OONTRIBUTIONS TO PALIEONTOLOGY I

NEW MIDDLE PLIOCENE RODENT AND LAGOMORPH FAUNAS FROM OREGON AND CALIFORNIA

BY ROBERT w. WILSON

With three plates

[Preprinted from Carnegie Institution of Washington Publication No. 487, pages 1 to 19. June 30, 1937] CONTRIBUTIONS TO PALEONTOLOGY I NEW MIDDLE PLIOCENE RODENT AND LAGOMORPH FAUNAS FROM OREGON AND CALIFORNIA

BY ROBERT w. WILSON

With three plates

[Issued June 30, 1937] NEW MIDDLE PLIOCE1 FAUNAS FROM 01

INT The purpose of this paper CONTENTS in the collections of the Califo1 PAGE coming from widely separated Introduction ...... 3 mately the same age. The fi Rome Fauna ...... 3 Pliocene strata near Rome, : Systematic Description of Fauna ...... 4 some diversity of type althot Rodentia ...... 4 known by very fragmentary re Mylagaulus? cf. iµonodon Cope ...... 6 from the Kern River deposits Dipoides stirtoni Wilson ...... 8 Kern River rodents and lagor Castor? sp ..•.••.•....•...... ••••...... •.•...... •.....••. 8 with them, share with other 'I Goniodontomys disjunctus n. gen. and sp ...... 9 Valley the important task of Lagomorpha .....•...... 12 tween the nonmarine deposits Hypolagus vetus (Kellogg) ...... 12 ~ ard marine sections of the Pa< I Hypolagus sp...... 13 Kern River Fauna ...... •...... 13 The illustrations for this pi Systematic Description of Fauna ...... 14 H. Wm. Menke, and have bee1 Rodentia ...... 14 plates by John L. Ridgway. Citellfis? sp...... 14 Peromyscus pliocenicus n. sp ...... 15 RO Lagomorpha ...... · .. . 17 The Rome fauna occurs in Hypolagus near limnetus Gazin ...... 17 Creek drainage, tributary to t Hypolagus small sp...... 19 of Rome, Malheur County, Or the section consists of light pebble beds, and mudstones. a narrow band of sediments s exposed along the eastern sic Crooked Creek. Except for B entire rodent fauna was obtain The following rodent and Iai the Rome fauna:

Rodentia Mylagaulus1 cf. mon-0don C Dipoides stirtoni Wilson Castor'! sp. Goniodontomys disjunctus n. Lagomorpha Hypolagus vetus (Kellogg) Hypolagus sp. NEW MIDDLE PLIOCENE RODENT AND LAGOMORPH FAUNAS FROM OREGON AND CALIFORNIA

INTRODUCTION The purpose of this paper is the description of two rodent faunas in the collections of the California Institute of Technology. Although coming from widely separated areas, these assemblages are of approxi­ PAGE mately the same age. The first fauna to be discussed is that from ····· ·· ·········· ·········· ·· 3 3 Pliocene strata near Rome, Malheur County, Oregon. It exhibits ················· ············ some diversity of type although the associated larger mammals are ·· ··· ···················· ··· ·· 4 ·················· ········· · 4 known by very fragmentary remains. The second rodent fauna comes 6 from the Kern River deposits, San Joaquin Valley, California. The ··· ··· ·········· ·· ·· ········· 8 Kern River rodents and lagomorphs, and the larger mammals found ····· ······················ 8 with them, share with other Tertiary assemblages of the San Joaquin sp ...... 9 Valley the important task of determining the time relationships be­ ···· ·········· ····· ····· ····· 12 tween the nonmarine deposits in which they are found and the stand­ ····························· 12 ard marine sections of the Pacific Coast. ······· ··················· ··· 13 The illustrations for this paper are from photographs by the late ····························· 13 14 H. Wm. Menke, and have been carefully retouched and arranged into plates by John L. Ridgway. ·· ················ ···· ······· 14 ···························· · 14 ··· ··· ·········· ········ ·· ··· 15 ROME FAUNA ····························· 17 The Rome fauna occurs in lake beds exposed along the Crooked ····························· 17 Creek drainage, tributary to the Owyhee River, five miles southwest 19 ················ ··········· ·· of Rome, Malheur County, Oregon. Locally, in the fossiliferous area, the section consists of light green to white tuffaceous sandstones, pebble beds, and mudstones. Most of the fossils were obtained from a narrow band of sediments situated near the base of the section as exposed along the eastern side of Dry Creek, a small tributary of Crooked Creek. Except for an isolated cheek-tooth of Castor?, the entire rodent fauna was obtained from this limited stratigraphic zone. The following rodent and lagomorph types have been recorded from the Rome fauna:

Rodentia Mylagaulus1 cf. monodon Cope Dipoides stirtoni Wilson Castor1 sp. Goniodontomys disjunctus n. gen. and sp. Lagomorpha Hypolagus vetus (Kellogg) Hypolagus sp. 3 4 CONTRIBUTIONS TO PALlEONTOLOGY NEW PLIOCENE RODENT FAU

Associated with the above assemblage are rather fragmentary mam­ place the female at a decided malian remains representing a diverse fauna. These remains have fornia Institute of Technology not been studied in detail, but the following forms have been recog­ cene and Pliocene mylagaulids nized in the fauna: presence of horns if such struct no horns were noted, and it is e Sca'/)anus(?) sp. Felid large sp. and similar specimens found E Pliohippus near spectans (Cope) Felid small sp. (possibly Machrero- from published descriptions, h1 Teleoceras fossiger? (Cope) dont) Prosthennops sp. Camelid large sp. rare. Sphenophalos nevadanus Merriam Camelid small sp. It can be shown rather satisf Lutra sp. Cervid sp. goes considerable change durin1 Plionictis near ogygia (Matthew) Mastodopt sp. ation in the skull is not known Lutravus cf. halli Furlong Turtle, fish, reptile, and bird remains question of individual variatior ?JElurodon sp. Canid (in JElurodon-Osteoborus cies the individual variation of group) probably n. gen. same stage of wear is not grea· by this statement is applied ,. Work on the Rome fauna has not been intensive enough to de­ species in proportion to the nu termine the exact time relations of the assemblage. However, a mid­ to be recorded. Moreover, soi dle Pliocene age or a stage of evolution comparable to that represented Inetitute collections points to by the Rattlesnake-Thousand Creek faunas seems certain. More specimens from the same zone : particularly, the Rome rodent fauna suggests by the presence of By means of relatively unw1 Castor? and of a vole that it is at least as advanced as the Thousand in the upper premolar the der Creek fauna. basins in the unworn teeth and The presence, in the Rome assemblage, of a mole, beavers, an otter, nal areas from particular cusp! and fish remains points to a rather moist environment with bodies of cheek-teeth of such forms as water in the immediate vicinity. The limited collecting area, the lithol­ susceptible to a similar treatn ogy of the sediments, and the large number of specimens of Dipoides countered in tracing the deriv: suggest that the deposits are lacustrine in origin. upper premolar. Use of this appreciation of the differences SYSTEMATIC DESCRIPTION OF FAUNA than can be obtained by citing arrangement in varying numbE RODENTIA of study combined with sectio1 Mylagaulide ally lead to an understanding The Mylagaulidre present many problems which need to be solved grinding teeth. Such work w before a satisfactory understanding of the group is reached. Not only thorough study of species whicl are the various species in considerable confusion, but even the genera mens and in which teeth in all are in an uncertain state of definition. The relation of horned to Doubtless a greater number hornless types has not been determined, nor have the limits of indi­ existence than have been desci vidual and age variation been fixed. nently burrowing types might It has been suggested that the presence or absence of horns is a sex and that consequently a numt character. However, as Matthew has pointed out, no other known in each faunal stage. However rodent possesses a like amount of sex distinction, although, as he states group, the establishment of n further, this argument is partly vitiated by the fact that no other procedure. rodent possesses horns. Such a variation, in a burrowing form in Pliocene mylagaulids are kn which the horns were presumably used as digging implements, would Possible exception is Epigault EONTOLOGY NEW PLIOCENE RODENT FAUNAS FROM OREGON AND CALIFORNIA 5 are rather fragmentary mam­ place the female at a decided disadvantage. Moreover, the Cali­ fauna. These remains have fornia Institute of Technology collections embrace a number of Mio­ 1wing forms have been recog- cene and Pliocene mylagaulids complete enough to demonstrate the presence of horns if such structures were actually present. However, no horns were noted, and it is extremely unlikely that all these skulls :l large sp. and similar specimens found elsewhere represent females. Judging :I small sp. (possibly Machrero- from published descriptions, horned types are everywhere relatively mt) rare. ielid large sp. It can be shown rather satisfactorily that the tooth-pattern under­ ielid small sp. rid sp. goes considerable change during the life of the individual. The vari­ ;todopt sp. . . ation in the skull is not known, nor is there a suitable answer to the tle, fish, reptile, and bird remams question of individual variation in premolar pattern. In certain spe­ cies the individual variation of teeth representing approximately the same stage of wear is not great. However, if the conclusion implied by this statement is applied widely, an extremely large number of been intensive enough to de­ species in proportion to the number of known specimens would have assemblage. However, a mid­ to be recorded. Moreover, some of the mylagaulid material in the iomparable to that represented Institute collections points to considerable individual variation in faunas seems certain. More specimens from the same zone and locality. suggests by the presence of By means of relatively unworn teeth it is possible to demonstrate , as advanced as the Thousand in the upper premolar the derivation of the lakes from the original basins in the unworn teeth and, moreover, the origin of certain denti­ re of a mole, beavers, an otter, nal areas from particular cusps, exactly as this has been done in the lst environment with bodies of cheek-teeth of such forms as Equus. The lower premolar appears aited collecting area, the lithol­ susceptible to a similar treatment, although greater difficulty is en­ mber of specimens of Dipoides countered in tracing the derivation of the adult pattern than in the upper premolar. Use of this method of studv leads to a clearer in origin. appreciation of the differences or similarities in" two distinct species than can be obtained by citing the number of lakes present and their ION OF FAUNA arrangement in varying numbers of rows. It is felt that this method l of study combined with sectioning of individual teeth might eventu­ lie ally lead to an understanding of the characters exhibited by isolated blems which need to be solved grinding teeth. Such work would have to be based primarily on a the group is reached. Not only thorough study of species which are known by a large number of speci­ confusion, but even the genera mens and in which teeth in all stages of wear are present. n. The relation of horned to Doubtless a greater number of species and perhaps genera are in ~d, nor have the limits of indi- existence than have been described. It is conceivable that such emi­ nently burrowing types might have a rather limited geographic range nee or absence of horns is a sex and that consequently a number of distinct species might be present ; pointed out, no other known in each faunal stage. However, with our present understanding of the istinction, although, as he states group, the establishment of more types is decidedly not a desirable ~ted by the fact that no other procedure. ation, in a burrowing form in Pliocene mylagaulids are known only by fragmentary material. A d as digging implements, would possible exception is Epigaulus hatcheri, the type of which consists 6 CONTRIBUTIONS TO PAL..£0NTOLOGY NEW PLIOCENE RODENT FA1

of a fairly complete skull and skeleton. Gidley referred the beds numerous, elongate, and tend to (Republican River) from which the type was obtained to the upper possess an external investment of 1 is a functional part of the tooth. Miocene. The type locality is near Long Island, Kansas. Stirton transversely below the premolar has recently referred a fauna, termed by him the Long Island fauna, lakes. The single specimen exhit to the middle Pliocene. Mylagaulus monodon and M. sesquipedalis is relatively unworn and indicate also have been recorded from Pliocene beds. The type localities of takes place with further wear. 'I these two species are in the Republican River beds, and hence the eight lakes in a well-worn tootl: exact age of the specimens is not known, although they are usually figure might be increased to ten. gaulus lrevis but smaller than in referred to the lower Pliocene. A referred specimen of M. monodon gaulid humerus is distinctly sr apparently comes from the same beds as E. hatcheri. Matthew has E. hatcheri. suggested that M. monodon may have to be referred to Epigaulus and The Rome species does not agr states that M. monodon and E. hatcheri are probably identical.2 may be closest to M. monodon. However, Matthew in describing the type of M. monodon states that the Rome species possesses a lar, tinctly bulged by the premolar. P4 has no cement outside the external enamel ring. The type of a larger number of lakes. The I E. hatcheri, as described by Gidley, possesses premolars in which an ment. A P4 referred by Cope to investment of cement forms a functional part of the teeth. This dis­ the Oregon type but the occlusa crepancy can be removed only by assuming: (1) that cement was the type and referred specimen ol present originally on the type of M. monodon and has dropped away; Rome specimens are rather close tional part of the wearing surface (2) that the type was incorrectly described (not probable); or differ from M. monodon as descri (3) that the presence or absence of cement is not a distinctive char­ Epigaulus hatcheri Gidley,1 alt] acter. Perhaps the explanation given under (1) is the most likely. from Rome in the presence of an 1 M. sesquipedalis is distinguished from M. monodon by its smaller size, The ramus of the Rome species l fewer, less elongate lakes, and more irregular arrangement of lakes. portiops, but is thick and deep l The tooth-row is slightly shorter Mylagaulus? c£. monodon Cope what in pattern in the two speciee Pliocene mylagaulids from the Great Basin have been known hitherto specific, the type from Rome wm . only by two isolated grinders. A P.i has been recorded by Miss Kellogg 3 upper premolars of the Oregon ty from Thousand Creek, Nevada, and E. Raymond Hall has described a species of mylagaulid from Bartl lower premolar from the Fish Lake Valley beds of Nevada.4 In comparison, well-preserved skull of the latter mylagaulid material from Rome is relatively abundant. The most perfectly which it differs distinctly from E1 preserved specimen is an almost complete lower jaw, C.I.T. No. 72 (Plate 1, M ylagaulus species from Fish figs. S, Sa, Sb). The extreme tip of the coronoid is missing, as is the posterior single isolated P4. This species : portion of the angle, but the dentition is complete including all three molars. what different occlusal pattern. U A second ramus, No. 1951, which was also obtained is more fragmentary given by Hall do not agree very than No. 72. The dentition in this specimen lacks MI. Isolated premolars, the species is probably distinct f1 Nos. 1952-195S (Plate 1, figs. 1-7), both upper and lower, and various The only other Pliocene myla fragmentary limb elements are present also in the Rome collection. The Basin is an unworn P4 from the two rami represent young individuals in which practically none of the Kellogg. This specime~ was refe1 enamel inflections has become isolated, thus making comparisons with other Kellogg compared this tooth direc1 specimens difficult. All the remains have been considered as representing a thew in the belief that the latter single species, although this is a doubtful procedure with reference to one Thousand Creek form is probabl~ or two specimens. However, in the case of the V The Rome species is characterized by premolar teeth which become Kellogg to M. monodon, the figt elongate in a fore-and-aft direction with wear. The enamel lakes are M. cf. lrevis from Skull Spring t 'R. A. Stirton, Univ. Calif. Pub., 'Bull. Dept. Geol. Sci., vol. 23, no. 8, 284, table 3, 1934. referred to M. pristinus, may also •w. D. Matthew, Bull. Amer. Mus. Nat. Hist., vol. 50, art. 2, 75, 1924. Valley mylagaulids in the Califo 'L. Kellogg, Univ. Calif. Pub., Bull. Dept. Geol., vol. 5, no. 29, 429, fig. 10, 1910. ence of a species distinct from th • E. R. Hall, Univ. Calif. Pub., Bull. Dept. Geol. Sci., vol. 19, no. 12, 307-308, figs. 18-19, 1930. 'J. W. Gidley, Proc. U. S. Nat. Mus.

• ONTOLOGY NEW PLIOCENE RODENT FAUNAS FRO!\I OREGON AND CALIFORNIA 7

Gidley referred the beds numerous, elongate, and tend to arrange themselves in rows. The teeth , was obtained to the upper possess an external investment of cement which in some specimens at least 1 is a functional part of the tooth. The ramus is very heavy and deep, thick g Island, Kansas. Stirton transversely below the premolar. Pi possesses from six to usually nine him the Long Island fauna, lakes. The single specimen exhibiting six lakes, No. 1954 (Plate 1, fig. 2), nodon and M. sesquipedalis is relatively unworn and indicates that an increase in the number of lakes eds. The type localities of takes place with further wear. The lower premolars possess a minimum of River beds, and hence the eight lakes in a well-worn tooth and a maximum of nine, although this figure might be increased to ten. The teeth are larger than those of Myla­ , although they are usually gaulus lcevis but smaller than in the type of Epigaulus hatcheri. A myla­ :d specimen of M. monodon gaulid humerus is distinctly smaller than the comparable element in E. hatcheri. Matthew has E. hatcheri. Je referred to Epigaulus and The Rome species does not agree exactly with any described species, but ~ri are probably identical.2 may be closest to M. monodon. Compared with the type of M. monodon, the Rome species possesses a larger ramus, which is deeper and more dis­ i of M. monodon states that tinctly bulged by the premolar. P4 is narrower transversely and possesses enamel ring. The type of a larger number of lakes. The latter show a somewhat different arrange­ esses premolars in which an ment. A P4 referred by Cope to M. monodon agrees in size and shape with part of the teeth. This dis­ the Oregon type but the occlusal pattern appears to be distinct. If both ning: (1) that cement was the type and referred specimen of M. monodon are actually conspecific, the 1don and has dropped away; Rome specimens are rather close to them. However, cement forms a func­ tional part of the wearing surface of the premolar, a character in which they ;cribed (not probable); or differ from M. monodon as described by Matthew. -nt is not a distinctive char­ Epigaulus hatcheri Gidley,1 although with larger P4/4, resembles the type o.der ( 1) is the most likely. from Rome in the presence of an external coat of cement on the cheek-teeth. monodon by its smaller size, The. ram us of. the ~ome species apparently possesses slightly different pro­ gular arrangement of lakes. portions, but is thick and deep below P4, a character seen in E. hatcheri. The tooth-row is slightly shorter in the Oregon form, and P4/4 differ some­ •n Cope what in pattern in the two species. If E. hatcheri and M. monodon are con­ ;in have been known hitherto specific, the type from Rome would be close to that species. However, the ,en recorded by Miss Kellogg 3 upp~r premolars of the Oregon type show a close agreement with those of a Lymond Hall has described a species of mylagaulid from Bartlett Mountain, Oregon (Pliocene). A very ds of Nevada.4 In comparison, well-preserved skull of the latter is without bony horn-cores, a character in abundant. The most perfectly which it differs distinctly from Epigaulus. er jaw, C.I.T. No. 72 (Plate 1, Mylagaulus species from Fish Lake Valley, Nevada, is represented by a id is missing, as is the posterior single i.solated P4. This species is smaller than the Rome type with some­ >lete including all three molars. "'.hat different occlusal pattern. Unfortunately, the dimensions of the figures obtained is more fragmentary given by Hall do not agree very well with his measurements. In any case lacks MI. Isolated premolars, the species is probably distinct from that found at Rome. lipper and lower, and various Tpe .only other Pliocene mylagaulid material on record from the Great in the Rome collection. The Basm is an unworn Pi from the Thousand Creek beds, described by Miss vhich practically none of the Kellogg. This specimen was referred by her to M. monodon. Although Miss naking comparisons with other Kellogg compared this tooth directly with figures of the type of M. lrevis Mat­ m considered as representing a thew in the belief that the latter was still referred to M. monodon Cope, the rocedure with reference to one Thousand Creek form is probably closer to Cape's species than to M. lrevis. However, in the case of the Virgin Valley specimen, referred by Miss premol!tr teeth which become Kellogg to M. monodon, the figure shows a tooth more closely resembling wear. The enamel lakes are M. cf. lrevis from Skull Spring than M. monodon. U.C. No. 12580, a P4 Sci., vol. 23, no. 8, 284, table 3, 1934. referred to M. pristinus, may also represent the Skull Spring species. Virgin 50, art. 2, 75, 1924. Valley mylagaulids in the California Institute collection indicate the pres­ •l. 5, no. 29, 429, fig. 10, 1910. ence of a species distinct from the Skull Spring type. Whether this second i., vol. 19, no. 12, 307-308, figs. 18-19, 'J. W. Gidley, Proc. U.S. Nat. Mus., vol. 32, no. 1554, 627-636, pls. 5!Hl5, 1907. 8 CONTRIBUTIONS TO PAL..EONTOLOGY NEW PLIOCENE RODENT FAU species is referable to M. ']>'l'istinus was not determined. An isolated P4 present author has not observed in the Institute collection from Thousand Creek appears to be distinct from specimens of Castor. the Rome species. There may be some doubt as to The fragmentary limb material from Rome is too incomplete to add any­ the rest of the rodent fauna. Th thing to our knowledge of the skeletal structure of the Mylagaulidre. isolated from that in which the otho It is quite possible that Mylagaulus? cf. monodon from Rome represents of beds of possible upper PlioceDE an undescribed species of rodent. In view of our very incomplete knowledge this question. However, not only~ of the Mylagaulidre nothing would be gained by establishing a new species. with Pliohippus teeth obviate th characters exhibited by the specir Measurements (in millimeters) of Mylagaulus7 cf. monodon Cope True beavers from North Amer No. 72 No. 1951 the upper Pliocene. Two teeth fr Pi-M3, alveolar length ...... 19.l 17.8 and the present one are all that l Length of diastema, I-P4...... 10.6 range of Castor is not accurately < Depth of ramus beneath PI.. . • ...... 18.9 types are known from the early I Depth of ramus beneath diastema...... 12.6 Thickness (transverse) of ramus beneath P4...... 11.8 M easurementa * (in Maximum Minimum ~?, antero-posterior dian Several isolated P!'s: ~?, transverse diameter. Greatest antero-posterior diameter...... 12.0 11.0 *Measured at accusal sur Greatest transverse diameter...... 6.4 5.6 Several isolated PI's: Greatest antero-posterior diameter...... 13.3 10.4 M Greatest transverse diameter...... 6.0 5.2 Goniodontomys Geological Age and Locality­ Castoridz Crooked Creek drainage, tributar: Dipoides srirroni Wilson west of Rome, Malheur County, Genotype-No. 1959, C.I.T. Cc Remains of this species of aberrant beaver are very common in the Rome bearing MI-M2 (Plate 2, figs. 2, 2 fauna. A detailed account of the fossil material has already been published.1 Referred Specimen-No. 1960, 1 of ramus with left M2 in place. Castor? sp. The true beaver is represented by a single isolated M2?. This specimen, GENERIC AND No. 1961 (Plate 1, figs. 9, 9a), represents an individual comparable in size Mandibular incisor passing fror to the existing Castor canadensis. The base of the tooth is somewhat dam­ probably extending well up into t aged, but practically the full height of the cheek-tooth seems to be preserved. dont, prismatic, and flat-crowned, Three internal and one external lateral grooves are present, as is normal for angles of cheek-teeth generally OJ Castor. The external groove extends to the base of the portion preserved, sides normally in contact. Enan while the three internal inflections extend somewhat less than half the dis­ tracts. MI with posterior loop, t tance to the base of the tooth. In a measure, the extent of the internal Second external salient angle of : grooves is a function of the amount of wear the tooth has undergone, but Ramus heavy and apparently shm undoubtedly the grooves are much less persistent than in Recent species. as in Mimomys primus. The posterior groove (metastriid) is slightly longer than the other two (mesostriid and parastriid), which are equal in length. The occlusal pattern is that of a normal Castor except for the presence of an enamel lake in the D i anterior enamel loop. Such a lake is present in the first molar of a ramus of Inferior Dentition-The mandil the Asiatic C. andersonni, as figured by Schlosser.2 However, in this species side of the tooth-row under ( ?) 1\ MI and M2 are shorter transversely than is the Rome molar, although M2 ascending ramus. The first mola has the same antero-posterior diameter, and MI only a slightly greater 1 The system used by M. A. C. Hint• measurement. C. andersonni, together with the Asiatic species zdanskyi angles has been adopted in this paper. I and broilii, has been placed by Young in a new genus Sinocastor. The (Microtina:) Living and Eilltinct, vol. 1, the teeth it is customary to enumerate 1 1 R. W. Wilson, Carnegie Inst. Wash. Pub. Ko. 453, pt. 3, 1934. backwards in upper molars, and from 1 • M. Schlosser, I'alre. Sin., ser. C, vol. 1, fasc. 1, pl. 2, fig. 43, 1924. angles on each side being formed by ti 1NTOLOGY NEW PLIOCENE RODENT FAUNAS FRO:\'I OREGON AND CALIFORNIA 9 determined. An isolated P4 present author has not observed the above-mentioned lake in any other ik appears to be distinct from specimens of Castor. There may be some doubt as to the association of the Rome Castor? with is too incomplete to add any­ the rest of the rodent fauna. The occurrence of the specimen in an area e of the Mylagaulidre. isolated from that in which the other rodents were obtained, and the presence onodon from Rome represents of beds of possible upper Pliocene or Pleistocene age in the vicinity, raise •Ur very incomplete knowledge this question. However, not only does the association in the field of Castorf by establishing a new species. with Pliohi'JYP'US teeth obviate this possibility to a large extent, but the characters exhibited by the specimen itself point in the same direction. ilus? cf. monodon Cope True beavers from North America are extremely rare in beds older than the upper Pliocene. Two teeth from the upper Snake Creek (Pi and P4) No. 72 No. 1951 19.l 17.8 and the present one are all that have been so far recorded. The geologic 10.6 range of Castor is not accurately determined, but the genus or closely allied 18.9 types are known from the early Pliocene of Europe and Asia. 12.6 11.8 Measurements* (in millimeterB) of CaBtor1 ap.

Minimum l\ft?, antero-posterior diameter...... • ...... 8.2 Maximum l\ft?, transverse diameter...... • ...... 7.1 12.0 11.0 •Measured at occusal surface. 6.4 5.6 Microtime? 13.3 10.4 6.0 5.2 Goniodontomys disjunctus n. gen. and sp. Geological Age and Locality-Middle Pliocene beds exposed along the Crooked Creek drainage, tributary to the Owyhee River, five miles south­ lson west of Rome, Malheur County, Oregon. Genotype-No. 1959, C.I.T. Coll. Vert. Pale., an incomplete left ramus are very common in the Rome 1 bearing MI-M2 (Plate 2, figs. 2, 2a, 2b). al has already been published. Referred Specime11r-No. 1960, C.I.T. Coll. Vert. Pale., a small fragment of ramus with left M2 in place.

isolated M2?. This specimen, GENERIC AND SPECIFIC CHARACTERS individual comparable in size Mandibular incisor passing from lingual to buccal side of tooth-row, and of the tooth is somewhat dam­ probably extending well up into the ascending ramus. Cheek-teeth hypso­ ek-tooth seems to be preserved. dont, prismatic, and flat-crowned, rooted and without cement. Re-entrant es are present, as is normal for angles of cheek-teeth generally opposed. Re-entrant angles from opposite base of the portion preserved, sides normally in contact. Enamel not differentiated into thick and thin mewhat less than half the dis- tracts. MI with posterior loop, three triangles, and complex anterior loop. ure, the extent of the internal Second external salient angle of MI opposite third internal salient angle. 1 • the tooth has undergone, but Ramus heavy and apparently shortened. Length of tooth-row approximately ~istent than in Recent species. as in Mimomys primus. ly longer than the other two in length. The occlusal pattern DESCRIPTION ~sence of an enamel lake in the in the first molar of a ramus of Inferior Dentition-The mandibular incisor crosses from lingual to buccal )sser.2 However, in this species side of the tooth-row under ( ?) M&, and probably extends well up into the the Rome molar, although M2 ascending ramus. The first molar (Plate 2, fig. 2) is composed of a pos- id MI only a slightly greater •The system used by M . .A. C. Hinton for enumeration of the re-entrant and salient n the Asiatic species zdanskyi angles has been adopted in this paper. See Hinton, Mo·oograph of the Vole1 and Lemminga Sinocastor. (Microtinre) Living and E111tinct, vol. 1, 22, 1926. To quote from Hinton: "In describing a new genus The the teeth it is customary to enumerate the salient angles and re-entrant folds from before i3, pt. 3, 1934. backwards in upper ,molars, and from behind forwards in lower molars, the first salient 2, fig. 43, 1924. angles on each side being formed by the transverse loop." 10 CONTRIBUTIONS TO PAL1EONTOLOGY NEW PLIOCENE RODENT FAl terior loop, three triangles, the anterior two opposed to form a somewhat Relationship to the microtines i tetragonal loop of enamel, and a complex anterior loop. The anterior loop enamel of the cheek-teeth, the ti is complicated by a pronounced inflection on the internal side, a similar relatively large number of elemen but less pronounced inflection buccally, and an antero-median inflection in in the ramus. 1 f the intermediatt front. A very shallow fold of enamel is also present on the buccal side and triangles of MI of Goniodontor. just anterior to the main external fold mentioned above. The posterior pattern would result quite close t< triangle communicates broadly with the posterior loop, but the commisures If Goniodontomys is to be assi1 connecting the median pair of triangles with the anterior and posterior por­ and not a lemming. Enough of 1 tions of the tooth are narrow. M2 is a less complex tooth which may be type specimen to demonstrate thi characterized as possessing three loops of enamel connected by narrow com­ Reference of Goniodontomys t misures. The median loop is apparently analogous to the two posterior microtin~s seems less likely tha1 alternating triangles in normal voles. A second, less worn specimen of rodent families possess members v Goniodontomys, No. 1960, possesses slightly more triangular salient angles, quite removed structurally as weI and in addition some slight angulation of the anterior loop suggestive of N eotoma and its relatives sugge the two more or less alternating triangles which replace this loop in the certain features of the dentition. normal vole dentition. M2 is distinctly shorter than MI. relatively thick enamel borders, Ramus-The ramus (Plate 2, figs. 2a, 2b) is rather deep and is apparently ramus is slimmer, and the ascend: shortened. The juncture of ascending and alveolar portions of the ramus farther back than in Goniodont01 occurs opposite the posterior root of MI as in voles, rather than opposite tomys possess a more complicat the posterior root of M2 as in N eotoma. The ridge for attachment of the Neotoma-like form from the midd masseter muscle terminates anteriorly about opposite the posterior surface the case in the former genus. Th of the anterior root of MI. The masseter medialis scar is slightly less de­ ingly like that of Goniodontomys veloped than the preserved portion of the masseter lateralis scar. The the otlier hand, in our genus the E mental foramen is situated close to the superior surface of the ramus, and internal salient of MI form a loo almost directly beneath the anterior tip of MI. A second small foramen responding loop in N eotoma is r is found just above the tip of the ridge for the attachment of the masseter slightly forward and outward. muscle. The area for symphyseal attachment extends back well under the Ameghino from the Pampean of posterior portion of MI. In this character our genus resembles Microtus more closely in pattern of MI th and differs from N eotoma. In the latter genus the symphyseal area is much different and the resemblance is r more limited posteriorly. The cape-rats possess extreme] quite similar to that of microtin RELATIONSHIPS recorded from North America a Although the systematic position of Goniodontomys is not entirely clear, Goniodontomys less than does tha it has been assigned to the Microtinre. The genus is known by such frag­ forms are known from Asia, and 1 mentary material that it is difficult to eliminate from consideration some pattern than are existing types. ' other groups of rodents with hypsodont teeth. Goniodontomys is widely a pronounced tendency to isolate 1 separated from most microtines by the more or less complete opposition of this respect Goniodontomys is clc the usually alternating prisms of the cheek-teeth. The second molar has an condition is limited to extreme WE occlusal pattern somewhat like that exhibited by some species of Hyperacrius Other groups of rodents with h and Eothenomys. The first molar, however, is quite unlike the usual micro­ blance to Goniodontomys than d tine MI, especially in the nearly opposite second external and third internal Sigmodo-n, the jerboas, and the salient angles. In voles of more normal type, when the triangles are op­ characters. posite one another, the second external triangle or salient angle is opposed If Goniodontomys is a vole, it i: by the second internal triangle. An approach to the pattern of MI in the it is not very close to existing for: Oregon genus is made in a specimen of Prometheomys schaposchnikowi,1 type. It is surprising that the figured by Hinton, but the resemblance is remote. Moreover, the anterior Tertiary, as the subfamily must termination of M2 in Goniodontomys is not angular, as is usually the case Poamys Matthew from the lower in Microtinre. No. 1960, the dentition of which is somewhat less worn than possibly a structural ancestor of the type, suggests this angular termination, and some Recent voles possess pothesis cannot be determined on second molars with a more or less rounded appearance. of Goniodontomys and Poamys iJ 1 M. A. C. Hinton, Monograph of the Voles and Lemmings, vol. I. Brit. Mus. Nat. Hist., 1 F. Ameghino, Contr. al Conocimient 86, fig. 55, 1926. Acad. Nae. de Ciencias Republ. Argent.

, .ONTOLOGY NEW PLIOCENE RODENT FAUNAS FRO'.\I OREGON AND CALIFORNIA 11

opposed to form a somewhat Relationship to the microtines is suggested in Goniodontomys by the thin terior loop. The anterior loop enamel of the cheek-teeth, the triangular shape of the salient angles, the m the internal side, a similar relatively la.i,:ge number of elements comprising MI, and various characters an antero-median inflection in in the ramus. If the intermediate-external triangle of M2 and the internal present on the buccal side a?d triangles of MI of Goniodontomys were shifted anteriorly , an occlusal ntioned above. The posterior pattern would result quite close to that in many microtines. erior loop, but the commisures If Goniodontomys is to be assigned to the Microtinre, the genus is a vole the anterior and posterior por- and not a lemming. Enough of the mandibular incisor is preserved in the 1 complex tooth which may be type specimen to demonstrate this conclusion. mel connected by narrow com­ Reference of Goniodontomys to some group of rodents other than the nalogous to the two posterior microtin~s seems less likely than the present assignment. A number of :econd, less worn specimen of rodent families possess members with hypsodont teeth, but most of them are more triangular salient angles, quite removed structurally as well as geographically from the Oregon genus. .he anterior loop suggestive of N eotoma and its relatives suggest a relationship with Goniodontomys in 1Vhich replace this loop in the certain features of the dentition. However, the dentition of Neoto-ma has orter than MI. relatively thick enamel borders, the salient angles are less triangular, the s rather deep and is apparently ramus is slimmer, and the ascending ramus rises from the horizontal ramus alveolar portions of the ramus farther back than in Goniodontomys. Moreover, not only does Goniodon­ in voles, rather than opposite tomys possess a more complicated MI, but it is hardly likely that any he ridge for attachment of the Neotoma-like form from the middle Pliocene would be so high-crowned as is , opposite the posterior surf ace the case in the former genus. The second lower molar of N eotoma is strik­ iedialis scar is slightly less de- ingly like that of Goniodontomys in the general aspect of the pattern. On masseter lateralis scar. The the otlier hand, in our genus the second external salient angle and the third irior surface of the ramus, and internal salient of MI form a loop directed forward and inward. The cor­ MI. A second small foramen responding loop in Neotoma is more transverse or, if oblique, is directed ;he attachment of the masseter slightly forward and outward. A fossil species of Oxymicterus, impexus nt extends back well under the Ameghino from the Pampean of South America, resembles Goniodontomys our genus resembles Microtus more closely in pattern of MI than does N eotoma. However, M2 is quite us the symphyseal area is much different and the resemblance is probably superficial.1 The cape-rats possess extremely hypsodont teeth with occlusal patterns quite similar to that of microtines. However, the group has never been s recorded from North America and its dental pattern resembles that of 1dontomys is not entirely clear, Goniodontomys less than does that of some voles. Moreover, early Pliocene , genus is known by such frag­ forms are known from Asia, and these are no closer to our genus in occlusal iinate from consideration some pattern than are existing types. The early Pliocene Asiatic specimens show ~th. Goniodontomys is widely a pronounced tendency to isolate the re-entrant folds of enamel as lakes. In ~ or less complete opposition of this respect Goniodontomys is closer to normal microtines, in which such a eeth. The second molar has an condition is limited to extreme wear. by some species of Hyperacrius Other groups of rodents with hypsodont teeth have decidedly less resem­ is quite unlike the usual micro­ blance to Goniodontomys than do those mentioned above. Such forms as ~ond external and third internal Sigmodon, the jerboas, and the Gerbillinre differ in one or more major ·pe, when the triangles are op- characters. 1gle or salient angle is opposed If Goniodontomys is a vole, it is the oldest so far recorded. Unfortunately, ch to the pattern of MI in the it is not very close to existing forms, and apparently represents an aberrant ·ometheomys schaposchnikowi, 1 type. It is surprising that the Microtinre are not more common in the ·emote. Moreover, the anterior Tertiary, as the subfamily must have had a considerable geologic range. . angular, as is usually the case Poamys Matthew from the lower Snake Creek was regarded by Matthew as iich is somewhat less worn than possibly a structural ancestor of the Microtinre. The validity of this hy­ and some Recent voles possess pothesis cannot be determined on available evidence. However, the presence .ppearance. of Goniodontomys and Poamys in beds older than the upper Pliocene sug- mming1, vol. 1. Brit. Mus. Nat. Hist., • F. Ameghino, Contr. al Conocimiento de los Mam. Foe. de la Republ. Argent., Actos Acad. Nae. de Ciencias Republ. Argent. en Cordoba, Atlas, pl. 4, fig . 3b, 1889. NEW PLIOCENE RODENT FAl 12 CONTRIBUTIONS TO PALlEONTOLOGY gests the possible presence of other microtine-like forms in the late Miocene A large series of topotype ma· and early Pliocene. parison. It should be noted tha Since Goniodontomys is an aberrant type, it does not furnish much evi­ the characters of P3 as well as sor dence as to the evolution of the Microtinre. The opposition of the triangles that have been used to establish ·1 in this genus does not necessarily demonstrate that this is the primitive is too scanty to permit one to s microtine condition, although the perfect alternation of prisms seen in many from those from Thousand Creek voles is probably a specialized and advanced character. this to a certain extent. Howev« Creek are closely comparable to Me08Urements (in millimeters) of Goniodontomys disjunctWJ n. gen. and sp. specific designation is justified. Cl.T. No. 1959 (genotype). Middle Pliocene, Rome, Oregon Measurements (in millim1 Length of crown, MT-M2...... 5.5 MI, antero-posterior diameter...... 3.0 P"3"-MI, occlusal length ...... MT, transverse diameter...... 1.5+ Depth of ramus below MI, meast M2, antero·posterior diameter...... 2.2 M2, transverse diameter...... 1.5 H Depth of ramus beneath MI...... 6.3 A left ramus of Hypolagus wi· LAGOMORPHA appears to represent a species c characterized by a much lighte Leporide although the ramus is as deep aw Hypolagus vetus (Kellogg) form teeth have convex postero-i Several fragmentary lower jaws and a number of isolated teeth, both P4 and MI. This surface is flat uppers and lowers, appear to be referable to Hypolagus vetus (Kellogg) .1 Rome may be intermediate in t The Rome specimens are of approximately the same size as topotype representing individuals smaller material of H. vetus from Thousand Creek. Characters exhibited by P3 have also been referred tentative agree with those of the latter material except that this tooth in the Rome which is smaller than the compa1 form may be slightly broader on the average, with flatter or more gently ence in size, the tooth is close to rounded anterior face, which serves to widen this portion of the tooth. Some Comparison of the Rome spE topotype specimens exhibit these characters, but the average specimen is handicapped by lack of suitable probably somewhat different. The lower molariform teeth, as seen in No. Rome form would seem to distin; 1962 (Plate 2, figs. 3, 3a), fragmentary left ramus with P3'-MI, and No. dentition is similar in size. 1963, fragmentary left ramus with P3'-MI, C.I.T. Coll. Vert. Pale. from It should be stated that suffici1 Rome, may be slightly larger than comparable material from Thousand possibility that this second specii Creek. or immature stage of H. vetus. A single isolated P2 is present in our collections. If this tooth is referable mostly on a single ramus. A sec to the same species as the rami, it differs from the Thousand Creek H. vetus tition, but this is still distinctly in a lengthening of the tooth transversely. However, the specimen shows Until more material becomes av: a deep antero-internal fold and a shallower antero-external fold as in H. types from the Rome locality. vetus. The upper molariform teeth do not seem to show such heavy external Measurements (in ribs as in the Nevadan species. Pl- 11"3", occlusal length ...... C. L. Gazin has referred a large species of Hypolagus from Hagerman to Depth of ramus below MI, meas H. near vetus.2 In shape of P3 and in size of some specimens at least, the Rome material is close to that from Hagerman. If the Rome fauna is KERN actually advanced over that from Thousand Creek, as is perhaps suggested by the presence of Castor?, the Thousand Creek, Rome, and Hagerman The Kern River deposits ar specimens may form a progressive series. However, it is doubtful if the Bakersfield, in Kern County, Rome fauna as a whole is more advanced than that from Thousand Creek. and lagomorph fauna obtainec Compared with other known species of Hypolagus, the Rome specimens ing forms: are readily distinguished by their large size as well as by characters in the dentition. Rodentia Citellus'l sp. 1 L. Kellogg, Univ. Calif. Pub., Bull. Dept. Geol., vol. 5, no. 29, 436-437, fig. 20 , 1910; Peromyscus pliocenicus n. L. R. Dice, Univ. Calif. Pub., Bull. Dept. Geol., YOl. 10, no. 12, 181- 182, figs. 4-5, 1917. 1 C. L. Gazin, Proc. U. S. Nat. Mus., vol. 83, no. 2976, 112-114, fig . l, 1934. Lagomorpha Hypolagus near limnetus Hypolagus small sp. NEW PLIOCENE RODENT FAUNAS FROM OREGON AND CALIFORNIA 13 NTOLOGY ike forms in the late Miocene A large series of topotype material of H. vetus was available for com­ parison. It should be noted that this material shows decided variation in t does not furnish much evi­ the characters of P3 as well as some variation in most of the other characters he opposition of the triangles that have been used to establish species of Hypolagus. The Rome material .te that this is the primitive is too scanty to permit one to say that the specimens average differently iation of prisms seen in many fr?m those fro~ Thousand Creek, al~hough study of the collection suggests this to a certam extent. However, smce certain specimens from Thousand character. Creek are closely comparable to individuals from the Oregon locality the .ys disjunctus n. gen. and sp . specific designation is justified. ' .ocene, Rome, Oregon Measurements (in millimeters) of Hypolagus vetus (Kellogg) 5.5 No.1962 No.1963 3.0 P!J-MI, occlusal length...... 9.1 9.3 1.5+ Depth of ramus below MI, measured on inside ...... 12.7 2.2 1.5 Hypolagus sp. 6.3 A left ramus of Hypolagu~ wit? ~4-M3', No. 1964 (Plate 2, figs. 1, la), appears ~o represent a spec.1es d1stmct. ~rom H. vetus. This specimen is characterized by a. much hghter dent1t1on than is present in H. vetus, although the ramus is as deep and apparently as robust. The lower molari­ 1>gg) form teeth have convex postero-internal ribs, most noticeably developed in 1mber of isolated teeth, both P4 and MI. This surface is flattened in typical H. vetus. H. vetus from Hypolagus vetus (Kellogg) .1 Rome m~y b~ i~t~rmediate in this respect. Several additional specimens the same size as topotype representmg md1v1duals smaller than those of H. vetus in the collection Characters exhibited by P3 ha~e a~so been referred tentatively to this type. One isolated P3 is present t, that this tooth in the Rome which is smaller than the comparable tooth of H. vetus. Aside from differ­ a, with flatter or more gently ence in size, the tooth is close to the vetus type. ;his portion of the tooth. Some Co!Ilparison of the Rm:~e species with small species of Hypolagus is but the average specimen is handicapped by lack of smtable material. However, the robust jaw of the lariform teeth, as seen in No. Rome form would seem to distinguish this species from others in which the ramus with P3-MI, and No. dentition is similar in size. C.I.T. Coll. Vert. Pale. from It should be stated that sufficient material i& not available to obviate the :i.ble material from Thousand possibility that this second species of Hypolagus represents merely a young or immature stage of H. vetus. Comparisons, as stated above 1 are based ions. If this tooth is referable ~~stly on a ~in~le r!"mu.s .. A second jaw apparently possesses a larger den­ the Thousand Creek H. vetus t1t10n, but this is still d1stmctly smaller than that in H. vetus from Rome. However, the specimen shows Until more material becomes available it seems desirable to recognize two antero-external fold as in H. types from the Rome locality. m to show such heavy external Measurements (in millimeters) of Hypolagus sp. No.1964 P4- }I!J, occlusal length...... 8.4 H ypolagus from Hagerman to Depth of ramus below MI, measured on inside ...... 12.9 >f some specimens at least, the ·man. If the Rome fauna is KERN RIVER FAUNA Creek, as is perhaps suggested Creek, Rome, and Hagerman The Kern River deposits are located about nine miles northeast of [fowever, it is doubtful if the Bakersfield, in Kern County, California. A middle Pliocene rodent :m that from Thousand Creek. and lagomorph fauna obtained from these beds consists of the follow­ 1.Jpolagus, the Rome specimens ing forms: :i.s well as by characters in the Rodentia Citellus? sp . .-ol. 5, no. 29, 436-437, fig. 20, 1910; Peromyscus pliocenicus n. sp . . 10, no. 12, 181-182, figs. 4-5, 1917. Lagomorpha 6, 112-114, fig. l, 1934. Hypolagus near limnetus Gazin H ypolagus small sp. 14 CO~TRIBUTIONS TO PAL.'EONTOLOGY NEW PLIOCENE RODENT FAU:!'

SYSTEMATIC DESCRIPTION OF FAUNA Ci RODENTIA Peromyscus Sciuridre Type-Fragmentary right ramus Pale. (Plate 3, figs. 2, 2a). Citellus? sp. Paratypes-Fragmentary left ra Sciurid remains from the Kern River beds are limited to a right ramus No. 1967 (Plate 3, fig. 4) ; fragmen with P4-M2 and an alveolus for MB", No. 1965 (Plate 3, figs. 1, la) ; a plete M2, No. 1968 (Plate 3, fig. 3) second right ramus without dentition; and fragments of an upper and a lower Geological Age and Locality-] incisor which may be referable also to this family. Specimen No. 1965 is County, California. C.I.T. Localit the basis for the doubtful generic determination given above. The dentition: of No. 1965 is extremely worn and part of the first molar SPECIFIC is missing. As a consequence, comparisons are difficult to make. No. 1965 is relatively small, agreeing in size with Recent specimens of Callospermoph­ Cheek-teeth hypsodont, but cro· ilus lateralis certus. The molars do not show the pronounced fore-and-aft faces; without accessory folds al compression and the high trigonids characteristic of typical Citellus, and present or absent. MI with divi( the specimen certainly does not represent the genus in the restricted sense. internal re-entrant angle becomei P4 is rather triangular in outline and the anterior two cusps were probably shallow re-entrant. M3 relatively closely appressed in their originally unworn condition. The mandibular bone lateral to anterior root of 1 incisor is relatively broad and lacks pronounced furrows. Thus, No. 1965 known fossil species, but approximi appears to represent the genus Citellus in a broad way, and to be related to such genera or subgenera as Callospermophilus and Otospermophilus, which Dn occupy a position intermediate between typical CitellttS and typical Sciurus. Peromyscus pliocenicus is distin No. 1965 seems to be more closely related to Otospermophilus gidleyi the genus, except P. antiquus and (Rattlesnake), Citellus sp. (Thousand Creek), and Citellus? sp. (Smiths these two species are somewhat sr Valley) than to any other Pliocene sciurids. It is about the size of these Kern River type. species and moreover agrees in general character of the dentition, at least Compared with P. antiquus Kell in so far as may be observed in the well-worn dentition of the Kern River northern Nevada, P. pliocenicus is specimen. the following selected differences: Otospermophilus gidleyi Merriam, Stock, and Moody 1 from the Rattle­ resulting in 'an external cusp rathe snake formation of eastern Oregon approximates the Kern River species Nevadan species; (2) MI relative in length of tooth-row, but the ramus of the former is deeper. The talonid M2 may be slightly less developt rjm in M2 of No. 1965 may be more angulate than in the Oregon species. variably developed, whereas they a Degree of angulation in No. 1965 is difficult to determine because of break­ tively large intermediate tubercles ing away of the inner margins of the molars in the specimen. Citellus? sp.2 (Plate 3, figs. 2, 2a), but are abi from Smiths Valley, central Nevada, also approaches No. 1965 in size, al­ mentary intermediate tubercles are though the tooth-row in the Smiths Valley species may be slightly longer. demonstrated in the paratype speci M2 of the former is slightly larger than the comparable tooth in the latter Peromyscus nesodytes Wilson,3 l and may be somewhat more compressed antero-posteriorly. In addition, approaches P. pliocenicus more clo the mandibular incisor of the Nevadan species is heavier and perhaps Kern River species differs from the slightly more compressed, although degree of compression is difficult to (1) more hypsodont; (2) internal determine accurately with the present material. Citellus sp.3 from the (3) intermediate tubercles variably Thousand Creek beds of northern Nevada is too inadequately known to appear to be absent. (4) MI, alth permit comparisons. No. 1965 agrees with the Thousand Creek specimen posterior diameter, is perhaps less tr internal re-entrant angle shows a te1 in size. Other Pliocene Sciuridre appear to be clearly distinguishable from becoming very shallow and open in < the Kern River species. notch in P. nesodytes. In the rel: For purposes of comparison it may be stated that the alveolar length, P4-M3, in No. 1965 is 8.4 millimeters. 1 It is possible that the character of P. plioce11icua than to the individual as a 1 terial does not suggest necessarily that t J. C. Merriam, C. Stock, and C. L. Moody, Carnegie Inst. Wash. Pub. No. 347, pt. 3, teeth. 68-69, fig. 23, 1925. • R. W. Wilson, Carnegie Inst. Wash. Pub. No. 473, pt. 2, 19-20, pl. 1, fig. 5, 1936. • L. Kellogg, op. cit., 432-433, fig. 16, 19 • L. Kellogg, Univ. Calif. Pub., Bull. Dept. Geo!., vol. 5, no. 29, 427-428, fig. 8, 1910. 'R. W. Wilson, Jour. Mamm., vol. 17, 4 NEW l'LIOCENE RODENT FAUNAS FRO:\l OREGOX AND CALIFORNIA >NTOLOGY 15

OF FAUNA Cricetid:e Peromyscus pliocenicus n. sp. Type-Fragmentary right ramus with MI-M2, No. 1966, C.I.T. Coll. Vert. Pale. (Plate 3, figs. 2, 2a). Paratypes-Fragmentary left ramus with MI-M2 and alveolus for M3, are limited to a right ramus No. 1967 (Plate 3, fig. 4); fragment of right maxillary with Ml and incom­ l965 (Plate 3, figs. 1, la) ; a plete M2., No. 1968 (Plate 3, fig. 3); C.I.T. Coll. Vert. Pale. ments of an upper and a lower Geological Age and Locality-Middle Pliocene Kern River beds, Kern amily. Specimen No. 1965 is County, California. C.l.T. Locality No. 49. m given above. ~ n and part of the first molar SPECIFIC CHARACTERS ·e difficult to make. No. 1965 t specimens of Callospermoph­ Cheek-teeth hypsodont, but crowns show tendency to wear to flat sur­ 'V the pronounced fore-and-aft faces; without accessory folds although intermediate tubercles may be ristic of typical Citellus, and present or absent. MI with divided antero-median cusp; tip of antero­ genus in the restricted sense. internal re-entrant angle becomes isolated with wear leaving a broad, erior two cusps were probably shallow re-entrant. M3 relatively unreduced. Pit or foramen in maxillary bone lateral to anterior root of Ml. Size large, slightly exceeding any condition. The mandibular 1 ~ed furrows. Thus, No. 1965 known fossil species, but approximating that of Peromyscus nesodytes. road way, and to be related to IS and Otospermophilus, which DISCUSSION ,1 Citellus and typical Sciurus. Peromyscus pliocenicus is distinguished from all other fossil species of d to Otospermophilus gidleyi the genus, except P. antiquus and P. nesodytes, by its large size. Even .{), and Citellus? sp. (Smiths these two species are somewhat smaller, in length of tooth-row, than the It is about the size of these Kern River type. Lcter of the dentition, at least Compared with P. antiquus Kellogg 2 from the Thousand Creek beds of n dentition of the Kern River northern Nevada, P. pliocenicus is not only larger but is distinguished by the following selected differences: (1) antero-median cusp of MI divided, md Moody 1 from the Rattle­ resulting in 'an external cusp rather than a narrow sloping ridge as in the nates the Kern River species Nevadan species; (2) MI relatively longer; (3) antero-posterior ridge of former is deeper. The talonid M2 may be slightly less developed; and ( 4) intermediate tubercles are te than in the Oregon species. variably developed, whereas they are entirely absent in P. antiquus. Rela­ ~ o determine because of break­ tively large intermediate tubercles are present in the type of P. pliocenicus .n the specimen. Citellus? sp.2 (Plate 3, figs. 2, 2a), but are absent in the paratype No. 1967. Rudi­ proaches No. 1965 in size, al­ mentary intermediate tubercles are also present in the upper cheek-teeth as pecies may be slightly longer. demonstrated in the paratype specimen, No. 1968. 3 comparable tooth in the l.a~ter Peromyscus nesodytes Wilson, from the Santa Rosa Island Pleistocene, tero-posteriorly. In add1t10n, approaches P. pliocenicus more closely in size than does P. antiquus. The ecies is heavier and perhaps Kern River species differs from the island form in the following characters: of compression is difficult to (1) more hypsodont; (2) internal cusps of cheek-teeth perhaps narrower; 3 (3) intermediate tubercles variably developed, whereas in P. nesodytes they :erial. Citellus sp. from the appear to be absent. ( 4) MI, although of approximately the same antero­ is too inadequately known to posterior diameter, is perhaps less triangular in outline; the tip of the antero­ the Thousand Creek specimen internal re-entrant angle shows a tendency to become isolated, the remainder >e clearly distinguishable from becoming very shallow and open in contrast to the more permanent V-shaped notch in P. nesodytes. In the relatively unworn MI, the antero-external ated that the alveolar length, 1 It is possible that the character of size is more applicable to the cheek-teeth of P. plioce11icus than to the individual as a whole. However, the present fragmentary ma· terial does not suggest necessarily that this individual possessed relatively large cheek­ igie Inst. Wash. Pub. No. 347, pt. 3, teeth. • L. Kellogg, op. cit., 432-433, fig. 16, 1910. pt. 2, 19-20, pl. 1, fig. 5, 1936. 1 R. W. Wilson, Jour. l\famm., vol. 17, 408-410, 1 fig., 1936. ,1, 5, no. 29, 427-428, fig. 8, 1910. NEW PLIOCENE RODENT FA 16 CONTRIBUTIONS TO PALJEONTOLOGY exhibits rudimentary tubercles. re-entrant angle is very deep, extending almost to the anterior margin of the cheek-tooth pattern. the tooth. With wear the tip of this re-entrant also becomes isolated. In l'viembers of the subgenus H P. nesodytes this re-entrant angle apparently never was so deep, although supplementary tubercles or with wear may have obliterated any enamel islet resulting from isolation. development. The largest speci (5) Judging from the alveolus, M3 is less reduced. tinctly smaller than P. plioceni< the dentition of No. 1966 exceed Comparative measuremenl8 (in millimeters) It is not to be assumed, on th· plete comparisons given above, t Peromysc11s pliocenicus P. nesodytes to any Recent subgenus. As a r P.antiquus C.I.T. No. 1780 any of the modern subgenera 'i U.C. No. 12571 (type) C.l.T. No. 1966 C.l.T. No. 1967 (type) Santa Rosa Pliocene. Limited comparisons (type) (paratype) Thousand Creek Island as the simplest means of elimin: Kern River beds Kern River beds Pleistocene of Recent species. However, i1 common in the middle Pliocene Ml-ID, ah-eolar for P. californicus, the large perc length ...... ·· ·· ····· ·· ··· 6.1 5.2 5.8 tion. This fact may indicate a MI-m, alveolar the Pliocene forms with, of cou: length ...... 4.3 ...... MI-m, occlusal ·········· ·· ·· ··············· the passing of geologic time . length ..••...... u u 3.9 ...... Ml, antero-posterior diameter ...... 2.2 2.3 2.0 2.5 Li MI, transverse diameter ...... 1.4 1.5 1.5 1.6 m, antero-posterior Hypolagu diameter ...... 2.0 2.0 1.8 1.9 m, transverse Leporid remains are rather abl diameter ...... 1.6 1.6 1.6 1.6 remains include fragmentary ra ID, antero-posterior diameter ...... 1.6 1.5 and skeletal parts. Two speciee ID, transverse ······ ··· ······ is near Hypolagus limnetus of t diameter ...... 1.2 1.3 material representing this form i men with the palatal portion of Peromyscus pliocenie111 on both sides ; a left maxillary ' C.I.T. No. 1968 (paratype). Kern River beds right ramus with P3"-M3, No. ML antero-posterior diameter. . . • . . . • • • . . . . • . • . . • . . . . . • • ...... 2.3 P3-M2, No. 1973 (Plate 2, fig. ML transverse diameter...... • . . • . • . . . • . • • • ...... • . . • 1.5? No. 1974. Although the Kern Ii with H. limnetus in approximatE Other known fossil forms either are much smaller than P. pliocenicus or depth of re-entrant folds in this are closely related to living representatives of the genus. internal border rounded (Plate ~ The Kern River species is clearly separable from Recent species of the tively deep, a point of resembla genus. It is sharply marked off by its large size from all except members of the plications on the mediar of the subgenera Megadontomys, Peromyscus, and Haplomylomys. Some teeth is apparently similar to tha members of the subgenus Megadontomys apparently exceed P. pliocenicus in size. However, in the former, supplementary enamel loops and tubercles less deep and complex on the ave regard in the Kern River form. are highly developed, whereas in our species the accessory tubercles between 2 the primary cusps are apparently not always present, and, most important, H ypolagus furlongi Gazin no accessory loops or folds of enamel enter into the cheek-tooth pattern of H. limnetus. However, the Ke either the superior or the inferior dentition. M egadontomys is an inhabitant by a less triangular P3 with dee of southern Mexico and Central America. of H. furlongi the antero-intern Many of the southern species of the subgenus Peromyscus are character­ Californian material is more likE ized by relatively large size. These species are all smaller than the Cali­ Other species of H ypolagus d fornian type and, moreover, possess upper cheek-teeth with supplementary Kern River type. H. vetus, al tubercles at the buccal margin which wear to form accessory enamel folds 1 C. L. Gazin, Proc. U. S. Nat. Mus., in the cheek-teeth. These supplementary folds may be small but are ap­ "Ibid., 118-119, fig. 4, 1934. parently always present. No. 1968, a maxillary fragment of P. pliocenicus, NEW PLIOCENE RODENT FAUNAS FROM OREGON AND CALIFORNIA 17 ~ONTOLOGY exhibits rudimentary tubercles. However, these cuspules do not enter into nost to the anterior margin of the cheek-tooth pattern. ·ant also becomes isolated. In Members of the subgenus HapLomylomys possess cheek-teeth without y never wa.s so deep, .altho~gh supplementary tubercles or with such tubercles in a rudimentary state of islet resulting from isolation. development. The largest species of the subgenus, P. californicus, is dis­ :luced. tinctly smaller than P. pliocenicus. Moreover, the accessory tubercles in the dentition of No. 1966 exceed those of any specimen of H aplomylomys. 'n millimeters) It is not to be assumed, on the basis of the brief and by no means com­ plete comparisons given above, that P. pliocenicus bears a close relationship P.n.~od11te1 to any Recent subgenus. As a matter of fact, it is highly improbable that P. antiquus C.I.T. No. 1780 any of the modern subgenera were differentiated as early as the middle - U.C. No. 12571 (type) Pliocene. Limited comparisons with Recent subgenera are given merely Santa Rosa ~7 (type) as the simplest means of eliminating from consideration the large number Thousand Creek Island Pleistocene of Recent species. However, it is worth noting that large species were ds common in the middle Pliocene of the United States, whereas now, except for P. californicus, the large peromyscine forms are all southern in distribu­ 5.2 5.8 tion. This fact may indicate a general southward movement of many of the Pliocene forms with, of course, considerable evolutionary change with ··············· ·············· the passing of geologic time. 3.9 ·············· LAGOMORPHA 2.0 2.5 Leporidre 1.5 1.6 Hypolagus near limnetns Gazin 1.9 1.8 Leporid remains are rather abundant in the Kern River collection. These 1.6 1.6 remains include fragmentary rami and maxillre as well as isolated teeth and skeletal parts. Two species appear to be present, the larger of which ... l.6 1.5 is near HypoLagus Limnetus of the Hagerman fauna.1 The more complete ... 1.2 1.3 material representing this form includes No. 1969 (Plate 3, fig. 5), a speci­ men with the palatal portion of the maxillre preserved with Pa-M.2 present on both sides; a left maxillary with P2--M.2, No. 1970 (Plate 3, fig. 6); a 1kua right ramus with P3-M3, No. 1972 (Plate 2, fig. 5) ; a left ramus with (ern River beds 2.3 P3-M2, No. 1973 (Plate 2, fig. 4); and a second left ramus with P4-M2, ······················· 1.5? No. 1974. Although the Kern River form is from an older stage, it agrees ······················· with H. limnetus in approximate size and in the shape of P3, as well as in h smaller than P. pLiocenicus or depth of re-entrant folds in this tooth. P3 is rather broad with the antero­ of the genus. internal border rounded (Plate 2, fig. 4). The antero-external fold is rela­ ,ble from Recent species of the tively deep, a point of resemblance to the Hagerman type. The character Te size from all except members of the plications on the median re-entrant fold of the upper molariform ;us, and HapLomykYmy~. S~me teeth is apparently similar to that in H. limnetus. The fold may be slightly ~pparently exceed P. pLwcenicus less deep and complex on the average, but there seems to be variation in this 1tary enamel loops ·and tubercles regard in the Kern River form. , the accessory tubercles between Hypolagus furlongi Gazin 2 from Grand View is closely related to rs present, and, most important, H. limnetus. However, the Kern River form is apparently distinguished into the cheek-tooth pattern of by a less triangular P3 with deeper antero-external fold. Moreover, in P2 M egadontomys is an inhabitant of H. furlongi the antero-internal fold is deep and crenulated. P.2 in the Californian material is more like that of H. limnetus and is not crenulat.ed. ~enus Peromyscus are character­ s are all smaller than the Cali­ Other species of Hypolagus do not appear to be closely related to the cheek-teeth with supplementary Kern River type. H. vetus, although from an approximately equivalent to form accessory enamel folds •c. L. Gazin, Proc. U.S. Nat. Mus., vol. 83, no. 2976, 114-117, figs. 2-3, 1934. folds may be small but are ap­ 1 lbi4., 118-119, fig. 4, 1934 . .llary fragment of P. pLiocenicus, 18 CONTRIBUTIONS TO PAL&ONTOLOGY NEW PLIOCENE RODENT FA'C horizon, is larger and possesses a P3 in which the antero-external fold is Hypo shallower and the outline of the tooth somewhat narrower anteriorly. 1 Several of the H ypolagus spech Hypolagus edensis Frick from the middle (?) Pliocene Eden beds is represent a second, smaller speciei smaller than H. near limnetus; the inner borders of the lower molariform teeth are rounded, not angulate as in the Kern River form; and the antero­ by No. 1975, a fragment of left external fold in P3 is more anterior in position. teeth. In contrast to the larger K Hypolagus? apachens'is Gazin 2 compared with H. near limnetus is from what elongate and triangular. C an older stage. It is a smaller form with the median -fold of the upper collection appear referable to th4 molariform teeth less complexly plicated. In addition, P3 is relatively All these teeth are somewhat di longer and the antero-external fold in that tooth is shallow. postero-internal groove on the sid Lastly, H. browni (Hay) 8 appears to be more advanced in character of on the occlusal surface of the tc the postero-external fol!l of P3 than is the Kern River species. However, the fairly deep but shows a tendency antero-external fold is relatively less deep than in H. near limnetus. This 1975, the anterior limb of the folc species is smaller and from a distinctly later stage (early Pleistocene?). lower molariform teeth, as compar The Kern River material agrees more nearly with H. limnetus from relatively a little longer antero-p Hagerman than with any other species of the genus. However, it comes able make this character doubtful. from an earlier stage and may represent a distinct species. Since seven or by isolated teeth, may possess me eight species of Hypolagus, representing a wide variety of types, are already plexly plicated than in the larger known, nothing is gained by separating t.he Kern River species as a distinct in view of uncertainty as to the I type. since individual variation is seen i are available for comparison. 1\. Measurements (in millimeters) of Hypol,agus near limnetus Gazin below they appear to be more cor Remains of this species are nc No. 1971 No. 1970 No. 1969 designation. As a matter of fact dividuals, and difference in tooth P~-M.3,. alveolar length ...... 13.1 limnetus from the same locality known small species of H ypolagt P~-Mg, occlusal length ...... 9.3+ material available, proved inadeq ~-Mg, alveolar length ...... · 1· ...... I 11.9 Measurements (in milll Pj!-Mg, occlusal length...... 8.7- , ...... 8.4 P:J-PI, antero-posterior length ...... Greatest width across antero- ventral prominence of zygomat- ic arches ...... , ...... 30.5 Least antero-posterior length of, bony palate ...... 5.5

No. 1972 No. 1973 No. 1974

P:J-~. alveolar length ...... 13.6 ...... 14.l (a) P:J-~. occlusal length ...... 10.8- ··············· Depth of ram us beneath MI, measured on outside ...... 10.6 ...... lo.4

(a) .Approximate.

1 C. Frick, Univ. Calif. Pub., Bull. Dept. Geol., vol. 12, no. 5, 348, figs. 52-53, 1021. • C. L. Gazin, Carnegie Inst. Wash. Pub. No. 404, 67-69, pl. 3, figs. 1-4, 1930. • O. P. Hay, Proc. U. S. Nat. Mus., vol. 59, no. 2391, 63o-631, 1921; L. R. Dice, Papers Mich . .Acad. Sci., .Arts, and Letters, vol. 16, 379-382, figs. 8-11, 1932. ~TO LOGY NEW l'LIOCENE BODENT FAUNAS FROM OBEGON AND CALIFORNIA 19

'1 the antero-external fold is Hypolagns small sp. at narrower anteriorly. Several of the Hypolagus specimens from the Kem River beds appear to ~ ( ?) Pliocene Eden beds is represent a second, smaller species of the genus. The species is reprosented lers of the lower molariform by No. 1975, a fragment of left ramus with P3-MI, and several isolated River form; and the antero- teeth. In contrast to the larger Kern River species, P3 in this type is some­ what elongate and triangular. Only three third lower premolars in the ith H. near limnetua is from collection appear referable to the species of Hypo"lagus under discussion. ie median fold of the upper All these teeth are somewhat different, including one, No. 1976, with a n addition, P3 is relatively postero-internal groove on the side of the tooth in evidence as a slight fold th is shallow. on the occlusal surface of the tooth. The antero-external fold in PH is ore advanced in character of fairly deep but shows a tendency to be quite wide. In one specimen, No. . River species. However, t~e 1975, the anterior limb of the fold is parallel to the axis of the tooth. The n in H. near limnetus. ThIS lower molariform teeth, as compared with those of H. near limnetus, may be tage (early Pleistocene?}. relatively a little longer antero-posteriorly, but the limited remains avail­ arly with H. limnetua from able make this character doubtful. The upper molariform teeth, represented e genus. However, it comes by isolated teeth, may possess median re-entrant folds which are less com­ tinct species. Since seven or plex!y plicated tha.n in the larger sp~~ies. This state~ent is als? doubtful ~ variety of types, are a~e~dy in view of uncertainty as to the position of the teeth m the maxillary, and ~rn River species as a distinct since individual variation is seen in this character when large series of teeth are available for comparison. Moreover, if these teeth are viewed .from us near limnetus Gazin below they appear to be more complexly folded. Remains of this species are not complete enough to warrant a specific No. 1970 No. 1969 designation. As a matter of fact, the specimens may represent young in­ dividuals, and difference in tooth-pattern between this form and H. near limnetus from the same locality may be due to age. Comparisons with 13.l known small species of Hypolagus were made, but because of the limited 9.3+ material available, proved inadequate to reveal any real relationships. 11.9 ··············· Mell8Urements (in millimeters) of Hypolagus small sp. 8.4 No. 1975 P~PI, antero-posterior length ...... · · · 5.2

-...... 30.5

5.5

No. 1973 No. 1974

14.l (a) 10.8-

10.4

12, no. 5, 348, figs. 52-53, 1021. --69, pl. 3, figs. 1-4, 1930. l, 630-631, 1921; L. R. Dice, Papers figs. 8-11, 1932. CARNEGIE INST. "\VASHINGTON Pue. 487

PLATE l

M11lagauluat cf. monodon Cope Fms. 1, 4, 5-Isolated right P.i's; Nos. 1955, 1952 (reversed), and 1953; X 214. Fm. 2-lsolated left Pi (reversed); No. 1954; X 214. Fm. 3-lsolated left Pi (reversed); No. 1956; X 2%. FIG. 6-lsolated right Pi; No. 195S; X 214. 4 Fm. 7-Isolated Pi; No. 1957; X 214. FIGS. S, Sa, Sb-Left ramus with Pi-M3 inclusive; No. 72. Fig. S, occlusal view, X 214; fig. Sa, intero-lateral view, X %. ; fig. Sb, extero-Iateral view, X %. Oastort sp. FIGS. 9, 9a-Isolated left M2?; No. 1961; X 2% . Fig. 9, occlusal view (reversed); fig. 9a ~ intero-lateral view. Sizes of all figures approximate. Calif: Inst. Tech. Coll. Vert. Pale. Middle Pliocene, Rome, Oregon. - 7 ~

Sa C AR:-l"EGIE I x sT. 'YASHINGTOX Pus. 487, PAPER I - WILSON PLATE 1

- • . 2 3

m Cope eversed), and 1953; X 2%.. %.. l. 4 5 6

~o. 72. Fig. 8, occlusal view, X 2%.;. ateral view, X %. ig. 9, occlusal view (reversed); fig. 9a,

1, Rome, Oregon.

7

!J CARNEGIE INST. WASHINGTON Pue. 48'

PLATE 2

Hypolagus sp. FIGS. 1, la- Incomplete left ramus with P4- Ml; No. 1964. Fig. 1, occlusal view, X !?'4; fig. la, lateral view, X l'h. ~Iiddle Pliocene, Rome, Oregon. ~ Goniodo11tomys disjunctus n. gen. and sp. FIGS. 2, 2a, 2b--Fragmentary left ramus with MT- M2', genotype specimen; No. 1959. Fig. 2, occlusal view, X 9; fig. 2a, intero-lateral view, X 4'h; fig. 2b, extero-lateral view, X 4'h. Middle Pliocene, Rome, Oregon. Hypolagus vetus (Kellogg) FIGS. 3, 3a- Incomplete left. ramus with P3- MT; No. 1962. Fig. 3, occlusal view, X 2%; fig. 3a, lateral view, X 1 'h . Middle Pliocene. Rome. Oregon. - Hypolagus near limnetus Gazin FIG. 4--lncomplete left rnmus with P3-M2'; No. 1973; X 2 14; occlusal view. Middle Pliocene, Kern River Beds, California. FIG. 5-Right ramus with 1'3- M3; No. 1972; X l 'h; lateral view. Middle Pliocene, Kern River Beds, California. Size& of all figures approximate. Calif. Inst. Tech. Coll. Vert. Pale.

«

3a C ARNEGIE INST. WASHINGTON Pue. 487, PAPER I-WILSON PLATE 2

2

1964. Fig. 1, occlusal view, X 2% ; me, Oregon. gen. and sp. - M'l, genotype specimen; No. 1959. view, X 4 1h; fig. 2b, extero-lateral logg) 1962. Fig. 3, occlusal view, X 2%; •me. Oregon. s Gazin l73; X 2%; occlusal view. Middle

lateral view. :\Iiddle Pliocene, Kern

4 ~ I

s CARNEGIE INST. WASHINGTON Pue. 487

PLATE 3

Citellusl' sp. Fiils. 1, la- Right ram us with P4- :i\I2"; :Ko. 1965; X 3. Fig. 1, occlusal \•iew; fig. la, lateral ,·iew.

Peromyscus plioce11icus n. sp. FIGS. 2, 2a- Incomplete right ramus with )JT- )12°. type specimen; Ko. 1966. Fig. 2, occlusal view ( reversed). X 6%; fig. 2a. lateral view. X 4'h. FIG. 3- Fragment of right maxillary with )Il and incomplete Ml ( reYersed ), paratype specimen; No. 1968; X 6%. FIG. 4- Left ramus with MI- M'Z (re\·ersed). paratype specimen; ~o. 1967; X 6%. Hypo/agus near limnetus Gazin FIG. 5- Incomplete maxillre with right P;FM2 and left Pi!.-l\12. ; Ko. 1969; X l 'h. FIG. 6- Fragmentary left maxillary with P~-MZ; ~o. 1970; X 2'A.. Sizes of all figures approximate. Calif. Inst. Tech. Coll. Vert. Pale. Middle Pliocene, Kern River Beds, California. 3

5 CARNEGIE INST. WASHINGTON PUB. 487, PAPER I - WILSON PLATE 3

2

X 3. Fig. 1, occlusal view; fig. la, s n. sp. ~- type specimen; Xo. 1966. Fig. 2, al view. X 4%. incomplete :M:1 (reversed ), paratype pe specimen; X o. 1967; X 6 % . us Gazin eft PQ.- :M1; Xo. Hl69; X I'h. ·o. 1970; X 2%.

~ . Kern River Beds, California. 3

4

6