Subfamilies and Genera of the Soricidae

GEOLOGICAL SURVEY PROFESSIONAL PAPER 565 Subfamilies and Genera of the Soricidae

By CHARLES A. REPENNING

GEOLOGICAL SURVEY PROFESSIONAL PAPER 565

Classification, historical zoogeography, and temporal correlation of the shrews

UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1967 UNITED STATES DEPARTMENT OF THE INTERIOR STEW ART L. UDALL, Secretary

GEOLOGICAL SURVEY William T. Pecora, Director

Library of Congress catalog-card No. GS 67-175

For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 - Price 50 cents (paper cover) CONTENTS

Page Diagnoses and contents of subfamilies Continued Page Abstract.___-_------__-______-____ 1 Subfamily Soricinae Fischer von Waldheim, 1817.____ 27 Introduction.______1 Tribe Soricini Fischer von Waldheim, 1817.______29 Evaluation of characters...-______3 Genus Crocidosorex Lavocat, 1951______29 Diagnoses and contents of subfamilies.______7 Crocidosorex piveteaui Lavocat.______29 Subfamily Heterosoricinae Viret and Zapfe, 1951.____ 7 Crocidosorex antiquus (Pomel)______29 Genus Domnina Cope, 1873______7 Genus Antesorex Repenning, n. gen______30 Domnina thompsoni Simpson._-_-_--_.___ 8 Antesorex compressus (Wilson)_____---_-_- 31 Domnina gradata Cope._____--.---.._.._ 8 Genus Sorex Linnaeus, 1758.______31 Domnina greeni Macdonald______9 Genus Drepanosorex Kretzoi, 1941 ___-____-____- 32 Domnina n. sp______9 Genus Microsorex Baird, 1877_-___-_-_-_-_---_- 33 Genus Paradomnina Hutchison, 1966______10 Genus Alluvisorex Hutchison, 1966___-___-___-_- 33 Genus Trimylus Roger, 1885.______10 Genus Petenyia Kormos, 1934______34 Trimylus compressus (Galbreath)______11 Genus Blarinella Thomas, 1911____--__--__----- 34 Trimylus aff. neumayrianus (Schlosser)__ 11 Blarinella kormosi (Schlosser)______35 Trimylus neumayrianus (Schlosser)______12 Blarinella quadraticauda (Milne-Edwards). 35 Trimylus dakotensis Repenning, n. sp_.__ 12 Soricini incertae sedis______-_------35 Trimylus roperi Wilson.______13 "Sorex" stehlini Doben-Florin.______------35 Trimylus mawbyi Repenning, n. sp______13 "Sorex" hibbardi SulimskL______-_____--_-_- 36 Trimylus sansaniensis (Lartet)______14 Genus Zelceina Sulimski, l962_-_-_------36 Genus Ingentisorex Hutchison, 1966.______15 Tribe Blarinini Stirton, 1930______37 Subfamily Crocidurinae Milne-Edwards, 1868-1874. _ _ 15 Genus Adeloblarina Repenning, n. gen______37 "Sorex" pusilliformis Doben-Florin______15 Adeloblarina berklandi Repenning, n. sp 37 "Sorex" dehmiVirei and Zapfe______15 Genus Cryptotis Pomel, 1848______- - - - - 39 Genus Crocidura Wagler, 1832..______16 Cryptotis adamsi (Hibbard)______39 Crocidura sp., Butler and Hopwood______17 Cryptotis"? meadensis Hibbard- .__-_____ 40 Corcidura wongi Pei.______17 Cryptotis kansasensis Hibbard-______40 Crocidura taungsensis Broom.______17 Cryptotis parva (Say)___-_-_-_-__--_----- 41 Crocidura kornfeldi Kormos__-___-_-__-_ 17 Cryptotis mexicana (Coues). ______41 Crocidura obtusa Kretzoi______' 17 Genus Paracryptotis Hibbard, 1950 ______41 Crocidura zorzii Pasa______17 Genus Shikamainosorex Hasegawa, 1957______42 Genus Diplomesodon Brandt, 1853_-__------_-_- 17 "Sorex" dehrieli KowalskL______----_---_--- 42 Diplomesodon fossorius Repenning.______18 Genus Blarina Gray, 1838__._..__------42 Genus Feroculus Kelaart, 1852______18 Blarina gidleyi Gazin ______-_-_------43 Genus Miosorex Kretzoi, 1959.______18 Genus Blarinoides Sulimski, 1959.___-_--__---_- 44 Miosorex grivensis (Depe"ret)______19 Blarinini? incertae sedis.______------44 Genus Myosorex Gray, 1838.______19 "Sorex" kretzoii Sulimski______-_--_-- 44 Myosorex robinsoni Meester.______20 Peisorex pohaiensis Kowalski and Li ______45 Genus Paracrocidura Balsac, 1956______20 Tribe Neomyini Repenning, new name.____-_----- 45 Genus Praesorex Thomas, 1913_-_-_-__-___-_-__ 20 Genus Neomys Kaup, 1829___------_------46 Genus Scuti orex Thomas, 1913______21 Genus Petenyiella Kretzoi, 1956_____-_------46 Genus Solisorex Thomas, 1924______21 Genus Episoriculus Eilerman and Morrison-Scott, Genus Soricella Doben-Florin, 1964.______21 1951.___--__.___------47 Genus Suncus Ehrenberg, 1832__-_---____-_____ 21 Episoriculus gibberodon (Petenyi) ___-__--- 48 Suncus sp., Meester______22 Genus Chodsigoa Kastschenko, 1907______-_- 48 Genus Surdisorex Thomas, 1906.______22 Chodsigoa bohlini (Young)______-___-- 50 Genus Sylvisorex Thomas, 1905______22 Genus Beremendia Kormos, 1934_____------50 Crocidurinae? incertae sedis______23 Genus Nesiotites Bate, 1945.______-----_------51 "Limnoecus" micromorphus Doben-Florin.______23 Genus Nectogale Milne-Edwards, l870__-_----__- 51 "Sorex" collongensis Mein______24 Genus Soriculus Elyth, 1854.______--_- 52 Genus Chimarrogale Anderson, 1877__---_-_----- 53 Crocidura?, Mein.______24 Genus Anourosorex Milne-Edwards, 1870. ______53 Subfamily Limnoecinae Repenning, new name___. ____ 24 Genus Amblycoptus Kormos, 1926____------54 Genus Angustidens Repenning, n. gen______25 Genus Hesperosorex Hibbard, 1957__-_------55 Angustidens vireti (Wilson)______25 Genus Notiosorex Baird, 1877_------55 Genus Limnoecus Stirton, 1930.______26 Genus Megasorex Hibbard, 1950.______---_--- 56 Limnoecus tricuspis Stirton______26 Neomyini? incertae sedis______-__------57 Limnoecus niobrarensis Macdonald-______27 Deinsdorfia franconia Heller.______----_------57 m IV CONTENTS

Diagnoses and contents of subfamilies Continued Page Diagnoses and contents of subfamilies Continued Page Subfamily Allosoricinae Fejfar, 1966.______57 Names of fossil genera not used in text______58 "Sorex" gracilidens Viret and Zapfe.______57 Ancestry of the Soricidae______58 Paleozoogeography and correlation______59 Genus Allosorex Fejfar, 1966.______57 Key to living and fossil genera of the Soricidae____ 63 Soricidae incertae sedis______57 References.______66 Podihik kura Deraniyagala.______57 Index______71

ILLUSTRATIONS

[Figs. 1-41 are drawings of the Soricidae] Page Page FIGURE 1. Taxonomically significant differences. 22. Alluvisorex chasseae- ______33 2. Domnina thompsoni. 23. Petenyia hungarica-.------34 3. Domnina gradata. ______9 24. Blarinella quadraticauda ______35 4. Domnina n. sp______10 25. Adeloblarina berklandi-.------38 5. Trimylus compressus. 11 26. Cryptotis adamsi,.------39 6. Trimylus dakotensis _____ 12 27. Cryptotis parva------_ 1 40 7. Trimylus mawbyi ______14 28. Paracryptotis rex...------41 8. Crocidura russu'a. ______16 29. Blarina brevicauda- ______43 9. Diplomesodon pulchellum- 17 30. Blarina gidleyi------43 10. Miosorex grivensis______18 31. Neomys jodiens------46 11. Myosorex varius. ______19 32. Episoriculus caudatus______47 12. Praesorex goliath. ______20 33. Chodsigoa hypsibia,------49 13. Suncus caeruleus.. ______22 34. Chodsigoa salenskii - ______49 14. Surdisorex norae. ______23 35. Beremendia fissidens______50 15. Sylvisorex lunaris. ______23 36. Nectogale elegans ------51 16. Angustidens vireti. ______25 37. Soriculus nigrescens. ______52 17. Limnoecus tricuspis. 26 38. Chimarrogale himalayica,------53 18. Limnoecus niobrarensis_ _ 27 39. Anourosorex squamipes. ______54 19. Antesorex compressus.... 30 40. Notiosorex crawfordi.------56 20. Sorex araneus. ______31 41. Megasorex gigas______56 21. Microsorex hoyi- ______32 42. Diagram of a phylogeny of the Soricidae ___ 61 SUBFAMILIES AND GENERA OF THE SORICIDAE

By CHARLES A. KEPENNING

ABSTRACT largely on the basis of masticatory modifications; this basis of The shrews constitute a distinct and uniform but phylo- separation is supplemented by characters in the teeth. The genetically diverse family in which five subfamilies, one sub­ tribe Soricini contains at least 3 extinct genera, about 30 ex­ divided into three tribes, can be recognized by consistent mor­ tinct species, and 4 living genera. It is characterized by lit­ phologic differences. These differences are traceable back tle modification of its mandibular articulation or teeth and by a prominent entoconid crest on lf±. The tribe Blarinini con­ through the late and middle Tertiary shrews and point to a hypothetical common ancestor. The trends of the morphologic tains at least 4 extinct genera, 10 extinct species, and 2 living change within the subfamilies rather clearly indicate what this genera. It is characterized by slight to great modification of ancestor should look like. Because the great majority of fos­ its mandibular articulation with a broad interarticular area, sil shrews are represented only by teeth and mandibles, the no entoconid crest on M» and by a trapezoidal M2. The tribe morphologic characters here used to classify most shrews per­ Neomyini contains 4 fossil genera, 13 fossil species, and 9 liv­ ing genera. It is characterized by great modification of its tain to these parts of the animals. The extinct subfamily Heterosoricinae is known from the mandibular articulation with a narrow interarticular area, an entoconid crest on Mt in most forms, and a prominent bifid Oligocene to the Pliocene and in both the Old and the New Worlds. Four genera are recognized in North America; one upper incisor in most Old World forms. of them is known also from the Old World. At least 12 spe­ The extinct subfamily Allosoricinae is known from the late cies are included in these genera. The subfamily is not closely Miocene to the late Pliocene in Europe. It is characterized by related to other shrews, and all shrews in the subfamily lack unusual carnivorelike cheek teeth that are unequaled in spe­ an internal temporal fossa on the mandible, have extreme lin­ cialization by any other shrew. Two genera (one unnamed) gual offset of the lower mandibular articulation, and retain are included in this subfamily. at least some pigmentation in the teeth. INTRODUCTION The living subfamily Crocidurinae is known from the middle Miocene to the Recent in the Old World. In the subfamily are The living shrews are generally placed in two sub­ 11 to 13 fossil species and 10 living genera. The. subfamily families with the exception of the African armored appears most likely to be of Austro-Tethyan origin (mainly shrew Scutisorex, which some workers (for example, Ethiopian Africa and peninsular India) and is not well rep­ resented in the known fossil record. It is characterized by an Alien, 1917") have placed in a separate subfamily. The internal temporal fossa on the mandible, by little or no sepa­ Soricinae are holarctic in distribution, and the Crocid­ ration of the articular facets on the mandibular condyle with urinae are entirely of Old World distribution and a labial emargination of the interarticular area, by the reten­ largely tropical. As stated by Ellerman and Morri- tion of a primitive triangular P4, and by no pigmentation of the son-Scott (1951, p. 41), this division into two sub­ teeth. The extinct subfamily Limnoecinae is known from the early families "seems to be based chiefly on the Soricinae Miocene to the middle Pliocene in North America and closely having teeth pigmented at the tips, and the Crocid­ parallels the Old World crocidurines in morphology. Two urinae having teeth entirely white." These authors, genera and three species are recognized. The subfamily is Friant (1947), and several others have pointed out characterized by crocidurine features except that P4 is modi- that, although the Soricinae normally have pigmented fled from a triangular tooth to one in which a central crest runs from the apex of the single cusp to the posterior cingu- teeth, pigmentation in the subfamily is variable lum; at least some pigmentation is present in the teeth. very light in several genera and entirely missing in The living subfamily Soricinae is known from the late Oli­ a few genera. Because of this variability, several gocene to the Recent in the Old World and from the early zoologists, including Ellerman and Morrison-Scott Miocene to the Recent in North America. It is characterized by an internal temporal fossa on the mandible, by little to great (1951) and Heim de Balsac and LaMotte (1956, p. separation of the articular facets on the condyle of the mandi­ 141-142), have not recognized subfamilies of the ble with a lingual emargination of the interarticular area, by Soricidae; however, others, including Friant, have the modification of the primitive triangular cusp of Pt into a recognized them. labial shearing blade with a posterolingual basin, and by re­ tention of tooth pigmentation in most, but not all, genera. In considering fossil shrews, there has been a This subfamily is very large and is divisible into three tribes similar tendency to abandon the subfamilies of living 1 SUBFAMILIES AND GENERA OF THE SORICIDAE shrews (Hibbard, 1950, p. 129; and Viret and Zapfe, Soricinae, representing the surviving subfamilies; and 1951, p. 414). With the study of fossil shrews, how­ the Heterosoricinae, the Allosoricinae, and the Lim- ever, greater attention has been directed toward ad­ noecinae (new name), representing extinct groups of ditional criteria to recognize suprageneric taxa with­ subfamily rank. The living subfamilies are of ap­ in the Soricidae. Although he preferred to consider proximately equal size; there are 21 soricine species his groups as of less than subfamily rank, Stirton and 21 crocidurine species in the Palearctic and Indian (1930, p. 223-225) reviewed and grouped many of regions (Ellerman and Morrison-Scott, 1951), no the living shrews to evaluate a Miocene fossil. He soricines occur in Africa, and there are no crocid- selected five divisions of the family, which he called urines in North America. groups, on the basis of eight features (10 features Specimens of nearly all North American fossil for the Anourasorex group). Those features that shrews older than Pleistocene were examined during Stirton used to distinguish his groups include tooth this review as were specimens representing all but pigmentation, shape of the internal temporal fossa two x living genera of the world; however, specimens of the mandible (fossa pterygoidea, intertemporal of only six Old World fossil species were examined. fossa, or posterointernal ramal fossa of authors), In this review the description of Old World fossil degree of separation of the condyles of the mandible forms is largely from published accounts. Such a (or articular facets on the condyle), height of tooth review has obvious disadvantages as the criteria here cusps, separation of the tips of the protoconid and utilized are not always discussed or clearly figured the metaconid, position of the foramen ovale, form in published descriptions. For this reason little at­ of the talonid of Mz, and nature of the entoconid and tempt to revise the nomenclature of the Old World basin of the talonid of J/i and M2. fossil shrews is made, although they are placed in the Earlier Schlosser (1924, p. 5) suggested that the subfamilies which seem most appropriate as based absence of secondary cuspules on the cutting edge of upon published descriptions. The specimens examined the lower incisor and the shape of the ascending are listed in the discussion of the taxa they represent. ramus of his species Crocidura, korm^osi were distinc­ Subfamily, tribe (where applied), genus, and many tive of the genus.. Still earlier Miller (1912, p. 30, extinct species diagnoses are given in terms of the 86) pointed out that the Crocidurinae could be sepa­ criteria here used. For many shrews, particularly rated from the Soricinae on the basis of the structure for the living genera, other criteria are available in of P4, Lavocat (1951, p. 23-24), following Miller's the literature to enlarge the diagnoses but these crite­ criterion, seems to have been one of the first to place ria are not mentioned here because they cannot be much significance on the morphology of P4 in fossil applied uniformly to both living and fossil shrews. shrews. He regarded his Crocidosvrex piveteaui as This does not imply that other criteria are not valid intermediate between Crocidura and Sorex largely on or useful. the basis of the form of P±. Wilson (1960, p. 35, 36, The illustrations are somewhat stylized but were 38) also placed emphasis on the form of the F4 in done carefully with a camera lucida attached to a Sorex. The slightly more anterior average position microscope. There is an advantage in having all of the mental foramen in the Crocidurinae has also forms illustrated by the same technique and by one been mentioned by several authors as of possible value person. The illustrations are not drawn to a uniform in distinguishing this subfamily from the Soricinae. scale because size has appeared to be of no more than Therefore, it seems that there are a large number specific significance. When possible, P^ MI, and Ms of dental and mandibular features that may be use­ are shown in both occlusal and buccal view; P 4 and ful in separating the Crocidurinae from the Soricinae M 1 in occlusal view. M 2 almost always parallels besides tooth pigmentation, which has been considered the development of M t and hence is not shown in the to be a fallible criterion. Cranial and postcranial illustrations, even of types. In the few shrews where skeletal features, as well as criteria based upon the M 2 shows noticeable difference from M i, the difference soft parts, may also be useful, but with very few ex­ is noted in the text. The mandible is drawn in lingual ceptions these are not known in the fossil record. A view, and the condyles of the mandible are shown in review of the above features, of the dental formulae, posterior view. and of the morphology of P4 and J/1 in most fossil Embryological studies to determine the dental species and in most living genera leads to the con­ formulae of the soricids in terms of incisor, canine, clusion that subfamilies are recognizable in both living premolar, and molar teeth (Xrnbach-Christie-Linde, and fossil shrews. Five subfamilies are clearly recog­ i Feroculus and Solisorex from Ceylou. Photographs of British nizable in the fossil record: the Crocidurinae and the Museum (Natural History) specimens of these genera were examined. EVALUATION OF CHARACTERS

1912; Kindahl, 1960) are not in entire agreement and characters may be mutually supporting of subfamily the formulae seem to vary with the genus under study. assignment, and some are of subfamily importance Although this work is certainly important in evaluation when developed to an extreme degree. Most appear of the phylogeny of the Soricidae, it has not been con­ to represent evolutionary trends in several shrew sidered in this review. Dental formulae given in this lineages that are parallel in direction but not equal in report are arbitrarily based upon the frequent use of rate of change. the term "antemolar" in published descriptions of During this review it has been found that in the shrews. The term applies to those teeth between the characters here examined the shrews are remarkably soricid incisor and the molars. The formula used is: stable in comparison with other mammals. Unless Total=soricid incisor-antemolars including P 4-molars/ otherwise noted, the reliability of these characters soricid incisor-antemolars including P4-molars does not seem to be reduced by individual variability. Acknowledgments. The following persons have Atavism is rather rare although it does occur in living loaned or donated specimens used in this review and and presumably in fossil forms. are here sincerely thanked: Kazimierz Kowalski, Structure of P4 Polish Academy of Science, Krakow; Oldrich Fejfar, Domnina gradata Cope (Patterson and McGrew, Central Geological Survey of Czechoslovakia, Prague; 1937) is, in several respects, including the structure R. W. Hayman, British Museum (Natural History), of the P4, ideal as the prototype of all soricids. In London; David H. Johnson and C. Lewis Gazin, U.S. other respects it is clearly a heterosoricine shrew. In National Museum, Washington; Barbara Lawrence, occlusal view the P4 of D. gradata has a simple equi- Harvard University, Cambridge; Malcolm C. Mc- laterally triangular outline and a single large tri­ Kenna and Sydney Anderson, American Museum of angular cusp (figs. 1, 3). There is a posterior basal Natural History, New York; Claude W. Hibbard, cingulum that extends nearly to the anterior corner University of Michigan, Ann Arbor; Rainer Zangerl, of the tooth on both the lingual and the labial sides. Philip Hershkovitz, and Joseph C. Moore, Chicago Although the cingulum is prominent at the postero- Natural History Museum, Chicago; William A. labial corner, it does not overhang the root and cover Clemens, University of Kansas, Lawrence; Peter the alveolus as in most modern shrews. The posterior Robinson, University of Colorado Museum, Boulder; surface of the crown slopes uniformly to the cingulum Richard H. Tedford, University of California, River­ and is slightly concave between the posterolateral side; J. Reid Macdonald and Charles A. McLaughlin, corners of the crown. The posterolateral corners of Los Angeles County Museum, Los Angeles; Seth B. the triangular cusp are sharp so that slight ridges Benson and Ruben A. Stirton, University of Cali­ run from the apex of the cusp to the cingulum at the fornia, Berkeley; and J. Arnold Shotwell, University posterolabial and posterolingual corners of the tooth. of Oregon, Eugene. The -P4 is the largest tooth of the antemolar series with S. B. Benson, C. W. Hibbard, J. H. Hutchison, the exception of the procumbent incisor and possibly Miklos Kretzoi, D. E. Savage, R. A. Stirton, Heinz the first antemolar, which is unknown. The antemolar Tobien, and F. C. Whitmore, Jr., have reviewed the alveoli of some species of Trimyhis are similar to those manuscript and have offered many helpful sugges­ of Domnina gradata, but P4 is not the largest post- tions. The author has profited from extended dis­ incisor antemolar in Trimylus. cussions of the crocidurine shrews by P. M. Butler, Domnina is here placed in the subfamily Heteroso­ University of London. ricinae on the basis of features other than the crown EVALUATION OF CHARACTERS of P4. The P4 in the genera Trvmylus and Ingentiso- Of the features observed in the dental and mandib- rex appear to be similar to that of Domnma except ular morphology during this review, the structure of that details of the cingulum and cusp morphology the P4 is most useful in the recognition of subfamilies. have been lost in the development of a rounded, bul­ (See fig. 1.) The structure of the mandibular condyle bous tooth. The P4 of Paradomnina is like that of is nearly as useful, and pigmentation of the teeth Domnina. seems significant but is subject to exceptions. The The basic triangular pattern that is retained in the absence of the internal temporal fossa is a subfamily Crocidurinae differs only slightly from that in Dom­ character of the Heterosoricinae. The other morpho­ nma,. In the Soricinae, however, the crown is con­ logic characters, when considered singly, appear to be verted into an oblique blade by emphasis of a crest of taxonomic value only after the identity of the connecting the anterior cusp with the cingulum at the subfamily has been established. In combination, most posterolabial corner of the tooth and by suppression SUBFAMILIES AND GENERA OF THE SORICIDAE

Lingual Posterolingual basfn

Anterior Posterior

Heterosoricinae Limnoecinae Soricinae and Crocidurinae

OCCLUSAL PATTERNS OF LEFT P.

Labial emargination Lingual emargination

Limnoecinae Soricinae Heterosoricinae and Crocidurinae

Subfamilies of the Soricidae

Narrow interarticular' Broad area interarticular area

Soricini Blarinini Neomyini

Tribes of the Soricinae

POSTERIOR PATTERNS OF LEFT MANDIBULAR CONDYLES FIGURE 1. Taxonomically significant differences in structure of P 4 and mandibular condyles. of the anteroposterior lingual crest and development are clearly recognizable in all shrews considered in of a basin in its place. In most soricines the labial this review. The exceptions are all within the group blade combines with the posterior cingulum to form a of heavy-toothed and stout-jawed soricines where continuous L-shaped crest which encloses the postero­ other characters, such as widely separated mandibular lingual basin on both labial and posterior sides. The condyles, leave no doubt of subfamily affinity. labial crest of the triangular P4 is also accentuated in No soricid has ever developed a molariform P4, and the Limnoecinae, but it has shifted lingually nearly to for this reason the late Eocene Scrtwrnimia gracttis a median position. The lingual crest of the triangle Stehlin (1940), which has a well-developed trigonid is almost completely suppressed but no posterolingual and talonid in its P±, cannot be considered ancestral basin has formed because of the medial position of the labial crest. to any shrew; as recognized by Stehlin (1940, p. 306), These basic patterns of the P4 are diagrammatically Saturnima, appears to be an erinaceoid similar to shown in figure 1. With few exceptions these patterns Leptacodon. EVALUATION OF CHARACTERS

Pigmentation of no more than generic importance as far as diagnosis Pigmentation in soricid teeth seems to be remark­ of most shrews is concerned. ably stable insofar as fossilization is concerned and The soricid internal temporal fossa is a consequence is usually preserved even in Oligocene specimens. It of the loss of the zygomatic arch and much of the may be obscured, however, by coloration of the entire masseter muscle. The internal part of the temporal tooth during fossilization or altered to such a degree muscle assumes the function of the masseter in this that it is not visible in normal light. In some genera, situation, and its insertion on the medial side of the when its presence is not detectable in ordinary light, coronoid process is enlarged and is shifted ventrally, it can be seen under ultraviolet light as described by to a level approximating that of the masseter in other Patterson and McGrew (1937, p. 247). Hibbard mammals, for a more masseterlike force and leverage. (1950, p. 128) noted that reaction of pigmentation to Loss of the zygomatic arch and development of the ultraviolet light is not consistent in either Recent or internal temporal fossa also are associated with loss fossil specimens that he examined. of the masseteric fossa on the lateral side of the In many shrews, pigmentation is present only at mandible, for the same reason. With the exception the tips of the cusps and along the crests of the lophs of heterosoricines, no shrew has a masseteric fossa; so that heavy wear removes all pigmented parts of also, the heterosoricines do not have a soricine internal the tooth. Although pigmentation is usually of a temporal fossa and have retained a reduced zygomatic very conspicuous color, the teeth of some shrews, like arch. Nectogale, are pigmented over much of their surface, Patterson and McGrew (1937, p. 256-257), appar­ but the pigmentation is of such a light value and ently in a lapse of memory while thinking of the slight saturation that it may be overlooked. Apparent internal temporal fossa, stated that a deep masseteric lack of pigmentation in a fossil shrew, therefore, must fossa "extending downward much nearer the ventral be evaluated cautiously, particularly if the teeth have margin of the jaw" (than in the bat Mvniopterw) been darkened by fossilization. is a soricid character. This and several other features Nevertheless, pigmentation is an important char­ led them to place the genus Mystipterus Hall in the acter in the recognition of the subfamilies of the Soricidae rather than in the Chiroptera. The deep Soricidae. The teeth of Domnina gradata are ex­ masseteric fossa of Mystipterus clearly indicates, how­ tensively pigmented (the pigment is a pale yellowish ever, that the animal has a normal mammalian mas­ orange), indicating that pigmentation was acquired seter system because the fossa cannot form without early in the history of the Soricidae (Patterson and the masseteric muscle. McGrew, 1937, p. 247) or was inherited from the Hall's illustration (1930, pi. 38b) shows that the soricoid ancestor of the family. In Trwvylus, pig­ masseteric fossa of Mystipterus is well developed, and mentation is present but restricted; it is certainly his opinion that the genus is a bat seems much more present on the lower incisor, on PI, and on the upper reasonable than other opinions which label it a shrew. incisor. In the Crocidurinae and in the Allosoricinae, The lack of reduction in the size of the talonid of MS tooth pigmentation is lost. In the Limnoecinae it (the only tooth known) and the full development of seems to be present in most specimens, but Macdonald the V-shaped pattern on the talonid and on the (1947, p. 125) and also Hibbard (1957, p. 329) stated trigonid of M3 suggest, however, that Mystipterus may that Limnoecus niobrarensis Macdonald shows no pig­ be a talpid. It resembles Mydecodon martini Wilson mentation. James (1963, p. 36) pointed out, however, (1960, fig. 32) and is somewhat less like the living that in Macdonald's type the teeth are worn to a level Neurotrichus. beneath the base of the pigmented area on other speci­ Mandlfoular condyle mens of the genus. In the Soricinae, pigmentation is characteristic but is very light in some genera and Domnina (D. thompsoni Simpson, see McDowell, absent in a few; some soricines have been classified as 1958, p. 181-182, fig. 30E) has virtually one broad crocidurines because of the lack of tooth pigmentation. mandibular condyle which is constricted and offset in its center foreshadowing the doubled articular Internal temporal fossa of the mandible facets of later soricids. Separation, to form two This fossa (also called fossa pterygoidea, intertem- articulations, has chiefly taken place in the early poral fossa, and posterointernal ramal fossa by other Miocene shrews although in some the articular facets authors) is not pocketed in the Heterosoricinae (fig. may still be joined. This situation is little changed 2). It is deeply pocketed in all other subfamilies in the later limnoecines and in the modern crocid­ (figs. 8-41), and in these groups its form seems to be urines. Some of the living soricines and the extinct

240-305 O-67 2 6 SUBFAMILIES AND GENERA OF THE SORICIDAE allosoricines have greatly accentuated the separation tions being in the extinct crocidurine Diplomesodon of the facets. fossorvus Repenning (1965) and on the extinct al- On many mandibular condyles it is difficult to losoricine AUosorex stenodus Fejfar (1966). clearly distinguish the extent of the articular facets. Height of cusps Compare, for example, the variations in interpretation of the facets on Limnoecus niobrarensis made by Hib- There is a general trend throughout the soricids bard (1957, fig. 3A), by James (1963, fig. 9c), again to increase the crown height in the evolution of the by James (1963, fig. 11), and in this report (fig. 18). separate lineages, except in the Heterosoricinae; how­ All were made by careful examination of the same ever, relative cusp height is difficult to judge. In this specimen. review it has been found easier, in most shrews, to There are three basic patterns to be seen in the estimate the stage of evolution by observing the rela­ mandibular condyles of the shrews (fig. 1). In nearly tive height above the labial cingulum at which the all shrews these are recognizable regardless of the bottom of the reentrant valley between the protoconid degree of separation of the articular facets, regardless and the hypoconid opens on the labial side of M±. of the vagueness of the limits of the articular sur­ In Domnina and the many earlier Miocene shrews, faces, and regardless of the antiquity of the shrew this reentrant opens at or very close to the level of considered insofar as the present record is known. the cingulum. In many of the more advanced forms, In the subfamily Heterosoricinae the lower articular there is an appreciable distance between the reentrant facet of the mandibular condyle is not much offset and the cingulum. In many living shrews, however, ventrally from the upper facet (more so in Trimytus this is not so and the value of this character is quite than in Donwdna) but is placed lingually to an ex­ variable with different lineages. Although partly treme degree so that virtually no part of the lower related to cusp height, this feature may also be related facet lies below the upper facet. This situation is to a tendency of the metalophid-protolophid junction not found in any other subfamily. to shift slightly labially in geologically later genera. In the subfamilies Crocidurinae and Limnoecinae, Heel of M3 the lower facet of the mandibular condyle is beneath Reduction of the talonid of the M3 in Croddura to the upper, and the two facets are continuous along a single trenchant cusp has frequently been cited as the lingual side of the posterior face of the condyle. a character distinctive of the crocidurines. It is pre­ If separated, an elevated and usually rounder bridge valent in this subfamily but is also found in some on the lingual side of the posterior face of the condyle, soricines, like Cryptotis, It also occurs in the Lim­ the interarticular list of Sulimski (1959, p. 128-129), noecinae but not in the Heterosoricinae. The heel of connects the two facets, and there is an excavation Ms is virtually lost in a few shrews, and the entire of the interarticular area on the labial side of the Mz is missing in five genera. posterior face of the condyle. This condylar structure is invariable in these two subfamilies but in a few Position of the mental foramen shrews, as in the giant species of Suncus, the articulat­ The mental foramen lies beneath the middle of Mi or ing surface connecting the upper and lower facets the anterior root of M2 in the Heterosoricinae; beneath is so broad that it is impossible to decide whether P4 to beneath the anterior root of Mj in the Crociduri­ the interarticular excavation is on the lingual or labial nae; beneath or slightly ahead or behind the anterior side of the posterior face of the condyle. root of MI in the Limnoecinae; and in the Soricinae In the subfamily Soricinae the lower facet is below beneath the middle of M± in most Miocene to Recent the upper in posterior view, and the two facets are species examined but beneath the F4 in Croddosarew and continuous along the labial side of the condyle. If in Antesorex, new genus, of the late Oligocene to early separated, as they are in all known forms younger Miocene. In all lineages it tends to have shifted than middle Miocene, the two articular facets are posteriorly in geologically younger forms. The position connected by an interarticular list that is on the labial side of the posterior face of the condyle, and the of the foramen seems fairly constant within the sub­ lingual side of the interarticular area is excavated. families but the overlap in position between subfamilies This pattern is also invariable, but the position of the is great. In addition there have been a very few articular surfaces is otherwise greatly variable in the individuals of living shrews noted during this review subfamily, largely in degree of anterior placement of that have either doubled mental foramina or have the the lower facet. This variation in articulation is not mental foramen significantly anterior to its position in known in other subfamilies, the most extreme excep­ other individuals from the same local population. DIAGNOSES AND CONTENTS OF SUBFAMILIES

In the figures, the mental foramen is usually shown outline of M2, and the position of the zygomatic process with the labial view of the lower teeth. These teeth are of the maxillary relative to M2, the nature of develop­ generally separated for clearer illustration, and a line is ment of the superior pterygoid fossa of the mandible, shown between the foramen and the tooth (either P4 or the extent of the external temporal fossa of the mandi­ MI) with which it is properly oriented. ble, and a few other features are characters of tribal significance in the Soricinae. Closeness of the protoconld to tlie metaconid on Ma Stirton (1930, p. 244) cited the closeness of these two DIAGNOSES AND CONTENTS OF SUBFAMILIES cusps in his "Crocidura group" as being distinctive. Subfamily HETEROSORICINAE Viret and Zapfe, 1951 Although it seems that it might reflect the height of cusp P± has triangular cusp, little or no posterior sulcus on to some degree, it does appear to be characteristic of the cusp, and root transversely broadened or even doubled living crocidurines and particularly of the extinct in most species; MI low cusped reentrant between limnoecines. I have had considerable difficulty in mak­ protonid and hypoconid opens at level of cingulum; Ms ing the comparison with even moderately worn teeth has double-cusped or crescentic talonid (this tooth is and consider it virtually impossible to judge the close­ absent in Ingentisorex from the late Miocene of Ore­ ness of these cusps in figures of worn teeth. As a conse­ gon) ; P4 and M1 have no emargination of posterior quence, the character has not been considered to be of basal outline and no posterior projection of hypoconal much value in this review except where the two cusps are flange; in Trimylus, M2 is trapezoidal and the zygomatic obviously close. In many shrews where it seems of process of maxillary originates opposite it, but in the importance, an anterior view of MI is figured. less specialized Domnina, M2 is rectangular and the Structure of P* and M1 zygomatic process originates behind this tooth; tooth There appears to be no subfamily distinctiveness in pigment present at least to some degree; masseteric fossa these teeth other than for the Heterosoricinae. This well developed; internal temporal fossa lacking or very subfamily is alone in the absence of an emargination of shallow and never pocketed; ventral condyle of mandi­ the posterior basal outline of the tooth in all forms and ble offset lingually much farther from plane of mandible in the consequent failure to develop a posterior projec­ at lower sigmoid notch than in any other subfamily of tion of the hypoconal flange. This structure is other­ shrews and at approximately right angles to vertical wise characteristic of most, but certainly not all, genera axis of horizontal ramus; mental foramen below middle of the Soricidae. of MI to below anterior root of M2 and usually associated with depressed area on labial face of mandible or with Miscellaneous features groove running anterodorsad from foramen to region Secondary cuspules on the cutting edge of the lower of P4, or both; a small foramen of unknown function incisor seem to have no significance in subfamily charac­ (Mawby, 1960, p. 955) on lingual face of mandible near terizations or in evolutionary patterns. The dental inferior margin below MI or M2 ; medium to very large formulae show a broad trend toward reduction in num­ size; early Oligocene to late Miocene, North America; ber of antemolars, but the trend cannot be correlated middle Oligocene to early Pliocene, Europe. "Sorex" with subfamily distinctions. Overlap of the alveolus kinkelini Zinndorf (1901) appears to be from middle and part of the ramus by the labial cingulum of P4 is Oligocene (Eupelian) deposits. The specimen is a most extreme in some soricine genera but is present to lower incisor that clearly represents Trimylus, but the some degree in both limnoecine and crocidurine shrews. species is not definable. As noted by Zinndorf (p. 134), A prominent entoconid on MI is characteristic of most it is closest to Trimylus neumayria/rms. shrews, but its variation in prominence and in position seems to be a tribal or generic character. A depression Genus DOMNINA Cope, 1873 on the labial surface of the horizontal ramus in the Genotype: Domnina gradata Cope, 18Y3 region of the mental foramen is also a generic character in the Heterosoricinae and in some Soricinae. A bifid P4 triangular in occlusal view with sharp-cornered or fissident upper incisor is conspicuous in the Heter­ triangular cusp; MI has very prominent entoconid dis­ osoricinae and in the tribe Neomyini; and is slightly tinctly separated from hypolophid and united to meta­ developed in many incisors in the tribe Soricini. It is conid by high ridge (entoconid crest); teeth not bul­ known in only one genus of the Crocidurinae. This bous ; cingulum strong but not inflated; mental foramen feature is formed by a smaller cusp branching from the below middle of MI and associated only with minor medial side of the principal (anterior) hooklike cusp of depression on labial face of mandible. Dental formula the falciform first upper soricid incisor. The occlusal 1-5-3/1-5-3 or 1-5-3/1-4-3. 8 SUBFAMILIES AND GENERA OF THE SORICIDAE Domnina is the oldest and most primitive shrew here than in any other subfamily. The presumed reduction recognized. In form of P4, height of crown of M^ in number of antemolars in D. gradata, their conspicu­ relatively slight reduction of the heel of M$, little sepa­ ous crowding between the molars and the insicor, and ration of the mandibular condyles, lack of a deep and the transversely broadened or doubled roots of these pocketed internal temporal fossa, lack of a posterior antemolars are marked differences between these two emargination on P4 and the upper molars, pigmentation species. of the teeth, and other features related to the structure Specimen examined. The type, American Museum of of the skull (see McDowell, 1958, p. 182-184, and Pat- Natural History 32647. terson and McGrew, 1937, p. 255), Domnina could easily Domnina gradata Cope, 1873 be considered ancestral to most other shrews. Domnina is precocious, however, in the posterior position of the Figure 3 mental foramen and in the extreme lingual placement of A species of Domnina with dental formula 1-5-3/ that part of the mandibular condyle homologous to the 1-4-3; lower antemolars crowded together, four in num­ lower condyle of other shrews. It is atypical, in com­ ber, with transversely broadened or incipiently doubled parison with genera of the other subfamilies, in its trans­ roots; somewhat larger size than D. thompsoni; heavily versely broadened or double-rooted premolars. These pigmented teeth nearly all of MI above cingulum is last features clearly indicate that Domnina is a spe­ pigmented. Middle Oligocene (Orellan) of Colorado cialized shrew belonging in the Heterosoricinae and not and Nebraska and Oligocene of Wyoming. a generalized ancestor to all shrews. This is a very well-known shrew (see Patterson and McGrew, 1937). The first postincisor antemolar of the Domnina thompsoni Simpson, 1941 lower dentition is not known. The similarity in alveolar Figure 2 pattern between this species and Trimylus from the late A species of Domnina with lower antemolars not Miocene of Oregon suggests that the P4 of D. gradata crowded anteroposteriorly; five (?) antemolars in addi­ may not be larger than the first postincisor antemolar. tion to soricid incisor; roots of antemolars not trans­ The second and third lower antemolars are of a form versely broadened; somewhat smaller size than D. similar to, but much smaller than, P4. gradata. Early Oligocene (Chadronian) of Montana. The P4 is typically soricid but differs in certain This is the only known specimen of Domnina on which characters. The protocone, typically small in most the mandibular condyles are well preserved. They are shrews, is either not present or is represented by a small drawn apart but not separated (see McDowell, 1958, bladelike elevation on the hypoconal cingulum. The fig. SOD, E). It is interesting to note, even in this paracone is more prominent and of a more conical form early stage of evolution of the soricid doubled condyle, than in most shrews. The parastyle is minute and that the lower anticular facet is well displaced lingually hardly more than a cingular irregularity. Finally, in the heterosoricine fashion and to a far greater extent there is no emargination of the posterior basal outline

FIGURE 2. Domnina thompsoni Simpson, type, AMNH 32647. Left man­ dible. Medial view from Simpson (1941); posterior view from McDowell (1958). DIAGNOSES AND CONTENTS OF SUBFAMILIES 9 although the basin between the parametaconal blade metaconid is lower than in D. gradata, and the labial and the hypoconal cingulum is very deep. The P4 is cingulum is not present below the protoconid (although known only on USNM 12841. his fig. 5A seems to show that it is). His measurements The M1 is typically soricid except that it lacks the indicate a shrew the size of D. gradata. Early Miocene posterior emargination and the protruding hypoconal (Arikareean) of South Dakota. flange which results from this emargination. This specimen was not seen. Consistent with its slightly younger geologic age, the Domnina n. sp. mental foramen is slightly posterior of the position in D. thompsoni and is a little behind the midpoint of MI Figure 4 rather than a little behind the anterior root of J/i. Material. U.S. National Museum 23095, isolated left On the basis of USNM 12841 from Wyoming, P± and MI with posterior half of the talonid missing. McDowell (1958, fig. SOB, C) concluded that Domnina These two teeth were found in situ within 3 inches of gradata lacked the zygomatic arch as do most shrews. each other and have the same degree of wear; it is Further preparation of the damaged skull of this speci­ assumed that they belong to the same individual. Col­ men leads me to feel not only that the lack of an arch lected by C. A. Repenning in 1959. is not demonstrable on this specimen but that the size Locality and age. U.S. Geological Survey vertebrate and shape of the broken end of the maxillary root of the loc. M1033, eastern Caliente Range, San Luis Obispo zygoma suggests that a relatively large and horizontally County, Calif. Early Miocene. Found in association broad arch or zygomatic process was present. This with ArcJiaeolagus and ParaMppus in beds which con­ conclusion is almost mandatory when one considers the tain, in the same area, mammals believed to belong in very large, deep, and extensive masseteric fossa on the the Arikareean North American Provincial Age mandible. It should be noted, also, that McDowell's (Repenning and Vedder, 1961). illustration (fig. 30C) shows a slight posterior emargi­ Diagnosis. A small species of Domnina with P4 that nation on P4 and M1. This emargination is not present has transversely broadened root; size one-fourth smaller on the specimen, but its absence could not be established than D. gradata; pigment reduced to only tips of cusps. until some matrix and a thick coating of glue were Description. USNM 23095 is a shrew differing from removed from the teeth. Domnina gradata, in the parts represented, only by its Specimens examined. Chicago Natural History Mu­ size and the amount of pigmentation of its teeth. P4 is seum P25750, P15320, P25754, and P25749; U.S. Na­ a triangular tooth with a single triangular cusp and a tional Museum 12841. posterior cingulum which extends forward on both sides Domnina greeni Macdonald, 1963 nearly to the anterior apex of the tooth. The cingulum According to Macdonald (1963, p. 168), the ridge does not overhang the root and would not have hidden (the entoconid crest) connecting the entoconid to the either the root or the alveolus. The root of this speci-

.Entoconid

Entoconid crest

Occlusal-lateral view to show antemolar alveoli

1-5-3 1-4-3

FIGURE 3. Domnina gradata Cope, USNM 12841. Left mandibles; P 4 and M1. M 3 is from CNHM P15320, and lower antemolar alveoli are from CNHM UC 1552. 10 SUBFAMILIES AND GENERA OF THE SORICIDAE men was lost, but on the basis of the area of attachment the mandible; antemolar roots compressed antero­ P4 was either double rooted or at least transversely posteriorly as in most heterosoricines; cheek teeth widened and anteroposteriorly compressed to about half elongate and not bulbous, mandible elongate in contrast of the width. Only a faint trace of yellowish pigment to Trimylus. Late Miocene (Barstovian) of Oregon. can be seen near the cusp apex. Maximum width of P4 Only one species, the genotype, is known. is 0.8 mm, and the anteroposterior length is about Specimens examined. The type and referred speci­ 1.0 mm. There is a cavity formed by caries on the mens listed by Hutchison (1966, p. 5). cingulum near the posterolingual corner of the tooth. Genus TRIMYLTTS Roger, 1885 2 The posterior half of the talonid is missing from the MI, and the tooth is well worn. The cingulum is strong Genotype: Trimylus schlosseri Roger, 1885 = from the paraconid to its termination at the fracture Sorex sansaniensis Lartet, 1851 across the talonid. The entoconid is prominent and is P4 triangular in occlusal view (lacking other details connected to the metaconid by a crest that appears to because of bulbous form), smaller than first antemolar be relatively as high as that in D. gradata. The meta- behind incisor (but first antemolar known only in late lophid joins the protolophid slightly labial of the midline Miocene and early Pliocene forms); MI very low of the tooth. The reentrant between the protoconid crowned, swollen, with inflated cingulum, these char­ and the hypoconid opens on the labial side of the tooth acters increasing in prominence in geologically younger at the cingulum and actually blends with the cingulum forms; entoconid present on M\ in earlier forms but as in D. gradata. A faint trace of yellowish pigment is more detached from metaconid than in Domnina and detectable on the labial face of the paralophid. The closely attached to hypolophid, decreasing in impor­ MI is about 1.5 mm in length (including an estimation of tance and becoming a distal part of hypolophid in the missing part of the talonid) and about 1.0 mm wide geologically youngest forms; talonid of M3 with cres- across the protolophid. D. n. sp. (USNM 23095) is centic loph and not distinctly bicuspid as in Domni/na; about as much smaller than D. thompsoni as D. thomp mental foramen below middle of MI to middle of M2 soni is smaller than D. gradata,: (more posterior than in any other shrew) and placed D. gradata* in a depression on labial face of mandible which leads Mi D. n. sp. D. thompsoni* (mean] anterodorsally to premolar region; horizontal ramus Length...______1. 5± 1.8 2.3 Width_____-_-_____---___- 1.0 1.2 1.5 massive and increasingly so in geologically younger *Data from Simpson (1941, p. 2). forms; mandibular condyles widely separated in heter­ osoricine pattern (lower condyle displaced more lin- The closer approach in size of D. n. sp. to D. thomp­ gually than ventrally); pigmentation reduced and in soni would suggest a closer relationship. The antero­ most forms only on incisors and anterior antemolars. posteriorly compressed root of the P4 of D. n. sp., how­ Dental formula 1-5-3/1-4^3 or 1-^3/1-3-3. ever, and the relative age of the three species indicates There are at least three species of Trimylus in Europe a closer relationship between D. n. sp. and D. gradata. and at least four in North America. The relationship Except for its alveolus, P4 is not known in D. thompsoni. between these is obscure but suggests two lineages. One Genus PARADOMNINA Hutchison, 1966 is confined to North America, and the other to Europe. Genotype: Paradomnina r&lictw Hutchison, 1966 If two lineages are represented, they are not readily A heterosoricine shrew resembling Domnina but lack­ recognizable because of parallel, but not identical, evo­ ing a prominent entoconid crest joining the entoconid lutionary trends on each continent. The presence of two to the metaconid on MI and J/2 ; postentoconid valley 2 In a recent publication, Doben-Florin (1964, p. 15-16) traced the history of the names Heterosorex Gaillard (1915) and Trimylus Roger lacking on M\ and M? so that hypolophid unites with (1885) and unquestionably established the priority of Trimylus. See entoconid; five postincisor antemolars including P4 in synonomy under Trimylus sansaniensis.

Anterior Lingual

FIGURE 4. Domnina n. sp., USNM 23095. Right P ^ and MI. DIAGNOSES AND CONTENTS OF SUBFAMILIES 11 lineages is suggested by the position of the conspicuous Natural History Museum UM400 and was collected mental foramen which is below the trigonid of M2 in all north of Pine Kidge, Sioux County, Nebr. Neither European forms for which illustrations have been seen. specimen has the antemolars, but the type shows the In North American forms the foramen is below the alveoli of the posterior two and part of a third (Mawby, talonid of M± except in Trimylus roperi in which it is p. 955) from which an estimate of the crowding between below the trigonid of M^ as in the European forms. the incisor and MI can be made. The Nebraska speci­ Progressive changes throughout the genus include in­ men shows a very light orange pigmentation on the crease in massiveness of the mandible, increase in upper half of the paraconid blade of M^ and, thus, bulbousness of the molars, increase in fusion of the indicates greater tooth pigmentation than is present in entoconid to the posterior loph of the talonid of Mi and younger species. T. compressus has been compared to M2 with ultimate loss of this cusp as a distinct feature T. aff. neumayrianus (described by Viret and Zapfe, in geologically youngest forms, and decrease in the space 1951, p. 425-426) by Mawby (1960, p. 955). Perhaps for the lower antemolars between the incisor and M^ the most significant comparison that can be made which results in reduction in size of the posterior ante- between the two is the slight crowding between the molars and (in European forms) loss of an antemolar incisor and MI relative to later species on both in both upper and lower jaws. The suggested relation­ continents. ships cannot be further evaluated without study of Specimen examined. Chicago Natural History Mu­ the European specimens. seum UM400. Trimylus compressus (Galbreath), 1953 Trimylus aff. neumayrianus (Schlosser), 1887 Figure 5 Heterosorex aff. neumayrianus (Schlosser). Viret and Zapfe, 1951, Eclogae geol. Helvetiae, v. 44, no. 2, p. 425-426, Dommna compressa Galbreath, 1953, Kansas Univ. Paleont. fig. 12. Contr., no. 13, Vertebrata, art. 4, p. 46-49, fig. 14. Heterosorex compressus (Galbreath). Mawby, 1960, Jour. Viret and Zapfe described a mandible of Trimylus Paleontology, v. 34, no. 5, p. 954-956, fig. 2. from upper Oligocene (upper Stampian) rocks of A little specialized species of Trimylus with molars Kickenbach, Switzerland. Their illustration shows distinctly more bulbous and cingula more inflated than three antemolar alveoli that are not at all crowded in Domnina but less so than in other North American between the incisor and Mlf The posteriormost alveolus species of Trimylus; entoconid prominent on M^ con­ is long anteroposteriorly and, as suggested by Wilson nected to posterior lophid (or hypolophid) of talonid (1960, p. 31), may have held the roots of two antemolars, and separated from metaconid; lower antemolars not which would bring the antemolar count to four. greatly crowded between incisor and MI although more Domnina gradata and all North American species of so than in D. gradata; mental foramen below talonid Trimylus in which the premolar count is reasonably of MI. Middle Oligocene (Orellan) of Colorado and well established have four antemolars behind the incisor. middle or late Oligocene of Nebraska. Insofar as I am aware, however, this is the only Only two specimens are known, both right mandibles. Trimylus of Europe that may have had four antemolars. The type, University of Kansas 8154, has been described Wilson's suggestion that there were roots of two ante- by Galbreath (1953, p. 46-49) and by Mawby (1960, p. molars in the last alveolus of T. aff. neumayrianus is 954-955, text fig. 2). The referred specimen is Chicago supported by the strong anteroposterior compression of

FIGURE 5. Trimylus compressus (Galbreath), CNHM UM400. Right mandible and dentition. 12 SUBFAMILIES AND GENERA OF THE SORICIDAE the roots of all heterosoricines except D. thompsoni The masseteric fossa (Schlosser, 1887, plate 2, fig. 50) that the elongate alveolus should have held just one root seems to be larger, deeper, and more extensive than in seems contradictory to a very well established character any other shrew, including the Arikareean and older of the subfamily. species from North America, but the fossa is not pre­ Trimylus aff. neumayrianus is also characterized by a served in the earlier European specimens. The mental very posterior mental foramen, a characteristic of all foramen is below the trigonid of M2 as in all Old World European species of the genus but not of any American forms. Early Miocene (Burdigalian) from Germany species except T. roperi. In size T. aff. neumayrianus is (Weisenau and Wintershof-West) and possibly middle comparable to T. compressus but it has a stouter hori­ dligooene (Rupelian) as "/Sotrex" kinkelini Zindorf, zontal ramus. Viret and Zapfe's illustration seems to 1901. No specimen seen. show low teeth on the cutting edge of the lower incisor, Trimylus dakotensis Repenning, n. sp. a characteristic of earlier European species. Figure 6 This specimen was not seen, and the only illustration Heterosorex. Macdonald, 1951, Soc. Vertebrate Paleontology and description seen are those of Viret and Zapfe. [Guide Book], 5th Field Conf., p. 72. Trimylus neumayrianus (Schlosser), 1887 Heterosorex. Mawby, 1960, Jour. Paleontology, v. 34, no. 5, Sorex neumayrianus Schlosser, 1887, Beitr. Palaontologie Geol- p. 951, 952, 955, 956. ogie Osterreich-Ungarns, v. 6, p. 122, pi. 2, figs. 46-50. Holotype. University of California Museum of Heterosorex neumayrianus (Schlosser). Gaillard, 1915, Soc. Paleontology 37270; a nearly complete left mandible linne'ene Lyon Annales, v. 62, p. 83. lacking the angular process and the antemolars. Col­ Heterosorex neumayrianus (Schlosser). Viret and Zapfe, 1951, Eclogae geol. Helvetiae, v. 44, no. 2, p. 425. lected by R. A. Stirton, 1934. Type locality and age. University of California Mu­ Viret and Zapfe have compared their Stampian seum of Paleontology loc. V 3417, Flint Hill fauna, Trimylus with Schlosser's description of his Aquitanian Bennett County, S. Dak. Early Miocene (Arikareean). species as well as possible from the published record. Diagnosis. A small species of Trimylus with mental The lower antemolars, as well as their alveoli, according foramen below talonid of M\\ MI has entoconid flush to Schlosser (1887, p. 122), are unknown. To judge against but not confluent with hypolophid, labial from his figure 60 of plate 2, however, the proximal cingulum heavy and inflated but absent anterior to termination of enamel on the labial side of the lower protoconid; M3 has inflated cingulum and tooth unre­ incisor seems to be very nearly below the anterior edge duced in contrast to late Miocene species; all cheek of MI, a fact indicating greater anteroposterior com­ teeth moderately bulbous and low crowned. pression of the antemolars than in Trimylus aff. neumayrianus from Rickenbach. The incisor to MI Description. T. dakotensis is comparable in size to relationship seems to be very close to that in the T. aff. neumayrianus and T. compressus; it is smaller Arikareean specimen from South Dakota, and there is than T. maiobyi, n. sp., and slightly smaller than the less anteroposterior compression of the antemolars than smallest specimens of T. roperi described by Wilson in the late Arikareean Trimylus roperi from Colorado. (1960, p. 32). Although the antemolars and that part The incisor bears two prominent cusps along its cutting of the mandible containing their alveoli are destroyed, the distance from the incisor to MI would indicate con-

FIGURE 6. Trimylus dakotensis Repenning, n. sp. Left mandible and dentition. DIAGNOSES AND CONTENTS OF SUBFAMILIES 13 siderably greater anteroposterior compression of the the late Miocene T. sansaniensis and T. mawbyi, n. sp. antemolars than is evident in either T. aff. neumay- All are strikingly like Domnina, and all have bifid upper rianus or T. compressus. T, dakotensis differs from T. incisors similar to the living genus Neomys. compressus in greater inflation of the labial cingula and These specimens were not seen but are well illustrated in greater bulbousness of the cheek teeth. These same and described by Wilson. characters are further accentuated in the late Miocene Trimylus mawbyi Repenning, n. sp. species to be described in the following pages. T. dako­ Figure 7 tensis differs only slightly, however, from the nearly contemporaneous T. roperi. The chief differences Heterosorex sp. Mawby, 1960, Jour. Paleontology, v. 34, no. 5, p. 950-956, text fig. 1. appear to be the somewhat smaller size, the absence of a labial cingulum anterior of the protoconid of M\, Type. University of Oregon 10486, right mandible slightly less massive mandible below the molars, and, with incisor, antemolars 1 and 4, M^z, and part of the perhaps most importantly, the anterior position of the ascending ramus. Collected by F. B. Van Houten in mental foramen in T. dakotensis. Measurements of the about 1955. type specimen have been given by Mawby (1960, table Type locality and age. Five miles southwest of the 1, column e). south end of Guano Lake and about 24 miles southwest The coronoid and articular processes of the mandible of Beatty Butte, Lake County, Oreg. The exact locality of T. dakotensis are very well preserved. In typical is not known to me; however, there are many outcrops heterosoricine fashion the lower articular facet is offset in the general area that could represent the locality, and lingually so that no part of it lies beneath the upper all are of the same strata. The referred material comes facet. Although similar in pattern to Domnina thomp- from these strata at localities from within 10 miles to soni, the articular facets are separated and this pattern about 40 miles from the type locality. matches all known species of Trimylus for which the Late Miocene (Barstovian). The associated fauna is articular structure is known. virtually identical to that from the Mascall Formation. The type of Trimylus dakotensis appears to be the Referred material. U.S. National Museum 23096, a first specimen of this genus found in North America left mandible with incisor, alveoli for four antemolars, (there is no record of the date of discovery of CNHM and MI and M2, U.S. National Museum 23097, a right UM400). Certainly it was the first Trimylus to be M1 ; both from U.S. Geological Survey loc. M1041, recognized from North America and was described in Beatty Butte fauna. Also five lower incisors, three manuscript many years ago by Joseph T. Gregory. upper incisors, and a mandible fragment with M\ from Unfortunately, the manuscript was never published and the following localities: Beatty Butte (loc. M1041), thanks are here given to Dr. Gregory for permission to Harney County; Guano Eanch (loc. M1042), Lake describe the specimen in this report. County, 15 miles southwest of Beatty Butte; and Snyder Creek (loc. M1043), Lake County, which is 37 miles Trimylus roperi (Wilson), 1960 west-northwest of Beatty Butte. All referred speci­ Heterosorex roperi Wilson, 1960, Kansas Univ. Paleont. Contr., mens were associated with Barstovian mammals. Vertebrata, art. 7, p. 27-33, figs. 14-22. Diagnosis. Of all known species of Trimylus, T. A large species of Trimylus with dental formula mawbyi has the greatest reduction of Ms relative to MI, 1-5-3/1-4-3; four antemolars closely appressed between the greatest bulbousness of MI, the greatest inflation of the lower incisor and MI ; entoconid of MI distinct in the cingulum of M!, the lowest cusp height of MI, and lingual view but completely fused to hypolophid in the greatest massiveness of the mandible relative to the occlusal view; MI lower crowned than Oligocene and molar size. The metalophid joins the protolophid of early Miocene forms and higher crowned than late M\ far more labially than in any other species. Also Miocene forms from North America; mental foramen conspicuous is the fact that the stout molars are farther below or slightly behind trigonid of M2 in North to the rear, relative to the anterior edge of the ascending America a character unique to this form; P4 and M1 ramus, than in any other heterosoricine. very similar to Domnina. gradata but more massive; Description. In addition to the above characters, size large, comparable to T. sansaniensis and T. mawbyi, T. mawbyi has the mental foramen below the talonid of n. sp. (see measurements in Mawby, 1960, and Wilson, MI, a position typical of most American members of this 1960). Late early Miocene (late Arikareean) of subfamily. The entoconid of MI is a blunt cuspid Colorado. thoroughly merged with the hypolophid. The upper This is the earliest Trimylus in which the upper teeth incisor is bifid. The M1 is quite comparable to T. roperi are known (Wilson, 1963). They are also known from only more stout and bulbous. The antemolar alveoli of

240-305 O-67 3 14 SUBFAMILIES AND GENERA OF THE SORICIDAE

UO 10486

Antero-occlusal view to show antemolar alveoli USNM 23096 1-4-3

FIGURE 7. Trimylus mawbyi Repenning, n. sp., type UO 10486 and referred speci­ mens. Type: right mandible and dentition. USNM 23096: left mandible and dentition. USNM 23097: right Ml. the lower jaw, USNM 23096, show clearly that four Trimylus sansaniensis (Lartet), 1851 transversely widened or doubled roots were present Sorex sansaniensis Lartet, 1851, Annuaire dept. Gers, Auch, much as in Domnina gradata only more closely com­ p. 13. pressed between the incisor and MI. The alveoli of the Trimylus schlosseri Roger, 1885, Naturw., Ver. Schwaben u. Neuburg Ber., v. 28, p. 106-107. first and last antemolars are nearly the same size and Heterosorex delphinensis Gaillard, 1915, Soc. linne"ene Lyon much larger than the intervening two. The type speci­ Annales, v. 62, p. 14. men, however, shows that the crown of P4 was much Trimylus schlosseri Roger. Stromer, 1928, Bayer. Akad. smaller than that of the first antemolar. These alveoli Wiss. Abh., Math.-naturw. Abt, v. 32, no. 1, p. 13 ( = are nearly identical to those shown by Wilson for T. S. sansaniensis and H. delphinensis). Heterosorex sansaniensis (Lartet). Viret and Zapfe, 1951, roperi (1960, fig. 20), which he feels (p. 31) are indica­ Eclogae geol. Helvetiae, v. 44, no. 2, p. 419. tive of four antemolars behind the incisor. Measure­ Trimylus sansaniensis (Lartet). Doben-Florin, 1964, Bayer. ments of the type have been given by Mawby (1960, Akad. Wiss. Abh., Neue Folge, v. 117, p. 15. table 1, column d). A species of Trimylus with postincisor antemolars The general mandibular and dental structure of T. reduced to three in the lower jaw, of which the most mawbyi immediately gives the impression of a nut­ anterior is largest (Crusafont Pairo, 1957), and cracker type of action rather than the usual soricid tightly compressed between incisor and M^ entoconid condition which could be compared to the weak but of MI completely fused to hypolophid; mental fora­ quick action of chopsticks. Perhaps the heterosoricine men below trigonid of M^ size comparable to T. mandibular characters in general reflect a molluscan, roperi; molars not so bulbous and not placed so far crustacean, or thick elytrum coleopterous diet. In posteriorly relative to ascending ramus as in T. addition, many of the mandibular characters are very mawbyi; four upper postincisor antemolars (including similar to the living Chinese Mole- or Burrowing- P4 ) are clearly indicated in figure 11 of Viret and Shrew, Anourosorex; thus fossorial habits are implied. Zapfe (1951, p. 423). The lower incisor does not have The type of T. mawbyi was described in detail by subsidiary cusps along its cutting edge (Zapfe, 1951, Mawby (1960) with excellent illustrations. He felt p. 464, fig. 7). Later Miocene (Vindobonian) and (p. 954) that a new species was represented but pre­ ferred not to name it until other specimens were early Pliocene (early Pontian) of Europe. known. Now that these are known, the species is No specimens were seen. The Vindobonian speci­ named in his honor. mens from La Grive St. Alban are well described Specimens examined. The type and referred speci­ and illustrated by Viret and Zapfe (1951, p. 419^24, mens mentioned above, which include all known figs. 10, 11), and those from the Vindobonian of specimens. Valles-Penedes by Crusafont Pairo (1957). DIAGNOSES AND CONTENTS OF SUBFAMILIES 15

Genus INGENTISOREX Hutchison, 1966 interarticular depression (shared only with the Lim­ Genotype: Ingentisorex tumididens Hutchison, 1966 noecinae). The first species discussed are those in­ A heterosoricine shrew with bulbous dentition like cluded in noncrocidurine genera in previous publica­ TrimylMS except Mj- missing; P4 very large and larger tions but whose subfamily affinities appear to be cor- than first postincisor antemolar; lower incisor short, cidurine. These are not reassigned to crocidurine not stout at base, one very prominent cusp on its genera, however, because the present review has been cutting edge; labial cingulum of MI and M2 strong detailed enough to establish subfamily rank only, and but not greatly swollen; trigonid of MI and M2 notice­ all available specimens should be examined before ably wider than talonid; mandible slender and not generic reassignment is made. The remainder of both massive. Late Miocene (Barstovian) of Oregon. fossil and living species are discussed under their ap­ Only one species, the genotype, is known. propriate genera, in alphabetical order. Specimens examined. The type and referred speci­ "Sorex" pusilliformis Doben-Florin, 1964 mens listed by Hutchison (1966, p. 9). "Small shrew, tooth cusps unpigmented * * * M2 Subfamily CROCIDUKINAE Milne-Edwards, 1868-1874 about as large as M-L. Incisor toothed [cutting edge P4 has triangular cusp and a sulcus down posterior of lower incisor scalloped], first scallop located far face of cusp, slight or no overhang of labial cingulum from the tooth base., Two antemolars in the lower over root and alveolus, one or two roots; pigmentation jaw, the posterior one [P4] two-rooted, its crown bear­ absent; dental formula 1-6-3/1-4-3 to 1-6-3/1-3-3 in ing two small secondary cuspids. MI and M2 with late Miocene and 1-3-3/1-2-3 to 1-5-3/1-3-3 in mod­ hypoconulid [entoconid separated from hypolophid], ern forms; articular facets of mandibular condyle talonid of M3 one- to two-cusped * * * Mental foramen separated but not widely, articulation surfaces usually below the second antemolar [unicuspid. Late Mio­ near to protoconid; M3 has remnant of basin remain­ cene ( ?) to Recent, Africa; early Pleistocene to Recent, ing in talonid as in Miosorex grivensis and not trench­ Europe; Pleistocene to Recent, Asia. ant as in modern crocidurines; mental foramen below During this review no characters were noted which P4 ', dental foramen usually far forward; P* and M* would serve to separate Crocidura from Suncus on have slight emargination of the posterior border quite the basis of lower dentition or structure of the man­ comparable to Miosorex grivensis, according to Mein dible. Prof. P. M. Butler (written commun. 1964) (1958, figs. 20, 21). Late Miocene (Vindobonian) or suggested that there appears to be taxonomic value possibly somewhat earlies (late Burdigalian) (Mein, in the "cingulumlike groove running along the length 1958, p. 106) of Europe. of the medial side of the first lower incisor. In all No specimens were seen. They are, however, well Suncus except S. lixus, this continues above the notch figured and described by Viret and Zapfe (1951, p. in the basal border of the tooth as a posterolingual 415-417) with a description of the upper dentition by cingulum. This does not happen in Crocidura." The Mein (1958, p. 25-26). This species is as distinct additional upper antemolar in Suncus is, of course, from other known fossil and living shrews as the diagnostic. Ellerman and Morrison-Scott (1951, p. living genera are distinct from each other. It should 65) stated that in their opinion Suncus is no more than be placed in a new crocidurine genus or with Miosorex a subgenus of Cnocidura. Throughout the Soricidae, grivensis as a separate species. This should not be dental formula is a significant character of generic done without study of the large sample now available stature, and the opinion held in this review is that (Mein stated that he has 100 mandibles and 14 maxil- Crocidura and Suncus must be retained as distinct laries from Vieux-Collonges). "Sorex" dehmi has an genera on this basis or else an immense amount of interesting combination of features, some of which synonomy is in order throughout all living shrews. seem more progressive than M. grivensis and most of It is not so easy to defend the generic distinctiveness which seem more primitive than its contemporary. of Crocidura and Praesorex by proclaiming custom to Genus CROCIDURA Wagler, 1832 be guiding dogma of soricid taxonomy. In this com­ Figure 8 parison the above diagnosis is not diagnostic. Cro­ Genotype: Sorex leucodon Hermann, 1780 cidura differs from Praesorex in smaller size (which Dental formula 1-4-3/1-2-3; lower incisor short to is variable), in having an inflated pterygoid region moderate length and curving upward, cutting edge on its skull, in having great interorbital skull breadth slightly serrate to smooth, groove along length of relative to the rostral breadth at the infraorbital medial side curves downward below notch in basal foramina, and in having a basined fossa for the in-

Medial

FIGURE 8. Crocidura russula Hermann, Belgium, author's collection 60186. Left mandible and dentition and left upper dentition. DIAGNOSES AND CONTENTS OF SUBFAMILIES 17 sertion of the superior pterygoid muscle on the man­ Crocidura taungsensis Broom, 1948 dible (all of which can easily be shown to vary with A small species lacking the anterior cusp present on size). There are two characters which may be con­ the P4 of living species, according to Meester (1955, stant differences between the genera but which have p. 274-275). Early Pleistocene, South Africa. not been checked for consistency in this review nor, as far as I am aware, in any study because of the Crocidura kornfeldi Kormos, 1934 lack of an adequate sample of Praesorex. These are More delicate than living C. suaveolens, according the more anterior position of the mental foramen and to Friant (1947, p. 264). No specimen was seen. the shorter, more velvety, fur of Crocidura. In Cro­ Early Pleistocene of Europe. cidura the mental foramen lies below the P± to below Crocidura obtusa Kretzoi, 1938 the protoconid of Mlf In the two available specimens Elongate first lower antemolar and "blunt" P4 with of Praesorex, it lies below the hypoconid of Mlm Ac­ dimensions comparable to living C. leucodon, accord­ cording to Meester (1953, p. 213), who had only two ing to Kretzoi (1938, p. 92). No specimen was seen. specimens of Praesorex but a large series of Crocidura, the difference in pelage is diagnostic. For this review, Pleistocene of Europe. therefore, Praesorex is retained as a separate genus, Crocidura zorzii Pasa, 1948 and the above characters are considered diagnostic of Citation from Sulimski (1962, p. 484). No speci­ Crocidura. men seen. Middle Pleistocene of Italy.

Crocidura sp., Butler and Hopwood, 1957 Genus DIPLOMESODON Brandt, 1853 This specimen (Butler and Hopwood, 1957, p. 2) Figure 9 seems too incomplete to defend the generic assignment. Genotype: Sorex pulchelbus Lichtenstein, 1823 Late Miocene, Kenya. Dental formula 1-3-3/1-2-3; first upper antemolar Crocidura wongi Pel, 1936 larger than second; M2 trapezoidal; lower incisor More primitive than any living crocidurine in that without or with single accessory cuspule on cutting the talonid of M3 is not reduced to a single trenchant edge; P4 slightly larger than preceding antemolar, cusp although it is practically so, according to Pei crocidurine posterolingual crest very weak; postero- (1936, p. 22). No specimen was seen. It is well labial crest virtually lacking; Mv has longitudinally illustrated by Pei (figs. 10,11). Pleistocene of China. shortened talonid, entoconid present and either com-

1-3-3 Medial 1-2-3

(JAWS) FIGURE 9. Diplomesodon pulchellum (Lichtenstein), Kazakhstan, U.S.S.R., AMNH 206558. Left mandible and dentition and left upper dentition. 18 SUBFAMILIES AND GENERA OF THE SORICIDAE pletely fused with hypolophid or very close to this Genus FEROCULUS Kelaart, 1852 lophid, entoconid crest very low, labial cingulum con­ Genotype: Sorex feroculus Kelaart, 1852 tinuous from paraconid to behind hypoconid but not Dental formula 1-5-3/1-2-3; lower incisor with two inflated or enlarged; MI and P± distinctly anteroposte- lobes on cutting edges as in Sylvisorex but inclined riorly compressed to adjust to short snout (in upper forward rather than rounded; foreclaws enlarged. dentition this adjustment is made by loss of antemolars Specimens of this shrew are very scarce. Through rather than by marked anteroposterior compression of the kindness of R. W. Hayman I have been able to teeth); trigonid of M3 slightly reduced but talonid examine photographs of the specimen in the British greatly reduced to single cusp with entoconid crest Museum (Nat. History) (BM 28.10.27.6). The sim­ leading directly to metaconid or entirely lost; zygo- ilarity to Sylvisorex is striking. From the photo­ matic process of maxillary originates just behind para- graphs I am unable to see any differences that would style of M2 to opposite metastyle of M-^; mental fora­ appear significant except possibly one less cusp and men below protoconid of M± to below anterior tip of the forward inclination of the cusps on the cutting Mi and posterolabial corner of P±\ ascending ramus edge of the lower first incisor of Ferocukis. The high, vertical and narrow; internal temporal fossa posterolingual emargination of the P4 and M1 is triangular and rather small; mandibular condylar broader and deeper in Feroculus. No specimen was process and articular facets somewhat more widely seen. Recent, Ceylon. separated than in Crocidura to extreme separation for Genus MIOSOREX Kretzoi, 1959 subfamily, articular surfaces continuous (or nearly so) along lingual margin of interarticular area. Figure 10 Kecent of south-central U.S.S.R. (Aral Sea area) and Genotype: Sorex pusillus var. grivensis Deperet, 1892 early Pleistocene of South Africa. Dental formula 1-6-3/1-3-3; P± has slight over­ Specimen examined. American Museum of Natural hang of labial cingulum over root and secondary History 206558, a male Diplomesodon pulchellum from cuspule along posterolabial crest; MI with moderate west Kazakstan, U.S.S.E. cusp height reentrant between the protoconid and hypoconid emerges on labial face a short distance Diplomesodon fossorius Repenning, 1965 above cingulum; talonid of MB not as reduced as in A large species with greatly reduced M-^- and with modern crocidurines but retains slight basin enclosed extreme specialization of the mandibular articulation, by high labial and very low lingual ridge; P* with paralleling that of many soricine shrews. Type and only slight emargination of posterior border; M1 with referred specimens listed by Repenning were seen. moderate emargination; mental foramen below root Early Pleistocene, South Africa. of /Y Late Miocene (Vindobonian) of Europe.

Hypolophid Protolophid

Medial

(JAWS) FIGURE 10. Miosorex grivensis (Deperet), UC 24859 with P4 and incisor restored from Viret and Zapfe (1951). Right mandible and dentition. DIAGNOSES AND CONTENTS OF SUBFAMILIES 19 The slightly reduced and basined talonid of M3, the bilobed so that two rows of lobate cusps run from slight emargination of P4 and M1, the retention of nipping part of incisor back to P4 ; P4 with accessory six upper antemolars, and possibly the lesser height cuspule on both posterolingual and posterolabial of crown of M^ are all that separate this crocidurine crests; P% minute and on the lingual side between the shrew from related living genera. two major lower antemolars; early Pleistocene to Miosorex grivensis (Deperet), 1892 Recent, Africa. Sorex pusiltus, var. grivensis Deperet, 1892, Mus. d'histoire Ellerinan, Morrison-Scott, and Hayman (1953, p. nat. Lyon Archives, v. 5, mem. 2, p. 48. 19) stated that the Ps is so vestigial as to be very Sorex (s.l.) grivensis (Deperet). Viret and Zapfe, 1951, rarely present and so considered Myosorex a subgenus Eclogae geol. Helvetiae, v. 44, no. 2, p. 411-412, figs. 1-3. Miosorex grivensis (Deperet). Kretzoi, 1959, Vertebrata hun- of Suncus. Eleven specimens of M. varius from Cape garica, v. 1, no. 2, p. 248, footnote 1. Province, South Africa (three localities), are in the This is the only species assigned to the genus. Museum of Vertebrate Zoology, University of Cal­ Specimens examined. University of California ifornia, Berkeley. Of these 11 specimens, 9 have the Museum of Paleontology 42672, a right mandible with vestigial P3 on both mandibles, 1 (UCMVZ 117080, M2 and Mz, and University of Colorado 24859, a right a male) has it on one mandible only, and 1 (UCMVZ mandible with J/i to Mz from La Grive St. Alban. 117085, a very old male) has the Ps missing on both More complete material is well figured and described mandibles although the upper antemolars are as de­ by Viret and Zapfe. scribed above. I am inclined to feel that the tendency Genus MYOSOREX Gray, 1838 to have three postincisor lower antemolars is very Figure 11 strong in this species and that this tendency, com­ Genotype: Sorex varius Smuts, 1832 bined with the unique structure of the upper ante- Dental formula 1-5-3/1-3-3; lower incisor short molars, is of generic importance and comparable in and curving upward, cutting edge smooth or has two magnitude to differences found between other genera. low cuspulids; first upper antemolar larger than fol­ Specimens examined. Myosorex warms, 11 speci­ lowing ones except P4, second very small, third nearly mens in the University of California Museum of as large as first, and fourth minute (lost in Myosorex Vertebrate Zoology. Also, early and late Pleistocene polli Balsac and LaMotte, 1956, p. 155-157), all bi- specimens of Myosorex in the University of Colorado cusped with cusps forming parallel longitudinal Museum from Cave of Hearths and Limeworks, blades; posterior lobe of falciform upper incisor also Makapansgat, Transvaal, South Africa.

Medial

(JAWS)

FIGURE 11. Myosorex varius (Smuts), Cape Province, Union of South Africa, UCMVZ 17086. Right mandible and dentition and left upper dentition. 20 SUBFAMILIES AND GENERA OF THE SORICIDAE Myosorex robinsoni Meester, 1955 a crocidurine shrew. The structure of the mandibular According to Meester (1955, p. 272-273), this species condyle has not been described. No specimen was lacks subsidiary cuspules on the posterolabial and seen. posterolingual crests of P±. Several specimens were Genus PRAESOREX Thomas, 1913 examined. Early Pleistocene, South Africa. Figure 12 Genus PARACEOCIDURA Balsac, 1956 Genotype: Orocidura goliath Thomas, 1906 Genotype: Paracrocidura schoutedeni Balsac, 1956 Dental formula 1-4-3/1-2-3; largest known shrew; Dental formula 1-4-3/1-2-3; upper incisor with teeth as in Orocidura; pterygoid region of skull not expansion of anterointernal border producing chisel- inflated; interorbital breadth not great relative to like cutting edge rather than point as in most shrews; rostral breadth at infraorbital foramina; mandibular P4, If1, and M2 with squared posterior borders lack­ insertion of superior pterygoid muscle not basined and ing any emargination; first lower postincisor ante- pterygoid spicule developed; mental foramen below molar very large and longitudinally bicuspid. Kecent, hypoconid of J/i; fur long, coarse, with prominent Africa. guard hairs. Recent of equatorial Afria. Relative to soricine shrews, most crocidurine shrews This ermine- to weasel-sized genus is the largest tend to have enlarged and elongate first lower post- known shrew, living or extinct. It is not, however, incisor antemolars. The elongation of this tooth in greatly larger than the giant species Sunous caeruleus. Paracrocidura exceeds any shrew examined during Except for size it is hardly more than a species of this review, however, and greatly exceeds the most Crocidura, and some species assigned to Orocidktra are closely comparable genera, Sylvisorex and Feroculus. nearly as large. With the exception of the position The structure of P4, M1, and M2 is unique for the of the mental foramen and the nature of the pelage, subfamily. The development of medial crests on the all characters listed in the above diagnosis are known tip of the upper incisors is also unique for the sub­ to vary with size in other shrews. Its similarity to family and appears to be comparable to the fissident Orocidura is discussed under that genus. Balsac and incisor found in the Heterosoricinae and in most Old LaMotte (1957, p. 29) consider Praesorex goliath, World genera of the tribe Neomyini (new name; see the only species included in this genus, to be a species p. 45). The structure of the P4, which has both pos­ of Orocidura. terolabial and posterolingual crests (with prominent Specimens examined. Harvard University Museum secondary cuspules), indicates that Paracrocidura is of Comparative Zoology 14718, a female Praesorex

FIGURE 12. Praesorex goliath (Thomas), Cameroons, MCZ 14718. Left mandible and dentition and right upper dentition. DIAGNOSES AND CONTENTS OF SUBFAMILIES 21 goliath from Cameroons; British Museum (Nat. Genus SORICELLA Doben-Florin, 1964 History) BM. 11.5.5.7, a male Praesorex goliath from Genotype: Soricella discrepant Doben-Florin, 1964 Cameroons. The following quotations are from Doben-Florin Genus SCUTISOREX Thomas, 1913 (1964, p. 48): Diagnosis. "Middle-sized soricid, teeth appear un- Genotype: Sylvisorex somereni Thomas, 1910 pigmented." Dental formula 1-5-3/1-2-3; first upper antemolar Lower dentition. "Doubly-scalloped incisor, scallops slightly larger than subequal following three; lower inclined forward, the first [most posterior] near the incisor short, curving upward at tip, with smooth tooth base. Three single-cusped antemolars, the last upper margin; in dorsal aspect articular facets of [most posterior or P4] two rooted, the middle tooth mandibular condyle much closer to being normal to smaller. Molars decreasing in size from front to back. longitudinal axis of the mandible than in other cro- MI and Mz without hypoconulid [entoconid fused to the cidurines; angular process of mandible enlarged, stout, hypolophid], M% rather large with double-cusped and attenuated posteriorly in comparison with other talonid. Labial cingulum strong." crocidurines; vertebral column with remarkable spe­ Upper dentition. "Incisor with posterior accessory cialization (Alien, 1917). No specimen was seen. It cusp, four antemolars, all well visible; the second is well illustrated by Alien. Kecent, Africa. smaller than the rest. Between the last [most posterior] Genus SOLISOREX Thomas, 1924 and P4 lies a small diastem. Ml and M2 are concavely Genotype: Soli&orex pearsoni Thomas, 1924 emarginate on their posterior side; their trigonbasins Dental formula 1-4-3/1-2-3; lower incisor with closed posteriorly for the most part by a flat enamel smooth cutting edge as in Suncus; foreclaws enlarged. border." The more significant parts of the description of Soli- Condyle. "Rather small with wide, almost straight, sorex given by Thomas (1924, p. 94) are as follows: undivided articular surface." "Articulation with lower jaws unusual in various From Doben-Florin's illustration of the posterior aspects. Most noticeable is the great size and trans­ view of the mandibular condyle (pi. 5, fig. 5: the condyle verse extent of the mandibular condyles, each of these of specimen 11702, shown on table 34, p. 53, to be a left being broader than the space which separates them mandible), the articular structure is distinctly cro- from each other across the middle line * * * Teeth cidurine and has prominent lingual confluence of the large and heavy * * * Upper 'unicuspids' only three two articular surfaces. Her figures of the occlusal view in number, of which the anterior one is almost unique of P4 are not clear enough for me to be certain that they in the family by being slightly bicuspid, in size the have the crocidurine triangular cusp. The high dental first is much the largest, the second about one-sixth formula, 1-5-3/1-3-3, the uniquely sharp cusps on the its bulk, and the third rather more than twice the cutting edge of the lower incisor, and the tall, narrow size of the second. First two molars more square coronoid process of the mandible, which is very remi­ than in Pachyura [Suncus], * * * with less salient niscent of Diplomesodon, appear to be the most distinc­ posterointernal lobes * * * Lower molars also sim­ tive characters of this genus. plified, without accessory basal lobes or ledges, the One species, Soricella discrepans Doben-Florin, has posterior (M$) remarkably reduced, consisting only been described. It is from the older Burdigalian of of a smooth lunate cuspless crest corresponding to the Wintershof-West. No specimen was seen. tuberculo-sectorial trigonid of the other shrews, the unusual heel completely aborted." Genus SUNCUS Ehrenberg, 1832 Solisorex is distinct from all known shrews. It is Figure 13 unique among the crocidurines in the loss of the labial Genotype: Suncus saoer Ehrenberg, 1832 cingulum on the lower teeth; the reduction of M3 and Dental formula 1-5-3/1-2-3; lower incisor short to the specialization of the mandibular articulation is moderate length and curving upward, cutting edge matched only by the extinct Diplomesodon f&ssorius; only very slightly serrate at most; groove along length the dental formula and other features of the dentition of medial side curves above notch in basal border and are similar to Praesorex. forms posterolingual cingulum; first upper antemolar Through the kindness of R. W. Hayman, I have larger than subequal second and third, fourth antemolar been able to examine photographs of a specimen in much smaller, all unicuspid. The genus contains the the British Museum (Nat. History), BM.31.8.14.4. smallest and the next to largest living shrews. Early No specimen was seen. Recent, Ceylon. Pleistocene to Recent, Africa; Recent, Europe and Asia.

240-305 O-67 4 22 SUBFAMILIES AND GENERA OF THE SORICIDAE

Posterior

Medial

(JAWS) FIGURE 13. Suncus caeruleus Kerr, Ceylon UCMVZ 81267. Left mandible and dentition and left upper dentition.

Ellerman and Morrison-Scott (1951, p. 65) stated posteriorly convex articular surface with labial emargi- that in their opinion this genus is no more than a sub- nation. Recent, Africa. genus of Grocidura. Although the structure of P± of this shrew closely Specimens examined. A small series including parallels that of the soricine shrews, other features, Suncus caeruleus and S. et-ruscus in the University of particularly the mandibular condyle, clearly indicate California Museum of Vertebrate Zoology. that it is a crocidurine shrew. The condyle itself is Suncus sp., Meester, 1955 also interesting in that it more closely approaches the normal mammalian single condyle than does that of any The genus has been reported from several South other living shrew examined in this review, although African localities. Several specimens have been seen. in Myosorex the articular surfaces are only slightly Early Pleistocene, South Africa. separated. Genus SURDISOREX Thomas, 1906 Specimens examined. U.S. National Museum 182593, Figure 14 a male of /S. norae, from Kenya; American Museum of Genotype: Surdisorex norae Thomas, 1906 Natural History 187262, a female from Kenya. Dental formula 1-4^3/1-2-3; upper antemolars as in Genus SYLYISOBEX Thomas, 1905 Crocidura except P* distinctly more elongate (also seen Figure 15 in Myosorex blarina zinki Balsac and LaMotte, 1956, Genotype: Crocidura morio Gray, 1862 fig. 2); lower antemolars elongate; first lower ante- Dental formula 1-5-3/1-2-3; lower incisor long and molar with two-cusped crest; P4 two rooted with acces­ sory cuspule on both posterolabial and posterolingual projecting straight anteriorly (except S. megalura, see crests, posterolingual crest incomplete behind accessory Heim de Balsac and LaMotte, 1957, fig. 10c), cutting cuspule, posterolabial crest normal for subfamily to edge serrate with three prominent notches separating accessory cuspule but behind this cuspule crest swings rounded cuspate lobes comparable to those of many lingually and merges with posterior cingulum at mid­ species of Sorex; first upper antemolar behind incisor point of tooth or even as far lingually as posterolingual larger than equal-sized second, third, and fourth, all corner of tooth; metalophid-protolophid junction of J/i unicuspid; P4 with accessory cuspule on the posterolin­ placed remarkably close to labial side of tooth; other gual crest only. Recent, Africa. dental characters as in Grocidura; ascending ramus very Ellerman, Morrison-Scott, and Hayman (1953, p. 19) broad; articular condyle of mandible of crocidurine treated Sylvisorex as a synonym of Myosorex* and pattern but unique in closeness of upper and lower Myosorex as a subgenus of Suncus. The construction of articular surfaces and basically a single, triangular, the lower incisor alone would indicate that DIAGNOSES AND CONTENTS OF SUBFAMILIES 23

FIGURE 14 Surdisorex norae Thomas, Kenya USNM 182593. Left mandible and dentition and left upper dentition.

FIGURE 15. Sylvisorex lunaris Thomas, Uganda, CNHM 26260. Left mandible and dentition and left upper dentition. is distinct and possibly closer to Feroculus than to metaconid stands slightly behind the protoconid. All Suncus. three molars with a delicate, but distinct labial Specimen examined. Chicago Natural History Mu­ cingulum * * * Two antemolars, the posterior one two seum 26260, Sylvisorex lunaris, from Uganda. rooted. Mental foramen below the trigonid of MI, condyle with undivided articular surfaces, Crocidura- Crocidurinae ? Incertae Sedis like. Mandibular foramen behind the middle of the "limnoecus" micromorphus Doben-Florin, 1964 internal temporal fossa." (Doben-Florin, 1964, p. 44). "Very small shrew, teeth probably pigmented, similar In her description of the holotype, Doben-Florin to L. niobrarensis [on which no pigment is visible] * * * further stated (p. 45) that a "pronounced, somewhat M3 with a single-cusped talonid. M± and M2 with a concave chewing surface, on the three-cornered outline hypoconulid [hypolophid separated from entoconid] [of P±\ slopes from the tooth cusp [apex]." On page and very strongly developed entoconid. In the J/i the 48, she makes the following comparison: "The chewing 24 SUBFAMILIES AND GENERA OF THE SORICIDAE surface of the two antemolars are stretched out in L. Three living genera which are customarily placed in niobrarensis and run to a median, small, posterior cusp. the Crocidurinae are not included in this subfamily in In the Wintershof form, however, the chewing surface this review. The main criteria used to remove them is three cornered, somewhat concavely excavated, the from the Crocidurinae and to place them in the elevated posterior margin allows the tooth to appear Soricinae are structure of P4 and structure of the two cusped from the labial side; there is, liowever, no mandibular condyles. In addition, some have pig­ corner behind the cusp as formed for example, in the mented teeth, at least in some individuals. In one Recent 8orex alpinus" From this description, as well genus, pigmentation is only at the very tips of the cusps as from what can be seen from the single illustration and is removed in early wear; in another, it is of a very of the occlusal surface of P4 (pi. 4, fig. 9a), the P4 of light color so that it is sometimes unnoticed. One genus "Z." micromorpJius appears to be crocidurine rather is unpigmented. In the past few decades many people than limnoecine. have noticed the fact that some crocidurines have pig­ Although difficult to judge from figure 9a of plate 4, mented teeth, and there have been two distinct reac­ the distance between the protoconid and metaconid of tions: One is that the subfamily criteria break down MI appears relatively much greater than in any known and hence the subfamily classification is not real; the limnoecine shrew. other is that the shrews in question have pigmented Doben-Florin's statement about possible pigmenta­ teeth and are not crocidurine by definition. tion and the posterior position of the mental foramen The first reaction is typified by Ellerman and suggest that the species is not crocidurine, however. It Morrison-Scott (1951, p. 41) who made note of the seems, therefore, that no conclusive subfamily assign­ inconsistencies in pigmentation and abandoned sub­ ment can be made from the published description. The family classification without further investigation. The species is here questionably referred to the Crocidurinae second reaction is perhaps best typified by Bate (1945, because of the form of the P4. The objections to this p. 761), who removed those shrews with pigmented referral may be reduced by noting that Doben-Florin teeth from the Crocidurinae but also cited several other was not certain that the teeth were pigmented. skull and skeletal differences that suggest closer rela­ "Z." micromorphu-s is from the older Burdigalian tionships to the Soricinae. fauna of Wintershof-West. No specimen was seen. Genera which have been included in the Crocidurinae (for example, by Simpson, 1945, p. 51-52) but are "Sorex" collongensis Mein, 1958 included in the Soricinae in this review are: To judge from Mein's figure 22, this shrew has a Anourosorex; Chimarrogale; Nectogale. crocidurine P4. Its teeth have no pigmentation, and the In addition, North American fossil shrews generally mental foramen is below P4, which also suggest the assigned to the genus Limnoecus are removed from the crocidurine shrews, in late Miocene or more recent time Crocidurinae and placed in a new subfamily, the at least. Mein (1958, p. 26) stated that the species also Limnoecinae, on the basis of the structure of P4, pig­ has a mandibular articulation that is enormous for the mentation, structure of Ms, and position of the mental size of the mandible but does not figure the articulation foramen; the mandibular articulation of the Lim­ or further describe it. His description of the internal noecinae appears identical to that of the Crocidurinae. temporal fossa as "semicircular" suggests the Soricinae. Subfamily LIMNOECINAE Repenning, new name Based upon the criteria of the present review, this P4 with little or no lingual segment or posterolingual shrew is only questionably placed in the Crocidurinae. ridge of primitive triangular cusp, labial segment of According to Mein, it has an anteriorly placed dental primitive cusp and its posterolabial crest form a foramen comparable to "$." dehmi and four lower medially shifted longitudinal ridge, labial shearing antemolars. No specimen was seen. Late middle or blade (as in Soricinae) not developed, posterolingual late Miocene of France. basin (as in Soricinae) little or not developed, single Crocidura?, Mein, 1958 root, slight overhang by labial cingulum over root and alveolus but never as extreme as in most soricines; MI Only two maxillary fragments are reported (Mein with protoconid and metaconid as close or closer 1958, p. 27). According to Mein, they resemble "/Sorex" together in unworn teeth than in any other known dehmi from the same fauna but have fewer antemolars. shrew, entoconid strong, entoconid crest low; MS lacking Based upon the criteria considered in the present report, double-cusped talonid in earliest form and an appre­ these specimens are not assignable to a subfamily. No ciable talonid basin lost through emphasis of metalophid specimens were seen. Late middle to early late Miocene, by late Miocene time, becoming very nearly a simple France. trenchant heel like that in modern crocidurine shrews DIAGNOSES AND CONTENTS OF SUBFAMILIES 25 and more trenchant than in late Miocene crocidurines; crests; MI with protoconid and metaconid distinctly upper teeth unknown except for incisor and Ma - pig- close together, entoconid strong and detached from mented heavily in earliest form but only at tips of cusps hypolophid but connected to metaconid by very low in later forms; lower dental formula 1-3-3 to 1-2 ?-3; entoconid crest, reentrant between protoconid and articular facets of mandibular condyle as in Cro- hypoconid emerges on labial face of tooth just above cidurinae, internal temporal fossa present; mental fora­ cingulum, labial cingulum very subdued to absent below men below trigonid of MI (more posterior than in most protoconid of MI but strong both anteriorly and crocidurine shrews of comparable age). Early Miocene posteriorly from this point; M3 with reduced talonid (Arikareean) to middle Pliocene (Hemphillian),3 but still retaining minute basin and low entoconid North America. crest; mental foramen below trigonid of MI ; pigmenta­ This is a small but nevertheless distinct subfamily. It tion extensive, comparable to Sorex; lower dental for­ is rather well known in the Miocene and early Pliocene mula 1-3-3 or possibly 1-4-3, in which case PS was faunas of western United States and is represented from vestigal and represented in type mandible by a minute faunas of both ages in California (Stirton, 1930; James, alveolus on anterior margin of alveolus of P4. 1963, p. 31-47), from the Miocene of Colorado (Wilson, Etymology. Latin: Angustidens, narrow tooth; 1960, p. 33-35), Wyoming (Keed, 1960, p. 6-7), and refers to the closeness of the protoconid to the metaconid. Nebraska (Macdonald, 1947), and from the early Angustidens vireti (Wilson), 1960 Pliocene of Nevada (J. E. Mawby, oral commun., 1963) Figure 16 and middle Pliocene, Oregon. Sorex vireti Wilson, 1960, Kansas Univ. Paleont. Contr., Ver- Genus ANGUSTIDENS Repenning, n. gen. tebrata, art. 7, p. 33-35 and 37-38. Genotype: /Sorex vireti Wilson, 1960 Limnoecus vireti (Wilson). James, 1963, California Univ. Diagnosis. P± triangular in occlusal view and pub. Geol. Sci., v. 45, p. 36. basically crocidurine in pattern except strong postero- A fairly large shrew slightly smaller than living labial crest meets (but does not merge with) cingulum Blarina brevicauda. Size, location of the mental fora­ distinctly medial of posterolabial corner and postero- men in a depression that leads anterodorsally to the lingual crest reduced and detached from apex of pri­ antemolar region, location of the dental foramen about mary cusp, sulcus on posterior face of cusp between on midline of and well below internal temporal fossa, and weakness or absence of the labial cingulum of M\ 3 A right mandible of Limnoecus sp. cf. L. tricuspis, bearing un­ below protoconid may prove to be of specific importance worn Mi to Ms, was collected by the author from the Hemphillian fauna at Rome, Oreg. This represents the only known middle Plio­ when other shrews are found that belong to this genus. cene record. Early Miocene (Arikareean) of northeastern Colorado.

KU 10037

Medial

FIGURE 16. Angustidens vireti (Wilson), type KU 10037 and referred specimens KU 10044 and 10304. Right mandible and lower dentition. 26 SUBFAMILIES AND GENERA OF THE SORICIDAE James (1963, p. 39) placed this species in the genus Limnoecus tricuspis Stirton, 1930 Limnoecus. The relationship is clear but the differences Figure 17 between Limnoecus of the late Miocene and Pliocene and this species seems to be greater than comparable Medium-sized shrew comparable in size to living differences between living shrew genera, as between /Sorex araneus; teeth more slender in occlusal view Sorex and Microsorex. Wilson's species, therefore, is and more delicate than L. niobrarensis ; P* retaining placed in a new genus. In the form of P* and M$, slight remnant of posterolingual crest; Ms retaining Angustidens vireti is exactly intermediate between the slight talonid basin and entoconid crest with trenchant condition believed to have been present in a hypothetical metalophid curving posterolingually to merge with soricid progenitor (typified in these two teeth, at least, hypoconid; lower dental formula 1-2?-3. Late Mio­ by Domnina) and Limnoecus. cene (Barstovian), early Pliocene (Clarendonian), Specimens examined. The type, University of California, and middle Pliocene (Hemphillian), Kansas Museum of Natural History 10037, and para- Oregon. types listed by Wilson (1960, p. 33). This is the genotypic species. In the type descrip­ Genus LIMNOECUS Stirton. 1930 tion Stirton (1930, p. 218) considered the lack of a Genotype: Limnoecus tricuspis Stirton, 1930 metaconid on M3 to be a diagnostic feature of P* with posterolabial crest of primitive triangular Limnoecus. James (1963, p. 31-47) reprepared and /*4 forming medial crest which hooks lingually where restudied the type and concluded that the metaconid it merges with posterior cingulum, posterolingual crest had been broken off of the specimen. He also described nearly or entirely lost; reentrant valley between pro- new material from the early Pliocene. Known mate­ toconid and hypoconid of MI emerges on labial face rial is not adequate to rule out the possibility of a well above cingulum; talonid of J/3 reduced to vestigial P^ being present. trenchant heel formed by posteriorly directed metalo- Specimens examined. The type, University of phid and hypoconid, talonid basin virtually lost California Museum of Paleontology 31047; topotypic although faint trace of entoconid crest (not continu­ material in the University of California at Riverside; ous with hypoconid) usually present on lingual edge and referred specimen, University of California Mu­ of talonid in unworn teeth; pigment reduced to tips seum of Paleontology 55515, from Cuyama Valley, of cusps or at most to upper half of paraconid blade Calif. Uncataloged specimen from USGS vertebrate of Mi] lower dental formula 1-3-3 to 1-2 ?-3. loc. M1078, Rome, Oreg.

Anterior view of M,\ moderate wear

Medial

FIGURE 17. Limnoecus tricuspis Stirton, type, UCMP 31047. Right mandible and lower dentition. DIAGNOSES AND CONTENTS OF SUBFAMILIES 27

Limnoecus niobrarensis Macdonald, 1947 Specimens examined. The type and paratype, Uni­ Figure 18 versity of California Museum of Paleontology 36171 Medium-sized shrew with teeth broader in occlusal and 36172. view, bearing stouter cusps, than those of L. tricuspis; Subfamily SOBJCINAE Fischer von Waldheim, 1817 PS minute, vestigial, and largely overlapped by P4 ; P4 with no lingual segment of primitive triangular P4 nearly square in occlusal outline with no trace of cusp, labial segment of primitive cusp and its postero- posterolingual crest; Ms with metalophid directed labial crest retained as labial shearing blade, in more posteriorly to hypolophid, very slight remnant of advanced forms this blade is continuous with elevated entoconid crest but no trace of talonid basin; mandi­ posterior cingulum forming L-shaped crest enclosing ble stouter than in L. tricuspis; lower dental formula posterolingual basin (which has evolved from posterior 1-3-3. Late Miocene (Barstovian), Nebraska. sulcus between crests of primitive triangular cusp of The differences between L. tricuspis and L. niobra­ P4 ), posterolingual crest greatly reduced or absent rensis are slight but are of the same sort and of same (more usual in advanced forms), cingulum overhangs magnitude as are observable in the mandibles of living root and lateral surface of mandible below postero- species of Sorex, for example, between Sorex cinereus labial corner of P4 in most genera more than in con­ and Sorex merriami. Based on this analogy with temporary forms of other subfamilies, single rooted living species, these differences seem to be of specific except in atypical Miocene form; molars variable; pig­ value. Many of the specific characters given by mentation usually present, sometimes extreme, but Macdonald (1947, p. 124-125) were based upon fea­ absent in some genera of one tribe; dental formula tures that James (1963, p. 32-33, tables 6, 7) pointed 1-6-3/1-4-3 in early Miocene forms (higher than in out resulted from damage of the type of L. tricuspis Domnina] and 1-6-3/1-2-3 to 1-3-3/1-2-3 in modern not known to Macdonald. James further concluded forms (1-4-2/1-2-2 in Amblycoptus] ; mental foramen that the differences in tooth proportions between the in many genera farther back than in contemporary two species could represent those of sex rather than crocidurines and limnoecines; articular facets of species and synonymizes L. niobrarensis with L. tri­ condyles little separated (or even joined in earlier cuspis. The possibility that the differences are sexual forms) to extremely separated in most living genera, rather than specific has not seemed great enough in facets are joined along labial 'side of condyle, rather the present review to abandon Macdonald's species, than along lingual side as in crocidurine-limnoecine and it is here retained until a larger sample of both pattern, with lingual emargination of interarticular species can shed more light on their variation. area; internal temporal fossa present. Late Oligocene

Anterior view t (paratype) slight wear

Medial

FIGURE 18. Limnoecus niobrarensis Macdonald, type and MI of paratype, UCMP 36171 and 36172. Right mandible and lower dentition. 28 SUBFAMILIES AND GENERA OF THE SORICIDAE (Stampian) to Recent, Europe; early Miocene (Ari- this the lower articular facet, on which the mandible kareean) to Recent, North America; early(?) Pliocene is hinged, is moved forward to shorten the lever arm to Recent, Asia; Recent, South America; not known of the horizontal ramus. The temporal muscle is also from Africa or peninsular India. Romer (1928, p. strengthened, and its area of origin is spread upward 100) stated that Sorex vulgaris (= Sorex araneus] and forward over the cranium. Its greater and more now lives in Mediterranean Algeria but is not known transverse force increases the tendency for the poste­ in the late Pleistocene deposits of that area; however, rior part of the mandible to be pulled into the side of this is the only publication I have found that mentions the braincase when biting. Consequently, the upper soricines from Africa. articular facets, the single function of which is to A diagnosis of this subfamily is difficult to outline counteract the transverse force of the temporal muscle, briefly because of great evolutionary change and is changed in orientation in order to remain approxi­ because of remarkable differences in contemporaneous mately normal to the average force resultant of the genera of the subfamily, particularly later in the his­ temporal muscle. As the upper articulation does not tory of the subfamily. Consideration of the subfamily function as a hinge but rather slides back and forth as a whole and throughout its history leaves only two over a strong articular boss on the cranium, this does criteria which are entirely definitive: the structure of not impair the hinge action of the mandible which is /*4 and of the articular condyle or condyles. In some controlled by the position of the lower condylar facet. genera even these are difficult to recognize but never Primitively the condyloid process supporting the has there been any question of subfamily recognition upper articular facet is a thin plate of bone alined when more variable criteria have been used in support with the plane of the ascending and horizontal rami, of the characters of P4 and of the condyles. It should as in Sorex, As the force of the temporal muscle in­ be emphasized, however, that other characters vary creases, the plane of the condyloid process curves away with the stage of evolution and with separate lineages from the plane of the horizontal ramus and toward the within the subfamily Soricinae. braincase in order to parallel the force resultant of Although the number of living species of crocidu- the temporal muscle to attain greatest possible strength rine shrews about equals that of the soricine shrews, in resisting the tendency of this muscle to pull the there are twenty-nine soricine genera recognized in mandible into the braincase. This is actually accom­ this review and only ten crocidurine genera. To re­ plished by an outward warping of only the ascending view this large subfamily with some system compara­ ramus rather than of the entire proximal part of the ble in detail to that used for the other subfamilies, the mandible. Soricinae have been broken down into three tribes. The most readily recognized changes in the proxi­ These subdivisions are based on the structure of the mal part of the mandible resulting from a strengthen­ mandibular articulation and associated musculature ing of the bite area are, therefore, separation of the which appears to remain constant within the tribes articular facets by the forward placement of the lower despite the presence of genera otherwise very diverse facet and outward deflection of the ascending ramus in form. The subfamily is readily classified into two to allow the inward curving of that part of the con­ tribes on the basis of the mandibular articulation. A dyloid process that supports the upper facet. third tribe is set up for those few genera (but many As these changes take place, an emargination of the species) that are of primitive structure in their articu­ lation and cannot easily be assigned to either of the lingual side of the interarticular area develops and two more specialized tribes. In varying degrees other the insertion of the superior pterygoid muscle on the features support this tribal grouping. lingual side of the condyloid process, between the two Within the tribes further subdivision, of a subtribe articular condyles, becomes basined. This condition is magnitude, is suggested by features found to be exclu­ found in Neomys and is the articular structure which sive in only part of a tribe, and these features seem characterized the tribe Neomyini (fig. 1). The action to have geographic significance, at least on a conti­ of the superior pterygoid muscle upon the mandible nental scale. These subtribes are discussed and serve is directly forward, consequently the lingual condylar as a basis for the order of discussion of genera where emargination weakens its insertion and a pterygoid such lesser subdivisions seem to be present. They are spine or boss is developed to strengthen it in many not sharply defined, however, and are not formally forms. The spine is at the dorsal margin of the supe­ set up as classification units. rior pterygoid fossa just below the upper sigmoid The modification of the mandibular articulation is notch. The tribe Neomyini is also characterized by a developed to increase the strength of the bite. To do bifid upper incisor and a prominent entoconid crest DIAGNOSES AND CONTENTS OF SUBFAMILIES 29 on J/i of most forms, as well as several other dental ticulation. Among living genera of this tribe, there and cranial features. are two distinct lineages. The first lineage, exempli­ In other shrews the insertion of the superior ptery- fied by the living Sorex, is primitive in nearly all goid muscle is strengthened by the development of a respects in addition to the articular structure and has bone plate across the lingual condylar emargination a known history dating from the late Oligocene of which also strengthens the condylar process support­ Europe and the early Miocene of North America. ing the upper condyle. Except in forms where the The second lineage, exemplified by the living Blari- bone plate is weakly formed, no pterygoid spine is nella, is advanced in most respects except in the ar­ required or developed. In posterior view the inter- ticular structure and has a known history dating from condylar area appears to lack any emargination al­ the middle Pliocene of eastern Asia, and from the late though the area filled by the bone plate behind the Pliocene of Europe. basined insertion of the superior pterygoid muscle is Genus CROCIDOSOREX Lavocat, 1951 usually depressed from the rest of the interarticular area. This is the condition found in Blarina and is Genotype: Crocidosorex piveteaui Lavocat, 1951 the articular structure that characterized the tribe Dental formula 1-6-3/1-4-3, including a minute Blarinini (fig. 1). The tribe Blarinini is also char­ vestigial P3 ; P4 with soricine labial shearing blade acterized by an uncleft upper incisor and by no ento- which bears a secondary cuspule but (in contrast to conid crest on the M^ as well as several other dental Sorex) does not merge with posterior cingulum and and cranial features. does not enclose posterolingual basin, overhang by The articular structure is unmodified in the tribe cingulum at posterolabial corner of tooth very slight Soricini (fig. 1). The articular facets are not far and single root not covered; M\ with cingulum only apart and the condylar process, the coronoid process, along paraconid blade and with reentrant valley be­ and the horizontal ramus all lie in the same plane. tween protoconid and hypoconid emerging on labial An entoconid crest is present on J/1? and a bifid upper face of tooth well above enamel base in marked con­ incisor, generally a feature of a variant individual, trast to contemporary Dormiina,; Ms as in /Sorex and may or may not be present. Donwina with unreduced talonid enclosed lingually by entoconid crest; mental foramen below P4 ; man­ Tribe SORICINI Fischer von Waldheim, 1817 dibular condyle with facets only slightly more sepa­ Articular facets of mandibular condyle continuous rated than in Domnina but lower facet placed more or only slightly separated, both facets approximately ventrally than in heterosoricines and not offset lin­ normal to longitudinal axis of cranium; when facets gually beyond plane of mandible at lower sigmoid are separate, interarticular area broad with no con­ notch, condyle with soricine pattern of labial connec­ spicuous lingual emargination seen in posterior view, tion between upper and lower facets and lingual lower facet not offset lingually beyond plane of mandi­ emargination between facets rather than opposite as ble at lower sigmoid notch or shifted appreciably for­ in crocidurines and limnoecines (see pi. 2, fig. 73, in ward (fig. 1) ; superior pterygoid fossa of the mandi­ Schlosser, 1887, for illustration of condyle); internal ble a shallow pit; external temporal fossa of mandible temporal fossa deep and narrowly triangular; masse- extends no lower on ascending ramus than upper sig­ teric fossa absent; P4 and M1 with moderate emargi­ moid notch; internal temporal fossa of mandible large nation of posterior basal outline between hypoconal and triangular; coronoid process not deflected labially flange and metastyle. but in plane of horizontal ramus and of condyloid Crocidosorex piveteaui Lavocat, 1951 process; zygomatic process of maxillary originates to Medium-sized shrew comparable in size to living rear of M2 ; M2 rectangular in occlusal view; M^ with Sorex ara/wus; M\ with prominent entoconid con­ strong entoconid crest; Mz unreduced or with only nected to metaconid by relatively low entoconid crest; talonid reduced, roots separate by distance equal to or upper teeth unknown; teeth unpigmented according greater than transverse diameter of anterior alveolus; to Lavocat (1951, p. 24). No specimen was seen. Late teeth pigmented. Oligocene (upper Stampian) of France according to This tribe includes four living genera; one is Sorex, Rene Lavocat (oral commun., 1963). containing by far the largest number of species of any Crocidosorex antiquus (Fomel), 1853 genus of shrew and the only truly holarctic genus now living. The tribe also contains the oldest fossil rep­ Sorex antiquus Pomel, 1853, Annales sci., litt, Indus. [Au- vergne], v. 26, p. 81-229. resentatives of the subfamily Soricinae because, by Sorex antiquus, Stehlin, 1940, Eclogae geol. Helvetiae, v. 33, definition, it has the least specialized mandibular ar­ p. 298, figs. Ib, 2b, 3, 4b, and 5b.

240-305 O-67 5 30 SUBFAMILIES AND GENERA OF THE SORICIDAE Small shrew, smaller than C. piveteaui and com­ nosis basically conforms to that here used for Cro- parable in size to living Microsorex hoyi; MI with ddosorex. Study of the specimens, however, could prominent entoconid connected to metaconid by well- well lead to the conclusion that the two genera are developed entoconid crest; teeth pigmented; upper distinct. teeth and mandible as described for genus. No speci­ Crocidosorex is structurally suitable as an ancestral men seen. Excellent material is well illustrated and form for all later soricine shrews. It. differs little described by Stehlin (1940, p. 298-301). Early Mio­ from living Sorex. the differences being those cited in cene (Aquitanian) of France (Montaigu-le-Blin) and the generic diagnosis. Germany (Weisenau). O. piveteaui and C. antiquus are similar shrews, but Genus ANTESOREX Repenning, n. gen. the differences in pigmentation and in the entoconid Genotype: /Sorex compressus Wilson, 1960 crest seem to warrant their retention as separate spe­ Diagnosis. P4 and mandibular condyle as in Sori- cies until such time as the specimens can be compared cinae. P4 with weak shearing blade that merges directly. Lavocat (1951, p. 23-24) makes no compari­ with posterior cingulum, shearing blade with slight son between the two. secondary cuspule, posterolingual basin not developed, ll/Sorex pusillus von Meyer" of Tobien (1939, p. 165) two roots arranged longitudinally, labial cingulum may belong to this genus, but the published description slightly overhangs anterior root and not posterior is not adequate for certain assignment. Stehlin (1940, root; MI with reentrant valley between the protoconid p. 299, footnote 4) stated his opinion that "Sorex" and metaconid emerging on labial face of tooth well pusillus and "Sorex" a/tribiguus are probably about above oingulum (much higher than in the limnoecine the same as "Sorex" antiquus. The same possibility shrew Angustidens vireti which was collected in the is also expressed by Kretzoi (1959b, p. 247); however, same deposits as the genotype of Antesorex} , well- Kretzoi also noted that "Sorex" pusillus von Meyer developed entoconid connected to metaconid by a high was preoccupied by Sorex pusillus Gmelin, and he entoconid crest, labial cingulum fairly strong along replaced the name with a new genus and species (ibid., paraconid blade but very weak and difficult to see p. 247, 248, footnote 1) Oligosorex meyeri to apply to along remainder of labial side of tooth; Ms as in /Sorex the shrew from Weisenau (without reference to other with talonid basin completely enclosed by the entoco­ shrews he would include in the genus) Kretzoi's diag­ nid crest; emargination of posterior basal outline of nosis for the genus Oligosorex is "late Oligocene and F4 about as in Crocidosorex but possibly slightly Miocene Soricids of small dimensions, retaining three greater in M1 ; teeth pigmented; mental foramen be­ unicuspids between P4 and the relatively small and low posterior root of F4 ; internal temporal fossa primitive front incisor in the lower jaw"; this diag­ strong and narrowly triangular as in Crocidosorex',

KU 10051

P4 and Afj-KU 10050 M-KU 10052

Posterior Medial

FIGURE 19. Antesorex compressus (Wilson), type KU 10050 and referred specimens. Left mandible and lower dentition and referred left upper dentition. DIAGNOSES AND CONTENTS OF SUBFAMILIES 31 masseteric fossa absent; mandibular condyle with Genus SOREX Linnaeus, 1758 facets only slightly more separated than in Domnina Figure 20 but lower facet placed more ventrally than in hetero- soricines, articular surfaces of facets continuous along Genotype: Sorex araneus Linnaeus, 1758 labial side of condyle and not along lingual side of Dental formula 1-6-3/1-2-3; lower incisor usually condyle as in limnoecines and crocidurines; number with ventral margin having little dorsal curve to­ of antemolars unknown. ward tip (all Old World species appear to have three Etymology. Latin: Ante, before, and the genus well-developed scalloped cusps on cutting edge of Sorex. lower incisor, but there are exceptions to this in New Antesorex compressus (Wilson), 1960 World species) ; MI with entoconid crest connecting Figure 19 entoconid to metaconid; Ms as in Domnina and Cro­ Sorex compressus Wilson, 1960, Kansas Univ. Paleont. Contr., cidosorex with unreduced and basined heel; P4 and Vertebrata, art. 7, p. 35-38, figs. 29-31. M1 with strong emargination of posterior basal out­ Small shrew comparable in size to Microsorex hoyi; line; protocone of P4 not as lingual in position as P± anteroposteriorly compressed and with constriction hypoconal flange; teeth pigmented; articular facets of around crown that may have been deformity during mandibular condyle slightly to moderately separated, growth; MI with protoconid remarkably close to the lower facet extends labially to or beyond plane of metaconid in unworn teeth. Late early Miocene (late mandible at lower sigmoid notch; interarticular area Arikareean), northeastern Colorado. of condyle not appreciably constricted in posterior The genus Antesorex from North Americans very view in most species but extremely variable and ap­ close to Cr&cidosorex from Europe, and A. compressus proaching the pattern of "other tribes in some species; is most like C. antiquus. Antesorex is slightly more mental foramen usually below protoconid of MI or advanced in the union of the labial shearing blade to anterior of this point. the posterior cingulum of F4, in the continuous (al­ though weak) cingulum along the labial side of J/i, The genus is the most generalized and, in many and in the somewhat greater emargination of the respects, the most primitive of living soricine shrews. posterior basal outline of M1. It seems less advanced No species are known that are unquestionably of early in the retention of two roots on P4, a feature that may Pliocene age, but there are a large number of fossil be an atavism. species included in the genus of later Pliocene and Specimens examined. The type, University of Pleistocene age. These have been briefly reviewed Kansas, Museum of Natural History 10050, and para- (mostly from published accounts) for conformity to types listed by Wilson (1960, p. 35). the generic diagnosis but not for synonymy or for

Medial

FIGURE 20. Sorex araneus Linnaeus, Belgium, author's collection 60182. Right mandible and lower dentition and left upper dentition. 32 SUBFAMILIES AND GENERA OF THE SORICIDAE specific diagnosis. The following list is of species Sorex kennardi Hinton, 1911. (?) Germany (Heller, 1958, p. here included in the genus, some questionably: 18-19) ; the type, from England, is late Pleistocene. No specimen was seen. Pliocene and earliest Pleistocene species Sorex laoustris (Hibbard), 1955 [Neosorex lacustris Hibbard]. Sorex dixonensis Hibbard, 1956. To judge from the published Kansas. No specimen was seen. illustrations, the low-crowned molars, the wide lower man- Sorex praealpinus Heller, 1930. Germany and Poland. No dibular condyle, the narrow interarticular area, and the specimens were seen. very strong entoconid crest on M2 are exceptional for the Sorex praearaneus Kormos, 1934. Hungary, Germany, and genus. The generic and tribal assignment is here consid­ Poland. No specimen was seen. ered doubtful. No specimen was seen. Earliest Pleistocene Sorex runtonensis Hinton, 1911. Austria, Czechoslovakia, Eng­ (late Blancan), Kansas. land, Germany, and Poland. Two specimens, author's collec­ Sorex alpinoides Kowalski, 1956. Poland. No specimen was tion 6235 A and C from Czechoslovakia were examined. seen. Sorex subaraneus Heller, 1959. Germany. No specimen was Sorex leahyi Hibbard, 1956. The lower condyle is more elon­ seen. gate lingually than in most species of Sorex. Generic and tribal assignment here considered in doubt. No specimen Species known only from the late Pleistocene was seen. Earliest Pleistocene (late Blancan), Kansas. Sorex perminutiis Kretzoi, 1959b, equals Sorex minutissimua Sorex mimitus Linnaeus, 1776, central Europe, late Pliocene Heim de Balsac, 1940, France. No specimen seen. (Sulimski, 1962, p. 461) to living. Recent specimens seen. Sorex frankstownensis Peterson, 1926. Pennsylvania. No Sorex rexroadensis Hibbard, 1953. No specimen was seen. Late Pliocene (early Blancan), Kansas. specimen was seen. Sorex sandersi Hibbard, 1956. No specimen was seen. Earliest Many living species have been reported from late Pleistocene (late Blancan), Kansas. Pleistocene deposits. Sorex subminutus Sulimski, 1962. No specimen was seen. Late Pliocene (pre-Villafranchian), Poland. Genus DREPANOSOREX Kretzoi, 1941 Sorex taylori Hibbard, 1937. No specimen was seen. Late Pliocene and middle Pleistocene (early Blancan and Irving- Genotype: Sorex (Drepanosorex] tasnadii tonian), Kansas. Kretzoi, 1941 Early Pleistocene species Diagnosis as Sorex except that the upper soricid Sorex araneoides Heller, 1930. Germany. No specimen was incisor is fissident in all known examples (Miklos seen. Kretzoi, oral commun., 1966); lower molars stout with Sorex cinereus Kerr, 1792. Kansas (Hibbard, 1944, p. 719- 720). A living species. Only Recent specimens were exam­ heavy cusps that are lower crowned than those of ined. Sorex ; teeth with pale-orange-yellow pigment rather Sorex cudahyensis Hibbard, 1944. Kansas. The type, Uni­ than dark red; mandibular condyles large with versity of Kansas Museum of Natural History 6513 was examined. greater vertical separation than in Sorex, and dorsal Sorex fallax Heller, 1936. Germany. No specimen was seen. condyle extended much farther back approximating

Posterior

FIGURE 21. Microsorex hoyi (Baird), Alaska, author's collection 6059. Left mandible and dentition and left upper dentition. DIAGNOSES AND CONTENTS OF SUBFAMILIES 33 the condition found in Blarinella, Middle Pleisto­ Soricini. In addition to the unspecialized mandibular cene of Europe. Four species have been named: articulation, they have dental characters that do not Drepanosorex margaritodon (Kormos), 1935. Middle Pleisto­ differ greatly between the Oligocene forms and the cene, Hungary. No specimen seen. living forms. Nor do the dental characters differ Drepanosorex savini (Hinton), 1911. Middle Pleistocene, Aus­ greatly from those found in the earliest members of tria, Czechoslovakia, and England. One specimen seen, au­ other subfamilies, like the heterosoricine Domnina, thor's collection 6234 from Czechoslovakia. Drepanosorex tasnadii (Kretzoi), 1941. Middle Pleistocene, the crocidurine Miosorex grivensis, and the limno- Hungary and Germany. No specimen seen. ecine Angustidens, although these earliest members of Drepanosorex pachyodon (Pasa), 1948. Original designation the other subfamilies clearly show the beginnings of Sorex pachyodon. Middle Pleistocene, Italy. No specimen dental modification characteristic of their subfamilies. seen. The remaining genera of the tribe Soricini, in con­ Genus MICROSOREX Baird, 1877 trast, have considerable dental modification from the Figure 21 primitive pattern although the mandibular articula­ Genotype: Sorex hoyi Baird, 1858 tion is still clearly primitive. Dental formula as in Sorex 1-6-3/1-2-3, but third Genus ALLTTVISOREX Hutchison, 1966 and fifth upper antemolars minute; relative to trigo- Figure 22 nid, talonid of Ms somewhat more reduced than in Sorex; because shortening of upper antemolar arcade Genotype: Alluvisorex arcadentes Hutchison, 1966 is reflected in lower antemolars, F4 somewhat broader A soricine shrew with primitive mandibular articu­ in occlusal view than in Sorex; in contrast to nearly lation comparable to that of Antesorex in outline but all species of Sorex, F4 with curving J-shaped labial with small interarticular area separating upper and crest rather than an angulate L-shaped crest, slight lower articular facets; dental formula ?/l-3-3 (in A. posterolingual basin; MI as in Sorex; mental fora­ arcadentes} to 1-6-3/1-2-3 (in A. chasseae) ; cheek men rather far to rear and nearly beneath hypoconid teeth stout and low cusped; mandible heavy; lower of MI. Only Recent specimens examined. Two fos­ incisor with three low rounded cusps on cutting edge; sil species have been named: F3 minute when present; MI with entoconid very close Microsorex pratensis Hibbard, 1944. Middle Pleistocene (Irv- to metaconid and connected to it by a high and promi­ ingtonian), Kansas. nent entoconid crest, postentoconid valley very wide, Microsorex minutus Brown, 1908. Late Pleistocene (Rancho- metalophid joins protolophid close to protoconid, labial labrean), Arkansas. cingulum prominent but not inflated; occlusal outline The five genera just discussed are the known rep­ of Ml and M2 distinctly rectangular; M3 with reduced resentatives of the unspecialized lineage of the tribe talonid consisting of longitudinal metalophid ter-

Mental foramen

LACM 4265

Less worn topotypic specimen UCR 10032

Medial

FIGURE 22. Alluvisorex chasseae (Tedford), paratype, LACM 4265 and topotype UCR 10032. Right mandible and lower dentitions. 34 SUBFAMILIES AND GENERA OF THE SORICIDAE minating in strong hypoconid, no entoconid, very low tocone to form a continuous endoloph; M2 rectangular entoconid crest, more on the lingual side of the talonid with zygomatic process of maxillary originating pos­ than on the occlusal surface and missing on one terior to M2 ; teeth pigmented and stout; mandible specimen of A. chasseae from Nevada (UCR 10237); heavy; mental foramen below hypoconid of MI ; artic­ mental foramen below metalophid of M^ upper in­ ular structure of mandible as in Blarinella; coronoid cisor not fissident; P4 and upper molars intermediate spicule is well formed and virtually as in living between Sorex and Blarinella. Late Miocene (Bar- Blarinella. Late Pliocene to middle Pleistocene of stovian) of Oregon and early Pliocene (Clarendonian) central Europe. of Nevada and California. Four species, Petenyia hungarica Kormos, Petenyia Two species have been named: suavensis Pasa (1948), Petenyia neglecta Kretzoi Alluvisorex arcadentes Hutchison, from the late Miocene of (1943), and Petenyia stehlini Kretzoi (1943) are southeastern Oregon. This species retains a minute P3. known. The genus is very close to the living Blar­ The type and referred material listed by Hutchison (1966, inella and differs chiefly in its loss of minute fifth p. 18) were seen. upper antemolar. Alluvisorex chasseae (Tedford), originally described from the Specimens examined. Central Geological Survey lower Pliocene of California (Tedford, 1961) and subse­ quently described from the lower Pliocene of Nevada (Suth- of Czechoslovakia SUtJG, OF 65566, a complete right ard, 1966). The para type, Los Angeles County Museum lower jaw, and StJtJG, OF 65565, a complete left 4265, and topotypes in the University of California, River­ lower jaw, both from the earliest Pleistocene (Villa- side, as well as the specimens from Fish Lake Valley, Nev., franchian) of southern Slovakia. in the same institution were seen. Genus PETENYIA Kormos, 1934 Genus BLARINELLA Thomas, 1911 Figure 23 Genotype: Sorex quadraticauda Milne-Edwards, 1872 Dental formula 1-6-3/1-2-3; lower incisor rela­ Genotype: Petenyia hungarica Kormos, 1934 tively shorter than typical Sorex, upward curving, Dental formula 1-5-3/1-2-3; lower incisor with two with two low scalloped cusps on upper cutting edge; low cuspules on its cutting edge; M± with prominent lower cheek teeth stout with deep cingula not inflated, entoconid crest about as in living species of Blarinella, labial cingulum strong but not inflated and continuous MI and M2 with entoconid close to metaconid and to posterior margin of tooth behind hypoconid; both cusps connected by high entoconid crest that is talonid of M3 greatly reduced and no trace of ento­ more prominent than in any other shrew; M3 with conid crest; third upper antemolar reduced, fourth trigonid unreduced in size but talonid reduced to very reduced, fifth upper antemolar absent in Petenyia single low nearly conical but slightly bladelike hypo­ (present but minute in Blarinella quadraticauda) ; P4 conid; P4 with very slight emargination of posterior and M1 with slight emargination of posterior basal basal outline and with lingual cingulum very promin­ outline; hypoconal crest of M1 and M2 united to pro- ent medial to protocone; M1 with very slight posterior

Medial

FIGURE 23. Petenyia hungarica Kormos, StJtJG, OF 65565. Left mandible and dentition. DIAGNOSES AND CONTENTS OF SUBFAMILIES 35 emargination; M2 rectangular; teeth heavily pig- kretzoii. The mandibular structures shown on Schlos­ mented as in Blarlna; zygomatic process of the maxil­ ser's illustration seem typical of the genus, as Miller lary originates posterior to M2 ', mental foramen pointed out. The talonid of Mj_ does not seem antero- beneath middle of 3/j; mandible stout; internal tem­ posteriorly shortened as in the Blarinini nor does the poral fossa tending to be oval; condylar process and mental foramen appear to be far enough back on the articular facets relatively large but otherwise as in mandible for this tribe. Middle Pliocene (Ertemte Soreso. fauna) of Mongolia. No specimen seen. One fossil and one living species have been named. Blarinella quadraticauda (Milne-Edwards), 1872 They range from the middle Pliocene to Recent in Figure 24 China. The genus Petenyia from the late Pliocene and early Pleistocene of Europe is very closely re­ This living species of Blarinella, bears considerable lated and differs chiefly in the loss of one upper ante- superficial resemblance to Blarina in its dental form­ molar. ula, reduction of upper antemolars, pigmentation, Blarinella kormosi (Schlosser), 1924 stoutness of lower teeth, depth of cingula of lower Croddura kormosi Schlosser, 1924, Palaeontologia sinica, ser. teeth, and in the stoutness of mandible. It clearly C, v. 1, no. 1, p. 5-6, pi., figs. 1, 2. differs from Blarina in its exceedingly prominent ento­ Blarinella kormosi (Schlosser). Miller, 1927, Palaeontologia conid crest and posterior continuation of the labial sinica, ser. C, v. 5, no. 2, p. 8-9. cingulum to the rear of the hypoconid on Mt and M2, From Schlosser's description and illustrations (1924, manner of reduction of the talonid of Ms, lack of p. 5-6; pi. 1, figs. 1, 2) and from Miller's discussion anteroposterior shortening of the talonid of MI and (192Y, p. 8-9), I am not sure that the very distinctive M2l development of a cingulum medial to the pro- entoconid crest of MI and single-cusped talonid of M3 toconid of /**, rectangular M2, more anterior mental are on this species. Photographs of Crocidura kormosi foramen, origin of the zygomatic process of the maxil­ on file in the U.S. National Museum, which were sent lary behind M2, and most importantly it differs from by Professor Wiman to Dr. Miller and served as a Blarina in its Sorex-like mandibular articulation and basis for Miller's revision, show that between the time features of the posterior part of the mandible and of of Schlosser's study and the time of Wiman's photo­ the cranium related to this type of articular structure. graph the specimen had lost its Ms. In addition, Specimen examined. Chicago Natural History Mu­ although the photographs are not clear in this respect, seum 3621T, a male from Szechwa,n, China. the entoconid crests on MI and M2 appear to be rather Soricini Incertae Sedis low, if present at all. The species could be a member "Sorex" stehlini Doben-Florin, 1964 of the tribe Blarinini and related to "/Sorex" kretzoii "Small Svrex, tooth cusps lightly colored yellowish Sulimski, a possibility further discussed under "S." orange. MI M3 =3.13 3.25 mm; Mz somewhat

Medial

FIGURE 24. Blarinella quadraticauda (Milne-Edwards), China, CNHM 36217. Left mandible and dentition and left upper dentition. 36 SUBFAMILIES AND GENERA OF THE SORICIDAE smaller than Mlf I [lower incisor] doubly scalloped, This shrew is intermediate between Sorex and Blar­ first scallop quite close to the tooth base. Three single- inella in almost all its characters. To judge from rooted antemolars [including P±~\ in the lower jaw, Sulimski's description, the species should be assigned the second and third very steeply placed. Molars with to a new genus rather than be used as the basis for wide cingulum and massive cusps. Ml and M2 with major enlargement of the definition of either estab­ hypoconulid [entoconid separated from hypolophid] lished genus. This genus should not be described, and a lingual metaconid placed behind the protoconid. however, without study of the specimens. No speci­ MS single-cusped in its talonid. Mental foramen beneath men was seen. to behind the protoconid of M^. Condyles with large articular surfaces." (Doben-Florin, 1964, p. 41). Genus ZELCEINA Sulimski, 1962 The articulation of the mandible is primitive and Genotype: Neomys soriculoides Sulimski, 1959 similar to that of Antesorex or Crocidosorex. Hence, Dental formula 1-5-3/1-2-3; upper incisor not bifid; the species appears to belong in the tribe Soricini. P4 and M1 appear to have moderate posterior emargi- Unlike these two earlier Miocene genera, "Sorex" nation; M2 rectangular, and zygomatic process of max­ stehlini has three quite specialized features: the talonid illary originates posterior to M2; lower incisor short of Ms is reduced to a single cusp, suggestive of the with two low scalloped cusps on cutting edge much as talonid of Petenyia, Blarwella, or some genera in the in Neomys ; MI with strong labial cingulum, but no tribe Blarinini; the mental foramen is far to the rear published information indicates presence or absence of in comparison to the other two earlier Miocene Soricini entoconid crest; Ms described as one-half smaller than genera; and the teeth as a whole are massively con­ MI, strongly reduced talonid without entoconid; teeth structed and have much heavier cingula than in these pigmented at tips; mental foramen below hypoconid two genera. These specializations, three antemolars, of ML and the more primitive mandibular articulation, rela­ The F4 is described by Sulimski (1959, p. 150; 1962, tive to that on Sorex, all indicate that the species does p. 476) as having a posteriorly projecting protocone not belong in Sorex. It would be possible to infer an the only published illustration of this tooth (1959, fig. ancestral relationship to Petenyia for this species 5C) suggests to me that the protocone is anteriorly except that Doben-Florin's illustrations do not seem placed close to the parastyle but has a prominent cingu- to indicate the presence of the prominent entoconid crest characteristic of Petenyia, and related genera. lar shelf posteromedial to the protocone as in Blari­ The absence of the entoconid crest would suggest a nella. There is no published figure of the occlusal view relationship to the largely North American tribe Blar­ of the lower teeth, and it is not possible to say whether inini and possibly an ancestral relationship to late the Ms has both talonid and trigonid reduced as in Miocene Adeloblarina through shortening of the rela­ some species of SoricuLus or has only the talonid re­ tive anteroposterior length of the talonid of M± and duced as in Petenyia^-Sulimski's (1959) figures 4a and M2 and specialization of the mandibular articulation. 4b of plate 3 seem to show that just the talonid is re­ "Sorex" stehlini is from the older Burdigalian fauna duced. The posterior part of the mandible is distinctly of Wintershof-West. No specimen was seen. more like Blarinella than Neomys except that the inter- articular area is a little too narrow for Blarmella:, it "Sorex" hibbardi Sulimski, 1962 is clearly too broad for Neomys, and in addition to the Lower incisor has the two strong scalloped cusps labial edge seems swollen in Sulimski's (1962) figure on upper cutting edge more strongly developed than 3b of text-plate 2, much as in living Blarinella. in Blarinella, and the incisor is relatively longer; cingula on lower molars strong and of uniform thick­ Definite assignment of this genus to any tribe is de­ ness rather than deepened; Ml and M2 with entoconids pendent upon the presence or absence of the entoconid well developed but without very prominent entoconid crest on MI, the nature of reduction of M%, and con­ crest of Blarinella and Petenyia; Ms with reduced sidered study of the articular structure of the mandi­ talonid lacking entoconid as in earlier species of Blar- ble. Assignment to the tribe Neomyini seems question­ wiella; coronoid process low and curved backward at able because of the lack of the bifid upper incisor in an tip; condyloid process very heavy with articular struc­ Old World form of this little specialization and be­ tures similar to Blarinella; teeth distinctly pigmented; cause of the posterior position of the mental foramen internal temporal fossa large and triangular but with and the two cusps on the lower incisor. Assignment well-developed limula; late Pliocene (lower part of to the tribe Blarinini is not supported by the rectangu­ Breccia), Poland. lar M2, the position of the zygomatic process of the DIAGNOSES AND CONTENTS OF SUBFAMILIES 37 maxillary, and the narrowness of the interarticular Genus ADELOBLARINA Repenning, n. gen. area of the mandible. Assignment to the tribe Soricini Genotype: Adeloblarina berklandi, n. sp. near the genus Petenyia is favored by the general simi­ Diagnosis. A soricine shrew with stout molars com­ larity of all features that have been described or illus­ parable to Blarina but with mandibular structures more trated but is left uncertain because of those features like those in Sorex; lower dental formula 1-2 ?-3; MI not described or illustrated. with crown having stout base, dense pigmentation, If Zelceina is related to Petenyia, in which there is talonid shortened anteroposteriorly, entoconid dis­ a strong entoconid crest and only the talonid of M3 is tinctly separated from hypolophid and relatively close reduced, it probably should be recognized as a distinct to metaconid but separated from metaconid by a pro­ genus, at least on the basis of differences in the articu­ nounced valley, hypolophid extends past the entoconid lar structure of the mandible. to the lingual cingulum, posterolabial part of labial One species has been named Z. soriculoides (Sulim- cingulum (near hypoconid) enlarged and attenuate ski) from the late Pliocene (lower part of W§ze Brec­ posteriorly as in Paracryptotis and not as in Blarina cia), Poland. No specimen was seen. so that labial embrasure between M± and M% would be less open than in Blarina; alveoli for M& are close to­ Tribe BLARININI Stirton, 1930 gether ; mandible moderately stout with depressed area Articular condyles of the mandible widely separated; that leads anterodorsally from mental foramen to re­ lingual condylar emargination at least partially filled gion of the antemolars; mental foramen below middle by a bony plate so that interarticular area is broad; of MI; coronid process narrow, its anterior edge form­ lower condyle enlarged, offset lingually, moved ante­ ing a 90° angle with alveolar plane of horizontal ramus riorly, and facing more ventrally relative to Sorex: and its tip not deflected buccally as in Neomys or condylar process supporting upper condyle curved to­ Blarina; inferior condyle oriented 90° from verticle axis ward skull; coronoid process deflected outward (fig. 1) ; of horizontal ramus and 45° from longitudinal axis, superior pterygoid fossa moderately basined but with articular facet facing ventrally and offset lingually little or no pterygoid spicule; internal temporal fossa of from plane of mandible at inferior sigmoid notch, ante­ the mandible small and oval or with a well-developed rior displacement of inferior condyle not as extreme as in Blarina, Cryptotis, or Notiosorex, slightly less than limula; external temporal fossa of mandible high and in Chimarrogale, about equal to Neomys, greater than extending down coronoid process no farther than upper in Sorex ; interarticular area presumably with moderate sigmoid notch or less, coronoid spicule prominent but lingual emargination not filled with bone plate, but very high on ascending ramus; mental foramen varies condyloid process anterior to interarticular area trans­ in position but usually beneath hypoconid of M± ; zygo- versely thickened more than in most genera of the matic process of maxillary originates opposite posterior Neomyini; base of condyloid process, at its junction part of M2 in all forms; M2 trapezoidal in occlusal view with ascending ramus, vertically narrow in comparison in most forms; most conspicuously M± with no trace of with all soricine shrews not included in tribe Soricini; an entoconid crest but strong entoconid (with one ex­ superior pterygoid fossa of mandible has very shallow ception), talonid ajiteroposteriorly shortened; entire basin with poorly developed boss for insertion of supe­ Ms reduced, talonid tending to be further reduced to rior pterygoid muscle; external temporal fossa of man­ form crescentic crest with V of hypoconid weak or ab­ dible not extended down coronid process as far as supe­ sent and reduced to minute shelf with small cusp in one rior sigmoid notch; internal temporal fossa of mandible genus; roots of Ms separated by distance less than narrowly triangular with well-developed limula. Etymology. Greek: Adelos [aSr/Aos], obscure, and transverse diameter of anterior alveolus; teeth heavily BUrina; related to Blarina to an unknown degree. pigmented and very stout. Blarina and Cryptotis are the only living genera in­ Adeloblarina berklandi Repenning, n. sp. cluded in this tribe. Blarina contains only 2 species Figure 25 (Hall and Kelson, 1959, p. 52) confined to the eastern This species was named in honor of James O. part of the United States; Cryptotis contains at least 25 Berkland, who found the type specimen while collect­ species (p. 55-56) in the eastern half of North America ing with the author in southeastern Oregon during and in Central America. It is the only shrew to have the summer of 1963. entered South America. The tribe contains a fair num­ Holotype. U.S. National Museum 23098; a right ber of extinct genera in both North America and Eurasia mandible lacking the superior mandibular condyle and is known back to the late Miocene. and bearing ML Collected by J. O. Berkland. 38 SUBFAMILIES AND GENERA OF THE SORICIDAE

Coronoid spicule

External temporal fossa

FIGURE 25. Adeloblarina berklandi Repenning, n. gen. and n. sp., USNM 23098. Right mandible and M\.

Type locality and age. U.S. Geological Survey fication with these features strongly suggest assign­ vertebrate loc. Ml 040; in poorly exposed and sparsely ment to the tribe Blarinini; the shallow basin of the fossiliferous orange tuffaceous mudstone at the south superior pterygoid fossa, the width of the inter- end of the Bowden Hills, Malheur County, Oreg. articular area, where preserved, just above the lower Late Miocene (Barstovian, possibly late Barstovian). condyle, and the elevation of the bottom of the exter­ Diagnosis. A species of Adeloblarina of a size nal temporal fossa above the upper sigmoid notch are greater than Cryptotis parva, smaller than Blarina evolutionary modifications of the Blarinini not found b

FIGURE 26. Cryptotis adamsi (Hibbard), topotype, UM 28411. Right mandible and dentition. 40 SUBFAMILIES AND GENERA OF THE SORICIDAE and, hence, has the dental formula of living Blarina. Cryptotis? meadensis Hibbard, 1953 The assignment of this species to Cryptotis in this Less reduced heel on Ms, more lingual metalophid- report is based upon the belief that the following protolophid junction, and somewhat less forward differences between it and Blarina represent phylo- placement of lower articular facet than in living genetically significant characters and that the presence Cryptotis. of P2 is an expectable primitive feature. From species These features appear less advanced but related to assigned to Blarina, Cryptotis adamsi differs in (1) those of the living forms. CA meadensis, however, greater anteroposterior shortening of the talonid of is very unusual in not having a distinctly isolated MI, (2) more posterior placement of the metaconid of MI, relative to the position of the protoconid, (3) entoconid on MI and in having a P4 that is very simi­ greater reduction of the heel of MS, (4) retention of a lar to the P4 in the Limnoecinae. The articular struc­ primitive blarinine mandibular articulation and associ­ ture of the mandible of C.1 meadensis clearly could ated jaw structure, and (5) a rectangular M2. Items not belong to a limnoecine shrew. Late Pliocene (4) and (5) could be considered primitive characters (early Blancan) of Kansas. that represent an ancestral condition in the Blarina Specimen examined. University of Michigan Mu­ lineage, but items (1), (2), and (3) are clearly spe­ seum of Paleontology 27266, the type, from the Rex- cializations in the late Pliocene species not yet achieved road fauna of Kansas. in the living species of the Blarina lineage. A fragmentary specimen assigned to this species has Cryptotis kansasensis Hibbard, 1957 been found in the middle Pliocene (Hemphillian) The teeth and mandible, as determined from pub­ Christmas Valley local fauna of Lake County, Oreg. lished descriptions of both species, are similar to those (U.S. Geol. Survey vertebrate loc. M1061). The in the living species C. magna. No specimens of either geologic range is, therefore, middle Pliocene (Hem­ species have been examined during this review, and I phillian) of Oregon and late Pliocene (early Blancan) am unable to see how the two species differ. From the of Kansas. living C'ryptotis parva, they differ in greater size, lack Specimens examined. Topotypes, University of of a posterior emargination of the upper molars and Michigan Museum of Paleontology 28411, 24352, and P4, more trapezoidal form of M2, a more rounded 27310, all well-preserved mandibles from the Kexroad anterointernal corner of P4, stouter lower molars, and local fauna, Kansas, and USGS M6511, a fragmentary greater filling by bone of the lingual emargination of mandible from the Christmas Valley local fauna, the interarticular area of the mandible. C. kansasensis Oregon. is from the early Pleistocene of Kansas.

Posterior

FIGURE 27. Cryptotis parva (Say), Kansas, UCMVZ 40144. Left mandible and dentition and left upper dentition. DIAGNOSES AND CONTENTS OF SUBFAMILIES 41

Cryptotis parva (Say), 1823 most of range found in living Blarina', mandibular Figure 27 articulation as in Blarina with broad interarticular This living species has also been reported as fossil area, but inferior facet not placed as far forward and from the middle Pleistocene deposits of Kansas (Hib- slightly visible behind lower sigmoid notch in labial bard, 1963, p. 203-204). Only Recent specimens were view; superior pterygoid fossa a shallow basin with seen. no pterygoid spicule; internal temporal fossa small Cryptotis mexicana (Coues), 1877 and circular; external temporal fossa reaching superior sigmoid notch; coronoid process moderately deflected This living species has been reported as fossil from labially. Hibbard (1950, p. 122-127) compared this late Pleistocene deposits of Mexico (Findley, 1953, p. genus with Cryptotis and with Blarina. Paracryptotis 633-639). No specimen was seen. appears to be closely related to Blarina on the basis of Genus PARACRYPTOTIS Hibbard, 1950 the reduction in size of the entire M^ and the trapezoi­ Figure 28 dal occlusal outline of M*, in addition to the articular structure of the mandible and other similarities to Genotype: Paracryptotis rex Hibbard, 1950 Blarina mentioned by Hibbard. In the dental formula Dental formula 1-5-3/1-2-3; Ms well reduced rela­ and possibly in the structure of the talonid of M$, tive to MI with well-developed crescentic talonid crest; Paracryptotis is more closely related to Cryptotis. In M! with deep labial cingulum from protoconid to the lingual position of the metalophid-protolophid posterior part of tooth (behind hypoconid), where junction and the development of the cingulum on M^ it is better developed than in living Blarina, meta- Paracryptotis is less advanced than living Blarina but lophid joins protolophid slightly labial to position of is about equal in this feature to the late Pliocene this junction in Adeloblarina but well lingual to posi­ Blarina gidleyi and to Cryptotis. In the mandibular tion of junction in living Blarina, base of crown broad, articulation, Paracryptotis is somewhat less advanced hypolophid does not extend lingually as far as in than Blarina but more advanced than Cryptotis. living Blarina, and entoconid is slightly more The genus includes only one species, Paracryptotis posterior; P3 smaller than preceding upper ante- rex Hibbard, from the late Pliocene (early Blancan) molars; P4 and M1 with no posterior emargination; Rexroad fauna of Kansas and the middle Pliocene M2 trapezoidal and zygomatic process of maxillary (Hemphillian) Rome fauna of Oregon. originating opposite metacone; pigmentation slightly Specimens examined. Two topotypic mandibles less than in Blarina; mental foramen from beneath labeled University of Michigan Museum of Paleon­ center of M\ to below posterior root of M\ anterior tology 47760. Uncataloged specimen from USGS to average position in living Blarina but overlapping vertebrate loc. M1078, Rome local fauna, Oregon.

FIGTTKE 28. Paracryptotis rex Hibbard, topotype, UM 47760. Right man­ dible and dentition. 42 SUBFAMILIES AND GENERA OF THE SORICIDAE Genus SHIKAMAINO SOREX Hasegawa, 1957 The mandibular articulation compares favorably with Genotype: Shikamainosorex densicingulata Paracryptotis in form and stage of advancement. Hasegawa, 1957 The knoblike structure of the apex of the coronoid Articular structure typical of Blarinini; differs from process seems unique. Because of the mandibular ar­ Paracryptotis in greater inflation of labial

4 Although Kowalski (1956, p. 346) stated that the mental fora­ Sulimski, who prepared Kowalski's illustration of the type, appar­ men is below P4, his figures (pi. 1, figs. 5, 6) clearly show it below ently took Kowalski's statement at face value without referring to the middle of M,. The talonid of Mi in the holotype is broken, and the specimen or his own artistry because he (1962, p. 471) made the the isolated trigonid resembles, at first glance, a P4 ; this situation same mistake in thinking the trigonid of Mi was the P4 when he suggests that Kowalski inadvertently mistook this talonid for P4 while compared his "8orea>" kretzoii with "Sorex" dehneli. describing the position of the mental foramen. He is not alone, for DIAGNOSES AND CONTENTS OF SUBFAMILIES 43

M, of 6212 \ I1 ' ,// showing variation of talonid pattern

Superior pterygoid fossa

Coronoid spicule External temporal

Depressed plate posterior to superior pterygoid fossa filling lingual emargination of interarticular area

FIGUBE 29. Blarina brevicauda (Say), author's collection 6188 from Michigan and MS of 6212 from New Jersey. Left mandible and dentition and left P4. ner in occlusal view not greatly different from that Blarina gidleyi Gazin, 1933 in Sorex. The range in position of the mental fora­ Figure 30 men in the species Blarina brevicauda is nearly as MS within more /Sorex-like variation of living Blar­ great as that seen within the subfamilies throughout ina ~brevicauda', pigmentation of the teeth and struc­ their geologic history. Other features, as the position ture of mandibular articulation identical to those fea­ of the junction of the metalophid with the protolo- tures in living species; MI with metalophid-protolo- phid on MI, the posterior emargination of .P4 and phid junction placed far lingual to position in living M1, or the structure of the mandibular articulation species and similar to Paracryptotis, hypolophid does seem as stable as in any other shrew genus. Four spe­ not extend as far lingual as in living species, and cies have been recognized as fossil. cingula as in Paracryptotis and Adeloblarina rather

Posterior

Medial

FIGURE 30. Blarina gidleyi Gazin, topotype, UM 3304. Right mandible and dentition. 44 SUBFAMILIES AND GENERA OF THE SORICIDAE than reduced in region of hypoconids and deepened Kretzoi (1962, p. 353) placed this species in the below protoconids as in living species; upper teeth genus /Shikamainosorex. Kowalski and Li (1963, p. unknown; mental foramen below midpoint of MI; 143) did not agree with such a generic assignment and horizontal ramus without sharply angulate inferior pointed out that Shikamainosorex has a greater re­ border of living species but possibly stouter through­ duction of MS and a broader cingulum than Blari­ out. noides. These differences appear real to me and in Most of these characters are clearly primitive, and addition the MS of Blarinoides clearly retains a small the species is intermediate between the living species entoconid that is not present on the reduced talonid of and Adeloblarina on the basis of these differences. It MS of Shikamainosorex. From the published de­ is possible the differences are of generic importance scriptions it appears that Shikamainosorex is clearly and that, therefore, this species should not be in­ related to Paracryptotis and that Blarinoides is re­ cluded in Blarina. lated more to Blarina, as the authors of both genera Late Pliocene (early Blancan) Hagerman fauna of originally supposed. Id:Jiu. Specimen examined. University of Michigan Mu­ Blarinini? Incertae Sedis seum of Paleontology 3304, as well preserved mandi­ "Sorex' kretzoii Sulimski, 1962 ble. This shrew appears, from Sulimski's description The other species of Blarina reported as fossil are and illustrations, to have a heavy lower articular all from the late Pleistocene. None of these were facet on the mandible that is basically horizontal and seen. notched in its upper edge and is shifted forward only Blarina fossilis Hibbard, 1943. Kansas. slightly from its position in Sorex, a reasonably broad Blarina simplicidens Cope, 1899. Pennsylvania. interarticular area, an upper articular facet that is Blarina brevicauda (Say), 1823. A living species in ,the east­ rotated well toward the longitudinal axis of the man­ ern half of North America and reported from the late Pleis­ tocene from Oklahoma and Kansas. dible, a labially deflected coronoid process, no ento­ conid crest on the lower molars, and a relatively little Genus BLARINOIDES Sulimski, 1959 reduced J/3. The combination of these features sug­ Genotype: Blarinoides mariae Sulimski, 1959 gests the tribe Blarinini, but unfortunately most of Dental formula 1-6-3/1-2-3; P4 very large and in­ the characters cannot be clearly verified in the pub­ flated; MS greatly reduced but talonid not reduced lished description. The vagueness results not so much relative to trigonid, distinct entoconid; P4 and Ml from the fault of the description, which is better than with moderate emargination of posterior basal out­ the description of many fossil shrews, but from the line as in Blarina; M2 trapezoidal and zygomatic proc­ generalized nature of "Sorex" kretsoii. This is par­ ess originating opposite metacone as in Blarina', teeth ticularly so in its mandibular articulation, which pigmented and stout; mandible very heavy; mental could be either a Blarinella-tjpe in the tribe Soricini foramen between roots of MI ; articular facets of man­ or a very primitive articulation in the tribe Blarinini. dible widely separated and about as in Blarina. From Sulimski's figures (text-pi. 2, figs. 9, 10), the Although very similar to Blarina, according to Sul­ articulation appears to belong in the Blarinini, but if imski (1959, p. 147) Blarinoides differs in its larger so it is closer to the Soricini in lacking a forward size, weaker coronoid spicule, inflated P4, presence of shift of the lower articular facet than is any member entoconid on M3 of all individuals, and relative sizes of the Blarinini except Adeloblarina from the late of the upper antemolars. In labial position of the Miocene. metalophid-protolophid junction and development of Tribal assignment must be decided on the basis of the cingulum on MI, Blarinoides appears much closer the lingual offset of the lower articular facet from to the living Blarina than to the late Pliocene Blarina the lower sigmoid notch, rotation of the upper articu­ gidleyi. lar facet toward the longitudinal axis of the mandi­ One species named: Blarinoides mariae Sulimski. ble, degree of curvature toward the skull of the con- Known from the late Pliocene of Poland and Hungary dyloid process supporting the upper facet, labial de­ (Sulimski, 1962, p. 483). Kowalski (1960, p. 170) flection of the coronoid process, position of the origin suggested that his species "Sorex" dehneli, from the of zygomatic process of the maxillary, and the occlu- early Pleistocene of Poland, may belong in the genus sal outline of M2. None of these are determinable Blarinoides, a possibility that seems logical to me from the published description. One strong point in from a review of the published accounts. No speci­ favor of assignment to the Blarinini is the positive men was seen. statement of Sulimski (1962, p. 469) that the ento- DIAGNOSES AND CONTENTS OF SUBFAMILIES 45 conid crest ("crest between the entoconid and meta- basined with pterygoid spicule in most forms; exter­ conid") is lacking; no shrew assigned to the Soricini nal temporal fossa of mandible low with ventral is yet known that lacks the entoconid crest. margin varying from level of superior sigmoid notch There is a rather striking resemblance between to level of lower articular facet; coronoid spicule high Schlosser's (1924) figure 1 of plate 1 of Blarinella to very low; coronoid process with very strong tend­ kormosi (Schlosser) and Sulimski's (1962) figure 9a ency to curve anteriorly at tip but tip spatulate in of text-plate 2 of "Sorex" kretzoii Sulimski. If a few forms; mental foramen beneath middle of J/i Blarinella kormosi does not have the very distinctive as in Sorex and not placed so far to rear as in many entoconid crest on J/i and M2 and the reduced talonid genera of Blarinini; M2 rectangular in occlusal view on M% of Blarinella, as Schlosser's illustrations sug­ and zygomatic process of maxillary originates poste­ gest and as might be implied by Miller's (1927, p. 9) rior to M2 in nearly all forms except those having words "the characters in which * * * Blarinella kor­ extreme reduction of last molars; protocone of P4 mosi differs from the living Chinese B. wardi appear tends to be farther separated from parastyle and to approach, in every instance, those of Blarina * * *" placed farther medially than in Sorex; upper incisor (Blarina does not have the entoconid crest or re­ bifid or fissident, as in Trimylus, in most Old World duced talonid of Blarinella}, it seems probable that forms; bifid upper incisor not present in New World Blarinella kormosi is very closely related to "xSWea?" forms; M\ with prominent entoconid crest in most kretzoii, and is not Blarinella. forms but weak in some and missing in those having "/Soreoc" kretzoii is from the late Pliocene (lower extreme reduction of last molars; MS variable and part of the We.ze Breccia), Poland. No specimen was may be unreduced as in Sorex, with talonid only seen. reduced as in BlarineUa, greatly reduced as in Blarina, or entirely lost; labial cingula on lower molars never Feisorex pohaiensis Kowalski and Li, 1963 deep as in Blarina, well developed to absent; teeth The authors of this genus and species do not men­ pigmented over less extensive area or lightly pig- tion whether the entoconid crest is present or absent mented relative to Sorex or colorless, slim to stout. on the .3/1-2 of the single known mandible, but it ap­ The Neomyini include nine living genera, more than pears to be absent in their illustrations. The unusu­ any other tribe of the subfamily Soricinae; however, ally specialized manclibular articulation could belong each of these genera include very few living species, to either the tribe Blarinini or the tribe Neomyini, to many only one. The fossil record of the tribe is judge from the published description and illustrations. clearly very incomplete. The oldest specimens are of The deep pit (superior pterygoid fossa) in the area early Pliocene age, and some of these are the most just anterior to the superior condyle of the mandible, specialized forms known, a situation illustrating how however, strongly suggests that Peisorex belongs in far the history of the Soricidae is from being com­ the tribe Blarinini. pletely documented. Because the bifid upper incisor According to the authors (Kowalski and Li, 1963, is also known in the Heterosoricinae, is sometimes p. 142), Peisorex pohaiensis is a large shrew charac­ present in the Soricini, and is found in one crocidu- terized by a connection between the internal tempo­ rine shrew, it would seem to be a primitive character ral fossa (their "fossa pterygoidea") and the man- inherited from a common ancestor. dibular (dental) foramen, by a simple coronoid proc­ The Neomyini are largely a Palearctic tribe al­ ess lacking any spicule, and by the prominent basin of though there are three genera from North America. the superior pterygoid fossa. The shrew is known Except for the living Anourosorex and fossil rela­ from the middle Pleistocene of China. No specimen tives of this genus, most Old World forms have bifid was seen. upper first incisors; the New World genera do not Tribe NEOMYINI Repenning, new name although they are very close to the eastern Asian Diagnosis as for tribe Blarinini except as follows: Chodsigoa in many other characters. Two lineages Lingual condylar emargination of mandible open, are thus suggested. Within the Old World forms no consequently interarticular area very narrow; lower suggestion of further phylogeiietic breakdown has condyle elongate, offset lingually from plane of lower been noted, but rather, the known forms fit rather sigmoid notch to greater extent than in Blarinini, neatly into a uniformly gradational series from the more anteriorly placed and facing more ventrally little specialized Neomys to the extremely specialized than in Soricini, separated in many specimens from Anourosorex. This graded series is entirely morpho­ lower sigmoid notch by groove because of extreme logic and is not related to an evolutionary sequence lingual offset (fig. 1) ; superior pterygoid fossa deeply insofar as the present fossil record reveals. 46 SUBFAMILIES AND GENERA OF THE SORICIDAE

Coronoid spicule External temporal fossa

Boss for insertion of superior pterygoid muscle

FIGURE 31. Neomys fodiens (Pennant), France, UCMVZ 30044. Left mandible and dentition and left upper dentition.

Genus NEOMYS Kaup, 1829 Neomys newtoni Hinton, 1911. Early Pleistocene, England and Figure 31 Czechoslovakia. One specimen was seen, author's collection 6340, from Czechoslovakia. The specimen was returned to Genotype: Sorex daubentonii Erxleben, 1777 the Czechoslovakian Geological Survey since preparation of Dental formula 1-5-3/1-2-3; upper incisor bifid in this report; the new catalog number is unknown. about half of the specimens examined; fourth upper Neomys intermedius Brunner, 1952. Middle Pleistocene, Ger­ antemolar small; P4 with protocone shifted more me­ many. No specimen seen. dial from parastyle than in Sorex; lower incisor with Neomys castellarina Pasa, 1948. Middle Pleistocene, Italy. No specimen seen. one weakly defined and irregular cusp on cutting edge; labial cingulum strong on lower molars; pig­ One living species other than N. fodiens is recog­ ment extensive for tribe but less so than is typical of nized by Ellerman and Morrison-Scott (1951, p. 61). 8orex\ pterygoid boss of mandible moderately devel­ Genus PETENYIELLA Xretzoi, 1956 oped; upper articular facet of mandible oval; lower Genotype: Sorex gradlis Petenyi, 1864 articular facet not placed so far forward that it is This genus was named in a footnote (Kretzoi, 1956, hidden in labial view; coronoid spicule high; external p. 160), and Sorex gracUis Petenyi (1864, p. 70) was temporal fossa high and terminating at superior sig- cited as the genotype. Kretzoi (1959a, p. 239) gave moid notch; tip of coronoid process narrow with a brief diagnosis. prominent anterior curve. Early Pleistocene to Recent Sulimski (1962, p. 479^80, 483) recognized the of Europe, Kecent of Asia. genotypic species Petenyiella gradlis and a new spe­ Except for dental formula, upper and lower incisor, cies Petenyiella zelcea from the W§ze Breccia on the and P4, there appears to be no way to separate Neo­ basis of direct comparison with the Hungarian speci­ mys from Sorex on the basis of dentition alone. mens. Unfortunately most of his description is in Among European species of Sorex, Neomys appears comparison to the genotype, which is poorly de­ distinguishable by the higher cusps in unworn teeth scribed and illustrated in available publications. From (Miller, 1912, p. 68). Sulimski's discussion (1962, pages just cited, and 1959, Five species have been reported as fossil: p. 157-159 under the genus Suncus], the following Neomys browni Hinton, 1911. Late Pleistocene, England. No features place the genus Petenyiella in the Neomyini. specimen was seen. Condyles fairly widely separated, lower facet elon­ Neomys fodiens (Pennant), 1771. Late Pleistocene, England; a living species. Only Recent specimens were seen from gate, and interarticular area narrow; lower incisor the Museum of Vertebrate Zoology, University of California. with two rather indistinct lobes on its cutting edge; DIAGNOSES AND CONTENTS OF SUBFAMILIES 47 MI "similar in structure" (Sulimski, 1959, p. 158) to 1959, pi. 3, fig. 2, and pi. 4, fig. 1; 1962, pi 2, fig. 1, Suncus and therefore assumed to have well-developed and text-pi. 2, fig. 1) most closely resembles P. gra­ entoconid crest; Ms "strongly reduced talonid with cilis \P. pann&rdca of Kormos, 1934, fig. 48] and has poorly recognized hypoconid" also similar in structure the upper articular facet well above the upper sig- to Suncus in which the talonid is more reduced than moid notch and the lower facet well posterior to the in Crocidura in European forms of Suncus and Cro- lower sigmoid notch. The specimens which Sulimski cidura (Miller, 1912, p. 84); teeth pigmented at tips; referred to P. gracilis from the W§ze Breccia (Sulim­ size very small. ski, 1959, pi. 3, fig. 1, and pi. 4, fig. 2) have the upper Kretzoi (1959a, p. 239) gave the following charac­ articular facet below the level of the upper sigmoid ters in diagnosis of P. gracilis: soricine of Suncus notch and the lower facet well forward and almost size with four upper unicuspids which are comparable hidden behind the lower sigmoid notch in lateral view. in form to the unicuspids of Sorex, two-lobed lower This is a much greater difference than has been ob­ incisor, and delicate yellow pigmentation. served in any other genus of shrew. The possibility Kormos (1934, p. 306, fig. 38) gave a good illustra­ that two genera are represented and also that Sulim­ tion of the mandible under the new name Pachyura ski has referred the wrong one to P. gracilis is here pannonica. He also suggested that the species might suggested. be considered a new genus, for which possibility he The genus is known from the late Pliocene of provided the new generic name Allopachyura. Be­ Hungary and from the late Pliocene and Pleistocene cause he did not apply this name to the species, it of Poland (mostly from the lower part of the W§ze would appear to have no validity, and Petenyiella is Breccia, according to Sulimski, 1962, p. 483). No here used. specimen was seen. As Kretzoi said (1962, p. 353), the genus is close to Sorictdus [Episoriculus as here used]. It is a Neo- Genus EPISOBICTTLTTS Ellerman and Morrison-Scott, 1951 myini shrew characterized by its small size, by its Figure 32 rather unspecialized mandibular condyles with the Genotype: Sore® caudatus Horsfield, 1851 small lower condyle farther to the rear than in Neo- Dental formula 1-5-3/1-2-3; upper incisor bifid as mys, and by a greatly reduced and single-cusped talo­ in Neomys, fourth upper antemolar very small; P* nid on MB- with protocone placed close to parastyle as in Sorex-, There appear to be significant differences in the lower incisor as in Neomys but shorter; labial cingu- development of the upper condylar process in Peten­ lum on lower molars as in Neomys ; Ms with heel very yiella. P. zelcea from the W§ze Breccia (Sulimski, slightly reduced relative to Neomys and V of hypo-

FIGUEE 32. Bpisoriculus caudatus (Horsfield), India, CNHM 35415. Left mandible and dentition and left upper dentition. 48 SUBFAMILIES AND GENERA OF THE SORICIDAE conid not so acute; pigmentation less than in Neomys Episoriculus gibberodon (Petenyi), 1864 lacking on J/3 and only at very tip of hypoconid on Equals Crocidura gibberodon Petenyi, 1864; Sori­ talonid of MI and M2 ; pterygoid boss of mandible al­ culus kubinyi Kormos, 1934; Soriculus gibberodon most undetectable; upper articular facet of mandible (Petenyi) of Kretzoi, 1956; and Asoriculm gibbero­ oval as in Neomys; lower articular facet placed far­ don (Petenyi) of Kretzoi, 1959. Elevation of the ther forward than in Neomys and barely visible be­ subgenera of Soriculus to generic stature necessitates hind lower sigmoid notch in labial view; coronoid reevaluation of the generic assignment of this fossil spicule lower than in Neomys and directed postero- species. The following characters bar its assignment ventrally toward lowest point of upper sigmoid notch to either Soriculus or to Chodsigoa and indicate its rather than above it as in Neomys; external tempo­ affinity with living Episoriculus: dental formula ral fossa slightly lower than in Neomys and extending 1-5-3/1-2-3; lower incisor short; Ms only slightly re­ well below upper condyle. Recent of eastern Asia duced with distinct entoconid; complete labial cingu- and late Pliocene of Europe. lum on lower molars; pigment present on tips of Episoriculus was proposed as a subgenus of Sori­ cuspids of the lower molars; no anterior inclination culus by Ellerman and Morrison-Scott (1951) be­ of tip of coronoid process; relative narrowness of cause of marked differences between species previ­ ascending ramus at level of upper sigmoid notch; ously included in Soriculus. They pointed out (p. 57) coronoid spicule higher than in Chodsigoa or Soricu­ that the external differences between these species are lus', lower articular facet not placed as far forward greater than those between Sorex and Blarinella. Fol­ as in Soriculus (not as far as in the living Chodsigoa lowing Osgood's suggestion of the subgeneric impor­ or Episoriculus} ; and upper articular facet oval. tance of the extra tooth in Chodsigoa, they also con­ Late Pliocene of Hungary and Poland. No speci­ sidered this genus a subgenus of /Soriculus. In the men was seen. present review, mandibular and dental characters also Genus CHODSIGOA Kastschenko, 1907 seem to separate these three subgenera to a significant Figures 33, 34 degree. Although these differences are not as great Genotype: Soriculus hypsibius deWinton, 1899 as those between Sorex and Blarinella, they are as Dental formula 1-4-3/1-2-3; upper incisor bifid as great as those that separate many genera of the Neo- in Neomys ; upper antemolars of uniform size; P4 with myini. Because of these differences the three sub- protocone as in Sorex and with extreme emargination genera which Ellerman and Morrison-Scott placed in of posterior basal outline one species, at least, exceed­ the genus Soriculus are here given full generic status ing all other shrews in this feature except Notiosorex for the sake of maintaining the degree of morphologic jacksoni; emargination of M1 and M2 also extreme; distinctiveness noted between genera of the Neomyini lower incisor as in Episoriculus ; P± variable in angu­ and between most genera of the Soricidae. larity of junction of labial shearing blade with crest Episoriculus differs from Chodsigoa in having one formed by posterior cingulum; Ms variable in reduc­ more upper antemolar, in more extensive pigmenta­ tion of talonid from nearly unreduced (fig. 34) as in tion, in a less anterior pitch of the tip of the coro­ Episoriculus to noticeably reduced (fig. 33) with noid process, in a lesser width of the coronoid process talonid shortened, hypolophid crescentic, entoconid lost at the level of the superior sigmoid notch, and in its as distinct cuspid, and entoconid crest almost lost but oval upper articular facet. These differences are not present as very low lingual crest; pigmentation very reduced but present on tips of antemolars, tip of para- great but are of the same sort and magnitude that cone of P4 and in one species on very tip of protoconid separate Episoriculus from Neomys. If Chodsigoa of MI ; pterygoid spicule moderately developed on ridge and Episoriculus are to be considered subgenera un­ running from internal temporal fossa to superior con­ der one genus, it should be under the genus Neomys', dyle well below upper sigmoid notch; upper articular the characters of Episoriculus and Chodsigoa differ facet triangular as in Anourosorex and Notiosorex,' much more from those of Soriculus. lower articular condyle as in Episoriculus and slightly Ellerman and Morrison-Scott recognize two living visible behind lower sigmoid notch in labial view; coro­ species of Episoriculus. noid spicule low entirely at or below level of upper Specimens exmnined. Two specimens of E. cauda- sigmoid notch; external temporal fossa lower than in tus (Horsfield) : Chicago Natural History Museum Episoriculus and terminating downward closer to level 39632, a male from Szechwan, China, and 35415, a male of lower condyle than to upper condyle. Pleistocene from Sikkim, India. These represent two subspecies. and Recent of eastern Asia. DIAGNOSES AND CONTENTS OF SUBFAMILIES 49

FIGURE 33. Chodsigoa hypsibia (deWinton), China, CNHM 35760. Left man­ dible and dentition and left upper dentition.

FIGURE 34. Chodsigoa salenskii (Kastschenko), China, CNHM 39614. Left man­ dible and dentition and left upper dentition.

Three species in this group are recognized by Eller- pigment on the very tip of the protoconid of M\ in man and Morrison-Scott (1951, p. 60-61) of which only N. crawfordi and none on the specimen of G. hypsibia two have been seen. Chodsigoa hypsibia, with a uni­ available for study, but there is pigment on the tip of formly curving labial crest of P± and a reduced heel the MI protoconid of G. salenskii. Both species of Chod­ on MZ, is almost indistinguishable from the living sigoa have greater emargination of P4 and the upper Notiosorex crawfordi of North America on the basis molars than N. crawfordi but not greater than the late of its dentition. Typical of the Old World Neomyini, Pliocene Notiosorex jacksoni. There is also great simi­ the trigonid of MI is longer relative to the talonid than larity in the mandibular structures although the coro- in Notiosorex, and the upper incisor is bifid. There is noid spicule is clearly lower and the ascending ramus 50 SUBFAMILIES AND GENERA OF THE SORICIDAE noticeably broader at the level of the upper sigmoid lack of pigment and the minimal emargination of M1. notch in Chodsigoa. Hibbard (1950), p. 128) made note Kowalski (1958, p. 14) also noted that this species is of this similarity. not Neomys. Specimens examined. Chicago Natural History Mu­ Pleistocene of China (Choukoutien, loc. 3). No seum 35760, a female C. hypsibia (deWinton) from specimen was seen. Shansi, China, and 39614, a female C. salensTdi (Kast- Genus BEREMENTIA Kormos, 1934 schenko) from Szechwan, China. Figure 35 Chodsigoa bohlini (Young), 1934 Genotype: Crossopiis fissidens Petenyi, 1864 Neomys bohlini Young, 1934, Palaeontologia sinica, ser. C, v. 8, Dental formula 1-5-3/1-2-3; upper incisor bifid; no. 3, p. 16-18, pi. 1, figs. 9-11. third upper antemolar small, fourth very small; P4 Pei (1936, p. 18-19) stressed the point that this fossil about as in SoricuLus with parastyle placed somewhat differs from Neomys in that pigment is found only on farther forward than in Neomys, and posterior emargi­ the lower incisor and antemolars and not on the lower nation moderate to strong; lower incisor without ir­ molars and that the talonid of Ms "is reduced to one regularities on its cutting edge; P± with somewhat single pointed cusp." From Pei's illustrations the lower inflated cingulum; Mt stout with complete and inflated condyle appears too far forward, the external temporal labial cingulum, entoconid crest high and prominent; fossa appears much too low, and the coronoid spicule MZ with unreduced trigonid but talonid very reduced, also appears too low to be Neomys. The Mz appears lacking entoconid and entoconid crest as in Chodsigoa to match that of Chodsigoa hypsibia. All these fea­ hypsibia', teeth with dark red to black pigment about tures are those that separate Chodsigoa from Neomys, to same extent as Blarinella, about equal in area to and therefore the species is here placed in Chodsigoa. Soriculus or N&ctogale but much darker in color; upper The coronoid spicule is too low, the paraconid-metaconid articular facet oval; lower facet placed farther forward distance on MI appears to be too great, and the geo­ than in any other genus of Neomyini except Anouro^- graphic distribution does not indicate that it should sorex and its allies and not visible in labial view; inter- be classified as Notiosorex. The species itself could articular area very broad for Neomyini and very similar stand on the basis of the narrowness of the ascending to Blarina, except superior pterygoid fossa is a deep ramus at the level of the upper sigmoid notch. The pit and has not filled in with bone to thicken the con- questionably associated upper dentition (figured by dyloid process as in Blarina ; coronoid spicule high; ex­ Pei) seems to be indeterminate; it does not seem to ternal temporal fossa moderately low below upper sig­ belong to either Chodsigoa or Neomys because of the moid notch and meets superior condyle; anterior curve

Medial composite of SUUG, OF 65563 and SUUG, OF 65564

FIGURE 35. Beremendia fissidens (Petenyi), SUtJG, OF 65563, 65564, and 732818. Right mandible and dentition and left upper dentition. DIAGNOSES AND CONTENTS OF SUBFAMILIES 51 on tip of coronoid process extreme. Late Pliocene to low cusp on cutting edge; P4 paralleling that in Lim- middle Pleistocene of Europe (where it is a very com­ noecinae with labial blade centrally located on tooth, mon fossil) and probably Pleistocene of China. posterolingual basin present; teeth with or without pig­ In several features, Beremendia closely parallels ment; articular structure of mandible as in Neomys; Blarma and Blarinoides. The stout teeth with very coronoid process relatively low and neither spatulate nor dark pigment, the large size, and the articular struc­ curved forward. tures of the mandible seem to show, at first glance, a No specimen of this genus has been seen, and this very close relationship between these genera. The bifid diagnosis is based upon Bate's description and illustra­ upper incisor, the entoconid crest on M^ anterior flex tions but is not the same as hers. Her diagnosis also of the tip of the coronoid, and even details of the con- included the following: Extinct; medium size; deep dylar structure clearly indicate, however, that Bere­ snout; P4 and M1'3 as in Sorex, and M3 larger than in mendia belongs in the tribe Neomyini. Three species Recent Chimarrogale or Soriculus. She made no note have been named: of the conspicuously less emargination of the posterior Beremendia flssidens (PetSnyi), 1864. Europe. The following border of P4 and M1 or of the relatively great lingual specimens have been examined: Central Geological Survey shift of the protocone of P4, in comparison with Sorex, of Czechoslovakia, and StJtJG, OF 732815, 732816, 732817, that seems to be indicated in her figure Tb. and 732818, all from west Slovakia, and StJtJG, OF 65563 and 65564 from south Slovakia. These specimens include Three species are described, all from the Pleistocene most of the upper and lower dentition with complete mandi­ of the Mediterranean Islands: bles. Nesiotites hidalgo Bate, 1945. Mallorca and Minorca. Beremendia sinensis (Zdanski), 1928. Choukoutien fauna of Nesiotites corsicanus Bate, 1945. Corsica. China (generic assignment by Kretzoi, 1956). No specimen Nesiotites similis (Hensel), 1855. Sardinia. was seen. Beremendia ucrainica (Pidoplichko), 1953. Middle Pleisto­ Genus NECTOGALE Milne-Edwards, 1870 cene of Europe (generic assignment by Mikl6s Kretzoi, oral Figure 36 commun., 1966). No description or specimen seen. Genotype: Nectogale elegans Milne-Edwards, 1870 Genus NESIOTITES Bate, 1945 Dental formula 1-4-3/1-2-3; upper incisor bifid; Genotype: Nesiotites hidalgo Bate, 1945 P4 with protocone fairly close to parastyle; lower teeth Dental formula 1-5-3/1-2-3 or 1^-3/1-2-3; upper as in Neomys except stouter cuspids and secondary incisor bifid; fourth upper antemolar minute when cuspids present on P4 and M±; no cuspids on cutting present; P4 and M1 with moderate emargination of edge of lower incisor; labial cingulum absent on P4 posterior basal outline; lower incisor with one or two and present only at base of paraconid blade of lower

FIGURE 36. Nectogale elegans Milne-Edwards, China, USNM 254812. Left man­ dible and dentition and left upper dentition. 52 SUBFAMILIES AND GENERA OF THE SORICIDAE molars; entoconid crest continuous between entoconid edge; P4 elongate with labial cingulum; labial cing­ and metaconid on molars; pigmented enamel extreme ulum of lower molars reduced, moderately strong in extent, covering about three-fourths of labial side below parastylid, weak below protoconid, and absent of M\, but very pale flesh color that contrasts very little posterior to protoconid; entoconid crest very low on with unpigmented enamel pigment more obvious in MI but still present (in contrast to tribe Blarinini); ultraviolet light; mandibular condyles as in Neomys entire M3 somewhat reduced and talonid further except lower articular facet not as long. reduced as in pattern of Blarina; pigment varying This genus is unique in the great extent of its very from somewhat more extensive than in Neomys to light pigment, equaled in extent by Beremendia, and undetectable but when present light orange (as in the accessory cuspids of the P± and lower molars; Domnina and some other ancient shrews) rather than Nectogale, Nettiotites, and Hesperosoresc appear to be dark reddish brown as in Sorex and Blarina\ no the only genera in the tribe that do not exhibit either pterygoid spicule or boss in specimens seen and su­ a spatulate or anteriorly deflected tip to the coronoid perior pterygoid fossa a shallow basin because of process. This is the first genus of the Neomyini dis­ thickening of condylar process; upper articular facet cussed which shows a reduced labial cingulum, a fea­ of mandible oval; lower facet barely visible behind ture characteristic of Anourosorex and its allies. There lower sigmoid notch in labial view; coronoid spicule is no fossil record of this genus and only one living prominent and low on ascending ramus; external species, Nectogate elegans from eastern Asia, according temporal fossa about as in Chodsigoa but with deep to Ellerman and Morrison-Scott (1951, p. 88). basin beneath coronoid spicule; tip of coronoid process The genus has been considered a crocidurine shrew. broadly spatulate with little to no anterior deflection. Specimen examined. U.S. National Museum 254812, Recent of eastern Asia. a male from Szechwan, China. Ellerman and Morrison-Scott (1951, p. 58) recog­ nized only one living species that would be in this Genus SORICULUS Blyth, 1854 genus as recognized here. In addition to a reduction Figure 37 of the labial cingulum of the lower molars, this genus Genotype: Cosira nigrescens Gray, 1842 is the first discussed in the tribe Neomyini to show a Dental formula 1-5-3/1-2-3; upper incisor bifid; significant reduction of the entire Ms and of the ento­ fourth upper antemolar small; P* with protocone conid crest on MI and is the first discussed to show a shifted medially and parastyle shifted anteriorly away strongly spatulate tip of the coronoid process; all from paracone to lengthen tooth, posterior emargina- these trends reach maximum development in Anouro­ tion moderate; lower incisor with one cusp on cutting sorex and its allies. There is no known fossil species

FIGUKE 37. Soriculus nigrescens (Gray), India, CNHM 82563 (teeth) and 82565 (mandible). Left mandible, right lower dentition, and left upper dentition. DIAGNOSES AND CONTENTS OF SUBFAMILIES 53

Lingual and anterior views of M,

FIGURE 38. Chimarrogale himalayica (Gray), India, UCMVZ 119400. Right mandible and lower dentition and left upper dentition. which could be assigned to this genus as here under­ ture is unique in the Neomyini although approached stood. by Soriculus) ; pterygoid boss moderately prominent /Specimens examined. Soriculus nigrescens (Gray), just below upper sigmoid notch; upper articular facet Chicago Natural History Museum 82363, a female oval; lower facet slightly visible in labial view but from Lachen, India; 82564, a male from Lachen, offset lingually from lower sigmoid notch by groove India; and 82565, a female from Ghoom, India. These as in CJiodsigoa, Notiosorex, and Megasorex; coronoid specimens represent two subspecies. spicule slight and very high on process; external temporal fossa fairly low, dropping to about mid­ Genus CHIMARROGAIE Anderson, 1877 way between condyles with a relatively deep basin in Figure 38 lower part opposite pterygoid insertion on medial sur­ Genotype: Crossapus Mmalayanus Gray, 1842 face and partly responsible for lack of basined ptery­ Dental formula 1^-3/1-2-3; upper incisor some­ goid fossa; tip of coronoid process small and knoblike, times bifid; P4 with protocone placed far medially with moderate anterior curve. Kecent of eastern Asia. from parastyle, posterior emargination strong; M2 Ellerman and Morrison-Scott (1951, p. 87-88) listed trapezoidal; lower incisor with smooth cutting edge one living species, CMmarro

Medial

FIGURE 39. Anourosorex squamipes Milne-Edwards, China, CNHM 36970. Left mandible and lower dentition and left upper dentition. style subdued, consequently anterior and medial basal North America (J. H. Hutchison, written commun. outlines form approximately a right angle, emargina- 1964) and Europe. tion of posterior basal outline very slight; M1 with Ellerman and Morrison-Scott (1951, p. 87) recog­ exaggerated parastyle and subdued mesostyle; M2 nized only one species, Anourosorex squamipes Milne- reduced and almost triangular; MB minute; lower Edwards, although some authors appear to still con­ incisor with smooth or very slightly irregular upper sider A. assamensis Anderson as a distinct species. cutting edge; M1 with exceptionally elongate trigonid, Specimen examined. Chicago Natural History Mu­ entoconid crest lacking, metalophid-protolophid junc­ seum 36970, a male from Szechwan, China. tion located very labially and comparable to Blarina, There are two fossil species included in the genus: labial cingulum totally lacking as on all lower teeth; Anourosorex inexpectatus (Schlosser), 1924. Placed in the Mz reduced; Ms minute with very slight talonid re­ genus Neomys by Schlosser and compared with Beremendia maining; teeth unpigmented and bulbous; zygomatic flssidens from Europe. As Miller (1927, p. 10) noted, the process of maxillary originates opposite M2 ', mental species agrees well with living Anourosorex. From Schlos- ser's illustration (1924, pi. 1, fig. 4), the presence of Ms is foramen in depression below middle of M^ coronoid indicated by its alveolus and thus eliminates assignment of process broad and spatulate, very broad at level of the species to the genus AmblycopPus. Middle Pliocene of upper condyle; coronoid spicule strong and fairly low; Mongolia (Ertemte fauna). No specimen was seen. external temporal fossa of mandible low, extending Anourosorex japonicus Shikama and Hasegawa, 1958. Mz about to level of lower condyle; superior pterygoid triangular; M3 and Ms further reduced than in living spe­ fossa of mandible as deep basin with prominent ptery­ cies. Middle Pleistocene of Japan (Siraiwa fauna). No goid spicule; upper articular facet triangular as in specimen was seen. Chodsigoa; interarticular area very narrow; lower Genus AMBLYCOPTTTS Kormos, 1926 condyle placed well forward of lower sigmoid notch. Genotype: Amblycoptus oligodon Kormos, 1962 Middle Pliocene to Eecent, eastern Asia. Dental formula 1-1-2/1-2-2; differs from Anouro­ In dental and mandible structures, this is the most sorex in the following characteristics: One more upper specialized of living shrews. Nectogale, Soricuhis, antemolar; M1 with greater emphasis of parastyle; and Chimarrogale all have features that suggest a relationship to Anourosorex, but it does not seem pos­ M2 completely reduced to triangle and resembling M3 sible to decide which is most closely related from the of other shrews; M\ missing; jaw stouter; inter- characters considered in this review. Although all articular area somewhat broader with oval upper these are eastern Asian shrews of the modern fauna, articular facet. Early Pliocene (Polgardi fissure- close relatives are also found in the Pliocene record of Hipparion fauna), Hungary. One species named, DIAGNOSES AND CONTENTS OF SUBFAMILIES 55 Amblycoptus oligodon Kormos. No specimen was tures could represent a primitive condition, not far seen. removed from the Soricini, that could be expected in Genus HESPEROSOREX Hibbard, 1957 the Neomyini. These suggestions, however, cannot be Genotype: Hesperosorex l&vei Hibbard, 1957 evaluated from the published descriptions. Labial cingulum of M± prominent, somewhat in­ The genotypic species, Hesperosorex lo'vei Hibbard, flated, not deepened, and continuous around hypoconid is from the middle Pliocene Teewinot Formation of to posterior border of tooth; entoconid of MI prom­ Wyoming. No specimen was seen. inent, close to metaconid (closer than on M2) and con­ Genus NOTIOSOREX Baird, 1877 nected to metaconid by an entoconid crest of moderate Figure 40 height; hypolophid of M± (inadvertently called the paraconid by Hibbard) well developed and extending Genotype: Sorex (Notiosorex} crawfordi Coues, 1877 past the entoconid to the lingual cingulum (not so Dental formula l-4r-3/1-2-3; upper incisor not extensive lingually on M2 to judge from Hibbard's bifid; P4 with protocone about as in Sorex, posterior figure); MI and M2 stout and "tips of the molars show emargination strong to extreme; M2 rectangular; lower evidence of pigmentation," suggesting noticeably less incisor short with one low cusp on cutting edge; P4 pigmentation than in Blarina; mental foramen below with curving junction of labial blade and crest of middle of M^ coronoid spicule 5 not well developed posterior cingulum; Mi with metaconid and paraconid and fairly low, as in Chodsigoa; tip of coronoid close together relative to Chodsigoa, labial cingulum process narrow, not curved anteriorly and not spatu- strong, entoconid and entoconid crest strong; J/3 with late; articular structures most generalized of tribe; reduced heel, entoconid lost, entoconid crest sup­ upper facet narrowly oval; lower facet broadly oval pressed, V of hypoconid not developed, and metalophid and hypoconid form curved trenchant crest; tip of and placed only moderately forward but offset far lingually (characteristic of Neomyini) and separated protoconid of M^ of paracone of P*, and tips of more from lower sigmoid notch by a groove; interarticular anterior teeth pigmented; teeth slender; zygomatic process of maxillary originates opposite metastyle of area broad relative to all other genera of Neomyini M2 ; mental foramen below middle of M\\ tip of except Beremendia and comparable in general develop­ coronoid process narrow and inclined slightly forward; ment to specialized Soricini as Blarinella, or general­ coronoid spicule moderately low; external temporal ized Blarinini as Cryptotis and Adeloblfirina,. Middle Pliocene (Hemphillian) of North America. fossa of mandible moderately low and extending down This genus is somewhat questionably referred to the to about midway between condyles; superior pterygoid Neomyini because the genotypic specimen, a mandible fossa of mandible deeply basined, pterygoid spicule with MI and M2, was not available for examination. strong or barlike; condyloid structure as in Chodsigoa Classification with the Soricini seems to be eliminated with triangular upper facet and groove labial to by the condyloid and coronoid structures but not by lower facet. Late Pliocene (early Blancan) to Recent, great or obvious differences. The tribe Blarinini is North America. eliminated by the entoconid crest of M± and by the One living species, Notiosorex crawfordi (Coues), is extreme offset of the lower condyle. The absence of a recognized and is also known from the late Pleistocene lingual emargination of the interarticular area is the of Kansas and California. The one extinct species, chief objection to placing Hesperosorex in the tribe Notiosorex jacksoni Hibbard, is known from the late Neomyini, but this objection seems to be outweighted Pliocene (early Blancan) of Kansas (Hibbard, 1950) by the distinctive lingual offset of the lower condyle and possibly from the middle Pleistocene (Irving- and the resulting groove between the condyle and the tonian) of California (Vallecito Creek fauna, unpub. lower sigmoid notch. The absence of the emargination record, J. A. White, oral commun., 1963) and differs of the interarticular area seems to be due to a minimal lingual shift of the superior condyle resulting from in being intermediate in size, bet ween living Notiosorex minimal curvature toward the cranium of the condy­ and Megasorex and in having extreme emargination loid process supporting the superior condyle and thus of the posterior basal outline of P4 and the upper suggests that the coronoid process is not deflected molars, thus resembling Chodsigoa salenskii. labially as in more typical members of the Neomyini. Specimens examined. Four specimens of the living If these suggestions are true, the mandibular struc- species from Texas and California in the University of California Museum of Vertebrate Zoology; one 5 As here used; Hibbard reserved the term, possibly correctly, in specimen possibly of N. jacksoni from the Pleistocene the sense used by Gaugrahn, for a needlelike projection rather than a ridge. of California in the Los Angeles County Museum. 56 SUBFAMILIES AND GENERA OF THE SORICIDAE

FIGURE 40. Notiosorex crawfordi (Coues), Texas, UCMVZ 98040. Left mandible and dentition and left upper dentition.

FIGURE 41. Megasorex gigas (Merriam), Mexico, LACM 13568. Left mandible and dentition and left upper dentition.

Genus MEGASOREX Hibbard, 1950 entoconid crest of M\ very low; talonid of Ms little Figure 41 reduced and as in Chodsigoa scdenskii; zygomatic Genotype: Notiosorex gigas Merriam, 1897 process of maxillary originates behind J/2 ; mandibular Dental formula 1-4-3/1-2-3; upper incisor not structures virtually as in Notiosorex and Chodsigoa bifid; P4 about as in Notiosorex but with only slight except tip of coronoid more inclined anteriorly and posterior emargination; M2 rectangular; lower incisor more spatulate than in Notiosorex; one living species, as in Notiosorex', P4 with more angulate junction of Megasorex gigas (Merriam), Recent of Mexico. labial shearing blade with crest of posterior cingulum Specimen examined. Los Angeles County Museum than in Notiosorex and more like Chodsigoa salenskii; 13568, a female from Colima, Mexico. DIAGNOSES AND CONTENTS OF SUBFAMILIES 57

Neomyini? Incertae Sedis incisor fissident, teeth not pigmented. The unpig- mented, carnivorelike lower teeth are most character­ Deinsdorfia franconia Heller, 1963 istic of this subfamily. This genus and species was defined (Heller, 1963, p. 5-7) on the basis of a toothless right mandible from "Sorex" gracilidens Viret and Zapfe, 1951 the Pleistocene fauna of Deinsdorf. In Heller's P4 with single, narrow, high cusp that has narrowly photographs it is characterized by rather extreme triangular occlusal section surrounded by a basal posterior placement of the mandibular condyles but cingulum, anteroposteriorly two rooted; MI very sug­ otherwise appears not unlike Neomys. Heller's (fig. gestive of F4 in Martes: narrow, paralophid nearly 2c) drawing of the posterior view of the condyles, parallel to axis of mandible, metaconid reduced and however, does not appear similar to Neomys at all. fused to protoconid; metalophid nearly parallel to Fejfar (1966, p. 243) interpreted Heller's drawing as axis of mandible, hypoconid and entoconid strong indicative of the tribe Soricini, but I have some ques­ and forming square corners on posterior border of tion about the accuracy of the drawing. If the draw­ short talonid; MB also similar but reduced and talonid ing is correct then I suspect that Deinsdorfia belongs reduced to rounded heel apparently with only one in the Blarinini. In any event, I am unable to place cusp; mental foramen below P4, internal temporal the genus in any soricid tribe from published descrip­ fossa well developed; upper and lower articular facets tions but feel that it most likely belongs in the clearly formed and fairly widely separated, but there Neomyini. appears to be no interarticular excavation. Late Mio­ cene (Vindobonian) of Vienna Basin. Subfamily ALLOSORICINAE Fejfar, 1966 From published descriptions, this shrew does not This subfamily was recently established by Fejfar to clearly belong to any subfamily except the Allosori­ include a most unusual shrew, Allosorex stenodus. In cinae. No specimen was seen. Fejfar (1966, p. 240- his discussion of this shrew, Fejfar (1966, particularly 241, figs. 13, 14) described new material and discussed p. 240-241) pointed out the similarities of Allosorex to the possible ancestral position of "Sorex" gracilidens "Sorex" gracilidens and suggested that the latter could to the genus Allosorex. well be the late Miocene ancestor of Allosorex. He did not, however, include "/Sorex" gracilidens in his new Genus ALLOSOREX Fejfar, 1966 subfamily. To me, the similarities are strong, how­ Genotype: Allosorex stenodus Fejfar, 1966 ever, and, being faced with a problem of subfamily Large; dental formula l-?-?/l-2-3; sectorial de­ assignment that Fejfar did not have, I include "Sorex" velopment of cheek teeth extreme, paralophid and gracilidens in the family Allosoricinae, rather than metalophid parallel to axis of mandible; lower incisor in "Soricidae, Incertae Sedis," for it clearly does not with hooked tip and trough along lingual side of belong in any other subfamily of shrews. The follow­ cutting edge; upper incisor very weakly fissident; ing diagnosis of the subfamily Allosoricinae is revised coronoid process very low; condyles very large and from Fejfar (p. 223) with modifications to allow inclu­ widely separated so that they ar.e not connected by an sion of "Sorex" gracilidens. interarticular area. Late Pliocene, Czechoslovakia. Lower antemolars high cusped with conspicuous This genus is, easily, the most specialized and un­ sectorial adaptation, P4 apparently derived from a usual shrew known. crocidurinelike triangular crown but a clear tendency Specimen seen. STOG, OF 652366, a toothless to form a high sharp anteroposterior shearing blade mandible, and STOG, OF 652370, a mandibular frag­ with no overhang of alveoli by enamel cap; lower ment with ./I/i, both topotypes from Ivanovce, near molars narrow and high cusped, very carnivorelike, Trencin, Czechoslovakia. trigonid much longer than talonid, paralophid blade knife edged and conspicuously oriented anteroposteri- SORICIDAE Incertae Sedis orly in comparison to most shrews, metaconid close to Podihik kura Deraniyagala, 1958 and much lower than protoconid, metalophicl united This living shrew was defined on the basis of two to protoconid and forming a continuous crest with specimens from Ceylon and was considered to be a hypolophid, hypolophid united with entoconid, ento- soricine shrew because the teeth were pigmented pale conid crest absent and talonid basin open lingually reddish gray. I am informed by Dr. Deraniyagala between entoconid and metaconid particularly in Allo­ (written commun., 1965), however, that the coloration sorex, cingulum weak to not present; coronoid process of the teeth has entirely disappeared in both preserved low and anteroposteriorly wide; mandibular condyles specimens. This suggests to me that the reddish color separated and upper one placed far posteriorad; upper was a stain rather than actual tooth pigment, which 58 SUBFAMILIES AND GENERA OF THE SORICIDAE is very stable. From published descriptions and tooth or possibly closer to the metaconid, entoconid supplemental notes provided by Dr. Deraniyagala I clearly separated from hypolophid and connected to am unable to assign this shrew to a subfamily. If it is metaconid by a high and well-formed entoconid crest, a soricine shrew, its occurrence would be most unusual, no hypoconulid, labial cingulum probably strong, la­ and the only one in peninsular India or Ceylon and bial reentrant between the protoconid and hypoconid the only one south of the ancient Tethys seaway. emerges at level of cingulum; M2 like MI but para­ conid closer to metaconid; M3 with trigonid like that Names of fossil genera not used in text on Ma but talonid reduced (to condition as in Sorex}, Allopachyura Kormos, 1934 ; included in Petenyiella. entoconid connected to hypolophid; all teeth pig- Asoriculus Kretzoi, 1959 ; included in Episoriculus. Heterosorex Gaillard, 1915 ; included in Trimylus. mented light orange; zygomatic process of maxillary Miothen Cope, 1873; included in Domnina. originates behind M2 ] zygomatic arch probably in­ Oligosorex Kretzoi, 1959 (Oligosorex meyeri) ; possibly valid; complete but maxillary process strong; mental fora­ in Crocidosorex ( Crocidosorex antiquus ). men below P± or /*3 ; internal temporal fossa of man­ Plesiosorex Pomel, 1848, not a soricid. dible strong but not pocketed; external temporal Protosorex Scott, 1895; included in Domnina. fossa of mandible weak and confined to tip of coro- ANCESTRY OF THE SORICIDAE noid process, its posterior margin above upper sig- moid notch; masseteric fossa of mandible moderately In this review it has been shown that the Soricidae strong, insertion on angular process marked by crest; is a family of diverse phylogeny and that subfamilies superior pterygoid fossa of mandible on anterior side can be recognized by consistent morphologic differ­ of condylar process; condylar process delicate; con- ences which are traceable back through the late and dyle small and single, articular facet probably cov­ middle Tertiary toward a common ancestry. The ers dorsal, posterior, and ventral sides of condyle to family is, however, distinctly differentiated from all fit into pocketed glenoid; small size. other mammals. There is no shrew known that is This hypothetical ancestor would unquestionably sufficiently generalized morphologically to be this com­ resemble a shrew despite the lack of the pocketed in­ mon ancestor. The dental and mandibular characters ternal temporal fossa, of the doubled mandibular con­ of the oldest known member of each of the four sub­ dyle, of the emargination of the posterior border of families and the directions of modification of these P* and the upper molars, and of the reduced number characters in younger forms suggest what this origi­ of antemolars. The characters that would make it nal shrew should look like. Its diagnosis may be appear shrewlike, together with the occurrence of these about as follows: characters in mammals with similar dentition, are as Dental formula 1-6-3/1-5-3; upper incisor falci­ follows: form and bifid; upper antemolars except P4 small, 1. Procumbent and enlarged lower incisor. Perhaps simple in form, equal in size or anterior ones larger, this is the most typically soricid feature. It is only slightly crowded; P4 nearly equilaterally tri­ shared with no similar mammal. angular in occlusal view, protocone close to anterior 2. Falciform upper incisor. This is shared to a slight apex of triangle and close to incipient parastyle at degree by some erinaceoids, as Saturninia, but still anterior apex, posterior emargination very slight or is a typical soricid feature. In connection with this lacking; J/1 square in occlusal view with W-shaped tooth, it should be noted that the bifid upper incisor ectoloph, protocone strong, hypoconal shelf and cingu- found in the Heterosoricinae, in the tribe Neomyini, lum strong, emargination of posterior outline slight and in some members of the tribe Soricini is also or lacking, no protoconule or metaconule; M2 square found in some Chiroptera but has not been noted and similar to Ml ; M3 triangular, metastyle and hypo­ in any other mammal with closely comparable den­ conal shelf lacking; lower incisor large with extreme procumbency typical of Soricidae, possibly with two tition. rounded cuspules on upper cutting edge; lower ante- 3. Entoconid separated from the hypolophid (except molars with a single triangular cusp, the first and last on M3 ). This is a typical soricid feature but is (P4 ) largest, only slight tendency to be crowded; found in a few Chiroptera and also in erinaceoids P4 not even slightly molariform and clearly single when the hypoconulid breaks the continuity of the cusped, posterior and lateral cingula prominent, two hypolophid. In the erinaceoids, however, the pres­ rooted; MI with trigonid and talonid well developed, ence of the hypoconulid is a far greater dissimilarity metaconid and paraconid more widely separated than than the separation of the entoconid is a similarity on M2, metalophid meeting protolophid at midline of to the shrews. PALEOZOOGEOGRAPHY AND CORRELATION 59 4. Entoconid crest on MI (and M2). This is a typical It appears, therefore, that Soricidae were certainly soricid feature but is shared with most Chiroptera represented in the Oligocene by the subfamilies Het- and with Nesophontes. erosoricinae and Soricinae. The subfamily Croci- 5. A strong hypoconal shelf and cingulum depressed durinae and the Limnoecinae have features that could from the level of protocone and no distinct hypo- not have been entirely derived from either the Sori- cone on M1 and M2. This is a typical soricid fea­ ciniae or the Heterosoricinae, as these subfamilies are ture that is shared in typical development only by here understood from their earliest known rec­ the Chiroptera. The Talpidae either lack hypo­ ords ; the features appear to have been inherited from' conal structures or have a hypoconal shelf that is on an ancestral form possessing some characters of both a level with the protocone and has a fairly distinct Oligocene subfamilies. Therefore it has to be as­ hypocone. The same can be said for most forms in­ sumed that the Crocidurinae and Limnoecinae were cluded in the Erinaceoidea. also present in the Oligocene or that an unknown 6. Protoconule and hypoconule absent. This is a group was present from which both diverged. The character that is shared with the Chiroptera, Tal­ geographic isolation of these two subfamilies, the pidae, and Nesophontes. Crocidurinae in the Old World and perhaps only 7. W-shaped ectoloph. This is a character that is south of the Tethys seaway before the Miocene (Re- shared with the Chiroptera, Talpidae, and perhaps penning, 1965, p. 195) and the Limnoecinae in the Nesophontes, as well as many other dialambdont New World after the early Miocene, suggests that they insectivores. were derived from the native soricid stock sometime 8. M3 reduced by loss of metastyle and hypoconal before the earliest record of the soricines. This would structures. This is a character shared with the certainly be in the Oligocene for the origin of the Chiroptera and Talpidae but not with typical Eri- Crocidurinae of the Old World and probably in the naceidae. Oligocene for the origin of the Limnoecinae in the 9. P4 not molariform. This is a character shared New World. with the Chiroptera, Talpidae, and Nesophontes. The Oligocene subfamilies of the Soricidae diverged 10. No hypoconulid on MI and M2. This is a char­ from an unknown, preearly Oligocene, very shrew- acter shared with the Chiroptera and Talpidae, as like soricid ancestor. This common ancestor appears well as with some erinaceoids and possibly Neso­ to have evolved, along with the Chiroptera and the phontes. Talpidae, from an unknown soricoid. If one assumes 11. Zygomatic arch (or process of the maxillary) orig­ that the slow rate of morphologic change evidenced inates behind M2. This is a character shared with in the middle to late Tertiary fossil record of the the Chiroptera and the Talpidae. shrews reflects earlier rates of change, divergence of 12. Metalophid-protolophid junction near the meta- these groups from the unknown soricoid must have conid of MI and M2. This is a character shared taken place before the Eocene. It would seem the and exceeded by the Talpidae and shared by the Soricoidea-Chiroptera stock could not have diverged Chiroptera. from the remaining insectivores after these had a 13. Metaconid farther from paraconid on MI than on molariform P4, and this may well take the ancestry M2. This character is found in some forms of most of the Soricoidea (at this point including the Chirop­ insectivore groups. It is mentioned here because it tera) back to the Mesozoic and possibly to the origins has not been seen in the Chiroptera during this of placental mammals. review. PALEOZOOGEOGRAPHY AND CORRELATION Although it represents only an approximation based Before the late Pliocene the fossil record of the upon the inferred dental characters of the hypothetical shrews is quite inadequate. Even after that time the ancestor of the four subfamilies of the Soricidae, it is known specimens are not abundant except from a few interesting to note the relative degree of resem^ance localities where either the manner of fossil accumula­ of the shrews to other insectivores and bats. Of the tion or the method of collection has been favorable. 13 characters listed as shrewlike, the Chiroptera share Furthermore, most shrews are highly endemic ani­ 11, the Talpidae share 7, the Erinaceoids share 4+, mals not only geographically limited in both living and the problematic sub-Recent genus Nesophontes and fossil species but tending to remain so through­ from the West Indies shares 4+. Of course consid­ out appreciable spans of their evolutionary histories. eration of other morphologic characters, as in the Finally, rates of evolution vary greatly as illustrated forelimbs of bats, may outweigh dental similarities. by the conservatism of the tribe Soricini contrasted 60 SUBFAMILIES AND GENERA OF THE SORICIDAE

with the exorbitant adaptive modification and vari­ the loss of one, two, or three upper and lower ante- ety seen in the tribe Neomyini, both within the sub­ molars, in greater reduction of the talonid of Ms, and family Soricinae. These conditions all indicate that in a greater emargination of the posterior basal outline interpretation of the significance of a fossil shrew of P4 and M1. In all other features, morphologic dif­ should be made with caution. ferences seem to be greater between living crocidurines The small subfamily Heterosoricinae contains only than between living forms and late Miocene forms. four known genera and is characterized by little ram­ The Crocidurinae are largely tropical Old World ification of its phylogeny. Domnina°is known only forms and are the only shrews, fossil or living, known from North America, where it ranges through the from Africa, peninsular India, and Ceylon. In Africa Oligocene and early Miocene. During this time it the subfamily shows its greatest diversity in living tends to increase the crowding of its antemolars and genera. The apparently late appearance of the croci­ possibly to reduce its pigmentation, but stages of evo­ durines in the Tertiary scene in Eurasia and their obvi­ lution cannot be clearly defined on the basis of the ous African affinities suggest that they may have known material. Unless a specimen can be identified evolved largely in Africa. Their distinctly soricid fea­ with a known species, it appears that it may be use­ tures, however, do not suggest that they are derived ful in establishing geologic age only in terms of from independent African sources but rather that they closeness to one limit of the temporal range of the are descendant from the same stem as all other shrews genus rather than to the other. and in many features are not far removed, morphologi­ The genera Paradomnina and Ingentisoreso are cally, from this ancestry. known only from the late Miocene of North America. Closely paralleling the Crocidurinae in lack of spe­ Trimylus is better represented in number of species cialized morphology but entirely distinct in distribu­ and in greater temporal and geographic range. In tion, the small New World subfamily Limnoecinae is North America and possibly in Europe the evolution of known by a few species from the early Miocene to the this genus appears to be almost straight line with a middle Pliocene. The earliest known species is associ­ uniform rate of change. It seems quite possible at ated with soricine shrews from which it is clearly dis­ present to recognize middle or late Oligocene and early, tinct on the basis of all subfamily criteria. These middle, and late Miocene stages of evolution in North earliest forms are clearly intermediate between the late American Trimylus. The possibility of similar useful­ Miocene to middle Pliocene Limnoecus and the hypo­ ness in the European forms cannot be evaluated with­ thetical soricid ancestor. Thus, stage of evolution out examination of the specimens. within the Limnoecinae is recognizable to the extent of Although Trimylus is known only back to the middle distinguishing a late Miocene to middle Pliocene spe­ Oligocene, it seems reasonable to assume that the genus cies from an early to middle Miocene one. Assuming became a part of the fauna of both North America and comparable rates of evolution, earlier (but now un­ Europe during the early Oligocene mammalian inter­ known) forms should be equally as definitive of geo­ change. Oligocene species show a greater similarity logic age, although perhaps not distinguishable from than do more recent species of both hemispheres; since the Crocidurinae except in their geographic distribu­ the beginning of the Miocene, possibly since the early tion. Oligocene, endemic Old and New World lineages ap­ The similarity of the dental and mandibular char­ pear to have been in existence. Trimylus roperi from acters of the Crocidurinae and the Limnoecinae result the early Miocene of Colorado may represent an immi­ largely from the retention of those features believed gration of the European stock during the early Miocene to have been typical of the hypothetical soricid mammalian exchange or may be a part of the American ancestor. It appears to be reasonable that both of stock that has an otherwise unrecognized variability in these subfamilies evolved in isolation, removed from the position of the mental foramen. The Heterosoricinae the competitive pressure of the closely related Sori­ is unique in several features and seems much more dis­ cinae. Although the earliest known limnoecine genus, tantly related to the other soricid subfamilies than all Angustidens, is associated with the earliest known the other subfamilies are to each other. North American soricine genus, Antesorex, it seems The fossil record of the Crocidurinae is very meager. likely both that the Limnoecinae had an older history The oldest specimens known are of middle Miocene age in North America and that Antesorex represented a from Europe (Doben-Florin, 1964) and Africa (Butler recently introduced stock closely similar to Crocido- and Hop wood, 1957, p. 2). The subfamily is character­ sorex of comparable and even greater age in Europe. ized by the retention of primitive characters, and the Furthermore, the extinction of the North American modern forms differ from those of the late Miocene in Limnoecinae in the middle Pliocene was at the time PALEOZOOGEOGRAPHY AND CORRELATION 61

SORICINAE CROCIDURINAE NEOMYINI

Crocidura Diplomesodon Anourosorex Chimarrogale Feroculus Chodsigoa Myosorex Episoriculus Praesorex Scutisorex Nectogale Solisorex : Neomys Suncus Notiosorex Surdisorex '. Soriculus Sylvisorex

{ .Neomys Crocidura : : Notiosorex-/ Myosorex ':' ' Nesiotites :

: Notiosorex '. Crocidura ' Chodsigoa ' Suncus Anourosorex Shikamainosor >,x ' Beremedia

Myosorex Crocidura Neomys Suncus Episoriculus ' Diplomesodon

Beremendia Blarina Blarinoides Notiosorex

Hesperosorex. Anourosorex

; ;. '.' '.:'. Amblycoptus[:]' ::':" .'. ' .' ' :'

'. '. ' : ' aff. Anourosorex ; :". '. '. '/:'

\ Crocidura -Pigment reduced or Soricella lost in a few genera

Bifid upper incisor retained in most Old World forms

Pigment reduced- to North America Pigment lost- to Africa Pigment retained

Unknown form with internal temporal Unknown form with internal fossa pocketed into mandible temporal fossa of man­ dible not pocketed Hypothetical soricide ancestor

FIGURE 42. A phylogeny of the Soricidae. 62 SUBFAMILIES AND GENERA OF THE SORICIDAE of apparently explosive diversification of the Sori- These early soricines (Orocidosorex and Antesoreae) cinae. This may also coincide with the approximate differ from the living Sorex in their greater number of time of extinction of the Heterosoricinae in both the teeth, more anterior position of the mental foramen, a Old and New Worlds, although there is no record of more primitive P4, less emargination of the posterior this subfamily younger than late Miocene in North border of P4, M1, and Mz, and in the slightly less America. In the Old World the Crocidurinae have separation of the articular facets on the mandibular survived. With the exception of a few species of condyle. These early soricines are also believed to be Suncus and Crocidura, however, the Crocidurinae of ancestral to other shrews of the Soricinae. Africa and peninsular India and the Soricinae of The Blwrinella generic group diverged from those Palearctica (excluding Mediterranean Africa and in­ with this primitive pattern presumably in the Miocene cluding northern India, Burma, and Indochina) are by enlargement of the mandibular condyle (but with mutually exclusive in their ranges. Those species of no relative increase in separation of the facets as was Suncm and Crocidura which have successfully invaded done in other tribes), accentuation of the entoconid Palearctica seem to have done so by being adaptable to crest of MI and M2, and reduction of the talonid of a great diversity of environments from dry grasslands M3. This generic group is represented in the late to damp forests and shores; many are commensal with Miocene and early Pliocene of North America, in the humans. middle Pliocene of China, and the late Pliocene of The boundary between those areas occupied entirely Europe (where Petenyid continues until the middle by the living crocidurine shrews and those areas occu­ Pleistocene), and now lives in the Chinese highlands. pied largely by the living soricine shrews coincides The tribe Blarinini is first known from the late with the position of the Miocene Tethys seaway to a Miocene in the genus Adeloblarina of North America. remarkable degree and suggests an incredible degree This genus is not far removed from the basic stock of of endemism with these shrews. the Soricinae, which is first known in North America The Soricinae is the largest subfamily of the shrews from the early Miocene, but has several features and has a fossil record known from the late Oligocene clearly indicative of the Blarinini. The middle Plio­ of Europe and from the early Miocene of North cene to living Cryptotis is not much advanced over this America. The earliest members of the subfamily are stage in several respects, but the living Blarina is very generalized, and this primitive condition persists greatly specialized in most characters except for its today in Sorex and closely related genera, here in­ retention of five upper antemolars. The number of cluded in the tribe Soricini. Two other tribal lineages upper antemolars, the nature of reduction of the M$, diverge from this unspecialized stem: the tribe and the degree of specialization of the mandibular Neomyini, apparently of Old World origin and still articulation seem to indicate that the tribe is divisible largely of Old World distribution, and the tribe into three generic groups: Two are represented by the Blarinini, apparently of New World origin and now living genera Blarma and Cryptotis and the third by confined to this area. Morphologically this subfamily the extinct Paracryptotis from the Pliocene. is so diverse that the stage-of-evolution concept has Blarina and related shrews are derived from a form no meaning at more than tribal level and in many similar to Adeloblarina by retention of five upper groupings is apparent only in smaller taxonomic units, unicuspids, by lack of reduction of the talonid relative here informally referred to as generic groups. In the to the trigonid of M3 (although the entire tooth may tribe Neomyini the fossil record is inadequate to be reduced), and by extreme separation of the man­ chronologically define any evolutionary picture al­ dibular condyles. Only two genera have this charac­ though the tribe is rather well represented in the fossil terization: Blarinoides from the late Pliocene of record. Europe and Blarina from the late Pliocene to Kecent The tribe Soricini includes two generic groups: the of North America. genus Sorex and closely related forms and the genus Cryptotis and Paracryptotis are derived also from BlarineUa with fossil genera related to it. The Sorex a form similar to Adeloblarina by the retention of group is conspicuously unspecialized and uniquely only four upper unicuspids except the earliest known successful among all shrews, as exemplified by /Sorex, species and by the reduction of the talonid of M3 the only living holarctic genus. Except for the Hetero­ relative to the trigonid. In Cryptotis, reduction of soricinae, the earliest known shrews belong in the the M3 talonid is greatest but separation of the man­ Sorex group by virtue of their lack of specialization. dibular condyles is relatively slight. The genus is KEY TO LIVING AND FOSSIL GENERA OF THE SORICIDAE 63 known from the middle Pliocene to the Kecent in larity of habits and the presence of this specialized North America. Paracryptotis has a somewhat less member of the Neomyini in the lower Pliocene of the reduced talonid on Ms but has much greater separa­ Old World, at the approximate time of the extinction tion of the mandibular condyles, nearly as extreme as of Trimylus, seems to provide a plausible explanation in Blarina, and is known from the middle and late for the extinction of the Heterosoricinae in that area. Pliocene of North America. /Shikamainosoreos from Within the Crocidurinae the extinct Diplomesodon the middle Pleistocene of Japan is very closely related fossorius and to some extent the living Solisorex have to Paracryptotis but is somewhat more advanced in also made comparable adaptations of their mastica­ reduction of the talonid of J/3 and in separation of tory structures. the mandibular condyles. "/Soreiv" kretzoii and "Sorex" In summary, the shrews are useful in correlation of dehneli from the late Pliocene of Poland may be re­ geologic age, but the degree of their usefulness de­ lated to Cryptotis or Paracryptotis, but published in­ pends on the lineage represented and on the geo­ formation is not sufficient to clearly indicate that these graphic considerations involved in such correlation. generic groups are represented in the Cenozoic of The subfamily Limnoecinae clearly originated in Europe. North America. The subfamily Crocidurinae almost Although largely a North American tribe, the Blar- certainly is of Indo-African origin although it must inini clearly were introduced once, and possibly twice, have been derived from a soricid stock of holarctic into the fauna of the Old World, where they do not distribution which was also the source of the Limno­ seem to have been too successful. Possibly they found ecinae and the Soricinae. The large subfamily Sori- the Old World Neomyini strong competition. Crypto­ cinae seems to be of Eurasian origin although diversi­ tis is the only shrew which has entered South America. fication of the subfamily into tribes took place The tribe Neomyini is first known from the early throughout Holarctica: the tribe Blarinini of New Pliocene of Europe and North America. The earliest World origin, the tribe of Neomyini of Old World known European form, Amblycoptus, is the most spe­ origin, and the tribe Soricini representing the primi­ cialized shrew of the tribe. Therefore, it appears that tive Old World stock. no pattern of evolutionary stages is detectable from The Soricidae offer, with varying certainty, evidence the information available for this review, but more of intercontinental faunal exchanges. Tentatively, likely that the diversification in morphologic types of these may be listed as follows: the tribe Neomyini was as great in the early Pliocene 1. Late Eocene or earliest Oligocene. Dispersal of as it is today. Trimylus between North America and Eurasia; pos­ The tribe is also represented in North America by sible dispersal of ancestral Limnoecinae from Eu­ Hesperosorex from the middle Pliocene, by Notiosorex rasia to North America; possible dispersal of an­ from the late Pliocene to the Kecent, and by the liv­ cestral Crocidurinae from Eurasia to Africa. ing Megasorex. These genera differ from most of the 2. Early Miocene. Dispersal of Soricinae from Eu­ Old World Neomyini in their lack of a bifid upper rasia to North America; possible dispersal of Trimy­ incisor. Inasmuch as the bifid falciform upper incisor lus to North America. is also found in the Heterosoricinae, it is considered 3. Middle Miocene. Dispersal of some Crocidurinae to be a primitive feature inherited from a common from Africa to Eurasia. ancestor. It therefore seems more reasonable to postu­ 4. Late(?) Miocene. Dispersal of Neomyini from late that the Eurasian Neomyini, with its great diver­ Eurasia to North America. sity of forms, represents the parental stock and that a 5. Middle and late Pliocene. Dispersal of Blarinini few forms entered North America before the early from North America to Eurasia. Pliocene and were also separated from the Eurasian 6. Late Pleistocene or Recent. Dispersal of Blarinini population before that time. from North America into adjacent South America; In establishing the subfamily Heterosoricinae, Viret introduction of some crocidurines to Pacific islands and Zapfe (1951, p. 426) suggested that the genus by man. Amblycoptus may belong to that subfamily because KEY TO IIVING AND FOSSIL GENERA OF THE SORICIDAE of mandibular similarities. Although clearly not a [Based on mandible and lower dentition] heterosoricine shrew, Amblycoptus and fossil and liv­ la. Internal temporal fossa of the mandible not pocketed ing Anourosorex do closely parallel Trimylus in the (Subfamily HETEROSORICINAE), p. 7 massive masticatory structures. The suggested simi­ 2a. Entoconid crest present; teeth delicate. Domnina, p. 7 64 SUBFAMILIES AND GENERA OF THE SOKICIDAE

la. Internal temporal fossa of the mandible not pocketed Con. Ib. Internal temporal fossa of the mandible pocketed Con. 2b. Entoconid crest absent. 4a. Interarticular area of mandible prominent lingually Con. 3a. Teeth delicate, cingulum not inflated; talonid of 5a. Teeth not pigmented Continued MS not reduced and with rectangular crest 6c. Mandibular dental formula 1-2-3 Continued through hypoconid and entoconid 8b. Mental foramen beneath protoconid of M\. Paradomnina, p. 10 16a. Cingulumlike groove along medial 3b. Teeth stout to bulbous, cingulum inflated; talonid side of incisor continues above of Ms reduced with crescentic crest through hypo­ notch in basal border of tooth conid ______Trimylus, p. 10 Suncus, p. 21 3c. Teeth stout, cingulum strong but not inflated; Ms 16b. Cingulumlike groove along medial lost______Ingentisorex, p. 15 side of incisor continues below Ib. Internal temporal fossa of the mandible pocketed. notch in basal border of tooth. 4a. Interarticular area of mandible prominent lingually, 17. Talonid of MI moderately com­ or articular facets confluent lingually, and excavated pressed anteroposteriorly; cor- labially. onoid process broad at base 5a. Teeth not pigmented; P4 tetrahedral with promi­ Crocidura, p. 16 nent anterior, posterolateral, and posterolingual 18. Talonid of M\ greatly compressed crests radiating from apex of principal cusp; anteroposteriorly; coronoid range Old World process narrow at base (Subfamily CROCIDURINAE), p. 15 6a. Mandibular dental formula 1-4-3. Diplomesodon, p. 17 "Sorex" dehmi, p. 15 19. First post-incisor antemolar very 6b. Mandibular dental formula 1-3-3. elongate and anteroposteriorly 7a. PS nearly as large as P4 ; talonid of Ms double-cusped unreduced, double-cusped with well- Paracrocidura, p. 20 developed basin; incisor short, curved, 8c. Mental foramen beneath hypoconid of MI, with two sharp, anteriorly inclined cusps size gigantic ______Praesorex, p. 20 on its cutting edge___ Soricella, p. 21 5b. Teeth pigmented at tips; P4 a longitudinal blade or 7b. PS minute; talonid of Ms slightly reduced, tending that way, posterolingual crest from cusp single-cusped but with entoconid crest apex reduced or lost; range North America and weak basin; incisor with two very (Subfamily LIMNOECINAE), p. 24 low, rounded cusps on cutting edge 20a. Posterolabial crest of P4 terminates against Miosorex, p. 18 posterior cingulum midway between pos- 7c. PS minute; talonid of Ms reduced to single terolabial corner of tooth and midline of cusp; incisor short.____ Myosorex, p. 19 tooth-----.-..------Angustidens, p. 25 6c. Mandibular dental formula 1-2-3. 20b. Posterolabial crest of P4 terminates against 8a. Mental foramen beneath P4 to para- posterior cingulum at midline of tooth conid of MI. Limnoecus, p. 26 4b. Interarticular area of mandible prominent labially or 9a. Incisor with prominent scallops on facets confluent labially, interarticular area excavated cutting edge. lingually; P4 with prominent anterior and postero- lOa. Three rounded scallops as in Sorex labial crests radiating from apex of principal cusp but Sylvisorex, p. 22 posterolingual crest essentially lacking, posterolabial lOb. Two anteriorly inclined pointe crest curves lingually and merges with posterior cusps as in Soricella cingulum of tooth enclosing a posterolingual basin Feroculus, p. 18 (Subfamily SORICINAE), p. 27 9b. Incisor not prominently scalloped. 21 a. Condyles of mandible not widely separated and 11. Talonid of Ms virtually lost, no lower condyle not moved greatly forward on cingulum on molars, condyles the mandible; entoconid crest present on very specialized MU -_-_------_-(Tribe SORICINI), p. 29 Solisorex, p. 21 22a. Mandibular condyles confluent. 12. Pt very elongate, condyles united 23a. Posterolabial crest of P4 does not merge Surdisorex, p. 22 with posterior cingulum 13. Talonid of MI anterosposteriorly Crocidosorex, p. 29 compressed, minute PS present 23b. Posterolabial crest of P4 merges with pos­ in some individuals terior cingulum. -----Antesorex, p. 30 Myosorex, p. 19 22b. Mandibular condyles separated.. 14. First antemolar behind incisor 24a. Ms unspecialized, virtually as in Domnina elongate relative to P* as in with distinct entoconid on rectangular Sylvisorex...-Scutisorex, p. 21 hypolophid; pigmented dark red; con­ 15. P* two-rooted siderable variation in pattern of "Sorex" pusilliformis, p. 15 mandibular condyles ...-Sorex, p. 31 KEY TO LIVING AND FOSSIL GENERA OF THE SORICIDAE 65 lb. Internal temporal fossa of the mandible pocketed Con. lb. Internal temporal fossa of the mandible pocketed Con. 4b. Interarticular area of mandible prominent labially or 4b. Interarticular area of mandible prominent labially or facets confluent labially Continued facets confluent labially Continued 21a. Condyles of mandible not widely separated and 21b. Condyles of the mandible widely separated Con. lower condyle not moved greatly forward on the 26c. Lower condyle entirely anterior to lower mandible Continued sigmoid notch and not visible in labial 22b. Mandilar condyles separated Continued aspect Continued 24b. Ma unspecialized, teeth heavy and low 28b. PI not bulbous; entoconid present or absent cusped; pigment orange; pattern of on My when absent hypolophid cres­ mandibular condyles as in Sorex centic; labial cingulum on M\ deep araneus but large, stout, and dorsal but not inflated-____ Blarina, p. 42 condyle extended further back 28c. Pt not bulbous; entoconid and hypolophid Drepanosorex, p. 32 absent on M3 talonid reduced to meta- 24c. M3 slightly reduced with entoconid conid crest; labial cingulum on M\ lacking and crescentic hypolophid; inflated. _____ Shikamainosorex, p. 42 pigmented dark red; Pt somewhat 21c. Condyles of mandible widely separated and lower compressed anteroposteriorly; condyles condyle moved conspicuously forward on the relatively large and close together mandible and located lingually so that groove Microsorex, p. 33 separates lower condyle from lower sigmoid 24d. M3 with greatly reduced talonid; teeth notch, interarticular area narrow; entoconid stout, entoconid crest very prominent crest present on M\ except in Anourosorex and and postentoconid valley very wide Amblycoptus___ (Tribe NEOMYINI), p. 45 on M\. 29a. Upper mandibular articulation with oval 25a. Metalophid of Ms forms median facet. crest with only minor remnant of 30a. Ms unreduced, entoconid present, hypo­ entoconid crest; mandibular con­ lophid rectangular. dyles separated but pattern not 31. Incisor long; pigment conspicuous greatly different than in Antesorex Neomys, p. 46 Alluvisorex, p. 33 32. Incisor short; pigment slight 25b. Metalophid of Ms forms single Episoriculus, p. 47 elongate cusp on talonid, faint 33. Incisor short; pigment variable and trace of entoconid crest; labial light orange; entoconid crest very cingulum on MI deep below pro- low on M\; labial cingulum not con­ toconid; mandibular condyles well tinuous around hypoconid on MI separated___ -.-Petenyia, p. 34 Soriculiis, p. 52 25c. Metalophid of M3 forms single 34. Incisor short to intermediate; pigment conical cusp on talonid, no trace slight to absent; labial shearing of entoconid crest; labial cingulum blade of Pt on longitudinal midline on MI deep below protoconid; of tooth______Nesiotites, p. 51 mandibular condyles widely sepa­ 35. Incisor long; pigment very light but rated. ___ -.-.-Blarinella, p. 34 21b. Condyles of the mandible widely separated and widespread; labial cingulum present lower condyle moved conspicuously forward only below paraconid of MI on the mandible, interarticular area broad; Nectogale, p. 51 entoconid crest lacking on MI 30b. Ms reduced with crescentic hypolophid. (Tribe BLARININI), p. 37 36a. Pigment not visible in ordinary light; 26a. Lower condyle not hidden behind lower labial cingulum weak on MI be­ sigmoid notch in labial aspect. hind protoconid 27a. Mental foramen below protoconid of Chimarrogale, p. 53 MI______-__ Adeloblarina, p. 37 36b. Pigment prominent; labial cingulum 27b. Mental foramen below hypoconid of strong and deep around MI MI-- ______Cryptotis, p. 39 Beremendia, p. 50 26b. Lower condyle meets lower sigmoid notch 30c. M3 reduced with single cusp on talonid and is barely visible in labial aspect Petenyietta, p. 46 "/Sorex" dekneli, p. 42, and 29b. Upper mandibular articulation with triangu­ Paracryptotis, p. 41 lar facet. 26c. Lower condyle entirely anterior to lower sigmoid notch and not visible in labial 37a. M 3 unreduced. aspect. 38a. Pigment prominent; entoconid crest 28a. ?4 bulbous; entoconid present on MS; of MI strong_ ..-Neomys, p. 46 labial cingulum on M\ inflated 38b. Pigment slight; entoconid crest of Blarinoides, p. 44 MI very low._ Megasorex, p. 56 66 SUBFAMILIES AND GENERA OF THE SORICIDAE

Ib. Internal temporal fossa of the mandible pocketed Con. Bate, D. M. A., 1945, Pleistocene shrews from the larger western 4b. Interarticular area of mandible prominent labially or Mediterranean islands: Annals Mag. Nat. History, ser. 11, facets confluent labially Continued v. 11, p. 738-769 (art. 63), 8 figs. 21c. Condyles of mandibles widely separated Con. Bobrinsky, N. A., Kuznetsov, B. A., and Kuzyakin, A. P., 29b. Upper mandibular articulation with triangular 1965, Opredelitel Melkopitayushchikh S.S.S.R. [Guide to facet Continued the mammals of U.S.S.R.]: Moscow, Prosveshchenie, 382 37a. M3 unreduced Continued p., 40 pis., Ill maps, 129 figs. 38c. Pigment slight; entoconid crest of M i Brown, Barnum, 1908, The Conard Fissure, a Pleistocene bone strong_____._ --..Chodsigoa, p. 48 deposit in northern Arkansas; with description of two new 37b. M3 reduced. genera and twenty new species of mammals: Am. Mus. 39a. Entoconid lacking but entoconid Nat. History Mem., v. 9. pt 4, p. 157-208, 11 pis. crest and talonid basin present Brunner, Georg, 1952, Die Markgrabenhohle bei Pottenstein on M3; very little pigment. (Oberfranken); Neues Jahrb. Geologic u. Palaontologie Monatsh. 1952, no. 9, p. 457-471, 4 figs., 7 tables. 40a. Tip of coronoid process mas­ Butler, P. M., and Hopwood, A. T., 1957, Insectivora and sive, coronoid spicule large Chiroptera from the Miocene rocks of Kenya Colony, no. Chodsigoa, p. 48 13 of Fossil mammals of Africa: British Mus. (Nat. 40b. Tip of coronoid process slight, History), 35 p., 10 figs. coronoid spicule small Chasen, F. N., 1940, A handlist of Malaysian mammals. A sys­ Notiosorex, p. 55 tematic list of the mammals of the Malay Peninsula, Suma­ 39b. Entire talonid of M 3 virtually lack­ tra, Borneo, and Java, including the adjacent small islands: ing. ______Anourosorex, p. 53 Bull. Raffles Mus. Singapore, Straits Settlements, no. 15, 209 p.,1 map. 39c. M 3 lacking.-.--.Amblycoptus, p. 54 Cope, E. D., 1873, Third notice of extinct vertebrata from the 4c. Mandibular articulations widely separated and upper Tertiary of the plains: Paleont. Bull. [Philadelphia], articulation placed very far to the rear; teeth unpig- no. 16., 8 p. mented and high cusped with carnivorelike appearance; 1899, [Posth.] Vertebrate remains from Port Kennedy MI with trigonid very long, talonid very short, para- bone deposit: Philadelphia Acad. Nat. Sci. Jour., v. 11, lophid oriented essentially anteroposteriorly, metaconid p. 193-286, 4 pis. close to protoconid, metalophid unites protoconid to Crusafont Pair6, Miguel, 1957, Sobre la f6rmula dentaria de hypoconid, hypolophid rectangular with no post ento­ Heterosorex sansaniensis (Lartet): Inst. "Lucas Mallada" conid valley, entoconid crest lacking and talonid basin Cursillos y Conf., no. 4, p. 77-80, 4 figs. opens lingually between metaconid. and entoconid Depe"ret, Charles, 1892, La faune de mammif£res miocenes de la Grive-Saint-Alban (Isere) et de quelques autres locality's (Subfamily ALLOSORICINAE), p. 57 du bassin du Rh6ne Documents nouveau et revision 41a. Mandibular condyles not extremely separated; ge'ne'ral: Mus. des Sciences nat. Lyon Archives, v. 5, mem. labial cingulum present on cheek teeth 2, p. 1-95, 4 pis. "Sorex" gracilidens, p. 57 Doben-Florin, Ursula, 1964, Die Spitzmause aus dem alt- 41b. Mandibular condyles extremely separated; labial Burdigalium von Wintershof-West bei Eichstatt in Bayern: cingulum lacking on cheek teeth Bayer. Akad. Wiss. Abh., Neue Folge, v. 117, 82 p., 11 figs., Allosorex, p. 57 7 pis. Ellerman, J. R., and Morrison-Scott, T. C. S., 1951, Checklist REFERENCES of Palaearctic and Indian mammals 1758 to 1946: British Authors of names for living taxa have been omitted from this list and may be Mus. (Nat. History), London, 180 p., 1 fig. obtained from the following works: Alien, 1939 (Africa); Bobrinsky, Kuznetsov, and Ellerman, J. R., Morrison-Scott, T. C. S., and Hayman, R. W., Kuzyakin, 1965 (U.S.S.R.); Chasen, 1940 (Malaysia); Ellerman and Morrison-Scott, 1953, Southern African mammals 1758 to 1951 a reclassi- 1951 (Palearctica and India); Ellerman, Morrison-Scott, and Hayman, 1953 (southern Africa); Hall and Kelson, 1959 (North America); Miller, 1912 (Europe); and Miller fication: British Mus. (Nat. History), London, 363 p., 3 figs. and Kellogg, 1955 (North America). Fejfar, Oldfich, 1966, Die plio-Pleistozanen Wirbeltierfaunen von Hajna'cka und Ivanovce (Slowakei), CSSR; V, Alien, G. M., 1939, A checklist of African mammals: Mus. Allosorex stenodus n.g. n.sp. aus Ivanovce A: Neues Comp. Zool., Harvard Univ., Bull., v. 83, 763 p. Jahrb. Geologic und Palaontologie Abh., v. 123, no. 3, Alien, J. A., 1917, The skeletal characters of Scutisorex Thomas: p. 221-248, 14 figs., 5 pis. Am. Mus. Nat. History Bull., v. 37, p. 769-784, 4 pis. Findley, J. S., 1953, Pleistocene Soricidae from San Josecito A'rnback-Cristie-Linde, A., 1912, On the development of the Cave, Nuevo Leon, Mexico: Kansas Univ. Pub. Mus. Nat. teeth of the Soricidae an ontogenetical inquiry: Annals History, v. 5, p. 663-669. Mag. Nat. History [London], ser. 8, v. 9, p. 601-623, 2 pis., Friant, Madleine, 1947, Les musaraignes (Soricidae) quaternaires 9 figs. et actuelles de PEurope occidental. Characteres dentaires et oste"ologiques, base de leur classification: Soc. Ge"ol. Balsac, Henri Heim de, 1940, Un soricide" nouveau du Pl&sto- Nord Annales, v. 67, p. 222-269, 23 figs. cene; considerations pal<5obioge'ographiques: Acad. Sci. Gaillard, Claudeus, 1915, Nouveau genre de musaraignes dans [Paris] Comptes Rendus, v. 211, p. 808-810, 1 fig. les de"p&ts Miocenes de la Grive-Saint-Alban (Isere): Soc. Balsac, H. Heim de, and Lamotte, M., 1956 and 1957, Evolution linneene Lyon Annales, v. 62, p. 83-98, figs. et phyloge"nie des Soricide"s africains: Mammalia [Paris], Galbreath, E. C., 1953, A contribution to the Tertiary geology v. 20, p. 140-167, 9 figs., v. 21, p. 15-49, figs. 10-20. and paleontology of northeastern Colorado: Kansas Univ. REFERENCES 67

Paleont. Contr., no. 13, Vertebrata, art. 4, 119 p., 26 figs., James, G. T., 1963, Paleontology and nonmarine stratigraphy 2 pis. of the Cuyama Valley Badlands, California; Part 1, Geology, Gazin, C. L., 1933, A new shrew from the upper Pliocene of faunal interpretations, and systematic descriptions of Chirop- Idaho: Jour. Mammalogy, v. 14, no. 2, p. 142-144, 1 fig. tera, Insectivora, and Rodentia: California Univ. Pub. Gulino, Guiseppe, 1939, Nota di mammalogia Gli Insectivori Geol. Sci., v. 45, p. 1-154, 53 figs., 8 pis. Italian!: Mus. Zoologia Anatomia comp. Torino Boll., Kindahl, M., 1960, Some aspects of the tooth development in v. 47, ser. 3, p. 131-155. Soricidae: Acta odont. scandinavica, v. 17, p. 203-237. Hall, E. R., 1930, A new genus of bat from the later Tertiary of Kormos, Theodor, 1926, Amblycoptus oligodon, n.g. und n. sp. Nevada: California Univ. Dept. Geol. Sci. Bull., v. 19, eine neue Spitzmaus aus dem ungarischen Pliozan: An- no. 14, p. 319-320, 1 pi. nales hist.-nat. Mus. natl. hungarici, v. 24, p. 352-370, 1 pi. Hall, E. R., and Kelson, K. R., 1959, The mammals of North 1934, Neue Insektenfresser, Fledermause, und Nager America: New York, The Ronald Press Co., 2 vols, 1083 p., aus dem Oberpliozan der Villanyer Gegend: Foldtani 79 p. of index, 553 figs. kozlony, v. 64, p. 296-321, 9 figs. Hasegawa, Yoshikazu, 1957, On a new Insectivora from the 1935, Die perlzahnige Spitzmaus (Sorex margaritodon upper Kuzuii formation in Japan: Yokohama Natl. Univ. Korm.) und das Anpassungsproblem: Allattani k6zle"m- Sci. Repts., sec. 2, no. 6, p. 65-69, 2 figs., 12 pis. enyek, v. 32, p. 61-79, 3 figs. Heller, Florian, 1930, Eine Forest-Bed-Fauna aus der Sack- Kowalski, Kazimierz, 1956, Insectivores, bats, and rodents from dillinger Hohle (Oberpfalz): Neues Jahrb. Mineralogie, the early Pleistocene bone breccia of Podlesice near Kroczyce Geologie, u. Palaontologie, Abt. B., v. 63, p. 247-298, 27 (Poland): Acta pa'aeont. polonica, v. 1, p. 331-394, 2 figs., figs., 4 pis. 4 pis. 1936, Eine Forest-Bed, Fauna aus der Schwabischen 1958, An early Pleistocene fauna of small mammals from Alb: Heidelberger Akad. Wiss. Sitzungsber.; no. 2, p. 1-29, the Kadzielnia Hill in Kielce (Poland): Acta palaeont. 15 figs, polonica, v. 3, p. 1-47, 24 figs. 1958, Eine neue altquartare Wirbeltierfauna von -1960, Pliocene insectivores and rodents from Rebielice Erpfingen (Schwabische Alb): Neues Jahrb. Geologie u. Kr61ewskie (Poland): Acta zool. cracoviensia, v. 5, no. 5, Palaeontologie Abh., v. 107, p. 1-102, 75 figs., pis. 1-3, p. 155-194, 8 figs., 3 pis. 3 tables. Kowalski, Kazimierz, and Li, Chuan-Kuei, 1963, A new form of 1963, Eine altquartare Wirbeltierfauna des unteren Soricidae (Insectivora) from the Pleistocene of North China: Cromerium aus der no-dlichen Frankenalb: Neues Jahrb. Vertebrata palasiatica, v. 7, no. 2, p. 138-143, 1 fig. Geologie u. Palaontologie Abh., v. 118, no. 1, p. 1-20, 10 figs. Kretzoi, Mikl6s, 1938, Die Raubtiere von Gombaszog nebst Hensel, R. F., 1855, Beitrage zur Kenntniss fossiler Saugethiere; einer Ubersicht der Gesamtfauna (ein Beitrag zur Strati- Insectenfresser und Nagethiere der Diluvialformation: graphic des Altquartaens): Annales hist.-nat. Mus. natl. Deutsch geol. Gesell. Zeitschr., v. 7, p. 458-501, 1 pi. hungarici, v. 31, p. 88-157, 5 figs., 3 pis. Hibbard, C. W., 1937, An upper Pliocene fauna from Meade 1941, Weitere Beitrage zur Kenntnis der Fauna von County Kansas: Kansas Acad. Sci. Trans., v. 40, p. 239-265, Gombaszog: Annales hist.-nat. Mus. natl. hungarici, v. 34, 5 pis., 2 figs. p. 105-139, 2 figs., 1 pi. 1943, The Rezabek Fauna, a new Pleistocene fauna from 1943, Bemerkungen iiber Pete"nyia: Foldtani kozlony, Lincoln County, Kansas: Kansas Univ. Sci. Bull., v. 29, v. 73, p. 607-608, 1 fig. pt. 2, no. 2, p. 235-247, 3 figs., 1 pi. 1956, Die altpleistozanen Wirbeltier-Faunen des Villdnyer 1944, Stratigraphy and vertebrate paleontology of Gebirges: Geologica hungarica, Ser. palaeont., v. 27, p. 1- Pleistocene deposits of southwestern Kansas: Geol. Soc. 264, 4 figs., 1 table. America Bull., v. 55, no. 6, p. 707-754, 20 figs., 3 pis. 1959a, Insectivoren, Nagetiere, und Lagomorphen 1950, Mammals of the Rexroad formation from Fox der Jurigstpliozanen Fauna von Csarnota im Villanyer Canyon, Meade County, Kansas: Michigan Univ. Mus. Gebirge (Sudtfngarn): Vertebrata hungarica, v. 1, no. Paleont. Contr., v. 8, no. 6, p. 113-192, 23 figs., 5 pis., 9 tbls. 2, p. 137-246. 1953, The insectivores of the Rexroad fauna, upper 1959b, New names for soricid and arvicolid homonyms: Pliocene of Kansas: Jour. Paleontology, v. 27, no. 1, p. 21-32, Vertebrata hungarica, v. 1, no. 2, p. 247-249. 5 figs. 1962, Fauna und Faunenhorizont von Csarn6ta: Hun­ 1956, Vertebrate fossils from the Meade formation of garian Geol. Inst. Ann. Rept., 1959, p. 344-395, 10 figs., southwestern Kansas: Michigan Acad. Sci. Papers, 1955, 5 pis. v. 41, p. 145-203, 2 pis., 16 figs. Lartet, Edouard, 1851, Notice sur la colline de Sansan: Annu- 1957, Notes on late Cenozoic shrews: Kansas Acad. Sci. aire de"pt. Gers, Auch, 42 p. Trans., v. 60, no. 4, p. 327-336, 3 figs. Lavocat, Rene", 1951, Revision de la faune des mammiferes 1958, Summary of North American Pleistocene mam­ oligocenes d'Auvergne et du Velay: Paris, "Sciences et malian local faunas: Michigan Acad. Sci., Arts and Letters, Avenir," 153 p., 26 pis., 1 fig., map. Papers, 1957, v. 43, p. 3-32, 1 table. Macdonald, J. R., 1947, New shrew from the Niobrara River 1963, A late Illinoian fauna from Kansas and its climatic upper Miocene of Nebraska: Am. Jour. Sci., v. 245, p. 123- significance: Michigan Acad. Sci., Arts and Letters, Papers, 126, 2 figs. v. 48, p. 187-221, 8 figs. 1951, The fossil Vertebrata of South Dakota, in Soc. Hinton, M. A. C., 1911, The British fossil shrews: Geol. Mag., Vertebrate Paleontology [Guide Book] 5th Field Conf., ser. 5, v. 8, p. 529-539, 11 figs., 1 pi. Aug.-Sept. 1951: p. 63-74. Hutchison, J. H., 1966, Notes on some upper Miocene shrews 1963, The Miocene faunas from the Wounded Knee from Oregon: Oregon Univ. Mus. Nat. History Bull., No. area of western South Dakota: Am. Mus. Nat. History 2, 23 p., 17 figs. Bull., v. 125, art. 3, p. 139-238, 30 figs., 2 maps, 31 tables. 68 SUBFAMILIES AND GENERA OF THE SORICIDAE

McDowell, S. B., Jr., 1958, The Greater Antillean insectivores: Romer, A. S., 1928, Pleistocene mammals of Algeria. Fauna of Am. Mus. Nat. History Bull., v. 115, art. 3, p. 115-214, the paleolithic statron of Mechta-el-Arbi: Logan Mus. 46 figs. Bull., v. 1, no. 2, p. 80-163. Mawby, J. E., 1960, A new occurrence of Heterosorex Gaillard: Schlosser, Max, 1887, Die Affe, Lemuren, Chiropteren, In- Jour. Paleontology, v. 34, no. 5, p. 950-956, 2 figs. sectivoren, Marsupialer, Creodonten und Carnivoren des Meester, J., 1953, The genera of African shrews: Annals Trans­ europaischen Tertians und deren Beziehungen zu ihren vaal Mus., v. 22, p. 205-214. lebenden und fossilen aus europaischen Verwandten: 1955, Fossil shrews of South Africa: Annals Transvaal Palaontologie Geologic Osterreich Ungarns Beitr. v. 6, Mus., v. 22, p. 271-278, 2 figs., 1 table. pts. 1 and 2, p. 1-227, 9 pis. Mein, M. Pierre, 1958, Les mammife'res de la faune sidero- 1924, Tertiary vertebrates from Mongolia: Palaeonto­ lithique de Vieux-Collonges: Mus. d'histoire nat. de Lyon logia sinica, ser. C, v. 1, no. 1, p. 1-119, 5 figs., 6 pis. Nouv. Archives, v. 5, p. 1-122, 172 figs. Shikama, Tokio, and Hasegawa, Yoshikazu, 1958, On a new An- Miller, G. S., 1912, Catalogue of the mammals of western ourosorex from the Ryugasi formation (fissure deposits) in Europe (Europe exclusive of Russia) in the collection of the Japan: Yokohama Natl. Univ. Sci. Rept., sec. 2, no. 7, p. British Museum: British Mus. (Nat. History), London, 105-112, pi. 16, 1 map. 1019 p., 213 figs. Simpson, G. G., 1941, A new Oligocene insectivore: Am. Mus. 1927, Revised determinations of some Tertiary Mammals Nat. History Novitates, no. 1150, p. 1-3, 1 fig. from Mongolia: Palaeontologia sinica, ser. C, v. 5, no. 1945, The principals of classification and a classification 2, p. 5-20. of mammals: Am. Mus. Nat. History Bull., v. 85. 350 p. Miller, G. S., and Kellogg, Remington, 1955, List of North Stehlin, H. G., 1940, Zur Stammesgeschichte der Soriciden: American Recent Mammals: U.S. Natl. Museum Bull. Eclogae geol. Helvetiae, v. 33, p. 298-306, 5 figs. 205, 954 p. Stirton, R. A., 1930, A new genus of Soricidae from the Barstow Patterson, Bryan, and McGrew, P. O., 1937, A soricid and two Miocene of California: California Univ. Dept. Geol. Sci. erinaceids from the White River Oligocene: Field Mus. Bull., v. 19, no. 8, p. 217-228, 2 figs. Nat. Hist., Geol. Ser. 6, p. 245-272, 15 figs. Stromer, Ernst, 1928, Wirbeltiere im obermiocanen Flinz Miin- Pasa, Angelo, 1948, I mamiferi di alcune antiche brecce veronesi: chens: Bayer. Akad. Wiss. Abh., Math. nurw. Abt., v. 32, Mus. civico storia nat. Verona Mem., v. 1, p. 1-111. no. 1, p. 1-71, 3 figs., 3 pis. Pei, W. C., 1936, On the mammalian remains from locality 3 at Sulimski, Andrzej, 1959, Pliocene insectivores from We.ze: Acta Choukoutien: Palaeontologia sinica, ser. C, v. 7, no. 5, palaeont. polonica, v. 4, no. 2, p. 119-175, 5 figs., 4 pis. p. 1-120, 59 figs., 6 pis. 1962, Supplementary studies on the insectivores from Pete"nyi, S. J., 1864, A Beremendi Me"szkobdnya terme"szetrajzes We.2e I (Poland): Acta palaeont. polonica, v. 7, nos. 3-4, 6s osle'nytanilag Pete"nyi Salamon Altai Leirva: H£traha- p. 441-502, 2 figs., 2 pis. gyott Munkdi, Magyar Tudomdnyos Akademia, Budapest, Suthard, J. A., 1966, Stratigraphy and paleontology in Fish no. 1, pt. 2, p. 37-81, 2 pis. Lake Valley, Esmeralda County, Nevada: California Univ., Peterson, O. A., 1926, The fossils of the Frankstown Cave, Blair Riverside, M.A. thesis. County, Pennsylvania: Carnegie Mus. Annals, v. 16, no. 2, Tedford, R. H., 1961, Clarendonian Insectivora from the Ricardo p. 249-314, 10 figs., 9 pis. Formation, Kern County, California: Southern California Phillips, W. W. A., 1928, Ceylon shrews: Spolia Zeylanica, ser. 14, Acad. Sci. Bull., v. 60, pt. 2, p. 57-76, 4 figs. v. 1, p. 295-331. Thomas, Oldfield, 1906, New African mammals of the genera Pidoplichko, I. G., 1953, Amvrosievskaia paleoliticheskaia Ceropithecus, Scotophilus, Miniopterus, Crocidura, Georychus, stoianka i ee osobennosti [The Amvrosiev Paleolithic station and Helipohobius: Annals Mag. Nat. History, ser. 7, v. 17, and it particularities]: Karatkie Soobshcheniya Inst. p. 173-179. Arkheologii Akad. Nauk URSR, v. 2, p. 65-68. 1913, On African bats and shrews: Annals Mag. Nat. Pomel, Auguste, 1853, Catalogue me"thodique et descriptif des History, ser. 8, v. 11, p. 317-321. verte'bre's fossiles de"couverts dans les bassins de la Loire et -1924, A new genus and species of shrew from Ceylon: de FAllier (suite et fin): Annales sci., litt., indus. [Auvergne], Spolia zeylanica, v. 13, pt. 1, p. 93-95. v. 26, p. 81-229. Tobien, Heinz, 1939, Die Insektenfresser und Nagetiere aus der Reed, K. M., 1960, Insectivores of the middle Miocene Split aquitanen Spaltenfullung bei Tomerdingen (Ulmer Alb): Rock local fauna, Wyoming: Harvard Coll. Mus. Comp. Naturf. Gesell. Freiburg Ber., v. 36, p. 159-180, 1 pi. Zoology, Breviora, no. 116, 15 p., 2 pis. Viret, Jean, and Zapfe, Helmuth, 1951, Sur quelques Soricides Repenning, C. A., 1965, An extinct shrew from the early Pleisto­ miocenes: Eclogal. geol. Helvetiae, v. 44, no. 2, p. 411-426, cene of South Africa: Jour. Mammalogy, v. 46, no. 2, p. 12 figs. 189-196. Wilson, R. W., 1960, Early Miocene rodents and insectivores Repenning, C. A., and Vedder, J. G., 1961, Continental verte­ from northeastern Colorado: Kansas Univ. Pub., Paleon. brates and their stratigraphic correlation with marine Contr., Vertebrata, art. 7, p. 1-92, 131 figs. mollusks, eastern Caliente Range, California, in Short -1963, Notes on North American species of Heterosorex papers in the geologic and hydrologic sciences: U.S. Geol. Gaillard: S. Dak. Acad. Science Proc., v. 42, p. 79-83, Survey Prof. Paper 424-C, p. C235-C239, 2 figs., 1 chart. 2 figs. Roger, Otto, 1885, Kleine palaontologische Mitteilungen; II, Young, C. C., 1934, On the Insectivora, Chiroptera, Rodentia, Saugethierreste aus der Reischenau (Zusamthal in Sch- and primates other than Sinanthropus from Locality 1 at waben): Naturw. Ver. Schwaben u. Neuburg Ber., v. 28, Choukoutien: Palaeontologia sinica, ser. C, v. 8, no. 3, 93-118, 3 figs. p. 1-160, 51 figs., 10 pis. REFERENCES 69

Zapfe, Helmuth, 1951, Die Fauna der miozanen Spaltenfullung Zinndorf, Jakob, 1901, Mitteilungen iiber die Baugrube des von Neudorf a.d. March (CRS); Insectivora: Osterreich. Offenbacher Hafens. Ein Beitrag zur geologischen und Akad. Wiss. [Vienna] Sitzungsber., Math, naturw. Kl. ser. 1, palaeontologischen Kenntnis der Cyrenenmergelschichten im v. 160, no. 5, p. 449-480, 1951. nord-ostlichen Teile des Mainzer Beckens, nebst einem Zdansky, Otto, 1928, Die Saiigetiere der Quartarfauna von Fundlbericht iiber bearbeitete Baumstamme aus prae- Chou-k'outien: Palaeontologia sinica, ser. C, v. 5, no. 4, historischer Zeit: Offenbacher Vereins Naturkunde Ber., p. 1-146, 16 figs., 16 pis. no. 42, p. 87-146, 3 figs., 4 pis.

INDEX

[Italic page numbers indicate both major references and descriptions]

A Page Page adamsi, Cryptotis...... 39,40,flg. 26 Brachysorex. See Cryptotis. Cryptotis--...... 37,39,40,41,55,62,63 Adeloblarina.... 36, 57, 38, 39, 41, 42, 43, 44, 55, 62 breaicauda, Blarina...... 25,38,42,43,44, fig. 29 adamsi.______39,40, fig. 26 berklandi...... 37.38, fig. 25 browni, Neomys...... 46 kansasensis-...... ______. 40 Africa.-- . 2, (50,62 Burma.. 62 magna...... 40 Allopachyura...... 47,58 meadensis ...... _____ 40 Allosorei--...... __ 67 mexicana..-- ______41 stenodus---...... 6,57 caeruleus, Suncus...... 20,22,fig. 13 parva...... 38, 39, 40, 41, fig. 27 Allosoricinae.. ______2,5,57 California-.- 9,25,26,34,55 cudahyensis, Sorex...... 32 Alluvisorex...... S3 castellarina, Neomys...... 46 Cusps, height. -----______-_____ 6 arcadentes...... 33,34 caudatus, Episoriculus...... 48,fig. 32 Czechoslovakia- ... 32,33,34,46,51,57 chasseae...... _____...._.. 33, 34, fig. 22 Sorex...... ____ 47 alpinoides, Sorex...... 32 Ceylon.. . 2,18,21,57,60 D alpinus, Sorex...... 24 chasseae, Alluvisorex...... 33,34,fig. 22 dakotensis, Trimylus- ______12,13, fig. 6 ambiguus, Sorex...... 30 Chimarrogale...... 24,37,51,53,54 daubentonii, Sorex...... ______46 Amblycoptus.-...... 27,54,63 himalayica...... 53,fig. 38 dehmi, "Sorex"...... 15,16,24 oligodon ...... 54,55 platycephala.-...... 53 dehneli, "Sorex"...... 42,44,63 Amphisorex. See Sorex or Neomys. China--._ 17,35,45,48,50,51,52,53,54,62 Deinsdorfia...... -______57 Ancestry of the Soricidae.______.. 58 Chiroptera..-_-_. ._. __ 5,58,59 delphinensis, Heterosorex...... ____ 14 Angustidens.-. ______25,33,60 Chodsigoa...... 45,48,49,50,52,53,54,55,56 densicingulata, Shikamainosorex...... _____ 42 vireti...... 25, 26, 30, fig. 16 bohlini.. ______50 Dental formulae of the soricids. -.-_____ 2 Anotus. See Blarina. hypabia...... 49,50,fig. 33 Diplomesodon...... 17,21 Anourosorex...- 2, 14, 24, 42, 45, 48, 50, 52, 63, 54 salentkii...... 49,50,55,56,fig. 34 fossorius...... 6,18,21,63 assamensis...... 54 cinereus, Sorex. ______27,32 pulchellum...... 18, fig. 9 inexpectatus...... 54 collongensis, Sorex -...... 24 discrepans, Soricella...... 15,21 japonicus...... 54 Colorado.- 8,11,13,25,31 .dixonensis, Sorex.-...... ___ 32 squamipes______.. _ 53,54, fig. 39 compressa, Domnina.. ______11 Domnina---...... _____.. 3, Antemolar, definition...______..__ 3 compressus, Antesorez...... 31,fig. 19 5, 6, 7, 8, 9, 10, 11, 13, 15, 26, 27, 29, Antesorex...... 30,31,33,36,60,62 Heterosorex..-...... ______11 31, 33, 52, 58, 60. compressus.-...... _ 31, fig. 19 Sorex...... 30,31 compressa...... 11 antiguus, Crocidosorex...... 15,30,31,58 Trimylus- _____...... __.. 11,12,13, fig. 5 gradata.. .. 3, 5, 8, 9, 10, 11, 13, 14, fig. 3 Sorex...... 29.30 Condyle, mandibular______._ 5 greeni ...... ______.--.-. 9 araneoides, Sorex...... 32 basic patterns..-______..-. 6,fig. 1 thompsoni...... 5, 8, 9, 10, 12, 13, fig. 2 araneus, Sorex...... 26,28,29,31,fig. 20 function______28 n. sy...... 9,10, fig.4 arcadentes, Alluvisorex...... 33,34 Congosorex. See Myosorex. Drepanosorex. -...... 32 Archaeolagus...... ______..__.. 9 Coronoidspicule--..____ 37,45,55, figs. 25, 29, 31 margaritodon...... 33 Arkansas. .._____.______.. 33 Correlation of the Soricidae____.__.... 69 pachyodon.-..______.... 33 Asoriculus...... 58 Corsica______51 savini...... 33 gibberodon...... 48 corsicanus, Nesiotites...... -... 51 tasnadii...... 33 assamensis, Anourosorex...... 54 Cosira. See Sorex. (Drepanosorex) tasnadii, Sorex...... 32 Atavism.. -.. -- ..__...... 3,31 nigrescent...... 52 Atophyrax. See Sorex. crawfordi, Notiosorex...... 49, 55, fig. 40 E Austria-__....._ . 32,33 Sorex (Notiosorex)------...... 55 elegans, Nectogale...... 51,52, fig. 36 Crocidosorex...... 29,30,31,36,58,60,62 Embryologic studies .. . 2 Endemism...... __.. 59,60,62 antiquus______._____ 15,30,31,58 Beremendia...... 50,51,52,55 piveteaui...... 2,29,30 England. 32,33,46 fissidens ...... 51, 54, fig. 35 Crocidura...... 2,6,15, 16, 17,18,20,22,23,47,62 Entoconidcrest..----.. --. 7, 10, 28, 29, figs. 2,3 sinmsit ...... 51 gibberodon.---...... 48 Episoriculus...... ______.______47,48,58 ucrainica...... 51 goliath...... 20 caudatus.. -...... 48, fig. 32 berklandi, Adeloblarina...... 37,38, fig. 25 kormosi______...... 2,35 gibberodon---...... 48 Blarina...... 29,35, Erinaceoidea...______.. .. 59 37,39,40,41, 42, 44,45,50,51,52,53,54,55,62,63 kornfeldi______...-..-. 17 brericauda 25,38,42,43,44, fig. 29 leucodon..______._..._ 17 etruscus, Suncus...... ______.... 22 fossilis...... _____.______44 morio______22 Eurasia..______60 Oidleyi...... 41,42,43,44,fig. 30 obtusa.-...... 17 Europe._....____._-_-_..._.. 60,62,63 simplicidens...... 44 russula-. ______... fig. 8 Evaluation of characters______3 blarina zinki, Myosorex...... 22 suaveolens.-.-..... ______...___ 17 External temporal fossa_- 29, 37, 45, figs. 25, 29, 31 Blarinella...... 29,33, 34, 35,36,38,44,45,48,50,55,62 taungsensis... ______.. 17 kormosi...... 35,45 wongi...... ______.. 17 zorzii-...... 17 quadraticauda______34,35, fig. 24 fallax, Sorex 32 sp_ ...... 15,17,24 wardi...... 45 Faunal exchanges, intercontinental. _____ 63 Blarinini- . . -- 29,35, Crocidurinae.. . 1,2,3,5,6,7, 16, 24,25,59,60,62,63 African aflBnities ____.... _... 15,60 Feroculus...... 2,18,20,23 36,37,38,39,42,44,45,52,53,55,57,62,63 feroculus, Sorex.. ______. 18 incertae sedis______. 44 incertae sedis.______.... 23 fissidens, Beremendia...... 51, 54, fig. 35 Blarinoides...... 44,51,62 Crossogale. See Chimarrogale. Crossopus...... 50 mariae...... 42,44 Crossopus. See Neomys. bohlini, diodsigm...... 50 fissidens...... ______.__. 50 fodiens, Neomys...... 38,46, fig. 31 Neomys..-...... 50 himalayanus...... 53 Fossa pterygoidea. ______2,5 71 72 INDEX

Fossil localities: page I Page Page Beatty Butte, Oreg ______..... 13 Idaho 44 Microsorex...... 26,33 Bowden Hills, Oreg...__.____..._ 38 Illustrations, contents and preparation.-.__ 2 hoyi...... 30,31, fig. 21 Choukoutien, China.... 50,51 India 48,53,60,62 minutus-...... 33 Christmas Valley, Oreg-----______40 Indochina.______62 pratensis ...... 33 Cuyama Valley, Calif .______. 26 inexpectatus, Anourosorex...... 54 Miniopterus...... 5 Ertemte, Mongolia.____ 35,54 Inferior pterygoid fossa______fig. 11 Minorca______51 Flint Hill, S. Dak. .. . 12 Inferior sigmoid notch______... fig. 8 minutissimus, Sorex...... 32 Quano Ranch, Oreg-______13 Ingentisorex...... _.-..._-..-..-- 3,7,15,60 minutus, Microsorex...... 33 Hagennan, Idaho...______--___ 44 tumidtdens...... 15 Sorex...... 32 Ivanovce, Czechoslovakia.-..._____ 57 intermedius, Neomys...... 46 Miosorex...... 18 La Qrive St. Alban, France.--..._-.-.. 14,19 Internal temporal fossa of the mandible...__ 2, grivensis--- ...... 15,16,19,33, Bg. 10 Montaigu-le-Blin, France______30 5, figs. 2,10 Miothen...... 58 Polgardi, Hungary..-_____ ... 54 Intertemporal fossa... . . 2,5 Mongolia ...___._____-...._.__ 35,54 Rexroad, Kansas...______.____ 40,41 Italy--.-...... 17,33,46 Montana______8 Rome, Oreg--______26,41 mono, Crocidura...... 22 Snyder Creek, Oreg--_.______13 Vallecito Creek, Calif.-- .. 55 jacksoni, Notiosorex...... 48,49,55 Morphologic characteristics, Allosoricinae. _ 57 Blarinini...... 37,62 Valles-Penedes, Spain...... 14 Japan.--...... 42,54,63 Vieux-Collonges, France____------__ 16 japonicus, Anourosorex...... _. 54 Crocidurinae-. -___..-.-..__... 2,15,60 Weisenau, Germany______12,30 Heterosoricinae______7,60 We.ze Breccia, Poland-,__.... 36,37,45,46,47 K hypothetical shrew ancestor..___ ... 58 Wintershof-West, Germany.-... 12,15,21,24.36 Kansas... -. . - 32,33,40,41,44,55 Limnoecinae-______24,60 fossilis, Blarina. ______44 kansasensis, Cryptotis...... 40 Neomyini...^______45,63 fossorius, Diplomesodon...... __- 6,18,21,63 kennardi, Sorex...... 32 shared with insectivores and bats 59 Soricinae. ______....___.. 27 France--...... 14,16,19,29,30,32 Kenya.... 17,22 frankstownensis, Sorex_____..._.___. 32 kinkelini, "Sorex"...... 7,12 Soricini...... __.. 29,62 subfamilies-..._ ____... 3 kormosi, Blarinella...... 35,45 G Crocidura...... 2,35 Musaraneus. See Sorex. kornfeldi, Crocidura...... 17 Mydecodon martini...... 5 Germany ___ -...--. 12,15,21,24,30,32,33,36,46,57 kretzoii, "Sorex"...... 35,42,44,45,63 Myosictis. See Neomys. gibber odan, Asoriculus ______. 48 kubinyi, Soriculus______..._ 4§ Mysorex-...... 19,22 Crocidura...... 48 blarina zinki...... 22 Episoriculus ...... 48 kura, Podihik...... 57 potti...... 19 Soriculus...... 48 robinsoni...... 20 gidleyi, Blarina...... 41, 42, 43, 44, fig. 30 varius...... 19, fig. 11 gigas, Megasorex...... - 56, fig. 41 lacustris, Neosorex...... 32 Mystipterus...... 5 Noliosorex...... 56 Sorex...... 32 goliath, Crocidura...... 20 leahyi, Sorex...... 32 N Praesorex...... 20,21,fig. 12 Leptacodon...... ------4 Nebraska 8,11,25,27 gracilidens, "Sorex"...... 57 Leucodmi. See Crocidura. Nectogale...... 5,24,50,51,52,54 gracilis, Petenyiella ...... 46,47 leucodon, Crocidura--. __ 17 elegans...... 51,52,fig. 36 Saturninia ...... ___ 4 Sorex. ______16 neglecta, Petenyia...... 34 Sorex ...... ___ 46 Leucorrhynchus. See Neomys. Neomyini 20,28,36,37,38,39, gradata. Domnina...... 3,5,8,9,10, 11, 13, 14, fig. 3 Limnoecinae- - - . 2, 3, 5, 6, 15, S4, 40, 59, 60, 63 46, 46,47,48,49,50,51,52,53,55,57,58,60,62,63 greeni, Domnina. ______9 parallel evolution ______60 incertae sedis______57 grieensis. Miosorex...... 15, 16, 19, 33, fig. 10 Limnoecus ...... 24,26,60 Neomys... 13,28,36,37,45, 46, 47,48,50,51,52,53,54,57 Sorex...... 19 bohlini ...... 50 pusittus...... 18,19 micromorphus ..... ______. __ 23,24 niobrarensis ...... 5, 6, 23, 24, 26, 27, fig. 18 browni ...... 46 castellarina-...... 46 H tricuspis...... 25, 26, 27, fig. 17 vireti...... 25 fodiens ...... 38,46, fig. 31 Heel of -Ma-.-- 6 intermedius...... 46 Heliosorex. See Crocidura. sp.. _____ ...... 25 lixus, Suncus...... 16 newtoni...... 46 Hesperosorex ______. 52,55,63 soriculoides-...... 36 lovei...... 55 lovei, Hesperosorex.--..-...... 55 Heterosorex.- ...... 10,12,58 lunaris, Sylvisorei ...... 23, fig. 15 Neosorex. See Sorex. compressus ...... 11 lacustris---...... 32 delphinensis ...... 14 M Nesiotites...... 61,52 neumayrianus. .. ______11, 12 M' and P*, structure. ______7 corsicanus-...... 51 roperi ...... 13 magna, Cryptotis...... 40 hidalgo...... -~~ 51 sansaniensis...... ______14 Mallorca.-- __ ...... _ . 51 similis...... 51 sp.-- . ... _ _..-.. ___ .--...... 13 Mamblarinaus. See Blarina. Nesophontes..- .- .---- 59 Heterosoricinae ______2, margaritodon, Drepanosorex _ _ . ______33 neumayrianus, Heterosorex...... 11,12 Sorer 12 3, 5, 6, 7, 8, 15, 20, 45, 58, 59, 60, 62, 63 mariae, Blarinoides...... 42,44 hibbardi, Sorex...... 35 Trimylus ...... 7,11,12,13 hidalgo, Nesiotites. ______51 Maries...... ______57 Neurotrichus...... - . .-.. 5 himalayanus, Crossopus ...... 53 martini, Mydecodon...... 5 Nevada---- .. - 25,34 himalayica, Chimarrogale ...... 53, fig. 38 Masseteric fossa.- _ --- - 9, 58, fig. 2 New World distribution, Blarinini-. 62 Hipparion ...... 54 mawbyi, Trimylus...... 12, 13, 14, fig. 7 Limnoecinae------__---_----- 60 Holarctica ______63 meadensis, Cryptotis...... 40 newtoni, Neomys...... 46 Homalurus. See Sorex. megalura, Sylvisorex...... 22 nigrescens, Cosira...... 52 hoyi, Microsorex ...... 30,31, fig. 21 Megasorex...... 53,55,56,63 Soriculus...... 53, fig. 37 Sorex ...... 33 gigas...... 56, fig. 41 niobrarensis, Limnoecus...... 5,6,23,24,26,27,fig. 18 hungarica, Petenyia...... 34, fig. 23 Mental foramen, position... ____ . __ __ 2,6 norae, Surdisorex...... 22, fig. 14 Hungary-.. ______..... 32,33,44,47,48,54 merriami, Sorex ...... 27 North America _ _ - 2,60,62,63 Hydrogale. See Neomys. Metaconid, closeness on Mi to protoconid. . - . 7 Notiosorex...... 37,48,49,50,53,55,56,63 Hydrosorex. See Neomys. mexicana, Cryptotis-...... 41 crawfordi--.- 49, 55, fig. 40 Hypothetical Soricid ancestor ______58 Mexico... . 41,56 gigas...... 56 hj/psitria, Chodsigoa ...... 49, 50, fig. 33 meyeri, Oligosorex...... ___ . _ -... 30,58 jacksoni... 48,49,55 hypsifnus, Boriculus ...... 48 micromorphus, "Limnoecus"...... 23,24 (Notiosorei) crawfordi, Sorex...... 55 INDEX 73

0 Page Rhinomus. See Crocidura. Page Sorex Continued obtusa, Crocidura.--..___ . 17 robinsoni, Myosorex-..______20 stehlini---- -.--...... 36 Oklahoma __ -__ . 44 roperi, Heterosorex----...... 13 subaraneus...... __. 32 Old World distribution, Crocidurinae. 1 Trimylus ...... 11,12,13,14,60 subminutus.______. . ... 32 Neomyini..__.___.. . 62 runtonensis, Sorea...______32 talpoides...... - .-... 42 Old World fossil shrews...___ ... 2 russula, Crocidura---. __._..----______.. fig. 8 taylori.- ______...... 32 ctigodon, Amblycoptus...... 54,55 varius ...... 19 Oligosarex ...... 30,58 vireti- ...... 25 meyeri...... ____- ..... 30,58 sacer, Suncus--- ______21 vvlgaris-...... __. .- -.-... 28 (Drepanosorex) tasnadii- ...... 32 Oregon 10,13,15,25,26,34,37,38,40,41 salenskii, Chodsigoa...... 49,50,55,56, fig. 34 Otiosorex. See Sorei. (Notiosorex) crawfordi...... 55 Oi-yrhin. See Sorer. sandersi, Sorex.-- ...... 32 Soricella.. ______. .. ..-... SI P sansaniensis, Heterosorex... ______14 discrepans.-. -______15,21 Pt, structure..... --...... 2,3, fig. l Sorex...... 10,14 Soricidae...___..__ 1,2,3,5,7,16,45,48,58,59,63 P*and M1, structure..______.__ 7 Trimylus-...... -- 10,13,14 incertae sedis.______57 Pachyura. [See Suncus.]...... 21 Sardinia . 51 Soricidus. See Sorex. pannonica.------...... 47 Saturninia...... ___ 4,58 Soricinae.-.______1, pachyodon, Drepanosorex...... 33 gracilis...... ______4 2,3,5,6,7,24,27,28,29,30,38,45.59,60, Sorer.___.___---.__.--. .-__. 33 savini, Drepanosorex-...... ______. 33 62,63. Palearctica.- . .__. . . .. 2,62 schlosseri, Trimylus...... ______10,14 ancestral characteristics_...______62 Paleozoogeography. -.____...... 59 schoutedeni, Paracrocidura______20 Soricini __-.-.- ______29, 33,36,37,38,39,44,45,55,57,58,59,62, pannonica, Pachyura-...... -... 47 Scutisorex...... 1,15,2. Paracrocidura...__._____.... --.-. 20 63. schoutedeni______...__ 20 Shikamainosorex...______42,44,63 incertae sedis.______35 Paracryptotis 37,39,4*, 42,43,44,62,63 densicingulata...... 42 Soriciscus. See Cryptotis. rez . 41,42,flg. 28 similis, Nesiotites- ...... __ 51 Soricoidea..______59 Paradomnina...... 3, 10, 60 simplicidens, Blarina...... 44 soriculoides, Neomys...... 36 relictus...... 10 sinensis, Beremendia...... __ 51 Zelceina...... 37 Paradoxodon. See Suncus. Solisorex...... 2,2^,63 Soriculus...... 36,47,48,50,51,52,53,54 Parahippus...... 9, pearsoni-...... _____ 21 gibberodon..-...... 48 parva, Cryptotis...... 38, 39, 40, 41, fig. 27 somereni, Sylvisorex...... 21 hypsibius.-...... 48 pareus, Sorex...... 39 Sorea; 2,15,22,25,26,27,29,30, SI, 32, kubinyi...... 48 Paurodus. See Crocidura. 33,34,35,36,37,38,39,42,43,44,45,46, nigrescens...... 53, fig. 37 pearsoni, Solisorex...... 21 47,48,51,52,55,58,62. SouthAfrica...... 17,18,19,20,22 Peisorex...... ------45 alpinoides...... 32 South America... ______63 pohaiensis...... 44,45 alpinus...... 24 South Dakota...____.._____....__. 9,12 Pennsylvania_....______.... _.. 32,44 ambiguus...... 30 Spain...... 14 perminutus, Sorex...... 32 antiquus...... 29,30 s, Anourosorex...... 53,54,flg. 39 Petenyia...... __...... 34,35,36,37,62 araneoides...... __ 32 Stability of morphologic characteristics. ___ 3 hungarica...... ___.. 34,flg. 23 araneus.-...... 26,28,29,31,fig. 20 stehlini, Petenyia...... 34 neglecta ...... 34 caudatus ...... 47 Sorez.____...__._ ... 36 stehlini-...... 34 cinereus...... 27,32 stenodus, Allosorex...... - 6,57 suavensis...... 34 collongensis. ______...__ 24 suavensis, Petenyia...... 34 PmtmyieOa...... 46,47,58 compressus.. ______30,31 gracilits...... 46,47 cudahyensis.-. ....-...-...... 32 suaveolens, Crocidura______.. .._ 17 zelcea...... _ 46,47 daubentonii...... ______46 subaraneus, Sorex...... 32 Pigmentation_.______.--.-__ 5 dekmi...... 15,16,24 Subfamilies.. - _ 2 Pinalia. See Neomys. dehndi...... 42,44,63 origins...... _ . 63 piveteaui, Crocidosorez...... _ .. 2,29,30 dixonensis...... 32 subminutus, Sorex-... __._.. . 32 Plerodus. See Suncus. faUax...... 32 Suncus...... 6,16,19, SI, 22,23,46,47,62 Plesiosorex______...... _.. 58 feroculus- ...... 18 caeruleus...... 20,22,fig. 13 Podihik hum...... 57 frankstownensis.______.__ 32 etruscus ...... 22 pohaiensis, Peisorex. ______....._.. 44,45 gracilidens...... ___ 57 lixus ------.--...---. ------16 Poland...... 32,36,37,42,44,45,46,47,48 gracilis...... -... 46 sacer...... 21 polli, Myosorex.---. ______.. _.. 19 grivensis-...... ___ 19 sp...... 22 Posterointernal ramal fossa. .-.---. .._ 2,5 hibbardi-...... 35 Sunkus. See Suncus. praelpinus, Sorex... ______.__.. 32 lioyi...... 33 Superior pterygoid fossa..___.- figs. 11,29,31 hennardi-...... -..-...... 32 praearaneus, Sorex...... ______32 Superior sigmoid notch.__ fig- 8 kinkelini...... 7, 12 Suprageneric taxa, characteristics.- -. 2 Praesorex...... 16,17,20,21 kretzoii...... 35,42,44,45,63 goliath...... 20,21, fig. 12 Surdisorex- .-.. 22 lacustris...... 32 norae...... 22, fig. 14 pratensis, Microsorex...... 33 leahyi...... 32 Protoconid, closeness on Mi to metaconid.... 7 leucodon-...... _ __ 16 Switzerland... . 11 Protosorex.------...... 58 merriami---- ...... 27 Sylvisorex...... - 15,18,20,22 lunaris.-.- . --- 23, fig. 15 Pterygoid spicule (or spine)__ 28,37,45, figs. 33,34 minutissimus...... 32 megalura.------22 pulchellum, Diplomesodon...... 18, fig. 9 minutus...... ____ 32 somereni.... 21 pulchellus, Sorex...... 17 neumayrianus...... 12 pachyodon.--...... 33 pusiUiformis, Sorex...... 15 parvus ...... 39 pusillus, Sorex______._ 15,30 perminutus...... ____ 32 Talonid of Ms.. 6 grivensis, Sorex______. ____.. 18,19 praealpinus...... 32 Talpidae.--.. . -... ------59 Pygmura. See Anourosorex. praearaneus...... -.-..__ 32 talpoides, Sorex..-. 42 pulchellus...... 17 Talposorex. See Blarina. Q pusittifofmis-...... -... 15 tasnadii, Drepanosorex- ._ .. ------33 quadraticauda, Blarinella______34,35, fig. 24 pusillus...... 15,30 Sorex (Drepanosorex)...... 32 Sorex. ______34 grivensis-. ______...... 18,19 taungsensis, Crocidura------17 quadraticauda______..._ 34 taylori, Sorex. . 32 R rexroadensis...______...... 32 relictus, Paradomnina.-.______10 runtonensis______...__. 32 Tethys seaway.. S9, 62 rex, Paracryptotis.---..... _____. 41,42,fig. 28 sandersi...... ______32 thompsoni, Domnina...... - 5,8,9,10.12,13, fig. 2 rexroadensis, Sorex. ______32 sansaniensis-.-.______. .-.-. 10,14 tricuspis, Limnoecus. - - 25,26,27,flg. 17 74 INDEX

Page Page W Page Trimylus . . . . - 3,5,6,7,8, 10, ucrainica, Beremendia. 51 wardi, Blarinella_.______. 45 11,12,13,14,15,42,45,58,60,63 Uganda.______23 West Indies..______59 compressus______11,12,13, fig. 5 U.8.8.R...... 18 wongi, Crocidura______17 dakotensis. _..------..___ 12,13,fig. 6 Wyoming...... 8,25,55 mawbyi. . 12,13,14,fig. 7 neumayrianus __.---__------7,11,12,13 Z roperi . - .-. .---- 11,12,13,14,60 Variation of rates of evolution...... __... _ 59 zelcea, Petenyietta______46,47 varius, Myosorex...... _.. 19,fig. 11 sansaniensis--...... 10,13,14 Zelceina. ___...______.. 36, 37 Sorex...... _...... _._------_.- 19 schlosseri....______.______... 10,14 vireti, Angustidens--...... 25, 26, 30, fig. 16 soriculoides ...... ______37 habits. --.-. ------14 Hmnoecus______25 zinki, Myosorex Marina... ____...._.__ 22 two lineages. ______10 Sarex...... 25 zorzii, Crocidura----- ______. 17 tumididens, Ingentisorex...... 15 vulgaris, Sorex-.- ____.______28 Zygomatic process..______7,9,29,37,45,58

U.S. GOVERNMENT PRINTING OFFICE : 1967 O MO-SOS