\ \\ < AA \ \\ \\ SS . A \ WN AN WW WY YY S \ A WY \\ S . \\ \ » ACO KK \) QQ GW \\\ SMITHSONIAN INSTITUTION UNITED STATES NATIONAL MUSEUM ButieTin 169 - THE FORT UNION OF THE CRAZY MOUNTAIN FIELD, MONTANA AND ITS MAMMALIAN FAUNAS BY GEORGE GAYLORD SIMPSON American Museum of Natural History New York City “i UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1937 For sale by the Superintendent of Documents, Washington, D.C. = - - - - - - - - - Price 45 cents ie Fe Paar yeas eA es tee ats 7 7 es | ee ees ce aha uae ree - i a wi eo vat craw a 7 i e" } nireet: . we i ae Vs _ - hs ; Li, o* ae oa i ; ee 7 W's v7 =~ . - th me) i ’ 7 : a, Pa rT 7 us i : / ~ , wo aay “HE. (OT. ee : iw aa . uk i a . soto Ne ede AGO AR we THOME QTE, ce Lace Var WSK 7 ae ce oar wth nee ary Bocas ae a - ; - Pena: fi - 7 7 a a ‘ ; : | ; \ , _ 7 “a _ 7 ; + Cpe He UAE; Lavi a ales Sie a = ! ~ al is itt at I wth — 7 , : DN we ot 7 . - »* i ae 7 oop 4 ‘ 7 > Lt os 7 : ; i 7 ; ; : gi ee Pes - mj aa any i pas : | ,. bd * ee . : 7 Ree a) wal f i : are << = ee ok en San Ne erage aan _ : ere 7 7 ; - “Sheen -_.' & ; a . o > 4 4 7 a _ al Mies $4 eas is ye ate Wei Pl vray uo Ton rs ¥ ate Dedicated to the Memory of James Williams Gidley, Ph. D. 1866-1931 ADVERTISEMENT The scientific publications of the National Museum include two series, known, respectively, as Proceedings and Bulletin. The Proceedings series, begun in 1878, is intended primarily as a medium for the publication of original papers, based on the collec- tions of the National Museum, that set forth newly acquired facts in biology, anthropology, and geology, with descriptions of new forms and revisions of limited groups. Copies of each paper, in pamphlet form, are distributed as published to libraries and scientific organizations and to specialists and others interested in the different subjects. The dates at which these separate papers are published are recorded in the table of contents of each of the volumes. The Bulletin series, the first of which was issued in 1875, consists of separate publications comprising monographs of large zoological groups and other general systematic treatises (occasionally in several volumes), faunal works, reports of expeditions, catalogs of type speci- mens and special collections, and other material of similar nature. The majority of the volumes are octavo in size, but a quarto size has been adopted in a few instances in which large plates were regarded as indispensable. In the Bulletin series appear volumes under the heading Contributions from the United States National Herbarium, in octavo form, published by the National Museum since 1902, which contain papers relating to the botanical collections of the Museum. The present volume forms No. 169 of the Bulletin series. ALEXANDER WETMORE, Assistant Secretary, Smithsonian Institution. WasuinatTon, D. C,, June 16, 1987. IV Introduction History CONTENTS of this study and acknowledgments___._-....------------. IPTEVIOUS) WOLD KS: 22 oe eee ee a te ee ee ee pannel Geology anu aumas a2 s2 28M 9 oon oe eee oe ace Geomraia iy sane a ees ey EN a ae ect Geology DStratiprapliye oo ae 8 ee a ote ee ee Saas oa eae Generaltstrationraphicexco lumi sis 2 ae ee re Cretaceoussertiarystransltionesoa= se eee eee ee eee TOT WTO eee ee ee 2 ey nt SHUG Rs eT SR ee ee ee eee Hossilslocalitiesvandstaunalelistsiess 5-6-2 ee ee eee General'occurrence of fossilimammals=s___ == = = ee a= The Gidley-5=2-5---g--—-2---- Genus Sitipnodon Simpson. 2 <- {5552-22 eee eee amily Pantolestidae: Cope 2222220 Us Sao a ee et ee Pantolestinae, new sublamily== 2 22282 22-5 ance Genus: Béessoecetor Simpson@2--— = — ae eee = eee eee Pentacodontinae; ‘new subfamily2--- 22-222. -22222225252 2 2 Genus Aphronorus Simpsons 2.222522 5--eee eee es Family Mixodectidse Coopers 2a- 2 ance ee eee eee Genus Hudaemonema Simpsons -~ Ss. eae Genus Eipidophorus Simpson-— 2-22 = So aa eee TINSECTLV.OTa, in CCrtAe SECISse sao em ao ne ae ee a ee eee Picrodontidaes newramily= saa ee a es Genus*Picrodus Douglass= =. -- -- =e re ere ee Family uncertaims<225s5 52. Soe a eee saosaeea Genus"and species‘ undetermined2=:22- =~ 222 eS Se Order serimates Linnaeus 22252 Sa soe es ee Se ee ee Family 7Anaptomorphidae Copes2=4--22 25-5. ee ee ae eee ee Genustearomomys Gidley: 2322 soe se eee eee ee Genus and species undetermined-_-___..-..--.=-._-----.. Genus*Palaechihon: Gidley <2 eos ee ee ee Genus Palenochiha simpsonso-2s2"--s2ee.- eee oe ee ee Family Carpolestidae:pimpson 22> - ee ee eee Genns Zl pnedotarsius Gidley = 22-225 22a ee eee Family “Plesiadapidae Trovuessart. jo252. 22 0el se eee Genus: Pronothodectes Gidley 2~ J22222 Soe S2 2-3 eee Genus?Plestadapis Gervais 2-2 ee Order, Faeniodonta! Cope. 22 an tee et eet eee eee eee eee Family Stylinodontidac: Marsh22—- 2 2 ee 2 ee ee Subfamily Conoryctinae Matthew_-—..-_----..----.-------- Genus: Conoryctes| Copes= 328 a oe ar ee ee eee Psittacotheriinae. Matthew oo 2222-22-82 See a ee eee Genus: siiacothersum: Cope 2. ao a ee ee nee Order Carnivora Vieq:d’Azyro2 =o 2b Se ea ee eee suborder Creodonta; Cope: 5-2 eee see se cece ee eee eee Family Arctocy onidae Murray... ano een oe eee Subfamily Arctocyoninsae Giebelo_ =. == aie a eee Genus Ciaenodon Scottue 2282-0 eee ek eee Genus Deuterogonodon Simpsonis- 22-222 sees eee Subfamily Oxyclaeninae Matthew- 2-22 eee es _ Genus “Prothrypiacodon' Simpson eo -aaeeee see eee eee eee Genus 'Chriatus’Copes] 25 ese ee ae ee ee Genus Metachrtacus Simpson 3} —- oan eee eee eee Genus Spernoryodon Simpson. 2 owe beens SUE ee ee Mamotricentes: Tew @enuse 22 2 ae ee eee ee Family: Miacidae"Cope {2 2) S222. oe een, Donen. s 2 en Subfamily Viverravinae Matthew 22-22-0822 2 2 Lee ee Genus: Didymictiis Copess ee a eo ee Genus Ictidopaypus Simpsoni 2222s eee eee eee Family Mesonychidae ‘Cope ss 2 as Bee SOLE Je Genus “Dissacus* Cope 42 2a fe eee nee ee eee ee CONTENTS Part 2: Classification and description of mammals—Continued. Order Cond yiartnranCopes=-— cso eae a See as See Family Hyopsodontidae: Lydekker-.--...-.-...-...-<--=-+.-- Subfamily Mioclaeninae Matthew---.--.------------------- Choeroclaenus Newekenuss== 23a = ee eee ee eee eee Genlshhlinsodon SCOvvese ome nn a oS ose esa oe es Genus bg aletes Sin psOn se a a ee occas Subfamily Hyopsodontinae Trouessart__------------------- Gaenuss 2iomiulius Simpsons. 8 2 oth ts oa eee GenusvHanlatetes: Simpson = a nt So a oo aes Family Phenacodontidae Cope 5. co---2 32e = ae wee Genus Neiracitenodon Scout. =. 2256 = = ot oem Genus Gidleyina:Simpsons=- -< =.= +---2:--2-----5s— Hamily Peripiychidse Cope. - 2.2 sao St oss eet aeons Subfamily Anisonchinae Osborn and Earle_----------------- Genus: Cert phagus, Doupglass= 2. 2. 52s eet meee ae Genus Antsonchus, COpe=s22_ <= 3052S o oe Orden BantodontasCope: 2.662 oe eee ne hoe ee toe ina Hamily, Pantolambdidse, Cope... -—-- =. + ss oso oaa Se Genus, Pantolambaa Cone. ss. en ee Pantolambda or allied genera, species undetermined _~------ RG tRereL LUPE CULOG Aire ot Sa lee a ee eo ote te ILLUSTRATIONS FIGURES . Diagram showing the relative abundance of identified individuals of the various orders and families of mammals in the National Museum collection: fromthe: Gidleys#@uarrys2 se" Sees eee ee ee . Comparison of relative abundance of families of mammals at various localities in the Crazy Mountain Field, Mont__-_..-------------- Histograms illustrating Cabrera’s law as applied to the Gidley Quarry . Histogram of measurements of second lower molars of all Lebo speci- IMECTHISFOLM LCLOCHTCACUS Bae ae a ae a ee oer a ane eee . Histograms of typical variates of multituberculates from the Gidley ang silperlingOQuarmicst a5 52 232 ee ee ae ee ee eee . Histogram and corresponding roughly fitted normal curve of length of P, in Ptilodus montanus from the Gidley and Silberling Quarries-_--- . Scatter diagram of length of M; and length of P, for all Gidley and Silberling Quarry specimens of multituberculates that show both oftthese:measurementsM ae aes eo. 2 en a ee ee ee ee . Left lower jaw of Ptilodus montanus Douglass_—-..--.-------------- . Lower dentitions of Ptilodus douglassi Simpson, P. gidleyt Simpson, and IPEStNCLE1ITTESIM PON sees a ee Se ee eee ae ee eee eee . Lower dentitions of Ectypodus grangeri Simpson, EH. russelli Simpson, E. silberlingi Simpson, and Parectypodus jepsent Simpson__-------- . Part of lower incisor and fragment of jaw of Hucosmodon sparsus Simpson- » Right lower jaw of Gelastops porcus Simpson-—-5---2--=2--22e5-25=— . Right lower teeth and alveoli and right lower jaw of Gelastops parcus SOTINT SOAS es ae oer eee per ns ee ee . Left lower jaw of Prodiacodon concordiarcensis Simpson------------- . Right lower jaw of Leptacodon ladae Simpson-__-------------------- . Left lower jaw of Leptacodon munusculum Simpson_---------------- . Left lower jaw of Myrmecoboides montanensis Gidley__..------------ . Right upper P!-M3 of Myrmecoboides montanensis Gidley__---------- . Left lower jaw of Stilpnodon simplicidens Simpson__..-------------- . Left lower jaw of Bessoecetor diluculi Simpson_...--.-------------- . Left upper jaw of Bessoecetor diluculi Simpson__..--.-.------------- . Left lower jaw of Aphronorus fraudator Simpson_-_.---------------- . Right lower jaw of Aphronorus fraudator Simpson..---------------- . Left upper jaw of Aphronorus fraudator Simpson___---------------- . Left lower jaw of Hudaemonema cuspidata Simpson_-_--------------- . Left upper molars of Hudaemonema cuspidata Simpson_------------- . Left lower jaw of Elpidophorus minor Simpson-_-------------------- . Right lower jaw of Picrodus silberlingit Douglass. .-.--------------- . Humerus of an unidentified fossorial mammal-____----------------- . Right lower jaw of Paromomys maturus Gidley__..----------------- . Left upper jaw of Paromomys maturus Gidley_...------------------ . Left lower jaw of Paromomys depressidens Gidley_----------------- . Left upper molars of Palenochtha minor (Gidley) ------------------- Vill Page 60 60 66 66 74 100 103 110 110 112 113 113 118 118 119 122 123 123 125 125 132 132 134 137 140 150 150 154 160 ILLUSTRATIONS IX Page S450 Upper molar of Conoryctes comma Cope. _- <2. ~~ ---2=--=--2-=--- 169 35. Histogram of length of M2 of Claenodon from the Torrejon of New Mexico,in, the American: Museums = 525-575 o-- 3 Sot uk oe 178 36. Tentatively referred specimens of Claenodon ferox (Cope) from the Melville (Fort Union No. 3): Left Mg, right My, left M»-3 and heel of Maarighibye stan dlehtyMcsen ke Ses res ee Se ee le 180 37. Skull and jaws, left side, of Claenodon montanensis (Gidley) ----------- 183 38. Skull of Claenodon montanensis (Gidley), palatal view__-..----------- 183 39. Foot bones of Claenodon montanensis (Gidley) ..-.------------------ 184 40. Left upper jaw of Claenodon silberlingi (Gidley) --------------------- 186 41. Right lower jaw of Claenodon latidens (Gidley) --------------------- 188 42. Left M2-3 of Claenodon vecordensis Simpson__..-.------------------ 189 A Samet MV nnOtee CLOCNOUONISD sate eee ay tet ee ee ee ee 189 44, Right M?-3 of Deuterogonodon montanus (Gidley) ------------------- 191 45. Right lower jaw of Prothryptacodon furens Simpson-_---------------- 196 46. Right lower jaw of Chriacus pugnaz Simpson_—----------------=---- 196 47. Left Mi-3 of Metachriacus punitor Simpson_-_-_--------------------- 198 48. Right lower jaw of Metachriacus punitor Simpson------------------ 198 49. Weft M'>* of Metachriacus, punitor Simpson. = --———_ =. = 198 50. Left lower jaw of Metachriacus provocator Simpson----------------- 202 51. Left M2-3 of Metachriacus provocator Simpson-_--_------------------- 202 52. Left M!— of Metachriacus provocator Simpson..------~------------- 202 53. Left lower jaw of Spanozyodon latrunculus Simpson-_---------------- 204 54. Left lower jaw of Mimotricentes latidens (Gidley) ---_--------------- 204 55. Left lower jaw of Mimotricentes angustidens Simpson-_-------------- 206 56. Left lower jaw of Didymictis microlestes Simpson------------------- 211 57. Left P?-M?2 of Didymictis microlestes Simpson -....----------------- 211 58a duett lower jaw of Didymictis tenuts Simpson 222. - 2225-25-22 225552- 212 59. Referred specimens of Didymictis haydenianus Cope from the Lebo: Left upper jaw with carnassial and right upper carnassial_______---- 213 60. Right lower jaw of Ictidopappus mustelinus Simpson-_-_--------------- 214 61. Left upper jaw of Ictidopappus mustelinus Simpson------------------ 214 62. Right lower jaw of Hilipsodon aquilonius Simpson------------------ 235 63. Left upper jaw of Ellipsodon aquilonius Simpson-_-_----------------- 235 64. Right Ps-Ms3 and right lower jaws of Litaletes disjunctus Simpson----- 240 65. Right upper jaw of Litaletes disjunctus Simpson----------- Bi ae dee 240 66. Left lower jaw of Litomylus dissentaneus Simpson ------------------ 242 67. Right M'-3 of Litomylus dissentaneus Simpson--------------------- 242 68. Right lower jaw of Haplaletes disceptatrix Simpson_---------------- 244 69. Right upper jaw of Haplaletes disceptatrix Simpson-_-_--------------- 244 70. Right lower jaw (with M;) of Tetraclaenodon symbolicus Gidley - ------- 247 71. Right Mi-,and right M?-3 of Tetraclaenodon symbolicus Gidley ------ -- 247 72. Left upper jaw of Gidleyina montanensis (Gidley) ------------------ 252 73. Left lower jaw of ?Gidleyina silberlingi (Gidley) .------------------- 254 74. Left lower jaw of ?Gidleyina superior (Simpson) -------------------- 255 75. Left lower jaw of Coriphagus montanus Douglass--_----------------- 260 76. Left upper jaw of Coriphagus montanus Douglass_------------------ 261 77. Right lower jaw of referred specimen of Anisonchus sectorius (Cope) fromunesWebOsea = see 8 fae eke ae ee a= ae ee eae eS 261 78. Right upper jaw of referred specimen of Anisonchus sectorius (Cope) franc aesepotams eee eS Ne BP oe Sse ee nos 261 79. Histogram of length of M; in Anisonchus sectorius (Cope): Torrejon andelebospecimens. 252504 0522 cans Soa ee eset eee ste obec se 263 80. Right lower jaw of Pantolambda intermedius Simpson---------------- 270 Hm CO bo 10. BULLETIN 169, UNITED STATES NATIONAL MUSEUM PLATES Following . Reconnaissance map of the mammal-bearing Fort Union group of the Crazy Mountain Field, Wheatland and Sweetgrass Counties, Mont - - . Bear and Lebo formations, Crazy Mountain Field, Mont__------_-_-- . Air views of Fort Union group, Crazy Mountain Field, Mont_-------- . Air views of Gidley Quarry, Fort Union group, Crazy Mountain Field, . Air view and site of the Silberling Quarry, Fort Union group, Crazy IMO um bari ie) eI ra Ge a creas ee epee ee ee rn . Melville formation, Crazy Mountain Field, Mont_-_--.-_------------ . Fort Union primates: Palaechthon alticuspis Gidley and Paromomys mauris Gidley, richt lower jawsies 2622-5 22 aaane ee ee . Fort Union primates: Pronothodectes matthewi Gidley and Paromomys maturus Gidley Meitlower IaWSsno2 scones o seach eee ee ee eee Fort Union primates: Plesiadapis gidleyi (Matthew), Pronothodectes matthewi Gidley, Plesiadapis rex (Gidley), Palaechthon alticuspis Gidley, Paromomys depressidens Gidley, and undetermined species, (GE3 0H GFK) 0 Weapere che eels Doel erties Ea A 8 eR mete a ee SPE a Fort Union primates: Palenochtha minor (Gidley), Elphidotarsius florencae Gidley, and Pronothodectes matthewi Gidley, lower jaws----- page THE FORT UNION OF THE CRAZY MOUNTAIN FIELD, MONTANA, AND ITS MAMMALIAN FAUNAS By Grorce GayLorp Simpson American Museum of Natural History, New York City INTRODUCTION Tu1s work is chiefly devoted to the description and discussion of a large collection of Paleocene mammals, from the Fort Union of central Montana, belonging to the United States National Museum. The first part of the memoir is geological, chiefly stratigraphic and paleontological. The location and general characteristics of the mammal-bearing area are described, followed by a résumé of its stratigraphy and geologic Structure. The general areal geology is only briefly outlined, and details are largely confined to the beds in which the mammals occur. In the paleontological section, the various fossil localities are listed, and the mammalian fauna of each is given, together with a summary of all fossil mammals found in the field. Faunal succession and faunal correlation, supplementing the remarks made in the previous section on more purely stratigraphic correlation, are discussed in detail. The general aspect and ecologic relation- ships of the mammalian faunas are also discussed. Nonmammalian fossils are incidentally mentioned, without any attempt at exhaustive treatment. The second and longer part of the memoir is zoological and is devoted to definitions of all taxonomic groups which were originally described from this field or the conception of which is markedly affected by material from here, to discussion of the relationships and phylogeny of the mammals represented, and to description of the specimens in the collection. Geology outside the area of reference and mammals other than those definitely identified here are mentioned only for the sake of comparison. In addition to the National Museum collection, specimens from this field now at Princeton University are included. Reference is also made to a small collection in the Carnegie Museum, but since this includes nothing not better represented in the National Museum collection, none of it is treated in detail. Some, but not all, of a large collection in the American Museum of Natural History from 1 2, BULLETIN 169, UNITED STATES NATIONAL MUSEUM this same field is also mentioned or discussed. In a few cases classi- fication has demanded some detail concerning fossils from other col- lections and areas, especially the San Juan Basin, in New Mexico, notably in dealing with the genera Claenodon and Anisonchus. Most of the new genera and species recognized in the course of this work were named, with brief diagnoses, in a preliminary note abstracted from the first draft of part 2 (Simpson, 1935d). These diagnoses are here repeated, in some cases in emended form, and ac- companied by the extended discussion and description omitted in the preliminary paper. There is included a full discussion of the fossil Primates, even though these have already been published in some detail (Gidley, 1923). In the course of preparation of this bulletin, I have had access to and have made constant use of a memoir on the Paleocene of the San Juan Basin by the late Dr. W. D. Matthew. This has now been published by the American Philosophical Society, but it was available only in manuscript throughout the course of my work. Reference to it is made by the citation ‘“(Pale. Mem.).” Other citations are by author and year and refer to the ‘Literature cited” at the end. As far as possible without destroying the comprehensive and unified character of this work, I have avoided any duplication of material contained in Matthew’s memoir. For groups that he also discusses I have generally omitted diagnoses and have confined discussion to a brief summary and to those points wherein I differ from him. Throughout this work, wherever they proved useful, statistical methods have been employed. These are all summed up in Fisher (1925) and also in a paper soon to be published (Simpson and Roe). The methodology is outlined here in connection with the first group treated, the Multituberculata. In all the statistical figures given, the stated error is standard, not probable. The following abbrevi- ations are consistently used throughout for brevity, along with the universally understood symbols for teeth. L= Length. W = Width. N=Number of specimens in a given sample. R= Kange. M= Mean. o=Standard deviation. d/o = Deviation from a specified mean, divided by the corresponding stand- ard deviation. V=Coeflicient of variability —< P= Probability. } [as defined t=A constant for comparing small samples lby Fisher. =(d?) =Sum of the squares of deviations from the mean. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 3 In the systematic zoological part of this work, it has not seemed necessary in each case to give the evidence or arguments for associa- tions of upper and lower dentitions. With very few exceptions these collocations are not based on actual association of upper and lower jaw remains of one individual, which is very rare in this field. It is based rather on occlusion, occurrence at the same localities, relative abundance, comparison with related forms from other fields, and similar indirect but usually conclusive criteria. In the few cases where there is any serious doubt about the association, this fact is mentioned. This manuscript was completed on March 15, 1936. Slight changes: have since been introduced, but no later general revision has been made, and with few exceptions statements made are to be taken as of that date. HISTORY OF THIS STUDY AND ACKNOWLEDGMENTS Work for this memoir was started in 1908, when Dr. T. W. Stanton commissioned Albert C. Silberling to collect Fort Union fossils for the United States Geological Survey and the United States National Museum. In 1909 Dr. J. W. Gidley visited the field, where he di- rected and planned further work by Silberling, and he also then under- took the laboratory and office work on the collection. In view of the extremely fragmentary and few fossil mammals that had been collected in the Fort Union up to 1908 and the general scarcity of mammals of comparable age from any formation, the magnitude to which this work would grow could not have been anticipated when it was started. In 1911, when Mr. Silberling finished his most inten- sive work on this collection (to which he also added periodically until 1932), he had brought together one of the largest collections of Paleo- cene mammals ever made. Furthermore, this collection consisted almost entirely of new species, more than half the genera were new, and it greatly extended morphological and distributional knowledge of the families and orders represented. The collection represented, potentially, the greatest single contribution to knowledge of early mammals that had ever been made. The very magnitude of the results achieved was embarrassing. All these hundreds of specimens had to be prepared, the majority of them by tediously working off the matrix grain by grain under a microscope. Concurrently with many other duties, this was undertaken by Dr. Gidley, and it occupied much of his time over a period of 12 years. Then the identification of the material presented great diffi- culties, as it must in any Paleocene collection, for on one hand all the genera and species are clearly variable and on the other the really significant characters are often in such small details that it may be difficult to distinguish forms properly classified in different orders. 4 BULLETIN 169, UNITED STATES NATIONAL MUSEUM These difficulties were enhanced by the fact that the most pertinent comparative material was in another institution and that Dr. Gidley’s other duties did not permit his spending the months, or even years, of comparative study necessary under these circumstances, except by short visits or the loan of a few specimens as opportunity presented. In spite of the- really tremendous amount of work that he had accomplished on the collection, Dr. Gidley was able to complete only a relatively small part of the final preparation and publication of manuscript before his death on September 26, 1931. Up to the end, he looked forward to the completion of the work, and remarked, in conversation, that the collection was the most important ever in his hands and that its publication would be his greatest contribution to science and his most enduring monument. This it is, despite the fact that he was not spared to complete it with his own hands. In 1932, Dr. Alexander Wetmore and C. W. Gilmore invited me to undertake the completion of this study. The officers of the American Museum of Natural History permitted the use of my time, as a coop- erative undertaking with the United States National Museum. The whole collection was shipped to New York, where it could be studied under the best possible circumstances and compared at first hand with almost all the other types of American Paleocene mammals. Knowl- edge of the field, and further accessions to the collections, were made possible by work with Mr. Silberling in Montana for the National Museum in 1932 and for the American Museum in 1935. The great extent of Dr. Gidley’s contribution to this work should be explicitly stated.! In the first place, the existence of this splendid collection is in large part due to him. He collected some of the best material, and he directed and encouraged the collection of most of it. Second, he prepared and cataloged the whole collection so that it came to me in almost perfect condition for immediate study. Third, he prepared and published four preliminary papers (as listed in the bibliography and discussed in the proper places in the text). It is inevitable that some differences in point of view and more particularly the lapse of time make complete agreement impossible, but his pre- liminary work greatly facilitated study of the groups involved. Fourth, a number of unpublished illustrations had been prepared under Dr. Gidley’s direction, and most of them appear in this publi- cation. Finally, he left a few notes on the unpublished parts of the collection. There is reason to believe that Dr. Gidley had the major outlines of the classification of the collection, and probably also many of its details, wellin mind. This preliminary orientation is one of the long- est and most difficult parts of research, but unfortunately Dr. Gidley did not find it necessary for his own use to reduce it to writing and 1 For a review of Dr. Gidley’s life and work see Lull (1932). FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 5 could not foresee how important this would have been for his suc- cessors. The available notes are thus fewer and less useful than was anticipated. Specifically, they are as follows: 1. A field notebook containing a list of 43 specimens collected by Gidley, with locality, horizon, and date, but no other data. 2. Illustrations and proofs of his paper on the Primates, with no unpublished data. 3. Notes on multituberculates and claenodonts, with no unpublished observations. 4. Twenty-seven pages of notes and of manuscript in various stages of preparation, with considerable duplication and different drafts of treatments of the same subject. These, the only unpublished original observations left by Dr. Gidley, include brief preliminary diagnoses of three new species of Tetraclaenodon, one of Protogonodon, one of Mio- claenus, one of Tricentes, one of Mixodectes, and one of a genus probably considered as new but not named or defined, and also a sketch diag- nosis of a new genus and species of phenacodonts (here called Gidleyina montanensis). As far as they can be deduced from these notes, I have mentioned Dr. Gidley’s opinions in the present text. In some cases they warrant detailed discussion and quotation. In others, his notes were clearly of the most preliminary sort and would surely have been modified before publication, and in these cases it has seemed unjust to do more than mention them briefly. As regards the recognition of species, I first studied the collection independently and then ascertained whether any new species (or genera) recognized by me were antici- pated in Dr. Gidley’s notes. If they were, I credited them to Gidley, ex ms.”, and quoted sufficient of his diagnosis to establish his author- ship. The species so recognized are Deuterogonodon montanus, Mimo- tricentes latidens, Tetraclaenodon symbolicus, and Gidleyina montanen- sis. Dr. Gidley’s notes also include diagnosis of the genus Gidleyina but under a preoccupied name. Some other new forms are recognized in the notes, but without a name or without a diagnosis or definite indication, so that Dr. Gidley could not technically be established as their author, but his recognition of them is mentioned. In a few cases I have been unable to agree that a form tentatively designated as new by Dr. Gidley is so, and then have pointed out this fact but have omitted his names in order not to create useless synonymy. None of Dr. Gidley’s notes were in such shape that it would have been just to him to publish them without revision, and in any event only a small part of this memoir is affected by his unpublished notes. "Gioue tatarnntional Rules of Zoological Nomenclature, Article 21, state that the author ofa scientific name is he who first publishes it with a valid definition ‘‘unless it is clear from the contents of the publication that some other person is responsible for said name and its indication, definition, or description.’”’ This validates Gidley’s authorship of these species published by me. 6 BULLETIN 169, UNITED STATES NATIONAL MUSEUM That this work has been written at all is largely due to Dr. Gidley, but he is not responsible for its errors or for any matter in it not explicitly stated to be his. It is further desired to acknowledge how much this work owes to Albert C. Silberling, whose contributions to it go far beyond those usual from a collector. He discovered most of the fossil localities, including all those of real importance, and made far the greatest part of the collection, with skill, persistence, and devotion that cannot be too highly praised. He established a system of field records of the greatest accuracy, used throughout this publication. Even beyond this he so thoroughly examined the field and so carefully studied and correctly interpreted its geology that much of what is here written about it, and of what has appeared in publications by others, is merely reducing to writing observations made by or with him. Heis thusina sense an author and authority of the geologic part of this work, although he has not actually written any of it and is not responsible for misinterpretation of his views or observations. Aside from the long periods when he was definitely employed as a collector, Mr. Silberling has spent every spare moment for the past 35 years working in this field, and this memoir is in a real sense the outcome of this lifetime vocation and avocation. The following illustrations were drawn by Rudolph Weber under Dr. Gidley’s direction: Figures 22, 30-34, 37-41, 44b and b’, 59, 63a, 71a and a’, 72, 736, 80.2 The other drawings were made by Sydney Prentice, under my direction. The accompanying map is based on field work by Silberling and me and was drawn by John C. Germann. The field photographs were taken by me. The late Prof. W. J. Sinclair, of Princeton University, lent and per- mitted the description of the specimens collected for that institution by Douglass and by parties under Farr. Dr. G. L. Jepsen has facili- tated comparisons with specimens collected under his leadership for Princeton in the Fort Union of Wyoming. The Carnegie Museum, through J. J. Burke, lent material in their collection from this field. I am much indebted to these institutions and colleagues, as well as to officials of the United States National Museum, particularly Dr. Wet- more and Mr. Gilmore, for their support of the work and constant aid and encouragement, and of the American Museum of Natural History for the time and facilities provided. PREVIOUS WORK Discovery of Paleocene* mammals lagged behind that of typical faunas of any later epoch, or even of the Mesozoic. This slowness of discovery and the factors of scarcity of fossils and difficulties of study 3 In a few cases Prentice has modified these drawings slightly. «“Paleocene” is granted to be an epoch of the Tertiary, including post-Lance or post-Danian and pre- Gray Bull or pre-Sparnacian time. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 7 to which it was due have made the subject of Paleocene mammals an obscure one, generally avoided, but they have also made it a particu- larly fruitful field for research in recent years. The first Paleocene mammalian fauna to be discovered was that of the Thanetian, or particularly of its subdivision the Cernaysian, in France. Arctocyon primaevus was described by Blainville in 1841, but knowledge of the fauna really dates from Lemoine’s publications beginning in 1878. It has only recently been revised and placed on a satisfactory basis by Teilhard (1916, 1921). The American Puerco formation of New Mexico was named in 1875 but yielded no mammals until 1881, when the famous collector David Baldwin began a long collecting campaign there for Cope. This work has been followed, since 1892, by a series of expeditions to this field under Wortman, Granger, Sinclair, Simpson, and others for the American Museum of Natural History, which also acquired the Cope- Baldwin collection. This stratigraphic sequence in the San Juan Basin of New Mexico has become the standard of comparison for the Paleocene of the world, and its faunas are far the best known and represented by the most nearly perfect specimens (although in variety they do not exceed those to be described in this memoir). They have been described by Cope, Osborn, Earle, Wortman, Matthew, Granger, Simpson, and others and are thoroughly revised in a memoir by Matthew (Pale. Mem.), published while this bulletin was in press. Three quite distinct formations and faunas have been recognized, Puerco, Torrejon, and Tiffany, the first two each with two well-marked separate faunal zones of different facies but nearly the same age. The Fort Union group was originally defined by Meek and Hayden (1861) as occupying “‘the country around Fort Union,® extending north into the British possessions to unknown distances; also southward to Fort Clark. . . . Seen under the White River group on North Platte River above Fort Laramie. Also on west side of the Wind River Mountains. . . . and also occupying extensive areas of country in Nebraska . . . and beneath the White River group at several distant localities.” Although the designation of the type locality leaves no doubt as to the inclusion of certain strata in the group or as to its general position in the scale, the name has been used in many different ways, and in keeping with the spirit of the original description it has been applied to any or all strata at about this part of the geologic section, that is, in what we now call Paleocene, over a very large area in the Northwest. Thanks to this widespread occurrence or use of the name, and to the economic value of these rocks, especially as a coal- bearing series, few geologic formations have been more intensively and extensively studied. No general review of this work is here attempted 5 Near the present site of Buford, N. Dak. 11921237 9 _ 8 BULLETIN 169, UNITED STATES NATIONAL MUSEUM (see résumé and references in Thom and Dobbin, 1924) aside from the Crazy Mountain area and briefer mention of mammal discoveries elsewhere. The earliest geological work in the particular area here under dis- cussion appears to be that of W. H. Weed. In 1893 he mentioned the presence of about 4,000 feet of Fort Union sediments overlying his Livingston formation east of the Crazy Mountains. In the Livingston (Weed, 1894) and Little Belt Mountains (Weed, 1899) folios, he in- cluded the extreme western part of this area and referred to the presence of Fort Union here but mapped it all as Livingston. In an article on the Fort Union (Weed, 1896) he also discussed the Fort Union area in a general way and gave a section of part of these and of lower beds along Lebo Creek. The first more detailed consideration of this area and the first ade- quate reference to its more important northeastern part was by Earl Douglass. In 1900 he worked here, principally in the Cretaceous but also in the Fort Union, where he collected leaves and invertebrates. In 1901 a Princeton University expedition was sent out under Dr. M.S. Farr, and Douglass accompanied this party and also continued work after they had left. The Princeton party traversed the Fort Union exposures, but most of their work of that season was in the Cretaceous. In August 1901, Douglass found the first mammals in this field, and the first ever discovered in the Fort Union, at two localities (5 and 6 of the lists given on a later page) near the northern end of Bear Butte. Although fragmentary and few, these sufficed to show the equivalence of these beds to the Torrejon of New Mexico (Douglass, 1902a). Douglass soon published detailed descriptions of the mammals and a discussion of the geology of the whole area (Doug- lass, 1902b). The Princeton party was also accompanied by Albert Silberling, then a boy of 16, who was destined to play the major role in the development of this field. Princeton field parties, also under Dr. Farr and accompanied by Silberling, revisited the area in 1902 and 1903, and they then examined the Fort Union in more detail and found fossil mammals at a number of scattered localities and at widely different levels, from near the base of the formation to about 4,350 feet above the base and from the northeastern part of the field to the vicinity of Cayuse Butte, near the western margin of the principal mammal-bearing area. The fossils collected and observations made in 1902 and 1903 have never been published. In 1905 Douglass made a long reconnaissance through Montana, and in the course of this he went northward from Bigtimber to Cayuse (or Melville) Butte. He noted the presence of the Fort Union over much of this traverse but did not add significantly to exact knowledge of the field. Apparently he had not been informed of the unpublished FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 9 discoveries of 1902 and 1903 on and around Cayuse Butte, and he found no mammals on his brief visit. His account of this trip (Doug- lass, 1909) includes a résumé of the geology of the area, involving also his observations in 1900 and 1901. In the meantime Silberling had continued prospecting the field and had located the two richest known deposits, later developed as the Silberling and Gidley Quarries. He was for a time associated with the Carnegie Museum, to which Douglass had now also gone, and he sent in to it a small but important collection, principally from the Silberling Quarry. This was described by Douglass a few years later (Douglass, 1908) and was the basis of the first exact definition of mammals from this field, including the types of Ptilodus montanus, Picrodus silber- lingi, Coriphagus montanus, and Megopterna minuta,’ as well as many less exactly identifiable specimens. In 1907 (see Stone, 1909) R. W. Stone worked in the northern part of this area for the United States Geological Survey, and in 1908 W. R. Calvert worked south of the area. They connected their ob- servations west of the Crazy Mountains, and in 1909 also by recon- naissance east of the mountains and later published an important discussion of the Fort Union here and of its relations to the Livingston, applying the name Lebo to the lower part of the Fort Union (Stone and Calvert, 1910). In 1908 Dr. T. W. Stanton visited the field, where he was accom- panied by Silberling, who had in the meantime left the Carnegie Museum and started ranching in this area. They made detailed observations, including a section of the Lebo which I give on a later page, and Silberling was employed under Stanton’s direction to collect for the United States Geological Survey and National Museum (Stanton, 1909). Knowlton had taken some interest in observations here, principally in connection with his placing of the Hell Creek and synchronous beds in the ‘Lower Fort Union” (see Knowlton, 1909), a view now universally abandoned and requiring no consideration here. In 1909, Stanton, Stone, Calvert, Knowlton, and M. R. Campbell briefly visited the area again to check certain critical localities, the results of their observations being principally given by Stone and Calvert (1910). Stanton (1914) and Knowlton (1914) later returned to their argument regarding the Cretaceous—Tertiary boundary, but with only incidental reference to this area. Silberling continued the collecting begun in 1908 and spent much time in 1909, 1910, and 1911 collecting for the National Museum. It was in these years that he obtained most of the specimens described in this work. A few specimens collected in later years were also acquired by the National Museum, and Silberling has continually spent much time in the field, even when not commissioned by any 6 Which is, however, invalid. 10 BULLETIN 169, UNITED STATES NATIONAL MUSEUM particular institution. In 1908 he had worked principally in the Silberling Quarry. In 1909 Gidley visited the field and then opened the Gidley Quarry at a site, Loc. 4, previously discovered by Silberling. Gidley himself collected many fine specimens here, and the work was continued by Silberling in 1909-11, and later, along with repeated prospecting of surface localities. In 1911 Gidley made another visit to the field.’ In 1909 Gidley described the splendid skull, jaws, and partial skele- ton of Ptilodus that had been found in 1908 by Silberling in the Silber- ling Quarry. This is still the best multituberculate specimen in any museum, and the best single mammal specimen ever found in this field. Gidley continued the preparation of the collection, work done entirely by himself, as time and other duties permitted. This was completed in 1920. In the meantime two preliminary papers had been published, one on A/yrmecoboides (Gidley, 1915) and one on the clae- nodonts (Gidley, 1919). After the whole collection was prepared he began monographing it, but only the section on the Primates (Gidley, 1923) was completed or published. The extent of his unpublished work has already been noted. In 1930 Silberling made a renewed examination of the field and also reopened the Gidley Quarry and made a collection that was purchased by the American Museum. In 1932, under the auspices of the Na- tional Museum (with the donation of my services by the American Museum), Silberling and I went over the whole area and adjacent regions, with the present work in mind. We then made the map (pl. 1) that accompanies this memoir and also made detailed strati- eraphic observations. A small amount of material was collected, incidental to our visits to all the fossil localities, but no intensive collecting or quarrying was undertaken. In 1935 the Third Scarritt Expedition of the American Museum, consisting of Mr. Silberling, a camp man, and me for the entire season, and Mr. and Mrs. Fenley Hunter, Dr. Walter Granger, and Albert Thomson for shorter periods, spent four months in this field, pros- pecting most of the surface localities, reopening the Gidley and Silber- ling Quarries, and developing a new quarry, named the Scarritt Quarry. The resulting collection, about equal in size to that here described, is in the American Museum. The material from the Scar- ritt Quarry has been described (Simpson, 1936b), and the results are here included in the general sections but not in the detailed descrip- tions. The surface material, which is relatively abundant and im- portant but does not include any species not also present in the National Museum collection, has also been identified and is mentioned where apropos in the present work. The very large Gidley Quarry 7 Gidley’s statement (1923, p. 1) that he visited the field in 1908 and 1909 is probably a misprint, for it is contrary to Silberling’s memory and to letters and other records that seem to show that his visits were in 1909 and 1911. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 11 collection, and the few specimens from the Silberling Quarry, in the American Museum are not yet studied and do not enter into this study. For the most part they duplicate the collection described here. Of other mammal discoveries in the Fort Union, the first of any im- portance was made by J. F. Lobdell in 1926 in a coal mine at Bear Creek, Mont. Collections were later made here for the Carnegie Museum and the American Museum and have been described by me (Simpson, 1928a, 1929a, b). The small but interesting fauna is of Upper Paleocene age, about equivalent to the Tiffany. Discoveries in northern Wyoming, west of the Bighorn Mountains, have been of outstanding value. Sinclair and Granger (1911, 1912; also Granger, 1914) had found a fauna of Paleocene aspect here in beds which had previously been considered to be true Eocene and which they named Clark Fork. In 1927-29, Dr. G. L. Jepsen, work- ing for Princeton University, found three distinct faunal horizons in (nominal) Fort Union strata beneath the Clark Fork in this area. He has shown that these correspond in age to the Puerco, Torrejon, and Tiffany, and hence has for the first time established a definite sequence of four distinguishable Paleocene mammalian faunas in a single continuous stratigraphic section (Jepsen, 1930). The faunas so far described are small but typical. Subsequent collecting under Jepsen at the same localities has yielded much more and better material, but the results have not yet been published. Sporadic discoveries of one or two specimens have been made at a few other Fort Union localities, but none is of much importance. The type of Titanoides primaevus came from near the type locality of the group, Buford, N. Dak. (Gidley, 1917). Typically Middle Paleocene forms, Yetraclaenodon and Pantolambda, were found in Billings County, N. Dak. (Lloyd and Hares, 1915). In Fort Union or Kingsbury beds of the Bighorus a jaw identified by Gidley as Tricentes was found (Stanton, 1909, p. 268). Eocene fossils were found in supposed Fort Union beds in the Powder River Basin (Wege- mann, 1917). The Jast-mentioned fossils, and perhaps that from the Kingsbury, are not really from the Paleocene, or from true Fort Union. The Paskapoo of Alberta, which may be considered in a general way a northern extension or equivalent of at least the upper part of the Fort Union, has yielded a few mammals, the first found by Brown in 1910 (Simpson, 1927), with later discoveries mostly by Russell (1926, 1929, 1932), all extremely fragmentary. Recently Patterson, working for the Field Museum of Natural History, collected fine skele- tons of Barylambda, an ally of Titanoides, at an Upper Paleocene level in the Plateau Valley formation of western Colorado (Patterson, 1933, 1934, 1935, 1937). PART 1: GEOLOGY AND FAUNAS GEOGRAPHY The region here primarily considered is comprised in Tps. 3-6 N. and Rs. 13-16 E. (Montana principal meridian), an area of about 575 square miles. It is approximately included in latitude 45°55’ to 46°20’ N., longitude 109°40’ to 110°10’ W., and is in central Montana, Tps. 3-5 N. in Sweetgrass County and T. 6 N. in Wheatland County. The principal commercial center is Harlowton, seat of Wheatland County, 8 miles due north of the designated area. This town, popula- tion about 1,500, is on the north side of the Musselshell River and is a division point on the main line of the Chicago, Milwaukee, St. Paul, and Pacific Railroad. About an equal distance south of the area is Bigtimber, population 1,200, on the south bank of the Yellowstone River. The only settlement in the limited region here primarily treated is Melville, an old town once of some importance but now reduced to little more than a post office and crossroads store. It is in secs. 1l and 14, T.4 N., R. 14 E. There are main east-west motor highways and railroads in the Musselshell and Yellowstone Valleys, north and south of this area, but just here east and west through travel is blocked by the impassable mountains immediately to the west. There is, however, a secondary north-south highway from Harlowton to Bigtimber through the area, and this is now being improved and graveled (1935). There are several county roads, graded but unsurfaced, as shown on the map, and also many private roads and wagon trails, which permit approach to within a mile or less of any point in the field. Much of the land has been plowed for dry-land wheat raising, but a great deal of it was not suitable for this purpose and so most of the wheat farms have been abandoned. A little wheat and some hay or alfalfa are still raised, but most of the region is given over to grazing, mainly as winter range for sheep and some cattle. The population is sparse and in recent years has not been prosperous. There are a few dude ranches in and near the mountains in the western part of the field. There are no known mineral resources. Topographically the area lies immediately east of the Crazy Moun- tains, a very rugged and beautiful isolated range 25 or 30 miles long, north to south, and 10 to 15 miles in width, rising to 11,178 feet in Crazy. Peak. The area specifically treated extends from the foothills of the mountains eastward about 25 miles. The divide between the Musselshell and Yellowstone Rivers runs through the field. The northern part is drained mostly by Fish Creek, here running eastward, 12 21AWN0I39 ES SE pees See = 1 PORCUPINE BUTTE RISE TO BIG ELK CR EEK SECTION ALONG LINE R.I4E R.ASE | TO HARLOWTON — = Ay 1B: SCALE tg2000 Eioee ° 1 2 (Horizontal and Ve i ELEVATION WHEATLAND CO. bee a aa co. iat a 43342 RISE 3 SCALE IN MILES at TO BIGTIMBER RI4E RISE RECONNAISSANCE MAP OF THE MAMMAL BEARING FORT UNION GROUP OF THE CRAZY MOUNTAIN FIELD WHEATLAND AND SWEETGRASS COUNTIES MONTANA XX MAMMAL LOCALITY FORMATION CONTACTS eae ¢ wi singer ug : CRAZY MOUNTAIN. FIELD ue — AWD oon orale COUNTIES | ; _—r ne _ —— / Z FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 13 and its southern part mostly by Sweetgrass and Otter Creeks and their tributaries, here running mainly southeastward. The extreme northwestern part is drained by American Fork and Lebo Creeks, tributaries of the Musselshell, and the extreme southwestern part by Bigtimber Creek, tributary to the Yellowstone. Lebo and Fish Creeks rise in this area and carry little water in ordinary weather, while the other streams mentioned rise in the mountains and have larger and more nearly permanent flow. Apart from the actual mountain area and its flanking zone of pediments, moraines, and outwash, here poorly developed, this region is a hilly one with mature topography almost entirely determined by the hardness and structure of the underlying rocks. The harder sandstones form ridges and scarps, and the softer beds are eroded into valleys or flats. The most striking elevations in the northeastern part of the field, like Bear Butte and Lion Butte, are sandstone- capped erosion forms. Cayuse Butte, a prominent but relatively low and irregular mass, and Porcupine Butte, the most prominent eleva- tion outside the mountains, also are supported by intrusive igneous rocks (as are the mountains). The southern part of the field is hilly but has no prominent topographic features. There are three small lakes in the area, Lebo Lake in the northwest and Lakes Adam and Wal- voord (more generally known as the Glass Lindsay Lakes) in the southeast, all artificial and developed for irrigation. The highest point outside the mountains is Porcupine Butte, 6,970 feet, and the lowest is on Fish Creek where it leaves this area, about 4,000 feet. Climate and vegetation vary from mountain to high-plains types. The mountain flanks have forest of lodgepole pine and other conifers, with aspens along the streams and some colorful meadows, while the main part of the field is very sparsely timbered, with evergreens on the higher sandstone ridges and cottonwoods in wet valley bottoms. The broad intervening areas, except for the relatively small irrigated patches, are characterized by sagebrush, pricklypears, small yuccas, and grass. Abandoned plowed land is taken over by weeds, especially Russian thistles. In spite of the rather sparse nature of the vegeta- tion, almost the entire area is sodded over. Shale exposures are small and relatively few, and there are no badlands. The topography along the northern edge of the area (beyond the Fort Union exposures) is very different, for here there are large, well- developed, high, gravel-covered terraces cutting across the oes Cretaceous beds without regard for their hardness. ag This general Fort Union area has received various ae hole such as ‘‘Fish Creek”’, ““Bear Butte’’, “Sweetgrass County”’, or ‘“‘Mel- ville.” Fish Creek ie a minor stream draining less than half of the region especially considered. Bear Butte is a small and marginal 14 BULLETIN 169, UNITED STATES NATIONAL MUSEUM topographic feature. Sweetgrass County does not include some of the most important localities, and does include a vast area, much of it south of the Yellowstone, far outside that here treated. Melville is an unimportant settlement local to one small part of the field. This whole region, a great synclinal area with Fort Union rocks filling it, may be called the Crazy Mountain region, for these mountains occupy its approximate center and are the most prominent topographic feature in this part of the State, and the major structure has already been called the Crazy Mountain Syncline. The smaller area desig- nated and discussed above is, in general, the eastern half of the Crazy Mountain region, and is that where the Crazy Mountain Fort Union is best and most clearly developed and where it has yielded mammals. This area will herein be called the Crazy Mountain Field. GEOLOGY STRATIGRAPHY GENERAL STRATIGRAPHIC COLUMN Rocks exposed in the area between the Musselshell and Yellowstone Rivers and east of the Crazy Mountains are from Lower Cretaceous to Recent in age. Aside from terrace gravels, moraine and outwash deposits, valley fill, and other unconsolidated young deposits, not to be considered here, the sedimentary rocks are Cretaceous and Paleocene, as far as definitely established. ‘There are numerous igneous intru- sions, all younger than the Paleocene sediments, and likewise omitted from this discussion. The whole stratigraphic column, including some rocks not exposed in the area of the map (pl. 1) but all within a few miles of these and undoubtedly underlying this area, may be summed up in a general way as shown in the column on page 15. From exposures outside this area, especially in the Big Snowy and Belt uplifts, it appears that the sedimentary rocks here exposed are probably underlain by many thousands of feet of earlier Mesozoic, Paleozoic, and pre-Cambrian sediments, probably one of the thickest piles of sediments in the world. The oldest surface beds of this region are exposed only in the centers of domes north of the mapped area, for instance in sec. 34, T. 7 N., R. 16 EK. Dr. Barnum Brown has recently obtained some interesting dinosaurs from this locality. ‘There is some question whether these beds are closer to the Kootenai or to the approximately equivalent Cloverley, but here they have generally been called Kootenai. The series here called ‘Undifferentiated Colorado” is probably susceptible to definite subdivision and correlation, but this has not yet been clearly accomplished and is outside the scope of the present study. The lower part was given the local name ‘‘American Fork”’ by Douglass (1909), who called the upper, marine, beds ‘‘Fort Benton”’, FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 15 Dark shales, greenish or No. 3 | Melville gray, with numerous gray | 5,000+ feet. to yellow sandstones. Mammals, ete. Fort Soft somber greenish sandy Union No. 2 shales and gray sand- | 850 feet. stones. Mammals, ete. Lebo rr Somber shales with some No. 1 hard brown sandstones. | 500 feet. Mammals, etc. euore Alternating pale cross-bed- nion ded sandstones and shales. (or Bear Rare turtle bones, ete. 500-600 feet. Lance) No dinosaurs or mammals. Pale, variegated clays with Hell Creek some gray sandstone. Di- | 2,000 feet. nosaurs. Brown and yellow sandstone and some somber shale. Lennep Transitional marine-fresh- 250-450 feet. water. Soft dark shales. Marine. Bearpaw (Also some dinosaurs.) 700-1,100 feet. : Cross-bedded gray sand- Judith River stone and sandy shale. | 400-800 feet. Dinosaurs. Brownish or yellowish thin- bedded sandstones and Claggett shales. Marine to brack- 400-800 feet. ish invertebrates. Three prominent coarse 100-250 feet. Bogle white sandstones. Coal. Upper part soft dark shales, locally sandy. Lower aloe part with much _ thin- =o (Undifferentiated) bedded and ripple-marked | 1,300—2,000 feet. sandstone. Upper part marine, lower with dino- saurs, turtles, and plants. Variegated red, green, and brown shales and sand- Kootenai stones. Dinosaurs and | 200+ feet (base 5 D:¢ e nonmarine mollusks. not exposed) probably an inexact correlation. Probably equivalents of the Mowry, Thermopolis, Frontier, and some other Wyoming formations are included. A prominent but local sandstone member has been called “Big Elk.” Douglass at first (1902) called the overlying beds ‘‘Niobrara’’, but their more exact correspondence with the Eagle and Claggett is 16 BULLETIN 169, UNITED STATES NATIONAL MUSEUM now wellestablished. In 1902 he applied the name “Fish Creek Beds” to the next higher division but later accepted their equivalence to the Judith River as established by Stanton and Hatcher (1905). Douglass (1909) continued to use the name “Fish Creek” for beds transi- tional between Judith River and Bearpaw, but this name is not now in general use. Lennep is a local name applied by Stone and Calvert to beds perhaps equivalent to the Fox Hills but not definitely correlated. In physical character they resemble the Lebo, and seem to gradeinto the Livingston laterally. CRETACEOUS—TERTIARY TRANSITION It is not proposed to discuss here the general evidence for drawing the boundary between the Cretaceous and the Tertiary, a problem to which this field contributes only one small part, and not the most important part. For present purposes it is assumed that beds up to and including the true dinosaur-bearing Lance and Hell Creek and their equivalents belong in the Cretaceous and that overlying beds without dinosaurs (except by redeposition) and with mammals of Tertiary type (including carnivores, condylarths, etc.), from the Puerco and its equivalents upward, are to be placed in the Tertiary. It is also assumed that the Paleocene is accepted as a separate epoch of the Tertiary, of equal rank with the Eocene, Oligocene, Miocene, and Pliocene, and that its lower limit is taken as the base of the Puerco or equivalent, or the top of the (restricted) Lance or Hell Creek or their equivalents, and its upper limit at the top of the Clark Fork, or equivalent, or base of the Sand Coulee and Gray Bull, or equivalents. On these assumptions, which I believe to be based on the most reasonable and useful interpretation of all the evidence, it is here pro- posed only to discuss briefly the beds in this field that must include the transition from Cretaceous to Tertiary as thus defined and to suggest where the boundary may occur in this local series. The following is a somewhat generalized section from the Hell Creek into the Fort Union No. 1 in sec. 26, T.6 N., R. 16 E., passing through Loc. 65 and near Loc. 78, about 2 miles northeast of the northern end of Bear Butte: Soe a ce eee Hori zoniof@ioces (8). oan ee eee Mostly greenish shales and sandy shales________-_---------- 30 Somber cross-bedded sandstone, the upper part bard and ridge- Fort Rormmiinig 1S 8 ER IE ee CEE 40 Union Somber greenish shale and sandy shale____--- -------------- 85 No. 1. |Same, with bone fragments including unidentifiable mammals of Mertiary aspect; boc: 652222 .2 25 Shee ees ae ee ee ee 7 name; withoubdossus 2-2. joe Pe a ee 35 FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 17 Zone of hard brown concretions, with numerous fresh-water in- vertebrates and some turtle bones___.___________-_-_---_-- +1 eee Transitional beds with shales somewhat darker than those below and with local lenses of brown sandstone_________-_-_-_-- 15 Alternating white cross-bedded sandstones and pale shales__about 485 about 500 Hell : : Gray Softer beds, pale sandstone and clay, with dinosaurs._. (Not measured) Definition of the new name ‘‘Bear’’ is given on page 20. All these beds are tilted here, but there is no evidence of angular unconformity. The contact beneath the lowest hard sandstone of the Bear is sharp but not more so than beneath other sandstones through- out this series. Such a contact could, but does not necessarily, represent a parallel or erosional disconformity. The change from Hell Creek to Bear is, however, rather more abrupt than from Bear to Fort Union No. 1, the latter being transitional through a thickness of 15 to 20 feet, the boundary here taken by convention at a local shell lens. Dinosaurs are found in place, apparently as originally deposited, at almost all levels in the Hell Creek, although no very good specimens have been found in this field. JI have seen no dinosaurs in the Bear, but Mr. Silberling informs me that he has found isolated, rare, and very fragmentary specimens in the lower part of that formation, the highest being 80 feet above the base in this section. It is entirely possible that these few fragments were redeposited and derived from the Hell Creek. No dinosaurs have been found in the upper 420 feet of the Bear, but a few champsosaur and turtle bones and some in- vertebrates occur. Invertebrates from the lens here considered as marking the top of the Bear have been identified by Dr. L. S. Russell as follows: ?Fusconaia danae (Meek and Hayden). | Viviparus formosus Meek. ?Nedionidus senectus (White). Campeloma nebrascense whitei Russell. Elliptio priscus (Meek and Hayden). Campeloma limnaeiforme (Meek and Viviparus trochiformis (Meek and Hay-| Hayden). den). Physa cf. canadensis Whiteaves. Dr. Russell remarks that the three pelecypods range widely in the Northwest, Judith River to Fort Union, and that the species of Vivi- parus characterize the Fort Union and equivalents (listing mainly Upper Paleocene levels), the two Campelomas are likewise from the Fort Union or equivalents, and the Physa is of little significance. He concludes that the ‘fauna contains nothing characteristic of the Lance and includes several species restricted to the Fort Union. It can be regarded with some confidence as Paleocene in age.”’ 18 BULLETIN 169, UNITED STATES NATIONAL MUSEUM As already noted, there are no exactly identifiable mammals from Loc. 65, but the fragments found here are of Tertiary rather than Cretaceous aspect, and the mammals from Loc. 78 are certainly Paleocene and probably Middle Paleocene. It seems probable that the whole Fort Union No. 1 here is of Middle Paleocene age and all but certain that it does not include the earliest Paleocene. There are two possible interpretations of this section. First, the usual view, in analogous cases, that the Bear is Cretaceous, equivalent to part of the Lance, with the necessary corollary here that part of the Paleocene, a Puerco equivalent, is probably lacking. This in- terpretation is possible, but it seems to me decidedly the less likely of the two. The invertebrates indicate Paleocene and probably not earliest Paleocene. Even if it be decided that they should be con- sidered basal Fort Union rather than uppermost Bear, they are inti- mately associated with the latter and in beds that grade into it in- sensibly. This gradation itself opposes this interpretation, for it makes the presence of an unconformity here, or a gap representing Puerco time, seem unlikely. If a break exists it is more likely to be within or below the Bear. The absence of dinosaurs, with the pres- ence of such reptiles as did survive into the Tertiary, in most of the considerable thickness of the Bear is also evidence against this view, negative but of some weight. The second possible interpretation is that most of or all the Bear belongs in the Paleocene, probably representing the Lower Paleocene, and that the Cretaceous—Tertiary boundary is in its basal part or below it. All the items of evidence mentioned above favor this view. Op- posing it are the possible presence of dinosaur bones in the basal Bear and the absence of mammals. The dinosaur fragments are, however, such as could very readily be redeposited from erosion of the Hell Creek, and they are confined to the lower 80 feet of a series about 500 feet thick. It is entirely possible that there is a disconformity, or at least a valid epoch boundary, in this series at the base of a sandstone above the level of these bones, and still more likely that the dinosaur bones are not really original fossils in these beds. The absence of mammals has no weight. Mammals are rare throughout the field as a whole, are generally excessively rare in the Lower Paleocene, in this field almost never occur in sandstone, which is all that is well exposed here, and even if present would be small forms easily overlooked. The absence of dinosaurs, however, may well be significant, because their bones are large, occur in most terrestrial Cretaceous formations, and are very likely to occur where, as here, there are remains of other types of reptiles. Another summary section through the whole Hell Creek and Bear was taken in sec. 15, T.6 N., R.15 E. This is continuous with the lower Fort Union section of Stanton and Silberling given elsewhere. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 19 ae ees Baserorororbauniont Now) “alec wi aerrag 8 led vote Clays with thin and inconstant sandstones, a more prominent sandstone at the top not well exposed here_______________ 280 Bear {Gray to buff platy sandstones in beds usually one to three feet thick, with thicker shale partings_____.........2_L_ 222 L- 330 610 Pale variegated shale with occasional soft blocky sandstones_ 765 Aer rane RATIOS CONC! atc) eee Re Eee ea 10 Mainly pale variegated sandstone___.___.......1._________- 270 Hell Heavy gray sandstone, with some shale________________.__- 145 Creek | Pale white to greenish clays with brown concretionary layers, sandstones increasingly prominent toward the top________- 660 1, 850 fy oil a hie B* Sp Base oie Creek » f= oe ee ee Yellow Sandstone and Somber Clay______________- (Not measured) Dinosaurs are found in the Hell Creek in this region also, but no animal fossils were found in the Bear. The discrepancy in thickness of the Bear between this section and that previously given, about 110 feet, seems too great to be due entirely to difficulty of accurate measurement and is probably at least in part a real difference. The localities are about 8 miles apart in a straight line. It is also possible that the limits taken do not exactly correspond in the two sections, as all the formations concerned are highly variable. Knowlton (in Stone and Calvert, 1910, p. 749) reports the following leaves from “200 feet below the top of the Lance Formation”’, that is, in the Bear at this locality: Sapindus affinis, Sapindus grandifoliolus, Platanus aceroides, Platanus sp.?; and the following from a level still lower by 200 feet, probably still in the Bear: Sapindus ?grandifoliolus, Sapindus sp., Platanus raynoldsii. Knowlton unreservedly called these Fort Union, but this has no bearing on the question here con- sidered, since he also called true Hell Creek and Lance floras Fort Union. All these species are reported from beds probably of Lance age, but they are all equally characteristic, or considerably more abundant, in the true Fort Union. Throughout the northern part of the field at least, and generally except where all the subdivisions of this part of the section tend to merge indistinguishably into the Livingston, the beds between the Hell Creek and the Fort Union No. 1, or basal Lebo, are a good lithologic unit. The scanty data also suggest a distinctive paleon- tological character. These facts and the possibility, or as I believe probability, that the beds should be classified as Paleocene, or asso- ciated with the Fort Union, rather than as Cretaceous and associated with the Hell Creek or ‘‘Lance” make it highly advisable to distinguish 20 BULLETIN 169, UNITED STATES NATIONAL MUSEUM them by a separate name, and I have here called them ‘‘Bear’’,® from Bear Butte around which they are typically developed. They may perhaps be equivalent, approximately, to the Tullock, but this would be an assumption that might result in serious misapprehension. They are far removed from and discontinuous with the type Tullock, into which it will never be possible to trace them, and their lithologic character is not the same. It is improbable that they are exactly equivalent to the Tullock, and even if this should prove to be the case it would seem warranted and necessary to retain for them a local name. At present correlation with the Tullock would be only a guess, which might well prove to be incorrect. Typical exposures of the Bear are shown in plate 2. Fort UNION The use of the name ‘‘Fort Union” has been so loose and ill-defined that it has become necessary for every writer who uses it to propose his own individual definition or to run the risk of being completely misunderstood. In general it has been applied to beds in the Dakotas, Montana, and Wyoming, in the northwestern high-plains region (and in part intermontane areas) that are, or are supposed to be, later than the Lance and earlier than the Wasatch. This apparently satisfac- tory definition is in fact most indefinite. In the first place, there has not always been general agreement even to this extent, Knowlton, for instance, placing beds generally referred to or correlated with the Lance in the ‘‘Lower Fort Union.” In the second place, the definition is dependent on that of Lance and of Wasatch, which are themselves very ill-defined. It is certain in some areas (notably Polecat Bench in northern Wyoming, as shown by Jepsen) and probable in most or all that strata generally referred to the Lance, often under the name of Tullock but not necessarily equivalent to the type Tullock, are in reality distinctly later than the typical Lance or the equivalent Hell Creek and both faunally and stratigraphically are more nearly related to the overlying beds, that is, to the Fort Union. Thus even aside from the question of accurate recognition of the boundaries and correlation of members of these formations and groups, there is often doubt as to which group should include a given member. The time- honored name ‘‘Wasatch”’ is still more ambiguous, to the point that very few of the beds called ‘“‘Wasatch” are really equivalent to any part of the type Wasatch. Granted that the use of the name is usually intended to imply approximate correlation with the Gray Bull, there still remain many doubts as to its proper contents, for there is generally a thick series of beds, Tiffany, Clark Fork, and so on, that are some- "Tan ndevied to Miss M. Grace Wilmarth, of the U. S. Geological Survey, for informing me that ‘‘Bear’’ and also “Melville’””. (defined on a Jater page) are not included in her records as ever having been used previously as the names of stratigraphic units. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 21 times included in ‘‘Wasatch’’, sometimes in Fort Union, and sometimes are separated from both as an intervening stage. As a matter of personal opinion, I would prefer to use “Fort Union group” for the whole Paleocene series of this northwestern plains. area, to include all strata of age equivalent to or intermediate between the Puerco and the Clark Fork. This usage is very broad, but the designation ‘‘Fort Union” has already been so loosely applied that a more restricted usage would be very difficult to frame or to maintain. It at least has the virtue of being on the whole a natural subdivision of the Tertiary and of including practically all the beds that have ever been called Fort Union, except in out-and-out error or in such wholly untenable classifications as that of Knowlton. For more precise work it will in any case be necessary to define and use more local names for particular formations and members included in the Fort Union group. In this field, the earliest workers recognized as Fort Union only the beds from the basal No. 3 sandstone (as defined on a later page) upward. Douglass, Stone and Calvert, Stanton, and all later workers, however, have also included the underlying andesitic beds called Lebo by Stone and Calvert. The name “Fort Union” is used throughout the present study for the Lebo and all higher early Tertiary strata in the area here considered. ‘The uppermost part of this series is of unknown age and may possibly be as young as the Gray Bull, in which case it should be removed from the Fort Union, but at present no evidence warrants this step. In addition there is a series of strata, the Bear, hitherto always considered as Cretaceous and referred to the Lance, but in my opinion possibly Tertiary. If it should prove to be Tertiary, I would place it in the Fort Union, but this is now doubtful, and in this study the word ‘‘Fort Union” is not intended to include the Bear. The local Fort Union, thus defined, includes three mappable litho- logic units of very unequal thickness. Mr. Silberling, who first recog- nized these units, has applied numbers to them, with Fort Union No. 1 at the base and No. 3 at the top, and his field designations have been employed in publications by Stanton, Osborn, Gidley, and others. In accordance with the general rules of stratigraphic nomenclature, local geographic names are here applied, but throughout this discus- sion I shall also employ Silberling’s numerical designations. The correspondence is as follows: No. 3 = Melville (new name). Fort Union; No. 2 ap led No. 1} _ 3 The No. 1 and No. 2 beds are generally similar, and both are included in the Lebo of Stone and Calvert, but they are easily sepa- 22 BULLETIN 169, UNITED STATES NATIONAL MUSEUM rable throughout the field, and their distinction makes discussion and records more exact. Both are said to be in large part andesitic, and both are characterized by their somber color, the shales gen- erally greenish and the sandstones dark brown, gray, or gray-green. The No. 1 beds, or lower Lebo (shown in pl. 2), are characterized by numerous lenses and beds of hard and resistant dark sandstone, so that this unit is generally topographically positive and forms a ridge or series of ridges. This characteristic is visible throughout the field; for instance, typically in the gentle anticline in the northern part of T. 5 N., Rs. 15-16 E., where the No. 1 forms an elevated hilly area surrounded by a horseshoe valley developed on the No. 2 beds. The upper limit of the No. 1 is taken at the top of the highest and most persistent of its hard sandstones. The thickness, as measured in secs. 15-22, T. 6 N., R. 15 E., is 496 feet (Stanton and Silberling). It has not elsewhere been accurately measured. It may be somewhat thicker in the western and thinner in the eastern parts of the field but apparently does not vary greatly. The No. 2 beds are topographically negative (see pl. 3). They form valleys between the No. 1 and No. 8 sandstones, or slopes beneath the latter. The characteristic material is greenish shale, often rather coarse and sandy, with lenses and beds of gray sandstone. When unweathered, these sandstones may be hard, for instance in the overburden of the Gidley Quarry, but they weather rapidly and are not resistant to erosion. It is this nonresistant nature of its sandstones, and generally their somewhat lighter color, that dis- tinguish this member most sharply from the No. 1. The thickness in the measured section mentioned above is 840 feet and probably averages 800 to 900 feet throughout the field. Stone and Calvert (1910, p. 753) give a total thickness of 2,080 feet for the Lebo on Lebo Creek. I did not measure the beds here (where they are not very well exposed and have a variable and uncertain dip) but estimated the thickness at not much over 1,500 feet, with about 600 feet in the No. 1 and 900 in the No. 2. They give a total thickness of the Lebo in T.6 N., R. 16 E. (that is, near the north end of Bear Butte) of only 463 feet, which I think is surely much too small. It is highly unlikely that this persistent formation thins out from 1,334 to 463 feet in less than 8 miles. The dip in this region changes rapidly, as the beds are around the Bear Butte syncline, and exposures are not continuous, so that exact measurement is not possible, but it is unlikely that the thickness is much if any less than 1,200 feet here, with some 700 or 800 in the No. 2 and 500 or 400 in the No. 1. The Lebo as a whole forms a narrow band along the northern edge of the field, from the northeastern end of the Crazy Mountains almost due east, but with some sinuosity, to sec. 23, T.6 N., R. 15 E. Here they turn abruptly southward, and their exposure widens greatly FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 23 under the influence of the Widdecombe Creek anticline, the axis of which is mostly on the lower Lebo, with the upper Lebo forming a horseshoe valley around it, the eastern limb being occupied by Widde- combe Creek. The exposures then swing around the northern end of Bear Butte (with the No. 2 also continuous through the gaps across the axis of the Bear Butte syncline) and then swing southward and widen until they form a very wide area of low relief in the vicinity of the Glass Lindsay lakes (Lake Adam and Lake Walvoord). The upper limit against the No. 3 swings abruptly westward in sec. 34, T. 4 N., R. 15 E., and runs over to the Crazy Mountains, which it reaches some 17 or 18 miles south of its northern departure from those moun- tains. The exposure as a whole thus forms a great loop closed on the west by the mountains, as shown in plate 1. The No. 1 and No. 2 beds seem to be perfectly conformable and to intergrade, forming a natural unit, sandier in its lower part, so that the single name Lebo for both is acceptable. In the northern part of the field they are generally sharply separable, but in the south it is often difficult to distinguish them. It is to be noted, however, that in this southern area they are nearly horizontal and are very poorly exposed and that in this direction all subdivisions, even between undoubted Cretaceous and Tertiary, tend to break down or to be very difficult to distinguish. The best exposed section showing the whole thickness of the Lebo is In secs. 15—-22,° R. 6 N., T. 15 E., where it has been measured by Stanton and Silberling who give the following section (Stanton, 1909, p. 263, here slightly modified): No. 3{ MASSIVE SATUSLOM Cae cetera cece eerie arn earn eae re aes See (Not measured) Feet Shale with a few thin bands of sandstone____________-_----_- 112 RGAE BAO SL OTIC B.S ee me Bey che a Coa ST al ly Ge ed 5 BSI URS San SF =, oe oa rs EE ter a pnp 18 Gray limestone, weathering brown [a concretionary layer, not a Norte) “continuous stratum] 2209? eeGeen ) of tie eek Ae ee 3 Greenish-gray somber shale with much soft sandstone of same color and brown .... concretions in lower third, a few thin bands of sandstone in upper part and several concretionary ZOWMEB TRC ATA UO [0 eye ee a ye a ee an 700 838 Brown, thinly cross-bedded sandstone forming... . ridge____- 32 Newd [Somibermereenisil Rs hyall ewer eeses ree em cee re mere 200 ‘er "Greenish-eray sandstanes 2) a4 ee EY tee a 15 Gréchish=pray elialess). 2252 eee _ Gers yt eee 249 496 A section of the underlying beds at this same locality is given on a previous page. 9 Stanton says sec. 15, but the greater part of the beds are here in the adjacent sec. 22. 119212—37 3 24 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Stone and Calvert and others have given other sections, but they are too generalized or include too many unexposed parts to be very helpful. The Fort Union No. 3 beds (see pl. 6) are a great mass at least 4,000 feet thick, and possibly as much as 6,000 if the highest strata of this series in the Crazy Mountains be included. Determinable fossils are known only from the lower 3,000 feet, and this is the part to which attention has been particularly directed. In spite of the great thick- ness and heterogeneity, there is no convenient lithologic or paleonto- logical basis for subdividing the beds. They consist of shales and sand- stones, most of them lenticular and highly variable. The usual topo- graphic expression is a series of ridges on the sandstones and valleys on the shales. The shales, particularly at the mammal localities, do not differ greatly from those of the No. 1 and No. 2 in appearance, but the sandstones are lighter in color and are generally more resistant than those of the No. 2. Since the shales are poorly exposed, the general impression is of a much paler formation than the underlying Lebo. The massive basal No. 3 sandstone is the most continuous and proms inent horizon marker in the Fort Union of this field and has been remarked by everyone who has worked here. It almost everywhere forms a prominent scarp, and its resistance to erosion is the cause of the elevation of Bear Butte, Lion Butte, and numerous less notable hills and escarpments. The base of this sandstone forms a natural division plane throughout the field and is evidently an erosional disconformity, although it is unlikely that it marks any considerable time gap. The sandstone tends to become less massive in the southern part of the field toward the south end of Lion Butte and where it swings westward to the mountains, and it is here generally more platy and formed by numerous thin beds with shale partings but can be traced continuously almost to the mountains. The fluviatile nature of these beds and the presence of numerous channel and flood sandstones make it possible that there are discon- formities at almost any level, but none can be detected as of any regional significance, and the whole series seems to be essentially continuous and without any noteworthy or sudden change in general type of sedimentation. The base of the No. 3 is further made noticeable, especially from the air, by almost everywhere supporting a growth of evergreens (pls. 3, 5). This is the more noteworthy because the sandstones of the No. 1 beds, even where almost identical with those of the No. 3 in topography and elevation, do not support any trees, and the Lebo in general is almost devoid of large vegetation except for cottonwoods along the stream courses. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 25 The No. 3 beds occupy an irregular area, somewhat more elongate east and west than north and south, bounded by the Crazy Mountains on the west and by the loop of Lebo exposures on north, east, and south (pl. 1). From their position over the Lebo, it may be supposed that the No. 3 beds correspond in their lower part to the Tongue River and perhaps in their upper part to the Sentinel Butte, but such a correla- tion is at present totally unwarranted. They are very distant from the typical or from any unquestionable exposures of the Tongue River or Sentinel Butte, their lithology is no more like either Tongue River or Sentinel Butte than like many other formations and certainly is not close enough, in itself, to warrant correlation without tracing them laterally continuously or nearly so into true Tongue River and Sentinel Butte, which is impossible. Their age is well established, in part, as brought out below under ‘‘Correlation’’, but that of typical Tongue River and Sentinel Butte is not, and paleontological correlation with those members is not now possible. Under these conditions it is certainly less helpful than conducive to serious error to adopt the frequent practice of assuming that a corre- lation exists. Still worse is the practice, also exemplified by some work on the Fort Union, of assuming that both Tongue River and Sentinel Butte must occur here and dividing the beds, in which no natural division has been established in the field, according to the proportionate thickness of these members in a widely different area. I therefore propose the local name Melville for the lithologic unit, from the town of Melville, which is situated on these beds and is surrounded, within a few miles, by excellent and typical exposures of them. The lower boundary of the formation is well established, as shown on the accompanying map. The upper boundary is still uncertain. The name is proposed to include at least the lower 3,000 feet of the No. 3 beds, to about the top of Cayuse Butte, and tenta- tively for the whole No. 3 of this field, to the highest sediments on Porcupine Butte. Later discoveries might make it advisable to remove some of these uppermost strata from the formation, and the pertinence of still higher beds around the mountain flanks is wholly dubious. Concretions are common throughout the Fort Union here. Limy concretions, weathering rusty yellow, may be of great size, up to 10 feet or more in greatest diameter, and locally characterize a definite stratum, but these appear to be truly concretionary and not a true sediment or limestone. Shell limestones do occur locally in the No. 3 beds, but in the No. 1 and No. 2, while shells may be fairly abundant in relatively small lenses, they are generally in a shale matrix and do not form a true limestone. In the No. 3 beds there are a few thin bands of comminuted shells, or shell breccia (notably at the Scarritt Quarry), generally mixed with clay. 26 BULLETIN 169, UNITED STATES NATIONAL MUSEUM The relatively small quantity of carbonaceous material is note- worthy in ail three subdivisions. There are a few thin impure and local lenses of coal in the No. 3, and a little prospecting has been done on these, but none are of any commercial value. Aside from these thin seams, generally only an inch or two thick, there are a few car- bonized tree trunks, occasional very local lenses of coal a few feet in diameter and less than an inch in thickness, and locally many minute carbon fragments and filaments. In marked contrast with the Fort Union of most other areas, these rocks can be classed as not coal- bearing. Several workers, especially Stone and Calvert and Silberling and I, have attempted to follow out the development of this Fort Union series into regions beyond the local field, but with indifferent success. To the eastward it ends against the Cretaceous, and the Fort Union reappears, considerably modified in thickness and character, after a long gap, in the Bull Mountain Field (see Woolsey, Richards, and Lupton, 1917; also Ellis and Meinzer, 1924). The lower part of the Fort Union is there predominantly a shale member, dark and greenish in color, with some sandstone (not closely similar to that of the No. 1 in the Crazy Mountain Field) and coal, about 200 to 300 feet in thickness. This is commonly correlated with the Lebo, which is probable on lithologic grounds, but there is no evidence that it repre- sents the whole Lebo or that it is not a lithologic facies of somewhat different span in time. The correlation is here more probable than in any other area where the Lebo is supposed to occur, but it cannot be considered as established beyond doubt. The upper part of the Fort Union is here only 1,650 feet thick and contains more pale shale, more and different limestone, more numerous and persistent hard sandstones above the base, and much more coal than does the No. 3 of the Crazy Mountain Field. The Bull Mountain Fort Union is an isolated mass, completely surrounded by Cretaceous outcrops. The Fort Union encircles the Crazy Mountains, which are formed principally by intrusive masses thrust into it, and has been followed in some detail by Stone and Calvert (1910). (Silberling and I have also made a reconnaissance around the mountains on which, as well as on Stone and Calvert, my remarks are based.) The Upper Creta- ceous formations and the Lebo, steeply folded and much disturbed and altered by later igneous activity, swing around and into the northeastern end of the mountains. The Hell Creek here appears to contain more andesitic debris and comes to resemble more closely the Lennep and the Lebo. Toward the southwest, and southward on the western flank of the Crazies, the Hell Creek, Bear, and Lebo have not been distinguished. It seems probable that they here merge litho- logically with the Livingston, as Stone and Calvert believe, or they may possibly be absent. There remains, however, the possibility FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 24 that more detailed study and discovery of fossils, especially verte- brates, would permit their differentiation here. West of the moun- tains the late Cretaceous and early Tertiary strata are thrown into a series of folds, roughly parallel to the mountain flank and with north- south trend. There is no wide zone of Fort Union, as east of the mountains, but strata lithologically similar to the Fort Union do occur in narrow zones, as folded. Some of them are almost identical in character with the Melville beds of the eastern flank and include abundant leaves. This Fort Union facies, overlying the Livingston, also occurs in the southern part of the mountain area, but it is much disturbed and poorly exposed and not enough work has been done to trace its continuity with the well-differentiated series in the north- eastern area. It is probable, as Stone and Calvert suppose, that it includes only the upper part of the Fort Union of the latter region. No mammals have been found in it.!° There is a large area litho- logically like the Melville around the southeastern flank of the mountains also, but here again exact correlation is impossible at present, and considerable search has not yet revealed any mammals or other fossils indicative of exact age. The Fort Union exposures are surrounded by older rocks throughout the northern and western areas, and no extension or correlation of the beds in these directions is possible. Following the beds southward, the division between the No. 1 and No. 2 tends to break down, and this is probably true also of that between the Hell Creek, Bear, and No. 1 Fort Union near the Yellow- stone River. The beds here take on a violet hue, especially the shales, a color also seen in the northern area but there rare and here predominant. The indivisible beds so colored apparently include the Lebo, probably the lower part of the Melville, and also extend down into the Cretaceous—that is, this part of the series is blending into the undivided Livingston, as reported by Stone and Calvert. There is a dark sandstone that may represent part of the No. 1 Fort Union skirting this area and visible near the 4,700-foot contour on the south of the divide between the Yellowstone and Stillwater Rivers, south of Reed Point, and also apparently at Absarokee, but the identification is uncertain. If this is the No. 1, it represents the farthest point to which a definite subdivision of the Crazy Mountain Fort Union can really be traced, even in this highly dubious fashion, as far as I have been able to ascertain. We were unable to establish any continuity with the Red Lodge or Polecat Bench areas, and lithologically the beds in those areas are not comparable to those of the Crazy Mountain Field, nor are similar subdivisions recognizable. 10 A bout 1910 a sheep herder showed Mr. Silberling a jaw of an animal similar to Pantolambda that he said was found near the head of Shields River, or between it and Potter Creek. It was impossible to obtain the specimen for exact identification or to verify its reported origin. 28 BULLETIN 169, UNITED STATES NATIONAL MUSEUM They include nothing really comparable with the Lebo, and the re- semblance to the Melville is too vague to have any real correlative value. Paleontological correlation alone seems to have any sig- nificance between this region and that near the Crazy Mountains. STRUCTURE The beds of this field are almost nowhere exactly horizontal and in places are steeply tilted. Even in the course of purely paleontological work it is necessary to take strict account of the structure, since relative levels between the numerous isolated exposures can seldom be measured directly but have to be calculated from the structural data. Along the north side of Fish Creek, from the mountains eastward to sec. 23, T. 6 N., R. 15 E., there is a strong monoclinal flexure dip- ping south at angles of about 40° to 75°. North of this fold only Cretaceous beds are exposed. The flexure itself involves the upper- most Cretaceous, Bear, Lebo, and basal Melville. The strike is sinuous but is mainly east and west to the end of the Crazy Mountains, where the Melville beds mostly run into that range, while the lower beds swing around its northern end. In the section noted, this fold ceases to affect the Fort Union beds, which are affected south and southeast of here by a broad anticline with northeast-southwest trend and low dips up to about 12°. Erosion along the axis of this anticline has formed a great embayment, about 6 miles wide at its mouth and of about the same depth, north and south, surrounded by a high rim on the basal Melville. Widdecombe Creek flows along the soft No. 2 zone on the eastern limb of the anticline. Directly east or southeast of this anticline is a parallel syncline the axis of which is occupied by remnants of the basal Melville sandstone rising abruptly above the No. 2 slopes and valleys. Puet Creek cuts deeply into this, separating the axial elevation into the long isolated Bear Butte to the northeast and Lion Butte, to the south and connected with the great mass of No. 3 beds extending westward to the mountains. The Hell Creek, Bear, and Lebo swing around the north end of Bear Butte, dipping toward it. The dip increases in intensity here away from the actual Butte, until in the upper Hell Creek northeast of the Butte it reaches about 30°. The Lebo flanks the long east scarp of Lion Butte with low dips, usually 4° or 5°, toward the latter, vari- able and affected by slight local disturbances. South of here, toward the Glass Lindsay Lakes, the Lebo is nearly horizontal, with dips up to 1° or 2°, erratic in direction but oftener to the west. Over the greater part of the Melville beds area, south of the strong monocline and west of the Widdecombe Creek anticline, the dips are prevailingly westward and fairly consistent at 4° or 5° over a large FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 29 area, to the western part of T. 5 N., R. 14 E. Here the wide pass between Porcupine Butte and Cayuse Butte is in the very broad, poorly marked syncline or basin. The beds around Porcupine Butte dip northeasterly into this area, although in the Butte itself the dip (about 5°) is nearly north. There are a few visible small faults with a throw of a few feet, such as one at the Silberling Quarry, but these are negligible in beds of such great thickness. No faults of sufficient importance to show on the map or to affect stratigraphic leveling significantly were detected. The deformation appears to have been entirely post-Fort Union, and nowhere in this region was any angular discordance detected between the various Cretaceous and Tertiary formations, even down into the Lower Cretaceous (Kootenai) in the domes north of this area. Itis areasonable assumption that the deformation was approx- imately contemporaneous with the post-Paleocene igneous intrusions. FOSSIL LOCALITIES AND FAUNAL LISTS GENERAL OCCURRENCE OF FOSSIL MAMMALS Fossil mammals, represented at least by material adequate to show its mammalian nature, have so far been found at 57 localities in this field, of which about 35 have yielded material sufficiently well pre- served for generic identification, and about 25 material specifically identifiable and of some real value in correlation and faunal studies. The great bulk of the collections comes, however, from the three locali- ties where quarries have been developed. Although some of the more resistant sandstones, such as those in the No. 1 beds or the basal sandstone of the No. 3 beds, are well ex- posed and form more or less continuous outcrops wherever they occur in the field, the finer and less resistant sandstones and the shales and clays are on the whole very poorly exposed. Bones are occasionally found in the sandstones, but they are there very rare and are generally of no value. Only one identifiable mammal has ever been found in a true sandstone in this field. The mammal localities are therefore almost entirely on the rarer shale exposures, which occur where coulees have cut the shale slopes or where the wind has developed blow-outs. Such exposures, seldom as much as a hundred yards in diameter and generally much less, are limited in number. In the productive area in this field there are probably not over 400 of them, and all of these have been prospected by Mr. Silberling, many of them also by me or others. The mammal occurrences may be grouped under two categories, as surface localities or as (actual or potential) quarries. At the surface localities, much the more numerous of the two, the mammal remains are rare, as far as known, and are so sparsely scattered through the matrix that only accidental finds or concentration from long weather- ing and wind erosion leads to any production. The ideal conditions 30 BULLETIN 169, UNITED STATES NATIONAL MUSEUM at such localities are deep weathering in situ without erosion or surface drift or wash, combined with gentle deflation, which removes the small weathered shale particles but leaves the larger or heavier fossils. Such conditions are relatively rare, and it is also noteworthy that numerous localities that were in good collecting condition when first found 30 years or so ago have now lost much of their value or promise by drifting over of windblown material or surface wash, by the spread of vegetation onto their weathered surfaces, or by active erosion, which removes the surface concentrates and leaves only a clean, hard shale exposure in which there is little chance of finding a fossil in situ. The quarry localities are those where fossils are so concentrated in a local pocket that it is profitable to work the bed as a whole and recover fossils in place. The difference is, of course, of degree and not of kind but is one of great practical importance. Were it not for its three principal quarries, which have been called the Gidley, Silberling, and Scarritt Quarries, this field would be of relatively little importance. There are marginal localities also of some importance, intermediate between quarries and straight surface localities. Thus Loc. 81 (as listed below) was literally a quarry, but the whole bone pocket and therefore quarry was only about a yard in diameter. Loc. 25 has yielded so much surface material that a concentration of probable quarry calibre is indicated, but the rather limited attempts made to quarry at that locality have not in fact developed a profitable bone layer. An outstanding characteristic of the field, regardless of level or geographic position, is the fragmentary nature of the material. In the hundreds of specimens collected, there are so far known only four or five mammal specimens complete enough to be called skulls, and only two of these really adequately reveal most of the skull structure. Only about 10 specimens include associated upper and lower teeth, and only three any surely associated limb bones. Nothing approach- ing a complete skeleton has ever been found. This fragmentary con- dition is seen not only in the surface specimens but equally in those found in situ deep in the quarries. The quarry specimens commonly show fresh breaks that look recent and yet abut against undisturbed matrix. It is also peculiar that most of the quarry specimens had lost some of their teeth before burial and that jaws quite devoid of teeth are relatively abundant. For his own records and in connection with the National Museum collecting, Mr. Silberling has numbered every locality where any fossils were found. These serial numbers are here adopted and are those used throughout the present work. There are now 82 numbered localities; 25 of these are not mammal localities, but for completeness and the convenience of later workers they are all given in the serial list on a later page. The map (pl. 1), however, shows only mammal localities. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 31 THE GIDLEY AND SILBERLING QUARRIES The greater part of the National Museum collection, about four- fifths of the identifiable specimens, is from the Gidley and Silberling Quarries, of which the Gidley Quarry is far the more important, with nearly seven times as many specimens as the Silberling Quarry. The Silberling Quarry, Loc. 1, was located, as a surface prospect, by Mr. Silberling in 1902, and he subsequently made a small collection here that formed the basis of Douglass’ publication in 1908. In 1908 Silberling opened the quarry for the National Museum and then collected most of the specimens known from this locality. In 1909 the much richer and more easily worked Gidley Quarry drew his attention away from the Silberling Quarry, and little work has since been done there. Silberling has done some further prospecting, however, and in 1935 the Third Scarritt Expedition reopened the quarry and worked it for a few days, but abandoned it on finding the bed bone poor, the stripping very difficult, the fauna essentially duplicating that of the Gidley Quarry, and the matrix unsatisfactory from the point of view of preparation. This quarry did, however, produce the splendid Ptilodus skull and partial skeleton described by Gidley, and it is the only locality in this field where Psittacothervum or Elpidophorus minor have been found. All its other genera and species are represented by about equally good or better material from other localities. The Silberling Quarry is in NESW sec. 4, T.5 N., R. 16 E., in an embayment near the middle of the east side of Bear Butte (see pl. 5). The fossil horizon is about 75 feet below the base of the No. 3 sandstone and is the highest level in the No. 2 that has yielded identifiable mammals. The bone layer is not well defined by any visible criteria but is limited to a zone 1 or 1% feet in thickness. The matrix is a fine greenish tuff or shale, very tough and harsh, extremely abrasive to handle, and difficult to work in preparation. It grades laterally into a bed with numerous fresh-water bivalves, among which mammals also occur, but this matrix is so hard that preparation of fragile specimens would be almost impossible. The Gidley Quarry, Loc. 4, is in NWNEY sec. 25, T. 5 N., R. 15 E. (see pl. 4). It is immediately adjacent to the county road, on its east side, where it descends the steep hill from the basal No. 3 rimrock to the valley of the upper part of Widdecombe Creek, on the relatively unresistant and topographically negative No. 2 beds. A small coulee here descends the slope, in a westerly direction, and cuts the bone bed apparently near the middle of the rich pocket in which the Gidley Quarry is developed. The locality was discovered by Mr. Silberling as a surface prospect in this coulee in 1905. When Dr. Gidley visited the field in 1909, Mr. Silberling showed him this locality as the most promising of any in the field. The surface 32 BULLETIN 169, UNITED STATES NATIONAL MUSEUM material was unusually abundant but so disintegrated as to have little value. Dr. Gidley dug in along the bone stratum until firmer material was found in place, made a good collection from this bed, and thus started the quarrying operations that Silberling and others have continued intermittently ever since. Mr. Silberling’s opera- tions for the National Museum, resulting in the bulk of the collection here described, consisted in stripping and working out an area of about 1,400 square feet (a very rough estimate). The other important operation, that of the Third Scarritt Expedition in 1935, extended over an area of about 1,750 square feet (also a rough approximation, as the area was irregular and not all uncovered at once). The quarry consists of two cuts, one north and one, of smaller size, south of the coulee and original surface exposure. The material collected by Silberling was carefully labeled as to the cut from which it came, but in this study there was found to be no significant differ- ence between the two parts of the collection, and it is all treated as a unit. The greater part of the worked-out area has been filled in again in later stripping, and the cuts left open weather and fill rapidly, so that the form of the quarry is not apparent in the field. The north section has probably been about worked out, except for a probably very rich corner left under heavy overburden. The south section was still rich along the margin as left in 1935 and probably would produce over an area of at least 1,000 square feet, and possibly 1,500, before the bone layer ran out into the hillside, so that the locality as left in 1935 is, as far as such things are predictable, still capable of producing another collection about as large as either of the two se far made there. The quarry has so far produced about 800 good identifiable mammal specimens and perhaps 1,500 single teeth and other unimportant fragments. The areal distribution of the fossils is very erratic and patchy. In places it is possible to work for 1 or 2 days without finding any jaws, and in others one man can collect ten or more jaws in a day. In general, however, the fossils seem to be distributed in an elongated area, about 5 to 20 feet in width, usually nearer the smaller figure, and with a tested length of at least 150 feet, which probably will continue to a much greater distance. The general trend is north- northeast to south-southwest. In many places the fossils are concentrated in a single and fairly well defined layer an inch thick or even less in which fragments of bone may be so numerous as nearly to make a bone-bed. ‘This material is, however, very fragmentary, and good jaws are exceptional in it. In other places the bone layer is less definite, and the fossils are scattered more sparsely but generally in better preservation through a thickness of about a foot, or up to about 18 inches at most. Oc- casional fragments are found in the nearly barren material above and FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 33 below this more definite fossil horizon but (as far as observed) only within a few inches of it. The level is about 125 or 150 feet below the base of the No. 3 sandstone. Some additional remarks on the occurrence of fossils here are made in discussing the facies and ecology of the quarry fauna. Table 1 gives a complete list of the fossils so far identified from the Gidley and Silberling Quarries, with the number of upper and of lower jaws of each species, based on the National Museum collection only except Elpidophorus minor, the one specimen of which is in the Princeton Collection. It is probably, but not quite surely, from the Silberling Quarry. TaBLe 1.—Fossil mammals so far identified from the Gidley and Silberling Quarries, Crazy Mountain Field, Mont. | Gidley Quarry Silberling Quarry Species Upper Lower Upper Lower jaws jaws jaws jaws MULTITUBERCULATA: Ptilodontidae: PATEL OGIESH IL OTILON) 1G tees en a en ee a ee 39 25 6 ll Pe tOUUSIQOUGUISS awa o ee a ae Re ee ne = | a eae See g LA ates ae eae 1 CETOMUSIGIIEYT== ae oon a he Need Eeael eee ees O92 nen eee 3 CAPE ENTITES Oe Hy a a LE ee eee ae oe eee neeee 143) 2 eee 3 4 JEST TONG] ays oe fel ses Be ae RSs A Ne be ela nap ee OA Ps Se ee (ea 1 PEREUDOUUISIOTON TCT see ke eee ees ese em. Pk eer! Ls 5 PEUCLYU DOGUSMUSSEL te nee 2 See eee ee aes | eee Pj Wns See ee 1 PEE DOUUS CILOCTUNG aoe eee ee a en ee ee | 1 PIOTECY DOLUS I ODSCN tease a PR nen eee eS eo FEUCOSTLORON ES DOAT SIS ha ae ete ee ee ee ee od ei ee eee 1 INSECTIVORA: ? Deltatheridiidae: Gelastops Pancuess= tJ Sk eles see ie oh ie 28 1 Rs eee 1 Leptictidae: PF OdICOMOTE CONCONULAT CEN StS see eae ea en nae eae eeee |e ee eee | 1 SRE DUILCOCOTINLG UM Canam ane e mee e eaa oa aee e as SE de 13 EDLACONOTUINAUTLSCULU Maan Oe oe eae a a seo no aoe sae ceo 1 Myrmecoboides, montanensis............-.--.------------ 1 2 Nyctitheriidae: ISHLDNOGON: SUNUDLICILENS -- aan aco oe ee teen oe eee cea asa ceee 1 Pantolestidae: PBESSOLCELOTECIUUCILL =e mee ee ree eee ee eee 2 sl 1 ZAM NTON OT1L8 fF OLA LON ae ee ee 8 1 | aes ee ee 3 Mixodectidae: HALA EMONEMA CUSPIAN OG sas eee een eae ene ceases 1 TlLgos. ee 1 PE DIAC DROTUS NINO 22 2oe = a= ee ee es See Ae ae oe |S oa oe selon eee eae 1 Picrodontidae: Picrodirs-sil benlingyt se ae es ee tes ss) d/o eee ee ty) cee es 1 PRIMATES: Plesiadapidae: iPronothodectesimattheiwt. 22 = tees as eae 3 eee 1 8 Carpolestidae: EAC OLUT SUIS LOT ENC On a= aoe ee ee ae ee ee 1 ? Anaptomorphidae: IPATOMOMYS MAUL US xa cee oa ae = ee ee 6 32 1 PUTONLONIGS CE DT CSSILEN Se a eee ee ee 3 4 ‘Palaechthomalticicepige: se) ieee Bie be rae ee 2 ee ee 1 PROLENOCHUR AWOL aoe oe aoe = 2 ee a ee 2 6 34 BULLETIN 169, UNITED STATES NATIONAL MUSEUM TABLE 1.—Fossil mammals so far identified from the Gidley and Silberling Quarries, Crazy Mountain Field, Mont.—Continued Gidley Quarry Silberling Quarry Species Upper Lower Upper Lower jaws jaws jaws jaws TAENIODONTA: Stylinodontidae: (COMONUCLESICOIM Nid eon ae 9 eee et eee 18 Pes eee 1 POStULOCOLRETIRLTI TIL LILI UTI ee as ee ee eee 1 CARNIVORA: Arctocyonidae: Clacnodonumontanensis 2212 2 sae ee 1 2 Claenodonisilberlingi=se es ee ee ee ee | 1 Claenodon:latidens24o% Us ER Se INGE OLINGH < fae eee 15 20 res INGE Of INR ee rs 13 eee | ee 2k ER ee See SS 16 77) SINAN) GEST fee Be 34 9A} eae Near middle of east half- 8 ami ae st INIWROLOINI Woe = 2 222 2 2h meee TS WrOlObes- oo sweet 35 2G eres | Re ee es 31 Pe sa INIEOR Shee ee ce eee 30 S28 Sees INIWHOfS\Wessees eee 29 Vs ae ING Win OfINIWiese nee 29 SOs == INGW ena eee ene es 30 ie ee eee 26 SUA es Sel Ree ee eae Cee 25 ahaa ee S\Wik oes ee ees ee 17 Sib en 2e |e eee ee See 36 BH eee INA Oy COS HIS) De eee eee 33 OOnse es Near center of north line_ 20 | 3 ee INDO IS De eee es eee 33 | ACs Se 8 SIM OfNiWee ee eee 17 3}! Eee SIP Gt Shite ene ae eee 15 40 SAWVMOLINIW see e a ee 29 CT SINOfNiWettee ae ee 29 Ab INGROLN Wiese ee aoe ee 20 CRW eee INIWVOLANI Wenn =e ana 20 |! Town- ship North oom ao oo ome On ons oon Oo AAAMWMD FP WON oro on Feet above base of Fort Union No. 1 (In No. 2)-_- (In No. 2)--- 1,200. = 2-22 —2 (In No. 2, a few feet be- low Loc. 3) (In No. 3)-.- (In No. 3)_-- (In No. 3)_-- (In No. 3)--- (In No. 3)--- KOOTENAI_-_ (In No. 3)-_-- Remarks SILBERLING QUARRY. “Various.”’ Ses faunal list. GIDLEY QUARRY. See faunal list. Do. (Not relocated—not a fossil locality). See faunal list. Do. Mammal fragments. See faunal list. Do. Do. Mammal and other frag- ments. See faunal list. Invertebrates and mammal fragments. Invertebrates. See faunal list. Mammal and other frag- ments. Invertebrates and Clcenodon (fide Silberling). (Not relocated.) brates. Plants. Footprints, plants. See faunal list. Do. (Not relocated.) Plants. See faunal list. Do. Invertebrates. Champsosaur and mammal fragments. (Not relocated.) brates. Do. Invertebrates. (Not relocated.) Plants. Plants. See “‘Princeton Localities.” Plants, invertebrates. See ‘‘Princeton Localities.” See faunal list. See ‘‘Princeton Localities.” Do. Do. Do. Inverte- Inverte+ 44 BULLETIN 169, UNITED STATES NATIONAL MUSEUM TABLE 2.—Serial list of localities, Crazy Mountain Field, Mont.—Continued No. Quarter 7 ee SH chiS Mulan soaks oe Yu el SE of SW of SW-------- 46. 22-23 8) NAW Of Ni Wels eee ce ese A fest ces © INGEVOTONID) -=----22 228522 ARE so INI eae oe eae eae CAS paps pag ies NGWiotN Wee. ee 50a Near middle of south half. Biles see SiWeweeeone ose sasewes Hotere eA INIWi OLN Be 2s0025282222= ies NW of NW of NW----- patra INSWOfsN Rie. eseecss- bosesese= INI Wess eet aat eee DOesee os SIRSOLN IW. 2 ae oe oe eee iiyfeee oe SiIWronSie e222 22 32 Lia ee Near center line in north half. S9l A Wastnaltse. 2. ae k 60S Se SE of NW of SW...----- (ee See ett) Fe erecta 622222222 SiWokSiW 222222252458 63a ING Wp eat os See ae (yee SIWik ete 2 eee (fe ee SEtopSiw--=----ses5->=- 662-5 SIWe eissae 22 eee ces (Gla see SiBoLS Bi... sete se2S 68oeann Near center of north line- (es SiWirie tof ee 2 eee Ree Omen Near middle of east half_ (ieee Se MOR TOtS Mice soc see ose ce (pe ee Near middle of north half. ome INGE OLS Wi22 i eseseeces (4 sets INJBVOLIN IN 22 Soe sence ae (pee SiWite 2 oe nee eee Ol. Saas SWaot(Sib 222 S-- ae sane Tesaee INiBiseset. see aoes yee SIWNOLS Wie = a- ese e-se2- ih eee A Ti) DORE ea pee et SOLES: ING ae eons or Se Slyeeey Near middle line in south half. S22e res SW of SW of SW------- Town- Section! ship ae North a 5 15 17 5 15 24 5 14 10 5 14 25 5 14 23 5 14 9 3 15 24 4 15 23 5 15 14 4 15 23 5 15 29 5 15 13 5 14 V7, 5 15 20 5 15 20 5 15 19 5 15 1 5 14 7 5 15 18 5 15) 34 6 16 26 6 16 14 5 13 10 5 13 20 5 15 18 5 15 31 6 15 36 6 14 12 5 14 34 6 16 15 5 15 6 4 16 11 5 15 32 5 15 26 6 16 10 5 13 14 5 13 23 6 15 11 4 15 Feet above base of Fort Union No. 1 GngNios3)2=2 (In No. 8)--- @mn.No:'3) 22. 2,500 See O;B00 22825 52= (In No. 3)--- (In No. 3)--- (In No. 3)--- 1,350 22 es25=5 (Lower part of No. 2). (Lower part of No. 2). (In No. 3)--- (High in No. 3). (High in No. Remarks See “‘Princeton Localities.” Do. Invertebrates. Plants. Do. Mammal fragments. See faunal list. Do. Do. Do. Do. Invertebrates. SCARRITT QUARRY. See ‘Princeton Localities.’” Invertebrates. See ‘Princeton Localities.” Do. Do. Do. Do. Mammal fragments. See faunal list. Do. Plants, tiles. See ‘Princeton Localities.” Do. See faunal list. Mammal fragments. See ‘‘Princeton Localities.’” invertebrates, rep- See faunal list. Fish. Mammal fragments. Do. Do. See faunal list. Champsosaur. Plants. See faunal list FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 4h INDEX TO LOCALITIES BY SECTION NUMBERS Townsuipe 5 Nortu, RANGE 13 East Section 10: Locs. 67, 79. Section 14: Locs, 66, 80. TOWNSHIP 5 NORTH, RANGE 14 East Section 1: Loc. 61. Section 15: Loe. 19. Section 9: Loc. 15. Sections 15-16: Loc. 10. Section 10: Loc. 47. Section 22: Loe. 13. Section 12: Loc. 72. Section 23: Loes. 11, 49. Section 13: Loes. 20, 56. Section 24: Loc. 46. Section 14: Loe. 18. Section 25: Loe. 48. TOWNSHIP 6 NoRTH, RANGE 14 EAstT Section 36: Loc. 71. TOWNSHIP 3 NORTH, RANGE 15 East Section 2: Loc. 24. Section 9: Loc. 50. TownsuHip 4 NorRTH, RANGE 15 East Section 14: Loes. 53, 82. Section 24: Loc. 51. Section 15: Loe. 39. Section 35: Loc. 25. TOWNSHIP 5 NorTH, RANGE 15 EAst Section 7: Locs. 44, 62. Section 19: Loc. 60. Sections 11-14: Loes. 3, 14. Section 20: Loces. 36, 42, 43, 58, 59, 68. Section 11: Loc. 76. Section 23: Loes. 4, 52, 54. Section 15: Loes. 17, 74. Section 25: Loc. 32. Section 16: Loc. 21. Section 29: Loes. 40, 41, 55. Section 17: Loes. 33, 38, 45, 57. Section 32: Loe. 77. Section 18: Locs. 63, 69. TOwNsHIP 6 NorTH, RANGE 15 EAST Section 23: Locs. 8, 81. Section 31: Loc. 70. Section 26: Loc. 31. Section 34: Loc. 16. Section 29: Loes. 28, 29. Section 35: Loe. 9. Section 30: Loes. 12, 27, 30. TownsuHIP 4 NortH, RANGE 16 EAst Section 5: Loc. 75. Section 7: Loc. 2. TOWNSHIP 5 NORTH, RANGE 16 EAST Section 4: Loe. 1. Section 8: Loc. 23. TOWNSHIP 6 NORTH, RANGE 16 EAst Section 25: Loc. 7. Section 33: Loes. 5, 35, 37. Section 26: Locs. 65, 78. Section 34: Loes. 6, 16, 22, 64, 73. Section 31: Loc. 26. TOWNSHIP 7 NORTH, RANGE 16 East Section 36: Loc. 34. 46 BULLETIN 169, UNITED STATES NATIONAL MUSEUM FAUNAL SUCCESSION AND CORRELATION COMBINED FAUNAL LIST Table 3 lists all the genera and species of mammals so far identified in this field, with the approximate level in the Fort Union series at which each is known to occur. The only omissions are a few speci- mens of uncertain level. They include no forms not otherwise recorded here but might have extended the vertical distribution were their horizons accurately known. The list incidentally emphasizes the extraordinary richness of this field and its known faunas. It shows the presence of 51 genera (5 tentatively identified, but different from the 46 definitely known) and 79 species (15 tentatively identified or now nameless, but different from the 64 definitely known and named). The field is richer in variety of fossil mammals than the Puerco and Torrejon together, indeed almost as rich as all the other North American Paleocene combined. The figures given for each column are the approximate stratigraphic levels in feet above the base of the Fort Union No. 1. In the 575 column, fossils from Loc. 51 are listed. The 800 column includes Locs. 24, 25, 50, and 81. The 1,150-1,200 column includes Locs. 5, 6, and 54 and the Gidley Quarry. The 1,265 is the Silberling Quarry. The 1,350 includes Locs. 35, 53, and 82, and 1,700-1,750 is Locs. 12, 27, and 28. The ‘Princeton localities’? are included in the 2,250-2,350 column, and Loc. 18 and the Scarritt Quarry in the 3,300-3,500 column. The last column, 4,350, is for Locs. 11 and 13. TABLE 3.—Combined faunal list, Crazy Mountain Field, Mont. No. 1 No. 2 No. 3 a (At | 575 | g00 [1150-11 965 | 1,350 |1:700 | 2.250-|3,350—) 4 359 level | feet | feet | 1 feet | feet | 12750 | 2,350 | 3,500 | “foot shown) feet feet | feet | feet Feet PHIOUUSIINONLANUS Ae men ae ae oe eel ee x x x x PPGlOdUSdOUgIASSt ae a ee eee | Baers ole ee x x PPiilodus gidleytetssss fen ON ea ee ees |e >.< axe CELOMUSISINCLOhIn =a eee ae eee Bilis PS ae x d.« Plodussp22- 2.2522 oe ee OE ee alle UE re x PECLY DOCUSIOTONG CT Ie a ae ne ee ee | ee ene | ene a | eee x PEclypodus 7ussella=_ aoe+ a2 eS | ees SE ee x axG Pictypodus silberlingi= oe ss eee | ea | x PUCLYDOGUSINAUNLET I sae See ee a en | oe eee || ee ee ee ee | ears a xX PE OTECLY DOGS TED SCM tare a eee ee | ee ee eee x | ERUCOSTOGOMEPOTSUss2 2. nese ee eae xX DG al (See ee |e O.S CQELASLODS DAT CUS rea eee ee ere | ee ee | eee ee x x Prodiacodon' concordiarcensis.._...--._|_---& = -a\e. 2. |e _= x UGepla codon laddese soe ane eee ee ee eee eee x Leptacodon muntuscultm.-—_----- 2-2 |2o--- |) no ne xX ieptacodonictl.tenenprcs 2. ate soe tes nae cones ce eal ceeeee (eee a ee eee |p eel eee | eae x FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 47 TABLE 3.—Combined faunal list, Crazy Mountain Field, Mont.—Continued No. 1 Species Feet Myrmecoboides montanensis - ._..-----|-------- Stilpnodon simplicidens.._.....-.-.-.--|-------- Bessoecetor diluculi OPER GITEOSITLOPUIS Peete ses oe ee eee eee ee PA DRTONOTIULS |RAUGQLON- = oe ene ene oe oo Eudaemonema cuspidaia._....--------|-------- Elpidophorus patratus_—2.....-.------|=--..--- UDIODROTUSLININ OF an ae eee eae eee ee ACTOUIR SILUCNIVILG Pattee nee ee | ee CU CHANIGASA pics re ae eee a ene | ae ee ‘Pronothodectesmatthewt: = 2 s225=.2222| ees IPlestadapis AnCepss na. 3. sama nono nee | ee eee ae PleSiQdA TSI er oe eee = eens cna e eee a | sees eco aL Betas Pl DhidOLansruslOnenCihe=-= s— =~ - see won| pe oee ne Ae ee Carmolanies N@zelaes == - sane aaa ae [onan ee eee FPOTOMOMPSINACUILUS T= onan eee ee | eee nee eee Paromomiysidepressidens. == 2. 52 eae Cf. Paromomys, genus and species Nd eLanmined' = _ hav canes ca seases|soescess|peeess IPOlOCCRthONAUICUSDIS oe a oe onan eal aoa a= | sateen PPAlenOChih@uliNOlraae 22 noe ee en ee| ba nessa | seen oe PRENACOleMUT TUG WONUS 228 Soe cee ono eee | oo Conory Ctes|COMME= sa 2 = 35-5 --ee==-=|---55-+= |e = =8 @iAConoryctesispeio2- = oo 5 2 2255252) BIL LS Psittacotherium multifragum........--|--------|------ CIAENOCONIMONTANENS1S so sae = aaa ee ea ee aeee eee ClnenodonisilberUinginn oes ase oO eee ae Claenadoniiatidens== >. * ees =e eee | ee GClaenodon’ terror sto... ssa ane ace sseen)| 2 2ebons sacs Claenodonisperes2: sos 2s=- 222 scent ase |ecteess|oocaes Deuterogonodon montanus--.......---|----.---|------ PDCULENOGONOGONISDi2 = = seen ae eee eae ee ee nee IPT OUT DLACOCONYUTENS =e canoe oe aoe oa ees oe eee Prothrypiacodon, (/Urens = === -. 2 Sees eaten No. 2 No. 3 ! 1,150- 1,700-| 2,250- La ih ON fare 1,350 | 1,750 | 2,350 feet feet | feet ufeles a) x Nanas x ei el 5) HR 1 | aenele ns oa ne tte dana Ke fx ee ee x ieee oe nee axe ss Xx ER Eee x xX ieee xX pas aXe X aefs 1 Xe foe sees! Xe x Pomel [er Wes MEDOAN eer x Ses Ae x A See aXe ee mS x peeve. XS EW. SEE ER en ereeacd XK LCS Xe ie as Xe x x Lee >:€ ox x a eECY x Exe x Aa a) eh ee x eared x vay ees aXe >, Sei ees >.€ LER notes IX aXe x ax asl x XG Cee | x Lae: x Stee xX Xx oo ee aXG Xe aX 3,350- E 00 | a feet x x x xX x x pees xX x poe xX x See Sse x 4 x x 48 BULLETIN 169, UNITED STATES NATIONAL MUSEUM TABLE 3.—Combined faunal list, Crazy Mountain Field, Mont.—Continued No. 1 No. 2 No. 3 Species 5 : revel | 25 | 800 | 7599 | 1,265 |1,350 | 772) | 35%) [329 | 4,350 shown) feet | feet feet feet | feet feet | feet | feet feet we Feet LALO MY US CISSENLONEUS= a sa snes neo e ee |e ee oe x | Eloplaletesidisceptatnine. 22428 cee e | | ae x | CLibaplaletes'Sp ase an tes kan an ee Pesos Alans Ai HL Bele Cee oe NE ee fexse fe yoann x Hitolestes.motissimius: 2. 2222) sae Wal owe a [a ea | ee Se eecese x Tetraclaenodon symbolicus...------.--|_...--__|__--_- am EUR xg | Tetraclaenodon ef. symbolicus .._..-__- 300 | X Tetraclaenodon' ci. puercensis.- -—--224|2 222. a e- | oc - oe EX Alsen x REtTACLACNOG ONS) D eee =e a ke | ee rc tonmen | (Sg fae ee | oe | sede oe | ee x x Gidleyinavmontanens7s Cae ert ee | SAEs | reed 0A | CRON Sel (pat Sede ea eae xX PGidleyinarsilberlingisse. sete tek ran cee I ee olla Ae ee | Ae ete ae oe x PGCIGLEUITIORSU DETt Once es eee tes ae eel | Coe wee | he kN en | ee eee ae SNE es pares Sie x PGidleyiNna Spy Sse es Se SSE ee a UE ee A Eine es 2 Cori PRAGUSMNONIANUS 22 ton ee ane ane ee oe a x x Anisonchus'sectorius-_. 2.2.20 2-3-} 55.222 22 >.< x SKA | eer dl eee x x Pantolumod@UnterMedius sans nae aa ee ee | eee el x PPantolambdarsp 3 35 ewer see | hee oly oe red x Pantolambdidsundetermined <.4-.2-|2 =| ee eee Xe xX x x FAUNAL SUCCESSION The No. 1 beds have yielded only ?Ptilodus sinclairi from near the base; Clacnodon vecordensis, Chriacus pugnaz, ?Mimotricentes sp., and Tetraclaenodon ?symbolicus from near the middle part; and Mimotri- centes angustidens from the top. The first and last of these forms are inseparable from species occurring well up into the No. 2, and so is Tetraclaenodon ?symbolicus. Claenodon vecordensis and Chriacus pugnax are distinct species but cannot be considered to be clearly ancestral to anything known from the No. 2. The evidence is inade- quate to demonstrate either occurrence or absence of distinct faunal advance within the No. 1-No. 2 series. It does suggest, however, that such advance must have been relatively slight and that it was probably of less than generic rank, if it occurred. The data definitely, if not altogether conclusively, oppose the hypothesis that the difference between No. 1 and No. 2 faunas can have been as great as between Puerco and Torrejon. Within the No. 2 beds the first important point is that the Gidley and Silberling Quarry faunas are identical in degree of evolutionary progress and show no significant difference in age despite a small difference in level, about 65 feet. There are 23 species definitely identified in the Silberling Quarry. Seventeen of these also occur in the Gidley Quarry and, as shown-in the systematic descriptions, these suggest no difference, even of less than specific rank, between the forms in the two quarries. Four Silberling Quarry species do not occur in the Gidley Quarry but do occur at lower levels, so that their FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 49 absence in the Gidley Quarry is not due to time differences but to facies or accidents of discovery. These species are Hucosmodon sparsus,” Mimotricentes angustidens, Didymictis haydenianus,” and Tetraclaenodon symbolicus. Only two species, Psittacotherium multi- fragum and Elpidophorus minor, known only from one specimen each, are confined to the Silberling Quarry, and there can be no doubt that this is the result of chance. The Gidley and Silberling Quarry faunas are, then, a unit as far as appreciable differences in time are concerned. The 575-foot level is represented by eight definitely identified species, of which seven also occur in the Gidley or Silberling Quarries or both, and the other, Metachriacus provocator, reappears at the 800-foot level. The 800-foot localities have 10 definitely identified species, 8 of which also occur in the Gidley or Silberling Quarries or both, while Deuterogonodon montanus is confined to this level and Metachriacus provocator appears also 225 feet lower. All the material of Deuterogonodon montanus is from a single locality, and it must be considered as a sporadic and chance occurrence with no probable validity as an index of time relationship, particularly as no probable or possible ancestral or descendent forms occur in this field. The distribution of Metachriacus prevocator is peculiar. It is relatively very abundant at No. 2 levels below the Gidley Quarry, being the only species that occurs at all these localities (except for Loc. 52, which is only 50 feet below the Gidley Quarry and must be considered as at essentially the same level). At the Gidley and Silberling Quarry levels it has not been found, but an allied species, M. punitor, is present and fairly common. WM. provocator is the larger and in some other respects probably the more specialized of the two species, despite its earlier appearance. Thus the relationship is not that of ancestor to descendant, and no temporal difference is demonstrated. The facies of these earlier localities are demonstrably quite different from those of the quarries, and the most reasonable conclusion is that the distribution of the two species of Metachriacus is conditioned by facies and not by time. It thus appears that the whole Fort Union No. 2 shows no probable significant faunal advance and that it forms a unit as far as time and correlation are concerned. It is particularly unfortunate that the three lower fossil zones of the No. 3 beds are represented by few and poor specimens, for this is a crucial point in the sequence, probably covering (as shown under “Correlation’’) the transition from Middle to Upper Paleocene. The lowest level, 1,350 feet, yields Claenodon ?ferox, Tetraclaenodon cf. puercensis, and indeterminable pantolambdids. The latter have no particular importance. Claenodon ?ferox occurs almost throughout 12 These probably are present in unprepared Gidley Quarry material. 50 BULLETIN 169, UNITED STATES NATIONAL MUSEUM the No. 3 beds but is not recorded in the No. 2 where there are smaller and probably somewhat more primitive species of the same genus. The Tetraclaenodon is not distinguishable from one in the Gidley Quarry, or from fragmentary specimens from high in the No. 3. The 1,700-foot level has cf. Conoryctes sp., ?Gidleyina silberlingi, Anisonchus sectorius, and pantolambdids. The first and last have no apparent significance for this discussion. ?Gidleyina silberling: marks the first (probable) appearance of that genus, which seems to be confined to the No. 3 and more progressive than anything in the No. 2. Anison- chus sectorius ranges throughout the No. 2 and far up into the No. 3. The 2,250 to 2,350 levels (chiefly ‘Princeton localities’) yield Claenodon ?feror, Gidleyina montanensis and G. spp., Anisonchus sectorius, and pantolambdids that are undetermined but are distinct from anything in the No. 2 beds. Loc. 15, which 1s somewhere in the lower No. 3 although not determined exactly as to level, yields an ally of Paromomys not demonstrably distinct from the No. 2 repre- sentation of that group. These lower No. 3 beds as a whole appear to be characterized by the survival of some genera, including Claenodon, Tetraclaenodon, Anisonchus, and perhaps Paromomys and Conoryctes, and of at least one species, Anisonchus sectorius, from the No. 2 beds, and by the appearance of a distinctive advanced species, Claenodon ?ferox, and a similarly distinctive and relatively specialized genus, Gidleyina. These two latter may not be real cases of faunal advance, for it is possible that these animals were already present in No. 2 time despite the lack of discovery, yet it seems tentatively acceptable to consider them as suggestive of some slight faunal progress. The data are too scanty for any definitive conclusion, but the indications are that these strata have a transitional fauna between that of the No. 2 beds and that of the Scarritt Quarry, about 1,000 feet above these lower No. 3 levels. The Scarritt Quarry, at about 3,350 feet, provides good data and for the first time shows decisive and incontrovertible evidence of faunal change in the series. Ectypodus hunteri is evidently related to some species from the Gidley Quarry but is certainly distinct and probably of later aspect, and the same is true of Leptacodon cf. tener. Bessoecetor thomsoni also has a Gidley Quarry ally but is probably not a descendant of the latter. The relative degree of specialization is not clear. Elpidophorus patratus is more specialized than its fore- runner ?£. minor in the Silberling Quarry. Plesiadapis anceps is very primitive in its genus yet is more advanced than Pronothodectes, from which it could well have been derived. The relation between Carpodaptes in the Scarritt Quarry and Elphidotarsius in the Gidley Quarry is closely analogous to that between Plesiadapis and Pronotho- dectes. Phenacolemur has no definitely recognizable ancestor in the FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 5t lower beds but is a progressive and specialized form. Litolestes is allied to Haplaletes and might have been, but probably was not, derived from the latter (but surely not from the known species). In any case, it is a distinctive genus abundant here and unknown in the No. 2 beds. Tetraclaenodon occurs in both series, and the species are not well enough known for useful comparison. Finally, Anisonchus sectorius, which ranges throughout the No. 2, is also present here (not in the Scarritt Quarry but at Loc. 18, which is not appreciably different in level). If we discount the influence of facies and chance as far as possible, it then seems clear that there is definite faunal advance from the Gidley and Silberling Quarry levels to that of the Searritt Quarry, that this advance is in some cases of generic rank (Pronothodectes— Plesiadapis and Elphidotarsius—Carpodaptes) but in others only specific (probably in the genera Ectypodus, Leptacodon, Bessoecetor, Elpidophorus, and Tetraclaenodon) and in part of even less degree (Anisonchus sectorius).¥ The progressive difference in the two faunas compared is real, but it is not very marked, much less than would be expected from a difference in level of some 2,000 feet. As nearly as such an imperfect parallel can be drawn, the difference seems to be definitely less than that between the Torrejon and the Tiffany and probably not much greater than between the Tiffany and Clark Fork. The highest faunal level, 4,350 feet, has provided little good evidence bearing on this discussion. Plesiadapis rex may be more progressive than P. anceps but is not definitely shown to be so. The Paromomys- like form is too poorly known for useful comparison. Tetraclaenodon still occurs at the higher level and so probably does Gidleyina, their degree of advance, if any, over comparable forms lower in the No. 3 beds not being clear. Claenodon ?ferox still occurs here and seems to be quite as in the lower No. 3, although the material is scanty. Thryptacodon, identi- fied with much probability but not certainty, 1s progressive over Prothryptacodon of the No. 2, but the intermediate stages in the lower No. 3 are unknown. An ally of Haplaletes here is too poorly known to give reliable evidence. Hlpidophorus, probably patratus, carries through from the Scarritt Quarry level. The data are too inadequate to say that there is no advance over the Scarritt Quarry, or indeed over the lower No. 3, but they suggest that the difference will probably prove to be slight if it exists at all. 13 The tremendous vertical range of Anisonchus sectorius in this field, nearly 3,000 feet, is extraordinary, This species and to still greater degree this genus are of unusually wide geographic range in the Paleocene, the species known from New Mexico to Montana and the genus also in Louisiana, and also are unusually conservative—the genus is one of the two or three that pass through from Puerco to Torrejon in New Mexico. 52 BULLETIN 169, UNITED STATES NATIONAL MUSEUM CORRELATION OF MAMMALIAN FAUNAS Data are not at hand for any exact correlation of the No. 1 beds. Of their mammals ?Ptilodus sinclairi, identified here with some prob- ability, is a No. 2 species. Claenodon is typically a Middle Paleocene genus, ranging in this field into the Upper Paleocene. Chriacus is a very long-lived genus, Lower Paleocene to Lower Eocene, and is here rather dubious but apparently of more progressive aspect than its Lower Paleocene allies. A/imotricentes is otherwise known only in the Middle Paleocene of this field, and M. angustidens (which, however, is known only from the highest No. 1 beds) ranges to the top of the No.2. Tetraclaenodon is typically Middle Paleocene, perhaps ranging into earliest Upper Paleocene in this field, and the species may be the same asin the No.2. As identified by Russell (personal communica- tion), invertebrates from immediately below the No. 1 beds, or perhaps in their base, are definitely of Paleocene aspect, and most of his com- parisons are closely with Middle to Upper Paleocene forms. On the whole the scanty evidence favors reference of the No. 1 beds to the Middle Paleocene, and none of it definitely suggests or warrants reference to the Lower Paleocene. The physical stratigraphic evi- dence, while even less decisive, is also consistent with belief that the No. 1 beds are not markedly different from the No. 2inage. Nothing suggests or justifies correlation with the Puerco. As already shown, the fauna of the No. 2 beds is a unit as far as correlation is concerned. From his first discovery (Locs. 5 and 6), Douglass (1902) already recognized equivalence to the Torrejon. Matthew (1914) accepted this fauna as of the same age as that of the Torrejon.* Gidley (1909, p. 616, footnote, apparently not else- where explicitly stated, but his definitive opinion) considered it ‘about equivalent to ... or perhaps somewhat older than the Torrejon.”’ The detailed evidence of the mammalian fauna is as follows: Ptilodus: A Middle Paleocene genus, P. montanus very near P. mediaevus of the Torrejon. Ectypodus cae Upper Paleocene genera, but species here quite distinc- Parectypodus|"| tive and generic assignments not certain. Eucosmodon: Forms indistinguishable generically on the basis of similar material range through the Paleocene. Gelastops: Probably allied to, but distinct from, Didelphodus of the Lower Eocene. Probably allied to but less specialized than Acmeodon of the Torrejon. Prodiacodon: Middle Paleocene genus, identification here not certain and species highly distinctive. Leptacodon: Genus otherwise Upper Paleocene, species distinctive. Myrmecoboides: Unknown elsewhere, no value in close correlation. Stilpnodon: This genus and its family too imperfectly known for close comparisons. Bessoecetor: Otherwise known only in the No. 8 beds in this field, allied to the typically Lower Eocene Palaeosinopa, but more primitive. 14 His correlation chart, fig. 2, seems to correlate it with the Puerco, but his text shows that this was not intended. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 53 Aphronorus: Allied to Pentacodon of the Torrejon but apparently more primitive. Eudaemonema: Allied to Mixodectes in the Torrejon and (perhaps more distantly) to other genera in Upper Paleocene and Lower Eocene, more primitive than any of these allies in some respects, but not on the same line of descent as any of them. Elpidophorus: An Upper Paleocene genus but here with a distinctive primitive species, which may well prove generically separable. Picrodus: Probably allied to Zanycteris of the Upper Paleocene, relative degree of specialization not determinable. Pronothodectes: Allied to Plesiadapis of the Upper Paleocene but distinctly more primitive. Elphidotarsius: Allied to Carpodaptes and Carpolestes of the Upper Paleocene but distinctly more primitive than either. eer amnus tence primitive genera, but exact lines of descent not clear Palaechthon : : : enough for more precise use in correlation. Palenochtha Conoryctes: Genus otherwise Middle Paleocene as far as definitely known, species imperfectly represented but indistinguishable from C. comma of the Torrejon. Psittacotherium: Genus otherwise certainly known only in Middle Paleocene, species indistinguishable from P. multifragum of the Torrejon. Claenodon: Genus otherwise chiefly Middle Paleocene but extending into Upper Paleocene in this field, species more primitive than that commonest in Torrejon, more comparable to but perhaps slightly more primitive than the second Torrejon species. Deuterogonodon: A distinctive genus not known elsewhere, perhaps especially allied to, but if so distinctly more advanced than, Protogonodon of the Puerco. Prothryptacodon: Allied to Thryptacodon, Upper Paleocene and Lower Eocene, but more primitive. eee ests , se unknown elsewhere, of Paleocene aspect but not more Spanoxyodon|'| closely useful in correlation. Mimotricentes: Perhaps especially allied to Tricentes, Middle Paleocene, and if so slightly more primitive, but probably on a different line of descent. Didymictis: Genus Middle Paleocene to Lower Eocene, relationship of two species not clear enough for more exact correlation, but a third indistinguishable from D. haydenianus of the Torrejon. Ictidopappus: Unknown elsewhere, perhaps especially allied to Viverravus of the Eocene, but if so more primitive. Dissacus: Genus Middle Paleocene to Lower Eocene, species not identifiable. Ellipsodon: Genus otherwise only positively identified in Middle Eocene, species close to #. acolytus, Torrejon, perhaps slightly more advanced or on different line of descent. Litaletes Genera unknown elsewhere, general aspeet Paleocene, perhaps Litomylus rather more Middle Paleocene, but not exactly determinable as Haplaletes to evolutionary position. Tetraclaenodon: Genus typically Middle Paleocene, probably reaching Upper Paleocene in this field, two species close to those from Torrejon, one of them indistinguishable. Coriphagus: Genus otherwise only in Torrejon, species closely allied. Anisonchus: Genus Lower and Middle Paleocene, also Upper Paleocene in this field only, species same as in Torrejon, but probably also reaches Upper Paleocene here. Pantolambda: Genus certainly identified only in Middle Paleocene, species allied to those of Torrejon. 54 BULLETIN 169, UNITED STATES NATIONAL MUSEUM In résumé, there are 13 genera all of Paleocene aspect but con- fined to this fauna and of little use in exact correlation. Of the eight other genera confined to this fauna, four have their closest allies in the Torrejon and four have their closest known allies in the Upper Paleocene, but three of these here appear to be distinctly more primi- tive than their Upper Paleocene relatives. Eight genera are other- wise typical of the Torrejon, and these here include two species indistinguishable from those of the Torrejon. The other species of these genera are in most cases closely allied to those of the Torrejon, mostly without definite evidence of being more or less advanced, but one may be less and one may be more specialized than the most comparable Torrejon species. Four genera are common to this fauna and to that of the Torrejon, but they also appear in later beds. Their species are here generally closer to the Torrejon species and in two cases appear to be identical. Two genera probably range nearly or quite throughout the Paleocene, one of them here represented by a typically Torrejon species, but one that does range into some- what later beds. Five of the genera are otherwise known only in Upper Paleocene (in some cases also ranging into later) beds, but in each case the species here present is decidedly distinctive, and in no case is a Middle Paleocene ancestor or close ally otherwise known. This evidence conclusively proves the fauna to belong to the Middle Paleocene and to be close to the Torrejon in age. It does not preclude a slight difference in age from the Torrejon, but dif- ferences of facies and geographic position are adequate to explain the faunal distinction without supposing the age to be different, and if the age is slightly different they obscure the evidence for this. There is nothing suggestive of the Puerco. There is a greater resemblance to the Upper Paleocene than is shown by the Torrejon fauna, but the evidence suggests that this is largely or wholly due to the discovery in this place and facies of forerunners of later groups not themselves really later here than the Torrejon. The only other very closely correlatable horizon is that of the Scarritt Quarry. The correlation has been discussed elsewhere (Simpson, 1936b), where it was shown that the quarry fauna itself is very close to that of the Tiffany in age but may be slightly earlier. From a more general point of view, there is some suggestive but inconclusive evidence strengthening the probability of slightly earlier age for this than for the principal Tiffany horizon. The probable occurrence of Tetraclaenodon at an even higher level, replaced in the Tiffany and all later beds by the more advanced and possibly descendant genus Phenacodus, weights the evidence in this direction, as does also the occurrence at about this level of a specimen indis- tinguishable from Anisonchus sectorius, a Torrejon species. The occurrence of Claenodon indistinguishable from the Torrejon C. ferox FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. i) far up into the No. 3, even higher than the Scarritt Quarry, is less dependable but suggests the same conclusion. It is also suggestive that the age difference between the Gidley and Silberling Quarry faunas, which are very close to the Torrejon in age, and that of the Scarritt Quarry appears to be less pronounced than the age difference between the Torrejon and the Tiffany. The Scarritt Quarry and equivalent levels may be correlated with the Tiffany, with the reservation that it is possibly a little earlier, and may be placed in the earliest Upper Paleocene. ‘The highest faunal level, that of Locs. 11 and 13, is likewise surely Upper Paleo- cene, probably about Tiffany. The lower levels in the No. 3 cannot be well correlated on present data. They are bounded by Torrejon and (probably early) Tiffany equivalents. There is some slight evidence that they are in fact transitional between these two and represent most of the gap in time known to exist between Torrejon and Tiffany, rather than being more definitely associated with one or the other. If there is a dis- tinct faunal break in this series, it would be logical to look for it at the most obvious change in sedimentation, at the base of the No. 3. The scanty fossil evidence does not warrant such a positive conclusion. The only known elements in the basal No. 3 that look toward the later faunules are Claenodon ?ferox of little or no value, since this is, in fact, a typical Torrejon species, and Gidleyina, which is more suggestive but in itself far from conclusive. Posonspe NEW Wyomine: MOonrmTANA: Sanenariaasin Bighorn Basin Crazy Mountain Field (No mammals : (No mammals known) Clark Fork known) RT Re Ree ie “Tiffany-Bear Tiffan 5 Creek” y (Princeton Quarry)! |-------- PAIGE SORES PALEOCENE (Faunal break) (Faunal break) Je aay pie Bere eS: ee a : : Gidley “ oe 5 J Torrejon Torrejon Quarry Lebo (Faunal break) (aunalibreak) |22=s2 22" iy aires Tt: yy ea Puerco Puerco (Age uncertain) CRETACEOUS (Unconformity) Lance ‘Hell Greek 1 Dr. Jepsen has not yet applied valid stratigraphic names to these levels, distinguished by him. 119212—37. 5 56 BULLETIN 169, UNITED STATES NATIONAL MUSEUM The known faunas of this field thus seem to cover most of or all the Middle Paleocene and to continue, probably without a marked break, into the early Upper Paleocene. There are here at least 2,000 feet of sediments of about the same type above the highest level with identified mammals, and sediments of somewhat different character in contiguous areas might raise this figure to as much as 4,000 feet. These beds doubtless represent most of or all the con- siderable span of later Upper Paleocene time not represented by faunas here, and they may well run into the true Eocene. The tentative arrangement based on these correlations is summed up in the diagram on page 55. Such a chart is necessarily excessively simplified, so that it does not adequately represent abundant and exact evidence where this is available; at the same time it must be explicit, so that in places it does not reveal the extent of uncertainty and lack of evidence. It is a correlation of strata by their known faunas, never entirely coextensive with the strata themselves. The “faunal breaks’’ listed definitely imply intervals of time not represented by known faunas, but it does not follow that strata of these ages are necessarily lacking in the region in question. The same is true of the parts outside the known distri- bution of faunas that are labeled ‘no mammals known.” ECOLOGY NOTES ON THE NONMAMMALIAN BIOTA Despite the abundance of nonmammalian fossils in the field and the existence of good collections made by Silberling and others, this part of the biota has not been adequately studied for this area. It is beyond the scope of the present work to include any critical exami- nation of the material or to attempt complete floral and faunal lists, nor has it seemed necessary to delay publication until such special studies could be made. In the present section a few previous records are mentioned, and beyond this the only purpose is to suggest the broad outlines or general character of the biotic background of the mammalian faunas. Plants—A number of small collections of leaves have been made in this field, but as far as I know no attempt to collect and to classify its whole flora has been made. Knowlton (1902) has published a special note on one small collection, and several other species are listed incidentally in the literature. Knowlton (1909, p. 194) listed a few plants from the ‘lower mem- ber of the Fort Union” (by which he usually means the Hell Creek or Lance), 1,000 feet below the top. Asitis not clear what he considered “the top”’, this is not explicit. He states that they were ‘‘about 1,200 feet below the small mammals now being studied by Mr. Gidley’’, which probably means the Silberling Quarry and would (by my esti- FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 5G mate) place the plants near the top of the Bear. They are said to be from sec. 29, T. 6 N., R. 16 E., a section that is mainly on the Bear but includes a small area of No. 1. The identifications are: Salix sp. Aralia notata. Populus amblyrhyncha. Leguminosites arachioides. Populus cuneata. Vitis xantholithensis. Populus genetrix. Credneria daturaefolia. Populus daphnogenoides. Phyllites cupanoides. Populus sp. Carpites sp. These are all common Fort Union species. Knowlton has also re- ported them all as Lance (e. g., Knowlton, 1919), but on examining the evidence it is found that Vitis (or Ampelopsis) xantholithensis, Cred- neria daturaefolia, and Phyllites (or Pterospermites) cupanoides are true Fort Union species reported in the Lance only on the strength of this occurrence, and that there is also some question as to whether Aralia notata really occurs in the Cretaceous, whereas it is abundantly characteristic of the Fort Union.’® In short, this is a Fort Union flora, which would be expected since it is probably from the Bear, which may well be Paleocene or even properly Fort Union, or possibly from the Lebo, which everyone now agrees to be Fort Union. Other small collections more definitely from the Bear and older than any of the known mammals of this field are mentioned above. A collection surely from the Lebo, ‘‘close under the massive gray sandstone of the Fort Union” (i. e., the base of the No. 3 beds), in “Sec. 15, T.6 N., R. 15 E.” (actually sec. 22, as later surveys have shown; sec. 15 includes only the very base of the Lebo), was made by Campbell, Stanton, Stone, and Calvert, and identified by Knowlton as follows (in Stone and Calvert, 1910, p. 755): Plaianus haydenit. Sapindus grandifoliolus. Populus cuneata. Grewia obovata. Populus amblyrhyncha. Grewtopsis platantfolia. These are all Fort Union species, and all occur in the Glendive region. The majority also occur in the Lance but are certainly in the Fort Union in this section. Knowlton (1902) described a small flora collected by W. H. Weed from ‘‘the sandstone series above the bend of the Sweet Grass, west of Porcupine Butte, Montana.’ This locality is in this field, at the extreme western edge of the mapped area, at a high horizon, far above any identified mammals. The following species were recorded: Glyp- tostrobus europaeus,s Onoclea sensibilis fossilis, Aralia ?notata, and Tilia weedii. The last has not been recorded elsewhere, but the others are Fort Union (and questionably earlier) species, and Knowl- 15 This incident casts further doubt on the supposed community of the Lance and Fort Union floras, for there are many other areas where beds probably or surely Paleocene and definitely post-Hell Creek have been arbitrarily called ‘‘Lance.’’ 16 G. e. ungeri in the original list, but later emended by Knowlton to the form here given. 58 BULLETIN 169, UNITED STATES NATIONAL MUSEUM ton concludes that the beds are Fort Union. Incidentally this is some slight evidence that these upper beds are still in the Paleocene, but this perhaps cannot be taken very seriously. It seems probable that the Fort Union flora of this field is closely similar to that of the Bull Mountain field (lists in Woolsey, Richards, and Lupton, 1917) and also to the very large and well-known flora of the Glendive area (lists and references in Knowlton, 1919), in short, that it is the typical, wide-ranging Fort Union flora, which seems to have shown little geographic or stratigraphic differentiation from Upper Cretaceous to Upper Paleocene and throughout the West. Leaves are found at almost all levels and throughout the field, although really well-preserved specimens can seldom be recovered. Specimens apparently of the handsome species Platanus nobilis were particularly noticeable in the sandstones stripped from above the bone layer in the Gidley Quarry. The general occurrence of leaves here and elsewhere in the field suggests that the whole region was heavily forested, chiefly by deciduous trees, during much of or all the Paleo- cene. It demonstrates the presence of a well-developed arboreal habitat and of abundant food for browsing and frugivorous animals and suggests (but by no means proves) that the more open type of plains habitat was here relatively restricted or absent. Knowlton (1927, pp. 184-186) has summed up the Fort Union flora in general, and probably his remarks apply in large measure to the plants of this field throughout Lebo and Melville time. About a dozen species of small ferns and the sensitive fern Onoclea are known, as well as horsetails, ‘a beautiful little selaginella”’, and abundant grasses and sedges. Only one palm, and that rare, is recorded. ‘It was, however, a very large-leaved fan-palm, showing that conditions were not altogether unfavorable.” Conifers are abundant, with three sequoias, a yew, bald cypress, two or three pines, and an arbor- vitae that was particularly abundant. There is a rare ginkgo. The most conspicuous and abundant dicotyledonous trees were poplars, with leaves suggestive of the quaking aspen. Sycamores also were abundant, some with very large leaves. Viburnums were also very common and were ‘‘presumably small trees or shrubs, known at the present day as arrowwood, blackhaw, tree cranberry, etc.”’ There were also oak, alder, chestnut, hazelnut, maple, elm, magnolia, hick- ory, walnut, birch, beech, ampelopsis, bittersweet, and rare figs and laurels. Invertebrates—Mollusks occur throughout the field and often in extraordinary abundance. The majority of them are fresh-water mussels and gastropods, although a few may be terrestrial snails. The common types are several different species each of ‘Unio’, Viviparus, and Campeloma. A large collection from just below the Lebo has been listed page 17. Dr. Russell has also identified the FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 59 following from Loc. 24 Gmmediately above the mammals): Viviparus retusus, V. planolatere, and Lioplax nebrascensis; and the following from Loc. 67, high in the No. 3: ?Nedionidus senectus, ?N. s. declivis (type locality; Russell, 1934), Vieiparus retusus, and V. cf. planolatere. (There are also large collections not adequately studied, nor is it necessary to discuss them here where only the general ecological aspect is in question.) Altogether, the invertebrates show that fresh water was abundantly present and thickly populated in this area throughout the deposition of the Fort Union. Fishes.—Fish remains are abundant in the field but usually very fragmentary. A few relatively good specimens have been recovered but have not yet been prepared or closely studied. Gar scales (Lep- isosteus sp.) are common at most surface localities and also occur in all three quarries. A still commoner fish in the quarries, seldom found on surface exposures, probably because of its more fragile character, appears to be an ally of, or to belong to, Stylomyleodon Russell, and Russell has already recorded its presence in this field (1928, p. 107). Specimens recently collected will add greatly to knowledge of thisform. Despite the almost complete lack of associa- tion of mammal remains in the Gidley Quarry, it is the rule rather than the exception to find fishes, apparently mostly Stylomyleodon, represented by associated strings of vertebrae, despite the loose articulation of these, and not infrequently also with parts of the skull in association. Reptiles—Champsosaur and crocodilian remains, generally too imperfect for close identification, are common at all levels. Isolated teeth suggest that Allognathosuchus was the common crocodilian. Champsosaurs occur up to at least 3,000 feet, and possibly more, above the base of the No. 3. Turtle remains are also common, but they also are usually too fragmentary for ready identification. Hay (1908, p. 498) has, however, described Aspideretes nassau from a specimen found by one of the Princeton parties in the No. 3 beds. Gilmore (1928b) has mentioned three fragmentary specimens of Peltosaurus sp. from the Fort Union No. 2 Silberling Quarry and has also named Harpagosaurus excedens, the type of which is from the same locality. Lizard remains are fairly common in both Silberling and Gidley Quarries, and it is probable that these animals were abundant in this region in the Middle Paleocene. MAMMALIAN FACIES The percentage composition of the principal No. 2 surface locali- ties, all near the same level and similar in facies, and of each of the three quarries is given in table 4 and shown graphically in figures 1 and 2. 60 BULLETIN 169, UNITED STATES NATIONAL MUSEUM FIGURE 1.—Diagram showing the relative abundance of identified individuals of the various orders an families of mammals in the National Museum collection from the Gidley Quarry. LOCALITIES GIDLEY SILBERLING SCARRITT soar QUARRY QUARRY QUARRY ORDER FAMILY AG PERIPTYCHIDE_] C.PHENACODONTIDZ D.HYOPSODONTIDA. E.,MESONYCHIDA Uh \ G.ARCTOCYONIDA CONDYLARTHRA H.STYLINODONTIDA L?ANAPTOMORPHIDA J,CARPOLESTID K.PLESIADAPID& L,PICRODONTIDA~ M. MIXODECTIDZE O. LEPTICTIDA. QIDELTATHERIDIDA. MULTITUBERCULAT es R. PTILODONT) D4 CP.NYCTITHERIID4E IN GIDLEY QUARRY LESS THAN Y4 PER CENT) FiGcuR£ 2.—Comparison of relative abundance of families of mammals at various localities in the Crazy Mountain Field, Mont. Abundance is indicated for each locality by percentages of identified individuals in the National Museum collection. Four Lebo surface localities (combined) and the three principal quarries are compared. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 61 TABLE 4.—Composition (in percentages) of the principal No. 2 surface localities and of each of the three quarries, Crazy Mountain Field, Mont. Localities Gidley Oo ly ’ y rder and fami aoe 24, 25 : Quarry Quarry Quarry MULTITUBERCULATA: 9 24 44 36 iPrilodontidaente 2-2-2 32. Se eee ae oe sh 9 24 44 36 INSECTIVORA: 1 18 12 26 aWeltatherididae =e =te sos eee ees Se 0 1 2 0 Weptictidsesases = 6 ee. Ss a eA ee ena 0 5 0 1 Ny ctitheriidaes {2s See a Sees ease one ee 0 (1) 0 0 IPANITOLEStIG BOs een at ao ees Ee ee ee 1 8 tf 15 IVIIxOGOCHIGNG 2 eso s= aaa Be ae ee 0 2 2 9 IPICTOGONUIO AOS see ae ee ee ns see ees 0 2 2 0 Wnicertaeisedis: 2Sese 2 sana Se Skt 32522 os See she 0 0 0 1 PRIMATES: 0 20 4 13 Plesiadapidae.="--25-==-22= + 22s. aes eee 0 2 0 7 @arpolestiGaevese= nes een ea oe nee eee 0 (1) 0 4 tAnaptomorphidae::2: -22-2433522252 S222 Ss See 0 18 4 0 Uncertaepsedis == so=Sa s eS ea tee 0 0 0 1 TAENIODONTA: 0 (1) 4 0 Stylinodontidae =) 22-5) = 3 SS. essen ee 0 (1) 4 0 CARNIVORA: 53 13 19 1 Anctocyonidaess=~! 32 hehe ee ceo peer obs 49 7 12 1 IVI Seid Aes! see Ss isa ee eee Se eG SSS 4 5 5) 0 Mesonychidaesc2=2-5s-22s--20L eens esses en ee esos 0 (1) 2 0 CONDYLARTHRA: 37 24 18 22 Ebyopsodontidsae £222 422-222 42-2 ee ea oe 4 20 14 21 iIPhenacodontidaes= = = ses. ee 16 1 74 Peripty chidaes. =: =- -2 =. 2. se ee ee AS 16 4 2 0 PANTODONTA: 0 1 0 2 iPantolambdidae- 2322 = - fee ee ees Se ee 0 1 0 2 1 Represented by 1 specimen (less than 0.5 percent). The figures represent the composition of the identified collections. They are somewhat biased in favor of the multituberculates, since practically all the specimens of these are identified, whereas there are numerous isolated teeth of other orders that are not identified, but in general they are representative of the faunas as they occur. The Gidley and Silberling Quarries agree well, within the probable limits of chance, except for the much smaller proportion of primates and greater of multituberculates in the Silberling Quarry, a differ- ence probably representing slight local facial distinction. The surface localities differ remarkably from the quarries and sug- gest very different environmental conditions. This result is doubtless somewhat biased by the fact that minute forms, like insectivores and primates, are more likely to break up and are harder to find at sur- face localities, but this does not explain the difference. Some of the so-called surface material was, in fact, found in place. These locali- 62 BULLETIN 169, UNITED STATES NATIONAL MUSEUM ties have been very closely examined and fragments as small as the smallest isolated insectivore or primate teeth recovered, so that the almost total absence of those groups must really result from their great rarity here. Furthermore, the arctocyonids are really much more abundant at these localities, for not only the relative but also the absolute number of specimens is greater for these localities than for the Gidley and Silberling Quarries, despite the much larger col- lections from the latter. This “surface” fauna, 90 percent carnivores and ungulates, is of more normal type, in comparison with Tertiary faunas generally, than are the quarry faunas. Its members average larger than do those of the quarry faunas, and they are probably terrestrial for the most part. This appears to be a normal flood-plain facies, rather closely analogous to that of the Torrejon.” Its most marked peculiarity is the higher percentage of carnivores than of herbivores, a condition for which no probable explanation is seen. Even at the surface localities there is a surprisingly low percentage of animals really of large size for the Paleocene. The phenaccdonts are of average size for that group, but the periptychids (all Anisonchus) are moderate in size, much smaller than the contemporaneous Peripty- chus, and most of the carnivores are also of middle size, with Deutero- gonodon very rare and other large carnivores absent, although they were common at this time in the Torrejon. This rarity of large animals is still more obvious in the quarries. Phenacodonts and pantolambdids are relatively very rare, Claenodon is uncommon, and other large mammals do not occur. The most abun- dant species, Ptilodus montanus, ?P. sinclairi, Leptacodon ladae, Aphro- norus fraudator, Paromomys maturus, Palaechthon alticuspis, Meta- chriacus punitor, Didymictis microlestes, and Ellipsodon aquilonius, are moderate to minute in size. In food habits the multituberculates were rodentlike, the insecti- vores doubtless insectivorous (as the word is usually used, not signify- ing a diet composed of insects), the primates probably mainly or ex- clusively frugivorous, the creodonts in part omnivorous (Claenodon), omnivorous-carnivorous (other arctocyonids), and predaceous-carniv- orous (miacids), and the condylarths probably browsing, perhaps in part frugivorous, or even partly insectivorous for the smallest forms. The known fauna apparently consumed every type of food known to have been present in the area, with the possible exception of the mollusks. The skeletal structure is too poorly known for any of these animals to give much direct insight into their locomotion or general habitus. By analogy and comparison with allied species and genera, the multi- tuberculates and insectivores were unguiculate and probably in good 17 Correlation of faunal types and collecting methods is real but indirect. Flood-plain deposition and facies would not normally result in concentration of fossils sufficient to permit profitable quarrying. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 63 part arboreal in habitus. The primates, also, may well have been mainly or entirely arboreal. The creodonts were probably mainly terrestrial, but it is probable that some of them were at least semi- arboreal. The abundant hyopsodontids were probably, judged from Torrejon and later allies, more unguiculate than ungulate in general habitus, and the smaller forms may well have been arboreal. The less common larger condylarths and the very rare pantodonts were prob- ably terrestrial. The evidence is not at all conclusive, but it warrants the tentative conclusion that this fauna is largely arboreal, which is well in accord with the evidence that the quarries were in a swampy and heavily forested area and would go far toward explaining the unusual facies of the quarry faunas. There is, indeed, a decidedly fossorial humerus (of unknown association with teeth) in the collection; the facies association of arboreal and fossorial animals is not uncommon and is in accord with a forest environment. The ordinal composition of the Scarritt Quarry collection differs significantly from that of the Gidley Quarry only in the almost com- plete absence of carnivores. Within the other orders, the family com- position is as nearly similar as would be expected in deposits of similar facies but different ages except among the Primates. The abundant Gidley Quarry types, Paromomys and its closer allies, are not repre- sented in the Scarritt Quarry collection, and instead of them the more specialized, perhaps more strictly frugivorous, plesiadapids and car- polestids have become fairly common, although the first were un- common and the latter very rare in the Gidley Quarry. The Gidley Quarry is also interesting from the unusual occurrence of its fossils and the indications of the possible conditions surrounding death and burial of its animals. The remains are invariably frag- mentary, and with extremely rare exceptions there is no association of specimens. The bones seldom show any signs of weathering or rolling but are usually fractured, and even when they abut against wholly undisturbed matrix these fractures are clean, fresh breaks. Some further fracturing and dissociation have resulted from the com- pacting of the bed and development of slip planes, but for the most part these preceded fossilization. Most of the jaws have lost some teeth before burial, and many have lost all the teeth. These isolated teeth (clearly lost after death but before burial) are common in the collection. There are many bone fragments, but it is clear that the quantity of skeletal material present, even in the most fragmentary state, cannot by any means represent all the bones of the animals represented by their jaws and teeth. The rather abundant presence of fish remains, often in articulation, and of aquatic reptile fragments and the presence of aquatic mollusks (rare in this quarry, but present), together with the nature of the sedi- ments, suggest that the deposit was formed in sluggish water, perhaps a 64 BULLETIN 169, UNITED STATES NATIONAL MUSEUM swampy stream course, ox-bow lake, or bayou. From the great va- riety of mammals present this evidently was not the site of a single or selective catastrophe, like many quarries that seem to represent quick- sand or quagmire traps, but must have made a fairly complete sample of the mammals of the surrounding forest and (to a less extent) glades. Regardless of whether the mammals came here to drink, swam into the water, dropped from trees, or were occasionally washed in, it seems likely that the breaking and scattering of their bones, and perhaps commonly their deaths also, were the result of activities of the carniv- orous fishes and reptiles. Such a history would probably explain the small ratio of bones to teeth (the former eaten and digested and com- minuted, the latter less palatable and more resistant), the many clean breaks, lack of association, and also the common intervention of macer- ation, without apparent weathering (perhaps in part digestive, and otherwise subaqueous) between death and burial. THE GIDLEY QUARRY AND ECOLOGICAL INCOMPATIBILITY Matthew (1930) has stated that ‘‘we should expect to find in a single fossil quarry that the material of each genus represents a single ecologic niche, or, if more than one, that they are quite distinct. We should not, in other words, expect to find two or more closely related species living together at the same time, within the same area, and with the same habitat, causing their remains to be preserved together in the same quarry . . . Hither there would be two or more species so widely different as to belong in obviously independent ecologic niches, or else there would be one more or less variable species.” This is an application to paleontological data of the general principle summed up by Cabrera (1932) as the Law of Ecologic Incompati- bility in these words: ‘‘Las formas animales afines son ecologicamente incompatibles, siendo su incompatibilidad tanto mayor, cuanto mas estrecha su afinidad.”’ !§ The Gidley Quarry fauna is ideally adapted to the application, on one hand, and to the exemplification and corroboration, on the other, of this law and of Matthew’s remarks on the taxonomy of quarry faunas. The species present in it were certainly contemporaneous, and it is highly unlikely that any of the remains can have been brought from a point so distant as to have inhabited distinctly different areas. The general environment was probably essentially the same for all, although unquestionably it included distinct ecologic niches. It is possible that deposition extended over a period of years and that there was some seasonal or other periodic change in the species of the 18 I do nct wish to claim for Cabrera a degree of originality that he disclaims for himself. Cabrera’s law has been recognized in various forms by many authors and for a long time, but as far as I know it has never been more exactly explained and exemplified or placed more clearly on its true ecologic (not geographic) basis than by Cabrera. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 65 neighborhood, but this is purely hypothetical. It is most reasonable to conclude that all these animals did live together, at the same time, within the same area. It is therefore to be expected that genera present in this quarry will either have only one species each or will have species not inter- grading at all and reflecting structurally their pertinence to different ecologic stations in the area. With this in mind, the assumption was made in dealing with each genus that it did include (in this quarry fauna) just one species unless the contrary could be proved beyond reasonable doubt. Matthew, in the paper cited above, and most other writers on the question of species making in paleontology have insisted on making due allowance for variation, or using for taxonomy only nonvariable characters, but they have adduced no real, objective criterion as to what ‘due allowance” should be, and they sometimes seem to over- lock the fact that there is no such thing as a truly and completely ‘nonvariable” character. Not merely as mechanical, mathematical procedures but as a general system of logic and a grouping method useful both explicitly and as an implicit background for dealing with both numerical characters and attributes, the methods of statistics provide the desired means of measuring variation accurately and the necessary criterion as to whether this variation is or is not of the sort normal within a species. These tests and this logical background have been the basis for taxonomy in this study. If the specimens pertaining to one genus could not indubitably be separated into differ- ent groups, the conclusion has been that the fundamental hypothesis of one species to each genus was correct. If they necessarily had to be separated into different groups, and these groups could not be interpreted as based on nontaxonomic differences (such as age or sex), then and only then has the hypothesis been discarded. Since this largely objective testing has intervened, it is not arguing in a circle to start the study with the assumption that Cabrera’s Law applies, and then to consider the results as a test of the validity and an example of the operation of that law. (See figs. 3 and 4.) The actual results are as follows: Thirty genera are represented in the Gidley Quarry by one species each. Since this includes the greater part of the quarry fauna, in general it is true here that the related (congeneric) animals living at that time in this area were of the same species. The apparent exceptions belong to six genera, each of which is worthy of brief special consideration. The multituberculates, with four species tentatively referred to Ptilodus and three to Ectypodus, are the most striking apparent exceptions. In the first place, however, the generic designations are very dubious. It is not at all certain that the species referred to 66 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Ellipsodon aguilonius Litomylus dissentaneus Litaletes disjunctus Haplaletes © disceptatrix 2.35... 275 3.15 3.55 395 4.35 LENGTH Mi, HYOPSODONTIDAE P.maturus 205.7 M245" 2.85 S125 LENGTH M2,PAROMOMYS FiGcurE 3.—Histograms illustrative of Cabrera’s law as applied to the Gidley Quarry fauna: a, Length of M: in hyopsodcntids: in this and some other measurements the species nearly or quite intergrade, and they lived in the same place at the same time, but they are morphologically very distinct and belong to four different genera; b, length of M2 in Paromomys: the two species are morphologically similar, probably are congeneric, and are found together, but in this dimension and in many other characters they do not intergrade. M. provocator M. punitor “at atararararata Number of individuals — odo BUN Zo 48, 4649 52 55 55 61 LENGTH Moe S$ _Quorry Specimens. M. punitor Other Upper Lebo Specimens, Or Peer, EEA One specimen in each sample. 40 43 46 49 5.2 WIDTH M2 Number of individuals FicurE 4.—Histogram of measurements of second lower molars of all Lebo specimens of Metachriacus. The distributions of these measurements, which are distributed much as are most of the characters of the samples, are clearly bimadal, and two entities are shown to be present. These entities intergrade, but they are separated by provenience, the two coming from different horizons and localities within the Lebo. They are thus interpreted as closely allied but separable species. If they occurred together, a more probable interpretation would be that they represent the two sexes of a single species. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 67 Ptilodus or to Ectypodus are really congeneric in each case. There may well be one or more other genera represented here, although it does seem unlikely that each species could belong to a different genus. Aside from this possibility, the species referred to one genus are in every case sharply distinct from each other, with no intergradation shown. ‘This is true of many different characters, and is well shown by table 5 of observed ranges of a few numerical characters: TasLe 5.—Ranges of numerical characters in multituberculates, Gidley Quarry, Crazy Mountain Field, Mont. Genus and species LP, LM; LPy:LM;, Serre oue Cusps M,; Ptilodus: Mm Mm MLOMUEANALS = See Sees Renee ees 7.1-9.1 3. 2-3. 7 2. 3-2. 4 13-15 5-6:4-5 GCOUGLASSin eee eee Ne 6. 5-6. 8 3.7 1.8 13 6:4 Gidley cae. -Ba kek cei SS SDT A 5. 9-6. 1 2.5 2.4 14-15 SUUCLOU INP ee eR Naat ee oe ee ee 2. 5-3. 7 1. 7-2.0 1,3-2.0 10-13 6-7:4 Ectypodus: OT ONGC eae eee eee 5. 2-5. 4 3. 3-3. 4 1. 5-1..6 13-14 8:6-7 NLSSCLLY See ARE hee ee ee 28 4,9-5.1 2.9 nef 13-15 10-11:6 STL OES LIG tee eee ee e e 3.3 2.3 1.4 12 9-10:5-6 Thus these species probably represent more than two genera and in any case are so sharply distinct that each must have had its own ecologic niche. They do not intergrade, but in some cases, notably P. montanus, there are known species, in this case P. mediaevus, with which they do nearly or quite intergrade but which did not live together with them. In short, these are not exceptions to but are striking exemplifications of Cabrera’s law. Lepiacodon is here credited with two species. Their tooth dimen- sions do not intergrade, although the degree of variation is well established for one of them, and there are structural differences such that they may well prove to be distinct subgenerically, or even generically, when both are more completely known. This is even more strikingly true of the two species referred to Paromomys, which are so different that I was for a time inclined to separate them generically (see fig. 3, 6). Didymictis also is represented by two species perhaps not really congeneric, not intergrading at all and one reaching a size nearly twice that of the other. In the case of Claenodon, Gidley has been followed in listing three species from the Gidley Quarry, all of about the same size and general character. Differences between them do exist, but the material is inadequate to establish the extent of variation, and when this is established I suspect that the supposed three species will prove to be variants of a single species. If not, this will be the unique example of the occurrence of two or more closely related and apparently ecologically similar species in the quarry. 68 BULLETIN 169, UNITED STATES NATIONAL MUSEUM EXTENT OF KNOWLEDGE OF MIDDLE AND UPPER PALEOCENE FAUNAS Aside from details concerning only special students, the impor- tance of such collections as those here described lies, from a more general point of view, in the knowledge they give of the broader out- lines of mammalian life in their time and area. To permit the proper drawing of inferences in this broader field, it is essential to consider the adequacy of the collections and their probable relationship to the faunas as a whole represented by them. The adequacy of a paleon- tological sample depends principally on three quite distinct factors: 1. The adequacy of the collection studied as a sample of the fos- sils actually preserved in the rocks. 2. The adequacy of the preserved fossils (collected or uncollected) as a sample of the whole fauna that actually lived in the area. 3. The adequacy of the real local fauna as a sample of the regional fauna of the whole land mass on which it lived. Probably the best criterion of the adequacy of a collection as a sample of the preserved fossils is that of repetition. When collecting begins to pile up mainly or only duplicates, it probably has achieved sampling adequacy for the local deposit, but as long as many species remain very rare in collections, it probably has not. Of the 51 surely separate species known from the Gidley and Silberling Quarries, 15 are here represented by only one specimen each, but of these four are known by other specimens from elsewhere in the field. Seven here have only two specimens each, but one is fairly common elsewhere. Six are here known from three specimens each, but two are also represented elsewhere. The other 23 species are represented by five or more specimens each from these quarries. In the field as a whole, of the 79 species, 19 are represented only by one specimen. Four of these represent genera still unknown elsewhere, and hence known from but one specimen each: Stilpnodon simplicidens, Unuchima asaphes, Elphidotarsius florencae, and Spanoxyodon latrunculus. Of the other 15 species here represented by only one specimen each, three (Leptacodon cf. tener, Psittacotherium multi- fragum, and Thryptacodon ?australis) are inseparable from species well known in other fields, and the others all belong to genera well known from other species, some of them abundant. For broader studies of morphology and faunal succession, local species are not very important, and of the 51 genera known to occur in this field, not more than five can be considered as very poorly represented in Middle Paleocene faunas generally.” From these data, as well as the general make-up of the collections and other considerations, the collections appear to represent the real 19 At least one of these, Hlphidotarsius, and possibly one or two others are closely allied to well-known genera. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 69 fauna * of the region fairly well but far from completely. The high proportion of species known here from only one specimen each, about 25 percent, shows that the collections, large as they are, are not yet adequate to give a proper sample of the species that may be pre- sumed actually to be available here as fossils. Such a high percent- age of very rare forms indicates clearly that any further collecting is almost certain to add to the number of known species. As far as inference can be drawn from these data and from ecologic and similar considerations, the situation regarding the whole pre- served (but only partly collected) sample, that is, the forms that were actually fossilized, is much better, but not perfectly satisfactory. It may reasonably be assumed that there was in this general region some variety among the larger and possibly plains-dwelling types, such as are common in the Torrejon and also in the later Paleocene and early Eocene in adjacent areas, but their great rarity or absence here and the general facies present strongly suggest that the condi- tions of deposition were such that some and perhaps many of these were not preserved and so will never be known in this area, no matter how large the collections made. On the other hand, knowledge of the general composition of the Middle and Upper Paleocene mammalian faunas of North America as a whole may now be considered very good. It is probable that we have representatives of almost all the orders and families and a large majority of the genera,”! that occurred on this continent during that time. The combined area represented by collections is now very considerable, on the order of 1,000 square miles of actual collecting territory, representing many times that in the ranges of sampled faunas. The environmental variety represented is apparently great, for the sediments yielding mammals of these ages are of many dif- ferent sorts, many genera are represented by several well-defined species in each, and the inferred habits of the various known mam- mals include almost every possible terrestrial mammalian habitus. The collecting areas certainly were part of a unified North American land mass in the Paleocene, extend more than 1,200 miles north and south, and were probably central on that land mass, ideally situated for a representative sample of the whole North American fauna. 20 It would be rather hopeless to crusade against the universal and careless habit of calling a collectiona fauna. By “real fauna’’ is meant what should properly be called simply‘“‘fauna’’; that is, the totality of mammals that actually lived in this area, and not merely those that happen to be known or the collection, a sample, on which this knowledge is based. 21 In accordance with the tentative views as to adequacy of local samples expressed elsewhere (Simpson, 1936a). PART 2: CLASSIFICATION AND DESCRIPTION OF MAMMALS Order MULTITUBERCULATA Cope, 1884 Douglass’ first collection included no multituberculates, but in his second collection (see Douglass, 1908) there were several teeth of this group. ape dike tee alm pee 1% Lt c% 8'P T@ LG GZ o% 1% 1 AR Asa | cmp nae oe et 9209 ‘on "WN'S'1 Wy wy Wy Uypy wy Up wy Uf Wy Uy wy Up M T AA a M Al M Al M T M Al WW sdsno}i jw sdsno| +g sdsno | eq sdsng | | ee uemoedg W WA rd ed td Id (oa 4 ¥Or 22. |e eae a cit o% FEZ (aad 9% 9'T g's C Su [ule [ae aay ee Sea aR a Se a an ee 6209 “OU "FNS" F:9 A Ae 0% Lh 5 alae (i twice oa axl! FE CUR: Home| (Waa gn ta gids Os ce Co. eel SEO ee = tnd Dako €L9T “OU "JA a1seuIB +) oP $9 as PAE 0% FeaGy “Slee aad nee (are 9'T (a3 a (aie ai Bae Mba Sees Sa! ok eee aT 9209 “00 "WoN'S' uy Wy Uy Uy Wy 2 rg suo | WT TAM AT tw sdsno|'W sdsno} “Soii9g ar ae ae ee ee IT uamjoedg ‘WH TAL snueyUOU snpollyg fo suawrads yonprarpur fo sjuawmainspayy—'g] AIAV FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 95 ?7PTILODUS DOUGLASSI Simpson FIGURE 9, a ?Ptilodus douglassi Stupson, 1935d, p. 225. Type.—U.S.N.M. no. 9795, right lower jaw with Ps-M>. Collected by A. C. Silberling. Horizon and locality —Gidley Quarry (and two referred specimens from Silberling Quarry), Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis.—Mean length P, (three specimens) 6.6, deviation from mean of P. montanus —1.4, 3.3 times standard deviation of the latter. Length M, (type) not significantly deviating from that of P. montanus. Ratio length P,:length M, (type) 1.8, deviation from mean P. mon- tanus —0.5, 12.5 times standard deviation of the latter. Length M,:length M, (type) 1.8, deviation from mean P. montanus+-0.4 (standard deviation not calculable, but difference almost certainly significant). Serration and cusp number not significantly different from P. montanus. Remarks.—Expressed in other and less definite words, the species is structurally closely similar to P. montanus but differs significantly in its smaller size, relatively much smaller P,, and relatively smaller M, (or, much larger M, relative to Py and M,). P, also appears to be somewhat lower in lateral contour, but this cannot be adequately checked. No species other than P. montanus resembles this closely enough to demand further comparison. TasLe 14.— Measurements of individual specimens of Ptilodus douglassi a a LP M, | LM; | 8 5 4 | LM Mi | *!Ta | Cusps | Cusps U.S.N.M. no. LP, LM, | WM, | LM Hons M, M: L W L W Mm | Mm} Mm| Mm | Mm Viti see ee 6.5 3.7 UBT 2.0 2.1 1.76 Die 1.8 13 6:4 4:2 ORSS Reese ee eee GGh eee ae ere aA ee ee |e ee Ee cone lees aeee 713 OS88p mee ee eee 658) | Ses | Eo See ee SSS ee ee |e a eee eee oe 13 2PTILODUS GIDLEYI Simpson FicureE 9, 5 ?Ptilodus gidleyt Simpson, 1935d, p. 225. Type.—U.S.N.M. no. 9763, left lower jaw with P, and broken Mj. Collected by A. C. Silberling. Horizon and locality —Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis —Length Py, type, 6.1 (two other specimens 5.9 and 6.1). Length My, type, about 2.5. Ratio length P,:length M, about 2.4, 96 BULLETIN 169, UNITED STATES NATIONAL MUSEUM that is, not significantly different from P. montanus or mediaevus but much higher than in any other known species of this or closely related genera. Serrations P,, type, 14 (two other specimens 14 and 15). P, very short and high in lateral contour, rising well above the grinding plane of M;. Remarks.—Except for its somewhat more elevated Py, this closely resembles P. montanus in its known proportions and other morpho- logical features, but it is much smaller (deviation of length P, from mean of P. montanus —2.0, nearly five times standard deviation of the latter), too much so to consider it a small variant of montanus, despite the considerable variability of the latter. a b FIGURE 9.—Ptilodus, comparative outline drawings of lower dentition: a, P. douglassi Simpson, U.S.N.M. no. 9795; b, P. gidleyi Simpson, U.S.N.M. no. 9763; ¢, P. sinclairi Simpson, U.S.N.M. no. 9770. Crown and external views, all drawn as if left lower jaws (a and ¢ reversed from right lower jaw). Three times natural size. P, of this species is of the same size as in P. trovessartianus. Ma- terial is insufficient for full analysis, but if we assume the variability of both not to be much greater than in P. montanus, the shorter M, of ?P. gidleyi * and the resulting larger ratio length P,: length M, are surely significant. These same differences distinguish it even more sharply from the somewhat larger species ?P. douglassi. No others resemble it closely. TaBLE 15.— Measurements of individual specimens of ?Ptilodus gidleyi LP : U.S.N.M. no. EP; LM, cM Sores Cusps M, Mim Mm | Gap aite see ea See: BM REPRE FP IE 550 || eee | ew oe 15 1g 3 a pe el eS E.Sun 6.1 ea. 2.6 ca. 2.4 14 76:? it (1 aed tara enn OR PON. Bia | any arn ee oe SPE I 14 46 This is not exactly determinable from the specimen, but the error of measurement can hardly exceed 0.2 mm and is almost surely less. Alveoli in a referred specimen also indicate a very short Mi. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 97 An isolated P, from Loc. 50 (American Museum collection) measures 6.1 mm in length, has 15 serrations, and closely resembles the type of this species in form. There can be little question that it belongs here, as it is so close to the known mean for this species and far outside the range of any other species recognized in this field. ?PTILODUS SINCLAIRI Simpson FIGURE 9, c ?Ptilodus sinclairi Simpson, 1935d, p. 225. Type.—U.S.N.M. no. 9770, left lower jaw with P,-M>2. Collected by A. C. Silberling. Horizon and locality—Gidley Quarry (referred specimens from Silberling Quarry), Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis—Length P,, mean 3.1+40.07, standard deviation 0.29+0.05. Length M,, mean 1.9+0.03 standard deviation 0.08. Ratio length P,: length M,, mean 1.6+0.08 standard deviation 0.22 + 0.06. Ratio length M,: width M,, mean 2.2+0.06, standard devia- tion 0.17+0.04. Serrations P, 10-138, mode 12. Cusps M, 6-7:4, mode 7:4. Remarks.—Some of the peculiarities of this species, second only to P. montanus in abundance, have already been discussed above. Its very small size at once distinguishes it from any species of related genera except ?Ectypodus silberlingi (diagnosed below), and Parecty- podus tardus, with which it cannot be congeneric, as it has P3. The two specimens (6089 and 6090, and also a third, 6149, so labeled but not published by number) that Gidley at first (1909, p. 623) re- ferred to “Ptilodus formosus?”’ (Marsh) (=Halodon formosus Marsh) belong to ?Ptilodus sinclairi.® Adequate comparison with the frag- mentary Cretaceous types of Marsh is impossible, but in view of the very different age and of the fact that when close comparison is pos- sible not only the species but also the genera are very distinct, it may, I think, be assumed that the Fort Union forms do not belong to Cretaceous species. As noted above, it is improbable that this species belongs to Ptilodus, but it cannot at present be clearly distinguished from that genus. 45 Gidley did not change their labels, but it is practically certain that he recognized their pertinence to a distinctive species. He recognized ? Ptilodus sinciaizi (under a different, unpublished name), and so labeled about half the specimens that I piace here, covering almost the same range of variation, so that in this case his specific criteria and mine lead to nearly the same result. He also included, however, one or two speci- mens that I place in other species. 98 BULLETIN 169, UNITED STATES NATIONAL MUSEUM The principal numerical data on this form are as follows: TaBLE 16.—Measurements of ?Ptilodus sinelairi: Continuous variates Variate N R M o Vv | 2 i ee ae 15 2. 5-38. 7 3. 130507 |-0. 29 +0. 05 Oro =e 116 1 ign Se 8 1. 7-2. 0 1. 88+0. 03 | 0. 083+0. 021 a Sea ig (ee | LPs My! 8 1. 32-1. 95 1.61+0. 08 | 0.22 +0. 06 LM,: WM,_-_- 8 1. 9-2. 4 2.25+0. 06 | 0.17 +0. 04 Tasie 17.—Distributions of discontinuous variates: ?Ptilodus sinclairi fi Internal ‘cusps: Mis = 2Ss58e2 Ais Wea, P 1 eo eEixternall cuspse Motes 2s 25 ae An ela3 Serr is Py: a SO 1225. valintermalecusps Moss. a ae =e 22 ES [een (G22 223 ct ‘usps M | ee ae ea h External cusps M abe’ i Mi Ms Serra- U.S.N.M. Dp LP, LM: LM; : Cusps | Cusps no. ue DM, | WialeGMy | oe) Mie lees ral! We ic A ee sine menal Petes Mm} Mm| Mm)| Mm| Mm OO TEESE Soe 2.6 1.9 0.8 0.9 0.9 L387 2.4 Ppa 10 7:4 4:2 LAYS ae ee ane es 229i 9 0.8 i 0.9 1253 2.4 Mei 12 ? ? DFO oe ea ee Boa) Dad 0.9 0.9 0.8 1. 94 1.9 1.9 11 6:4 4:2 O77 OLS sese aoe eee ee Me 3 Heir 1.9 0.8 0.9 0.9 1.95 2.4 Pal 12 6:4 4:2 These four best specimens include practically all the extremes in these characters. A P, collected by A. C. Silberling on January 21, 1903, now in the Princeton University collection, measures 3.0 mm in length and has 12 clear serrations and 1 obscure. It is indistinguishable in any way from some of the smaller variants of Ptilodus sinclairi from the Gidley Quarry and is referred to that species. Nothing like it has been dis- covered at a Lower Paleocene horizon, nor is it exactly matched in any of the many Lance specimens (Cimolomys, sensu lato) known to me. It bears the horizon designation ‘‘Puerco”’ (by which was then meant approximately the series later called Fort Union No. 1), the number 14, and the locality ‘“Moen and Beck”’ (a very general locality including a large area along Bear Butte). Mr. Silberling informs me positively that this is from Loc. 65, and this can be taken as correct. The specimen is thus the oldest mammal ever found in this field. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 99 Genus ECTYPODUS Matthew and Granger, 1921 ?7ECTYPODUS GRANGERI Simpson Figure 10, a ?Ectypodus grangert StmPson, 1935d, p. 226. Type.—U.S.N.M. no. 9801, left lower jaw with P,-M,. Collected by A. C. Silberling. Horizon and locality —Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis—Length Py, type and mean of four specimens, 5.3. Length M,, type and mean of three specimens, 3.4. Ratio length P,: length M,, type 1.56, referred specimen 1.53. Length M,: width M,, type 2.6, mean of three specimens, 2.4. Serrations P,, type 14, other specimens, two 14, one 13. Cusps My, type, 8:7, other specimens one 8:7, one 8:6. Apex of P, nearly on a level with grinding surface of Mi. Remarks.—Like the other two new species tentatively referred to ?Hctypodus, below, this differs from all species referred to Ptilodus in the large length: width ratio and large cusp number of M,, characters that appear to be surely significant and make closer comparison with species of Ptilodus unnecessary. ‘This, and the next two species, are excluded from Parectypodus by the presence of P3. The present form compares very closely with Hctypodus musculus in all its visible characters. In spite of the small size of the samples, the considerably greater size of ?E. grangeri is almost certainly significant, and its association with wide geographic separation and with a dis- tinct difference in age shows the species to be distinct. The principal measurements are given in table 19. TasLe 19.—Measurements of individual specimens of ?Ectypodus grangeri U.S.N.M. xe LPs LM, | Sera | Gusps aay LP, LM wM tions M, 16 Ww 1 1 Py Mm Mm Mm O80 0 eee ot oi Lo weet nS 5:2 3.4 1.6 1.53 2.1 14 8:6 OS OTE Renee Eh. 5 Sik ees 5.3 3.4 1.3 1. 56 14 8:7 OU a SAS eee 3 ee edo 5 see SC. 5 ee | ears eee aed (ep ene wee ee 13 OS 2a eee ee a anes ae 4 ease 2S eee 2 Bee ee bees oe 14 tf SS Se es ee ee ee (ope te 3.3 DBs Ye eee ae 255i) 2k Ae 8:7 | ?ECTYPODUS RUSSELLI Simpson Ficure 10, b ?Ectypodus russellt Simpson 1935d, p. 226. Type.—U.S.N.M. no. 9765, left lower jaw with PM. Collected by A. C. Silberling. 100 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Horizon and locality.—Gidley Quarry, Fort Union, Middle Paleo- cene horizon, Crazy Mountain Field, Mont. Diagnosis —Length P,, type and mean of three, 5.0. Length M,, type and referred specimen, 2.9. Ratio length P,y:length M,, type and referred specimen, 1.7. Length M,:width M,, type 2.4, referred specimen 2.6. Serrations P,, type 14, others one 15, one 13. Cusps M,, type 10:6, referred 11:6. Cusps Mg, type and referred specimen 5:2. Crest of P, elevated well above M,. MITT) Ca / Dereewy IG p d FIGURE 10.—Ectypodus and Parectypodus, comparative outline drawings of lower dentitions: a, H. grangeri Simpson, U.S.N.M. no. 9801; 6, EH. russelli Simpson, U.S.N.M. no. 9765; c, EH. silberlingi Simpson, U.S.N.M. no. 9798; d, P. jepseni Simpson, U.S.N.M. no. 9769. Crown and external views, left lower jaws. Three times natural size. Remarks.—This species considerably resembles the preceding one, and I was at first inclined to consider its smaller size as due only to variatiow, as it is well within the possible size limits. That would, however, make the range in M, external cusp number for the com- bined species twice as great as has been demonstrated for any known species of this or allied genera, even in much larger samples, and the probability is very greatly against this occurring in a single species and against the association of larger cusp number with smaller size being due to accidents of sampling. It is possible also, although less clear, that association with a slightly larger length P,:length M; ratio is significant. The added character of a more elevated P, makes the specific distinction practically certain. P, of this species very closely resembles that of Ectypodus coch- ranensis. However: (a) It here has a definitely larger notch for P;, (b) the localities are well separated, (c) the ages are distinctly FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 101 different, (d) only P, can be compared, and (e) extremely slight differ- ences in this tooth are commonly associated with specific or even generic differences in other parts of the dentition.” In size this species is a little closer to H. musculus than is E. grangeri, although still definitely larger, but its other distinctions from £. grangert are distinctive in about the same degree from EF. musculus. A tooth of this species (no. 6088) was referred by Gidley to ‘‘Ptilodus serratus’? (=Halodon serratus Marsh) in his preliminary publication (1909, p. 622) but with the reservation that it might later be placed in a new species. Had he completed his work, Dr. Gidley would unquestionably have placed this specimen in a new species, as is done here. The remarks made above, regarding specimens of ?Ptilodus sinclairi referred tentatively to ‘‘Ptilodus formosus’’ are equally apropos here. TaBLE 20.— Measurements of individual specimens of ?Ectypodus russelli Mi M; Serra- pena sect [ee ei ay |g Bl 2 LM LM : Cusps | Cusps U.S.N.M. no. LP, en tions Te peel eee epyrt) HM 9) TOW east tops, M1 M: Mm | Mm | Mm | Mm | Mm | Uses 4.9 2.9 1 1.6 1.5 1.6 1.8 13 11:6 5:2 O(G0)ssee sees een 5.0 2.9 1.2 | 1.5 1.4 1.72 2 1.9 14 10:6 5:2 BOSS eee el ese 5 fll eee | Wee eles eee [ovoescoe fecseseee) eee 15 | ?ECTYPODUS SILBERLINGI Simpson Figure 10, c¢ ?Ectypodus silberlingt Simpson, 1935d, p. 226. Type.—U.S.N.M. no. 9798, left lower jaw with incisor and Py-Mg;. Collected by A. C. Silberling. Horizon and locality Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis.—(Only one specimen.) Length Py, 3.3. Length M,, 2.3. Ratio length P,:length M,, 1.4. Length M,:width M,, 2.6. Serrations P,, 12. Cusps M,, 9-10:5-6.% Crest of P, relatively low. Remarks.—In most of its characters, this species falls well within the range of ?Ptilodus sinclairi, with which it would be confused in casual examination, but its deviation from the mean of sinclairi in length of M, is 5 times the standard deviation of the latter, and in the ratio length M,:width M,, 2.3 times, and the marked difference in cusp number of M, is also surely significant. The species cannot be, and the genera probably are not, the same. 47 Russell (1929, p. 173) in describing EF. cochranensis was doubtless referring to my ?E. russell! when he wrote, ‘‘Undescribed premolars of about the same size as EH. cochranensis occur in the Fort Union beds of Montana, but these teeth differ markedly from the present specimen in having a prouounced undercutting in the anterior margin of the crown.” 48 That is, clearly at least 9:5, and in each row rudiments of another cusp so that it is doubtful whether the count should be 9:5 or 10:6. 102. BULLETIN 169, UNITED STATES NATIONAL MUSEUM Among other species, it most closely resembles Ectypodus musculus, but aside from the different horizon and locality it differs in being somewhat, perhaps significantly, smaller and in having more cusps on M,, as well as other minor distinctions that may prove also to be significant when a larger sample permits their evaluation. Measurements of the type other than those given above: Width M,, 0.9; length M>;, 1.2; width Ms, 1.0; ratio length M,:length Mb, 1.9; cusps Ma, 25:3. ?ECTYPODUS species U.S.N.M. no. 9772 includes three unassociated last upper premolars, one broken, from the Gidley Quarry. They have one complete row of 11 cuspules, or serrations, and a rudimentary anteroexternal basal row of 2 cusps, both strong and distinct, the more posterior larger and opposite the third or fourth cusp of the main row. These teeth suggest Hetypodus musculus, but the elevation of the posterior end of the main crest is less than in that species, and the apex is not on the most posterior cuspule but on the fourth or fifth from the posterior end. Neoliotomus conventus is inadequately known, but apparently these specimens are less symmetrical in lateral contour. Size and cusp number surely distinguish the species from any comparable form, and the generic reference is doubtful. The lengths of the two more complete teeth are 4.8 and 4.5 mm. These probably represent the upper dentition of one of the species bere named from lower jaws, and by inference somewhat more prob- ably one referred to ?Hctypodus, but the association cannot be estab- lished. From the ratio length Py,:length P* in Eetypodus musculus (1.48), the lower P, should be 6.7—7.1 mm in length, about as in ?Ptilodus douglassi, but the latter has no characters suggesting Ectypodus. ?2Ectypodus grangeri may be based on the corresponding lower teeth, but it seems somewhat too small. Genus PARECTYPODUS Jepsen, 1930 ?PARECTYPODUS JEPSENI Simpson FicureE 10, d ? Parectypodus jepsent SIMPSON, 1935d, p. 227. Type.—U.S.N.M. no. 9769, left lower jaw with P,-M,. Collected by A. C. Silberling. Horizon and locality —Gidley Quarry, Fort Union No. 2, Crazy Mountain Field, Mont. Diagnosis.—(Only one specimen.) Length Py, 4.3. Length M,, 3.1. Ratio length Py:length M, 1.4. Length M,:width M,, 2.2. Serrations Py, 11. Cusps M,, 7:6. P,long andlow. No notch for P3. Remarks.—This can be compared only with the later Parectypodus simpsoni. Its longer M,, lower length P,:length M, ratio, smaller FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 103 serration number, markedly different cusp formula of M,, and other lesser differences show it to be markedly distinct. It differs even more, but in somewhat different ways, from Parectypodus tardus. The generic reference is very dubious. Measurements of the type other than those given above: Width M,, 1.4. Genus EUCOSMODON Matthew and Granger, 1921 EUCOSMODON SPARSUS,‘! new species Figure 11 Type —U.S.N.M. no. 10113, part of lower incisor, with a small adherent jaw fragment. Collected by A. C. Silberling. Horizon and locality —Loc. 25, and referred specimens from Loc. 51 and the Silberling Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis.—Type incisor, maximum transverse diameter 4.0, mini- mum 1.4 mm, ratio 2.86. FIGURE 11.—Eucosmodon sparsus, new species, U.S.N.M. no. 10113: Part of lower incisor and fragment of jaw. External view and cross section at anterior end of incisor as preserved, with enamel band shown in heavier outline. Four times natural size. Remarks.—This is an interesting form worthy of formal record despite the imperfection of the material. Poor as this is, it fulfills the practical requirements of demonstrating distinction from any comparable species and ensuring that better specimens, when found, can be securely determined as of this species.°° The occurrence of three essentially similar specimens from three different localities and horizons, but all within this field and all in the No. 2 beds, covering a short span of time, also seems to demonstrate the validity and con- siderable range of the species. U.S.N.M. no. 9861 is a broken incisor from the Silberling Quarry and no. 9705 a similar specimen from Loc. 51. Their dimensions, given below, agree closely with those of the type. In maximum diameter these teeth approach FE. americanus primus and are significantly smaller than other comparable species. In minimum diameter they are somewhat less and in compression ratio 4° Sparsus, scattered, from the dispersion of the scantly known remains of the species. 50 Such specimens are almost surely included in the American Museum collection, but they are not yet prepared, and in any event it is proper that types be from the older collection, as far as possible. 119212—37——_8 104 BULLETIN 169, UNITED STATES NATIONAL MUSEUM significantly more than F. a. primus. The minimum diameter is about that of E. teilhardi, but the maximum diameter and the compression are significantly less. HH. americanus and E. molestus have the incisors significantly larger in both diameters and the compression less. Exact comparison with E. gratus Jepsen is not possible, but the incisor alveolus of that species measures 2.7 by 1.1 mm, ratio 2.45. The alveolus, and of course therefore the incisor as well, are much smaller than the incisor of EL. sparsus. No known species aside from the several now mentioned could well be so closely related to EH. sparsus as to demand differentiation. TaBLE 21.—Comparison of data on lower incisors of Eucosmodon Diameters Species and specimen ——_—_—_—_______—_————_} Ratio Maximum |} Minimum Mm Mm RC INETUCON US AU YD Cs ee er ey eee 6. 0 296 Dar, MOT DTLMUS ApalTaAly pes. a2 ee ae eee ee eee am 4,2 1.9 2. 21 EE AMmolestus, MOLY Pe! -< = ashe as oe ee Soe 5. 8 2.6 2. 23 Evtedhardt, paratypes. Use to seals ees eee 52 1.6 3.25 Mean Sih i274 Sees tae ree eg Mal 4,3 15 2. 70 OURS Seo Bt ek Le a ree oe eee 4.0 1.4 2. 86 PODORMEN Gin: eee Yh Oc A A pe ee 41 16| 2.56 OSGi oe nico be ee Sans cern seas 4.0 easy 2. 67 (#.. graius, type, from alveolus) sa. eee (O26) (1. 1)/(2. 45) The original of Douglass’ plate 1, figures 18 and 20, probably belongs to this genus. If the scale of the figure is accurate (which is not invariably true in cases where J have studied Douglass’ originals), the maximum diameter is about 3.7 mm, smaller than other specimens referred to this species, to which the specimen may nevertheless belong. Order INSECTIVORA Gray, 1827 Forms that are at least nominally referable to the Insectivora con- stitute an important element in Paleocene and Eocene faunas, and this is particularly true in faunas, like that here described, in which the microfauna is well or disproportionately represented. ‘They are abundant in the present collection, including at least 10 species, all of which occur together in the Gidley Quarry. Despite diverse specializations, such as in the incisors of the Sori- coidea or the limbs of the Talpoidea, the recent insectivores are on the whole the most primitive of living placental mammals. The characters common to all of them, excluding the specializations of various particular phyla or larger groups, are in general those unques- tionably primitive for all placental mammals. The conception of the order and its definition thus do not so much depend on special FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 105 characters developed within the order as on characters also basic for other orders but generally lost or more profoundly modified in those orders. Indeed the outstanding ordinal character of the Insectivora, paradoxically, is the absence of ordinal characters, in the sense that these are developed in other groups. Although the order is thus necessarily more loosely knit than is usual, it does not follow that it is altogether artificial. The retention of so many primitive characters is in itself some indication that this conservative group may have still been somewhat unified after the more progressive orders had acquired their distinctive characters. Furthermore, there are a few characters, such as those noted by Matthew in the astragalus, not of wholly generalized placental pattern and distinctive of the Insectivora as against many or most other mammals. There are also some characters, like the peculiar specializations of the incisors, that do not occur in all insectivores and are occasionally paralleled in other orders but that nevertheless appear so frequently among insectivores that they seem to reflect a certain genetic tendency and to help to bind the group together. Every individual living insectivore is a specialized animal, each species in its own way, and none can be considered, even structurally, as really representing the general placental ancestry in any very exact sense of the words. Yet the abstract conception Insectivora based on all known forms is such that it would, almost perforce, include the most primitive placental mammals. Probably the most remote ancestors (in the Cretaceous) of most, perhaps of all, the placental orders would be referable to the Insectivora by definition. In this sense the order Insectivora is prototypal and ancestral to all others among the Placentalia. Because the ordinal characters of the Insectivora are mainly primi- tive and because most of the Paleocene mammals are primitive, almost all known Paleocene forms resemble the Insectivora in many respects. If we knew the archaic mammals of the Paleocene but had no knowledge of any of the forms that lived between that time and the Recent, it would be a much more logical and practical system to refer almost all Paleocene mammals to the Insectivora,” rather than to distribute them in numerous different orders as is now the usual practice. This distribution, in accordance with a “vertical” or so- called evolutionary conception of classification, is accomplished by the recognition of fossils intermediate between the Paleocene groups and the more distinctly separated later orders and of incipient speciali- zations within the Paleocene groups themselves that point toward groups later to become so distinct that they are granted ordinal rank. 31 This is approximately the sense of Cope’s Bunotheria, a broad group including the Insectivora and -various other primitive mammals. Cope did, however, separate and distribute among other orders some Paleocene mammals, such as the Condylarthra, that were on the whole as primitive as those included in -the Bunotheria. 106 BULLETIN 169, UNITED STATES NATIONAL MUSEUM The concept Insectivora as it is applied to the Paleocene (and Eocene) faunas may thus include four different sorts of lesser groups: 1. Very primitive placentals whose ancestral relationship to other later groups is not now recognizable. Naturally we do not know what groups may be included here, for the only way in which this situation could be shown to exist would be by recognition of the rela- tionships, but probably some so-called insectivores are of this char- acter. 2. Animals that are in fact in or near the ancestry of later more specialized insectivores and that are therefore Insectivora sensu stricto. The Nyctitheriidae probably belong in this category, al- though the relationship is not definitely established. The other families in the present fauna almost surely are not Insectivora in this special and most limited sense. 3. Animals not structurally representative of the ancestry of later insectivores but sharing certain specializations with them that seem to indicate that they arose from a common stock with the later groups after the definite differentiation of that stock. These are also to be considered as Insectivora in a strict sense, even though the usage is broader than it is as applied to the second category. The degrees of such collateral relationship vary greatly. Thus the Leptictidae fail definitely in this category, as they share many and apparently sig- nificant special characters with the Erinaceidae, although clearly not ancestral to that family. The Pantolestidae likewise show evidence of a special collateral relationship to the later insectivores, but the resemblance is less particular and the relationship evidently more remote. 4. Groups that were derived from the nominally insectivore pro- toplacental stock but that had begun to diverge markedly from any other groups, without, however, having a sufficiently long history, being sufficiently important faunal elements, or acquiring sufficiently striking special characters to warrant the erection for them of a special order. Such groups are clearly Insectivora only in a very broad sense, yet their exclusion from the order would be a needless complication of taxonomy. The Mixodectidae appear to belong to this category. In this fauna there are 10 genera that are referable to the Insectivora in the general sense here accepted. Gelastops is evidently related to Didelphodus of the Lower Eocene and Acmeodon of the Middle Paleocene. It is perhaps a modified survivor of the protoplacental stock. Prodiacodon and Leptacodon are typical leptictids, Myrmeco- boides is an aberrant member of that group, and Bessoecetor is a primi- tive but typical pantolestid, while Aphronorus constitutes with the Torrejon Pentacodon a more aberrant group probably of pantolestid origin. Eudaemonema seems surely to be a mixodectid, although FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 107 phyletically distinct from any other known genus. Picrodus is really of unknown affinities and is placed in the Insectivora only in default of other evidence. Family ?7DELTATHERIDIIDAE Gregory and Simpson, 1926 Subfamily DIDELPHODONTINAE Matthew, 1918 Matthew placed this subfamily in the Leptictidae, pointing out that the affinities of Didelphodus are uncertain, but that it ‘‘can not easily be placed in any other family.” Except for the general state- ment that ‘‘the teeth are in many respects not unlike those of the Leptictidae’’, he gave no definite reason for placing Didelphodus in that family, where it was decidedly anomalous. In defining the Deltatheridiidae, Gregory and Simpson (1926) stated that Didelphodus might belong in that group, and I still later (Simpson, 1928) gave in somewhat more detail the reasons for this tentative assignment. Gelastops of the present fauna is sufficiently close to Didelphodus to warrant their tentative collocation in one subfamily but, as brought out more fully in describing Gelastops, this is not certain. In any case Gelastops is poorly known and adds little to evidence for the affinities of this group. Its more aberrant premolar structure, still more striking in the probably allied Acmeodon, does to some extent argue against close affinities with the much more primitive Cretaceous forms but this may merely indicate an incipient line of specialization within the Deltatheridiidae. The data do not warrant a more positive conclusion. On present evidence it seems well to retain Matthew’s subfamily Didelphodontinae, for Didelphodus, Gelastops, Acmeodon, and probably Phenacops. This necessitates the proposal of a new subfamily Deltatheridiinae, defined, among other characters, by the less progressive premolars, less separated paracone and metacone, and narrower talonids. Genus GELASTOPS Simpson Gelastops Stimpson, 1935d, p. 227. Emperodon Simpson, 1935d, p. 229. Type.—Gelastops parcus Simpson. Type of Emperodon.—Emperodon acmeodontoides Simpson. Mistribution—Middle Paleocene, Fort Union, Mont. Diagnosis.—Canine large and erect. P, intermediate in structure between Didelphodus and Acmeodon, with paraconid high on crown, metaconid nearly as high as protoconid and partly confluent with latter, a vertical crest descending posteriorly from the metaconid and another from the protoconid, and a small, bicuspid, basined talonid. Molars leptictid or didelphodontine, paraconids large and more inter- nal than in Prodiacodon or similar leptictids, trigonids elevated, that 108 BULLETIN 169, UNITED STATES NATIONAL MUSEUM of M, large relative to talonid and those of M,.3; compressed antero- posteriorly. M, and particularly M3; reduced relative to M,, heel of M; much reduced but with projecting hypoconulid. Discussion—By an unfortunate chance the National Museum material of this unusual genus does not include associated P, and M,. Since P, seems clearly related to Acmeodon, while M, seems just as clearly related to Didelphodus, and since the possibility of relationship between Acmeodon and Didelphodus had never been suggested, I was led to believe that two different animals were represented, one related to each of these genera. Specimens collected since this manuscript was first completed show that the P, supposedly characteristic of Emperodon and the M, supposedly characteristic of Gelastops really belonged to the same animal and this has made possible a last-minute correction in the present work. There is little doubt that Gelastops is related to Acmeodon, although it is, on the whole, more primitive in structure. It had the anterior premolars less reduced than in the latter. P, is superficially quite different in the two genera, but the differences appear to be modifica- tions of the same fundamental structure, which is unlike that of any other genera known tome. In Gelastops the paraconid is larger and more internal and the metaconid is distinct, but the latter is prob- ably represented in Acmeodon by the cuspule on the posterointernal crest descending from the protoconid, in which case the distinction is the relatively minor one that in Gelastops this cuspule is merely more emphasized and shifted slightly anteriorly. The peculiar protostylid is larger in Acmeodon but is also present in Gelastops. In Acmeodon the two crests run into the talonid rim and the valley between them into the talonid basin in such a way that the talonid is poorly differen- tiated, while in Gelastops the talonid is well set off by notches, but the parts seem to be entirely homologous in the two cases. The large and internal paraconid and talonid markedly narrower than trigonid, which so strikingly separate Acmeodon from the Leptictidae, are developed in almost exactly the same way in Gelastops. On the other hand, there is also good evidence for the relationship of Gelastops to Didelphodus. The general aspect of the jaw, canine, and cheek dentition is much the same in the two genera. In Gelas- tops P, is probably more reduced and P,-, more elevated. Py, is definitely more specialized in Gelastops than in Didelphodus, having a higher crown, more elevated (but little larger) paraconid, larger and much more elevated metaconid, and more prominent posterior crest on the metaconid. This crest is, however, distinctly present in Didelphodus and in general all the structural features seem to corre- spond in the two genera, the differences involving only relative proportions and prominence of the various parts. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 109 The lower molars of Gelastops so closely resemble those of Didel- phodus as to make a hypothesis of relationship most reasonable. The only really clear difference is the greater elevation of the trigo- nids in Gelastops and this is not extreme. The upper molar men- tioned below also adds to the evidence, but the reference is not certain. Gelastops thus resembles both Acmeodon and Didelphodus and is structurally almost intermediate between the two. Without pro- ducing absolute proof, it strongly suggests that the three genera Didelphodon, Gelastops, and Acmeodon belong to a natural group, and they are here tentatively recognized as constituting the sub- family Didelphodontinae. There are still difficulties in the way of this interpretation that perhaps can be resolved only by the discovery of upper dentitions of Gelastops and Acmeodon. The outstanding difficulty is that the series Didelphodus-Gelastops-Acmeodon is an excellent structural sequence in that order, on the basis of the parts now known in all, but that the most primitive genus, Didelphodus, is the latest in time, a relationship that is fully possible but that re- quires better evidence for definitive acceptance. GELASTOPS PARCUS Simpson Figures 12, 13 Gelastops parcus Simpson, 1935d, p. 227. Emperodon acmeodontoides Simpson, 1935d, p. 229. Type-—U.S.N.M. no. 6148, right lower jaw with canine, M,, Msg, and alveoli. Collected by A. C. Silberling. Type of Emperodon acmeodontoides.—U.S.N.M. no. 9850, right lower jaw with P,, M;, and part of P;. Collected by A. C. Silberling. Horizon and locality—Type probably from Silberling Quarry,” other specimens from Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis—Sole known species of genus. Measurements in table 22. Remarks.—Besides the type and that of the synonym, there are two referred specimens from the Gidley Quarry in the National Museum collection: U.S.N.M. no. 9601, a right lower jaw with M,_; and alveoli, and U.S.N.M. no. 9446, a right lower jaw with M>-3. From the alveoli it appears that a reduced P; was present; this tooth is absent in Acmeodon and unreduced in Didelphodus. Po-3 were present and each had two well-separated roots. The structure of P, has already been sufficiently described except to note that the protoconid tip bends outward (labially) in a peculiar way, approach- 82 There is some inconsistency in the available records. It is certain that the type is from the Fort Union No. 2, and from its preservation it must be from one of these two quarries, but some error or omission had been made in recording the field number. Being collected in 1908, it is more likely to be from the Silberling Quarry. 110 BULLETIN 169, UNITED STATES NATIONAL MUSEUM ing Acmeodon, its outer wall looking as if it had been plastically de- formed by pressing the apex outward. Except for details in the generic diagnosis or evident from the measurements, the lower molars so closely resemble those of Didelphodus absarokae, described by Matthew (1918), that further description of them is not necessary. There is a single left M?, U.S.N.M. no. 9554, from the Gidley Quarry, that may belong to this species, although it differs more markedly from Didelphodus absarokae than do the lower molars. It is remarkably short and wide, or extremely transverse, and further differs from Didelphodus in the great extension of the parastyle spur, the slightly more external position of paracone and metacone, and the vestigial character of the metaconule. It resembles Didelphodus in other respects, such as the emarginate outer border, shelf-like meta- FIGURE 12.—Gelastops parcus Simpson, U.S.N.M. no. FIGURE 13.—Gelastops parcus Simpson: 6148: Right lower jaw, external view. Twice natural a, Crown view of right lower teeth size. and alveoli, U.S.N.M. no. 9601; 6, internal view of right lower jaw, U.S. N.M. no. 9850. Twice natural size. stylar extension, connate bases of paracone and metacone, and entire absence of hypocone or of anterior or posterior cingula. The trans- verse extension and the reduction of the metaconule are, furthermore, Didelphodus-like but here more extreme and the nonmarginal position of paracone and metacone are also Didelphodus-like but less pronounced. TABLE 22.— Measurements of individual specimens of Gelastops parcus. 1a M, M3 M; M2? U.S.N.M. no. L W L W L WwW L WwW L Ww Mm|Mm|Mm|Mm|Mm|Mm)|Mm|Mm|Mm\|Mm O48 Yee Sree se oe ae ose Sree | ee cae 3. 5 7 Yall eee | eee 2.9 180252228 Sacer O44 6 Be 2 rate an Sees ot Ea een See. Ree eee ee eed ee oe 2.5 2.0 PA 1.9 OR SORE Se ee ee es ee ee Ee ees Fe. 2.8 JQ) [sae healer ae 2.9 2.2 10) LS pa es ee ee ee ell (ee See | eee BLOM) 253) | 289) ) = 19. O55 aes hee, OU Sees Pay i ee Ee en ke Oe 2k | Ee eee eee Pst 5.0 FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. skal Family LEPTICTIDAE Gill, 1872 Unlike many of the Paleocene and Lower Eocene groups referred to the Insectivora, the Leptictidae are Insectivora sensu stricto. That is, they are not merely primitive forms that are presumed to have been derived from an undifferentiated insectivore stock and that do not enter well into any other order, but are definitely insectivores in a special sense, clearly related to recent insectivores. Their affinities seem to be with the erinaceoids, although here they constitute an extinct side line, not ancestral to the true Erinaceidae. More exact elucidation of their affinities depends on full analysis of the characters of the Oligocene forms, which are known from nearly complete skeletons but have never been adequately studied. Matthew (1918) has pointed out that there is a group of genera undoubtedly leptictid (now about nine genera) and that in addition to these there are several diverse genera placed here without much positive evidence but in default of other indications. Didelphodus, Phenacops, and Acmeodon, then placed here by Matthew, are now TABLE 23.—Comparison of lower dentition of five genera of Paleocene Leptictidae Genus Canine Py Trigonids Molar paraconids Ms; talonid Diacodon_-_.--- Small_.| With sharp, low, median | Trigonids of Ps- | Lower than other | Elongate. paraconid. Protocon- M; moderate- trigonid cusps, id and metaconid op- ly elevated. small, submedi- posite and _ nearly an. equal. Heel basined, with 3 cusps, hypo- conid largest and near- ly confluent with hy- poconulid. Prodiacodon_..| Large..| With trigonid similar to | Trigonids of Psi- | About asin Diaco- | About as in Diacodon. Talonid M; strongly don. Diacodon. with hypoconid rela- elevated. tively less dominant, hypoconulid more sep- arate, sometimes with 4 distinct talonid cus- pules. Leptacodon__.-| Large_-| Closely similar to Diaco- | Little elevated..| Nearly internal, | Short. don, but metaconid otherwise more smaller, usually pos- like Diacodon. terointernal to proto- conid. Talonid small. Myrmecoboides_| Moder- | Elongate, trigonid sim- | Moderately ele- | Small but nearly | Very elon- ate. ilar to Diacodon. Tal- vated. as high as meta- gate. onid long, narrow, ba- conids, fully in- sined, without dis- ternal, fusing tinet hypoconulid, en- with metacon- toconid about equal to ids. hypoconid. Xenacodon_..-- Small__| With small basal para- |_-___ (0 (oe ee About intermedi- | Short, strong- conid, large metaconid. ate between ly reduced. Talonid very short, Diacodon and not basined. Leptacodon. 112 BULLETIN 169, UNITED STATES NATIONAL MUSEUM tentatively removed to the Deltatheridiidae. The genus of most doubtful affinities still retained in this family is Yenacodon Matthew and Granger, of the Tiffany. The more positively leptictid genera Diacodon, Prodiacodon, Leptacodon, and Myrmecoboides also occur in the Paleocene, the last three being represented in this fauna. Table 23 shows some of the more striking distinctions in the lower dentitions of the five genera recorded from the Paleocene. Of these genera, Diacodon, Prodiacodon, and Leptacodon are typical leptictids and are evidently very closely allied, to such a point that they are rather difficult to distinguish. Myrmecoboides seems to be a true leptictid but is highly distinctive and perhaps not closely related to any of the more typical genera. Xenacodon is a distinctive but poorly known and rather dubious form. Genus PRODIACODON Matthew, 1929 PRODIACODON CONCORDIARCENSIS Simpson FicureE 14 Prodiacodon concordiarcensis Stmpson, 1935d, p. 228. Type.—U.S.N.M. no. 9637, left lower jaw with P., P., M3, and alveoli. Collected by Dr. J. W. Gidley. Horizon and locality —Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. ’ , ‘ V7" \ A H) SD FIGURE 14.—Prodiacodon concordiarcensis Simpson, U.S.N.M. no. 9637, left lower jaw: a, Crown view; b, internal view. Five times natural size. Diagnosis —Much smaller than Prodiacodon puercensis. P, with paraconid more secant and projecting more anteriorly, talonid with three conical cusps, hypoconid largest and entoconid smallest. M3 with trigonid more slender and markedly elevated, talonid less elongate, with three subequal cusps. Remarks.—P, has a high but small paraconid and low posterior cuspule. The horizontal ramus is very long and slender, and the anterior teeth were evidently procumbent. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 113 This delicate species differs so much from Prodiacodon puercensis that the generic reference is not at all certain, but the only alternative would be erecting a new genus, which is highly undesirable at present. The present form is apparently a typical leptictid, peculiar only in minute details, and as close to Prodiacodon as to any other defined genus. It is almost equally close to Leptacodon, but its reference there would make the generic definitions almost impossible to frame adequately. The type is the only known specimen. Its dimensions are as follows: Length P., 1.3; width P,, 0.5; length P,, 2.0; width P,, 1.1; length M;, 1.9; width Ms, 1.2. Genus LEPTACODON Maithew and Granger, 1921 LEPTACODON LADAE Simpson Figure 15 Leptacodon ladae Simpson, 1935d, p. 228. Type.—U.S.N.M. no. 9640, right lower jaw with Py-M;. Collected by A. C. Silberling. Horizon and locality Gidley Quarry, Fort Union, Middle Paleo- cene horizon, Crazy Mountain Field, Mont. FIGURE 15.—Lepiacodon ladae Simp- FIGURE 16.—Leptacodon munuscu- son, U.S.N.M. no. 9640, right lum Simpson, U.S.N.M. no. lower jaw: a, Crown view; ), in- 9819, left lower jaw: a, Crown ternal view. Five times natural view; 6, internal view. Five size. times natural size. Diagnosis.—Slightly larger than L. tener or L. packi and slightly smaller than L. siegfriedti, structurally closer to the former two species (subgenus Leptacodon) than to the latter (subgenus Leipsano- lestes). P, elongate, paraconid median, metaconid very small but in the same postion as in L. tener, talonid as in that species. Molar paraconids smaller than in ZL. tener but distinct and internal. Hypo- conulids of M,-3; more projecting than in L. tener. Talonid of M; more elongate and entoconid smaller. Remarks.—This species is referable to Leptacodon with very little doubt. Although fairly common in the Gidley Quarry, the speci- mens are all very fragmentary, and they add nothing to knowledge of the genus beyond making it a little more varied. 114 BULLETIN 169, UNITED STATES NATIONAL MUSEUM The dimensions of the type are as follows: Length Py, 1.4; width P,, 1.0; length M,, 1.5; width M,, 1.2; length M,, 1.4; width Mg, 1.3; length M3, 1.5; width Ms, 1.1. The available numerical data on the whole sample are given in table 24. TABLE 24.—Numerical data on specimens of Leptacodon ladae Variate N R M o V WGP ese eae 5 1.4—1.6 1. 48 Wile oe eee 5 0.8—1.0 0. 90 ING See ae Ss i 1.4—1.6 1. 53+0. 03 0. O07 40. 02 4,.6+1. 2 WIM joo. = 2 6 1.0—1.3 1.17+0. 04 0. 09+0. 03 8 142.3 1 NY ee pees ee 9 1.38—1.6 1.438+0. 03 0. 08 +0. C2 5 fe ded Wags ee ee 8 0.9—1. 4 1. 20+0. 05 0. 14+0. 04 11.7+2.9 WMeo 2 Se ae 8 1.4—1.6 1. 51+0. 02 0. 06+0. 02 4.0+1.0 \ WNW iy ees € 1.0—1. 2 1. 09+0. 02 0. 06+0. 02 5. 941.6 1 BN Ge ee 3 4,5—4.7 4. 57 The two available specimens of L. tener are at or slightly below the observed lower limits for Z. ladae in all molar dimensions. The deviations of the four available dimensions of the type of L. tener from the means of L. ladae, divided by the corresponding standard deviations of the latter (d/c) are as follows: LM;: —6.1 LM:: —4.0 WM;,:—5.0 WM:: —2.9 The size difference, slight as it is, thus appears to be significant. Individual measurements of ZL. packi are not available, but that species is very close to L. tener in size, hence probably also signifi- cantly smaller than L. ladae. LEPTACODON MUNUSCULUM Simpson Figure 16 Leptacodon munusculum Simpson, 1935d, p. 228. Type.—U.S.N.M. no. 9819, left lower jaw with M, and M;. Col- lected by A. C. Silberling. Horizon and locality Gidley Quarry, Fort Union, Middle Paleo- cene horizon, Crazy Mountain Field, Mont. Dragnosis.—M, of about the same size as in L. tener, M; more reduced. Paraconids smaller and more strictly internal. Talonid of Ms; relatively narrower. Remarks.—This poorly known species is the smallest mammal in the collection and is one of the smallest mammals known. It is ‘3 The original diagnosis says ‘‘Slightly smaller than Leptacodon tener’’, which was probably true of the dentition as a whole and perhaps of the animal, but M1 is as large as in L. tener. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. £5 almost certainly referable to this genus but is shown to be a dis- tinctive species even from the one very fragmentary specimen avail- able. M, is 1.2 and M; 1.1 mm in length. The comparison with L. ladae by d/o is thus as follows: LM;:4.7 LM;:6.8 The ratio LM,/LM; compares as follows: L. tener (referred specimen) : 0.86. L. ladae (three specimens) :1.00—1.07, mean 1.05. L. munusculum: 1.09. M; is thus more reduced in ZL. ladae and in L. munusculum than in L. tener. The difference between L. munusculum and L. ladae may be, but is not shown to be, significant. Genus MYRMECOBOIDES Gidiey Myrmecoboides G1pLEY, 1915, p. 395. Type.—Myrmecoboides montanensis Gidley. Distribution.—Middle Paleocene, Fort Union, Montana. Diagnosis —Canine of moderate size. Premolars well spaced. P, elongate, with strong but low, median paraconid, metaconid well separated and nearly as high as protoconid, talonid long and narrow with small, subequal hypoconid and entoconid and hypoconulid barely suggested. Molars, particularly M3, with long talonids and with moderately elevated trigonids. Paraconids smaller than meta- conids but nearly as high and partly connate, so that paraconid and metaconid together form a twinned apex that is higher than the protoconid. Remarks.—Gidley stated that the name Myrmecoboides was “given to the ancient form on account of its likeness to Myrmecobius rather than as a positive assumption of real relationship.” He did, however, state that the genus was marsupilalian and probably related to the Myrmecobiidae, and he went into much detail regarding its bearing on marsupial evolution. Abel (1919) placed Myrmecoboides in the Myrmecobiidae. Osgood (1921) stated that Myrmecoboides might well be ancestral to Caenolestes, if it wasa marsupial. Schlosser (1923) classified it in the Myrmecobiidae but noted that it might not be mar- supial. Longman (1924) gave a résumé of Gidley’s view, stating (as had Gidley) that the resemblances to Myrmecobius were suggestive but not conclusive. Matthew (1916) had, however, quickly pointed out that the characters of Myrmecoboides are not myrmecoboid or marsupial, but leptictid and placental. Gidley also came to accept this conclusion (personal communication) but apparently did not publish this fact. Students of Paleocene mammals have long realized that Myrmecoboides cannot be a marsupial, but aside from Matthew’s brief statement, which has been overlooked by almost all later com- 116 BULLETIN 169, UNITED STATES NATIONAL MUSEUM mentators, the evidence has not been clearly stated in publication and Gidley’s abandoned view is becoming entrenched in the literature. It is therefore necessary to consider it here. Gidley did not explicitly discuss the evidence for referring Myrmeco- boides to the Marsupialia. It is clear that the reference was based almost entirely on the molariform character of the fourth cheek tooth, and in part simply on a general resemblance to the one marsupial genus Myrmecobius. Gidley mentions the three possibilities regarding the molariform tooth: that it is a true molar, that it is a deciduous molar, and that it is a true premolar. He cites the narrower crown, smaller talonid, and large, anteriorly directed paraconid as seemingly preclud- ing the possibility that the tooth is a true molar. This evidence is valid, and to it may be added (as first pointed out by Matthew) the fact that the protrusion and wear of the tooth definitely prove that it was erupted long after the following tooth, which could not be true of a first molar. Gidley states that the very complex structure in com- parison with P; opposes the supposition that this 1s P, and supports his conclusion that it is dm,, retained in the adult dentition. But the relative time of eruption, as first shown by Matthew, is impossible for dm,, which in all known mammals is erupted before M, rather than long after, as the tooth in question is in Myrmecoboides. Further- more, the fourth cheek tooth in marsupials, whether it be M, or dm,, is erupted before the following tooth or at almost the same time. Gidley also seems to have overlooked, momentarily, the fact well known to him that in the Leptictidae and some other unquestionable placentals P, is characteristically as molariform as this tooth, or more so, even though P; may be as simple as in Myrmecoboides. It may be con- cluded, and Gidley also did conclude on more mature deliberation, that this is certainly P,, and hence that its evidence is entirely opposed to marsupial and in favor of placental affinities for the genus. Gidley analyzed all the resemblances to and differences from Myrmecobius. The only resemblances that could be supposed to carry much weight are the spacing of the premolars and the elevation of the internal cusps of the molars. These characters also do occur in placentals and are highly aberrant among marsupials. These, and all the other lesser points of resemblance, would perhaps tend to link Myrmecoboides to Myrmecobius if the former were otherwise proved to be a marsupial, but they have no value as to the primary question of showing it to be marsupial. There are indeed profound differences between the fossil and Myrmecobius. Gidley recognized these but con- cluded that degeneration could lead to this great transformation, which is true but is not evidence that it did so. On the other hand, Myrmeco- boides has the basic and diagnostic characters of the Leptictidae, and the Myrmecobius-like characters can only be interpreted as superficial, aber- rant, habitus characters bringing about a slight degree of convergence. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 117 The fact that Myrmecoboides is not a marsupial obviously deprives it of any bearing on whether the fourth cheek tooth of marsupials is dm, or M,, a question that Gidley properly discussed when under a misapprehension as to the affinities of the genus. It also removes this much support from Gidley’s thesis, which was, however, also sup- ported by other evidence, that the Australian families were differen- tiated outside Australia and at a very early date.™ Among the Leptictidae, Myrmecoboides is aberrant and does not belong with such typical genera as Diacodon or Leptacodon, although its less direct relationship with them is highly probable. MYRMECOBOIDES MONTANENSIS Gidley Figures 17, 18 Myrmecoboides montanensis G1pLEY, 1915, p. 395. Type.—U.S.N.M. no. 8037, left lower jaw with canine and P,—Ms3. Collected by A. C. Silberling. Horizon and locality —Gidley Quarry, Fort Union, Middle Paleo- cene horizon, Crazy Mountain Field, Mont. Diagnosis—Sole known species of genus. Dimensions below. Remarks.—Gidley (1915) has accurately described the lower denti- tion, and the slightly different emphasis demanded by transfer to this family is supplied by the revised generic diagnosis. The second spec- imen mentioned by Gidley is U.S.N.M. no. 9418, which has only M,.» and these so worn that the cusp structure cannot be made out. U.S.N.M. no. 9552 is a right upper jaw with P*-M? that is so well suited to be the upper dentition of Myrmecoboides montanensis, and not of any other known species in the fauna, that it may be referred here. Dr. Gidley has noted that this is probably the upper dentition of Myrmecoboides—further evidence of his later recognition of the affinities of the genus, for this upper jaw is entirely leptictid in char- acter and does not at all resemble Myrmecobius. These upper teeth are very close to those of Prodiacodon throughout. P* seems to have had the paracone and metacone better separated in Myrmecoboides, but is imperfect in this region. Ml! has the external shelf more pro- nounced, the parastylar and metastylar lobes more projecting, a deep §4 When Gidley wrote, it was a reasonable a priori assumption that diverse marsupials would be the micro- faunal elements most likely to appear in the Paleocene, although subsequent discovery has shown this not to be the case. He was inevitably predisposed toward this view (as was also Matthew and as were other con- temporaneous students), and his preliminary note on Myrmecoboides suffered from this preconception. It is further exemplified by his manuscript notes on several other placental genera labeled as ‘‘Marsupial No. 1” and so on; it has not seemed necessary to cite this first impression in each case, as Dr. Gidley would certainly have abandoned it before completing his studies. Dr. Gidley also had a strong and more personal predis- position, strikingly exemplified in the Myrmecoboides paper, to believe that the modern mammalian families were of extremely remote origin. This is a legitimate thesis, and the erroneous nature of part of the supposed evidence does not remove the possibility or vitiate Gidley’s whole argument. His conclusion unquestion- ably contains an important truth, but I believe, apparently with the consensus of recent students, that he overemphasized its importance and extent. 118 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Se, 2 “I Wines BG UK BN Rs eB AN YS By Ego =P 5 Zp” FIGURE 17.—Myrmecoboides montanensis Gidley, U.S.N.M. no. 8037, left lower jaw: a, Crown view: 6, in- ternal view. Four times natural size. FiGuRE 18.—Myrmecoboides montanensis Gidley, U.S.N.M. no. 9552, right upper P‘-M3, crown view. Four times natural size. median notch in the external border, and the hypocone more internal. M? has the outer portion analogously modified, but this is less dis- tinctive from Prodiacodon than is M!. M* is of almost identical struc- ture in the two genera. TABLE 25.—Measurements (in mm) of the type and of the referred upper jaw of Myrmecoboides montanensis Pi P2 P3 Py Mi M2 M3 Pt MI! M? M3 p(wiulwiolwie|wielwi[s]w[olw/o|wlelwie w]e] w ee | se | | | J | | | | i [J | es | eee lt 1.1) 0.6} 1.8] 0.8) 2.3) 0.9) 2.7) 1.3) 2.4) 1.5) 2.3 4a 2. 6) 1. 5) 2.6] 3.3) 2.3] 3.7 24 3.8} 2.0) 3.7 | Family NYCTITHERIIDAE Simpson, 1928 This family was established to include a number of rather poorly known Paleocene and Eocene genera, all represented by species of small or minute size, that probably represent a single varied group and that seem to be primitive shrewlike insectivores, without any of the specializations characteristic of the true shrews. In this respect the dentition is molelike rather than shrewlike, but there is no evi- dence of definitely talpid specializations. As Matthew (1909) has pointed out (in other words), speaking of Eocene genera that I have since placed in this family, they seem to combine all the most primi- tive characters of the moles and the shrews. To this extent the evi- FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 119 dence suggests an undifferentiated group from which these late insec- tivores arose, but annectant types are lacking and knowledge is too incomplete to test this hypothesis adequately. Matthew (1918) has also pointed out that Nyctitherium itself may be a chiropteran, and this may be true of the whole group, or it may be a primitive complex allied or ancestral to all three groups, Talpoidea, Soricoidea, and Chir- optera, although there is reason to suppose that these three groups were already distinct before the end of the Paleocene. It is likewise possible that two or all three of these groups had acquired their skel- etal specializations but not their dental characters in the Paleocene and that the nyctitherids are an artificial assemblage based on dental resemblance only. Despite this possibility, it seems most practical to continue to associate them in this family until some broader basis for classification becomes available. FIGURE 19.—Stilpnodon sjmplicidens Simpson, U.S.N.M. no 9629, left lower jaw: a, Crown view; 6, internal view. Five times natural size. In the National Museum collection only one nyctitheriid specimen is available, but it represents a distinctive genus, Stilpnodon. It is not very close to any other known genus, and its reference to the family is not certainly established, but it is most conveniently placed here. It shares with Leptacodon munusculum the distinction of being the smallest known mammal of this fauna. Genus STILPNODON Simpson Stilpnodon Simpson, 1935d, p. 229. Type.—Stilpnodon simplicidens Simpson. Distribution.—Middle Paleocene, Fort Union, Mont. Diagnosis.—P, with very high, slender main cusp, minute rudimen- tary anterior basal caspule, no metaconid, simple nonbasined talonid with one cuspule. M; reduced, with distinct, low, nearly median paraconid, trigonid erect and moderately elevated above talonid, protoconid large, trigonid nearly as long as talonid, talonid short and much narrower than trigonid. 119212—387 9 120 BULLETIN 169, UNITED STATES NATIONAL MUSEUM STILPNODON SIMPLICIDENS Simpson Figure 19 Stilpnodon simplicidens Simpson, 1935d, p. 229. Type.-—U.S.N.M. no. 9629, left lower jaw with P34, M3, and alveoli. Horizon and locality —Gidley Quarry, Fort Union, Middle Paleo- cene horizon, Crazy Mountain Field, Mont. Diagnosis.—Sole known species of genus. P,length, 1.0mm. Ms; length, 1.2 mm. Remarks.—P; is very simple, without anterior basal cuspule and with the sloping heel not forming a distinct cusp. From its roots, P, was of about the same size as P;. The posterior mental foramen is beneath the anterior root of P,. Family PANTOLESTIDAE Cope, 1884 In his Bridger memoir Matthew (1909) united various genera pre- viously widely scattered in the system and placed them in the Pantoles- tidae. From a study of Pantolestes, a relatively specialized but well- known Middle Eocene genus, he showed that at least the typical members of the family have basic insectivore characters, overlain by peculiar specializations not closely paralleled in recent insectivores. He interpreted the majority of these specializations in Pantolestes as adaptations to aquatic, or to amphibious, life. At that time he pointed out that Palaeosinopa, then known from the Lower Eocene, is closely related to Pantolestes and that Pentacodon, of the Middle Paleocene, might tentatively be placed in this family. In 1918 Matthew revised and carefully described the Lower Eocene Palaeosinopa, again emphasizing its close relationship to Pantolestes, and reviewed the evidence for reference of these genera to the Insectivora. This evidence, which still appears to be adequate for such a conclusion is, in briefest possible outline, that the pantolestid dentition is, indeed, more creodontlike than like any recent insectivore but is also closely similar to the unquestionably insectivore leptictids, and that numerous characters of skull and skeleton (especially the astragalus) are quite unlike any known creodonts or other carnivores but do resemble the Leptictidae and other Insectivora. A special point of the evidence (singled out by Schlosser, see Matthew, 1918, for undue emphasis) is the position of the posterior mental foramen beneath M,, a character known only in Insectivora, although not characterizing all members of that order. Various fragments suggested the presence of this family in the Upper Paleocene (e. g., Simpson, 1927), and Jepsen (1930a) referred a FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. P20 partial jaw from the Middle Paleocene to Palaeosinopa. I have described a genus, Bessoeceter, from the Upper Paleocene (Simpson, 1936b). The present fauna includes a distinctive species that seems surely to be closely allied to or in the Palaeosinopa—Pantolestes line and is tentatively referred to Bessoecetor, thus definitely projecting the typical pantolestid phylum back into the Middle Paleocene. It also includes in Aphronorus an ally of Pentacodon, represented by much better material than is that Torrejon genus and much improving our knowledge of this small group. The probable relationship of Pentacodon and Aphronorus, on one hand, and Bessoecetor, Palaeosinopa, and Pantolestes, on the other, is confirmed and strengthened by the present evidence. In both the lower jaw has a large semiprocumbent canine followed by three small elongate premolars of similar pattern in the two groups and then by a relatively enlarged premolar, the pattern of which is, however, unlike in the two lines. The molars, upper and lower, are of almost exactly similar pattern, with only such minute differences as are encountered among very closely allied genera. The posterior mental foramen is beneath M, in both cases and the other known osteological characters, although few and not very distinctive, are generally similar. The only characters opposing such relationships are those of P*j, which are more and differently specialized in Pentacodon and Aph- ronorus than in Bessoecetor or even the much later genus Pantolestes. These marked differences do not seem to exclude a family relation- ship, but they do show that two divergent and sharply defined groups are included, especially when the distribution of the various genera is taken into account. In the present state of knowledge it seems convenient to formalize this distinction by placing the known genera in two subfamilies, the Pantolestinae with Bessoecetor, Pantolestes, Palaeosinopa, and perhaps some less well known forms, and the Pentacodontinae with Pentacodon and Aphronorus. For the present these are immediately and adequately distinguishable from the characters of P*, as given in the diagnosis below. PANTOLESTINAE, new subfamily Type.—Pantolestes Cope, 1872. Distribution —Middle Paleocene to Middle Eocene, North America. Lower Eocene, Europe. Diagnosis —P*, relatively little enlarged. P, compressed, no metaconid, semishearing, heel essentially unicuspid and little or not basined. P* with compressed amphicone, no metacone, low, sharp styles, small protocone without cingula. 122 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Genus BESSOECETOR Simpson Bessoecetor Simpson, 1936b, p. 9. Type.—Bessoecetor thomsoni Simpson. Distribution —Middle and Upper Paleocene, Fort Union, Montana. This genus was based on an unusually good suite of specimens from the Scarritt Quarry. Itis evidently very closely allied to Palaeosinopa but seems to be slightly more primitive in several respects. The species Palaeosinopa diluculi, previously recorded from the present fauna (Simpson, 1935d, p. 230) appears to be more nearly related to Bessoecetor and may be provisionally placed in that genus, although it is too poorly known to reveal all the desired generic characters. Most of the characters in which it differs from Lower Eocene species of Pal- aeosinopa and which were therefore given as specific characters in the original diagnosis are now shown by the fine material of Bessoecetor wey , ha NZ) ) " ST {i " FIGURE 20.—Bessoecetor diluculi (Simpson), U.S.N.M. no. 9810, with parts in outline added from U.S.N.M. nos. 9312 and 9539, left lower jaw: a, Crown view; 0, internal view. Three times natural size. thomsoni to be generic characters of Bessoecetor. This includes, most notably, the strongly trenchant P, with large anterior basal cusp and incipient basining of talonid, the relatively small expansion of the hypocone shelves, and the less reduced M*. The Middle Paleocene form also has the sharply distinguished molar talonid cusps character- istic of Bessoecetor. BESSOECETOR DILUCULI (Simpson) Figures 20, 21 Palaeosinopa diluculi Simpson, 1935d, p. 230. Type.—U.S.N.M. no. 9810, left lower jaw with P,-M,. Collected by A. C. Silberling. Paratype—U.S.N.M. no. 9553, left upper jaw with P*-M®? (some- what broken). Collected by A. C. Silberling. Horizon and locality —Gidley and Silberling Quarries, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 123 Diagnosis. —P, and M,-; significantly larger than in B. thomsoni, My about same size. elongate. Heel of P, relatively larger, talonid of M; more Amphicone of P* more compressed. Remarks.—The dentition is closely similar to that of B. thomsoni, fully described elsewhere (Simpson, 1936b). TABLE 26.—Dentition measurements (in mm) of Bessoecetor diluculi Py Mi M2 M3; U.S.N.M. no. L Ww L WwW L Ww L Ww OSI eres. 222 ck | eee 1S) 2.4 1 ¢/ 2.5 2.0 OSM eee tot |e eee Ree ee RL Ree a 2. 6 2. 2 By Os 73S OA A ee a ee | eee 1.4 2.3 be 2.7 2.0 Oi3)) a eee ee Ss 3. 2 1.3 27 1.8 Ps M! M? M3 U.S.N.M. no. L WwW L WwW L Ww L w Oo See Betas S| Beal pei PAC 3. 4 2.8 3. 8 OS neers ee eee Bae 2Grtlee Paya Al he se 2. 6 3. 5 2. 0 3.5 FIGURE 21.—Bessoecetor diluculi (Simpson), U.S. N.M. no. 9553, with part in outline added from U.S.N.M. no. 9585, left upper jaw: a, External view; 6, crown view. Three times natural size. y DS =) & aN CANN 6) See ‘3 yy N S As Ay if ZZ TI NUN SUT a We FIGURE 22.—Aphronorus fraudator Simpson, U.S. N.M. no. 6177, left lower jaw: a, Crown view; 6, external view. ‘Three times natural size. PENTACODONTINAE, new subfamily Type.—Pentacodon Scott, 1892. Distribution Middle Paleocene (and doubtfully Upper Paleocene), North America. Diagnosis —P*, much enlarged. P, with very heavy protoconid, sloping backward, well-developed metaconid, and basined heel. P4 with massive, conical, paracone, smaller but sharply distinct metacone, 124 BULLETIN 169, UNITED STATES NATIONAL MUSEUM styles small or lacking, and large, low protocone with widely expanded anterior and posterior cingula. Remarks.—Aside from its probable relationship to the Pantolestinae, this group shows some resemblance to the Anisonchinae and to the Apheliscidae, but in neither case is this resemblance close enough to suggest real affinity. The enlargement of P*, and the general but primitive stamp of the molars is not unlike the Anisonchinae but does not include any detail such as might indicate a real relationship. P*, still more resemble those of Apheliscus but differ in cusp structure, and the molars are widely different. It is not impossible that Aphelis- cus is an offshoot of the same stock, but real evidence for such a hypoth- esis 1s lacking. Genus APHRONORUS Simpson Aphronorus Simpson, 1935d, p. 230. Type.—Aphronorus fraudator Simpson. Distribution.—Middle Paleocene, Fort Union, Montana. Diagnosis.—Generally similar to Pentacodon. P, with anterior end less produced downward than in Pentacodon, talonid more distinctly basined, with second cuspule more distinct. M,_3; less reduced relatively toM,. Trigonids of M,_, relatively shorter and entoconids relatively higher than in Pentacodon. Three talonid cusps of Mg, more distinct. P* with metacone well differentiated, protoconule distinct. M! and to less degree M? slenderer and more transverse than in Pentacodon, more leptictid in aspect. Remarks.—This genus is evidently closely allied to Pentacodon, and its distinctive characters, taken together, do not definitely suggest that it is necessarily either an earlier or a later stage in the Pentacodon phylum. Evolution could have proceeded in either direction, or the two may have diverged from a common ancestry. Nevertheless Aphronorus does seem to resemble Palacosinopa in more details than does Pentacodon and to this extent may be supposed to retain more of of the primitive characters of the family as a whole, despite its possible divergence in other respects. APHRONORUS FRAUDATOR Simpson Figurms 22-24 Aphronorus fraudator Simpson, 1935d, p. 230. Type —U.S.N.M. no. 6177, left lower jaw with Py-M;. Collected by A. C. Silberling. Horizon and locality —Gidley and Silberling Quarries, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis—Sole known species of genus. Dimensions given in table 27. Remarks.—The lower canine, known from its alveolus only, is large and semiprocumbent. P,is very small and appears to have a single FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 125 root. P, is likewise small but more elongate and with two roots. The only specimen that shows P,_2 is senile and these teeth are truncated by wear at the bases of the crowns. P,; is larger than P, but also small in relation to the whole dentition. When unworn it has a triangular protoconid followed by a heel with a slight internal basin and two rudimentary posterior cusps. P,is much enlarged and has a massive protoconid, which slopes backward, and a lower distinct metaconid, closely applied to the posterointernal base of the protoco- eee FIGURE 23.—Aphronorus fraudator Simpson, U.S.N.M. no. 9448, with parts in outline added from U.S.N.M. nos. 9291 and 9519, right lower jaw: a, Crown view; b, internal view. Four times natural size. Wan) / : Git \4 ev STN FIGURE 24.—Aphronorus fraudator Simpson, U.S.N.M. no. 9560, with tooth in outline added from U.S.N.M. no. 9565, left upper jaw: a, External view; 6, crown view. Four times natural size. nid. Anterior to this there is a sharp internal cingulum, turning into a vertical crest at the anterior midline of the tooth, but there is usually no paraconid or anterior cuspule, although in one or two specimens there is a slight tendency to develop a cuspule here. The large heel has its main cusp external to the midline, connected by a sharp crest to the middle of the trigonid base. Internal to this crest is a basin, and a second, smaller and less distinct cusp is near the posterointernal corner of the tooth. 126 BULLETIN 169, UNITED STATES NATIONAL MUSEUM The lower molars decrease in size from M, to M;. The paraconids are small and tend to assume a conical shape but are still connected to the anterior protoconid crest. The paraconids are slightly internal to the midline. Metaconids and protoconids are nearly equal. The talonid is wider than the trigonid on M,, about equal on M3, and narrower on M3. On M,.; the three talonid cusps are about equal in height but the hypoconulid is smallest. On M,; the hypoconulid is enlarged and projects posteriorly in the usual manner. The mandible is lighter than in Pentacodon and very closely ap- proaches Palaeosinopa in all respects. P? is 3-rooted and has a single outer cusp like that of P* on a smaller scale save for the absence of a metacone. The protocone is small, and there are no cingula. P* is a peculiar, large, heavy tooth. Its con- tour is hourglass-shaped, with median constriction and expanded equal inner and outer lobes. The paracone is nearly circular in sec- tion and is heavy and tall and slopes slightly backward, like the pro- toconid of P,. It is followed by a smaller metacone. There is only the vaguest rudiment of a parastyle, but a very small metastyle is present. The protocone is crescentic and large but lower than the paracone. There is a small protoconule and rudimentary metaconule. Approximately equal and expanded anterior and posterior cingula occur on the protocone base, and each tends to develop into a cuspule (protostyle and hypocone). The upper molars are of leptictid type, strongly transverse, with emarginate external shelves, projecting parastylar and metastylar lobes (especially on M*), metacones slightly smaller than paracones, conules distinct, anterointernal cingula, and low hypocones slightly more internal than the protocones. M® is reduced in the usual manner, with metastyle and hypocone absent and metacone small. TaBLE 27.—Numerical data on Aphronorus fraudator Variate N R M o Vi | i ey ee eee 10 3. 2-3. 8 3. 52+0. 05 0.17 +0. 04 4.9+1.1 WiRie so one 10 2. 0-2. 4 2.15+0. 05 0.14 +0. 038 6. 741.5 1 SY ee ae eee 10 2. 8-3. 1 2. 88 +0. 03 0. 098 + 0. 022 3.4+0.8 WiMisoss. 5... 10 1. 8-2. 4 2.18+0. 05 0.17 +0. 04 7.641. 7 Wipe ee 12 2. 5-2. 9 2. 6740. 04 0.13 +0. 03 4.7+1.0 WiiMote 232 11 1. 9-2. 4 2.17+0. 03 0. 114+0. 024 5. 2+1.1 BMeousreoet Uf 2. 6-2. 9 2. 79+0. 04 0.10 +0. 03 3. 5+0. 9 WM, .=-=-—- 7 1. 7-2. 1 1. 90+0. 05 0.12 +0. 03 6.3+1.7 1 Ey tat a ie 8 Py 2. 84+0. 06 0.18 +0. 05 6. 341.6 WIR Ace ne cece oe 8 3. 5-4. 3 3. 88+0. 08 0.22 +0. 05 56 641.4 Mts = = 8 2. 5-2. 8 2. 68+0. 03 0. 097 +0. 024 3.6+0. 9 Wail aa ea ee Gi 3. 6-3. 9 3. 80-40. 04 0.11 +0. 038 2.8+0. 8 Mtn oo seele 3 4 2. 2-2. 6 2. 43 Wiese oe 3 3. 9-4. 1 4. 00 CNS ee 2 1. 5-1. 9 1bea70) Wi Meet ctr eas 2 3. 3-3. 6 3. 45 FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 127 Family MIXODECTIDAE Cope, 1883 Previous views as to the affinities of this peculiar group have been summed up by Matthew (1909, 1915, and Pale. Mem.). In briefer résumé, Cope considered the mixodectids as primates. Matthew (1897) concurred formally but suggested that Mizodectes itself might rather be a rodent. Osborn (1902) accepted and expanded this view, defining for the mixodectids a rodent suborder Proglires. Wortman (1903) argued for their return to the Primates. Finally, Matthew (1909 and subsequently) concluded that they probably belong in the Insectivora. Two very distinctive genera, Hudaemonema and Elpidophorus, have recently been added to the family. They add to the known variety and have an interesting bearing on relationships within the family, as brought out below, but they do not much alter the evidence for ordinal relationships. Skeletal remains referred to Indrodon by Osborn and Earle and to Microsyops by Wortman, in each case considered as indicating primate affinities, have been shown (Matthew, 1909) to be doubtfully or not associated. An astragalus and other fragments referred to Mizxodectes were at first said by Matthew (1897) to be rodentlike and later (1909) to be equally insectivorelike and in any case not similar to any known primate. The dentition is said, even by Matthew in rejecting primate affini- ties, to be primatelike. There is, indeed, some resemblance to various primate genera in a few details, for instance the upper molars suggest Shoshonius in general proportions and in the strong mesostyle, and the enlargement of one anterior tooth and development of a diastema by loss of other teeth are also seen in Tetonius and some other genera. Such resemblances seem to have no value in the determination of affinities, since they refer to single characters of various different primates that are, in just these characters, highly aberrant among the primates as whole. Aside from such points, which can almost be discarded categorically as more likely to be convergent than not, I detect no primate resemblances in the teeth that go beyond the general Paleocene tuberculosectional pattern common to many different orders at this time. As set forth in the section of this paper dealing with the Primates, the Paleocene and Eocene primates, despite their primitive character, do have a distinctive stamp in molar pattern that is common to all of them and that is not seen to occur in any other order. The mixodectids do not have any of these truly distinctive and (at this time) ordinal primate characters. On the contrary, their high sharp cusps (notably in the talonids), elevated trigonids and internal lower cusps, displaced hypoconulids, and many other dis- tinctive details are quite unknown among any primates and some of 128 BULLETIN 169, UNITED STATES NATIONAL MUSEUM their family characters, such as the ectoloph construction, are ap- proached only as decidedly aberrant generic characters by one or a few primate genera. If we grant Matthew’s second opinion that the astragalus of Mizo- dectes does not have diagnostic ordinal rodent characters, the evidence for rodent relationships is little more than the presence of enlarged in- cisors and (in some but not all genera) of more or less molariform pre- molars. Both these characters also appear independently in insecti- vores, numerous different lines of ungulates, primates, marsupials, end other orders. Nor are they really rodentlike in precise detail in this group. Indeed, there now seems to be no actual evidence that the mixodectids are related to rodents. Granting the usage of Insectivora not only to include the recent groups but also numerous extinct forms that necessitate definition on primitive characters only, we may well call the mixodectids insecti- vores. Negatively, it may be said that no other defined order could receive them, and positively that their dentition is insectivorelike at least in habitus, that the astragalus is more like that distinctive of insectivores than like any other group except rodents (which are excluded by the dentition), and that in general they have the primitive features by which the Insectivora sensu lato are defined. It is probable that the mixodectids include a related group of phyla that diverged from the primitive placental stock, and apparently from the Insectivora in a more limited sense, at a very early date. Had such a sideline evolved more rapidly, or had it run a longer span and occupied a more important place in mammalian history, it would be more conveniently defined as an order, as, for instance, are the tillo- donts, which probably had a very similar history but developed more striking specializations. Since, in fact, the mixodectids were a short- lived and relatively unimportant group, it is most convenient simply to classify them in the order Insectivora, from which they probably arose. Within the Mixodectidae there have been included two apparently distinct groups: Mizodectes and Indrodon of the Torrejon, on one hand, and Cynodontomys and Microsyops of the Lower and Middle Eocene, on the other. The Torrejon forms are certainly closely re- lated, indeed the distinction between them is not clear, and the Eocene genera are also closely allied and rather difficult to distinguish. Matthew (1915c) defined the two groups as subfamilies, Mixodectinae and Microsyopinae, and he repeatedly expressed doubts as to their really being related to each other, rather than merely convergent. In the lower jaw (the upper being uncertainly known in this respect) the ‘‘Mixodectinae” retain a canine, and the enlarged tooth is an incisor, while in the ‘‘Microsyopinae”’ there is only the enlarged tooth FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 129 anterior to P*. The enlarged teeth are thus not rigidly proved to be the same, and with his customary caution Matthew stressed the fact that they might not be homologous. The premolars are different in the two groups, although I believe that the difference has been over- emphasized. In the ‘‘Microsyopinae” P*, are less elevated and more molariform than in the ‘“Mixodectinae.’”’ The molars are almost identical in the two groups, except that in the earlier forms there is a distinct and markedly internal hypocone, while in the later the hypocone is anomalously small and is less internal. The two genera discovered since Matthew’s work, especially Eudaemonema, alter this situation. Hudaemonema has the anterior (lower) dentition less specialized than in Mizodectes, and Elpidophorus has it either closely similar to Mixodectes or slightly less reduced. In both cases there is no reason to believe that the enlarged incisor is not homologous with that of Mixodectes and also with that of Cyno- dontomys. In both genera P, is submolariform. In Hudaemonema it is very similar to that of Cynodontomys and Microsyops, rather than to the contemporary Mizodectes. In Elpidophorus it is aberrant in having a strong, projecting paraconid. P* is not known in Hudae- monema; in Elpidophorus it closely resembles that of the Eocene genera. In Hudaemonema the upper molars have strong, internal hypocone, comparable to Mirodectes, while Elpidophorus is more like the Eocene genera in this respect. These genera thus mingle characters of the “Mixodectinae” and the ‘‘Microsyopinae”, and they make it impossible to maintain a consistent separation between these groups. At the same time they strongly support the reality of a relationship between them. For one reason or another the known Paleocene genera of this family cannot be considered ancestral to each other or to the Eocene forms. Thus Eudaemonema is the most primitive as regards the retention of anterior teeth, but its premolars are more advanced than in Mizodectes and Indrodon, and its strong, projecting hypocones seem to exclude it from the ancestry of Elpidophorus or of Cynodontomys and Microsyops. Mirxodectes and Indrodon cannot be ancestral to the later forms for the same reason and cannot be structurally ancestral to the contemporary Hudaemonema because of their reduced anterior dentition. Elpidophorus has aberrant specializations in P, and in details of molar structure that seem to exclude it from consideration as the ancestor of the Eocene genera. Evidently four different groups, intimately related but all on different lines of phyletic descent, are represented. These can be contrasted, among other details, by the characters listed in table 28. 130 BULLETIN 169, UNITED STATES NATIONAL MUSEUM TABLE 28.—Comparison of dentition of six genera of Mixodectidae Genus Mizodecies and Indrodon. Eudaemonema-- Elpidophorus-.- - Cynodontomys and Microsy- ops. Formula AES ER hs Se 7) I eae 1(+?).1.3.3- 1:03:37 — Py Elevated, paracon- id minute, meta- conid rudimen- tary or absent, talonid poorly basined with 2 cusps. About as tall as mo- lars, paraconid small, metaconid large, talonid well basined, with 3 cusps. With paraconid large and project- ing, otherwise comparable to Eudaemonema. Comparable to Eu- daemonema. Internal] lower molar cusps Slightly or not taller than ex- ternal. Meta- conid opposite or slightly pos- terior to proto- conid. About as in Mizo- dectes. More definitely higher than ex- ternal. Meta- conid — slightly anterior to pro- toconid. About as in Mizo- dectes. Pi Elevated, ex- ternal cone simple. (Unknown) --_-- Less elevated, with separate paracone and metacone. Comparable to Ilpidophorus. Upper molars With strong hypocones, projecting internally. About as Mixodectes. in With weak hy- pocones, not projecting internally. Comparable to Elpidopho- Tus. A possibility worthy of serious consideration is that the Plagio- menidae may be related to the Mixodectidae. ing principal characters suggestive of such a relationship: They show the follow- Molarization of premolars somewhat similar to mixodectids (except Mixodectes and Indrodon). Upper molars with feeble hypocone shelf and pronounced and peculiar median transverse valley (asin Elpidophorus and to a less degree some other mixodectids), and otherwise generally but more vaguely similar. Tendency to emphasize external shelf of upper molars and to develop cuspules on it (but see below). Elevation of internal over external cusps of lower molars (a tendency in all mixodectids, pronounced in Elpidophorus). Lines joining protoconid to metaconid and hypoconid to entoconid parallel and anterointernal-posteroexternal (tendency in mixodectids, pronounced in Elpi- dophorus). Paraconids, and trigonids generally, similar. Molar talonids markedly broadened. These, and a few minor details, produce a general type of dentition common to plagiomenids and mixodectids and not, as far as I know, to any other groups.” 85 This was evidently partially noticed by Matthew (1918, p. 600), for he notes the resemblance to Plagio- mene of “an undescribed genus from the Paskapoo beds’’, which was undoubtedly Elpidophorus, not then named and not until very recently known to be an aberrant mixodectid. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 131 On the other hand, the following characters show that the relation- ship cannot be very close, if it exists at all: Rather than a single mesostyle developed as a fold of the ectoloph, the plagio- menids have two independent median cusps on the outer shelf of the upper molars. The anterior incisors (in the lower jaw, at least) are not reduced in number and none is markedly enlarged. The hypoconulid does not have the characteristic mixodectid displacement toward the entoconid. The cheek teeth are all deeply furrowed and tend to proliferate cuspules. The inconclusive evidence of possible relationship of the plagio- menids to the Dermoptera has not been significantly altered since Matthew wrote (1918). The chain of evidence thus tending to link the mixodectids with the Dermoptera is so weak at every point as not to merit serious consideration at present. Genus EUDAEMONEMA Simpson Eudaemonema Simpson, 1935d, p. 231. Type.—Eudaemonema cuspidata Simpson. Distribution Middle Paleocene, Fort Union, Mont. Diagnosis.—Dental formula, 3773; Median incisor enlarged. Canine larger than lateral incisor or P;. P;_;small, 1-rooted. P, submolari- form, with small paraconid, large, high metaconid, and basined, tricuspid talonid. Lower molars as in Mizodectes, but trigonids more elevated and all six cusps sharper and more distinct. Upper molars with prominent hypocones, projecting strongly internally. EUDAEMONEMA CUSPIDATA Simpson Ficures 25, 26 Eudaemonema cuspidata Stupson, 1935d, p. 231. Type.—U.S.N.M. no. 9314, left lower jaw with C, P,-Ms, and roots or alveoli of all other teeth. Found by Dr. J. W. Gidley. Horizon and locality—Gidley Quarry (referred specimen from Silberling Quarry), Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis.—Sole known species of genus. Dimensions in table 29. Remarks.—Alveoli in the type and roots in another specimen (no. 9317) show clearly that there was an enlarged median incisor with a procumbent, laterally compressed root, flanked by one much smaller lateral incisor. The canine root is oval, implanted nearly vertically, nearly as large as that of the median incisor, and much larger than that of either of the adjacent teeth. The crown is low, with a simple, spatulate, recurved tip. P, is not known, but its alveolus in three different specimens suggests that it was constantly present, although absent in all other known members of this family. Its root is single, small, and circular in section. 132 BULLETIN 169, UNITED STATES NATIONAL MUSEUM P, has a single but larger root. Its crown is simple, with one cusp, convex on the outer face and excavated anterointernally near the tip, followed by a small, mainly internal heel, not forming a definite cusp. P; is larger and has two poorly separated roots. The anterointernal excavation is accentuated and tends to pinch off a very rudimentary paraconid, but there is no trace of a metaconid. The heel rises to a single, distinct cusp. P, is nearly molariform. The paraconid is small but distinct and is median. The metaconid is nearly as high as the protoconid, with which it is partly confluent. The talonid is well developed, of about the same width as the trigonid, and has a large hypoconid and smaller entoconid and hypoconulid, the latter median. > sD Z A \) en FIGURE 26.—Eudaemonema cus- pidata Simpson, U.S.N.M. no. FIGURE 25.—Eudaemonema cuspidata Simpson, U.S.N.M. no. 9314, 9558, left upper molars: a, Ex- left lower jaw: a, Crown view; 6, internal view. Three times natu- ternal view; 6, crown view. ral size. Three times natural size. M,-» are closely similar to each other. The trigonid is moderately elevated and is narrower than the talonid. The paraconid is low and shelflike but is distinct and is united by a crest to the crescentic pro- toconid. The paraconid is about intermediate between the median and internal positions. The metaconid is slightly higher than the hypoconid, and the hypoconulid is lower and near the entoconid but more distinct from it than in Hlpidophorus. All three talonid cusps are unusually lofty and distinct. A cuspule tends to develop on the hypoconid-metaconid ridge. Ms; has the talonid narrower, about as wide as the trigonid, and the hypoconulid projects posteriorly and is higher than the other talonid cusps. Of the four specimens that show the posterior mental foramen, it is beneath P, in two (including the type), is double with both under P, in another, and in the fourth is beneath P, but is followed by another of much smaller size under M,. In all cases the anterior mental foramen is larger, pronounced, and beneath P,. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 133 U.S.N.M. no. 9558, from the Gidley Quarry, is a left upper jaw with M!3. It closely resembles Mixodectes, occludes well with some of the lower jaws of Hudaemonema cuspidata, and is not what would be expected in the upper jaw of any other known species, so that I place it here with some confidence. Aside from details of proportion, as shown in the figures, the outstanding differences from Mizodectes are the better development of the conules, especially the metaconule (almost lacking in Mizodectes), and the even greater internal displace- ment of the hypocone of M?’. TABLE 29.— Measurements (in mm) of lower teeth of Eudaemonema cuspidata P2 P3 P4 Mi M2 M3 U.S.N.M. no. | ee | ee) Se a ae L WwW L WwW L WwW L WwW L Ww L WwW EK: eS a ee Oe 1.8 3 2.0 1.4 Perl 1.9 3.3 257 OEE: ag 4.0 2.6 (BY as he a ee el ee ae 2.6 TOS 525 [bee 3:5 Ai OS eee ee eae S ene ee SE See 1.9 155 2.9 2.0 By E 2.6 3.6 2.8 SHI to Pe 5 Se oe er eae | (eee | oy (Bee Se | ee) ae eee ee eee 3.8 2.9 Bul} Se See eee Le eee 1.9 1.6 PIT PP) 3.0 prc! 3.4 2.9 YB SV ee ee el [mete pel ee Ee eee | [ns Pari 1.9 3.2 7 af 355 2.9 Genus ELPIDOPHORUS Simpson, 1927 ELPIDOPHORUS MINOR, new species FiGureE 27 Type.—Princeton no. 14201, left lower jaw with P;-M». Collected by A. C. Silberling. Horizon and locality—Probably Silberling Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis —Smaller than EF. elegans or E. patratus. P? slenderer than in E. patratus, paraconids P,-M, more strictly internal, heel of P, smaller and less strongly basined, elevation of inner cusps P,-M, distinct but slightly less pronounced than in EF. patratus. Remarks.—The type and only known specimen was found by Silber- ling on January 18, 1903, and is labeled as from the Torrejon, Locality No. 2 (not of the serial list later started and employed in this work). Mr. Silberling states that these data mean the No. 2 beds at or near the Silberling Quarry. It is extraordinary that no further material of this peculiar form seems to occur in the much larger collections made subsequently. The species is very distinct from the later E. patratus of this field, and it will probably prove to be generically different, but they are certainly allied and the present data do not seem to warrant generic definition. EH. minor resembles Eudaemonema more than does £. 134 BULLETIN 169, UNITED STATES NATIONAL MUSEUM patratus, but it is closer to the latter than to E. cuspidata and shows the Elpidophorus line to have been distinct at this time. Dimensions of the type are as follows: Length P3, 2.0; width Ps, 1.4; length P,, 2.7; width P,, 1.9; length My, 2.8; width M,, 2.6; length M,, 3.0; width Mg, 2.8. INN 2 te FIGURE 27.—Elpidophorus minor, new species, Princeton Univ. no. 14201, left lower jaw: a, Crown view; 6, internal view. Six times natural size. ?INSECTIVORA, incertae sedis PICRODONTIDAE, new family Type.—Picrodus Douglass, 1908. Distribution.—Middle and Upper Paleocene, North America. Diagnosis—Minute insectivorelike or batlike forms of doubtful ordinal affinities, with one pair of greatly enlarged incisors (at least in lower jaw), muzzle long and slender, canines reduced, premolars small with no tendency to molarization, molars large, brachyodont, with shallow, expanded basins and indistinct cusp structure, adaptively resembling molars of the recent Phyllostomatidae. Discussion.—At present only two genera, Picrodus and Zanycteris, are referred to this family, and these are not directly comparable with each other. The evidence for their close relationship is, however, impelling, as discussed below. The separation of family characters from those merely generic is not entirely practicable in this stage of knowledge, but there can be little doubt that these two genera do belong to a family otherwise unknown, and the characters given in the above diagnosis distinguish them from any other family. The FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 135 molars invite comparison with the Phyllostomatidae, although they do not prove that a real relationship exists. The character of the antemolar dentition distinguishes the Picrodontidae sharply not only from the Phyllostomatidae but also from all other Chiroptera. The affinities of the Picrodontidae are wholly dubious at present. They compare in a very broad and general way with the Insectivora, Chiroptera, and Primates. Reference to the Primates is merely a possibility, with no positive evidence to commend it. Evidence for reference to the Chiroptera is seen in the phyllostomatidlike molars but is really very tenuous and does not at present warrant the extraordi- nary conclusion that the Chiroptera had already in the Middle Paleo- cene achieved this peculiar and aberrant molar pattern and at the same time had lost, or not yet acquired, characters otherwise universal among chiropterans. Such references, even when circumspectly expressed, are moreover likely to be misleading, for they inevitably are restated in more general works by authors not acquainted with the original material, in some such form as “Specialized phyllostomatid bats were already present in North America in the Middle Paleocene”, without the necessary addition that the evidence actually falls far short of proof. It is more conservative and less prejudicial to future work to refer the Picrodontidae to the ?Insectivora, using Insectivora in its scrap-basket sense, pending discovery of more conclusive indica- tions of affinity. When these are discovered, they are (as far as can be foreseen) as likely to point to the Insectivora as to any other order, Genus PICRODUS Douglass, 1908 Picrodus Douatuass, 1908, p. 17. Megopterna Dovaeuass, 1908, p. 18. Type.—Picrodus silberling: Douglass. Type of Megopterna.—Megopterna minutia Douglass. Distribution.—Middle Paleocene, Fort Union, Montana. Diagnosis —Enlarged, procumbent anterior lower tooth, followed by three or four small teeth, the most posterior (Ps?) 2-rooted but small and simple. My, much enlarged, with a small, elevated, and procum- bent trigonid with three poorly differentiated cusps, heel elongate and large, with a curving crest and two vague internal cuspules, basin not closed. M, with lower, subquadrate but 3-cusped trigonid, large, oval, basined talonid with crest and two internal cusps. Enamel of both talonids papillated. Remarks —From Douglass’ specimens and, still more, his some- what diagrammatic figures 1t would appear altogether impossible that Picrodus and Megopterna should be synonymous, but this is shown to be true beyond any question by the larger series of specimens now available. The type of Picrodus included P, and M,, the latter im- perfect, and that of Megopterna included M, and a small fragment of 1192123710 136 BULLETIN 169, UNITED STATES NATIONAL MUSEUM M, that Douglass mistook for a complete tooth. The type species are synonymous. Douglass referred Picrodus questionably to the Epanorthidae (=Caenolestidae) and Megopterna questionably to the Insectivora, without family reference. The resemblance to caenolestids is con- fined to a vague adaptive similarity to some fossil forms with enlarged M, and is not indicative of affinity. Picrodus is almost certainly a placental mammal. Among placentals, however, I am not acquainted with any genus with which close and direct comparison is possible. There is, indeed, a vague resemblance to certain highly specialized recent bats, but this does not extend to structural details, is contra- dicted by the quite different arrangement of the anterior dentition, and is more likely to be misleading than not. There is one known genus, Zanycteris, with which Picrodus is almost certainly closely related, although direct comparison is impossible since in Picrodus only lower and in Zanycteris only upper teeth are known. As I have elsewhere noted (Simpson, 1935a), Zanycteris (like Picrodus) resembles some recent bats, particularly the phyllo- stomatids, in adaptive characters of the cheek teeth but is different in details probably of more importance as indices of affinity and in the structure of the anterior dentition, as far as it is known. The great probability of affinity between Picrodus and Zanycteris is inde- pendent of the possibility that they are related to the Chiroptera. In Zanycteris the reduction and complete lack of molarization of the premolars, the enlargement of M' and reduction of M?, and the peculiar papillated coronal enamel are all unusual specializations analogously developed in the lower dentition of Picrodus. Furthermore, even in detail the shapes of M'!~? in Zanycteris adapt them perfectly for occlu- sion with lower teeth like those of Picrodus. Zanycteris paleocena will not occlude with Picrodus silberlingi, being a smaller species, but probably a dentition structurally the same as that of Zanycteris but of different size would occlude with Picrodus silberlingt. Zanycteris is known only from one specimen found in the Tiffany, Upper Paleocene, of southwestern Colorado. Its type is certainly not the same species as that of Picrodus, and the genera are probably distinct, but not surely. Knowledge of their exact affinities must await discovery of upper teeth of Picrodus or lower teeth of Zanycteris. PICRODUS SILBERLINGI Douglass Figure 28 Picrodus silberlingi Doucuass, 1908, p. 17. Megopterna minuta Douatass, 1908, p. 18. Type.—Carnegie Mus. no. 1670, right lower jaw with P,-M,. Collected by A. C. Silberling. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 137 Type of Megopterna minuta.—Carnegie Mus. no. 1675, left lower jaw with M, and part of talonid of M;. Collected by A. C. Silberling. Horizon and locality —Types from Silberling Quarry, most referred specimens from Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis.—Sole known species of genus. Dimensions given below. Remarks.—From alveoli it is clear that this species had a much enlarged, procumbent anterior tooth, probably an incisor, with a compressed root. U.S.N.M. no. 9866 includes M, of P. silberlingi and also a loose tooth, which probably is an associated lower incisor. It has a completely enameled crown, curving sharply to a point. The whole crown has a series of eight or nine ridges or angulations, diverg- ing posteriorly from the apex, so that in transverse section it is irregularly polygonal. Aside from these, one side is more convex, the other somewhat excavated, with a slight basal cingulum. Pp GE: a LES G Z ek f YZ FE el y (I \ SEZ FIGURE 28.—Picrodus silberlingi Douglass, U.S.N.M. no. 9622, right lower jaw: a, Crown view; 6, external view. Four times natural size. The anterior tooth is followed by three small, closely spaced, ap- proximately equal alveoli. The material does not indicate whether these were for three separate teeth or for one 1-rooted tooth and one 2-rooted. The next tooth, presumably P,, has two roots in the several specimens that show it, not one as stated by Douglass for the type. It has a simple main cusp, more procumbent than shown in Douglass’ figure, followed by a small heel. The next tooth, pre- sumably M,, is the largest in the jaw and is very peculiar. It differs considerably from Douglass’ figure, although I believe that the pres- ent specimens do belong to his species and that the discrepancy is due to the worn and broken nature of his specimen and the impos- sibility of accurate observation except under a binocular microscope at magnifications of 15 or 20 X. This tooth consists of a trigonid and talonid, but both are greatly modified. The trigonid is small and is produced and procumbent, as if drawn forward and upward in a plastic condition. The protoconid lies near the midline of the tooth as a whole. The metaconid is slightly lower, poorly separated from the protoconid, and internal and slightly posterior to the latter. The 138 BULLETIN 169, UNITED STATES NATIONAL MUSEUM still lower and likewise poorly separated paraconid is almost directly anterior to the protoconid but slightly more internal. The talonid is wider than the trigonid and is very long. It has a crest that begins rather indefinitely on the external side against the base of the trigonid and curves back to the posterointernal corner of the tooth. Its highest part is where it swings internally and across the midline of the tooth, and here it bears two or three vague cuspules. There are also two small cuspules on the inner margin of the talonid, separate from the crest and at a lower level. The more definite of these is just anterior to the posterointernal corner, and the other vague cusp lies between this and the metaconid base. The sloping and volute surface of the talonid, from the crest down to the inner margin, Is. finely papillated and wrinkled. The lowest point of this surface is. at the posterointernal corner, where there is an almost spoutlike exit. from the vague talonid basin. Mz was figured by Douglass (“‘Megopterna minuta’’), but his draw- ing makes the cusps appear more upright, sharp, and distinct than they really are. The trigonid suggests that of M; but is much less. elevated, the cusps are better separated, the paraconid is more inter- nal, and the trigonid is given a more quadrate form by the angulation of the crest connecting protoconid and paraconid. The talonid is broad and oval, less sloping and more distinctly basined than that of M,. The crest defines the posteroexternal angle, instead of curving obliquely across the tooth as on M,, and there are two distinct inter- nal cusps, the more posterior of which is connected to the crest. The basin surface is papillated as on M,. M; is not preserved on any specimen in the collection. From its alveoli, it was smaller than M,. Upper teeth have not been recog- nized. From the downward curvature of the lower margin posterior to the dental region, it is evident that the angie was of placental type. The mental foramina are numerous and variable. There may be a cluster of three or four in the general region of Py. The most constant appear to be a fairly large foramen approximately between P; and P, and a smaller one about between P,; and M,. The type is not well preserved, and I have not remeasured it. It may be slightly larger than the other specimens, but there is no doubt that all are conspecific. The most reliably measurable dimension is the oblique maximum diameter of My, that is, a dimension in the midline in a vertical (but not also a horizontal) plane, from paraconid to base of posterior end of talonid.*® The constants of this dimen- sion in the National Museum sample are: N, 8; R, 2.6-2.9; M, 2.7140.04; o, 0.105+0.026; V, 3.9+1.0. The width of this tooth 86 The ordinary length, a horizontal between transverse, vertical tangential planes, would show great subjective and accidental variation in such a small and peculiarly oblique tooth. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 139 is not so accurately measurable but is recorded within the range 1.0 to 1.3, mean 1.16, for these eight specimens. P,, poorly preserved and very difficult to measure, has a maximum oblique diameter of about 1 mm and a width of 0.6 or 0.7 mm. M4, is well preserved only in the type of Megopterna minuta, where it is 1.4 mm in length and about 1 mm in width. The less completely preserved specimens of this tooth do not suggest a deviation of more than 0.1 mm from this. The one Silberling Quarry specimen in the National Museum col- lection has M, with the dimensions 2.8 and 1.1, well within the range of the Gidley Quarry material. Family Uncertain Genus and species undetermined Figure 29 U.S.N.M. no. 9777, from the Gidley Quarry, is a right humerus of a fossorial mammal of about the size of a recent Scalops. It is imperfect but preserves highly characteristic features. ‘The laterally compressed head, short, stout, twisted shaft, and flattened, widely flaring distal end are disposed in such a way that if the head was directed posteriorly, the entepicondyle was anterior and only very slightly internal in position, and the ectepicondyle similarly posterior, so that the lower arm was thrown outwards almost at right angles to the body. The bicipital groove, mainly occupying the proximo- internal quarter of the posterior face, is deep and narrow, bounded by sharp crests and elevated tuberosities, which are, however, broken off. The pectoralis major insertion is broad and shallow, occupying most of the proximal half of the anterior face of the shaft, and not sharply bounded distally. The deltoid process is broken, but from its base it was more prominent and more internal than in Scalops, more proximal and heavier than in Arctoryctes. The notch between head and ectepicondyle is nearly semicircular. The distal end has greatly produced epicondyles, the extension of the ectepicondyle far beyond the globular capitulum being especially striking, in comparison with Scalops. The other distal articulations are poorly differentiated or preserved. The entepicondylar foramen is strangely developed as a long, small canal, running from the posterior face near the internal margin to the middle of the anterior face of the broad distal end. This peculiar humerus resembles those of recent moles in many respects but also differs throughout in detail. Unquestionably the resemblance bespeaks similarity of habits. Whether it also indicates phylogenetic affinity is quite uncertain. In some respects resemblance is closer to Arctoryctes from the Oligocene (see Schlaikjer, 1933), but there are also numerous differences: The distal end is more nearly parallel to the long (anteroposterior) axis of the head; the deltoid 140 BULLETIN 169, UNITED STATES NATIONAL MUSEUM process is less internal, stouter, and more proximal; the ectepicondyle is more produced; the entepicondylar foramen, or rather canal, is longer; and other differences of proportion and detail are seen. Never- theless a relationship seems probable. The Arctoryctes humerus was supposed by Matthew to belong to a chrysochlorid, but Schlaikjer has shown that the evidence is all against this view and favors talpid affinities. Arctoryctes may belong with the dentitions and skulls. known as Proscalops. FIGURE 29.—Humerus of an unidentified fossorial mammal: a, Anteroexternal face; b, posterointernal face. Four times natural size. On the basis of the teeth, no genus known from the Gidley Quarry would seem to be closely allied to Proscalops. Some nyctitherids have been supposed to be talpids or at least talpoids, but the only probable nyctitheriid in this fauna, Stilpnodon simplicidens, is too small to have had this humerus. The humerus does not belong to a multituberculate nor to any other order known in this fauna save the Insectivora. It is not leptictid and cannot belong to Aphronorus if that genus is really a pantolestid, but might if the genus does not belong with Palaeosinopa and Pantolestes. It might belong to Gela- stops but probably does not if that genus is correctly considered an ally of Didelphodus. Eudaemonema shows some, but only very distant, resemblance to such dentitions as Proscalops and might conceivably have had a fossorial humerus. The dentition of the animal represented by this humerus may be unknown, although this is improbable in view of the many jaws and few humeri collected from the quarry. In any event the presence of such a specialized fossorial animal in this ancient fauna is of great interest. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 141 Order PRIMATES Linnaeus, 1758 The Fort Union primates are of exceptional interest as the oldest known members of the order to which man belongs, and any light that they can cast on the early history of this great group is highly important. Dr. Gidley fully recognized these facts, and when the arduous task of preparation was finally completed, he turned first to the primates in beginning his definitive work. His previous papers were all preliminary and provisional, but he completed the primate section of his proposed memoir and published it in 1923 as a separate paper, later to be united with the other proposed sections into a single monograph. No other section was ever finished, and the pri- mate paper was Gidley’s last contribution to the Paleocene.*’ When these primates were discovered they were far the oldest known. Many primates were known from the Eocene of Europe and North America, but only one, Plesiadapis, was known from the Paleocene, and this was considered as only very doubtfully primate and is considerably younger than the Fort Union primates of Gidley’s collection. After Gidley’s discovery, but before its publication, Matthew (1915) added Nothodectes (=Plesiadapis) also from the Paleocene, and in 1921 Matthew and Granger added several more genera from the Tiffany, but these are all younger than Gidley’s material. Jepsen and I have made recent additions to the known upper Paleocene primates, but only Plesiolestes Jepsen, 1930, is of an age comparable to Gidley’s genera, and there is no reason to suppose that it is older.® It is in accord with Gidley’s intention that these forms are here redescribed in connection with the whole fauna, despite their publica- tion previously. This is the more necessary because since Gidley’s publication knowledge of early primates has been greatly increased both by discovery and by revision, calling for reconsideration of many points that he mentioned. For this reason, his diagnoses and discus- sions are not quoted in full, but are revised in the light of the wider knowledge of today and of the somewhat different conclusions to which this has led me. Gidley recognized four new genera in this fauna and placed six species in them. The fact that two species are based on upper jaws, with lowers referred, and the other four on lower jaws, with uppers referred to three of them, introduces a slight element of doubt, but on the basis of the lower jaws, at least, it is certain that six species are represented and that Gidley’s identifications of all these specimens are "With the minor exception of a very brief note on the Tiffany. 58 Abel (1931) lists Plesiolestes as from the Lower Paleocene and suggests that it is the oldest known primate, but it is from the middle Paleocene and not appreciably, or not at all, older than the genera here discussed, one of which it closely resembles. 142 BULLETIN 169, UNITED STATES NATIONAL MUSEUM correct. The only taxonomic change to be introduced is that a fifth genus is made to receive a species that Gidley placed doubtfully in Palaechthon. Family 7ANAPTOMORPHIDAE Cope, i883 In this fauna the three genera Paromomys, Palaechthon, and Palenochtha are very tentatively listed with the Anaptomorphidae, although (as will appear) the relationship is not clear and this whole complex of early primates is highly polyphyletic and very confusing. Paromomys, Palaechthon, and Palenochiha evidently belong to slightly divergent lines, but they have certain characters in common. The most important of these are: 1. An enlarged, semiprocumbent lower incisor, its root not extending beneath P,. . Other lower incisors vestigial and variable or absent. . Lower canine present and only slightly reduced. P, and probably sometimes P, absent.*? . Py little or not enlarged, trigonid simple, elevated, with low, 2-cusped talonid. . Molar trigonids with small, generally short and quadrate basins, paraconids meneralle: distinct but reduced, cusps marginal or nearly so. 7. Heels of M,-» large, simple, broadly basined. 8. M; with third lobe, which, however, differs greatly and characteristically in the three genera. 9. P3 (as far as known) 2-rooted and not transverse. 10. P4 transverse, strong protocone, paracone and metacone little or not differ- entiated, conule feeble or absent. 11. Upper molars without mesostyle, protostyle, or hypocone; two small but distinct conules; posterointernal corner of crown expanded and basined to varying degrees; inner face of molar with vertical groove at least on M?. Oop wry These suggest a possible fairly immediate common origin for the three genera, but they are diverging from each other, principally as follows: Paromomys: Antemolar dentition of unmodified basic type as listed above. Molar trigonids very short and quadrate, paraconid almost disappearing by fusion with metaconid on M23. Third lobe of M3 very strong, with at least two distinct cusps. Inner base of M?, at least, more or less bilobed. Posterointernal expan- sion of upper molars very marked. Palaechthon: Ps more progressive, with distinct paraconid and metaconid. Molar trigonids less quadrate and paraconids more distinct. Third lobe of Ms; weaker but still with two cusps. Upper molars not bilobed, posterointernal expan- sion less. Palenochtha: Anterior dentition further modified by loss of another tooth, prob- ably P,. Py much asin Paromomys. M3; with weak, l-cusped third lobe. Upper molars not bilobed, posterointernal expansion slight. Except for the absence of P, (which, however, is not absolutely cer- tain), Palenochtha seems definitely the most primitive of the three and its general structure is such as might be expected in the ancestry of 59 Certain homologies, discussed om a later page, are here assumed. Exact identification is impossible, but as the teeth are probably homologous between the genera in question, if they are labeled consistently it does not matte from the point of view of determining affinities whether the labels prove to be correct or not, FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 143 both the other genera. Paromomys and Palaechthon show crossing specializations. The general molar structure of Palaechthon is specialized more or less in the direction of Paromomys but is less aber- rant, while its P, is definitely more progressive. Paromomys shows distinctly the most aberrant molar structure but has P, still relatively unproeressive. The only known primate of comparable age is Plesiolestes Jepsen, from a Torrejon equivalent in the Fort Union of northern Wyoming. Its age is not appreciably different from that of the Gidley Quarry spec- imens, and the geographic locality is not very distant, all occurring in the same widespread formation. Jepsen tentatively referred his genus to the Plesiadapidae but noted (1930a, p. 506) that “there are many structures on the two specimens which are not like those of other Plesiadapids.”” He did not compare with Gidley’s previously pub- lished genera, which Plesiolestes resembles in many ways. The anterior alveoli show an enlarged semiprocumbent incisor and a smaller, less procumbent canine, as in all three of Gidley’s genera here discussed, and also a moderate P,, as in Paromomys and Palaech- thon. Ps is also closely similar, but relatively larger, being about as high as P,, whereas in Gidley’s genera it is lower. P, closely resembles that of Palaechthon, the only difference clear from the available data being that in Plesiolestes the heel is wider and the paraconid and meta- conid stronger, especially the latter. The molars are very similar to those of Palaechthon and seem to me to show no difference of probably generic value. While Plesiolestes may be provisionally accepted as valid, chiefly on the basis of the more progressive P,, it is almost surely very closely related to Palaechthon, and the distinction of the genera is not at pres- ent wholly satisfactory. If, as is possible, Plesiolestes is somewhat younger, it could well be a slightly modified and progressive descend- ant of Palaechthon. Its diagnostic features, as against Palaechthon, seem to me modifications away from as much as toward the plesiada- pids. In any case it is surely closer to Palaechthon than is either genus to any undoubted plesiadapid, and if a plesiadapid relationship exists at all, Palaechthon is probably less removed from that line than is Plesiolestes. Palenochtha, the least aberrant of the present genera as regards comparison with an abstract protoprimate dentition, seems to resemble the Eocene tarsioids (in the broadest sense) more than any other known mammals. The specialization of the anterior lower teeth is not exactly as in any later tarsioid but is within the apparent poten- tialities of the group. Omomys, from the Bridger, has two enlarged 60 The American lower Eocene species placed in Omomys are very doubtfully congeneric with the Bridger genotype. As noted by Teilhard, the European lower Eocene specimens are very distinctive and might be, in my opinion definitely are, representative of a different genus but one close to and perhaps structurally ancestral to true Omomys. 144 BULLETIN 169, UNITED STATES NATIONAL MUSEUM anterior teeth followed by three premolars, but in it there is a small tooth between these two enlarged teeth, which is either much more reduced or wholly absent in the much older Fort Union genera. Absarokius may very closely resemble the Fort Union genera in the anterior teeth, although this is very dubious, as they are known in Absarokius only from poorly preserved alveoli of one specimen, which seems to show less disparity between the incisor and the ?canine. Tetonius and the European Necrolemur have a single enlarged anterior tooth,® a condition that could be derived from that of the middle Paleocene genera, although there is no adequate evidence that it was so derived. P, is more primitive in Palenochtha than in any later tarsioid genus known to me, but the difference in such forms as Anaptomorphus or “Omomys”’ belgicus is not marked, and as the increasing and diverg- ing specialization is in keeping with the relative ages it has no crucial bearing on general affinities. The lower molars of Palenochtha are much like those of ?Omomys vespertinus, “Omomys”’ belgicus, and simi- lar forms, that is, those Eocene tarsioids in which the molars are least specialized. The same may be said of the upper molars: those of Palenochtha show distinctions by which the genus may be recognized (such as the internal groove and more inclined protocone), but they very closely resemble the least aberrant Eocene tarsioids. Compari- son with ?Omomys vespertinus is especially suggestive of affinity. Most later genera differ in the manifestly progressive development of sec- ondary internal cusps. It is, incidentally, worthy of note that Palenochtha has no known character that would exclude it from ancestry to TYarsius, itself, although of course the absence of intermediate stages makes this ob- servation unworthy of being advanced except as an interesting but wholly untested possibility. The more advanced P, of Palaechthon does not call for detailed con- sideration. It is in line with progressive changes in many tarsioids and some other primates, although it shouid be noted that the devel- opment of P,in the Tetonius, Carpolestes, Apatemys, Plesiadapis, and some other groups lies along distinctly different lines. The peculiar molar structure suggested in Palaechthon and fully developed in Paromomys is more distinctive. The short quadrate trigonid and marginal paraconid approximated to the metaconid appear among tarsioids in only one or two later genera. Absarokius has very similar trigonids on M,3, but that of M, is more elongate, probably secondarily in connection with the shearing development of P,. Other American tarsioids are more distinctive. Among Kuropean forms, only Necrolemur and Microchoerus are similar, and they are 61 Commonly called the canine, but it seems to me more probable that it is an incisor in both cases. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 145 smore advanced in the reduction of the paraconid, but are much younger. The broad third lobe of M; and double hypoconulid also -appear in Necrolemur and Microchoerus, but not in Absarokius. Among American tarsioids only Washakius has a similar talonid on M3, and its trigonids are quite different. The characteristic posterointernal upper molar expansion and basin- ing of Paromomys are suggested in many later tarsioids, such as Absarokius, Hemiacodon, Shoshonius, Tetonius, and, in Europe, Nanno- pithex (most marked in Absarokius and Nannopithex), but in all these the structure is much less marked. In some cases (e. g., Tetonius) it is so slight that it is noted only by special search with Paromomys in mind, and in all the instances mentioned there is the characteristic distinction that a posterior cingulum passes internally beyond the jimits of the incipient (or vestigial) basin and tends to form a hypocone at its inner end. This could be a specialization from the Paromomys condition, but the difference is clear and there is no evidence of cer- tainly intermediate stages. In the European Necrolemur (and its highly modified ally, Microchoerus), however, the hypocone is on the rim of a basin much like that of Paromomys, and structural ancestry as regards this character is quite possible but hypothetical. In summary comparison with the acknowledged tarsioids, there are resemblances throughout and every separate structure of the Fort Union genera is approached in some later genus. The fundamental ‘similarity is most clear in Palenochtha, but even in this most general- ized type the anterior dentition is too specialized for ancestral rela- tionship to any known later genera but Tetonius, Necrolemur, Micro- choerus, and (still more doubtfully) Absarokius, and in these cases the minor morphological differences are also marked and annectant forms unknown. Paromomys and Palaechthon also resemble various later genera, but in most cases crossing specializations make any approach to direct phyletic connection impossible. Absarokius is, on the whole, the most similar American form, but in several respects it is apparently less specialized; for instance, in the simpler heel of M;, probably less enlarged incisor, and smaller protocone of P* (which may, however, be secondary), despite its younger age. The European genus Necrolemur ® compares more closely with Paromomys than does any known tarsioid to the extent that it exhibits all the principal specialized characters of Paromomys and that while it has numerous additional specialization of its own, no crossing special- zation is involved. Its dental formula is probably 7{35. Stehlin (1916) has placed this upper formula beyond any serious question. He gives the lower formula as 5;q3, but the evidence is very uncon- vineing. A priori it is highly improbable that an enlarged median 62 The following remarks apply equally to Microchoerus, except that the latter is much more highly special- ized in the dentition. If Necrolemur comes from Paromomys, then, ipso facto, Microchoerus does also but has eyolved more rapidly. 146 BULLETIN 169, UNITED STATES NATIONAL MUSEUM tooth should be a canine, that P, should be present in an upper Eocene tarsioid when it is absent, as far as surely known, in every other known tarsioid even in the Paleocene, or that an enlarged lower canine should occlude against an enlarged first upper incisor and anterior to a second upper incisor. The occlusion in itself is so suggestive of an incisor that only the strongest contrary evidence would warrant any other conclusion, and the other considerations seem to place this almost beyond doubt. It is true that in Tarsius the largest anterior mandi- bular tooth is the canine, but as Stehlin (1916) himself has shown the analogy with Necrolemur is very distant, and in Tarsius the lower canine occludes between I? and C as would be expected, not between I' and I*?. Furthermore, the actual formula in Tarsius is 33, as I believe it was also in Necrolemur.® If this formula be accepted for Necrolemur, its anterior dentition could be derived from that of Paromomys by further enlargement of the already enlarged incisor and great reduction of the canine, still of moderate size but not enlarged in Paromomys. The premolars of Necrolemur are broader, lower, and more proclivous than in Paro- momys, and there is a distinct metaconid on Py. The lower molar structure is closely similar throughout except for details in Necrolemur like the complete loss of separate paraconid on M,.3, which are the logical continuation of tendencies clearly present in Paromomys. In the upper jaw, the 3-rooted P® of Necrolemur is also progressive. P* and the molars are less transverse than in Paromomys, a feature of no clear significance. The protocone of P* may be smaller in the later genus; if so, it is the only character that suggests, and it does not prove, that Necrolemur could not be derived from Paromomys. Aside from their proportions, the upper molars of Necrolemur differ in having stronger conules (the metaconule double) and distinct hypocones, but the basic plan is remarkably similar, as already suggested. As far as the dentition goes, it must be concluded that Necrolemur and Paromomys are probably rather closely related, and the latter could be ancestral to the former. The conclusion is obviously un- proved and open to doubt. Corresponding with their great separation in space and in time, the genera do differ markedly and annectant types are unknown, but the fundamental similarity is striking. Trogolemur, Uintalestes, and Phenacolemur are all ineertae sedis, but all show some special resemblances to Paromomys and its Fort Union allies. Trogolemur has the same dental formula as Paromomys, tia3- The incisor is relatively larger, and the next tooth, presum- ably the canine, is very small. The premolars are more expanded 63 Although the case is less clear, by analogy it seems probable that the enlarged median mandibular teeth of Tetonius are also a pair of incisors, not canines as generally supposed following Matthew. Ishould write the Tetonius formula either 1.0.3.3 or 1.1.2.3, of which the second is perhaps slightly more probable. (Matthew wrote 0.1.2.3 but this is an evident lapsus, and he clearly meant to write 0.1.3.3-) FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 147 transversely but otherwise similar. P, has a metaconid. My,» are much like Paromomys. The third lobe on M; is narrower than in P. maturus but closely approached in P. depressidens. The cheek teeth throughout are very close to Palaechthon but lower, broader, and heavier. Palaechthon is an admirable structural ancestor for Trogolemur, although the time gap is too great for definite decision. Uintalestes is very poorly known but is evidently related to Tro- golemur from which it differs essentially only in the further dental reduction, having only seven teeth in the lower jaw, and the narrower heel of P,. The lower molars of Phenacolemur could readily be derived from the Paromomys type but are heavier and more quadrate. P, is also similar but is much enlarged and likewise heavier and more quadrate. The much heavier incisor and the complete loss of all teeth between it and P, sharply distinguish Phenacolemur, however, and the time gap is far too short for derivation from Paromomys. The molar resemblance may, therefore, be misleading. P* has a much stronger posteroexternal cusp than in Paromomys, and the internal groove is absent on the upper molars, but they have an equally and similarly expanded posterointernal basin and in general are as close to those of Paromomys as are the lower molars. Resemblances to the Carpolestes and the Plesiadapis phyla are dis- cussed in dealing with the contemporary members of the latter, Elphidotarsius and Pronothodectes, but the adaptively related group Apatemyidae has no known representative before the upper Paleocene. They may, however, be summarily dismissed as possible close relatives of the Fort Union forms, as none of their peculiar distinctions are foreshadowed in the latter. The apatemyids, as redefined by Jepsen (1934), have an enlarged incisor, larger than in Carpolestes or Plesia- dapis, which more nearly resemble Paromomys and its allies in this respect, and early lose all teeth between this and P3, at least two of which are retained in Paromomys and in the other two groups men- tioned. P; becomes 1-rooted and peculiarly bladed. P, is markedly reduced even in the upper Paleocene and becomes vestigial in later forms. The molars have a quadrate trigonid, as in many early primates, but, especially on M,, it is much more elongate anteropos- teriorly than in Paromomys and its allies. The upper teeth are equally divergent. Gidley (1923, pp. 3-4, 8-9) noted the resemblance of Paromomys and Palaechthon to the Notharctinae in the lengthened heel of M3, the trigonids consisting chiefly of protoconid and metaconid connected by a loph and with an anterior shelf, and the posterointernal expan- sion and basining of the upper molars. He added, however, that these are not exclusively notharctine characters and concluded that they did not indicate close affinity in this case. The resemblance is, 148 BULLETIN 169, UNITED STATES NATIONAL MUSEUM indeed, very close, and as far as the posterior teeth are concerned the differences involve only a few minor details of apparent generic or lesser value. The anterior teeth, however, are very different and are much more specialized in the earlier genus. Pelycodus has the formula. siag aS against yy, In Paromomys and Palaechthon, its incisors are small and its canine large. Furthermore, the Paromomys-like molar characters of Pelycodus are not seen in Adapis, but the Notharc- tinae, to which Pelycodus belongs, and the Adapinae show a funda- mental resemblance in skull and skeletal structure, which leads (Gregory, 1920, and elsewhere) to their association in one family. And this general structure is very unlike that of the supposedly tarsioid genera, such as Jetonvus or Necrolemur, with which there is. equal or greater reason for supposing Paromomys to be related. To propose as a tentative solution of this extraordinarily intricate problem that Paromomys is a derivative of a tarsioid-notharctine ancestry is not fully satisfactory, for a corollary would be that the Notharctinae were descended from a different tarsioid or prototarsioid. ancestry from the Adapinae, one with more Paromomys-like molars. and this is wholly unsatisfactory on the basis of the mutual relations of Notharctinae and Adapinae as inferred from their own much better known structures. It is much more probable either that the marked resemblance in the molars of Paromomys and Pelycodus is wholly convergent or that Paromomys is really a divergent offshoot. of the earliest notharctine ancestry toward which various tarsioids. have converged in one way or another. The second alternative would imply extreme antiquity of the Pelycodus pattern and the very remote separation of Notharctinae and Adapinae, to a degree that seems: improbable. The very tarsioid, and not particularly Pelycodus-like, pattern of Palenochtha, which nevertheless seems almost surely to be: a fairly close relative of Paromomys and Palaechthon, and the many distinctly tarsioid characters of the latter genera, however electic in their combination, also suggest that the former alternative is more probable, as Gidley concluded. I must confess, however, that I see no way of forming a really strong and reasonable opinion on this. problem from the present evidence. Genus PAROMOMYS Gidley Paromomys GIDLEY, 1923, p. 3. Type.—P. maturus Gidley, 1923. Distribution.—Middle Paleocene, Fort Union, Mont. Diagnosis—Dental formula probably 33. Lower incisor enlarged, root extending beneath P,. Canine normal or slightly reduced. P, present, 2-rooted. P, not enlarged, paraconid and metaconid very rudimentary or absent, trigonid apex slightly higher than M;, heel low, bicuspid. Molars with short trigonids, with closed, small,, FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 149: quadrangular, transverse trigonid basins, paraconids small -and closely approximated to metaconids, especially on M>-3; cusps.sub- marginal; no metastylid. Talonid of M; greatly enlarged, with strong third lobe with (at least) two distinct, transversely paired cusps. P? 2-rooted. P* 3-rooted, with strong protocone, no distinct meta- cone. Upper molars primitively tritubercular, without mesostyle, protostyle, or hypocone, but with a ridge from the protocone swinging around the posterointernal corner, which is much expanded and basined. Internal bases generally bilobed. The morphology of the genus is described under its type species, and the distinctive characters of the second species mentioned under it. PAROMOMYS MATURUS Gidley Figures 30, 31; Puats 7, Figures 2, 2a, 3, 3a; Puats 8, Ficurss 2, 2a, 3, 3a: Paromomys maturus GipLEy, 1923, p. 3. Type.—U.S.N.M. no. 9473, right lower jaw with P,-M; and anterior alveoli. Collected by A. C. Silberling. Horizon and locality —Al\l known specimens from Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis —Trigonids M,-, notably narrower than talonids. P* strongly transverse. Internal bases of M!~ strongly bilobed. Meas-- urements and derived statistical data given below. Dental formula 't2° (but see below). All teeth closely placed, without diastema. Discussion.—There are six specimens in which the anterior alveoli are all shown, although in none are their rims unbroken. One of these, as noted by Gidley (1923, p. 9—Gidley mentions two, but the other is not of this species), has a minute pit external to and between the incisor and canine, which might be an alveolus for a vestigial incisor. It could, however, be a mere break, and as none of the other five specimens shows it this is a more probable explanation. I, was thus probably absent, and if ever present was vestigial and oftener lacking. The incisor and canine are unknown except by their alveoli. The incisor was large, its root slightly compressed laterally, and was semi- procumbent, its root extending to beneath the posterior end of P» or anterior end of P;. The canine was considerably smaller, its nearly circular root with about half the (maximum) diameter of J,. It is less procumbent, and its root is shorter than that of I,, the root of which passes beneath it. The one canine root occupies about the same- space as the two of P». P, has two separated, divergent roots. The crown is high, slender, very slightly procumbent, andsomewhatrecurved atthe tip. Theouter- face is convex, the inner excavated anteriorly and posteriorly, adjacent to curving, vertical sharp anterointernal and posteroexternal crests... 150 BULLETIN 169, UNITED STATES NATIONAL MUSEUM On one specimen (9479) there is a very minute anterior cuspule, high on the crown, which is only very vaguely suggested on a second speci- men surely of the same species (9676). There is a low and very small heel, with a minute cusp at the posteroexternal corner of the = FIGURE 30.—Paromomys maturus Gidley, U.S.N.M. no. 9473, right lower jaw: a, Crown view; 6, internal view. Four times natural size. (After Gidley, 1923, fig. 1.) FIGURE 31.—Paromomys maturus Gidley, U.S.N.M. no. 9540, left upper jaw: a, Crown view; 0, external view. Four times natural size. (After Gidley, 1923, fig. 2.) tooth, at the base of the posterior crest, from which a ridge passes downward and internally to the posterointernal corner. P,; has about the same length and height as P, but is considerably wider and more robust. The anterior crest is median near the apex and turns inward below this, an accentuation of the slighter curve of FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 151 this crest on P;. The posterior crest is less curved and is only slightly external to the midline, the tooth being much more swollen external to it than is P;. The heel is more definite, much wider, the cusp less external, and the ridge less sloping. P, is much longer and wider and somewhat higher than P;, shorter, slightly narrower and somewhat higher than M,. Aside from its greater size, it differs from P; chiefly in the much stronger heel. A _ small cusp appears at the posterointernal angle, and a ridge running anteriorly from this tends to close a small basin. A minute cuspule may appear about halfway up the crown on the inside of the anterior edge, but this rudiment is often lacking even on unworn teeth. Simi- larly, a very vague rudiment of a metaconid appears on the most progressive variants (e. g., 9545) but is oftener absent. The molars have small trigonids and large basined heels. On M, the trigonid is well elevated and is directed somewhat forward, while on M,_; it is progressively lower. On M,-, the talonid is considerably wider than the trigonid, the inner face of the tooth base being along a straight anteroposterior line and the outer face strongly oblique. On M; the trigonid and anterior half of the talonid are of about equal width. The enamel is nearly smooth, but on completely unworn teeth the basin is somewhat wrinkled. Variable, crenulated external cingula are developed on all the molars except on the third lobe of M;. There are no internal cingula. The paraconid is present on all the molars and is anterior and slightly external to the metaconid. On M, it is definite and well separated from the metaconid, although small. On M,-,; it is much closer to the metaconid and is almost fused with the latter, disappearing with slight wear. On M, the metaconid is about equal to the proto- conid, and on M,-; it is higher. On all, the metaconid is internal and slightly posterior to the protoconid, and the two are connected by a notched crest. Another, less prominent crest runs forward and slight- ly inward from the tip of the protoconid to the anteroexternal angle of the tooth, then internally along the anterior rim to the paraconid, enclosing a short, transverse, very shallow and small trigonid basin. M,-2 have typical hypoconid and entoconid of about equal height. The sharp basin rim is vaguely expanded in the hypoconulid region, but no definite apex is here formed. There are no metastylids. The heel of M3; is very elongate, with two definite lobes each pri- marily with two large cusps, one external and one internal. The pos- terior, or hypoconulid, lobe may be further complicated by the incipi- ent fission of one or both of its cusps, and adventitious cuspules may even appear in the basin, the exact structure of this part being highly variable, although its basic features, the extension of the basin into a third lobe and the strongly double hypoconulid, are constant. 119212—37—11 152 BULLETIN 169, UNITED STATES NATIONAL MUSEUM The symphysis is short, unfused, and relatively deep. There is a larger mental foramen beneath P, or the posterior part of the canine and another, smaller, beneath P,. The dental foramen was far back of the molars and slightly below the alveolar level. The angular process is not completely preserved in any case but was directed decidedly downward, as well as backward, and evidently was strong and more or less styliform. Condyle and coronoid are not preserved. The upper canine is represented only by part of its alveolus in one specimen (9540). This suggests that it was strongly reduced, the portion of alveolus preserved indicating a root not larger than the pos- terior root of P?. P?-* are known only from alveoli. Each had a small anterior and large posterior root. On P? the disparity is greater, and the posterior root more transverse, but even it apparently does not have a third root, and the inner heel, or protocone, must have been small. P*is a large, transverse tooth with three separate roots. The high outer cusp, paracone or amphicone, is vaguely triangular and is single, only very slight inner and outer vertical depressions suggesting the incipient appearance of a metacone on its posterior slope. There is a small, distinct parastyle and a much less distinct metastyle higher on the crown than the parastyle. The posterior half of the outer face has a narrow, sharp basal cingulum. The protocone is large and definite but lower than the amphicone, and its apex is anterior to a median transverse line across the tooth. From it a small sharp, cingular crest runs to the parastyle. Another crest falls away directly posteriorly from its apex to the expanded posterointernal angle of the crown, where it turns nearly at right angles and becomes a well- developed but simple posterior cingulum. A minute, isolated cuspule appears in the position of a metaconule. M! and M? are almost identical in structure, differing only in outline and proportions. Paracone and metacone are strong, distinct, and nearly equal, the paracone very slightly larger. There is a strong external cingulum, rising at the anteroexternal corner without definitely forming a parastyle cusp. The metastyle is likewise small and vague but is more nearly cuspidate. There is no trace of a mesostyle. The inner face of the tooth is flattened and has a distinct, median, vertical groove that divides it into two basal lobes, the posterior lobe being on M! slightly and on M? distinctly larger. These lobes, however, do not correspond to distinct cusps. There is only one cusp, the protocone, which is on the posterior lobe very near the distal end of the groove. Although larger, the protocone has the same structure as on P* with the addition of a ridge to the metaconule, departing not from the apex but from the posteroexternal slope. The anteroexternal ridge, as on P‘, runs to the anteroexternal corner, or to the parastyle (here less distinct) but at its midpoint here has a cuspule, a protoconule, the FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 153 posteroexternal base of which also meets a small ridge descending the inner face of the paracone. This protoconule is very small, and the metaconule is still smaller and indistinct. The posterior crest from the protocone apex forms a sharp loop around the expanded posteroin- ternal corner of the crown, thus making the teeth distinctly quadrate although no hypocone is present. M$ is shorter than M'~?, with the whole posterior half, most notice- ably the metacone and posterointernal loop, much reduced. The internal base has only one lobe. On one specimen (9542) there is no groove, and on the other (9540) it is very slight and does not reach the base. TABLE 80.—Numerical data on lower teeth of Paromomys maturus Variate N R M o V Ibs eee | 18 2.3—2.8 |2.556+0. 026 |0. 112+0. 019 4.4+0. 7 Wily eese e 18 1.6—1.9 |1. 778+0. 022 |0. 092+0. 015 5. 240.9 TDN bree ae ee a 21 2.8—3.2 /3.005+0. 020 0. 090+0. 014 3.0+0. 5 \T 0 7 ee Pall 2.0—2.3 /2. 288+0. 018 |0. 084+0. 013 3. 8+0. 6 TEA (heelys a 20 2.9—3.2 |3.040+0. 024 |0. 108+0. 017 3.5+0. 6 WWEMiG ee. 20 2.2—2.5 |2.355+0. 017 |0. 07440. 012 3.1+0.5 Migs a leer 14 3.3—3.8 |3.59 £0.03 |0. 130+0. 025 3.640. 7 WWalVigean eS 175 1.8—2.1 |1.994+0. 006 |0. 025+0. 004 1.3+0. 2 WPM ess 13 0. 77—0. 89 |0. 88740. 009 |0. 03440. 007 LM,:WM,2_--| 19 1. 25—1. 41 /1. 298+0. 011 |0. 046-0. 007 EM; LM,____| 10 J. 16—1. 31 |1. 215+0. 014 |0. 046+0. 010 TaBLe 31.—Measurements of individual specimens of Paromomys maturus TYPE LOWER Jaw! Ps Mi M2 M; | M LP, LM2 LMs: “3 | IM: | WM: | LMs L W L W L W L W Mm Mm Mm Mm Mm Mm Mm Mm 7Ast%) ies ond 222 | oul 2.3 3.6 1.9 9.0 0. 81 1.35 1.16 CHARACTERISTIC UPPER DENTITIONS P4 M! M? M3 U.S.N.M.no. peer vont (tly SE sole heer ill cont gyi fy Mag LM! LM} LM? L Ww L W L W L W Mm | Mm} Mm | Mm | Mm | Mm | Mm | Mm pai | ee enna ae Sea 2.9 3.4 2.9 4.0 2.9 4,1 2.3 4.1 8.3 1.00 1.41 0. 79 OAD Nae eer 85. *h 225) 3.0 2.6 3.8 2.6 4.0 2.0 3.6 7.4 0. 96 1, 54 0.77 1 These measurements have been taken, and ratios calculated, on 30 different specimens, but these are most usefully summed up by the statistical data derived from them and wherever possible in this memoir I do not publish the long tables of raw measurements. Data for Mii-3 have not been calculated. Although at first sight this would appear to be the best single size measurement, in fact it depends to so great a degree on crushing and other extraneous factors as to be highly inaccurate and hence of little value. 154 BULLETIN 169, UNITED STATES NATIONAL MUSEUM The infraorbital foramen is high and narrow and lies immediately anterior to the anterior root of P*. Very little of the orbital rim is preserved, and I see no basis for supposing it larger or smaller than in any possibly related group. It seems probable that it extended little, if any, farther forward than P*. The principal numerical data on lower teeth of this species are given in table 30 (see also fig. 3). The highly homogeneous character of the sample, and by inference the only slightly variable nature of the species, is very striking. All the coefficients of variation are remarkably low and even the highest, 5.16 £0.86, is very commonly exceeded in races that are pure in the strictest sense. There are too few upper jaws to calculate derived statistical data. PAROMOMYS DEPRESSIDENS Gidley Figure 32; PLATE 9, FIGURE 7 Paromomys depressidens GipLEY, 1923, p. 4.% Type.-—U.S.N.M. no. 9546, part of right upper jaw with P*-M?. Collected by A. C. Silberling. Horizon and locality—AlL known specimens from Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. FIGURE 32.—Paromomys depressidens Gidley, U.S.N.M. no. 9485, left lower jaw, crown view. Four times natural size. (After Gidley, 1923, fig. 3.) Diagnosis —Trigonids of M,_2 nearly as wide as talonids. P* less transverse than in P. maturus. Bases of M'* less strongly bilobed. Size notably smaller; length M, (mean of three specimens), negative deviation from mean in P. maturus about eight times standard devia- tion of latter. See also measurements in table 33. P, larger relative to M,; ratio LP,:LM, (one specimen), positive deviation from mean in P. maturus over three times standard deviation of latter. M, wider relative to its length; ratio LM,:WM, (mean of three specimens), negative deviation from mean in P. maturus nearly three times stand- ard deviation of latter. Discussion.—l,—-P3; are known only from their alveoli, which are developed about as in Paromomys maturus. Of the two specimens showing these alveoli, one (9416) has a possible alveolus for a vestigial I,, and the other (9482) does not, so that, as in P. maturus, this tooth was either absent or inconstant. P, is slenderer and somewhat less 64 In referring to Gidley’s description, note that the text has been transposed. The text in Gidley, 1923, from p. 5, line 12, beginning ‘‘Several upper-jaw portions . . .’’ to p. 6, line 3, ending ‘‘. . . above the junc- tion of P3 and P4”’ is made part of the description of P. depressidens but in fact refers to P. maturus and evi- dently was meant to follow p. 4, line 12, of text, after ‘‘. . . as in Notharctus nunienus (Cope).” FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 155 progressive than in P. maturus, being to some extent intermediate in structure between P; and P, of that species. The supposed depression of the molar trigonids, stressed by Gidley, involves slight differences that defy accurate measurement. To my eye they do not appear at all less elevated than in P. maturus. The trigonids of M>_; are rela- tively slightly shorter, and the paraconids may be still more nearly connate with the metaconids. More definite is the fact that on M, the trigonid is nearly as wide as the talonid, while in P. maturus it is definitely narrower. Perhaps in keeping with the smaller size, the heel of M; is rather simple, and although it has the basic structure of P. maturus in some variants the third lobe is less wide and its two main cusps less distinct. The available upper teeth of this species are all deeply worn, and I believe that the characters given by Gidley, ‘‘cusps and lophs depressed and basins shallow; protoconules present but less well defined than P. maturus; metaconules absent’’, are all due to this wear, or at least that the wear makes it impossible to know whether these are true morphological characters or not. P* is much less transverse than in P. maturus, and the inner sides of the bases of M'~? are less strongly bilobed. Otherwise the structure seems to be very similar, as far as it can be surely determined. The species is decisively distinct from the genotype, and there can be no question as to its validity in spite of the fact that P. maturus and P. depressidens were absolutely contemporaneous and both are known only from the same very limited locality. These facts and the very TABLE 32.— Measurements of known specimens of Paromomys depressidens LOWER JAws LP, LM2 LM; WEE NeMomoasiriny rea lah alee ii cade cee Pe Pee aay A SLs Mm | Mm | Mm | Mm | Mm | Mm | Mm | Mm O48 petrvees ge beser | ie | bees 2.0} 1.6) 2.0) 1.8) 26) 1.5 (EAT 2 oes Sc 1G 1.30 OC ieee OE OU Ma ii baal ate -29h bee a es he J a 2.8 | 1.6 6.9 0. 95 O28 Zeca eS dee LS Shee see p50 | Nios) ses je! [game 5s | hae Li) a ees EIS Te 1, 22 1. 23 OF1G1S 2 ees 22 eee 1 tah |e) 9.77 eae | Se eS Ss PECIH a Sth ee ened Pe ee eee Sed eee ere Ts 1, 21 UPPER JAWS 4 A 2 3 P M M M a | bet | wM?| ums ESS NGWE. 0. 3a rl ate aga Ie Paull Gail REL Eee aL PE ATi IGeol oop Mm | Mm|Mm|Mm|Mm | Mm | Mm | Mm eee 159) 2305) 250) 256) 220 ees. 0 ERS ey / 5.7 0.95 1.50 0.75 156 BULLETIN 169, UNITED STATES NATIONAL MUSEUM marked specific distinction suggest that complete, unworn dentitions might prove generic distinction, but the material actually in hand does not warrant such a conclusion. As the samples are small, measurements of all known specimens are given in table 32 (see also fig. 3). Genus and Species Undetermined Cf. PAROMOMYS In 1932 Silberling and I found at Loc. 13 a single left M', apparently representing an otherwise unknown primate. It resembles Plesiadapis anceps but not very exactly and is still less like other species of that genus. It is probably too small to belong to Plesiadapis rex, which occurs at the same locality. It resembles Paromomys maturus more closely than any other species with which comparison has been made, but it is more transverse, has the inner face even longer and more sloping, and has the anterointernal, not posterointernal, basal part definitely more projecting. It also resembles the most primitive species of Pelycodus but could not belong in that genus. Such an isolated tooth is inadequate for generic, or even for certain family identification, but its presence seems worth recording. Genus PALAECHTHON Gidiey Palaechthon Gipury, 1923, p. 6. Type.—P. alticuspis Gidley. Distribution.—Middle Paleocene, Fort Union, Montana. Diagnosis —Dental formula probably ~->3. Lower incisor and canine about as in Paromomys, or possibly incisor slightly larger and canine slightly smaller. Roots of P, less divergent or incompletely divided. P, larger relative to M,, talonid as in Paromomys, but trigonid with distinct, subequal paraconid and metaconid. Molar trigonids very similar to Paromomys but on M,_3 paraconid slightly more distinct, lower on the crown, and less marginal. M; with double hypoconulid, but third lobe less strong than in Paromomys. Trigo- nids more elevated. Upper molars somewhat more transverse than in Paromomys maturus, internal bases less distinctly bilobed, and posterointernal expansion jess marked. PALAECHTHON ALTICUSPIS Gidley Puate 7, Figure 1; Puate 9, Ficurss 5, 6 Palaechthon alticuspis GipLuy, 1928, p. 6. Type.—U.S.N.M. no. 9532, right lower jaw with P.-M,. Collected by A. C. Silberling. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 157 Horizon and locality—All known specimens from Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis.—Sole known species of genus as redefined. See mor- phological and numerical data (table 33). Discussion.—From their alveoli, in two specimens (9532 and 9534), J, and C are much as in Paromomys maturus, but the incisor is perhaps relatively a little larger and the canine still smaller, its alveolus occupying distinctly less space than the alveoli of P,. There is no evidence of a second incisor. The alveoli of P, are confluent, and in one specimen the roots are fused, in another barely separate. P, and P; are similar to each other and to those of P. maturus. Pz» is slenderer and slightly higher than P; and P; has the incipient anterior cuspule more nearly distinct than Po». P, is much more progressive than in Paromomys, having small but distinct subequal paraconid and metaconid in all cases, whereas in Paromomys these cusps are either barely incipient or entirely absent on P;. The talonid and general structure are, however, as in Paro- momys. M, almost exactly resembles that of P. maturus. M3_; are also closely similar but have the paraconid more distinct than is usual in Paromomys, lower on the crown and also a little more external, or less marginal. On both M, and M, the trigonid is nearly as wide as the talonid. On M,_; the trigonid is more elevated than in Paromomys, and the external cingulum is weak or absent on the talonid. The third lobe of M3 is much less developed than in Paromomys maturus, but the hypoconulid is bifid or, in one case (9430) approximately trifid. The mandible is also much as in Paromomys. In one specimen (9450) most of the posterior part is present, although the ends of the three processes are broken. The inner face is nearly plane. The coronoid is broad and apparently rose little above the articular process. Its anterior border is nearly straight and at right angle to the alveolar border. The condyle is far above the molar level. The large dental foramen is far posterior to the teeth, beneath the corono-condylar notch, and is above the alveolar level. The angle is long, slender, styliform, and thickened, and extends backward and slightly down- ward. Its tip was posterior to a vertical from the condyle. P* is not known, but it has three roots and was nearly as wide as M?. The upper molars closely resemble those of Paromomys maturus except in being somewhat more transverse, with slenderer sharply pointed cusps, and in the much less marked posterointernal expansion and basining. There is an internal vertical groove, but on M! the base is not bilobed and on M? this is barely indicated. Several of these points are resemblances to Paromomys depressidens, and it has already been noted that the teeth on which that species is based are much worn and of doubtful detail. As the two species are of about 158 the same size, the status of these upper jaws is dubious, although the lower jaws are quite distinct, and it is unfortunate that one type is an upper jaw and the other a lower. The upper jaws are, however, probably distinct and correctly associated with the lower jaws as determined by Gidley. Those placed in Palaechthon alticuspis are very slightly smaller, molars definitely more transverse, and the pos- terointernal basin probably less developed. It seems justifiable, in the absence of definite evidence to the contrary, to accept the identi- fications established and thus avoid changing the nomenclature, con- sidering the separation of the species and genera as validated by the lower jaws and, for the present, overlooking the fact that one type is, in fact, an upper jaw and of doubtful status. BULLETIN 169, UNITED STATES NATIONAL MUSEUM TABLE 33.—Numerical data on Palaechthon alticuspis Variate N R M o Vv 19 Fre ener ets 10 1.9 -2.1 2. 030+0. 025 | 0. O78+0. 017 3.8+0. 9 Wa Vie eras See 9 ER als 76 1. 567+0. 022 | 0. 066+0. 016 4.2+1.0 1 DAI) Cm ck ek a 12 1.9 —2. 2 2. 067 +0. 024 | 0. 085+0. 017 4.1+0.8 WEMGe ee Sees: 11 1.5 -1.8 1. 682+0. 025 | 0. 084+0. 018 5. Ose Tl LM.2:WM2_-_-!| 11 1. 17-1. 40 | 1. 28140. 021 | 0. 070+0. 015 TABLE 34.—IJndividual measurements of Palaechthon alticuspis TYPE AND CHARACTERISTIC LOWER JAWS Py Mi M2 M3 | LP, LM, LM; U.S:NsVE n0s46 | TERE a GOR CLG eo LL sa Ses M1-3 LM: | WM: | LM: L Ww L Ww L Ww L Ww Mm | Mm | Mm}Mm|Mm | Mm | Mm | Mm O58 2E SC a a rete 1.9 2 1.9 1.5 2.0 We Wiecsn See ee ee es 1.00 1.18 430s te Be at Rr tae 2.1 1.2 2u1 1.6 2:1 1 7. 2.4 1.3 6.5 1.00 1. 24 1.14 O48 se. Se eee 2.0 12 2.1 1.5 2.1 1.6 2.3 1.4 6:3 0.95 1.31 1.09 UPPER JAWS (ONLY TWO ARE KNOWN) M! M? M3 U.S.N.M. no. M!-3 “ ae L Ww L WwW L WwW Min Mm Mm Mim Mm Mm Q550 Fee Se vee te 1.9 3.0 Ny / 3.0 1.3 2.5 5.3 1.76 0. 76 OOD IEE Sear eee See 1.9 3.0 1.9 5 976 fe aa le F< Mi ante 1. 68 Genus PALENOCHTHA Simpson a lenochtha Simpson, 1935d, p. 281. Type.—Palaechthon minor Gidley. Distribution.—Middle Paleocene, Fort Union, Montana. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 159 Diagnosis.—Dental formula probably 3. Anterior lower denti- tion shorter than in Paromomys or Palaechthon and apparently with one tooth absent, probably P.. Py, of about the same length relative to M, as in Palaechthon alticuspis but relatively higher, with no sign of the metaconid and only vague rudiment of the paraconid. My,-», similar to those of Palaechthon, but M; with smaller third lobe and single hypoconulid. Upper molars similar to those of Paromomys and Palaechthon but very slender, transverse, and more triangular. Posterointernal expansion weak. Inner base not bilobed. M# shorter relative to M?. Discussion.—In describing Palaechthon minor, Gidley (1923, p. 8) said: ‘“‘Most of the differences noted above suggest for the species just described a slightly different line, or direction, of development than is indicated in P. alticuspis. It is possible, therefore, that more complete materials may prove that these two species do not form a natural generic group.”” Although more complete materials are still lacking, I do not see how this species can be placed in Palaechthon. The further reduction of the anterior teeth, the absence of a metaconid and of a distinct paraconid on P, (not explicitly mentioned by Gidley), and the single hypoconulid on M; are just such differences as are used to distinguish genera among all early primates. While it is true, as Gidley notes, that the upper molars differ less from Palaechthon alticuspis than do the lowers, still the differences are rather more marked than are those between the latter species and Paromomys depressidens. Unfortunately, the anterior upper teeth, which often show more marked generic characters, are unknown, but I think there can be no doubt that the genus is distinct. PALENOCHTHA MINOR (Gidley) Figure 33; Puate 10, Ficure 1 Palaechthon minor Gip.Ey, 1923, p. 7. Type.— U.S.N.M. no. 9639, right lower jaw with P,-M; and anterior alveoli. Collected by A. C. Silberling. Horizon and locality—All known specimens from Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis—Sole known species of the genus as described above. See description below and measurements in table 35. Discussion.—Teeth anterior to P, are represented only by alveoli in the type and in no. 9631. Both show an alveolus for a large, some- what compressed, procumbent incisor, followed by a smaller, more erect alveolus, evidently for a canine. Between this and P, there appears to be only a small double alveolus, or two very small alveoli confluent at their mouths. It is highly probable that this lodged one tooth, P;, and in this event P,-, must have been missing. There 160 BULLETIN 169, UNITED STATES NATIONAL MUSEUM appears, however, to be no diastema, and the antemolar region of the jaw is relatively shorter than in the other genera here considered, what- ever may have been the exact number and homologies of the teeth. P, is relatively higher, slenderer, and simpler than in Palaechthon. In the two available specimens there is no clear trace of a metaconid, and the paraconid is represented only by a very rudimentary and scarcely visible angulation of the anterior edge. The heel is short but is relatively broad and has a rudimentary basin and two very small and poorly differentiated posterior cuspules. M,_» are almost identical in structure with those of Palaechthon. The trigonid of Ms; is also closely similar, but the talonid is different. It is reduced, more pointed posteriorly, with the third lobe de- Ficure33.—Palenochtha cidedly narrower and less clearly differentiated, and minor(Gidley),U.SiN.Ma the hypoconulid apparently single. no. 9590, left upper mo- ae lars, crown view. Four In the upper dentition, only M!~* are now known. rMidley, 102, het) Lhese are basically similar to those of Palaechthon but give quite a different superficial impression by reason of their more delicate structure, more transverse and triangular outline, and the accentuated forward twist of the protocone, present in the other primate genera but here most strongly developed. Para- style and metastyle are distinct and subequal, as are protoconule and metaconule. The inner face of the protocone is flattened and bears a faint vertical depression, but the base is not at all bilobed. The posterointernal basin or expansion is only faintly indicated, less de- veloped than in the other genera. M® is short, rather strongly tri- angular, and developed analogously to that of Palaechthon. TABLE 35.—Individual measurements of Palenochtha minor TYPE AND PRINCIPAL REFERRED LOWER JAWS P4 Mi M2 M3 U.S.N.M. no. |—————_|—______|—________|______| yw, Es lh reas LM: | WM: | LM; igh Bae ee eae LSP OLS ysl aT Desa aL 15 O29 4.3 1.00 1.08 1,15 GGdIoL 251 Ace Lie DEMOS Sle 2 al eel OP eal See etOni se soen |e I2e22)| € eee ee 0. 92 1.30 9636-52 o aoe net See aeons WSuliq LaOuiy 283 | ol Aad at 059 Cea a eee 1.18 1. 08 UPPER JAWS (ALL KNOWN) Mi! M? M3 WM?| LM U.S.N.M. no. SS | ey I te LM? } LM? 9590) seb piss 2. fe a see 1.3 2.0 1.2 2.0 0.8 1.6 3.8 1. 67 0. 67 FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 161 Family CARPOLESTIDAE Simpson, 1935 As was pointed out in describing Carpolestes (Simpson, 1928, p. 10), Hlphidotarsius supplies a good morphological ancestry for the very peculiar and aberrant genera Carpodaptes Granger and Matthew of the Tiffany and Carpolestes Simpson from Tiffany or slightly later equivalents in the Fort Union. The present opportunity to compare the genotypes of the three genera at first hand fully confirms this and leaves no doubt that they are closely related. Elphidotarsius and Carpolestes show the same highly characteristic basic structure throughout P-M;. Carpodaptes has P, more enlarged than in Elphidotarsius, its apical cuspules all in a straight line and one more in number. On M, the trigonid is still more elongate, and the paraconid almost directly anterior to the protoconid. Other struc- tural distinctions are very slight and unimportant. In Carpolestes P, is still larger, its cuspules increased to seven or eight, its heel elevated to the trigonid level of M,;. On M, the paraconid and protoconid are exactly in the same longitudinal line and continue without a break the cuspule series of P,. The structural sequence Elphidotarsius— Carpodaptes—Carpolestes is almost perfect (also in the size of the known species) and may be a direct phylogeny, although the possible age difference between the last two genera seems too small to permit such a marked structural advance in a direct descendant, and it is more likely that some collateral evolution is involved. Upper teeth are as yet known only in the genera Carpolestes (Car- polestes dubius Jepsen; see Jepsen, 1930a) and Carpodaptes. The molars are of primitive tritubercular type, with distinct hypocone, more or less closely paralleled in some primitive Eocene primates (e. g., Omomys, Caenopithecus, Pseudoloris, and others). P*~*, how- ever, are very extraordinary and unlike anything known in any other primate or indeed any placental mammal, to such a degree that when the first isolated example of one of these teeth was found J hesitantly referred it to the Multituberculata (‘‘Litotherium” © Simpson, 1929, p. 9), and this remarkably bad guess was only corrected when Jepsen found associated premolars and molars. Like the last premolar of Piilodus, both P? and P* have three longitudinal rows of cusps. These premolars are much unlike those of any tarsioid, as is Ps, but it may confidently be predicted that P** of Elphidotarsius, when found, will distinctly approach the normal tarsioid type, as does P, of that genus. P** of Carpodaptes are indeed closer to Carpolestes, but they show some approach to more normal structure. 65 A strict synonym of Carpolestes Simpson, 1928. The retention of Carpolesies as the definitive name is not only preferable, as Jepsen suggests, but also the only possible course in accordance with the rules of nomenclature, 162 BULLETIN 169, UNITED STATES NATIONAL MUSEUM The dental formula is also most completely known in Carpolestes dubius, in which it is ==, as in Paromomys. The incisor is en- larged but does not extend beneath Py, also as in Paromomys, but the canine is more reduced. In both Carpolestes and Carpodaptes P.-; are reduced to 1-rooted vestiges with buttonlike crowns. Noting these divergent specializations, but also the strong hint of tarsioidlike upper molars, in the most specialized genus of the phylum in characters not known in Elphidotarsius, we may expect the latter to cast more light on affinities as far as its more scanty remains go. P, in this genus could easily be derived from one like that of, say, Palaechthon, but it is already too specialized, and its structural ancestor must have been too generalized, to cast any real light on affinities. M, also shows what may be taken as the beginning of a narrowly phyletic specialization in its elongate trigonid. Otherwise it is much like that of many tarsioids but of too generalized a heritage to give decisive evidence. M,_; are almost exactly like those of Pronothodectes, so much so that were this form known from those teeth alone it would have to be defined as a species of Pronothodectes. They also resemble to a marked degree those of Eocene primates of other groups, such as ?Omomys vespertinus and, in less degree, Pelycodus. As far as I know they do not so closely resemble any genus not now considered as primate. The characters that are distinctive from the most generalized tuberculosectorial pattern and that are not clearly habitus characters or otherwise neomorphs of this very restricted phylum all appear to me to be definitely primate. To this extent I cannot agree with Jepsen (1930a, p. 523—he was, however, dealing only with the much more aberrant terminal genus Carpolestes and had not recognized the relationship to Elphidotarsius) that “it is possible to select suites of characters which, taken by themselves, would place Carpolestes in any one of several orders.’’ I do, of course, recognize that a really defini- tive determination of affinities is in such cases practically impossible from teeth alone, but since teeth are, in fact, all we have I see no useful alternative to classifying them at least tentatively as belonging to the group they most resemble, that is, to the Primates. The family in which these related, aberrant animals are placed was defined and discussed in revising the Tiffany fauna (Simpson, 1935c). Genus ELPHIDOTARSIUS Gidley Eiphidotarsius GipLEy, 1923, p. 10. Type.—E. florencae Gidley. Distribution.—Middle Paleocene, Fort Union, Montana. Diagnosis —Three lower molars and at least one premolar, dental formula otherwise unknown. P, enlarged, equaling or exceeding M, in 66 Jepsen writes 1.0.4.3. It is, of course, impossible to say which is correct, but the form I give seems to me slightly more probable, and it facilitates comparison by being consistent with the other formulae here used. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 163 every dimension. Apex with four cuspules, second highest, first three anteroposterior, and fourth slightly internal. Talonid very short, with one cusp. M, with elongate trigonid, paraconid far from meta- conid, no trigonid basin. M,_; with closed trigonid basins and para- conid small, near metaconid, but distinct. Talonid of M, with well- differentiated third lobe, posterointernal rim elevated and vaguely including two or more apices. Protoconid reduced on M2_3. Trigonid cusps all well in from the margin. ELPHIDOTARSIUS FLORENCAE Gidley Puate 10, FiaursEs 2, 2a Elphidotarsius florencae GIDLEY, 1923, p. 10. Type.—U.S.N.M. no. 9411, left lower jaw with P,-M;. Collected by Dr. J. W. Gidley. Horizon and locality — Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis —Sole known species of genus as diagnosed above. See also description and measurements below. Discussion.—P, is a very peculiar tooth, longer, wider, and higher than M, although not greatly exceeding the latter in any dimension. Both sides are almost smoothly convex, but the apex is formed by four cuspules, or serrations, of which the second is highest although the third is slightly larger. The first three are in a straight anteropos- terior line, the fourth slightly more internal. The very short heel has one cusp, from which a small sharp crest descends vertically along the posteroexternal edge of the tooth and turns into the external cingulum. On M, the distinct but small paraconid and the larger metaconid are widely separated and there is no trigonid basin. The metaconid is posterointernal to the protoconid. In the talonid the hypoconulid cannot be distinguished and the basin is not completely closed, as there is a deep narrow notch between the entoconid and the trigonid. On M, the trigonid is much shorter and wider, the metaconid less posterior, the paraconid close to the metaconid although still distinct, asmall trigonid basin present. The talonid is like that of M, but larger. The trigonid of M; is like that of M,. The heel is modified by the addition of a well-differentiated third lobe, the elevated posterior and posterointernal rim of which shows some tendency to split into two cusps, although these are not distinctly developed. On M, the proto- conid and metaconid are of nearly equal height. On M, the protoconid is slightly and on M; decidedly lower than the metaconid To an even greater degree than is common in primitive mammals the alveolar border slopes outward, so that the external faces of all the cheek teeth are much higher than the internal. Presumably an enlarged incisor was present, but its root did not extend under P,, and the jaw is broken off at this point. 164 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Measurements of the only known specimen are as follows: Length P,, 1.5; width P,, 1.2; length M,, 1.2; width M,, 1.2; length Mb, 1.2; width Mo, 1.3; length M3, 1.7; width M3, 1.0; M,_3, 4.4; ratio length P, : length M,, 1.25; ratio length M.: width Mb, 0.95; ratio length M; : length Ma, 1.42. Family PLESIADAPIDAE Trouessart, 1897 Although quite distinctive and manifestly in the plesiadapid line, Pronothodectes shows a definite resemblance to the other primate genera in this fauna. Elphidotarsius represents the beginning of a divergent line, especially in the first stages of specialization of P,, but there is a remarkably detailed resemblance in the molar structure. The paraconids are more distinct in Elphidotarsius and the trigonid of M, more expanded anteroposteriorly. The talonid cusps of M,_» are less distinct. Except for the here very slight difference in the para- conid, the highly characteristic M; is almost identical in the two gen- era. The resemblance to Paromomys, especially P. depressidens, in the lower teeth is also very marked, the noteworthy differences aside from the divergent emphasis in the anterior teeth being in the some- what more progressive, or slightly different, specialization of the molar trigonids and the aberrant heel structure of M,; of P. depressidens. The trigonid structure, but not that of the heel of M; is somewhat more closely approached in Palaechthon. The very incomplete knowledge of Pronothodectes upper teeth sug- gests a basic resemblance to those of other genera here described, but makes it easier to see a few outstanding differences: the better para- cone-metacone separation and strong conule of P* and the less- marked posterointernal molar expansion in Pronothodectes, all of which are resemblances to Plesiadapis. There can be no question that Gidley was right in considering Prono- thodectes as closely related to Plesiadapis (‘‘Nothodectes’’). The prin- cipal differences are clear from the description. Pronothodectes has the enlarged incisor less procumbent, tooth reduction considerably less advanced and diastema not developed, cheek teeth less depressed and of somewhat simpler detail. In all these respects and also in the smaller size of its species, Pronothodectes is more primitive than Plesia- dapis, to which it seems surely to be ancestral in a structural, and perhaps also in a literal sense. Pronothodectes represents the earliest known member of a primate phylum analogous to the Elphidotarsius-Carpodaptes-Carpolestes phylum but with a greater known range in space and time. Its principal terms are Pronothodectes, middle Paleocene of North America, Plesiadapis, upper Paleocene of North America and Europe and, prob- ably, lower Eocene of North America, and Platychoerops, lower Eocene FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 165 of Europe. The European Cheiromyoides Stehlin is a contemporary and close relative of Plesiadapis. Nothodectes Matthew is clearly a synonym of Plesiadapis, as pointed out by Teilhard and accepted by all subsequent students.* Jepsen (1930 and 1934) tentatively placed Plesiolestes in this group, but the belief that it does not belong here has already been expressed above. He also (1934, p. 290) rejects my redefinition (Simpson, 1929c) of Platychoerops and its separation from Plesiadapis, but I have already defended this at some length (1935c). Cheiromyoides Stehlin, accepted as valid by Abel, Jepsen, and others but rejected as a synonym of Plesiadapis by Teilhard, differs less from typical Plesiadapis than does typical Platychoerops. It evi- dently represents an only slightly divergent and, as far as known, sterile side branch of the phylum. Genus PRONOTHODECTES Gidley Pronothodectes GrpLHyY, 1923, p. 12. Type.—P. matthewi Gidley. Mistribution.—Middle Paleocene, Fort Union, Montana. Diagnosis.—Dental formula so Lower incisor mucb en- larged, semiprocumbent, root laterally compressed. Canine (or possibly P,) small, slightly procumbent. P, 1-rooted. P, with quadrate base, short, high trigonid portion, no paraconid or meta- eonid. Talonid large with a single cuspule. Paraconid distinct on all molars, anteroexternal to the metaconid and progressively nearer the latter from M, to M3. No metastylid, but a vague cusp on the hypoconid-trigonid crest. M; with third lobe and elevated postero- internal rim with two or more poorly differentiated apices. P* with separate subequal paracone and metacone apices, their bases con- fluent, large conule mass, and strong protocone, which is, however, less expanded than in Plesiadapis. Upper molars Plesiadapis-like but simple, without major crenulations or secondary cuspules, and probably lacking the mesostyle (although this may be removed by wear in the known material). PRONOTHODECTES MATTHEW I Gidley Puate 8, Figure 1; Puate 9, Fiagurus 2, 11, 12; Pyare 10, Ficurss 3, 3a Pronothodectes matthewt Gidley, 1923, p. 12. Type.—U.S.N.M. no. 9547, part of right maxilla with P*-M?. Collected by A. C. Silberling. Paratypes.—U.S.N.M. no. 9332, left lower jaw with incisor root and crown of ?C, P2, Pu, and M,-3. Collected by Dr. J. W. Gidley. 67 It is, however, certain that Nothodectes gidleyi Matthew is specifically distinct from the European Piesiadapis triscuspidens Gervais. 166 BULLETIN 169, UNITED STATES NATIONAL MUSEUM U.S.N.M. nos. 10005 and 10044, isolated upper incisors (pertinence dubious). Collected by A. C. Silberling. Horizon and locality All known specimens from Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis—Sole known species of genus. See also description below. Discussion —The lower dentition is crowded, with no diastema. I, is a very large, laterally compressed, nearly procumbent tooth. Its crown is not known. Immediately above and behind the root of I, is a shallow, very small, and somewhat doubtful alveolus, probably for a vestigial I,. The following tooth, probably a reduced canine, is also small and 1-rooted. The root is slightly procumbent, the rather formless crown more so, as it projects obliquely upward and forward from the root. While Dr. Gidley did not discuss these anterior teeth, he gave the dental formula as jyoroa3; from which he evidently considered the first alveolus as doubtful and the tooth just mentioned as a first pre- molar. While the question cannot be answered definitely, the formula yor1133 seems to me much more probable. The morphology is indecisive, but in later plesiadepids P14, are always lacking and in probably related groups (especially the tarsioids) are apparently among the first teeth to be lost, while the canine is more tenacious, being still present in the upper, although not in the lower, jaw of the much more advanced genus Plesiadapis and seldom or never absent in the tarsioids even though it may be reduced.® P, is a simple tooth with one vertical root and a slightly procumbent crown excavated on the inner side and with a small 1-cusped heel. P; has two roots and is not reduced relative to Py. Its crown is not known. P, is similar to that of Plesiadapis, but its base is more quadrate, the trigonid portion is relatively shorter and higher, the heel is at least as large, relatively, or a little larger, but its transverse posterior crest rises to one apex, rather than two as usual in Plesiadapis. The paraconid is distinct on all the molars and is anteroexternal to the metaconid. From M, to M; it is progressively closer to the meta- conid and relatively smaller. The protoconid is about as high as the metaconid on M,, and on M; is somewhat and on M; much lower than the metaconid. The talonids of M,-2, which are considerably wider than their trigonids, are simple and basined. My, PLL EH jib FIGURE 38.—Claenodon montanensis (Gidley), U.S.N.M. no. 8362, skull, palatal view. Three-fourths natural size. (After Gidley, 1919, fig. 6.) ratio of total widths is 0.47, or about half, as stated by Gidley, in C. montanensis type, and 0.46 in the specimen of C. feror, which again is a wholly insignificant difference. Computation of the exact figures also shows Gidley’s impression that the vertical depths of the anterior faces of these bones are relatively less than in C. ferox to be mistaken. They are in fact slightly, but not significantly, greater than in this specimen of C. feror. 119212—37——18 184 BULLETIN 169, UNITED STATES NATIONAL MUSEUM In Gidley’s description of the caleaneum (1919, p. 552), ‘‘tibial face” is probably a misprint for “fibular facet.”” The type calcaneum is somewhat damaged in this region. A specimen collected after Dr. Gidley’s paper was published shows that in C. montanensis the fibular facet on the caleaneum is relatively quite as well developed as in CO. feror. His statement, ‘cuboid with facet for the astragalus, navicular and ectocuneiform arranged horizontally, nearly parallel and merging into each other” also appears to involve a misprint or lapsus, since it is inconsistent with his accurate figures and is either not clear or not correct. As his figures show, the astragalar facet is at an angle of nearly 90° to the navicular facet, and the latter and the ectocunei- FIGURE 39.—Claenodon montanensis (Gidley), U.S.N.M. no. 8362, foot bones: a, Lunar and scaphoid, dorsal view; b, part of tarsus, dorsal view; c, parts of the three median digits of pes, dorsal view. Natural size. (After Gidley, 1919, figs. 7 and 8.) form facet are approximately in the same vertical plane. Comparison of several specimens of C. ferox does not confirm the supposed differ- ence in this species in the separation and different outline of the two last mentioned facets on the cuboid. These facets are much less definite in the available specimens of C. ferox *! than in C. montanensis, but differ little in outline. Dr. Gidley’s important conclusion that the present limb bones are closely similar to those of Claenodon ferox is certainly correct and is only emphasized by these slight corrections of details. The C. mon- tanensis material is little over half of the size of that of C. ferox, and it differs in details of proportion, strength of processes, or rugosities, & The artist has made their outline far too distinct in Gidley, 1919, pl. 28, fig. 2a. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 185 such as are normally specific functions of size, but not in any essential structure. As already suggested, Dr. Gidley’s comparison with Ursus seems. to me to be beside the point and to involve no features not to be expected by convergence in heavily built plantigrades not more closely related than as members of wholly different groups of the same order. The species is surely distinct from C. ferox, being excluded from that group by the smaller size, reduced hypocone of M!, and reduced: size and more transverse proportions of M*. (Some of the other characters given by Gidley are not of probable specific value.) The size is within the probable range of the small unnamed Torrejon species. For instance, the length of M, is 9.2 in the largest Torrejon specimen of the small species and 9.4 in the type of C. montanensis. The size distribution doubtless overlaps, but were a larger series available it is probable that the mean for the Fort Union specimens would be found to be significantly greater. The reduction of the posterior part of M* is also within the extreme limit of variability of the Torrejon species but probably is a specific character varying about a different, but not widely separate, mode. The somewhat smaller hypocones of M'~? and the less conical protocone of P* seem to be beyond the limits of the Torrejon specimens and thus still better specific characters, on present data. Possibly correlated with the differences in P* is the somewhat narrower, more distinctly unicuspid character of P,, also apparently a good specific character.®” U.S.N.M. no. 9634 is an isolated M, from the Gidley Quarry, measuring 9.4 by 6.5 mm and referable to this species with little doubt. No. 6159, from Loc. 52," includes right and left M>, appar- ently associated, and other fragments, also referable to C. montanensis. CLAENODON SILBERLINGI (Gidley) Fiacure 40 Neoclaenodon silberlingi G1IDLEY, 1919, p. 552. Type—U.S.N.M. no. 8363, part of left®* maxilla with P?-M? and alveoli for canine and P!-*. Collected by Dr. J. W. Gidley. Horizon and locality —Gidley Quarry, Fort Union, Middle Paleo- cene horizon, Sweetgrass County, Mont. Diagnosis —Gidley (1919, p. 553): “A slightly smaller species than N. montanensis from which it differs as follows: cusps of all the molars seemingly! more depressed; m* and pm’s*:* somewhat less reduced 82 These, of course, substantiate the validity of the still unpublished Torrejon species, rather than of C. montanensis, which has priority and is surely distinct from any species previously named. 8 Labeled ‘‘Gidley Qu.’ (Recorded Loc. ‘No. 52’=Loe. No. 4).’’ Loc. 52 is very near the Gidley Quarry and about 50 feet lower stratigraphically. The difference does not appear to have any importance, but the localities are not exactly the same. § “Right” in the original designation of type (Gidley, 1919, p. 552) is a lapsus. “1 Though found in its original bed, the enamel of all the teeth is considerably damaged through weather- ing or leaching by surface water which had reached the specimen through cracks in the matrix.” 186 BULLETIN 169, UNITED STATES NATIONAL MUSEUM although more reduced than in Claenodon: all the cheek teeth, ex- cept p’*, relatively wider; distance between p* and the canine relatively greater indicating a somewhat more elongate face; infraorbital foramen approaching nearer to the alveolar border above p’. “This species in size approximates C. protogonioides (Cope) but is apparently clearly distinguishable from the Puerco species by the much greater reduction and more oval contour of m*, and in the relatively wider proportion of all the cheek teeth. “MEASUREMENTS OF N. silberlingi citoymts ts bocce tsodie set 254 ee eee ie a ee ssa 60.0 mm DORMS sac om a i ea ee 34.5 mm TMstO Meee ea ee ee ae ee eee ee Soe as see eos 21.4 mm mblengthess 22. sacs ost et eas cana nas + Rene aee cone Sees 7.0 mm. ATE WICC Bon ose eae Se eee Re SE eed ees ee PEs eee 8.8 mm. meen gties, 2602. 4. ts hee eee Be ee eae ane 7.4 mm. metwidthestne yo8 sel Se Sh soa eh ese Sake te es ese S 11.5 mm. any wen Pheer 2S. ok SE a ae nS bor eee oes ae ON A Se 5.0 mm, MieaWiGl this see oe ee oe eee ee eee ee re iora ae 9.0 mm. 5.0 mm 5.5 mm 7.5 mm,” FIGURE 40,—Claenodon silberlingi (Gidley), U.S.N.M. no. 8363, left upper jaw: a, Crown view; 0, external view. Natural size. (After Gidley, 1919, fig. 9.) Discussion —Allusion to ‘‘C. protogonioides” implies comparison rather with the unnamed Torrejon species of Claenodon than with the different Puerco species to which the name properly applies. The type of this species is so poor, the cusp structure being very much obscured and in part destroyed by corrosion and breakage, that it cannot surely be distinguished from the small Torrejon species, on one hand, or from C. montanensis, on the other, although these two are distinct from each other. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 187 The relative size and contour of M? are, in fact, well within the range of variation of the Torrejon species, and the teeth are not significantly more transverse. The size is large for that species but not beyond its presumable range and hence of doubtful significance. It is impossible to give good clear diagnoses separating the two, although it is my opinion that they are probably distinct. It is, on the other hand, probable that C. silberlingt is synonymous with C. montanensis (adding to the probability that it is not the same as the Torrejon species). The size is about the same, and both types are from one quarry. Mlis a little shorter in C. silberlingi, and some- what more transverse, but these are doubtfully real, since the speci- men is so poorly preserved, and if real are not marked enough to prove any taxonomic distinction. M?® is definitely larger than in C. mon- tanensis. The ratio of their lengths, the dimension in which they differ most, is 1.10. In Torrejon specimens of the C. feror group the variation in absolute dimensions is much greater than this, and the size of M’ relative to M! or M? also varies quite as much as the differ- ence between C. silberlingi and C. montanensis, although this is a more constant figure. The distance between P? and the canine is almost exactly as in C. montanensis. In calling it “Telatively greater’, Gidley must have meant relative to the length of M!, but as the other tooth dimensions are as great as in C. montanensis this is simply to repeat that the length of M! is relatively small and is not a character of the diastema. M!*? may have had larger hypocones than in C. montanensis, but this is almost hypothetical, and the protocone of P* may have been stronger, also rather dubious. I retain the name tentatively, on these very doubtful characters, but believe that the species will probably prove to be invalid. CLAENODON LATIDENS (Gidley) Ficure 41 ? Neoclaenodon latidens GipuLEyY, 1919, p. 554. Type-—U.S.N.M. no. 8388, right lower jaw with M,-3, a small fragment of M,, and the broken lower part of the ramus from the canine alveolus to the angle. Collected by Dr. J. W. Gidley. Horizon and locality. Gidley Quarry, Fort Union, Middle Paleocene horizon, Sweetgrass County, Mont. Diagnosis.—Gidley (1919, pp. 554-555): “Size approximately that of N. montanensis, but with decidedly wider molars; jaw relatively longer, much straighter, and more slender. Since the upper dentition of N. latidens and the lower dentition of N. silberlingi are not known, these species can not now be compared, but the difference in size seems sufficient to distinguish them. 188 BULLETIN 169, UNITED STATES NATIONAL MUSEUM “MEASUREMENTS Tengthiof mpreLo eas 5 Le Se ee 2 eee Se Soe cce ea eeees 9.7 mm. WWaACthiOt ingot ee Sa ee ee ea NE eee oe 8.4 mm. Wengiheiigss Goes eee ee es Ne ee ees) Sele ee eee 9.6 mm. Wiidlithrofarnises=, tate rep tees et eens BITE N ble SCRE Sere) Yank ee ea ee 7.6 mam. Depthvol jaweal Mee = ase ee ee ee eee 16.0 mm. “Unfortunately, as in the type of N. silberlingi, the enamel of the inolars has been considerably damaged through weathering or leach- ing by surface water, which somewhat obscures the detailed structure. “The generic reference is provisional, since certain features, as the straight and more slender proportions of the jaw and relatively greater width of the lower molars, so sharply distinguish N. latidens from all other species of this genus or of Claenodon. They suggest that its affinity to the group to which I here assign it may be, after all, not very close. More complete and better preserved material may, therefore, necessitate placing it in a distinct genus.”’ WO ES FIGURE 41.—Claenodon latidens (Gidley), U.S.N.M. no. 8388, with parts in outline probably adapted from other species of Claenodon, right lower jaw: a, Crown view; 6,external view. Naturalsize. (After Gidley, 1919, fig. 10.) Discussion.—The apparent differences from C. montanensis in the suape of the mandibular ramus are, in my opinion, illusory and due to the different preservation of the two specimens. ‘The jaws appear to have been almost identical or at least well within the possible range for asingle species. So far as available material goes, the species depends wholly on the notably wider M,_3._ That thisis of specific value is not certain, especially as the size is otherwise that of C. montanensis and both are from the same quarry, but the species may be tentatively retained. Isee no reason to suspect that a new genus is represented. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 189 CLAENODON VECORDENSIS Simpson Figure 42 Claenodon vecordensis Stimpson, 1935d, p. 232. Type.—U.S.N.M. no. 13781, left M?-3. Collected by A. C. Silber- ling and G. G. Simpson, 1932. Horizon and locality —Loc. 9, 300 feet above the base of Fort Union No. 1, Crazy Mountain Field, Mont. Diagnosis —M? similar to that of C. silberlingi in outline but 10-20 percent larger and somewhat more transverse; hypocone vestigial, strong crenulated internal cingulum. Mj? relatively as large as in C. ferox and similar in structure except for smaller metacone and more evenly rounded external border; vestigial hypocone present. Discussion.—This species is smaller than the smallest known variants of C. feror, has a smaller hypocone on M?, and the contours of M? and M? are different and beyond the known range of variation of that species. C. silberlingi is somewhat smaller and has different tooth proportions. M?* is much larger and less transverse than in C. mon- tanensis and has the posterior part better developed. The small unnamed Torrejon species is much smaller and has less wrinkled enamel, and the internal cingulum is feebler or absent. FIGURE 42.—Claenodon FIGURE 43.—? Claenodon vecordensis Simpson, sp., U.S.N.M. no. 6158, U.S.N.M. no. 13781, left left Mi, crown view. M?-3, crown view. Nat- Natural size. ural size. It is unfortunate that another poorly known species must be added to this genus, already burdened with several species of doubtful status, but the present specimen is identifiable and surely cannot enter into any species previously established so far as their range of variation is known or can be fairly inferred. It is, furthermore, from a very dif- ferent geological horizon from the other Fort Union claenodonts and one from which little material has been obtained, so that its strati- graphic importance also necessitates some convenient designation for it. Measurements: Length M?, 9.0; width M?, 13.5; length M°, 6.7; width M?, 10.0. ?CLAENODON species FIGURE 43 U.S.N.M. no. 6158, from the Gidley Quarry, is a left lower jaw frag- went with M,, Among Dr. Gidley’s notes are two sheets devoted to a 190 BULLETIN 169, UNITED STATES NATIONAL MUSEUM description of this specimen, in which it is placed in a new species. Apparently it was at first placed in Claenodon with a query, and later the generic name was erased and another, which appears to be new, substituted, but no corresponding change was made in the description, and the genus was not defined. I do not believe this to be adequate for the definition of either genus or species and feel obliged to suppress these manuscript names. The specimen is aberrant (with respect to species of Claenodon) in several details, but their significance cannot be judged and comparative Fort Union material is too scanty for good diagnosis. It is, for instance, entirely possible that this belongs to Claenodon latidens.® Genus DEUTEROGONODON Simpson Deuterogonodon Simpson, 1935d, p. 232. Type.—D. montanus (Gidley). Distribution.—Middle Paleocene, Fort Union, Montana. Diagnosis.—Dentition basically arctocyonid in type, and resembling Protogonodon and Claenodon. Small, distinct, cingulum hypocone on M?-3 (at least), cingula almost completely circling these teeth. Small but well-defined mesostyle present. Parastyle of M* a distinct cusp, crowning a lobe projecting strongly externally. Lower molars with trigonid only slightly higher than talonid, metaconid smaller than but as high as protoconid. Paraconid very small but distinct, subconical, on slope of metaconid directly anterior to its apex. Talonid basin open, crescentic lophid continuous but crest differentiated into bypo- conid, hypoconulid, and entoconid, progressively smaller in that order. Enamel wrinkled, but all cusps clear-cut and little or no tendency to form crenulations or accessory cuspules. Discussion.—This seems to be a very distinctive genus, at once distinguished from any similar form by the presence of a mesostyle. The combination of the other characters given is equally distinctive, although individually they are less so. The genus could be a deriva- tive of Protogonodon, although it is too incompletely known and too distinctive to establish this as a definite theory. I know of no Torre- jon genus that compares more closely than the probably related but manifestly distinct Claenodon, and none of the lower Kocene arctocyo- nids could be derived from Deuterogonodon. The possibility that Phenacodus was derived not from Tetraclaenodon, as commonly sup- posed, but from Protogonodon by way of a form something like Deu- terogonodon is worthy of consideration but cannot be very seriously 8 Dr. Gidley may have had the same idea, and his manuscript is perhaps older than the publication of 1919. The specimen was collected before the type of C. latidens and was obviously in Dr. Gidley’s hands when he wrote his claenodont paper, so that its omission may well be due to his having decided that the specimen did not warrant a new name. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 191 upheld on present evidence. Aside from the presence of a mesostyle, Deuterogonodon is much less like Phenacodus than is Tetraclaenodon.® Dr. Gidley’s manuscript notes include two drafts of a description of the type of this genus, in both of which it is referred to Protogonodon. On one, however, almost surely the second, the words ‘‘new genus” have later been written under ‘‘Protogonodon.’”’ Gidley thus came to recognize the clear-cut distinction of this genus from Protogonodon, but his notes do not contain any generic diagnosis or new generic name, and I have been forced to supply these. DEUTEROGONODON MONTANUS (Gidley) Ficure 44 Deuterogonodon montanus (GIDLEY) Simpson, 1935d, p. 233. Type—U.S.N.M. no. 6160, part of right maxilla with a fragment of M!, M? lacking the paracone and parastyle, and M* complete, with a left lower jaw fragment, possibly of the same individual and almost surely of the same species, with the talonid of M, and most of M,. If these should prove not to be one individual, the upper teeth constitute the type, and the lower teeth are a paratype. Collected by A. C. Silberling. Paratype-—U.S.N.M. no. 6161, isolated right Mo. Horizon and locality —All material is from Loe. 25, about 300 feet above the base of Fort Union No. 2, Sweetgrass County, Mont. FIGURE 44.— Deuterogonodon Diagnosis.—Gidley: ‘‘SSomewhat larger than P, anus (Gidley): a, U.S. N.M. no. 6160, right M?-3, [Protogonodon] pentacus (Cope).’’ *” crown view; 6, U.S.N.M. Simpson: Sole known species of the genus as 7° 6161, right Ms, crown view; 6’, same, internal view. defined above. Natural size. Measurements are as follows: IMe2tinediantwid tinct ty Wes eter ne dete hh ee 14.6 Ne Neng thie. * sy hee ee es 2 ess he eee: See ee: Be 10 Mies (paratype)! wach hes Sel elo pe eee re ae ee 10.5 IMGs (Araby pe) tenet ne 2 a ao See ro ee 12. 6 8 A new species of Profogonodon from the Puerco, which I have described in a note published as a supplement to Matthew’s Paleocene Memoir, suggests that within the genus Protogonodon there was a tendency to develop along two different lines, one leading (or related and collateral) to Telraclaenodon and one more definitely creodont and Claenodon-like. If derived from Protogonodon, Deuterogonodon probably arose from a species of the latter, rather than of the former, group. 87 I quote only enough of Dr. Gidley’s diagnosis to establish his authorship of the species. The rest of the diagnosis compares with the Puerco species of Protogonodon and is hence rather generic than specific, and among the few characters given I cannot agree as to the reality or value of some, and others seem to involve slips of the pen that I cannot correct with any certainty that Dr. Gidley’s thought is being followed. The rough manuscript was far from completion. 192 BULLETIN 169, UNITED STATES NATIONAL MUSEUM ?DEUTEROGONODON species U.S.N.M. no. 9653 includes part of a left maxilla with P*-M? and part of a right mandible with the heel of M3. All the teeth are incomplete and much battered, and none is directly comparable with the types of D. montanus, so that pertinence to that genus and species cannot be established but is probable, at least as far as generic identity. The principal characters exhibited are that P* has a strong conical protocone, somewhat smaller than the external cusp or cusps, and that in the talonid of M; the three cusps, especially the entoconid, are more distinctly separated than in M,-» of the type of D. montanus. These specimens are from Loc. 18, the horizon of which is in doubt but is higher than that of the types of D. montanus and in Fort Union No. 3. There are likewise a few tooth and limb fragments, from the type locality, that probably belong to this genus but are of no particular value at present. Subfamily OXYCLAENINAE Matthew, 1937 (Oxyclaenidae Scott, 1892) The oxyclaenine genera of this fauna, Metachriacus, Mimotricentes, Spanozyodon, Prothryptacodon, and (somewhat doubtfully) Chriacus, belong with several other Paleocene genera in a very confusing com- plex. The size ranges of their species do not differ greatly, and their morphology is markedly stereotyped in general pattern. Within the limits of this general type, however, they seem to ring almost every possible change in combinations of detail, so that they are difficult to distinguish yet are amazingly varied in minutiae and cannot be grouped into one or a few broad but natural genera. The variations involve, among other features: 1. Placing of the canine (from vertical in Tricentes, etc., to strongly procumbent in Prothryptacodon, etc.). 2. Number and crowding of premolars and associated develop- ment of diastemata, varying from such types as Metachriacus to Spanoxyodon. 3. Molarization of Py, from a wholly premolariform type, as in Tricentes (and several other genera) to a submolariform condition as in Spanoxyodon. 4. Reduction of the paraconid, from strong and distinct (e. g., Deltatherium) to almost indistinguishable (e. g., Mo-3 of Metaehriacus). 5. Placing of the paraconids, from almost strictly internal (e. g., Mimotricentes) to almost strictly median (e. g., Tricentes). 6. Elevation of the trigonids and their shearing character, from high and strongly shearing (e. g., Chriacus) to low and bunodont (e. g., Tricentes). 8 Dr. Gidley labeled them as of the same species. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 193 7. Proliferation of accessory cuspules, from practically nil (e. g., Chriacus) to considerable and distantly approaching the multicuspid ‘“arctocyonines” (e. g., Metachriacus). 8. Development of hypocones on upper molars from practically nil (e. g., Deltatherium) to pronounced (e. g., Chriacus). Other characters, known in fewer genera, also seen to be highly distinctive, such as the molarization of upper premolars, width of external upper molar shelf, shape of paracone and metacone, and many other characters. Table 38 contrasts the genera of this complex that occur in this fauna, and those most likely to be confused with them, as regards these characters. TABLE 38.—Comparison of dentition characters in six genera of Oxyclaeninae a Genus Canine Pi-4 Diastemata Ps GChriacus_ 2. .-.-- Moderately procum- | Present..__| Slight or none...--_- Slender, with small, dis- bent. tinct metaconid. Metachriacus-_--_- Ci Chiiacusiesa oe ae |e G022222= ChiGhriacue-2222_- Cf. Chriacus. Dricentes--=--=—-- Crown vertical__...___- P; absent__} Short, C-P:__-----.- Stouter, no metaconid. Mimotricentes....| Cf. Tricentes__..------- Presentes.-| sNones =. == 2-2-2 =e Cf. Tricentes. Spanoxryodon.--_-- Cin Chrideus 2-222 P;-2 absent_| Long, C-P3._.__---- With metaconid larger than in Chriacus. Prothryptacodon__| Strongly procumbent, | Present_.__| Very slight, around | Slender, metaconid root extending be- Piva. barely incipient or neath premolars. absent. | rr ————————————————— eel Genus | Paraconids of Mo-3 Chrigcus.— .----=- | Distinct, near meta- conids, internal. Vestigial, placed about as in Chri- acus or slightly less internal. Metachriacus----- Vestigial, low on crown, nearly me- dian. PMICENLES === Mimotricentes_._.| Vestigial, higher on crown than in Tricentes and in- ternal. About asin Chriacus or slightly less in- ternal. Prothryptacodon..| Comparable to Me- tachriacus, but slightly better differentiated. Spanoryodon...-- Trigonids of M1-3 in still Lower than Chriacus, shearing. Still, lower, buno- dont. WiChetmecentesaa = o— About asin Chria- cus or slightly lower. Cf. Chriacus___..-- Elevated, shearing_ Cuspules of My-3 No accessory cus- pules. Tendency to de- velop cuspules on anterior rim _ of trigonid, conules on both hypoconid wings, and slight metastylid. Enamel rugose or papillated but few or no definite cus- pules. Chitnicentese= = Cf. Chriacus. - Ch Chridcuss2 2 === Mi-3 With well-developed hypocones, some tendency to devel op protostyles Sharp cingulum around protocones of M!-3 but hypo- cones rudimentary on M?, absent*on M3, no distinct protostyles. Hypocones small or absent, no proto- styles. Cf. Tricentes. (Unknown.) (Unknown.) 194 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Renewed study has necessitated some modification of the arrange- ment given in a preliminary paper (Simpson, 1935d). In the first place the species there called Ciriacus pusilius was wrongly placed in Chriacus. It belongs in Metachriacus, where it is indistinguishable from the genotype, M. punitor. This makes possible some redefini- tion of Metachriacus, which proves to be even more distinct from Chriacus than was at first supposed. The species called Tricentes latidens (quoted from Dr. Gidley’s notes) is now seen to be distinct from Tricentes, despite a very marked adaptive resemblance, for it has P;, apparently always absent in Tricentes, and the paraconids are in quite a different position although reduced about asin Tricentes. It is necessary to erect a new genus for this species, since redefinition of Tricentes to include it would make that genus so broad as to destroy all balance in the generic arrangement of the family achieved by Matthew and others. Spanoxryodon is a peculiar type with aberrant specialization, and Prothryptacodon is evidently a very primitive form slightly but de- finitely modified in the direction of Thryptacodon. For convenience in identifying fragmentary material a summary (table 39) of lower tooth dimensions is here given. TABLE 39.—Measurements (in mm) of lower dentition in seven species of Oxyclaeninae! Py Mi | M; | M3 Species and number of specimens ——————— L WwW L Ww L WwW L WwW Prothryptacodon furens (2)...-------- 4.9 Qa 5.2 3.7 5.2 4,2 6.3 4.0 TLCRTMACUS DUGNETI() i. 2. see} seseos| so2= 22 |e 55522 (pe! 5.9 Metachriacus punitor (9) ...-----.---- 4.3 27 4.7 Sat 5.0 4.3 Seal 3.7 Metachriacus provocator (12) __..----- 5.0 2.9 54 4.2 5.8 4.9 6.7 4.1 Spanorydon latrunculus (1) ---.------ 5.0 2.8 Baz 3.9 5.8 4.5 Mimoiricentes latidens (2)..-----.---- 5.4 3.9 5.6 4.7 6.0 5.5 6.1 4.7 Mimotricentes angustidens (4)_.------ 4.4 3.0 4.8 3.6 5.6 4.6 5.5 3.6 1 The figures in this table are means for the number of specimens (in all, not necessarily for every dimen- sion) shown in parentheses. The range of variation is known reasonably well only in Metachriacus punitor and M. provocator, as given hereinafter. The dimensions in themselves are not necessarily distinctive, for instance between Metachriacus provocator and Spanoryodon latrunculus, but in such cases there are, of course, well marked nonnumerical distinctions. Some worn and incomplete specimens might, however, be unidentifiable in such cases. Genus PROTHRYPTACODON Simpson Pr othryptacodon Stmpson, 1935d, p. 233. Type.—Prothryptacodon furens Simpson. Distribution —Middle Paleocene, Fort Union, Montana. Diagnosis.—Canine semiprocumbent, root extending beneath pre- molars, as in Zhryptacodon. P,-, spaced widely. P, similar to Thryptacodon. Molar trigonids higher than in Thryptacodon, para- FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 195 conids reduced and in nearly the same position as in Thryptacodon but more distinct, higher on crown, trigonids less basined, and with fewer accessory cuspules. Only one distinct inner talonid cusp (entoconid), as opposed to two in Thryptacodon. Discussion.—This genus could well be ancestral to Thryptacodon and in any case is evidently allied to it. The adaptive characters are somewhat intermediate between the more primitive types, such as Oxyclaenus, and the more complex type seen in Thryptacodon, involv- ing some flattening of the molar crowns and proliferation of cuspules. In these characters Thryptacodon and to a less extent Prothryptacodon parallel or converge toward the Claenodon—Anacodon line. The latter group, however, is much earlier and more highly specialized in this rather bearlike direction. Prothryptacodon is contemporaneous with Claenodon in the Middle Paleocene, and Claenodon is already more specialized than Thryptacodon in this direction. Thryptacodon appears in the Upper Paleocene and survives into the lower Eocene, where Anacodon appears as a highly aberrant survivor of the Claenodon group. The canine is slender, laniary, compressed, with the root con- siderably larger than the crown and meeting the latter at an angle of about 135°. The root is implanted almost horizontally and extended at least to the anterior end of P,. P,, from its alveolus, had a single root and was well spaced, slightly nearer to the canine than to P,. P, and P; were 2-rooted, and P, is preceded and followed by a short diastema. PP, has a very slight internal swelling, high on the crown, that may indicate an incipient metaconid. In one specimen M, has a faint external trigonid cingulum, and in another this is more definite. There are two mental foramina, one beneath P, and one beneath P3 or the anterior end of P,. The other characters of genus and species are adequately given in the diagnosis and figures. PROTHRYPTACODON FURENS Simpson Figure 45 Prothryptacodon furens Simpson, 1935d, p. 234. Type-—U.S.N.M. no. 9260, right lower jaw with P,-M; and alveoli. Collected by A. C. Silberling. Horizon and locality—Gidley Quarry (referred specimen from Silberling Quarry), Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis—Sole known species of genus. Measurements in table 40. Remarks—Only two specimens are as yet known, both partia] lower jaws. U.S.N.M. no. 9262 is from the Silberling Quarry. 196 BULLETIN 169, UNITED STATES NATIONAL MUSEUM TaBLE 40.—Individual measurements (in mm) of Prothryptacodon furens Py Mi Ma; M3 Specimen L WwW L WwW L WwW L WwW Types Neeley ae ore raze! 4.9 207 5.2 317 5.2 4,2 5.3 4.0 UiSeNeMisn0: 9262424. a. tessa eso ne seco ls See Ue Eee eh aoe 5.1 4.2 Ni RR Riser ° Cat | hos FS cg a —- SS Ser FIGURE 45.—Prothryplacodon furens Simpson, U.S.N.M. no. 9260, right lower jaw: a, Crown view; 6, internal view. Twice natural size. FIGURE 46.—Chriacus pugnar Simpson, U.S.N.M., no. 13782, right lower jaw, crown view. Twice natura size. Genus CHRIACUS Cope, 1883 This genus typifies the Torrejon but has been reported also in the Puerco and in the Almagre. It is improbable that a single genus had this enormous, almost unique length of distribution, and the earliest and latest forms, known only from fragments, may well prove to be distinct. The pattern is a simple and generalized one, easily confused with that of allied forms, and requiring close study and good material for its certain distinction. The appearance of Chriacus in the present fauna would be expected, from its age relationships, and this genus was reported by Douglass (1908) and by me (1935d). The earlier reports, however now prove to be doubtful or erroneous. As elsewhere noted, one species hitherto referred to Chriacus is now placed in Metachriacus. The other, C. pugnaz, is retained here, but the specimen on which it is based is so incomplete that the generic reference is not certain. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 197 CHRIACUS PUGNAX Simpson FIGURE 46 Chriacus pugnax Stmpson, 1935d, p. 235. Type.—U.S.N.M. no. 13782, right lower jaw with M,_, and alveoli. Collected by A. C. Silberling and G. G. Simpson. Horizon and locality—Loc. 78, Fort Union, Crazy Mountain Field, Mont. Diagnosis —About the size of C. pelvidens, but molars wider, tri- gonids less elevated, talonids of M,_, notably wider than trigonids. M, length 7.1, trigonid width 4.9, talonid width 5.9. Remarks.—In addition to the characters cited in the diagnosis, there is a diastema anterior to P;in the unique specimen, but the length and significance of this cannot be established. The specimen is from one of the lowest horizons that have yet yielded mammals in this field, only 200 feet above the base of the Fort Union No. 1. This suggests comparison with Mimotricentes angustidens, also from a low (but not such a low) horizon, but the latter is markedly smaller and otherwise less Chriacus-like. Genus METACHRIACUS Simpson Metachriacus Simpson, 1935d, p. 235. Type.— Metachriacus punitor Simpson. Distribution —Middle Paleocene, Fort Union, Mont. Diagnosis ——Canine and premolars about as in Chriacus. Molar trigonids lower than in Chriacus but still shearing rather than bunodont (as in Tricentes). Paraconids reduced and poorly distinguished, near metaconids, placed about as in Chriacus or slightly less internal. Accessory cuspules developing on anterior rim of trigonid, on hypoconid wings, and a slight metastylid. Upper molars with sharp cingulum around the protocone, but no protostyle. Hypocone present on M?’, rudimentary or indistinguishable on M?, and absent on M®. Remarks.—This genus is hardly distinguishable from Chriacus on the basis of the premolars, which distinguish it from almost all other known genera, but the molars are of quite a different adaptive type, the lower crowns, blunter cusps, and proliferation of accessory cus- pules in the lower molars representing convergences in varying degree toward Tricentes and toward Claenodon. METACHRIACUS PUNITOR Simpson Figures 47-49 Metachriacus punitor Simpson, 1935d, p. 235, Chriacus pusillus Simpson, 1935d, p. 234. Type.—U.S.N.M. no. 9288, left lower jaw with M,-3. Collected by A. C. Silberling. 198 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Paratype.-—U.S.N.M. no. 9286, right lower jaw with P;—-M; (M, and M; broken). Collected by A. C. Silberling. Type of Chriacus pusillus—U.S.N.M. no. 9270, right lower jaw with P.-M;. Collected by Dr. J. W. Gidley. Horizon and locality —Gidley and Silberling Quarries, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. : te ‘ED " EE FIGURE 47.—Metachriacus punitor Simpson: a, U.S.N.M. no. 9288, left Mi-3 crown view; 6, U.S.N.M. no. 9286, with parts in outline from nos. 9282 and 9486, right lower jaw, external view. ‘Twice natural size. H \ 4ff Z \\ \ At y I, SR Sey (2 (ee FIGURE 48.—Metachriacus punitor Simpson, small variation, U.S.N.M. no. 9270, right lower jaw, internal view. Twice natural size. FIGURE 49.—Metachriacus punitor Simpson, U.S.N.M. no. 9331, left M!-3, crown view. Four times natural size. Diagnosis.—Heel of P, expanded, basined, squarely truncated pos- teriorly. Molar crenulation moderate. My,_». less wedge-shaped. Measurements given in tables 42 and 43. Discussion.—My earlier publication on the two species here united was confused and incorrect. By a misinterpretation of a small worn specimen and by the acceptance of an incorrect association, I was led to place a few of these specimens in Chriacus and therefore I did not FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 199 properly compare them with the better materials manifestly belonging in a distinct genus, named Metachriacus. With the removal of the extraneous material and renewed comparison of a large number of specimens, it is clear that Metachriacus punitor and ‘‘Chriacus pusillus” both belong in Metachriacus. The specimens previously referred to “Chriacus pusillus”’ are smaller than those placed in Metachriacus pumitor, and M; is slightly more reduced. Nevertheless, with recog- nition that they are congeneric and in view of the fact that all are from one horizon and locality it appears that no sharp division between the small and large specimens can be made and that they are merely individual variants of one species. This is borne out by the statistical constants given below. Of I,-P, only the alveoli are known. Judged from these, there were three small, subequal, closely crowded incisors, the canine was moderately enlarged and procumbent (but less so than in Prothrypta- codon), and P, was a small tooth implanted vertically by a single root. P23 are similar, but P; is larger. Both are slender, 2-rooted, pointed teeth, the outer face convex, with a sharp anterior crest and the posterior and anterointernal faces excavated. There is a small barely cuspidate heel and a tiny anterior basal cuspule. PP, is con- siderably longer than P; although barely higher. The anterior cuspule is more distinct and the talonid much more developed, its internal half basined. There is a high rudimentary metaconid, not well differ- entiated from the protoconid, closely similar to that of Chriacus. The lower molars are of generalized arctocyonid type except for the special characters already listed. The symphysis is shallow, weak, and unfused, the horizontal ramus long and slender, the mental foramina beneath P, and P3. Upper teeth have not been found associated with lower jaws of Metachriacus punitor, but there are several isolated upper teeth and jaw fragments, including representatives of the three upper molars, that are from the same level and locality, are harmonious in size and structure, and may be referred to this species with some assurance. In outline they closely resemble Chriacus but are distinguished by the entire absence of a protostyle and the considerably lesser development of the hypocone. TABLE 41.—IJndividual measurements (in mm) of Metachriacus punitor P: P3 | Py Mi M: M; U.S.N.M. no. | | Ss | S| MH —_ | ——_——_ L WwW L WwW L WwW L Ww L WwW L Ww PAS Se re Papa tee ee | eee | | eee el oe eee eS] MES) | ech el c hes PAGE) 3.8 OUS6 eee ace aaa e eens | eee ae ea| ee eee ee eee Ca. 14714 || 2:9! (2s 222 3.9] 52) 4.4 O27 OLN ssa Eases ee 2.8} 1.6] 3.5] 2.0 4.4) 2.6) 49] 3.7] 50] 43 119212—37——14 200 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Measurements of the three principal specimens and statistical con- stants of the whole series are given in tables 41 and 42 (see also fig. 4). TABLE 42.—Numerical data on lower dentition of Metachriacus punitor Variate N R M o V Joye son as 6 4, 1-4. 7 4, 32+90. 08 0. 20+0. 06 4,541.3 WAR eee eae 6 2. 4-3. 0 2. 72+0. 08 0. 20+0. 06 (silts Daves iMG ste eae 5 4,4-4.9 4. 6640. 07 0. 16+0. 05 3. 5#1.1 WieM ewe ote Se 8 3. 5-3. 9 3. 71+0. 05 0. 16+0. 04 4.1+1.0 Mig eee 8 4, 6-5. 4 4. 98+0. 09 0. 27+0. 07 5. 441.3 Wiens Seiad 9 4, 1-4, 5 4, 31+0. 04 9, 11+0. 03 2.6+0.6 AIM ee eee 6 5. 1-6. 1 5. 72+0. 14 0. 354.0. 10 6.2+1.8 IWiMig ees 6 3. 6-3. 8 3. 688+0. 04 0. 09 +0. 03 2.4+0.7 Three specimens from the Silberling Quarry, all very incomplete, are referred to this species. Each is somewhat aberrant but not consistently among the three and not beyond the established range of variation. They are not included in the figures in table 42, which are based on a pure sample from the Gidley Quarry. METACHRIACUS PROVOCATOR Simpson Figures 50-52 Metachriacus provocator Stmpson, 1935d, p. 235. Type.—U.S.N.M. no. 9278, left lower jaw with P,-M3;. Collected by Dr. J. W. Gidley. Horizon and locality—Type from Loc. 51, referred specimens from Locs. 81, 50, 25, 24 (all below Gidley Quarry), Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis.—Heel of P, less expanded, little or not basined, more pointed posteriorly. Molars markedly crenulated. M,-, more wedge-shaped, i. e., talonids markedly wider than trigonids. Slghtly but significantly larger than M. punitor in most dimensions (see table 43). Discussion.—This species was originally based on three specimens (Simpson, 1935d), but since then five more, less complete, have been identified in the National Museum collection, and 10 have been col- lected for the American Museum, so that it is now one of the best- known species in the fauna. It does not occur in the quarries, where it is replaced by the allied but certainly distinct M. punitor, but occurs at the widely scattered Locs. 24, 25, 50, 51, and 81 (specimens from 25 and 81 only in the American Museum, but used in this study), all of which are somewhat below the quarries, although all in Fort Union No. 2. Both upper and lower jaws are known from each locality. It is probable that the distribution, below the quarry levels, is accidental, FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 201 that is, that VM. provocator is not a significantly older species replaced in time by M. punitor, but that they may well have been really contem- poraneous but living in different facies. In addition to the characters cited in the diagnosis, the two speci mens of M. provocator that show the alveoli of P, had this tooth relatively more reduced than in M. punitor, with more definite diaste- mata before and behind it. The upper molars, well known in this species, have the hypocones more definite and more projecting internally than in M. punitor, and M3 is less reduced and less transverse. There is also a tendency to develop a rudimentary protostyle on M?. These characters make the upper teeth closer to Chriacus than are those of M. punitor, and the upper dentition of M/. provocator would not perhaps in itself be sepa- rated generically from Chriacus, but its lower dentition shows the generic characters even more clearly than does that of M. punitor. There are three lower and three upper jaws, some associated, from each of Locs. 25 and 51, three lower and one upper from Loc. 81, two lower and one upper from Loc. 50, and one lower and one upper from Loc. 24. Deviations exist, of course, between the material from dif- ferent localities, but these are not consistent and are not statistically significant. The samples are too small to demonstrate racial dif- ferences, if such exist. In fact the whole combined sample does not exceed the variety usual in a homogeneous species but on the con- trary shows unusually small variation, as shown by the figures in table 43 (see also fig. 4). TaBLE 43.—Numerical data on upper and lower dentition of Metachriacus provocator Variate N R M z(d?) o V Mies 22S 3 | 4.8-5.1 | 4. 97 0. 0467 WP ee Soc s}) 63) (289 2. 90 0 2h Ee 7.| S.2—5 S |Pbrot 2 0008nieS 2522 0. 205+0. 055 | 3.84+1.0 Winineson sos 7 |) 359-45 63 a4. 20.09) |b=s5-5=" 0.24 +0. 06 o. 7£1.5 ih oe (ta G-OR0 2 057 93-05 Os |e oee = 0.12 +0. 03 2. 2+0. 6 i. ho 1. \245G-55 214s Ole OS OF (eee seoee 0.19 +0. 05 3.8+1.0 oie | 4 6 TOF |" 6.70 0. 1400 Wise. 5 | 4.0-4.3 | 4. 12 0. 0680 LE yee 2 | 4.6-5.0 | 4 80 0. 0800 iG Sa ae 2) 4.8 4, 80 0 1G) Ee 5 | 5..0-5. 4 |. 5. 18 0. 1080 END re 2 ra 5 | 6. 1-6. 7 | 6. 30 0. 2600 iia ae 7 oe ao. iMon.)O7 2-0, OO: | 225 ee 2 0.24 +0. 06 4.2+1) i NV NES ee oe etl) ee Seed lero Os OF wee oe ee | 0.18 +0. 05 2.4+0.6 1S). OS ee 4|4.3-4.8 | 4 55 0. 1700 sid Sale eee 4 | 5.8-6.4 | 6.15 0. 2700 In only one case does the coefficient of variation exceed 5 (this would be true also if this coefficient were calculated for all the variates). 902 BULLETIN 169, UNITED STATES NATIONAL MUSEUM ign FIGuRE 50.—Metachriacus provocator Simpson, U.S.N.M. no. 9278, left lower jaw: a, Crown view; 6, internal view. ‘Twice natural size. Figure 51.—Metachriacus provocator Simpson: a, U.S.N.M. no. 15126, left M2-3, crown view; 6, U.S.N.M. no. 9259, right lower dentition, crown view. ‘Twice natural size. FIGURE 52.—Méetachriacus provocator Simpson, U.S.N.M. no. 9259, left M!-*, crown view. Four times natural size. The difference from M. punitor is so obvious that its significance requires no proof. In every case the deviation between the means is significant, and in the available samples the observed ranges overlap, barely, only for the widths of Py and M,. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 203 The individual dimensions of the type lower jaw are as follows: Length Py, 5.0; width P,, 2.9; length M,, 5.5; width Mj, 4.1; length M2, 5.7; width M,, 5.2; length M3, 6.5; width Ms, 4.0. Genus SPANOXYODON Simpson Spanoxyodon Simpson, 1935d, p. 236. Type.—Spanoxryodon latrunculus Simpson. Mstribution.—Middle Paleocene, Fort Union, Montana. Diagnosis.—P_2 absent and long diastema between canine and P3. P3_, much as in Chriacus, but Ps with metaconid larger, higher, and more distinct. M,-, about as in Chriacus or with trigonids slightly lower and paraconids slightly less internal. Remarks.—Only one specimen referable to this genus is yet known, and its characters are adequately summed up in the diagnosis. SPANOXYODON LATRUNCULUS Simpson Ficure 53 Spanoxyodon latrunculus Stimpson, 1935d, p. 236. Type.—U.S.N.M. no. 9287, left lower jaw with canine alveolus and P;-M;. Collected by Dr. J. W. Gidley. Horizon and locality.— Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis.—Sole known species of genus. Measurements of type as follows: Length P3, 3.8; width P3, 2.3; length Py, 5.0; width P,, 2.8; length M,, 5.2; width M,, 3.9; length Mz, 5.8; width Mg, 4.5. MIMOTRICENTES,*® new genus Type.—Tricentes latidens Gidley. Distribution.—Middle Paleocene, Fort Union, Montana. Diagnosis.—Adaptively and structurally closely similar to Tri- centes, but P, present, no marked diastema, and molar paraconids higher on crown and internal, not median, in position. Discussion.—In my preliminary paper I left the type of this genus in Tricentes, where Dr. Gidley had placed it. On further study, how- ever, I am forced to erect a new genus for it. The resemblance to Tricentes is close, but in the latter P, is apparently invariably absent (the name Tricentes refers to the presence of only three premolars, striking in this primitive fauna), and the paraconids of M2-; are quite different. The former might be only a primitive character and not surely of generic rank if unaccompanied by other differences. The character of the paraconid, however, as reduced as in Tricentes but in a different way, removes this form from the Tricentes lineage and may even mean that the relationship is not closer to that genus than 89 Miyos, an imitator-+ Tricentes. 204 BULLETIN 169, UNITED STATES NATIONAL MUSEUM to other primitive arctocyonids and that the resemblance is purely convergent, aside from the fact that both have the stereotyped pattern of all these ancient forms. The canine, preserved in the type, has a large root and is curved in a pronounced arc. The crown is unusually erect, as it is also in Tri- centes. P, is known only by its alveolus, which is single and rather large. The premolars very closely resemble those of Tricentes, except FIGURE 53.—Spanozyodon latrunculus Simpson, U.S.N.M. no. 9287, left lower jaw: a, Crown view; 8, inter- nal view. Twice natural size. Smads ( ada. uy pe uy We FIGURE 54.—Mimotricentes latidens (Gidley), U.S.N.M. no. 9269, left lower jaw: a, Crown view; 0, internal view. Twice natural size. that the heel of P, is more expanded transversely and its posterior border is more nearly a straight transverse line—a distinction prob- ably not generic, as it is closely similar to that between Metachriacus punitor and M. provocator. The molars are even flatter and broader than in Tricentes but otherwise closely resemble those of that genus except for the difference in the paraconids. Upper teeth are known only in one specimen, mentioned below. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 205 MIMOTRICENTES LATIDENS (Gidley) Figure 54 Tricentes latidens GipLEY, Simpson, 1935d, p. 236. Type.—U.S.N.M. no. 9269, left lower jaw with canine and P,-M3. Collected by Dr. J. W. Gidley. Horizon and locality Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis —Length: width ratio (type only): M, 1.09, M; 1.30. Measurements given in table 44. Paraconids vestigial and trigonids short. Discussion.—This species is known principally from two good specimens from the Gidley Quarry, the type and U.S.N.M. no. 9276, a right lower jaw with P,-M, (M, broken), confirming but not adding to knowledge derived from the type. There are two other specimens from the Gidley Quarry, probably of this species but of no value, and two, also very fragmentary, from the Silberling Quarry. One of the latter, U.S.N.M. no. 9672, with P, and M,, is smaller and less robust than the type, with the paraconid of M, more distinct, but it might be a variant of this species. The only known upper teeth of this genus, and probably but not surely this species, are M? from Loc. 81, in the American Museum collection. Like the lower molars, they closely resemble those of Tricentes, the only clear difference, and this of doubtful value, being that the internal cingulum does not circle the protocone and that on M? the external cingulum does not cross the paracone. TABLE 44.— Measurements (in mm) of the two principal Gidley Quarry specimens of Mimotricentes latidens P2 P3 Py Mi M2 M3 U.S. N. M. no. a a le eS E | L W L W L W L WwW L W L Ww 260E ese saan sess seek 3.6 2.8 4.6 3.3 5.4 4.0 655) 4.7 6.0 5.5 6.1 4.7 EIA Ohio a se ee BEET fees iiac eael| ESR | ate eo RRSIMl bas 7/6| ae oa MIMOTRICENTES ANGUSTIDENS," new species Figure 55 Type-—U.S.N.M. no. 9277, left lower jaw with P,-M:. Collected by A. C. Silberling. Horizon and locality —Type from Silberling Quarry. Referred speci- mens, not identified with complete certainty, from Locs. 51 and 73. Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. % Angustus, narrow+dens, tooth. 206 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Diagnosis.—Very slightly smaller than M. latidens. P, relatively smaller, slender, and with paraconid much stronger. Molars rela- tively narrower. Length: width ratio of M, (of type) 1.24. Para- conids less reduced and trigonids less compressed than in M. latidens, trigonids about equal to talonids in size. Discussion.—The difference in size is slight and may not be sig- nificant, but the differences in structure and proportions are marked and sharply distinguish this from M. latidens. The type has another distinction, not cited in the diagnosis, in the fact that the talonid of M, is definitely narrower than the trigonid, while in M. latidens it is of equal width or slightly wider than the trigonid. There are, however, two isolated M.,’s from Loc. 50 that closely resemble M. angustidens but that have the talonid in one equal to the trigonid and in the other slightly wider. At present it seems more probable that this is a variable character in the species than that these teeth rep- resent a third species. Princeton no. 13758 is a left lower jaw with M,_3, from Loc. 73, especially inter- esting because it is from the Fort Union S) Ne, LKS ay Wee iy ANS Wy, & a { \ na Ks Ze St Aton pw nr a FIGURE 55.—Mimotricentes angusti- dens, new genus and species, Prince- ton Univ. no. 13758, left lower jaw, a, Crown view; b, internal view. Twice No. 1, where fossils are very rare and be- cause it is, as far as I know, the only fossil mammal ever found in sandstone in this natural size. field. Its M, agrees very closely with the type of M. angustidens except that the talonid is as wide as the trigo- nid. M,; differs from that of M. latidens in the same way as does M3, being relatively narrower, with stronger paraconid and _ longer trigonid. The dimensions of the four specimens mentioned are given in table 45. TaBLE 45.—Individual measurements (in mm) of lower dentition of Mimotricentes augustidens Pa Mi M; Ms Specimen L Ww L WwW L Ww L w UESSNeMin on O27 72s. ase oo ee 4.4 3.0 4.8 3.6 5.6 4.5 WES Ne Meno $9706 See eee Se a SA os Sit ee | 5.6 4.6 UcS ANGIE OS O05 sss 2 Bee 8 IE Eee ee SUE AE | ee ae 5.6 4.5 FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 207 ?MIMOTRICENTES species U.S.N.M. no. 6178 is a left M, from Loc. 9, at a low level in the formation. It apparently belongs in this or a closely allied genus. Its dimensions, 6.2 by 5.2 mm, are perhaps not significantly different from those of M. latidens, although it is relatively somewhat narrower. The trigonid is distinctly longer than in the type of the genus and about equal to the talonid in size, not distinctly smaller as in the type. Although inadequate for identification, the occurrence merits mention in view of the horizon represented. Family MIACIDAE Cope, 1880 The known history of the Miacidae is anomalous and emphasizes the inadequacy of some of our knowledge of details in this early epoch and the probably erroneous character of some negative conclusions regarding it. The miacids (so carefully and fully defined by Matthew in many works that diagnosis here is unnecessary) are a specialized group, for in them the carnassial shearing teeth are very well devel- oped, despite their absence in all other known Lower and Middle Paleocene mammals. Furthermore, they are an adaptive and potent group, for their carnassials are M, and P*, as in the Carnivora (vera) or Fissipedia and there is every reason to believe that they are, in a broad sense, ancestral to all the latter. The appearance of this ap- parently modernized group in a fauna otherwise almost wholly archaic is extraordinary. The known distribution within the family is also noteworthy. The first genera to appear, and the only ones known before the true Hocene, are not the most primitive and generalized and are not ancestral, even structurally, to the majority of later types. All have lost M%, unquestionably present in the ancestry. Even aside from the fact that these teeth are present in most later miacids, they are almost universally present in Middle Paleocene mammals of other families. This anomalous history must involve, first, rapid progressive evolu- tion of the group generally, the Miacidae, which is not surprising in view of later history, which shows this general type to be probably the most plastic and adaptive of allmammals. Second,itmustinvolve the early, minor differentiation of a special line, the Viverravinae, which entered the regions known to us paleontologically at about the beginning of the Middle Paleocene, while the more varied adaptive Miacinae were confined, until the great Eocene invasion, to some facies or region still unknown to us. In speaking of the Miacidae as specialized, it is important to em- phasize the relative value of the words. They are specialized in com- parison with the extraordinarily archaic contemporaneous Arctocyo- nidae, which are not far from being generalized primitive placentals, but in comparison with the other carnivores, specifically with the 208 BULLETIN 169, UNITED STATES NATIONAL MUSEUM fissipedes, they are extremely primitive, much more so than any Tertiary dogs or other true fissipedes. The anomaly is thus not so striking as might appear at first sight. Furthermore, within the family the Paleocene forms are distinctly more primitive than Eocene Viver- ravinae and are certainly not ancestral to the Miacinae, so that com- parison with the latter is misleading. On a small scale, within the family, the Viverravinae are a miniature “archaic” radiation of Miacidae and the Miacinae a later “progressive” radiation, much as, on a far grander scale, the peculiarly specialized periptychids are an archaic radiation and the basically more primitive hyracotheres are a progressive radiation among the ungulate cohort. There are two distinctive genera of miacids in this fauna, Didymictis, evidently an abundant form with several species and long known from the distant Torrejon (as well as from numerous later horizons), and Ictidopappus, a rarer type known from only two specimens in this fauna and as yet unknown elsewhere. Subfamily VIVERRAVINAE Matthew, 1909 (Viverravidae Wortman and Matthew, 1899) Viverravus, Didymictis, and Ictidopappus evidently form a closely related group characterized, among other features, by the prominent anteroexternal cuspule of P*, the elongate oval outline of M?, and the absence of M’; (Matthew, 1915). For this group the name Viverra- vinae is available, contrasting with the typical miacids, the Miacinae. As mentioned below, Didymictis may be a compound genus, but if so its components are very closely related. Viverravus, also, is a some- what doubtfully bounded genus. Its earlier, lower Eocene species are very close to Didymictis, while some of its later, middle Hocene species, perhaps including the genotype, are so markedly advanced over the early forms that they might not ordinarily be placed in the same genus. This point is not here apropos, but Viverravus is of some present interest because of the possibility of special relationship to Ictidopappus. Viverravus and Didymictis were separated by Matthew (1915) on the basis of the crested heels of M,_, in the former, basined in the latter. In fact the early species of Viverravus (e. g., V. acutus, V. politus, and even the slightly later V. dawkinsianus) do have basined talonids, although they are obviously becoming crested by emphasis of the hypoconid and marked reduction of the entoconid. In Jcti- dopappus the talonids are quite as basined as in Didymictis, a condition doubtless ancestral for Viverravus also but partly or completely lost in species definitely referable to that genus. Aside from the difference in the heel, relatively slight at the beginning of the Eocene, the early species of Viverravus also differ from Didymictis in the longer, lower trigonid of M,, which is a striking resemblance to Ictidopappus. 209 FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. Were it not for the unusual proportions and simple structure of Py, Ictidopappus would make an ideal ancestor for Viverravus. Some species of the latter, e. g.,V. acutus, have P, small and simple, but much more trenchant than in Ictidopappus, strongly compressed laterally, the posterior cusps well developed, and all cusps in a straight line, not subtriangular. The Ictidopappus P, might be ancestral to this, but there is no good evidence that it was, and Viverravus may after all be an offshoot of Didymictis in which the trenchant P, was already highly developed in the Middle Paleocene. Genus DIDYMICTIS Cope, 1875 This exceptionally long-lived genus is recorded from all levels from the Torrejon to the end of the lower Eocene. During this period there is not only definite evolutionary advance but also evidence of the presence of several different phyla. It is quite possible that one or more of these can be and should be separated generically, but the criteria for doing so are poor at present. A useful step was Matthew’s (especially 1915) revalidation of Viverravus, which was long confused with Didymictis or considered a synonym. Now Matthew (Pale. Mem.) has further separated the sole Torrejon species, D. hay- denianus, and placed it in a new subgenus, Didymictis (Protictis), with the suggestion that this may prove to be of generic rank. The most abundant species in the present collection, D. microlestes, introduces some difficulty in this arrangement. In its more important morphological characters it compares with the types of Didymictis (Protictis) and of Didymictis (Didymictis) as shown in table 46. TaBLE 46.—Comparison of dentition characters of three species of Didymictis Species P3 Protocone of M! Proportions of M!-4 D. (Protictis) hay- | With rudimentary proto- | High, angulate, posterior | Strongly transverse, M? denianus. cone. wing weak. reduced. D. microlestes_-__-- Intermoediates=--22 5223-25 AsineD iprotenussss-cs----—- Asin D. protenus. D. (Didymictis) | Compressed, no proto- | Broader, wings equal-_--_---- Less transverse, M? less re- protenus. cone. duced. SSS ————EoEEESESESESESESEaEaESESaSaSpapaSSSEESESESESESESESESESESESESESESESESESSSS=S=S=SS=—=" Species P3 P41 Ma D. (Protictis) hay- denianus. With very minute anter- ior cuspule, heel crested without distinct cusp. With relatively large an- terior cusp, heel with small basin and crest with rudimentary cusp. Witb minute anterior and large posterior accessory cusp, no basin. D. microlestes_---- D. (Didymictis) protenus. With distinct, nonshearing anterior cusp, 2 talonid cuspules, small basin. With anterior cuspule large and shearing, 1 talonid cuspule, small basin. With trigonid somewhat elevated and shearing, heel reduced. Trigonid comparable to D. haydenianus, heel less re- duced. With anterior cuspule very | With trigonid low, tuber- small, nonshearing, 1 tas lonid cuspule, basin very poorly developed or ab- sent. cular, heel not reduced. 210 BULLETIN 169, UNITED STATES NATIONAL MUSEUM The upper teeth, although much smaller and lighter in construction, resemble D. protenus rather than D. haydenianus. These are probably primitive characters, and D. haydenianus is slightly aberrant. The development of P3_, is also somewhat closer to D. protenus, and it is not clear that this is wholly in primitive characters. The presence of only one talonid cusp on P, may be, but it is unlikely that the more cuspidate P3 is primitive. The development of relatively large and shearing anterior cuspules on these teeth is different from either D. haydenianus or D. protenus and seems to be a specialization. M; is perhaps merely primitive with respect to either of the other two species. It is possible that D. microlestes stands nearer the D. protenus ancestry than does D. haydenianus, in which case the subgeneric separation might well be maintained, and D. microlestes would belong to the typical subgenus. The evidence for this is not very good, however, and for the present it seems best not to attempt a subdivision of the genus. There is another Didymictis-like species in the fauna, D. tenuis, of very diminutive size. At present it is known from a single specimen, and, as noted below, there is some doubt as to the charcters of Py. If they were confirmed the species could hardly be placed in Didy- mictis, but knowledge is now so imperfect that it seems best to leave it here until further evidence is at hand. Table 47 gives an idea of the very considerable differences in the length of M,; among the three Middle Paleocene species referred to Didymictis: TaBuie 47.—Length of Mé, in three species of Didymictis Species N R M D;-haydentanus—Torrejon25. 22-28. 2225. 222 6 7. 5-8. 0 7. 72 LD NECT OLCSLCR ores Ss Oe StS ee he 8 4, 3-5. 1 4, 66 Dissteniient aes oS 2 oe oe et SN oe AGREE Ae oe 2.9 There are also a few specimens too imperfect for exact identification that are comparable in size to D. haydenianus and cannot at present be distinguished from that species. DIDYMICTIS MICROLESTES Simpson Figures 56, 57 Didymictis microlestes Stmpson, 1935d, p. 238. Type.—U.S.N.M. no. 9301, left lower jaw with P,-M>. Collected by Dr. J. W. Gidley. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 211 Horizon and locality —Gidley Quarry (one referred specimen from Silberling Quarry), Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis —Much smaller than D. haydenianus or any known later species (see measurements below). P3 similar to P,, but cuspules less well developed. Py, with large anterior cuspule, developed into a small shearing blade, talonid relatively broad and less rounded than in‘other species, with one main cusp. M, with slightly elevated and shearing trigonid, talonid reduced. < i rl Peele = A ry ee FIGURE 56.—Didymictis microlestes Simpson, U.S.N.M. no. 9301, with parts in outline supplied from U.S.N.M. nos. 6146 and 9306, left lower jaw: a, Crown view; 6, internal view. Three times natural size. FIGURE 57.—Didymictis microlestes Simpson, U.S.N.M. no. 9299, with part in outline supplied from no. 6147, left P’-M?4, crown view. As preserved, P? is reversed (or rotated 180°) from the position shown in the drawing, but this is believed to be accidental. Three times natural size. Discussion.—The morphological characters of this elegant little species have been fully brought out in the comparison and diagnosis above. The principal available numerical data on the lower dentition are given in table 48. TaBLE 48.—Numerical data on lower dentition of Didymictis microlestes Variate N R M o V Tig seest ON aE 7 4.1-5.0 | 4.49+0. 11 0. 28 +0. 07 6.2+1.7 Wires) sk i 2.0-2.5 | 2.17+0. 06 0.17 +0. 04 7. 742.0 IL ee ee Te 8 4, 3-5.1 | 4. 66+0. 09 0.26 +0. 06 5.641. 4 WIMg eee 2 8 2. 9-3. 2 | 3. 0O8+0. 03 0. 097+0. 024 3.1+0.8 1G Foe ee 4 3. 4-3. 8 | 3. 65 [= (d2) =0. 0900] Wie 2 4 PFI By | PA PAD [=(d?) =0. 0100] 212 BULLETIN 169, UNITED STATES NATIONAL MUSEUM The type happens to be an unusually small specimen. It measures: Length P,, 4.2; width P,, 2.1; length My, 4.3; width M;,, 3.2; length Ma, 3.4; width Mp, Didi Only two upper jaws and an isolated carnassial are at hand, measure- ments of which are given in table 49. TaBLeE 49.—Individual measurements (in mm) of upper dentition of Didymictis microlestes ps M! M2 U.S.N.M. no. L W L Ww L Ww ODO O mc, ee Nae a een oe 3. 9 4.2 5. 4 2. 4 4.0 G3 Rae uteri attic yas 4.9 ‘atte! 4.3 Sell 2.5 3. 9 GA ip eek eke 2 Laie ol oual 4,1 FIGuRE 58.— Didymictis tenwis Simpson, U.S.N.M. no. 9297, left lower jaw: a, Crown view; 6, internal view.. P, is broken, and it may not be correctly placed. Six times natural size. DIDYMICTIS TENUIS Simpson Fiaure 58 Didymictis tenuis Simpson, 1935d, p. 238. Type.—U.S.N.M. no. 9297, part of left lower jaw with M, and broken ?P;. Collected by Dr. J. W. Gidley. Horizon and locality.— Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis —Much smaller than any known comparable miacid. M, with very elevated trigonid, hypoconid and entoconid about equally high and distinct. M, length 2.9, width 1.8. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 213 Remarks.—As the specimen is preserved there is a small tooth anterior to M, that has been cemented to the specimen without any clear or certain contact. Knowing the care with which Dr. Gidley worked, I have no doubt that this tooth was found with the specimen, but not knowing his exact evidence of association I am not certain that it is in fact P, of this individual. It is, furthermore, somewhat incomplete. It appears to be a very small and simple tooth with a short, high, conical main cusp and a single conical posterior cusp. If these are its true characters, it is very much unlike P, in any species certainly referred to Didymictis, but this is too uncertain to draw any conclusion, and the lower carnassial is sufficiently Didymictis-like to leave the species in that genus at least until better material is found. DIDYMICTIS HAYDENIANUS Cope, 1882 Fiaure 59 U.S.N.M. nos. 6143 and 6145, each including an upper P*, represent a species inseparable from Didymictis haydenianus. Their most reliable dimension, the (oblique) length of the straight shearing edge, is 10.7 and 9.7 mm, respectively. In Torrejon specimens referable to D. haydenianus this dimension is 9.2-11.3 mm. No constant morpho- logical difference is seen. The ma- terial is inadequate to establish definitely that the Fort Union form is exactly the same as that from the a bd,» a me eT Pes POPE op Lae acl ear separable. no. oe left upper jaw with carnassial, crown Be map ican Museum eollee- (hak on tee ence ae tion there is a specimen from Loe. same, anteroexternal view; 6’, same, postero 81 ae broken PY anduNitiwihe internal ie a ae ae ere cross- oblique length of the shearing crest cannot be measured exactly, but it was about 11.2 mm near the known upper limit for D. haydenianus. The specimen is more robust than the two mentioned above and might be a large variant of the same form or a different subspecies or species. There is also a fragment of a Py, including the heel, U.S.N.M. no. 9930, from Loc. 51, that has the size and cusp structure of D. hay- denianus, quite unlike D. microlestes. Genus ICTIDOPAPPUS Simpson Ictidopappus Stimpson, 1935d, p. 2387. Type.—Ictidopappus mustelinus Simpson. Distribution.—Middle Paleocene, Fort Union, Montana. 214 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Diagnosis.—Differing from Didymictis in the relatively smaller and much simpler P3;-, and relatively lower and longer trigonid of M,, from Viverravus in the wider and more triangular P,; and more defi- nitely basined talonids, and from other known miacids in the absence of M3. FIGURE 60.—Ictidopappus mustelinus Simpson, U.S.N.M. no. 9296, right lower jaw: a, Crown view; 3, in- ternal view. ‘Three times natural size. Figure 61.—Ictidopappus mustelinus Simpson, U.S.N.M. no. 9295, left upper jaw, in two fragments but associated, crown view. ‘Three times natural size, ICTIDOPAPPUS MUSTELINUS Simpson Fiaures 60, 61 Ictidopappus mustelinus Stmpson, 1935d, p. 237. Type.-—U.S.N.M. no. 9296, right lower jaw with P;—-M, and talonid M2. Collected by A. C. Silberling. Horizon and locality—Gidley Quarry, Fort Union, Middle Paleo- cene horizon, Crazy Mountain Field, Mont. Diagnosis.—Sole known species of genus. Measurements below. Discussion.—From the alveoli the canine was a large procumbent tooth and was followed by a diastema. P, may have been absent and P, 1-rooted, but this is not certain. P;is a very small and simple tooth consisting of a somewhat compressed main cusp followed by a slight unbasined heel. P, is shorter than M, but nearly as high and is sub- FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 215 triangular. There is a small anterointernal cusp, resembling a rudi- mentary paraconid rather than the anterior basal cuspule of Didymictis, and there is a very slight and uncertain indication of a rudimentary basal metaconid. The talonid is very short and wide and vaguely cusped. There are no other cuspules or cingula. M, has a large but, in comparison with Didymictis, low trigonid elongate anteroposteriorly. The talonid is very small, with distinct but not prominent hypoconid, hypoconulid, and entoconid developed on its raised rim. Its well-developed basin opens internally, between entoconid and metaconid base. My, was evidently much reduced. The talonid is small but is elongate and basined, with the three cusps poorly differentiated. The specimen is broken immediately posterior to this, but from the shape of the talonid of M, it cannot have been followed by another tooth. Dimensions are as follows: Length P3, 2.0; width P;, 1.4; length P,, 2.9; width P,, 1.9; length M,, 3.8; width Wi 200 There is a specimen, U.S.N.M. no. 9295, a left maxilla with P!, P°- M?, and the alveolus of the canine, that is probably the upper jaw of Ictidopappus mustelinus. Its size is exactly right for occlusion with the type, it has P? much smaller than in Didymictis, harmonious with the smaller P, of the type, and it has M'-? markedly shorter than in Didymictis, harmonious with the shorter M, and heel of M, of the type. The only feature suggesting distinction is the embrasure between P* and M!, the outer angle of which is much more acute than the outer angle of the trigonid of M,, which fits into it. This, however, does not prevent normal occlusion, as it might at first sight appear todo. In Didymictis, also, the embrasure angle is more acute than the occluding trigonid angle. Occlusion is not strictly orthal but is oblique, in part ectal (in fact nearly analogous to the triconodont occlusion but, unlike triconodonts, interlocking). At the close of the bite the external trigonid angle is internal to the embrasure angle, and the trigonid fits into the wider internal part of the embrasure and does not coincide with it. It cannot be proved that this upper jaw is of Ictidopappus, but it is highly probable, and it should not be assumed to be distinct. The generic distinction of the upper jaw is very marked. The canine was relatively large, its alveolar wall swollen, much as in the most advanced and quite unlike the primitive species of Didymictis. P! is present and a small simple tooth probably with two roots but with alveolar mouths confluent. (P? is represented by one root, the specimen being broken here.) P* is much smaller than in Didymictis. Its ectoloph is similar in form. The pronounced inner spur, worn but probably not cuspidate, is median and has a separate and strongly divergent root. The ectoloph of P* is also similar to Didymictis but has a more decided notch in the external contour and is without a cingulum. The protocone spur is slenderer, perhaps less definitely 1192123715 216 BULLETIN 169, UNITED STATES NATIONAL MUSEUM cusped (worn) and projects almost directly internally, not at all forward, in marked distinction from Didymictis. M! have the same cusp structure as in all miacids, but are highly peculiar in proportions, being very short and wide, markedly tri- angular, not at all rounded, with sharply emarginate external borders. These characters are more nearly approached by Viverravus than by Didymictis but sharply distinguish Ictidopappus from any other miacid, Measurements of this specimen are as follows: Length P’, 2.6; length P*, 4.8; width P*, 3.7; length M!', 3.2; width M}, 5.2; length M?, 2.0; width M?, 3.7. Family MESONYCHIDAE Cope, 1875 This family, so widespread elsewhere in the Middle Paleocene and on into the Eocene, is represented in the present fauna only by two broken teeth. Genus DISSACUS Cope, 1881 DISSACUS, species undetermined U.S.N.M. no. 9692 from the Gidley Quarry is the base of a tooth with the heel preserved. It agrees in size with M, of Dissacus nava- jgovius, of the Torrejon, but differs in having a very vestigial basin internal to the heel crest. Another specimen, from the Gidley Quarry, is perhaps M,, lacking the anterior portion. The metaconid is very small and is relatively anterior, as in some species of Dissacus, but the heel is relatively shorter than in other specimens of that genus known tome. It is clear that these fragments are mesonychid, and there is nothing to distinguish them certainly from Dissacus, but they are not really identifiable. Order CONDYLARTHRA Cope, 1881 As with most of the major groups here considered, Matthew has thoroughly reviewed the taxonomic history of the order Condylarthra. Since, however, I propose a marked change in the current arrange- ment of the order, an outline of this history must now be given. Cope proposed the name Condylarthra (then supposed to be a suborder of Perissodactyla) in 1881, basing it on the Phenacodontidae and princi- pally on Phenacodus but with Periptychus (‘“‘Catathlaeus’’), Mioclaenus, and Tetraclaenodon (‘‘Protogonia’”’) probably and Anisonchus possibly included. Subsequently (1882-1884) the family Periptychidae was distinguished, but retained in the Condylarthra, the family Menisco- theriidae was added, and Mioclaenus was excluded, being considered first an artiodactyl, then a creodont. On the latter point, Scott (1892) showed that many species placed in Mioclaenus by Cope were FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 217 generically distinct and were indeed creodonts, while Afioclaenus, sensu stricto, might be a condylarth, in which case it would form a very distinct family. In 1895 Osborn and Earle followed Scott’s suggestion and defined the Mioclaenidae as a family of condylarths. They retained the three condylarth families of Cope, Periptychidae, Phenacodontidae, and Meniscotheriidae. They divided the Periptychidae into Anison- chinae and Periptychinae and stressed the resemblance of the former to the Mioclaenidae. They also noted that the Periptychidae had resemblance to the Amblypoda and suggested the possibility of their belonging there, but left them in the Condylarthra. In one of his last papers, in 1897, Cope adopted this suggestion and associated the Periptychidae with the Pantolambdidae in the division Taligrada of the Amblypoda. In the same year, but with Cope’s paper before him, Matthew rejected this transfer and adduced new evidence and reasons for retaining the Condylarthra as a broad group including Periptychidae, Phenacodontidae, Mioclaenidae, and Meniscotheriidae. In 1898 Osborn adopted Cope’s arrangement of the Amblypoda, in- cluding the Periptychidae in the Taligrada and hence excluding it from the Condylarthra.*! He seems subsequently to have adhered constantly to the conception of the Condylarthra as including only the Phenacodontidae, Mioclaenidae, and Meniscotheriidae, and this authoritative view has since been the most widespread. Matthew continued for a time to include the Periptychidae but after about 1914 agreed with Cope and Osborn in placing that group in the Amblypoda. The relationship of the hyopsodonts to this order was not estab- lished until relatively recently, and even now the conclusive evidence for it does not appear to be widely known. From the time of its dis- covery by Leidy in 1870 until 1963 Hyopsodus was universally con- sidered to be allied to Notharctus, Pelycodus, or similar genera. It was therefore generally considered to be a primate, occasionally an insectivore, but in these instances largely on the evidence of supposed allies, which are in fact primates. In 1903 Wortman definitely dis- tinguished Hyopsodus from the early lemuroids and referred it, in the family Hyopsodontidae (defined but incorrectly delimited by Schlos- ser in 1887 and recognized under an invalid name by Marsh in 1875), to the Insectivora, on its own characters, not those of lemuroid sup- posed allies. In 1909 Matthew thoroughly reviewed the Bridger hyopsodontids, pointed out their resemblance to the mioclaenids, and suggested that the two families might eventually prove to be synony- mous. He remarked that they lack diagnostic insectivore specializa- tions and expressed belief that they are closer to the Condylarthra than to the more typical Insectivora. Nevertheless, he then placed %1 He inadvertently cites Osborn and Earle (1895) as Osborn (1892) and gives the impression that the transfer of the periptychids to the Amblypoda was then proposed. 218 BULLETIN 169, UNITED STATES NATIONAL MUSEUM them in the Insectivora because of their primitive character and because of certain marked differences from the phenacodonts. In 1915 Matthew reviewed the lower Eocene hyopsodonts, which, for the first time, gave him a good knowledge of their foot structure. He then considered them to be condylarths, confirming his tentative suggestion of 1909. He carefully redefined the Condylarthra and included these five families, the last provisionally: Mioclaenidae, Hyopsodontidae, Phenacodontidae, Meniscotheriidae, and ?Pleura- spidotheridae. In his last contribution (Pale. Mem.) Matthew retained this ar- rangement, except that the Mioclaenidae and Hyopsodontidae are reduced by further study to two subfamilies of Hyopsodontidae, as already tentatively foreseen in 1909. I now propose to return to Cope’s arrangement of 1884, with the only change the inclusion of the Hyopsodontidae (with Mioclaeninae), that is, to his classical conception of a group based on both the phen- acodonts and the periptychids. It seems to me, after careful and long consideration with practically all the pertinent original specimens (including a good deal even unknown to Matthew), that comprehen- sion of this group has been retarded and taxonomy has been in a blind alley since the rise in the nineties of the idea of close periptychid- pantolambdid affinities, an idea to which even Matthew finally sub- scribed after some years of resistance. This reactionary view, which at this late date will rather seem radical, requires an outline defense even though much of the crucial evidence is not drawn from the pres- ent fauna. The original suggestion that Periptychus might be an amblypod (Osborn and Earle, 1895, p. 47) was based on the facts that its tarsus is not serial and that “it has the strictly trigonal molar of the Ambly- poda.”’ It may at once be noted that these arguments have since proved to be valueless. It is now known that the primitive condy- larth tarsus was not serial, and the molars of Periptychidae are not, as a rule, strictly trigonal, those of some condylarths are, and the periptychid molars are otherwise decidedly more condylarth- than amblypod-like. Cope (1897, p. 335) stated that he had anticipated that the perip- tychids, with their astragalo-cuboid contact (nonserial tarsus), might be the bunodont ancestors of the Amblypoda, but he awaited dis- covery of their carpus and evidence that it, also, was nonserial. The carpus had not been discovered, but the continuing failure to discover any other possible amblypod ancestry led Cope then to assume the presence of a nonserial carpus in the periptychids and to consider them as this ancestry. He was also influenced by the suggestion of Osborn and Earle. This is too theoretical to warrant much con- FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 219 sideration. The periptychid carpus is, in fact, “alternating”, but so is that primitive for and apparently fairly common among (other) condylarths. Osborn (1898, pp. 177-179, 184-186) gives a long list of taligrade characters as defining that group and common to pantolambdids and periptychids. The great majority of these are, as he states, primitive characters. As far as confirmed among the so-called Taligrada by later research, they are also found to occur among or to be equally typical of Condylarthra and hence have no bearing on the particular question here considered. The only progressive taligrade character given is “molars triangular (tritubercular), selenodont”, which is decidedly untrue of the Periptychidae and opposed to Osborn’s thesis. Indeed, I cannot see that Osborn then advanced any actually valid evidence in favor of the conclusion given, which has since become taxonomic and phylogenetic dogma, largely on his authority. Matthew (1897) had already shown that even in the supposedly typical condylarths, the phenacodonts, the early forms have alter- nating, not serial, carpus and tarsus and that the Condylarthra there- fore could not be defined and were not characterized by the mooted primitive serial carpus and tarsus as had previously been supposed. He therefore found no difficulty in retaining the Periptychidae in the Condylarthra and gave a lucid and valid argument for doing so, even though, as I now think, he minimized his evidence by much over- stressing the resemblance of the periptychids to the pantolambdids in limb structure and their difference from the phenacodonts and mio- claenids in dental pattern. Upon transferring the Hyopsodontidae to the Condylarthra, Mat- thew (1915b, p. 311) gave a long diagnosis of the Condylarthra, in- volving the whole bodily structure. His intention at the time was to exclude the Periptychidae, since he did so in earlier and later general classifications, although this point was not then specifically mentioned, since it was foreign to the fauna he was revising. It is therefore re- markable and significant that his definition of the Condylarthra clearly excludes the Pantolambdidae but applies exactly to, and hence includes, the Periptychidae with a single exception: ‘‘tarsals serial.”” This one point was, in fact, an error or lapsus, for the forms he explicitly meant to include do not have strictly serial tarsals, and in some the approach to the periptychids in this respect is very close. Matthew’s subsequent defense of the collocation of Periptychidae and Pantolambdidae was based almost entirely on the limbs, especially on the astragalus. When evidence drawn from the dentition was at variance with that drawn from the limbs or astragalus, he almost in- variably followed the latter. Without quarreling with this principle of research, it will appear below that the evidence is not necessarily at variance in the present instance. The dental evidence certainly 220 BULLETIN 169, UNITED STATES NATIONAL MUSEUM favors condylarth rather than amblypod affinities for the periptychids, and the limbs might perhaps support either view equally well and cer- tainly do not oppose condylarth affinities. The recognition of the affinities of Coriphagus, as discussed under that genus, adds to the evidence for the opinion supported here. It has been generally recognized that the teeth of the periptychids could not give rise to those of pantolambdids or coryphodonts. They are in fact aberrant and developing along a line, or series of lines, of their own. It has sometimes been recognized, and can readily be shown, that their greatest resemblance is with the Condylarthra and that they could all be immediately derived from types well known in that order. Indeed, they intergrade with certain mioclaenids to such a point that the families are difficult to distinguish on this basis. The pantolambdid-coryphodont dentitions, on the other hand, are widely different. Union of periptychids and pantolambdids has, then, rested entirely on limb, and especially on foot structure. Indeed, without slighting the fact that other resemblances occur, it has depended more on the astragalus than on any other point. If this arrangement is a natural one, it seems necessarily to imply that the ‘‘taligrade”’ astraga- lus and limb structure arose in a stock with extremely primitive teeth and that the widely divergent periptychid and pantolambdid denti- tions developed later. Such a thesis seems a priori rather improbable, but certainly it is not impossible. The apparently, but I think falsely, analogous case of the divergence of, say, suid and camelid dentitions after the artiodactyl foot structure arose suffices to demonstrate that such a history is conceivable. The analogy is probably false and the thesis indefensible because in the case of the artiodactyls the teeth, followed back in time, dis- tinctly converge and are rather plainly derivable from a common type possessed by animals that already had all the essential artiodacty] limb characters. This is not true of the periptychids and pantolambdids. Even within the limits of the Periptychidae there are forms with hardly any suggestion of the ‘“‘taligrade” foot, but with teeth much too dis- tinctly periptychid to give rise to the pantolambdids. A common ancestor, if it existed at all, can hardly have had taligrade feet but must almost certainly have been a condylarth and a very primitive condylarth. A review (table 50) of typical astragali of the groups conceived will make the situation clearer. These genera are all of about the same age. The Hyopsodontidae are represented by isolated Gidley Quarry specimens surely of this group but not exactly determinable, as there are several hyopsodontids of about this size in the quarry. All five genera have numerous other characters of the astragalus in which they are closely similar. These are nondifferentiated primitive protoungulate characters, most of them disappearing in more advanced forms. 221 FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. ‘os IV'T *“SUIOJ surpedeid Ul UvYy} JOAIVT Ai qeqoid ‘yoursip ynq |[eurg ‘esplyuOposdoA HY UI sv ynoqy ‘Iodsie] Apqsys Apueiwddy *[[eus A190 A. *x9AT09 SST puB [81SIp Sse] ‘pessoidep ALWYSITS *pesseidurod ssey yng ‘wop -0WaD}IDAJAL 0} Ie[UIS ATESOTD ---"wopouanjovi7aL ULSse ynoqy *xeAU0D AlOSIOASUBIL oyVsuU0[9 BdBJINS AB[NOIY -18 ‘peyuv[dosiop pesseidu0g *po10lIyst0o Jou ‘yues -qv Ay[voyovid ‘y10qs A109 A "~~ snanjypwazy Ul SB yNogy *paqotdys ~~} -u0d ssey ‘1ey10ys AYSITS ~-wopouanjonsjaz Ul SB yNoqy -peyolysuoo ‘enbiyqo ‘su0'7 “yooey prloqnod YIM Jeng -09 ‘podieljue A[qeviepIsuog *snyohidisagq 10 eeplyuoposdods Wf UIse ynoqy |-~~~--"-777>- DpquinjojUung ‘popuvdxe pus poueyiey 4 el0m 8 AAAS ‘uwopouanpjn.1ja,, Ul Se ynogy |-eepryuoposdoAH urseynoqy |---~--- Sa ie snyohidilag Rigs SepaaeeR Ra Op-~~~~|"-uopowanjon.1ja,, UL SB yNOqY |~---------7""-snanpynwazy *poeuluiiejepun ““uwopouan7jIn17a,[, UL Se ynogy |-~-peonpoad sseq ynq ‘re[IUAIIg *Po7eT -Ost ‘xoAu0d ATJUOS ‘[BAO snues ‘pryaoposdosy *[eu10}.x0 daeys qeory ploqno *yoou 03710 UMOp ‘Bo[qO0I} 0} [BISIP SUIPUGIXY od ‘od ‘wopouanpoD4jaZ, Ul SB ynoqy *BO[T001} 04} SB ATI[EISIP IvJ OS ZUIPUI}xX9 JON 4oov} IB[OO][VUL [eUINJUT “yep 4SOM][B ‘IOMOT[VYS ATPoyIVIAL "IOMOT[VYS {1} YSIS 10 uopouanjIvija,, Ul sve ynoqy ~-uopouanjID17a, UL SB UOGW ae ae Ss ~MOT[BYS AOA aan cnaenenenn- en aan-- MOTTBYS eal oo, *peyea -9]9 SSOT [[TS ING snydAjdisag JO OBUINOBLIOI| UL sv ynoqy ~-eepryuoposdoA UI sv ynoqy “--"wopowanpjovJay, UL SB INogy *peyBAde Ssof ‘uopouanov41jay, Ul sv ynoqy ‘pepunod ynq 4ours!p [eus9}Ul ‘davys 4soro [eus09yx a] $1S019 4oouy rBpnoeyue}sng *(19Z1B[ WOUIBIO}) uopowanjId17a,, Ul se Inogy “OBpl} -uoposdoAH Ul sB ynoqe UOI}VUIsIBMIe PUB UBUIBIOT ~-wopouan7In1ja, UL SB Noqy “pls [BU19}x9 UO OOM ‘M014 -euIdIzWe peoUNOUOId uy “BOY O01} JO pue [eurxoid 48 uOl)eUIs -IVU19 [VI pet psounouoId Uy UOUIBIO] Ie[esBIISV ‘ese, PUB [eUAOJUL [[BuUIg |~--~"------ UOpOUId)I0.L79,L, sosseooid reyur[g snuey *JOpBoIq APYSTIS 10 [euolsuoutpmnby |-~----- 77-7 -- Dpquinj0ojuDy pap ee [vuolsuemIpinbe ynoqy |-~------ ~~~ ~~ snyoAidi4ag Beas Tal yqoys pu peoiq AoA |----->7- ">>> ~snappypwmazy *poulmiejep aero [eUOIsuSTAIpInbe ynoqgy | -un snues ‘pryuoposdoA --q1OYS va[qo0I} ‘pRoiq Apog |--------77- wopowanjo017a,.L eo[qo014 pue Apoq jo suons0odoig snuey quopojund auaooajvg VD pun vuy.Uvpipuoy auaooappg fo puauab unof ur yoboujsy yoordhy fo mamay—0g¢ AIAV, 222 BULLETIN 169, UNITED STATES NATIONAL MUSEUM All five groups are basically similar in the astragalus. The early Hyopsodontidae probably are most primitive in this part, and surely the characters that they share with Tetraclaenodon must be taken as primitive. From this point of view, Pantolambda is much the most divergent. Tetraclaenodon and the early hyopsodontids resemble each other very closely, almost the only differences being the greater elevation of the crests and excavation of the trochlea and probably the reduction of the cuboid contact in Tetraclaenodon Both these characters may be incipient specializations, and both are much em- phasized in later phenacodonts and not in later members of other groups. The longer, or relatively narrower, body in the hyopsodontids is probably of slight significance. Hemithlaeus is very close both to Teiraclaenodon and the hyopsodontids. Its slightly shorter neck, almost its only peculiarity with respect to the more primitive con- dylarths, can hardly be supposed to make this a ‘“‘taligrade” astragalus, especially as the shortness is only relative and the neck is, in fact, well developed and typically constricted. The same statement applies to Periptychus, the neck of the astragalus being about the same in that genus and definitely more condylarthran than ‘‘taligrade” in character. All the other characters of the astragalus are condylarth- ran except that the cuboid facet is about intermediate between the most primitive known condylarth and amblypod conditions. Pantolambda has a much more primitive astragalus than Coryphodon, yet the table clearly shows that it diverges farther from the primitive condylarthran condition than does Periptychus. This divergence consists chiefly of the appearance in rudimentary form of characters greatly emphasized in Coryphodon. Despite the fact that he himself abandoned it, Matthew’s argument of 1897 in favor of considering Periptychus as a condylarth and Pantolambda as an “amblypod” seems to be as valid now as when he wrote it, indeed more so, for he was not then fully aware of the distinctions between these two genera now brought out. Periptychus does, of course, make some approach toward the so- called amblypods in limb structure, but this is far from reaching identity, and, being only vaguely or not at all seen in smaller con- temporaneous allies of Periptychus, may indeed be only convergent and largely conditioned by size and mode of locomotion. Conver- gence is the more likely in such forms that have not in any case come far from a purely primitive type of ungulate limb structure. Similarly Pantolambda is much more primitive than Coryphodon or other, later allies in limb structure, but it shows the beginning of the so-called amblypod type, and the approach is as much toward all or any primi- tive ungulates as specifically toward Periptychus and its allies. Pat- terson (1934) has also pointed out that the limb structure of Bary- lambda tends to link Pantolambda with the coryphodonts. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 223 The later history of the astragalus and feet in general in these groups is not entirely pertinent but may be mentioned. In the phenacodonts, culminating in Phenacodus itself, the limbs became considerably specialized in an inadaptive cursorial direction. Side toes were moderately reduced, the limbs became or remained moder- ately slender, and the feet digitigrade, carpus and tarsus serial. The astragalus differs markedly even from that of the closely allied but earlier Tetraclaenodon. 'The trochlea becomes very long, the foramen and emargination are lost, the crests both become high and sharp, and the head becomes more spherical and loses contact with the cuboid.” The hyopsodonts were remarkably conservative. As far as we know them, the limbs of Hyopsodus differed extremely little from those of its long antecedent Middle Paleocene relatives. An astragalus of Hyopsodus from the Eocene is almost identical with that of a Paleocene hyopsodontid here described except for the quite unimportant details of having the body somewhat less elongate and the head slightly more spherical. The Anisonchinae and Periptychinae have no known descendents after this stage (except for a few scraps in the early Upper Paleocene apparently not generically different from those of the Middle Paleocene). The amblypod astragalus became very markedly modified in Coryphodon. Its limbs are highly graviportal throughout, and the astragalus is profoundly modified and convergent toward some other graviportal types. The present conception of this order is as follows: ® Order CONDYLARTHRA: Family Hyopsodontidae: Subfamily Mioclaeninae Subfamily Hyopsodontinae group. Small, possibly insectivorous (in habits, not affinities) ani- mals with simple, low-crowned, bunodont teeth and clawlike unguals. This longest-lived group is also in almost all respects the least specialized. Lower Paleocene to upper Eocene. North America. Family Phenacodontidae: A progressively more cursorial and probably more strictly herbivorous group, generally analogous within this much more primitive and nonadaptive radiation to the early progressive ungu- lates (especially perissodactyls) that replace them in the Eocene. Teeth brachyodont, becoming polybunous with some slight tendency toward lophiodonty. Lateral toes becoming somewhat reduced and unguals flattened into hoofs. Middle Paleocene to lower Eocene. North America, South America, Europe. | Members of a persistently very primitive 82 Some of these characters were supposed to be typical of the Condylarthta, because Phenacodus waS vhe only adequately known genus when the order (or suborder, then) was first defined, and it may seem strange to consider Phenacodus as a peculiar and in many respects atypical condylarth. It is, in some sense of the word, technically a ‘‘type’’ of the Condylarthra, but it is definitely not typical throughout. Taxonomy has many such cases in which a natural group was recognized and named, even though in the first instance it was largely based on a form later found to be margina! in it. % Matthew has given excellent diagnoses of all the groups concerned. These characteristizations are meant to be explanatory, not formally diagnostic. 224 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Order ConpYLARTHRA—Continued. Family Periptychidae: Characterized by persistently plantigrade feet and by teeth markedly bunodont, the premolars not becoming molariform (as in other families) but evolving independently into large swollen crushing teeth, the molars relatively small, with the primary cusps conical and crowded together, developing a peculiar type of polybuny especially in the upper teeth by the development of new cuspules largely internal to the primary cusps. Subfamily Anisonchinae: Small and slender forms with the basic dental characters of this family, but the teeth relatively simple and general structure apparently closely similar to the Hyopsodontidae. Lower to Middle Paleocene. North America. Subfamily Periptychinae: Larger subgraviportal forms developing heavy limbs and somewhat amblypodlike feet, with complex, polybunous molars. Lowerto Upper Paleocene. North America. Family Meniscotheriidae: Hyracoidlike animals of middle size with lopho- or buno-selencdont teeth, serial carpus and tarsus, and narrow, hooflike unguals. Their early history is unknown and their relationships doubtful. They may not be very close to the other condylarths. Their dental evolu- tion seems to have been in a direction distinct from any other primitive ungulates, and almost opposite that of the periptychids, but could have started from a common basis in the Paleocene. Subfamily Meniscotheriinae: Typical, more lophiodont forms. Upper- most Paleocene and lower Eocene. North America. Subfamily Pleuraspidotheriinae: More bunodont forms, of still more dubious position. Upper Paleocene. Europe. In the, present fauna the Hyopsodontidae are very abundant and varied. Phenacodonts are present but are not abundant, being espe- cially rare in the quarry facies. Anisonchines are not uncommon but are limited in variety, only two genera and species being recog- nized, and are much less common or varied than in the Puerco and Torrejon. The Periptychinae, so common in the San Juan Basin faunas, appear to be wholly lacking. Meniscotheres are absent, as would be expected since this group is known only from younger strata. Family HYOPSODONTIDAE Lydekker, 1889 The small Paleocene animals now believed to be condylarths allied to Hyopsodus have had a confusing and complex history, which is here to be sketched only in its more essential points. Cope’s genus Mioclaenus was at first referred by him to the Condylarthra, but he later removed it to the Creodonta on the basis of the skeletal characters of ‘‘Mioclaenus”’ feror. He referred many species to the genus, mak- ing it a sort of dumping ground for unspecialized dentitions of more or less bunodont, tubercular-sectorial type. Schlosser, in 1886, sug- gested that Mioclaenus might really be a condylarth (as Cope origi- nally supposed). Scott (1892) separated out a number of Cope’s species and placed them in distinct genera. ‘M/.” feroxr was then made the basis for the genus Claenodon, a true creodont. Scott then considered that true Mioclaenus, really allied to the type AZ. turgidus, FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 225 might be a condylarth genus, in which case “it will form a very distinct family of that order.” Osborn and Earle (1895) placed Mioclaenus in the Condylarthra and proposed a new family Mioclaenidae. They discussed only M. turgidus and evidently intended to include in the family only Mioclaenus and in that genus only the few species not definitely removed by Scott. Matthew (1897) hesitated in regarding the Mio- claenidae as condylarths but did leave them in that group. He placed in the genus Mioclaenus the species Tricentes inaequidens Cope, which Scott (1892) had made type of the genus Hilipsodon, and he proposed a new genus Protoselene for Mioclaenus opisthacus Cope. In their 1895 paper, Osborn and Earle described a new genus Oxyacodon, listed as incertae sedis but in the vicinity of the creodonts. Matthew (1897) left the genus as incertae sedis and transferred to it Anisonchus agapetillus Cope. In 1914, in a faunal list, Matthew transferred Oxyacodon to the Mioclaenidae, with a footnote that it might be a periptychid. In his subsequent work (see Pale. Mem.) Matthew confirmed the association of Oxyacodon with Mioclaenus and transferred to it the other Puerco species, “‘Mioclaenus”’ turgi- dunculus, thus confining the genus Mioclaenus (sensu stricto) to the Middle Paleocene. He also revived Ellipsodon Scott and placed in it Mioclaenus acolytus Cope and Mioclaenus lemuroides Matthew. The arrangement reached by Matthew is thus as follows: Type: O. apiculatus. Referred: O. agapetillus, O. turgidunculus, O. priscilla. Type: M. turgidus. Referred: M. lydekkerianus. J Type: E. inaequidens. |Referred: EH. lemuroides, E. acolytus. Protoselene: Type: P. opisthacus. Puerco: Oxyacodon| M ioclaenus} =n Etlipsodon The history of the Eocene group Hyopsodontidae has been suffi- ciently noticed in connection with the discussion of the Condylarthra as whole. Matthew early recognized that the hyopsodontids and mioclaenids were related and in 1909 suggested that future discovery might result in merging the two supposed families. In his Paleocene memoir he took this step, retaining the earlier double grouping in the form of two subfamilies, Hyopsodontinae with Haplomylus and Hyopsodus and Mioclaeninae with the earlier genera listed above. After Matthew’s work numerous discoveries of new hyopsodontids have been made. Jepsen (1930) described Litolestes and Phena- codaptes from the Upper Paleocene of Wyoming. He placed the former doubtfully in the Insectivora and the latter doubtfully in the Artiodactyla, but I have already suggested (Simpson, 1936b) that they are probably hyopsodontids. Finally there are three new 226 BULLETIN 169, UNITED STATES NATIONAL MUSEUM genera of hyopsodontids in the present fauna, Haplaletes, Litomylus, and Litalestes. The number of known genera has nearly doubled since Matthew completed his work, and the whole unwieldy group requires reconsideration. The genera now known and considered hyopsodontid are as follows: Choeroclaenus (defined below): Type C. turgidunculus. Lower Paleocene. Oxyacodon: Type O. apiculatus. Lower Paleocene. Mioclaenus: Type M. turgidus. Middle Paleocene. Ellipsodon: Type EH. inaequidens. Middle Paleocene. Protoselene: Type P. opisthacus. Middle Paleocene. Iitaletes: Type L. disjunctus. Middle Paleocene. Intomylus: Type L. dissentaneus. Middle Paleocene. Haplaletes: Type H. disceptatrix. Middle Paleocene. Litolestes: Type L. ignotus. Upper Paleocene. Phenacodaptes: Type P. sabulosus. Upper Paleocene. Haplomylus: Type H. speirianus. Uppermost Paleocene and Lower Eocene. Hyopsodus: Type H. paulus. Lower to Upper Eocene. The types and many other specimens of these genera have been eexamined for the present work ™ in order fully to analyze the affini- ties of the Fort Union forms here described and their contribution to knowledge of the family. In the first instance, generic designations were disregarded and the 18 well-defined Paleocene species referable to this family were graphi- cally compared by a tabulation of all their known characters. They were found to fall naturally into groups that correspond very well with the various genera recognized by Matthew and _ those defined after his work. The principal characters distinguishing these generic groups are shown in the following key, which also gives (in parentheses) the species now placed in each genus, the type being indicated by an asterisk. KEY TO THE PRINCIPAL GENERA OF HYOPSODONTIDAE KNOWN IN THE PALEOCENE I. Paraconids internal, fusing with metaconids; entoconids indis- tinct, fusing with hypoconulids. A. P*%, enlarged, inflated, few accessory cuspules. 1. Paraconids less internal, entoconid distinct on M2, few or no crenulations, M; large, with projecting hypoconulid, M? with large metacone___________-- Choereclaenus % (turgidunculus*) 2. Paraconids wholly internal, entoconid indistinguishable on Mz, crests crenulated. M3; much reduced, with rounded heel. Metacone vestigial on M’________----- Mioclaenus (turgidus*) %{ Material of Litolestes and Phenacodaptes through the courtesy of Dr. G. L. Jepsen. The American Museum collections includemostof the types and many excellent specimens of all the generaexcept Phenaco- daptes. The type specimen of Hyopsodus paulus was not seen, but many good specimens of this and other species of Hyopsodus were examined. 95 Two or three poorly known supposed species of doubtful status were omitted. % New genus, defined on p. 232. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 224 B. P44 cuspidate, more or less enlarged but not inflated. 8. Ps without distinct paraconid, metaconid absent or rudi- mentary, talonid relatively wider, M*; much to somewhat reduced. Ellipsodon (inaequidens*, priscus, lemuroides, acolytus, aquilonius) 4. P, with rudimentary paraconid, metaconid relatively large, talonid narrow, M?’; relatively large___-____-- Litaletes (disjunctus*) II, Paraconids median to subinternal, not fusing with metaconids, entoconids distinct, molar talonids basined. C. Teeth more lophiodont, P, bicrescentic, paraconids distinct, P4 with metacone, M!~? with strong mesostyle, M’; large. ee a ee en ee RR eee ee ee Protoselene (opisthacus*) D. Teeth more bunodont, paraconids reduced or vestigial, no mesostyles, M*; generally somewhat reduced. a. Py, without distinct paraconid, but relatively elongate and trenchant, Ms; large, with projecting hypoconulid, molar cusps acute. 6. Py with rudimentary metaconid, molar paraconids rela- tively large_-_-_-- Cxyacodon (apiculatus*, agapetillus, priscilla) 7. Ps with distinct metaconid, with a pit between this and protoconid, molar paraconids reduced___Litomylus (dissentaneus*) b. Py with distinct small paraconid but relatively wide and heavy, M3; somewhat reduced, molars bunodont, generally broad and low. 8. P, with smaller metaconid and talonid, talonid of Ms; less elongate, upper molars more transverse and angulate, outer cusps of P’-4 more compressed. Litolestes (ignotus*, notissimus) 9. Metaconid and talonid of P, larger, M3; more elongate, P-4 with more conical outer cusps, upper molars rounded and less-transverses32_ erat tee 8! ees Haplaletes (disceptatrix*) c. Py elongate, with strong paraconid, M%; reduced, prominent cingulum descending from protocone tip. TL ag ah ge i en y= tape Scola gh ee ac a Haplomylus (speirianus*) Like all keys, this is artificial, but it is based on an extensive analysis of the characters of all the species listed, and an effort has been made to select characters that have clear taxonomic value and are probable or possible indications of phyletic relationships. Thus it is believed to be probable that the capital letters indicate four natural groups of genera. The primary division, indicated by Roman numerals, also appears to me (but with somewhat less probability) to be a natural dichotomy of the whole group. Matthew has suggested that Choeroclaenus turgidunculus, which he referred to Oxyacodon, might be ancestral to one or more species of Ellipsodon. The resemblance is certainly close, and I believe a relationship to exist. It seems to me, however, to be that indicated above by group J and hence more general than he suggested and inclusive of Mioclaenus, sensu stricto. In more exact phyletic relation- ship, Choeroclaenus turgidunculus appears to be closer to Mioclaenus turgidus The resemblance in the molars is as close as to Ellipsodon, 228 BULLETIN 169, UNITED STATES NATIONAL MUSEUM and the premolars are practically those of Mioclaenus in miniature and unlike those of Ellipsodon. Furthermore, the rather poorly known Ellipsodon priscus carries that genus, or something very like it and probably closely related, back into the Lower Paleocene, contempo- raneous with Choeroclaenus. The group of species referred to Hllipsodon is rather heterogeneous, as discussed under that genus. The presence of so many varied species shows that several divergent minor phyla are present, but all appear to be rather closely allied. Hllipsodon priscus represents a possible an ‘estral type of structure, without being clearly allied to any particular one of the Middle Paleocene species. The other species are all approximately contemporaneous and so represent a spreading out of the group without permitting the discernment of any special lines of descent. Oxyacodon represents the second major group in the Lower Paleo- cene. Its distinctive characters are almost entirely primitive and it affords a structural ancestry for its general group, D. Itis improbable that the ancestry of Protoselene would enter into Oxyacodon, and the case of Haplomylus is also dubious. Upper teeth of Oxyacodon are unknown, and might considerably modify the present conception of the genus. Litomylus very closely resembles Oxyacodon but is in at least two respects, molarization of P, and reduction of molar paraconids, a more advanced form. As far as the scanty data go, it could be a relatively unprogressive descendant of Oxyacodon. Haplaletes and Litolestes, both possible structural derivatives of Oxyacodon, are successive, Middle and Upper Paleocene, respectively, and appear to be close relatives, but they cannot be along exactly the same line of descent, at least in the known species. Litolestes, the later genus, is probably more specialized in the reduction of M*; and perhaps in the more transverse upper molars, compression of P*~*, and some other details, but its premolars seem to be slightly but dis- tinctly less progressive than in Haplaletes. (Its known species are also somewhat smaller than Haplaletes disceptatriz.) Haplomylus appears at the end of the Paleoce..2 and runs into the lower Eocene. It is clearly a member of this general group, but none of the older genera is enough like it to suggest any very close structural ancestry. Its general premolar and molar structure, although somewhat more advanced as would be expected, is of the type of group D of the foregoing key and is such impelling evidence of relationship that the genus has been classed with that group. At the same time some important details are not foreshadowed in any of the other genera. The most striking point, the development of the posterointernal part of the upper molars as a broad cingulum sweeping down from the protocone, curiously reminiscent of some of the ancient primates, is approached (but not very closely) in FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 229 Ellipsodon, but the other characters of the dentition almost exclude the possibility of special relationship. Protoselene is a more sharply defined genus than any of the others here considered. It has the general characters of a primitive hyopso- dontid but is evidently becoming specialized throughout the dentition in a way hardly suggested by any other genus. Hyopsodus is not inserted in the above key, because it is not known in the Paleocene and because it is so distinctive that it can be recog- nized at a glance, and confusion with the Paleocene genera is impos- sible. This distinctive character, however, is entirely in features demonstrably progressive, and anyone who studies the whole structure of Hyopsodus, particularly with reference to the evolution that occurred within that genus, can hardly fail to endorse Matthew’s conclusion that it is an ally of the Paleocene forms here discussed. The genera discovered since Matthew’s work still more strongly substantiate the reality of this relationship, for it may now be said that Hyopsodus has no known structural character not clearly developed or adumbrated in the Paleocene hyopsodontids.” Hyopsodus most nearly resembles group D of the foregoing key, and in a general way this group has every essential requirement for the structural ancestry of the Eocene genus. Its most exact resem- blance in details appears to be with Haplaletes. The sequence Oxyacodon—Haplaletes-Hyopsodus is, as far as it is known, one in which no difficulty opposes its acceptance as a structural phylum. _ There are no ‘‘crossing specializations’, and all characters seem to be modified uniformly and in one direction in accord with the relative ages of the genera. At the same time it is, of course, apparent that the data are inadequate to prove that this is an exact genetic phylum, and, as in most cases, the probabilities are very much against our having in collections the exact members of the true line of descent. There remains for discussion only Phenacodaptes Jepsen, 1930. This has not been inserted in the key because its affinities with the other genera are not definitely established, and Dr. Jepsen has material that he has not yet described and that may give a_ better basis for decision. - From his published data, the genus appears to me to enter into the Hyopsodontidae. It has characters strongly suggestive of the dichobunid artiodactyls, but so has the whole Paleocene group of hyopsodontids. It is, indeed, almost impossible to frame a diagnosis, on dental characters alone, that will surely distinguish hyopsodontids and artiodactyls, yet such skeletal parts as are known show that they were quite distinct, at least in the lower Kocene, and even the dentitions give a definite feeling, supported 97 Mile. Friant’s recent reference of Hyopsodus to the Insectivora and strange discussion of the derivation of insectivore, especially erinaceoid, molar patterns can be quite ignored. She seems to be wholly unaware of any of the evidence for the true affinities of Hyopsodus. 230 BULLETIN 169, UNITED STATES NATIONAL MUSEUM by distinctions open to exception but fairly distinctive with the whole group in mind, that they are different groups. Whether it be con- sidered as a dichobunid or as a hyopsodontid, Phenacodaptes is a peculiar form. Yet all its characters known to me are either dupli- cated or rather closely approached by various hyopsodontids, and its reference to that group is at least as probable as any other view. The fact that no artiodactyl, or no other artiodactyl, is known from the Paleocene in itself carries no great weight as regards the affinities of Phenacodaptes, except from the point of view of logical procedure in the special case. If, as seems to me to be true, Phenacodapies re- sembles a group that is known to have been abundant and varied when it lived at least as closely as it resembles another that has never been found in deposits of that age, it seems preferable to refer it to the former group pending discovery of decisive evidence. If Phenacodaptes should prove to be a hyopsodontid, it will not very closely enter any of the categories of the key, not so much that it has any nonhyopsodontid character as that it is a synthetic type. The lower premolars are somewhat more suggestive of group J, although they could well appear also in group JJ. The molar tri- gonids, as far as I can judge by the data known to me, may be either Ellipsodon- or Haplaletes-like, probably the former, but the entoconids are more as in Haplaletes and its allies. Supergeneric grouping of these forms has always been based on a separation of Hyopsodus from all other known forms. Historically it is easy to see how this arose and that it was logical to the point of being the only arrangement permitted by the data. Hyopsodus is, within this group, an advanced genus with pronounced modifications, tending to conceal its relationships to the very primitive forms. Even within the genus, knowledge was principally based on relatively late (especially middle Eocene) and specialized species. Further- more the only Paleocene forms adequately known were from the Lower and Middle Paleocene and were typified by such a form as Mioclaenus turgidus, which lies rather far from the structural ancestry of Hyopsodus. Kven Matthew necessarily based his conception of the genus on forms that suggest marked separation from Hyopsodus within the family. Aside from Mioclaenus he knew only Oxyacodon, Ellipsodon, Protoselene (with Haplomylus in latest Paleocene and early Eocene). Oxyacodon is so ancient and primitive that intermediate stages were necessary to show its probable phyletic position. Ellipsodon now appears to lie nearer Mioclaenus, at least in its typical species, than to the more Hyopsodus-like genera. Protoselene is curiously divergent and not very near any other genus. He regarded Haplomylus as to some extent intermediate between the earlier forms and Hyopsodus, and probably this largely influenced him in uniting the Hyopsodon- FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. Zou tidae and Mioclaenidae, but he saw that Haplomylus could not be ancestral to Hyopsodus and was not clearly derivative from any known Paleocene form, so that the evidence was inconclusive. The discovery, since Matthew, of several Middle and Upper Paleo- cene genera that are clearly related to Mioclaenus and Ellipsodon but that approach Hyopsodus much more closely than do those two genera has much altered the conception of this family. It is now seen that in the Paleocene there are a less Hyopsodus-like and a more Hyopsodus-like group. The distinction of the Hyopsodontinae, with Hyopsodus only, depended in part on progressive characters, sure to become uncharacteristic when forms of intermediate age and structure were found, and in part on what appear to be valid phyletic characters which separate Hyopsodus from one group of Paleocene genera, but associate it with the other. On present data, it seems preferable to base supergeneric classification on these latter characters and to draw the line not between the Paleocene forms and Hyopsodus but between those of the Paleocene forms that are less and more like Hyopsodus, grouping that genus with its closer relatives among the older genera. Subfamilies drawn upon this basis are defined below. This arrange- ment is still only tentative, and it is clear from the discussion of generic relationships above that a great deal must yet be learned before a really well-founded classification within the family will be possible, but the new arrangement perhaps represents a step toward thisend. The most doubtful points, as regards the forms now known, are the affinities of the more atypical species placed in Ellipsodon and of Litaletes, the true place in the system of the rather isolated genus Protoselene, that of the apparently aberrant Haplomylus, and the relationships of Phenacodaptes. The new evidence substantiates without greatly altering the grounds for considering the hyopsodontids as condylarths, sufficiently set forth by Matthew. Discovery of intermediate forms makes the family more coherent than it seemed to him and improves the evidence for considering the relatively well known Hyopsodus as indicative of the affinities of the Paleocene genera, and so strengthens his conclu- sions. The resemblance of the early hyopsodontids to the dichobunids in the dentition is so remarkably close that it is difficult to ascribe it entirely to convergence. Although the known skeletal parts are not of artiodactyl type, it is quite possible that some branch of the earliest hyopsodontids did give rise to the Artiodactyla, but this can be proved or disproved only by further discovery. Even if this should prove to be the case, the hyopsodontids as a whole would probably be best classified as Condylarthra, since they had the general characters of that order, and retained them after the ancestral artiodactyls were distinctly differentiated in the skeleton. 1192123716 202 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Subfamily Mi0cLAENINAE Matthew (ex ms.) (Mioclaenidae Osborn and Earle, 1895) Revised definition.—Paleocene hyopsodontid condylarths with P*, more or less enlarged and sometimes inflated. P, generally relatively simple, with small talonid. Molar paraconids reduced, internal, fusing with metaconids. Molar talonids generally open, entoconids reduced, fusing with hypoconulid and becoming vestigial. MM’, often more or less reduced. M!~? with very weak or no hypocone, posterior cingulum tending to run to tip of protocone. This subfamily is redefined to include Mioclaenus, its structural ancestor Choeroclaenus, Ellipsodon, and (doubtfully) Litaletes, and to exclude Protoselene, Oxyacodon, and some other genera formerly placed in it. Choeroclaenus does not occur in the fauna here under discussion, but it 1s defined below because reconsideration of the whole family demands the proposal of this new name. Ellipsodon and Litaletes have species in this fauna and are further considered in con- nection with these species. CHOEROCLAENUS 88, new genus Type.—Mioclaenus turgidunculus Cope, 1888. Distribution.—Lower Paleocene, Puerco, New Mexico. Diagnosis.—P*, bulbous, inflated. P, without anterior cuspule, paraconid or metaconid, talonid very small, with one faintly crested cuspule. Molar paraconids small but distinct, nearly confluent with metaconids but not wholly internal. Entoconids distinct and about as high as hypoconulids, molar talonids basined. M4, little or not reduced. My, with projecting hypoconulid. M!~* transverse, with sharp external, anterior, and posterior cingula. Hypocone indistinct, posterior cingulum tending to connect with protocone tip. M? with well-developed metacone. Conules small, distinct. Cusps low but sharp and clear-cut, crenuiations and proliferation of minor cuspules slight or absent. Discussion.—The type species rather closely resembles the type of Mioclaenus, M. turgidus, and has almost invariably been referred to that genus.’ It is, however, sharply distinguished by the characters given above and in the key on a previous page. Most of these dis- tinctions are primitive characters, and they tend to link this form, more nearly than the later and more aberrant Mioclaenus turgidus, to the small and more generalized early hyopsodontids. Matthew, the only person who had critically examined the specimens since Cope, recognized this and recorded it (Pale. Mem.) by transferring the species to the primitive genus Oxyacodon, a structurally defensible and reason- %8 Xozpos, pig, + claenus, a combination meant to suggest the dental resemblance to bunodont artiodactyls, relationship to Mioclaenus, and derivation from the Puerco beds. % The single exception, previous to Matthew’s last work, seems to be Roger, who placed it in Protogonodon in his catalog, but the reason for this is not apparent. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 233 able assignment. Our present greatly expanded knowledge of Middle Paleocene hyopsodontids, however, throws emphasis on certain charac- ters that now appear to be more important phyletically than those previously apparent. The premolar inflation is unlike Ozyacodon both in kind and degree and points toward Mioclaenus, s. s., and not Ellipsodon. In the lower molars the position, shape, and connections of the paraconid are very much unlike Oxyacodon and point toward Mioclaenus and Ellipsodon rather than toward Litomylus and Hap- laletes (described since Matthew’s death) as does Oxyacodon. The entoconids are, indeed, distinct, an Oxyacodon-like character, but they are definitely fusing with the hypoconulid and do not suggest continued independence as in the group to which Oxyacodon belongs. (No Middle Paleocene genus known to Matthew belonged to this group, although it is now richly represented, and so he could not evaluate the impor- tance of this character.) The upper teeth are not known in true Oxyacodon, but the characters of the lower teeth sharply distinguish Choeroclaenus from that genus. Genus ELLIPSODON Scott, 1892 This genus is revised in Matthew’s memoir. It is unusually varied in structure, and the species here referred to it increases this variety. Ellipsodon aquilonius, of this fauna, closely resembles E. acolytus. The latter is fairly close to H. lemuroides, which in turn approaches £. inaequidens. There is no logical or convenient separation, probably of more than specific rank, in this series of four species, yet EL. aqguilonius is markedly unlike H#. inaequidens, which is the type of the genus. Whether any generic or subgeneric separation is proper, must depend on future discovery. At present it seems that the genus may be too broadly drawn, but this is not demonstrated inequivocably, and it probably is a natural genus in the sense that the species referred to it are related to each other. The type species, HL’. inaequidens, is poorly known. The type speci- men is a palate with P°-M? of one side or both, all the teeth deeply worn, and the palate encased in hard concretion. Another palate is less worn, but even more obscured by concretion, and a third shows M'~? fairly well preserved. Lower jaw fragments reveal P.-M;. All the few surely referable specimens were collected by Baldwin for Cope in 1882 to 1885, and the great collections made subsequently contain only one specimen possibly referable to the species, and this one is highly dubious. As far as their localities are recorded, Baldwin’s specimens all came from Gallegos Canyon, and it is possible that they are from some local pocket worked out by him or not well exposed when later collectors visited the area. The unusual circumstance that the type is the poorest known species of the genus and has not turned up in new collections for over 50 years makes the status of the 234 BULLETIN 169, UNITED STATES NATIONAL MUSEUM genus somewhat uncertain. The structure of Py, the extreme reduc- tion of M;, and perhaps some less clear details are peculiar with respect to LH. lemuroides or E. acolytus, and the fact that the genus is known principally from these atypical species may mean that its true nature, as based on EL. inaequidens, is now seriously misunderstood. In general aspect, . aquilonius resembles some of its associates. such as Litaletes disjunctus more than it does Ellipsodon inaequidens, but in structural detail it seems closer to Ellipsodon acolytus and is conservatively associated with that species generically (rather than definitely with the type of Ellipsodon). ELLIPSODON AQUILONIUS Simpson Figures 62, 63 Ellipsodon aquilonius Stmeson, 1935d, p. 242. Type.—U.S.N.M. no. 9280, right lower jaw with P;—M; and alveoli. Collected by A. C. Silberling. Paratype —U.S.N.M. no. 9567, right upper jaw with P?—-M?. Collected by Dr. J. W. Gidley. Horizon and locality—Types from Gidley Quarry, surely referable specimens from Silberling Quarry and one, more doubtful, from Loc. 51, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis.—Close to E. acolytus in size and structure, but teeth generally slightly slenderer, P, relatively shorter and with metaconid more distinct. M’*; only moderately reduced. PP? without protocone. M!~? with rudimentary hypocone not connected to protocone apex. Measurements given below. Discussion —This is one of the commonest species in the quarries and is represented by a fine series of specimens revealing its dental morphology and variation in detail, although in no case is the anterior dentition preserved. The number of incisors is unknown. The post-incisive dentition was complete numerically. The upper canine and P! are unknown. P? is a small simple tooth with a small anterior and a larger posterior root. The crown 1s com- pressed, trenchant, with a median cusp, minute posterior cuspule, and posterointernal cingulum. P® has three roots but is longer than wide and has only one distinct cusp, which is central and is triangular in section, with minute anterior and posterior basal cuspules and a sharp, continuous, but not cuspidate internal cingulum, stronger on the posterointernal than on the anterointernal face. P* is wider than long and has a strong protocone, which is, however, lower than the amphicone. There is no separate metacone. The amphicone is triangular and has a sharp posterior and a weak anterior crest from FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 235 the tip. There is a distinct parastyle and small vague metastyle. The entire crown is circled by a cingulum which bears a cusp, topo- graphically a metaconule, near the middle of the posterior border. Aa hs? hud f FIGURE 62.—Ellipsodon aquilonius Simpson, U.S.N.M. no. 9280, right lower jaw: a, Crown view; 6, in- ternal view. Two and one-half times natural size. FIGURE 63.—Ellipsodon aquilonius Simpson, U.S.N.M. no. 9567, with parts in outline supplied from U.S.N.M. nos. 9571 and 9576; left upper jaw: a, External view; 6, crown view. Four times natural size. M' and M? have the same structure but differ in size and proportions. Although distinct, the paracone and metacone are crested and the crests tend to form a simple ectoloph. The protocone is crescentic and its wings bear definite protoconules and metaconules, just internal to paracone and metacone. There is a sharp, continuous external cingulum, small definite parastyle, very vague metastyle, and no mesostyle. The pronounced anterior cingulum stops short at the anterointernal corner without rising or rounding the long internal slope of the protocone. The otherwise similar posterior cingulum 236 BULLETIN 169, UNITED STATES NATIONAL MUSEUM rises toward the protocone apex at the inner end and terminates in a definite point, a rudimentary hypocone on the posterior protocone slope and near, but separate from, the protocone tip. M?® is oval, with rounded corners, and is reduced but much less so than in £. inaequidens. It has a vestigial but distinct metacone. The lower canine is known only from its single, cylindrical root, which indicates a small tooth (but larger than P,) only slightly procumbent. P, has one small root and a slightly procumbent and recurved simple crown, excavated on the posterointernal face. P. is considerably larger, 2-rooted, and with a single distinct heel cusp. P; is transitional to P, in structure, with the heel considerably expanded and a curved crest and excavation at the anterointernal angle. P,, although sharply distinct from the molars, is more nearly molariform than in any other species referred to this genus. There is a low distinct metaconid on the posterointernal protoconid slope, at about two-thirds of the distance from the base to the apex of the crown. The anterointernal protoconid slope is excavated, and the anterior protoconid crest curves inward and then posteriorly around it, gener- ally without forming a cusp but in a few specimens with a very rudi- mentary and low paraconid. The talonid has a slight crest ending in a cusp at the posterior margin, somewhat external to the midline, with a vague, open internal basin and a small posterointernal cuspule. M, has the protoconid and metaconid opposite, the metaconid slightly the larger of the two. The paraconid is distinct, fully internal, partly connate with the metaconid, and smaller and lower than the latter. The talonid is nearly as high as the trigonid and is well basined, but with the basin open in a narrow notch between entoconid and metaconid. The hypoconid is large, distinct and crescentic. The small hypoconulid and larger and equally high entoconid are poorly separated. When quite unworn, three small cuspules are seen, one on the anterior hypoconid wing, one on the posterior metaconid slope, and one on the anterior entoconid crest. The talonid is wider than the trigonid. Mg is similar to M;, but the trigonid is larger, absolutely and relatively, and is as wide as or wider than the talonid. The entoconid is reduced in size, in height, and in distinction and the talonid basin more open. Mg is reduced, the trigonid decidedly the widest part of the tooth, the entoconid indis- tinct, and the hypoconulid large and sharply projecting as a well- defined spur. U.S.N.M. no. 9686 preserves dm,. Its talonid closely resembles that of M, but is smaller. The trigonid is much more elongate and narrow than on the permanent molars, and the paraconid is larger and well separated from the metaconid. The horizontal ramus is slender, with a long symphysis, which seems not to have fused even in old age. The posterior mental FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. Ie foramen is generally beneath the posterior end of P,; or anterior end of Py. The coronoid process seems to have been high and broad. The masseteric fossa is shallow and simple, its only sharp boundary anterior. The angle is poorly preserved but evidently was rounded and not sharply projecting. The dental foramen is about on a level with the alveolar border. The condyle is elevated well above the teeth and its transversely oval, gently convex articular surface faces equally posteriorly and superiorly. The infraorbital foramen is above P*, and the zygoma arises chiefly above M?”. The statistical constants of the principal tooth dimensions are given in table 51, based entirely on the sample from the Gidley Quarry. TaBLE 51.—Numerical data on upper and lower dentition of Ellipsodon aquilonius Variate N R M o V (iP ees ery ay cect: 10 2. 8-3.5 | 3.22+0.06 | 0.19 +0. 04 6.0+1.3 \ileh ne eee eae ee 10 1. 7-2. 1 | 1. 88+0. 03 | 0. 108+0. 024 3. 740. 8 Uh] Bp etic yet ae a 1 3. 2-3. 7 | 3.4140. 05 | 0.16 +0. 04 4.6+1.0 \W ee ae ea 11 2. 1-2.5 | 2.3140. 04 | 0. 116+0. 025 5.0+1.1 LM ae Ss oa ae 19 3. 1-3. 6 | 3. 40+0. 03 | 0. 14340. 623 4.2+0.7 TI Dep eae le ele 19 2. 7-3.2 | 2.90+0. 04 | 0. 172+0. 028 5.9+1.0 Vistas cio seers = 28 3. 1-38. 9 | 3.58+0. 03 | 0. 17940. 024 5. 0O+0. 7 VA ig es ae ee 29 2. 8-3. 6 | 3. 30+0. 04 | 0. 19040. 025 5. 8+0.8 TAN ey ga ge apd 24 3. 5-4. 1 | 3.82+0.03 | 0. 134+90. 019 3.540. 5 WAM. grate th 24 2. 4-3.0 | 2. 75+0.03 | 0. 16140. 023 5.8+0.8 DD EREAS a Sea eee a 6 3. 1-3. 3 | 3.2340. 03 | 0. 07540. 022 2.340. 7 We Seee eer 7 2. 5-2.9 | 2.7140. 06 | 0.15 +0. 04 5.44+1.4 Up semete et pee be 7 3. 0-3. 2 | 3. 1140. 024! 0. 06440. 017 2.1+0. 6 AWG REL ie ee ae poy a of 3. 8-4.1 | 3.9640. 04 | 0.12 +9. 03 3.0+0.8 TEIN Uae ge el satarplen 10 3. 1-3. 4 | 3. 21+0.03 | 0. 104+0. 023 3. 30+0. 7 VV ie 2 a SROs ae 10 4.1-4.5 | 4.29+0.05 | 0.14 +0. 03 3.4+0. 8 Ie eee 13 3. 1-3. 7 | 3.4340. 05 | 0.17 +0. 03 4.9+1.1 WaViz eine Sie anne 13 4.9-5.4 |] 5.20+0.05 | 0.17 +0. 03 3.2+0. 6 TE LNs eee ee 8 2. 3-2.6 | 2.40+0. 04 | 0. 100+90. 025 4.2+1.0 Weiser mi g 3. 5-4. 1 | 3. 7140.07 | 0.21 +0. 05 5.6+1.3 Despite the fact that the sample probably includes both sexes and certainly includes teeth in many different stages of wear, the figures are very consistent and show remarkably little variation. The high- est V is only 6.0 and the average is 4.35. The consistent distribution of the V’s is also striking, for of the 20 values only two differ from the average by as much as twice the corresponding standard error. The two exceptions (for length of P? and length of P*) are based on scanty data and are abnormally low values, 2.3 and 2.1. The unusually adequate sample of E. aquilonius from the Gidley Quarry and the presence of seven lower jaws apparently of the same species from the Silberling Quarry afford the best opportunity to detect any minor differences that might occur between these two horizons and localities. The possible association of morphological 238 BULLETIN 169, UNITED STATES NATIONAL MUSEUM variants with one locality or the other and possible differences in mean dimensions have been carefully compared. The greatest differ- ence in mean dimensions is only 0.2 mm (for width of P3), which is not shown to be significant (it being mathematically demonstrable that a difference as great would arise in random sampling of a homo- geneous sample about once in 20 trials or oftener). The other differ- ences are far from any probable significance. Only one specimen from the Silberling Quarry falls outside the observed range for the Gidley Quarry in a single dimension, having M, 3.8 mm in length, but this is far within the probable range of the Gidley Quarry popu- lation, the deviation being only 1.4 times the standard deviation. It is unnecessary to give the results of the numerous other detailed comparisons made, since all were negative, showing no significant difference between the samples from the two quarries. Since the samples are so good, this warrants the positive affirmation that a single race of this species occurs in both quarries. A single specimen from Loc. 51, U.S.N.M. no. 9709, a lower jaw with M, and broken M,, has these two teeth above the average size for the Gidley Quarry sample, but within the range of the latter. ELLIPSODON species U.S.N.M. no. 9662, from Loe. 18, is a partial right lower jaw with Py. This tooth resembles that of Hllipsodon aquilonius but is rela- tively higher, the paraconid is more distinct, and the length, 4.2 mm, is significantly greater than in that species (d/o=4.9), although the width, 2.6 mm, is not (d/s=2.5). This is probably another species, but the material is inadequate for its exact determination. Genus LITALETES Simpson Litaletes Stmpson, 1935d, p. 242. Type.—Lutaletes disjunctus Simpson. Distribution —Middle Paleocene, Fort Union, Montana. Diagnosis.—P, with bladelike main cusp, distinct paraconid and relatively large metaconid, talonid small with narrow, rudimentary, open basin. Molar paraconids distinct, small, internal. Entoconids indistinct, fusing with hypoconulids, especially on My. P* with rudi- mentary metacone, strong metastyle. M?!~° with distinct hypocones. M:’; not reduced, M? with strong metacone. Discussion.—This genus is very distinct from Ellipsodon, but Litaletes disjunctus and Ellipsodon aquilonius do not differ very greatly. It could hardly be supposed that Ellipsodon inaequidens, type of that genus, is congeneric with Litaletes disjunctus, for their whole adaptive tendency seems different, and each tooth has definite and pronounced structural distinctions. The question then is not whether Ellipsodon and Litaletes are distinct genera, but where the FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 239 more or less transitional Ellipsodon aquilonius belongs. It is closer to Ellipsodon acolytus than to Litaletes disjunctus or any other species known to me, and, as already stated, that is the reason for referring it to Hllipsodon at present. It is, however, probably closer to Lita- letes disjunctus than to Hllipsodon inaequidens. Perhaps it will be necessary to transfer /. aquilonius and E. acolytus to Litaletes at some future time, but that introduces a great difficulty as regards the generic position of E. lemuroides, and for the present the system adopted here seems equally natural and more convenient. LITALETES DISJUNCTUS Simpson Fiaurss 64, 65 Litaletes disjunctus Simpson, 1935d, p. 242. Type.—vU.S.N.M. no. 9323, right lower jaw with C—M; (M; broken). Collected by A. C. Silberling. Paratype—U.S.N.M. no. 9324, right upper jaw with P%—M?. Collected by A. C. Silberling. Horizon and locality —Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis.—Sole known species of genus. Measurements given in table 52. Discussion. —P** and M! of the species are now known only from the paratype, so that their variation is not established. On this specimen both P’ and P* have distinct parastyle and metastylie and on both the posterior amphicone crest bears a rudimentary metacone, largeron P*. P*hasarudimentary protocone. On P* the metaconule is not developed on the cingulum but more normally, on the protocone wing, and the cingulum does not cross the inner face of the protocone. On M!* the parastyle is unusually prominent and the hypocone is larger than in Ellipsodon aquilonius and not so near the protocone apex. M® is less reduced, less oval, and the metacone, although smaller than the paracone, is large and distinct. TABLE 52.— Available numerical data on lower teeth of Litaletes disjunctus from the Gidley Quarry Variate N R M =(d)? | [jd BAS = sx SS SAE URE aa Sees Sis a 3 3. 3-3. 5 Sell 0. 0267 Vi der! Leas Ae Te a ot 0 ERE ae fet nae re 3 1. 9-2. 1 2. 00 0. 0200 lib, ss ets Se Ee Oe Et ieee ee 5 3. 5-3. 9 3. 68 0. 1080 VOY TP aoe spre: NONE AS ine eae TURN ESS. 5) 2. 4-2. 8 2. 54 0. 1120 a Vinmeeearsatee ed CP ea ee 0 3. 9-4. 2 4, 07 0. 0743 VIAL ie 2 Sete ee, Role REI St ne OY Oa Py SRR 5 3. 2-3. 5 3. 30 0. 0890 LEN Ti este es oe en er aie 6 4, 3-5. 2 4. 67 0. 5134 WIM [pp kaos oe Sek Be ee ees ee 5 3. 7-4. 5 4, 04 0. 3520 TE iM open Sven's Sie ene Se ene Say ae gE 3 5. 0-5. 3 op als} 0. 0467 NI ING lye i a i a peed ees. 2 as 4 3. 3-3. 8 53 ts 0. 1700: 240 BULLETIN 169, UNITED STATES NATIONAL MUSEUM f Sank } Liles nN i} me FIGURE 64.—Litaletes disjunctus Simpson: a, U.S.N.M. no. 9281, right Py-M3, crown view; 6, U.S.N.M. no. 9323, right lower jaw, external view; c, U.S.N.M. no. 9338, right lower jaw, internal view. ‘Twice natural size. FIGURE 65.—Litaletes disjunctus Simpson, U.S.N.M. no. 9324, right upper jaw: a, External view; 0, crown view. Three times natural size. The lower canine is a small but tali, erect, spat- ulate tooth. P, is low, l-rooted, with a minute heel. P._3 are more ad- vanced than in Ellipsodon in that each has a distinct paraconid. The heel is, however, relatively short, and only half its width is formed by the incipient basin. The anterior blade of the protocone is dis- tinctively modified into a sort of shearing crest. The lower molars close- ly resemble those of Hlip- sodon aquilonius, but My has trigonid and talonid of nearly equal width, M, has trigonid generally markedly wider, and M; is less reduced. U.S.N.M. no. 6179 from Loc. 51, includes an M, morphologically compa- rable to this species and measuring 4.7 by 3.9 mm, near the means for the Gidley Quarry specimens. There is, however, an upper jaw with M?~* from the same locality, dimen- sions given in table 53, that is morphologically very close to this spe- cies but notably smaller than the available Gid- ley Quarry specimens. Since, however, these are only two in number it cannot be assumed that a real difference exists. The greatest relative dif- ference, in length of M?’, FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 241 does not necessarily imply a coefficient of variation higher than 5 on the hypothesis that a single species is present, and this is a very moderate degree of variation for one species. TaBLE 53.— Measurements (in mm) of available upper teeth of Litaletes disjunctus P3 Ps Mi! M2 Mi U.S.N.M. no. L Ww L WwW L W L lel ered by WwW 9324:(Gidley/Quary) == =22 22. 2-3 3.4 2.8 33 4.4 3.9 5.0 4.6 6.2 3.6 5.3 Phsza(GidleysOuarry) ae ae I eh ree Oe Se 4.8] 6.5] 3.4 5.6 HEU sooty Se ae | Bem a | pele | bln £0( 66) 30 Rd Subfamily HYOPSODONTINAE Trouessart, 1879 ! Revised definition——Paleocene and Eocene hyopsodontid condy- larths with P*,; not inflated, generally moderate in size, cuspidate, somewhat molariform but never exactly so, P, generally with a meta- conid and wide but imperfect talonid basin. Molar paraconids median to subinternal, not fusing with metaconids, entoconids dis- tinct and high, talonid basins closed. M®%, little or not reduced. M!~? with definite hypocone, small in earlier and large in later forms, sharply distinct from tip of protocone. Remarks.—Aside from the type genus, I place here Oxyacodon, Litomylus, Litolestes, and Haplaletes with some assurance, Protoselene and Haplomylus very doubtfully. Litomylus and Haplaletes represent the subfamily in the present collection. Genus LITOMYLUS Simpson Litomylus Stimpson, 1935d, p. 243. Type.—Litomylus dissentaneus Simpson. Distribution.—Middle Paleocene, Fort Union, Montana. Diagnosis.—P, trenchant, paraconid rudimentary, metaconid dis- tinct, separated from the protoconid by a small pit. Molar cusps rather bunodont but acute, paraconids reduced and median, M’*; little reduced, hypoconulid of M3; sharply projecting. Hypocones of M!* relatively large and internal. M®* transverse, triangular. LITOMYLUS DISSENTANEUS Simpson FiGure 66, 67 Litomylus dissentaneus Stmpson, 1935d, p. 243. Type.—U.S.N.M. no. 9425, left lower jaw with P;-M;. Collected by A. C. Silberling. 1 “Hyopsodinae”’ in the original publication, but the emendation can hardly be claimed to change authorship. DAD BULLETIN 169, UNITED STATES NATIONAL MUSEUM Horizon and locality —Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis.—Sole known species of genus. Dimensions given ir table 54. Discussion.—This is the rarest of the four species of hyopsodontids in the quarry, but it is distinctive and fairly well known. FIGURE 66.—Litomylus dissentaneus Simpson, U.S.N.M. no. 94235, left lower jaw: a, Crown view; b, internal view. Four times natural size. => EV i tr cy) > RY » \ b wy ui FIGuRE 67.—Litomylus dissentaneus Simpson, U.S.N.M. no. 9557, with tooth in outline from U.S.N.M. no. 9580, right M!-3: a, External view; b, crown view. Four times natural size. No upper teeth anterior to P! are known, and P* is represented only by an uncharacteristic fragment. M'~? have sharp, subequal, nearly conical paracone and metacone. The protocone is likewise sharp and smaller than in the other species of this group. The conules are large and equal. The external cingulum is sharp and even forming equal angulations, rather than distinct cuspules, at the parastylar and meta- stylar corners. There is no mesostyle. The hypocone is larger than in any known contemporaneous species and is quite distinct from the protocone and equally internal, but smaller. M°* is markedly trans- verse and is triangular, not rounded or oval, without a hypocone but with a sharp, distinct metacone. P 3_, are long, low, narrow, trenchant teeth, unlike any others known in this family. Each has a rudimentary, median, basal paraconid. The talonids are poorly developed in each case and have only a single posteromedian cuspule and a very rudimentary posterointernal basin. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 243 The sharp posterior protoconid crest bears a very slight thickening or cuspule. The anterointernal region is somewhat excavated and has a cingulum below it. P3 has no metaconid, but P, has a small papilla about halfway up the crown, and above and external to it, between it and the protoconid apex, is a small pit or pocket. The lower molars, with their acute, well-separated major cusps, resemble those of Oxyacodon. The paraconids are vestigial, forming a small shelf or crest connected with the anterior protoconid wing but not with the metaconid. On M, this nearly reaches the inner border, on M; it is submedian, and on M; fully median. Protoconid and metaconid are of nearly equal size. The heels are well basined and the entoconids are sharp and distinct, nearly as high as the hypoconids and on M,-, larger and higher than the small hypoconu- lids. On M; the hypoconulid is as high as the hypoconid and projects sharply posteriorly. Only five specimens, ail from the Gidley Quarry, are recognized as of this species. Measurements are given in table 54. TaBLE 54.—Jndividual measurements (in mm) of upper and lower dentition of Litomylus dissentaneus P3 P, Mi M2 M3 U.S.N.M no. L W L WwW L WwW L W L W 9425 eee ee ee eee 33 14 3.5 1.9 2.9 2.4 3.0 27 352 252 OST See o ons 22s oo 5 Son eee |e Se ee eS ees ha 2.9 21 3.0 225 3.4 2.2 ih a 3.6 1.6 3.9 BY, M! M? M3 U.S.N.M. no. L W L WwW L WwW BD fete Oe Stoo oe se. Slee ee eh st ooo oe 3.2 4.2 3.3 4.7 RS ai a5 Se ee nes ee ee ee eee 2.9 3.9 3.2 4.5 2.6 4.1 Genus HAPLALETES Simpson Haplaleies Simpson, 1935d, p. 243. Type.—Haplaletes disceptatriz Simpson. Distribution.—Middle Paleocene, Fort Union, Montana. Diagnosis.—P, not trenchant, with distinct paraconid and meta- conid and relatively large basined heel. Molars bunodont, cusps rather rounded and low. Paraconids reduced and median. M3 somewhat reduced, hypoconulid of Ms; slightly projecting, markedly less than in Litomylus. Amphicones of P?-* more rounded than in Litolestes. P* with rudimentary metacone. Upper molars rounded. Hypocones of M!~? distinct, but smaller than in Litomylus. M® less transverse and more rounded. 244 BULLETIN 169, UNITED STATES NATIONAL MUSEUM HAPLALETES DISCEPTATRIX Simpson Fraurss 68, 69 Haplaletes disceptatriz Stimpson, 1935d, p. 244. Type.—U.S.N.M. no 9500, right lower jaw with P;-M;3. Collected by A. C. Silberling. Paratype —U.S.N.M. no. 9555, right upper jaw with P?-M?. Col- lected by Dr. J. W. Gidley. FIGURE 68.—Haplaletes disceptatrix Simpson, U.S.N.M. no. 9500, with parts in outline supplied from U.S.N.M. no.5,9600, right lower jaw: a, Crown view; 6, internal view. Four times natural size. TC WW cfN: FIGURE 69.—Jlaplaletes disceptatrix Simpson, U.S.N.M. no. 9555, right upper jaw: a, External view; b, crown view; c, internal view. Four times natural size. Horizon and locality—Gidley Quarry, Fort Union, Middle Paleo- cene horizon, Crazy Mountain Field, Mont. Diagnosis.—Sole known species of genus. Dimensions in tables]55 and 56. Discussion.—This delicate little species is the smallest condylarth in the quarries, and only Litolestes includes smaller known species in this order. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 245 TaBLE 55.—Principal available numerical data on lower dentition of Haplaletes disceptatrix, all from the Gidley Quarry Variate N R M = (d?) Lg) PO oes 2 RN is MR ee Peg ee ane Sr oe 3 2. 1-2. 3 252.0 0. 0200 VE eee pace we rete area re en II 3 Teel 1. 40 0. 0200 Ji) yee gk ee se SRB OE EES alee eee bec Bbw em 2 D5 Dey 0 AA eee eee « ae RR ey eee ee 2 1. 6-1. 7 1. 65 0. 0050 TEI Cos, cg 2k nd a Sie eee i ian Me 4 2. 4-2. 6 2. 48 0. 0276 AVA IG EE RSE ERE ON Ee hee Path 2 8 4 2. 1-2. 3 DANS 0. 0276 TAY Los TRS Oe ee oe Sper Lect 6 2. 4-2. 9 2.65 0. 1350 VAIN Geese ees, Meet eee ee a 6 2. 3-2. 7 2) oil 0. 1066 IL) Nilcs: el ee etn ABA Bee Oe, COLE ot peg USES RE 6 2. 6-2. 9 2. 80 0. 9800 VESTN Ub a I et oO 6 2. 2-2. 4 Desi 0. 0286 TABLE 56.— Measurements (in mm) of the two available upper dentitions of Hapla- letes disceptatrix, both from the Gidley Quarry P? ps Ps M! M3 M! U.S.N.M. no. L W L W L W L WwW L W L Ww Cte ee ee 1.5 Vat 22 2.1 2.4 2.9 2.4 3.2 2.6 3.7 1.9 3.1 Q55 Gee ee Bees Be re ee ES ere a PN eee 2.5 3. 1 2.9 Ould: P? is a small simple tooth resembling an anterior lower premolar. P? is peculiar in having a low and small but sharp and distinct proto- cone. P? ‘resemble those of Litaletes in their rudimentary metacones, but the metastyles are less developed, and P* has no metaconule. The upper molars are similar to those of Litomylus, but are more rounded in contour and have the hypocones of M'~? smaller. P; has a distinct paraconid and a heel about as well developed as that of P, in most contemporaneous species. P, has definite paraconid and metaconid and a wide, large, basined heel with two posterior cusps. The lower molars resemble those of Litomylus, but the crowns are lower, the cusps blunter, and the internal cusps more distinctly elevated over the external. The talonid of M; is less elongate. Family PHENACODONTIDAE Cope, 1881 Most of the manuscript notes left by Dr. Gidley refer to this group, and it was evidently his intention to publish a preliminary paper on it. There are three different drafts of part of his brief account, evidently written at different times and in part not consistent with one another. The draft that seems to be most recent is still incom- plete and has still later memoranda written on it for other changes that were never made. The integral publication of this manuscript is impossible, as it does not form a connected whole and does not fairly 246 BULLETIN 169, UNITED STATES NATIONAL MUSEUM represent the final opinions of Dr. Gidley. I have, however, quoted parts of the manuscript directly, have mentioned some other of Dr. Gidley’s opinions in indirect quotation, and have used his manuscript names and followed his disposition of the specimens as far as possible. The material has been much increased since Dr. Gidley’s study of it, but it is still rare. He tentatively identified only 8 specimens, and 11 have since been collected (3 for the National Museum and 8 for the American Museum). This family is abundant in the Torrejon, and also in the lower Eocene, but it is extremely rare in the quarries in the Crazy Mountain Field and can be considered as common only at one surface locality, no. 25, from which 11 specimens have been obtained. In the Torrejon the only genus of this family is Tetraclaenodon.? It is there very abundant and highly varied and has therefore received numerous specific names, but from Matthew’s work it seems probable that only two valid species occur in the Torrejon: TYetraclaenodon puercensis, a larger, more common, and varied form, and T. pliciferus, a smaller, rarer, and perhaps less varied species. In the Crazy Mountain Field, also, there are indications of two species, one of about the size of 7. puercensis, but very rare and not exactly identifiable, the other smaller, although generally larger than T. pliciferus. There is also a second genus, Gidleyina, apparently characteristic of the higher levels in this field. TETRACLAENODON Scott, 1892 TETRACLAENODON SYMBOLICUS Gidley Fiaurss 70, 71 Tetraclaenodon symbolicus GipLEY, Simpson, 1935d, p. 239. Type.—U.S.N.M. no. 6169, part of right lower jaw with M, and alveoli of P3.4 and My. Collected by A. C. Silberling.® Paratype.—U.S.N.M. no. 6168, jaw fragment with right M,_,, and a separate left P, probably this species but probably not associated. Collected by A. C. Silberling. 2 This genus is still often called Hwprotogonia. It was originally described as Prologonia Cope, 1881. Cope later considered this as preoccupied by Protogonius Hiibner, 1816, and replaced it by Ewprologonia Cope, 1893, the type of both being P. (or H#.) subquadrata. In the meantime Scott had proposed the genus Teira- claenodon Scott, 1892, for Mioclaenus floverianus Cope. Scott did not recognize the relationship, but his Tetraclaenodon was certainly the same genus as Protogonia Cope, and it therefore includes as a synonym Euprotogonia Cope. Matthew in 1897 preferred Euprotogonia Cope, 1893, to Tetraclaenodon Scott, 1892, on the ground that the latter was based on an error, and through Matthew’s work Euprotogonia became the familiar name for the genus. Matthew later recognized that his action had been invalid, and he used the name Tetraclaenodon in all his more recent work. Now Cabrera (1935) has insisted that Protogonius Hiibner does not preoccupy Protogonia Cope, since they differ in termination, and he calls the genus Protogonia. Without taking a decisive stand, I shall tentatively continue to use Tetraclaenodon, which has the cardinal virtue of being generally and correctly understood and of being unambiguous. Huprotogonia is certainly invalid, and Protogonia is of dubious validity, is ambiguous, and is unfamiliar to present-day students. 3 IT retain this specimen as type, since it is clearly that intended by Gidley. No. 6168, here made paratype, would be a better type. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 247 Horizon and locality.—Type from Silberling Quarry, paratype from Loc. 25, about 250 feet lower stratigraphically, Fort Union, Middle Paleocene horizon, Montana. FIGuRE 70.—Tetraclaenodon symbolicus Gidley, FIGURE 71.—Tetraciaenodon symbolicus Gidley; U.S.N.M. no. 6169, right lower jaw (with M)): a, a, U.S.N.M. no. 6168, right Mi-2, crown view; Crown view; 6, internal view. One and one-half a’, same internal view; 6, Princeton Univ. no. times natural size. 13757, right M23, crown view. One and one- half times natural size. Diagnosis.—Gidley: ‘‘This species is smaller than FE. [Tetraclaenodon, G. G. S.] puercensis, being about intermediate in size between that species and FE. minor [= Tetraclaenodon pliciferus, G. G.S.]. The lower molars are proportionately narrower transversely than those of the former species,* and the lower jaw is much shallower. This last character may be due in part, however, to a less mature condition of the specimen, which represents a young individual with the first true molar just coming into use.» The striking similarity in detail of the lower molars with those of E. [T., G. G. S.] puercensis is a notable feature of the species and separates it clearly from EH. minor [T. pliciferus, G. G. S.].. The more notable points of similarity are the slight roughening and wrinkling of the enamel surface and a tendency of the lophs of the teeth to break up into small cuspules.”’ ® Simpson: Intermediate between TJ. pliciferus and JT. puercensis in size, but nearer the former both in size and in structure. The only constant difference from T. pliciferus is the greater size, inadequate for specific differentiation were it not constantly correlated with the 4 And within the range of T. pliciferus in this proportion.—G. G. S. 5 A specimen of T. pliciferus of comparable age has a deeper jaw, despite its smaller teeth but a referred specimen of T. symbolicus also hasa deep jaw. Thisis probably a highly variable character, and also depends on crushing to a considerable degree.—G. G. S. 6 From figures of the Torrejon specimens this would seem a striking and good distinction, but the speci- mens themselves show that 7’. pliciferus also has wrinkled enamel and a tendency for lophs to break up into cuspules. These may be functions of size, to a limited extent, and slightly less pronounced in T’. pliciferus than in T. puercensis. T. symbolicus is about intermediate between the two in these characters, as in size. 119212—37——_17 YAS BULLETIN 169, UNITED STATES NATIONAL MUSEUM widely different geographic distribution, as far as known. Crenula- tions possibly slightly more developed and paraconid weaker on type and paratype of 7. symbolicus, but these are variable characters and other specimens suggest that they are not of specific value. Discussion.—This species seems to be variable, and it is difficult to separate it from 7’. pliciferus, with which it must be closely related. Its smallest variants, indeed, could not be separated from T. pliciferus were they found together, but the homogeneous sample from Loc. 25 averages larger than 7’. pliciferus. Since all these individuals are from one horizon and locality they evidently represent either one herd or an actually interbreeding stock, the character of which is thus slightly different from the Torrejon species and may be given taxo- nomic distinction. The size difference is statistically significant. I therefore accept Dr. Gidley’s species, but consider it as much closer to T. pliciferus than he believed. It happens that the type and to less degree the paratype have the enamel unusually crenulated and the paraconids small, characters slightly closer to 7. puercensis than to T. pliciferus, although those species intergrade in this respect. The other specimens from Loc. 25, however, have the enamel somewhat smoother and the paraconids more distinct, almost exactly as in T. pliciferus.’ The more recently discovered specimens include one with dm,, representatives of all the lower molars and of P, and Py, and also M?*. These are all morphologically within the range of 7. pliciferus. They might be grouped into three subdivisions, large, medium, and small, but I think the grouping would be subjective and that the variation is individual and approximately normal. TABLE 57.—Numerical data on lower molars of Tetraclacnodon symbolicus Variate N R M o V Mg ent Sie ek 6 7.4-7.9 | 7.6340. 07 | 0. 18+9. 05 2.4+0.7 AWi Nie es Se 6 5. 6-6. 6 | 6. 2340.13 | 0.3140. 09 5. 0+1. 4 HN ig aercesee 2 saees Le 6 7. 5-8.2 | 7.77+0. 10 | 0. 24+0. 07 3. 0+0. 9 Wiig I Ge 6 6. 2-7.0 | 6. 75+0. 12 | 0. 30+0. 09 4,5+1.3 ORY Es ae 6 7.3-8.1 | 7. 7040.10 | 0. 25+0. 07 3.340. 9 WiAVIg Stk ae ea 6 5. 2-5. 7 | 5.87+90. 08 | 0. 20+0. 06 3.8+1.1 The available material (including that in the American Museum) provides only six well-preserved examples of each of the three lower molars, and it is not entirely homogeneous since the type and one other specimen (as given below) are not from Loc. 25, whence all other 7 One of the most extreme specimens in this respect, U.S.N.M. no. 6167, was referred to 7. symbolicus by Gidley, in his notes, so that his conception of the species was the same as mine despite the differences in the diagnosis, FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 249 specimens were derived. Nevertheless the material is well unified, and data on it can be more adequately summed up in statistical form than otherwise. (Table 57.) The dimensions of the two specimens not from Loc. 25 are given in table 58. TaBLe 58.—-Individual measurements (in mm) of lower molars of Tetraclaenodon symbolicus M; M2 M3 U.S.N.M. no. ——_———_____—_— L W L WwW L WwW GG OS ae ot a ee eee 7. 9 6.3 OOD jee sere: Dee ie ES eee Co OF. | eaten a ee eee Co Tpi7/ These are to some extent marginal, as might be expected. The type, from the Silberling Quarry, has the largest M, recorded, but a specimen from Loc. 25 approaches this within 0.1 mm. No. 9925, from Loc. 3, has the widest M, and M; recorded, but these two dimen- sions are approached within 0.2 and 0.1 mm, respectively, by Loc. 25 specimens. The variation is low for the whole series in any event, and there is no reason to believe it heterogeneous as to race. The close approach of this species to T. pliciferus is shown by the dimensions of M, of the type of the latter, length 7.5, width 6, within the range of 7. symbolicus but slightly below the mean. The most readily measurable of the types of T. “minor” = T. pliciferus, Amer. Mus. no. 3897, has the following dimensions: Length M,, 6.8; width Mu, 5.6; length Mz, 7.0; width Ma, 6.2. This is somewhat more representative of the smaller Torrejon species than is the type of T. pliciferus, which is a large variant. The lengths are below and the widths at the observed lower limits for T. symbolicus, and the differences are significant. Princeton no. 13757, from Loc. 9, one of the rare No. 1 Fort Union specimens, includes M!~? evidently of Yetraclaenodon and closely comparable to 7. symbolicus, although the variability of this species, lack of adequate material of the upper dentition, and some differ- ences from other specimens, perhaps individual and perhaps of minor taxonomic value, make the reference uncertain. TETRACLAENODON cf. PUERCENSIS (Cope, 1881) A few fragmentary specimens demonstrate the presence of a larger species of Tetraclaenodon. These include a left upper M? from the Gidley Quarry (U.S.N.M. no. 9620), associated left dm* and M? from Loc. 6 (American Museum), associated left M? and part of 250 BULLETIN 169, UNITED STATES NATIONAL MUSEUM M? from Loc. 82 (American Museum), and specimens figured by Douglass (1902b, p. 222) from Loc. 5 or 6. These specimens vary considerably among themselves, and they are not clearly distinguish- able from variants of 7. puercensis, but they are inadequate for specific determination and do not definitely establish the presence of that species in this field. The original of Douglass’ 1908, pl. 1, fig. 4, is perhaps a right P? of this same form, but this is uncertain, and the other isolated teeth referred to Tetraclaenodon by Douglass seem still more dubious. There are also preserved with U.S.N.M. no. 11913 a right and a left M! (possibly M?) probably of Tetraclaenodon and, at least in their worn condition, closely resembling 7. puercensis. They certainly are not associated with no. 11913, since they are from a much older indi- vidual, and it is very improbable that they are of the same species, and not at all clear that they are congeneric. No. 11913 is recorded as from Loc. 11 or 13. These localities are at about the same level and are the highest that have yielded identifiable mammals. A note by Silberling with the specimens seems to leave little doubt that these specimens were derived from that level except in the highly improbable case that they have accidentally been substituted for two other upper molars in the collection. Tetraclaenodon has not otherwise been reported from beds as late as this, and these teeth are inadequate to establish its presence although they make it probable. Genus GIDLEYINA Simpson ® Gidleyina Simpson, 1935d, p. 240. Type.—Gidleyina montanensis (Gidley). Distribution. Upper Paleocene, Fort Union, Montana. Diagnosis.—Gidley : ‘Cheek teeth bunolophodont; first and sec- ond upper molars subquadrate, consisting of four principal cusps, two intermediates, and a well-developed mesostyle, conules con- nected by continuous lophs with the summit of the protocone; pre- molars 3 and 4 with well-developed protocones, but with metacones rudimentary; thus superficially they each consist of two principal transversely placed cusps.” 8 The teeth themselves are not marked, as are most specimens in the collection. §In one draft of his manuscript Dr. Gidley referred the type of this genus to Huprotogonia, in another to Ectocion, and in a third, presumably the most recent, to Proectocion, new genus. His new generic name is, however, preoccupied by Proectocion Ameghino, 1904, and therefore it cannot be used. He intended to change it, for he had made a pencil notation, ‘‘change name, not related to Hctocion, but rather to Pro- togonodon’”’, but I find no other name in his notes or on his labels and so have been forced to supply one. It is highly appropriate that the genus should be named for Dr. Gidley. (Gidleya Cossman, 1907, is a fossil bovid.) 10 Quoted from what is probably the most recent draft of Dr. Gidley’s manuscript, the only one in which & new genus is proposed. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 251 Simpson: Closely resembling Ectocion, but upper premolars with much smaller metacones, first and second molars with slightly smaller mesostyles and hypocones, protoconules of P*-4 and M!? slightly more united by lophs to protocone. Discussion.—The molars of this genus can hardly be distinguished generically from Ectocion, although unlike any known species in details. The premolars, however, are distinctly less molariform and at once distinguish Gidleyina from Ectocion. It is well known that from partial dentitions alone it is often difficult or impossible to determine even the ordinal affinities of genera in these ancient faunas, but in this case every indication is that Gidleyina is in fact related to Ectocion, and perhaps ancestral to it." If this is correct, it is clear that the Ectocion line was already distinct from that of Phenacodus in the Middle Paleocene and had already acquired a mesostyle and slightly more lophiodont pattern, although in other respects, such as the complication of the premolars, not more advanced than Tetraclaenodon.” Among Torrejon genera, Gidleyina most closely resembles Pro- toselene in many respects but is at once distinguished by the large and posterointernal protocone on P?, distinct conules on P*, and other lesser details, suggesting that the relationship is not very close. The dis- tinctions from Protoselene are resemblances to Ectocion. In addition to the type, based on an upper jaw, I tentatively refer two species based on lower jaws to this genus. They are described below. GIDLEYINA MONTANENSIS (Gidley) Figure 72 Gidleyina monianensts (GipLEY) Simpson, 1935d, p. 240. Type.—Princeton no. 12048, part of left maxilla with P°-M? and a probably associated right P?. Horizon and locality—bLoc. 68, about 1,000 feet above Gidley Quarry, Fort Union, Sweetgrass County, Mont.¥ il This is Dr. Gidley’s opinion in all three drafts of the manuscript on this form, but still later he noted that affinity is closer with Protogonodon. This seems to me highly improbable and was perhaps noted rather as a point to check than as a conclusion. 12 In one of his manuscripts Dr. Gidley proposed placing the Tetraclaenodon-Phenacodus and the Gidleyina- Ectocion phyla in different subfamilies. Even if we grant that the phyla were distinct from Middle Paleo- cene to lower Eocene, they are so similar that considering them as two subfamilies seems to me dispropor- tionate to the classification of other groups of mammals. 13 There are now no locality data with the specimen. One of Dr. Gidley’s manuscripts says ‘‘Near sec. 23, R.15E., T.5 N.. . near top of Fort Union No. 2 of Silberling.”’ Localities 4, 52, and 54 are the only ones in (or near) this section—the Gidley Quarry and a nearby exposure near the same level. As far as I can determine, no Princeton material came from anywhere near here. Another of Dr. Gidley’s drafts, and apparently the latest, says ‘“‘From the vicinity of Bear Butte. ... Exact level not known, but probably from near the middle of the section of this locality.” Mr. Silberling, however, remembers the discovery of the specimen and positively states that it was found at the locality now numbered 68. T'wo other speci- mens perhaps of this species are from the cluster of localities in the western part of T. 5 N., R. 15 E., where most of the Princeton specimens were found, some 1,500 feet above the base of No. 3. 252 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Diagnosis.—Gidley * ‘‘P? to M?=31-+ mm; M!'?=13.1 mm; length of M'=6.4 mm, greatest width=9.1 mm, width across hypocone and metacone=8.3 mm, greatest width of M?=9.8 mm, other measure- ments of this tooth same as those of M'; parastyle and mesostyle prominent, mesostyle angular and continuous with the ectoloph; P# with uninterrupted internal cingulum, and with low but well-defined lophs connecting the summit of the protocone with the protoconule and base of the metacone respectively.” FIGURE 72.—Gidleyina montanensis (Gidley), Princeton Univ. no. 12048, left upper jaw: a, External view; 6, crown view. One and one-half times natural size. Discussion.—The P? probably associated with the type is a simple 2-rooted tooth with one laterally compressed external cusp, somewhat anterior on the crown, and a small posterointernal expansion of the base but not true protocone. My measurements of the type are as follows: P*-M?, 19.6; M'?, 13.8; length P?, 4.5; width P?, 3.1; length P®, 5.9; width P’, 5.8; length Pt. 5.7;-width.P*,, 7.2; leneth M', 7:0: width M®, 9:0; length: IMe: 6.9; width M?, 9.9. Two other specimens from the same cluster of localities all at about the same level are in the Princeton collection. Princeton no. 14195 is an M? similar to that of the type but slightly larger and with the hypocone more internal. These could be individual variations. Princeton no. 14190 is an isolated P, 7.2 mm in length and 4.8 mm in width. Its size is almost exactly that of the corresponding tooth of 2G. silberlingi, but the protoconid and metaconid are closer to each other and the metaconid is relatively more posterior. 14 J quote Dr. Gidley’s diagnosis from the draft in which he placed the species in a new genus. His other two diagnoses view it as a species of Tetraclaenodon and of Ectocion, respectively, and are therefore inappro- priate. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 253 ?GIDLEYINA SILBERLINGI (Gidley) '5 Ficure 73 ?Gidleyina silberlingit (GiDLEY), Simpson, 1935d, p. 240. Type—U.S.N.M. no. 6166, partial left lower jaw with P;—Ms. Collected by A. C. Silberling. (In the same lot are a partial right lower jaw with M,_3 and another right lower jaw fragment with M, and the heel of M,;. They probably belong to the same species but include parts of one or two different individuals and are excluded from the type material.) Horizon and locality —tLoc. 27, about 400 feet above the base of Fort Union No. 3, Wheatland County, Mont.'® Diagnosis.—Gidley: ‘© ... About the size of or a little smailer than EL. minor [= Tetraclaenodon pliciferus, G. G. S.J... Jaw rela- tively long and slender, especially anteriorly ; the teeth proportionately narrow transversely ... with a decided tendency to selenodonty ... The paraconid in the molars is vestigial or wanting, and P, is submolariform . . . the heel . . . having the crescentic form of that of the molars, while the metaconid is large and as high as the protoconid.” Discussion.—It is possible that this is the lower dentition of Gid- leyina montanensis. Since, however, it cannot be demonstrated to belong even to this genus and since among lower dentitions it is a distinctive and interesting type that requires some means of reference until its association with upper teeth can be established, it seems quite proper to accept Dr. Gidley’s decision to define it as a species, which can be reduced to synonymy later, if necessary, with no great con- fusion. In comparison with other known lower jaws, this is generically dis- tinct from any previously described. ctocion is similar but has a simple longitudinal crest on P3, instead of an incipient crescent, while P, is more complicated and molariform, with a distinct posterointernal cusp absent in the present specimen. The molars offer no contrast definitely of generic value, unless it be the somewhat larger and more definitely closed trigonid basins and less distinct vestigial paraconids of ?Gidleyina silberlingi. The possibility that Gidleyina is not really ancestral to Ectocion or, on the other hand, that ?G. silberlingi does not belong to Gidleyina is enhanced by the fact that whereas the upper 15In a draft of the manuscript on this family, Dr. Gidley describes this as a species of Euprotogonia (= Tetraclaenodon). Onthespecimen label he has crossed out ‘‘Euprotogonia”’ and written ‘‘Ectocion.’’ It is thus evident that he recognized the probable relationship of this jaw to the new genus I have named Gid- leyina, since this was also successively identified by Dr. Gidley as Huprotogonia and as Ectocion before its generic distinction was recognized. I have not quoted his diagnosis in full, giving only enough to validate his claim to authorship of the species, since it was written before he had recognized the genus here named Gidleyina and therefore is not fully apropos. My comparisons following the diagnosis suffice for the expres- sion of more fully studied opinion as to diagnosis and affinities. 16 Given on labels, etc., as “Sweetgrass County’’, but, as can be seen on the map, this is one of several localities slightly north of the county line. 254 BULLETIN 169, UNITED STATES NATIONAL MUSEUM teeth of Gidleyina montanensis are almost ideally prototypal to those of Ectocion, the lower teeth called ?G. silberlingt seem to be progressing either more rapidly or in a different direction in the development of P; and the molar trigonids. This is not certain, however, as these characters are highly variable and an apparent reversion of this sort is not inconceivable. FIGURE 73.—?Gidleyina silberlingi (Gidley), U.S.N.M. no. 6166, left lower jaw: a, Crown view; 5, external view. One and one-half times natural size. P, differs markedly from that of Tetraclaenodon in its incipient cres- cent and basin, but P, is basically similar. The melar paraconid is much less distinct in ?G@. silberlingi than in most specimens of Tetra- claenodon, although approached by a few extreme variants of the latter, and the enamel is much less rugose, the crests less crenulated. These characters suggest Protoselene, but in the latter even P, is much less molariform, with the metaconid strong in ?G. silberlingi, barely incipient at best and the talonid very different. The following measurements are from the type: Length P3, 6.7; width P;, 3.9; length Py, 7.2; width Py, 4.7; length My, 7.0; width My, 5.4; length Mo, 7.3; width Mg, 5.4; length Mg, 7.3; width, Ms, 4.6. ?2?GIDLEYINA SUPERIOR (Simpson) Figure 74 ?Tetraclaenodon superior Simpson, 1935d, p. 239. Type.—U.S.N.M. no. 11913, part of left lower jaw with talonid of M,, Me, and M; still in capsule. Collected by A. C. Silberling.” Horizon and locality.—Loc. 11 or 13, about 3,000 feet above the base of Fort Union No. 3, Sweetgrass County, Mont. Diagnosis.—Lower molars with paraconids vestigial, broad trigonid basins with crenulated anterior margin, crenulations otherwise slight. 17 The two upper molars apparently of Tetraclaenodon, discussed on a previous page, are preserved in the same lot of material but are not associated with the lower jaw and were definitely excluded from the type material of this species. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 2p External cingulum absent, talonids incipiently lophoid. Talonid of M, markedly narrower than trigonid. Lower molars of about the size of those of Tetraclaenodon symbolicus but slightly narrower relatively. Somewhat longer and distinctly wider than those of ?G. silberlingi. M, length 7.7, trigonid width 6.2, talonid width 5.5. FIGURE 74.—?Gidleyina superior (Simpson), U.S.N.M. no. ii913, left lower jaw: a, Crown view; 6, internal view. One and one-half times natural size. Remarks.—This distinctive but imperfectly known species was at first referred, with a query, to Tetraclaenodon, but with the comment that it might belong to Gidleyina. The crucial evidence of the pre- molars is lacking, but after further study it seems probable that it is congeneric with ?Gidleyina silberlingi. The relatively slight enamel crenulation, the structure of the trigonids, and the incipiently lophio- dont talonids are distinctions from species surely referred to Tetra- claenodon and points of resemblance to ?G@. silberlingi. Reference to Gidleyina depends on that of the last-named species, discussed above. Family PERIPTYCHIDAE Cope, 1882 Subfamily ANISONCHINAE Osborn and Earle, 1895 Anisonchines are among the commonest fossils in the Puerco and Torrejon, but in the present fauna they are neither abundant nor varied. One form cannot at present be distinguished from the Torrejon species Anisonchus sectorius. Only one other form, Cori- phagus montanus, is recognized. This genus also occurs in the Torre- jon, but the species is distinct. The recognition that MJizoclaenus is a synonym of Coriphagus and that these animals are not oxy- claenids, as generally supposed, but primitive anisonchines clears up a decided taxonomic anomaly and also casts important light on the origin and affinities of the Anisonchinae. The five genera of this subfamily so far distinguished may be recog- nized by criteria presented in table 59. BULLETIN 169, UNITED STATES NATIONAL MUSEUM 256 ‘od ‘od ‘od *9m00 4ouTY -sIP 8 plqnuoo -odAYq = ‘payesuolq *peqye -queloytp A,i00d Pplfnuovod 4y ‘yous SW JO [90H “fry UO juesqe JO 9uIes “jy UO UvIpeur -qns pue |{[vus “ay uo yours ‘uRIpetmqns ‘youry “SIP JQ = [BISTYSOA “URIpPOUL -qns “jourystp AT[e -1oues yNq [RIZTISOA tr gate ie ee yuosq vy *[eu104 “ul Ajiwou ‘4ourySsTC. spluoovieg *pautseq JOU JO poulseq fq JO yey} ‘pourseq +g Jo 4yeyy ‘ezeiapou 30 [[eus speey ‘soyndsno JOl1ojUB OU JO [[BUIS ‘ayesuoja 10 snoqing ‘you ATV -Ioues tg 4nq poulseq Ajjeioues =Fq ‘[[eurs sjeey ‘serndsnd 101104 -uB OU IO [[vUs ‘snoqing “pourseq A[[ei9ues sypaoy eyeviepotm ‘serndsno IOLIOJUB VdIvT ‘oYVSUO[T ‘sao poulseq 3uoijs ‘sojndsno Jof1ejueB OU ‘eyBsUOLy *s[ooy poulseq o}yviepou ‘sajndsnod 101104ue OINUlU QIK ‘eVesuO[y ¥-¢ d “19448 Surdojaaua mnnsui uo 4urod se pedoyeaep ‘au0004 -O1d 04 Jol1eysod uvyy [BUJO}UI o10m BUOD0d 4 FY] *‘Tenbeqns ‘*pooeryd Ay[BonyjetmmAs = ‘yewurg *9m090401d 04 [BU19}UIOI8}Ssod ‘Suo01s euo000dsAy ‘4ueS -qB 1O [[vuUIs o[A}sojOIg “9u0o -0yOI1d 0} [eUIOZUIO -Ia4sod ‘8u013s ouo000d. -AW ‘[[euls 93[44s0}02g *10}}B] SUICOTOA -ua A[iveu wnyNsurd uo jurod se pedojaaep ‘au000j01d 0} [eUI0Z -ulolejsod + emuo000d A ‘e[Aysojyoid yOULSID ON suov0d Ay pue o[f4s0}01g ‘OJVUISIVUe PUB aye[nsue AT}USI[S 9p -10q Jojno ‘ayvipenb -qnis 04 Jepnsuerd} AT -onbilqo ‘A[asIOASUBI I, “sap -10q 1J0jno popunod ‘TBoljoumids ‘as loa -SuBIy AlOJBIBPO TL “SOP -10q Jojyno pepunol ‘ayeipenbqns ‘asiaA “Suvi, A[OYVIOPO PW “slop -10q J9jno pepunod ‘repnsueny Arend, -qo ‘aSIOASURIY sso’ *"S1op1IOg 104NO 9}eUIZ -Ivula pue oyepnsue ya ‘ayeipenbqns ‘OSIOASUBIY A[PIYIC TY *gu000j01d [Bo -TMOOqNS O5IBl UTA *snuov0jd vy] Ut se dsno s0yno 04 petjdde Ajasofo os you ‘9u000401d 91} -U9dS81D OSIPT YIM ‘auooryd -me 9} SurIdo[aa -ua pure 0} pod -de ouo000jo1d 014 -W99S0I0 OSI8T UTM “9u00 -0j01d [eormooqns TTBUIS JOULISIP TLM “909 -0301d 4noq}IM |----7~ unpooDUud) Sas Sana Op-~~77|-- 7 "snanpypumary *9U000} -O1d Bu0IjS UIA "9100 -0401d JMOUyTM “‘{UOTILPNI 9u0904 -O1d OJNUIU ALAA eae snyouosiup aa snuoo0)dpyy SE Be snboydisog 1d apuryouosiup fo Diauab aayf ur saajavunyo uoyuuap fo Uuos1pdWwoj— 6G ATAV J, ed snuen FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. DASE Genus CORIPHAGUS Douglass Cortphagus Dovatass, 1908, p. 17. Mizoclaenus MATTHEW and GRANGER, 1921, p. 7. Type.—Coriphagus montanus Douglass. Type of Mixoclaenus.—Mizoclaenus encinensis Matthew and Granger. Distribution — Middle Paleocene, Fort Union, Montana, and Torre- jon, New Mexico. Diagnosis.—P, 1-rooted. P3-, subequal, somewhat swollen but elongate, with minute anterior basal cusps and small heels basined posterointernally. Molars relatively small, trigonids larger than talo- nids and notably higher, paraconids distinct and nearly internal, trigonids basined with cusps crested, poorly differentiated, and not conical. M,; much reduced, with hypoconulid distinguishable but not prominent. P® with very rudimentary protocone, P* with dis- tinct but small subconical protocone, M!~? transverse, subquadrate, outer borders angulate and emarginate, distinct hypocones postero- internal to protocones and on cingula nearly enveloping the latter. Discussion.—Douglass based this genus and its type species on a single but unusually complete lower jaw with P,-M3, found by Silber- ling in the Silberling Quarry. In 1913 and 1916 parties under Dr. Granger found three * specimens-of a similar form in the Torrejon, and in 1921 Matthew and Granger described these as Mizoclaenus. They then noted the resemblance of Mizoclaenus to Coriphagus but cannot have realized, from Douglass’ somewhat schematic figure, how close itis. They decided to hold AfMizoclaenus as distinct at least until discovery of the upper dentition of Coriphagus. Matthew’s fuller description of Mizoclaenus (Pale. Mem.) was written before 1921 (probably in 1917) and was not corrected. It does not mention the resemblance to Coriphagus. The upper dentition of Coriphagus montanus is now partly known, and it has been possible to compare original specimens of that species and of Mizoclaenus encinensis. The conclusion is that the two species are unquestionably congeneric and hence that Mizoclaenus is a synonym of Coriphagus. They compare very closely in every known part, and the type species of the two supposed genera differ only in size and doubtfully in slight variations of proportions. Douglass (1908) referred Coriphagus to the ?Insectivora, without family assignment. Matthew and Granger (1921) placed “Mizo- claenus’”’ in the Oxyclaenidae but noted resemblances to Mioclaeninae and Anisonchinae.” In his longer work Matthew (Pale. Mem.) has 18 Matthew mentions four, describing only two of them, but I can only find three in the collection. 19 Hay (1930) followed them in placing Mizoclaenus (which he wrongly ascribed to the Tiffany) in the Oxyclaenidae, but he placed Coriphagus in the Plagiomenidae, a family with which it has practically nothing in common. Schlosser (1923) placed Mirociaenus (which he wrongly ascribed to the Puerco) in the Oxy- claenidae and Coriphagus in the Mioclaenidae. These and other casual references are accompanied by no evidence and require no discussion. 258 BULLETIN 169, UNITED STATES NATIONAL MUSEUM outlined the evidence in more detail. (He is discussing Mizoclaenus, which now proves to be Coriphagus.) The only oxyclaenid character given is “upper molars resembling those of Chriacus’’, but he adds that they are wider transversely, more triangular, external angles more prominent,” hypocone less so, and M® much reduced and more transverse. With these, and other modifications, the resemblance to Chriacus is really quite attenuated. Matthew notes that the rounded condyle and other characters of the jaw and the small premolariform canine are not oxyclaenid but do not approach condylarths or in- sectivores. I add that they do, almost to identity, approach the Anisonchinae. Matthew also notes, but rejects as inconclusive, some resemblance to Didelphodus, leptictids, and Palaeosinopa in the molars, but adds that the premolars suggest the Mioclaeninae but are more like the Anisonchinae. This genus has, in fact, all the diagnostic characters of the Anison- chinae and nothing that decisively indicates pertinence to any other group. ‘The upper and lower premolars are of fully anisonchine type and are especially suggestive of Conacodon cophater.2!_ They differ in such details, well within the morphological range of the Anison- chinae generally, as the incipient development of a protocone on P? and the less transverse P*. This last tooth is intermediate between the “round premolar cusp” type (Hemithlaeus and Conacodon) and the “flat premolar cusp” type (Haploconus and Anisonchus), adding to the evidence already given by Matthew that these are not, as Osborn and Earle thought, major phyletic divisions of the Anison- chinae. The lower premolars still more closely resemble those of Conacodon cophater, the only definite differences being that they are slightly less inflated and have the anterior basal cuspule a little smaller (but larger than in Conacodon entoconus). Matthew (Pale. Mem.) mentions the heavy and peculiar wear on these teeth in Coriphagus (‘‘Mizxoclaenus’’), truncating them obliquely. This wear occurs in all Anisonchinae and is almost diagnostic of the group. The molars are on the whole more primitive or generalized than those of other anisonchines, which is what induced Matthew to refer the genus to the Oxyclaenidae. Yet they have the basic anisonchine characters. The upper molars markedly resemble those of Anison- chus gillianus and Conacodon cophater, apparently the most primitive in this respect among other anisonchines. From the former they differ chiefly in the less rounded outer contour, shorter internal slope, and development of the hypocone on a cingulum around the protocone. The first and last of these characters are resemblances to Conacodon cophater in which, however, the internal slope is also long and the hypo- cone is more internal, with respect to the protocone, than in Cor- 20 This does not seem to me to be quite certain. 21 Conacodon entoconus differs greatly from the smaller species and might almost be distinguished ‘gin it generically. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 259 phagus. Mis markedly reduced and very transverse, a resemblance to Conacodon cophater but here somewhat intensified. In the lower molars the large swollen trigonids with the tips of the cusps pinched together, giving an aspect difficult to describe but characteristic when seen, is diagnostic of Anisonchinae and typically developed in Coriphagus. The paraconids are retained and are closely similar to those of Anisonchus gillianus, possibly a trifle more internal, but hardly more so than in some specimens of that species; in A. sec- torius they are more definitely median. The molar talonids are rela- tively smaller and their cusps less conical and distinct, especially on M;. This is perhaps the most aberrant feature of Coriphagus, con- sidered as an anisonchine, but the difference from such a form as Conacodon cophater is really slight. The characters of the mandible mentioned or shown by Matthew are almost identical with those of other anisonchines about the same size. Coriphagus is in many respects the most primitive known anison- chine, representing, in view of its age, an unprogressive surviving type. Its closest comparisons are with Conacodon cophater and Anisonchus gillianus, both Puerco species and older than the known species of Coriphagus. Coriphagus carries still closer the marked resemblance already noted between the anisonchine and the hyopsodontid denti- tions. Were no other anisonchines known, it could very well be classed as a hyopsodontid representing another incipiently divergent line in addition to the several already known in that group. But all these divergent characters are in the direction of the more specialized anison- chines, and these in turn show marked resemblance to the still more specialized periptychines. The whole hyopsodontid—periptychid com- plex seems to bear the definite stamp of divergence from a common ancestry. CORIPHAGUS MONTANUS Douglass Figures 75, 76 Coriphagus montanus Dovatass, 1908, p. 17. Type.—Carnegie Mus. no. 1669, left lower jaw with P,-M;3. Col- lected by A. C. Silberling. Horizon and locality —Type from Silberling Quarry, several referred specimens from Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. Diagnosis.—Smalier than C. encinensis (see measurements). Lower teeth relatively narrow. M, less reduced relative to other teeth. Discussion.—The differences in proportions given are not entirely certain, since they depend on only one specimen of C. encinensis, and these characters are variable. The size difference is slight and in itself might not warrant full specific status, but its constant association with different provenience makes it certainly significant. Every di- 260 BULLETIN 169, UNITED STATES NATIONAL MUSEUM mension of all three specimens of C. encinensis is larger than the cor- responding dimension of any of the seven available specimens of C. montanus. The best single comparison is of the length of M,, which compares as follows, by Fisher’s formula: LMi, C. montanus: N 6, M 3.25) 3 TUNE AG eicnicn ee INO SMO San eee Sy, iG Z Aa? = FIGURE 75.—Coriphagus montanus Douglass, U.S.N.M. no. 9334, with parts in outline supplied from U.S.N.M. nos. 9599 and 9685, left lower jaw: a, Crown view; 0, internal view. Four times natural size. The difference is certainly significant although not great.” Table 60 gives ranges and means for the National Museum speci- mens of C. montanus and corresponding dimensions of the paratype of C. encinensis, Amer. Mus. no. 17074. The material is not sufficiently abundant for the calculation of other statistical constants. TaBLE 60.—Numerical daia on lower dentition of Coriphagus montanus and C. encinensis C. montanus Dimension 71RD cea Nera N R M TUR yc ee eh Ee ie ees 3 2. 7-2. 8 Ded 3.2 AVE he icy pee ER 3 1. 4-1.7 Linas Zeal: | Wy ey pees eyo eerie eee ae 2, 2. 3-2.'°5 2. 40 on AW) Digits Seek ene st Ber eR 2 1. 4-1. 6 1. 50 223 1 Del eee ge Bye yea rey ces 6 3. 1-3. 5 3: 25 Said NV IN Te Ae oe DES 6 222 DEAS 2.8 1 BY cae RS rap ne Si to 6 2. 5-2. 9 Ete, 3. 4 Walger ec Sak Fae. ey 6 2. 0-2. 4 2.18 2.8 TE IN IGRY Sey) iS AD Rests tones 3 2. 3-2. 8 2. 60 2.9 AAI les eee te pe ti on en 3 1. 7-1. 9 1. 80 22, ENT yeah ee ss a oe ite 2 8. 2-8. 7 8. 45 9.9 32 The mean of C. encinensis is only 14 percent greater than for C. montanus, and it is to be stressed that this is not in itself and stated in this way of specific value. The largest specimen of C. montanus is 13 percent greater than the smallest, and the paratype of C. encinensis is only 6 percent larger than the largest of our specimens of C. montanus in this dimension. With large series the two species would doubtless intergrade in size, yet they are certainly distinct. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 261 FIGURE 76.—Coriphagus montanus Douglass, U.S.N.M. no. 9591, left upper jaw, crown view. Four times natural size. Ala GZS a i WR beh A > ig TNC 1 BNO Rg ale (cali AUD TAS IN We AN = ¥ eases FIGURE 77.—Anisonchus sectortus (Cope), referred specimen from the Lebo, U.S.N.M. no. 9267, right lower jaw: a, Crown view; 6, external view. Twice natural size. : S a —£ SS fIGURE 78.—Anisonehus sectorius (Cope), referred specimen from the Lebo, U.S.N.M. no. 9263, right upper jaw: a, External view; 6, crown view. Twice natural size. 262 BULLETIN 169, UNITED STATES NATIONAL MUSEUM U.S.N.M. no. 9591 is a left upper jaw with P?-M! and about half of M?, from the Gidley Quarry, which is certainly referable to Cori- phagus montanus. It is deeply worn but well shows the generic characters. P? may be slightly less transverse and the external cingulum of P* weaker than in C. encinensis, and all the teeth are slightly smaller, but the agreement in structure is very close. Genus ANISONCHUS Cope, 1881 ANISONCHUS SECTORIUS (Cope, 1881) Fiaures 77-79 Douglass (1902b, p. 222) described and figured an Anisonchus that he compared with A. sectorius but mentioned the possibility that it might be distinct. The National Museum collection includes a series of excellent specimens of this genus, and their pertinence to Anisonchus sectorius can be rather positively established, although they may well pertain to local races as will be pointed out.” Anisonchus is the only genus represented by good material in this fauna that seems to be represented here by the same species as that occurring in the Torrejon. It therefore is a special point of attack for considering the relationships of these two widely separated Middle Paleocene deposits, and the material has been subjected to detailed and lengthy analysis. The results are not entirely conclusive, largely owing to the small size of the available pure samples, but they never- theless are of considerable interest, and they also provide data that must be useful in future work. The full analysis would fill many pages with numerical and morphological data and calculations, and so it is not published here in extenso, but only such figures as are most necessary to illustrate the general conclusions reached. A study was first made of the Torrejon specimens themselves to see whether more than one species or race could be distinguished, par- ticular attention being paid to possible distinction between material from the two principal fossil levels of the Torrejon. The results of this analysis were negative: From the data at hand it is not possible to subdivide the Torrejon material, all of which is referable to Anison- chus sectorius.4 Despite considerable variation, there is only one specimen, Amer. Mus. no. 3533, that stands out as strongly aberrant. It was collected by Baldwin in 1885, and the exact horizon and locality are not recorded. Even this specimen, however, is so close to typical A, sectorius that it would be methodologically incorrect to discard it from the general sample. 2 Labels show that Dr. Gidley referred some of the Fort Union specimens to Anisonchus sectorius and some to a new species, but he left no diagnosis or discussion. I have carefully endeavored to visualize his concept of the new species, thinking that it might correspond with one of the inconclusively indicated local races, but this does not seem to be the case, and I am unable to ascertain the characters relied on by him. 2 It may be noted, however, that Matthew is incorrect in believing one of the cotypes of A. mandibularis to belong to Anisonchus sectorius. Whatever the position of this very dubious species, it is not a synonym of A. sectorius and probably does not belong in Anisonchus. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 263 The general statistical data on the simple dimensions of the lower teeth of all Torrejon specimens of Anisonchus sectorius in the American Museum are presented in table 61. TABLE 61.—Numerical data on lower dentition of Anisonchus sectorius Variate N R M o V TEP ON UT 2) 8 5. 8-6.8 | 6. 2840.11] 0.30+0,08 4,841.2 1 a 8 2A 1 (pS 22-00 14 0. 40+-0. 10 1D. OFt Sad. 11 CE aes he ee imi ile¢ 5. 3-6. 6 | 5. 86+0. 07 0. 30+0. 05 5. 2+0. 9 Wikis ond fy 17 3.5-4.3 | 3.8840.05 | 0. 22+0. 04 5. 74£1.0 Eee 17 5. 0-5. 8 | 5.4740. 05 0. 22+0. 04 ey WEN eee 17 3. 5-4. 2 | 3. 87+0. 04 0. 17+0. 03 4.5+0.8 Magee eh 21 4,8-5.6 | 5.3440.04| 0.20+40. 03 3. 8+0.6 WM... ih! 23 3. 7-4.4 | 4.0640.03 | 0.1740. 02 4,140.6 eee 4. 14 5.1-6.1 | 5.58+0.08| 0. 28+40.05 5. 041.0 WM S5i 6222. 14 3.2-4.0 | 3.544005 | 0.2040, 04 5.741. 1 The lower dentition material from the Crazy Mountain Field now available consists of three specimens from the Gidley Quarry, three from Loc. 25, four from Loc. 51, and one each from Locs. 50 and 18. The highest number of comparable specimens from a single horizon and locality is only three, for P, from the Gidley Quarry. bh .o) ol pats ‘ J Number of individuals 4.9 Bi2 fH iSSii! SB Ga LENGTH Mi FiGuReE 79.—Histogram of the length of Mi in Anisonchus sectorius (Cope): In solid outline, Torrejon specie mens (American Museum); in dotted outline, Lebo specimens (National Museum). By inspection of the specimens and of their dimensions, it is sug- gested that there is local differentiation. Thus in the series from the Gidley Quarry, Loc. 25, and Loc. 51, each local sample seems to be reasonably homogeneous and to differ slightly from that of the other two localities. In the jaws from the Gidley Quarry the lower molars are longer, but not noticeably wider, than those from Loc. 25. The very poor material from Loc. 51 suggests closer agreement with the Gidley Quarry at least in P, and M,, but two isolated M,’s from there 1192123718 264 BULLETIN 169, UNITED STATES NATIONAL MUSEUM are closer to those from Loc. 25. This impression of slight heterogene- ity cannot now be considered dependable or formally recognized. The differences could hardly be of more than subspecific scope in any event and perhaps are merely those of separate strains within one subspecies, and the data are too few to establish them as real. The best data for material from one locality, dimensions of P, and M, from the Gidley Quarry, may be compared with the Torrejon sample by Fisher’s t-test, previously mentioned, with the following results: VARIATE N,+ N32 t P LP, 20 9 99. <0.05>0.02 WP, 20 1.60 0:1 LM; 19 2.83 <0.02>0.01 WM; 19 74 <0;2--051 Thus these two teeth are probably significantly longer, but not wider, than those of the Torrejon sample as a whole. Since the latter is heterogeneous in origin and perhaps as to race, it does not neces- sarily follow that the Gidley Quarry race does not occur in the Torre- jon, but it probably does not. Comparison of the whole Fort Union sample, however, shows no significant difference, as the following figures for the only variates probably significant in the Gidley Quarry sample show: VARIATE N,+ Nz t Pe LP, 25 1.95 <0.1>0.05 LM, 22 1.66 << 0,201 In short, the evidence now is that the Fort Union sample may include more than one local or temporal genetic group of minor scope and the same may be true of the Torrejon material. At least one of these minor groups in the Fort Union is distinct from the Torrejon sample as a whole and probably from any group included in the latter. But the definitive separation of these minor groups can- not be accomplished from the data now available, and there is no significant difference between the Fort Union Anisonchus as a whole and that of the Torrejon as a whole. All are referable to a single species, A. sectorius. TaBLe 62.—Numerical data on P, of Anisonchus sectorius Variate N R M o V 5. 3-6. 6 5. 94+0. 05 0. 3140. 04 5SZ2O7T 5-4. 3 3. 920. 04 0. 20+ 0. 03 5. 2+£0. 7 FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 265 As an indication of the variation of the whole species as thus known, data for Py, a rather variable and characteristic tooth best represented in the combined collections, are here presented (table 62). Individual measurements of the many specimens studied, and many other data, are on hand and will be permanently filed. The uniformity of these results with those based on Torrejon specimens only is striking. Despite the great increase (one and a half times) in the size of the sample by the addition of specimens of widely different provenience, the two means are increased by only 0.08 and 0.04, respectively, the first figure only 0.01 more than the corresponding standard error for the smaller sample and the second less than the corresponding standard error. The standard deviations are altered by amounts considerably less than their standard errors in the smaller sample and the same is true of one coefficient of variation, while the other is not changed at all. The interest of the coefficients of variation for the Torrejon sample should also be pointed out. One of them, for WP; is unusually high, but this is largely caused by the single aberrant or abnormal specimen previously mentioned.”®> If we accept P3 as abnormally variable or as represented by some extraneous material, the other eight coefficients of variation range from 3.8 to 5.7 and average 4.8. The accumula- tion of such figures is of great importance in view of our almost com- plete lack of any exact knowledge of the variability of fossil species in samples collected under the usual field conditions. The preceding discussion is based on lower teeth. The upper dentitions have also all been examined and compared, but they merely substantiate the evidence of the lower dentitions, and the samples are less satisfactory in all respects. Order PANTODONTA Cope, 1873 (as suborder), new usage The order Dinocerata (emended from Dinocerea) was proposed for the uintatheres by Marsh in 1872. In 1873 Cope proposed to reduce this to subordinal rank and with the new suborder Pantodonta, for the coryphodonts, placed it in the order Proboscidea. In 1875 the two suborders were transferred by Cope to a new order Amblypoda. In 1883 he added to this order the suborder Taligrada, based solely on Pantolambda, and in 1897, as stated on an earlier page, added the periptychids to the Taligrada. Marsh in 1884 proposed ‘“‘Ambly- dactyla” and “Coryphodontia” as strict synonyms of Cope’s names (which Marsh claimed to be essentially preoccupied) Amblypoda and 35 Omission of this specimen would reduce V from 12.3 to 7.5+2.0. 3% This is within twice the standard error (that is, within the range of probable true values) for all the ‘single coefficients and is within less than the standard error for all but two. If the aberrant individual be omitted, it is also well within the range of probable true values for P3. 266 BULLETIN 169, UNITED STATES NATIONAL MUSEUM Pantodonta, respectively, but these names have not been generally accepted. The classification thus achieved is well summed up by Osborn (1898): Order AMBLYPODA: Suborder Taligrada: Periptychidae. Pantolambdidae. Suborder Pantodonta: Coryphodontidae. Suborder Dinocerata: Bathyopsidae. Uintatheriidae. This arrangement is now classic and with slight modifications has since come into all but universal use. Nevertheless, in the light of later discovery and research, it has little to recommend it. The probable affinities of the periptychids with the condylarths, rather than with the pantolambdids, have been discussed on a pre- vious page. On the other hand, all recent work (see especially Simp- son, 1929d, and Patterson, 1934) tends to emphasize the essential unity of Pantolambda and Coryphodon and their respective allies. The known pantolambdids are not ancestral to the known corypho- donts, and family separation is warranted, but they are so simular in structure aside from primitive or progressive features generally cor- related with greater or lesser age that there seems no reason to place them in separate suborders, and the distinction between Taligrada and Pantodonta is unwarranted. The uintatheres, on the contrary (Simpson, 1929d and elsewhere), seem to be a group independent of the pantolambdids and cory- phodonts from a very remote time and linked to them only through a prot- or perhaps even pre-ungulate, non-“amblypod” ancestry. The classic arrangement was undoubtedly influenced by the belief that taligrades, pantodonts, and dinoceratans represented offshoots of a single stock appearing successively in time with correspondingly progressive specializations. Now it is clear that this simple picture does not correspond to the facts. Among the supposed ‘‘taligrade’’ periptychids the more advanced members are the only ones that show any considerable resemblance to the pantolambdids in foot structure, but they cannot possibly be ancestral to the latter not only because they are contemporaneous but also because aside from the feet (and in part including them) their structure is very different. The idea of successive offshoots does apply to the pantolambdids and coryphodonts, but it decidedly breaks down again with the cory- phodonts and uintatheres because these groups are not successive but contemporary phyla, and it is the latest and most advanced members of each that show some resemblance, which hence is only convergent, and the earlier members are even more decisively dissimilar. FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 267 The classification seems, in fact, to have been based on grades of foot specialization, over the prototypal ungulate foot, rather than on characters peculiar to natural phyla. It thus united periptychids and pantolambdids because they had advanced relatively little in limb structure, even though their phyletic relations must have been distant. It separated (subordinally) pantolambdids and cory- phodonts because the former were less and the latter more advanced in limb specilization, despite the clear evidence that these groups are closely related phyletically. And finally it united coryphodonts and uintatheres because both have specialized limb structures, but over- looked their marked phyletic separation. The revision of nomenclature for the accurate representation of these newer points of view is difficult. The following diagram shows the relationships of the classic names to the groupings now considered natural: @ondylarthra®: 2-222. - eee ; = | FAM DLy DOCas2 2 ene eae eee fi rs Taligrads. 2222 2* 30 e As originally proposed. iPantodonta s--eoesee= = ; Dinoceratas-.--sseceeoe—e Condyilarthra=- == = 5-24. =-- AM Dy pOGRss-sesee seen =e F = As subsequently modified by alioradas 225 [sets ete | 4 t Cope and accepted by most Pantodonta==---+-s.e.—e 1 other authors. Dinoceratas=----2— eee Condylarthral= 225. = i > Pantodontass ate seoo eae sen Dinocerata-. is. 502 _ Arve a, As here modified. 8 3 as EF aaa g Bras ae os oO Lo} = ago ea ie al Mec =a Ue ° | g g ES go 5 a a3 | am ios is 3 a Sw 3 I | e A in| aigrt | neolne: olds Since both periptychines (‘‘Catathlaeus’’) and anisonchines (Anison- chus) were explicitly cited among the genera belonging to the Con- dylarthra when that group was first proposed, since most definitions of Condylarthra need little or no alteration to include the periptychids, and since no other ordinal or subordinal name has been based primarily on the periptychids, the removal of the Periptychidae to the Condy- larthra raises no nomenclatural problem. 268 BULLETIN 169, UNITED STATES NATIONAL MUSEUM The name Dinocerata was originally based on the uintatheres alone and has always been taken as referring to them and including only them,” so there can be no question as to the propriety of continuing this name in this usage. What remains is to settle on a name for the Pantolambdidae and Coryphodontidae. Some students (including me, 1931) have used “‘Amblypoda” essentially in this sense, usually including the Perip- tychidae. There is good precedent for such a restriction of a name to one of several groups formerly included in it, and it is generally more advisable than coining a new name. In the present case, however, it should be avoided if possible. “‘Amblypoda”’ was based about equally on the coryphodonts and the uintatheres. To exclude the uintatheres from it is not quite the removal of the type group but certainly is a radical change in usage and one not well justified. The name “Taligrada” might be expanded to this usage, but this also is objectionable. As originally defined it was carefully drawn so as to exclude and contrast with the best-known members of the group for which a name is now sought, that is, the coryphodonts. Further- more, in the past 40 years it has almost invariably been taken to include or even to be typified by the periptychids. “Pantodonta” has none of these objections. It was proposed and has always been used for typical members of the group now in question. No animals foreign to this group have ever been called pantodonts. Its original definition, although brief, offers a good contrast with both Condylarthra and Dinocerata, even as those groups are now under- stood, and would include the pantolambdids (not known when the name was proposed), so that we are using the name exactly in the sense of the original author, in fact more so than he did later. The fact that a group that he later excluded from the Pantodonta is now included seems to be of no particular importance, especially as his original conception is not thereby changed. It is entirely proper in taxonomy to extend a name formerly applied to one group to include another later found to be closely related and is open to much less question than would be the exclusion from a named group of a sub- division on which it was originally largely based (as in excluding uintatheres from the Amblypoda). The name “Amblypoda” I would discard altogether, as not pertaining to any group acceptable as natural or convenient in modern taxonomy. The present conception of the group Pantodonta may be summarized as follows: 27 Marsh did suggest synonymy with Cope’s broader ‘‘Amblypoda’’, but in fact nothing but the uinta- theres was meant to be included in the original description and Marsh later (1884) accepted this restriction, for he proposed ‘‘Amblydactyla” to replace ‘‘Amblypoda”’ and to include both coryphodonts and uin- tatheres, with only the latter listed as ‘‘Dinocerata.” FORT UNION OF CRAZY MOUNTAIN FIELD, MONT. 269 Order Pantoponta: A group of middle-sized to large, very archaic ungulates (or “subungulates’”’). The dentition is little or not reduced in number, and remains practically brachyodont, but exhibits a high degree of lophiodonty even in the earliest members. The primitive pattern is strongly selenodont but this is secondarily masked to some extent in later forms. The canines are apparently always large, and may develop into great tusks. Skull and jaws are generally massive with very powerful muscle attachments. The brain remains very small and primitive. The limbs are massive and strong, ambulatory to graviportal, and retain many primitive features such as separate radius and ulna, tibia and fibula, and five toes on each foot. Carpus and tarsus retain and strongly accentuate the alternating arrange- ment. Family Pantolambdidae: Relatively less advanced forms, with the teeth fully selenodont, the skull roof little or not flattened, the tail long and heavy, and other primitive characters. Subfamily Pantolambdinae: Lighter, more ambulatory types of smaller size and with astragalus still retaining some condylarth characters. Middle Paleocene. North America. Subfamily Barylambdinae: Heavy graviportal types of larger size, with fully specialized pantodont type of astragalus. Upper Paleocene. North America. Family Coryphodontidae: Highly specialized forms, large in size, all with graviportal limbs, teeth with primitive crescents considerably modified, skull roof broad and flat, tail reduced. Upper Paleocene—lower Eocene in North America. Lower Eocene in Europe. Upper Eocene to Middle Oligocene in Mongolia. Family Pantolambdodontidae: A somewhat dubious group known from lower jaws only, which suggest relationship with Pantolambda but have numerous differences in details. Upper Eocene. Mongolia. The present fauna contains few remains of pantodonts, but Panto- lambda is represented by various fragmentary specimens, some of which indicate a species first defined, and at present known only, from this fauna.” Family PANTOLAMBDIDAE Cope, 1853 Genus PANTOLAMBDA Cope, 1883 Douglass (1902b, p. 224) described and figured an upper premolar perhaps of this genus. It is about the size of P? of P. bathmodon, but differs somewhat in form, the main cusp being more central and the external margin less sharply notched. The National Museum ma- terials does not serve to define this form. The second Pantolambda mentioned by Douglass (1908, p. 24) probably belongs to the species defined below. 28 The present conception of the Pantodonta, especially as regards its essential unity as here defined, Owes much to Patterson’s discovery of complete skeletons of Barylambda and to his fine studies of them (Pat- terson, 1933, 1934, 1935, 1937). Although not from the Crazy Mountain Field, the type species and specimen of Titanoides was found in the Fort Union and named and described by Gidley (1917), and it was originally proposed to include a discussion of it in this work, but Patterson’s studies make this quite unnecessary. 270 BULLETIN 169, UNITED STATES NATIONAL MUSEUM PANTOLAMBDA INTERMEDIUS Simpson Figure 80 Pantolambda intermedius Simpson, 1935d, p. 244. Type.—U.S.N.M. No. 8384. Left lower jaw with M,_, and alveoli of C-P,, associated with symphysis fragment with right I,_. and alveoli of left I,_;. Collected by Dr. J. W. Gidley. Horizon and locality—Gidley Quarry, Fort Union, Middle Paleocene horizon, Crazy Mountain Field, Mont. A as a ¢ a - niga eh vo basret olaaaig, : txfara, ae | ee, ee: 4 die . F hte: a aa Hee Fig Sie id NF tag a Ti aks ae “i ¢.. Ath, fia‘ ¢ 1 , 5) : t1o-Las,, ne al eer Ree soared iobanaor ralaee phys wit oe bie ¥ a aA, Aims) We i alalend (3 ray a ses eniaay pa te ie Ton we = A SEs Ural) Fa) aa i aaa bee AOD tie ee “9 en es alt ; uf eet oe ‘ti a a soba: eit Ha tat: rela ae py ie oe st waar roa ie eis: i" agialy az pid La. oe Fe udt 410. gkiasd Tio bie ae ree . a at td o o) Ke BAS Dips ve 4 A BAL act road s ne wont) Hn, ae eet re = a Gh nD rey aie a [mish F sural "90H npr ‘Yara *) a th ue oh ae cs a ote a ae ies be a8, pie aa q ee fete se i Posts pho T yoltastlog egies od, Wie mo atts ani veh, Yon guittnte de "004 (00 (ils rate i, Be Pats SA. ab (lo He ue . PH nn Pa a cae =? Cis: ane Dy poe Si. Pre sks Seg BY Joy lo eae He ee: Oe at aie , f Oi: Aaehy no maeie vue : te ~t aS at at asi, cubally be nnd Feet oyhibclin 7 ache EA, Bree head sea she ag A) Rati nil ‘ial GP) en . Bt ATE? iif iy conte ABU 8 fi se aes oe ae ae git ye ie Tot ee had atelier ih Corey re caveat sete A ; ned cine a % tne rie shngstl Sen Wie eee | Dae aes MO es _ heey: [ae ae. Ane tocihyy tec ety Sei * eee ae a ae trina eae ‘Bboh Pat « PR, Ag Wh oF. Wie tah Si. v pertae ir, ba : Th Ue bia 2 169 PLATE 2 BULLETIN U.S. NATIONAL MUSEUM Joddn aie punoise10j Ul safeys puB seTOyspues fed SuvUIE}[V OY, “WOTJOOS JO OUT] YINOs Iwau ysaM Ayo VUNIXOIdde Suryoo] “wy OTH “N 9° ‘6 ‘00'T Avo “GT “USN 9 VL ‘Sg (09g UI BUOYSpUes ([ ‘ON WOTUL WO) OGa'T 1AMOT ‘7 *ydeisojoyd oy} Aq potoAod Bor’ BY} JO (YINOS) 4Jo] oY 07 A[ayeIpeurUT! [[TY Sty} JO eseq OY} AveU SI ‘ploy SITY) UL PuNoy Useq OAVY S[VULUIBUL YOIYM 7B JSOMOT OY] ‘GQ *0O'T ‘UWOISTATP YVY JO [voIdA} OUOJSpUS JequIOS JURASISeI B AQ poddea st puv ‘; ‘ON UOIUA 1104q 10 ‘OGeT IOMOT OY} AQ PoullOJ SI pUNOLsaI0J Yo UI [[IY OYT, “JoejJUo0d OqeT—-Iveg ey} uO AyoyVUTTXOIdde st ainjord ey} JO Javed reddn ssosoe Youeq sy} pue ‘Iveg oy} Jo 41ed L ‘9Z (00g ‘T Ww “LNOW ‘OQ13I4 NIVLNNOW AZVYD ‘SNOILYVWYHO4Y O83] GNV Yuva U.S. NATIONAL MUSEUM BULEEERIN 169. PEATIE:S FORT UNION GROUP, CRAZY MOUNTAIN FIELD, MONT, FOR EXPLANATION OF PLATE SEE OPPOSITE PAGE, PLATE 3 1, Air view, looking approximately south, with Sec. 33, T. 6 N., R. 16 E., near the middle of the picture (prominent but small isolated timber butte in this section). The meandering watercourse is Widdecombe Creek and _ is developed on the nonresistant Upper Lebo (Fort Union No. 2), as are the other low sodded areas through the central part of the picture. The main timbered ridge across the photograph is the northwest side of Bear Butte. The broken area between the small outlying butte and the patch of timber (on Bear Butte) farthest to the left, just beyond the road, is Loe. 5, one of Douglass’ two localities where mammals were first found in the Fort Union. Part of Lion Butte forms the skyline in the center and right parts of the picture. 2, Air view of the east side of the north end of Bear Butte, looking approximately south in See. 34, T.6 N., R. 16 E. The rimrock of Bear Butte, along the upper edge of the picture, is the basal Melville (No. 3) sandstone, and the rest of this area is all on the Upper Lebo (No. 2). The shale exposure in the coulee above the road in the upper left of the picture is Loe. 6, one of Douglass’ discovery sites. PLATE 4 1, Air view of the Gidley Quarry and vicinity, looking slightly north of east, the foreground in See. 28, T. 5 N., R. 15 E. The slope in the foreground, on which the quarry is visible, is on the Upper Lebo (No. 2), as is also the broad valley of the Widdecombe Creek in the upper left. The higher level in the upper and right parts of the picture, and along the horizon, is supported by the basal Melville (Fort Union No. 3) sandstone. It is typically marked by evergreen timber, but in the foreground this timber extends sparsely well down onto the Lebo, but only on talus derived from the Melville. Bear Butte is dimly visible along the left part of the horizon, and a small corner of Lion Butte is seen in the upper right corner. Between these, small hills upheld by isolated patches of the Melville may be seen. 2, Air view of the Gidley Quarry, a closer view, looking more to the north, of part of the foreground of fig. 1. The picture was taken in 1935 and shows an advanced stage in the American Museum quarrying operations. The approximate area covered by the U. 8S. National Museum work (and since filled in) is indicated by dotted lines. The larger section (A of Silberling’s notes) is on the far (north) side of the small coulee, and the smaller (B) on the near side. The original discovery was made in the coulee near the spot marked by the cross. U.S. NATIONAL MUSEUM BULLETIN 169 PLATE 4 FORT UNION GROUP, CRAZY MOUNTAIN FIELD, MONT. FOR EXPLANATION OF PLATE SEE OPPOSITE PAGE. U.S. NATIONAL MUSEUM BULLETIN 169" (PEATESS FORT UNION GROUP, CRAZY MOUNTAIN FIELD, MONT. FOR EXPLANATION OF PLATE SEE OPPOSITE PAGE. PLATE 5 1, Air view, looking approximately west-northwest, the foreground in Sec. 4, T. 5 N., R.16 E. The small drainage basin in the foreground is on the Upper Lebo (Fort Union No. 2). The Silberling Quarry is in the upper part of the main right (northwest) branch of the coulee. The sparse timber marks the basal Melville (No. 3) sandstone and the top of Bear Butte, which here, near the middle of its length, is at its narrowest point. Beyond this is the valley of Widdecombe Creek, hidden by Bear Butte, and beyond this, dim in the photograph, rise low hills, without timber, developed on the Lower Lebo (No. 1) along the axis of the Widdecombe Creek anticline. 2, Site of the Silberling Quarry, looking approximately north in Sec. 4, T. 5 N., R. 16 E. The two figures near the middle of the picture stand near the ends of the main section of the quarry. The slope is on the uppermost part of the Lebo (upper No. 2), but the basal Melville (No. 3) lies immedi- ately above, and numerous talus blocks from it are seen. The quarry is not discernible as such, since the picture was taken in 1932, 24 vears after intensive work there. PLATE 6 1, Typical exposure of the Lower Melville sandstone, looking approximately north in Sec. 23, T.5 N., R. 15 E. The valley in the upper right corner is on the Upper Lebo (No. 2), with low barren hills of the Lower Lebo (No. 1) beyond it. 2, Shell imestone in the Melville at Loc. 40, Sec. 29, T.5 N., R. 15 E. Such beds characterize the middle part of the Melville, although the formation is predominantly of sandstone and shale. The invertebrate beds are gen- erally not so thick or so well exposed as they are at this locality. The apparent nodules are gastropod shells. 169 PLATE 6 BULLETIN U.S. NATIONAL MUSEUM ‘A9Vd ALISOddO AAS ALV1d AO NOILYNW1dxa HOA “LNOW ‘O13I14 NIVLNNOW AZVYHO ‘NOILVWHOS ATIIATAW U.S. NATIONAL MUSEUM BULLETIN i169 PLATE 7 3 FORT UNION PRIMATES. 1, Palaechthon alticuspis Gidley, part of right lower jaw with P2-M», type (U.S.N.M. no. 9532), external view, Gidley Quarry; 2, Paromomys maturus Gidley, part of right lower jaw with P:-Ms, paratype (U.S.N.M. no. 9545), external view, Gidley Quarry; 2a, same, crown view; 3, P. maturus, part of right lower jaw with Py-M;, type (U.S.N.M. no. 9473), external view, Gidley Quarry; 3a, same, crown view. (All figures about four times natural size; after Gidley.) U."S. NATIONAL MUSEUM BULLETIN 169 PLATE 8 3 FORT UNION PRIMATES. 1, Pronothodectes matthewi Gidley, part of left lower jaw with Ps-Ms, paratype (U.S.N.M. no. 9531), external view, Gidley Quarry; 2, Paromomys maturus Gidley, part of right lower jaw with P:-Ms, paratype (U.S.N.M. no. 9475), external view, Gidley Quarry; 2a, same, crown view; 3, P. maturus, part of left lower jaw with P:-M3, paratype (U.S.N.M. no. 9337), external view, Gidley Quarry. (All figures about four times natural size; after Gidley.) U.S. NATIONAL MUSEUM BULLETIN 169 PLATE 9 FORT UNION PRIMATES. FOR EXPLANATION OF PLATE SEE OPPOSITE PAGE. PLATE 9 Plesiadapis gidleyi (Matthew), left upper molar, U.S.N.M. no. 10765, crown view. From the Tiffany beds near Ignacio, Colo.; for comparison with Pronothodectes. Pronothodectes matthewi Gidley, right P*-M2?, type (U.S.N.M. no. 9547), crown view. Gidley Quarry. Plesiadapis gidleyt (Matthew), right Pt, U.S.N.M. no. 10659, crown view. From the Tiffany beds near Ignacio, Colo.; for comparison with Prono- thodectes. Plesiadapis rex (Gidley), left lower molar, type (U.S.N.M. no. 9828), crown view. Gidley Quarry. Palaechthon alticuspis Gidley, right M!—, paratype (U.S.N.M. no. 9550), crown view. Gidley Quarry. Palaechthon alitcuspis Gidley, right M'!—, paratype (U.S.N.M. no. 9551), crown view. Gidley Quarry. Paromomys depressidens Gidley, right P*-M*, type (U.S.N.M. no. 9546), crown view. Gidley Quarry. Undetermined, probably primate upper incisor, U.S.N.M. no. 10090, lingual view. Gidley Quarry. Undetermined, probably primate upper incisor, U.S.N.M. no. 10010, lingual view. Gidley Quarry. Undetermined, probably primate upper incisor, U.S.N.M. no. 9928, lingual view. Gidley Quarry. Probably Pronothodectes matthewt Gidley, upper incisor, U.S.N.M. no. 10005, lingual view. Gidley Quarry. Probably Pronoihodectes matthewt Gidley, upper incisor, U.S.N.M. no. 10044, lingual view. Gidley Quarry. Undetermined lower incisor, possibly Plestadapis rex, U.S.N.M. no. 9827, lateral view. Probably Loc. 138. Undetermined, possibly multituberculate lower incisor, U.S.N.M. no. 9552, lateral view. Gidley Quarry. Undetermined, pair of upper incisors, U.S.N.M. no. 9917, lingual view. Gidley Quarry. Plesiadapis gidleyt (Matthew), poorly preserved upper incisor, U.S.N.M. no. 10639, lingual view. From Tiffany beds, 514 miles east of Bayfield, Colo. (All figures about four times natural size; after Gidley.) U.S. NATIONAL MUSEUM BULLETIN 169° PLATE, 10 FORT UNION PRIMATES. 1, Palenochtha minor (Gidley), part of right lower jaw with Ps-Ms3, type (U.S.N.M. no. 9639), external view; la, same, crown view; 2, Llphidotarsius florencae Gidley, part of left lower jaw with P;-Ms;, type (U.S.N.M. no. 9411), external view; 2a, same, crown view; 3, Pronothodectes matthewi Gidley, part of left lower jaw with base of incisor, canine (?), P2 and Py-M3, U.S.N.M. no. 9332, external view; 3a, same, crown view. (All figures about four times natural size; after Gidley. All from Gidley Quarry.) INDEX (Principal references are given in boldface) Abel, O., 86, 115, 141, 165. absarokae, Didelphodus, 110. Absarokee, Mont., 27. Absarokius, 144, 145, 168. aceroides, Platanus, 19. Acmeodon, 106-111. acmeodontoides, Emperodon, 109. acolytus, Ellipsodon, 53, 225, 227, 233, 234, 239. Mioclaenus, 225. acutus, Viverravus, 208, 209. Adam, Lake, Mont., 13, 23, 37. Adapinae, 148. Adapis, 148. admirabilis, Ptilodus, 84, 85. affinis, Sapindus, 19. agapetillus, Anisonchus, 225. Oxyacodon, 225, 227. Alder, 58. Allognathosuchus, 59. Almagre formation, 196. alticuspis, Palaechthon, 33, 47, 62, 156 seq., 159. Amblydactyla, 265, 268. Amblypoda, 217, 218, 265-268. amblyrhyncha, Populus, 57. American Fork, Mont., 13. American Fork formation, 14. americanus, Eucosmodon, 83, 104. americanus primus, Eucosmodon, 83, 103, 104. Ampelopsis, 57, 58. Anacodon, 180, 195. Anaptomorphidae, 33, 60, 61, 142 seq. Anaptomorphus, 144. anceps, Plesiadapis, 35, 47, 50, 51, 156, 168, 169. angustidens, Mimotricentes, 34, 36, 37, 47-49, 52, 194, 197, 205 seq. Anisonchinae, 124, 217, 223, 224, 255 seq. Anisonchus, 2, 50, 62, 216, 256, 258, 262 seq., 267. agapetillus, 225. gillianus, 258, 259. mandibularis, 262. sectorius, 34, 37-41, 48, 50, 51, 53, 54, 255, 259, 261, 262 seq. Apatemyidae, 147. Apatemys, 144. Apheliscidae, 124. Apheliscus, 124. Aphronorus, 106, 121, 124 seq. eee 33, 38, 47, 58, 62, 124 , 140. peiediatus, Oxyacodon, 225-227. 119212—37—_19 aquilonius, Ellipsodon, 34, 37, 39, 47, 62, 66, 227, 233, 234 seq., 238, 240. arachioides, Leguminosites, 57. Aralia notata, 57. Arbor-vitae, 58. Arctocyon, 171, 180. primaevus, 7. ee ueeygnidee 34, 60, 61, 170, 171 seq.,. 207. Arctocyonides, 172. Arctocyoninae, 170-172, 173 seq. Arctoryctes, 139, 140. Artiodactyls, 220, 225, 229-231. asaphes, Unuchinia, 35, 47, 68. Aspideretes nassau, 59. australis, Thryptacodon, 42, 48, 68. Baldwin, D., 7, 233, 262. Barylambda, 11, 222, 269. Barylambdinae, 269. bathmodon, Pantolambda, 269-271 Bathyopsidae, 266, 267. Bats, 136. Bear Butte, Mont., 13, 16, 22, 28, 24, 28, 31, 38. Bear Creek, Mont., 11, 170. Bear formation, 15, 17 seq., 21, 26-28, 55; 57. Bearpaw formation, 15, 16. Bears, 180, 182. Beech, 58. belgicus, “Omomys’’, 144. Belt uplift, Mont., 14. Bessoecetor, 51, 52, 106, 121, 122 seq. diluculi, 33, 47, 122 seq. thomsoni, 35, 47, 50, 122, 123. Big Elk sandstone, 15. Bighorn Basin, Wyo., 55, 83. Bighorn Mountains, Wyo., 116 Ue Big Snowy Uplift, Mont., 14, Bigtimber, Mont., 12. Bigtimber Creek, "Mont., 13. Billings County, N. Dak., 11. Birch, 58. Bittersweet, 58. Blainville, H. M. de, 7. Bolodon, 70. Bridger formation, 143. Broom, R., 71, 72, 86, 89-92. Brown, Be He 14, Buford, N. Dak., Tey dle Bull Mountain Field, Mont., 26, 58. Burke, J. J., 6. Cabrera, A., 64, 246. Cabrera’s Law, 64 seq. Caenolestes, 115. Caenolestidae, 136. Caenopithecus, 161. 279 280 Calamodon, 169. Calvert, W. R., 9, 16, 19, 21, 22, 24, 26, 27, 57. Campbell, M. R., 9, 57. Campeloma, 17, 58. limnaeiforme, 17. nebrascense whitei, 17. canadensis, Physa, 17. Carnivora, 34, 60, 61, 170 seq. Carpites, species undetermined, 57. Carpodaptes, 50, 51, 58, 161, 162, 164. hazelae, 35, 47. Carpolestes, 53, 144, 147, 161, 162, 164. dubius, 162. Carpolestidae, 33, 60, 61, 161 seq. Catathlaeus, 216, 267. cavirictus, Pantolambda, 41, 270, 271. Cayuse Butte, Mont., 8, 9, 18, 25, 29, 35, 41. Champsosaurs, 17, 36, 59. Cheiromyoides, 165. Chestnut, 58. Chiroptera, 119, 135, 136. Chirox, 70. Choeroclaenus, 226, 228, 232 seq. turgidunculus, 226, 227. Chriacidae, 171, 172. Chriacinae, 172, 173. Chriacus, 52, 171-173, 192-194, seq., 198, 199, 201, 203, 258. pelvidens, 197. pugnax, 36, 47, 48, 194, 196, 197. pusillus, 194, 197-199. species undetermined, 35, 47. Chrysochlorid, 140. Cimolomys, 98. gracilis, 85. Claenodon, 2, 42, 50, 52-54, 62, 67, 173, 174 seq., 190, 191, 195, 197, 224. corrugatus, 174-181. ferox, 39-42, 47, 49-51, 54, 55, 174-179, 180 seq., 182-185, 189. latidens, 34, 47, 176, 187 seq., 190. montanensis, 34, 38, 47, 175, 176, 181 seq., 187-189. procyonoides, 176. protogonioides, 175-177, 181, 186. silberlingi, 34, 42, 47, 176, 185 seq., 196 188, 189. species undetermined, 34, 47, 189 seq. species unnamed, 176, 178. vecordensis, 36, 47, 48, 176, 189. Claenodonts, 10. Claggett formation, 15. Clark Fork Basin, Wyo., 81, 83. Clark Fork formation, 11, 16, 20, 21, 81, 83. Cloverley formation, 14. cochranensis, Ectypodus, 77, 82, 83, 100, 101. Colorado group, 14, 15. comma, Conoryctes, 34, 39, 47, 53, 169. Conacodon, 256, 258. cophater, 258, 259. entoconus, 258. BULLETIN 169, UNITED STATES NATIONAL MUSEUM concordiarcensis, Prodiacodon, 33, 46, 112 seq. Condylarthra, 34, 60, 61, 173, 216 seq., 267, 268. Conoryctes, 50, 53, 169. comma, 34, 39, 47, 53, 169. species undetermined, 40, 47, 50. Conoryctidae, 169. Conoryctinae, 169. conus, Microcosmodon, 83. conventus, Neoliotomus, $3, 102. Cope, E. D., 7, 71, 127,,170, 176,,.216- 218, 224, 233, 246, 265, 267, 268. cophater, Conacodon, 258, 259. Coriphagus, 53, 173, 220, 256, 257 seq. encinensis, 259, 260. ‘ montanus, 9, 34, 48, 255, 257, 259, seq. corrugatus, Claenodon, 174-181. Coryphodon, 222, 223, 266. Coryphodontia, 265. Coryphodontidae, 266-269. Coryphodonts, 220, 265. Crazy Mountain Field, Mont., 14, 55. Crazy Mountain Syncline, Mont., 14. Crazy Mountains, Mont., 9, 12, 14, 22- 26, 28. Crazy Peak, Mont., 12. Credneria daturaefolia, 57. Creodonta, 170 seq., 224. Crocodiles, 36, 59. cuneata, Populus, 57. cupanoides, Phyllites, 57. cuspidata, Eudaemonema, 33, 47, 131 seq., 134. Cynodontomys, 128-130. Cypress, bald, 58. danae, Fusconaia, 17. daphnogenoides, Populus, 57. daturaefolia, Credneria, 57. dawkinsianus, Viverravus, 208. declivus, Nedionodus senectus, 59. Deltatheridiidae, 33, 60, 61, 107 seq., 112: Deltatherium, 172, 192, 193. depressidens, Paromomys, 33, 47, 66, 147, 154 seq., 157, 159, 164. Dermoptera, 131. Deuterogonodon, 53, 62, 173, 190 seq. montanus, 5, 37, 47, 49, 191, 192. species undetermined, 37, 47, 192. Diacodon, 111, 112, 117 Dichobunids, 230, 231. Didelphodontinae, 107 seq. Didelphodus, 106-111, 140, 258. absarokae, 110. Didymictis, 53, 67, 208, 209 seq., 214- PAL OY haydenianus, 34, 37, 38, 47, 49, 53, 209-211, 213. microlestes, 34, 37, 47, 62, 209, 210 seq., 213. protenus, 209, 210. tenuis, 34, 47, 210, 212 seq. diluculi, Bessoecetor, 33, 47, 122 seq. Palaeosinopa, 122. INDEX Dinocerata, 265-268. Dinocerea, 265. Dinosaurs, 16, 17. disceptatrix, Haplaletes, 34, 48, 66, 226— 228, 244 seq. disjunctus, Litaletes, 34, 37, 47, 66, 226, 227, 234, 238, 239 seq. Dissacus, 53, 216. navajovius, 216. species undetermined, 34, 47, 216. dissentaneus, Litomylus, 34, 48, 66, 226, 227, 241 seq. Djadochtatherium, 72. Dogs, 180, 208. Douglass, E., 6, 8, 9, 14-16, 21, 31, 38, 52, 70, 104, 135-138, 169, 196, 250, 257, 262, 269. douglassi, Ptilodus, 33, 46, 67, 74, 75, 82, 83, 95, 96, 102. dubius, Carpolestes, 162. Eagle formation, 15. Farle, C., 7% 127, 01717250217, 298; 225, 258. Ectocion, 250-254. Ectypodus, 51, 52, 65, 67, 73, 81-83, 99 seq. cochranensis, 77, 82, 83, 100, 101. grangeri, 33, 46, 67, 74, 75, 79, 82, 83, 99, 101, 102. hunteri, 35, 46, 50. musculus, 77, 79, 82, 83, 99, 101, 102. russelli, 33, 46, 67, 74, 75, 79, 82, 83, 99 seq. silberlingi, 33, 46, 67, 74, 75, 79, 82, 83, 97, 101 seq. species undetermined, 83, 102. elegans, Elpidophorus, 133. Elftman, H. O., 71, 92. Ellipsodon, 58, 225, 227-232, 233 seq., 238-240. acolytus, 53, 225, 227, 233, 234,239. aquilonius, 34, 37, 39, 47, 62, 66, 227, 233, 234 seq., 238, 240. inaequidens, 225-227, 233, 236, 238, 239. lemuroides, 225, 227, 233, 234, 239. priscus, 227, 228. species undetermined, 41, 47, 238. Elliptio priscus, 17. Ellis, A. J., 26. Elm, 58. Elphidotarsius, 50, 51, 53, 68, 147, 161, 162 seq., 164. florencae, 33, 47, 68, 163 seq. Elpidophorus, 51, 53, 127, 129, 1380, 133 seq. elegans, 133. minor, 31, 33, 47, 49, 50, 133 seq. patratus, 35, 42, 47, 50, 51, 133. Emperodon, 107, 108. acmeodontoides, 109. encinensis, Coriphagus, 259, 260. Mixoceclaenus, 257. entoconus, Conacodon, 258. Epanorthidae, 136. 281 Erinaceidae, 106, 111. Eucosmodon, 52, 70, 72, 73, 92, 103 seq. americanus, 83, 104. americanus primus, 83, 103, 104. gratus, 83, 104. molestus, 83, 104. sparsus, 33, 37, 46, 49, 83, 103 seq. teilhardi, 83, 104. Eudaemonema, 53, 106, 127, 129, 130, 131 seq., 140. cuspidata, 33, 47, 131 seq., 134. Euprotogonia, 246, 250, 253. minor, 247, 253. puercensis, 247. subquadrata, 246. europaeus, Glyptostrobus, 57. europaeus ungeri, Glyptostrobus, 57. Eutheria, 72. excedens, Harpagosaurus, 59. Falconer, H., 71. Farr, M. S., 6, 8, 40. Ferns, 58. ferox, Claenodon, 39-42, 47, 49-51, 54, 55 174, 179, 180 seq., 182-185, 189. Mioclaenus, 224. Figs, 58. Fish Creek, Mont., 12, 28, 40, 41. Fish Creek beds, 16. Fisher, R. A., 2, 260, 264. Fishes, 59. Fissipedia, 170, 207, 208. Flora, 57 seq. florencae, Elphidotarsius, 33, 47, 68, 163 seq. floverianus, Mioclaenus, 246. formosus, Halodon, 97. Ptilodus, 97, 101. Viviparus, 17. Fort Benton formation, 14. Fort Union group, 7, 15, 16, 20 seq. fossilis, Onoclea sensibilis, 57. Fox Hills formation, 16. fraudator, Aphronorus, 33, 38, 47, 53, 62, 124 seq., 140. Friant, M., 229. Frontier formation, 15. frugivorus, Phenacolemur, 35, 47. furens, Prothryptacodon, 34, 38, 47, 194, 195 seq. Fusconaia danae, 17. Gallegos Canyon, N. Mex., 233. Ganoids, 36. Gelastops, 52, 106, 197 seq., 140. parcus, 33, 46, 109 seq. genetrix, Populus, 57. Germann, J. C., 6. Gidley, J. W., 3-6, 10, 21, 31, 32, 41, 52, 56, 67, 70-72, 78, 86, 90-92, 97, 101, 115-117, 141, 147-149, 154, 155, 159, 166, 168, 174-176, 180-185, 187, 189- 192, 194, 245-248, 250-253, 262, 269. Gidley Quarry Mont., 9, 10, 22, 30, 31 seq., 35, 37-39, 46, 48-51, 55, 58 seq., 68, 74-76, 93. 76, 81-83, 282 gidleyi, Nothodectes, 165. Plesiadapis, 168. Ptilodus, 33, 38, 46, 67, 74, 75, 82, 83, 95 seq. Gidleyina, 5, 39, 50, 51, 55, 246, 250 seq. montanensis, 5, 41, 48, 50, 251 seq., 253, 254. silberlingi, 40, 48, 50, 253 seq., 255. species undetermined, 41, 48, 50. superior, 42, 48, 254 seq. Giebel, C. G., 171. gillianus, Anisonchus, 258, 259. Gilmore, C. W., 4, 6, 59. Gingko, 58. Glass Lindsay Lakes, Mont., 13, 23, 28. Glendive, Mont., 57, 58. Glyptostrobus europaeus, 57. europaeus ungeri, 57. gracilis, Cimolomys, 85. Ptilodus, 70, 74, 84. grandifoliolus, Sapindus, 19, 57. Granger, W., 7, 10, 11, 70-72, 92, 141, 257. grangeri, Ectypodus, 33, 46, 67, 74, 75, 79, 82, 83, 99, 101, 102. Grasses, 58. gratus, Eucosmodon, 83, 104. Gray Bull formation, 16, 20, 21, 79, 81, 83. Gregory, W. K., 107, 148, 170. Grewia obovata, 57. Grewiopsis platanifolia, 57. Halodon formosus, 97. serratus, 101. Haplaletes, 51, 53, 226-230, 233, 241, 243 seq. disceptatrix, 34, 48, 66, 226-228, 244 seq. Haploconus, 256, 258. Haplomylus, 225-228, 230, 231, 241. speirianus, 226, 227. Hares})'@.i3i:, 11. Harlowton, Mont., 12, 40. Harpagosaurus excedens, 59. Hatcher: J? B.,1'6; Hay, O- Ps, 59, 257. Hayden, F. V., 7. haydenianus, Didymictis, 34, 37, 38, 47, 49, 53, 209-211, 213. Didymictis (Protictis), 209. haydenii, Platanus, 57. hazelae, Carpodaptes, 35, 47. Hazelnut, 58. Hell Creek formation, 15-20, 26-28, 55, 56. Hemiacodon, 145. Hemithlaeus, 221, 222, 256, 258. Hickory, 58. Horsetails, 58. Hunter, Mr. and Mrs. F., 10. hunteri, Ectypodus, 35, 46, 50. Hyopsodinae, 241. Hyopsodontid, aff. Haplaletes, 42, 48. genus undetermined, 221. Hyopsodontidae, 34, 60, 61, 66, 217, 220, 222, 223, 224 seq. BULLETIN 169, UNITED STATES NATIONAL MUSEUM Hyopsodontids, 259. Hyopsodontinae, 223, 225, 231, 241 seq. Hyopsodus, 217, 223-226, 229-231. paulus, 226. Ictidopappus, 53, 208, 209, 213 seq. mustelinus, 34, 47, 214 seq. ignotus, Litolestes, 226, 227. inaequidens, Ellipsodon, 225-227, 233, 236, 238, 239. Tricentes, 225. Indrodon, 127-130. Insectivora, 33, 60, 61, 104 seq., 217, 2185-225. 229) 250 intermedius, Pantolambda, 34, 48, 270 seq. Invertebrates, 58 seq. Jepsen, G. Li, 6, 11, 20, 55;:79,¢83, 120, 141, 148, 147, 161, 162, 165, 225, 226, 229. jepseni, Parectypodus, 38, 46, 74, 75, 82, 83, 100, 102 seq. Judith River formation, 15-17. Kingsbury formation, 11. aahn re Lise Knowiton, F. H., 9, 19-21, 56-58. Kootenai formation, 14, 15, 29. ladae, Leptacodon, 33, 46, 62, 113 seq., 115. Lance formation, 15-21, 55-57. latidens, Claenodon, 34, 47, 176, 187 seq., 190. Mimotricentes, 5, 34, 38, 47, 194, 205, 206. Neoclaenodon, 187, 188. Tricentes, 194, 203, 205. latrunculus, Spanoxyodon, 34, 47, 68, 194, 203. Laurels, 58. Lebo Creek, Mont., 8, 13, 22. Lebo formation, 9, 15, 16, 21 seq., 24-28, 3651305 ,0 0, nOlts Leguminosites arachioides, 57. Leipsanolestes, 113. Lemoine, V., 7. lemuroides, Ellipsodon, 225, 227, 238, 234, 239. Mioclaenus, 225. Lennep formation, 15, 16, 26. Lepisosteus species undetermined, 59. Leptacodon, 51, 52)) 67, 106; 111,» 112; 113 seq., 117. ladae, 33, 46, 62, 113 seq., 115. munusculum, 33, 46, 114 seq. packi, 113, 114. siegfriedti, 113. tener, 35, 46, 50, 68, 113-115. Leptictidae, 33, 60, 61, 106, 107, 111 seq., 120. limnaeiforme, Campeloma, 17. Lion Butte, Mont., 13, 24, 28. Lioplax nebrascensis, 59. Litaletes, 53, 226, 227, 231, 232, 238 seq., 245. disjunctus, 34, 37, 47, 66, 226, 227, 234, 238, 239 seq. INDEX Litolestes, 51, 225-228, 241, 243. ignotus, 226, 227. notissimus, 35, 48, 227. Litomylus, 53, 226 228, 233, 241 seq., 243, 245 dissentaneus, 34, 48, 66, 226, 227, 241 seq. Litotherium, 161. Livingston formation, 8, 9, 16, 19, 26, 27. Lizards, 59. Lloyd, E. R., 11. Lobdell, J. F., 11. lobdelli, Psittaeotherium, 170. Longman, H. A., 11 Loxolophus, 171. Lull, R. S., 4. Lupton, C. T., 26, 58. lydekkerianus, Mioclaenus, 225. Magnolia, 58. mandibularis, Anisonchus, 262. Maple, 58. Marsh, O. C., 71, 217, 265, 268. Marsupials, 70-72, 115-117, 128. Matthew, W. D., 2, 7, 52, 64, 65, 70, 72, 105, 107, 111, 115-120, 127-131, 140, 141, 146, 169-177, 180, 194, 207, 209, 216-219, 222, 223, 225-227, 230, 232, 233, 246, 257, 258, 262. matthewi, Pronothodectes, 33, 47, 165 seq. maturus, Paromomys, 33, 39, 47, 62, 66, 147, 149 seq., 154-157. mediaevus, Ptilodus, 52, 67, 70, 77-80, 82, 83, 85, 96. Meek, F. B., 7. Megopterna, 135, 136. minuta, 9, 136-139. Meinzer, O. E., 26. Melville, Mont., 12, 14, 25, 40. Melville formation, 15, 20, 21, 25, 27, 28, 39; 5D: Meniscotheriidae, 216-218, 224. Meniscotheriinae, 224. Mesonychidae, 34, 60, 61, 170, 216. Metachriacus, 53, 66, 172, 178, 192-194, 196, 197 seq. provocator, 37, 38, 47, 49, 66, 194, 200 seq., 204. punitor, 34, 47, 49, 62, 66, 194, 197 seq., 201, 202, 204. species undetermined, 37, 47. Metatheria, 72. Miacidae, 34, 60,61, 170, 171, 182, 207 seq. Miacinae, 207, 208. Microchoerus, 144, 145. Microcosmodon, 73, 81. conus, 83. microlestes, Didymictis, 34, 37, 47, 62, 209, 210 seq., 213. Microsyopinae, 128, 129. Microsyops, 127-130. Mimotricentes, 52, 538, 173, 192, 193, 203 seq. angustidens, 34, 36, 37, 47-49, 52, 194, 197, 205 seq. 283 Mimotricentes latidens, 5, 34, 38, 47, 194, 205, 206. species undetermined, 36, 47, 48, 207. minor, Elpidophorus, 31, 33, 47, 49, 50, 3 seq. Euprotogonia, 247, 253. Palaechthon, 159. Palenochtha, 33, 47, 159 seq. Tetraclaenodon, 249, minuta, Megopterna, 9, 136-139. Mioclaenidae, 217, 218, 225, 231, 232, 257. Mioclaeninae, 218, 223, 225, 232 seq., 257, 258. Mioclaenus, 5, 216, 217, 224-228, 230- 233 acolytus, 225. ferox, 224. floverianus, 246. lemuroides, 225. lydekkerianus, 225. opisthacus, 225. turgidunculus, 225, 232. turgidus, 224-227, 230, 232. Mixoclaenus, 173, 255, 257, 258. encinensis, 257. Mixodectes, 5, 53, 70, 127-131, 133. Mixodectidae, 33, 60, 61, 106, 127 seq. Mixodectinae, 128, 129. molestus, Eucosmodon, 83, 104. Monotremes, ly wis 92. montanensis, Claenodon, 34, 38, 47, 175, 176, 181 seq., 187-189. Gidleyina, 5, 41, 48, 50, 251 seq., 253, 254. Myrmecoboides, 33, 47, 117 seq. Neoclaenodon, 174-176, 181, 188. montanus, Coriphagus, 9, 34, 48, 255, 257, 259 seq. Deuterogonodon, 5, 37, 47, 49, 191, 194. Ptilodus, 9, 33, 37, 38, 46, 52, 62, 67, 70, 74-80, 82, 88, 84 seq., 95-97. Mowry formation, 15. multifragum, Psittacotherium, 34, 47, 49, 58, 68, 169 seq. Multituberculata, 2, 33, 60, 61, 70 seq. Multituberculate undetermined, 42. munusculum, Leptacodon, 33, 46, 114 seq. musculus, Eetypodus, 77, 79, 82, 83, 99, 101, 102. Musselshell River, Mont., 12, 14. mustelinus, Ictidopappus, 34, 47, 214 seq. Myrmecobiidae, 115. Myrmecobius, 115, 116. Myrmecoboides, 10, 52, 106, 111, 112, 115 seq. montanensis, 33, 47, 117 seq. Nannopithex, 145. nassau, Aspideretes, 59. navajovius, Dissacus, 216. nebrascense whitei, Campeloma, 17. 284 nebrascensis, Lioplax, 59. Necrolemur, 144-146, 148. Nedionodus senectus, 17, 59. senectus declivus, 59. Neoclaenodon, 173-176. latidens, 187, 188. montanensis, 174-176, 181, 188. silberlingi, 175, 185. Neoliotomus, 81. conventus, 83, 102. ultimus, 83. Niobrara formation, 15. nobilis, Platanus, 58. notata, Aralia, 57. Notharctinae, 147. Notharctus, 217. Nothodectes, 141, 164, 165. gidleyi, 165. notissimus, Litolestes, 35, 48, 227. Nyctitheriidae, 33, 60, 61, 106, 118 seq. Nyctitherium, 119. Oak, 58. obovata, Grewia, 57. Omomys, 148, 161. belgicus, 144. vespertinus, 144, 162. Onoclea, 58. sensibilis fossilis, 57. opisthacus, Mioclaenus, 225. Protoselene, 225-227. Ornithorhynchus, 71. Osborn, H. F.,.:7) 215 73; 86;:127,:170- 172, 217-219, 225, 258, 266. Osgood, W. H., 115. Otter Creek, Mont., 13. Owen, R., 71. Oxyacodon, 225-230, 232, 233, 241, 248. agapetillus, 225, 227. apiculatus, 225-227. priscilla, 225, 227. turgidunculus, 225. Oxyclaenidae, 171, 257, 258. Oxyclaeninae, 170-173, 192 seq. Oxyclaenus, 171, 172, 195. packi, Leptacodon. 118, 114. Palaechthon, 53, 142-145, 147, 148, 156 seq., 159, 160, 164. alticuspis, 33, 47, 62, 156 seq., 159. minor, 159. Palaeosinopa, 52, 120-122, 124, 126, 140, 258. diluculi, 122. species undetermined, 35, 47. Palenochtha, 53, 142-145, 148, 158 seq. minor, 33, 47, 159 seq. paleocena, Zanycteris, 136. Palm, 58. Pantodonta, 34, 60, 61, 265 seq. Pantolambda, 11, 27, 58, ‘221, 222, 265, 266, 269 seq. bathmodon, 269-271. cavirictus, 41, 270, 271. intermedius, 34, 48, 270 seq. species undetermined, 34, 39, 48, 271. BULLETIN 169, UNITED STATES NATIONAL MUSEUM Pantolambdidae, 34, 60, 61, 217, 219, 266-268, 269 seq. Pantolambdids, 35, 40, 41, 48, 49, 50, 62, 218-220. Pantolambdinae, 269. Pantolambdodontidae, 269. Pantolestes, 120, 121, 140. Pantolestidae, 33, 60, 61, 106, 120 seq. Pantolestinae, 121 seq., 124. parcus, Gelastops, 33, 46, 109 seq. Parectypodus, 52, 81-83, 99, 100, 102 seq. jepseni, 33, 46, 74, 75, 82, 83, 100, 102 seq. simpsoni, 77, 79, 82, 83, 102. species undetermined, 83. tardus, 77, 79, 83, 97, 1038. Paromomys, - 50,,:51,2 53; 63;...66,.167, 142-147, 148 seq., 156, 157, 159, 162, 164. depressidens, 33, 47, 66, 147, 154 seq., 157, 159, 164. maturus, 33, 39, 47, 62, 66, 147, 149 seq., 154-157. Paromomys, ef., genus and species unde- termined, 42, 47, 156. Paskapoo formation, 11, 81, 83, 130. patratus, Elpidophorus, 35, 42, 47, 50, ole ios: Patterson, B., 11, 222, 266, 269. paulus, Hyopsodus, 226. Pearson, H., 78. Peltosaurus, species undetermined, 59. pelvidens, Chriacus, 197. Pelycodus, 148, 156, 162, 217. Pentacodon, 53, 106, 120, 121, 124, Pentacodontinae, 121, 123 seq. Periptychidae, 34, 60, 61, 216, 217, 220, 224, 255 seq., 266-268. Periptychinae, 217, 223, 224. Periptychines, 259. Periptychus, 62, 216, 218, 221, 222. Perissodactyla, 216. Phenacodaptes, 225, 226, 229-231. sabulosus, 226. Phenacodontidae, 34, 60, 61, 216-218, 223, 245 seq. Phenacodonts, 62. Phenacodus, 54, 190, 191, 216, 223, 251. Phenacolemur, 50, 146, 147. frugivorus, 35, 47. Phenacops, 107, 111. Phyllites cupanoides, 57. Phyllostomatidae, 134, 135. Physa, 17. canadensis, 17. Picrodontidae, 33, 60, 61, 134 seq. Picrodus, 53, 107, 134, 135 seq. silberlingi, 9, 33, 47, 136 seq. Pines, 58. Plagiaulax, 70. Plagiomene, 130. Plagiomenidae, 130, 257. planolatere, Viviparus, 59. Plants, 56. 126. 219, INDEX platanifolia, Grewiopsis, 57. Platanus aceroides, 19. haydenii, 57. nobilis, 58. raynoldsii, 19. species undetermined, 19. Plateau Valley formation, 11. Platychoerops, 164. Plesiadapidae, 33, 60, 61, 148, 164 seq. Plesiadapis, 50, 51, 53, 141, 144, 147, 164-166, 167 seq. anceps, 35, 47, 50, 51, 156, 168, 169. gidleyi, 168. rex, 42, 47, 51, 156, 167 seq. tricuspidens, 165. Plesiolestes, 141, 148, 165. Pleuraspidotheriidae, 218. Pleuraspidotheriinae, 224. pliciferus, Tetraclaenodon, 246-249, 253. Polecat Bench, Wyo., 20, 27. politus, Viverravus, 208. Poplars, 58. Populus amblyrhyncha, 57. cuneata, 57. daphnogenoides, 57. genetrix, 57. species undetermined, 57. Porcupine Butte, Mont., 18, 25, 29, 41, 42, 57. Potter Creek, Mont., 27. Powder River Basin, Wyo., 11. Prentice, S., 6. primaevus, Arctocyon, 7. Titanoides, 11. Primates, 2, 10, 33, 60, 61, 127, 128, 83, 135, 141 seq., 217. primus, Eucosmodon americanus, 103, 104 Princeton localities, Mont., 40, 46. Princeton Quarry, Wyo., 55. priscilla, Oxyacodon, 225, 227. priscus, Ellipsodon, 227, 228. Elliptio, 17. Proboscidea, 265. procyonoides, Claencdon, 176. Prodiacodon, 52, 106, 107, 111, 112 seq. concordiarcensis, 33, 46, 112 seq. puercensis, 112, 113 Proectocion, 250. Proglires, 127. Pronothodectes, 50, 51, 53, 147, 162, 164, 165 seq. matthewi, 33, 47, 165 seq. Proscalops, 140. protenus, Didymictis, 209, 210. Prothryptacodon, 51, 53, 173, 192, 193, 194 seq., 199. furens, 34, 38, 47, 194, 195 seq. Protictis, 209. haydenianus, 209. Protochriacus, 171. Protogonia, 216, 246. subquadrata, 246. protogonioides, Claenodon, 181, 186. 175-177, 285 Figueouee 5, 53, 172, 190, 191, 250, 251. Protoselene, 225-232, 241, 251, 254. opisthacus, 225-227, provocator, Metachriacus, 37, 38, 47, 49, 66, 194, 200 seq., 204. Pseudoloris, 161. Psittacotheriinae, 169 seq. Psittacotherium, 31, 53, 169 seq. lobdelli, 170. multifragum, 34, 47, 49, 53, 68, 169 seq. Pterospermites, 57. Ptilodontid undetermined, 35. Ptilodontidae, 33, 60, 61, 73, 74, 80 seq. Ptilodus, 10, 31, 52, 65, 67, 70-73, 81-83, 84 seq., 99, 161. admirabilis, 84, 85. douglassi, 33, 46, 67, 74, 75, 82, 83, 95, 96, 102. formosus, 97, 101. gidleyi, 33, 38, 46, 67, 74, 75, 82, 83, 95 seq. gracilis, 70, 74, 84. mediaevus, 52, 67, 70, 77-80, 82, 83, 85, 96. montanus, 9, 33, 37, 38, 46, 52, 62, 67, 70, 74-80, 82, 83, 84 seq., 95-97. serratus, 101. sinclairi, 33, 36, 46, 48, 52, 62, 67, 74-76, 78, 79, 82, 83, 96, 97 seq., 101. species undetermined, 33, 46, 83. trovessartianus, 77, 82, 83, 96. puercensis, Euprotogonia, 247. Prodiacodon, 112, 113. Tetraclaenodon, 34, 39, 48, 49, 246-248, 249 seq. Puerco formation, 7, 11, 16, 18, 21, 46, 52-55, 81, 83, 173, 186, 191, 196, 224, 255, 259. Puet Creek, Mont., 28. pugnax, Chriacus, '36, 47, 48, 194, 196, 197. punitor, Metachriacus, 34, 47, 49, 62, 66, 194, 197 seq., 201, 202, 904. pusillus, Chriacus, 194, 197-199. raynoldsii, Platanus, 19. Red Lodge, Mont., 27. Reed Point, Mont., 27. Reptiles, 59, retusus, Viviparus, 59. rex, Plesiadapis, 42, 47, 51, 156, 167 seq. Tetonius, 167, 168. Richards, R. W., 26, 58. Rodents, 127, 128. Roger, O., aoe Romer, A. Se Russell, L. gs. v1. Nie O29; 190, LOL: russelli, Ectypodus, oe) ‘46, Gteed4, 975; 79, 82, 83, 99 seq. sabulosus, Phenacodaptes, 226. Salix, species undetermined, 57. San Juan Basin, N. Mex., 2, @; 00; Say. 224, 286 Sand Coulee formation, 16, 81, 83. Sapindus affinis, 19. grandifoliolus, 19, 57. species undetermined, 19. Sealops, 139. Scarritt Quarry, Mont., 10, 25, 30, 34 seq., 39, 41, 46, 50, 51, 54, 55, 60, Gls 63: Schlaikjer, E. M., 139, 140. Schlosser, M., 86, 115, 120, 217, 224, 257. Scott, W. B., 86, 216, 217, 224, 225, 246. sectorius, Anisonchus, 34, 37—41, 48, 50, 51, 53, 54, 255, 259, 261, 262 seq. Sedges, 58. Selaginella, 58. senectus, Nedionodus, 17, 59. senectus declivus, Nedionodus, 59. sensibilis fossilis, Onoclea, 57. Sentinel Butte formation, 25. Sequoias, 58. serratus, Halodon, 101. Ptilodus, 101. Shields River, Mont., 27. Shoshonius, 127, 145. siegfriedti, Leptacodon, 113. Silberling, A. C., 3, 4, 6, 8-10, 17, 18, 21-23, 26, 27, 30-32, 34, 38, 40—42, 56, 70, 86, 98, 183, 168, 250, 251, 257. Silberling Quarry, Mont., 9, 10, 29, 30, 31, 33 seq., 37, 39, 46, 48-51, 55, 56, 59-62, 68, 74, 75, 93. silberlingi, Claenodon, 34, 42, 47, 176, 185 seq., 188, 189. Ectypodus, 33, 46, 67, 74, 75, 79, 82, 83, 97, 101 seq. Gidleyina, 40, 48, 50, 253 seq., 255. Neoclaenodon, 175, 185. Picrodus, 9, 33, 47, 136 seq. simplicidens, Stilpnodon, 33, 47, 52, 68, 119, 120, 140. Simpson, G. .G., 7, 54, 70, 71,-92; 107, 120, 121, 136, 161, 194, 225, 266. simpsoni, Parectypodus, 77, 79, 82, 83, 102. Sinclair, W. J., 6, 7, 11. sinelairi, Ptilodus, 33, 36, 46, 48, 52, 62, 67, 74-76, 78, 79, 82, 83, 96, 97 seq., 101. Soricoidea, 104, 119. Spanoxyodon, 53, 173, 192-194, 203. latrunculus, 34, 47, 68, 194, 203. sparsus, Eucosmodon, 33, 37, 46, 49, 83, 103 seq. speirianus, Haplomylus, 226, 227. Stanton, T. W., 3, 9, 16, 18, 21-23, 57. Stehlin, H. G., 145, 146. Stillwater River, Mont., 27. Stilpnodon, 52, 119 seq. simplicidens, 33, 47, 52, 68, 119, 120, 140. Stone, R. W., 9, 16, 19, 21, 22, 24, 26, 27, 57 Stylinodontidae, 34, 60, 61, 169 seq. BULLETIN 169, UNITED STATES NATIONAL MUSEUM Stylomyleodon, 59. subquadrata, Euprotogonia, 246. Protogonia, 246. superior, Gidleyina, 42, 48, 254 seq. Tetraclaenodon, 254. Sweetgrass County, Mont., 12, 14. Sweetgrass Creek, Mont., 13, 41, 57. Sycamores, 58. symbolicus, Tetraclaenodon, 5, 34, 36, 37, 39, 43, 48, 49, 246 seq., 249, 255. Taeniodonta, 34, 60, 61, 169 seq. Taeniolabis, 71, 99. Taligrada, 217-219, 265-268. Talpid, 140. Talpoidea, 104, 119. tardus, Parectypodus, 77, 79, 83, 97, 103. Tarsioids, 144, 145, 148. Tarsius, 144, 146. Teilhard de Chardin, P., 7, 148, 165. teilhardi, Eucosmodon, 83, 104. tener, Leptacodon, 35, 46, 50, 68, 113- 115 tenuis, Didymictis, 34, 47, 210, 212 seq. Tetonius, 127, 144-146, 148, 168. rex, 167, 168. Tetraclaenodon, 5, 11, 39, 41, 43, 50—54, 190, 191, 216, 221-223, 246 seq., 251-255. minor, 249. pliciferus, 246-249, 253. puercensis, 34, 39, 48, 49, 246-248, 249 seq., 250. species undetermined, 35, 42, 48, 250. superior, 254. symbolicus, 5, 34, 36, 37, 39, 43, 48, 49, 246 seq., 249, 255. Theria, 72, 92. Thermopolis formation, 15. Thomson, A. C., 10. thomsoni, Bessoecetor, 35, 47, 50, 122, 123. Thryptacodon, 51, 53, 172, 173, 194, 195. australis, 42, 48, 68. Tiffany formation, 7, 11, 20, 54, 55, 81, 83, 112, 136, 141, 161; Tilia weedii, 57. Titanoides, 269. primaevus, 11. Tongue River formation, 25. Torrejon formation, 7, 11, 46, 52-55, 62, 70, 74, 81, 88, 121, 128, 169-171, 173, 175, 176, 178, 180, 181, 185-187, 189, 196, 208-210, 218, 216, 224, 246-249, 255, 257, 262-265, 271. Tricentes, 5, 11, 53, 171, 172, 192-194, 197, 203-205. inaequidens, 225. latidens, 194, 203, 205. tricuspidens, Plesiadapis,165. Triisodon, 171. Triisodontidae, 171. Triisodontinae, 170, 172. Tritylodon, 71. trochiformis, Viviparus,17. Trogolemur, 146, 147. INDEX 287 trovessartianus, Ptilodus, 77, 82, 83, 96. | Viviparus, 17, 58. Tullock formation, 20. formosus, 17. turgidunculus, Choeroclaenus, 226, 227. planolatere, 59. Mioclaenus, 225, 232. retusus, 59. Oxyacodon, 225. trochiformis, 17. : : Walnut, 58. eee seme Ia cen Walvoord, Lake, Mont., 13, 23. Wasatch formation, 20, 21. Uintalestes, 146, 147. Wiashalan 4 Uintatheres, 265, 266, 268. Weber ae Uintatheriidae, 266, 267. Weed, W. H., 8, 57. ultimus, Neoliotomus, 83. weedii, Tilia, 57. ungeri, Glyptostrobus europaeus, 57. Wegemann, C. H., 11. Unio, 58. Wetmore, A., 4, 6. Unuchinia asaphes, 35, 47, 68. Wheatland County, Mont., 12. Ursidae, 182. whitei, Campeloma nebrascense, 17. Ursus, 185. Widdecombe Creek, Mont., 23, 28, 31, vecordensis, Claenodon, 36, 47, 48, 176, 36, 38. 189. Wildeat Creek, Mont., 37. vespertinus, Omomys, 144, 162. Wilmarth, M. G., 20. Viburnums, 58. Woolsey, L. H., 26, 58. Vitis xantholithensis, 57. Wortman, J. L., 7, 127, 170, 217. Viverravidae, 208. xantholithensis, Vitis, 57. Viverravinae, 170, 207, 208 seq. Xenacodon, 111, 112. Viverravus, 53, 208, 209, 214, 216. Yellowstone River, Mont., 12, 14, 27. acutus, 208, 209. Yew, 58. dawkinsianus, 208. Zanycteris, 53, 134, 136. politus, 208. paleocena, 136. O . ait ig ‘new, 4 ae _— ee it lattes Fae | 8 Soa Lebolnt ilentiesedeons : va vl Th peers le VEE 1h (2 eat Fonte r H ide a a Syie? Mtably a | fe ass, ae [Revita Singlets iacans =i eet wi eset i ABS Coe ees» s % 5 Ly aIETG Sees CPi aay te weir iit a >) ae On - F Er oo fil luve | | ee pa vey 8 Te i aah santa AL at, ult Pir aes a par: pa i, “ik 0S olka i <7 “OR” A “aol yigd Pree mi, ei We L By ata aaa thf else 4 a ee CRA ATT 4 TRS ath B tei. . - ee! - ! é ¢ } aint ea ah mek ad ne a AE 4s LAY |) 2 bP geht, bes uy) 0, pike if bbe heie a ae aa. it bac We: 4 ie! OS pave A ae a iS hub isp 2 a iD, oli paris Mh ve rs eLiaat Uoiluoys Siete ert oe iris M yt we, a 4 ies petites A its a) itis Lee a ‘doohee i Diag aa ae 2 hy foo ee agierie ! tae ae! rt, OL Hao, tases husltess ad rs A aa oh oa es ak aul dopmatl tthg ay Peciiwidan aarabiqre ) ane : el hue AA - ads PME eee ie “at ai . Tanke AD. Sassi Diy ir dite tte a ae Eel aie. an ‘Fh Ha “~ ena Fas anu mid : oe ave se capbage 9 Ramet Rot! tar ale i ; W fet fisi i. tapi! aun ot" te lt, a “Rel Diglee h rubles nA Fad lente pa AGE «Sag ait lg pena au 26 EAL och SURO Acie ig murray | Estey 79 aa wre ae mie Ay Ahir ia é 6 teu artim Hi jie View! NS ay senate thi ail inane (SE vex OY AOR BAD hee Ae Gee e SELL: Ail apepnnaae ye be Seda pee arog TS silanes f mas hi ae pbnaie: eis ee he fk ES ae seu Goer sO hy TIS t Ae ein ay ave Coase R a sas oe he da seine a8 = FR rca! abe T 8G40% Ieretss a {20g 7 Vatelen ts Hea ( it Si, 7 e08s oY NTE ls ue g iad - Tri > mY eG ig ray i i j 7 i in Marry Woeeer ash) 1 e a yee a prers ov Se" -- 4 nr ~ Ww 7 a7 ot — cor : : q . ia 2 aoe rr / ro a Pde} i en i an he ' OA fbr, * A gribs ve q i j fi fi ; a : AGP t > % i , im co § Uns i) é Talis ‘1 _ - i - = - ’ AD tae A ly t : oh Sh i . | a ey € Bh, : . 7 ; ; y hes vi et ' i ; ; ; pd maith : % mos yrs e,. * ais as ‘ K ‘ >. : \« ; vt ies Mak, bp okie : y : ree) Ys = 1) i - so 7 shtiar: 23 #é : a j hi Ore AES peat a f He & ut PHD ° \ \ 7 ‘ ie f ihe - i” : ; y “pi 4 i. LP ccd, her wie Te, 4 iS oe oo + dd oe “7 rs 7 & Le Live - f = ea - Huldgs ; i ; Mat ais ae ® veS} nt. Ty 4 1) iii” hatha! (LY Pe Gi = 2 mee 2a eal! WA ae 4 a 7 - a te ht, & aL ‘a : a lr i 7 ak : Fi . J hp My pila oy aeioe p tt a — ris fed rea] . = oh, ye. SeyD takes - ALA, aioe iG, Seve air ee 6. Ni? pag rave. ital D a vj ay TY «tf - : = , ‘arts wT Pon T \ ee ' 4 Ly @ ' “ Ma] } Mm ea i A eve re) CP, j Sor by ' Ca te S "a et j ys mt he he Dee ae te ] oS lee ee wie ae es (UAL eae oy 4 , tt m4. sh * vi 7 chy eeniine 3 Vi 2) ae GH LS ayy i »* OG, Aas q 0 TLD ay. aa AT Hh 7 Pye er) ee ge ore eee i : 7 & deve Th 1) tile " rye a i, 7 ) Ler ie Atos! »4 F Sait ame ke aad, Ty tae, - = as Stee en emi Ps Oo) yeh, dea ee nn, a ry eae rit: Gite e : ; oa . ei retains ‘ eer 0 iy . ® i U s e WiniArn sation, § i ve T)3 10 ‘taal t Pecigye ais Hie tet - * 7 - a 7 - = - Uy “7 , c j — ee SS niente - a SMITHSONIAN INSTITUTION LIBRARIES “HINT TTA 9088 01421 2534 WY \ » ty Yj \ AW WY