the McV facet but extending onto the proximal surface of the bone is a small, triangular, concave facet for the pisiform. In the Gebel Zelten mastodont unciforms the cuneiform facet approximates to a rhomb and is longer than broad, the maximum breadth being achieved near the rear of the bone. In Elephas this facet is appreciably wider than long, the caudal edge being wider than the cranial, although the latter is relatively wider than in Mastodon. The distal facets of the Gebel Zelten mastodont unciforms are arranged much as in Prodeinotherium but are more elongate anteroposteriorly. These facets are of similar length in Prodeinotherium and the Indian elephant but in the latter the facets for McIV and McV are subequal in size and that for McIV is almost flat. TABLE 18 Unciform measurements Parameter P. hobleyi D. giganteum = M.angustidens E. maximus M.21882 (Howenegg) M.26655 U.B.20189 Maximum length 9:2 16-4 10-6 8-4 Maximum width 10°3 17°5 11-6 973 Maximum height 8-9 10:9 8:9 6:6 Length cuneiform facet 7:6 14°8 8-3 6°5 Width cuneiform facet 9:0 16:2 8:5 Ql Length McV facet 6:8 8-9 6:5 5°9 Width McV facet 4:2 53 5:0 39 Length McIV facet 6:7 12:2 7:0 5:2 Width McIV facet 4°7 8-7 6°5 54 Length McIII facet 51 1371 5°5 59 Width MclIII facet 1-6 3°6 I°3 2:0 Pelvis. One half of a pelvis (M.26667g) was associated with the immature deino- there skeleton from Gebel Zelten site D. It may have belonged with the juvenile skull (6404 : 44), the adult skull (6404 : 14) or neither. Unfortunately the pelvis was shattered during transportation from the site. Femur. A right Prodeinotheriwm femur (M.26667d) was associated with skull 6404: 44. It lacks both the head and the distal epiphysis but the shaft is approxi- mately the same length as that of the extant Indian elephant. The shaft is com- pressed craniocaudally and is virtually straight. The greater trochanter was developed much asin Mastodon or Elephas. The lesser trochanter is poorly preserved. The distal half of the cranial surface of the shaft bears a large ridge that becomes more swollen distally as it extends towards the patellar surface. This ridge is comparatively larger than in Elephas and suggests a larger area of attachment for the vastus intermedius muscle. The caudal surface of the Prodeinotherium femur, where preserved, strongly resembles that of the Indian elephant. 340 PRODEINOTHERIUM TABLE I9 Femur measurements Parameter Prodeinotherium Elephas hobleytz maximus M.26667d U.B.20189 Length shaft 78-7 80 Width shaft at proximal epiphysis 2575+ 22:9 Depth shaft at proximal epiphysis 12-24 10°7 Width shaft at distal epiphysis 20°8+ 16-4 Depth shaft at distal epiphysis 176+ 13°0 Width at centre of shaft II3 8-6 Depth at centre of shaft 9°6 65 Fibula. The distal third of the shaft of a Prodeinotheriuwm fibula (M.26667a) was associated with skull 6404 : 44. The anterointernal border or interosseous ridge is separated from the anteroexternal border in the proximal half of the specimen by a narrow and flat anterior surface. The two borders converge in the distal half and the external border fades out before reaching the distal extremity. The posterior surface is round and broad but forms a sharp ridge immediately above the distal extremity. The lateral face is convex outwards and presumably served for the attachment of the peroneous brevis muscle. The medial face is concave and afforded attachment for the flexor longus hallucis muscle. From the distal epiphysis the bone decreases in width proximally, the change being gradual along the anterior edge but with a fairly sharp genuflexion along the posterior edge just below the dorsal limit of the specimen. From the shape of the posterior edge of the Gebel Zelten specimen it would seem that the bone was originally about as long as that of the Indian elephant. It differs from the latter, however, in that the distal extremity of the Gebel Zelten specimen is half as wide again. Tarsal bones. MacInnes (1942) described an astragalus, two naviculars and an external cuneiform from Rusinga and attributed them to the taxon Prodeinotherium hobleyi. Comparison of the astragalus (M.21880) with mastodont astragali from Gebel Zelten and East Africa and with the astragalus facets of mastodont and deinothere tibiae confirms MacInnes’ identification. The external cuneiform and navicular differ both in size and in some morphological features from the equivalent bones of the Indian elephant but are likely to have belonged to Mastodon angustidens. Astragalus. The deinothere astragalus is more equant than that of Mastodon. The dorsal and medial surfaces bear a rectangular convex facet for articulation with the tibia. This facet is as long but less wide and more convex than that of M. angustidens. The anterior border of the Prodeinotherium astragalus from Rusinga is broken but the missing portion was probably of similar appearance to that of Mastodon except that the groove on the craniomedial surface, separating the tibial facet from the navicular facet, was less deep and less wide than in Mastodon. There are three facets on the distal surface—a triangular and convex navicular facet and two calcaneum facets. In Prodeinotherium the navicular facet extends to the craniomedial border of the astragalus, the medial edge of the facet being concave. The medial tip of the navicular facet of Mastodon is also concave but extends less FROM GEBEL ZELTEN, LIBYA 341 far medially. The groove separating the navicular facet from the lateral calcaneal facet is deeper but less wide than in Mastodon. The lateral calcaneal facet is tri- angular and is flatter than that of Mastodon. The caudomedial calcaneum facet is oval and concave. It is separated from the lateral facet by the continuation of the same groove that separates the latter from the navicular facet. The medial calcaneal facet is separated from the navicular facet only by a very shallow groove and extends posteriorly onto the under surface of the massive posteromedial projection. On the lateral surface of the astragalus the facet for articulation with the fibula appears to be as long as, but much wider than, that of Mastodon. TABLE 20 Astragalus measurements Parameter P. hobleyi D. giganteum = M. angustidens E. maximus M.21880 (H6wenegg) M.26658 U.B. 20189 Length lateral edge 79 — 74 7°4 Length medial edge I1l‘o 16:0 7:84 10:2 Maximum width Io:9+ 17°7 10-0* II°5 Depth medial edge pas 9°5 7X 7:9 Depth lateral edge 4°4 7:4 BZ 4°4 Length tibial facet 6-9 — 73+ 6-9 Width tibial facet FoR — 8-3 8-3 SKELETAL EVIDENCE ON THE LOCOMOTION OF DEINOTHERES The limb material of the Gebel Zelten deinotheres is too fragmentary to permit a detailed functional analysis. Nevertheless, comparison of these specimens with other Prodeinotheriwm and Deinotherium remains affords some information about the locomotion of these animals. The most important skeletal elements in this respect would appear to be the scapula, femur, manus and pes. Smith and Savage (1956) determined that in graviportal mammals the scapula tends to be high and broad to afford sufficient surface area for the serratus and other shoulder muscles. They noted that the scapula blade of graviportal mammals is frequently expanded posteriorly to increase the mechanical advantage of the main retractor muscle of the forelimb (the teres major) and that the spine is usually heavy and may bear a posteriorly projecting flap in the region overlying the infraspinous fossa (Rhinoceros, Brontops). Alternatively the spine may end in a large acromion (Loxodonta, Uintatherium). The trapezius muscle originates on the spine and metacromion, the deltoid on the acromion. Both these muscles aid in abducting the forelimb. Compared with the scapula of Elepbhas maximus, the scapula of Prodeinotherium hobleyi (text-fig. 14) is less tall (51 v 62 cm) and less broad (25 v 42cm). The caudal angle is sited more dorsally above the head of the scapula (25 v 15 cm) and is less elongate posteroventrally from the spine (26 v 30 cm). Because the caudal angle of the scapula is sited more dorsally in Prodeinotherium than in Elephas, the mechanical advantage of the teres muscle is thereby decreased. Smith and Savage (1956) have shown that the mechanical advantage of the teres muscle is equivalent to the length of the moment arm of the teres muscle divided by the perpendicular distance 342 PRODEINOTHERIUM from the glenoid to the ground. The mechanical advantage of the teres muscle of the Indian elephant is nearly 1/13 but that of P. hobleyi is 1/16. The supraspinous fossa of Prodeinotherium hobleyi is relatively smaller than that of Elephas, and the scapular spine rises less abruptly and more dorsally. An acromion process is developed on the spine of the P. hobley: scapula and is sited more dorsally than in Elephas. A metacromion is also present, sited level with the base of the acromion and forming a flap overhanging the infraspinous fossa. Both acromion and metacromion are reduced in comparison with those of the Indian elephant. Behind the metacromion the spine of P. hobleyi is broader and more robust than in the Indian elephant. The supraspinous fossa of D. gigantewm from Valladolid, Spain, is even more reduced than that of the Gebel Zelten deinothere. The caudal angle of the Deino- therium scapula is less posteroventrally produced than in the extant elephants and, significantly, there are no processes diverging from the relatively slender spine. The absence of acromion and metacromion was also noted by Stefanescu (1899) on the scapula of D. ‘gigantissimum’ from Manzati, Rumania. Reduction of the spine and supraspinous fossa, and the absence of acromion and metacromion would in artiodactyls be regarded as evidence of cursorial modification (Smith & Savage, 1956). The retention of a rigid body axis is essential in animals of elephantine bulk in order to support the weight of the body. The method of rapid locomotion is thus limited to a fast amble or trot. The more rapid action of the teres muscle of deino- theres when taken in conjunction with other features of the scapula does, however, suggest that Deinotherium was potentially capable of more rapid locomotion than the elephantoids. Such an interpretation is supported by other features of the appendi- cular skeleton and tallies with Tobein’s (1962) interpretation of the function of the Deinotherium manus and pes. Prodeinotheriwm, although more typically graviportal than Deinotherium, shows some signs of cursorial adaptation. The humerus of the Gebel Zelten Prodeinotherium must have been slightly shorter than that of the Indian elephant but was still appreciably longer than the ulna. The shaft of the humerus is compressed laterally and elongated anteroposteriorly in direct contrast to the condition in Elephas. The lateral epicondyle is shorter than in Elephas but projects farther laterally. The condyles of the distal epiphysis have a greater degree of curvature in Prodeinotherium than in Elephas suggesting that the ulna of P. hobleyi may have been capable of being flexed through a greater angle. The ulna of the Gebel Zelten deinothere is of similar length to that of the Indian elephant but is proportionately stouter. As in the elephantoids the nearly vertical alignment of the forelimb requires an olecranon process that diverges at right angles from the shaft. There is no evidence to suggest that the ulna was fused to the radius at its distal end although fusion at this point would seem likely. The distal epiphyses of both the ulna and radius were proportionately more massive in Deino- therium than in Prodeinotherium. The radius of Prodeinotherium is best known from Maboko Island, Kenya, that from Gebel Zelten being represented only by the head. Prodeinotheriwm radius M.15429 is again about the size of that of Elephas maximus. It is less straight proximo-distally and craniocaudally than in the elephantoid proboscideans. The FROM GEBEL ZELTEN, LIBYA 343 radial tuberosity is less well developed than in Mastodon angustidens and the interos- seous ridge diverges halfway down the shaft. The radius shaft of both Mastodon angustidens and P. hobleyi is relatively wide laterally and compressed craniocaudally, but, whereas the mastodont shaft retains this shape throughout its length, that of Prodeinotherium becomes progressively more equant distally. Only the large carpal bones of Prodeinotheriwm have been examined. These are serially arranged and are smaller but similar in shape to those of D. gigantewm from Hoéwenegg, Germany (Tobein, 1962). The morphology of the proximal surface of the cuneiform suggests that the medial malleolus of the ulna extended farther distally in deinotheres than in elephantoids. No metacarpals have been collected from Gebel Zelten but a limited amount of information about the metacarpals may be gleaned from the distal row of carpal bones. In the unciform the facet for McV is proportionately larger and less laterally orientated than in the Howenegg Deinotherium specimen and the facet for McII1issmaller. The magnum of Prodeino- therium bears a large facet for McIII and a smaller and more laterally aligned facet for McII. Ehik (1930) described metacarpal fragments of P. bacaricwm from Kotyhaza, Hungary. These were laterally compressed, but less so than the metacarpals of Deinotherium giganteum from Pikermi, Greece, described by Dietrich (1916). The metacarpals of P. bavaricum are less elongate than those of D. gigantewm and resemble more closely those of Elebhas. Whereas Dietrich (1916) and Tobein (1962) stated that the first metacarpal of D. giganteum is sufficiently reduced to suggest functional tetradactyly of the manus, Ehik alleges that in P. bavaricum the fifth metacarpal shows the greatest reduction. The distal articulation surface of the fourth meta- carpal of D. giganteum is concave but that of P. bavaricum is concavoconvex. Ehik (1930 : 12) implies that the manus of P. bavaricum was more plantigrade than that of D. giganteum. The metacarpals of D. gigantewm are dolichopodous (Tobein, 1962). Whereas the proximal epiphysis of the fourth metacarpal of D. gigantewm is orientated at right angles to the longitudinal axis of the shaft of the bone, the proximal epiphysis of McIV of P. bavaricum is aligned obliquely to the longitudinal axis of the shaft. The digits of P. bavaricum must therefore have spread farther laterally, were probably more uniform in size and are likely to have been less digitigrade than those of D. gigantewm. The trapezium and trapezoid bones of Prodeinotherium have not been examined and it is therefore impossible to say if the first metacarpal was reduced as in D. giganteum. The long bones of the hind limb of the deinotheres are similar in many respects to those of the elephantoids. The head of the femur is vertically aligned and, as in other graviportal mammals, the femur is longer than the tibia. Osborn (1929 : fig. 670) formulated the general rule that, relative to the length of the tibia, the femur of cursorial mammals shortens with increase in potential speed while the pes elements become more elongate. The femur and tibia of Prodeinotherium are of comparable lengths to those of Elephas maximus but the femur of D. gigantewm from Valladolid is relatively shorter with respect to the tibial length. The tibia/femur ratio of E. maximus is 0:55 whereas that of D. gigantewm is 0-71, indicating that the femur of 344 PRODEINOTHERIUM D. giganteum is about 30% shorter than that of the extant Indian elephant. Reduction of the first metatarsal in Deinotherium was noted by de Pauw (1908), and the narrow elongate pes of D. gigantewm in conjunction with the dolichopody of the digits convinced Tobein (1962) that D. gigantewm was capable of swifter locomotion than its contemporary elephantoids. The reduction in length of the Deinotherium femur would appear to support Tobein’s interpretation. The equant tibial facet of the astragalus of all deinothere taxa from which this bone is known is a distinctive character that reflects the comparatively narrow width of the pes of deinotheres compared to that of elephantoids. The tibial facet of Prodeinotherium astragali is more convex than in astragali of its contemporary elephantoids but that of Deinotherium is flatter. The large posteromedial projection from the Prodeinotheriwm astragalus is not present in either Mastodon or Elephas and is reduced in Detnotherium. The surfaces of the Prodeinotherium astragalus that articulate with the fibula and medial malleolus of the tibia are larger and more vertically orientated than in Mastodon angustidens. IV. CONCLUSIONS The family Deinotheriidae is known from Africa, Europe and western Asia. All representatives of this family are large. Features of the skull and dentition serve to distinguish the deinotheres from the elephantoids and distinctive characters may also be seen in the postcranial skeleton. On their dental characters the deinotheres appear to be most closely related to the barytheres and moeritheres and may have been derived from ancestral Barytherium stock during the early Tertiary. The family Deinotheriidae is represented by two genera, the primitive Prodeino- therium and the later and larger Deinotherium. The Gebel Zelten specimens are assigned to Prodeinotherium hobleyi. They include the earliest deinothere skulls yet recovered and the fine condition of preservation of the skulls permits their use for reinterpreting the morphology of less well preserved Deinotheriwm skulls. The Gebel Zelten specimens exhibit a number of primitive deinothere features including a steeply downturned rostrum, narrow and more anterior external nares, a skull roof that is proportionately wider and longer, less elevated occipital condyles, shorter paroccipital processes and a more vertically inclined occiput. The facial region of P. hobleyi was modified for the attachment of a proboscis as was that of Deinotherium. The cranial region was adapted for greater movement of the skull on and with the neck than in elephantoids. The most important modi- fications enlarge the paroccipital processes and elevate the occipital condyles to increase the downward thrust of the mandibular tusks. Reversed curvature of the zygapophyses of the cervical vertebrae (compared to that of elephantoids) also aids in the downward movement of the head. In detail the dental characters render the deinotheres unique. Loss of the superior pre-cheek teeth and retention of a single pair of downturned inferior tusks is a combi- nation not exhibited by any other mammal. The main function of the tusks may have been for self defence and feeding, both involving a violent downward thrust of FROM GEBEL ZELTEN, LIBYA 345 the head. It would seem unlikely that the tusks were used for digging and their main function in feeding was probably for clearing and stripping vegetation. Wear facets on the anteromedial surface of the tusk tip in a number of deinothere tusks would appear to support this interpretation. Lophodont cheek teeth have been evolved in several mammalian groups but the trilophodont anterior molars of the deinotheres are unique. The deciduous fourth premolar is also trilophodont and served as a shearing tooth in contrast to the crushing function of the second and third milk premolars. The anterior permanent molar initially serves as a shearing tooth but after the posterior molars have been erupted it becomes part of the anterior crushing battery. The second and third molars are bilophodont and persist as shearing teeth until the lophs are removed by wear. The angle of shear changes from nearly vertical in newly erupted teeth to almost horizontal in greatly worn teeth. The shearing surface is maintained by thegosis. There is no increase in hypsodonty or brachyodonty throughout the recorded history of the deinotheres, and the length versus width parameters of the teeth of all deinotheriid taxa remain constant. Deinothere vertebrae are unfortunately rare but the axis and atlas vertebrae are known. They may readily be distinguished from those of the elephantoids and are modified to provide large attachment areas for the muscles of the neck. The zygapophyses of the cervical vertebrae have reversed curvature relative to those of the elephantoids. The neural spines of the anterior thoracic vertebrae were ap- parently longer and more posteriorly inclined than those of contemporary elephan- toids. Differences from elephantoid structure may also be seen in the appendicular skeleton, the scapula and foot bones being most distinctive, the long bones less so. Advanced characters shown by Deinotherium include reduction of the scapular spine, elongation of the carpals, tarsals and metapodials, elevation in posture of the manus and pes and reduction of the first digit leading to the functional tetradactyly of the feet. Many of the limb modifications point to a cursorial adaptation of Deinotherium, but Prodeinotherium resembles Elephas in the size and proportions of the long bones. Prodeinotherium and Deinotherium are separated on the basis of size, minor modi- fications of the dentition and characters of the skull and postcranial skeleton. Six species are recognized: P. hobleyi, P. bavaricum, P. pentapotamiae, D. giganteum, D. indicum and D. bozast. Each species is based primarily on geographic distribu- tion. The conservative dental characters of deinotheres are of little use for specific differentiation but the skulls and skeletal elements may prove more diagnostic when these are better known. Deinotheres must have originated in Africa during the early Tertiary but are first known from the early Miocene. During the early Miocene Prodeinotherium migrated from Africa via the Middle East (and perhaps the western Mediterranean) into Eurasia. By the late Miocene Deinotheriumhadappeared. Possibly two species evolved separately, D. giganteuwm from P. bavaricum in Western Europe and D.indicum from P. pentapotamiae in western Asia. More probably Deinotheriwm was derived from P. bavaricum and then spread eastwards. Deinotherium and Prodeinotherium coexisted in Eurasia during the late Miocene. Pvodeinotheriwm had become extinct in Eurasia by the end of the early Pliocene and Deinotheriwm by the middle Pliocene. 346 PRODEINOTHERIUM P. hobleyi was present in East and North Africa during the early Miocene but was thereafter restricted to East Africa. It survived into the Pliocene but is not known to have coexisted with D. bozasit which had appeared by the beginning of the Pleisto- cene. Whether D. bozasi evolved in situ or represented the southerly migration of the Eurasian deinotheres cannot yet be determined but the former seems more likely in view of the limited distribution of African deinotheres after the early Miocene. Both Prodeinotherium and Deinotherium were browsing mammals that used their tusks and probosces to gather food. The proboscis of Prodeinotherium was smaller, and the skeleton lacked many of the cursorial adaptations of Deinotheriwm. Prodeinotherium probably inhabited more densely vegetated regions while Deinotherium frequented more open territory. There is no obvious successor to Deinotherium in the middle or late Pleistocene of Africa. Whether it failed to adapt to changing environmental conditions or was supplanted through competition from a group or groups of smaller herbivorous mammals must remain a matter for conjecture. V. REFERENCES ANDREWS, C. W. i191t. On a new species of Dinotherium from British East Africa. Proc. zool. Soc. London, Dec. 1911 : 943-945. 1921. A note on the skull of Dinotherium giganteum in the British Museum. Tb7d., Sept. 1921 : 525-534. ARAMBOURG, C. 1934. Le Dinotherium des Gisements de l’'Omo. C. R. Soc. géol. Fr. Paris, 1934 : 86-87. 1961a. Prolibytherium magniert, un Vellericorne nouveau du Burdigalien de Libye. Ibid, 1961, no. 3 : 61-62. 1961b. Note preliminaire sur quelques vertébres nouveaux du Burdigalien de Libye. Ibid, 1961, no. 4 : 107-108. 1963a. Le genre Bunolistriodon Arambourg 1933. Bull. Soc. géol. Fr. Paris, (7) 5 : 903- git. 1963b. Continental vertebrate faunas of the Tertiary of North Africa. In Howell, F. C. & F. Bourliere (Edit) African ecology and human evolution. Chicago, 1963 : 55-64. —— & P. MacnierR, 1961. Gisements de vertébres dans le bassin Tertaire de Syrte (Libye). C. R. Acad. Sci. Paris, 252 : 1181-1183. BERGouNIOUx, F. M. & F. CrouzELt. 1962a. Les Deinotherides d’Europe. Avnls. Paléont. Paris, 48 : 13-56. 1962b. Les Deinotherides d’Espagne. Bull. Soc. géol. Fy. Paris, (7) 4 : 394-404. 1963. Migrations des Mastodontes Miocenes. C.R. Congr. nat. Soc. Sav., Sect. Sct., Paris, 87 : 867-872. De BLAINVILLE, H. M. D. 1837. Note sur la téte de Dinotherium giganteum actuellement a Paris. C.R. Acad. Sci. Paris, 4, no. 12 : 421-426. Brives, A. 1919. Sur la decouvert d’une dent de Dinotherium dans la Sabliere du Djebel Kouif pres Tebessa. Bull. Soc. Hist. nat. Afy. N. Alger., 10 : 90-93. Craupius, M. 1865. Das Gehorlabyrinth von Dinotherium giganteum nebst Nemerkungen iiber der Werth der Labyrinthformen fiir die Systematik der Saugerthiere. Paldeonto- gvaphica, Stiittgart 13 : 65-74. Deum, R. 1963. Paldontologische und geologische Unterschuchungen im Tertiar von Pakistan: 3. Dinotherium in der Chinji-Stufe der Unteren Siwalik Schichten. Bayer. Akad. Wiss. Wien, (n.s.) 114 : 1-34. DietricH, W. O. 1916. Uber die Hand und Fuss von Dinotherium. Zeitschr. dt. geol. Ges. Berlin, 68 : 1-44. FROM GEBEL ZELTEN, LIBYA 347 Desio, A. 1935. Studi geologio sulla Cyrenaica, sul Deserto Libico, sulla Tripolitania e sul Fezzan Orientali. Mission Scient. R. Acc. d'Italia a Cufra (1931). Roma, 1. Doust, H. 1968. Palaeoenvivonment studies in the Miocene of Libya. Vol. 1 of unpub. Ph.D. thesis, Imperial College, London. Eutk, J. 1930. Pvodinotherium hungaricum N.g., N.sp., Geol. Hungarica Ser. Palaeont., fasc. 6, Jan. ’25 : I-24. FaLconerR, H. 1845. Description of some fossil remains of Dinotherium, Giraffe and other Mammalia, from the Gulf of Cambay, western coast of India. Q. Jl geol. Soc. Lond. 1 : 356— 37/2: Graf, I. E. 1957. Die Prinzipien der Artbestimmung bei Dinotherium. Palaeontographica, Stuttgart, Abt.A, 108 : 131-185. Hamitton, W. R. 1973. The Lower Miocene ruminants of Gebel Zelten, Libya. Bull. Br. Mus. nat. Hist. (Geol.), 21, 3 : 73-150 JAYNE, H. 1898. Mammalian anatomy part 1: the skeleton of the cat. Philadelphia, 816 pp. Kaup, J. J. 1829. Deinotherium giganteum. Isis (XXII), Jena, 4 : 401-404. Kuipstein, H. & Kaup, J. J. 1836. Beschreibung und Abbildungen von dem in Rheinhessen aufgefundenen colossalen Schedel des Dinotherii gigantet mit geognostischen Mittheilungen uber die knochenfuhrenden Bildungen des mittelrheinischen Tertiarbeckens. Darmstadt, 38 pp. LasKAREV, V. 1944. Uber die Dinotherien reste aus Serbien. Neues Jb. Min. Monatsch, Stuttgart, 1944, 2/3, abt.B : 67-77. Macinnes, D. G. 1942. Miocene and post-Miocene Proboscidea from East Africa. Tvans. zool. Soc. London, 25, part 2 : 33-106. Maenigr, P. 1962. Etude géologique du gisement de vertébres du Gebel Zelten (Libye). C.R. Soc. géol. Fy. Paris, 1962, fasc. 2 : 55-57. 1968. Etude lithostratigraphique du Miocene inferieure du Bassin de Syrte (Libye). Proc. Comm. Medt. Neog. Strat; Geol. Bologna, 35 : 119-130. von MEYER, H. 1833. Beitrage zur Petrefaktenkunde. Das Dinotherium Bavaricum, mit Ruck- sicht auf die riesenmassigste fossile Thiergattung der Dinotherien uberhaupt, und auf die Struktur der Mahlzahne in den Tapiren. Nova Acta Acad. Leop. Carol., Halle, 16 : 485-516. Osporn, H. F. 1907. Evolution of mammalian molar teeth to and from the triangular type. New York. 250 pp. 1929. The titanotheres of ancient Wyoming, Dakota and Nebraska. U.S. geol. Su. Washington, Monograph 55, 2 : 703-953. 1942. Proboscidea, volume 2. American Museum of Natural History: 803-1675. OwEN, R. 1866. On the anatomy of vertebrates, vol. II. London: 437-444. PACHECO-HERNANDEZ, F. H. 1930. Fisiografia, geologia y paleontologia del territorio de Valladolid. Comm. Invest. pal. y prehist., no. 37 (ser. pal. no. 9) : 1-192. PALMER, R. W. 1924. On an incomplete skull of Dinotherium, with notes on the Indian forms. Mem. geol. Surv. India, [Paleont. Indica], (n.s.) 7, mem. no. 4 : I-13. DE Pauw, L. F. 1908. Trois mois au Musée de Bucharest ou quelques considerations sur le Dinotherium de Manzati. Brussels. 35 pp. Rapinsky, L. 1967. Relative brain size; a new measure. Science, Feb. 1967, 155, no. 3674 : 836-838. SAHNI, M. R. & Tripatui, C. 1957. A new classification of the Indian deinotheres and description of D. ovlovii sp. nov. Mem. geol. Surv. India, [Paleont. Indica], 33, no. 4 : 1-33. SavaGE, R. J. G. 1967. Early Miocene mammal faunas of the Tethyan region. Syst. Assoc. Publ. no. 7 : 247-282. 1973. Megistotherium, gigantic hyaenodont from Miocene of Gebel Zelten, Libya. Bull. Br. Mus. nat. Hist. (Geol.), 22, 7 : 483-511 — & Hamitton, W.R. 1973. Introduction to the Miocene mammal faunas of Gebel Zelten. Libya. Bull. Br. Mus. nat. Hist. (Geol.), 22, 8 : 513-527 —— & WuitE, M.E. 1965. Two mammal faunas from the early Tertiary of central Libya. Proc. geol. Soc. Lond., no. 1623 : 89-01. 348 PRODEINOTHERIUM SELLEY, R. C. 1966. The Miocene rocks of the Marada and Jebel Zelten area, central Libya; a study of shoreline sedimentation. Guidebook Petr. Expl. Soc. Libya. 1966. 30 pp. —— 1967. Paleocurrents and sediment transport in near shore sediments of the Sirte Basin, Libya. J. Geol., Chicago, 75; no. 2 : 215-223. 1968. Near shore marine and continental sediments of the Sirte Basin, Libya. Proc. geol. Soc. Lond., no. 1648 : 81-90. SuHinpbo, T. & Mori, M. 1957. Musculature of the Indian Elephant Part III: Musculature of the trunk, neck and head. Ak. Fol. Anat. Jap., 28 : 18-39. Sisson, S. & GRossMAN, J. D. 1947. The anatomy of the domestic animals. 3rd edit. revised. New York. 972 pp. SLIJPER, E. J. 1946. Comparative biological-anatomical investigations on the vertebral column and spinal musculature of mammals. Kon. Ned. Akad. Wet. Verh. (Tweedie Sectie), 42, no. 5 : 1-128. SmitH, J. MAYNARD & SAvaGE, R. J. G. 1956. Some locomotory adaptations in mammals. J. Linn. Soc. London (zoo), 62, no. 288 : 603-622. STEFANESCU, G. 1897. Dinotherium gigantissimum Stef. trouve a Munztac. Cong. geol. Int. (7th session), 1897, St. Petersburg: cxc—cxci. 1907. Quelques mots sur le Dinotherium gigantissimum Stef. trouve a Munztac. Cong. geol. Int. (10th session), 1906, Mexico, 1 : 427-430. Stott, N. R. ef al. 1961. International Code of Zoological Nomenclature adopted by the XVth International Congress of Zoology. London. 176 pp. STROMER, E. 1938. Huftier-Reste aus dem unterspliocanen Flinzsande Munchens. Bayer. Akad. Wiss. Wien, Abh. A, 44 : 1-39. ToBEIN, H. 1962. Uber Carpus und Tarsus von Dionotherium giganteum Kaup. Palaeont. Z., H. Schmidt-Festband, Nov. 1962 : 231-238. J. M. Harris, B.Sc., Ph.D. CENTRE FOR PREHISTORY NAIROBI East AFRICA PLATE 1 Fic. a. Prodeinotherium hobleyi skull (M.26665), anterior view. Fic. b. Prodeinotherium hobleyi skull (M.26665), dorsal view. Bull. By. Mus, nat. Hist. (Geol.) 23, 5 PLATE 1 Ta tb IGN, 2 Fic. a. Prodeinotherium hobleyi skull (M.26665), left lateral view. Fic. b. Prodeinotherium hobleyi skull (M.26665), right lateral view. Bull. Br. Mus. nat. Hist. (Geol.) 23, 5 IPILAN IIE 2 2b E* PLATE 3 Fic. a. Prodeinotherium hobleyi skull (M.26665), posterior view. Fic. b. Prodeinotherium hobleyi skull (M.26665), ventral view. Bull. By. Mus. nat. Hist. (Geol.) 23, 5 PLAT Ie 3 PLATE 4 Fic. a. Prodeinotherium hobleyi mandible (6412 : 10), right lateral view. Fic. b. Prodeinotherium hobleyi mandible (6412 : 10), dorsal view. Bull. Br. Mus. nat. Hist. (Geol.) 23, 5 PLATE 4 4a TINIE, 5 Fic. a. Prodeinotherium hobleyi mandible (6404 : 13), dorsal view. Fic. b. Prodeinotherium hobleyi right maxilla (6401 : 4) ventral view. Bull. By. Mus. nat. Hist. (Geol.) 23, 5 PLATE 5 A LIST OF SUPPLEMENTS TO THE GEOLOGICAL SERIES OF THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) . Cox, L. R. Jurassic Bivalvia and Gastropoda from Tanganyika and Kenya. Pp. 213; 30 Plates; 2 Text-figures. 1965. 6. . EL-Naccar, Z. R. Stratigraphy and Planktonic Foraminifera of the Upper Cretaceous—Lower Tertiary Succession in the Esna-Idfu Region, Nile Valley, Egypt, U.A.R. Pp. 291; 23 Plates; 18 Text-figures. 1966. {r10. . Davey, R. J., Downtz, C., SARGEANT, W. A. S. & Wittiams, G. L. Studies on Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 248; 28 Plates; 64 Text- figures. 1966. £7. . APPENDIX. Davey, R. J., Downig, C., SARGEANT, W. A. S. & WiLLIaMs, G. L. Appendix to Studies on Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 24. 1969. 8op. . Evtiotr, G. F. Permian to Palaeocene Calcareous Algae (Dasycladaceae) of the Middle East. Pp. 111; 24 Plates; 17 Text-figures. 1968. £5.12. . Ruopes, F. H. T., Austin, R. L. & Druce, E. C. British Avonian (Carboni- ferous) Conodont faunas, and their value in local and continental correlation. Pp. 315; 31 Plates; 92 Text-figures. 1969. {11. . Caitps, A. Upper Jurassic Rhynchonellid Brachiopods from Northwestern Europe. Pp. 119; 12 Plates; 40 Text-figures. 1969. £4.75. . Goopy, P. C. The relationships of certain Upper Cretaceous Teleosts with special reference to the Myctophorids. Pp. 255; 102 Text-figures. 1969. £6.50. . Owen, H. G. Middle Albian Stratigraphy in the Paris Basin. Pp. 164; 3 Plates; 52 Text-figures. 1971. {6. . Sippigui1, Q. A. Early Tertiary Ostracoda of the family Trachyleberididae from West Pakistan. Pp. 98; 42 Plates; 7 Text-figures. 1971. £8. « BC Printed in England by Staples Printers Limited at their Kettering, Northants establishment _ CIRRIPEDES FROM THE UPPER -*°"".. CRETACEOUS OF ALABAMA — Lem “REGION, USS.A. I. PALAEONTOLOGY | J. S. H. COLLINS >) GEOLOGY eee 'B, MELLEN Vol. 23 No. 6 LONDON : 1973 GULE REGION, ..U.S.A. IL PALAEONTOLOGY BY JOSEPH STEPHEN HENRY COLLINS Dulwich, London Pp 349-380; 5 Plates, 3 Text-figures Il. GEOLOGY BY FREDERIC FRANCIS MELLEN Jackson, Mississippi, Seen el Pp 381-388; 2 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) GEOLOGY Vol. 23 No. 6 LONDON : 1973 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) instituted in 1949, 1s issued tn five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 23 No. 6 of the Geological (Palaeontological) series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. World List abbreviation Bull. Br. Mus. nat. Hist. (Geol.). © Trustees of the British Museum (Natural History), 1973 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued 13 July, 1973 Price £2.25 GIRRIPEDES FROM THE UPPER CRETACEOUS OF ALABAMA AND MISSISSIPPI, EASTERN GULF REGION, U.S.A. PART I. PALAEONTOLOGY By J. S. H. COLLINS CONTENTS Page INTRODUCTION é : : : : 5 ‘ : . = 351 HISTORICAL REVIEW . 5 : : 5 : : 6 : 3 352 SYSTEMATIC DESCRIPTIONS : ‘ A : . ° : . 354 ACKNOWLEDGEMENTS ‘ : : ‘ : : 5 > 5 386 REFERENCES . 5 . : : : 5 . . . ° 387 SYNOPSIS From the Upper Cretaceous Selma Chalk and Ripley Formation of Mississippi and Selma Chalk of Alabama, U.S.A. valves of nine new species and one new subspecies of cirripedes are described and figured: Cretiscalpellum harnedi, C. macrum, C. vallum, C. venustum, Arcoscal- pellum bakeri, A. campus, A. hubrichti, A. witherst, Brachylepas angulosa, Virgiscalpellum gabbi apertus. Further valves belonging to Virgiscalpellum gabbi (Pilsbry) and two other Virgiscal- pellum valves are described and figured from the Ripley Formation. They occur in silty, finely sandy and highly foraminiferal and ostracodal horizons within the ‘Selma’ which represents an ‘outer neritic’ environment of the continental shelf. Three biostratigraphic horizons are recognized: (1) zone of Vivgiscalpellum of Ripley and basal Prairie Bluff (Maastrichtian) age; (2) zone of Cretiscalpellum vallum of basal Annona or upper Coffee (Campanian) age; and (3) zone of Arcoscalpellum hubrichti of middle Mooreville (Santonian) age. These three zones supplement known faunal zones within the Upper Cretaceous. INTRODUCTION BETWEEN 1935 and 1938 F. F. Mellen corresponded with T. H. Withers of the British Museum (Natural History) and sent numerous small collections of cirripede valves to him for study and for description in a paper Withers had planned. The new forms recognized were received too late for inclusion in the monographic Catalogue of Fossil Cirripedia, Vol. 2, Cretaceous, 1935. The holotypes were to be retained in the British Museum (N.H.) and paratypes were to be returned to the Department of Geology and Geography at Mississippi State University. The intervention of World War II, the retirement of Withers and his death in 1958, prevented the completion of the study of the Mississippi barnacle collections which, according to present tabulation, numbered 146 valves from seven localities (see below). Bull. By. Mus. nat. Hist. (Geol.) 23, 6 352 CRETACEOUS CIRRIPEDES Between 1967-1969 many valves were secured from previous collections and additional hundreds were collected by F. F. Mellen and others, not only from those of the earlier sites that still existed, but from new localities scattered over several Counties in Alabama and Mississippi. A further group of 36 valves, from Greene County, Alabama, were lent to J. S. H. Collins by the Paleontological Research Institute, Ithaca, New York. HISTORICAL REVIEW The present material contains only one, Virgiscalpellum gabbi (Pilsbry), of the eight species of cirripedes previously recorded from the Cretaceous of the United States of America. Of the other seven species, the generic position of one, ‘Scalpel- lum’ inaequiplicatum Shumard 1862, from the Upper Senonian of Texas, was con- sidered by both Shumard and Withers (1935) as provisional, owing to the poor pre- servation of the specimen. The inadequate description of Scalpellum sp. listed by Adkins 1928, from the Upper Austin Chalk of Texas, was also considered to be equivocal by Withers (1935). Of the remaining species, Loriculina ?texanum Withers is recorded from the Middle Albian of Texas; Calantica (Titanolepas) martini Withers, Stramentum haworthi (Williston) and Squama spissa Logan are known from the Senonian of Kansas. Stvamentum canadensis (Whiteaves) (Withers, 1935) has been described from the Fort Benton group of Manitoba and recently Russell (1967) added a further Canadian species, ?Calantica saskatchewanensis, from the Bearpaw Formation of southwestern Saskatchewan. Arcoscalpellum conradi (Gabb, 1876) from the Vincentown Formation of New Jersey, must now be regarded as a Palaeocene form (Richards, 1958), which substantiates Withers’ (1935) opinion that the scutum and carina of this resembled certain Tertiary species. Students and collectors of the rich fauna of Upper Cretaceous sediments of Missis- sippi, Alabama and Tennessee made great use of Wade (1926) in the identification of fossils, despite its limitation to basal Ripley strata of Coon Creek, McNairy County, Tennessee. It was soon recognized that Virgiscalpellum gabbi, first recorded from Coon Creek, was present in the Ripley strata of Oktibbeha County, Mississippi, and that other barnacle types were also present. The collections to be considered afford an aggregate of some 1493 valves ; from these 46 different valves have been distinguished and are included in the following species: Cretiscalpellum harnedt sp. nov. C. macrum sp. nov. C. vallum sp. nov. C. venustum sp. nov. Arcoscalpellum bakeri sp. nov. A. campus sp. nov. A. hubrichti sp. nov. A. withersi sp. nov. Virgiscalpellum gablt gabbi (Pilsbry) V. gabbi apertus ssp. nov. V. sp. V. sp. Brachylepas angulosa sp. nov. ALABAMA & MISSISSIPPI 353 This material not only allows nine new species and one new subspecies to be des- cribed, but also includes the first member of the sub-order Brachylepadomorpha, to be recorded from the Western Hemisphere. The new material of V. gabbi allows us to add the scutum, tergum, carinal and rostral latera to the known valves of that species. Moreover, these valves show that V. g. gabbi is related to the Maastrichtian species V. hagenowianum (Bosquet) which occurs only at Maastricht. Both Pilsbry (1933) and Withers expressed doubts whether or not the upper latus, figured by Withers (1935, p. 39, fig. 2, erroneously called a scutum by Wade, 1926) should be ascribed to V. gabbi. In his notes on the valves present in the original collections sent to the British Museum (Nat. Hist.), Withers had considered that this valve really belonged to the species here named A. withers: but in view of the additional material in the present collections, which bear a striking similarity to the upper latus of V. darwinianum (Bosquet) figured by Withers (1935, pl. 38, fig. I) there seems little to support this opinion. Until such time as more material becomes available, these particular upper latera are here retained, with reservation, in V.g. gabbi. The present material from the Ripley Formation of Mississippi shows that the carina of V. gabbi is very variable in structure and carinae similar to the one Withers set aside in his notes as a new species are here regarded as a subspecies of V. gabbi. Also, a distinct scutum and tergum ostensibly from the Ripley Forma- tion and Annona chalk respectively are described, but not named. These anomalies indicate that much remains to be learned of the genus Virgiscalpellum, particularly from the American Cretaceous. It is felt that extensive new collections are desirable to allow a re-study of the genus to be made. The valves of Arcoscalpellum hubrichti have characters in common with both of the European Groups of A. maximum (J. de C. Sowerby) and A. fossula (Darwin) ; both A. campus and A. withersi have affinities with the Group of A. fossula. Of the four species of Cretiscalpellum present, C. vallum is the best known; not only is it represented by numerous carinae, but also by eight other different valves. C. macrum and C. venustum, the geologically oldest species of the genus in the present collection, are known from only fifteen and four valves respectively, while C. harnedi, the youngest of the genus in this study, is represented by thirteen valves. Withers (1927, 1935, 1953) and Cheetham (1963) have fully discussed the mor- phology and classification of fossil lepadomorph cirripedes and the latter has also provided a comprehensive survey of North American Tertiary species. Following the practice adopted by Darwin, a carina, when present, has been selected as holotype. ABBREVIATIONS The following abbreviations relating to Museum and other collections have been used in the text: BMNH, British Museum (Natural History); MSU, Mississippi State University; PRI, Paleontological Research Institution. 354 CRETACEOUS CIRRIPEDES SYSTEMATIC DESCRIPTIONS Series CIRRIPEDIA Burmeister, 1834 Order THORACICA Darwin, 1851 Suborder LEPADOMORPHA Pilsbry, 1907 Family SCALPELLIDAE Pilsbry, 1916 Genus CRETISCALPELLUM Withers 1922 Cretiscalpellum Withers : (9), 9, 374- 1935 Cretiscalpellum Withers : 2, 144. Diacnosis: Scalpellidae with seventeen valves including four pairs of large and little differentiated, much overlapping lower latera; subcarina much larger than rostrum ; carina with umbo apical. TYPE SPECIES: C. unguis (J. de C. Sowerby). RANGE: Aptian (Lower Greensand) to Maastrichtian. Cretiscalpellum macrum sp. nov. (Plate 1, Figs 1-6) Diacnosis: Carina thin, moderately arched with fine apico-basal ridge, acute basal angle. Tergum and scutum thin, tergum with carinal margin divided into upper and lower portions. HoLotypPe: A carina. BMNH In.64414. (PI. 1, fig. 1), Upper Senonian, Moore- ville Chalk; 3 miles E. of West Greene and 3 miles W. of Clinton, Greene County, Alabama. MATERIAL: 15 valves: British MusEUM In.64414. Carina. Holotype. Pl. 1, fig. 1. 385 ACKNOWLEDGEMENTS : c . : : 2 - 5 : 386 REFERENCES . : 5 : . : : . : o 387 STRATIGRAPHY Hitherto, cirripedes from the ‘Selma chalk’ have all but been ignored stratigraphi- cally in favour of other more easily recognized macroscopic and abundant micro- scopic index fossils. It was thought that cirripedes might be of biostratigraphical importance from the taxonomic study of the Mississippi/Alabama collections. The horizons of the various species were tabulated and it was seen that they are ap- parently naturally grouped with no inter-horizon overlapping. The cirripedes of the ‘Selma chalk’ may be summarized as follows: Ripley Formation—Navarroan (Maastrichtian) : C. harnedi, A. bakeri, A. campus, A. withersi, V. g. gabbi, V. g. apertus, B. angulosa. Basal Annona—Tayloran (Campanian): C. vallum. Middle Mooreville—Austinian (Santonian): C. macrum, C. venustum, A. hubrichiz. The genus Virgiscalpellum is restricted to the Ripley and the basal part of the over- lying Prairie Bluff. This is overlain almost directly by deposits which contain Diploschiza mellent, a small bivalve used as an index fossil in the Prairie Bluff. Two new species of Arcoscalpellum and one of Cretiscalpellum found with Virgiscalpellum also appear to be restricted to this horizon. The Ripley Formation of Oktibbeha County, Miss., seems to be the most prolific cirripede unit yet described from North America. An attempt has been made to zone this unit on the basis of its cirripede fauna, but other than the one specimen of B. angulosa, from the mid-Ripley Barr Pasture Locality (M4) there seems to be no significant variation, apart from an apparent increase in abundance of all forms in the middle and upper parts of this unit. C. vallum appears to have a restricted vertical range. It is found in deposits of outer neritic sediments at the base of the Annona chalk (which is the equivalent of the upper part of the Coffee Sands) some 20-30 ft below the lowest range of Diplo- schiza cretacea, an important zonal indicator. A. hubrichti and two new species of Cretiscalpellum seem to be restricted to a zone which lies near the mid portion of the 250 ft thick Mooreville chalk (marl). Bull. Br. Mus. nat. Hist. (Geol.) 23, 6 Cc 382 CRETACEOUS CIRRIPEDES Approximately 250 ft of chalk, and chalky marl separates the zone containing Virgiscalpellum from that with C. vallum, while about 200 ft of chalky marl separates the zone containing C. vallum with that of A. hubrichti. From these observations it is clear that there are three distinct cirripede biostrati- graphic ‘zones’ within the ‘Selma’. These, designated in descending order, are: Zone of Virgiscalpellum (Ripley and basal Prairie Bluff—Maastrichtian). Zone of Cretiscalpellum vallum (basal Annona or upper Coffee—Campanian). Zone of Arcoscalpellum hubrichti (middle Mooreville—Santonian). Electrical Log MABEN OIE WATER WELL BIOSTRATIGRAPHIC (PALEOCENE) TERTIARY DIPLOSCHIZA MELLE NAVARRO aie GROUP ONE OF V/RGISCALPELLUM MAASTRICHT/IAN EXOGYRA CANCELLATA TAYLOR | coonewax 1 GROUP BEC * |$DIPLOSGHIZA GRETACEA ‘ANNONA nf = NS = Q = x G <4} CRETISCALPELLUM VALLUM MOORE- ARCOSCALPELLUM AUSTIN VILLE HUBRICHTI SANTONIAN Fic. 4. Correlation and biostratigraphic zones in ‘Selma chalk’ represented on a selected electrical log, Maben Water Well, SE. 4+ of SW. 4 of NW. 4 of Section 31, T. 20 N., R.12E., Oktibbeha County, Mississippi. Elevation 466 feet. Position of Zone of Cvretiscalpellum vallum is interpreted at 1000 feet (—534). This Zone crops out at elevation 200 feet on south valley wall of Tibbee Creek, NE. } of SE. 4 of Sec. 6, T. 19 N., R. 16 E., Clay County, 244 miles to the east. The westerly dip of 734 feet is at an average rate of 30 feet per mile, slightly less than the rate of ‘true’ dip, which is west-southwesterly, about 35 feet per mile. ALABAMA & MISSISSIPPI 383 GEOGRAPHICAL RANGE: The zone of Virgiscalpellum has been established over an outcrop distance of 155 miles from Coon Creek, McNairy County, Tennessee, to south central Noxubee County, Mississippi. Although most, if not all, of the six species recorded appear possibly to be restricted to this zone, the designation is chosen because V. gabbi (Pilsbry) was first reported (by Wade) from the lower Ripley Coon Creek beds of Tennessee; also, this species is the dominant one throughout the Ripley of Oktibbeha County, Mississippi. The zone of Cretiscalpellum vallum has been established over an outcrop distance of 564 miles from Tibbee Creek, Clay County, Miss., to one mile west of Gainesville, Sumter County, Alabama. The zone of Arcoscalpellum hubrichti has been established at many localities in an outcrop belt about two miles wide in Greene County, Alabama; in 1970, Dr E. E. Russell (Mississippi State University) located the species 4 miles west of Columbus, Lowndes County, Miss. (M12), in the Mooreville, in a position estimated to be 50 ft above its base, thus extending the lateral outcrop over a distance of 65 miles between EUROPEAN TEXAS MISSISSIPPI ALABAMA STAGES PRAIRIE BLUFF PRAIRIE BLUFF maastricurian| NAVARRO ee GROUP McNAIRY RIPLEY RIPLEY GOON CREEK CAMPANIAN NCOR DEMOPOLIS DEMOPOLIS GROUP = = ARCOLA ARCOLA ANTONIAN Aafecsre MOOREVILLE MOOREVILLE CONIACIAN RODE EUTAW EAGLE FORD EAGLE FORD TER OMIAS GROUP (LOWER EUTA =" U. TUSCALOOSA GORDO WOODBINE | CENOMANIAN GROUP |M. TUSCALOOSA POKER EOLINE L. TUSCALOOSA! ¢ ot ToNDALE Fic. 5. Correlation chart showing European stages, Texas Group terminology and Upper Cretaceous stratigraphic nomenclature in Mississippi and Alabama. No attempt was made to show range of hiatuses in Texas or Alabama or in the Mississippi subsurface. GAS ROCK = x = is} = Ly ry AUSTIN 384 CRETACEOUS CIRRIPEDES the States. Barnacle valves were also found at Cochrane, Pickens County, Alabama, some 15 miles distant from West Green. Recent search has not re-discovered this particular locality, but as Cochrane is on Mooreville terrain, it is probable that the valves recorded were of A. hubrichtt. Valves of A. hubrichti outnumber those of the accompanying species of Cretiscalpellum by a ratio of 200 : I. A detailed examination of cores and electrical loggings are needed for greater refinement in the stratigraphical placing of the zones of C. vallum and A. hubrichtz. Cirripedes of the ‘Selma chalk’ may well be found in sediments of widely varying bathymetric depths, but the present collection seems to have been preserved best in deposits consisting of thin-shelled molluscs, comminuted shell fragments, fine sands and silts. It is suggested that these deposits were formed on outer neritic bottoms of the continental shelf—in depths of 200-600 ft. A notable exception is V. gabbi from highly glauconitic and extremely fossiliferous marls of the basal Ripley, Coon Creek beds of McNairy County, Tennessee, which indicates very shallow marine conditions. The various formations composing the ‘Selma chalk’ are tabulated for Alabama, Mississippi and Texas, and correlated with the European Stages (Fig. 5). Below the ‘Selma’ the lower and upper Eutaw are shallow marine and estuarine sediments and would probably yield a cirripede fauna with intensive collecting. Beneath, the sediments of Tuscaloosa Age are almost entirely continental except in the deepest sub-surface. CRETACEOUS DEPOSITION All the present cirripede collections were made from deposits on a platform area lying north of a tectonic hinge which marks the northern limit of the Mississippi Interior Salt Basin. The structure of these deposits indicate a massive step by step transgression of the Mesozoic seas from the south. During the late Upper Cretaceous the area of north Mississippi, west Tennessee, west Kentucky, southeast Missouri and northeast Arkansas was covered by marine water for the first time since the retreat of the shallow Upper Carboniferous (Pennsylvanian) seas, Alleghany Revolu- tion (Rodgers, 1967). The Appalachian Mountain Chain to the north and northeast and the Nashville- Ozark and Ouachita Uplifts to the north and northwest were probably the proven- ances of the sediments. The initial dip (Paleoslope) of the strata of ‘Selma’ age was in a south easterly direction, except in central Alabama and eastwards, although a period of intense volcanism reached its climax during late Taylor or early Navarro times in west- central Mississippi, southeast Arkansas and northeast Louisiana. This produced the highly uplifted and truncated Jackson Dome and the Sharkey-Monroe Uplift from which sediments and igneous detritus have been eroded. Scattered biotite flakes, pyroclastic granules and pebbles, and bentonite deposits in the shallow water Tuscaloosa, Eutaw, Coffee and Ripley Formations which became increasingly abundant near the volcanic uplifts, would seem to be evidence of the importance of contributions of the late Mesozoic volcanism to the geological history of the area. ALABAMA & MISSISSIPPI 385 It is further suggested that the axis of the Mississippi Embayment and its lateral Desha Basin which extends west into southeast Arkansas were a subsidence caused by the initial uplift of the Jackson Dome and the Sharkey-Monroe Uplift to the west. LOCALITIES AND COLLECTIONS MISSISSIPPI ‘Original Collections’ Cirripede valves, from seven localities and totalling 146, were from the Ripley Formation of Oktibbeha County (140) and from the Tayloran chalk of Clay County (6). Of these, 81 were collected by H. H. Harned, Jr., and 65 by Mellen. At the time these collections were made there were no good maps of the area, consequently the collections were designated: OC.1. 2 miles E. of State College. (Probably NW of NW}, Sec. 5, T. 18 N., R. 15 E., Oktibbeha County.) OC.2. } mile E. of State College (H.H.H., Jr.) (Probably Bardwell Pasture, S.W. corner NEi of NW}, Sec. 6, T. 18 N., R. 15 E., Oktibbeha County.) OC.3. 4 mile E. of State College. (Probably near Centre W4 of NW} Sec. 6, T. 18 N., R. 15 E., Oktibbeha County.) OC.4. Catalpa Creek. (NW34 of NE#t of Sec. 25, T. 18 N., R. 14 E., Oktibbeha County: basal Prairie Bluff.) OC.5. 5 miles NW. of State College (Stoney Point). (SW4 of SWH of Sec. 16, T. 19 N., R. 14 E., Oktibbeha County: basal Prairie Bluff and Upper Ripley.) OC.6. Tibbee Creek (44 miles S. of West Point). (NE# of SEH, Sec. 6, T. 19 N., R. 16 E., Clay County: basal Annona or upper Coffee.) OC.7. 14 miles E. of State College—Barr Pasture (H.H.H. Jr.) (Probably N.W. corner of NE+ of NW} of Sec. 6, T. 18 N., R. 15 E., and other outcrops lying to the northwest and northeast.) Serial numbers have been included to simplify locality recordings in the text. Other collections made by Harned during the 1930’s were included in material sent directly to Collins later, are marked ‘Barr and Bardwell Pastures’, ‘Barr’, and ‘Sand Creek and Barr’. It would be imprecise to attempt to assign more specific location descriptions to the old collections due to the fact that some of the old chalk gullies have been completely filled and obliterated in recent years. MISSISSIPPI ‘Recent Collections’ Mr. ‘Oktibbeha County, from Dunn Seiler Museum’—exact localities unknown, but all undoubtedly from the Ripley of the MSU campus area. Mz. Barr & Bardwell Pastures: NE} of NW} Sec. 6, T. 18 N., R. 15 E., Oktibbeha County (H. H. Harned, Jr., Coll.) (middle Ripley). M3. Barr Pasture: N.W. Corner of NE} of NW} of Sec. 6, T. 18 N., R. 15 E., Oktibbeha County (H. H. Harned, Jr. Coll.) (middle Ripley). M4. Barr Pasture: cf. ‘M3’. (middle Ripley) Ms. Bardwell Pasture: S.W. Corner of NE} of NW} of Sec. 6, T. 18 N., R. 15 E., Oktib- beha County (H. H. Harned, Jr. Coll.). These gullies have been filled and com- pletely obliterated. (middle Ripley) Ato. AIl. CRETACEOUS CIRRIPEDES Sand Creek & Barr: Probably SW Sec. 31, T. 19 N., R. 15 E. & N.W. Corner of NE} of NW Sec. 6, T. 18 N., R. 15 E. Oktibbeha County (H. H. Harned, Jr. Coll.). (middle & lower Ripley) 6 Miles N.W. of Starkville (H. H. Harned, Jr. Coll.) Locality indeterminable, but prob- ably upper Ripley. N. of Evans Hall: NW} of NW3 Sec. 1, T. 18 N., R. 14 E., Oktibbeha County. Chapel Hill Church locality of S. valley wall Catalpa Creek, 5 miles S. & 1-75 miles E. of Barr Pasture locality, NE} of NE Sec. 32, T. 18 N., R. 15 E., Oktibbeha County. Trim Cane—Josey Creeks confluence; SW of SW} Sec. 20, T. 19 N., R. 14 E., Oktib- beha County. 4% miles S. of West Point, S. valley wall of Tibbee Creek, E. of U.S. Highway, 45—W.: NE of SE} Sec. 6, T. 19 N., R. 16 E., Clay County. (basal Annona or upper Coffee.) Cuts in Highway 82 bypass, c. 4 miles W. of Columbus, Secs. 26 & 27, T. 19 N., R. 17E. Lowndes County. ALABAMA Approx. S.W. corner SE} of NE} Sec. 27, T. 23 N., R. 1 W., on public road (through W. M. Steele land), 14 miles N. of West Greene, Greene County. c. 1 mile N.E. of ‘Ar’, c. SW of SEH, Sec. 23, T. 23 N., R. 1 W., Greene County. c. 14 miles N.W. of ‘Ar’, near centre NW3 Sec. 22, T. 23 N., R. 1 W., Greene County. 2 miles E. of West Greene, 4 miles N. of Clinton, near Centre SW Sec. 36, T. 23 N., R. 1 W., Greene County. 3 miles E. of West Greene & 3 miles W. of Clinton, c. miles N.E. of church, c. Centre SW3 Sec. 31, T. 23 N., R. 1 E. Greene County. Mt. Olive Church, around the common quarter corner, Secs. 30 & 31, T.22 N.,R.1E., Greene County. 1:2 miles due E. of Mt. Olive Church and in SW} of SW} Sec. 32, T. 22 N., R.TE., Greene County. I mile S.E. of Mt. Olive Church locality; 14 miles N.E. of Mt. Olive Church, north of Highway, Greene County. Shallow washes in brushy area W. of paved road 44 miles S.S.W. of Eutaw in SW} of NE# of Sec. 20, T. 21, N., R. 2 E., Greene County. 1 mile W. of Gainesville, NE}? Sec. 10, T. 21 N., R. 2 W. Sumter County. ‘Bonanza’, SW of SW}, Sec. 32, T. 22 N., R. 1 E., Greene County. ACKNOWLEDGEMENTS The writers are greatly obligated to many people for assistance in the collecting, preservation and study of the cirripedes covered in the present work: to Professors Troy J. Laswell, Ernest E. Russell and Donald M. Keady of the Department of Geology and Geography at Mississippi State University for assistance in collecting and encouragement of student participation, and for affording a repository for barn- acle collections and literature in the Dunn-Seiler Geology Museum; to Andrew Martin, Terry Christian, William Honey and John Kline for contributed speci- mens; to Stephen Hook who collected with one of us (F.F.M.) and made numerous other contributions during 1968; to Horace H. Harned, Jr., and Wentworth V. Harned, who, as students of geology at Mississippi State University, collected and supplied considerable material from the Ripley of Oktibbeha County: to ALABAMA & MISSISSIPPI 387 Leslie Hubricht, of Meridian, Miss., collector of the type material of Arcoscalpellum hubrichti and to Dr K. V. W. Palmer for kindly allowing access to this material; to the Keeper of Palaeontology, British Museum (Natural History), for making available Withers’ notes and the original material; to Dr W. T. Dean, R. Baker (retired), S. F. Morris and W. R. Smith, also of the British Museum (Natural History) for kindly allowing access to specimens in their charge, and to C. W. Wright, of London, for allowing us to examine specimens in his collection and for much valuable advice. We wish to record our warmest thanks also to Dr James W. Ward of the Anatomy Department of the University of Mississippi School of Medicine and Drs Gordon Gunter and J. Y. Christmas of Gulf Coast Research Laboratory, who furnished specimens of recent cirripedes and expressed much interest in this work; to Dr Wayne E. Moore, Division Paleontologist of Chevron Oil Company, for his critical review of the geological parts of this paper; to the Photographic Department, British Museum (Natural History) and to E. Kentish, of London, who prepared the photo- graphs. Support for extensive field work and other geological studies by Mellen has been afforded by Germany-Gulmon Oil Company; the Mississippi Geological Survey; and by the Mississippi Research and Development Center under its Grant Number 1-68-70. Also, many palaeontologists and stratigraphers were of assistance and encouragement during the investigation. REFERENCES Apxins, W. S. 1928. Handbook of Texas Cretaceous Fossils. Univ. Texas Bull. 2838, 1-304, pls., I-37. BRAUNSTEIN, J. 1950, 1959. Subsurface Stratigraphy of the Upper Cretaceous in Mississippi. In Cretaceous of Mississippi and South Tennessee: Eighth Field Trip Guidebook, Mississippi Geol. Soc.; In Upper Cretaceous Outcrops Northeast Mississippi and West Central Alabama: Fourteenth Field Trip Guidebook, Mississippi Geol. Soc., Jackson. Capitan, W. M. 1954. Subsurface Geology and Oil and Gas Possibilities of Northwestern Arkansas; Bull. geol. Surv. Avk., Little Rock, 20. CHEETHAM, A. H. 1963. Gooseneck Barnacles in the Gulf Coast Tertiary. J. Paleont., Tulsa, Oklahoma, 37 : 393-400, pl. 46. Coiiins, J. S. H. 1965. Arcoscalpellum comptum (Withers), a cirripede new to the Gault. Palaeontology, Lond., 8 : 629-633, pl. go. Darwin, C. R. 1851. A Monograph on the Fossil Lepadidae, or Pedunculated Cirripedes of Great Britain. vi—88 pp., 5 pls., Palaeontogr. Soc. Monogy., London. Fisk, H. N. 1944. Geological Investigation of the Alluvial Valley of the Lower Mississippi River. Mississippi River Comm. Grouskorr, J. G. 1955. Subsurface Geology of the Mississippi Embayment of Southeast Missoun: Rep. Mo. geol. Surv. Wat. Res., Rolla, 37, Sec. Ser. Hoek, P. P.C. 1907. The Cirripedia of the Siboga Expedition: Siboga Exped. Rept., 314, Cirripedia Pedunculata, 127 pp., Io pls. MELLEN, F. F. 1958. Cretaceous Shelf Sediments of Mississippi, Bull. Miss. geol. Surv., Mississippi, 85. Monroe, W.H. 1941. Notes on Deposits of Selma and Tipley Age in Alabama. Bull. geol. Suvv. Ala., Montgomery, 78. 388 CRETACEOUS CIRRIPEDES PirsBry, H. A. 1916. The Sessile Barnacles (Cirripedia) contained in the Collection of the U.S. National Museum; including a Monograph of the American Species. Bull. U.S. nat. Mus., Washington, 93 : xii—366 pp., 76 pls., 99 text figs. 1933. An unusual Cretaceous Cirriped. Science, 77 : 1994, 283-284. Pirssry, H. A. & Orssen, A.A. 1951. Tertiary and Cretaceous Cirripedia from Northwestern South America. Pyvoc. Acad nat. Sci. Philad., 103 : 197-210, pls. 6-11. Pryor, W. A. 1960. Cretaceous Sedimentation in Upper Mississippi Embayment. Bull. Am. Ass. Petrol. Geol., Chicago 44 : 9. RAINWATER, E. H. 1963. Geological History and Oil and Gas Possibilities of Mississippi; In Mississippi Geologic Research Papers—1962, Bull. Miss. geol. Surv., 97. Ricuarps, H. G. 1958. The Cretaceous Fossils of New Jersey. Bur. Geol. and Topogr., Trenton, N.J., Pt. 1., vi-266, 46 pls. RODGERS, J. 1967. Chronology of Tectonic movements in the Appalachian Region of Eastern North America, in Symposium on the Chronology of Tectonic movements in the United States. Amer. Jour. Sct., New Haven, 265, no. 5, pp. 408-27, table. RUSSELL, L. S. 1967. A Pedunculate Cirripede from Upper Cretaceous Rocks of Saskat- chewan. J. Paleont., Tulsa, Oklahoma, 41 : 1544-1547, 5 text figs. SHUMARD, B. F. 1862. Descriptions of new Cretaceous fossils from Texas. Pyvoc. Boston Soc. Nat. Hist., 8 : 188-205. STEPHENSON, L. W. & Monror, W.H. 1940. The Upper Cretaceous Deposits (of Mississippi). Bull. Miss. geol. Surv., 40. STEPHENSON, etal. 1942. Cretaceous Correlation. Bull. geol. Soc. Amey., New York 53 : 435- 448. Wave, B. 1926. The fauna of the Ripley Formation on Coon Creek, Tennessee. U.S. Geol. Surv., Prof. Paper, 137, 1-272, 62 pls. WitTHeErRS, T. H. 1928-1953. Catalogue of Fossil Civripedia in the Department of Geology, British Museum (Natural History) : 1928 Triassic and Jurassic. xii-154, 12 pls.; 1935 Cretaceous. xvi-434, 50 pls.; 1953 Tertiary. xv—396, 64 pls. London, British Museum (Natural History). 1945. New Cretaceous Cirripedes and Crab. Ann. Mag. nat. Hist., London, Ser. 2, 5 : 552-561, pl. 2. J. S. H. CoLiins FREDERIC F. MELLEN 63, OAKHURST GROVE 1202 STANDARD LIFE BUILDING E. DuLwicH JACKSON Lonpon, S.E.22. MISSISSIPPI 39201 DESsAr PLATE 1 Cretiscalpellum macrumi sp. nov. Page 354 Upper Senonian, Mooreville Chalk. Fic. 1. Carina. Outer view. Holotype. In.64414. A5 : Greene County, Alabama. Fic. 2. Carina. Outer view. A8. Paratype. MSU 1311. Fic. 3. Scutum (right). Outer view. Ai. Paratype. In.64415. Fic. 4. Tergum (right). Outer view. A4. Paratype. In.64416. Fic. 5. Tergum (left). Outer view. A8. Paratype. MSU 1312. Fic. 6. Tergum (left). Outer view. A8. Paratype. In.64417. Fics. 1-6 x2 diam. Cretiscalpellum venustum sp. nov. Page 356 Lower or middle Mooreville Chalk, Upper Senonian. Fic. 7. Carina. a, outer view; b, side view. Holotype. In.64418. A7 : Greene County, Alabama. Fic. 8. Carina. Outer view. A8. Paratype. MSU 1313. Fic. 9. Tergum (left). Outer view. A8. Paratype. In.64419. Fics. 7-9 x2 diam. Cretiscalpellum harnedi sp. nov. Page 357 Maastrichtian, Middle Ripley Formation. Fic. 10. Scutum (right). a, outer view; b, inner view. M8 Paratype. In. 64420. Fic. 11. Carina. a, side view; b, outer view; c, inner view. Holotype. In.64421. M3: Oktibbeha County, Mississippi. Fic. 12. Tergum (right). a, outer view; b, inner view. M8. Paratype. In.64422. Fics. 10-12 x4 diam. Cretiscalpellum venustum sp. nov. Page 350 Fic. 13. Transverse section of carina, MSU 1313 at apex, c. x8 diam. Cretiscalpellum vallum sp. nov. Page 358 Upper Senonian, basal Annona or upper Coffee Formation. Fic. 14. Carina. a, outer view; b, inner view, x2 diam. Mir. Paratype. MSU 1316. Fic. 15. Carina. a, outer view; b, side view, x2 diam. Holotype. In.64423. Murr: Clay County, Mississippi. Fic. 16. Carina, with small bivalve on the outer surface, x2 diam. Mur. Paratype. MSU 1355. Fic. 17. Transverse sections of carina, In.64423, at a, c. 7-0 mm and J, c. 14-0 mm from apex. c. x5 diam. Fic. 18. Tergum (right). Outer view, x2diam. M11. Paratype. In.64426. Fic. 19. Scutum (left). Outer view, x4 diam. Mir. Paratype. MSU 1317. Fic. 20. Scutum (left). a@, outer view; 6, inner view, x2 diam. M11. Paratype. In. 04425. Fic. 21. Upperlatus. Outer view, x4diam. M11. Paratype. MSU 13109. Fic. 22. Subcarina. a, inner view; b, outer view, x2 diam. Mrr. Paratype. In.64424. Bull. By. Mus. nat. Hist. (Geol.) 23, 6 PEATE 1 10b 12b Ila ie) Ile | a 15a 17b 20a 20b D* 22a 22b PLEAD E, 2 Cretiscalpellum vallum sp. nov. Page 358 Fic. 1. Subcarina. Outer view, x4 diam. M1r. Paratype. MSU 1320. Fic. 2. Tergum (left). a, inner view; b, outer view, x2diam. Mir. Paratype. In.64428. Fic. 3. Tergum (left), with small bivalve on the outer surface, x2 diam. Mir. Paratype. In.64435. Fic. 4. Tergum (left). Outer view, x4diam. M11. Paratype. MSU 1318. Fics 5-6. Carinal latera (left & right). Outer views, x4 diam. M11. Paratypes. In. 64433-34. Fic. 7. Rostral latus (left). Outer view, x2 diam. Mir. Paratype. In.64429. Fic. 8. Subcarinal latus (right). Outer view x4 diam. Mir. Paratype. In.64430. Fic.9. Inframedian latus. Outer view, x4diam. Ato. Paratype. In.64431. Fic. 10. Upper latus. Outer view of juvenile valve, x6 diam. OC.6. Paratype. In. 64432. Fic. 11. Upperlatus. Outer view, x4 diam. Mii. Paratype. In.64427. Fic. 12. Subrostrum. Outer view, x4 diam. Mr1r. Paratype. I[n.64483. Arcoscalpellum hubrichti sp. nov. Page 363 Upper Senonian, Mooreville Chalk. Fic. 13. Carina. a, outer view; b, side view; c, inner view, x2 diam. At : Greene County, Alabama. Holotype. - 6072PRI. Fic. 14. Carina. a, outer view; b, side view, x2 diam. A6. I[n.64438. Fic. 15. Carina and associated left tergum which has become orientated through 180°. Outer view, natural size. Az. MSU 1358. Fic. 16. Scutum (right). a@, outer view; b, inner view, x2 diam. A7. Paratype. In. 64436. Fic. 17. Scutum (left). Outer view, x2 diam. A7. Paratype. In.64437. Fic. 18. Scutum (left). Outer view, x2 diam. A7. Paratype. MSU 1325. Fic. 19. Upper latus (right). Outer view, x4diam. A3. Paratype. MSU 1359. Fic. 20. Transverse sections of carinae; at a, 1o-o mm & b, 15-0 mm from the apex, c. x 4 diam. Paratypes. 6082a—6082bPRI. Bull. By. Mus. nat. Hist. (Geol.) 23, 6 ILA WIS, l6a-b 20a-b PLATE 3 Arcoscalpellum hubrichti sp. nov. Page 363 Fic. 1. Tergum (right). Outer view, x2diam. Ar. Paratype. 6074PRI. Fic. 2. Tergum (right). a, outer view; 6, inner view, X2diam. A3. MSU 1324. Fic. 3. Tergum (right). Outer view, x2 diam. Ar. In.64439. Fic. 4. Upper latus (left). Outer view, showing well developed growth beyond the umbo, x2 diam. A7. Paratype. In.64443. Fic. 5. Upper latus (right). Outer view, x2 diam. A7. Paratype. In.64444. Fic. 6. Carinal Jatus (right). Outer view, x2diam. A7. Paratype. In.64440. Fic. 7. Carina. Side view of apical part, x2diam. Az. Paratype. 60873aPRI. Fic. 8. Carinal latus (right). a, outer view; b, inner view, x2 diam. A7. Paratype. In.64441. Fic. 9. Carinal latus (left). Outer view, x2diam. A7. Paratype. In.64442. Arcoscalpellum bakeri sp. nov. Page 367 Maastrichtian, Ripley Formation. Fic. 10. Carina. a, outer view; b, side view; c, inner view, x2 diam. My4 : Oktibbeha County, Mississippi. Holotype. In.64446. Fic. 11. Transverse sections of carina taken at, a, c. 5-o mm and Bb, c. 18-0 mm from the apex x5 diam. Paratype. In.64448. Fic. 12. Carina. a, outer view; b, inner view, x4 diam. M3. Paratype. In.64447. Fic. 13. Tergum (right). Outer view, x4 diam. Mz. Paratype. In.64449. Arcoscalpellum campus sp. nov. Page 369 Maastrichtian, Middle Ripley Formation. Fic. 14 Carina. a, outer view; b, side view, x2 diam. M5 : Oktibbeha County, Missis- sippl. Holotype. In.64452. Fic. 15. Scutum (left). a, outer view; b, inner view, x4 diam. M5. Paratype. MSU 1328. Fic. 16. Carinal latus (right). a@, outer view; b, inner view, x4 diam. M4. Paratype In.64453- Arcoscalpellum withersi sp. nov. Page 371 Maastrichtian, Ripley Formation. Fics 17-18. Rostral latera (left & right). Outer view, x4 diam. M5. Paratypes. In.64454-64455. Bull. By. Mus. nat. Hist. (Geol.) 23, 6 PAs 3 PLATE 4 Arcoscalpellum campus sp. nov. Page 369 Fic. 1. Scutum (left), Outer view, x4 diam. M5. Paratype. In.64450. Fic. 2. Tergum (left). a, inner view; b, outer view, x4 diam. M5. Paratype. In.64451. Fic. 3. Tergum (left). Outer view, x4 diam. M4. Paratype. MSU 1329. Fic. 4. Transverse section of carina, In.64452, at c. 4.0 mm from the apex, x6 diam. Arcoscalpellum withersi sp. nov. Page 371 Fic. 5. Carina. Outer view, x2 diam. OC.7. Paratype. MSU 1330. Fic. 6. Carina. a, outer view; b, side view; c, inner view, x2 diam. M3: Oktibbeha County, Mississippi. Holotype. In.64456. Fic. 7. Transverse sections of carina, at, a, c. 6-0 mm and 8, c. 11-0 mm from the apex, x5 diam. M3. Paratype. In.64463. Fic. 8. Scutum (right). Outer view, x2 diam. M4. Paratype. In.64459. Fic. 9. Scutum (right). a, outer view; b, inner view, x2 diam. M3, Paratype. MSU 1331. Fic. 10. Tergum (right). Outer view, x2diam. M2. Paratype. MSU 1332. Fic. 11. Tergum (left). Outer view, x2 diam. M3. Paratype. In.64457. Fic. 12. Tergum (left). Outer view, x2 diam. Mz. Paratype. MSU 1425. Fic. 13. Scutum (left). Outer view, x2 diam. M4. Paratype. MSU 1426. Fic. 14. Carinal latus (right). a, outer view; b, inner view, x4 diam. M4. Paratype. In.64461. Fic. 15. Upper latus (right). Outer view, x4diam. M4. Paratype. In.64462. Virgiscalpellum gabbi gabbi (Pilsbry) Page 374 Maastrichtian, Ripley and basal Prairie Bluff Formations. Fic. 16. Upper latus (right). Outer view, x4 diam. OC.1. In.64474. Fic. 17. Carina. a, outer view; b, side view, x2diam. M8. MSU 1334. Fic. 18. Carina. a, side view; b, inner view, x2diam. OC.1. In.64464. Bull. By. Mus. nat. Hist. (Geol.) 23, 6 PLATE 4 PLATE 5 Virgiscalpellum gabbi gabbi (Pilsbry) Page 374 Fic. 1. Tergum (left). a@, outer view; b, inner view, x2 diam. M2. In.64472. Fic. 2. Tergum (left). Outer view, x2 diam. Mz. MSU 1335. Fic. 3. Scutum (right). a, outer view; b, inner view, x2diam. Mz. Paratype. In.64467. Fic. 4. Tergum (left). Outer view, x2diam. OC.2. In.64469. Fic.5. MRostrallatus (left). Outer view, x4diam. M3. Paratype. MSU 1337. Fic. 6. Rostral latus (left). Outer view, x2 diam. M3. Paratype. In.64475. Fic. 7. Scutum (left). Outer view, x2 diam. Mz. Paratype. In.64470. Fic. 8. Carinal latus (right). Outer view, x4 diam. OC.1. Paratype. In.64473. Fic. 9. Transverse sections of carinae: a, open lower limb 4:0 mm from umbo, In.64465; infilled lower limb 8-0 mm from umbo, In.64466. 5 diam. M2. Fic. 10. Carina. a, side view; b, inner view, x2 diam. With the umbo and part of the lower limb infilled. M8. MSU 1416. Fic. 11. Scutum (right). Outer view, x2 diam. Mz. MSU 1336. Virgiscalpellum sp. Page 379 Upper Senonian, basal Annona or upper Coffee Formation. Fic. 12. Scutum (right). Outer view, «x4 diam. OC.6: Clay County, Mississippi. In. 64476. Virgiscalpellum sp. Page 379 Maastrichtian, Upper Ripley Formation. Fic. 13. Tergum (left). Outer view, x4 diam. OC.3 : Oktibbeha County, Mississippi. In.64477. Virgiscalpellum gabbi gabbi (Pilsbry) Page 374 Fic. 14. Carina. a, outer view; b, side view; c, inner view, x2 diam. With thickened lateral margins. OC.1. In.64487. Virgiscalpellum gabbi apertus sub-sp. nov. Page 378 Maastrichtian, Ripley and basal Prairie Bluff Chalk. Fic. 15. Carina. a, outer view; b, side view; c, inner view, x2 diam. OC.4 : Oktibbeha County, Mississippi. Holotype. In.64478. Brachylepas angulosa sp. nov. Page 380 Maastrichtian, Ripley Formation. Fic. 16. Rostrum. a, outer view; b, side view; c, inner view, x4 diam. M4 : Oktibbeha County, Mississippi. Holotype. In.64479. Bull. By. Mus. nat. Hist. (Geol.) 23, 6 PLATE 5 A LIST OF SUPPLEMENTS TO THE GEOLOGICAL SERIES OF THE: BULLETIN. OF THE BRITISH MUSEUM (NATURAL HISTORY) . Cox, L. R. Jurassic Bivalvia and Gastropoda from Tanganyika and Kenya. Pp. 213; 30 Plates; 2 Text-figures. 1965. {6. . Et-NaceGar, Z. R. Stratigraphy and Planktonic Foraminifera of the Upper Cretaceous—Lower Tertiary Succession in the Esna-Idfu Region, Nile Valley, Egypt, U.A.R. Pp. 291; 23 Plates; 18 Text-figures. 1966. fro. . Davey, R. J., Downtz, C., SARGEANT, W. A. S. & Witttams, G. L. Studies on Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 248; 28 Plates; 64 Text- figures. 1966. £7. . APPENDIX. DAVEY, R. J., Downtg, C., SARGEANT, W. A. S. & WILLIAMS, G. L. Appendix to Studies on Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 24. 1969. 8op. ie Se . Ettiott, G. F. Permian to Palaeocene Calcareous Algae (Dasycladaceae) of the - Middle East. Pp. 111; 24 Plates; 17 Text-figures. 1968. {5.12%. . Ruopes, F. H. T., Austin, R. L. & Druce, E, C. British Avonian (Carboni- ferous) Conodont faunas, and their value in local and continental correlation. Pp. 315; 31 Plates; 92 Text-figures. 1969. {1I1. . Cuitps, A. Upper Jurassic Rhynchonellid Brachiopods from Northwestern , Europe. Pp. 119; 12 Plates; 40 Text-figures. 1969. £4.75. . Goopy, P. C. The relationships of certain Upper Cretaceous Teleosts with — special reference to the Myctophorids. Pp. 255; 102 Text-figures. 1969. 6.50. . OwEN, H. G. Middle Albian Stratigraphy in the Paris Basin. Pp. 164; 3 Plates; 52 Text-figures. 1971. {6. . Sippigui, Q. A. Early Tertiary Ostracoda of the family Trachyleberididae from West Pakistan. Pp. 98; 42 Plates; 7 Text-figures. 1971. £8. 0-05) between the spacing of carinal nodes on the two kinds of branches. Cross-sections of some heavily encrusted branches showed that although the reverse surface may be smoothly rounded and without striae, secondary laminae FROM FERMANAGH, IRELAND 487 within the branch become progressively more corrugated towards the axial region. This tendency reaches its maximum at the contact with the strongly ribbed under side of the basal plate, a part of the primary skeleton. The ribbing seen in cross- sections is merely a reflection of the striated or longitudinally ridged condition in solid specimens, and it therefore seems that although the presence or absence of ‘striae’ on the reverse may be an indication of the age, or the state of weathering of a specimen, it is unlikely to be of taxonomic value. Genus SEPTOPORA Prout, 1859 DiaGnosis. Zoarium a fenestrate, flabellate or leaf-like expansion; primary branches numerous; increasing by bifurcation or interpolation; the lateral branches unite with those from the adjacent primary branches; apertures in two rows on primary and lateral branches; reverse usually with fine striae and scattered dimorphic pores (Nickles & Bassler 1900: 41). TYPE SPECIES. S. cestviensis Prout 1859. Mississippian, Illinois. Septopora hibernica sp. nov. (Pl. 26, figs 1-8) MATERIAL. Only ten specimens were available. They are not in such good condition as those of other species, being encrusted with secondary tissue to an unusual degree, so that surface features are commonly obscured. Some branches lack the entire apertural face, presumably owing to partial replacement, so that only the base and sides remain. It is possible that all specimens belonged to a single colony, and this should not be overlooked in using the following data for comparative purposes. The largest fragment was 42 x 17mm. PD.4885-91, PD.5336-8. Measurements (N = 10): B € D LAE I-20-1°50 1-400 0-100 7-14 F.w. I*50-2:00 1-640 0-156 g°51 L.ap.s. 0:35-0:40 0°375 O-O1I 3:00 B.w. 0:72-0:87 0-770 0-043 5°58 (primary) B.w. 0:45-0:67 0-515 0:070 13°59 (secondary) Note: Data for zooecial apertures are not included as these structures are com- monly broken, abraded, or so reduced by coarse grained silicification that original dimensions are not evident. Micrometric formula: 6—8/7—9//12—14/none. Apertures per fenestrule. Range of specimen modes: 3 Distribution of specimen modes: 10 DiacGnosis. Septopora of unknown zoarial shape, lacking carinal nodes and ‘dimorphic pores’. With well developed keel on primary, but not secondary branches. The latter have a variable growth habit, causing fenestrule shapes to be G 488 VISEAN FENESTRATE BRYOZOA irregular. Primary branches may show a Ptylopora-like pattern and bear two rows of small, closely-spaced zooecial apertures. Chamber base shapes triangular or hemi-hexagonal with short lateral walls. Structural data as above. TYPE SPECIMENS. Holotype: PD.4885. Paratypes: PD.4886 to PD.48q0. DESCRIPTION. There is no certain evidence of zoarial shape though, judging from the structure of available fragments, it was probably a more or less upright, foliaceous or fan-shaped growth. Two of the larger fragments have a pronounced Ptylopora-like branch pattern and are only distinguishable from that genus by the cross-bars (secondary branches) which bear zooecia. Primary branches are straight or gently flexed. They are stout and somewhat flattened on the obverse, which bears a prominent, rib-like median keel. In shallow, longitudinal grooves on either side of this are two rows of zooecial apertures. The reverse is broadly rounded and longitudinally striated, the striae serving to separate a number of ridges, each of which bears on its crest a row of minute, rounded papillae. These are about 0-01 mm in diameter and spaced about 0:04 mm apart. From eight to ten parallel ridges are visible on the reverse of a branch. Secondary branches (corresponding to dissepiments in the Fenestellidae) are variably developed, and their original shapes are commonly obscured by heavy secondary deposits. Where two unite in the chevron pattern characteristic of Septopora, they leave the main branches at angles of about 50°. This arrangement is uncommon in the present material in which, owing to space restriction and irregularities of growth, others predominate. Close to the junction of primary branches secondaries may be short and straight, rather like stout fenestellid dissepi- ments. In other cases a secondary appears to have developed independently of its partner, forming an oblique cross bar between adjacent primary branches. Bow- shapes are of fairly common occurrence, the angular junction between components having been rounded by later schlerenchymal growth. In other examples irregular shapes have resulted from asymmetrical unions between opposing secondary branches. Well developed secondaries have two rows of alternating apertures, but in many cases the arrangement is irregular. Some appear to lack apertures alto- gether, but this is probably due to sealing by schlerenchymal overgrowth rather than to a genuine absence of zooecia. Secondary branches do not have a median keel and the cross-section is approximately circular. The reverse shows the same characteristics as in primaries. Fenestrules in Septopora are typically chevron-shaped with rounded extremities, but these are rare in the present specimens. The irregular growth of secondary branches inevitably results in many variations of fenestrule shape, which may be round, oval, rectangular or even triangular. Variable amounts of secondary encrustation cause still further diversity and commonly convert regular to irregular shapes. The relatively small size of fenestrules and their variable configuration might almost be said to characterize the specimens examined. Zooecial apertures are in two alternating rows close to the keel and away from branch margins. They may be circular, with a diameter of about 0-15 mm, or (more commonly) oval, measuring about 0-225 by 0-125 mm. Some are relatively small FROM FERMANAGH, IRELAND 489 and more than their own diameter apart; others are much larger and separated by only a thin bar. Much depends on the amount of secondary encrustation present, for there is no doubt that this has in many cases reduced the size of apertures or even sealed them altogether. On the other hand, many have been enlarged by breakage or incomplete replacement of their rims. Apertures on both kinds of branches were probably of about the same size. Few zooecial base shapes are distinguishable in the specimens, but it is possible that there is a slight difference in this respect between primary and secondary branches. Those of the former seem to have hemi-hexagonal shapes with a length of about 0-32 mm and maximum width of 0-2mm._ The short lateral walls are about 0-05 mm long. On secondary branches chambers appear to be of similar size but triangular, with the maximum length (parallel to the branch margin) about 0-3 mm and maximum width about 0:25 mm. Only three chamber bases of each kind were seen, however, and it is uncertain whether the observed difference is general or not. Discussion. This is the first positive record of Septopora in the British Isles, though Whidborne (1895, p. 183) mistakenly assigned a poorly preserved specimen (probably of Ptiloporella) to the genus. Nor has Septopora been found in other European countries, with the exception of Russia where a number of species have been reported in recent years (Nikiforova 1938; Shishova 1952, 1957; Morozova 1955). The genus was introduced by Prout in the United States, and several species were described by Ulrich (1890). A possible reason for the previous absence of Septopora from faunal lists is that, due to lines of weakness provided by the angular junctions between secondary branches, the colonies were particularly liable to breakage. The resultant fragments typically consist of a primary branch with a number of secondaries on each side, all broken at the junction angle. Such fragments are exceedingly difficult to distinguish from Penniretepora stipes (always abundant in Carboniferous bryozoan faunas) and could easily have been described as such in the literature. Among described forms of Septopora the present specimens are closest to the type species, which is from approximately the same horizon (lower Chester) in Illinois. Common features include a clear median keel on the obverse, about three zooecial apertures per fenestrule and the absence of small accessory apertures (‘dimorphic pores’). There are also important differences, however, S. cestriensis having a rather finer mesh (ten or eleven branches and fenestrules in 10 mm, according to Prout), carinal nodes and apertural peristomes. Ulrich (1890; 628) recorded the species from the lower and middle Chester of Illinois and Kentucky, but his speci- mens had an even finer meshwork than those of Prout. He also mentioned the presence of accessory pores on both obverse and reverse, a feature not shown by Prout’s specimens or the present ones. The average branch width of Ulrich’s material was 0-5 mm, considerably less than that of the Irish specimens, though this might be accounted for by the thick secondary encrustation of the latter. Shishova (1952; 162) found, in the Dinantian of the Moscow region, specimens that she assigned to S. cestriensis. She stated their formula as: 14/103—114//22-223, which differs widely from that of the Fermanagh specimens. Also her material had 490 VISEAN FENESTRATE BRYOZOA lozenge-shaped (rhomboidal) chamber bases, and there were accessory pores on the reverse. The only other described species which the present specimens resemble is S. subquadrans Ulrich, from the upper Chester. This differs from S. cestriensis mainly in fenestrule shape, and has the formula: 7-12/10}—12//21. The number of apertures per 5 mm is greatly in excess of that shown by the Fermanagh specimens, however, and there are accessory apertures on both surfaces (Ulrich 1890; 6209). In view of the lack of correspondence with existing species it is clearly necessary to introduce a new name for the present material. S. hibernica seems appropriate for the first recorded occurrence of the genus in Ireland. Other Genera Note: In addition to the above forms, three further fenestrate bryozoans belonging to the Carrick Lough fauna have been described separately elsewhere by the author. In each case the account given follows the basic pattern used here. The additional species and location of their descriptions are: 1. Ptilofenestella carrickensis gen. et sp. nov. In 1965, Palaeontology 8 : 478-401. 2. Muinilya nodulosa (Phillips). In 1965, Geol. Mag. 102 : 135-142. 3. Polypora stenostoma sp. nov. In 1971, Palaeontology 14 : 178-187. “REFERENCES Bass_erR, R. S. 1935. In: Fosstlium Catalogus, 1: animalia. Ed. W. Quenstedt. Pars 67: Bryozoa, 1-229, s’Gravenhage. 1953. Bryozoa: in Tveatise on Invertebvate Paleontology, ed. Moore: i-xili, G1—253. Kansas. BELL, W. A. 1929. The Horton—Windsor District, Nova Scotia. Mem. geol. Surv. Can. 155 : 25-84, pl. 12-13. Brunton, C. H. C. 1966. Silicified productoids from the Visean of County Fermanagh. Bull. Br. Mus. nat. Hist. (Geol.), London 12 : 173-243, pls 1-109. CAMPBELL, K.S.W. 1961. Carboniferous fossils from the Kuttung rocks of New South Wales. Palaeontology, London 4 : 428-474, pls 53-63. Canu, F. 1900. Révision des bryozoaires du Crétacé figurés par d’Orbigny. Bull. géol. Soc. Fr., Paris (3), 28 : 334-463. CuRonIic, J. 1949. In: Newell, N. D., Chronic, J., and Roberts, T. G., The Upper Paleozoic of Peru. Mem. geol. Soc. Am., Washington 58 : 1-241, 44 pls. Conpra, G. E. & Exias, M. K. 1944b. Study and revision of Archimedes (Hall). Geol. Soc. Am. Spec. Pap. 53 : i-vili, 1-243, 41 pls. Cumincs, E. R. 1904. Development of some Paleozoic bryozoa. Am. J. Sci., New Haven 17 : 49-78. 1906. Description of the bryozoa of the Salem Limestone of southern Indiana. Rep. Dep. Geol. Nat. Res. Indiana, Indianapolis 30 : 1274-1296, pls 27-40. Dawson, J. W. 1878. Acadian Geology (3rd edition). i—xxviii, 1-797, London. EICHWALD, E. 1860. Lethaea Rossica; Paléontologie dela Russe: Ancienne Periode. Bryozoa : 355-494, pls 23-32, Moscow. In French. Exias, M. K. 1937a. Depth of deposition of the Big Blue (late Paleozoic) sediments in Kansas. Bull, geol. Soc. Am., Rochester, N.Y. 48 : 403-432. —— 1937b. Stratigraphic significance of some late Paleozoic fenestrate bryozoans. J. Paleont., Tulsa, Okla. 11 : 306-336. FROM FERMANAGH, IRELAND 491 Exias, M. K. 1950. List of identified species in Hunterian collection of Carboniferous bryozoa, with brief notes on thety most characteristic chavactervs. Unpublished manuscript: 1-4. 1956. A Revision of Fenestella subantiqua and related Silurian fenestellids. J. Paleont., Tulsa, Okla. 30 : 314-332, pl. 43. 1964. Stratigraphy and paleoecology of some Carboniferous bryozoans. C.7., Cinquiéme Congr. internatn. de stratigvaphie et de la géologie du Carvbonifére : 375-382, pls 1-5. Extras, M. K., and Conpra, G. E. 1957. Fenestella from the Permian of West Texas. Mem. geol. Soc. Am., Washington 70 : IX and 158; 23 pls. Forrste, A. F. 1887. Flint Ridge bryozoa. Bull. Denison Umiv. Sci. Lab. 2 : 71-88, pl. 7. GerorGE, T. N., and OswaLp, D. H. 1957. The Carboniferous rocks of the Donegal syncline. Q. Jl geol. Soc. Lond., London 113 : 137-183. GoLtprFuss, A. 1826-44. Petrefacta Musei Universitatis Regiae Borussicae Rhenanae Bonnensis, I. 252pp. Diisseldorf. Hatt, J. 1881. Bryozoans of the Upper Helderberg and Hamilton groups. Absty. Trans. Albany Inst. 10 : 1-26. Hatt, J. & Simpson, G. B. 1887. Palaeontology of New York, 6: Corals and bryozoa, i-xxvi, 1-298, 66 pls, Albany, N.Y. HaArRMER, Sir S. F. 1934. The polyzoa of the Siboga Expedition. Pt. III Cheilostomata, Asco- phora. Family Reteporidae, i—viii, 503-637, pls 25-41, Leiden. Hinks, T. 1880. A history of the British marine polyzoa 1: i-cxli, 1-601, Vol. 2, 83 pls, London. HIsINGER, W. 1837. Lethaea Svecica seu Petrificata Sueciae Iconibus et Chavacteribus Illustrata. 124 pp., 38 pls, Stockholm. Kaisin, F. 1942. Les bryozoaires fenestrellinidés et acanthocladiidés du Tournaisian de la Belgique. Mém. Inst. Géol. Louvain, 13 : 93-141, pls 1-8. In French. Kine, W. 1850. The Permian fossils of England. Palaeontographical Society i-xxxvii, 1-258, 28 pls, London. KoeEnic, J. W. 1958. Fenestrate bryozoa in the Chouteau Group of central Missouri. /. Paleont., Tulsa, Okla. 32 : 126-143, pls 21-22. Lacaalj, R. & GAUTIER, Y. V. 1965. Bryozoan assemblages from marine sediments of the Rhone delta, France. Micropaleontology, New York 11 : 39-58. LikHAREV, B. 1934. Upper Carboniferous of Ferghana. Bull. geol. Soc. China, Peking 13 : 155-172. LonsDALE, W. 1839. In Murchison, R. I., The Silurian System 2 : 675-677. London. LowenstaM, H. A. 1950. Niagaran reefs of the Great Lakes area. J. Geol., Chicago, etc. 58 : 430-480. MapDGEFRAU, K. 1933. Zur entstehung der mitteldeutchen Zechstein-Riffe. Centralbl. Mineralogie, Jahrg. 1933 (Abt. B, 11) : 621-624, Berlin. In German. M’Coy, SirF. 1844. A synopsis of the character of the Carboniferous Limestone fossils of Iveland : i-vill, I-207, pls 1-29, Dublin. McFarian, A. C. 1942. Chester bryozoa of Illinois and western Kentucky. J. Paleont., Tulsa, Okla. 16 : 437-458, pls 65-68. MitterR, T. G. 1961. Type specimens of the genus Fenestella from the Lower Carboniferous of Great Britain. Palaeontology, London 4 : 221-242, pls 24-27. 1962a. North American species of Fenestella from the Carboniferous of Great Britain and Ireland. J. Paleont., Tulsa, Okla. 36 : 120-125, pls 23-24. 1962b. On Hemitrypa hibernica M’Coy. Geol. Mag. London, Hertford 99 : 313-321. 1963. The bryozoan genus Polypova M’Coy. Palaeontology, London 6 : 166-171, pls 23-24. Moore, R.C. 1929. A bryozoan faunule from the Upper Graham Formation, Pennsylvanian, of north central Texas. J. Paleont., Tulsa, Okla. 3 : 1-27, pls 1-3. NEKHOROSHEY, B. P. 1926. Lower Carboniferous bryozoa from the Kuznetsk Basin. Bull. Com. géol. Leningrad 43 : 1237-1290, pls 19-20. In Russian, with English summary. 1956. Nizhni Kamennougolnye Mshanki Altaia y Sibiri. Tvudj vses. nauchno-issled. Inst. Geol.; Min. Geol. Och. Nedr. S.S.S.R. (n.s.) 13 : 1-419, pl. 57, Moscow. In Russian. 492 VISEAN FENESTRATE BRYOZOA NEWELL, N. D. and others. 1953. The Peymian reef complex of the Guadalupe Mountains vegion, Texas and New Mexico. i—xix, 1-236, 32 pls, San Francisco. NickLes, J. M. & Bassler, R. S. tIg00. A synopsis of American fossil bryozoa. Bull. U.S. geol. Surv., Washington 173 : 1-663. Nixirorova, A. I. 1926. Lower Carboniferous bryozoa of Turkestan. Bull. Com. géol. Leningrad 45 : 175-192, pls 4-5. In Russian, with English summary. 1927. Lower Carboniferous bryozoa from the Donetz Basin. Bull. Com. géol. Leningrad 46 : 245-268, pls 12-14. In Russian, with English summary. 1933a. Contribution to the knowledge of the Lower Carboniferous bryozoa of Turkestan. Trudy uses. geol-vazv. Obed. NKTP Leningrad 207 : 1-26, pl. 12. In Russian, with English summary. 1933b. Stratigraphic range of the Carboniferous bryozoa of the U.S.S.R. Tvrudy vses. geol.-vazv. Obed. NKTP Leningrad 268 : 1-32. In Russian, with English summary. 1938a. Types of Carboniferous bryozoa in the European part of the U.S.S.R. Akad. Nauk. S.S.S.R., Inst. Paleontology, Paleontology U.S.S.R. 4 : 1-290, 55 pls. In Russian, with English summary. 1938b. Stratigraficheskoe raspredelnie mshanok v neftenosnykh rifovykh isvestnyakakh Ishimbaevskogo raiona. Leningr. vses. nauchno-issled. geol.-vazv. Inst. (Ser. A) 101 : 76-89. In Russian. Oakey, K. P. 1948. In Muir-Wood, H. M.: Malayan Lower Carboniferous fossils. Brit. Mus. (Nat. Hist.) : 1-118, 17 pls, London. D’ORBIGNY, A. 1850. Pyvodvome de paleontologie stratigraphique, 3 vols : LX and 394; 1-427; 1-196, Paris. In French. OswaLp, D.H. 1955. The Carboniferous rocks between the Ox Mountains and Donegal Bay. Q. Jl geol. Soc. Lond., London 111 : 167-186. PHILLIPS, J. 1836. Illustrations of the geology of Yorkshive, Pt II : XII and 185, 14 pls, London. PRANTL, F. M. 1934. Carboniferous bryozoa from Dobsina (Slovakia). Bull. int. Acad. Sci. Prague 35 : 225-243, pls 1-2. Prout, H. A. 1859. Third series of descriptions of bryozoa from the Paleozoic rocks of the western states and territories. Tvans. Acad. Sci. St Louis, St Louis, Mo. 1 : 443-452, pls 17-18. Ross, J. P. 1961. Ordovician, Silurian and Devonian bryozoa of Australia. Bull. Miner. Resour. Surv. Aust., Canberra 50 : 1-171, pls 1-28. SARLE, C. J. Ig01. Reef structures in Clinton and Niagara strata of western New York. Am. Geol., Minneapolis 28 : 282-299. SCHWARZACHER, W. 1961. Petrology and structure of some Lower Carboniferous reefs in north-western Ireland. Bull. Am. Ass. petrol. Geol., Chicago, etc. 45 : 1481-1503. — 1963. Orientation of crinoids by current action. J. sedim. Petrol., Tulsa, Okla. 33 : 580-586. SuisHova, N. A. 1952. Podmoskovnye i dono-medveditzkie kamennougol’nye mshanki roda Septoporva. Trudy paleont. Inst., Moskva 40, 159-176, pls I-V. SHRUBSOLE, G. W. 1879. A review of the British Carboniferous Fenestellidae. Q. Jl geol. Soc. Lond., London 35 : 275-281. 1880. A review and description of the various species of British Upper Silurian Fenestelli- dae. Q. Jl geol. Soc. Lond., London 36 : 241-254. 1881. Further notes on the Carboniferous Fenestellidae. Q. Jl geol. Soc. Lond., London 37 : 178-189. SHULGA-NESTERENKO, M. I. 1941. Lower Permian bryozoa of the Urals. Akad. Nauk. S.S.S.R., Paleont. Inst., Paleont. U.S.S.R. 5 : 1-226, 67 pls. In Russian, with English summary. 1951. Kamennougolnye fenestellidy Russkoi platformy. Akad. Nauk. S.S.S.R., Paleont. Inst. 32 : 1-61, 34 pls. In Russian. 1955. Kamennougolnye mshanki Russkoi platformy. Akad. Nauk. S.S.S.R., Paleont. Inst. 57 : 1-207, 32 pls. In Russian. FROM FERMANAGH, IRELAND 493 Simpson, G. B. 1895. A handbook of the genera of the North American Paleozoic bryozoa. A. Rep. State Geol. (N.Y.) for 1894 : 407-608. Simpson, I. M. 1954. The Lower Carboniferous stratigraphy of the Omagh syncline, Northern Ireland. Q. Jl geol. Soc. Lond., London 110 : 391-408. SPJELDNAES, N. 1957. The name Fenestella. J. Paleont., Tulsa, Okla. 31 : 675-676. Stacy, L. W. 1936. Correlation of zoarial form with habitat. J. Geol., Chicago, etc. 44 : 60-65. STUCKENBERG, A. 1895. Korallen and bryozoen der steinkohleablagerungen des Urals USW. Mém. Com. géol. St Petersburg 10 : 1-244, 24 pls. In German. TAVENER-SMITH, R. 1965. A new fenestrate bryozoan from the Lower Carboniferous of County Fermanagh. Palaeontology, London 8 : 478-491, pl. 66. —— 1966a. Ovicells in fenestrate cryptostomes of Visean age. J. Paleont., Tulsa, Okla. 40 : 190-198, pl. 25. 1966b. The micrometric formula and the classification of fenestrate cryptostomes. Palaeontology, London 9 : 413-425. 1969. Skeletal structure and growth in the Fenestellidae (Bryozoa). Palaeontology, London 12 : 281-309, pls 52-56. 1971. Morphology and significance of unusual frontal features in the bryozoan Polypora stenostoma. Palaeontology, London 14 : 178-187, pl. 25. TuHeEopoRIvicH, G.I. 1941. Migration of Upper Paleozoic reefs in south Bashkiria. Akad. Nauk. S.S.S.R. Doklady 32 : 499-502. In Russian, with English summary. Toots, H. 1951. Uber einige cryptostome bryozoen aus dem Mitteldevon des Rheinischen Scheifergebirges. N. Jb. f. Geol. u. Pal., Abh. Bd. 93 : 233-246. In German. Trizna, V. 1939. New species of the Upper Paleozoic Fenestellidae and Acanthocladiidae from the Bashkirian Urals. Geol. Oil Inst. Tvans. (sey. A.) 115 : 102-144. In Russian, with English summary. Urricu, E.O. 1886. Descriptions of new Silurian and Devonian fossils. Conty. Am. Paleont. Cincinnati 1 : 3-35. 1890. Paleozoic Bryozoa. Bull. geol. Surv. Illinois 8, Vol. 1 : 283—688; Vol. 2: pls 29-78. Vine, G. R. 1879. On Palaeocoryne and its development on Fenestella. Sci. Gossip, London 15 : 225-229 and 247-249. Waters, A.W. 1896. On Mediterranean and New Zealand Reteporae and fenestrate bryozoa. J. Linn. Soc. (Zool.), London 25 : 255—271. WHIDBORNE, G. F. 1895. A monograph of the Devonian fauna of the South of England. Palaeontogr. Soc., Vol. 2 : i-xi, 1-219, 24 pls, London. Wuittincton, H. B. & Evitt, W. R. 1953. Silicified Middle Ordovician trilobites. Mem. geol. Soc. Am., Washington 59 : 1-137, 33 pls. Youne, J. & Younc, J. 1874. On Palaeocoryne and other polyzoal appendages. Q. Jl. geol. Soc. Lond., London 30 : 684-689. R. TAVENER-SM1TH, Ph.D. Dept. of Geology UNIVERSITY OF NATAL, KING GEORGE V AVENUE, DURBAN, SOUTH AFRICA. PLADE 1 Fenestella frutex M’Coy Fic. 1. Obverse of large zoarial fragment, with an encrusting stenoporid colony. PD.5001.* 2:9. Fics 2, 3, 5,6. Obverse of small zoarial fragments to illustrate common variations in mesh- work appearance. PD.5002 to PD.5005, x6-0; 5:0; 6:7; 5-6, respectively. Fic. 4. Stout, barbed spines from the obverse surface to the meshwork. PD.5006. 4:5. Fics 7, 9. Detail of the obverse surfaces of morphological variants within this species. PD.5005, 18-0; and PD.5002, x2I-o. Fic. 8. Reverse side of a zoarial fragment showing longitudinal ‘striae’ and minute nodes. PD.5007. x5°8. * The serial numbers of specimens refer to the Bryozoa catalogue numbers of the collections of the British Museum (Natural History), where the material is stored. I PLATE Bull. By. Mus. nat. Hist. (Geol.) 23, 7 , co ee te ie ee _P a - > =, te Fa EAI Fenestella ivanovi Shulga-Nesterenko Fics 1, 3, 5, 7. Obverse of zoarial fragments to show variations in meshwork pattern. PD.5008, 6-0; and PD.4682 to PD.4684, «5:0; 5:3; 5:0, respectively. Fic. 2. Proximal part of a zoarium with basal holdfast and lateral supports; the last developed from marginal branches and dissepiments. PD.4685. 5:2. Fic. 4. Zoarial fragment with elongate spinose structures which are ‘infertile’ branch continuations. PD.4686. 95:5. Fic. 6. Delicate spines with minute barbs growing from the obverse sides of branches. PD.4687. 5:0. Fic. 8. Reverse surface of zoarial meshwork. PD.4684. 17:0. Fic. 9. Detail of obverse side showing the distinctive shape of fenestrules. PD.4685. < 16:0. PLATE Bull. By. Mus. nat. Hist. (Geol.) 23, 7 PLATE 3 Fenestella cf. multispinosa Ulrich Fic. 1. Reverse of zoarial meshwork showing longitudinal ‘striae’ and nodes on the mid-line of branches. PD.4688. 5:7. Fic. 2. Another view of the reverse surface. PD.4689. 5-0. Fics 3, 7. Obverse of zoarial fragments with slightly different meshwork characteristics. PD.4689, x 5:0; and PD.4690 7:1, respectively. Fic. 4. Obverse side showing normal branches passing distally into sterile lateral processes. JID) AGeie, SK FD Fics 5, 8. Obverse of meshwork showing localized development of thick, striated, secondary skeleton. PD.4692, 8-0; and PD.4693, 7-0. Fic. 6. Reverse of zoarial fragment showing branches passing distally into stout, barbed, lateral spines. PD.4694. 5°4. Fic. 9. Detail of obverse surface. PD.4680. x 18-3. PLATE 3 23,7 Bull. By. Mus. nat. Hist. (Geol.) PLATE 4 Fenestella modesta Ulrich Fics 1, 2, 4, 6. Obverse of zoarial fragments illustrating variation in meshwork pattern. PD.4696, «6-6; PD.4698, «5:0; PD.4695, «6:8; and PD.4697, 6-4, respectively. Fic. 3. Detail of obverse surface. x 25-0. Fic. 5. Reverse of zoarial fragment. PD.4695. 6:8. Fenestella hemispherica M’Coy Fic. 7. Obverse of zoarial fragment. PD.4699. x 13°5. Fic. 8. Reverse surface showing uniserial nodes along mid-lines of branches. PD.4704. 46:3. Fic. 9. Large fragment showing typical meshwork pattern. PD.4701. 4:7. PLATE 5 Fenestella hemispherica M’Coy Fics 1, 2. Obverse surface of zoarial fragments. PD.4700, x15:0; and PD.4702, x6-6. Fic. 3. Reverse side showing nodes. PD.4699. «58. Fic. 4. Obverse of meshwork showing characteristic mode of branch division. PD.4704. « 13°8. Fenestella parallela Hall Fic. 5. Reverse of zoarial fragment showing striated pattern. PD.4710. 7:3. Fics 6, 8, 9. Obverse surfaces showing meshwork characteristics. PD.4707 to PD.4709 X70; X7:0; 9:7, respectively. Fic. 7. Detail of obverse side. PD.4708. 14°5. PLATE 5 / Bull. Br. Mus. nat. Hist. (Geol.) 23, * ; = La) nd —_ —~ @ . BA 1D PAIN AILIS, (6) Fenestella rudis Ulrich multinodosa subsp. nov. Fic. 1. Detail of obverse of a zoarial fragment. 14:2. Fic. 2. Obverse of meshwork with localized presence of thick, secondary skeleton which has coated branch surfaces and reduced or sealed zooecial apertures. PI).4714 (paratype). 8-8. Fic. 3. Reverse side of zoarial fragment. PD.4714 (paratype). 8-2. Fics 4, 6, 7. Obverse of fragments showing slight differences of meshwork pattern.PD.4713 (paratype), 5:0; PD.4716 (paratype), «7-4; and PD.4712 (holotype), 5-0. Fic. 5. Obverse of partly silicified fragment showing hemi-hexagonal zooecial chambers. PD.4715 (paratype). 7-9. Bull. Br. Mus. nat. Hist. (Geol.) 23, 7 PLN ALLE, (6) PALAIS, 7 Fenestella plebeia M Coy Fic. 1. Obverse of zoarial fragment showing meshwork pattern. PD.4719. 3-6. Fics 2, 3. Detail of obverse surface. Fig. 3, from proximal part of colony shows thick secondary skeletal envelope. Minute punctae along branch medial lines are carinal node bases. PD.4719, X 11-4; and PD.4717, x 11-3, respectively. Fic. 4. Obverse of zoarial fragment from proximal part of colony. PD.4717. x 4:5. Fic. 5. Obverse of partly silicified fragment showing triangular zooecial chambers. PD.4718. 6-0. Fic. 6. Zoarial fragment with inflated peristomes projecting into fenestrules. PD.4721. x 6-4. Fic. 7. Reverse surface of zoarial fragment. PD.4717. 4:5. Fenestella ci. arthritica Phillips Fic. 8. Detail of obverse surface. PD.4722. 15:0. Fic. 9. Large zoarial fragment giving an impression of the colonial growth form. PIDFA72 8. ez: PLATE 23,7 ul. Hist. (Geol.) Bull. Br. Mus. ne PAA aS: Fenestella ci. arthritica Phillips Fics 1, 3, 4. Obverse of zoarial fragments showing characteristics of meshwork pattern. Stout carinal nodes are visible along proximal part of middle branch in Fig. 3. PD.4727, x 3:9; PD.4726, x5:0; and PD.4724, 5:1. respectively. Fic. 2. Reverse surface of meshwork. PD.4727. 3:9. Fenestella praemagna Shulga-Nesterenko Fic. 5. Proximal part of a colony with basal holdfast and lateral supporting struts. PD.4732. 5-0. Fic. 6. Strong carinal nodes, which divide at their distal ends, originate along the obverse mid-lines of branches. PD.4735. 5:1. Fic. 7. Proximal part of a colony attached to a Pennivetepova branch. PD.4733. X47. Fic. 8. Obverse of colony showing basal holdfast and supporting processes. PD.4734. x 10-4. Fic. 9. Detail of obverse. PD.4731. 14-0. Fic. 10. Reverse side of zoarial fragment. PD.4730. xO6:1. Fic. 11. Obvyerse of proximal part of an old colony, with thick secondary skeletal cover. PID AG PR). >< 7/50) Bull. Br. Mus. nat. Hist. (Geol.) 23, 7 PLATE 8 PLATE 9 Fenestella fanata Whidborne carrickensis subsp. nov. Fic. 1. Obverse of meshwork showing characteristic pattern at branch division. PD.4736 (holotype). 10-1. Fic.2. Zoarial fragment with stout lateral spine developed as a branch continuation. PD.4737 (paratype). 35. Fics 3, 5, 7, 8. Obverse of zoarial fragments, showing general meshwork characteristics. Specimen shown in Fig. 3 (x3:9) is not preserved. The others are numbered: PD.4738 (paratype), x 5:0; PD.4739 (paratype), x 3-8; and PD.4736, x 3:5, respectively. Fic. 4. Zoarial fragment with ovicells visible in the distal part. PD.4740 (paratype). 3:5. Fic. 6. Reverse side of a fragment, showing evidence of damage and subsequent repair during the life of the colony. PD.4741. 3:3. Fic. 9. Detail of obverse surface. PD.4742 (paratype). 9:9. Bull. Br. Mus. nat. Hist. (Geol.) 23, 7 PLATE 9 PLATE io Fenestella cf. spinacristata Moore Fic. 1. Large zoarial fragment showing overall meshwork pattern. PD.4743. 3:8. Fic. 2. Reverse surface of meshwork. PD.4746. 5-0. Fic. 3. Detail of obverse side. PD.4743. 49:5. Fic. 4. Proximal part of a colony with holdfast attached to a Pennivetepova branch. PD.4747. X61. Fic. 5. Obverse of zoarial fragment. PD.4744. x7'1. Fic. 6. Fragment bearing evidence of structural damage and repair during the life of the colony. PD.4750. 4:5. Fenestella cf. funicula Ulrich Fic. 7. Reverse side of fragment from proximal part of a colony, showing a stout spine developed from obverse mid-line of a branch. PD.4758. 5°5. Fic. 8. Obverse of a zoarial fragment. PD.4751. 95:0. Fic. 9. Reverse side of above. PD.4751. 5:0. Fic. 10. Fragment from proximal part of an old colony showing thick secondary skeletal encrustation. PD.4759. 2:9. PLATE 11 Fenestella cf. funicula Ulrich Fic. 1. Obverse of zoarial fragment showing ovicells in the distal part. PD.4759. 4:3. Fics 2, 3. Obverse surfaces showing general meshwork characteristics. PD.4753, 6:8; and PD.4755, X7:0. Fic. 4. Obverse surface of meshwork with strong carinal nodes. PD.4756. x 4:0. Fenestella filistriata Ulrich Fics 5,7, 9. Obverse of zoarial fragments showing meshwork pattern. PD.4763, 5:3; PD.4761 x3:8 and PD.4762, x5.0, respectively. Fic. 6. Reverse surface. PD.4763. 5:5. Fic. 8. Reverse of zoarial fragment showing longitudinal ‘striae’ and hemi-hexagonal bases of zooecial chambers. PD.4764. 7-0. Fic. 10. Detail of obverse.. PD.4761. 12:2. IIL INARI Fenestella cf. filistriata Ulrich Fic. 1. General view of obverse side of a zoarial fragment. PD.4761. 5-2. Fenestella subspeciosa Shulga-Nesterenko Fic. 2. Reverse surface. PD.4766. 5:7. Fics 3, 4, 5. Obverse of zoarial fragments showing general meshwork characteristics. PD.4765, <5:0; PD.4769, x 4:0; and PD.4767, 5-6. Fic. 6. Obverse at higher magnification, showing prominent cowl-like peristomes indenting fenestrule margins. PD.4766. 5-2. Fic. 7. Reverse of zoarial fragment showing a stout spine which resulted from continued growth from an abnormally directed dissepiment. This spine later re-united with the mesh- work, as shown. 7:2. Fic. 8. Obverse of above fragment showing a local concentration of secondary skeletal tissue where the aberrant spine (coming from below) rejoined the meshwork. 7-2. PLATE 13 Fenestella pseudovirgosa Nikiforova Fic. 1. Detail of obverse. PD.4770. 15:0. Fic. 2. Reverse side of meshwork. PD.4771. 2:9. Fic. 3. Zoarial fragment showing general meshwork characteristics. PD.4770. 5:0. Fic. 4.. The stout spine developed from a branch end in the distal part of this specimen EID 72, KB), grew back and re-united with the reverse side of the meshwork. Fic. 5. This strong spine resulted from the continued growth of an abnormal dissepiment. PD.4773. X3°7- Fics 6, 9. Zoarial fragments from proximal parts of colonies, showing thick deposits of secondary skeletal substance. PD.4774, «5:6; and PD.4775, 5:0. Fic. 7. Obverse of fragment, showing a strong laterally directed spine. PD.4776. 4:1. Fic. 8. Reverse surface showing longitudinal ‘striae’, variably developed nodes, and the triangular base of a zooecial chamber. PD.4777. 4:3. PA AGE a4: Fenestella cf. albida Hall Fics 1, 3. Obverse of zoarial fragments showing general meshwork characteristics. PD.4783, x 6-7; and PD.4780 x 5:0, respectively. Fic. 2. Obverse surface showing chain-like pattern due to projection of inflated peristomes into fenestrules. PD.4781. 6-4. Fic. 4. Detail of obverse. PD.4780. x 15-0. Fic. 5. Proximal part of a colony with basal attachment to a Penniveteporva branch. PD.4779. X3°4- Fic. 6. Reverse side of meshwork. 5:8. Fenestella oblongata Koenig Fic. 7. Obverse of a zoarial fragment. PD.4790. 7-0. Fic. 8. Reverse surface showing longitudinal ‘striae’, small nodes and the triangular base of a zooecial chamber. PD.4791. 7°5. PLATE 15 Fenestella oblongata Koenig Fic. 1. Detail of obverse of zoarial fragment. PD.4786. 14:8. Fics 2, 3, 4. Obverse views of fragments showing general aspect of meshwork. PD.4787, 5:0; PD.4789, x 4:2; and PD.4786, x 6-3, respectively. Fenestella cf. delicatula Ulrich Fics 5, 6. Obverse surfaces of zoarial fragments illustrating variation in meshwork charac- teristics. PD.4797, x6:3; and PD.4794, x7-0. Fic. 7. Zoarial fragment with strong spines from the reverse side of a branch. The spines bear minute barbs. PD.4799. 5:0. Fic. 8. Detail of obverse side. PD.4798. x 16-0. 15 PLS NAE Led / (Geol.) 23, Mus. nat. Hist. Bull, Br. PLATE 16 Fenestella cf. delicatula Ulrich Fic. 1. Reverse side of meshwork. PD.4798. 7-5. Fic. 2. Obverse of zoarial fragment. PD.4795. 4:8. Fic. 3. Fragment with strong lateral spine developed asa branch continuation. PD.4798. K 7:5. Fenestella polyporata (Phillips) Fic. 4. Reverse of meshwork showing longitudinal ‘striae’. PD.4802. 7-2. Fics 5, 7, 8. Obverse surfaces of zoarial fragments showing general features of meshwork. PD.4805, x 7:0; PD.4802, «3:7; and PD.4804, 5-0, respectively. Fic. 6. Detail of obverse. PD.4804. <14°9. Bull. Br. Mus. nat. Hist. (Geol.) 23, 7 PLATE 16 PA ASE, 7 Fenestella polyporata (Phillips) Fic. 1. Obverse of zoarial fragment. PAD SOR, <7 Fenestella irregularis Nekhoroshev FIGS 2, 5. Obverse of zoarial fragments, showing meshwork features. and PD.4811, x 3:9, respectively. PDYASesx< 7-1 Fic. 3. Detail of obverse. PD.4810. x 13:0. Fic. 4. Reverse side of zoarial fragment. PD.4809. 7:5. Fic. 6. Obverse of incompletely silicified fragment showing casts of zooecial chambers PD.4813. x85. Fic. 7. Oblique view of obverse showing strong carinal nodes. PD.4812. x 4°5. Levifenestella undecimalis (Shulga-Nesterenko) Fic. 8. Reverse surface with longitudinal ‘striae’ and grooves excavated by ramifying Condvanema (ctenostome) stolons. PD.4815. 5:0. Fic. 9. Detail of obverse. PD.4816. x 16-0. ATE IPL Mus. nat. Hist. (Geol.) 23, 7 Bull. Br. PATE aS Levifenestella undecimalis (Shulga-Nesterenko) PD.4817, x6:6; PD.4814, 4-0; and PD A816 2 << 7-3: FIGS 1, 2, 4. Obverse surfaces of zoarial fragments. PD.4820, x 6-2, respectively. Reverse of meshwork showing closely spaced longitudinal ‘striae’. Obverse of large zoarial fragment which gives an idea of the colonial growth form. Fia. 3. FIG. 5. PD.4818. x 2:3. Minilya plummerae (Moore) Fic. 6. Small fragment with a stout, barbed spine developed from the reverse surface. RD A8275 Os: Obverse of zoarial fragments showing general aspect of meshwork. PD.4825, Fics 7, 8. x 6:0; and PD.4824, x7-5, respectively. PLATE x73 23,7 (Geol.) Bull. Bry. Mus. nat. Hist. Pe ACs. 1G) Minilya plummerae (Moore) Fic. 1. Detail of obverse of meshwork. Positions of carinal nodes are indicated on a part of one branch. PD.4823. 22-0. Figs 2, 4. Obverse of zoarial fragments, showing general meshwork characteristics. PD.4823, 6-6; and PD.4822, 5-0, respectively. Fic. 3. Reverse surface. PD.4822. 5:0. Minilya binodata (Condra) Fics 5, 7. Obverse of zoarial fragments. PD.4829, «7-3; and PD.4848, 8-5. Fic. 6. Reverse surface. PD.4828. 5:8. PEALE, 19 uw. Hist. (Geol.) Bull. Br. Mus. ne > & ie 6 = = > . 3 * cs 2. 4 > -o: 6.5. J ¥,' a ei ‘ } sy i ‘ Finny: Hh en CS s é a ¢ 5 er S 2 ) t re x ~~ b ~ : A , < A aS s - ee ad $,' : > 2% : a weeeie PEALE 20 Minilya binodata (Condra) Fic. 1. Detail of obverse, with the positions of some carinal nodes indicated. PD.4828. x 16°5. Minilya oculata (M’Coy) Fic. 2. A fragment from the proximal part of an old colony, with thick secondary skeletal investment. PD.4833. 6:2. Fic. 3. Detail of obverse showing the biserial arrangement of carinal nodes. PD.4838. If) Fics 4, 6, 8. Obverse of zoarial fragments, showing meshwork pattern and growth habit. PD.4834, «5:0; PD.4837, «6-4; and PD.4835. x 6-9, respectively. Fic. 5. Reverse side of zoarial fragment. PD.4839. 6:5. Fic. 7. Proximal part of a colony with holdfast attached to a Rhabdomeson fragment. PD.4832. x 3:9. PLATE AB) 7) t. Hist. (Geol.) sy. Mus. na Bull. I PACE 2a Ptiloporella varicosa (M’Coy) Fics 1, 2, 3, 6, 7. Obverse of zoarial fragments showing prominent development of primary and secondary branches. PD.4840 to PD.4844 inclusive, x 4:0; X4:-7; X5:0; X2:6; X2°7, respectively. Fic. 4. This specimen, with a midrib and symmetrically placed lateral branches, resembles Ptylopora M’Coy in its growth pattern. PD.4845. 4-6. Fic. 5. Small fragment with strong spinose structures which originate on the obverse mid- line. PD.4846. 4:3. Fic. 8. Reverse Surface, showing the relationship between primary and secondary branches. IDFA SAV 30: Fic. 9. Detail of 6bverse. PD.4842. 135. PAA TE 23,7 (Geol.) uw. Hist. Bull. Br. Mus. ne PLATE 22 Hemitrypa hibernica M’Coy Fic. 1. Proximal part of a colony attached to a Rhombopora fragment. Ends of supporting spines are visible. PD.4848. 3:2. Fic. 2. Reverse surface of zoarial fragment showing a secondarily thickened ‘dissepimental ALC ED PASAG sass: Fic. 3. Proximal part of a colony with enclosing superstructure and stout supporting spines. PD.4850. 5:0. Fic. 4. Obverse side with superstructure removed. PD.4851. 5-0. Fic. 5. Obverse with superstructure partly removed to show abnormal branching pattern, possibly a pathological condition. PD.4852. 5:0. Fic. 6. Lateral view showing superstructure supported by carinal pillars. Underside of meshwork is encrusted by a Fistulipova colony. PD.4853. 6-5. Fic. 7. Reverse surface with distal parts of a long, branched supporting spine which re- unites with the colonial meshwork. PD.4854. 3:2. Fic. 8. Obverse of a large fragment, with young Fistulipora colony encrusting the super- structure. PD.4855. 3:0. Fic. 9. Obverse of meshwork showing ovicellular concavities. PD.4856. 4-4. Fic. 10. Detail of obverse. PD.4857. x 18-0. PLATE 2B (Geol.) nat. Hist. Bull. By. Mus. ~? a@@e@egee e+e @ eee @ ‘¢ , ENC PNICIS 23} Polypora dendroides M’Coy Fics 1,5. Obverse of zoarial fragments from the proximal parts of colonies. Thick secondary skeletal tissue is particularly evident at the bases of broken supporting spines, x 5-4; PD.4865, and PD.4866, x 2-6. Fics 3, 4. Reverse of meshwork. Fig. 4 shows the proximal parts of supporting spines. PD.4867, 5-0; and PD.4868 x 2-6. Fics 2, 6. Obverse surfaces showing general characteristics of meshwork. PD.4867, x 5-0; and PD.4869, x 4:3. Fic. 7. Detail of obverse side. PD.4867. 13-0. Bull, Br. Mus. nat. Hist. (Geol.) 23, 7 PLATE 23 / PLATE 24 Polypora verrucosa M’Coy Fic. 1. Detail of obverse. PD.4870. x12°8. Fics 2, 3,5, 7. Obverse of zoarial fragments showing general features of meshwork and branch surface. PD.4872, x 3-4; PD.4873, x3:8; PD.4870, x6-6; and PD.4874, x6:2, respectively. Fic. 4. Fragment from proximal part of a colony, showing prominent peristomes with terminal apertures, some of which are sealed by the thick secondary skeletal investment. PD.4871. 6-6. Fic. 6. Reverse side showing incipient beekitization resulting from the silicification of thick secondary skeletal accretions. PD.4877. 6-0. > Mus. nat. Hist. (Geol.) Bull. Br. PLATE 25 Ptylopora pluma M’Coy parva subsp. noy. Fic. 1. Detail of obverse showing midrib and lateral branches. PD.4880 (paratype). < 18-6. Fic. 2. Reverse surface with prominent ‘striae’. PD.4880 (paratype). 1.056: Fics 3, 5, 8, 10. Obverse of zoarial fragments showing general meshwork characteristics. PD.4878, «5:0; PD.4879, x 6:0; PD.4881 and PD.4882, 4:1; 4-1 (all are paratypes). Fic. 4. Proximal part of a colony with thick outer secondary investment. The stalk-like holdfast has subsidiary attachment structures at its lower end. 5-0. Fic 6. Midrib of a colony with vestiges of secondary branches. PD.4883 (holotype). x 2-7. Fics 7, 9. Reverse sides of zoarial fragments. Fig. 9 shows a short lateral spine developed from a branch end. PD.4884 and PD.4882 (paratype) respectively. Both x 4-1. PLATE 7 23, (Geol.) Bull. By. Mus. nat. Hist. =F 2 © © & @& o> ik ia a wm \S F PLATE 26 Septopora hibernica sp. nov. Fics 1, 7, 8. Obverse of zoarial fragments. In many places branch surfaces are obscured and zooecial apertures sealed by thick secondary skeletal deposits. PD.4885 (holotype), «2:2; PD.4886 (paratype), x 6-1; and PD.4887 (paratype), x 3-6. Fic. 2. Reverse surface of meshwork. The difference between primary and secondary branches is very marked. PD.4885 (holotype). x 2:2. Fic. 3. Obverse with thick investment of secondary skeleton. The ‘striae’ traverse both branches and dissepiments when traced away from the bases of large spines. PD.4889 (paratype). 4:5. Fic. 4. Fragment with Ptylopora-like meshwork of midrib and lateral branches. PD.4888 (paratype). 1-9. Fic. 5. Zoarial fragments with irregular growth habit. PD.4890 (paratype) and PD.4891. X4°5- Fic. 6. Reverse surface showing thick deposits of secondary skeleton. PD.4889 (para- type). 6-4. / Mus. nat. Hist. (Geol.) 23, Bull. Br. A LIST OF SUPPLEMENTS TO THE GEOLOGICAL SERIES OF THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) . Cox, L. R. Jurassic Bivalvia and Gastropoda from Tanganyika and Kenya. Pp. 213; 30 Plates; 2 Text-figures. 1965. 6. . Et-NaGcGar, Z. R. Stratigraphy and Planktonic Foraminifera of the Upper Cretaceous—Lower Tertiary Succession in the Esna-Idfu Region, Nile Valley, Egypt, U.A.R. Pp. 291; 23 Plates; 18 Text-figures. 1966. {1to. . Davey, R. J., DowntE, C., SARGEANT, W. A. S. & WitiiaMs, G. L. Studies on Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 248; 28 Plates; 64 Text- figures. 1966. £7. . APPENDIX. Davey, R. J., DowntE, C., SARGEANT, W. A. S. & WILLIAMS, G. L. Appendix to Studies on Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 24. 1969. 8op. . Ertrott, G. F. Permian to Palaeocene Calcareous Algae (Dasycladaceae) of the Middle East. Pp. 111; 24 Plates; 17 Text-figures. 1968. 5.124. . Ruopes, F. H. T., Austin, R. L. & Druce, E. C. British Avonian (Carboni- ferous) Conodont faunas, and their value in local and continental correlation. Pp. 315; 31 Plates; 92 Text-figures. I969. II. . Cuitps, A. Upper Jurassic Rhynchonellid Brachiopods from Northwestern Europe. Pp. 119; 12 Plates; 40 Text-figures. 1969. £4.75. . Goopy, P. C. The relationships of certain Upper Cretaceous Teleosts with special reference to the Myctophorids. Pp. 255; 102 Text-figures. 1969. £6.50. . OwEN, H. G. Middle Albian Stratigraphy in the Paris Basin. Pp. 164; 3 Plates; 52 Text-figures. 1971. £6. . Sippigur, Q. A. Early Tertiary Ostracoda of the family Trachyleberididae from West Pakistan. Pp. 98; 42 Plates; 7 Text-figures. 1971. {8. Ber Printed in England by Staples Printers Limited at their Kettering, Northants establishment SAS ere rer C9 t Chee . ANE ce ene ee Rai ecae ata ty at #4, y So ante sagas pateeaticgeas Heit es ata t Re roe tie ; Ear ahr acer ary aa fwiets eat gtg lt ea ae: sledge tits nine Seles K Roath: ais x ; 4, ate atid eel) as ah ateeat fe ~ 4 edhe Viton s timed ae) phase! PRISM tate rite® eee ++ + + + * we ne eye ‘4 fe i ane esety; shes x nn te