;v. -^; ^::m^u,f,^s)-m^^^mcM^!^^^^sm^s^K,umi L I B FIAR.Y OF THL UNIVERSITY or ILLINOIS bb05 REMO . c^ GEOLOGICAL SERIES OF FIELD MUSEUM OF NATURAL HISTORY Volume VI Chicago, December 28, 1937 No. 18 A SORICID AND TWO ERINACEIDS FROM THE WHITE RIVER OLIGOCENE By Bryan Patterson Assistant Cukator, Paleontology AND Paul O. McGrew Walkbr Museum, The University op Chicago The study presented in the following pages is due to a fortunate coincidence of collecting. For several years past, one of us (P. 0. McG.) has been working certain lower Brul^ exposures in north- western Nebraska for small mammals. Field Museum recently received from Mr. James Mellinger of Longmont, Colorado, a representative sample of the microfauna found in a lower Brul^ channel deposit in northeastern Colorado. Each of these two col- lections contained good specimens of new or little-known insectivores, the material from one locality supplementing that from the other in a most gratifying way. Study of these specimens has established two genera known hitherto from very incomplete material, and has made possible certain interesting taxonomic reassignments. We wish to express our sincere thanks to the Staff of the Depart- ment of Vertebrate Paleontology of the American Museum for their many kindnesses during our visits there, and particularly to Dr. Walter Granger for permitting us to borrow and redescribe the holotype of Domnina gradata Cope. We are indebted to Mr. James Mellinger for his gift of specimens to Field Museum. Mr. Michael S. Chappars of Walker Museum kindly permitted the use of his Leica camera for taking the photomicrographs. Most of the Ne- braska specimens were collected by Winifred Cook McGrew. Some rodent specimens from Mr. Mellinger's locality have been examined by Dr. Albert Elmer Wood. He has kindly informed us that they indicate lower Brul6 age. The technique employed in making the accompanying photo- micrographs will be described elsewhere by McGrew. No. 401 245 . tfT'-Mrv Survey 246 Field Museum of Natural History — Geology, Vol. VI Soricidae Gray Domnina Cope Domnina Cope 1873, Pal. Bull., No. 16, p. 1. Protosorex Scott 1894, Proc. Acad. Nat. Sci. Phila., p. 446. Genotype. — Domnina gradata Cope 1873. Distribution. — Middle Oligocene, Lower Brul^ beds, Colorado, South Dakota, Nebraska. Emended diagnosis. — If, C§, P|, M|. Teeth heavily pigmented. P- lacking. P^ with small protocone and hypocone. Upper molars with hypocones resembling those of Blarina, posterior cingula ex- tending externally to metastyles. I^ less procumbent than in recent genera, without accessory cuspules, alveolar rim not extending poste- riorily beneath premolars. Py-? small; Py larger than Pij_^; Py the largest of the series with single, conical, anteriorly projecting cusp, and relatively strong transverse posterior cingulum. Lower molars progressively reduced in size from My to M^, protoconids and hypoconids sharply angulate externally, strong antero-external cingula; entoconids of My_ij high, and connected by ridges to poster- ior slopes of metaconids. Talonid of M^ reduced, hypoconid small, entoconid greatly reduced or absent. Horizontal ramus heavy, deep, straight beneath molars, mental foramen beneath My. Remarks. — This interesting genus has had a peculiarly checkered taxonomic history — a somewhat surprising fact — for when proper comparisons are made, the fundamental resemblances of the lower molars of the holotype to those of the soricids are at once apparent. In his type description. Cope (1873a, p. 1) made no statement as to affinities. Subsequently (1874, p. 465) he stated that "Domnina Cope, and probably Embassis Cope [= Peratherium], present affinities to the Soricidae, so far as known," but in 1884 (pp. 788, 810) he changed his earlier view and referred the animal, with some doubt, to the Chiroptera. This assignment was followed without comment by several early workers. Zittel (1893, p. 562) placed Domnina in the Ictopsidae (=Leptictidae), and in this he was followed by Hay in his first bibliography. Matthew at first (1899, p. 55) referred to it as "Insectivora inc. sed.," and later (1909a, p. 105) placed it in the Talpidae. Haug (1922, p. 1528) went so far afield as to include it in his Oligocene faunal list as a rodent. In recent years the con- sensus has been that it is a member of the Talpidae (Hay, second bibliography. Cook and Cook 1933, Scott and Jepsen 1936). 5b0.5 ^^'^•-^' FT Oligocene Soricid and Erinaceids 247 The type of Protosorex crasstis Scott was formerly in Walker Museum, but is now unfortunately lost. The specimens collected by Mellinger and McGrew in Colorado and Nebraska agree with the description of this form, and also with the holotype of Domnina, thus leaving no doubt that Protosorex must be relegated to synonymy. Scott made no mention of pigment on the teeth. The holotype of Domnina gradata appeared to show no pigment, whereas our material is richly pigmented. There appeared to be a possibility, therefore, that the new specimens might represent a distinct genus distinguished from the two previously described specimens by the presence of pigment. To test this possibility the holotype of D. gradata, the referred specimens, and a specimen each of Blarina bremcauda and of Peratherium sp. (collected in the locality at which part of the referred material of Domnina was found) were examined under ultra-violet light. Under this light the pigmented portions of the enamel of Blarina did not fluoresce, in sharp contrast to the unpigmented areas which fluoresced a bluish violet color. The enamel of Peratherium fluoresced a yellowish white color, uniform on all parts. Our specimens of Domnina, on which the pigmentation is very apparent in ordinary light (see figs. 61, 62, 64), fluoresced dull orange on the pigmented areas and yellowish white, as in the Per- atherium specimen, on the unpigmented. The holotype was then examined with results identical to those obtained for the referred specimens. From this evidence, and in the absence of the holotype of P. crassus, it is legitimate to assume that the specimen described as Protosorex was pigmented also. The heavily pigmented teeth of Domnina indicate that colored enamel was acquired early in the history of Soricidae. In view of the interesting results obtained from this material by the use of ultra-violet light, it would seem desirable to examine all fossil soricids in a similar manner. The dental formula followed is that proposed by Arnbach- Christie-Linde (1912), and employed by Cabrera in his catalogue of the recent insectivores (1925). ^ The teeth immediately following the upper and lower incisors have been regarded by most authors as canines. In a study of the development of the anterior teeth in Sorex, Arnbach found, however, that in embryos there were present the germs of five upper and four lower incisors, an upper and lower canine, and four upper and lower premolars. Of these only the last three upper and the last lower incisors, the upper premolars, and Py 1 This able and useful work, and its companion volume on the recent mar- supials, does not appear to be as widely known among paleontologists as it deserves. 248 Field Museum of Natural History — Geology, Vol. VI and Pt persisted, the rest disappearing before birth. A vestigial milk dentition was also observed. Similar results, save for some de- tails, were obtained for Neomys. Domnina, in which the lower pre- molars are all present and Py and P^ larger than Pij-^, provides some support for Arnbach's views, as noted by her (1912, p. 623). Discovery of still earlier soricids should provide further tests of this formula. Domnina gradata Cope (figs. 60-65) D. gradata Cope 1873, Pal. Bull., No. 16, p. 1. Miothen crassigenis Cope 1873, Synopsis Vertebrata Colorado, p. 8; D. crassi- genis Cope 1874, Ann. Rept. U. S. Geol. Geog. Surv. Terr., 1873, p. 470. Protosorex crassus Scott 1894, Proc. Acad. Nat. Sci. Phila., p. 446. Holotype. — ^Amer. Mus. No. 5353. Portion of right horizontal ramus with My--?. Horizon and locality of holotype. — Lower Brul6 beds. Cedar Creek, Logan County, Colorado. Referred specimens. — F.M. No. P15320, incomplete left mandible with I^, Py, Mt-7, alveoli of Py-^; collected in the lower Brul6 beds. Sec. 12, R. 65 W., T. 11 N., three miles northwest of Gault School- house, Weld County, Colorado, by Mr. James Mellinger. Univ. Chicago No. 1547, portion of right ramus with My,^; Univ. Chicago No. 1550, portion of left ramus with Mt^; Univ. Chicago No. 1551, portion of left maxillaiy with M-~-; all collected by McGrew in the lower Brul^ beds, twelve miles north-northwest of Crawford, Dawes County, Nebraska. Univ. Chicago No. 1552, portion of left ramus with My_^ and alveoli of Py_y; Univ. Chicago No. 1553, portion of left ramus with My_3f and alveoli of Py-^; both collected by McGrew in the lower Brul^ beds ten miles northeast of Harrison, Sioux County, Nebraska. Univ. Chicago No. 1554, incomplete facial region with left P-, lingual portion of P^, and alveoli of I-"-, P-; collected by McGrew in the lower Brul^ beds two and one-half miles north of Chadron, Dawes County, Nebraska. Diagnosis. — As for the genus; for measurements see p. 256. Remarks. — The description and measurements given by Scott for P. crassus indicate that this species must be placed in the synonymy of D. gradata. The holotype of D. crassigenis (Cope) consists of a pair of incomplete mandibles with heavily worn lower molars. The specimen shows no characters that might serve to separate it from D. gradata. Cope's "Miothen" gracile (1873b, p. 8), referred to Domnina in 1874 (p. 470), was later regarded by him (1884, p. 796) as a synonym of Peratherium huntii (Cope). Oligocene Soricid and Erinaceids 249 Description. — The incomplete facial region shows the alveoli of the large I-, the small l-~^, and the small P^. P^ is much larger than P^, and consists of a single anteriorly placed cusp. External, internal, and posterior cingula are present; a low, rather sharp crest runs posteriorly from the apex of the cusp to the posterior cingulum. Fig. 60. Domnina gradaia Cope. Univ. Chicago No. 1551. Left M^ ^ in maxilla fragment; external, crown, and internal views. X 5 P- is lacking. This tooth is greatly reduced in some modern genera, and absent in others. The lingual portion only of P- is preserved. The protocone and hypocone are not well differentiated, and appear as anterior and posterior swellings on an internal ridge. M-~- are very similar to corresponding teeth oiBlarina, the only observable difference being that the strong posterior cingulum ex- tends externally to a point immediately posterior to the metastyle. 250 Field Museum of Natural History — Geology, Vol. VI The other existing forms differ in having the posterior border of the molars indented. I^ is strikingly different from that of other soricids. It is large and only moderately procumbent, forming an angle of about 30° with the ventral border of the ramus. The tip is broken off; the remaining part shows a very slight upward curve. The tooth is roughly oval in cross-section, the dorso-ventral diameter being greater than the transverse. The alveolus is anterior to that of Py, Fig. 61. Domnina gradata Cope. F. M. No. P16320, incomplete left mandible with I^, P^-M^j crown view. X5 and the ventral border of the tooth therefore does not extend back beneath the premolars — a decided contrast to the living forms. No accessory cusps are present, their places being taken by a low but sharp crest on the labial side of the dorsal face, at the base of which is a slight groove for the overlapping Fj. The outer side of the crown is heavily pigmented. The presence of P^-^ distinguishes Domnina from all other known shrews, living and fossil. The alveoli of these teeth are considerably smaller than those of Py and Py. The teeth themselves were in all probability greatly reduced. Py is slender and high. It is composed of a triangular cusp with its apex extending anteriorly, and a low Fig. 62. Domnina gradaUx Cope. F. M. No. P15320, incomplete left mandible with I^, p^M^; external and internal views. X5 251 252 Field Museum of Natural History — Geology, Vol. VI broad cingulum heel. Slight external and internal cingula are present. The cusp extends well forward over the alveolus of P^. The heavily pigmented molars decrease in size posteriorly. The trigonids are longer and higher than the talonids. The paraconid is a large, well separated cusp connected by a low ridge to the anterior Fig. 63. Domnina gradata Cope. Amer. Mus. No. 6353. Holotype. Portion of right ramus with right MY_g; external, crown, and internal views. X 6 ' slope of the protoconid. The metaconid is directly lingual to the protoconid. The latter cusp is sharply angulate externally. A prominent antero-external cingulum is present. An external cin- gulum is present on the hypoconid. The anterior slope of the hypo- conid extends downward and forward to the protoconid-metaconid crest, its internal slope (possibly including the hypoconulid) ► Fig. 64. Domnina gradata Cope. Univ. Chicago No. 1552. Portion of left mandible with M^_7j; external, crown, and internal views, crown view with specimen tipped to show the premolar alveoli. X6 253 254 Field Museum of Natural History — Geology, Vol. VI extends inward behind the entoconid to the Ungual margin of the tooth. ^ The entoconid of Mx_tj is a large, transversely com- pressed cusp, which is not united to the posterior wall of the tooth. It is connected by a sloping ridge to the posterior face of the meta- conid, thus forming a lingual wall to the talonid basin. On Mx the Fig. 65. Domnina gradata Cope. Univ. Chicago No. 1554. Incomplete facial region with left F*, lingual portion of P*; lateral, ventral, and dorsal views. X4 talonid is wider than the trigonid, on M^^ trigonid and talonid are about equal in width, and on M^ the talonid is greatly reduced. 1 We believe that the cusp regarded by Stirton (1930, fig. 1) as the entoconulid on the molars of Limnoecus is the entoconid, and that the rather slight projection immediately posterior to it is either the hypoconulid or merely a projection of the ridge which runs lingually from the hypoconid. An apparent inconsistency may also be noted in this paper. The well developed postero-internal cusp in Mj and Mj of Crocidura russula russula is labeled "end" on Fig. 2, whereas on p. 220 the following statement appears: "In the Crocidura group the entoconid is absent, and the entoconulid is well developed on Mx-ir." Oligocene Soricid and Erinaceids 255 The entoconid is either absent on M^ (F.M. No. P15320), or reduced to a low, inconspicuous ridge (Univ. Chicago No. 1552). The skull fragment adds nothing to the description already given by Scott. The rostrum is relatively wider than in most of the recent shrews examined, but is approximately equalled in this respect by Nectogale. The mandible has a sta-ikingly deep horizontal ramus. The ventral border is straight beneath the molars, but turns upward at an angle of about 30° beneath the premolars. The alveolar border is nearly horizontal from M^ forward to P^, turning downward toward the incisor under the anterior premolars. On the inner face of the ascending ramus of F.M. No. P15320 there is a small rounded notch above which the jaw is broken away; this may be the ventral border of an intertemporal fossa. This notch is situated below a line projected back from the cusps of the molars and is just slightly above a line projected along the alveolar border. It is possible that this apparent border is an angular break which has been rounded by the erosional action of water. In Heterosorex, a primitive form sharing several characters with Domnina, the intertemporal fossa is absent, a fact which might argue for absence in Domnina. More complete material will be necessary before the presence or absence of the fossa in Domnina can be proved definitely. The holotype of "P. crassus" lacked the condyle and angle; Scott made no mention of the fossa, the specimen presumably being broken in this region also. Relationships. — It is not possible at present to arrive at a satis- factory conclusion regarding the position of Domnina in soricid phylogeny. The genus is primitive (for the Soricidae), and as such might be regarded as structurally ancestral to many of the living forms. That it is actually ancestral to any of them seems doubtful. The occurrence of Miocene species referred to living genera (See the catalogue of fossil shrews given by Stir ton 1930, pp. 220-223.) would appear to indicate that the differentiation of the family into the groups now surviving may have taken place as early as the Oligo- cene, if not earlier. Domnina and the Miocene genus Heterosorex (Gaillard 1915) have several characters in common, as stated above. The molar patterns and proportions of the two genera are very similar, and the horizontal rami of both are deep and heavy, and of comparable proportions. Heterosorex, however, retains part of the zygomatic arch, a structure already lost in Domnina (Scott 1894, p. 446). The antemolar dentition of Heterosorex is, on the other hand, much more specialized than that of Domnina. There are but two upper and 256 Field Museum of Natural History — Geology, Vol. VI lower premolars, and Py is much larger than Py. The two forms may be members of a group of related genera, or the resemblances between them may indicate nothing more than a common retention of certain heritage characters. Among living genera, Blarina closely approaches Domnina and Heterosorex in upper and lower molar structure. Stirton (1930, p. 219) considers Heterosorex to be more closely related to the "Blarinae" than to other living shrews. This may be the case, but it will require much more fossil material than is now available before any satisfactory phylogenetic arrangement of the family can be made. MEASUREMENTS IN MILLIMETERS u. c. 1554 P^a.-p 1.0 P^ tr 0.9 M^, a.-p M^ tr M^, a.-p u. c. 1551 Holotype A. M. 5353 PI 5320 U. C. 1549 U. C. 1552 U. C. 1547 'P.erassus"* M^ tr. Length, Pt-Ms 7.0 P;i, a.-p 1.3 P5, tr 1.2 Length Mt_u 5.8 5.7 ... 5.4 5.8 5. Of Mt, a-p 2.2 2.5 2.2 2.4 My.tr 1.4 1.4 1.5 1.5 1.5 M 5, a.-p 2.1 1.9 1.9 1.8 2.0 M^, tr 1.2 1.3 1.4 1.4 1.4 Mg, a.-p 1.6 1.6 ... 1.5 1 M^.tr 1.1 1.1 ... 1.1 1 Depth of ramus under Mt 2.4 2.5 2.5 2.2 2.5 J * Measurements from Scott. t The lower figure is perhaps due to less precise methods of measuring. X 2.5 under M-j. The measurements given for D. crassigenis by Cope in 1874 do not agree with those given by him in 1884; consequently neither set is included here. NOTE ON MYSTIPTERUS VESPERTILIO HALL This form, based on a fragment of the left mandible with M^, from the Esmeralda, lower Pliocene, of Nevada, was described by Hall (1930) as a bat and referred to the Vespertilionidae. He com- pared the specimen with Miniopterus and noted the following dis- tinctions, regarding them as of generic significance only: (1) para- conid-protoconid blade distinctly notched and with strong cingulum at its base; (2) each of the five principal cusps present but without accessory cusps; (3) trigonid longer than talonid, even on medial side; (4) talonid basined; (5) masseteric fossa deep. All of these characters, however, are typically soricid and appear to us, therefore, to be of ordinal significance. Oligocene Soricid and Erinaceids 257 Other important differences from Miniopterus and resemblances to the Soricidae are (a) masseteric fossa extending downward much nearer the ventral margin of the jaw; (b) that part of the jaw behind the anterior margin of the masseteric fossa notably less upturned; (c) talonid relatively shorter; (d) ramus relatively heavier. From the evidence available we believe that Mystipterus vespertilio must be transferred from the Vespertilionidae to the Soricidae. Erinaceidae Bonaparte The genus described below has not hitherto been placed in this family. The reasons for the assignment are given in the discussion of affinities on p. 267. Metacodon Clark MetcLcodon Clark 1936, in Scott and Jepsen, Trans. Amer. Phil. Soc, N. S., 28, p. 22. Genotype. — Metacodon magnus Clark 1936. Distribution. — Lower Oligocene, Chadron beds, South Dakota; Middle Oligocene, Lower Brul4 beds, Nebraska and Colorado. Emended diagnosis. — I^, Cy, P^, M|. P- submolariform with parametocone, strong parastyle, metastyle crest. Upx)er molars narrow transversely, protoconules absent, metaconules absent or very small, external shelves wide with stylar cusps varyingly de- veloped, parastyles strong, metastyle crest on M-~-, hypocone of M- small. Lower C small, erect. Py two rooted; P^ small in com- parison with Pif-ir; Pt submolariform with metaconid not completely separated from, directly lingual to, and lower than protoconid; anterior cusp and talonid small, low. Trigonids of lower molars much higher than talonids, base of paraconid situated high on talonid. Horizontal ramus slender, shallow. Metacodon magnus Clark M. magnus Clark 1936, in Scott and Jepsen, Trans. Amer. Phil. Soc, N. S., 28, p. 23, pi. 2, figs. 5, 5a. Holotype. — Princeton Univ. Mus. No. 13835, portion of right mandible with P^^-M^. Horizon. — Chadron beds. Big Corral Draw, South Dakota. Diagnosis. — Larger than M. mellingeri, Mx_^ 6.0 mm.^ Upper molars relatively wider transversely, hypocone more prominent. P^^ with metaconid larger and better separated from protoconid. 1 We are indebted to Dr. Clark for supplying this measurement from his MS. 258 Field Museum of Natural History — Geology, Vol. VI M^ relatively larger and with narrower talonid, hypoconulid forming sharp posterointernal angle on talonid basin rim, rim continuous between entoconid and metaconid. Remarks. — Knowledge of the upper molars of M. mellingeri per- mitted identification of upper molars of M. magnus among miscel- laneous lots of Chadron specimens in the Princeton University Museum. Comparisons between the two species in the postdental region of the mandible showed that the character of M. magnus upon which Clark chiefly relied in erecting the genus — the reduction of the angle and lowering of the condyle — cannot be regarded as established. The holotype was found in fragments, and the contacts in the posterior region, although apparently good, were found on close re-examination to be inexact. This matter will be discussed by Dr. Clark in his forthcoming paper on the Chadron beds. The genus is fully validated, however, by the new material described here. Metacodon mellingeri sp. nov. (figs. 66, 67, 69-72) Holotype. — F.M. No. P15321, incomplete right mandible with Px, Pt> My_^, and alveoli of F^-^. Horizon and locality. — Lower Brul^ beds. Sec. 12, R. 65 W., T. 11 N., three miles N.W. of Gault Schoolhouse, Weld County, Colorado. Collected by Mr. James Mellinger, in whose honor the specific name is given. Paratypes. — F.M. No. P15322, incomplete left mandibular ramus with P^_4 and alveoh of It-Pi?; F.M. No. 15323, portion of right mandibular ramus with C, Py-i?; Univ. Chicago No. 1555, left max- illary with P^-M^; Univ. Chicago No. 1556, M^-"^ and portion of right mandible with M-^^s ; Univ. Chicago No. 1557, portion of left mandible with P^-M^; Univ. Chicago No. 1558, portion of right man- dible with P4-M7J; Univ. Chicago No. 1559, left mandible with P4- M.^; Univ. Chicago No. 1560, portion of left maxillary with P^-M^. The Field Museum paratypes were collected by Mr. Mellinger at the same horizon and locality as the holotype. Univ. Chicago Nos. 1555, 1556, 1557, 1558, and 1560 were collected by McGrew from the lower Brul6 (Middle "Oreodon" level), twelve miles N. N.W. of Crawford, Dawes County, Nebraska; No. 1559, from the same horizon eight miles N.E. of Harrison, Sioux County, Nebraska. Diagnosis. — Smaller than M. magnus, My-^ 4.9 — 5.3 mm. Upper molars relatively narrower, hypocone slightly less developed. Px with metaconid only slightly separated, anterior cusp lower and smaller. M^ relatively smaller with narrower talonid, hypoconulid Oligocene Soricid and Erinaceids 259 projecting posteriorly as on My_^, basin rim not continuous between entoconid and metaconid. Description. — In strong contrast to the leptictids, P- is not molariform. The parametacone of P- is large and high, wider trans- FiG. 66. Metacodon mellingeri sp. nov. Univ. Chicago No. 1655. Paratype. Right P^-M'" in maxilla fragment; external, crown, and internal views. X 5 versely and higher than either the paracone or metacone of M^. The metastyle crest is well developed. The protocone is prominent and sharp, but is smaller than that of M-~^; it is somewhat inclined anteriorly. A poorly defined ridge runs from its apex to the base 260 Field Museum of Natural History — Geology, Vol. VI of the antero-internal slope of the parametacone. The parastyle is large, cuspidate, directly anterior to and well separated from the parametacone, and strongly protruding. It forms a right angle with the ridge running anteriorly from the protocone. The meta- style is a stout cusp inclined posteriorly. It bears two small stylar cuspules arranged antero-posteriorly on its anterior slope, the posterior being the larger of the two. The external cingulum con- nects the parastyle and metastyle, but is very weak external to the parametacone. The anterior cingulum is merely a slight swelling at Fig. 67. Metacodon mellingeri sp. nov. Univ. Chicago No. 1556. Paratype. Left M"^ • portion of right ramus with M^_-^; crown views. X 5 the base of the anterior slope of the protocone. The posterior cingulum is more extensive than on the molars, extending from the metastyle inward to the postero-internal corner of the protocone, where it thickens to a slight hypocone. On M- the paracone and metacone are subequal and lower than the parametacone of P-. They are fairly close together, much as in Tupaiodon, and have a short thick ridge connecting their bases. The protocone is large, high and notably inclined anteriorly, this inclination imparting a somewhat twisted appearance to the tooth Ridges from the protocone apex run externally around the bases of the para- and metacone to the para- and metastyle, the anterior Oligocene Soricid and Erinaceids 261 ridge being considerably the higher of the two. In contrast to Tupaiodon neither ridge shows any proto- or metaconule thickening. To judge from P-, it would appear that the developing posterior ridge may have captured the external portion of the posterior cingulum. The parastyle is large, low, and blunt. It is strongly protruding, but not so much as that of P-, and is situated anterior and slightly external to the paracone. The parastyle bears on its posterior slope a stylar cuspule, which is demarcated from it ex- ternally by a small shallow groove. The metastyle projects externally to a much greater degree than the parastyle. It is rather wider transversely than that of P-, and lacks stylar cuspules. The meta- style crest is prominent, a decidedly erinaceid character. The external cingulum forms a moderate shelf which turns rather sharply Fig. 68. view. X 5 fMetacodon sp. Univ. Chicago No. 1561. Left M in maxilla fragment; crown outward and backward around the metastyle. The anterior cingulum is similar to that of Tupaiodon, extending internally to the antero- intemal slope of the protocone. The posterior cingulum also resem- bles that of the Mongolian form, but the hypocone is not so robust. The paracone and metacone of M- differ from those of M- in that the notch between them is somewhat deeper, and the metacone is postero-internal to the paracone. The protocone is less inclined anteriorly, and its posterior ridge is not so robust. The very large parastyle is antero-extemal to the paracone. As in most erinaceids, it is larger and more salient than the metastyle on this tooth, the reverse of the conditions seen on M^. Its stylar cuspule is larger than that of M^^ and is demarcated externally by a posterior groove in addition to the median one. The metastyle crest is shorter and notably lower than that on M-. The external shelf is wider than on the preceding tooth and has a deep median cleft. The anterior 262 Field Museum of Natural History — Geology, Vol. VI cingulum extends around to the lingual slope of the protocone, just meeting the posterior cingulum. The latter is similar to that on M-, but the hypocone is slightly smaller. The paracone of M- is slightly higher than the metacone, and rather more antero-external to it than on M-. The protocone is relatively higher than on M-"- and actually higher than the para- FiG. 69. Metacodon mellingeri sp. nov. F. M. No. P15322. Paratype. Portion of left ramus with symphysis, P-!r_x', external, crown, internal, and anterior views. X5 cone. The posterior ridge is higher than on the preceding teeth enclosing a moderately deep central basin. The parastyle is ex- tremely large and projects externally to a marked degree. Its stylar cusp is very slight. The metastyle is greatly reduced and the crest is lacking. The external cingulum forms a much smaller shelf than on the preceding molars. The hypocone is merely a slight enlarge- ment of the internal extremity of the posterior cingulum. Oligocene Soricid and Erinaceids 263 The description given above has been taken from Univ. Chicago No. 1555, which is by far the best upper dentition so far obtained. The few differences exhibited by other specimens may be briefly noted. On Univ. Chicago No. 1556, and on this specimen only, there occurs a small ridge running directly inward from the base of the metacone to the ridge extending posteriorly from the protocone apex. This ridge is important since it appears to represent the metaconule (fig. 67). A specimen collected in the lower Brul^ north- east of Harrison and consisting of M^~^ in a maxilla fragment (Univ. Fig. 70. Metacodon mellingeri sp. nov. F. M. No. P15323. Paratype. Fragment of right ramus with portion of symphysis, canine, Py_7; external, crown, and internal views. X 5 Chicago No. 1561) differs in several characters from the remainder of the material and is probably not referable to M. mellingeri and possibly not to Metdcodon (fig. 68). Both teeth are notably narrower than in the paratypes, and have the anterior cingulum somewhat thickened internally. The metastyle of M^ is external rather than postero-extemal to the metacone, and the external cleft between the para- and metastyle is much deeper. On M- the hypocone is prominent, and extends posteriorly to a slight extent. When more material is obtained, this specimen may prove to represent a dis- tinct genus. 264 Field Museum of Natural History — Geology, Vol. VI The lower incisors are not preserved. The alveoli show that they were subequal, moderately procumbent, and arranged in a nearly antero-posterior row. The roots were long. The only canine preserved is considerably worn. It is a small, blunt, erect, single rooted tooth, apparently lower than P4 and the molars. The labial side is convex, the lingual slightly fluted anteriorly. The transverse diameter is slight. The single rooted canine of this form and of Tupaiodon morrisi indicate that Leche may have been incorrect in his belief that double-rooted canines were primitive for the family. All the premolars are double-rooted. Py is notably procumbent and consists of a robust blunt cusp and a minute posterior cuspule. Pif is a much larger tooth than either Py or P^. It is erect and long with a large cusp at the anterior extremity and with a small heel. P^ is very small in comparison with P^; it has a stout central cusp, a small anterior cuspule, and a slight heel. P4 is as high as, or in some cases even slightly higher, than the molars. The metaconid is partially separated from the protoconid, but not to such an extent as in M. magnus, and is definitely the lower of the two cusps. A variable antero-external cingulum is present which extends upward to the low, small anterior cusp. The heel is fairly well developed and somewhat excavated transversely; it is rather high on the lingual side and slopes down toward the labial, forming a semi- trenchant heel. The first and second lower molars are nearly subequal, M^ being almost imperceptibly the larger of the two; M^ is somewhat smaller. The trigonid and talonid of My are approximately equal in width, but the trigonid is much higher and slightly longer. The protoconid forms a sharp external angle, and is somewhat larger and higher than the metaconid. The paraconid is lower than the meta- conid and is situated in front of and close to it. The bases of these cusps are high on the trigonid and consequently the cusps them- selves become obliterated by moderate wear. The antero-external cingulum is well developed but does not reach the paraconid. The talonid cusps together form an elevated posterior rim, the hypo- conid being postero-external and angulate externally, the entoconid postero-internal. The hypoconulid is in the middle of the posterior rim and projects posteriorly to a moderate degree, less, however, than in Ictops. Ridges from the apices of the entoconid and hypo- conid slope down anteriorly to the base of the trigonid. The talonid basin is well defined and quite deep. 265 266 Field Museum of Natural History — Geology, Vol. VI On M^ the trigonid is wider transversely and narrower antero- posteriorly, with the paraconid even closer to the metaconid than on My. The hypoconulid is rather more internal in position, and more closely connected to the entoconid. The latter cusp is smaller and more anterior in position. The talonid of M^ is relatively nar- rower than on My_^; the hypoconid is not as angulate externally, and the entoconid is notably reduced in size. The horizontal ramus of the mandible is shallow, slender, and long; the ventral border curves upward beneath the ascending ramus. The ligamentous symphysis extends posteriorly to a point ),^^-JgSfBri&f^ Fig. 72. Metaeodon mellingeri sp. nov. Univ. Chicago No. 1559, portion of left ramus with P^-My U. C. 1567, portion of left ramus with P^-M^. U. C. 1558, portion of right ramus with P^-M™ Crown views of paratypes. X 5 beneath the anterior half of P^. The ascending ramus arises well behind M^ and is moderately heavy. The deep masseteric fossa is sharply defined anteriorly by a strong ascending ridge. The anterior mental foramen is either beneath Py (P15323) or beneath the an- terior root of Pif (P15321, P15322), the posterior beneath the pos- terior root of P^. Affinities. — Clark (in Scott and Jepsen 1936, p. 22) referred Metaeodon to the Leptictidae, and regarded it as a probable descen- dant of Leptacodon. This reference was fully justified by the material at his disposal, since the lower molars are very similar to those of typical leptictids. The new material described here shows, how- ever, that the genus can hardly be placed in that family. The Oligocene Soricid and Erinaceids 267 undivided parametacone of P-, the wide external shelves, strong styles, metastyle crests, and more centrally placed paracone and metacone of the upper molars sharply distinguish Metacodon from the Leptictidae. Members of this family had already acquired the molariform P^ in the upper Cretaceous (e.g. Gypsonictops). The holotype of Leptacodon tener, the genotypic species, is typically leptictid. The genus therefore cannot be considered as ancestral, or even closely related, to the form under consideration. There is a possibility that some of the species that have been placed in Leptacodon may not belong in the genus, but thS question is out- side the scope of the present paper. Since Metacodon is excluded from the Leptictidae, its family position remains to be determined. In all the characters in which the genus differs from leptictids, it shows a decidedly close resem- blance to the Erinaceidae. The various genera of this family, living and extinct, appear to be fundamentally similar in dental characters to the form under discussion. The likeness of Pf appears to be especially significant. The upper molars of the recent genera are wider transversely, the hypocones are larger, and the external cingula and metastyle crests are less prominent. Such differences as these, however, are to be expected between an Oligocene genus and existing forms. Tupaiodon, of the Oligocene of Mongolia, tentatively regarded as a tupaiid by Matthew and Granger (1924, p. 1, fig. 1), but transferred to the Erinaceidae by Simpson (1931, p. 12), and Proterixoides (Stock 1935) of the upper Eocene of Cali- fornia are partly intermediate between Metacodon and the more advanced members of the family, particularly as regards the trans- verse diameter of the upper molars. The lower molars of Metacodon differ from those of most erinaceids in having higher and more antero- posteriorly compressed trigonids, and distinct hypoconulids; char- acters which approach those of the Leptictidae. Although in part primitive, these differences may, and probably do, indicate that the genus is a representative of a somewhat aberrant side branch of the Erinaceidae,' but we do not believe that they are of sufficient im- portance to warrant exclusion from the family. The new genus Ankylodon described below, appears to be structurally intermediate in lower molar pattern between Metacodon and the more typical erinaceids. ITupaiodon minutus has a very distinct hypoconulid on M^ (fig. 73), as has Proterixoides davisi (Stock 1935, pi. 1, fig. 2), ^ The occurrence of Proterix, a more typical member of the family, in the Brule tends to support this view. 268 Field Museum of Natural History — Geology, Vol. VI and we have detected a small hypoconulid on My_^ of a specimen of Hemiechinus in the Field Museum collection (No. 35811). Metacodon shows similarities to the tupaiids, but the resem- blances are not as close to this family as to the erinaceids. The upper molars are rather like those of early pantolestids, e.g. Bes- soecetor, but the very different Fi seem definitely to preclude close relationship. The position of Metacodon within the Erinaceidae is uncertain. The living membgrs of the family are customarily divided into two subfamilies, the Erinaceinae and the Echinosoricinae.^ The former is certainly a natural group, but the latter may include more than one stock. In any event, Metacodon appears to stand in about the same structural relation to the living echinosoricines as they in turn Fig. 73. ?Tupaiodon minutus Matthew and Granger, with M^ ■,■; crown view. X5 F. M. No. P14124. Portion of left ramus stand to the erinaceines. Further than this it is impossible to go at present. Entomolestes, from the middle Eocene Bridger and lower Eocene Gray Bull beds, was provisionally placed by Matthew (1909b, pp. 541-542, pi. 50, fig. 2) in the Talpidae, although he stressed its resemblances to the Tupaiidae, and later (1918, p. 597) referred it tentatively to this family. Simpson at first referred it to the Adapi- soricidae (1928, pp. 5-6), but later (1931, p. 17) considered it to be an erinaceoid and possibly a leptictid. We believe that the non- molariform P^^ indicates that it should be referred provisionally to the Erinaceidae. One rooted V^^^ occur in the living genus Hylomys. The heel of P^ is very similar to that of Metacodon, and M^^.^ are close to those of ITwpaiodon minutus (fig. 73). 1 Cabrera (1925, p. 58) has shown that the widely used named Gymnura Lesson is unfortunately preoccupied and must be replaced by Echinosorex Blain- ville 1838. Oligocene Soricid and Erinaceids 269 MEASUREMENTS IN MILLIMETERS u. c. 1555 P* a.-p.* 2.0 P*tr.* 1.0 Length M^-^ 3.8 M^a.-p 1.2 M^-tr 2.3 M^a.-p 1.0 M^tr 2.3 M^ a.-p 9 M^tr 1.9 C, a.-p C, tr Length Px-M^ Length Pt-t Pt, a.-p Pt tr Pt. a.-p P^ tr Ptj, a.-p Putr Pt. a.-p Pi tr Length Mr-g Mt, a.-p Mxtr M J, a.-p M^ tr M^, a.-p M^tr Depth of ramus under Py Depth of ramus under M^ u. c. 1556 Holotype P15321 P15322 P15323 U. C. 1567 U. C. 1558 U. C. 1559 10.9 6.1 1.2 .5 1.6 .9 4.9 1.8 1.1 1.9 1.2 1.5 .9 2.1 2.7 1.6 * The antero-poeterior diameters of the upper cheek teeth have been measured across the bases of the lingual slopes of the paracones and raetaeones, the transverse across the centers of the teeth. Ankylodon^ gen. nov. Genotype. — Ankylodon annectens sp. nov. Distribution. — Middle Oligocene, Lower Bnil^ beds, Colorado. Diagnosis. — P^ submolariform, protoconid and metaconid curv- ing backward, metaconid lower than protoconid. My_^ with trigo- nids extending forward in advance of the anterior root and with a moderate backward curvature; trigonids lower, and teeth as a whole broader than in Metacodon; external sides of protoconids and hypoconids strongly convex; paraconids notably reduced; hypo- conulids distinct, connected with hypoconids and entoconids. Ankylodon annectens^ sp. nov. (fig. 74) Holotype. — F.M. No. P15326, portion of left ramus with Fx-M^. ^ In allusion to the curved trigonids. * In allusion to its structurally intermediate position between Metacodon and more typical erinaceids in lower molar structure. 270 Field Museum of Natural History — Geology, Vol. VI Horizon and Locality. — Lower Brul^ beds. Sec. 12, R. 65 W., T. 11 N., three miles N.W. of Gault Schoolhouse, Weld County, Colorado. Collected by Mr. James Mellinger. Diagnosis. — As for the genus, for measurements see below. Fig. 74. Ankylodon anneetens gen. et sp. nov. F. M. No. PI 5326. Holotype. Portion of left ramus with P^-M^; external, crown, and internal views. X5 Description. — The anterior cusp of P4 forms a low shelf across the front of the tooth. The metaconid is well separated from the protoconid, rather more so than in Metacodon mellingeri. The heel is very similar to that of Metacodon, but is definitely lower on the tooth. Cingula are not developed. My_7j^ are almost identical in Oligocene Soricid and Erinaceids 271 structure; both extend anteriorly, and display the same degree of backward curvature of the trigonid and strong convexity of the labial faces of the proto- and hypoconids. The trigonids are narrow antero-posteriorly; the paraconids are reduced to mere ridges. Weak antero-external cingula are present. The protoconids are higher than the metaconids. The entoconids are larger and rather higher than the hypoconids. As in Metacodon, the three talonid cusps form an elevated posterior rim. The hypoconulids are more nearly in transverse alignment with the hypo- and entoconids. The tri- gonids and talonids are sharply separated by a deep transverse groove. The small portion of the ramus present is nearly uniform in depth throughout. The small posterior mental foramen is an- terior to Pt, presumably beneath the posterior root of P^ as in Metacodon. Ankylodon is definitely erinaceid in structure. Its various points of similarity to Metacodon therefore strengthen the reference of the latter genus to the Erinaceidae, as noted above. MEASUREMENTS IN MILLIMETERS P^a.-p 1.6 P^tr 1.2 Mya.-p 1.8 Mrtr 1.4 M^a.-p 2.0 M^tT 1.6 Depth of ramus under Mx 2.8 REFERENCES Arnbach-Christie-Linde, a. 1912. On the Development of the Teeth of the Soricidae: an Ontogenetical Inquiry. Ann. Mag. Nat. Hist., (8), 9, pp. 601-625, pis. 18-19, text-figs. 1-9, Cabrera, A. 1925. Genera Mammalium. Insectivora, Galeopithecia. Mus. Nac. Cienc. Nat. Madrid. Pp. 1-232, pis. 1-17. Cook, H. J. and Cook, M. C. 1933. Faunal lists of the Tertiary Vertebrata of Nebraska and adjacent areas. Nebr. Geol. Surv. paper No. 5, pp. 1-58. Cope, E. D. 1873a. Third notice of extinct Vertebrata from the Tertiary of the plains. Pal. Bull. No. 16, pp. 1-8. 1873b. Synopsis of new Vertebrata from the Tertiary of Colorado, obtained during the summer of 1873. Pp. 1-19. Washington, Government Printing Office. 1874. Report on the Vertebrate Paleontology of Colorado. Ann. Rept. Geol. Geog. Surv. Territories for 1873, pp. 427-533, pis. 1-8. 1884. The Vertebrata of the Tertiary formations of the West. U. S. Geol. Surv. Territories, Book 1, pp. i-xxxv, 1-1009, pis. l-75a. 272 Field Museum of Natural History — Geology, Vol. VI Gaillard, C. 1915. Nouveau genre de musaraignes dans les D6p6ts miocfenes de la Grive Saint-Albans (Is^re). Ann. Soc. Linn. Lyon, 62, pp. 83-98, text-figs. 1-8. Hall, E. R. 1930. A new genus of Bat from the later Tertiary of Nevada. Univ. Calif. Publ. Bull. Dept. Geol. Sci., 19, pp. 319-320, pis. 38. Haug, E. 1922. Traits de G6ologie. II; Les p6riodes g^logiques. Pp. 1153-2024, pis. 72-135, text-figs. 346-485. Paris, Armand Colin. Hay, O. p. 1902. Bibliography and Catalogue of the fossil Vertebrata of North America. Bull. U. S. Geol. Surv. No. 179, pp. 1-868: 1930. Second Bibliography and Catalogue of the fossil Vertebrata of North America. Carnegie Inst. Wash. Publ. No. 390, Vol. 2, pp. 1-1074. Matthew, W. D. 1899. A provisional classification of the fresh water Tertiary of the West. Bull. Amer. Mus. Nat. Hist., 12, pp. 19-75. 1909a. Faunal lists of the Tertiary Mammalia of the West. Bull. U. S. Geol. Surv. No. 361, pp. 91-138. 1909b. The Carnivora and Insectivora of the Bridger Basin, Middle Eocene. Mem. Amer. Mus. Nat. Hist. 9, 291-567, pis. 42-52, text-figs. 1-118. Matthew, W. D. and Granger, W. 1924. New Insectivores and Ruminants from the Tertiary of Mongolia, with remarks on the correlation. Amer. Mus. Novit. No. 105, pp. 1-7, text-figs. 1-3. Scott, W. B. 1894. A new Insectivore from the White River beds. Proc. Acad. Nat. Sci. Phila., 1894, pp. 446-448. Scott, W. B. and Jepsen, G. L. 1936. The Mammalian Fauna of the White River Oligocene, Part 1, Insec- tivora and Carnivora. Trans. Amer. Phil. Soc. N. S., 28, pp. 1-153, pis. 1-22, text-figs. 1-7. Simpson, G. G. 1928. A new Mammalian Fauna from the Fort Union of Southern Montana. Amer. Mus. Novit. No. 297, pp. 1-15, text-figs. 1-14. 1931. A new Insectivore from the Oligocene, Ulan Gochu horizon, of Mongolia. Ibid. No. 505, pp. 1-22, text-figs. 1-5. Stirton, R. a. 1930. A new genus of Soricidae from the Barstow Miocene of California. Univ. Calif. Publ. Bull. Dept. Geol. Sci., 19, pp. 217-228, text-figs. 1-2. Stock, C. 1935. Insectivora from the Sespe uppermost Eocene. California. Proc. Nat. Acad. Sci., 21, pp. 214-219, pi. 1.