fipitrisry perssa canes esses oreasaiyisg his recettesert ess Siissotaten tsa pr - secon pasha pean nasaraannesaoesesiepas JEAN Se tren terete sre rms ete tenet BE wird se eg we Fs Lowe ie beatae 19 onl ream g aatmy oe ADMD, eee * c Nay, a Z Fesen Fr 5 ee he > i 4 y ae oS I uy oy A ‘, if ~ aj X: ~ ea % rei y f s t ‘ S ‘ ¥ ’ § vi res F ce She . 3 z vl s Fi . . ss / _ v re \ c } ‘ g ~ Mee I i ¢ s : . 4 4) fe ; * ; * § Vy, " ) Ang 3) F oy 3 \ 7 % ‘ 4 , a; j ; Pe . : i] * ’ . . ‘a ny : sie * od ‘ ira i 7 r . L% ae ; / 1 eee ea KU 7 ~ ly , i ‘ iar oi © OT ee eee a >) * ’ ’ < ai? . . a ° - ’ . 4 i ‘eq ‘dingsiiiq ‘wnesni) eisewied ul jeUlsC jUusiuy “y sejseyg Aq uoljoesip sueyoieH wp depun spew suijuied wo14 "JONVLSIG SHL NI NOGOHOVYL ‘SdOLVYSOINL 4O NOILVHYOLSSY ) "Id XI1X HdVYSONOW AAAYNS TVOIDSO104S *S “Nn DEPARTMENT OF THE INTERIOR MONOGRAPHS OF THE UNITED STATES GEOLOGICAL SURVEY AOE BOM dye -Gi ba to¢ WASHINGTON GOVERNMENT PRINTING OFFICE L907 UNITED STATES GEOLOGICAL SURVEY CHARLES D. WALCOTT, Direcror Mise CMe ATORSLA BY JOUSUN, Je) JsCAsXOle0 Mix BASED ON PRELIMINARY STUDIES BY OME EUN Whe CS UVEACESS Tali MDITHD AND COMPLETHD BY TRILCISUA TPAD) Se AGA AL. WASHINGTON GOVERNMENT PRINTING OFFICE 19:07 a ¥ Page Page Page Page Page Page Page Page IE tai 39. 57. 65. Page 115. Page 170 Note: Concerning statements on pages 139 and 140 relative to the condition of skull and jaw of specimen No. 4928, it should be said that no record has yet been found of the receipt at the They probably became separated from the other material in packing and in their fragmental condition can not now be accurately identified. National Museum of the missing parts. ERRATA. [Monograph XLIX of the United States Geological Survey. ] Eighth line of text from bottom, for ‘*1871 E” read ‘‘ 4739.” Third line of text from bottom, after the word ‘*Type” insert ‘* No. 5457, U. National Museum.” To list of genera and species add ‘‘34. 7. su/catus Marsh. Am. Jour. Sci., vol. + May, 1890, p. 422. Type No. 4276, U. S. National Museum.” In legend of fig. 30, for ‘‘ 2116” read ‘* 1201.” In legend of fig. 32, for ‘* 2065” read *‘ 2416.” In legend of fig. 34, for ‘2065 ” read ‘* 2416.” Fig. 62, left pubis is not that of type. Nos. 1, 2, and 3 of fig. 74 are of specimen No. 5793, U. S. National Museum. Eleventh line from top, for ‘‘ 1871 E”- read ‘* 4739.” . Tenth line from top, for ‘‘1871 KE” read ‘' 4739.” CONTENTS. Page IBID IDI ORDD,, Toy? JEL, ID Ops ovals se caesar eS Rep TI re, eee ey ge XII Worx or JoHN Bett Hatcuer, by H. F. Ospon Ne Ate oe Se AAA E ie ee I CM tetas Bene Foon nak XVI IDO S WRIA CIN 2 Meda ceuce eee Steers Ges Sete gee ate Cts eta aia ee tet i oe ee O.S'00 F ANGRTEIOII/S, TIRIDIMAGIN Wes ae dee sae Re ee SG 8 SRC CNA IE ea era eed elo dn a XXIX Parr I. History or Discovery, CLASSIFICATION, OSTEOLOGY, AND SYSTEMATIC DESCRIPTIONS, BY J. B. HatcHer. CHaprEeR 1. History of the discovery of remains of the Ceratopsia._.........._...-..----------..-------------- 3 Hires RGSc OWelaves lige rom Bra Vpembl on ens cers tryst. ot ey ence ee ei Sl ee nd aioe gece 3 DiscovienyaouLypeolAgathaumasisylvestmis oy; bro. HB Meekis==)s- 6220-9 2s ee ee 4 @eraropsiamemaimsidiseoveredsiny Colorado byalbroty Hi. Ds \Cope sess 4 2eee so 422-28) gee ee 2 eee eee 5 Wisconenlesuo vera GaeMem Da wsombam) Canadas Moe ees See eS eee ee So ee oe 5 Bxplorations mm Judihyviner beds byalbrotesson! Copesses-4--52=- 222 seen. 2 oe eet aes] eee eee 5 Remains of Ceratopsia discovered by G. L. Cannon, G. H. Eldridge, and Whitman Cross.........-..-.----- 6 Discovery of well-preserved remains of Ceratopsia in Converse County, Wyo...-...---.------------------- a Discovery of Ceratopsia remains in Canada by L. M. Lambe...-...........---....! SUmanaie er ete MeN ae 2 9 Ceratopsia remains discovered in the Laramie of Montana by Barnum Brown and R. sik Lil aneaauhamec seg a 9 CHarrrer II. Classification and representation of the Ceratopsia......................-------------+-------:-- 10 Claceiieniiom GP Yoe Comino seins see ceed saree ote Stee es meee ee See ee rene dea ae 10 AN poabencalslistsoimeenerayand ispecleskereea ery yrs tee Ween ninee Ay RMU cee Eee oS au anes 11 Supposed European representativesiom the! Ceratopsia-.-.55-2-22422 222-2. 20.5- sess 8-2 oss. 2ee cose eee ee 12 CHAPRER Mle Osteo locayponthes CeratopsiassMmeee ea 4y- ee maine @ eines ee eens Sek he ei eke 14 TMO® SG. oc ce. 505 SOO wo ets A SEL Rene cs ea ete 2 pe ed NE ep Sa 14 AN ies CURVE NUTT AS scsehc: oae t Sleii ERe ces is ey Shes ea Se TRE ET ete ERI mo Cae 15 Mnerocerpitalysccmentias ss ee es Par na aL aN ay Vad te ese MIE AUB LEN AVON co Rest on 15 Niersphenoidalesecomwenitmey ie ee terme mene sam eau) OEM ne Mee) aes a oh AS aaae pase! eo Lay 17 Wne® incl OP PEM Onesie eso a a eR ee Lie ee ae Sg eee 19 Wine Ciueclre ion WEP ANOS Ae es oe Cee eek ar ae ais ca ee ast eps a es a 22 sites inoncaleeero Maemere treed oem Peter page hag sens te een WC ya FL ML Vad Gere nt eh 24 semnasall Sommers ere eee rape mernoe eh PeiCs ee, ween eg LS 8 A LaF e HS ett Ca te ea Pan ann ANN UR on tray A 25 ‘Qhlove samenpeiMMNT ASSL oe ee AE i ia i Ae SAI EN ce ey ote a a 25 IN AYG). (NTSC? OMNES 5 Se ie hc aoe Gs HOLA Ble Oe St ese RS tee Sires el eat ae ee 26 WN@ (SHA YON Bena saaSabia qes'Sae aoe esS ee i ot ee ce na ac Lee, an a eo ee 26 ATlavs TORMENT: Occ G Get eg Oe ees SG ons RS ae ee 27 IH R@ SHON < se BaSbos ele dees Hae Sas NS Oe Peet ae ee che es an ee ate et ee ae 28 INOS joraaiee cbse sods Ses Sau Shoe MRSS Re Be EE Seo Sey e I enn Re pee a eS 30 Nae Genanall BapTaVAPTOLs 205 se SS pete se CHET Be ae ey EE 31 IH eIstipracHo val enonmnCOLesHe pe rere amnesia se Tame Meo -utbe aliens Maen Wee eee Lae 32 ANOS RAGE IOYOVAM CORE coe Goelc aos Se, Mae Mneicl ab eT CL CS Ee EC ES an ne IN, ha 32 ERITGETOR Uae OM Capra pap pare te AD Atm Ce gM eo. AL IL MIE SEE RS eee i kee Lo SR 33 AN EXG) Gh avy DEPN ISS Screen pe SS cele eS a as Se ce eR 34 WINS COCCO OMIM eon mete SoS Hole cle meV Sere Ae eee err ei nS ee SIE in 34 Etenmalno peninosmmat nex Cran iiiiminese rt apaen ws een i Meee oh epee sn yee 2 2 Sao ie = teie ertheless aie 34 Phere up tarenaiporaletossse Wasane a mmenee Uy Ne tenes ures Sere Sn SE i Scien Ae elit ats helr cis bile «ciate cate 34 iiespastinontalponain Crime emer mer tereer i Seep teetnee Beyer ulty Nena 5 <5 mete eins Shiajai he's Geese. Scene 35 pueaAberaleconmOralmOsscen a amen anes Saleen ys a Ree ee ake an ayie accitine eleieieinn = 35 INN QM oncac clin nerd PaO Cle ee BET om Cte ee ee UNE A Bec ev ea ge ge 35 INKS) ENaC Oe EME Bis tone ule cD © SOI Se EEE ret re Pe Se ea Sr 36 Woe PORIGHOPMECE Cpoebinva Ns Neo. oon cies cs SCE See ade Does Nemes ao sOC ce ce eee aeeen ness 36 Res CSHeHLOLap Ala UUNEmy AGUILUL CS am eset nee tts ee evar SNE es ep la eo ees eNove.w Sy eiavelsls 2 5 ion/2 ajatelatwtene 36 PUISE ALE ELO hay Alain CM Ona U Ln eee Merman Arenas eee Raye exes ele] eis Siecle yefsislele lc = ine del= =a 36 ae loneninenaval lost pliar openly zecd ecteee sila be less coco ce tee oot ene ar ERA eens ann era 36 Mi SORES @YsbiN. UH b in. co odes daAL UO O6G TOO ROG RE Se GOR CO Suen IME is aaa eae aaa eee 38 VI CONTENTS. CuHapter III. Osteology of the Ceratopsia—Continued. The skull—Continued. The lower Jaws: 2225 Soe Bs ee eo a le a The predembaisy. 2. 12.205 Nhe Ws fn Se Suet es a er ge Oe a The demtary: Segoe loci Se Se ee a Te a on The splentals: sates ol eggs TS Sa eee ep Prete pee ee Dive anti@ullar ec 25 sc SS a ars Tos Rt Sars eI gta a ath Ug Phe ‘teeth. : atsaeiace a See Sea ghey eee es Poe allen Seni Me eg 2 Vee Aa ee ante Arrangement ofjteethuin) tines \aiyoeee ees eee eee ee ee SoMa Nac 20) oth ee Replacement: of thie teethi loi 250, No Nae EE cr sake A ge eyes ek oe ye an Functions of thesteethiew ssa oS tot ip a i SEL ST tae ae en ke The vertebral :coltimatay irs. 3150S ae Ec a ee are Du D a N See E The vertebral. formally. 2.5 See eee 3 a ie UIE OR eS COE AP ne alae The anterlorCervical ss woe eae ta ee RU A MO a The-fiith, sixth and Seventy ceryiicals: 2. ea ue ae a) yt ene ease ee The dorsal, vertebrae cote oo or a) aia Fl ae OE aro Rey ec) The first! dorsal 3 5:52) a BO RE Rs A ee Rt) SN SPA a The second dorsal ct) ie ee ee es RTE ae ee iano ae ea The thind dorsal cic Si ah sone te ee pe ena Oe eee ae Ce te IDroriseils) 4b i) Wa eos ee cee oe SMEAR a tha 1s cae AROS Leer 2 ae cS ee re aa eae ° Whew pelvis 256 siokile oo eye cre sews Sree es MS SS Se ee ee A Wave psl bob sate ee eee ieee lea name ae Cece Ryn mali: I ne et ODE eke ao aes ooo cS eine ns Mine pullis\. 2 eared SA Sis Sis a tency cy Seen eit sR SS a ee ee Phie ischial lee ee eae eae ea oat ge eee ee Mhaishould ereurd leks yes ee yh ee eee ee so fc ts IEE 55 SIN» SIT GS ge The scapula. -....-- aya ea ee See ee SI oe se PVG: COPOCOUG: 4 (i Rs ef asie, cin ee See eae Seis, So Re Ie IE RIN Se A a Phe wibsie O20 Pes ee Ce ns Be nl oe gE ee nae The’ cervical ribs 2.5 pos ae es ae ae Pe, Gy oe ed ot eee ny Phe dorsal ribs. 36 ey See ean Ie ey Say nears SS 8 a Thecaudalliribs ss ge oe Lee yh I US A Ct a eee Sg a en ee 5. Phe fore limi: and Toobas aaa so ee NG Ue Oe Fe ATE RB 0, Paap UAL rag OS ARH GHC ETUS seen eens Sep ast ea ee ee eh ee te Se RIES eis aS aac Sin Se mcieictal = The radius (2 oo. ole oT cee eee ee 1 Nene) aU aretha ee a PIG) Capus: x6. 225..3 ss isihe Lp alee See ay Ie eg PS Be poe ge eg Oe Tien etacanpuss iso stele (Pee eed 7 ee SE a ee ee Mhesphalang esis. Safle ses ca SE Se 8 SES yep ep =, a ee eee ET Tine invael Inman eaavel WOW 2s 55 os 2 2cessecessssescceceee- De Mee Hie LSE oh eee oe 5 FING RT ETA ULT = ase sis2 oSe CN aee oa8 He Seae wie 3 scien eo TUES oe Pi ee, rm Oe ae Thesmletatarsus tc cc SS Ra ee gees oP) bial Teepe a 1eL a era g epoR g The phalanges....--..-- SP Aparna eo a aee “Sha ate a clvetk ete, SI ais DS We oes a a The ‘exoskeleton. /s. sujet eee wets ae peiore Rea eve ee Po Aaa E ee ie anaes 8 Se SSO CHArrun IV. Systematieldescriptions of cenera and) speciess = a2 990 ee se Genera and species described by Professor Cope... .....----.------------ ics Schl a teed AN 9 a eae DSS AMIS wfoce, iscyissrate ns hse Nera seco arm rae ote ory yt a eee DENG mCaUStus ss Se le ee rc tates Soe eer en LR Ne alee TD) Snasy clematis rs ces eee a) ites eee a RAT os CERNE SIH as es oid ts Ds bicarimatuss — sees Mee yee hs tile a Ley ss Fe ree nr ID eno eee eee Re ee Reo aie he Seb eRe nooner mooqan abetoocceractobdcodeasidsachoss “ Monoclomius edo Se Fo ea eee Ne 5 Vea ea ae SN eh Pane ES nn ae eo Original description. ..-....----. sess oe SVE oS See SE ass eres ae eee IOIRRA EN ANN Genel eA aaAtaO reals Gc se ee CEM Roh eSAbs Rooke ovoedandtebndneconassdcandogsT Deas CONTENTS. VII Cuaprer IV. Systematic descriptions of genera and species—Continued. Page. Revision of species of Judith River Ceratopsia—Continued. Genera and species described by Professor Cope—Continued. Monoclonius—Continued. M. crassus—Continued. DESCrip Onno iatyperspecimn emisseere ris ae hae 5.) eee ps RRs aies, cee yeree dN Sem bel Spe SP 72 Erincipalenleasubemecnitss awn skate Mase 0m NOS Aime yamine, Pit ee MRE Eh 80 VIBBLISS US Peyeme earn rem eyet en es Sia Spee ee a Be 7 elec meh d DNS gute S Beeete nee ha Sea Rade 81 IML SRECCTGUE UCONTOUIS Jo Ss ee Es SP eh as ces PAR ces Ne pe Eo ea 81 INE S[OIVSC OCS eo THe & Swe Geis Ses Sa eye eee octal ee eae ee Oe gD ane ee ee Bie beh 87 Genera and species from the Judith River beds of Canada, described by Lambe..............-------- 88 Monocloniustda wsOnilneater wet ferry Chieti pean ah URN Sane aa aiinae al ec Suber ae 89 SME, CHIDAICIITSG § LcsecniBing elas lak AB 5 eR RCI Ree ECVE ere CI ate ea cee ge 93 AVP Oe itmeee erecta een cey ec eee sts Sew ay Te se Bel wet ee Lee Ree, Oe NM Qk ee 96 UCC OC CLAS te er ENE er income yer cy Ul aie ed Lyte ac a tenn NE HC hive a TN ns ey aN A Ch iN oe 97 So WENIGIU io Beciss Rose's S SIGE Sete Meee oe ert ae Eee ee, amo eae Re nee a Se 98 (Generarandaspectesid escnibeds yale rotessorilar's nee =e ae oe ee = eee or ene eee eee esr 100 Cerato ps apace ee ee eer ene ee cme yar ee wet atieh inh Rieitg Se SPOS get guites ee eens 100 : CAIMONUATISRE IEE ee Cn en Rs, py mummies YOER Er Seley bo.) cul ye oe eh RRA la lh is 100 Cp POMOGONS... asbsedenkot e485 e6 Gaderer eae BIS ea ei ins eaten ate ee eres eyes Sees 103 SHoeoles Saroncouslhy MEKAAEC! WO (COO cossecagdadassanee Gao enosnecon sess cea enson asoceasesese 104 Revie wgomuneispeciesioimbaramilen@eratopsiae seas -1 se cois ee aces ee eisc so caleues co eene es ees cise oie 104 Generayandispecies describediby brofessom Copes= 55.522 2-2 oe ee eee ee 104 AGENT TION io 58/5 5 BASS Bea S HOMES A MAAS SOS re eS SC ae ee See eee eee ara ee 104 Jo SSIVERTI Se cosnekasSd.ao SSBC SS SESE G8 SUSIE OO gS ree CaO Sea ot ane eee 105 Ho caliiiy ieee sass eee ee teckls racine eer cera sy Rare ear ee Pe Mae Me Me Lan oe RUS Sal dal 105 iD eScriptlongo ye COD CM eEr emer err ec Teer etre ey Retry an Aad ea hE a Se 106 IMairenvalldan oon evil lol Ot eerie yek tae lerne apie hein ea aah oa wa aie eee ua nee Peon aise eiels 108 WetailledMdescuiy toner eee yee) eit Ne Cy eet ene foe Sk oe ORR Lene as 109 Brin elpalymieasuneMmCItses sae sey: Se san eee nee elcome nae Meets ee ee 111 Ang HONGiss clas sila oAe OBS ee SEE Ot Se Set Ce Te Oa cate Ser Cee korea eg eR ea 111 JRGVODD SR 6 6:9 d.cro civ, bo Hid ORR G OE ON oe EA I pee aT eae rc en eee RR eg ei aera 111 IP, MOTUS S seca’ Seeswoce ide aaa es oe Se cy Nee reyes i aes a Me in cee Rea me ee 112 WinraosjaararchAlnis Crayne es. ese Ge oh ces ek teenie Res aS cach ye ee 113 Mamospondiylsroleasmee teehee rar mie Piss ais mise niet ee eerie she cc WOM rs Se oee ee 113 Clerord ny navelnnvGhs Sasee eo ae se acre eee aera Se a Fe eh AA pA BIE owe YO fie a Re aes 114 Cl, HRIAGChAIE so 35 die do aleuisde aap eG mee mer eet ct Rs ett ene crenata a meet Rees eis re Ay wee 114 Genera and species described hy Marsh and Hatcher, chiefly from Converse County, Wyo-.-..--------- 115 Corstopsp Mears hineer et epi rereh oe eee er term ree we te neereh Sn Uae Ne a Ne at Si NEN ce 115 COMATOSE QUREOMMI Ss ooo odr oss oaG TARAS SURES SOS eee SIAR ener rae er ee aaa ea 115 IACOICIE 2.a43 oon ago Searoeke Goda DEE Sco Gra BAUE A BOI saan ere Ee esti eee aren ae eae 116 I, (CEOS INOIECUE) Is Sig bce ode eo ta dee Sante dao tare aoe hn Seem Asn ertne are 117 IDVECON Ga Scio 4 adla Bane a5 6 DRO oe ON CES Coie Seen Mer Ge Herein Ua acne ea Bers a aa a 117 Oncinalidescnip none ane sme er ere tee hte ee ye ioe Bete a) tye nS 2) Neg ote Shae 118 Ceratops horridus made the type of the new genus Triceratops. ......--..-------------- 118 ID SSM OO OH AADCSSceb ac So chase OSce a oa eee N Clo 6 eae a Sie sneer ers moe Peete ie Mats ca ete 119 Ae COTEINOS 3.5 cr oA GSS BD SOSA eI EPAS TSE STOR a EI et ea 122 lLe@gallline gHacl INOMVACil 559455 55.58 OSS BCS AEE O ee eee see oe ee ns Sree eee se cereai e Sie 122 Oniginaledeseninnt omer tee eye alee eee Ore ienc os Met ererc seers = ist sieiasebelit e ahetWare ovale Sictels Borg 122 RID eHIC CSELIptlONe eeeP Pr ere eer rie tie cr bitte NEM Te Lia te oe oe 123 IRE clpalentcasiibenl CntSeerererim a2 ocr e.s- ee iscjee meu ee ciaciech a= Se cetera Set seemercte 126 AN, (OOS 3 a aa a BAS e oes SA ebe Sui CaO OD Sa AE Se ae ere ee eens ee I eee 127 Oxigimalgclescriptionpeye eree cee aoe ere clen cot one oe ats Seale es ee coe Meet emieee 127 EO CAG ER eer eee Pere rp er ae one ha ier eimalde dele a hab teens dle yee see se 127 Detailed description of type specimen. ...........------ Ban 5 OCGA AA SSPE SAAS AAS 128 (Erie palumeasineMmentsee see ease ss sharma nace: | iets cee s mer Sees Seen epee ey 131 AD, PRN MUES ore eb chose altich ies eRe Ore rT Ee RO EO Pe 132 STs (UL Cabs pee eee cre erate heii te See ala ls oS ddaare cialaltia meals SoiieG ne a) aeu oa BES 6 133 HD elatusten aes = ert eee Se Byte winle ete Sgidid AOC ae Ce ARIE A ce SA SAO AM AAT Tei eee 134 VOCE MIN Hs Sele Lemos he cone ec che ous Pacer or Yer SERE ees Cnet Seer a I aT gE ea 135 OMEN! CEATGHNUOM .. -mecncdoadedde jAddreaseo pre HOS cmendor oo poe GUEaoeE Spon: boaone oe 135 SPECI CHCaLUNCS MP Eee seme nye Re Sina ciah a cles Sein oiselsletelearslady Seth ats pine 135 Enimenpalemleraunomlenttsloiiyip Caer teeta ciate =e alee eters alate as oe efole ota tetas a ole cial (elm aso 136 WescunpuLOnMO ls DLESLO Ut Camere iste a miejers roe ael\yataei ie al al aia ofate alate abeiaisie v\ala\s aialatene/oln cialele 136 MCA SUMEMTEKUS TOL LCSLO Lys eterna erate tetetetnyafetey terete) x elele la ie /e afayeiafole|ateferatan\aia\clefm = s)<\cfeie s wielalaie\a|n'ale 138 iy VALLI CONTENTS. Cuaprer IV. Systematic descriptions of genera and species—Continued. Review of the species of Laramie Ceratopsia—Continued. Genera and species described by Marsh and Hatcher, cheifly from Converse Come Wyo.—Continued. Triceratops—Continued. TN :calicormis 2.42.2 sos a ee BSS i oS ale ate, ies ae ayo ee ap Locality sig sos 5 2 Sues se Ps sige eS oS So ree ore ea BOrieinal deseripbions l= 54492 seen! eee ee ead hate ees eee Lhe Description of typer =e ee TT. Obtusust2..< se soos Je eee se pate Bees Se i ES eI argent ek ee Pe WR WLEVICOLNUS! 36 ae ee eas ers net ae ee IE Eh Ne ais cies otro cla oa 43 Locality and horizom.222 55.252 3 as 6 nee ee oe ee ae Description of type: 22-2522. 2 fre se cS See ee SE ee Ee a Principal-measurem entise: 352.5585 ase seek Te ae See Sterol op hiss asses WY SEs are Bete yal M68 SO EO Ts 8 S: flabelilatus. 26 14. s)s 22a Se te Ea ee ene ate ape RR Ss eth ed ceo Owicnnall Cles@ayoiim..---2cesss5ssecee gna ds ee ye tine sreoeete aoa Gale ee eae Lotality and dhorizon:: 222) So) Shes 5S Resa eee Ge ener ae = Characters of acu 6 apis ole, See Re ee eS ear ae Diceratops: sss. 5 Sos sc 2 is ec Sra NS ae Pay tis es oP “ue Di hatcherics. 2/2 athe 8. 58s eee ee a Os eye eee Pee sek INod osaurtiss.ece LSS Sete ee ee a Es aes LR a I f=>.4. 11 b eee ea ens Maen fy pees Ae eh emir M rie ON oe eT Ji Sake pa i ee Part II. PHYLOGENY, TAXONOMY, Disrersurion, Hasits, AND ENVIRONMENT OF THE CERATOPSIA, By R. 8. Lunt. Generic.and specific summariys 222 NS I Se es Se cs Generalidiscussion’ of ‘the phylogeny 4:22. 05 Seek ao ee een oe ea ee nee ee ee See ae Revision ‘of genera g2 2) 2 4 ceis 8s oS cosh ae se eis Sistas a Se ale St eae SING ere ae ER a Mes Monoclompus-iriGerabopsinp yz ur ase eee eae Monoclomitis ss 2 senso hei Bo Gee Sine pera SE Bis ote ease ei era a Pee on Cenmtrosaunrst: 725 I ee cas Vee eae sD ee een ee en Pee Lis Aan STAs he a a Apart atin ase s2 2525 Sue Sis aise Sk IES Se SB epee cleans wo moans Os rT Bae et DICEPATOPS: seo Ge Ree CEE SEC se os Se eae ae ee ee ee ec ee OES orev ene Hepes rs ye Ee ae WPLCOTALOPS Ve Necks Asse Sirels, eee ORNS = Rises Sree we tts Sees See ee teaere sia ek ere r= eo ee The Ceratops-Torosaurus phylum.-............-...-.-.----- RS ieee eam eras Sie i Sc Ceratops).: 28-2 5.2 cee Rae el tat ete hie ss aE eee ey eet Ee ee MOVOSAUNU So o2 2 Dipad ions SSS SS eel RS cL Se EI ec Age ec Hhiminated ‘eénerasscc oak esate eet e SS ae pale = GS RIESE ere Se Claorhymehis en See pS See ied PS Se Bete AEN Seth cal oe ede Ore ce es ST ee exer en fey ee a Dy Se eUINULS Sh ayetyei eee te ete apr Shane aye oa Se ee ett He RNs ORS uo OE Pe pe Manospondylus: 22-205 555.2 se Pisses 2 he gk oe Re oo ho eee See elon ee ee es e Polyona x 5) 2ha By SRS ede Fiera eye Screwless Se ae a Oe Stepocerase 25 vak 2 SSN Go oR Se RS Soe Sota ini, Sere Oe eel Sterrholophuss:03:5 ss eas wee eee are Gees ee ery ae Sree eA ee ee Revision (of species’, 22-2 She tas setae ac yeas are ete ere Ue SP EEO kas ean ae Z The Monoclonius-Triceratops Ce PARSON ae Bis BOG cae Nou SUL Gos CeEe See oe oe eee Monoclonius \crassusio: 5: 5st ant ee Bre a eee. IN) I ee ll eae eee SMe Gawsomilj essed es 5 See DEEN, iret i he ae a ae A Mi sphenoGerls: 2iiss is doses Sie. Sree oe ote ie sash. eae See le ae eee ee Centrosaurus apertus).isc8 20 ie Slo ae See Renee oe Se eyo ee eee ee ee ee ‘Agathaumeasisyl Vestris 22 sai me Gh eis Seen wel a Do ee ee oe Ae eee a Diceratops hatcheris. csi Feel. Salk tive cesses sos eeeSee eeeae 31 2 emNcicolla lO Ga CORCIOleeUt CCRALOPSHDNONSUSM aan oe Mae ee Seey= ae ene A ee nen ee toate l= ee aaea: 33 SO MBVOSiLalnOttnyperOlaelntccrato DSiClatisiet = ts ey a | a ese ese one eee foes ee cles 33 3. Horizontal section of brain case of type of Triceratops flabellatus.........-....---/----+--------:------- 37 3), 1Bieiiin CHS OF INGARTOMS SERRE, TNIOBOSP AVION /oaacueacoss Soebeu sae doccoe sous slab eee eEeeeeee Seasooses 37 Soon PUGINASeCHONNOMSs KUO mintCeRatO psn /lDELLAIUS see = ese ae ee 2 ee ee 38 See ane CAST OLMPNICCHOLOP Sas CinG tis = laveralla wil @Wjaneysra= 2) 1 relay er eee Reyes See Seer ee esl 39 335, JARECKeItENAy OL IS Ol TN MUGRHOD janes Saaanaccs Sakura se caee seers saqeecse doses eeaspesaeseeee se esee 40 36. Right dentary of type of Triceratops TAD CH UUs ese Wet a Mate! ek Ree Ne Clas beret 0s Haked 41 Sipm belinsplenialuotmthy py CL OMMnICEenatopSyDTONSUS = =2 maestro eae ool is nei a ee eels Soe re eee oes eae Seas 41 Sametime arom merolelnicenato pss abellatiseee rere een a see ee eee ee ee eee en 42 Som eui pbiims tina Pu ater OlatayI) ChOlMUINICENOLO DSpfLQOCLLCLUS» = amen === se 5 see eee eee eee 42 40. Right angular, surangular, and articular of type of Triceratops prorsus..........--..-..--------------- 42 41. Lower jaw of type of Triceratops prorsus...-.-.------ side GIG BOEIS OBST e CES DEE CT eee eae Ele 43 12, SODOAOP WOON CINTA CKO SCT) Sacer Sectenceee kos a Baer oge bee eeepc eera Aaa ae ee eee 43 43. Vertical series of teeth from lower jaw of Triceratops flabellatus-.........----------------------------- 44 Teme TG Mle e UNO lM nT CenCLOpsp) LAD CLLACUS anaes cet Ie a niyeeareiaatere Seine laisse Jpee a oe ence st 45 45. Sections through roots of incipient and growing teeth of Triceratops.......-...-..---.-.--2+---++------ 45 AG, Siujociior gine! watertore ieee Ot MMe GhiOySsao55coscocdseqseedaccoasaqpoa4deea5 es a Jone eee Ee Saou Le ee 45 47. section of maxillary. of Triceratops, showing dental grooves...........-.---..:-----+--.+---+---------- 46 48. Presacral vertebral series of type of Triceratops brevicornus Hatcher.........--.----------------------- 47 LOM NULCMOMGeIVICA Il mVeLuc Duce Omi Mi CEnato Ds iPNONSUSHeme eee a a)se Seta eset ee ee ee eee a 47 DO MAUL IOn Cena erie bee OlMintCenatOps) PROTSUS sa— =e) -2 = esas e a oieis eer = oe ne elles eo ence 48 51. Anterior dorsal vertebra of Triceratops prorsus.......------------- BH oo CU CC Oe ep eae eye cea 50 DROS eH OM COIs lev enuc bla OUminiCenatopsiDNOnSUSm emit «/=is/ald'n) oleate mlb le aa/als/o' = 0) 57 62 Wet pubisioltypeot Limcenatops jlabellatus..-..--....---......----3------- Stee aAOEe a lpr ayn Age 57 (ah, Ugcliike. Ot WAM nope) (on Osta cas od peace no Ue oad o een Obpe BES eCner ee meas agen iota ce Chet 3 58 GAMMVCKIAG TOL ONC OLACOLGNOMUILCENALO DR UPTONEUS ant sees) oem cle we ole we Sale ecinlwie Se woes new ose va clo n=- 59 GHAMELTIIO RU SIOMMTCENGLO DEN PROTEUS AUCTION VIGWo seis enai = see eae 2 Notice of new fossil mammals: Am. Jour. Sci., 3d ser., vol. 34, October, 1887, pp. 323-324. ¢ Notice of gigantic horned Dinosauria from the Cretaceous: Am. Jour. Sci., 3d ser., vol. 37, August, 1889, pp. 173-175. HISTORY OF DISCOVERY. 7 For two or three years after the discovery of the type of Ceratops alticornis Messrs. George L. Cannon, George H. Eldridge, and Whitman Cross continued to find fragmentary remains of horned dinosaurs in the Denver and Arapahoe(?) beds in the vicinity of Denver, especially on and about the slopes of Green Mountain, an elevation rising about 1,000 feet above the plains, just without the foothills of the main eastern range of the mountains, about midway between Golden and Morrison and just west of the city of Denver. None of these remains were suffi- ciently complete to permit a determination of even the more important characters of these dinosaurs, and their true nature still remained unknown. DISCOVERY OF WELL-PRESERVED REMAINS OF CERATOPSIA IN CONVERSE COUNTY, WYO. In the summer of 1888, under the direction of Prof. O. C. Marsh, I undertook an expedi- tion for the purpose of collecting vertebrate fossils in the Judith River badlands of the upper Missouri River, a locality already rendered classic by the researches of Hayden, Leidy, and Cope. While this expedition was not especially successful, 14 boxes (about a ton) of rather fragmentary material was procured, including the skull fragments figured by Marsh and made the type of the new genus and species Ceratops montanus,* for which a new family, Ceratopside, was at the same time proposed. Notwithstanding their fragmentary nature the remains demonstrated the presence of horned dinosaurs in these deposits, a fact which had previously been suspected by Cope.? At the same time they threw much new light on the affinities: of the material previously collected by Hayden, Cope, Cannon, and others. After passing some two months in the Judith River badlands with very indifferent success, in the early autumn of the same year (1888), at the instance of Professor Marsh, I proceeded to southern Wyoming to investigate the remains of a fossil vertebrate discovered by Mr. Louis Lamotte, just south of the Seminole Mountains, on the west side of the North Platte River, about 1 mile from that stream and 40 miles below Fort Steele. These remains proved to belong to the Ceratopside, although this was not then suspected by me. There were present parts of the skull, vertebrz, ribs, and other portions of the skeleton, all in such a fragmentary and decomposed condition as to render their determination impossible, in the light of what was then known of these dinosaurs. After spending some time in this vicinity in a fruitless search for more perfect remains, I decided to abandon this locality and proceed to the White River badlands, lying east of the Black Hills in South Dakota, and pass the remainder of the season in making still further collections of the fossil sarmanelie so abundant in that classic locality. While en route I stopped at Douglas, Wyoming, and there met a former acquaintance, Mr. Deforest Richards, afterwards governor of Wyoming, now deceased, who introduced me to Mr. Charles A. Guern- sey. This gentleman, having a general but snilusiestic interest in matters relating to natural history, and especially to geology sind paleontology, had, through a long residence in the country as manager and owner of the ‘‘Three-Nine”’ cattle npn siceeedien in bringing together a senegal collection of fossils. On inspecting this calllscttinn: through the kindness and at the request of Mr. Guernsey, I was impressed with its value, G03 it ponteincd many specimens of great perfection and beauty, and only a glance was needed to show that the entire lot had been brought together with great judgment and discrimination, such as are rarely seen in amateurs and such as might with profit be emulated even'in some of our public museums, especially in their exhibition series. Among the many interesting things in this collection I was at once struck with the frag- ment of a very large horn core. This fragment was about 18 inches long and perhaps 8 inches in least diameter at the base, which was hollow, the cavity being filled with a hard, brown sandstone closely resembling the sandstone concretions that are so abundant in the Laramie. On inquiry Mr. Guernsey informed me that the specimen had been taken from a skull several feet in length which had been found by his ranch for eman, Mr. Edmund B. W ilson, completely a A new family of horned Dinosauria, from the Gre Am. Jour. Scei., 3d ser., vol. 36, December, 1889 pp. 477-478. b See Bull. U.S. Geol. and Geog. Surv. Terr., vol. 3, p. 592. 8 THE CERATOPSIA. embedded in a hard sandstone concretion, weighing not less than 2,000 pounds, that lay in the bottom of a deep canyon about 35 miles north of Lusk, Wyo. Observing my interest in the specimen, Mr. Guernsey very kindly assured me that if I wished to see the skull he would at some future time conduct me to the locality, from the immediate fulfillment of which favor he was then prohibited by pressing business engagements. Having completed my season’s work in the White River badlands, I returned to New Haven on January 3, 1889. In the meantime I had written Professor Marsh several letters concerning this peculiar horn core and he had published, in December, 1888, his description of Ceratops montanus, from the material which I had collected in the early part of the season in Montana and which for the first time demonstrated the presence of horned dinosaurs in the Judith River beds. This, together with the fact that Cannon, Eldridge, and Cross had already recovered undoubted dinosaur remains from the Denver beds, where the type of Bison alticornis had been found in situ, caused Professor Marsh to doubt the mammalian nature of the pair of horn cores which constituted the type of the latter and of which I had as yet only seen figures. When I examined these horn cores, soon after my arrival in New Haven, I at once recognized. the striking similarity between them and the horn core in the collection of Mr. Guernsey. I also expressed an opinion that the latter could have come only from Upper Cretaceous beds. I immediately wrote Mr. Guernsey requesting him to send on his specimens for further exam- ination and comparison. He very kindly and promptly complied with this request, and on its arrival in New Haven Professor Marsh at once recognized the remarkable similarity between the two specimens, and, after his characteristic nature, became immediately possessed with a burning desire to secure the skull and learn the exact geological horizon from which it came. ' Accordingly, on February 20, 1889, I left New Haven for Lusk, Wyoming, and although long delayed by inclement weather, the season being midwinter, succeeded in securing the remainder of the skull, our party having been conducted to the exact locality by Mr. Wilson, the original discoverer of the specimen. The skull was found embedded in a large sandstone concretion at the bottom of a deep canyon, exactly as had been described by Mr. Guernsey. The incidents connected with the procurement of this skull are here thus fully related for the twofold purpose of giving full credit to all concerned and of illustrating the manner in which one of the most important localities for vertebrate fossils was made known, for it was in this immediate vicinity, in Converse County, Wyoming, that the remarkable collection of Ceratopsia and the scarcely less remarkable collection of remains of other reptiles, as well as several thousand isolated jaws and teeth of diminutive mammals, were procured by me and my assistants, Messrs. Peterson, Utterback, and Sullins, during the years 1889-1892, while working in the interests of the United States Geological Survey and under the direction of Prof. O. C. Marsh. As the work of exploration in this newly discovered locality progressed the exact nature of this remarkable group of dinosaurs was rapidly brought to light and the real affinities of those remains, which had for a long time proved so puzzling to both Cope and Marsh, became apparent. The supposed ischia of Polyonax mortuarius, described and figured by Cope, were seen to be portions of the supraorbital horn cores, while the element considered by the same author as an episternal of Monoclonius crassus proved to be a parietal. The horn cores described by Marsh as pertaining to a bison proved to belong to a dinosaur, and the supposed dermal plates mentioned by him in his description of the type of Ceratops montanus became the squa- - mosals of that animal. Nor should these errors in identification be taken as a reflection upon the sagacity of either of these authors, but rather as additional evidence of the remarkable nature of these newly discovered dinosaurs, so different in so many osteological characters from anything hitherto discovered or suspected among representatives of that group. They are, however, striking examples of the many pitfalls that beset the path of the paleontologist when attempting to describe from insufficient or fragmentary material new genera and species— and especially new families—of extinct animals. They are, moreover, striking examples of that axiom so often disregarded in vertebrate paleontology, namely, that one observed fact is worth any amount of expert opinion. HISTORY OF DISCOVERY. 9 Professor Marsh was not slow to recognize the importance of the Converse County locality, and, with the tenacity and enthusiasm that were so characteristic of him, for four years he continued the work in that region, which was carried on for him by me, often under most _ discouraging circumstances, but which in the end resulted in the accumulation of the splendid i collection which forms the basis of the present monograph. This locality has since been visited by collecting parties sent out from the American Museum of Natural History, Princeton, Chicago, and the Kansas State universities, and from the Carnegie Museum. All of these have met with some success. DISCOVERY OF CERATOPSIA REMAINS IN CANADA BY L. M. LAMBE. For many years the fresh and brackish water deposits of the Upper Cretaceous on Milk, Red Deer, and Belly rivers in Canada have been known to contain remains of dinosaurs. It was not, however, until after the recent publication by Prof. H. F. Osborn and Mr. L. M. Lambe, based on the material brought together by the latter during the seasons of 1897, 1898, and 1901, through explorations carried on in the interests of the Canadian Geological Survey in the Edmonton and Belly River series of the Red Deer River district, that anything like an adequate knowledge of the nature of these dinosaurs was made known. The publication of this joint memoir by Osborn and Lambe“ has not only made known a new and highly impor- tant locality for the Ceratopside, but has extended their known geographical range and has brought to light a number of new forms. a his, ee ete 2 ede See eS een ee CERATOPSIA REMAINS DISCOVERED IN THE LARAMIE OF MONTANA. In 1902 the American Museum party, consisting of Messrs. Barnum Brown and R. S. Lull, discovered a nearly complete skull and other material of one of the larger members of the Cer- atopsia in beds.belonging to the Laramie on Hell Creek Canyon, 135 miles northwest of Miles City, Mont. The skull was briefly described by Lull.? a On Vertebrata of the mid-Cretaceous of the Northwest Territory: Geol. Surv. Canada, Contr. to Can. Pal., vol. 3, pt. 2, 1902, pp. 1-81. > Skull of Triceratops serratus: Bull. Am. Mus. Nat. Hist., vol. 19, pp. 685-695. = ay hee i 4 aeons ’ . Ra SS Cylal Bele sleet eile CLASSIFICATION AND REPRESENTATION OF THE CERATOPSIA. CLASSIFICATION OF THE CERATOPSIA. Crass: Reptilia. Suspco.ass: Dinosauria Owen. OrpDER: Predentata Marsh SUBORDER: Ceratopsia Marsh. Famity: Ceratopsidze Marsh (Agathaumide Cope). In the above scheme of classification I have regarded the Dinosauria as deserving the rank of a subclass, as was proposed by Marsh. While the Dinosauria embraces forms which it — appears to the present writer are sufficiently diverse in character to entitle the group to be ranked as a subclass,? still, on the other hand, there are a number of important characters common to all dinosaurs which would appear to justify the recognition of the Dinosauria as a natural group, although this has been denied by some authorities, and especially by the late Dr. George Baur, who held that the Dinosauria was not a natural group, but that it was com- posed of three distinct groups having no characters in common and showing no close relation- ships. These three groups were, according to Baur, the Iguanodontia (—Predentata Marsh), Cetiosauria (=Sauropoda Marsh), and Megalosauria (=Theropoda Marsh). In the light of our present knowledge concerning the structure of these extinct animals the position taken by Baur does not seem tenable. Without entering upon an extended discus- sion of the classification of dinosaurs, which would be quite out of place in this connection, it may be well to state briefly just what are the views held by the author of this monograph and _to set forth some of the more important reasons therefor. The following characters, first pointed out by the late Professor Marsh as possessed in common by all dinosaurs, would seem to be of sufficient importance to warrant their being considered as a natural group distinct from all other reptiles. These are: 1. Teeth with distinct roots, either fixed in more or less distinct sockets or in longitudinal grooves, never ankylosed; no palatal teeth. 2. Skull with superior and inferior temporal arches. . Double-headed cervical and thoracic ribs. . Sacral vertebrz coossified and more numerous than in other reptilia; seldom less than five. . Ilium extended in front of acetabulum, in the construction of which latter the ilium, ischium, and pubis take part. . Fibula complete. . The reduction in number of digits commences with the fifth. ND Of LW On the other hand, since we find within this group great diversity in form, structure, and habit, some of its members being carnivorous while others are herbivorous, some quadrupedal and others functionally bipedal, some unarmored while others are heavily armored, with all the many and diverse anatomical characters shown in their osteology which might reasonably be expected from such diversity of habits, there would seem no good reason why they should not form a subclass of the Reptilia comparable, for example, with the Metatheria among the Mammalia, and divisible into three orders, for each of which several names have been proposed by various authors. Of all these those proposed by Marsh appear the most appropriate. These are: _ 1. The Theropoda, embracing all the carnivorous dinosaurs. 2 The Sauropoda, embracing all the herbivorous forms in which the predentary is wanting. 3. The Predentata, embracing all the herbivorous forms in which the predentary is present. a Osborn considers the group a superorder within the subclass Diapsida Osborn.—R. S. L. 10 : ) ; 7 CLASSIFICATION. 1 In accepting the terms Theropoda and Sauropoda rather than Megalosauria and Opistho- celia or Cetiosauria, I do so out of regard for the more comprehensive nature of those terms as used by Marsh. The latter terms as used originally by Fitzinger (Megalosauri), 1843, Owen, 1859, and Seeley, 1874, respectively, I consider of subordinal rank only. Predentata of Marsh is preferable to Orthopoda Cope, because it is in no sense coordinate with the latter, but is a much more comprehensive term.. Cope’s Orthopoda and the Ornithopoda of Marsh (not Huxley) are more nearly synonymous. The Ceratopsia constitute a well-defined and rather compact group, which falls naturally within the Predentata as that order has been defined by Marsh. It embraces forms which in many respects are the most highly specialized representatives of the Dinosauria yet discovered. Marsh has considered the group as a suborder and defined it as follows: Premaxillaries edentulous; teeth with two distinct roots; skull surmounted by massive horn cores; a rostral bone, form- ing a sharp, cutting beak; expanded parietal crest, with marginal armature; a pineal foramen (?). Vertebrze and limb bones solid; fore limbs large, femur longer than tibia; feet ungulate; locomotion quadruped. Dernral armor. The above characters may be amended and supplemented as follows: Incipient teeth with single roots, which later become completely divided into two branches by replacing teeth; no true pineal foramen; dermal armor exceedingly imperfect; three [four, R. S. L.]@ anterior cervicals coossified; blade of iltum horizontal; post-pubis much reduced; ulna with well-devel- oped olecranon process; rudimentary fourth trochanter on femur. There is nothing new or original in the above scheme of classification. It is that used by Marsh in his ‘‘ Dinosaurs of North America,’’’ and may be taken as representing his latest and most mature views on that subject, with which, in the later years of his life, he was most inter- ested. It is used here in preference to the more recent classifications of other authors, because, in my opinion, it represents more nearly than any of these what is at present actually known _ of the structure and relations of the various groups of the Dinosauria to one another and to the Reptilia in general. All the Ceratopsia at present known pertain to one family, Ceratopside Marsh (—Aga- thaumide Cope). I accept the family name Ceratopside of Marsh rather than Agathaumide of Cope, because the former has priority and the genus Ceratops,° upon which it was founded, can not be shown to be a synonym of any of Cope’s genera. ALPHABETICAL LIST OF GENERA AND SPECIES. For convenient reference I give below a complete list, arranged alphabetically, of all the various genera and species that have at different times and by different authors been referred to the Ceratopsia, with a reference to the original descriptions of each and present location of the type. A. Agathaumas Cope. Proc. Am. Philos. Soc., vol. 12, 1872, pp. 481-483. 1. A. sylvestris. Proc. Am. Philos. Soc., vol. 12, 1872, p. 483. Type No. 4000, American Museum Natural History. 2. A. milo Cope. Bull. 1, ser. 1, U. S. Geol. and Geog. Sury. Terr., p. 10. Type in American Museum Natural His- tory; not identifiable. B. Ceratops Marsh. Am. Jour. Sci., 3d ser., vol. 36, Dec., 1888, pp. 477-478. 3. C. (Bison) alticornis Marsh. Am. Jour. Sci., 3d ser., vol. 34, Oct., 1887, pp. 323-324, and Am. Jour. Sci., 3d ser., , vol. 37, Aug., 1889, pp. 174-175. Type No. 1871 E, U.S. National Museum. 4. CO. horridus Marsh. Am. Jour. Sci., vol. 37, Apr., 1889, p. 334.—Triceratops horridus, Am. Jour. Sci., vol. 38, Aug., 1889, pp. 173-174. Type No. 1820, Yale Museum. 5. C. montanus Marsh. Am. Jour. Sci., volt 36, Dec., 1888, pp. 477-478. Type No. 2411, U.S. National Museum. 6. CO. (Hadrosaurus) paucidens Marsh. Am. Jour. Sci., vol. 37, Apr., 1889, p. 336, and Am. Jour. Sci., vol. 39, Jan., 1890, p. 83. Type in the U. S. National Museum. C. Claorhynchus Cope. Am. Naturalist, vol. 26, p. 757. 7. C. trihedrus Cope. Am. Naturalist, vol. 26, pp. 757-758. Type No. 3978, American Museum Natural History. a See p. 47. b See Sixteenth Ann. Rept. U.S. Geol. Survey, pt. 1, 1896, pp. 237-244. ¢ See p. 100. 12 . THE CERATOPSIA. D. Dysganus Cope. Proc. Acad. Nat. Sci. Phila., vol. 28, 1876, p. 250. 8. D. encaustus Cope. Proc. Acad. Nat. Sei. Phila., vol. 28, 1876, pp. 250-251. Type No. 5739, American Museum — Natural History. : 9. D. haydenianus Cope. Proc. Acad. Nat. Sci. Phila., vol. 28, 1876, p. 251. Type No. 5738, American Museum Natural History. : 10. D. bicarinatus Cope. Proc. Acad. Nat. Sci. Phila., vol. 28, 1876, p. 252. Type No. 3975, American Museum Natural History. 11. D. peiganus Cope. Proc. Acad. Nat. Sci. Phila., vol. 28, 1876, p. 252. Type No. 3974, American Museum Natural History. E. Manospondylus Cope. Am. Naturalist, vol. 26, pp. 756-757. 12. M. gigas Cope. Am. Naturalist, vol. 26, p. 757. Type No. 3982, American ee Natural History. F. Monoclonius Cope. Proc. Acad. Nat. Sci. Phila., 1876, pp. 255-256. 13. M. belli Lambe. Contr. Canadian Pal., val 3, pt. 2, pp. 66-67. Type No. 491, Geol. Sury. Canada. 14. M. canadensis Lambe. Contr. Canadian Pal., vol. 3, pt. 2, pp. 63-66. Type No. 1254, Geol. Surv. Canada. 15. M. crassus Cope. Proc. Acad. Nat. Sci. Phila., 1876, pp. 255-256. Type Nos. 3998, 3998?, 3998d, 3997, American Museum Natural History. 16. M. dawsoni Lambe. Contr. Canadian Pal., vol. 3, pt. 2, pp. 57-63. Type Nos. 1173 and 971, Geol. Surv. Canada. 17. M. fissus Cope. Am. Naturalist, vol. 23, 1889, p- 717. Type 3988, American Museum Natural History. 18. M. recurvicornis Cope. Am. Naturalist, vol. 23, 1889, p.716. Type No. 3999, American Museum Natural History. 19. M. sphenocerus Cope. Am. Naturalist, vol. 23, 1889, p. 716. Type No. 3989, American Museum Natural History. G. Nodosaurus Marsh. Am. Jour. Sci., Aug., 1889, p. 175. 20. N. textilis Marsh. Am. Jour. Sci., Aug., 1889, p. 175. Type No. 1552, Yale Museum. H. Polyonax Cope. Bull. U. S. Geol. and Geog. Surv. Terr., 1874, No. 2, pp. 24-25. 21. P. mortuarius Cope. Bull. U. S. Geol. and Geog. Surv. Terr., 1874, No. 2, pp. 24-35. Type No. 3950, American Museum Natural History. I. Stegoceras Lambe. Contr. Canadian Pal., vol. 3, pt. 2, pp. 68-69, 1902. 22. S. validus Lambe. Contr. Canadian Pal., vol. 3, pt. 2. Type No. 515, Geol. Sury. Canada. J. Sterrholophus Marsh. Am. Jour. Sci., vol. 41, Apr., 1891, p. 340. 23. S. flabellatus Marsh. Am. Jour. Sci., vol. 41, Apr., 1891, p. 340. Type No. 1821, Yale Museum. K. Torosaurus Marsh. Am. Jour. Sci., vol. 42, Sooty 1891, p. 266. 24. T. gladius Marsh. Am. Jour. Sci., vol. 42, Sept., 1891, p. 266. Type No. 1831, Yale Museum. 25. T.latus Marsh. Am. Jour. Sci., vol. 42, Sept., 1891, p. 266. Type No. 1830, Yale Museum. L. Triceratops Marsh. Am. Jour. Sci., vol. 38, Aug., 1889, p. 173. 26. T. calicornis Marsh. Am. Jour. Sci., vol. 6, July, 1898, p. 92. Type No. 4928, U.S. National Museum. 27. T. elatus Marsh. Am. Jour. Sci., vol. 42, Sept., 1891, p. 265. Type No. 1201, U.S. National Museum. 28. T. flabellatus Marsh. Am. Jour. Sci., vol. 38, Aug., 1889, p. 174; Sterrholophus flabellatus, Am. Jour. Sci., vol. 41, Apr., 1891, p. 340. Type No. 1821, Yale Museum. 29. T. galeus Marsh. Am. Jour. Sci., vol. 38, Aug., 1889, p. 174. Type No. 2410, U. S. National Museum. 30. T. horridus Marsh. Am. Jour. Sci., vol. 38, Aug., 1889, pp. 173-174. Type No. 1820, Yale Museum. 31. T. obtusus Marsh. Am. Jour. Sci., vol. 6, July, 1898, p. 92. Type No. 4720, U. S. National-Museum. 32. T. prorsus Marsh. Am. Jour. Sci., vol. 39, Jan., 1890, p. 82. Type No. 1822, Yale Museum. 33. T. serratus Marsh. Am. Jour. Sci., vol. 39, Jan., 1890, pp. 81-82. Type No. 1823, Yale Museum. SUPPOSED EUROPEAN REPRESENTATIVES OF THE CERATOPSIA. All of the above enumerated genera and species are from. the Cretaceous deposits of the United States, and all genera and species are included that have at any time been referred by any authority to the Ceratopside, although some of these, as will be shown later, pertain to families quite distinct from the Ceratopside, belonging even to different suborders. In addition to the above mentioned American genera and species several Kuropean forms have at various times and by various authorities been neganded as pertaining to the Certo. These are: A. Struthiosaurus Bunzel. Trans. Imp. Royal Geol. Inst. Vienna, 1871. 1. S. austriacus Bunzel. Trans. Imp. Royal Geol. Inst. Vienna, 1871. ; = B. Cratzomus Seeley. Quart. Jour. Geol. Soc. London, vol. 37, 1881, p. 637. 2. C.lepidophorus Seeley. Quart. Jour. Geol. Soc. London, vol. 37, 1881, pp. 660-667. 3. C. pawlowitschu Seeley. Quart. Jour. Geol. Soc. London, vol. 37, 1881, pp. 642-660. The above mentioned genera and species are from the Gosau beds (Upper Cretaceous) of Austria. In addition to these, various other dinosaurian remains from Europe and elsewhere have at different times been described by various authors as showing characters indicating EUROPEAN REPRESENTATIVES. 13 possible relationships with the Ceratopsia. Among these may be mentioned, as of special interest, (1) the peculiar horn-like dinosaurian bone from the Wealden of the Isle of Wight, described and figured by Richard Lydekker;* (2) a fragmentary bone, also from the Wealden of the Isle of Wight, figured and described by J. E. Lee® as pertaining to Polacanthus, and reprinted in his ‘‘Note Book of a Geologist;’’ (3) the various spines and dermal ossicles that have at different times been figured and described by various authors, usually as pertaining to Hylzosaurus, also from the Wealden.* Of these, the last all seem without doubt to pertain to Hylzosaurus or other allied genera, the affinities of which are with the Stegosauride rather than the Agathaumide. The same may be remarked, though with somewhat less emphasis, of the element described by Lee, but from the description and figure of the bone referred to by Lydekker it would appear at first glance that that element might very likely have pertained to a member of the Ceratopsia. From an examination of a cast of the specimen kindly sent me by Dr. A. S. Woodward, of the British Museum, I am convinced, however, that the original represented an ungual phalanx from the first or second digit of a large sauropod dinosaur, in which the proximal end with its articular surface had weathered away in such manner that the softer and more cancellous inner portion, by decaying more rapidly than the denser external wall, produced a cavity slightly resembling that found at the bases of the frontal horn cores in the Ceratopsia. The homologies of these various elements and the position which the different genera to which they have been assigned occupy in relation to the Ceratopsia will be more fully discussed in that portion of this volume relating to the revision of the genera and species. a Quart. Jour. Geol. Soc. London, 1890, pp. 185-186. b Ann. and Mag. Nat. Hist., 1843, p. 5. c See Owen, Foss. Rept. Wealden Form., pt. 4, Tables IV and IX. 2 De rh ia ED Be tt AER it Say NR SR SSN OS a CHARS Re alae Q OSTEOLOGY OF THE CERATOPSIDA. ‘The following rather detailed description of the osteology of the Ceratopside is for the most part based on the genus Triceratops. I have selected this genus, both because its osteology is more completely represented in the collections at my disposal than that of any of the other genera of the group, and because it is, on the whole, less divergent and more fairly representa- tive of the family than most of the other genera. While this description of the osteology will in the main be based on remains pertaining to the genus Triceratops, reference will occasionally be made to material pertaining to other genera, more especially where striking and important structural or morphological differences are displayed. THE SKULL. All the more important elements of the reptilian skull are present in the Ceratopsia, though — in many instances they are so much modified as to bear little resemblance to their homologues in the skull of most other members of the Reptilia. In the Ceratopsia perhaps more than in any other group of reptiles, fossil or necent, the ‘skull has become greatly modified and specialized in certain directions. The chief specializa- tion has been in the direction of affording increased protection and in the development of more efficient organs for procuring food. Specialization along the former line has resulted in the development of the powerful armature seen in the nasal and frontal horns, the enormous expan- sion of the parietals and squamosals, and the development of the epijugals and epoccipitals. The development of the predentary and rostral bones doubtless increased the animal’s ability to obtain food, while they at the same time served as additional protective organs. The com-_ pact nature of the skull as a whole is a striking feature in this group of dinosaurs and contrasts strongly with the loose and open structure so characteristic of the theropod and sauropod dinosaurs and of the skulls of most other reptiles, living and extinct, where the different cranial elements are, for the most part, reduced to more or less slender rods, usually separated by large vacuities and loosely attached to one another, the union between the different elements often being only cartilaginous. The inclosed and compact nature of the skull in the Ceratopside was of the greatest value as a means of protection. It parallels that seen in most mammals and is shown, though in a less perfect condition, in most of the Chelonia, as also in Parevasaurus and the pacuncdlont reptiles. The more striking features just noticed as being present in the skull are well shown in fig. 5, taken from a photograph of the type of Triceratops prorsus mounted in the Yale Museum. The brostl hood-like parietal crest; the powerful supraorbital horn cores laterally placed and projecting forward and upward; the somewhat less powerful median nasal horn, projecting almost directly forward; the formidable cutting beak formed by the opposing noeine and predentary bones, borne respectively by the premaxillaries and dentaries; the large laterally placed orbits and narial orifices and the comparatively small laterotemporal fontanelles are . conspicuous and striking features, at once distinguishing the skull of the Ceratopside from that of any other known family of reptiles. Seen from above, the skull appears wedge-shaped with the apex directed anteriorly. : 14 THE SKULL. 15 THE CRANIUM. Although the skull in Triceratops and the allied genera is larger than that of any other known land animal, living or extinct, the cranium proper, or brain case, is not unusually large, while the brain itself is remarkably small as compared with the size of the animal as a whole or of the skull alone. Only the bones of the occipital and parietal segments take any considerable part in the formation of the brain case proper, and some of these are ex¢luded. The post- frontals do, however, in some cases form a very small part of the anterior portion of: the superior wall of that part of the brain case inclosing the olfactory lobe. The different bones of the cranium proper were the first in the skull of the Ceratopsia to become coossified with one another. It thus happens that the anatomy of this region of the skull is somewhat difficult to interpret in adult specimens. Fortunately there are in the Yale Museum collections skulls of a number of young individuals in which the sutures are still sufficiently distinct to show most of the characters of the different bones of this region. One of the skulls (No. 1821), which a SSeS EOEEEOSSNSNSNMNNNS. oo oe oaa——amiaws='’ oe ‘See, Fie. 5.—Skull of Triceratops prorsus (type), No. 1822, Yale Museum, as seen from left side. About one-sixteenth natural size. After Marsh. was made the type of Triceratops (Sterrholophus) flabellatus by Professor Marsh, is especially valuable in this respect, being essentially a disarticulated skull. Unless otherwise stated in the text, the following description of the skull of Triceratops will be based on this specimen. THE OCCIPITAL SEGMENT. All the bones of the occipital segment are present, though the supraoccipital is somewhat reduced in size. \ In fig. 6 there is shown an oblique posterior view of these bones, with the parietals and squamosals in position as seen from beneath the latter elements. The basioccipitals coossify early with the exoccipitals and leave no indication of the sutures even in the skulls of comparatively young individuals. This is the case in the type of T. flabellatus. In the occipital condyle of a younger individual (No. 1831, Yale Museum) forming the type of Torosaurus gladius, however, the sutures are still open and show that the condyle was formed by the union of the basioccipital and exoccipitals. The basioccipital constituted about one-third of the occipital condyle, but formed no part of the border of the foramen magnum. The basioccipital is massive, and the processes are extremely heavy and terminate in broad, rounded, rugose extremities, which have the appearance of articular sur- ~ 16 THE CERATOPSIA. faces, but which in life were doubtless covered with heavy cartilaginous pads. The basioc- cipital processes are connected distally by a rather thin lamina, and posteriorly there is a deep pit just beneath the occipital condyle. In front they abut against the expanded posterior borders of the processes of the basisphenoid, with which they are united by suture only except in very old individuals, in which they become coossified and the sutures become more or less perfectly obliterated as the age of the animal increases. The basioccipitals are pierced by no foramina. The exoccipitals are large and massive. They entirely inclose the foramen magnum and expand laterally into greatly elongated processes, which project outward and backward, over- lapping the blades of the quadrates and articulating distally with the squamosals by a deep groove on the inferior surface of the latter. The superior border of the distal end of each exoccipital process is expanded and projects into a corresponding notch on the inferior surface of the antero-inferior angle of the parietal. From this notch the suture between the parietals_ and exoccipitals describes a gentle curve until the supraoccipital is reached, when the superior AY iss Lei. | Gag i ' fii Wy Ye Ss nN Vig I Sey OTT Wy UY W Yitzzzan = wea BEZZE_A Ly, SE EEA ZL Fic. 6.—Occipital region of Triceratops flabellatus Marsh (type), No. 1821, Yale Museum. Inferior view as seen obliquely from behind. a, Postfrontal buttress of supraoccipital; so, supraoccipital; pa, parietal; sq, squamosal; exo, exoccipital; fm, foramen magnum; 0c, occipital condyle; bo, basioccipital; bs, basisphenoidal processes; j, jugal; gj, quadratojugal; x, x1, foramen for exit of tenth and eleventh nerves; XII, foramen for exit of twelfth nerve. One-eighth natural size. border of the exoccipital is produced into an obtuse angle embraced by the parietal and supra- occipital. The exoccipitals and supraoccipital unite by a nearly horizontal suture. The pos- _ terior surface of these bones is deeply excavated, inclosing two deep pockets separated by a thin median partition or keel, which becomes stronger above the supraoccipital suture. On the inferior surface and at the base of the lateral or transverse process of the exoccipital there are two foramina, situated one above the other when the skull is in its normal position. The upper and smaller of these two foramina pierces the exoccipital and enters the foramen mag- num just within the external opening of the latter. It doubtless served to transmit the hypo- glossal, or twelfth nerve. The more inferior but larger foramen divides, one branch, the posterior, opening into the foramen magnum a little in advance of that for the hypoglossal nerve, prob- ably served for the transmission of the eleventh, or accessory, nerve, while the more anterior branch, which opens into the foramen lacerum posterius, may have served for the passage of the pneumogastric nerve. The sutures between the exoccipitals and alisphenoids evidently \ THE SKULL. 12 were early obliterated. They are not visible in any of the skulls known to me. The foramen lacerum posterius is large, as shown in fig. 8, and is situated at the base of and just anterior to the processes of the basioccipital and exoccipital. It is probable that the basioccipital and exoccipital and the basisphenoid and alisphenoid all took part in inclosing this foramen. The suture between the supraoccipital and exoccipital remains open until late in life, so that the form and character of the supra- occipital are easily determined. It is proportionally small and composed of two rather thin wing-shaped portions articulating inferiorly on either side with the exoccipitals and alisphenoids. Each of these is produced backward, so as to inclose and overhang the deep cavity already mentioned in describing the exoccipitals, and they unite medially to form the strong superior portion of that keel which divides this cavity vertically into two lateral cavities. Superiorly the supraoccipital sends upward two pillars, one on either side of the median line, to give support to the parietals at their union with the postfrontals, thus contributing to the firm support __ Ne c : aie Fic. 7.—Occipital condyle of type of of the supraorbital horns. In this manner the supraoccipital Horeconrus igladiue ENowIs3i. vale unites with the widely expanded processes of the exoccipitals in HUIS, WO) Betas, cet giving support from beneath to the enormous frill-like occipital ig Bs Coat eae crest formed by the squamosals and parietals. The condyle is pedunculate. The peduncle, or neck, is rather short, circular in cross section, and is usually somewhat constricted. The basi- occipital and the two exoccipitals enter about equally into its construction, though the former is excluded from the foramen magnum. The articular surface of the condyle has the form of a nearly perfect hemisphere, and indicates considerable freedom of motion, notwithstanding the depth of the cup in the atlas and the enormous posterior projection of the frill. The condyle of the type of Tricera- tops prorsus (No. 1821, Yale Museum) is 91 millimeters in vertical diameter and 102 millimeters in transverse diameter. The foramen magnum is small and somewhat ellipti- cal in outline, the vertical axis being the longer. It has a vertical diameter of 36 millimeters and a transverse diameter of 32 millimeters. The superior border of its Fic. 8.— Occipital region of skull of Triceratops flabel- exit is somewhat in advance of its inferior, so that it looks latus (type), No. 1821, Yale Museum, as seen obliquely from beneath, infront. pfp, Pillar for upward and backward. support of parietals and postfrontals; so, supra- occipital; ex, exoccipital process broken away; THE SPHENOIDAL SEGMENT. as, alisphenoid; ab, alisphenoid buttress for post- ; frontal; ptf, postfrontal; oc, occipital condyle; bo, basioccipital process; bs, basisphenoid; pt, surface The basisphenoid is firmly coossified with the alisphe- on, basiphenoid process for contact with ptery- goid; x1, foramen for twelfth nerve; x-x1, fora- men for exit of tenth and eleventh nerves; jflp, foramen lacerum posterius; eam, internal audi- tory meatus?; ca, foramen for carotid artery; fo, foramen ovale; z, undetermined foramen; pf, fora- mina entering pituitary fossa; sf, sphenoidal fis- sure; of optic foramen; fr, foramen rotundum; ff, small foramina entering brain cavity just back of olfactory lobe. One-eighth natural size. noids. The basisphenoidal processes are produced some- what beneath the basioccipital processes and present in front at their extremities rugose surfaces for contact with the pterygoids. Anteriorly and superiorly the basisphe- noid is compressed, and forms a stout median interorbi- tal septum. The external opening of the middle eusta- chian canal is situated between and at the base of the basisphenoidal processes, as shown at mec in fig. 9 and in PI. XLVI. It is entirely within the basisphenoid instead of being situated between that bone and the basioccipital, as in the croco- dile. Two large foramina, situated one on either side of the skull at the bases of the basisphe- noidal processes, pierce the basisphenoid and enter the brain case near the base of the olfactory MON XLIx—07——2 18 THE CERATOPSIA. lobe through a deep fossa, which doubtless lodged the pituitary body. They probably trans- mitted the carotid arteries. One of these foramina is shown at ca in fig 8. Anteriorly the basisphenoidal processes are received into deep pockets cn the posterior surfaces of the thin but widely expanded posterior wings of the pterygoids. The alisphenoids, including also the parasphenoids, with which they are so completely fused, even in young individuals, as to render the latter elements indistinguishable, are extremely irregular in form. They are firmly coossified with one another and with the exoccipitals and the basisphenoid. Together with the latter element they usually form the entire anterior portion of the brain case, save only the extreme anterior portion of the superior border, which in some instances is formed by the anterior projection of the united postfrontals. Supero-posteriorly the alisphenoids articulate with the supraoccipital and supero-anteriorly with the postfrontals. Just beneath the lateral union of the supraoccipital and post- frontals the alisphenoids are developed into a strong lamina or Fic. 9.—Occipital region of type of Tri- buttress, which gives greater support to this region. Anteriorly ceratops flabellatus, No. 1821, Yale Mu- . : . ' : coi aod Galtanly chow helo the coossified alisphenoids and basisphenoid are embraced by Frontals; pf, postirontals; as, alisphe- the vomers and the posterior projections of the palatines, as is well shown in fig. 24, from the type of Triceratops horridus, No. noids; exo, exoccipitals; bo, basioccipi- tals; bs, basisphenoids; spt, surface for 5 contact with pterygoids; o/, exit for 1820 of the Yale Museum collections, though not so apparent See) eae Gear Ean in the type of T. flabellatus, No. 1821 of the same museum, in which these parts are less perfectly preserved. In the type of T. flabellatus the olfactory nerves, as shown in fig. 9, leave the brain case by a single large median foramen inclosed entirely by the alisphenoids, while in the type of 7. horridus this foramen is subdivided by a strong but short median partition of bone, as shown in fig. 27. As will be shown later, however, this differ- ence is probably due to age, and is of no specific or generic im- portance. In the latter species the alisphenoids are continued for- ward and form the entire roof of the olfactory lobes, entirely ex- cluding the postfrontals and giving origin to that partition which has just been mentioned as dividing the foramen for the exit of the olfactory nerves into two fora- mina, conditions well shown in the figure last cited. Near and an-’ Fic. 10,—Skull of type of Triceratops flabellatus, No. 1821, Yale Museum, seen from left terior to the base of the ascending side. h, Supraorbital horn core; c, supratemporal fontanelle; b, orbit; e, epoccipi- lamina of the alisphenoid two small foramina (ff, fig. 8), situated tral; p, predentary. One-twelfth natural size. After Marsh. one in advance of the other, pierce the bone and enter the brain cavity just back of the olfactory lobe. At a distance of 38 millimeters behind the more posterior of these two foramina is the anterior opening of the alisphenoid canal, or foramen rotundum. The alis- phenoid canal is inclosed externally by a heavy bridge of bone, and is continued posteriorly for 35 millimeters, when it enters the foramen ovale. The latter foramen is large and is sepa- rated from the foramen lacerum posterius by a strong bony partition, which is pierced by a small foramen that may possibly represent the external auditory opening. Just beneath the tal; q, quadrate and quadratojugal; a, anterior nares; h’, nasal horn core; 1, ros- d THE SKULL. 19 anterior opening of the alisphenoid canal there is a rather deep pocket, at the base of which arefourforamina. The two larger of these foramina occupy a more elevated position than the two smaller, and are placed one anterior to the other. The anterior and slightly larger of the two is the optic foramen, while the posterior is doubtless homologous with the sphenoidal fissure.. The two smaller foramina (pf, fig. 8) lead directly into the comparatively deep and elongate pituitary cavity, and their homologies are not definitely known. They may have served to transmit nerves to the large muscles of the orbit. The positions of the various foramina mentioned above are well shown in. fig. 8, which represents the right side of the external wall of the brain case with the supraoccipital and postfrontal in position, as seen obliquely from beneath and in front. In so far as I have been able to determine, the position of these foramina is fairly constant in the various genera and species of the Ceratopside. THE FRILL OR PARIETAL CREST. The parietals and squamosals are enormously expanded and form a great hood-like frill, which projects far beyond the occipital segment of the skull, entirely inclosing both superiorly and laterally the cervical region. In life the frill doubtless afforded considerable protection to the neck, especially in the genus Triceratops, though it could not have been so effective as a protective armor in Torosaurus. The special- ization which has taken place in the bones of the frill has not been uniform in the different genera and species of the Ceratopsidz, so that this region of the skull is of the greatest importance in determining genera and species. In Tricera- tops it consisted of a continuous sheet of bone, composed of the closely applied squamosals and parietals. The inferior surface was deeply concave transversely, and the whole projected backward and extended for a distance of nearly or quite a meter behind the occipital condyle, roofing over the entire cervical region above and extending well down on either side. To increase its effectiveness as a protective shield, and perhaps also at the same time to give to it a cer- 7 ios tain value as an offensive weapon, its external margin or als periphery was curved outward and supported a series of ie. 11—Posterior view of skull of Triceratops elongated, acuminate, triangular ossicles (epocel pitals), eas Sine) No! Bo Sea daar which have the appearance of having been covered in life peparietalsls, squemosall q) quadrate: pd) pre- with some hard chitinous substance. When in position, — entary; @ dlentary. OED N E nee these give to the border of the frill a scalloped or serrated appearance. In the genus Triceratops these ossicles are derived from separate centers of ossification and do not ankylose with the squamosals and parietals until late in life. In the type of T. flabellatus, which pertained to a young individual, they were for the most part found in position, but free; while in- the type of 7. prorsus, representing an old animal, they are firmly coossified with the squamosals and parietals.. They are well shown in figs. 10 and 11, as is also the general form of the frill. In the genera Monoclonius and Torosaurus there are no epoccipitals, but the margins of the squamosals and parietals each present a series of more or less distinct prominences, giving a serrated appearance to the margin of the frill, similar to that seen in Triceratops. In the first-mentioned genera, however, these promi- nences are not derived from separate centers of ossification, as are the epoceipitals in Tricer- atops, but are present even in young individuals, firmly attached to and forming a part of the squamosals and parietals, as shown in figs. 12 and 13 and in Pls. If and III. The variation in the form and structure of the frill in the various genera and species of the Ceratopsid is well shown in Pl. II, where a comparative view of the different types of frill is given. fa 20 THE CERATOPSIA. The squamosals vary ereatly in form in the different genera and species. In Triceratops: flabellatus the form of the squamosal is intermediate between that seen in Ceratops montanus Fic. 12.—Superior view of frill of Torosaurus gladius (type), No. 1831, Yale Museum. , Parietal; sq, squamosal; pf, parietal fontanelle sf supratemporal fossa; pfs, surface for contact with postfrontal; z, foramen analogous to the postfirontal foramen in Triceratops. One-sixteenth natural size. : and Torosaurus latus or T. gladius, as shown in the accompanying figures. It is broad anteriorly and narrows posteriorly. The posterior and inferior border is undulating and supported six small pointed dermal ossicles (epoccipitals), as shown in fig. 10. The external surface for some distance inside this border is very gently concave, but above this it is uniformly convex. The parietal suture is open throughout the entire length of the bone, and the parietal border is heavy. Anteriorly the squamosal is in contact with the quadrate, the exoccipital, the jugal, and the postfrontal, and for a limited extent with the quadratojugal, besides sending a strong process underneath the parietal. It overlaps the external surfaces of all these bones and has a very extended contact, especially with the postfrontal, through ened ; o ; 3 Fig. 13.—Superior view of parietal of type of Monoclonius crassus, No. 3998, — which it forms the chief support posteriorly American Museum of Natural History. sqs, Surface for articulation and interiorly to the massive supraorbital with squamosal; pfs, surface for postfrontal. One-eighth natural horn. On the inner side, near its anterior a: end, the surface of the squamosal is crossed obliquely by a heavy ridge of bone, presenting anteriorly an abrupt shoulder, against which the distal extremity of the exoccipital process THE SKULL. Dal abuts, while the broad, thin end of the quadrate is wedged in between the exoccipital process and the anterior end of the squamosal and is rather firmly held in a groove on the under surface of the latter. The position occupied by this groove for the quadrate, as well as the form of the squamosal, varies greatly in the different genera and species. On the inferior side of the squa- mosal, just posterior to the transverse buttress for the exoccipital process, there is a broad, shallow, concave surface, while be- yond this the inferior surface of the bone is convex throughout. The line marking the boundary between these convex and concave areas is approximately that which separ- ates the external free portion of the squamosal from the internal por- tion, which in life was embedded in flesh. The dividing line between these two areas is indicated by a difference in texture and surface markings, that of the external and free area being much more rugose, : soe : : 2 ake Fic. 14.1, Right squamosal of Torosaurus gladius (type), No. 1831, Yale Museum; especially im old individuals, where inferior view. 2, Right squamosal of Triceratops flabellatus (type), No. 1824, Yale it was probably covered with a Museum; inferior view. 3, Right squamosal of Ceratops montanus (cotype), No. Bers 4802, U. S. National Museum; inferior view. e, Epoccipital; s, parietal surface; q, horny substance similar to that quadrate groove. All one-twentieth natural size. After Marsh. which invests the skulls of recent turtles. In the skulls of young individuals like that at present under consideration (No. 1821, Yale Museum), these two surfaces are less distinctly differentiated than in the skulls of older individuals, though still quite apparent, as shown in figs. 6 and 15. Marsh has considered the Fic. 15.—Inferior view of right squamosal of Triceratops flabellatus (type), No. 1821, Yale Museum. e, Epoccipital; sp, surface for contact with parietal; ex, buttress for exoccipital; gg, quadrate groove; p, process overlapping quadratojugal; /, margin forming posterior border of lateral temporal fossa; pfs, postfrontal border; jp, process interposed between jugal and postfrontal; exs, surface overlapped by exoc- cipital process. One-eighth natural size. partial absence of this character as of generic importance, and has made it distinctive of the genus Sterrholophus, of which the present specimen was the type. To the present writer it would seem, however, that the difference in the rugosity of the external surface of the squamosal is due rather to a difference in the age of the different individuals, while it may also have been, in part at least, a sexual character. In support of this view it may be mentioned that the rugosities on this surface are less distinct in those specimens where the sutures are less obliterated DD, THE CERATOPSIA. than in those where they are more perfectly closed, regardless of the genus to which the species may pertain. This is especially noticeable in Nos. 1821 and 1823, Yale Museum, the types, respectively, of Triceratops (Sterrholophus) flabellatus and Triceratops serratus, in both of which most of the cranial sutures are still open,indicating that the animal was in each instance com- paratively young. The great diversity in size and form displayed by the squamosals in the Ceratopsidz is well shown in PI. III. The parietals, though perhaps separated in extremely young individuals, early became perfectly united, leaving no trace of a suture. They are imperfectly preserved in the type of Triceratops flabellatus, but are quite complete in a number of other specimens in the Yale and U.S. National Museum collections, and especially so in the type of T. serratus, No. 1823, Yale Museum, as shown in fig. 16 and in Pl. XXVIII. The following description of these elements will be based largely on that skull. Superiorly they are convex laterally and somewhat concave longitudinally, especially along the median line. Together they form about one-half the crest or frill of the skull. Laterally and posteriorly they articulate with the squamosals throughout about two-thirds of their length on the superior surface, when this articulation or contact is interrupted on either side by the large supratemporal foramina or fosse. On the inferior side or surface the parietals and squamosals are in contact throughout their entire length save at the point where the exoccipital processes are interposed between them, as shown in fig. 6. Anteriorly the parietals are overlapped by the postfrontals and underlapped by the supraoccipitals and exoccipitals. They are much expanded posteriorly but contract rapidly ante- riorly. On their superior surface along the median line or crest there is a series of rugose prominences, usually three innumber. The posterior border sup- ported seven low pointed epoccipitals, - one median and six lateral arranged Fic. 16.—Superior view of skull of Triceratops serratus (type), No. 1823, Yale three on either side. Though there is Museum. s, Squamosal; p, parietal; c, supratemporal fossa; e, epoccipital ; Smee j d, epijugal; x, postfrontal foramen; fp, postfrontal; pf, prefrontal; f, frontal; apparently some variation in the num- iD jugal; h, supraorbital horn core; 1, nasal; h’, nasal horn core; pm,premax- her of these ossicles it hardly seems illary; 7, rostral. One-twentieth natural size. After Marsh. Spe ee ; probable that such variations are of specific importance or that the number was constant in different individuals for any given spe- cies. Seven seems also to have been the number present in No. 1822, Yale Museum, the type of T. prorsus, though the parietal is not entirely perfect in that specimen and it is therefore impossible to determine the number with certainty. THE QUADRATOJUGAL ARCH. The lower temporal or quadratojugal arch is formed by the jugal, quadratojugal, and quadrate, and these bones are united by suture only even in comparatively old individuals, as shown in fig. 10. The quadrate, fig. 17, A and B, is triangular in outline. When in position the longer border or base of the triangle is almost parallel with the vertical axis of the skull, though directed somewhat backward and upward. The inferior border is nearly straight and heavy throughout its entire length, and might very appropriately be called the shaft, to distinguish it from the extremely thin angular superior border. The anterior extremity of the quadrate, although much narrower than the posterior, is heavier and presents an elongated and trans- _ versely placed articular surface for articulation with the lower jaw. The posterior end of the quadrate is wider but much thinner than the anterior and is overlapped by the squamosal and — ag THE SKULL. 23 underlapped by the exoccipital process. On the internal side near the superior border and toward the anterior end of the quadrate there is a deep pocket into which fitted the posterior process of the pterygoid, while the thin posterior blade of the pterygoid overlapped the delicate superior angle of the quadrate when these bones were in position, as shown in fig. 24, from the type of Triceratops horridus. The external border of the quadrate anteriorly, throughout half its length, is opposed to the quadratojugal. The latter element sends backward a keeled process, which curves round the inferior border of the thickened shaft of the quadrate and is continued upward and backward along the external surface of that bone, finally com- ing in contact with and passing beneath a sharp-pointed anterior projection from the squamosal, as is beautifully shown in PI. XXVII, from the type of Triceratops serra- tus, No. 1823 of the Yale Museum collection. Marsh has stated that the quadratojugal does not articulate with the squamosal in the genus Ceratops, and has considered this as a character distinguishing that genus from Triceratops, in which genus he states that these two elements are firmly united with one another by sutures. While there is a considerable difference in the character Fic. 17.—A, Internal view of right quadrate of type of Triceratops flabel- s s latus, No. 1821, Yale Museum; B, external view of same. sq, Surface of the eae of these elements ek these two fitting into quadrate groove of squamosal; pt, suriace overlapped by genera, the union being apparently much pterygoid; ptn, notch for process of pterygoid; sa, surface for articu- stronger in Ceratops, it is clear that they do LE ae ae SL ik ae ee articulate with one another in Triceratops. The articular surfaces for the lower jaw on the quadrates are in nearly the same vertical plane as the occipital condyle in the Ceratopsia instead of being placed far back as in the crocodile. The quadratojugal, fig. 18, A and B, is irregularly triangular in outline, very thick medially, but thin about the margins. It is wedged in between the jugal and quadrate and also has,. as stated above, a limited contact with the squamosal. The squamosal and quadrate articula- tions have already been described and need no further mention here. The surface for articula- tion with the jugal is irregularly elliptical in outline and occu- pies the anterior half of the external surface of the bone. It bears a number of deep grooves and intervening ridges, which correspond to similar inequalities on the opposing surface of the jugal, thus giving great strength and rigidity to this region. The quadratojugal is the smallest of the three ele- ments which in the Ceratopsia form the quadratojugal arch. The jugal is the largest bone of the three entering into BE ee ead Gta actoi eit of the construction of the quadratojugal arch. Seen from the type of Triceratops flabellatus, No. 1821, Yale Side when in position it appears T-shaped, the stem being ee ee eattocccotatey ad formed by that portion of the bone which overlaps the quad- Wao ieata Orieeal nie natural sive. © ratojugal, as shown in fig. 10 and in Pls. XX VII and XXXIV. The posterior branch of the crossbar at the top sends back- ward a rather slender bar which interlocks with an anterior projection from the squamosal, as shown in P]. XLIV, and forms most of the supero-external border of the lateral temporal fossa. Anteriorly and superiorly the transverse portion of the jugal articulates with the maxillary and sends forward a narrow process which is interposed between the lachrymal and the superior bor- der of the maxillary. At the extreme top the jugal forms the antero-inferior border of the orbit, ‘inclosing somewhat less than one-third of that opening. Antero-superiorly it articulates with D4 THE CERATOPSIA. the lachrymal and postero-superiorly with the postfrontal. The jugal thus forms the external surface of the skull throughout the anterior portion of that area lying between the orbit and the infratemporal or lateral temporal fossa. The relations of the various elements constituting the quadratojugal arch to one another and to the surrounding bones are well shown in figs. 6, 10, 24, and 26, and in Pl. XLIV. THE FRONTAL REGION. The prefrontals, frontals, and postfrontals are all present, though the relative importance of each is somewhat different from that which*is usual in the reptilian skull. The characters of these elements are best seen in the type of Triceratops serratus, No. 1823 of the Yale Museum collection, and I shall base my description of these elements on that skull. The frontals are small and become coossified in early life medially with one another, later- ally with the prefrontals, and posteriorly with the postfrontals, so that in the skulls of old indi- viduals the sutures between all these bones are completely closed, the whole forming then a single bone roofing over the entire top of the skull between the parietals and nasals, constituting about two-thirds of the orbital borders and culminating in the transversely placed and powerful pair of supraorbital horn cores. In the types of Triceratops flabellatus and T. serratus, Nos. 1821 and 1823 of the Yale Museum collections, most of the sutures between the different ele- ments of the frontal region are still open, so that in these skulls it is still possible to determine the form of the different frontal bones. The true frontals are abruptly truncated posteriorly in the type of 7. serratus, No. 1823, but anteriorly they are deeply emarginate and receive the posterior median projection of the nasals. The frontals form no portion of the orbital border, nor do they contribute to the formation of the supraorbital horn cores. The prefrontals are placed alongside of and external to the frontals, which they about equal in size. Supero-posteriorly they articulate with the postfrontals and immediately beneath the postfrontal suture they form the extremely thick anterior border of the orbit, which in this region is projected far beyond the general surface of the skull. Inferiorly they articulate with the lachrymals, while anteriorly they are in contact with the nasals. The postfrontals in the Ceratopsia have become of unusual importance. They are several times larger than the frontals and prefrontals combined. They form more than one-half the orbital border and are extremely massive posteriorly. They alone give origin to the so-called frontal horn cores, which might therefore be more appropriately called the postfrontal or supra- orbital horn cores. The latter name is preferred by the present author. Anteriorly they are in contact with the frontals and prefrontals, posteriorly with the parietals and squamosals, while immediately beneath the center of the orbits they oppose the jugals. From beneath the postfrontals receive support from both the supraoccipital and the alisphenoid. On the median line and just in advance of the parietal suture a large foramen is usually found piercing the external wall of the postfrontals. It has been called by Professor Marsh the pineal foramen. This foramen appears to have communicated with the large cavities in the postfrontals at the base of the horn cores and with certain other smaller cavities in these bones and between them and the supraoccipitals. Its functions are not known, but it is not unlikely that it served for the transmission of nerves and nutrient blood vessels. In some skulls of very old individuals it is almost or entirely closed; the latter is the case in the type of Triceratops prorsus, No. 1822, Yale Museum. The base of the supraorbital or postfrontal horn core is invaded by a very large cavity with comparatively thin walls of bone. Between the horn cores there are a number of smaller cayi- ties situated within the body of the postfrontals, as shown in fig. 24. A number of somewhat larger cavities lie between the supraoccipital and the posterior portion of the postfrontals and anterior portion of the parietals and squamosals. There appears to be considerable individual variation in the number and relative size, form, and position of these various cavities. | a This is inno sense a “‘ pineal foramen,” as it leads into the postfrontal sinuses and not into the brain case. It isa temporary opening, gen- erally closing in older individuals, and hence is analogous to the fontanelle in the skull of the humaninfant. A more accurate term would be postfrontal fontanelle, and as it was evidently not permanent the idea of its transmitting nerves and blood vessels seems hardly tenable. (See fig. 33.)—R. 8. L. THE SKULL. YD The lachrymals form a part of the lower portion of the anterior border of the orbits. This portion of the orbital border is much thinner and is less projected than that immediately above, which is formed by the prefrontals. In front of the orbit the lachrymal is continued forward between the prefrontal above and the jugal and maxillary below. The antero-inferior angle of the lachrymal is lodged between the nasal and the superior branch of the maxillary, and just at the union of these three elements there is in the type of Triceratops serratus a foramen which passes between rather than through any of these bones. The position of this foramen varies in the different species; in general it is comparable with that of the infraorbital foramen in the mammalia, and its function may have been similar in the two groups. The internal opening of this foramen is partially bounded by the palatine. THE NASALS: In the skulls of old individuals the nasals of opposite sides are firmly united by suture. In the types of Triceratops flabellatus and T. serratus, however, they are still separate. Along the median line of the skull the sutural border is very thick throughout the entire length of the nasals. The external surface of the nasals is regularly convex. Each nasal consists of a broadly expanded posterior portion, an elongated superior portion closely ap- plied to, and in old age coossified with, that of the opposite nasal, and a shorter inferior branch, as shown in fig. 19, A and B. The elongated superior branches of the nasals form the roof of the nasal passage anteriorly and support the nasal horn, which, however, has its origin in a distinct and single center of ossification that does not coossify with the nasals until late in life. At their anterior ex- tremities the rasals each send down- ward a short process which overlaps the superior border of the premaxillary. Fig. 19.—A, External view of right nasal of type of Triceratops flabellatus, : : No. 1821, Yale Museum; B, internal view of same. a, Inferior process; b, The postero-inferior branch of the nasal process beneath nasal horn overlapping premaxillary; c, surface for contact is shorter and more slender than the with premaxillary; 7, surface for lachrymal; /, surface for frontal; no, ante- superior. It is directed downward and Pee ere caitearios: of nasal horn core; n, surface for forward and overlaps the ascending pos- terior branch of the premaxillary, which is wedged in between this branch of the nasal and the maxillary. The free borders of the nasals form the superior, posterior, and most of the inferior borders of the anterior nares, which open laterally. Posteriorly the nasals are over- lapped by the anterior edges of the frontals, prefrontals, and lachrymals. The nasals pass far back under the anterior border of the frontals and prefrontals, but the surface covered over by the lachrymals is more limited. THE MAXILLARIES. The general outline of the maxillaries, as shown in fig. 22, is that of an irregular triangle, of which the inferior border forms somewhat the longer side. Posteriorly the maxillary is produced into a superior ascending branch and an inferior and horizontal branch. Anteriorly and superiorly the maxillary has an extended articulation with the premaxillary, and supero- posteriorly it is in contact with the jugal, the lachrymal and for a short distance with the nasal, as shown in fig. 10. Internally it has an extended union with the palatine and the lachrymal, as seen in fig. 20. The long inferior and posterior process of the maxillary is embraced at its extremity by the anterior of the two inferior processes of the pterygoid, and supports on its superior and external surface the flat and rather rudimentary transverse, which latter bone, as a“ eS a re 'goids, are reduced to little more than 26 THE CERATOPSIA. _ shown in fig. 21, is also in contact with the inferior process of the palatine for a short distance along its upper and inner margin. The external surface of the maxillary is perforated by a number of smaller and larger foramina, placed rather irregularly; and on the internal side, about 2 inches above the alveolar bor- der, there is a series of small foramina extending throughout the entire length of the dental series. These doubtless served for the transmission of nutrient blood vessels to the teeth.¢ The general form and appear- ance of the maxillary when de- tached from the surrounding ele- ments is well shown in fig. 22, drawn from the type of Triceratops fiabellatus. In this specimen only a few teeth protrude from the alveolar groove, and in this respect Fig. 20.—Internal view of maxillary, palatine, and lachrymal, with pterygoid and quad- the drawing gives a very Imper- rate drawn inin outline. From type of Triceratops flabellatus, No. 1821, Yale Museum. feet idea as to the appearance of 1, Lachrymal; #f, infraorbital foramen: Vv, vacuity between temporal fossa and mouth; Bre mx, maxillary; pl, palatine; x, foramen between pterygoid and palatine; pt, ptery- the dentition when complete and goid; a, surface for contact with basisphenoid process; g, quadrate; df, dental in position. The position of the foramina; ab, alveolar border. One-eighth natural size. i infraorbital foramen differs from that in the type of 7. serratus, where it is located between the superior border of the maxillary . and the inferior margin of the nasal. THE TRANSVERSE BONES. The transverse bones, or ectoptery- flattened ossicles on the superior and external surface of the posterior and inferior process of the maxillary, as shown at ¢r in figs. 21 and 24. Above they are in contact with the palatines and below, on the inner side, with the pterygoids. They are very rudimen- YZ Yi} tary, have no contact with the jugals, iy AL -\\ => ———— —— == SS ———— — = WS WSS SSS ——SS— = = \ SSS SSS LSS SS SSS == S== — = SSS ——o SSS . Marsh, No. 4342, U. S. Na- or arrangement of phalanges in the manus of the i ee Ceratopsia. From the nature of the metacar- ess; r, surface for proximal pals, however, there can be no reasonable doubt ae poll cr eae that the manus was mesaxonic in structure, and that the phalanges of the third digit were larger than those of the lateral digits, while it is quite possible, also, that the middle digit may have borne one phalanx more than digits II and IV and perhaps two more than did the first digit, though these are mere conjectures and have not as yet been con- firmed by direct observations. No manus even fairly complete has as yet been discovered. The pha- langes of the proximal series are of moderate length, with the longitu- dinal diameter exceeding the trans- verse. The intermediate phalanges are short, the transverse diameter exceeding the longitudinal. The unguals are broad and depressed. They are laterally expanded distally and constricted proximally. They Fia. 69.—Anterior view of metacarpals of T'riceratops serratus, No. 970, American exhibit a coarse cancellous struc- Museum of Natural Iistory. One-fourth natural size. Drawn by R. 8. Lull. ture, and during the life of the animal were evidently incased in a horny substance resembling hoofs. The form and character of the various phalanges are well shown in fig. 70. Fig. 68.—Radius of Triceratops serratus Marsh, No. 970 American Museum of Nat- ural History. d, Distal extremity; p, proximal ex- tremity. One-eighth nat- ural size. II : I FE ES EES ee Aga is alr A DM ie TR ET Te ee 62 THE CERATOPSIA. THE HIND LIMB AND FOOT. The hind limb and foot in T: riceratops is longer than the fore limb and somewhat more slender, though it also may be regarded as robust. THE FEMUR. The femur is half as large again as the tibia. The shaft is constricted and somewhat elliptical in cross section, with the transverse diameter the longer. It is much expanded at either extremity. Proximally the head is well differentiated from the shaft and the greater trochanter by a marked constriction or neck. It is directed upward and inward at an angle of about 45° to the longitudinal axis of the femur. The articular surface of the head is rugose and exhibits a convoluted structure, which is continued out upon the superior surface of the greater trochanter. The latter is expanded antero-posteriorly, inclosing externally a deep digital fossa. About midway between the head and the distal end of the femur there is a rudimentary fourth trochanter situated on the postero-internal border of the bone. The distal end of the femur is rather more expanded transversely than antero-posteriorly. The external condyle is larger than the internal. The intercondylar notch is very deep and narrow. The external condyle sends back- ward a considerable projection, which is narrow and constricted just opposite the bot- tom of the intercondylar notch. The prin- cipal characters of the femur are well shown in fig. 71 and Pls. XIV and XV. THE TIBIA. The tibia is short, much constricted medi- ally, but greatly expanded at either extrem- ity. At the distal extremity this expansion is almost entirely in a transverse direction, there being no marked antero-posterior ex- pansion. There is no well-defined internal malleolus on the inner side, but the external Cc D Fig. 70.—Phalanges of the manus of Triceratops serratus, No. 970, Ameri- A : can Museum of Natural History. A, B, Intermediate phalanx, border 1S developed into a very prominent viewed from the front and side; C, D, ungual phalanx, viewed from front and side. One-half natural size. Drawn by R. S. Lull. DEOCeSS; forming somewhat less than one-half the distal extremity of the tibia. This process falls a little below the inferior border of the astragalus and closely embraces that element on its external surface. It may be regarded as having performed the same functions as did the external malleolus or distal end of the fibula in the Mammalia. Its anterior surface was in contact with the distal extremity of the fibula. The principal characters of the tibia are shown in fig. 71 and Pl. XVI. THE FIBULA.a I have not observed this bone in the genus Triceratops or any of the Laramie forms of the Ceratopsia. It was doubtless present, however, though much reduced. In describing the genus Triceratops Marsh® says of it: ‘“‘The fibula is very slender and the distal end was closely applied a An admirably preserved fibula was found with the skull (No. 970, American Museum of Natural History) described by Lull as pertain- ing to Triceratops serratus Marsh. This bone, which is here figured (fig. 72), islong and very slender, with a subeylindrical shaft and flattened, expanded extremities. Proximally the bone is crushed so as to exaggerate the flattening to some extent. The face which was applied to the tibia is plane, while the outer surface is convex. The articular face is rounded and rugose. The distal articular extremity is much flattened and was evidently closely applied to the front of the tibia. The articular faceis concave and oblique with reference to the long axis of the shaft. A somewhat roughened area of considerable size, for muscular attachment, occurs on the outer surface of the shaft about one-third the length of the bone from the distal end.—R. S. L. i b Dinosaurs of North America, p. 214. THE HIND LIMB AND FOOT. 63 to the front of the tibia.’”’ This is probably correct, though I am unable to ascertain that this description was based upon an actual specimen of the bone in question. In Monoclonius, an earlier and smaller form, the fibula was present but rather slender, as shown in fig. 88. Fic. 71.—A, Anterior view of left femur of Triceratops prorsus Marsh, No. 4842, U. S. National Museum, in mounted skeleton: h, Head; t, greater trochanter; c, internal condyle. B, Front view of left tibia of same. No. 4842, U. S. National Museum, in mounted skeleton: c, Cnemial crest; f, surface for fibula; a. astragalus. C, Distal end of same: a, Astragalus. All one-eighth natural size. After Marsh. THE TARSUS. The astragalus is the only element of the tarsus known to the present writer, though others were doubtless present. It was closely applied and early became coossified with the tibia, covering over the internal two-thirds of the distal extremity of that bone, as shown in fig. 71 and Pl. XVI. I am unable to say anything further regarding the structure of the tarsus in the Ceratopsia, no elements other than the astragalus referable to it being known to me. THE METATARSUS. Three functional metatarsals were present in the pes of Triceratops. These were the second, third, and fourth. It is possible also that vestiges of the first and fifth were present, but as yet no pes has been discovered sufficiently complete to determine this point. The metatarsals 64 THE CERATOPSIA. are much longer and stronger than the metacarpals. They interlock with one another at their proximal ends and are closely applied throughout most of their length in order to give greater rigidity and strength to the foot. The median or third is the largest of the three, but the second and fourth are also quite well developed; their shafts are somewhat constricted, but they expand at the extremities. At the distal end the lateral expansion is greatest, while at the proximal end they are expanded most antero-posteriorly. Their form is well shown in fig. 73. P Fic. 72.—Fibula of Tri- Fic. 73.—a, b, c, External, posterior, and internal views of metatarsal of Triceratops prorsus, ceratops serratus, No. U.S. National Museum. d, e, f, Dorsal, lateral, and palmar views of ungual phalanx of the 970, American Muse- pes of Triceratops horridus, U. S. National Museum. All one-eighth natural size. After um of Natural His- Marsh. tory, anterior view. One-eighth natural size. Drawn by B.S. Lull. THE PHALANGES. The phalanges of the pes are very similar to those of digits II, III, and IV in the manus, though somewhat larger. The proximal phalanges are rather long, but the intermediate ones are very short and broad. The terminal phalanges are broad and rugose or cancellous. They were incased in flat hoofs instead of compressed claws as in the sauropod and theropod dinosaurs. It is probable that the third digit was provided with four or five phalanges, while the secon. and fourth digits would each have had one less than the third. THE CERATOPSIA. 65 THE EXOSKELETON. : Various spines and dermal plates have been found, more especially in the Laramie, associated with remains of the Ceratopsia and other dinosaurs. None of these have yet been found in such association as would dem- onstrate conclusively that they pertained to any member of the Ceratopsia, and nothing is posi- tively known as to the positions occupied by any of these ossifica- tions in the anatomy of the ani- mal. A few of these are shown here in fig. 74, but considering the limited knowledge we at present possess concerning them it seems scarcely worth while to speculate as to the positions they occupied. The asymmetri- cal spines, one of which is shown in fig. 74, 1, 2, 3, may have been arranged in pairs at the base of the tail, and it is not impossible that the plates shown in 4-10 and others similar to them may have been so embedded in the skin as to have formed a cuirass or armor about the throat and over certain regions of the back. The curious group of little ossi- cles coossified about a common base, shown in 11 and 12, were found associated with a consid- erable portion of a skeleton of a member of the Trachodontide, though not in such manner as to demonstrate conclusively that it pertained to that skeleton. No Ceratopsia remains were found Fic. 74.—1-3, Side, front, and top views of dermal spine; 4-7, top, bottom, side, and end views of dermal plate; 8-10, top, side, and bottom views of dermal plate. All the above figures are one-eighth natural size. 11-12, Side and front views of dermal ossifications found with portion of skeleton of Diclonius. One-half natural size. After Marsh. with it, however, and it would seem probable that it pertained to the associated skeleton were it not for the fact that no such ossifications were found associated with the two nearly complete skeletons of that trachodont collected by the writer in the same deposits. MON XLIx—07——5 CEA ae lee SYSTEMATIC DESCRIPTIONS OF GENERA AND SPECIES. METHOD OF TREATMENT. In the following systematic descriptions all the various genera and species that have at any time been referred to the Ceratopsia will be included, and all specimens that have ever been — used either as types of new genera and species or as pertaining to hitherto unknown portions of forms previously described will likewise be included, regardless of the present author’s opinion relative to the synonymy of the various generic and specific names or of the correct identifica- tion or reference of any of the elements that have at various times been described by different authorities from materials found in such isolated position as to render their correct reference uncertain or largely, at least, a matter of individual opinion. In all such instances the original descriptions and figures of the various authors will first be given in full, even at the risk of tediousness, and will be followed by a presentation of the present author’s views. In the diagnoses and descriptions of the various genera and species a reference will first be given to the original description of each. This will be followed by references to all later publi- cations by the original author in which a further elucidation of the characters of the genus or species under consideration may have been undertaken, based either on the subsequent discovery of new and more perfect material or on a more complete description of the original type. A reference will also be made to the more important literature by other authors. In the treatment of each genus and species, when possible, the parts forming the original type will be listed and definitely located, and the name of the museum to which each now belongs will be given, as well as the distinctive numbers which have been assigned to them in such museums, so that in each instance a permanent record may be available as to the character and location of all types which have served as the basis for new genera or species. In the descriptions of genera and species the original description of each will be given verbatim. This will be followed by such additional extracts from subsequent publications, either by the founder of the genus or species under discussion or by other authorities, as shall appear to the present author to be important. All original descriptions of type specimens will be given as quotations and all types accessible to the present author will be carefully figured regardless of his opinion as to synonymy, the aim being to put the reader in possession of all the facts upon which the opinions of the present and all previous contributors have been based, so that each student may be able intelligently to arrive at his own conclusions wherever differences of opinion may exist. While the distinctive characters of the various genera and species must of necessity be very largely based on the original types, yet wherever a genus or species is accepted as valid no oppor- tunity will be omitted further to elucidate its characters by the introduction and description of such additional material as may appear to the writer to be referable to the same, care always being taken, however, to state distinctly that such material does not pertain to the type and definitely to locate it by its appropriate number and a reference to that museum to which it belongs. . Although the very large collections brought together by the late Prof. Othniel Charles Marsh will serve as the basis of the present volume, free use will also be made of the material collected by the late Prof. Edward Drinker Cope, as well as of that brought together by the 66 ’ 4 _ DYSGANUS. 67 Canadian Geological Survey, chiefly through the efforts of the vertebrate paleontologist, Mr. Lawrence M. Lambe. The more recently collected materials of the American Museum of Natural History in New York, of the Carnegie Museum at Pittsburg, the Museum of the University of Chicago, and the Museum of the State University of Kansas have also been freely placed at the disposal of the writer. The earlier, smaller, more primitive, and less specialized mid-Cretaceous forms from the Judith River and Belly River beds will be first considered, followed by a description of the larger and more specialized forms from the later Laramie deposits of Converse County and Black Buttes, Wyoming, and from the Denver beds of Colorado, which latter have been consid- ered by some as of post-Laramie age. REVISION OF SPECIES OF JUDITH RIVER CERATOPSIA. GENERA AND SPECIES DESCRIBED BY PROFESSOR COPE. DYSGANUS Cope. 1876. Type species, D. encaustus. Original description in Proc. Acad. Nat. Sci. Phila., vol. 28, 1876, p. 250. Cope, E. D., Bull. U. S. Geol. and Geog. Surv. Terr., vol. 3, 1877, pp. 572, 596; Proc. Acad. Nat. Sci. Phila., 1883, p. 99; Am. Nat., vol. 24, 1890, p. 571. Nopesa, F. Baron, Féldtani Kézlény, Budapest, 1901, vol. 31, p. 270. Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, p. 20. This genus was proposed by Professor Cope in 1876, when he also gave brief descriptions of four supposedly new species (D. encaustus, haydenianus, bicarinatus, and peiganus) which he considered as pertaining to it. Neither in his original nor in any subsequent description did Cope state definitely what he considered the affinities of the representatives of this genus. Hay, in his Bibliography and Catalogue of the Fossil Vertebrata of North America, has placed this genus in the Trachodontidx, while Nopésa, in his Synopsis und Abstammung der Dinosaurier, has considered it as pertaining to the Ceratopside. Since the specimens upon which was based the type species of the genus and the types of any of the other species are no longer deter- minable@ in the Cope collection, we have now to rely entirely upon his descriptions for the deter- mination of those characters which might throw light upon the question of the relationship of this genus. Before discussing it further, therefore, it would seem advisable to quote in full Professor Cope’s original description of the genus and its several species. DysGaNnus ENCAUSTUS Cope. 1876. Type (No. 5739, American Museum of Natural History) consists of detached teeth. Original description in Proc. Acad. Nat. Sci. Phila., vol. 28, 1876, pp. 250-251. Cope, Bull. U.S. Geol. and Geog. Sury. Terr., vol. 3, 1877, p. 572. Nopesa, F. Baron, Féldtani Kézlény, Budapest, 1901, vol. 31, p. 270. Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, p. 14. _ Char. gen.—A large number of teeth exhibit the characters of this genus, which is a peculiar form of herbivorous Dino- sauria. The crowns are compressed so that the fore-and-aft diameter much exceeds the transverse. The body of the crown is a flattened shaft of dentine, one face of which is the denser and produces the cutting edge. This face is flat or weakly keeled, while there are two other faces uniting at an open angle, thus giving a subtriangular section. On each of these faces is adherent a shaft of cementum-like material of a dense character, whose external face is longitudinally concave. These inclose between them on the median line a deep groove, which expands below into a wide concavity, which appears to be enlarged as the age of the tooth increases preparatory to shedding. The other parts of the base of the crown below the cutting face are inclosed in a rather thick deposit of rugose cementum, which rises a distance on the sides of the tooth. The method of replacement of the teeth in this genus appears to resemble that of Cionodon, except that there is no indication of the existence of as many series in the transverse direction. The longitudinal grooves in the anterior and pos- terior cement columns are probably occupied by the borders of the apices of successional teeth. The presence of these columns, etc., distinguishes this genus from that and other allied genera. Char. specif —The cutting face is more or less concave and is impressed or sunken, its lateral borders and the cement of the basis projecting beyond it. The inferior border is also usually oblique, that of one of the sides rising diagonally. In the same proportion a weak keel is also unsymmetrically placed, lying close to the opposite border and dividing the face into a wide and a narrow concavity. The oblique border is also incurved, the edge of the posterior cement column curving round a All types of Ceratopsia in the American Museum of Natural History, except that of Agathawmas milo, have now been identified. —R.S. L. ‘4 Brewerdsne' Fe ee ry lt SP p Hight es lg _ . = bated Si iH set te te BIE I ec ae RR a Oe I ee 68 THE CERATOPSIA. the cutting face of the dentine. The latter is delicately rugose in unworn specimens. The external basal cementum rises highest on the incurved border of the crown; its surface is minutely rugose, the rugosity being generally punctiform. It is also of a different color from the dentine in the specimens as preserved, and is occasionally found nearly worn away. The edge of unworn teeth is not serrate. . Measurements. M. hength of basis of toothy: cas2 2c) Seles eee CII oer rte atlas tale rape) hee vey ae eR 0. 012 Diameter of crown, antero-posterior... -...---- pee elicdiel GAs | he abe eMeaamed ars hatte eee aan . 009 Diameter of:crown; trans verse=e'2: ae a.s2 fot ae Se as Pace re SOE ee ee ee . 004 Transverse diameter below ‘cro wane n:. ee sais say oy cee ot cet Ser eS SL age age RNR ()() The teeth are rather smaller than those of ae foulke. The borders present no indication of the crenation seen in that and other species, either in worn or unworn specimens. DYSGANUS HAYDENIANUS Cope. 1876. Type (No. 5738, American Museum of Natural History) consists of detached teeth. Original description in Proc. Acad. Nat. Sci. Phila., vol. 28, 1876, pp. 251-252. Cope, E. D., Bull. U.S. Geol. and Geog. Surv. Terr., vol. 3, 1877, pp. 572, 594, 596. Nopesa, F. Baron, Féldtani Kézlény, Budapest, 1901, vol. 31, p. 270. Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, p. 14. Represented by a number of teeth found in such relation that they are supposed to belong to two individuals. They differ materially in form from those of the D. encaustus and exceed any of them in size. The base of the tooth possesses the thick investment of rough cementum, and has a slope away from the base of the crown. The form of the crown is peculiar in possessing a lateral face placed at a strong angle to the usual face, and separated — from it by a strong protuberant angular ridge. This angular cutting face would resemble that of the Dicloni were it not ’ that the body of the dentine of which it is composed is a flat plate instead of a triangular segment of a subquadrate prism. Each face has a separate plate, which is separated from the other by a suture. A solid mass fills the angle between them, which is divided by a groove produced by the pressure of the angle of the face of the succeeding tooth which fits it. The wider of the “front’’ faces is divided by a low longitudinal ridge. Both of the faces are bounded by an external incurved ridge, which causes them to have a concave surface. A tooth of a size equal to that of the one just described, and found with it, has a form more nearly like that of D. encaustus in the less degree of prominence of the lateral angle. It displays but a single posterior cementum-like mass, which presents considerable lateral faces as well as a posterior one, as in the first-described tooth. Measurements. ma Tuength: of base of crowmem=s 22052 eye cise le ee ee ee ee eee 0. 010. Mlevations ob remaining party ot: Crowe ees eee ese ane eee ee . 006 Diameter;of crown, antero-posterior 2. 2 2s fet sel eee ae ree 015 Diameter-ofcrowny transverse. total’j..< ses cei eee oe eee ee eee ee ee .010 Diameter'of crown, transverse dentine, 3222 25/5 e ee sake ese ee eee ee eee 004 Dedicated to Dr. F. V. Hayden, U.S. Geologist. DYSGANUS BICARINATUS Cope. 1876. Type (No. 3975, American Museum of Natural History) consists of detached teeth. Original description in Proc. Acad. Nat. Sci. Phila., vol. 28, 1876, p. 252. Cope, E. D., Bull. U. S. Geol. and Geog. Surv. Terr., vol. 3, 1877, p. 572. Nopesa, F. Baron, Féldtani Kézlény, Budapest, 1901, vol. 31, p. 270. Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, p. 14. This dinosaurian is represented in the collections by some of the teeth of three individuals. Two of the teeth represent immature stages, while the others are worn by continued use. They all present characters not found in the D. encaustus. from which they differ in a direction the opposite of that which characterizes the D. haydenianus. The crowns present a nearly flat face without incurved lateral angles nor prominent median keel. The basis is wide, — projects in a rim beyond the face, and is invested with rough cementum. The face is peculiar in being divided into three planes by two low angular ridges, and its surface is smooth. The dentinal column is triangular, and there are two posterior columns separated by a fissure in mature teeth. The absence of the lateral incurved angle and the presence of the two median ones distinguish this species from the D. encaustus. Measurements. M. Toemp th, Of Wasis iyi) cesar: 3/5, e seve Sea ic Sa oie rap T I oe eee 5 eed eee EE 0. 009 Width. of basis js %.0 20 i252 SNe age ee oa Se aan ea a ane ee Pa O11 Length: ofsworn faces. .0 22:2 eSeers nb Sen eine seein Seats oe Ce eee ee nee ee eee ee . 006 Diameter of crown, antero-posterion == 22-5 =--2-552 5252-25525 2 5255255" wre 1g ie Re Te eer re O11 Diameter ‘of ‘crown; ‘transverse sass ec ie os 2s See ee Sere ae ee flere ae eee eee . 007 5 , j DYSGANUS. 69 DyYSGANUS PEIGANUS Cope. 1876. Type (No. 3974, American Museum of Natural History) consists of detached teeth. Original description in Proc. Acad. Nat. Sci. Phila., vol. 28, 1876, p. 252. Cope, E. D., Bull. U.S. Geol. and Geog. Surv. Terr., vol. 3, 1877, p. 572. Nopesa, F. Baron, Féldtani Kézlony, Budapest, 1901, vol. 31, p. 270. Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, p. 14. In the typical tooth of this species the form approaches the genus Palxoscincus Leidy, in the compression of the crown and the contraction of the base; it is a limital species of Dysganus if really properly placed in that genus. The widest portion of the crown is above the base; from this expansion it contracts in both directions, and in the unworn tooth forms an angular median apex. This is not the case in D. encaustus, which is regularly rounded. The margin of the crown is narrowed, expanding but little toward the expansion, and is quite rugose. From these rugosities low ridges descend on the face of the tooth, whose surface is also minutely rugose. The face is divided by a prominent median rib, which extends to the apex. No cementum is visible on the basis in the only specimen in which this part is preserved. Measurements. M ILemada OF CROWMa ssecsotd co bop es SHO SORE AS BEBO DOE Aa EOS Sees ES ana eee en na Rae cee ea 0. 008 Wiameters ofacrowmmtraAnsVeIse same mete en ceria ayalthe ec ioj acraciee aie soeesc es eee Ss soles 005 Wiameber om crown aantero-posterion, ral) DaSes= sees ee eae ae ese yee . 008 Wiamletertolicrownsanteko-posteniok woreatestesmr aa aerate ae eee tye ee oe a O11 From the above descriptions of this genus and the various species placed in it by Professor Cope it will readily appear that he included in it teeth pertaining both to the Ceratopsidee and to the Trachodontidex, while the type of D. peiganus should be considered from his descrip- tion as pertaining to some member of the Stegosauride, and perhaps not specifically different from the tooth figured by Lambe on page 57 of his Vertebrate Fauna of the Mid-Cretaceous of the Northwest Territory, and provisionally referred by him to Stereocephalus tutus. In attempting to determine the affinities of the genus, however, we should rely chiefly upon the characters presented by the types of the typical species, D. encaustus. Since, however, these types are no longer available, we have to rely upon their characters, as described by Cope. If, as Professor Cope has stated in his definition of the genus, in the teeth ‘‘the body of the crown is a flattened shaft of dentine” and “‘the crowns are compressed, so that the fore-and-aft diameter much exceeds the transverse,’ then these teeth can not pertain to the Ceratopsia, for in no member of that group known to the present writer have the teeth either the form or the propor- tionate diameters given by Cope. Neither am I acquainted with any species of the Ceratop- sid in which there are no crenulations on the margins of either the worn or unworn teeth, as is stated to be their condition by Cope in D. encaustus, the type species. Neither does the method of replacement of the teeth in any of the Ceratopside resemble that of Cionodon, as the latter has been described by Cope on page 448 of his Report on the Vertebrate Paleontology of Colorado, which is as follows: The teeth are rod-like, the upper portion subcylindric in section, with the inner face flattened from apex to base, while the lower half is flattened externally by an abrupt excavation to the middle for the accommodation of the crown of the suc- cessional tooth. The inner face of the tooth, from apex to base, is shielded by a plate of enamel, which is somewhat elevated at the margins and supports a keel in the middle, thus giving rise to two shallow longitudinal troughs. The remainder of the tooth is covered with a layer of some dense substance, possibly cementum, which overlaps the vanishing margins of the enamel. The outer inferior excavation of the shaft presents a median longitudinal groove to accommodate the keel of the closely appressed crown of the successional tooth. The apex of the tooth being obtusely wedge-shaped, the functional tooth is pushed downward and transversely toward the inner side of the jaw. The tooth slides downward in a closely fitting vertical groove of the outer alveolar wall. The inner wall is oblique, its section forming, with that of the outer, a V; it is furrowed with grooves similar and opposite to those of the cuter wall, but entirely disconnected from them. The base of the shank of the functional tooth, on being displaced by the successional, slides downward and inward along the groove of the inner side, each lateral movement being accompanied by a corresponding protrusion. At the most, three teeth form a transverse line, namely, one new apex external, one half-worn crown median, and the stump or basis of a shank on the inner. The new crowns are, however, protruded successively in series of three, in the longitudinal direction also. Thus, when an apex is freshly protruded, the shank in front of it is a little more prominent, and the third stands beyond the alveolar border. As each shank increases somewhat in diameter downward in the C. arctatus, the section increases in size with protrusion; hence, before the appearance of a new crown outside of it, there are but two new functional teeth in a cross row. Thus, in the outer longitu- dinal row, only every third tooth is in functional use at one time; in the middle series all are in use, while in the inner, every third one is simultaneously thrown out in the form of a minute stump of the shank, if not entirely ground up. ries 70 | THE CERATOPSIA. From the order in which he took up ane described the various genera of dinosaurs in his original description of this genus it is clear that Cope did not at that time regard Dysganus as - being closely related to Monoclonius, for in the text we find it separated from that genus by Diclonius and the three new species of that genus, then proposed by Cope. We might infer, therefore, that at that time Cope considered Dysganus as more closely related to Diclonius than to Monoclonius. Indeed, the only character mentioned by Professor Cope in his original description of the genus which in the present writer’s opinion might be considered as indicating — for this genus any relationships with the Ceratopsia is the indefinite statement that ‘‘the longi- tudinal grooves in the anterior and posterior cement columns are probably occupied by the borders of the apices of successional teeth,’ conditions similar to those which are known to obtain in the Ceratopsia. In a critical note published by Professor Cope in the American Nounelis! of June, 1890, page 571, he would seem, however, to have considered Dysganus as pertaining to the Ceratopside. For in referring to the ‘‘two-rooted teeth,” described by Marsh as peculiar to that family of dinosaurs, he remarks: The ‘‘two-rooted teeth,” described by Professor Marsh * * * are not such in point of fact. The appearance of two roots is produced by the absorption of the middle part of a single root by the crown of the successional young tooth. * * * Teeth of this kind were figured by Leidy as belonging to Trachodon, and were described by me as representing the new genus Dysganus. In his list of the species of the family Ceratopside (Agathaumide), published on pages 715-717 of the American Naturalist for August, 1889, Cope makes no mention of the genus Dysganus nor of any of its included species. Hay, i in his Bibliography and Catalogue of the Fossil Vertebrata of North America, remarks that it is a genus of uncertain affinities, though, as we have already remarked, he scat it in the Trachodontide, while Nopesa has, as we have pointed out, referred it without question to the Ceratopside. After a careful study of all the literature on the genus and its included species I am con- vinced that the genus Dysganus was based on teeth pertaining to two or more genera belonging in part to the Trachodontide and in part to the Ceratopside. In view of this and the absence of the type specimens? and the imperfect nature of the material upon which the genus was based, as well as the lack of any figures or sufficiently exact description of the particular teeth, considered by Cope as the type of the genus, to fix their true nature, I feel warranted in exoledine it from the recognizable genera of the Cera- topside. It shold), I believe, be considered as a nomen nudum. MONOCLONIUS Cope. 1876. Type species, I. crassus Cope. Original description in Proc. Acad. Nat. Sci. Phila., vol. 28, 1876, pp. 255-256. Cope, E. D., Bull. U.S. Geol. and Geog. Surv. Terr., vol. 3, 1877, p. 573; Proc. Acad. Nat. Sci. Phila., 1883, p. 99; Am. Nat- uralist, vol. 23, 1889, pp. 715-717 and 906, and vol. 26, 1892, p. 757; Am. Geol., vol. 8, 1891, p. 56. Baur, G., Science, vol. 17, 1891, pp. 216-217; Am. Naturalist, vol. 24, 1890, p. 570; Am. Naturalist, vol. 25, 1891, pp. 448, 450. Dana, J. D., Manual of Geology, 1895, p. 847. Hatcher, J. B., Am. Naturalist, vol. 30, 1896, p. 113. Lambe, L. M., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 9, 20, 21. Marsh, O. C., Am. Jour. Sci., 3d ser., vol. 41, Feb., 1891, p. 176; 3d ser., vol. 43, Jan., 1892, pp. 83-84; 3d ser., vol. 50, Dec., 1895, p. 497; Sixteenth Ann. Rept. U.S. Geol. Survey, pt. 1, 1896, pp. 217, 243. Nopesa, F. Baron, Féldtani Kézlény, Budapest, 1901, vol. 31, p. 270. Woodward, A. S., Outlines of Vertebrate Paleontology, p. 213. Zittel, K. A. yon, Text-book of Paleontology, English translation by C. R. Eastman, vol. 2, p. 245. The chief distinctive characters of this genus are its small size, the diminutive supenonaitem horn cores pointing directly upward; the short, broad, and saiels fenestrated parietals; the short squamosals, as indicated by the squamosal border on the parietals in the type. Fa aSee footnote on p. 67. | | MONOCLONIUS. “I — MONOCLONIUS CRASSUS Cope. 1876. Type (No. 3998, American Museum of Natural History) consists of teeth?, sacrum, anterior dorsals, and parietal. Original description in Proc. Acad. Nat Sci. Phila., vol. 28, 1876, pp. 255-256. Cope, E. D., Bull. U. S. Geol. and Geog. Surv. Terr., vol. 3, 1877, pp. 573, 594; Am. Naturalist, vol. 14, 1880, p. 511; ibid., vol. 20, pp. 153-154; vol. 23, 1889, pp. 715-717, 905. Lambe, L. M., Contr. to Canadian Pal., vol. 3 (quarto), pt. 2, 1902, p. 68. Nopesa, F. Baron, Féldtani Kézlény, Budapest, 1901, vol. 31, p. 270. Osborn, F. H., Contr. to Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 14, 20. ORIGINAL DESCRIPTION. Cope’s original definition of this genus and species was as follows: Char. gen.—Teeth with obliquely truncate face and distinct root, which is grooved for the successional tooth on the front. No external cementum layer, caudal vertebre biconcave, and brim narrow. Fore limbs large and massive. The teeth of this genus resemble those of Hadrosaurus, and like them are replaced from the “front,” an arrangement which precludes the possibility of more than one series of teeth being in functional use at one time. The robust fore limbs and elongate ilium distinguish Diclonius « from Hadrosaurus. From Trachodon it differs in the absence of the rough cementum layer on the back of the tooth. Char. specif—The faces of the teeth are acuminate oval in form and are divided by an elevated keel, which is median above, but turns to one side at the base. Margin crenate, the grooves extending more or less on the crown “back,” which is otherwise smooth. Sacrum with ten vertebrz, the last centrum much compressed, the diapophyses extending horizontally from the neural arch above, and connected by a vertical Jamina with the iliac supports; length, 27.33 inches. The bones of the limbs are robust, the hinder the longer, but not so much so as in some other genera. Length of femur, 22 inches; width, proximally, 7.4 inches; distally, 6 inches. Length of tibia, 20 inches; greatest diameter, proximally, 8 inches; distally, 7.25 inches. The three anterior dorsal vertebre are coossified, and the first exibits a deep cup for articulation with the preceding vertebra. The episternum is a T-shaped bone, thin and keeled on the median line below. Length of transverse portion, 21 inches. Subsequent discoveries have shown that the above description is erroneous in many par- ticulars. First, Cope’s description of the teeth seems to have been based, not on the teeth of Monoclonius, but on those of Trachodon, since he says that they resemble those of Hadrosaurus and are replaced in a similar manner from in “‘front,’’ while we now know that they differ from the teeth of both Hadrosaurus and Trachodon in that they are fixed in the jaw by two roots inserted transversely instead of by one root, as in both the latter genera, and that they are replaced from below in much the same manner as the deciduous teeth are replaced by the permanent set in the Mammalia; second, the statement that “‘the three anterior dorsal verte- brz are coossified’’ has been shown by subsequent discoveries to apply to the three [four]? anterior cervicals, Cope having mistaken the cervicals for dorsals; third, the T-shaped bone, described as an episternum, we now know to be the coossified parietals. Notwithstanding these errors, which, owing to the pioneer nature of the work and the rather fragmentary material upon which it was based, we may excuse, the size and form of the parietals, squamosals, and supraorbital horn cores and the characters of the sacrum and the relative proportions of the fore and hind limbs, as described by Cope, may be considered as fairly diagnostic of the genus and as placing it on a valid foundation, while little difficulty will be found in identifying the species (MV. crassus), either from the original and subsequent descrip- tions by Professor Cope or from actual comparisons of other material with that of the type, which latter method should always be the final course of arbitration in determining the synonymy of all closely related and apparently identical forms. LITERATURE. In 1886°¢ Cope described and figured as a probable sternal element the parietal of JMono- clonius crassus, and discussed at some length the structure of the sternum in the Dinosauria. Two years later,’ after Professor Marsh had described and figured the type of Ceratops montanus, Cope identified as a frontal horn core the element figured by him in 1877.° He also @ This is certainly a misprint. Cope undoubtedly meant to say Monoclonius. See also Proc. Acad. Nat. Sci. Phila., 1883, p. 100. b See footnote on p. 76. ¢Am. Naturalist, vol. 20, Feb., 1886, pp. 153-155. d Am. Naturalist, vol. 22, Dec., 1888, pp. 1108-1109. ¢ Pl. 34, fig. 8, Bull. U. S. Geol. and Geog. Surv. Terr. ig nig AiR ce Nh ea a a 72 THE CERATOPSIA. at that time (December, 1888) was inclined to the opinion that Ceratops Marsh was a synonym of Polyonax Cope, which latter had been founded in 1874 on material collected in the Laramie of Colorado. The following year,* in a note entitled “‘The horned Dinosauria of the Laramie,” which did not appear, however, until early in the year 1890 (hence the apparent conflict in dates), Professor Cope was enabled, through the publication by Professor Marsh of a paper (with fig- ures) entitled ‘‘The skull of the gigantic Ceratopside,’’? to determine definitely the homolo- gies of certain bones pertaining to the type of Monoclonius crassus and to define further that genus and species. In this paper Cope proposes the family name Agathaumide to take the place of Ceratopside, previously proposed by Marsh and based on the genus Ceratops, which Cope considered a synonym of some one of the previously established genera pertaining to the family. He concludes this paper by giving a list of the genera and species already known, adding three new species, Monoclonius recurvicorms, M. sphenocerus, and M. fissus, all of which he describes with more or less detail, unfortunately omitting, however, to include the various forms already described by Marsh. The additional characters then given as diagnostic of the present genus and species, Monoclonius crassus, are so important that I can not refrain from quoting him in full in this connection. After giving due credit to Professor Marsh for the assistance rendered by the latter’s article just referred to, Cope continues: The most complete skeleton in my collection is that of the Monoclonius crassus Cope. This includes representatives of all the elements except the bones of the feet. The posterior part of the skull is preserved, including the left frontal bone. This bears a horn over the middle of the orbit, of small dimensions, and with the apex antero-posteriorly compressed. : The parietal bones are enormously expanded and are interrupted on each side of the middle line by a huge foramen, which causes the remaining parts of the bone to resemble the corresponding parts of Chamzleo, depressed in a horizontal plane. The squa- mosals are lateral, and consist of a wide plate with convex external border with a slightly undulating outline. The ilium is remarkably elongate both anterior and posterior to the acetabulum, appropriate to the ten vertebrae which constitute the sacrum. It and the sacrum resemble very closely those of the Agathawmas sylvestre Cope, which fact, with the evidence derived from the other vertebrz, leaves no doubt that the Agathawmas is to be referred to the horned herbivorous Dinosauria, with Monoclonius and Polyonax. This family is calted by Marsh the Ceratopside, but as it is not certain that Ceratops Marsh is distinct from one of the genera previously named I shall call it the Agathaumide (or hellenice Agathaumantidz) from the longest known genus Agathaumas. In a later article entitled ‘‘Notes on the Dinosauria of the Laramie,”* Cope considers Ceratops Marsh a synonym of Monoclonius Cope, and distinguishes this genus from Polyonax Cope (Triceratops Marsh, according to Cope) by the relative lengths of the frontal and nasal horn cores, Monocloniwus being characterized by short frontal and elongate nasal horns and Polyonax by long frontal and comparatively short nasal horns. In this paper Cope also iden- tifies as cervicals the three coossified vertebra which he had described in 1876 as anterior dor- sals. He also here announces for the first time the presence of a postpubis, which had been supposed by Professor Marsh to be absent in these dinosaurs. DESCRIPTION OF TYPE SPECIMEN. With this brief historical review of the synonymy and original and subsequent descriptions of Monoclonius, I will proceed to a detailed description of the type specimen (No. 3998) as it now exists in the collections of the American Museum of Natural History, New York City. In Cope’s original description of this genus and species he mentions the teeth, caudal verte- bree, the three anterior dorsals (now known to be the cervicals), the fore limbs, the femur, tibia, ilium, sacrum, and the episterum (now known to be the parietals). Owing to the fact that Professor Cope failed to employ any definite marks or symbols for distinguishing type specimens, and the confusion existing in his collection prior to its removal from Philadelphia to New York, I am unable to identify all the material mentioned by him as pertaining to the type of the present species. In his remark that the most complete skele- aAm. Naturalist, vol. 23, Aug., 1889, pp. 715-717. bAm. Jour. Sci., vol. 38, Dec., 1889, pp. 501-506, Pl. XII. cAm. Naturalist, vol. 23, Oct., 1889, pp. 904-906. MONOCLONIUS CRASSUS. 13 ton in his collection included representatives of all the elements except the feet, I am convinced that he had in mind material pertaining to more than one individual. Moreover, it does not seem at all unlikely, but, on the other hand, from a study of the material it appears quite probable, that the actual type material described by Cope was of a composite nature and per- tained to two or more individuals. In the following detailed description of the type I shall include only such material as I can positively identify as having been associated by Cope with the type either in his original or subsequent descriptions: The skull_—Those parts of the skull preserved and at present identifiable as certainly asso- ciated by Cope with the type consist of the greater portion of the parietals (first described and later figured by Cope as the episternal) and the left frontal. Among the several squamo- sals in the Cope collection from the Judith River beds I am absolutely unable to identify that mentioned by Cope in the passage quoted above; neither am I able to determine positively that any squamosal in the collection pertains to this species. The parietals are preserved almost entire. Their general form is well shown in fig. 75. They are firmly united along the median line, and may be described as consisting of a straight, broad median portion, somewhat expanded in front and presenting throughout the an- terior three-fourths of its length a mark- edly convex superior surface supporting three low, rounded, rugose, median promi- nences appearing at intervals of about 75 millimeters. The inferior surface of this median bar is concave throughout the anterior one-half of its length, but plane posteriorly. The anterior extremity shows sutural markings, and its inferior portion is more expanded than the superior, both anteriorly and laterally, for articulation with the postfrontals, and perhaps also to some extent laterally with the squamosals. The specimen distinctly shows that the union between these two cranial elements, the parietals and postfrontals, was by overlap and underlap, like the shingles of a roof, the inferior border of the parietals SASS ; ; Fig. 75.—Superior view of parietal of type of Monoclonius cras8us, No. being overlapped by the superior border of 3998, American Museum of Natural History. sqs, Surface for articu- lation with squamosal; pfs, surface for postfrontal. One-eighth the frontals. Such an articulation would neat er permit a continuous growth of the skull to an almost indefinite age. Posteriorly this median bar expands laterally, giving rise on either side to two broad bars, which curve first outward, then forward, and at last inward toward the anterior lateral extremity of the median bar, with which, however, it probably never came in actual contact, although its present imperfect condition does not permit this character to be determined with certainty. These lateral parietal bars have an average width of about 10 cm., though they are decidedly more slender anteriorly than posteriorly. They describe about one-half the circumference of an ellipse, and, together with the median bar, they inclose on either side a large parietal foramen, elliptical in outline, perhaps not entirely inclosed anteriorly, with the longer diameter directed antero-posteriorly and having a dimension of about 315 mm., while the shorter transverse diameter is about 220 mm. in length. Posteriorly and medially the parietal border is broadly emarginate, and at a distance of 10 cm. on either side of the median line there is located the first of a series of undulations which continue at intervals all along the periphery of the parietal, decreasing in prominence anteriorly and becoming almost obsolete just before the squamosal suture commences. The squamosal suture occupies the external (anterior) border of the anterior portion of the lateral 74 THE CERATOPSIA. bar of the parietal. It is but 170 mm. in length, and indicates that the squamosals were short, not very broad, and rather inconspicuous as compared with the same elements in some other contemporaneous and later forms. Among the latter may be mentioned more especially Tri- ceratops. Cope’s description of the squamosals is somewhat indefinite and does not appear to have been based on any material pertaining to the type. The left frontal bone (No. 3997, American Museum of Natural History) mentioned by Cope does not, I am satisfied, belong to the same but toa smaller and younger individual than the pari- etals just described. This conclusion is borne out by the fact that Cope made no mention of this" remarkably characteristic element in his original description of the species, and only associated it with the type in his subsequent paper published thirteen years later. Not only is the border for contact with the parietal shorter than that in the parietals, but the color, texture, surface markings, and general conformation of the bone all indicate that it pertained to an individual distinct from that to which the parietals pertained. I believe that any characters it may present should not be considered as certainly diagnostic of the present species, for its specific . identity must, for the present at least, re- main uncertain. I describe and figure this element in this connection not out of re- gard for any certain additional characters it may furnish distinctive of the present genus and species, but rather for the infor- mation which it affords relative to the homologies of certain cranial elements in the Ceratopsia as a group. It may be de- scribed as consisting of both those elements which have been described and figured by Marsh? as the frontals and _ postfrontals, but without the slightest indication of a fronto-postfrontal suture, although the sutures for the parietal, squamosal, jugal, prefrontal, and opposing frontal appear in their proper positions on the different mar- gins of the bone. In general form it is tri- angular, with the apex of the triangle di- rected anteriorly. Posteriorly its superior i <<) ii i} | 1 1 FI} =f A Am ct : his ni \: 2 WNT. s ON SSS \ Ges Fic. 76.—A, Oblique front view of frontal and postirontal (No. 3997, American Museum of Natural History) of Monoclonius crassus?, with supraorbital horn core; B, superior view of same. a. Anterior end; Pp, posterior; h, supraorbital horn; 0, orbit. Both figures are one- fourth natural size. surface is broad, flat, and elevated. Ante- riorly it curves sharply downward and termi- nates in a narrow-pointed projection, which was interposed between the opposite frontal and the left prefrontal. Externally it is much deflected posteriorly, and medially it forms the superior two-thirds of the circumference of the orbit. . Immediately above the orbit, at the external border of the postfrontal, there rises a low horn core, broad below and pointed above, with the antero-posterior diameter considerably exceeding the transverse. The external lat- eral surface of this prominence is plane, the internal strongly convex, giving to the horn core the form of one-half of the apex of a cone that has been longitudinally bisected. The orbital border is thick and rugose.. The inner border throughout the anterior one-half of its length presents a suture for articulation with the frontal of the opposite side. Posteriorly, however, this border is somewhat emarginate, and instead of presenting a sutural margin it shows a thin, free edge, which formed the lateral border of an elongate foramen, homologous with that which Marsh has designated as the pineal foramen. This expands below into a large cavity, partially inclosed laterally by a vertical septum-but apparently communicating with the orbit by one or more small foramina. The posterior border presents an irregular margin for articu- aAm. Jour. Sci., 3d ser., vol. 44, 1891, pp. 167-168, Pl. I. Fd clef) Ee cee | ' ) ; ee ee ee ae ee ee eee eee ee ee a ee ee ee eee MONOCLONIUS CRASSUS. re) lation with the parietal. Externally this assumes the nature of a wide, deep pocket formed by the posterior extension of the superior and inferior walls of the bone. Into this pocket the lateral angle of the median bar of the parietal fitted. At about the middle of the posterior margin of the postfrontal for a very short distance it presents a free, thin border, indicating the presence of a small foramen opening into a large cavity beneath the surface of the post- frontal, which is here comparatively thin. The posterior margin of the deflected lateral portion of this bone presents in part a fractured surface and in part a suture for articulation with the squamosal. Inferiorly there is a short sutural surface for articulation with the jugal, though for the most part this is not shown. The usual articulation between this bone and the parietal, squamosal, and jugal seems to have been by overlap and underlap rather than by direct interlocking sutural contact. The antero-external border of the frontal is thin and presents a free edge, which appears to have overlapped the prefrontal. The sacrum.—This (No. 3998a, American Museum of Natural History) is composed. of 10 vertebre, if we include all those coossified with the true sacrals. If, however, we determine the number of sacrals by the number of sacral ribs there are but 8 sacrals, and of these only 4 Fic. 77.—A, Oblique inferior view of sacrum of type of Monoclonius crassus Cope (No. 3998a, American Museum of Natural History); B, oblique superior view of same. a, Anterior end; p, posterior; ab, acetabular bar; f, foramina between sacral ribs, centra, and acetab- ular bar. One-eighth natural size. or perhaps 5 had the sacral ribs united distally into a longitudinal bar, both giving support to the ilium and entering into the construction of the acetabulum. The general form of the sacrum is much obscured by distortion, due to crushing. The posterior sacral ribs are for the most part incomplete, and none of the neural spines are entire. Many of the diapophyses are also injured, but otherwise the sacrum is complete and in a good state of preservation for ¢ dinosaurian sacrum. The centra are all of medium length and somewhat constricted medially. Commencing with the fourth and terminating with the ninth, the inferior surface is broad and shows a median shallow groove somewhat more marked at the junctions of the several centra. Save the first and last all the centra have the vertical diameter exceeding the transverse, though these proportions have undoubtedly been materially affected by crushing. The first three centra are decidedly heavier than any of the succeeding ones, and the last is much com- pressed and proportionately stouter than those immediately preceding it. The first two sacrals do not bear ribs, although the first sacral rib springs jointly from the second and the third = _——s 76 THE CERATOPSIA. sacrals, and all the succeeding sacral ribs spring from the point of union of the succeeding centra except the last, which is given off directly from the middle of the centrum of the tenth sacral; while the next preceding springs jointly from the eighth and ninth centra, it is for the most part borne by the ninth. The diapophyses of the first and second vertebre are short and weak, much more so than in Triceratops, and are without inferior plates. Those of the suc- ceeding vertebre are longer, stouter, much expanded superiorly, while inferiorly they are united throughout their length with the corresponding sacral ribs by thin, bony plates, which in the region of the acetabulum give a firm support to the longitudinal bar mentioned above, formed by the expansion and union of the extremities of the sacral ribs of this region and constituting a portion of the wall of the acetabulum. This bar, together with the coalesced sacral ribs and diapophyses, incloses three foramina which are left open both above and below. Next to the skull the most important distinctive characters in the Ceratopside are to be found in the sacrum. It is unfortunate that as yet the best sacra have been found either isolated or at most associated with only fragmentary or indifferent skull material, so that for the most part it has been difficult to correlate the various forms of sacra and crania. The same may be said of the present specimen, for although Cope in his description has referred it to the same skeleton with the parietals and postfrontal described above, there would seem from such general characters as size, color, degree yf | Dr of petrifaction, etc., to be little doubt that all Lol Yi three pertained to as many different individuals. if fu A Moreover, associated with them in the same lot LEZ ‘at ya de Po, > (designated Lot II by Cope) are representatives of many different individuals pertaining not only to various species, but to genera, families, orders, and even classes of reptiles, and almost without exception unaccompanied by any distinctive marks or labels giving data as to their association or as to the geological horizon or geographical position in which they were found, other than the Judith River beds of Montana. The cervicals.*—Only the three [four]? an- Il Iv terior cervicals are known. The three coalesced Fie. 78.—Atlas, axis, and third and fourth cervical of type of vertebrse described by Cope as anterior dorsals Monoclonius crassus Cope (No. 3998, American Museum of = 2 F F Natural History), as seen from left side. a, Anteriorend; are IM reality the atlas, axis, and third [and OE ake oe Sane (Drawing altered fourth] cervical. All of these [except the atlas] bore double-headed cervical ribs. They are firmly coossified by their centra. The atlas [and axis] has the lateral and vertical diameters nearly equal, it bears no spine, and at its anterior extremity there is a deep cup for the reception of the condyle of the skull. The vertical diameter of the second [third] and third [fourth] cervicals exceeds the transverse, although in the present specimen the proportions have evidently been much altered by crushing. Both these vertebrae bear spines, and that of the axis [third cervical] is much extended antero-posteriorly. These spines are closely applied to one another, though not coossified, save at the base, where they are separated by a large circular foramen 28 millimeters in Thanneter The dorsals.—There are in the collection a number of Agee all incomplete and evidently found isolated, and there is at present no record available as to their associations. In size a The coalesced anterior cervicals of Monoclonius crassus, asin Triceratops, undoubtedly consist of four bones, the atlas being reduced to a narrow ring-like element, the line of demarcation being indicated by a series of pit-like depressions ranging from the inferior surface of the bone upward and fading out a little beyond the mid-distance to the summit. The fore and aft diameter of the atlas as thus indicated is much greater inferiorly. b Numbers in brackets added by R. S. Lull. c Fig. 78 as prepared under Mr. Hatcher’s direction failed to indicate these depressions. The drawing has been subsequently corrected to agree with the specimen. The axis, Hatcher’s atlas, is rendered thus somewhat less in anterior extent, though its posterior limitations are as before. The anterior margin of the spine, which rises abruptly above the suture between the axis and the third cervical, extends forward in a gentle curve fading out on the dorsal portion of the axis. The axis also bears two distinct rib facets as indicated in the figure. Herein it differs from that of Triceratops, in which no ribs are borne in the axis. (Compare fig. 50, p. 47.)—R. S. L. | MONOCLONIUS CRASSUS. 77 and general characters they agree well with what we might expect to obtain in MV. crassus, and some of them may pertain to the same skeleton as the type. The centra of all are short, somewhat constricted medially, with nearly plane or slightly biconcave extremities. The neural arches are of moderate height and somewhat constricted just above the neural canal, which is of only moderate di- mensions. Theneural spines are of moderate length and compressed. The transverse processes are triangular in cross section. There is a tubercular rib facet at the extremity of each transverse process and a capitular facet situated on the side of the centrum in the anterior dor- sals, on the neural arch in the median dorsals, and on the inferior side of the trans- verse processes in the poste- rior dorsals. The anterior and poste- rior zygapophyses of Oppo-_ Fic. 79.—A, Anterior view of posterior median dorsal of type of Monoclonius crassus Cope, No. site sides are distinct, though 3998, American Ai seumr Natural History, B, side view of same. c, Capitular rib facet; x 3 az, anterior zygapophysis; pz, posterior zygapophysis. Both one-fourth natural size. not widely separated in the vertebre of the anterior and middorsal regions. In the posterior dorsals, however, they are confluent. : The caudals.—In the collection there is a single caudal, which, though not the first of the series, I refer to the anterior caudals. It is biconcave, with a small transverse process springing from near the middle of the side of the centrum. The centrum is short and constricted medially below the transverse processes. The neural arch is low and the zygapophyses of opposite sides well separated. The spine is want- ing in the present specimen. The pelvis.—In his original description Fig. 81.—A, Anterior view of an anterior caudal of type Fic. 80.—A, Anterior view of anterior dorsal of type of Monoclonius crassus of Monoclonius crassus Cope, No. 3998, American Mu- Cope, No. 3998, American Museum of Natural History; B, side view of seum of Natural Ilistory; B, side view of same. az, same. az, Anterior zygapophysis; pz, posterior zygapophysis; c, capit- Anterior zygapophysis; pz, posterior zygapophysis. ular rib facet. Both figures one-fourth natural size. Both figures one-fourth natural size. of the genus and species Cope simply mentions the ilium as being elongate. There are in the collections from the Judith River beds two ilia. The larger of these is nearly complete, and although a little small in comparison with the sacrum described above, it corresponds very well in character and state of preservation with the parietals; hence I have referred it to Per ee u ina ee ty OO ———_—_—_———— ee 78 THE CERATOPSIA. the same skeleton, though with a query. It differs little from the same bone in Agathaumas, except in its smaller size. It is proportionally longer and more slender in the present genus Vie. 82.—A, Inferior view of right ilium (No. 3998? American Museum of Natural History) of Monoclonius crassus Cope; B, superior view of same. a, Anterior extremity; p, posterior; is, ischiac peduncle; pb, pubic peduncle; ss, surface for contact with sacrum. One-eighth natural size. and has the deflected external margin posterior to the ischiac peduncle more marked and pro- duced into a somewhat angular prominence. The extreme ante- rior end of the blade is wanting, but otherwise the bone is essen- tially complete. Superior and inferior views of this illum are given here in fig. 82. I have been unable to detect any portion of either pubis among the collections from the Judith River beds. There is a left ischium nearly complete, but lacking the extremity of the anterior process with the articular surface for the pubis and the extreme distal end of the shaft. This agrees very well in size and general characters with the other Fig. 83.—External view of ischium of Mono- Material that has been referred f clonius crassus Cope, No. 3998, American by Cope to M. crassus. I there- i Museum of Natural History. il, Proc- - nies A ess for contact with ilium; p, process fore describe and figure it in this for contact with pubis. One-eighth connection. In general form the natural size. M p i 5 Fe ¥ ischium in the Ceratopsia is not Fic. 84.—External view of right scapula very unlike a rib, and after a casual glance it might be mistaken and coracoid of Monoclonius crassus for one of those elements. More careful study, however, will at cae Pee eee cena once reveal its identity. Proximally it presents a moderately C, coracoid: g, glenoid cavity; d, ru- gosity for attachment of deltoid Bis expanded subcircular articulation for contact with the ischiac } one 4 ahyteh muscle; sc, scapulo-coracoid suture. f peduncle of the ium. Just below this it sends forward a rather One-eighth natural size. . slender process for articulation with the pubis. The extremity of this is wanting in the present specimen. The shaft of the ischium is subcircular in cross PRS —

Greatest length jot masal horncore: ae ame ee ee ee en acta eR eee 302 Mransverse diamerersonsMornlcore) ati ase bey ee ee eae eee ee ee 80 i Antero-posterior diameter of horn core at base._....---------- i SRC A? Nee oh yang sae NY tee te 5 Gentes 160 Height of apex of horn above premaxillary suture....-....-..--.-----------------------+-------- 465 GENERA AND SPECIES OF CERATOPSIA FROM THE JUDITH RIVER BEDS OF CANADA, DESCRIBED BY LAMBE. Having concluded this review of the different species of Monoclonius that have been proposed by Cope upon remains of these dinosaurs from the Judith River beds of northern Montana, I will next consider the forms described by Prof. H. F. Osborn and Mr. Lawrence M. Lambe and based, for the most part at least, on material collected by the latter in the Belly River beds of the Northwest Territory, Canada. MONOCLONIUS DAWSONT. 89 The treatment of this material by the above-mentioned authors forms part 2 of volume 3 (quarto) of Contributions to Canadian Paleontology, published by the Geological Survey of Canada and entitled “On Vertebrata of the Mid-Cretaceous of the Northwest Territory.”’ This memoir, consisting of 81 pages of descriptive matter, with 21 plates and numerous text figures, is divided into two parts. The first of these, entitled ‘‘ Distinctive Characters of the Mid-Cretaceous Fauna,” pages 7—21,is by Professor Osborn, and is devoted largely to a discussion of the relations of the vertebrate fauna of the supposed Belly River beds to those of the Judith River beds of Montana and the Laramie of Converse County and central southern Wyoming, and of the Denver beds in Colorado. The second part of this memoir, ‘‘New Genera and Species from the Belly River Series (Mid-Cretaceous),” is by Mr. Lambe, and is for the most part purely descriptive. Three new species of Ceratopside are recognized and described as pertaining to the genus Monoclonius. These are Monoclonius dawsoni, canadensis, and belli. A new genus and species, Stegoceras validus, is also described, and erroneously referred to the Ceratopside. The species described by Lambe will now be considered. MonocLonius DAWSONI Lambe. 1902. Type (Nos. 1173 and 971,¢ Geol. Sury. Canada) consists of parts of two skulls from Red Deer River, Canada. Original description in Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 57-63. Lambe, L. M., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 68, 75, 76, 81. Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 14, 20. The type of the present species was collected by Mr. Lambe in 1901, in the Judith River beds on Red Deer River be- tween Berry Creek and Dead Lodge Canyon, Alberta, no more exact locality or horizon being given by that author.’ Mr. Lambe’s description is as fol- lows: The remains of an apparently unde- scribed species of this genus, consisting of the skull of one individual and the pos- terior crest of another, are of especial interest. The skull when found lay on its right side, and although very much crushed, certain parts of it supply defi- TG: nite information as to its structure and size. The two orbits, the right maxilla, Fic. 92.—Head of Monoclonius dawsoni, from a sketch in the field. One-sixteenth natural ap d tk eral aks size. p, Parietal; or, orbit: oc, occipital condyle; g, quadrate; m, maxilla (inner side Bh Glee rahe and the occipita con Yes showing a row of foramina); h, nasal horn core; s, squamosal. After Lambe. were conspicuous and apparently in place, with a large posterior crest extending to the rear. Somewhat in advance of the orbits a horn core, of large size and appar- ently symmetrical form, occupied a position suggestive of a nasal origin, the nasal bones and the frontals being probably represented by the fragments filling the space between the orbit and the horn core (see fig. 14 [fig. 92], from a measured drawing made before the parts of the skull were removed). The large posterior crest forms the back part of the skull above. Its exact shape is fortunately supplied by the admir- ably preserved specimen shown in outline, from beneath, in fig. 15 [fig. 93]. The surface of the bone, above the orbit and from there inward for a short distance toward the median line of the skull, is moderately smooth and shows no trace of a horn core. With the separate posterior crest was found a horn core, similar in shape to, although not as well preserved as, the one belonging to the skull. The posterior crest is composed of the parietals and squamosals coalesced. The former are represented by a flat, thin, smooth median portion that expands laterally both in front and behind. Anteriorly it is deeply concave below and broadly convex above, but posteriorly it thickens gradually, and dividing to either side is continued forward as the squamosals in a broad curve to meet the anterior expansion. On either side of the median element is included a large supratemporal vacuity or fontanelle. a This specimen, No. 971, here figured (fig. 93) and described has since been made the type of a new genus and species. See footnote, p. 93.—R. 8. L. b See p. 179. en et a my i < Lyncrs = ee Lae eee eB s mS OND f wv OO — 90 0 THE CERATOPSIA. a The posterior crest is somewhat saddle-shaped. Its sides are wavy, with a slight thickening of the bone in the posterior five of the seven corresponding conyexities of the periphery, while a pair of inwardly directed spurs of bone, with their points turned slightly downward, are developed on its posterior border, one on each side of the median line. The intervacuital ‘element, is thickened along its median length and a more decided strengthening of the bone occurs along the posterior border, resembling in this respect the corresponding part of Monoclonius belli, described farther on. In all other parts of the crest the bone is thin, more particularly near and at the margin of the fontanelles, whilst along the sinuous curves of the sides the edge is sharp, except in the emarginations, where it is rounded. Vascular markings occur on both surfaces, more particularly on the peripheral projections. Measurements. Mm. Height of orbit ews) e saps eo ape a reed ee oe ee poe ae eae 0. 110 Widthvof orbit.2.522 223. 2258 2 so cee Sa a tee dein eee eee ie Se .095 IBIS Teel Che NON COR oe wakes coSe as oheseoonSs so sae meadseces cose ecsscomandedsacsescesasede ts . 331 Circumberencelatbasevolstornh Cole Sere = es ee tate nee eee euler Slee . 343 Antero-posteriors diame ter oly loase tof cena eae ete ete ley pele ae ae eee 5 11835) Transverse diameter of paserofisarmne® ham cee sere tegee yey cyar ete ce Vere oars ae eg ese . 092 Diameter’ of ocerpitalcondyle sys. scant Shee oo as essere ey set ete fens coe oe a . 060 beara Or medi, AyojOrsM. 612k 34525546 os saoscd as sadozaen ss ae caso se sone seco Sseossi5: . 350 JEG, OF KEWAS, KYONKOPTINNs Asoo bs ye ssa Snsescanasssao en $uTe dooaesbs sacs sssnassossese . 120 luong diameter of lower face on quardialies === ee seen eae eee ae Je O80! Short diameter of lower face of quadrate.......-..---------- fittest RRS et 3) SOR semen 036 The skull and the posterior crest were collected on Red Deer River in 1901. With this species are provisionally associated a scapula and coracoid, a sacrum, an ilium, a rostral bone, and a predentary bone, described or referred to in the next succeeding pages. The scapula with coracoid is figured on Pl. XIX, fig. 4 [Pl]. XX, fig. 1], viewed from its inner side. The scapula is long and narrow, slightly concave inward in the direc- tion of its length, stout below, thin- ning rapidly upward, upper end ter- minating squarely, breadth decreas- ing toward mid-length, slightly ex- panded above, front margin thin, back margin broad below, narrowing to its mid-length, then continuing thin upward. A rounded ridge ex- tends upward, on the outer surface, diagonally across from the upper end of the glenoid cavity to the front margin, continuing as a decided thick- ening of the front margin above. The coracoid is broader than high, emarginated below the glenoid cavity and produced backward below, lower border turned inward, inner surface . decidedly concave, back border at F1a. 93.—Posterior crest ¢ referred by Lambe to Monoclonius dawsoni, probably pertaining toa emargination thick, border elsewhere distinct species; viewed from beneath. Slightly less than one-eighth natural size. The num- rather thin, rounded. Foramen tray- bers give the thickness of the bone, in centimeters, at the points indicated. P, Parietal; S, 2 3 =< squamosal; F, fontanelle. After Lambe. DENS thickness of upper part di- rected obliquely downward and out- ward with an enlarged outer opening. A small foramen occurs, below the glenoid cavity, in the emargination of the posterior border. Glenoid cavity higher than broad, its curve forming almost a semicircle. In the specimen figured the coracoid was probably firmly united with the scapula, the suture between them, extending from the mid-height of the glenoid cavity forward, being only slightly indicated. The union of the two bones may be regarded as an evidence of age in the individual. The left scapula and coracoid from the Red Deer River district, so similar, in most respects, to that of Triceratops prorsus Marsh, as figured in Sixteenth Report of the United States Geological Survey, differs in one important particular, viz, in having the lower border of the coracoid turned inward instead of outward. a This is the type of Centrosaurus apertus Lambe. See Pl. XXIV and footnote on p. 93. ee Eee MONOCLONIUS DAWSONI. Measurements of scapula and coracoid. Scapula with coracoid (left). Cat. No. 506: Mm. Extreme length of scapula with coracoid in line with back edge of shaft-----__- ee eseetee ONS879 Length of scapula-.----.-.------ Ete ee ke See eet bc a oe tee See ee ee 711 LUG HEFAN XOROSS Sate NON! GOAT Pace SAS ese A Se ae ee . 150 Wencibrofelenoidecaviltyralongecuivess=== sass es a eee es eo eee oe ee ee ee . 204 Breadth of glenoid cavity at suture hetween scapula and coracoid.-_-.-_---.--------------- .078 Bresaibeowmelenodl cay byanedre1tnek ends eee se sa ee aa ae ee oe ee ee ease . 096 Breadth of scapula at junction with coracoid, inner surface.........-.-------.------------- Ble Breadth of scapula at junction with coracoid, outer surface.........-.-.------------------- . 149 Breadth ofscapularaimuppemendon clenoidicavitiy.ss===- 2526-2 25) - asso ee . 238 Ipread untomrscapu lays Galen e thee meee cei ee eos ne oe Ree en LE ee plits IBywemelils Ol Sceya by Dit Why ner Cua ee ae ee ee a ee ee ee ee ee . 184 Breadthvof coracoid'at lower end. of glenoid cavity_.- 2. ..222-..-.-.2.-.----------+-2+------ -223 ihicknessrotiseapula atjupper endimear tront) borders 2------ = 2225-5 5255-22--- 2-555 --- . 025 Thickness on base of ridge above upper end of glenoid cavity... -.-.----------------------- . 060 eliiticknesstatiowenendeoh plenoidicavitye 5-1-5 345 a= Jo ee eee Ries 2 . 060 Mhickness‘of coracoid in concavity, below, toramens—- -=+----- 5. =-2-24=---2+2----5--245---- . 020, Wridtuhtofstorame»nsinnewend ape erirteecwer pt eee by eee ey ie se es Sie ee .014 leith trofisamemnnenten deen re Pee ee aI ee FAS oP Ng a ee Se cs . 030 Widuhtofisamemoutenend: sseererter ernst ery oh ee eo Se eet es shee 025 Eeishtrofesame s0uteren dienes emma eee fase eas Were ws ne ed ee corse eeoee cele 040 91 The rostral bone figured on Pl. XX and the small predentary bone (Pl. XIX, figs. 5 and 6) [Pl]. XX, figs. 2 and 3] were found separately, and may with some probability of correctness be referred to this species. A large ilium is figured toward the end of this report. This species is named in honor of Dr. George M. Dawson, C. M. G., late director of the Geological Survey of Canada. From the above description it will readily appear that the fragmentary skull (No. 1173, Geol. Surv. Canada) must be considered as the actual type of the present species, on which alone we must depend for the determination of its distinctive characters. The posterior crest and associated horn core, as well as the scapula and coracoid, ilium, rostral, and predentary bones provisionally associated with the type may have pertained to the same species, but of this we can not be certain, and since they were all found disassociated they may possibly have pertained to one or more different species. However this may prove to be, any characters which this collateral material may show can not be taken as certainly diagnostic of the species until such time as the dis- covery of additional and more complete material shall conclusively demonstrate that these particular types of parietal crest, scapula, etc., were associated with that pattern of skull shown in the actual type. The type (No. 1173, Geol. Sury. Canada) consists of a very fragmentary skull with large and somewhat compressed nasal horn core curved backward, ovate in Pia. 94.— Trachodon sacrum provisionally associated with cross section, with the broad end of the oval directed Monoclonius dawsoni by Lambe; superior view, less anteriorly. The orbit is nearly circular, and there tary supraorbital horn core. This rudimentary horn than one-seventh nutural size. nc, Neural canal; d, E : 5 : diapophysis; np, neural platform; ns, neural spine; still remains in the type the base of a small rudimen- i, interspace. After Lambe. core is flattened on its external surface as in the cotype of Monoclonius crassus Cope. About 1 inch of the apex of this horn is wanting in the present specimen. is preserved showing three or four of the marginal undulations. A fragment of the parietal These resemble those seen in the type of Monoclonius crassus Cope, and I am inclined to regard the present species as Pa aes TS ge a 92 THE CERATOPSIA. closely allied to if not identical with the M. crassus of Cope. The nasal horn and orbit are very large when compared with the occipital condyle, the maxillaries, and quadrate. The peculiar parietal (see footnote b, p. 93) associated by Lambe with the type of MW. dawsoni I regard as pertaining to a distinct species and perhaps also to a distinct genus. . The median parietal bar is very heavy and deeply emarginate posteriorly, where on either side it gives off an elongated process which is pointed and curves inward and slightly downward. The bases of these processes are separated from one another by a distance of 300 mm. They are each 109 mm. in length, have a breadth of 112 mm. at the base, and a thickness of 30 mm. They are acutely pointed, have rugose surfaces with deep vascular grooves, and in life were evidently insheathed in horn. The posterior border of the parietal between the bases of these processes is very thick and round. On the superior surface, in the middle just in front of the posterior border, there is a rather deep concavity. The median bar of the parietal is very thick throughout its entire length. Its superior surface is marked by a number of very gentle elevations; it is rugose and shows shallow vascular impressions. Beside the processes already mentioned as present on the posterior portion of the parietals, there is on either side a series of seven emarginations on the parietal borders separated by as many prominences. The Fic. 95.—The same sacrum as in fig. 94; right lateral aspect, about one-seventh natural size. a, a, Facet for ilium; d, diapophysis; np, neu- ral platform; ns, neural spine. After Lambe. posterior of these prominences are the larger, and each bears evidence of having supported a distinct epoccipital bone during the life of the animal, save perhaps the two anterior, which appear to have been overlapped by the squamosal, since the parietal in this region presents a rather distinct sutural surface for contact with the squamosal. The parietals on either side of the median bar present a large fontanelle, and the margin of bone inclosing this is very thin. These fontanelles have a length of 292 mm. and a breadth of 254 mm. The parietal bar is concave on its inferior surface, especially anteriorly. At. its anterior extremity it presents a — number of cavities and articulates with the postfrontals. ¢ The sacrum described and figured by Lambe and provisionally associated with the present species does not belong to the Ceratopsia. It is clearly the sacrum of one of the Ornithopoda and belongs without doubt to some species of Trachodon. The nasal horn (No. 190, Geol. Surv. Canada) figured by Lambe and reproduced here in Pl. XVIII, figs. 1 and 2, though referred by him with a query to M. dawsoni, resembles more nearly the same element in M. recurvicornis Cope. It differs from that element in I. dawsoni by curving forward as in M. recurvicornis instead of backward as in the type of M. dawsoni. a See fig. 93 and Pl. XXIV. MONOCLONIUS CANADENSIS. 93 It is extremely massive and only moderately compressed. It curves rather more strongly forward than does the nasal horn of J. dawsoni in the opposite direction or backward. One half of a horn core, split longitudinally* and found with the peculiar parietal (No. 971, Geol. Surv. Canada) referred by Lambe, though I think erroneously, to IM. dawsoni, appears to represent one half of a nasal horn of the type described by Cope as M. sphenocerus. Owing to the imperfect nature of this specimen it is not possible to determine positively whether it represents a nasal or a supraorbital horn core. It seems more than probable that _the type of parietal shown in fig. 91 was associated with a nasal horn similar to that of JZ. sphenocerus. If this horn core should prove to be a supraorbital the evidence would be equally in favor of excluding it and the associated parietals from the present species in which the supra- orbital horn core is very rudimentary. It seems quite probable that this peculiar parietal may pertain to MW. sphenocerus. From the nature of the associated horn core it can not pertain to M. dawsom.° MOoNOCLONIUS CANADENSIS Lambe. 1902. Type (No. 1254a, b, ¢c, d, e, Geol. Surv. Canada) consists of anterior dorsal and fragments of skull. Original description in Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 63-66. Lambe, L. M., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 68, 81; Ottawa Naturalist, vol. 18, pp. 81-84; Trans. Roy. Soc. Canada, 2d series, vol. 10, sec. 4, pp. 3-9. Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 14, 20. a When Mr. Hatcher saw this horn core it was partially embedded in the matrix; on being entirely removed from the latter, the bone was found to be complete and not split longitudinally, as both Hatcher and Lambe supposed. (See Pl. XXIV, fig. 3.)—R. S. L. b CENTROSAURUS APERTUS Lambe. 1904. Type No. 971, Canadian Geological Survey, consisting of a parietal crest and nasal horn core. Original description in Ottawa Naturalist, vol. 18, July, 1904, pp. 81-84. Lambe, L. M., Trans. Royal Soc. Canada, 2d ser., vol. 10, 1904, sec. 4, pp. 3-9. The material upon which this genus and species is based was originally referred to Monoclonius dawsoni and was described and figured as such by Lambe in Contributions to Canadian Paleontology, vol. 3, pt. 2, 1902, pp. 58-59. The type was found in the Belly River (Judith River) series of Red Deer River, Alberta. The author’s original descriptionis as follows: “The crest of C. apertus is composed principally of the coalesced parietals which form an expansion having somewhat the shape of a saddle, broader than long and much more robust posteriorly than in front where the bone is decidedly thin. The squamosal is not known. The parietal part of the expansion, figs. 1 and 2, Pl. I [Pl]. XXIV], is longitudinally ridged in the median line, is broadly expanded laterally on either side, and ends posteriorly in a robust transverse bar that is concave in outline behind as viewed from above. The fontanelles occur one on either side of the median line and are of large size; they are bounded behind by the transverse bar that forms the posterior border of the crest, and laterally and in front by the thin side extensions. Along the median line the bone is transversely concave beneath. The separate ossifications, named by Marsh epoccipitals, are well developed in four pairs, with, in addition, the pair of hooked processes, already mentioned, which are regarded as specially developed epoccipitals. The alar extensions are referred to in the original description as the squamosal por- tion of the crest, the squamosals being then regarded as having coalesced with the parietals. Near the anterior border of the right extension, however, there is a definite line of demarcation, a, figs. 1 and 2, which can be considered only as the suture for the squamosal. * * * The postfrontal suture, b, figs. 1 and 2, extends from the inner side of the anterior end of the fontanelle obliquely forward and inward to the median line in front. Numerous impressions of blood vessels are present on and in the neighborhood of the epoccipitals and hooked processes and on the upper surface along the median ridge. “The horn core, fig. 3, found with the crest is presumably a nasalone. It is straight and laterally compressed so as to be lenticular in cross section, presenting a sharp edge to the front and rear. A somewhat similarly shaped nasal horn core has been described by Cope under the name Monoclonius sphenocerus. One side, that figured, is deeply channeled longitudinally; the other is more regularly convex; vascular markings are conspicuous on both sides. There is apparently no great distortion, if any, of the specimen, which is 380 centimeters long and imperfect at the tip and below. “We may conclude from the above that Centrosaurus apertus had a broadly expanded squamoso-parietal crest composed mainly of the coalesced parietals, the squamosals being confined to the antero-lateral edge of, and taking but little part in the formation of, the frill; that the large oval fontanelles were included entirely within the parietal part of the expansion and that the epoccipital bones were well developed, of which the hinder pair were greatly modified so as to form large hooks or spurs of bone on the hinder border; that a closely fitting integu- ment was present, as is indicated by the many impressions of blood vessels on the upper surface, with the probability that the projections of the periphery at the sides and behind were sheathed with horn.” In his second paper on this species Lambe gives the following: Principal measurements. Mm. Extreme length from anterior end of crest (imperfect), medially, to line touching posterior edge of specimen oneither side. 616 Length on median line, from anterior end to posterior border.........-.-.-.-------- +++ +++ +++ ss eee eee e esses etree seen eens esses 486 Semibreadth of specimen on curve of under surface .........-..---.---- 22-2 scene ce eee eee rete teeter nee eens 470 Semibreadth of specimen horizontally .....-......---2- 2-22 -----0e eee e eee ence rete teen e eee nee etre scene erences 439 Vertical drop of lateral edge of specimen below median line of upper surface at mid-length.......-------- +++ ..esseseeeesees 186 Antero-posterior diameter of fontanelle.........-...-.-.------- sec ABR ne EEE ettce PEAS SoC ECE MAD SRE CODES acne 296 Transverse diameter of fontamelle. ....-...---- 5.2 cee ee cence eee ee ee ce nee eee eee nee enw e ce setereesenestseereseatas 248 Circumference of base of left posterior Spur............-.--2----22- eee eee eee ee eee teen terete eens terete sseceeenes 172 Mr. Lambe’s description of this interesting type is mainly due to Hatcher's suggestion that it represents a distinct genus and species from Monoclonius dawsoni.—R. 8. L. 94 THE CERATOPSIA. Red Deer River, Alberta, between Berry Creek and Dead Lodge Canyon. Lambe’s original description is as follows: This species is founded on a squamosal, part of a parietal, a jugal, a supraorbital horn core, the left ramus of the lower "jaw, and an anterior dorsal vertebra, with some other parts of the skull, not yet fully determined, of one individual. A right ramus of another individual is shown on PI. XVIII [XIX] and a separate horn core on Pl. XVII [XVIII]. A right mandibular ramus, referred to this species on account of its resemblance in form to the one shown above, is described further on. 2 The horn core (fig. 18 [fig. 96, Pl. XVIII]) rises above the orbit frome the postfrontal, of which it formsa part. The postfrontal unites behind, by suture, with the squamosal and below with the jugal. The orbit is oval, with the longer diam- ~ eter vertical, its upper curve lying close under the base of the horn core, its margins not ridged. The horn core is small, about 21.6 centimeters long from the upper edge of the orbit to its summit and 22.8 centimeters in circumference near the base, circular in section, and solid. Squamosal [P]. XXIII], somewhat triangular in shape, flat, moderately thin, its outer edge smooth, rounded, wavy in out- line, so as to produce six minor convex curves. Shorter and more pronounced near the front. Its outer front edge is deeply emarginated, with a shallower concavity limiting the outer ter- mination of the jugal suture (see fig. 18 [fig. 96]), inside of which is the suture for the union with the postfrontal. The inner bor- der is slightly concave. The lower surface near and parallel to the inner posterior end is broadly and shallowly grooved for the reception of a long, slender bone, triangular in section, that projects backward and inward, its outer edge continuing the curve of the squamosal. Probably this slender bone represents Fig. 96.— Monoclonius canadensis; part of the skull fromthe right the anterior end of a forward bent side extension of the parie- lateral aspect. One-sixteenth natural size. fp, Pdstfrontal; h, ta] such as occurs in the species Monoclonius belli, in which case horn core; 0, orbit; s, squamosal; p, right lateral extension E e ay ie irom wanicvals jijueal Guisplaccdl naselayaualowen jane atrermee fontanelle of moderate size might be expected on the inner Lambe. side of the squamosal. In fig. 18 [fig. 96] the underlying bone (imperfect posteriorly) is indicated by a dotted outline under the squamosal, beyond which it projects; its outer free edge shows a round-edged con- vexity in continuation of the sinuosities of the squamosal. The proximal inner margin of the squamosal is bent at right — angles to the plane in which the remainder of the bone lies and its under surface is deeply excavated in its inner front part for some distance back from the postfrontal suture. Measurements of squamosal, etc. Mm ILengthvon\ curve otfoutersborden ees sete eee Mena iene ee eee a enone 0. 576 Length: onccurverofanner, borders: 0c) goon oe ee a ge ee a oe 973 ene thi iromlpostenlorremd ston Gerber ago rity pine 110s yee eae ee ee ene eee ee . 533 Mhickness!mear outer order art ai d= ler cules ae ee yer ees re ee See eS . 028 Breadthsacross:fromb: margin.” tee hss oo hey et ev ae le ee ee 399 Ani ckavess ive ary MATE Ty Oe e repent torn 1 0 es a . 038 Length of bone underlying the squamosal (imperfect).......-.---.--- wel te eM Seles Bate reme ager . 502 Breadth*ot same; at mid-lengths 29: 4a 9-2 oe eee Seats ee Ese . 064 Greatestithicksmess) a taal oer ort nesses ee res Fe Sipe s We ac Sr Sees z eee See Sree ie Sea . 030 With the parts of the head shown in fig. 18 [fig. 96] was also found an anterior dorsal vertebra, fig. 19 [fig. 97], of rather small size. The faces of the centrum of this vertebra are slightly concave. Next following is the description of a right mandikular ramus, found separately in 1897, but agreeing in size with the one depicted in fig. 18 [fig. 96]. Ramus of lower jaw (right), Cat. No. 284. Ramus of lower jaw (Pl. XVIII [XTX], figs. 1 and 2) stout, with an inward bend at mid-length, low and thick behind, elevated and laterally compressed in front where the inner surface is shallowly concave. Excayated posteriorly below for nearly one-third of its length, the excavation extending upward along the back surface of the coronoid process and anteriorly as the mandibular canal leading forward to the mandibular groove in the lower border. The dentary canal between the outer alveolar wall and the outer surface enters from the upper and anterior part of the excavation by a large opening. Coronoid process stout, upright, hooked forward and flattened laterally above, its outer upward surface rugosely striated. A broad, low ridge, least defined toward the center, runs at about mid-height along the outer side, the surface, m a general way, above and below, retreating obliquely inward. The dental chamber, straight, starting at a low level behind, inclined strongly upward and slightly outward toward the front, its lower edge making an angle of about 20° with the lower border. Alveolar a Referred to as a nasal by Lambe in Trans. Roy. Soc. Canada, 2d ser., vol. 10, sec. 4, p. 7. Ee oe — I J j f % ‘ ~ MONOCLONIUS CANADENSIS. 995 grooves in outer wall of dental chamber deeply impressed toward their upper ends by a second series of groove terminations, an evidence of two roots in the teeth belonging to this jaw, such as are characteristic of some of the species of the Ceratopside (Agathaumide). Height of dental chamber much reduced forward. A number of large foramina present in the outer sur- face. Front border, as viewed from the side, sinuous, rugose for its union with the predentary bone. Twenty-three alveolar grooves are present in the dental chamber (imperfect posteriorly) of the specimen figured. A small symphyseal surface is present in the front lower border. Measurements of ramus of lower jaw. Mm. Bxtnemepleng tive tanatd lel ch tee smear ater per Retr eye eC re rar ees NN IT or SoS Se 9. 398 IDyeyoyilen aie teow el exe Noo cisd Sea cei sees SS ee Ee SS re A Oe he . 116 Distance from upper border, a little in advance of front end of dental chamber to lower posterior border Giles ywmphysealusurkacemmee ete. mya pam ian amine ek dal a 2 ak 2 Ee al ade. aa ee S137 Heizhtofetacetonarticulation! ofjpredentany bones=-25--- 2-422-445-2452 52522 == 2-2 = ose ee . 096 Distance from top of coronoid process to lower border.--.-.-.-.------------------------------- ~ 9 ith Breadth of coronoid process from point of anterior hook backward.---.-.----------------------- . 097 Thickness at center of upper coronoidal expansion.-.-.---..--.------------------ ge eye . 024 thicknesstoficoronoid processiatitssmid-nelgh tse sass eee ee eee . 038 Antero-posterior diameter of symphyseal surface ..-...-.-.-.--------------------------------- . 055 EDEN OF GME Seon no Gada ae ae eee Re Be NS Oe Oe ey EIS BO ae ae eas eee eos 025 Width of larger alveolar grooves at middle of dental chamber------------- Bc eee nae eae eee . 009 SPxd PTOOVES UN) ay SPACe Ol mae as ae eee eee a ee See ea tee eA Ge Seas CSE a eaeetoe . 072 Height of grooves from their base to upper edge of outer alveolar wall, at middle of dental chamber. .028 Heichtrotysamerantenionlyern aemae oe eee Oe eee ee see ee eae Chea Mae we Boel See . 044 A maxillary bone (not figured) with teeth that are double fanged is réferred to this species. One of the teeth is shown on Pl. XVIII [XTX], figs. 3 and 4. A separate tooth, presumably from the lower jaw, is also figured on Pl. XVIII [XIX]. It was found separately, but on account of its having two roots, agreeing thus with the evidence of the alveolar grooves of the mandibular ramus just described, it is likewise referred to M. canadensis. The discovery of this specimen is of the greatest importance, since it affords the first defi- nite information regarding the character of the parietals and squamosals that are associated with the type of frontal horn cores shown in Marsh’s type of Ceratops montanus. A com- parison of the frontal horn of the type of the present species with that of Marsh’s type of C. montanus makes it apparent that they are essentially the same, and I do not hesitate to remove the present species from the genus Monoclonius Cope and place it in that of Ceratops Marsh. Whether or not it should be regarded as specifically distinct from C. mon- tanus 1 am unable to say without further study, and this question will be left for that portion of the present volume devoted to a revision of the genera and species. It is true that Professor Marsh had referred the Fig. 97.—Anterior dorsal vertebra of Monoclonius canadensis. One- broad squamosal shown in Pl. Ta fig. il. to fourth natural size. A, Front view; B, left side view. a, Anterior / , 1] : > face of centrum; d, diapophysis; h, facet for head of rib; n, neural oy, ‘ JU: ; aS m ; \ ; Ceratops montan US, but since that Mee canal; s, neural spine; ¢, facet for tubercle of rib; z, prezygapophy- was found at a place about 15 miles distant sis. 2’, postzygapophysis. After Lambe. from the locality that yielded the type, Marsh’s correlation, though possibly correct, is purely conjectural, and to Lambe’s fortunate discovery we are indebted for positive knowledge concerning the actual association of these important elements of the skull in this type of the Ceratopsia. The squamosal (No. 1254a, Geol. Surv. Canada) found associated with the type of the present species exceeds in length that of any squamosal yet found from the Judith River beds. The portion posterior to the quadrate groove is more than twice the length of the anterior por- tion. Posteriorly it is much contracted and elongated and is already clearly assuming the form shown by the squamosals in the later genus Jorosaurus, from the Laramie. The quad- 96 THE CERATOPSIA. ratojugal notch is deep and the inferior border of the squamosal posterior to this notch is pro- duced downward and forms a rather acute triangular projection. Back of this the external border of the squamosal presents six gentle prominences separated by intervening sinuosities. - These are more pronounced and less separated from one another in the anterior region than in the posterior. The border throughout is rather heavy. [See Pl. XXIII, figs. 1, 3.] On its internal and inferior border throughout the posterior one-half of its length the squamosal presents a deep and broad groove for the reception of the slender bone shown by Lambe and seen here in fig. 96. [Also in Pl. XXIII, figs. 2, 4.] This bone is imperfect at the posterior extremity, but it is clear that it represents a portion of the external bar of the parietal. Its general form is very similar to that element as seen in the type of Torosaurus gladius Marsh, and it articulates with the squamosal in the same manner. The distal portion of that part of this element preserved in the present specimen presents on its external margin beyond the surface for contact with the squamosal an elongated prominence which, when the bone is placed in its proper position relative to the squamosal, is seen to form a continuation of that series of prominences already described as being present on the border of the squamosal. The internal margin of this external bar of the parietal is very thin throughout most of its length and evidently formed the external border of the parietal fontanelle. I have little hesi- tancy in asserting that the squamosal and frontal horn cores of the present species were associated with a parietal of the same general type as that described by Lambe and referred to Monoclonius belli, and I am of the opinion that the two may be specifically identical, although from the material at hand it is impossible to determine this point with certainty. The supra- orbital horn core and postfrontal found associated with the other remains are similar in form to those which formed the type of Ceratops montanus Marsh and distinct from those associated with the type of Monoclonws crassus Cope. The bone figured and described by Lambe (see fig. 96) as a jugal* proves on examination to be a right nasal. It shows posteriorly the proper articular surfaces for contact with the frontals and prefrontals, and sends downward a process for articulation with the ascending branch of the maxillary and premaxillary. Anteriorly it sends downward a process for contact with the premaxillary. Superiorly on its inner surface it presents a broad and elongated articular surface for contact with the opposing nasal. The nasal horn was detached, indicating that the animal was not fully adult. The inferior border is concave and describes a nearly complete semicircle which formed the superior border of the nasal opening. In the present species the anterior portion of the squamosal may be described as being composed of a broad vertical portion and a narrower horizontal portion. The horizontal por- tion, together with that of the opposite squamosal, formed most of the superior surface of the skull in this region, reducing the anterior extremity of the parietals to a narrow point, as In Torosaurus. The angle formed by the union of the vertical and horizontal plates at the anterior extremity of the squamosal is produced into a strong ridge and shows a series of rugose promi- nences separated by concave surfaces. One of these latter is very deep. Posterior to these the superior margin of the squamosal is deeply emarginate, and this emargination probably led to the supratemporal fontanelle. Monoctonius BELLI Lambe. 1902. Type (No. 491, Geol. Sury. Canada) consists of a parietal. Original description, Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 65-67. Lambe, L. M., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 59, 64, 68, 81. Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 14, 20. The type (No. 491, Geol. Surv. Canada) of the present species consists of the greater portion of the coalesced parietals, as shown in Pl. XXJ. The bone was found by Mr. Lambe in 1898 in the Belly River series on the Red Deer River. aSee footnote, p. 94. —.- o> —- iw SS ee —e eS STEGOCERAS. on His original description is as follows: The bone, figured on the above plate [Pl. X XI], is interpreted as representing the coalesced parietals of the posterior crest of an undescribed species of Monoclonius, probably ancestral to such later forms as Torosaurus latus and T.gladius of Marsh from the Laramie of Wyoming. To facilitate an understanding of the view held as to the position the FF parietals probably occupied relative to other bones of the head, a drawing i of the bone has been applied to the figure, slightly modified, of the skul] of T. gladius, as given by Marsh in the Sixteenth Annual Report of the : United States Geological Survey. a The parietal element from Red Deer River is symmetrical, T-shaped, } with a subcylindrical shaft expanding rapidly both in front and behind. Anteriorly the expansion is concave below, strengthened above by a median, rounded ridge in continuation of the central shaft, and thinning out later- ally. Posteriorly the shaft divides, nearly at right angles to itself, to either side, so as to form a strong transverse bar slightly concave at mid- length above and convex below, thin at its front edge and thickest behind. The posterior border is angularly rounded. The space on either side of the shaft represents the inner halves of the supratemporal fontanelles. The bone missing from the specimen would complete the outer border of the fontanelles and effect a union with the inner margins of the squamosals. The lower face of the anterior expansion, on either side of the median line, is striated by distinct furrows that follow down the lower lateral sides of the shaft as deep grooves and curve outward on to the transverse bar. The anterior upper surface also F16-_98-—Posterior crest of Monoclonius belli, trom ae pa 2 Red Deer River. One-eighteenth natural size. exhibits similar grooves that do not, however, pass beyond the mid-length Tie ted i cclare trota hedrawinitot the Read ‘ = =" ale if} of the shaft. * of Torosaurus gladius Marsh, as seen from above. The parietal, imperfect at its anterior end, is about one-third the size P, Parietal; S, squamosal; F, fontanelle. Aiter of that of 7. gladius, and would probably represent a proportionately Lamibe- smaller animal, an earlier and more generalized form of the genus with larger fontanelles than its later Laramie successors. Measurements of parietal bone. Mm Extreme length of specimen (imperfect anteriorly) along median line.__._.........------------- 0. 584 Breadth of front expansion from median line to left edge of specimen_--.-.--------------------- . 173 Breadth of posterior border from median line to left edge of specimen. ----.------------:------- . 805 Crconicreneeotsshaticatuic-lenoth eens aoe S82 see ee oe aes ea es Saree teieie lee ive Baleceyeys . 180 Broadr nlotmsaniera und lenoihiw re are aan i eee RE ese Sh tol re ales . 065 pickrecsoisamernusinid— lene tienen ees ere eee We ST eS ae chine . 053 Thickness of anterior expansion at center on median line................---------------------- . O41 Thickness at anterior end of specimen on median line._____....-....----- yah eee HOLS Thickness on median line midway between posterior border and TROT sea of aie. LI eee es 035 AnieKo-poOsvenormaiamevarortontanel less! 2 sey Se eke eee eo ee oe . 416 Belly River series, Red Deer River, 1898. This species is dedicated to Dr. Robert Bell, the administrative head of the Canadian Geological Survey. After a careful study of the type of the present species, together with that of J. canadensis Lambe, one can not avoid being convinced as to their generic identity with Ceratops montanus Marsh, while at the same time the great dissimilarity shown in the parietals and squamosals of these species when compared with the same elements in Monoclonius dawsoni Lambe affords evidence additional to that already pointed out as obtaining in the frontal horn cores, in favor of the generic distinction of the three former species from that of the last-mentioned species. STEGOCERAS Lambe. 1902. Type species, S. validus. Original description in Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 68 and 69. Nopesa, F. Baron, Centralblatt fiir Mineralogie, 1903, p. 266. Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), pt. 3, 1902, p. 21. This genus was founded by Lambe on some problematical reptilian remains from the Judith River beds of Canada. Although they are probably not referable to the Ceratopside they will be considered here for the reason that Lambe referred them to that family. MON XLIx—0O7——7 ee — - 98 THE CERATOPSIA. STEGOCERAS VALIDUS Lambe. 1902. Type (Nos. 515¢@ and 1423, Geol. Surv. Canada) consists of two cranial fragments pertaining certainly to different species: and possibly to different genera. Original description in Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 68-69. Lambe, L. M., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, p. 81. Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, p. 15. The present genus and species are based upon the fragmentary remains of three different individuals. Unfortunately Lambe failed to designate any one of these remains as the type of the genus or species,” and he remarks in speaking of the two better preserved specimens that they may pertain to different species. His description is as follows: The two symmetrical, compact bones, represented on P]. XXI [XXII], were found separately. The lower portions of their sides, as well as their ends, consist of sutural surfaces, indicating that other bones were firmly united to them and completely surrounded them. A transverse suture divides each almost equally into an anterior and a posterior half. On the lower surface there is evidence of a line of coalescence in a longitudinal direction and extending from end to end. The upper’ surface of each specimen is dome shaped. In the larger specimen (Cat. No. 515)a the anterior end is produced forward and is slightly elevated, terminating in two: projections; the surface is here distinctly nodose. In the lateral posterior upper surface a similar rugosity is apparent. The surface of the central convexity is smooth. In the smaller specimen (Cat. No. 1423) the upper surface is smooth and pitted throughout. It is trilobed posteriorly and is not produced forward in front, where, however, two small nodes occur, one on each side of the median line. The structure of the lower surface is marked by a number of smooth, concave areas, as represented in the reproductions,. from photographs of the specimens, in figs. 2 and 5 of Pl. XXI [XXIJ]. - It is probable that these bones were situated in the median line of the head, in advance of the nasals. They may have: belonged to a species of dinosaur not otherwise represented in the collections from Red Deer River and, judging from the difference in shape of the two specimens, more than one species may be indicated. Marsh in his figure of the head of Triceratops’ serratus shows a nasal horn core (divided both transversely and longitudinally by sutures) that may correspond to the speci- mens from Red Deer River. A third specimen (Cat. No. 1594), similar to the anterior half of the larger of the two bones, was collected in 1901. _ It has: separated from its posterior half along the line of the transverse suture. For these bones the name Stegoceras validus is proposed with the hope that future discoveries may aid in a clearer under- standing of their affinities. (See also Cat. No. 1075.) Belly River series, Red Deer River. 1898, 1901. After a careful study of the materials upon which the present genus and species were based I am fully convinced that they pertain to a reptile hitherto unknown. I can not, however, agree with Lambe’s determination as to the position which these elements occupied in the skull,¢ nor do I believe that they pertain to any member of the Ceratopsia. And it has yet to be shown that they belong to any dinosaur. Ishould not be surprised if, when the true nature of the ani- mal to which they pertain is known, it would prove to belong to some other reptilian order. As. suggested by Lambe I am inclined to the opinion that the two better preserved specimens belong to different species, but I can not agree with him in interpreting them as prenasals. They appear to me rather as representing the superior portion of the occipital, parietal, and frontal segments of the skull of some reptile in which, as shown in fig. 99, the different bones of this region have been greatly reinforced from above by the coalescence of dermal ossifications — which rise in a dome-shaped mass above the brain case, completely enveloping the cranial elements of this region and thus giving great additional strength to the cranium. | This ossified dermal covering, as it might be called, differs in structure from the ordinary membrane bones forming the roof of the brain case to which it is attached in being extremely dense and in exhibit- ing a columnar structure, so that about the margins this dome-like mass of bone, overlying and firmly coossified with the ordinary cranial elements, is seen to be composed of innumerable a Mr. Lambe considers No. 515 as the type.—R. 8. L. b These sutures separate the nasals and premaxillaries; the horn core is absent.—R. 8S. L. ce In a paper published since the death of Hatcher (Trans. Roy. Soc. Canada, 2d ser., vol. 10, sec. 4, p. 24), Mr. Lambe says: “ Stegoceras- validus is based on portions of the skull from the median line of the head, with indications on the upper surface of the presence of an unpaired horn. These parts were supposed to.be prenasal, but, as pointed out by Nopcsa (Ueber Stegoceras und Stereocephalus, von Franz Baron Nopesa, jr., Centralblatt fiir Mineralogie, etc., No. 8, 1903, Stuttgart), they probably represent the frontal and nasal elements of the skull. In Stegoceras we have an entirely new type, a unicorn dinosaur remarkable in that it bore a horn springing from the fronto-nasal region,. recalling a somewhat similar development in the mammals A ceratherium incisivum and Elasmotherium sibiricum.”—R. 8. L. STEGOCERAS VALIDUS. 99 minute columns of bone arranged perpendicular to and radiating from the external surface of the various bones of the roof of the skull, with which they are firmly coossified at their bases, as shown in fig. 100. If my interpretation, which differs only in a few unimportant particulars from that of Nopesa, is correct, Lambe was not only wrong in considering these elements as prenasals,” but he also possibly mistook the anterior for the posterior extremities. I would interpret the large median cavity shown on the inferior surface in fig. 99 as the upper portion of the brain cavity, while the transverse suture which divides it into ante- ; rior and posterior moieties I would consider the parieto-frontal suture, rather than the fronto-nasal, as believed by Nopcsa. Thus the bone in front of the suture becomes, according to my interpretation, the frontal rather than the nasal, as regarded by Nopcsa, though faint lateral sutures just in advance of the interorbital constriction may mark the posterior borders of the nasals. Like Nopesa, I regard the lateral cavities as orbital. Posteriorly in the larger of the two specimens there are two deep lateral cavities which opened externally by a small circular foramen, only the internal border of which latter is rep- resented in either specimen. The deep lateral cavities I inter- pret as infratemporal fossz, while the circular foramina opening into them may possibly represent the supratemporal fosse, though from the position of these foramina this interpretation appears hardly probable. On the larger of the two more complete specimens just behind that cavity Fic. 99.—Inferior view of Stegoceras validus & Lambe, No. 1423, Geol. Surv. Canada. which I have considered a, Anterior; p, posterior; m, nasal; f, the brain cavity there is a frontal; pe, parietal; so, supraoccipital. Natural size. broken or sutural surface, ¥1G. 100.—Side view of type specimen, No. ss, {0llowed by a rather deep median excavation terminating Geol. Surv. Canada, showing columnar struc- anteriorly in two shallow pits separated by a median eee ee) septum. ~ This may have formed the roof of the foramen magnum. Without more perfect material it is quite impossible to determine definitely either the homologies of these elements or the nature of the animal to which they pertained. In the present writer’s opinion, however, there is no good reason for considering them as horn bearing, or the animals to which they belonged as pertaining to the Ceratopside. I, therefore, do not include them in that group. Principal dimensions of bones of Siegoceras validus. For the larger (type specimen No. 515, Geol. Surv. Canada): Mm. (Crepe MEME 5 os Lee Sees pete Me SERS Ot ore ee ee 106 Cremiesin |aienll 25S se bee ce ele eae oo 8 eee es oY et eee ee ee 60 (Cayeriegt, Gayla en tees eo cee ee oe Cer Sl ely ee OI Rs ne 41 For the smaller (No. 1423, Geol. Surv. Canada): (Gnesi Radio Sera bet be ee oes oo aan 24 ee cee a 64 (Caroniecte iG aalilsh 2 olan an 22 Seen ci SD Se cs reed Sea 44 CHR ish CY DUN ern Se sleet cao HFS pe Coe COSA Oe Brot a ea eistet oes 31 Taken as a whole the collection of Ceratopsia made by Lambe from the Belly River beds is a most important one, since it furnishes much valuable evidence not only regarding the dis- tinctly generic characters of these earlier and less specialized members of the family, but relating also to the ancestry and phylogeny of the later, more highly specialized, and much better known a See footnote c, p. 98. i i | | i ‘a \ t : h 4 100 THE CERATOPSIA. forms from Converse County, Wyo. The affinities of Monoclonius, as shown in the type species M. crassus Cope and in MM. dawsom of Lambe, are apparently with the later genus Tricera- tops of Marsh, while Ceratops montanus Marsh, C. recurvicornis Cope, C. canadensis, and C. bella Lambe would seem to be ancestral to Torosaurus. These questions will be more fully discussed when we come to treat of the phylogeny and taxonomy of these dinosaurs. GENERA AND SPECIES DESCRIBED BY PROFESSOR MARSH FROM THE JUDITH RIVER BEDS. CERATOPS¢ Marsh. 1888. : Type species, C’. montanus. Original description in Am. Jour. Sci., 3d ser., vol. 36, Dec., 1888, pp. 477-478. Marsh, O. C., ibid., Apr., 1889, pp. 327, 334-335; Aug., 1889, p. 175; Dec., 1889, pp. 505, 506; Jan., 1890, pp. 81-83; Feb., 1891, pp. 171, 176; Apr., 1891, pp. 340-341; Sept., 1891, p. 266; Jan., 1892, p. 83; Dec., 1895, p. 497; Sixteenth Ann. Rept. U.S. Geol. Survey, 1896, pt. 1, pp. 145, 206, 210, 216, 219, 243. Ed. Am. Geologist, vol. 8, 1891, p. 56. Baur, G., Science, vol. 17, 1891, p. 216; Am. Naturalist, vol. 24, 1890, p. 570; Am. Naturalist, vol. 25, 1891, p. 450. Hatcher, J. B., Am. Naturalist, vol. 30, 1896, p. 113. Lydekker, R., Nature, vol. 48, 1893, p. 304. Nicholson and Lydekker, Manual Pal., 1889, vol. 2, p. 1163. Nopesa, F. Baron, Féldtani Kézlény, Budapest, 1901, vol. 31, p. 270. Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 9, 20. Woodward, A. S., Outlines Vert. Pal., pp. 213, 216. Zittel, K. A. von, Text Book Pal., trans. by C. R. Eastman, vol. 2, p. 245. The present genus may be distinguished from Monoclonius Cope, based on material from the same beds in Montana, by the greater development of the supraorbital horn cores, the longer and narrower squamosals, the enlarged fontanelles, by which the parietals are reduced to slen- der median and lateral bars. The nasal horn cores are very probably quite different also in the two genera, though we can not as yet be certain as to their character in Ceratops. From our present kaowilades of the skull of Ceratops it seems to have been a precursor of Torosaurus ‘Marsh, while Monoclonius Cope appears to have been ancestral to Triceratops Marsh. CERATOPS MONTANUS Marsh. 1888. Type (No. 2411, U. S. National Museum) consists of occipital condyle and pair of frontal horn cores. Original description in Am. Jour. Sci., 3d ser., vol. 36, 1888, pp. 477-478. Hatcher, J. B., Am. Naturalist, vol. 30, 1896, p. 113. Marsh, O. C., Am. Jour. Sci., vol. 37, Apr., 1889, p. 327; Jan., 1890, p. 83; vol. 43, 1892, p. 84; Sixteenth Ann. Rept. U.S Geol. Survey, pt. 1, p. 216. Nopesa, F. Baron, Féldtani Kézlony, Budapest, 1901, vol. 31, p. 270. ; Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 9, 14, 18, 20. Walcott, C. D., Science, new ser., vol. 11, 1900, p. 23. In December, 1888, Professor Marsh described the new genus and species Ceratops mon- tanus. The type (No. 2411, U. S. National Museum) consists of an occipital condyle and a pair of frontal horn cores, found together and pertaining to the same skull. These remains were found by me in 1888. The horizon was near the top of the Judith River beds, and the exact locality was on the northwestern slope near the summit, about 300 yards from the point of the first hogback that projects into the valley of Cow Creek from the west, just below where the old Cow Island and Fort Benton freight road descends into the valley of Cow Creek, about 10 miles above the confluence of that stream with the Missouri River. Marsh’s original description of the above genus and species was as follows: The present genus appears to be nearly allied to Stegosaurus of the Jurassic, but differs especially in having had a pair of large horns on the upper part of the head. ‘These were supported by massive horn cores firmly coossified with the occipital crest. The latter are probably attached to the parietal bones, but, as the sutures in this region are obliterated, they may be supported in part by the squamosals. aThe name Ceratops was used by Rafinesque in 1815 to designate a genus of birds. As no description of the genus was published, and as no forms were mentioned as pertaining to it, it becomes a nomennudum. Should expert nomenclaturists decide that Rafinesque preoccu- pied the name Ceratops I have suggested that the term Proceratops be used in its stead. (See Am. Jour. Sci., 4th ser., vol. 21, p. 144:)—R.S. L. CERATOPS MONTANUS. 101 The horn cores in the type specimen are subtriangular at base, but nearly round in section in the upper half. Their position is represented approximately in the figures of the accompanying plate [shown here in fig. 100]. These horn cores are slightly hollowed at the base, but are otherwise solid. The exterior texture and markings show that they were evidently covered with true horns, and these must have formed large and powerful offensive weapons. In position and direction these horn cores are somewhat similar to the large posterior pair of protuberances in Metolania,2 one of the extinct Testudinata, and to the corresponding ones of the existing Phrynosoma. The only known example of similar structure in the Dinosauria is the single median horn core on the nasals of Ceratosaurus, from the Jurassic. It is not improbable that there were other horn cores on the skull in the present genus, but of this there is at present no positive evidence. A detached median prominence resem- bling a horn core was found with some similar remains, but may pertain to an allied genus. The resemblance in form and position of the posterior horn cores to those of some of the ungulate mammals is very striking, and, if detached, they would naturally be referred to that group. The basioccipital found in place with these horn cores and represented in Pl. XI [fig. 104, 1b and 2 5,] is much elongated, and formed the entire occipital condyle. Its exact position with reference to the horn cores could not be determined. ; Fig. 101.—Upper view of cranium of Meiolania Teeth, vertebree, and limb bones, which probably belonged to the pres- platyceps Owen, drawn from a cast (No. 208) in ent genus, were all secured in the same horizon. They indicate a close affinity the U. 8. National Museum. One-fourth nat- with Stegosaurus, which was probably the Jurassic ancestor of Ceratops. ee Se Among other remains referred to the present reptile, but not found with the type specimen, are some peculiar large der- mal plates, in pairs, that indicate a well-ossified armor. These plates show indications of being covered, in part, at least, with scutes, as in turtles. Their position can not at present be determined. The type specimen on which the present genus and species are based was found in place, in the Laramie deposits of the Cretaceous in Montana, by Mr. J. B. Hatcher, of the U.S. Geological Survey. Other specimens, apparently pertaining to the same species, were secured in the same horizon of the same region. Remains of the same reptile, or one nearly allied, had previously been found in Colorado in deposits of about the same age by Mr. G. H. Eldridge, also of the U. S. Geological Survey. The associated fossils found with the present specimens are remains of other dinosaurs, crocodiles, turtles, and fishes, mostly of Cretaceous types. The mollusks in the same beds indicate fresh-water deposits. The fossils here described indicate a reptile of large size, 25 or 30 feet in length, and of massive proportions. With its horned head and peculiar dermal armor, it must have presented in life a very strange appearance. The remains at present referred to this genus, while resembling Stegosaurus in various important characters, appear to represent a distinct and highly specialized family that may be called the Ceratopsidee. They will be described more fully in a later number of this journal. As is now well known, the above description is erroneous in many particulars. The beds were not Laramie, but Judith River, a distinctly older formation. The horn cores in the Cer- atopside are not firmly coossified with the occipital crest, and they were not di- rectly attached to either the parietals or squamosals. Marsh’s suggestion that there were probably other horn cores on the skull proved correct. The basiocci- pital does not form the entire condyle, but that element is made up of the coossi- Fic. 102.—Side view of the coalesced terminal segments of the dermal bony fied basioccipital and exoccipitals. The tail sheath of Meiolania platyceps Owen, from a cast (No. 207)in the U.S. -Jarge plates mentioned by Marsh as der- National Museum. One-fourth natural size. 5 C mal plates are in reality squamosals. The size of ‘the animal indicated by the type material was considerably overestimated by Marsh, and his suggestion that Stegosaurus was the Jurassic ancestor of Ceratops might at present be questioned by many. a Meiolania platyceps was described and figured by Sir Richard Owen in the Philosophical Transactions of the Royal Society, 1888, B, pp 181-191, pls. 31-37. Meiolaniais a turtle, one of the Cryptodira, found on Lord Howes Island in the Pacific, and in at least two features is sug- gestive of the Ceratopsia. The superior view of the hinder portion of the cranium here shown (fig. 101) exhibits, in addition to the low hind- ermost pair of horns, ‘‘larger cores, which rise in advance of and exterior to these; the length of this core is 2inches’’ (Owen). These cores are the ones to which Marsh refers as being similar to those of Ceratops montanus, and, like those of the dinosaur, were evidently sheathed with horn. Fig. 102 shows the osseous tail sheath of Meiolania, which also bears vascular markings similar to those borne on the ceratopsian frill.—R. S. L. 102 THE CERATOPSIA. I can add nothing of importance to Professor Marsh’s original description of the type, except to give the more important measurements. These are as follows: Measurements of Ceratops montanus. Mm. Height’ of frontal horn.core:! 222.58 2 Roe aye Crees een) eR ear Oey an a 220 ANMEDEOHONATO? CHITAGUP Ot SENNG HUF OREO. = os soso dome ceased ea seocaac sso soccosessbserecezsess 93 Mransverse diame ter OlgSai ey ett ASCs sesame es meee ea ase te ane a aa 88 Aransversexdiame bet OlOCClpltalla cond iy ees ee ae 71 _ The principal characters of the type material are well shown in figs. 103% and 104, and from these and the above description it is clear that the genus Ceratops must rest on the char- acters displayed by the occipital condyle and the frontal horn cores, neither of which elements are represented in the actual type of the genus Monoclonius of Cope. If the frontal shown in fig. 75, bearing the diminutive horn core which was subsequently associated by Cope with the type, though evidently pertaining to a much smaller individual, be admitted as a cotype, and therefore to some extent, at least, diagnostic of the genus Monoclonius, there would still be good reasons for considering Ceratops montanus as both generically and specifically distinct Wie Fic. 103.—Supraorbital horn cores of the type of Ceratops montanus Marsh (No. 2411, U. S. National Museum) in their proper position. A, O blique back view; B, anterior view. One-fourth natural size. = from Monoclonius crassus, for such striking differences as are shown in these two types of frontal horn cores are certainly suggestive of other and even more important structural differences in the skull and other portions of the skeleton. However, the full discussion of such questions must be left for that portion of this volume devoted to a revision of the genera and species. The squamosal figured by Marsh® as pertaining to Ceratops montanus, but originally described in the above quotation as a dermal plate, was found by the present writer many miles from the locality. which furnished the type. It may or may not pertain to that genus and species. Judging, however, from the character of the squamosal in Ceratops (Monoclo- mus) canadensis Lambe, where the supraorbital horn cores are strikingly similar to those of the present species, it seems probable that Marsh was in error in referring this squamosal to C. montanus. aJn figuring the type of Ceratops montanus, Marsh mistook the external lateral view of the supraorbital horn core for the posterior. For- tunately a portion of the orbit is present, making it possible to determine the position of the horneores with certainty. In fig. 103 these horn cores are shown in their proper positions, which is somewhat different from that assigned them by Professor Marsh. bAm, Jour. Sci., vol. 43, 1892, pl. iii, fig. 3. This monograph, Pl. I, fig. 1. ae 2 a a Pe Pe BA 3 ; . is CERATOPS. 103 CERATOPS PAUCIDENS Marsh. 1889. Type consists of left maxillary and a premaxillary (in U. S. National Museum), Judith River beds, Dog Creek, Montana. Originally described as Hadrosaurus paucidens in Am. Jour. Sci., 3d ser., vol. 37, Apr., 1889, p. 326. Later placed in the genus Ceratops (Am. Jour. Sci., 3d ser., vol. 39, Jan., 1890, p. 83). Nopesa, F. Baron, Féldtani Kézlény, Budapest, 1901, vol. 31, p. 270. The type of this species (now in the U. S. National Museum), a left maxillary and a premaxillary, was found by the present writer in the upper Judith River beds, on Dog Creek, Montana. The exact locality was on the western slope, very near the summit, of a rounded badland hill about 20 rods east of the spring situated about a one-quarter of a mile east of the freight road running from Judith to Maiden, Mont. The locality is locally known as the “Bad Place” by freighters, having acquired this name from the nature of the road, which here follows the winding divide between Dog Creek and _ the Judith River. The divide at this place is both narrow and crooked. It is inclosed on either side by badlands which extend on either side to the neighboring streams. It is dis- tant about 12 miles from Judith post-office, on the Missouri River, and, on account of the spring mentioned above, dur- ing certain seasons of the year the locality is a favorite camp- ing place for freighters bound LEAN nl to and from either Judith or — Big Sandy, a station on the BA oe Great Northern Railway. In his original description 7% Professor Marsh referred this species to the genus Hadrosau- ae the year following, how- Fic. 104.—1, Supraorbital horn core and ies condyle of type (No. 2411, U. 8. National ever, he placed it in the genus Museum) of Ceratops montanus Marsh, side view: a, Horn core; b, condyle. 2, Posterior Ceratops. His original descrip- views of occipital condyle and supraorbital horn cores of same: a, Left horn core; c, right; b,condyle. One-fourth natural size. After Marsh. Yi ; j ”, y dy Vif ] Z bp, AAI LN \\\ \\ \< 4 VW AN ANN \ WW © (\ i 4 i ri \ tion is as follows: In strong contrast with the species above described (Hadrosaurus breviceps) is another from the same region and same formation. The best preserved specimen that now represents it is a left maxillary, nearly complete. With this was found some other portions of the skull, but the maxillary affords the best distinctive characters. All, however, indicate a skull of extreme lightness and delicacy of build for one of the Ornithopoda. The maxillary is especially slender, and the anterior and posterior extremities are pointed. The middle of the bone is more massive, but yet very light for this portion of the skull. The teeth are of the general type of those in this genus, but are comparatively few in number and only one row appears to have been in service. The maxillary preserved is about 10 inches in length, and 3 inches high near the center. The row of teeth in use contains about thirty. The remains on which the present species is based were found in 1888, in the Laramie formation of Montana, by Mr. J. B. Hatcher, of the United States Geological Survey. aAm. Jour. Sci., vol. 37, 1889, p. 36. 104 THE CERATOPSIA. The following year? Marsh recognized the true affinities of this species and placed it in the genus Ceratops. His remarks at this time were as follows: The specimen recently described by the writer under the name Hadrosaurus paucidens should probably be referred to the genus Ceratops, as a comparison with more perfect specimens indicates a much closer affinity with that genus than at first. supposed. In addition to the maxillary described, one of the premaxillaries is in good preservation. This agrees in general features with the corresponding bone in Triceratops, but is less specialized. Its inner surface is deeply concave, showing that. the two premaxillaries did not meet each other closely, as in Triceratops, but apparently only in front. This species, as well as the type of the genus Ceratops montanus, represents smaller, less specialized forms of the family, and may be from a lower geological horizon than the gigantic reptiles which the writer has recently made known. It is not at all unlikely that the type of the present species pertained to one of the several species of Ceratopside already described as from the Judith River beds. Since, however, the teeth, the maxillaries, and the premaxillaries of all these are imperfectly known, it is at present impossible to determine to which of them this specimen should be referred. Marsh’s statement that this species, as well as the type of the genus Ceratops montanus, represents smaller, less. specialized forms, from a possibly lower geological horizon than the gigantic Ceratopside from the Laramie of Converse County, Wyo., is significant, considering that the present type was recovered from near the summit of the Judith River beds, and should be taken as additional evidence in favor of the view long held by the present author, that the Judith River beds repre- sent a horizon decidedly older than the beds of Converse County, Wyo., the correctness of which was conclusively demonstrated in 1903 by the stratigraphic work of Messrs. T. W. Stanton and J. B. Hatcher.” SPECIES ERRONEOUSLY REFERRED TO CERATOPS. The two species Ceratops (Bison) alticornis and CO. horridus, although originally referred by Marsh to Ceratops, certainly do not pertain to that genus. Their affinities are more nearly with Triceratops, in which genus the latter was subsequently placed by Professor Marsh, though for some reason unknown to the present writer, or perhaps by oversight, the former was left by Marsh in the genus Ceratops. In the present volume they will be considered as belonging to the genus Triceratops, and their description will be deferred until we come to treat of the species of that genus. From the above description of these earlier, smaller, and more primitive forms from the Judith River beds we will pass on to the larger, later, and more specialized forms from the Laramie of southern Wyoming and of Converse County in east-central Wyoming, from north- western South Dakota, eastern Montana, and from the possibly later deposits, known as the Denver beds, of Colorado. REVIEW OF THE SPECIES OF LARAMIE CERATOPSIA. GENERA AND SPECIES DESCRIBED BY PROFESSOR COPE. AGATHAUMAS Cope. Type species is A. sylvestris. Original description of genus, Proc. Am. Philos. Soc., vol. 12, pp. 481-483. Baur, G., Science, vol. 17, 1891, pp. 216-217; Am. Naturalist, vol. 25, 1891, pp. 448, 450, 452. Cope, Ann. Rept. U. S. Geol. and Geog. Surv. Terr. for 1873, pp. 435, 438, 442, 444-446; Bull. U.S. Geol. and Geog. Sury. Terr., vol. 1, No. 2, Ist ser., pp. 16, 17; Rept. U.S. Geol. and Geog.,Surv. Terr., vol. 2, 1875, pp. 41, 53-54, 248; Am- Naturalist, vol. 12, 1878, p. 246; Proc. Acad. Nat. Sci. Phil., 1883, p. 99; Am. Naturalist, vol. 23, 1889, p. 715; Syl. Lectures on Pal., Univ. of Pennsylvania, 1891, p. 43; Am. Naturalist, vol. 26, 1892, pp. 757-758. Dana, J. D., Manual of Geology, 1895, p. 847. Editor Am. Geologist, Am. Geologist, vol. 8, 1891, p. 56.. Lydekker, R., Nature, vol. 48, 1893, p. 304. Lambe, L. M., Sum. Rept. Geol. Surv. Canada, 1898, p. 187. Marsh, O. C., Am. Jour. Sci., 3d ser., vol. 41, Feb., 1891, p. 176; vol. 43, Jan., 1892, pp. 83 and 84; vol. 50, Dec., 1895, p- 497; Sixteenth Ann. Rept. U.S. Geol. Survey, 1896, pt. 1, pp. 217, 248. Nopesa, F. Baron, Féldtani Kézlony, vol. 31, Budapest, 1901, p. 270. a Am. Jour. Sci., vol. 39, 1890, p. 83. > Bull. U. S. Geol. Survey No. 257, 1905. ! | AGATHAUMAS SYLVESTRIS. | 105 Osborn, H. F., Bull. Am. Mus. Nat. Hist., vol. 5, 1893, p. 326; Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 9, 20- Woodward, A. S., Outlines Vert. Pal., p. 213. Zittel, K. A. von, Text-book Pal., trans by C. R. Bastman, vol. 2, p. 244. This, the earliest known genus of the Ceratopside, represents one of the larger of these horned dinosaurs. Unfortunately nothing is known of the skull in this genus. This is especially unfortunate in the present instance, since in this group of dinosaurs the skull affords by far the best generic and specific characters, and most of the genera and species are based very largely and often entirely upon cranial characters. From the material at hand it would be hard to distinguish the present genus from any one of the several genera that have been proposed for the reception of these later Laramie Ceratopside. AGATHAUMAS SYLVESTRIS Cope. 1872. Type (No. 4000, Am. Mus. Nat. Hist.) consists of about sixteen vertebrz, including the sacrum, the right ilium, frag- ments of ribs, etc. Originally described in Proc. Am. Philos. Soc., vol. 12, pp. 481-483. Marsh, O. C., Am. Jour. Sci., 1873, pp. 230-231. Baur, G., Science, vol. 17, 1891, p. 217. Cope, E. D., Am. Naturalist, 1872, p. 670; Proc. Acad. Nat. Sci. Phila., vol. 24, 1872, p. 279; Ann. Rept. U.S. Geol. and Geog. Surv. Terr. for 1873, pp. 435, 438, 442, 444-446, 447; Bull. U. S. Geol. and Geog. Surv. Terr., vol. 1, No. 2, Ist ser., pp- 9, 11, 16-18, 20; Rept. U.S. Geol. and Geog. Surv. Terr., vol. 2, 1875, pp. 31, 34, 40, 54-56, 57, 248; Bull. U.S. Geol. and Geog. Surv., vol. 3, 1877, p. 594; Am. Naturalist, vol. 12, pp. 245-246; Am. Naturalist, vol. 23, 1889, pp. 715-717; Syl. Lectures on Pal., Univ. Pa., 1891, p. 43; Am. Naturalist, vol. 26, 1892, p. 758. Dana, J. D., Manual of Geology, 1895, p. 847. Hatcher, J. B., Am. Naturalist, vol. 30, 1896, p. 113. Nopesa, F. Baron, Féldtani Kézlény, Budapest, 1901, vol. 31, p. 270. Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, p. 14. / The first of these larger and more specialized members of the Ceratopsia to be discovered and described was the Agathawmas sylvestris’ of Cope. Indeed this was the first of all the genera and species of this group of dinosaurs to be described and named. LOCALITY. The type (No. 4000, Am. Mus. Nat. Hist.) of the present genus and species was discovered in 1872 by F. B. Meek, of Dr. F. V. Hayden’s Geological and Geographical Survey of the Territories. It was found in southern Wyoming, not far from Black Buttes station on the Union Pacific Railroad, 52 miles east of Green River. According to Professor Cope, who unearthed the remains, they were taken from a bed of sandstone occupying a stratigraphic position just above the thinner or lower stratum of the Bitter Creek series of coals and over- lain by two other coal seams. This bed of sandstone in which the bones were found “crops out high on the bluffs” lying east of Black Buttes station. Dr. T. W. Stanton, who is well acquainted with the country about Black Buttes, writes me as follows: I can not say that I have ever identified the exact spot from which the bones were taken, but the horizon and the general location within 200 or 300 yards is easily found. Meek, who discovered the bones, and Cope, who collected and described them in 1872, were not very definite in their statements as to the locality, though they gave its exact position in the section. The only natural inference from Cope’s description is that the place is east of the railroad, which for a short distance here has a course almost south. In 1873 Lesquereux visited the locality and gave its position as follows: now called, is at the top of the ridge facing the depot, at a short distance, half a mile, east from it. The débris taken out in digging the bones of the animal is still mixed with a quantity of fragments of those bones, and some of the specimens are remarkably interesting, bearing, as they do, fragments of bones on one side and fossil leaves on the other.”” White evidently overlooked this definite statement when he published his belief that the type locality of Agathaumas sylvestris is probably about a mile south of Black Buttes station and about 100 feet above the principal horizon for invertebrate fossils. ¢ The Black Buttes section was examined by Messrs. Stanton and Knowlton in 1896, who described ¢ it as follows: “The most prominent feature of the section at Black Buttes is the massive bed of sandstone somewhat over 100 feet thick at the base of the exposure, forming steep hills and cliffs northeast of the railroad opposite the station and passing ‘ ‘The Saurian bed, as it is aThe original spelling was Agathawmas sylvestris, later variations by Cope were Agathaumus for the genus and sylvestre and silvestre for the species. b Ann. Rept. U.S. Geol. and Geog. Surv. Terr. for 1873, p. 373. cAnn. Rept. U.S. Geol. and Geog. Surv. Terr. for 1877, p. 223. d Bull. Geol. Soc. America, vol. 8, p. 143. 106 ~THE CERATOPSIA. beneath the surface by its dip of 9 or 10 degrees near the coal mine. The upper portion of it is also exposed on the south side of Bitter Creek Valley, about a mile from the station. All of the Laramie fossils, whether plants, invertebrates, or vertebrates, that have hitherto been described or listed as coming from Black Buttes were obtained from the overlying beds ’ within about 100 feet of the top of this massive sandstone. The original specimen of Agathaumas sylvestris was found about 20 feet above it, and the plants that have been described came from the same horizon and from several higher bands up to the bed overlying the principal coal, some 60 or 75 feet higher. The invertebrates from this locality have about the same range. Most of the beds vary considerably in character and thickness within short distances, but the fossiliferous and overlymg portions of the section may be described in general terms as a series of variable sandstones, clays, and coal beds exposed in low hills and ridges with a dip of 9 or 10 degrees eastward at the base, but decreasing in the upper portions to 5 or 6 degrees, which is about the same as the dip of the overlying Wasatch beds.” DESCRIPTION BY COPE. Cope’s original description of the present genus and species is as follows: # During the present season F. B. Meek, of Dr. F. V. Hayden’s Geological Survey of the Territories, discovered some large bones near Black Buttes station, on the Union Pacific Railroad, 52 miles east of Green River and near the Hallville coal mines. Shortly afterwards I visited the spot with a branch expedition, and commenced excavations with a view to the recovery of the remainder of the animal. The position was‘discovered to be between the thinner or lower strata of the Bitter Creek series of coal, which at this point, occupy a position of elevation and crop out high on the bluffs. _ Two strata appear above the sandstone in which the bones occur and one below it. The portion of the skeleton found rested in the midst of vegetable débris, as sticks and stems, and was covered with many beautiful dicotyledonous leaves, which filled the interstices between the bones. The plant bed gradually passed into a shell bed, containing numerous thin dimyaria, and close by some oysters were found. The whole question as to geologic age and aqueous conditions during which these beds were deposited being unsettled, I gave especial attention to the recovery of the bones, with the view of reaching a definite conclusion on these points. We succeeded in recovering 16 vertebre, including a perfect sacrum, with dorsals and caudals; ‘both iliac and other pelvic bones, those of one side nearly perfect; some bones of the limbs, ribs, and other parts not determined. The vertebre are large. The dorsals are short, with vertically oval centra and small neural canal. The diapophyses originate well above the neural canal, diverge upward, and are triangular in section. The neural spine is very much elevated and the arch short antero-posteriorly. The zygapophyses are close together in both directions, those of the same aspect being separated by a narrow keel only. They do not project, but consist of articular surfaces cut into the solidspine. The latter is flat and dilated distally. The articular faces are nearly plane with a slight median prominence. The ribs have two articular surfaces, but I found no capitular pit on the dorsal centra. Elevation of centrum 7.5 inches; width of same, 5 inches 7.5 lines; length of same, 3 inches 8.5 lines. Total elevation of a dorsal vertebra, 28 inches 3 lines. The sacrum consists of five vertebre, the anterior centrum not depressed. They give out huge diapophyses which are united by suture. They are themselves united distally in pairs, each pair supporting a longitudinal convex articular face for the ilium. Each pair incloses a perforation with the centra. The first diapophysis goes off from the point of junction of the first and second vertebra, the second from the third only, and is more slender. The total length:is 25 inches, and the width 30 inches. Its vertebre are flat below, with latero-inferior angles. The last centrum gives off a simple diapophysis. Another vertebra exhibits a diapophysis as low as the floor of the neural canal and united by coarse suture. Others posterior to the sacrum are more elongate, with slightly compressed centrum and with diapophysis opposite floor of canal and not united by suture. Centra flat below; no chevron bones discoverable. Length of centrum, 4 inches 4 lines; depth of articular face, 4 inches; width of same, 4 inches 3 lines. The iliac bone is extended antero-posteriorly. One extremity is thick and rather obtuse, but of little depth. There is a large protuberance above the acetabular sus. The other extremity is dilated into a flat, thin plate of rather greater length than the stouter extremity. From one of its margins a rod-like element projects. Its total length is about 4 feet, of which the acetabular sinus measures about 8.10 inches. A short bone pertaining to the limbs has the articular surfaces at a strong angle to each other, hence the shaft is twisted. It is deeply grooved on one side near the extremity. The other extremity bears a rather flattened hourglass-shaped articular face, and below it on one angle is a crest. The convexity of the surface is not great, and this extremity resembles that of a dinosaurian or crocodilian reptile. Its length is, however, only 8} inches; apparently too small for a humerus, though this is not certain, while it is decidedly too small for a metatarsal of such an animal. From the above description it is evident that the animal of Black Buttes is a dinosaurian reptile, the characters of the sacral and iliac bones alone sufficing to demonstrate this point. If the reader will compare the measurements given for species of this group already known he will observe that those of the present animal exceed those yet described from North America. It is possible that if the corresponding parts of Hadrosaurus tripos Cope, or Thespesius occidentalis Leidy, are discovered, they may approach it. It is thus conclusively proven that the coal strata of the Bitter Creek Basin of Wyoming Territory, which embraces the greatest area yet discovered, were deposited during the Cretaceous period, and not during the Tertiary, though not long preceding the latter. It appears that the forests that intervene between the swamps of epochs during which the coal was a Proc. Am. Philos. Soc., vol. 12, pp. 481-483. eel AGATHAUMAS SYLVESTRIS. 7 formed were inhabited by these huge monsters; that one of them lay down to die near the shore of probably a brackish- water inlet, and was soon covered by the thickly fallen leaves of the wood; that continued subsidence of the level submerged the bones, which were then covered with sand. The form of the ilium differs very materially from that of Hadrosaurus and the vertebre are plane, thus differing from Thespesius. The limb bone is distinct from anything in Lelaps, which, moreover, probably resembles Thegalosaurus in its ilium. The present form recalls rather Cetiosaurus. As it is evidently new to our system, it may be called Agathaumas sylvestris. In his Cretaceous Vertebrata Cope describes and figures the type in detail. His description is as follows: ¢ The characters of this genus are derived from the typical species A. sylvestris, which is represented by dorsal and lumbar vertebre and an entire sacrum, with the ilia, one nearly entire, ribs, and a number of other bones the characters of which have not yet been positively ascertained. One of these resembles the proximal part of a pubis, others portions of the sternum, etc. On eight (and perhaps nine) vertebre, anterior to the sacrum, there is no indication of the capitular articular facet for the rib. This facet is found, as in Crocodilia, at or near the base of the elongate diapophyses. The centra are slightly concave posteriorly, and still less so on the anterior face, with gently convex margins. The neural canal is very small, and the neural arch short and quite distinct from the centrum, having scarcely any suture. The neural arch has a subcubical form, partly truncated above by the anterior zygapophyses. In like manner the base of the combined neural spine and diapophyses are truncated below by the square-cut posterior zygapophyses. The diapophyses are long, and directed upward; they are triangular in section. There are eight (and perhaps nine) sacral vertebrze, which exhibit a considerable diminution in the diameters of the centra. The diapophyses and neural arches are shared by two centra, the anterior part of a centrum bearing the larger por- tion of both. The diapophyses are united distally in pairs, each pair inclosing a large foramen. The anterior is the most massive and rests on the ilium; the posterior pair the most expanded; the superior margins of its posterior edge form an open V, with the apex forward on the neural arch of the fifth vertebra. On the last sacrals the diapophyses rise to the neural arch again. The exits of the sacral spinal nerves are behind the middles of the centra, and continue into grooves of the sides in all but the last vertebra. The reduced and rather elongate form of the last sacral vertebra induces me to believe that this animal did not possess such large and short caudal vertebre as are found in the genus Hadrosaurus, and that the tail was a less massive organ. The ilium is much more elongate than the corresponding element in Hadrosaurus, Cetiosaurus, or Megalosaurus. Its upper edge is turned and thickened inward above the anterior margin of the acetabulum, and here the middle of the conjoined diapophyses of the second and third sacral vertebree was applied when in place. In front of this point the ilium is produced in a straight line and a stout flattened form with obtuse end. Posterior to it its inner face is concave to receive the second transverse rest of the sacrum, and the superior margin is produced horizontally toward the median line like the corresponding bone in a bird. The posterior part of the bone is the widest, for it is expanded into a thin plate and produced to a consider- able length. From one of the margins (my sketch, made on the ground, represents it as the upper) a cylindric rod is pro- duced still farther backward. This it is believed is only the shaft of a displaced rib. The base of the ischium is coossified with the ilium and is separated behind its base from the iliac portion of the acetabulum. There is no facet nor suture for the pubis at the front of the acetabulum. The ribs are compressed. There are no bones certainly referable to the limbs. The form of the ilia distinguishes this genus from those known heretofore. The last nine dorsal vertebre have rather short centra, the most posterior the shortest. They are higher than wide; the sides are concave, the inferior face somewhat flattened. The neural arch is keeled behind from the canal to between the posterior zygapophyses, and a similar keel extends from the base of the neural spine to between the anterior zygapophyses. The neural spine is elevated and compressed, the diapophysis is convex above and concave along the two inferior faces, most so on the posterior. The articular face of the first sacral vertebra is wider than deep. The eight sacral vertebra are flat- tened below, in all except the first, by a plane which is separated from the sides by a longitudinal angle. The neural spines of the anterior five sacral vertebrae are mere tuberosities. A large sutural surface for attachment of a transverse process is seen on the posterior third of the eighth sacral vertebra, which descends nearly as low as the plane of the inferior surface. On the ?tenth sacral there is no such process, but its neural arch and that of the ?ninth support transverse processes. These are more Jike those of the dorsals in having three strong basal supporting ribs, the anterior and posterior extending for some distance along the arch. Hither naturally or in consequence of distortion, the plate of the ilium is at a strong angle to the vertical axis of the acetabulum, and at the posterior part of it its plate presents a free margin on the outside as well as the inside of the femoral articulation. a Pp. 53-56; also Bull. No. 2, U.S. Geol. and Geog. Sury. Terr., pp. 17-19. 2 EE eR ee Sy eee tere ee ee Sa SS 108 THE CERATOPSIA. Measurements. Mm Wengthyotthemine posteriori donsallinve nic bias sae eee aes eee ee 0. 880 henge thyolthemineysacraligvente breed (Oj nt cles) ene ere ales ea . 930 © Menethrotriehtalimme 2ipiecessO's4—Ol22) rq (Galena ches) seep ae 1. 060 Length of eighth dorsal from the sacrum._.........-.-..-- Bee ers estes eae eee wa Saad = . 090 Gene thyoti the base olebhe me uray Op lity.s ls see eee are . 085 Depth of the articular faces) 22.5 sos sm lee ht toes te See ee ee a . 153. Widthcof the articular faces: sass 22 bible. Seer ee neg BE aed a ae . 123 Iheng thos the second tromys aerials eae) eevee ieee eee ede . 070 Depth: of the articular'tace! 252) - re se eee ee te oe seine Tae taka te epee aise ee ee ae WES5) Widthvolithevarticullartta ce = see se eee eee Re ae: PERL Ie re Mees ee SB clei 137 Elevation of the neural canal e222 72 <2 Syaeesne eye ee Ne ee ee 045 Widthvot the:meurall:camsl: 2205 3 oo secei se 2 eos SS oe 2 arp Seen eee ele At ne en ee 028 Mlevationotithertaceror therzy cap op liv Ses see eee re eterna ea eae ee . 104 Mlevationtor thebase of theimeurall isprme sya) segs ee nt mete ey Seis eer ol Ae le ge . 150 Wengthvoiuwercdiap opliysis stro nmmt hey O wera ase ems ae ee =e aye eee eee ee ea eno een . 200. Length of the diapophysis from the capitular articulation._.________._.......-...-.-.--------- - 125. ‘Antero=posterior widthiabovels. 52 Sui soe Sess SaaS sh ey tee ee een le ne eC - 050 Antero-posterior width of the base of the neural spine.-....--___-.--__-------_---------------- . 070 PATHE ENRCO= |) OSHELLO Tawa Ct nm eat mt Loe W297 20 | OTD LAGS eee a en . 078 Wena thyofsthenmeurallts pier (ie moe nity) pees eee ae ar ees . 200: Wiadthvoficenbrummao fe they fins tits acre teem pe aye eset ey are eye ee so ne rea . 160. Depth of the centrum of the first sacral (to the neurapophysis)...-......----------------------- . 145 hens thy ofathexcembruiaare far ulae tits isso Crt beeen ae a . 100 Wengtb) of the centrum of the seventh sacral ets) ees =e. amen sey see an eee eae eee . 100 Depth of the centrum of the}seventhysaerall (belt) eae se a eee eee 085 Width of the centrum of the seventh sacral (behind)-_.._.-...-._._..--.-.---.-.-.-_-.-.-.--- . 100 EXxpanse of the second sacral transverse support (22 inches)......---.-.----------------------- . 560 Wenethvofithe nlm: nite ortonp hey sice ta bo vil vi eee aes ae ea - 470 Length ofthe acetabulumc2-—: oh ec ee NE eae oN) Mengthvor thehlrm posterior: tothe se: ce telnet ae seen ean en . 390 Whidthiotabhre ilirumayerta thre erate rt ore xctaice innit ese eae een ee . 140 Wadthvor the tumararbEtley trait otab bere cet efor OT ee aa . 210 Wadthwot thentlinma¥ertat be sp Oster ory ex. 210s [0 21a aa . 250 Mhickness:abovexthevacetalbulunns. 2s ses Se tee ete ap eee yee ee Se ee ee . 060 Widthvofsthe acetabullumazss 2. ae fee eves = 2 ete ere eee eer wee meee eae Ra ee AUP eyes 105 ‘Width: of the basesjofithe-ischiwinm secon ss Sele a oc ees a egal el apt ae a 085 Width of the'shaft of a.ribi:cc,S. 82 ee See SPS se = Sve Se ve Stee a ean ea 062 Other bones, not yet determined, will be included in the description in the final report. This species was no doubt equal in dimensions to the largest known terrestrial saurians or mammals. MATERIAL NOW AVAILABLE. Of this material there remains accessible only the right ilium and a mere fragment of the: anterior extremity of the left, one rib nearly complete but lacking the capitulum and the distal extremity, a proximal portion of a second rib without either capitulum or tuberculum. The sixteen vertebre mentioned by Cope consist of the nine posterior dorsals, for the most part: represented only by centra. There are, however, three complete but detached neural arches. One of these still supports the transverse process of the left side and this shows at its extremity the tubercular rib facet, while the capitular rib facet may also be seen on the inferior border, near the point where the transverse process unites with the neural arch. There is also pre- served a left one-half of a fourth neural arch, which likewise bears a transverse process, com- plete except at the very extremity, where the tubercular facet is wanting. In the second from the anterior dorsal of the series figured by Cope the neural arch is still in position. It shows only a sutural union with the centrum, being held in position by the inclosing matrix. Both transverse processes, as well as the neural spine, are wanting in this and the three pos- terior dorsals figured by Cope. The sacrum which he has described as being perfect and. consisting of five, sometimes eight or perhaps nine, vertebre is also present. The three centra of the anterior caudals or sacro-caudals are present in identically the same condition as figured. — ee AGATHAUMAS SYLVESTRIS. 109 by Cope, as is also the fragmentary coossified neural arch shown by him in PI. IV, fig. 18, of his Cretaceous Vertebrata. The undetermined bones shown by him in PI. IV, figs. 19, 19a, and 20, are also present. No pelvic bones, other than the ilium mentioned above, and no limb bones pertaining to the type are to be found, and since they have nowhere been described by Cope it is doubtful whether they were ever recovered except, perhaps, in an extremely frag- mentary and hopeless condition. DETAILED DESCRIPTION. The ilium.—This, after the sacrum, is the most important element of the original type. The right ilium is present and complete save the anterior and posterior extremities and a portion of the pubic peduncle, which are wanting. The blade of the ilium is much extended, both anterior and posterior to the acetabulum. The anterior extremity appears broad and somewhat truncated when compared with the pointed, narrow, and elongate posterior extremity. When adjusted to the sacrum the expanded blade of the ilium is more nearly horizontal than vertical, just the opposite from that which obtains in the sauropod dinosaurs and more closely resembling that of the Mammalia. In this position the external edge, when viewed from above, presents a compound curve, gently concave ante- riorly and more decidedly convex throughout the posterior two-thirds of its length. When viewed from in front the external border continues in the same horizontal plane until just above and a little posterior to the ischiac peduncle, when it bends abruptly downward. This, the external margin, is moderately thick throughout its entire length, but it is greatly thickened near the posterior and anterior extremities and in the deflected region just above the ischiac peduncle. The broad superior surface of the ilium is convex transversely throughout its middle region, but concave at either extremity; antero-posteriorly it is gently convex through- out its entire length. The internal margin is rather thin at either extremity, but considerably thickened and otherwise modified medially for contact with the sacrum. The inferior surface of the ilium gives origin to the ischiac and pubic peduncles, which spring from near its internal margin. From the base of the pubic peduncle a broadly convex ridge of bone extends diago- nally across the anterior blade of the ilium to its antero-external angle, inclosing in front the deep concavity which occupies the inferior surface of this bone just external to the acetabulum. The inferior surface of the posterior blade is a nearly flat surface. The ischiac peduncle is low, broad, and much expanded both transversely and antero-posteriorly. The articular sur- face of the pubic peduncle is wanting; that portion of the base remaining indicates, however, that this element, although more slender, was actually longer than the ischiac peduncle. In the Ceratopsia the ilia are never coossified with the sacrum, and in the present specimen the surface for attachment with the sacrum is but poorly indicated. In Cope’s description of this ilium he mistook the anterior for the posterior extremity, erroneously described the base of the ischium as coossified with the ilium, and stated that there was no facet for the pubis. Some of these errors he corrected after the descriptions of more complete material by Professor Marsh. The vertebre.—The nine posterior dorsals figured by Cope are all accessible. These, save the second from the first of the series, are represented only by the centra, for the most part in a much damaged condition. The exact association of the detached neural arches mentioned above is uncertain. With the second, held in place by the surrounding matrix, there is the neural arch, but without the spine or transverse processes. This arch, as well as those of the other vertebr represented in the series, were united only by suture with their respective centra, thus indicating that the individual had not yet reached maturity. All the dorsal centra are short, deeper than broad, plano-concave or slightly biconcave, constricted medially and obovate in cross section. The neural canal is small, its vertical diameter exceeding the transverse, and is nearly inclosed by the neural arches, but inferiorly it is bounded by the superior surfaces of the centra. The neural arch is rather high, rising about 85 mm. above the neural canal before giving origin to the transverse processes. Of the detached neural arches, each bearing a transverse process, mentioned above as still associated with the type material, one has been a —— 110 THE CERATOPSIA. referred by Cope to the eighth and another to the ninth of the series. These transverse proc- esses are triangular in cross section and bear, near their union with the neural arches, capitular rib facets, while the one best preserved shows at its extremity a portion of the tubercular rib facet. The capitular facets are each supported inferiorly by a single strong vertical lamina or buttress, which springs from the posterior lateral border of the neural arch. The zygapophyses on either side are placed very close together, both anterior and posterior, those of one side being separated from those of the other by a very narrow median ridge or keel. The articular surface of each occupies a single horizontal plane, the anterior looking upward, the posterior downward. The remarkable simplicity exhibited in the articulation of these vertebre con- trasts strongly with the complicated condition which obtains in the same region of the verte- bral column of the Sauropoda. I can not understand Cope’s reason for describing the neural arch as ‘‘short and with scarcely any suture,’ save that he meant antero-posteriorly. The sacrum.—The sacrum as figured and described by Cope is in a fair state of preserva- tion. The three sacro-caudals figured by him as Nos. 15, 16, 17 of the vertebral series are evidently posterior sacrals or sacro-caudals and represent 7, 8, and 10 of the functional sacrals, 1 and 9 being wanting, not having been preserved. In one of these the neural arch is preserved, but in the other two only the centra remain, though fortunately in a better state of preserva- tion than are those of the dorsal region. They are somewhat more elongate than the dorsal centra and have the vertical and transverse diameters more nearly equal. The inferior surface of each is broad and slightly concave. Superiorly they expand laterally at either extremity in order to give support to transverse processes, which, however, are wanting in the present specimens. These centra are not so regularly constricted medially as are those of the dorsals, and this is especially true of that centrum numbered 17 in Professor Cope’s figures. The neural arch is compressed and low when compared with that of the dorsals. The two coos- sified neural arches referred to and figured by Cope as sacrals are present and appear to have occupied a position in the vertebral column similar to that shown in Cope’s figure. I believe they pertain to two of the posterior sacrals or sacro-caudals. The five true sacrals are present and decrease regularly in size from the first to the last, as shown in Pl. XXV. The centra of the first and second of these vertebre are comparatively broad and short; those of the three posterior are more elongate. All the sacral ribs or trans- verse processes spring more directly from the union of the centra in the present sacrum than in that of Triceratops, as will be seen from a comparison of the figures, a condition apparently somewhat intermediate between that which obtains in Monoclonius and Triceratops. In the latter genus, as has already been pointed out, only the anterior of the sacral ribs takes its origin about equally from two vertebre, the three posterior pairs of sacral ribs taking their origin almost entirely from the anterior half of a single centrum. The sacral ribs are strong and they unite distally to form a strong bar, which articulates with the ilium and forms a portion of the acetabular wall. Together the four sacral ribs inclose three large foramina. It would — perhaps be better to consider only the four posterior of these vertebre as true sacrals and treat the anterior as a sacro-lumbar. It is probable also that had the animal been fully adult there would have been attached to this still another sacro-lumbar, as in the sacrum of Tricera- tops. The total number of vertebre united in the sacrum would then be 10, as in the last- - mentioned genus. There are represented in the type 18 instead of 16 vertebrz, as stated by Cope, and his figures show 17 centra and portions of 18 vertebree. The undetermined bone shown by Cope in figs. 19 and 19a, Pl. IV, of his Crevhedeue Verte- brata I believe to be a lateral metapodial, shove) somewhat oerened sind distorted. The ribs.—Only portions of two ribs pertaining to the type specimen are to be found. One of these consists of a fragment about a foot long, representing the proximal portion of a rib from the mid-dorsal region, with both the head and tubercle wanting. It is much-flattened proximally, but becomes triangular in cross section beyond the tubercle. The other rib is nearly complete and is from the posterior region, though not the last of the series. The head and extreme distal end are both wanting. It is flattened proximally, becoming subovate beyond the tuberculum and subelliptical toward the distal extremity. POLYONAX. 111 Principal measurements of the type. Mm Greatestalensthyotportionvotmilimmypresenved sere fears eer a pe ae eee ee sk ae 1,142 Hsiimaredslenrt hwo tatliumewihenkcom~plete sass epee amas eee ee eee ee oe 1, 392 Antere-posterior expanse of ischiac and pubic peduncles. ....-..-..-.--.--.---------+--------- 342 Greatest transverse diameter of ischiac peduncle. 222.5... ).- 2242-22-22 45522------25. 22-2 +-2 ee 150 Distance from superior border of ilium to extremity of ischiac peduncle......_.......-.-----.--- 212 iransverserdiamererot erchthudorsal tromysacrume se eyes eee eee ees es ane 132 Weniioall ciammaier Or saints 5.52 sy eB a AS A ROE ae aca sae Seat Se ar eS 167 Length of VET a els SE SS a, Bes Mie FS ON OO UP teh CHR Ope a a me 91 Height of anterior zygapophyses above suture for centrum.....-...-.------------+------------ 111 Height of base of neural spine above suture for centrum.-...........-.-.-----/--+---+5------- 155 Distance from capitular to tubercular facets........-...-----:------ Rehan Mae ella ee 130 The above description of the type of Agathawmas sylvestris, together with the original and later descriptions by Professor Cope, makes it clear that the generic and specific distinc- tions must rest on such characters as are to be found in the ilium, sacrum, posterior dorsals, and ribs, and it will be well to, bear this point in mind when we come to a discussion of the synonymy, which will be taken up in that part of the present volume devoted to a revision of the genera and species. AGATHAUMAS MILO Cope. 1874. Type (No. ?, American Museum of Natural History) consists of a sacral centrum and fragment of proximal end of tibia, no longer determinable. Original description Bull. U. S. Geol. and Geog. Surv. Terr., vol. 1, No. 1, 1874, Ist ser., p. 10, footnote. Cope, E. D., Bull. U. S. Geol. and Geog. Surv. Terr., vol. 1, No. 2, Ist ser., p. 21; Rept. U.S. Geol. Surv. Terr., vol. 2, 1875, p- 58. Nopesa, F. Baron, Féldtani Kézlony, Budapest, 1901, vol. 31, p. 270. In a brief footnote on page 10 of Bulletin No. 1, first series of the Bulletins of the U.S. Geological and Geographical Survey of the Territories, published in 1874, Professor Cope names, without any description whatever, this and a number of other supposedly new Mesozoic Reptilia. His remarks in this connection are as follows: In examining a collection from this formation, made by one of my assistants, I find a series of Mesozoic genera of verte- brates as follows: Dinosauria, Cinodon [a misprint for Cionodon|] arctatus, gen. et sp. nov.; Polyonax mortuarius, gen. et sp. noy.; Agathaumas milo, sp. nov.; * * * On page 21 of Bulletin No. 2, first series, United States Geological and Geographical Survey of the Territories, Professor Cope identifies the very fragmentary material constituting the type of the present species as pertaining to Hadrosaurus occidentalis. His remarks in this connection were as follows: Remains of species of Dinosauria were obtained at two localities in Colorado not many miles apart, the greater number at one of them, from which also the crocodilian and turtle remains were derived. Those from other deposits consist of portions of limb bones apparently of a single individual of gigantic size. The more abundant fragments are referable to three species. A fragment of limb bone is very similar to portions from the other locality, and associated is a sacral vertebra of appropriate size and characters. All of these were therefore referred provisionally to a single species under the name of Agathawmas milo, but are here described under Hadrosaurus occidentalis. It is thus clear that, according to Cope, Agathawmas milo Cope became a synonym of Hadrosaurus occidentalis Leidy. Whether or not this latter determination was correct the fact remains that A. milo is a nomen nudum, and the fragmentary nature of the type is such as to preclude an adequate description. The species should therefore be discarded. POLYONAX Cope. 1874. Type species P. mortuarius. Original description Bull. U. S. Geol. and Geog. Surv. Terr., vol. 1, No. 1, 1874, p. 10, footnote. Baur, G., Science, vol. 17, 1891, p. 217. Cope, E. D., Ann. Rept. U. S. Geol. and Geog. Surv. Terr. for 1873 (pub. 1875), pp. 448, 451-452; Bull. U.S. Geol. and Geog. Surv. Terr., vol. 1, No. 2, 1st ser., 1874, pp. 7, 21, 24; Rept. U.S. Geol. and Geog. Surv. Terr., vol. 2, 1875, pp. 58, 63-64; Am. Naturalist, vol. 22, 1888, p. 1109; Science, vol. 13, 1889, p. 290; Am. Naturalist, vol. 23, 1889, p. 715. alte THE CERATOPSIA. Dana, J. D., Manuai of Geology, 1895, p. 847. Lambe, L. M., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, p. 68. | Marsh, O. C., Am. Jour. Sci., January, 1892, p. 83; December, 1895, p. 497; Sixteenth Ann. Rept. U.S. Geol. Survey, 1896, Peal ; pt. 1, pp. 217, 243. wan Nicholson and Lydekker, Manual Pal., 1889, vol. 2, p. 1163. 1 Hi Nopcsa, F. Baron, Féldtani Kézlony, Budapest, 1901, vol. 31, p. 270. P| Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, 1902, pp. 9, 20. a Woodward, A. S., Outlines Vert. Pal., p. 213. a Zittel, K. A. von, Text-book Pal., trans. by C. R. Eastman, vol. 2, p. 244. eat This genus, like the last-mentioned species of Agathawmas, was proposed but not described eave in the footnote a portion of which is quoted above.* The genus is described as follows: Char. gen.—A species considerably larger than the last (Cionodon arctatus), represented by vertebre and numerous fragments of limb bones. The most characteristic of the former are two probably from the posterior dorsal region, which are somewhat distorted by pressure. The more anterior is shorter than the other and exhibits both anterior faces slightly concave, the one more so than the other. They are higher than wide and the border is scalloped above for the capitular articulation for the rib. There are numerous nutrient foramina and some ligamentous pits on the articular surfaces. The inferior face is rounded. In the larger vertebra both faces are more strongly concave, and at each end of the lower side there is an obtuse hypopophysial tuberosity. The sides of the centra of both vertebra are concave. The neural canals are relatively small and the neurapophyses coossified. A third vertebra without arches is similar in specific gravity, though without the white surface layer of the others. It is appropriate in size and form to this species, and is peculiar in its flat form, resembling the anterior dorsals of the Hadrosaurus. In this respect it is related to the shorter vertebra of the two above described as the latter is to the longer. The surface of the posterior articular surface is damaged; it was not concave, and is now slightly convex; the anterior is preserved and is concave. Only one species of Polyonax has been described, namely, P. mortuarvus Cope. This will now be noticed. POLYONAX MORTUARIUS Cope. 1874. Type (No. 3950, American Museum of Natural History) consists of fragments of horn cores, vertebre, etc. Original description in Bull. U.S. Geol. and Geog. Surv. Terr., vol. 1, No. 1, Ist ser., 1874, p. 10, footnote. ‘Cope, E. D., Ann. Rept. U.S. Geol. and Geog. Surv. Terr. for 1873 (pub. 1875), pp. 448, 451-452; Bull. U.S. Geol. and Geog. Sury. Terr., vol. 1, No. 2, 1st ser., 1874, pp. 7, 21, 25; Rept. U.S. Geol. Sury. Terr., vol. 2, 1875, pp. 26, 59, 64-65; Am. Naturalist, as in Triceratops, but posteriorly this gives origin to two deep channels, one on either side. These run backward along the surface of the parietal and terminate in two small circular fontanelles, conditions very similar to those which obtain in Torosaurus. Measurements of the type. Mm Distance from anterior end of rostral to posterior of squamosal...-.--..-.-..------------------ 1,990 Distance from anterior end of rostral to anterior of orbits ---2..---.-------.-.----------------- 845 Distance from inferior border of orbit to lower end of jugal...............-+-..---------------- 363 Distance from posterior border of nasal opening to extremity of beak......-.....-..--.--------- 5 it! Distance from posterior border of orbit to posterior surface of horn core... -....---------------- 175 IDISHATICeRDE LW EenNamLer Oran OnGers| OLN Onis mere ent ee Sete = eee ae eee 340 Circumference otsupracubitalihormicoresyat)base)=-22 2-42-4222 -- -2--- 45-22-52 2sh acess. 610 Circumference of supraorbital horn cores 200 mm. above orbit. -..-.-.-.----------------------- 340 Wentiee!! chimgior ©! OHSS. bags se octie Sea ecuree oe OBE Bae Sa See ee ae 165 FAMUENOs DOSUCHIOMGIALE DEIN OIMOLDIUS awe anes pie eae See cis gee hae ete sce elec aes 125 TOROSAURUS Marsh. 1891. Type species, 7’. latus. Original description in Am. Jour. Sci., 3d ser., vol. 42, 1891, p. 266. Marsh, O. C., Am. Jour. Sci., 3d ser., vol. 43, p. 81; Nature, vol. 48, 1893, p. 438; Am. Jour. Sci., 3d ser., vol. 50, p. 497; Six- teenth Ann. Rept. U.S. Geol. Survey, 1896, pt. 1, p. 214. Dana, J. D., Manual of Geology, 1895, p. 847. Lambe, L. M., Contr. to Canadian Pal., vol. 3 (quarto), pt. 2, p. 66. Osborn, H. F., ibid., pp. 9, 14, 15, 20. Zittel, K. von, Text-book Pal., trans. by C. R. Eastman, vol. 2, p. 245. a This form was described by Hatcher, but left without a name. Diceratops is suggested as a generic name, from the absence of the nasal horn core, while the appropriateness of the specific name is self-evident.—R. 8. L. b See footnote on p. 24.—R. S. L. cSee footnote on p. 163.—R. S. L. *- 150 THE CERATOPSIA. The present genus appears to be distinct from any of the others that have been proposed for these larger members of the Ceratopside, either from this or other localities. The more important differences are to be found in the parietal and squamosal regions of the skull. In the present genus that portion of the skull posterior to the frontal horn cores is considerably elongated, while the anterior portion is correspondingly abbreviated and more compressed than in Triceratops. The most prominent and distinctive characters, however, are to be found in the parietal crest, which is pierced by a pair of large foramina, situated one on either side of the median line, as shown in fig. 12, and designated as the supratemporal fontanelles or parietal foramina. The squamosals are also characteristic, being long and slender, assuming somewhat the form and pattern of a broadsword, instead of short and broad, as in Triceratops (see fig. 14; 1, 2, 3). As already remarked, the parietals in the present genus most nearly resemble the same clements in the type of Monoclonius (M. crassus), and there seems little doubt that the last-mentioned genus was ancestral to Torosaurus.% TorosauRgus LATuS Marsh. 1891. The type (No. 1830, Yale Museum), from the Laramie of Converse County, Wyo. Original description in Am. Jour: Sci., 3d ser., vol. 42, September, 1891, p. 266. Marsh, O. C., Am. Jour. Sci., 3d ser., vol. 43, p. 81; Sixteenth Ann. Rept. U. S. Geol. Survey, 1896, pt. 1, p. 214. Lambe, L. M., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, p. 66. Osborn, H. F., ibid., p. 14. LOCALITY. ‘ The type (No. 1830, Yale Museum) was discovered at the locality marked +19, Pl. LI. It lay embedded in an extremely hard bluish calcareous sandstone concretion near the summit of the bluff on the north side of Lightning Creek, about 2 miles above the mouth of that stream, in the bottom and near the extreme head of a small, dry watercourse emptying into Lightning Creek. When discovered the summits of the frontal horn cores and most of the parietal crest had already weathered out and been carried away by the current of water which at times filled the channel of the small watercourse, due either to occasional heavy rains or the melting of the winter’s accumulation of snow. Associated with the skull in the same concretion there were, in considerable numbers, plant impressions and small lignitized stems. ORIGINAL DESCRIPTION. Professor Marsh’s description of the type of the present species was as follows: Another well-marked species of this group, which may be referred to a new genus, is represented by one skull, and parts _ of the skeleton, from nearly the same horizon as the specimen above described. One of the most striking features of the present species is seen in the posterior crest, which, instead of being complete, as in the skulls hitherto found, is perforated by a pair of large openings. These are in the parietals, but they have the inner margin of each squamosal for their outer border. They are well behind the supratemporal fossz, but doubtless were originally connected with them. They may be called the supratemporal fontanelles. The squamosal bones, moreover, are very long and slender, and distally only show near the ends sutures for union with the parietals. Another distinctive character is seen in the main horn cores, which are placed well back of the orbit. The nasal horn core is short, with the apex compressed, and directed forward. This genus is of much interest, as it represents an earlier and less specialized form than either Ceratops or Triceratops, both of which have the posterior crest complete. The existing chameleons show the other extreme, where the outline only of the parietal crest has been attained. Some of the principal dimensions of this skull are as follows: Inches. Length from apex of nasal horn core to extremity of squamosal.__._......----------------------- 80 Distance from same apex to front of orbit---- - PE eae gt eee Senne AE ey eh Mele Te ee SS Hi Goh < 21 Distanceriro nay same) foc O iy .O legjo sual e tral cop ra ees ee en 54 Width between posterior extremities of squamosals._--__ 2 =) 2-222 2-22 28 ee ee 56 This important specimen was discovered by Mr. J. B. Hatcher in the Laramie of Wyoming. a This section of the manuscript is as originally written, not having been revised by the author, who later changed his views with regard to the ancestry of the genus (see p. 100).— R. 8. L. TOROSAURUS LATUS. 151 DESCRIPTION OF THE TYPE. The type of the present genus and species consists of a skull (No. 1830, Yale Museum) without the lower jaw. The posterior portion of the parietals, the summits of the supraorbital horn cores, the rostral, premaxillaries, and a considerable portion of the maxillaries are also wanting. . The frill or posterior portion of the skull is greatly expanded, both transversely and antero- posteriorly, and was pierced by two large supratemporal fontanelles. Only the antero-external border of the left one of these fontanelles is preserved in the type, so that it is not possible to determine with certainty either their form or dimensions. Marsh has figured them as elliptical, and his figure is reproduced here as fig. 118. According to Marsh the squamosals form the postero-external borders of the fontanelles, though I am unable to tell, after an examination of the type, whether or not this is the case. It may be that, as in the fol- lowing species of this genus, these fontanelles were entirely inclosed by the parietals. In the present genus the parietals form a considerably greater portion of the frill than in Triceratops, and the squamosals are correspond- ingly smaller, being reduced to long, narrow, blade-like bones on the external mar- gins of the parietals, very narrow behind but broader in front. The skull is ex- tremely broad and massive between the orbits and the anterior extremities of the squamosals, while in front of Fic. 118.—Dorsal aspect of skull of type of Torosaurus latus Marsh, No. 1830. Yale Museum. the orbits the facial region Onestyentieth natural size. ec, Supra-temporal fossa; c’, anterior temporal foramen‘ é f’, parietal fontanelle; h, supraorbital horn core; h’, nasal horn core; p, parietal; s, squa- narrows rapidly and appears mosal. After Marsh. very short and compressed when compared with the broad and elongate posterior crest or frill. In front the parietals pre- sent two large irregular-shaped fossz, situated one on either side, which are doubtless connected by foramina with the large cavities in the postfrontals at the bases of the horn cores. Just within and near the antero-internal borders of these two large fosse two foramina pierce the parietals. From the external openings of these two foramina two broad, shallow grooves run forward and upward on the surface of the skull, rapidly converging and terminating in two foramina, separated by a thick partition of bone, which marks the median line of the skull. These two foramina are doubtless homologous with the single median one usually found in this region in skulls pertaining to the genus Triceratops and known as the postfrontal foramen or pineal foramen?’ of Marsh. Owing to the age of the individual, the sutures in this region are so much obliterated that it is quite impossible to determine whether these foramina are located in parietals or postfrontals. In the type of the following species, which pertains to a younger but larger individual, they appear to have been located entirely - within the parietals. This being the case, it would not be technically proper to call them post- frontal foramina, even though they are homologous with that element in the genus T'riceratops. a See footnote on p. 24.— R.S. L. 152 THE CERATOPSIA. The squamosals are very long and slender and are united with the parietals only by an open suture. At their anterior extremities the squamosals are firmly coossified with the post- frontals and the jugals. The inferior and anterior angle of the squamosal is not so much produced ~ asin Triceratops, and the quadratojugal notch is not so deep. The squamosal is more expanded in front of the groove for the quadrate in the present genus than in Triceratops. The transverse diameters of the quadrate at or in front of the groove for the quadrate and just behind the quadratojugal notch are subequal in the present genus. The frontal region is broad and the supraorbital horn cores are well separated at their bases; they are much compressed, ovate in cross section, with apex of oval directed anteriorly. The horn cores are directed upward, outward, and forward. The orbits are elliptical in outline, with the longer diameter nearly horizontal and not nearly so vertical as in those species of Triceratops which have elliptically shaped orbits. The orbits in the present genus occupy a rather more anterior position than in the genus just mentioned. — The sutures are so completely closed in this region that it is impossible to determine the extent to which the various elements enter into the construction of the orbit. The lachrymals, pre- frontals, and frontals are so fused with one another and with the postfrontals that their outlines can not be distinguished. A very large round foramen pierces the prefrontal on the right side just within the strong orbital border. There is no corresponding foramen on the left side, and the presence of this foramen of the right side is therefore probably of pathologic origin and of no specific importance. Beneath the orbits the jugals are much expanded and the external surface of the skull in this region slopes downward and outward and is much less nearly vertical than in the genus Triceratops, making the skull appear extremely broad when seen from above. In front of the orbits the facial region contracts very abruptly and appears short and much compressed when seen from above, with a short nasal horn core, broad at the base and sharp above and directed upward and forward. The nasals are completely fused with one another and with the adjoining elements. There is a large lachrymal foramen, situated well in advance of the orbit, but owing to the age of the individual it is impossible to tell whether it is situated entirely within the maxillary, as in Triceratops (Sterrholophus) flabellatus, or between that bone and the nasal, as in Triceratops prorsus. The superior surface of the skull in front of the paired supraorbital horns presents a gently concave fore-and-aft surface, while behind these horn cores the post- frontals fall away rather abruptly to meet the parietals and squamosals. Principal measurements of the type. Mm Distance from posterior border of orbit to extremity of squamosal _-._..-.----.--.------------- 1,640 Distance from anterior border of orbit to front of nasal horn core..........-.------------------- 440 Expanse of jugal ss... .2 See eee ee I nee aR aaa he eee ee oP 810 Greatest: length: of skulls(estumated) = 25 aps = oe seve te ete ae a pe 2, 200 Fore-and-aft diameter of supraorbital horn core at base_.......-_....------------------------- 266 Transverse diameter of supraorbital horm core at base..........-...--------------------------- 153 Transverse diameter of supraorbital horn core 6 inches above base.-.-.-_..--------------------- 175 Fore-and-aft diameter of supraorbital horn core 6 inches above base___.--.-.-----.------------- 108 Circumference of supraorbital horn core 6 inches above base_..--.-.--------------------------- 490 Circumferencerohsupraorbital§horayicore siti aSc see eae ae ee 638 Heightvor masalshornsalb oy emvaler, O19 OG CTC lea eis ell se eae eee 148 Transverse diameter of nasals at base of nasal horn core..._.-..-...-.-..-+2---2-2-:------------ 90 Distance’ betweem orbits iss or ae Soles wees Oe eee ee een GL sii a ee ea pe ee eo ris co aaa 432 Antero=posterior diameter lof orbibe See fe 5s Ae or Fe At le Aang Deron yee ee 166 Vertical diameter of orbit... iss 225222552 oes See ee ae oe = Sas Sa ee 119 ToROSAURUS GLADIUS Marsh. 1896. Type (No. 1831, Yale Museum), from the Laramie of Converse County, Wyo. Original description in Am. Jour. Sci., 3d ser., vol. 42, Sept., 1891, pp. 266-267. Marsh, O. C., Am. Jour. Sci., 3d ser., vol. 43, p. 84; Sixteenth Ann. Rept. U. S. Geol. Survey, 1896, pt. 1, p. 215. Dana, J. D., Manual of Geology, 1895, p. 846. Lambe, L. M., Contr. Canadian Pal., vol. 3 (quarto), pt. 2, p. 66. Osborn, H. F., Contr. Canadian Pal., vol. 3 (quarto), p. 15. EE TOROSAURUS GLADIUS. 153 The type (No. 1831, Yale Museum) of the present species was found at the locality marked +19 A, Pl. LI. The horizon was considerably lower than that which afforded the type of the preceding species. The fragmentary remains upon which it was based by Professor Marsh were found in a thick bed of shale which forms the gentle northern slope of the low divide between Lightning and Cow creeks, midway between and at a distance of about 1 mile from the mouths of those creeks. Professor Marsh’s description of the species was as follows: A second species of apparently the same genus is represented by various portions of a skull in good preservation. In this specimen the nasal horn core is short and obtuse, and nearly upright. The main horn cores are elongate, oval in outline, and in position resemble those of the skull above described. The most remarkable features in the present specimen are the squamosal bones, which are greatly elongated and so attenuated as to have the general shape of the blade of a sword, thus suggesting the specific name. These bones, moreover, show but slight evidence at their distal extremity of union with the parietals. As the inner margin is rounded for nearly half the length, this feature will distinguish the present species from all others hitherto described. The following are some dimensions of portions of this specimen: Inches. WenothvombonnscoresiromEtoprom orbit to;suMmnity nse eee = soot ee ee ee eee SL 27 Antero-posterior diameter of same horn core at base... ........--------------------+------------ 8 INMGRS, Glee GWaP Orin se2 doask Shoe ae Se Ms eee ee he Se ee eee Se a ee 5 henge iho tsq wan osalybebindrexocceipitalecnooviess== sees oe5 9s a4 8e 2s ieee ee eee 55 (Creme, Wath. < c2c5 ssc LOSS Sees Be Soa os See Se eS A ee ER a 15 VV thre sitanaricl dll segetey pem ee te eee g ws paper ey eo lage ip oe ee eet are na) am he 9 These interesting specimens were also found in the Laramie of Wyoming by Mr. J. B. Hatcher. The type of the present species consists of a nearly complete parietal, left squamosal, a supraorbital and nasal horn core, an epijugal, the occipital condyle, and other skull fragments of less importance. The type of the present species represents the extreme development of the form of parietal crest that is peculiar to this genus. The squamosals, as shown in fig. 14, are extremely narrow and elongate. The parietals are very broad and long and are interrupted by a pair of huge supraorbital fontanelles, well shown in fig. 12. The posterior border of the parietals was emarginate at the median line. The median bar of the parietals was broad, thin, and smooth, especially at the posterior extremity. Its superior surface is nearly flat at the posterior extremity, but becomes more convex anteriorly, where it presents three low, rugose prominences arranged in a longitudinal series. In front of and posterior to the supratemporal fontanelles the parietals expand into broad, thin plates, which form respectively the anterior and posterior borders of the fontanelles. Below, these plates are united by a thin bar of bone, which unfortunately is not complete in the present specimen, but which, when in position, overlapped the superior border of the squamosat and formed the inferior lateral margin of the large fontanelle. There were no epoccipital bones on the posterior border of the parietals such as are present in other genera of the Ceratopsia, but the sharp, thin posterior border of these elements presents on either side of the median line a series of nine elongated prominences alternating with eight shallow emarginations, which give to the periphery of these bones the same peculiar scalloped effect that in Triceratops is produced by the epoccipitals. The narrow and elongate sqguamosals for 25 centimeters at their posterior extremities present a rather thick concave inner lateral border, against and into which fits the posterior external lateral border of the parietal. In advance of this for a distance of nearly a meter the external margin of the parietal overlaps the internal margin of the squamosal. This over- lapping and underlapping is carried to the greatest extent in the region opposite the middle of the supratemporal fontanelles, where it amounts to as much as 6 centimeters. Anteriorly the inner borders of the squamosals are much thickened and the inferior surface of each is produced into a long, narrow plate which becomes wider anteriorly. This plate passes beneath the margin of the anterior plate of the parietal and forms a part of the floor of the large supra- temporal fossa, while the thickened superior marginal border of the squamosal in this region forms the external lateral boundary of this fossa. The postero-external lateral border of the squamosal is without either sinuosities or epoccipitals, but anteriorly this border is marked by a series of four gentle prominences separated by three elongate and shallow emarginations. a - i 154 THE CERATOPSIA. | : < a | The last of these four prominences forms the antero-inferior angle of the squamosal and marks Wi - the posterior border of the quadratojugal notch, which is much shallower than in other genera Wi - of Ceratopsia. The groove for the quadrate is far in advance of the position which it occupies Wit in other genera. The more important characters of the squamosal are well shown in fig. 14, 1. ie The supratemporal fosse were deep and elongate and the floor of each appears to have been Wy stronger than in Ceratops or Triceratops. As in the type species of the genus, there was a small wat foramen on either side of the parietal just within the anterior lateral borders of these fossz. | From these foramina two broad, shallow channels run forward along the superior surface of the parietal. These converge and meet in the median line at the anterior extremity of the parietals, where there seems to have been a single median postfrontal (pineal) foramen instead of two placed one on either side of the median line, as in the preceding species of this genus. } Fie. 119.—Dorsal aspect of skull of type of Torosaurus gladius Marsh (No. 1831, Yale Museum). c, Supratemporal fossa; e’, epijugal bone; bil f’, parietal fontanelle; h, supraorbital horn core; h’, nasal horn core; p, parietal; s, squamosal; z, postfrontal foramen. One-twentieth ii natural size. After Marsh. | He Owing to the imperfect condition of the specimen in this region, it is not possible to determine with accuracy whether there were one or two of these foramina. Taken as a whole the entire frill, notwithstanding its great size, which considerably exceeds © that of any member of the Ceratopsia yet discovered, was remarkably light and frail and could have formed but a very inefficient protection to the cervical region. The surfaces of the bone are throughout smooth and entirely devoid of those grooves, channels, and rugosities so char- | acteristic of the parietals and squamosals in the genus Triceratops. In the present specimen — the entire frill seems to have been embedded in or covered by soft, muscular tissue instead of bearing a thick and dense dermal or horny sheath, as would appear to have been the case in — Triceratops and fully adult specimens of Sterrholophus. = p Only one of the swpraorbital horn cores is preserved and a portion of the base of this is — | wanting. It shows these horn cores to have been rather long, slender, and somewhat com- — pressed transversely, with a very rugose surface deeply furrowed with vascular groves, indi- | cating that in life they were covered with formidable horns. They were directed forward, — Hh upward, and outward, and, assuming that the facial region was correspondingly as short in the present as in the preceding species considering their length, they must have extended nearly or quite as far forward as the nasal horn. The nasal horn core, which was found detached, is very short and stout, compressed, and 1 has a very sharp-pointed apex. Its surface is extremely rugose and shows a number of deep | vascular grooves for the protection of the blood vessels which lay between the bone and the — | heavy horn with which in life it was insheathed. The surface of this horn core is broadly _ ———— eC ert NODOSAURUS TEXTILIS. iD) rounded in front and more contracted behind, so that in cross section it is ovate, with the apex of the oval directed posteriorly. The anterior surface of the horn core seems to have occupied a plane nearly perpendicular to that of the longitudinal axis of the cranium. The bone is comparatively dense without, but the central mass shows a coarse cancellous structure. A broad, low, disarticulated epijugal was found associated with the remains. This differs from the same element in Triceratops in being proportionally much broader and in having a concave inferior surface. The external buitncd of the bone is marked by ee vascular grooves, and it was united with the jugal by a very irregular suture. The general aspect of this skull as seen fom above is well shown in fig. 119, from a restora- tion of it by Professor Marsh. An examination of this figure will at once show the striking contrast between the enormously elongated posterior region and the abbreviated facial region, which becomes more especially emphasized when compared with similar views of crania per- taining to the genus Triceratops." The detached occipital condyle was found with the type. Considering the size of the parietal crest and the total length of the skull, the occipital condyle was small and the sutures between ‘the different elements composing it were not completely closed; when found it had parted along these sutures and lay in three subequal parts. The lower and median of these was formed by the basioccipital, while the exoccipitals formed the upper two-thirds of the condyle, as shown in fig. 7. The size of the condyle shows that, notwithstanding the enormous size of the frill, the skull proper was not unusually large for a member of the Laramie Ceratopsia, being rather under than over the average size. Trincipal measurements of the type. Mm enc ihvompanetalsysloneymedianiline ashe eee sees yeti LS ari Ea) Le 1, 432 Cremestpenothxofspanictallss sae eeiee cov meyers eee saci Mires alias 3 oe EOE ie ne 1, 568 Greatest length of squamosal.......---..--- a aU E ole ls ROM A eat eee aes es Oo ee eae 1, 400 CreanestMUneacuneolks tlaldtOsalle area ie Meer Me ai HE toa Ee ee ue as oe bye Stes 430 TDesitrumatenierel Iesaveailon aye lem ig ea le ee as on ie Sa ap Cet Sie etal a eae eae eae ee PD. S35 Greatest distance between external borders of squamosals measured over curved surface. ___.-.--- 932 Transverse diameter of supraorbital hom core near base......-...-..-2---22.-:4-------------- 133 Transverse diameter of supraorbital horn core at middle. .---.-....-...-.---.----------------- 82 Fore-and-aft diameter of supraorbital horn core at middle.............-.--------------------- “Ace walla Fore-and-aft diameter of supraorbital horn core near base ...---.-.-.-.-----------------+----- 191 ore-and-att diameter of nasal horm core mear base. 22-2 26222- 22.222 12-2 2242-22222 222. 2-8. 130 iransverse diamereror nasal morn:coremealybaser asses nee se get ose ee ese ee 81 Circumference tof, supracrbitall born’ core near base: 2-2. -2-.-5-.2----2-+--.22--+21--+-----+-+--- 510 Cincumitexence of supraorbitall hormkcore atimiddles = 2232-22 2.26. 225--- 22-2282. 8202 5252222. 125 Here tromepljU ec aleen ee ena eee meat ears atoms State RtinEtae Mas needs dae Swale saad ie 59 IBAEHYS Kay Gd! Coe alla eA oe 2 See eae Ce ie ea ee 105 WircineterrOmoccipitAlkcond yoni seer ee eta eee oes Oe jaye See ass lotrne soba ee Sale ees ee 85 NODOSAURUS Marsh. 1889. NoposaurRus TEXTILIS Marsh. 1889 Type No. 1815, Yale Museum. Original description in Am. Jour. Sci., 3d ser., vol. 38, Aug., 1889, p. 175. Marsh, O. C., Am. Jour. Sci., 3d ser., vol. 39, May, 1890, pp. 424-425; ibid., vol. 50, Dec., 1895, p. 497; Sixteenth Ann. Rept. U.S. Geol. Survey, pt. 1, p. 225 (dermal ossicles shown in fig. 5, Pl. LX XV). Nicholson and Lydekker, Manual Pal., 1889, p. 1164. Zittel, K. A. von, Handbuch Pal., p. 754. The type (No. 1815, Yale Museum) of the present genus and species was found, according to Professor Marsh, in the ‘‘middle Cretaceous of Wyoming.’’ Marsh’s original description was as follows: Another new member of the Stegosauria, from a lower horizon in the Cretaceous, was discovered several years since in BESOIN end is now in the Yale Museum. The skull is not known, but various portions of the skeleton were secured. One a Manet? s restoration, fig. 119, does not include the portion of the skullin front of the nasal horn; if this were added the disproportion would be less marked.—R.S. L. 156 THE CERATOPSIA. characteristic feature in this genus is the dermal armor, which appears to have been more complete than in any of the American forms hitherto found. This armor covered the sides closely and was supported by the ribs, which were especially strengthened to maintain it. In the present specimen portions of it were found in position. a series of rounded knobs in rows, and these protuberances have suggested the generic name. Near the head the dermal ossifications were quite small, and those preserved are quadrangular in form and arranged in rows. The external surface is peculiarly marked by a texture that appears interwoven like a coarse cloth. This has suggested the specific name, and is well shown in the cut below [fig. 120]. : The fore limbs are especially massive and powerful, and are much like those of the Jurassic Stegosaurus. There were five well-developed digits in the manus, and their terminal phalanges are more narrow than usual in this group. The ribs are T-shaped in transverse section, and thus especially adapted to support the armor over them. The caudal vertebre are more elongate than those of Stegosaurus, and the middle caudals have a median groove on the lower surface of the centrum. The animal when alive was about 30 feet in length. The known remains are from the middle Cretaceous of Wyoming. From the above quotation it is clear that Marsh at that time considered this genus and species as belonging to the Stegosauria. From the title of his paper (‘Notice of Gigantic Horned Dinosauria from the Ciiacome and the context of the text, however, it is also evident that he then included Nodosaurus in the Ceratopside, which ine at that time con- sidered as a family of the Stegosauria. A few months later, however,“ he removes this genus from the Ceratopside, making it the type of a new family, the Nodosauride. He now includes It was regularly arranged in ‘the families Ceratopside and Nodosauride in a single group, the Ceratopsia, which in the introductory paragraph of his article he regards as a suborder, but in the classification of American Cretaceous dinosaurs with which he closes his paper he makes this group of ordinal rank and makes no mention of the Stegosauria. It is clear, therefore, that he at that time regarded the Ceratopsia and Fic. 120.—Dermal ossicles of Nodosaurus textilis Marsh. Sles osetia os pertains iO Gerace orders onj=ul ORO Re CN Taare orders, and that he included the Nodosauridz in the Ceratopsia rather than the Stegosauria. His classi- fication of American Cretaceous dinosaurs as then proposed is as follows: The American Cretaceous dinosaurs now known represent several well-marked families, which may be arranged as follows: 4 Order Theropoda. Carnivorous. (1) The Dryptosauride, including the large carnivorous forms, of which only imperfect specimens have been found but sufficient to mdicate that they are distinct from the Megalosauridxe of the Pimopeaks Jurassic. Limb bones hollow. Fore limbs very small. Feet digitigrade, with prehensile claws. Order Ornithopoda. UHerbivorous. (2) The Trachodontide, herbivorous forms of large size, with teeth of the Hadrosaurus type, in many rows vertebra opisthoceelian. Limb bones hollow. Fore limbs small. Feet digitigrade. (8) The Claosauridx. Only a single row of teeth in use. Cervical vertebrze opisthoccelian. limbs small and feet ungulate. Genvical Limb bones solid. ore (4) The Ornithomimide. Limb bones hollow. Fore-limbs very small; hind limbs of avian type. Feet digitigrade and unguiculate. Order Ceratopsia. Herbivorous. (5) The Ceratopsidz. Highly specialized forms fully defined above. (6) The Nodosauride. Heavy dermal armor. Bones solid. Fore limbs large. Feet ungulate. No Sauropoda are known from the American Cretaceous. Although in the above classification Marsh treated a Ceratopsia as a distinct order, yet a year later,” in the text of a paper accompanying a restoration of the skeleton of Triceratops, he romance This group, so far as at present investigated, is very distinct from all other known dinosaurs, and whether it should be regarded as a family, Ceratopside, as first described by the writer, or as a suborder, Ceratopsia, as later defined by him will depend upon the interpretation and value of the peculiar characters manifested in its typical forms. a Am. Jour. Sci., 3d ser., vol. 39, May, 1890, p. 425. > Am. Jour. Sci., 3d ser., vol. 41, Apr., 1891, p. 340. NODOSAURUS TEXTILIS. 157 This last-mentioned paper was followed in a few months by another? accompanying a restoration of Stegosaurus, and in this it is clear that Marsh regards the group as deserving the rank of a suborder only, coordinate with the Stegosauria. The two together he then regarded as comprising the order Ornithopoda. Three years later,” however, Marsh restricted the term Ornithopoda to its original limits and made the group coordinate with the Stegosauria and Ceratopsia, all three of which he considered as suborders belonging to a single order for which he proposed the very appropriate name Predentata. The following year Professor Marsh presents a revised classification of the Dinosauria, and in this he places the Nodosauridez and the genus Nodosaurus in the Stegosauria rather than the Ceratopsia. This is also the classification followed in his “‘Dinosaurs of North America,’¢ which may be regarded as the most comprehensive expression of his views on this subject. It would seem, therefore, that Professor Marsh’s final decision was in favor of excluding Nodosaurus from the Ceratopsia and including it in the Stegosauria. After a personal examination of the type I am convinced that it is not a member of the Ceratopsia. [Hatcher’s manuscript ends here.] a Am. Jour. Sci., 3d ser., vol. 42, Aug., 1891, pp. 179-181. ce Am. Jour. Sci., 3d ser., vol. 50, Dec., 1895, pp. 485-498. 5 Am. Jour. Sci., 3d ser., vol. 48, July, 1894 pp. 89-90. d Sixteenth Ann. Rept. U. S. Geol. Survey, pt. 1, p. 243. JPRS Ales _ HABITS, AND ENVIRONMENT OF THE CERATOPSIA BY Ihe Ase Se eels 159 FS PHYLOGENY, TAXONOMY, DISTRIBUTION, HABITS, AND ENVIRON- MENT OF THE CERATOPSIA. GENERIC AND SPECIFIC SUMMARY. GENERAL DISCUSSION OF THE PHYLOGENY. Of the fourteen genera which have at sundry times and by various authors been referred to the Ceratopsia but seven have survived the process of elimination to which they have been subjected. Of these, three more primitive genera were found in the earlier Judith River beds, and the remaining four, which are much more specialized, were found in the Laramie deposits. That there is a genetic relationship between the earlier forms and their successors is evident, and, as Osborn, Lambe, and Hatcher have suggested, there seem to have been at least two lines of descent, one leading to Triceratops, with its entire bony frill, and the other to Torosaurus, with persistently open parietal fontanelles. Professor Osborn’s® statement is as follows: It is not at all improbable that the horned dinosaurs will prove to be diphyletic, one line, with persistent open fosse, leading from Monoclonius to Torosaurus, the other leading to Triceratops with closed fosse. Lambe,? in speaking of the parietal of Ceratops (Monoclonius), says: The parietal is about one-third the size of that of 7. gladius, and would probably represent a proportionately smaller animal, an earlier and more geveralized form of the genus, with larger fontanelles than its later Laramie successors. In Hatcher’s unrevised description of the genus Torosaurus (p. 150) he says: As already remarked, the parietals of the present genus most nearly resemble the same elements in the type of Mf ono- clonius (M. crassus), and there seems but little doubt that the last-mentioned genus was ancestral to Torosaurus. In the revised portion of his manuscript, however (p. 100), Hatcher reverses this earlier decision in the following statement: The affinities of Monoclonius, as shown in the type species, M. crassus Cope and M. dawsoni of Lambe, are apparently with the later genus Triceratops of Marsh, while Ceratops montanus Marsh, C. recurvicornis Cope, C. canadensis, and C. belli Lambe would seem to be ancestral to Torosaurus. Mr. Lambe’s statement is fully in accord with this, while that of Professor Osborn may be reconciled to it if by ‘*‘ Monoclonius”’ he had special reference to M. canadensis and M. belli, which Hatcher has removed to the genus Ceratops. If, on the contrary, Osborn had in mind the type species, Monoclonws crassus, his opinion as expressed would be at variance with Hatcher’s final idea of these relationships. The weight of evidence to be reviewed below is certainly in favor of Hatcher’s final theory, according to which the genera may be provisionally arranged in two phyla in the following sequence: , Phyla of the Ceratopsia. Geological horizon. Triceratops phylum. Torosaurus phylum. Denver beds. Triceratops. ; q Diceratops. Torosaurus. Laramie of Converse County, Wyo. Triceratops. Laramie of Black Buttes, Wyo. Agathaumas. Centrosaurus. Ceratops. Judith River (Belly River). Monoclonius. aContr. Canadian Pal., vol. 3 (quarto), pt. 2, p. 31. b Ibid., p. 67. MON XLIx—07——11 161 162 — THE CERATOPSIA. Of the two most primitive genera Monoclonius seems to be the more generalized and repre- sents the earliest known stage in the evolution of the Ceratopsia. Because of the gap between the Judith River beds and those of the Laramie the series is by no means complete, - nor are we yet aware of the characteristics of pre-Judith River ancestors. The earliest known Ceratopsia are endowed with the main distinguishing characters, the horns and parietal crest. With the exception of the genus Agathawmas, of which the skeleton only is known, the more important taxonomic characters are found in the cranium, and the genera can readily be differentiated without recourse to the skeletal features; in Boer in four out of the seven genera these are unknown. REVISION OF GENERA. Taking the six genera recognized by Hatcher and one erected later by Lambe, which Hatcher would undoubtedly have accepted as valid, the main generic comparisons and contrasts are as. follows: I. THE MONOCLONIUS-TRICERATOPS PHYLUM. _ 1. Monoclonius, the most primitive genus, is comparatively small, and has three horn cores, the nasal being long, powerful, and curved backward. The supraorbitals are diminutive, flattened on the outer surface but very convex on that toward the median line, so that the basal section is almost triangular. The horns are sharp pointed. The orbit is nearly circular and in at least one species has a heavy rugose border. The frill is made up largely of the coalesced parietals, the squamosals being somewhat triangular, plate-like bones, short and proportion- ately broad. The margin of the crest is crenulated, but the prominences do not seem to have arisen from separate ossifications, as in the succeeding genera of this phylum. The parietals are widely fenestrated—in fact, they do not entirely surround the fenestre, and the median bar is broad and thin in contrast to the much narrower bar in Ceratops. Of the skeletal characters those of greatest taxonomic value are found in the sacrum. The number of coalesced vertebree in the entire sacrum is ten; of these eight bore sacral ribs, of which four pairs united distally on either side into a longitudinal bar for the support of the ium. The centra were of medium length, constricted medially, and the sacral ribs arose directly opposite the points of articu- lation, so that each rib bore equally on two contiguous centra. The ilium is rather long and slender, and the deflected margin posterior to the ischiac peduncle is produced into a somewhat angular prominence. The blade is rounded anteriorly, while posteriorly it is elongate and narrows presumably to a point, though the extremity is not preserved. The blade, when the bone is in position, is horizontal, as in other Ceratopsia. In contrast with the ilium of Aga- thaumas that of Monoclonius is smaller and proportionately more slender, and the deflected external margin is more pronounced. 2. Centrosaurus.—In Centrosaurus the nasal horn is straight, laterally compressed so as to be lenticular in section, and somewhat similar to that of Monoclonius sphenocerus. The coalesced _parietals onal surrcund the large, oval fontanelles and the median bar is very heavy, especially between the peculiar inward-curved processes at its posterior border, but. thins perceptibly toward the anterior portion of the frill. On either side the fontanelle is inclosed in a thin extension of bone. In addition to the two curved posterior prominences there. are seven others, separated by emarginations, the posterior ones bearing separately ossified epoccipital bones as in Triceratops. The squamosals are short, as in Monoclonius, bemg confined to the antero-external angles of the frill. 3. Agathaumas.—The genus Agathaumas, so far as known from skeletal fragments, is intermediate in character between Monoclonius and Triceratops. It comes from the lower Laramie, near Black Buttes Station, Wyoming. In Agathawmas the main distinctive charac- ters are of necessity derived from the sacrum, the ilia, the dorsal vertebre, and the ribs. The sacrum has five true sacral vertebre, probably with one presacral and four caudals, making ten in all, as in Monoclonws. In the type specimen, which was not fully mature, there are four pairs of sacral ribs, arising more directly from the articulations than in Triceratops, yet, with the exception of the first, not so much so as in Monoclonius, being in a condition GENERIC AND SPECIFIC SUMMARY. 163 transitional between that in each. The four sacral ribs coalesce distally to form the horizontal bar for articulation with the ila, but in the present instance the bar of each side is not continuous, the ribs being, as Cope expresses it, united in pairs. It is possible that continuity would have come later in life had the creature reached maturity. The centra of the true sacrals decrease in size regularly from first to last, as in Monoclonius and Triceratops, the first and second being particularly broad and short. They are somewhat more elongate than the dorsal centra, the vertical and transverse axes being more nearly equal. Perhaps the most important distinction is the poimt of origin of the sacral ribs. There was evidently a pleuro-diapophysial connection with the ilium, as in Monoclonius and Triceratops, but the diapophyses are not preserved. The dorsal vertebre exhibit no marked distinction from those of the other genera. In Agathaumas the blade of the ilium is much extended fore and aft, the anterior end being truncated and the hinder extremity narrow and elongate. The external margin is moderately thick throughout its length and much thicker near the extremities and in the deflected region above the ischiac peduncle. The internal margin is thin except in the region of articulation with the sacral diapophyses. A broad, convex ridge of bone extends diagonally across the anterior blade of the illum to its antero-external angle, inclosing in front a deep depression just external to the acetabulum. The ilium is actually larger and relatively shorter and broader than that of Monoclonius, and the deflection of its external margin is less pronounced. It differs from the ilium of Triceratops in being more slender and in having occasional thickenings of the external border, whereas in Triceratops the margin is more uniformly thin. The posterior end of the Aga- thaumas iltum is thin, while that of Triceratops is thickened. The ilium of Agathaumas is transitional between those of the contrasted genera in general proportions, especially the relative length and breadth. 4. Diceratops.—This genus is known from a complete skull and, although unquestionably related to the Monoclonius- Triceratops phylum, its serial order is somewhat difficult to conjecture. By the absence of a nasal horn core it resembles Triceratops obtusus, though evidently not synonymous therewith. The fenestrated parietals would seem to point to primitive conditions until one notes the presence of similar fenestrae” in the squamosals, a character which here appears for the first time. This, together with the fact that the squamosal fenestre are of unequal size—which may also have been true of those of the parietals, as only the right is preserved—leads one to conjecture whether they may not have been secondarily acquired and together with the vestigial nasal horn, may not be evidences of high specialization from some Triceratops-like ancestor. The main diagnostic characters wherein Diceratops differs from Monoclonius are the reduced nasal horn, the vastly increased supraorbital horns, the elliptical orbit, the well-developed parietals with their small fontanelles(?), the more elongate fenestrated(?) squamosals, and the separately ossified epoccipital bones. From Triceratops, its nearest ally, it may be distinguished mainly by the much smaller rostral bone; by the absence of the nasal horn, which in all of the species of Triceratops except TJ. obtusus is fairly well developed; by the very erect, short, robust supraorbital horn cores, which seem to take their origin much farther back with relation to the orbit; by the concavity of the frontal region between the orbits; and, finally, by the peculiar form of the persistent postfrontal (pineal) fontanelle suggestive of that of the genus Torosaurus. The parietals are Triceratops-like except for the small fenestrae on either side of the median line, while the squamosals, aside from the unique fenestra, differ from those of Triceratops in the conformation of the inferior border, which lacks the quadrate notch. Another distinctive feature is in the very erect position of the descending process of the jugal, which is directed slightly forward instead of downward and backward as in T’riceratops. aThe writer is now firmly convinced that all of these apertures through the frill of Diceratops are pathologic, having been caused either by wounds or by disease (Am. Jour. Sci., 4th ser., vol. 20, pp. 419-422). Similar perforations occur in the right frontal of the type of Triceratops serratus (p. 124) and in the squamosal of that of 7. elatus (p. 136). Mr. ©. W. Gilmore, who prepared the type, is not sure of the parietal perforation, but as no bone adhered to the matrix at that point the opening was allowed to remain. 164. THE CERATOPSIA. 5. Triceratops (including Siemnelonhn) —In Triceratops this phylum reaches its culmina- tion in specialization and size, including some most formidable species. The supraorbital horns reach ties maximum development in TZriceratops, while the nasal horn is in process of reduction and the broadly expanded frill, though not so large actually or relatively as that of Torosaurus, was much more efficient for protection because of its vastly greater strength. The rostral bone is better developed than in Diceratops, being especially heavier in older individuals, and there is generally a sharp cutting edge on its ‘atetion face. The nasal horn, while semaine to reduce, is of moderate length in the more generalized species and is directed forward, so that its posterior face rises but little above the level of the upper surface of the nasal bones, in sharp contrast to that of Monoclonius. The nasal horn core is generally very rugose. The supraorbital horn cores are slender to robust, ovate in section as contrasted with the triangular cross section of the rudimentary horn cores in Monoclonius and the nearly circular section of those of Diceratops. In Triceratops the horn cores point upward and forward at an angle of 45°, whereas in Diceratops they are nearly erect. They range in length from short in T. brevicornus to very long in T. calicornis. The orbit is generally elliptical in shape with the long axis somewhat diagonal, running from above downward and forward, therein agreeing with Diceratops rather than with Monoclonius. The parietals are convex laterally and somewhat concave along the long axis, much expanded posteriorly, but narrowing toward the forward end. They are thick along the more or less elevated and rugose ancl line and around their posterior border, but very thin in the center of each side, though there are no traces of fontanelles. It is mainly in this last feature that the parietals differ from those of the other genera. The squamosals are stout and broad, constituting at least half of the area of the frill. The superior surface of the frill, and, in some aged individuals, the margin of the lower side as well, bears vascular impressions, showing it to have been sheathed in a horny covering. These impressions are much more pronounced than in either Monoclonius or Centrosaurus. Hatcher has made no final disposition of the genus Sterrholophus either in his completed manuscript or in his notes, but one is justified in the assumption that he believed the genus to be. synonymous with Triceratops on the ground that all of the distinctive features which would serve to separate the genera he considered as juvenile characteristics. The type consists of disarticulated skull and other skeletal portions of an immature individual and differs from a typical specimen in the entire absence of vascular impressions from the frill. Those on the under surface are seen only in aged individuals, as in the type of — Triceratops prorsus (Pl. XX XIII), and it may well be that their appearance on the superior surface of the skull is a strictly adult feature. Another distinction between the skulls of Triceratops horridus and Sterrholophus flabellatus is shown in figs. 9 and 27, pages 18 and 30. In the former the olfactory foramina are paired, whereas in the latter there is but a single foramen. But, as Hatcher says, “it is probable that the median septum was present as cartilage and was lost in maceration. It is very unlikely that there was in any of these dinosaurs a single olfactory nerve.’ Another distinction between Sterrholophus flabellatus and some species of Triceratops is in the position of the lachrymal foramen, which is entirely within the maxillary, not as in other species between the maxillary and nasal. In this feature Triceratops prorsus agrees with Sterrholophus flabellatus, while in T. serratus the other condition obtains, T. brevicornus having the foramen between the nasals and maxillary but embraced by an ascending process of the maxillary in its anterior moiety, a condition transitional between that of the other two. This character can thus be considered of specific value only. It would therefore seem wise, in the light of our present knowledge, to consider the newer generic name Sterrholophus as a synonym of Triceratops. The main skeletal comparisons between Triceratops, Agathaumas, and Monoclonius which it is possible to make lie in the sacrum and ilia, in which Triceratops shows a further advance- ment over Agathaumas. The general structure of the sacrum agrees with that of the latter in having ten vertebra. Of these four bear sacral ribs, which coalesce to form the longitudinal GENERIC AND SPECIFIC SUMMARY. 165 bar for articulation with the ilium. The first pair of ribs agrees with those of both above-men- tioned genera in springing from the point of articulation between vertebre 2 and 3, bearing equally upon each, while the succeeding sacral ribs arising behind the articulation bear very little upon the preceding centrum at all. A reference to figs. 53 and 77 (pp. 52 and 75) and Pls. X and XXV (pp. 218 and 248) will make clear these comparisons and contrasts. The ilium in Triceratops is broad and elongated, much expanded anteriorly and tapering to a somewhat thickened point behind, whereas in Agathawmas the posterior extremity is thin. The external border is thin and smooth in contrast to the variably thickened margin in the latter genus. The outline as viewed from above is less regular than in Agathawmas, which is due to a greater prominence of the deflected border above the ischiac peduncle. In Triceratops the internal margin is greatly thickened, forming the superior border of the acetabulum and strengthening the peduncles. Compare figs. 55, 61, and 82. Ii. THE CERATOPS-TOROSAURUS PHYLUM. Of the second phylum but two genera are as yet known, Ceratops and Torosaurus, agreeing in the possession of large, persistently open parietal fenestre and in the general conformation of the squamosal, which, while not so saber-like in Ceratops as in Torosaurus, exhibits a strong tendency to become so, in sharp distinction with the short, broad, triangular squamosals of Monoclonus. The members of this phylum also agree in the absence of separately ossified epoccipitals, the crenulated effect of the margin of the frill being produced by prominences which do not arise as separate ossifications. An exception to this is found in the type specimen of Ceratops (Monoclonius) recurvicorns Cope, referred to Ceratops by Hatcher but possibly representing a new genus, a member of the Triceratops phylum. 1. Ceratops, a Judith River genus, is known only from portions of the skull, no other skeletal elements having been discovered. Our knowledge of the nasal horn core is imperfect, as it is derived only from the specimen of Ceratops recurvicornis mentioned above. In this specimen the horn core is massive, compressed distally, and curves forward instead of back- ward, as in the contemporary genus Monoclonius (see fig. 3). Another isolated horn core, figured in Lambe’s memoir (Pl. XVII, fig. 1=Pl. XVIII, fig. 1 of this monograph), was referred by that author with a query to Monoclonius dawson Lambe. Hatcher claims that it curves forward instead of backward, which would seem to place it in the genus under con- sideration, though under what species can not be determined. The supraorbital horn cores show a greater degree of development than in Monoclonius, but are apparently still smaller and less massive than the nasal core. They are more nearly circular in cross section, being subtriangular near the base, and are curved forward and upward and, in one species, C. montanus, strongly outward (fig. 103). The orbit is oval, with its long diameter nearly vertical as contrasted with the circular orbit of Monoclonius and the more nearly horizontal oval of Torosaurus. The coalesced parietals are reduced to aslender, median bar, subcylindrical in mid-length, and narrow postero-lateral processes, which only partially surround the large elliptical fontanelles. The squamosals, which overlap the outer portions of the parietals, are triangular and more elongate than in Monoclonius, tending toward the form of squamosal found in Torosaurus. 2. Torosaurus.—The anosennons of Te Torosaurus cranium at once separate it from that of any other genus, especially from its immediate contemporaries, for while in Triceratops the frill and facial region are well balanced, in Torosaurus the huge crest entirely overshadows the abbreviated anterior portion of the shill The nasal horn core in the type species is broad at the base and the sharp apex is directed upward and forward. The supraorbital horn cores are larger, as in all Laramie Ceratopsia, and ovate in basal section, with the broader end of the oval to the rear. The cores are directed upward, outward, and forward. Thus the nasal horn core is relatively smaller while the supra- 166 _ THE CERATOPSIA. orbitals are larger than in Ceratops. The orbits are elliptical, with the long diameter more nearly horizontal than in other genera, and are placed farther forward relative to the position of the horns than in Triceratops, while the skull is broad and massive between them, narrowing very rapidly toward the snout. The skull is extremely broad in the region beneath the orbits, as the external surface slopes downward and outward as compared with the more nearly verti- cal sides of the skulls referable to the genus Triceratops. There is a large lachrymal foramen well in advance of the orbit. The postfrontal fontanelle seems to have been paired, at least in T. latus, the type of T. gladius being too imperfect to permit one to judge. From this run two shallow grooves directed outward and backward, each ending in a small foramen. There are similar grooves in the type of Diceratops, and in the specimen (Nos. 1203 and 1206-1210, U. S. National Museum) referred by Marsh to Triceratops sulcatus (Pl. XX XVII, fig. 1) the grooves are present but as the posterior portion of the skull is broken away the terminal foramina can not be seen. In each of these forms, however, the postfrontal foramen is unpaired. The parietals are broad and long, with large supratemporal fontanelles. The parietal fontanelles are not so large proportionately as in Ceratops, and the median bar is broad, thin, and smooth, especially toward its posterior end. The squamosals are very elongate and sword- like, readily derivable from such as are found in Ceratops. The quadratojugal notch is much less pronounced than in Triceratops; the groove for the quadrate being far in advance of its position in other genera. The posterior margin of the frill is emarginate, but there are no separately ossified epoccipitals as in Triceratops. III. GENERA ELIMINATED OR REMOVED FROM THE CERATOPSIA. The genera abandoned or removed from the Ceratopsia by Hatcher are named below: 1. Claorhynchus Cope, with its one species, trihedrus, founded upon a rostral and pre- dentary bone, and now lost sight of, Hatcher supposes, from Cope’s description, to pertain to the Trachodontide rather than to the Ceratopsia. 2. Dysganus Cope, with its four species, D. encaustis, D. haydenianus, D. bicarinatus, and D. peiganus, is abandoned on the ground that it was “based on teeth pertaining to two or more genera belonging in part to the Trachodontide and in part to the Ceratopside.” The type material was very imperfect and the descriptions were inexact and unaccompanied by figures, hence the genus should, in Hatcher’s opinion, be considered a nomen nudum. 3. Manospondylus Cope, with its single species, gigas, is based upon the centrum of a single vertebra (fig. 105). From its general form, its coarsely cancellated internal structure, and the deep fossa of the superior lateral surface it resembles most closely the vertebre of Dryptosaurus, a theropod dinosaur, hence one is justified in removing the genus in question from the Cera- topsia. 4. Another of Cope’s genera, Polyonax, with the species mortuarius, erected upon “ frag- ments of horn cores, vertebre, etc.,’’ is abandoned because of “the extremely fragmentary and totally inadequate nature of the material upon which the genus and species are based.” 5. Finally, Lambe’s genus Stegoceras, with its single species, validus, is based upon two cranial fragments (figs. 99 and 100 and Pl]. XXII), which Lambe called “‘prenasals,” but which Hatcher believed are the “‘superior portion of the occipital, parietal, and frontal segments of the skull” of a reptile which may represent an order new to science. He further adds that “there is no good reason for considering them as horn bearing or the animals to which they belonged as pertaining to the Ceratopsia.”” Hence Hatcher does not include the genus in that group. 6. Sterrholophus Marsh is considered synonymous with Triceratops, for the reasons already mentioned. GENERIC AND SPECIFIC SUMMARY. 167 REVISION OF SPECIES. I. THE MONOCLONIUS-TRICERATOPS PHYLUM. MONOCLONIUS. Of the seven species referred to the genus Monocloniwus by their several authors, Hatcher retains but three, MW. crassus Cope, M. dawson Lambe, and M. sphenocerus Cope. He abandons M. fissus Cope, because of insufficient type material, while the remaining three, canadensis Lambe, recurvicornis Cope, and belli Lambe, are removed to the genus Ceratops. 1. Monoclonius crassus Cope (pp. 71-80, figs. 75-88) is the type species and is known from the remains of at least two individuals (type No. 3998, American Museum of Natural History), including the parietals, the frontal and postfrontal bones, a supraorbital horn core, and other portions of the skeleton. Of this material only that pertaining to the skull can be used in specific contrast, for of the remaining species no skeletal parts are known. The supraorbital horn core (fig. 76) is low, broad below, and pointed above, with a flat outer face and a strongly convex inner surface. The main point of contrast with that of IM. dawsoni seems to be one of size, as in the latter the horn cores are extremely diminutive, so small, indeed, as to have been overlooked by so careful an observer as Lambe. The orbit is nearly circular, with a thick, rugose border. The complete parietals (fig. 75) are known only in this species, though fragments of parie- tals included in the type specimen of M. dawsoni show no specific distinctions fom those of the species under consideration. As the parts preserved in MV. crassus are unknown in MM. sphenocerus one can not contrast them. 2. Monoclonius dawsom Lambe (pp. 89-93, fig. 92) is known from cranial fragments of specimen No. 1173, Geological Survey, Canada. The second specimen, consisting of parietals and a nasal horn core (No. 971), referred to by Lambe in his original description, has been made the type of a new genus and species, Centrosaurus apertus Lambe, in a paper published since Hatcher’s death (see p. 93, footnote b). The nasal horn core is lense. somewhat compressed, backward curving, and ovate in section, with the broader end in front, differmg materially in shape from that of M. sphenocerus. The supraorbital horns are described as very diminutive and triangular in section, as in MV. crassus. About 1 inch of the apex is not preserved. The parietals are known only from a fragment with four marginal undulations, as in M. Crassus. The orbit is large and circular. Both orbit and nasal horn core are very large in proportion to the occipital condyle, maxillaries, and quadrate. Hatcher was ‘inclined to regard the present species as closely allied to if not identical with J. crassus of Cope.” Some of the remainder of the material provisionally associated with the type by Lambe, consisting of a sacrum, a scapula and coracoid, a predentary, and a rostral bone, may pertain to other genera and species. 3. Monoclonius sphenocerus Cope (pp. 87-88, fig. 91, A, B, C) consists of portions of pre- maxillary, nasals, and nasal horn of a large though scarcely adult animal. It is distinguished from M. dawsoni in the form of the nasal horn, which is much compressed, with the anterior margin acute and the posterior rounded, the reverse of that of the last-mentioned species. In M. sphenocerus the horn is straight and directed upward and backward instead of being curved backward, as in M. dawsoni. It is, as Hatcher says, the largest and most powerful nasal horn core observed in any of the Ceratopside (fig. 91). Thus JM. sphenocerus is in sharp contrast with M. dawsoni, and probably also with M. crassus, though a direct comparison with the latter can not be made. The nasal horn resembles most closely.the one associated by Lambe with the parietals which constitute the type of his new genus Centrosaurus, and the true affinities of M. sphenocerus may prove to lie with that genus. a Ottawa Naturalist, vol. 18, 1904, pp. 81-84. This paper was issued July 7, four days after Ilateher’s death. 168 THE CERATOPSIA. CENTROSAURUS. 1. Centrosaurus apertus Lambe (p..93, footnote bd; fig. 93; Pl. XXIV) is based upon a parietal erest and an associated nasal horn core (No. 971, Canadian Geological Survey), and no specific: distinctions need be reviewed. AGATHAUMAS. In the genus Agathawmas Hatcher retains but one species, A. sylvestris Cope, the other, A. milo Cope, having been referred later by its author to Hadrosaurus occidentalis. Hatcher sums the matter up by adding: ‘‘Whether or not this determination was correct the fact remains that A. milo is a nomen nudum, and the fragmentary nature of the type precludes. adequate description. The species shorllél therefore be discarded.”’ 1. Agathaumas sylvestris Cope.—The characteristics of Agathawmas Minion (pp. 105-111, Pl. X XV) have been discussed under the generic summary. DICERATOPS, 1. Diceratops hatcheri Lull.—This genus contains but one species, D. hatcheri, described but left unnamed by Mr. Hatcher, based upon skull No. 2412, U. S. National Museum (p. 149,. Pls. XLVIL and XLVIIT). The specific characters need aint be cited in this ae but. one species being thus far known. TRICERATOPS. The present writer recognizes ten species under the genus Triceratops Marsh, a number which he believes to be in harmony with Hatcher’s views, though in the case of T. (Sterrholo— phus) fiabellatus no final statement by Hatcher has been Rowand. These species are as follows: 1. T. (Ceratops) horridus Marsh. 6. T. sulcatus Marsh. 2. T. prorsus Marsh. 7. T. elatus Marsh. 3. T. brevicornus Hatcher. 8. T. calicornis Marsh. 4. T. serratus Marsh. 9. T. (Sterrholophus) flabellatus Marsh. 5. T. (Bison, Ceratops) alticornis Marsh. 10. T. obtusus Marsh. Thus three species have been brought into the genus, two from Ceratops and the only species of Sterrholophus, which becomes synonymous with Triceratops. One species, T. galeus Marsh (p. 132, fig. 111), which was based upon a single horn core that resembles most the same structure in Torosaurus gladius, Hatcher decided to abandon on the ground of the frag- _ mentary nature of the specimen. 1. Triceratops horridus Marsh (pp. 117-122, figs. 24, 25,27, 107, Pl. XX V1), the type species, is based upon the imperfect skull of a huge individual, No. 1820, Yale Museum, fully adult. and very old. The rostral bone is very heavy, not so sharp along its inferior border as in some species,. downward curved toward the tip, with deep vascular impressions. The nasal horn core (Pl. XX VI) is very broad at the base, short and blunt in contrast to that of the most nearly allied species, T. prorsus, in the type of which the nasal horn is long and directed forward, and T. brevicornus, in which it is short and very stout but not highly rugose. This contrast is themore interesting because the types of all three species were aged individuals. The supraorbital horns in T. horridus are exceedingly stout and rugose, long, and directed forward as in T. prorsus,in contrast to the extremely short horns of T. brevicornus. The great. ‘size of T. horridus as compared with either T. prorsus or brevicornus is an important distinction. 2. Triceratops prorsus Marsh (pp. 127-132, figs. 35, 37, 40, 41, 49-58, 63-67, 71, 109, 110, Pls.. VI-XVII, XXX-XXXVI) is an example of the opposite extreme in size from T. horridus,. the type skull, No. 1822, Yale Museum, that of an aged individual, being one of the smallest. of the Laramie Ceratopsia. The rostral bone is contrasted with that of 7. horridus in being somewhat less massive and having a sharp cutting edge, as contrasted with the blunt margin in the other. The inferior margin curves downward toward the point, in agreement with TJ. horridus and T. brevicornus.. GENERIC AND SPECIFIC SUMMARY. 169 The nasal horn core is long and is directed forward, the-anterior border extending forward and upward at an angle of 30° instead of being perpendicular to the long axis of the skull as in T. horridus and T. brevicornus. The horn much exceeds that of each of the contrasted species in length, the tip of the horn in the type specimen being just over that of the rostral bone CEE OXOGDY)). The supraorbital horn cores are slender, and are directed upward, forward, and outward at an angle of 45° for about half their length, then curve gently inward. Here the contrast with the allied species is evident in the stoutness of the horns as compared with their slenderness in T. prorsus. The horns of T. prorsus are much longer proportionately than those of T. brevicornus. The orbits in the present species are nearly circular as contrasted with the elliptical orbits of brevicornus. The form of those of T. horridus can not be ascertained, as but a quarter of the margin is preserved, but they would seem to agree more nearly with those of T. prorsus. The lachrymal foramen is entirely within the maxillary bone, as in T. (Sterrholophus) fiabellatus, in contrast to its position between the maxillary and nasal as in T. serratus, T. brevi- cornus being in a sense transitional between the two types, while the condition which obtained in T. horridus can not be determined, as this part of the specimen is lacking. The frill is deeply arched transversely, ranging through an are of 27°, with seven lateral and one median epoccipitals, making fifteen in all. The quadratojugal notch is deeper than in any other known species, and the postfrontal (pineal) fontanelle is entirely closed. Posteriorly the frill border was free, for vascular impressions occur on its inferior face for a distance inward of 20 cm. from the margin. This feature, together with the closure of the postfrontal fontanelle, may be a characteristic of old age, but in some old skulls of other species the fontanelle seems to be persistently open. 3. Triceratops brevicornus Watcher (pp. 141-142, Pls. XL—XULII) presents another instance of an aged individual, the type being No. 1834 of Yale Museum. The rostral bone is proportionately very heavy, with a deeply excavated inferior surface. The nasal horn core is short and very stout, the antero-posterior diameter much longer than the transverse, and the anterior border vertical, as in 7. horridus. The supraorbital horn cores are short and stout and, in contrast to the much longer cores of horridus and prorsus, more nearly circular in section than in any other species. The orbit is an irregular ellipse, and the lachrymal foramen lies between the nasal and maxillary bones, but is partially embraced by an ascending process of the latter. The infratemporal arcade is formed from the quadrate, with overlapping processes from the jugal and squamosal, as in 7. serratus. The frill is elevated somewhat sharply toward the posterior margin, as in 7. prorsus, and bears nineteen epoccipitals, six pairs of which are borne on the squamosals, as contrasted with fifteen in the last-named species. It is doubtful whether or no this will prove a specific rather than an individual distinction (Hatcher). 4. Triceratops serratus Marsh (pp. 122-127, figs. 16, 26, 32, 34, 42, Pls. XXVIJI-XXIX) is founded upon the skull of an immature individual (No. 1823, Yale Museum), but one in a remark- able state of preservation. The rostral bone is rather small, lighter and less rugose than in other species; this, however, may be either a juvenile or possibly a sexual character. The nasal horn core is wanting in the type, having been lost at the suture between it and the nasals. It must, however, have been considerably compressed transversely. The supraorbital horn cores are slender and much more erect than in most species, some- what elliptical in section at the base and more nearly circular in their mid-length. The orbit is large and irregularly elliptical in outline, its long axis running obliquely down- ward and forward. The position of the orbit is in advance of and superior to that of 7. (Sterrho- lophus) flabellatus. The lachrymal foramen lies between the nasal and maxillary, as in the last-mentioned species. 170 EE) CMRATORSIAS The general form of the skull is long and low, the frill, which is twice as broad as long, being but little elevated behind, in contrast to most of the other species except flabellatus. - The median ridge of the parietal region is elevated and bears four rugose prominences, wherein 7. serratus differs from flabellatus and elatus, but agrees with prorsus and brevicornus. The number of epoccipitals is seventeen, five pairs being borne on the squamosals. In this, as in other features, the type specimen agrees with No. 970 of the American Museum, a much larger though still immature skull which the present writer? has referred to the same species. 5. Triceratops (Bison, Ceratops) alticornis Marsh (pp. 115-116, fig. 106) is known only from a pair of remarkable supraorbital horn cores (No. 18718, U. S. National Museum), first referred by Marsh to Bison and later to Ceratops. Hatcher has removed the species to the present genus for reasons which are at once evident when one contrasts the generic characters of Ceratops and Triceratops, especially with reference to the relative development of the elements represented by the type. The supraorbital horn cores are long, with slender-pointed ends, curving forward and outward, then upward. These horn cores resemble possibly those of serratus more than any other species and rep- resent the most highly specialized type within the genus. The frontal region is broad, somewhat convex, and very rugose. 6. Triceratops sulcatus Marsh (pp. 133-134, figs. 112, 113, Pl. XX XVII) is based upon some skull and skeletal fragments (No. 4276, U.S. National Museum) of a very large, fully adult animal. In the original description Marsh describes a deep groove on the posterior surface of the upper half of the supraorbital horn core. Unfortunately but one horn core is now available, and it is incomplete, evidently having been broken off and healed over during the life of the individual. This core does not show the distinctive groove, and Hatcher is inclined to give the presence or absence of the groove but little specific weight. The horn was ovate in cross section and as preserved shows no distinctive characters. This is also true of the humerus and vertebre which are preserved. On the whole there seem to be no characters in the fragmentary material representing the type which afford a basis for a true specific diagnosis. It would be well, therefore, to await the discovery of additional material before deciding as to the validity of this species. 7. Triceratops elatus Marsh (pp. 134-138, Pl. XLIII) is based upon a scarcely adult specimen of large size (No. 1201, U.S. National Museum). The rostral bone is of moderate size and is not coossified with the premaxillaries, the upper branch.being free. The nasal horn core is short and stout, its posterior border not being continuous with the upper surface of the nasals as in most species. The apex rises but little above the superior border of the nasals. Another skull (No. 4805, U.S. National Museum) has a horn intermediate between that of the type and that of T. calicornis, evidently the nearest ally to the species under discussion. The supraorbital horn cores are long and massive, curving strongly forward as in T. cali- cornis, much compressed laterally except toward the apex, where they become nearly circular. The orbit is large, elliptical, with a somewhat oblique axis. The inferior process of the jugal is directed downward and backward at an angle of 45°, in contrast to other species, wherein it is more nearly vertical. The frill is much elevated toward its posterior margin and differs from that of other species in the absence of undulations along the margin of the squamosal, this bone being rather sharp edged and regular in outline. Epoccipitals were borne on the parietals and the posterior half of the squamosals only. The number of epoccipitals was fourteen, four pairs on the squamosals and six on the parietals, without the usual median one. a Bull. Am. Mus. Nat. Hist., vol. 19, pp. 685-695, Pl. LIX. GENERIC AND SPECIFIC SUMMARY. tel 8. Triceratops calicorms Marsh (pp. 138-139, Pls. XX XVIII—-XL) is based upon a large but immature skull, jaw, dorsal vertebre, ribs, alin, sacrum, etc. (No. 4928, U.S. National Museum). The specific affinities would seem to be with the preceding species, 7. elatus, from the similarity of both nasal and supraorbital horns and of the frills. The main distinctions lie in the proportions of the skull of the present species, as the premaxillaries are here much longer and the rostral bone larger, with a much fuller anterior curve, thus making that portion of the skull in advance of the nasal horn proportionately much more pronounced. The nasal horn core is low, almost vertical, convex in front and concave behind. T. calicormis agrees with T. elatus in having the small nasal horn core sharply marked off from the nasals, not apparently continuous with them as in most species. The dleseandhine process of the jugal differs from that of T. elatus in eine more nearly vertical, as in other species. This may, however, as Hatcher says, be due to difference of pressure in the two types. The frill is only in part preserved, but the squamosals are almost entire. They agree with those of 7. elatus in lacking marginal undulations and apparently bore but three or four epoccipitals on the posterior portion, which are not ankylosed. The squamosals are longer and narrower, a fact which aids in giving to the entire skull a greater proportionate length than that of 7. elatus. The lachrymal foramen is between the nasal and maxillary in both species, thus agreeing with serratus and flabellatus. Hatcher expresses the opinion that 7. elatus and T. calicornis may prove synonymous. 9. Triceratops (=Sterrholophus) jflabellatus Marsh (pp. 143-148, figs. 6, 8-11, 15, 17-23, 28, 31, 33, 36, 38, 39, 43, 44, 60-62, Pls. XLIV-XLVI) is known from the nearly complete though disarticulated skull of a very young individual (No. 1821, Yale Museum). This species Marsh made the type of a new genus, Sterrholophus, because of the character of the frill, but, as shown above (p. 164), the peculiar absence of vascular impressions from the latter may be Rone dered an adolescent character, which may also be said of each presumably generic feature exhibited by the skull. The nasals are massive, but not yet coossified, nor was the nasal horn core ankylosed, so that this important feature is lacking. The supraorbital horn cores are rather long, laterally compressed near the base but more nearly circular in section toward the summit. They are inclined forward and outward at a less angle than in other species. The orbit is elliptical and the lachrymal foramen lies entirely within the maxillary bone as in T. prorsus. The long axis of the frill exhibits less of an upward curve than in other species except T. serratus. The squamosals are rather short and broad, as in 7. serratus, though the entire frill viewed from above is proportionately much narrower because it is more highly arched. T. flabellatus bore upon its frill the maximum number of epoccipitals known, nineteen, of which six pairs are attached to the squamosals, the remaining seven upon the coalesced parietals. 10. Triceratops obtusus Marsh (pp. 140-141, figs. 116, 117) is known from portions of a skull (No. 4720, U.S. National Museum) representing a large, adult animal. : The nasals are very broad and the nasal horn core is reduced to a broad, rounded, rugose prominence marked with a number of deep vascular grooves. The dentary is described as exceptionally deep and massive, and the teeth are unusually large. The extreme forward extension of the mandibular fossa seen here is exceptional among Ceratopsia. a Le 172 THE CERATOPSIA. RELATIONSHIP OF SPECIES, If one were to group the species of Triceratops according to affinities the result would be expressed something as follows: Group 1. Group 2. T. horridus. T. elatus. T. prorsus. T. calicornis. T. brevicornus. T. serratus and T. flabellatus would each stand alone, although in some features they suggest each other. Of the species alticornis, sulcatus, and obtusus the skulls are too fragmentary for a fair comparison. T. obtusus, because of its greatly reduced nasal horn, and alticornis, from the development of the long, slender supraorbitals, seem to be the most specialized. Il. THE CERATOPS-TOROSAURUS PHYLUM. CERATOPS. Of the four species mentioned under this genus in the alphabetical list on pages 11-12, two, C. (Bison) alticornis and C. horridus, have been removed to the genus Triceratops, with which their affinities clearly le, and one, C. (Hadrosaurus) paucidens (pp. 103-104), is of question- able validity, for, as Hatcher says: Tt is not at all unlikely that the type of the present species [a maxillary and premaxillary] pertained to one of the several species of Ceratopside already described as from the Judith River beds. Since, however, the teeth, the maxillaries, and the pre- mazxillaries of all these are imperfectly known, it is at present impossible to determine to which of them this specimen should be referred. Thus there remains of the original four but one, the type species O. montanus, to which _ Hatcher has added three which he removed from the genus Monoclonius. The list of species as revised by Hatcher is as follows: 1. Ceratops montanus Marsh. Type. 2. Ceratops canadensis Lambe. 3. Ceratops recurvicornis® Cope. 4. Ceratops belli Lambe. 1. Ceratops montanus Marsh (pp. 100-102, figs. 103 and 104), the type species, is based uponan — occipital condyle and a pair of supraorbital horn cores, No. 2411, National Museum. The supraorbital horn cores were of moderate length, subtriangular in section at the base, __ but nearly circular in the upper half. They are not compressed, but curve strongly outward and slightly forward. The frill is unfortunately unknown in the type species. 2. Ceratops (Monoclonius) canadensis Lambe (pp. 93-96, figs. 96, 97, Pls. XVIII, XIX, XXIID is known from portions of a skull and an anterior dorsal (No. 1254 a, b, c, d, e of the Canadian Geological Survey). Hatcher removed this species from the genus Monoclonius — because of the similarity of the supraorbital horn core with that of the type species of the present genus. He does not, however, give us his final decision as to whether it is to be con- sidered specifically distinct from C. montanus or not, as that point was left for the present discussion, which Hatcher unfortunately did not live to undertake. The only distinction must lie in the character of the supraorbital horn cores, which, although Hatcher considers them “‘essentially the same,’’ seem to differ in that the base is subtriangular in section in C. montanus and circular in C. canadensis. Another distinction lies in the curvature of the horn. In canadensis the horns curve well forward, then upward, but in montanus they are more nearly straight and, if Hatcher’s arrangement expressed in fig. 103 is correct, flare outward strongly in a most peculiar manner. | The nasal horn core was lost before fossilization, as the creature was young, so that its char- acters can not be determined. a The present writer questions the reference of recurvicornis to this series for the reasons given on page 173. ee ee ee ee ee ee ee ee ee ee ee eee ll ee i i ti Be ee ee a | GENERIC AND SPECIFIC SUMMARY. ie The squamosal is somewhat triangular, flat, and moderately thin, the outer edge rounded and wavy in outline. One can not contrast it with those of the other species, for with them the element is unknown. (Pl. XXIII.) All that is preserved of the parietals is a slender bar, triangular in section, which formed the external boundary of the parietal fontanelle. This outer bar is very similar to that of Torosaurus gladvus. Hatcher says with reference to this species: T have little hesitancy in asserting that the squamosal and frontal horn cores of the present species were associated with a parietal of the same general type as that described by Lambe and referred to Monoclonius belli, and I am of the opinion that the two may be specifically identical, although from the material at hand it is impossible to determine this point with certainty. 3. Ceratops (Monoclonius) belli Lambe (pp. 96-97, fig. 98, Pl. XX) is “woumeed! upon a pair of parietal bones (No. 491, Canadian Geol. Survey). The species was decribed as Monoclonius by Lambe, but was removed to the present genus by Hatcher, for, as he says: After a careful study of the type of the present species, together with that of I. canadensis Lambe, one can not avoid being convinced as to their generic identity with Ceratops montanus Marsh, while at the same time the great dissimilarity shown in the parietals and squamosals of these species when compared with the same elements in Monoclonius dawsoni Lambe affords evidence additional to that already pointed out as obtaining in the frontal horn cores, in favor of the generic distinction of the three former species from that of the last-mentioned species. The character of these parietals has already been sufficiently discussed, especially since a specific comparison can not be made, as the elements are unknown elsewhere in the genus except for the outer parietal bar of the type of canadensis. The material now available affords scant ground for specific distinctions between C. mon- tanus, C. canadensis, and C. belli, and future discoveries may show that these are synonymous. INCERTZ SEDIS, 4. Ceratops ( Monoclonius) recurvicornis Cope (pp. 81-87, fig. 90) is based upon fragments of the skull of a young individual (No. 3999, American Museum of Natural History). Nasal horn core massive, curving strongly forward, heavier than the supraorbitals and much more compressed distally. The nasal bones narrow rapidly anteriorly, so as to appear wedge shaped when viewed from above. The supraorbital horn cores are short, stout, abruptly pointed, and compressed at the sub- triangular apex. The horns are straight, almost erect, in sharp contrast with those of other Ceratops species, and are relatively much smaller, though they show a decided advance over those of Monoclonius. There seems to have been a lesser horn core just in front of the supra- orbital horn, subtriangular in cross section and directed forward. This feature is unknown elsewhere. The squamosal is stout and of considerable size, with no vascular grooves, evidently a juve- nile character, and the edge bears prominences which in turn bear epoccipitals separated by sutures from the squamosal bone. This species has been removed by Hatcher from the genus Monoclonius mainly because of the supraorbital horn cores, which resemble those of Ceratops montanus much more than those of Monoclonius crassus. The squamosal is so different from that of the type species J/. crassus that, when taken into consideration with the structural differences found in other portions of the skull, especially the horn cores, Hatcher does not hesitate to refer it to the genus Ceratops. The presence of separately ossified epoccipital bones in this species, which do not seem to occur elsewhere either in the genus or phylum, together with the size, which is decidedly greater than that of most of its contemporaries, added to the differences between the supraorbital horn cores and those of other members of the genus Ceratops, leads the present writer to doubt the correctness of Hatcher’s conclusions in the matter. It should doubtless be removed from Mono- clonius, but it seems rather to represent a new genus in the Triceratops phylum. In the light of our present knowledge, however, this matter can not be settled. _ —— SS tCS:S 174 THE CERATOPSIA. TOROSAURUS. 1. Torosaurus latus Marsh (pp. 150-152, fig. 118) is based upon a specimen (No. 1830, Yale Museum) consisting of an incomplete skull of an aged individual. Nasal horn core broad at base, sharp above, directed upward and forward. Supraorbital horn cores much compressed, ovate, with apex forward. Horn cores directed upward and out- ward and forward. Supratemporal fossz relatively larger than in T. gladius. Skull broad and massive between the orbits. Parietal fontanelles not entirely within the parietals, in contrast to those of T. gladius. Hatcher says that he can not verify this feature in latus. The postfrontal fontanelle paired, as opposed to single median one in T. gladius. 2. Torosaurus gladvus Marsh (pp. 152-155, figs. 7,12, 14, and 119) is based upon a type specimen (No. 1831, Yale Museum) consisting of detached portions of a skull. Nasal horn core very short, stout, compressed, with a sharp apex. It is very rugose, the section being an oval with the rounded portion in front. The supraorbital horn cores are rather long, slender, somewhat compressed laterally, and very rugose. They were directed forward and outward. The postfrontal fontanelle seems to have been single, as opposed to the paired ones in T. latus, but the broken condition of the skull in this region leaves this point somewhat in doubt. The supratemporal fossz are proportionately smaller than in T. latus, and the fontanelles — axe entirely surrounded by the parietal bones. Hatcher says: “‘The type of the present species represents the extreme development of the form of parietal crest that is peculiar to this genus,” though 7. gladius is geologically the older of the two species. GEOLOGY AND PHYSIOGRAPHY OF THE VARIOUS CERATOPSIA LOCALITIES.4¢ The remains of Ceratopsia, though referred by various authors to each of several horizons, are referable either to the Judith River of Montana and its equivalent, the Belly River of Canada, or to the Laramie of Wyoming and Montana and the Denver and Arapahoe formations of Colorado. The principal geographical localities, as well as one or two minor places from which — Ceratopsia remains have been reported, are shown on the accompanying map (PI. L). The vertical range of the Ceratopsia is limited to the upper half of the upper Cretaceous. JUDITH RIVER LOCALITIES OF MONTANA. CORRELATION. The conclusions reached by Messrs. T. W. Stanton and J. B. Hatcher?’ concerning the — stratigraphic position of the Judith River beds and their correlation with the Belly River forma- tion are of prime importance in considering the phylogeny of the Ceratopsia. These general conclusions, based upon a careful resurvey of the upper Cretaceous formations in Montana and the adjacent parts of Canada having special bearing upon the problems in hand, are as follows: 1. The Judith River beds are distinctly older than the Laramie, being separated from the latter by at least several hundred feet of marine shales, identical in their faunal and lithologic features with the Pierre, to which we have given the local name of Bearpaw shales, from the Bearpaw Mountains, about which they are well exposed. “2. The Belly River beds of Canada are identical with the Judith River beds of Montana. The name Judith River beds having priority should be the accepted name for this formation, and the terms Belly River and Fish Creek beds should be dropped. 3. The marine sandstones and shales immediately underlying the Judith River beds do not represent either the Benton, as some Canadian geologists have supposed, or the Fox Hills and upper Pierre, as most geologists of the United States who a Based upon the writings of Hatcher, Stanton, Lambe, Eldridge, Cross, and upon the writer’s own explorations. b Geology and paleontology of the Judith River beds: Bull. U. S. Geol. Survey No. 257, p. 66, 1905. ———- -— —- , | f j CERATOPSIA LOCALITIES. 179 have examined them have believed, but they constitute a distinct horizon within the Montana group, which we have called the Claggett formation, from old Fort Claggett, at the mouth of the Judith River, near which they are well developed. 4. The Eagle formation, from its stratigraphic position and faunal relations, marks the base of the Montana group in this region. 5. The Bearpaw shales, the Judith River beds, the Claggett, and the Eagle formations all belong to the Montana group, and together probably form the equivalent of the Pierre, as that term is generally understood, though the possibility is recog- nized that in the typical area the Pierre may have more restricted limits. The following correlations are shown: * Sections in South Dakota, Montana, and Assimboia. South Dakota Central and northern Southern Assiniboia section. Montana section. section. Laramie. Laramie? Laramie? Fox Hills. Fox Hills? Fox Hills? Bearpaw. Bearpaw. outialars Eoup. Judith River. Belly Riven Pierre. (Judith River). Claggett. Claggett. Eagle. (2) (ase Niobrara. Colorado group. | | Benton. Benton. (2) | | i Dakota. Dakota? (2) The Montana and Colorado groups are generally recognized as larger subdivisions of the strata lying between the Dakota and the Laramie. The South Dakota and Nebraska section is the Meek and Hayden section with the Laramie added, while the other two columns represent the sections studied by us [Stanton and Hatcher]. The queries in the lower part of the columns indicate formations not seen by us, and the queries in the upper part of the columns indicate our doubts as to the correlation of any particular horizon in these sections with the Fox Hills and as to the limits of the Laramie and its relationship with overlying formations that have been described in this region. DOG CREEK. The first Judith River locality mentioned by Stanton and Hatcher is at Dog Creek, a small stream emptying into the Missouri from the south, about 2 miles below the mouth of Judith River. This creek rises in the Moccasin Mountains and flows for 25 or 30 miles northward, first ‘in a narrow and shallow valley, through grass-covered table-lands, and for the last 12 or 15 miles through a deep, rugged canyon, with wild, deeply dissected badlands on either side. The walls of the upper valley are composed of Judith River sandstones and shales; those of the deeper lower canyon are mainly the light ash-colored sandstones and darker shales and clays of the Judith River beds above and ‘‘the darker bufl-colored sandstones and dark sandy or black clay shales of the underlying Claggett formation’’ below. For 5 to 8 miles from the mouth of Dog Creek the beds in the bluffs of the stream are undisturbed, and in this stretch the Claggett underlies 300 or 400 feet of Judith River formation. A little farther up Dog Creek, in undisturbed areas, the Claggett formation disappears beneath the bed of the stream, and the bluffs of the canyon are formed entirely of the Judith River beds, which have a maximum thickness here of perhaps 500 feet and are composed below of alternating layers of light, ash-colored sandstones and darker shales, abounding in numerous fresh-water Mollusca. a Loc. cit., p. 63. b Stanton and Hatcher, op. cit., pp. 37-38. 176 THE CERATOPSIA. Among the ceratopsian remains mentioned are those of Ceratops and Monoclonius, and it H| was here, or in this vicinity, that type material referred to Ceratops paucidens Marsh was found, the exact locality bemg— | ~ on the western slope, very near the summit, of a rounded badland hill about 20 rods east of the spring situated about one- 1! quarter of a mile east of the}freight road running from Judith to Maiden, Mont., * * * about 12 miles from Judith post- office, on the Missouri River. [See this monograph, p. 103.] it The type of Ceratops (Monoclonius) recurvicornis Cope was found in a bluff on the north i side of Missouri River nearly opposite the mouth of Dog Creek,in Montana. According to Hi Cope the geological horizon was near the base of the Judith River beds as these are represented {i in this locality. BIRCH CREEK. Birch Creek rises in the Bearpaw Mountains and flows southward, emptymg imto the | Missouri River nearly opposite the mouth of Dog Creek. The walls of the canyon of Birch Creek | are composed of Judith River and underlying formations, while the hills on either side of the | | Fie. 121.—Undisturbed Judith River beds, with overlying Bearpaw shales, on Cow Creek, Montana. After Stanton and Hatcher. The | | contact is below the pine trees on ridge in middle distance. } } i} canyon are formed of the Bearpaw shales. It was from the sandstones near the base of the == i Judith River beds near the mouth of Birch Creek that Professor Cope, in 1876, secured the type of Monoclonius crassus. a The type of Monoclonius sphenocerus Cope was found by Charles H. Sternberg on the Missouri River near Cow Island, Montana; level not recorded. COW CREEK. i Cow Creek, which yielded the type of Ceratops montanus Marsh, flows south from the — Wit Bearpaw Mountains, emptying into the Missouri about 30 miles below Judith post-office. Pp ) ptyims 1» ; Cow Creek flows in a deep, rugged canyon from a place a short distance below the point where it leaves the Bearpaw | r Mountains to its confluence with the Missouri. * * * The bluffs on both sides of Cow Creek for several miles above its mouth are made up largely of rocks belonging to the Claggett formation, overlain by the lighter colored materials of the Judith River beds. Hi These exposures are similar to those of the same formations on Dog and Birch creeks, | and frequent faulting exposes here and there the underlying Eagle sandstones and Benton i shales. ¥ ss 5 4 2 CERATOPSIA LOCALITIES. uCett About 10 miles above the mouth of Cow Creek, at the point where the old Fort Benton and Cow Island freight road leaves the creek and turns westward toward the Bearpaw Mountains, there is a conspicuous fault in the Judith River beds. Immediately south of the fault * * * a prominent ridge composed of sediments belonging to the Judith River beds projects into the valley of the creek. The type of Ceratops montanus [pp. 100- 102] was obtained near the summit of this ridge. Accompanying Ceratops montanus were Trachodon mirabilis, crocodiles, fishes, turtles, and a number of fresh-water invertebrates. JUDITH RIVER (BELLY RIVER) LOCALITIES IN CANADA. It is to the explorations and consequent publications by Mr. L. M. Lambe, vertebrate paleontologist to the Geological Survey of Canada, that we are indebted for a considerable enriching of our knowledge concerning the earlier Ceratopsia. = = 2 sa poo ee { Fic. 122.—View of the west side of the valley of Red Deer River, Alberta, showing the lower Belly River beds. After Lambe. Stanton and Hatcher, in speaking of the Belly River exposures, say: The exposures examined by us in Canada are all in the southeastern portion of the large, continuous area of Belly River beds mapped by Dawson, McConnell, and Tyrrell. They include both the top and bottom of the formation, as well as good exposures of the overlying and underlying beds, and hence give a fair idea of the formation as described by Dawson. The principal localities that have yielded the Belly River vertebrate fossils described by Lambe are on Red Deer River some distance north of the most northern point visited by us, but we have no doubt that they are on the same horizons which we studied. The genera and species of Ceratopsia described by Lambe were collected from the Judith River (Belly River) beds in the Red Deer River district of Alberta, in the summers of 1897, 1898, and 1901. Mr. Lambe says of these explorations: In 1897 the writer descended the Red Deer River, starting from the village of Red Deer (in Alberta), and made collections from the Edmonton subdivision of the Laramie, between Red Deer village and Willow Creek, and from the Belly River series between Bull Pound Creek and Dead Lodge Canyon. aContr. Canadian Pal., vol. 3 (quarto), pt. 2, p. 25. MON XLIx—07——12 178 THE CERATOPSIA. He says, further: In this year, however, it was found that the best results were obtained in the Belly River series in the vicinity of Berry Creek. Accordingly this locality was revisited in 1898 and again in 1901 and collections made from the Belly River series ' only, in an extensive area of “badlands” on either side of Red Deer River between Berry Creek and Dead Lodge Canyon. | This locality is in Alberta, just west of the border between that Province and Assiniboia. In speaking of the Belly River series Dr. G. M. Dawson? says: In the region of the Bow and Belly rivers the Pierre is underlaid by an extensive fresh and brackish water series, consisting of sandy argillites and sandstones; the upper portion is characteristically pale in tint, the lower generally darker and yellowish or brownish. This has been called the Belly River series and appears to correspond precisely to that occupying a similar stratigraphical position on the Peace River and there designated the Dunvegan series. These indicate the existence of a. prolonged interval in the western Cretaceous area, in which the sea was more or less excluded from the region and its place occupied for long periods by lagoons or fresh-water lakes. | Fig. 123.—View in the valley of Red Deer River, Alberta, upper (primitive mammal) Belly River beds, on east side of the stream south of Berry Creek. After Lambe. Mr. Lambe, in a letter to the writer dated May 30, 1905, writes: All my specimens are from near the mouth of Berry Creek, on the Red Deer River, and the rocks there exposed, I was of the opinion, belonged to the upper or “‘pale” portion (of Doctor Dawson). * * * On either side of the Red Deer River below the mouth of Berry Creek there is an area of “badlands” about 6 to 8 miles, roughly, in diameter. * * * I camped : on the river bank on both sides and obtained all the types in this area. In this upper half of the series there is a great similarity in the beds all through, but for convenience of reference in my field notes I refer to further subdivisions which I have named lower, middle, and upper (primitive mammal) beds. The upper photograph in my memoir,? facing page 25 [fig. 122], shows a characteristic view of the lower beds. The other, the lower photograph [fig. 123], is taken from the level of the upper or primitive mammal beds which at this locality [Red Deer River, at and below mouth of Berry Creek] apparently reach the prairie level, which is seen in the photograph as the distant horizontal line. It is from this upper level that I obtained the type of Ptilodus primevus. a Descriptive sketch of the physical geography and geology of the Dominion of Canada, by A. R. C. Selwyn and G. M. Dawson, Montreal, 1884, p. 40. b Contr. Canadian Pal., vol. 3 (quarto), pt. 2. oar > ‘ CERATOPSIA LOCALITIES. JETS) The specimens representing the species of Ceratopsia may be placed with regard to their horizons as follows: Catts IN@L OW, COR TOSITIS CTE 6 Seon Se see OR es sae aaa See SSS eel a een Ta ertieds: Cat Nowsl 2 54 aponoclonisicanadens isin evar Se ae fee ee ee anaes 1) be Cate Nosmlilii3ueMonocloniusidaiwsontaer ese erise a aaa eee ee cays aoc seo melee ee Sas ees Middle beds. (These beds snould possibly be included with the upper beds.) Cat. No. 491, Monoclonius belli... .-- vnc nee eee eee eee ee eee eee sees eee sees Reet i aneaanedse Gini; NOs, GG, WVRY Cie, SUQTOEaRS CONGNS <= Spon sogeenee Se seta sS Seo eee Se See aaa The types, as regards their geographical position, are as follows: 971 (Centrosaurus apertus). West side of Red Deer River, July 26, 1901. Not far from it were remains of Trionyx foveatus, Leidy. 1254a, etc. (Monoclonius canadensis). East side of Red Deer River, August 2, 3, and 20, 1901. At same level and near remains of Trachodon. 1173 (Monoclonius dawsoni). East side Red Deer River, August 15, 1901. 491 (Monoclonius belli). East side Red Deer River, below the mouth of Berry Creek, August 13, 1898. Not far distant and at the same level were found remains of Trachodon, Cimoliasaurus, and turtles. 515 (Stegoceras validus). East side Red Deer River, below the mouth of Berry Creek, August 15, 1898; 1423, east side, August 24, 1901. At about the same level as M. belli, Ormthomimus altus, Adocus variolosus, Baéna antiqua, Trionyx foveatus, etc., Ptilodus primevus, Myledaphus bipartites, Lepidotus occidentalis, Diphyodus longirostris, etc. There seem to be no very distinct lithological differences on which to base these subdivisions of the Red Deer River rocks, but the lowermost beds seen in the area below the mouth of Berry Creek include some yellowish sandstones, which may be the uppermost beds of Doctor Dawson’s lower yellow portion. LOCALITIES FOR LARAMIE CERATOPSIA. The main localities for Laramie Ceratopsia are three in number—one near Black Buttes station on the Union Pacific Railroad, in southwestern Wyoming; a second, by far the most notable, in the northeastern part of Converse County, Wyo., in the area shown on the map (Pl. LI), and a third in the canyon of Hell Creek, Montana, a tributary of the Missouri, about 135 miles northwest of Miles City. THE BLACK BUTTES, WYOMING, LOCALITY. The Black Buttes locality is mteresting historically as having produced one of the first specimens of ceratopsians, the type specimen of Agathawmas sylvestris Cope. It lies in Sweet- water County, in southwestern Wyoming, not far from Black Buttes station on the Union Pacific Railroad. Stanton and Knowlton? describe this locality as follows: The most prominent feature of the section at Black Buttes is the massive bed of sandstone, somewhat over 100 feet thick at the base of the exposure, forming steep hills and cliffs northeast of the railroad opposite the station and passing beneath the surface by its dip of 9° or 10° near the coal mine. * * * ‘The original specimen of Agathaumas sylvestris was found about 20 feet above it. The bones were found, according to Cope, in a bed of sandstone that lies just above the lower stratum of the Bitter Creek series of coals and is overlain by two other coal seams. This bed of sandstone “‘crops out high on the bluffs” a half mile east of the station. From the molluscan fauna it is judged that the beds below and above the dinosaur bed consisted mainly of deposits from brackish water, with alternations of fresh-water deposits and of coal seams, probably implying coastal swamps with abundant vegetation in which frequent slight changes of level occurred, bringing in brackish waters during periods of subsidence and fresh waters during periods of elevation. This would produce physical conditions in keeping with our conception of the environment of the Ceratopsia. (See p. 194.) THE CONVERSE COUNTY, WYO., LOCALITY. The history of the discovery by Mr. Hatcher of this, the most important Ceratopsia locality, has already been stated (pp. 7-9), and a description of the location and character of the deposits is given in an earlier paper by Hatcher (Am. Jour. Sci., 3d ser., vol. 45, Feb., 1893, pp. 135-144), from which I shall quote at length. The beds were reached by Hatcher by going north from the town of Lusk, Wyo., on the Fremont, Elkhorn and Missouri Valley Railroad; now they a Bull. Geol. Soc. America, vol. 8, p. 143. (See also pp. 105-106.) Hi 180 THE CERATOPSIA. are more readily accessible from Edgemont, S. Dak., on the Burlington and Missouri division of the Chicago, Burlington and Quincy. The Ceratops beds first appear about 25 miles north of Lusk— occupying the summit and northern slope of a yellow sandstone ridge extending in a westerly direction from Buck Creek to Lance Creek and crossing the latter stream near the mouth of Little Lightning Creek, a small tributary from the west. A short distance west of Lance Creek the Ceratops beds pass under other beds composed of very similar material and presumably of Cretaceous age. From Buck Creek the eastern border of the Ceratops beds has been traced in an almost continuous exposure extending northeastward to the Cheyenne River and crossing this stream a short distance below the mouth of Lance Creek. From this point it takes a more northerly direction and, skirting the western slope of the Black Hills, it has been . traced to the north line of Converse County and on into Weston County. * * * The Ceratops beds were originally confined to the western slope of the Black Hills and of the less elevated series connecting the latter with the Rawhide Range. * * * all In no instance have the Ceratops beds been observed east of the Black Hills or their less elevated continuation to the southwest. | The Ceratops beds proper—that is, those beds containing remains of the Ceratopside—are known to have a surface | exposure in that portion of Converse County embraced within their eastern and southern border, as defined above, and a — line extending from that point on the latter where it passes under the overlying beds a short distance west of Lance Creek, 1) nearly due north to Weston County, i. e., the country drained by lower Lance, Lightning, Cow, Doegie, and Buck creeks, and that portion of the Cheyenne River and its tributaries between the mouth of Lance Creek and the north line of Converse County. ii Most of these creeks are shown on PI. LI, which was originally drawn by Hatcher to accom- Ii pany the paper just cited, but was not published until later. ° | The Ceratops beds are made up of alternating sandstones, shales, and lignites, with occasional local deposits of limestones Hit and marls. The different strata of the series are not always continuous, a stratum of sandstone giving place to one of shales | ° and vice versa. ‘This is especially true of the upper two-thirds of the beds. The lack of continuity has rendered it well nigh | impossible to establish any definite horizons in the upper members of the series. All the deposits of the Ceratops beds of this | region bear evidence of having been laid down in fresh waters. Among the invertebrate fossils found in them, only fresh- | water forms are known. * * #* tH The sandstones largely predominate in the lower members of the beds. They are always fine grained, massive to well ih] stratified, and nearly white to yellowish brown in color. They are occasionally compact and hard, but for the most part quite | soft and friable. * * * Almost everywhere in the sandstones are numerous concretions of varying size and shape. Some are almost perfect spheres and vary from the size of a marble to 18 to 20 feet in diameter. Others are from a few inches to several feet in transverse diameter and sometimes several hundred feet in length, a cross section forming a nearly perfect circle. Others still are very irregular in form. These concretions usually show no concentric structure, and while they some- times inclose foreign objects, as a Triceratops skull or a single bone as a nucleus, they are for the most part simply centers of solidification and not true concretions. This is frequently shown by the cross bedding in them, so often seen in the sand- stones themselves. * * * ttl The lignites occur in thin seams, never more than a few inches thick, of only limited extent, and with many impurities. At no place in the Ceratops beds of this region have workable coal beds been found. | The exact localities of the type specimens from the Converse County beds are as follows: iti Triceratops horridus at the point marked +1, Pl. LI, on the south side of a canyon entering Hil Buck Creek from the west and about 5 feet from the bottom of the canyon, contained in a | concretion formerly embedded in a light-yellow, soft, heavily bedded sandstone. i Stratigraphically T. horridus came ‘‘from midway between the Fox Hills and Fort Union” of Converse County. Hatcher estimates the stratigraphical range from the locality of the nih type of Ceratops montanus to that of the present species as 3,500 feet, allowing 2,000 feet for _— slit the Bearpaw shales and true Fox Hills sandstones and 1,500 for the Laramie below the I skull level. (See p. 119.) The type of Triceratops (Sterrholophus) flabellatus was found at the point marked +2 in Pl. LI, where ‘‘it lay in a bed of arenaceous shale, at the summit and the extreme western point HH of a high and rocky ridge about half a mile in length, running westward from the main divide i} between Buck Creek and Lance Creek.”’ Hi) Stratigraphically it lay above the position of any other type, with the possible exceptions HiIl of 7. sulcatus and T. brevicornus. a Triceratops prorsus, type, lay at the locality marked +3 in Pl. LI. It was situated on Dry Creek, which empties into Lance Creek from the west, about 3 miles above the mouth. It lay on the north side, about 100 yards above the location of the type of 7. serratus, in a hard i a Am. Naturalist, Feb., 1896, vol. 33, Pl. III. , CERATOPSIA LOCALITIES. 181 concretion of calcareous sandstone which had weathered out of a thick stratum of soft and almost white sandstone. It was at a slightly lower level than T. serratus. The skeleton 4 referred to 7. prorsus probably comes from nearly the same horizon as the type. The type of Triceratops serratus lay about 20 feet above the bottom of the same side of the draw at the point marked +4, Pl. LI. This skull was in the usual calcareous concretion at the base of a stratum of sandstone. The horizon was slightly above that of T. prorsus and considerably above the type of 7. horridus. Triceratops sulcatus, type, was discovered at the poimt marked +5, on the divide between Dry Creek and Lance Creek. Its horizon was a little above that of T. serratus, but below the level of 7. flabellatus. The locality of the type of Triceratops obtusus is at +9, Pl. LI, about 1 mile east of Lance Creek, near the southern border of the Ceratops beds. The horizon ‘‘would be about the middle of the Laramie, as those deposits are represented in this region.’”’ This would seem to bring the species below the level of Triceratops horridus, which is about the middle of the upper half, and hence make it the lowermost species thus far recorded. Triceratops elatus, type, was found at the point marked +16 on the map (Pl. LI). It lay in loose arenaceous shale a quarter of a mile east of Lance Creek and opposite the mouth of Lightning Creek. The horizon given is about the middle of the Laramie series, which would bring 7. elatus and T. obtusus at about the same level, though, as Hatcher says, owing to the absence of exposures of overtying or underlying deposits in this immediate vicinity it is difficult to determine exact horizons with even moderate precision, and this difficulty is augmented by the frequency with which the sandstones and shales of these deposits replace one another both vertically and laterally, making it extremely difficult to trace any given stratum for any considerable distance. Stanton says the position of T. elatus ‘‘would not be more than 300 to 400 feet below the highest Ceratopsia remains of this area.” Triceratops brevicornus, type, was discovered embedded in a hard sandstone concretion in the divide between Lance and Lightning creeks, 3 miles above the mouth and 14 miles south of Lightning Creek, indicated by +22, in Pl. LI. Stratigraphically it lay near the summit of the Laramie deposits, hence is possibly the highest of the known species. Stanton thinks that “its position can not be very much higher than TY. prorsus, serratus, and sulcatus.” Triceratops calicornis, the remaining Converse County species, was found at +29, Pl. LI, about a mile east of the abandoned U-—L ranch, which is at the junction of Dry and Lance Creeks. It was embedded in a stratum of rather hard sandstone. Hatcher gives us no clue as to the horizon of this specimen. Stanton says, ‘‘It is apparently from about the same horizon as T. sulcatus.’ Of the genus Diceratops, with its single species hatcheri, the type specimen was found in a hard sandstone concretion about 3 miles southeast of the mouth of Lightning Creek, at the point marked +25, Pl. LI. No statement as to the stratigraphical position is given. From its locality it can not be far from the level of 7. flabellatus. Of the two species of the genus Torosaurus, one (T. latus) was found in an extremely hard bluish-colored calcareous concretion near the top of the bluff on the north side and about 2 miles above the mouth of Lightning Creek. It lay in the bottom and near the extreme head of a small, dry watercourse at the point marked +19, Pl. LI. From its geographical position Stanton believes 7. latus to be the highest Ceratopsia specimen from this region. The type of Torosaurus gladius came from a horizon considerably lower (200 feet) than that of the preceding species and lay in a thick bed of shale on the northern slope of the divide between Cow and Lightning creeks at + 19A, Pl. LI. Hatcher does not give the stratigraphical position of the two Torosaurus species with relation to those of the genus Triceratops. a The mounted skeleton, No. 4842, U. S. National Museum. 182 THE CERATOPSIA. THE HELL CREEK, MONTANA, LOCALITY. This locality * lies in the northern part of Dawson County, Mont., along the canyon of Hell i] Creek, a tributary flowing northward into the Missouri River about 30 miles above the mouth iit of the Milk River and 150 miles east of the Judith River localities. The country has an altitude iM up to 3,000 feet above sea level and consists of grassy table-lands with occasional flat-topped buttes and, in places along the stream courses, wild dissected badlands. There seems to be no continuous bone-bearing layer, but occasional localities where speci- mens, mainly fragmentary, may be found, some in joint clay, some in unconsolidated sandstone, and again in concretions so typical of the Laramie formation. The American Museum party of 1902 found the remains of thirteen or more skulls, pre- i sumably of Triceratops, but all had weathered out and disintegrated but one, which was intact i except for the nasal horn core, the nasals, and the distal portions of the sigur ome horn i cores, which had weathered away. This specimen (see p. 185, fig. 26) was found on the extreme point of the divide separating Hell Creek from a tributary which entered it from the west about 15 miles from the Missouri River. It was about 35 feet from the bed of the canyon and lay in its natural position. The precise horizon was not ascertained. | Few of the Ceratopsia found in this region were in concretions, although the party unearthed HL portions of the skeleton of an enormous theropod dinosaur (Tyranosaurus rex Osborn), which Mal was contained in several bluish calcareous concretions of extremely hard, homogeneous texture. i The skull, No. 970 of the American Museum, is doubtless referable to Triceratops serratus Marsh. In adlihinor., the party secured portions of the skeleton of another specimen at a point il | about a mile south of the place where the first was found and an equal distance away from Hell Hi Creek Canyon. This was also Triceratops, but the species has not been determined. It was embedded in joint clay, at the base of a large table butte, the upper portion of which had been baked to a terra cotta from the accidental burning out of ‘dhe lignite seams. . Two splendid skulls, one probably referable to Triceratops brevicornus, were found near a Hell Creek, in 1904, by Mr. W. H. Utterback, of the Carnegie Museum in Pittsburg. The | T. brevicornus skull lay in soft sand and was in perfect condition.? Mi DENVER LOCALITIES.c Ceratopsia remains have been found in the vicinity of Denver, Colo., in beds known as the Arapahoe and Denver, considered to be of post-Laramie age. The Arapahoe, the older of the two formations, occupies the site of an ancient lake of considerable extent. i] Along the northern and northwestern edges the formation now appears only as a thin horizontal sheet, or in scattered Mil} outliers upon the uneven surface of the underlying Laramie. Along the western outcrop, where the strata are highly inclined Hi : and confined between underlying and overlying terranes, the formation is 600 to 800 feet thick. * * * The total thickness -of the Arapahoe as originally laid down is undeterminable. The Arapahoe is divisible into two well-marked series of beds; a lower, of sandstones and conglomerates, 50 to 200 feet thick, and an upper, of clay, 400 to 600 feet thick.d The lower member is composed of débris derived from the underlying Carboniferous, the Triassic, the Jurassic, and from the lower divisions of the Cretaceous, up to and including the Hi Laramie. The shales of the overlying member of the Arapahoe are light gray and arenaceous | and contain a few ironstones similar to those of the Laramie. iy The vertebrate remains ‘‘occur in the conglomerate along the foothills and in the basal HI sandstones and overlying clays beneath the prairies.’’ The few specimens from the conglom- erate are worn, while the abundant remains in the clays are finely preserved. The Arapahoe formation is distinguished from the Laramie by the sandy nature of its clays, by the comparative paucity i of its ironstones, by the generally brighter colors, and by the vertebrate remains. From the overlying Denver the Arapahoe is readily distinguished by the eruptive nature of the material composing the former. é a Lull, R.S., Bull. Am. Mus. Nat. Hist., vol. 19, Art. XXX, Dec., 1903. See also this monograph, p. 185. b Ann. Rept. Carnegie Museum for 1905, p-. 24, figure facing p. 64. | c Emmons, Cross, and Eldridge, Geology of the Denver Basin: Mon. U. S. Geol. Survey, vol. 27, pp. 150-254. | d Hldridge, op. cit., pp. 151-152. 1} e Eldridge, op. cit., p. 154. CERATOPSIA LOCALITIES. 183 Within the area of the Denver Basin the Arapahoe formation rests unconformably upon the Laramie, although along its upturned western edge the break is recognized only through change in sedimentation. The Denver formation, which lies unconformably upon the Arapahoe, is very different from the latter in the composition of its rocks, though textually they are similar. The materials of which the Denver beds are composed may be classed as the débris of Archean, of sedimentary, and of eruptive rocks. The Archean débris, which is confined to the upper portion of the series, consists of bowlders, pebbles, and sands, and is similar to that of the underlying Arapahoe. The material undoubtedly came from the Archean lands lying to the west. The sedimentary débris, which is mingled with the Archean in the upper portions of the formation, consists of small sandstone and limestone pebbles derived from the upturned edges of the Mesozoic strata. There are also conglomerate bowlders from the Dakota Cretaceous, while on the plains there is some quartz and feldspar, originally from the Archean but derived immediately from the soft Arapahoe beds. The eruptive débris is the most characteristic feature of the Denver formation, especially . in the lower half, though it ranges to the summit of the beds. These eruptive materials are almost entirely andesitic and imply a period of violent eruptions before the beginning of Denver time. The total thickness of the Denver is estimated at 1,449 feet. Fragmentary remains, which doubtless represent several species of Ceratopsia, have been found in the vicinity of Denver; from the Arapahoe is one identified as Ceratops montanus, that is ‘‘the remains of the same reptile or one nearly allied to it.’’ This identification seems hardly possible, as Ceratops montanus is a Judith River type and is vastly older than the Arapahoe. The fragmentary nature of the fossil precludes accurate determination. The type of Tricera- tops galeus is also from the Arapahoe, but Mr. Hatcher has rejected the species on the ground of inadequate material. From the Denver two species are reported, one referred by Marsh provisionally to Tricera- tops horridus, which may be correct; the other, the type of Triceratops alticornis, by far the most notable ceratopsian from this locality. This specimen was found by Mr. Cannon in the rocks of the largest of the tributaries, which rises on the eastern slope of Green Mountain and enters the Platte River near the Larimer Street Bridge, Denver. The exact locality was about 1 mile from the mouth of the smaller stream. RELATIONSHIP OF THE BLACK BUTTES, CONVERSE COUNTY, AND DENVER BEDS. Stanton and Knowlton” thus summarize our knowledge concerning the relative age of the ‘‘Ceratops beds:”’ Until a few years ago it was the custom to include in the Laramie all of the beds between the Fox Hills and Wasatch formations. In the Denver region the detailed studies of Cross and Eldridge . . . have resulted in the recognition of the Arapahoe (Willow Creek) and Denyer beds separated from the Laramie and from each other by unconformities and distin- guished by marked lithologic features. A revision of the fossil floras of that region has also shown that the Denver beds contain a flora composed of species a large proportion of which are not found in the underlying Laramie. . . . The Denver and Arapahoe beds have yielded representatives of a remarkable reptilian fauna consisting largely of horned dinosaurs of the family Ceratopsidx. The presence of this family in the Ceratops beds of Converse County and probably at Black Buttes has suggested the very reasonable query whether the beds containing them at these places also are not younger than the true Laramie. The facts we have presented relative to the stratigraphy and paleontology of the Black Buttes dinosaur horizon seem to us convincing that it is in the Laramie and near the base of that formation. It is less than 200 feet above the marine Cretaceous, and there is no evidence of a break nor of any abrupt lithologic change. The character of the flora and of the invertebrate fauna also, so far as the species have a distribution in recognized horizons elsewhere, favors its reference to the Laramie. If the dinosaur bed of Black Buttes is not Laramie, then the Laramie is either absent or is represented only by about 100 feet of sandstone. The overlying beds up to and including strata with a Fort Union flora seem to form a continuous series that is indivisible either structurally or lithologically, and we can see no reason for placing the top of the Laramie lower than the base of the lowest bed with a Fort Union flora. Closely similar conditions are seen in Converse County, the principal difference being a greater development of the beds. The sandstones at the base, overlying the Fox Hills, are a few hundred feet thick, and the variable more argillaceous, higher a Bull. Geol. Soc. Am., vol. 8, p. 155. 184 THE CERATOPSIA. beds, with a fresh-water fauna in large part identical with that at Black Buttes and a flora that also indicates the same horizon, have a much greater thickness. Here again there seems to be no break in a series that has Fort Union plants: in its upper member. The abundant occurrence of such a species as Campeloma multilineata throughout all but the lowest portion of the series argues strongly for continuous sedimentation. The difficulty of recognizing unconformities in beds so little disturbed and the possibility that there may be such undis— covered breaks in these two areas is freely admitted, though it does not seem to us probable. From the facts now available it seems most probable that in Converse County and in the Bitter Creek Valley the time representatives of the Denver and Arapahoe beds are undifferentiated portions of a continuous series and can not be separated from the Laramie. The Fort Union beds are apparently distinguishable by means of their flora, and these mark the upper limit of the Laramie in the areas in question. The Triceratops skulls from Hell Creek represent species whose position stratigraphically is about the middle of the Converse County series of specimens, which is evidence in favor of considering the former deposits as contemporaneous with the latter. : Geological sequence of the Ceratopsia. Formation. Locality. Species. Triceratops alticornis. TPL Pe Nes ae Cae ? Triceratops horridus. , [rector alticornis. Denver=.2222 S$. ee8= Near Denver, Colo Arapahoe..........- Near Denver, Colo Triceratops galeus.a ? Ceratops montanus.® Torosaurus latus. Torosaurus gladius. Diceratops hatcheri. Triceratops elatus. : Triceratops flabellatus. Lance Creek beds, Converse County, |{ Triceratops brevicornus.¢ Wyo. Triceratops sulcatus and calicornis. Triceratops serratus.c Waramie: seasceseeee Triceratops prorsus. Triceratops horridus. . ? Triceratops obtusus. Near Black Buttes Station, W yoming.| Agathaumas sylvestris. SES ao ROS GRE ACC asEeaasc paca ates o No Ceratopsia. PN oath SARA, oN en GAR PN kA tes REO ads No Ceratopsia. Ceratops montanus. Ceratops paucidens. Near sadiGhys\ton t= esse aaenen Monoclonius crassus. Ceratops recurvicornis. Tmdbitin Raver hede. . Monoclonius sphenocerus. @ Centrosaurus apertus. f Ceratops canadensis. f Monoclonius dawsoni. 9 Ceratops belli. % Ido ISHN ee ee 5 Bearpaw sesso cee Red Deer River, Alberta ¢ a Species rejected by Hatcher because of the fragmentary condition of the type. > This form is placed amsng the Arapahoe Ceratopsia by Whitman Cross on page 230 of the anoangerepln on the Geology.of the Denver Baan’ : (Mon. U.S. Geol. Survey, vol. 27). It must be a case of mistaken identity (see p. 183 of this monograph). c Also Hell Creek, Montana. d Horizon not recorded. ¢ Belly River beds. The Montana and Alberta species were contemporaneous, f From upper beds. g From middle beds. h From lower beds. Although the formations above the Fox Hills have been placed in regular sequence in — the above table, it may be that the Ceratopsia-bearing strata in Converse County include the equivalents of those. near Denver. THE CERATOPSIA. 185 COLLECTING CERATOPSIAN MATERIAL. In the chapter of the present memoir devoted to the history of discovery reference is made to the various expeditions which have searched for these interesting forms. While not the pioneer ceratopsian collector, Hatcher brought to light by far the major part of all of the known material pertaining to this group. His work was mainly among the larger genera of the Laramie, especially in the Converse County locality, and his experience was such that he could have written a most interesting and instructive chapter upon the difficulties and dangers incident to the collection of such huge fossils. In one specimen collected in Converse County by Hatcher the concretion containing the skull weighed 6,850 pounds when received at the Yale Museum. This is by far the largest speci- men of Triceratops yet found, having an estimated length of 8 feet from the tip of the rostral bone to the hinder margin of the frill. In the same concretion were many other bones and a large fragment of vegetable origin, further evidence in favor of Hatcher’s conception of the habitat of the Ceratopsia. The present author’s experience in the field in search of Ceratopsia, though limited to a single eventful season, was of such a nature as to present a full measure of experience from the variety of conditions met with. This expedition will be taken as an illustration of field methods and the problems to be solved. During the summer of 1902 Mr. Barnum Brown, of the American Museum of Natural History, accompanied by the author, was sent into the northern part of Montana to explore a new locality for Laramie dinosaurs. The locality was made known to Professor Osborn through Mr. William Hornaday, director of the zoological gardens in Bronx Park, New York, who, in company with Mr. L. A. Huffman, a photographer, of Miles City, Mont., had taken a trip up into the Hell Creek region the season before. While there these gentlemen found a ceratopsian horn core, which showed that there were prospects of finding these fossils in this locality. The great majority of the Laramie Ceratopsia found by Hatcher were incased in hard sand- stone concretions, which, while vastly increasing the weight of the specimen and the difficulty both of collecting and of subsequent preparation for exhibition and study, generally insured its preservation from destruction by the action of the elements. As many of Hatcher’s finest specimens had already weathered out of the bank in which they had been embedded, the impor- tance of this fact can readily be imagined. Our party found no fewer than thirteen skulls, but all of these except one had entirely dis- integrated, and the portion of this one that protruded on the surface of the ground was destroyed. This specimen was in unconsolidated sandstone; another was in joint clay; while a third fossil, not a ceratopsian, was inclosed in concretions and gave us a very perfect idea of such collecting. As each of these three matrices presented its special problems for solution, they will be described in some detail. At quarry No. 1, that opposite the camp, a specimen of a huge carnivorous dinosaur (Tyra- nosaurus rex Osborn) was found embedded in a number of separate concretions. The largest of these was but partially buried, and from the exposed end a broken portion of a huge tibia pro- truded. By searching down the hill most of the fragments which had been broken away were recovered and afterwards restored to their original position. The excavation of the concretion itself was easily accomplished, but it was far too heavy for shipment, and the work of reducing it to a more moderate bulk was arduous enough, as the rock was hard and the facilities for sharpening and retempering dulled and broken tools were very crude. The block, which con- tained a tibia and the coossified pubes, weighed 1,700 pounds when all of the matrix possible was removed, and the task of loading it upon the wagon presented a serious engineering problem. A road was cut along the face of the butte to the little valley which divided it from the neighboring hill, to which place the wagon could be brought. The block, carefully swathed in strips of burlap dipped in liquid plaster to protect the protruding bone, was placed upon an 186 THE CERATOPSIA. improvised stone boat and hauled along the road by means of the block and tackle and one of the heavy horses. : When the load reached the wagon the latter was run into two parallel trenches, so as to bring the wagon bed on a level with the ground, when the specimen was hauled into the wagon with comparative ease. The other concretions were all smaller than the first, most of them containing a single vertebra or one or more small bones, but the deeper lying ones were extremely hard, being composed of a perfectly homogeneous bluish sandstone, and had to be removed bedily to New York, where better tools were available, before their final reduction could be accomplished and the contained bones freed from their matrix. Our first Triceratops specimen (quarry No. 2), consisting of vertebree, ribs, the sacrum, the lower jaws, and a few limb bones, lay in a peculiar joint clay of a bluish color, though occa- sionally stained a rusty brown along the joints. The bones were poorly preserved, presenting in this respect a marked contrast with the admirable condition of those of the carnivore, as the joints in the clay ran through the bones as well, which therefore required the most delicate manipulation. The process consisted in the removal of the overlying material, first with a heavy pick and spade, a process technically known as “stripping,” then with a light prospector’s pick, and finally with a harness awl and whisk broom. The exposed bone surface was then treated with a solution of gum arabic to harden it and then covered, first with tissue paper and finally with strips of burlap dipped into flour paste. The bones were then excavated still far- ther, the exposed surface being covered as before, and finally were lifted from their age-long bed and the lower side treated in the same manner. While the smaller bones were now ready for packing, the larger ones had to be provided with a plaster of Paris jacket, sometimes with splints of wood for further support, just as a surgeon prepares a broken limb. The bones were then packed in hay in heavy boxes ready for shipment. Quarry No. 3 contained the great skull of Triceratops serratus, No. 970, before referred to and figured in this monograph (fig. 26, p. 29). This was found, together with some limb and foot bones, about 1 mile down Hell Creek from the camp, and presented a third aspect of Laramie collecting. The specimen in this instance was in unconsolidated sandstone, the skull lying in normal position, with the supraorbital horn cores protruding on the surface and sustaining an abun- dant growth of vegetation, which aided largely in the disruption of the bone. . The nasals, with their horn core, were eroded away, and the small rostral bone lay dis- placed on the right side of the snout, while one complete dentary, with perfect dentition, and other portions of both jaws lay beneath the skull, as though still attached by ligaments when the specimen came to its last resting place. As is usual with Ceratopsian skulls, the upper teeth had almost entirely disappeared, which leads one to believe that the huge head must have had great powers of flotation, owing to the cavities in which the expansive gases incident to decay could develop, and was as a consequence probably the last portion of the creature’s frame to become buried in the sand. Our specimen was in fine condition where it was completely buried, though extremely fragile, and the process of excavation was carried out with the utmost caution, the exposed portions being covered at once with plaster bandages to guard against possible injury. As the broad expanse of frill, measuring in this instance 5 feet in width by nearly 3 feet in length, would not bear its own weight, it was supported from beneath by vertical props as fast as the earth was removed. It was decided to retain the matrix under the palatal portion of the skull for safety’s sake, so it was necessary to spray this repeatedly with gum arabic solu- tion, which hardened it, so that it was self-sustaining. The under side of the frill was then plastered, and by occasional tunneling the matrix was bound fast to the skull by means of the burlap strips. Next a complete trestlework was built of scantling underneath the entire struc- ture, and the whole was bound fast with the plastered strips, so that the resultant fabric was f extremely solid. The box, which was built beneath and around the specimen without disturb-_ ing it in the least, was of the smallest possible dimensions, yet measured 7 feet in length by 53 _ ‘feet in breadth, and about 4 feet in depth, and weighed, with its contents, 3,100 pounds. EVOLUTIONARY SUMMARY. 187 The skull location was on the extreme point of a divide between Hell Creek and one of its tributaries, with a precipitous descent on either side, so that here again it became necessary to construct a roadway along the edge of the canyon wall to the broader part of the divide above, to which the wagon could be driven. The tackle and horse were used, as before, boards being this time laid along the roadway, with small logs for rollers. Even with a large, wide-gage wagon it was necessary to remove the wagon box, as it could not possibly contain the speci- men. Three trunks of trees were laid on the fully extended wagon frame, which was backed up to a convenient bank, on which the hinder end of the logs rested. It was then found that the skull box was too wide to pass between the rear wheels, so they were undermined, and thus lowered beneath the level of the logs. The box was then slid into position with little difficulty, the hind axle was raised with a lifting jack, and earth was thrown beneath the wheels until they again bore the load. The protruding tree trunks were then sawed off, the specimen box was lashed fast to the wagon, and the task of loading was thus completed. The skull was hauled out to Miles City, a distance of 135 miles, whence it was shipped directly to! New York. PROBABLE APPEARANCE, HABITS, AND ENVIRONMENT OF THE CERATOPSIA, AND THE CAUSES THAT LED TO THEIR EXTINCTION. EVOLUTIONARY SUMMARY. The earliest known Ceratopsia are from the Judith River beds, and the race continues upward until the close of the Cretaceous period, during which time it underwent a striking evolution, largely one of size and armament. Of the ancestors of the Judith River forms we have no record, probably because, as Matthew believes (see p. 194), they were dry-land types, notwithstanding the swamp-living habits of their successors. : The Judith River ceratopsians already exhibit the characteristic horns and frill, but the relative proportions of nasal and supraorbital horn cores are the reverse of those of the Laramie types. The nasal horn was the first to develop. It varied somewhat in form, being straight and compressed in Monoclonius sphenocerus and curved strongly backward in M. dawsoni. In the other Judith River genus, Ceratops, it seems to have curved forward instead of backward, and the supraorbital horn cores, which were rudimentary in Monoclonius, are much more advanced in development, though the nasal horn was in all probability still the larger. The frill in the Judith River types is by no means so well developed as in the Laramie forms, and in Monoclonius and Centrosaurus consisted largely of the coalesced parietals, the squamosals taking but little part in its formation, but in Ceratops, though the crest still consists mainly of the widely fenes- trated parietals, the squamosals become a more prominent factor, tending toward the form of the latter bone found in Torosaurus. There is evidence that the teeth of the Judith River forms were somewhat more primitive than those of the later Ceratopsia in that the peculiar method of replacement was not fully assumed in the earlier types. The stratigraphical range from the Judith River beds to those containing the earliest Laramie type, disregarding for the moment Agathawmas, is 3,500 feet according to Hatcher’s estimate (p. 119). This implies a great lapse of time, during which the most striking evolutionary changes occurred, so that Triceratops horridus, the earliest recorded Laramie species, exhibits a marked advance in cranial features over its Judith River predecessors. Here the supraorbital horn cores are vastly larger, while the nasal horn is much reduced, tending to disappear almost entirely in 7. obtusus or Diceratops. Correlated with this change of offensive armor is an increased development of the defensive frill, which in Triceratops is no longer fenestrated, though in Torosaurus it still retains large paired fontanelles in the parietal portion of the crest. The advance in the structure of the Laramie types seems, therefore, to lie in (1) the larger size of the indiwiduals, (2) the preponderance of the supraorbital horns over the nasal, (3) the more perfectly developed parietal crest, and (4) the perfection of the type of dentition peculiar to the group. 188 THE CERATOPSIA. Agathaumas, from the lower Laramie, is known only from certain portions of the skeleton and seems to be transitional between Monoclonius and Triceratops, while Triceratops alticornis, from the Denver beds and therefore the most recent of all, exhibits a type of supraorbital horn core which may readily be considered the final stage in the evolution of these weapons. In the table of stratigraphical sequence (p. 184) the relative positions of the various type specimens are given, except where the level is unrecorded. The phylogenetic series corresponds approximately with the stratigraphic sequence, although of the two Torosaurus species Hatcher distinctly states that gladius ‘‘represents the extreme development of that type of parietal pecu- liar to this genus,”’ thereby implying that J. gladiws, which is geologically the older, is the more specialized of the two. PROBABLE APPEARANCE OF THE CERATOPSIA. THE JUDITH RIVER TYPES. Our knowledge of the general form of the earlier Ceratopsia is still very vague, for with the exception of Monoclonius crassus, among Judith River types, most of the remains are of incom- plete skulls only. But two attempts have been made to restore Judith River forms, consisting of a statuette and a painting, both by Charles R. Knight. The restorations, which differ somewhat from each other, have been called Agathawmas sphenocerus, the supposition being that Monoclonius and Agathaumas were synonymous genera, the latter term having priority. The restoration is really that of Monoclonius sphenocerus Cope. The statuette, the property of the American Museum of Natural History, is figured as Pl. I, fig. 1, in the Catalogue of casts, models, photographs and restorations of fossil vertebrates, by Professor Osborn,’ 1898. The painting, also in the American Museum, was reproduced in the Century Magazine? for November, 1897. The conception of the tall, straight nasal horn and the much smaller supraorbital horns is undoubtedly correct, though in Monoclonius sphenocerus the compressed nasal horn slants some- what backward, while in other species of the genus it curves to the rear. The writer questions whether the hinder margin of the frill was quite so prominent as it is represented in the model and painting, for in the earlier types the whole crest is much less strongly developed than in the later forms, and although the edge is crenulated yet there is little reason to suppose that the marginal armature was as well developed as it appears to be in the restoration. The general bodily proportions are based upon ‘‘a reconstruction of a possibly identical and prior restoration of Triceratops prorsus (Marsh)”’ (Ballou). -It is now known that Marsh was in error in making the presacral series of vertebrz too numerous in his restoration of Triceratops (fig. 125), hence the probabilities are that here again the back is unduly elongated, though of this we haveno positive knowledge. The Judith River Ceratopsia were smaller than their Laramie successors, and of lighter and less muscular build, and from the development of cranial armature differed in their offensive — a and defensive tactics. It is difficult to conjecture in what way Ceratops would differ in external appearance from its contemporary Monoclonius except in the much greater development of the supraorbital horns. The nasal horn core is unknown except in one species, Ceratops (Monoclonius) recurvi- cornis Cope, in which it curves strongly forward, in contrast to that of Monoclonius. As this © species differs from the others that Hatcher refers to Ceratops in an important anatomical feature—namely, in the presence of separately ossified epoccipital bones—the present writer is loth to consider it as typical of the genus, if, indeed, it belongs with Ceratops at all. There are no data concerning the bodily proportions or contour of Ceratops. a See also Science, vol. 7, p. 842, fig. 1. b Ballou, W. H., Strange Creatures of the Past, p. 18. APPEARANCE OF THE CERATOPSIA. 189 THE LARAMIE TYPES. Marsh has given us the first skeletal restoration of Triceratops in the Sixteenth Annual Report of the United States Geological Survey (Pl. LX XI), reproduced here as fig. 125. This restoration is very accurate except for the number of presacral vertebre, which has been overestimated. As Hatcher has shown, in Triceratops brevicornus at least (p. 46, fig. 48), the number of presacrals is not more than twenty-two, while in Marsh’s figure there are twenty- one without the cervicals, which probably number at least eight more. Another feature about which doubt may be expressed is the correctness of the restoration of the feet, for while odd foot bones have been brought to light, no com- plete manus or pes of this interesting animal has yet been found. Knight has essayed at least three restora- tions of Triceratops, the first two, a statuette and a drawing, being based largely, if not wholly, upon that of Marsh. The statuette was made for the American Museum, and a view of it is shown as fig. 26 in F. A. Lucas’s Animals of the Past (McClure, Philips & Co., New York, 1901), while the drawing was also published by Lucas.* The restoration of Triceratops which forms the frontispiece of the present volume was made under Hatcher’s personal supervision, and is one of the most successful of Knight’s artistic reconstructions. The general proportions are markedly different from those of the preceding conceptions of the animal, notably in the shorten- ing of the back and in placing the highest point in the arch of the vertebral column over the sacrum rather than farther forward. The propor- tion of head to trunk is also greater, the skull be- ing almost one-third of the entire length of the animal, including the tail. The length of the tail is conjectural, as a complete caudal series is unknown and it is possible that it may have been reduced somewhat, as it no longer subserved the function of a counterpoise, as in bipedal dinosaurs. The skeleton of Triceratops prorsus which has recently been mounted at the National Museum is that upon which Marsh based his Fic. 124.—Restoration of Triceratops. Rear view of the restoration, and a comparison of the results is of mounted skeleton in the U. 8. National Museum. (See = also Pl. XLIX.) great interest. The skeleton was mounted by Mr. C. W. Gilmore, of the Museum staff, and his description of the specimen is here given: Sais rey a Sone The skeleton of Triceratops prorsus, recently placed on exhibition in the court devoted to vertebrate paleontology, is the first one of this extinct genus to be mounted. As all of the specimens [of Triceratops] referred to above were more or less fragmentary, the most complete one (No. 4842c) (Sk. C, 2082 and 2084 ¢] was used as a basis for the present restoration. The missing parts [including the skull (No. 2100)] were substituted from other individuals of about the same size and belonging to the same species. Where suitable bones were not available, as was the cas in a few instances, these parts were restored in plaster, colored to somewhat resemble the bones, but having the shade differ sufficiently to be easily recognized. Thus we have been able to present a fairly accurate representation of the skeletal structure of this peculiar reptile. Every bone used in aAnn. Rept. Smithsoni n Institution for 1901, p. 644, PI. I. b Gilmore, C. W., A mounted skeleton of Triceratops: Proc. U. S. Nat. Mus., vol. 29, 1905, pp. 433-435. Some words in brackets have been since added by Mr. Gilmore. ¢Catalogue number of the U. S. National Museum. 4d Marsh’s original numbers. 190 THE CERATOPSIA. the skeleton bears its catalogue number, and all plaster bones are marked with a red +. There is thus preserved a definite record of all the associated material comprising the composite skeleton. * * * The skeleton as mounted is standing on a base of artificial matrix intended to represent the color and texture of the Laramie sandstone in which the remains of these animals are found. From the tip of the beak to the end of the tail the skeleton as restored is 19 feet 8 inches in length. The skull, which is6 feet long, equals nearly one-third of this length [a re- markable proportion]. At the highest point (the top of the sacrum) the back is 8 feet 2 inches above the base. The mounted skeleton presents several features which would other- wise be lost to the observer if seen in the disarticulated condi- tion. The short body cavity, the deep thorax, the massive limbs, and the turtle-like flexure of the anterior extremities are characters appreciated only in the mounted skeleton. The position of the forelimbs in the present mount appears rather remarkable for an animal of such robust proportions, but a study of the articulating surfaces of the several parts precludes. an upright mammalian type of limb, [such] as was represented by Marsh in the original restoration. * * * The [tail and] forefeet are perhaps the most conjectural parts of the whole restoration. [The tail is restored almost wholly from Marsh’s drawing of this animal.] Mr. Hatcher, after a careful study of all the forefoot material [of this group] known, was unable to arrive at a satisfactory conclusion as to the arrangement or the number of bones comprising the manus. In constructing these parts we have largely followed Marsh’s drawing, assisted some-— what by forefoot material kindly loaned by Dr. H. F. Osborn, of the American Museum of Natural History, New York City. [It seems probable that a similar condition prevailed in this animal to that found in the other better known dinosaurs, so instead of introducing a full complement of carpal and tarsal bones, as was done in the first restoration, only two elements. were modeled to represent the carpus, while the astragalus. alone comprises the tarsus. ] The nasal horn [core] of the skull used in the present skele- ton appears to be missing, and on account of the unsatisfactory evidence as to whether the horn is wholly or only partly gone,. it was decided not to attempt a restoration at this time. This will account for the absence [or rather the apparent lack of development] of one of the important features upon which the name of the animal is based, Triceratops meaning three-horn face, in allusion to the presence of the two large horns above the eyes and a third smaller horn on the nose. It may be of interest to mention here that Professor Marsh used this skeleton (No. 4842), supplemented by other remains. now preserved in the collections of the Yale University Mu- seum, for the basis of his restoration of Triceratops prorsus, published as Pl. LXXI in the Dinosaurs of North America [fig. 125].¢ Pls. LXIV-LXVIII in the same work were also. largely reproductions of parts of this same individual. A comparison of the above-mentioned restoration by Marsh [fig. 125] with the mounted skeleton [P]. XLIX of this work] shows several differences in points of structure, due chiefly to the better understanding of these extinct forms. The most striking dissimilarity is in the shortening of the trunk by a reduction of the number of presacral vertebrae. Marsh’s error was due to an overestimate of the length of this region, { a mistake also made in his restoration of Brontosaurus I (=A patosaurus), as has been shown by Riggs. Mr. Hatcher determined from a well-preserved vertebral column in the Yale University Museum the number of presacrals as twenty-one, ¢ this being six less than was ascribed to the animal by Professor Whi Marsh. Q i i) ff fj ‘es LJ 2 a After Marsh. Fie. 125.—Restoration of Triceratops prorsus. i) aSixteenth Ann. Rept. U. S. Geol. Survey, pt. 1, p. 386. i b Science, new ser., vol. 17, March 6, 1903, pp. 393-394. | c Doctor Lull now recognizes twenty-two presacrals (see p. 47). APPEARANCE OF THE CERATOPSIA. Oat: [Im mounting this skeleton an attempt was made to embody all of the discoveries and changes resulting from Mr. Hatcher’s thorough study of the Ceratopsia, which he freely communicated in advance of publication. It must be understood, however, that there are still many undetermined points in the anatomy of this group, and the present writer, who mounted the skeleton with the able assistance of Mr. Norman Boss], alone must be held responsible for any anatomical inaccuracies that may be detected in the reconstruction. One of the most notable features in this skeleton is the great breadth of the pelvis and hind quarters. This impresses one most when the skeleton is viewed from the rear, and is a strong point of contrast between the Ceratopsia and other Dinosauria. (See fig. 124.) This is in the main due to the position of the expanded blades of the ilia, which are horizontal rather than vertical, as in other dinosaurs. The body is so broad that the creature is much better propor- tioned than one would be led to expect from viewing the skull alone. Measurements of the mounted skeleton. Meters. Juength; following: the curve of the vertebral centra (about). ...-....----.---=-.-----.---------+--------- Un PD Skull: eM GMO 3 vn He so Ges Da UOR eon see SOD ee oe Sey ene nS re ae 1.88 Wenzthyomicondyleitoniprotnestoved rostrallee = passer eee Se eee ees oo ese 1.10 Weustheiromsposteriommancinyotstnlli tomasallhormee esses asses soe 2 2 eens en eae ye 1. 46 Bea Ores oe en Se bh oe cise re Gee eR a ee Se ag ena IL Gill IBrenchila Cit CHOWSee: ee Le oeerSucic SUA Mere abe Sots GAS SAMS © hee eels ee reece aoe ae eee 42 Distance benweeninips OlSUpLAOn pba Orns er eee ee eee ee (ee eee ee eens eee . 83 Length of right supraorbital horn core from upper edge of the orbit-.........---.-.------------------- .70 Cinthia horlcoresyus ma poye un Corbites ae ee ne eae See DPS 2 be Sey e scien . 66 Vertebral column: 4 enctihvotmourntcoalescedicenvicalavertebre= seems se- ee 222 aes aee eo oe sees eee ae nese oe nee 415 ene beoteentinercebyl Caleselics may ssa eis aera yan eens He aetna eases See cheek e Soe ose 88 Meno hmo fad orso- limp ase ries tees anes ea eee eye Ee See cee tains ene esle esse ee sec sae 1.725 ILGRGiln OF ACMI 2. se S cae Pelee SSSA eS See SE Sere a ee oe eee ma RE cea at 1.10 Ibercin OF masons), GUGM SEES fee's seems Sie Eos es ae ete ee 2. 41 srencunvomunoraxses serene et neta ewe we Il ee eae ltt oH Reda dt load c! Bye i, 155 Wepunkoithormx (tojlowerendrofmibsiomliy,) essen yas a eee eyes ee = eke sees se es ees vaste pe 1. 525 Length of longest rib (measured along the curve) from tubercle to end ._..........--.--------------- 1.45 Meno thyotcapiinlamntoybuberculumpy sears eee eto eer ates SayStyat os alsyais es eee ee yee atti Sec . 24 BreacuoROlpsach lute = sme peer ys aan Tame aera ae ee Na ge see ee eh bee coe a . 64 BieAGIIN Oi GAH WANE. co sored Un SOOO SAO aS ears Os ee at a ee Pe a a 1. 24 Right fore limb: sWenethvotscapuln and coracoid)2-- +. 5.2200 2.2-en-4----322--4 22 24 So oro aR eee tee ee leo Ree ee 1.35 Length of coracoid only ....-.....------- So ibcn Slope bd doae tee bee BANS: Spat NEOs S Smee aes 38 BGAGKIN. Ol COCO! cn De se ietcc Bose oe ent oo SE ae es Acree a en He ene ie ee . 39 Brescunkotab) sderaiscapullaat Up pemen (sameeren say leases = one ee ee olen ys eee es ae . 26 Ginernieetiy, loncrarc lila corr Green nuill, n, A els ee ac eis OC RAR ES el ORIG ee eee 36 ILeraieailay Ot lnIHUEL 3 2 SHES wee Sot cae on gel bae 2 5 ae COU Eee ee ee ee ioe ete sae 5 fl Brcnakile or proramtll iC sae 5on8 om AaSe Bare e Seke ele eA aaa 1 EER tee ne eae ee . 40 Bramall: or Glial GG os pooa nc eanebe Soe Ree AOS on ee eee Cer he Bee eee aaa eames 36 (Chitin OF Ditilittas 2 asses cooe decd Os GOES OU COSTE OE Herat ee tee eA rs ee ee . 43 Temengiln nt Whe, Gana se da Os Cou bE Bes HenoSS on deat es OP SR Se ee ARE ee eee ieee ae 65 Teale OF P@RATE EMG Some to eos aaa ne ne Boe eS ee aa eRe sae eS See ere .38 Teiceyahiln or GlEfiGll Gite as cei ea see done o Da ee OUR te ise Se eee ere Aa ee ee 19 (rita Oh MOMS SS lew oan tee se eek oe OS ae OE eee Se er eee tee an ee ee ee . 36 Ibe saree: Gi HECKER. 2 26 Aone Geos eece taddoe POLS a Seco DANE BOG Hebe Cle Bene ee ae epee ese a esae nee 41 ica @ prersiwell GaGlo secascd Sasacequoods GocU RRS Ee CBnE ae See ee oes be Seer cee son ae ea mee .18 Raahin oh Gbistitall Cavl 2538 as otodon godt bcoenehe Sor ees ese ees = See! | Se aeee Sese ee cSeeasre ear 14 (Cimill OFM Se ono le atch tos tate 2c 6shous 6 acoe DBD BOCES Ons Sean eee nOO Aenea ESS ae eee ee 205 (Norr.—As the manus is entirely restored no measurements were taken.) Right hind limb and pelvis: Tenge of Miia oo So5dcorlae cess Bote de aoe ae Hop eee One Bn HOE DE GSee Bee Dao na Sc enor Osan aras 1. 50 Chieti, Gkyatilat (Ui) enit(e PLEUNIE)) ynn oo bene oe bos see oe ecie a CaS eee BBE Ep AeSp Space ao neeremeass = 32 Boeagily @fi ljitls oo scotesc.eeehesdd. cetc Decco ON eOEe CE Hee OE Seas e eee eee oe Seer iar ecrse 32 lugar: i [UM Bhe oe obecterare Hoppa dene Re RO ce See = COBO CU EE EE ee on OS ee ee cornea c Bar ah Aer eaaarcs 85 MaGnatiy gyi. jialiguianr Cul oo acee a keer ened ep eo Secor EG Sone HEE ECS e een emo ree terre .28 2 THE CERATOPSIA. | Right hind limb and pelvis—Continued. Meters. Menethvonascehimmy (measumedsomyoutbe ci cilia ye) meee se eee eae ase te eee 1.50 IBYRSAGUIN Gt joroRaIMMe! CMG. oe 5 saa 2 see ae oo se Sse sees obsess ee se sass ossesoecsss252052502252 . 40 Tero OE WHNUTES 2 oe elo Se ee se Se Se ee ee ee ee ce sees coon oo ec de sees eect ae so easssseesesossserete: 1.15 Breadth of proximal ends... 2.325020 05 222 eo) ee ee 42 Breadth of distal end......------ cade a HAUL Sestng Sic Chet EE Lie eS wea Sec Se SR ea Ae . 43 Girth of shaittes 22 oes 2S eS erp Se ets eS eee no ee eer oe en . 485 Thgaradn Of wllove, Guorel mstimeOMOG. 2 esse cee se se cee cece ee soe doce eeeeoasccaneses SE AE recht 72 Breadthvor-proximal emdlofitiliame yeep oye ee ere a ee . 895 Breadth ‘of distal emd*of tibia: 22222 2s 22 ae See epee Ree a Pea 39 Mibullal restored )\ =. 5/2s hSees R ete ae eae nS ne A ere Mien gbhieo fem e bat curse Teese age a eget ee 29 Teng Ghote nme tert ears ell SMITE sey see Pe ee . 355 Tenethvof digit, Wee. 65 A eel se os De R Grersta cee ceelene eee ate er nee ees . 325 Length of ungual phalanx dV ios 252.2 5 ces Soe eee te ac are ee Tee o Hal Breadth of ungwal phalanx PV 25 22 28 fos 5ee Lek ees ee ey ear ene eee .12 Measurements depending on posture: Height to: sumamat) of back o's sical S sles ore ae ee et set as erarre ye ees ayy ety ee 2.47 Breadth: ofishoulders atielenord fossee2 2 3°- sete 5 ee ee fee ae ee eee eee eee ae 125 Bréadth: of elbows... 22 402225 bi Nee tessa SORES: 2S oe Se SE ae ep eT ee 2.16 | Breadth across outersdigits: = )..222 aj... 0625 sos Seneca ee ce Cee Seen ee eee ere IL 7@ Breadth across heads of femora..\2-).2 4... 2.24 Sees Sees Se see eee oe ee ee ee 1.50 Bréadth: across *kneeéss:..2/2 = Seer ce ae oa ee oe ene ee ore ee ee eee eee 1.93 IBYRCEKGIUIN GXOWOSS) CUE CHAS. 52 Be ee So ee = see See oe esos oe 2 Ss sss Sse es ebeess:- 2.04 This skeleton has been admirably prepared and mounted by Mr. Gilmore. Two points, however, are open to question, as is clearly recognized by Mr. Gilmore—the position of the skull ome that of the ischia. a: In the skull here mounted, as well as in the skulls of the type of Diceratops hatcheri, of Triceratops calicornis, and of oil species of Triceratops, the condyle, which is supported on : neck, is strongly deflected downward in such a way as to cause the head to be carried low i front if the line of the condyle is continuous with that of the cervicals, which it evidently shoul be. The hemispherical articulating surface of the condyle is of such extent that the heac might have been raised to a somewhat higher position than that given it in the mount or, on the other hand, depressed until the beak reached the ground, but the average position woul be with the samazle about 20 inches (0.51 m.) lower than in the mounted position. In the ischia the articulating extremities are of such a character that their exact method of union — with the ilia and pubes is highly conjectural, and further evidence from other specimens may _ necessitate a radical departure from their position in the present mount. It seems probable that four is the correct number of digits in the manus rather than fiv as in the mounted specimen. The cranial armature at once shows a sharp contrast in development and in mode of 1 with that of the ancestral Monoclonius, for in Triceratops the frill was complete and heavy a undoubtedly flared upward and outward to a greater extent than in the former genus, atfordi not only leverage for muscular attachment, but, as a helmet-like structure, serving to protec the neck region from the horns of the adversary. In Triceratops, too, the temporal horns large and are strongly curved forward, with a corresponding reduction of the forward-direct nasalhorn. The restoration(Pl. I, frontispiece) expresses well the appearance of the lower: head, in which all three horns are brought to bear against the enemy at the moment of impa In Monoclonius the offensive stroke must have been an upward thrust in which the erect o backward-curved nasal horn would prove a most efficient weapon, while Triceratops woul charge with lowered head, seeking either to impale his enemy or to bear the latter down by the impetus of his great areteli ae The expansion of the frill and the development of great protecting ridges around the or! would be such as to best protect the most vital points, the neck and the eyes, from such a a of attack. PROBABLE HABITS. 193 Triceratops was extremely deficient mentally, and was probably of comparatively peace- able disposition except, perhaps, at the breeding season. Then the combats between rival males which probably took place must have been prompted and carried out by blind, unrea- soning instinct solely. This would make such weapons and defensive armor very efficient, for the Ceratopsia were evidently not intelligent enough to use weapons requiring skill in their manipulation. The question of other skin protection can not yet be settled, for while certain dermal scutes and ossicles have been found which may have been borne by Triceratops, we have no knowledge of their position or arrangement. It is unreasonable to suppose that the skin was naked, for such condition is found only among exclusively aquatic reptiles, such as the ichthyosaurs. The limbs were doubtless somewhat elephantine except that the ulna, with its huge olecranon process, gives evidence of having been flexed to a greater degree, as shown in the restoration. Of the feet we have but little trustworthy knowledge. The hoof-like claws are clearly indicated by the form of the ungual phalanges and, if Hatcher’s very reasonable conception of the creature’s habitat be correct, one would expect a somewhat spreading foot, which would bear the animal up in soft ground. : Of Triceratops only have we any idea of general form and proportions, its contemporaries Diceratops and Torosaurus being known only from the skull. Diceratops, it will be remembered, had no nasal horn, and the supraorbitals were erect instead of forward projecting as in Triceratops. Its mode of fighting must have differed some- what from that of the latter, probably in lowering the head much more. The peculiar fenestra in the squamosals are unequal in size, and the one in the right parietal, near the margin of the frill (the corresponding portion of the left having been destroyed), may have been due to wounds caused by the penetration of an adversary’s horn. In the Yale Museum there is a scapula of Diclonius with a clean-cut perforation, the edges of which have healed so as to give the appear- ance of anormal foramen. It is absent in the other scapula of the same animal, and Professor Marsh used to say that it was made by the horn of a Triceratops. This is certainly very suggestive of the Diceratops fenestre. The left squamosal of the type of Triceratops elatus also shows a perforation near the parietal suture, which is of pathologic character. (See Pl. XLIII, p. 284.) ; ‘In Torosaurus the cranium was of proportions so different, with its immense though weak frill and its wedge-shaped facial region, that the aspect of the head must have differed greatly from that of Triceratops. The upper surface of the frill does not bear the deep vascular impres- sions of the last-mentioned genus nor are there marginal ossicles, indicating that instead of being somewhat free with a dense horny or tegmentary covering, the crest was more or less buried in the flesh of the neck. It was evidently used to obtain leverage for the head and not like the neck guard of a helmet, to protect the cervical region. The presence of the large vacui- ties is further evidence in favor of this belief. The two known species of Torosaurus are huge creatures, larger than the average Triceratops though of less proportions than the giant specimen of the latter genus alluded to on page 185. The supraorbital horns of Torosaurus were well developed, though the nasal horn was proportionately reduced and acutely pointed. PROBABLE HABITS. / The feeding habits of the Ceratopsia are manifest from the tooth structure and from the character of the vegetation preserved with ceratopsian remains. The forward part of the mouth was edentulous and was sheathed on both the upper and the lower jaw, with a cutting beak like that of a turtle. Within the mouth were the magazines of teeth, each series presenting a vertical though slightly twisted wearing surface toward that of the opposing series, the worn surface of the lower teeth facing outward, that of the upper row inward. There is no possibility of a lateral grinding movement, as in herbivorous mammals; the lower jaw must have been moved entirely in a vertical plane. The beak probably served for cropping the more succulent leaves and shoots of low trees or shrubs, while the teeth were used to chop the food into short pieces before it was swallowed. As such pieces would naturally fall outside of the teeth of the MON XLIx—07——13 194 THE CERATOPSIA. i lower jaw, the food must have been retained within the mouth by the muscular walls of the cheeks. Unless the teeth also subserved the function of food getting as well as of mastication, which is questionable, it is doubtful whether the gape of the mouth had a greater backward extent than the anterior end of the tooth series. This would bring the corner of the mouth decidedly in advance of the position indicated by Knight in Pl. I. (Compare fig. 5.) PROBABLE ENVIRONMENT. T. W. Stanton in a note? says: It is difficult to reconstruct the physiographic conditions which prevailed in the Middle West. during later Mesozcic time, but it should be remembered that in that-region: there: was’then a great shallow continental or mediterranean sea, and that there were large areas so near sea level that very slight movements would bring them beneath the sea or partly or wholly drain. them, so that it is probable that shallow-water and nonmarine conditions were often extended over large areas Cory rapidly. It would seem as though some such elevation, occurring at the close of the Claggett, gave rise to conditions under which the fresh-water Judith River deposits could be formed and that the Judith River period was succeeded in turn by a subsidence which caused an encroachment of the sea upon the land, giving rise to the Bearpaw shales. Next a second diastrophic movement caused a recession of the salt waters and inaugurated the uml tions which characterized the Laramie. Hatcher ® thought that the period of elevation which brought about the close of the marine Cretaceous was followed, during the Laramie, by a period of subsidence not sufficient to cause a return to marine conditions, but such as to allow continual shallow-water deposition, as is evidenced by the great aaaner of lignite seams in the Ceratops beds and. by the absence of continuity of strata and the frequent cross- -bedding which prevailed. Hatcher says: The Ceratops beds are thought to afford evidence in themselves of having been deposited not in a great open lake, ue in a vast swamp, with occasional stretches of open waters, the whole presenting an appearance similar to that which now exists in the interior of the Everglades of Florida. This condition would account for the frequent changes from one material to, another in the same horizon. ... In some places in the beds these changes are quite frequent, strata of sandstones and shales replacing one another in great confusion. It would also explain the cross-bedding so often seen in the sandstones of this region, in localities remote from the present border of the beds, and hence far removed from the shore of the ancient lake or swamp. This cross-bedding could hardly occur in offshore deposits of a great fresh-water lake of any considerable depth. The conditions that prevailed over this region during the period in which the Ceratops beds were deposited were prob- ably those of a great swamp with numerous small, open bodies of water connected by a network of watercourses constantly changing their channels. The intervening spaces were but slightly elevated above the water level or at times submerged. The entire region where the waters were not too deep was covered by an abundant vegetation, and inhabited by the huge dino- saurs (Triceratops, Torosaurus, Claosaurus, etc.), as well as by the smaller crocodiles and turtles and the diminutive mammals, all of whose remains are now found embedded in the deposits. S ~ The frontispiece admirably depicts such a scene as Hatcher has described. Dr: W. D. Matthew,° in a recent paper, speaks of three modes of life available during the Mesozoic for land vertebrates, ‘‘the amphibious-aquatic, the arboreal, and the aerial, the terres- trial being subordinate because the upland flora was largely undeveloped or inedible as compared with its present condition.’’ The three provinces Matthew believed were peopled by reptiles, mammals, and birds respectively. With reference to the dinosaurs in particular Matthew’s — views are equresellt in a letter to the writer, dated June 6, 1905, as follows: I believe that they (the dinosaurs) were a—in fact the—land group of reptiles, but that nearly all we know of them is a number of aberrant amphibious or aquatic specialized branches; that the great arid subglacial period of the Perm-Trias gave ~ them their initial trend on lines parallel to the evolution of the Mammalia during the Tertiary-Quaternary; that in the late Jura and the late Cretaceous a reaction to moist, torrid climate caused a great expansion and specialization of amphibious swamp-living forms, adapted to the marshy jungles then prevalent. These are the dinosaurs we know. Of the dry-land forms we know very little. A few Triassic types, possibly some of the Jurassic ones, like Ornitholestes and Hallopus and Laosaurus, and the jungle-living carnivorous types departed less than the others from the primitive dry-land type. The Sauropoda I regard as exclusively water-living—the larger forms at least.@ 5 5 ‘ a Proc. Am. Philos. Soe., vol. 43, p. 364, 1904. b Am. Jour. Sci., 3d ser., vol. 45, 1893, p. 142. c Am. Naturalist, vol. 38, Nov.—Dec., 1904, p. 816. d Herein Matthew and Hatcher disagreed, as the latter considered the Sauropoda also as “ terrestrial reptiles with amphibious habits; passing much, perhaps most, of their time in shallow water, where they were able to wade about in search of food.’’ The evidence is strongly in favor of Matthew’s belief.—R. S. L. Us CAUSES OF EXTINCTION. 195 The stegosaurs, claosaurs, and ceratopsians may have been more or less land haunting, but not upland, and they all impress me as amphibious adaptations from a type highly specialized for land locomotion. In the later Cretaceous the terrestrial province was greatly expanded by the development of the upland flora which provided for a corresponding spread of terrestrial types. These were derived mainly from the previously arboreal mammals, the birds maintaining their aerial habitat, while of the reptiles, the lizards and snakes only were able to adapt themselves to these new conditions. At the time of the expansion of the upland realm there was great dwindling of the amphibious-aquatic province, due to the orogenic movements occurring at the close of the Mesozoic, which drained the Cretaceous sea and its adjacent swamps-and river deltas and caused the Reptilia to undergo a corresponding diminution. PROBABLE CAUSES OF EXTINCTION. Several theories have been advanced as to the probable causes of extinction of the Ceratopsia, some authors maintaining that the horned herbivorous types were in part destroyed by the large carnivorous dinosaurs. There is always, however, a balance in nature, an offsetting of Carnivora or parasitic forms against their plant-feeding contemporaries and, though the latter may have been held in check ‘by the former, it is extremely improbable that strictly contem- poraneous forms which have evolved in the same environment could ever exterminate one another. It seems that animals of another race, or hordes of creatures which emigrated from another region, would be more likely to exterminate their predecessors. The mammals fulfill the requirements of a new foe, and the development of the frill in the Ceratopsia has been consid- ered as meeting the necessity for a better protection of the neck blood vessels from the weasel-like attack of small but bloodthirsty quadrupeds. Another notion advanced by Morris and ampli- fied by Cope was that the Cretaceous mammals sought out the eggs of the dinosaurs and destroyed them—Cope even going so far as to suggest the Multituberculata, with their long, sharp anterior teeth, as the probable offenders. Matthew, however, has given the Mesozoic Mammalia a totally different habitat from that of their dinosaurian contemporaries in the belief recently expressed that the mammals were distinctively arboreal, while we are led to believe that all dinosaurs were either terrestrial or possibly amphibious, the Ceratopsia at least inhabiting the lowlands in a swamp or delta region. 2 By far the most reasonable cause, and the one which Hatcher himself believed, seems to be that of changing climatic conditions and a contracting and draining of the swamp and delta regions caused by the orographic upheavals which occurred toward the close of the Cretaceous. - The Ceratopside and their nearest allies, the Trachodontide, both highly specialized plant feeders, were unable to adapt themselves to a profoundly changed environment because of this very specialization, and, as a consequence, perished. : That the Ceratopsia made a gallant struggle for survival seems evident, for they lived through the first series of upheavals at the close of the Laramie and also the second series at the close of the Arapahoe, which were accompanied by great volcanic outbursts in the Colorado » region; but the changes accompanying the final upheavals which formed most of the great western mountain chains and closed the Mesozoic era gave the death blow to this remarkable race. iz 3 BIBLIOGRAPHY. American Geologist. [Editorial comment on Dr. G. Baur’s “Remarks on the reptiles generally called Dinosauria.”] Am. Geologist, vol. 8, 1891, pp. 55-56. 3 Ballou, W. H. Strange creatures of the past. Century Magazine, vol. 55, 1897, pp. 18, 20-21. Baur, G. Professor Marsh on Hallopus and other dinosaurs. Am. Naturalist, vol. 24, 1890, p. 570. Baur, G. The horned saurians of the Laramie formation. Science, new ser., vol. 17, 1891, pp. 216-217. Baur, G. Remarks on reptiles generally called Dinosauria. Am. Naturalist, vol. 25, 1891, pp. 443, 447-448, 452. ii Beasley, W. L. A remarkable fossil discovery. Sci. Am., vol. 89, 1903, p. 87. i Bunzel, E. Reptilien der Gosauformation. Abh. Geol. Reichsanstalt, Vienna, 1871, bd. 5, ae =i), Cope, E. D. On the existence of Dinosauria in the transition beds of Wyoming. Proc. Am. Phil. Soc., vol. 12, 1872, pp. 481-483. - Cope, E. D. Remarks on geology of Wyoming. Proc. Acad. Nat. Sci. Phila., vol. 24, 1872, p. 279. i Cope, E. D. Report on the stratigraphy and Pliocene vertebrate Eaeeeicne of northern Colorado. Bull. U. S. Geol. i} and Geog. Surv. Terr., vol. 1, no. 1, 1874, p. 10. i’ Cope, E. D. The eee age of the coal of Wyoming. Am. Naturalist, vol. 6, 1872, p. 670. I = Cope, E. D. Report on the vertebrate paleontology of Colorado. Ann. Rept. U. S. Geol. and Geog. Sury. Terr. for 1873, AW 1874, pp. 429-533. Cope, E. D. Review of the Vertebrata of the Cretaceous period found west of the Mississippi River. Bull. U. S. Geol. i and Geog. Surv. Terr., vol. 1, 1874, no. 2, pp. 9-21. Cope, E. D. The Vertebrata of the Cretaceous formations of the West. Rept. U. S. Geol. Surv. Terr., vol. 2, 1875, pp. 26, 31-41, 53-56, 64-65, 248. ‘ Wy - Cope, E. D. Descriptions of some new vertebrate remains from the ont Union beds of Montana. Proc. Acad. Nat. Sci. - Phila., vol. 28, 1876, pp. 248-261. Cope, E. D. Report on the geology of the Judith River, Montana, and on the vertebrate fossils obtained on or near the Missouri River. Bull. U. S. Geol. and Geog. Surv. Terr., vol. 3, 1877, pp. 565-579. Cope, E. D. [Review of Prof. L. Lesquereux’s contributions to the fossil flora of the Western Territories. Pt. 2, The Ter- tiary Flora.] Am. Naturalist, vol. 12, 1878, pp. 243-246. i] Cope, E. D. The geological record. Am. Naturalist, vol. 14, 1880, p. 511. - Hil Cope, E. D. On the characters of the skull in the Hadrosauride. Proc. Acad. Nat. Sci. Phila., vol. 35, 1883, pp. 99-100. } i} Cope, E. D. The sternum of the Dinosauria. Am. Naturalist, vol. 20, 1886, pp. 153-155. | | Cope, E. D. A horned dinosaurian reptile. Am. Naturalist, vol. 22, 1888, pp. 1108-1109. p | Cope, E. D. The age of the Denver formation. Science, vol. 13, 1889, p. 290. i Cope, E. D. The horned Dinosauria of the Laramie. Am. Naturalist, vol. 23, 1889, pp. 715-717. Hi Cope, E. D. Notes on the Dinosauria of the Laramie. Am. Naturalist, vol. 23, 1889, pp. 904-906. I z Cope, E. D. [Note on the teeth mentioned by Professor Marsh.] Am. Naturalist, vol. 24, 1890, p. 571. i Cope, E. D. Syllabus of lectures on geology and paleontology. Univ. of Penn., 1891, p. 43. i ; Cope, E. D. Fourth note on the Dinosauria of the Laramie. Am. Naturalist, vol. 26, 1892, pp. 756-758. i | Cope, E. D. Syllabus of lectures on the Vertebrata. Philadelphia, 1898, p. 70. HI Cross, Whitman. The Denver Tertiary formation. Proc. Colorado Sci. Soc., vol. 3, 1888, pp. 119-133. See also Emmons, Cross, and Eldridge. | | i} Dana, J.D. Manual of Geology, 4th ed., New York, 1894, pp. 846-847. oh Eldridge, G. H. On some stratigraphical and structural features of the country about Deurers Colorado. Proc. Colorado Hi Sci. Soc., vol. 3, 1888, pp. 86-118. See also Emmons, Cross, and Eldridge. ! ih Emmons, S. F., Cross, W., and Eldridge, G. H. Geology of the Denver Basin. Mon. U. 8. Geol. Survey, vol. 27, i 1896. AHI Fiirbringer, Max. Zur vergleichenden Anatomie des Brustschulterapparates und der Schultermuskeln; iv Teil. Jena’ ische Zeitschr. Naturwiss., Jena, vol. 34, 1900, p. 351. 196 Se es BIBLIOGRAPHY. 197 Gilmore, C. W. The mounted skeleton of Triceratops prorsus. Proc. U.S. Nat. Mus., vol. 29, 1905, pp. 483-435. Gilmore, C. W. Notes on some recent additions to the exhibition series of vertebrate fossils. Proc. U.S. Nat. Mus., vol. 30, 1906, pp. 608-610. Hatcher, J. B. The Ceratops beds of Converse County, Wyoming. Am. Jour. Sci., 3d ser., vol. 45, 1893, pp. 135-144. . Hatcher, J. B. Some localities for Laramie mammals and horned dinosaurs. Am. Naturalist, vol. 30, 1896, pp. 112-120. Hatcher, J. B. The genera and species of the Trachodontide (Hadrosauride, Claosauride) Marsh. Annals Carnegie Mus., vol. 1, 1902, pp. 377-386. Hatcher, J. B. Two new Ceratopsia from the Laramie of Converse County, Wyo. Am. Jour. Sci., 4th ser., vol. 20, 1905, pp. 413-419. Hatcher, J. B., T. W. Stanton and. See Stanton and Hatcher. Hay, O. P. On some recent literature bearing on the Laramie formation. Am. Geologist, vol.-32, 1903, pp. 115-120. Hutchinson, H. N. Extinct Monsters. New York, 1893, pp. 105-108. Lambe, L. M. On reptilian remains from the Cretaceous of northwestern Canada. Ottawa Naturalist, vol. 13, 1899, pp. 68-70. : Lambe, L. M. Sum. Rept. Geol. Surv. Dept. Canada for the year 1898, 1899, pp. 186-187. ' Lambe, L. M., Osborn, H. F., and. On Vertebrata of the Mid-Cretaceous of the Northwest Territory. 2. New genera and species from the Belly River series. Geol. Surv. Canada, Contr. Can. Paleontology, vol.3 (quarto), part 2, 1902, pp. 57-69. Lambe, L.M. On the squamoso-parietal crest of two species of horned dinosaurs from the Cretaceous of Alberta. Ottawa Naturalist, vol. 18, 1904, pp. 81-84. Lambe, L. M. On the squamoso-parietal crest of the horned dinosaurs Centrosaurus apertus and Monoclonius canadensis from the Cretaceous of Alberta. Trans. Roy. Soc. Canada (2), vol. 10, sec. 4, 1904, pp. 3-10. Lambe, L. M. Progress of vertebrate paleontology in Canada. Trans. Roy. Soc. Canada (2), vol. 10, sec. 4, 1904, pp. 23-24. Lee, J. E. Notice of saurian dermal plates from the Wealden of the Isle of Wight. Ann. and Mag. Nat. Hist., vol. 11, 1843, p. 5. Leidy, J. Notices of remains of extinct réptiles and fishes discovered by Dr. F. V. Hayden in the badlands of the Judith River, Nebraska Territory. Proc. Acad. Nat. Sci. Phila., vol. 8, 1856, pp. 72-73. Leidy, J. Extinct Vertebrata from the Judith River and Great Lignite formations of Nebraska. Trans. Am. Phil. Soc., 2d ser., vol. 11, 1859, pp. 139-154. Lesquereux, L. Age of the North American Lignite. Ann. Rept. U. S. Geol. and Geog. Sury. Terr. for 1873, 1873, pp. 367-378. 2 _ Lucas, F. A. Animals of the Past. New York, 1901, pp. 100-104, 121-126. Lucas, F. A. The dinosaurs, or terrible lizards. Ann. Rept. Smithsonian Inst. for 1901, 1902, pp. 641-646. Lucas, F. A. Constructing an extinct monster from fossil remains. [Z7riceratops.] Sci. Am., vol. 86, 1902, p. 43. Lull, R. S. Skull of Triceratops serratus. Bull. Am. Mus. Nat. Hist., vol. 19, 1903, art. xxx, pp. 685-695. Lull, R. S. [Editorial notes on Two new Ceratopsia from the Laramie of Converse County, Wyo.] Am. Jour. Sci., 4th ser., vol. 20, 1905, pp. 413, 418-419. Lull, R. S. Restoration of Diceratops. Am. Jour. Sci., 4th ser., vol. 20, 1905, pp. 420-422. Lull, R. S. A new name for the dinosaurian genus Ceratops. Am. Jour. Sci., 4th ser., vol. 21, 1906, p. 144. Lydekker, R. On a peculiar horn-like dinosaurian bone from the Wealden. Quart. Jour. Geol. Soc. London, 1890, pp. 185-186. Lydekker, R. Some recent restorations of dinosaurs. Nature, vol. 48, 1893, p. 304. Lydekker, R., H. A. Nicholson and. See Nicholson and Lydekker. Marsh, O. C. Notice of some new fossil mammals. Am. Jour. Sci., 3d ser., vol. 34, 1887, pp. 323-324. Marsh, 0. C. A new family of horned Dinosauria from the Cretaceous. Am. Jour. Sci., 3d ser., vol. 36, 1888, pp. 477-478. Marsh, O. C. Notice of new American dinosaurs. Am. Jour. Sci., 3d ser., vol. 37, 1889, pp. 334-336. Marsh, 0. C. Notice of gigantic horned Dinosauria from the Cretaceous. Am. Jour. Sci., 3d ser., vol. 38, 1889, pp. 173-175. Marsh, O. C. Skull of the gigantic Ceratopside. Am. Jour. Sci., 3d ser., vol. 38, 1889, pp. 501-506. Marsh, O. C. Description of new dinosaurian reptiles. Am. Jour. Sci., 3d ser., vol. 39, 1890, pp. 81-83. Marsh, O. C. Additional characters of the Ceratopside, with notice of new Cretaceous dinosaurs. Am. Jour. Sci., 3d ser., vol. 39, 1890, pp. 418-426. Marsh, O. C. The gigantic Ceratopsidx, or horned dinosaurs, of North America. Am. Jour. Sci., 3d ser., vol. 41, 1891, pp- 167-178. Marsh, O. C. Restoration of Triceratops [and Brontosaurus]. Am. Jour. Sci., 3d ser., vol. 41, 1891, pp. 339-341. 198 . THE CERATOPSIA. Marsh, O. C. Restoration of Stegosaurus. Am. Jour. Sci., 3d ser., vol. 42, 1891, p. 181. Marsh, O. C. Notice of new vertebrate fossils. Am. Jour. Sci., 3d ser., vol. 42, 1891, pp. 265-267. Marsh, O. C. On the gigantic Cératopside (or horned lizards) of North America. Rept. Brit. Assoc. Adv. Sci., 60th meet- ing, Leeds, 1890, 1891, pp. 793-795. % Marsh, O. C. The skull of Torosaurus. Am. Jour. Sci., 3d ser., vol. 43, 1892, pp. 81-84. ; = Marsh, O. C. Restorations of Claosaurus and Ceratosaurus. Am. Jour. Sci., 3d ser., vol. 44, 1892, pp. 343-346. i ; Marsh, O. C. Some recent restorations of dinosaurs. Nature, vol. 48, 1893, pp. 437-438. , Marsh, O. C. Restoration of Camptosaurus. Am. Jour. Sci., 3d ser., vol. 47, 1894, p. 245. | Marsh, O. C. The typical Ornithopoda of the American Jurassic. Am. Jour. Sci., 3d ser., vol. 48, 1894, p. 90. | - Marsh, O. C. Restoration of some European dinosaurs, with suggestions as to their place among the Reptilia. Am. Jour. | Sci., 3d ser., vol. 50, 1895, pp. 407-412. ; | Marsh, O. C. On the affinities and classification of the dinosaurian reptiles. Am. Jour. Sci., 3d ser., vol. 50, 1895, pp. 483-498. Marsh, O. C. Dinosaurs of N orth America. Sixteenth Ann. Rept. U. S. Geol. Survey, pt. 1, 1896, pp. 237-244. zi Marsh, O. C. Vertebrate fossils of the Denver Basin. Mon. U.S. Geol. Survey, vol. 27, 1897, pp. 509-516, 527. Marsh, O. C. New species of Ceratopsia. Am. Jour. Sci., 4th ser., vol. 6, 1898, p. 92. Matthew, W. D. The arboreal ancestry of the Mammalia. Am. Naturalist, vol. 38, 1904, p. 816. Nicholson, H. A., and Lydekker, R. Manual of Paleontology, 3d ed. Edinburgh, HS vol. 2, p. 1163. i Nopesa, F. aes Foldtani Kozlony, Budapest, 1901, vol. 31, p. 270. i Nopesa, F. Baron, Ueber Stegoceras und Stereocephalus. Centralblatt fir Mineralogie, 1903, p. 266. j Osborn, H. F. Fossil mammals of the upper Cretaceous beds. Bull. Am. Mus. Nat. Hist., vol. 5, 1893, p. 326. Osborn, H. F. Models of extinct vertebrates. Science, new ser., vol. 7, 1898, pp. 842-844. 3 Osborn, H. F. Casts, models, photographs, and restorations of fossil vertebrates. Am. Mus. Nat. Hist., 1898. ) Osborn, H. F., and Lambe, L. M. On the Vertebrata of the Mid-Cretaceous of the Northwest Territory. 1. Distinctive characters of the Mid-Cretaceous fauna. Geol. Survey Canada, Contr. Canadian Paleontology, vol. 3 (quarto ), 1902, pt. 2, pp. 9, 14-15, 19-21. : Seeley, H. G. The reptile fauna of the Gosau formation. Quart. Jour. Geol. Soc. London, vol. 37, 1881, pp. 620, 637-667. | Selwin, A. R. C., and Dawson, G. M. Descriptive sketch of the physical geography and geology of the Dominion of | Canada. Montreal, 1884, p. 40. Stanton, T. W., and Knowlton, F. H. Stratigraphy and paleontology of the Laramie and related formations in Wyo- \ ming. Bull. Geol. Soc. America, vol. 7, 1898, pp. 143, 155. i Stanton, T. W. The stratigraphic position of the Judith River beds. Science, new ser., vol. 16, 1902, pp. 1031-1032. } Stanton, T. W. [Note in J. B. Hatcher’s ‘‘An attempt to correlate the marine with the nonmarine formations of the Mid- if dle West.”] Proc. Am. Philos. Soc., vol. 43, 1904, p. 364. i) Stanton, T. W., and Hatcher, J. B. Geology and paleontology of the Judith River beds. Bull. U. S. Geol. Survey No. ii 257, 1905. i Steinman, G., und Déderlein, L. Elemente der Paliontologie. Leipzig, 1890, p. 665. ! i ; Walcott, C. D. Correspondence relating to collection of vertebrate fossils made by the late Prof. O. C. Marsh. Science, new ser., vol. 11, 1900, p. 23. White, C. A. Section of Laramie strata at Black Buttes station. Ann. Rept. U. 8. Geol. Surv. Terr. for 1877, pp. 222-223. White, C. A. (Bibliographical sketch of Dr. F. V. Hayden.) National Academy of Sciences, piece Memoirs, vol. 3, pp. 399-400. Williston, S. W. The Laramie Cretaceous of Wyoming. Science, new ser., vol. 16, 1902, p. 952. 4 Woodward, A. S. Outlines of Vertebrate Paleontology. Cambridge, 1898, pp. 213-216. : Zittel, K. A. von. Handbuch der Palaeontologie. Munich and Leipzig, 1887-1890, vol. 3, pp. 749-754. Zittel, K. A. von. Text-book of Paleontology. English translation by C. R. Eastman. London, vol. 2, 1902, pp. 243-245. i { J ee NT gO TOA a Fie. Fig. Fic. Fic. Fig. Fig. Fic. NO OP WN IP A Wa, IC IE COMPARATIVE VIEWS OF CRESTS. (See p. 19.) Plate prepared under the direction of Lull. . Type of Monoclonius crassus Cope. No. 3998, American Museum of Natural History. . Type of Centrosaurus apertus Lambe. No. 971, Canadian Geological Survey. After Lambe. . Type of Triceratops flabellatus Marsh. No. 1821, Yale Museum. After Marsh. . Type of Triceratops serratus Marsh. No. 1823, Yale Museum. After Marsh. . Type of Diceratops hatcheri Lull. No. 2412, U.S. National Museum. . Type of Ceratops (Monoclonius) belli Lambe. No. 491, Canadian Geological Survey. After Lambe. . Type of Torosaurus gladius Marsh. No. 1831, Yale Museum. ep, epoccipitals. sq, squamosal. pa, parietal. sqf, squamosal fenestra. paf, parietal fontanelle. sqs, squamosal suture. ptf, postfrontal fontanelle. All one-sixteenth natural size. 202 MONOGRAPH XLIX PL. II U. S. GEOLOGICAL SURVEY COMPARATIVE VIEWS OF CRESTS. eee 4 ELA Uo Feige s, JIL COMPARATIVE VIEWS OF SQUAMOSAL BONES. (See p. 20.) Plate prepared under the direction of Lull. Fic. 1. Ceratops montanus? Marsh. No. 4802, U.S. National Museum. After Marsh. a, Dorsal aspect; b, end view; c, ventral aspect. Fic. 2. Ceratops sp. (young). No. 2415, U.S. National Museum. After Marsh. a, Dorsal aspect; b, end view; c, ventral aspect. Fic. 3. Torosaurus gladius Marsh. No. 1831, Yale Museum. After Marsh. Ventral aspect. Vic. 4. Triceratops flabellatus Marsh. No. 1821, Yale Museum. After Marsh. Ventral aspect. Fic. 5. Monoclonius sp. No. 3394, American Museum of Natural History. a, Dorsal aspect; b, ventral aspect. Fie. 6. Monoclonius sp. No. 3996, American Museum of Natural History. a, Ventral aspect; b, dorsal aspect. Fic. 7. Monoclonius sp. No. 3995, American Museum of Natural History. a, Dorsal aspect; b, ventral aspect. Fig. 8. Ceratops canadensis Lambe. No. 1254, Canadian Geological Survey. After Lambe. a, Ventral aspect; 6, dorsal aspect. All one-twelfth natural size. 204 MONOGRAPH XLIX PL. Ill U. S. GEOLOGICAL SURVEY COMPARATIVE VIEWS OF SQUAMOSAL BONES. 200 « Jie SNe IV Fie. Fie. PAPA Nee COMPARATIVE VIEWS OF NASAL HORN CORES. (See p. 32.) PJate prepared under the direction of Lull. 1. Type of Monoclonius dawson Lambe. No. 1173, Canadian Geological Survey. Modified from Lambe. 2. Type of Monoclonius sphenocerus Cope. No. 3989, American Museum of Natural History. 3. Type of Centrosaurus apertus Lambe. No. 971, Canadian Geological Survey. After Lambe. 4. Type of Ceratops (Monoclonius) recurvicornis Cope. No. 3999, American Museum of Natural History. After Cope. . 5. Ceratops sp. referred with a query to Monoclonius dawson Lambe, by Lambe. No. 190, Canadian Geological Survey. Modified from Lambe. . Type of Triceratops horridus Marsh. No. 1820, Yale Museum. . Triceratops prorsus? Marsh. No. 2100, U. S. National Museum. . Type of Triceratops brevicornus Hatcher. No. 1834, Yale Museum. 6 .7. Type of Triceratops prorsus Marsh. No 1822, Yale Museum. After Marsh. 8 9 . 10. Type of Triceratops calicornis Marsh. No. 4928, U.S. National Museum. 1. 11. Type of Triceratops obtusus Marsh. No. 4720, U.S. National Museum. . 12. Type of Diceratops hatcheri Lull. No. 2412, U.S. National Museum. The figures are all viewed from the left side and are all one-eighth natural size. 206 U. S. GEOLOGICAL SURVEY COMPARATIVE VIEWS OF NASAL HORN CORES. MONOGRAPH XLIX PL. IV pene WIND ye 207 ; { j ' / ier Fie. Kia. Kia. Fic. Fie. Fie. Fia. Fic. Fie. 12 Ib a Id OW COMPARATIVE VIEWS OF SUPRAORBITAL HORN CORES. (See p. 32.) Plate prepared under the direction of Lull. . Monoclonius crassus. No. 3997, American Museum of Natural History. . Type of Ceratops montanus Marsh. No. 2411, U.S. National Museum. oblique back view. . Type of Ceratops (Monoclonius) canadensis Lambe. No. 1254, Canadian Geological Survey. Modi- fie] from Lambe. . Type of Ceratops (Monoclonius) recurvicornis Cope. No. 3999, American Museum of Natural History. Modified from Cope. 5. Type of Triceratops prorsus Marsh. No. 1822, Yale Museum. After Marsh. 6. 7 8 Triceratops prorsus? Marsh. No. 2100, U.S. National Museum. . Type of Triceratops brevicornus Hatcher. No. 1834, Yale Museum. . Type of Triceratops calicornis Marsh. No. 4928, U. S. National Museum. ®, 10. Type of Diceratops hatcheri Lull. No. 2412, U.S. National Museum. Type of Triceratops (Sterrholophus) flabellatus Marsh. No. 1821, Yale Museum. After Marsh. The figures, except fig. 2, are all viewed from the left side and are all one-eighth natural size. 208 MONOGRAPH XLIX PL. V U. S. GEOLOGICAL SURVEY COMPARATIVE VIEWS OF SUPRAORBITAL HORN CORES. 209 12 bp Ta WY IL | LOWER JAW OF TRICERATOPS SULCATUS MARSH. (See p. 39.) | Plate prepared under the direction of Marsh. Specimen No. 4276, U.S. National Museum. Fre. 1. Superior view. Fic. 2. Internal view. Fic. 3. External view. ang, Angular; art, articular; ep, coronoid process; d, dentary; s, splenial; sang, surangular. One-sixth natural size. 210 a U.S. GEOLOGICAL SURVEY MONOGRAPH XLIX PL. VI se = “4. ys " PODAATP DANS LOWER JAW OF TRICERATOPS SULCATUS MARSH. PLATE VIL. a) 211 i Ss ey Kt | Hi I | i F it 4 = i e | Nall | iI | a (ans | | : wal | i, Hi | | ) l il y s a (oe ’ a a 4 * a SS ae | i | “si iH ce i ely Vlei 4 | POSTERIOR DORSAL VERTEBRA OF TRICERATOPS PRORSUS MARSH. iF q (See p. 51.) ; ih Plate prepared under the direction of Marsh. if | Specimen No. 4842, U.S. National Museum. ‘a i Fic. 1. Anterior view. 3 || Fic. 2. Left lateral view. i Fic. 3. Superior view. | Fig. 4. Posterior view. a. | Fic. 5. Inferior view. b ql) One-fourth natural size. "Ne i 212 1 | i} Hi : Vi) Hil | | 2 i} | 4 a = i aaa Hi i) il . | : U.S. F. BERGER, del GEREOLOGICAL SURVEY MONOGRAPH XLIX PLATE VII Se MONOGRAPH XLIX PL. XL \ \ \ ey = ( Ir aas f Haya ESS ist n 2 ER CH } i A Za : i SSS7 z=. Mt s rr A) x WZ = fil J Ms 9 eR = 1 AY 2Z2~>N AM y yy, \ Aa, Mt FRI VAs UB S.P S HATCHER AND T. CALICORNIS MARSH. U. S, GEOLOGICAL SURVEY MONOGRAPH XLIx PL. XL ni ; i ORNUS HATCHER AND T. CALICORNIS MARSH. VERTEBRE OF TRICERATOPS BREVIC Pb ee 2b IP AWS, XC IL I SKULL OF TYPE OF TRICERATOPS BREVICORNUS HATCHER. (See p. 141.) Plate prepared under the direction of Hatcher. Type specimen, No. 1834, Yale University Museum. Left lateral view. ang, angular. no, nasal opening. art, articular. o, orbit. cp, coronoid process. pa, parietal. d, dentary. ' pd, predentary. ep, epoccipital. pmex, premaxillary. ju, jugal. qu, quadrate. Uf, lachrymal foramen. r, rostral. mex, maxillary. sang, surangular. nas, nasal. Sq, squamosal. nh, nasal horn core. - soh, supraorbital horn core. — One-eighth natural size. 280 "M3IA ACIS ‘YSHOLVH SMNYOOIASYE SdOLVYSOINL 4AO T1NMHS 1X “Id XI1X HdVYSONOW AZAUYNS 1V9ID01039 *S “N ie i PIL ASIN 2 GIO. IP yA WB, OK IGAL IL SKULL OF TYPE OF TRICERATOPS BREVICORNUS HATCHER. (See p. 141.) Plate prepared under the direction of Hatcher. Type specimen, No. 1834, Yale University Museum. Palatal view. Hj | bo, basioccipital. pa, parietal. i dc, dental channel. pal, palatine. Hit ep, epoccipital. pmz, premaxillary. . exo, exoccipital. pt, pterygoid. qu, jugal. qu, quadrate. mx, maxillary. r, rostral. oc, occipital condyle. sq, squamosal. One-eighth natural size. 282 MONOGRAPH XLIX PL. XLII U. S. GEOLOGICAL SURVEY My , y yf yy Miia aH ARS t Ti iy \\ ohh Mp oy y \\ ‘\ if j j q Me ANY A : - “i A il a | Wil AY ia H| ‘| i \ * i a I \ Qa WAKA ! oe it Wh te We i, \ iH Hala i iW nA Mt i aly \ iI!')/ y i | yf } y} YY SY KI Yes SA ip y) Yj Mf ff < h Yj g YY ef 4 i NG 4 ' \ Pe | Qe ‘i ey A i w= Ww a = KS YY SA FS Ws WF \ SS WZ % Lig S \ RY SG A PALATAL VIEW. SKULL OF TRICERATOPS BREVICORNUS HATCHER, ees PLAIN. 2SEIEOE IP ba Jd; SCI TI, SKULL OF TYPE OF TRICERATOPS ELATUS MARSH. (See p. 135.) Plate prepared under the direction of Hatcher. Type specimen, No. 1201, U. S. National Museum. Left lateral view. ju, jugal. pf, postfrontal. lac, lachrymal. pm, premaxillary. Uf, lachrymal foramen. pif, pretrontal. mz, maxillary. qj, quadratojugal. nas, nasal. qu, quadrate. nh, nasal horn core. r, rostral. no, nasal opening. sq, Squamosal. 0, orbit. soh, supraorbital horn core. pa, parietal. One-eighth natural size. 284 "‘MAIA AGIS “HSYVIN SNLVIE SdOLVYsOIYL JO T1INMS WAX “Id XITX HdVYSONOW A3BAYNS WW9ID010359 *S “N PLAT, SIDI PMA SLI. : SKULL OF TYPE OF TRICERATOPS FLABELLATUS MARSH. il (See p. 143.) il Plate prepared under the direction of Marsh. Type specimen, No. 1821, Yale University Museum. Left lateral view. il ang, angular. 0, orbit. ill cp, coronoid process. pa, parietal. i d, dentary. pd, predentary. i) . ep, epoccipital. pf, postfrontal. A | ju, jugal. pmex, premaxillary. H lac, lachrymal. prf, prefrontal. Mi | Uf, lachrymai foramen. qj, quadratojugal. ii mex, maxillary. qu, quadrate. | nas, nasal. sang, surangular. mh, nasal horn core. sq, Squamosal. | no, nasal opening. soh, supraorbital horn core. ill One-fourth natural size. ill 286 act > ee Sees : Renee) i 2 a i ; : oS | Soa : oes - Sones x / ! Bs ‘ 8 = Ww > c Pm | n al 4 o o o a ° WwW o 7) 5 SIDE VIEW. LU, SKULL OF DICERATOPS HATCHERI IPLAC, 2G WIG IP ye Jd; CI WIE SKULL OF TYPE OF DICERATOPS HATCHERI LULL. + (See p. 149.) Plate prepared under the direction of Hatcher. Type specimen, No. 2412, U. S. National Museum. Superior view. ep, epoccipital. paf, parietal fenestra? f, frontal. pf, postfrontal. Uf, lachrymal foramen. pff, postfrontal fontanelle.. na, nasal. r, rostral. no, nasal opening. sf, squamosal fenestra? 0, orbit. sq, Squamosal. pa, parietal. soh, supraorbital horn core. One-eighth natural size. 294 ‘MAalIA dOL ‘11N1 IYSHOLVWH SdOLVYSOIG 40 11NxHS LE SS SS 8 QR ww = INATX “Id XIX HdVYSONOW A3SAYNS 1V9I901039 Ss “Nn U. S. GEOLOGICAL SURVEY MONOGRAPH XLIX PL. XLIX RESTORATION OF TRICERATOPS. Views of mounted skeleton in United States National Museum. A, Side view; B, oblique rear view, U. S. GEOLOGICAL SURVEY MONOGRAPH XLIX PL. L ALE? Ass eRe EK, MAP SHOWING PRINCIPAL CERATOPSIA LOCALITIES. Compiled by R. S. Lull. @YoeH *g “ff JeyYy "OAM ‘ALNNOO SASYSANOO JO dv Oe RE ti - i a ' yu uostarapyVry ~Y DS Pa7 D\ ae" ¢ umes YOPIOGT SAOZOI2ZD —m— me SIULe IVT IAC Y SP2q //SSOf 07 PCOY .«-.-----------_. --- 2 PUNCL U2PAY PACY 2EP/INCSOIPE}Y a a S82E2/Pu/ O a> en PUTOy uVI2ag ea1ey S/EUW/WweU “ “ S22ED/Puy/ oo a PYUNOL UI2G PACY SY/NYS $IOZEIZII4 PISYM SAIZI/CEIO{ SIZED/PU] + « = a a l a | AY AM | S ow aN x ou | We ~ i Q q0P ‘] ie J a ! LILO 3% \2 SD gf setsnoq¢ Dy ped Sl AG & Ws alo 0 5 MA+ ZVI 7¥02 S & YR Bil 19 9 JINMVHS or iy? aL {hs US gnY P[[tAUeW \ KY aadse9 ) d | Qnp des O 9 Sighs Ss (on UeM1947EeT°F *, (A AB Aw Oo 1 LLY IE 2D sao | | |} Ee 4) “Gan ! Zs ‘sva's | 25S ONS 1 Og fe) < ! a o|z ? [ P12 Arve TS po 1D WN m | | yy 1 49 b I ! ' 1 “Id XITX Hd VYSONOW AZAUNS WVOIDO1OSD ’*S "N eh teee iNeD TX: [Names in italic are synonyms; figures in black-face type indicate descriptions; figures in ztalic denote illustrations.] A. Pag>. Aceratherium incisivum PAGINA aa A ao SSIS COO XXIV Acetabular DATs oo- - -<---s-ac 8 =~ 4= 75 PXCOtS DUN sate ace a nonto= ces 66 Acknowledgments to those aiding... -.- XXVII, XXIX-XXX AGOCUs WHTIOIOSHS seen ee ee ea eee = enn 179 Agathaumantjde.............-------------------+----- 72 Agathaumas Cope -....-- sere ae cheese ses 11, 72,78, 104-105, 105-111, 112, 113. 161, 162-163, 164, 165, 168, 187, 188 TV OLCODOE ee ans series tae oe aaa a 11, 67, 111, 112, 168 SPILENOCEVUS sen ana nel eee inlnl= sna iene = 1-1 = 188 SUINESTEC COD eich a ita arate nein = 72, 84 Sylvestris Cope.......- 4,11, 104, 105-111. 168, 179, 184, 248 Agathaumide Cope.............-.----------- 10, 11.13, 72,113 Alberta, fossil remains discovered in.........-.- 177-178, 184 PPLISPUCNOUAS seen en noe as serln cine te = 17, 18, 17, 18, 28, 30, 120, 148 AripiaT- 2s ses 5 sone ae 41, 42, 43, 130, 210, 252, 262, 266, 280, 286 Ants, aid of, in collecting fossils........-......-.--- XIX, 144 IM SIG SEN I Rigs ooo shag sess se eseghe se sancrmacnes seomae 190 IN) GIES Ghose fae Als 55 SS Sa6r op one oS aninae XXIV Arapahoe beds, description of ......-.....---.------- 182-183 fossil remains discovered in.........-..--.- 7,174, 182-183 woke nee oSs- a ssceicten en se easeoste ages 156 42,4248, 43, 130, 210. 252, 264, 266, 280 $2 Sass Ae Asso cen rigesc ae noseeS 175 Joe Or pean Gee BOL erIcHOsoe 63,79, 230 “cos esas eae Se sea ens ee Sage ac Con Ee aeeaae 47, 48.76 LAG! tiny CP TOICE ee Sane Sau stn Joa socee besa: capeesteas 387 ANIGIGONY IC RUUS ies ctoirien clon ma aie ine lene = ain 17, 28, 37,37 mUShIIAACOXAbODNIA TrOM-. oc enh cee esc cloc se qe epics 12 JCA Ee” 2c BE ADO Dee aE REO OA SDE OEE ep aE ea aaES 48,76 B. LRG 0 [0 eee a 8 Oe Oe Sen Soe OS ape oes 179 Basioccipitals . 6, 15-17, 15, 17, 18, 38, 82, 121, 264, 270, 272, 282, 290 Basisphenoid 16, 17,17, 18, 18, 28, 82, 148, 290 Lyn 10 1 i See SE oe ae ee Bearpaw shales, correlation of Beecher, C. E., fossil remains discovered by........... 122 Belly River beds, correlation of.......................- 174 fossil remains discovered in.. 9, 67, 88-100, 161, 174, 177-179 WI GWIOM Meena eter a! ai acum caciaa Sawant cakes 177,178 RSPR et ere eee ose ut aioe an ok oie asia isiawie se wee e's XVII Berger; Frederick, drawings by : Xxx Berry Creek, fossil remains discovered in.............. 178 Bibliography of Ceratopsia.......................--- 196-198 of Hatcher's works...<....... XX, XXIII, XXIV, XXV-XXVI Birch Creek, fossil remains discovered in............... 176 location and description of........................ 176 SESE OTB aicsetier eels Ae e Oy Seite chal oriole! a: wha Billoo «whole a6:0, fol 115, 116 ELIOORTUED WEAR A LS ere Boa mide aie ane aoe din lav ce cata XVIII, 6-8, 11, 104, 115-116, 1145, 168, 170, 172, 183, 184, 188, 234 Page. Bitter Creek basin, Wyo., age of...............-.---- 106-107 fossil remains discovered in..................---- 106-107 Black Buttes station, Wyo., fossil remains discovered MCAT yaaa ee eeee sees 4, 67, 105, 161, 162, 179, 184 IBOSUMICRS Es ALS NVOLKIOR cco gsi oa bcke sede eonemen 141 IBID Ge se Bees 2s ee ee SRP a ort tere 31, 115, 116, 118 Brain; description Of-----.-..5--2---s-ee-- Be saint 36-38 IBTONCOSRULUS Sree sac en ane conn os ae soe ale eae Xxv, 190 Brontotheriids, work on.............-. XIII, XIV, XVII, XVIII Brown, Barnum, and Lull, R.S., fossil remains discoy- Sled Dye eae abe eee wea ee 9, 185-187 Buckman, S. S., and Schuchert, Charles, on ‘ plesio- UN a Ae SA ESO ie CIC a a Me SRS) aaa 136 BiNzel EON DINGSAULIAe ema woe as Gee y ete ane oon 82 C. Camara Salirisseeneee eter yaaa eae ntes Bae wclen serene XX Canada, fossil remains discovered in...... 5,9, 88-100, 177-179 Cannon, G. L., fossil remains discovered by.......-.- 6, 7,116 Carotid artery 387 Campuses feccetecosei-ce ee 61 Catidaliribsts2 en sane tees 59 Catidalivertebreea = tec esc elcceee a a ceeae 53-55, 65, 77,77 CentTOSa INS se eee eee eae cei 161, 162, 164, 167, 168, 187 apertus Lambe.........- 93, 167, 168, 179, 184, 202, 206, 246 Gentrirm scence mma jit ice cle we sing a ecm a ceo inynlelmienc less 114 ISAT AT ODS WLarS iis ett Sin cis Wists tis sine wie a ais tere crate ore XVIII, 11, 23. 32, 72, 86, 87,95, 100, 101-104, 115-116, 117, 119, 122, 143, 150, 154,161, 165, 166, 168, 172-173, 176, 187, 188 GB78071) AIcOTNTS; MATSH 2. sate wre wi- moses 6-8, 1i, 104, 115-116, 115, 168, 170, 172, 183, 184, 188 (Monoclonius) belli Lambe .................... 12, 89,90, 4, 96-97, 97, 100, 161, 167, 172, 173, 179, 184, 202, 240 (Monoclonius) canadensis Lambe.... 12,89, 98-96, 94, 95, 97, 100, 102, 161, 167, 172-178, 179, 184, 204. 208, 234, 236, 244 ROPTLOi Sua shee. areten cee ctei= 2 11, 12, 18, 23, 27, 28, 29, 30, 38, 64, 104,116, 117-122, 121, 124,127, 128, 131, 148, 164, 168, 169, 172, 180, 181, 183-184, 187, 206, 250 MmOnpAMiUsiMaTShE fae ee sapien ase eee oe 7,8, 11, 21, 71, 87, 95-97, 100, 100-102, 102, 103, 118, 119, 161, 165, 172, 173, 176, 177, 183, 184, 204, 208 (Hadrosaurus) paucidens Marsh. 11, 103-104, 172, 176, 184 (Monoclonius) recurvicornis Cope ............. XVUI, 5, 4, 6,12, 32,72, 81-87, 86,92, 100, 161, 165, 167, 172, 178, 176. 188, 206, 208 SP eines tacete vc cciincm ba anas oalecsleteta terion estieise 204 CarntopsibEds ae Oleneccce. eek urcen ste eccaasus scotuc XXI HIGDOBL CLONE Of ep he siatee wnla wlelecs ie aieiaie niet ote cinia ciate oYa (sia 194 location and description of........................- 180 Ceratopsia, appearance. of. .-. 2. 2... cece eee ewes 188-193 bibliography of................ ened nc ARN 196-198 CIBRSINCHULODIO Neen eee wien olan sie ale sta eta w ict alk lates 10-12 COUBCTION OL ss osere S.ncu cist bein cane nienstotes XIX, 144, 185-187 descriptions, systematic, of -....-....00.00.nsens- 66-157 CUBCOVORVIOl MIBLOTV) Ole aun cn cmc sais vwwwicacicie = 3-9 295 ‘ 296 INDEX. 4 ; a Hy “ Page. i 4 Page Ceratopsia, distribution of s25_ 2-2 oe 174218495 Dental paz ine ee eee 46,130 distribution of, map showing --....---.----------- 294 | Dentary CnvironmentiOlsee esses eee eer eee 194-195 40-1, 41, 43, 130, 140, 140, 171, 210, 252, 262-266, 280, 286 European representatives of.............--------- 12-13 | Denver, Colo., fossil remains discovered-near._.._. 6, 182-184 evolutionsof 2225 eet ae yas hee eee ae ROC 187=188)) | DenverbedsiarelOf=se res —see eee eee eee eee 116 extinction of, causes of..-........-- 195 description o iaesenes ease ener eee aoe 183 genera and species of, list of 11-12 fossil remains discovered in---..-.........-- gee 6, habits'oh. Ao ekhoee sae eee ee ee ee eee 193-194 67, 115-116, 132, 161, 174, 182, 183 Osteologys0he- Meas Gceea cee eee eee Seer 14-64 | Dermal spines and! platess= 52 --5--222-s2soee eee eee 65 phylogeny Of serene ee eee 161-162 | Diapophyses ...............- 48, 48, 49, 52, 58, 91,92, 95,106,107 — Tesborations(Ob.:éces ce eee eee 188=190s | SDiceratheriumissesss5 see eee ee eee ye eee XXIV Wie WS Ofte ets eee eee Frontispiece, 189, 190,294 | Diceratops...... XXVII, 149, 161, 163, 164, 166, 168, 181, 187, 193 ‘taxonom yori. 25> vee cease: sates cee a eee 162-174 hatcheri Lull.... 149, 168, 181, 184, 192, 202, 206, 208, 292, 294 Ceratopsia localities, geology and physiography of .. 174-184 | Diclonius (Claosaurus)............-....... XVIII, 65, 70, 71, 82. IMAP SHOWiNg ses Ta eee ee eee ere eree 294 calamarius............- fost - Sse eee Ceratopside Marsh................- ,----- 5,7, 10,11, 12, 67, 69 perangulatus Bina A. LS 101. Dimocerataca-s-2--<--- eee ee ee 3 23 Brie Aa at OR ay ER eves ALY Stade! = es .. 89 | Dinosauria Owen.. 48, 59, 76 46-49, 47, 48, 76, 76, 80; 278 Dinos aneciteritiies wi Soe iace Scene. S ca sak else Ae eee See een 10,11 classification of 1075 DiphyodusjlonupirOstris ee na ee 179 Seneee asa eee 72 | Diplodocus Marsh. ------- Sse ateeee BOG SOSH OO RR aE ee ene ee ee Tl ee earmepiels » 58) Seats see es ee eee ee Kou PAS oA EN Gy ele Se Myer pe AR ea ol te I eS 179 | Discovery of Ceratopsia, history of.................... 3-9 Sates, eS SSE ARS BRO DR on a EO 69 | Dog Creek, fossil remains discovered at..... 5,81, 103, 175-176 RR ets eae es oe Pi. Ses i oe Ua 111, 112, 113 location and description of... 522--------.-.----. 175-176 Claggett formation, occurrence of..............----- 175s. olichorhints ses -eete sees eee ee 2 Claorhynchus! Copeseeee serrate eee eeee ener Id, V14,'166.,| Dorsalribseases sess eek Sees 2 ee ee itrihedrus'Cope..07- 02 ae ee eee ee 11,114,166 | Dorsal tendons, ossified Claosauridsel: See ee ree Bae ees Oe ate ena 156) Dorsal vertebrasoss 2. sess ae aaaen ee ee 49-51, CIB OSR ITU a aR ene ele Hey pte XVIII, 194, 195 50, 51, 76-77, 77, 95, 106, 107, 109, 212, 214, 278 Climate, change of, extinction of Ghatopotn due to.... 195 | Dry Creek, fossil remains discovered on...-.-......... 180 Cnemisliones tz as oe ee eee eee GE} |) Dini ORPIbINC ES ooo ene scesocssacsoecccestesccosese7= NGG Collectime*anethodSlotese esse ea ae eee 185-187) || DGYDLOSAUUS ea ee ees te ee ee eee eee 114, 166 Colorado, fossil remains discovered in ......- Phere haere By || DNS mANDIS (COI). = seco tecsssecasesozsoccs 12, 67, 67-70, 70, 166 6, 67, 101, 111, 115-116, 118, 132, 174, 182-184 bicarinatus Cope Bd ae abe sss obeaselscesocenans 12, 67, 68, 166 Colorado} eroupy section ofsee- ee eee ase seen eee une eEnealsius; Copess seas ss see see 12, 67, 67-68, 69, 116 Converse Co., Wyo. See Wyoming Ine hyoleronEVOIS (COI) 5 iat ees sseeesssss5- 12, 67, 68, 166 Cope hb explora onsite ease eee 5-6 DELS ams) CONC seem ere esas ee es 12, 67, 69-70, 166 HU SUTES OY iaes cyatenisic coe eee ee Oe nee ae 248 = 3 ' HOS COULEAHONG OF. job sowecdddeccooe=sessccscacac * 66 ye fossil remains described by...........-- “4,5, 67-88, 104-114 F fossil remains discovered by.............--------- 5,176 | Eagle formation, occurrence of......-.-.....-..-..-.--- 105 Coracoid..........-.-- eke We ke 58, 59, 78, 79,90-91, 238 | Elasmotherium sibiricum.....--...........-------.---- 98 Goronotienaeen Fcc 5c nicpetiel ct etn Re eceiee 137, 137 Eldridge, G. H., fossil remains discovered by.. 7,101, 116, 132 Gorondid: process, Se422 oe Ba 40, Environment of Ceratopsia, character of.... 187-188, 194-195 41, 48, 94, 130, 137, 137, 140, 210, 252, 266, 280, 286 | EPijugals.........-.......-- 14, 22, 82, 154, 15, 252-256, 266, 270 Cow Creek, fossil remains discovered Ole ee 176,177 Eipoccipitals Shape 14, 18, 19, 20, 21, 22, 22, 84, 126, 135, 141-142, location and description of.................s.... 176-177 202, 252-266, 270, 274, 276, 280, 282, 286, 288, 292, 294 Cow Island, Mont., fossil remains discovered near.... 87,176 Europe, Ceratopsia from.-....-.....-.-.-...--2-2+.+-- 12-13 Cranial armature usiachian: Canale se tse ese eee ees 18, 27, 29, 81, 290 (Crrniattitaal se dec ko oe ete Rs Ceratopsia, outline of-------.--22.--.- oe Cratzomus Seeley rxoccipitals ........-- 2.2.2.2 2022e22ee2eeeeeee esos 6, 15-1 lepidophorus Seeley 16, 17, 18, 37, 38, 82, 121 141, 264, 270 272, 282, 250 pawlowitschii Seeley Exoskeleton........-.-.---- onescarteoseaocrssssccso025 65 Crests eee ee Extinction of Ceratopsia, causes of._...............-.- 195 Garailan isin Eyes, development of. . Fee oc nm een ene ne ee oneness anne . 38-39 Crocodilia css wesacchiccee PAS eee ; 107 P Cross, Whitman, fossil remains discovered by 7 : D Feeding, methods of-...........-. He ae Eee ee 46, 193-194 3 MemMUTs = 01306. eeca aeons seeSte aoe See 62, 63, 79, 80, 226, 228 Dawson, G. M., fossil remains discovered by BMI UO Ein Hoc caacbosaabasonccbewig sa sonccanones 62-63, 64, 80, 80 onyBellly sR era) G0 Seen ee eee eee ee eens 178 | Fish Creek beds, correlation of.....................--- 174 Deltoid:ridgerc. west Ges eee ee oe a ee 79 | Fontanelle, interpremaxillary..................-....-- 81 Denials tora me eee ease eee 26,41, 41,130 | Fontanelle, parietal...............--..--.--- --- 20,90,151, 202 Dentaligroovesss jana -os- eee ee ence eee ee 29,80, 46,282 | Fontanelle, postfrontal............. 20, 24, 38, 202, 272, 276, 294 INDEX. ; 297. Page. Page. Fontanelle, supratemporal.....-.-.....--------------- TSO tea SEP LLeHe ets LD ID Yas cate eee Sects aa = oe 294 HOOUMCHATACLOLON sane aera wane sane a 46, 193-194 on) geologic depositions =. 422.26 =-45-- 405. 2ee=- es SRT WOOP MOTO sence sen ss tse obec edene 59-60, 61, 62, 190, 232, 294 on Hadrosaurus paucidens: -2--.-22-<-- sess snene== 172 BEG OL) in Oars eaeet esa Ja 63-64. 64, 189, 190, 232, 294 Qn WUGhURIVEr DEAS som es eee ane eeeee nesses 175 Foramen, anterior palatine. .-.......------------------ 31,36 on Mesozoic:conditions. |. ..... 222 2-25.2025-55.¢5- 194 Horamen,imtraorbitaly.- eo. - 0. =e eee see ac\-2n. Oper, 129) | OT NTOTIOCLONIULS er ese ee men tore Eps 172-173 See also Foramen, lachrymal. on stratigraphic range of titanotheres.-........--. XVIII Foramen, interpterygoid--.-.---.....----------------- 27, 28 | GUYELOLOSATITIES eae eee oe Boe ie eee 161, 174 Foramen, lachrymal eee AL 64e on Triceratops 169, 171, 258, 260, 266, 268, 280, 284, 286, 292,294 | prelace Dyess ee ae ae eee ene ee ca? Foramen, olfactory 2 18, 18, 28, 30, 37, 164 | scientifie contributions of. .-........-......... XX-XXVII WGErAWIEN, OPLiCs == .a2eseeccsee ss eesse< See < 17,19, 28, 37, 38 OTK Obeeret oe hance ee Unt ae teats Remo XIV, XVU-XXVII Horamoenpanietale mance ease ances dean Sab ce 35,151 | Hatcher, J. B., and Stanton, T. W., on Judith River Foramen, pineal, description of ......-..----.--------- 24,35 | [pad Sane ee tas See oe ee SS 2 174-175, 177 See also Fontanelle, postfrontal. lmiiave ON Pe Gr LD ySe ANUS week sae ees een er crerte cs eoe 7 Foramen, postirontal...-..-----.-....-------- 22, 35,151, 154 | Hayden, F. V., fossil remains discovered by.-..:..---- 3 See also Fontanelle, postfrontal. WROLOLe ete Serer ee NET: ae eben eae hn etree 3 Foramen, pterygo-palatine.............--.-..---.-- LEReVALS amen Ng N SENSE Ole seak a eae ce See Noe eee oe Seen 39 Foramen, temporal....-.-.......-.------------------- 35,151 | Hell Creek, Mont., fossil remains discovered on.......- 9, Foramen lacerum posterius.......--. 17, 17, 18, 28, 36-37, 37,39 | 179, 182, 184-187 Foramen magnum ...........--.-------- 6, 16, 36, 121, 264,272 | Horn core, nasal......-- 14, 18, 22, 28, 32-33, 33, 38, 85, 86, 88, 88, IoramensOValOsss << .0.-- so acne een ee cneees 17,18, 18, 37, 37,39 89, 92, 93, 115, 121, 129, 129, 132, 136, 137, 139, 140, OTIS TO ML Ness sae oe 2 ae cet eee 17,18 140,141, 151, 154, 154-155, 162-171, 173, 174, 187, 206, Woraminardentaless oc asaa-cacnen sae ece awe 26, 26,41, 41 | 234, 246, 250, 258-252, 266, 268, 274, 276, 280, 284, 286 Roraminga, miscellaneous 5-0-6 2-2-2 elem - 16-18, 16, | Horn core, supraorbital.......-- 5, 8, 14, 18, 19, 22, 24, 28, 80, 31, 17, 26-28, 26-28, 30, 31, 36-37, 37, 41, 41, 42, 73, 82, 89 | 32, 38, 74, 85, 86, 86, 94, 94, 102, 103, 115, 121, 121, 128, OGRE: UGH HOE Goce ein oo eee sense sace sone = Sarees 35, 161 | “ 131, 133, 183, 134, 135, 141, 151, 152, 154, 154, 162-165, Fossx, supratemporal..---....---.- 20, 22, 34-35. 151,154, 154 | 167-174, 187, 208, 234, 252-276, 280, 284-288, 292-294 Fossils, collecting of. aid of ants in................. X1X,144 Hoxmicores: discovery of- =. = - 220 n=222<22-= eee - Xvi, 7,8 LOM “Sc Roc casa ene eee ones dseqpe ore eeececere 14, 16, 20,19-22, | Hulke, J. W., on Dinosauria ............-.---2--2---+- 82-83 125-126, 131, 154, 162, 164, 166, 169, 170, 171,187,202 | tfymerus...........----------- ee 60, 60, 79, 79-80, 220, 222 LN) Wife OS sao oe Shee see sone segs Sade sees Zt 20S) aiiy oe OSAUEUSS ae Saecee sere aise sete eae gate aes oe oe 13 On TAIN oes cc's « eb aee ern badsea-heanocec 18, 22, 24, 30, 38, 74, | 74-75, 80, 99, 124, 145, 242, 254, 256, 268, 276, 288, 294 | I. G. |) EAB ROC NENG Secon oom aoc Se seas seb Sen SascarcseaS 10 | i(Ghiteaie ee oe 54, 56, 56-57, 77-78, 78, 106, 107, 109, 162, 163, 165 Genera and species, systematic descriptions of.....-- GO Tmlelapace dN Genworkofeestatenss os ake cet e oe ss 5 revision Of......-..-..--------++++-++20e+0222ee ee: 161-174 | Ischium...........--------+20s2seeee0+- 56, 57-58, 58, 78, 78-79 Gilmore, E. W., on Triceratops prorsus 189-191 | TOBE fey see a ete ee eee eters Hees 163,189 | J. (CSIC) GUD IN peo ain oon Sabo me ee oe leg apelawmeridecen scence cece c cmctecan se 39-48, 40-43, 94, 94-95, Gosau beds, Ceratopsia from 126, 186-137, 137, 139, 142, 210, 236, 252, 266, 280, 286 Green Mountain Creek, Colo., fossils found on......... 6 | Judith River beds, age of........<......-.... XX,XXII, 104, 174 Guernsey, C. A., fossil remains collected by...... 7-8, 117, 118 | fossil remains discovered in....................------- 35 i. | 5-6, 67-104, 119, 161-162, 165, 174-179, 184, 187 , | CORGUMNOUCAIOQtee eee mest tea cenh eae eecinwtcicte 174-179 Habits of Ceratopsia, sketch of..........-....----.-- 193-194 ICO LANE eer ee MU eneiomio’s erate meiaec emer tars wee 76 FIBA TOSAUTIGE 2. oe oon a oa oe meen oe we ew innn eno 113 EY DEMON) AP PEATHNCE.Ole ese e ee ecsiee eee se nee ee 188 TERSCCHNC Eg PLU UIA se atm = stec eieca alot olin inl v = DAU Od RULe LOCI ONT Ole ett eee acter oleae a oP tel os sete aaa 25, 28, 36, | MOUS NYs 2 Seaceee sa eeeee oso seec ore i el ' 38, 129, 130, 250, 252, 256, 266, 274, 280, 284, 286, 292, 294 Lusk, Wyo., fossil remains discovered near... XVIII, 8, 179-180 i Tevdekeker. Richardson Coston ciinoniieloGr viene 13 Nasals..-.-----.- 22,25, £5, 28, 88, 86, 87, 8, 99, 123, 129, 140, 140, vigor , BUCP SIA IES. SE Nae ae 146, 171,250-258, 266,268, 274,276, 280, 284-288, 292, 294 Nebraska, \workaimes< cf ies sence ease cee See eee XIX-XXI as Neural canal............2.--- _... 48,50, 51, 91, 95, 106, 107, 109 McGee, W J, on J. B. Hatcher.......------------- xvu,xxmu | Neural platform....-.-...-..-----.-..-.-+---------- +-- 91,97 Mandible. Sce Jaw, lower. Neural Some) ssacaeostias s2sctes 48, 60-52, 65, 91, 9, 95, 106, 107 MERCH HOSS. os 5.cs noc sone adoseesosee 41, 187, 140,171 | Nodosauride.....- peste ec ese e eee ee cesses eee 156, 157 Mandibular ramus. See Jaw, lower. ‘| Nodosaurus Marsh..-......----.-..----.--- XXVIII, 12, 155-157 Manospondylus)|Copes=--— seo. e- see ee sees 12,118, 166 textilis Marsh ....-.-........----.------- 12, 155-157, 156 ISAS CODE! seem ia hee es eee 12, 113-114, 114, 166 Nopcsa, F. Baron, on Dyseenus Man feoceras. a 2535 sae as aes eee ee eee eee canes _XXIV on Stegoceras..----.--22- 22-22-22 2202022 eee eee Marsh, O. C., figures by -.-.-.- 15, 18, 19, 21, 22, 33, 39, 40, 43, 48, 50, 52, 53, 55-58, 60, 61, 63-65, 103, 115, 121, 129, 130, O. ' 151, 154, 156, 190, 210-232, 252-256, 262-272, 286, 288 | Occipital condyle -..-.---.--..-.---------------------- 6, 16, TOSS CoOMechONStOhe ee eae eee ee ee 66 17, 17, 37, 84-85, 89, 103, 131, 155, 264, 270, 272, 282, 290 fossil remains described by.............---.---- 100-104, | Occipital segment--__-.---__.__.__._.- 6, 16, 15-17, 17, 18, 242 115-141, 143-148, 150-157 | Olfactory foramen. .........2:.02-2¢02--0e-eneeee-ve-e- 28 TO TO Lea Pets hoe San ha alge nn i nee ns SON Gro. | WOMbEXe rovny Molo seo se Ses a 37,39, 39 Mair 1170.1 all ama pviO TEP Ta ees ean ea na ar Soe Ophir dia yy) se Meena tess he IS oe seh ah eer yee es 83 Matthew, W. D., on Mesozoic conditions..........-- 194195) |e Opis thoccell aes se ee essere ee esse eee ese eee 11 on Triceratops......-- ish ig saat eee ee OEE ES 124 | Optic foramen 17,19, 28, 37, 38 Maxillariess/s.5 fees sonaseense 25-26, 26-30, 88, 46,89, 123-124, | Optic nerve. -..-....-.---..-..---------------++-------- 39 146-147, 250-262, 266-270, 274, 276, 280-286, 290, 292 | Orbits........ 18, 26, 30, 835-36. 74, 86, 89,94, 94, 141, 162, 164, 167, Medullaoblongataege sss eee Aaa eee seeer eee 87,39 169-171, 250-252, 256, 258, 266, 272-274, 280, 284, 286, 292, 294 Meek, I. B., fossil remains discovered by --.------- As105 e106)" Ornitholestes=. @-sess-ae2 See eee eae eee eee see 194 Mecalosauriaies \sj.cseasesee ae eee eee ee eee 10,11 | Ornithomimide-.-...--....-.- Cee : 156 Megalosauridze 1560) Onnithomimus se essss=seeeeeeoeeeee 179 Meralosalitus ia ssece2 see cee ee eee EEA ete hae 107 , Ornithopoda Marsh 11, 103, 156, 157 Meiolania platyceps Owen........-.----------------- 1015107 | Ornithosauniaosece sec esnen cee seen eee eee sees 83 Motacarpusiss 2.82 ees soe cee to cee eee 61, 61,232 | Orthopoda Cope....-...-.......-.-._._.-.- gactasorssce 1S Metatarsals(c ss. yess tc sete e eee ane eee 64,282) Osborn) Herk tOoreword bye re staan sates eee XIII-XV Metatarsus onidescentsof Ceraitopsiaessa se sen eae ee eee eee ees 161 Metatheria on mid-Cretaceous fauna........-.-..--.---.------- 89 Missouri River, fossil remains discovered on sketch of J. B. Hatcher’s work by..-......--- XVII-XXVI , INDEX. 299 Page. Page Osborn, H. F., and Lambe, L. M., fossil remains de- Red Deer River, fossil remains discovered on....-...- 89-100, BOVIMEGUD Yrs ae a eee ee eee 9,88 177-179, 184 Osteolozyiot Ceratopsiss - 5.2 Sop -t a es ce ea eawee ceeeee 14-65 WIGW? ODE sito a. oe sess oe re ees seo ae ose 177,178 : * PROD OUR a2 Satie asic since redeem cle oa cae 10,11, 83, 111 E- RDS es ee eee ak Beet ae ee 48,59, 110-111 Palatine...... fe ees 26, 27-28, 28-30, 38, 120, 147, 282,290 | Restorations of Triceratops ......--.-----..--------- 188-190 Palatine vacuities, posterior.................-.-------- 36 | _ Views of ........-.--.-...-. oeces Frontspiece, 189, 190,294 Paleontology, vertebrate, work on, condition of .-.... x11 | Rose, Doctor, on dental foramina.........----.----.-- 26 TG G ogakh eon Concee Se Bae ae Bene ae xm-xtv | Rostral Palsosiicnsiliciivaeee ete ee a 69 22, 28, 29, 31, 33, 33-34, 38, 121, 123, 129, 164, 168, 169, Petpet ae ees. Wr ys Co 48-49 170, ou 256, 260, 262, 266-270, 274, 276, 280-284, 290-294 LP RUS IGT OWS oe soe e a= SORE Senne ae eee OB aS eoc os 29 EPEC IS ULES eee pa rene nai ait ae iene Pelee sie 14 Ss. Parietal fenestra 165, 202, 294 a aged see abaieanietaltontarielle: Sacral vertebrae =.~...25-.--.:+.- 31-53, 52, 75-76, 106, 110, 248 IE SriCtAMONEANEN Canes oo eine cae ests cmcee cess 20,151 SACKUM. -- 2-9-5 Fate a ela ae a ie RUE PT Ee eA 51-53, TEP EGIL Gey i noe oe eee ee ee 35 52-54, 75, 75-76, 91, 92, 92, 106, 110, 162-165, 216, 218, 248 Parietals............... 14, 16, 19, 19-22. 20, 21, 22, 31, 88, 73,73 Sauropoda Marsh........- XIV, XX, XXIV-XXV, 10, 110, 118, 156 £0, 89, 90,92, 93, 97, 97, 99, 121,125, 145, 151, 161,153, | SCAPUIA----------------------------+-- 58, 58,78, 79, 90-91, 238 154, 162-166, 202, 240 242, 252-270, 274, 276, 280-29} Schuchert, Charles, and Buckman, S. S., on “ plesio- , 162-166, 202, 240, 2 270, 274, 276, 2: Ea f XIX, XXI-XX11I | : WAGE Sai 2S 22a pets goede sia esa ab 56, 56, 77-79, 139 SCOLL pWietb +00 Js bee atChenee ae see ene meece © a eeee XIX ; ; g4 | Seeley, H. Gi fons mosalnisiee ae eee te ee ee §2-£3 feo ie helg bees oy 61.62, 65, 64, 232 | Seminole Mountains, fossil remains discovered near... 7 ee Cue a SGHSOIOL EATS Se seer eee era te ee en ER Seger 38-39 ie i poy Le Oa ee a 161-162 Shouldexpindlet sss cuteswn ere ls sean SISA 2 ey Ue BS 175 Sight, powers of, in Geratopsial.. so. 2-22 snn--o--255 38-39 24.35 Skulls... ..- 14-46, 15, 18, 19, 22, 28, 29, 38, 73-75, 115, 119-132, 129, Be Loe bic malr h e 37.38 134, 1388-139, 141-142, 144-145, 149, 150-153, 151, 154,185 Pituitary The ees ey en Py AP LI SS 37. 39 WelehtiOfe= -atca2 = cae acecce sets oes ences XVMI, XXIX, 185 Plesiotype, definition of 136 Rel) \SOUSO LOL a eencc ingen nro somnp nen area n= 38 Fae) rTST ATA Coa ole ai Sears oe lah el ale 13 South Dakota, fossil remains discovered in........-- 113-114 Polyonax Cope.........-.---------++- 12, 72, 111-112, 113, 166 Og Og tite aitgnccips ssaaceene echo s8~Ese at Ph ass aes mortuarius Cope. . eS tle ul Lo 113, 166 Species and genera, revision Qiemsrtgo= S26 seaseass2see 161-174 Postfrontal fontanelle.............. 20.24, 38, 202, 272, 276,204 TSO AEE C9 S010 OTS) Ses Seon Coker eS See also Foramen, postfrontal. Sphenoidal SBM 2 Smo orisedeses oe erttcteee 17-19, 18,82 Rostirontal foramen ys < - soe a- ene ens oon =~ 228, 855154 SUID SSCS ROE CES no ee Sa eect ea SOs Mang ge ae See also Fontanelle, postfrontal. pee ak we Re ee a as Ae GP many ta ea Ws Boattcontals osc. se 5, 17, 18, 22, 24, 28, 30,38, 74, | Sines, A EE Note ee ae oA 74-75, 94.125, 145, 250-256, 266, 268, 274, 276, 284-288, 294, 2 eee Rosana uses Ge ghee aD ne Aled; as Fotomacvtormation, ape off... 2 se. sone cc cc nee eee XVIII STEER Sa CSURSS generac eag Sat Tes one ce seear sat Baedenfary fi ek 18.19 SOUAMOSA See e ee seni g- eee -- 14,16, 19-22, 20-22, 31, 31,38, 89-40, 40, 43, 130, 130, 238, 252, 262, 264, 274, 280, 286 Be aoe ise pede geal g et teey hoy Ly Bradentata Marsh 10 1 157 154, 162-166, 171, 173, 202, 204, 252-270, 274, 276, 280-294 Prefrontals.... 22, 24,124, 146, 202-256, 266, 268, 274, 276, «84-288 Stanton, 'T. W., on Agathaumas sylvestris..........-. ion Premaxillaries ........-0.-2...0-----..--. 22,28, 29,80. 31,38, SO eae COL IODSEonpcer et 3: See pat Bs a 87, 88, 123, 129, 130, 147, 250, 256, 258, 262, 266-270, 280-292 on stratigraphic position of Ceratopsia .....-..-.. 180-181 Prentice, Sydney rarines by ; ame Stanton, T. W:, and Hatcher, J. B., on Judith River i MOXA WINES OW Sala nin aversion emit XX) + ExGnACrHl Ore DLibesee eee cece uc. ae eu ctee vcece HIG 218 | = beds ee ge ee BARC it GEC a anal emus Ei a ee Oe ee i ee 100 Stanton, T. W., and Knowlton, F. H., on Agathaumas Pterygo-palatine foramen .............-..-.....2--- 26, 27, 28 —1 EN NGun geese ane argc J Boe abc eae st Pimpin tee ae, nah 23, on Ceratops beds......:.....-.--..+-+.+---------- 183-184 26-27, 26-30, 81,120, 121, 121, 147-148, 270, 289 290 Stegoceras UaMbere. ssccn nsec ce~=see == 12,97, 98-100, 166, 242 Ptilodus primevus "178, 179 validus Lambe!...-22..------ 2.... 12, 98-100, 99, 166, 179 Veh; ToC ae yD akg oh DT 56.57 5) 78 StEPOSAUTI Mes etse icone ws cestode steers XIV, 118, 155, 157, 195 nee tee acta a ake a ape oe oe, 57, é : POROPMGMIIE WOLEMON