ae tan Teiidy CAL hd bh bath ay sie je ha hy, yeaa t Yet Ret te i 2s . : vee pebs > sees eu Ky d > Gat ae ate Sys eee SISA erty At Tea Pg Annan JR, er ; re es Pe ¥ * = ee ee ey rere a =O beg DR ee ee Se HARVARD UNIVERSITY Gi LIBRARY OF THE Museum of Comparative Zoology « ym ‘ ; i 7 ‘ 7 if 4 = 4 7 a A ra ij ai i o / , a ~ ae andl’ Mea Rae Here Ge nivale fe ae a %, i i bik 1 al j \\ Tul!) OU Tertiary Pongidae of East Africa: Pram Evolutionary Relationships and Taxonomy David R. Pilbeam Bulletin 31 PEABODY MUSEUM OF NATURAL HISTORY YALE UNIVERSITY Tertiary Pongidae of East Africa: Evolutionary Relationships and Taxonomy DAVID R. PILBEAM Peabody Museum of Natural History and Department of Anthropology, Yale University BULLETIN 31 e DECEMBER 1969 PEABODY MUSEUM OF NATURAL HISTORY YALE UNIVERSITY NEW HAVEN, CONNECTICUT Bulletins published by the Peabody Museum of Natural History, Yale Uni- versity, are numbered consecutively as independent monographs and appear at irregular intervals. Shorter papers are published at frequent intervals in the Peabody Museum Postilla series. The Peabody Museum Bulletin incorporates the Bulletin of the Bingham Ocean- ographic Collection, which ceased independent publication after Vol. 19, Article 2 (1967). Publications Committee: A. Lee McAlester, Chairman Theodore Delevoryas Willard D. Hartman John H. Ostrom Keith S$. Thomson Thomas M. Uzzell Alfred W. Crompton, ex officio Editor: Jeanne E. Remington Asst. Editors: Nancy A. Ahlstrom, editor of this volume Elizabeth G. Weinman Communications concerning purchase or exchange of publications should be ad- dressed to the Publications Office, Peabody Museum of Natural History, Yale University, New Haven, Connecticut 06520, U.S.A. Printed in the United States of America CONTENTS APT oO EMG WIRES 32h cites tntunsteheneh tects ties atstas reas ahaa oa ea ee vi US ale OI CIP AUB IGIES 9 203 32 FG AGUS oh dacs eos nea hee ne Ral end eee Vii eS ie Oi APPENDICES ote tala k iy oktas tek oe clete kako we a Oe ee 1x ABSTRACTS (ENGLISH, GERMAN, RUSSIAN) .:....:s0c0c:eue-+ eee ] ACUI NERO) OTe LOIN eis whee Be ie ai steel Dane ML Rains WEA ears La 5 INCKNOWEEDCEMENTS: Ritch duce sueG £ue wey Sei ie WIR we ohne Se anon aoe 6 Pee MATERIATES AND METHODS ssc cause tc hales) cursed cave one ameres 7 IV TIEHAIS RINE TLS ays 12 Bese os ares ETS GE SRA tat ses Te te eel reek rae vi PINS Wil ULLOUAS vay Saad ever ire ese rot aee ues tates Oust tee iene Naat ce sh cua ae ce nye nape 8 ONG ARTA. SAIS TICS movie ee asi cleiare aun he Piate Lowe seein oo Gee Oe Diets he 9 Py ORE VAL OIISsrstey sere cea k, Gia cre Ren RPA EMCI: eae a eile Seer tee awly Sane hehe 9 PUA RT VPAIRVACER a ANAT STS sees) oaieira 215i aida ota tots Abate aa RISER IN ORO daa eee SINE 11 III. GEOLOGIC AND PALEOECOLOGIC CONSIDERATIONS OF mane TeAST? AbRIGAIN SED ES ik oh ok ere oak oe fare tice aia bakes Asa cee Seo Se as see Re dBi eG ORs Ge ee 47 UMP: 66-41) 3 oi65 Ve ean oe Bice OPA 48 Second molars 8 sdk. bo oa eee ee Ge oye nee On eee 50 INEP O20 8 se 2s mex SM Gonieedomen ora tek ele Sie ewe Bie eas ie eee ee 50 WINER 62209" sect ti Ses eis GS ww UR ee ee eon 50 Lower dentition andimandiblés') 20.07 2. 4c.) ssc oe eee 2 ie eee 51 UMP 62-06. Si iskue fesate ee ec Oa ae eee eee 51 UMP sO 22153. soe ieee pete lnicia ecco w/e nae 53 IMP O21 A sats 5 sand rc laps Sn One eta doe coins ot 54 UMP O22 bo i ocis os od aie Be eee ok eee ee eee 55 WIMP °6 2216 B52 os 5d ik Rta AIT toes coe tence Detera ee ene 55 WMP (66202) ss. 2.05 noseenh head Oe VER ress eye 56 Discussion ‘of ‘the Jower dentition 421.25... 420i eee eee 56 DISCUSSION; OF THE UGANDA, PONGIDS)): 2.002 6454 $0 c6 es tee ee ee ee 57 VI. DRYOPITHECUS (PROCONSUL) MAJOR FROM KENYA ....... 59 BACKGROUND. OF “THE IMIATERIAT | 55 dineie se oe oe eke Oe ee 59 ICARGE) SPECIMENS \q.cuseteioe ievaitn. c%s Sass and Means eee anes eee eee 60 IBM(NET) MIGG48 225.8 5 cts, homies ee ae (Proconsul) <.... 22... +0 eee eee 89 lie bivarate plots of €; and P, of D. (Proconsuly’ 32 t sas. ee eee 90 12; Bivariate plots of P, and M, of D. (Proconsul) ....- 2034. :- 2 eee 91 13: (Bivariate ‘plots of M, and M, of D. (Proconsul) 23) ..%. 6.6 eee 92 14. Bivariate plots of P, of Dryopithecus species, Pan troglodytes ang (GOnlla COMMU G ho. .iooa sos Bae sone nous Gece © ose cee a> mee teenie 140 15. Bivariate plots of M, of Dryopithecus species, Pan troglodytes AIG GORA: COTTA vax. oon cs oveas e-scovenn: pare etee a agave See enerne ee ihe, 16. Bivariate plots of M, of Dryopithecus species, Pan troglodytes AMG AG OUND TC OTIULA och sre got dias «0h oat Rc ee OI So Oe eee 142 17. Bivariate plots of M; of Dryopithecus species, Pan troglodytes ANG IGORMUG COMME % coscic cds eee Vous bons eee ote Cee een Ee) 182 Occhusal views of UMP 62210) 5.5.3 es yeni oo ot aiers tects eto ee eee ake 19. Lateral and anterior views of UMP 62-I1 ..........--.. 2242522055 145 20. Occlusal views of UMP 62-10 and UMP 66-01 ........:.-.--.... 5. 146 21. Occlusal views of UMP 62-07, UMP 66-41, UMP 62-08 and INIVERG AOD SET oii cee 6 ieee iole ae Sra eee ote cts clekolt Sieve ae earlier oe Ere 147 22. Occlusal views of UMP 62-06, UMP 62-13, UMP 62-14, UMP 62-15, WIV0UR 6216) amd: WMP GG=02) 2 ec eee bic 2 eee oe telethon eee 148 23. Occlusal views of UMP 62-16 and BM(NH) M14086; BM(NH) M16648; BM(NH) M14086; and BM(NH) M16648 .................. 150 24. Buccul and ligual views of UMP 66-41 and UMP 62-14; lateral view OF WIMP, 62-08) and. NIMK 19050 225207 elec eae ee ce eee 15 25. Occlusal view of BM(NH) M14084 ..........222-2-- 2225 ss se ee wean 152 26: Occlusal view ot NIMK. 190,10 2.0. oh bbe eee obese er sem eerie 152 27. Occlusal and buccul views of BM(NH) M16647 ...........--..++--: 153 28. Occlusal view of BM(NED) Mi1lG649 22025. oy fie oe clas eh eee ee: 29. Reconstruction of Ramapithecus punjabicusS ....... +. eee ee eee e eee 155 30. Occlusal views of UMP 66-41, UMP 62-07, and YPM 13834 ........ 155 31. Comparative views of Dryopithecus africanus, Pan paniscus and Ramapithecus PUnjavicus . ... cee eines cable Hee or eames eee rls 156 eel ee a ee oe | oOmeanN OD oP bo nO NO IS HN NM NN SIO oF oF NN = © no bo oOo ©9 ©8 ©9 ©9 ©9 9 or BP Cfo NO — © ©9 ©9 69 coonwI me . Potassium—argon dates . Facial measurements of UMP 62-11, Gorilla and Pan . Alveolar and palatal measurements of UMP 62-11, Gorilla and Pan ... . Measurements of central incisors of UMP 62-11, Gorilla and Pan .... . Measurements of lateral incisors of UMP 62-11, Gorilla and Pan ...... . C! measurements of UMP 62-11 and Gorilla ee et oN ESO OND OH OO LIST OF TABLES le} 1@) 109. .ei).6) 0, (6: a) (w' let vie ee (6 1s fe) =e) 6) okie P23 measurements of UMP 62-11, Gorilla and Pan © 0 (6/ 70.10) ef. 0) elie! (@) e1e) eo) 0) «1 hele 4 .P4 measurements of UMP 62-1, Gonlla and Pan 3.256.004 ete nun: . M1! measurements of UMP 62-11, Gorilla and Pan . Crown and root dimensions of UMP 62-11, Gorilla and Pan ........ . M2 measurements of UMP 62-11, Gorilla and Pan ................+. * M® measurements of UMP 62-1), Gorilla and Pam .... 4. 6.025655 . Measurements on the upper dentition of UMP 62-11, Gorilla and | EAH he oe WOR RAIS Co ELLER PUES Te ee APR AL BOM FEMME ged Rene echt geet CRI Ce . Comparative measurements of C1 in D. (P.) major specimens ........ . Comparative measurements of M! in D. (P.) major specimens ........ . Comparative measurements of M2 in: (a) D. (P.) major, (b) G. gorilla ~ My measurements of ).(P.) major and (G.ontlla rience ter sete eck . Comparative measurements of P, in: (a) D. (P.) major, (b) G. gorilla .. . Comparative measurements of P, in: (a) D. (P.) major, (b) G. gorilla and HERE TOL OCU CSi as, 2 aie a tetayerrie eves ey sie lese Waele se shee ey eeikey eno ety ena oss Hata ere iet- . Comparative measurements of M, in: (a) D. (P.) major, (b) G. gorilla .. . Comparative measurements of M, in: (a) D. (P.) major, (b) G. gorilla . . Comparative measurements of Mg in: (a) D. (P.) major, (b) G. gorilla .. . Comparative measurements of C, in D. (P.) major and G. gorilla ...... . Measurements of D. (P.) major and G. gorilla mandibular dentitions .. . Measurements on upper and lower dentitions of D. (P.) major ........ . Ratio (B 1,/L P3-M;) x 100 in G. gorilla and P. troglodytes ........ . Measurements of mandibular tooth rows of G. gorilla and P. EOC OCS CCS ae ael ls wctn port ated cae er tLe tarog oeaTa cue) sarees seas Oh sielion hae (Stee Gone . Ratio (B C,/L P3-M;) x 100 in G. gorilla and P. troglodytes ........ . Comparative measurements of dP, in: (a) D. (P.) major, (b) G. gorilla INC 1es LOCI OGY LES ex fat. opoiet otese ta gege tote ieee fe she one ee ot eatery oney sy aatioe . Measurements of D. (P.) major mandibles ............++++eeeeee eee . Measurements of D. (P.) major mandibular dentitions ............--. mE OWwermolars ol ulcertaim status =o. see ee where ei iense ste 2 ete ee Ct measurements OL a (F2 \AfseCAnUs 7. as). 2 eye Sar eie airs 2s eieini= = siete . Ratios of MDi lengths of D. (Proconsul) specimens, Gorilla and Pan ... . Position of the range for 24 dental dimensions of D. (Proconsul!) SISCCIES ss a5; eter ses op foile cieicp stocks apsealia! Oss Sue peeve wntngs sik meg es tia, S35), eve feedel ete! ses . Facial measurements of BM(NH) M16647 am O) WER G2 eis ce coer . Dental measurements of BM(NH) M16647 and UMP 62-11 ......... . Measurements of upper teeth of RGMAPUNCCUS! @v.0 2 = oe =e ee a - Ratios (Minimum BuLi B/Maximum BuLi B) x 100 of Ramapithecus, D:(S.) swvalensis and BM(NEN) MNG649) net eee eee . Measurements of lower teeth and mandibles of Ramapithecus ........ . Ratio (M; T/M; D) x 100 of fossil hominids and pongids, Gorilla and PGI seecesectane ts (erate tase te eM Eee tobantaa tole ie ote latcnt ate Mtoe eee . Ratio (crown area of I'/crown area M1 + M2?) x 100 in Dryopithecus ANG RAMA P UNE CUS 4 choi. e.g oss s sie deed Gale OSD AA pales 3» ake Oe . Measurements of East African Pongidae, non-Dryopithecus (Pro- CONSUL 1 Carre ee scien see oe Oe es ak Sk Re ee . Ratio (Max L P?/MDi L P*) x 100 in fossil species, Gorilla and Pan .. . Mandibular measurements of East African early Miocene mandibles . . . Mandibular measurements of R. punjabicus and D. sivalensis ......... - Symphyseal measurements ol Dmyopithecis, 2): 8.0.24. 4 ee eee . Ratios of molar MDi Loi Pengidae and Hominidae.....-2. 2). .5-eee . Ratio of M; (Tal B/Tri B) x 100 in fossil and living pongids......... . Ratio (crown area of I!/crown area M! and M?) x 100 in living PORGIMAC i shoe. oa. 9 sharia td aed sb Ae ee sane Se cleo ome eee 107 109 111 112 117 118 nS 119 121 129 130 LIST OF APPENDICES Gontila gorilla male-maxallae.. 46 as. cislsssi0hc ee ote oye De eeeeeers 158 Gonilla conliatemaleamaxillaes.:. hisses ining bso © occeistea ihke dates ae ee 160 Gorilla. coriia, male: mandibles: an566 sei < 7s oles nose donts mo pide cers 162 Gonlia gortla female mandibles: cit obrie eae eis eye. cleo Ree sist 164 Ran troglodytes male maxilla sic elie ef + cites see ere 168 Panctroclodyies temale amaxallae 25, lie nue oisis «ators s bl iseiols «ey seie eres 170 Pan woe ogyies male mvamdi bless .2ic 1 acts toacrs «mo wsleielte hie cies fe clieneiay tele 172 Lone nocvodyies wemale, manGiblese acinar tele a cern tin er 173 Gorillaicornlla malevcrantal 5. 2ocie 8 so caGe Oi wie aie ioe ace ila kets 174 Conllaorilla female cramial tne. 6c arg seis oe one systole sere o| eiesiei are Save. 175 OM trOClODVLES: GRAMMAL y uepsi nin os toate heist amas cree at ee eee 176 Diryopithecus (Proconsul) ajricanus maxillace. 66 fe. se eG skye sre ie Dryopithecus (Proconsul) africanus. mandibles)... 222. .2.....0..--- 178 Dryvopithecus (Proconswul) myanzae maxillae 2403... .52 6.4. sone 180 Dnyopithecus (Proconsuly myanzde, mandibles (7.1). 14)-)-16)- 2215 = 22 ee 18] Dryopithecus(Proconsul) gajor maxilla: amu ce ae eee 182 Drvyopithecius (broconsi jomajon mandibles 2. ayy) nei ata ses a 183 Statistics of Dryopithecus (Proconsul), Gorilla gorilla and Pan troglodytes 184 ab nee Vip y +8 ©... ABSTRACT Recently recovered pongid specimens of Dryopithecus (Proconsul) major from Moroto and Napak in Uganda are described and compared to material of the same species from Songhor and Koru in Kenya. This species is probably ancestral to Gorilla gorilla. Although smaller than the living species, with rela- tively smaller anterior teeth, less lophodont and hypsodont cheek teeth, shal- lower palate, less prognathous face, and without a simian shelf, D. (P.) major may already have been adapting to the predominantly herbivorous diet typical of the living species. D. (P.) major is found in deposits of Early Miocene age dated to around 18 to 20 million years. The species exhibits sexual size dimor- phism similar to that shown by its putative descendant. It was a quadrupedal form, probably capable of arm-swinging like some of the Ceboidea. A second species, D. (P.) africanus, from deposits at Songhor, Koru, and Rusinga, is related to D. (P.) major and may be ancestral to Pan troglodytes. The postcranial remains of D. (P.) africanus indicate that this was a lightly built, quadrupedal, mainly arboreal species with some adaptations suggesting arm-swinging behavior. A third species, D. (P.) nyanzae, is discussed. It is the most primitive of the three species and may resemble closely the species ancestral to D. (P.) major. The Late Miocene and Early Pliocene hominids from Kenya and India are described in detail, and it is concluded that they represent a single genus, Ramapithecus, and probably the same species. This taxon, although retaining some morphologically primitive features, resembles later hominids in the pro- portions of its incisors, canines, premolars, and molars, and in its shallow and robust mandible. Hominid species earlier than this are not definitely known, the material classified by Leakey as Kenyapithecus africanus being more probably a pongid, perhaps with some relationship to Dryopithecus fontani and D. (Siva- pithecus) stvalensis of Eurasia. The interrelationships of the living pongids, their Mid-Tertiary ancestors, and the Mid-Tertiary hominids are discussed, and it is concluded that the dif- ferentiation of hominids and pongids, and of individual pongid species, occurred earlier in time than is generally supposed. ZUSAMMENFASSUNG Neue entdeckte Pongid Exemplare der Dryopithecus (Proconsul) major von Moroto und Napak in Uganda werden beschrieben. Sie werden mit Material derselben Art von Songhor und Koru in Kenya verglichen. Diese Art ist wahr- scheinlich ein Vorfahr des Gorilla gorilla. Obwohl kleiner als die heutige Art, mit relativ kleineren vordern Zahnen, weniger lophodonte und hypsodonte Backenzihne, flacherer Gaumen, weniger vorspringendes Gesicht, und ohne Torus transversus inferior, hat D. (P.) major viellicht schon angefangen sich an der vorwiegend pflanzenfressende Ernahrung typish von heutige Art anzupassen. D. (P.) major wurde in unteren Miocinablagerungen gefunden, die ungefahr 18 1 Z PEABODY MUSEUM BULLETIN 31 bis 20 Millionen Jahre alt sind. Die Art zeigt einen geschlechtlichen Gréssen- dimorphismus ahnlich dem seines vermeindlichen Nachkommen. Es war eine vierfiissige Form die vermutlich in der Lage war die Arme zu schwingen wie einige der Ceboidea. Eine zweite Art, D. (P) africanus, der in Ablagerungen bei Songhor, Koru, und Rusinga gefunden wurde ist verwahnt mit D. (P.) major und mag eine Vorfahre der Pan troglodytes sein. Der hinteren Schadelreste von D. (P.) africanus zeigen dass dieser eine leicht gebaute, vierfiissige, meist auf Baumen lebende Art war mit einigen Anpassungen, die ein Arm-schwingendes Verhalten vermuten lassen. Eine dritter Art, D. (P.) nyanzae, wird erwahnt. Es ist die primitivste der drei Arten und mag der vorangehenden Art des D. (P.) major sehr ahnlich sein. Die obere Miocin und untere Pliocin Hominiden aus Kenya und Indian werden im einzelnen beschreiben. Es wird daraus gefolgert dass sie eine Einzelgat- tung, Ramapithecus, und wahrscheinlich derselber Art darstellen. Obwohl dieser Taxon einige morphologisch primitive Eigenschaften behalt, ahnelt es den spat- eren Hominiden in den Proportionen der Schneide-, Eck-, vordere Backen- und hintere Backenzahne und in seinem flachen und kraftigen Unterkiefer. Friihere Hominidarten sind nicht sicher bekannt. Das Material, das von Leakey als Kenyapithecus africanus klassificiert wird, ist vermiitlich eine Pongid; etwas Ver- wahntschaft mit Dryopithecus fontani und D. (Sivapithecus) sivalensis von Eurasian ist vielleicht vorhanden. Die Zusammenhangen der heutigen Pongiden, ihrer mitteltertiaren Verfah ren, und der mitteltertidaren Hominiden werden erwahnt. Es wird deraus gesch- lossen dass die verschiedenartige Entwicklung von Hominiden und Pongiden, und von einzelnen Pongid Arten friiher stattfand als allgemein vermiitet wird. TERTIARY PONGIDAE OF EAST AFRICA 3 PE3SIOME HeyaBHoO BBHKOMAHHBIe MOHTHIHEIe 9ksemiIapn Dryopithecus (Proconsul) major ¥3 Mopoto u Hanaka B Yranje onucaHb H cpaBHeHbI ¢ MaTepuasoM, npHHay- JexkaluM TOMY ke Buy, 43 Conropa u Kopy B Kenuu. dor Bux BeposATHO mpexoK Gorilla gorilla. XOTA MeHbINe WUBYMeTO BUA, © PeATHBHO MeHbINMMH NepexqHuMH 3yOaMM, MeHee JOOAOHTHBIMH M THICOAOHTHBIMH KOPeCHHBIMH syOaMu, MeHee ray- OokHM HeOOM, MeHee HPOTHaTHHIM AM0M H 6e3 torus inferior transversalis, D. (P.) major MOT yike IpHclocoOAATbCA MpeoOsazawMme TpaBoATHOMy oOOpasy NuTaAHHA, THIMYHOM JIA ikuByMero Bua. D. (P.) major HaXO{AT B HUKHEMMOMeHOBEIX OTIO- HKeEHHAX; OmpezeteHo, YTO HX BOspacT — 18-20 muaIM0HOB Jet. Buy moKasbBaerT MOIOBOK TUMOP(HsM B BelWaMHe, NOAOOHEI TOMY NOKa3syemMomMy ero mpesuortaraeMbIM MOTOMKOM. OH ODI YeTBEPOHOTHUM, H BEPOATHO, CHOCOOHDIM WepeABULaTbesA Kauasch HA pykax, 110,00HO HekOTOpHIM Ceboidea. Bropoft Bugz, D. (P.) africanus, u3 oTaomenui B Conrope, Kopy u Pycuure, B pogctBe c D. (P.) major 4 MoxeT OnITS WpexKom Pan troglodytes. NocrKpanuasbuble octaTku D. (P.) africanus 10Ka3bBal0T, YTO 9TO OBIT BU AeCTKOTO CTpoeHHA, UeTBEpO- HOTHM, NPeUMyMecTBeHHO JpeBecHOTO OOpasa iKM3HU, C HeEKOTOPLIMH Upucmocodse- HHAMH HaBOJAUIMMH Ha MUICIb O WepeABHKeHHU, Kayasdch Ha PyKaXx. Tpetuii Bug, D. (P.) nyanzae, Toe paccMaTpuBaercsa. ITO caMblit IpMMUTUB- Hii U3 TpeX BUJOB H MOET OBITS OYEHD NOXOKUM Ha pesKa D. (P.) major. BepXHeMHOMeHOBLIe H HUKHEMIMONeHOBEIe TOMUBN AL Kenun u Waanu onucannt OOCTOATEIBHO MU 3aKIIOUAeTCA, YTO OHM UPNHAsAexaT OHOMY posy, Ramapithecus, u BePOATHO, OJHOMY BUJY. ITOT TAKCOH, XOTA OH HM COXPaHAeT HEKOTOpPHIe MOpdosoruye- CKH IPHMHTHBHbIE OCOOCHHOCTH, MOXO%# HA 103 ,HHX TOMHHMOB pasMepaMH eFO pesi{oB, KIBIKOB, IPCMOIAPOB HM MOAAPOB, eFO HeTAyOOKOH uM TOACTOH YercTbIO. O ToMMHH- KOBBIX BUaX [peBHee YTOTO He 3HaeM ONPeseeHHO, TAK Kak MaTepHad, Klaccuunn- popaHHHt JIuKu kak Kenyapithecus africanus 601ee BepOsTHO ABAACTCA MOHTHIOM, BO3MO‘KHO B KakoOll-To PopMe pogzcTBa c¢ Dryopithecus fontani u D. (Sivapithecus) siva- lensis Eppasun. PaccMaTpHBawTCA B83aHMOOTHOMICHHA KUBYNIMX MOHTUOB, UX CpeHeTPeTHYHEIX Ipe{KOB H CpeHeTPeTHYHEIX TOMHHUAOB. SakIWUAeTCA, UTO TUPPepenmMaluAa TOMU- HOB H MOHTHOB, a TAakKe HHAUBUAYAIbHEIX MOHTHAHEIX BUOB, COBepmMuAach paHee yeM STO OOHYHO IpeqMoAaraercs. CHAPTER I. INTRODUCTION In a preliminary study, Elwyn Simons and I (1965) outlined the history of study of the various species of fossil pongids classified in the Dryopithecinae. One prob- lem which concerned us was the great proliferation of specific names, many of which were based on the most fragmentary material. During the course of our analysis, it became clear that the number of valid names was rather small, certainly less than a dozen; these species could be accommodated comfortably in a single genus, Dryopithecus. We did not examine at any length the inter- relationships of the various species of Dryopithecus, nor did we discuss in great detail the evolution of the living pongids from particular Dryopithecus species. Thanks to the generosity of Dr. W. W. Bishop, and through the courtesy of Prof. E. L. Simons, a sample of pongid material from the Miocene of Uganda was made available to me for study. The detailed analysis of these specimens forms the core of this study and has required further work on the extensive series of pongids recovered from the Early Miocene of Kenya. The Ugandan material was classified originally as Proconsul; this genus was first diagnosed by Hopwood in 1933 (Hopwood, 1933a). Simons and I (1965) transferred species of Proconsul to Dryopithecus, and throughout this study these species are de- scribed as Dryopithecus (Proconsul). The most comprehensive study of D. (Proconsul) species is to be found in the monograph by Clark and Leakey published in 1951. These workers expressed the opinion that D. (Proconsul) species were pongids and that many of the characters in which they differed from the living great apes were merely primi- tive and should not be used to exclude them from the Pongidae. Leakey, however, no longer classifies these African Miocene species in the Pongidae, but has proposed a separate family for them, the Proconsulidae. He regards other Dryopithecus species as pongids. He has listed in recent publica- tions some of his reasons for advocating a new family. “I have suggested the setting up of this family for certain ape-like fossil creatures which differ from the true Pongidae in lacking a simian shelf, in the structure of their canine teeth, in the position of the root of its molar-process maxillary [sic], in the absence of a torus, in the nature of their mandibular condyles and the way their teeth are arranged in the mandible, etc.” (Leakey, 1963, p. 40). “By contrast to any living pongid, Proconsul has a rounded, smooth forehead and a low, rectangular orbit rather than a high, round one. In not just one but many specimens of four species, Proconsul has a relatively small canine with a molar-premolar series converging forward, as opposed to the backward convergence in pongids. .. .” (Leakey, 1965, p. 13). Leakey therefore regards these species as belonging to a genus distinct from the genus Dryopithecus to which the Eurasian species belong. He also believes that the species of D. (Proconsul) are fundamentally very similar to one another. The present study involves an examination of new material of D. (Proconsul) major and a comparison of this material with other D. (Proconsul) species; it 5 6 PEABODY MUSEUM BULLETIN 31 shows that the subgenus is by no means as homogeneous in some features as Leakey has suggested. Further specimens of Dryopithecus from East Africa distinct from D. (Pro- consul) have been described by MacInnes (1943) and by Clark and Leakey (1951). Clark and Leakey placed these specimens in Sivapithecus africanus, a species transferred by Simons and myself (1965) to Dryopithecus. Leakey (1967) has since proposed that these specimens, together with other newly described individuals, constitute a new species within the Hominidae. A detailed analysis of Leakey’s paper and of the material indicates that this allocation is dubious and that the East African specimens are better classified at the moment in Dryopithecus. The interrelationships of the African Dryopithecus species are discussed, as well as the light they throw on the evolution of the African great apes. Their relationship to Eurasian species of Dryopithecus and to the earliest hominid, Ramapithecus punjabicus, is also examined in some detail. ACKNOWLEDGEMENTS Grateful thanks go to the Curator and Trustees of the Uganda Museum for permission to study material in their collections. I should like to thank the following for allowing me access to material under their care: Drs. L. S. B. Leakey, J.-P. Lehmann, J. Biegert, K. P. Oakley, M. Crusafont Pairo, M. C. Mc- Kenna, and E. L. Simons. I am most grateful to the following for their advice and help: Drs J. C. Cutbill, D. B. Williams, E. H. Ashton, W. W. Bishop, A. H. Schultz, C. P. Groves, J. R. Napier, B. G. Campbell, and A. C. Walker. Mr L. P. Morley and Rosanne Pilbeam took some of the plates; some were kindly loaned by Sir W. E. Le Gros Clark, others by Dr E. L. Simons and Dr A. C. Walker. Graphs, tables, and figures were drafted by Rosanne Pilbeam. The research was carried out during tenure of Fellowships from the Graduate School of Yale University, and the Wenner-Gren Foundation; a grant was also received from the Boise Fund of Oxford University. This Bulletin has been published with the aid of a National Science Foundation Publication Grant No. GN-475. Finally, I should like to thank Rosanne Pilbeam, whose help as typist, drafts- woman, photographer, and general editor has made the completion of this dissertation possible; and Elwyn Simons, my friend and supervisor, who has guided me through the evolutionary biology of the higher primates and who has made the work involved so enjoyable. CHAPTER II. MATERIALS AND METHODS To make a meaningful contribution, adequate comparative material is essential in a study of this type. Since this is predominantly a study of African fossils, the living African apes have been selected for the comparisons. Samples were measured of both sexes of Western gorillas, subspecies Gorilla gorilla gorilla (Groves, 1967), and of the chimpanzee subspecies Pan troglodytes troglodytes. These are generally referred to as G. gorilla and P. troglodytes in the text, tables, and appendices. The measurements on their skull and dentitions are included in Appendix I. Some statistics such as means, ranges, standard devia- tions, and so forth, are available in the literature (Schuman and Brace, 1954; Ashton and Zuckerman, 1950), but more of this sort are included here. However, it is difficult to gain access to raw data and so the measurements used in this study are included in full. Other workers may then calculate, for example, their own indices, correlation coefficients, and regression coefficients, as well as any other statistics they may require. MEASUREMENTS Dental and mandibular measurements were taken on 20 male and 20 female gorillas and 14 male and 12 female chimpanzees. Palatal and facial measure- ments were taken on 20 male and female gorillas and 10 male and female chimpanzees. The facial measurements are either self-explanatory or are ex- plained in the text. Specimens came from the Anthropologisches Institiit der Universitat, Zurich; British Museum of Natural History, London; Anatomy Mu- seum and Duckworth Laboratory of Physical Anthropology, Cambridge, England; and the Museum of Comparative Zoology, Cambridge, England. Dental measurements and their abbreviations are explained below; they were taken to the nearest tenth of a millimeter with engineer’s straight-jaw vernier calipers with specially sharpened points. All measurements are in millimeters unless otherwise stated. Good estimates of measurements are marked with one asterisk; poor estimates are marked with two asterisks. Generally, two basic measurements are given for each tooth, a length and a breadth. The length is usually mesiodistal and is measured between the mid- points of the mesial and distal borders with the calipers held vertical to the occlusal plane. Canines and first lower premolars have maximum lengths. First upper premolars have both a mesiodistal and a maximum projected length. Breadths are measured normal to lengths and are projected; they are termed transverse in the case of canines and lower first premolars, labiolingual for incisors, and buccolingual for upper premolars and molars. These breadths are maxima. Lower second premolars have buccolingual breadths. They also have trigonid and talonid lengths, measured from the midpoint of the talonid basin to the mesial and distal margins. Lower molars have three breadths: trigonid, / 8 PEABODY MUSEUM BULLETIN 31 talonid, and maximum. Heights are measured from the midpoint of the enamel line labially or buccally to the tip of the crown or the midpoint of the occlusal surface or edge. Certain bilateral measurements have been taken on the tooth row as a whole. Four breadths are listed; between the distal alveolar margins of upper and lower lateral incisors (I1B and I,B), and on the crown across canines, second premolars, and third molars (C1 B, C, B; P* B, Py B; M? B, M; B). The length of the cheek tooth row from the mesial border of the first premolar to the distal border of the third molar was also taken (P?-M® L, P3;—Msz; L). The depth of the mandible from the alveolar margin is measured at the mesial margins of the second premolar and third molar (P, D, M; D). Where possible, mandibular thickness normal to the long axis of the cheek teeth is taken at the mesial border of the second premolar (P, T), and at the mesial border of the third molar (M; T). The depth of the mental foramen is measured below the alveolar margin (Ment for D). A maximum of 18 measurements could be taken on upper jaws and teeth and 28 measurements on the lower. As complete a set of measurements as possible was taken on each fossil; extra measurements other than those listed here are explained in the text. Abbreviations for museum numbers are also included here. ABBREVIATIONS L length B breadth H height D depth T thickness MDi mesiodistal La labial LaLi labiolingual Li lingual BuLi or BLi _ buccolingual Tri L trigonid length Tri B trigonid breadth Tal L talonid length Tal B talonid breadth I incisor (I! = upper incisor, I, = lower incisor) C canine (C1 = upper canine, C, = lower canine) P premolar (P* = first upper premolar, P, = second lower premolar) M molar dP deciduous premolar MNHN Muséum National d’Histoire Naturelle, Paris UMP Uganda Museum, Kampala BM(NH) British Museum of Natural History, London NMK National Museum Centre for Prehistory and Palaeontology, Nairobi, Kenya YPM Peabody Museum, Yale University, New Haven, Conn. TERTIARY PONGIDAE OF EAST AFRICA 9 GSI Geological Survey of India, Calcutta AMNH American Museum of Natural History, New York CGM Cairo Geology Museum, Cairo The object of taking morphological measurements is to reflect the genotypic variability in the population. Measurements for bilaterally symmetrical struc- tures are taken on one side of the jaw only; there is no extra information to be gained from taking the length of M3, for example, on both sides of the jaw, since both teeth are “reflections” of the same genotype. Any differences in size are phenotypic, and statistically speaking can be ignored (Ashton and Zucker- man, 1950, p. 474). Similarly, if statistical constants are calculated for living or fossil populations, pairs of teeth from opposite sides of the same jaw should not be included, since this will mean that one genotype is sampled twice and will therefore bias the calculations. Measurements can also be altered by wear, weathering, and by compression from adjacent teeth. These are dentally non-genotypic factors. Although the degree of compression, for example, may be an interesting phenomenon and one which should be studied, such factors ought, as far as possible, to be ignored. Attempts have been made here to compensate for tooth wear, rotation, and other variables, thus reproducing best estimates of the measurements of the tooth as it was when newly erupted. Estimates of this sort are also marked with an asterisk. UNIVARIATE STATISTICS The standard univariate statistics have been calculated from the gorilla and chimpanzee samples. Calculations were performed on the Titan Computor in the University Mathematical Laboratory at Cambridge University, using a pro- gram written by Dr. J. L. Cutbill. The statistics included in Appendix 1 are the mean and its standard error; the standard deviation; the coefficient of variation corrected for small samples (Simpson, Roe and Lewontin, 1960, p. 101-102); the range; and 95 per cent confidence limits. The standard errors in no case exceed 5 per cent of the mean and it can therefore be assumed that samples are of adequate size. ABBREVIATIONS n sample size xX mean SE standard error SD (or s) standard deviation Win corrected coefficient of variation OR range 9577, GL, 95 per cent confidence limits Sample standard deviations are calculated using the formula, 10 PEABODY MUSEUM BULLETIN 31 where x is any individual sample variate. The corrected coefficient of variation is calculated using the formula, S 4n+1 | esto) The sample confidence limits are calculated using the formula X = ts, where t is Student’s t and has n — 1 degrees of freedom at the appropriate level of probability, in this case 0.05. Other workers may calculate their own con- fidence limits using this formula; all other necessary information is given in Appendix 1. The comparative tables in the text include means and confidence limits for raw measurements, and means and ranges for indices and ratios. Strictly speak- ing, ratios and indices cannot be treated as single variates for the purposes of calculating confidence limits. Simpson, Roe and Lewontin (1960, p. 164) give a method for calculating confidence limits when numerator and denominator are uncorrelated. “If the upper and lower limits for the numerator are written Ay and Ay and the limits for the demoninator symbolized as By and B,, confidence limits for the ratio are given by A,/By and A,y/B,. The confidence value for this interval is the product of the individual confidence levels. If the confidence limits for numerator and denominator are those for a 95 per cent interval, the confidence interval for the ratio has a confidence level of 90.25 per cent (.95?). If a 95 per cent level is wanted for the ratio, the confidence limits for numerator and denominator must represent the 97.5 per cent limits (\/.95 = .975).” However, since ratios and in- dices are only of use when the variates involved are correlated, they should not be used if numerator and denominator are uncorrelated (in such cases, ratios contribute no more information than the original variates). When variates are correlated, ratios are less variable than the raw measurements and confidence limits calculated in the manner outlined above will then be conservative, a de- sirable property. In all the cases in which ratios have been calculated from fossil material, the samples involved are very small. For this reason, no attempt has been made to test the significance of differences between ratios. The smallness of the samples should constantly be borne in mind when evaluating the fossil data. In a few cases, equiprobability ellipses have been calculated according to the method outlined by Defrise-Gussenhoven (1955). These ellipses encompass any desired percentage (in this case 95 per cent) of a bivariate population. Accurate confidence limits for ratios can be calculated by noting the points on the ellipse at which the values of the ratio are highest and lowest. The positions of these points vary with the slope of the major axis of the ellipse and the product moment correlation coefficient between the two variates. These in turn depend on the variances of the variates, and their covariance. In cases of fossil samples smaller than eight, ranges rather than confidence limits have been given and standard deviations have not been calculated. TERTIARY PONGIDAE OF EAST AFRICA i MULTIVARIATE ANALYSIS It was considered desirable to use a multivariate technique which dealt with individuals rather than with samples, one which sorted individuals into groups rather than one which required individuals to be placed a priori into groups before analysis could begin. In this particular case where the sorting of individ- uals is important, and where fossil sample sizes were small and in some cases clearly biased towards one sex of a size dimorphic species, such a technique has been particularly useful (see Chap. VI). The type of analysis used was Principal Coordinates, a modification, by J. CG. Gower (1966), of Principal Components Analysis. The computer program was written by Dr. J. L. Cutbill, for whose help (and mathematical tuition) I am greatly indebted. Principal Components can most easily be explained in two dimensions. Consider a bivariate sample plot surrounded by, say, a 95 per cent equiprobabil- ity ellipse. The original individual points are in x, y space. If each point is now projected onto the major and minor axes of the ellipse, the points can be given new coordinates along these axes (say, a and b) instead of along x and y. If x and y were originally fairly strongly correlated, the coordinates along the major axis will account for most of the variance in the sample. The coordinates along the minor axis will account for the remainder of the variance. As a general case s,? + s,? = s,” +s,?, where s,? is the sample variance of x, etc. The sum of the original variances is equal to the sum of the new variances. The total variance remains unchanged, although the distribution of the variances changes, the tendency being to “squeeze” as much of the variance into one dimension as possible. ‘The new variances are the latent roots or eigenvalues of the 2 xX 2 covariance-variance matrix of the original variates. One valuable property of the new coordinates a and b is that they are orthogonal and the new variates are uncorrelated. This model can now be considered in more than two dimensions. If p measure- ments are taken on each of n individuals, the sample can be represented as a swarm of n points in p-dimensional space. This swarm can be surrounded by a p-dimensional hyperellipsoid. The major axis, or first principal component, of this ellipse can then be found, and this axis will account for “most” of the variance in the system. The next axis, or second principal component, will account for “most” of the remaining variance and so forth. Thus, the original n points can now be represented along the first few new coordinates, and the sample can be plotted on one or more two-dimensional graphs. Using the first few axes only, some accuracy is inevitably lost, since only a certain percentage of the total variance is accounted for by the first few dimensions. However, if this total percentage is reasonably high, the remainder can be safely ignored. Principal Components Analysis generally takes the p sets of raw data on n individuals and produces a p X p matrix of variances and covariances. This matrix is then manipulated to produce the new axes and coordinates along these axes. This is known as an R mode method of analysis, in which attention is concentrated on the relationships between variables. The type of Principal 12 PEABODY MUSEUM BULLETIN 31 Coordinates Analysis used here differs from most Principal Components Analyses in that it is a Q mode method, utilizing relationships between individuals. ‘This type of analysis is ideal for sorting individuals and groups of fossils. In any multivariate method it is important to standardize the variates used. If this is not done, large dimensions like P;-M, L will dominate smaller ones such as MDi L P,. Generally, variates are transformed to dimensionless quanti- ties by dividing by the sample standard deviation. However this has no mathe- matical superiority and other normalizers can be used, for example, the mean or the range. In Principal Components Analysis the product moment correlation coefficient is used as a measure of association between variates (R mode). Principal Co- ordinates uses here a similarity coefficient for the Q mode analysis of individuals. This coefficient is calculated according to the following formula, Be, CRM peal oe p The similarity coefficient s,, between individuals y and z is obtained as follows. The difference between the values of variate a for individuals y and z is calculated (ja, — a,|) and standardized by dividing by the sample range for character a (OR,). This operation is performed for each of p characters and the total summed from the first to the pth character. The total is subtracted from p and this value is then divided by p to give the coefficient. When two individuals have identical values for each variate, will be zero, and the coefficient will then be 1. When two individuals in a sample are at the opposite extremes of the range for every variate, the coefficient is 0. The similarity coefficients can be regarded as a measure of distance between each pair of points. These n individuals can be represented in n — | dimensions without distorting the inter-individual distances. The problem is to produce n — 1 new dimensions, so that the first m dimensions, where m < n — I, contain “most” of the variance. The program produced new values for each individual in each of the new coordinates and also produced the percentage of the total variability accounted for by each coordinate. The assemblage was then plotted in these m dimensions with a minimum of distortion. The amount of time taken to extract latent roots from an n X n matrix is proportional to n?. In order to save machine time and to deal with large sample sizes, the program used extracted only the first four latent roots and hence coordinates are available only in the largest four dimensions. Graphs have been plotted in the first two dimensions, these accounting normally for between 60 and 70 per cent of the variability. The third and fourth dimensions con- tribute two or three per cent only and are not used on these graphs. The variates used are the 28 mandibular dimensions described above and listed in Appendix 1. The program was tested first with data from living material, to discover whether or not it could sort individuals of known species into specific groupings; this it did. In the case of gorillas and orangutans it TERTIARY PONGIDAE OF EAST AFRICA 13 sorts according to sex, too. Fossils were introduced into the analysis and the amount of clumping or scattering noted; this was then compared with the range of variation to be found within and between living species (see Figs. a0). Since fossil material is rarely so complete as to provide a full complement of 28 mandibular variates, the program was adapted to cope with gaps in the data. By experimentation it was noted that for adequate comparison between two individuals, they should share values for at least two-thirds of the variates. Measurements on fossil and Recent specimens are listed in the Appendices. Where necessary, measurements have been utilized to illustrate various points. CHAPTER III. GEOLOGIC AND PALEOECOLOGIC CONSIDERATIONS OF THE EAST AFRICAN SITES The African pongid material discussed in subsequent chapters is derived mainly from five areas in East Africa: Moroto and Napak in Uganda; and Rusinga, Songhor, and Koru in Kenya (Fig. 1). During the Miocene these deposits were associated with a number of volcanic centers. Those at Rusinga contain material from the Rangwa volcano and those at Songhor and Koru material from Tin- deret volcano. The sites at Moroto and Napak take their names from the respec- tive volcanoes on the slopes of which the sites are found. These sites have been dated as Early Miocene or “Burdigalian”, although the application of European faunal stage names is probably not strictly appro- priate (Andrews, 1911; Clark and Leakey, 1951; Bishop, 1962). It is, however, unlikely that all these sites are contemporaneous. The potassium-argon dating of the sites will be discussed briefly later (p. 17). The geology of the Uganda de- posits has been covered most adequately by Bishop (1962, 1963a, 1964, 1966, and 1967) and Bishop and Whyte (1962). This description draws heavily on their work. Bishop is now working on a monograph on the stratigraphy and paleoecology of the Uganda sites which will cover more thoroughly the subjects of this chapter. MOROTO II KARAMOJA DistrRicT, NORTHEASTERN UGANDA This site was first discovered in 1961 by Bishop and Whyte (1962) and has yielded an excellent pongid palate, UMP 62-11. The fossiliferous deposits consist of water-laid sediments infilling steep-sided valleys cut into the Precambrian basement gneisses. Most of the sedimentary material is derived from the gneisses, although some is of volcanic origin; fossiliferous horizons have a calcareous cement. Bishop (1966, p. 157) noted that some of the lavas at Moroto II may be later than Early Miocene in age. One potassium-argon date on the capping lavas has been obtained by Bishop, Fitch, and Miller (1969, in press). It is 14.3 + 0.3 x 106 years. This gives a latest possible date for the deposits. The associated fauna is rather sparse and poorly preserved, and apparently does not differ from that at the other Miocene sites (Bishop, 1967). However, as noted, the potassium-argon date is a latest age estimate and so there is a possibility that the site is somewhat younger than the other sites discussed, perhaps middle Miocene. NAPAK KARAMOJA DisTRIcT, NORTHEASTERN UGANDA A number of fossil sites have been found in the vicinity of the dissected remnant of the Napak volcanic cone, some 55 miles southwest of Moroto. The sites have 14 TERTIARY PONGIDAE OF EAST AFRICA 15 FIGURE |. FOSSILIFEROUS LOCALITIES LOSODOK © UGANDA + KENYA KIRIMON, TAMBACH %} on BARINGO N) MARIWA ONO ee eKORU °FORT TERNAN LAKE VICTORIA After Bishop (i963) been divided by Bishop into two groups on the basis of lithology (Bishop and Whyte, 1962). The first includes sediments made up of sands, gravels, and tufts occurring beneath volcanic sediments and resting on basement rocks (sites II, III, VI, VII, and VIII). Moroto II is also a site of this type. The second group consists of bedded tuffs and agglomerates within the volcanic pile, approxi- mately 500 feet above the basement (sites I, IV, V, IX, and X). The pongids from Napak come only from sites I, IV, and V; these sites are treated as a contemporaneous group. The pyroclastic sediments are usually medium to fine tuff and the fauna is generally very well preserved due to its burial in such sediments. The high incidence of teeth, jaws, and cranial fragments, some with signs of rodent and carnivore gnawing, suggests that these faunas largely represent less edible bones left by predators on the temporary land surfaces. Fortunately, secondary calcification has meant that the fossils have been remarkably well 16 PEABODY MUSEUM BULLETIN 31 preserved. Resistant fossils were secondarily concentrated in patches as they were weathered out. The fauna has been discussed by Bishop (1964, 1967) and appears to be broadly similar to those from Rusinga and Songhor in Kenya. Detailed analysis of faunas from the various sites at Napak shows interesting differences between the faunas, presumably due to ecological variation be- tween contemporaneous assemblages (Bishop, 1967). Bishop (1964, p. 1331) reported a date of 19.0 + 2.0 x 10° years for biotite from Napak I. A further sample from the same site yielded an age of 17.8 + 0.4 x 10% years for three age determination runs [Bishop, Fitch, and Miller, 1969 (in press)]. The second age however is perhaps less accurate; the Napak deposits are probably latest Early Miocene. As stated above, it is possible that the Moroto II material is a little younger than that from Napak. KORU KAVIRONDO DistricT, KENYA Primate material from Koru was first reported by Hopwood (1933a and b). The deposits are mainly of subaerial volcanics like those of Napak I, IV, V, and IX. SONGHOR KAVIRONDO DisTRIcT, KENYA The main site was discovered in 1932 by Leakey and MacInnes. Like Napak I, IV, V, and Koru, the primate-bearing deposits at Songhor consist of subaerial pyroclastics. Both Koru and Songhor occur at the base of the same pyroclastic sequence, and they may be regarded as contemporaneous. Sediments in these sites origi- nated from the same volcanic crater. Like Napak I, IV, and V, the deposits at Songhor and Koru probably represent a relatively short period of time. The following radiometric dates have been obtained from material collected at Koru and Songhor (Bishop, Fitch, and Miller, 1969, in press): Koru, 19.5 + 0.3 X 10° years and 19.6 + 0.3 x 106 years; Songhor, 19.7 + 0.5 x 108 years and 19.9 + 0.6 x 10 years. The two sites can be treated as contemporaneous. ‘They fall at the end of the Early Miocene. RUSINGA ISLAND LAKE VICTORIA, KENYA Leakey and MacInnes first located hominoids on Rusinga Island on the south side of the Kavirondo Gulf of Lake Victoria in 1931. Since then the island has been searched extensively for fossils and a large amount of material has been recovered. The island is 170 miles south of Napak and 75 miles west-south-west of Songhor. These deposits have yielded faunas, including primates, more ex- tensive than those from the other East African sites. Their geology has been TERTIARY PONGIDAE OF EAST AFRICA iLi/ described by Shackleton (1951), Whitworth (1953), Bishop (1967), and Van Couvering and Miller (1969). Recent stratigraphical work on Rusinga Island has been completed by John Van Couvering (Van Couvering and Miller, 1969). Briefly, he concludes that the main fossiliferous horizons at Rusinga are some 18 million years old. Some specimens may date back to 20 or 22 million years, but most of the pongid material is probably between 18.5 and 17.5 million years old. PALEOECOLOGY The various East African Early Miocene sites are almost certainly not exactly contemporaneous; more careful faunal collecting and radiometric dating should in the future enable us to build up a chronological sequence of sites in East Africa. The known potassium-argon dates are set out in tabular form in ‘Table 1. Dates for the European sequence are drawn from von Koenigswald (1962), Table 1 Potassium -argon dates (1 0° years) EUROPE EAST AFRICA LA = V2Z.or0.4 SARMATIAN 14.0 514. 340.3 14. 0+0. 2 TA, 7007, 14.6+0.5 14,7+0.5 16. 24+0.4 VINDOBONIAN 17.8+0.4 19. 640.3 | 19. 7+0.5 19.9+0.6 18 PEABODY MUSEUM BULLETIN 31 and those for East Africa from Evernden et al. (1964), Van Couvering and Miller (1969), and Bishop, Fitch, and Miller (1969, in press) . The faunas of the earlier Miocene sites are broadly similar, although careful studies have emphasized differences in relative frequency of various groups (Whitworth, 1958; Bishop, 1967). With the sites arranged in chronological order, no doubt evolutionary changes within lineages will be detectable. Studies of this sort, however, are not possible with the present state of knowledge of these faunas. The faunal differences between sites are generally interpreted as being due more to ecological contrasts between more or less time-synchronous assemblages than to faunal evolution. Bishop (1967) has shown marked fluctuations in the relative abundance of rodents, proboscideans, and artiodactyls at four of the Napak sites (I, IV, V, IX). These differences, he assumes, are due to variations in local ecologic conditions between contemporaneous sites which are geographi- cally close together. Whitworth (1958, p. 46-47) compared the well-known faunas of Rusinga and Songhor. The differences between the two generally take the form of the presence of one species at Rusinga and its absence from Songhor. For example, the ruminants Dorcatherium chappuisi, D. pigotti, and D. parvum are abundantly present at Rusinga but absent from Songhor. Conversely, D. songhorensis is absent from Rusinga and present at Songhor. Another large ruminant Propalaeoryx nyanzae is present at Rusinga and not at Songhor. The lagomorph Kenyalagomys (two species, K. rusingae and K. minor) is like- wise present in abundance at Rusinga yet absent from Songhor; Megalohyrax championi and other hyracoids follow this same pattern. Whitworth believes that these differences are best regarded as ecologic and concludes (1958, p. 46) that the “fossil assemblage at Rusinga seems to be a representative savannah fauna, and it has been suggested . . . that in early Miocene times Rusinga was principally an area of parkland and steppe. At Songhor, we are dealing with the fauna of an associated, but more restricted, habitat, perhaps analogous to the isolated, tree-clad jebels which rise, here and there, above the Recent African savannahs, and support a mammal fauna distinct from that of the surrounding parklands.” While sympathizing with this general viewpoint, Leakey has recently (1967, p. 163) suggested an alternative explanation. “It is, however, possible that the Songhor beds may only represent the upper part of the Rusinga series, and that both may be of Middle Miocene age, leaving the main Rusinga deposits in the Lower Miocene.” As we have already seen, this thesis is unlikely, the radio- metric dates weighing against it. Chesters (1957) has studied the available palaeobotanical material from the East African Miocene. Most of her material comes from Rusinga and the neighboring island of Mfwanganu. The majority of her Rusinga material consists mainly of fruits rather than leaves. This has made comparison with other Tertiary African floras difficult, since they have been composed predominantly of leaves. The flora of tropical Africa apparently has changed relatively little during the Tertiary. That of the East African Miocene shows that the ecology and climate of East Africa would not have differed greatly from the present day. Considering those species most similar to living forms, and for which habitat can be inferred, the setting suggested to Chesters by the flora is one of a tropical TERTIARY PONGIDAE OF EAST AFRICA 19 rain forest with tall trees and a multiplicity of climbers. Many of the trees are species which grow close to, or at, the forest margin. Fruits adapted for water dispersal are also present in large numbers in the deposits. The flora is indicative of a gallery forest overhanging lakes, rivers, and streams, with trees covered with climbers. The fauna at Rusinga contains forest, marshland, and savannah elements, and presumably the savannah forms inhabited the local grasslands. Gallery forest was present along water courses and around lakes and marshes. The slopes of the volcano were probably thickly wooded. Animals characteristic of the various habitats are found as fossils in lake-shore deposits presumably because they would have visited the lake for water and died there. The larger primates certainly would have inhabited the wooded slopes of the volcanoes. Here, as Bishop (1963a, p. 259) has said, ‘one is tempted to draw analogies with the present habit of the mountain gorilla in Kigezi, southwest Uganda... . This is between 7,500 and 9,500 feet above sea level (1,500-3,500 feet above the surrounding country) on the slopes of volcanoes which have been active until very recently and which are in areas where lakes impounded by volcanic activity abound.”’ This description would probably fit the sites at Songhor and Koru, Napak and Moroto; regions of tree-clad relief standing above local areas of savannah and open woodland. Verdcourt (1963) has studied the nonmarine Mollusca in a valuable con- tribution to the paleoecology of the area. Fortunately, land mollusca are good indicators of climatic conditions. Like the flora, the Miocene mollusc fauna is similar to that of the Recent, emphasizing the fact that this area has been climatically stable. The majority of the specimens studied by Verdcourt again came from Rusinga. Verdcourt concludes (1963, p. 35) that the predominant type of vegetation represented in these Early Miocene deposits is evergreen rain forest receiving a fairly heavy annual rainfall (35-70 inches), heavier than that of the present. Evidence from Rusinga indicates that climatic fluctuations also occurred, for some of the molluscs are indicative of drier habitats such as riverine forest and savannah woodland. The aquatic habitats evidently consisted of swamps and rivers, although not necessarily lakes. ‘““The impression gained is that the area around Rusinga and Karungu was covered with a series of swamps.” (Verdcourt, 1963, p. 35). The predominant vegetation at the Koru and Songhor sites is said by Verd- court to have been evergreen forest, with an implied rainfall of 40-80 inches, hardly surprising for elevated areas. Similar habitats and climatic conditions probably obtained at Napak and Moroto, too. In conclusion, the Rusinga sites contain faunal elements from a variety of habitats. The hominoids found in these sites may have lived near the rivers and swamps, in the savannah, or on the forested slopes of the nearby volcanoes. The sites at Songhor, Koru, Napak, and Moroto are more likely to contain a homogeneous fauna, representative of the local habitat. This habitat would have been not dissimilar to that of the presentday mountain gorilla, Gorilla gorilla beringet. CHAPTER IV. DRYOPITHECUS (PROCONSUL) MAJOR FROM MOROTO II, UGANDA PRESENTATION OF MATERIAL The next three chapters contain descriptions of the Miocene ape Dryopithecus major from Uganda and Kenya. First I shall discuss the morphology and possible relationships of the excellent material from Moroto Il in Uganda. Next, in Chapter V, I shall attempt to demonstrate that all pongid specimens from another Ugandan site, Napak (see Fig. 1), are best assigned to a single species, and that this material is conspecific with that from Moroto II. Finally in Chapter VI I attempt to show that the Ugandan apes belong in the species Dryopithecus (Proconsul) major, first described from Songhor and Koru in Kenya. This species is gorilla-like in a number of features and exhibits—like the gorilla— a high degree of sexual dimorphism in dental, facial, and mandibular characters. HISTORY OF MOROTO DISCOVERIES The history of recovery of the material from Moroto II in Uganda is of partic- ular interest. In August 1961, Dr. W. W. Bishop and Mr. F. Whyte, while in- vestigating a number of mid-Tertiary sites in Uganda and Kenya, discovered the locality known now as Moroto II (Bishop and Whyte, 1962). They recovered part of the left maxilla of a large hominoid, lacking tooth crowns (specimen UMP 62-11), and a fragment of right mandibular horizontal ramus (UMP 62- 10), also of a large hominoid. Four months later, in December 1961, Prof. David Allbrook led a party of students from Makerere University College to the same site. They found a right maxilla and two other fragments of upper jaw, all from the same individual as UMP 62-11. The reconstructed specimen was described by Allbrook and Bishop (1963) and provisionally assigned by them to Proconsul major. Allbrook’s party also recovered a left upper canine of this species (UMP 62-12) and some more mandibular fragments. Collecting ceased until Bishop returned to the site in December 1963. Sieving surface material downslope from the original point of exposure of UMP 62-11 resulted in the collection of 25 further pieces, including most of the teeth and parts of the inferior facial region. To judge from the sharp edges of the broken pieces and the lack of weathering, the maxilla had evidently only recently been exposed on the side of the erosion gulley. Seven further mandibular fragments were also found. One of these belonged to the specimen previously numbered UMP 62-10. The other six made up a rather weathered portion of a left mandibular horizontal ramus (UMP 66-01) belonging to the same individual as UMP 62-10. The fragments of the upper jaw and face, UMP 62-11, came from one side of the gulley, the mandibular fragments, UMP 62-10 and 66-01, from the other. 20 TERTIARY PONGIDAE OF EAST AFRICA D2) The mandibular fragments show a higher degree of weathering, suggesting that they had lain rather longer on the surface before collecting. It is possible that the maxillary and mandibular remains come from the same individual. SPECIMENS UMP 62-11 (Fics. 18 AND 19) This is one of the most complete specimens of a mid-Tertiary dryopithecine known, and has been described and figured by Allbrook and Bishop (1963) and Bishop (1964). One especially pleasing feature is the almost total absence of postmortem deformation. The maxillae, premaxillae, nasals, palatines, and dentition have been reconstructed from a total of 36 fragments found over a period of two and a half years. A further 17 fragments cannot be fitted into place. Most of the lateral walls of the maxillae are missing, as are the lateral parts of the nasal meatus, and the maxillary sinuses are exposed. The premaxillae are complete as are the adjacent parts of the maxillae as far as the distal border of P*, Narrow strips of the maxillary frontal processes are preserved superiorly between the infraorbital foramina and the lacrimal fossae, but they do not reach the frontal. The nasals, which are deficient inferiorly, run superiorly as far as the nasofrontal suture above which is preserved a tiny strip of frontal. A number of isolated fragments can be identified. These include the lateral mar- gin of the left orbit and a portion of what is possibly the left inferior malar region. The dentition is complete except for the right central incisor. FACE NASALS AND NASAL REGION The nasofrontal suture is convoluted. Superior to this, the frontal is repre- sented only by a very small fragment of bone which does not reach the level of glabella. The fragment continues the smooth sagittal curve of the nasal. The nasals are preserved superiorly, but are extremely fragmentary inferior to the level of the infraorbital foramina. The contour of the internasal suture is concave anteriorly. A tangent drawn to the curve at the level of nasion makes an angle of approximately 60 degrees with the occlusal plane of the premolars and molars. At the level of reconstructed rhinion the angle is approximately 30 degrees. (The occlusal plane of the cheek tooth is used here as the plane of orientation since none of the accepted planes are available.) The nasomaxillary sutures are not convoluted and diverge slightly inferiorly. The distance between the two sutures at the nasofrontal suture is approximately 10 mm and at the level of the infraorbital foramina 12.5 mm (a vertical distance of some 40 mm). The sutures continue to diverge and about 15 mm below the level of the infraorbital foramina the sutures are 16 mm apart. The inferior parts of the nasals are not preserved, although the premaxillary part of the nasopremaxillary suture is preserved at its most superior point. The distance 22) PEABODY MUSEUM BULLETIN 31 between the sutures at this level—assumed to be approximately equivalent to the level of rhinion—is about 25 mm. The length of the nasals from nasion to rhinion has been estimated at about 65 mm, within the range for Gorilla gorilla and outside that of Pan troglodytes. In cross-section the nasals form an almost flat plane at the level of nasion but are convex anteriorly at the level of the lacrimal fossae and at this point are set at an angle of approximately 160 degrees to each other. The nasals are therefore long, rather narrow bones, gradually expanding inferiorly from nasion to between the infraorbital foramina and rhinion, and then rather more rapidly as far as their inferior border. In general shape they resemble those of the Cercopithecidae rather than Pongidae, although in size they are matched only by the gorilla. Chimpanzees tend to have approximately parallel nasomaxillary sutures like this specimen, although this is a variable feature. In Gorilla gorilla, the nasals are often constricted superiorly, then expanded greatly in their inferior half (Vogel, 1965, p. 308). Superiorly the margins of the nasal aperture are not preserved. Laterally, the crista nasalis is smoothly rounded. Premaxillary-maxillary sutures are pres- ent 3.5 mm lateral to the lateral borders. These sutures diverge inferiorly to the alveolar border. The crista nasalis on each side curves gently towards the midline, so that the outline of the inferior half of the border of the piriform aperture is shaped like a half-circle when viewed anteriorly. There is a flattened prenasal area anteroposteriorly some 10 mm long, running from the anterior margin of the incisive fossa to the superior margin of the nasoalveolar clivus. There are no traces of nasal spines. This area shows superficial similarities to that of male gorillas, although in gorillas the prenasal area is much more extensive and a nasal spine is present. Tables 2 and 3 contain some nondental measurements of UMP 62-11, to- gether with means and ranges of small samples of male gorillas, female gorillas, and chimpanzees. Some of these measurements are rather unorthodox, because the fragmentary nature of parts of the specimen means that standard reference points are lacking. Appendix 1 contains full details of all measurements taken, together with the statistics calculated from these samples. Nasal height, listed in Table 2, is a convenient measurement of the height of the superior part of the face. In the Ugandan specimen it is 98.5 mm. Such values are found only in male gorillas, and this value is in fact above the mean (95.8 mm) for ten male lowland gorillas. The nasal index is low and in this feature the fossil resembles gorillas rather than chimpanzees. However, the high value of the index in Pan troglodytes is due to the low value of nasal height, whereas the difference between the values in male gorillas and the fossil are due to differences in nasal breadth. In the case of female gorillas differences are due both to height and breadth. (This illuminates one of the dangers of using the absolute value of an index too uncritically.) The low nasal breadth, 30.2 mm is a reflection of general upper facial narrowness in this specimen, presumably a primitive feature correlated with small anterior tooth size. INTERORBITAL AREA AND ORBITS Only the anterior parts of the interorbital region are preserved. Neither maxillofrontale nor dacryon are present, and therefore an unorthodox inter- Table 2 Nasal height Nasal breadth Index B/H x 100 Breadth across anterior lacrimal crests Thickness of lateral orbital margin Orbital height from zygomaticofrontal to inf. border Nasospinale- alveolare Breadth across infra- orbital foramina Left infraorbital for. to left anterior lacri- mal crest Table 3 Alveolar L Alveolar B at M? Alveolar B at C* Palatal D at C? -p® Palatal D at M* TERTIARY PONGIDAE OF EAST AFRICA Facial measurements of UMP 62-11, Gorilla and Pan UMP 62-11 +9855 3052 UMP 62-11 95.8 62.5 66.5 65.3 94, 0 #4,0 270 X OR x OR X OR x OR X OR X OR X OR G. ge gorilla males (n= 10) 95. 8 89. 5-106.3 37.9 33.4- 40.7 39.6 S55 a) CSias) 26.9 16.9- 34.9 15.8 135 0= 1703 28.9 23, 2— 3455 34.1 ZOO Ziel 61.9 Dseo= 69.2 39. 7 34.,5= 43.7 females (n=10) 77.7 68. 6-85. 2 32.6 ZE 0-395 7 42.2 3367—-O0.7 US)S 7 1S. 3=23). 4 11.0 65, 748 29.4 27. 1-31. 4 28.1 PANS k=) oy x5) 53.2 49, 0-59. 6 34.0 29.1-41.4 G. g. gorilla males (n=10) 104.3 98. 5-109.0 12.7 68. 2-76.2 75.2 64. 4-84.5 70.1 64. 5-73. 8 97.4 86. 3-108. 2 OF 7. Gog— 16.2 16.:9 11.0- 24.3 females (n=10) 85. 0 80. 6-89. 5 65.7 62, 1-69. 1 58.2 52. 2-64. 1 77.4 74. 4-81.4 113.4 103. 0-125. 7 10.3 8.2-14.4 15.9 11. 0-22. 4 25 P. troglodytes (n=10) 54.0 42.1-59.8 Zatz 21 0=28))6 46.6 Al 7-61. 7 20.6 135 6=2'6.3 Jodi Lao L058 ZA WG PSIG 8) 29.3 Zao ee 50.2 42.9-57.0 27.0 PAM SJE} hee) Alveolar and palatal measurements of UMP 62-11, Gorilla and Pan P. troglodytes (n=10) 67.7 63. 0-75. 9 58.0 54. 8-63. 2 56.2 50. 3-62. 4 85.8 81. 2-393. 8 103.5 95. 7-116.4 6.2 Fo = “Ss (0 12.4 10. 0-14. 8 orbital breadth measurement has been taken, utilizing terminals on the anterior lacrimal crest at the level of the most superior part of the lacrimal fossa. Table 2 includes this measurement for UMP 62-11, as well as for gorillas and chimpan- zees. The fossil yields a value of approximately 20.0 mm and falls therefore very close to the means of chimpanzees and female gorillas and within the range of male gorillas. 24 PEABODY MUSEUM BULLETIN 31 In living gorillas, particularly males, the interorbital breadth increases pos- teriorly. This is associated with the pneumatization by the frontal sinuses of the interorbital area. In UMP 62-11 there is a tiny indentation behind the posterior lacrimal crest on the left side which is probably part of the frontal sinus, suggesting that interorbital width would have increased posteriorly. Anterior and posterior lacrimal crests are preserved superiorly on the right side and delimit the superior part of the lacrimal fossa. On the left both crests remain and surround a small fossa which opens into the nasolacrimal duct. The lacrimal fossa is small compared to that of modern pongids. The maxillo- lacrimal suture is preserved in each fossa. The only other fragment of orbit is an isolated piece from the left lateral margin. The fragment is broken superiorly at the zygomaticofrontal suture and inferiorly at the inferolateral corner internally and externally on the frontal process of the zygomatic externally. Table 2 includes two measurements on this fragment: the thickness of the lateral border taken at the (presumed) midpoint between the superior and inferior orbital margins, and the vertical distance from the zygomaticofrontal suture to the horizontal plane of the most inferior part of the inferior orbital margin. The fragment resembles most closely in size and morphology a gracile female gorilla. It is in fact very similar to the homologous area of a female lowland gorilla Pr 52.0.4 in the Duckworth Labora- tory of Physical Anthropology at Cambridge, England. The lateral margin of UMP 62-11 is even more slender than in the living form. As far as can be inferred, the orbit was probably relatively high with respect to breadth as in many female gorillas. MAXILLAE AND PREMAXILLAE The maxillae and premaxillae are well preserved anteriorly. The alveolar processes are intact, although their most posterior portions, the maxillary tuber- osities, are deficient superiorly. Alveolar bone adherent to tooth roots and not invaded by sinuses is intact. Thus on either side of the piriform aperture and nasals remains of the maxillae extend superiorly above the canines and pre- molars almost as far as the inferior margin of the orbits. However, these frontal processes are confined to narrow strips, as the majority of the bone forming the lateral walls of the maxillary sinuses has broken and been lost, although a number of pieces from this lateral maxillary area remain among the unattached pieces. The anterior part of the face is dominated inferiorly by the alveolar juga of the canines which extend superiorly as prominent ridges lateral to the nasal aperture. In UMP 62-1], these juga are associated with large canine roots and so are relatively large and massive. At the level of the upper part of the nasal aperture, the juga disappear and the frontal processes of the maxillae become gently concave. The architecture of this part of the face is strongly reminiscent of that of male gorillas, although built on a somewhat smaller scale. The premaxillary area is short from nasospinale to alveolare (see Table 2) and this may be regarded as a lack of alveolar prognathism alone or as part of the general shortness of the maxillary alveolar processes. In Recent Pongidae, the alveolar processes project inferiorly to a greater extent (note the dimensions TERTIARY PONGIDAE OF EAST AFRICA DS of palatal depth in Table 3), while as part of the general increase in size of the dentition the incisors have been carried far anterior to nasospinale. Thus in modern Pongidae the great length from nasospinale to alveolare is a function of two related factors. Apparently, neither factor was important in UMP 62-l1. The slope of the nasoalveolar clivus is at an angle of 30 to 40 degrees to the occlusal plane of the cheek teeth, somewhat steeper than that of most male gorillas, and about as steep as that of females. The difference is apparently correlated with the fact that the length of the clivus is less than in male gorillas. A rather strong jugum is associated with the root of the left central incisor (the right had been lost some time before death). The area between this and the ipsilateral canine jugum is rather flattened. Similar juga are associated with the mesiobuccal roots of the anterior premolars. These juga merge superiorly with the canine juga. Immediately posterior to the premolar juga are slightly depressed areas—the canine fossae. Although the relevant parts of the face are missing, it is unlikely that the depressions were of any great extent posteriorly, for the zygomatic processes of the maxillae begin at the level of the posterior border of M!. Regions immediately below the infraorbital foramina were prob- ably predominantly flattened, although a slight groove perhaps ran inferiorly from the infraorbital foramen to the fossa. The zygomatic processes of the maxillae are broken off, but eversion of the maxillae some 10 mm above the alveolar margins indicates that the inferior margins of the zygomatic processes were situated at about the level of M?. An isolated fragment of what is probably the left zygomatic bone consists of a small fragment of the inferior border, some 15 mm long, including the zygomatic surface of the zygomaticomaxillary suture. Superiorly, 10 mm from the inferior border there is part of the zygomatic process of the maxillary sinus. The inferior margin of the fragment is roughened for the origin of the superficial laminae of the masseter. The area is not excavated as in many gorillas, nor even as greatly roughened as in many living chimpanzees (and some gorillas), suggesting that the portion of the masseter originating therefrom was relatively poorly developed. Anteromedial to this is a tuberosity, in size and morphology gorilla- rather than chimpanzee-like. Although the fragment is’small, it indicates that in all probability the part of the zygomaticomaxillary region represented was similar in build to that of a small female gorilla. The medial margins of the infraorbital foramina are bilaterally preserved as “lipping” on the lateral margins of the broken maxillary frontal processes. As a measurement of facial breadth, the distance between the mediosuperior edges of the infraorbital foramina has been taken. This is not a particularly satisfactory measurement, but is the only one which can be taken. It is impossible to measure the depth of the foramina below the inferior orbital margin, so the distance between the infraorbital foramen and the superior end of the anterior lacrimal crest has been measured on the left side. Both measurements are listed in Table 2. The distance between the infraorbital foramina is less than that for 20 gorillas and 10 chimpanzees. The specimen clearly had a rather narrow face. One gets the impression of a long, gorilla-like upper face above a dentition which is a little smaller than that of a female gorilla, except for the canines. The large canines, and the relatively massive portions of the facial buttress 26 PEABODY MUSEUM BULLETIN 31 system associated with these, suggest that UMP 62-11 is a male. The specimen probably resembled a “scaled-down,” long-snouted male gorilla, with a gracile upper face rather like those of female gorillas. The nasal floor is formed by the palatine processes of the maxillae. Just posterior to nasospinale the maxillae drop 5 mm or so to the nasal floor. The inferior nasal meatus is restricted anteriorly. The incisor fossa is almost 10 mm broad and takes up the entire width of this anterior part. The incisive canal slopes forward at a very oblique angle to the plane of the cheek teeth (and to the nasal floor). Part of the posterior lip of the incisive fossa is broken, thus increasing the anteroposterior fossa length, but even so the incisive canals were very large, much larger than is the case in the living African apes. This is probably a primitive feature. Traces of the intermaxillary crest are present along the midline of the nasal floor. The lateral walls of the fossa are present as far back as the distal borders of the posterior premolars. Posterior to this the thin septum separating the fossa and the maxillary antrum has been broken leaving only ridges inferiorly. The palatine bone is deficient and the most posterior parts of the nasal floor are therefore absent. The inferior conchal crests are present on the lateral walls at the front of the nasal fossa some 13 or 14 mm above the floor, extending approximately 20 mm posteriorly. The fragmentary remains of the nasal bones and associated parts of the maxillary frontal processes do preserve some features of interest on their poste- rior surfaces. On each side the medial half of the lower 10 mm of a single infraorbital canal is preserved. Medial to this is part of the roof of a superior lobe of the maxillary sinuses. Farther medial still to this on each side, but best preserved on the left, is the nasolacrimal canal. This is a delicate bony tube, some 4—5 mm in diameter. On the left side it is 25 mm long. The canal runs anteriorly as well as inferiorly from the lacrimal fossa towards the nasal floor at an angle of approximately 50 degrees to the cheek tooth row. At its broken inferior end the canal is 25-30 mm above the nasal floor. Just superior to the break the canal becomes part of the lateral wall of the nasal fossa. Unfortun- ately, the manner in which the nasolacrimal duct opens into the inferior meatus is unknown, but it is unlikely that it would have expanded into the dilated nasolacrimal bulb typical of Gorilla gorilla. In adult gorillas the bulb forms almost immediately inferior to the lacrimal fossa (Cave, 1961). In the chimpanzee the expansion is much less marked, but in neither does one find 25-30 mm of un- dilated nasolacrimal canal. ‘The specimen is in general less pneumatized than are living pongids of comparable age and the absence of an expanded nasolacrimal bulla may be another manifestation of this general underdevelopment. Medial to the nasolacrimal canal are low ridges corresponding to the broken middle conchae and the superior portion of the bony nasal septum. SINUSES MAXILLARY Fortunately, the surface of the nasal fossa and the lining of the sinuses have a distinctive pink color, thus enabling the extent of sinuses to be judged from the isolated identifiable fragments. TERTIARY PONGIDAE OF EAST AFRICA 27 The maxillary sinus occupies the body of the maxilla from the level of the distal border of P+ anteriorly to some 30 mm behind and 40 mm above the distal border of M3 posteriorly. The mesiodistal root of M? is bilaterally exposed in the sinus. Inferiorly, the sinus floor is depressed slightly between M1 and M?, and between M2 and M32. However, the antrum cannot be termed multilocular as it is in Recent pongids. In living apes, with the extension inferiorly of the alveolar processes, the sinus fills deep recesses between the roots of the cheek teeth and these recesses are often separated by sharp partitions. In UMP 62-11 the alveolar processes are relatively poorly developed and the sinus is not expanded. In living apes, since the lengthening alveolar processes have carried the floor of the maxillary sinus inferior to the palatine processes, the palatine processes in turn become pneumatized. This has not occurred in UMP 62-11. Superiorly, there is at least one anterior process of the sinus. The sinus extends superiorly to the lacrimal fossa. It seems therefore to have been as extensive superiorly as in living pongids. The sinus has invaded the zygomatic process at least as far as the lateral border of the orbit. Both the lateral orbital and the inferior zygomatic fragments preserve patches of the distinctive pink sinus lining on their inferior and superior aspects respectively. FRONTALETHMOIDAL A multilocular frontalethmoidal sinus was present at the level of the fronto- maxillary suture and below and posterior to this. It extended medial to the lacrimal and ethmoid bones. The bone surrounding the sinus was probably a little thicker, and the sinuses themselves less extensive, than in living African pongids. ALVEOLAR PROCESSES The alveolar processes of the maxillae are preserved intact. The external or buccal and labial portions of the processes will be considered first. The external alveolar margins diverge anteriorly. This is undoubtedly correlated with the large size of the canines. Table 3 gives some external dimensions of the alveolar region. In the sample of gorillas, bicanine breadth exceeds that across the ectomolaria (at M2) only in males. The relative breadths of the alveolar processes at the level of canines and second molars has been expressed by the ratio (Bimolar B/Bicanine B) x 100. In UMP 62-11 the value of this ratio is 94.0, indicating that bicanine breadth exceeds bimolar breadth. In male gorillas this ratio varies from 86.3 to 108.2, in females from 103.0 to 125.7. The value of this ratio is governed almost entirely by the size of the canine roots, which affects the bicanine breadth. In the mountain gorilla, Gorilla gorilla beringet, this tendency to anterior expansion in males is often even more marked. Thus a specimen of a young adult male mountain gorilla in the Duckworth Laboratory has a bicanine breadth of 83.7 mm and a bimolar breadth of 64.7 mm, giving a ratio of 77.3. The length of the alveolar region from prosthion to the plane of the posterior edges of the maxillary tuberosities had to be estimated for UMP 62-11 since the tuberosities are deficient posteriorly. The estimated length is 28 PEABODY MUSEUM BULLETIN 31 95.8 mm which exceeds that of 10 gorilla females (range 80.6—89.5) and falls short of the range for 10 male gorillas (98.5—-109.0). The ratio (Bimolar B/Alveo- lar L) x 100 has been calculated as an expression of relative alveolar length. The ratio ranges from 64.5 to 73.8 for the male gorillas and from 74.4 to 81.4 for the female gorillas. Male gorillas therefore have relatively longer and narrower alveolar regions than do females. In UMP 62-11, the value of this ratio is 65.3, indicating a long narrow muzzle. As far as proportions are concerned, the specimen is once again best interpreted as a male. PALATE A few palatal measurements are included. Palatal length cannot be taken, since staphylion is lacking. The palatal breadth between endomolaria at M? is 33.4 mm and at P* between 33 and 34 mm. Palatal depth has been taken from the plane of the alveolar margin to the midline of the palate (Table 3). At M? the depth is about 7.5 mm and between C! and P? about 4 mm. The alveolar processes project considerably less than in gorillas and less than in many chimpanzees. The intermaxillary suture is still visible, but the premaxillary-maxillary su- ture is almost entirely obliterated. On the left side, a small portion of the horizontal part of the palatine is preserved, and the maxillopalatine suture is just visible. The greater palatine foramen is present on the left side. Anteriorly the incisive canals open in a broad palatine incisive fossa. The median partition has been broken and is absent. Broad, shallow grooves run anteriorly from the fossa on either side of the midline. PTERYGOID AREAS The right and left pterygoid areas are preserved posteriorly. The medial and lateral plates have been broken off. Their basal thickness is similar to that in female gorillas, although the lateral plate lamina was probably thicker than the medial. DENTITION CENTRAL INCISORS The left central incisor is intact, although it had been broken at the alveolar border. The crown is heavily worn, the pulp cavity being exposed in an area of 5 mm by | mm. The tooth is rather procumbent, its long axis forming an angle of approximately 45 degrees to the plane of the cheek teeth, and resembles male rather than female gorillas in this respect. Just below the cervix there is a thickening of the enamel which represents the cingulum. To facilitate comparison with the living great apes, incisor crown measure- ments have been taken using the methods described by Ashton and Zuckerman (1950). ‘The dimensions taken are the labial mesiodistal and labiolingual diame- ters. ‘These are compared in Table 4 with those of adult gorillas and chimpanzees which had worn dentitions. TERTIARY PONGIDAE OF EAST AFRICA 29 Table 4 Measurements of central incisors of UMP 62-11, Gorilla and Pan UMP G. g. gorilla P, troglodytes 62-11 males females males females (n=12) (n = 11) 10.8 X 12.4 a 10.0 9.9 MDi L (n=13) (n=8) 95%CL 7.0-17.8 8.2-14.0 5.2-14.8 5,.4-14.3 Ont x 10.6 9.2 8.9 8.8 LaLi B (n=18 (n=13) 95%CL 6.0-15.2 8.5-11.0 6.3-11.2 6.2-11.4 The tooth is closest in size to those of female gorillas. Morphologically, there are no particularly distinctive characters, the tooth being similar to those of gorillas, although smaller. The root is approximately 20 mm long and appears as a rounded ridge on the nasoalveolar clivus. Just above the cervix, the root has a mesio distal diameter of 9.4 mm and a labiolingual diameter of 8.7 mm. Similar dimensions can be found in small specimens of Gorilla gorilla. The right central incisor was lost during life and the root socket filled with cancellous bone. LATERAL INCISORS Both lateral incisors are present and in both the roots are missing for a few millimeters inferior to the alveolar margin. On each root there is a large elliptical wear facet on the lateral (distal) side which in its lower part has eroded the cervical margin of the crown. The facets were formed by the lower canines and are a little unusual. In male gorillas facets are sometimes found in a similar position, but rarely are these facets as well developed as in UMP 62-11. In the male gorilla, as the lower jaw moves laterally from centric occlusion, one upper canine occludes with the ipsilateral lower canine; this acts as a stop preventing further lateral movement. The contralateral lower canine may or may not have reached the contralateral lateral incisor and so the presence or absence of wear facets is variable. The factor affecting presence or absence is the relative breadth across the outside of the lower canines compared to the distance from the internal margin of one upper canine to the distal margin of the contralateral lateral incisor. The distance across the outside of the lower canines will depend on the bicanine breadth at the cervical margin and on the relative divergence of the axes of the crowns. Thus very divergent lower canine lingual borders, as in many male gorillas (e.g. Duckworth Pr 62.0.1), may prevent formation of wear facets on the upper incisors. As the tip of the lower canine occludes with the distal border of the upper incisor, the anterior margin of the upper canine occludes with the distal and labial surface of the lower canine, forming there a characteristic wear facet. Very approximately, the distance between the midpoint of the upper lateral incisor wear facet and the midpoint of the upper canine anterior wear facet corresponds to the distance between the tip of one lower canine and the midpoint of the 30 PEABODY MUSEUM BULLETIN 31 labiodistal wear facet of the other lower canine. This distance in UMP 62-11, measured on the upper dentition, is approximately 48 mm. This gives us some guide to the dimensions of the lower dentition. The incisor crowns are moderately worn, more so on the left where the occlusal wear facet is approximately 5 mm by 2 mm. On the right the facet is about 3 mm by 1 mm. The loss during life of the right central incisor resulted in the exposure of the front teeth to differential chewing and thus to differential wear. On the left, the mesial border of the crown is vertical, while the distal border slopes distally to the cervix. The labial border is smoothly convex, the enamel being rather heaped up just below the cervix. The labial surface of the tooth is set at a less acute angle to the plane of the cheek teeth than is the root, the values being approximately 60 and 40 degrees respectively. The lingual border (which strictly speaking also includes the distal aspect) is coarsely wrin- kled as in Gorilla gorilla and is almost parallel to the plane of the cheek teeth. The lingual surface is thickened at the cervix and from there an enamel “pillar” extends to the occlusal surface. The pillar gets thinner occlusally. Measure- ments are given in Table 5. Table 5 Measurements of lateral incisors of UMP 62-11, Gorilla and Pan UMP G. g. gorilla P. troglodytes 62-11 males females males females L R 66 We7 ok 10.6 9.0 8.7 8.8 MDi LL (n=11) (n=13) (n=5) (n=5) 95%CL Sa7=12.0) 4.5-13.5 6.9 -L055 7.0-10.6 10,2 DOs Ty ye Oley 9.0 8.3 7.9 LaLi Bt (n=11) (n=13) — (n=18) (n=13) 95%CL 7.5-13.9 5.6-12.4 6.1-10.5 5.7-10.1 Breadth across 3901 > x 41.7 So t/ Bio 3652 jateral alveolar (n=19) (n=20) (n=14) (n=12) margins (I*B) 95%CL 38.6-46.7 32.7-42.6 33.8-41.1 32,9-39.5 1comparative data from Ashton and Zuckerman (1950). Metrically, the lateral incisors are relatively narrow mesiodistally but broad labiolingually. In the second character the teeth resemble those of Gorilla gorilla, particularly males. The relatively small mesiodistal diameter is but one reflection of the fact that a number of this specimen’s teeth have relatively short mesio- distal diameters when compared to those of the living African apes. Table 5 also includes measurements taken between the external borders of the lateral incisors at the alveolar margin. Measurements are taken at the incisor alveolar margin to facilitate comparisons between living and fossil ma- terial in which the incisor crowns are often lost. The fossil falls between the means for male and female gorillas. Considering the overall small size of UMP 62-11 when compared to the male gorillas we may conclude that, relative to overall size, the fossil has a rather broad incisor region. This suggests again that the specimen is a male. TERTIARY PONGIDAE OF EAST AFRICA om Small diastemata are present on each side. At the level of the alveolar margin the left diastema is 5.2 mm long. The measurement cannot be taken on the right. CANINES The left canine is broken approximately at the level of the cervix although part of the crown is preserved in the distal part of the tooth. The mesial part of the tooth is broken above the cervical margin. Distally, the portion of the root between the cervix and interdental septum is missing. A slight cingulum is present distolingually. The inferior distal 15 mm of the tip of the labial section of the crown was recovered. The root of the right canine is broken at the alveolar margin. However, part of the crown is preserved —the proximal two-thirds—and this reaches as far proximally as the cervical margin on the mesial half of the lingual side of the crown. Since the middle section of the canine crown, missing on the left, is present on the right, and the tip, although absent on the right, is present on the left, both crowns can be restored with some accuracy. The reconstructed canines are typically pongid, resembling those of gorillas. A well-marked mesial groove runs almost to the tip of the crown. There are smooth areas of wear on either side of the groove and running parallel to it. Measurements are given in Table 6. In all three dimensions UMP 62-11 falls within the 95 per cent confidence limits for male lowland gorillas and outside the limits for females. For labiolingual breadth and labial height, the fossil falls very close to the means for males; in mesiodistal diameter the fossil is relatively small. The difference in proportion is reflected in the length-breadth index, the Uganda specimen being relatively short mesiodistally. We have already noted that the mesiodistal length of the lateral incisors is also relatively short. The roots of the canines are approximately 40 mm long and are relatively massive. Associated with the large roots are the strong canine juga described above. Morphological and metrical features of the crown and roots indicate that UMP 62-11 represents a male. Table 6 ro measurements of UMP 62-11 and Gorilla UMP G. g. gorilla_ 62-11 males females. *18.6 x hi 14.5 Max L (n=20) (n=20) 95%CL 18, 3-24. 1 12./0-16.2 *15.7 X 16.0 1.2 Trans B (n=20) (n=20) 95%CL 13: 7-18. 2 9,5-12.8 *27.6 x 29.1 Tot La H (n=7) (n=11) 95%CL 19. 8-38. 4 14,4-19.8 118.4 x 133n2 130.1 Index L/B (n=20) (n=20) x 100 OR 115. 4-151.3 112. 5-145. 4 32 PEABODY MUSEUM BULLETIN 31 FIRST PREMOLARS The protocone is broken away in the left P®. The crown is broad bucco- lingually and the roots well-developed. There are two buccal roots and a single lingual root. A premolar jugum is associated with the mesial buccal root and extends superiorly 18-20 mm above the cervical border. The mesiobuccal angle slopes sharply inferiorly and distally and bears a wear facet from the distal face of the main cusp of P3. ‘This same tooth has left a wear facet in the middle of the mesial border of P?. A wear facet for the mesial face of P; can just be seen on the distal aspect of the cervical region of the left upper canine. In normal centric occlusion, the paracone-protocone crest of P? generally lies directly superior to the line of contact of P; and Py. The distance from the paracone-protocone crest of P® to the distal wear facet on C1 is between 7 and 8 mm. This is approximately equivalent to the distance between the tip of the main cusp of P; and its distal border. The tip of this cusp generally lies about half way along the mesiodistal (projected) length of P;. We may conclude there- fore that the projected length of P3 parallel to the long axis of the cheek teeth would have been between 14 mm and 18 mm and its maximum length perhaps between 16 mm and 20 mm. A groove runs superiorly, between the two contact facets described, along the middle of the mesial face of P*. There is a slight cingulum on the mesial half of the labial border running obliquely inferiorly and distally. The paracone is moderately worn, a small area of dentine being exposed. The distolingual surface of the paracone shows wear facets from articulation with the protoconid of P4. The crown of the right P* is intact, the paracone is larger and higher than the protocone; paracone height above cervix, *11.5 mm, and protocone height above cervix, *8.0 mm. Both cusps show moderate wear, dentine being exposed on the paracone tip. A high crest runs between the cusps and is itself traversed by a mesiodistal sulcus running from the small pit which represents the anterior fovea to the talon basin. A small groove, almost obliterated by wear, runs across the tooth parallel to the distal border. The distal border is convex and has a wide (*4.0 mm long) contact facet with P*. The buccal and lingual surfaces of the crown slope relatively sharply towards the occlusal surface. The buccal surface in particular is less vertically oriented than the normal condition in gorillas and chimpanzees. Measurements are given in Table 7. Metrically UMP 62-11 is closest to Gorilla gorilla; in maximum length and buccolingual breadth the fossil is just a little short of the means for female gorillas. However, the mesiodistal length of the crown measured in the midline of the tooth is 2 mm shorter than the female means and is outside the 95 per cent confidence interval for females. The shape of the crown is more triangular in the fossil than in gorillas (and chimpanzees), the mesiobuccal corner of the crown projecting some way mesial to the rest of the mesial border, particularly at the level of the cervix. If the fossil is to be regarded as ancestral to living gorillas, then the following changes have occurred in P®* structure since the Miocene. The tooth has become larger, the tips of the cusps farther apart, and the buccal and lingual surfaces more vertical. The shape of the paracone in TERTIARY PONGIDAE OF EAST AFRICA 358) Table 7 P* measurements of UMP 62-11, Gorilla and Pan UMP G. g. gorilla P. troglodytes 62-11 males females males females L R (n=20) (n=20) 830) 852 x, 10.9 1022 Fas 7.0 MDi L (n=14) (n=12) 95%CL 9.4-12.4 9.0-11.4 6.2- 8.3 6.2- 7.9 10.6 10.7 x 12.3 TI10 8.1 7.9 Max L (n=14) (n=12) 95%CL 10.8-13.8 9.5-12.4 6.8- 9.3 7.0- 8.9 - 14.6 X 15.6 14.8 10.5 10.5 BuLi B (n=13) (n=11) 95%CL 13, 4-17. 8 12. 8-16. 8 9.3=L1516 S52-11.9 particular has changed so that the mesial border no longer slopes backwards but has become more nearly vertical. The cusps have become higher and more pointed, the crests sharper and more clearly defined. SECOND PREMOLARS Each P# is a very broad tooth with strongly sloping buccal surfaces like P%. The paracone is somewhat higher than the protocone. Roots are a little smaller than in P%, the buccal root being approximately 12 mm long. There is a small cingulum at the mesiolingual corner and the rest of the lingual surface is etched with vertical grooves. On the buccal surface there are two vertical grooves, one on the mesial moiety, the other on the distal. Such features are seen occasion- ally in gorillas. The paracone and protocone are united by a low crest. There is a small anterior fovea which is displaced towards the buccal side of the crown. The talon is similar to that of P? and bulges distally, there being a rather long area of contact with M! (5 mm) which has worn a concave facet on the mesial border of Mt}. The crown is moderately worn with dentine exposed on the protocone. Mesial to the paracone-protocone crest there are wear facets for the distal surfaces of the protoconid and metaconid of P,, and distal to the crest wear facets for the mesial cusps of M,. Measurements are given in Table 8. In breadth UMP 62-11 falls within the gorilla range, in length between gorillas and chimpanzees. Like the lateral incisor, canine, and first premolar, Table 8 P*measurements of UMP 62-11, Gorilla and Pan UMP G. g. gorilla P. troglodytes _ 62-11 males females males females L R 7.9 8.0 x alls 10.6 7.0 720 MDi L (n=20) (n=20) (n=14) (n=12) 95%CL 9.9-12.2 9.1-12.0 6.4- 7.7 6.4- 7.6 14.8, 1457 me 15.3 14.4 10.2 10.2 Buli B (n=20) (n=19) (n=13) (n=11) 95%CL 13. 7-16.8 12.5-16.4 8. 4-10. 9 9.6-10.8 34 PEABODY MUSEUM BULLETIN 31 this tooth is relatively short mesiodistally. The cusps are lower and the crests less sharp than is the case in living gorillas. The antemolar teeth as a whole are rather broad when compared to the molars. ‘Thus the range of the ratio (BLi B M!/BLi B P*) x 100 in gorillas is as follows: for 20 males, the mean is 97.5 and the range is 89.4—108.3; for 19 females the mean is 97.8 and the range 92.7—110.7. The value of this ratio for the left P* and M1! of UMP 62-11 is 116.5, the P* breadth being relatively greater than in any of 39 gorillas. An explanation is difficult to find. Perhaps the broadening of P* is associated with the breadth of the anterior dentition and alveolar region in general. The great cervical breadth of P* and the closeness of the tips of the paracone and protocone result in the markedly sloping buccal surface, such a slope being absent in living gorillas where the surface is more nearly vertical. Once again, if the specimen is regarded as a proto-gorilla, phylogenetic trends have been towards increase in cusp size and height; changes in shape are also involved, (e.g., slope of buccal and lingual surfaces becoming steeper correlated with a relative widening of the intercuspal occlusal area). FIRST MOLARS The left tooth is rhomboidal, with a marked distal convex bulge. Paracone, metacone, and hypocone are subequal; the protocone is the largest cusp. The fine features of occlusal morphology have been largely obliterated by wear. Small patches of dentine are exposed on the paracone and metacone. A some- what larger area on the hypocone is 2.5 mm in diameter, and the largest area occurs on the protocone, where it is 4.5 mm in diameter. The distal bulge has worn a wear facet on the mesial face of M?, the area of contact being 4.5 mm long. The groove between paracone and metacone is still patent. ‘There is a marked buccal cingulum between the two buccal cusps and a slight cingulum at the buccodistal corner. The cingulum runs onto the distal surface. In lateral view the buccal cingulum gives the appearance of a row of inferiorly projecting, isolated spurs. The lingual cingulum has been largely obliterated by wear ex- cept for a fine groove running parallel to the mesial margin of the protocone and 1 mm internal to it. This cingulum probably ran only as far as the mesial margin of the hypocone and not around the hypocone. All the roots are present. The lingual root diverges lingually at an angle of 30 degrees to the buccal roots and has a vertical groove on its lingual aspect. The lingual root is 9.4 mm thick mesiodistally just above the cervix and has a vertical length of about 15 mm. The buccal roots are rather more delicate and perhaps not much more than 12-13 mm in vertical length. The right M? is essentially similar to the left, although the area of dentine exposed on the hypocone is somewhat larger on the right. Measurements are given in Table 9. In size, the first molars of UMP 62-11 fall between those of chimpanzees and gorillas. Length/breadth indices are a little lower than those of gorillas and rather higher than those in chimpanzees. Like the other teeth, this tooth is relatively small mesiodistally when compared to those of gorillas. Few details of the occlusal surface can be determined, although the unworn crowns were probably similar to UMP 66-41 from Napak IV and UMP 62-07 TERTIARY PONGIDAE OF EAST AFRICA 35 Table 9 M* measurements of UMP 62-11 , Gorilla and Pan UMP G. g. gorilla P. troglodytes 62-11 males females males females L R (n=20) (n=20) cubits 7 cod lale (5) x 14.8 14.1 9.9 9.8 MDi L (n=14) (n=12) 95%CL 13.0- 16.6 1Z2o- 16.6 ~~ 8..38-1150 5 la) 12.7 12.6 x 15.6 14.8 2 TT 2 (n=12) (n=11) BUEN E 95%CR, .13.7- 17.5 13.2= 16.6 10. 1-12. 2 9,9-12.5 ap Q2F 159221 x 94,7 95.2 89.2 88.1 Eee (n=12) (n=11) as OR 84.0-105.0 85.8-105.1 82.6-94.5 79.5-95.5 from Napak V. In all three lingual cingula are present and the unworn molars show some resemblances to those of Gorilla gorilla. These resemblances are discussed in the following chapter. The degree of development of the roots of M! in UMP 62-11 is midway between that seen in gorillas and that of chimpanzees and appears to be cor- related, at least to some extent, with tooth crown size. Measurements of UMP 62-11, a female gorilla, and a male chimpanzee first molars, selected at random, are listed in Table 10. SECOND MOLARS As in the first molar the protocone is the largest cusp, the paracone, meta- cone, and hypocone being smaller and subequal in area. The lingual and buccal sides converge very slightly from front to back. There is a distal bulge, as in P’, P4, and M}, and since the lingual side is longer than the buccal, the bulge appears to be more prominent in the lingual half of the distal border. The crown shows some wear, not as much as M!, but enough to expose dentine in a small area on the protocone. ‘The crests between paracone and protocone, and metacone and hypocone, have been worn so that the occlusal surface now shows very little relief. A trace of the anterior fovea is present, closed lingually by the protoconule. Table 10 Crown and root dimensions in UMP 62-11, Gorilla and Pan Gorilla UMP Chimpanzee female 62-11 female DPr52.0.4 ID) Beaye}q (58) MDi L *13.9 call ba Waser / *10.2 BuLi B 1505 WAS *11,2 Crown area *201.55 *148, 59 *114, 24 Vertical L lingual root *18.0 *15.0 *13.5 MDi T lingual root thea 9.3 USU Vertical L buccal roots (max) *15.0 *13.0 *12.0 36 PEABODY MUSEUM BULLETIN 31 The buccal cingulum is similar to that of M!, appearing as a series of spurs. It is most prominent between the two main cusps and on the buccodistal corner. A sulcus runs parallel to the distal border in its buccal half and separates the metacone-hypocone crest from the flattened most distal part of the crown. The lingual cingulum runs from the midpoint of the mesial surface of the protocone, around the mesiolingual corner, along the lingual border, and around the hypocone onto the distal border. The cingulum is beaded and the lingual slopes of protocone and hypocone are etched with vertical grooves which merge into the beading. The roots are less widely splayed than is the case for M!. The buccal roots have a maximum vertical length of 13.0 mm and the lingual root a length of about 16-17 mm. The latter has a mesiodistal thickness of 11.5 mm just above the cervix. The mesiolingual corner of the right M? is missing. Apart from this, the features of crown and roots are the same as for the left side. Measurements for M2 are given in Table 11. UMP 62-11 falls within the 95 per cent confidence Table 11 M* measurements of UMP 62-11, Gorilla and Pan UMP G. g. gorilla P. troglodytes 62-11 males females males females ip R (n=20) (h=20) =13.7 ¥13,1 X 16.0 14,8 10. 2 10.0 MDi L (n=14) _ (n=12) 95%CL I329— LSat 12.8- 16.8 8,8-16.8 857-1153 1450) 1453 x 16.7 15.6 als 7/ ite! BuLi B (n=13) (n=11) 95%CL 14.7- 18.5 1357- 17.5 10.6-12.8 10. 0-12.9 ives 33.5 92.9 X 96.2 94.7 87.4 86.7 (n=13) (n=11) 1/8 x 100 OR 91.1-106.2 87.4-104.5 80.2-94.0 77.4-95.1 limits for female gorillas and is a relatively small tooth when compared to those of male gorillas. The cusps are relatively lower and their sides less steeply sloping than in gorillas. In general, though, the Uganda fossil is reminiscent of gorillas in molar morphology. THIRD MOLARS The crowns are less worn than in the other molars and the trigon crests are still well marked. The protocone is the largest cusp. Buccal and lingual cingula are well developed as in M?, the lingual cingulum being especially prominent around the protocone, thus giving the mesial moiety of the tooth a greater breadth than the distal. The contact facet with M? is a little over 3 mm long. Measurements are given in Table 12. Although relatively broader than M% of female gorillas, the lengths and breadths of M® fall within the 95 per cent confidence limits. As a general feature of this dentition, mesiodistal lengths are reduced when compared to gorillas, particularly male gorillas. The increased TERTIARY PONGIDAE OF EAST AFRICA 37 Table 12 M* measurements of UMP 62-11 Gorilla and Pan UMP G. g. gorilla _P, troglodytes 62-11 males females males females ib R (n=20) (n=20) (n=11) M28 ela 7, x 15.0 1357. 9.2 9.0 MDi L (n=10) 95%CL TS0— 175 2 1is6= 15.8) = 79-1005 Tet LOS 14.4 14.3 x 15.7 14.8 11.0 10.9 BuLi B (n=9) 95%CL 13.,.6— 17.8 Wag aes Baelo 5 Slee 74 88.9 88.8 Oe 95.8 92.7 84.0 83.0 Index (n=9) 1/8 x 100 OR 87.8-111.5 82.1-103.8 76.2-97.7 75.0-89.6 lengths in male gorillas are probably associated with a general increase in head length, as pointed out by Gregory (1916, p. 276). THE Uprer DENTITION AS A WHOLE Table 13 sets out a number of measurements for the upper dentition as a whole. The length of the cheek tooth row—from P? to M%—is 54.4 mm. This falls towards the lower part of the 95 per cent confidence interval for female gorillas. This measurement reflects the relatively short mesiodistal lengths of the cheek teeth when compared to those of gorillas. The tooth row is a little compressed—the combined mesiodistal lengths of cheek teeth are 56.1 mm, 1.7 mm greater than P°—M® length. The crowding is best described as moderate when compared to the condition in the living great apes (see Appendix 1). The other measurements listed in Table 13 are taken across the tooth rows and are maximum external measurements. The narrowing from canine, to second premolar, to third molar is a reflection of the relatively great breadth of the canine region. The proportions are similar to those seen in male gorillas (see Appendix 1). (In overall size the specimen is closest to female gorillas.) The Table 13 Measurements on the upper dentition of UMP 62-11, Gorilla and Pan UMP G. ge gorilla _P, troglodytes 62-11 males females males females (n=20) (n=20) (n=14) (n=12) p? -v? 1 54.4 xX 67.6 61.6 43.6 43.6 95%CL 6103=73.9 52.0-7152 39. 8-47. 4 40. 1-47.1 C2 B *69,1 X 75.0 58.2 58.3 54.6 95%CL 65.7-84.4 45.0-71.4 54, 1-62. 5 48, 7-60. 6 p+ 3 64.8 a 713 62.0 57.6 57.5 95%CL 63.8-78.7 52.5-71.6 52. 5-62. 8 52. 3-62.8 mM? B 60.5 x 71.4 63.9 55.7 56.8 95%CL 64.3-78.4 56.8-71.1 49.9-61.5 50. 8-62. 9 38 PEABODY MUSEUM BULLETIN 31 tooth row is therefore convergent posteriorly, yet there are clearly a number of points which need to be made about this statement. ‘The measurements listed are external breadths. Because of the relatively great breadth of P4, the breadth at P# is much greater than the breadth at M®%, or indeed at M!. Yet another estimate of divergence could be the minimum internal breadths between homol- ogous teeth, measured at the cervix. Thus for UMP 62-11 the values are: be- tween C!, *38.5 mm, between P‘4, 35.7 mm, and between M3, 32.3 mm. The internal narrowing between C! and M® is 6.2 mm, in contrast to an ex- ternal narrowing of 8.6 mm, and is a reflection of the fact that C! and P* are broader teeth than M?. However, this still does not satisfactorily answer our problem. Since the func- tion of the cusps and basins of the cheek teeth are occlusion, the best index of convergence—the best way of comparing the relevant parts of the upper tooth row and the relevant parts of the lower—is to utilize distances between cusps and basins. Accordingly, the distances between the tips of the protocones of the homologous pairs of cheek teeth have been measured. When we come to discuss the type of lower dentition with which UMP 62-11 would have occluded, we shall then know the absolute (and relative) positions of the central basins of P,-M;. The interprotocone breadths for UMP 62-11 have to be estimated be- cause of occlusal wear. They are : at P4, *41.8 mm, at M!, **41.2 mm) at M2? *40.0 mm, and at M?, *40.4 mm. Values for hypocones show a similar trend. Thus there is a gradual convergence from front to back, although M® is ro- tated a little so that the protocone has swung distobuccally. When we come to discuss the mandible of Dryopithecus (Proconsul) major from Songhor (NMK 190,1), we shall be able to examine the breadths between the basins of the lower teeth with which the protocones occlude, and thus to compare these mandibular and maxillary tooth rows. It seems that this is the most satisfactory functional measure of tooth row convergence and divergence. DISCUSSION OF THE MAXILLA UMP 62-11 was adult; the third molars had been in occlusion, but not for very long. The specimen is fairly clearly a male, at least unless this Miocene species exhibited no (appreciable) sexual dimorphism. This is unlikely, for even in those hominoid species which exhibit little size dimorphism, the canines and associated structures are larger in males than in females. The only clear exceptions are species of Hylobates and Symphalangus. The canines are as large as those of male gorillas and clearly do not resemble the canines of female gorillas or chimpanzees. Associated with these large canines is the broad anterior palatal and alveolar region. The shape of the alveolar region and palate and the morphology of the paranasal areas are typical of male gorillas. The height of the face from nasion to nasospinale is also absolutely large and falls within the range for male gorillas. However, although the pro- portions are different, in overall size UMP 62-11 is closest to the female gorilla. The facial breadth measurements show that the specimen was relatively narrow faced, even considering the small size of the fossil. In fact, indications are that the fossil was narrower faced and of more delicate upper facial build even than many chimpanzees. TERTIARY PONGIDAE OF EAST AFRICA 39 Thus we have an individual with a high, rather narrow face, a face which is at least as prognathous as those of female gorillas. From what is known of the lateral orbital and malar regions, and from what can be deduced about the masticatory musculature from the size of the cheek teeth, the fossil had an upper face similar again to that of a female gorilla; a supraorbital torus was probably present. Since we can make but a poor estimate of the size of the masticatory muscles, and since we have no idea at all as to brain size, it is difficult to say whether or not the specimen possessed a sagittal crest. Crests occur in gorillas only when the dentition is proportionately very large; it is possible that UMP 62-11 had a Sagittal crest. Incisor tooth wear is perhaps slightly heavy when compared to premolar and molar wear, but it should be remembered that the single central incisor bore the brunt of the anterior chewing stresses. The postcanine dentition is relatively small compared with that of gorillas, most of the dimensions falling within the range of female gorillas. The mesiodistal lengths tend to be relatively smaller than buccolingual breadths and length/breadth indices are accordingly low. Thus, the mesiodistal lengths of P*, P#, and M? fall below the lower 95 per cent confidence limit for female gorillas. M! is the only tooth in which both di- mensions fall outside at least the female confidence intervals. _ The morphology of the dentition is gorilla- rather than chimpanzee-like. This applies particularly to the anterior dentition. The cheek teeth show some similarities, but the cusps are not as high, conical, steep-sided, and sharp as in G. gorilla. However, it is probable that UMP 62-11 is sampled from a species which is ancestral to, or close to the ancestry of, G. gorilla. The living gorilla is predominantly a shoot- and leaf-eater rather than frugivorous like the chim- panzee (Schaller, 1963, 1965). This is at least a partial explanation for its lophodont cheek teeth with high, steep-sided cusps, and sharp intercuspal crests. UMP 62-11, together with other material from Napak, discussed below, show that the probable Miocene forerunners of G. gorilla also had somewhat lopho- dont teeth, although this trait was not as well developed as in the living form. Two further points remain to be discussed: the depth of the palate and the degree of pneumatization of the maxilla. It is perhaps a little misleading functionally to speak of palatal depth. What is really meant is alveolar height—the projection inferiorly of the alveolar tooth-bearing processes. The height of the process is correlated first with the size of the tooth roots; as roots get longer and more robust, the alveolar processes project more, and the palate gets deeper. However, this is not the whole story. Considering M1, we have noted that the crown- and root-size of UMP 62-11 are intermediate between those of a female gorilla and a male chimpanzee. The palatal depth of the female gorilla (Duckworth Pr 52.0.4) at M1 is about 20 mm, and of the male chimpanzee (Duckworth Pr 53.0.9) about 10 mm. However, palatal depth in the fossil is only some 6 mm. Also, the vertical lengths of the lingual roots of premolars and molars are fairly constant in UMP 62-11, although the palate deepens somewhat from front to back. Palatal depth be- tween C! and P? is about 4 mm and at M? about 7.5 mm. Clearly, some factor other than root size is involved. Davis (1964) has recently discussed the mechanics of chewing in the case of 40 PEABODY MUSEUM BULLETIN 31 the herbivorous carnivore, Ailuropoda, the Giant Panda. Figure 2 (a), which is adapted from Davis, represents the situation in which the point of mandib- ular articulation is at the level of the plane of the cheek teeth. Such a system is seen among the Carnivora with the clear exception of Ailuropoda. It is also found in the Prosimii. A point p,; on the lower dentition travels along the arc pi! — pi during centric occlusion. A tangent to the arc drawn at py, is per- pendicular to the tooth row—this is also the case for p. and for any other point along the tooth row. During occlusion there is little or no mesiodistal movement. Figure 2 (b) represents a mandible with the point of articulation raised above the tooth row. The tangent at p, to the arc py — p,! crosses the tooth row obliquely and the obliquity of the tangent increases from front to back, from B to C. Thus, from front to back there is an increasing component of horizontal, FIGURE 2. Representation of occlusal relationships when the point of mandibular articulation is (a) at the level of, and (b) above, the plane of the tooth row (from Davis, |964; 71, fig-32). mesiodistal, movement as the teeth occlude. Even without anteroposterior or rotatory movements at the temporomandibular joint (and these would certainly have occurred), any food material placed between the upper and lower teeth will be subject to horizontal as well as vertical chewing movements. “In Ailurop- oda... an anteroposterior grinding action is achieved by elevating the articula- tion” (Davis, 1964, p. 69). Similarly in the herbivorous great apes, the distance between the mandibular articulation and the plane of the cheek teeth is relatively great. There are two ways of increasing this distance: by raising the plane of the temporomandibular joint, and by lowering the tooth row. The tooth row is lowered by increasing the inferior projection of the alveolar processes, carrying the cheek teeth further inferiorly, which results in a deepening of the palate. TERTIARY PONGIDAE OF EAST AFRICA 41 If this line of argument is correct, Miocene forms show evidence of being less well adapted than modern gorillas to chewing tough vegetable material. As we have seen, the morphology of the cheek teeth and of the roots also suggest this. Davis (1964, p. 69) has also pointed out that increasing the depth of the skull increases the volume of the temporal fossa thus providing greater space for masticatory muscles, improves the muscular efficiency for producing pressure at the occlusal surfaces, and increases the magnitude of the vertical stresses that the skull can withstand. As was noted above, the presence of marked wear facets on the upper lateral incisors suggests a less tight interlocking of the upper and lower canines than in G. gorilla. The “anchoring” of the anterior dentition in the living species may be associated with an increase in the grinding efficiency of the cheek teeth. (The development of overbite in post-Neolithic human populations is said to serve a similar function.) As the alveolar processes lengthen and carry the dentition inferior to the plane of the palate, the palatal processes can become pneumatized by extensions of the maxillary sinuses. As the bony parts of the face become more extensive, pneumatization will spread, too, in order to lighten the face. The maxillary sinuses of UMP 62-11 are probably as extensive superiorly (as far as the orbits) and laterally (into the zygomatic processes) as in living gorillas. However, in- feriorly—into the palate and between the tooth roots—the sinuses are not as extensive in the fossil. Neither do they extend as far anteriorly, for in UMP 62-11 they stop at the level of the distal border of P*. In living gorillas they spread much further forward. But this lack of pneumatization is no doubt also correlated with the brevity of the alveolar processes. In conclusion, the specimen represents, with a high degree of probability, a male individual, which may well be sampled from a species ancestral to, or close to the ancestry of, the living gorilla. However, the Miocene form is smaller than its putative descendants, and less well adapted to a diet consisting almost exclusively of tough vegetable material. UMP 62-10 and UMP 66-01 (Fic. 20) The specimen consists of two fragments of the horizontal ramus of a large an- thropoid. The portion of the right ramus, UMP 62-10, has been fully described by Allbrook and Bishop (1963) and figured in Bishop (1963b). This frag- ment contains the roots and extensively worn crown of M,, the roots of M., and the mesial root of M3. The fragment from the left side, UMP 66-01, added since Allbrook and Bishop’s description, consists of several adjacent pieces of the ramus. These pieces have never been fully described before. Opposite M, of UMP 62-10 the internal surface of the mandible slopes lingually at an angle of about 60 degrees to the occlusal plane. More distally, opposite M, and Ms, the internal border falls away more sharply. Externally, the buccinator crest is present opposite the mesial root of Mz; and the distal root of Mz where it begins to merge into the horizontal ramus as the superior lateral torus. 42 PEABODY MUSEUM BULLETIN 31 No attempt has been made to estimate the crown size of M, and M3; these teeth are present in more complete form in the left fragment, UMP 66-01. Part of the crown of right M, has been preserved at the distobuccal corner. Traces of a large wear facet are preserved on the hypoconid. Crown dimensions of M, have been estimated by measuring root dimensions at the cervical border and com- paring these with the root dimensions of intact crowns of other molars from East African Miocene deposits. They are: MDi L, **11.8 mm, Tri B, *10.8 mm, and Tal B, *11.0 mm. One piece of UMP 66—01* comes from immediately to the right of the sym- physis and consists of the most anterior 20 mm of the planum alveolare. An- teriorly the most distal 15 mm or so of the roots of the incisors is preserved. These roots are very narrow mesiodistally and more elongate labiolingually. It is im- possible to estimate dimensions since the roots are broken some distance below the alveolar margin and obliquely to their longitudinal axes. The medial border of the fragments forms the right side of the symphyseal joint. An approximate estimate of the bilateral distance between the distal borders of lateral incisors is not more than 23 mm. Posteriorly on the distal margin of the fragment, a small portion of the canine alveolus is preserved. Although there is some difh- culty in orienting the fragment accurately, it is probable that the anterior 20 mm of the planum alveolare sloped gently posteriorly and inferiorly at an angle of approximately 30 degrees to the occlusal plane. The remaining piece of the left side of this mandible is reconstructed from several fragments, also from the UMP 66-01 lot. Crowns or roots of the tooth row from canine to last molar are present, together with the adjacent parts of the horizontal ramus and two small pieces of the ascending ramus. One piece of the ascending ramus represents what is probably the most supe- rior portion of the anterior border. The crista endocoronoidea and the anterior portion of the crista endocondyloidea are present on the internal surface and these circumscribe the inferior corner of a planum triangulare. Where the cristae join, the ramus is 11 mm thick. Laterally, the horizontal ramus runs from canine to last molar. The buc- cinator crest is preserved opposite M3. The ramus is 24.5 mm thick at the level of the mesial moiety of M3. Anteriorly the crest merges into the superior lateral torus which finally disappears at the level of M,. Between M, and the canine and premolar juga the lateral face of the mandible is slightly concave. The root of the canine and the mesial root of P; have associated with them prominent alveolar juga which fuse inferiorly. At the level of the canine the broken mandi- ble is approximately 30 mm deep. Parts of the mental canal are exposed in vari- ous places on the inferior portion of the fragment. The position of the mental foramen is indicated by the presence of lipping anteriorly and by the most distal part of the canal posteriorly. The foramen was probably single and is situated below the midpoint of P; approximately 29 mm below the cervical border. Internally, the ramus is preserved between the distal border of the canine and the back of the trigonum postmolare. Posteriorly, the most inferior 13 mm of the crista pharyngea merges into the superior alveolar margin. The most superior 15 mm or so of this margin is moderately well preserved. Opposite M3 *T should like to thank Prof. P. V. Tobias for help with the interpretation of this fragment. TERTIARY PONGIDAE OF EAST AFRICA 43 the medial border falls abruptly away from the tooth, while opposite M, the slope is approximately 60 degrees to the occlusal plane. The slope becomes shallower anteriorly. Opposite P; the most superior 25 mm of the margin is preserved and the morphology indicates that anteriorly a long and gently slop- ing planum alveolare was present, sloping backwards to the superior transverse torus. The most distal 24 mm of the canine root is preserved. The root is massive and is responsible for the lateral convexity at this part of the mandibular ramus. Length and breadth measurements of the crown would have been approxi- mately 16 mm and 12 mm. The root curves medially at an angle of between 70 and 80 degrees to the occlusal plane. This angle increases proximally, suggesting that the longitudinal axis of the crown would have been at 80 or more degrees to that plane. The mesial root and the alveolus for the distal root of Ps are pre- served. The most mesial point on the root is level with the most distal part of the canine root. The tooth is massive and large, the long axis being set at about 40 degrees to the axis of the other cheek teeth. Length and breadth of the crown have been estimated at about 16 mm and 9.5 mm. From the roots of P4 very approximate estimates of crown length and breadth can be made. These are 9.5 mm and 10.0 mm. No estimate was possible for M,; dimensions of M, were estimated from the opposite side of the jaw (UMP 62-10). The crown of My, has been broken off inferiorly and distally from the mesial border of the occlusal surface to the distal cervical margin. Fragments of weathered enamel remain on the mesial border, on the buccal border of the protoconid, and on the lingual border of the metaconid. Dentine remains exposed on the protoconid wear facet. This facet was some 3 mm in diameter, a little smaller than the wear facet on right M,. Since the mesial border of both M, and Ms are present, the length of M, can be obtained. Buccolingual crown diameters have been estimated from root dimensions and are therefore less reliable. The measurements are: MDi L, =e Oimm, Grub, **12:2 mm,and. Val B;4*12.0 mm. M; is the best preserved tooth crown. Enamel is absent from all but the most mesial and distal parts of the lingual border, and from the protoconid. An area of dentine barely 2 mm in diameter is preserved on the protoconid. Wear facets of similar size are also present on hypoconid and hypoconulid. The tooth nar- rows distally and is triangular in occlusal view. The dimensions are: MDi L, 15.0 mm, Tri B, 12.5 mm, and Tal B, *10.4 mm. The total length of the cheek tooth row is approximately 63 mm. UMP 67-28 Remains of several large chimpanzee-sized vertebrae from Moroto II have been described and figured by Walker and Rose (1968). They most probably come from the same specimen as the maxillary and mandibular specimens UMP 62-11, UMP 62-10, and UMP 66-01. Particularly well-preserved is a middle lumbar vertebra, UMP 67-28. Morphologically the vertebra is like that of living homi- noids rather than cercopithecoids. Walker and Rose note that the vertebra re- sembles those of Pan and Gorilla, with a few hominid features. Resemblances to Pan and Homo may well be primitive characters, or those associated with body 44 PEABODY MUSEUM BULLETIN 31 size, while those to Gorilla could indicate that this Miocene pongid is ancestral to the living ape. Walker and Rose (1968, p. 981) conclude that there are “quite eclectic resemblances to modern hominoid vertebrae to be seen in the Moroto vertebra, but none of them would necessarily preclude Dryopithecus (Proconsul) major from being a direct ancestor of the genus Gorilla.” DISCUSSION OF THE MOROTO II PONGID UMP 62-10 and 66-01 with a high degree of probability represent the same individual. The relationship between the maxilla (UMP 62-11) and these man- dibular parts remains to be discussed. Both represent a species of large Miocene pongid. Although little of tooth crown morphology (particularly mandibular) can be seen, from what is known of other Miocene pongid species and of living Pan troglodytes and Gorilla gorilla, the two sets of material can plausibly be referred to the same species. In 1963, Allbrook and Bishop assigned UMP 62-10 tentatively to Proconsul major [Dryopithecus (Proconsul) major according to Simons and myself]. The presumed morphology of the horizontal and ascending rami would have resembled broadly those of moderately robust female gorillas, although canine size suggests that the individual was male. Unfortunately, the inferior portion of the symphysis is not preserved, and it is not possible to tell whether or not the inferior margin projected posteriorly as a simian shelf. Can the maxillary and mandibular material (UMP 62-11, 62-10, and 66-01) be referred to a single individual? One way to answer this is to consider the probable dimensions of the mandibular teeth which would be associated with UMP 62-11. Very approximate lower cheek teeth lengths can be gauged by measuring between the hypocone tips of adjacent upper cheek teeth. Thus, the distance between the hypocones of P* and M! is approximately equal to the length of M,, and so on. The length of Mz can be estimated by measuring from the hypocone of M? to the distal border of M3. Estimated lengths from UMP 62-11: P,, 9.0 mm; M,, 12.0 mm; M., 13.0 mm; and Ms, 15.0 mm. Measured lengths of UMP 62-10 and 66-01: P,, **9.5 mm; M,, **11.8 mm; Mg, 13.0 mm; and Ms, 15.0 mm. The correspondence between estimated and measured lengths is good. As is the case for UMP 62-11, the lower teeth are closer in size to those of gorillas than to those of chimpanzees. If the fossil lower tooth measurements are com- pared to those of female gorillas, it will be noted that the fossil breadths lie close to or within the confidence interval of the living pongids, while the lengths of all but Mz lie generally below the lower confidence limit. There is a contrast in the state of preservation of the maxilla and the man- dibular fragments, this could be due to earlier weathering out of the latter. Although the two sets of material were found separated by some distance on opposite sides of an erosion gully, there is no geological reason why they should not represent the same individual. However, at least one further individual of the same species is represented at Moroto by UMP 62-12. Accordingly, the alloca- tion of UMP 62-11, 62-10, 66-01, (and 67-28) to the same individual should be regarded as tentative. TERTIARY PONGIDAE OF EAST AFRICA 45 Using casts of mandibular remains of D. (P.) major from Songhor, a man- dible has been constructed to occlude with UMP 62-11. This mandible is dis- cussed in more detail in a later chapter. From this reconstructed mandible, esti- mates of various mandibular dimensions have been made. Although the Moroto maxillary and mandibular remains may not belong to the same individual, the overall mandibular dimensions inferred from the reconstructed mandible of the undistorted maxilla have been used for estimates of the dimensions of the fossil mandible from Moroto II. These dimensions are listed in Appendix 2. Represented at Moroto, therefore, is at least one individual of a species of large pongid, morphologically and metrically suitable for consideration as an ancestor of Gorilla gorilla. ‘The species represented is most probably D. (P.) major, as Allbrook and Bishop (1963) suggested. However, Leakey (1963, p. 32) disagrees with this allocation and states that, “in my opinion it is not . . [Dryopithecus major] at all. It is pongid, possibly ancestral to Gorilla.” The taxonomic position of these specimens will be discussed in detail in subsequent chapters. UMP 62-12 This tooth, a left upper canine described by Allbrook and Bishop (1963), dem- onstrates the existence of at least a second individual of D. (P.) major at Moroto II. It is similar to Napak canines UMP 62-03, 62-04, and 62-05 in all features of both crown and roots. Measurements are given in Table 14, p. 47, and show the metrical similarities within this sample of canines. There is very little size difference between the various specimens, and, except for the length of UMP 62-12, all measurements fall within the 95 per cent confidence intervals for a small sample of male gorillas and outside those for females. CHAPTER V. DRYOPITHECUS (PROCONSUL) MAJOR FROM NAPAK, UGANDA HISTORY OF DISCOVERY Primate material from Napak in Uganda was first described in 1958 by Leakey (in Bishop, 1958). Further material from Napak was discovered and described in subsequent years. For the purpose of this study, the three primate-yielding sites Napak I, Napak IV, and Napak V are regarded as stratigraphically con- temporaneous. The pongid material recovered from these sites belongs, in all probability, to a single species, D. (P.) major. It will be discussed tooth by tooth, rather than site by site. SPECIMENS Urprrer DENTITION INCISORS UMP 66-03. A right upper lateral incisor from Napak V, the specimen was men- tioned by Bishop (1964, p. 1329). The tip of the root is missing, together with most of the crown, except for the proximal portion of the lingual face. The cer- vical border extends farther towards the occlusal surface on the mesial and distal faces than on the lingual. There is a slight lingual cingulum, but this merges into the wrinkled lingual face of the crown. On the mesiolingual corner there is an oval wear facet approximately 7 mm by 4 mm, worn by the lower canine. ‘The measurements are: MDi L, *8.2 mm, and BulLi L, *10.1 mm. In the limited parts preserved, UMP 66—03 resembles very closely indeed the homologous tooth of UMP 62-11 (p. 29). Wear facets of similar size and shape are found in both specimens. CANINES UMP 62-03. A right upper canine from Napak I, the specimen was described by Leakey in Bishop (1962, p. 8) as a left upper canine. “The specimen, including the enamel structure, is so typical [of Dryopithecus major] that I have no hesitation in my identification.” Allbrook and Bishop (1963, p. 1190) described it as a right upper canine. It was figured by Bishop (1963b). Measurements are given in Table 14. The tooth is similar in overall morphol- ogy to others of this sample from Uganda, although the root is somewhat more slender than those of UMP 62-04 and 62-05 from Napak V. The crown tapers fairly rapidly to a point, more so than in many G. gorilla. The buccal border is increasingly convex towards the tip; the lingual border is decreasingly concave towards the tip. Metrically, like UMP 62-11 and 62-12, 62-03 falls within the 95 per cent confidence limits for male gorillas. Morphologically UMP 62-03 is very similar indeed to the Moroto specimen. The tip of the crown was broken prior 46 TERTIARY PONGIDAE OF EAST AFRICA 47 Table 14. Comparative measurements of C* in D. (P.) major specimens UMP UMP UMP UMP UMP 62-11 62-03 62-12 62-04 62-05 Max L +1356 JIS 7/ 17.6 19.3 18.8 Trans B eS. 7 415.7 16.1 16.0 16.1 Labial H 210 *25.0 ZO) *22.0 eZ 2i619 Index L/B x 100 118.4 Thal), al 09. 2 120. 6 116.8 to fossilization and the dentine is exposed. Enamel has been chipped from vari- ous parts of the crown, particularly mesiobucally. Mesially a groove runs from the cervix towards the tip of the crown. Lingual to the groove there is a broad (approximately 3.0 mm) wear facet worn from contact with the lower canine. Damage to the tip prevents the estimation of total facet length. The buccal border at the cervix is mostly concave, although distally it is flat; the lingual border is similar. There is a slight lingual cingulum. Buccal and lingual surfaces meet at the distal border to form a distinct ridge, at least in the superior half of the crown, and on this ridge, beginning some 6 mm below the cervix, is a 2mm broad facet worn by P3. The root has been damaged after fossilization, the enamel being chipped away lingually in the proximal half. A marked groove runs along the mesial aspect, just lingual to the mesial border. There is a slight depression running from root tip to cervix just buccal to the distal border. UMP 62-04 and UMP 62-05. The specimens are a right upper canine and a left upper canine respectively. They are both from Napak V and have been described by Allbrook and Bishop (1963) and figured by Bishop (1963b). These two canines probably come from the same individual. Morphologi- cally and metrically they are almost identical and they exhibit similar degrees of wear. In both cases the crown tips have been broken subsequent to fossilization. There are barely discernable wear facets mesially and distally. The crowns and roots are larger in length and breadth dimensions than is the case for other canines in this sample (see Table 14). The crown height however is relatively small, giving them a “stubby” appearance. The length and breadth measure- ments suggest that the specimens are from a male individual—certainly there must be few, if any, present day female gorilla canines with dimensions as large as those of UMP 62-04 and 62-—05—yet it would be difficult to find male gorillas with canine crowns as low as these. FIRST MOLARS UMP 62-07 (Fig. 21). A right upper first molar from Napak V, the specimen was described by Allbrook and Bishop (1963) and figured by Bishop (1963b). The buccal roots have been broken subsequent to fossilization. The mesiobuccal corner of the crown has also been broken off. This has been restored, although a small segment of the buccal surface and part of the paracone are missing. The occlusal surface is rhomboidal in outline, paracone and metacone being rela- tively mesial to protocone and hypocone. There is a marked distal bulge. The protocone is the largest cusp, the hypocone, metacone, and paracone being sub- equal and smaller. The mesial moiety is also somewhat broader buccolingually 48 PEABODY MUSEUM BULLETIN 31 than the distal. The cusps tend to be pyramidal rather than rounded, particu- larly the relatively unworn buccal cusps. Crests are similarly rather sharply defined. Moderate wear facets are present on the occlusal surface and a tiny area of dentine is exposed on the tip of the protocone. A protoconule is developed just lingual to the midpoint of the mesial border and is lingual to the sulcus separating paracone from protocone. There is no marked crest between paracone and protocone. The protoconule is worn and merges into the protocone, although a small furrow remains between them on the lingual side. The metacone-protocone crest is crossed by a sulcus separating metacone from protocone. A small cuspule lies between metacone and hypo- cone. There is also an accessory cuspule distal to the hypocone. The lingual cingulum runs lingually from the protoconule round the pro- tocone and hypocone to the accessory distal cuspule. It is much less prominent on the hypocone than on the protocone. It is beaded, particularly between pro- tocone and hypocone, and the beading is carried onto the lingual surfaces of these cusps as vertical wrinkles which have been variably worn. A distal ridge or cingulum surrounds the posterior fovea distally. The buccal cingulum is rep- resented by a distinct spur-like mesostyle between paracone and metacone and by slight lipping at the mesiobuccal corner of the paracone. There is a mesial wear facet some 3 mm long, worn from contact with P‘, and a distal facet approximately 4 mm long. The lingual root diverges lingually at an angle of about 60 degrees to the occlusal plane (approximately 30 degrees to the buccal roots) as in UMP 62-11. It is characterized by a lingual longitudinal groove. Crown and root measurements are included in Table 15. Allbrook and Bishop assigned this tooth tentatively to D. (P.) major. UMP 66-41 (Fig. 21). The specimen is a right upper first molar from Napak IV and is mentioned by Bishop (1964, p. 1329). The crown and lingual root are per- fectly preserved. The buccal roots are broken 8 mm (mesial root) and 4 mm (distal) below the cervical border. The lingual root is splayed lingually as in UMP 62-07 and 62-11. Crown features are essentially similar to those of UMP 62-07, although the cusps are somewhat more pyramidal and the crests more sharply defined. The metacone is relatively better developed than in UMP 62-07. There is some occlusal wear, and a small area of dentine is exposed on the protocone tip. The smaller cuspules—protoconule and accessory hypocone cusp—are pres- ent, although both are rather less distinct than in UMP 62-07. The paracone- Table 15 Comparative measurements of M?in D. (P. ) major specimens UMP UMP ‘UMP 62-11 62-07 66-41 Left MDi L *11.7 *11,4 1210 BuLi B WAS 7) *13.9 TARO Index L/B x 100 CHA dl 82.0 85.7 Crown area *148. 59 *158. 46 168.0 Vertical L Ling. root *15.0 *15.0 *15.0 MDi T Ling. root, maximum Cbe 8.3 9.5 TERTIARY PONGIDAE OF EAST AFRICA 49 protocone crest is better developed than in UMP 62-07 and the two cusps are not separated by a sulcus. The crest merges with the protocone just lingual to the tip of that cusp and is joined at the point of merging by a small distal extension of the protoconule. The lingual cingulum is better developed than in UMP 62-07 and is clearly demarcated around the hypocone as far as the accessory distal cuspule. Although reduced by wear, the beading of the cingulum and the vertical wrinkling of the lingual faces of protocone and hypocone are still clearly visible. The distal cingulum reaches the distobuccal corner and is ex- tended onto the buccal border mainly as heaping of the enamel rather than as a distinct ledge. The mesostyle is poorly developed. Measurements are given in Mable 15. Although the crown features of UMP 62-11 right M1! from Moroto IT have been largely obliterated by wear, the essential similarities to these other right upper first molars, UMP 62-07 and 66-41, are immediately obvious (Figs. 18 and 21) and all can be treated as one species. All are rhomboidal in crown outline, with a marked distal bulge. Mesiodistal lengths are similar, ranging only from 11.4 to 12.0 mm. The buccolingual breadth varies rather more, from 12.7 mm to 14.0 mm. The differences in shape are shown by the crown indices, varying 10.1 units from 82.0 to 92.1. Such a range is certainly not greater than those found within species (see Table 9, p. 35). The details of cuspal and sulcal mor- phology are very similar, as far as can be ascertained within this small sample. The lingual cingulum is best developed in UMP 66-41, where it is as well marked on the hypocone as it is on the protocone. In UMP 62-07 it is less well developed on the hypocone and in UMP 62-11 the hypoconal cingulum is probably less developed than in UMP 62-07. In contrast the mesostyles of UMP 62-11 and 62-07 are more prominent than those of UMP 66-41. The morphology and placement of the roots are similar in all three teeth. There is no reason to place these teeth in more than a single species. All four main cusps are well-developed, prominent, and rather pyramidal in form. The intercuspal crests are also well developed. ‘The tendency towards hyp- sodonty and lophodonty is not as marked as in Gorilla gorilla and the tooth crowns are not so large. The paracone-protocone and metacone-hypocone crests are generally better developed in living gorillas. The possible functional ex- planations for such changes have been noted in the preceding chapter. The cingulum is generally quite well developed in gorilla upper molars (Korenhof, 1960; Frisch, 1965; Remane, 1960). It is rarely as clearly marked as in these Miocene specimens, although the overall morphology—the beading and vertical wrinkling—is otherwise very similar. In G. gorilla, the buccal cusps, especially the metacone, are well developed and the distance between the tips of paracone and metacone is generally very similar to that between protocone and hypocone. In the Miocene species the distance between the lingual pair is generally greater, making the distal bulge more pronounced on the lingual side of the crown. These specimens from Uganda make reasonable morphological precursors for living gorillas. If they are indeed ancestral, the main trends since the Miocene would have been towards increase in crown size producing cusps of subequal volume, increased hypsodonty and lophodonty, and some cingular reduction. 50 PEABODY MUSEUM BULLETIN 31 SECOND MOLARS UMP 62-08 (Fig. 21). A left upper second molar from Napak V, the specimen has been described by Allbrook and Bishop (1963, p. 1190) and figured by Bishop (1963b). The roots have been broken off since fossilization, exposing an extensive pulp cavity. The cavity extends furthest occlusally below paracone and metacone. The crown is rhomboidal, the lingual border being longer than the buccal. The four main cusps are subequal in size, the greater length of the lingual moiety being accounted for by accessory cuspules and mesial and distal cingula. Although moderately worn, the cusps, particularly the buccal, still re- tain their distinct pyramidal form. The metacone-protocone crest is well de- veloped. Paracone and protocone are separated by a sulcus, as are metacone and hypocone. There is a prominent protoconule on the mesial border, from which two small crests run buccally to the paracone and one distolingually to the protocone. The lingual cingulum is well developed, beginning at the lingual border of the protoconule, and running around the protocone and hypocone to a distal accessory cusp, distal to the hypocone. It is beaded and the lingual surfaces of protocone and hypocone are vertically wrinkled. Around the hypocone beading and wrinkling merge and the cingulum is not as clearly demarcated as it is on the protocone. The vertical wrinkling is continued on the buccal border. A small mesostyle is present. A distal cingulum or marginal ridge runs from the distal accessory cusp to the distobuccal border. A small posterior fovea separates this from the metacone-hypocone crest. The fovea sides are also wrinkled. Con- tact facets 4 mm long are present on the mesial and distal borders. Measurements are given in Table 16. UMP 62-09. A left upper second molar from Napak I, the tooth has been described by Allbrook and Bishop (1963, p. 1190). The specimen is part of an unerupted molar crown. The crown is broken and only the buccal three-fifths or so has been preserved. The paracone, metacone, and half of the hypocone are present, together with the metacone-protocone crest, and a pair of lingual crests from the paracone. Comparison with UMP 62-08 indicates that this is, with a high degree of probability, a second molar. The occlusal surface is covered with wrinkles, varying from relatively fine in the trigon fossa to relatively coarse in the posterior fovea. Measurements are given in Table 16. UMP 62-11 and 62-08 are very similar (Figs. 18 and 21) and can be regarded as representatives of a single species. Allbrook and Bishop placed them provi- sionally in D. (P.) major. Apart from the stronger mesostyle of UMP 62-11 (mesostyles are well developed on all three molars of the individual from Moroto II), there is hardly a morphological feature present in one which is absent in the other. UMP 62-08 is considerably larger than the homologous tooth of UMP 61-11. The size and shape range for male gorilla M?, shown in Table 11, p. 36, is great and there is no reason to doubt that UMP 62-11 and 62-08 are to be placed in one species only. UMP 62-11 is almost certainly a male and, because of size, UMP 62-08 probably is, too, indicating that in this species there were individuals with larger teeth than UMP 62-11 and quite possibly larger maxillae and faces. Like M! the second molars show resemblances to those of living gorillas. There are also a number of differences, but none of the differences rules out the TERTIARY PONGIDAE OF EAST AFRICA 51 Table 16 Comparative measurements of M’ in: (a) D. (P.) major, (b) G. gorilla UMP UMP NMK NMK NMK UMP (a) 62-11 62-08 102, 405, Tol, 62-09 Left CMH35 381 CMH34 MDiL cikeie al lye MAS / 14.4 12.0 U5s2 BuLi B 14.0 W/S83 T5o3 Oars 13.6 - Index L/B x 100 93.5 87a 83.0 88.3 88.2 - (b) G. gorilla G. gorilla males females (n=20) (n=20) x 16. 0 14,8 MDi L 95%CL U3eo TOs 12. 8-16.8 me 1657, 15.6 ae 95%CL 14, 8-18.5 13. 9-725 X 96.2 94,7 Index 1/B x 100 OR 91. 1-106.2 87. 4-104.5 probability that the fossil species represented is close to the ancestry of G. gorilla. Lower DENTITION AND MANDIBLES UMP 62-06 (Fig. 22). The specimen, from Napak V, has been described by Allbrook and Bishop (1963, p. 1190) and figured by Bishop (1963b). It consists of the anterior portion of the left mandibular horizontal ramus of an infant. It was tentatively assigned to D. (P.) major by Allbrook and Bishop (1963, p. 1190). The anterior surface of the mandible and the alveolar margin have been eroded, exposing the crown of an unerupted incisor, the alveolus of the decidu- ous canine and the alveolus for the mesial root of the deciduous anterior pre- molar. The distal root of dP, and mesial root of dP, are preserved in the jaw; dP, was elongated mesiodistally. The unerupted crown of P, (not M, as Allbrook and Bishop tentatively suggested) was exposed on the lateral surface of the ramus and this was carefully removed for study. Openings to gubernacular canals are present lingual to the incisor and canine. The specimen is broken mostly lateral to the symphysis, although the actual symphysis is represented in the inferior and posterior parts of the region, there being some local excavation sug- gesting incomplete fusion. The posterior symphyseal surface slopes gently inferiorly to the superior transverse torus, forming a planum alveolare, the most posterior point of which is opposite the distal root of dP;. The posterior symphyseal surface then slopes slightly anteriorly and inferiorly to the inferior transverse torus. Thus the most posterior point of the inferior transverse torus is just anterior to that of the superior torus. Between the tori is a small genial pit. Although a simian shelf or basal plate is absent inferiorly, the symphyseal structure of UMP 62-06 is very similar to that of infant Gorilla gorilla, where a gently sloping planum is gen- 52 PEABODY MUSEUM BULLETIN 31 erally present, in contrast to the usual condition in Pan troglodytes where the planum normally slopes more abruptly. (It was inferred in Chapter IV that a long, gently sloping planum would have been present in UMP 66-01/62-10 from Moroto II.) At the level of dC,, the horizontal ramus is about 28 mm deep. The body shallows posteriorly. Some very approximate measurements are possible for the tooth sockets and roots. The depth of the canine alveolus from the interdental septum is approxi- mately 20.0 mm; and the buccolingual breadth of the distal root of dP; is about 5.0 mm and that of the mesial root of dP, about 5.5 mm. The incisor was identified by Allbrook and Bishop as a lateral incisor. Mor- phologically, however, the crown resembles those of pongid central incisors. Careful inspection reveals two gubernacular openings distal to the exposed in- cisor and mesial to dP; (for C, and I,). It is unlikely, too, that there would be sufficient space between the mesial border of the incisor and the midline to accommodate another permanent tooth. The tooth is therefore regarded here as a central incisor. In sagittal section the crown is elongate and triangular, the labial surface being gently convex and the lingual more strongly concave. The labial surface is symmetrical, mesial and distal labial margins being of equal length in contrast to the condition found in pongid lateral incisors where the distal border is shorter. Lingually, too, the crown is symmetrical. There are slight marginal ridges and a hint of central thickening in the proximal two- thirds of the crown. The crown narrows towards the cervix more markedly on the lingual surface. The crown is broken lingually just above the cervix. The measurements are: MDi L, 6.4 mm, LaLi B, *8.0 mm, La H, *14.0 mm, and Li H, *13.0 mm. In overall shape the tooth resembles those of gorillas rather than chimpanzees, the crown being relatively narrow mesiodistally, high, and symmetrical. The mesiodistal diameter at the lowest level of the cervical margin is 4.4 mm. The mea- surements of I, roots of the D. (P.) major mandible from Songhor described by Clark and Leakey (1951, p. 59) are MDi, 4.0 mm, and Lali, 7.2 mm. ‘These measurements are taken on roots below the cervix and will be some- what smaller than the dimensions at the cervical level. Metrically, therefore, the Napak and Songhor specimens are very similar. The P, was removed from its crypt. In position the most superficial part of the tooth would have been at least 8.0 mm below the alveolar margin. Resorption of dP, roots had probably not begun. The crown is relatively more molariform than in other posterior lower premolars in this sample (Fig. 22). The crown is clearly not fully formed, the buccal surface in particular being less fully convex than in adult specimens. There is a good deal of relief on this surface too. The protoconid and metaconid are subequal in height, the protoconid being mar- ginally greater in area. The two cusps are joined by a crest, itself crossed by a mesiodistal sulcus. Mesial to the crest is an anterior fovea, surrounded in front and to the sides by a mesial cingulum or marginal ridge. The mesial portion of the tooth is well developed relative to others in this sample. Small buccal and lingual cusps are present on this ridge just mesial to protoconid and metaconid. Distal to the protoconid, and separated from it by a sulcus, is a small hypoconid. From the hypoconid a ridge surrounding the posterior fovea runs distally and then lingually to the metaconid. The ridge is beaded. The sulcus between proto- TERTIARY PONGIDAE OF EAST AFRICA 33 conid and hypoconid is continued down the buccal surface as a prominent groove. An unerupted P, at an approximately similar stage of development to UMP 62-06 was dissected from an infant gorilla (Duckworth Pr 52.0.5). In this gorilla the main cusps are relatively more massive, and the main crest and the mesial and distal marginal ridges less well developed—possibly due to the greater im- maturity of this specimen. In general features, however, the two teeth are similar. The only reliable measurements possible are those of length. The mesiodistal length would probably not have exceeded 10.3 mm. The trigonid length was about 6.9 mm and the talonid length 3.4 mm, giving a talonid/trigonid length index of 49.3. In full development the buccolingual breadth would have been at least 10.0 mm and possibly several tenths of a millimeter more. UMP 62-13 (Fig. 22). The specimen from Napak I, figured in Bishop (1963b), consists of the anterior part of the right horizontal ramus from an infant mandi- ble. Apart from a small portion of the symphysis, only the superior parts of the ramus are preserved. The roots of dC,, dP;, and dP, are preserved, the crowns being broken off at the junction of crown and root. The fragment is broken mainly lateral to the symphysis, except inferiorly, and the unerupted and broken crown of I, is exposed. The first permanent molar had erupted through the gum, wear facets show that the tooth had come into occlusion, although eruption was probably not complete. Openings to the gubernacular canals are present lingual to dC, and dP3. The symphyseal morphology is similar to that of UMP 62-06. A gently sloping planum alveolare was present, leading backwards to a superior transverse torus situated opposite the distal root of dP. Inferior to this torus is a shallow area of excavation, the genial pit. There was apparently no posteriorly projecting simian shelf, although the superior transverse torus projects only slightly posterior to the inferior transverse torus. A few measurements are comparable with UMP 62-06. The depth of the ra- mus opposite dC, is approximately 26 mm, some allowance being made for supe- rior and inferior breakage. The buccolingual diameter of the distal root of dPs is 4.6 mm and of the mesial root of dP, 6.5 mm. UMP 62-13 is a little smaller and more gracile than UMP 62-06, but the differences are no more than individual or possibly sexual, and the two specimens are assignable to a single species. Crown dimensions have been estimated from the roots of the deciduous teeth. They are: dC,, Max L, *7.5 mm, Trans B, *5.0 mm; dP;, MDi L, *8.5 mm; and dP,, MDi L, *9.8 mm, BuLi B, *7.8 mm. The unerupted posterior premolar was carefully removed from its crypt by drilling. The most superficial part of the crown was some 4.0 mm from the alveo- lar margin. The crown may not be fully formed, although it is certainly more developed than UMP 62-06 and the dimensions are probably close to what they would have been in the erupted condition. The buccal surface is fuller and more rounded than in UMP 62-06 and is less vertical than the lingual surface, thus the cusps appear to be situated more towards the lingual edge of the occlusal surface. The anterior fovea is less well developed than in UMP 62-06; protoconid, metaconid, and hypoconid are similar to those of UMP 62-06. There is a slight buccal groove running from the protoconid-hypoconid sulcus. The enamel ex- 54 PEABODY MUSEUM BULLETIN 31 tends farthest inferiorly on the buccal side. The lingual part of the distal border projects farthest distally. The measurements (also included in Table 19, p. 61) are: MDi L, 8.6 mm, BuLi B, 9.1 mm, Index L/B X 100, 93.5; Wri 1) 59", Tal L, 2.7 mm, and Index Tal L/Iri L x 100, 45.8. The first molar is slightly worn. There is no distal wear facet. Morphologically this tooth is very similar to others in this sample. The metaconid, entoconid, and hypoconid are subequal in size, the protoconid somewhat smaller, and the hypo- conulid smallest. The metaconid is slightly distal to the protoconid and the hypo- conulid is shifted a little buccal to the midline. Metaconid and hypoconid are in contact for about 2 mm, and hence the “Dryopithecus Y” pattern is clearly developed. The tail of the Y is only slightly mesial to the tip of the hypoconid, and the two arms diverge at approximately equal angles from the tail. The buccal cingulum is present between protoconid and hypoconid, closing buccally the groove between these two cusps. This part of the cingulum is faintly beaded. There is a pit at the mesiobuccal corner and a buccal extension and deepening of the sulcus between hypoconid and hypoconulid, representing the mesial and distal limits of the buccal cingulum. Mesial and distal marginal ridges enclose the anterior and posterior fovea. Crests unite adjacent cusps around the margin of the occlusal surface and are particularly well developed between metaconid and protoconid, and entoconid and hypoconulid. The cusps are well defined and pyramidal. There is a tiny protoconulid mesial to the protoconid. Grooves on the distal faces of metaconid and proto- conid and on the mesial face of metaconid demarcate tiny cuspules, too small to warrant naming. The measurements (also included in Table 20, p. 62) are: MDi L, 10.8 mm, Tri B, 9.0 mm, Tal B, 9.2 mm, Index L/B x 100, 117.4, and Index Tal B/Tri B x 100, 102.1. UMP 62-14 (Fig. 22). This lower first molar from Napak I was described by Leakey (in Bishop, 1958, p. 1481) as representing a “quite typical lower right second molar’ of Dryopithecus (Proconsul) nyanzae. It was figured by Bishop (1958, 1963b). The specimen consists of the very well-preserved crown of what is probably a first molar (rather than a second molar). The roots are broken some 5 mm below the cervical border. There is a mesial wear facet from contact with dP, or Py; a small distal area of wear on the hypoconulid suggests that M, had just erupted. Occlusal wear has not proceeded very far although farther than in UMP 62-13; a tiny patch of dentine has just appeared on the tip of the proto- conid. Although described as a M, by Leakey, comparisons with other D. (P.) major M,’s from Uganda (Fig. 22) and from Songhor in Kenya show that this is a first molar. Morphologically it is very similar to UMP 62-13 and others in this series. However, isolated first and second molars of Dryopithecus (Proconsul) species are often difficult to tell apart on morphology alone. Size is not a uni- formly helpful criterion either. It is possible, within single sex gorilla series, to find with some frequency M, and M,j of similar size and shape, and this tendency is accentuated when pooled male and female series are used (see Appendix 1). The largest cusps are metaconid and hypoconid, with hypoconulid the small- est. The hypoconulid is just buccal to the midline and the metaconid just distal to the protoconid. The Y pattern is again well developed. In all these features TERTIARY PONGIDAE OF EAST AFRICA a9 UMP 62-14 resembles UMP 62-13. The cingulum is prominent and quite strongly beaded, the beading running onto the protoconid and hypoconid as vertical wrinkling. The buccal cingulum is almost continuous from the mesiobuccal corner to the hypoconulid, although it is interrupted for a short distance on the protoconid; the pit at the mesiobuccal corner is very similar to the condition noted in UMP 62-13. In almost all features of morphology, extending even to the most trivial of occlusal wrinkles and accessory cuspules, the crowns of UMP 62-13 and 62-14 are very similar. The measurements (also included in Table 20, p. 62) are: MDi L, 12.0 mm, Tri B, 10.2 mm, Tal B, 10.5 mm, Index L/B x 100, 114.2, and Index Tal B/Tri B x 100, 103.0. UMP 62-15 (Fig. 22). The specimen, a left lower first molar from Napak I, was figured by Bishop (1963b). This tooth had apparently only just erupted. There are a few traces of wear on the mesial moiety of the crown and what is possibly a tiny mesial facet. The roots have been broken off at the cervical border. The occlusal surface has the rather coarse wrinkles characteristic of unerupted or unworn teeth. The metaconid is more markedly distal to protoconid and the hypoconulid more centrally placed, than on either UMP 62-13 or 62-14. Both features point to this tooth being M, rather than M, (Frisch, 1965, p. 79-94). Unlike UMP 62-13 and 62-14, the buccal cingulum is continuous from the mesio- buccal corner to the hypoconid. It is represented by vertical wrinkling on the more distal and buccal parts of the hypoconid and by the distobuccal deepening of the intercuspal sulcus between the hypoconid and hypoconulid. The measure- ments (also included in Table 20, p. 62) are: MDi L, 12.7 mm, Tri B, 10.6 mm, Tal B, 10.9 mm, Index L/B x 100, 116.5, and Index Tri B/Tal B x 100, 102.8. UMP 62-16 (Fig. 22). This specimen, from Napak I and described by Leakey (in Bishop, 1962, p. 8), consists of the left lower last premolar and first molar, joined by adherent alveolar bone. The buccal root of Py has been broken off sub- sequent to fossilization and the crown rotated slightly in a distobuccal direction. The crown of M, has a small crack running lingually from the hypoconid to the lingual border. This has increased the trigonid diameter to 9.5 mm, although originally it was probably 9.3 mm or 9.4 mm. Both teeth are moderately worn, with dentine being exposed on the protoconid, metaconid, and hypoconid of M,; the buccal wrinkling and cingulum beading are practically removed. Leakey tentatively identified this specimen as Sivapithecus africanus? Clark and Leakey, though he has since transferred it to Kenyapithecus (1967). His description in Bishop (1962, p. 8) is here quoted almost in full: The measurement of the molar tooth is 11 x 9 mm, which is not nearly large enough for Proconsul major. Moreover the tooth in question does not look like that of a Proconsul major. On the other hand it is far too large for the first lower molar of Proconsul nyanzae, from which it also differs in the complete absence of a cingulum. The 4th premolar also does not seem to resemble that of Proconsul nyanzae. Unfortunately we know nothing definite about the lower dentition of Sivapithecus africanus, but if we tentatively place this specimen in this genus and species we shall not be far wrong. 56 PEABODY MUSEUM BULLETIN 31 The last premolar is similar in size and morphology to UMP 62-13, although it is not as long (Fig. 22). It also resembles BM(NH) M14086 from Koru, described by Clark and Leakey (1951) as Proconsul nyanzae, in size and shape (Fig. 23) and those specimens of D. (Proconsul) from Songhor figured by Clark and Leakey (1951, pl. 4, figs. 19 and 22; pl. 5, fig. 35). The anterior fovea and the morphology of the buccal surface, with mesial and distal vertical grooves, are also similar to other premolars in the Uganda and Kenya samples. The measure- ments (also included in Table 19, p. 61) are: MDi L, 7.9 mm, BuLi B, 9.0 mm, Index L/B x 100, 87.8, Tri L, 5.3 mm, Tal L, 2.6 mm, and Index Tal L/Tri L x 100, 49.0. Although the M, of UMP 62-16 is more extensively worn than others in this sample, the occlusal morphology is again remarkably similar. The metaconid is slightly distal to the protoconid and the hypoconulid is just buccal to the midline. The accessory cuspules on protoconid and metaconid are identical to those on UMP 62-13 and 62-14 (Fig. 22). A pit is present on the mesiobuccal corner, as in UMP 62-13 and 62-14. Leakey’s statement that there is “complete absence of a cingulum”’ is rather puzzling. Although not as prominent as that of UMP 62-14, the cingulum be- tween protoconid and hypoconid is nevertheless quite distinct. In recent Pon- gidae and those Dryopithecinae where cingula are reduced or lacking [many specimens of Dryopithecus (Sivapithecus) species for example], the sulcus be- tween protoconid and hypoconid descends the buccal face uninterrupted, some- times almost to the cervical border, before shallowing and merging with the enamel surface. In all the East African mid-Tertiary pongid material the sulcus is terminated buccally by this part of the buccal cingulum. Although reduced in height and smoothed out by wear in UMP 62-16, it is nevertheless clear that a buccal cingulum is present, and was as extensive as that on UMP 62-14 and 62-13. This also applies to the cingulum or marginal ridge between hypoconid and hypoconulid. The measurements (also included in Table 20, p. 62) are: MDwhL, 107 mm, Tri B, *9.4 mm, Tal B, 9:4:mm, Index L/B x 1007 *ila8; and Index Tal B/Tri B x 100, *100.0. UMP 66-02 (Fig. 22). This specimen, a slightly worn and rootless crown of a right lower last premolar from Napak V, was mentioned by Bishop (1964, p. 1329). There are small mesial and distal contact facets. The distolingual corner projects a little distally. In this feature UMP 66-02 is intermediate in morphology between UMP 62-13 and 62-16. In other features of occlusal surface morphology UMP 66-02 resembles the P, of UMP 62-16 closely. The cervical border extends inferiorly at the mesiobuccal corner, as in all other premolars in this sample, and also those from Songhor and Koru (described in Chap. VI). The measurements (also included in Table 19, p. 61) are: MDi L, 7.8 mm, BuLi B, 9.0 mm, Index L/B x 100, 86.7, Tri L, 5.2 mm, Tal L, 2.6 mm, and Index Tal L/Tri L x 100, 50.0. DISCUSSION OF THE LOWER DENTITION The last premolars show some variation in size and morphology, but this is not as great as that seen within the presently living African pongid species (see Table 19). Similarly the first molars exhibit no more metrical and morphologi- TERTIARY PONGIDAE OF EAST AFRICA 57 cal variability than might be expected within a single species, particularly if that species exhibits some sexual dimorphism in size (see Table 20). The cusps are pyramidal rather than rounded, particularly on the lower molars. Like the upper molars considered before, these lower molars are morphologically acceptable as ancestral to G. gorilla. In the living form, the homologous teeth are very much larger and the cusps are often more projecting. In terms of overall tooth size the cusps on these fossil teeth are no lower than those of most gorillas. However, the relief of the gorilla molar is much more striking. The sulcus between metaconid (generally the highest cusp) and ento- conid, for example, usually approaches the cervical margin much more closely in gorillas. The newly erupted M, crown from Napak I, UMP 62-15, was com- pared with an unerupted, although fully formed, M, of G. gorilla (Duckworth Pr 52.0.5), see Table 17. Table 17 M, measurements of D, (P.) major and G. gorilla UMP 62-15 ProZ.Qs Se Tri B 10.6 2s . Metaconid H Tit 2, 8.6 Index H/B x 100 67.9 70.5 Metaconid/Eniocd. sulcus H 55 4,0 Index cusp H/ sulcus H x 100 130.9 Aalisy 50) Although protoconid height, measured from the cervical border, is relatively little greater in the gorilla, the cusp relief is more marked because the intercuspal sulci and basins are deeper. Both last premolars and first molars are more loph- odont in gorillas than in the fossil species. DISCUSSION OF ‘THE UGANDAN PONGIDS There is no reason to assume that the pongid material from Moroto II, Napak I, IV, and V described above should be referred to more than a single species, particularly if the individual, sexual, geographical, and temporal factors affect- ing variability within a single evolving lineage are considered. It has already been established, at least, that the upper and lower dentitions fall into single species groupings; can the two dentitions be associated? Unfortunately, only one site, Moroto II, has yielded remains which might possibly include associated upper and lower dentitions. From Napak I we have only lower first molars, while uppers are known only at Napak IV and V. However, treating the Napak sites as a single assemblage and including Moroto II as well (although this site could prove to be geologically somewhat younger), we have for consideration the upper and lower dentitions of a large pongid. Cuspal and cingular morphology is very similar in the two sets of material. ‘These maxillary and mandibular samples resemble each other as closely as do upper and lower dentitions of P. troglodytes, G. gorilla, D. (P.) africanus, or D. (P.) nyanzae. It has also proved possible to occlude UMP 66-41 from Napak IV 58 PEABODY MUSEUM BULLETIN 31 (M1) with UMP 62-14 from Napak I (M,), producing an extremely good fit (Fig. 24). It is of interest to note here the functional role of some of the crown struc- tures. With the upper molar stationary, when the lower molar moves buccally from centric occlusion the paracone and metacone tips occlude between the buccal cingulum and the buccal cusps of the lower molar. When the lower molar moves lingually, the lower molar lingual cusps occlude between the lingual cin- gulum and cusps of the upper molar. In both cases the cingula act as extra cusps. During side-to-side movements the protoconule, a prominent feature of the upper molars, occludes with the protoconid-metaconid crest. The upper and lower dentitions from Uganda and the associated maxillary and mandibular material, are probably sampled therefore from a single species. The species includes those specimens described by Allbrook and Bishop (1963) as Proconsul major, by Leakey (in Bishop, 1958 and 1962) as Proconsul nyanzae and Sivapithecus africanus, and by Simons and me (1965) as D. (P.) major. Whether or not this assemblage does represent D. (P.) major, known before 1958 only from Kenya, will be considered in the following chapter. CHAPTER VI. DRYOPITHECUS (PROCONSUL) MAJOR FROM KENYA BACKGROUND OF THE MATERIAL As noted in Chapter III, the collections from Songhor and Koru in Kenya may be considered as representing relatively short periods of time. Koru and Songhor are geographically close together and probably temporally close, too. The ma- terial from the two main sites can be considered here as forming a single as- semblage. Napak, Moroto II, Songhor, and Koru represent stratigraphically a relatively uncomplicated set of sites, unlike the Rusinga series. ‘This assemblage is probably broadly contemporaneous, although as noted Moroto II may be a little younger than the others. Quite possibly all these sites were deposited in similar environments. The history of recovery of primate material from the East African Miocene sites has been discussed by Clark and Leakey (1951, p. 1-3). The first pongids to be described came from Koru (Hopwood, 1933 a and b). Specimens which concern us here are the maxilla which Hopwood made the holotype of a new genus and species, Proconsul africanus BM(NH) M14084 (see Fig. 25) and a man- dible BM(NH) M14086 (see Fig. 23). In 1951 Clark and Leakey described two further species in this subgenus, P. nyanzae and P. major, and included the referred mandible in P. nyanzae. All but one of the specimens assigned by Clark and Leakey to P. major came from Songhor. This specimen, NMK 71, CMH 142*, from Rusinga has since been transferred by Leakey (1967) to his proposed species Kenyapithecus africanus. In 1952 Clark described further ma- terial from East Africa, including several more specimens of P. major from Songhor. Two also came from Rusinga, one a dPy (NMK 129,257). The other specimen was part of a right clavicle which came, according to Clark (1952, p. 280) from the surface, and according to the Museum catalogue from site R3a. This specimen is similar to the clavicle from Rusinga, site R106 (NMK 145,604) which Clark and Leakey suggested might belong to P. major (1951, p. 99). This clavicle from R106 has since been assigned tentatively by Leakey to his proposed species Kenyapithecus africanus (1967, p. 163). In 1965 Simons and I trans- ferred all Proconsul species to Dryopithecus. The dental and mandibular material from Kenya assigned with a reasonable degree of certainty to D. (P.) major therefore comes almost exclusively from Songhor. Most of the more complete remains have already been described in some detail in the literature. However, dimensions and certain further relevant observations are included below. Unless otherwise stated, specimens come from Songhor. * National Museum of Kenya primate specimen numbers consist of two sets of figures, sep- arated here by commas. The first set lists the primates consecutively; the second set is appar- ently a field number and varies with locality and date of collection. a9 60 PEABODY MUSEUM BULLETIN 31 LARGE SPECIMENS BM(NH) M16648 (Fig. 23). This specimen, the holotype of D. (P.) major, has been described and figured by Clark and Leakey (1951) and consists of part of a right mandibular horizontal ramus with the roots only of P; and the roots and crowns of P, through M3. The second molar is damaged distally and in the distobuccal quarter of the crown. In the mandibular parts preserved, BM(NH) M16648 is very similar to UMP 66-01 from Moroto I. The roots of Ps; show that the long axis of this tooth was set at an angle of about 40 degrees to the axis of the distal tooth row. The crown size of P, has been estimated from the roots. The large size of BM(NH) M16648 suggests that this is probably a male specimen. Compared with individuals of G. gorilla (see Appendices), this D. (P.) major has relatively small P;, Py, and My, and a relatively large M3. The combined specimen from Moroto II (UMP 62-10 and 66-01) probably has a rather larger P, and a somewhat smaller M;. Could these variations in size and proportions occur within a single species? In a sample of 20 male lowland Gorilla gorilla (subspecies G. g. gorilla), the ratio (MDi L P3/MDi L M,) x 100 varied over 26 units from 99.4 to 125.8 with a mean of 111.8. This is one method, albeit unsatisfactory, of measuring the relative proportions of P3, M, being chosen as standard. The ratio for BM(NH) M16648 is 115.7 (14.0/12.1 x 100), just a little above the gorilla mean. In the Moroto II mandible the ratio is 135.6 (16.0/11.8 x 100), 20 units higher. It should be noted, however, that there is evidence from Songhor (see below) of lower first premolars at least as long as those from Moroto II and no evidence of substantially larger mandibles (with larger M,’s) in which these P3’s could have been set. Considering the morphological features of similarity, it is unlikely that this difference in ratio between UMP 62-10 and 66-01, and BM(NH) M16648 is to be regarded as significant. A number of authors (e.g., Clark and Leakey, 1951, p. 37, 56, 57, 62) have used the ratio (MDi L M,/MDi L M3) x 100 to measure the relative lengths of the last two lower molars. In the sample of male gorillas, this ratio varies from 86.6 to 109.9, a typically large range of variation. The value of the ratio in BM(NH) M16648 is 82.2 (14.8/18.0 x 100) and in UMP 66-01, 86.7 (13.0/15.0 x 100), only 4.7 units difference. The actual length of Mz; in BM(NH) M16648 is 18.0 mm and in UMP 66-01 15.0 mm. In the gorilla sample the mean value of this character is 17.6 mm and the 95 per cent confidence limits are 15.5 mm and 19.7 mm. Once again, considering both relative and absolute lengths of M3, there is no reason to assign these two specimens to more than a single species. Although the crowns of P, and M, are rather worn in BM(NH) M16648, enough detail is preserved to show that the dental material from Napak is very similar morphologically and metrically to that from Napak. Measurements are set out in Tables 18-22 which follow. NMK 190,1 (Figs. 24 and 26). This specimen, described by Clark and Leakey (1951) and figured in Clark and Leakey (1951) and Leakey (1967), is the main portion of a right canine crown, found in 1932. The symphyseal region, TERTIARY PONGIDAE OF EAST AFRICA 61 Table 18 Comparative measurements of P, in: (a) D. (P.) major, (b)G. gorilla UMP NMK BM(NH) NMK NMK NMK BM(NH) (a) 66-01 190,1 M16648 382,3 383,4 604,69 M14086 Left Right Max L **16.0 **13.0 **14,0 *16.0 16.0 14.4 ee, Trans B **9,5 **Q, 2 **9,0 *9.0 9.0 9.8 752. G. gorilla G,. gorilla (b) males females (n=20) (n=20) x 17215 14.9 niet 95%CL 15. 4-19. 6 13. 1-16.7 # . x Ded 10.4 zens 95%CL 9. 7-14. 4 ey ta Re Table 19 Comparative measurements of P, in: (a) D. (P. )major, (b) G. gorilla and P, troglodytes (a) UMP UMP UMP UMP UMP BM(NH) NMK NMK BM(NH) 62-06 62-13 66-01 62-16 66-02 Ml6648 190,1 272,56 M14086 MDi L *10,3 8.6 **9,5 7.9 7.8 9.2 9.0 9.0 728 BuLi B - 9,1 **10.0 9.0 ath Say 9:7 0.6 9.0 Index 1/B 93.5 95.0 87.8 86.7 94.8 92.8 84.9 86.7 x 100 Tri L *6.9 5.9 = ee SoZ 6.9 6.3 = = Tal L *3,4 a, = 2.6 2.6 228 5] = - Index Tal If ae 100 *49.3 45.8 7 ASHO 950510) 9.3303 42.8 = = _G. gorilla G. gorilla P. troglodytes (b) males females (n=20) (n=20) (n=20) x Die 7, 10.8 - ee G54Ch | 10,212. 2 OnG-1159 - ‘ x 13.4 1255 ~ BuL . . Ee Q95%CL 12.0-14.7 10. 3-14. 7 A xe S72 86.6 - Index 1/B x 100 OR 79. 7-94. 8 70. 8-100.9 = x - - S704 Index Tal L/friL x 100 OR = 54 29. 3-49. 0 62 Table 20 UMP (e) 62-13 MDi L 10.8 Tri B 9.0 Tal B 9.2 Index L/B x 100 117.4 Index Tal B/ neh swany ese (b) X MDE 95%CL x pil 95%CL X sais 95%CL Index 1/B x 100 x OR Index Tal B/ x Tri B x 100 OR Table 21 (a) MDi L Tri B Tal B Index I/B x 100 Index Tal B/Tri B x 100 (b) MDiL Tri B Tal B Index L/B x 100 Index Tal B/Tri B x 100 UMP 62-14 12.0 10.2 10.5 114.2 103.0 UMP 62-10 esol tS *10.8 eal (0) 107.2 101.9 NMK 190.4 13.9 13.1 12.5 106.1 95.5 xX 95%CL X 95%CL X 95%CL x OR x OR PEABODY MUSEUM BULLETIN 31 Comparative measurements of M, in (a) D. (P.) major, (b) G. gorilla UMP UMP BM(NH) NMK _ BM(NH) 62-15 62-16 M16648 190,1 M14086 WG 2 IO) 7/ WA ak 11.8 *10.9 10.6 *9.4 LOZ 10.9 Gls} 10.9 9.4 TON, 10.8 S55) LVGo) “A LTaS9 ILS} (0) 108.2 114.8 102.8 *100.0 104.9 99R W025 2 G. gorilla G. gorilla males females (n=20) (n=20) IS 7 14.9 14.4- 17.0 13 3'6=) L6n2 13.4 WAS (5) 12.3- 14.6 11.0- 14.3 13.4 WAS 7/ 12.6- 14.2 PIiN2— 142 TANG 53 Hs 7/ 110. 0-122.4 108. 1-124.0 99/55 100.3 95. 1-102.3 95.3-108.1 BM (NH) M16648 UMP 66-01 14,8 12.8 *12.8 115.5 *100.0 *13.0 LAMA 2 SAL) 106.5 97.5 G. gorilla males (n=20) 753 ily UE) 8) 15.4 14, 0- 17.0 5 (0 key 5= Gis 6 AG CA 104. 8-119.8 97.6 92.3-111.0 NMK 271,55 15.6 14,1 13.5 110.6 95. 7 Comparative measurements of Mg in: (a) D. (P.) major, (b)_G. gorilla BM (NH) M14086 *12.6 10.8 10.2 IIS 7/ 94.5 G. gorilla females (n=20) 16.2 145 wader 14,4 12.0- 16.8 14.0 11.8- 16.1 alg &) 103, 0-125.4 97.8 89. 6-103.3 TERTIARY PONGIDAE OF EAST AFRICA 63 Table 22 Comparative measurements of M, in: (a) D. (P.) major, (b)_G. gorilla BM (NH) NMK UMP BM (NH) M16648 190,1 66-01 M14086 MDi L 18.0 17.3 15.0 *13.5 Tri B 13.4 14,2 12.5 «10.5 Tal B 12.0 12.0 «10.4 2 index L/B x 100 134.3 121.9 120.0 128, 6 Index Tal B/Tri B x 100 89.5 84.5 86.7 G. gorilla G. gorilla (b) males females (n=20) (n=20) X 1726 16.0 MDit 95%CL 15,52 19.7 13.7- 18.4 . x 15.3 14,2 eke 95%CL ‘Ee ail Nleee aes x 13.6 DQe7, Tal 95%CL 12. 0- 15.2 10,1- 15.3 X 141 112-9 Index 1/B x 100 OR 108, 0-122.8 105. 8-123. 3 x 89.0 89.2 Index Tal B/Tri B x 100 OR 79.8- 99.3 82.5- 95.7 with the roots of the left canine, incisors, right premolars, and first molar, to which this canine crown belonged, was recovered in 1947 and the tip of the right canine in 1948. The specimen was first described by Clark and Leakey in 1951 (p. 59-60). In 1962 the horizontal ramus from the left side of the mandible was recovered and fitted onto the earlier symphyseal piece. The more complete specimen is figured in Leakey (1967, fig. 4e). The left dentition from P, to M; is very well preserved and is almost unworn. Comparison with the type of D. (P.) major shows that this is also a member of that species. The size of the dentition, particularly of the canine, indicates that NMK 190,1 is most prob- ably a male. Morphologically the canine crown resembles male G. gorilla, al- though its long axis is more vertical than in the living ape. Estimates of tooth size have, where necessary, been made from the roots preserved. The first premolar was probably a little smaller than that of the type specimen. Measurements are included in Table 18. The second premolar is also similar to that of the type (measurements in Table 19). Measurements for Mj, M,, and Ms; are given in Tables 20, 21, 22, and 23. Metrically as well as morpho- logically, there is no reason to assume that the specimens listed in these tables form more than a single species. The overall proportions of the mandible have been estimated by reconstruct- ing both tooth rows. A cast of BM(NH) M16648 was used to build up the right mandibular ramus and teeth. Great care was taken to make the rami and tooth rows symmetrical. Measurements are given in Table 24. Considering external measurements, the mandibular tooth row does diverge posteriorly, as it does in male gorillas. The specimen from Songhor is very similar 64 PEABODY MUSEUM BULLETIN 31 Table 23 Comparative measurements of C, in D, (P.) major and Gorilla NMK UMP G. gorilla 190.1 66-01 male female (n=20) (n=18) 15.1 **16, 0 x 18.1 130i Max L 95%CL 15,2-21.0 11, 1-14.9 12.5 **12.0 me 14,1 10.3 Trans B 95%CL 12, 2-16.0 8.9-11.7 Table 24 Measurements of D. (P.) major and_G. g orilla mandibular dentitions G, gorilla. NMK UMP ~~ males 190,1 62-10/66-01 (n=20) CIDP **23.0 X 28.4 aes 95%CL 24, 8-32. 1 **46.0 **49,5 me 56.3 eae 95%CL 47. 4-65. 2 **51.0 **54.5 x 61.1 Sie 95%CL 56. 6-66. 6 rs **58.5 **60.5 Xx 67.1 ae 95%CL 59, 4-75. 8 in these dimensions to the reconstructed mandible from Moroto II (men- tioned in Chap. IV, p. 45). The general size and proportions of the Moroto mandible were determined as follows. A reconstructed mandible of combining BM(NH) M16648 and NMK 190,1 was sawn in half at the symphysis and the half mandibles were oriented beneath the maxilla from Moroto II (UMP 62-11) until their correct relative occlusal positions were obtained. The small gap which separated the two rami at the symphysis was then filled with plaster and a new mandible cast. This new reconstruction provides estimates of the overall dimensions of the mandible associated with UMP 62-11. As can be seen from Table 24, in these dimensions the Moroto mandible is broader by 3.5 mm to 4.0 mm than the combined one from Songhor. Compared with the sort of size variation found in the sample of living male gorillas, the differences are rela- tively unimportant. Although the tooth rows of the reconstructed mandible NMK 190,1 diverge posteriorly when external dimensions are considered, this is not the case for internal dimension. Internal measurements are: inside P,, **32.0 mm, inside M,, **32.0 mm, inside Mo, **31.0 mm, and inside Ms, **31.0 mm. In Chapter IV, page 45, the problem of divergence and convergence of tooth rows was discussed and it was concluded that the most meaningful way of ap- proaching the problem was from a functional point of view. Protocones of the upper premolars and molars occlude with talonid basins in the lower premolars and molars. ‘The distance between the tips of M, protocones is therefore equal to the distance between the talonid basins of the M,’s. The distances between the talonid basins of pairs of cheek teeth of NMK 190,1 were estimated first from the composite mandible and second from photo- graphs. A photograph of NMK 190,1 was taken and a reversed print of the mandibular tooth row prepared. This reversed print was then oriented on the TERTIARY PONGIDAE OF EAST AFRICA 65 original in the following way to produce, by way of a mirror image, a composite photograph. A tangent AB was drawn to the most distal points of the two canines. This line was normal to the anteroposterior midline axis of the symphysis. A further tangent CD was drawn to the most lingual points of left P, and left M3. The angle between AB and CD was measured. The right tooth row could then be oriented so that the tangent C’D’, drawn to the most lingual points of right P, and right M3, subtended the same angle with AB. Distances between basins were then measured from the photograph. Measurements between the talonid basins of NMK 190,1 were obtained and are shown compared to the interprotocone di- mensions of the Moroto palate, UMP 62-11, in Table 25. Table 25 Measurements on upper and lower dentitions of D. (P. )major NMK 190,1 UMP 62-11 Photograph Cast (interprotocones) Across P, **4059 **39.3 *41.9 Across M, **40.3 **38.7 **41,2 Across Mz *%39),16 **38.0 *40.0 Across M, 2319/10 **38.0 *40.4 There is a slight convergence posteriorly in both upper and lower dentitions. Although the Moroto II specimen is 2 mm or so larger in each measurement, the relative proportions of the two are remarkably similar. (These characters, like many others, show a wide range of variation in living gorillas.) Once again we have confirmatory evidence that we are dealing with individuals sampled from a single species, D. (P.) major. Leakey has stated that he does not regard the specimen from Moroto Il (UMP 62-11) as D. (P.) major. His views are summarized as follows. First, “in not just one but many specimens of four species, Proconsul has a relatively small canine with a molar-premolar series converging forward, as opposed to the backward convergence in pongids, including Bill Bishop’s . . . fossil pongid from Moroto —a contemporary of Proconsul’” (Leakey, 1965, p. 13-14). In addition, it has been “claimed that it should be possible to fit together a Proconsul major mandi- ble with the maxilla found by Bill Bishop and attributed to Proconsul major. I have tried it with this backward converging maxilla, and I simply do not see how they could be made to articulate normally and with proper tooth contact” (Leakey, 1965, p. 106). The mandibular specimen of D. (P.) major to which Leakey refers is NMK 190,1. As noted above, the tooth rows do diverge posteriorly if the external di- mensions are used, and this is also the case for male gorillas (and for other living pongids—see Appendix 1). This is not the case, however, for internal dimen- sions, nor for the interbasin measurements. The internal contours of the mandi- ble converge anteriorly, as do the external contours, and a superficial examina- tion of the mandibular rami certainly makes it seem as though the mandible is convergent anteriorly. This feature is also to be seen in male gorillas. It seems clear, however, that Leakey is talking not about mandibular rami but about tooth rows. The tooth rows of NMK 190,1 and UMP 62-11 are very slightly convergent posteriorly using the interprotocone and intertalonid basin criteria. 66 PEABODY MUSEUM BULLETIN 31 They will not articulate with each other because the bicanine diameter of NMK 190,1 is a little too small to encompass the lateral incisors of UMP 62-11. This is hardly surprising since they come from different individuals from sites separated geographically and possibly temporally, too. What is perhaps more surprising is their similarity. The mandibular symphysis of NMK 190,1 is of some interest. ‘This was de- scribed by Clark and Leakey (1951, p. 60) and figured by them (p. 59). A longitudinal cross section of this symphysis is figured in Leakey (1967, p. 160, fig. 5a) and is also shown in Figure 3. The symphysis is approximately triangular in cross section, with a long planum alveolare gently sloping to the superior trans- verse torus, the most posterior point of which is opposite Py. The internal contour then slopes inferiorly and anteriorly to the inferior transverse torus. The anterior border of the symphysis is rather straight. The length of this an- terior border from infradentale to gnathion is approximately 52 mm and the symphyseal thickness normal to this dimension approximately 28 mm. It should be emphasized that this is the only complete symphysis of a male D. (P.) major known. The other less complete mandibular remains referred to this species share certain features with this specimen. Thus UMP 62-10/66-01 and BM(NH) M16648 show the thickening of the internal aspect of the mandible anteriorly and the gradual shallowing of the superior internal slope; in addition they have long, gently sloping plana. The infant mandibles from Napak also have these features and show that inferiorly the internal contour of the mandible slopes anteriorly and does not project backwards to form a simian shelf. This symphyseal morphology contrasts with that seen in another African species of Dryopithecus, D. (P.) africanus (Fig. 3). Here the planum slopes Fig.3 Mid-symphyseal sections of living and fossil Pongidae (not to scale) ON (PR) major D. fontoni D.(P) africanus N.M.K. I90,| B.M.(N.H)I4086 MNHN. 1872-2 = ee PF troglodytes DP 53.03 G. gorilla D. Pr. 52.0.3 B.M.(N.H.) 933 TERTIARY PONGIDAE OF EAST AFRICA 67 more steeply to the superior transverse torus. Although there is no posterior projection in the inferior symphyseal region, the inferior border is in some cases barely anterior to the superior transverse torus. The absence of a long, sloping planum is coupled with a relatively gracile internal mandibular contour, with- out marked lingual thickening. ‘The two species are similar in having no simian shelf, and therefore a roughly triangular symphyseal cross section; they differ in the relative proportions of the triangle (this is shown clearly in Leakey, 1967, figs. 5a, 5b). If we consider the superior two-thirds of the symphyseal area of D. (P.) major, it is of some interest that the contour of this region is very similar to that of many G. gorilla. It should be emphasized here that, as in all other features, gorillas are variable in symphyseal structure (Vogel, 1961; Schultz, 1963). In gorillas, there is generally a long planum alveolare, sloping gently to a superior transverse torus, the most posterior part of which is opposite P,. The similarities of this and the anterior symphyseal contour to D. (P.) major are striking. Inferiorly, however, the gorilla mandible usually projects posteriorly as a simian shelf or basal plate, although this feature is quite variable in expression (Vogel, 1961; Schultz, 1963). The development of a simian shelf in G. gorilla has also pro- duced the inferior rounding of the anterior symphyseal contour. Both anterior rounding and shelf are absent in D. (P.) major. It has been suggested by a num- ber of workers that the simian shelf acts as a buttress to strengthen the anterior mandibular region, particularly in recent pongids where the anterior dentition is relatively broad (see, for example, Scott, 1963). Table 26 contains values of the ratio, (I, B/P;-M; L) x 100 for living Table 26 Ratio (BI,/LP, —M, x 100 inG, gorilla and P. troglodytes G. gorilla P. troglodytes Males Females Males Females (n=20) (n=20) (n=14) (n=12) x 36.8 S702 57.8 58.1 OR 34. 6-40. 8 33. 0-42. 0 53. 4-61. 8 54. 0-62.3 chimpanzees and gorillas. This ratio has been calculated in an attempt to relate the breadth of the incisor region to the size of the cheek teeth and hence to chewing stresses. In gorillas the ratio is low mainly perhaps because the length of the tooth row is relatively great. In chimpanzees with a much shorter tooth row, the incisor region is as broad as in gorillas (see Table 27). Presumably the incisors are of similar importance in the two species, or perhaps of greater impor- tance to the frugivorous chimpanzees; grinding and crushing are more important to the herbivorous gorilla, hence the great size of the cheek teeth. The value of this ratio in D. (P.) major is: UMP 62-10/66-01, 35.9, and NMK 190,1, 33.6. The incisor region is, relatively speaking, no smaller than in gorillas. If the simian shelf in gorillas is a phylogenetic response to anterior chewing stresses, presumably these stresses act as selection pressures producing simian shelves only when the absolute values of the dimensions reach certain critical thresholds. The gorilla symphysis is buttressed not only by the simian shelf, but also by the superior transverse torus (which is very much smaller in chimpanzees). The symphysis of the type observed in D. (P.) major could plausibly have 68 PEABODY MUSEUM BULLETIN 31 Table 27 Measurements of mandibular tooth rows of G. gorilla and P. troglodytes G. gorilla G. gorilla P. troglodytes P. troglodytes males females males females (n=20) (n=14) (n=12) x 28.4 26.4 29.0 28.9 978 (n=20) : 95%CL 24.8-32.1 22.5-30.3 26. 3-31.8 25. 7-32. 1 X 56.3 44,7 47.6 45.1 C, B (n=20) 95%CL 47.4-65.2 37.9-51.5 43. 1-52.0 39. 6-50.5 x 61.1 B22 51.4 50.9 RB (n=19) 95%CL 56. 6-66.6 42. 2-62.2 45.9-56.9 44, 6-57.2 xX 67.1 62.0 54.6 55.1 M, B (n=19) 95%CL 59.4-75.8 54.5-68.5 49.3-60.0 50. 0-60.3 x Wied 71.0 50.3 49.7 PME (n=20) 95%CL 70.1-84.7 64. 0-78.0 44,5-56.1 47.0-52.5 evolved into the gorilla type of symphyseal morphology. Until we understand fully the functional anatomy of the symphysis it may well be impossible to evalu- ate the importance of the simian shelf. It has probably evolved separately in all three living pongids, presumably as some sort of phylogenetic response to chew- ing stresses. Too great an emphasis should not be placed on its degree of de- velopment, particularly in phylogenetic and taxonomic studies. Table 28 shows a further ratio calculated in an attempt to interrelate jaw size, via cheek tooth length, to mandibular breadth; the ratio is (C; B/P;—-M3 L) x 100. Once again, the relatively short cheek tooth length in chimpanzees is to be noted. Male gorillas have relatively broader bicanine breadths than females and therefore have somewhat higher ratios. The values of this ratio for D. (P.) major are: UMP 62-10/66-01, 77.3, and NMK 190,1, 73.0. Enough is preserved of the horizontal ramus from Moroto II for comparison with NMK 190,1 to be useful. The external surface of the ramus is virtually identical in size and morphology in the two individuals. Canine and premolar juga are present in both, making the anterior part of the body convex. Posterior to this is a concave region below P, and M,, before the body becomes thick and convex again opposite M, and Ms. Table 28 Ratio (BC./L P,—M,) x 100 inG, gorilla and P. troglodytes G. gorilla _P. troglodytes Males Females Males Females (n=20) (n=18) (n=13) (n=12) LOST 6205, 93.6 90.6 O I ee) 64. 6-77. 8 57. 3-68.5 85. 0-100. 8 83. 6-98.2 TERTIARY PONGIDAE OF EAST AFRICA 69 In summary, as far as dental and mandibular morphology, dimensions, and proportions of these relatively complete D. (P.) major mandibles from Uganda and Kenya are concerned, all three fall within a range of variation to be expected in a sample drawn from a single species. It is probable, too, that all three speci- mens are male. It is particularly important to note that the palate from Moroto II and the most complete mandible from Songhor (NMK _ 190,1) do fall in all probability within the same species. Further specimens of D. (P.) major from Songhor were described by Clark and Leakey (1951, p. 60-61) and Clark (1952, p. 279-281). As noted, some of these have been removed by Leakey (1967). ‘Those that are, in my opinion, def- initely D. (P.) major are briefly noted here and their measurements given. NMK 271,55. The specimen is a left lower second molar, described and figured by Clark and Leakey (1951). They noted that this tooth, an unerupted, weath- ered crown, was morphologically similar to the M,’s of D. (P.) nyanzae, al- though they recognized it as a D. (P.) major. It is very similar to the M,’s of D. (P.) major specimens BM(NH) M16648 and NMK 190,1, although it is a little larger. Measurements are included in Table 21. NMK 272,56. The specimen is a left lower second premolar, described and figured by Clark and Leakey (1951). Morphologically this tooth is similar to the others from Songhor and Napak. Measurements are included in Table 19. NMK 382,3 and 383,4. The specimens are a left and a right first lower premolar, described and figured (NMK 383,4) in Clark (1952). They are the crowns and roots of two first premolars, probably from the same individual. The left tooth is broken mesially. The right crown is well preserved. Morphologically, the crowns are similar to those of G. gorilla. Although said to contrast morphologically with the Ps;’s of D. (P.) nyanzae, these teeth are similar to the D. (P.) nyanzae from Koru [BM(NH) M14086] described below. Measurements are given in Table 18. NMK 604,69. The specimen is the weathered and worn crown of a D. (P.) major first premolar. Measurements are included in ‘lable 18. NMK 102, CMH35. The specimen is a right upper second molar, described and figured by Clark and Leakey (1951). This tooth, tentatively identified as a sec- ond molar, was only provisionally assigned by Clark and Leakey to D. (P.) major because of its large size and its similarity to teeth of other species of that sub- genus. Measurements of this specimen are included in Table 16, p. 51. The crown morphology is similar to molars of D. (P.) africanus and nyanzae in hav- ing a well-developed cingulum. This cingulum is beaded and the lingual surfaces of protocone and hypocone are wrinkled. A distinct protoconule is also present mesial and buccal to the protocone. In all these features, NMK 102, CMH35 resembles the upper molars from Moroto II and Napak referred here to D. (P.) major. NMK 405,381 (Fig. 21). The specimen is a left upper second molar, described and figured in Clark (1952). The morphology of the crown is very similar indeed to that of UMP 62-08 from Napak (see Fig. 21). There can be little doubt that 70 PEABODY MUSEUM BULLETIN 31 this is a second molar of the species to which UMP 62-08 and 62-11 belong. Mea- surements are included in Table 16, p. 51. Clark and Leakey (1951) and Clark (1952) had at their disposal only two upper teeth which might possibly be referred to D. (P.) major. Since these molars resembled the mandibular teeth of D. (P.) major in much the same ways that the upper molars of D. (P.) nyanzae and D. (P.) africanus resembled their respective mandibular teeth, a tentative assignment to D. (P.) major was made. At Moroto II, Napak and Songhor, upper molars of this type are found with mandibles referable to D. (P.) major. As noted already, the Moroto II maxilla has the dental morphology and proportions to be expected of a D. (P.) major palate. The M? from Napak V, UMP 62-08, occludes well with NMK 190,1 from Songhor (see Fig. 24). It is perhaps a little too big and would probably occlude best with the slightly larger BM(NH) M16648. However, it is not possible to test this as UMP 62-08 is from the left side and BM(NH) M16648 from the right. Therefore, unless we are dealing, at three separate localities, with the upper dentition of one large pongid and the lower dentition of another large pongid, both species having similar proportions and morphologies, it seems best to assume that we are dealing with a single, variable species. As Bishop has pointed out (1964, p. 1331), it would be illogical to do otherwise, particularly since the new evidence accumulated since 1952 strengthens rather than weakens the hypothesis that only one species is involved. So far in this review, no mention has been made of unequivocal D. (P.) major material from the deposits at Rusinga. In 1952, Clark described the only well- identified specimen, NMK 129,257, from site R3 on Rusinga Island. NMK 129,257. The specimen, a left deciduous second premolar from Rusinga R3, lower Hiwegi Series, was described by Clark (1952). Clark assigned this tooth to D. (P.) major because it was clearly that of a pongid, and because of its large size. Measurements are given in Table 29. It was said by Clark (1952, p. 280) to be “intermediate in general dimensions between chimpanzee and the gorilla.” ‘This statement is not correct, Table 29 shows the specimen falling within Table 29 Comparative measurements of dP, in: (a)D. (P. )major, (b)G. gorilla and P troglodytes NMK NMK UMP (a) 129,257 277 58'48(S) (62-13) (Napak) MDi L 10.3 9.7 *9.8 Max B 8.1 Ted *7.8 G. gorilla P. troglodytes b) (n=12) (n=17) 1 x 13.6 9.0 ae 95%CL 11, 4-15.8 7.3-10.7 1 X 10.3 7.6 oe 95%CL 7.9-12.7 6.0- 9.2 TComparative data from Ashton and Zuckerman (1950) TERTIARY PONGIDAE OF EAST AFRICA 71 the chimpanzee confidence limits. The estimated crown dimensions of dP, of UMP 62-13 from Napak are only a little smaller than those of the Rusinga speci- men. DISCUSSION OF THE LARGE SPECIMENS The material from Songhor and Moroto II, and presumably some from Napak, probably represents male specimens—the large size of the canines, for example, suggests this. It also seems a reasonable hypothesis that D. (P.) major is ancestral to or close to the ancestry of the living species G. gorilla. It is well known that this living species shows considerable sexual dimorphism in dental and mandibular size, particularly canine size (see Appendix 1 and Figs. 4-6). This is in contrast to Pan troglodytes where, in the dental and mandibular di- mensions that we have been discussing, there is little size dimorphism. One is tempted, therefore, to ask where the females of D. (P.) major are, for one should expect to find them; one would expect them to be morphologically similar to, though smaller than, the males of D. (P.) major. Some of the smaller specimens from Napak are probably females. (It should be cautioned here, however, that smallness is not invariably to be equated with “‘femaleness,” as a glance through Appendix 1 clearly shows.) Clark and Leakey published the diagnosis of D. (P.) major in 1950. The diagnosis included the following sentence (1950, p. 261). “The lower teeth re- semble those of P. nyanzae in their relative proportions and cusp pattern, but are of a much larger size.’ As noted above, large upper teeth were assigned to this taxon because they also resembled morphologically those of D. (P.) nyanzae. The specimens assigned to D. (P.) major in 1951 and 1952 are therefore prob- ably almost entirely those of males. It has been suggested already that female D. (P.) major are present at Napak. To discuss other probable females we must return to the two Kenya sites, Koru and Songhor. SMALL SPECIMENS BM(NH) M14086 (Fig. 23). The specimen, from Koru, has been described and figured by Hopwood (1933b). It consists of right and left mandibular rami, broken at the symphysis on the right and just lateral to it on the left. The left side contains the root of C, and crowns of P; to Ms. The roots of Mg are pre- served. The mandible has been dissected internally to expose the tooth roots. On the right side the same root and crowns are preserved, together with the buccal three-fourths of the mesial moiety of the M3; crown. This individual was referred originally by Hopwood to the species Proconsul africanus and later transferred to Proconsul nyanzae [D. (P.) nyanzae| by Clark and Leakey (1951, p. 39). The mandible and dentition were described in detail by Hopwood (1933b) and only a few details will be noted here. The symphysis is thick, with a triangular cross section and a long, gently sloping planum alveolare. The ramus thickens anteriorly, so that the internal contours converge. In all features preserved, the rami and symphyses of BM(NH) M14086 and NMK 190,1 are identical (see Fig. 3). Comparative measurements are given in Table 30. he. PEABODY MUSEUM BULLETIN 31 Table 30 Measurements of D. (P.) major mandibles NMK BM(NH) NMK ibS)(0) 75a M14086 198,28 Symphyseal H from infradentale to gnathion *52.0 *42.0 *41.0 Symphyseal T *23.0 *23.0 2305 Rameal D at P, 41.0 36.5 *33.0 Rameal T at P, 21.0 17.0 = Rameal D at M, 34.5 *32.0 - Rameal T at M3 Zoe *20.0 20.5 If D. (P.) major is ancestral to G. gorilla and is assumed to be dimorphic like the gorilla, and if NMK 190,1 is accepted as a male, a reasonable hypothesis is that BM(NH) M14086 is a female of the same species. Measurements of the edentulous mandible NMK_ 198,28 from Songhor, described and figured by Clark and Leakey (1951, p. 48 and pl. IV, fig. 21) are also included in the above table. The symphysis has the same morphology as BM(NH) M14086. The size and proportions of the cheek teeth, as judged from their roots, were probably similar to those of BM(NH) M14086, and this specimen, too, can be regarded as a female D. (P) major. It is possible to estimate some measurements from the tooth rows of these three specimens (see Table 31). If reference is made to Table 27, it will be seen that these differences are no greater than might be expected within even a moderately dimorphic species. The dentition of BM(NH) M14086 is not particularly well preserved. Both canines are broken at the alveolar margin. Their lengths and breadths are approximately 13.5 mm and 9.5 mm, expectable dimensions if this is a female. Both first premolar crowns are in good condition. The mesial border slopes straight back to the main cusp. There is a mesial wear facet worn by the upper canine and a groove lingual to this; the groove has a small cingulum at its mesial-lingual extremity. A moderate talonid is developed distally. The buccal surface is moderately convex in its upper third, then rather flattened. Morpho- logically, the P3’s are similar to NMK 382,3 and 383,4 from Songhor discussed above. Length and breadth are 12.7 mm and 7.5 mm. The rest of the dentition is worn and eroded and almost all occlusal details have been lost. It has been possible to estimate with some accuracy the lengths and breadths of right P, and M, and left M,. Since these differ from the values given by Hopwood (1933b, p. 447), they are listed here: right P,, MDi L, *7.8 mm, Bul B, *9.0 mm; left M,, MDi L, *10.9 mm, Tri B, 9.3 mm, Tal B, 9.5 mm; right Mz, MDi Table 31 Measurements of D, (P.) major mandibular dentitions NMK BM (NH) NMK 190i, 1 M14086 198,28 Tg op B cea 2 **20.0 = Canae **46.0 2x42 30 = Pe iB canta b 0) **48.0 = P,-M, L *63.0 *54.0 *52.0 TERTIARY PONGIDAE OF EAST AFRICA 75 L, *12.6 mm, Tri B, 10.8 mm, Tal B, 10.2 mm; and, left M;, MDi L, *13.5 mm, Tri B, *10.5 mm. The P, and M, of UMP 62-16, the fragment assigned originally to Sitvapithecus africanus by Leakey (in Bishop, 1962), is very similar to BM(NH) M14086 in dimensions and, as far as can be seen in M14086, morphology (Appendix 2 and Fig. 23). It is probable that UMP 62-13, 62-16, and 66-02 from Napak in Uganda are also female specimens. Three more small specimens are probably smaller individuals of D. (P.) major. NMK 101, CMH34. A left upper second molar from Songhor, the specimen has been described by Clark and Leakey (1951). The crown is moderately worn. Morphologically, the occlusal surface resembles other M*’s in this species. Meas- urements are included in Table 15, p. 51. The metacone is somewhat reduced in size. The lingual cingulum runs from protoconule to the distal accessory cuspule and the buccal cingulum is rather rudimentary. NMK 198,28. The specimen, from Songhor, has been described and figured by Clark and Leakey (1951). This edentulous left horizontal ramus has already been mentioned during the discussion of BM(NH) M14086 from Koru. NMK 277,58. The specimen is a left deciduous lower second molar from Songhor. The morphology of this specimen is similar to that of NMK 129,257 from Rusinga, although the cusps are a little less massive and the length and breadth a little smaller. Measurements are included in ‘Table 29. It should be borne in mind, however, that future finds may reveal that some or all of these individuals belong in a taxon separate from D. (P.) major. The evidence at present available makes the hypothesis outlined here rather more likely. Since much of the medium-sized material from Songhor and Koru has either been transferred by Leakey (1967) out of D. (Proconsul) to his proposed species Kenyapithecus africanus (transfers with which I agree), or reassigned by me to D. (P.) major, there is a clearly marked metrical gap between D. (P.) africanus and D. (P.) major at Songhor and Koru (see Appendix 2, Figs. 4-6 and 14-17). Two lower molars from Songhor, NMK 389,11 and NMK _ 390,12, described by Clark (1952, p. 277) as second molars, apparently bridge this size gap (Table 32). The difficulty of differentiating small M,’s and large M,’s has already been noted. There are several possible explanations for these two teeth. 1) They are both small M,’s of D. (P.) major. 2) They are M,’s of D. (P.) major. 3) They are Table 32. Lower molars of uncertain status NMK 389,11 NMK 390,12 MDi L 10.8 Ub 7 Tri B 10.3 10.5 Tal B 10.2 LOE 7 Index L/B x 100 104.9 109.3 Index Tal/Tri x 100 99.0 101.9 74 PEABODY MUSEUM BULLETIN 31 M,’s of D. (P.) nyanzae. 4) They are lower molars of some other hominoid spe- cies. I hesitate to choose between these alternatives, although the last is perhaps least likely. DISCUSSION OF THE D. (P.) MAJOR MATERIAL FROM EAST AFRICA The methods involved in Principal Coordinates Analysis have already been outlined in Chapter II. Briefly, Principal Coordinates Analysis enables one to take a swarm of points in multidimensional space (each point representing p standardized measurements on an individual) and reduce the number of dimensions to a much smaller number, two or three, thus enabling the points to be visualized and plotted on one or two graphs. During the dimensional reduction, the interindividual distances in n-1 space are preserved with the minimum amount of distortion. Analysis was performed first on material of known species and sex. Data on sets of male and female chimpanzees and gorillas were analyzed. Using 28 mandibular dimensions, the two species are clearly separated; so, too, are male and female gorillas, although male and female chimpanzees are not (Fig. 4). In these dimensions, therefore, gorillas show a great deal of sexual dimorphism while chimpanzees do not. When the most complete specimens of D. (Proconsul) from Moroto, Songhor, and Koru are analyzed with the chimpanzees and gorillas Fig.4 Principal Coordinates Analysis of Fan froglodyfes and Gorilla gorilla all Ee Pon tradodytes (femates) r Fan troglodytes (males) De® + Gorilla g. gorilla (females) © Gonllg g. gorilla (males) E Gorilla g. beringei (mad@ TERTIARY PONGIDAE OF EAST AFRICA 75 Fig.5 Principal Coordinates Analysis of Pan troglodytes, Gorilia gorilla, and Oryopithecus major. Pon troglodytes (females) Fan troglodytes (males) Garilla g. gorilla females) Gorilla g.gorila (males) Gorilla g. beringei (male) Dryopithecus major aE + 6 x F Ope t (Fig. 5), it can be seen that they lie between the two living species and are a little dispersed (see Figs. 14-17). They could, of course, represent two species, but since the larger specimens are probably males, it might be expected that the females of this species would fall on the graph approximately where the Koru specimen is plotted. In order to test this hypothesis further it would clearly be desirable if another species of pongid, smaller than G. gorilla but also sexually dimorphic in these features, could be included in the analysis. Pongo pygmaeus, the orangutan, is a species which fulfills these requirements. A small sample is included in the analysis shown in Figure 6. The two sexes can clearly be separated. The material from Moroto, Napak, Songhor, and Koru assigned here to D. (P.) major could therefore belong to one species. As noted above, this conclusion is strengthened by the great morphological similarities within the sample of fossils. Before fossil material can be classified in one species, two basic questions must be answered. First, can the individuals be regarded as drawn from a single, evolving lineage? Second, starting at the time level of the oldest known fossils, where in the evolving continuum should the (arbitrary) boundary between this species and the next youngest be drawn? The first question has been answered already, in that all the material discussed so far appears to be sampled from a single lineage. The possibility remains, however, that material 76 PEABODY MUSEUM BULLETIN 31 Fig.6 Principal Coordinates Analysis of Aun froglodyies, Gorilla gorilla and Pongo pygmaeus. Ban troglodytes (females) Pan trogodytes (males) Gorilla g. gorilla (females) Garilig g. gorilla (males) fengo pygmaeus (females) Pengo pygmaeus (maes) Gorilla g. beringel (male) wo tOOO DP O+ § from Moroto II may be younger than that from the other sites, and may show some evolutionary advances over that from Napak, Koru and Songhor. This possibility has been rendered considerably less likely, however, by the recent discovery at Napak of P, and M, of D. (P.) major (the specimens are as yet un- described). Both specimens are almost identical to homologous teeth in the Moroto palate. If this species is ancestral to the gorilla, the palate from Moroto II can give us information about the sorts of changes which have occurred in this lineage since the Miocene. If we then indulge in a little more speculation, it is possible to extrapolate back through time and make hypotheses about the as yet un- known and undiscovered parts of earlier individuals in this, and preceding, species. One might conclude that the anterior dentition and supporting structures of the face would be smaller in pre-Early Miocene forms. The absolute size of premolars, particularly upper ones, would have been less, and so too would the lengths of some of the posterior cheek teeth, particularly M, and Mg. In conclusion, the material discussed in this and the preceding chapters can safely be assigned to a single species D. (P.) major which is probably ancestral to, or close to the ancestry of, the gorilla and which is probably, like the gorilla, sexually dimorphic as far as tooth and jaw size are concerned. This species is found at the Ugandan sites, and at Koru and Songhor in Kenya. Only one specimen of a tooth has been found at Rusinga, a left deciduous second TERTIARY PONGIDAE OF EAST AFRICA Ig) premolar, NMK 129,257, from site R3. The majority of the dental remains therefore have been collected at deposits for which an elevated, forested en- vironment has been inferred. If D. (P.) major is a species confined mainly to such a habitat, or even to lowland forest, this might account at least in part for its under-representation at Rusinga. The presence at Rusinga, Songhor, and perhaps other sites, too, of another medium-sized species, called by Leakey (1967) Kenyapithecus africanus, de- mands that we exercise a certain amount of caution in the consideration of the postcranial material from these sites. However, some skeletal remains could represent D. (P.) major. Walker (personal communication) reports that two distal femoral fragments from Napak probably belong in D. (P.) major. They are approximately chimpanzee-sized. No other species of large hominoids have been described from the Napak sites. These femoral specimens resemble one described by Clark and Leakey (1951, p. 97) from site R1 on Rusinga Island (NMK 89, R1932). These fragments, if they do indeed represent D. (P.) major, contrast with Pan and Gorilla in that they have shorter femoral necks than the living forms; this feature, together with the disposition of the trochanters, suggests perhaps some leaping locomotor behavior patterns. A talus (NMK 111, CMH145) and associated calcaneum (NMK_ 112, CMH146) from Songhor, matching in size large adult male chimpanzees, prob- ably represent D. (P.) major, too. The talus has been compared to living man on the one hand, and Pan and Gorilla on the other (Day, personal com- munication), using the multivariate statistical technique of Canonical Analysis. The Songhor talus matches that of the gorilla. Walker (personal communication) believes, however, that functionally this talus may also be similar to colobines, a conclusion which would tie in with the femoral evidence. The Moroto vertebrae (p. 43) reinforce the dental, facial, and gnathic evidence indicating that we are dealing here with a probable gorilla ancestor, and the distribution of this species in sites which sample a forested upland habitat like that of the living mountain gorilla also supports this view. How- ever, the total postcranial evidence indicates that D. (P.) major may well have been a more active and less terrestrial form than the living great ape. There is nothing to suggest that D. (P.) major was a “brachiating” animal, like Pongo for example. It might have been a knuckle-walker. However, arm-swinging quadrupedal behavior as in Ateles or Lagothrix is a distinct possibility. CHAPTER: Vile DRY OPIGIEIEC USe(PROCON Sin) AFRIGANUS AND DRYOPITHECUS (PROCONSUL) NYANZAE BACKGROUND OF THE MATERIAL Hominoid fossils were first recovered from Rusinga Island and Songhor in 1931 and 1932 by L. S. B. Leakey and D. MacInnes. In 1933 Hopwood described a new ape, Proconsul africanus, from the Miocene of Koru. The mandible described by him has been transferred here to D. (P.) major (p. 71). Collecting has continued steadily since then, due principally to the magnificent efforts of Dr. and Mrs. Leakey, and now several hundred hominoid specimens are preserved in the National Museum Centre for Prehistory and Palaeontology in Nairobi. A large number of these are pongids. MacInnes (1943) described many of the Rusinga pongids as Proconsul africanus. In 1950 Clark and Leakey published diagnoses of two further species of this genus, P. nyanzae and P. major, together with a new species of Sivapithecus, S. africanus. This they followed in 1951 with a detailed monograph. Clark described additional speci- mens in 1952, and more recently Napier and Davis (1959) have discussed parts of the postcranial skeleton of P. africanus. Simons and I (1965) transferred the species of Proconsul to Dryopithecus, retaining them in a subgenus, (Proconsul). No attempt will be made to describe in detail individual specimens of D. (P.) africanus and D. (P.) nyanzae in this chapter, since these have been covered adequately by Hopwood (1933b), MacInnes (1943), Clark and Leakey (1951), Clark (1952), and Napier and Davis (1959). However, the opportunity will be taken to compare the various species of the subgenus. DRYOPITHECUS (PROCONSUL) AFRICANUS D. (P.) africanus was first described by Hopwood in 1933 (a and b), the type specimen being BM(NH) M14084 from Koru, noted in the previous chapter. In these papers Hopwood made clear his view that D. (P.) africanus was ancestral to Pan. The specimen consists of part of a right maxilla with the upper dentition preserved from C1 to M® (Fig. 25). Measurements are included in Appendix 2. The canine is relatively long, narrow, and high; its dimensions are given in ‘Table 33 (comparative data from Ashton and Zuckerman, 1950). ‘The measurements fall squarely within the limits for the chimpanzee. The relatively great size of the canine, particularly its height, compared to the relatively small size of the cheek teeth, suggests that BM(NH) M14084 is a male. The premolars are fairly simple bicuspid teeth with the paracone being higher than protocone, particularly in P?. Both cusps are rather sharp. ‘The buccal border of the tooth is longer than the lingual, due to the mesial 78 TERTIARY PONGIDAE OF EAST AFRICA 719 Table 33. C* measurements of D. (P.) africanus BM(NH) P. troglodytes -P, troglodytes M14084 males females 11.6 xX 13.6 T13 Max L (n=22) (n=16) 95%CL Ce SSN aey/ 8. 0-14. 6 9.1 x 10.9 9.0 Trans B (n=22) (n=16) 95%CL 7.4-14.4 69) 15.2 X 20.1 14.7 La H (n=5) (n=7) 95%CL 8.5-30.7 9. 8-19.6 projection of the mesiobuccal corner, and the enamel line is prolonged up- wards in the mesial half of the buccal surface. The mesiobuccal edge of the paracone is straight and slopes downwards and distally from cervix to tip. There are faint buccal ridges mesially and distally. Mesial and distal marginal ridges are present, although not especially marked. The second premolar is similar, although there is not such a marked disparity in height between the cusps, nor is there a mesiobuccal extension. The first two molars have well-marked trigons with low, moderately rounded, subequal cusps interconnected by rather rounded crests (Fig. 25). Mesial and distal ridges from the tips of the buccal cusps produce what is almost a con- tinuous buccal crest. The hypocone is as large as the protocone and is shifted lingually. The trigon occupies a relatively small area of the occlusal surface, there being prominent mesial and distal marginal ridges and a lingually projecting shelf-like lingual cingulum which runs distally as far as the hypocone. On M?, the largest of the three molars, this cingulum is faintly visible onto the hypocone. The buccal surface slopes strongly. There is a small buccal cingulum between paracone and metacone. The third molar is greatly reduced, particularly in the distal moiety. There is a decreasing development of the metacone from M! to M? and the protoconule is poorly developed on all three molars. The zygomatic process begins to diverge 5 or 6 mm above the alveolar margin; it is set above the distal half of M1. The maxilla is rather full above C1 and P® due to the juga associated with the roots of these teeth. Above P‘ there is a slight concavity, indicating the presence of a small canine fossa. A number of other maxillary specimens of this species are known from Songhor and Rusinga. In general details they are similar to the type, although the teeth are often smaller. One almost complete skull of D. (P.) africanus is preserved, NMK R1948,50 from site R106 on Rusinga. The dentition is very similar to that of the type, although the canine height is less. The incisor region is small and this, associated with the relatively small canines, means that the face is not particularly prognathous and that the tooth rows tend to diverge posteriorly. The skull is described fully by Clark and Leakey (1951, p. 16-28). It has been crushed and was restored by these workers. Robinson (1952) pointed out that the reconstruction was perhaps a little too prognathous. 80 PEABODY MUSEUM BULLETIN 31 The vault is thin, delicate and rounded, there being an almost total absence of cranial superstructures such as tori and ridges. This is presumably because the dentition and facial skeleton are delicately built, and therefore facial and cranial buttressing for powerful masticatory musculature are not required. A similar combination of factors is found in the pygmy chimpanzee, Pan troglodytes paniscus, although here the anterior dentition is larger than that of D. (P-.) africanus, and the skull is somewhat larger and more robust. I (1965, p. 41) suggested that, “the fact that africanus is a small form like the pygmy chimpanzee accounts for the .. . rounded cranial vault, and the lack of brow ridges.” Leakey replied to this (1965, p. 105), “if you enlarge photographs of the skulls of pygmy chimpanzees up to gorilla size, as I have done, you find that the pygmy chimpanzee has just as marked a torus relative to the size of the skull.” All illustrations and specimens of pygmy chimpanzees that I have located show clearly that the facial skeleton is relatively smaller than that in other “normal-sized” chimpanzees; the cranial vault is relatively larger. Correlated with these differences in proportions, the face in pygmy chimpanzees is less prognathous, the cranial vault more rounded, and the supraorbital tori less well developed. Figure 7, taken partly from Remane (1959) shows this clearly, Fig.7 Sagittal sections of Fan troglodytes and D.(P) africanus. (c) and (b) from Remane (1959); (c) modified from Clark and Leakey (I95!). (a) Fon troglodytes (b) Pan troglodytes paniscus (c) DP) africanus TERTIARY PONGIDAE OF EAST AFRICA 81 as do the excellent illustrations in Weidenreich (1941, figs. 18-20). The im- portant points are the relative proportions of face, torus, and vault; no amount of photographic enlargement will alter these proportions. Certain features of the brain of D. africanus are primitive; for example, the relative underdevelopment of cerebral sulci, the small size of the frontal lobes, and the presence of a subarcuate fossa indicating a primitive cerebellar morphol- ogy. This combination of features has been described as cercopithecoid or hylo- batid rather than pongid (Clark and Leakey, 1951), although these authors noted (p. 113) that these are best regarded as primitive catarrhine features which have been retained in some cercopithecoids and hylobatids, and lost (or rather overlayed) in the living pongids. Mandibles and the mandibular dentition of D. (P.) africanus are fairly well represented in the collections. The morphology of the mandible is rem- iniscent of the chimpanzee, although the incisors and incisor region are smaller than in the living species. The bicanine breadth is also low, and hence the rami converge anteriorly. The rami are low (less than 27.0 mm deep at P+), although relatively robust. The symphysis is primitive, the simian shelf being absent. The height of the symphysis from infradentale to gnathion is rarely more than 28 mm and symphyseal thickness rarely more than 12 mm. The anterior margin slopes backward and downward at an angle of about 60 degrees to the occlusal plane; the most inferior part of the anterior border curves gently until it is parallel to the alveolar plane as it merges into the posterior border. The posterior border slopes fairly steeply backward as a planum alveolare. The superior transverse torus is rather poorly developed and low down, only 8 or 9 mm above the inferior transverse torus, which projects posteriorly in the midline almost to the same level as the most posterior point on the superior torus. Apart from the absence of a simian shelf, the significance of which was discussed in the previous chapter, this symphyseal morphology is similar to that of the chimpanzee (see Fig. 3). The postcranial skeleton of D. (P.) africanus is best known from the juvenile material recovered in 1951 by Dr. T. Whitworth at Gumba on Rusinga Island. This material is described in detail by Napier and Davis (1959); they described a representative set of bones from the forelimb, a few from the foot, the maxilla, mandible, and occipital bone of the same juvenile individual. Davis and Napier (1963) reconstructed the skull of this specimen and showed that the face was relatively orthognathous when compared to those of D. (P.) major and of the living pongids. The forelimb skeleton of this D. (P.) africanus was discussed by Napier and Davis principally from the viewpoint of functional anatomy and locomotor behavior rather than of phylogenetic relationships. Their conclusions, together with some modifications suggested by Walker (personal communication), are briefly summarized here, bone by bone. The proximal extremity of the humerus in general features is intermediate between those of the Old World arboreal quadrupeds like Presbytis, classified as quadrupedal semibrachiators (Napier, 1963a; Ashton, Healy, Oxnard, and Spence, 1965; Napier and Walker, 1967), and Pan. The distal extremity is very similar to that of Pan. In particular, the increased surface of origin for the brachialis muscle, the prominence of the keel, the absence of a sharp lateral 82 PEABODY MUSEUM BULLETIN 31 border to the olecranon fossa, the high trochlear index, and the roundness and backward extension of the capitular surface, are all characteristic of the living chimpanzee. Presumably these features imply a forelimb that was mobile, that could be utilized to support the body’s weight during locomotion, and an elbow joint providing freedom of movement and stability for pronation and supination during suspension. The radius is basically that of an arboreal quadruped, although the lateral curvature of the shaft is reminiscent of that in Pan. Walker (personal communication) believes that some features of the radius suggest ground-living adaptations. The distal extremity of the ulna is not similar to that of Pan, but rather to those of New and Old World monkeys. The brachial index (length of radius/length of humerus x 100) cannot be calculated exactly, but lies between 83 and 88. This specimen is juvenile and the adult brachial index would have been somewhat higher. The brachial index in Papio is 104, in Cercopithecus 96, in Pan 93, in Lagothrix 89, in Gorilla 80, and in Homo sapiens 75 (all sample mean values). In general, brachial indices of fossil primates, where known, tend to be lower than those of their putative descendants (Napier and Davis, 1959, p. 38). In this case D. (P.) africanus allies most closely with Pan. Although the hand is relatively shorter than that of Pan when considered as a percentage of total forelimb length, the relative proportions of the different parts of the hand (carpus, metacarpus, and phalanges) are rather similar to those of Pan. It is tentatively concluded that the thumb of D. (P.) africanus was relatively longer than that of Pan. There was a free os centrale, unlike Pan. The set of the trapezium resembles that of Pan and results in the presence of a deep carpal tunnel. This feature is associated with the well-developed flexor tendons of pongids (for example, Pan) and the most arboreal of the New World quadrupedal semibrachiators (for example, Ateles). According to Napier and Davis (1959, p. 60) the articular surface of the trapezium would not permit axial rotary movements of the first metacarpal (and hence true opposability). In general, the carpal bones resemble those of arboreal monkeys, although some features characteristic of Pan are present. Walker (personal communication) believes that some characters of D. (P.) africanus may indicate ground-living behavioral patterns. ‘The metacarpals and phalanges resemble those of arboreal quadrupedal species. The foot is represented by relatively few bones. The medial cuneiform, although morphologically reminiscent of Cercopithecus, is relatively short. Such shortening is connected by Napier and Davis with the general tarsal shortening thought to be associated with an arboreal way of life. On the other hand Walker sees the shortening of the anterior tarsal segment as a terrestrial specialization. ‘The phalanges resemble those of Pan. The morphology of the postcranial skeleton suggests an arboreal animal whose locomotor habits would have been similar to those of the New World forms like Lagothrix or Ateles, quadrupedal animals in which arm-swinging and body-suspension below flexible supports nevertheless play an important part in climbing and other locomotor activities. According to Walker (personal communication), D. (P.) africanus may also have been at least partially adapted to ground-living. Very tentatively, the evidence could be interpreted as in- TERTIARY PONGIDAE OF EAST AFRICA 83 dicating that D. (P.) africanus was a small, active “quadrupedal” form, arm- swing occasionally like some of the New World species. However, it may also have ventured to the ground. Ecologically, the species could have filled much the same sort of arboreal/terrestrial niche as Cercopithecus aethiops. Special resemblances to living primates are almost exclusively to Pan. If D. (P.) africanus is the ancestor of the chimpanzee, or is similar to that ancestor, an increase in body size, more ground-living behavior, and the de- velopment of knuckle-walking adaptations (Tuttle, 1967), would convert the Miocene form into living Pan. In this way, we need not violate the actual fossil evidence, nor is there any necessity to postulate an improbable long-armed, gibbon-like, “‘brachiating” phase in chimpanzee ancestry. THE AFFINITIES OF D. (P.) AFRICANUS Dental and mandibular measurements of D.(P.) africanus are set out in Appendix 3, together with basic statistics for samples of more than eight individuals. There are no detectable differences between samples from Rusinga, Songhor, and Koru. Material from the three samples has been pooled in the calculation of statistics and all specimens can be regarded as being drawn from a single lineage. Leakey (personal communication) believes that some material previously assigned to D.(P.) africanus belongs to another, as yet unnamed, species of this subgenus. The specimen to be described as the holotype of this new species has not been included here. The possibility should be borne in mind that further material, at present classified in D. (P.) africanus, does not in fact belong there. The general resemblances, particularly dental, between D. (P.) africanus and D. (P.) major are probably enough to warrant their retention in the same subgenus. In both species the upper molars and occasionally premolars have prominent cingula, particularly on the lingual side. The first molar is smaller than the second (see Table 34 for ratios of molar lengths). Cingula are also present on the lower molars although these are less marked. The lower molars Table 34 Ratios of MDi lengths of D. (Proconsul) specimens, Gorilla and Pan Jape tgs) Dew (25) Deals) G. P. africanus nyanzae major gorilla troglodytes a 6 3 1 40 26 oe X 90.6 77.9 89.3 94.0 98.9 OR 87.0- 97.4 72.3- 84.0 i 83.5-101.1 82.0-109.8 Bis be! fsa 2 3 1 40 19 EL X 112.9 105.6 102.2 107.3 112.3 OR 112.2-113.6 100.0-109.8 2 96.6-122.8 100.0-122.0 n 4 4 4 40 26 Be Se X 85.2 78.2 86.0 89.3 96.5 OR 80.0- 89.9 75.4- 82.3 81.7- 90.8 85.3- 97.5 90.0-104.9 n 2 4 4 40 25 ae X 84.1 87.5 Baue 98. 4 107.7 OR 83.7- 84.5 83.2- 92.8 80.3- 93.3 86.5-106.9+ 96.5-12201 84 PEABODY MUSEUM BULLETIN 31 increase in size from first to last. Simian shelves are lacking, and the most posterior part of the symphysis is the superior transverse torus. There are other less obvious similarities, for example, the mesial and distal grooves on the buccal surface of the upper premolars, the mesiobuccal extension of P3, and the mesostyles of the upper molars. Many of these features are probably primitive and for that reason would not be of great significance. The differences between the two species are also quite marked. Compared with D. (P.) africanus, the upper canines of D. (P.) major are relatively broader, and the upper premolars more massively built with lower cusps. It should be remembered however that only one individual of D. (P.) major is known which retains the upper premolars, and these should most certainly not be regarded as necessarily typical of that species through space and time. In the upper teeth the molar cingula of D.(P.) major are less shelf-like and projecting, although generally more extensive. The buccal surfaces of the cheek teeth are more vertical, the metacone is not as greatly reduced from first to last molar, and the hypocone does not project as far lingually. ‘This means that, apart from the lingual cingulum, most of the molar occlusal surface is enclosed within the area circumscribed by the tips of the four main cusps. The cusps are less rounded and somewhat higher, and more pyramidal, while the intercuspal crests are rather sharper. The protoconule is better and more consistently developed. The third molar of D. (P.) major is less reduced compared to the second than in D. (P.) africanus (see Table 34 for ratios of tooth lengths). In this characteristic and other features, D. (P.) africanus resembles Pan troglodytes while D.(P.) major resembles Gorilla gorilla. It has already been suggested that D. (P.) major is probably ancestral to G. gorilla (p. 76); there is a fairly good possibility that D.(P.) africanus is in or near the ancestry of P. troglodytes. Table 35 sets out for each dimension listed in Appendix 3 the position of the range in these Miocene species in relation to the ranges of the gorilla and chimpanzee samples included in the same appendix. In 15 out of 24 dimensions, the range of D. (P.) africanus overlaps that of the chimpanzee; in the remaining nine the ranges fall below those of the chimpanzee. If Miocene protochimpanzees, like pregorillas, were generally smaller than their descendants, D. (P.) africanus is suitable metrically for consideration as the ancestor of P. troglodytes. It is of interest to note that in 13 dimensions the ranges of D. (P.) major overlap those of the gorilla, in seven Table 35 Position of the range for 24 dental dimensions of D. (Proconsul) species Range overlaps Range intermedi- Rangeover- Range be- G. gorillarange ate between or laps P. low P. overlaps both P. troglodytes. troglodytes troglodytes and range range G, gorilla ranges D, (P.) africanus - - 15 9 D, (P.) nyanzae Z 5 17 - D. (P.) major ils} 9 2 - TERTIARY PONGIDAE OF EAST AFRICA 85 they overlap both gorillas and chimpanzees, and in two they are intermediate between the ranges of the two living pongids. The proportions of the molars of D. (P.) africanus (Table 34), particularly the reduced M®, resemble those of the chimpanzee, as does their morphology. Loss of cingula seems to be a rather general trend in the Hominoidea (Frisch, 1965), and if it is assumed that protochimpanzees had cingula, then the low, rather rounded molars of D. (P.) africanus are what might be expected of such an ancestor. Clearly, however, many profound changes have occurred since the Miocene. All the teeth have increased in size, a change associated with increase in body size; the teeth anterior to M? have increased the most, the anterior parts of the dentition becoming greatly hypertrophied presumably in response to changes in feeding behavior or diet or both. This increase in the anterior dentition is also found in gorilla phylogeny (but to a lesser degree), and seems once again to be a trend found independently in all three living pongids. The protoconule is less well developed in P. troglodytes then in G. gorilla and in this feature the fossil species parallel their possible descendants (Korenhof, 1960). The maxillary sinus is preserved in only one specimen of D. (P.) africanus, the juvenile from Gumba (Napier and Davis, 1959, p. 13). It extends as far forward as the midpoint of P*. The deepest part of the antrum is between M! and M2. There is a trace of the lacrimal duct anteriorly opening into the inferior meatus; it would presumably have been tubular and rather short. In these features, D. (P.) africanus is very similar to D. (P.) major although built on a considerably smaller scale. The maxilla of D. (P.) major is, however, much more prognathous than those of D. (P.) africanus. The mandible of D.(P.) africanus is similar to, though smaller than, that of P. troglodytes. In particular, the incisor region is very much smaller. (The distance between the distal borders of the lateral incisor alveoli is never more than 16 mm in D. (P.) africanus. In 26 individuals of P. troglodytes the value of this dimension does not fall below 26.4 mm.) Like the maxillary tooth rows, those of the mandible converge anteriorly because of the narrowness of the anterior part of the mandible. The symphyseal cross section contrasts with that of D. (P.) major in that the slope of the anterior surface is shallower in D. (P.) africanus and the anterior and posterior surfaces do not meet sharply at an angle. In D. (P.) africanus the planum alveolare slopes more steeply than in D. (P.) major. As D. (P.) major resembles G. gorilla in symphyseal morphology, so too does D. (P.) africanus resemble P. troglodytes (Fig. 3). The ascending ramus of the mandible is rather low relative to the horizontal ramus, in contrast to the condition seen in Pan troglodytes, D. (P.) nyanzae and, probably, D. (P.) major. This indicates that the face would have been less deep in D. (P.) africanus, indicating a system less adapted to transmitting chewing stresses. The significance of this feature was discussed in Chapter IV. The mandibular dentitions of D. (P.) africanus and D. (P.) major are similar despite the great difference in size. The proportions of molar crown lengths in the two species are also generally similar (see Table 34, p. 83). When the cranial, dental, and postcranial evidence is considered together, 86 PEABODY MUSEUM BULLETIN 31 it seems likely that D. (P.) africanus may well be the early Miocene ancestor of Pan, albeit a much more primitive ancestor. Dentally and cranially the changes in this lineage are probably correlated mainly with changes in diet associated with increasing body size. However, until further work is available comparing the feeding behavior of the predominantly frugivorous chimpanzee with that of smaller cercopithecines with similar diets, little more can be said. Postcranially, D. (P.) africanus was neither a knuckle walker like P. trog- lodytes, nor a brachiator like the gibbons (species of Hylobates). In terms of the rather inconvenient modern locomotor categories, it was probably a quad- ruped. However, D. (P.) africanus cannot be written off as a mere “monkey” because of this, at least not as a cercopithecoid-like monkey. The closest locomotor equivalent would probably have been the living New World quadrupedal semibrachiators or arm-swingers of subfamily Atelinae, perhaps with some ground-living features added. No gibbon-like “brachiator” phase need be posited for chimpanzee evolution. As noted before, similar conclusions concerning the relationship of D. (P.) africanus and P. troglodytes were drawn by Hopwood (1933 b). The postcranial remains of D. (P.) africanus described above indicate that this species was a small, lightly-built, arboreal quadruped, yet a quadruped showing adaptations suggesting that arm-swinging was becoming a major com- ponent of its locomotor activities. Therefore, D. (P.) africanus and D. (P.) major contrast not only cranially but postcranially, too. However, the two species might well have shared a common ancestor in the earliest Miocene, an ancestor which possibly resembled D. (P.) africanus. For the moment they are best retained in the same subgenus. DRYOPITHECUS (PROCONSUL) NYANZAE As noted in Chapter VI the representation of D. (P.) nyanzae at Songhor and Koru is meager, and a number of specimens previously described as D. (P.) nyanzae have been transferred here to D. (P.) major. As far as Early Miocene deposits are concerned, the great majority of D.(P.) nyanzae specimens come from Rusinga Island. Measurements of specimens of this species are included in Appendix 2. There are no noticeable differences between the samples so far collected from the various time levels represented at Rusinga; sample sizes are too small for any trends to be detectable, and the levels appear to span only a short period of time. The samples have consequently been pooled for the calculation of the statistics included in Appendix 3. Material assigned to D. (P.) nyanzae has been described fully by MacInnes (1943), Clark and Leakey (1951), and Clark (1952), and only brief points need to be made here. Dentally, facially, and gnathically, this species is generally larger than D. (P.) africanus and generally smaller than D. (P.) major. Bivariate plots for six upper and six lower teeth are included in Figures 8-13. Since the samples are so small, it is highly probable that the overlap between the three species would be increased were sample sizes also increased. (With such small samples, the effort involved in the construction of confidence limits or equiprob- ability ellipses would produce little worthwhile return.) Like the other two TERTIARY PONGIDAE OF EAST AFRICA Fig8 Bivariate plots of C and P3 of D (Proconsu/). u i) i 13 IS 7 I9 * DIP) major DIP) nyanzae QP) africanus 4 6 8 om || Max. length 88 PEABODY MUSEUM BULLETIN 31 Fig. 9 Bivariate plots of P4 and MI of 2.(FProconsul). a> (S46) hee DIP) major M-D length o D(P)nyanzae x D(P) africanus M-D length TERTIARY PONGIDAE OF EAST AFRICA Fig.10 Bivariate plots of M2 and M3 of D. (Proconsul). M2. wo 9 tl 13 IS 6 M-D_ length « DIP) major o D(P) nyanzae x DIP) africanus M-D length 89 90 PEABODY MUSEUM BULLETIN 31 Fig. 1! Bivariate plots of C and P3 of D. (Proconsul). Max length ¢ D (2) major o D(P)nyanzae *% DIP) africanus 7 9 il 13 I5 17 Max. length TERTIARY PONGIDAE OF EAST AFRICA Fig.12 Bivariate plots of P4 and MI of D. (Proconsul). ¢ DAP) major o D(P) nyonzae x DIP) africanus M-D length ot 92 PEABODY MUSEUM BULLETIN 31 Fig.13 Bivariate plots of M2 and M3 of D.(Proconsul). 9 tl 13 15 17 M-D length « DIP) major o DP/nyanzce x DE) africanus M-D_ length TERTIARY PONGIDAE OF EAST AFRICA 93 species, D. (P.) nyanzae is variable and this variability is presumably due, at least in part, to sexual dimorphism. Although D.(P.) nyanzae is in overall dimensions similar in size to the chimpanzee, the relative proportions of various parts of the dentition and facial skeleton are rather different. As in all early Dryopithecus the incisors are relatively small in the fossil form; consequently the premaxillary region is also small, and subnasal prognathism is not marked. The first molars are small compared with teeth distal to them (see Figs. 9 and 12). Table 34 includes ratios of molar lengths and shows that D. (P.) nyanzae has the lowest values of any Dryopithecus (Proconsul) species for the ratios M1/M? x 100 and M,/M. X 100. The value for D. (P.) major given in this table is from a single specimen and the value for this species should therefore be treated with caution. Apart from these differences in proportions, the three species of D. (Pro- consul) are probably similar enough to be classified in a single subgenus, although many of their similarities are primitive characters. Common features in the lower dentition of D. (P.) africanus and D. (P.) major have already been noted and D.(P.) nyanzae shares these similarities, too. The mandible of D. (P.) nyanzae is more similar in size and proportions to those of D. (P.) major than to D. (P.) africanus (see Clark and Leakey, 1951, p. 105), the horizontal and ascending rami being high. The symphysis lacks a simian shelf and is broadly similar to that of D. (P.) major, although it is less massive. Even in the case of the “1942 mandible’, NMK 1, CMHI1 (Leakey, 1943; MacInnes, 1943, p. 171-174; Clark and Leakey, 1951, p. 45-47), the mandible is large, although the mandibular teeth are small, in size often overlapping D. (P.) africanus (see Appendix 2). This particular specimen is a little problematical; the dentition is crowded and the lower molars, particularly M,, differ in some features from others of this species. The symphysis, although long from infradentale to gnathion, is not particularly thick. The ratio of symphyseal thickness to symphyseal length varies from 53.8 to 57.3 in three specimens of D.(P.) major (NMK 190,1; BM(NH) M14086; and NMK 198,28), and from 43 to 48 in three specimens of D. (P.) africanus (NMK_ 262,635; NMK R1948,50; NMK 640,417). The ratio in NMK 1, CMH1 is 43. Unfortunately data on other specimens of D. (P.) nyanzae are lacking. Specimen NMK 1, CMH1 has been treated here as a D. (P.) nyanzae. Clark and Leakey (1951, p. 54-55) have summarized the main differences between D. (P.) nyanzae and D. (P.) africanus as follows: Apart from the absolute dimensions of the teeth of the premolar-molar series . . . other features of the dentition which distinguish P. africanus from P. nyanzae are as follows: the greater M'/M? and M?/M® ratios, the marked regression of M®, the relatively smaller and more slender canines, the tendency to caniniform development of the buccal cusp of P%, the relative height in the upper molars of the internal cingulum which tends to form a projecting shelf, the better definition of the trigon by clear-cut crests joining the protocone, paracone and metacone, the greater tendency for the hypocone to merge broadly with the internal cingulum (in M? and M2), the relatively large size of the hypocone which in the first two 94 PEABODY MUSEUM BULLETIN 31 molars approximates to that of the protocone, the weak development of the protoconule which forms little more than a slight thickening at the forward angulation of the ridge joining the protocone and metacone, the broader anterior cingulum, the weaker definition of the posterior cingulum, and the absence of the elaborate degree of coarse beading of the internal cingulum which appears to be so characteristic of P. nyanzae. At the time of this description, D. (P.) major, particularly the upper dentition, was poorly known. Now that larger samples of this species are available it is clear that the upper teeth resemble those of D. (P.) nyanzae rather than D. (P.) africanus, for example, in the morphology of the trigon and cingulum, the relative proportions of the cusps, the slope of the buccal surface, and the well-developed protoconule. The upper molars, particularly M? and M3, are very similar in nyanzae and major. One of the most complete speci- mens known of D. (P.) nyanzae is the holotype, BM(NH) M16647, first described by MacInnes (1943, p. 164-168) as Proconsul africanus (Fig. 27). It consists of much of the face and upper dentition of what is probably, from the evidence of the canines, a large adult male. The dimensions of the left canine are: Max L, 15.8 mm, Trans B, 12.0 mm, and La H, *21.5 mm. All these measurements fall outside the values of the 95 per cent confidence limits for female chimpanzees listed earlier in this chapter and close to the mean values for males. The specimen was described further by Clark and Leakey (1951, p. 44-45) and figured by them (pl. IV, figs. 16 and 20). It is distorted and damaged. The incisors are missing and the premaxillary region between alveolare and naso- spinale is almost entirely absent. The left tooth row from canine to first molar is preserved; the last two molars are intact, although displaced upwards. The right canine and first premolar crowns are damaged; the second premolar and first molar are intact and in position; the second molar is intact but displaced a little upwards and buccally; the third molar is broken in half mesiodistally, the buccal half facing almost sideways, the lingual half being pushed upwards. The palate is intact to M1 on the left and M? on the right. The suture between the palatine processes has been invaded by matrix, and this has forced the maxillae slightly apart. With respect to the right maxilla, the left has also shifted mesially. The separation of the palatine processes increases gradually from opposite the canines to the level of M?, and thus the tooth rows appear to diverge, although in fact they probably may not have done so. The distance between the lingual surfaces of the P#’s is 32 mm; corrected for distortion the distance would be approximately 30 mm. The facial region is crushed and distorted, although well enough preserved on the right for some details and measurements to be determined. This specimen, the holotype of D. (P.) nyanzae, will be compared very briefly here with the only complete maxillary specimen of D. (P.) major, UMP 62-11 from Moroto II. BM(NH) M16647 is built on a rather smaller scale than UMP 62-11, the alveolar breadths being at least 10 mm smaller. The alveolar juga associated with the canines are similar; there is no premolar jugum in BM(NH) M16647 since the buccal roots of P? are not as massive as TERTIARY PONGIDAE OF EAST AFRICA 95 in UMP 62-11. Posterior to the jugum is a concavity running upwards to the infraorbital foramen, the canine fossa. In BM(NH) M16647 the midpoint of this depression is 21.5 mm above the alveolar margin and is situated above the distal border of P*. In UMP 62-11 the depression is similarly situated but is 25 mm above the alveolar margin. The canine juga of BM(NH) M16647 surround the piriform aperture and gradually flatten superiorly to become long and narrow maxillary frontal processes. ‘The nasal aperture was long and narrow; this much at least can be determined in spite of the fact that the right side is crushed inwards and the left side pushed forwards. The nasal bones were long and narrow and their curvature similar to those of UMP 62-11. The two specimens are very similar in the general architecture of the face and the internal parts of the nasal aperture, although BM(NH) M16647 is generally smaller. The zygomatic pro- cesses of both specimens are situated mainly above M?. Some facial and other dimensions are included in Table 36 below. Many of the measurements of BM(NH) M16647 are approximations; distances between bilaterally symmet- rical structures are generally measured from one side to the midline and doubled. Table 36 Facial measurements of BM(NH) M16647 and UMP 62-11 BM (NH) M16647 UMP 62-11 Nasal H *75. 0=*80. 0 *995/5 Nasal B *2,050 30.2 Nasals: L Minimum *45.0 60.0 Nasals: B at infraorbital foramina alg ZO B across antr. lacrimal crests *20.0 20.0 B across infraorbital foramina *40.0 40.0 Infraorbital for. to antr. lacrimal crest +2950 *30.0 Palatal D: at C*-P®* *4,0 *4.0 at M* *7.0 *7.5 Measurements of the dentition and tooth rows of the two specimens are set out in Table 37. Length/breadth indices are also included. The diastema be- tween lateral incisor and canine is 5.2 mm in UMP 62-11, 5.0 mm in BM(NH) M16647. The canine of BM(NH) M16647 is more slender than that of UMP 62-11 (L/B index 131.6 compared with 118.4). Other than this, the two teeth are morphologically similar. The premolars are more massively built in UMP 62-11. The paracone of P? in BM(NH) M16647 has been worn sharp on its mesial surface by attrition from P;. The P#’s are similar, although those of BM(NH) M16647 are smaller. The buccal surfaces of P? and P*4 are similar in the two specimens, having mesial and distal vertical grooves. In UMP 62-11 the distal borders of P? and P* are convex distally; this bulging is absent in BM(NH) M16647, the distal borders being flatter, particularly in P?. The first molars differ. The surface details of both have been largely obliterated by wear, although it is probable that the morphology of cusps, crests, cingula and so forth were broadly similar. The mesiodistal length of BM(NH) M16647 is 96 PEABODY MUSEUM BULLETIN 31 Table 37. Dental measurements of BM(NH) M16647 and UMP 62-11 D. (P. )major D. (P.)nyanzae UMP 62-11 BM (NH)M16647 (left, except for P*) Max L *18.6 15.8 c Trans B calls ¥/ 5 (0) LaHH *27.0 *21.5 Index L/B x 100 118.4 131.6 MDi L 8.2 7.0 ps Max L 10.7 8.2 BuLi B 14.6 11.8 Max L/B x 100 7303 69.5 4 MDIL 7.9 6.8 Pp BuLi B 14.8 12.0 Index L/B x 100 Geo 56.6 MDi L ealiley, 9.6 M2? BuliB ye Y/ n220 Index L/B x 100 92.1 80. 0 MDi L *13,1 12.4 M* Bulli B 14.0 13ai7, Index L/B x 100 93.5 90.5 MDi L 12.8 Ws M* BuliB 14,4 14,1 Index L/B x 100 88.9 80.1 Po=Me - 54.4 *47.5 i. 8 39.1 *30. 0 (GRE 3) *69. 0 *55.0 pe. 8B 64. 8 *54.0 much less than that of UMP 62-11 and this is reflected in the crown index. The mesial and distal borders of BM(NH) M16647 M? are straight and parallel. The crowns of all the cheek teeth of UMP 62-11 are marked by their distal bulges; these bulges are absent in P*, P#, and M! of BM(NH) M16647, and these teeth are compressed mesiodistally. The last two molars, however, are very similar in UMP 62-11 and M16647, although the distal bulges are more marked in the former specimen. The protoconule is well marked on the right M? of BM(NH) M16647, and this is illustrated in Figure 27. The buccal cingulum of the molars forms a series of spurs in BM(NH) M16647, just as in UMP 62-11 (see Figs. 19 and 27). The face of BM(NH) M16647 would have been narrower and less prog- nathous than UMP 62-11. The dentition anterior to M? is relatively smaller, less long mesiodistally, and more compressed in BM(NH) M16647. Nonetheless, the similarities between these two maxillae and upper dentitions are marked, and although they clearly are drawn from two different species, the smaller form could well represent a species morphologically—if not temporally—an- cestral to D. (P) major. D. (P.) nyanzae is more similar to D. (P.) major than to some of D. (P.) africanus. If D. (P.) major and G. gorilla are related, then some of the TERTIARY PONGIDAE OF EAST AFRICA 97 dental trends within this lineage are as follows: increase in size of the more anterior teeth, particularly P*,, P*,, and M1,, but including incisors and canines, too. If these trends are extrapolated back through time from D. (P.) major, the ancestor of D. (P.) major might well have resembled D. (P.) nyanzae. It is even possible that some of the oldest D. (P.) nyanzae could be ancestral to the larger species. A tentative hypothesis is that populations of D. (P.) nyanzae (or an immediately preceding species) became progressively isolated ecologically, perhaps in elevated and forested terrain, and evolved into D. (P.) major. More fossil material, with adequate stratigraphic control, good dating, and from a variety of sites representing a variety of environments, is necessary before this hypothesis can be any more, or less, than a possibility. POSTCRANIAL REMAINS OF D. (P.) NYANZAE The forelimb and foot skeleton of D. (P.) africanus from Gumba, Rusinga Island and the postcranial material of D. (P.) major from Uganda and Kenya have already been reviewed. The remainder of the material from the East African sites is listed below; much of it is D. (P.) nyanzae, although some could be Leakey’s new species “Kenyapithecus africanus.” 1) Left talus (NMK 74, CMH147) from Kalim, Rusinga, described by Mac- Innes (1943, p. 174-177) and Clark and Leakey (1951, p. 87-92). 2) Left talus (NMK 511,234) from Kulu Waregu, Rusinga, described by Clark (1952, p. 277). 3) Distal end of left tibia (NMK 553,924) from R2-4, described by Clark (1952)sp. 278): 4) Most of right femur, proximal end of left femur, shaft of left humerus, and part of right clavicle, all associated (NMK 120, 1933), from Maboko Island, described by Clark and Leakey (1951, p. 93-98). 5) Fragment of lateral part of left clavicle (NMK 254,394) from R105-106, Rusinga, mentioned by Clark (1952, p. 278). 6) Fragment of lateral half of right clavicle (R313) from R3, Rusinga, de- scribed by Clark (1952, p. 278). The two clavicular fragments are said by Clark (1952, p. 278) to be “very similar to (but slightly larger than) the clavicular specimen previously described from Maboko Island and provisionally referred to P. nyanzae. In general dimensions they are comparable to the clavicle of an adult male chimpanzee.” These postcranial specimens, if indeed they all belong to D. (P.) nyanzae, indicate that, like other species of this subgenus, nyanzae was a relatively lightly built, actively arboreal, quadrupedal form. Leakey (1967) has suggested that some of the postcranial remains from East African Miocene sites may belong in “Kenyapithecus africanus.” At present there is no way of proving this, although the material assigned to D. (P.) africanus and D. (P.) major seems to have been correctly classified. CHAPTER VIII. THE EARLIEST HOMINIDS In this chapter I shall be discussing the relationships between the late Miocene/early Pliocene hominid Ramapithecus from deposits in India and Kenya. Recently Leakey (1967) has suggested, among other things, that earlier hominids still have been found in East Africa. In order to discuss these matters in detail it is necessary to review some of the other species of Dryo- pithecus. A large number of fossil apes have been recovered from the Siwalik Hills in India. Simons and I (1965) classified most of this material in Dryopithecus (Sivapithecus). Within this subgenus were two species, D. (S.) indicus and the smaller D. (S.) stvalensis, the latter containing species described previously in the genus Sugrivapithecus by Lewis (1934, 1937a). My recent work (Pilbeam, in preparation) suggests that a majority of D. (S.) indicus and D. (S.) stvalensis are merely size variants within a single species D. (S.) sivalensis. Certainly the metrical and morphological variation among members of this proposed species would be less than that encountered within Pongo pygmaeus, the orangutan, to which the Tertiary form is possibly related. A number of other workers (Gregory and Hellman, 1926; Lewis, 1937b; Simons, 1963) have suggested that some of the Siwalik pongids represent a species ancestral to the orangutan. Simons and I (1965) discussed the ancestry of Pongo pygmaeus and concluded then that its Tertiary ancestors were either unknown or unrecognized. It now seems rather more possible (see Pilbeam, 1966, 1967, and in preparation) that D. (S.) sivalensis could be ancestral to P. pygmaeus; the latter’s occlusal specializations would then have been acquired mostly during the last 10 million years. D. (S.) stvalensis has been recovered from deposits of Miocene and Pliocene age in the Siwalik Hills (Chinji and Nagri zones), approximately equivalent in age to the period from European Sarmatian (or even Late Vindobonian) to Pontian (Simons and Pilbeam, 1965, p. 95-97). Absolute ages for these deposits are not available, although by analogy with European and African dated sites the ages of these beds would be between 14 or 15 million and 8 or 9 million years. If Hipparion evolved in North America about 12 million years ago, and if Hipparion is confined to the Nagri zone and is absent from the Chinji, then the Chinji- Nagri boundary would be drawn at around 12 million years. The molars and premolars of D. (Sivapithecus) species differ from those of D. (Proconsul) in lacking buccal or complete lingual cingula, in having relatively rounded cusps in some specimens, and in exhibiting markedly sloping buccal and lingual surfaces. These and other diagnostic criteria, together with illustrations, are listed in Gregory and Hellman, 1926; Lewis, 1934, 1937a and b; and Gregory, Hellman and Lewis, 1938. In 1943, MacInnes described a series of specimens collected in preceding years at Rusinga and Songhor. Among the specimens was a left maxillary fragment from R106, Rusinga Island containing P3, P* and M1 (1943; ‘pl: 98 TERTIARY PONGIDAE OF EAST AFRICA 99 24, fig. 2), which was assigned by MacInnes to Proconsul africanus. In 1950 this specimen was made the type of a new species, Sivapithecus africanus, by Clark and Leakey (p. 261). It is at present in the British Museum (Natural History), number M16649. This specimen is illustrated in Figure 28. In 1951], Clark and Leakey described in detail this specimen, as well as two isolated upper molars (NMK 20, CMH27 and NMK 121, CMH26) assigned by them to the same species (Clark and Leakey, 1951, p. 62-67). Their conclusions were as follows (p. 67): In summary, the upper teeth of S. africanus differ from those of all the other large apes from the early Miocene deposits of East Africa in the absence of an elaborate internal cingulum and in the relative simplicity of the cusp pattern. They closely resemble the Asiatic species of Sivapithecus ...in the large size of the upper premolars, the relative size of the main cusps of the upper molars, the limited extent and sharp definition of the trigon, the absence of coarse secondary foldings of the enamel, the absence of a large posterior fovea, and the clear separation of the hypocone from metacone and protocone. The specific separation of S. africanus from the Asiatic types may be justified by certain features such as the flatness of the palate ..., the persistence of a trace of the antero-internal cingulum on the upper molars (which, however, is also indicated in an upper molar, D.176, of S. indicus . . .), the development of the internal cingulum on P‘, and the slightly smaller width of the first and second upper molars. Moreover, it is extremely improbable that representatives of an identical species would be found so widely separated in time and space. They also stated (p. 110) that the East African species might be ancestral to D. (S.) stvalensts. Although in 1951 he had agreed in the assignment of BM(NH) M16649 to a species of pongid, in 1953 (p. 178), Leakey wrote the following: The evidence of such teeth as we have, however, most strongly suggests that the Kenya Sivapithecus may prove a candidate for the role of a direct ancestor of man in the Lower Miocene. The genus Proconsul, and especially the species nyanzae, cannot be wholly ruled out of the picture, but the Proconsuls have, in particular, upper molar teeth which tend to suggest specialization away from the direction taken by man, whilst Sivapithecus has upper molars strongly suggestive of man himself. It seems likely that it was the presence of molar cingula in D. (Proconsul) species which enabled Leakey to rule them out of the ancestry of man. The absence of cingula in BM(NH) M16649 does not mean, however, that it is necessarily related to Hominidae. Other undoubted pongids of genus Dryo- pithecus also lack cingula. In 1962, Leakey described a primate from Late Miocene deposits at Fort Ternan, Kenya (see Chap. III) as Kenyapithecus wickeri. This specimen, he implied, might be a hominid, and it has since been accepted generally as such. Leakey, however, did not place it in any primate family, but wrote (1962, p. 696), “I strongly suggest that when the new genus is allocated to a family, Sivapithecus africanus will have to join it there and be removed from 100 PEABODY MUSEUM BULLETIN 31 the Pongidae.” In 1963 (p. 42) he wrote, “It now seems possible that Sivapithecus africanus should be treated as a species within the new genus Kenyapithecus, but one which is more primitive and perhaps ancestral to wicker.” In the same year, Simons suggested that ‘“Kenyapithecus wickeri” was a junior synonym of a hominid described by Lewis in 1934, Ramapithecus brevirostris. In 1964, Simons described this and other material as R. punjabicus. In 1965, Simons and I transferred BM(NH) M16649 from Sivapithecus africanus to D. (S.) sivalensis. Both reassignments have been strongly criticized recently in a paper by Leakey (1967), in which he maintains “Kenyapithecus wickeri” as a species distinct from Ramapithecus, and also transfers BM(NH) M16649 to Kenyapithecus to become K. africanus, ancestral to “K. wickeri’. To this species he has also assigned a number of specimens, some newly recovered, from Rusinga and Songhor either previously undescribed or (mainly) classified as D. (P.) nyanzae or africanus; he regards “K. africanus’ as the earliest hominid. These taxonomic changes will be discussed in detail in the following pages, together with a certain amount of background material. For the remainder of this work the Fort Ternan hominid will be described either as ‘“Keny- apithecus wickeri”, without being italicised, or by its specimen numbers. In 1934, G. E. Lewis described a new primate genus from the Siwaliks, Ramapithecus. When first described, this genus contained two species, R. brevirostris and R. hariensis. The latter came from Nagri beds of Late Miocene or Early Pliocene age (very approximate dates, from about 12 to 8 million years). The former was originally thought by Lewis to come from younger beds, but in 1937 (a and b) he corrected this and ascribed R. brevirostris instead to the Nagri zone (see discussion in Simons, 1964). R. hariensis consisted of a distorted maxillary fragment with worn Mt? and fractured and broken M2. In 1937 Lewis (1937a) transferred the speci- men to Sivapithecus sivalensis (later regarded as D. (S.) sivalensis by Simons and myself, 1965). It is a fragmentary specimen and may in fact be a Ramapithe- cus. Its exact taxonomic position cannot at present be established. R. brevirostris consisted of a right maxilla, YPM 13799, with P3—M2? preserved (Fig. 29). The canine alveolus, the root of the lateral incisor, and the lateral part of the central incisor alveolus are also present. In his first paper, Lewis (1934) gave a minute description of this specimen and concluded that it was either a manlike pongid, or a very primitive hominid. In his unpublished doctoral dissertation (1937b, p. 45-46) Lewis stated his view that R. bre- virostris was ancestral to the Pleistocene genera Australopithecus and Homo. “The lack of diastemata; the relatively small incisors, canines, and premolars; the low relief and elliptical horizontal cross section of the upper premolars; and the bluntly rounded cusps and other morphologic features are common to Ramapi- thecus, Australopithecus, and Homo. It would seem therefore, that Ramapithecus is the structural, and possibly direct, ancestor of these two genera. In the year following the original diagnosis of Ramapithecus, Hrdlicka (1935) published a highly critical note. His contribution is disappointingly subjective and should be ignored. Unfortunately, it seems to have had a con- siderable and stultifying influence upon paleoanthropological thinking. Until the 1960's, Ramapithecus brevirostris continued to be treated as the most TERTIARY PONGIDAE OF EAST AFRICA 101 manlike of the Middle Tertiary pongids, but no more than that. In a taxonomic revision of this and other Indian material Simons (1964) transferred it to R. punjabicus. The premaxillary and anterior maxillary regions of YPM 13799 are relatively small and the juga associated with the tooth roots delicately built. Subnasal prognathism is not marked, and the distance between nasospinale and alveolare would have been little more than 10 mm. Fortunately, the maxilla forming the base of the lateral border of the nasal aperture and the lateral wall of the inferior meatus is preserved for 15 mm or so anteroposteriorly. The maxillary fragment and its mirror image can be mutually positioned using the planes of the lateral nasal walls and the orientation of the canine alveolus. The labiolingual axis of the canine alveolus is oriented at an angle of some 55 degrees to the long axis of the premolars and molars in both YPM 13799 and later hominids, suggesting that the dental arcade was parabolic. The shape of the canine alveolus is oval, with the long axis labiolingual as in hominids, and not mesiodistal as in pongids. The alveolus measures 7.5 mm labiolingually (restored) by 6.5 mm mesiodistally. The canine root would have been some 15 mm long. The zygomatic process is situated above M1. Generally in living and fossil pongids the process is above M2? or the distal root of M?, although individual specimens, particularly of Pongo pygmaeus, may have processes situated further forward. Anterior to the process is a marked canine fossa, a tucking-in of the maxilla medial to the plane of the buccal tooth roots and posterior to the canine jugum. Although each of these features may be found individually in living or fossil pongids, as a total morphological pattern they occur infrequently outside the Hominidae. The hominid characters are to be seen particularly in metrical features and proportions. The morphology of the cheek teeth is of interest. The first premolar is roughly triangular in occlusal view. The buccal length is a little greater than the lingual, due to the mesial extension of the mesiobuccal corner. A distinct fovea is present at this corner and is circumscribed by the mesial marginal ridge and the more mesial of the two paracone-protocone crests. The mesio- buccal corner does not extend superiorly, as it does in Middle Tertiary pongids; for example D. (P.) major (see Chap. IV and Fig. 18), and D. (S.) sivalensis (Gregory, Hellman, and Lewis, 1938, pl. 1, fig. 2a; pl. 7, fig. 1). In these cases, also, the triangularity is more marked and the mesial border is concave mesially. In YPM 13799 the mesial border of P® is straight, any slight concavity being due to interproximal attrition with C! (Lewis, 1934, pl. 1, fig. la; Simons, 1961, fig. la; and Fig. 29). Robinson has noted (1956, p. 56) similar features in Australopithecus (Paran- thropus) robustus. ‘In occlusal view the crown is a little asymmetrical on account of the slight projection of the buccal half of the mesial face.” For Australo- pithecus africanus he notes (1956, p. 58), “The mesiobuccal groove commonly is a definite groove which lies in a triangular-shaped depression which is wide at the occlusal margin. In some cases . . . this effect is so marked that there appear to be two mesiobuccal grooves converging as they pass upwards from the occlusal margin.” 102 PEABODY MUSEUM BULLETIN 31 The relief of the little-worn cheek teeth is less marked than in Dryo- pithecus. The lingual surfaces of buccal cusps slope less steeply, and, particularly in contrast to D. (S.) sivalensis, the occlusal fovea occupies a relatively greater proportion of the total surface of the crown. In D. (S.) sivalensis the buccal and lingual surfaces of the cheek teeth slope gently and the minimum buccolin- gual breadth between the tips of paracone and protocone is often less than 50 per cent of the maximum across the same cusps (see Table 39, p. 107). ‘The meta- cone-protocone crest is also more pronounced in R. punjabicus molars and is crossed by a less pronounced sulcus than in D. (S.) stvalensis (compare Simons, 1961, fig. 1A with Gregory, Hellman, and Lewis, 1938, pl. 7, fig. F; and Fig. 29 with Fig. 30). Morphologically, the molars and premolars of Ramapithecus differ from those of Dryopithecus in a number of features, and resemble Australopithecus. The relative simplicity of the occlusal surface and the absence, or at best, poor development of lingual cingula gives D. (S.) sivalensis a superficial resemblance to R. punjabicus. This means that the allocation of isolated teeth or small fragments to one or the other sometimes becomes difficult. The two upper molars originally assigned by Lewis to R. hariensis are a case in point. A number of authors have noted the essential basic similarity of the cheek teeth of all hominoid species (for example, Gregory and Hellman, 1926; Korenhof, 1960). Schuman and Brace (1954) remarked on the difficulty in discriminating isolated human and chimpanzee molars; this has also been the case with worn human and orangutan molars. Clearly, if the living forms are often difficult to tell apart, the problem is likely to be more acute in dealing with their Middle Tertiary ancestors. Figure 30 compares two molars of D. (P.) major and one of D. (P.) sivalensis. Apart from the variable develop- ment of the lingual cingulum, the three teeth are very similar and could easily have been sampled from a single variable primate species. It is to be expected therefore that certain of the cheek teeth of Miocene and Pliocene hominids and pongids will appear morphologically similar, particularly in the absence of obvious discriminating characters like large cingula. In 1938, Gregory, Hellman, and Lewis described (p. 21-22) and figured (pl. 2, fig. 8) a right horizontal mandibular ramus (GSI D168). This specimen preserved crowns of P; to M, and the canine and incisor alveoli. They assigned it tentatively to Ramapithecus (in fact to R. cf. brevirostris). The mandible is gracile, with small cheek teeth worn flat; the P, is elongated and sectorial, and the inferior transverse torus is large and projects posteriorly as a simian shelf. This mandible is clearly not that of a hominid (Simons, 1961, 1964; Simons and Pilbeam, 1965) and is not to be associated with Ramapithecus. The tentative allocation of this specimen seems also to have confused later workers and hampered the correct interpretation of Ramapithecus. In 1961, Simons wrote an important paper which began the long-delayed rehabilitation of Ramapithecus as the earliest known hominid. In this he discussed the dating of YPM 13799, the polemical nature of Hrdlicka’s con- tribution, the misclassification of the mandible GSI D168 as well as a number of previously unmentioned morphologic and metric characters. For the first time in a published paper, a paleontologist noted the illogicality of regarding TERTIARY PONGIDAE OF EAST AFRICA 103 the maxilla as that of a manlike ape rather than as a hominid. In short, when they occur in sufficient numbers and in the right combination, similarities should be treated as homologies rather than parallelisms, unless or until such an interpretation can be disproved. In 1962, Leakey described* a new primate from Fort Ternan in Kenya, from deposits approximately equivalent in age to the Chinji beds of the Siwaliks. They have been dated radiometrically to 14 million years (see Table 1, p. 17). The new find was composed of a left maxilla NMK FT 1272, with canine, roots of P3, and crowns of P+ to M?; a right maxilla NMK FT 1271, with M1, M2, and roots of M3; and a right Mz, and was made the type of a new genus and species, ‘““Kenyapithecus wickeri” (Fig. 29). The new specimen showed many similarities to YPM 13799, as can be seen from a comparison of Leakey’s paper with Lewis’s original detailed description. Simons (1968) has recently reviewed this material in detail. The canine crown and 10 mm of root are preserved intact in NMK FT 1272, and are set almost vertically in the maxilla; the canine jugum in YPM 13799 is set at a similar angle to the occlusal plane. The canine crown is small, as in hominids, although morphologically different from the earliest Pleistocene hominids. Both mesial and distal borders slope downwards towards the tip of the crown. In Australopithecus both borders are vertical for much of their length. Leakey (1962, p. 693) describes the canine as follows: It is a relatively compressed canine, bucco-lingually, and has a consequent slightly exaggerated anterior-posterior length, which is further stressed by a swelling on the lower anterior margin at the point where the internal cingulum ends. There is a shallow groove on the anterior face running from this swelling to the tip, much as may be seen in upper canines of the Proconsulidae and in some upper canines of the Asiatic Stvapithecus, but it is much less strongly developed. The lingual face of the canine is slightly convex but lacks the strong medium [szc-median?] ridge seen in the upper canines of the genus Sivapithecus, even in small females. On the posterior face there is a narrow area of wear extending longitudinally from the tip to a point just above where the cingulum ends posteriorly. This wear facet seems to postulate a lower third pre-molar of semi-sectorial type. The inner face of the crown of the canine is marked by a cingulum. The canine is illustrated in the plate included in Leakey’s paper. In this plate it is compared (fig. c) with “a cast of the smallest published Stvapithecus upper canine” (1962, p. 696). This is not the smallest published D. (Siva- pithecus) C1; that had been described (p. 12-13) and figured (pl. 1, fig. 1; pl. 2, figs. C, D, E; pl. 4, figs. A, B) by Gregory, Hellman, and Lewis (1938). This small canine is part of an upper dentition assigned by them to D. (S.) sivalensis and is numbered in the 1938 paper K29/466; it is now GSI D299/300. It is similar to the canine of NMK FT 1272 although 1 or 2 mm larger in each dimension. The mesial groove is shallow and there is no median lingual ridge, contra Leakey. The presence of a distolingual wear facet in the Kenya C! * The description of this specimen was first published on 22nd May 1962, in Ser. 13, vol. iv, Annals and Magazine of Natural History for November, 1961. 104 PEABODY MUSEUM BULLETIN 31 implies that P; protoconid was prominent, not that it was necessarily sectorial or semi-sectorial, or that the canine projected more than in Leakey’s recon- struction. The dimensions of the canine of ‘‘K. wickeri’ are, Max L, *10.0 mm, Trans B, *8.0 mm, and La H, *11.5 mm. The root is compressed, having the fol- lowing cross-sectional dimensions at the (inferred) alveolar margin, Max L, *8.0 mm, and Trans B, *6.0 mm. The canine crown of YPM 13799 is not preserved, although the alveolus is. The labial wall of this alveolus has been crushed lingually. The maximum diameter of the alveolus is a labiolingual one and is set at an angle of approximately 55 degrees to the mesiodistal line of the cheek teeth. The dimensions of the alveolus are, Max L, *7.5 mm, and Trans B, *6.5 mm. If the long axis of the crown of the C! paralleled the mesiodistal line of the dentition from P? to M3, the C! crown of YPM 13799 would have been set at approximately the same angle to the root as in NMK FT 1272, although the crown dimensions would presumably have been a millimeter or so smaller. The crown of P? is missing in “K. wickeri’”. The tooth had three roots, the mesiobuccal being larger than the distobuccal, as in YPM 13799. The crown was evidently only slightly longer buccally than lingually and the mesial border was probably straight. Another tooth, P*, also from NMK FT 1272, is described thus by Leakey (1962, p. 693-694): This fourth upper pre-molar is specially interesting for a number of reasons: it differs markedly in its morphology from the corresponding tooth in Sivapithecus africanus described by Le Gros Clark & Leakey, and it is clearly of a different species. In the present specimen there is no cingulum at all, whereas in Sivapithecus africanus there is an anterior and a posterior cingulum with the latter extending to the lingual face. The crown of this fourth upper pre-molar of Kenyapithecus wickeri is very wide bucco-lingually .. . having almost the same diameter as the first upper molar; the surface of the crown is much more complex than in Sivapithecus africanus. In addition to the well defined anterior and posterior fovea, there is a mesial [sic] fovea formed as follows: a ridge runs from the tip of the lingual cusp towards the centre of the tooth and then divides into two parts which enclose the mesial fovea. This is a structure which I have not seen in any Proconsul or in any of the Asiatic Stvapithecus, Ramapithecus, or other Siwalik primates. A slightly similar structure is to be seen in some of the less worn upper pre-molars of Australopithecines. This fourth upper pre-molar is also very much lower crowned than in the Asiatic genus Sivapithecus, and is comparable, in this respect, to Ramapithecus. It is distinctly lower crowned than the corresponding tooth in Sivapithecus africanus. The contact facet between the fourth upper pre-molar and the third upper pre-molar is long, even though a part of it is missing on the chip at the external corner. ‘This facet must have been at least 5 mm long when TERTIARY PONGIDAE OF EAST AFRICA 105 intact, a feature in which it resembles Ramapithecus brevirostris, and a character which is certainly connected with a shortening of the face and of the molar pre-molar series becoming compressed from end to end. This type of compression is not seen in Sivapithecus africanus, nor in any of the Indian species of Sivapithecus about which I can find record. The ‘“‘mesial’’ fovea should be termed “central”, since a “mesial” fovea would be the anterior fovea. An exactly similar structure is formed in YPM 13799; apart from mesial and distal marginal ridges in YPM 13799, there are three further transverse ridges joining paracone and protocone, making five transverse ridges in all. The third and fourth (distalward) are crossed by a central mesiodistal sulcus. These features are described in detail by Lewis (1934, p. 165) and are very similar to those seen in “K. wickeri.” Contact facets between adjacent cheek teeth are up to 5 mm long in “K. wickeri” and 6 mm or more in YPM 13799. As in the latter, the buccal and lingual surfaces of P* are much more vertical than in D. (S.) stvalensis. The first molar in NMK FT 1271 and 1272 is low crowned, with enamel smoothed by wear. The occlusal fovea occupies most of the area of the crown and cingula are lacking, except for a mesiolingual cingular remnant. The metacone-protocone crest is prominent and is barely crossed by a faint sulcus which has been almost removed by wear. The lingual surfaces of the buccal cusps slope much more gently than in D. (S.) sivalensis or D. (Proconsul) and the cusps are less prominent. The second molar is similar, though a little larger. The metacone and hypocone are relatively a little smaller than in M!, and the buccolingual breadth across the distal moiety is accordingly smaller than that across the mesial. In all these features, “K. wickeri” is closely similar to YPM 13799. The third molar roots, preserved on the right maxilla, NMK FT 1271, indicate that this tooth was shorter than M2. The zygomatic process of the maxilla is situated above the distal half of M!, and there is a canine fossa, between the (presumed) canine jugum and the process, which is situated above the buccal roots of P? and P4. The palate was arched much as in YPM 15799. A right M, was also preserved (Leakey, 1962, plate, fig. d). There is a small cingulum between protoconid and hypoconid and the hypoconulid is situated in the midline. The tooth is low crowned and the cusp tips are situated close to the margins of the tooth. The buccal cusps are lower than the lingual. Leakey concluded (1962, p. 695), “In all of these characters it [“K. wickeri’’] shows a greater or lesser approach towards the structures we associate with the Hominidae. For this reason, I have purposely refrained from suggesting any family into which this new species should be placed.” In subsequent publications (for example, Leakey, 1963, 1967) he has made it clear that he believes this specimen to be a hominid. In spite of Remane’s recent comment that the canine is more like those of female pongids than of hominids (1965, p. 281), it is doubtful whether the combination of features seen in ‘‘K. wickeri’’ would occur outside the Hominidae. From the published descriptions of Lewis (1934), Simons (1961) and 106 PEABODY MUSEUM BULLETIN 31 Leakey (1962), it is clear that YPM 13799 and “K. wickeri” are extraordinarily similar and that their similarities often extend to the most minute details (Simons, 1968). They are best classified in the same genus, probably in the same species (see Pilbeam, 1966, 1968). Any other classification would obscure the very great similarities between these specimens. It is difficult, indeed, to see which morphologic characters could be used to support species distinction. This point was made by Frisch (1962), and Simons (1963). However, the available material is not extensive and it might well prove that such limited remains simply do not permit identification at the species level. For example, similar fragments of two species of Macaca might well be difficult to tell apart. One solution to the problem would be to describe the Kenyan specimen as R. cf. punjabicus. Simons also pointed out that specimen GSI D185, the right maxilla from the Siwaliks which Pilgrim (1915) had assigned to Dryopithecus punjabicus, was in fact a Ramapithecus. This specimen comes from the Nagri zone and has P3_M? preserved and the distal part of the canine alveolus. The alveolus indicates that the permanent canine was small and vertically implanted. The premolar and molar occlusal morphology is very similar to YPM 13799. The zygomatic process is above M! and there is a canine fossa as in the other two specimens. The maxilla is described in detail by Pilgrim (1915, p. 16-22), and figured by him (pl. 3). Remane (1954, p. 124) has compared this specimen to the Australopithecus africanus maxillary fragment from Garusi in East Africa and has shown them to be very similar. The dimensions of these three specimens are given in Table 38. The teeth of YPM 13799 in particular are mesiodistally shortened, and the mesio- distal lengths are accordingly estimates (see Chap. Il). The L/B indices of YPM 13799 are also somewhat lower than those of the other specimens. An attempt has been made to quantify one feature of occlusal morphology mentioned previously (see Table 39). The ratio of minimum to maximum buccolingual breadth across paracone and protocone has been calculated for Table 38 Measurements of upper teeth of Ramapithecus YPM GSI NMK 13799 D185 FT1272 (from cast) MDi L 7.0 5.9 p? Max L end 6.6 BuLi B 10.5 9.0 Index Max L/B x 100 Tale?! Wes 8) MDi L 720 6.9 6.8 P* BuliB 10.0 9.2 10.1 Index L/B x 100 70.0 75.0 67.3 MDi L 10.0 10.2 TO M* BuliB 10.9 10.3 10.4 Index L/B x 100 91.7 99.0 ClyPs al MDi L 10.6 10.3 11.6 M* BuLi B VIS7 10.8 11.6 Index L/B x 100 89.8 95.3 100.0 TERTIARY PONGIDAE OF EAST AFRICA 107 Table 39 Ratios (Minimum BuLi B/ Maximum BuLi B) x 100 of Ramapithecus, D, (S.) sivalensis and BM(NH) M16649 R. punjabicus D. (S..) sivalensis isch BM(NH)M16649 p* 62. 0-66. 4 46.0-50.0 (n=2) 43.9 M 67. 0-68. 5 48. 8-57.0 (n=5) 46.6 M* 66. 6-69.7 46.4-61.3 (n=5) z P4, M}, and M2, in Ramapithecus and in D. (S.) stvalensis (including BM(NH) M16649). A high ratio will reflect the fact that lingual and buccal surfaces are more vertical, a low ratio that they slope and that the occlusal fovea is relatively restricted. The molars and second premolars are morphologically clearly distinguishable in the two species. The resemblances which they do have are due to their relatively simple cusp patterns and their lack of cingula. In 1963, Simons discussed briefly various synoptic views of higher primate classification. He comments (p. 888), “Sivapithecus . . . is not easily separated from Rama- pithecus.” This presumably means that the two groups are superficially similar in some features, not that there is necessarily any special phylogenetic re- lationship between the two. In 1964, Simons wrote another important paper on Ramapithecus. In this he re-emphasized the fact that the three maxillae discussed above are, on known parts, inseparable at the species level, and that the mandible GSI D168 was not a Ramapithecus at all. He also pointed out that mandibles of R. punjabicus were in fact probably already known. In 1934, Lewis had described several species other than R. brevirostris. One of these was Bramapithecus thorpei, a species having as type specimen YPM 13814, a left horizontal ramus with M3, Ms, roots of M,, and the distal part of the distal alveolus of Py. The specimen came from the Chinji zone, and thus is probably as old as the Fort Ternan material. The molars are short, broad, and closely approximated. The distal border of My has worn a crescentic facet in the mesial border of Ms and in turn has had worn in its mesial border a facet by M,. The three molars are subequal in length. P4 was also close to M,. Clearly, the entire tooth row was compressed. The association of short teeth and a compressed tooth row has already been noted in YPM 13799. The internal contour of the mandible also indicates that the tooth row and face were short anteroposteriorly. The ramus is shallow and thick at Ms, thinner at P,. An arcuate tooth row and mandible is indicated. Lewis (1934, p. 174) noted that, “The sculpture of the crowns is highly suggestive of many human molars. . . . [This species] may very well lie near to the stem which led to the Hominidae proper.” The morphology of the molar crowns (see Lewis, 1934, pl. 1, fig. 4) contrasts with D. (S.) sivalensis. The cusp tips are relatively further apart and hence the occlusal fovea is wide. The crowns are low; the buccal cusps are lower than the lingual cusps. The hypoconulid is central on M, and M3. In combination these features make this mandible hominid-like and parallel 108 PEABODY MUSEUM BULLETIN 31 many characters of the upper teeth of R. punjabicus. The My, of “K. wickeri’” is similar to that of YPM 13814, although the tooth from East Africa is longer, just as the upper molars of NMK FT 1272 are longer than those of YPM 13799. ‘There is a slight buccal cingulum between protoconid and hypoconid in both M,’s and in the Mz of YPM 13814. Lewis also described in 1934 a new species of Dryopithecus, D. sivalensis, having as type YPM 13806, a right mandibular ramus with M, and Mz (p. 171-172, pl. 1, fig. 5). Lewis later transferred this species to Bramapithecus (1937a). The teeth are similar to those of YPM 13814 and specific distinction is not warranted (Simons, 1964; Simons and Pilbeam, 1965). These mandibles and dentitions are strongly reminiscent of shallow mandibles of Early Pleistocene hominids, for example, the Australopithecus africanus figured by Boné, 1955, and that formerly known as “Telanthropus capensis” (see Pilbeam, 1967; and Table 41). The species D. punjabicus discussed above had been established in 1910 by Pilgrim, the types being GSI D118 and 119 consisting of a portion of right horizontal ramus with Mg, and a portion of the left with M,. These were described comprehensively in 1915 by Pilgrim (p. 10-16). The tooth crowns are low, the occlusal fovea broad, and the hypoconulids central. In 1937a, Lewis transferred these mandibular fragments to Bramapithecus; they are also members of the same species as YPM 13814 and 13806. Now, in 1963 Simons had referred the maxilla, GSI D185, to R. brevirostris; the mandibles GSI D118 and 119 were transferred by Lewis to Bramapithecus. It has already been shown that these mandibles are hominid-like, particularly in ways which parallel the hominid features of R. punjabicus maxillae. That these maxillae and mandibles belong in one and the same species seems most logical, for much the same reasons that the upper and lower dentitions of D. (P.) major (Chap. VI) and D. (S.) stvalensis (Gregory, Hellman, and Lewis, 1928) were associated. By extrapolating from living hominoid species it is possible to estimate which upper teeth are likely to go with which lowers; trial occlusions can raise the probability. When only one species of a given group is present (one species of hominid maxillae and one of mandibles) the probability is still further strengthened. The lack of associated material need not prevent the correct allocation of upper and lower dentitions of a species. In fact, associated dentitions are by no means common, at least as far as fossil primates are concerned. The material described above has therefore been assigned to one species, the binomen of which is Ramapithecus punjabicus. The arguments outlined here are expanded in Simons (1964, 1965b, and 1968), Simons and Pilbeam (1965), and Pilbeam (1966). Dimensions of the mandibles and lower dentitions are given in Table 40. R. punjabicus is known therefore from East Africa and India, from deposits of Late Miocene and Early Pliocene age. It is possible that other specimens should be assigned to this species (see Simons, 1964, p. 534; Simons and Pilbeam, 1965, p. 137). A number of paleoanthropologists have accepted the synonymy of “K. wickeri” with R. punjabicus, among them Campbell (1966) and Clark (personal TERTIARY PONGIDAE OF EAST AFRICA 109 Table 40 Measurements of lower teeth and mandibles of Ramapithecus YPM YPM GSI "K. wickeri" 13814 13806 118 & 119 MDi L *10.0 MDi L eile al atid PS 8 1ELS(0) Tri B 10.4 9.6 i}5 al 95/0 Tal B VO. a 5 7/ OS 8.9 Index L/TriBx100 106.8 Ales 4 116.9 Wall MDi L collab (& Thabe al 2S 4a 3) 10.8 8.9 10.6 TalB 10.0 (sq) 10.0 D ATS ae eZ D 24.5 ZOD 24.0 a Ze 0 21.0 2052 communication and 1964). Leakey, however, has not agreed with this view. He wrote (1964, p. 49) in reply to a letter of Clark’s in Discovery of London, So far as Kenyapithecus is concerned, Sir Wilfrid states—as though it was an established fact—‘Kenyapithecus proves to be indistinguishable from Ra- mapithecus.’ I challenge this ex cathedra statement. What I believe Sir Wilfrid means is that a young scientist at Yale has recently tentatively suggested that Kenyapithecus and Ramapithecus may belong to one and the same genus. This is a very different thing from ‘proving that they are indistinguishable.’ Moreover, in order to give a semblance of likelihood to this suggestion, the Yale scientist first had to claim that the lower jaw and dentition, which the original finder attributed to Ramapithecus, must be dissociated from the maxilla which constituted the type of that genus. Having done so, he then links some mandibles of another genus, Bramapithecus, with the Ramapithecus maxilla as representing its lower dentition. Such armchair juggling may, or, of course, may not, prove to be justified, in due course, when someone finds an associated upper and lower dentition of Ramapithecus. Meanwhile, since we have a lower molar of Kenyapithecus, and since it does not at all agree with the lower molars which Lewis described as those of Ramapithecus, the words ‘proves to be indistin- guishable’ are, to say the least of it, inappropriate. I myself pointed out, in my original description, that there was some similarity between Kenyapithecus and Ramapithecus but there is no scien- tific evidence upon which to go further than that, at the moment. “Armchair juggling’, as we have noted, is a not unusual method of allocating isolated upper and lower dental material to one species. We have also noted the “Bramapithecus” mandibles are suitable for consideration as hominids, complementary to Ramapithecus maxillae, while GSI D168 (the mandible associated by Gregory, Hellman, and Lewis with Ramapithecus) is not a hominid. 110 PEABODY MUSEUM BULLETIN 31 Lewis has recently stated his belief that GSI D168 is a pongid (see Simons, 1964). Leakey’s 1962 description in fact mentioned two points of similarity between Ramapithecus and “Kenyapithecus” and one point of difference which does not exist. There is a large amount of evidence indicating synonymy, at least at the generic level. Finally, in 1967, Leakey produced a detailed statement of objections to Simons’ (1964) and Simons’ and my diagnosis (1965) of Ramapithecus and to our synonymy of the Indian and Kenyan specimens. In this paper he proposed that BM(NH) M16649 be transferred from D. (S.) sivalensis (S. africanus of Clark and Leakey, 1950 and 1951) to Kenyapithecus, this genus standing apart from Ramapithecus. His objections to the diagnosis given by Simons and me for Ramapithecus are as follows (1967, p. 156): a) It makes size a generic character in that it states ‘slightly smaller overall size than Dryopithecus and Australopithecus, but immediately refers to an exception. Moreover, should a larger species of the genus be discovered, it would automatically be excluded by this diagnosis. It must, therefore, be emphasized that size is never really valid as a generic character. b) It states that the mandible is ‘shallower’ than in Australopithecus and Dryopithecus. It does not say whether this is in relation to the overall size, or whether it is shallower only relative to the length of a tooth row. ‘This feature again makes size, for example, ‘shallower’, a generic character, which is taxonomically unsatisfactory. Several points need to be made here, the first of which is that R. punjabicus as defined in Simons (1964) and Simons and me (1965) is a monotypic genus. The diagnosis of the species therefore inevitably becomes the diagnosis of the genus, since, in the absence of other species in the genus, it cannot be known in advance which characters are “generic” and which “specific”. This leads to another point which needs to be stressed. In the study of fossil forms (as in studies of the evolution, ecology, and population genetics of living populations) it is the species to which our attention is directed. We assign individuals to species and it is the species which is most important since it is the unit which evolves. Although formal species are often arbitrary (particularly in paleontology), nevertheless an approach to true species groupings can be made, particularly using multivariate statistical techniques. In addition, genera are arbitrary groupings of species, bracketed together because they are similar. Only when species have been grouped into genera can generic characters be listed. Finally, Leakey objects to the use in diagnoses of characters to which exceptions can be found. It should be emphasized here that even with species, and particularly with genera, we are dealing with groupings or sets which are polythetic rather than monothetic (Sokal and Sneath, 1963). Polythetic sets are composed of individuals each of which possesses some, though not necessarily all, of a number of characters which define the set; one character is rarely possessed by all members. In contrast, to qualify for membership of a monothetic set each individual must possess a definite number of characters and _ all characters are found in all members of the set. It has for some time been TERTIARY PONGIDAE OF EAST AFRICA 111 realized that species, and other taxa in the hierarchy, are polythetic rather than monothetic sets, and therefore exceptions are bound to occur. To expect otherwise is to revert to typology. To clarify one point, the known mandibles of R. punjabicus from the Late Miocene and Early Pliocene of India are absolutely shallower than most members of species of Dryopithecus except some D. africanus and D. laietanus. If width be taken into account, the mandibles are relatively shallower than all species of Dryopithecus (see Simons, 1964, fig. 1, and p. 531-532; and Table 41). The relative thickness has been calculated using mandibular thickness (M; ‘T) and depth (M; D) measured at M3. Table 41 shows clearly that, out of 19 mandibular specimens of Dryo- pithecus, only one is more robust at the level of M, than the least robust of R. punjabicus. R. punjabicus is similar to Early Pleistocene A. africanus, although a little less robust. Table 41 Ratio (M,T/ M, D) x 100 of fossil hominids and pongids, Gorilla and Pan _ n x OR Australopithecus africanus 3 89.8 87.5-92.9 R. _punjabicus 3 83.0 736 2-O0e 4 D. (P.) africanus 4 66.1 5300-7065 D,. (P.) nyanzae 3 61.6 57. 1-68.0 D. (2.) major 2 68.2 62.5-73.9 D. (S.) silvalensis 3 74.8 69.5-82.5 D. fontani 3 65.7 61. 0-70. 6 D,. laietanus 4 67.9 62° 6=7 719 P. troglodytes (males) 10 65.7 58. 0-74.5 P. troglodytes (females) 10 59.0 49.7-66.5 G. gorilla (males) 10 66.8 59. 4-76. 2 G. gorilla (females) 10 7055 59. 7-82. 0 Pongo pygmaeus 6 67.4 55 Wa7ha5 0) Leakey’s objections to the diagnosis continue (1967, p. 156), c) It goes on to say ‘shorter face’, but because no complete face of Dryopithecus (other than in the Proconsul group) has been found and certainly no complete face from gnathion to nasion in Ramapithecus, this seems to be a most unwise diagnostic character to attribute to the genus, in the present state of our knowledge. It is obviously possible to misinterpret the diagnosis given by Simons and me (1965, p. 136). The shortness referred to in “shorter face’ is antero- posterior and not superoinferior, as was inferred by Lewis when he named YPM 13799 “brevirostris.” Finally, (Leakey, 1967, p. 156), d) The character ‘incisors and canines reduced in relation to cheek teeth when compared with Dryopithecus’ may be valid for Asia but it is not true 12 PEABODY MUSEUM BULLETIN 31 for specimens from East Africa which Simons and Pilbeam wish to include in Ramapithecus. While it may be true in respect of Ramapithecus pun- jabicus (brevirostris) (if indeed incisors of this genus and species are known in Asia, of which I am not aware) . . . it is certainly not true... in respect of comparisons of the incisors as between Kenyapithecus and the Proconsul group, which Simons and Pilbeam insist on including in the genus. The canine of YPM 13799 must have been small both absolutely and relatively (as is the canine of “K. wickeri” which Leakey, 1962, p. 694 noted). In 1934 (p. 156), Lewis wrote of the incisors: “I2-The root is present in the alveolus, but the crown has been lost. A small incisor close to the hominid type is indicated. —Most of the alveolus has been preserved, and indicates that the size of the two incisors was approximately equal. The roots indicate that the incisors more nearly approached the vertical than is the case in the Simiidae known hitherto; the attendant slight prognathism is hominid.” To this account can be added the following brief comments. The lateral incisor root was no more than 16 mm long from cervix to tip. The root is broken above the cervix and is no more than 4 mm long mesiodistally and 5.5 mm broad labiolingually. The central incisor alveolus is preserved distally; the tooth root would have been no longer from cervix to tip than that of the lateral incisor. Such a root is both absolutely and relatively very small. The crown of the central incisor could hardly have been more than 9 mm long mesiodistally and 6.5 mm broad labiolingually. These incisors are clearly small in absolute terms. The only pongid con- temporary with YPM 13799 for which I have incisor measurements are for GSI D196, the palate originally described by Pilgrim as Sivapithecus orientalis (1927, p. 549), transferred by Lewis (1937a, p. 144) to S. indicus, and by Simons and me (1965, p. 125) to D. (S.) indicus; I now believe this to be a male D. (S.) sivalensis. The cheek teeth and lateral incisor root are preserved, as is the alveolus for the central incisor. An isolated upper central incisor of a large pongid is preserved in the Yale collections (YPM 16919). This specimen has only recently come to light, and has yet to be described. It fits the alveolus of GSI D196 and has been used as the incisor of D196 in the calculations. It may, in fact, represent the same individual as GSI D196. In order to determine the relative size of the incisors, the crown area of the upper central incisors has been calculated (product of length and breadth) and expressed as a ratio of the sum of the crown areas of M! and M2? (Table 42). The first two molars only were chosen since third molar crowns are not present in YPM 13799 and the palate of “K. wickeri’. The central incisor 2 Table 42 Ratio (crown area of I}/ crown area M*+ M )x 100 in Dryopithecus and Ramapithecus R. .punjabicus YPM 13799 *29.9 D. (S.) sivalensis GSI D196/YPM 16919 *32.2 D. (P.) major UMP 62-11 29.5 D, (P.) nyanzae NMK 5, CMH3 212 D. (P.) africanus 'NMK R1948,50 24.6 TERTIARY PONGIDAE OF EAST AFRICA 113 crown area has been estimated for YPM 13799 and this is clearly only a very rough approximation. If pongids contemporary with R. punjabicus are considered, only one spe- cimen is available, the composite of D. (S.) sivalensis. Here the incisor is relatively larger than in R. punjabicus. The palate of D. (P.) major is from older, early Miocene deposits and the ratio is slightly lower than in GSI D196, though still higher than R. punjabicus. The values for D. (P.) nyanzae and D. (P.) africanus are slightly below that for R. punjabicus; these in- dividuals are geologically considerably older than the Indian form. If a trend towards incisor hypertrophy has occurred in the pongids then lower values for older species might be expected. Since relatively small incisors are primitive for Pongidae, it is hardly surprising that early Miocene apes and late Miocene/early Pliocene hominids should both have relatively small anterior teeth. Seemingly, the Hominidae have retained the small incisors of their earlier Tertiary ancestors. (It should be remembered here that but a single specimen is recorded for each species noted above and the range of variation of incisor size within the species is unknown.) In 1967 (p. 157) Leakey described an upper central incisor from the same level at Fort Ternan as the maxilla of “K. wickeri’. Leakey believes that the incisor and maxilla belong to the same species. ‘The dimensions of the incisor are: MDi L, 10.0 mm, and LaLi B, 6.5 mm, giving a crown area of 65 mm?. The ratio of crown areas (if the incisor and maxillae are treated as coming from a single individual) is 26.3, 0.8 units above the estimate for YPM 13799 and below the values for the contemporary GSI D196 and the slightly earlier UMP 62-11. The ratio is greater than that for one D. (P.) nyanzae and 1.6 units above that of one D. (P.) africanus. Both these specimens are 5 million or more years older than East African R. punjabicus. Figure 31 compares the maxillary dentitions of D. (P.) africanus, a pygmy chimpanzee Pan troglodytes, and the Fort Ternan Ramapithecus. All three have been brought to the same M1 length. The incisor from Fort Ternan is much smaller than that of P. troglodytes and is less hypsodont than geologically older D. (P.) africanus. So, considered in terms of crown height, the incisors of Ramapithecus, like the canines, are relatively smaller than those of Miocene to Recent apes. There is little difference between the East African and Indian specimens of R. punjabicus as far as incisor size is concerned, and there are no other significant differences, as has already been shown. The claim that they do not represent the same genus or species cannot be sub- stantiated in this way. The incisor of Ramapithecus from Fort Ternan is of interest, since it is the only complete one known. The specimen is described by Leakey (1967, p. 57): This tooth is remarkably like that of Homo and differs in a number of characters from the corresponding teeth of Proconsul. . . . Moreover, it has characters in common with the australopithecines, in particular in the cross- section of the root where it meets the crown. In the upper central incisors of Proconsul, the root at the junction with the crown is approximately trihedral, and has a greater diameter from the labial to the lingual aspect 114 PEABODY MUSEUM BULLETIN 31 than from side to side. In the Kenyapithecus wickeri specimen, the cross- section is not trihedral but much more oval with the maximum diameter from side to side, while the labial-lingual diameter is reduced. The height of the crown of this tooth, which is practically unworn, is less than the width, and it is more compressed labio-lingually compared with the cor- responding teeth of Proconsul. The region near the cutting edge is very thin providing a fine chisel edge very different from the upper incisors of Proconsul, where a marked medial thickening extends almost to the top of the crown. The following dimensions are given (1967, p. 157), Crown width (MDi L), 10.0 mm, and Crown height (La H), 10.25 mm. Diameters of root at junction with crown, Bilateral (MDi), 7.5 mm, and Labio-lingual, 6.5 mm. These dimensions do not agree with the statement that “the height of the crown ... is less than the width.” For comparison; dimensions for the central incisors of D. (P.) nyanzae and africanus are given in Appendix 2; their MDi lengths are lower, although the LaLi breadths are relatively closer to the Fort ‘Ternan specimen. In D. (P.) major (UMP 62-11) the crown dimensions of I' are, MDi L, 10.8 mm, and LaLi B, 9.1 mm. Unfortunately, the crown is so worn that features of the lingual surface cannot be seen. Like R. punjabicus from Fort Ternan, the MDi diameter of the root at the cervix (9.4 mm) exceeds the labiolingual diameter (8.7 mm). In this feature D. (P.) major differs from the two smaller species. At the cervix, the root of the incisor of D. (S.) sivalensis (YPM 16919) is 8.6 mm long mesiodistally and 8.1 mm broad labiolingually. Presumably the rela- tive proportions at the cervix alter as the mesiodistal crown diameter increases. The central incisor from Fort Ternan resembles those of hominids in its “shovel-shaped” morphology. However, not all hominids have central incisors in which the lingual surface lacks a median pillar. Robinson (1956, fig. 6a) has illustrated such a structure in Australopithecus robustus. In the only speci- mens of A. africanus (1956, fig. 7f) two small processes run centrally down the lingual surface from the gingival eminence. The central incisor of D. (S.) stvalensis from the Chinji (YPM 16919) was mildly shovel-shaped when un- worn, although the lingual mesial and distal marginal ridges have almost dis- appeared through wear. There is a pronounced lingual bulging just below the cervix; there is no medial lingual thickening running down the crown. The central incisors of the subfossil orangutan, Pongo pygmaeus palaeosumatrensis, illustrated by Hooijer, show variations in crown lingual surface morphology. Some are moderately shovel-shaped with no central pillar (Hooijer, 1948, pl. 1, figs. 6 and 11) while others have these well developed (ibid., pl. 1, figs. 7, 8, 9, and 14). Finally, the composite D. (P.) nyanzae figured by MacInnes (1943, pl. 24) has incisors similar to those described by Leakey for Ramapithecus. Accord- ingly, the presence or absence of mesial, distal, or central ridges is variable in species of Hominoidea. In his description of the East African R. punjabicus, Leakey wrote (1962, p. 695), “It must suffice to point out that, while it clearly represents a creature related, to some extent, to the Lower Miocene species, which Leakey and Le Gros Clark called Sivapithecus africanus, it is equally clearly not a true Sivapithe- TERTIARY PONGIDAE OF EAST AFRICA LS cus, as may be seen by the morphology of the upper canine and fourth upper pre-molar, and also by the presence of the canine fossa and the position of the root of the malar element of the malar maxillary process.” He concluded (p. 696), “when . . . [the Fort ‘Ternan specimen] is allocated to a family, Sivapithecus africanus will have to join it there and be removed from the Pongidae.” In the preceding paragraphs he had given no details of the similarities between “K. wickeri” and “S. africanus’? which might necessitate the reclassification of the latter. Before continuing with this account it is important to restate the fact that the material from Fort Ternan cannot be separated taxonomically from Indian Ramapithecus and that the discovery of the upper central incisor does not alter this conclusion. The generic nomen “Kenyapithecus” is therefore no longer available. In his most recent papers, Leakey (1967 and 1968) has proposed that new material be added to the hypodigm of the species proposed by Clark and Leakey in 1950, S. africanus. Leakey regards this material as being distinct from D. (S.) sivalensis. The original holotype of ‘‘S. africanus’ has already been men- tioned and is BM(NH) M16649 from site R106, Rusinga (Fig. 28). The zygo- matic process of this maxilla is situated above M}? and is closer to the alveolar border than in R. punjabicus. This is probably a primitive character. It is most unlikely that the canine fossa was better developed than in other species of Dryopithecus, particularly in the presence of a presumed large canine jugum. The lowness of insertion of the zygomatic process seems to be a primitive character and is found in D. (P.) major, D. (P.) africanus, and, as far as can be seen in the holotype, D. (P.) nyanzae. In D. (Proconsul) the process is set either above the distal moiety of M! or the mesial part of M?; the process is more distal in D. (P.) major UMP 62-11, possibly because this specimen has a sizeable anterior dentition and face, and also possibly because it is younger and therefore more evolved towards the condition seen in G. gorilla. A more forward position of the zygomatic process may be a primitive character for some Pongidae; without earlier material this cannot be determined. It is not confined to Hominidae, as Straus (1963, p. 149) has shown in his study of Oreopithecus, and the process is often situated above M1! or M? in Pongo pygmaeus. The condition seen in BM(NH) M16649 resembles that of R. pun- jabicus, although the morphologies of the lateral maxillary areas above the cheek teeth contrast strongly in the two groups. In BM(NH) M16649 there is none of the “tucking in” to be observed above the buccal roots of the premolars as in R. punjabicus. The way in which the zygomatic process merges with the alveolar area is also different. The teeth of BM(NH) M16649 have been described in detail by Clark and Leakey (1951, p. 63), quoted above. All three preserved cheek teeth (P?—M?) have marked buccal and lingual slopes (see Table 39, p. 107). The cusps and crests are rather rounded; this roundness is increased by wear. It is not possible to tell whether or not this character is typical or otherwise for the population from which this individual is sampled. The first premolar has two well-developed buccal roots (Fig. 31). The mor- phology of the buccal surface is very similar to that of D. (P.) major, UMP 62-11 (Fig. 18), particularly in the slope of the mesial buccal ridge and the 116 PEABODY MUSEUM BULLETIN 31 superior extension of the enamel margin in the mesial half of the buccal surface. The mesial extension of the mesiobuccal corner of the crown has already been noted for UMP 62-11, GSI D299/300 (Gregory, Hellman, and Lewis, 1938, pl. 1, fig. 2a) and YPM 13837 (ibid., pl. 1, fig. 10a, pl. 7, fig. I). The condition is similar in BM(NH) M16649 and the mesial border is concave mesially. The occlusal slopes of the protocone and paracone are both steep and a central sulcus terminates the ridges running from these cusps towards the center of the occlusal surface. In these features it resembles D. (P.) major and D. (S.) sivalensis P®’s and contrasts strongly with R. punjabicus. The morphology of P* is generally similar to P?, although the mesiobuccal extension is not present. It contrasts with that of R. punjabicus, as Leakey (1962) noted, and compares with those described for D. (S.) stvalensis: GSI D299/300 (Gregory, Hellman, and Lewis, 1938, pl. 1, fig. 3a), GSI D1, and GSI 18065 (Prasad, 1964, pl. 20, fig. 3b). There is a slight distolingual cingulum. The first molar is simple and low crowned. As in P*, the buccal and lingual surface slope so that the cusp tips are relatively approximated (see Table 39, p- 107). It is morphologically similar to D. (S.) stvalensis first molars: GSI D299/ 300, BM(NH) M13365, and YPM 13834 (Fig. 30). A protoconule is present; the lingual cingulum is present only at the mesiolingual corner, as in YPM 13834. Figure 28 shows that the enamel is heaped-up on the lingual surfaces of all three preserved cheek teeth of BM(NH) M16649. This suggests either that other individuals within this species have cingula, or that individuals in preced- ing species had cingula, or both. The crown is somewhat lower than in Indian D. (S.) sivalensis, possibly a primitive feature. Remane (1965, p. 281) has written of this specimen, “Die Zuordnung zu dem asiatischen Genus Sivapithecus ist provisorisch, die Reste kénnten gut in die Gattung Gorilla oder vielleicht auch als Extremvariante zu Proconsul major gestellt werden. Aus diesem Rest diirfen daher vorlaufig noch nicht palaonto- logisch-tiergeographische Schliisse gezogen werden, etwa in dem Sinne: ‘Die Gat- tung S:vapithecus erscheint zuerst im Miozan Afrikas.’ ” Since a complete upper dentition of D. (P.) major, UMP 62-11, is now known, it can be seen that this specimen, BM(NH) M16649, does not represent any species in the subgenus (Proconsul). However, Remane’s remark is reasonable in that it does make the point that BM(NH) M16649 has many pongid-like features. One important pongid feature of BM(NH) M16649 which was first mentioned in 1965 by Simons and me (p. 113) is the presence above the roots of P? of the distoinferior part of a large canine alveolus curving back over the premolars (Fig. 28). The breadth of the mesial border of the first premolar also indicates that the canine would have been large. In summary, this palate shows many pongid features and no unequivocal hominid characters. If this is a hominid ancestral to R. punjabicus, the number of similarities are very few; the anterior position of the zygomatic process is one, although this is not necessarily a hominid character. ‘The number of resem- blances to Late Miocene and Early Pliocene D. (S.) stvalensis, however, are more numerous. Dimensions for BM(NH) M16649 have been given by MacInnes (1943, p. 168) and Clark and Leakey (1951, p. 67). Leakey gives two different sets of TERTIARY PONGIDAE OF EAST AFRICA 117 measurements (1967, tables 1 and 2). Measurements as defined in Chapter II are listed in ‘Table 43. A number of other maxillary specimens are referred by Leakey to the same species as BM(NH) M16649. ‘These are left (NMK 300,111) and right (NMK 274,52) maxillae of one individual from Songhor, the left with P* and roots of C1 and P%, the right with P?-M!? and C! root. These specimens resemble BM(NH) M16649 in general features; NMK 300,111 has a canine fossa, although this is no better developed than in BM(NH) M16647 (D. nyanzae), GSI D196 (D. stvalensis), or in many Pongo pygmaeus. The palate of NMK 300,111 is shallow, like that of BM(NH) M16649. The canine root is preserved in both specimens; it is said to be set vertically in the maxilla of NMK 300,111, although no more so than in GSI D196, for example. The dimensions of the canine of NMK 274,52 would probably have been *11.9 mm (Max L) by *9.0 mm (Trans B) at the Table 43 Measurements of East African Pongidae, non-Dryopithecus_ (Proconsul) BM(NH) NMK NMK NMK NMK NMK M16649 300, 274, 333, 748, 1377 DG.) 11 52 404 vA sivalensis MDi L 7a8 6.2 6.9 7.8- 8.0 (2)4 3 Max L 9.0 8.3 8.6 8. oe 89-9) 4. (2) S Bull B 1255 10.7 115 — 1OL0 4, Ti 0-14) Index Max 1/Bx100 72.0 77.5 74.8 85.0 80.9-82.5 (2) MDi L 8.1 7.8 7.0 als eo 0% Fla 728 (3) P* Buli B 123° leg de? 11.7 10:02 ° aeS=12/08(4) Index L/B x 100 6500) 6702 6205 64.9) 60.0), 16056-6615 06) MDi L 10.5 9.4 10.3-11.3 (7) M* Buli B 11.8 10.9 U1, 419267) Index 1/B x 100 89.0 86.2 78. 6-93.4 (7) ,MDi L **11,0 10.07 10.8-13.3 (9) M’ Buli B **12.0 11,5 12.6-15.4 (9) Index I1/Bx 100 **91.6 87.0 80. 6-95.4 (9) lNumber of specimens measured 2From Leakey (1967) cervical border. The long axis was mesiodistal, as in pongids. For seven specimens of D. (S.) sivalensis from the Siwaliks the ranges for these dimensions are 10.5 to 16.6 mm and 8.7 to 12.5 mm. Two further maxillary fragments from Songhor have been assigned with the two described above. A right maxilla (NMK 748,7) contains P? and P*, and a left fragment (NMK 1377) also contains these two teeth; a right M2 (NMK 333,404) from Songhor has also been included. Of these, NMK 300,111 and 274,52 were originally classified as D. (P.) nyanzae (Clark and Leakey, 1951, p. 53), and NMK 333,404 as D. (P.) africanus (Clark and Leakey, 1951, p. 36). Such fragmentary specimens are very difficult to assign correctly. One interesting contrast between R. punjabicus and Dryopithecus species is in the proportions of the premolar lengths. In the hominid species, the two premolars generally have closely similar lengths; in pongids P? is longer than 118 PEABODY MUSEUM BULLETIN 31 P+ mainly because of the mesial extension of the buccal face of P?. The ratios Max L P*/MDi L P+ x 100 are given in Table 44 below. The ratio in three specimens of R. punjabicus is lower than in earlier or contemporary pongids. Leakey (1967, p. 160-161) has transferred an upper right central incisor (NMK 9, CMH9) from D. (P.) nyanzae (Clark and Leakey, 1951, p. 39-40) to his proposed new species. The specimen comes, according to Clark and Leakey (p. 39) from “Rusinga’”’, no site being given. No locality data is available in the National Museum of Kenya Catalogue. Leakey (1967) states on p. 158 that the specimen comes from site R106 although in the description on p. 160-161 no exact locality is given. The tooth is large, dimensions 10 x 7.25 (p. 161) or 10.0 x 7.5 (Table 2, p. 162), and similar in morphology to the one which he referred Table 44. Ratio (MaxL P/ MDi L p*) x 100 of fossil species, Gorilla and Pan n X OR D, (P.) africanus Z 110. 9-129.2 D. (P,) nyanzae 3 110, 2-120. 7 D, (.) major 1 134, 1 D, (S.) sivalensis ul 118.6 BM (NH)M16649 ahaha Jk NMK 274,52 PS. 9 NMK 748,7 113.1 NMK 1377 141.7 G. gorilla (males) 20 110.9 103. 5-122.2 G. gorilla (females) 20 AOR RY ¢ 91. 8-110.8 P. troglodytes (males) 14 114.5 102. 6-130. 0 P. troglodytes (females) a2 he ys 103. 0-121.4 RB. punjabicus 3 100. 0-107.3 A. robustus_ 4 90.0- 94.3 A. africanus 5 88. 9-120.9 "Ke. wickeri® = **100.0 to R. punjabicus. It should be emphasized again that the absence of a median lingual pillar does not prevent this incisor from being pongid, nor make it hominid; it is quite possible that incisors of female Siwalik D. (S.) sivalensis would have had similar dimensions and a similar morphology. Leakey has also included in his proposed species three mandibular fragments, all of them containing part of the symphyseal area, and a more complete mandi- ble, NMK Rs 1967, 394 (Leakey, 1968). These fragmentary specimens are so in- complete as to render their correct taxonomic assignment almost impossible. A piece of right ramus (NMK 71, CMH142) from site R106, Rusinga, which comes from the same site as BM(NH) M16649 (Leakey, 1967, p. 158, figs. 3a and 4a) is also listed by Leakey. This specimen was tentatively referred by Clark and Leakey (1951, p. 60) to D. (P.) major. The fragment is sectioned mesially TERTIARY PONGIDAE OF EAST AFRICA 119 just mesial to the right C, and distally midway through M,. The roots of C, to P, are present and the mesial root of M,. The specimen differs from D. (P.) major in a number of features. Leakey lists these (1967, p. 158). The first one is: a) At the level of P,M, the depth of the corpus is only very slightly less than that seen in Proconsul major, but the thickness of the mandibular [stc], at this point, is totally different; the corresponding measurements at P4M, are: Proconsul major Depth 41 mm Thickness 22 mm Kenyapithecus africanus Depth 39.5 mm Thickness 12 mm My own measurements plus some from Leakey are in the following table. Table 45 Mandibular measurements of East Africanearly Miocene mandibles D. (P. )major NMK NMK BM (NH) NMK71 Rs 1967,394 190,1 M4086 CMH142 (fom lealeyses) D 41.0 36.5 38.0 34.0 ae: 21.0 17.0 Lee 13.5 * IndexYD 5) 5 46.5 40.8 39.7 x 100 A few sets of measurements are available for D. (S.) stvalensis from the Siwaliks and for R. punjabicus (Table 46). Table 46 Mandibular measurements of R. punjabicus and D. sivalensis R. punjabicus D. (S.) sivalensis YPM 13814 YPM 13811 GSI D197 GSI D199 D 2275.9 26.0 S55 5) = Dee ea WAG) USS 14.5 13.0 peer adaaa 44,2 40. 8 = x 100 As far as other Early Miocene mandibles are concerned, NMK 71, CMH142 is relatively thin and gracile, as can be seen from the thickness-depth ratio. However, mandibles of D. (S.) stvalensis from the Silwaliks have in the region of P, thicknesses as low as 11.5 mm. The lowest recorded index is that of GSI D197 (“S. himalayensis”’ of Pilgrim, 1927). There is no evidence to indicate that this non-D. (Proconsul) feature is a hominid characteristic. Leakey continues (p. 158): b) In the known mandibles of Proconsul major there is a clearly defined diastema between the lower 3rd premolar and the lower canine; in this speci- men the anterior lingual root of the lower 3rd premolar is set far forward and extends well beyond the posterior rim of the alveolus of the canine, and there is no diastema. . c) In Proconsul major (as also in both the other species of the genus) the furthest posterior projection of the symphyseal region is set high in the mid-line and extends well back, so that its limit is almost in line with the lower 4th premolar. In this specimen the most backward part of the symphy- sis is not further back than the level of the front of the 3rd premolar, and may be even further forward. 120 PEABODY MUSEUM BULLETIN 31 d) In Proconsul major the inner wall of the corpus, in the region of the premolars, slopes slightly inwards, and the corpus itself is very thick. In this specimen the outer wall is at first straight, then turns slightly outwards. e) The area of the root of the canine in Proconsul major is marked by a strong surface swelling of the anterior face of the mandible, giving a clear line of demarcation between the “chin” region and the lateral wall of the corpus. In this specimen, there is no such swelling and the root of the canine is much less massive. f) The roots of the lower canines are more laterally compressed than in any Proconsul, and are orientated more antero-posteriorly and less trans- versely than in any species of Proconsul or Pongid. g) The whole mandibular structure is gracile even though the corpus is very deep (39.5 mm). h) The mental foramen lies relatively low on the corpus, and is situated beneath the 4th premolar, instead of below the 3rd premolar as in Proconsul major. The following comments can be made. The specimen is not a D. (Proconsul). However, this does not automatically make it a hominid. For point b, a similar compression of the anterior part of the tooth row can be seen in some specimens of D. (S.) sivalensis (for example YPM 13811, see Lewis, 1934, pl. 11, fig. 1c), and also in some specimens of both sexes of Pongo pygmaeus. ‘The symphyseal region, point c in Indian D. (S.) stvalensis, has a variable simian shelf. The most posterior part of the symphyseal midline may be as far forward as the distal border of P3. The superior transverse torus is set farther forward than this. If it is assumed that ancestors of D. (S.) stvalensis and D. fontani (a closely re- lated species from the later Miocene and early Pliocene of Europe), were gener- ally without posteriorly projecting inferior transverse tori, some of them may have had a symphyseal morphology like that of NMK 71, CMH142. Point d also applies to certain D. (S.) stvalensis specimens from the Siwaliks, as does e. Point e applies to YPM 13811, as well as to AMNH 19411 (Gregory and Hellman, 1926, p. 22). Point f is certainly not a hominid feature. It applies to D. fontazni. Finally, the depth of the mental foramen below the alveolar border is very vari- able in pongid species and is of dubious taxonomic value. If this depth be ex- pressed as a percentage of total rameal depth at the same point, the value of the ratio varies from 54.3 to 72.8 in a sample of 40 G. gorilla and from 59.2 to 73.9 in 26 P. troglodytes. Although many of the features present in NMK 71, CMH142 also occur in specimens of D. (S.) stvalensis from the Silwaliks they do not occur in this com- bination in large males (NMK 71, CMH142 is male according to Leakey). How- ever, this specimen is very similar to D. fontant, particularly the type specimen, Paris Museum No. AC 36 (see Fig. 30). The roots of right C, to P, are retained in NMK 71, CMH142 and it is pos- sible to obtain estimates, albeit poor ones, for some tooth dimensions. ‘They are: for C,, Max L. = **14 mm, Trans B= **11 mm; for Ps; Max E,— 4*12 ium Trans B = **7 mm. The canine therefore was large, the first premolar roots were long, and, presumably, the crown was sectorial. Leakey also describes another mandibular fragment, NMK 227,276, from the TERTIARY PONGIDAE OF EAST AFRICA 121 surface at Rusinga (1967, p. 158-159, figs. 3b, 4b, 5e). This specimen was de- scribed originally as D. (P.) nyanzae (Clark, 1952, p. 276) and may well be a member of that taxon. The specimen consists of the symphyseal region contain- ing the roots of left P; to right P3. Again, there is overlap between C, and Ps. The anterior surface of the symphysis is almost vertical in its superior half (Leakey, 1967, fig. 3b). In his figure 5e the section appears not to be oriented quite correctly, the infradentale being rotated posteriorly and gnathion ante- riorly. There is no simian shelf; the symphysis is not as thick as in D. (P.) major. Symphyseal length is *37 mm and thickness *19.5 mm, giving a thickness-length ratio of *52.7; three D. (P.) major varied from 53.8 to 57.3. According to Leakey (1967, p. 159), the symphyseal cross section “recalls that of the more primitive members of the genus Homo and of some australo- pithecines.” Individuals of Australopithecus species, however, generally have projecting inferior tori (Scott, 1963). The similarity of NMK 227,276 to Pleis- tocene hominids may not imply a phylogenetic relationship, particularly since 17 million years separates this specimen and the Olduvai Gorge specimen il- lustrated by Leakey (Hominid 13, see 1967, fig. 5a). A number of symphyses of D. (S.) sivalensis from India are known, the one which preserves the midline most accurately is AMNH 19411 (Gregory and Hellman, 1926, p. 22, fig. 11). The main axis of the symphysis is probably somewhat less vertical than in NMK 227,276; there is no simian shelf, but instead ‘“‘a nearly vertical area for the in- sertion of the digastric muscles” (Gregory and Hellman, 1926, p. 24). This specimen is from the Lower Chinji zone and is thus one of the oldest Siwalik pongids. Measurements of NMK 227,276 and AMNH 19411 are compared in fable 47. Table 47 Symphyseal measurements of Dryonithecus NMK 227,276 AMNH 19411 Symphysis D *37 43.0 He *19.5 cath 7/5 te: Index T/D x 100 Las Ac If *41,4 I B *18.7 *18.5 Ge 33 *35.0 X 3:51.95 The third mandibular fragment is NMK 417 from Songhor (Leakey, 1967, p- 160, figs. 4c and 5f). According to Leakey, it is very much smaller than NMK 71, CMH142 and NMK 227,276. Symphyseal dimensions are approximately 26 mm by 13 mm. This is almost certainly not conspecific with either NMK 71, CMH 142 or NMK 227,276. It probably represents D. (P.) africanus. The more complete mandible described by Leakey (1968) again shows many differences from D. (Proconsul) species. However, these differences do not in- dicate hominid affinities. The symphyseal section resembles D. (Sivapithecus); the inferior transverse torus projects posteriorly a little more than the superior torus [see Leakey, 1968, Fig. 1b(i)]. The canines are large and project above the occlusal plane of the cheek teeth. The P is sectorial and unicuspid. The long axes of the cheek tooth rows are concave buccally, rather than con- vex as in hominids. The contour of the dental arcade resembles that of primitive hominoids (Aegyptopithecus, Aeolopithecus) and is probably therefore a prim- 122 PEABODY MUSEUM BULLETIN 31 itive trait. There are no features of the mandible which are specifically homi- nid, and none which would indicate any special affinity with Ramapithecus. This mandible has been assigned by Leakey to “K. africanus,” although the association of the mandible with the type maxilla, BM(NH) M16649, is by no means certain. No associated maxillary and mandibular specimens of “Kenya- pithecus africanus” are known; consequently the mandibles have been referred to the same species as BM(NH) M16649 by “juggling.” The mandible and lower dentition, as far as it is preserved, shows pongid features—large canines and elongate first premolars. The tooth row is compressed, although this feature can be observed in some D. (S.) sivalensis and in D. (P.) major (Fig. 18), as well as in dentitions of R. punjabicus. Symphyseal morphology is variable, as in D. (Proconsul), D. (S.) sivalensis, and D. fontani, and resembles the Eurasian species of Dryopithecus. The maxillae also show pongid features, particularly in the size of the canine and the morphology of the first premolar; these characters are probably primi- tive. Features which could indicate hominid affinities are the canine fossa and the relatively anterior position of the zygomatic process. Canine fossae of this sort can, however, be observed in some D. (Proconsul) and D. (S.) sivalensis, and occasionally in living pongids, particularly the orangutan. As noted above, the relatively anterior position of the zygomatic process can also be observed in Pongo pygmaeus as well as in D (S.) sivalensis and is another reflection of facial shortness. In his discussion, Leakey (1967, p. 161-162) states, It is, however, important to bear in mind that the differences between Sivapithecus indicus and Sivapithecus sivalensis appear to be marked, and that what is known as Sivapithecus sivalensis is possibly not a Sivapithecus at all. There seems to be no doubt that the species Sivapithecus sivalensis in India may stand much closer to the genus Ramapithecus than it does either to Sivapithecus indicus or to the true European Dryopithecus speci- mens, or even to... Proconsul. ... This can be clearly established by examin- ing the morphology of the molar and premolar teeth. Leakey illustrates his article with a photograph of casts of the C!—M§ series of D. (S.) stvalensis, GSI D299/300 (1967, fig. 6c). The excellent illustrations given by Gregory, Hellman, and Lewis (1938), Gregory and Hellman (1926), Lewis (1934), and Prasad (1964) show that, as far as the upper dentitions are concerned, most “S” sivalensis and “S. indicus’ are so similar as to constitute a single species. ‘This species is no more variable than either subfossil or living Pongo pygmaeus (Pilbeam, in preparation). D. (S.) sivalensis is no more similar to R. punjabicus than is “K. africanus” from East Africa. They are super- ficially similar in that their cingula are either very reduced or totally absent. Leakey (1967) concludes that the synonymy of Kenyapithecus with Rama- pithecus and of Sivapithcus africanus with D. (S.) stvalensis must be rejected for the following reason (p. 162); ‘““The diagnosis of Ramapithecus which has been given by Simons and Pilbeam states that, in the genus Ramapithecus the mandible is ‘shallower’ than in Dryopithecus and Australopithecus. ‘This is certainly not true of Kenyapithecus africanus, where we have a mandibular fragment which is 39.5mm. deep at the level of the 4th premolar.” TERTIARY PONGIDAE OF EAST AFRICA 123 This statement involves some peculiar logic. The diagnosis of R. punjabicus, which is the diagnosis of a species, covered a Late Miocene and Early Pliocene group. Three mandibles are known which can be assigned to this species; these are all three shallow, both absolutely and relatively (see Table 41, p. 111). Al- though this species may contain individuals with deeper mandibles, this cannot be determined until more individuals are known. “Kenyapithecus africanus” is a different species and is from the Early Miocene. At least one of the three mandibles referred to this species (NMK 417) is as shallow as any of those re- ferred to R. punjabicus, which invalidates the objection. Since the synonymy of “K. wickeri’ with R. punjabicus has already been demonstrated, the important point is to establish the relationship of “K. africa- nus” to Ramapithecus; is it ancestral or not? If it is ancestral, ought they to be placed in one genus? The exact relationship between the Early Miocene and Pongidae discussed in this chapter and Late Miocene D. (S.) sivalensis and D. fontani cannot yet be accurately established. An African origin for Eurasian Dryopithecus is to be ex- pected and some of these Early Miocene specimens make morphologically suit- able ancestors. Until more complete material from the Early, Middle, and Late Miocene has been recovered, however, the solution to this problem, and the correct taxonomic allocation of the East African material, must remain open. Simons and I (1965) assigned some of the specimens to D. (S.) sivalensis, at that time believing the East African and Indian samples to be contemporane- ous. This is now known to be unlikely, the Kenyan material probably being latest Early Miocene in age (see Table 1, p. 17). There are a number of mor- phological differences too, particularly in the mandibles. At the moment these pongids from East Africa are perhaps best described as Dryopithecus sp. indet. Phylogenetic relationship with R. punjabicus is improbable; a more probable association is with D. (S.) sivalensis. CHAPTER IX. PONGID EVOLUTION: DISCUSSION AND CONCLUSIONS AFRICAN DRYOPITHECINES The raison d@etre of this study is the pongid material from Uganda, described in Chapters IV and V. It has been shown that this material is unlikely to be derived from more than a single species, Dryopithecus (Proconsul) major. The known fossil remains of this species show many features reminiscent of Gorilla gorilla, and thus the species is probably ancestral to the living ape. Comparison of the Ugandan material with that from Songhor and Koru in Kenya has demonstrated that this species is also represented at these sites. Its earliest known occurrence in East African sites is about 20 million years ago, and it probably persisted until at least 16 million years ago. This species, like G. gorilla, shows-marked sexual size dimorphism; the males are characterized by large, projecting canines and relatively prognathous faces. The dentition exhibits a number of features antic- ipating those to be seen in gorillas, particularly in the morphology of canines and incisors. Some of the cheek teeth, however, show trends towards the lopho- donty and hypsodonty characteristic of G. gorilla. The teeth also show many resemblances to other contemporary species of African Dryopithecus. D. (P.) major retains cingula on upper and lower molars. Of the three living pongid species, G. gorilla retains molar cingula most frequently (Korenhof, 1960; Frisch, 1965). Dental and cranial changes in this lineage since the Miocene appear to have involved a general increase in size, a small differential hypertrophy of the anterior cheek teeth, increased lophodonty and hypsodonty of the cheek teeth, the loss of cingula, the development of long maxillary alveolar processes surrounding a deep palate, increased facial prognathism, the broadening of the anterior mandibular region, and the appearance of the simian shelf. The postcranial material of D. (P.) major from East Africa shows that this species was approximately chimpanzee-sized, although somewhat more lightly built and active than the living species. The Miocene form was probably an arboreal quadrupedal form, perhaps with some knuckle-walking behavior. Mor- phologically the forelimbs, thorax, scapular, and vertebral column might have resembled those of quadrupeds like Ateles or Lagothrix rather than Old World forms like Colobus or Hylobates. As body size increased, more time would be spent on the ground and knuckle-walking behavior and adaptations would evolve fully. The gorilla is a mainly herbivorous vegetarian living in thick, lush, rain forest at a variety of altitudes (see Schaller, 1963; Groves, 1967). In particular, the Eastern gorilla (G. g. beringei of most authors; G. G. beringet and G. g. manyema of Groves, 1967) inhabits upland country. D. (P.) major has been recovered from deposits at Napak, Moroto, Koru and Songhor. These sites preserve animals which presumably died on the forest-covered slopes of active volcanoes in habitats similar to those occupied by Eastern gorillas today. It is 124 TERTIARY PONGIDAE OF EAST AFRICA 125 tempting to think that D. (P.) major lived in such an environment and that this species was already adapting to the more selectively vegetarian way of life of its Recent descendant. Traces only of D. (P.) major have been recovered from the Rusinga Island deposits. The sites on the island represent a variety of environ- ments; savannah, swamp, lacustrine, and riverine. The pongid material from Rusinga consists mainly of two further species of D. (Proconsul), D. (P.) africanus and D. (P.) nyanzae. D. (P.) africanus is much smaller than D. (P.) major, in size falling between Symphalangus syndacty- lus and Pan troglodytes paniscus. It is similar enough to D. (P.) major for them to be classified for the moment in the same subgenus. There are differences, however, in the relative proportions of the molars, in the morphology of the symphysis, and in the smoothness of the cranial vault. This last feature is cor- related with the orthognathous face and the lightness of the masticatory appara- tus in D. (P.) africanus. Although the cranial vault is not known in D. (P.) major, it is most unlikely that it was also rounded and delicately constructed. The forelimb of D. (P.) africanus is known from one tolerably complete speci- men from Rusinga, and the species was probably mainly arboreal. In a number of features—cranial, dental, and postcranial—D. (P.) africanus resembles P. troglodytes and is possibly ancestral to the living species. ‘This relationship has already been tentatively suggested by a number of workers since Hopwood first described this species in 1933. It is found in Kenya at sites other than Rusinga, but it has not yet been recovered from Uganda. D. (P.) africanus is found with D. (P.) major at Songhor and Koru and, like living chimpanzees, probably lived in a variety of habitats ranging from savannah and open woodland to forested mountainside. In the Miocene, therefore, both proto- chimpanzees and proto-gorillas were probably small, lightly built, mainly arbo- real creatures, the former probably lacking knuckle-walking adaptations. For the moment these two ancestral species are classified in one subgenus, although further material may indicate that they are better separated at the subgeneric level. The third species, D. (P.) nyanzae, in size is midway between the other two species; the smallest individuals overlap D. (P.) africanus and the largest D. (P.) major. In overall morphology D. (P.) nyanzae is most similar to D. (P.) major, although it differs in having relatively smaller anterior teeth in general and first molars in particular (see Table 35). In these features D. (P.) nyanzae is more primitive than D. (P.) major. D. (P.) nyanzae seems to be contemporary with D. (P.) major, yet it is probable that this species, or a very similar species, was ancestral to D. (P.) major. D. (P.) nyanzae is found in abundance through- out the Rusinga beds, while so far it is virtually absent from the sites at Koru, Songhor, Moroto, and Napak which have yielded D. (P.) major. These differences may be due simply to random sampling fluctuations, or may be because there were genuine ecologic differences between the two species. The oldest of these East African sites may be sampled from a time when the evolution of D. (P.) major from a species very similar to D. (P.) nyanzae was barely completed. In a number of publications, Leakey (see especially 1963) has proposed that the three species discussed above should be included in a family, Pronconsulidae, separate from the Pongidae. Pongidae would include the living great apes and other species of Dryopithecus. The classification of D. (Proconsul) has been 126 PEABODY MUSEUM BULLETIN 31 discussed by Simons and me (1965, p. 105-111), and it has been shown there that the differences between Dryopithecus Lartet and Proconsul Hopwood cannot be regarded as significant at the generic level. The characteristics of the various species of D. (Proconsul) which differ from those of the living pongids are to be regarded as primitive. Thus, the relatively orthognathous face, narrow symphyseal region, the absence of a simian shelf, the presence of well-developed molar cingula, the relatively small first molars, and the generalized cranial and postcranial morphology can be regarded as foreshadowing conditions seen in the living African great apes. As Remane comments (1965, p. 281), “Proconsul unterscheidet sich von den _ rezenten Pongidae weniger als die miozinen Vertreter vieler anderer Sdugertierfamilien von ihren rezenten Arten.” Until recently, in the absence of known Miocene Hominidae, D. (Proconsul) has been regarded as some kind of ‘‘model’ of the pre-Pleistocene hominids. Napier has stated (1963b, p. 185), “It is inevitable that Proconsul . . . should be considered in the search for the Tertiary ancestors of . . . Villafranchian hominids. Whatever the final status of Proconsul in the systematics of hominid evolution may turn out to be... it represents an important structural and functional stage in the phylogeny of hominid locomotion.” Since it now seems that two of the D. (Proconsul) species are ancestral to the living African apes, or close to that ancestry, and since these Early and Middle Miocene species are almost certainly not ancestral to Late Miocene Hominidae, the viewpoint expressed by Napier should be re-examined. It is true that the forelimb and skull of D. (P.) africanus (nondentally, the best known of the species) are in many ways generalized and unspecialized compared with those of living apes. Whether or not Early and Middle Miocene hominids were generalized and unspecialized in the same ways is unknown, and an answer must depend on the recovery of Miocene hominid postcranial material. One point that is clear; if the relationships suggested in this monograph are correct, then the lineages leading to P. troglodytes and G. gorilla were probably specifically distinct in the Early Miocene, twenty million years ago. Clark and Leakey (1951, p. 111-112) suggested that D. (Proconsul) species were ancestral to the African great apes. The similarities shown by these two living species have led some workers (i.e. Sarich and Wilson, 1967) to assume that their separation is relatively recent. In my view, at least some of these similarities ought now to be regarded as more probably parallelisms. The ancestral species, D. (P.) africanus and D. (P.) major, are sufficiently similar dentally and facially to be classified in the same subgenus; they are much more similar in these areas than are P. troglodytes and G. gorilla. This probably applies postcranially, too. It seems taxonomically most sensible to retain the two present-day species in separate genera for the moment, as Schultz (1966) has suggested. The absence in D. (Proconsul) of the posteriorly projecting inferior trans- verse torus (basal plate or simian shelf) is an obvious point of difference be- tween the Miocene species and their living descendants. As noted above, Leakey (1963, 1965) regards this feature as a major character separating D. (Proconsul) species from other pongids. However, if this character is viewed functionally and regarded as an adaptation to stresses in the anterior mandibular region associated with an increase in size of the dentition, particularly the incisors, its significance TERTIARY PONGIDAE OF EAST AFRICA 127 can be viewed in evolutionary perspective. A fossil species ought not to be re- moved from the ancestry of a living form simply because it differs in certain relatively minor points. If the differences have reasonable functional explana- tions, such a course becomes even more absurd. If evolution has occurred, de- scendants inevitably differ from ancestors. EURASIAN DRYOPITHECINES The African dryopithecines come from deposits between 20 and 18 million years old. Younger Dryopithecus have been found in European and Asian sites. Those from India have already been mentioned. The first European dryopi- thecine, D. fontani, appears in deposits of late Middle Miocene (Tortonian) age in France and Czechoslovakia; these sites are probably 15 or 16 million years old (see Table 1, p. 17). A number of specimens of D. fontani have been de- scribed from south-western France and a few teeth from the Vienna Basin in Czechoslovakia. These specimens were noted by Simons and me (1965, p. 83- 85). Two specimens, the type mandible of D. fontani from St. Gaudens de- scribed in 1856 by Lartet, and a referred mandible from the same site described in 1890 by Gaudry, are in the Muséum National d’Histoire Naturelle, Paris. Through the courtesy of the director, M. J.-P. Lehman, I was able to examine these specimens. This work is not an appropriate place for a comprehensive discussion of the European dyropithecines. However, a number of relevant notes on these specimens will be included here. The type specimen, Paris Museum no. AC 36, consists of right and left horizontal rami, the left with P,-M, and Mz alveolus, the right with broken C,, P;—-M,, and Mz alveolus. A small fragment of symphysis containing the incisor alveoli is also preserved. This specimen is figured in Piveteau (1957, p. 198). At P,, the horizontal ramus is deep (34.5 mm) and thin (12.5 mm); posteriorly it shallows and at Mz is 28.0 mm deep. The tooth row is crowded anteriorly and the overlap between the mesial border of right P; and the distal border of C, is marked. The canine, first premolars, and second molars fall within the range of D. (P.) nyanzae, although P, and M, are somewhat larger. In view of the inferred trends within the several pongid lineages, this increase in relative size of P, and M, might be expected in a late Middle Miocene form. In fact, D. fontani M,’s are relatively no larger than those of D. (P.) africanus and D. (P.) major specimens (a point previously overlooked), although they are relatively larger than those of D. (P.) nyanzae (see Table 49). The dentition of AC 36 is very similar to that of D. (Proconsul); the molars retain distinct buccal cingula. In both first and second molars the hypoconulid is shifted bucally. The symphyseal fragment contains the incisor alveoli. The total incisor breadth was not great, 17 or 18 mm only. The incisors were not procumbent. The most superior 20 mm of the anterior part of the symphysis is preserved for some 7 mm on either side of the midline. This part of the anterior symphyseal contour was probably close to being vertical to the occlusal plane. The posterior border of the symphysis is poorly preserved, only the most superior 7 or 8 mm being present. Neither symphyseal thickness nor morphology can be gauged from this fragment. 128 PEABODY MUSEUM BULLETIN 31 The second specimen, Paris Museum no. 1872-2 or 1902-12b, has also been figured by Piveteau (1957, p. 198 and 199); an excellent photograph is in- cluded by Remane (1965, p. 282). The dentition, although not as crowded as in the type, is nevertheless morphologically and metrically very similar. This specimen has generally not been fully described (except by Remane, 1922), and a number of misconceptions about it are currently in the literature. For example, in 1953 (p. 179), Leakey wrote of D. fontant, “The simian shelf is well developed and the rows of cheek-teeth are parallel (for instance, in the specimen from the Middle Miocene of St. Gaudens in France).” The symphysis of this specimen is fractured anteroposteriorly, and the right tooth row has been pushed lingually. This has affected in particular the more distal teeth. The left horizontal ramus has also been fractured between P, and M, and the distal part of the molar row rotated lingually. Although the tooth rows appear to be parallel, this is due to distortion. Originally, the internal contours of the tooth rows probably diverged slightly posteriorly; the outer contour would have been more markedly divergent. The tooth rows were probably no more parallel than in D. (P.) major, NMK 190,1 for example. The symphysis is well preserved (see Fig. 3). The anterior margin slopes backward and downward as in D. (P.) africanus. The planum alveolare is long and gently sloping like that of D. (P.) major. The symphysis is 39.5 mm long and 21.5 mm thick in the midline and is therefore a little smaller than symphyses of supposed female D. (P.) major, although the proportions are similar (see dimensions in Chap. VI). What appears to be an inferior lingual torus is present on both horizontal rami below M, and M,; the torus is present in a less de- veloped form in the type and the thickness is probably emphasized in 1872-2 by lingual crushing of the basal parts of the rami. An inferior transverse torus is absent from the midline. This region has been examined microscopically and the absence is almost certainly not due to breakage or crushing. ‘The most pos- terior part of the cross section is at the level of the superior transverse torus; the cross section of the symphysis is therefore very similar to that of D. (P.) major. These specimens show mandibular similarities with D. (Proconsul), and also to NMK 76, CMH 142, assigned by Leakey (1967) to ““Kenyapithecus africanus.” A third mandible of D. fontani from St. Gaudens, a left horizontal ramus with C, to Ms, was described in 1898 by Harlé. This specimen is illustrated by Piveteau (1957, p. 201). The resemblances between this specimen and some Asian D. (S.) sivalensis (e.g., AMNH 19412, see Gregory and Hellman, 1926, p. 23) are particularly marked. Cingula are more marked in the European speci- mens, however, and they are in this respect more primitive, resembling D. (Pro- consul) species. The upper dentition of Middle Miocene D. fontanz is poorly known. What is preserved resembles that of D. (S.) stvalensis (see Gregory and Hellman, 1926, fig. 6B), although lingual cingula may be somewhat better developed. Some D. (S.) sivalensis from the Siwaliks may be as old as Middle Miocene, although the majority of the known specimens are most probably younger than this. D. (S.) sitvalensis molars tend either to lack cingula entirely or to have only small remnants preserved. ‘The lingual borders of the upper molars bulge lingually, this bulging presumably representing the remains of the lingual cingu- lum. The buccal borders of the lower molars bulge in a similar way. TERTIARY PONGIDAE OF EAST AFRICA 129 THE RELATIONSHIPS OF DRYOPITHECUS SPECIES Although the three subgenera of Dryopithecus show varying degrees of expres- sion of the cingulum, this feature is variable within genera and species of a number of mammals (e.g., Hylobates, see Frisch, 1965) and the differences do not seem to warrant generic distinction. In basic morphology, the teeth of the various Dryopithecus species are closely similar (see Fig. 30). Table 48 includes ratios of lower molar lengths for various species of fossil hominoids. The first molar is relatively shortest in D. (P.) nyanzae, and this is presumably a primitive trait. The ratios in other species of Dryopithecus, al- though similar, are a little higher. If the ratios be compared to those in living Table 48 Ratios of molar MDi L of Pongidae and Hominidae (M, /M,) x 100 (M,/M,) x 100 n x OR n X OR Propliopithecus haeckeli i HOso 1 1 88.9 Aegyptopithecus zeuxis 2 84.0 S45 CI" A815) 7/ 1 78.8 D, (P,) africanus 4 85.2 80.0- 89.9 Z 84.1 83.7- 84.5 D. (P.) nyanzae 4 78.2 75.4- 82.3 4 87.5 83.2- 92.8 D, (@.) major 4 86.0 81.7- 90.8 4 85.6 80.3- 93.3 D, sp. (Rs 1967,394) 1 sly 2 1 93.8 D, fontani 2 84.5 84.0- 85.0 Z 93/0 92.3- 93.7 D. (S.) sivalensis 9 86. 0 80.0- 93.7 7 9257 83.8- 99.3 P. troglodytes 26 Jo. 5 90.0-104.9 ZS LOFa7. 96.5-122.1 G, gorilla 40 89.3 Bhs S746) 40 98.4 86.5-106.9 Pongo pygmaeus 20 93.3 Eis o> SIG 4 20 100.1 86. 2-108. 2 R, _punjabicus 1 *90.1 3 96.7 94. 4-100.0 Australopithecus africanus 3 88.5 86.4- 90.5 2 104.5 99.3-109.6 Australopithecus robustus 7 s)als 7/ 86. 8- 97.3 4 95.4 92. 5-101. 2 Homo erectus 13 98.3 88. 6-104.8 ial 103.9 9250-11950 pongids, it can be seen that in all lineages the first molars have increased rela- tively in size. The ratio of second and third molar lengths is low in D. (Proconsul) species, the third molar being relatively elongated. Living pongids have approximately subequal second and third molar lengths, although in P. troglodytes M3; is gen- erally shorter than M,. This is probably due to a reduction in size of M3, perhaps quite recently. In G. gorilla and Pongo pygmaeus the ratio is approximately 100; trends in the lineages leading to these species seem therefore to have included an increase in size of My relative to M3 (see Appendix 3 for D. (P.) major—G. gorilla changes). Dryopithecus species from the Middle and Late Miocene and Early Pliocene have somewhat longer M,’s relative to M; than in D. (Proconsul) species. The 130 PEABODY MUSEUM BULLETIN 31 differences are probably due to the greater geological age of the African species; no doubt the later Miocene African contemporaries of Eurasian Dryopithecus would have produced similar ratios. It is of interest to note that the Early Mio- cene Dryopithecus species from Rusinga (Leakey, 1968) has a relatively small M,, although the relative proportions of M, and Ms; are as in Eurasian Dryopithe- cus. Table 49 includes ratios of M, talonid to trigonid breadths. This ratio gives some indication of the distal narrowing of M3. Examination of the values ob- tained for living pongids and comparison with their putative ancestors suggests that a more triangular Mz (hence giving a lower ratio) is primitive. The shape of Mj is not constant in D. (Proconsul), D. (P.) major retaining the more primi- tive shape compared to the other two species. G. gorilla is also the most primitive of the living Pongidae in this character. D. (S.) stvalensis, D. fontant and Pongo pygmaeus are very similar to each other. Table 49 Ratio of Mg (Tal B/ Tri B) x 100 in fossil and living pongids n X OR Propliopithecus haeckeli 1 93er Aegyptopithecus zeuxis 1 87.4 D. (P.) africanus 8 93.4 90.0- 96.4 D. (P.) nyanzae 5 92.4 89.6- 97.8 D. (P.) major 3 86.9 84.5- 89.5 D. fontani 1 94.2 D. (S.) sivalensis 12 93.7 90.0-100.0 P. troglodytes 10 97.8 97.3—L0lee G. gorilla males 20 89.0 Uso Disc Pongo pygmaeus 6 93.4 8974 — S960 R. punjabicus 3 94,2 9210-89 Seno The functional significance of the changes reflected in these various ratios is not well understood. Presumably, an increase in surface area of the cheek teeth has been selected for in these primates in which crushing and grinding are important activities. As the more anterior smaller molars have increased rela- tively in size, the three molars have tended to subequality in length. It is not only cheek teeth which have become enlarged, but incisors too. Table 41, p. 111, included values of the ratio comparing upper central incisor area to the combined areas of first and second molars. ‘The value of this ratio for living pongid species was calculated using mean measurements from Ashton and Zuckerman (1950). These ratios show that the central incisors are relatively larger in living pongids than in Dryopithecus species. However, D. (P.) major (UMP 62-11) is very close to the mean values for G. gorilla. The lineages lead- ing to chimpanzees and orangutans have shown a much greater increase in in- cisor size since the Miocene, and these living species have relatively much larger incisors than gorillas. One explanation of this might be that the Miocene ances- tors of G. gorilla were becoming adapted to an herbivorous diet of shoots, stems, and leaves. Increasing adaptation since then has involved a general increase in TERTIARY PONGIDAE OF EAST AFRICA 131 tooth, jaw, skull, and body size, and some relative increase in size of the anterior cheek teeth. The incisors have increased little in relative size when compared with the rest of the dentition. Morphologic changes of cheek tooth crowns, how- ever, have been important. In both chimpanzees and orangutans, however, the incisors are very large relative to the cheek teeth, see Table 50. Both species are frugivorous rather than herbivorous and large anterior teeth may have been selected for piercing the tough outer covers of fruits. Present-day chimpanzees often use their incisors for this purpose (Walker, personal communication). However, these phylogenetic and functional hypotheses can only be tested with the recovery of more, adequately dated, material, with proper stratigraphic and ecologic controls. Although the Eurasian Dryopithecus species are at present classified in separate subgenera, D. fontani from Europe and D. (S.) sivalensis from the Siwaliks are closely related and should probably not be separated at the sub- Table 50 Ratio (crown area of I*/crown area M and M*) x 100 in living Pongidae P. troglodytes (males) 43.1 P. troglodytes (females) 47.0 G. gorilla (males) yale 74! G. gorilla (females) S0N5 Pongo pygmaeus (males) 50.3 Pongo pygmaeus (females) syle generic level. Similar pairs of genera and species of other mammals are dis- tributed in Europe and Asia during this period, and the two continents should be treated as a continuous land mass (or rather Europe should be regarded as a small westerly projection of Asia). Eurasian Middle Miocene Dryopithecus spe- cies almost certainly originated in Africa and presumably reached Eurasia no earlier than the Middle Miocene, some 16 or 17 million years ago. None of the known African species of D. (Proconsul) are obvious ancestors of the Eurasian species; however, there is a hint of such an ancestor in some of the East African sites, as I noted in the preceding chapter. PONGID EVOLUTION Sarich and Wilson (1967) have attempted to assess the relationships of living hominids and individual pongid species using immunological resemblances in serum albumins. They argue that albumins have evolved at a constant rate and that these proteins can be used as a “clock” which, once calibrated, will give the time of divergence of each living species. Sarich and Wilson conclude that, as- suming a cercopithecoid/hominoid dichotomy in the Oligocene (the “clock” is calibrated with paleontological data), the hylobatids diverged from other hominoids ten million years ago, the proto-orangs eight million years ago, while the Pan, Gorilla, and Homo lineages have split only during the last five million years. More recently Sarich (1968) has revised this date to 3.5 million years. 132 PEABODY MUSEUM BULLETIN 31 Sarich and Wilson suggest that living apes and man are descended from a small Miocene dryopithecine which was already a “brachiator,” like the gibbon —an arm-swinger with elongated upper limbs and other correlated scapular, thoracic, and lumbar features. They state (1967, p. 1202): “If the view that man and the African apes share a Pliocene ancestor and that all the living Hominoidea derive from a late Miocene form is correct, a number of the prob- lems that have troubled students of this group are resolved. The many features of morphology, particularly in the thorax and upper limbs, which man and the living apes share in varying degrees, but which were not present in the Miocene apes, such as Dryopithecus, ... and Pliopithecus, are then seen as due to recent common ancestry and not, as generally accepted, to parallel or convergent evolution.” Several objections can be made to this viewpoint. First, many of the biochemical assumptions may be rather too sweeping. For example, Barnicot, Jolly, and Wade (1967, p. 343-344) note: “An antiserum prepared against a given protein of one species can be used to test sera of other species and the reactions give information about antigenic similarities and differences. However, the relation between antigenic properties and protein struc- ture is by no means fully understood so that the information provided by such immunological tests is much less explicit than chemical analysis.” Until amino acid sequences of albumins are better understood, caution in interpreting im- munological resemblances should be urged. Also the structure of other proteins should be considered in deriving phylogenies, for not all tell the same story. Barnicot, Jolly and Wade (1967, p. 352-353) have expressed this very well. The evidence provided by the structure of a single kind of protein, unique though it may be in the insight it allows into events at the gene level, does not necessarily give an unbiased view of the affinities of whole organisms. An obvious remedy for this is to look at more proteins, however laborious this may be. It is to be anticipated that in some lineages particular proteins will be found to have changed considerably whereas others have changed less. Indeed protein structures will probably prove to be a microcosm in which well-recognized evolutionary phenomena such as convergence and differential rates of change will often be exemplified. Students of molecular evolution naturally turn to paleontology for a time scale when they wish to calculate rates of change of proteins. To reverse this procedure and seek in proteins an evolutionary clock from which time can be read is likely to be fallacious and no more justifiable than in dealing with morphological characters. Although regularities in rates of structural change of proteins may emerge empirically there is no good reason to assume a priovi that the underlying forces of mutation and selection pressure have remained constant. Paleontological evidence should ideally be used as a framework in which the evolution of particular proteins, or morphological characters, can be studied. Second, the evidence indicates that all Miocene hominoids so far recovered were basically quadrupedal. None of them can be described as “brachiators” in the sense that Hylobates is a brachiator. Third, the paleontological evidence discussed in this paper indicates that, with some probability, proto-Gorilla and proto-Pan lineages can be traced back TERTIARY PONGIDAE OF EAST AFRICA 135 to the early Miocene, while early hominids are known from late Miocene de- posits. If Sarich and Wilson are correct, these species became extinct, to be replaced later in time by another set of proto-Gorilla, proto-Pan and pre-Homo lineages. This would be a more unlikely set of parallelisms than those which Sarich and Wilson wish to explain away. Fourth, the trend towards a forelimb-dominated type of locomotor-feeding behavior (“brachiation’’) in a small branch niche (Napier, 1967) is documented not only in Hominoidea, but in other primate taxa, too. For example, an extinct prosimian, Palaeopropithecus (Walker, 1967), and an extinct catarrhine, Oreo- pithecus (Straus, 1963), were both ‘“‘brachiators.” Ateles shows parallelistic changes also, particularly in thorax, scapula, and lower trunk. Evidently the trend towards this particular locomotor-feeding adaptation is widespread in primates and progresses to different levels in a number of living and extinct lineages. The various hominoid lineages could plausibly have evolved their “brachiating” adaptations in parallel from a quadrupedal ancestor already com- mitted to the small branch niche. Fifth, if we assume that the early Pliocene hominoid ancestor (after Sarich and Wilson, 1967) was a small “brachiator” and that the proto-orang of the middle Pliocene was a large “modified brachiator” or “fist-walker’” (Tuttle, 1967) we are faced with the problem of transporting these highly arboreal animals, with few if any terrestrial adaptations, from Africa across the Arabian peninsula and continental India to their present forested habitat of Southeast Asia. It is highly unlikely that a forest, or even woodland, connection existed in East Africa and Arabia during the Pliocene (Howell, 1967). In the unlikely event that a common ancestor of the Hominoidea existed in the early Pliocene of Africa, the dispersal of highly arboreal forms like Hylobates and Pongo across arid terrain at this time seems improbable, to say the least. In conclusion, the paleontological evidence suggests that the radiation of living hominoids had already occurred in the Miocene. Hominids are known from the late Miocene and one might infer that their ancestors are to be found in still earlier times, already differentiated from ancestral pongids. However, none of the early Miocene species discussed here show any dental or gnathic characters specially typical of hominids, and at present there is no evidence connecting any known early Miocene species with Ramapithecus. Nevertheless, hominid dental, gnathic, and (presumably) behavioral characters appear to be more ancient than often supposed, just as the differentiation of living pongid species occurs at an earlier time than has been generally thought. As a final word I should like to point out that the fossil record of ancestral pongids is by no means as scanty as some recent workers have suggested (see Preface of Washburn and Jay, 1968). Nor can Dryopithecus species be written off as merely dental apes, with bodies like those of quadrupedal monkeys (Wash- burn, 1968). There are ‘“quadrupeds” and then there are “quadrupeds.” Ateles and Cercocebus may both be classified as arboreal quadrupeds, but they differ markedly both in behavior and in morphology. The Cercopithecoidea are a much more homogeneous group in locomotor terms than the Ceboidea, and to believe that, because Dryopithecus was a quadruped, it was therefore a cerco- pithecoid-like quadruped, is to make an unfortunate mistake. All evidence points to a much more flexible, ateline-like, type of quadrupedalism for Dryopt- 134 PEABODY MUSEUM BULLETIN 31 thecus. Thus hominoid characters of scapula, forelimb, thorax, and vertebral column might already have been to some extent present in “quadrupedal” Dryopithecus species. This seems also to have been the case for Pliopithecus (Walker, personal communication). This scheme would rule out a gibbon-type Miocene stage in hominoid evolution and would imply that Pan and Gorilla never went through a long-armed “‘brachiating” phase. EITERATURE CrreD Allbrook, D., and Bishop, W. W., 1963, New fossil hominoid from Uganda: Nature, v. 197, p. 1187-1190. Andrews, C. W., 1911, On a new species of Dinotherium (Dinotherium hobleyi) from British East Africa: Zool. Soc. London Proc., v. 2, p. 943-945 Ashton, E. H., Healey, M. J. R., Oxnard, C. E., and Spence, T. F., 1965, The combination of locomotor features of the primate shoulder girdle by canonical analysis: Jour. Zool., v. 147, p. 406-429. Ashton, E. H., and Zuckerman, S., 1950, Some quantitative dental characteristics of the Chim- panzee, Gorilla and Orang-Outang: Roy. Soc. [London] Phil. Trans. B., v. 234, p. 471-484. Barnicot, N. A., Jolly, C. J., and Wade, P. T., 1967, Protein variations and primatology: Amer. J. Phys. Anthrop., v. 27, p. 343-355. Bishop, W. W., 1958, Miocene Mammalia from the Napak Volcanics, Karamoja, Uganda: Nature, v. 182, p. 1480-1482. 1962, The mammalian fauna and geomorphological relations of the Napak Volcanics, Karamoja: Geol. Survey of Uganda Records, 1957-1958, p. 1-18. 1963a, The later Tertiary and Pleistocene in eastern equatorial Africa, in Howell, F. C., and Bourliere, F. (eds.), African ecology and human evolution: Aldine, Chicago, p. 246-275. 1963b, Uganda’s animal ancestors: Wildlife and Sport, v. 3, 7 p. 1964, More fossil primates and other Miocene mammals from North-East Uganda: Na- ture, v. 203, p. 1327-1331. 1966, Stratigraphical geomorphology, in Dury, G. H. (ed.), Essays in geomorphology: Heinemann, London, p. 139-176. 1967, The later Tertiary in East Africa—volcanics, sediments and faunal inventory, in Bishop, W. W., and Clark, J. D. (eds.), Background to evolution in Africa: Univ. of Chicago Press, Chicago, p. 31-56. Bishop, W. W., Fitch, F. J., and Miller, J. A., 1969, New potassium-argon age determinations rele- vant to the Miocene faunal and volcanic sequence in East Africa. (In press) Bishop, W. W., and Whyte, F., 1962, Tertiary mammalian faunas and sediments in Karamoja and Kavirondo, East Africa: Nature, v. 196, p. 1283-1287. Boné, E. L., 1955, Une clavicle et un nouveau fragment mandibulaire d’Australopithecus pro- metheus: Palaeont. Africana, v. 3, p. 87-101. Campbell, B. G., 1966, Human evolution: Aldine, Chicago, 425 p. Cave, A. J. E., 1961, The frontal sinus of the gorilla: Zool. Soc. London Proc., v. 136, p. 359-373. Chesters, K. I. M., 1957, The Miocene flora of Rusinga Island, Lake Victoria, Kenya: Palaeont- ographica, v. 101, B, p. 30-71. Clark, W. E. Le Gros, 1952, Report on fossil hominoid material collected by the British-Kenya Miocene Expedition, 1949-51: Zool. Soc. London Proc., v. 122, p. 273-286. 1964, The evolution of man: Discovery, London, v. 25, July, p. 49. and Leakey, L. S. B., 1950, Diagnoses of East African Miocene Hominoidea: Geol. Soc. London Quart. Jour., v. 105, p. 260-263. 1951, The Miocene Hominoidea of East Africa: Fossil mammals of Africa, no. 1, Brit. Mus. (Nat. Hist.), 117 p. Davis, D. D., 1964, The giant panda: a morphological study of evolutionary mechanisms: Fieldi- ana, v. 3, 339 p. Davis, P. R., and Napier, John, 1963, A reconstruction of the skull of Proconsul africanus: Folia Primat., v. 1, p. 20-28. Defrise-Gussenhoven, E., 1955, Ellipses equiprobables et taux d’eloignement en biometrie: Inst. Sci. Nat. Belg. Bull., v. 31, no. 26, 31 p. Evernden, J. F., Savage, D. E., Curtis, G. H., and James, G. T., 1964, Potassium-argon dates and the Cenozoic mammalian chronology of North America: Amer. Jour. Sci., v. 262, p. 145-198. Frisch, J. E., 1962, Neuer Fund eines fossilen Hominoiden in Kenya: Anthrop. Anz., v. 25, p. 298-301. 1965, Trends in the evolution of the hominoid dentition: Bibliotheca Primat., v. 3, 130 p. Gaudry, A., 1890, Le Dryopithéque: Soc. Geol. France Mem., v. 1, p. 5-11. Goodall, Jane, 1965, Chimpanzees of the Gombe Stream Reserve, in DeVore, I. (ed.), Primate behavior: Holt, Rinehart and Winston, New York, p. 425-473. US}s) 136 PEABODY MUSEUM BULLETIN 31 Gower, J. C., 1966, Some distance properties of latent root and vector methods used in multi- variate analysis: Biometrika, v. 53, p. 325-338. Gregory, W. K., 1916, Studies on the evolution of the primates: Amer. Mus. Nat. Hist. Bull., v. 35, p. 239-355. and Hellman, M., 1926, The dentition of Dryopithecus and the origins of man: Amer. Mus. Nat. Hist., Anthrop. Papers, v. 28, 123 p. Gregory, W. K., Hellman, M., and Lewis, G. E., 1938, Fossil anthropoids of the Yale-Cambridge India Expedition of 1935: Carnegie Inst. Washington Publ. no. 495, 27 p. Groves, C. P., 1967, Ecology and taxonomy of the gorilla: Nature, v. 213, p. 890-893. Harlé, E., 1898, Une machoire de Dryopithéque: Bull. Soc. Géol. Fr., 3e série, v. 26, p. 377-383. Hooijer, D. A., 1948, Prehistoric teeth of man and of the orang-utan from Central Sumatra, with notes on the fossil orang-utan from Java and Southern China: Zool. Meded., v. 29, p. 175-284. Hopwood, A. T., 1933a, Miocene primates from British East Africa: Ann. Mag. Nat. Hist., ser. 10, v. 11, p. 96-98. 1933b, Miocene primates from Kenya: Linn. Soc. Lond. Jour., v. 38, p. 437-464. Howell, F. C., 1967, Recent advances in human evolutionary studies: Quart. Rev. Biol., v. 42, p. 471-513. Hrdlicka, Ales, 1935, The Yale fossils of anthropoid apes: Amer. Jour. Sci., ser. 5, v. 29, p. 34-40. Koenigswald, G. H. R. von, 1962, Potassium-argon dates for the Upper Tertiary: Koninkl. Nederl. Akademie Van Weterschappen, ser. B, v. 65, no. 1, p. 31-34. Korenhof, C. A. W., 1960, Morphogenetical aspects of the human upper molar: Uitgeversmaat- schappij Neerlandia, Utrecht, 368 p. Lartet, E., 1856, Note sur un grand singe fossile qui se rattache au groupe des singes superieurs: Acad. Sci. [Paris] Comptes rendus, v. 43, p. 219-223. Leakey, L. S. B., 1943, A Miocene anthropoid mandible from Rusinga, Kenya: Nature, v. 152, p- 319-320. 1953, Adam’s ancestors: Methuen, London, 235 p. 1962, A new Lower Pliocene fossil primate from Kenya: Ann. Mag. Nat. Hist., 13th ser., v. 4, p. 689-696. 1963, East African fossil Hominoidea and the classification within this super-family, in Washburn, S. L. (ed.), Classification and human evolution: Aldine, Chicago, p. 32-49. 1964, The evolution of man: Discovery, London, v. 25, August, p. 48-50. 1965, in DeVore, P. (ed.), The origin of man: Symposium distributed by the Wenner- Gren Foundation, New York, 150 p. 1967, An Early Miocene member of Hominidae: Nature, v. 213, p. 155-163. 1968, Lower dentition of Kenyapithecus africanus: Nature, v. 217, p. 827-830. Lewis, G. E., 1934, Preliminary notice of new man-like apes from India: Amer. Jour. Sci., 5th serve 27) p- 61-179: 1937a, Taxonomic syllabus of Siwalik fossil anthropoids: Amer. Jour. Sci., 5th ser., v. 34, p- 137-147. 1937b, Siwalik fossil anthropoids, Yale Univ. Ph.D. Thesis, unpubl., 125 p. MacInnes, D. G., 1943, Notes on the East-African Miocene primates: East Afr. Uganda Nat. Hist. Soc..Jour:, ve 7,.p) 141—-181° Napier, J. R., 1963a, Brachiation and brachiators: Zool. Soc. Lond. Symp., no. 10, p. 183-195. 1963b, The locomotor functions of hominids, in Washburn, S. L. (ed.), Classification and human evolution: Aldine, Chicago, p. 178-189. 1964, The evolution of bipedal walking in the hominids: Archives Biol. (Liege), v. 75, p. 673-708. 1967, Evolutionary aspects of primate locomotion: Amer. Jour. Phys. Anthrop., n. ser., v. 27, p. 333-342. Napier, J. R. and Davis, P. R., 1959, The fore-limb skeleton and associated remains of Proconsul africanus: Fossil mammals of Africa, no. 16, Brit. Mus. (Nat. Hist.), 69 p. Napier, J. R., and Walker, A. C., 1967, Vertical clinging and leaping—a newly recognized cate- gory of locomotor behaviour of primates: Folia Primat., v. 6, p. 204-219. Pilbeam, D. R., 1965, in DeVore, P. (ed.), The origin of man: Symposium distributed by the Wenner-Gren Foundation, New York, p. 39-42. 1966, Notes on Ramapithecus, the earliest known hominid, and Dryopithecus: Amer. Jour. Phys. Anthrop.; n. ser., v. 25, p. 1-5. 1967, Man’s earliest ancestors: Sci. Jour., v. 3, no. 2, p. 47-53. 1968, The earliest hominids: Nature, v. 219, p. 1335-1338. Pilbeam, D. R., and Simons, E. L., 1965, Some problems of hominid classification: Amer. Sci., v. 53, p. 237-259. Pilgrim, G. E., 1910, Notices of new mammalian genera and species from the Tertiaries of India: Geol. Survey of India Records, v. 40, p. 63-71. TERTIARY PONGIDAE OF EAST AFRICA 137 1915, New Siwalik primates and their bearing on the question of the evolution of man and the Anthropoidea: Geol. Survey of India Records, v. 45, p. 1-74. 1927, A Sivapithecus palate and other primate fossils from India: Geol. Survey of India Memoirs, n. ser., v. 14, 24 p. Piveteau, J., 1957, Traité de paléontologie; Tome VII, Primates: Masson et Cie, Paris, 675 p. Prasad, K. N., 1964, Upper Miocene anthropoids from the Siwalik beds of Haritalyangar, Hima- chal Pradesh, India: Palaeontology, v. 7, p. 124-134. Remane, A., 1922, Beitrage zur Morphologie des Anthropoidengebisses: Arch. Naturgesellsch., v. 89, p. 1-179. 1954, Structure and relationships of Meganthropus africanus: Amer. Jour. Phys. Anthrop., MSEDn Ve U2; 9p. 123-126: 1959, Die primitivisten Menschenformen (Australopithecinae) und das Problem des tertidren Menschen: Naturwiss. Vereins fiir Schleswig-Holstein, v. 29, p. 3-10. 1960, Zahne und Gebiss, in Hofer, H., Schultz, A. H., and Starck, D. (eds.), Primatologia: Karger, Basel, v. 3, T. 2, p. 637-846. 1965, Die Geschichte der Menschaffen: in Menschliche Abstammungslehre: Gustav Fischer, Stuttgart, p. 249-309. Reynolds, Vernon, and Reynolds, Frances, 1965, Chimpanzees of the Budongo Forest, in DeVore I. (ed.), Primate behavior: Holt, Rinehart and Winston, New York, p. 368-424. Robinson, J. T., 1952, Note on the skull of Proconsul africanus: Amer. Jour. Phys. Anthrop., n. ser, Vi llO} ap. /—12- 1956, The dentition of the Australopithecinae: Transvaal Mus. Mem., no. 9, 179 p. Sarich, V. M., 1968, The origin of the hominids: an immunological approach, in Washburn, S. L., and Jay, P. C. (eds.), Perspectives on human evolution: Holt, Rinehart, and Winston, New York, v. 1, p. 94-121. Sarich, V. M., and Wilson, A. C., 1967, Immunological time scale for hominid evolution: Science, v. 158, p. 1200-1203. Schaller, G. B., 1963, The mountain gorilla: Univ. of Chicago Press, Chicago, 431 p. 1965, The behavior of the mountain gorilla, in DeVore, I. (ed.), Primate behavior: Holt, Rinehart and Winston, New York, p. 324-367. Schultz, A. H., 1963, Age changes, sex differences, and variability as factors in the classification of primates, in Washburn, S. L. (ed.), Classification and human evolution: Aldine, Chicago, p. 85-115. 1966, Changing views on the nature and interrelations of the higher primates: Yerkes Newsletter, Yerkes Primate Center, v. 3, no. 1, p. 15-30. Schuman, E. L., and Brace, C. L., 1954, Metric and morphologic variations in the dentitions of the Liberian chimpanzee; comparisons with anthropoid and human dentitions: Human Biol., v. 26, p. 239-268. Scott, J., 1963, Factors determining the skull form in primates: Zool. Soc. London Symp., no. 10, p. 127-134. Simpson, G. G., Roe, A., and Lewontin, R. C., 1960, Quantitative zoology: Harcourt, Brace, New York, 440 p. Shackleton, R. M., 1951, A contribution to the geology of the Kavirondo rift valley: Geol. Soc. London Quart. Jour., v. 106, p. 345-392. Simons, E. L., 1961, The phyletic position of Ramapithecus: Postilla (Peabody Museum, Yale University), no. 57, 9 p. 1963, Some fallacies in the study of hominid phylogeny: Science, v. 141, p. 879-889. 1964, On the mandible of Ramapithecus: Natl. Acad. Sci. Proc., v. 51, p. 528-535. 1965a, New fossil apes from Egypt and the initial differentiation of Hominoidea: Nature, v. 205, p. 135-139. 1965b, The hunt for Darwin’s third ape: Med. Opinion and Rev., November, p. 74-81. 1967, The earliest apes: Sci. Amer., v. 217, p. 28-35. 1968, A source for dental comparison of Ramapithecus with Australopithecus and Homo: S. Afr. J. Sci., v. 64, p. 92-112. Simons, E. L., and Pilbeam, D. R., 1965, Preliminary revision of the Dryopithecinae (Pongidae, Anthropodea): Folia Primat., v. 3, p. 81-152. Sokal, R. R., and Sneath, P. H. A., 1963, Principles of numerical taxonomy: Freeman, San Fran- cisco, 359 p. Straus, W. L., Jr., 1963, The classification of Oreopithecus, in Washburn, S. L. (ed.), Classifica- tion and human evolution: Aldine, Chicago, p. 146-177. Tuttle, R. H., 1967, Knuckle-walking and the evolution of hominoid hands: Amer. Jour. Phys. Anthrop., n. ser., v. 26, p. 171-206. Van Couvering, J. A., and Miller, J. A., 1969, Miocene stratigraphy and age determinations, Rusinga Island, Kenya: Nature, v. 221, p. 628-632. 138 PEABODY MUSEUM BULLETIN 31 Verdcourt, B., 1963, The Miocene non-marine mollusca of Rusinga Island, Lake Victoria and other localities in Kenya: Palaeontographica, v. 121, A, p- 1-37. Vogel, C., 1961, Uber den phylogenetischen wert von Mandibelmerkmalen bei héheren Primaten: Zeitschr. Morphol. Anthrop., v. 51, p. 275-288. 1965, Uber den Nahtverlauf im Gesichtsskelett catarrhiner Primaten einschliesslich des Menschen: Anthrop. Anz., v. 29, p. 301-312. Walker, A. C., 1967, Patterns of extinction among the sub-fossil Madagascan lemuroids, in Martin, P. S., and Wright, H. E. (eds.), Pleistocene extinctions: Yale, New Haven, p. 425-432. Walker, A. C., and Rose, M. D., 1968, Fossil hominoid vertebra from the Miocene of Uganda: Nature, v. 217, p. 980-981. Washburn, S. L., 1963, Behavior and human evolution, in Washburn, S. L. (ed.), Classification and human evolution: Aldine, Chicago, p. 190-203. 1968, Perspectives and prospects: Amer. Jour. Phys. Anthrop., v. 27, p. 376-373. Washburn, S. L., and Jay, P. C. (eds.), 1968, Perspectives on human evolution: Holt, Rinehart and Winston, New York, 287 p. Weidenreich, F., 1941, The brain and its role in the phylogenetic transformation of the human skull: Amer. Phil. Soc. Trans., n. ser., v. 31, pt. V, p. 321-442. Whitworth, T., 1953, A contribution to the geology of Rusinga Island: Geol. Soc. London Quart. Jour., v. 109, p. 75-96. 1958, Miocene ruminants of East Africa: Fossil mammals of Africa, no. 15, Brit. Mus. (Nat. Hist.), 50 p. TERTIARY PONGIDAE OF EAST AFRICA 139 Figures 14—31 140 PEABODY MUSEUM BULLETIN 31 FIGURE 14. Bivariate plot of Pa Dryopithecus species, Pan troglodytes, and Gorilla gorilla with 95% equiprobability ellipses. 6 M-D length 9 12 5 TERTIARY PONGIDAE OF EAST AFRICA 141 FIGURE 15. Bivariate plot of M, Dryopithecus species, Pan troglodytes, and Gorilla gorilla with 95% equiprobability ellipses. > ¢ M-D length 142 PEABODY MUSEUM BULLETIN 31 FIGURE I6. Bivariate plot of M5 Dryopithecus species, Pan troglodytes, and Gorilla gorilla with 95% equiprobability ellipses. M2 G. gorilla ° D(E)maor x Datricanus | P troglodytes IS I7 I9 9 ut 13 M-D tength TERTIARY PONGIDAE OF EAST AFRICA 143 FIGURE 17. Bivariate plot of M3 Dryopithecus species, Pan troglodytes, and Gorilla gorilla with 95% equiprobability ellipses. M3 7 & gorilla 15 13 Max B-L Pf troglodytes 0.(P)major x D(P)africonus 9 Il 13 M-D length 15 \7 19 144 PEABODY MUSEUM BULLETIN 31 FIG. 18. a and b, occlusal views of UMP 62-11 from Moroto II. x 1.5. TERTIARY PONGIDAE OF EAST AFRICA 145 Fic. 19. a, lateral view of UMP 62-11 from Moroto II. b, anterior view of UMP 62-11 from Moroto II. 146 PEABODY MUSEUM BULLETIN 31 FIG. 20. a, occlusal view of UMP 62-10. b, occlusal view of UMP 66-01. TERTIARY PONGIDAE OF EAST AFRICA 147 FIG. 21. Occlusal views of: a, UMP 62-07 from Napak V, stereophotograph. x 2.2. b, UMP 66-41 from Napak IV, sterephotograph. x 2.1. c, UMP 62-08 from Napak V, stereophotograph. x 1.9. d, NMK 405,381 from Songhor. x 2.6. 148 PEABODY MUSEUM BULLETIN 31 FIG. 22. Stereophotographs, occlusal views of: a, UMP 62-06 from Napak V. x 1.6. b, UMP 62-13 from Napak I. x 1.8. c, UMP 62-14 from Napak I. x 2.3. TERTIARY PONGIDAE OF EAST AFRICA 149 FIG. 22. Stereophotographs, occlusal views of: d, UMP 62-15 from Napak I. x 2.4. e, UMP 62-16 from Napak I. x 3.0. f, UMP 66-02 from Napak V. x 3.0. 150 PEABODY MUSEUM BULLETIN 31 FIG. 23. Occlusal views of: a, UMP 62-16. x 14. b, BM(NH) M14086. x 1.4. c, BM(NH) M14086. x 1.2. d, BM(NH) M16648. x 1.2. TERTIARY PONGIDAE OF EAST AFRICA 151 FIG. 24. a, buccal view of UMP 66-41(top) and UMP 62-14. x 2.4. b, lingual view of UMP 66-41(top) and UMP 62-14. x 2.4. c, lateral view of UMP 62-08(top) and NMK 190,1. x 2.2. 52 PEABODY MUSEUM BULLETIN 31 FIG. 25. Stereophotograph, occlusal view of BM(NH) M14084. x .85. FIG. 26. Occlusal view of NMK 190,1 from Songhor. x 1.1. TERTIARY PONGIDAE OF EAST AFRICA 153 S FIG. 27. BM(NH) M16647: a, M2, M3. Occlusal view. b, C1, P3, P4, M1. Occlusal view. c, M2, M3. Buccal view. d, C1, P3, P4, M1. Buccal view. PEABODY MUSEUM BULLETIN 31 154 ‘OL X ‘6FOOLIN (HN)WA JO Mata [esnpI0 ‘yders0j0ydoaiais “gg ‘914 TERTIARY PONGIDAE OF EAST AFRICA 155 sSaonenal, 52 x FIG. 29. Reconstruction of Ramapithecus punjabicus. Right maxilla, YPM 13799. Left maxilla, NMK Ft1272. Left M3, GSI D186. x 1.6. FIG. 30. Occlusal views of: a, D. (P.) major, UMP 66-41. b, D. (P.) major, UMP 62-07. c, D. indicus, YPM 13834. PEABODY MUSEUM BULLETIN 31 156 “yisuzy cW ‘TW owes 94} OF 7Yy8no1q ITV ‘snaiqvlund snaayjidvuvy ‘9 ‘(snoayuiqvuvy jo 1ostsul yam) snasiupd uvd ‘Gq ‘snupaifv (qd) qd ‘& “1g “914 i 2. Wo{)olulan Wil TERTIARY PONGIDAE OF EAST AFRICA 157 APPENDICES The following abbreviations have been used: Spec no Comp no Specimen number Computer card number Material from the following institutions has been used in this study. CODE A. B C SOURCE Anatomy Museum, Cambridge University Zoology Department, British Museum (Natural History), London Museum of Comparative Zoology, Cambridge University Duckworth Laboratory of Physical Anthropology, Cambridge University Anthropologisches Institut der Universitat, Zurich 158 PEABODY MUSEUM BUELETIN’ 31 APPENDIX la Gorilla gorilla male maxillae Comp No 11001 LOO? 11003 11004 11005 11006 11007 Spec No 7 6598 2 4802 26503 2 4902 2 6599 Z 6672) Meco ci Max L 19. 4 20.4 23250 DA? 20.6 27 19.0 Trans B 16.8 14.0 15.2 16.6 16.4 17.8 15.6 MDi L 10.5 10.6 Pie 12.0 10.8 11.4 10.5 P= Max L gee 12.3 1322 13.0 12.8 1226 11,4 BLi B 17.0 15.5 14.8 16. 6 16.0 14.3 16.2 pe MDiL 10.8 10.8 11.6 le? 1027 V4 11.0 BLi B 15.8 14.7 15.4 16.8 15.0 14. 6 16.0 vi MDiL 13.6 14.4 1559 15.3 14.3 14.4 14.6 BLi B WGai2 14.3 15.9 15.5 14.3 14.8 16.5 m2 MDiL 16.3 15.4 17.9 16.8 14.7 162 15.8 BLi B 16.8 16.4 16.9 17.8 15.5 17510 GA (0) m3 MDiL 15 eu5 14.6 16.6 15. 6 ies? 14. 6 15.0 BLi B 16.3 15.5 16.1 16.9 14.6 14. 8 15.8 PM oi 68.5 67.0 73.0 70.5 63.5 G73 677 Sea: 39.5 40.4 *44,0 44,5 39.5 *43.0 42.3 C8 78.3 69.4 *76.0 76.5 T225 *74,0 79.4 Poeae 70. 8 67.8 75.5 73.5 68.3 64.5 78.6 M*? B 75.8 70.8 76.5 70.2 68.7 68.5 73 Comp No 11008 11009 OO Laon LOT? 11013. Manoa Spec No Z 6603 2.6602 2 6504 B 1939 9B AZ5 B 9JomBezsee 913 As a2 ct MaxL 2352 22.3 20.8 19.9 19.2 2G 23,2; Trans B 16. 6 1723 16.0 14.8 13.3 16. 2 LOe MDi L 1220 1054 11.4 10. 6 10.2 10.9 10e0 P> Max L 1225 12.0 Vie7 12.3 11.0 12.2 11.6 BLi B 15.2 15.4 14.0 15.0 14.3 17.0 16.8 pt MDiL 11.8 10.5 10.5 10. 6 *10.3 *11.5 *10.8 BLi B 15.1 15.0 15.2 14. 4 1453 16. 0 *15.3 mi MDiL 15.8 14,4 14.3 #9327 *14,4 15.1 *14,1 BLi B 16.4 15.8 15.6 14.5 16.0 15.8 15.9 uz? MDiL 17.0 15.3 15.3 *15.0 14.7 16. 4 *15.4 BLi B 16.7 16.6 16.8 16.1 16.3 16. 4 16.8 um? MDiL 16.5 UO 7 14.4 14.7 *14,3 *15.6 15.2 BLi B 16.8 15.6 16.0 15.4 15.6 16.4 16.6 Pp? mw? L Fan8 65.8 65.2 64. 0 63.5 68.5 65.0 TW» AB = 44,5 44,8 38.5 2623 40.5 40.7 Con: *82.0 75.2 77.9 68.0 69.0 *68.3 78.0 pe: 1B (ales *74,0 *73.0 68. 8 69.0 66. 6 Ue 3 M B (ORS Fle 3 O77, 70,0 68, 1 68.3 (Mot TERTIARY PONGIDAE OF EAST AFRICA 159 APPENDIX la, continued Comp No 1 FOS 11016 110,27 11018 11019 11020 Spec No B 48. Bo4igi D Pr5i2, A G2 A G28 Rope? 436 603 On ne ee Se gn ee i ee eee ci Max L 22.0 22.3 20.8 2352 20.1 20.3 Trans B 16.5 16.0 USS 7/ 16.8 15.4 WSS 7/ , MDIL 10.0 11.4 9.7 Zhe 10.4 Wt, 2 P MaxL 11.5 13,2 122 12.9 12.6 12.0 BLi B 14.3 16.4 14.6 16.4 14.9 iy ae pt MDIL ileal *12.0 10.5 1222 10.7 11.5 BLi B 14.4 15.5 14.0 16.1 15.3 16.3 1 MDiL 14.8 16.2 14.0 16.7 14,8 16.1 BLi B Ara 16.0 14.7 1724 15.6 17.2 m2? MDIL 15.2 17.5 14.5 16.5 16.9 W7e0 BLi B 14.3 16.5 15.4 L728 17.5 18.3 me MDiL 14.5 17.0 15.0 14.7 *14,0 lyf BLi B 13.0 16.8 Wy 15.3 15.0 72 p°_M® L 64.0 72.0 65.3 68.5 68. 1 72 r B 39.2 42.6 42.6 43.3 *41.3 44.0 BE? 8B 67.5 75.8 78.2 *79.0 80.1 *75.5 P B 66.0 Vac4 70.4 74.4 72.9 Wael M B *70.5 76.6 65.8 ice 75.3 71.5 n X + SE SD Vie OR 95% CL IN Geet iw ee eps oe Se Se ee 1 MaxL 20) © 29,2020. 31 1397 6.551) 19.,0=2852 18. 32-24. 08 Trans B 20 15.9740. 24 1.07 6,814.7 13.3=17.8 13. 72-18, 22 MDi L 20 10,90+0.16 0.72 6.68 fae) 9, 39-12. 41 p> Max L 20 12. 28+0.16 0270. 5.75 1T.0-13.2 10. 81-13. 75 BLi B 20 15.59+0. 23 1.04 6.79. 140-17. 1 13. A1=17. 77 p+ MDiL 20 11,08+0.12 0.55, 502. “10, 3—125.2 9,92-12. 24 BLi B 20 15.26+0.17 0.74 4,94 14,0-16.8 13-71-16, 81 mi MDiL 20 14,82+0.19 0.85 5.83 13.6-16.7 13. 03-16. 61 BLi B 20 15.62+0.21 0.92 5.98 14,1-17.4 13. 69-17. 55 m2 MDiL 20 16.00+0.22 1.00 6.34 14.5-17.9 13. 90-18. 10 BLi B 20 16.65+0.20 0.90 5.49 14.3-18.3 14, 76-18. 54 vm? MDiL 20 15. 02 +0. 22 0.97. 6.53 13.2-17.0 12. 98-17. 06 BLi B 20\ 15. 72:40; 22 100 6.46 13.0-17.2 13. 62-17. 82 P?—M 20 67.6040. 67 3.00 4.50 63.5-73.0 61. 30-73. 90 mB 19 41. 6640555 2.38 5.78 36.3-44.8 38. 64-46. 68 eB 20 75.03+41.00 4.46 6.02 67.5-82.0 65. 65-84. 41 P’, B 20° 7.270279 3.55 5.05 64.5-78..6 63. 81-78. 73 M B 200) | 7t209210..75 32316 04s 7710 OSe0-7 7008 64, 34-78, 44 160 PEABODY MUSEUM BULLETIN 31 APPENDIX 1b Gorilla gorilla female maxillae Comp No 12001 12002 12003 12004 12005 12006 22009 Spec No Z 1224: 26592 °2 56985 26595 2 6593. 2 fr, @aggea Cam. ct MaxL 14.6 14.3 14.6 15.9 14.4 14.6 12.5 Trans B 10.7 127, 12.5 TAS 11107 es, 10.3 MDi L 10.0 eA 9.5 10.7 10.9 10. 2 1052 P> MaxL 11.4 11.8 9.9 1156 11.2 11.0 10.3 BLi B T5443 16.3 14.3 16.4 16.2 15.6 13.5 pt MDiL 10.3 110 9.5 10.7 na ie 10.5 9.7 BLi B 14.6 1sF4 13.8 15.3 15.6 15.5 12.8 wt MDiL 14.3 14.5 1237 15.4 15.3 13e5 13.4 BL B 14.8 15.5 14.8 15.8 15.5 1527 13.4 m2? MDiL 14.8 15.5 13.5 16.8 15.4 13.8 14.2 BLi B 15:7 16.8 15.3 16.9 172 15.8 14.0 um? MDiL 14.7 13.8 T2NO 15.9 15.3 13.0 12.6 BLi B 15.2 16.8 14.6 15.4 16.2 15.8 13.0 P°_M* L 62.3 65.6 Siete 67.8 67.7 S2ac 59.8 Tr B 37.7 39.7 33.8 *37.0 40.2 37.7 36.3 Gis 59.0 74.0 65.3 58.8 52.7 57a2 45.2 me B 63.8 792 66.7 *63.0 68.3 61.5 57.3 M B 66.4 68.4 65.3 67.3 67.4 *66.7 56.6 Comp No 12008 12009 22020 12012 422012 22013 enone Spec No % 6676 Z 6840 Z 6594 B 1939 B 1939) (8B T9e89NR mers 935 925 933 936 lL MaxL 14.4 TAs? 14.3 15.5 15.1 13.2 14.1 Trans B TA 11.2 17.5 HG 2 12.0 10.1 11.0 , MDIL 1022 10.4 9.7 10.5 1023 9.8 9.2 P~ MaxL 10.3 11.0 10.6 eee 1207 10.3 10.1 BLi B 13.9 14.3 14.6 15.5 15.6 13.5 15.0 pt MDiL 10.4 10.3 10.3 *11.6 *11.9 9.7 11.0 BLi B 13.9 14.2 14.2 15.6 = 1362 15.6 vit MDiL 14.0 1355 13.5 14.5 15.0 1354 14.3 BLi B 15.4 14.2 15.0 1e7 15.2 1S53 14.1 ve MDiL 14.1 15.4 13.4 15.6 16.2 14.7 15.0 BLi B 15.7 16.4 14.7 16.6 15.5 14.6 16.0 um? MDiL 12.8 14.4 122 14.5 13.2 1325 14.4 BLi B 13.4 15.6 13.6 16.0 *14.0 13.6 15.9 PML 6123 64.4 50.5 *65.0 66.0 *62.0 62.7 TB 37.:8 37.6 39.5 *38.0 42.6 34.0 86.7 ct OB 52.2 57.2 61.4 60. 0 63.6 bans 60.3 P* BB 52.5 61.7 61.2 62.8 *66.5 55.6 64.5 M* B *66.0 63.4 58.3 *62.0 *65.0 58. 2 67.1 TERTIARY PONGIDAE OF EAST AFRICA 161 APPENDIX 1b, continued Comp No 12015 12016 12017 12018 12019 12020 Spec No B 1939 A G20 A G21 A G22 A G23 A G0 927 ct MaxL 16.0 14.0 14.1 sae 14.3 14.6 Trans B 11.0 10.5 10. 2 11,2 10. 3 10,2 MDi L 10. 6 9.4 9.5 10.9 10,7 9.8 P® Max L 11.6 10.9 10.3 hee, {1,3 10.5 BLi B 15,1 14.3 13.0 14.6 14.8 14.6 p* MDiL 10.9 10.3 10.4 1167, 10. 2 9.8 BLi B 14.9 13.8 12°65 14.6 14.6 14.2 wit MDiL 14.4 1351 14.3 14.7 14.1 13.4 BLi B 150 14.0 13.6 14.8 15.1 14.7 wm? MDIL 15.2 1327 15.2 15.0 15.0 13.8 BLi B 16. 0 12,7 14. 6 16. 0 15.5 15.1 wv? MDiL 14.0 12.8 13.8 14.1 13.4 1373 BLi B 16.3 13.8 13.3 14.9 14.3 14.2 P? —M* L 64.1 59.3 60.8 64.5 *60. 0 58.0 7. B 39.0 Aly poae" | 35.5 Aly. a7 36 34.1 ec .B 55.0 *59.5 53-2 63.3 54.5 57.3 r B 62.0 58.6 57.4 62.3 58.2 65.5 M B 63.8 60.8 63.5 63.6 *62.0 66.8 n "‘X +SE SD V OR 95% CL cor ci MaxL 20 14.50+0.18 0.82 5.72 12.5-16,0 12. 78-16, 22 Trans B 20 =11.18+0.18 0:75) 7212 10, 1=12-7 9, 52-12. 84 ,; MDIL 20 10.20+0.13 0.58 5.79 9.2-11.4 8.98-11.44 P MaxL 20 10.96+0.15 0.69 6.35 9,912.7 9,51-12,41 BLi B 20 14.82+0.21 0.95 6.50 13.0-16.4 12. 82-16. 82 p* MDiL 20 10.57+0.15 0.68 6.50 9.5-11.9 9. 14-12. 00 BLi B 19 14.44+0,22 0.94 6.62 12.6-15.6 12. 46-16. 42 wt MDiL 20 14.07+0.16 0.74 «5,31 (12.7=15:4 12. 52-16. 62 BLi B 20 14.79+0,17 0.78" 5332) “13,3-15.6 13. 15-16. 43 wm? MDiL 20 14.82 +0.20 0.92 6,27 13.4-16.8 12. 84-16. 75 BLi B 20 15.63 +0.20 0.90 5.86 14.0-17.2 13. 74-17. 52 m? MDLL 20 13.69+0.23 TOU -7.45 1220-1559 11, 57=16, 81 BLi B 20 14.8040.26 1.16 7.95 13.0-16.8 12. 36-17. 24 P>—mv? L 20 61,59 £1.02 4.57 7.51 50.5-67.8 51.99-71.19 th 2B 20 ° 37.68 £0.53 2.36 6.34 33.8-42.6 32. 72-42. 64 cB 20 58.16+1.34 6.00 10.45 45.2-74.0 44. 96-71. 36 pe SB 20 +62. 0821, 02 4-64-7241 * 52.5-71.2 52. 48-71. 58 M* B 20 63.93 £0.76 3.39 5.38 56.6-68.4 56,91-71.05 PEABODY MUSEUM BULLETIN 31 162 Goce Commerce Ose mG ee= secl7 ~2ree = S°22 S"eZ. — OOfe —“S*Pe- “Ste =e"TZ” oOr6ze" -qaopqueyr Poccuae 0 7a OMY ame Ocoee ac'27—, GecOce) sOscea. “S70 -." 9°92 lOrSza aS te = e4ce Sez Sez L ew Cie V re ecco a Gp 98cer = S407 e°Sr O0re Sly O86 21053 « BLiB 124 213. 45 20 s0 1257 A352 Lies 120) Sie 2562.3 MDI eo. 0) 1556" ©1453 MaxsBe 12.1 “13577 + 11. 6 ° terri 8 UZ SASS HL 6 WZ Tal B Zee 13.7 4 M MDI SUS. 7 C17 e.-2 MaxiBiEel47 clsadi) 13.2 5 eeecrieB IN, alas © iSi5. 24 14, Tal B 1336) 1459" > 13.0 MDItGe pLow2) Alle 5.4 M MaxiB 113.9 (1450 13.53 14. Setri 8 13.9 14.0 13.3 0 Tal B 1201 13c4 12. PyoM, 1 Ws. = Tess) (a68) 7055 7525 6856 69.2 68.8 68.6 65.7 I B de) E2018) Zone Zag — Alfeli AYGS PRIA PRN AS) GY) G) B 44.3 50.1 42.1 = 43.8 *43.8 41.4 *42.0 44.3 *44.0 RB B 54.2 58.1 49.1 94.3, 50.7 )s9l8 90.3: 51.6 #5150 51.0 Mz B 61.1 62.8 62.8 64.8 64.7 60.5 62.6 58.6 *64.0 60.4 me oD 3629/5 39.9 31.10 33.0, 0.1 134.8 34.6 34.3 33.0 36.4 M D 36.38 §37.,0 = SGC Osos ace! 31,7 (33.0 29.2 (32.6 Sr PASO PASO AN 7/ RY KES PRG al 26510!) °235 28 2257 2657 MentforD 22.0 26.4 21.0 210 22.0) 2056 21.2 22.1 20.7 25.4 166 PEABODY MUSEUM BULLETIN 31 APPENDIX ld, continued X + n x ESE SD Vee OR 95% CL Cc. Max L 18 13.00+0.21 0.88 6.88 11.3-14.6 11. 13-14. 87 a; trans B 18 10.33+0.16 0.66 6.50 9. 1-11.15 8. 93-11. 73 Pp Max L 20 14,.87+0.19 0.85 7.24 13.7-16.4 13. 08-16. 66 ® Trans B 20 10.36+0.21 0.94 9.19 8.2-11.5 8. 39-12.33 Pp MDiL 20 10.76+0.12 0.54 5.05 9.8-12.1 9. 63-11.89 @ BLiB 20 12.52 +0. 24 1.06 8.61 10.5-14.4 10. 30-14. 74 MDiL 20 14,88 +0.14 0.63 4.29 13.6-16.5 13. 56-16. 20 M Max B 20 12.80+0.17 0.75 6.00 11.6-14.0 11. 22-14.38 2rd B 20 V2 ot 0218 0279" 16.38° Tr s=1450 10. 95-14. 27 Tal B 20 12,.68+0.16 03.73) 7. 5286) 24-13 57 11. 15-14. 21 MDiL 20 USS 2474 Se (WG alts} 0.81 5.09 14.7-17.8 14, 52-17. 92 M Max B 20 14,41+40.25 Tass H/aqtisy AAS tao Hays 5) 11. 93-16. 89 ae irl B 20 14.36+0.25 1.14 8.01 12.6-16.5 11. 96-16.76 Tal B 20 13.99) 320523 104 7658 L2.T-1650 11. 77-16. 13 MDiL 20 16.00+0.25 Yol2Z 7.08 1452-1852 13. 65-18.35 M Max B 20 14,19 +0.24 PO8se eZ 4—llGeZ 11. 92-16. 46 3 Tri B 20 14.19 10524 1508 7.72 * 22.4-1602 11. 92-16. 46 Tal B 20 125.67 20.28 124° 92:97) TOS8=15.0 10. 07-15. 27 PF, —M, L 20 TOA ORT Sos 4.79 60.7-77.0 64. 04-78. 04 TY 3B 20 26.39+0.41 1386) 9 s7eliZ) 232-2958 22. 49-30. 29 C, B 20 44,74+0.74 3.24 7.35 40.8-54.0 37. 94-51. 54 i 18 19 H2520 209 4.76 9.23 42.0-62.0 42. 23-62. 23 M, B 19 61.96+0.71 3.10 5.06 53..2-66.4 54. 45-68.47 1 3D) 20 36.50+0.62 2:77; 7.69 31.0-42.0 30. 68-42.32. M D 19 35625060 Penh xe) Ae ER) 29. 77-41. 73 3 T 20 24.73 £0.40 W749, Wen, led =—Z28.10 20. 97-28. 49° Ment for D 20 23.70+0.69 3. Ul 13527 195-3016 17. 17-30. 23 TERTIARY PONGIDAE OF EAST AFRICA 167 168 PEABODY MUSEUM BULLETIN 31 APPENDIX le_ Pan troglodytes male maxillae Comp No 21001 - 21002 21003 21004 212005 21006 ~2%0ne Spec No Z Z Z Z Z Z Z ci MaxL 12.9 14.9 15.3 14.0 12.5 13.6 15.6 Trans B 10.6 9.6 13.0 10.4 10.0 10.5 13ha MDi L 6.8 7.6 7.4 72 7.6 72 7.4 P? Max L 7.0 8.3 734 .0 8.6 8.0 728 BLi B 11.0 10.4 11.3 - 10.3 10.2 10.7 p+ MDiL 6.6 6.8 129, 6.6 7.3 7.5 726 BLi B 9.8 9.6 10.7 - 10.4 9.8 10.5 wi MDiL 9.5 9.2 10.4 9.8 1022 9.6 10.6 BLi B 10. 6 10.4 alee - 10.8 10:7 11.7 m2 MDiL 9.6 9.6 10.0 9.3 10.4 9.8 10.5 BLi B 11.2 11.0 7 - 11.5 11.5 12.3 wm? MDiL = 9.0 8.2 - = * 9.8 9.4 BLi B - 11.0 10.5 - - 12.0 ob ee P> —M? L 41.6 42.4 43.3 42.4 47.3 43.3 46.2 = 8 37.6 *37.0 36.4 35.6 39.9 34.4 40.2 Gy 8 57.0 56.5 56.6 58.8 56.3 59.0 61.6 Paawe 55.7 58.4 55.8 56.6 53.8 55.0 62.3 M? B S72 60.4 53.2 53.5 49.5 53.8 57.8 Comp No 21008 210095 21010 Z1011 22012 21078) ron Spec No Zz Zi, az B 87 D pr53, A PiS:) (eee 122 0.2 7URTA ci MaxL 15.4 15.4 13.4 *12.0 12.3 14.3 14.0 Trans B 12.2 11.9 V4 * 8.7 9.6 10.8 1037 MDi L 735 7.4 6.5 #'S5.0 8.0 6.4 6.7 P® MaxL 8.5 7.4 7.5 8.7. 9.1 se 8.4 BLi B 10.8 a7) 10.2 *10.0 i ee * 9,4 10.0 p* MDiL 702 Fan 6.7 +-7=() 710 6.8 7.0 BLi B 10.2 10.5 1on2 *10.0 10.5 10.3 * 9.8 wi MDiL 10.4 9.5 10,2 *10.6 * 9.5 10:2 9.0 BLi B U1 7 11.2 13 *11.6 - 7 10.9 m2 MDiL 11.0 10.0 IRS} 10.4 ih ee | 9.3 10.0 BLi B V7 11.6 12.4 11.6 1228 11.6 Lise m? MDIL 9.3 9.5 9.3 8.6 1053 ox7 9.3 BLi B 12.2 1087 12.0 * 8.8 Ui7 10.3 10.8 pP?—mM? L 44,3 42.3 ASu7: 44,1 46.0 41.6 42.3 ee 38.6 28.2 36.5 38.6 38.7 38.5 34.9 Cc" Bp 57.9 60.2 57.8 55.7 59.2 62.3 57.8 p* 8B 60.5 56.5 58.3 59.4 *60.5 57.6 56.6 M? B 56.2 55.5 57.2 56.8 58.2 56.1 54.2 TERTIARY PONGIDAE OF EAST AFRICA 169 APPENDIX le, continued n X + SE SD ee Neor OR 95% CL ct MaxL 14 =: 13. 96+ 0.33 1223) 38295) 1250-1556 11. 33-16. 59 Trans B 12> 10,902 0595 3.3% 12-21" 6e7=1322 8. 10-13. 70 MDiL 14 7.264+0.13 0.50 7.05 6.4- 8.0 6.19- 8.33 P> MaxL 14 8.06+0.16 0.59 7.46 7.0- 9.1 6.80- 9.32 BLi B 13 «1047 £0005 0553 95.14 97A=1053 9, 32-11. 62 p* MDiL 14 7.04+0.08 0.32 4.56 6.6- 7.6 6.35- 7.73 BLi B i3) S10) 19:40,.10) 9 0235" 6.47. 9.6-1057, 8. 42-10.94 mt MDiL 14 9.91+0.14 0.53 5.43 9.0-10.6 8. 78-11. 04 BLi B 12 11,15 +£0.13 0:47 4.27 \10.4-1127 10. 12-12. 18 m2? MDiL 14 “205164017, - 0.65 6,47 9. 3-113 8.77-11.55 BLi B 13 11.70+0.14 0.51 4.47 11.0-12.8 10. 60-12. 80 me MDiL 11- 9.22+0.18 0.58 6.45 98.2-10.3 7.94-10.50 BLi B 11 10.98+0.30 1.00 9.30 8.8-12.2 9. 78-12. 18 P°—M°L 14 43.63+0.48 1.78 4.16 41.6-47.3 39. 82-47. 44 = 8 14 37.464+0.46 1.72 4.66 34.4-40.2 33. 78-41. 14 GC B 14 58.34+0.53 1.99 3.47 55.7-62.3 54, 08-62.50 P* B 14 57.64+0.64 2.39 4,23 53.8-62.3 52.53-62.75 M? B 14 55.69+0.72 2.69 4.92 49.5-60.4 49, 93-61. 47 22004 22005 22006 22003 Pan troglodytes female maxillae 22002 PEABODY MUSEUM BULLETIN 31 APPENDIX lf 22001 170 Comp No Spec No Max L Trans B oy “N mo OK RN NM COC CADDO en RO MOD FO ee. « | peta J SoS [ Jeet } . Ce eed | . . . (ya I . ° NO OH ON DH © DHODSO Oo mR SM ROD ND ei ae are ei sv mmwm wo N et KOR x * Nc HDwo ON NO at HH DOD COOW ae © NO mNO OF >. @ O77 ae. . ° [ Yea} . Li te at) e ° e ° . C©S90 DRO DADH DADO NHNADL onm aAMnD ROW NRO eS re ret fe = munmnwumnwm Neto eit ei ce * * a Q fos) SOM NO NN NM © RADSHO A ON MMO Hx! (eval ) es. ¢ S18 s e . eh. ee e ° . * e °. MRO OH OH DO Ht HMB ° NO nna NO et ae ari et = mumwum Ww N an | [on | eo * NN ; roy on Om DBO FTN MO SHON OO Oo mon mot ON * «@« . *@« *. « *. ese e e. 8 . . eo @:°e - 8¢@ COO DH ODN DHA NN COHODO ° OD DVDWOO WOO qo a) ae) at mMmnnNno N o = a NN fo) mn NR wre mo oOmMmm oO st no ro] e.hCOUCs e.|.|Um6 é: . .@ e oe CY a) . e e RO on an ! 1! mM nunoaw ° val nol ni re et ei b= mOumWwm WwW N Ce | * * AN N ron TR RN NO MH HMOORN Oo ODMR COM NRW e e« . « . . . ¢ . . . ° Oo . . e « ° . OD AH DO HOM tHnANw SCO RDO OD a at mnnw N eo 4 NN © | | P| 4 pes | H 4 4 Ae ga 6 sae eycee — eo eae, 0 Qa Ga Ga ga 2 | és Gan az ao AQ oa iT aamN em aie ei a ee ea Eo A n er re) et J o Q, ”% + ee a a HOMES oa re) Au 4 50. 9.4 11.2 10.4 14 9.4 10.8 9.8 10.7 43.0 54.8 6 -8 .0 10. 36.8 51.2 -6 10.4 LS7 tite *42.0 36.3 BLi B BLi B TERTIARY PONGIDAE OF EAST AFRICA 1741 APPENDIX lf, continued n X + SE SD Vv OR 95% CL cor oc: MaxL 12 “J1.2840.15'. 0253 "4384 10,5-1232 10. 11-12. 45 Trans B iI 8.9640.10 06:39 3277 8.2-'943 8.12- 9.60 MDiL 12 702 £0501" (0538 5.54." 6S3=725 6. 18- 7.86 Pp? MaxL 12 7.93+0.13 0.44 5.62 6.8- 8.5 6.96- 8.90 BLi B Jl 10.54+0.18 0.59 5.71 9.8-11.8 9. 22-11. 86 pt MDL 12 6.98+0.08 0.28 4.13 6.6-7.4 6.36- 7.60 BLi B TY ows 0.12, (0227 985389" 9. 5—1078 9.55-10.75 wt MDiL 12 9.83+0.15 0.53 5.48 9.3-10.6 8. 66-11. 00 BLi B TY DDE Oss «= «0.59 5.41 |) 99-1157 9.89-12.53 wm? MDiL 12 9.974+0.17 0.59 6.01 9.0-10.8 S5G67—-1le 27 BLi B Jl 11.44+0.20 0.65 5.83 10.2-12.4 9,99-12.89 me MDiL 10 9,01+0.19 0.59 6.67 #8.4-10.3 7. 68-10.34 a, BIB 9 10.8740.20 0.59 5.56 10.0-12.0 9751-02028 p® -mM? L 12 43.58+0.46 1.60 3.74 40.8-46.0 40. 06-47.10 T 2B 12. 36.1840,44 “Ins 4.27 ~ 34/3-39:6 32. 86-39. 50 cB 12 54.61+0.78 2.70 5.05 50.0-58.0 48. 67-60.55 Pe B 12 57.53+0.69 2.39 4.25 52.6-60.6 52, 27-62. 79 M B 12 56.8340.79 2.74 4.92 50.7-60.8 50. 80-62, 86 PEABODY MUSEUM BULLETIN 31 172 PI 92-82 “EL Z°SZ-1T ST €4°M68°2 42°0+96 61 FI SP Glec te Ove Goce Bl S22 ose 821 6 21 8°61 S°8T S°6T Cioy qua Gyelénos ato e690 hieZ bel Se 106oset Ti TeG1 G91 921 9 0c0e0¢ rol O02 2°21 e° 21 ©) ZN Si ala St Eh Ley 8h VE-9S EZ S*VE-V'PZ €6°8 SS°Z 89°0+20°62 FI Gee 1a 6cad O80 Ze eS 8c Oe ss eae, Iene 162 Bae a ZS 9E-Zh EZ Z°9E-S SZ BETOLIO"E 28°C 44662 FI 8°62 ZTE O°ZE S°ZE VEZ O'S Z°9E 2°82 S°LZ v°L2 a ‘a O68 GS-97 16r 16 25-0) US 99nv0S 2% 29 019 7S 71 Ber SiuG 25" 8 79S te eSeceeSe SOS 0 TSa0) Ts 79S Vesa dW Hee OS-t6 8cs €°Ps Ges 6 °SP ap Gl ES-OS=L9s68" O8t=c- LY BiG Ce On OMS ea aS eaCVe Gas is eG. OV 0°87 9°vT 0°87 q ‘9 All ALA ASA (0) IPE IeG SHG Pie MO ea ey eA Aly M0) te Teel ts, wtOlts €°62 6°82 vy °972 camel [bie PASM Air? (0 er SSS) a/4i7 Vou O04 74 by i Zecy 6 2y 60S 8°67 8'8h S*6r 8'°8r T N—T SP ‘OI-T6°8 CaO 46 S€°0 O1T°0 +89°6 GN Sil AN 15} xf) 66 Ty A Gy YA a1eL VOT /h U=9) 255 9€°O O1T°O +S6°6 Gi 7a YAN 11S 9°6 SiG OnGen 9s : dMlewy vL°OT-02°6 Z°0I-9°6 se°0o OT°0 +26°6 Cine Onn ONO O ONG) 270i 86) 926) Z6e = aEXEIN OSS OL=2.e76 900-556 SExO> OL O6OMOL Gl. 7 0l. SOL 836 O-0n 86> 6.0L 2 6 SOL” GIN 6S *L1-S6°8 TSGE=S 46 OS Oe Ome 0 eC ahs Ole OO O01 GaiG ae Oaioe eG) sOm ence Oile aGad6 a1eL 09 °ITI-88°8 9°TI-£°6 Com Ome eC GO cri ceOL, mace 9) 260 aoc eOlLee Gee OL oy (Se A OE 7A COE SRS d BL ey COME i= a6 SEE Sia6 9$°0 9T°O+6E°OT ZI TOT Z‘0T 9°OT CirG Om Ore C Oba e/a Ol See6) cles Al S6eELa ce e6 ETE, {0} Cre Owe OL 0G sob GicC8E=E 92 @ 88-2 °29 TO %S6 b°22-0°IT v°PI-2°8 levoeoncs TE 69=1 “29 $°68-9 °08 Y Lyat, 62 9°6S-0 "67 g°LE-€°T? v°TE-T °LZ 8°PI-2°9 PxeG> cy sic L°6E-0 °8Z 'Z°S8-9 °89 2°te SE 0°22 2°% Shy4 Wags 9°€ €&°% Z£°S T° ZO “95 Cedimec sc 6°ST °P 6: 75k cls €°2 | at A Seta 4 8°8 8°S Oe Oe & Bote) A ds 9° +6°ST O°T}€°OT 6°T+72°8S 0°T+2°S9 7°T+0°S8 9°T+0°PE S-L+e cS 0°2 +1 °82 9°0 +7 °62 £°0+0°IT 8°0 +2°6T 6°0 +9 ‘°ZE S°T +2°Z2 as + X OT OT OT OT OT OT OT OT OT OT u L°bT $°8 L°6S 8°89 0°88 Z°LE O-Ss y°92 9°62 8°PT y°es 9°2E z°S8 ST DY DY DV BY OY oS S°2t 8 °cr 978 GSS E79 Too) 0-99 VCoa 20 Z°EE E°VE o°CS* Z ES €°Te 0°62 T°Z42 2°TE Scr 3-0T Vote Leot 9°0€ 1 °ZE €°8L 0°22 se T€ 0°22 0°€T eck 399 $°68 T°9€ 9°05 S°Zé pte 8°8 2 °6T L°6E T°08 0€ PPULIO OTLUIEJ ET [FIOD ST [Oy 8°2T O9°PT O°ST¥O°TT+6°ET 9°02 2°8 0°6+ O°OT*S°6x 6°8 FFT Vivs ©1689 0519-225 18S. perg OSS G79 2299. 9-59) 9°Co et 6) 9°08) O-T8*s 7.98 6.58 1°28 098 Ess G:66 “160 Pty Ole 0 1s O567 82 tS 6579 9°65 0.0Se 198 U2e FOF € Se c'9¢ 19 6 kz P°OS S°8S O°OE*S°6S 9°42 0°6Zx OCikeen eo [har Aih, O7AcsT Ui Game ed) Zap €°St 702 9°02 S°6L O°C2x2 Te O°8¢ <7e 8 Ve 9° ce © 6c Lcs OSs) ~O°22Z4°9"89" L682, S92 Le2Z £2 6T Co ple, Oc. v0 DY DY OW OW cad ff XIGNdddV eW +e uidep Tel eled ed /,0 12 ujdep [e}elTed Upeelq JeTOSATY eW 72 UWpeelq Je[OSsATY Y bust Ie[OsATY }SO1O [OWTIOLT *IJUL }JOT OF UdUIeIOT [EI qIOeIZUT OT PUTWLIOJ TeFGIO -PIJUT SsoJoR yy peelg BILTOSATR -oTeuTdsOseNn Jeploqg *JUT 0} TeUOIJOOT] eUIOBAZ ay HTey TEGO uTH1IeW TeITqIO TeIS}LT SSOUROTUL $}S910 [RWTIOePT ‘que ssoloe ypeslg yypeeiq TeseNn qybyey TeseN i PEABODY MUSEUM BULLETIN 31 176 CLel=chGe eo orUsOlee tte cel Oty cle OL cect OrCle OF Ole ¢ Ul O°Cly. (Q0Clw 6 Tl 7 cL 8 7h “Tcl INGE ujdep TeqeTed ed/7O 3 LADS OG =00¢ ~“O-ce -O0"c 9°0 +2 9 OSS 06% Oece — oa Se OS9% ~_OG¥r OSes Oa8e5 SoaZ yydep Te1eTed ° ° ° . ° . . mye ° . . ° ° ° e ° ° e . 0) ye Pyo-aely “yizs=e 10S 209 © 2°e 2 F2°9S OL €°0S O'PSs6"SS “B°ZS B°es 8°29 26S Y'Z9x 6°99 ¥°09 oper serdenty = eW 22 CAVES c0S) oe ue9=6 a7 Se TS GG GO) EO) OS PeeOlh OSS GSS) 59 -So™ ae OS ee Soee (87S. a. 99S a7 [29R ciccoh S98 YIpesigq JeTOoATY 0°SZ-P°09 6°SL-0°S9 6'P Z°€ OTL F2°LZ9 OT 0°€9%0°S9%T°L9 0°89¥0°99x $°99 8°69 6°SL P°L9 8°L9 yybuey JejoaaTy }SO1O [PWTIOLT ONPERON OG AG eC sOsliceOnTL =L-6 lt Oce OL GS-27 G°7e L°S¢ 8°22 Z2°¥c 6°92 8 22 Se Sle 6 TE “UP jJ9T 0} US IO} [ETqQIOLIJUT OT ° . e ° . ° °*T TL ° © e ° e ° o } e ° ° r) PUTUIEIOF TeiFqio BreceQutys OZc=6'2y v4 S76 TT ee 0S Of 20S Sey e°z8 OS yey “E°2S “Edy 02S G°zy 70S Cy seco meenn ° ° ° . ° . or . ° e ° ° . e ° ° . . oreToOsATe SeGeceker ziee-0 cc GG) pp pre ec Ot €°Se cae ose @°S¢ 2°97 G°6c €°C6 Zee 2°ez site onsen Jep1og *JUT O} ecole OCR Col Cbeoelh pee 8 OO te Ol 2°20, 9°22 0'6L 1°02 ¥°02 9°22 8 SceOEcs 0 Tce 8i-e2 TeUOIJOOTJLWO HAZ 3YBTeY TeHGIO ee eG : : Sao ‘ . 5 é : a d . . ‘ uTHieW Te}TqIO Talla onOla-Uedecehia nO b 6 0h 6 OF 8-8 6°88 678 64 2:8 6°6 i Me) az 6°6 [e190] PT SSOUAOTYL ° oes ° ey . . oT . ° ° ° . . . . ° r) . sjsoio Teupioey Tee Ue OMemeeOcerec (Or sce) Ort 40702 OT Zeel a'0c Ost Tre Toe O:1Z S12 icc 6°92 0°0c) eee, Sr Oe=t 02) “9 87=0' T2406 CeaLO tg Se OF OFT, O-9¢ Site, Lore Se —“7Se 0eS7-5 SZ 8 -2e = ONE Yipesiq TeseN O29SSO sarees OcoecrenO Ol mc 0°SS5> 8 6S) 9) 65h LSS’ Siees 74Hfey TeSeN oe Sie MOze" 2 oO) SVAII 4g ha nOuGun eto TO %S6 r-(@) IW AGS eesti sory 9, < U==6d V dv dV dV dvV af1O) dW °estad *eStdd Sdv eyueio Sojkporbou Weg AL XIGNIddV TERTIARY PONGIDAE OF EAST AFRICA 177 APPENDIX 2a Dryopithecus (Proconsul) africanus maxillae Site Kathwanga Spec NoNMK: 220, 547, 569, 570, 593, 686, 34,CMH 688, I, CMH 138 670 1403 1404 1988 313 101 315 112 1499 jc MDiL - Tas - 7.5 - - - 6.9 8.7 LaLi B - 5.7 - 5.7 - - ~ 5.5 6.8 - ee Max L - - - - 8.0 8.0 *9.0 - - - Trans B - - - - 7/50) (3583 *7.3 - ~ - MDi L - - - - - - *4.6 - - - P? MaxL - ~ - - - - *5.0 ~ - ~ BLi B - - - - - - *8.6 - - - pt MDiL - - - - - - *4,4 - - 5.2 BLi B - - - - - - *8.0 - - 9.1 mi MDiL - - 7.8 - - - *7,4 - - 8.0 BLi B - - 9.3 - - - *8.5 - - 9.8 vw? MDIL 8. 8 - - - - - - - - 8.6 BLiB 10.5 - - ~ - ~ - - ~ 10.4 3 MDiL - - “ ~ - - - - ~ - MBB - - - - - - - - - - Site Kiahera Hiwegi R100-112 Spec NoNMK: 218, is; 45077) 557 ¥ 74s, 1948, 216, 531, 550, 103 342 94 1040 156 50 48 286° 9 <77:7, EB MDiT 8.1 Ta7, - - - 70 7.6 - - TS TeliB 6.1 6.0 = 2 = 6:3 Roi 2S = ci MaxL - 9.5 8.8 9.8 - 10.4 - - - TransB ~ 7.6 6.3 Sas - SoZ - - - MDi L - - - - - 5.4 - ~ - PP? MaxL - ~ - - - 6.1 ~ - - BLi B - - - - - 10.8 - - - pt MDiL - - - - - 5.5 - - - BLi B - - - - - loyal - - - mi MDiL - - - - ~ 8.2 - Tok 75 BLi B - - - - - 9.6 - 8.7 8.9 m2 MDiL - - - - - 9.2 - 7.6 - BLi B - - - - - 11.0 - 9.5 - Sy NAD) Ie - - - - 8.3 Sar - - - Eli B = = = ins 11.0 = = = ee Site Koru Songhor Spec NoBM(NH:M14 M14 M14 NMK: 95, 191, 409; 8'62° 245) “201 SOT, 081 084 085 CMH24 2 385 62 45 112 ye MDiL - - - - - - - - - - ais - - - - - - - - - ct Max L a Ge - - - - = “GA = Trans B - S)5 al - - - - = = 7.5 = , MDIL - 6.0 - - - - - - - ~ P™ Max L - 7.5 = - - ~ - ~ - - BLi B - 9.7) = - - - - ~ ~ - p+ MDiL - 5.8 - - - 5.4 - 5.6 - - BLi B — 9.3 — C _ 7.8 oo, 95:7. J aed 2 MDiL Te 8.0 7.9 - 9.0 - 70) y= = - BLi B 8. 10.0 9.2 - 9.9 - 8.0 - - - 2 MDiL i fs 9.2 - 9.4 - - - - - - BLi B 9. 11.5 -= 10.4 - - - - - - ms MDiL - 8.2 - - - - - - - 8.8 BLi B zs 10.6 - - - - - - - 114 178 PEABODY MUSEUM BULLETIN 31 APPENDIX 2b Dryopithecus africanus mandibles Site Kathwanga SpecNoNMK: 134, 183, 268, 690, S'Si 58, 51 465 1097 893 Sh IL7/ 1228" CMH129 1499 Max L WSS) - - - - - 8.7 1 Trans B 5.3 - - - - - 6.8 Max L - = - - - - 8.5 3 Trans B - - - - - - 5.8 P MDi L - - 710 - - - 6.0 4 BLIB - - 6.5 = = - 6.4 MDiL - - - 9.0 ce) al 7.7 8.9 Max B - = - 7.8 tins) 7.5 7.7 BX Tri B = = = 7.8 7.8 = 7.4 Tal B = 2 = 7jae 7.9 = 77 MDi L - - - - - 9.5 9.9 M Max B - - - - - *8.6 8.4 Tri B - ~ - - - *8.6 8.3 Tal B rs = = = = 7.9 8.4 MDiL - 10.8 - - - - - Max B - 8.8 - - = - - ®Tri B ~ 8.8 - - - - - Tal B - 8.3 - = = - - P—Ms; L = = = = = = ri fe 2 zZ 2 ~ = = CiB - - - - - - - 1, B i aes — = = = = Ms B — C— = = = = as D my - - -_ - = = Py uh — = — = —_ - - D - - - - - 19.6 - Mary z E = = z *15.0 - Site Kiahera Hiwegi SpecNoNMK: 262, 8, CMH8 35,CMH ibs} Ib 542), 545, 573, 578), 635 102 342 583 628 1558 1778 Max L - - - S23 - - = - 2 Trans B - - - 6.5 - = - - Max L a - - - - - 8.0 - 3 TransB - - - - - - 4.8 - p, MDiL - - 6.2 - - - 5.3 - * BLi B - - 6.6 - - ~ - - MDi L - - (i 2 - - - *8.0 - Max B - - Teo - = - 7.3 - M) tri B _ 4 7.1 = as - 7.3 2 Tal B - - Thee) - - - Uae) - MDi L ~ - 9.8 - - - *10.0 - Max B ~ - Seb - - - - - 2Tri B - - 8.4 - - - - - Tal B - - 315 5) - - - - - MDi L - - Dlce7 - TONS elaG - 9.5 u, Max B = - 8.5 - 8.9 8.8 - 8.2 3Tri B - - 8.5 - 8.9 8.8 - 8.2 Tal B - = 7.9 - 8.4 8.3 - Hos: ie M,L ~ ~ 43.0 - - - = - tT 8B = = = - - - - - C,B = = = = = s - - RB - - - - - - - = M,B = = = = = - = = PD Oia 2360 - - - OS = a *10.0 Hogs - - - - 11.0 = D - Pas Tf - - - - - May - 11.6 . - - ~ - - TERTIARY PONGIDAE OF EAST AFRICA 179 APPENDIX 2b, continued Site R100-112 Koru Spec No NMK: 140, 147, 1948, 514, 63:77, 640, BM (NH) NMK 599 645 50 282 260 417 M14087 (syltsy 574 Max L 8.5 Ba2 9.5 - 8.7 - - 10.4 1 TransB 6.0 6.2 6.4 - T30 ~ - 82 Max L *9.0 - 9.3 9.6 - - - 3 =6TransB 5.0 - 6.0 7.0 - - - p MDIL - - 6.4 - - - - 4 BLiB - - 7A - ~ - - MDi.L - - Seo - - - - Max B - - OS - - - - + Tri B - - 7.5 - ~ - Tal B - - Ua - - = - MDi L - - 1053 - - - - Max B - - 9.0 - - - - 2 TriB - - 9.0 - - - - Tal B - - 8.6 - - - - MDi L - - ey - - - 11.9 Max B = - 9.0 - - = 9.4 Ms ‘tri B z = 9.0 = = = 9.4 Tal B - - 8.4 - = ~ 8.6 P—M, L ~ - 43.5 ~ - - - it B ~ - - - - 15.6 - Gc. 8B - = = = = = = = - - - - - = - M, B = = = = = a oD 25.5 - 20.0 - - - - oT LTS6 - 10.0 - = = = D = as = = = 2 = M; T = = = = = = Site Songhor SpecNoNMK: 90, 91, 94, 113, 196, 334, 380, 381, 406, 139, 143, 144, 384, D4 9 21 “F304 We 405 1 2 382) "+62 "62 "62 5 Max L 8.5 - - = 8.3 = - - = 053 - - *11.0 + TransB 6.8 - = - 7.2 - - - =) 705, (= = 7.5 Max L - - - - - - - = = = = = = * TransB - ss = = = = = = es Be = E MDi L - 6.0 - - - _ = = = = a §6BLiB 5.0 = = = = = = x ie MDiL - - 8.0 - - - 8.2 - 8.8 - Sa5ueseas - Max B - - 6.0 - - - 6.5 - oo - 65,05 Feo - M, TiB ee ee Bilis 6 de ee ed Tal B - - - - - - 6.4 - Thais} ~ - Tak - MDiL = = Cee | 2 = he = = a . = mM. Max B 745 = - - Sew - - - - - = “tri B - = = = zs Gl = As ps x : = Tal B - - - - 7.5 = = ps = = MDi L - - - *10.0 = 955 ts (0) ake - - - - - Max B = = = = = GaiQ Sao ee0 - - - - ~ SP intB - = = z= S 8.9 = ¥ = is Tal B = =e - = = 8.1 - = e 2 5 a ie —M., I - = = = = = = = = = = = = io .B - *15.0 - - = = = = = = = 2 = C, B EME) = 2 E 2 E x 3 = = Zs Pa B - = - = = = = = = a = = ae M, B - - - = = = = = S = a = jan ay) 8) Saag) - = orf ata) = iB = = = = 7 - 10.4 - - - = Fi2K0 = a = = = ie D a eS Pat iF veO2ROMI5 0 = ee ce ee ee ~ = = 12.0. - - *14,.5 - - - - - - PEABODY MUSEUM BULLETIN 31 180 eL[[TXew Sezuedu (TnsuodoIg) SNOStFAoATG °@ XIGNdddV¥ - = = Ol Tvl 8%ct = = = = atid ow = = = Git “stl. 70505 = a = S TIGN © = = Se ae ec eee Oks ae Sue 0°rT = b°2T = arid = oI = eeclgn wecl., 20) -01 € Or ZeLE = 6°OT = TTIGW e = = = Cth 02h 9 8 08 8°6 ag = = qld iy 2 = 2 640) oe v's 0°6 8°6 . = TIGWN t = = = ORL De Onl. 20° Tr = = = = atid = = = Bio ne: .9 8°S = = = = TIGN ”* = - = SOL 8ctt ~ SOL = = = = @ Fd - = = Seg" 28 O°Z = = - - TXeEN od = - = 1)? Lies ~ = = = TIAN = = m 6°0L O°2T = = - - - q suely 5 o = = OnVis 265 SE e = = = = T xe vl v°L £°9 8°9 = = = = = 8°9 qrier 8 °8 v°6 9 *6x v°8 = = = = = 9°6 LIGN 2 Lv9I TN SITHNO'?P ZvSZ'v69 ZPeT‘99S SHINO’S (HN)IWG ZIZ‘SST €66'99T O0S9'22Z TZTHINO'0S ZEHINO'SZ 696T'P6S ‘XINN ONOSdS ZTT-O0TU phomtH eIOueIy 0 i (“i;*‘“‘(SDt*;*C« OEM EY eH 141, 144, 151 R100-112 a, 9.8 Hiwegi I, 4, 6, 6, 560, 38, - 2.9 8.9 10.0 Kiahera 181, 538, 539, 108 48 487 CMH] CMH4 CMHS5 CMH136 106 CMH10CMHI35 60 GB & 12.0052 Dryopithecus (Proconsul) nyanzae mandibles 39, 374, 2, €20, APPENDIX 2d aA, Kathwanga 23, 143 1229 CMH7 CMH30 CMH31 CMH110 @2 1145 1 7, Spec NoNMK:125, 186, Site TERTIARY PONGIDAE OF EAST AFRICA 181 ft SEeeGeronGe Tg it 0 1 ateripeat 1 Sete ees tab tot Dm OO © wo tO (Sy ce De = Oe UP eh TALI te mel Ve ar ree A Nae et Tad Jas Lae! ost onnnm U GSS Gea cee UU GL Une vite 1 Cee Tera Via) te et ORO I ! to st 7 Se ee a 1 Stet) Se 1 Nadedq teat aot I Wet Hoa thet ty eh ht Mt a I aie 50 aaa | | ODM oO 1 rb Sec 2 oa vt Ny ater] I Aerie 8 in | tot 1 | (ea Dare x ! oP a 0 I i] I eet 1 OT Go) 1 1 2) S| SS pCO RGR One) COIS teste See : oho mn to ot ONYN CHMHM OTTS © me ee Soa eth GO elec (epee) She) She LD) SCO) ST CONLO CO NES KON: ort nan no nwWnWon°o NE WO e .- @ . . . ° eo 8 ° . >. @ eo @ . ° o Uf . . e ° Ld a ¥ ABIB 84 B aR Pal ese tat Bieta a ae tet A a} 1; iyettign iat tot M71O Or} aire lire) axe 1 Beas SP oll satie tt hey Saag I toate tae 1 1 ot oOo oMm~ ao [ ee} . @ a TS Servis SS ey an 1} eater ee se at. GG m4 a | fo) foe) = ae , e . e otros te tues acy eer tial = vt 1 Veatenteet cL t “a So MOO WwO Ot Oo EE N a a (=) (>) oO So e 7. e (oe eee) . pad e ° e (Jee) e e . . e e . SRR mone egg ant BAY SN 82 fas te PIR a ee LONE we 1 Pees te it onntnNn t heel Sil of] - pe it a Te ee ile ay to ou outed 1. pr arias Atom. taect ! PU alee tt at 14 5g 2a SS) Ss) Si eT MONS a tte cee a Thad 0 Oa as S3°°°38 onan oO oO acs e os Peco Lom ge mig Bay a) a eG ! Wet Oh 2 tot tot nw Ot © [Poti Je eG She a are iC Hata i es 0 to Jet om oO Oo st teeta! ei rtm Pca JECU? 0 SsHiteheyey Ui 1 (ep ine = =) R faa) faa) Hm c 1 Bee aK OO eM pa] E ea Se easeken She ep Pe) Pet ey SS ee ES Sanaam Aun Ae rt mR fv] | Pee iy] o r) yy ees P= a a PpOas ef & PEABODY MUSEUM BULLETIN 31 182 e = = é; = - = - = - T‘OLs - f°0v c"0l d Tel I “a = = = = - - - - - 0% ~ L293. 25-9 TIan T A e _ = - 2 2 = - - - - eae AKA ald oy "i 2 a = = = - - - - - - L°eL 8°20 TTGW & 9°€T €°9I €°ST - €°ZT - - - - - - - LAL (oe Ae atid iy 0°21 Vel Zeer Z°St — 1°St - - - ~ ~ - - TCLs Tels TTICN e - - - - - O°VT 6°ETx - - - - ~ 9°21 £22 atid - - - - - OZ =F Ulse ~ - - ~ - 9°IT* L°ITx TIGN ¢t = es S = = = = = - - - - L°vl 8°PT arid ss = ee 2 oe - - - - - - - 0°98) e672 ITGN ”* = s: = 2 < = “ = - - - - 9°FI - d T1q = a ze a « = = ~ - - - - LOT 900 TXEN 4d 2 3 a = s es = = = - - - 2°93 048 T IGN - - - ~ = = = T°Ot O°9t Z2°Sts. = Tor = acl ey eee - - - - = - - 8°ST 9O°6T L°8Te - 9°LT - 9°8Tx JT xe t = = 2 s = = = = = - - - - 6 ISIS = S we = = ss as a = = - - - 8g°OL T TaN PeHNO I8€ SeHNO Y T TOT ‘sop =‘ ZOL *XINN 60-29. 80-29 1-99 40-29 S0-29 0-29 ¢€0-z29 £€0-99 Z1-29 11-29 ?dINN ON 9edg = =" == sybuog = aa eT Cae = IcGN — =. OJOIO/ alts ee[[Fxeu JOPeW ({NsuodoIg) SnoeujfaoAIq 82 XIGNdddV 183 TERTIARY PONGIDAE OF EAST AFRICA - - - o = §°S2 O°F2 0°02» = - - ~ a - $°2 BA en - - = = - c°’rse = 0 °ZEx oa - - - = o- - qd - - ~ - =- 68S, = - > - = - - = $°09%% aq "Ww - - - - ~ O°ISe = 0°8hx - - - - ~ - S°PSes € wt - - - -~ - 0°97 - 0 °2Px - - - - = - G°67e% d o) - - = - - 2°12 - 0°02 - ~ - - - ~ 0°SZx% @ I - 7 - - - 0°S9xx 9°S9xx O°FSe - ~ - - a ys O'roex TW “eZ - - ~ - - Of2r 2 0°2ct = - - - - - - P°OTs d 1eL - - - - - Z°vL Pet SOT - - - - - ~ S°2ZI GHL ey - - ~ ~ ~ CPL P°St S"OTs - - - - - - $°ZT q xeyy - - - - - €°ZE O°St S°ETs ~ - - ~ - - 0°ST IT IGN - - - ~ Cro §S°zr “oc7ts. 2 °OT - - ~ - - - 6°ITs qd TeL - “ ~ ~ TPlo oT cls o-oo OT - - ~ ~ - ~ 2°21 GL ey - ~ - - T-7l TOS \/8°Zts 98°0T - - - ~ - - 2°ZIx @ Xe - - - ~ O°St. 6250 .:8 7h. 9°2Ts - - - - - - O°SIx TIAW - ~ - - - "Ore reL Ole aeSic6 ~ v°6 6°OLT SOT 26 == O°ITs q TL - - ~ - = 60b sscs0t . 2536 - v'6es 9°OT Z°OT 06 - 8°OTs GML ty - “ ~ - ~ 6:00 2500 S°6 - v°6s 6°OIT S‘OT 7°60 O°Ilxs a xeW - - ~ - - Sil [°2ZE 6°OTs - EO eLe2t 02k 8°0IT - 8°Il+e ‘LIAN - ~ ~ 9°OL - LG 26 0°6 0°6 0°6 - - T‘6 O°Oles Td 4, - - - 0°6 - 0°6- “26 e°Z 6°2. (6°Z - - 9°8 €°0Ix S°6 xe TIGN 8°6 0°6 O% x = - Z7°S xx 00°C xe Z°L - - - = - ~ S°6 + G suet, ¢, v'vET O°9T O°9Tx = - O°Slxx O°PLxx L°2T ~ - - - - = O°Oles TxXeW - - ~ - ~ $°ZT - G°6 # - - = - - - O°ZIxx @ SueIy To - - ~ - - T°ST - G ‘els - - ~ = - - O°9Txe TxeW SP99IN 980PFIIN 10-99 69’709 ese E'78E 99'7ZL4Z SS'LZZ T'O6T (HN)NG (HN)Wa 20-99 91-29 ST-29 I-79 €T-29 90-29 01-29 ‘dN ON dedg AGUbUOGuE san. 7a nioy eS — eden ko soTqrpuew Jofeu ({nsuosoig) SNSSeyfaoAICE 3@ XIGNddd¥ OJOIOVL ©+S 184 Statistics of Dryopithecus (Proconsul), APPENDIX 3.. Gorilla gorilla and Pan troglodytes Ds (Po) africanus OR = _D. (P.) nyanzae xX + SE sD V cor n SE sD Vege x Wf 6-19, 3 15.7-16.1 4 19.5 4 15.9 14,0-15.8 Z 14.9 2 1.5 1262 80-1156 1,14 Max L 9 9,440.38 Trans B 9 7,.5+0. 32 10..9-12.0 OS 96) 1352) Ged OL Gr PEABODY MUSEUM BULLETIN 31 oe ened e se ora 6 Nt WOM ar coir tt | i | 11 nt Cx (to) Cyto) s e e e SAN Nico eiri rir HOD SEC, Ne kas 1h et aaa No CoO RIN OM e. s « Ld . e. es e Ot Ot Ho mW cir ei cir air Stiri wir Om ow NO 00 OO tO ee e es e e es ze or ors Oa AN a! [on | ei ri nr Va | ha | tee | ean (=>) [¥ey feo) (Sh, Sayieo} (a) Ko} . s e. . . e sz. @ | a) wre oo} Kao} (>) C1 ei ei ci cir eae Ag Po fae ee es et icUa ree ean iat) an Oo in © N st ww e. s. C Wale) es es e.0UmUCU Oe Nr or or rim [oe | ei ri coir mm mon Yio mw nm co or Oo © © 10 . e e : e e - ¢ ) en) wo oO an Oo or et rei [os | air 1 yl Wey il Fmt | (>) {te} = ©€ (Sys) ow Wo yN No No or N Tike Ractive ec pepe orm (=) ri ot on i eS (=) {=} oo Dm oO riw air On ao oo +1+l +) +1 N MN mom Oo aon m WwW e °: . ° ° . e ° oan mon lps op) ao © rei oo oO Wo NN co co ari | pa H ~ 0 ain an Ay i | or ed od sa aos ea a am a O a re o J Dy Ay = = 12.8 14,4 1 uh 10.0-11.5 12.8-14.1 3 13.5 10.6-11.5 4 8.4 4 ial MDi L BLi B M3 on OD MoO RO OF O e e C te. 7 e @ e e s-.hUs e On Oc GOO Nei 10) (co ar rei ari ar nr ri | oe | Lent Lik tot 1st eae own wn oO a | wo oO wD ° e e e ° se s e J J co On NM ~~ DD Lan i ©) N Oo lop) ei ci ri ari ri mo to 1 ey Wt Smee ih ot) C nwo oOo-n-mM on NO oo nwhm ua wa so les a ate (Ss) for (Sj (S) | es ee ol NA NN es e es e oo (ea (=>) +I14l +14 Oo sto ale eS ions on J . s s e se s s se s oe; "2 @ tr tO OD FO HN ON cir ei cir ei ri ar mMnmMn OMO DD DOOM NOW Hs! on ore ott? OR OW W«? s e e es e se e. e es s os es NO NKR ODO OD MOF AHA coir ri ei ari ane Loa Lt 1 1 ee | ; a | 1! ot On m1nmOM wn SHO FO wo st e e e s e e se e e e ss. 6 @ OM OO ON OR OD FD ci te | [oo | ei fs 29 way C2) See myst oo bm 23 85 3 tt be 8.0 SS eee (e\{ey fejis) ©) YN rir Oo oH MO oO e se e e se o.oo oo 16 +i+l +141 +l AD OoMnonwame nM DOM OO se . se s e«- e es e J es s es ND OCHO RNR OO FO Mr a ri Po oe es | ry NO OO Net DA ON cir on oOo © ort of MOC Oo) : J e e se e > « C em a) s . rio On NR OR OD FD [| ei ri et tit ee au 1 st 1 ot Ve Ri a) oo MO | a ow Toma) e s e eo . es e es. J s . . ~wmWonrt? mn Reo OnmR OR Re nO DD bs a) Le Grant | Fre | ee e I a) vie Con OP) in c;O hry oo er ei om oO co wo Oo om oO ce — Ww ~m co oo 5 lie iy Lest a a ee 55 to i) (ails jt >) (Se) (=) ma st wsTMée 0 Bo NN rir N NN oo coo so oo +1 +1 +l+I +I +1 +1 om~ om N oO On w st n~ WwW s es es s e e o-.hUmt se e . s Dm oO om 0 on~ De® oS om oO [Yori¥e)) (oy}Yey Ta} Gal weep Wey ta) le) air riri cir faa] Q Hn Oa ae Ha WM wim Rie} RA el Se a ad od GO OD Ax aa as rae Se Sem sees ee eae a R o Cy P Py M M M 185 TERTIARY PONGIDAE OF EAST AFRICA es won DD So nn ant wae rir J e e [o> as | no Of aN on on ND OD NX ANN Ke 1 e 1 OM ODON ND MO NO WOO WD AW a « se | ei ° ! ° Ce | ei =< are nar ant ri 1 rei 1 ot non os ° e foe} ei i] 7 e SOD OD OD AGH DO 0M CHO ON WN FADO OD OR MD YAN YAN FWD ODA NO OM W-«! e wD ot I ei e e NY ON MO MONMR NH AHO DODhR OM OM HM MF Fo ot ON Hm MHOt OM ON Frnt Youonwmiimnmnnmn MoO oOWM be 8 > as SoyApolbor *g TL°O+ OT ‘OT ZT O+ SZ °OT ZT °0+ LS °OT ZL °O+ 20 °IT OT°0+22°6 TI °O+ €9°OT 60°04 26°8 OT°0+S9°Z €T°0+82L°L OT°O+OT°IT €2°0+78°6 62°O + 81 °2T 8T°0 + €6°0T €l°O+2L°6 ZT°O +89 °IT ZT °0 + 20°0OL TT°O> 8T°IT OT°0+28°6 £0 °0 5 9T °OT 90°04T0°Z TL°O +09 °OT 0T°0+00°8 62°0 +00 °OT €6°O F2L°CT S207 coi ¢ 6L-¢ 7E 890-8 =S= BOUND DEC 1972 Ne ewe TH MeN REAR ir N ARS a een sie eee Ree Ce ee a nee eae te te age VEE Re TSR BS Seine am SERA te (etn, OF Sa Ne firs OD diy, 9 ed erates oo tS ee os ; NW uiesemeeeserdlie afr gees ace aera PE RR AE Mn wt NRE peewee a aaner aes EATER ETS “s - : - oP “a = “ ; . “= i : ee - a ae, - . . . : i i wets. - - _ he Wo pe ewe Pa Pv eny Preto! os 8 pe a4 ara ee wi og rue ma de tere. pine en args! Ogre epee wtp estan mora ata nase ta dey ay ee Mth etme FED ppetnniet bese uk sys ed reece Peers